EPA-600/2-76-137-^
May 1976
Environmental Protection Technology Series
           EVALUATION  OF SELECTED METHODS FOR
             CHEMICAL AND BIOLOGICAL TESTING  OF
                INDUSTRIAL PARTICULATE  EMISSIONS
                                 industrial Environmental Research Laboratory
                                      Office of Research and Development
                                     U.S. Environmental Protection Agency
                                Research Triangle Park, North Carolina 27711

-------
                RESEARCH REPORTING SERIES

 Research reports of the Office of Research and Development, U.S. Environmental
 Protection Agency, have been grouped into five series. These five broad
 categories were established to facilitate further development and application of
 environmental technology. Elimination of traditional grouping was consciously
 planned to foster technology transfer and a maximum interface in related fields.
 The five series are:
     1.    Environmental Health Effects Research
     2.    Environmental Protection Technology
     3.    Ecological Research
     4.    Environmental Monitoring
     5.    Socioeconomic Environmental Studies

 This report  has  been assigned to  the ENVIRONMENTAL PROTECTION
 TECHNOLOGY series. This series describes research performed to develop and
 demonstrate  instrumentation, equipment, and methodology to repair or prevent
 environmental degradation from point and non-point sources of pollution. This
 work provides the new or improved technology required for the control  and
 treatment of pollution sources to meet environmental quality standards.
                     EPA REVIEW NOTICE

This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication.  Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency,  nor does mention of trade
names or  commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service. Springfield, Virginia 22161.

-------
                                       EPA-600/2-76-137

                                       May 1976



   EVALUATION OF  SELECTED METHODS

FOR  CHEMICAL AND BIOLOGICAL TESTING

 OF  INDUSTRIAL PARTICULATE EMISSIONS
                         by

                      H. Mahar

                The Mitre Corporation
               Westgate Research Park
               McLean, Virginia  22101
           Contract No. 68-02-1859, Task 5
                ROAPNo. 21ADD-BG
             Program Element No. 1AB013
          EPA Project Officer: L.D.Johnson

      Industrial Environmental Research Laboratory
       Office  of  Energy, Minerals, and Industry
          Research Triangle Park, NC  27711
                    Prepared for

    U.S. ENVIRONMENTAL PROTECTION AGENCY
          Office of Research and Development
                Washington, DC 20460

-------
                             ABSTRACT

The report gives results of chemical analyses and cellular biological
assays performed on size-classified particulate material collected
at nine industrial sites using a new series cyclone sampling train.
The excercise was formulated to determine the performance of the train
and whether the chemical analyses or the bioassays, alone or in
combination, were sufficient to characterize the hazards associated
with particulate emissions.  This program lends support to the view
that size-classified particulate matter is needed for the various
chemical or biological tests.  Elemental analysis and partial organic
characterization of the particulate samples have been performed.  A
cellular bioassay, utilizing rabbit alveolar macrophages, has been
used to estimate the toxic potential of particulate samples in terms
of their observed acute cytotoxic activity.  A bacterial screening
technique, utilizing several histidine deficient Salmonella typhimurium
strains, has been used to study the mutagenic potential of the parti-
culate samples.  No strong correlation was observed between the
chemicals analysis and biological activity of the samples.
                                  ii

-------
                         TABLE OF CONTENTS

                                                               Page
ABSTRACT                                                        ±±
LIST OF ILLUSTRATIONS                                            v
LIST OF TABLES                                                 vii
ACKNOWLEDGMENT                                                viii

1.0  CONCLUSIONS                                                 1

2.0  INTRODUCTION                                                4
     2.1  Approach                                               4
     2.2  Objectives                                             6

3.0  SAMPLE COLLECTION                                           8
     3.1  Fabrication and Calibration of Sampling Train          °
     3.2  Sampling Support Procedures                           10
     3.3  Field Sampling                                        H
          3.3.1  Steel Mill, Open Hearth Furnace                13
          3.3.2  Steel Mill, Basic Oxygen Furnace               i3
          3.3.3  Steel Mill, Coke Oven Heater                   14
          3.3.4  Steel Mill, Iron Sintering Plant               14
          3.3.5  Oil-Fired Power Plant                          i4
          3.3.6  Copper Smelter                                 I5
          3.3.7  Aluminum Smelter                               15
          3.3.8  Paper Mill, Kraft Pulping                      15
          3.3.9  Ceramics Plant, Clay Aggregate Production      i6
          3.3.10 Municipal Waste Water Sludge Incinerator       16

4.0  SAMPLE ANALYSIS                                            i7
     4.1  Chemical Analysis                                     I7
          4.1.1  Spark Source Mass Spectrometry                 I7
          4.1.2  Gas Chromatography - Mass Spectrometry          19
          4.1.3  High Resolution Mass Spectrometry              19
     4.2  Biological Characterization                           20
          4.2.1  Cytotoxicity Evaluation                        21
          4.2.2  Mutagenicity Evaluation                        24

5.0  RESULTS                                                    27
     5.1  Chemical Analysis                                     27
     5.2  Bioassays                                             37

6.0  DISCUSSION                                                 53
     6.1  Sample Collection                                     53
          6.1.1  Sample Train Performance                       53
          6.1.2  Demonstrated Need for Size-Classified
                 Particles                                      54
                                 iii

-------
                         TABLE OF CONTENTS
     6.2  Chemical Analysis                                    55
          6.2.1  Spark Source Mass Spectrometry                55
          6.2.2  Gas Chromatography - Mass Spectrometry         56
          6.2.3  High Resolution Mass Spectrometry             56
          6.2.4  Comparison of Analytical Results              57
     6.3  Biological Characterization                          58
          6.3.1  RAM Cytotoxicity Bioassay                     58
          6.3.2  Mutagenesis Bioassay                          63
     6.4  Comparison of Chemical Analysis to Observed
          Biological Activity                                  65
     6.5  Considerations for Future Research                   71

7.0  REFERENCES                                                74
                                 IV

-------
                        LIST OF ILLUSTRATIONS

Figure Number

      1        Series Cyclone Train, Field Sampling Configura-
               tion                                                9

      2        Rabbit Alveolar Macrophage Cyctotoxicity Screen-
               ing Test Procedure                                 23

      3        Salmonella Typhimuriura Mutagenicity Screening
               Test Procedure                                     26

      4        Viability of Rabbit Alveolar Macrophages
               Exposed to Various Fractions of Particulate
               Samples Collected at the Open Hearth Furnace       39

      5        Viability of Rabbit Alveolar Macrophages
               Exposed to Various Fractions of Particulate
               Samples Collected at the Coke Oven Heater          40

      6        Viability of Rabbit Alveolar Macrophages
               Exposed to Various Fractions of Particulate
               Samples Collected at the Iron Sintering Plant      41

      7        Viability of Rabbit Alveolar Macrophages
               Exposed to Various Fractions of Particulate
               Samples Collected at the Ceramics Plant            42

      8        Viability of Rabbit Alveolar Macrophages
               Exposed to Various Fractions of Particulate
               Samples Collected at the Copper Smelter            43

      9        Viability of Rabbit Alveolar Macrophages
               Exposed to Various Fractions of Particulate
               Samples Collected at the Aluminum Smelter          44

     10        Viability of Rabbit Alveolar Macrophages
               Exposed to Various Fractions of Particulate
               Samples Collected at the Sludge Incinerator        45

     11        Viability of Rabbit Alveolar Macrophages
               Exposed to Various Fractions of Particulate
               Samples Collected at an Oil-Fired Power Plant      46

-------
                        LIST OF ILLUSTRATIONS

Figure Number                                           ,         Page

     12        Viability of Rabbit Alveolar Macrophages
               Exposed to Supernatant Fraction Collected
               from the Copper Smelter (3-lOy Sample)             47

     13        Viability of Rabbit Alveolar Macrophages
               Exposed to Particles Plus Supernatant
               Fraction Collected from the Sludge Incinerator
               (l-3y Sample)                                      48
                                vi

-------
                          LIST OF TABLES

Table Number                                                  Page

     1        Logistics of Sample Collection                   12

     2        Sample Disbursement                              18

     3        Elemental Analysis of Particulate Samples as
              Determined by Spark Source Mass Spectrometry     28

     4        Carcinogenic and Structurally Similar Poly-
              cyclic Organic Constituents in Particulate
              Samples as Determined by Gas Chromatography-
              Mass Spectrometry                                32

     5        Constituents of Particulate Samples as
              Determined by High Resolution Mass Spectrometry  34

     6        Semi-Quantitative Mass Spectral Analysis of
              Particulate Matter Collected at the Aluminum
              Smelter                                          35

     7        Possible Detection of Carcinogenic Polycyclic
              Organic Material from High Resolution Mass
              Spectrometric Analysis                           36

     8        Viability of Rabbit Alveolar Kacrophages
              Exposed to Sub-Micron Particle Filters           49

     9        Mutagenic Activity of Industrial Particulate
              Samples                                          50

    10        Dose Related Response for Mutagenically Active
              Particulate Samples                              51

    11        Mutagenic Activity of Three Selected Industrial
              Particulate Samples                              52

    12        Relative Cytotoxic Nature of the Industrial
              Particulate Samples                              62

    13        Relative Ranking of Identified Chemical
              Constituents Based on Acute Toxicity            67
                                vii

-------
                          ACKNOWLEDGMENT

A  research effort of this magnitude requires close cooperation and
support by all participants.  In addition to those affiliations
mentioned in  the text of this report, the author cites major con-
tributions offered by the Project Officer, Dr. L. Johnson, of the
Industrial Environmental Research Laboratory (EPA/RTP), and
Dr. M. Waters of the Health Effects Research Laboratory (EPA/RTP),
who provided pertinent comments on the rabbit alveolar macrophage
bioassay procedure that he was instrumental in developing.

The MITRE Corporation was assigned the role of coordinator in this
project, and has been responsible for the collation, analysis, and
interpretation of the results.  The author thanks Mr. G- Erskine
and Dr. N. Zimmerman for their assistance in this endeavor.

This work has been performed under Contract Number 68-02-1859 for
the Process Measurements Branch, Industrial Environmental Research
Laboratory,  Environmental Protection Agency, Research Triangle Park,
North Carolina  27711.
                                 viii

-------
                         1.0  CONCLUSIONS
In order to determine a rapid, effective, and inexpensive means for
evaluating the potential hazards associated with participate emissions,
two methods, chemical analysis and cellular bioassay, were carried out
on size-classified particulate material collected at nine industrial
sites.  The experiment was designed to test whether such methods,
alone or in combination, were capable of indicating potential hazards
associated with particulate emissions.

The rabbit alveolar macrophage (RAM) cytotoxicity bioassay and a
mutagenicity screening test, using three Salmonella typhimuriuiti
bacterial tester strains, were utilized to predict the acute
toxicity and mutagenic behavior of the particulate samples.  Partial
organic characterization, with emphasis on polycyclic hydrocarbons
of known carcinogenic potential, and inorganic elemental analysis
were performed on the same size-classified particulate samples.
Several conclusions can be drawn from the research as conducted:

      (1)  the series cyclone sampling train developed for this
          study has been shown to be a useful tool in collecting,
          within a one-to-five hour sampling period, sufficient
          quantities of size classified particulate material
          from a variety of industrial sources to permit further
          chemical and biological testing.  Furthermore, the
          need for size-classified particulate material has been
          demonstrated, since different size particles collected
          at the same industrial source do not necessarily possess
          similar chemical and biological characteristics.
     (2)  the rabbit alveolar macrophage (RAM) cytotoxicity
          bioassay can provide a consistent, ordinal ranking
          of particulate samples, based on their acute cellular
          toxicity.
                                 1

-------
      (3)   the mutagenic screening test  is capable of indicating
           that  some of the industrial particulate samples possess
           positive mutagenic activity.
      (4)   no strong correlation has been established between
           the  chemical characteristics  of the particulate samples
           (elemental  and  partial organic characterization)  and
           their observed  biological activities  (acute  cytotoxic
           and mutagenic behavior).

 The  chemical analyses have provided comparable results, both  from
 independent laboratory analysis as well as independent testing
 methods.   The expense and sophistication of each technique  reflects
 the  intent of the screening program.  The RAM cytotoxicity  bioassay
 is capable of providing a consistent, ordinal ranking of particulate
 samples based on their acute cellular toxicity.  If the desired rank-
 ing  should reflect proportional differences in observed cytotoxicities,
 then the  current RAM testing protocol must be enlarged to include more
 definitive concentration—response information.  For more intensively
 studied priority streams, additional RAM response parameters  (e.g.,
 functional impairment, membrane integrity) will supplement  the
 currently used  index of cell viability  (dye exclusion).

 The  mutagenic screening test has indicated that several particulate
 samples possess mutagenic activity, under the test conditions.  It
 should be  recognized that a positive mutagenic screening test using
£. typhimurium  is the first step in a battery of tests to evaluate
 the  mutagenic hazard of the particular sample.  Additional  solvent
vehicles  and microbial tester strains can be added to  the experimental
protocol,  as warranted.

-------
The sensitivity and specificity of a testing program must be compati-
ble with the test program's intent.  If a large number of pollutant
emissions are to be screened, so that the more hazardous ones can be
identified and intensely studied, then the initial screening sequence
need not be extremely sensitive or specific.   Both bioassay procedures
have indicated their potential utility in assessing the specific
biological activities of industrial praticulate emissions.

-------
                        2.0  INTRODUCTION

 The Process Measurements Branch of the Industrial Environmental
 Research  Laboratory  (EPA/RTP)* is developing a phased sampling and
 analysis  strategy  that provides for the environmental source assess-
 ment of industrial and energy processes.  In this phased approach,
 initial survey  testing (Level 1) is used to evaluate the potential
 environmental hazards of pollutant or process streams through examina-
 tion of their physical and chemical characteristics as well as their
 biological activity.  Those streams identified as potentially hazard-
 ous will  be subjected to more intensive testing procedures  (Level 2
 and 3) on a priority basis.

 At  all phases of the analytical program, the sophistication of each
 measurement technique is compatible with its companion techniques and
 with the  quality of the sample to be assayed.  The ultimate goal of
 an  environmental source assessment is to insure that the waste streams
 from a given process are environmentally acceptable or that adequate
 technology exists  for control.

This report presents several methods which have been used to charac-
terize the hazards associated with particulate material emitted from
industrial sources.  Although there are alternative approaches and
techniques which could be used to define environmentally hazardous
streams,  these selected methods were chosen to be both complimentary
and cost-effective.

2.1  APPROACH
The development of a rapid, effective, and inexpensive means for
evaluating the potential hazards associated with particulate emissions
*PMB/IERL, Environmental Protection Agency, Research Triangle Park
 North Carolina  27711                                             '

-------
is an essential part of PMB's environmental source assessment program.
To this end, two methods, chemical analysis and cellular bioassay,
were carried out on size-classified particulate material collected at
nine industrial sites.  The experiment was designed to determine
whether the chemical analyses or bioassay procedures, alone or in
combination, could assess potential hazards associated with particu-
late material.

Chemical analysis, alone, cannot provide sufficient data for complete
evaluation of pollutant emissions, because the biological activity of
the samples cannot be consistently predicted.  Interactive effects
(e.g., synergism, antagonism) between chemical constituents are
difficult to assess.  The biological availability of a given con-
stituent is likewise not easily predicted from chemical data alone.
The presence or absence of given toxic components neither precludes
nor indicates a relationship of the effluent to a suspected toxic
effect.  Chemical analysis should not be restricted to an a priori
determination of known hazardous or toxic compounds.  The possibility
of overlooking unanticipated, biologically active materials must be
avoided.

