x>EPA
           United States
           Environmental Protection
           Agency
           Industrial Environmental Research  EPA-600 9-81-01!
           Laboratory         April 1981
           Research Triangle Park NC 27711
           Research and Development
Proceedings: Second
Symposium on
Process
Measurements for
Environmental
Assessment

February 25-27,1980

-------
                                        EPA 600/9-81-018
                                        April 1981
                     PROCEEDINGS


                  Second Symposium

                         on

                PROCESS MEASUREMENTS
            FOR ENVIRONMENTAL ASSESSMENT
Philip L. Levins, Judith C. Harris, Karen D. Drewitz
               Arthur D. Little, Inc.
                     Acorn Park
           Cambridge, Massachusetts 02140
             EPA Contract No. 68-02-3111
       EPA Project Officer:  Larry D. Johnson
     Industrial  Environmental  Research  Laboratory
          Office of  Research and  Development
         U,S,  Environmental Protection  Agency
           Research  Triangle Park,  NC   27111

-------
                                PREFACE

These Proceedings consist of manuscripts for twenty-two papers and
abstracts for twelve poster sessions presented at the Second Symposium
on Process Measurements for Environmental Assessment held at the Sheraton-
Atlanta Hotel in Atlanta, Georgia on February 25-27, 1980.

The symposium focused on the state-of-the-art of sampling and analysis
techniques that are appropriate for process measurements in the context
of an environmental assessment program.  Methods for qualitative and
quantitative chemical characterization of organic and inorganic species
in process and discharge streams and biological assays of environmental
samples were included.  The symposium represents a continuing effort on
the part of the Process Measurements Branch of the EPA's Industrial
Environmental Research Laboratory at Research Triangle Park to share
recently developed methodology and encourage the interchange of ideas
among researchers in government, industry and academia.  Some specific
topics represented include:

     •  Use of sorbents for sampling

     •  Sampling of reactive species

     •  Sampling and analysis methodology for coal conversion
        processes

     •  Advanced inorganic analysis techniques

     •  Advanced organic analysis techniques

     •  Application of bioassay methods to complex samples

The General Chairman of the symposium was Mr. James A. Dorsey, Chief of
the Process Measurements Branch, IERL-RTP.  The symposium coordinators,
Dr. Philip Levins and Dr. Judith Harris of Arthur D. Little, Inc., wish
to express their gratitude to Mr. Dorsey for his active participation in
the preparation for this symposium.  They would like to acknowledge the
contributions of session moderators Peter W. Jones of the Electric Power
Research Institute, Michael R. Guerin of Oak Ridge National Laboratory,
David L. Stalling of Columbia National Fisheries Laboratory and Larry
Johnson of Process Measurements Branch, IERL-RTP.  Thanks are due to
Ms. Karen Drewitz, Mr. John Piecewicz and Ms. Debra Sorlin of Arthur D.
Little,  Inc.  who provided support services at the symposium.
                                   iii

-------
                           TABLE OF CONTENTS

                                                                  Page

     PREFACE                                                      iii

     MANUSCRIPT:

 1.   "The Integrated Approach to Chemical-Biological Analysis,"
        M.  R.  Guerin                                                1

 2.   "Description of Bioassay Results and Projection of Bio-
        assay  Study Needs in Support of a Major Synfuels
        Industry," A.  Kolber, M. B. Wilkie,  T.  J.  Wolff and
        D.  G.  Nichols                                              17

 3.   "Collection and Recovery of Organics from Water Using
        XAD-2  and XE-347 Resins," J. C. Harris and M. J. Cohen     41

 4.   "Approaches to Level 1 IR and LRMS Measurement and
        Spectral Interpretation," W. F. Gutknecht  and
        A.  Gaskill, Jr.                                             55

 5.   "Characterization of Coal Gasification By-products and
        Ambient Air Samples from a Lurgi Gasification Facility
        by Selective Detector Gas Chromatography," K. W. Lee,
        D.  S.  Lewis, C.  H. Williams and K. J.  Bombaugh             74

 6.   "Characterization of Process Streams from Liquefaction
        of Low-Rank Coal with Synthesis Gas,"  B. W. Farnum,
        S.  A.  Farnum and C. L. Knudson                             87

 7.   "Four-hour Algal Bioassays for Assessing  the  Toxicity of
        Coal-derived Materials," J. M. Giddings                   104

 8.   "Direct Determination of Polynuclear Aromatic Hydrocarbons
        in Coal Liquids and Shale Oil," A. P.  D'Silva, Y. Yang
        and V. A. Fassel                                          117

 9.   "Synchronous Fluorescence and Phosphorescence at Room
        Temperature for Levels 1 and 2 Organic Analysis,"
        R.  B.  Gammage, T. Vo-Dinh and P. R.  Martinez              119

10.   "Applications of Contaminant Enrichment Modules to
        Organic Trace Analysis," D__. L. Stalling, J. D. Petty,
        L.  M.  Smith and G. R. Dubay                               134

11.   "A Multichannel,  Remote Controlled Teflon and Glass
        Positive Displacement Apparatus for Collecting Trace
        Organics from Environmental Sources,"  D. C. Tigwell
        and D. J. Schaeffer                                       143


                                   iv

-------
Table of Contents (Cont'd)

                                                                   Page

12.  "Analysis of Organic Acids from Diesel Exhaust in Mine
        Air," I. Bodek and K. T. Menzies                           155

13.  "A New Electrochemical Approach to Trace Level Aldehyde
        and Ketone Analysis," R. P. Baldwin, J. F. Price and
        J. Siria                                                   169

14.  "A Computer Interfaced Toxicity Testing System for
        Variable Effluent Loading," J. Cairns, Jr. and
        K. W. Thompson                                             183

15.  "On-line Monitoring and Toxic Materials in Sewage at
        the Lawrence Livermore Laboratory," M. Auyong,
        J. L. Cate, Jr. and D. W. Rueppel                          199

16.  "Comparison of Four Leachate Generation Procedures,"
        D. W. Bause                                                207

17.  "Possible Effects of Collection Methods and Sample Pre-
        paration on Level 1 Health Effects Testing of Complex
        Mixtures," D. J. Brusick                                   226

18.  "The Dynamic Interaction Between Vapor Phase and Parti-
        culate Materials," D. F. S. Natusch                        240

19.  "Coal Fly Ash as a Model Complex Mixture for Short-Term
        Bioassay," G^ L. Fisher, C. E. Chrisp and F. D. Wilson     241

20.  "Elemental Analysis for Environmental Assessment Measure-
        ments," K. T. McGregor                                     255

21.  "Level 2 Inorganic Sampling and Analysis Methodology
        Applied to FGD Systems," R. F. Maddalone                   275

22.  "FTIR:  A Tool for Both Organic and Inorganic Analyses
        in Environmental Assessment Programs," R. L. Barbour
        and R. J. Jakobsen                                         305

     POSTER SESSION:

  1.  "Measurement Methods for SOX and NOX in the Presence of
        NH3," D. S. Chase and B. M. Myatt                          323

  2.  "Evaluation of Stable Labeled Compounds as Internal
        Standards for Quantitative GC/MS," B. N. Colby             324
                                   v

-------
Table of Contents (Cont'd)                                         Page

 3.  "Level 1 Environmental Assessment:  Fluidized-Bed
        Combustion," R. R. Hall, P. F. Fennelly, G. T. Hunt
        and R. J. Kindya                                           325

 4.  "Evaluation of Level 1 Analysis Procedures,"                  326
        R.  K. M. Jayanty, W. F. Gutknecht, A. Gaskill, Jr.
        and D. E. Lentzen

 5.  "Anomalous High Total Cyanide Results Due to Nitrite in
        Biotreater Effluents:  The Key to the 'Cyanide-
        Generation Syndrome' in Biotreater Technology,"
        R.  A. Johnson, J. C. Rapean and T. P. Hanson               327

 6.  "Design Modifications to the Source Assessment Sampling
        System," R. L. Campbell, K. W. Mason, W. R. Parker
        and T. J. Wagner                                           328

 7.  "Diagnosis of Metal Speciation in Aqueous Solutions,"
        B.  McDuffie and P. Figura                                  329

 8.  "Solid Sorbents for Air Sampling," J. F. Piecewicz,
        J.  C. Harris and P. L. Levins                              330

 9.  "Screening - Biological and Chemical Data Analysis,"
        N.  H. Sexton and L. I. Southerland                         331

10.  "Analysis of Coal Liquid Subfractions Which Exhibit
        Microbial Mutagenic Activity," B. W. Wilson,
        R.  A. Pelroy, M. R. Peterson and W. C. Weimer              332

11.  "Quantitation of Polycyclic Aromatic Hydrocarbons in
        Complex Mixtures by High Resolution Glass Capillary
        Gas Chromatography/Mass Spectrometry Using Selected
        Ion Monitoring," G. A. Gibbon and C. M. Wtvite^              333

12.  "Analytical Results of a PCB Test Incineration,"
        C.  D. Wolbach. W. F. Fitch, N. Flynn and B. Markoja        334

     MANUSCRIPT AND POSTER SESSION AUTHOR INDEX                    335

     APPENDIX:  Symposium Participants                             338
                                   VI

-------
         THE INTEGRATED APPROACH TO CHEMICAL-BIOLOGICAL ANALYSIS

                              M. R. Guerin
                      Analytical Chemistry Division
                      Oak Ridge National Laboratory
                       Oak Ridge, Tennessee  37830
                                ABSTRACT

Environmental and  health  assessments are  increasingly dependent on  the
results of biological testing.   Biological  testing,  in turn,  is  becoming
increasingly  dependent  on  physical  and chemical  assistance to  achieve
reliable and  interpretable  results.   This  is partly because  of  the  com-
plex and possibly changing compositions of the samples and partly because
little is known  about  the response  of  many bioassay  systems to  complex
mixtures.

A  currently  common interaction between disciplines  has been termed  the
"matrix approach".  Contributing groups maintain their disciplinary focus
but generate  data  or materials  needed  by  other groups  to  reach  a mutual
programmatic  objective.   Examples are  the USEPA  environmental assessment
projects and  the USEPA/USDOE synfuels research sample repository.   A more
intimate interaction occurs as the  contributing  groups focus more  com-
pletely on  the  program objective.   An example  is  the combined use  of
chemical class fractionation and biological  testing  to identify  constit-
uents responsible  for biological effects.   Further  integration occurs  as
the fundamental issues of delivery and  dosimetry  are addressed.   Systems
to deliver smokes  to experimental animals  for  inhalation  bioassay and  to
monitor the  exposures,  methods  for  determining the dose achieved,  and
studies  of   physical-chemical   influences   on  bioassay   responses   are
examples of current work in this area.

The most effective approach to  an integrated chemical-biological  program
requires that each of the  levels  of  interaction be  present  and  that  the
collaborating groups give the programmatic objective as much  attention  as
is given to disciplinary concerns.
                              INTRODUCTION

There are a number of approaches (Table 1) to meeting program objectives.
All  rely  on applying  the  required scientific  disciplines  to reach  the
objective.  In general,  the  task of meeting  a  program  objective  becomes
more  difficult  as  the  number  of  scientific  disciplines  involved  in-
creases.  The nature of the product required and the degree  of efficiency
with which the objective is reached depend  very highly  on the nature  and
degree of interdisciplinary interactions.   This is  particularly the  case
for Environment Assessment Programs because they deal with a wide  variety
of  sub-objectives  and  require the  input  of very many  scientific  disci-
plines .

In  the  "independent"  approach (Table  1),  each  group functions  indepen-
dently  and  interacts  with other  disciplines  only when it   is  somehow

-------
beneficial to  their  own work to  do so.   The groups  focus  on peers  in
their  individual  disciplines for  both  guidance and  reward.    The  prime
product  is  generally  knowledge  or methods  that constitute  fundamental
contributions to their disciplines.  It is usually a very inefficient way
to  reach an applied  objective  but there  is always  the  chance that  an
advance of such fundamental importance will be made as to solve the prob-
lem entirely or at least  greatly  reduce the time required for  its  ulti-
mate  solution.    The  primary  limitation  is  novel  ideas,   creativity.
Funding  sufficient for  the most sophisticated  tools  and  to  attract  crea-
tive talent is also a limitation.

There  are  a  number of  interactive (or "cooperative") approaches.   They
vary in  the  degree  of interaction and  interdependence.   Interaction can
range  from little more than a program  coordinator's  success  in bringing
the groups together to  share  results  to an interaction  that  could  class-
ify as an "integrated" approach.   The focus ranges from pure  disciplinary
to pure project orientation.  The result tends to be data.  As the  inter-
action grows stronger the effort results in "information" (a  synthesis of
the data)  and  eventually "knowledge"  (an  understanding  of the data and
its implications).  The  interactive  approach tends to be a  rather  effi-
cient  approach to meeting an  applied  objective.   The efficiency achieved
usually depends on the talents of  the project  officer.   There is no lack
of  immediate  problems which  can  be attacked  using available  tools and
knowledge so the current limitation tends to be funding available.

In  the "integrated"  approach,  the programmatic  objective  (e.g.  in-vitro
bioassay of complex mixtures, environmental  control  technology,  environ-
mental assessment)  becomes  the disciplinary objective.    There must  be
either much disciplinary interaction or "total" interaction—to the point
where  the  scientific  discipline  takes only equal (but not  less)  import-
ance to  the  applied discipline.   The  focus tends  to be on  the  project
(applied  discipline)   and  the  individual  scientific  disciplines   being
drawn  upon to  address  the problem.  The result  is  fundamental  knowledge
and methods as is the  case  for  the independent approach but  much  of the
contribution is  to the  applied  discipline  rather  than  the  fundamental
scientific discipline.   The integrated  approach can be either very in-
efficient (as when the  applied  objective is peripheral to the  issue)  or
exceptionally efficient (a highly  relevant  applied  objective).   Creativ-
ity is an  important  limiting factor here  as  it  was  for  the  independent
approach because  both  rely  on insight to  resolve  fundamental  issues.
Funding is somewhat more limiting because many of today's problems  can be
addressed using currently available tools and knowledge.

Examples of  on-going  programs related  to environmental  assessments are
presented as indicative of  the  nature of the  products generated at var-
ious levels of interaction.

                   INDEPENDENT-INTERACTIVE APPROACHES

Most USEPA  and USDOE  site-specific environmental  assessments may  be  in
this  category.    Much  of  the ultimate integration  of  results  is  the
responsibility of  the project officer  or  of  a disciplinary assessment
group  included  among  the contractors.   As examples  of  these activities

-------
are covered  extensively throughout these  proceedings,  they will  not  be
discussed further here.

Paraho/SOHIO Matrix Equipment

The Paraho/SOHIO study  (Table  2)  (2)  is an example of  a  loosely coordi-
nated interaction that  relies  most heavily on  the disciplinary expertise
of  its   participants.     The   approach  has  proven  exceptionally  cost-
efficient to its project offices  because  it draws upon  the mutual inter-
ests (and the funds) of several agencies.

Samples were provided  by  the Department of the  Navy  through Development
Engineering, Inc.  (DEI),  and  the  Standard Oil Company of  Ohio  (SOHIO).
DEI produced 50,000  barrels  of  shale oil  by  the Paraho  above-ground
retorting process  and  forwarded  the product  to  SOHIO for refining  on
behalf of the Department of the Navy.   The products are being  studied  as
military  fuels  and  for unforeseen environmental/health  effects  by  the
Department of Defense.

Samples provided for the matrix study have been  distributed  (Table 2)  to
approximately fifteen different groups  around  the country  expert  in var-
ious chemical,  biological, and ecological  analysis  areas.   Each  group  is
examining the samples as part  of  its basic research program.   The groups
publish their results as they see fit and meet  at least  annually  to share
experiences and results.

A highly  definitive  characterization  of the samples is thus obtained  in
exchange for highly relevant samples  for fundamental study.

Synfuels Research Materials Facility-Repository

The mechanism  to achieve  the  matrix study  involved  establishing  (1)  a
central materials "repository" or  "facility" to  ensure  that  the  investi-
gators were  provided with the proper  samples  and  that the sample sup-
pliers were provided with  the  research  findings.  While the matrix pro-
gram  is  loosely interactive  (except within  some  of  the  participating
institutions),   the  "facility" must  be highly  interactive and  program-
oriented.   Its  activities  extend  to  providing special  services  in samp-
ling  and characterization  and  in  bioassay   preparation.    Of   special
importance,  it  also seeks to fill gaps  in  the  data base and  to follow  up
significant  observations  by  soliciting  the  input  of  other  groups  in
exchange for samples.

Several independent and "matrix"  experiments are (7) underway in  the gen-
eral area of synfuels   technologies though none  are  as fortunate  to  be
working with such  relevant samples  as  is the  Paraho/SOHIO study.   The
studies  (Table  3)  involve a  variety  of sample  types,  endpoints,  and
institutions.  Samples  are described in the broadest  of terms  and do  not
emphasize process names or specific technologies.   The  reason  is  that  we
have  no  evidence  that the  samples  are   suitable  for  process-specific
assessments.   Operating conditions at  the time  of  sampling as   well  as
sampling  and sample  handling  conditions are seldom adequately specified
to  draw  process-specific  conclusions.    The samples  currently available
are best suited  for  generic research into  the  general  chemical  and

-------
biological properties of coal- and shale-derived petroleum substitutes as
compared to petroleum.   They are especially valuable  for intercomparing
chemical and biological properties,  comparing and developing methods,  and
as "training" or "validating" materials for new contractors.

As more  groups  examine a  given  sample and  as  the groups  interact  more
closely, the  samples  can  become very  precious.   This is  especially  the
case  for highly interactive  programs  where the  groups  seek  to  compare
mutagenicity-teratogenicity-carcinogenicity  or  to  identify carcinogens,
or to  compare new methods with  established methods.  A series  of  five
materials  is  currently being collected  and characterized  to serve  as
materials  suitable  for extended study.   They  differ  from  other  samples
primarily  in  that  30-50 gallons will  be available  to draw from  and  in
that they  are being  stored and  periodically monitored for stability  with
attention  to  long-term utility.   The  resulting  data base  will  period-
ically be published and distributed with the samples.

These materials, the  current "matrix"  type  experiments,   and  the  current
"facility" concept are targeted to generic research.  They serve to bring
the talents of independent research groups to bear on programmatic objec-
tives without  jeopardizing their primary  disciplinary responsibilities.
The samples and facility serve as a  "resource"  to the scientific  commun-
ity and  matrix  program.   Samples are  available to  any investigator  fol-
lowing approval of the sample supplier.

This is  a  unique route to  an interdisciplinary  objective because  it  does
not demand individual  initiative for a programmatic interaction.

                    INTERACTIVE-INTEGRATED APPROACHES

Specific  aspects  of  USEPA and  USDOE  environmental  assessments  require
highly  interactive  approaches involving  several  scientific disciplines.
Fundamental advances  in delivering  test materials  to biological  systems
for bioassay  and in  determining the  dose  achieved  or  causative agents
present  are most likely when the "integrated" approach is used.

Identifying Mutagenic  Constituents of Petroleum Substitutes

The traditional approach to  identifying bioactive constituents of complex
organic  mixtures  involves several  steps.   The  starting sample  is  bio-
assayed,  divided  into  chemically  or  physically  defined  fractions,  the
fractions  are bioassayed,  and the process is repeated centering on frac-
tions and  subfractions exhibiting the  desired biological response.   Con-
stituents  of  the  bioactive subtraction are  identified,  suspect constit-
uents  are  synthesized, and  the synthesized compounds are  subjected  to
bioassay to confirm their importance.

Knowing  the  determinant constituents  identifies  compounds  or properties
worthy of  chemical analysis  and monitoring.  It can also suggest measures
which might be  taken  to reduce  the  biological  activity of the materials.
The approach  generally provides fundamental information  on bioactivity-
chemistry  interrelationships and  almost  always  yields  new methods  of
chemical and biological analyses.

-------
The approach is designed to identify bioactive constituents but it yields
useful secondary  information  in the process.  Chemical  class  fractiona-
tion provides data on the general composition of  the  samples.   This is a
semi-preparative scale fractionation (typically 5+ gm of starting sample)
designed for bioassay  rather  than quantitation so recoveries  can be low
and precision only  fair.   Trends are  still evident.   The fractionation
can be carried out with quantitation in mind if desired.

Results are  summarized (Table 4) for  two different  petroleum  crudes,  a
crude  shale  oil,  a series  of  coal-derived products  (ASOH-atmospheric
still  overhead,  ASB-atmospheric still  bottoms,  VSOH-vacuum still  over-
heads, VSB-vacuum  still  bottoms), and  a  pair of  coal-derived  "distill-
ates" that are the same except  that  one had been  hydrotreated  with rela-
tively high  severity.    The  procedure  yields  a  "volatiles" fraction,  a
fraction insoluble in  ether,  ether  soluble  acids,  insoluble acids,  ether
soluble bases,  insoluble bases,  neutral  saturates,  neutral  aromatics,
neutral polycyclic aromatics,  and a  neutral polar fraction.  Although the
results are crude, they prove  very  useful for judging  the  nature of the
samples—volatility,   intractability,   acidity,    and  aromaticity   for
example.

The mutagenicities  of  the individual  fractions  as determined  using the
Ames  test  for bacterial mutagenicity  (all  with  TA-98  and Aroclor  1254
induced S-9 activation) are expressed  (Table 5)  as the number  of rever-
tant colonies per microgram of fraction tested.    The greater  the number
of revertants per microgram,  the  greater  the mutagenicity.   As  a further
benchmark, benzo(a)pyrene exhibits about 50 Rev/yg.

A cursory examination of this  data,  too, proves useful.   We note that the
volatile material never exhibits much  mutagenicity.   The  acidic constit-
uents of the oils tend not  to be mutagenic.  The  basic  constituents and
the polycyclic  aromatic  neutral  constituents, however,  are quite  often
mutagenic.  For any given petroleum  substitute, the  ESB fraction is more
mutagenic than the neutral polycyclic aromatic fraction.  Some  of the ESB
fractions rival or exceed the  mutagenicity of pure benzo(a)pyrene.

Finally, assuming (10) that the mutagenicity of  the  starting material is
the sum of the  mutagenicities of the  fractions (adjusted  for  the weight
percentage content of the fractions), one can estimate the total mutagen-
icities of the samples and the percentage of the  total contributed by the
individual fractions.  Using  this approach  (3-5)  we find  (Table  6) the
petroleum crudes only  very  slightly  mutagenic,  the shale  oil  to  be more
mutagenic, and the coal-oils to vary greatly in mutagenicity depending on
their volatility and processing.

There are also  other important  trends.  The most  unexpected of these is
the high contribution  to the  mutagenicities  of the petroleum  substitutes
by  their  alkaline constituents.   The  alkaline  and  polycyclic  aromatic
neutral constituents are not  only the most  active subtractions but they
are also the primary contributors to the mutagenicity of the samples as a
whole.

These observations constitute  important results  generated  in  the course
of  the primary  study which itself  focuses on identifying  the  bioactive
                                     5

-------
constituents.  The mutagenicity is found (Figure 1) to reside in the rel-
atively non-volatile  material  and concentrate  in  the basic  and  neutral
fractions.   The  neutral  mutagens  are (9) aromatic  rather  than aliphatic
and  are  concentrated  in the  3-5  fused ring  subtractions.    Polycyclic
aromatic hydrocarbons  (9), neutral nitrogen heterocyclics  (13),  and neu-
tral polar  constituents  all exhibit mutagenicity.   Both the  parent  and
simply  alkylated polycyclics  and  the  multialkylated  polycyclics  con-
tribute (6) to the mutagenicity.   The alkaline mutagens are found (10-12)
to  be  among the  more polar and  more  aromatic  constituents  and  reside
mostly in  the 2-4  fused ring fractions.   More  detailed study  (11,12)
shows the determinant alkaline mutagens  to be polycyclic aromatic primary
amines although there  is  also  some  contribution by the alkaline nitrogen
heterocyclics.

More detailed  study can  be  carried  out to identify  individual constit-
uents at the isomeric  level  if need  be.   Alternatively,  the  same general
approach can be applied  to other  sample types  or  used with other biolog-
ical endpoints.

This general  approach requires considerably  more  interaction  than does
the disciplinary matrix approach because chemical and biological observa-
tions dictate subsequent biological and  chemical steps.

Delivery and Dosimetry

Complete integration of  chemistry and biology  tends to  occur as the cen-
tral  issues  of   "delivery"  and  "dosimetry"  are  addressed.    This  is
especially  so  when  complex mixtures are  involved and when  the constit-
uents of interest are organic.

Among the most complex efforts to "deliver"  a  relevant  material for bio-
testing was  that  carried out by  the smoking and  health  research commun-
ity.  The  object  (13)  was to develop a  way to  smoke  a  cigarette  the  way
people do  (but more  reproducibly)  and  route  the  smoke  so  that  it  was
inhaled by experimental  animals  in a  form  not  significantly different
from that  inhaled by humans.  At one point  there were  some  thirty cri-
teria that had to be met by the device.

The work has resulted in systems costing a few thousand dollars and up to
expose dogs  or rodents,  to systems  costing more than $100K  to simultan-
eously expose  several hundred rodents.    In  the process,  many sophisti-
cated issues in  dosimetry,  aerosol  physics and monitoring,  aerosol chem-
istry, sampling and analysis, and inhalation bioassay were addressed.

An  example  somewhat more  relevant  to  the subject  of this  symposium is
study of  the beeswax  pellet implantation  model for  producing cancer in
animals.   In this model, the material to be tested is mixed or encased in
beeswax and  the  beeswax pellet is surgically  implanted  into the animal.
When smoke condensate was applied (14) to the lungs of rodents using this
technique,  it  was found that the condensate  from only  a  few cigarettes
was sufficient to produce lung tumors.

-------
The model  was studied  (15)  chemically  by  preparing  pellets  containing
cigarette  smoke  condensate  to  which  l^C  tracers  of various  compounds
were added.   Physiological  saline  was pumped across  the  pellets  and the
pellets  were  taken  for  analysis  of  ^C after  various  times.    It  was
found that  some  constituents are released very quickly while  others are
released more slowly.   Studies using  -^C-BaP  or dotriacontane  showed
that 98% or more of the compound remained in  the  pellet after  as  many as
28 days  of treatment.   The rate  of  release was  found  to  parallel  the
hydrophilicity of the individual compound.   The more hydrophilic the com-
pound,  the more rapidly it is released.

Using this  model,  the  physiological  test system  is  exposed  to   a  con-
stantly changing test material and  is  never  exposed  to a  material repre-
sentative of  the initial sample.   The  bioassay  model is  therefore  not
applicable  to complex mixtures  containing a wide  variety  of both hydro-
philic and  lipophilic constituents.

We have evidence to  suggest  that similar factors  operate  upon  subjecting
lipophilic  samples to in-vitro bioassays as  is  increasingly done  in sup-
port of environmental assessments.   When crude  oils  are subjected to the
Ames test,   the sample is frequently observed to distribute throughout the
plate as  discrete  droplets.   "Halos" of  discoloration  may be observed
around many of  the  droplets suggesting  a release  of at  least  some  con-
stituents into the  test  medium.    Fundamental  equilibrium considerations
would suggest that the material released first will be enriched in hydro-
philic  (e.g.  phenolic)  constituents.    Localized  high concentrations  of
such  constituents   may  be   responsible  for  the  "toxicity"  frequently
reported as interfering  in  in-vitro  bioassays.    The  inhomogeneous  dis-
tribution  of  the sample  in the plate  and  the effect that the  droplet
size, number,  and spatial distribution on ultimate dose could account for
much of the difficulty in reproducing the results  of the bioassays.

A  similar  effect is  likely to accompany  other in-vitro  bioassays  when
applied to  complex  mixtures.    The study of  these  phenomena require  a
truly integrated biological-chemical approach, sophisticated experimental
design,  and methods  and instrumentation that  may not  even exist.   The
results—interpretable and reproducible bioassays, logical limitations of
the bioassays, and  fundamental  delivery and dosimetry knowledge—may be
worth the effort.
                             ACKNOWLEDGMENTS

I wish to acknowledge the  support  of  the  National  Cancer Institute Smok-
ing and  Health  Program,  the Department of  Energy,  Office of  Health  and
Environmental Research,  and the Environmental  Protection Agency  Health
Effects Research Laboratory at Research Triangle Park, NC.

Research sponsored jointly by the U.S. EPA Health Effects Research Labor-
atory and the Office of Health and Environmental Research,  U.S.   Depart-
ment of  Energy,  under  contract W-7405-eng-26 with Union  Carbide  Corpor-
ation.

-------
                              REFERENCES
1.  Coffin,  D. L., Guerin,  M.  R.,  and Griest,  W.  H., "The  Interagency
    Program in Health  Effects  of  Synthetic  Fossil Fuels  Technologies:
    Operation  of a  Materials  Repository,"  Proc.  Symp.  on  Potential
    Health and Environ. Effects of Synthetic Fossil Fuel  Technologies,
    Gatlinburg, TN,  1978,  CONF-780903,  153-156  (1979).

2.  Environmental Protection Agency,  Proc.  Conf.  on A  Matrix  Approach  to
    Biological Investigation of  Synfuels,  Research Triangle Park, NC,
    1979 (in press).

3.  Epler, J.  L., Young,  J,  A., Hardigree,  A. A., Rao,  T.  K., Guerin,  M.
    R., Rubin, I. B., Ho,  C.-h.,. and Clark,  B. R.,  "Analytical  and  Bio-
    logical Analyses of Test Materials from  the  Synthetic  Fuel  Technol-
    ogies, I.  Mutagenicity of Crude Oils  Determined by  the Salmonella
    typhimurium/Microsomal  Activation  System,"  Mutat.  Res.,  57,  265
    (1978).

4.  Epler, J. L., Rao, T. K., and  Guerin,  M. R. ,  "Evaluation  of  Feasi-
    bility of Mutagenic Testing of Shale Oil Products,"  Environ.  Hlth.
    Persp.,  _30, 179-184 (1979).

5.  Epler, J.  L. , Clark,  B. R. ,  Ho,  C.-h.,  Guerin, M.  R. , and Rao,  T.
    K., "Short-Term Bioassay of Complex Organic Mixtures,  Part II, Muta-
    genicity  Testing,"  in  Application  of  Short-Term  Bioassays  in the
    Fractionation and Analysis  of  Complex Environmental Mixtures, M.  D.
    Waters,  S. Nesnow,  J. L. Huisingh,  S.   S.   Sandhu,  and  L.  Claxton,
    Eds., Environ.  Science  Res.,   15,  269-290,   Plenum  Press, New  York
    (1979).

6.  Griest,  W. H., Caton,  J. E.,  Guerin, M.  R.,  Yeatts, L. B.,  Jr., and
    Higgins,  C.   E.,  "Extraction   and  Recovery  of Polycyclic  Aromatic
    Hydrocarbons  from Highly Sorptive  Matrices  Such As Fly  Ash," Proc.
    4th  Intern.  Symp.  on Polynuclear  .Aromatic  Hydrocarbons,  Columbus,
    OH, 1979 (in press).

7.  Griest,  W.  H. ,   Coffin,  D.  L. ,  and Guerin,  M.  R.,  "Fossil  Fuels
    Research Matrix Program," Oak  Ridge National Laboratory  Report,  (in
    press).

8.  Guerin,  M. R. ,  Maddox,  W.  L. , and Stokely,  J.  R.,  "Tobacco Smoke
    Inhalation Exposure:   Concepts and Devices,"  in Proc.  Tobacco Smoke
    Inhalation Workshop on  Experimental Methods  in Smoking  and  Health
    Research, G.  B.  Gori,  Ed.,  DHEW Publication No. (NIH)  75-906,  31-44,
    (1975).

9.  Guerin,  M. R., Epler, J. L. , Griest, W.  H. ,  Clark, B. R.,  and  Rao,
    T. K. , "Polycyclic Aromatic Hydrocarbons  from Fossil Fuel Conversion
    Processes," in Carcinogenesis, Vol. 3:   Polynuclear Aromatic  Hydro-
    carbons , P. W. Jones  and R.  I. Freudenthal, Eds., Raven Press, New
    York, 21-33 (1978).

-------
10.   Guerin,  M. R. , Clark,  B.  R.,  Ho, C.-h., Epler,  J.  L. ,  and Rao,  T.
     K.,   "Short-Term  Bioassay  of  Complex  Organic  Mixtures,  Part  I,
     Chemistry,"   in   Application   of   Short-Term   Bioassays   in   the
     Fractionation and Analysis of Complex Environmental Mixtures, M.  D.
     Waters,  S. Nesnow,  J.  L.  Huisingh,  S.  S.  Sandhu,  and  L.  Claxton,
     Eds., Environ. Science Res.,  15,  247-268,  Plenum  Press,  New  York
     (1979).

11.   Guerin,  M. R., Ho,  C.-h.,  Rao,  T.  K. , Clark,  B.  R. ,  and Epler,  J.
     L.,   "Polycyclic  Aromatic  Primary  Amines  As  Determinant  Chemical
     Mutagens in Petroleum Substitutes," Environ.  Res,  (submitted).

12.   Ho,  C.-h., Clark, B. R. , Guerin, M.  R. , Ma,  C. Y. ,  and Rao, T.  K. ,
     "Aromatic Nitrogen  Compounds  in  Fossil Fuels—A  Potential  Hazard?"
     in  ACS,   Div.  of  Fuel  Chem. ,   24,  281-291,  Preprints  of Papers
     Presented at  Honolulu,  HI  (1979).

13.   Ho,  C.-h., Ma, C. Y. ,  Clark,  B.  R.,  and Guerin,  M. R.,  "Separation
     of  Neutral  Nitrogen  Compounds  from  Synthetic  Crude  Oils   for
     Biological Testing," Environ.  Res,  (in press).

14.   Stanton,   M.   F.,  Miller,  E.,  Wrench,  C.,   and  Blackwell,   R.,
     "Experimental Induction of Epidermoid Carcinoma in the Lungs of  Rats
     by Cigarette  Smoke Condensate,"  JNCI, 49,  867-873 (1972).

15.   Rubin,  I.  B.  and Guerin,  M. R.,  "Chemical Evaluation of  the Beeswax
     Pellet   Implantation Bioassay  Model  for  Studies  of  Environmental
     Carcinogens," JNCI,  58, 641  (1977).

-------
                   TABLE 1




Disciplinary Approaches to Program Objectives

Characteristic
Interaction
Focus
Product
Efficiency
Limitation

Independent
None-Little
Discipline
Knowledge,
Methods
Poor
Creativity
Fund ing
Approaches
Interactive
Little-Much
Discipline-
Project
Data-Informat ion-
Knowledge
Fair-Very Good
Funding

Integrated
Much-Total
Project-
Mult id is cipline
Knowledge, Methods
Poor-Excellent
Funding
Creativity
                     10

-------
                                TABLE 2




                  Paraho/SOHIO Shale Oil Matrix Study
Participants a, D
4-1
C -H
O >, >, O
•i-l 4J 4-1-1-1
4-1 -r-l CO ->-l C C
CO r-( O i-IO -HO) S E
r-l C CO -H -H -r-l ^ 00 >> dl 0)
coo -i-i c jcc en o 4-> u 4->
O-H r-i a) cxai C 4-1 T-I co a) co
•i-l 4J 0)00 OOO OJ-i-l CO >-,C>->
EO 4-1CO COCO COO 0) -rt T3 tfl -r-l CO
CUCO O4-I O4J 3(-l TJXCOMO
v ,C!-i C03 M3 Ocfl OOOOCOo
Material0 u^ « S OS So PSHPHWSW
Crude Shale Oil 5 1,2,5 9,10 1,5,8 55-
HDT Shale Oil 5 1,2,5 9 1,5,8 5 5
HDT Residue 5 1,5 9 1,5,8 5 5
Jet Fuel-Shale 5 1,5 9,10 1,5,8 - 5 -
Diesel Fuel Marine
(DFM) 5 1,5 9 1,5,8 - 5 4
Petroleum Jet Fuel 5 1,5 9,10 1,5 - 5
Petroleum DFM 5 1,5 9 8 5 4
al . DOE-Los Alamos Scientific Laboratory
2. DOE-Lawrence Livermore Laboratory
3. EPA-Research Triangle Park
4. EPA -Gulf Breeze Laboratory
5. DOE/EPA-Oak Ridge National Laboratory
6. API-Union Oil
7. API-Exxon
8. EPA-Kettering Laboratory
9. Brown University
10. University of Texas

Chemical
Analysis
1,5-7
1,5-7
1,5-7
1,5,6

1,5-7
1,5
1,5



"Partial listing.
                                 11

-------
                                    TABLE 3

               Synfuels Research Materials Facility Components3
    Materials
             a,b
  Endpoints
         Institutions
Coal Liquefaction
   Crude Product
   Process Stream
   Refined Product
   Solid Residue

Coal Gasification
   Aqueous
   Solid Residue

Coal Combustion
   Ash

Shale Liquefaction
   Crude Product
   Solid Residue
   Refined Product
   Refinery  Residue
   Aqueous

Petroleum
   Crude Oil
   Refined Product
Chemistry
  BaP,  PAH
  Class
  Physical

Mutagenicity
  Bacterial
  Drosophila
  Mammalian

Tumorigenicity
  Skin
  Intra-tracheal

Toxioity
  Mammalian
  Aquatic
  Marine

Teratogen-Lcity
Oak Ridge National Lab
Los Alamos Scientific Lab
Lawrence Livermore Lab
Pacific Northwest Lab
EPA-Gulf Breeze
EPA-Research Triangle Park
Univeristy of Texas
Brown University
University of Tennessee
University of Connecticut
Exxon Corporation
Gulf Oil
Union Oil Company
Mobil Oil Corporation
Kettering Laboratory
National Bureau of
  Standards
U.S. Department of Navy
Laramie Energy Technology
  Center
Bartlesville Energy
  Technology Center
Pittsburgh Energy
  Technology Center
Electric Power Research
  Institute
Hydrocarbon Research, Inc.
 aPartial  listing.

 ^Contributing  Technologies:   H-Coal,  Solvent  Refined  Coal,  Paraho  Shale,
   Synthoil,  COED,  Zinc  Halide,  LETC  Simulated In-Situ Shale
                                       12

-------
                               TABLE 4

           Weight Percent Distribution by Chemical Fraction
Weight Percent
Material V I ESA IA ESB IB
Petroleum Crude
5301 7.2 - 0.3 2.2 0.2 0.6
Petroleum Crude
5305 19.6 - 0.2 <0.1 0.6
Shale Crude 4601 0.3 - 0.4 0.3 3.3 0.5
Coal ASOH 1312 34.7 - 0.9 - 1.1 0.5
Coal ASB 1313 0.8 - 0.6 0.1 2.1
Coal VOSH 1314 0.3 - 1.2 0.3 2.6
Coal VSB 1315 - 59.5 0.2 0.8 1.8 9.6
Coal Dist 1601 4.2 - 1.4 0.1 1.5 0.1
Coal Dist HDT/H
1604 12.8 - <0.1 1.1 0.3 1.0
V = volatiles
I = ether insoluble
ESA = ether soluble acids
IA = insoluble acids
ESB = ether soluble bases
IB = insoluble bases
Ng = neutral saturates
N N N- N . T
S Ar Ply Pol

55.7 16.5 9.8 6.2 99

51.0 7.4 3.9 1.1 84
55.6 27.4 7.0 2.2 97
11.9 24.6 0.5 <0.1 75
22.9 55.8 6.9 <0.1 89
13.9 58.5 19.7 <0.1 97
1.3 8.0 18.4 0.3 100
12.7 48.9 6.3 0.2 76

18.5 42.3 1.4 - 77






    neutral aromatics
    neutral polycyclic aromatics
    neutral polar fraction
T = total recovery

-------
                                 TABLE 5



                     Mutagenicity of Class Fractions
Material
         Revertants per Microgram



I   ESA   IA  ESB   IB     N,,
N    N    N
 Ar   Ply  Pol
Petroleum Crude
5301
Petroleum Crude
5305
Shale Crude 4601
Coal ASOH 1312
Coal ASB 1313
Coal VOSH 1314
Coal VSB 1315
Coal Dist 1601
Coal Dist HDT/H
1604

0

0
0
0
0
0
-
0

0

0

0
0
0
0
0
0.2 3
0

0

0

0
0
0
0
0
0
0

0

0 0

0
8 0.1
0 0
7 1.7
40 0.1
300 3
9 0

0 0

0

0
X
0
0
X
X
0

0

X

X
0
0
0
0
0.1
0

0

X

X
1.6
0
0
16
3
4

0

0

0
0
0
0
0
0.7
0

0
  aSee Table 4 for definitions.



    "x" = <0.1
                                   14

-------

















co
r-l
•r-l
O

M-l
v£> O
W >>
r-? 4J
PQ *i-|
«U 0
H T-I
Q)
M
CO
»
























bC
^ e

4J >
o 







i— 1
CO
• r-l
l-i
0)
4-1
CO
^





in ^oooocooi^^o o
00 ^d" i — 1 i— 4 vj-
CO i— 1 CM CO CO
-d- vo

O OoOoOoO O

ro ocNOomcTit'i o
co in CM r*~ vo

vO OOOoOOO O
vO m

o OCMOOOOO o






O O^nOrnO<)-O O
V


o oo\OcTNininr^ o
vo CTN CM oo rn


o ooooooo o


O OOOOOOO 1

1 1 1 1 1 1 CM 1 1


O OOOOOOO O



1 — 1
0) 01 O S3
TJ T3 ^ CM -
-------
Q
 I
_J
Z
cc
o
^

(-
u
Z)
Q
O
CC
a.
cc
O
_J
O
UJ
Q

OC
O
^^«
Z
O
1-
cc
O
a.

>
LU
	 CC
0
z
0
t-
<
_ I
—
1 —
co
Q
    I-
    oc

    D.

— 9-
    U

    DC
    LU
    I
                                                                    o
                                                                    D_
                                                                     O
I
Q.
<
oc
a
o
                                                                            o
                                                                            cc
                                                                            I
                                                                                      O
                                                   O
                                                   CC
                                                   <
                                                   LU
                                                   OC
                                                   O
                  I
                  a.
                  <
                  oc
                  O
                  O
O
oc
I
O
                                                                                                        O
                                                                                                        z
                                                                                                        oc
                                O
                                >
                                                          z
                                                          o
                                                          co

                                                          CD

                                                          cr
                                                          LU
                                                          X
                                                                            cc x    |—
                                                                            I- O
                                                                            X oc
                                                                               O
                                                                     l/l
                                                                     Q
                          X
                          0_-

                          QC



                         I
                          O
                          oc-
                          I
                          o
                                                                                          X
                                                                                          Q_
                                                        o
                                                        o
                                                        o
                                                        oc
                                                                                                                                    LU
                                                                                                                                    I
                                                        CD
                                                        CC
                                                        <
                                                        CJ
                                                        o
                                                        CC
                                                        Q
E

J
O
                                                                          1
                                                                          &
                                                                          a
                                                                          d>
                                                                          CO
                                                                                                                                                     1
                                                                                                                                                      CO
                                                                                                                                                     "5
                                                                                                                                                      O
                                                                                                                                                     O
                                                                                                                                                     "o
                                                                                                                                                     S
                                                                                                                                                      £
                                                                                                                                                      c
                                                                                                                                                      o
                                                                                                                                                     o
                                                                                                                                                      o
                                                                                                                                                     'E
                                                                                                                   a.

                                                                                                                   .1

                                                                                                                    c
                                                                                                                    g
                                                                            16

-------
   DESCRIPTION OF BIOASSAY RESULTS AND PROJECTION OF BIOASSAY STUDY
             NEEDS IN SUPPORT OF A MAJOR SYNFUELS INDUSTRY

       A. R. Kolber, M. B. Wilkie, T. J. Wolff and D. G. Nichols
                      Research Triangle Institute
                   Research Triangle Park, NC  27709
                               ABSTRACT

Various coals, including peat, lignites, bituminous coals, anthracites,
as well as oil shales and tar sands are candidates for conversion to
synthetic fuels and/or chemical feedstocks.  We have tested the biotoxicity
of effluents from a fixed-bed, air-blown, coal gasifier producing 75
ft3/hr of low BTU gas (200 BTUs/ft3) from 3 Ibs/hr. raw coal, and gener-
ating 0.1 Ib/hr crude tar effluent.

Four coal types were tested in the gasifier, the crude tars collected,
fractionated into chemical classes using a solvent partioning procedure
and the fractions analyzed by GC/MS.  The crude tars and their chemical
fractions were subjected to mutagenicity and cytotoxicity testing using
a modification of the (Ames) Salmonella/his  and Chinese Hamster Ovary
Cell bioassays, respectively.

The integrated Chemical-Analytical and Biological Testing approach yielded
the following information.  Whereas the crude tars for certain coal types
(Wyoming sub-bituminous and North Dakota Lignite) exhibited borderline or
no mutagenicity and cytotoxicity, some of the chemical fractions (organic
bases and PNAs) which were present as minority components by percent toal
organic mass, exhibited substantial mutagenicity and cytotoxicity.  In
nearly every case, those crude tars and chemical fractions which were
mutagenic, were also cytotoxic, and vice versa.  Biotesting of crude
gasifier tars (and complex organic mixtures in general) may not be suffi-
cient when these mixtures contain highly bioactive minor substituents, or
when antagonistic reactions within the mixture mask actual mutagenicity
and/or toxicity.

In addition to the specific data presented, this paper summarizes recent
progress in sampling, chemical analysis, and bioassay methodology develop-
ment for synthetic fuel process effluents (complex mixtures).  The various
streams and effluents of environmental significance are characterized,
specifically regarding potentially hazardous (toxic and/or mutagenic)
substances.  Bioassay tests are specified for various types of samples.
                                     17

-------
                             INTRODUCTION

Until about 1910, nearly every major eastern U.S. city had its gashouse,
where gas for lighting and cooking was produced from coal by distillation.
This formerly widespread technology lost its markets to natural gas and
petroleum, but now, with both natural gas and petroleum oil increasingly
in short supply, the production of gas from coal is being reexamined as
a viable alternative for our energy requirements.  Coal is found in 30
of the 50 states, and represents more than 90 percent of the proved
reserve of all developed fuels.  At the present rate of consumption, if
coal were utilized for one-half our present energy requirements, the
proved reserve would last about 150 years.

There are more than one hundred different theoretical schemes for coal
gasifiers and gas producers (16,7), but none of these will produce a gas
product with a heating value equivalent to natural gas (1030 BTU/ft3).
Depending on the process, coal gasification produces a gas product with
a heating value from about 125 BTU/ft3 (low BTU) to about 560 BTU/ft3
(high BTU).  This gas has mainly industrial uses, as synthesis gas for
the manufacture of ammonia, plastics, and synthetic fuels, and, when
locally produced, as a boiler fuel.  This application is especially
important since more than 30 percent of the U.S. production of natural
gas is consumed by industry (1979 estimate), (12).  More important, a
high hydrogen-content gas (synthesis gas) produced by coal reaction will
be used as the substrate for synthetic liquid fuel (coal liquifaction).
High BTU gas which could substitute for natural gas can be produced by
methanation of synthesis gas using nickel catalysis.  A block diagram
illustrating the overall processes involved is shown in Figure 1.

Developments in coal gasification technology must take into account the
potential hazard to human health and to the environment presented by
fugitive emission and toxic byproducts and effluents.  Coal is a complex
carbonaceous material, containing sulfur and nitrogen in addition to
carbon and oxygen.  Theories of coal structure (16) produce models such
as that shown in Figure 2.  The four roost advanced, or most commonly
used gasifier processes each produce tars,  oil and/or ash and contaminated
quench water as environmental effluents, and considering that 10,000
tons or more coal may be reacted per day in a commercial gasifier, as
much as 7-8 tons per day of potentially hazardous waste products must be
disposed of.  Table 1 illustrates the commercial gasifiers and their
operating data, including effluent streams.

We report in this study the chemical characterization of and quanti-
fication of biotoxicity of effluents from an experimental coal gasifier
reacting various coals with a fixed mode of operation, using organic
extraction, chemical fractionation and analyses (with gas chromatograph/
mass spectra) combined with concurrent short-term in vitro toxicity and
mutagenicity bioassays.
                                    18

-------
                                METHODS

Description of Experimental Coal Gasifier. The effluents tested in our
laboratory were produced by a fixed-bed, air blown gasifier producing 75
fts/hr of low (200 BTU/ft3) BTU gas from 3 Ibs/hr. raw powdered coal,
generating 0.1 Ib/hr crude tar effluent, representing 0.3% by mass of the
input reactant.  The gasifier is so constructed as to permit variable
modes of operation (temperature, pressure, air-oxygen mixture, etc.).
The reactor is fitted with sampling ports for fugitive emissions and
product effluent traps.  A flow diagram of the reactor is given in Figure
3.  The reaction of four coal types was investigated:  Western Kentucky
and Illinois (Eastern coals), and Wyoming sub-bituminous and North Dakota
lignite (Western coals).

Chemical Extraction and Fractionation. The crude tar effluent collected
from the trap of the quench/condensation device is the major effluent
stream.  The crude tar is extracted with methylene chloride and fraction-
ated by an aqueous acid-base partitioning scheme modified from Novotny et
al. (9) by Pellizzari, et al., (10) to yield five organic classes:
organic acids, bases, polar neutrals (PN), non-polar neutrals (NPN), and
polynuclear aromatic hydrocarbons (PNA).  There is usually a 15 percent
loss of material during the fractionation.  This is a partitioning scheme
which is designed more for efficient partitioning and minimal spillover
of compounds from one chemical class to another than for maximal recovery.
To monitor chemical class overlap, the complex mixtures can be spiked
with deuterated standards representative of compounds in each chemical
class, and these compounds later identified in gas chromatograph/mass
spectra scans of the partitioned classes.

Characterization of the constituents of the five chemical classes, and
identification of known toxic substances present was performed by gas
chromatograph-mass spectra analysis.  A list of representative compounds
identified in the chemical classes is shown in Table 2.  A summary
diagram for this process is given in Figure 4; a detailed description is
as follows.
                                                                         (R
The methylene chloride extract of crude tar was filtered through a Teflon
filter 10.5 |J pore size).  The filtrate was evaporated to dryness using a
Rotovap , the solids retained by the filter taken up in 100 ml methanol
(B-J), filtered through the Teflon filter, and the filtrate taken to
dryness.  The filtrates were combined, dried, weighed and redissolved in
CH2C12, the solution spiked with internal standards to allow for quantifi-
cation and to serve as a quality control parameter for the overall parti-
tion/analysis scheme.  The standards used were qunioline-d7 (organic
base), phenol-d5 (organic acid), dodecane-d26 (non-polar neutral), and
anthracene-dlO (PNA).

This solution was extracted twice with equal volumes of 10% H2S04,
followed by 20% H2S04.  The aqueous phases were combined and washed with
CH2C12, which was combined with the original CH2C12 phase.  The aqueous
                                    19

-------
phase was cooled (ice bath) adjusted with 25% NaOH to pH 10, and extracted
three times with CH2C12 to generate the "organic bases."  The aqueous
phase was discarded.  The original CH2Cl2 phase was extracted three times
with 5% NaOH, the aqueous phases combined and washed with CH2Cl2-  The
CH2C12 phase was combined with the original CH2C12 phase.  The NaOH
phases were placed in an ice bath and acidified to pH 3 with 20% H2C04,
and extracted three-times with CH2C12 to generate "organic acids."  The
remaining aqueous phase was discarded.

The original CH2C12 phase was taken to dryness, reconstituted in cyclo-
hexane, and filtered through a Teflon  filter (0.5 Mm)•  The cyclohexane
filtrate was extracted three times with an equal volume of methanol: H20
(4:1) solution, the methanol:H20 extract concentrated, and the H20
removed by freezing-drying to generate "polar neutrals."  The cyclohexane
phase was extracted three times with an equal volume of nitromethane, the
nitromethane phases combined and evaporated to dryness to generate the
PNA's."  The cyclohexane phase was taken to dryness to generate the "non-
polar neutrals" fraction.

The partition scheme has been validated for efficiency and presence of
spillover of chemicals from one class to the other using mixtures of
deuterated known substances representative of the individual chemical
classes (as described above).  Known mutagens were included in the mixture
and the recovery of mutagenic activity during fractionation was monitored
by bacterial mutagenicity bioassay.

Bioassay

Tissue culture medium and serum was obtained from Gibco Corporation (New
York) and KC Biologicals, Kansas City.  DMSO was purchased from Fisher
Scientific.  Bacto-Agar and bacterial nutrient medium was obtained from
Difco Corp.  NADPH  (tetrasodium salt, Type 1) and known positive mutagens
(highest purity available) were obtained from Sigma Chemical Company.

Bacterial Mutagenicity. Raw coals, crude reactor tars, and chemical
fractions derived from the tars by the partitioning scheme were dissolved
or suspended in dimethylsulfoxide  (DMSO) and submitted to mutagenicity
bioassay using a modification of the Ames/Salmonella bacterial system
employing Aroclor-induced rat liver microsomal (metabolic) activation,
and bacterial strains TA98 and TA100.  Strain TA100 detects base-substitu-
tion and strain TA98 detects frame shift mutagens.

Ames Salmonella strains were cultured overnight in Oxoid nutrient broth
to stationary phase.  The suspension was diluted to an optical density
equivalent to 3.3 x 108 cells/ml with saline.  For the mutagenicity
assay, VL08 cells in 0.3 ml saline were added to 0.3 ml sample, 0.3 ml
NADPH or water and  0.3 ml 0.25 M sucrose or rat microsomal activation
preparation.  For quantitative plate incorporation studies, 2.5 ml molten
agar (45°C) was added to the bacteria and sample, the mixture poured onto
a 25 ml agar base layer  (1.5% Difco Bacto-agar in Vogel-Bonner medium E),
and incubated at 37°C for 72 hours.  For bacterial toxicity assay, the
bacterial suspension was diluted serially in saline to 1000 cells/ml, and
200-600 cells plated in a 2.5 ml molten agar overlay.  The assay protocol
followed was consistent with the suggestions of DeSerres and Shelby (2),

                                    20

-------
including assay of positive mutagen and solvent controls in each experi-
ment.  Triplicate plates were routinely run.

Preparation of S-9 Rat Liver Activation System. The procedure of Ames et
al. (1) was used.  Male Charles River rats, strain CDx (200 g) served as
the source of liver material.  The animals were housed one week before
initiation of induction.  Food and water were given ad libitum.  Induced
rat liver (minimum of three rats) was obtained from rats injected intra-
peritoneally on day one with 500 mg/kg of Aroclor 1254 in Corn Oil (0.5
ml of a 200 mg/ml for a 200 g rat).  On day five, the animals were sacri-
ficed.  The liver was removed, washed twice in cold, 1.15% KC1, blotted
dry with sterile paper, weighed, minced, brought to 200 mg/ml protein
with 0.25 M sucrose, and homogenized with four strokes of a cold Potter-
Elvenhjem apparatus with a Teflon  pestle.  The homogenate was centrifuged
for ten minutes at 9,000 x g, the lipoprotein layer aspirated off, and
the protein concentration of the supernatant adjusted to 30 mg protein/ml
with 0.25 M sucrose.  The stock solution was checked for sterility,
quick-frozen in small aliquots (2-5 ml), and stored for a maximum of two
months at -80°C.  For the assay, an aliquot of stock solution was slow-
thawed and adjusted to an appropriate concentration of protein with 0.25
M sucrose.  NADPH was added at 0.32 rag/plate.  Each batch was checked
against known positive controls using the plate incorporation assay.   S.
typhimurium strains TA98 and TA100 were obtained from Dr. Bruce Ames,
(Biochemistry Dept. University of California at Berkeley).  NADPH and
standard mutagens were obtained from Sigma Chemical Co. (St. Louis).

