United States
Protection Agency
Office of Solid Waste   Office of Air     Off ice of Research   EPA/530-SW-87-02lf
and Emergency Response  and Radiation    and Development   June 1987
Washington, DC 20460   Washington, DC 20460 Washington, DC 20460
Municipal Waste
Combustion Study

Sampling and Analysis of
Municipal Waste Combustors

This document has been reviewed and approved for
publication by che Office of Research and Development,
U.S. Environmental Protection Agency.  Approval does
not signify that the contents necessarily reflect the
views of the Agency.  Nor does mention of trade names
or commercial products constitute endorsement or
recommendation for use.

                                TABLE OF CONTENTS


List of Tables                                                              v

List of Figures                                                            vi

I.        INTRODUCTION                                                      1

          A.   OVERVIEW                                                     1

          B.   PURPOSE OF THIS DOCUMENT                                     2

          C.   SCOPE                                                        2

          D.   USE OF REPORT                                                4

II.       SAMPLING AND ANALYSIS STRATEGIES                                  5

          A.   PURPOSE OF DATA COLLECTION                                   5

          B.   SELECTION OF SAMPLING AND ANALYSIS METHODS                   9

          C.   QA AND QC OVERVIEW                                           9

III.      SAMPLING                                                         11

          A.   OVERVIEW                                                    11

          B.   SAMPLING METHODS FOR STACK GAS EMISSIONS                    16

          C.   SAMPLING METHODS FOR FLUE GAS                               26


          E.   SAMPLING METHODS FOR WASTE FEED                             31

IV.       SAMPLE PREPARATION PROCEDURES                                    33

          A.   OVERVIEW                                                    33

          B.   REPRESENTATIVE ALIQUOT FROM FIELD SAMPLES                   33

          C.   RECOVERY METHODS                                            38

          D.   EXTRACTION METHOD                                           39

          E.   DRYING AND CONCENTRATING OF EXTRACTS                        40

          F.   SAMPLE CLEAN-UP                                             40

          G.   DIGESTION                                                   41

V.        ANALYSIS PROCEDURES                                              42

          A.   OVERVIEW                                                    42

                                 TABLE OF CONTENTS ,CONT.)


          B.   PROXIMATE ANALYSIS                                           43

          C.   SCREENING ANALYSIS                                           44

          D.   DIRECTED ANALYSIS                                            45

VI.       CONTINUOUS MONITORING METHODS                                     47

          A.   OVERVIEW                                                     47

          B.   SAMPLE CONDITIONING                                          47

          C.   MONITORS FOR INORGANICS                                      49

          D.   MONITORS FOR ORGANICS                                        60

          E.   INDICATOR OR SURROGATE MONITORING                            61

          F.   SPECIAL QA/QC CONSIDERATIONS                                 61

          G.   POTENTIAL FOR PROCESS CONTROL                                63

VII.      QUALITY ASSURANCE AND QUALITY CONTROL                             64

          A.   OVERVIEW                                                     65

          B.   TITLE PAGE AND TABLE OF CONTENTS                             65

          C.   PROJECT DESCRIPTION                                          65

          D.   PROJECT ORGANIZATION AND RESPONSIBILITY                      67

          E.   QUALITY ASSURANCE OBJECTIVES                                 67

          F.   SAMPLING PROCEDURES                                          67

          G.   SAMPLE CUSTODY                                               70

          H.   CALIBRATION PROCEDURES AND FREQUENCY                         71

          I.   ANALYTICAL PROCEDURES                                        72

          J.   DATA REDUCTION,  VALIDATION, AND REPORTING                    73

          K.   INTERNAL QUALITY CONTROL CHECKS                              73

          L.   PERFORMANCE AND SYSTEM AUDITS                                75

                                 TABLE OF CONTENTS  (CONT.)

          M.   PREVENTIVE MAINTENANCE                                       76

               ACCURACY, AND COMPLETENESS                                   76

          0.   CORRECTIVE ACTION                                            77

          P.   QUALITY ASSURANCE REPORTS TO MANAGEMENT                      78

VIII.     REFERENCES                                                        80

APPENDIX A                                                                  84

APPENDIX B                                                                  8q

                                 LIST OF TABLES

Table                                                                      Page


2    STACK SAMPLING METHODS                                                 17


4    SAMPLING METHODS FOR SOLID AN .  _IQUID EFFLUENTS                       ' 30

5    SUMMARY OF SAMPLE PREPARATION METHODS                                  34


     APPLICABLE TO MSW COMBUSTOR SAMPLES                                    46

8    CONTINUOUS ANALYZERS FOR CARBON MONOXIDE                               53

9    CONTINUOUS ANALYZERS FOR CARBON DIOXIDE                                54

10   CONTINUOUS ANALYZERS FOR OXYGEN                                        55

11   CONTINUOUS ANALYZERS FOR SULFUR DIOXIDE                                56

12   CONTINUOUS ANALYZERS FOR NITROGEN OXIDES                               57

13   CONTINUOUS ANALYZERS FOR HYDROCHLORIC ACID                             58

14   CONTINUOUS ANALYZERS FOR HYDROGEN CYANIDE                              59

15   ESSENTIAL ELEMENTS OF A QA PROJECT PLAN                                66

     OBJcuTIVES                                                             69

                                 LIST OF FIGURES

No.                                                                    Page

          SCHEMATIC                                                    12

          SAMPLE LOCATIONS                                             13

          LOCATIONS                                                    14

          SHOWING SAMPLING LOCATIONS                                   15

5         SCHEMATIC OF STANDARD EPA METHOD 5 TRAIN                     22

6         MM5 TRAIN SCHEMATIC DIAGRAM                                  23

7         SASS SCHEMATIC DIAGRAM                                       25




     This document is intended to serve as a resource document outlining
recommended sampling, analysis and monitoring procedures for municipal solid
waste combustion facilities.  The effort reported here has focused on gathering
information on probable measurement requirements and on available methods that
may meet those requirements.  Critical evaluations of alternative methods and
recommendations to fill gaps in available methodology were outside of the scope
of this assignment.

     The Arthur D. Little, Inc. effort is part of a major program to develop an
EPA report to Congress on Municipal Waste Combustion.  The overall effort is
being directed by Radian Corporation (G. Wilkins,  Project Manager).  The Arthur
D. Little contribution is being performed under EPA subcontract with Dynamic
Corporation (C. Matkovich, work assignment director).  L. D. Johnson is the EPA
work assignment director for the Arthur D. Little effort.  Key Arthur D. Little
staff involved in this work are J. C.  Harris (to whom comments should be
addressed),  D. L.  Cerundolo, and K. E.  Thrun.  The Arthur D. Little reference
numbers for this work are 55464 and 55465.

                                 I.  INTRODUCTION


     This report is an assessment of sampling and analysis methods for municipal
waste combustors.  The information presented in this report was developed during
a comprehensive, integrated study of municipal waste combustion.   An overview of
the findings of this study may be found in the Report to Congress on Municipal
Waste Combustion (EPA/530-SW-87-021A).   The Technical volumes issued as part of
the Municipal Waste Combustion Study include:
          Municipal Waste Combustion Study:
          Report to Congress
          Municipal Waste Combustion Study:
          Emissions Data Base for Municipal
          Waste Combustors
          Municipal Waste Combustion Study:
          Combustion Control of Organic
          Municipal Waste Combustion Study:
          Flue Gas Cleaning Technology
          Municipal Waste Combustion Study:
          Costs of Flue Gas Cleaning
          Municipal Waste Combustion Study:
          Sampling and Analysis
          Municipal Waste Combustion Study:
          Assessment of Health Risks
          Associated with Exposure to
          Municipal Waste Combustion Emissions

     •    Municipal Waste Combustion Study:
          Characterization of the Municipal
          Waste Combustion Industry                         EPA/530-SW-87-021H

     •    Municipal Waste Combustion Study:
          Recycling of Solid Waste                          EPA/530-SW-87-021I

     The purpose of this document is to provide guidance on sampling and
analysis methods to assist federal,  state, and local environmental authorities
in reviewing plans for operations and testing of MSW combustors.   The sampling
and analysis procedures outlined here are intended to represent state-of-the-art
methods that may be useful in determining the regulatory compliance status of
MSW incineration facilities and in assessing their environmental  impacts.  These
same methods may be useful in research and development programs related to MSW
combustion technology, standard setting,  etc.


     This document provides an overview of available state-of-the-art methods
for sampling and analysis to address testing and monitoring of MSW combustors.
For purposes of this report, testing and monitoring are defined as follows:

     "testing" means performing periodic  sampling and analysis
     by EPA-approved or recommended  methods to:   confirm compliance with any
     limits that have been imposed by regulatory agencies as permit conditions;
     generate data that may be used  as inputs to environmental risk assessments;
     and/or support research and development in municipal waste combustion.

     "monitoring" means obtaining continuous instrumental measure-
     ments of key process parameters and selected pollutants in the
     emissions to verify that the facility continues to operate "in control".

     The testing can be expected to focus on several categories of potential

     Criteria pollutants:  particulates,  CO, SC-   NO ,
                                               £.    X

     Acid gases:  HC1, HF

     Trace metals:  Cr(III),  Cr(VI),  Cd,  As, Hg, Pb, Be, etc.

     Organic pollutants:  chlorina "   dibenzo-p-dioxins, chlorinated
     dibenzofurans, and other trace level species that may be indicators of
     potential environmental impacts

     The parameters for which continuous monitoring may be required during
routine operation or requested as part of special purpose programs include:

               carbon monoxide
               carbon dioxide
               nitrogen oxides
               sulfur oxides
               hydrochloric acid
               "total hydrocarbon"
               performance "indicator" species (e.g., CO)

     Procedures for measuring the above parameters in ultimate effluents from
MSW combustion facilities (stack gas,  solid residues such as bottom- and
fly-ash, and liquid effluents such as scrubber water) are described in this
report.  Two other aspects of MSW combustor sampling and analysis--measurements
on the MSW feed and measurements on flue  gases upstream of air pollution control
devices--are also addressed,  but procedures appropriate for these media are

less well developed or defined.  In both cases, the principal difficulty is in
obtaining a representative sample.  MSW is highly heterogeneous and some
individual items present are so large or bulky as to preclude effective use of
compositing procedures.  (See Section HIE for further discussion).  Flue gas
sampling (discussed in Section IIIC) is complicated by uneven flow conditions
and by components (e.g., high particulate material and acid gas loadings) that
clog or corrode conventional sampling equipment.

     This document speaks only to issues of MSW combustor source sampling and
analysis.  Procedures for assessing ambient air impacts,  either by dispersion
modelling or by direct ambient air sampling and analysis in the vicinity of the
facility, are not described.  Sampling and analysis of fugitive emissions (e.g.,
from transfer or storage operations) are not covered in this document.


     It is anticipated that this report will serve as a useful reference point
for MSW combustor owner/operators and for federal, state and local authorities
involved in facility permitting process.  The sampling and analysis methods
described and/or recommended here are not to be construed as representing unique
or mandatory requirements for emissions testing or continuous monitoring at MSW
combustion facilities.  Inclusion in this report does not mean that a sampling
or analysis method is an official EPA method.  The procedures described here
should be reviewed against those in the current Code of Federal Regulations and
in official methods manuals, such as "Test Methods for Evaluating Solid Waste"
(SW-846)   ,  before any MSW combustor test program is finalized.



     The principal purpose of generating data from MSW combuster sampling and
analysis is to determine the effects of emissions from a facility on health and
the environment.  These data can be used to determine compliance with applicable
criteria or to perform an assessment of the health risk associated with the
emissions.  In some cases, the purpose of the sampling and analysis may be
primarily to support research and development in MSW combustion technology.  In
those instances, there may be a requirement for some types of measurements
(e.g., characterization of waste feed or in-furnace measurements) that are more
extensive than those needed for compliance testing or risk assessment.

     Table 1 indicates some examples of the types of compounds for which
measurements have been made in various test programs.    Most of these are not
regulated pollutants in stack gas emissions, and inclusion in the Table is not
meant to imply that measurement of these potential pollutants should be
required.  However, this document presents methods that would be applicable to
the sampling and analysis of these types of pollutants if such measurement were
determined to be necessary or useful.

     In every case, it is essential that critical decision limits for each
parameter be established in specific quantitative terms prior to selection of
sampling and analysis methods.  This may require, in many cases, that a
preliminary air quality modelling run be performed in order to calculate what
stack gas concentration corresponds to a specific ambient air quality criterion.
This exercise allows calculation of the sampling and analysis method detection
limit that will be necessary to meet the project objectives.

     The precision and accuracy criteria for the measurement method will also
depend on the specific uses to be made of the data.  It is important to
establish whether it is each of n measurements,  or the mean of n measurements,

                                     TABLE 1

Carbon Tetrachloride
Semivolatiles:  Benzo(a)pyrene
               Polychlorinated Biphenyls
                              Others:    Hydrogen Chloride
                                        Hydrogen Fluoride
                                        Sulfur Dioxide
Source:  Arthur D. Little, Inc., conversations with regulatory authorities.
    /L Arthur D. Little, Inc.

 or  the  95%  confidencelimit  around  the mean  that  is  to be compared with  the
 critical decision  limit.

      The general purposes  of collecting data by testing and monitoring of MSW
 combustors  are  discussed below.  The  specific purposes (and thus the critical
 decision limits) must  be established  on a case-by-case basis.

      Criteria Pollutants.  The criteria pollutants (particulate material, carbon
 monoxide, nitrogen oxides, ozone, and sulfur dioxide) are those species  for
 which. EPA has established  primary and/or secondary National Ambient Air  Quality
 Standards (NAAQS).  As the name  implies, these standards apply to the cumulative
 impact  from all sources, not the source-by-source emissions, on ambient  air
 concentrations.  However,  it may be necessary to measure emissions of criteria
 pollutants  from an MSW combustor in order to assess  its incremental contribution
 to  the  total ambient level.  This is  especially likely in the case of new MSW
 facilities  in areas where  current levels are close to or exceed the NAAQS. The
 criteria pollutants are also of concern with regard  to Prevention of Significant
 Deterioration (PSD) determinations.

     Acid Gases.   In addition to SO   (discussed above), hydrochloric acid and a
 lesser  quantity of hydrogen  fluoride  are acid gases  that can be expected to be
 present in MSW combustor off-gases.   Although there  are not federal EPA
 standards relating to HC1  emissions from facilities  other than hazardous waste
 incinerators, many state environmental agencies (e.g., Maine, Massachusetts, New
 Jersey, California) consider that acid gas removal from MSW incinerator
 effluents represents Best  Available Control Technology (BACT).  Testing  of stack
 emissions of HC1 and HF may  be required to demonstrate that a degree of  acid gas
 control consistent with BACT has been achieved.

     Trace Metals.   Trace  metal species, in addition to lead, that may need to
be measured in MSW incinerator effluents include:   arsenic, beryllium, cadmium,
 chromium, and mercury.  In general, air emissions of trace metals from MSW
 sources are not currently  addressed in EPA regulations, but may be of concern
with regard to potential health effects.  In some cases,  it may be necessary to
differentiate Cr (III) from  Cr (VI) in order to assess the magnitude of  health

effects.  The metals content of MSW incinerator aqueous effluents (e.g.,
scrubber or quench water) and of solid residues (e.g., bottom ash, fly ash) may
need to be ascertained in order to determine disposal status vis-a-vis NPDES or
RCRA regulations, respectively.

     Organic Pollutants.  Data on chlorinated dibenzo-p-dioxins and
dibenzofurans in stack gas effluents may be required as inputs to air quality
models in order to estimate whether the risk of exposure to these chemicals from
this source is within the range of acceptability.   Permitting authorities may
 *                                                                         ,
request data on other trace organi -  such as polynuclear aromatic hydrocarbons
or phenols, in order to perform similar risk calculations.  Again, NPDES or RCRA
regulations may require the determination of specific organic chemicals in
aqueous effluents or solid residues, depending on "he disposal or treatment
alternatives that are to be applied to these streams.  Selected organics may
also be measured during MSW combustion research and development as indicators of
the efficiency of the combustion process and/or pollution control devices.

     Monitoring.   Continuous monitoring of MSW combustor emissions provides a
mechanism for tracking the performance of the system in real time.  The
parameters that can be monitored continuously are generally those (e.g.,
temperature,  oxygen, carbon monoxide) that indicate the overall combustion
efficiency or those (e.g., opacity, SO ) that indicate air pollution control
device (APCD) performance.  These measurements are important for confirming that
a facility is continuing to operate under controlled, steady-state conditions
and is in compliance with specific emissions limitations written into its
operating permits.   In addition, the on-line monitor data can serve as an "early
warning system" to detect combustor/APCD system upsets due to mechanical
failures and/or gross changes in waste composition (e.g., very wet waste).
Thus,  even if an on-line monitor is not sufficiently accurate or precise for
determining compliance with emissions limitations (e.g., commercially available
opacity meters do not reliably determine particulate matter emissions below 0.03
gr/dscf),  data from such an instrument is nevertheless useful as an indicator of
potential upset conditions.


     Once the uses of the data have been defined and the consequent data quality
objectives have been established, it is possible to make an informed selection
among alternative sampling and analysis (S/A) methods.

     The principal criteria to be used in making this selection are the

     •    regulatory status.  Is the S/A method chosen approved by EPA for this
          measurement purpose?

     •    sensitivity.  Has the method been proven to have a detection limit
          that is sufficiently low to allow accurate quantification of the
          pollutant of interest at concentrations corresponding to the critical
          decision limit?

     •    selectivity.  Will other species that are likely to be present in MSW
          combustor effluents interfere in the determination?

     •    reliability.  Is the method sufficiently rugged to be applicable in
          the hostile  nvironment represented by MSW incineration?

     It is important to emphasize that these criteria must be applied to the
overall sampling--sample preparation--analysis system,  not just to the analysis
method per se.


     Quality Assurance (QA)  and quality control (QC) are vitally important
components of any sampling and analyses program.  QC procedures, including the
analysis of standards, blanks, replicate samples,  and spikes, provide on-going
confirmation that the sampling and analysis methods are "in control" and that

the data generated are valid for the intended purpose.  QA procedures ensure
that all data, including QC data, are reviewed and that any necessary corrective
actions are instituted in a timely fashion.

     In order to ensure the effective implementation of QA/QC functions, it is
necessary that specific quantitative Data Quality Objectives (DQOs) be
established for each measurement.  Guidance in developing a QA/QC plan for MSW
combustion testing/monitoring is discussed further in Section VII of this report
and in the EPA report, "Interim Guidelines and Specifications for Preparing
Quality Assurance Project Plans."

                                 III.  SAMPLING


     Sampling of MSW combustion facilities for purposes of determining
compliance with regulatory limitations or of assessing potential environmental
risks is usually focused primarily on air emissions (stack gas).  Solid residues
(bottom ash and fly ash) are also commonly sampled.  Some facilities may also
have aqueous effluents from wet scrubbing or ash quenching operations.

     Figure 1 shows a generalized schematic of the MSW combustion process, with
sampling points indicated generically,  Figures 2, 3,  and 4 present,
respectively, schematics of a mass burn, refuse-derived-fuel (RDF),  and starved
air MSW incinerator, again with generic identification of sampling locations.
Exact sampling points must be specified on a case-by-case basis for each
facility, taking into account accessibility, temperature and flow conditions,
and operational modes (e.g., batch or continuous discharge of residue).

     The critical considerations in selection of a sampling method are:   (1) the
representativeness of the sample and  (2) compatibility of the sampling procedure
with the intended analytical finish.

     To ensure representativeness, emissions samples are usually integrated over
time.  A 1-4 hour time-integrated sample is usually taken to smooth out the
perturbations that occur due to the inherent variability of MSW.  In the case of
stack sampling, it is often possible  to accomplish this directly by adjusting
the sampling rate to cover the desired time period.  In the case of liquid or
solid residues from MSW combustors, the time integration is usually accomplished
by preparing a composite sample, combining equal-sized aliquots collected at
intervals over the duration of a test.  Sampling of solid waste, if requested in
a particular facility test,  is usually done by compositing of subsamples,  in
accordance with procedures described  in SW-846.  In all cases,  the
representativeness of the sampling method can be checked, but not proven,  by
acquiring duplicate samples representing the same time period of facility

                               :*. Plant  *»
                               * Trash  ••«
                                                               J	I
                                     Flue Gas
                                               Stack Gas
                                             Treatment Train
                       Ash Pile
                                                 Fly «id Bottom
                                                 Ash Samples *
                                                                        Bottom Ash *
                                                    Fly Ash *
                                Ash Transfer Hopper
                                                          *. To Landfill
                                 FIGURE  1
                                           (Adapted from original by  S-Cubed,  a Division of
                                            Maxwell Laboratories, Inc.)
                      *In many  facilities,  fly ash and bottom  ash  are collected  in  a single  hopper.

                            Samples *
                                                                                             Samples *
              (Adapted from  original by C.  R.  Brunner, "Incineration Systems Seminar,"  (198?))
                       tiff* - f 1

                                                                                                               Fabric Filter
                                                 Heavy Ends
                                 Fluidizing Blower   to Disposal

                                      (Adapted from  original by  C. R.  Brunner, "Incineration

                                            •Stack Emission Samples
         (Adapted from original by C.  R. Brunner,  "Incineration Systems Seminar,"  (1982))

     Compatibility is a complex issue.  Most analytical methods (e.g., GC/MS for
trace organics,  AAS for trace metals) cannot be applied directly to the sample
but require extensive sample preparation.  Even methods that are inherently
"on-line" procedures (e.g., continuous instrumental monitoring of CO or SO ) may
require sample clean-up to remove particulate material, water vapor, etc.  or
other forms of sample conditioning prior to analysis.  It is important that the
QAPP (quality assurance project plan) for any MSW combustion sampling and
analysis program contain procedures to assure that no unacceptable losses are
introduced by whatever procedures are required to ensure compatibility.


     Table 2 lists EPA-approved and/or recommended state-of-the-art stack
sampling methods for various pollutants and/or categories of pollutants. (Note:
Table 2 includes only extractive sampling methods.  Continuous monitoring
methods are discussed in Section VI.)  Most of these procedures have been
subjected to extensive method development, validation and/or round-robin
collaborative testing.   Although they may not all have been validated at MSW
combustion facilities specifically, there is every reason to believe that they
will be directly applicable for the purpose of sampling MSW stack emissions.

  In fact, a number of these methods have been employed in recent MSW combustion
                              (3 4)
sampling and analysis programs  '   and appear to provide reliable data. For
example, Table 3 shows the results obtained in a series of 6 replicate stack gas
emission tests for polychlorinated dioxins and furans at one MSW combustor using
the ASTM sampling and analysis protocol    (see below and Appendix B for
description).(6) These data show the precision (relative standard deviation
(RSD))  that can be achieved when sampling and analysis is carried out according
to the recommended methods by experienced personnel  (in this case, Radian
Corporation for sampling and Triangle Laboratories for analysis).  The RSD
values in the table reflect the combined effects of actual variations in
emissions, sampling variance, and sample preparation/analysis variability for
this test series.  RSD's could be expected to be somewhat higher at lower
PCDD/PCDF emission levels.  However, the data indicate that the state-of-the-art
sampling and analysis methods are workable in experienced hands.  It should be

                                                  TABLE 2

                                          STACK SAMPLING METHODS

Criteria and Conven-
tional Parameters
Isokinetie collection of a 1 hr. sample
on glass fiber filter at 120+ 14°C.
Train includes:  T-controlled probe,
optional cyclones,  heated filter,
impingers,  flow control and gas volume
metering system.

Visual determination of opacity

Instrumental measurement of opacity
(optical density)

Collection in isopropanol (SO,.) and
hydrogen peroxide (S0?) impingers of
M5-type train.

Integrated gas bag or direct interface
via air-cooled condenser.
Collection in evacuated flask
containing sulfuric acid and hydrogen
Designed to meet 0.08 gr/SCF
standard.  Probably adequate
down to 0.01 gr/SCF, especially
if sampling period is increased
to 2 hrs.
 EPA Meth.  5
Not reliable for quantifica-
tion at 0.03 gr/SCF or below.

Low ppm to percent
Water vapor, carbon dioxide
are interferences; need
silica gel, ascarite traps to
20-1000 ppm

Grab sample (not
ppm levels
Does not differentiate
NO from NO,,
                                                                                                EPA Meth. 9
EPA Meth. 6,8
EPA Meth. 10
EPA Meth. 7.7A

                                                   TABLE 2 (cont.)

                                          STACK (FLUE GAS) SAMPLING METHODS

Collection in aqueous NaOH impingers in
M5-type train.

Collection on paper or membrane (not
glass fiber) filter and aqueous
impingers in M5-type train.
ppm to percent range
Low ppm range.


EPA Meth. 13B
     Trace Metals


M5 or SASS train, glass fiber filter
and nitric acid or ammonium persulfate

Collection on glass fiber filter and
nitric acid impingers in M5-type train.
ppb to ppm levels if
is  A 0.75 M
of stack gas
                                 EPA Meth. 12
Collection in iodine monochloride or
acidic permanganate impingers in
M5-type train.

Collection on glass fiber filter and
aqueous impingers in M5-type train.

Collection on millipore AA filter and
aqueous impingers in M5-type train.
Probe must be glass- or
ppm levels. Other reagents
also possible.
Probe must be glass or
EPA Meth. 101
                                 EPA Meth. 108
EPA Meth. 104

                                              TABLE 2  (cont.)

                                     STACK (FLUE GAS)  SAMPLING METHODS
Trace Organics

Specific Volatile
Collection on Tenax-GC
1 LPM for 20 minutes.
and charcoal at
Semi-volatile       M5 train modified to include XAD-2 trap
organics, including for organic collection between filter
dioxins, furans     and irapingers.

                    5-fold scale up of MM5 system.
ppb-ppm levels;multiple

ppb-ppm levels; multiple
                                          sub-ppb levels for
                                          dioxins/furans if dedicated
Vinyl chloride*

Integrated gas bag
                  0.1-50 ppm
Collection on DNPH-coated sorbent or in   ppn levels
aqueous DNPH impingers.
                                 EPA Meth.  106

Gaseous Hydro-
carbons, total

Gaseous Hydro-
carbons, total
Integrated bag sample or direct

Evacuated stainless steel or aluminum
tank behind chilled condensate trap.
                  ppm levels
                                 EPA Meth.  18
                                                   EPA Meth.  25
* VOST can also be used for vinyl chloride, but collection and recovery efficiency may be low.

                                         TABLE 3
                         Data from 1986 Test as Saugus Resource  Recovery Plant
                         STACK GAS (Corrected to 12% C02)
                                  Avg cone (n = 6 HM5 runs)

















                         Total         579      126

                         GRAND TOTALS, DIOXINS + FURANS



/fc. Arthur D. Little. Inc.

recognized that these procedures require an exceptionally high level of quality
control to ensure that adequate recoveries and detection limits have been
achieved in each instance.

     The Method 5 (M5) type train represents the principal method for sampling
of criteria pollutants, acid gases, metals, and semi-volatile organics.   The
basic Method 5 train, shown schematically in Figure 5, includes a probe with
buttonhook nozzle and pitot tube, a filter section, an impinger train and a
metering system.  The probe and filter sections are maintained at a temperature
of 250 F; material recovered from these train components and dried to constant
weight is defined as particulate material.  For quantification or chemical
analysis of particulate material, the sampling must be done isokinetically
(i.e., gas velocity into the probe equals the gas velocity within the stack).