Bioassay techniques can be used effectively to assess the biological
activity of particulate samples.  Classical whole animal, in vivo
bioassay methods, as well as cellular in vitro tests, have been
developed to monitor the potential effect of pollutants on living
systems.  The advantages of cellular bioassay include its relatively
low cost, small sample requirement, and short experimentation time;
hence, its appeal for rapid evaluation of numerous, potentially
hazardous compounds.  Criticism of cellular bioassay suggests that
unsuspected effects may be missed since the whole animal with
potential target organs is not being considered.  Aspects of this

-------
problem can be overcome by judicious choice of cell types and critical
selection of cytological and biochemical test parameters.

Whereas cellular bioassay takes into account the biological activity
of a given sample, it cannot specify the compounds in a crude sample
responsible for the observed effects.  A cost-effective screening
method could involve the use of cellular bioassay to determine which
effluent samples are biologically active, together with the use of
chemical fractionation and analysis to ascertain which agents are
responsible for the observed effects.

The rabbit alveolar macrophage (RAM) cytotoxicity bioassay and
mutagenicity screening (using three Salmonella typhimurium bacterial
strains) were utilized to predict the acute toxicity and mutagenic
behavior of size-classified particulate samples collected at nine
industrial sites.  Partial organic characterization, with emphasis
on polycyclic hydrocarbons of known carcinogenic potential (or
structurally similar compounds), and inorganic elemental analysis
were performed on the same size-classified particulate samples.
These techniques should not be construed to represent a Level 1
sampling and analysis strategy, but rather a composite of selected
protocols from both Levels 1 and 2.

2.2  OBJECTIVES
The purpose of this research effort is to evaluate the effectiveness
of selected testing methods in accomplishing the following objectives:
     (1)   To determine whether the sequential cyclone sampling
          train can provide sufficient size-classified samples
          for chemical and biological tests, and whether this
          classification is useful;

-------
(2)   To determine whether the RAM bioassay can provide a
     reliable estimate of the acute cellular toxicity of
     particulate samples;
(3)   To determine whether the mutagenic screening test is
     capable of indicating a positive mutagenic response
     to the particulate samples;
(4)   To determine whether the chemical analyses can be
     correlated to the observed biological activity of
     the samples.

-------
                     3.0  SAMPLE COLLECTION
 3.1  FABRICATION AND CALIBRATION OF SAMPLING TRAIN

 A  series cyclone sampling train was developed to provide the capabil-
 ity  to collect sufficient quantities of size-classified particulate
 material (>_300mg per size range) from a variety of industrial sources
 so that subsequent chemical and biological characterization of  the
 sample could be achieved.  The cyclone train design called for  col-
 lection of samples according to particle aerodynamic diameter,  in
 ranges of >lOy, 3-lOy,  l-3y, and 
-------
                                                     VACUUM GAGE
PITOT
SAMPLING
PROBE
                              I	
          VAC.
          LINE
 HEAT
 TRACE
 WRAPPING
                                                                              VACUUM GAGE

                                                                              COARSE ADJ. VALVE

                                                                              FINE ADJ.
                                                                              BYPASS VALVE
X = LOCATION OF THERMOCOUPLES
D = DISASSEMBLY POINT FOR CLEANING
AIR TIGHT
VAC. PUMP
                                             DRY TEST
                                             METER
                                 ORIFICE AP
                                 MAGNEHELIC GAGE

                                          FIGURE 1
                       SERIES CYCLONE TRAIN, FIELD SAMPLING CONFIGURATION

-------
 Due  to time constaints, TRW Systems Group was only able to calibrate
 the  sampling train as follows:
      1)  generate a nebularized quantity of polydispersed talc  and
         pull it through the train;
      2)  measure the amount of talc collected in each cyclone;
      3)  assume an "S"-shaped efficiency curve for each cyclone with
         corresponding unspecified constants; and
      4)  determine the efficiency curve constants which produce the
         cyclone residues from the original samples of known  size
         distribution.
 The  design criteria of the sampling train were best met when  the flow
 rate through the train was 3 scfm.  The computed cyclone cut-off
 diameters (Dsr|s) were 9.5y, 2.0y, and 0.5]i for the three cyclones,
 respectively.
The  sampling train, as prepared by TRW Systems Group,  should  not  be
considered optimal since the intent was to provide a  fail-safe  sam-
pling  system capable of collecting a variety of particulate samples
                                                          (2 3)
under  a variety of operating conditions.  In its reports,   '    TRW
Systems Group urges that the train be redesigned and  recalibrated to
minimize weight, to reduce the number of components,  to  reduce  the
amount of sample wall-loss, and to identify more precisely  the
minimum critical flow that gives acceptable size classifications.
In addition, the use of a more suitable filter material  was advised,
since  the type used in this study was subject to thermal degradation
during sampling train operation.
3.2  SAMPLING SUPPORT PROCEDURES

Since particulate material collected would undergo  biological testing
and chemical analysis, an intensive effort was  expended during field
                                 10

-------
operations to insure sample integrity.  All equipment  that would  come
into contact with the particulate samples was cleaned,  inspected, and
packaged according to procedures demonstrated under  the VBLI Program.
To avoid unnecessary field cleaning, appropriate components of  the
sampling train, storage containers, and transfer equipment were
pre-cleaned and packaged in the laboratory clean room which met
Class 10,000 requirements of Federal Standard No. 209.     Nylon
drapes were used to isolate the sample containers (pre-weighed
Nalgene) from the rest of the environment during field  transfer of
the sample material.
Upon return to the TRW facilities, the samples were transferred to
pre-cleaned, tared, high density polyethylene storage bottles.  All
                                                           (4)
transfers were performed in a Class 1000 laminar flow bench    to
prevent particulate contamination from the surrounding environment.
This transfer technique was also followed for sample disbursement.
When necessary, only polyethylene utensils were allowed to touch
the sample.  All samples were stored in a limited access safe situ-
ated in a suitably controlled environment.
3.3  FIELD SAMPLING

Industrial sites sampled were selected to provide particulate material
possessing a variety of physical and chemical characteristics, as
well as an anticipated range of cellular toxicity.  In order to
evaluate the cyclone train, a wide range of sampling conditions under
which the cyclone train would operate was chosen.  Table 1 provides
the sampling logistics at the ten industrial sites.*  The sampling
*Sample collected at one industrial site was insufficient for further
 analysis.
                                  11

-------
                                                       TABLE 1



                                          LOGISTICS  OF SAMPLE  COLLECTION
Source
Sampling
location







T° stack*
T° lOy cyclone*
T oven*

Flow rate
through
sampling
train
Total
sampling
time


Open
Hearth
Furnace
Electrostatic
precipitator
(4 duct diam-
eters down-
stream from
ESP, 4 duct
diameters
upstream of
stack)
350°-425°
350°
350°-400°

4.8 scfm



5 hours
Coke
Oven
Heater
Base of
stack







400°
380°
380°-
400°
3.5 scfm



5 hours
continuous continu-
Basic
Oxygen
Furnace
Downstream
of ESP,
downstream
of induc-
tion fan,
upstream
of stack


150-225°
270°
270°

4.6 scfm



5 hours
continuous
ous
( ;
:
Iron
Sintering
Plant
Inlet to
baghouse







400°
350°
390°

5 scfm



2 hours
intermittent
over 5 hour
period

Oil
Fired
Power
Plant
Wet
scrubber
inlet






170°
195°
250°

4.5 scfm



2 hours
continuous



Clay
Aggregate
Plant
Between
primary
and
secondary
cyclones




510°
400°
400°

3.8 scfm



1.25 hours
intermittent
over 5 hour
period

Copper
Smel ter
Outlet
of
roaster
rever-
berator
inlet
to bag-
house

250°
275°
300°

4.7 scfm



1 hour
continuous


Aluminum
Smelter
Inlet to
bag-
house






210°
300°
300°

5 scfm



2 hours
intermittent
over 5 hour
period
Minicipal
Sludge
Incinerator
Duct
between
furnace
and
water
quench



1100°
410°
380°

4.7
scfm


5 hours
continuous


! i
Kraft
Mill
Process
Stack
effluent
from ESP






335°
335°
350°

4.8
scfm


5 hours
continuous



*Degrees fahrenheit

-------
locations at the industrial sites were not  consistent with  respect  to
control devices and the particulate material  collected  should  not
be construed to represent  the actual  emissions of  that  particular
site.
The specific sampling location at each  industrial  facility was
selected to provide a representative  sample of  the  effluent  stream.
Total sampling time was limited to a  maximum of five hours;  in  those
instances where particulate grain loading was high, intermittent
sampling periods over the  five-hour interval were  utilized.  Site
characteristics, as well as more specific sampling  locations are
presented in the following sections.


3.3.1  Steel Mill, Open Hearth Furnace

The open hearth furnace system sampled  in this  study was of  the
oxygen-lanced variety with a production capability  of 225 tons
per heat.  The sampling port was located approximately four  duct
diameters downstream from  the electrostatic precipitators and about
four duct diameters upstream from the entrance  to  the base of the
stack.  The sampling port  itself was  located sixty  feet above ground
level.  Sampling time was  five hours, continuous.


3.3.2  Steel Mill, Basic Oxygen Furnace

The basic oxygen furnace sampled in this study  has  a production
capability of 109 tons per heat.  The sampling  port was located
downstream from the electrostatic precipitator  and  induction fan,
upstream from the stack, and about six  feet above  the site floor.
Total elapsed sampling time was five  hours.

                                 13

-------
 3.3.3   Steel Mill,  Coke Oven Heater

 The  coke  oven  operation sampled  in this  study  consisted  of  a number of
 coke oven batteries,  each battery containing forty-five  ovens.   The
 particulate sample  was collected at  the  base of  a  stack  serving one
 of  the  batteries.   Effluent from this  stack is representative of
 the  combustion of the by-product coking  gas used to  heat the coke
 ovens.  The by-product coking  gases, having been passed  through
 electrostatic  precipitators, cooled  and  compressed,  are  burned  in  the
 fire boxes to  heat  the ovens.  The sampled stream  is representative
 of only a minor portion of the coking  process.   The  particulate
 samples were obtained over a five-hour period.


 3.3.4   Steel Mill,  Iron Sintering Plant

 The  sintering  process sampled  in this  study employed two Dwight Lloyd
 sintering machines  with combined capacity of 850 tons per day.   The
 sampling  site  was located upstream from  the baghouse entrance.   The
 gas  stream itself was generated  by the sintering process.   The
 particulate samples were obtained during two hours of intermittent
 sampling  over  a five-hour period.


 3.3.5   Oil-Fired Power Plant

 The  oil-fired  power plant sampled for  this study was a horizontally
 fired unit, fired at  right angles to the walls of  the rectangular
 firebox.  The  sampling port was  located  on the inlet duct to the
 low-energy wet  scrubber.  High-sulfur  fuel oil was being combusted
 during  sampling.  The sample was collected continuously  over a  two-
hour period.   A power plant shutdown precluded a longer  sampling
period.

                                 14

-------
3.3.6  Copper Smelter

Particulate samples were collected  from  the  side  of  a  tear-shaped
horizontal duct issuing from the  roaster/reverberator  process  up-
stream from the baghouse.  The  sampling  effort  experienced  extremely
heavy grain loading; the sampling probe  became  clogged after one
hour of operation, at which time  the  sampling effort was  terminated.
Sampling at the copper smelter  was  the only  instance of a high grain
load process that did not utilize an  intermittent sampling  schedule
to obtain an integrated sample  over a five-hour period.


3.3.7  Aluminum Smelter

The duct sampled at the aluminum smelting complex contained effluent
resulting from the electrolytic process.  The sampling probe was
inserted into the horizontal effluent duct upstream from the bag-
house.  Since high grain loading conditions  existed, fifteen-minute
sampling intervals, once every  hour for  five hours, provided the
representative sample.


3.3.8  Paper Mill, Kraft Pulping

Sample collection occurred in the recovery furnace effluent stream,
downstream from -the electrostatic precipitator.  Accumulated sample
mass after five hours of operation was negligible (O.OOSg), with
approximately ninety percent being deposited in the  1-3 micron
cyclone.
                                  15

-------
3.3.9  Ceramics Plant, Clay Aggregate Production

Particulate effluent from a ceramic aggregate production plant rotary
kiln furnace was collected.  The sampling probe was situated between
the primary and secondary cyclones approximately fifteen feet above
ground level in the side wall of a vertical circular duct.  Since
heavy grain loading conditions existed, fifteen minute sampling
intervals, once every hour for five hours, provided the sample.


3.3.10  Municipal Waste Water Sludge Incinerator

The sludge incineration system sampled in this study utilized a
three-stage spiral design, where sludge is injected at the top and
is directed downward through successive stages of incineration,
until the final ash product is removed at the bottom.  Due to extreme
temperatures,  direct sampling of the incinerator was not feasible.  A
water-cooled sampling probe was inserted on the pre-cooler portion of
the incinerator outlet.  Sampling time was five hours, continuous.
                                  16

-------
                         4.0   SAMPLE ANALYSIS


The  size-classified  particulate  material  collected  at  the industrial
sites was  subjected  to  three chemical  analyses  as well as two  cellular
bioassays.   Particulate samples  collected in  the largest  cyclone  (>10y)
were not analyzed  because  it was felt  that these particles would
settle  out  of  the  atmosphere in  a short period  of time and thus,
represented a  minor  air pollution hazard.   Table 2  provides the
disbursement schedule for  the samples  collected.  All  chemical and
biological  tests were not  run on every sample collected,  and Table  2
must be consulted  for the  actual tests conducted on a  particular
sample.  Analytical  methodologies for  those tests follow.


4.1  CHEMICAL  ANALYSIS

Three types of chemical analysis were  performed on  the  size-classi-
fied particulate material.   Elemental  composition of the  samples was
determined  by  spark  source mass  spectrometry.  Partial  organic char-
acterization,  emphasizing polycyclic hydrocarbons, was obtained by
gas  chromatography-mass  spectrometry and high resolution mass
spectrometry.


4.1.1  Spark Source Mass Spectrometry

Spark source mass  spectrometry (SSMS)  was  performed on  the  industrial
particulate samples by Accu-Labs  Research.*  Assays were  performed
*11485 W. 85th Avenue, Wheat Ridge, CO 80033

                                 17

-------
                                                      TABLE 2
                                              SAMPLE DISBURSEMENT

SOURCE
Sample Size (u)
TOTAL SAMPLE
COLLECTED (rag)
Spark Source
Mass Spectrometry
Gas Chromatography-
Mass Spectrometry
High Resolution
Mass Spectrometry
Cytotoxicity
Bio assay
Mutagenicity
Bio assay

STEEL PLANT
Open Hearth
Furnace
1-3 3-10 Fa
1,718.6 243.2
X X
X XX
X X
X XX
X

STEEL PLANT
Coke Oven
Heater
1-3 3-10 F
297.9 14.5
X
X X
X
XXX


STEEL PLANT
Basic Oxygen
Furnace
1-3 3-10 F
1,254.5 116.2
X X
X XX
X
X XX
X

STEEL PLANT
Iron Sintering
1-3 3-10 F
1,054.4 1,774.7
X X
X XX
X X
X XX
X
OIL FIRED
POWER PLANT
1-3 3-10 F
90.6 0.4
X
X

X X


QOTIRPF
Sample Size (y)
TOTAL SAMPLE
COLLECTED (mg)
Spark Source
Mass Spectrometry
Gas Chromatography-
Mass Spectrometry
High Resolution
Mass Spectrometry
Cytotoxicity
Bioassay
Mutagenicity
Bioassay

CERAMICS PLANT
Clay Aggregate
Rotating Kiln
1-3 3-10 F
4,221.4 12,788.4
X X
XXX
X X
XXX
X X
COPPER
SMELTER
1-3 3-10 F
598.9 2,193.6
X X
XXX
X X
XXX
X X
ALUMINUM
SMELTER
1-3 3-10 F
983.3 1..613.6
X X
XXX
X X
XXX
X X
WASTE WATER
TREATMENT PLANT
Sludge
Incinerator
1-3 3-10 F
1,613.5 13,357.3
X X
XXX
X X
XXX
X X
 F - filter (>ly)

Note:   In addition, a pulp and paper mill was sampled; unfortunately, the sample mass collected was to small for
       any subsequent analysis to be performed.
                                                           18

-------
for  seventy-five  different  elements.   The SSMS technique provided a
lower  detection limit  of  0.1  ppmw* for each element.   Additional
semi-quantitative scans  (geo-scans) were run to identify concentra-
tions  of  those elements  present  in quantities greater than one
percent by weight.   No repetitive  analyses of individual particulate
samples were  performed.   The  quantitative SSMS results are estimated
to be  accurate within  two hundred  percent while the geo-scan's
accuracy  is within approximately five  hundred percent.