Chinese Hamster Cell Cytotoxicity. General cytotoxicity was studied in
vitro using Chinese Hamster Ovary (CHO) cells in tissue culture.  CHO
cells, clone Kl, were obtained from the American Type Culture Association
at passage 5, passaged once in our laboratory, frozen in 10% DMSO, and
stored at -70°C.  Cytotoxicity studies were performed between passage 7
and 15.  Cells were maintained in Ham's F12 medium supplemented with 10%
calf serum, were passaged at 1:20 dilution (using 0.01% trypsin) and
incubated at 37°C in a 5% C02-air atmosphere.  Cells were split at conflu-
ence (about 4 days).  For cytotoxicity studies, 10s cells were explanted
into 35 mm culture dishes (Corning) with 2 ml medium and allowed 24 hours
for attachment.  The sample was added in 25 pi DMSO, incubated with the
cells for 24 hours, the medium was replaced, and the cells incubated
further.  At 24 hour intervals, the monolayers from triplicate plates
were trypsinized and the cells counted.  The results were expressed as
the fraction inhibition of normal cell growth during the logarithmic
phase produced by the toxic material.  The effects of raw coals and
gasifier tars was compared in each experiment to the growth inhibition
produced by a standard cytotoxin, CdCl2 (illustrated by Figure 5).
Clonal growth experiments were performed by explanting 200-2000 cells per
35 mm culture dish in 3 (Jl medium, incubating 24 hrs.  for attachment,
adding the sample in 25 (Jl DMSO, incubating 24 hrs for exposure, replacing
the medium and incubating 7-10 days.  When the colony size reached about
100-200 cells, the medium was discarded, the colonies stained with methyl-
ene blue and counted.  The results are expressed as fraction reduction of
colonies grown in a control medium.
                                    21

-------
                                RESULTS

Mutagenicity. The raw coals were non-mutagenic to doses exceeding 10
mg/plate.  The crude reactor tars generated from the Western coals (Wyoming
subbituminous and North Dakota lignite) were weakly mutagenic; the crude
tars from the Eastern coals (Western Kentucky and Illinois) were strongly
mutagenic.  Mutagenicity in each case required metabolic activation and
was confined to frame-shift mutagenesis (active with strain TA98 only).
The dose-responsive nature of the mutagenic response and the relative
toxicity of the crude tars is shown in Figure 6.  Of the chemical classes
tested, the organic base was generally the most mutagenic; in some cases
generating in excess of 30 revertants/per microgram.  The mutagenicity
profiles were generally reproducible for separate gasifier reactions of
the same coal.  This is illustrated by Table 3, which summarizes the
crude tar and chemical class mutagenicity for each coal reacted, and by
Figure 7.  A dose-response mutagenicity profile for the strongly active
base fraction is shown in Figure 8.  Inspection of Table 3 indicates that
for certain coals (Western co^ks in this case) the crude tar is weakly
mutagenic.  However, some chemical classes derived from the crude tar,
which represent perhaps only a small fraction by mass of the tar, (the
base fraction represents ca. 4% of the tar mass) are very strongly
mutagenic when tested separately.

Cytotoxicity. The raw coal dusts were non-cytotoxic to doses as high as
10 mg/dish, although the cells seemed to ingest the dust particles, the
cell cytoplasm appearing filled with particles after 24 hours.  The CHO
cytotoxic activity of the crude tars was similar to mutagenic activity.
Those crude tar extracts which were strongly mutagenic (Eastern coals)
were also cytotoxic, those tar extracts which were weakly mutagenic were
also weakly cytotoxic.  The basic organic fractions derived from the tar
extracts were strongly cytotoxic, and in this regard, the cytotoxicity
results were also similar to the mutagenicity results; those fractions
which were strongly cytotoxic were also mutagenic.  Table 4 summarizes
the cytotoxicity data.
                                     22

-------
                              DISCUSSION

This study investigated the toxic potential of coal gasification byproducts
and effluents.  It appears that simple bioassay of crude complex mixtures,
such as effluent tar and oils may not provide an accurate estimation of
potential health risk.  This conclusion is clearly illustrated by inspec-
tion of Table 3, where it can be seen that testing of the crude tar
generated from the reaction of Western coals would not reveal the potent
mutagenicity and cytotoxicity of the organic base fractions.  The toxicity
of chemical classes representing a small fraction by mass of the crude
tar, such as the organic bases, would be masked by the greater mass of
non-toxic material present.  We conclude that fractionation of such
complex organic mixtures into chemical classes be performed, and the
fractions then submitted to bioassay.  The fact that these toxic materials
are present in small mass fractions does not preclude their potential
environmental impact, or human health risk, since commercial coal gasifi-
cation plants reacting 10,000 tons of coal per day will produce about 2-4
tons/day of the strongly toxic and/or mutagenic base fraction alone.

Synthetic fuels systems and effluents having a history of biotesting
include fluidized-bed combustion, fixed-bed coal gasification, and shale
oil conversion.  More toxicological research has been conducted on
fugitive emissions and process effluents than on the commercial products
themselves.  Liquid fuels produced by methanation of product gas have
been studied by Epler, et al., (3).  Synfuels produced from shale-oil
conversion by either surface or undergound retorting have been recently
studied for their toxic potential, (3).  Although end-organ toxicity,
(e.g., neuro, pulmonary, renal, hepatic, etc.), is certainly of great
concern, attention has been given mainly to carcinogenic/mutagenic risk.
This emphasis is attributable to the relatively sophisticated bioassay
technology available for carcinogenesis/mutagenesis determination.

For complex environmental mixtures, methods for chemical characterization
are vastly better developed than for biological testing.  Chemical
analysis of synthetic fuels, byproducts and effluent streams has been
performed in detail (4,13).  Analytical chemical analyses include proce-
dures for gas, semivolatiles and volatiles, for fugitive and stack emis-
sions, and chemical extraction and partitioning schemes have been developed
for tars and oils (14,6,10,9).  Similar acid-base extractive chemical
partitioning schemes have been developed for complex organic mixtures
such as ambient air and cigarette-smoke particulate (8).  More recently,
gentler chemical fractionation methods using gel chromatographic procedures
for separating non-polar constituents in complex environmental mixtures
have been developed for shale-oil (6,4), and have been suggested for
other complex organic mixtures to avoid artifact generation by treatment
with strong acids and bases.  "Class" fractionation procedures all have
some overlap of compounds, or "spillover" of certain compounds from one
class to another, but experimentation with the fractionation schemes
illustrated has minimized this overlap.  Chemical identification in all
cases has been accomplished with gas chromatographic retention and mass
spectra (4,10).  Inspection of chemical analytical results helps to
identify known mutagenic and/or carcinogenic compounds in complex environ-
mental mixtures.


                                   23

-------
Epler e_t al. (3) and Guerin e_t al. (4) have published evidence that
shale-derived oil contained chemical classes which exhibited mutagenicity
in bacterial bioassay.  Nitrogen-heterocycles appeared to be responsible
(3), the organic base fraction having the greatest activity.  As for the
study reported here, mutagenic activity required metabolic activation and
was confined to Ames strain TA98, indicating that frame-shift mutation
was predominant.  These authors found that chemical class fractions
derived from natural petroleum crude oil were very weakly active; less
than 0.1 revertant/pg as compared to 30-35 revertants/pg for the same
chemical class derived from synfuels from shale-oil or coal gasification.
Pelroy and Wilson (11) performed chemical-class fractionation by solvent
extraction of high-boiling coal liquids, and the mutagens detectable in
the Ames assay system were concentrated in the organic base fractions,
which contained nitrogen-heterocycles.  Aromatic amines were concentrated
in the fractions of a TLC eluate having the greatest mutagenicity in the
Ames assay system.  Timourian et al. (15) have chemically fractionated
effluents and products from underground coal gasification and oil shale
processing.  Bioassay included mutagenicity in bacteria, toxicity and
mutagenicity in different mammalian cultured cells, induction of sister
chromatid exchanges (SCE) in vivo and in vitro, and germ cell toxicity in
vivo.  Again, in this case, organic bases were mutagenic for TA98, although
the organic neutrals were also effective mutagens.  Crude shale oil is
highly mutagenic for bacteria, but only slightly mutagenic for Chinese
Hamster Ovary cells.  In vivo, crude shale oil is negative for SCE induc-
tion.

Hsie, et al. (5) have investigated the mutagenicity of the basic (N-
heterocycle-containing) fraction of Synfuel A (prepared by the method of
Epler, et al.,)(3) in the Chinese Hamster Ovary cell forward mutation
system of Hsie, et al. (5).  The acetone-extracted ether-soluble base
fraction of Synfuel A, which was shown to be the most mutagenic fraction
for bacterial bioassay (3) also exhibited the greatest mutagenicity to
Chinese Hamster Ovary cells.  This acetone-extracted fraction has been
shown by Epler et al., (3) to contain higher molecular weight azaarenes,
benzacridines and azabenzpyrenes, as compared to weakly mutagenic benzene
and isopropanol extracts of the same shale-derived synfuels.
                                     24

-------
                            RECOMMENDATIONS

Coal gasification and other forms of synthetic fuels production may well
provide at least partial relief for North America's energy problems.
But, in light of the results of preliminary toxicity studies environmental
and health impact from these industrial processes must not be ignored.
This is especially so at a stage of our society when both industrial and
population density is high.

We would like to recommend a battery of bioassays to be conducted on
synfuels products, byproducts and effluents, including fugitive emissions.
Young, et al. have described possible occupational hazards associated
with coal gasification.  These include handling possible wastes and
exposure to fugitive emissions, which might include vapor-phase mutagens.
The testing battery is designed to be rapid, informative, and cost-
efficient, and for this reason will consist of in vitro tests.   The
testing is divided into health and eco-effects, and the battery paradigm
is designed for the mode of exposure of the environmental stream.  Table 5
illustrates a sample battery testing program which includes mutagenicity,
DNA damage (clastogenic effects), and cytotoxicity.  In vitro battery
biotesting is a screening procedure designed to identify potential health
risk, but must be substantiated by more complete testing, including skin
painting and whole animal carcinogenicity testing, tumor production, and
chronic end-organ toxic effects.  Other sample battery testing protocols
for synthetic fuels byproducts have been suggested, notably by Epler, et
al., (3).  It is further recommended that chemical fractionation and
analysis be interfaced with biotesting when environmental complex mixtures
are investigated.
                                    25

-------
                             REFERENCES
 1.   Ames,  B.  W.,  J. McCann,  and E. Yamasaki, "Methods for detecting
     carcinogens and mutagens with the Salmonella/mammalian microsome
     mutagenicity  test," Mutation Res., 31:347-364  (1975).

 2.   DeSerres, F.,  and M.  Shelby, "The Salmonella Mutagenicity Assay:
     Recommendations," Science, 203:563-565, 1979.

 3.   Epler,  J.,  B.  Clark,  C.-h. Ho, M. Guerin, and  T. Rao, "Short-term
     bioassay  of complex organic mixtures:  Part II, Mutagenicity Testing."
     In:  Application of short-term Bioassays in the Fractionation and
     analysis  of complex environmental mixtures, EPA publication EPA-
     600/9-78-027,  pp 24-269, 1978.

 4.   Guerin, M., B. R. Clark, C.-h. Ho, J. Epler and T. Rao,  "Short-term
     bioassay  of complex organic mixtures; Part I,  Chemistry."  In:
     Application of short-term bioassays  in the fractionation and analysis
     of complex  environmental mixtures, EPA publication EPA-600/9-78-027,
     pp 24-269,  1978.

 5.   Hsie,  A., P.  O'Neill, J., San Sabastian, D. B. Couch, "Quantitative
     Mammalian cell genetic toxicology:   study of the cytotoxicity and
     mutagenicity  of seventy individual environmental agents  related to
     energy technologies,  and three subtractions of a crude synthetic oil
     in CHO/HGPRT  system," In:  Application of short-term bioassays in
     the fractionation and analysis of complex environmental  mixtures,
     EPA publication EPA-600/9-78-027, pp 293-315,  1978.

 6.   Jones,  A.,  M.  Guerin, and B. Clark,  "Preparative-scale liquid chroma-
     tographic fractionation of crude  oils derived  from coal  and  shale,"
     Anal.  Chem.,  49:1766-1771, 1977.

 7.   Kent,  J.  A.,  Riegel's Handbook of Industrial Chemistry,  Van Nostrand
     Reinhold  Co.,  New York, p 45  (1974).

 8.   Lofroth,  G.,  "Mutagenicity assay  of  combustion emissions," Chemosphere,
     7:791-798,  1978.

 9.   Novotny,  M.,  M. Lee,  and K. Bartle,  "The methods for  fractionation,
     analytical  separation and identification of polynuclear  aromatic
     hydrocarbons  in complex mixtures," J. Chromatog. Sci., 12:606-612,
     1974.

10.   Pellizzari, E. D.,  L. Little, C.  Sparacino, T. Hughes, L. Claxton
     and M.  Waters, "Integrating microbiological and  chemical testing
     into the  screening  of air samples for potential  mutagenicity," In:
     Application of short-term bioassays  in the fractionation and analysis
     of complex  enviornmental mixtures, EPA publication EPA-600/9-78-027,
     pp 333-351,  1978.
                                    26

-------
11.  Pelroy, R. ,  and B.  Wilson, "Mutagenicity analysis of coal liquids,"
     In:  Proc. Second Symposium on applications of short-term bioassays
     in the analysis of complex environmental mixtures,  Abstract (avail-
     able from Dr.  M.  Waters, U.S.  EPA,  HERL, Research Triangle Park,  NC
     27709).

12.  Perry, H., "The Gasification of Coal," Scientific American, 230:3,
     1974.

13.  Rubin, I. B.,  M.  Guerin, A. Hardigree and J.  Epler, "Fractionation
     of synthetic crude oils from coal for biological testing," Envir. Res.,
     12:358-365,  1976.

14.  Swain, A., T.  Cooper,  and R. Stedman, "Large  scale fractionation  of
     cigarette smoke condensate for chemical and biologic investigations,"
     Cancer Res.. 29:579-583, 1969.

15.  Timourian, H.,  J. Felton, D. Stuermer, F.  Hatch, S. Healy, P.
     Berry, A. Carrano,  L.  Thompson, and J. Carver, "Comparative Biolo-
     gical Activity of complex effluents and products from coal gasifi-
     cation and oil shale processes," In:   Proc. Second Symposium on
     applications of short-term bioasssays in the  analysis of complex
     environmental  mixtures, (Abst.) (available from Dr. M.  Waters,  U.S.
     EPA, HERL, Research Triangle Park,  NC  27709).

16.  Young, R., W.  McKay, and J. Evans,  "Coal gasification and occupational
     health," Amer.  Indus.  Hyg. Ass. Jour., 39:985-997,  1978.
                                    27

-------
   HIGH B.I.U, GAS
                  ttVOATIZATIW
                      cm.
FIGURE 1    Steps  in Coal Gasification
                         28

-------
29

-------
%
X
<
Ul
"\
K
lAnmstuiit RESUIIUO»
1

F1 !
_/
=i:r3" 5 =
\
i=
o
o i
. 
-------
                                Extracted  Tar Organics
                                             I WASH 3X WITH 1M NiOH
                             CH2CI2 LAYER ........ N»OH LAYERS
                                   WASH 2X
                                   WITH H2O
                                                    1 WASH WITH CH2CI2
                                            CH2CI2 LAYER
                       CH?Cl  LAYER"-; H20 LAYER
                         2            (pfi - 10)
              WASH 2X
              WITH 10%

              1XWITH   j
              MKHjSQ,  if
               NaOH LAYER

                   II
                   I


WITH   JORGANtC ACIDS!
                                                   ADJUST TO pH 2 WITH 6N HCI
                                                   EXTRACT WITH CH2CI2
                                                         —
                                         WASH
                                         CYCLOHEXANE
                  I               I
             CHjClj LAYER	 HjSC^ LAYER
                                    CYCLOHEXANE LAYER  HjO LAYER
                                    (II EVAPORATE TO       I
                                      DRYNESS             ADJUST TO pH 2 WITH 1N HCI
                                    (2) DISSOLVE IN CH2CI2    | EXTRACT 3X WITH CH2Cl2
                                                  I ORGANIC ACIDS!
                                                                     AQ
                   WASH 2X
                   WITH H^O
                                   WASH2XWITHCH2CI2
                              * I              I
                            CH2CI2 LAYER    H2SO4 LAYER
                                              I ADJUST TO pH 12 WITH 10% N*OH
                                              I EXTRACT 3X WITH C
            I             I
                YER.._..H2O LAYER
                        j(pH-6)
              EVAPORATE I
                            Ju
                        [ORGANIC BASES


TO DRYNESS !  I WASH 5X WITH CH2C(2
              EXTRACT  CH2CI2 LAYER

              (1) ADDED CYCLOHEXANE
              (2) WASH 3X WITH 4 1
                          H20LAYER

I                             ADJUST TO pH 12 WITH 1M N»OH
                             EXTRACT 3X WITH CH2CI2
                                I ORGANIC BASES!
                                                       I
                                                      AQ
 CYCLOHEXANE LAYER
                         CH2OH/H2O LAYER
        (1) CONCENTRATE        I
        121 WASH 6X WITH CH3NO2   WASH 4X WITH CYCLOHEXANE
                CYCLOHEXANE LAYER
             CYCLOHEXANE LAYER
                            CH3OH/H2O LAYER

                                  FREEZE DRY
EVAPORATE
TO DRYNESS
                 (EVAPORATE TO DRVNESS    I POLAR NEUTRALS!
           JNON
               POLAR NEUTRALSl
             FIGURE  A    Fractionation Scheme for  Gasifier Tars
                                          31

-------
          GOO

          400
          100
          n
          •o
          70
          M
          n
          40
        s
        5 w
        i  •
                                n     M
                                 TIME (HOURS)
FIGURE 5
          Effect of CdCl- on  the  Growth of  Chinese Hamster Cells
          in Culture
     Triplicate cultures  of  CHO  cells are explanted at 10  cells, 35
mm dish, incubated 24 hours  at 37°C in a 5% C02  atmosphere, and CdCla
at 10"e M  ( A  ), 10~7 M  ( A ), and  10~6 M ( • ), was added in 25 yl
DMSO, and  the  cultures incubated 24 hours (arrows), at which time the
medium was replaced with  fresh medium and incubation continued, counting
cells at 24 hours intervals  using an automated cell counter.  Control
cultures were  incubated with 25  yl  DMSO ( Q .  Q ) •
                                   32

-------
                           CRUDE TAR
                            TA98







600

GOO

400
300

200
100











600

500
400
300
200
100

ILLINOIS **
I
\
\
\
1
\

. \
\
• \
\

« \ ^''* "" 	 —
X •** .^'^^""v..
^ — ^^.
*\.
Ill 1
0 100 260 600
N. DAKOTA
O. RUN *S1
t 	 	 	
— — — •
.

-

-
.

-
.
• S^""
*
III
WYOMING

~^~~» '

-

"
.

.

•
-

.^— * 	 ' -
i-X
ill I
10 100 260 GOO
W. KENTUCKY
0. RUN *41
1
\
\
\
\
X

\.

r *
:/ \\ :
' N*~~— \_
XC*
j X
100

M

60 m
8
70S
60 S
r
so je
X
40 |
f-
30

20
10




100

90
S
80 1

70 Jj
m
60 F
g
GO 5
40 1
30
20
10

       10  100
                250
                          600  10
                                 100
                                      260
                                               500
FIGURE 6  Mutagenicity of Four Coal  Gasification Crude Tars
          to Salmonella Strain TA98

     Crude tars suspended in DMSO were  added  at  the doses indi-
cated to 107-108 Salmonella TA98 and 3  mg  S-9 microsomal activa-
tion preparation, and plated on a histidine deficient agar plate
in 2.5 molten agar.  For toxicity studies, the bacteria-sample
mixture is diluted to 1000 cells/ml  and 200 and  600 cells plated
in 2.5 molten agar on histidine-containing agar.  Mutagenicity
results are expressed as total revertants  (-  A -) uncorrected
for toxicity; toxicity results are expressed  as  percent control
cells plated with no sample but including  solvent giving rise
to colonies on histidine-containing  agar (- • -).
                               33

-------
                           WYOMING SUB-BITUMINOUS
                                TA98
      2000

      1100

      woo

     Z MOO

     5 1200

     1000
       MO

       400

       200
                                                rauui NCUTIIAL
         100 128 260 376 600
                              1000 60*2.5 100 126  260

                              DOSE p«/pl«w
                                                376
                                                      600
FIGURE 7  Reproducibility  of Mutagenicity Determinations  for
           Wyoming Subbituminous Coal  Gasification Fractions

      The crude  tars and chemical fractions from  these tars,
derived from 3  separate reactions of  Wyoming subbituminous coal
were  assayed for  mutagenicity as described in the legend  to
Figure 6 and the  text.  Mutagenicity  results are expressed as
total number of revertants/plate uncorrected for toxicity.
Numbers (35, 33,  47) refer to gasifier  reaction  run.
                                 34

-------
                          ORGANIC BASE COMPARISON
                               TAM

2000
1100

1MO
EVERTANTS
i I
K 1000

too
too
400
200
ILLHMMI 1AM
0. RUN '44


-
/^ - -
/'
S
i
i
- i
i
O RUN»33

/'
/'
'
1
i
/
/
/
/
100
to
s
to »
ENT CELL SU
K 8
to a

40 >
30
20
10
                             100
                                  1226 «26   126

2000
itoo
MOO
1400
< 1200
K
I'™0
too

too
400
200

N DAKOTA (AM
0. DUN »«1


•^'X-^
•' V*~~ •— .
. ""'"""••
- /
- '
/
"/
7
A

•.KENTUCKY BASE
8. HUN *41

*"^
*^^^
^\
.
/ ""•••^4__3 	 ~*
/ V
/ \
^ /
" ! \
i \

100
to
•0
70
to

(0
40

30
20
10

                             BOO    10   100   260
                              DOSEf
FIGURE 8  Mutagenicity of Organic Base Fraction of Gasification
          Tars  to  Salmonella Strain TA98

     Organic base  fractions partitioned from crude tars of four
coal types were  tested for mutagenic potential to strain TA98
with 3 mg S-9 microsomal activation preparation, as described
in the text and  legend to Figure 6.  Mutagenicity results are
expressed as total revertants per plate (-A-) uncorrected
for toxicity; toxicity results are expressed as percent control
cells plated with  no  sample (but including solvent), giving rise
to colonies on histidine-containing agar (- • -).
                               35

-------
                        TABLE  1

      Energy,  Tar  and  Oil Yields  From
     Fixed-Bed Gasification Reactors
REACTOR TYPE
 (coal type)
GAS PRODUCTION       ENERGY VALUE            YIELDS
   (SCF*/lb)         (BTU**/SCF*)    (Percent of Coal Feed)
                                    TAR       OIL
LURGI

lignite
sub-bitiminous
bituainous

SLAGGING LURCI

lignite
bituminous

WELLMAN-CALUSHA

sub-bituainous
bituminous

RILEY-MORGAN

lignite

CHAPMAN-WILPUTTE

bituminous

WOODALL-DUCKMAN

bituainous
     NA***
     35
     37
     NA
     35
      45
      65
     •\,30
     •\-20
                         25
NA           1.2
300          1.5
298          3.0
NA
375
150
164
                        150
                        170
                                           200
1.7
7.0
3.4
6.0
                                    3.7
                                    10
                                                        7.5
         4.0
         0.7
         1.0
5.6
NA
NA
NA
                                             4.0
                                             NA
  tSC¥ - standard cubic foot
   BTU - British Thernal Unit
   NA
       Data not available
                                36

-------
                        TABLE 2

Quantitative Results for Selected Compounds in Tar Fractions
(expressed in /ig of compound/gm coal dropped)
Fraction
PNA
Organic
Bases
Organic
Acids

Naphthalene
Alkylnaphthalene
Fluorene
Dibenzofuran
Anthracene +
Phenanthrene
Fluoranthene
Pyrene
Chrysene
5-Ring Compounds
Alkylpyridines
Quinoline
Alkylquinolines
Alkylbenzoquinolines
Acridine
Phenol
Cresols
Xylenols
o-isopropylphenol
trimethylphenol
Run 47
Wyoming
42
44
17
16
36
16
15
6
Trace
18
4
2
ND
1
431
787
954
10
86
Run 51
North Dakota
384
134
59
53
145
36
29
28
9
25
25
4
4
2
263
799
112
170
32
Run 52
Illinois No. 6
1,490
230
167
191
801
276
191
139
112
20
8
8
111
2
15
32
10
<1
1
GC/MS Conditions:
LKB-2091 GC/MS, 15-20 mWCOT capillary column, 1%Se-30/
BaC03, 100°/2 min/8° min/265°.
                              37

-------
                              TABLE 3

Mutagenicity of Coal Gasification Crude  Tars  and  Fractions
                      Percent of     Specific Activity        Activity
                      Crude Tar       (revertants/lJg)       (revertants)
No th Dakota Lignite
Ra crude tar
Ta Bases
Ta PNA
*T r Acid
Nonpolar Neutral
Wyoming (Subbitumlnous
Raw crude tar
Tar Bases
Tar PNA
Tar Polar Neutral
*Tar Acid
*Tar Nonpolar Neutral
Wyoming Subbitumlnous
Raw crude tar
Tar Bases
Tar PNA
*Tar Polar Neutral
*Tar Acid
*Tar Nonpolar Neutral
Wyoming Subbitumlnous
Raw crude tar
Tar Bases
Tar PNA
Tar Polar Neutral
*Tar Acid
*Tar Nonpolar Neutral
Western Kentucky^ No. 9
Raw crude tar
Tar Bases
Tar PNA
Tar Polar Neutral
*Tar Acid
Tar Nonpolar Neutral
Illinois No. 6
Raw crude tar
Tar Bases
Tar FNA
Tar Polar Neutral
*Tar Acid
*Tar Nonpolar Neutral

100.0
A. 55
36.26
23.61
17.1
(33)
100.0
3.33
40.0
7.50
26.67
20.83
135)
100.0
3.44
35.40
0.28
29.55
18.21
(47)
100.0
2.54
36.26
7.76
30.70
27.82
100.0
6.98
61.79
5.32
5.32
12.62
100.0
6.67
66.67
3.63
6.67
11.25

1.915
17.91
.852


1.189
8.504
1.558
1.417

0.399
15.62
.652

0.265
5.8
0.36
0.30

8.69
33.79
7.72
1.766
"
11.29
6.23
1.965
37.46


191.5
81.49
30.89
127,2

118.9
28.32
62.32
10.63
101.28

39.9
53.73
23.08
76.81*

26.5
14.73
13.05
2.33
30. 11*

869
235.85
477.02
9'4 +
722.27

1129
41.55
131.0
135.98
30B.53

 .Specific Activity -  (revertants with (ample - spontaneous  rev«rtants)/dose
  Activity - specific  activity x fraction mass per 100 ug of composite material
 t(corrected for toxicity)
  Nonmutagenic
  Additive Mutagenic Activity (fractions comprising the crude tar)
 ( )  gasification reaction number
                                   38

-------
                          TABLE 4

  Cytotoxicity of Coal Gasification Tars and Chemical
Class Fractions to Chinese Hamster Ovary Cells in Culture.
Sample
Western Kentucky
CRUDE TAR
PNA
POLAR NEUTRALS
ACIDS
BASES
NON-POLAR NEUTRALS
CYCLOKEXANE
INSOLUBLES
XAD-2
STEADY STATE
XAD-2
SURGE
XAD-2
CONTROL
H20 CONDENSATE
Wyoming Subbltumlnous
CRUDE TAR
PNA
POLAR NEUTRALS
ACIDS
BASES
S.,0 CONDE1JSATC
Illinois No. 6
CRUDE TAK
PfcA
ACIDS
BASES
POLAR NEUTRALS
NON-POLAR NEUTRALS
I Total
Crude Tar
100.00
61.79
5.32
5.32
6.98
12.62
7.97
-

100.0
36.26
7.76
30.7
2.54
-

100.0
66.67
6.67
6.67
3.63
11.25
Concentration
mg/ml
10.0
10.0
10.0
5.0
15.0
10.0
10.0
5.0
5.0
5.0
10.0

5.0
10.0
1.0
1.0
0.1
CONCENTRATE

10.0
10.0
5.0
0.1
2.5
10.0
Mg Sample
Per 2 ml
250
100
250
100
250
100
125
50
375
150
250
100
250
100
125
50
125
50
125
50
250
100

125
50
250
100
25
10
25
10
2.5
1.0
23

250
100
250
100
125
50
2.5
1.0
62.5
25
250
100
Growth
Kinetics
X Inhibition
83
69
69
43
0
5
69
52
90
74
96
90
22
22
79
63
22
22
0
0
0
0

36
36
94
63
78
64
83
71
6
28
45
0

72
72
77
60
0
0
0
0
0
0
72
44
Clonal
Efficiency
1 Inhibition
56
45
23
12
0
12
12
0
53
50
23
5
20
17
36
38
30
9
26
6
45
0

26
5
80
40
38
56
54
49
12
10
19
0

100
98
95
70
15
18
12
22
42
50
88
98
                            39

-------
                                       TABLE 5

              A Sample In Vitro  Biotoxicity Screening Battery
Test
Hut agenesis
Salmonella /hi s+

Chinese Hamster Ovary Cells
(HGRPT)
House Lymphoma 1.5178Y
(TK-1)
Salmone 1 la / 8-az aquanine

Saccharomyces (yeast)

Tradescantia

Transformation
C3H 10T.5 cells

Syrian Hamster Cell
Enhancement of Viral
Transforamtion

Clastogenic (DNA Damage)
Sister Chroma t id Exchange
(Mammalian Cell)
Unscheduled DNA synthesis
(Mammalian Cell)
Pol A+/?ol A (c. coll)

Parameter Assayed
(Bacteria)
Reverse mutation (point)
Forward mutation (point)

Reverse mutation (point)

(Bacteria)
Forward nutation (point)
Eukaryotic mutation
(point)
Plant nutation (point)
*De tacts vapors

oncogenic (Focus-forming)
chemical transformation
oncogenic transformation
oncogenic event
*degects inorganic
carcinogens

DNA breakage and
re J oining
Increased DNA repair
activity
Increased DMA repair
activity (Bacteria)
Sample Size
100 mg-1 gn

100 mg-2 gm

100 mg-2 gm

100 mg-1 gm

100 ng-1 gm

10 Mg-1 mg


2-5 gm

2-5 g>
2-5 gm



2-5 g»

2-5 g*

100 mg-1 gm

Time Required
2-4 veeks

6-8 veeks

6-8 weeks

2-4 veeks

2-4 veeks

1 week


12-14 weeks

12-14 weeks
12-14 weeks



4-6 veeks

4-6 veeks

2-4 veeks

Route of
Administration
variable

variable

Variable

variable

variable

respirable


variable

variable
variable



variable

variable

variable

Rabbit Alveolar Macrophage
Chinese Hamster Ovary Cell
lununotoxicity to lung   25 mg-1 gtn
Cytotoxicity           500 ugm-100 i
2 veeka
3 veeks
respiraMc
variable
                                              40

-------
                COLLECTION AND RECOVERY OF ORGANICS FROM
                  WATER USING XAD-2 AND XE-347 RESINS

               J. C. Harris, M. J. Cohen and M. J. Hayes
                         Arthur D. Little, Inc.
                                ABSTRACT

Sampling and analysis of organic materials in aqueous industrial effluents
has most frequently been approached by procedures using solvent extraction
for isolation of the organics.  This approach has some well known drawbacks
when applied to materials that are highly polar or present at very low
concentration.  The use of a solid adsorbent material as an in situ ex-
tractive sampler appears to be an attractive alternative approach.

A systematic laboratory investigation has been conducted to determine the
applicability of macrorecticular resins for general and compound specific
sampling of organics in water.  The experimental approach involved quanti-
tative characterization of sorbent/sorbate systems by frontal chromato-
graphic analysis of breakthrough and measurement of recovery.  Effects of
sample flow rate, sorbent type, organic compound type and, to a limited
extent, sample matrix have been considered.  A mixed resin cartridge,
incorporating XAD-2 for collection of non-polar organics and XE-347 for
collection of polar organics, has been tested and appears promising for
full scale sampling applications.

                              INTRODUCTION

The qualitative and quantitative analysis of trace levels of organic
compounds in aqueous industrial effluents has been the object of increasing
interest and study.  Current EPA sampling and analysis procedures for
organics are directed towards comprehensive characterization, as in Level 1
(9), or towards specific compound quantification as in Level 2 (6) or in
priority pollutant analysis  (2, 3).  Liquid-liquid extraction with an
immiscible organic solvent is the approach that has been most widely used
for isolation of organics.  This approach, which has its roots in the
methodology of synthetic organic chemistry, is capable of giving quanti-
tative recoveries of a wide range of non-polar organic materials.  On the
other hand, transport and extraction of large volume (1-10L) aqueous
samples, required to meet detection limit constraints, is cumbersome and
frequently complicated by the formation of emulsions.  Furthermore, very
polar, water-soluble organic species are not always extracted efficiently.

An attractive alternative approach for collection of organics from large
quantities of water is the use of an in situ continuous extractive
sampler containing a solid sorbent material.  An early device of this
type, once widely used in the field, was the EPA's carbon adsorption
module (CAM) (15) which had generally good collection efficiency, but
low recoveries, for organics in water.  A number of previous workers had
investigated other sorbent media including macroporous resins (7, 8, 13,
14, 16, 17), ion exchange resins (12) and polyurethane foam plugs (10).
Based on information available in the literature, it appeared that sor-
bent approaches offered potential advantages over solvent extraction
by a number of criteria:
                                    41

-------
     1.   Convenience:  A  compact sorbent cartridge  is more easily
         handled and shipped than a  large volume aqueous  sample.

     2.   Sample Integration:  A sorbent approach is  inherently
         compatible with  time-integrated sample collection
         while solvent extraction is more commonly  associated
         with grab or composite/grab aqueous samples.

     3.   Range of Applicability:  Sorbent approaches appear to
         be  suited for ppm  and lower levels of organics.  Solvent
         approaches are good for ppm or higher levels.

     4.   Recovery of polar  organics:  A sorbent approach  can
         potentially be tailored to  allow good recovery of
         polar organics as  well as hydrophobic species.

This paper describes and presents some results of a systematic laboratory
investigation of the applicability of macroreticular resins for general
and compound specific sampling of organics,  based on small scale, chromo-
tographic experiments.   The objective was to develop screening procedures
and a data base that would allow selection of the resin(s)  best suited
to a particular sampling application and selection of a sampling module
size/configuration that gives quantitative collection of contaminant(s)
from the aqueous stream of interest.

                        EXPERIMENTAL APPROACH

The basic approach used in this work was the generation and analysis of
the frontal breakthrough curve produced by challenging a sorbent car-
tridge with an aqueous solution of organic analyte at constant flow rate.
This methodology has been described in some detail elsewhere (4, 5).  The
approach, which has also been successfully applied to air sampling sys-
tems (1, 4,  11), involves determination of the volumetric capacity (Vr0)
of the sorbent for the specific pollutant.  The relationship of V,-n
to a characteristic point in the frontal breakthrough curve is
illustrated in Figure 1.   The volumetric capacity is a quantitative
measure of the breakthrough characteristics of the sampling device and
is the maximum volume of sample stream that can be pulled through the
resin while still retaining the pollutant.  The units of V,_n are mL (of
aqueous sample)  per cc (of resin bed).

The experimental apparatus employed for initial chromatographic screening
studies was a small scale  (3.2 cm x 0.5 cm I.d.)  sorbent cartridge con-
nected to an LDC UV monitor and Linear Instrument recorder.  The cartridge
used in the full scale studies was constructed of 15 cm x 3.8 cm i.d.
stainless steel tubing.  Test solutions were prepared using degassed
distilled deionized water and were introduced using a Milton Roy mini-
pump or a Fluidmeter, Inc. Pump.
                                    42

-------
                      o
                     Q.
o
o
                                   O
                                o .y
                                        vp
                                        O^
                                        O
                                        LT)
                                                                            O
                                                                            in
                                                                                           0)
                                                                                           E
                                                                                          b
                                                                                           «
                                                                                           E
                                                                                                                     0)

                                                                                                                     £
                                                                                                                     3
                                                                                                                     60
                                                                                                                     3   >.
                                                                                                                     O  4J
                                                                                                                     !-i  -H
                                                                                                                    ^   O
                                                                                                                    4-1   Cti
                                                                                                                    ^J   ft
                                                                                                                     nJ   «)
                                                                                                                     a)  u
                                                                                                                     M
                                                                                                                    PQ   O
                                                                                                                        •H
                                                                                     4J  0)
                                                                                     c  a
                                                                                     O  3
                                                                                     t-t H
                                                                                     fn  O
                                                                                          o
                                                                                          LO
                                                                                                                     o  >
                                                                                                                        CO
                                                                                                                    w
                                   a;
                                   o
                                   c
                                   o
                               CO  O
03
0)
                                                  43

-------
In the recovery studies, the exposed resins were Soxhlet extracted with
methylene chloride or methanol for 24 hours.  The solvent extracts were
concentrated by Kuderna Danish evaporators to a convenient volume,
prior to analysis by GC/FID on a 10% OV-101 column with temperature
programming.

The sorbent resins studied were obtained from Rohm & Haas and cleaned
by sequential twenty-four hour Soxhlet extractions with distilled, de-
ionized water, methanol and methylene chloride.  The cleaned resins
were stored in methanol.  All other chemicals and solvents used were
Reagent Grade.  The "real" sample utilized in evaluation of the full
scale cartridge was an aqueous effluent associated with a coking
facility and generously provided by the facility operator for this re-
search program.

                        RESULTS AND DISCUSSION

Effect of Sorbent Type

The initial studies in this program were done using XAD-2, the macro-
reticular resin most extensively previously studied for collection of
organic species from water.  The results of these studies (5) essen-
tially confirmed the conclusions reported by others:  XAD-2 has high
volumetric capacity for non-polar organics, but low capacity for species
such as phenol or benzoic acid.

Several other macroreticular resins were tested to explore their poten-
tial for use as alternatives or supplements to XAD-2 for collection of
polar species.  These included:  XAD-7, an acrylic ester polymer;
IRA-904 and IRA-93, anion exchange resins; XE-340, XE-347 and XE-348,
carbonaceous adsorbents.  The XAD-7 was found to have higher affinity
for phenol, but lower affinity for benzoic acid, than XAD-2.  This resin
was therefore not considered a good candidate for collection of polar
organics in a comprehensive sampling system.  The ion exchange materials,
IRA-904 and IRA-93, were excluded from consideration because it was
found that breakthrough on these resins was virtually instantaneous if
the aqueous solution contained as much as 250 ppm of sodium chloride.
The Ambersorb series of carbonaceous resins, XE-340, XE-347 and XE-348,
showed high affinity for a range of polar compounds tested.  The volu-
metric capacity of XE-347 for phenol, benzoic acid and benzyl alcohol
at 10 ppm in water is 75 times that of XAD-2.  This can be attributed in
part to the higher surface area of the XE-347; the XE-347 resin used in „
this study was found to have a BET surface area of 700 m /g versus 330 m /g
for XAD-2.  However, the 75-fold higher V _ value found for XE-347
almost certainly reflects stronger specific interactions between  resin
and sorbate(s) as well.  Within the XE series, the V   values were
generally comparable and decreased in the order:  XE-347, XE-348, XE-340.
Based on these resin screening studies, XE-347 was selected as the most
attractive candidate to supplement XAD-2 in a general purpose system
for collection of organics from water.
                                     44

-------
Effect of Flow Rate

A series of experiments was performed to determine the effect of flow
rate on the V,-,. value.  The data in Table 1 show V _ values observed for
a 10 mg/L acetophemone solution using XAD-2 resin in a small scale
(0.5 cm i.d.) cartridge.  Mass transfer effects are pronounced at the
high face (linear) velocities corresponding to rapid sample flow rates.
The net effect of the kinetic non-equilibrium is a decrease in the
effective capacity of the column; the V,-n value at 25 cm/min (6 mL/min)
is less than half that at 4 cm/min (1 mL/min).

In sampling applications it is possible to compensate for the decreased
efficiency at high flow rate by using a much greater quantity of sorbent
than that required for quantitative collection under equilibrium condi-
tions.  Such compensation may account for some published reports (e.g.,
14) of "no breakthrough" at face velocities of 100 cm/min or higher.

In the limit of low flow rate, <4 cm/min, V,._ is independent of the
face velocity.  This information can be used to design a sorbent car-
tridge that is scaled appropriately for a particular sampling applica-
tion, as illustrated in Table 2.  For example, to collect organics from
a 10 L water sample in a 4 hour period (40 mL/min) would require a
sorbent cartridge 1.5 inches in diameter or larger.  The total quantity
of sorbent resin required and thus the necessary depth of the resin bed
can be calculated from the sample volume and the V,-n for the particular
sorbent-solute(s) of interest.

Full-Scale,  Two-Resin Cartridge

From the earlier screening studies it was apparent that no single resin
would suffice to adequately sample a wide range of organic chemical
classes.  Therefore, a two-stage cartridge with an XAD-2 resin bed up-
stream of an XE-347 resin bed was designed and constructed.  The intent
was to allow collection of non-polar, readily adsorbed organics on the
XAD-2, from which they could be efficiently recovered.  The more polar
species, poorly retained by XAD-2, would be trapped by the downstream
XE-347 section.

The diameter of the cartridge was set at 3.8 cm (1.5 in) based on a pre-
sumed sample size criterion of 10 L in 4 hours as in Level 1 procedures
(9).  The minimum depth of each resin bed, based on the retention
characteristics of the resins used, was calculated to be 3 cm (1.2 in).
In order to minimize possible channeling and back eddy effects, however,
a bed depth of about 7.6 cm (3 in) was used for each resin.  The car-
tridge is thus, in a sense, overdesigned for this application and could
be used for a sample of up to 30 L (12 hr) without breakthrough.
                                    45

-------
                        TABLE 1
            Sorbent Capacity as a Function
                  of Sample Flow Rate
Flow Rate               Face Velocity            T,     T ,
  T / •                        / •                  Vcn> mL/cc
 mL/min                    cm/ mm                 50	
   6.1                      31                       60
   3.1                      16                      150
   1.6                       8.1                    280
   0.88                      4.5                    430
                              46

-------
                        TABLE 2
             Scale Up of Sorbent Cartridge

  Diameter of                         Maximum Flow
Resin Bed (in.)                           Rate
      1/4                               1 mL/min
      1                                16 mL/min
      1.5                              40 mL/min
      8                                 1 mL/min
                           47

-------
The performance of the cartridge was tested and compared to solvent
extraction using both a mixture of 7 model compounds in distilled water
solution and a real complex aqueous effluent sample.  In each case, the
sample was pumped through the cartridge at 40 mL/min.  The direction of
flow was up (against gravity) through the XAD-2 section and then through
the XE-347 section.  The effluent was monitored by UV (254) to confirm
that <1% breakthrough occurred.  The sorbed organics were recovered
from the resins either by separate batch extractions of the two sections
after removal from the cartridge or by continuous extraction of the in-
tact cartridge, using methylene chloride.

Table 3 presents some of the data for recovery of model compounds via
the cartridge and via solvent extraction.  The batch extraction data
show that the non-polar compounds are collected on the XAD-2 as expected,
and do not breakthrough into the XE-347 section.  (The low recovery of
alkanes such as dodecane has been observed consistently in this work
and others (17).  It is postulated that this is due to losses from the
original aqueous sample onto container walls.)

Phenol, however, is distributed roughly 50:50 between the XAD-2 and the
XE-347.  Pyrrole, which has even higher water solubility than phenol,
is found only in the XE-347 portion of the sampling system.

It was found that methylene chloride and methanol were equally effective
at recovery of sorbed phenol from XE-347.  Because of this, and because
phenol, at least, was distributed in both sections of the cartridge,
there was no necessity for or advantage to extracting the two resin
sections separately.  For convenience, therefore, a continuous extrac-
tion procedure applicable to the intact cartridge was adopted.  The
hardware used was that developed for cleaning XAD-2 resin.  (See
Reference 9, Appendix D.)  Table 3 shows that the total recoveries for
batch and continuous extractions of the cartridge are equivalent.

Table 3 also includes data for solvent extraction of a 1 L portion of
the model compound test mixture by Level 1 procedures (9).  It is clear
that the cartridge approach is roughly comparable to, though perhaps
slightly less effective than, the solvent approach for recovery of the
non-polar organics.  (Dodecane is an exception, as noted above.)  For
phenol and pyrrole, the sorbent approach gives significantly higher
recoveries than does solvent extraction.

When the sorbent cartridge was challenged with a real complex aqueous
effluent (a coking plant waste), the results were entirely consistent
with expectations based on the model compound studies.  Figure 2 shows
the chromatograms of the concentrated sample extracts obtained by the
XAD-2/XE-347 cartridge (2A) and by solvent extraction (2B).  The XAD-2/
XE-347 extract clearly contains more of the shorter retention time—
presumably lower molecular weight, more water soluble—species than does
the direct liquid-liquid extract.  The peak marked with an asterisk in
Figure 2A, for example, which  is probably due to phenol  (based on reten-
tion time and sample source) is only about one-tenth as intense in the
                                    48

-------
     c
     o
AJ  -H
 a  ^
 0)   O
 >   tfl
                                        oo
                                                        VO
 W   fi
 3   O
 O  -H
 3  4-1
 C   O
•H   rt
    w
C3>    VO
\o    oo
IT)    rH
00    CM














ro

W
pq
^3
H













to
C
3
O
ft
B
o
u
—1
r~l
QJ
*~U
0
IS
IH
0

^">
J^|
QJ
j>
0
o
<1>








G
o
•H
4-1
O
cd
4-1
X
w
J-
o
4-1
cd
pq






r-l
cd
4-1
O
H


ro LO 00
i — r^ r~-


C
•H
tn
01
pd
B
o r**~
£ "*
i
Pd
X






o o o







c
•r-l
to
0)
E M
0
S-I CM
ta§





ro in oo
r*^ r**^ r^*»

                                             oo
                                             00
                                             CO
13
3
O
D
B
0
u
Trichloropropane
i
ro
«*
CM
r*
r-l
alene
f-^
4-1
f~]
cx
cd

Iphthalate
K^
^3
4-1
0)
•H
Q
0)
•H
T)
•H
SJ
O


r-l
0
c
0)
I-C
PM
Ol
G
qj
0
0)
T3
0
Q
0)
r-l
0

H
!>v,
PM
                                                                                OJ
                                                                                4J
                                                                                w
                                                                               4-J
                                                                                O

-------
 o
•H
4-1
 O
 rt
w
TJ
•H
cr
•H
•H
3
CT1
                     60
                    T3
                    •H
                     nj
                     O
                                                                    /d
                                                                               C-J

                                                                               W
                                   50

-------
solvent extract (2B).   The overall recoveries of organics from the aqueous
effluent sample by the two approaches are compared in Table 4.   The two
approaches give equivalent recoveries of the high boiling, high molecular
weight (GRAV) range organics.  For TCO range, lower boiling,  lower
molecular weight materials, however, the cartridge approach yields a
50% higher recovery.  It is significant that most of the enhanced recovery
is apparently due to the types of compounds (e.g., phenol) that  are col-
lected on the XE-347 section of the cartridge.  However, since this
particular sample had  a preponderance of GRAV range material, the total
recovery for the XAD-2/XE-347 cartridge is only 5% higher than that for
the solvent approach.

                              CONCLUSIONS

The conclusions that have been drawn from the work reported here include
the following:

     1.  Small scale studies have predictive value for sample
         system design.

     2.  A slow sample flow rate through the resin is important
         for quantitative uptake.

     3.  A combination of resins, XAD-2 for non-polar organics
         and XE-347 for polar organics, appears to be the most
         promising general approach.

     4.  For the XAD-2/XE-347 cartridge, the recovery of most
         model compounds is comparable to solvent extraction.

     5.  Phenol and pyrrole recoveries are higher for the
         cartridge than for the solvent approach.

     6.  For a real complex aqueous effluent, XAD-2/XE-347
         gives better  recovery of species  such as phenol.

     7.  For a real complex aqueous effluent, sorbent and
         solvent give  comparable recoveries of GRAV range
         organics.

     8.  High levels of suspended particulate in the
         aqueous effluent will require modifications of
         sorbent cartridge.

                            ACKNOWLEDGEMENTS

This work was performed under EPA Contracts No. 68-02-2150 and No. 68-02-3111
from the Process Measurements Branch, Industrial Environmental Research
Laboratory, Research Triangle Park, N.C., under the direction of Dr. Larry
Johnson.  The authors thank Rose Fasano, Zoe Grosser, Philip Levins, and
Karen Weaver for their contributions to this work.
                                     51

-------
                            TABLE 4

              Overall Recovery of Organics From
                    Aqueous Effluent Sample
                             TCP* mg/L    GRAV** mg/L      Total mg/L

Sorbent Cartridge               8.4            47             55

Solvent Extraction              6.1            46             52
 Total Chromatographable Organics = Materials in about 100°C to
 300°C boiling point range.

 *
  GRAVimetric Organics = Materials with boiling points above
  about 300°C.
                                  52

-------
                                REFERENCES

 1.   Adams,  J.  W.,  K.  T.  Menzies,  and P.  L.  Levins,  "Selection and
     Evaluation of  Sorbent Resins  for the Collection of Organic
     Compounds," EPA-600/7-77-044  (April  1977).

 2.   Environmental  Protection Agency, "Sampling and  Analysis Procedures
     for Survey of  Industrial Effluents for Priority Pollutants," EMSL,
     Cincinnati, OH (April 1977).

 3.   Environmental  Protection Agency, "Guidelines Establishing Test
     Procedures for the Analysis of Pollutants," Federal Register, 44,
     69464-69575, Dec. 3, 1979,  and Federal Register, .44, 75028-75052,
     Dec.  18,  1979.

 4.   Gallant,  R. F., J. W. King, P. L. Levins, and J. F. Piecewicz,
     "Characterization of Sorbent  Resins  for Use in Environmental
     Sampling," EPA-600/7-78-054 (1978).

 5.   Grosser,  Z. A., J. C. Harris, and P. L. Levins, "Quantitative
     Extraction of  Polycyclic Aromatic Hydrocarbons  and Other Hazardous
     Organic Species from Process  Streams Using Macroreticular Resins,"
     in Polynuclear Aromatic Hydrocarbons, P. W. Jones and P. Leber,
     eds.,  Ann Arbor Science Publishers,  Inc., Ann Arbor, MI, 67-79 (1979),

 6.   Harris, J. C., M. J. Hayes, P. L. Levins, and D. B. Lindsay,
     "EPA/IERL-RTP  Procedures for  Level 2 Sampling and Analysis of
     Organic Materials," EPA-600/7-79-033 (February 1979).

 7.   Junk,  G.  A., J. J. Richard, J. S. Fritz, and H. J. Svec, "Resin
     Sorption Methods for Monitoring Selected Contaminants in Water,"
     in Identification and Analysis of Organic Pollutants in water,
     edited by L. H. Keith, pp.  135-158,  Ann Arbor Science, Ann Arbor,
     MI (1976).

 8.   Junk,  G.  A., J. J. Richard, M. Grieser, D.  Witiak, J. Witiak,
     M. Arguello, R. Vick, H. Svec, J. Fritz, G. Calder, "Use of
     Macroreticular Resins in the  Analysis of Water for Trace Organic
     Contaminants," J. Chromatogr., 9_9, 745-762 (1974).

 9.   Lentzen,  D. E., D. E. Wagoner, E. D. Estes, and W. F. Gutknecht,
     "IERL-RTP Procedures Manual:   Level 1 Environmental Assessment,"
     (Second Edition), EPA-600/7-78-201 (Oct. 1978).

10.   Navratil, J.,  R. Sievers, and H. Walton, "Open-Pore Polyurethane
     Columns for Collection and Preconcentration of Polynuclear Aromatic
     Hydrocarbons from Water," Anal. Chem., 49,  (14), 2260-2263 (1977).
                                      53

-------
11.  Piecewicz, J. F., J. C. Harris, and P. L. Levins, "Further
     Characterization of Sorbents for Environmental Sampling,"
     EPA-600/7-79-216 (Sept. 1979).

12.  Pollio, F., and R.  Junin,  !'Sorption of Phenols by Anion
     Exchange Resins," Environ. Sci. and Technol., 1, (2), 160-
     163 (1967).

13.  Stepan, S., and J.  Smith,  "Some Conditions for Use of Macro-
     reticular Resins in the Quantitative Analysis of Organic
     Pollutants in Water," Water Res.,  11, (4), 339-342 (1977).

14.  Strup, P. E., J. E. Wilkinson, and P. W. Jones, "Trace
     Analysis of Polycyclic Aromatic Hydrocarbons in Aqueous
     Systems Using XAD-2 Resin and Capillary Column Gas Chroma-
     tography-Mass Spectrometry Analysis," in Polynuclear Aromatic
     Hydrocarbons:  Second International Symposium on Analysis,
     Chemistry and Biology, P.  W. Jones and R. E. Freudenthal,
     eds.,  131-138, Raven Press, NY  (1978).

15.  Taras, M. J., A. E. Greenberg, R.  D. Hoak, and M. C. Rand,
     eds.,  "Organic Contaminants," Standard Methods for the
     Examination of Water and Wastewater, 13th ed., American
     Public Health Association, Washington, DC, 259-270 (1971).

16.  Van Rossum, P., and R. Webb, "Isolation of Organic Water
     Pollutants by XAD Resins and Carbon," (1977).

17.  Webb,  R. G., "Isolating Organic Water Pollutants:  XAD
     Resins, Urethane Foams, Solvent Extraction," EPA-660/4-75-003
     (1975).
                                 54

-------
                   APPROACHES TO LEVEL 1 IR AND LRMS
                MEASUREMENT AND SPECTRAL INTERPRETATION

                  W. F. Gutknecht and A. Gaskill, Jr.
                   Systems and Measurements Division
                      Research Triangle Institute
                         Post Office Box 12194
                     Research Triangle Park, N. C.  27709
                               ABSTRACT

Approaches to measurement and  interpretation of  infrared  (IR) and
low resolution mass spectra (LRMS) taken by four Environmental Protec-
tion Agency (EPA) contractors following the guidelines of the Level 1
Environmental Assessment Program were evaluated.  The mechanism of
evaluation involved supplying each contractor with test spectra and
samples for analysis and interpretation.  Using IR, the contractors
identified 55 to 85 percent of the structural moieties present in the
various samples or indicated by the various test spectra, with 85
percent being approximately the maximum attainable.  Sources of error
in IR measurement included analyst-to-analyst variation in signal
location of ±5-10 cm   and failure to optimize signal intensity
through proper loading of the salt plates or KBr pellets.  IR interpre-
tation errors disclosed were errors of omission and assignment.   To
improve IR data quality, all significant signals must be interpreted,
including those complementary signals that support or refute other
assignments; also, reasonable, alternative interpretations must be
considered.  The contractors identified 50 to 90 percent of the compound
classes represented using LRMS.  LRMS interpretation errors disclosed
were failure to find molecular ions, reporting molecular ions as
fragment ions and vice versa, incorrectly identifying molecular ions
and incorrectly assigning ions to homologous series.   To improve LRMS
data quality, the reference series Eight Peak Index of Mass Spectra
should be used, intensity and mass/charge range criteria set to simplify
spectra, the LC fraction data used with caution, and IR and LRMS data
analyzed independently to avoid bias.

                             INTRODUCTION

Infrared (IR) and low resolution mass spectrometry (LRMS) are corner-
stone techniques in the Level 1 Environmental Assessment program.
Infrared spectrometry is used to identify the organic functional
groups present in an environmental assessment sample; when sufficient
sample is available, LRMS is used in conjunction with the IR technique
to determine the principal categories of organic compounds present in
these samples.  (1)

It is apparent though, from various evaluations of the Level 1 program
that these techniques are not always being used by environmental
assessment contractors to maximize the output of useful data.  To
determine the source of contractors' problems with IR and LRMS tech-
niques, a study was undertaken with the following objectives:

                                   55

-------
          To identify IR and LRMS measurement techniques that either
          promote or deter the collection of accurate analysis results.

          To identify IR and LRMS spectral interpretation techniques
          that either promote or deter the collection of accurate
          analysis results.

          To determine the importance of intangibles such as experience
          to the collection of accurate analysis results.

The study was conducted in two stages.  In the first stage, six IR and
six LRMS test spectra were sent to four Environmental Protection
Agency (EPA) contractors for interpretation.  These spectra were
prepared with six samples that varied both in complexity and functional
group content.  The samples were to represent Level 1 liquid chromat-
ographic (LC) fractions 4, 5, and 6 and were prepared using compounds
chosen on the basis of categories expected in these fractions and
compounds reported in actual environmental assessment studies.  The
compounds are shown in Table 1.

The contractors were requested to interpret the spectra as normal
Level 1 spectra; i.e., no special efforts were to be made with these
spectra beyond that expected for Level 1 work.  However, the contrac-
tors were asked to clearly indicate the spectral characteristics
leading to the identification of functional groups, structural charac-
teristics, or, when possible, individual compounds or compound classes.
Using the analysis sheets provided, the analyst was to spell out
clearly the reasoning behind each assignment, including information
about signal (i.e., peak) wavelength or mass, signal magnitude or
shape, patterns of signals,  or correlation with LC fraction number.