     As noted in Table 2,  the M5 train is used with minor modifications to
determine many inorganic species, such as sulfur oxides or trace metals.  These
modifications do not involve substantial changes in train geometry but relate to
the use of:  glass-  or quartz probe liners; special filter media; or selective
reagents in the M5 impingers.  For example, use of a glass lined probe,  glass
fiber filter and 0.1 N nitric acid in the impingers of a standard M5 train
provides good collection efficiency for most trace metals.  To quantify heavy
metals in the stack emissions, the probe catch, filter and/or impinger solutions
are digested as described in Section IV G and analyzed as described in Section V

     For sampling of semivolatile organics in stack emissions, a more
significant modification of the standard M5 train is required.  The Modified
Method 5 (MM5) train (SW-846 method 0010), shown in Figure 6, incorporates a
condenser/cooler section and a module filled with a solid adsorbent between the
exit of the filter and the entrance to the first impinger.  The adsorbent
                                                    (' ~j 8 ^
recommended is Amberlite XAD-2, which has been shown  '   to give good
collection efficiency for vapor phase organic compounds with boiling points
above about 100°C.  This category, "semivolatile" organics, includes many
species that are of potential interest for risk assessment (Table 1), including
chlorobenzenes,  chlorophenols, polycyclic aromatic hydrocarbons, and chlorinated

                                     IMPINGER TRAIN OPTIONAL, MAY BE REPLACED
                                          BY AN EQUIVALENT CONDENSER

                                     FILTER HOLDER

                     DRY CAS METER

                    Filter Holder
                      Stack Wall
     Thermocouple  |J       probe

"S" Type
  Piwt             FT        _—-y
                     Ma4»o meter
Recirculation Pump
                                  Water Jacketed Condenser

                                Sorbent Trap
                                                                  Check Valve
* L
-J *

•**• In.

                                                        I mpinger

                                                         By-Pass Valve
                  Dry Gas Meter   Air-Tight
                                                                                    Vacuum Line

dibenzodioxins/furans. The MM5 train is recommended for sampling of
dioxins/furans in MSW combustion effluents in the joint ASME/EPA Environmental
Standards Workshop draft Protocol.    (See Appendix B.)  The MM5 approach has
also been used by other agencies involved in sampling for dioxins/furans,
including the New York State Department of Environmental Conservation (9) and
Environment Canada (10).   There are some differences in configuration between
the MM5 train designs currently in use; these may produce somewhat different
distribution of collected pollutants across the train components (i.e.,
different fractional collection on filter vs. sorbent vs. condensate).   However,
all contain the same basic components and should provide comparable overall
collection efficiencies.

     The MM5 train, like the M5 train, is used to sample isokinetically.  Also,
as with the MS, a variety of impinger reagents can be used for selective
collection of species such as trace metals (Table 2).

     The Source Assessment Sampling System (SASS) (SW-846 method 0020)  is an
alternative to the MM5 train.  The SASS,  shown schematically in Figure 7, is
essentially a fivefold scale-up of the MM5 system.   Use of SASS is recommended
when calculations indicate that the sampling flow rate of the M5/MM5 train
(0.75-1.0 cu.  ft. per min.) would not collect a sufficient quantity of pollutant
in a reasonable time period to meet data quality objectives for detection
limits.  Note that the design criteria for the SASS train cyclones are based on
a constant sampling rate,  not on a variable rate as used in true isokinetic
sampling.   The SASS may therefore be most suitable for sampling semi-volatile
organics and trace metals that are present in the stack as vapor phase
materials.  The variation from isokineticity may introduce errors when the SASS
is used to collect particulate material;  however, the variations and any
resultant errors are generally small compared to other sources of variance in an
MSW combustor sampling and analysis program.

     The operation of both the MM5 and the SASS train is described in reference
11.  Note that each of these stack sampling trains generates multiple components
for subsequent sample preparation and analysis (Sections IV and V).  Note also
that,  while the M5-type approach can be modified to meet a number of sampling


                                                                          BALL VALVE
                                                                                                 GAS COOLM
                                                                                       IMC/COO ll«
                                                                                       ritACC CICMENT
                                     MY GAS MJ!£ It/Oil IF ICf METER
                                      CENTRALIZED TIMPtltATURE
                                       AND mSSURI «ADOUT
                                          CONTROL MODULI
                             	TWO Itth3/nwn VACUUM PUMPS


IERL-RTP Procedures  Manual:  Level 1 Environmental Assessment

     (Second  Edition), El'A-600/7-78-201

objectives (particulate material, trace metals, semi-volatile organics, acid
gases), it is generally recommended that one not attempt to meet more than one
(or two) of these goals with a single stack sample.  In theory, for example, one
could  run a MM5 train, dry the probe wash and filter to constant weight for
weighing of particulate material, then split the particulate catch--digesting
one-half for trace metal determination and extracting the other half for
semi-volatile organic analysis.  In practice, such an approach is likely to
result in sample losses and in unnecessarily high limits of detection for
particular pollutants.

     The Volatile Organic Sampling Train (VOST) was developed^  ' to sample
organic vapors that are too volatile for efficient collection in the MM5 or
SASS.  This includes most organics with boiling points of about 100 C or lower.
This includes a number of species that may be of interest such as carbon
tetrachloride, benzene and vinyl chloride.   A schematic of the VOST is shown in
Figure 8.  As shown, the VOST system includes a probe, glass wool roughing
filter, a gas cooling/condensing section, and two solid adsorbent tubes in
series.  The first tube contains 1.6g of Tenax-GC   sorbent and the second
contains l.Og of Tenax backed up by  l.Og of charcoal.  The VOST is used to
collect a series of 20L stack gas samples at a flow rate of 1 LPM through 5
successive pairs of fresh traps.  After sampling, the tubes are sealed and
returned to the laboratory for thermal desorption and analysis of volatiles by
GC/MS  (Section V).   A protocol for use of VOST has been published by EPA   ' and
is also available in SW-846 as method 0030(1).

     The VOST system has been extensively validated in the laboratory     and
applied successfully in the field    .   In addition, EPA has sponsored the
development and testing of a series of VOST audit cylinders, containing known
mixtures of volatile organics, that are available to regulatory agencies as
checks on overall VOST sampling and analysis performance by users of the system


     The methods described in Section B, above, are also applicable to sampling
of flue gases upstream of air pollution control devices (APCDs).  However, flue

 Glass Wool
(or test System)
                 Heated Probe
Isolation Valves
         Carbon Filter
                                 ice water
                     Trap Impinger

                                                            Silica Gel
Dry Gas

gas is typically hotter, wetter, dirtier (higher particulate loadings, higher
levels of condensable organics) and more corrosive than stack gas.  Thus
modifications are frequently required to adapt stack gas methods to flue gas
sampling.  For example:

     •    When the flue gas temperature is in excess of about 500 C, a
          water-cooled jacket must be placed around the outside of the heated

     •    The flow conditions wit.    the flue and difficulty in accessing
          sampling points may preclude isokinetic sampling and/or traversing of
          the duct.  A preliminary velocity traverse will usually allow
          selection of a fixed sampling point of "average" velocity that is away
          from the walls of the flue.  The probe can then be located at this
          position and the sampling rate adjusted to be as close to isokinetic
          as practical.

     •    High loadings of particulate material and/or condensable organics may
          cause frequent interruptions of sampling for replacement of filters to
          avoid excessive pressure drops across the train.  Use of a roughing
          filter (in-flue) or insertion of a cyclone component upstream of the
          M5-type filter may be necessary to allow reasonably convenient
          continuity of sampling.

     •    Glass,  quartz, or Teflon liners may be required to protect all
          stainless steel surfaces from corrosion.

     The principal reason for sampling the flue upstream of the APCD is to allow
assessment of control device removal efficiency.  An alternative approach is to
sample and analyze the material collected in the device, rather than the flue
gas challenge concentration, and compare this to the quantity emitted in the
stack gas.  The efficiency can then be estimated as:

        .        collected	      x  100
          "  0          + Q
              collected    emitted

where 0  ,,     ,    -    concentration in fly ash x rate of fly ash collection
                         in APCD

      0             -    concentration in stack gas x volume flow rate of stack
       emitted                                  °

     This "mass balance" is probably comparable in uncertainty to that based on
direct flue gas measurements for "major" pollutants in MSW combustion, such as
particulate material or acid gases.  It cannot be recommended for trace level
pollutants such as individual metals or organics,


     The methods specified in SW-846    and in "Sampling and Analysis for
Hazardous Waste Combustion (First Edition)"     are generally directly
applicable to MSW combustor solid and liquid effluents.  Table 4 summarizes the
relevant methods.  In addition, EPA is currently in the process of developing
specific recommendations for sampling MSW combustor ash.

     For both liquid and solid effluents, the sampling strategy will depend on
whether the stream is generated continuously (as in once-through scrubber water)
or in a batch process (as in a fabric filter).  In the former case, composite
samples are generated by collecting equal-sized aliquots at regular time
intervals over the course of the test run (e.g., from a tap on a discharge line
(liquid) or from a conveyor (solid).)  In the latter case, composite samples are
prepared from subsamples from statistically selected points that represent the
horizontal (area) and vertical (depth) extent of the batch.

     Particular problems may be encountered in sampling MSW combustor bottom
ash,  which may include bulky items such as metal containers.  This issue should
be specifically addressed in the samp\ing and analysis plan for each test.  It
must be recognized that any decision to "sample around" such bulky objects could
compromise the overall validity of the data collected, unless it could be
established that they are negligible sources of the compounds to be determined.

                                                            TABLE 4
                                       SAMPLING METHODS FOR SOLID AND LIQUID EFFLUENTS

     Liquids in pipelines.
     Liquids in sumps,  tanks,
     or open drain.

     Wet or dry ash on
     conveyors or in bins.
     Dry fine ash on conveyors,
     in bins or piles.

Attach TFE line to tap.   Flush
container with fresh sample,  then

Glass or TFE beaker on rod.
Stainless steel trowel or lab
scoop.  Obtain random sample below
surface level.

Tube within a tube, rotated to
align slots for sampling.




     Wet ash in bins or
o    piles.
Section of tube cut in half
lengthwise.  Insert into waste and
      Reference 1

     This stream is grossly inhomogeneous.   None of the sampling techniques
promulgated for hazardous waste sampling CSW-846)    is directly applicable to
MSW as-received at a mass-burn combustion facility.  (Some refuse-derived-fuel
(RDF) processes may produce a waste stream that is sufficiently homogeneous and
finely divided to be amenable to scoop or trier/thief sampling.)

     Researchers have chosen one of two approaches towards sampling MSW:
stratified random sampling at the source (socio-economic demography)     or by
subjectively-modified random sampling at the repository, be it an incinerator or
          / 1 -I TO 1 Q\
a landfill   '   '    .   For purposes of evaluating the performance of an MSW
combustor, or of understanding the relationship between emissions and waste feed
characteristics, the latter is the preferred approach.

Procedures used for sampling and testing of refuse, and the resulting
variability in measurements of properties of MSW, have been summarized by
Hasselriis (20).  Based on this review, it is recommended that 3 to 10 samples
of MSW should be collected over at least a 2 week sampling period.  The smallest
sample should be 1 "unit" of MSW (i.e., one trash bag or barrel of domestic
waste at the source; one truck hopper load or crane bucket load at an MSW
combustor plant.  Gone and quarter procedures can be applied to reduce the unit
to a manageable sample size (80-130 kg); this sample can then be shredded to
achieve a reasonably homogeneous material for analysis. As an example, the MSW
sample processing protocol reported by Bell     is included in Appendix A. If
any element of subjective sampling is imposed (e.g., exclusion of bulky items or
hospital wastes) this must, at a minimum, be explicitly noted and an estimate of
the quantity of rejected material be provided.

     The EPA, Department of Energy(DOE),  and American Society for Testing of
Materials (ASTM) have been collaborating in the development of protocols for RDF
sampling, including sampling accuracy, bias, and reproducibility estimates.  A
draft procedure has been approved by an ASTM subcommittee and is awaiting ballot
approval by the main E-38 Committee on Resource Recovery.  This protocol should
be a useful resource for designing waste sampling plans for MSW combustors in

                            IV.  SAMPLE PREPARATION PROCEDURES


     The sample preparation procedures for use on MSW combustor samples involve
a number of steps.  In the field, the collected samples must be transferred to
appropriate, clean containers  (generally glass or TFE for organic analysis /.nd
high-density linear polyethylene f,/  inorganic analysis) and appropriately
preserved and stored.  In the  laboratory, the sample must be converted (via
digestion, extraction, etc.) into a matrix which is compatible with the final
analysis methods needed.   Table 5-presents a summary of the sample preparation
procedures that will commonly be required for MSW combustor samples.   This table
indicates how samples collected by procedures in Section III are converted to
forms amenable to analysis by procedures in Section V.

     In some cases (e.g., analysis of chloride in caustic impinger solutions)
the sample preparation may be minimal (e.g., diluting an aliquot to a known
volume).  In other cases  (e.g., analysis of dioxins/furans in an MM5 stack gas
sample) the procedures may be complex,  requiring extraction of multiple
components, concentration and clean-up of extracts.

     The use of surrogate or standard addition methods is strongly recommended
as a QC check on any losses in the sample preparation steps.  For this purpose,
the additions should be made to the sample prior to any sample preparation.


     Combination and preparation of representative aliquots of collected samples
is appropriate for all MSW combustor solid and liquid effluent samples.  The
collected sample is homogenized prior to withdrawal of aliquots for analysis.
Individual aliquots are composited to form a single sample (or replicate QC
samples) for subsequent preparation and analysis procedures.  Table 6 summarizes
these procedures.

                                                          TABLE 5

                                           SUMMARY OF SAMPLE PREPARATION METHODS
                                                                                            Cross  References
Stack Gas
M5, MM5 or

Sampling grep Analysis
                     - probe wash
                     - filter

                     - probe wash
                     - filter
                     - impinger

                     - probe wash
                     - filter
                     - sorbent

                     - condensate
                     - sorbent
                      - condensate
Dry to constant weight   III B
Standard addition to     III B
split samples.  Digest
in acidic oxidizing

Add surrogate, Soxhlet   III B
extract with CH_Cl
Concentrate. Clean-up
as necessary

Add surrogate            III B
extract at pH 2 and
pH 11 with CH C12,
Concentrate. Clean-up
as necessary

Spike with internal      III B
standard.  Thermally
desorb onto analytical
trap.  Desorb this trap
into GC/MS.

Spike with internal      III B
standard.  Purge onto
analytical trap.  Desorb
into GC/MS.

                                                          TABLE 5

                                       SUMMARY OF SAMPLE PREPARATION METHODS (cont.)
                                                                                             Cross  References
Flue Gas

Bottom Ash
Fly Ash
(same as for

Procedure Sampling
stack gas) III C

Standard addition to III D
split samples. Digest
in acidic medium in
Prep Analysis

                                      Parr bomb.

                                      Add surrogate.   Soxhlet  III D
                                      extract with CH^Cl,
                                      as necessary,
Standard addition to     III D
split samples.  Digest
in acidic, oxidizing

Spike with internal      III D
standard.  Purge onto
analytical trap. Desorb
into GC/MS.

Add surrogate.  Liquid   III D
liquid extract at pH 2
and pH 11 with CH.Cl
Concentrate.  Clean-up as
                                                        IV C-F
                                                                           IV G
                                                                           IV C-F

                                                          TABLE 5

                                       SUMMARY OF SAMPLE PREPARATION METHODS (cont.)
                                                                                             Cross References

Grind or mill to
reduce particle size.
£rejj Analysis

                                                            Take subsaraples.
Same as for ash
Spike with Internal
standard.  Dilute in
reagent water or poly-
ethylene glycol in purge
cell.  Purge, trap and
desorb into GC/MS,



       Gross references are to sections of this document.
        Reference:  Miller, N.G., R.W. James and W.R. Did   n, "Evaulated Methodology for the Analysis
        of Residual Waste," Report prepared under EPA Contract No. 68-02-1685 (December 1980).
        Also see SW-846 method 8240.

                                                     TABLE 6

                                  SUMMARY OF PROCEDURES FOR COMPOSITING  SAMPLES


Shake well and pour

Combine aliquots in
clean container and
shake to mix well.
  Minimum Quantity
of Composited Sample

 1 L (semi-volatile
 1 L (metals)
                       NOTE:   For volatilea analysis, composite just prior  to analysis
                               by adding aliquota from multiple VOA bottles  to purge
                                                                       5 mL (volatile
Stir or shake well;  use
dipper to take > 3
Grind, if necessary,  to
reduce particle size  (20
mesh screen) using agate
or alumina equipment;
riffle through steel  or
aluminum riffler.
Combine aliquots in
clean container and
stir or shake to

Combine aliquots,
cone-blend three
times, roll-blend,
cone and quarter.
 100 mL (semi-vol
 100 g (metals)
 50 g (organics)
 100 g (metals)

     Samples of stack (flue) gas collected  with an extractive sampling train
(M5, MM5, SASS or VOST) already represent time-averaged sample collections.  In
effect, the sampling approach has composited the gas on a time-weighted basis.
It is generally inappropriate to further composite such samples.  However, in
some cases where ultra-low levels of detection are required (e.g.,  dioxin/furan
analysis), it may be necessary to pool the entire extracts from multiple
sampling runs.


     The specific sampling methods listed in Table 2 contain explicit procedures
for the physical recovery of samples from the train.  However, they do not
necessarily specify procedures for monitoring the chemical recovery achieved in
the sample preparation process.  There are two basic approaches that can be used
for this purpose:  (1) addition to each sample of surrogate compounds, which are
chemically similar to the species of interest but not expected to be present in
the sample (e.g., for GC/MS analyses, stable isotope-labelled analogs of the
target compounds); or (2) standard addition (spiking) of the target species
themselves to selected split samples.

     The spiking levels used in each instance are selected after consideration
of the target detection level for each analyte and the expected concentration of
the species in the sample.  For MSW incinerator effluent samples, it is
generally desirable to select spiking levels that correspond to 2 to 10 times
the target detection limit or to 2 to 4 times the critical decision limit.  The
level chosen should be explicitly stated in the QAPP for each test program.

For example,  assume that the critical decision limit for a semi-volatile organic
pollutant (e.g., 2, 3, 7, 8-teterachlorodibenzodioxin, TCDD) is 1 ng/m  in stack
gas, and a 5 m  sample of stack gas is to be collected using the MM5 train.  An
appropriate level of surrogate (e.g.,   Ci2~2> 3' 7> 8-TCDD) to be spiked into
the MM5 train components prior to extraction can be calculated as:

     4x1 ng/m  x 5 m  - 20 ng surrogate

Assuming that the organic extract is concentrated to a final volume of 1.0 ml,
and that 5 pL are injected into the GC/MS system, this spiking level would yield
100 pg of surrogate on-column if recovery through the sample preparation steps
were 100%, and 50-pg on column if recovery were only 50%.  It must be confirmed
by analysis of calibration solutions that these levels are within the analytical
detection limit of the GC/MS system.  Recovery of 60% or higher can be
        ( 21)
expected     if the laboratory procedures are under control.


     Solvent extraction with methylene chloride is the procedure most broadly
used to prepare samples for organic analysis. In the case of stack sampling
trains, several separate extractions (of probe wash and filter, sorbent trap,
and of condensate) will be required.  It is general practice to combine these
extracts prior to analysis so that a single value is obtained for the stack gas
concentration of each species.  This approach allows lower detection limits to
be achieved.  Further, there is no intrinsic advantage to differentiating
between material collected in the "particulate" (front half) vs "vapor" (back
half) portions of the train, since these catches do not necessarily indicate the
                                                   /4 22 23)
particulate/vapor distribution present in the stack  '  '

     Liquid-liquid extraction, using a manual, separatory funnel method or a
continuous extractor, applies to:  (1) spent scrubber water or other waste water
effluent from the MSW combustor and (2) aqueous condensate collected from the
stack gas effluent in the MM5 or SASS trains.  The volumes of sample/extracting
solvent and any necessary pH adjustment of the sample should be as specified in
EPA Method 625 if the purpose of data collection is NPDES compliance
determination.  In the case of stack gas condensate, these parameters can be
scaled down to accommodate the actual quantity of sample collected.

     Soxhlet extraction applies to: (1) solid effluents from the MSW combustor
and (2) probe wash, filter and sorbent components of the MM5 or SASS trains.
Quantities of sample and extraction solvent should be as specified in Method
P024 b (Sampling and Analysis for Hazardous Waste Combustion)     or Method 3540

     Some very wet ash samples or sludges may require alternative extraction
procedures (e.g., homogenization, Methods P022 a,b or  P024 a,c).      Also,
solid or slurry samples that are not suitable for direct introduction to the
purge cell of a purge-and-trap apparatus may require a micro-extraction (e.g.,
Methods P022b, P024c)     prior to determination of volatiles.

     If the MSW combustor test program includes evaluation of solid residues
vis-a-vis the RCRA hazardous waste characteristics, a separate sample of the
solid material must be extracted by the EP (Method 1310, SW-846) or the new TCLP
(Fed. Reg. Vol. 51, No. 114, June 13, 1986), when promulgated as a final method.

     Unless an alternative procedure is specified in a particular analytical
procedure, solvent extracts should be passed through a short column of anhydrous
sodium sulfate into a Kuderna-Danish evaporative concentrator apparatus.  In
most cases, rapid concentration to a final extract volume of 1-10 mL will
provide adequate detection limits without unacceptable losses of semi-volatile
organic species.


     For some samples,  the level of interfering compounds is sufficiently high
to preclude successful analysis for the species of interest.  For such samples,
one or more clean-up steps must be included in the sample preparation procedures
Because of the wide variation in sample matrices and in the physical/chemical
properties of the species that may be sought, no single method or set of methods
can be recommended for MSW combustor samples.  This is one of the reasons why
use of surrogate spiking of MSW combustor samples is strongly recommended
(Section IV C).  If it can be demonstrated that recovery and detectability of
the surrogate are adequate, then no clean-up steps are necessary.  However, if
interferences overload the analytical column or detector and the surrogate is
not detectable, clean-up will be required.  This must be established empirically
for each project.  The QAPP should specify the criteria for acceptable surrogate

recovery/detectability and corrective actions to be taken if the criteria are
not met.

     Some analysis methods (e.g., Methods 8010-8250, SW-846( J) specify that
silica gel, Florisil or alumina column clean-up be applied.  In some cases three
or more sequential clean-up steps may be required.  For example, the ASME
methodology for analysis of dioxins and furans    specifies sequential passage
of the concentrated sample extract through 1) a combination column containing
silica gel and acid and base-modified silica gel, 2) a basic alumina column, 3)
a PX21 carbon/celite 545 column and 4) a silica/diol column.  Such an extensive
clean-up is usually not necessary unless the critical decision limit is in the
ng/mg  range.  However, using at least one column clean-up step (silica gel,
alumina) may significantly improve detection limits when critical concentrations
are in the pg/m  range.

     The preparation method for all samples requiring metals analysis includes a
digestion step.  Its purpose is to convert all of the metal-containing species
into inorganic form.

     For most sample types, an acidic, oxidizing medium is specified for
digestion.  Solid samples can be digested using HF and HNO. in a Parr bomb.
Solutions can be digested using HNO. and H«0.    .   In some instances, repeated
digestion using HNO, alone    may give adequate recovery.  For mercury in solid
materials, Method 105 (40 CFR Part 61) calls for aqua regia digestion followed
by potassium permanganate oxidation.

     A relatively new development is the use of microwave energy, rather than a
conventional hot plate, for acid digestiors.  This reduces the time and acid
quantity required for complete digestion

                             V.  ANALYSIS PROCEDURES


     The overall strategy for analyzing samples from MSW combustion must reflect
the multiple possible purposes of data collection discussed in Section II  A.

     When the purpose of the data collection is to determine compliance with a
specific regulation concerning air, water, or solid effluents (or their
disposal) it is essential that the analysis method used be one recognized by the
appropriate agency.  These include, but are not necessarily limited to:

     EPA  NSPS Methods 1-5              40 CFR Part 60
     EPA  NESHAPS Methods 101-108       40 CFR Part 61

     EPA Clean Water Act Methods        40 CFR Part 136
     601-612; 624, 625; 1624, 1625;

     EPA RCRA Methods                   SW-846
     7040-7951; 8010-8310

     One exception to the above generalization relates to measurement of
criteria pollutants in combustor effluents.  The EPA NAAQS methods for criteria
pollutants are ambient air monitoring methods that are suitable for source
monitoring only with modifications (gas conditioning; dilution).  However, there
are NSPS methods applicable to direct measurement of criteria pollutants in
combustion sources.

     Even when the purpose of data collection is to provide input for more
general environmental assessment, or risk assessment, rather than regulatory
compliance,  the above EPA-approved methods are frequently useful.  In
particular,  the nearly-identical GC/MS-based methods designed for determination
of priority pollutants (624,  625; 1624, 1625) or Hazardous Substance List  (HSL)
compounds (RCRA 8240, 8250, 8270) have wide applicability for organic analysis.

The corresponding procedures for metals analysis by atomic absorbtion
spectroscopy (AAS) or inductively coupled plasma spectroscopy (ICP) are also
directly applicable to MSW combustor samples.

     For some assessment purposes,  official federal EPA methods may be
inappropriate or unavailable. In these cases, methods promulgated by ASTM and
ASME should be used if available.  For example, there is at present no official
method for determination of dioxin/furan congeners that applies specifically to
MSW stack emission sample analysis.  However, a procedure developed to supuort a
joint ASME-EPA project on MSW comb  , -.ion is recommended    .   In addition,
SW-846 method 8280 may be used for homolog-specific analysis.
     Similarly, there are presently no official methods suitable for screening
MSW combustor samples to identify and/or quantify other pollutants that may be
present as a result of incomplete combustion.  Research in this area is underway
as well.  The selection of appropriate analysis methods for such pollutants must
be established on a case-by-case.  Part C of this section provides some
preliminary suggestions.

     Potentially useful methods for direct analysis of selected species in MSW
combustor effluents using continuous instrumental monitors will be discussed in
Section VI of the report.


     "Proximate analysis" is a term used in "Sampling and Analysis for Hazardous
Waste Combustion (First Edition)"     to describe procedures for determining the
approximate (gross) composition of hazardous waste.  The methods recommended for
that purpose include:

     % moisture, solid and ash determination
     elemental analysis (%C, N, S, P, F, Cl, Br, I)
     total organic carbon
     total organic halogen

     heating value

These methods are applicable to MSW.


     As noted above, screening analysis is intended to identify and/or quantify
species chat are present in a sample and are of potential concern with regard Co
effects on health and/or environment but were not specifically pre-selected for
analysis (see Section D, Directed Analysis).  There are no official procedures
for screening analysis.  In practice, the approach that has most commonly been
used in research and development programs involving waste combustion has been as

     For organic species, perform full-mass range scanning (e.g., 40-500 amu)
under appropriate GC/MS conditions  (usually capillary column and El ionization).
In most cases, the GC/MS conditions specified in EPA. methods 624/1624/8240 for
volatiles and in EPA methods 625/1625/8250/8270 for semivolatiles will be
applicable to screening analysis of MSW combustor samples.  In addition to
providing data for directed analysis, the results of these analyses can be used
to identify unknown or unexpected compounds by comparison to a library of
reference spectra or by spectral interpretation.  A typical criterion for
screening analysis is to attempt to identify the 20 most intense unknown peaks
or those peaks whose intensity exceeds 10% of the internal standard intensity
(whichever is the lesser number of unknowns).

     For inorganic species, ICP analysis can be used to screen for about 20-23
metals that are known or suspected  to be of environmental concern.  Alternacive
approaches sometimes recommended to address a broader range of elements include:
X-ray fluorescence, neutron activation analysis, particle-induced X-ray
emission, and spark source mass spectrometrv.  All of these are multielement
methods and require minimal sample preparation.  However, the instrumentation
required is expensive and not widely available.  Also, the methods are more

amenable to qualitative screening than to quantification unless standards and
samples are carefully matched to eliminate sample matrix effects.

     It is probable that screening analysis is more likely to be requested for
MSW combustor air emission samples than for liquid or solid effluents.  These
methods can be applied to samples of MSW combustor stack gas collected with the
MM5, SASS or VOST (volatile organics only) after suitable sample preparation
techniques have been applied.  They should also apply to liquid or solid
effluent samples after sample preparation.