4.1.2  Gas Chromatography-Mass Spectrometry

 Gas chromatography-mass  spectrometry  (GC-MS) analysis of twenty-four
particulate samples was  performed  by Battelle Columbus  Laboratories.**
The  analysis  focused on  polycyclic hydrocarbons of known carcinogenic
potential or  structurally similar  compounds.   Each sample  was  subjected
to ultrasonic extraction  with methylene  chloride at 50° C, and the  re-
sulting solution  was recovered by  centrifugation.  Internal standards
were added before the  volume  of  the solution was reduced to 200  yl;
the  added standards were  9-methylanthracene,  9-phenylanthracene, and
9,10-diphenylanthracene.  All samples  were analyzed using  GC-MS with
quantification by specific  absolute ion  current integration.   The
detection limit for individual polycyclic  organic species  was  slightly
less than 10 ng.  No repetitive analyses  of individual particulate
samples were performed.


4.1.3  High Resolution Mass Spectrometry

High resolution mass spectrometry  was  performed  by a United States
Energy Research and Development  Administration  facility (ERDA/PERC)
 *parts per million by weight
**505 King Avenue, Columbus, OH 43201
                                 19

-------
 in  Pittsburgh, Pennsylvania*.  High resolution mass  spectrometry has
 the capability of determining the precise mass of various  polynuclear
 hydrocarbons  from which  the chemical formula can be  derived, but
 isomeric  identification  cannot be performed by HRMS  alone.  High resolu-
 tion mass spectrometry can be used for the preliminary screening of
 complex mixtures for  the possible presence of several hundred hazard-
 ous and/or toxic compounds.


 The particulate samples  were vaporized** and the components observed
 in  the mass spectra.  High resolution mass spectra were  recorded on
 photographic  plates,  and the data processed by computer.


 Semi-quantitative mass spectral analyses of two particulate samples
were  obtained by successive scans over the period of time during which
 the sample yielded vaporization products.  Mass spectra  from all
 fourteen  particulate  samples analyzed were screened  by computer for
 nine  precise masses that would indicate the possible presence of
 carcinogenic  polynuclear aromatic hydrocarbons.


 4.2  BIOLOGICAL CHARACTERIZATION

 In  this study, two in vitro bioassays were utilized  to determine the
acute toxicity and the mutagenic potential of the 3-lOy,  l-3y and  >ly
particulate samples from industrial sources.  The rabbit  alveolar
macrophage  (RAM) has been used to determine the potential  acute
cytotoxicity of the samples.  The alveolar macrophage exists as a
pulmonary free cell and  provides an early line of defense  against
 *Pittsburgh Energy Research Center, 4800 Forbes Avenue,  Pittsburgh,
  PA 15213.
**@300"C, 10~6 torr

                                  20

-------
 inhaled foreign bodies.  Because of its phagocytic activity, it is
 particularly useful in the toxicologic evaluation of airborne particu-
 late matter.  The mutagenic bioassay utilizes several bacterial
 indicator strains (histidine deficient Salmonella typhimurium strains
 TA-1535, TA-1537, and TA-1538), with reversion to prototrophy indica-
 tive of mutation.  Both bioassays have indicated their utility in
 studying the effects of certain pure compounds, but neither has been
 used extensively on complex mixtures.

 4.2.1  Cytotoxicity Evaluation

 Northrop Services, Inc.,* under contract with the Experimental Biology
 Laboratory** (EBL/RTP), performed the rabbit alveolar macrophage (RAM)
 cytotoxicity test, according to a modification of the procedure
 developed by Waters et al.  *    The RAM culture medium was added to
 pre-weighed particulate samples to achieve a desired final particulate
 concentration in the medium.  The complete culture medium consisted
 of Medium 199 in Hanks' salts supplemented with 20 percent heat-
 inactivated fetal bovine serum, 100 units/ml penicillin^G, 100 yg/ml
 streptomycin sulfate and 100 yg/ml kanamycin.  The particulate samples
 were incubated with continuous agitation on a rocking platform (12
 oscillations per minute) for 20 hours in the culture medium (less
 serum but including antibiotics) to allow for dissolution of any
 soluble components of the particulate matter.  Macrophage .cells were
 then added (along with the bovine serum) to achieve a final concen-
 tration of approximately 5 x 10  cells/ml.  The cultures were returned
 to the roller platform and incubated for 20 hours (@ 37° C in a
 humidified 5 percent CO  atmosphere).  The culture medium containing
 *Box 1484, Huntsville, AL  35804
**EBL, Environmental Protection Agency, Research Triangle Park, NC
  27711

                                  21

-------
 unattached cells was  then poured off and retained.  Cells remaining
 attached to the culture vessel were removed using trypsin and recom-
 bined with the original culture medium.  Cell number was determined
 by direct count using a hemocytometer.  Cell viability was estimated
 by light microscopy on the basis of trypan blue dye exclusion.

 An initial cytotoxicity screening of the particulate samples was
 performed at a final  particulate concentration of 1000 yg/ml of
 culture medium.  Samples found in the initial screening to produce
 net cell death of  greater than 15 percent, as compared with controls,
 were retested in a preliminary concentration-response test using
 particulate concentrations of 1000 yg/ml, 300 yg/ml, and 100 yg/ml
 of culture medium.  The pH values of the cultures were monitored
 throughout,  and if shifts below 6.8 or above 7.6 occurred, the sample
 was tested under both unadjusted and adjusted conditions.

 In addition,  an attempt was made to ascertain whether the toxicity
 of a given particulate sample was due to the particles  themselves
 and/or soluble component(s) released into the medium.  The particulate
 matter was incubated  in the culture medium (less serum but including
 antibiotics)  for 20 hours and then removed from the medium via
 centrifugation and filtration through a 0.22y Millipore filter.   The
 filtered  supernatant  and centrifuged particles (resuspended  in fresh
 medium) were  then  independently tested for cytotoxicity.  Figure  2
 indicates  the testing sequence used in this study.

The filters used to collect the sub-micron particles were cut  into
quarters, desiccated, weighed, and pre-incubated with sterile  deionized
water plus antibiotics for twenty hours.  RAM cells were added  to
half of the pre-incubated filter samples (final concentrations:
5 x 10  cells/ml)   after addition of the Medium 199  concentrate plus
                                 22

-------
        PARTICULATE  SAMPLE
     INITIAL SCREENING TEST
         (at lOOOpg/ml)
 viability after 20 hrs exposure
  to particles plus pre-incuba-
 tion supernatant (in triplicate)
  "PARTICLES PLUS SUPERNATANT"
(if viability of  initial
screen <85% of controls)
                           PRELIMINARY CONCENTRATION-RESPONSE TEST
       CONCENTRATION:
         lOOOyg/ml
    CONCENTRATION:
       300yg/ml
         CONCENTRATION:
             lOOyg/ml
    VIABILITY AFTER 20 HR
    EXPOSURE TO PARTICLES
     PLUS PRE-INCUBATION
         SUPERNATANT
       (in triplicate)
"PARTICLES PLUS SUPERNATANT"
 VIABILITY AFTER 20 HR
 EXPOSURE TO PARTICLES
WITHOUT PRE-INCUBATION
      SUPERNATANT
     (in triplicate)
      "PARTICLES"
•VIABILITY AFTER 20 HR EXPOSURE
  TO PRE-INCUBATION SUPERNATANT
         (in  triplicate)
         "SUPERNATANT"
                                               FIGURE 2
               RABBIT ALVEOLAR MACROPHAGE CYTOTOXICITY SCREENING TEST PROCEDURE

-------
 serum to reconstitute the complete incubation medium described
 previously.   The cultures were incubated  with agitation for twenty
 hours,  and the cell number and viability  noted ("filter plus super-
 natant" fraction).   The remainder of  the  pre-incubated filter samples
 were removed from the solution and RAM cells  were added to the
 supernatant  (final  concentration:   5x10  cells/ml),  incubated for
 twenty  hours,  and  cell number  and  viability determined ("supernatant"
 fraction).   The filters that were  removed were dried,  weighed,  and
 pre-incubated  with  sterile deionized  water  plus antibiotics a second
 time.   After addition  of medium  concentrate  and  serum,  RAM cells  were
 added  (final concentration:   5x10   cells/ml),  incubated  for  twenty
 hours,  and  cell  number and viability  noted  ("dried  filter" fraction).
 4.2.2   Mutagenicity  Evaluation

 Litton-Bionetics,  Inc.,* under  sub-contract  to Research  Triangle
 Institute**  performed  the mutagenic  screening tests according  to  a
 modification of  the  procedure developed by Ames et al.          Three
 histidine deficient  Salmonella  typhimurium strains (TA-1535, TA-1537,
 and TA-1538)  were  used, with reversion to prototrophy indicative  of
 mutation.  The TA-1535 strain is most likely to undergo  base pair
 substitutions, and the TA-1537  and TA-1538 frameshift reverse muta-
 tions.  All  three strains have  defective DNA excision repair systems
 as well as defective lipopolysaccharide coats, thereby increasing
 the sensitivity  of the strains  to observable mutational  events.


All particulate  samples were dissolved in a  single solvent, dimethylsul-
foxide (DMSO).   Exposure of the bacterial populations to the particu-
late material, using DMSO as the vehicle, occurred on plates by the
 *5516 Nicholson Lane, Kensington, MD 20795
**Box 12194, Research Triangle Park, NC 27709
                                 24

-------
agar overlay method.  In the event that the compound(s) might require
metabolic activation in order to exhibit mutagenic activity, all
bacterial tests were run in the presence and absence of a mouse liver
activation system.      Positive control tests were run on the
bacterial systems by exposing each to known active mutagens.  Since
DMSO was the solvent vehicle for each test, appropriate solvent con-
trols were run.  The toxicity of each particulate sample to the
bacterial populations over a range of exposures was determined prior
to the mutagenicity testing.  The highest doses used in the mutagenic
tests were restricted to those levels lethal to not more than twenty-
five percent of the exposed population in the toxicity screening.
This restriction minimized the potential for growth of non-mutant
cells (phenocopies-) utilizing histidine released from dead cells,
while it allowed for a reasonable exposure level in order to detect
only moderately mutagenic samples.  The lowest mutagenic test dose
was at least two orders of magnitude lower than a concentration
which produced detectable toxicity.  The experimental procedure used
in the mutagenicity screening test is summarized in Figure 3.
                                 25

-------
  PARTICULATE SAMPLE
       SALMONELLA TYPHPflJRIPM

 Strains:  TA-1535,  TA-1537, TA-1538
 Each strain tested  in the  presence
  and absence of mouse liver activa-
  tion system
     dissolved
        in
 DIMETHYLSULFOX1DE
      (DMSO)
Solvent  Controls
      Test
  (A* and NA**)
Positive Controls
      Test
  (A* and NA**)
                        TOXICITY DETERMINATIONS
                           for each strain,
                      activated and non-activated
                        MUTAGENIC SCREENING TEST
                           for each strain,
                      activated and non-activated
                        (with suitable controls)
 *Activated
**Non-Act ivated
                                FIGURES
  SALMONELLA TYPHIMURIUM MUTAGEIMICITY SCREENING TEST PROCEDURE
                                   26

-------
                          5.0  RESULTS
 5.1  CHEMICAL ANALYSIS

 The elemental   composition of  the industrial  particulate  samples,  as
 determined by spark  source mass  spectrometry,  is  provided in  Table 3.
 In addition  to  the Accu-Labs Research analyses, two  particulate
 samples, the coke oven  heater  l-3y sample and  the copper  smelter  l-3y
 sample, were analyzed for  major  and minor species by the  Analytical
 Chemistry Branch of  the Environmental Monitoring  and Support  Laboratory
 (EMSL/RTP).* All results  are  reported as parts per  million by weight,
 except as noted.  As evidenced in Table 3,  the elemental  concentrations
 of both particulate  samples collected from the same  site  are  not  always
 similar (e.g.,  sodium in the aluminum smelter  samples, beryllium  and
 aluminum in  the copper  smelter samples,  silver and uranium in the
 basic oxygen furnace samples.


 Results of gas  chromatographic-mass spectrometric analyses of certain
 samples are  provided in Table  4.   Polycyclic hydrocarbons are reported
 as parts per million by weight,  except for  filter analyses.   Due  to
 thermal degradation of  the  filter,  the amount of  particulate  material
 deposited on any filter was undetermined,  so filter  analyses  are
 reported as  total nanograms detected.   Three samples were not analyzed
 because of insufficient sample (i.e.,  coke  oven heater l-3y sample,
 and the oil-fired power plant  l-3y  and 3-lOy samples).  The GC-MS
 analysis focused on hydrocarbons  of known carcinogenic potential,  or
of structurally similar compounds.   When examining GC-MS  and  HRMS
*Environmental Protection Agency,  Research  Triangle  Park, NC  27711
                                  27

-------
                                              TABLE 3
                        ELEMENTAL ANALYSIS OF PARTICULATE SAMPLES AS
                       DETERMINED BY SPARK SOURCE MASS SPECTROMETRY8
E 1 ewent
Aluminum
Antimony
Arsenic
Barium
Beryllium
Bismuth
Boron
Bromine
Cadmium
Calcium
**Carbon
Cerium
Cesium
Chlorine
Chromium
Cobalt
Copper
Dysprosium
Erbium
Europium
Fluorine
Gadol inium
Gallium
Germanium
Gold
Hafnium
Holmium
"Indium
Iodine
Iridium
Iron
Lanthanum
Lead
Lithium
Lutecium
Magnesium
Manganese
Mercury
Molybdenum
Neodymi urn
Steel Plant
Open Hearth
Furnace ,
3-10 Micron
68
27
190
6,3
<0.24
11
47
10
5.4
2.0%
NR**
1.2
0.26
270
880
120
=3600



790

90
7.1



rs*
1.7

40. OX
0.21
900
59

1.0*
0.52*
NR
35
0.6S
Steel Plant
Open Hearth
Furnace,
1-3 Micron
140
70
480
12
0.34
21
31
43
44
0.83%
NR**
1.1
2.5
260
60
110
=2900



760

87
18



IS*
3.9

40.0%
0.20
=1900
110

1.0%
=4400
NR
79
0.65
Steel Plant
Coke Oven
Heater,
1-3 Micron
5.3 (I?)1
3.7
3.9 (200)
1.0 (15)
0.25
48 (135)
3.3 (15)
1.1
9.4 (131)
=1100 (220)
NR**
0.10
0.53
17
94 (170)
2.6 (11)
170 (190)