The test spectra interpretation results returned by the contractors
were studied with respect to accuracy, thoroughness of interpretation,
approach to interpretation,  etc.  On the basis of these results, two
synthetic samples were prepared and sent to the contractors for both
IR and LRMS analysis.  The samples for this second stage of the inves-
tigation were prepared to represent LC fractions 5 and 6.  They were
prepared principally from compounds that were used to prepare the
stage one spectra and that proved difficult to identify previously.
This was done, in part, to determine if measurement of their own
spectra would improve the contractors' abilities to make accurate
analyses.  Included with the two synthetic mixtures was a sample of
coal tar from a coal gasification unit.  This sample was to be separated
using the Level 1 LC scheme, and fractions 4, 5, and 6 were to be
analyzed using IR and LRMS.   The composition of the two synthetic
mixtures and a partial listing of the contents of the tar sample are
presented in Table 2.

                        RESULTS AND DISCUSSION

The IR and LRMS results returned by the contractors were studied with
regard to accuracy and sources of error.  Approaches to spectral
                                    56

-------
interpretation were investigated through personal visits and/or tele-
phone conversations with the contractor spectroscopists.

                              IR RESULTS

The accuracy of the analysis results was measured in terms of the
percentage of structures in the test mixtures correctly identified by
the contractors.  These percentages are shown in Table 3.   An accuracy
level of 80 percent is considered very good; in general, identification
at this accuracy level means that only the structures most difficult
to identify have been missed.  Contractors A and C achieved 60-percent
levels on two sample types principally because of low quality spectral
data.  Contractors B and D achieved low levels principally because of
inaccurate and incomplete spectral interpretations.  Details regarding
these sources of error are presented in the following sections.

                            IR MEASUREMENT

Several factors of IR measurement that affect final interpretation
results were observed.  One of these was variation in signal location
from contractor to contractor for signals from the same structural
source.  Such variation was in the range of ±5 to ±10 cm  .   A second
factor was failure to optimize signal intensity.  Weak signals cannot
always be optimized because of insufficient sample quantity; however,
overloading of the salt plate or KBr pellet, which can and must be
controlled, was observed and had the effect of weak or medium signals
being falsely identified as medium or strong signals.  A more serious
factor observed was the weakness or absence of signals in the 4000 cm
to 2000 cm   range in spectra obtained with KBr pellets compared to
those obtained with salt plates.  This may be due to interaction with
the KBr; it may also be due to loss of volatile compounds during the
pellet preparation process.  This particular factor needs further
investigation.

                           IR INTERPRETATION

Sources of difficulty in IR interpretation include interpreting signals
incorrectly, making incorrect though reasonable interpretations, over-
looking important signals, and not using complementary signals that,
taken together, support or refute identification of a particular
structure.

Examples of assignment errors are shown in Table 4.  Some of these
errors may be subject to debate, though the assignments for signals at
1805 cm" , 1600 cm" , 1245 cm~ , and 3052 cm~  clearly seem in error.

Erroneous assignments that are, nevertheless, reasonable occur fairly
often.  These result from the simple fact that there are different
structures giving rise to signals at very nearly the same wavenumber.
Examples of such errors are shown in Table 5.  Esters, ketones, alde-
hydes, and acids all have carbonyl signals at about 1700 cm   and they
are often difficult to differentiate, especially in mixtures.  A large
                                    57

-------
number of structures such as C-0, C-N, C-C(=0)-C, and C-O-C give rise
to signals in the 1100 cm   to 1400 cm   range and these, too, are
often difficult to differentiate.  These erroneous but reasonable
assignments arise from sources such as the disagreement among the
contractor spectroscopists and the reference  works they use for
ranges of signals arising from particular structures.  This disagree-
ment is illustrated in Table 6.

The nature of the spectra may also contribute to erroneous assignments.
Figure 1 compares the spectra from synthetic mixture LC6 (see Table 2),
which contained six compounds, and the spectra from the LC1fraction 6
of the coal tar.  The envelope of signals between 1800 cm   and 1000 cm
is not atypical for a real-world sample.  The broad, overlapping sig-
nals indicate the presence of a variety of different compounds having
the same functional groups and/or a considerable amount of intermolec-
ular bonding, e.g., H-bonding.  One can readily see that accurate
signal assignment with the tar sample spectrum will be difficult.

Errors in signal assignment were observed.  However, the majority of
errors noted were apparently errors of omission; that is, not making
optimum use of the available data.  Table 7 shows examples of signals
simply not assigned.  Another error of omission observed is failure to
make optimum use of supplementary data.  Table 8 shows examples of
such errors.  A signal at 1590 cm  ,  for example, was described as be-
ing due to the NH of an amine.  Signals at 1235 cm   and 1180 cm  ,
which could be due to C-N of the amine and thus support the 1590 cm
assignment, were not described as such.  In another case, a signal at
1240 cm   was described as being due to the C-O-C structure of an
ester.  The spectroscopist did not note, however, the absence of an
ester carbonyl signal at the expected wavenumber value of 1735 cm  ,
which would refute the 1240 cm   assignment.

           ADDITIONAL FACTORS AFFECTING IR ANALYSIS ACCURACY

Factors other than errors in assignment and errors of omission that
affect analysis accuracy include approach to interpretation, time
given to measurement and interpretation, and experience of the spectros
copist.

All spectroscopists approached interpretation in similar ways.  They
relied on memory to identify most of the major signals, and then used
reference texts and tables to identify the minor signals and their
possible relationships to the major signals.  The time given to
interpretation of the spectra dealt with in this study ranged from 1
to 4 hours per spectrum, though the 4 hours included time for assign-
ment of all IR signals to compounds already identified by LRMS.  The
spectroscopists involved in this study had levels of interpretation
experience ranging from 6 months to 25 years.   Upon examining these
factors and the accuracy levels attained, no strong correlation between
accuracy of interpretation and time given to interpretation was found.
There was a correlation between accuracy of interpretation and experi-
ence of the spectroscopists, though this correlation was also weak.
                                  58

-------
The accuracy levels attained by spectroscopists with 25 years, 22
years, 11 years, 7 years, 3 years, and 6 months experience (one con-
tractor had three spectroscopists involved in the study) were in the
relative order 1, 3, 4, 2, 6, and 5, respectively.

Upon examining all factors affecting accuracy of the IR analyses, it
appears that thoroughness of interpretation is most important.  In
order to achieve a high level of accuracy, the experienced, qualified
spectroscopist should (1) interpret all signals of significant magni-
tude; (2) search out and identify complementary signals that support
or refute other signal interpretations, and (3) make alternative
interpretations of signals when such interpretations are reasonable.

                             LRMS RESULTS

The interpretation of Level 1 low resolution mass spectrometry was
first studied by supplying each of the four contractors with six low
resolution mass spectra obtained with a heated probe sample inlet.
Each of the mixtures used to generate the test spectra was represented
by a single spectrum like the spectrum for LC6 shown in Figure 2.
These spectra were taken under conditions that would optimize as many
of the mixture component signals as possible in a single spectrum.  In
the spectrum shown in Figure 2, the molecular ions for each of the
five components are indicated.  In the actual test spectra, though, no
peaks were singled out.  The axes in Figure 2 show that the percent
intensity of each ion relative to the most intense ion, 224 (fluorenone
carboxylic acid), is plotted vs. the mass to charge ratio as is typical
in data-system-generated spectra.  One contractor noted that, in their
laboratory, a series of spectra are usually taken over varying probe
temperatures and at both 70 eV and 15 eV in order to fully character-
ize a sample.  All of the test spectra were taken at 70 eV and only a
single probe scan was given in each case.  Thus, for this contractor,
the test spectra omitted a substantial amount of information.

The contractors' rationale behind the assignment of individual compounds
was the focus of both this phase of the study and the test mixture
phase.  Although one of the goals of Level 1 is to identify categories
of compounds as well as individual compounds, the importance of the
contractors' category assignments in this study was deemphasized for
two reasons.  First, although the correct category assignment was made
in several instances, it was based on an incorrect compound identifica-
tion.  This could be somewhat misleading to the investigator interested
in evaluating the approach to interpretation being taken.  Second, in
general,  Level 1 category assignments tend to be based on the individual
compounds identified.

Some of the results of the contractors' analyses of the test spectra
are shown in Tables 9, 10, and 11.  Each table shows the compounds
present,  their molecular weights, and the contractors'  intensity
assignments according to Level 1 (either 1, 10, or 100) for those
compounds correctly identified.  The daggers indicate correct compound
assignments for which the contractor did not report the intensity.  In
                                   59

-------
general, the assignments of isomers of the actual compounds were
considered correct assignments because isomers usually cannot be
differentiated by LRMS alone.   For example, ethyl anthracene and
dimethyl anthracene were acceptable assignments for 9,10-dimethyl
anthracene.  The results for test spectrum LC4b (Table 9), a mixture
of alkylated anthracenes and naphthalenes, show that 22 out of 24
possible correct identifications were made (91 percent correct).
Anthracenes and naphthalenes are probably easily identified in mixtures
because most analysts are familiar with their spectra arid because
these compounds have intense,  easily recognized molecular ions.  The
results for spectrum LC5a (Table 10), a mixture of indoles, carbazoles,
and fluorenes, indicate that several contractors had difficulty iden-
tifying the carbazoles and fluorenes.  The results show that 12 out of
24 possible correct identifications were made (50 percent correct).
The results for spectrum LC6b (Table 11), a mixture of a carboxylic
acid (1-fluorene carboxylic acid), keto-acid (9 fluorenone-4-carboxylic
acid), ether-acid (o-phenoxybenzoic acid), and a phenol, show that
very few of these compounds were identified correctly (5 out of 20; 25
percent correct).  Contractors A and D did not identify any compounds
correctly.

The results of the contractors' analyses of the test mixtures were
somewhat improved over the test spectra results.  The results for
mixture LC6 (Table 12) show that every contractor identified at least
one compound correctly and one contractor, C, identified all of the
compounds correctly.  In all, 11 out of 24 possible correct identifi-
cations were made (45 percent correct).  A comparison of categories
identified in mixture LC6 vs.  the results for spectrum LC6b (Table 13)
indicates that only ketones were found more easily in mixture LC6.  It
is also interesting to note that one of the compounds common to both
test spectra and mixtures, fluorenone carboxylic acid, can be listed
as either a ketone or an acid.  In this study it was listed as an
acid.

Unlike the interpretation of the synthetic spectra and mixtures, the
contractors' assignments of compounds in the coal gasification tar
sample were generally based on molecular ions only, with supportive
evidence given by the finding of an alkylated homologous series such
as alkyl naphthols.  In general, the contractors reported the series
rather than a single compound.

                      LRMS INTERPRETATION ERRORS

Five general types of interpretation errors observed in the results
from the test spectra and mixtures are:

     1.   Failing to find molecular ions.

     2.   Reporting molecular ions as fragment  ions.

     3.   Reporting fragment ions as molecular  ions.
                                    60

-------
     4.   Incorrectly identifying a molecular ion.

     5.   Incorrectly assigning ions to homologous series.

Not finding the molecular ion for a compound invariably resulted in
the compound not being identified.  This was a problem for contrac-
tor B in general and for all contractors in the interpretation of
spectrum LC6b.  Failure to find molecular ions accounted for almost
half the compounds missed in the test spectra and test mixtures.  The
reporting of molecular ions as fragment ions of other compounds and
vice versa was also a significant error observed in the results for
certain spectra.  For example, fluorene in spectrum LC5a (Table 11)
was not correctly identified by any contractor but two contractors
reported the molecular ion as a fragment ion for carbazole.  In the
case of fluorene, the error is understandable because many of the most
abundant ions for fluorene, including the molecular ion, were also
fragment ions of the more dominant components, carbazole and ethyl
carbazole.  Also in LC5a, contractors A and B incorrectly identified
the m/e 180 fragment ion for ethyl carbazole as the molecular ion for
fluorenone, a compound not present in the spectrum.  In several cases
the molecular ion was identified incorrectly as belonging to another
compound.  Contractors A and B incorrectly identified the molecular
ion for ethyl carbazole in LC5a as acridone.  The final kind of error
observed was that resulting from the spurious identification of an
alkylated homologous series based on the identification of a nucleus
molecular ion and the subsequent assignments of ions at the appropriate
intervals (14 m/e units) to the series members.  Contractors B and D
made this error in incorrectly identifying the molecular ion for
phenazine and 1,10-phenanthroline in spectrum LC5b as being due to
fluorenone.  Based on the assumed presence of fluorenone, the molecular
ion for anthrone was incorrectly identified as belonging to methyl
fluorenone.  Clearly, although the presence of a homologous series in
a sample could potentially simplify the interpretation, the incorrect
identification of such a series can lead to many subsequent errors as
illustrated by the previous example.

          ADDITIONAL FACTORS AFFECTING LRMS ANALYSIS ACCURACY

Factors other than errors in assignment and errors of omission that
affect analysis accuracy include the approach to interpretation taken
and the interaction with the IR analyst to produce a composite report
on the sample.  The experience of the spectroscopist, the time given
to measurement and interpretation, and the instrumentation employed
appeared to have little bearing on the results.  Contractors using
analog oscillograph recorders achieved essentially the same quality of
results as did those employing digital data recording systems.

Following the evaluation of the results of the test spectra and mixture
interpretations, the spectroscopists for each of the four contractors
were contacted and asked to describe in detail the steps each normally
takes in interpreting spectra.  A composite of the major factors
listed by the contractors that correlated, either positively or nega-
                                  61

-------
tively, with the more successful interpretations of the spectra and
mixtures was then prepared.  A successful interpretation was one in
which the majority of the compound and category assignments were
correctly made.  Some of the significant inputs are listed in Table 14
along with the analysts' evaluation of their utility in interpretation
expressed either as a "check mark" for agreement with the input shown
or as "no" for disagreement.  There was little agreement among the
analysts.  Of the inputs shown in Table 14, all but number 7 (finds
homologous series important) and number 9 (uses LC data to verify and
exclude) were positively correlated with successful interpretations.
All four analysts look for the molecular and at least a few significant
fragment ions that can be correlated with the molecular ion.  Several
look for the eight most abundant ions for a compound and attempt to
match these ions to those in reference spectra or lists in the refer-
ence Eight Peak Index of Mass Spectra (2).   Several analysts try to
account for all peaks above a given intensity and mass/charge value,
although this will vary from spectrum to spectrum.  Several look at
the peak intensity ratios of ions suspected of belonging to the same
compound to see if these ratios agree with those given in the reference
listing.  Several find homologous series very important in making
assignments.  Until this study, none of the LRMS analysts gave the IR
data equal weight in making compound/category assignments.  All of the
analysts make use of the LC fraction in which a suspected compound
elutes to verify the compound's presence, but three out of four of the
analysts also use the LC data to exclude the presence of compounds not
expected in a particular fraction.  However, the actual LC fraction in
which a compound elutes can be different than expected.  The results
of this study showed that the exclusion of compounds based on the LC
fraction is not always warranted.  Contractor A deleted the listing of
a fused nonalternate hydrocarbon (benzofluorene) from the results for
spectrum LC5a because the LC fraction was two fractions too late for
such compounds.  Contractor A also felt that spectra taken at low
ionizing voltage were helpful in identifying molecular ions.  The
feasibility of using low ionizing voltage is, however, dependent on
the type of mass spectrometer being used.  In general, an examination
of the inputs that lead to successful interpretations suggests that a
systematic approach and a thorough approach go hand-in-hand.

IR data also can improve the overall quality of the LRMS interpretation.
After reviewing the IR interpretation results of the test spectra and
samples, the LRMS analysts in many cases revised their observations,
often by adding new compounds or categories and by confirming the
presence of compounds already listed based on the LRMS alone.  Table 15
shows several examples of how the IR data were used to produce a more
accurate composite set of results.  LRMS analyst A originally and
correctly listed anthrone as part of mixture LC5 but when the analyst
could not find a reference spectrum of anthrone with which to verify
the LRMS observation, he deleted anthrone.   A more gratifying result
of the IR's contribution occurred when the same LRMS analyst changed
the assignment of hydroxyanthraquinone in mixture LC6 to that of the
correct assignment, fluorenone carboxylic acid.  This, in turn, lead
to the analyst finding fluorene carboxylic acid, which was also in the
                                  62

-------
sample.  LRMS analysts B and D initially found nothing interpretable
in spectrum LC6a and test mixture LC6, respectively, until the IR data
were consulted.  Then both were able to make correct assignments of
pyridone (B) and aromatic acids (D) in these cases.  Analyst B also
correctly changed the interpretation of LC5a from acridine to carbazole
and fluorenone to ethyl carbazole after reviewing the IR results.
Analyst C used the IR data to confirm carbazole in LC5a and to determine
that anthracene was a major component in LC4a.  Finally, analyst D
added the categories aromatic acids, esters, and ketones to the results
reported by LRMS alone in tar sample LC fraction 6.  There were many
other examples of how the IR aided the LRMS interpretation.  One
immediate consequence of this study is that at least one of the con-
tractors is now going to utilize the IR data on a more equal basis
with the LRMS data.

            RECOMMENDATIONS FOR IMPROVED LRMS DATA QUALITY

Based on the results from the analysis of the test spectra and samples
and discussions with the spectroscopists, several recommendations were
made to improve the quality of the compound and category assignments
made in Level 1 studies.  Major recommendations are:

     1.   The reference series Eight Peak Index of Mass Spectra (2)
          should be used to interpret and verify spectra.   Other
          references tend to be incomplete and/or inaccurate.

     2.   Intensity and mass/charge range criteria must be set to
          ensure a systematic and efficient interpretation.

     3.   The IR and LRMS analysts should first analyze their data
          independently and then consult with one another to produce a
          composite report.  The initial independent analyses will
          prevent biasing of the interpretation.

     4.   The IR and LRMS data should be used as primary inputs to
          spectral interpretations.  Other inputs, such as the LC
          fraction, the sample source, and previously generated data
          on similar sources, should be used with caution as aids in
          spectral interpretation.

     5.   Polyfunctional compounds like fluorenone carboxylic acid and
          phenoxybenzoic acid should be listed in both possible cate-
          gories to avoid confusion and because both functional groups
          may be active from a health effects standpoint.

     6.   Suggestions for utilization of the information collected in
          this study include the preparation of guidelines for mass
          spectral and IR interpretation for analysts beginning Level 1
          work, sets of worked out examples of spectra, workshops in
          which analysts can exchange information, and spectral inter-
          pretation tests.
                                    63

-------
                              REFERENCES

1.   Lentzen, D. E., D. E.  Wagoner,  E.  D.  Estes,  and W.  F.  Gutknecht,
     "IERL-RTP Procedures Manual:   Level 1 Environmental Assessment
     (Second Edition)," EPA-600/7-78-201 (1978).

2.   Imperial Chemical Industry,  Ltd.,  Eight Peak Index  of  Mass  Spectra,
     First Edition, Mass Spectrometry Data Center,  Alder Maston,
     Reading, United Kingdom, 1970.

                            ACKNOWLEDGMENT

The work described in this paper was performed under Contract No.  68-
02-2688, Task 101 with the Industrial Environmental Research Laboratory
of the U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina.  Special acknowledgment is given to the  EPA Project
Officer, Dr. Larry D. Johnson, and to the organizations  that partici-
pated in this study, Arthur D. Little,  Inc., Battelle Columbus Labora-
tories, Southern Research Institute, and TRW Defense and Space Systems
Group.
                                  64

-------
                                   TABLE 1
                   Mixtures Used to Prepare Test Spectra*

Mixture LC4a                                 Mixture LC4b
naphthalene (6.2)                             naphthalene (35.0)
anthracene (37.1)                             anthracene (23.3)
phenanthrene (6.2)                            phenanthrene (23.4)
9-methyl anthracene (25.1)                     9-methyl anthracene (6.2)
9,10-dimethyl anthracene (25.4)                9,10-dimethyl anthracene (6.2)
                                             2,3-dimethyl naphthalene (6.3)

Mixture LC5a                                 Mixture LC5b
fluorene (9.1)                                 7,8-benzoquinoline (16.1)
2,3-benzofluorene (4.9)                         phenanthradine (30.5)
N-ethyl carbazole (52.1)                        1,10-phenanthroline (15.5)
2-methyl indole (20.0)                          phenazine (30.2)
2,3-dimethyl indole (4.9)                        anthrone (7.7)

Mixture LC6a               p                 Mixture LC6b
2-hydroxypyridine (67.9)                        9-fluorenone (22.6)
2,6-dimethyl phenol (16.5)                      o-phenoxybenzoic acid (22.5)
o-o'-biphenol (15.6)                            1 -fluorene carboxylic acid (22.2)
                                             9-fluorenone-4-carboxylic acid (21.7)
                                             2,6-dimethyl phenol (11.2)

'Weight percent of each component in mixture shown in parentheses.
                                   TABLE 2
                                 Test Mixture


Mixture LC5                                  Mixture LC6
fluorene (4.5)                                  9-fluorenone (26.1)
2,3-benzofluorene (4.1)                         1-fluorene carboxylic acid (11.1)
7,8-benzoquinoline (19.4)                       9-fluorenone-4-carboxylic acid (41.5)
1,10-phenanthroline (31.0)                      2-methyl-1-tetralone (5.2)
phenazine (16.8)                               2-t-butyl phenol (6.3)
2,6-dimethyl quinoline (16.8)                    2-methyl-1,4-naphthoquinone (9.8)
anthrone (7.5)

   Coal Gasification Tar Sample; Compounds Expected in Each Fraction*

LC4                         LC5                         LC6
dibenzofuran                  overlap from                  benzidine
carbazole                    LC4 and LC6                 alkyl pyridines
                                                          quinoline
                                                          alkyl quinolines
                                                          alkyl benzoquinolines
                                                          acridine
                                                          phenol
                                                          cresols
                                                          xylenols
                                                          o-isopropylphenol
                                                          trimethylphenol
'Limited analysis based on GC measurements.

                                      65

-------
                                TABLES
              Accuracy of Structure Identification Using IR*
                                        Contractor

Test
Spectra
Test
Samples
Tar
Samples
A
85%

55%

80%

B
65%

55%

70%

C
85%

80%

65%

D
60%

80%

85%

* All percentages shown have an uncertainty of about ± 10 percent.

                               TABLE 4
         Selected Examples of Errors in Assignment of IR Signals

                  Apparently                   Apparently
Signal, cm ~1
3052
1805
1600
1600
1480, 1470
1510
1245
1190,1160
Erroneous Assignment
OH
Carbonyl
Unsaturated CH
$C = 0
Unsaturated CH
C-N
$C = 0
C-O-C
Correct Assignment
Unsaturated CH
Overtone or summation band
Aromatic C = C
Aromatic C = C
CH3
Aromatic C = C
C - O of aromatic alcohol
C - O of alcohol
                               TABLE 5
   Selected Examples of Reasonable, Though Erroneous, Assignments
                                             Apparently
Signal, cm ~1      Erroneous Assignment         Correct Assignment
1715,1690        Carbonyl                    Overtone or summation bands
1710              C = Oof ester                C = Oofketone
1710              C = O of acid                 C = O of ketone
1280,1240        C - O of aromatic alcohol       CH3
1235              C - O of aromatic alcohol       C - N of carbazole
1300              C - N of aryl amine            OH/C-O coupling
1310              C - N of secondary amine      C-C(=O)-C of ketone
                                  66

-------
                                TABLE 6
    Wavenumber Ranges Indicated for Different Structures as Found in
                   Three Different Reference Sources

                                     Reference Source
Structure
OH bend of
alcohol
N. B. Colthup*
1460-1260 cm-1
Dow Chemical5
—
Nakanishi0
1500-1200 cm-1
C - O stretch
 of alcohol
NH bend of
 primary amine
Aromatic C = C

C-O-Cof
 aliphatic ether
C = N stretch
1290-1000 cm
                               -1
1260-980 cm
                               -1
1650-1550 cm-1
1540-1450 cm-1
1175-1055 cm
1700-1590 cm
            -1
            -1
1665-1615 cm-1
1515-1370 cm ~1
1140-1090 cm~1
1630-1590 cm~1
1200-1000 cm
                               -1
1640-1560 cm-1     1665-1625 cm-1     1640-1560 cm-1
1600-1450 cm-1

1200-1040 cm~1

1690-1640 cm~1
aN. B. Colthup, J. Opt. Soc. Am., 40:397,1950.
bIR spectral assignment chart, Dow Chemical Company, Midland, Michigan.
°Koji Nakanishi, Practical Infrared Absorption Spectroscopy, San Francisco: Holden-Day, Inc.,
 1962.
                                TABLE?
    Selected Examples of Unassigned Signals Found in Various Spectra
Signal Not Assigned, cm
                     -1
                       RTI Assignment
2930,2850
1680,1670
1430,1240
1450,1465,1390,1380,1364
1240,1160,1100
1015
                       Saturated CH
                       Carbonyl
                       Coupling of acid OH/C- O groups
                       CH3
                       C - O of alcohol
                       C - O - C of ether
                                     67

-------
 Wavelength, fim
 5           6
                                  9   10
12     15
            20
2000
  Wavenumber, cm
                   1500
                                     1000
                                                        500
       Figure 1a  IR spectrum, LC6, synthetic mixture.
 2000
                   1500
                                      1000
                                                         500
   Wavenumber, cm
             -1
         Figure 1 b  IR spectrum, LC6, coal tar sample.
                                 68

-------
                               TABLES
                 Incomplete Utilization of Available Data

                                          Complementary Data
Signal, cm ~1        Assignment             Not Used
1737
1690
C = O of ester
C = Oof acid
1160 cm-1, 1100 cm-1
(C - O - C of ester)
3300 cm-1 (OH)
                                          1440 cm -1 (OH/C - O coupling)
                                          940-920 cm ~1 (OH bending)

1590                NHofamine            1235cm-1,1180cm-1
                                          (C - N of primary amine)

1240                C-O-C of ester        No signal at 1735cm-1
                                          (C = O of ester)
TABLE 9
Compounds in Test Spectrum LC4b Coi
Weight
Compound* MW %
Naphthalene
Anthracene
9-Methyl anthracene
Phenanthrene
9,10-Dimethyl anthracene
2,3-Di methyl naphthalene
128
178
192
178
206
156
35.0
23.3
5.9
23.4
6.2
6.3
rrectly Identified by LRMS
Contractors
A
10
100
10
100
10
10
B
NR
t
t
t
t
NR
C
10
100
10
100
10
10
D
100
100
100
100
10
10
 Credit given for assignment of isomers of LC4b compounds.
^Compound reported but intensity not given.
NR = not reported.
                                TABLE 10
      Compounds in Test Spectrum LC5a Correctly Identified by LRMS

                                   \i/~:_u*             Contractors
Compound
Fluorene
2,3-Benzofluorene
N-ethyl carbazole
2-Methyl indole
Carbazole
2,3-Di methyl indole
MW
166
216
195
131
167
145

9.1
4.9
52.1
9.0
20.0
4.9
A
NR
NR
NR
10
100
10
B
NR
NR
NR
NR
NR
NR
C
NR
10
100
10
100
10
D
NR
NR
100
10
100
10
 NR = not reported.

                                  69

-------
                                                                                                  t-S
                                                                                                     CM
§
o
00
o
to
                                         I
                                        o
                                                                                                    o
                                                                                                    •00
                                                                                                    (VI
                                                                                                         (0

                                                                                                         o
                                                                                                         i    E
                                                                                                     o
                                                                                                    -CN
                                                                                                              CM

                                                                                                               £


                                                                                                              I
                                                                                                   —o
o
CM
                                               A};suaiu|
                                                         70

-------
                              TABLE 11
     Compounds in Test Spectrum LC6b Correctly Identified by LRMS

                                 uf~i«ut            Contractors
Compound
9-Fluorenone
2,6-Dimethyl phenol
o-Phenoxy benzoic acid
1-FluoreneCOOH
9-Fluorenone4-COOH
MW
180
122
214
210
224

22.6
11.2
22.5
22.2
21.7
A
NR
NR
NR
NR
NR
B
NR
NR
NR
t
t
C
NR
NR
10
10
100
D
NR
NR
NR
NR
NR
"l"Compound reported but intensity not given.
NR = not reported.
                              TABLE 12
      Compounds in Test Mixture LC6 Correctly Identified by LRMS


                                 ... . ut            Contractors
Compound
9-Fluorenone
1-FluoreneCOOH
9-Fluorenone-4-COOH
2-Methyl-1-tetralone
2-t-Butyl phenol
2-Methyl-1,4-
Naphthoquinone
MW
180
210
224
160
150
172

%
26.1
11.1
41.5
5.2
6.3
9.8

A
100
NR
NR
NR
NR
NR

B
100
NR
40
NR
NR
NR

C
10
10
100
10
10
10

D
10
NR
NR
NR
10
NR

NR = not reported.
                                   71

-------
                                TABLE 13
            Comparison of Categories In Test Mixture LC6 and
                 Test Spectrum LC6b Identified by LRMS
                                                      Contractors

Mixture


Spectrum


Category
LC6 Ketone*
Acid*
Phenol
LC6b Ketone*
Acid*
Phenol
Weight %
41.1
52.5
6.3
22.6
43.9
11.2
A
200
NR
NR
NR
NR
100
B
100
40
NR
NR
t
NR
C
30
110
10
NR
110
NR
D
10
NR
10
111
NR
NR
 9-Fluorenone-4-carboxylic acid was placed in acid category.
t Compound reported but intensity not given.
NR = not reported.


                                TABLE 14
         Low Resolution Mass Spectral Interpretation Inputs and
                   How They Are Used By Contractors

	A	B	C	D

The analyst:

 1. Looks for molecular ions              s         s          s          s

 2. Looks for some fragment ions         s         ^          **          *•

 3. Looks for eight most                 ^         no         ^          no
   abundant ions
4. Uses eight-peak indices often
5. Accounts for all peaks above
given intensity and m/e value
6. Uses peak intensity ratios
7. Finds homologous series
very important
8. Gives IR data equal weight
9. Uses LC data to verify
& exclude
10. Uses LC data to verify only
s no
• no
s no
s no
no no
"
no no
s no
.x no
no s
no ,x
no no
no s
s no
»x = agrees with input shown.
no = does not agree with input shown.
                                        72

-------
                           TABLE 15
Examples of How IR Data Were Used to Verify and/or Revise LRMS Results
Contractor
A
B
C
D
Sample
LC5
LC6
LC6a
LC5a
LC4a
LC5a
LC6
TarLCG
Before IR
anthrone
hydroxy-
anthraquinone
nothing
acridine
fluorenone
anthracene
carbazole
nothing
only heterocyclic
N compounds
After IR
deleted anthrone
9-fluorenone-4-COOH,
added fluorene-COOH
pyridone
carbazole
ethyl carbazole
anthracene- major
component
carbazole confirmed
aromatic acids
heterocyclic N
compounds, aromatic
acids, esters, ketones
                                  73

-------
                 CHARACTERIZATION OF COAL GASIFICATION
               BY-PRODUCTS AND AMBIENT AIR SAMPLES FROM
              A LURGI GASIFICATION FACILITY BY SELECTIVE
                     DETECTOR GAS CHROMATOGRAPHY

                K. W. Lee, D. S.  Lewis, C. H. Williams
                          and K.  J. Bombaugh
                          Radian Corporation
                               ABSTRACT

Samples of primary by-products produced by the Lurgi Gasification Plant
in the Kosovo Region of Yugoslavia were taken for organic characterization.
Ambient air around the plant was monitored for particulates and organic
vapors.

Primary by-products, organic vapors, and particulate-borne organics were
analyzed for sulfur- and nitrogen-containing organic compounds using
selective-detector gas chromatography.  Selective detection revealed a
positive correlation between sulfur- and nitrogen-containing compounds
in the primary by-products and the ambient air samples.  Relative quanti-
fication and tentative identification of the major sulfur- and nitrogen-
containing compounds was established by comparison to standards and by
peak enrichment techniques.  These identifications were confirmed by
GC/MS.

                             INTRODUCTION

An ambient air monitoring program was conducted at the Kosovo industrial
complex in Yugoslavia.  This complex contained a Lurgi gasification plant
that produces medium BTU gas from lignite.  This site was selected because
Lurgi technology is proposed for indirect liquefaction and SNG plants in
the United States.  The main problem within this program was to correlate
the emissions from the gasification plant to the components found in the
ambient air samples.  The samples collected were particulates on HiVol
quartz fiber filters and organic vapors on Tenax resin.

A phased analytical approach was developed to solve the correlation problem.
The analytical techniques in this program were verified before analyses
were begun.  Extraction and thermal desorption efficiencies were determined
and reproducibility verified.  Vapor breakthrough on the Tenax resin traps
was determined for various organic compounds.

Within 'this program the analytical procedures involving gas chromatography
were used for screening of samples and preliminary identification of or-
ganic compounds.  Retention times and spiking with known compounds were
used for tentative identifications.  A Hall detector in the nitrogen and
sulfur specific modes was used to profile the components in both the gasi-
fier by-products and air samples.  These profiles were compared in order
to provide the correlation between the gasification plant emission sources
and the air samples.

                                     74

-------
                                 SAMPLING

HiVol Filters

Ambient air samples for participates were taken on HiVol quartz filters.
The schematic diagram shown in Figure 1 illustrates where the filter is
located in the sampling apparatus.  Typical parameters for particulate
sampling were:

     Sampling rate = Im3/min.;

     Sample volume = 1400m3/24 hr. (25°C, 1 atm.).

The filters were weighed before and after sampling for particulate loading.
However, for organic analysis, the filter and particulates were stored in
organic-free aluminum foil for organic analysis.

Tenax Samples

The following is a description of the Tenax samples.  Figure 1 illustrates
the location of the Tenax traps in the sampling apparatus.  The traps con-
tained ^3g of 60/80 mesh Tenax resin.  The second trap was used as a back-
up to indicate breakthrough of compounds from the first trap.  Typical
sampling parameters for the Tenax traps were:

     Sampling rate = 4&/min.;

     Sample volume = 5.8m3/24 hr. (25°C, 1 atm.).

The Tenax traps after sampling were sealed with Teflon end caps in the
field.  When the samples arrived at the laboratory, the Tenax resin was
transferred to a clean, 40 m£ glass vial with a Teflon-lined cap.   The
weight of the resin was obtained during the transfer.

                           TREATMENT OF SAMPLES

A phased approach was used to select and analyze the ambient air samples.
This approach is outlined as follows:

     1)  Secure and identify all samples;

     2)  Interpret meteorological data and select samples;

     3)  Selectively screen several upwind, downwind and
         blank samples for organics using a gas chromato-
         graph (GC) equipped with a flame ionization
         detector (FID);

     4)  Using an FID and several selective detectors,
         obtain GC profiles of selected by-products from
         the gasification plant;
                                   75

-------
FIGURE  1   CONFIGURATION  FOR COLLECTING AEROSOL SAMPLES
            FOR ORGANIC ANALYSIS
        FILTER SHIELD
                                                  HIGH VOLUME GLASS
                                                     FIBER FILTER
   HI-VOL
  SAMPLER
FRONTTENAX
VAPOR TRAP
                                                            REAR TENAX
                                                            VAPOR TRAP
                   TO PUMP, METER, ETC
                                            TO PUMP, METER, ETC.
                                   76

-------
     5)  From screening data, carefully select samples to
         be profiled and quantified by gas chromatography
         using an FID and Hall detector (sulfur and nitro-
         gen modes);

     6)  Compare profiles and establish correlations if
         possible;

     7)  Input all data with tentative identifications to
         the mass spectroscopist for identification, con-
         firmation, and quantitation of major compounds.

In the final treatment of data, the concentrations of compounds in the air
samples were plotted versus the percent downwind of that sample.  Percent
downwind was defined as the percent of the total sampling time in which
the sampling site was within a 45  corridor of the downwind direction.
From the plots, a correlation coefficient can be calculated.

An important aspect of this program was the verification of methods.  The
methods, verifications, and quality control are discussed in the following
sections.  After these sections, the application of these methods to the
Lurgi gasification facility are outlined.

Extraction of Samples

Both Tenax and HiVol filter samples were Soxhlet extracted for 24 hours.
A 1.0 to 1.5 gram portion of a Tenax sample was extracted.  Cyclopentane
was the solvent used with the Tenax samples, and dichloromethane was used
for the HiVol filters.  Since the dichloromethane produces noise in the
Hall detector, the solvent was exchanged with cyclopentane before injection.

Thermal Desorption of the Tenax Samples

Thermal desorption was used as a second, independent method for separating
the organic vapors collected on the Tenax resin, prior to the GC-FID pro-
filing.  Approximately 0.05g aliquots of the homogenized Tenax resin were
placed in a thermal desorption injector mounted on a GC-FID.  The adsorbed
vapors were thermally desorbed from the Tenax and collected in a pre-column
trap held at dry ice/acetone temperature.   At the completion of desorption,
the coolant was removed from the cold trap and the GC oven program was
initiated.

Gas Chromatographic Conditions

Table 1 contains a list of conditions and materials used for the various
detectors on the gas chromatographs.  Note that similar chromatographic
conditions have been used so that chromatographic profiles can be compared.

Two packed columns were necessary for this work.  For most hydrocarbons
and sulfur-containing compounds the OV101 phase worked well.  The polar
nitrogen-containing species tailed on this phase; therefore, a 10% Apiezon
L plus 2% KOH mixed phase was used.
                                    77

-------
                                 TABLE  1

                Conditions Used with  the Three Different
                             Detectors  (Modes)
Detector (Mode) :
Condition
Carrier Flow
Inj ector Temperature
Detector Temperature
(Reactor)
Flame lonization
Detector (FID)
Na , 30 ml/min.
250°C
300 °C
Hall-Detector
Sulfur (HECD-S)
He, 35 ml/min.
250°C
225°C
(850°C)
Hall Detector
Nitrogen (HECD-N)
He, 35 ml/min.
250°C
225°C
(850°C)
Oven Program
Column
50°C for 5 min;    50°C for 5 min;
20°/min to 250°C,  to 250°C, hold
hold for 60 min.   for 60 min.
                   50°C for 5 min;
                   to 250°C, hold for
                   60 min.
10% OV101 on 80/
100 mesh Supelco-
port; 10'x2 mm
ID glass column
10% OV101 on 80/   10% Apieson L, 2%
100 mesh Supelco-  KOH on 80/100 mesh
port; 10'x2 mm ID  Chromosorb, W-AW,
glass column       10'x2 mm ID glass
                   column
                                    78

-------
Tentative Identification

Two techniques were used to assign tentative identifications to organic
compounds in the samples.  The first involved spiking known compounds into
samples for peak enrichment and identification.  This technique was
successful in the tentative identification of benzene, toluene, xylenes,
several of the multiple ring aromatics and four of the alkyl substituted
thiophenes and benzothiophene.  The second technique was a simple plot of
boiling point versus retention time of normal alkanes.  Using this techni-
que to predict the boiling point of unknown compounds coupled with the
knowledge of what to expect as emissions from the gasifier, tentative
identifications were made.  This technique was especially successful with
non-polar compounds.
                           METHODS VERIFICATION

Both thermal desorption and Soxhlet extraction were tested with model com-
pounds spiked onto both the fiber filters and the Tenax resin.  The spiked
samples provided data on reproducibility and efficiency of recovery.
Ambient air Tenax samples provided data on breakthrough during sampling as
well as reproducibility of the thermal desorption technique.

Recoveries for Solvent Extraction and Thermal Desorption

Blank glass fiber filters and blank Tenax resin aliquots were spiked with
five aromatic compounds before solvent extraction:  toluene, naphthalene,
anthracene, 1,2-benzanthracene and benzo(a)pyrene.  The recoveries of
these compounds were measured by GC-FID analysis of the extracts and the
spiking solution.  All measured recoveries fell in the range of 80-120%,
except for toluene.  The toluene recovery was generally in the range of
20-70%, and was probably due to losses occurring in the solvent evapora-
tion step.

Blank Tenax resin was spiked with benzene, toluene, naphthalene, anthra-
cene and 1,2-benzanthracene prior to thermal desorption.  The recoveries
for these compounds by thermal desorption and GC-FID were in the range
of 90-110%.

Vapor Breakthrough on the Tenax During Sampling

The front and rear Tenax traps were screened for sample vapor breakthrough
by GC-FID.  Equal quantities of benzene and tolune were found on the
front and rear traps indicating complete vapor breakthrough for these two
species under the conditions used in this sampling.  For the C2~alkyl-ben-
zenes,  the quantity on the rear trap was only about 15% of that on the
front trap and for heavier species, the quantity on the rear trap was
negligible.  Thus, breakthrough was a problem for benzene and toluene
and compounds of similar or higher volatility.
                                    79

-------
Reproducibility of the Thermal Desorption Technique

The precision of the thermal desorption method was determined by quanti-
fying thirteen corresponding peaks in a duplicate measurement by GC-FID
of weighed Tenax sample aliquots.   The peaks selected included benzene,
toluene, the C2~alkyl benzenes and naphthalene plus eight other prominent
peaks not identified at the time.   The reproducibility in this duplicate
determination was good, with an average percent standard deviation for
all thirteen components of 8.7%.

                               APPLICATION

The phased scheme was applied to an ambient air sampling program involving
a Lurgi coal gasification facility.  The Hall detector scans in both the
sulfur and nitrogen modes will be discussed as specific examples of results
from this type of approach.  The details of the methods and a complete
discussion of the results will appear in the final contract report to the
Environmental Protection Agency.

Selection of Samples

The phased approach outlined in the Treatment of Samples section was
followed.  After identifying all the samples, they were categorized as to
upwind, downwind and crosswind samples from the meteorological data.  For
screening purposes, two downwind Tenax samples with the highest percent
downwind during the sampling period were chosen.  These were coupled with
their corresponding upwind samples, one low downwind sample (crosswind)
and two blanks.  The flame ionization detector (FID) scans of these
samples confirmed the choices made as far as organic loading was concerned.

Analyses

After screening, approximately Ig of each of the selected Tenax samples
was Soxhlet extracted for 24 hours with cyclopentane.  The solvent was
reduced in volume to 0.5 ml in a Kuderna-Danish apparatus using a 3-ball
Synder column.  The samples along with standards and by-products were pro-
filed on gas chromatographs equipped with FID, Hall-sulfur and Hall-nitro-
gen detectors.  The conditions for these analyses were discussed in the
Treatment of Sample section.

Results:  Sulfur Species

Two representative profiles of the sulfur-containing species are illustrated
in Figure 2.  The chromatograms show the distribution of sulfur species in
the "medium oil" by-products from the gasification plant and a typical
downwind air sample.  The retention times are slightly different for the
compounds due to variations in the initial temperature of the gas chroma-
tograph, but the profiles have a high degree of correlation.

The two large peaks at about 6.5 minutes in the "medium oil" were tenta-
tively identified by spiking.  The compounds, 2- and 3-methylthiophene,
were spiked into the sample and the peaks which increased relative to


                                    80

-------
FIGURE 2   GC-HECD SULFUR COMPOUND PROFILES FOR A DOWNWIND
           TENAX VAPOR TRAP EXTRACT AND FOR MEDIUM OIL
                                          DOWNWIND  SAMPLE
                 10
                                     20
                                       MEDIUM OIL
                  10
               Retention Time (min)
                              81

-------
the others were identified as the spiked compounds.  This technique was
also used for thiophene (RT ^ 3 min.), 2,4-dimethylthiophene (RT ^ 9 min.),
ethylthiophene (RT % 9 min.), and benzothiophene (RT ^ 13 min.).  Thio-
phene was not detected in the Tenax sample,  perhaps because of its high
volatility and subsequent low collection efficiency on Tenax resin or loss
in solvent evaporation.

The resin and/or solvent contained an interference which masked part of
the profile around 10 minutes.  The upwind site Tenax sample and the blank
sulfur profiles are given in Figure 3 and are essentially identical.  That
is, no sulfur species were detected in the upwind vapor sample correspond-
ing to the downwind sample.  We concluded that the profiles of sulfur
species exhibit a positive correlation between the downwind air samples
and the gasification plant.

Results:  Nitrogen Species

Very similar results were found with the nitrogen components in the "medium
oil" by-product and the air samples.  In Figure 4, the profiles of "medium
oil" and a downwind sample are illustrated.   The profiles are similar;
however, the large, early peak in "medium oil" at 6.7 minutes is not pre-
sent in the air sample.  Again, this may indicate that the more volatile
compounds had low collection efficiencies on the Tenax or they were lost
during concentration of the solvent.

Most of the nitrogen compounds in the "medium oil" and the air samples
appeared to be alkyl substituted pyridine(s) or quinoline(s) from spiking
and boiling point identifications.  One of the peaks at about 15.7 minutes
retention time and another at about 18.5 minutes were also seen in the
upwind samples but not in the blank resin (Figure 5).  The peaks were
either low in concentration in the "medium oil" or not in the "medium oil"
at all.  This indicated that the source of these compounds was not the
gasification plant.

Results:  GC/MS

Most compounds tentatively identified by GC were confirmed by GC/MS.  How-
ever, nitrogen compounds in  the "medium oil" were not confirmed initially
because of the carbon  13 abundance in carbon-containing analogs.  An acidic,
aqueous extraction of  "medium oil" and a neutral back-extraction separated
the carbon-containing  analogs from the nitrogen-containing compounds.

Scans with an HECD-N of the basic fraction from "medium oil" gave an iden-
tical chromatogram to direct injection.  Therefore, no major nitrogen-con-
taining compounds were lost in the extraction process.  This also indicated
that all the major chromatographable compounds which contain nitrogen are
basic.  GC/MS was able to confirm all tentatively identified peaks in the
"medium oil" using the basic fraction.
                                    82

-------
FIGURE 3   GC-HECD  SULFUR COMPOUND PROFILES FOR AN UPWIND
           AND A BLANK TENAX VAPOR TRAP EXTRACT
                                   UPWIND SAMPLE
                              10
                                          15
                                                      20
                  Retention  Time  (ruin)
                                      BLANK
                              10
                                          15
                                                     20
                                83

-------
FIGURE 4   GC-HECD NITROGEN  COMPOUND PROFILES FOR A DOWNWIND
           TENAX VAPOR TRAP  EXTRACT AND FOR MEDIUM OIL
                                         DOWNWIND SAMPLE
                     10
                               15
                                         20
                                                    25
                                           MEDIUM OIL
                              15
                                         20
                                                   25
                     Retention Time  (min)
                                84

-------
FIGURE 5   GC-HECD NITROGEN  COMPOUND PROFILES FOR AM UPWIND
           AND A BLANK  TENAX VAPOR TRAP EXTRACT
                                           UPWIND  SAMPLE
                  10
                             15
                                        20
                                           BLANK SAMPLE
                  10
                             15
                                         20
                                                     25
                    Retention  Time (rain)
                            85

-------
                               CONCLUSIONS

The following conclusions were obtained from the application of the phased
analytical approach:

     1)  Positive correlations between the ambient air samples
         and the gasification plant products were established
         from the nitrogen and sulfur profiles.

     2)  The phased approach was effective in characterizing
         samples containing a wide range of organic compounds
         including hydrocarbons, sulfur compounds and nitrogen
         compounds.

The phased approach should be applicable to other processes where selective
detection can be applied.  By using selective detectors,  GC/MS time can
be more effectively used with a decrease in analytical costs.

                             ACKNOWLEDGMENTS

This program was funded by the United States Environmental Protection
Agency under contract numbers 68-02-2608 (Task 42) and 68-02-3137.  Ronald
K. Paterson (EPA) was most helpful during all phases of this work.  Without
the support of T. Kelly Janes (EPA) and William J. Rhodes (EPA),  the
analyses of the gasification by-products would not have been possible.
                                     86

-------
          CHARACTERIZATION OF PROCESS STREAMS FROM LIQUEFACTION
                   OF LOW-RANK COAL WITH SYNTHESIS GAS

         Bruce W. Farnum, Sylvia A. Farnum and Curtis L. Knudson
                  Grand Forks Energy Technology Center
                        U.S. Department of Energy
                    Box 8213, Grand Forks, ND  58202
                                ABSTRACT

Beulah (ND) seam lignite was liquefied with 1:1 carbon monoxide-hydrogen
at  460°  C and 27.5 MPa  in a 2.3 kg  coal/hour  continuous  flow reactor.
Process streams obtained  included C  to C, hydrocarbon gases, water ef-
fluent, light  oils,  vacuum distillate, THF soluble  vacuum bottoms,  and
THF  insoluble  fraction.   Analysis of  tail  gases  was  by on-line  GC.
Light  oils  (volatile at  27.5  MPa and 300° C)  were  separated into phe-
nolic, basic,  hydroaromatic  and hydrocarbon  fractions by solvent  ex-
traction.   These were  analyzed by 200 MHz  H and 50 MHz   C NMR, GC-MS,
low voltage MS, IR,  UV,  fourth derivative UV,  and  capillary  GC.  Water
inherent  in  the lignite  formed  a separate layer in the light oil col-
lector.   The water  was characterized for pH, NH«,  alkalinity, cyanide,
sulfide,   TOC,  phenol and  cresol  content. Prior  to  vacuum distillation
the bulk  product  was  sampled  and  analyzed  by NMR, % water, % ash,  %
CHNS, oxygen by neutron activation, THF solubility,  viscosity, molecular
weight distribution of  the THF solubles by gel permeation HPLC, and ash
analysis  by  x-ray fluorescence. Major  inorganic  components were deter-
mined by  inductive  coupled argon plasma spectrometry and atomic absorp-
tion  after wet ashing.  Vacuum distillate  was  characterized  by column
chromatography  on  alumina,  solvent extraction,  HPLC,  NMR,   IR,  UV  and
LVMS.  Solids  deposited on  the reactor  walls  (coke)  were examined  by
optical microscopy,  scanning and analyzing electron microscopy,  and  x-
ray fluorescence.

                              INTRODUCTION

Pioneering work on the liquefaction of a finely ground slurry of lignite
utilizing  synthesis  gas (1:1 carbon  monoxide-hydrogen)  was  carried  out
by Appell  (1)  and  co-workers at the Pittsburgh Energy Technology Center
using a rocking bomb  autoclave.  A continuous process liquefaction unit
was also  operated at  Pittsburgh  (2).   In 1975 the study of liquefaction
of  low-rank  coal  was  transferred to the  Grand Forks  Energy  Technology
Center (GFETC).  A hot charged time sampled stirred autoclave system was
developed and used to study the reaction kinetics of the process (3).   A
2.3 kg coal  per hour continuous process liquefaction unit was then con-
structed  at  GFETC,  and  is  presently  being used  in  the  study of lique-
faction of low-rank coal by adaptations of various current processes (4,
5).

Extensive  deposits  of lignite  are  found  in  the  Northern  Great Plains
Province  (including  western North Dakota, eastern Montana and southern
Saskatchewan) and the  Gulf Coast Province (a narrow belt  across Texas,
Louisiana and Arkansas).   The  lignites differ from higher rank coals by
containing high moisture,  low  sulfur  and alkaline inorganic components.
                                    87

-------
Much of the  inorganic  matter consists of alkali and  alkaline earth ca-
tions bound as humate salts at ion exchangeable sites on the coal struc-
ture.  Minerals  such as pyrite,  clay and  silica  are  finely dispersed
throughout the organic matrix.  Lignites also have  a much higher organic
oxygen content than higher rank coals.  The unique  properties of lignite
are believed  to  affect its liquefaction behavior so  that  processes de-
veloped for bituminous coal are not directly applicable to lignite with-
out further research.

                           REACTION CONDITIONS

Current emphasis in coal liquefaction is toward formation of distillable
products.   Experiments with the continuous reactor at GFETC have includ-
ed recycling  of  the  heavy ends including their mineral content, and re-
moval  of  the  volatile  oils  as  products  of  each  pass.   A  4.55  liter
stirred autoclave  reactor  was operated  at 460°C and  27.5 MPa   .  The
pasting solvent  for  the first pass was  redistilled  anthracene oil (IBP
296° C @  1.3  KPa),  a  phenol-free  distillate by-product  of bituminous
coke production.  Slurry was prepared for each pass through the reactor
by adding 30% by weight,  pulverized  (100% minus  60  mesh) as-received
(ca.  30%  moisture)   lignite  to the  heavy product (BP  300°C @ 27.5 MPa)
from the  previous  pass  including the mineral matter  and  any unreacted
lignite.  The feed  coal  was a Beulah,  ND  seam  lignite  with unusually
high mineral  content.   Table 1 contains the  analyses  of  the feed coal.
Slurry was  fed to  a 4.55 liter autoclave acting as a continuous stirred
tank reactor  (CSTR)  to obtain a nominal residence time of one hour.  CO
and H?  gas  (1:1) were added  at 14.2  1/min.   Further details of the re-
actor conditions have been presented by Farnum et.  al.  (6).

The  yields  of various process streams  produced during  product slurry
recycle operation are illustrated in Figure  1 as a function of pass num-
ber.  The overall yields during 34 passes were: C..-C, hydrocarbons, 19%;
high pressure volatile  oils,  19%;  vacuum  distillate,  36%; THF soluble
vacuum bottoms,  9%;  and THF insoluble organic polymers, 11%.  These were
calculated  on a  moisture- and mineral matter-free coal basis.  The pro-
cess was  upset  during  pass  16 by  plugging  of  the  reactor with "coke"
consisting  of mineral  particles dispersed in an organic polymer matrix,
causing the  yield  of organic THF insolubles exiting the unit to drop to
zero during  the  pass.   The unit was  restarted using the heavy oils pro-
duct as pasting  solvent, and  operation was  continued through 34 passes.
Figure  2  illustrates the approach to  constant  composition of the light
oils  in  terms of  aromatic-to-aliphatic  proton  ratio  measured  both by
proton NMR and infrared  absorbance. Samples  from the thirtieth pass were
selected  for  detailed  chemical characterization studies  as  representa-
tive of lignite-derived products.

                           WATER EFFLUENT

Lignite seams often serve as  aquafiers  and  contain 35  to  40% water when
freshly mined.  After  handling  and  pulverizing, the  feed coal in this
experiment  was  29%  water  by  weight.   During liquefaction the moisture
makes  an  undesirable contribution to  total  pressure within the reactor,
but  is available  for conversion of  CO to  hydrogen via  the water gas

                                     88

-------
shift reaction.  Thirty-one percent  of the water in the feed slurry was
recovered as a layer which separated from the light oils which condensed
from the  gas phase  in  the product  receiver.   This  effluent  water was
characterized by standard  analytical methods for waste  water  and stan-
dard EPA  methods  (7, 8).   The average  of analyses  of  four samples of
condensate  water  produced  under  lined-out  conditions  is   reported  in
Table 2.
                               LIGHT OILS

The light oils which were condensed from the product in the gas phase at
300°C and  27.5 MPa  were  fractionated by  solvent extraction  into  four
chemical classes using the method of Fruchter et al. (9) modified by the
use of  pentane rather  than isooctane.  The  use of methylene chloride
following pentane extraction was also explored and found to give simpler
mixtures in the  case of the phenolic fraction.  Characterization of the
light oils by Farnum et al. is discussed in a forthcoming preprint (10).
The fractionation scheme is illustrated in Figure 3.

The light  oil  fractions were separated by capillary GC using a 50 meter
OV-101  glass  column.   The  number  of  components  detected is  listed in
Table 3.   There  are  less than 85 components above 1%,  allowing for some
overlap of components among fractions.

The ultraviolet  spectrum  of the light oil from the thirtieth pass (lab-
eled LO 28-14) is displayed in Figure 4.  Subtraction of the spectrum of
the hydroaromatic  fraction  (HA) yielded a  spectrum displaying many of
the features of that of the phenolic fraction.  The fourth derivative UV
spectra displayed in Figure 5 show a maximum at 275 nm which is assigned
to phenol, and a maximum at 268 nm assigned to the cresols.  The maximum
at 250 nm  in  the HA fraction is assigned  to phenanthrene.  Fourth der-
ivative UV is useful in the analysis of complex mixtures since the maxi-
ma in  the fourth  derivative correspond to the maxima  in the underiva-
tized UV.  There  is  a  small shift  in wavelength during derivatization,
in this case about 4 nm towards shorter wavelength.

                                                13          1
The investigation of the light oil fractions by   C NMR and  H NMR along
with gas chromatography alone gave considerable information.

The two  phenolic fractions, >20%,  were  well  characterized.   The CH2C12
fraction contained only the four most water soluble phenols, phenol it-
self, o-cresol,  m-cresol  and  p-cresol,  with a small amount  of an uni-
dentified  phenol.    The IR was  consistent  with  this  observation.   The
exact  amounts  present were determined  by  GC.    The  pentane soluble
phenolic fraction  also contained the same four phenols in large amounts
with other alkylphenols making up  the rest of the material (Figure 7a).
There were no guiacols  and no naphthols found.