     As noted in the overview to this Section, analysis methods are available
for most species that are likely to be pre-selected as targets for analysis in
MSW incineration.  Table 6 lists suitable analysis methods for organics and
metals that may be of concern.  Specific methods for distinguishing between
species such as Cr (VI) and Cr (III), suitable for use with combustion effluent
samples, based on selective extraction of Cr (VI) by an alkaline reagent, are
under development and in the process of validation by EPA's Emission Measurement
Branch at Research Triangle Park, North Carolinav  ' .   In addition, SW-846 lists
methods both for Cr(VI) and for total chrome;  these methods are potentially
applicable to MSW combustor samples providing that suitable procedures for
solubilization of the chromium are applied.

     Note that, for organic analysis, Table 7 lists only the GC/MS alternatives.
Methods that use GC or HPLC with detection principals less specific than MS
(e.g.,  flame ionization (FID) or electron capture (ECD) detection for GC,
ultraviolet (UV) or refractive index (RI) for HPLC) are less likely to be useful
for MSW combustion samples, because of the variety and quantity of potentially
interfering substances likely to be present.

                                                 TABLE 7

                                   APPLICABLE TO MSW COMBUSTOR SAMPLES
       Species                                       Method                                  Reference

Volatile Organics                          Packed column GC/MS; full mass range                1,  15
                                           scanning 20-260 amu.

Semivolatile Organics                      Capillary column GC/MS; full mass range             1,  15
                                           scanning 40-500 amu.

Dioxins/Furans                             Capillary column GC/MS; selected ion                1,  5

Metals                                     Flame (high levels) or furnace (low                 1,  24
                                           levels) AAS.

                                           Inductively coupled plasma spectroscopy             1,  24
                                           (not for mercury, lead, arsenic)

                       VI.  CONTINUOUS .MONITORING METHODS


     Continuous monitoring systems include _in. sjLEu measurements, in which a
sensor is mounted directly in a stack or flue, and extractive methods, in which
a sample of stack or flue gas is pumped through an interface to the measuring
device.  In situ monitoring offers the advantage that no alteration in flue gas
composition is introduced by the sy^'-em.  Conversely, however, an in situ sensor
must be physically resistant to st,    conditions (temperature, moisture,
particulate matter) and be chemically selective (blind to potential
interferents).   An extractive approach is amenable to gas sample conditioning to
remove substances that might interfere with the desired measurement or damage
the instrument.  It also allows instruments to be placed in a sheltered location
where maintenance and calibration are more convenient.  However, this approach
provides opportunity for loss of target species in transfer lines and in
components of the gas conditioning system.

     Many continuous monitors are based on rather sophisticated chemical
analysis principles.  Manufacturers have aa.de considerable efforts to
"ruggedize" the commercially available systems so that they are capable of
unattended operation for periods of days to weeks and can be
maintained/calibrated by plant operating personnel.  Despite these advances, it
is vital that a rigorous QA/QC program be established for continuous monitors to
ensure that misleading data are not recorded and that operating problems are not


     An extractive sampling system interface typically includes the following

          *    probe
          »    coarse filter
          *    transfer lines

          •    pump
          •    moisture removal system
          •    fine filter

     The probe must be located at a point in the stack or flue that allows a
representative sample to be withdrawn.  A multi-port probe may be useful in this
regard.  Depending on stack or flue temperature and corrosivity,  materials of
construction may be stainless steel or ceramic.  The coarse filter, usually
located in-stack to minimize effects of condensation, is typically sintered
metal or ceramic, depending on temperature.   Ceramic is more resistant to high
temperature corrosion effects but also more  susceptible to cracking.  Plugging of
the probe and/or coarse filter can be a severe problem, especially when - ,;npling
flue gas upstream of particulate control devices.   Backflushing the system
regularly with clean carrier gas may help but frequent replacement of the
in-stack filter will probably still be required.

     The transfer lines must be constructed of inert materials that are
resistant to corrosion and do not absorb the species to be measured;  ceramic
(probe) and Teflon   are materials of choice for most applications. Stainless
steel would be an alternative if the stack or flue gas is not highly corrosive.
The pump must also be constructed of, or coated with, inert material such as
Teflon .   Especially when samples are to be  monitored for organics, any
potential contamination of the sample with lubricants must be avoided.
Diaphragm or bellows pumps can meet this constraint.

     Moisture removal can be accomplished in several ways:  adsorption (e.g.,
silica gel); condensation; dilution to below dew point; membrane permeation
system.  Issues that must be addressed in selection of the drier component
include:   adequate capacity for water removal given the moisture content of the
stack or flue gas; and minimal coincident losses of target species along with
the water.  A condensation approach, for erample, may lead to unacceptable
losses of acid gases such as HC1 and SO .

     The fine particulate filter is usually located close to the analyzer inlet.
It usually must achieve virtually complete removal of all particles larger than

1 micron.  When sampling flue gases, these filters require replacement whenever
the pressure drop across the filter approaches the limit of pumping capacity.


     The following is based on information provided by instrument suppliers and
by references 26 and 27.

     Continuous temperature measurements in combustor flue or stack gases are
generally accomplished by using type J, K, or RT thermocouples.  The
thermocouples must be shielded from radiation and protected against mechanical
damage and corrosion by shielding inside a ceramic or metal protection tube or
in a thermowell.

     Continuous monitoring of particulate material is generally accomplished
using an in situ opacity meter.  Typically, these devices measure changes in
optical density, OD (percent transmittance),  due to scattering and/or adsorption
of light by particles in the stock.  The OD reading obtained depends not only on
the mass loading of particulates that are present, but also on their size
distribution, the particle shape, particle composition, the system's
temperature, the presence or absence of water droplets and the configuration of
the stack.  The lack of measurement specificity may render opacity monitors less
reliable at MSW combustors than at other stationary sources, as waste feeds at
MSW locations are highly variable, probably causing emission levels and
compositions to vary over time as well.  Also, commercially available opacity
meters for stack monitoring may be uncertain by a factor of two or more at
particulate loadings below 0.03 gr/SCF.

     Extractive sampling, rather than in situ monitoring, is most commonly used
for inorganic gases, although in situ monitors are available for CO, C02, 02,
NO ,  and SO .  Typically, the detection of pollutant species of concern is
  X        X
accomplished using one of the detection principles identified in the descriptive
information presented below.  There is relatively little history of application
of these instruments to MSW combuster stack gas and, especially,  flue gas.

                             for continuous monitoring at other types of
                             detection principle used by continuous analyzers is
                                Amenable for use in the determination of CO,
                              >y NDIR is based on the principle that each of
                              lers, will absorb specific signature wavelengths of
                              rating matched beams at the correct signature
                              h a clean "reference" cell while the other is
                              r.  sample gas, and determining the difference in
                               beams, it becomes possible to measure the amount
                              Les that is present.

                               icomplished using devices such as a. thermistor, a

                               f NDIR based instrumentation is the fact that it
                               i that the technology is applicable to a wide
                               Vlso, instruments are relatively rugged and
                               lave been in use in field monitoring situations
                               ire associated with this detection principle are
                                iecies that will absorb similar signature
                                nd the fact that optical systems needed to
                                he generated infrared light may degrade due to
                                 NDUV) analyzers employ much the same philosophy
as uo cne analyzers based on NDIR absorbtion, only in this instance, the  light
source emits in the ultraviolet or visible regions of the spectrum and a
reference cell is generally not used.  Typically, a reference wavelength  in a

region where the pollutant species of interest has minimal absorptive capacity
is generated and quantitation is completed by differential analysts.

     An important advantage of the NDUV analyzers is that water vapor is not an
interference, as water does not absorb light in the ultraviolet region of the
spectrum.  As is the case with most extractive monitoring techniques, however,
particulates which will absorb or scatter generated light must be removed from
the sampled gas stream.

Poljirographic Analyzers

     Numerous pollutant species of potential interest at MSW combustors may be
measured continuously using polarographic analyzers.  Polarographic analyzers
operate on the principle of selective diffusion and chemical reaction of a
pollutant species of interest which induces a current flow that is measured
electronically.  The sensing device intrinsic to the operation of all
polarographic analyzers is commonly called the electrochemical transducer.

     In operation,  the pollutant species of interest enters the transducer
through a selective, semi-permeable membrane.  Once in the transducer, the
pollutant is oxidized or reduced by reaction with an electrolyte solution which
induces a current flow.  The induced current flow is proportional to the
concentration of the pollutant species in the gas stream.  The selectivity of
electochemical transducers is dependent upon the selection of membrane
materials,  electrolyte, and sensing electrode chemistry or composition.

     The polarographic analyzers offer several advantages over other analyzers,
including multi-pollutant capability by switching transducer, fast response and
simplicity of operation.   Principle disadvantages of this technique are that
transducers must be replaced or rejuvenated periodically, and the instrument
must be frequently calibrated because the response of the transducer does
deteriorate as the electrolyte solution is consumed.

Electrocatalytic Analyzers

     Electrocatalytic analyzers are currently available for oxygen
determination.  This type of instrumental determination is based on the
principle that current flow is induced when two solutions containing similar
materials at differing concentrations are brought into contact.  Thus,  by having
two cells, one a reference and the other the unknown, separated by a porous
zirconium oxide barrier (acting as both an electrolyte and catalyst),  an
electron current flow can be promo" d, measured, and related directly to o>vgen

Paramagnetic Oxvgen Analyzers

     Oxygen content may also be determined using a paramagnetic analyzer.
Detection in a paramagnetic analyzer utilizes the fact that oxygen molecules
(and the molecules of a few other compounds) are attracted by a magnetic field.

     Two different philosophies of paramagnetic detection are commonly used in
instrumentation.  One, called thermomagnetic or magnetic wind instruments are
based on the principle that the magnetic attraction of oxygen decreases as
temperature increases.  The second, called magneto-dynamac, uses the
paramagnetic attraction of oxygen to swing a specialty torsion balance.


     Certain inorganic pollutants, most notably nitrous oxide and ozone, may
also be detected using analyzers based on chemiluminescence.  Within this type
of instrumentation, the pollutant species of interest is mixed with a second
reactant to generate light.  Ideally, the generated spectrum of light produced
is specific only to the pollutant of interest,  but more commonly, optical means
are used to isolate and quantitate the intensity of specific wavelength.  By
measuring the intensity of the generated light, a direct estimate of the
concentration of pollutant species present in the gas sample may be obtained.

     Information of currently available instrumentation for inorganic pollutant
species in MSW combuster exhausts are presented in Tables 8 through 14.


                                                          TABLE 8

                                        Continuous Analyzers for Carbon Monoxide


|0-50 PPM
|0-100 PPM
|0- 200 PPM
| 0-1000 PPM
JO- 5000 PPM
|0-1 %
|0-5 %
|0-50 PPM
| 0-100 PPM
|0-250 PPM
JO- 500 PPM
JO- 999 PPM
|0-4 %

Interferences Comments
| Water Vapor and
| Particulates ; other
(species with similar
| infrared absorbance
| characteristics
(Unsaturated Hydro -
| carbons and ammonia

In-situ or remote; |Anarad, Inc.
Water Vapor and |Dynatron
Particulates should (Horiba
be removed; corro- | Infrared Ind.
sive gases may etch |Rosemont (Beckman)
optics (Servomex

Requires sample
conditioning to
remove particulate
and reduce tempera-

Energetics Science

                                                        TABLE 9
                                        Continuous Analyzers  for  Carbon Dioxide
|Nondispersive |0-10 PPM
Infrared | 0-100 PPM
j 0-500 PPM
|0-5000 PPM
|0-0.5 %
|0-2.5 %
|0-5 %
|0-20 %
|0-100 %
Polarographic 0-50 PPM
(Electrochemical) | 0-100 PPM
JO-250 PPM
|0-500 PPM
| 0-999 PPM
0-4 %

Water Vapor and | In- situ or remote;
Particulates; other | Water Vapor and
species with similar | Particulates should
infrared absorbance |be removed; corro-
characteristics |sive gases may etch
| optics
Unsaturated Hydro- (Requires sample
carbons and ammonia | conditioning to
(remove particulate
| and reduce tempera-
( ture
Anarad, Inc.
Infrared Ind.
Rosemont (Beckman)
Syconex Corp.


                                               TABLE  10
                                   Continuous  Analyzers  for Oxygen
|0-5 %
(0-10 %
|0-25 %
|0-35 %
 0-100 %
                    |Gas must be cooled  [Horiba
                    |and particulate re- |MSA
                    |moved               |Neutronics
 0-10 %
 0-22 %
                    (Good for high tern-  |Anarad,  Inc.
                    (perature applicationjDynatron
                                         |Lynn Products
                                         I MSA
                     |0-5 %
                     |0-10 %
                     (0-25 %
                     |0-50 %
                                                          (Rosemont  (Beckman)

                                                TABLE 11
                              Continuous Analyzers for Sulfur Dioxide
|Nondispersive jO-500 PPM (Water Vapor and |In-situ or remote; |Anarad, Inc.
[Infrared JO- 2000 PPM
j | 0-10000 PPM
1 |0-1 %
|0-5 %
| |0-10 %
i |0-30 %
(0-100 %
1 1
1 1
Nondispersive |0-50 PPM
(Ultraviolet j 0-100 PPM
(0-500 PPM
|0-1000 PPM
| 0-5000 PPM
Particulates ; other (Water Vapor and (Dynatron
species with similar) Particulates should
infrared absorbance (be removed; corro-
sive gases may etch
(optics; Generally
(selected for percent

Polarographic (0-5 PPM (Methyl and Ethyl
(Electrochemical) (0-10 PPM (Mercaptans, Hydrogen
(0-50 PPM
(0-100 PPM
Sulfide, Ammonia

Infrared Ind.
Rosemont (Beckaan)
S iemens
Syconex Corp .

Anarad , Inc .


                                               TABLE 12
                              Continuous Analyzers for Nitrogen Oxides
|Nondispers ive
 (for NO only)
0-500 PPM
0-2000 PPM
0-10000 PPM
0-2 %
0-10 %
(Water Vapor and     [Water Vapor and     (Horiba
 Particulates;  other (Particulates should (Rosemont  (Beckman)
 species with similar|be removed; corro-   (Siemens
 infrared absorbance |sive gases may etch (Syconex Corp.
 characteristics     (optics
|Nondispers ive
 (for N02 only)
0-50 PPM
                                           Anarad Inc
 Po 1arographic
 (for NO only)
0-25 PPM
0-50 PPM
0-250 PPM
(Methyl Mercaptan,
(Ammonia,  Nitrogen
 Dioxide,  Sulfur
(Requires  sample
(conditioning to
 remove particulate
 and reduce tempera-
Energetics Science
0-2.5 PPM
0-10 PPM
0-25 PPM
0-100 PPM
0-1000 PPM
0-2500 PPM
0-10000 PPM
 Possible Ammonia,
 Carbon Dioxide,  and
 Need to dry and
 remove particulates
 from sample
Monitor Labs
Rosemont (Beckman)
Thermo Electron

                                                            TABLE 13

                                       Continuous Analyzers for Hydrochloric Acid
                (Water  Vapor and     |In-situ  or  remote;   (Flakt
                (Particulates;  other  (Water Vapor and      |Syconex Corp
                (species  with similar|Particulates should  (Teco
                 infrared absorbance  |be removed; corro-
                 characteristics     (sive gases  may etch
(0-5  PPM         Chlorine  gas,  Methyl(Requires  sample
|0-10 PPM        and Ethyl mercaptan,(conditioning  to
jO-20 PPM       |Hydrogen  Sulfide,    (remove  particulate
(0-50 PPM       (Ammonia,  Nitric  Ox-  (and  reduce  tempera-
j0-200 PPM       |ide,  Hydrogen  Cyan-   ture
|                jide,  and  Sulfur  Di-
|                (oxide
I                I
I                I

                                            TABLE 14
                             Continuous Analyzers for Hydrogen Cyanide
|0-5 PPM
(0-10 PPM
|0-20 PPM
(Chlorine gas,  Methyl(Requires  sample     (InterScan
(and Ethyl mercaptan,(conditioning to     (
                                    (Hydrogen Sulfide,
                                    (Ammonia, Nitric Ox-
                                    (ide, Hydrochloric
                                    (acid, and Sulfur Di-
                                    j oxide
                      remove  particulate
                      and reduce tempera-


     Total hydrocarbons  (or total non-methane hydrocarbons) at ppm to percent
levels in stack gases can be monitored continuously with a flame ionization
detector (FID) or infrared (IR) detector.  These detectors are relatively rugged
and are quite sensitive  to hydrocarbons.  The response factor is generally lower
for organics that incorporate functional groups such as haxides, hydroxyl,
carbonyl, carboxylate.

     The photoionization detector (PID) is applicable to many of these organic
categories,  but experience with this detector as a continuous monitor is more
limited.  There is some  evidence that maintenance is more of an issue with PID
than with FID or IR instruments.

     The electron capture detector (ECD), which has high sensitivity and
selectivity for halogenated organics under laboratory conditions, is not rugged
enough for routine continuous monitoring in the field.  Also, because these
detectors contain radioactive materials, NRG permitting regulations govern their
installation and use. The Hall detector, also specific for halogenated species,
has been used at hazardous waste incineration sites, but with difficulty.

     Catalytic combustion (hot wire) and thermal conductivity detectors are also
used for continuous monitoring of organics.  However, most commercially
available instruments based on these principles are generally designed for
percent level concentrations corresponding to explosive limits of combustible
gases and vapors.  A few low-level instruments suitable for MSW combustion
monitoring are available, however,

     Monitoring of specific organic compounds, rather than total organics,
requires that chromatographic separation be accomplished prior to detection,
Instrumental monitors that interface a gas chromatograph to an FID or PID are
commercially available.  These operate in a semi-continuous basis, since the
.chromatographic separation imposes a cycle time of (typically) 5-30 minutes
between measurements..   The GC/FID or GC/PID analyzers are vulnerable to false

positive interferences because the retention time is an imperfect means of
compound identification.

     Instruments based on more selective detection principles (e.g., GC/MS or
GC/FTIR) are beyond the present state-of-the-art for stack monitoring, except in
research installations.  Instruments using these detectors may be sufficiently
expensive to install and demanding to operate that they are not suitable for
routine continuous monitoring.  Most require, for example, more stringent
control of temperature, humidity and power supply than is likely to be practical
at -.n operating MSW plant.


     A conceptual ideal for continuous monitoring of MSW combustors would be a
inexpensive,  rugged, and simple instrument that provides continuous measurement
of a surrogate parameter indicative of total system performance.  Identification
of a suitable surrogate parameter, which can be reliably correlated with
emissions of any and all pollutants of potential concern, is one of the
objectives of on-going research in MSW combustion technology.  At the present
time, the carbon monoxide level in the air emissions is generally regarded as
the best available surrogate for chemically-based monitoring of overall
combustion efficiency.  Monitoring of combustion temperature is also relied on
as an indicator that the process is in control.  Methods for monitoring these
indicator parameters are discussed in Section VI C.


     All of the general provisions that will be discussed in Section VII apply
to continuous monitoring.  A few special considerations are worth noting,

     First,  calibration with zero and span gases should be performed on the
system as a whole, not just on the analyzer/detector itself.  This may, however,
be literally impossible for in situ monitors; generating a stack or flue full of
calibration gas of known composition is not a realistic approach.  In the case

of extractive monitoring systems, provision should be made to allow introduction
of calibration gas at, or just behind, the probe. This approach to "total
system" calibration has been incorporated into all of the federally referenced
continuous monitoring methodologies that are currently applicable to stationary

     The EPA has also established guidelines for instrumentation that is used to
continuously monitor sulfur dioxide, nitrogen oxides, carbon dioxide, oxygen,
and carbon monoxide emissions from stationary sources.  These guidelines fall
into two categories including both   strumental design criteria and performance

     Instrumental design criteria establish minimum acceptable levels for items
such as instrumental response time, interference rejection ratios and ranges.
Performance specifications are established to assure that the developed data is
accurate and precise.  Specifications for calibration drift and the relative
accuracy of calibration, as well as requirements for duration of unattended
operation are established.  Specific criteria applicable to monitors used to
measure CO, C02, 02, NO , and SO  are provided in Title 40 Code of Federal
                       X        &
Regulations, Part 60.13 and Appendix A and B.

     Second, QA/QC procedures for continuous monitoring methods must take into
account the fact that operating personnel may have had minimal training and
experience with chemical measurements and analytical instrumentation.  The
procedures must, therefore, place minimum reliance on operator judgment and be
as explicit and simple as possible concerning QC criteria and corrective

     Third, the question of data reduction and data maintenance for continuous
monitor output deserves special consideration.  A continuous monitor, in one
sense,  generates an infinite numbe" of data points on pollutant concentration
vs. time.   It must be determined in advance whether daily, weekly, monthly, etc.
data are to be archived and/or reported.  Are running averages desired?
Commercially available data loggers are able to conduct these functions once

decisions are made with respect to these parameters.  How should/must short-term
excursions from average values be recorded/reported?


     The acquisition of continuous monitoring data affords opportunities for
process control, either by operator intervention or by use of automatic feedback
loops.  In current MSW combustion practice operator intervention is the more
common response mode, but newer facilities are moving toward more automated

     Catastrophic failure of the system (e.g., plugging of nozzles in a dry
scrubber or breakage of a fabric filter) is generally readily detectable by
operators even in the absence of continuous monitoring data.  Temperature and,
secondarily, ojcygen data from continuous monitors can be (and are) used to
adjust process operating parameters such as MSW feed rate or over/underfire air
supply to the incinerator.

     Use of chemical monitoring data for process control is more complicated,
both in theory and in practice.  In general, the relationship of MSW combustion
operating conditions to the emission rates of pollutants, especially trace
organics, is not well enough understood to allow systematic process control in
response to monitoring data.



     A vital part of any sampling and analysis program is the provision for
procedures which maintain the quality of the data obtained throughout the
sampling and analysis exercise.  These procedures, termed quality assurance and
quality control (QA/QC),  serve to (a) document the quality (i.e., accuracy,
precision, completeness,  representativeness and comparability) of generated
data; (b) maintain the quality of data within predetermined control limits for
specific sampling and analysis procedures; and (c) provide guidelines for
corrective actions if QC data indicate that a particular procedure is out of

     The following definitions, which represent interdependent activities, serve
to differentiate between the complementary activities of quality assurance and
quality control.

     •    Quality Assurance (QA) activities addresses the delegation of program
          responsibilities to appropriate individuals, documentation, data
          review,  and audits.  The objective of the QA procedures is to allow an
          assessment of the reliability of the data.

     •    Quality Control (QC) activities address the maintenance of facilities,
          equipment, personnel training, sample integrity, chemical analysis
          methods, and production and review of QC data.  QC procedures are used
          continuously during a. sampling and analysis program to maintain the
          quality of data within control limits.   QC data should be evaluated
          immediately by the analysts; if the QC data fall outside a set of
          specified control limits,  corrective actions, as specified in the work
          plan, must be taken.

In this section, specific QA/QC procedures are described.  For an individual
sampling and analysis program, these procedures and/or others may be selected to
reach the goal of obtaining high-quality data.  At a minimum, the procedures

which are selected must be consistent with the standard ooerating procedures
and/or good laboratory practices of the sampling crew and analytical laboratory

     The following discussion of QA/QC procedures is based upon a. guidelines
document    issued by the Office of Monitoring Systems and Quality Assurance of
the EPA Office of Research and Development.  This document, QAMS 005/80,
entitled "Interim Guidelines and Specifications for Preparing Quality Assurance
Project Plans," and the references cited therein provide an extensive resource
in selecting appropriate QA/QC procedures for sampling and analysis efforts.
The QAMS-005/80 document identifies sixteen essential elements of a QA Project
Plan (QAPP).   These elements are listed in Table 15 and described briefly in the
following discussion.  A QAPP of some kind would generally be required for most
sampling and analysis efforts.


     These elements are self-explanatory.  The title page should indicate
individuals with QA responsibility for the sampling and analysis efforts.
Approval signatures should be required prior to the start of the sampling and
analysis efforts.  The Table of Contents should include a distribution list for
the QAPP.


     A general description of the project, including the experimental design,
and intended use of data should be provided.  The description may be brief but
should have sufficient detail to allow the individuals responsible for review
and approval,  of the QAPP to perform their task.  Flow diagrams, tables and
charts should be included, as appropriate.  A schedule, with anticipated start
and completion dates, should also be specified.

                                    TABLE  15


     1.        Title Page

     2.        Table of Con"  nts

     3.        Project Description

     4.        Project Organization and Responsibility

     5.        QA Objectives

     6.        Sampling Procedures

     7.        Sample Custody

     8.        Calibration Procedures and Frequency

     9.        Analytical Procedures

    10.        Data Reduction,  Validation,  and Reporting

    11.        Internal Quality Control Checks

    12.        Performance and System Audits

    13.        Preventive Maintenance

    14.        Specific Routine Procedures Used to Asses Data Precision,
               Accuracy and Completeness

    15.        Corrective Action

    16.        Quality Assurance Reports to Management
Source:    U.S.  Environmental Protection Agency/Office of Monitoring Systems and
          Quality Assurance, Office of Research and Development, Washington,
          D.C.,  "Interim Guidelines and Specifications for Preparing Quality
          Assurance Project Plans," QAMS-005/80 (December 29, 1980)
    /k Arthur D. Little, Inc.


     A table or chart which shows the project organization and line authority
should be  included.  An example project organization chart is shown in Figure 9.
Key individuals who are responsible for ensuring the collection of valid
measurement data and the routine assessment of measurement systems for precision
and accuracy should be identified; the responsibilities of each individual
should also be delineated.  It should be noted that the QA coordinator should be
organizationally independent of the project organization so that the risk of
conflict of interest may be minim_ 'i.


     For each major measurement system, numerical QA objectives for accuracy,
precision and completeness should be established.  These objectives may be
generally based on previous experience in applying comparable procedures to
similar sample matrices as well as on the requirements of the program.  The QA
objectives for precision,  accuracy and completeness should be summarized in a
table or chart; an example of a summary table is shown in Table 16.

     All measurements should be made such that results are representative of the
media (e.g., waste feed,  stack emissions) and conditions being measured.  Any
factors considered within the experimental design to ensure representativeness
should be described.

     All data should be calculated and reported in units consistent with other
organizations reporting similar data to allow for comparability of data bases
among organizations.   Units for all measurement parameters should be specified.


     A detailed description of all sampling procedures to be used is an integral
part of any QAPP.   Where applicable,  the following information should be

     •    Description of techniques or guidelines used to select sampling sites.


                                                                  Section Manager
                                             Corporate Quality
                                             Assurance Officer
                                                                  Program Manager
                                             Quality Assurance
Quality Control
 Data Manager

                                    TABLE 16

                          AND COMPLETENESS OBJECTIVES3
                                               b             b                 c
            Parameter                 Precision      Accuracy      Completeness

 Flue  gas dioxin/furans                     ±          50-100            100
 (Modified Method  5)

 Ash dioxins/furans                         +          50-100            100

 Velocity/volumetric flow rate              +           +20             100
 (EPA  Method 1 & 2)

                                             f             e
 Fixed gases/molecular weight             ±10          +20             100
 (Modified EPA Method 3)

 Moisture (EPA Method 4)                  + 2Qd         + 10d            100
                                             f             e
 Flue  gas SO, (continuous monitor)        +20          +20             100

 Flue  gas temperature                    + 10°F        + 20°F            100

 Ash (NAA)                                 ND            ND              100

 Particulate Mass                         +10          ±12             100

 Flue  gas HC1 (1C Analysis)               +10          +10             100

 Flue  Gas Lead/Cadmium (AA/AAF, NAA)       ND            ND              100

 Flue  Gas Chromium/Nickel (AA/AAF,        +10          +10             100
 NAA,  Colorimetry)
.All objectives are expressed in terms of percent  (%).
 Precision and accuracy estimiated based on results of EPA collaborative  tests.
 .Valid data percentage of total tests conducted.
 Relative error (%) derived from audit analyses, where
               Measured Value - Actual Value   ,«„„
     Percent -	. „ n	 x 100%
                      Actual Value
 Coefficient of variation (CV) determined from daily analysis of a control
 sample where

     CV - Standard Deviation

 Percent difference for duplicate analyses, where

     o     ,.   First Value - Second Value   iriri.
     Percent - T—-—~	    , ,, -,   xx 100%
               0.5 (First + Second Values)

ND   -    not determined for this method

Source:   "Revised Sampling and Analytical Plan for the Marion County Solid
          Waste-to-Energy Facility Boiler Outlet Salem Oregon," EPA Contract No.
          68-02-4338, DCN: 86-222-124-02-05, September 16, 1986.