220 (100)

4.0
0.98



IS*
1.8(5.8)

>U (4500)
0.03
940 (1500)
6.1

170
250 (82)
NR
"(12)

Steel Plant,
Basic Oxygen
Furnace
3-10 Micron
73
17
67
68
<0.26
5.0
130
110
9.7
4.0%
NR**
5.9
2.8
3.0%
220
130
1100



5.0%

97
10



IS*
19

>80%
2.2
170
27

2.0%
1.0%
NR
38
0.73
Steel Plant,
Basic Oxygen
Furnace
1-3 Micron
21
9.7
52
40
<0.15
2.9
50
65
12
2.0%
NR**
0.73
1.2
3. OX
170
57
800



6.0%

57
6.0



IS*
2.6

40. OS
0.28
210
16

=2500
0.77%
flR
10
0.1C
Steel Plant,
Iron
Sintering,
3-10 Micron
=1300
2.5
35
60
0.57
5.7
61
150
14
40.0%
NR**
38
18
=3300
76
86
«1000
1.8
0.53
0.98
1.0%
0.41
6.3
2.9

0.63
0.13
IS*
29

6.0%
7.3
=2100
18
0.06
6.0%
640
NR
11
9.5
Steel Plant,
Iron
Sintering,
1-3 Micron
890
3.5
120
140
0.39
100
76
270
22
30.0%
NR**
46
25
1.0%
52
59
=2200
1.3
0.37
0.67
0.7%
0.23
4.4
9.2

0.44
0.09
IS*
69

2.0%
10
0.82%
28
0.24
4.0%
321
NR
17
6.5
Oil Fired
Power Plant,
1-3 Micron
340
7.8
9.3
280
8.1
1.1
16
10
0.97
1.0%
NR**
14
4.6
33
= 1000
220
300
0.31

0.33
31
0.10
190
4.7



IS*
0.61

1.0%
0.52
450
13

0.59%
220
NR
814
0.68
Copper
Smelting
1 -3 Micron
250 (940)+
8.0% (5.8%)
3.0% (21%)
170 (31)
5.9
2.0% (5700)
11
750 (52)
0.49% (3800)
0.75% (800)
NR**
3.9
1.9 (60)
610 135)
190 (153)
180 (64)
5.0% (2.9%)



210

6.5 (18)
140 (120)
25
1.3

IS*
160 (25)

3.0* (It)
4.1
35.0% (11.2*)
43

2.0* (1.1*)
310 (73)
***
0.51* (570)
9.8
  Reported as parts per million by weight,
  except as noted
  *IS:  Internal  Standard
 "Not Reported
***Mercury Observed
  T( ) Reported by independent laboratory (EMSL/RTP)
                                                     28

-------
                                             TABLE 3
                        ELEMENTAL ANALYSIS OF PARTICULATE SAMPLES AS
                       DETERMINED BY SPARK SOURCE MASS SPECTROMETRY a
                                           (CONTINUED)
Element
Nickel
Niobium
"Nitrogen
Osmium
"Oxygen
Palladium
Phosphorus
Platinum
Potassium
Praseodymium
*Rhenium
Rhodium
Rubidium
Rutheni urn
Samari urn
Scandium
Selenium
Silicon
Silver
Sodium
Strontium
Sulphur
Tantalum
Tellurium
Terbium
Thallium
Thorium
Thullium
Tin
Titanium
Tungsten
Uranium
Vanadium
Ytterbium
Yttrium
Zinc
Zirconium
Steel Plant
Open Hearth
Furnace ,
3-10 Micron
490
0.93
NR**

NR**

0.5%

0.69%
0.13
IS*

32


<0.16
2.7
0.89%
41
2.0%
4.8
2.0%
0.82
0.15

0.53
<0.23

340
33
11
0.33
290
0.78
2. OX
0.64
Steel Plant
Open Hearth
Furnace ,
1-3 Micron
470
0.9
NR**

NR**

=4900

3.0%
0.13
IS*

95


<0.16
2.3
• =4300
65
8.0%
11
4.0%
0.79
0.40

1.5
<0.22

610
15
10
1.5
330
0.51
2.0%
1.4
Steel Plant
Coke Oven
Heater,
1-3 Micron
51

NR**

MR**

41 (82)f

a!700 (1500)

IS*

21 (340)


>1%
9.8
200 (2500)
4.2 (35)
=2000 (700) '
1.2 (6.5)
>\t (7250)
0.18
0.85

7.5 (84)
0.24
'
66 (81)
8.0 (125)
0.87
3.4 (16)
38 (15.5)
0.08
=1100 (850)
0.66
Steel Plant,
Basic Oxygen
Furnace
3-10 Micron
93
2.3
NR**

MR**

=3500

4.0%
0.15
IS*

63


<0.18
0.51
1.0%
440
9.0%
43
=3500
0.44
0.15

0.78
<0.25

140
0.44
12
6.2
40
0.56
0.58%
0.69
Steel Plant,
Basic Oxygen
Furnace
1-3 Micron
46
2.7
NR**

NR**

= 1200

2.0%
0.04
IS*

74


<0.10
0.40
0.56%
43
.077%
14
5.0%
0.52
<0.08

0.97
<0.14

40
9.6
6.8
<0.14
23
0.71
=4600
0.40
Steel Plant,
Iron
Sintering,
3-10 Micron
26
6.5
NR**

NR**

= 1300

1.0%
0.95
IS*

180

1.1
1.1
89
3.0%
6.1
3.0%
160
0.31%
0.29
0.58
0.10
11
2.3
0.03
4.4
290
3.6
2.3
48
0.57
65
760
10
Steel Plant,
Iron
Sintering,
1-3 Micron
64
4.5
NR"

MR**

890

3.0%
2.8
IS*

=950


1.1
310
2.0%
35
5.0%
110
1.0%
0.20
2.0
0.15
35
3.7
0.09
6.1
370
2.4
3.7
33
0.21
25
=1100
14
Oil Fired
Powsr Plant,
1-3 Micron
5.2
0.22
NR**

NR**

=1600

=3500
0.14
IS*

29


<0.08
63
=4500
4.4
20.0%
57
7.0%
0.41
0.12
0.04
0.78
1.3

14
210
13
1.3
2.0?
13
900
2.2
Copper
Smelting
1-3 Micron
1300
31 (5.3)*
NR**

MR**

=1300 (180)

10.X (1600)
2.0
IS*

140 (105)


0.78 (<15)
V" F V \ ' ** /
0.55% (865
1.0% (3000
=1100 (420
3.0% (1700
29 (45)
2.0%
1.3
=<3600- (1700)

=3900 (875)
0.56

4.0% (3700)
560 (170)
180
5.6
49 (128)
19 (16)
1.0% (1.7%)
11 (9.8)
 aReported as parts per million by weight,
 except as noted
 *IS:  Internal Standard
**Not Reported
 f( ) Reported by independent laboratory (EMSL/RTP)
                                            29

-------
                                             TABLE 3
                       ELEMENTAL ANALYSIS OF PARTICULATE SAMPLES AS
                       DETERMINED BY SPARK SOURCE MASS SPECTROMETRY a
                                           (CONTINUED)
Element
Aluminum
Antimony
Arsenic
Barium
Beryllium
Bismuth
Boron
Bromine
Cadmium
Calcium
**Carbon
Cerium
Cesium
Chlorine
Chromium
Cobalt
Copper
Dysprosium
Erbium
Europium
Fluorine
Gadolinium
Gallium
Germanium
Gold
Hafnium
Holmium
*Indium
Iodine
Iridium
Iron
Lanthanum
Lead
Lithium
Lutecium
Magnesium
Manganese
Mercury
Molybdenum
Neodymi urn
Copper
Smelting
3-10 Micron
=2500
3.0%
6.0%
81
0.59
1.0%
23
750
=1400
4.0%
NR**
20
3.7
=1100
79
480
20.0%


1.0
370
0.31
13
24
58
0.65

IS*
160

20.0%
7.9
4.0%
18

1.0%
880
***
-2500
9.8
Al umi num
Smel ter
1-3 Micron
2.0%
130
=4100
5.2
2.9
270
72
86
16
=1300
NR**
0.45
7.5
690
90
43
=2000



>50%

900
3.4



IS*
31

4.0%
0.37
330
9.8

330
20
NR
67
1.1
Aluminum
Smel ter
3-10 Micron
=3500
66
7 DO
25
1.8
56
27
23
1.3
2.0%
NR**
0.14
0.27
140
56
27
=4600



20.0%

200
2.1



IS*
16

2.0%
0.05
73
13

870
6.3
NR
51

Ceramics
Plant
1-3 Micron
3.0%
10
890
480
2.3
140
100
14
13
15.0*
NR**
180
6.0
330
150
3.4
210
1.7
0.50
1.3
9.0?
0.89
60
6.3

0.9S
0.27
IS*
11

6.0%
37
60
390
0.32
10.0%
-1600
NR
23
45
Cerami cs
Plant
3-10 Micron
3.0%
5.3
110
290
1.2
79
230
7.6
6.6
15. 0*
NR**
200
6.6
200
120
18
93
4.5
1.0
1.8
1.0%
0.99
36
1.4

1.3
0.30
IS*
5.9

6.0%
27
33
430
0.35
6.0%
=*1300
NR
12
49
Sludge
Incinerator,
1-3 Hicron
0.9%
2.5
100
-1700
2.4
=1500
82
720
630

MR**
190
18
=4600
-4000
850
3.0%
1.8
0.26
0.97
-4300
1.5
130
2.8
84
2.7
0.13
IS*
280

7.0%
39

50
0.72
6.0%
850
NR
110
47
Sludge
Incinerator,
3-10 Hicron
0.69%
30
43
=2000
14
=1900
65
420
790
20.0%
NR**
220
11
=2700
=2300
=1000
3.0%
2.5
0.31
1.3
=2500
0.55
74
1.7
120
3.7
0.13
IS*
330


O* Ojo

DC
O9
0.37
7.0%
-1000
NR
33
55
  Reported as parts per million by weight,
  except as noted
  *IS:  Internal Standard
 "Not Reported
***Mercury Observed
  +(  ) Reported by independent laboratory (EMSL/RTP)
                                               30

-------
                                         TABLE 3
                   ELEMENTAL ANALYSIS OF PARTICIPATE SAMPLES AS
                  DETERMINED BY SPARK SOURCE MASS SPECTROMETRY a
                                      (CONCLUDED)
El ement
Nickel
Niobium
**Nitrogen
Osmium
**0xygen
Palladium
Phosphorus
Platinum
Potassium
Praseodymium
*Rhenium
Rhodium
Rubidium
Ruthenium
Samari urn
Scandium
Selenium
Silicon
Silver
Sodium
Strontium
Sulphur
Tantalum
Tellurium
Terbium
Thallium
Thorium
Thullium
Tin
Titanium
Tungsten
Uranium
Vanadium
Ytterbium
Yttrium
Zinc
Zirconium
Copper
Smelting
3-10 Micron
760
6.7
NR**

NR**

=1300

1.0*
2.0
IS*

43

3.6
1.7
a: 3400
3.0%
=2200
7.0*
82
2.0*
1.3
si 300

1100
1.1

=4600
=1100
79
2.4
98
8.1
0.7*
n
Al umi num
Smelter
1-3 Micron
=2800
1.8
NR**

NR**

0.76*

=4100
0.08
IS*

98


<0.17
22
740
2.0
8.0*
9.4
3.0*
1.5
26

3.0
<0.24

100
130
21
1.3
0.51*
2.2
160
5.3
Aluminum
Smelter
3-10 Micron
=3800
0.48
NR**

NR**

=2000

840
0.03
IS*

6.1


<0.08
14
230
1.1
27
50
0.6*
0.91
4.3

0.19
<0.12

33
140
20
0.40
=1500
1.4
56
1.5
Ceramics
Plant
1-3 Micron
88
29
NR**

NR**

0.61*

4.0*
13
IS*

460

2.5
15
42
40.0%
0.57
20.0*
=1200
=3800
0.54
2.7
0.36
10
10
0.12
36
0.51*
17
10
540
1.1
61
360
35
Ceramics
Plant
3-10 Micron
36
32
NR**

NR**

0.67*

4.0*
9.9
IS*

370

3.6
40
20
30.0*
0.64
15.0*
=1400
=4700
0.60
1.3
0.53
5.3
11
0.14
9.2
0.67*
3.7
5.6
210
1.8
38
220
38
Sludge
Incinerator,
1-3 Micron
=4500
30
NR**

NR**

10.0*

1.0*
9.4
IS*

22

1.1
3.8
73
10.0*
600
10.0*
=1300
0.8*
1.4
0.29
0.21
2.5
2.3
0.10
1.0*
0.53*
35
11
94
0.2?
18
0.56*
210
Sludge
Incinerator,
3-10 Micron
=3000
31
NR**

NR**
16
10.0*

1.0*
11
IS*

24

2.6
4.4
52
8.0*
500
8.0*
=1500
0.57*
3.4
0.34
0.30
5.3
3.1
0.08
0.5*
-3800
44
31
55
0.37
21
0.56*
120
 'Reported as parts per million by weight,
 except as noted
 *IS:  Internal Standard
**Not Reported
 f( ) Reported by independent laboratory (EMSL/RTP)
                                        31

-------
                                               TABLE 4
                              CARCINOGENIC AND STRUCTURALLY SIMILAR
                                POLYCYCLIC ORGANIC CONSTITUENTS IN
                             PARTICULATE SAMPLES AS DETERMINED BY GAS
                               CHROMATOGRAPHY-MASS SPECTROMETRY a
SOURCE
PARTICLE SIZE
IDENTIFIED COMPONENT
ANTHRACENE/PHENANTHRENE
METHYL ANTHRACENES
FLUORANTHENE
PYRENE
METHYL PYRENE/FLUORANTHENE
CHRYSENE/BENZ(A)ANTNRACENE
METHYL CHRYSENES
BENZO FLUORANTHENES
BENZO (A)PYRENE
BENZO (E)PYRENE
3-METHYLCHOLANTHRENE
INDENO (1,2,3, -CD)PYRENE
BENZO(GHl)PERYLENE
DIBENZ(A,H)ANTHRACENE
DIBENZO (c, G)CARBAZOLE
DIBENZO(A, 1 AND A,H>PYRENES
CORONENE
OPEN HEARTH FURNACE
1-3H

2,7

1.3
0,1

0,1

1,6
0,9
3

1,1
0,1
1,3

1,5

3-10,,

0,1

0,2




0,5
0.3
2







FILTERC

6785 ng

2155
222



281
29
98







COKE OVEN HEATER
1-3H

3,9
















3-10H

(B)
















FILTER0

830 rtg

132
190

200

110
73
500







BASIC OXYGEN FURNACE
1-3H

21.8

7.1
2,9



1.6

1.8







3-10,,

6.9

1,1
1













FILTER0

163ng




131

111

685

118
321




IRON SINTERING
1-3,

289,5
27.1
31.2
16,1
18,0
1,9

1,8
1,3
9,0

100
0,5




3-10H


















FILTER0

5929n9

1982
1512

127

633
191
2030

620
313
193

538

OIL FIRED POWER PUNT
1-3,,

(B)
















3-lOn

(B)
