                                                  13
The hydrocarbon fraction, 48%, is seen to exhibit   C resonance lines in
the aliphatic  region that coincide with the  shifts  seen  for the normal
paraffin  series  (Figure 6b).   These may be  long chain substituents on
benzene or may be paraffinic hydrocarbons.  Tetralin and decalin may be
present.   Methyl  benzenes  and methyl naphthalenes do not appear in this
fraction.

-------
The hydroaromatic  fraction,  22%,  Figure 6a«and Figure  7b,  shows  promi-
nent naphthalene lines in both the  H and   C  NMR spectra.   Phenanthrene
is present as  indicated  by the low-field 4, 5 protons (Figure 7b).   The
IR shows  the probable presence of ethers; however, the ethers are  not
aliphatic--i.e. , raethyoxy--in  nature.   The    C NMR does show  the  pos-
sible presence of  dibenzofuran in very small  amount.   Carbazoles  may be
present.

The base  fraction,  5%,  contained  phenolic contaminants since it was ex-
tracted first  from an aqueous layer.   It  did concentrate  most  of the
nitrogen  compounds  except  for  carbazoles.   The quinolines  and pyridines
present in this fraction appear to have mainly methyl substituents.   Two
types of N-H stretching are seen in the IB, 3400cm   and 3485cm  .  Pro-
tonated pyridine is seen at about  3400cm

                                HEAVY OIL

The product  oil (less light  oils) was characterized by gel permeation
HPLC to ascertain the degree of molecular size reduction as a measure of
liquefaction process  efficiency.   Figure 8 illustrates  the  approach to
constant  molecular  weight distribution as  the product  oil  is  recycled
with added  coal.   The use of the  ratio  of UV absorbance at 950 MW vs
250 MW as a  liquefaction product  parameter is illustrated in Figure 9,
and has been previously  discussed by Knudson  et  al.  (11).   Examination
of  the  character  and content  of the vacuum  bottoms from  coal  lique-
faction has been discussed by Farnum et al. (12,  13).

Proton NMR  in  d,.   pyridine has been used to  determine  the  aromatic (%
Har), benzylic  (/0  Ha)  and other aliphatic (% Ho) proton distribution in
the  undistilled product  slurry  according  to the  method  of Brown  and
Ladner  (14).   Table 5 lists  the  results along  with some computed  pa-
rameters,  f  (the fraction aromatic carbon), a (the degree  of substition
of aromatic  sites)  and Haru/Car  (the H/C  ratio  of  the hypothetical un-
substituted aromatic material).  The approach to constant composition is
apparent.   Mass  spectral  analysis  of  vacuum distillates from lignite
liquefaction have been discussed  by Schiller  (15).  Further  work is in
progress on characterization of light oils and vacuum distillate by LVMS
and GC-MS.

                              BIBLIOGRAPHY

 1.  Appell,  H.R.  and  I.  Wender,  "Hydrogenation of Coal with Carbon Mo-
     noxide,"   Div. of Fuel Chem. Preprints,  American Chemical Society,
     12 (3),  220-222 (1968).

 2.  Del  Bel,  E.,  S.  Friedman, P.M. Yavorsky and I. Wender,  "The  Lique-
     faction  of Lignite  by  the   CO-Steam Process,"  Presented  at  the
     A.I.Ch.E.  National Meeting, Houston, TX,  March 16-20,  1975.

 3.  Sondreal,  E.A.,  C.L.  Knudson,  J.E.  Schiller and T.H.  May,  "Devel-
     opment  of the  CO-Steam  Process  for  Liquefaction of  Lignite  and
     Western Subbituminous Coals," Proceedings of the 1977 Symposium on
     Technology and Use of Lignite,  GFERC/IC-77/1,   Grand  Forks  Energy
     Technology Center, Grand Forks, ND 58202,  May 1977.

                                    90

-------
 4.   Sondreal,   E.A.,  C.L.  Knudson,  R.S. Majkrzak,   and  G.G.  Baker,
     "Liquefaction  of  Lignite  by  the   CO-Steam  Process,"  A.I.Ch.E.
     National Meeting, Miami,  Florida,  November 12-16,  1978.

 5.   Willson, W. ,  G.G.  Baker,  C.L. Knudson,  T.C. Owens and  D.E.  Sever-
     son,  "Application  of  Liquefaction  Processes  to  Low-Rank  Coals,"
     Ind.  Eng.  Chem.  Prod.  Res. Dev.  18,  297-310 (1979).

 6.   Farnum,  B.W.,  C.L.  Knudson and D.A.  Koch,  "Products  of Liquefaction
     of Lignite with  Synthesis Gas by Product  Slurry  Recycle,"  Div.  of
     Fuel  Chemistry Preprints,  American Chemical Society,     24     (3),
     195-203  (1979).

 7.   American Public  Health Assn.   Standard Methods  for Examination  of
     Water and Wastewater,   14th Ed., APHA,  AWWA, and  WPCF,  Washington
     (1975).

 8.   Environmental Protection  Agency,  "EPA Methods for  Chemical Analysis
     of Water and  Wastes,"   EPA #5501-0067 (1976).

 9.   Fruchter,   J.S.,   V.C.  Laul,  M.R. Peterson,  P.W. Ryan,  and  M.E.
     Turner,  "High-Precision Trace Element and  Organic  Constituent Anal-
     ysis  of Oil Shale  and  Solvent-Refined Coal Materials,"  in Analyti-
     cal Chemistry of Liquid Fuel Sources, American  Chemical  Society,
     Washington, B.C.  (1978).

10.   Farnum,  S.A., E.S. Olson,  B.W.  Farnum, andW.G. Willson,  "Charac-
     terization of Light  Oils  from  Liquefaction of Lignite,"   Div.  of
     Fuel  Chem.  Preprints,  American Chemical Society,  Vol. 25  (1980)  In
     Press.

11.   Knudson, C.L., J.E.  Schiller and A.L. Ruud, "Temperature Effects  on
     Coal  Liquefaction;  Rates  of Depolymerization and Product Quality  as
     Determined by Gel  Permeation Chromatography,"  Div.  of Fuel Chem.
     Preprints,  American Chemical Society, 22  (6), 49-58  (1977).

12.   Schiller,  J.E.,  B.W.  Farnum and E.A. Sondreal, "Viscosity  of  Coal
     Liquids  -  The Effect  of  Character  and Content of  the  Non-distil-
     lable   Portion."    Div.  of Fuel Chem.  Preprints,  American Chemical
     Society, 22 (6)  33-48  (1977).

13.   Farnum,  B.W.  and C.L.  Knudson, "CO-Steam Process:   Functional Group
     Analysis of Non-distillables  in Lignite-derived  Liquids,"  Div.  of
     Fuel  Chem.  Preprints,  American Chemical Society,    23   (2)    67-71
     (1978).

14.   Brown,  J.K. and  W.R.  Ladner,  "A Study of  the Hydrogen Distribution
     in Coal-like  Materials by  High-resolution Nuclear Magnetic  Reso-
     nance Spectroscopy,"  Fuel (London),  39, 87-96 (1960).

15.   Schiller,  J.E.  "Composition of Coal Liquefaction  Products,"  Hydro-
     carbon Processing  56,  147-152  (1977).


                                    91

-------
                                  TABLE 1

                           Analyses of Coal for
                      Beulah 3,  North Dakota,  Lignite
Basis of reported analysis:
                                         Coal Analysis (GF-79-2147)
                                     As-received
               Moisture-free
                 Moisture-  and
                   ash-free
Proximate analysis, pet.:
  Moisture	
  Volatile matter	
  Fixed carbon	
  Ash	
    TOTAL	
Ultimate analysis, pet.:
  Hydrogen	
  Carbon	
  Nitrogen	
  Oxygen	
  Sulfur	
  Ash	
    TOTAL	
29.48
30.21
29.58
10.73
100.00
--
42.84
41.94
15.22
100.00
—
50.53
49.47
--
100.00
  6.20
 42.87
  0.48
 37.91
  1.81
 10.73
100.00
  4.15
 60.79
  0.68
 16.59
  2.57
 15.22
100.00
  4.89
 71.71
  0.80
 19.57
  3.03
100.00
  An "atypical" Beulah Lignite chosen specifically for its high ash content.
                                           92

-------
                                   TABLE 2
                                                   o
                         Effluent Water Composition
               (Averages of Recycle Passes 21, 24, 29 and 32)
                                          (Concentrations in ppm)
                 pH at analysis
                 Alkalinity as CaCO,
                 Ammonia
                 Total sulfur
                 Total carbon
                 Inorganic carbon
                 Organic carbon
                 Phenol
                 o-cresol
                 m,p-cresols
                                  8.6
                             80,400
                             27,300
                              2,380
                             31,200
                             12,400
                             18,800
                              6,430
                                579
                              1,640
  a.  Analysis carried out under contract by Stearns-Roger, Inc.
                                TABLE 3

             Capillary G.C. Analysis of LO 28-14 Fractions
   Fraction
                  // Components
Basic
Phenolic
Heteroaromatic
Hydrocarbon
12
26
21
26
0.5-1.0%

    7
   23
   18
   14
0.25-0.5%

   21
   10
   19
   20
Total

  48
  84
 124
 121
                                     93

-------
                                   TABLE 4

                Elemental Analysis Distribution of Light Oil
                 28-14 and its Fractions Based on LO as 100%

28-LO-14
Basic
Phenolic
Heteroaromatic
Hydrocarbon
% c
84.34
19.94
19.77
22.38
22.26
% H
10.69
2.64
2.49
2.21
3.35
% 0
0.37
0.09
0.25
0.035
0.001
X*
3.70
3.35
0.03
0.32
0.00
%Q
O
0.22
0.01
0.01
0.20
0.01
%H20
0.23
—
--
--
_ •.
                                  TABLE  5
                     NMR  Analysis  of Unfiltered Product'
Sample
27-Y1
27-B2B
27-B2D
28-2B
28-4B
28-6B
8B
10B
12B
13B
14B
15B
16B
Pass No.
1
3
5
18
20
22
24
26
28
29
30
31
32
%Har
33.4
35.8
38.0
42.6
43.4
42.7
43.1
43.5
43.0
43.9
43.1
42.5
44.1
%Hoc
28.6
28.5
28.6
28.7
29.3
28.8
28.3
27.1
28.6
27.7
28.3
28.0
28.2
%Ho
38.0
35.7
33.5
28.7
27.3
28.4
28.6
29.5
28.4
28.5
28.6
29.5
27.7
fa
0.65
0.67
0.69
0.73
0.74
0.73
0.74
0.74
0.74
0.75
0.75
0.74
0.75
a
0.30
0.29
0.30
0.30
0.28
0.29
0.29
0.27
0.28
0.28
0.29
0.29
0.29
Haru/Car
0.76
0.77
0.76
0.75
0.74
0.73
0.74
0.72
0.72
0.71
0.70
0.71
0.70
a.  Analyses performed on a 90 MHz Spectrometer located at the University of
    North Dakota Department of Chemistry.

                                     94

-------
    100



    90



    80
  o
  8  TO
    60
 O
 QL
    50
    40
 O  30
    20



    10



    0
	  WATERJ_H2S1_NH3,_C02J_AND_OTHERS (BY DIFFERENCE!
               C,-C4  HYDROCARBON GASES
                        HIGH PRESSURE
                          LIGHT OILS
                                                     (IBP - 300° C AT 4000  PSIG)
                        VACUUM DISTILLATE  (IBP-250° C AT I TORR)
       THF  SOLUBLE VACUUM BOTTOMS 0
                                          ORGANIC THF  INSOLUBLES
                       8   10   12   14   16   18   20  22   24   26  28   30   32

                                   RECYCLE PASS

                    RUN  27 	4-	 RUN  28 	H
FIGURE 1    Runs 27 &  28:   Distribution  of MAF coal conversion products

             versus recycle pass.
                                     95

-------
         O
                                                                -0.12  9
                                                                     ro
                                             m
                                             oo
                                             CD
                                             CM
10       15      20

 RECYCLE  PASS NUMBER
25
30
                                                                 010   CE

                                                                      cc.
                                                                 0.08
                                                                 0.06
                                                                 0.04
FIGURE  2    Aromatic to aliphatic  proton ratio  in light oils by NMR and IR

            versus recycle pass  number.
                                     96

-------
                            LIGHT OIL 28-14 (IN PENTANE)
                                          EXTRACT WITH
                                          IN HCI
                    PENTANE
                        I IN NoOH
              PENTANE
                 lOMSO
                          DMSO
       PENTANE
          JEVAP.
PENTANE EXTR.
EVAP.
    "HYDROCARBON"   "HETEROAROMATICS"
       (HC) 48%          (HA) 22%
                            AQUEOUS
                                 NEUTRALIZE w/NoOH
                                 PENTANE EXTR., CH,CI2
                                 EVAP.


                           "BASES"  (5%)
    AQUEOUS
         NEUTRALIZE w/HCI
         PENTANE EXTR.,
         EVAP.


"PHENOLICS" (25%)
FIGURE 3    Separation and recovery of  the light oil fraction by extraction.
                                          97

-------
UJ
o
                                    CPU Light Oil UV Spectra
       240  250 260 270  280 290 300  310  320  330 340  350

                       WAVELENGTH (nm)
         FIGURE h    UV spectra of LO 28-lA and fractions.
                                98

-------
                           4th Derivative UV CPU
                           Light Oil 28-14
                                               phenolic fraction
                                                   LO 28-14
    240 250 260  270  280  290  30O  310  320 330 34O 350
                     WAVELENGTH (nm)

FIGURE  5    Fourth derivative UV of light  oil 28-lU and  fractions.
                            99

-------
    160
120
   80
Shift (ppm)
40
                                                     20
                                          10
                                  Shift (ppm)
FIGURE 6    50 Mz ^3c NMR,  a)  of the hydroaromatic fraction, completely
            proton decoupled  in CDC13,   b) aliphatic and aromatic regions
            of the hydrocarbon fraction (without NOE, Cr(AcAc)3, 00013),
            of light oil  28-1 IK
                                      100

-------
                             5       4
                              Shift (ppm)
FIGURE 7
                              5       4
                               Shift (ppm)
200 MHz 1H EMR  of light oil 28-lU,  a) phenolic  fraction,
b) hydroaromatic  fraction.
                                   101

-------
FIGURE 8
              ELUTION  VOLUME, ml

Changes in the molecular weight distribution with recycle
(Run 27).   Detection was at 365 nm.
                      102

-------
       3.5
        3.0
     I"
     O
     oo
     (VI
     \
     1

     8  2.0
     u
     o
     <
     m
     CO
     00
     LL
     O
        1.0
        0.5
    \
      \

D 0 %   \
 TETRALIN \
                                       RUNS PLOTTED
                     \
BATCH

 23
 26
 27
 28
 30
 39
 51
 53
 CPU

10-2
11-1,2,3
12-3,4,5
13-1,2,3
14-1,2,3
15-1,2,3
16-1
26-1 to 12
 D 4.7%
                                  — CSTR CONTINUOUS UNIT
         400       420      440      460       480
                       REACTOR TEMPERATURE, °C
                                              500
FIGURE  9     Ratio of U.V.  absorbance  at 25\ nm of  950 MW to 250 MW
             material versus reactor temperature for  reaction times
             under 30 minutes.
                                  103

-------
               FOUR-HOUR ALGAL BIOASSAYS FOR ASSESSING THE
                   TOXICITY OF COAL-DERIVED MATERIALS

                             J. M. Giddings
                     Environmental Sciences Division
                      Oak Ridge National Laboratory
                       Oak Ridge, Tennessee 37830
                                ABSTRACT

As part of an overall program to assess the potential environmental
hazards of coal conversion technologies, researchers at Oak Ridge
National Laboratory have developed a rapid bioassay to measure the
effects of coal-derived materials on algal photosynthesis.  Algal cul-
tures or natural algal communities are exposed to the test materials for
four hours.  Photosynthesis is determined by the l "*C-bicarbonate method
during the final two hours of exposure and compared with controls for a
measure of toxicity.  In bioassays with individual aromatic compounds,
quinones and aromatic amines were found to be particularly toxic to
algae; azaarenes and thiophenes were the least toxic classes tested.
Experiments with the water soluble fractions (WSFs) of more than twenty
natural and synthetic oils showed that coal liquefaction products are
considerably more toxic than petroleum products; shale oils are inter-
mediate in toxicity.  Further studies with particular subfractions of
several WSFs have identified ether-soluble bases as the major contributors
to the toxicity of coal-derived oils.  The four-hour photosynthesis test
has several advantages over the Algal Assay Bottle Test for measuring the
toxicity of complex materials.

                              INTRODUCTION

Environmental scientists at Oak Ridge National Laboratory are engaged in
a broad program of research concerning the environmental fate and effects
of products and wastes from coal liquefaction and gasification processes.
The algal bioassays described in this report are included in the environ-
mental effects research program because algae are the dominant primary
producers in most lentic (non-flowing) aquatic ecosystems.  Since algae
are at the base of most lentic food webs, changes in the productivity or
structure of the algal community could have important consequences for
the rest of the ecosystem.

                        ALGAL PHOTOSYNTHESIS TEST

The algal bioassay was developed to meet two principal objectives:

     1.  Rapid screening of coal-derived materials (liquid products
         and fractions thereof, solid waste leachates, and waste-
         waters) for acute toxicity to freshwater algae.

     2.  Determination of suitable concentrations for further
         effects testing.


                                    104

-------
The primary function of the bioassay is to compare the toxicities of
different materials.  Comparisons are based on a simple toxicological
response:  short-term inhibition of algal photosynthesis.  The bioassay
can be used with any algal culture or natural algal community.  In our
experiments, the test organisms are Selenastrum capricornutum, a green
alga, and Microcystis aeruginosa, a non-nitrogen-fixing blue-green alga.
These are two of the species recommended for the Environmental Protection
Agency's Algal Assay Bottle Test (AA:BT) (9).  The medium and growth
conditions for these cultures are specified by Miller et al.  (9).

Aliquots (18 or 20 ml) of the cultures are distributed to 20 x 120 mm
screw-capped culture tubes.  The test material is then added in aqueous
(2 ml) or acetone (0.1 ml) solution, and the cultures are placed under
a bank of fluorescent and incandescent lights (1700 yWatts cm~2 irradia-
tion between 400 and 700 nm) in an environmental chamber at 24°C.  The
cultures are pre-incubated for 2 hr before photosynthesis measurement is
begun.  Pre-incubation is included in the procedure because some materials,
such as naphthalene (12), do not inhibit algal photosynthesis immediately
but require 60-90 min to produce a full effect.  After pre-incubation,
cultures are spiked with 0.01 ml (0.1 yCi) of a 14C-labelled sodium bi-
carbonate solution and incubated for another 2 hr.  Photosynthesis is
stopped by the addition of 0.05 ml formalin at the end of the incubation
period.  The amount of 1I+C that has been converted to organic form by
photosynthesis is determined by acidifying 5-ml samples with 0.1 ml of
0.1 N^ HC1, bubbling with air for 20 min to remove inorganic 1'*C, and
measuring the remaining activity by liquid scintillation spectrometry
(6).  A set of controls is included in each experiment, and the response
of the algae to the test material is assessed by comparing treated cul-
tures with controls.  When test materials are added in acetone solution,
a set of acetone controls is also run; acetone has never had a detectable
effect on photosynthesis in our experiments.

Some of the results presented below are expressed as ECzo values.  The
ECao is the concentration of test material causing 20% inhibition of
photosynthesis (i.e., photosynthesis is 80% of controls), as determined
by interpolation on a plot of photosynthesis vs. log concentration.

                      STUDIES WITH SINGLE COMPOUNDS

In the course of developing the photosynthesis test, we compared the
toxicity to S^. capricornutum and M. aeruginosa of more than 30 aromatic
compounds, including unsubstituted and methylated hydrocarbons, phenols,
quinones, amines, azaarenes, and thiophenes (6).  Some of the results for
M. aeruginosa are shown in Figure 1.  For each class of compounds except
quinones, toxicity increased with increasing number of aromatic rings.
Quinones were by far the most toxic class of compounds tested; amines
were the next most toxic.  Thiophenes and azaarenes were the least toxic.
These generalizations also applied to ^. capricornutum (6).  Information
on the relative toxicities of these common aromatic compounds has been
useful in interpreting results of bioassays with complex coal-derived
mixtures.
                                     105

-------
The toxicity of most of the single compounds and complex mixtures we
have tested was approximately the same (within a factor of 5) for E^.
capricornutum and M. aeruginosa.  The responses of both species were
consistent with those of natural algal communities collected from a
shallow pond (Table 1).  Eight aromatic compounds and three oil extracts
were ranked in essentially the same order by natural communities, S^
capricornutum, and M. aeruginosa.

The 4-hr algal bioassay also ranked 30 aromatic compounds in approxi-
mately the same order of toxicity as 48-hr Daphnia magna LCso tests (11).
The interpolated ECao values for most of the aromatics tested were within
one order of magnitude of the I), magna 48-hr LCso's (Figure 2).  The fact
that concentrations causing slight photosynthetic inhibition in algae
generally caused extensive mortality in I), magna could be considered
evidence that algae are less sensitive to toxicants than zooplankton
(8).  However, it is difficult to compare the true ecological signifi-
cance of the responses of the two taxa, particularly since the bioassays
were of different lengths.

                      STUDIES WITH COMPLEX MIXTURES

The applicability of the 4-hr algal bioassay to testing complex coal-
derived materials is illustrated by experiments with the water soluble
fractions (WSFs) of coal liquefaction products.  We chose to test WSFs
rather than whole oils because the acute toxicity of oils to aquatic
organisms is generally caused by the oils' water soluble components
(3,10).  By testing WSFs we also avoid the dosage problems caused by the
immiscibility of oils in water.  We prepare WSFs by floating oil on dis-
tilled water in a 1:8 (oil:water) ratio and stirring very gently for 16
hr in the dark.  The WSFs are then separated from the oils and filtered
(Whatman No. 41) before testing.  Dilutions are made in fresh algal
growth medium (9).  Test solution concentrations are expressed as
percentages of full-strength WSF.

The initial objective of this work was to compare the toxicities of WSFs
prepared from coal liquefaction products, shale oil products, and
petroleum products.  Selected dose-response curves for one oil of each
type are presented in Figure 3.  The petroleum-derived marine diesel fuel
(DFM) was the most toxic of the petroleum products we tested, including
No. 2 diesel fuel, a jet fuel, and three residual fuel oils.  Crude shale
oil was more toxic than petroleum DFM  (Figure 3); refined shale oil
products, however, were similar in toxicity to petroleum products.  Coal
liquefaction products were consistently more toxic to algae than shale
oil or petroleum (Figure 3).  The 11 unrefined coal-derived oils we
tested were all toxic to S_. capricornutum at WSF concentrations of 1% or
less, whereas petroleum WSFs were generally nontoxic even at 10% (7).

We suspected that the difference in toxicities of coal-derived and
petroleum WSFs was due to the abundance of phenols and aromatic amines
in coal liquefaction products.  To investigate this possibility, we
fractionated the WSFs of several liquefaction products into ether-soluble
                                      106

-------
acid, ether-soluble base, and neutral components, and tested each sub-
fraction at one-tenth of its concentration in the original, full-strength
WSF.  A consistent pattern emerged (Figure 4).  The ether-soluble bases
were the most toxic subfraction in nearly every case.  The neutral sub-
fractions, with one exception, were the least toxic to both species.  The
effects of ether-soluble acids varied considerably among oils and between
species.  Our conclusion, based on these results, was that the bases
(mostly aromatic amines) and, secondarily, the acids (mostly phenols),
were the major toxic constituents of WSFs of coal liquefaction products.

                COMPARISONS WITH ALGAL ASSAY BOTTLE TEST

The Algal Assay Bottle Test (AA:BT) (9), originally developed during the
late 1960's to study the causes of eutrophication in lakes, has recently
been applied to the testing of toxic materials.  The Environmental
Protection Agency is considering the AA:BT as a standard method for
testing chemicals regulated under the Toxic Substances Control Act (2),
and a modified version of the AA:BT appears in the latest edition of
Standard Methods for the Examination of Water and Waste Water (1).  Since
widespread adoption of the AA:BT appears imminent, we have begun testing
selected materials by this method for comparison with the 4-hr photo-
synthesis test.

Briefly, the AA:BT involves inoculating test solutions with a small
number of algal cells, incubating the cultures for 14 days, and measuring
the final yield of algal biomass.  Our AA:BT experiments were conducted
as prescribed by Miller et al. (9) except that the culture flasks were
capped with foil to reduce the loss of test compounds by volatilization.
Algal biomass was determined after 14 days of growth by vacuum-filtering
3 20-ml aliquots from each culture onto pre-rinsed, dried, tared, 0.4-y
Nucleopore filters; drying the filters at room temperature over anhydrous
calcium sulfate; and weighing each filter on a Cahn microbalance.  Net
dry weights of the treated cultures were compared with dry weights of
controls to determine the effect of each treatment.

The results of the first AA:BT experiments are presented in Table 2, with
results of photosynthesis tests on the same compounds included for com-
parison.  The two methods agreed closely in some cases and diverged widely
in others.  Neither method was consistently more sensitive than the other.
The discrepancies may reflect different measured responses, different
exposure times, or experimental artifacts.  We are now running additional
AA:BT experiments to permit more conclusive comparisons.

Because of its speed, small scale, and technical simplicity, the 4-hr
photosynthesis test is more efficient than the AA:BT for testing large
numbers of materials.  The photosynthesis bioassay is, obviously, more
rapid than the AA:BT; the entire procedure for the photosynthesis test,
including preparation, exposure, and analysis, can be completed in less
than 2 days.  The 4-hr test takes less space in an environmental chamber
than the AA:BT, so more treatments can be included in a single experiment
and thus can be compared with a single set of controls.  Measurement of
photosynthesis by the 14C method is no more technically demanding than

                                     107

-------
measurement of biovolume with an electronic particle counter, and much
less tedious than direct determination of dry weight.

The AA:BT is not easily applied to testing volatile or unstable materials,
including most of the single compounds and complex mixtures of concern to
coal conversion.  This is because the concentration and chemical composi-
tion of such materials may change drastically during a 14-day experiment.
Photolysis, hydrolysis, and biodegradation of the test material are
unavoidable.  Volatilization can be reduced by sealing the flasks, as
we did in our AA:BT experiments, but this also shuts off the C02 supply
to the algae and can prevent the cultures from reaching their normal
maximum biomass (4).

Not only the test substance, but also the exposure conditions, may change
during an AA:BT experiment.  "Bottle effects", that is, effects of
enclosure on an algal population or community, become more pronounced
with increasing incubation time.  As an algal culture ages, changes occur
in pH and other chemical variables, nutrient levels decline, algal
metabolites accumulate in the medium, bacteria develop on the walls
of the vessel, and the algae undergo a sequence of profound physiological
changes (5).  Most of these factors can affect the response of a test
organism to a chemical.

An objective of most standardized bioassay procedures is to control the
test substance, the exposure conditions, and the physiological state of
the test organism as closely as possible.  Since all of these variables
change continuously in an AA:BT experiment, unequivocal interpretation of
the results is practically impossible.  For example, the AA:BT cannot be
expected to produce a valid dose-response relationship if the dose cannot
be unambiguously defined.  There are two ways to minimize or avoid these
difficulties.  One solution is to adopt continuous culture techniques (5)
if long-term exposures are desired.  The other strategy is to minimize
the enclosure time, as we have done with the photosynthesis test.

One of the  greatest challenges of environmental toxicology is relating
laboratory  test results  to a natural ecological context.  This is
especially  difficult with the AA:BT; the maximum yield of algal biomass
is relevant to the problems of eutrophication but not to the effects of
toxic substances.  Algae in nature rarely experience a stationary phase
comparable  to the measured endpoint of the AA:BT.  Even a high standing
crop of algae can sustain animal populations only if the algae that are
consumed are continually replaced by new production.  The rate of primary
production  sets an upper limit to the rate of secondary production in most
lentic ecosystems.  Since the 4-hr photosynthesis test detects effects on
algal productivity directly, if only for a short period of time,  it is
more ecologically meaningful than the AA:BT.

                               CONCLUSIONS

Because it  is simple,  rapid, and inexpensive, the 4-hr algal photosynthe-
sis test has proven useful in screening large numbers of coal-related
substances.  The responses of S_. capricornutum and M. aeruginosa were
similar to  those of natural algal communities in our experiments, and
                                     108

-------
also correlated well with ID. magna acute  toxicity tests.  The 4-hr  algal
photosynthesis test is a convenient, ecologically relevant alternative  to
the Algal Assay Bottle Test for  toxicity  screening of simple and  complex
materials.

                            ACKNOWLEDGEMENTS

Research sponsored by the Office of Health and Environmental Research,
U.S. Department of Energy, under contract W-7405-eng-26 with Union  Carbide
Corporation.  Publication No. 1515, Environmental Sciences Division, Oak
Ridge National Laboratory.  Technical assistance was provided by  J. N.
Washington.  B. R. Parkhurst, R. M. Cushman, C. W. Gehrs, and W.  Van Winkle
made helpful suggestions on the manuscript.

                               REFERENCES

 1.  American Public Health Association,  Standard Methods for the
     Examination of Water and Waste Water, 14th ed., APHA, Washington
     (1976).

 2.  Environmental Protection Agency, "Toxic Substances Control," Federal
     Register, ^4_, 16240-16292 (1979).

 3.  Evans, D. R. and S. D. Rice, "Effects of Oil on Marine Ecosystems:
     A Review for Administrators and Policy Makers,"  Fish. Bull.,  72,
     625-638 (1974).

 4.  Fitzgerald, G. P., "Factors Affecting the Algal Assay Procedure,"
     Report for the Office of Research and Monitoring, U.S. Environmental
     Protection Agency, Project No. P5J11912-J (1975).

 5.  Fogg, G. E., Algal Cultures and Phytoplankton Ecology, University  of
     Wisconsin Press, Madison (1965).

 6.  Giddings, J. M., "Acute Toxicity to  Selenastrum capricornutum  of
     Aromatic Compounds from Coal Conversion," Bull. Environ. Contain.
     Toxicol., 23_, 360-364 (1979).

 7.  Giddings, J. M. and J. N. Washington, "Coal Liquefaction Products,
     Shale Oil, and Petroleum:  Acute Toxicity to Freshwater Algae,"
     submitted to Science (1980).

 8.  Kenaga, E. E. and R. J. Moolenaar, "Fish and Daphnia Toxicity  as
     Surrogates for Aquatic Vascular Plants and Algae," Environ.  Sci.
     Techno1., 13, 1479-1480 (1979).

 9.  Miller, W. E., J. C. Greene and T. Shiroyama, "The Selenastrum
     capricornutum Printz Algal Assay Bottle Test," EPA-600/9-78-018
     (1978).

10.  Moore, S. F. and R.  L. Dwyer, "Effects of Oil on Marine Organisms:
     A Critical Assessment of Published Data," Water Res. , 8^, 819-827
     (1974).
                                     109

-------
11.   Parkhurst,  B.  R.,  Environmental Sciences Division,  Oak Ridge
     National Laboratory,  Oak Ridge,  TN,  personal communications,
     October 1,  1979.

12.   Soto,  C., J.  A.  Hellebust and T.  C.  Hutchinson,  "Effects of
     Naphthalene and Aqueous Crude Oil Extracts on the Green Flagellate
     Chlamydomonas angulosa.  II.   Photosynthesis and the Uptake and
     Release of Naphthalene," Can. J.  Bot.,  53, 118-126 (1975).
                                    110

-------
                                TABLE 1

 Acute Toxicity of Aromatic Compounds (ECzo in ppm) and Water Soluble
 Fractions (EC2o in %) to Algal Cultures and Natural Algal Communities
                           Selenastvwn    Miarocystis   Natural Algal
      Aromatics           capricornutum   aeruginosa      Community
Phenanthrene
1-Naphthylamine
2-Naphthol
Naphthalene
Acridine
Phenol
Quinoline
Benzene
0.7
1.7
4
5.7
1.3
174
25
144
0.15
0.22
0.75
0.85
3.6
54
117
50
0.25
0.23
1.0
1.4
3.3
102
110
250
Water Soluble Fractions
Petroleum DFM                  13                             4.5
Coal-Derived Fuel Oil           0.4            0.5            0.2
Coal-Derived Distillate         0.4            0.9            0.4
                                    111

-------
                      TABLE 2

   Toxicity of Aromatic Compounds to Two Algae:
 Comparison of Algal Photosynthesis Bioassay (APB)
      and the Algal Assay Bottle Test (AA:BT)
   Species
  Compound
-EC20 (ppm)—
APB     AA'.BT
   capricornutum    acridine          1       0.5

        "           naphthalene       6      30
M. aeruginosa
phenanthrene      0.2     3

2-naphthol        0.8    20

quinoline       100       7

acridine          4      20
                        112

-------
                                        ORNL-DWG 80-8873
             ACUTE TOXICITY OF AROMATIC COMPOUNDS  TO
                          Microcystis aeruginosa
           io4
                            NUMBER OF RINGS

FIGURE 1   Acute  toxicity of aromatic compounds to Microcystis  aeruginosa.
           ac  = acridine; an = aniline; ap = 9-aminophenanthrene; b =
           benzene; bq = benzoquinone; db = dibenzothiophene; ran = 1-
           methylnaphthalene; n = naphthalene; na = 1-naphthylamine; nl =
           2-naphthol; nq = naphthoquinone; p = phenanthrene; ph = phenol;
           pq  = phenanthraquinone; py = pyridine; qu = quinoline; th =
           thiophene; tn = thianaphthene; to = toluene.   The line connecting
           pyridine and quinoline is estimated; pyridine was not toxic at
           1000 ppm,  the highest concentration tested.
                                  113

-------
o.
o.
o»
o
 o
 CM
O
UJ
   -1
   -2
                                               ORNL-DWG 80-7786
                                                      	Si
                                      •  SELENASTRUM
                                      O  MICROCYSTIS
                                                     I
     -2
-1012
DAPHNIA MAGNA 48-hr LC50 (log ppm)
   FIGURE 2   Comparison of acute  toxicities of aromatic compounds to
              algae and to I),  magna.  Each point represents one com-
              pound.  Broken diagonals enclose compounds for which
              algal ECat and I),  magna LCso differ by  less than one
              order of magnitude.  D. magna data from Parkhurst (11).
                                  114

-------
                                     ORNL-DWG  80-7783
  100
   80
c
o
o
CO

-------
ro
sj-
sT
00
 I
O
00
O

 I
  O
                                                                   116

-------
       DIRECT DETERMINATION OF POLYNUCLEAR AROMATIC HYDROCARBONS
                     IN COAL LIQUIDS AND SHALE OIL

                A. P. D'Silva, Y. Yang and V. A. Fassel
          Ames Laboratory - USDOE and Department of Chemistry
                         Iowa State University
Elaborate analytical methodologies are usually utilized to determine
the carcinogenic benz[a]pyrene and other potentially carcinogenic poly-
nuclear aromatic hydrocarbons (PAHs) present in coal liquifaction products
and shale oil, the liquid fuels of the future.  Existing methodologies
require that individual compounds or a small group of compounds be iso-
lated by utilizing high performance liquid chromatography (1), capillary
column gas chromatography (2, 3) or thin layer chromatography (4).  Detec-
tion and quantitation of the isolated PAH compounds is usually performed
by mass spectrometry or luminescence spectrometry.

Coal liquids and shale oil are very complex mixtures of organic com-
pounds and their analysis by conventional techniques is relatively diffi-
cult.  In contrast, laser excited Shpol'skii effect spectroscopy  (LESS)
is a simple and less elaborate approach that has been utilized for the
direct determination of benz[a]pyrene and other PAHs in SRC II, a coal
liquid and a sample of shale oil.  The determinations are accomplished
without prior isolation of the compounds.

For the analysis of complex mixtures of PAHs, the spectral resolution
in the fluorescence spectra can be enhanced by adopting low temperature
luminescence techniques.  Currently, the most promising techniques are
Shpol'skii effect spectroscopy in n-alkane hosts (5, 6), fluorescence
line narrowing spectroscopy (FLNS) in organic glasses (7,  8) and matrix
isolation (MI) spectroscopy in a frozen nitrogen matrix (9,  10).  Typical
bandwidths (full width at half maximum, FWHM) in the electronic spectra
of PAHs in organic glasses,  frozen N, and n-alkane matrices are ^200 cm"-'-,
'vlOO cm~l, and vLO cm~l, respectively.  In FLNS and MI, the resolution
in the luminescence spectra is achieved by the application of narrow
bandwidth tunable dye lasers.  The inherent, relatively narrow bandwidths
in the electronic spectra of PAHs in n-alkane solvents is  a significant
advantage in achieving high resolution luminescence spectra.  The resolu-
tion can be further enhanced in Shpol'skii effect spectroscopy,  as tunable
narrow bandwidth (^0.02 nm at 400 nm) dye lasers can be utilized to
selectively excite any given compound present in a highly  complex mixture
of compounds.  Thus,  laser excited Shpol'skii effect spectroscopy (LESS)
of PAHs in n-octane,  frozen to a solid at 15K, has been utilized to
achieve the following:  (1)  site selected luminescence of  a PAH compound,
(2) selectively excited luminescence of any individual alkylated benz[a]-
anthracene present in a mixture of multialkylated benz[a]anthracenes and
(3) quantitate selected PAHs including benz[ajpyrene in diluted samples
of SRC II and shale oil.

The apparatus for experimental procedures are described in a communication
submitted for publication in Analytical Chemistry (11). A comparison
of the data obtained  by LESS with that obtained by the NBS and by other
techniques at the Ames Laboratory is also presented in the communication.
                                     117

-------
                               REFERENCES

 1.   Wise,  S.  A.,  S.  N.  Chesler,  H.  S. Hertz, L. R. Hilpert, W. E. May,
     Anal.  Chem.,  49,  2406 (1977).

 2.   Guerin,  M.  R.,  J. L.  Epler,  W.  H. Griest, B. R. Clark, T. K. Rao,
     in Carcinogenesis,  Vol.  3, P.  W. Jones, R. I. Freudenthal, Eds.,
     Raven Press,  NY pp. 21-33 (1978).

 3.   White, C. M.,  A.  G. Sharkey, Jr., M. L. Lee, D. L. Vassilaros, in
     Polynuclear Aromatic Hydrocarbons, P. W. Jones, P. Leber, Eds.,
     Ann Arbor Science,  Inc., Ann Arbor, MI, pp. 261-275 (1979).

 4.   Hurtubise,  R.  J., J.  D.  Phillip, Anal. Chim. Acta, 110, 245  (1979).

 5.   Kirkbright,  G.  F.,  C. G. DeLima, Analyst, 99^ 338 (1974).

 6.   Khesina,  A.  Ya. ,  G. A. Smirnov, E. A. Ermakov, A. A. Knyazwa,
     Ya. I. Yashin,  Zh.  Anal. Khim., J53, 2032  (1978).

 7.   Brown, J. C.,  M.  C. Edelson, G. J. Small, Anal. Chem., 50, 1394  (1978),

 8.   Brown, J. C.,  J.  M. Hayes and G. J. Small, "New Laser  Based Methodo-
     logies for the Determination of Organic Pollutants via Fluorescence"
     in Lasers and Chemical Analysis, G. M. Hieftje, F. E.  Lytle and
     J. C.  Travis,  Eds., The Humana Press, Inc., Clifton, NJ  (1980).

 9.   Wehry, E. L.,  G.  Mamantov, Anal. Chem., 51, 643A  (1979).

10.   Dickinson,  Jr., R.  B., E. L. Wehry, Anal. Chem.,  51, 778  (1979).

11.   Yang.  Y., A.  P. D'Silva and V.  A. Fassel, Anal. Chem., submitted
     for publication.

                             ACKNOWLEDGEMENT

This research was supported by the U.S. Department of  Energy,  Contract
No.  W-74-5-Eng-82,  Office of Health and Environmental  Research, Physical
and Technological Studies, Budget Code GK-01-02-04-3.
                                   118

-------
           SYNCHRONOUS FLUORESCENCE AND  PHOSPHORESCENCE AT ROOM
             TEMPERATURE FOR LEVELS 1 AND  2  ORGANIC ANALYSIS

               R. B. Gammage, T. Vo-Dinh and P. R. Martinez
                    Health and  Safety Research Division
                       Oak Ridge National  Laboratory
                        Oak Ridge, Tennessee 37830
                                 ABSTRACT

By simple and rapid synchronous luminescence  techniques,  analysis  was
carried out on thirteen polynuclear components  in a  liquid  chromatography
fractionated XAD-2 extract from Source Assessment Sampling  System  runs.
This same sample was used in an interlaboratory evaluation  of  current
Environmental Protection Agency Level 1 organic analysis  procedures.   Com-
parative results show that synchronous fluorescence  and room temperature
phosphorescence provide data that are more  than adequate  for Level 1.
These techniques should also be seriously considered in the development  of
Level 2 organic analysis procedures.  An example is  given of the use of
luminescence for rapid screening of an unfractionated coal  liquefaction
product for a major PNA component.
                                INTRODUCTION

The interlaboratory examination of combined XAD-2 extracts  collected  by
Source Assessment Sampling Systems (SASS)  is part of a  study  (1)  to verify
the efficacy of the analytical  scheme outlined  in the InduA&iiat  EnvJJion-
me.ntal Re.4e.anck LaboiatofLy - R&Aeasich Tsuangle.  Path. PA.oce.duAe..6  Manual.:
Level 1 EnvJJiOWWLYVtal A&AZAAm&nt  (2).  The study is coordinated by the
Industrial Environmental Research Laboratory -  Research Triangle  Park
(IERL-RTP).  Our contribution to this study is  to gauge how well  certain
rapid, easy to use, and cost-effective luminescence techniques  can be
employed for the analysis of polynuclear aromatic (PNA)  compounds.

By demonstrating success we hope to "popularize" the luminescence tech-
niques so they will be accepted into the tiered or phased approach to
analytical investigation (Levels 1 and 2)  that  is advocated by  the EPA  (3).

Two luminescence techniques are considered; synchronous luminescence  (SL)
spectroscopy of solutions at room temperature (4) and room  temperature
phosphorimetry (RTF) (5).

                                METHODOLOGY

Synchronous Fluorescence (SF) and Synchronous Phosphorescence  (SP)

In conventional luminescence spectroscopy, either the excitation  (A   )
or the emission wavelength (A   ) remains fixed.  With synchronous lumi-
nescence, the luminescence signal is recorded while A    and A   are scanned
simultaneously at a fixed wavelength interval,  AA, between  the  two.

                                   119

-------
The principal advantage of this technique is improved selectivity.  This
results from a narrowing of a targeted spectral band, or bands, and a
reduction or elimination of spectral emissions from other compounds that
would otherwise cause interference effects.  The theory and selection of
appropriate values of AA have been dealt with previously (4).  The con-
ventional and synchronous fluorescence spectra of a synthetic mixture of
five PNA compounds (4) are shown in Figure 1 for the purpose of familiar-
izing the reader with the degree of spectral simplification that is possible.

The concept of synchronous excitation is applicable to phosphorimetry (5),
as well as fluorimetry (4).  In the case of SF the Stokes'  shift determines
the optimum value of AX which is often set at 3 nm (4).  For SP it is the
singlet-triplet energy difference that determines the optimum value of AA
(6); for most PNA compounds AA is the range of 100 to 300 nm.

A strong point of the synchronous technique is the simplicity of instru-
mentation.  The spectrometer employed for synchronous measurements is often
the same instrument used for measurement of conventional fixed-excitation
luminescence.  Several instruments with interlocking capability of the two
monochromators, as a standard feature, are commercially available.  Even
with an already existing spectrometer that does not have the interlocking
feature, a simple and inexpensive switch can be added for synchronous
scanning.

In some cases, further enhancement in selectivity is achieved by deriva-
tization of the luminescence signal output (7).  This method, referred to
as derivative luminescence spectroscopy, consists of measuring the first,
second, or higher derivative of a spectrum with respect to wavelength.  It
is used mainly to enhance minor spectral features superimposed on a more
intense background.  In the second-derivative (d2) mode, where one measures
the curvature of a peak, broad bands are suppressed while sharp bands are
intensified.  An illustration of this effect, for the RTP spectra of fluorene
(7), is illustrated in Figure 2.

The application of synchronous luminescence to the analysis of complex
organic mixtures is not without its limitations and potential pitfalls.
The user needs to be aware of these.  The range of accessible compounds is
limited.  Some PNA compounds fluoresce or phosphoresce too weakly to make
their measurement practicable.  Alternately they absorb or luminesce at
wavelengths not easily accessible.

Sample dilution is often necessary to avoid anomalies caused by inner filter
effects, energy transfer and/or quenching.  One needs to work within a range
where the response varies linearly with concentration.  The dilution neces-
sary to achieve this, however, can reduce the concentration of trace, or
weakly luminescing constituents below their limits of detection.  Discus-
sions of these problems, as they effect analyses by synchronous luminescence,
have appeared recently (8,9).
                                   120

-------
Room Temperature Phosphorimetry (RTF)

The relatively new technique of RTF is based on the detection of phospho-
rescence from organic compounds adsorbed on solid substrates at room temper-
ature.  This type of phosphorimetry has been reviewed (10).

Normally the solid substrate is filter paper.  The sample solution is
spotted on the filter paper and after drying, phosphorescence is measured
by a spectrofluorimeter equipped with a rotating phosphoroscope.  The
simplicity of the procedure is a distinct advantage.  Expanded experimental
details can be found in reference 11.

To improve the sensitivity and lower the detection limit to the nanogram
or subnanogram range, one resorts to the use of heavy-atom perturbers.  For
PNA compounds, a variety of heavy atom salts, such as cesium iodide, sodium
bromide, lead acetate and thallium acetate, are very efficient in enhancing
the phosphorescence quantum yields of most PNA compounds in a selective
manner (12).  The filter paper can be pretreated with the perturber or the
heavy atom salt can be added after spotting the paper with the PNA solution.

The fact that these operations are rapidly and easily carried out, makes
the technique of RTF attractive for the screening of large numbers of PNA
bearing samples.  The ease with which automated sampling and analysis (11)
could be carried out is also attractive.  All of these features are attrac-
tive in the context of Level 1 analysis.

                                  SAMPLES

As part of the interlaboratory study to evaluate Level 1 organic analysis
(1), we were provided with two samples.  The first sample was an artificial
mixture of 16 compounds in methylene chloride, sample number 5 in Table 15
of reference 1.  The real-life sample was pooled XAD-2 resin extract from
three SASS train runs.  The LC fraction in methylene chloride was designated
I-TX-3 (1).  The composition of this sample was determined by the collabo-
rators of the interlaboratory study, including ourselves.

A third sample to be mentioned briefly is a coal liquid product from the
Synthoil coal desulfurization process.  This material was dissolved in
methylene chloride and examined by RTF without any fractionation.

                          RESULTS AND DISCUSSION

Artificial Mixture

This sample was used to check the efficacy of the d2 - SF technique.  The
identity and quantity of the 16 aliphatic and aromatic compounds that were
present are listed in Table 1.  By SF, phenol and four of the PNA compounds
could be detected readily.  The measured quantities are in good agreement
with the actual concentrations (13).  The aliphatic compounds are not
susceptible to analysis using this technique.  Quinoline fluoresces weakly
but can be detected with high sensitivity using RTF (12) .  For the five
compounds that could be detected, the analysis is accurate enough for
Level 2 (14).
                                    121

-------
SASS Sample

After 1000-fold sample dilution, thirteen and ten PNA compounds could be
quantified by SF and RTF, respectively.  Quantitative analysis yielded the
results in Table 2 (13).

The four compounds present in relatively high concentrations (~10~6 M) were
phenanthrene, fluoranthene, chrysene and pyrene.  Trace constituents  (~10~7
to 10~8 M) were benzo[e]pyrene, benzo[a]pyrene, 2,3-benzofluorene, 1,2,5,6-
dibenzanthracene, dibenzothiophene, perylene, fluorene, anthracene and tetra-
cene.

It is worthwhile to note the sometimes complementary nature of SF and RTF.
On the other hand benzo[e]pyrene, dibenzothiophene and phenanthrene were
measured with better sensitivity and accuracy by RTP, while anthracene, pery-
lene and tetracene were detected only by SF.

Data quoted with two significant figures, e.g. fluorene, are sufficiently
accurate for Level 2.  The standard deviation on these data is generally
within ± 30%.  The less accurate results listed in Table 2 are still accept-
able for the ± 300% requirement for Level 1.  These data have been reduced
to a relative decade scale in Table 3.  This enables the data to be compared
to low resolution mass spectroscopy (LRMS) data obtained by another labora-
tory for Level 1 analysis.  At this low level of accuracy the degree of
matching is good.

To illustrate the type of spectral specificity that can be obtained, Figures
3 and 4 show RTP spectra of the SASS sample using two different excitation
wavelengths and two different heavy atoms.  The excitations at 365 and 343
nm are optimum for fluoranthene and pyrene, respectively.  The degree of
compound specificity is highlighted by inclusion of the spectra of pure
pyrene and pure fluoranthene.  These spectra provide convincing evidence
for the unambiguous identification of major PNA components in real-life
samples using RTP.

The analytical curves (log intensity vs log concentration) for fluoranthene
and pyrene in Figure 5 show that linearity of response is possible over
three decades of concentration.

Synthoil

Some data for Synthoil are given to show the possibilities of using synchro-
nous luminescence to screen a completely unfractionated sample for a major
PNA compound.

Pyrene is the major PNA compound in this very complex sample in a concen-
tration of 4300 ppm (15).  The RTP spectrum of the unfractionated Synthoil
using lead acetate heavy atom perturber, is shown in Figure 6.  Comparison
with the RTP spectrum of pure pyrene indicates that a rough quantitative
analysis could easily be made.  The spotting on filter paper, the drying,
and the recording of the spectrum took only a few minutes.  The technique,
therefore, can provide a means for true rapid screening of a major PNA
compound, or compounds.
                                    122

-------
                                PROGNOSIS

The two techniques of synchronous luminescence and room temperature phos-
phorescence are demonstrably useful for quickly characterizing PNA bearing,
real-life samples.  Without any fractionation, it is possible to measure
a major PNA component, such as pyrene in Synthoil, within a few minutes.

After coarse fractionation, the quantitative analysis can be extended to
include a dozen or more PNA compounds.  The results from synchronous
luninescence can clearly reinforce or expand the knowledge forthcoming
from LRMS and infrared, the two other techniques that are the basis of
Level 1 organic analysis.

There are additional advantages.  Once calibration curves have been con-
structed, the analyses are simple to conduct, rapid, inexpensive and can
be performed with commercial instrumentation.  The methodology is also
amenable to automation.

We recommend that the synchronous luminescence techniques be seriously
considered for incorporation in the EPA/IERL-RTP phased approach to
sampling and analysis for environmental assessment.

                               REFERENCES

1.  "Level 1 Environmental Assessment Performance Evaluation," EPA/
    IERL-RTP, Report No. ADL 79347-24, November 23, 1977.

2.  "EPA/IERL-RTP Procedures Manual:  Level 1 Environmental Assessment,"
    EPA-600/7-78-201, October, 1978.

3.  Dorsey, J. A., C. H. Lochmuller, L. D. Johnson and R. M. Statnick,
    "Guidelines for Environmental Assessment Sampling and Analysis
    Programs; Historical Development and Strategy of a Phase Approach,"
    Draft Revision March 9, 1976.

4.  Vo-Dinh, T., "Multicomponent Analysis by Synchronous Luminescence
    Spectroscopy," Anal. Chem., 50, 396-401  (1978).

5.  Vo-Dinh, T., R. B. Gammage, A. R. Hawthorne and J. H. Thorngate,
    "Synchronous Spectroscopy for Analysis of Polynuclear Aromatic
    Compounds," Environ. Sci. Technol., 12, 1297-1302 (1978).

6.  Vo-Dinh, T. and R. B. Gammage, "Singlet-Triplet Energy Differences
    as a Parameter of Selectivity in Synchronous Phosphorimetry,"
    Anal. Chem., _5£, 2054-2058 (1978).

7.  Vo-Dinh, T. and R. B. Gammage, "The Applicability of the Second-
    Derivative Method to Room Temperature Phosphorescence Analysis,"
    Anal. Chim. Acta, 107, 261-272 (1979).

8.  Latz, H. W., A. H. Ullman and J. D. Winefordner, "Limitations of
    Synchronous Luminescence Spectroscopy in Multicomponent Analysis,"
    Anal. Chem., 50, 2148-2149 (1978).

                                    123

-------
 9.   Lloyd, J. B. F., H. W.  Latz,  A.  H. Ullman and J. D. Winefordner,
     "Exchange of Comments:   Limitations of Synchronous Luminescence
     Spectroscopy in Multicomponent Analysis," Anal. Chem.,  52,  189-191
     (1980).

10.   Vo-Dinh, T. and J.  D. Windfordner, "Room Temperature  Phosphorimetry
     as a New Spectrochemical Mehtod of Analysis," Appl. Spectroscopy
     Rev.. 13, 261-294  (1977).

11.   Vo-Dinh, T., G. L.  Walden and J. D. Winefordner, "Instrument for
     the Facilitation of Room Temperature Phosphorimetry with a  Continuous
     Filter Paper Device," Anal.  Chem., 49., 1126-1130 (1977).

12.   Vo-Dinh, T. and J.  R. Hooyman, "Selective Heavy-Atom  Perturbation
     for Analysis of Complex Mixtures by Room Temperature  Phosphorimetry,"
     Anal. Chem., _51, 1915-1921 (1979).

13.   Vo-Dinh, T., R. B.  Gammage and P. R. Martinez,  "Analysis of a SASS
     Sample by Synchronous Fluorescence and Room Temperature Phospho-
     rescence,"  (to  be  published).

14.   "EPA/IERL-RTP Procedures for Level 2 Sampling and Analysis  of Organic
     Materials," EPA-600/7-79-033, PB 283-800, February 1979.