     •    Specific sampling procedures to be used (by reference in the case of
          standard procedures and by actual description of the entire procedure
          in the case of nonstandard procedures).

     •    Charts, flow diagrams, or tables delineating sampling program

     •    A description of containers, procedures,  reagents,  etc., used for
          sample collection, preservation, transport, and storage.

     •    Special conditions for the preparation of sampling equipment and
          containers to avoid sample contamination (e.g., containers for
          organics should be solvent-rinsed; containers for trace metals should
          be acid-rinsed).

     •    Sample preservation methods and holding times.

     •    Time considerations for shipping samples  promptly to the laboratory.

     •    Sample preparation (e.g., concentration,  dilution,  cleanup

     »    Forms, notebooks, and procedures to be used to record sample history,
          sampling conditions,  and analysis to be performed.


     It is essential that adequate chain-of-custody procedures be established
for each project.  The following sample custody procedures should be addressed
in the QA Project Plans:

          Field Sampling Operations:

     •    Documentation of procedures for preparation of reagents or supplies
          which become an integral part of the sample (e.g., filters and
          absorbing reagents).

     •    Procedures and forms for recording the exact location and specific
          consideration associated with sample acquisition.

     •    Documentation of specific sample preservation method

     •    Pre-prepared sample labels containing all information necessary for
          effective sample tracking.

     •    Standardized field tracking reporting forms to establish sample
          custody in the field prior to shipment.

     Laboratory Operations:

     •    Identification of the person responsible to act as sample custodian at
          the laboratory facility who is authorized to sign for incoming field
          samples, obtain documents of shipment (e.g., bill of lading number of
          mail receipt),  and verify the data entered into the sample custody

     •    Provision for a laboratory sample custody log consisting of serially
          numbered standard lab-tracking report sheets.

     •    Specification of laboratory sample custody procedures for sample
          handling, storage, and dispersement for analysis.


     For each critical measurement parameter, including all critical pollutant
measurement systems, the following information should be included:

     •    Calibration procedures and information.

     •    Applicable standard operating procedure  (SOP) or written description
          of the calibration procedure(s) to be used.

     •    Frequency planned for recalibration.

     •    Calibration standards to be used and their source(s), including
          traceability procedures ".nd verification of purity procedures.


     For each critical measurement parameter, the  applicable standard operating
procedure (SOP) should be referenced or a written description of the analytical
procedure(s) to be used provided. Officially approved EPA procedures should be
used when available and if applicable.


     For each critical measurement parameter, the  following items should be
briefly addressed:

     •    The data reduction scheme planned on collected data.

     •    All equations used to calculate the concentration or value of the
          measured parameter and the reporting units.

     •    The principal procedures that will be used to validate data integrity
          during collecting, transferring (if applicable), and reporting of

     •    The methods used to identify and treat outliers.

     •    The data flow or reporting scheme from collection of raw data through
          storage and validation of results.  A flowchart will usually be


     •    Key individuals who will handle the data in this reporting scheme.
          (If this has already been described under project organization and
          responsibilities, it need not be repeated here.)


     This section presents guidelines for the number and frequency of replicate
and spiked QC samples and calibration standards to be used, including
concentration of surrogate or spike compounds to be added to designated QC
samples.  The QA plan should document the objectives for number,  type,  and
frequency of QC samples.

     Quality control samples are analyzed in the same way as field samples and
interspersed with the field samples for analysis.  The results of analyzing the
QC samples are used to document the validity of data and to control the quality
of data within predetermined tolerance limits.  QC samples include blank
samples, analytical replicates, and spiked samples.

Blank Samples

     These samples are analyzed to assess possible contamination from the field
and/or laboratory, so that corrective measures may be taken, if necessary.
Blank samples include:

     •    Field Blanks--These blank samples are exposed to field and sampling
          conditions and analyzed to assess possible contamination from the
          field (a minimum of one for each type of sample to be collected and

     •    Method Blanks--These blank samples are prepared in the laboratory and
          are analyzed to assess possible laboratory contamination (a minimum of
          one for each lot of samples analyzed).

      *     Reagent and Solvent Blanks--These blanks  are prepared in the
           laboratory and analyzed  to  determine  the  background of each of  the
           reagents or solvents  used in  an  analysis  (a minimum of one for  each
           new  lot number of solvent or  reagent  used),

 Replicate  Samples.

      These samples are  analyzed in order to establish control and assess  the
 precision  of the  analytical methodology.   Replicate samples include:

      •     Field Replicates  -  These samples are  collected  in the  field and
           analyzed in order to  assess the  reproducibility of the  sampling
           program (a minimum of one for each sampling event per  sample type and
           measurement parameter).

      *     Laboratory Replicates -  These replicate samples are prepared in the
           laboratory in order to assess the reproducibility of  the  laboratory
           procedures used (a minimum  of one for each lot  of samples analyzed).

 In addition, replicate  analyses of specific samples  may be undertaken by  the
 analyst to check  on the  validity of any anomalous results.  Such  results  could
 be the result of  instrument or  data system malfunction, operator  error,
 laboratory contamination, etc.  Repeat  analyses of  the sample in  question and a.
 previous "normal"  sample  will serve to  indicate which of  the possible problems
 is, in fact, present.

 Spiked Samples

     Samples may be  spiked with one or  more selected surrogate compounds  prior
 to extraction and analysis.   "Surrogate" compounds are defined as species  that
 are chemically similar to the compounds being determined but that are not
 expected to be present in the samples (e.g.,  when GC/MS is the analytical method
 to be used, stable-isotope labelled analogs of  the compounds sought are
excellent surrogates).  The data on surrogate concentrations are used to

calculate surrogate compound recovery from each sample as one measure of the
accuracy (bias) of the sample preparation and analysis procedures.  To the
extent that an analytical method (e.g., GC/MS) is consistent with use of
surrogates,  this procedure allows recovery to be estimated for every sample at
trivial incremental cost to the testing. In some cases, e.g., trace metals,
GC/ECD analyses, it may be difficult to select an appropriate surrogate compound
that will mimic the behavior of the species sought but not lead to positive
interferences in the analysis.

     In addition to use of surrogate spiking (if possible), selected samples
should be spiked with target analytes at a predetermined concentration level(s).
This requires that each of the selected samples be carried through the entire
sample preparation and analysis procedure twice, once unspiked and once spiked.
Also, spiked blank samples for each measurement parameter should be analyzed in
order to assess the inherent accuracy (bias) of the analytical method. To ensure
that the per cent recovery of the spike can be determined with a reasonable  .
degree of confidence, the spiking level should be at least 2-3 times the
critical decision level (see Section II.A) and at least as high as the level
expected in the unspiked sample. Depending on the concentration of analyte in
the unspiked sample, these data may provide an estimate of the recovery of the
species of interest from the sample matrix.For difficult sample matrices,
multiple spiking levels may be used (method of standard additions).


     Each QAPP should describe the internal performance evaluation and technical
systems audits which are planned to monitor the capability and performance of
the system(s) to be used for obtaining critical measurements.

     The technical systems audit consists of an evaluation of all components of
the critical measurement systems to determine their proper selection and use.
This audit includes a careful evaluation of both field and laboratory quality
control procedures.   Systems audits are normally performed before or shortly
after systems are operational; however, such audits should be performed on a

regularly scheduled basis during the lifetime of the project of continuing
operation.  The on-site technical systems audit may be a requirement for formal
laboratory certification programs.

     After systems are operational and generating data, performance evaluation
audits are conducted periodically, as appropriate, to determine the bias of the
critical measurement system (s) or component parts thereof.  The plan should
include a schedule for conducting performance audits for each critical
measurement parameter, including a performance audit for all measurement
systems, should the nature of the work require that a performance audit be done.
As part of the performance audit process, laboratories may be required to
participate in the analysis of performance evaluation samples.


     The QA project plan for a trial burn should itemize the procedures for  .
preventive maintenance that are relevant to the sampling analysis and efforts
required in the project.   For example,  the following types of preventive
maintenance items should be considered and addressed in the QA Project Plan:

     •    A schedule of important preventive maintenance tasks that must be
          carried out to minimize downtime of the critical measurement systems.

     •    A list of any critical spare parts that should be on hand to minimize


     The data quality indicators (e.g., precision, bias, completeness, and
method detection limit (MDL)) should be routinely assessed for all critical
measurement parameters.  Specific procedures to assess precision, bias,
completeness, and MDL on a routine basis should be described in each QA Project
Plan, as applicable to the measurement parameter and the system being measured.
The QA Project Plan should also contain and discuss any statistical or
mathematical methods used to evaluate the measurement data.


     Corrective action procedures include the following elements and must be
described for each project:

     •    The predetermined limits for data acceptability beyond which
          corrective action is required.

     •    Procedures for corrective action.

     •    For each critical measurement system, identify the individual
          responsible for initiating the corrective action and also the
          individual responsible for approving the corrective action, if

Corrective actions may also be initiated as a result of other QA activities,

     t    Performance evaluation audits.

     9    Technical systems audits.

     •    Laboratory/interfield comparison studies.

A formal corrective action program is difficult to define for these QA
activities in advance and may be defined as the need arises.

     If long-term corrective action is necessary to eliminate the cause of
nonconfonnance,  the following closed-loop corrective action system may be used.
As appropriate,  the sample coordinator, analysis coordinator or the program
manager, ensures that each of these steps is followed:

     1.   The problem is defined.

     2.   Responsibility for investigating the problem is assigned.

     3.   The cause of the problem is investigated and determined.

     4.   A corrective action to eliminate the problem is determined.

     5.   Responsibility for implementing the corrective action is assigned and

     6.   The effectiveness of the corrective action is established and the
          correction implemented.

     7.   The fact that the corrective action has eliminated the problem is
          verified and documented.


     QA Project Plans should provide a mechanism for periodic reporting to
Management on the performance of critical measurement systems and data quality.
As a minimum, these reports should include:

     9    Changes to the QA Project Plan, if any.

     •    Limitations or constraints on the use or applicability of the data, if

     •    Quality programs, quality accomplishments, and status of corrective

     •    Results of QA systems and/or performance evaluation audits.

     »    Assessments of data quality in terms of precision, bias, completeness,
          representativeness, and comparability.

     •    Quality-related training.

The final report for each project must include a separate QA section that
documents the QA/QC activities that lend support to the credence of the data and
the validity of the conclusions.

                                VIII.  REFERENCES
1.   U.S. EPA, "Test Methods for Evaluating Solid Waste-Physical/Chemical
     Methods," SW-846, Third Edition (November 1982).

2.   U.S. EPA, "Interim Guidelines and Specifications  for Preparing Quality
     Assurance Project Plans," QAMS-005/80 (1980).

3.   Ozvacic, V., "A Review of Stack Sampling Methodology for PCDDS/PCDFS,"
     Chemosohere 15 1173-1178 (1986).

     Velzy,  C. 0.,  "ASME Standard Sampling and Analysis Methods for
     Dioxins/Furans," Chemosphere.  15,  1179-1185 (1986).

5.   American Society of Mechanical Engineers, Draft Protocols on "Sampling for
     the Determination of Chlorinated Compounds in Stack Emissions" and
     "Analytical Procedures to Assay Stack Effluent Samples," (December 1984).

6.   Radian Corporation, report to Massachusetts Department of Environmental
     Quality Engineering, "Final Emissions Test Report.  Dioxins/Furans and
     Total Organic Chlorides Emissions Testing.  Saugus Resource Recovery
     Facility.  Saugus, Massachusetts," (October 1986).

7.   Gallant, R. F.,  J. W. King, P. L.  Levins and J. F. Piecewicz,
     "Characterization of Sorbent Resins for Use in Environmental Sampling,"
     EPA-600/7-78-054, NTIS No.  PB-284347 (March 1978).

8.   Piecewicz, J.  W., J. C. Harris and P. L. Levins,  "Further Characterization
     of Sorbents for Environmental Sampling," EPA-600/7-79-216, NTIS No.
     PB80-118763 (September 1979).

9.   New York State Department of Environmental Conservation, "Emission Source
     Test Report, Sheridan Avenue RDF Plant,  'ANSWERS'," (1985).

                            VIII.   REFERENCES (cont.)
10.  Cianciarelli, D. J. and B. D.  Williams,  "A Summary of Source Sampling
     Methods Used to Evaluate the Flakt Pilot Plant at the Quebec City Mass
     Burning Incinerator," Environment Canada,  December 1986.

11.  Schlickenrieder, L. M.,  J. W.  Adams,  K.  E.  Thrun, "Modified Method 5 and
     Source Assessment Sampling System Operators Manual," EPA-600/8-85-003,  NTIS
     No. PB85-169878 (February 1985).

12.  Jungclaus,  G. A.,  P. G.  Gorman,  G. Vaughn,  G.  W.  Scheil,  F.  J.  Bergman,   L.
     D.  Johnson and D.  Friedman, "Development of a Volatile Organic Sampling
     Train (VOST)," Proceedings of the Ninth Annual Research Symposium on
     Incineration and Treatment of Hazardous Waste, EPA-600/9-84-015 (July

13.  Hansen,  E.  M., "Protocol for the Collection and Analysis of Volatile POHCS
     Using VOST," EPA-600/8-84-007, NTIS No.  PB 84-170042 (1984).

14.  Ross, R. W., F. C. Whitmore, R.  H. Vocqui,  T.  H.  Backhouse,  B.  M.
     Cottingham and R.  A. Carnes, "VOST Applications at the USEPA Combustion
     Research Facility," Proceedings of the Eleventh Annual Research Symposium
     on Incineration and Treatment of Hazardous Waste, EPA-600/9-85-028
     (September 1985).

15.  Harris,  J.  C., D.  J. Larsen, C.  E. Rechsteiner and K. E.  Thrun, "Sampling
     and Analysis Methods for Hazardous Waste Combustion (First Edition),"
     EPA-600/8-84-002,  NTIS No. PB84-155845 (February 1984).

16.  Bell, J. M., "Development of a Method for Sampling and Analyzing Refuse,"
     University Microfilms, Xerox University Microfilms, Ann Arbor,  MI (1963).

                            VIII.  REFERENCES (cont.)
17.  Churney, K. L.,  A. E. Ledford, Jr., S. S. Bruce and E. S. Domalski, "The
     Chlorine Content of Municipal Solid Wastes from Baltimore County, MD and
     Brooklyn, NY," U. S. Dept of Commerce, National Bureau of Standards,
     National Measurement Laboratory, NBSIR 85-3213 (1985).

18.  Domalski, E. S., K. L. Churney, A. E. Ledford, Jr. and S. S. Bruce,
     "Monitoring the Fate of Chlorine from MSW Sampling Through Combustion Part
     I:  Analysis of the Waste Stream for Chlorine," National Bureau of
     Standards, Chemical Thermodynamics Division, Center for Chemical Physics,
     Chemosphere. 15, Nos. 9-12, pp. 1339-1354 (1986).

19.  National Recovery Technologies, Inc. (NRT) DOE/SBIR Combustion Effects  -
     Study, Sommer, E. J., G. R. Kenny and J. A.  Kearley, "Municipal Solid Waste
     Profile Analyses at the Summer County Resource Authority for Week of
     October 1, 1984."

20.  Hasselriis, Floyd, "Refuse-derived Fuel Processing," Butterworth Publishers

21.  Cuiu,  C., R.  Halman, K.  Li, R. S.  Thomas, R. C. Lao, "Analytical Procedures
     to Assay Environmental Samples for PCDD/PCDF," Chemosphere.  15. pp. 1091-98
     (1986) and Stieglitz, L. ,  G. Zwick and W. Roth, "Investigation of Different
     Treatment Techniques for PCDD/PCDF in Flyash," ibid..  pp. 1135-40 (1986).

22.  Rappe, C., S.  Marklund,  L. Kjeller and M. Tyskland, "PCDDs and PCDFs in
     Emissions from Various Incinerators," Chemosphere, Vol. 15,  pp. 1213-17

                            VIII.  REFERENCES (cont.)
23.  Fairless, B.J., D. I. Bates, J. Hudson, R. D.  Kloepfer, T. T. Holloway,  D.
     A. Morey and T. Babb, "Procedures Used to Measure the Amount of 2, 3, 7,
     8-Tetrachlorodibenzo-p-dioxin in the Ambient Air Near a Superfund Site
     Cleanup Operation," Environ. Sci. Technol.,  21. pp.  550-555 (1987).

24*  Wagoner, D.  E., "Sampling and Analysis for Hazardous Waste Combustion,"  EPA
     Workshop Proceedings (1986).

25.  Butler, F.  E., J.  E. Knoll and M. R. Midgett,  "Chromium Analysis at a
     Ferrochrome  Smelter, a Chemical Plant and a Refractory Brick Plant," APCA
     Journal, 36,  581-584 (1986).

26.  Bonner, T.  A.  et al, "Engineering Handbook for Hazardous Waste
     Incineration," SW-889 (September 1981).

27.  Brenchley,  D.  L.,  C. D.  Turley and R. F. Yarmac,  "Industrial Source
     Sampling,"  Ann Arbor Science Publishers, Inc., Ann Arbor, Michigan (1973).

                              APPENDIX A

     Equipment and Procedures Used in Milwaukee Sampling Program


     During the twelve-month period beginning August 3, 1959, one collection a
week was made from each sample area.  The material picked up included all
garbage, ashes, combustible and n^combustible rubbish produced by the res -:ents
of the sample area.

     A special City crew collected the material and separated it during
collection into three categories:  ashes, combustibles, and noncombustibles.
Total weight, volume, density and moisture content determinations were made on
each of the three categories.  The percent that each component contributed  to
the total sample was then calculated.  In addition to these physical tests,
certain chemical analyses were made on the combustible and ash portions.  The
chemical tests included analyses for hydrogen, nitrogen, carbon, lipids,
potassium, and phosphorus.  Subsequent calculations for percent liquid content
and C/N were also made.  All tests were performed on each individual sample.


1.   A dump truck was outfitted with a specially constructed bed to provide
     separate compartments for the ash, noncombustibles, and combustibles
     portions of the combined refuse. An Allis-Chalmers forage harvester (shredder)
     complete with necessary appurtenances including a loading platform, a  25
     horsepower electric motor,  a discharge spout, and all required safety
     devices was used.
 Adapted from Reference 16.

2.   Six 1 1/2 cubic yard steel bins complete with castors and a removable end.

3.   A battery-operated fork lift.

4.   A 1000 pound net-capacity platform scale rebuilt to provide an accuracy of
     0.1 pounds.

6.   Three 55 gallon drums.

7.   A gram laboratory scale.

.8.   A forced air drying oven.

9.   A Wiley Mill.

10.  A sand splitter.

11.  Eleven drying pans.

12.  A hand tamp.

13.  A Parr oxygen-bomb calorimeter,

14.  Air-tight sample cans.

Sampling Procedure:

1.   Collected refuse from appropriate sampling area.  Separated material during
     collection into three separate components, ashes, noncombustibles,  and

2.   Trucked material to central testing point.  A corner of one of the cities
     incinerator  buildings was used as this location.

3.    Broke open the combustible packages and removed the unburnable items.   All
     combustible material was placed in the proper movable bin and all
     non-combustibles were thrown into their respective truck compartment.

4.    Processing of combustibles,

     a.    All movable bins containing combustible refuse were weighed
          and volume measurements were taken prior to running the material
          through the shredder.

     b.    The material was processed through the shredder four times.
          Two movable bins were placed under the discharge spout to halve the
          material each time it went through the machine.  By halving the refuse
          four times the samples was reduced to one-sixteenth of its original

     c.    The remaining material was quartered by a shovel, retaining about
          enough to fill a gallon container.

     d.    Weight measurements were taken of all the combustible refuse prior to
          disposal to determine the moisture lost during the shredding
          operation.                    »

     e.    The sample was transferred from the gallon container to a drying  pan
          and placed in the oven.

     f.    Weight measurements pertinent to moisture content calculations were

     g.    A sand splitter was used to quarter the sample after it was processed
          through a Wiley Mill where it was reduced to a maximum size of one
          millimeter.   The size of the final quarter was about 100 grams.

     h.    Fifty grams of this were sent for certain chemical analyses and the
          other 50 grams were returned to the drying oven.  It was necessary to

          redry the material to eliminate moisture added during the grinding

     i.    The calorimeter test was performed and the information recorded.

5.   Processing of Noncombustibles

     a.    Noncombustibles were placed in the proper metal bins, weighed,  and the
          volume determined.

     b.    This material was stored in an empty bin and separated periodically
          into three separate components - cans, bottles, and miscellaneous
          noncombustibles.  Data were recorded to enable calculation of the
          percent of total weight and volume represented by each component.  The
          material was then thrown away.

     c.    Once a week a grab sample was taken from the material produced by one
          classification.  This sample was taken from a different area each
          week, thereby each area was tested once every 10 weeks.

     d.    This sample was placed in a drying pan and then put into the oven.

     e.    Weight measurements pertinent to moisture content calculations were

6.   Processing of Ashes

     a.    Ashes were placed in the proper bin, weighed and volume measurements
          were taken.

     b.    The ashes were quartered by shovel,  retaining approximately a pint of

     c.    The pint  sample was placed in an air-tight container and stored until
          all such  ash samples for the week were placed therein.

d.   At the end of the week the sample container was emptied and the lumps
     were reduced with a. hand tamp.

e.   The material was quartered with the sand splitter, retaining about 2

f.   A moisture content was made by use of the drying oven and the
     pertinent weight measurements were recorded,

g.   The dried ash was reduced to a weight of about 400 grams with the sand

h.   Half of this was sent to Purdue for further analyses and half was
     returned to the drying oven.

i.   After a second drying period, a calorimeter test was performed.

              APPENDIX B



                      numtut ramou. rot UOOUIRC
                         rt&ATiMC MIA BMIM: SAMPLINC m
                   nnMioM •» cftxouiuw OACAMIC OMTOVM
               ui nucf nan MLUI mm OOUWMTIOM rum
     Tha  lafara*tlaa aa.  (ur»aca  aparatloa which chow I 4 a*. gathara4 4urlag
 (•  far  ttaca aalaalaa*  af  chtarlaata4 argaalc caaaaaaa1* !• a«Ma«rl«a4 la
   following  4ata far*.   It  will  ha Mta4 that acttMl  *•!»•• whtla aaapllag
 •a4arway la tha  far* •( c*fU« «f  «ctM«l  lit If clMfU U tlM
                        aatlMi I* r«qtM*t«4. It
            le* at  tlM  tt*» •< »••*  •*••  •••»!!•« U actually kat«f 4a«a.
 a>artaa«  katw«aa  tha •>r«ct*4 a^ actval valuaa aortM ••^•••§ ia a fvlAa
  t*  wkatlMC the  ayata*  waa «a«fatlaa ••raallf tidtla it «aa aalafl aaaylaJ.
  tka c*«BariaMi  au||««t*  tka af*t««  mmj  m»t  kava kaa* afaratlag •»t»*Hf
  lag aaayllBf. It My aa 4tf(lc«ilt ta Mllt»a4 t«a 4ata.

     It will  alaa ba a«ta4 that  aaoa af tha r*^«ilr«4 UlarMtlaa la ta ha
 ilaetaa' a*  naatly c*ntl>u*w(ly aa paaalala 4urlaf aaayll*g«  tkla la aa aa
  Mralt  Mtactlaa  af chaagaa.  IMaa4a4  a«4  U«a*«rt«at.  nMcb  nay hava
 :urta4 4urt*g  aaaallag.  Tha acc«irra«ca  or  ahtaaca af  a*y a«ch chaag** *• •
 :tor la tatafpratlag tha taavlta af tha aaayllag.

     ••cocalng  th*  aaaaatlat  lafara«ttam  ahall  hagla  at  laaat thraa haura
 fora •••»!!••  ta  wg«B **4 ha ca«tt«M4  tat at  laaat  aa howi altar aaayllag

     Tha  lallowlag  fan*  la Iata«4a4  ta accaapaay raaa't* •• *a«Kca aaaallag
  r«fn*a-ta-aaargy facllltlaa.
   Naa* a*4 a44raaa  af tha facility)
   N.«  ••<< l«l«pho««  nuabcr  of •!••(  f*n*r*l  ••••(•r  ar athar y>*r*oa ta
   contact  f«|ai4t*| aalaalo*  •••pilag rro(raai
I.   N«M  «n4  t«Uphon« auvbar  of aviciloa •••f>ll»| (••• >«n«|«rt

                                                                                     MM* aa4 talaah«aa  auM«r  at  pataoa [a«ao««lbl« for  Ukoratory aMlyala
                                                                                     af aaaylaai
                                                                                     Haa* aa4 talaahaaa Miahar af Mtaaialaglcal •OBltoil«( atatla* ••r*lcl»g
                                                                                     th« lacallty af tha alaatt
     Baalgaatta* af tha caafetatlaa ttala  actually •••»!••!

     	.	I
     »ata af aalaala* aaa»lla|>

     Tlaa Mhaft aaayltag atarta4i

     Tlaa whaa aaapllag taralaata4i
 • •  Iy»« af rafvaa aracaaalag ayataa  («.§., ••••-bum. tafuaa 4afl»a4 fual)i
 9.  Kwacttaa rafvaa acaaaratlaa a»tho4.  If apacaarlatai
10.  Tya« af furaaca (e.g..  Matarvall. tafractary-vall, hybr!4)i

                                                                                     la  auilllary  fital ragulacly flra4f  If aa. nhat  fuall

                                                                                     Bwalg* haat talaaaa rata (Matu/hr)t  	
                                                                                            •!••• aro4uctloa rata aa4 ca*41tloaai   	

                                                                                                *r. 	M'l. «^ C»* e«t«§rata4>
                                                                     Iba/hi  •

     fa«4 ft9f»rtl»u  far  which twit «•• 4ailgna4i

          Aaaga  •/  ha«Cl«|  »»lu«»       _ ta

          Ranga  af  •olvtuta content!    _ to

     -    kaoga  oj  a«h  content!         _ to

                                                               wt  X

                                                               wt  t
     ri«*aa  provide • cro««-«ect loncl dl«|f»a of the facility, preferably to
     • cele.  •bowing the ipcllel relctlonihtp between the  MJOI  eleaeiit* af the
                       •  ••      -•--  ------  <--j  ...,» ,„  ,h, f,mim: the

    o*4  reel4uc  renovel ayelen; ahepe el  Ik*  furnee«;  prlaury  e»4  aeco«4ery
    conbuetloo •!( porte;  the holler en4 It* flue ••• •••••!«•;  coot alower*!
    ••jar heat lianaler eurfacee; aeo*oa>lier; elr prehceter (If  epproprtete);
    •Ir  pollution control  eyeten;  I»4«ee4 4ieft fen;  e»4 •tack.   Indicate
    locotlona of  tenparatitre and fiooBiir* 4«t«ctor> •!••.

S.  DoccrlptlcM of tit* grot* ojr«toai

    -    (wf oil or i  __^	
         T»p« (•••-, t«eloroe«tl«i. rolUt. tra««ll«|. rotary)!

         •uaoor of st«f MctloM (If «f>f rof rl*ta)t  _

         Croto »r«o (ft ) (or
•.  •••crlptloo of tb« holler i
                 olMMloo* (ft)i  L

              of coot »low*r*i  	
                               (•ppro«laot« tlM*)l
         of cMbu*tlo« (e.g.. OBCOSO air, atar*«4 atr)i
II.   Owrfira and u«4arflra air

          Baacrlo* a«*ifo tut Ijr* •*
         Oaacrlba »o» total cooAoattoo air ao4 air aletrlbottoo la co«trolled
         Vailfy  that  air  *l»trlkutlo« ayataa* ara operating aa
                                                                                     19.  Tf»« of 4tafti
                                                                                         •aw is draft ragulataa:
                                                                                    20.  ftaecrlptleo of aalld neate feed lag and etokUg ayaten:

                                                                                         -    low la fending late controlled!