FILTER0









(








SOURCE
PARTICLE SIZE
IDENTIFIED COMPONENT
ANTKRACENE/PHENANTHRENE
METHYL ANTHRACENES
FLUORANTHENE
PYRENE
METHYL PYRENE/FLUORANTHENE
CHRYSENE/BENZ(A)ANTHRACENE
METHYL CHRYSENES
BENZO FLU08ANTHENES
BENZO (A)PYRENE
BENZO(E)PYRENE
3-METHYLCHOLANTHRENE
INDENO (1,2,3, -CD)PYRENE
BENZO(GHI)PERYLENE
DIBENZ(A, H)ANTHRACENE
DIBENZO(C,G)CARBAZOLE
DIBENZOtA, I AND A,H>PYRENES
CORONENE
CERAMICS PLANT
l-3p

22

6,2
1,1
3,9


5,8
5.8
13,8

3,6
0,8
1.5



3-10H


















FILTER0

126ng


36
82
23
57
19
29








COPPER SMELTER
1-3H

0.5
















3-lOn


















FILTER0


















ALUWNUK SMELTER
I-*

21.7
9
31.1
12.9
32.5
93.6
25
218
16.5
807.2
59.5
311.1
126
1387
81.7
352
87.3
3-lOp

172.6
51,3
137.2
155.8

287.3
79.7
786.3
105.6
1011
23.5
219.2
786.3
1213.3
77.6
216.1
50,2
FILTER0

11216ng
1706
20918
21738
27222
26198
7266
59016
121238
250692
27159
39959
19180
151225
6988
1172
6821
SLUDGE INCINERATOR
1-3H


















3-10H


















FILTER0


















"REPORTED AS PARTS PER MILLION BY WEIGHT

BNOT ANALYZED

°ONE QUARTER OF FILTER MATERIAL ANALYZED
 (NOTE: FILTER ANALYSES REPORTED AS NANOSRAMS PRESENT)
                                                   32

-------
results, one should note that if collection cyclones were operating
at temperatures above 350 F, much of  the  organic  fraction would have
been lost as vapor (see Table 1 for sampling  logistics).
Table 5 indicates results  from the high resolution mass spectra of
the vaporized constituents of the particulate  samples, and the
hydrocarbon structures  identified from those spectra.  Since studies
of pure 4-, 5-, and 6-ring aromatic hydrocarbons have shown that
their rate of vaporization varies with both the number of aromatic
rings and the type of condensation (peri or cata), the aluminum
smelter samples were subjected to additional mass spectral scans made
over the period of time during which  the samples continued to yield
vaporization products.  The  semi-quantitative  results, presented in
Table 6, are more representative of the hydrocarbon constituents
of those samples than obtainable by the routine analytical method.
Eighty-seven percent of the  aromatic  hydrocarbon content of these
two samples was concentrated in 4- to 6-ring aromatic systems.
Mass spectra from all samples analyzed by HRMS were screened for nine
precise masses indicating  the possible presence of carcinogenic
polycyclic organic hydrocarbons  (see Table  7).  Since HRMS alone
cannot determine the isomeric form of a compound with a given mass
number, the presence of a  compound with the precise mass corresponding
to a known carcinogen is not conclusive evidence of the carcinogen's
presence.  The mass spectra of both aluminum  smelter samples (l-3y
and 3-10 y) indicated the  possible presence of all carcinogens listed
in Table 7.  The precise masses  indicating possible carcinogens were
not detected in any of the other samples analyzed.
                                  33

-------
                                                           TABLE 6
                   CONSTITUENTS OF PARTICULATE SAMPLES AS DETERMINED BY HIGH RESOLUTION MASS SPECTROMETRY"
SAMPLE ORIGIN
Open Hearth
Furnace
Open Hearth
Furnace
Coke Oven Heater
Basic Oxygen
Furnace
Iron Sintering
Plant
Iron Sintering
Plant
Copper Smelter
Copper Smelter
Aluminum Smelter
Aluminum Smelter
Ceramics Plant
Ceramics plant .
Municipal
Incinerator
Municipal
Incinerator
PARTICLE SIZE
(MICRONS)
3-10
1-3
1-3
1-3
3-10
1-3
1-3
3-10
1-3
3-10
1-3
3-10
1-3
3-10
PERCENT OF SAMPLE
VAPORIZED
2.1
0.12
5.3
3.2
7.3
37.6°
n.a.d
d
11 4 a.
1.3
5.2
2.4
2.1
4.0
0.5
GASES EVOLVED
(S300°C, 10-6 torr)
HCN, CH3CN, HC1, C02, N02, EtOH,
CH3COOH, S02, COS, CO
HCN, CH3CN, HC1, NO, N02, H2S,
C02, EtOH, CO, COS, S02, CS2
HCN, CH3CN, CO, NO, H2S, C02,
N02, S02, COS, CsH6, CS2
HCN, CH3CN, CO, NO, HC1, C02,
N02, EtOH, S02, COS
HCN, CH,CM, CO, NO, K2S, HC1,
C02, K02, EtOH, SOj, COS
HCN, CHjCN, CO, NO, HC1, C02,
N02, EtOH, S02, COS, H2S
S02, CS2, CO, C02
HCN, CO, HC1, NO,, SO,, CS,,
COS 2 Z
CHN, CO, NO, HC1, CO , NO ,
SO COS H S
CHN, CO, NO, HC1, CO., NO ,
so2, cos * f
HCN, CO, NO. MeOH, HC1, SO,,
N02, H2S, COS '
HCN, CO, KO, HC1, CH CN, CO,,
EtOH, N02, S02, COS.BjS i
HCN, CO, NO, HC1, CH3CN, CO,,
N02, S02, CS2 i
HCN, CO, NO, H,S, HC1, CH.CN,
C02, N02. COS/SOj, CS2 3
HYDROCARBON STRUCTURES
IDENTIFIED11
Pyridine, Cs-C-j, aliphatic radicals, trace
oxygenates
Pyridine, Me-pyridine; aliphatic hydrocarbon
radicals through C-j
Six unidentified mass peaks (< mass 102;
trace oxygenates
Pyridine, Me-pyrldine, C7~Cg naphthenea,
aliphatic radicals through Cg: unidentified
mass peaks (< mass 109)
Pyrrole possible, trace oxygenates, trace
hydrocarbons through C_
C6H6' C10HB' C11C10' C14H10
C6V C10H8' C11H10' °UH10' C6H6°
As406
AB4°6
Aromatic hydrocarbons; nitro- and sulfur
hetero eye lies
Similar to 16-3 in composition; slightly
lower carbon number distribution for all
classes of compounds
.Trace organics through C.-
Pytidine, trace organica through C-.
Pyrrole, phenol, aromatics through C._
aliphatic radicals through CQ
8
Trace hydrocarbon
aFrom low ionizing voltage mass spectra data, as reported by ERDA/FERC

 All isomeric structures are possible

Measurement doubtful

 n.a.  - not available
                                                              34

-------
                               TABLE 6

       SEMI-QUANTITATIVE MASS SPECTRAL ANALYSIS OF PARTICIPATE
              MATTER COLLECTED AT THE ALUMINUM SMELTER*
Particle Size

Quantity Analyzed

Percent Vaporized
     l-3u

   100.9 mg

     4.5
  3-10y

100.0 mg

  3.7
Structural Types

     3-rlng aromatics
     Phenylnaphthalenes
     4-ring, peri-condensed
     4-ring, cata-condensed
     5-ring, peri-condensed
     Phenylanthracenes
     5-ring, cata-condensed
     6-ring, peri-condensed (mass 276)
     6-ring, peri-condensed (mass 302)
     7-ring, peri-condensed (Coronene)
     Dinaphthothiophene
     Azapyrene + Benzocarbazole
     Benzacridine
     Carbazole
     Acridine
     Dibenzocarbazole
     Dibenzacridine
     Azabenzo(ghi) perylene
     Azaperylene
Percent of Total lonization
2.3
2.3
6.5
7.0
20.4
3.5
12.6
17.1
3.5
1.1
1.5
12.2
1.3
1.0
1.2
1.4
0.8
0.03
4.3
0.8
5.4
7.5
7.0
12.4
12.0
13.3
3.7
2.5
0.9
1.6
12.5
5.7
1.9
3.2
0.4
3.1
2.1
4.0
n
 High resolution mass spectrometry; quantitation by integrated  peak
 height versus time
                                 35

-------
                               TABLE 7

    POSSIBLE DETECTION OF CARCINOGENIC POLYCYCLIC ORGANIC MATERIAL
          FROM HIGH RESOLUTION MASS SPECTROMETRIC ANALYSIS
                                                        CARCINOGENIC
MASS     FORMULA           POSSIBLE COMPOUNDS            POTENTIAL*

228      C.0H.„      Benzo(c)phenanthrene
          lo L2.

252      C2QH12      Benzo(b)fluoranthene
                     Benzo(j)fluoranthene
                     Benzo(a)pyrene

254      C9nHiA      Benz(j)aceanthrylene
             -1         (cholanthrene)

256      ConHn,      7,12-Dimethylbenz(a)anthracene
          2U lo

267      C20H13N     Dibenzo(c,g)carbazole                  -H-f

268      C21H16      3-Methylcholanthrene

278      C22H14      Dibenz(a,h)anthracene

279      CRN     Dibenz(a,j)acridine                    4-1-
                     Dlbenz(a,h)acridine                    -H-

302      C24H14      Dibenzo(a,h)pyrene
                       (Dibenzo(b,def)chrysene)
                     Dibenzo(a,i)pyrene
 + = carcinogenic; ++, -H-+, I I I I  = strongly carcinogenic  (Reference 11)
                                 36

-------
5.2  BIOASSAYS

Preliminary concentration-response  data  from  the  RAM cytotoxicity
bioassay are presented  in Figures 4-11.   In order to linearize the
concentration-response  data, cell response, expressed as viable cells
as a percent of control and reported  as  mean  values +1 standard
deviation, is plotted against  the common logarithm (base 10) of the
particle concentration  in the  culture medium.  The cell responses to
the three types of particle exposure  are provided (see Section
3.2.1):  particles plus supernatant fraction  (P+S), particle fraction
(P), and supernatant fraction  (S).  Linear dose-response regression
                                                              (12)
lines were fitted to the data, using  a least  squares solution.
Particulate samples collected  at the  basic oxygen furnace were found
to be nontoxic at the highest  concentration (i.e., lOOOyg particles/ml
medium), so no further  runs were performed.   Due  to sample mass limita-
tions, the particle plus supernatant  (P+S) fraction of the coke oven
heater 3-10y sample was tested only at the 1000 yg/ml concentration.
Additional concentration-response tests were conducted on selected
fractions of two particulate samples, to determine the cellular
response over a greater particulate concentration range than pre-
viously tested.  In Figure  12, the RAM viability is noted after
exposure to the supernatant fraction  (S) of the 3-lOy copper smelter
sample at concentrations ranging from 1 yg/ml to 40 yg/ml; the
preliminary concentration-response data (Figure 8) are also provided
for comparison.  In Figure  13, the RAM viability is noted after
exposure to the particles plus supernatant fraction (P+S) of the l-3y
sludge incinerator sample at concentrations ranging from 1 yg/ml to
200 yg/ml; preliminary concentration-response data from Figure 10
are included for comparison.
                                  37

-------
Results of the cytotoxicity bioassay on the filter material used to
collect the sub-micron particles at the industrial facilities are
provided in Table 8.  The mean value of two observations is presented.
Results of the mutagenic bioassay, as conducted by Litton-Bionetics,
Inc., are presented in Table 9.  The aluminum smelter l-3y particles
indicated mutagenic activity on two of the three bacterial strains
tested.  The copper smelter l-3y particles indicated possible
mutagenic activity on one of the three bacterial strains.  The
remaining particulate samples indicated no mutagenic activity on the
bacterial strains, under the test conditions.  Table 10 provides
dose-related response data for the two samples possessing mutagenic
activity.
Research Triangle Institute (RTI) conducted limited additional
mutagenic screening tests on aliquots of three of the eleven particu-
late samples tested by Litton-Bionetics, Inc. (i.e., copper smelter
3-10y sample, ceramics plant 3-lOy sample, and sludge incinerator
3-lOy sample).  Similar laboratory procedures were used, except that
RTI employed freshly prepared mouse liver microsomal fractions for
metabolic activation studies (rather than frozen preparations), and
conducted all screening tests (e.g., toxicity determinations, posi-
tive and sterility controls, and the mutagenic screening test) for a
given sample on the same day, rather than on separate days.  Results
of the RTI mutagenic screening tests are presented in Table 11;
the average of two observations per treatment level is presented.
                                 38

-------
  100
~ 80-
  60
  20
        STEEL PUNT, OPEN HEARTH FURNACE ( 3-10
            100
                                        300
                           CONCENTRATION (us PARTICLES /ML MEDIUM)
                          500
                                           1000
 100 r
  80
a
  20
      STEEL PLANT, OPEN HEARTH FURNACE ft-art
           100
            300           500
CONCENTRATION GIG PARTICLES /ML MEDIUM)
                                                                      1000
                                  FIGURE 4
         VIABILITY OF RABBIT ALVEOLAR MACROPHAGES EXPOSED
      TO VARIOUS FRACTIONS OF PARTICULATE SAMPLES COLLECTED
                      AT THE OPEN HEARTH FURNACE
                                     39

-------
     STEEL PLANT, COKE OVEN HEATER O-inH)
 1001-
 80 -
 60
 20
                                                               I	I   I
           100                           300           500
                           CONCENTRATION 
-------
 100
         STEEL PLANT, IRON, SINTERING (3-io.J
t
	I
i
  60
  10
  20
            100
                                         300            500
                            CONCENTRATION («G PARTICLES /ML MEDIUM)
                                 1000
 100 r
      STEEL PLANT, IRON SINTERING d-3iO
  80
  60
  40
  20
            100
                           CONCENTRATION
  300            500
PARTICLES /ML MEDIUM)
                                                                        1000
                                 FIGURE 6
        VIABILITY OF RABBIT ALVEOLAR MACROPHAGES EXPOSED
     TO VARIOUS FRACTIONS OF PARTICULATE SAMPLES COLLECTED
                     AT THE IRON SINTERING PLANT
                                      41

-------
 100 r
      CERAMICS PUNT (3-10n)
 80-
 : 60
|Uo
 20



/
KEY
©=s
0= p
©=P+S




, 1 	 1 	 1 	 1 	 1 	 1 	 1 	 1 	
' 100 300 500 1000
                          CONCENTRATION (ws PARTICLES /ML MEDIUM)
  100
             CERAHICS PLANT ( l-3n)
  80
  60
  '10
  20
           100
            300           500

CONCENTRATION (ce PARTICLES /ML MEDIUM)
1000
                                  FIGURE?
          VIABILITY OF RABBIT ALVEOLAR MACROPHAGES EXPOSED
       TO VARIOUS FRACTIONS OF PARTICULATE SAMPLES COLLECTED
                         AT THE CERAMICS PLANT
                                     42

-------
  100 r
   80
   60
   20 -
       COPPER SMELTER 6-10„)
       /    100
             300           500
 CONCENTRATION U PARTICLES /ML MEDIUM)
                 1000
  100 r
  80
  60
        COPPER SMELTER 0-3(0
§ 10
  20
            100
             300
CONCENTRATION (p s PARTICLES /ML MEDIUM)
500
                                                                      1000
                                  FIGURES
         VIABILITY OF RABBIT ALVEOLAR MACROPHAGES EXPOSED
      TO VARIOUS FRACTIONS OF PARTICULATE SAMPLES COLLECTED
                         AT THE COPPER SMELTER
                                     43

-------
      ALUMINUM SMELTER
 100
  60
£
	I
i
  20
           100
	300          500

 CONCENTRATION (CG PARTICLES  /ML MEDIUM)
                                                                  1000
 100
  80
  60
  20
           100
            300           500
CONCENTRATION ( 
-------
100
 80
 50
20
     SLUDGE INCINERATOR (3-10^)
          100
                                      300           500
                          CONCENTRATION («G PARTICLES /ML MEDIUM)
                                          1000
100
 80
 60
20
     SLUDGE INCINERATOR
          -§-
                                                        1.   1   1   1
          100
            300
CONCENTRATION 0*s PARTICLES /ML MEDIUM)
                                                                    1000
                                 FIGURE 10
         VIABILITY OF RABBIT ALVEOLAR MACROPHAGES EXPOSED
      TO VARIOUS FRACTIONS OF PARTICULATE SAMPLES COLLECTED
                      AT THE SLUDGE INCINERATOR
                                    45

-------
     OIL FIRED POWER PLANT
 100 -



 80

5
S
£60

—I

§1

i
<_>

 20
          100
           300          500
CONCENTRATION O/G PARTICLES /ML MEDIUM)
                                                              1000
                              FIGURE 11
        VIABILITY OF RABBIT ALVEOLAR MACROPHAGES EXPOSED
      TO VARIOUS FRACTIONS OF PARTICULATE SAMPLES COLLECTED
                    AT AN OIL FIRED POWER PLANT
                                 46

-------
 100
 30
£60
5 if)
 20
COPPER SHELTER (3-10.)