15.   Vo-Dinh, T., R. B.  Gammage and P. R. Martinez,  "identification and
     Quantification  of  Polynuclear Aromatic Compounds in Synthoil by
     Room Temperature Phosphorimetry," Anal. Chim. Acta, in press.
Research  sponsored by the Office of Health  and Environmental Research,
U.S. Department of Energy under contract  W-7405-eng-26 with the Union
Carbide Corporation.
                         By acceptance of this article, the
                         publisher or recipient acknowledges
                         the U.S. Government's right to
                         retain a nonexclusive, royalty-free
                         license in and to any copyright
                         covering the article.
                                      124

-------
                    TABLE 1

Artificial Complex Sample Quantitatively Analyzed
  by Second Derivative Synchronous Fluorescence
Compound
Squalene
n-Tridecane
Biphenyl
Chlorobenzene
Acenaphthene
Chrysene
Dihexylether
Dinitro toluene
Dibenzothiophene
Diethylphthalate
2-Ethylhexanol
Phenol
Quinoline
Palmitic Acid
Stearic Acid
Di-p-tolylsulf oxide
Actual Measured
concentration concentration
[mg/20ml] [mg/20ml]
12.7
14.4
13.6 15+3
20.1
14.0 13.7 ± 0.4
15.7 16.1+1.0
17.8
20.1
12.7 12.4 ± 0.6
14.4
15.4
16.7 12.0 + 1
12.6
13.3
12.2
20.4
d2 - SF
peak
[nm]


308

328
368


335


285




                        125

-------
                                TABLE 2


Concentrations of PNA Compounds in the SASS Sample Extract as Measured by

Synchronous Fluorescence (SF) and Room Temperature Phosphorescence (RTF)
Compound

Anthracene

Phenanthrene

Fluorene

Dibenzothiophene

1,2,5, 6-Dibenzanthracene

2 , 3-Benzof luorene

Fluoranthene

Chrysene

Pyrene
Tetracene

Benzo [ e ] pyr ene

Benzo[a]pyrene
Perylene


7.7

<2

1.2

<4

<5

3.3

1.9

3.2

3.0
1.9

<6

6.0
1.0

SF
-7
x 10
-5
x 10
-7
x 10
-7
x 10
_7
x 10
-7
x 10
-6
x 10
-6
x 10
-6
x 10
x 10~8
-6
x 10
_7
x 10
x 10~8
Concentration (M)*
RTF

	
-6
8 x 10
7
1.5 x 10
-7
3 x 10
_7
1 x 10
_7
3 x 10
-6
1.6 x 10
-6
3.6 x 10
-6
2.3 x 10
	
_7
4 x 10
-6
<10
	
*
 Measurements taken with a 1000-fold diluted solution.
                                    126

-------
                                TABLE 3
                Relative Concentrations on a Decade Scale
                   of PNA Compounds in the SASS Sample
Synchronous Low Resolution Mass Spectroscopy
Compound Luminescence* (Arthur D. Little, Inc.)
Anthracene /Phenanthrene
Pyrene
Benzof luoranthenes
Fluoranthene
Chrysene, Benzanthracene
Benzopyrenes
1,2,5, 6-Dibenzanthracene
2 , 3-Benzof luorene
Dibenzothiophene
Fluorene
Tetracene
Perylene
100
100
	
100
100
100
10
10
10
10
1
1
100
100
100
	
100
10
	
10
10
10
	
	
Concentration extrapolated to the original sample.
* > 1 x 10-3 M = 100
  < 1 x 10~3 but   1 x 10~4 M = 10
  < 1 x ID"1* M = 1
                                     127

-------
                                                ORNL-OWG TT-111S2R
         £ 4
         <

         o
         u
         u
         in
         o

         _l
         u.
         1  5
         $ 4
         2  3
         o
         at


         I 2
         O
         K

         U


         I 1
I  I  |  i  I  I  I  |  I  I  I  I  |  I  I  I I
                                                   1  1 T I
                                                            (a)
                      I  I  I  I  I  I  I  I  I  I I  I I  I  I  I  I  I  I  I  I
                r  i
      i  i  i  i  I  i  i  i  i  i  i  i  i  i i i i i  i  I  i
                   300        350       400       450

                                WAVELENGTH (nm)
                                          500
FIGURE 1  Conventional fixed-excitation and synchronous fluorescence

          spectra  of a synthetic mixture of five PNA compounds.
                                      128

-------
                          ORNL-DWG 78-1838 R
FLUORENE
d -SIGNAL
RTP
SPECTRUM
                  400     500       600
                    WAVELENGTH  (nm)
FIGURE 2  Comparison of the direct (d° - lower curves) and second-
       derivative (d2 - upper curves) RTP spectra of fluorene.
                       129

-------
    8
                                          ORNL-DWG 80-7717 R
 LU
 J-
 2
 UJ
 °= 3
    0
                 111(1
              RTP SPECTRA
                                        SASS SAMPLE
                                   	 FLUORANTHENE

                                   WITH LEAD ACETATE
                                     Xex = 365 nm
                                            T*  i    i   i
        500
  600
WAVELENGTH (nm)
700
FIGURE 3  Characterization of fluoranthene in the SASS sample by room
         temperature phosphorimetry (solid curve:  SASS sample; dashed
         curve:  fluoranthene;  X   = 365 nm).
                              130

-------
                                           ORNL-DWG 80-7718R
   8
u
i-
LL)
>
GJ  4
a:
  0
            I   |   I   f   I


            RTP SPECTRA
 	 SASS SAMPLE
 	 PYRENE


WITH THALLIUM ACETATE
     yex= 343 nm
                      I    I   I
              500              600
                        WAVELENGTH (nm)
             700
   FIGURE 4  Characterization of pyrene in the SASS sample by room
            temperature phosphorimetry (solid curve:  SASS sample;
            dashed curve:  pyrene;  X   =343 nm).
                         fj      ex
                            131

-------
I  10'
r>

o
k.
•»»
jQ
t  103
co
z
LJ
O.
   10
                                               ORNL-DWG 80-7716R

          1  I  I I I III)    I  I  I I I l|l|    I   I I  I I III]    I   I I  I I III


            RTP INTENSITY
            SASS SAMPLE
            o PYRENE (X,m=595nm)
              FLUORANTHENE(Xejn=545nm)
           I  I  I I Mill    I  I  I I Hill   I   I  I I Mill    I   I I I Nil
     1CT5
icr4          io~3           icr2
     DILUTION FACTOR (v/v)
1C'
   FIGURE 5  RTP analytical curves of pyrene and fluoranthene in the SASS
           sample.
                                132

-------
                         ORNL-DWG 79- 12927 R
CO
"Z.
uj  8
u
u
(T
   0
        I  I   I
                    I
I  I
         RAPID SCREENING OF SYNTHOIL BY
                       RTF
                UNFRACTIONATED SYNTHOIL
         	 PURE PYRENE
         \ex= 343 nm
                              I   I  I   I   I
              500         600

               WAVELENGTH ( nm )
                                       700
FIGURE 6  Rapid identification of pyrene in unfractionated

        Synthoil.
                      133

-------
            "APPLICATION OF CONTAMINANT ENRICHMENT MODULES
                      TO ORGANIC TRACE ANLYSIS"

      D. L. Stalling, J. D. Petty, L. M. Smith, and G. R. Dubay

           Columbia National Fisheries Research Laboratory
                   U. S. Fish and Wildlife Service
                     Route 1, Columbia, MO 65201
                               ABSTRACT

The assessment of environmental pollutants often requires the analysis
of  contaminants   present  at  trace  levels  in  complex  biological
matrices.   These  analyses  are greatly  facilitated  by  contaminant
enrichment  and concurrent removal of  biogenic  coextractives  before
analysis.   Optimum  use  of  sophisticated    analytical  systems  is
facilitated  by effective  removal  of  sample coextractives from con-
taminants.   Simplification of  complex  residues  is  achieved  using
enrichment processes that are also specific for chemical classes.   We
have  developed  a  series  of  modular  chromatographic    enrichment
procedures designed  to  be integrated  into an  automated  sequential
system,  controlled by a microprocessor,  for fractionation of complex
residues  extracted from environmental samples.   Initially removal of
most coextracted  biogenic compounds  takes place using gel permeation
chromatography (GPC), and then acidic contaminants in the GPC effluent
are retained by cesium silicate.  Another module,  carbon dispersed on
urethane foam or on glass fibers,  is used to retain planar aromatics.
The  unretained  neutrals  and basic  contaminants are  collected  and
further fractionated with a variety of adsorbents.   These  procedures
have enabled detection of chlorinated dibenzofurans,  terpenes,  naph-
thalenes,   and phenolics  in  addition to  the  common organochlorine
pollutants in fish samples by electron capture gas chromatography (EC-
GC), negative ion mass spectrometry (NI-MS), and conventional GC/MS.
                             INTRODUCTION

Complex multiclass residues are usually encountered in the analysis of
environmental samples for organic contaminants.  Analyses of this type
can be facilitated by a comprehensive approach,   including fractiona-
tion  of  contaminants  by  class.   In an effort  to  provide such  a
procedure,  we have developed and tested three modular chromatographic
procedures:   gel  permeation chromatography (GPC);   cesium  silicate
chromatography (CsSC); and dispersed-carbon chromatography.

These  modules were  designed  to  achieve  an  integrated  separation
process  suitable  for  automation.   Space  and manpower  limitations
encountered  in  many  laboratories  increase the  need for  automated
sample cleanup  and  analysis.   Improved  routine  purification  cap-
abilities  made possible  by  chromatographic  separation of  chemical
classes broadens the scope of contaminants amenable to analysis.

                                    134

-------
Two factors that have  limited the progress in the development of such
automated cleanup techniques are lack of module compatibility and lack
of applications using flexible electronic  controllers  for automation
of chromatographic procedures.   We have undertaken the development of
methods that can accommodate sequential chromatographic  processes for
fractionation   of complex chemical residues.   Contaminant enrichment
and fractionation  are achieved by sequential linkage of the described
chromatographic modules.  This approach was specifically developed for
the enrichment  of  contaminants  from aquatic organisms—particularly
fish tissue (1).

We have measured the effectiveness of these chromatographic enrichment
modules by determining a contaminant enrichment factor (CEF) (2). This
factor  is defined  as  the ratio  of the  weight of initial extracted
sample to  the weight of residual coextractives  present after proces-
sing multiplied by the fractional recovery of the contaminant(s) under
study.   We defined this ratio  as  the contaminant enrichment  factor
(CEF) (2).

Normally, the percent recovery of specific contaminants is stressed in
describing a cleanup procedure.  However,  the CEF associated with the
procedure(s)  involved in sample cleanup offers an additional means of
evaluating the  overall effectiveness of the methods.   The larger the
ratio the greater the matrix rejection.  Use of both recovery data and
CEF will give a better  comparison of  procedures  for  enrichment  of
complex  environmental residues,   enabling analysts to  evaluate pro-
cedural alternatives more accurately.

To automate contaminant  enrichment using our chromatographic modules,
we contracted for  the  development  and  construction  of a prototype
programmable    chromatographic separation  system (3).   This  system
consists of a solvent delivery subsystem,  a sample storage subsystem,
a low pressure chromatograph, a fractionation unit, and a controller.

The  controller is  a microcomputer  interfaced  to  a chromatographic
system through various  solenoid-activated valves.   The  computer  is
programmed to closely control the chromatography and still provide the
module selection to achieve the separations required. This flexibility
is provided  through  independent control of the solvent and flow rate
used for elution, introduction of sample,  which column(s)  used,  the
size and  number of fractions  collected,   and the order in which the
controller runs the chromatography steps.

                          MODULAR PROCEDURES

Gel Permeation Chromatography (GPC)

GPC  is a  form  of liquid  chromatography in which solute molecules -
pollutants and biogenic materials -  are selectively  retarded as they
permeate  the pores  in  the  column  packing.   Generally the  larger
nonaromatic  lipid molecules are excluded from all or a portion of the
pores by  virtue of their  size,   and therefore elute from the column
before the  aromatic or  smaller contaminant  molecules.   Long  chain
aliphatic hydrocarbons  also have little interaction with these resins
                                   135

-------
and are well separated from aromatic hydrocarbons. GPC gels have three
advantages:

    1.  they can be used with a wide range of solvents,
    2.  they do not require complex solvent gradients, and
    3.  they can be used repetitively for non-polar compounds
        without chemical regeneration.

The technique,  which has been automated (2),  has been applied to the
isolation of contaminants,  such as pesticides,  phenols,  PAHs,   and
chlorinated aromatics—including chlorinated  dibenzo-p-dioxins (CDDs)
and dibenzofurans (CDFs)—from fatty acids,  oil components,   trigly-
cerides, and other biological materials (2,3,4). The CEFs for GPC have
ranged from 50 to 200 for lipid-rich extracts (2).

Cesium Silicate Chromatography

Simplification and automation  of procedures for  isolation of  acidic
pollutants such  as phenols is critically  needed  for the analysis of
large numbers  of samples.   Our modification of a method published by
Ramiljak et al.   (5)   has circumvented many problems associated with
isolation of acidic materials (6). Application of this procedure using
columns  of silica  gel treated  with  cesium  hydroxide (Cs-silicate)
yielded significant improvements in the selective  recovery of phenols
from environmental samples.   These columns,   when placed in sequence
with  the  GPC eluate  (1:1,   v/v,   methylene  chloride/cyclohexane)
consistently retained  the  11  priority  pollutant phenols plus  2,4-
dichloro- and  2,4,5-trichlorophenoxy acetic acids  (4).    A  solvent
mixture,   5  mL  of  40%  methanol  in  acetone,   is  sufficient  to
quantitatively recover all of the priority pollutant phenols.  Neutral
compounds such as PCBs, CDFs,  and CDDs are not retained.  The columns
are regenerated after they  are washed with  40 mL of the GPC  solvent
mixture.

Phenolics  are  determined as their anisoles or their pentfluorobenzyl
ethers by electron capture GC or negative ion GC/MS. This procedure is
currently being evaluated for its ability to separate acidic compounds
in samples provided  from  our  laboratory's monitoring  programs  for
environmental contaminants in fish.

Dispersed-Carbon Chromatography

Our development of  a method for  dispersing finely divided carbon  on
the surface of shredded polyurethane foam (7)   is an improvement over
granular carbon columns which required large solvent volumes  to elute
strongly adsorbed contaminants.   The carbon particles adhere strongly
to the foam and are not dislodged during Chromatography. The increased
porosity of this support results in increased solvent  flow,   greatly
facilitating the use of carbon as an adsorbent.

Although  shredded urethane foam  is an excellent support  for carbon,
the  foam can be  degraded  by ultraviolet  light.   In  an effort  to
overcome these problems,  we examined several alternative supports for
the dispersal of carbon.
                                   136

-------
Sample interferences  are reduced by using  carbon dispersed on  glass
fibers.  Consequently,  a large-capacity contaminant enrichment module
based  on  carbon-glass  fibers  for  isolating  planar aromatic  con-
taminants was developed.   We also explored  the  combination of  this
module with  other auxiliary  cleanup procedures that could be readily
incorporated into a sequential configuration (2).

            APPLICATIONS OF CONTAMINANT ENRICHMENT MODULES

Analysis of CDDs and CDFs in Aquatic Samples

Recently,   we have  applied our modular approach to the enrichment of
CDFs and CDDs from extracts of  aquatic organisms.   CDFs are produced
when PCBs are subjected to incomplete combustion as in open burning or
pyrolysis,   and  CDDs  are  known contaminants in several widely used
chemicals (8).   Detection of CDFs and CDDs was sought because several
of their isomers have been shown to be a highly toxic to mammals (9).

Demonstrating the occurrence of CDFs and CDDs  in the aquatic environ-
ment  has  proved to be a formidable task for  analysts because  these
contaminants occur  in  very  low concentrations,     and  interfering
components are present in fatty tissues (10). Highly effective cleanup
techniques  are  required to  enrich  extracts  containing  these con-
taminants.   Using the dispersed  carbon module sequentially linked to
the GPC module,  we demonstrated that the CEF was approximately 10,000
for CDDs and CDFs (2).   Use of these modules thus provided a means of
isolating  these  compounds in quantities suitable  for  screening  by
negative ion mass spectrometry,   by direct probe or  GC/MS.   The  MS
analyses for CDD and CDF residues in selected  fish,   turtle and seal
fat samples were performed by Dougherty and coworkers (11).

In these analyses, CDFs having 4 to 8 chlorines were detected for  the
first time  in  aquatic  environmental samples.   The samples  studied
included relevant blanks,   as well as fish samples  from  the Hudson,
Ohio, and Connecticut rivers and Lake Michigan, containing PCBs in the
range of 20-200 ppm.

In addition,   fat  of snapping turtles from  the Hudson River  and of
Baltic  gray seal from the Gulf of Bothnia were examined.   The turtle
fat   was   found   potentially  contaminated  by  penta-  and   hexa-
chlorodibenzofurans  and  seal fat by  pentachlorodibenzofurans.   The
concentration of PCBs exceeded 750 ppm in  the turtle fat samples and
was  approximately 10 ppm in the seal  fat.   Chlorinated naphthalenes
were also detected in all samples analyzed.

A more definitive analytical  procedure was used  because  the  direct
probe high resolution MS technique  employed is only  a screening tool
for CDFs.   The method of choice was high resolution GC/MS which Buser
and Rappe in several publications (12,13,14)  have recently applied to
the separation of CDD and CDF isomers at levels corresponding to part-
per—trillion residue concentrations.

The  snapping turtle and seal  fat samples  were further  analyzed  by
Rappe et al.  (14).  Aliquots (2 ul) corresponding to 0.1 and 0.8 g of
                                    137

-------
fat of the snapping turtle and seal, respectively,  were injected into
a 50 m glass capillary column (OV-17 or Silar IOC) coupled by way of a
platinum  capillary  interface  directly  into  the  ion source  of  a
Finnigan 4000 quadrapole  GC/MS.   The instrument was  operated in the
negative chemical ionization mode and methane was used as the  reagent
gas.   Under  these  conditions  the  CDFs  exhibit  intense  negative
molecular  ions (M~)   with limited  fragmentation  resulting from the
addition of H and loss of Cl, to yield (M-34)~ ions.

Complete spectra (m/e = 80-800)   were recorded for the major GC peaks
observed.   Identification of a GC/MS peak as  a  CDF was based on the
presence of the proper M~ ions and intensities of the ion clusters due
to the chlorine isotopes.   Identification of isomers was based on the
retention times of authentic  standards on both  OV-17  and Silar  IOC
columns.

The total CDFs in the turtle fat sample was estimated at 3  ng/g.   Of
the individual CDF isomers,   the toxic  2,3,7,8-tetra- and 2,3,4,7,8-
pentachlordibenzofuran were estimated  to  be  0.045  and  0.62  ng/g,
respectively.   The  levels of CDFs in  the seal  fat sample were much
lower than  in  the turtle fat.   In this sample the  total  CDFs were
estimated  as  0.04  ng/g,   and  the  2,3,7,8-tetra-  and  2,3,4,7,8-
pentachlorodibenzofuran as 0.001 and 0.015 ng/g, respectively.

       ANALYSIS OF POLYCHLORINATED TERPENES (TOXAPHENE) IN FISH

Toxaphene is an  insecticide that  has  been  produced in  very  large
amounts primarily for use in controlling cotton pests (15). Because it
is  a  complex  multicomponent  mixture  of  over  150 polychlorinated
terpenes, it has often escaped detection in aquatic residue surveys.

We have examined the occurrence and nature of  polychlorinated terpene
(PCTerp)   residues in fish  samples collected as part of the National
Pesticide Monitoring Program.   Residues derived  from toxaphene  have
been  frequently encountered in fish from  the Southern United States,
at levels as high as 20 - 100 ug/g (4).

Our  awareness of  PCTerps  as  widely dispersed contaminants  in  the
aquatic environment began in 1976,   when extracts from lake trout fry
were referred to our laboratory by the  Great  Lakes  Fishery Research
Laboratory,  Ann Arbor,  Michigan,  for examination by electron impact
mass spectrometry.   Toxaphene-like residues  were  found,   yet it is
believed  that very little toxaphene had been  used in  the watershed.
These  data  posed  several questions as  to  the nature  of the  con-
taminants so detailed analysis was required.

Characterization  of  the  residues  in  these  samples  was difficult
because of the presence of many other contaminants (DDE,  DDT,   PCBs,
etc.)   which were not separated from  PCTerp's.   An  improved sample
cleanup  scheme and  means of  detection was  needed to eliminate  the
interferences to  toxaphene-like  contaminants.   A  cleanup procedure
using nitric acid- sulfuric acid treatment  of extracts was adapted to
this problem and  it  proved suitable  for removal of  most  compounds
(including  PCBs,     though   they   are   removed  by  silicic  acid
                                    138

-------
chromatography)   interfering with  the detection of  PCTerps  in fish
extracts. However, most chlordane components were not removed and they
are present as early eluting GC interferences to PCTerp analysis.

Screening  of  samples  for  PCTerps  can  be  done  with GC  by using
temperature programming  and an electron capture  detector.   Although
toxaphene or  PCTerps can be detected by electron  capture GC,   their
identity cannot be confirmed by this technique.  Electron impact-MS is
not  a  good  means of  confirmation for  PCTerp residues  in  complex
environmenal samples because  their fragmentation patterns are complex
and  response  factors are  low.   Before definitive  analyses of  the
residues in environmental  samples could be made by GC/MS,   both  the
sensitivity  and  specificity of  the  GC/MS  technique  needed to  be
enhanced.

The complex fragmentation of PCTerps  in  electron  impact mass spect-
rometry  makes it difficult to obtain clearly defined molecular weight
data.   CI-MS had  such poor  response to PCTerps that it  is not very
useful  in  the  analysis  of  environmental  samples (16).   Negative
ionization    mass spectrometry (NI-MS)   with glass capillary  chrom-
atography columns,  significantly improved our ability to characterize
PCTerp residues.

The  use  of  NI-MS,   in  contrast  to EI-MS,   yields  both  greater
sensitivity  and  less  molecular fragmentation (17).   Thus NI-MS has
distinct advantages over other techniques for analysis of toxaphene or
other PCTerps in environmental samples.

This  technique  was  applied  to several sample  extracts  containing
PCTerp residues.  Residues in fish from Llano Grande Lake, Texas, Lake
Michigan,   and in fat  from Baltic Sea seals were compared  with each
other and with toxaphene.  NI-MS spectra of PCTerps exhibited enhanced
sensitivity relative to chlordane components and revealed similarities
between the PCTerp residues in all of these samples.

Environmental residues contained  several  isomers  of the more highly
chlorinated PCTerps,   which were not detected in toxaphene.   Further
research  on PCTerp residue analysis  is  directed  toward obtaining a
more comprehensive profile of the PCTerp residues which are present in
aquatic ecosystems  and  obtaining  some  insight into  the  transport
mechanisms that have led to their global dispersal.

                               SUMMARY

We  have  defined  contaminant enrichment  as  concentration  of trace
environmental contaminants(s)   concomitant  with  at  least  partial
elimination of biogenic  material or other interfering  chemicals.   A
contaminant enrichment module,   then is  a process  unit  designed to
provide an increase concentration of the contaminant(s)   relative  to
the total material present.  Ideally, these modules are designed to do
this  in an   integrated fashon.   Mechanization and/or automation  of
contaminant  enrichment can  readily  result  from  the  ability  to
sequentially link different chromatography modules.

                                   139

-------
Measurement  of CEFs provides  insight  into the potential  utility of
individual  and integrated  cleanup procedures.   Modules should  be
optimized for  the isolation  of specific  contaminants  or classes of
chemicals from any other part of the residues that might interfer with
the quantitation of particular trace organic residues in environmental
samples.

We have developed  several  modules which provide an  automated system
for enrichment  and fractionation  of environmental contaminants found
in aquatic  ecosystems.   We  have have designed  and evaluated  three
modules:

   1. Automated gel permeation chromatography for removal of
      lipids and other biogenic compounds, and the consequent
      enrichment of a wide variety of contaminants

   2. Cesium silicate chromatography, for the enrichment of acidic
      contaminants or the removal of acidic compounds before
      carbon chromatography is used

   3. Carbon chromatography (Amoco PX-21 carbon) dispersed on
      shredded urethane foam or glass fibers, for the enrichment
      of planar aromatic contaminants such as chlorinated
      dibenzofurans, dibenzo-p-dioxins, and naphthalenes.

Application  of  improved  contaminant enrichment  processes  to trace
organic analysis of environmental samples has increased our ability to
undertake  difficult analyses for  low  levels of  contaminants (e.g.,
CDFs,  CDDs,  and chlorinated napthalenes)  and has provided new means
for isolating acidic  contaminants such as phenols and phenoxy  acids.
These modules have proved to be complementary to newly developed modes
of mass spectrometric detection in which NI-GC/MS is used.

Through  cooperation  with  other investigators  we have  extended our
knowledge of the occurrence of polychlorinated terpenes, dibenzofurans
and  naphthalenes in fish  and other  aquatic organisms.   We  used  a
combination  of  these  modules for the  enrichment  of  chlorinated
dibenzofurans in environmental samples that contained high  concentra-
tions  of PCBs.   Using NI-GC/MS,   we  detected numerous dibenzofuran
isomers  samples of fat from snapping turtles of  the Hudson River and
from seals of the Baltic Sea.   The highly toxic 2,3,7,8-tetra-,  and
2,3,4,7,8-pentachlorodibenzofurans were among  the isomers identified.
We  have  also  applied  negative  ion  GC/MS  to  the  detection  and
characterization  of polychlorinated terpenes  in  fish from the Great
Lakes and rivers of the Southern United States.
                                    140

-------
                          BIBLIOGRAPHY

1.  Stalling, D. L., L. M. Smith, and J. D. Petty, "Approaches to
    Comprehensive Analyses of Persistent Halogenated Environmental
    Contaminants," in Measurment of Organic Pollutants in Water and
    Wastewater. C. E. van Hall, (ed.), ASTM, Philadelphia, PA, 320-
    323 (1979).

2.  Stalling, D. L., J. D. Petty, L. M. Smith, and G. R. Dubay,
    "Contaminant Enrichment Modules and Approaches to Automation of
    Sample Extract Cleanup," in Environmental Health Chemistry.
    Chemistry of Environmental Agents as Potential Human Hazards.
    J. D. McKinney, (ed.), Ann Arbor Publishing Company, Ann ArboR
    MI, in press (1980).

3.  Hartley, J. W., "A Programmable Chromatographic Separation
    System: Specifications and Design," Ph. D. Dissertation,
    University of Missouri - Columbia (1979).

4.  Ribick, M. A., L. M. Smith, G. R. Dubay, and D. L. Stalling,
    "Applications and Results of Analytical Methods Used in Envir-
    onmental Monitoring," in Aquatic Toxicology, D. Branson and K.
    Dickson, (ads.), ASTM, Philadelphia, Pa, in press (1980).

5.  Ramiljak, Z., A. Sole, P. Arpino, J. Schmitter, and G.
    Guiochon, "Separation of Acids from Asphalts," Anal.
    Chem., 49, 1222-1225 (1977).

6.  Stalling D. L., J. D. Petty, and L. M. Smith, "Chromatographic
    Enrichment of Acidic Compounds using Alkali Metal Silicates,"
    J. Chromatogr. Sci., in press (1980).

7.  Huckins, J. N., D. L. Stalling, and W. A. Smith, "Foam-Charcoal
    Chromatography for Analysis of Polychlorinated Dibenzodioxins
    in Herbicide Orange," j_. Assoc. Off. Anal. Chem., 61, 32-38
    (1978).

8.  Rappe, C., H. R. Buser, and H. P. Bosshardt, "Dioxins,
    Dibenzofurans and Other Polyhalogenated Aromatics: Production,
    Use, Formation, and Destruction," in Health Effects of
    Halogenated Aromatia Hydrocarbons, W. J. Nicholson and J. A.
    Moore, Eds., New York Academy of Sciences, New York, NY, 320,
    1-18 (1979).

9.  Poland, A., W. F. Greenlee, and A. S. Kende, "Studies on the
    Mechanism of Action of the Chlorinated Dibenzo-p-dioxins and
    Related Compounds," in Health Effects of Halogenated Aromatic
    Hydrocarbons. W. J. Nicholson and J. A. Moore, Eds., New York
    Academy of Sciences, New York, NY, 320, 214-230 (1979).
                                  141

-------
10.  Zitko, V., 0. Hutzinger and P.  M. K. Choi,  "Contamination of
     the Bay of Fundy - Gulf of Maine Area with  Polychlorinated
     Biphenyls, Polychlorinated Terphenyls, Chlorinated Dib-
     enzodiozins, and Dibenzofurans," Environ. Health Perspect. 1,
     47-50(1972).                                             ~

11.  Dougherty, R. C., L. M. Smith,  D. L. Stalling,  C. Rappe,  and
     D. W. Kuehl, "Negative Chemical lonization  Studies of Polych-
     lorodibenzo-p-dioxins and Dibenzofurans in  Environmental  Samp-
     les," Abstracts of papers, 178th National American Chemical
     Society Meeting, Washington, DC, ISBN 8412-0517-5, Sept.  9-
     14th, Environmental Chemistry Section, No.  38 (1979).

12.  Buser, H. R., and C. Rappe, "High Resolution Gas Chromato-
     graphy of the 22 Tetrachlorodibenzo-p-dioxin (TCDD) Isomers,"
     Submitted to Anal. Chem. (1980).    ~

13.  Buser, H. R., "Analysis of Polychlorinated  Dibenzo-p-dioxins
     and Dibenzofurans in Chlorinated Phenols by Mass Frag-
     mentography," J_. Chromatogr. 107, 295-310  (1975).

14.  Rappe, C. H., R. Buser, D. L. Stalling, L.  M. Smith, and  R. C.
     Dougherty, "Polychlorinated Dibenzofurans  in Environmental
     Samples—A New Class of Toxic Pollutants Associated with
     PCBs," submitted to Nature (1980).

15.  Hooper, N. K., B. N. Ames, M. A. Saleh, and J.  E. Casida,
     "Toxaphene, a Complex Mixture of Polychoroterpenes and a  Major
     Insecticide, Is Mutagenic," Science. 205,  591-593 (1979).

16.  Stalling, D. L., and J. N. Huckins, "Analysis and GC-MS
     Characterization of Toxaphene in Fish and Water," EPA-600/3-
     766-076, pp. 43, (1976).

17.  Dubay, G. R., D. L. Stalling, M. A. Ribick, J.  D. Petty,  and
     C. Rappe, "A Comparison of Negative lonization- and Electron
     Impact-MS for analysis of Toxaphene-Like Residues in the
     Environment," presented at the 28th Annual  Conference on  Mass
     Spectrometry and Allied Topics, American Society for Mass
     Spectrometry, New York, NY, May 25-30, (1980)
                                    142

-------
      A MULTICHANNEL, REMOTE CONTROLLED, TEFLON AND GLASS POSITIVE
        DISPLACEMENT APPARATUS  FOR COLLECTING TRACE ORGANICS FROM
                          ENVIRONMENTAL SOURCES

                   D. C. Tigwell and D. J. Schaeffer*
                Illinois Environmental Protection Agency
                           2200 Churchill  Road
                      Springfield, Illinois   62706
                                ABSTRACT

The recent, rapid interest in isolating and identifying trace organic
compounds from environmental sources has focused largely on studies of
adsorbents, extraction techniques, and chromatographic methods.  As we
have recently shown (4). the most critical factors in the sequence are
the most difficult to control:  the reproducible capture of compounds
from a complex matrix and the subsequent recovery of compounds from
the adsorbent.  Further, the volume reproducibility and type of
sampling-grab, flow, or time-proportional-can substantially affect the
validity of the sample to represent the source (1).

We have designed an all glass and teflon multichannel positive
displacement sampler which is driven by 50 mL syringes held in
precision machined mounts.  Up to four combinations of sources and
collection methods, such as simultaneous sampling of a pre and post
process wastewater with each stream being independently collected on
e.g., XAD resin or carbon, can be applied.  Internal standards can be
precisely metered into the sample prior to the collection device, and
the effluent from an adsorbent bed can be collected for independent
anaylsis.

This apparatus has been used to collect samples of raw and finished
drinking water, samples from the Illinois River, and from municipal
and industrial effluents, including an untreated coal conversion
process wastewater.  Four radio-controlled units are being used in a
nationwide EPA-funded study of the movement of organics through
municipal treatment plants (8-10).

We will describe the design and construction of the apparatus and its
use for the simultaneous sampling of trace organic compounds on XAD
and granular activated carbon.
    *Address correspondence to this author.
                                  143

-------
                              INTRODUCTION
The ability to identify and quantify organic compounds present at
trace levels in the environment, and the use of these data for
assessing environmental impact, depends upon the least accurate or
reproducible segment of the sampling and analysis chain.  Although
sorbents for concentrating organic residues from the aqueous matrix
have been developed and analyses have improved, methods for collecting
samples are primitive.  Further, most of the developments have been
empirical since the theoretical constraints on the sampling process
and in the propagation of errors through the chain are only now
beginning to be identified.

Sampling Theory

Schaeffer and Janardan (1) and Schaeffer et al. (2) have examined the
nature and quality of information obtained from grab and composite
samples.  Thus, if the analyzed concentration of a single grab (or
sub-) sample is X, then X will have a certain distribution (of
concentrations) with a mean u and a variance O .
    (1)  E(X) = u, var(X) = a2.
The information content, termed the invariance (3), of a single grab
is given by the reciprocal of the variance,
    (2)  Iv = I/a2.
If X}, X2,  .  .  •• Xn are the values of n subsamples, and w],
W2,  .... wn their proportions in the composite such that 2,w£
= 1,  then the result of compositing n discrete  (sub)samples is
    (3)  Y = W! Xj + w2 X2 +  .  .  . + wn Xn.
From this it has been shown  (1,2) that the mean  and variance  of  the
composite are
     (4)  E(Y) = u, Var(Y) = a2  (1 + n20w2)/n,

                                    144

-------
        o
where Cw^ is the variance  of  the  compositing proportions W£.   Since

the  information content of the  composite  is Iy =  (var(Y))~l,  the
relative information of a  composite and grab sample  is
    (5)  Iy/lx = n/(l + n2 aw2).
When precisely equal subsample  sizes  form  the composite  so  that wj  =
                        r\
W2 = •  •  • = wn, then awz = 0 and  the composite  contains n  times  as
much information as does the single grab.  However,  if composites are
                                r\
formed  so that W£ = w;, then 0W   >0 and the advantage of

compositing is reduced.  If aw2 >  (n-l)/n  , the  information
content of a composite is less  than that of a grab sample.

Since the compounds of interest are assumed to be present at  low
levels, it is unlikely that toxicologic data in  the  range of  interest
exists.  Further, because the levels are low, the ability to  capture,
isolate, identify and quantitate the compounds present in a single
sample  is poor as shown by Janardan and Schaeffer (4) and Janardan  et
al. (5).  (The problem is increased up to n-fold by  taking  a  grab
sample rather than a composite  with w.[ = w;, i.e., aw  = 0).
Consideration of the objectives for monitoring the environmental
levels of organic compounds, coupled with theoretical results, and
backed by field experience described subsequently, shows that load and
not concentration information is required.

With large resources load information can be obtained without
compositing.  Since, for example,  the estimated  analytical  costs alone
for the priority pollutants exceed $2000/sample  (6), most sources will
probably only be characterized  by  a single sample; this  is  best
accomplished by a composite.

The the analyst ultimately reports the concentration, C, of the
compounds in the sample." Hence, the total flow, V,  over the  sampling
period must be measured in order to estimate load.   The estimated load
^
L is computed as
    (6)  L = C V.
                                    145

-------
Typically, flow proportional composites are taken.  However, such
composites provide biased estimates of load (7).  Thus, the concentration
of a component in the composite is
    (7)  C =
where w£ is the proportion of the ith subsample, with concentration
G£, in the composite.  The load is the sum of the subsample loads
    (8)
From Eqs.  6-8, the load obtained from a flow proportioned composite with
subsamples of size w^=W£ is estimated from the measured concentration
as
    (9)
Since the true load estimate L is a linear function of the subsample
loads (Eq. 8), while the load estimated from the concentration of the
flow proportioned composite is a non-linear function of subsample loads
(Eq. 9),  flow proportioning results in a biased load estimate.  This is
in addition to the loss of information due to compositing given by Eq. 5.

The conclusion from theory, which is supported by field studies, is that
the type  of information which will be most useful for assessing the
environmental impact of trace levels of organic compounds is  loading data
obtained  from constant volume (time proportioned) composite samples.
Taken together, the theoretical results define the sampling requirements
which consequently specify the operational requirements of the composite
sampling  device.  The purpose of this paper is showing how the sampling
and operational requirements are met by a new composite sampler.
                                     146

-------
                        SAMPLER DESIGN AND OPERATION

Design Requirements

The sampling system has to be field portable.  Since samples will be
taken over hours to days, it must be able  to operate unattended.  This
requires that it be reasonably immune to  suspended  solids, that it be
capable of operating at below freezing temperatures, or in hot weather
with sample cooling.  Further, since many  of the compounds are chemically
reactive or physically sorb on wettable surfaces, all surfaces in contact
with the sample have to be relatively inert and non-wettable, such as
teflon and glass.

No single concentration/recovery sequence  works for all compounds (5).
Also, it may be important to simultaneously characterize more than one
source.  These require equipment capable  of taking  equal volume samples
using any combination of sources and collection methods, simultaneously.

A review of commercially available composite wastewater samplers revealed
none ideally suited to our purposes.  Consequently, we designed, built,
and are using a multichannel sampler.  It  is light  weight and rugged,
easy to repair in the field, and it meets  the other constraints and needs
identified above.

Design and Operation

The sampler box and drive system are aluminum.  The pump and plumbing are
teflon and glass.  The drive mechanism is  shown in  Figure 1.  The
cylinder piston combinations are standard  50 cc infusion syringes which
are held tightly and accurately in machined mounts  (Figure la).  The
plungers are removably fitted to the moving drive arm which is connected
by a rod (Figure Ib) to a pinion (Figure  Ic) on the drive cam (Figure
Id).  Sample volume/cycle is determined by the position of the pinion on
the precisely machined drive cam.

The syringes are connected via teflon lines to the  common ports on the
four channel hydraulically isolated two position valve shown in Figure
2.  The body of the valve is machined teflon lined  with a precision bore
glass tube (Figure 2b) which has had ports diamond  drilled through it.  A
teflon valve spool which is relieved to allow bi-directional flow and
stiffened internally by a stainless steel  control rod, slides inside the
tube.  The control rod is fitted with a spring return and is connected to
a large solenoid (Figure 2a).

In typical operation, sample collection is triggered by a signal received
by the electronics from an internal or external source.  This latches a
relay which places the drive cam in motion.  As the drive cam turns
through the first 180 degrees, sample is  drawn through the intake port of
the valve (Figure 2b) and into the syringes.  A microswitch is then
tripped energizing the solenoid which drives the valve spool to the
sample expulsion position shown in Figure  2b.  At this point (180-360°)
the syringes are closing.  Fluid which has collected in the syringes and
in the lines between the syringes and the  valves is forced to the outlet
port of the valve.  As the drive cam completes its  360 degree cycle,
                                    147

-------
another microswitch is tripped which turns off the pump motor and
de-energizes the solenoid which allows the valve to return to its
original position.  On a given cycle, 10, 20, 30, or 40 mL of sample can
be passed through each channel of the system.  Ten or more samples of a
given size can be taken each minute.  This pump can, therefore, deliver
large volumes of sample in a short period of time.

Extremely accurate and reproducible sample volumes (i.e., Sw2 ~ 0)
can be delivered at pressures up to 50 psi, regardless of outlet
head.  This is particularly important when solid sorbent columns are
used since the head loss across the columns can increase dramatically
over the length of the run as suspended solids collect toward the
inlet side of the columns.  While differences in the column packing
(e.g., open channel, carbon, etc.) mesh can cause different delivery
pressures at each channel from the start of sampling, accumulation of
suspended matter during the course of the run can cause widely
differing outlet pressures on each channel of the valve.  Delivery
rates of peristaltic and bellows-type positive displacement pumps are
highly dependent on outlet head.  This results in a bias favoring the
earlier subsamples.  Piston type pumps, however, are largely immune to
delivered volume variations due to outlet head.

Sampler electronics are located remotely in a control box.  The motor
and solenoid valve are controlled by electro-mechanical relays.
Timing is accomplished through the use of microswitches located around
the drive cam.  Power is supplied through a simple regulated 12 volt
supply.

The sampler must be able to sample both on a time and flow
proportional basis.  This is accomplished by using a high impedance
relay driver which actuates the (relatively) high current relay coils
using the small current available from the digital timer or from the
dry contact read relays of commercially available flow meters.  The
interface, which is shown schematically in Figure 3, is capable of
initiating a sampler cycle from a positive logic signal from
transistor-transistor logic or from CMOS Logic at a signal level of
approximately 0.5mA.  Alternatively, an input is also provided for
extremely low current signals from dry contact reed relays installed
in most flow meters.

This high impedance input can be used for switching the sampler on
when a particular liquid level is reached by using a simple probe-type
level monitor consisting of two wires.  The high impedance input of
the interface requires a current of about 1 microamp to initiate
sampling.

The present version of the timer is PMOS/CMOS quartz controlled type
capable of timing in 16 ranges from 1 millionth of a second to 1 pulse
every 464 days over 4 decades.  It requires almost no power.
                                     148

-------
                              APPLICATIONS

Nationwide Study of Sewage Treatment Plants

A computer-based system of 4 samplers is being used by DeWalle and
Chian et al. (8-10) to study the movement of priority pollutants
through municipal wastewater treatment plants.  Samplers are placed at
strategic points in the plant:  i.e., at the headworks, after the
aeration basins, prechlorination, and post chlorination.  The sampler
control boxes are plugged into radio receivers.  An ultrasonic
flowmeter positioned near the headworks is plugged into the computer.
The computer queries the operator on the flowmeter interval desired,
and plant hydraulic data is entered through a standard teletypewriter
keyboard/CRT.  Upon keyboard command, sampling is initiated by the
computer.

The computer follows a 24-hour slug of water through the plant in real
time and samples it in a flow proportional manner.  The firmware
currently assumes plug flow, but any hydraulic model can be used.
Recirculation can be accounted for, etc.

Coke Plant Sample

In our studies (11), the sampler has been used for collection of
samples from industrial process effluents surface water, and POTW
effluents.  The most recent sample was a time proportioned composite
of an untreated coking plant wastewater.  Samples were simultaneously
collected over 20 hours on macroreticular resins and in glass
bottles.  An internal standard solution of 2-ethylhexanoic acid and
perdeuteroanthracene in absolute ethanol was continuously metered into
the intake line of the sampler.  The sampler performed well on the
very hot and heavily silted sample.  The data from this sample will be
reported elsewhere (12).

                               CONCLUSION

Based on experience gained over several years of sampling for trace
organics, and on the theoretical statistical aspects of the sampling
process, we have designed and built a unique water sampler.  It is
capable of meeting theoretical requirements such as accurate sample
volume, and the practical requirements of compactness, portability,
and ease of service.  Experience is now being gained with this
equipment by us (5, 11-12), and with a computer controlled system by
DeWalle and Chian et al. (8-10) in a nation-wide study.

This equipment brings the art of sampling up to the standards required
by the subsequent chemical analysis.

                            ACKNOWLEDGEMENTS

We thank LeVerne Hudson, Roger Kanerva, and Mike Mauzy.  Konanur G.
Janardan and Harold W. Kerster deveoped the mathematical sampling
theory.  Foppe DeWalle and Ed Chian enabled the development and
testing of the computer controlled version.  Satu Somani, William
                                    149

-------
Glave, Robert Teece, Gerald Mack and James Johnston contributed to the
testing of the original version.  William Lawrence and Owen Ray did
the glassblowing and machining.   Glen Berry prepared the figures and
Debbie Ray and Sheila Hinkley typed the manuscript.

                               REFERENCES

1.  Schaeffer, D. J. and Janardan, K. G. , "Theoretical Comparison of
    Grab and Composite Sampling Programs," Biom.  J., 20, 215-227
    (1978).                                          ~~

2.  Schaeffer, D. J., Kerster, H. W. and Janardan, K. G.,  "Grab Versus
    Composite Sampling:  A Primer for the Manager and Engineer."
    Environm. Management, In Press (1980).

3.  Finney, D. J., Probit Analysis, 3rd ed., Cambridge University
    Press, Cambridge (1971).

4.  Janardan, K. G. and Schaeffer, D. J., "Propagation of Random
    Errors in Estimating the Levels of Trace Organics in Environmental
    Sources," Anal. Chem. , 51_, 1024-1026 (1979).

5.  Janardan, K. G., Schaeffer,  D. J. and Somani, S. M., "Efficiencies
    of Liquid-Liquid Extraction, Carbon, and XAD-Absorption in
    Isolating Organic Compounds from Environmental Sources," Bull.
    Environm. Contain. Toxicol.,  24, 145-151 (1980).

6.  Environmental Protection Agency, "Guidelines  Establishing Test
    Procedures for the Analysis of Pollutants; Proposed Regulations,"
    Federal Register, 44, 69464-69575 (1979).

7.  Kerster, H. W., Schaeffer, D. J. and Janardan, K. G.,
    "Flow-Proportioned Composite Samples Bias Estimates of Load,"
    unpublished study (1979).

8.  Chian, E. S. K. and DeWalle, F. B., "Analytical Methods for
    Priority Pollutants in Municipal Sewage and Sludge," Water
    Pollution Control Federation, Abstracts 52nd  Annual Meeting,
    Houston, Session 25 (1979).

9.  DeWalle, F. B. Kalman, D. A., Perera, C. and  Chian, E. S. K.,
    "Priority Pollutant Removal Efficiencies in POTW's as Related to
    Their Physical-Chemical Properties," Water Pollution Control
    Federation, Abstracts 52nd Annual Meeting, Houston,Session 8
    (1979).

10. DeWalle, F. B. and Chian, E. S. K., "Presence of Priority
    Pollutants in  Sewage and Their Removal in Sewage Treatment
    PMnts.  First Annual Report to U.S. Environmental Protection
   159 >.ncy," University of Washington, Seattle (1979).

11. Somani,  S. M., Teece, R. G. and Schaeffer, D. J., "Identification
    and Promoters  in Industrial Discharges into and in the Illinois
    River,"  J. Toxicol. Environ. Health, £, 317-333 (1980).
                                    150

-------
12.  Tigwell,  D.  C.  Schaeffer,  D.  J.  and  Somani.  S.  M..  "Trace  Organic
    Residues:   Composite Sampling and Analysis  of Coke  Plant
    Effluent," American Chemical  Society,  Abstracts 14th Annual  Great
    Lakes Regional  Meeting,  Macomb,  Illinois (1980).
                                    151

-------
              J=L
FIGURE 1 Mechanical components of sampler
          (a is syringe mount, b is connecting drive, c is pinion, d is machined cam)
                            152

-------
153

-------
                                         O
S < i*  m
                                         LJJ
                                    CO

                                    LU
                                    o;
                                    =D
                                    C.3
                                         O
                                         co
     154

-------
           ION CHROMATOGRAPHIC ANALYSIS OF ORGANIC ACIDS IN
                      DIESEL EXHAUST AND MINE AIR

                      I. Bodek and K. T. Menzies
                        Arthur D. Little, Inc.
                              ABSTRACT

Low molecular weight carboxylic acids are among the highly water soluble
compounds that are difficult to quantify by conventional organic analy-
tical procedures.  The relatively new technique of ion chromatography
has the potential for analyzing these acids in complex matrices.  Con-
centrations of two organic acids found in samples of diluted diesel
exhaust and coal mine air were determined using two ion chromatographic
procedures.

Samples were collected in midget bubblers containing dilute sodium car-
bonate solutions.  Analysis of acetate, formate and fluoride ions in
these solutions was performed using a 500 mm anion separator column and
sodium borate eluent.  Analysis of formate, acetate and aerylate in
these solutions was performed using the ion chromatography exclusion
mode (ICE) using the DIONEX IE-C-1 column with dilute hydrochloric acid
eluent.

Retention time for formate ion using the borate system is 11 minutes at
a flow rate of 2.3 mL/min.  Fluoride, chloride, sulfate and acetate do
not interfere in the analysis.  Linear response is obtained over the
concentration range of 0.1-4 yg/mL.  The detection limit is estimated
at 0.05 yg/mL for an injection volume of 100 yL at 3 ymhos full scale
sensitivity.

The ICE system was used to confirm the identity of the acetate and for-
mate detected in the samples and determine their concentration.  The
retention times of formic and acetic acids using the ICE system at an
eluent flow rate of 0.7 mL/min are 16 and 23 minutes, respectively.
No interference is observed from chloride, sulfate and acrylate.  Linear
response is obtained for formic acid and acetic acid in the concentration
range of 1-20 yg/mL and 3-100 yg/mL, respectively.  The detection limit
for a 100 uL sample injection volume at 30 ymhos full scale sensitivity
is estimated at 0.5 yg/mL.  Concentration of samples by freeze-drying
affords better detection limits with minimal loss of formic or acetic
acid.  Acrylic and carbonic acids are also analyzable under these con-
ditions .

                             INTRODUCTION

Low molecular weight carboxylic acids may be present in emissions from
combustion sources at part per million concentration levels (1).  Formic
acid and acetic acid both act as irritants to mucous membranes, eyes
and skin (2).   The emission goals established for the EPA-IERL environ-
mental assessment programs are 5 ppm for formic acid and 10 ppm for acetic
acid in gaseous effluents (3).  Formic acid may also be produced in the
environment by oxidation of formaldehyde emitted from combustion sources (4)
                                    155

-------
Previous attempts to measure concentrations of formic acid at ppm levels
have been thwarted by inadequate detection limits.   Generally,  analytical
methods for formic acid employ collection in aqueous solutions and reac-
tion with oxidizing or reducing agents.   Measurement of formic acid at
high concentrations can be made by adding an excess of oxidizing agent
and titrating the remaining excess of oxidant with reducing agents (5).
Another method (6) relies on the distillation of a chloroform-formic
acid azeotrope to separate formic acid from higher carboxylic acids
and analysis of the formic acid by potentiometric titration with sodium
hydroxide.  Detection limits for this method are generally in the 0.1%
range (6).  Gas chromatography has been used to measure formic acid
both directly and after derivitization.   Direct analysis poses problems
of corrosion of metal surfaces and of high detection limits with a flame
ionization detector.  Derivitization (7) eliminates these problems and
achieves a detection limit of about 25 pg/mL.

Ion chromatography has recently been successfully used for analysis of
formaldehyde after oxidation to formic acid  (8) and thus can be used
for direct analysis of formic acid.

In order to prevent interference by such inorganic anions as fluoride,
chloride, and nitrate which occurs with Na2C03/NaHC03 eluents, a weak
aqueous eluent, i.e., Na2Bi+07, was used to achieve adequate separation.
The detection limit in such ion chromatographic analysis is limited by
the conductance of the suppressed eluents, but the development of ion
chromatography exclusion  (ICE), which permits separation of weak acids
in a lower conductance background, has circumvented this problem.  The
determination of weak acids in complex media  (9) has been reported with
this technique.

This paper describes the analysis of formic acid and other carboxylic
acids in diesel engine exhaust and mine air using ion chromatography
(1C) and ion chromatography exclusion (ICE).

                        EXPERIMENTAL PROCEDURES

Collection

Organic acids in diluted  (1:10) diesel exhaust or mine air were collected
by drawing sample through two fritted bubblers  (10) in series, each con-
taining 15 mL of 10~3 M Na2C03.  A flow rate  of 1.0 liter per minute
and collection time of 60 minutes for diluted diesel exhaust or 240 min-
utes for mine air was used.  A 37 mm glass fiber filter  (Gelman Type A/E)
was placed before the bubblers to remove particulates.

Analysis

The solution in each bubbler was transferred  to a 25 mL volumetric flask
and diluted to volume with the collection medium (10~3 M Na2C03).  In
the case of samples collected from diluted diesel exhaust, an excess
(3 mL) of this solution was flushed through a 100 yL sample loop of
                                    156

-------
a Dionex Model 14 ion chromatograph.  The sample was analyzed on the
anion system and with instrumental conditions presented in Table 1.
                             TABLE 1

                  Ion Chromatographic Conditions



       Instrument               Dionex Model 14 Ion Chromatograph

       Eluent                   0.005 M ^Bi+Oy

       Flow Rate                2.3 mL/min

       Detector                 Conductivity

       Sensitivity              30 ymho/cm full scale

       Anion Columns            3 x 125 mm Dionex Anion Pre-column
                                3 x 500 mm Dionex Separator (Borate Form)
                                6 x 250 mm Dionex Suppressor (H+ Form)

       Sample Volume            100 yL Sample Loop

       Recorder                 HP 7133A

       Chart Speed              0.5 cm/min
In the case of samples collected from mine air, the bubbler solution
was concentrated by freeze drying.  Samples were transferred to wide
mouth jars, frozen to -25°C and freeze-dried under 0.5 cm Hg vacuum
in a Vacudyne, Inc. Pilot Freeze Dryer.  Once reduced to dryness,
1 mL of distilled/deionized water was added to the samples.  After
shaking to ensure complete dissolution, 300 ML aliquots were placed
in micro vials available for use in a Waters Associates Autoinjector
Model 710 A.  One hundred yL sample volumes were automatically injected
into the ion chromatograph and analyzed on the ion chromatography
exclusion  (ICE) system and with instrumental conditions presented in
Table 2.
                                    157

-------
                                TABLE 2

            Ion Chromatographic Exclusion (ICE) Conditions




       Instrument             Dionex Model 14 Ion Chromatograph

       Eluent                 0.0001 M HC1

       Flow Rate              0.7 mL/min

       Detector               Conductivity

       Sensitivity            3 ymho/cm full scale

       ICE Columns            9 x 250 mm Dionex Exclusion
                              3 x 500 mm Dionex Halide Suppressor
                                 (Ag+ Form)

       Sample Volume          100 yL, Waters Associates Autoinjector
                                 (Model 710 A)

       Recorder               HP 7133A

       Chart Speed            0.5 cm/min




                        RESULTS AND DISCUSSION

Analytical Method

Initial experiments were conducted on the Dionex Model 14 ion chromato-
graph to determine the feasibility and sensitivity of ion chromatography
for analysis of formic acid.  The conditions of 1C analysis (Table 1)
chosen were based on the need for separation of formate ion from the
fluoride, chloride and acetate anions and for obtaining a suitable detec-
tion limit.  Thus, the weak borate eluent (0.005 M ^26^07) combined
with a 500 mm anion separator column (and a 150 mm pre-column) were
chosen.  Conversion of the pre-column and separator column to their
borate forms (from the normal carbonate form) was necessary, and the
process of continuously passing the new eluent through the columns until
a stable baseline was obtained required several hours.

Standard stock solutions of formate and acetate were prepared from
reagent grade sodium salts.  Mixed stock solutions were also prepared.
Injections of 3 mL were made tc fill a sample loop of 100 yL volume.

A typical ion chromatogram of formic and acetic acid is shown in Figure 1.
The retention times for acetic and formic acids under these conditions
are 6.6 and 8.8 minutes, respectively.   The preceding negative peak

                                  158

-------
                                     Formic
                                      Acid
 1.5 jug/ml
Formic Acid
   and
Acetic Acid
     Inject
 I
10 Minutes
                             Anion
                             System
FIGURE 1   ION CHROMATOGRAM OF FORMIC AND ACETIC ACID
                               159

-------
grouping seen in the 2-5 minute region is due to water whose conductivity
is lower than boric acid.  The peak height is linear with formic acid
concentration over the range from 0.1 to 35 yg/mL.  The corresponding
linear range for acetic acid is about 0.2 to 70 yg/mL.  The presence of
fluoride or chloride in the sample does not interfere with the formic
acid analysis.  Retention time for these ions under the stated conditions
are F~ (5.4 minutes) and Cl~ (29 minutes).  Sulfate ion is retained for
a longer time.  Ultimately, these anions are eluted and may interfere
with subsequent samples.  Generally, these highly retained anions show
very broad peaks, which are easily distinguished from the organic acid
peaks.

Due to this interference, the need for frequent regeneration of the
suppressor column and the ability of ion chromatography exclusion to
easily separate organic acids and Cl~ and S0[t=, an ICE system was tested
for similar analyses.  Typical ICE conditions  (Table 2) provide easy
separation of strongly ionized species (e.g., H2SOit, HC1, HN03) from
weakly ionized species  (e.g., organic acids) due to the greater reten-
tion of uncharged species in the interstitial fluid of the packing
material (9).  The weakly acidic nature of the eluent influences the
ionization of each organic acid in proportion to its pK and thus its
retention time.  A 10"1* M hydrochloric acid eluent offers adequate
separation of the strong acids and the organic acids of interest with
a reasonable retention period.  Standard stock solutions were prepared
as before and injected onto the ion chromatograph.  In order to utilize
very small sample volumes, the normal sample loop, which requires excess
sample, was by-passed and a Waters Associates Autoinjector  (Model 710 A)
installed.  One hundred yL injections were made directly into the flowing
eluent by the autoinjector and analyzed on the ICE system.  A typical
ion chromatogram of formic acid and sulfuric acid is shown  in Figure 2.
The retention time for the sulfuric acid and formic acid are 9.8 and
15.4 minutes, respectively.  The peak height is linear with formic
acid concentration over the range of 0.3 to 10 yg/mL.  The  corresponding
linear range for acetic acid is 3 to 100 yg/mL (Figure 3).  As well as
providing separation of these organic acids and precluding  interference
from strong acids, the ICE system permits continuous analysis of samples
for up to 30 hours without regeneration of the suppressor column.

The precision of the ICE method was determined by analyzing six repli-
cates of two standard solutions containing strong acids  (i.e., t^SO^)
and both formic acid, acetic acid and carbonic acid.  Carbonic acid is
present as a result of the use of Na2C03 in the standard solution matrix
 (as in the collection medium) and dissolution  of  atmospheric carbon
dioxide (Figure 4).  At formic acid concentrations of 5.0 and 10 yg/mL,
the measured mean concentrations  (Table 3) were 5.08 and 10.0 yg/mL,
respectively.  The relative standard deviation (CV) was 0.025 and
0.016, respectively.