                                                                                         -    frequency and length of feed ren atrakei  	
21.  B«acrl»a tba overall  pleat  control  eyeteo  logic  (e.g.. vket  aaaaur
     ara «aa4 aa tko aaela for controlling firing ratet)t
                                                                                    II.   Stack height  (ft)i
                                                                                         •tack  4lan*tat  at  tap  (ft)l
23.  Tjpa of atr pollution control ayatani  	

24.  If alactroatattc praclpltatori

          •pacific collection area (ft'/IOOO ACrn)j

     -    Boalgn taaparatnra at Inlet (*f)i  	

              ir of In4epan4ant hue aectlanai  	
                                                                                             Which In4apon4ant  hue aactlona were In aarvlce oWlng enieelon

     -    bealgn partlculate loading at Inlati  	

                                    at antlett  	

     Kapplng frequencyi  	

II.  If fahrlc flltari

          rahrtc type	Valght 	Weave

                                                                                             lag cleaning o*tho4 an4 frequency i
                                                                                             Alr-to-cloth retlo (ACTN/ft')s
                                                                                             •oatgn praaaura 4rop acroea beg* (la., W.C.):

                                                                                             Dealgn gee tenperature et Inlet (*F)i  	

                                                                                             Total nunber of he|t:	

     Actual o«»bor *f  baga  io  **r*lc«  at tta« ol aaopllagt

     bug* in preaaura 4roff> acroaa  baga Aufiog taatlatgi  __
IIMcfe fl«a •'* eooatooaoto  *t*  regularly  o*aautaa't  Indicate locatte* *{
•miter t
                     IM t n*M>t
                   Unit Serve*
                  j tkia Hoolter
C«I|MM HOOD* 1eretur«a  cod  location*:   	*F, 	

                                              -     Attach  eue-uiry  record of  furnace  temperature o*aeuree»nta duii«g

           Attack atrip dkart near* *f taaawratura *t top af furnace, Im fra«t
           af  a cram twaaa - •<•« lh* aaqallag aarta* Ctl t«cot4« «r«
           f*r mmt» ck«a •••
           their l*c«ti*M)
41.  flw*  ga»
          Attach  atrtf durt nc*r«* (*' c««c«mtt«tt«M «f ths f«li*»ia§ !!«•
                     t««at*. aBawifAi *»*r tM •••pliBf p«ri*tfs  car boa
                   ,  carba* 41aalaa. *Ky|aa, altrafaa amlaa, auKvr aloxlaa.
          tacal hyaracaraaaa, watar *aa«<> ••* (&•• ••• fl«M
     -    Attack atrip cKart mcarJ lar la-atack aa«city a»ar Cha

42.  Uiclu4a • capr  af tha oaaratar'a lag far tha aarl«4 af tka taat.

I.I Ptiaclpjat   Stack ga»aa tkat say caatala cfclarlaatal
    avgaatc caapaiiaaa •(* ttltkataiia fraa tka •(•cfc walag •
    •••pliag ttala.  Tka aaalyta la ealiaetad la tka caapllag
    trata.  tfca c*»pvu»4« af lataraat *rt 4«t*ral*«4 ky
    a*l*aat aitracttaa falla«»a4 ky §aa cktaa>ata(f ayhy/aaaa
    aaactraaeapy (GC/MS).

l.l Aaal |cafcy I |f t  tfcla aatba* I* appllcakl* 1*' *»* 4atat-
    a>laatta>a af eklarlaatatf atgaalc caapana4« I* •tack ••!•-
    •!•••.  Tk* ••apltag trata la ** iaalgaad »k«t aaly tfca
    tatal aaanat *f aaefc eklactaata* argaaic ea«p»n«4 !• tka
    •tack aataala** «ay k* aatar*laa4«  T» ••(•, •*
    k>»* k*a« f«rf»r««4 «• ••••••tr>t« tkat tk«
                   ckl«rlaat«4 *i§a*lc c ••••«••• c*ll*«l.ai !•
                   *f ik* ••••Hag trala *cc*tal*ly •••crtk««
    tk« actual aartltla* •! aack la (ka ataek ••Icalaaa.  If
    aapatata aacta af tfca •••alias (**!• *i« aaalf*a4 ••>•»-
    acaly, tka iaia akanl4 ka lacliiaai aa4 aa aata4 aa la
    Sacklaa 1 kala*.  tka aasfliag akall ka c*aa«et«4 ky
    caaavtaat fataaaaal a«ya«laaea4 wick Ckia taat ptaca^nta
    ••4 cagalcaat af lairlcaclaa af tka aparatiaa •' *••
    pr««ctlka4 •••pliag Ctala *a4 coaatratata ml tka aaalytt-
    cat tcckalavaa fat cklatiaatad ar§aalc caaaawatfa, aapacl-
    ally P»»a ••< PC»f«.

    Tfca raafa af tka aaalyclcal *«tkaJ ..r ».  .,MH«a4 caaalaar-
    akly tk*a«fk caacaattatlaa mmilmt «tl.tiB..   fh, total  Matk»4
    •aaaltlvlty la alaa fcl|kly iap«ai«at aa tki  valuaa af  atack
    |aa •••pl«4 aa4 tka aatacttaa Halt of tk« «*alytlcal  flalak,
    Tfca aaat akall aataialaa far tkali •y*t«.  ih. .i.i.M.  aatact-
    afcla atack |aa caacaatrat laa f«r tfca cklaria«ta4 araaalc ca«-
    aanaaa af lataraat.  Tfca »lalaua ••••ct.kl.  atach «••  caaca*-
    ttatlaa akaiiM faaatally ka la tka as/a1 (••aograa/cvklc
    •ataii at lavac faaga.
    •atai  Tkla •«tk»4 •••»••• th«t all *f (k« caaaavaaa af
    lataiaat ata «allacta4 altkar *a tka 1*0-1 raala at ia
    wpatiaa* ••••!!•( tvala caap/aaaata.  Blae* tka *atk*4 at
    •ka araaaat tl*a ka* aat kaaa *ail4ata4 ia tka ftaaaaca af
    all tka atkac eaaaaaaata ptaaast fICI, klfk *r|aaic laaa)
    I* tka ataek a«l*al*a, tt la racaaa«aa«4 tkat aaataprlata
    ^•ality aaatral («C) ataaa ka «*»l*f«4 **tll avek ••!!••-
    claa kaa fcaaa aa*alataa.  Tkaaa QC atapa *ay Iacln4a tka
    ••• at a ka«k«p taala trap ar tka aadltlaa af a ra«i«««a-
    tatlva lakalad ataa4ar4 f ilatlafulakakla fr»a tka tataraal
    • taaaati aaai tat a.ttaatltatlaal ta tka flltat aai/ai tka
    SAB-t ta tka flaM prlat ta tka ctart af •••aliaf.  Tkaaa
    •(•»• bill aravlaa lafataatlaa aa paaalkla kt*aktkvaa|k af
    tka caaaaaaia af latataat*


• •cognlilut tk«t *04tftc*tlaa af tk« •*tk*4 a*y k* r«^»ilr«4
lor »p«clfic apallcattaaB. tka flacl r«p«rt af • t*«t vkata
ck>..|.« •!• •••• skall Iacln4«i  (I) tka *«*ct mo4t I icat laa;
(2) the r»ilo«»l« I"' th* aoal f lot loa; *ai (1) •• «»tla»t« of
tit* *ff«c( tfca •o«lllc»lio« Mill pfniiic* om tk* «•(••
    Orgvatc caaaanaaa atfcar tkaa tha ceapau«4a  af  iataraat  »ay
    latacfata ultfc tfca aaalyala.  Appcaatlata aaapla claaa-af
    •tap* akall ka p*tfar*aa.  Thr«y|k •!!  •!•)••  •( ••••!«
    kaaallag aa4 aaalyala, caia •kawld ka !•••• ta a*ai4  caatact
    af aa«plaa aai aatracta with ayathstlc  argaate aatatlala atkac
    tkaa aalytatraf laaratfcy tan* (tri*).   Aakaatva* «h«ul4 aat ka
    •••4 ta k*14 TPt* tlaat* o  114* (fcvt,  If *«e«*aary,  appra-
    fllata klaaka aaat ka taai  . *4 lakclcatlag aa4 •••Itag
    |**aaaa muft aat ka aaa4 aa tka ••apliag  trala.


    Piaclalaa aai acearaey ••••ur«a*at«  fc«*«  not yat  kaaa kaaa aa
    PCM aai PCDP natal tfcl* aatkaa.  Thaaa aaaavraaaata  ara
    •••4«4.  lavavar, tacavary afflclaaclaa far aaurc*  aaa>plaa
  '  aplkai wltk caapaaa4a bava taaga* lr»a  10 ta 1IOS.'*


    *.l   S»aplt«i Iua« i   Tba •«••»• r af aa*pllag ritaa auat k*
    •afftclaat ta ara«!4a •taiaal  ctatlatlcal aata *«4  la aa
    eaaa akatl ka laaa tkaa Cfccaa  
      C  -  ?ha aaaipia race-wary (I)
      •  -  tke allowable etack aalaeleea (ag/e>  )
                                                                              •lack Wai
      lia.plei  A - O.OJO eg; • - lOlj C * J0t| aad 0 -  I a|/ai
               •f - O.OS ,

                        ti«*n«t -"
                                                                                                                    Ce*d*«ae« t
                                                                                                                    Naele Cat It Idea
    • a«pll*)g  Tralai   Ik* trate caeelate ef aosile, probe, boated
parttculata  filter,  aed eerbaat eiadula followed by fear leplagere
(fig. I).  »r**t*fe« te e>ade let tke addlllea ml (I) • cycles* la
tke h«*t*4 filter b*i wk«» t«*tl»g •••rc«i **i(cl*( Iklgli c*«c««-
cr*tl««« *f  parttcwlat* ••ttar. (I) * l«rf« water trap b*twa*a the
M«t*4  filter mm* the eerba*t aeettl* far eteck geee* vttli kigk
•ol«lHr« c*me*t. »«4 O) »4te are «eetetae4 la tke lellewtag aect !•<••.
tke *e»«le  ekall  ke mmtm te tke epeclf Icet leee *f IP* Metke4
The **««le  »ey  ke *a4e ef elckel plated atalalaaa eteel,
    or koroi lllcata glata.
1.1.2  Probe

    the pteke  ehall ka lleei er ••«• ef T««, kerael llcata , or
fuart* glaee*   The liner er probe e*te*4a peet tka reteleleg »«t
late the etech.   A ta*petatere ce»trelle4 Jeckat pievl4aa pcetec-
tlo« af the  lifter or probo.  Ike lleer er peek* ahall ka efelpped
wltk a coaeectiat flttleg that la capakle ef fara)l*g e leak-fee*.
vacMuat-tlgkC  eeeaoetloa Mlthewt eeallag graaeaa.

I.I.I  £a«ple  Traatfar Llgaa (eptloaal)

    the eaaple traeafer lleea. If •••«o4. ahall b* heat traced,
heavy veiled Tri* (1.1 «• fl/l l«.| O.K. » 0.1 cm |l/l te.| veil)
with ceaaectlag  flttlega that are capable of foralag leak-free,
*acitu*-tlgkt ceeeectfeea vltheet H*lag eealleg gteaeee.  Tk* lie*
ekewld ke *e ekett ea peeelkle aei •«•( be •atatelaed et ltO*G.

1.1.4  tllter  II el it I

    lereelllcate  glaae, wltb a glaaa frit flltar avppert aei e
|laaa-te-glaae eeal or Iff* gaakat.  A twbbar |a*ket ahall eat be
ua*4.  the holoor 4*al|a ehall prevlec a poaltlve aaal agalaat
leakage tiom  th*  outald* er areviii I h« fllt«r.  The holder akatl
bt artachod  le.**dlat*ty at  tka ewtlat  ol the probe (er cycloee, If
                                                                                  el* *l the Uodlllad ttfeeattote-Saillfc lype
                                                                                  1 e»d I Caalala 100 erf Walei •'.
                                                                               ef 4 Coal a to* BOO'SOO Oiaiaa SIMca Bel
                                                                                 t le lha Ha«ia*>fltae*d Back-Up le Ike flaal* Catltldpa

                                                                           T) • Tfceiei*ca«plo lecell**

                                                                             fig.  1.  Modified W* Hethod 5 Train for Orf*n!c* S*-pIln|
                                                               Source:  Method*
                                                                                                                     lor A»*«a»in| Oi|*nlc«
                                                                                                   Soulc*! In C>pO*U[t Ev«lu«C|oa DIvltlOB Siulll*

    Tke  cycleae  •hall  be cea>atr«ct*4  *f  baroalllcat*  giaea  «ltk
         g  flttlaga  (list a>re capable  »f  f«rat«|  leak-free.
*«c*«»-t Ifkt  c**«j*ctl»*c *)itk*Mt  walajg aaallag (raaaaa.
1. 1.*  filter
    Tke baatlag  •yet**  •••t  k*  capable  ef  oialst alnle|  •  tee>p«-re-
ture *tm*m* tk*  Illter  keltfer f*«**  «   .*»•.  if «a*4) imliig  •••pl
1*1 »f 110*14 C  fl4l*IS*r}.  A  t**p*f«(«|> •••••  c«p«bl*  of
•••••trial ?*ap*r**«r* t* vlthl* 3*C <5.4 F>  *k«ll  fee {••tailed  ••
                           4  tfca filtac
                                                                                   4S.t *•>
                                                                                   Itr IB.I
     •• tk* •*rfc**t
                                   . c**llMi  !•  a  itsp, ak*l)
                      ffk* ••*•••(  ••4«la •••!! »•  vad* •( glc
vlth c*Mft*ctl«§ llt(taj|« tk*t *f*  ••!• t« !»(» l**k-lt**»
ti|kl *««!• •ltk*«t ••• *l  •••!••! gf««*«a {tl|*.  i Mi J).
•*•-! ttap •>«•» k* !• • vertical p**ttl««.  It |«  ai(*c«4a4 kf
c*ll-cyp« c**4«*»«i, *!•* *fl«Bt«4 **rtlc*llf. vltk clrcvlatl
c»l| v*t«|.  £•• *ajtatl*| tk« **vk**t •*4nl«  mttft  k«
<10 C {••§}.  «•• t«ap«r«t
       (tap k«tk
                           •«•! k*
                                            (• *••! tk* ••r»«»t-

                               tk«  vartlcal p*»ltla*
    rout *t ••!• ta)fl»g«r* «llk c*a*«cttii| fitting* abl* t* tar*
laak-lra*. vaCMua-tlgkt  •••!• «l(fc««t acalcat |taaa«« wk*a CCH-
••ct*4 t«g«tk«c, akall •• «*«4.  All (••t«f*r« «r* cf cfc*
6>B*«kHfg-Sattk •••!•• **iifl«4 kr raplactag tka t If wttk 1.1 cm
(I/a la.} 1» gl»«« tMk*  ••t«*il*| t* 1.) c» (1/1 !••) tt»m (fa
       •! tk* fleck.
1.1,1  Matarl«M tl«t««

    Tka ••!•(!•! ay«t«03 *k«ll caaatit ml • vacima g«u§«, • l«ak-
fr*« P«"P, tkai«BOia*«»» capable *f •••••(lag tanparatvia t* *itbla
3*C (-i*»). a 4ry gaa »)atar vltb I p«tea«t •cc»r»ef at th«
i«^ut<*4 •••plt>| i«t«, aa4 r*lat«a a^wlpataat, or a^ulvalvat.

7.1.10   «roa>at«r
                                                                                     .• CM
                                                                                     I/I In. I
                      or «lh«r b«ro»*t*r» capable el aieaewrliig
         Ic ft»»matm (0 wltkla 2.3 Hg (O.I 1*. Mg > ckall be
                                                                                                                oo. 4.r
                                                                                                                 la  1/4  i


' S^


1 °


                                                                                        or it«ooa
                                                            To c*f off •4>«rb«nt (ub« a«4 tk« «tlt*r ••
                                                        tto«« •! Ik* teal*.  If TfK* *ci*w c«aa«ctlea« «t«
                                                        •c(«» cap* ahalt •«
                                                                   Tkr«« JOO •!. ••!(••• •• . 0011AS*. «r •^

                                                                  .»  froM ••< Tt«««l«r H«« imgfc
                                                                                                                       of  aulll-
                                                               ctoat loagtli Ifcat t« coapatlklo «ltk Ch« llaar or *r«»o a«4 traoa
                                                               fac ll«a.

                                                                   • ••1*4 flltar koldar or practaaaaa1. vl4a-aoutk aaioar  glaaa
                                                               caatalaara wick Ife*-ll«a4 aciaM ca»a or Mrapaa4  !• kaaaaa  claaa4
                                                           •»  •»!••<•

                                                            tiiflo kaaa. Okaua >o4al ISO), or

                                                           .i  AU»iiiii» rat!
                                                               J.J.I  fiacl»»aa4 Matai Ca»

                                                                   To raeavar »aa4 alllca |al.

                                                               I.I.I  >t»claa»a4 Cra4aata4 Crltaaar. a .1 . .  2SO
                                                                   2)0 •!, «itk I •! (coovatloas. koroalllcata  |laaa.

                                                               1.2.9  tttutj jaaula »tora«a Co»taio»fi
                                                                              •••at (laaa kottlo* or claar (laaa kottlo*  vraf»a<
                                                               !• apa^va aiatallal, I L. vlth Tf£*-J*aa< acrav capa.
                                                                      — rtbarilaaa »aawa-*ajal
                                                                                                           AH or Ia.ul»ajaa|
                                                                   filer to uaa la th* ll«la. aaclt lot al filter* ahalt  •« auk-
                                                                      to pracl«anlo| and a ^uatlty coatroi (QC) coat ••ln«( 1»»
                                                               chack to conflra tkat thara ara no contaolnanta piaacnt tkat «ll

  t*rf*r* «itk tfc*
  • It*.
                          t*tf*t 4*t*cti*a
   If  p*rf*r*«4. flll«r pi*cl**«l*g *b*ll c***l*t *l g«ibl*t
 rr«cll*«, 1* k*lch** »*t *• ••£••4 SO filler*, nick tb* *»i-
)•((•)  t* b« *pplt«4 t* Ik* M*I4 **«pl**.  A* • QG ek*ck, tk*
ttr«ctl*g ••!•••«(•) *h*ll k* *Mkj*ct*4 t* tk* •*•* **»e**tt«-
 • *.  cl***«p **>4 ***ly*i* pr*c*4«t** t* k* n**4 f*r Ik* 11*14
>•*!*••  Tk* k*ekgr*«»4 »l fcl**h *•!«• *k**t**4 *k*ll k* «*•-
irt*4  (* * f*| it it II k**l* **4 *k*ll k* c*rr«cl*d fat **y
iff*r«*c** I* c**e«att*ti»« f*ct*i ••!«•••• tk* <)C ck*ck
1 4  tk*  *ctu*l •*•*!• ***ly*l* *(*c**i>(*
                                                      M*t*»                 t*tr*ct  Kith lo tot  I  hr

                                                      M«tkyi  *lc*k*l        litr.ct  f*r 12 hr

                                                      M*tkyl***  «,bi*ri4*    Bitr*ct  t*r 21 hr

                                                      •*•***                g»tr*ct  f*r 22 hr

                                                   Ik* IAD-1  •«*!  k*  4ri*4 ky *•* *f tk* tullawtmg t*ck*t>u**,

.,,.   .
                                                                           ilt*r •**l«*tl*>« «|
                                                                       • •l***t,  * f lMMi«*4-k*4 t*ck«l*,»* k**
           Cf - inltjjl
           *i *pjfl»«tiit, *olv**t
                                                                                                 t*  k*  tk*  t«*t«*t
                                                       A  ...pi.  ..!.«  ...» *.lt.kl. ....
                                                   •111 **r»*  ** *  ••ti*(*et*ry c*lu>*.   A 10.2  cm (4  !•,) «l*«*t*c
                                                   ffyr*«  pip*  •.§ • (2  ft.  la«f) wltt  k*14 *lt «f  tk*  1AB-2 ft** tk*
                                                   •••kl*t  ••tv*ct*(. Mltk  *«fflci**t  *p*c* |*r  flut*l«tag tk* k«4
                                                   Mktl*  t***r*tl*| « •!•&•!•• IAK-2  l»*i  «t tk*  *«lt  *l  tk* c* !•>•>*.
                                c**c**tc*t*4 *iti*ct
    Tb* *.u**tit*tt*« crlt*ri*M f*r *cc*pt*kl« Iilt*r *.**Uty will
 •p**4 •• tk* **t«etl** Halt crttcrt* «*t*kl<«k*4 f*f tk* 11*14
 ••pli*g *»4 ***l»«i* pt*g»*«-  fllt*t* tk*t gl«* * k*ckgc*w*4 *r
 Immk  *lg**l per fllt*t gr*«t*r tk** *r •«•*! t* Ik* t*r|*t 4*t*c-
 I** Halt fat tk* •••lyt*(*> *f e**e«t* *k*ll k* i*J*ct*4 f*r
 1*14  M*«.  •*!« tk*t *cc*pt**c* «tlt*fl* f*f flltcc cl***ll«**«
 *p*«4* **t a*ly •• tk* l*k*t**t ••tcetl** Halt *f tk* •**ly*l*
 i*tk*4 »ut *!*» •• tk* ••p*ct*4 fl*14 •••pi* **iu»* *»4 *• tk*
 >**ti*4 Halt *f 4*t*eti*» I* tk* **api*4 *tt*«a.

    If tk* filter* 4* **t p*** tk* QC ck*ck, tk*y *k*ll k* r«-
 iBtr*ct*4 **4 tk* **l»**t *»tr*el« r*-***ly««4 N*til *• *ce*pl*kly
 LOM k*ckgr*»*4 i***l i* *cki***4.

 1.1.2   A*»»rilt« IAB-2 •••!»

    fk* cl***up pr*c«4ur* »*f k* c*rfi*4 *ut In * |t«*t •••ki*t
 i>ti*ct«r. wkick Hill c**t*l« *»evgk 4ak*rlit* IAD-2* r**i*
 (1AO-J) f*r **v*r*l •••pll*| trap*.  A* *il gl**» tbi»kl* SJ-tO ••
 ID i ISO •* 4**p t>*ct*t
cup Mltk * gi*** N**l plug *>4 *t*l*l*** *t**l *er*** *i>c* It
fl»*t* D* **l:kyl*** ckl*rt4*.  Tkl* pcec*** i**«l««* **f«**ti*l
           1* tfc* following *t4*i.
                                                       fli*  >(«  u**4  t*  !*•*•*  tk* *al**«t  !•  tk*  b*y t* pr***r«l*g
                                                   tk*  cl*«*ll***« of tk*  140-2.   Ltfuii *itK*g«* fr«t« • t«gul*f
                                                   c**«*ict*l  lt*«14 *ltio(**  cyli***r  k** i*«t(B*ly *i*.«« t* k* *
                                                   r*lf*kl*  **«rc* *f l*rg«  *«U*««  «f  g** lr*«  fro* *tg*«ic c**c*«l-
                                                   •••!*.   tk*  It*.«14 *ltt*g«* cyll*4«r 1* c****ct*4 t* tk* c*lu** ky
                                                   *  l**gtk  *f  pt*cl****4  O.fJ «• (1/g  i*.) c*pt*i tukt*g. c*li*4 t*
                                                   p.**  tki*ngk * k**t  **«ic*. A* *ltr*g*« I* kl*4 (**• tk* cylia-
                                                   4*r,  it  i* **p*tt**4 !• tk* k«*t  *«urc« **4 p***** tki*«tgk tk*
                                                   C*|M*B.   A c**«**l**t •«*•  •*«rc« 1* •  v*t*t  k*tk k**t«4 <*•• *
                                                   •!*•• !!•*.   Tk*  fl**l  *lti*g*« t*»p*r*tnt« *ki»ul4 **ly ** mmtm to
                                                   t*«  t*»ck *«4 »*t ***r  40 C.   •«p«rl«*c* k**  *koH* tk*t *k*«t SOO
                                                   |  *f  i**-2 «*y k* 4rt*4 •••(•tght c***ti«t*( *  full l»0 t c*ii*4*c
                                                   ml n,w|4 *itt*g**.
                                                      ** * *«C**4 ck*ic*, kigk  pttrity  t**h
                                                  4ry tk* SAt-2.  Tk* kigk  purity  •!((«••• mumt  flrct  k* p****«
                                                  tkr*«gb * k*4 •* «ctiv*t*4 ck*rc**l  *ppr*stB*t «ty  1)0 uL  tm
                                                  «*!••*.  Witfc *ltk*r  typ* of  4tyl*g  »«tko4.  tk*  r*t*  of  fl*«
                                                  *k**l4 |**tty *glt*t*  tk* k*4.   IBC***!**  flul4«tlo*  m»r  c**** tk*
                                                  psctlcl** t* kr**k up.

                                                  (•) AI •• *lt*t**t*.  if tk* *ltr*g** prac*** 1*  not  •**ll»kl*> tk«
                                                  lAt-2 »*y k* 4(1*4 i*  • «*cuu* «»«*. if tk*  t«*p*t*t«f«  **««c
                                                  *BC*«4* 20*C.

                                                      fk* IAB-2, *«*• if putch**«d el«««. marnt k*  cb«cli*4  tot  koch
                                                  n*thyl*** ckl*(t4* **4 k***** t«*l4n**. plu* •«!••! *!**••  »«lor*
       5 o I » ta I
Initial lint* with 1
                                            L H^O
                                                   for 1 cycl*.
                                                      Storage of CJ»»» IAI>-2;  IAD-I cl««n«4 mad drl.d  •
                                                  *»««* t* *Hlt*bl* for l***4lal« u«« 1* th* field,  jiowliti It
                                                  p«««** th* QC co»t ••ln«l torn check dticilfacd In (4), b*lo».  *
                                                  miimi . •t*cl**n*4 dry IAB-2 •*? d*««la* u*«cc*ptabl* !«>•)* *f

                                                       If praelaaaa4 1*0-1 ta not ta k« u««4  l»«41at*tf,  *»  akall
                                                   •tara4 aadat 4tat 1 1 l*4-la-glaaa aiatkaaal.   HQ  .„!•  lh*n t»
                                                   prlar ta taftlatlaa af II»I4 •••fling,  ch« ••<••• *atka*al akatl
                                                   ka 4acaat*4( tka IAD- 2 she 11 k« •••k«i  Mlth •  •••!! ••!<•** *f
                                                   ••tkfl«i>« ckivtii* «»4 4rl«4 vltk «!••• nltr*t«« •• 4«ccrlfe«i «•
                                                   (M ah***.  *• •lif«*t phall (»•• k« tain** lor tha  QC  c«»ta»l«»-
                                                   tt*« chaclt 4aacil»a4 !• (4). kalav.

                                                       II Ik* atarad IAB-1 falla tk« QC ckack. It mmf  ka  raclaaaa4 »;
                                                   ra»aaci*| tka fl*ai t«a ataaa af tka aatiaetiaa aa««aa>c* ak««ai
                                                   ••^•••tial a)*tkpl*«a cblarMa aai kaaaaa ••trsctlaa.   Tka fC
                                                   caataalaattaa ekacfc akall ka tapaataJ aftac tka I*»-l  I* r«cl*aa*4
                                                       QC Caataalaatlaa Ckacki  Tka IAB-2,  vkatkac  »»refc«aa4. "pva-
                                                   claaaa«"*, ar ciaaaa4 •• tfaacrlkai *k**a. akall k«  ankjaetai te a
                                                   QC ckack la eaaftr* tka akaaaca af aay caataal*aata  tkat atlgkt
                                                   cawaa latarlaiaac** la tka aakaatMaat  aaalyala af  fl«14 •«*»!*•.
                                                   AM all^aat al lift- 1, aqaivalaat im alia  ta  aaa flat* ••*»ti*.f taka
                                                   ckarfa* akall ka lakaa ta ckaractatlta a >l»«la  katck *f 11»-1.