           SUPERNATANT FRACTION
                              EXPANDED TESTS
                                                       PRELIMINARY TESTS'
                          S     10        30   SG '
                         CONCENTRATION («B PARTICLES/ML MEDIUM)
                              100
                                       300  500   1000
                               FIGURE 12
     VIABILITY OF RABBIT ALVEOLAR MACROPHAGES EXPOSED TO
           SUPERNATANT FRACTION COLLECTED FROM THE
                    COPPER SMELTER (3-10ju SAMPLE)

-------
                                 PARTICLES PLUS SUPERNATANT FRACTION
 100
 80
£60
                                              •—. Y      ^.PRELIMINARY TESTS

                                                    --..
 20
©PRELIMINARY TESTS

A EXPANDED TESTS
                        5     10        30   50    100

                        CONCENTRATION UG PARTICLES/ML MEDIUM)
                                           300   500
                                                     1000
                              FIGURE 13
      VIABILITY OF RABBIT ALVEOLAR MACROPHAGES EXPOSED TO
      PARTICLES PLUS SUPERNATANT FRACTION COLLECTED FROM
                THE SLUDGE INCINERATOR (1-3M SAMPLE)

-------
                               TABLE 8

         VIABILITY OF RABBIT ALVEOLAR MACROPHAGES EXPOSES  TO
                     SUB-MICRON PARTICLE FILTERS
FILTER SAMPLE
Oil Fired Power Plant
Copper Smelter*3
Aluminum Smelter
Aluminum Smelter
Iron Sinteringb
Copper Smelt erb
Iron Sinteringb
Paper Mill
Municipal Incinerator
Ceramics Plant
Open Hearth Furnace
Coke Oven
Basic Oxygen Furnace
Teflon Control Filter
FILTER PLUS
SUPERNATANT
20. 4a
26.3
28.2
31.5
33.0
37.2
56.1
56.1
92.1
93.0
95.9
97.6
98.3
NON- TOXIC
SUPERNATANT
10. 5a
29.5
24.9
32.1
55.2
30.6
58.8
76.7
71.8
53.6
98.8
98.8
98.4
NON-TOXIC
DRIED
FILTER
30. 13
41.0
86.3
60.8
95.7
35.6
99.5
97.9
93.8
98.4
97.4
99-6
99,6
NON-TOXIC
Percent viable cells (mean of two observations)

Original filters replaced because of heavy grain loading.
                                 49

-------
                                TABLE  9

          MUTAGENIC  ACTIVITY  OF  INDUSTRIAL  PARTICULATE  SAMPLES3


                                    Response With  Indicator  Strain

          Sample                  TA-1535        TA-1537         TA-1538

 Open  Hearth  Furnace,  l-3y           _               _               _

 Basic Oxygen Furnace,  l-3y          -               -

 Iron  Sintering Plant,  3-10y         -

 Ceramics  Plant,  l-3y                -

 Ceramics  Plant,  3-10y               -

 Copper Smelter,  l-3y                -               -              +/-

 Copper Smelter,  3-10y               -               -

 Aluminum  Smelter, l-3y              -               +              +/-

 Aluminum  Smelter, 3-10y                                            -

 Sludge Incinerator, l-3y                                           -

 Sludge Incinerator, 3-10y           -
a    =  As determined by Litton-Bionetics, Inc.
-    =  No response obtained
+    =  Positive mutagenic response showing a dose-related increase
+/-  =  Questionable response or non-dose related positive results
                                  50

-------
                                       TABLE 10
         DOSE RELATED RESPONSE FOR MUTAGENICALLY ACTIVE PARTICULATE SAMPLES3

PARTICLE CONCENTRATION
(yg/plate)
i^^— — 	 1 	 1— IT- 	 1- 1 -111 .-1111
Control
Aluminum Smelter (l-3y)
0.01
0.1
0.5
1.0
10.0
100.0
500.0
1000.0
Copper Smelter (l-3y)
0.01
0.05
0.1
0.5
1.0
10.0
100.0
NUMBER OF REVERTANTS PER PLATE
TA-1535
i
NAb Ac
•^^•••^^^•••••^^^•^^•••^^^^^^^^•^^^^^^^^^•^^^^^^^q^^^^
14.5 + 1.3d

13.0 + 2.8
13.0 + 2.9
14.3 + 1.7
13.8 + 1.9
10.5 + 2.7
-
-
-

15.0 + 2.2
14.5 + 2.9
15.0 + 4.1
13.8 + 2.1
13.8 + 1.0
-

14.0 + 0.8

11.8 ± 1.3
13.3 + 1.0
-
14.5 + 2.4
15.3 + 2.1
13.8 + 3.2
-
-

13.0 + 2.9
-
16.0 + 1.2
14.0 + 1.4
12.0 + 1.4
15.8 + 1.7

TA-1537
NA A
17.0 + 2.5

_
14.5 + 2.7
-
14.8 + 3.8
15.8 + 1.9
16.5 + 3.9
13.5 + 3.9
-

14.8 + 2.1
-
14.3 + 2.6
17.3 + 1.1
13.8 + 1.7
16.5 + 1.3

19.0 + 2.5

„
-
-
17.5 + 1.7
15.3 + 1.0
21.0 + 2.2
29.3 + 3.3
38.0 + 4.7

14.5 + 1.3
-
17.8 + 2.9
-
14.0 + 2.2
19.8 + 1.0
17.5 + 2.7
TA-1538
NA A
15.5 + 2.4

16.3 + 4.2
14.0 + 4.2
14.5 + 5.5
14.3 + 4.9
13.5 + 2.9
-
_
-

16.5 4- 1.7
16.5 + 2.7
14.8 + 1.9
14.3 + 4.6
16.3 + 3.4
-

39.3 + 1.7

_
40.8 + 6.1
39.8 + 6.3
46.5 + 7.3
48.8 + 8.2
63.8 + 11.5
_
-

39.5 + 1.7
_
43.8 + 6.1
44.0 + 2.6
49.5 + 5.2
66.0 + 5.0
"
SAs reported by Litton-Bionetics, Inc.
 Non-activated
 Activated
 iiean + ISO of four observations

-------
                                       TABLE 11
        MUTAGENIC ACTIVITY OF THREE SELECTED INDUSTRIAL PARTICULATE SAMPLES8
PARTICLE CONCENTRATION
(|j.g/plate)
Copper Smelter (3-10p)
Control
1
10
100
500
1,000
Ceramics Plant (3-10p)
Control
1
10
100
500
1,000
Sludge Incinerator (3-10y)
Control
1
10
100
500
1,000
NUMBER OF REVERTANTS PER PLATE
TA-1535
NAb AC

507d
385
362
466
259
11

322
282
283
202
263
233
490
382
438
452
346
305

174
138
165
120
100
59

47
58
66
54
54
78
144
156
132
142
132
132
TA-1537
NA A

10
12
13
11
12
10

6
8
8
11
10
4
4
8
8
3
11
7

12
22
16
18
8
3

8
9
8
12
9
6
10
11
14
12
10
8
TA-1538
NA A

31
20
22
16
13
0

3
10
11
14
9
8
6
9
8
9
7
10

45
42
45
36
16
8

18
16
14
12
14
12
12
16
12
10
14
17
, As reported by Research  Triangle Institute
 Non-activated
^Activated
 Average of two observations

-------
                          6.0  DISCUSSION

This research effort was undertaken  to provide field test validation
of selected sampling and testing procedures intended to characterize
potential hazards associated with airborne particulate emissions.
There are numerous alternate or additional tests that could be used
to accomplish this goal; however, only those tests utilized in this
research effort are discussed.  Any  conclusions or recommendations
concerning these selected tests are  based solely on the test proto-
col as performed.  In some instances, preliminary findings indicated
the need to modify the scope of some tests.  When time or fiscal con-
straints permitted, these modifications were undertaken.  Otherwise,
the test protocols were completed as initially defined and recommen-
dations for additional studies are offered for consideration.

In the subsequent sections, each of  the sampling and testing proce-
dures is discussed separately, followed by a section integrating the
entire research effort.

6.1  SAMPLE COLLECTION

6.1.1  Sampling Train Performance
The series cyclone sampling train was developed to collect, within
one-to-five hours of operation, sufficient quantities of size-
classified particulate material from a variety of industrial sources.
At the initiation of this project, it was decided that a sample mass
equal to 300 mg was sufficient for all subsequent chemical and/or
biological testing.  In most cases this collection criterion was
satisfied, with the exceptions being the open hearth furnace, the
basic oxygen furnace, and the coke oven heater (see Table 2).  Since
                                 53

-------
 the oil-fired power plant shut down after two hours of sampling train
 operation, collection of the minimum sample mass was not realized.
 The stack emissions at the paper mill were substantially reduced on
 the day of sampling, compared to stack emissions during the site
 inspection one day earlier, and a negligible amount of particulate
 material was collected.  The cause of this drastic emissions reduc-
 tion on the day of sampling was undetermined.

 The cyclone train also exhibited its versatility by operating
 successfully under a variety of circumstances.  In that instance
 where cyclone train operation at elevated temperatures was a problem
 (i.e., sludge incinerator), a water-cooled probe was used.  Success-
 ful train operation under heavy grain load conditions was also
 accomplished, although the filters required replacement and the
 total elapsed sampling time was short.  The actual sampling locations
 at each industrial process provided diverse conditions under which
 the cyclone train had to operate.

 6.1.2  Demonstrated Need for Size-Classified Particles
 The need for size-classified particulate material has been established,
 since the bioassays and chemical analysis have indicated that parti-
 culate fractions from the same industrial source do not necessarily
 possess similar characteristics.  The elemental composition (deter-
 mined by SSMS) of particulate samples collected at the same site
 often differed (e.g., Na in the aluminum smelter samples, Be and Al
 in the copper smelter samples, and Ag and U in the basic oxygen
 furnace samples).   The GC-MS analysis of the l-3u samples of both
 the iron sintering plant and the ceramics plant differed substantially
 from the analyses of the respective 3-10y samples.  The RAM bioassay
 indicated different particulate toxicities for the aluminum smelter
samples and the sludge incinerator samples (i.e., l-3y particles
                                 54

-------
compared to the 3-10y particles).  The mutagenic screening tests
conducted by Litton-Bionetics indicated possible mutagenic activity
for the aluminum smelter l-3y sample and the copper smelter l-3y
sample, while the respective 3-10y particles exhibited no mutagenic
activity.

6.2  CHEMICAL ANALYSIS
The size-classified particulate material collected at the nine
industrial sites was subjected to three types of chemical analysis:
SSMS, GC-MS, and HEMS.  Aliquots of each particulate sample were
analyzed according to the procedures described in Section 4.1.  No
repetitive determinations were made for any analysis, except for
two particulate samples analyzed by SSMS in independent laboratories.
The chemical analyses were conducted on intact particulate samples.

6.2.1  Spark Source Mass Spectrometry
Several SSMS scans were conducted on the particulate samples in
order to identify trace species  as well as major constituents (see
Table 3).  In addition, two samples (i.e., coke oven heater l-3y
and copper smelter l-3y) were analyzed for major and minor species
by an independent laboratory  (EMSL/RTP).  Since no repetitive
analyses were performed by each  laboratory, the precision of the
data presented in Table 3 has not been determined.  The SSMS tech-
nique is estimated to be accurate to within 200 to 500 percent.

The SSMS analysis identifies the elemental composition of a given
sample, and does not determine the chemical matrix of the sample,
nor does it reflect the availability of a constituent to a biological
system.  It is a useful survey tool, however, in that it can detect
the presence of over seventy elements in a particulate sample.
                                 55

-------
6.2.2  Gas Chromatography-Mass Spectrometry
The  GC-MS analysis  focused on polycyclic hydrocarbons of known
carcinogenic potential or structural analogues.  Results were
reported as nanograms of a certain species detected in the entire
sample, and converted to parts per million by weight (ng species/mg
sample).  Since an  accurate determination of sample mass collected
on the filter material could not be made, filter analyses are
reported as total nanograms present.  All major hydrocarbon peaks
were identified in  the mass spectra, but there were additional peaks
present (but not identified) that did not correspond to known
polycyclic hydrocarbon carcinogens.  Size-classified particles
collected at the same site did not always have comparable analyses
(e.g., iron sintering plant, ceramics plant).  The aluminum smelter
particulate samples contained by far the largest amounts of poly-
cyclic hydrocarbons detected in any of the samples analyzed.

6.2.3  High Resolution Mass Spectrometry
The  constituents identified by HRMS in the fourteen samples analyzed
are  presented in Table 5.  Only that portion of the particulate
sample that vaporized under the test conditions would be detected
in the mass spectra. Interpretation of the mass spectra indicates the
mass number of the compounds present but does not specify the isomer.
The  resulting hydrocarbon masses can then be screened for those
precise masses that would indicate the possible pjresence of known
hazardous and/or toxic compounds.  This method determines only the
presence of a specified mass, and cannot accurately quantify it.
The presence of a precise mass indicates only that the suspect
structure is possibly present, since the isomeric form having that
mass is undetermined.
                                 56

-------
Analysis of the copper smelter samples indicated an intense spectral
peak identified as As.O  , the dimer of As 0  .  The strong As,0fi peak
prevented further computer analysis of spectra from both samples
(l-3y and 3-10y).  Since studies of pure 4-  to 6-ring aromatic
hydrocarbon compounds, representative of those detected on the
aluminum smelter samples, have shown that their rate of vaporiza-
tion varies with ring number and type of condensation, the aluminum
smelter samples were subjected to an additional semi-quantitative
HRMS analysis.  Mass spectra from all fourteen particulate samples
were screened for those  precise masses associated with nine known
carcinogenic aromatic hydrocarbons  (Table 7).  Spectra from the two
aluminum smelter samples indicated  the presence of all nine precise
masses; none of the  remaining samples contained any of those nine
precise masses.

6.2.4   Comparison  of Analytical  Results
The elemental  composition  of  two particulate samples  (i.e., coke
oven heater l-3y,  copper smelter l-3p) analyzed by different labora-
tories  show reasonable  agreement,  considering the accuracy of the
technique.   The arsenic  concentration in both copper smelter samples
is  substantially higher  than other  particulate samples analyzed by
SSMS.   The  intense As 0.. spectral peaks obtained by HRMS analysis of
the  same two samples substantiate the SSMS results.  Results from the
HRMS of the particulate  samples  collected at  the iron sintering plant
appear consistent although the percent of each sample vaporized
differed substantially  (see Table 5).  The GC-MS analysis of the same
particulate samples  (i.e., iron  sintering plant, 1-3 and 3-10y)
detected pocyclyclic organic species in the  l-3y sample only.