Collection Method

The goal of the initial phase of our work was  to determine  the concen-
tration of formic acid in engine exhaust subject  to different forms of
control, e.g., catalytic oxidation.  Initially, samples were collected
from diesel engine exhaust diluted by a factor of 10 in a stainless
                                   160

-------
                                   O
                                   CM
                               _   o
Q

O
<

O

5
DC
O
                                          co
                                        5 O
                                        < ei
                                        ^ <
o£
UJ ^
a o
                                        LU
                                        cc
          — 'o
          E <
          O) O


          si
            LL
161

-------
                                  ppv
                                             |UJ/6nf
o
o
           o
           O)
o
CD
O
ID
O
CO
o
CN
1   I
 o   y
LL   <

•   O
                                                                                            o
                                                                                            CO
                                                                                             o
                                                                                             CN
                                                                                             O
                                                                                             O
                                                                                                —  cc
                                                                                                LU  <
                                                                                                LU  H
                                                                                                O.  C/}
                                                                                                             £

                                                                                                             LU
                                                                                                             O
                                                                                                             Ul
                                                                                                             tc.
                                                                                             O
                                                                                             CO
                                                                                             o
                                                                                             CN
                                                  162

-------
                                                                                                                 O
                                                                                                                 CO
                                                                                                                       Q.
                                                                                                                       UJ
                                                                                                                       Z
                                                                                                                       LL
                                                                                                                       O

                                                                                                                       <
                                                                                                                       tL
                                                                                                                       (3
                                                                                                                       O
                                                                                                                       o
                                                                                                                       QC

                                                                                                                       u

                                                                                                                       *
                                                                                                                       111
                                                                                                                       DC
                                                                                                                       D
                                                                                                                       O
2
                                                                                    Q.
                                                                                    E
                                                     163

-------
                         TABLE 3

                  Analytical Precision



                       Formic Acid


    Calculated                                  Observed
Concentration (yg/mL)                      Concentration  (yig/mL)


       5.00                                       5.23
                                                  5.18
                                                  5.18
                                                  4.95
                                                  4.95
                                                  5.01

                                        Mean      5.08
                                        Std Dev   0.127
                                        CV        0.025
      10.00                                      10.2
                                                 10.2
                                                 10.1
                                                  9.9
                                                 10.1
                                                  9.8

                                        Mean     10.0
                                        Std Dev   0.163
                                        CV        0.016
                               164

-------
steel dilution tunnel.  A collection medium of 15 mL of 10   M sodium
carbonate solution was utilized for two reasons.  First, such an aqueous
solution (7) is reported to provide good collection efficiency of soluble
organic acids at flow rates of 0.5 to 5 L/min.  Second, due to the vola-
tility of free formic acid, it was felt that a basic solution would pro-
vide improved stability of the samples over periods up to seven days.
The collection efficiency of this medium was determined by measuring the
amount of formic acid collected at 1.0 L/min in two bubblers connected
in series.   The amoung of formic acid found in the front and back bubbler
at three challenge concentrations are reported in Table 4.  The data
show that the collection efficiency is greater than 92%.


                              TABLE 4

                       Collection Efficiency
          Challenge            ug  Formic  Acid  Found       % Collection
    Concentration  (mg/m3)      Front          Back          in Front

            12                 153            <2.5            >98.4

             6.7                90.0          <1.0            >98.9

             2.5                32.5          <2.5            >92.3

             0.06               10.3           0.8             92.2
In order to provide at least 99% collectiQn of formic acid, two bubblers
were routinely used in series.  In a later phase of work, the formic
acid concentration in mine air subject to diesel emissions was measured.
The expected concentrations were about one hundred times lower than
those found in engine exhaust.  The efficiency of the collection scheme
was again measured under these conditions of challenge concentration
(0.06 mg/m3).   The collection efficiency was found to be 92.2% at this
level (Table 4).

Sample Analysis

For sample collection in both diluted diesel exhaust and a mine atmosphere,
the collection technique described previously was used, but the sampling
periods were 60 minutes and 240 minutes, respectively.  In the mine samples,
the amount of  formic acid collected was too small to be analyzed reliably
after dilution to 25 mL.  This problem could be precluded by collection
at a higher flow rate (e.g., 5 L/min) or longer time periods.  However,
in order to utilize the samples as collected, they were concentrated by
freeze-drying.   The pH of samples prior to freeze-drying was checked to


                                    165

-------
ensure that solutions were slightly basic.  Strongly acidic species,
e.g., t^SOij, collected during mine air sampling, slowly deplete the
Na2CC>3 in the collection media and weakly ionized organic acids may
be volatiled at low pH.  Recovery of formic acid during the freeze-
drying process was checked by carrying standard solutions through
the entire analytical procedure.  As is shown in Table 5, recovery
of triplicate samples of two concentrations averaged about 0.88.
                              TABLE 5

                            Formic Acid
                    Recovery After Freeze-Drying
    Initial Sample          Freeze-Dried Sample
    Concentration              Concentration
    	(pg/mL)             	(yg/mL)	         Recovery

          4.3                       3.9                   0.91

                                    3.7                   0.86

                                    3.9                   0.91


          9.7                       9.0                   0.93

                                    8.3                   0.86

                                    7.8                   0.80
Storage stability of these samples was checked by replicate analysis
after a period of seven days.  Losses of less than 2% were observed
after storage for this period of time.

The accuracy of the ICE method was assessed by using the standard
addition technique.  Four mine samples containing measured concentra-
tions of formic acid of about 0.8 to 1.1 yg/mL were spiked with a known
volume of formic acid standard solution sufficient to double the
sample concentration.  The observed concentration indicated agreement
of better than ±12% with an average agreement within ±5% (Table 6).
                                    166

-------
                             TABLE 6

                 Standard Addition of Formic Acid
          Initial                    With Spike
   Concentration (pg/mL)      Calculated     Observed       Agreement

            0.8                   1.6           1.8           1.12

            1.0                   2.0           2.1           1.05

            1.1                   2.2           2.3           1.04

            1.1                   2.2           2.2           1.00
Standard addition was also used to confirm the identification of the
chromatographic peaks based on comparison of retention time.   No inter-
ference was observed due to chloride,  sulfate or carbonate ion.  ^Both
formate and acetate were adequately resolved.  A small unidentified
peak did precede formic acid but did not significantly affect quanti-
tation.

                              CONCLUSIONS

Two ion chromatographic techniques were developed to quantify formic
and acetic acid in both diesel engine exhaust and mine air subjected
to diesel emissions.  A commonly reported anion separation system
utilizing a weak borate eluent adequately separated the acids of
interest in diesel exhaust.  It was, however, affected by the presence
of strong acids during subsequent consecutive analyses.

In order to preclude this problem and the necessary frequent regenera-
tion of the anion system's suppressor column, an ion chromatography
exclusion scheme was utilized.  Samples collected in a mine environ-
ment were reliably concentrated by freeze-drying and then analyzed on
an ICE system with dilute hydrochloric acid  eluent.  The precision
of the ICE method was experimentally determined to be ±2.5%.  The
accuracy was not independently determined but good precision and
recovery yield confidence that measured values are within ±5% of the
true value.  No interferences were observed  in the ICE system due  to
strong acids, carbonic acid or other water soluble species present
in mine air subject  to diesel emissions.
                                   167

-------
                              REFERENCES

1.   National Academy of Sciences, "Vapor-Phase Organic Pollutants,"
    (ISBN 0-309-02441-2), Washington, DC (1976).

2.   Patty, F. A.,  ed., Industrial Hygiene and Toxicology, Second Revised
    Edition, Volume II, Toxicology, Interscience Publishers (1963).

3.   Cleland, J. G. and G. L. Kingsbury, "Multimedia Environmental Goals
    for Environmental Assessment," Volume II, EPA-600/7-77-136b,
    November, 1977.

4.   Menzies, K. T., "Fate of Reactive Diesel Exhaust Contaminants,"
    Draft Final Report, U.S. Bureau of Mines Contract J0188061,
    June 1980.

5.   Peters, D. G., J. M. Hayes and G. M. Hieftze, Chemical Separations
    and Measurements, Saunders Co., 1974.

6.   Warner, B. R.  and L. Z. Raptis, "Determination of Formic Acid in
    the Presence of Acetic Acid," Anal. Chem., 27, 1798  (1955).

7.   Smallwood, A.  W., "Analysis of Formic Acid in Air Samples," Amer.
    Ind. Hyg. Assoc. J., _39, 151-153  (1978).

8.   Kim, W. S., C. L. Geraci and R. E. Kupel, "Solid Sorbent Tube
    Sampling and Ion Chromatographic Analysis of Formaldehyde," Amer.
    Ind. Hyg. Assoc. J., 41, 334-339  (1980).

9.   Rich, W., F. C. Smith, L. McNeil and T. Sidebottom,  "Determination
    of Strong and Weak Acids and Their Salts by Ion Chromatography
    Coupled with Ion Exclusion," unpublished, available  from Dionex, Inc.
                                     168

-------
                    A NEW ELECTROCHEMICAL APPROACH TO
                     TRACE LEVEL ALDEHYDE AND KETONE
                                ANALYSIS

                 R.P. Baldwin, J.F. Price, and J. Siria
            Department of Chemistry, University of Louisville
                                ABSTRACT

A new electroanalytical approach using chemically modified electrodes
for the determination of specific classes of organic compounds is
described, and tentatively evaluated.  The suggested method consists of
two principal steps:  the pre-concentration of the analyte at the electrode
surface via a selective chemical reaction with the modifying molecule
and the quantitation of the subsequent surface-bound reaction product
via differential pulse voltammetry.  By judicious choice of the chemically
modified electrode and the pre-concentration reaction employed, the
approach in principle may be applicable for the selective monitoring of
several important families of organic compounds.  In this work, however,
a model analytical system applicable specifically for the determination
of dissolved aldehydes and ketones is considered.  The chemically modified
electrodes employed for the analysis were formed by the adsorption of
allylamine onto the surface of a platinum wire.  The pre-concentration
reaction involved the formation of an imine by condensing the carbonyl
analyte with the attached amine.  For the model analyte ferrocene carbox-
aldehyde, a detection limit of ICT^M was observed using a pre-concentration
time of only five minutes.  No major interference was observed when the
analysis was carried out in the presence of a hundredfold excess of ferrocene.

                              INTRODUCTION

Environmental analysis systems are often called upon to perform quantitative
measurements on chemical species present in only minute amounts in complex
and sometimes highly interfering matrices.  In many cases, the analyst's
task could be compared to the job of locating the needle in the proverbial
haystack.  Furthermore, as the identity of the target analyte is sometimes
not known until the results of the analysis are completed, the analyst is
not even sure beforehand that a needle is what he should be looking for.
As a result, the use of any analytical technique for the purpose of making
environmental measurements, generally imposes severe demands on the performance
level required of the technique.

Among the most important criteria that must be considered in evaluating any
environmental analysis approach are both its sensitivity and selectivity.
Thus, because of high degree of specificity inherent in their separation
step, chromatography-based methods have ranked among the most widely used
for trace organic pollutant monitoring.  Similarly, although most electro-
analytical methods possess adequate sensitivity for such use, these methods
are generally not specific enough to be used alone and have had their
primary application as detectors at the end of gas or liquid chromatographs.
                                    169

-------
In the instances in which they are applicable,  however,  electroanalytical
methods do present several significant benefits.   By and large,  the
instrumentation involved in electrochemical techniques is comparatively
inexpensive and simple to operate; and, even though they are not generally
capable of ultra-trace level determinations, they routinely operate in
the ppm range for electroactive species.  In addition, the technique of
anodic stripping voltammetry (ASV) offers a specific opportunity for orders
of magnitude sensitivity enhancement by including a pre-concentration
step in which the intended analyte is electrodeposited at the electrode
surface prior to analysis.  Unfortunately, ASV is mainly useful only for
the analysis of metal ions since, unlike metals,  most organics cannot be
plated out and effectively pre-concentrated.

In this report, we will describe a new pre-concentration approach to the
electrochemical analysis of organics.  Our objectives are to add a dimension
°f chemical selectivity to the electrochemical approach while, at the
same time, maintaining or increasing its usual sensitivity.  The suggested
method proposes the use of recently developed chemically modified electrodes
(CME's) first to pre-concentrate a specific organic analyte at the electrode
surface via a selective chemical reaction and then to quantitate the
subsequent surface-bound reaction product via conventional voltammetric
techniques.  CME's are essentially ordinary electrodes whose surfaces have
been altered chemically by the attachment of molecules containing particular
functional groups.  As a result, CME's possess a dual set of properties-
electrochemical ones which are determined by the externally applied potential
and chemical ones which are determined by the nature and reactivity of the
attached molecule.  In this work, the variable chemical properties of CME's
will be utilized to allow the selective pre-concentration of organic analytes
at the CME surface.  In particular, the chemical reactivity of the surface-
modifying functional groups toward specific classes of organics will be
examined with respect to the sensitivity enhancement and the gross chemical
selectivity that can be imparted by means of the pre-concentration step.

The experimental system which is considered below is applicable in principle
for the determination of carbonyl-containing compounds.  We will specifically
employ ferrocene carboxaldehyde as a model analyte.  This choice was made
because ferrocene carboxaldehyde can react chemically (i.e., undergo
pre-concentration) as an aldehyde and at the same time can easily be
detected at the electrode surface by means of its distinctive and readily
accessible oxidation to the ferricinium species.

                               BACKGROUND

In recent years several groups have been actively engaged in research
directly  involving the design and characterization of CME's  (1-19,21-27)
Both normal conducting  (Pt, C, etc.) and semiconductor  (Sn02, Ti02, and
Ru02,  etc.) electrode surfaces have been altered by  the  attachment of
numerous  organic molecules yielding  electrodes possessing a variety of
interesting chemical properties.
                                     170

-------
In most cases, one of two general attachment approaches has been employed:
(1) the covalent bonding of the modifying molecule to functional groups
present on the native electrode surface or (2) the irreversible chemisorption
of the modifier onto the electrode material.  The former method, employing amide,
ether, and silane linkages, has been shown to be quite general in scope,
yielding surfaces that are permanently and reproducibly altered.  The
drawback of the covalent approach is that the chemical reaction employed
in the binding step is usually non-trivial in nature, requiring special or
severe reaction conditions and/or long reaction times (several hours to
several days).  The second approach, utilizing the direct adsorption of a
variety of species (e.g., n-alkenes, polymers, etc.), also appears to be
very general both in terms of the types of electrode meterials (Ft, C, etc.)
that can be modified and the types of functionalities that can be attached.
The adsorption itself is usually relatively permanent and can normally be
accomplished by simply immersing the "clean" electrode in a dilute solution
of the surfactant for a few minutes.

A listing could be given that encompasses each molecule and functional
group that has been thusfar either bound or adsorbed onto an electrode
surface, but it would be simpler for our purposes to summarize the compilation
by making one very general observation from it.  The reactivity-both chemical
and electrochemical-of species attached to electrode surfaces remains
roughly comparable to that of the analogous species present in bulk solution.
Chemical reactions characteristic of the attached molecule have been
demonstrated to occur at CME surfaces under relatively severe reaction
conditions and without noticeably affecting the attachment itself.
Examples include metal complexations, ligand substitutions, protonations,
amine-aldehyde condensation, amidizations, etc.  Similarly except for a
few sterically prohibited situations (15), species that are electroactive
in dissolved form have been shown to retain approximately identical redox
behavior when present in surface-bound form (19).  Thus, the response of
CME systems possesses two different components, a chemical component
determined by the nature and reactivity of the specific attached molecule
and an electrochemical component determined by the specific potential at
which the electrode is maintained instrumentally.

The main thrust of the work to date has been directed toward the development
of generalized modification strategies and the verification of these
modification procedures either by surface analysis of the resulting electrodes
or by electrochemical characterization of the redox behavior of the attached
species.  Furthermore, in the application areas that have thusfar been
reported for CME's, the primary goals that have been considered have been
the formation of an electrode capable of participating in selective (9,10),
catalytic  (7,13) , or photo-assisted (22,27) Faradaic processes.  But in
only a few cases have the special properties of CME systems been exploited
for the purpose of obtaining an electrode system of enhanced analytical
ability.  Lane and Hubbard  (17, 18) have employed a platinum electrode
coated with an adsorbed layer of iodide ion to prevent the electrode fouling
and high background currents that normally occur in the in vivo analysis
of dopamine and norephinephrine.  And, in work more directly related to
our own, Cheek and Nelson  (4) have utilized a carbon paste electrode
containing attached EDTA functionalities to pre-concentrate via chelation
and subsequently to analyze via cyclic voltammetry silver ion present at
concentrations as low as 10"-*-^.
                                      171

-------
                            GENERAL PLAN OF WORK

The general requirements for an acceptable CME/pre-concentration system
are the following:

     1.  The analyte involved should be one that is of direct
         analytical importance.

     2.  The CME involved should be one that can be formed rapidly
         and reproducibly.

     3.  The pre-concentration reaction should be rapid and straight-
         forward, uncomplicated by side products and as specific as
         possible for the analyte of interest.

     4.  The reaction product should remain bound to the electrode
         surface and should possess a characteristic electrochemical
         activity.

Finding organics which are of analytical importance is of course a simple
matter.  Similarly, because of the wide variety of CME's and modification
methods that have been developed, procedures for the attachment of many
common functional groups are already available; and, when the attachment
is made via one of the adsorption approaches, the procedure required is
elementary to perform.  The most difficult aspect of the proposed approach
appears to be the proper selection of the pre-concentration reaction.
Among the sources that could be consulted in this regard is the literature
of qualitative organic analysis which consists, in large part, of a
compilation of straightforward and relatively specific chemical reactions
suitable for functional group and family analysis.  Further, many of the
derivitization reactions developed for use in chromatography, mass spec-
troscopy, and optical analysis are also appropriate candidates for the pre-
concentration step.  Many of these reactions involve products that are
expected to be electroactive and therefore should be suitable for the sub-
sequent electroanalysis step.

A chemical system which meets several of these criteria and was consequently
chosen as a model system for the work described here was one which is appli-
cable for the analysis of dissolved aldehydes and ketones (25).  The pre-
concentration step utilizes the familiar acid-catalyzed condensation reaction
between carbonyl-containing compounds and primary amines to give imine-
containing products;

               9,                     H+      RV
             R-C-R'   +   R"NH0   <     >     JC=NR"   + HnO
                              irt   ^	     ^


This reaction has been employed in the trace electroanalytical determination
of dissolved amines  (20,26) and has previously been shown to take place at
amine-modified CME's  (26). Since the imine bond formed as a result of this
reaction is reducible at a readily accessible negative potential, it would
                                      172

-------
in principle be feasible to quantitate the reaction product directly for
any reactive aldehyde or ketone.  For this study, however, ferrocene
carboxaldehyde
was utilized as a model aldehyde analyte; and the characteristic oxidation
of the ferrocene to ferricinium ion was employed for product detection and
estimation.  The technique of differential pulse voltammetry was used for
electrode surface characterization and analysis purposes.  Our objectives
were to evaluate, first, the selectivity with which organic molecules con-
taining a specific functional group can be concentrated at an appropriate
CME surface and, second, the sensitivity with which such procedures can
be used for quantitative analysis of the surface-bound species.

                                PROCEDURE

In this study, CME's were constructed from platinum wire, one end of
which was sealed into glass tubing.  Rigorous cleaning and pre-treatment
of the platinum surface was essential for obtaining reproducible results.
Prior to each experiment, the electrodes were heated to incandescence in a
methane-oxygen flame and quenched in perchloric acid.  They were then
placed in an aqueous 0.1M LiClO^ solution, and the potential was scanned
slowly between +1.5 volts and -0.6 volts for several cycles and then held
at +0.40 volts for at least three minutes.  Figure 1A shows the typical
current-voltage behavior observed following this pre-treatment regimen.
This and all subsequent analyses of the coated and reacted electrodes
were performed in 0.1M LiC104/acetonitrile using differential pulse
voltammetry (potential scan rate: 2 mv/sec; initial potential: 0.0 volts;
pulse amplitude: 50 mv).

Electrode modification was performed by adsorbing allylamine onto the clean
platinum surface.  This was accomplished by immersing the clean electrodes
in a 0.5% allylamine solution for three minutes.  A differential pulse
voltammogram of the freshly modified CME is shown in Figure IB and is seen
to contain little dramatically different behavior from the clean electrode
surface.  Ferrocene carboxaldehyde pre-concentration was accomplished by
immersing the allylamine-coated electrode in an acidified ethanolic analyte
solution that was heated to 75°C.  The differential pulse voltammogram of
such a reacted electrode is illustrated in Figure 1C and shows the
presence of a reversible redox couple at approximately +0.32 volts.  The
observed peak potential closely matches that found for the imine condensa-
tion product formed when allylamine and ferrocene carboxaldehyde react in
bulk solution and the product is monitored at a clean platinum electrode.
The current observed for the reacted electrode decreased in amplitude
slightly when the differential pulse scan was repeated; hut, after the
first few repetitions, stable and reproducible currents were obtained.

                                     173

-------
                           0.32
Q.
E
TO
                                  0.4
                               0.6
                                                 (V v*. SCE)
0.8
+1.0
       FIGURE 1
Differential pulse voltammograms of  a clean platinum  electrode
(A),  an electrode coated with allylamine (B),  and an  electrode
coated with allylamine and reacted with 4.5 x  ICT-^M ferrocene
carboxaldehyde (C).   Added uncompensated resistance:   1100  ohms.
                                       174

-------
                                RESULTS

By analogy to the conventional ASV technique, it is expected that the
amount of ferrocene carboxaldehyde-allylamine condensation product
formed at the CME surface as described above should be dependent on two
experimental parameters: the concentration of ferrocene carboxaldehyde
in the reacting solution and the reaction time allowed.  Each of these
effects was evaluated separately.

When identical allylamine-coated platinum electrodes were immersed for
five minutes in reaction solutions containing a range of ferrocene car-
boxaldehyde concentrations, the series of differential pulse voltammo-
grams shown in Figure_2 were obtained.  For ferrocene carboxaldehyde
concentrations of 10" M and above, the voltammograms were virtually
identical, indicating similar and probably saturated surface coverage
for each case.  At concentrations less than 10~ M, the peak current
decreased gradually with concentration - with a detectable signal
still observed at 10"~'M,

The level of response obtained for low analyte concentrations may be
compared with data obtained using conventional electroanalytical tech-
niques.  Since ordinary differential pulse voltammetry is generally
recognized as one of the most sensitive polarographic methods for bulk
solution analysis, this technique was selected for the comparison.
Accordingly, data compiled using the conventional differential pulse
method (with the pulse height and scan rate chosen to match those used
in the CME pre-concentration and stripping approach) are compared in
Table 1 with data obtained using CME's.  At high ferrocene carboxalde-
hyde concentrations, the conventional differential pulse signal is
substantially larger than the CME response whose amplitude is limited
at all concentrations by the finite quantity of analyte which it is
possible to attach to the electrode surface.  The differential pulse
signal, however, depends on the diffusion of the electroactive species
to the electrode and thus decreases rapidly with concentration.  Hence,
at low concentrations, the pre-concentration and stripping method can
be expected to provide competitive performance.  Its observed detection
limit of 10~'M compares very favorably with that seen for the conven-
tional technique.

The voltammograms in Figure 2 represent behavior observed when allyl-
amine-coated electrodes were allowed to react for a fixed time period
in solutions of various ferrocene carboxaldehyde concentration.  When
the coated electrodes were immersed in identical 5 x 10~^M analyte
solutions and allowed to react for increasing time intervals, subse-
quent differential pulse analysis indicated that the +0.32 volt peak
was formed very rapidly upon immersion.  However, prolonged reaction
times beyond about five minutes did not lead to increased ferrocene
carboxaldehyde concentration at the electrode surface or to the in-
creased peak currents that were expected.  Rather, the peak height de-
creased gradually with time; and the peaks became distinctly broader
in shape.  Whether this observation is due to the further reaction of
                                    175

-------
                                                         20  juamps

                                                         T

-------
                                 TABLE 1
          Comparison of CMC Response with Conventional Differential
          Pulse Voltammetry
[Ferrocene Carboxaldehyde],  M       ip, CME, yampsa         ip, Bulk, yamps
            10"2                          120

            IO"3                          110                      510

            IO"4                           41                     60.5

            10~5                           10                      6.5

            10"6                            6                      0.73

            IO"7                            2                        0
a.  Reaction time - 5 min.

    Reaction temperature - 75°C

    lip = +0.52 volts vs. SCI;

    Added uncompensated resistance: 1100 ohms

b.  12p = +0.61 volts vs. SCE
                                     177

-------
the ferrocene or imine moieties or to the gradual alteration of the CME
surface under the reaction conditions employed has not yet been deter-
mined.

Finally, the selectivity of the concentration procedure was examined by
allowing the imine formation to take place in the presence of both
equimolar and hundredfold excess amounts of ferrocene itself.  The
voltammograms obtained under these circumstances were similar to those
seen for ferrocene carboxaldehyde when no ferrocene was present, and
the results are summarized in Table 2 .   It is interesting to note that
no drastic effect due to ferrocene was observed except for a slight sup-
pression of the imine peak current at very high interferent concentra-
tions.  Apparently, the concentration reaction is selective for the
carbonyl functionality, permitting, the selective reaction and quantita-
tion of samples containing this group in the presence of highly concen-
trated and otherwise identical species.

                               CONCLUSIONS

What we have described and tentatively evaluated here is a completely
new approach for the trace level analysis of selected families of or-
ganic compounds.  Its principal advantages appear to be three in number;

     1.  Sensitivity.  The technique described is analogous in many
         respects to that of anodic stripping voltammetry, one of
         the most sensitive analytical methods. ASV is able to
         achieve detection limits in some cases below the ppb level
         primarily because of a pre-concentration step similar to
         that which we propose to employ.  For the model analyte
         considered here, a detectable signal was observed at the
         10~7M or 25ppb level.

     2.  Selectivity.  Conventional electrochemical methods possess
         sufficient selectivity of response to make them useful by
         themselves in some environmental applications.  In addition
         to this selectivity based primarily on differences in ana-
         lyte redox potential, our approach also offers a gross
         chemical selectivity for specific classes of compounds that
         derives from the nature of the pre-concentration reaction
         employed.

     3.  Convenience.  Some existing analytical methods possess
         similar or greater selectivity and sensitivity than the
         proposed approach.  But, of those which are useful for
         organic analysis, few offer these advantages at the same
         level of experimental convenience.  Because of the selec-
         tive pre-concentration step, the extraction or chromato-
         graphic step usually required prior to trace organic
         analyses might often be unnecessary.  Additionally, the
         sensitivity obtained here utilized a pre-concentration
         time of only five minutes; and, as in most electrochemical
         methods, no sophisticated or expensive instrumentation
         was required.

                                     178

-------
                                  TABLE 2
          Effect of Ferrocene on CME Response to Ferrocene Carboxaldchyde
[Ferrocene Carboxaldehyde],  M       [Ferrocene], M        ip, yamps




          10"3                            0                  104




          lO'^                           IQ~*                 96




          10~3                           10'5                102
          10~5                            0                   11




          iO"5                           10~3                  8
                                         io-3
Reaction time - 5 minutes




Reaction temperature - 75°C




Hp = +0.32 volts




 Added  uncompensated resistance:  1100 ohms
                                    179

-------
                            ACKNOWLEDGEMENT

This work was supported by the National Science Foundation RIAS grant
#77-06911, by the University of Louisville Graduate School, and by the
University of Louisville Chapter of the Sigma Xi Scientific Research
Society.

                               REFERENCES

1.  Armstrong, N. R., A. W. C. Lin, M. Fujihira, and T. Kuwana,
     "Electrochemical and  Surface  Characteristics of Tin  Oxide  and
     Indium  Oxide Electrodes," Anal.  Chem.,  48,  741-50  (1976).

2.  Brown, A.P.  and F.C. Anson, "Cyclic and  Differential  Pulse
     Voltammetric Behavior of Reactants Confined to the Electrode
     Surface," Anal. Ghem. , 4,9., 1589-95 (1977).

3.  Brown, A.P., C. Koval  and F.C. Anson,  "Illustrative Electrochemical
     Behavior of Reactants Irreversibly Adsorbed or Graphite  Electrode
     Surfaces," J. Electroanal. Chem.,  72,  379-87 (1976).

4.  Cheek, G.T.  and R.F. Nelson, "Applications  of Chemically  Modified
     Electrodes  to Analysis of Metal  Ions,"  Anal. Lett.,  All, 393-402 (1978).

5.  Evans, J.F.  and T.  Kuwana, "Radiofrequency  Oxygen Plasma  Treatment
     of Pyrolytic Graphite Electrode  Surfaces," Anal.  Chem.,  49,
     1632-35 (1977).

6.  Evans, J.F.  and T.  Kuwana, "Introduction of Functional Groups  onto
     Carbon  Electrodes  via Treatment  with  Radio-Frequency Plasmas,"
     Anal. Chem., 51_, 358-65  (1979).

7.  Evans, J.F., T. Kuwana, M.T. Henne and G.R.  Royer,  "Electrocatalysis
     of Solution Species Using Modified Electrodes," J. Electroanal.
     Chem.,  80_,  409-16  (1977).

8.  Firth, B.E.  and L.L. Miller, "Oxidations on DSA and Chirally
     Modified  DSA and  Sn02 Electrodes," J. Am.  Chem.  Soc., 98.,  8272-73
      (1976).

 9.  Firth, B.E., L.L. Miller, M. Mitani,  T.  Rogers, J.  Lennox and
     R.W.  Murray,  "Anodic  and Cathodic Reactions on a  Chemically Modified
     Edge Surface  of  Graphite,"  J. Am. Chem. Soc.,  98_,  8271-72  (1976).

10.  Haller,  I.,  "Covalently Attached  Organic Monolayers on Semiconductor
      Surfaces."  J.  Am.  Chem.  Soc.. 100, 8050-55 (1978).

11.   Itaya,  K.  and  A.J.  Bard,  "Chemically  Modified  Polymer Electrodes:
      Synthetic Approach Using Poly(methacryl chloride)  Anchors,"
      Anal.  Chem.,  5iD,  1487-89 (1978).

                                      180

-------
12.   Kaufman, F.B. and E.M. Engler, "Solid-State Spectroelectrochemistry
      of Cross-Linked Donor Bound Polymer Films," J. Am. Chem. Soc.. 101,
      547-49 (1979).

13.   Kerr, J.B. and L.L. Miller, "Electrocatalysis of Organic Dihalide
      Reductions Using a Polymer Modified Electrode," J. Electroanal.
      Chem., 101, 263-67 (1979).

14.   Kuo, K., P.R. Moses, J.R. Lenhard, D.C. Green and R.W. Murray,
      "Immobilization, Electrochemistry, and Surface Interactions of
      Tetrathiafulvalene," Anal. Chem., 51, 745-48  (1979).

15.   Lane, R.F. and A.T. Hubbard, "Electrochemistry of Chemisorbed
     'Molecules. 1. Reactants Connected Through Olefinic Substituents,"
      J. Phys. Chem.. _77, 1401-10 (1973).

16.   Lane, R.F. and A.T. Hubbard, "Differential Double Pulse Voltammetry
      at Chemically Modified Electrodes for in vivo Determination of
      Catecholamines," Anal. Chem., 48, 1287-93 (1976).

17.   Lane, R.F. and A.T. Hubbard, "Electrochemistry of Chemisorbed
      Molecules. 5. Role of Nonaqueous Solvents," J. Phys. Chem., 81,
      734-39 (1977).

18.   Lane, R.F., A.T. Hubbard, K. Fukunaga and R.J. Blanchard, "In Vivo
      Determination of Dopamine and Homovanillic Acid," Brain Res., 114,
      346-56 (1976).

19.   Lenhard, J.R., R. Rocklin, H. Abruno, K. Willman, K. Kuo, R. Nawak
      and R.W. Murray, "Chemically Modified Electrodes. 11.  Predictability
      of Formal Potentials of Covalently Immobilized Charge-Transfer
      Reagents," J. Am. Chem. Soc., 100, 5213-15 (1978).

20.   McLean, J.D., V.A. Stenger, R.E. Relm, M.W. Long and T.A. Huller,
      "Determination of Ammonia and Other Nitrogen  Compounds by
      Polarography," Anal. Chem., _50, 1309-14 (1978).

21.   Merz, A. and A.J. Bard, "A Stable Surface Modified Platinum Electrode
      Prepared by Coating with Electroactive Polymer," J. Anr. Chem.
      Soc., 100, 3222-23 (1978).

22.   Miyasaka, T., T. -Watanabe, A. Fujishima and K. Honda, "Light
      Energy Conversion with Chlorophyll Monolayer  Electrodes,"
      J. Am. Chem. Soc., 100, 6657-65 (1978).

23.   Moses, P.R., L.  Wier and R.W. Murray, "Chemically Modified Tin
      Oxide Electrode," Anal. Chem., _47, 1882-86 (1975).

24.   Oyama, N., K.B.  Yap and F.C. Anson, "Spontaneous Coating of
      Graphite Electrodes by Amino Ferrocenes," J.  Electroanal. Chem.,
      100. 233-46  (1979).

25.   Price, J.F. and R.P. Baldwin, "Organic Pre-Concentration and Stripping
      Analysis at Chemically Modified Platinum Electrodes," submitted
      to Anal. Chem.

-------
26.   Sharp, M.,  M.  Petersson and K.  Edstrom,  "Preliminary Determinations
      of Electron Transfer Kinetics  Involving Ferrocene Covalently
      Attached to a Platinum Surface," J.  Electroanal.  Chem.,  95,
      123-30 (1979).

27.   Wrighton,  M.S., R.G.  Austin, A.B. Bocarsly, J.M.  Bolts,  0.  Haas,
      K.D. Legg, L. Nadjo and M.C. Palazotto, "Design and Study of a
      Photosensitive Interface:  A Derivatized n-type Silicon Photo-
      electrode," J. Am. Chem. Soc., 100,  1602-03 (1978).
                                      182

-------
             A COMPUTER INTERFACED TOXICITY TESTING SYSTEM
               FOR SIMULATING VARIABLE EFFLUENT LOADING

                   J. Cairns, Jr. and K.  W. Thompson
  University Center for Environmental Studies and Biology Department
          Virginia Polytechnic Institute and State University
                 Blacksburg, Virginia   24061  U.S.A.
                               ABSTRACT

Water quality criteria and standards are based primarily on toxicity
tests carried out with single chemicals whose concentration is as nearly
constant as possible.  In the "real world," however, organisms are
exposed to mixtures of chemicals which usually have markedly fluctuating
concentrations.  The primary difficulties in simulating "real world"
conditions in toxicity testing are:  (a)  a means of varying chemical
concentration to fit a predetermined set of conditions, and (b)  a
system which is capable of tracking and recording the response of aquatic
organisms to these variations and which is quantitative and suitable for
cross correlations of dose and response.  Mini- and microcomputer inter-
facing with a toxicity testing system provides a means of systematically
varying the concentration of a test chemical or chemicals in a continuous
flow system.  The same computer can also be used for the data acquisition
system to store the voluminous time-series biological response data
necessary for cross correlations with variable chemical concentrations.
A description of the apparatus, examples of its use, types of data
generated, and data analysis are discussed.

                               INTRODUCTION

Nearly all toxicity tests upon which water quality criteria are based
have been carried out in test containers in which the intent was to keep
the concentration of the toxicant constant.  The development of continuous
flow apparatus was in large part a consequence of the failure of batch
test containers to maintain constant concentrations of certain kinds of
toxic materials.  Despite all the difficulties involved in maintaining
constant concentrations and the rarity of tests in which constancy has
actually been accomplished, one should bear in mind that lack of
variability was the desired goal.  However, when one attempts to
transfer laboratory data to the "real world," one finds that constancy
is the exception, not the rule, and that fluctuations in waste quality
and quantity may occur from day to day, hour to hour, or even minute
to minute.  In spite of this lack of correlation between laboratory
toxicity data and "real world" situations, toxicity tests in which the
toxicant levels fluctuate according to a predetermined scheme have
received little attention.  This is probably due to the problems in-
volved with automation of the toxicant controlling machinery.  Conse-
quently, applying results from relatively constant tests carried out
in the laboratory to fluctuating "real world" situations presents
various difficulties.  Organisms are often apt to respond to thresholds
rather than averages.  On the other hand, duration of exposure is an

                                   183

-------
important factor in determining a response.  Some thresholds, therefore,
may not elicit a response if the duration of exposure is too short.
Water criteria and standards (for example, National Academy of Sciences
(12)) have been developed for years without substantial attention to
the fluctuating exposure.  Only circumstantial evidence of the effec-
tiveness of such standards exists, and no substantial body of evidence on
the relationship between fluctuating and constant concentration ex-
posures exists for aquatic organisms.   Standards developed without this
information might accidentally be correct, but might just as easily be
either over- or underprotective.  If either of the latter situations
exists, the environment may be damaged or excessive expenditures may be
made for unnecessary waste treatment systems.

One basic question is: in a qualitatively fluctuating effluent discharge,
when do the organisms respond to the peak concentrations as if they were
constant at the peak level?  Additional interesting questions are:

     1.  What amplitude and frequency of episodic exposures,
         individually and of short duration, appear to have no
         deleterious effects?

     2.  When there are sublethal responses to fluctuating con-
         centrations, are the organisms capable of returning to
         normal and, if so, how much time does it take?

     3.  At what concentration is amplitude a more important
         consideration than frequency of fluctuation, and vice-
         versa?

     4.  If a large retention pond with thorough mixing and
         delivery of a constant concentration and quality of
         waste is possible and conditions permit variable re-
         leases, is a pulsed delivery of the same total amount
         on a daily basis more or less harmful to the exposed
         organisms than a constant delivery of waste?

     5.  Is there some way in which regulatory standards could
         be modified to permit occasional nontoxic pulses of
         short duration when large holding ponds are not available
         and batch processing makes pulses quite likely?

This short paper could not possibly address all these questions; however,
the  instrumentation making such tests feasible, one of  the main obstacles
to answering these questions and other related questions, can be dis-
cussed.  Two major capabilities must be incorporated into the instru-
mentation used to solve problems:   (a)  a means of predetermining  the
amplitude and frequency of the changes in  toxicant concentration, and
(b)  a means of recording on a regular or  continuous basis the biolog-
ical response to these changes so that correlations between dose and
response may be achieved.  A statistical analysis that will enable a
quantitative evaluation of the data generated must also be developed.

The  key to resolving the problem is the interfacing of minicomputers

                                     184

-------
with biological systems.  This would permit the type of control necessary
to make specific alterations in toxicant concentration at prescribed
times, a means of handling large volumes of biological data and storing
it, and a capability for pre-programming the analysis of data so that
this can be accomplished in a cost-effective efficient manner.

The apparatus described herein represents an initial attempt to meet some
of these requirements.  This apparatus represents one of a series of uses
of minicomputers in the acquisition and analysis of biological data from
this laboratory (1,2,4,10,14).  A first attempt to resolve a problem is
usually imperfect.  Nevertheless, the apparatus and the accompanying re-
sults described in this paper provide compelling evidence that sufficient
promise exists for resolving some or all of the problems discussed pre-
viously and for justifying continued research and development of the
apparatus.  It is relatively inexpensive, simple, requires primarily
shelf items, and is within the reach of most industrial facilities and
nearly all major industrial research laboratories.  In addition, once
the apparatus is in place and functioning, it should be possible for
technicians to operate and maintain it.

                                THEORY

The electrical action potential of muscle movement in the branchial
region of a fish during ventilatory or "breathing" activity produces an
electronic signal.  This signal can be detected using electrodes sub-
merged at opposite ends of a fish holding compartment (Fig. 1).  When
this microvolt signal is amplified to the 4 to 10 volt range, it can be
recorded physically on a strip chart recorder (Fig. 2) or electronically
on magnetic tape with a minicomputer.

That the ventilatory behavior of a fish will change when various toxi-
cants, even at sublethal concentrations, are introduced into the water
has often been demonstrated (6,11,14).  However, the use of this principle
as a continuous water quality monitor has been handicapped by the huge
potential volume of data.  A single unstressed bluegill (Lepomis macro-
chirus) can produce as many as 100,000 ventilatory signals in a 24-hour
period.  Natural variability requires that many individual organisms be
monitored simultaneously for any analysis to be statistically valid and
to prevent false indications of environmental contamination.  Most
workers in the past have recorded ventilatory activity intermittently
with strip chart recorders and then analyzed data by personal observa-
tion, a tedious procedure at best (3,13).  A more efficient use of such
data is the continuous accumulation of ventilatory data from a large
number of individuals over a long period of time.  The most versatile
and efficient method of accomplishing this is interfacing the fish with
a minicomputer.

                               APPARATUS

An earlier version of the system now being used to investigate the
ventilatory responses of the bluegill to sublethal doses of heavy metals
has been described by Thompson et al. (14).  Cairns et al. (5) demonstrated

                                    185

-------
that, with slight modifications,  this system could be used with numerous
species, both freshwater and marine.

Sensing Chambers

A fish holding chamber with a total of 16 individual 700 ml capacity
monitoring compartments was constructed of plexiglas (Figs. 1 and 3).
The bottom of this chamber was constructed of 6.35 mm transparent plexi-
glas.  The ends and dividing baffles were of 3.18 mm transparent plexi-
glas, and the sides of each individual compartment were of 3.18 mm opaque
white plexiglas.  All joints were fused with methylene chloride.

The submerged electrodes were of type 304 stainless steel wire and were
attached with silicone aquarium cement to the baffles at either end of
the fish monitoring compartment.  The electrodes were on the side of
the dividing baffles opposite the fish to prevent contact between fish
and electrode.  Each electrode was formed in such a way that the wire
never passed through the flow of water from any of the perforations in
the dividing baffles (Fig. 1).  The electrode wire was insulated from
just below the water line, where the insulation was sealed to the wire
with silicone cement to prevent water intrusion, to the barrier strip
on the outside of the tank.  The only exposed wire was below the water
surface.

Since fish respond to visual stimuli and to vibrations in a manner
similar to their response to chemical stimuli, the entire fish holding
chamber was enclosed in a plywood isolation chamber.  The supports of
this enclosure were anchored in sand filled containers to protect
against vibration (Fig. 3).

Water Delivery

Diluent was dechlorinated tap water maintained at a constant tempera-
ture of 22 C.  This was delivered by gravity flow to three plexiglas
mixing chambers where the diluent could be mixed with various toxicant
solutions if desired (Fig. 3).  The water was then passed into delivery
chambers, also of plexiglas, equipped with 1 mm diameter standpipes.
A constant head pressure was maintained over these standpipes by delivery
in excess and the excess was drained through an adjustable standpipe.
This allowed delivery of water or water and toxicant at a constant rate
and concentration to each testing compartment (Fig. 3).

Data Accumulation

Each monitoring compartment was connected to an amplifier specially
designed for fish ventilatory investigations (8).  These amplifiers were
interfaced to a PDP-8E minicomputer through a 16-channel multiplexed
analog  to digital converter.  Data were accumulated for 15-minute periods,
printed on a Decwriter II computer terminal, and stored on magnetic
tape cassettes by a Decassette TU60 tape drive.  These data subsequently
could be transcribed by the minicomputer over a telephone line to the
Virginia Polytechnic Institute and State University's main computer  (an
IBM/370-1580) for analyses.

                                    186

-------
Computer Program

Although it is not apparent in Figure 2, the ventilatory signal is often
bimodal (7,14).  This bimodality is an artifact of the  amplifier design
and is more pronounced in more slowly ventilating individuals (14).   The
signal has been visually correlated with the ventilatory activity of L^.
macrochirus, and each waveform, bimodal or not, represents one complete
ventilatory cycle (7).  The computer program developed for this system
is capable of rejecting this artifact when it is present and of accu-
mulating the total counts as well as an approximation of the average
amplitude of the signal.  This program also uses a software controlled
variable threshold to reject low amplitude noise.  Rejection of real
signals is less than 2 percent, while extraneous counts are less than 1
percent.

Toxicant Delivery

An Integral Masterflex, variable speed, peristaltic pump drive with
remote control capability was used to introduce the toxicant into the
mixing chambers (Fig. 3).  As many as nine pump heads could be used
simultaneously for delivery of different toxicant solutions to each
mixing chamber or for producing various combinations in the mixing
chambers.  The toxicant solutions were pumped through 1.66 mm inside
diameter silicone tubing.  One of the three mixing chambers was used
for control and, thus, never received toxicant solution.

The minicomputer used a digital to analog converter to deliver a pre-
selected voltage to the pump drive.  A computer-pump interface was used
to convert the output voltage from the computer to the optimum range
for use by the pump (Fig. 4).  The computer output voltage could be
changed by the computer at preselected one-hour intervals to simulate
fluctuating "spills" or releases.  The concentration of toxicant in the
individual sensing compartments reached maximum concentration in 15-20
minutes, depending on the flow rate through each compartment.

                           TESTING PROCEDURE
                                             • I
Three tests were run to test the effect of Cu   on the ventilatory
activity rates of the bluegill.  During each test, 11 fish received
applications of CuCl« solutions, and 5 were left untreated as controls.
The fish were maintained under conditions of continuous light to reduce
diurnal variation.  The fish were not fed while in the chambers or for
24 hours before being placed in the chambers.

The fish were allowed 24 hours in the testing compartments to acclimate
before ventilatory signals were recorded.  Ventilatory rates were
recorded then for a total of 216 hours (9 days).  No toxicant was applied
during the first 96 or the last 24 hours.  Toxicant solutions were
applied in various patterns during the intervening 96 hours.  During
the first test, the toxicant was delivered at a constant rate for 96
hours at a concentration of 0.5 mg Cu   per liter (Fig. 5).  During the
second and third tests, approximately equivalent amounts were delivered
in intermittent patterns.  The toxicant in test two was delivered

                                   187

-------
during two variable 20-hour applications separated by a 20-hour recovery
period (Fig. 6).  The resulting concentrations of Cu   were 0.5, 1.0,
and 1.5 mg/1 in 4-hour increments.  Toxicant during run three was de-
livered in eight 5-hour applications with 1 hour increments (Fig. 7).
The average concentrations during the 1-hour increments were 0.5, 1.1,
and 1.3 mg/1.  The variation in concentration is due to slight fluctu-
ations in the voltage delivered to the pump by the computer.

                             DATA ANALYSIS

The total counts from four consecutive, 15-minute reporting periods were
summed and divided by 60 to produce the average rate per minute for each
1-hour period for each individual fish being monitored.  These average
rates were graphed for the entire 216 hours of the test.  Examples of
these graphs are in Figures 5, 6, and 7.

While it is true that consecutive hourly averages for a single individual
are not independent observations and that there is considerable natural
variation that makes comparison between two individuals difficult, the
application of a conventional one-way analysis of variance to data with
large sample sizes is valid.  This method was used to examine the
differences between the mean rates of the combined controls and the three
treatment patterns.  For the purpose of the analyses the various patterns
are indicated as follows.  Pattern 0 = controls, all three tests combined;
pattern 1 = one toxicant application; pattern 2 = two applications; and
pattern 3 = eight applications.  Separate analyses were performed on
the pretreatment, treatment, and post-treatment data.  Subsequently, a
post-mortem Duncan's multiple range test was also carried out to locate
possible sources of major differences.

The average rate for an entire period was calculated for each fish, and
the analyses were run with these averaged data.  The period for pre-
treatment was always 96 hours.  The time for treatment was from the
start of the first toxicant application until the end of the last applica-
tion and varied from run to run.  The time for the post-treatment period
was all the remaining time after  toxicant delivery ceased and was also
variable.

                                RESULTS

Five of the 48  fish are not included in the final analysis due  to various
physical and electronic failures.  Among the remaining  30 individuals
that received toxicant in  1 of the 3 patterns, 27 showed an average
increase of 16.4 counts per minute during the treatment period.  The  3
that decreased  did so by an average of 1.4 counts per minute.   Among  the
13 control  fish, 8 increased by an average of 3.9 counts per minute,
and 5 decreased by an average  of  3.5 counts per minute.

After  the  toxicant flow ceased, most of the treated  individuals  responded
with reduced ventilatory rates, and 29 of the treated fish  survived the
treatment  period.  Four of  these  had increased counts averaging  3.4 per
minute, and  25  decreased by an average of 17.0 per minute.  Among the

                                    188

-------
13 untreated controls, 3 increased by an average of 2.2 counts per minute,
and 10 decreased by a 5.7 count average.

The analyses of variance indicated that there was no statistical differ-
ence between the controls and the treated individuals during either the
pretreatment or the post-treatment periods; the corresponding probabil-
ities for no difference were 30 percent and 9 percent (Table 1).  Thus,
the null hypothesis was acceptable at the 0.05 alpha level (Table 1).
During the treatment period, the probability for no difference was 4
percent, indicating rejection of the null at the 0.05 level.

Duncan's multiple range test indicated no difference between controls
and any other fish during the pretreatment period.  During the treatment
period, there was no detectable difference between any of the three
treatment patterns.  The controls, however, could not be differentiated
from pattern 3, but they could be separated from patterns 1 and 2.
However, it should be noted that these analyses are based on the average
rate for the entire period from the start of the first toxicant applica-
tion to the end of the last application, including the intermittent
periods of no toxicant application.  It is quite apparent from Figure
7 that there is definitely a noticable reaction in some of the individ-
uals.  That this does depend to some extent on the pattern of application
can be noted from Figures 5, 6, and 7.  The results of Duncan's test
for the post-treatment period gave mixed results also.  The treated
fish from run one (pattern 1) could not be distinguished from the controls,
neither could the treated fish from run two (pattern 2) or three (pattern
3) be distinguished from the controls.  However, these could be differ-
entiated from those treated with pattern 1 (Table 1).

                              DISCUSSION

Some tentative conclusions can be drawn from this work even though
any final conclusions must await a larger data base.

     1.  There is a trend toward a reduction in the average rate
         of ventilatory activity over time.  This is probably due
         to the fact that the fish are not fed while in the sensing
         chambers.

     2.  The presence of cupric chloride in solution at sublethal
         levels causes a general increase in the ventilatory
         activity of most bluegills which can be statistically
         significant if the sample size is large enough.

     3.  With the analyses used, differences between treated
         and untreated fish can be detected, but differences
         between application patterns cannot.  However, the
         response to the different patterns can be distinctive
         as can be determined from visual inspection of the
         graphed data.

      4. There was apparent recovery with respect to ventila-
         tory rate in many individuals within 24 hours after

                                    189

-------
         toxicant delivery ceased.   However,  long-term chronic
         effects cannot be implied  from these data at present.

     5.  The apparatus provided the type of information it was
         designed to produce and merits additional attention.

                            ACKNOWLEDGMENTS

This research was supported by Grant R 80 42 740 10 from the United
States Environmental Protection Agency.  Manuscript preparation
was supported by a grant-in-aid from the Monsanto Fund.  We also
wish to thank J. D. Landers, G. L.  Nunn and W. J. Showalter for
their expert technical assistance without which this paper would
not have been possible.  The authors acknowledge Darla Donald
for the editing and preparation of this manuscript and Betty
Higginbotham for the final typed copy.

                              REFERENCES

 1.  Cairns, J., Jr., "Critical Species Including Man, Within the Bio-
     sphere," Naturwissenschaften,  62, 193-199 (1975).

 2.  Cairns, J., Jr., K. L. Dicksons and J. Slocomb, "The ABC's of
     Diatom Identification Using Laser Holography,"  Hydrobiologia,
     54, 7-16 (1977).

 3.  Cairns, J., Jr., K. L. Dickson, R. E. Sparks, and W. T. Waller,
     "A Preliminary Report on Rapid Biological Information Ssytems
     for Water Pollution Control," J_, Water Pollut. Control Fed.,
     •42, 685-703 (1970).

 4.  Cairns, J., Jr., and D. Gruber, "Coupling Mini- and Microcomputers
     to Biological Early Warning Systems," BioScience, 29, 665-669
     (1979).

 5.  Cairns, J., Jr., K. W. Thompson, J. D. Landers, Jr., M. J. McKee,
     and A. C. Hendricks, "Suitability of Some Freshwater and Marine
     Fishes for Use with a Minicomputer Interfaced Biological Monitoring
     System," Water Resour. Bull.,  (in press).

 6.  Drummond, R. A., G. F. Olson,  and A. R. Batterman, "Cough Response
     and Uptake of Mercury by Brook Trout, Salvelinus fontinalis,
     Exposed to Mercuric Compounds at Different Hydrogen Ion Concentra-
     tions," Trans. Am. Fish. Soc., 103. 244-249 (1974).

 7.  Gruber, D., J. Cairns, Jr., K. L. Dickson, and A. C. Hendricks,
     "A Cinematographic Investigation into the Fish's Bioelectric
     Breathing Signal," j;. Fish. Biol., 14_, 429-436 (1979).

 8.  Gruber, D., J. Cairns, Jr., K. L. Dickson, R. Hummel, III, A.
     Maciorowski, and W. H. van der Schalie, "An Inexpensive, Noise-
     Immune Amplifier Designed for Computer Monitoring of Ventilatory
     Movements of Fish and Other Biological Events, " Trans. Am. Fish.
     Soc., 106. 497-499 (1977).
                                   190

-------
 9.  Landers, J.  D., Jr., "Automated Biological Monitoring System Design,"
     Masters thesis, Virginia Polytechnic Institute and State University
     (1979).

10.  Lubinski, K. S., K. L. Dickson, and J. Cairns, Jr., "Microprocessor-
     based Interface Converts Video Signals for Object (Fish) Tracking,"
     Computer Design, Dec., 81-87 (1977).

11.  Morgan, W. S. G., and P. C. Kuhn, "A Method to Monitor the Effects
     of Toxicants Upon Breathing Rate of Largemouth Bass (Micropterus
     salmoides Lacepede).  Water Res., 8^, 67-77 (1974).

12.  National Academy of Sciences, Water Quality Criteria of 1972,
     Washington,  D. C. (1974).

13.  Sparks, R. E., J. Cairns, Jr., and A. G. Heath, "The Use of Bluegill
     Breathing Rates to Detect Zinc," Water Res., 6, 895-911 (1972).

14.  Thompson, K. W., K. L. Dickson, D. Hooley, J. Cairns, Jr., and
     W. H. van der Schalie,"A Computer Monitored Aquatic Bioassay
     System," Proceedings of Tenth Mid-Atlantic Industrial Waste
     Conference,  University of Delaware, 61-66 (1978).
                                    191

-------
                                TABLE 1

                Results of Analysis of Variance (ANOVA)
           Pretreatment period   Treatment period  Post-treatment period

              F     P     df      F     P     df      F     P     df
ANOVA       1.26   0.30* 3,39    3.16  0.04  3,39    2.29  0.09* 3,38
Results

Duncan's**
Test        0=1=2=3        1=2=3,0=3    1=0,0=3=2
Results
 * Null hypothesis may be accepted at the 0.05  a  level.