                                                       Tka IAB-I allfnat akall ka aakjactal ta tka  a«»a •mtiaetla*.
                                                   caacaatrattaa. claaawa. aai •••Iptlcal ptac«4ttr a(al  *• ta fata) te
                                                   ka afaliaf t* tk* flala* aaapla*.  Tka  a,«*«ti tat I va cittacla f*i
                                                   aceaatakla IAB-I ^aalltf  '11 4apaai aa  tk* aatactiaa Halt ett-
                                                   tarla aatakllaka4 fat t*.   ial< aa*ptlai *•* aaalyala »raft>«.
                                                   •AB-t wkiek ftalia a kackt(aaa4 at klaak alfaal  gtaatar. tkaa ar
                                                   *4«al ta tkat cat raapaia t •( ta a*a-kalf  tka MDL  far  tka aaalft«(«)
                                                   *f caacai* akall ka rajactatf far ftala* aaa .  H«ta  tkat tk* accapt-
                                                   aaca ll*it fat &A*-2 claaallaaaa iapaa^a aat ••!>  a« tk* lakataat
                                                   4atacclaa ll*lt af tka analytical «atka4 k«t alaa  ** tk* aa»acc«4
                                                   fla!4 aaapla ••!«•* aa4 *a tka 4a*ica4 Halt af  4atacti*a la tka
                                                   aaaplad atraa*.

                                                   t.l.l Ciaaa Maal

                                                       Ctaaaa4 ky tkarait|k rlaalni, l.a«. ••^uaatlal  taavralae la
                                                   tkraa aliqvata af kaxaaa, itlai 1* a  I IO*C  avaa. mm4 atata4 la. a
                                                   ka«*aja-«a*k*4 glaaa jar >ltk IPI*-tlaai  aciav cap.
                                                   •.!,* Hatat
    Bal**l*a4, tkaa |laa*-ilat Iliad, a«4 atarai
      caacataara vltk Tfl*>ll*aJ acraa capa.
                                                   • . 1.5  Ulff 6fi
fig. 4   1AD-I
    Ia41eatlag typa, *-l* aiaak.  If pra*la«alf «*a4, 4ry at
f*t t ki •  •*• atllca gal «af ka «a«4 aa tac«i*«4.

•.l.i Cfftkaj jca

    flaca ciwakai lea la i><* aatar katk *caua4 tka l*plagata
Jar lag aaapllag,

 PaatlcI4a Duality. attt4lck  and Jackaaa  ~Dlatllla4 la Claaa'
tvalaat. at«ra4 la original  caatalaara.   A  klaak •«•( ka
aaaa4 km tka analytical 4atactlaa aatko4.
                                                               • f
   ••atlcl4a Duality,  Iur41ck aaa1 Jackaaa *Bt*tlll«4 !• Claaa
 itvalaat.  atara* la arlglaal caataiaara.  A klaftk •«•( ka
  ••••4 ky  tka analytical 4atactl*n aatka4.

             Sactlaaa 10. J.I. I mm* 10. 1. 1. J shall ka 4ooa I* tka
             lakat atary .
 •' •*   ftataat  rraaaratlaa

   All  train coaipaaaata akall ka •atatalo«4 aa4 callkrata4
 cor4lag to tka araca4ura 4aacllka* 1* AfTO-OS}* ••!••• atkarwlaa
 actfla* karata.

   Ualgk aavaral 200 ta 100 t p»rtlo»» of alllca |«l !• alr-ti|kt
 •*•!••(• to tka oaataot O.J (.  K«coi4 tka total waifkt o| tka
 llea  |«1 alua cootalaat, •• aack co>talo«r.  A* •• altaraatl««.
 • ctltca gal ••} »• Mal|ka4 41r«ctly !• It* la»l«|at or ••••llag
 >14or  jitat ptior t« trala
    •alact • aoisla alca kaa*4 aa tka raa(a of valoclty kaa4a
•uck tkat It ta aat aacaaaary t* ckaaga tka oattla al«a !• or4«r
to aalatala laaklaattc •••attaf rataa.  »«.(!.« tka roa  4o aat
ckaaga tka aa»Ia alaa.  laaura tkat tka prapar 4Itfaraatlal
praaaura fauaa la ckaaaa for tka raaga af valaclty kaa4a
aaca««atata4 (aaa tactlaa 1.1 at IPA Hatko4 1).

    •alact a avltakla aroka laagtk auck tkat all travaraa polata
cao. ka aa«pla4.  Par larga atacha. caaa!4ar aa.pllag fro. a.poalta
a!4aa af tka atack ta ra4»ca tka laagtk af arakaa.

    • alact a tatal aaapllag tla>a graatar tkao or a^ual to tka
• !•!•••• tatal aaaipllag ttaa apactfla4 la tka taat araca4uraa for
tk« apaclflc ta4«atry a«ck tkat (I) tka aaaaltag ttaa par polat la
••t laaa tkaa 1 •!•., a«4 (1) tka aaapla valuaa takaa (carracta4
ta ataa4ar4 caa41tlaaa) vlll aicaa4 tka ra«y|ra4 •lala.ua tatal gaa
aa*pla valuaa 4ataralaa4 la (actlaa t.l.  Tka lattar la kaaa4 a a.
aa apftaBtaata avaraga aaapllag rata.

    It ta racaa>aa4a4 tkat tka anakar af atautaa aaapla4 at aack
polat ka a* latagar ac a* latagat pUa aaa-kalf alavta. la ai4ar
ta aval* tl«a-kaaplag arrora.
   Ckack Illtar* •lavally a|*loat ll|kt (o( I r cafular I t la«
 aw* or »lnk«la loaka.  faek tka Illtara flat IM a »raclaa*a4
iaaa coataiaar or wra^aad ba*aaa-rlaaa4 alualauai fall*

1. 1.1. 1  riaUataary Datgra)iaat iona

   (•lact tka aaa>alla.f alta ««4 tka olnlaua auak*( af •••plla|
atata accar4laf ta «FA Natko4  I.  DatarBlaa tka atack praaaura.
taiDaratura. aaa1 tka raa|a al valaclty ka*4i uato| cr* Matka4 1)
t  la i«eoo»«»4a4 tkat a laak-ckack of tko pltat liaaa (aaa irA
ttko4 I. Sac. 3*1) ka aarfaraia4.  Oatarailaa tka •olatura coataat
•!•( EPA Appro>l«at lea Hatko4  4 or Ita altaraactvaa far tka
itrpoaa of aaklai laoklaatlc aanpllnl r al «-•« t tlaga .  bicarala*
K« atack gaa 4ry aelacular wal|kt. •• 4*«crlka4 la EPA Hatko4 2.
• c . )••; If lata|ral*4 EPA Hatko4 3 »aaiplla| la ua*4 for aolacu-
ar w*l|ht 4a(ar«laat loa. tha lata|rac«4 ka| aaapla akall ka takan
I mu 1 1 *n«ou« I y wltk. aa4 tor tka aaaia total laagtk of tla« aa . tka
f>A M.cl.od 4 aaapl »»| •
                                                                       All glaaa parta af tka trala upatraaaj of aa4 lMelw4lng tka
                                                                   aarkaat aia^Mla aa4 tka llrat laplagar akau!4 ka claaaa4 aa
                                                                   4aacrlka4 !• lactlaa JA af tka 1*10 laaua af "Haaual af Aaalytlcal
                                                                   Natka4a far tka Aaalyala af raatlc!4aa la laajaaa aa4 Eavl raaaaat al
                                                                   Ia«plai.*   Ipaclal car* ako«14 ka 4a«ota4 to tka raaaval at raal-
                                                                   4«al alllcaaa graaaa aaalaata aa gioua4 glaaa caaaactlaaa af naa4
                                                                   llaaavara.  Tkaaa graaa* ra*14uaa akou!4 ka raao*a4 ky aaaklag
                                                                   aa»aral kaura la a ckro»lc ac!4 claaatag aalvtlaa artar ta raattaa
                                                                   claaalag aa 4aacdka4 akava.

                                                                   10. 1. 1.1  Aaifcttitti IAD-1 taala Trap

                                                                       Haa a auftlclaat aaouat (at laaat 10 gaia or S •••/•' of atack
                                                                   gaa ta ka aaa>pla4) af claaaa4 EAK-1 to fill coaplataly tka glaaa
                                                                   aackaat trap wklck kaa kaaa tkatougkly claaa«4 aa araacrlka4 aa4
                                                                   rla.a*4 vttk kaaaaa.  tallav tka IAA-2 wltk k«Maa«-r I*«a4 gla*a
                                                                   waal aa4 cap katk aa4a .  Tkaaa capa a haul 4 aot ka raajo«o4 aattl
                                                                   tka trap ta fltta4 lata tka trala.  Saa Pig. I l«(  4»talla.

                                                                       Tka 41aaaaloaa aa4 IAO-2 capacity af tka aorkaat  trap.  aa4 tka
                                                                   • oluaa af gaa ta ka aaaipla4. akou!4 ba *arla4 •• aacaaaary  ta
                                                                   aaaura afftclaat collactloa of tka apaclaa o< lataraat.  Coaa
                                                                   1 lluat latlva 4ata ara praaantad la takl« I.

                                                                   10.1.2  Praaaratloa of Collact|qi» Ttaln

                                                                       During praparatlon an4 aaacaibly ot  ih« •••pllng train,  keep
                                                                   • 11 trala opaa.ln|* what* cont aal n«l 1 oa. can antar co«
•» !• •«
 • is s
o    «•»    si
  o «•»  ••<•
            • .  *  •  •» •
                               .   2
                              3   1
•4   •
•8   f
s   5
!   =
S   M
-4   .•
1   S
•9   M
             M    J
 !  3
                                                            Jn*t prior to •••••bly or until ••••!!•( !• about to k«gl».
                                                                      Bo o.ot u«« »««t««t tc«>»«i j« •••«»blt«i th«
                                                           •••r*ila«t«ly  100 •••  of  wator  i*  ••ch  of  tko  flrot  two
                                                 l«ol«tor«  vltli  •  |t««u*t«4  cfliooor,  ••<  !••*•  the third  lopl«|«r
                                                 o«ftT.  Ploco •••roilaotolf 100  to  300  g  or  «or«. If ••c*«*icf.  of
                                                 • Illc*  |«t  !• th«
                                                                                                                  oo ttt« o«t«

                                                                *•••••!• tli* ttklo •• *li*Ma !• Fig. I.

                                                                •loco ctM*lio4 Ico to tko Motor both (round tho

                                                            10. I.)  took Choct rrocooaroo

                                                            10. 1. 1.1  JiitUi Loofc Cfcocfc
                                                    Tho  trolB.  iBcluoloi  tho  orooo.  will  ho  look  chockod  prior  to
                                                 ho log  locortod  lot*  tho  otock oftor  tho •••pllttg  trslo  hoo  hoao
                                                 •o*««hlo4.  fora  OB  oo4  oot  (If  oppllcohlo)  tho hoot log/coollog
                                                 o«otoa(a)  to  cool  tho  ooaplo  goo yot  roaolo  ot o  tovporotoro
                                                 oofflcloot  t* o*o!4  cooo«««otloo I*  tho >roho o>4 cooooctlog  lino
                                                 to  tho flrol  loplogtr  ( opproilaottly  120  C).  Alloo  tlao  for  tho
                                                 toaporotoro to  otohlllso.   Look  chock  tho trolo ot  tho  ooospllog
                                                 •Ito hf  plogglog  tho ootilo  «lth o Tff* plvg o«4  polllag  o  110  ••
                                                 •g  (II !•.  Ig)  «ocy«o).   A  lookogo.roto la ovcooo  of  41  of tho
                                                 ovorogo  ooBpllag  roto  or  O.OOJ7  B /B!B (0.02 elm) ohlchovor lo
                                                 looo,  to BBoccoptohlo.   Sooollog Bwot  coooo  If ptoooor* 4ortog
                                                 ooBpllag omcoo4o  tho look  chock  prooovro.

                                                    Tho  follotilag  look chock  lootrvctloa  for tho  ooapltag trol«
                                                 4oocrlbo4  la  AW-OSIt'  oo4  AfT»-OJil* Boy  ho holpfol.  ftort tho
                                                 p«ap with  hfpooo  »ol»o fwlly  opoa oa4  cooroo o4Juot  «olvo coo-
                                                 plotoljr  clooo4.   tortlotly  afoa  tho  cooroo  o4juot *«lvo oo4 olovly
                                                 clooo  tho  hypooo  «ol*o «atll  HO BB  Ig (12  la. Mg)  VOCMMB lo
                                                 roocho4.   >o  apt  rovoroo  tho  4lroctloa of tho hypooo  «*lvo.  Thlo
                                                 •til coooo ••tor  to  hock  «p  tato tho  probe.  If 3(0 BB  Ig (12 to.
                                                 •g) to o«coo4o4 4vrlog tho  toot, olthor look chock ot thlo  hlghor
                                                 VOCOOB or  oa4 tho  look chock  oo  4oicrlho4 holov oa4 otort tho toot

                                                    Hhoa tho  look  chock  lo  coBploto4.  flrot  olovly  roaovo tho TH*
                                                 plog froB  tho lalot  to tho  probo thoo  laBo4lotoly turo  off  tho
                                                 •OCBMB p«Bp.  Thlo pro*oat*  tho  «otor  ta  tho laplagoro  froB bolog
                                                 forco4 hock«or4 tato tho  probo.

                                          Look  Chocfco Ourtoj  o To»t

                                                    A  look chock  oholl ho  porforao4  hoforo oa4 oftor  o  choogo ol
                                                 port 4«rlag • toot.  A look chock oholl ho porforao4  hoforo oo4
                                                 oftor  •  coBpoaoat  (•••-,  flltor  or optloool Motor kaockout  tf«f)
                                                 to  choago4 4«rf*g  o  t«»t.

    Suck laak ckacka akall ha parfor««4 accof4lag ta tha pioc*aur*
 1*** I* tvctlaa 10.I.I.I af tbla Bathe* a»capt that It ahall ba
 irl*r«*4 at * »«tuu« aa.ua 1 t* *r graataf tha* tha blgbaat valiia
 >cor4«4 up ta that polat 1* tba taat.  If tba laakag* rata la
 >«*4 1* ba •* graatai tha* 0.000)1 *>/•!• (0.02 ft /«!*) *r 41 af
 ia  ***raga aaapllag tat* (Mklchavar |a aaallar) tha taault* ara
 scaptabla.  If. havavai, a hlghar laabaga rata la •baar»a4> tba
 i*t*r  ahall althari  (1) r*car4 tha laakaga rata a*4 tka* correct
 i*  valu*« *f gaa •••pl«4 alnca th* laat laak ckack aa akawa 1*
 ictlaa »f tbta •*th*J, ar (2) *a!4 tha taat.  >«a|-taat Laak Cbach

    A laak chack ta •aaa'atatjr at tha a*4 af a taat.  Thla laak
 hack akall ba parfar«a4 1* *cc*r4a*c« with tha ptaca4ura gt«a* 1*
 •ctloa aacapt that It ahall ba caa4uct«4 at * **cuu*
 a,ual to ar graatar tha* th* blghaat ••!«•• racar4*4 4urlng tha
 aat.  If tha l*akaga fata 1* Iaun4 t* ba *• graatai tba* 0.000)7
  /•la  (0.02 ft /•!•) ar 41 af tba avaraga aa*pllag tat* (whlck-
 •ar 1* an»llar). th* raculta ara acc*pt*bl*.  If, havavar. a
 Ighar  l*aka|* t*t* la *kaar«a4, th* taatar ahall althart  (1}
 •cor4  tha laakaga vat* aa4 carract tb* valuta aa gaa aa*pia4
 Inca tba laat laak chack *a abav* 1* lactla* It.1.1.4 af tbla
 •tho4, *r (1) »a!4 th* taat.  Corr*ctlnt far t«c«»fl»« 1-aakaga iitat

    Tha a^uatl** glvan In Sacttaa 11.3 af tbta •«tk«4 far calcv-
 atlag  V (*t4)( tb* c*rcact*4 *alu«* «f gaa •••pta4, ca* b* u««4
>a urlttan u*taaa tha taakag* rata *baar*a4 4«rlag a*y laak chaci
tftar tha atari af a taat a«c*«4«4 L^, tha •••!•»• accaptabla
i**kaga rata (aa* 4afl*ltl**a kalav).  If a* *kaar*a4 laakaga r*ta
i>caa4a t , tka* raplaca W  1* tb* aa.«atlaa t* gactt** II.1 «llh
ih*  f011«fll*g atpraaalani

    W  - Volun* of gaa •••»ta4 aa **aaura4 by tka 4ry gaa
         natar (4act).

    L  - Ma«l*u* accaptakla laakaga tata a^ual t» O.OOOS? •/•!•.
     *    .« «, ltj/.l«) ar 41 *| tk.
                                                                      *, * gaaatfag tlaa latar««t bctwaaa tit« aucc*»«l«a  laafc ckacka
                                                                           kagt**l*g ultk tka l*tat*al katwaaa tka  (lot  aa4 aaca*4
                                                                           laak ckacka, •!*.

                                                                      •  • Sa«fll*g tla>a latarval katvaaa th« laat  (a  tk) laak ckack
                                                                           a«4 tka a*i af tka taat, *la

                                                                  • ukatltuta aaly lat tkaaa laakag**

                                                                  10. I. 1.1  Tr»«« Oaaf«tla*
                                                                                                         at tf)  uklck  amc.a4a4
     wkicka»ai la aaallac.
L  -
                             tka aoafaga aa*alt*g fata,

                  okaatva4 4urln| ih« poat-taat laak ck*ch.
     •wrlag  tka  taajallag  tua,  a a**plt*g  rata ultkl*  101  of  tka
 aalacta4  aa*allag  tata akall  ka •al*talaa4.  ilata Mill ka  caa-
 a!4aia4 accaalakla If  taa41aga ara racar4a4 at  laaat a«*rf  J •!*.
 a*4  aat atata tkaa  101  af  tka  a»lat taa41aga ata  ta aacaaa  af *|OS
 aa4  tka avaraga *f tba f«l«t  raarftaga  la  wltkla  £101.  ftuflag'tka
 rwa.  If It  kacanaa aacaaaatf  ta cbaaga aay ayataat eanaaaaat la aay
 • ait  af tka trala. • laak ckack eiuat  ka  §arf*raa4 artar  t*

     far aacfc rua,  tacor4  tka  4ata  r*^ult«4 on  tba 4ata akaata.  aa
 amaa.pl* la  akan* 1* ilg.  4.   ta swca  ta  facor4  tka  laltlal  4rp gaa
 •atar  raa4l*g.   gac*t4 tka 4ff gaa «atac  taa4t*ga at  tka kaglaalag
 a*4  a»4 af  aack a**pli*g  ttata lacr«*«at  a*4 «kaa •••fling  la

    f* kagla •a«pll*g, raaova tka  *a««la  c«p.  varlfy  flf appllc-
 akla)  tkat  tka  ptaba aa4  a«rka*t *a4ula  taaparatura  caatral •»•-
 taaa  ara  warklag aa4 at ta*par*tHra aa4  that tba praba la praparlf
 paattla*a4.  raaltlaa tka  »raka at tka aa*pll*g  palat.   Ia*a41-
 ataljp  atert tba puap a*i  *4j«at Iba flaw  rala.

    If tba  atacb la ua4*r  alg*lflca*t  auk-anklaat praaaura  (kalgkt
 af taipl*gar ataa),  taka cara  t* claaa  tha caaraa a4|uat  ••!••
 baf*r* l*aartlag tba praka iat* tba atack ta avo!4 «atar backlag
 1st*  tba  ptaba.  If *acaaa«rjr,  tba pu»p *ay ba t«r*a4 a* vltfc  tka
 caaraa a4J*«t *al*« cl*aa4.

    Ourlag  tba  ta*t rua,  *aka  parlo41c a4}uat*aata ta baap  tba
 prafca ta*paratnrc  «t tb« pro»«r  valua.  «44 aara tea ••«, If
 •acaaaary.  aalt t* tba lea bath.   Alto. parl«41cally cback  tba
 l«»al «a4 aar*  al  tba »aaa»atar  aa4 *al*tala th* ta*parat«ra *f
 aarbaat *a4ula  at  ar laaa tba*  20  C but akava 0°C.

    If tka  praaaura 4rap acr«aa  Ika trala b«cea*a kljb aaougk to
• •ka tha  aa«pll*g  rata 41fftcult to ••latala,  ttta test rua ckall
 ba  tarat**ta4 u»l«»« tba  r«p^acln| of   th* lllt«r coiracta tk*
 proklaa.  If  tka ftltar !• raplac<4. • l**k  ch.ck akall  k«
 par f »(*«d .
t  - Lc*k*|*
                           4»rlo|  tht I**fc ck«ck r*t
                                    '* "  1,1.1.-.*).
    At tba a»4 ol CK« •••pic run, turn off th* pu»p,  r**o»* ih«
Bioba an4 noiil* fro* th* »t»ck. end f*cot4 th* Mnal  4ij  («•
•atat raa41ng.  f«i
                                                                       IO.Z  leaole >eco»«ry

                                                                           fraper cleanup prece4ura ke|lna ••  aeon  ••  eh*  probe to
                                                                       removed frea tke etack at lh« ead of tke  eaapllaf period.

                                                                           Wkea tka prate cea ka aafety kaadled.  Mlpe  off  all eateraal
                                                                       partlcalat* e>etter aear tk« tif »l th«  prok*.   Icao** tk«  pt*k«
                                                                       !(•• eh* tt»l« ••< el*»a •!( kath ••«•  «lth  b«n«ii*-il««a4  •!••!•«
                                                                       1*11.  •••! »ir tk« !•!•( (• tka trsla  wick  • |io««4 (!•••  cay «r
                                                                                          •«• 1*11.
Till* •»••
                                                           Hl< k« cl**B
                                                            •* !••!•( ck*
                                                                                                     •ncl*««4 •• tkat tk«  ck«ac«*  •(
                                                                                                     •pl« vlll k« •lnl
            tk« trcla prl«r t» *a4 Jurlng 4l«a«««akly and  *ot«
         c«««l«l«»», •.•.. broke* flitcci, color  •( tko  l«ol«|«r
        »te.  T«oot tko ••••!•• •• lollovoi

•O.J.I  fo«t>l»or PP. I

    Kltk«r coal tko •••• of tk« flltor ko!4or  or  corol«tly roaovo
tko flltor troa tko flltor koloor oo4 oloco 1C to It* I«oaclflo4
coacoloor.  ••• • pair of •rocloaoao' t*oo*ora  to  kao4la  tfco
flltor.  If K la aacaaaary to fold tko ftltor, oo ao oyek tkat
tko tortlcalato coko la I cat 4* tko fola.  Carofullr traoafor to
tk« coatalaar aoy acrtlCHlat* oiattat ••a/or ftltar flkor4  vklck
•akaio to tko ftltar koloor gaakat . kf «al«| a ary loort krlatla
kraak a«4/or a akarp-a4|o4 kla4a.  Soal tka cootaloar.

10. 1. 1  torkoat Mo4yloa

    •••••• tko oorkoot Bo4iilo fro* tko tralo ao4  caf  It  off.

10. I. 3  Crcloao Cotck

    If tko oftlooal cyclooa la «ao4. onaiit Itat I wolf racovor tko
partlcMlata lato o aaaola contoioor ao4 cop.

10.1.4  ianalo Coatalaot jo. I

    O.«oot Itatlvoly rocovor •atotlcl 4opoalto4  la  tko  ooicla.
praka. traaofar ilaa, tka fraat kalf at tko flltor bolder. oa4 tka
cyclaao. If »a«4. flrat by broaklaa, aa4 tkao by eee,iteot lal ly
rlaalaf wltk acotoaa aa4 tkaa koaoao tkrao tla*a  aack aa4  a44 all
tkaao tlaaoa to Coatalaor Mo. 2.  Nark loval at Honl4 aa  coa-

• 0.2.5  «aa»la Cootalmr Ho. j

    Klaao Cka kack kalf of tko filter bolder,  tke cooaoctlef llee
kotvooa tka filter oa4 tke coaaoaeer *a4 tke coo4eeeer (If oaieg
tka aeparato ceae'eaaer-aorbeat trap) tbree tlaea  aeck «ltk acetone

 • ana collacttag all rtnaaa  In Caatataar  ].   If uaiag tka
 • 4 coa4a»aar-aarkaat trap,  th*  rt*aa  af  tka  caa4aaaar «h«ll
 faia*4 I* tka i*k*v*l*iy altar  i*****l el tka IAO-1.  II tka
 >al watar kaeekaut trap kaa  kaaa «aploy«4. It ahall  »•
 >4 aa4 racai4a4 aa.4 It* eaataata piac«4 1* C»«t«t««v J (long
 h« rlaaaa *( It.  •!••*  II  tfctaa »!••• aack  wltk acataaa,
 >••••.  Mark !•••! *l l(«.u>4 mm  ceatalaar.
    aaatt. Cfaf » r Ho. i
                                «ff  th«
i»ova tka lltat
t«r t* !•«*•«
!•  chxt.  P»ut the caattftt*
!•••• 4lt«ctly !•<* G«»t*ta«V ••* 4.   •!••• tk« laplagat
itlallf thr«» (!••• with •c«toa«> ••<  k«»B*.  Mark t«**t ml
I  *• t»«l«l»«i.

I   S«i»t« CoiH»l»«r Ho. i

                                                                                - Bara»atrl« araaauta at tka •••plin(  ati«, •• Hg
                                                                                  C ta. Mg) .

                                                                                - AaaaUta ataek gaa acaaaura,  aa Hg (la. Mgi.

                                                                                " StaMaid aaaaluta **aaaura.  7*0 .. M(  |if.t3 |..
                                                                                  Hg) .

                                                                                - I4«al gaa  caaataat. O.Oallt •• Hg-B1/8^-.-.,!.
                                                                                  fll.ll I.. »g-U^*g-la-«ala),         *

                                                                                - Akaalnta avaraga 4ry gaa  a>atat ta«patatur* *I  C*»).

                                                                                • AaaoUta avarag* atack gaa ta*paratiira *t C°«).

                                                                                * Icaa4ai4 akaalvta taaaaratma.  Jtl°I (i«°f).
4 ««t|kt *• 4*c« «k««t.  laptf tk< «••*••!• *»i !!••*• !*(•
!••( Ma. S.  •!••• aaefc vttfc 4t*tlli«4 •! ««C«r thra« ttmmm,
                                                                                * fatal  aaaa  al
                                                                                  •lllca gal.
                           » taluaa al gaa aaapl
                             4c* (4cl).
                                                                                                      callacta4  la  laatagaia  a«i
                                                                                                          •aa*ur«4  ay  4tjr  ga*  aatar.
       tka laat laplagar. wlp* tka a«ta!4a ta  raaava aaeaaalva
  ••4 atkar 4akrla, walgk (ataa  Iaclu4«4). a«d t*cor4 watgkt
 ta akaat.  Maca tk* alllca fat lata tea aaik«4 caatataar.


 arry ant ealculatlaaa. ratalaiag at laaat ana a*tra 4aclaal
 a kayaa4 tkat al tka •*a,«t«a*> 4ata.  •aua4 oil Itgntaa altar


;        * fetal walgkt ol ckl«rlaata4 argaalc  coapou*4> la
 *        atack gaa aaapla. ag.

:        * Coacaatralloa af cklartaata4 argaalc caaa«wa4a la
 *        atach gaa, Mi'» • «*rtfcta4 ta ataa4ar4 co*4tttaaa
          •I 20*C, l*d *« *f (*• r, 19.92 la.  Hg) •• 4ry

l  *     - Cioaa-aacttonal araa af aoc*la, •  (ft ).
I        - U«t*r vapor la tka gaa •((*••, propoitloa by volo»«.