Data presented in Tables 4 and 6 permit comparisons between coronene
and dibenzocarbazole concentrations in the l-3y and 3-10y aluminum
smelter samples as determined by GC-MS or HRMS analysis.  In each
                                 57

-------
 case,  the HRMS analysis consistently indicated higher constituent
 concentrations than determined by GC-MS analysis.  With HRMS, coronene
 was  detected at  concentrations five times higher and six times higher
 in the l-3y and  3-10y samples, respectively, than levels determined
 by GC-MS.  Dibenzocarbazole was detected by HRMS at seven and two
 times  the level  determined by GC-MS for the l-3y or the 3-10y sample,
 respectively.

 6.3  BIOLOGICAL  CHARACTERIZATION
 The  biological activities of the size-classified particulate samples
 have been categorized using two in vitro bioassays, one to determine
 the  acute cytotoxic nature of the samples, and the other to determine
 whether any of the samples are mutagenically active.  Both bioassays
 have indicated their utility in studying the effects of certain pure
 compounds, but neither has been used extensively to study complex
 mixtures.  The following sections discuss the two bioassays, as
 performed, and note procedural problems that were encountered during
 the  testing of the complex particulate samples and evaluation of the
 results.  This research study examined the ability of each bioassay
 to assess specific biological characteristics of complex particulate
 samples.

 6.3.1   RAM Cytotoxicity Bioassay
 The  effects on rabbit alveolar macrophase (RAM) viability from
 exposure to particulate material (including soluble and/or insoluble
 fractions) are provided in Figures 4 through 13.  In those figures,
 RAM  cell viability, expressed as percent of controls, is linearly
 regressed with the common logarithm of the particle concentration
 in the  culture medium.  In an attempt to improve the linearity of
 the experimental data, additional transformations were applied  (e.g.,
probit, logit), but no improvement was evident.
                                 58

-------
Results from the basic oxygen  furnace  (l-3y and 3-10y) are not
reported, since they appeared  non-toxic at 1000vig/ml and were not sub-
jected to further testing.  The methods used to obtain the particles
plus supernatant  (P+S) fraction,  the particle  (P) fraction, and the
supernatant (S) fraction are explained in Section 4.2.1.  The number
of observations per exposure level varied from six observations at
1000 g/ml to three observations at 300yg/ml or lOOyg/ml.*  In
Figures 12 and 13, three observations per exposure level in each
expanded concentration-response trial were reported.

The cytotoxic effects of particulate material collected on the filters
are provided in Table 8.  During  the collection of the size-classified
particulate material, heavy grain loading conditions existed at three
sites and required the replacement of the filters.  Original filters
and their replacements were subjected to the RAM bioassay procedure.
The cytotoxic nature of the sub-micron particles cannot be directly
compared to these of the l-3y  or  3-10y samples collected at the same
industrial site, because the sub-micron particulate concentration to
which the RAM cells were exposed  could not be quantified (due to
thermal degradation of the filter material).

Figures  4-11  indicate  that  there  is  a  general  tendency  for the  l-3y
particles  to  be more  toxic  than the  respective 3-10y particles
collected  from the  same  site  (e.g.,  sludge  incinerator, Figure  10).

The toxic  nature  of  the  1-3^ and  3-10H- particulate  samples appears  to
be associated with  the particles  themselves,  since  the  particles  plus
*In a few instances, only two  observations were reported at lOOyg/ml.
                                  59

-------
supernatant (P+S) or the particle (P) fractions are consistently more
toxic than the respective supernatant (S) fraction for particles from
a specific industrial site (notable exceptions are the copper smelter
samples and the oil-fired power plant sample, Figures 8 and 11).


Figures 12 and 13 provide concentration-response data for two parti-
culate samples tested over a greater exposure range than the remainder
of the particulate samples.  Figure 12 (copper smelter 3-10y sample,
supernatant fraction) indicates that the log concentration-response
relationship is not linear over the entire particulate exposure
range, while the relationship expressed in Figure 13 (sludge incin-
erator l-3y sample, particles plus supernatant fraction) appears
consistent for both preliminary and expanded tests.  It would appear
that data obtained over a wider than 10-fold exposure range (e.g.,
lOOyg/ml to lOOOyg/ml) is required to adequately determine the con-
centration-response relationships for certain of the particulate
samples in the RAM bioassay.

Existing concentration-response data were insufficient to permit a
proportional ranking of the particulate samples based on their
observed cytotoxic nature.  Since the cytotoxic nature of the samples
differed widely, any comparison of the particulate samples at .a
specific level of response (e.g., LD^s*) required extrapolation of
the concentration-response lines for some of the samples.  Figures 12
and 13 have indicated that extrapolation is ill-advised.  If  all
concentration-response relationships are to be compared, both the
slopes and intercepts of the regressions must be considered.  Over
a limited exposure concentration/response range, the regression
*that concentration lethal to fifty percent of an exposed population
 within a specified length of time.
                                 60

-------
estimates of the slopes and intercepts may not be representative
of the relationship over a broader concentration range  (see Figure 12),
When regression equations describing several particulate samples have
unequal slopes (i.e., dose-response lines not parallel), an interpre-
tation of comparative toxicities of the samples depends on the level
of response considered.

Particulate samples tested by the RAM bioassay can be ranked, on an
ordinal scale, according to observed cytotoxicity, from most to least
toxic.  Separate rank orders, based on cell viability (mean + 1 SD)
can be established to include all samples tested at each particular
test concentration (i.e., lOOOyg/ml, 300yg/ml, 100vig/ml) and for
each test fraction (i.e., P+S, P, or S).  Non-parametric statistical
                                                     (13)
procedures  (Kendall's coefficient of concordance test    ) indicate
that the toxicity rankings established at each test fraction or at
each test concentration are not significantly different (a = 0.01).
furthermore, using Spearman's rank correlation test or Kendall's
   (13)
tau,    all pairs of ranks within test fractions at different particu-
late concentrations, or between test fractions at the same particle
concentration were found to be strongly associated (a = 0.01).  In
effect, these statistical tests indicate that the ordinal ranking of
industrial sites based on observed RAM cytotoxicity is independent
of test concentration, over the range tested, and independent of
testing procedure (i.e., test fraction).  All rank orders tended to
agree as to which industrial particulate samples were more toxic
and which samples were less toxic.  In Table 12, the particulate
samples have been organized into three toxic categories (i.e.,
relatively high cytotoxicity, intermediate cytotoxicity, and rela-
tively low cytotoxicity) that reflect the overall performance of each
particulate sample in the ordinal rankings for each particulate con-
centration tested or each test fraction.  The order in which each
                                  61

-------
                             TABLE 12
 RELATIVE CYTOTOXIC NATURE OF THE INDUSTRIAL PARTICULATE SAMPLES3
RELATIVELY HIGH CYTOTOXICITY
          Oil-Fired Power Plant, l-3y
          Copper Smelter, l-3y
          Copper Smelter, 3-10y
          Aluminum Smelter,  l-3y
          Sludge Incinerator, l-3y

INTERMEDIATE CYTOTOXICITY
          Aluminum Smelter,  3-10y
          Iron Sintering Plant, l-3y
          Open Hearth Furnace, l-3y
          Open Hearth Furnace, 3-10y
          Coke Oven Heater,  l-3y

RELATIVELY LOW CYTOTOXICITY
          Iron Sintering Plant, 3-10y
          Ceramics Plant, l-3y
          Ceramics Plant, 3-10y
          Sludge Incinerator, 3-10y
          Basic Oxygen Furnace, l-3y
          Basic Oxygen Furnace, 3-10y
<3
 Based on ordinal ranking of particulate samples at all test concentra-
 tions for all test fractions
 Order within categories are arbitrary
 Coke Oven Heater,  3-10y—insufficient testing to permit ranking
 Oil-Fired Power Plant, 3-10y, and paper mill samples—no bioassays
 conducted
                                62

-------
particulate sample appears within  each  of  the  three categories is
arbitrary and should not be construed to represent differences in
relative cytotoxicities.

The RAM bioassay, as performed  in  this  research effort, encountered
some procedural problems that required  attention.  Particles tended
to agglomerate when added to the test system.  Not only could the
effective particle size to which the RAM cells were exposed be
changed, but the degree to which various chemical constituents would
solubilize in the bioassay medium   could be  affected.  The agglom-
eration also led to difficulty  in  counting cells at the termination
of a trial (via hemocytometer).  The exclusion of trypan blue dye
was the criterion by which cell viabilities were estimated.  Addi-
tional response indicators, including measurements related to membrane
integrity and functional impairment, have  been developed to add sensi-
tivity to the RAM bioassay and are  available for future, more-intensive
studies.

6.3.2  Mutagenesis Bioassay
The mutagenic activities of eleven  particulate samples on three
JL' typhimurium bacterial strains under  the test conditions specified
in Section 4.2.2 are presented  in Tables 9 and 10.  The screening
tests, as conducted by Litton-Bionetics, indicated that one sample,
the aluminum smelter l-3y material, caused a weak dose-dependent
mutagenic response in two of the three  bacterial strains (TA-1537,
TA-1538), while a second sample, the copper smelter l-3y material,
indicated possible mutagenic activity in one strain (TA-1538).  Each
mutagenic response occurred with the mouse liver activation system.
All nine remaining particulate  samples  possessed no detectable
mutagenic activity under the test conditions.
                                 63

-------
The parallel mutagenic screening tests that RTI performed* on three
particulate samples  (i.e., copper smelter 3-lOja, ceramics plant 3-10|j.,
and sludge incinerator 3-10H- samples) are presented in Table 11,
where mutagenic activity is described in terms of the number of rever-
tants  (mutants) per plate.  Since only two observations per treatment
level were performed, the average number of revertant colonies are
reported.  No consistent increase in the number of revertants on
treated plates relative to controls is evident.  Slight increases in
the number of revertants per plate resulting from exposure to the
ceramics plant 3-lOp- sample are inconclusive.

The exposure levels that each laboratory utilized in the mutagenic
screening tests were different, often by a factor of 100.  Results
of those samples tested by both RTI and Litton-Bionetics are com-
parable.  Two samples (i.e., aluminum smelter l-3y and copper smelter
l-3y samples) were identified by Litton-Bionetics as possessing
weakly mutagenic activity under the laboratory test conditions.

The mutagenic activity of a compound is usually expressed as the
ratio of the number of revertants resulting from treatment relative
to the number of spontaneous revertants in suitable controls.  Muta-
genic activity is indicated as the ratio becomes significantly
greater than 1.0.  If exposure doses in mutagenic screening  tests
are sufficient to produce noticeable bacterial toxicity, the number
of revertants per plate can be normalized to reflect the anticipated
decrease in the number of viable cells exposed to the test compound.
If this normalization procedure is attempted, the preliminary
toxicity screening test must be sufficient to determine a representa-
tive dose-response toxicity relationship.
*RTI employed a modification of the Litton-Bionetics protocol
 (see Section 5.2).

                                 64

-------
The mutagenesis bioassay  procedure utilized in this  study should  be
viewed as an initial  attempt  to  screen complex particulate material
for mutagenic activity under  controlled laboratory conditions.  Because
of its broad solvent  characteristics,  DMSO was chosen  as  the  solvent
vehicle used to introduce all particulate samples  into the test system.
The solubility of a particulate  sample in DMSO varied  depending on the
laboratory mixing technique employed  (e.g., sonication versus manual).
The irregular dose-response relationships between  particulate exposure
and survival of bacterial populations  reflected the  difficulties  faced
when testing a complex sample.   The testing protocol employed by
Litton-Bionetics and  RTI  were similar  except for those points
mentioned in Section  5.2.  However, the additional effort  expended in
following the RTI modifications  was substantial.   The  benefits of each
test procedure should be  evaluated, and if  comparable,  the less expen-
sive one  (Litton-Bionetics) should  be  suggested for  routine, prelimi-
nary screening tests.

6.4  COMPARISON OF CHEMICAL ANALYSIS TO OBSERVED BIOLOGICAL ACTIVITY
An objective of this  research was  to determine whether  the observed
biological activity of the particulate  samples could be correlated
to the identified chemical composition  of each sample.  Since the
chemical analyses are intended for  routine  screening of large numbers
of samples, it is impractical to perform a  comprehensive chemical
analysis on a given particulate  sample,  until the  need  to do so has
been identified.  The chemical analyses utilized in  this research were
selected to offer a screening tool  with the capacity to provide a
substantial amount of chemical information  about the intact particu-
late sample for a relatively  small  investment.  The  SSMS technique
can rapidly assay for over seventy  elements, although  the  chemical
matrix of the sample  is not provided.   HRMS can suggest the presence
of numerous hazardous and/or  toxic  compounds  in the  particulate sample,
but without additional effort, those compounds cannot  be quantified.
                                  65

-------
The GC-MS analysis has focused on identifying polycyclic hydrocarbons
of known carcinogenic potential, and by inference, those with tnutagenic
potential.

Initial comparisons of the chemical data in terms of observed cyto-
toxicity involved the determination of the relationship of the
individual elemental concentrations in the particulate samples to
the observed cytotoxicity of that sample.  A best-fit regression
line was determined individually for each of the elements by plotting
the logarithm of the element's concentration in the total sample
versus the observed cytotoxicity (as percent viability) for that
sample for each test fraction (i.e., P+S, P, and S) at lOOOyg/ml
exposure.  This statistical approach assumes that no interaction
occurs between the various element's effects on observed cyto-
toxicity.  No strong correlation was found between individual elemental
concentrations and observed cytotoxicity in a given sample.   Since the
SSMS data do not provide information as to the biological availability
or chemical form of those elements present, strong correlation should
not be expected.   Two particulate samples could be identical in ele-
mental composition, but due to different chemical compounding, possess
vastly different biological activities.