** 0 = controls, 1 = continuous application, 2=2 applications,
   3=8 applications.
                                    192

-------

                                   FISH HOLDING CHAMBER

                                   ^.ELECTRODES

                        WATER
                       RECEIVING
                       CHAMBER
                               DRAIN
                 FiGUPvE 1    Individual monitoring compartment.
             +3-
           O

             OH
           UJ
           or
            -3J
                 r-

                 0
                                Lepomis macrochirus
                                 (Bluegill  Sunfish)
10
                                     SECONDS
                                                               30
FIGURE  2    An example of the ventilatory signal  as recorded  on a strip
            chart recorder.
                                      193

-------
FIGURE 3    The water  delivery and  data  accumulation system.
                            194

-------
        DAC INPUT
                                                  1.5 K
             6I2K
                                                  -15
                                                              > TO PUMP
FIGURE 4   Schematic  for  the computer/pump interface  (redrawn  from
           Landers  (9)).
                                  195

-------
4-1
CO
M
•H
M-l
CU
_(-*,
4-1

a
o
J-l
U-l

CO
cu
4-1
Cfl
J_|

^
J_|
O
4-1
o
c
•o
cu
^
•H
CU
O
cu
J-l

CU
c
cfl
t-H


4-1
CO
V-i
•H
&
4J
T3
c
cfl

*"O
pi
o
0
CU
CO

CU
^rj
4-1

C
o

OH
4-1
CU
a
4-1
CO
CU
^(
4-1

CU
r|
4-1

60
Pi
•H
rl
3



Cfl -H CO+
rH
•H
4-1
g
•*
01
60
cfl
i-l
CU
^
cfl

CU
t~|
4-1

4H
O

CO
CU
!— |
ft
a
CO

w
m
w
14-j

CU
4-1
C
o

r-t
CO
3
T)
•H
£>
•rl
T3
C
•H

CU
£1
H


•
4-1
CO
CU
4-1


0

4_)
cu
r-l
•H
rj
13

Pi
0
•rl
4-1
Cfl
O
•H
rH
ft

CO

4-1
rj
Cfl
O
•H
X
0
4-1


3
U

O
,_,
••••^
60
a

l/^
•
c

T3
cu
^
•H
CU
O
CU
^_]

en
cu
fi
cfl

ft


tilatory
c
CU
^

c
•H

CU
CO
Cfl
cu
rl
O
c
•H

c
CU
13
*rj
3
CO
CU
fi


•
/-*^
cfl
cu
^1
cfl

*o
cu
T3
cfl

CD
N«^

•13
o
•rl

CU
ft



rH
Cfl
3

cfl
^4
60

CO

r*.
4J
•H
fe

4-1
C
CU
a
4-1
Cfl
cu
4-1
cu
.c
4-1

l|H
O

4J
}^
CO
4-1
CO

CU
r*.
4J

4->
Cfl

01
4-1
Cfl



typical

CO
cfl
^

13
O
•rl
}_l
(U
ft

4-1
PI
CU
a
4-1
Cfl
CU
^
4-1

CU
4-1
MH
0

*"O
c

-------
C3
O
CJ
CU
en

cu
43
4-1

s
o
S-i
14H

en
CU
4-1
cd
S-i

[>^
S-i
o
4-1
cd
r-l
•H
4-1
£3
d)


d)
60
cd
S-i
d)
p»
cd

dl
43
4-1

m
O

en
0)
rH
cx
0
cd

w
w
£D
0
4-1
cd

en
cu
en
0
13

O
|5
4J

(3
•H

4-1
£3
cd
CJ
•H
X
o
4-1

13
d)
[>
•H
CU
a
d>
S-i

t~j
en
•H
14-1

^rj
dl
4-1
cd
d)
S-i
4J

d)
43
H


•
4-1
en
cu
4-1



a
o
•rl
4-1
CO
•H
^
cd


d)
43
4-1

CU
4J
O
S3


•
4-
-|-
^
u

rH
\^
60
s

LP|
•
tH

•^
C3
cd

*
o

iH

n
LO
•
o

<4H
0

en
rH
d)
^»
d)
rH



en
•rl
MH

13
d)
4J
cd
cu
S-i
4J

o
^
4J

dl

fH


•
(~1
en
•H
14H

13
cu
4-1
cd
cu

4-J

d)
43
4J

60
C3
0
0
cd

dl
en
a
0
cx
en
cu
S-i

<4H
O






d)
43
4-1

60
C
0
B


13
G
3
O
m

en
dl
g
dl
j»i
4-1
X


d)
43
4-1

4-1
f3
d)
en
d)
M
CX
d)
j_i

d)
^
a
60
•H
U-l

CO
•H
43
4-1

C
•H



d)
4J
4J
cd

d)

cd


c
•rt

a
o
•H
4-1
O

rrj
d)


•U
a
dl

cd
CX
cx
cd

cu

f— (


•
en
rH
cd
3

•H
^
•H
T)
a
•H

rrj
d)
4-1
cd
d)

4-1



d)
43
4-1
0
(3
cd

0
4-1

d)
3
Tt

en
•H

(3
O
•H
4-1
cd
o
•H
rH
cx
CX
n)

4-1
a
cd
o
•H
X
0
4J

13
C3
O
O
CU
en

cu
43
4-1

14-1
0

rs^
cd
cu
cx



0

d)
T3
3
4-1
•H
rH
CX
B
cd

QJ
43
4-1

C
•H

C
O
•H
4J
CJ
3
13
d)
J_)

cd

n
4-1
C
cd
u
•H
X
o
4J

d)
43
4J

O
4J

C
O
•H
4J
CJ
cd
d)




TJ
dl
13
•rl
cd
0
d)
S-4

^
rH
43
cd
42
0
S-i
cx

cu
4J
cd
^

cu
43
4-1

CU
rH
•H
f{
£2


•
rH
cd
C3
60
•H
en

^
}_i
o
4-1
cd
rH
•rl
4-1
C
d)
p>

d)
43
4-1



d)
a
cx >,
O r-l
S-I 4-1
13 a
di
d) S-i
13 S-i
3 3
W CJ
•H
rH en
CX-H

fl) C
o
d) -H
43 4J
4-1 O
CO
- d)
d) SJ
4-1
3 cn
a -H
•H 43


r4
dl •
CX fi
0
en -H
4-1 4J
fi 0
3 0)
0 4-1
O d)
13
O
CM 14H
rH O

MH 4J
O -H
B
0) vH
60 rH
c
cd CU
>-< 43
4-1
0)
43 £
W 0
rH
C d)
•H 42



later

cd

4-1
CO

rH
•H
CO
4-J
d)
id

d)
rl
O
B

£3
•rH

rrj
d)
4-1
S-I
O
CX
d)
S-i

CU
43

rH
rH
•H
^

13
rj
cd

13
d)
•H
13
3
4-1
cn

60
a
•H
dl
43






















































•
CU
4-1
cd
13



g  § g  o
              197

-------
               T3
                M     n3
               •rl      C
               x  4J  cfl
                4-1  h
                    O   «
                CU  42 rH
               43  to   •
                a 43   -
                o  60 m
                rl -H   •
               m  CD o
                w  C  , rH  01
rl  CX  >
O  CX  0)
4-1  cfl -H
       43
        O
                §
                CO
            0)  Ol
            fj  J_j
            4J  CX  0)
                CU 43
                rl H  13
                        C
                        O
                     •  O
                3  co  cu
                60 rH  CO
            0)  -rl  CO
            CO  4-1  3  CU
            S     13  43
            O  M -rl  4J
            CX-rl  >

            0)  4-1 13  O
        ff-
        O  01

        ca
                                C!
                                •H
 cfl
rH
•H  cfl
 4-1  &

 CU  4J
 >  C
    cfl
 cu  o
 60 >H
 cfl  X
 M  O
 O  4-1

 CO  CU
                                43   CU   CX
                                CO  4-1
                                •rl   Cfl   0)
                                14-1   0)  43
                                     ^_,  4J
                        W
            l-l  4J  CU  Cfl
            CO  cfl  43
            >  CU  4J  CU
                VJ      4-1
            01  4-J
     o
 CU  4J
43   cfl
 4-1  rH
    •H
 O  4-1
 4J   C2
     cu
 0   >
 o
•rl   0>
 4-J  43
 O  4J
 cfl
 cu  m
 S-i   O

 ^   cu
 Q)  13
43   r
 4-1  4-1

 S3

 *   I
 Q   CO
                                    4-1           Cfl
                                     O          tj
                                         CU
                                     CU  43      rl
                                     60  4-1  60  Ol
                                             C
                                     CO
                                         !? -H  cfl
                                         O  0) rH
                                     cu  cu
                                    43 43
                                     4-1
                                        13
                                     d  cu
                                    •H  PL
                                         CX
                                    T3  O
                                     0)  M
                                     (3 -O
                                    •H

                                                     0)
        CO  4-1  o  O
                    60 Cfl
                    C  rl
                 01 4J  O
                43     O
                DS  6  £3
            O)  4-1  cfl  -H
            4-1
            O  0)  13  C
            S3  43  fi  O
                H  3  vH
                    O  4J
             •      mo
 cu
 3
13

 CO -H
•H

 C3  O
 O -rl
•H 4-1
4-1  U
 Cfl  3
 O T3
•H  CU
                                                     o
                                                     J-i  0)
                                                     P.43
                                                         4-1
                                                     CU
                                                     4J   -
                                                     CO  CU
                                                     rf  4J

                                                     0)  C
                                                    43  -H
                                        4J   CO

                                        CU  rH
                                        >-<  -H
                                        rl   Cfl
                                        3  4-1
                                        O   0)
                                            13
                                        CO
                                        •H   CU
                                             S-i

                                        O   £3
                                        •H
                                        4J   C
                                        O  -rl
                                        cd
                                        CU  T3
                                        h   Ol
                                            4J
                                        CO   S-I
                                        •rl   O
                                        43   CX
                                        H   CU
                                             rl

                                          •   CU
                 Q) C
                rH-H
O  CU+  42  CO 13  CX
    £2  3  en  cu  a)  ex cd
    H O vH  S  t-i  cfl
           4-1  CU           ••
       rH      s_i  4J  4-1  4->
     . -^ T3  4J  C  G  0
    CO  60  CU  X  0)  cfl  cfl
    4-1  g  4-J  0)  "
    CO
                                                    rH  CU
                                                    •H  CX
                 CO CO
                 X 0)
                W 4-1
                r--
                w

                o
                                         cfl
                                         CX
                                             O   O
                                         §
                                         O
                                            •H rH
                                             4-J rH
                                             a -H
                                             CU  J3
                                             4J
                                             Q) T3
                                            13  0
                                                 CO
                 CO
            ro   cu   cu
              •   h  43   CX  O   6
            rH  4-1   4J   CO   4-1   4J
                                60 O
                                •H  og
                                W  rH
                                                 01
                                             4-1 -rl
                                            •H T3
                                             S  3
                                            •H 4J
                                            rH  CO
198

-------
             ON-LINE MONITORING OF TOXIC MATERIALS IN SEWAGE
                   AT  THE  LAWRENCE  LIVERMORE  LABORATORY

              M. Auyong,  J.  L.  Cate,  Jr.  and D.  W. Rueppel
           Lawrence Livermore Laboratory, Livermore,  Ca. 94550
                                 ABSTRACT

It is becoming increasingly important for industry to prevent releases
of potentially toxic material to the environment.  The Lawrence
Livermore Laboratory, a U.S. Department of Energy research facility,
has developed a system to monitor its sewage effluent on a continuous
basis.

A representative fraction of the total waste stream leaving the Plant
is passed through a detection assembly consisting of an x-ray
fluorescence unit which detects high levels of metals, sodium iodide
crystal detectors that scan the sewage for the presence of elevated
levels of radiation, and an industrial probe for pH monitoring.  With
the aid of a microprocessor, the data collected is reduced and analyzed
to determine whether levels are approaching established environmental
limits.  Currently, if preset pH or radiation levels are exceeded, a
sample of the suspect sewage is automatically collected for further
analysis, and an alarm is sent to a station where personnel can be
alerted to respond on a 24-hour basis.  In the same manner, spectral
data from the x-ray fluorescence unit will be routed through the
24-hour alarm system as soon as evaluation of the unit is complete.

The monitoring system has played an important role in averting
treatment plant problems when a spill has occurred.  The design of the
system and operational experience will be discussed.

                               INTRODUCTION

Since its beginning in 1952, the Lawrence Livermore Laboratory (LLL),
Livermore, California has supported an environmental surveillance
program to determine its impact, if any, on the local environment.  The
Laboratory, located in a suburban valley 65 km east of San Francisco,
is involved in widely varied research and development programs for the
U.S. Department of Energy.  The main efforts are in nuclear and
nonnuclear weapons, magnetic and laser fusion energy, biomedical
research, and nonnuclear energy technologies, such as geothermal power
and fossil fuel utilization.  These programs and the support efforts
involved in carrying them out generate waste products that could
present a negative impact on the environment if not properly managed.
Although the Laboratory has an extensive program to control and dispose
of all potentially toxic materials at the source, the environmental
monitoring program serves as a check on control procedures.  An
important part of this monitoring program is the on-line wastewater
monitoring system, which detects elevated radiation, metal
concentration, and pH levels.

                                   199

-------
                               BACKGROUND

The Laboratory's sanitary sewer system is a possible route for the
escape of toxic materials.  Its effluent, which is approximately
350,000 gallons per day, is processed by the municipal Livermore Water
Reclamation Plant.  The plant is  a secondary treatment operation, whose
output is discharged to several different sources.  Until very
recently, most of this effluent was returned to a creek which
eventually recharged a groundwater basin.  A portion of the effluent is
used for irrigating vegetation along roadways and a few local crops.
Earlier this month, however, a transport pipeline to the San Francisco
Bay was completed, and most of the treated effluent now flows to the
Bay.  The sewer monitoring system was designed and constructed to
detect toxic material releases and to facilitate immediate response
action when they occur.  This is  necessary in preventing damage to the
city treatment plant and regions receiving discharged effluent.  If
toxic materials are detected at levels that exceed predetermined LLL
alarm limits, signals are sent to a central alarm station that is
manned 24-hours a day and a sample of the suspect toxic effluent is
automatically collected.  Since two hours pass before LLL effluent
reaches the treatment plant, sufficient time is available to alert
emergency personnel, evaluate the situation and, if necessary, arrange
for diversion of the material to emergency holding basins at the
treatment plant.  The toxic waste can be treated in the basins or
removed without destroying or cutting down on the efficiency of the
treatment plant.

                       MONITORING  SYSTEM COMPONENTS

The on-line monitoring system consists of several components.  The flow
route to these components starts at a point in a manhole where all
Laboratory sewage discharge lines converge.  As the sewage flows
through a Parshall flow-measuring flume, approximately 40 L/min of it
is pumped to an above ground building where the detection instrumen-
tation is located.  Inside this building, the sample enters a tank
housing the high- and low-energy radiation detectors, pH probe, and a
sampling line leading to the metal analyzer unit.  After it is scanned,
the sample is returned to the sewer via an outlet pipe.

Radiation detectors

Radioisotopes being used  at LLL that could be released in quantities
that would exceed continuous exposure maximum permissible concentration
 (MPC) levels  are  9°Sr, 235Us 238U§ 237Np> 238pU} 239Pu
241Am and 244Cm.  All of these, with the exception of 9°Sr, emit
heavy element x-rays  and  low energy gamma rays during decay.  For
90Sr, low energy  bremsstrahlung photons  give an  indication of
specific activity.  With this  in mind,  the radiation  detection system
was designed  (1).  The primary detector, a 3 x 127 mm sodium  iodide
crystal separated from the sample by a  thin polycarbonate plastic
                                    200

-------
sheet, allows recognition of the low energy x-rays, gamma rays, and
bremsstrahlung photons in the energy range of 10 to 100 keV.  In
addition, a 50 x 50 mm sodium iodide crystal detector is mounted
immediately adjacent to the sample tank to provide detection of higher
energy (100 - 1000 keV) events.

The electronics associated with the detectors includes single channel
analyzers and an Intel 8008 microprocessor which analyzes the data and
decides whether preset levels have been exceeded.  The alarm system for
radiation detection has been set at a level that corresponds to less
than 2% of the allowable monthly discharge limit for a continuous
release of 9°Sr or less than 0.5% of the monthly allowable discharge
limit for a continuous release of 239pu (2).

pH monitor

A commercial industrial pH probe is housed in the tank holding the
radiation detectors.  Output from the pH sensor is sent to the Intel
8008 microprocessor where the data is evaluated.  If the pH drops below
3 or exceeds 11 for five minutes within a "floating" ten minute block
of time, the alarm system is activated.

Metal detection

The deleterious effect of excess concentrations of heavy metals on
sewage treatment plant operations has been demonstrated in a number of
studies.  The Laboratory has had two incidents in the past when
inadvertent releases of toxic metals have created operating problems at
the sewage treatment plant.  In one incident, copper cyanide solution
destroyed 50% of the bacteria in the trickling filter unit, and in the
other, excess chrome significantly reduced the efficiency of one of the
two digester units.  These events prompted a search for a continuous
metal detection system that would prevent the recurrence of plant
down-time caused by a metal release to the sewer.  When the search was
started about two years ago, commercial units available appeared to be
unsuitable for sewage analysis.  The units required pre-analysis
filtering, which would affect the representability of the sample and
which would require a considerable amount of maintenance.  A monitoring
unit capable of detecting hazardous concentrations of ions found to be
most harmful to the bacteria in the treatment plant process,
specifically copper,-nickel, chrome and zinc was then designed.  The
system  meets the requirements placed on a continuous system:  it is
fairly reasonable in cost, especially when contrasted to the cost of
reseeding a treatment plant; works on-line and in real time; does not
require extensive pretreatment of the sample, thereby preserving
representability; and requires minimal maintenance.

The unit providing metals detection is an x-ray fluorescence analyzer
(XRFA).  Its design is based on the principle that elements emit
characteristic x-ray lines when excited by a radiation source.  These
                                   201

-------
x-ray lines can be measured by energy dispersion techniques to
determine their energy, which permits species indentification, while
the intensity of the lines is proportional to concentration.

A portion of the sample stream that flows through the tank described
previously is routed through a macerator to reduce any solids to
particles 100 microns or less in diameter.  The flow is then introduced
through a nozzle into a flume inclined at 45°.  The flow spreads to a
sheet 1 mm in depth before it reaches the source/detector region of the
unit.  A thin plastic window backed by air is positioned under the
stream in this region to prevent detection of the chrome and nickel in
the stainless steel flume.  A 109Cd source is used to excite the
elements of interest if they are present in the stream, and a
commercial xenon-C02 mixture x-ray proportional counter is the
detector used.  The output from the detector goes to amplifiers, then
through an analog-to-digital converter interfaced to a Digital
Equipment Company LSI-11 microprocessor.

A dual register system for the data output allows both a rapid response
to high metal concentrations (500 sec count) and a more sensitive
response to low concentrations released over a longer time frame (5000
sec accumulated count).  The data counting system also provides for
advisories at concentrations that are elevated, but not at alarm
levels.  Minimum advisory and alarm levels are shown in Table 1.

Maximum permissible discharge concentrations for copper, chrome,
nickel, and zinc have been established at LLL based on studies of the
effects of toxic metals on sewage treatment and the dilution factors
resulting from the intermixing of LLL's effluent with the domestic
sewage from the Livermore city population.  These concentrations, as
well as the 500- and 5000-sec alarm limits for the elements of concern,
are shown in Table 2.  Although only zinc will cause an alarm in 500
sec if its LLL limit is exceeded, copper, chrome and nickel will cause
alarms in 5000 sec or less if their limits are exceeded.  This will
provide adequate response time.

System monitor

In addition to detecting high- and low-enerqy radiation, pH excursions,
and heavy metal releases, the monitoring system trips an alarm when an
equipment malfunction occurs.  A signal is sent to the central alarm
station if problems such as a count rate lower than background, power
failure, or flow malfunction occurs.

Grab sampler

An important part of the monitoring system is a mechanism that extracts
a sample from the waste stream as soon as an alarm is tripped.  This
sample is often valuable in identification of the toxic material
involved and in the evaluation of the situation.
                                     202

-------
                                TABLE 1

                   Minimum Advisory and Alarm Levels
                          for  the  XRFA Monitor
                       500 sec                        5000 sec
Element    Advisory, ppm    Alarm, ppm    Advisory, ppm     Alarm, ppm


  Cr            36             119              11              37
  Mn            23              78               7              25
  Fe            15              50               5              16
  Co            10              33               3              10
  Ni             6              21               2                7
  Cu             5              16               2                5
  Zn             4              14               1                4
  As             2               7               1                2
  Se             2               7               1                2
  Hg             5              17               2                5
  Pb             4              12               1                4
                                  203

-------
                                TABLE 2
     Comparison of ILL Maximum Permissible Metal Discharge Limits
                         and XRFA Alarm Limits

Single metal
Cr
Cu
Zn
Ni

LLL limit, ppm
100
10
50
10
XRFA al
500 sec, ppm
119
16
14
21
arm limit
5000 sec,
37
5
4
7

ppm




Combinations: total concentration to be less than 100 ppm
                                 204

-------
Teletype printouts

Both the LSI-11 and Intel 8008 are interfaced to teleptypes.  The
printouts provide responding personnel with current as well as historical
data on effluent releases.

                            OPERATING EXPERIENCE

For the most part, the existing monitoring system has proved to be very
effective.  Since it has been in operation, there have been several
occasions when diversion of the Laboratory's flow at the Livermore
treatment plant was deemed necessary.  In all instances, alarms were set
off because of inadvertent acid releases.  These incidents alone have
proved the worth of the system since restoring the treatment plant to
normal operation would have been more costly than the investment made in
designing, constructing, operating and maintaining the monitoring system.

In addition to the pH excursions, there have been a number of radiation
alarms.  In all instances, grab samples analyzed have shown that the
material triggering the alarm was a radiopharmaceutical, generally 99Tc
from brain scans or 131 j from thyroid treatments.  These alarms are
caused by the radiopharmaceutical being excreted by outpatients who return
to work following medical treatment.

Although the alarm system has not yet been hooked up for the x-ray
fluorescence analyzer, an inadvertent release of about 40 ppm chrome from
a plating operation was noted on the teletype printout.

                           SUMMARY  AND CONCLUSIONS

The on-line sewer monitoring system cannot in itself prevent the
accidental discharge of toxic materials into sewers.  However, by rapidly
detecting releases and sounding alarms to alert emergency personnel to
respond, actions can be taken to prevent damage to the environment.
Because it is a real-time system, Laboratory personnel can more easily
locate the source of a toxic discharge and take corrective action to
prevent its recurrence.
                                    205

-------
                                           REFERENCES

1.   Gate,  0.  L.,  Jr.  and T. D.  Hoeger,  "A Radioisotope Monitoring  System
     for Sewage  Effluent,"  American  Industrial  Hygiene Journal,  693-699
     (October  1972).

2.   U.  S.  Department  of Energy,  "Standards  for Radiation  Protection,"  DOEM
     0524  (1975).
                             "Work performed under the auspices of the
                             U.S. Department of Energy by the Lawrence
                             Livermore Laboratory under contract number
                             W-7405-ENG-48."
                                            NOTICE

                  This report was prepared as an account of work sponsored by the United
                  States Government. Neither the  United  States nor the United States
                  Department of Energy,  nor any  of their employees, nor any  of their
                  contractors, subcontractors,  or their employees,  makes  any warranty,
                  express or  implied, or assumes any legal liability or responsibility for the
                  accuracy, completeness or usefulness  of any  information, apparatus,
                  product or process disclosed, or represents that its  use would not infringe
                  privately-owned rights.

                  Reference to a company or product name does not imply approval or
                  recommendation of the product by the University of California or the U.S.
                  Department of Energy to  the exclusion of others that may be suitable.
                                               206

-------
                       COMPARISON OF tOUR LEACHATE
                         GENERATION PROCEDURES

                               D. E. Bause
                        GCA/Technology Division
                               ABSTRACT

Four leachate procedures, ASTM Method A, ASTM Method B, EPA-OSW Ex-
traction (EP), and Carbonic Acid Extraction (CAE), have been evaluated
for their general applicability, reproducibility,  Environmental Assess-
ment methods compatibility,  and leaching power.   The leachates generated
by these methods were analyzed for nine metals by atomic absorption
methods and for F , C& , and S0^~ by ion chromatography.   Seven energy
process wastes including oil shale, FBC waste, fly ash, boiler slag,
scrubber sludge, and hopper ash were extracted to evaluate the general
applicability of the leachate tests.  The ASTM methods had the best
reproducibility, while the EP had the poorest precision.   The EP and
CAE procedures leached the largest quantities of trace metals from
the wastes.  However, based on the total metal concentration in the
sample, the leachate methods generally extracted < 1%.  The EP and ASTM-
B methods caused some difficulty with flameless AA analyses.   Based on
the RCRA criteria, five of the energy wastes would be classified as
hazardous by at least one of the leachate procedures.  Selenium usually
exceeded the threshold value for the leachate.
                             INTRODUCTION

In order to fulfill the solid waste characterization needs of Environ-
mental Assessment (EA) programs, the Process Measurements Branch of
EPA/IERL-RTP has directed research to identify a suitable leachate
generation procedure.  In conjunction with this effort, the GCA/Tech-
nology Division has evaluated four leachate generation methods.  The
evaluations include the following methods as described:
                                   207

-------
     •     EPA-OSW Extraction Procedure (EP) - an open
           system, acetic acid extraction;

     •     ASTM Method A (ASTM-A)  - a closed system, water
           extraction;

     f     ASTM Method B (ASTM-B)  - a closed system, acetic
           acid-acetate buffer extraction;

     •     Carbonic Acid Extraction (CAE)  - a closed system,
           CO^-saturated water extraction.

These procedures are discussed briefly below.  The experimental parame-
ters for these methods are compared in Table 1.

The Extraction Procedure (EP)

The EP method has been proposed by the EPA-OSW to meet the RCRA guide-
lines in evaluating the hazards of solid waste disposal.  With the
addition of acetic acid to the aqueous solution, the procedure
presumably intends to simulate the first stage of anaerobic degradation,
involving the formation of volatile, organic acids at a disposal site.
The acidic medium also provides a more aggressive leaching test than
the purely aqueous leachate.

The experimental procedure suggests a minimum size of 100 grams for the
extraction.  The separation of any liquid fraction from the original
sample is accomplished by filtration or centrifugation methods.  After
sample preparation, consisting of either grinding or subjecting the
sample to the structural integrity test, the solid is placed in an extractor
which must be capable of thoroughly mixing the solid and the leaching medium.
A stirring device is suggested, but other agitation methods may also be used.
An amount of deionized water equal to 16 times the weight of sample is added.
The pH of the resulting leachate is monitored, maintained at 5.2 4^ 0.2
and adjusted with 0.5 N acetic acid, if necessary.  The extraction proceeds
for 24 hours with a maximum addition of 4 m£ of acid per gram of sample per-
mitted to maintain the pH.  After 24 hours, the mixture is filtered and
deionized water is added to adjust the volume to 20 times the weight of the
sample.

ASTM Methods

The American Society for Testing and Materials  (ASTM) has proposed two pro-
cedures to determine the leachable components of a solid waste.  Both methods
are "intended as a rapid means of obtaining a solution for evaluation of the
extractable materials in waste.  They may be used to produce solutions for
the estimation of the relative environment hazard inherent in the leachings
from the waste".  The wastes are to be used in the form which they are
disposed.  Where available, sampling is to proceed using ASTM sample
methods developed for the specific industry.
                                    208

-------









*l
<^
\
S
(H
w
 CO
*~o
S-i 0
O f.
([-.[ | i
CU
cn S
cu cu
4-J 4—1
CU cfl
B ^
ctf O
S-i cfl
cfl CU
PH nJ

-H W
	 . to 
j"^ ! p |
x w
w
^ M
0 1
a H
O CO
cn <;
•H
S-l
cO
ft
B
o
o






























o
CN
Pd
CU 60
4-1
cfl 0
W S-l O
"^ 3 *~~*
O 4J
cO
cn
i
CN
0
0





^
•H
O
60
CJ
•H 0
PM 4J o
W CU ^H
•Si

£75
m
•
o







cu
1 <4-l
CU 3
4_) pq
cfl
4-1 T3 60
PQ CU -H
1 CJ CJ O
jsil 
H
Q) 60
-, CN 0
S H ps m
H CO
CO ^
,
Q 4-1
4-1 -H
S-i
4-1 60
CJ CU
CU 4-1
,n H
3
CO tH
CO
S-l S-i
0 3
4J
T3 CJ
C 3
•H S-i
S-i 4-1
O CO












•H O

CU
CO Td
3 CU
cn
.., o
cu ft
C cn
O -i-l
& P
c B
•H 0
cu
cn Td
P cu
cn
.»- O
CU ft
C w
O -H
^ f~]
I
cd
S-l
cfl
ft
01
S-l
PM

cu

ft C
S o
cO *H
W J->






























4-1
CO
CU
H






















































O
CN

,— 1







actor
S-l
4-1
X
W
id
cu
•rf
4-1
•H
G
CU
ft
cn
£
JZ)


cu
o
•H
cu
Q

60
C
•H
S-l
S-i
•H
4-1
CO







Td
cu
4-1
cn
0)
60
60
3
CO
   3
   cr
  •H
  TD  O
  •H  -H
  iH  4-1
   O  tO
      &
 O
,£
 4J
 cu
               o
               •H
               4-1
               Cfl
                                 J-l
                                 ,c
                                 oo
                                                o
                                         m
                                CM
                                         LO
                                 cn
                                 S-i
                                00
                                o-
                                         LO
                                                               §
                                                               o
60
C
•H
4J
Cfl
O
O
S-i
ft
•H
CJ
cu








S-l
Q)
ry]
cO

CO


T3
CU
Td
C
cu
B
B
o
CJ
CU
&


•
cn
S-i
£1

CO
>J-




                                                              B
                                                              O
                                                              3
  CO

209
 cu

•H
H

 C
 O
•H
4J
 cfl
                  60
                                         ft
c
o
4J
3
(~n
0
CO

60
B
•H
t~°.
O
cfl
CU






M-t
O

S-i
CU
,0
B
3




CO
C
o
•H
4-1
O
tfl
j^j
4J
X




CU
S-i
3
4-1
Cfl
S-i
cu
ft
S
cu
                                                     W

-------
The water extraction (ASTM Method A), uses Type IV water for the
extraction, while the ASTM Method B employs an acetic acid-acetate
buffer solution to leach the metals from the wastes.  The experi-
mental procedure, however, is identical for each method.  A minimum
sample size of 350 g is recommended for each method.

The quantity of sample chosen for leaching is placed in a round,
wide-mouth bottle (constructed of material appropriate for the solid
waste and subsequent analyses) and mixed with the fUO or acetate buffer.
The volume, in milliliters of leachate added, is equal to four times
the weight in grams of the sample.  Mixing of the phases is accomplished
by any apparatus which is capable of producing the constant movement
equivalent to a reciprocating shaker operated at 60 to 70 one-inch
strokes per minute.   Agitation is continued for 48 hours, followed by
vacuum filtration of the liquid phase.

The Carbonic Acid Extraction  (CAE)

The Carbonic Acid Extraction  (CAE) was introduced as an alternative to
the leaching media of the previous methods.  The CCL-saturated water
was intended to simulate the aggressive leaching characteristics of
the acetic acid and acetate buffer solutions without having the problems
associated with the bioassay tests and analytical methods.

In order to maintain a CO^-saturated leachate, the extraction must be
performed in a closed system.  Thus, the CAE followed the basic pro-
cedure and agitation method  of the ASTM methods.  A minimum sample
size of 100 g is suggested for the extraction.  Round, wide-mouth
linear polyethylene  bottles are used to contain the leachate.   The liquid
to solid ratio is 16:1 and was chosen to minimize common ion effects,
which may affect the solubility of some species in the leachate.  It was
hoped that this liquid to solid ratio would be low enough to prevent the
trace elements from being diluted below the AA detection limits.

The carbonic acid solution is prepared by bubbling CO,, through the de-
ionized water until the pH reaches a minimum (approximately 3.9 to 4.0).
The mixture was agitated at the rate advocated by the ASTM methods (60
strokes/minute).  After shaking for 48 hours, the leachates were vacuum
filtered through a 0.45 ym filter.

Evaluations

The evaluations of the four leachate procedures were based upon the
following criteria:

     •     General Applicability - Any procedure employed as
           part of EA methodology must be amenable to a wide
           range of waste materials;
                                   210

-------
     •     Reproducibility - In order to make valid judgements
           regarding the potential hazard of a waste, the re-
           producibility of the generation procedure must be
           well defined;

     I     EA Methods Compatibility - It is highly desirable
           that the leachate produced not necessitate modifica-
           tions to the established analytical procedures;

     •     Leaching Power - To the extent practical in the
           laboratory, the leachate generation procedure
           utilized should simulate the anticipated fate of
           the waste.

                         RESULTS AND DISCUSSION

General Applicability of the Leachate Methods

To evaluate the general applicability of the leachate methods, the
four procedures were applied to a variety of energy process wastes,
including oil shale, fluidized bed combustion waste, bituminous coal
fly ash, bituminous coal boiler slag, lignitic coal scrubber sludge,
and hopper ash from a coal-fired power plant.  Except for the hopper
ash, the waste samples were supplied as part of an ASTM interlaboratory
test program to assess three extraction procedures, the ASTM-A, ASTM-
B, and EP methods.   Most of these wastes were essentially dry, since
no weight was lost when the percent moisture was determined for the
ASTM methods.  The only exception was the scrubber sludge, which had a
weight loss of 28% upon drying.  For the wastes tested, no procedural
problems were encountered with any of the leaching methods.  All of the
samples were extracted in the form in which they were received.  None of
the  wastes had to be ground or subjected to the structural integrity
test, as prescribed in the EP method.  Since the samples were basically
dry solids, the preparation schemes of the leachate methods have not
been tested thoroughly.  For more complex industrial wastes, a proto-
col for liquid-solid separation may not be adequately addressed by
some of the procedures.

Leaching Power

The final pH data for the leachates are summarized in Table 2, which
also indicates the number of extractions completed for each sample
and each leachate method.  As indicated in Table 2, the reproducibility
of the pH was quite good for replicate extractions by all four pro-
cedures.
                                   211

-------
                                     TABLE 2
                                                          •f
                     Summary of Final pH for Wastes Tested
Waste Sample
ASTM-A
ASTM-B
EP
CAE
Oil Shale



FBC Waste

Bituminous Coal
Fly Ash No. 1

Bituminous Coal
Fly Ash No. 2
9.88 5.09
11.13f 5.32
10. 74**
10.74*"
12.52 11.94
12.54
10.4 4.5


3.28
3. 51^
8.70
8.50


12.28
12.32
5.0
5.0
5.0


6.64
6.58


11.74




3. ll^
3.08^
Bituminous Coal
  Boiler Slag
  3.55
  4.27
           h.22^
Lignitic Coal
  Scrubber Sludge
  5.0
  4.5
5.1
5.43
Hopper Ash
12.13
12.16
12.16
11.03
11.04
11.02
9.44
10.37
10.22
7.30
7.25
7.33
  Unless otherwise noted, the agitation rate for ASTM-A, ASTM-B, and CAE
  was  60 strokes/min.
  Sample leached with bottle lying horizontally on shaker.
  Sanple leached with no agitation.
  TAsitation at  120 strokes/min.
    6                                212

-------
The release of trace metals has been correlated with the pH of the
aqueous extract.  The desorption of trace metals from fly ash sur-
faces decreases with increasing pH.  Additionally, fly ash has been
shown to affect the pH of the aquatic environment.(9)  The change
in pH is a function of iron and/or calcium in the  fly ash.  The
amorphous iron oxides produce an acidic solution while the lime
(Ca(OH)2) yields a basic extract in distilled water.  For many of
the wastes, the final pH of the 1^0 extraction correlates with the
predominance of Ca or Fe oxide.  An exception to this is the bitu-
minous coal fly ash No. 1 which gave a basic pH in distilled water
although the major oxide was iron.  For the FBC waste and hopper
ash, the predominance of the CaO offset the acidic media of the
ASTM-B, EP, and CAE leachates and produced a basic pH.

Some of the initial metals analyses were done by flame atomic
absorption.  This applies to the results for the bituminous coal
fly ash No. 1 and the lignite coal scrubber sludge.  Since many of
the results for these leachates were below the flame AA detection
limits, it was necessary to use the graphite furnace, with its
greater sensitivity, for the AA analyses.  Calcium was determined
by flame AA while mercury was analyzed by the cold vapor technique.
The anions, F , CH , and S0^~ , were determined by ion chromatography.

Typical leachate data for the FBC waste is shown in Table 3.  To
facilitate intermethod comparisons, the solution concentrations for
the leachates were converted to micrograms of contaminant per gram of
dry sample.  The results indicate that the major components in the
                Q j         y -_
leachates are Ca^"1"  and SO^ .  The fluoride and chloride could not be
determined in the ASTM-B and EP leachates because  the acetate ion present
in these leaching media interferes with the 1C analyses.  It appears that
most of the trace metals solubilized during the extractions exist as sul-
fate compounds.  These results are supported by the findings of a previous
study.(7)  The 1C analyses of oil-fired and coal-fired fly ashes  indicate
that the predominant anion in solution was SO, .  Fourier transform infrared
analysis of the water-soluble fractions supported  the assumption that the
soluble metals like nickel, vanadium, and magnesium are sulfate forms.

As an aid to evaluating the leachate data, the number of times each
leachate method gave the highest concentration or highest quantity (in
mass/g of dry sample) of an inorganic contaminant  is tabulated in
Tables 4 and 5, respectively.  The results were compared only for the
four wastes extracted by all four leachate procedures.  These four wastes
were oil shale, FBC waste, lignite coal scrubber sludge, and hopper ash.
When a waste was extracted more than once by a method, the results were
averaged before comparing the data.  In some cases, the results were below
the detection limits of the analytical technique, and an intercomparison
of the leachate  tests was not made.  This omission of some data sets is
reflected in the number of comparisons made for each inorganic species
(the maximum number of tests for each species would be 4).   No comparisons
were made for fluoride and chloride because these anions could not be
analyzed in the leachates generated by the ASTM-B and EP methods.
                                   213

-------
                          TABLE 3
                                          JL
     Comparison of Leachate Data (in yg/g)  for FBC Waste
Species
Ca (mg/g)
Ag
As
Ba
Cd
Cr
Hg
Pb
Se
F~ (mg/g)
Cl~ (mg/g)
SO^2" (mg/g)
ASTM
3.8
<0.002
0.15
<1.6
0.002
0.25
<0.004
0.036
0.16
0.005
0.02
3.8
-A}
4
<0.002
0.14
<1.6
0.002
0.12
<0.004
0.048
0.16
0.005
0.03
4.9
ASTM-B
10.8
0.002
0.25
0.22
0.068
0.44
<0.004
<0.004
<0.02
t
t
2.0
EP}
56
<0.01
0.17
7.2
0.004
0.284
<0.02
0.16
0.24
t
t
17
56
<0.01
0.13
7.7
0.018
0.30
0.02
0.21
0.24
t
t
25.6
CAE
4.3
0.010
0.40
1.1
<0.002
0.27
<0.016
0.093
0.45
0.010
0.048
2.6
 Unless otherwise indicated, concentration given in yg/g of
 dry waste.

 Acetate interference.

rLeached in duplicate.
                            214

-------
                                  TABLE 4


            Number of Times Each Leachate Test Gave the Highest
                 Concentration of an Inorganic Contaminant
        (Only for Wastes Extracted by All Four Leachate Methods)
                                                           Total
Contaminant    ASTM-A     ASTM-B     EP        CAE         Comparisons

Ca                          31                          4

Ag               1                              12
As               1          2                   11

Ba               1                                              1
Cd                          3                                   3

Cr                          3                                   3

Hg                          12                          3
Pb               1                   2                          3

Se               1          1                   24


SCv~             22                                   4
  4
Totals           7         15        5          4              31*
% of            23         48       16         13
Total
Comparisons
 For some leachates, the concentration is below the detection limit and
 no comparison is made.
                                    215

-------
                                  TABLE 5
            Number of Times Each Leachate Test Gave the Largest
          Quantity (mass/g of Sample) of an Inorganic Contaminant
          (Only for Wastes Extracted by All Four Leachate Methods)
                                                           Total
Contaminant    ASTM-A     ASTM-B    EP         CAE         Comparisons

Ca                                   44
Ag                                              22
As                                   134
Ba                                   213
Cd                          3                                   3
Cr                          12                          3
Hg                                   2                          2
Pb                                   33
Se                                              44
Totals           0          4       17         11              32*
% of
Total            0         13       53         34
Comparisons

 Comparisons are not made when "less than" values are reported for each
 leachate method.
                                    216

-------
It is evident from Table 4 that for most of the inorganic consti-
tuents in the leachates, the ASTM methods gave the highest
concentration.  The ASTM methods account for the highest leachate
concentration 71% of the time.  This fact can be attributed to the
low liquid to solid ratio (4:1), and the large amount of material
(350 g) used for the ASTM methods.  The ASTM-B method yielded the
highest leachate concentrations for many of the trace metals, and
probably indicates the ability of the acetate buffer solution to
effectively solubilize these cations.

The leachate concentration produced by the ASTM methods gives an
estimation of the maximum concentration similar to Ham's Procedure
C, using his synthetic leachate. (6)  With Ham's Procedure C, the
extraction was done three times, with the extracted waste being
replaced with fresh material after each elution.  The same leaching
solution was used throughout the three tests.

Based on the results in Table 5, some leachate procedures exhibit an
elemental selectivity.  More cadmium is extracted by the ASTM-B
method than by the other procedures.  This trend is also observed for
the extraction of selenium, arsenic, and silver by the CAE method.
The EP method extracts the largest quantity of materials, especially
the major components, Ca^  and S0^~.  This  is probably due to the
higher liquid to solid ratio and the EP's more aggressive agitation
method.  However, the more vigorous stirring in the EP, could cause
the particles to break up and expose new surfaces to the leaching
medium.  The results, then, might be artificially higher for the EP,
and unrealistic in predicting the environmental impact of the waste
disposal.

It is difficult to explain some of the elemental selectivity indicated
above; more samples need to be tested before being certain that these
trends do exist.  However, some patterns are apparent and are supported
by previous studies.  For example, cadmium is preferentially extracted
by the leachate solution which has the lowest pH.  In most cases, this
is the leachate generated by the ASTM-B method.  Cadmium, which exists
as a cationic species in fly ash, has been shown to be leached more
readily in acidic solutions.(3)

Arsenic and chromium are also known to concentrate on the surface of
fly ash particulate.  The study on trace metal solubility in coal fly
ash (3) demonstrated-that arsenic and chromium could be solubilized
in acidic media, but they were sparingly soluble in ^0.  This informa-
tion is reflected in the data for the fly ash and hopper ash samples.
When the quantity of arsenic and chromium leached by the ASTM-A method
is compared with the other three methods, more arsenic and chromium
have been extracted in the acidic solutions.  When arsenic and
chromium exceeded the RCSA threshold values, it was only in the fly
ash and hopper ash wastes and only for the leaching tests which used
acidic solutions for extraction (Table 6).
                                  217

-------
                TABLE 6
Wastes Classified as Toxic by RCRA Criteria

Waste
Oil Shale


Fluidized Bed
Combustion Waste


Bituminous Coal
Fly Ash No. 1

Bituminous Coal
Fly Ash No. 2


Bituminous Coal
Boiler Slag


Lignitic Coal
Scrubber Sludge


Procedure
ASTM-A
ASTM-B
EP
CAE
ASTM-A
ASTM-B
EP
CAE
ASTM-A
ASTM-B
EP
CAE
ASTM-A
ASTM-B
EP
CAE
ASTM-A
ASTM-B
EP
CAE
ASTM-A
ASTM-B
EP
CAE
Element (s) Exceeding
Threshold Value
Se
*
None
None
None
None
None
None
None
Se
As, Se
As, Se
NR1"
NR
As, Cr, Se
NR
As, Cr
None
None
NR
None
Se
Se
Se
Se
              218

-------
                                    TABLE 6  (Cont)
                Wastes Classified as Toxic by RCRA Criteria
    Hopper Ash                    ASTM-A                            Se



                                  ASTM-B                            Cr, Se



                                  EP                                Se



                                  CAE                               Se
*
 None  -  no elements exceeded the threshold value.

fNR    -  Not Run.
                                       219

-------
Based on the RCRA criteria, five of the energy wastes would be
classified as hazardous by at least one of the leachate methods, and
the findings are summarized in Table 6.  The hazardous leachates were
extracts of oil shale, bituminous coal fly ash, scrubber sludge, and
hopper ash.  Hazardous leachates for the scrubber sludge, and hopper
ash were produced by all four methods.  In most cases, the concentra-
tion of selenium exceeded the maximum acceptable concentration of 10
times the National Interim Primary Drinking Water Standards (0.1 mg/L
for Se).  In the ASTM interlaboratory program to evaluate the EP and
ASTM methods, (1) selenium levels were often in excess of the proposed
EPA limits.

The high solution concentrations cited for the ASTM methods in Table 4
are reflected in the number of hazardous leachates produced by the
ASTM methods.  Four toxic leachates were produced by each of the
ASTM methods, while the EP and CAE methods each yielded 3 hazardous
extracts.

It is also interesting to note that, regardless of the method used for
extraction, selenium often exceeded the RCRA threshold value.  In a
study of the solubility of trace elements in coal fly ash, (3) it was
determined that acidic, neutral, and basic solutions could solubilize
selenium from fly ash.  AIM HNO-j solution was the most efficient for
extracting the selenium, while the 1^0 and NH/OH extracts were compara-
ble in the amounts leached from the fly ash, but were much lower than
the acidic solution.  The anionic character of selenium in the fly
ash could account for its partial solubility in the ELO extraction
(ASTM-A method).  Selenium is probably present as the selenate ion
(SeO^j) which is leached more readily in t^O than a cationic species
such as cadmium.

The percentages of the metals leached from the hopper ash have been
calculated in Table 7.  Analysis of the hopper ash was conducted by
the GCA Analytical Laboratory.  After total digestion of the hopper ash,
the metals were measured by flame AA.  The results are reported in
yg/g, with the exception of calcium, which is listed as percent CaO.
No percentage is reported for results which were below the detection
limits for the AA analyses.

Many of the results indicate that less than 1% of the metal was leached
from the waste.  Chromium was extracted in a greater percentage than the
other metals.  The percentage of chromium extracted is especially high
for the fly ash samples leached by the ASTM-B and CAE methods.  This
reemphasizes the availability of chromium on the surface of fly ash and
its solubility in acidic media.  It appears that most of the other trace
metals may be bound to  the sample matrix in a manner which makes them
unavailable for leaching.  Another possibility is that the compound forms
of the trace metals are not solubilized by the leaching media.
                                     220

-------
                                  TABLE 7
                       Percentage Leached from the Hopper Ash

Concentration
of MetalA
in Waste ASTM-A ASTM-B EP
Ca (% as
CaO) = 22.9 1.2 5.4 27
Ag = - - -
As = 32.9 0.3 0.8 1.2
Ba = 4800 0.2 0.01 0.08
Cd = 6.06 0.4 0.08
Cr = 134 0.4 1.6 3.4
Hg = 0.46
Pb = 97 0.03 - 0.3
Se -


CAE

2.0
-
5.2
0.2
0.1
2.4
-
0.1
-
Results of AA analysis given in yg/g unless otherwise indicated.
                                    221

-------
Precision of Leachate Methods

Another means of evaluating the four leachate procedures is to compare
the precision of replicate extractions.  The relative standard devia-
tion (RSD) has been calculated for the replicate extractions of
hopper ash and is presented in Table 8.  A comparison of RSD indicates
that the precision for the ASTM methods is almost the same, with the
ASTM-A method having slightly better reproducibility.  The Carbonic
Acid Extraction (CAE) is a close third in precision behind the ASTM
methods.  The Extraction Procedure had the worst reproducibility for
the hopper ash extractions.  A comparison of the precision for other
wastes also indicated that the RSD for the EP was below that of the
other methods.

The precision of the methods tested in this study can be compared with
the results of previous investigators.  One study found that the intra-
laboratory precision for the EP was quite good with chromium and lead
having relative standard deviations of less than 5%. (8)  Another
investigation into the interlaboratory precision of the EP demonstrated
that the relative standard error ranged from +9.2% for arsenic in
refinery sludge extract to + 63% for arsenic in fly ash leachate.(2)
For the ASTM program, the leachate results from 18 laboratories indicated
extreme variability for the EP and ASTM methods. (1)  One conclusion
drawn from the ASTM data was that the three leachate tests exhibited no
consistent difference in precision.

A comprehensive study was undertaken by the Electric Power Research
Institute (EPRI) to evaluate the reproducibility of the proposed RCRA
Extraction Procedure. (4)  It is apparent that most of the variation in
the results for the metals in the leachates is due to the analyses and
not extraction variability.  Most of the variation in the results of
arsenic, barium, cadmium, and lead analyses by graphite furnace and
barium and selenium by flame AA could be attributed to inter-lab analytical
variability.  This conclusion would account for the poor precision cited
in the ASTM interlaboratory study.

Compatibility with Analytical Methods

The inherent toxicity of the EP leachate to some bioassay tests has been
documented by a study at Oak Ridge National Laboratory. (5)  Since the
acetate ion was the source of the bioassay problems, the ASTM-B method
would be expected to exhibit the same problems.

The acetic acid matrices of the EP and ASTM-B methods also caused some
analytical problems.  These leachate media caused rapid deterioration of
the graphite tubes.  Standard solutions had to be injected more frequently
to monitor the condition of the graphite tubes, resulting in increased
analytical time and, with more frequent tube replacement, increased
expense.
                                   222

-------
                               TABLE 8
               Calculation of Relative Standard Deviation
                 from Results of Hopper Ash Extractions


ASTM-A
Ca 7.9
Ag
As 5.5
Ba 12
Cd
Cr 4.9
Hg
Pb 4.7
Fe 13
F~ 1.5
CjT 0
SO," 15
Relative Standard Deviation (%)
ASTM-B EP CAE
2.6 0 7
_
8.5 39 14
4.5 11 1.
16 92 29
5.7 42 7.
_ _
14 0
14 18 19
t f 40
t t 4.
17 20 9.





1

8




0
1
Acetete interference
                                 223

-------
It is also possible that the acetic acid matrix caused some of the
variation in the results observed for the EP.   The injection of
strong acidic solutions into the graphite furnace can degrade the
analytical precision.  The strong acidic solutions "wet" the inner
surface of the graphite tube, and cause variable distribution.of
the injected sample.   This variable distribution will reduce the
precision of replicate injections.   However, the reduced analytical
precision for the EP does not sufficiently account for the differences
in precision among the EP and the other three leachate tests.

The acetate ion in the ASTM-B and EP solutions interfered with the
determination of fluoride and chloride in the 1C analyses.  The reten-
tion time of the acetate ion is comparable to the retention times of
the fluoride and chloride ions and its presence in large excess masked
the determination of the fluoride and chloride concentrations.

                       CONCLUSIONS AND RECOMMENDATIONS

When the four leachate procedures are evaluated in terms of the general
applicability, reproducibility, methods compatibility, and leaching
power, the ASTM-A and Carbonic Acid Extraction would be preferred for a
standard leachate test.  In terms of leaching power, the CAE extracts
a greater quantity of inorganic contaminants than ASTM Method A.  How-
ever, the reproducibility of ASTM-A is a little better than the CAE.
No analytical problems were encountered with either method and both
procedures could adequately handle the wastes tested in this program.
The leachates from the CAE would have to be subjected to the bioassay
tests to determine the compatibility of the CAE with the health and
ecological tests.

The major criticism of all the proposed methods is the lack of sufficient
detail in the procedures.  A standard leachate procedure should be
explicitly defined to avoid interpretation by the analyst and, consequently
increase the reliability of interlaboratory results.

The objectives of the leachate test should be of primary concern.  Is the
test attempting to simulate acid rain conditions, or degradation at a
disposal site?  Once the objectives of the test have been established,
then the procedure to achieve these goals can be designed.  The separation
of complex, multi-phase industrial wastes prior to leaching has not been
covered by any of the methods.  A separation scheme similar to that
suggested by Ham (6) could be adopted.  Sample preparation, or whether
to use  the waste in  its disposed form,  is another question which needs
to be addressed.  The portion of the leachate methods most susceptible to
interpretation by the analyst is the agitation method.  If the shaking
apparatus and agitation rate is rigidly defined, much of the variability
in the leachate results would be eliminated.  Finally, the means for
preserving the sample for subsequent analyses must be included in the
standard leachate test.
                                   224

-------
                           ACKNOWLEDGEMENTS

Assistance provided by the Process Measurements of EPA at the Industrial
Environmental Research Laboratory, Research Triangle Park,  North Carolina,
under U.S. EPA Contract No. 68-02-3129, is gratefully acknowledged.

                              REFERENCES

1.      American Society for Testing and Materials, interlaboratory
        evaluation of EP and ASTM methods, report submitted for
        publication (1979).

2.      Burd, R. M. and J. M. Riddle, "Final Report:  Evaluation of
        Solid Waste Extraction Procedures and Various Hazard Identi-
        fication Tests", NUS Corporation, Pittsburgh, Pennsylvania,
        U. S. EPA Contract No. 68-01-4725 (1979).

3.      Dreesen, D. R., K. E. Wangen, E. S. Gladney, and J. W. Owens,
        "Solubility of Trace Elements in Coal Fly Ash", Environmental
        Chemistry and Cycling Processes, proceedings of a symposium
        held at Augusta, Georgia, 1978, pp. 240-252 (1978).

4.      Electric Power Research Institute, "Proposed RCRA Extraction
        Procedure:  Reproducibility and Sensitivity", Electric Power
        Research Institute, Palo Alto, California (1979).

5.      Epler, J. L., W. H. Griest, M. R. Guerin, M. P. Maskarinec,
        D. A. Brown, N. T. Edwards, C. W. Gehrs, B. R. Parkhurst,
        B. M. Ross-Todd, D. S. Shriner, H. W. Wilson, F. W. Larimer,
        and T. K. Ra0r "Toxicity of Leachates", Interim Progress Report,
        April 1, 1978 to January 1, 1979, Oak Ridge National Laboratory,
        Oak Ridge, Tennessee (1979).

6.      Ham, R. K., M. A. Anderson, R. Stegmann, and R. Stanforth,
        "Comparison of Three Waste Leaching Tests", U. S. EPA Report
        No. 600/2-79-071 (1979).

7.      Henry, W. M. "Methods for Analyzing Inorganic Compounds in
        Particles Emitted from Stationary Sources", Interim Report,
        U. S. EPA Report No. 600/7-79-206 (1979).

8.      Meier, E. P., L. R. Williams, R. G. Seals, L. E. Holboke,
        and D. C. Hemphill, "Evaluation of the Procedures for
        Identification of Hazardous Waste", Interim Report - August
        1979, EPA Environmental Monitoring Systems Lab, Las Vegas,
        Nevada (1979).

9.      Theis, T. L., and J. L. Wirth, "Sorptive Behavior of Trace
        Metals on Fly Ash in Aqueous Systems", Environmental Science and
        Technology, 11_, 1096-1100 (1977).
                                    225

-------
             POSSIBLE EFFECTS OF COLLECTION METHODS AND SAMPLE
             PREPARATION ON LEVEL 1 HEALTH EFFECTS TESTING OF
                             COMPLEX MIXTURES

                              By D.J. Brusick

          Director of The Department of Genetics and Cell Biology
                   Litton Bionetics, Inc.  Kensington, MD
                                 ABSTRACT

Level 1 Environmental Assessment is oriented toward the development of a
data base that will permit a relative ranking of industrial streams with
respect to their potential biohazard.  The combination of both chemical
and biological evaluation of collected samples in the assessment increases
the level of data integration and coordination, but reduces the chance that
a potentially hazardous stream will go undetected.  During the initial phases
of the IERL Environmental Assessment Program, considerable time was given
to methods of sample collection, preparation and analysis for chemistry
assessment (IERL-RTP Procedures Manual:  Level 1 Environmental Assessment
(Second Edition) EPA-600/7-78-201,  October, 1978).  The recent introduction
of bioassays requires a similar appraisal of the functions of sample col-
lection, storage and pretest handling as they relate to the specific health
effects and ecological tests proposed for Level 1 biological assessment.

Because of the need to rank streams according to potential biohazard, the
initial approach taken by IERL was  to evaluate samples in the specific
bioassays in a state as similar as  possible to that found at the time of
sampling.  However, recent results  reported in the scientific literature
on analysis of complex environmental mixtures have shown that pretest
processing of samples (concentration of liquids, extraction of particular)
often results in enhanced biological activity (6,7,9,10,11).

The purpose of this discussion is to review several of the sampling and
pretest procedures and to illustrate how the application of these techni-
ques to Level 1 Environmental Assessment will affect the test responses
and ultimately the goals of this program.

                 GOALS OF LEVEL 1 ENVIRONMENTAL ASSESSMENT

The fundamental goals of IERL Level  1 Environmental Assessment are to reli-
ably detect potentially hazardous emissions from stationery source sites
and to rank the streams in order of  their need for control technology appli-
cation (Figure 1).  These goals must be attained in a cost-effective manner
so that many different emission sources can be sampled and evaluated and
repetitive samplings from the same source can be made.

The basic approach proposed for Environmental Assessment has been an evalu-
ation of emissions for the following five parameters:

                                    226

-------
                      TYPICAL STEPS INVOLVED IN ENVIRONMENT ASSESSMENT
                                 Release
                                                         Environment
                           Rate of Release Measured
                           and Sampling Performed
                            Sample Transported
                            to Sites of Analysis
                             Preanalysis Handling
                               •  Concentration
                               •  Extraction
                  Chemical Analysis
Biological Analysis
                               Coordinated
                         Potential for Environmental
                             Effects Oetermlned
FIGURE  1    An  overview of  the  steps  involved  in Level  1
               Environmental  Assessment.
                                          227

-------
     1.    The rate of release of the emission into  the  environment.

     2.    The distribution between physical states  of the released emission
          products in the environment.

     3.    Chemical composition of the emission.

     4.    Detection of potential impact of the emission on specific health
          effects.

     5.    Detection of potential impact of the emission on the ecosystem.

This document will focus on parameters 4 and 5 with less emphasis placed
on items 2 and 3.

The initial Level 1 methods manual for biological testing contained protocols
for five health effects tests and eight biological tests.  The types of
data generated from this array of tests were quite different and not amenable
to interpretation by non-biological scientists.   For their optimum use, a
method of developing uniform data sets was needed.   A review of Level 1
bioassays and data generated from several pilot studies resulted in recom-
mendations concerning uniform methods for data analysis and formatting (8).
It was also recognized in this review process that elimination and/or substi-
tution of test procedures would benefit the Level 1 assessment.  Table 1
identifies the current status of recommendations for Level 1 bioassays;
however, there are no current procedures approved for E5.  Test substitutions
or protocol modifications are under consideration for some of the other
tests as well.  These will be given in more detail in the second edition
of the Level 1 Manual for biological testing.