I        - r*tcant ol lanklnatlc  •••pi lag.

                •.. »«l«ht of u»t«r. II t/f-malt (It
                                                                      *B(»l4)  •  Valuaa  af  gaa  aaapla •aaa«ra4  ky  tha 4ry gaa aiatar
                                                                                 catract«4  ta ataa4ar4  caa41ttaaa,  4*ca  (4acl).

                                                                      *Mfati)  -  Valuaa  af  vatar  vapor  ta tka gaa  aaapla cartacta4 ta
                                                                                 ataa4a(4 caa41tfaaa, «ca (act).

                                                                      »,       "  St«ck gaa  valaclty, ealc«lata4 kjr  caaktiatlaa caicu-
                                                                                 lailaa. a./aac  flt/aac).





                           - Matat ka> carractlaa lact«v.

                           - Avataga araaauta 4Ilfataatlal acroaa tka arlllca
                             •atar, •• HjO (ta. ijO).

                           * fatal aaapllag tl*a, ala.

                           • tpactltc gravity al a«rcur*.

                           * lac/ala.

                           - Caavaralaa ta parcaat.
                                                                        Avaraga Pry Gaj Jittar T«ap«r«tuc«a»4 Ayar»g« Otlllea
                                                                        fraayura Drop

                                                                      Saa 4ata fkaat (rig. S).

II.1  Oil C».
    C.rt.ct tk. •••?!• *•!••• ••••ur*4 ky Ik* dry f.« B.t.r t.
•t.«4.(4 C..4III... IIO'C. '40 M «( <»•*/. 1>.»1 I.. HC» by
M.l.| EqM.tl*. I .
                                     jj                M
             f (•!') • » »  T.t4  fk«r * II.* " "l*.  %.t * TYT"
                        "                      ~ —
                           '•      .14

    B( - O.J.J) *«/•• Hg f.r ..trie «.lt.

       • 17.*) *B/I.. Hg far B.gll.k ••!(•

II.*  y.ju.t .1 M.t.r t...t
              „<««, -

    K. -  0.00114 .'/.I f.r ..tele

       -  0.0471 ft /.I f*r t.gll.k ..It.

•!(••• !•
                     « *r« yr*»«t In Ik* ••• «tr««a •••••• the
             ••(•r«<«4 «*4 ••• • pcycht •••ttlc ckatc C« •bt*la
             1 -
       - 0.0034)4 •• •(

       • 0.00144* !• If

      O«c»«tr«tio» a
                           J/«I - *K f«r mutrte <••!(•

                          tt'/al - *l t»t B«|ll*k ••!(•

                                   Orj««lc C>«>»u«4» !• tt«ck
                                *f ckl«il««t«4
                                                                           IS.  QHtLITf ASSUDAHCI (QA) r.OCCBUKEC
                                                                                                                          If  M.kl...  ...
                                                                               Tk. 11.14 kl..k.  .k,.!4  k.  ..k.ltt.4  ..
                                                                           c.ll.ct.4 .«  ..ck Mr»lc«l.r (..
                         S£PT.  II,  ltS4
                         0«. it. 19S4

Scope and Applicability of Method
The analytical procedures described here an applicable for the
rmlnatlon of polychlorlnated dlbenzo-p-dloxli«(PCOO) and dlbenzo-
ns(PCOF) In stack effluents from combustion processes.  These  methods
also applicable to residual combustion products such as bottom and
tpltator ash.  Ihe methods presented entail addition of Isotoplcally-
led Internal standards to all samples In known quantities, extraction
he sample with appropriate organic solvents, preliminary fractional Ion
cleanup of the extracts using a sequence of liquid chromatography
mns. and analysts of the processed extract for PCDO and PCOF .using
led gas chromatography • mass spectrometry (GC-MS).  Various
ormance criteria are specified herein which the analytical data
 satisfy for quality assurance purposes.  These represent mlnU
eria which must be Incorporated Into any program In which PCOO
PCDf are determined In combustion product samples.
    The toxicological data which arc available for  the  PCDO and PCOF arc
far from complete.  That Is. the toxicological properties of all of
the Isomcrs comprising the IS possible PCOO and IIS possible PCOF arc
not presently knoMi.  However, a considerable body  of toxkologlcal
data exists for which indicates that.  In certain animal
species, this compound Is lethal at extraordinarily low does and causes
• wide range of systemic affects. Including hepatic disorders. carcinoma
and birth defects.  Millc much less data Is available regarding the
toxicology of sufficient data Is available to fora the
basis for the belief that Is stiller  In Its lexicological
properties to 2.3.7.B-TCOO.  Relatively little Is known about the tornI-
cology of the higher chlorinated PCOD and PCDF (that Is. pent* through
octachlortnated rtOO/KOf). although there Is sow  data to suggest that
certain penU-. hexe-. and hepU- PCOO/PCOF tsomtrs are haiardous.  In
view of the extraordinary toxicIty of  and  In view of the
exceptional biological activity of this compound (on the basis of enzjmo
Induction assays I and of compounds having similar  oolecular structures.
••tensive precautions are required to preclude exposure U personnel
during handling and analysis of materials containing these compounds and to
prevent contamination of the laboratory. Specific safety and handling
procedures which art recommended are given In the Appendix to this protocol.
The method presented here does not yield definitive Information on
concentration of individual PCDD/PCOF Honors, except for
achlorodlbenio-p-dloxin (TCOO) »M 2,3.7,0-Tetrachlorodtbenzofuran
F).  Rather. U Is designed to Indicate the total concentration of
tsoeers of several chlorinated classes of PCOO/PCDF (that Is. total
a-, penta-. hexa-. hepta-. and octachlorinated dibenio-p-dloxins and
niofurans).  Of tho 75 separate PCDO and 135 PCOF tloners, there
22 ICOO. 38 1COF. 14 PeCOO. » PtCDF. 10 IUCOO. 16 HxCOF. 2 HpCOO.
COF. I OCOO and I OCDf.*

The analytical ncthod presented herein Is Intended to be applicable
determining PCDO/PCDF present In combustion products at the ppt to
level, tout the sensitivity which can ultimately be achieved for a
n sanple will depend upon the types and concentrations of other chemical
ounds in the sanple.

The method described here must be  Implemented by or under the   •
rvlsion of chemists with experience  in handling super toxic materials
analyses should only be performed  in rigorously controlled, limited
ss laboratories.  The quantltatlon of PCOO/PCOf should be accomplished
 by analysts experienced in utilizing capillary-column gas chromatogrephy-
 spectrometry  to accomplish quantltatioo of chlorocarbons and similar
ounds at very  low concentration.
    The abbreviations which are used to designate  chlorinated  dlbenzo-p-
    dloxlns and dlbonufurans throughout this document arc  as follows:

    PCOO - Any or til of the 75 possible chlorinated dtbenzo-p-dtoxin Isomers

    PCOF - Any or all of the 135 possible chlorinated dlbenzofuran Isomers

    TCOO - Any or all of the 22 possible tetrachlorlnated dibenzo-p-dtoxtn  Isomers

    TCDF - Any or all of the 138 possible tetrachlortnated  dlbenzofuran  Isomers

    PeCOO - Any or all of the 14 possible pentachlorlnated  dibenzo-p-dloxln Isomers

    PeCOF • Any or all of the 28 possible pentachlorlnated  dlbenzofuran  Isomers

    HxCOO - Any or all of the 10 possible hexachlorlMted dlbenzo p-dloxln  Isomers

    NICDf - Any or all of the 16 possible  hexachlorinated  dibensofuran  Isomers

    HpCOO - Any or all of the 2 possible heptachlortnated dlbenzo-p-o'ioxln  Isomers

    NpCOF - Any or all of the 4 possible heptachlorlnated dibenzofuran Isomers

    OCOO - Octachlorodibenzo-p-dtoxtn

    OCDf - Octachlorodibenzofuran

    Specific Isooers. - Any of the abbreviations cited above nay  be converted to
    designate  a specific Isomer by Indicating the exact  positions (carbon •!<>•*)
    where chlorines are located within the molecule.   For exanple, 2.3,7,8 ICOO
    refers to only one of the 22 possible ICOO Isoners -  that iso*er which  Is
    chlorinated in the 2.3.7,6 positions of the ditcn/o p-dioxin  ring structure.

    Reagents and  Chemicals
    Ibe fallowing reagents and chemicals are appropriate for use in these
 rocedures.   |n all cases, equivalent materials from other suppliers
 ay also be used.

    2.1  Potassium Hydroxide. Anhydrous, firanular Sodium Sulfate and
 4itfwic Acid (all Reagent trade):  J. I. laker Cfceaiical Co, or fisher
 clentlf ic Co.  Ike granular sodium tulfate Is purified prior to use
 iy placing a  beaker containing the  sodium sulfate In a 400 C oven for
 our hours, then  removing the healer and allowing It to cool In a desiccator.
 .tore the purified sodium sulfate Im • bottle equipped Mlth a leflon-
 itned screw cap.
    2.2  Hexane. He thy lane Chloride, lenteM. Methanol, Toluene.
Isooctane:  "Oittilled  I* fless* fcirdick and Jackson.
    2.3  TrJdecane (Reagent firade):  Sigma Chemical to.
    2.4  taste Alumina  {Activity trade 1, 100 • 200 mesh):  ICN
Pharmaceuticals.   Immediately prior  to use. the alumina Is activated by
beating for at least  li hour* at fttTC IN a ewfrie furnace and then
allowed to cool  f* a desiccator for  at least 30 minutes prior to use.
Store pre-conditioned alumina In a desiccator.

    2.S  Silica  (Ito-Sll A. 100/200 mesh);  lio-Rad.  The following
procedure Is reconmended for conditioning the fto-SIl A prior to use.
•lace an appropriate quantity of Jlo-SIl Atna30omx3QOmm long
glass tube (the  silica  gel  Is bald In place by glass wool plugs) which  <
Is placed In a tube furnace.  The glass tube Is connected to a pre-
purificd nitrogen  cylinder, through  a series of four traps (stainless
steel tubes. 1.0 cm 0.0. •  10 at long}":  1) trap Mo. 1 - Mixture
comprised of Chrooasarb H/AM (60/M  nesh coated with SS Aptcion I).
Graphite (UCP-J-IOO). Activated Carbon (SO to 200 nesh) In a I:1.S:1.S
ratio (Chromosorb  U/AW. Apleion I obtained from Supelco, Inc.. Graphite
obtained fro* Ultracarbon Corporation. 100 mesh. l-N-US»i Activated
Carbon obtained  fro* Fisher Scientific Co.)j 2) trap No. 2 - Molecular
Sieve 13 I (60/00  e»sb), Supelco, Inc.;  3) Irap Ho. 3 - Carbosieve S*
(80/100 mesh), obtained  from Supelco. Inc.; 4} Ihe tio-Sil A Is heated
in the tube for  30 minutes  at ltO*C  while purging with nitrogen (flow
rate 50 100 at/minute), and the tube Is then removed from the furnace
and allowed to cool to  room temperature.  Nethamtl (US mi) Is then
passed through the tube, followed by US ni methyl CM chloride.  Ihe
tube containing  the silica  is then returned to the furnace, the nitrogen
purge Is *g*ln established  (&0-100 ni flow) and the tube is heated at
SQ°C (or 10 minutes,  then the temperature Is gradually Increased to
in,»Sr .....r ?s Minutes »nd maintained at 180 C for SO minutes.  Heating
                                      — -- i« mntlnued until the tube
. cools to room temperature,  finally. the silica Is transferred to a clean,
  dry.  glass bottle and capped witb a teflon-lined screw cap for.storage.

     2.f  Silica tcl  Impregnated Utth Sulfuric Acid:  Concentrated sulfurlc
  acid  (44 g) Is combined wit* 100  g lio-Sil A (conditioned as  described
  above)  In a screw capped bottle and agitated to mix thoroughly.   Aggre-
  gates are dispersed with a stirring rod until a uniform mixture Is obtained.
  Ibe M2S04-slltce gel is stored In a screw-capped bottle (teflon-lined  cap).

     2.1  Silica Sel Impregnated with Sodium Hydroxide:  IN Sodium
  hydroxide (39  gl Is combined with 100 g Ifo-Sil A (conditioned as
  described above) In a screw capped bottle and agitated to mix throughly.
  Aggregates  are dispersed with a stirring rod until  a uniform  mixture Is
  obtained.   The KaOH-sflfca gel  Is stored In a screw-capped bottle
  (leflon-llned cap).

     2.1  Carbon/Cellte;
 Carbon:   AI-21 Carbon. Anderson Development Co., Adrian. Nlch. 49221
 Celtte S4S:  fisher  Scientific  Co

     Combine AI-21 Carbon (10.1 g) with CeMte S4S (124  g)  In a
 2SO ml glass bottle  fitted with a  Teflon-lined  cap.  Agitate  the  mixture
 to combine thoroughly.  Store In  the screw-capped bottle.

     2.9   Sepralyte Olol (40»):  Analytichem International

    2.10 nitrogen and Hydrogen  (Ultra High Purity):  M*theson Scientific

1.  Apparatus and Haterla Is

    1h*  following appiratus and materials are appropriate for use in these
procedures.   In all cases, equivalent Items from other suppliers may
also be  used.

    3.1  filassware used In the analytical procedures (including the
Soxhlet  apparatus and disposable bottles) Is cleaned by rinsing successively
three  times with methanol  and then three times with  methylcne  chloride,
and finally drying It in  a  100 C oven.   Bottles  cleaned in this manner
are allowed to cool  to room temperature  and  are  then  capped using  teflon-
lined  lids.  Teflon cap liners are rinsed as J«t described but  arc
allowed to air-dry.   More  rigorous  cleaning  of some glassware  with
detergent may be required  prior  to  the solvent rinses,  for  example, the
glassware employed for Soxhlet extraction of  simples.

                                                                             4.   Instrumentation
    3.1.1  Staple Vessels:   12* ML •ml 2SO el flint glass bottles fitted
with screw caps and teflon cap liners. and glass test lubes. VWH-Sclentlf Ic.
    3.1.2  leflon Cap Liners:  Scientific Specialities Service. Inc.
    3.1.3  Soxhlet Apparatus:  Extraction apparatus. Alllhn condenser.
Kimax Brand. American Scientific Products Cat. No. £6252-2A.
    3.1.4  Gravity Flow Liquid Chroootographlc Columns:  Custom
Fabricated (Details of the coliMis an provided IN later sections).

    3.1.S  Hlcro-vlalt (3.0 at):  fccltance CUM.
    3.2  Capillary (as Chroma tographic Columns:  TMD different coliMns art
required If data on both and 2.3.7,8-10*. as well as on
tbe total PCOO/PCOF by chlorinated class.art desired.  Ibe appropriate
coliMns are:  1)  A fused silica col MM (60 M x 0.2S mm 1.0.) coated
with DB-5 (0.2S v fllai thickness). JIM Scientific. Inc.. Rancho Cordovi.
Calif..  Is utlliied to separate each of the several tttra-through
octachlorlnatcd CDOs and COfs. as • froup. fro* all of the other groups.
While this col MM docs not resolve all of the I some.* within each
chlorinated group. It effectively resolves each of the chlorinated
groups from all el the other chlorinated groups and therefore provides
data on the total concentration of each group (that Is. total tctra-.
penta-. hexa-. hepta- and ocle CDDs and COfs).  Ibis column also
resolves from all of the other 21 1COO isomers and this
isomer can therefore no determined quantitatively If proper calibration
procedures are applied as described further in a later section.  This
coluMi does not completely resolve 2.3,7.1-ICDF fro* the other TCOF
tsomers. and  If a peak corresponding In retention time to
is observed in the analysis using this column, then a portion of the
staple extract nust.be reanalyzed using the second £C column described
below if Isomer • specific data on is desired.  2)  A
fused silica  column  (30 N x 0.2S am 1.0.) coated with 08-22S (0.25 »
filet thickness).  JIM Scientific.  Inc., Rancho Cordova. Calif.,  wist be
uttilled to obtain quantitative data on the concentration of 2,3.7.8-
1COF. since this column adequately resolves fro* the other
TCOf Isomers.

    3.3  Balance:  Analytical la lance, reidlbility. 0.0001 g.

    3.4  Nitrogen Slowdown Concentration Apparatus:  N-Evap Analytic*)
Evaporator Model  III. OrganoMtion Associates  Inc.
    Gas Chromatograph-Nass Spectrometer-Data  System  (GC/HS/OS):  The
instrument system used to analyie staple extracts  for rCDO/rCDF
comprises a gas chromatograph (fitted for ctplllary  columns) coupled
directly or through an enrichment device to a miss spectrometer which Is
•quipped with a computer-based data system.  The  Individual components
•f the GC/NS/OS are described below.

    4.1  Gas Chromatograph (GC):  The chromatogrtph must be equipped
with an appropriate Injector and pneumatic system to permit use of  the
specified glass or fused silica capillary columns.   It oust also  Incor-
porate an oven which can be heated In a reproducible, programmed
temperature cycle.  The Injector should be configured for splitless/
split Injections.  The GC column performance  should  be verified at  the
beginning of each 8 hour work period or at the beginning of coch  scries
of analyses If more than one set of samples is analysed during an 8
hour shift.  Extracts of complex combustion products and effluents  may
contain numerous organic residues even after  application of the exten-
sive prefractlonatlon/cleanup procedures specified In this method.
Those residues may result In serious deviation of GC column perfor-
mance and therefore, frequent performance checks  are desirable.   Using
appropriate calibration mixtures, as described below,  the  retention
time windows for each chlorinated class of COOs/COfs must be  verified.
In addition, the GC colu*   illlied must be demonstrated to effectively
separate fro* .11 other TCOO tsomers  If data on  2,3,7.8-
TCOO alone Is desired with at least 201 valley definition between the
2.3.7,8- Isomer and the other adjacent-elutlng TCOO  isomers.   Typically.
capillary column peak widths (at half-maximum peak height) on the order
of S-IO seconds arc obtained In the course of these  analyses. An
appropriate GC temperature program for the analyses  described herein
Is discussed In a later section (sec Table 1).

    4.2  Gas Chroamtograph-Mass Spectrometer  Interface:  The  GC-HS
Interface CM Include enrichment devices, such as a  glass  Jet separator
•r • siIIcon* membrane separator, or the gas  chromatograph can be
directly coupled to the MSS spectrometer source.  If the system ha*.
adequate pumplnc of tne source region.  The Interface may  Include a
dlverier valve for shunting the column effluent and  isolating the
MSS spectrometer source.  All components of  the  Interface should be
gloss or gloss-lined stainless steel.  The interface components must be
compatible with temperatures In the neighborhood  of  2SO*C. which  is UM
     nature at which the Interface Is typically maintained throughout
analyses for fCOO/PCOf.  The GC/MS Interface must be appropriately
configured so that the separation of from the other TCOO
isomers which Is achieved in the gas chramatographic column is not
appreciably degraded.  Cold spots and/or active surfaces (adsorption
sites) in the GC/MS Interface can cause peak tailing and peak broadening.
If the latter arc observed, thorough cleaning of the injection port.
Interface and connecting lines should be accomplished prior to pro-
     3.5  lube Furnace:   Llndberg Type 59344.

    4.3  Mass Spectrometer  (MS):  Ike mass spectrometer used for the
 inaiyses described IMC re  It  typically * double-(ocwv ing sector or
 •uadrupale  instrument •quipped Him •* electron  impact source (70 ev}.
 aainlatned  *t 2SO C, and •  standard electron Multiplier detector.
 If possible.  It  1» desirable to have both IBM »M high resolution
capability  with  IN mass spectrometer used, since confirmation of
date obtained by low resolution MS using high resolution MS Is sometimes
desirable.  Alternatively.  * combination of mas* spectrometers can be
used far this purpose.   Ike static resolution of the instrument Must
be maintained at  • minimum  of 1:SOO (with • 10» valley between adjacent
MSSCS) If  operating In  the IOM resolution US mode, and a minimum
resolution  «f 1:10.000 Is desirable far operation I* the bloh resolution
•ode.  The  MIS  ipcctroMter mitt also bo COM floured for rapid computer-
controlled  selected-loo  em I tor Ug to both hlfh end lew resolutloo
operating modes.  At a minimum, two Ion-masses characteristic of *ach
class of chlorinated dteaIns should be on*tiered, aod these are two
 tons 1* the molecular Ion Isotoplc cluster.  It  Is desirable for
 Increased confidence In  the data to also Monitor the fragment IMS
arising from the loss of COC1 from the molecular Ion,  U order to accomplish
the requisite rapid Multiple loo monitoring sequence during the tin*
period d«fined by a typically capillary ckromatographtc peak CUM base
of the chromatographlc ooak Is typically 15-20 seconds In width), the
 fallowing MS performance parameters art typically required (assuming
a 4-Ion monitor I nf sequence for oach class of PCDD/PCOf):  dwell time/
 ion-mass. olOO msec.; minimum number of data pototi/cbrematographtc
peak. 7 .   The mass seal* of the aass spectrometer Is calibrate** uslof
hlo> boiling perfluoroaeroseno and/or some other suitable mass standard
depending upon the re*u!romMts of the K-MS-05 system vt 11 tied.  The
actual procedures uttltiod  for calibration of the miss scale will bo
unique to the particular mass spectrometer being employed.  A list of
 the appropriate  IMS to  bo  monitored  to the rCOO/KOf analyses described
 herein Is presented In a laUr section (see Table 1).

    4.4  Data System:  A dedicated computer-bated data system, capable
of providing  the data described above. Is employed to control the rapid
 selected-loo monitoring  sequence and to acquire  the data,  toth dl«U«)
 data (peak  areas or feak kelfkts) as well as peak profiles (f'*P''p «»
 Intensities of  ton-masses monitored as a function of time) should he
 acquired during  the analyses. tMl displayed by the data system.  Ibis
 raw data (mass chromatogrtms) should be provided to the report of the data.

 S.  Calibration  Standards

    A recomwnded set of calibration  standards to be used In the analyses
 described  herein Is presented below.  Stock standard solutions of the
 various rtDO and KOf  isomers and mixtures thereof are prepared l» *
 olo*ebo»,  using  weighed  quantities of the authentic isomers."  Ihese
 stock solutions  are contained  In appropriate »alw»trlc flasks and are
 stored  tightly  stoppered,  in a refrigerator.  Altquols of the *«««
 standards  are removed  for direct use or for subsequent serial dilutions
 fo pr*paft wmking standards.  Ihese  standards must he checked regularly
               •   ...-_„» IP tetanic  factors for the« over a period of
tlmel to ensure that solvent evaporation or other losses have not occurred

                                         -*.>.».• 1COO.
      -  ; r —  £ia-OCOO, and 2Sng/t,lllCl,-OCOf.  Portions of MIS
 Isomer mixture are added to all samples prior to  analyses and serve
 as  Internal standards fer use in quantltatlon. Recovery of these
 standards Is else used to piage the overall efficacy of the analytical

    S.f  Standard •:  Prepare • stock solution containing 1.0 119 of
 •7CU-2.3.;.i-TCOOM of Isooctane.  This standard can be colnjected
 If desired, alone with ailquots of the final sample extract to reliably
 estimate the recovery of the "C,. 2.3.7.i-ICuO surrogate standard.

    S.I  Standard Mixture C:  Prepare a stock solution containing
 100 naM of Isooctane of each of the following PCOO  and PCOF; TCOf; 2.3.7.B 1COO; 1.3.4.i.l-PeCOf. PeCOf; |.2.4.?.i-
 PeCOO; l,2.).*.»-PeCOOi l.t,1.4.f,i-N*COF; 2.3.4.t,7.8 HxCOf; I.
 MxCDD. kxCOOi HpCOf; t.»-Hp£W;
 l.2.3.4,4.7.i-HpCOOi l.» HptOO. OCuf: and OCOO.  Ibis isemer
mixture Is used to define the  gas chromstooraphlc retention tine
 intervals or windows for each of the penta-. hexa-, hepta-. and
 octachlorlnated groups of PCOO  and PtOf.   Each  pair of Isomers of a liven
chlorinated class which is listed here corresponds to the first and
 last elutlitf Isomers of that class on the M-S caff Mary CC column
 (except fer ICBD and TCOf).  In addition, this isomer mixture is used
 to determine SC-MS response factors for representative tsomers of each
of the penta-. hexa-. hepta-, and octachlorlnated groups of PCOO  and
PCDf.   The later data are used In quantItat Ing the analytes In unknown

    S.4  Standard Mixture 0;  Prepare a stock solution containing
SO pg/nL ef Isooctane of each of the following ICOO isomers:  l.I.f.l-
 TCUD;; 2.3.?.t-TCOO; and 1.2.B.9-TCOO.  iwo of the
 Isomers In this mixture are used to define the gas chromatographlc
retention time window for ICOOs (M.fi.t-TCOO Is  the first eluting ICOO

    *• Some of the PCQO/PCDF isomer standards recommended for this mtthod
arc available from Cambridge Isotope laboratories. Cambridge. Massachusetts.
Other PCOO/PCDf standards are available from the  Irehm Laboratory, ttright
 State University. Dayton. Ohio, from the U.S.  [PA Standard Repository
at Research Triangle Park. North Carolina and possibly from other laboratories
Not all of the indicated isotopically-labelled PCOO/PCDf internal  standards
recommended here arc presently available In quantities sufficient  for
widespread distribution, but these arc expected to be available  in the near

isaner and  1,2.8.9 1COO Is  the  last eluting ICDR isomer on tht 88-5
CC column).   Ihe  remaining  Isomers serve lo demonstrate that Iht 2,3,7,1-
1CUO  turner  it  resolved fro* the other nearest eluting 1C00 isomers,
and that  the column therefore yields quantitative data (or the 2.3.7.S-
1COO  isoner  alone.
    5.5  Standard Mixture  I:  Prepare * stock solution containing SO pgM
of isooctane of  each of  the following ICOf IsoMrs: ICOf; 2,3.4,8-
ICUf; 2,3.?.8-UOf, ICOf; and ICOf.  Ihis isomer mixture
Is used to define the  ICOf gas chromatographtc retention time window
(I.3.6.B- and  ICOf are the first and last editing ICOfs on the
06-5 capillary column) and to demonstrate that 2.3.7,8-JCOf is uniquely
resolved treat  the adjacent-tinting HOT isomers.

6.  Procedures for Addition of Internal Standards and Extraction of Samples

    Both liquid  and solid  samples will be obtained for PCDO/PCDF
analyses as a result of  tht application of  an appropriate stack
sampling procedure.  Staples
resulting from the sampling train will  include the following (these
will be provided to the  analytical laboratory as separate samples in
the fore indicated):   1) particulate filter and particulates thereon;
2) particulates  from the cyclone (If used); 3) combined aqueous solutions
from the impingers; *,) the intact IAD-resin cartridge and the resin
therein; 5) combined aqueous rlnte (If used) solutions from rinses of
the noizie. probe, filter  holder, cyclone (if used), impingers, and
all connecting lines; i) combined acetone rinse solutions from rinses
of the noizle, probe,  filter holder, cyclone (if used), impingers. and
all connecting lines;  7) combined hexanc rinse solutions from rinses
of the nonle, probe,  filter, cyclone  (if used), impingers, and all
connecting  lines. In addition, samples of bottom ash, precipitator
ash. incinerator feed materials or fuel,  quench liquids, and materials
from effluent  control devices may also be provided for analyses.

    In general,  the volumes of all liquid samples received for analyses
are measured and recorded, and where appropriate, solid samples or
aliquot* thereof arc weighed.  Any samples which are homogeneous (as
for example, a single  liquid phase sample or a solid which can be
thoroughly  mixed) can be split prior to analyses, if desired, provided
that  this will still permit the attainment of the desired detection
limits  for  the analytes  of interest.  Samples such as particulates from
the sampling train which are generally collected in relatively small
quantity, are  preferably analyied in total.  .