Additional statistical evaluations considered the correlation of
observed cytotoxicity with a calculated toxicity index that reflected
the relative hazard, on a scale from one to ten, of the chemical con-
stituents (elemental and organic) detected in the particulate samples.
The scale presented in Table 13 is based on empirical determination of
                                            (14)
acute lethality in small laboratory animals.      The following equa-
tion represents the chemical toxicity index used:
                                  66

-------
                                                                                      TABLE 13
                                                   RELATIVE RANKING OF IDENTIFIED CHEMICAL CONSTITUENTS BASED ON ACUTE TOXICITY
O
VI
Species
Aluminum
Antimony*
Arsenic*
Barium
Beryllium*
Bismuth*
Boron
Bromine
Cadmium*
Calcium
Cerium
Cesium
Chlorine
Chromium*
Cobalt*
Copper*
Dysprosium
Erbium
Europium
Fluorine
Gadolinium
Gallium
Germanium
Gold
Hafnium
Holmium
Iodine
Iridium
Iron
Lanthanum
Toxicity
Rating3
3
7
7
4
10
6
3
3
7
3
4
1
6
3
5
3
3
3
3
7
3
3
2
1
3
3
1
3
3
7
TLVb

0.5
0.5

0.002



0.1




0.5
0.1
0.1














Species
Lead*
Lithium
Lutecium
Magnesium
Manganese*
Molybdenum*
Neodymium
Nickel*
Niobium
Osmium
Palladium
Phosphorus
Platinum
Potassium
Praseodymium
Rhodium
Rubidium
Ruthenium
Samarium
Scandium
Selenium*
Silicon
Silver*
Sodium
Strontium
Sulfur
Tantalum
Tellurium*
Terbium
Thallium*
Toxicity
Rating3
7
5
3
3
6
6
3
3
2
10
3
7
5
3
3
3
2
3
3
4
5
1
3
3
3
7
1
3
3
8
TLVb
0.15



5
5

1.0












0.2

0.01




0.1

0.1
^^^^^^^^^^v«^«ftBHB«MBB^^H^^^^^^^^^^_^^^^^^^M^HHV^
Thorium
Thullium
Tin*
Titanium*
Tungsten
Uranium
Vanadium*
Ytterbium
Yttrium
Zinc*
Zirconium


Anthracene /Phenanthrene
Methyl Anthracenes
Fluoranthene
Pyrene
Methyl Pyrene/Fluoranthene
Chrysene/Benz (a) Anthracene
Methyl Chrysenes
Benzo Fluoranthenes
Benzo(a)Pyrene
Benzo (e)Pyrene
3-Methylcholanthrene
Indeno (1,2, 3-cd)Pyrene
Benzo (ghi)Perylene
Dibenz (a ,h) Anthracene
Dibenzo (c , g) Carbazole
Dibenzo(a,i and a,h)Pyrenes
Coronene
Toxicity
Rating3
5
3
4
6
4
3
8
3
5
5
2


5
5
1
1
1
7
4
5
4
2
5
5
7
3
7
7
5
TLVb


2.0
10


0.05


5




















                       aToxicity rating based on most probable chemical form, most comparable exposure route, most representative

                        test species;  obtained  from ref. 14

                       bThreshold limit values  (expressed as mg/m3)  from ref. 15

                       *20 most influential elements

-------
                      TItr = Z TRi x [C]±                  eq. (1)
where:
                      XI    = Toxicity Index (calculated)
                      TR.    = Toxicity Rating for the element
                              or organic species
                      [c~].  = Concentration of individual
                              element or species identified in
                              a particulate sample
  Equation (1) represents an extremely simplified approach to a very
  complex problem.  In this equation, no consideration is given to the
  possible biological availability of the constituents, nor is the
  solubility of any component factored in.  Since specific inorganic
  species were not identified in the chemical analysis, the toxicity
  rating required flexibility in ranking the toxic components.  In
  constructing the toxicity rating, lethality data for the most prob-
  able form of an element was utilized.  The toxicity rating itself
  is a compromise since it is extremely difficult to compare toxic
  dosages of a given substance across test species, routes of exposure,
  or times of exposure.  It is acknowledged that a one-to-ten scale
  does not completely reflect the relative differences in toxicities
  between chemical compounds.

  The common logarithm of the toxicity index (eq. 1) for each particu-
  late sample was plotted versus its observed cytotoxicity, expressed
  as percent viability, tested at lOOOyg/ml.  A best-fit regression
  line was fitted to the data from all particulate samples tested for
  each bioassay test fraction (i.e., P+S, P, and S).  The resulting
  best-fit regression line possessed a correlation coefficient  (r) of
  -0.09 for the P+S fraction, an r=-0.23 for the P fraction, and an
  r=-0.23  for the S fraction.  Negative correlation coefficients
                                   68

-------
indicate the expected relationship between macrophage viability and
constituent concentration.  A correlation is stronger between the
variables as the correlation coefficient approaches +1.0.

In examining the data,  it was noted that approximately twenty elements
in Table 13 consistently provided the  largest proportionate impact on
the toxicity index  (TI) summation for  the particulate samples.  A
second iteration of equation (1) was performed, with the TI values
determined solely from  those twenty elements.  The twenty elements are
designated in Table 13.  The regression equations possessed the follow-
ing correlation coefficients:   P+S fraction, r= -0.63; P fraction,
r= 0,74; and S fraction, r= -0.62.

A second toxicity index was devised to assess the ranking system
employed in equation  (1), since the toxicity rating of one-to-ten
might not be sensitive  enough to reflect large differences in
,toxicity.  Using the  threshold  limit values  (TLVs)^ 5' of the
twenty influential  elements, the following was performed:
where
                       TItlv
                       TI     =  toxicity index (calculated)
                       [E] .    =  concentration of  individual  element
                          1      in a particulate  sample
                       TLV     =  threshold limit value* for the
                          G
                                element
The logarithm of  the  toxicity index (eq.  2)  for each particulate
sample was plotted  versus  the observed cytotoxicity (as percent
viability) for that sample.   The resulting best-fit regression
*TLVs represent the lowest  suggested value  for  the  element  (or
 most likely compound  (ref.  15)).
                                 69

-------
 line possessed a  correlation coefficient  (r) of -0.62 for the P+S
 fraction,  an r=-0.62 for the P fraction  and an r=-0.65 for the
 S  fraction, comparable to results using equation 1.  One should
 be cognizant of the  fact that suggested TLVs present the relative
 hazard associated with each element; however, the criteria to
 assess hazards include (among others) inherent toxicity, carcino-
 genic and  mutagenic  potential, odor thresholds, and tendency to
 accumulate in the body.

 Several  implicit  assumptions are made when interpreting the experi-
 mental data according to equation (1) or  (2).  The toxicity index
 determinations assume a linear dose-response relationship to
 quantify suggested TR or TLV with observed constituent concentra-
 tions.   The method assumes strictly additive effects or no inter-
 action between various constituents.  In  addition, this treatment
 assumes  that all  components of toxicological importance have been
 identified and properly quantified.

 Results  of SSMS and  GC-MS analyses were incorporated in this evalua-
 tion.  The results from the HRMS analysis could not be directly
 applied  since the presence of known toxic components are only sug-
 gested by  HRMS.

 The improved correlation obtained when considering a toxicity index
 based on the twenty most influential rather than the total chemical
 analysis suggest  inadequacies in the model.  However, this infor-
mation can be used to provide direction for more intensive studies
 to determine casual  relationships between chemical composition and
observed biological  activity.

The comparison of mutagenic bioassay results with the chemical
analysis of the particulate samples is somewhat limited.  Of the

                                 70

-------
 eleven samples tested for mutagenic activity, only two samples
 (i.e., copper smelter, l-3y sample, and aluminum smelter,  l-3y
 sample)  indicated positive findings.   Since the mutagenic  screening
 test  performed in this research is relatively insensitive  to
 inorganic mutagens,  only the organic analyses were considered.   The
 aluminum smelter samples did contain by far the greatest concentra-
 tions of polycyclic  hydrocarbons,  and both samples possibly con-
 tained all nine highly carcinogenic aromatic constituents  identified
 by HRMS  (see Table 7).  However, both aluminum smelter  samples were
 tested for mutagenic activity,  and only the l-3y  sample tested
 positive.   The copper smelter l-3y sample  indicated a weak mutagenic
 response,  while tests of the 3-10y sample  were negative.   Chemical
 analysis of the copper smelter  samples were again very  similar to
 each  other; however, a very high concentration of arsenic  was noted
 in each.   Arsenic compounds have caused neoplasms in experimental
         (14)
 animals,      and some investig
'mutagenisis and carcinogenisis
        (14)
animals,     and some investigators suggest a common mechanism for
 6.5   CONSIDERATIONS FOR FUTURE RESEARCH
 This  research effort should be evaluated in light  of  its  intended
 goals as well as its programmatic constraints.   Several analytical
 strategies found in PMB's environmental source  assessment program
 were  evaluated for their ability to characterize the  potential
 environmental hazards associated with selected  industrial particulate
 emissions.  The scope of each analytical tool was  designed to be
 compatible with its companion techniques and the quality  of the
 sample that was assayed.  To subject a sample that is not representa-
 tive  of a given source to extensive biological or chemical character-
 ization would be an inefficient use of limited  resources.  To subject
 a  sample obtained at great expense to a cursory biological testing
 regime would likewise not be advised.  However, suggestions that
                                 71

-------
reflect the overall compatibility of the sampling and analysis strategy
can be made to improve future environmental source assessment programs.

This research effort has recognized the need for a more suitable filter
material with which to collect sub-micron particulate material. An
ideal filter would be biologically and chemically inert, not subject
to thermal degradation, and capable of satisfying the engineering
features of the sampling train (e.g., pressure drop).  The possible
alteration of the particulate samples resulting from collection,
handling, and disbursement is being considered in ongoing studies.

Several observations can be made concerning the chemical characteriza-
tion of the particulate samples.   The SSMS technique can be considered
a useful survey tool for characterizing the particulate samples, and
has the ability to focus interest on those samples with unusually high
concentrations of potentially hazardous elements.  The GC-MS analysis
should not be considered a survey technique, but can quantitatively
identify constituents of suspected hazardous samples.  The HEMS
analysis, although not easily quantifiable, can screen particulate
samples for a wide variety of hazardous and/or toxic substances.
Inorganic speciation and the partitioning of chemical species within
the biological assay media would aid in assessing the potential bio-
logical availability of various constituents.  Although those deter-
minations were not made in this research effort, they are being con-
sidered for appropriate placement in the overall environmental source
assessment program.

The two bioassay procedures performed in this research effort should
be validated by comparison to currently recognized standard procedures
so that routine screening tests can be compared to other studies.   If
a proportional ranking system is required of the RAM bioassay screen-
ing test,  the protocol will require expansion to include extended
                                 72

-------
concentration-response data and additional response indicators (e.g.,
functional impairment, membrane integrity).

Additional solvent vehicles could be added to the mutagenicity
screening test so that chemical constituents with low solubility in
DMSO can also be evaluated for mutagenic activity.  If normalization
of exposed bacterial populations to reflect the inherent toxicity of
the sample is envisioned, then consistent concentration-response
data must be generated during testing.  A comparison of alternate
protocols (see Section 6.3.2) can define the better system for a
specific testing level (i.e., Levels 1, 2, or 3).

It should be recognized that a complete and comprehensive environ-
mental source assessment program dictates a sampling and analysis
strategy whose philosophy and structure permit a maximum use of
available resources.
                                 73

-------
                           7.0  REFERENCES
 1.  TRW Systems Group.  VBLI Mechanical Subsystem Cleanliness Controls.
        Cl #11982.  October 1974.

 2.  TRW Systems Group.  Interim Report for Fabrication and Calibration
        of Series Cyclone Sampling Train.  24916-6019-TU-OO.  April 1975.

 3.  TRW Systems Group.  Interim Report; Field Sampling for Cytotoxicity
        Test Samples.  24916-6026-RU-OO.  October 1975.

 4.  Federal Standard, Clean Room and Work Station Requirements, Controlled
        Environment.  #209B.  April 1973.

 5.  Waters, M. D., D. E. Gardner, C. Aranyi, and D. L. Coffin.  Metal
        Toxicity for Rabbit Alveolar Macrophages in Vitro. Envir. Res.
        9:32-47.  1975.

 6.  Waters, M. D., D. E. Gardner, and D. L. Coffin.  Cytotoxic Effects
        of Vanadium ,on Rabbit Alveolar Macrophages in, Vitro. Tox. Appl.
        Pharm.  28(2):253-263.  1974.

 7.  Ames, B. N., E. G. Gurney, J. A. Miller, and H. Bartsch.  Carcinogens
        as Frameshift Mutagens:  Metabolites and Derrivatives of 2-Acetyl-
        aminofluorene and Other Aromatic Amine Carcinogens.  Proc. Nat.
        Acad. Sci. 69:3128-3132.  1972.

 8.  Ames, B. W., F. D. Lee, and W. E. Durston.  An Improved Bacterial
        Test System for the Detection and Classification of Mutagens
        and Carcinogens.  Proc. Nat. Acad. Sci.  70:782-786.  1973.

 9.  Ames, B. W., W. E. Durston, E. Yamasaki, and F. D. Lee.  Carcinogens
        as Mutagens:  A Simple Test System Combining Liver Homogenates
        for Activation and Bacteria for Detection.  Proc. Nat. Acad.  Sci.
        70:2281-2285.  1973.

10.  Gletten, F., V. Weekes, and D. J. Brusick.  In Vitro Activation  of
        Chemical Mutagens:  I. Development of an In Vitro Mutagenicity
        Assay Using Liver Enzymes for Activation of Dimethylnitrosamine
        Mut. Res. 9:32-47.  1975.

11.  National Academy of Sciences.  Particulate Polycyclic Organic
        Matter,  pp. 6-12.  1972.

12.  Dixon, W. J., and F. J. Massey, Jr., Introduction  to Statistical
        Analysis, (3rd ed).  McGraw-Hill Co.  1969.
                                  74

-------
13.  Siegel, S.  Nonparametric Statistics for the Behavioral Sciences.
     McGraw-Hill, N.Y.  1956.  pp. 202-240.

14.  Christensen, H.E., T.E. Luginbyhl, and B.S. Carroll  (eds).
     Registry of Toxic Effects of Chemical Substances. U.S. Department
     of Health, Education and Welfare.  1975.

15.  American Conference of Governmental Industrial Hygienists.
     Documentation of Threshold Limit Values.  1971.
                                  75

-------
TECHNICAL REPORT DATA
(Please read Initructions on the reverse before completing)
11. REPORT NO 2.
EPA-600/2-76-137
4 TITLE AND SUBTITLE Evaluation of Selected Methods for
bhemical and Biological Testing of Industrial Parti-
culate Emissions
7. AUTHOR(S)
H. Mahar
9. PERFORMING OROANIZATION NAME AND ADDRESS
[The Mitre Corporation
[Westgate Research Park
McLean, Virginia 22101
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
3. RECIPIENT'S ACCESSION1 NO.
5. REPORT DATE
May 1976
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
1AB013; ROAP 21ADD-BG
11. CONTRACT/GRANT NO.
68-02-1859, Task 5
ftJX¥£^r W-W C°VERED
14. SPONSORING AGENCY CODE
EPA-ORD
ITS. SUPPLEMENTARY NOTES project officer for this report is L.D. Johnson, Mail Drop 62,
Ext 2557.
   ABSTRACTThe report gives results of chemical analyses and cellular biological assays
 |performed on size-classified particulate material collected using a new series cyclone
 sampling train at nine industrial sites. The exercise was formulated to determine
 performance of the train and whether the  chemical analyses  or the bioassays,  alone
 or in combination,  were sufficient to characterize the hazards associated with parti-
 culate emissions.   This program lends support to the view that size-classified parti-
 culate matter is needed for the various chemical or biological tests.  Elemental
 analysis and partial organic characterization of the  particulate samples have been
 performed.  A cellular bioassay, utilizing rabbit alveolar macrophages, has been
 used to provide a rank ordering of particulate samples  in terms of their observed
 cytotoxic activity.  A bacterial screening technique, utilizing several histidine
 deficient Salmonella typhimurium strains, has been used to study the  mutagenic
 potential of the particulate samples.  Attempts to correlate observed  biological
 activity with chemical analyses  are provided.
Jl7. KEY WORDS AND DOCUMENT ANALYSIS
|a. DESCRIPTORS
Air Pollution, Dust, Industrial Plants,
Chemical Analysis , Bioassay, Sampling,
Cells (Biology), Properties, Cytology,
Size Separation, Toxicity, Bacteria,
Selection, Mutagens, Mutations
- - 	 -
[18. DISTRIBUTION STATEMENT
1 Unlimited
b.lDENTIFIERS/OPEN ENDED TERMS
Air Pollution Control
Stationary Sources
Particulates
19. SECURITY CLASS (This Report)
Unclassified
20. SECURITY CLASS (This page)
Unclassified
c. COSATI Field/Group
13B, 11G
07D,06A/06O,
06C 14B
07A/13H,06T,
06E 06C
21. NO. OF PAGES
84
22. PRICE
EPA Form 2220-1 (9-73)
                                         76

-------