    6,1  Organic liquid  Samples (Acetone and Itexane Solution!)   Combine
the acetone and henane  rinse solution and concentrate  to a volume of about
1-5 ml using the nitrogen blowduwn apparatus (a stream of dry nitrogen)
   X!ch 2uB.Turr!;,nijji.7,B.r!^*-r-*	i"<  *•»""
   0* he.ane and add these  o  he colcenlrafed <' f * "Uh ***" P0rtlo'ls
   further to near dryness.  Ihis reside -HI {?'" lons *nd «~*»«r«te
       C.2  Aqueous Liquids
   Add an appropriate quantity of  the isotopically-labeled internal  standard
   •future (Standard Mixture A described earlier) to the aqueous  liquid
   sample (or an aliquot  thereof)  in a screw-capped bottle fitted with  a
   lafIon-lined cap.   Add approximately 251 by volume of bexane to the
   spiked aqueous  sample, seal the bottle and agitate on a  shaker for a
   period of  three  hours.  Allow the vessel to stand until  the aqueous and
   organic  layers separate, then transfer the organic layer to a separate
   sample bottle.   Repeat the hexane extraction sequence  two additional
   times  and combine the organic fractions  with that from the first ex-
   tract ION.  Proceed with the sample fractionation and cleanup procedures
   described below.

      63  Solid Samples
  Place a glass  extraction  thimble and ) g of silica gel and a plug of
  glass wool  Into  the  Soxhlet  apparatus, charge the apparatus with  toluene
  and reflux  for a  period of one hour,  iemove the toluene and discard  ft,
  retaining the silica gel. or if desired, retain a portion of the  toluene
  to check  for background contamination,   for extraction of particulates,
  place the entire  sample in the thimble onto the bed of precleaned  silica
  gel  (1  cm.  thick), and  top with the precleaned glass  wool retained
  from the  Initial  Soxhlet cleaning procedure.   Add the appropriate
  Quantity of the isotopically-labelled  internal  standard mixture
  (Standard Mixture A described earlier)  to the  sample  in the  Soxhlet
  thimble.  Charge the Soxhlet  with toluene and  reflux  for  a period of
  If hours. After extraction,  allow  the  Soxhlet to cool, remove the
 toluene extract,  and transfer it  to another sample vessel.  Concentrate
 the extract  ta  a  voli-me  of approximately 40 ml by using the nitrogen
 blowdown apparatus described  earlier.  Proceed with the sample fraction*
 tion and cleanup  procedures described below.

 1.  Procedures for Cleanup and Fractionation of Sample Extracts

     The  following  column chromatographic sample clean-up procedures
are used in  the order given,  although not  all  may  be required.   In
general, the silica and alumina  column procedures are  considered ta be a
minimum requirement.  Acceptable alternative cleanup procedures may be
used provided that they  are demonstrated to effectively  transmit  a

 ipresentallwe set of the analytes of  interest.  Ihe column chromato-
 raphic procedures listed here  have been demonstrated to be effective
 or • mixture consisting of ICOO. 2.3,7,i-ICDO, 2.3,6.8-1CBf,
 .2.4,i-ICuf. 2.3.M-KOf, 1,2,3.7.8  PeCOO. PeCOf.  1.2.3,4,7,i-
 xCOO, l,2,4.t,7.t-M«CaF, l.2.3.4.».7.8-NpCOu. M.M.M.i-HpCDf ,
 COO and OCOf

    An extract, obtained as described in the foregoing sections   is
 oncentrated to a volume of about 1 «L using the nitrogen blowdown
 pparatus. and this Is transferred quantitatively (with rinsings) to
 he combination silica §el column described be tow.

    7.1  Combination Silica Gel Column:  Pack one end of t glass
 :olu*n (20 mm. 0.0. * 230 mm In length) with glass wool (precleaned)
 ind add, in sequence. 1 g silica gel,  t g base-modified silica  gel,
 I g silica gel. 4 g acid-modified silica gel, and I g silica gel.
 f Silica gel and Modified silica gel are prepared as described  In the
 Reagents sections of this protocol.)   Preelute the column with  30 mt
 nexane and discard the eluate.  Add the sample extract in 5 mt  of mexane.
 to the column along with two additional S *1 rinses,  flute the coition
 •fth an *ddltional M nt af hexene and retain  the entire eluate.
 Concentrate this solution to a  VOIUM  of about I •).

    7.2  Basic Alumina Coliwn:  Cut off a 10 *t disposable Pasteur
glass pipette at the 4 nt graduation mark and pack the lower section with
glass wool (precleaned ) and 3  g of Moelm basic alumina (prepared as
described in the Reagent section of this protocol),  transfer  the
concentrated extract fro* the combination silica column to the  top of
 the column and elutc the column sequentially with li ml of beiane,
 10 ml of 81 methylene chloride- In-hcianc and IS mi of SOS methyl rue
 chloride-In-hexane, discarding  the first two eluate fractions and
 retaining the third eluate fraction.   Concentrate the latter  fraction
 to about 0.5 mL using the nitrogen slowdown apparatus described earlier.

    7.3  PI-21 Carbon/Cellte 545 Column:  lake a 9 inch disposable
Pasteur pipette and cut off a 0.5 Inch section from the constricted tip.
 Insert a filter paper disk at the top  of the tube, 2.5 cm. from the
constriction.  Add a sufficient quantity of PI-21 Carbon/Celite 54S
 (Prepared as described in the reagent  section of this protocol) to the
 tube to fora a 2 CM. length of  the Carbon-Celite.  Insert a glass wool
 plug.  Preelute the column in sequence with 2 ml of 501 benzene-in-*tbyl
acetate, 1 ml of 501 methylene  chloride-in-cyclohexane and 2 ml of hexane.
 and discard these eluates.  load the extract (in 1 ml of hexane) from
 the alumina column onto the top of the column, along with I ml  hexane
 rtn&e.  {lute the column with 2 ml of  501 methylene chloride-in-hexane
 and 2 at of SOS nentene-in-ethyl acetate and discard these eluates.
 Invert the column and reverse elute  it willi 4 ml o( toluene, retaining
 tins «luate.  Concentrate the eluate and transfer it to a Reacti-Vial
 for stoiage.  Store extracts in a freezer, shielded from light, prior
    7.4  Silica/Blot Micro Column Cleanup-
Steps small amounts of highly colored
                                                  th.  K
                                                      *t>01" elM«-«P
column:  Push a small plug of alass wool u
glass -asteur pipette, followed* by j^2 '„
International ). i m of silica oel LT it  it P
In. column li   -       "                 ''
                                                         fo1 »«•»«§
                                                          * " !  3*
    £S£Ut5SrjL£2£: {sag- * «»«»»- -
     1.1  Sample extracts prepared by the procedures described In the
 foregoing are analyzed by GC-NS utilising the following instrumental
 parameters.   Typically, I to 5 i>L portions of the extract are Injected
 Into the CC.   Sample extracts are first analyied using the OB-S  capillary
 6C column to  obtain data on the concentrations of total tetra-through
 octa-COOs and COfs, and on ICDO.   If tetra-COfs are detected
 In this analysis, then another aliquot  of the sample is analyzed In
 a  separate run, using the 00-225 column to obtain data on the concentration
 of 2.3.7.f-lCOf.
    t.2  Gas Chromalojjraph
    •-2.2  Carrier gas:  Hydrogen. JO Ib head pressure.
programmed (lifl'C for 1 min..  then Increase from loblC~to"240*C'§
hold  at 240*C for i min.)

    8.2.4  Interface lemperature:  2&0°C
    B.3 Hast Spectrometer
                                                                               8.3.1  loniiation Hode:  Outrun inipdct  (10  eV)

    8.3.2   Static  Resolution:   1:600  (101 valley) or 1:10,000 depending
upon requirement     Usually the sample extracts are initially analyzed
using low resolution MS.  then  If PCOO/PCOF are detected, it is desirable
to analyze  a  second portion of the  sample extract using high resolution
    8.3.3   Source lemperature:
    8.3.4   Ions  Monitored:   Computer-Controlled Selected-Ion Monitoring.
See Table I  for  list  of Ion masses  monitored and time intervals during
which tons characteristic of each class  of COOs and COfs are monitored.
    ft.4  Calibration Procedures:
    8.4.1   Calibrating the MS Mass Scale:   Perfluoro Kerosene, decafluoro-
triphenyl phosphine. or any other accepted mass marker compound must be
introduced  Into the MS. In order to calibrate  the mass scale through at
least m/z SOO.   The procedures specified by the manufacturer for the
particular  MS Instrument used are to be employed for this purpose.  The
mass calibration should be rechecked at least  at 8 hr. operating Intervals.
    8.4.2   Table I  shows the CC temperature program typically used to
resolve each  chlorinated class of KOO and PCOF  from the other chlorinated
classes, and  indicates the corresponding time  Intervals during which ions
Indicative  of each  chlorinated class  are monitored by the MS.  This
temperature program and ion monitoring time cycle must be established by
each analyst  for the particular instrumentation  use-* by Injecting allquots
of Standard Mixtures C. 0. and E (See earlier  section of this protocol
for description of  these mixtures).   It may be necessary to adjust the
temperature program and Ion monitoring cycles  slightly based on the
observations  from analysis of these mixtures.

    8.4.3   Checking GC Column Resolution for and
TCOF:  Utilize the  column-resolution  ICOO and  ICOF tsomer mixtures
(Standard Mixtures  0 and t.  containing 50 pg/pl. respectively of the
appropriate ICOO and TCOF Isomers) to verify that and
2.3,7.8-TCOF  are separated from the other ICOO and TCOF isomers.
respectively.  A 20J valley or less Bust be obtained between the mass
chromatographic peak observed for and adjacent peaks
arising from  other  ICOO Isomers.and similar separation of 2.3.7.B-TCOF
from other  neighboring TCOFs. is required.   Standard Mixture 0 is
utilized with the OBS coluwi and Standard Mixture I with the DB-225
column.  Analyze the column performance standards using the instrumental
parameters  specified in Sections 8.2  and 8.3.  and in Table |.  lite
column performance  evaluation must br performed  each time a new column
is Installed  In the gas chromatograph. and at  least once during each 8
hour operating period.  Providing that the same  column is employed for a
period of time. Its performance can also be  gauged by noting the peak
width (at 1/2 peak height) for ICOO  or for  If
this peak width Is observed to broaden by 201 or more as collared to
the usual width for satisfactory pperatlon,  then the column resolution
Is suspect and must be checked.  If the colun resolution is found to
be Insufficient to resolve and  2.3.7.B-TCDF fro* their
neighboring ICOO and ICOf Isomers. respectively, (as Matured OA the
two different columns used for resolving these two Isomers). then a
new M-S and/or 08-22S CC coluwi must be Installed.
                                                   •.4.4  Calibration of the GC-MS-OS system to accomplish quantitative
                                               analysis of and 2.3.7.B-TCOf. and of the total tetra-
                                               through octa-COOs and COfs contained In the sanple extract.is accomplished
                                               by  analyzing a  series of at least throe working calibration standards.
                                               fach of these standards Is prepared to contain the sane concentration
                                                  each of the  stable-lsotoplcally labelled Internal standards used
                                               here (Standard  Mixture A) but a different concentration of native
                                               PCOO/PCOf (Standard Mixture C).  typically. Mixtures will be prepared
                                               to  tint the  ratio of native PCOO and PCOf to Isotoplcatly-labelled
                                               KOO and PCOf will be on the order of 0.1. O.t and 1.0 In the three
                                               working calibration Mixtures.  The actual concentrations of both native
                                               and tsotoplcally-labelled PCOO and PCOF In the working calibration
                                               Standards will  be selected by the analyst on the basis of  the concen-
                                               trations to  be  Measured In the actual sanple extracts.  At the tine
                                               when allquots of each of the standards are Injected (and also when
                                               Injecting allquots of actual sample extracts). If desired, an aliquot
                                               •f  a standard containing typically 1 ng of  MCK-2,3,7.B-!CDO (Standard D)
                                               can be  drawn Into the micro syringe containing the calibration solution
                                               described above (or the sample extract) and this is then co-injected
                                               •long with the  sample extract In order to obtain data permitting
                                               calculation  of  the percent recovery of the *lC,,- internal
                                               Standard,  equations for calculating relative response factors fro* the
                                               calibration  data derived fro* the calibration standard analyses, and for
                                               calculating  the recovery of the "Ci,- and the other
                                               tsotoplcally-labelled PCOO and PCOF, and  the concentration of native
                                               PCOO and PCOf In the sanple (fro* the extract analysis), are summarized
                                               below.  In these calculations, as can be seen, Is employed
                                               as  the  Illustrative model.  However, the calculations for  each of the
                                               other native dloxlns and furans In the sample analysed are accomplished
                                               In  an analgous  manner.  It should be noted that in view of the fact
                                               that stable-tsotopically labelled internal standards corresponding to
                                               each tetra-  through octachlorinated class are not used here (owing
                                               to  limited availability at this time) the following approach is adopted:
                                               For quant I tat ion of tetrachlorinated dibeniofurans "C,,-2.3,7.8 TCOF
                                               is  used as the  internal standard.  For quantitation of tetrachloro-
                                               dlbenio-p-dloxlns, l*Cu- is used as the internal standard.
                                               For quantitation of PeCOO. HxCOO. PeCOf. and HxCOF. the corresponding
                                               Stablc-lsotoplcally labelled HxCOO and HxCOF internal standards are used.
                                               For quantitation of HpCDD,  OCOO. and  HpCOF. OCOF. the isotopically-
                                               labelled OCOO and OCOF, respectively, are used.  Inherent  In this
                                               approach is  the assumption that the response factors for each of  the  Isomeri

f each chlorinated class are the sane,  and  In the case of the penta-
td nepta-COOs and COfs, the assumption  is nade that the responses for
MS* two classes are equivalent to those for the tetra-lsoners and the
cta-isoners. respectively.

   1.4.S  Equations for Calculating Response factors. Concentration of
,3.7,1-ICOO la An Unknown Sanple. and Incovorias of Internal Standards.
quatIon 1:  Response factor (RRF) for native 2.3,7.8 1CDO using
            t*CM-2.3.7.t-1COB as an Internal standard.
    •tore:  A(  - SIM response far 2.3.7,•-TOW Ion at mji 320 * 312

            A,. * SIN response for l§C41-2.3,7.i-ICOO Internal standard
             11   Ian at aVi 332

            C.  • Concentration of the internal standard (pf./nt..)

            Ct  • Concentration of the 2.3.7.I.-TCOO (pg.M.)
Equation 2:  Response Factor (R*f) for "q,-2.3.7.i-ICOO, the co-injected
             external standard

     Mff-  (*i,e.§)/(*.,cul

     where:  A(§ • SIN response for "Ci.^.S^.i-TCOO Internal
                   standard Ion at
            A.. • SIN response for co-Injected "CI»-2.).7.«-lCDO eiternal
             **   standard at m/i 321 * O.OOf (SIN response for native
                  2.3.7,t-KOO at mji 322)

            C|g • Concentration of Ike Internal standard (pg-M-)

            C   * Concentration of the external standard (f»9.M.)
                                                                               Equation 3:  Calculation of concentration of native 2.3,7,1-rcOO usinfl
                                                                                            "£»a-2.3.?.l-ICOD as internal standard

                                                                               Concentration.  M /I- « (Ag) dt)/(*u)(Mf-)(u)

                                                                                    **ere:  A§  • SIN response for 2.3,7.1-fCOO ion at n/j  320 * 322

                                                                                            *1§ * SIN response for the  "C|(-2.3.7.I-1COD Internal
                                                                                             11   standard Ion at •/! 332

                                                                                            l§  * Anount of Internal standard added to each sanple (pg.)

                                                                                            H   • Iteliht ef soil  or waste In or***

                                                                                          ilF^  * Melatlve response factor fro* equation 1

                                                                               Equation 4;  Calculation of S recovery of llC»,-2.3.7.«-lCOO Internal  standard

                                                                               I Recovery -
                                                                                           A,  * SIN response  for **Cii-2,3.?,l-TCOO internal  standard
                                                                                            11   IM at m/t 332

                                                                                           A.. * SIN response  for "CI,-2.3.7.B-TCDO external  standard
                                                                                            •*   Ion at mjt 321 - 0.009     (SM Response  for  native
                                                                                                 2.3.7.a-TCOO  at mil 322)

                                                                                           f.  * Aeount of "CU-2.3.7.R-TCOO external  standard
                                                                                            *    co- Injected Mitk saey»le extract (n§.)

                                                                                           I.  * Theoretical eeaunt of "Cn-2,3.7.i-TCOO  Internal
                                                                                            1    standard In Injection

                                                                                         Mf.  » Relative response factor from Equation 2

                                                                                   As noted above, procedures  similar to these are applied to  calculate
                                                                              analytical  results for all of the other KOO/rtOf determined in  this net hod.

                                                                                  •.S  Criteria Which iC-NS Data Must Satisfy for Identification of
                                                                              PCOO/FCOf la San? I as Ana ly ted  and Additional Details of  Calculation Procedures,

                                                                                  In order to Identify specific PCOO/PCOf in sanples analyied.  the
                                                                              6C-HS data  obtained nust satisfy the foil 0-109 criteria:
                                                                                  •.S.I  Hail spectral  responses nust be observed at  both the Molecular
                                                                              and fragment ion nasses corresponding to the ions indicative of each
                                                                              chlorinated class of  PCDO/PCOF  identified (see Table 1) and Intensities
                                                                              of these ions nwst naxfntiie essentially simultaneously  (within « I
                                                                              second).  In addition, the chro*utogr*phic retention tines observed for
                                                                              **ch PCOO/PCOf signal Must be correct relative to the appropriate

 stable-isotopically labelled internal  standard and must be consistent
 with  the  retention time Hindoos  established  lor lite chlorinated group to
 which the particular PtOO/KDf  is  assigned.

    8.1.2  Ihe  ratio of the intensity  of  the molecular  ion (M)* signal
 to  that of  the  (H«2)* signal Mist  be within  * 101 of the theoretically
 expected  ratio  (for example. 0.?}  In the  case of irnu;  therefore
 the acceptable  range for thli ratio Is 0.69 to O.fib).

    B.5.3  Ihe  intensities of the  Ion  signals are considered to be
 detectable  If each exceeds the baseline noise by a factor of it least
 3:1.   Ihe Ion intensities are considered  ta  be quantitatively measurable
 tf  each ion intensity exceeds the.  baseline noise by • factor of at
 least 5:»c.
    B.S.4  for  reliable detection and quant I tat ion of PCOf It Is also
desirable  to awnI tor signals arising fro* chlorinated diphenyl ethers
which.  If  present  could give rise to fragment  Ions yielding Inn Masses
identical  to those Monitored as  indicators af  the fCOf.  Accordingly,
in Table I. appropriate chlorinated dtphenyl ether Masses arc specified
which awst be Monitored simultaneously with the PCOf Ion-Misses.  Only
when the response  for the dtphenyl ether  ion Mass is not detected at
the same time     the PC Of Ion Mass can the signal obtained for an
apparent PCOF be considered unique.

    B.S.S  HeasureMent of the concentration of the congeners in a
chlorinated class  using the Methods described  herein n based on the
assumption that all of the congeners are  identical to the calibration
standards  employed in term of their respective chemical and separation
properties and  in  tents of their respective gas chroMatographic and MISS
spec trwne trie responses.  Using these assumptions, for example, the
l§C||-2.3.?,i'!CDD internal standard is utiliied as the internal
calibration standard for all of the 22 ICDO  isomers or congeners.
Furthermore,  the  concentration of the  total ICDO present  in a sample
extract is determines by calculating, on  the basis of the standard
procedure  outlined above, the concentration of each KOO  isomcr peak
(or peaks  for Multiple ICDO isomers, where  these coelute) and these
individual concentrations are subsequently  suamed to obtain the concen-
tration of 'total' ICDO.
     c"  In practice, the analyst can estimate the  baseline noise by measuring
 the  eiteiision of  the baseline immediately prior  to  each  of  the two mass
 chromatographic peaks attributed to a given PCM)  or  PCOf.   Spurious signals
 nay  arise either  from electronic noise or from oilier organic compounds  in
 the  extract.   Since it may be desirable to evaluate  the  judgement of the
 analyst  in this respect, copies of original mass  chromatograms must be
 included in the report of analytical results.
    B.6  frequently, during the analysis of  actual sample extracts.
extraneous compounds which are present  in the  extract (those organic
compounds not completely removed during the  clean-up phase of the analysis)
can cause changes in the liquid and gas chromatographic elution characteristics
of the PCDO/PCOf (typically retention times  for the PCDO/PCW art prolonged).
Suck extraneous organic compounds, when introduced Into the mass spectro-
Meter source may also result in • decrease in the sensitivity of the MS
because of suppression of tonliatlon. and other affects such as charge
transfer phenomena.  The shifts in chroma tographic retention times are
usually general shifts, that is, the relative retention times for the
PCDO/PCOf are not changed, although the entire elution time scale Is
prolonged.  Ihe analyst's Intervention in the CC MS operating sequence
can correct for the lengthened GC retention  times which are sometimes
observed due to the presence of extraneous organics  in the sample
••tract,  far example, using the program outlined In Table 1. If the
retention time observed for 2.3.7.B-1CDO (which normally  Is 19.1 minutes)
Is lengthened by 30 seconds or more, appropriate adjustments  In the
programming sequence outlined In Table 1 CM be made, that If, each
selected Ion-monitoring program Is delayed by * length of time propor-
tionate to the lengthening of the retention  time for the  2.3.7,B ICDO
isomer.  In the case of foniiation suppression, this phenomenon Is
Inherently counteracted by the Internal standard approach.  However.
if loss of sensitivity due to lonlzation suppression is severe.
additional clean-up of the sample eitract may be required In  order to
achieve the desired detection limits.

9.  Quality Assurance/Quality Control

    9.1  Quality assurance and quality control are ensured by the  following

    9.1.1 fach sample analyzed is spiked with stable isotopically  labelled
internal standards, prior to extraction and  analysis.  Recoveries
obtained for each of these standards should  typically be  in  the range
from 60-90*.  Since these compounds are used as tru* internal standards
however, lower recoveries do not necessarily Invalidate the  analytical
results for native PCOO/PCOf. hut may result In higher detection limits
than are desired.

    9.1.2  Processing and analysis of at least one method blank sample
is accomplished for each set of samples (a set being defined as 20 samples
or less).

    9.1.3  It is desirable to analyze at least one sample spiked with
representative native PCDD/PCOF for each set of 20 or fewer  samples.  Ihe
result of this analysis provides an indication of the efficacy of the
entire analytical procedure,  the results of this analysis Mill be
considered acceptable if the detected concentration of each of the native

 :00/PCDF added ie Ike saayle Is within *SOI of the known concentration.
 In appropriate »tt of native isomers to~be added here ii • set  such
 t that indicated far Standard Mixture C.)

    9.1.4  At least OM of the samples analyzed out of each set  (of 20
 ample* or less) Is amalyted In duplicate and the rttults of  the duplicate
 •alysls arc Included in the report *f data.

    i.l.S  Performance evaluation staples prepared by (PA.or  other
 aboratorles.Hhicb contain representative KOO/rCOf in concentrations
 pproilMtlnf those present In typical field samples being enalyted
 but unknoMi to the analyilnf lab) should be perledlcally distributee*
 • laboratories accomplish Inn, Inns* analyses.

    i.l.i  Sources nf all calibration and perfonHnce standards  used In the
 nalyses and the purity of these Materials nust be specified  In  the data

 I.  Data Reporting

    10.1  Each report of analyses accomplished using the protocol
 (escribed herein Hill typically Include tables of results Mblcb  Include
 [be following:
 (or detection Halts) are reported.

    10.1.i  Ihe sane raw and calculated data which arc provided for the
 actual  samples Mill also be reported for the duplicate analyst}, the
 method  blank analyses, the spiked sanple analyses  and any other QA
 or performance sanples analyied in conjunction with  the actual sanple

    10.I.I  The recoveries of the internal standards in percent.

    10.1.1  Ihe recoveries of the native KOO/KOf from spiked saeples
 In percent.

    10.1.9  The calibration data. Including response factors calculated
 from the three point calibration procedure described elsewhere  in this
 protocol,  bate showing that these factors have been verified at least
 once during each • how period of operation or Mltb  each separate set
 of saaiples analyied oust be Included.

    10.1.10  Ihe weight or quantity of the original  sanple analyied.

    10.1.11  Documentation of the source of all PCOO/PCOf standards
 used and available specifications on purity.
                                                                       Lln/i I
    10.1.1  Complete Identification nf the sanples analyied (sample
lumbers and source).

    10.1.1  Ihe dates and times at which all analyses were accomplished.
this Information should also appear an each mass chroma togram included
tilth the report.

    10 1 3  Raw mass chromatographlc data which consists of the absolute
Intensities (based on either pe»k ktl»hl or «*«k *r**> •' thc
observed for the ion-masses monitored (See lable 1).
    10.1.11  In addition to the tables  described above, each report of
analyses Hill include all mass chromatograms obtained for all saayles
analyied, as well as for all calibration, GC column performance, and
6C SrindoH* definition runs and results of column performance checks.

    10.1.13  Any deviations from the procedures described In this protoea I
which are applied In the analyses of samples Hill be documented In
detail in tbe analytical report.

11. Typical Oata Indicative of Method Performance - Precision and Accuracy.
    10.1.4  Ihe calculated ratios of the intensities of tbe molecular
ions for all KOO/KOf detected.

    10  I t>  Ihe calculated concentrations of native 2.3,7.8-UDO and
2,3,7.8-ICOf, and  the  total concentrations of the congeners of nach
cl*s* of PCOO/KOf  for each sample analyzed, enpressed in nanograms
1COD per gr*« of sample  (that  Is, parti-per-btM Ion) as determined
>,«. th» r»w dau.   If no PCOOVPCDf are detected, the notation "Mot
                                                  '  Me roncentr**'""'
    II.I  Ihe method described herein has typically been employed to
quantitatively determine 2.3.1,1-lCOO in combustion product saa^les at
concentrations as low as 10 picograms/gram and as high at 100 1*9/9.
Concentrations 0f the other PCDO/PCDf which can be detected typically
fall within the range of 20 picograns/isomer/gram of sanple. to 100
plcograms/f of sample.  Of course, the Halts of detection which can
be practically achieved are dependent on the quantity of sample available

. J th     unt     :lnd     tber     rfer     orga      esli
present In the sanple.  With respect to precision,  the  avenge, deviation
of data obtained fro* the analyses of a nwfcer of allquots of the MM
sw*le containing the 2.3.l,t-1COO Isoner In the 250-300 ppb range
Is estlMted to be «1M or better.  Data on the precision of quantisation
of Multiple PCOO/PcBf In a finale ssaple are not as yet available.  As
yet, there Is Inadequate Interlaboratory and performance evaliMtlon data
available to specify  the accuracy which can be expected of the analytical
procedures described  herein.
a   ss   as
8   88   IB
r.-rrr? [is:
888888  i

                        TMU i.
wtt* IM*. pnfPM u nen
>i«v iMMtrtwrt t»l<

ItM Mi
Itin MMtt trcfria;

 IM P»», tlM/MII * e.JB MC.
iUrt feu H*fTMi

 1U »Ml tMtNMt * (.X MC,
»»«•« (*«.
               I* !7V


t»f i» MM. «r*«r** U MB*
Ut*B MMtritft «•'«
(Mi Utwn u IN*
                                                   »y MMt
••OH, ao»t.
             •K tn M*rt*lttttM

                wlf) »Mt*l*n

                   K CTll—> i
                                                                                                              ICCONHENKt SAFCIV AND NMMN.1NG PiOCtfiURtS
             Safety and Hand! iitl Procedures la Connection
  he Jaaj^ltcat Frtoca  far BfleraiiMt tan af P€m/KW  in £gifeMslt»
                                                                                                    are cagptctefi1 usinf a «Miciteii trash OMpactiw «se4
                                                                                  salf f«r hi«r