A summary of three pilot studies using a battery of Level 1 bioassays
employing this data-formatting system have been reported (8).  Employing
the recently proposed data-formulating procedures, the results from the
pilot studies are reported as high  (H), moderate (M), low (L), or nondetec-
table (ND) in a summary table (Figure 2) according to the scheme given in
Table 2.  One feature of the pilot  studies was that the  results from the
studies on Coal Gasification Fluidized Bed Combustion and Textile Plan
Liquid Effluents were evaluated with very little pretest sample handling
or sample history associated with the samples used in the bioassays.  For
example, water samples were not concentrated, leachates were not concen-
trated and particulate samples were not extracted with organic solvents.
Some filter or sorbent collectors were extracted with solvents or sonicated
with solvents, but in large part, most of the samples were placed in the
bioassays as received by the testing facilities.

Comparative analysis of the  chemistry and biology of these pilot programs
has been shown to be generally complementary between the severity assessment
based on chemicals detected  and biological responses (1).  In a few instances
toxicity was not predicted by the chemical analysis.

What effect pretest processing might have on such data is important since
the Level 1 procedures are being used more frequently.   Should samples be


                                    228

-------


U 0
3 £
S E
f £ I s
« u ^ -*-
£  J3 •*- t—
«T3 O) *J Z
< */> GJ CTi (/I •*— IQ
•*- c ^- »— c o ^-
I. U, — < (Q i-*  &.
OJ •—
to UJ UJ UJ L. UJ UJ UJ
QJ i-


(/) IQ
< w .w
<: •*-
in u
0> -i- O
CO H-
QJ -f—
CT X -M
fD O C
3 -0
E ^- O
> CC.

•
4-
4-
4-
4-
4-
+

tn
3
CT

4-
1
4-
•
'
I
1
I
1
I


0)
CD

I
I
I
4-
1
4>
4-
1
4-
4-
4-
4-
l/)
L.
3
•o
C
ra
•o
cr
"o
OJ
3
X

•
1
4-
I
•
'
I
4-
4-
4-
(tracts
UJ
1
o
V
u_

'
•
1
4-
'
1
1
1
4-
I
•
4-
) Chronic Toxicity
> Applicable tc
Result!
a/
t—
CM
1
1
4-
'
I
I
•
i
•
'
tn
C

1
4-
1
'
*
'
•
'
0)
*J
t/1
in Developmenl
Method
un



ra
i-
0)
i
o
£
3
'o
in
O
in
C
s.
?

t/1
i
o
fO

-------
' LU
i ,a
OKJ
< UJ
i/l t-
,O LU
OS (—
CTO


po

H*
,"«
'h-
,Z3
,UJ!O

ir
£ 'U.L
— a"


i

4-J
Ul
1

g 0
*— ^O CM
A A A

Q O O





,— O
T ^?
o o
*A
o S
5 o
o -o



3 C3
o *o
*
§o
CM



= = =
i" 1 1





O
Lf)


>>
^
j=

41
H-
-,I
-v. C
i'§

o
^
«s
a





a



^
d



_
o

a




-^^
1





o
3

>>
^_
5

ai

^
LU i—>
11

§ S S


a o a a o a
x z zz zz




— a *-o •— o
r— 1 r— 1 f— 1
-ai .in -ui

T T r
oo oo oo
•CM -CM -C\J
O 0 O



o o o
o o o o o o
v CM v CNJ v CM

'1 "| "I

t. s_ u
0) >

3d •*"•
a c o
1-4)  r-
c — *-^ o»
•^ LU r~ 4)
% £ S ^
/- Q. I
g. . s*. ^
o a •*- o 01
a. z +•> ^3
t*. M U 41 O
CO tJ •<-• 3 VI
O C *— Z
•^ ** C 4J fl S
 C •/»
3 C O) CJ LU 41
C O "^ w» 01 ^C
-2°-tI ^51 =
Wt*. 41 >.*-> t, 3
^ O <*- *J U 4- ~
.— " M. .^-01 =
— '3MUJ U1*-!* X
i/l dO T- *^ "O  O 01

^^OI*Q "— O •*-» ^
3^0^ MCC 1
c ^ a. o» * -^ -21 t/l
U LU X O T) 4t 9
4» "O UJ <— "O C
h- 4) C 11 3 V
-— -4J O C -M 
c +* S i» Z c
4) 41 41 C i- 41
»— OJUOt if) — - >
'Ot/iocoi-Mcr ^
(J O (J O f— H3 4) i/l
•— Q c_> -a i- .
— -o M i/> 41  c =
> II 11 II *J •*- 0> 3
U O LO O o»i— fg 'o
^Isaaifll T
230

-------
                           BIOASSAY SUWAftr
Technical Directive or Protect No..

Contract No	
                                             Freth Water    ManfM
                                                    I	I
                           I    i
                                      I    i
                        I   I
NO « No Detectable Toiiaty
 L • Low Toxlcity
 M . Moderate Toxldty
 H * High Toxicity
 FIGURE 2   Data from all tests  of one stream can  be  translated
             into toxicity definitions and  compared.
                                   231

-------
used directly from the sampling process or should an effort be made to reduce
each sample to its active constituents or to concentrate dilute samples?

In addition to pretest processing, other aspects of Level 1 assessment have
been identified as important components in the final data analysis and inter-
pretation.  These aspects are listed below:

     I.   Sampling Methods -- Selection of representative sample.

     2.   Storage -- Important in maintaining proper composition and
                     prevention of degradation.

     3.   Shipping (Transport) -- Same as 2.

     4.   Pretest Handling -- Alter chemical composition, physical state
                              or preferential extraction and concentration
                              of potential toxicants and mutagens.

Items 1, 2 and 3 can be grouped into a single parameter called sample history
(i.e., documentation of all handling of the samples up to receipt at the
testing laboratory).  Item 4 encompasses several pretest sample processing
techniques carried out at the testing facilities (see Table 3).

Another factor not listed above but of real significance to the final
assessment is one of environmental fate of the released emission and in
particular, bioaccumulation.  Significant bioaccumulation of weak or moder-
ately toxic substances in emissions might have considerable environmental
impact.  These considerations are examined in detail in the following
sections.

                              SAMPLE HISTORY

The reasons for emphasis on sample history relate to the final interpre-
tations of the test results and the ranking of emission sources with respect
to potential hazard.

The emission rate profoundly influences environmental assessment since  large
volume emissions  (e.g. fly ash from power plants) provide an opportunity
for significant human and ecosystem exposure  (3,7).  An emission with ex-
tremely low volume and rate of release may represent a negligible health
or ecological hazard even if it is highly toxic.

For example, if one identifies a potential hazard* as a function of release
rate  (volume/unit time), its toxicity and its environmental fate, a quanti-
fied  assessment may be possible.  Thus, if any single component is very
large, the potential hazard could be significant.  Attempts should be made
to document the history of each sample put into  the Level 1 bioassay program.
*    Potential hazard  is not synonymous with risk assessment.  The types
     of bioassays used in Level  1 are able to detect potential but are not
     capable  of  estimating  risk  either qualitatively or quantitatively.

                                     232

-------
oo
«
C;
QJ .


3 QJ «J -C
E • T3 0 U
0) C •*- Wl
•Q 4J 3 CT*
c re O 01
en QJ O •«-
O ui ••- .^
^— "^ re *ft c
x c oi =s re
o 3 — -c i-
*"* JS t. 3
**" <= 4-> O.2:

•2 g « "> ui
4-» E Cn C C
C i- O) O

01 C QJ re-«-
O- O 4J C U


c
o
f~


U QJ

C *—
o o
Ul 

+•* u
c •»-
0) C
> re

O S-
ui o

U 4J
•f- QJ
C r—

C" X

O t/1











c
0


















 (J OJ t/> C_> O 3
s» -^ t. re en 3 •*•• ^
OCQJOJC-OEO
•f~ O) QJ U JK: t- £
s- i_ c c •*-> (j o;
yi O o.-^- fO C -M
•M C S- — 3
"s "" c f z >>^-*^
QJ ^ re •»•- ^c E u
o-*J en E vi •£ c

co u c re f ^
re - •«- re u o QJ
L>ocencOJC4->
"c S o en u"fc 3
re ••- re QJ o
S-.COO3O— C
o CT>+J E re oi
3 C C
VI O -«J O >*• O >>4J
*j .c re 4J *-> c
re *J 4V re c re 
£ £
>
( O
Q I/I

X T3

.c re
en
3 C
o ••-
l_ -Q l/l





•— o t-
U. <+- M-










c
0


re
s_

c
QJ
(J
C
0
CJ

•a
>f~
cr

u
C O QJ <
QJ ^ l^ *J y> tt
CT 01 o re
m -M -D oj c
3 •*- O 3 QJ
S**-^- 0 S- (J
4J ^ •^«
•o o re >> x
c *> s- c re o
re QJ QJ E 4->
•0 4^ £
O QJ i— S U QJ
*X (J 1- 4J
o re -o 3 F— -CJ
4J re u **- 3
H- re o o
°x§o'S3
QJr-— 4J 4J
re re re c o -c
QJ •£ -O S -f- 4J
QJ c t. c re 
^» 0) cn^ u QJ
^ QJ i— ^
r- O> O Wl C tj
re **- -^- o H-
•»- QJ wn 4^
4-» i. • if) U
c o.*— r- re
QJ re re • a.


•*- c QJ E 4->«— re
u en re -C f- s */>
^. re E u *t- m re


c
QJ





U
C
re
^
0

c


c
o


re



c
on


QJ
£
c
o

4-"
u
QJ


O




i.


c
o


u
re
^j
x
UJ
0 i **-
^§c

C QJ •*-
3 > 4J

E re i.
u c x
QJ QJ
0) 1/1
3 O.
»— Q) Q}
O 1- ^—
:n
O vi
01 • i—
wi E re
re wi e
*E i>?
J- O
re l*-
*-» QJ
C en c «/>
Q) i- O O
i. ,0 ^.^_

*S **ii S
0) QJ v» en
w j= o :-
ex 4-> a. o
o
O) LO
Jj°
*>» *
^ OJ




o o>

r— • QJ
Q.C
e 0
re 4^
(/) QJ
u
*+» re
o

C OJ
O "d


-^ o •
-a r- o
•O -C 4J
<: u  a.

o >,
s^t,>
-^- 1/1 C 4-»
>)4-> J^ U
a.  &. Ul
c re -Or-
QJ r- ^> 0»
i 3 3 >
Q.^ O QJ
O ^- U ^
r- QJ
at u vi x:
0) O -C •*-
0 4J |— JZ

*—
"i
f"




cn
c
O *N
^- -^- C
4-> t/» QJ
U OJ
QJ *O W
^ CO
»— re u>
o
(_j CTI QJ
C S-

QJ "OS
4^ C

£ 'u c
u. o re






E:
J

o

•o
c
re

cn
c

ISJ



QJ

re

3
(J

i_
re
a.
                                  233

-------
Currently only the toxicity factor is employed in the Bioassay Summary Table
(Figure 2).   This factor is not tied back to the type of emission release
rate or to the volume of emission sampled.

                         PRETEST SAMPLE PROCESSING

Several laboratories in the U.S. (particularly at the EPA IERL in RTP and
the Oak Ridge National Laboratory) have initiated programs to analyze complex
mixtures for mutagenic activity (2,4).  The Ames Salmonella assay is typically
employed in these programs and is conducted on fractions of the complex
mixtures as a means to direct further fractionation tests in the pursuit
of biologically active components (pure substances).

The general procedures are to:  (a) collect the desired sample (particulate,
gas or liquid) on a filter or solid sorbent, (b) extract the organics from
the filter or sorbent by sonication or Soxhlet methods into an organic
solvent, (c) exchange solvents (to DMSO) or evaporate to dryness and resus-
pend in DMSO and (d) conduct the bioassay on the concentrated extract.
Chemical fractionation and further testing may eventually lead to specific
associations between chemicals or chemical classes with biological activity
(4).  Table 4 describes some of the pretest sample processing techniques
currently used and demonstrates how these methods could affect the interpre-
tation of Level 1 bioassay data using conventional data evaluation techni-
ques .

Level 1 toxicity assessments listed in the Bioassay Summary Table (Figure 2)
include several types of biological systems and phylogenetic levels.  If a
significant amount of pretest processing is anticipated, the resultant
extracts and concentrates should be evaluated in all the Level 1 tests and
not just the Ames or one or two selected tests.  If unprocessed samples
are used in the remaining tests on the site emission, the balance of test
results may be altered.

Pretest processing should be included in evaluation of all Level 1 test
results, otherwise the data balance will be upset and the ranking of the
test site might be biased by an abnormally toxic response obtained from an
extract or concentrate in a single assay.  This could erroneously exaggerate
the potential hazard.

The previous precaution is not meant to preclude pretest processing, but
as with the sample history factor (release rate) described previously, the
pretest sample processing must in some way be factored into the final assign-
ment of a toxicity value and its contribution to the potential hazard.
The mechanism to accomplish this need appears to be rather complex.  In
the simplest approach, one might want to divide the actual test response
by a concentration factor to normalize the effect back to the pre-processing
state.

                              RECOMMENDATIONS

1.   Documentation of sample collection, storage, shipping and pretest
     processing should be available for all test samples.  Examples of
     forms for this purpose are shown in Figures 3 and 4.

                                     234

-------
                                        Sample Collection Form No.

                      LEVEL 1 SAMPLE COLLECTION FORM

I.    SAMPLE HISTORY
     A.   Company Name 	
     B.   Sampling Manager
     C.   Contract No. 	
     0.   Sampling Date 	
     E.   Emission Source	
     F.   Approximate Rate of Emission (Volume/Time)
    G.   Proportion of Emission Sampled (Volume Captured)

II.  SAMPLE TYPE

    A.   Sample No. 	
    B.   Name
     C.   Sample Description
III.  HANDLING CONDITIONS

     A.   Storage

             Container	           Temperature	        Light

       D Amber Glass Bottle      D Ambient                D Keep in Dark
       D Polyethylene Bottle     D Refrigerate (0 to 4°C)
       D Coated Bag or Bottle    D Freeze (-20°C)

     8.   Approximate Time of Storage Before Shipping 	

IV.   SHIPPING HISTORY
    A.   Biological Contractor
    B.   Address:
     C.   Carrier	
     D.   Date  Shipped	  By
     E.   Special Packaging	
     Two  copies of this form must accompany each  sample.
    FIGURE  3   A possible  form on which  sample  history  can
                  be  collected to establish the conditions of
                  collection,  storage,  and  shipping.
                                235

-------
                                         Sample Processing Form No.

                      LEVEL 1 SAMPLE PROCESSING FORM
I.    SAMPLE IDENTIFICATION
    A.
    B.
    C.
    D.
Sample Collection  Form No.
Contract No. 	
Project Officer 	
Sample No. 	
II.   SAMPLE TYPE AND PROCESSING  REQUIREMENTS
     Basic Type

  D Solid
  D Liquids
                   Subtype
         O Solid Granular
         D Slurry (>50% Solids)
         D Participates from
           Filter
         Q Filter/Unit Participates
         D Suspensions (<50% Solids)
         D Effluent
         D Leachate
         D Extract
         O Condensate
         Processing
D  Grind to <5p size
D  Extract Participates with
   Organic Solvent
D  Remove Particulate
   from Fi Her
D  Prepare Water Leachate

n  Concentrate with XAO-2
D  Solvent Exchange
D  Evaporate to Oryness
  D Gas
         D Pressure Collection

         D Vacuum Collection
III.  BIOASSAYS REQUESTED

   D Ames, Salmonella
   D SAM Toxictty
   D CHO Clonal Toxicity
   D Rodent Quantal  Toxicity
                              D Freshwater Fish Toxicity
                              Q Freshwater Invertebrate
                              D Algal Test
                              O Insect Toxicity
                              D Plant Test
                              D Soil Test
     FIGURE  4   A possible  form  for use  in  describing the
                   pretest  processing  and types  of  bioassays
                   to  be  used  with  a sample.
                                   236

-------
2.   Pretest processing should be factored into the final determination
     for toxicity designations of H, M, L and ND.   Specific methods need
     to be developed to normalize data obtained from samples modified prior
     to evaluation.

3.   Emissions from a given site should be applied to the spectrum of Level 1
     bioassays uniformly and not in a manner likely to bias the final inter-
     pretation.  Data related to environmental fate should be included.

4.   Discharge severity (DS) calculations presently used to factor chemical
     and physical information should be expanded to include biological
     response and fate.  Specific methods need to be developed.

                            SUMMARY/CONCLUSIONS

Level 1 Environmental Assessment Bioassays should permit an accurate ranking
of emissions from stationary site sources with respect to their potential
hazard.  Moreover, the ranking must ensure that the potential hazard is
likely to be derived from the emissions as released into the environment
and that pretest processing should be kept to a minimum and applied uni-
formily across all Level 1 bioassays if performed.

Level 1 Assessment should include the environmental fate and rate of release
of the emission along with the chemical and bioassay toxicity determinations.
This approach would result in a second level determination called a severity
potential hazard.  A general scheme proposed to develop this assessment is
given in Figure 5.  If a potential hazard can be calculated with reasonable
accuracy the utility of Level 1 results will be greatly enhanced.
                                    237

-------
0

-------
                                REFERENCES

 1.   Biological Screening of Complex Samples  from Industrial/Energy Processes.
     EPA-600/8-79-021,  August,  1979.

 2.   Epler,  J.L.  Mutagenicity testing of energy-related compounds.   In,
     Energy  and Health, N.E. Breslow and A.S.  Whittemore,  eds.  Proceedings
     of a Conference Sponsored by Siam Institute for Mathematics  and Sociology,
     Siam, Philadelphia, pp. 17-36,  1979.

 3.   Fisher, G.L.,  C.E. Chrisp and O.G.  Raabe.   Physical factors  affecting
     the mutagenicity of fly ash from a coal-fired power plant.   Science,
     204:879-881, 1979.

 4.   Huisingh,  J.,  R.  Bradow, R.  Jungers, L.  Claxton,  R. Zweidinger, S.
     Tejada, J. Bumgarner, F. Duffield,  M.  Waters, V.F. Simmon,  C.  Hare,
     C. Rodriguez and L. Snow.   Application of bioassay to the  characteri-
     zation  of  diesel particulate emissions.   In, Application of  Short-Term
     Bioassays  and the Fractionation and Analysis of Complex Environmental
     Mixtures,  M.D.  Waters, S.  Neonow, J.L.  Huisingh,  S.S.  Sandhu,  L.  Claxton,
     eds. EPA-600/ 9-78-027, September,  1978.

 5.   IERL-RTP Procedures Manual:   Level 1 Environmental Assessment  Biological
     Tests.   EPA-600/7-77-043,  April, 1977.

 6.   Klekowski, E.  and D.E. Levin.  Mutagens  in a river heavily polluted
     with paper recycling wastes:  results  from field and laboratory mutagen
     assays. Environmental Mutagenesis, 1^:209-219, 1979.

 7.   Kubitschek,  H.E.  and L. Venta.   Mutagenicity of coal fly ash from
     electric power plant precipitators. Environmental Mutagenesis, 1^:79-82,
     1979.

 8.   Level 1 Bioassay Assessment and Data Formatting.   Final Report from
     Litton  Bionetics,  Inc. under EPA contract 68-02-2681.

 9.   Lofroth, G.   Mutagenicity assay of combustion emissions.   Chemosphere,
     10:791-798,  1978.

10.   Pitts,  J.N.  Jr.,  D. Grosjean, T.M.  Mischke, V. Simmon,  and D.  Poole.
     Mutagenic  activity of airborne  particulate organic pollutants.   Toxi-
     cology  Letters, 3.-.65-70, 1977.

11.   Teraniski, K.,  K.  Hamada and H.  Watanabe.   Mutagenicity in Salmonella
     typhimurium mutants of the benzene-soluble organic matter  derived
     from air-borne particulate matter and  its five fractions.  Mutation
     Res., 56:273-280,  1978.
                                    239

-------
  THE DYNAMIC INTERACTION BETWEEN VAPOR
     PHASE AND PARTICULATE MATERIALS

            D. F. S. Natusch
        Colorado State University
Manuscript not available for publication.
                    240

-------
   COAL FLY ASH AS A MODEL COMPLEX MLXTURE FOR SHORT-TERM BIOASSAY

                             G.  L.  Fisher
                    Battelle Columbus Laboratories

                                  and
                     C.  E. Chrisp and F.  D. Wilson
                   University of California at Davis
                               ABSTRACT

National energy policy has designated coal as the major energy source
for production of electricity throughout the remainder of this century.
Associated with the combustion of coal is the generation of vast quan-
tities of primary mineral matter, fly ash.  This report describes the
application of bioassays including bacterial mutagenesis, mammalian
lung cell function and mammalian hematopoietic cell maturation to the
study of the biologically significant physical and chemical properties
of coal fly ash.   The effects of sample collection, particle size,
solvent extraction, temperature treatment and irradiation are discussed.
                                 241

-------
                             INTRODUCTION

Combustion of coal for electric power generation has increased markedly
throughout the past decade and is expected to continue to increase
throughout the remainder of this century.  Associated with coal combus-
tion is the production of a variety of biologically active, inorganic
and organic compounds.  Of major concern is the release of oxides of
carbon, nitrogen, and sulfur, biologically active trace elements, sili-
ceous primary particulate matter and organic compounds.  Coal fly ash
represents the bulk of the primary particulate matter produced during
coal combustion.  Interaction of the fly ash particles with inorganic
and organic compounds formed during the combustion process results in
the formation of a unique complex mixture that may serve as a useful
model for other such mixtures resulting from interaction of a relatively
inert carrier particle with biologically active metals and organic com-
pounds.  In a similar manner, the bulk of atmospheric particulate matter
may serve as a matrix for subsequent interaction of airborne vapors and
condensable gases.

                           SAMPLE COLLECTION

The fly ash samples described in the following studies were collected
from a western U.S. power plant burning low sulfur (0.5%), high ash (20%)
coal (29).  Unless otherwise indicated, the samples were collected and
size fractionated in situ downstream of the plant's electrostatic pre-
cipitator (ESP).  On occasion, studies of ESP-collected ash were also
performed.  Four fractions of stack-collected material were obtained
with volume median diameters (VMD) of 20, 6.3, 3.2 and 2.2 urn, respec-
tively, and geometric standard deviations of approximately 1.8.  Size
classification was performed utilizing a specially designed thermostat-
ed (95°C) system containing two cyclones in series followed by centri-
peter with 25 parallel jets  (29).  Cyclone-separated material was de-
posited in a collection hopper while centripeter particles were collected
on fabric filters and removed for hopper deposition by cleaning with
reverse air jets operating at one-minute intervals.  Thus, in contrast
to standard filter collection techniques, collected fly ash samples
were not continually exposed to the reactive gases in the flue stream.
Such an exposure may lead to changes in the chemical composition and
biological activity of collected polynuclear aromatic hydrocarbons (4).
                                 242

-------
It should also be pointed out that our samples were collected continu-
ously over a 30-day period.  Thus, alterations in coal composition or
combustion conditions and other parameters of plant operation that may
affect chemical composition should be reflected in these samples.  In
this regard, Kubitschek and Kirchner (28) have demonstrated the impor-
tant effects on mutagenic activity of combustion conditions during
start-up and shut-down of a bench-model, fluidized-bed combustion system.

                 PHYSICAL AND CHEMICAL CHARACTERIZATION

Microscopic Studies

In previous studies (8,10,16), we have described the physical and mor-
phological properties of coal fly ash and generally find that coal fly
ash is an extremely heterogeneous, complex mixture with a variety of
morphological forms.  Variation in morphology generally represents both
matrix composition and exposure conditions during combustion.  Upon
heating, aluminosilicate inclusions in the coal initially become rounded,
and, through degassing, become vesicular (8).  Further heating will
result in sphere formation, either solid sphere, hollow sphere (ceno-
sphere), or sphere-within-sphere structures  (plerosphere).  Crystal for-
mation is a somewhat later event in morphogenesis with internal crystals
or quench crystals forming rapidly during the phase transition from
liquid to solid (10).  Such quench crystals have been identified as
mullite by Gibbon (20).  While internal (quench) crystal formation
occurs in a time frame of milliseconds, surface crystal formation appears
to be a much slower process taking days or months.  Such surface crystal
formation appears to be the result of sulfuric acid interaction with
metals found on fly ash surfaces.

Previous efforts on the analysis of individual fly ash particles demon-
strated extreme matrix heterogeneity between morphologically similar
particles (35).  We have more recently completed a detailed comparison
of light-microscopic morphology and individual-particle, elemental com-
position using SEM X-ray analysis.  The utilization of comparative,
light and electron microscopy provides a powerful tool for the assess-
ment of physical and chemical properties of coal fly ash (15).  Our
findings have demonstrated the presence of relatively pure mineral
phases although, certainly, the bulk of the ash is amorphous and appears
to have the composition of the clay minerals generally found to be asso-
ciated with coal.  Pure alpha-quartz, calcium phosphate, titanium diox-
ide, calcium oxide, iron oxide, and alumina have been observed (25).
The degree of pigmentation of spherical particles has clearly been shown
to be the result of varying iron content within the particles.  Statis-
tical cluster analysis has confirmed the utility of the light microscopic
morphological classifications (15).  In particular, we have shown that
the morphological classifications defined by light microscopy generally
possess a relatively high degree of homogeneity with regard to elemental
compositions of individual particles within the specific morphological
                                  243

-------
classes.  It should be pointed out, however, that the most morphologi-
cally amorphous particles tended to have the greatest diversity in ele-
mental composition.

We have also recently completed studies on the crystalline phases within
coal fly ash using X-ray diffraction analysis (24).  These studies dem-
onstrated that the predominant crystalline species in coal fly ash are
alpha-quartz, mullite, and magnetic iron oxides.  The highest concentra-
tion of crystalline material was found in the coarsest fly ash fractions
with quartz concentrations of 4 percent, mullite of 8 percent, and iron
oxide of 0.6 percent.  The concentrations of crystalline minerals de-
creased with decreasing particle size such that alpha-quartz in the
finest fly ash fraction was 1.3 percent, mullite 4 percent, and magnetic
oxides 0.03 percent.  It is hypothesized that the silica is the result
of quartz intrusions within the coal itself, whereas mullite and magne-
tite are formed during the combustion and cooling processes associated
with utilization of the coal.  It should be pointed out that the mullite
is generally associated with a quench crystal phase that occurs during
rapid cooling, and hence, is generally encapsulated within the alumino-
silicate matrix of the clay mineral.  This does not appear to be the
case for alpha-quartz or magnetic iron oxides.

As described in our earlier work, surface crystal formation may take
place (10,16).  This is apparently the result of chemical interaction or
formation of sulfuric acid on fly ash surfaces.  Leaching of minerals
and heavy metals from within the fly ash by the surface associated sul-
furic acid may result in crystal formation.  Electron microprobe analy-
sis of larger crystals found in fly ash has identified only calcium and
sulfur to be present; hence, the majority of crystals appear to be cal-
cium sulfite, either present as gypsum or anhydrite.  We postulate,
however, that the mechanism of sulfate crystal formation may also pro-
vide for the increased biological availability of refractory metal
oxides by conversion to the more soluble metal sulfates.

As described by many other investigators, we have found that the concen-
tration of the volatile (at coal combustion temperatures) trace elements
or their oxides are highest in the finest fly ash particles (7,34).  In
this regard, cadmium, zinc, selenium, arsenic, antimony, tungsten,
molybdenum, gallium, lead, vanadium, fluorine, and sulfur are most
enriched in the finest fly ash particles.  However, one also finds that
the relatively refractory elements, uranium, chromium, barium, copper,
beryllium, and manganese are enhanced in the finest fly ash fractions.
While it appears that enhancement of volatile trace elements in fine fly
ash particles is due to vapor phase condensation as described by Natusch
et al.  (33), other mechanisms may also be important in this phenomenon.
Filtration studies with neutron-activated coal fly ash indicate that
some elements are highly enriched in particles in the size range from
0.2 to 0.4 micrometers (17).  In particular, the elements antimony,
arsenic, tungsten, uranium, and chromium have a disproportionately high
concentration in particles less than 0.4 micrometers.  It has been
hypothesized that  the enhancement  in submicron particles is the result
of homogeneous nucleation and subsequent coagulation of primary
particles or the condensation of reaction products.  The presence,

                                    244

-------
however, of relatively high concentrations of biologically active trace
elements in the submicron mode may present special problems for particu-
late abatement technologies.

We have performed further studies to evaluate the relative distribution
of elements in fly ash associated with the aluminosilicate matrix com-
pared to those either in separate mineral phases or associated with the
particulate surface (23).  Preferential dissolution techniques utilizing
either hydrochloric acid or hydrofluoric acid indicated that greater
than 70 percent of the titanium, sodium, potassium, magnesium, hafnium,
thorium, and iron is associated with the aluminosilicate matrix.  On the
other hand, more than 70 percent of the volatile elements arsenic, se-
lenium, molybdenum, zinc, cadmium, tungsten, vanadium, uranium, and
antimony are associated with the particulate surface.  It also appears
that the majority of the calcium, scandium, strontium, lanthanum, and
the rare earth elements is associated with a separate mineral phase,
possibly an apatite phase, which has a particle size distribution simi-
lar to the aluminosilicate phase.  These findings, then, allow for com-
putation of the probable biological availability of trace elements in
coal fly ash.  In agreement with these observations, we have found rela-
tively high degrees of solubility for molybdenum, calcium, selenium,
barium, arsenic, tungsten, zinc, and antimony in coal fly ash treated
with a Tris-hydrochloric acid buffer at pH 7.4 (17).

                       IMMUNOTOXICOLOGY STUDIES

Although many sensitive in vitro bioassays exist for mutagenesis, few in
vitro assays are available for screening environmental toxicants for
their potential effects on host cellular defense factors.  The difficul-
ties associated with the development of assays for monitoring immune
effects are underscored by the extremely complicated nature of the
cellular and humoral factors involved in the host reaction to neoplasia.
To this end, we have initiated development of a variety of methodologies
for the study of environmental factors affecting cellular immunities.
Major effort has been dedicated to the development of assays reflecting
inhibition of macrophage function.

In Vitro Studies

Scanning electron microscopic X-ray analytical techniques have been ap-
plied to the investigation of particles contained in macrophages (26,35).
We have demonstrated .that the matrix composition of particles contained
within phagocytes may vary dramatically.  Furthermore, one may hypothe-
size that the variation in chemical composition of phagocytized parti-
cles may also represent a variation in toxicological potential of such
particles (26,27).   In this regard, a model has been proposed for the
exposure of individual lung cells to the foreign elements contained in
fly ash.  Segregation of elements in specific particles of fly ash re-
sults in a possibility for significantly higher exposure levels within
individual cells than that projected from the model of elemental concen-
trations distributed uniformly among all particles.  Furthermore, we
have demonstrated the feasibility of comparative microscopic techniques
for a correlation of viability of individual cells with the elemental
composition of phagocytized particles.  Light microscopic analysis of
                                    245

-------
macrophages deposited on SEM finder grids and treated with trypan blue
dye indicates the viability of the cells; subsequent electron-microscopic
analysis of the same cells provides indication of the elemental composi-
tion of the phagocytized particles.  Studies are now under way to evalu-
ate the relative toxicity of the various fly ash compositions.

To evaluate whether there is a potential for trace elements contained
within fly ash to alter the function or viability of macrophages, we have
calculated the elemental content of macrophages after phagocytizing a
few fly ash particles (27).  In general, our calculations indicate that
the concentrations of many biologically active trace element are in-
creased by one, two, or even three orders of magnitude.  These calcula-
tions, then, indicate that the trace elements in fly ash have a poten-
tial for producing cellular damage.  Further studies are necessary to
evaluate biological availability of trace elements within the fly ash.
In this regard, we have found that many trace elements within fly ash
are effective inhibitors of either lectin-induced or mixed-lymphocyte-
induced lymphocyte blastogenesis (37).  Relative to concentrations in
coal fly ash, chromium, lead, vanadium, and copper appear to be effective
inhibitors of the lymphocyte stimulation assays.  It is interesting to
note that both lymphocytes and macrophages appear to be extremely sensi-
tive to vanadium.

In vitro exposure of macrophages to particulate matter of similar size
distributions has provided a comparison of the relative toxicity of fly
ash, silica, and glass beads (13,38).  Silica was chosen as a positive
control since it is a well-documented, macrophage toxicant.  Glass beads
or aluminosilicate particles were chosen as a negative control because
of apparent inertness.  Initial studies of the in vitro effects of expo-
sure to these particles have been performed with both rat and mouse pul-
monary macrophages (13).  Particulate exposure was performed at a 40 to 1
particle-to-cell ratio comparable to concentrations of test particles
used in the phagocytic assay (12).  Our studies indicate that the phago-
cytic activity of macrophages increases with time in incubation media
(13).  Control phagocytosis tends to increase with increased time in
culture; the degree and rate of enhancement of phagocytosis is dependent
upon the species derivation of macrophages.  Rat macrophages tend to
show enhancement after two hours in culture whereas murine macrophages
tend to exhibit enhancement after two days in culture.  Exposure to fly
ash resulted in a lag time for increased phagocytic rates.  The lag ap-
peared at two and four hours for rat macrophages.  Similarly, fly ash
produced a lag in phagocytosis compared to controls at one and two days
in culture for murine macrophages.  Interestingly, although, silica expo-
sure did not produce a lag phase, the final phagocytic capability at
seven days was markedly below that of controls.  Further studies are
necessary to evaluate the significance of such a lag phase as well as to
define the nature of the enhanced phagocytosis associated with in vitro
culture.

We have also developed techniques for evaluation of the proliferative
capacity of lavaged cells from the lung (1,2).  The clonogenic technique
used was adapted from previous techniques utilized for quantitation of
bone marrow granulocyte-monocyte progenitors (40).  The basic culture

                                     246

-------
system involves the plating of lung-lavaged macrophages into a semi-
solid methylcellulose medium.  Colonies tend to be rather small and slow
to grow compared to the kinetics of hematopoietic cells.  Preliminary
studies indicate that all particle types, i.e., fly ash, silica or glass
beads, may have an effect on the ability of these lavaged cells to
divide (38).  We have generally found that the response of functional
assays is the opposite of the proliferative response.  Agents that tend
to stimulate phagocytosis appear to result in a decrease in prolifera-
tive capacity.  Similarly, we have found that agents that inhibit phago-
cytosis tend to stimulate proliferation.  These observations suggest
that particle-induced phagocytic functions may preclude differentiation
and subsequent division of progenitors.

We have also demonstrated the utility of cloning techniques in the study
of the in vitro dose-response characteristics of trace elements on lym-
phohematopoietic progenitors using semi-solid culture systems (39).
Because of the enhanced concentration of zinc and selenium in fine fly
ash particles, and because of the known biological activity of these
elements, we evaluated the response of murine spleen B-lymphocyte pro-
genitors and bone marrow granulocyte-monocyte progenitors to selenite
and zinc exposures.  Interestingly for both elements, a significant sup-
pression of cell proliferation from splenic B-lymphocytes was observed
at physiological concentrations.  This was not the case for the
granulocyte-monocyte progenitors.  The results demonstrate the feasibil-
ity of using lymphohematopoietic cloning techniques as sensitive short-
term bioassays to determine the effects of fossil fuel combustion prod-
ucts on cellular pathways involved in hematopoiesis and immunological
processes.  Further studies are presently under way to evaluate the
effects of in vivo exposure on the progenitor cell function.

In Vivo Studies

Exposure Conditions

We have also performed in vivo inhalation studies with mice acutely ex-
posed to fly ash and silica aerosols and chronically with rat exposure
to fly ash alone.  In this regard, we have developed technologies for
the generation of well-dispersed fly ash aerosols.  Our general approach
to aerosol generation involves the utilization of a Wright dustfeed
mechanism for fly ash deagglomeration and aerosolization followed by a
cyclone for separation of larger particles and a krypton-85 discharger
for reduction of particulate charge to Boltzman equilibrium (36).  In an
attempt to improve aerosolization procedures, we compared the efficacy
of a fluidized-bed generator to that of the Wright dustfeed system (30).
The results of the study indicated that aerosols resulting from the use
of the fluidized-bed generator are relatively unstable with time and
that deagglomeration is markedly less efficient than that associated
with the Wright dustfeed mechanism.  Aerosols produced by the Wright
dustfeeder had smaller aerodynamic size and broader sized distributions
than those resulting from the use of the fluidized bed with or without
use of the Wright dustfeed as a feed mechanism.  It is for these reasons
that we have continued to utilize the Wright dustfeed rather than a

                                    247

-------
fluidized bed for generation of stable, deagglomerated aerosols of fly
ash as well as other non-hygroscopic particles.

Acute inhalation studies have been performed with small exposure chambers
for periods up to two hours.  Such systems employ nose-only technology.
Chronic inhalation studies have been performed in immersion chambers for
periods up to 180 days for 20 hours per day (36).  Size distributions
are continuously monitored using a light-scattering particle counter.
Size distribution data and aerodynamic data are also obtained using
scanning electron microscopic analysis of point-to-plane ESP samples or
cascade impactor samples.  Total mass measurements are made periodically
utilizing filter samples.  For acute inhalation studies, we have uti-
lized the finest fly ash fractions of the size-classified, stack-
collected fly ash.  However, because of the need for relatively large
masses of material for chronic inhalation, we have employed size-
classified material collected from the hopper of the power plant's elec-
trostatic precipitator.

Macrophage Studies

In acute inhalation studies with mice, using stack-collected fly ash,
animals were sacrificed two, six, and fifteen days after exposure, and
macrophage function, pulmonary pathology, and progenitor cell kinetics
were evaluated (19).  Macrophage functional assays indicated a depres-
sion in phagocytic capacity of fly-ash-exposed mice at six days and fif-
teen days after exposure compared to controls.  A similar depressed
phagocytic activity was observed in silica-exposed animals.  Progenitor-
cell assays indicated an initial depression in pulmonary alveolar macro-
phage colonies at two days after exposure and a marked elevation at 15
days after exposure.  In contrast to the alterations observed in pulmon-
ary macrophage progenitors, macrophage precursors in bone marrow and
spleen were not significantly affected.  These data suggest that the
elevation in progenitor cell activity appears to be due to recruitment
of progenitors from the lung itself, that is, local production.  Further-
more, when compared to in vitro studies, the replicative capacity of
macrophages is readily observed.  For example, shortly after acute inha-
lation exposure, as observed with in vitro studies, a depression in pro-
genitor cells was observed.  However, two weeks after exposure a marked
enhancement was observed, in contrast to in vitro studies.  Similarly,
preliminary studies with i.p. exposure to zinc have indicated that
lymphohematopoietic progenitors are less sensitive to in vivo exposure
than cells exposed in vitro.

Further evidence to support the recruitment hypothesis is provided by
observations of changes in the relative distributions of particles with-
in phagocytes (19).  For the two-, six-, and fifteen-day observation
periods, the relative distribution of particles was observed to continu-
ally decrease within the cells, suggesting again an enhanced production
of phagocytic cells within the lung environment.  On the other hand, we
have not observed similar effects with long-term, low-level chronic ex-
posure to fly ash aerosols derived from the power plant's electrostatic
precipitator.  It is not clear whether the difference in biological

                                   248

-------
response reflects either the concentration,—200 milligrams per cubic
meter versus 2 milligrams per cubic meter; the difference in species—
rat versus mouse; or perhaps, most importantly, the difference of fly
ash sources,—stack-collected versus electrostatic-precipitator-hopper-
collected ash.

We have also developed techniques for quantification of fly ash deposit-
ed in rat lung during the chronic inhalation studies (18).  Because of
the difficulty in separating particulate matter from lung tissue, we
chose to evaluate the utility of elemental analysis as a measure of lung
deposition of inhaled fly ash.  Selection of the appropriate element
required that the following criteria be fulfilled:

     1.   the elemental analysis should be specific and detection limits
          should provide a sensitive indicator of lung burden;

     2.   the tissue content of the element should be low and relatively
          constant;

     3.   the dissolution of the element should parallel the particulate
          mass dissolution; and

     4.   the element should be uniformly distributed throughout the
          size range of the fly ash under study.

Under these criteria, only aluminum, silicon, and titanium were consid-
ered for further evaluation.  Because of the elemental detection limits
of atomic absorption spectrophotometry, only aluminum and silicon were
chosen for further evaluation.  It was found that the analysis of alumi-
num was more sensitive and less troublesome than that of silicon.
Application of the technique to inhalation studies indicated that the
lung content of fly ash calculated from the aluminum analysis was in
quantitative agreement with calculations based upon available deposition
and clearance data.

                         MUTAGENESIS STUDIES

We have continued to evaluate the mutagenic properties of coal fly ash
extracts using the standard Salmonella assay system (6,11).  Our studies
indicate that the finest fly ash fractions are indeed the most mutagenic
in keeping with a model of surface deposition (9).  However, the 3.2 ym
fraction of fly ash was found to be more mutagenic than the 2.2 ym
fraction.  At first this was assumed to be due to antimutagens within
the fly ash and studies were performed to evaluate the possible role of
selenium as the selenite or fluorine as the fluoride.  These elements
were chosen because of the relatively high concentrations of selenium
and fluorine in the finest fly ash fraction compared to that of the
3.2 ym fraction.  Also selenium, as the selenite has been demonstrated
to be an antimutagen for acetylaminofluorene and its derivatives.  Fluo-
ride is generally recognized as an enzyme inhibitor.  Addition of these
elements to extracts of the 3.2 ym fraction, however, did not alter the
mutagenic activity.  Thus, we do not have experimental support for the
hypothesis that antimutagens are present in the finest fly ash fraction.

                                    249

-------
It has been suggested by Natusch (31) that the difference in mutagenic
activity may be due to differences in chemical absorption of mutagens on
the two fly ash fractions.  Further evidence for the chemisorption of
mutagens on fly ash surfaces is provided by the demonstration of photo-
stability of these compounds.  We have irradiated fly ash samples with
ultra-violet light, with sunlight, and X-irradiation, and have not ob-
served loss in mutagenic activity.  However, heating of the fly ash at
temperatures of 200 to 250°C results in loss of approximately one-half
of the mutagenic activity, while heating above 300°C results in complete
loss of detectable mutagenic activity (11) .  The biphasic nature of the
loss in mutagenic activity with heating indicates the presence of at
least two mutagens or classes of mutagens.  Most recently, we have
demonstrated that the mutagenic loss is due to decomposition of surface-
associated materials as opposed to volatilization (22) .  This important
finding further supports the indication of chemisorption of mutagens on
fly ash surfaces.

Fly ash collected by the power plant's ESP does not appear to be muta-
genic in the Ames test (11).  Furthermore, even upon size classification
to a size distribution equivalent to our finest, stack-collected frac-
tion, we do not observe mutagenic activity of the electrostatic-
precipitator-collected materials.  Our studies indicate that tempera-
tures of approximately 100°C may be critical for absorption of mutagens
onto fly ash surfaces.  Further support for these observations is pro-
vided by the calculations of Natusch and Tomkins (32) which indicate
that, indeed, the deposition of vapor phase polynuclear aromatic hydro-
carbons on fly ash particles is extremely sensitive to temperature
changes around 100°C.

Studies of the comparative efficiency of mutagen extraction by a variety
of solvents have been performed.  We have found that normal saline is a
very poor extractant of mutagenic activity, whereas horse serum and
serum from other species are fairly efficient extractants of fly ash
mutagens (6).  Our data indicate the formation of a mutagen-serum pro-
tein complex that can be isolated from the serum, although it is avail-
able to the bacterial genome.  Further studies indicated that albumen
alone is nearly as efficient as the total serum in extracting mutagens
from the fly ash (5).  Direct extraction of fly ash with dimethylsul-
foxide and sonication results in the greatest detectable levels of
mutagens.

We have not identified the chemical composition of the mutagens of fly
ash, although most recent studies using acidic, neutral, and basic aque-
ous extracts further support the indication that the mutagens in fly ash
are not inorganic.  Acidic aqueous fractions were not mutagenic, whereas
basic aqueous fractions contained approximately one-half of the mutagens
extractable with dimethylsulfoxide.  These results indicate that a
significant portion of the mutagenic activity in coal fly ash can be
accounted for by the presence of weak organic acids.
                                   250

-------
To evaluate the carcinogenic potential of coal fly ash, we have devel-
oped a modification of the tracheal implant system described by Griesemer
et al. (21).  For these studies, we have packaged fly ash in 0.2 ym
Nuclepore filters.  This approach allows for the slow release of muta-
gens to the sensitive tracheal epithelial cells and minimizes risks for
local tissue damage and toxicity.  We have employed the bacterial muta-
genesis system to monitor the release of mutagens from the fly ash in
the tracheal implants (3).  Studies are now underway to evaluate the
carcinogenic potential of the fly ash using the tracheal implant system.

                               CONCLUSIONS

In conclusion, our results demonstrate the extreme complexity of coal
fly ash in terms of matrix composition, morphological appearance, sur-
face trace element and organic chemical composition.  Assays are now
evolving for the measurement of the potential immunotoxicity of fly ash.
Acute inhalation studies have demonstrated that coal fly ash may be
equally as toxic as alpha-quartz to the pulmonary alveolar macrophage.
The feasibility of application of sophisticated cloning techniques for
the evaluation of potential lympohematopoetic effects from complex mix-
tures has been demonstrated.  Mutagens in coal fly ash appear to be
absorbed to fly ash surfaces and hence may exist in the environment for
relatively long periods of time.  Techniques are now being developed to
evaluate the carcinogenic potential of coal fly ash utilizing a combina-
tion of bacterial mutagenesis assays and tracheal implant carcinogenesis
assays.  Detailed chemical, morphological, and toxicological analyses
indicate the utility of coal fly ash as a model complex mixture.

                             ACKNOWLEDGEMENT

This work was supported by the U.S. Department of Energy through the
Laboratory for Energy-Related Health Research.  This review summarizes
collaborative efforts with a number of excellent scientists and close
personal friends.
                                   251

-------
                               REFERENCES
 1.   Boorman,  G.  A.,  L.  W.  Schwartz,  F.  D.  Wilson.   "Formation of Macro-
     phage Colonies In Vitro by Free  Lung Cells Obtained From Rats,"
     J.  of the Reticuloendothel.  Soc.,  26,  855-866  (1979).

 2.   Boorman,  G.  A.,  L.  W.  Schwartz and F.  D.  Wilson, "In Vitro Macro-
     phage Colony Formation by Free Lung Cells During Pulmonary Injury,"
     J.  of the Reticuloendothel.  Soc.,  26,  867-872, (1979).

 3.   Chrisp, C. E., and G.  L. Fisher, "Mutagenesis  of Coal Fly Ash
     Linked to Tracheal Graft Assay for Carcinogenesis (Abstract)," llth
     Annual Meeting Environmental Mutagenesis  Society, March 16-19,
     1980, 74.

 4.   Chrisp, C. E., and G.  L. Fisher, "Mutagenicity of Airborne Particles",
     Reviews in Genetic Toxicology, submitted  to Mutat. Res.

 5.   Chrisp, C. E., and G.  L. Fisher, unpublished data.

 6.   Chrisp, C. E., G. L.  Fisher and J.  Lammert, "Mutagenicity of Respi-
     rable Coal Fly Ash," Science, 199,  73-75, (1978).

 7.   Coles, D. G.,  R. C. Ragaini, J.  M.  Ondov, G. L. Fisher, D. Silberman
     and B. A. Prentice, "Chemical Studies of  Stack Fly Ash From a Coal-
     Fired Power Plant," Environ. Sci.  and Technol., 13, 455-459, (1979).

 8.   Fisher, G. L., "The Morphogenesis  of Coal Fly Ash," in Proceedings
     of the Symposium on the Transfer and Utilization of Particulate
     Control Technology, Denver,  Colorado, 1978, IV, 433-440, (1979).

 9.   Fisher, G. L. , "Size-Related Chemical and Physical Properties of
     Power Plant Fly Ash," in Aerosol Generation and Exposure Facili-
     ties, (K. Willeke, Ed.), Ann Arbor Science Publisher,  Michigan,
     203-214,   (1980).

10.   Fisher, G. L., D. P.  Y. Chang and M. Brummer,  "Fly Ash Collected
     From Electrostatic Precipitators:   Microcrystalline Structures and
     the Mystery of the Spheres," Science, 192, 553-557, (1976).

11.   Fisher, G. L., C. E.  Chrisp and 0.  G. Raabe, "Physical Factors
     Affecting the Mutagenicity of Fly Ash From a Coal-Fired Power
     Plant," Science, 205,  879-881,  (1979).

12.   Fisher, G. L., K. L.  McNeill, C. B. Whaley and J. Fong, "Attachment
     and Phagocytosis Studies With Murine Pulmonary Alveolar Macrophages,"
     J.  of the Reticuloendothel.  Soc.,  ,24, 243-252, (1978).
                                    252

-------
13.  Fisher, G. L., K. L. McNeill, C. B. Whaley and J. Fong, "Functional
     Studies of Lavaged Pulmonary Alveolar Macrophages Exposed to Parti-
     cles In Vitro," in Radiobiology Laboratory Annual Report, UCD 472-
     124, University of California, Davis, California, 50,  (1977).

14.  Fisher, G. L. and D. F. S. Natusch, "Size-Dependence of the Physi-
     cal and Chemical Properties of Fly Ash," in Analytical Methods for
     Coal and Coal Products, III, (C. Karr, Ed.), Academic Press, 489-
     541, (1979).

15.  Fisher, G. L., B. A. Prentice, T. L. Hayes and C. E, Lai, "Compara-
     tive Analysis of Coal Fly Ash by Light and Electron Microscopy," in
     Air, American Institute of Chemical Engineers, in press, (1980).

16.  Fisher, G. L., B. A. Prentice, D. Silberman, J. M. Ondov,
     A. H. Biermann, R. C. Ragaini and A. R. McFarland, "Physical and
     Morphological Studies of Size-Classified Coal Fly Ash," Environ.
     Sci. and Technol., 12, 447-451, (1978).

17.  Fisher, G. L. , D. Silberman, B. A. Prentice, R. E. Heft and
     J. M. Ondov, "Filtration Studies With Neutron-Activated Coal Fly
     Ash," Environ. Sci. and Technol., 13, 689-693, (1979).

18.  Fisher, G. L., D. Silberman and 0. G. Raabe, "Chemical Characteri-
     zation of Coal Fly Ash in Rodent Inhalation Studies, Environ. Res.,
     in press, (1980).

19.  Fisher, G. L. and F. D. Wilson, "The Effects of Coal Fly Ash and
     Silica Inhalation on Macrophage Function and Progenitors," J. of
     the Reticuloendothel. Soc., 27, 513-524, (1980).

20.  Gibbon, D. L., "Microcharacterization of Fly Ash and Analogs: the
     Role of SEM and TEM," in Scanning Electron Microsc. (0. Johari,
     ed.), Scanning Electron Microscopy, Inc. Chicago, in press, (1979).

21.  Griesemer, R. A., J. Kendrick and P. Nettesheim,  "Tracheal Grafts,"
     in Experimental Lung Cancer (E. Karbe and J, F. Park, ed.), Springer-
     Verlag, Berlin, 539-547, (1974).

22.  Hansen, L. D., C. E. Chrisp and G. L. Fisher, unpublished data.

23.  Hansen, L. D., and G. L. Fisher, "Elemental Distribution in Coal
     Fly Ash Particles," Environ. Sci. and Technol., in press.

24.  Hansen, L. D., D. Silberman and G. L. Fisher, "Crystalline Compo-
     nents of Stack-Collected, Size-Fractionated Coal Fly Ash," submitted.

25.  Hayes,  T. L., C.  E. Lai, B. A. Prentice and G. L. Fisher, unpub-
     lished data.

26.  Hayes,  T. L., J.  B. Pawley and G. L. Fisher, "The Effect of Chemical
     Variability of Individual Fly Ash Particles on Cell Exposure," in
     Scanning Electron Microsc., 1978, (SEM Inc., AMF O'Hare, 111.),
     239-244, (1978).
                                    253

-------
27.  Hayes, T. L.,  J. B. Pawley, G. L. Fisher and M. Goldman, "A Toxico-
     logical Model of Fly Ash Exposure to Lung Cells," Environ. Res., in
     press, (1980).

28.  Kubitschek, H. E. and F. R. Kirchner, "Biological Monitoring of
     Fluidized Bed Combustion Effluents," in Second Symposium on Appli-
     cation of Short-Term Bioassays in the Fractionation and Analysis
     of Complex Environmental Mixtures, in press (1980).

29.  McFarland, A.  R., R. W. Bertch, G. L. Fisher and B.  A. Prentice, "A
     Fractionator for Size-Classification of Aerosolized Solid Particu-
     late Matter," Environ. Sci. and Technol., 11, 781-784, (1977).

30.  McFarland, A.  R., C. Ortiz, G. L. Fisher and 0. G. Raabe, "Aerosol-
     ization of Bulk Powders Utilizing Fluidized Bed and Wright Dust
     Feed Techniques," in Radiobiology Laboratory Annual Report, UCD
     472-124, University of California, Davis, California, 10, (1977).

31.  Natusch, D. F. S. , Colorado State University, Fort Collins, CO,
     personal communication, February 26, 1980.

32.  Natusch, D. F. S. and B. A. Tomkins, "Theoretical Consideration of
     the Adsorption of Polynuclear Aromatic Hydrocarbon Vapor Onto Fly
     Ash in a Coal-Fired Power Plant," in Carcinog., _3_, (P. W. Jones and
     R. I. Freudenthal, ed.), Raven Press, N. Y., 145-154, (1978).

33.  Natusch, D. F. S., J. R. Wallace and C. A. Evans, "Toxic Trace
     Elements: Preferential Concentration in Respirable Particles,"
     Science, 183, 202-204, (1974).

34.  Ondov, J. M., R. C. Ragaini, R. E. Heft, G. L. Fisher, D. Silberman
     and B. A. Prentice, "Interlaboratory Comparison of Neutron Activa-
     tion and Atomic Absorption Analyses of Size-Classified Stack Fly
     Ash," Proceedings of the Eighth Methods and Standards for Environ-
     mental Measurement Materials Research Symposium, National Bureau of
     Standards, Gaithersburg, Maryland, 1976, 565-572, (1977).

35.  Pawley, J. B., and G. L. Fisher, "Using Simultaneous Three Color X-
     ray Mapping and Digital-Scan-Stop for Rapid Elemental Characteri-
     zation of Coal Combustion By-Products," J. of Microsc., 110, 87-
     101, (1977).

36.  Raabe, 0. G., K. D. McFarland and B. K. Tarkington, "Generation of
     Respirable Aerosols of Powerplant Fly Ash for Inhalation Studies
     With Experimental Animals," Environ. Sci. and Technol., 13, 836-
     840, (1979).

37.  Shifrine, M., G. L. Fisher and N. J. Taylor, "Effect of Trace Ele-
     ments in Coal Fly Ash on Lymphocyte Blastogenesis," submitted.

38.  Whaley, C. B., F. D. Wilson, G. L. Fisher, M. Shifrine, and
     K. L. McNeill, "Growth of Canine Alveolar Macrophage Colonies," in
     Radiobiology Laboratory Annual Report, UCD 472-124, University of
     California, Davis, California, 47, (1977).

                                    254

-------
39.  Wilson, F.  D.,  G.  L.  Fisher and B.  A. Concoby, "Studies on In Vitro
     Dose-Response Characteristics of Trace Elements (Zn, Se) on Lympho-
     hematopoietic Progenitors Using Semisolid Culture Systems," in
     Proceedings of 19th Annual Hanford Life Sciences Symposium on
     Pulmonary Toxicology of Respirable Particles, October 22-24, 1979,
     Richland, Washington, in press, (1979).

40.  Wilson, F.  D.,  L.  O'Grady, C. McNeill and S. L. Munn, "The Formation
     of Bone Marrow Derived Fibroblastic Plaques In Vitro, Preliminary
     Results Contrasting These Populations to CFU-C," Exp. Hematol.,  ^,
     353-354, (1974).


                                FOOTNOTE

     This manuscript was also presented at the Second Symposium on
     Application of  Short-Term Bioassays in the Fractionation and
     Analysis of Complex Environmental Mixtures, Williamsburg, Virginia,
     March 4-7,  1980.
                                     255

-------
                 ELEMENTAL ANALYSIS FOR ENVIRONMENTAL
                        ASSESSMENT MEASUREMENTS

                             K. T. McGregor
                        GCA/Technology Division
                                ABSTRACT

Several recent studies have been initiated to evaluate specific areas of
the Inorganic Analytical Protocol for Environmental Assessment Measure-
ments.  The results of a portion of this work which focuses on the
elemental analysis scheme are described here.  Emphasis is placed on the
Level 1 scheme and the effects of organic matter on elemental determina-
tions.  In conjunction with this work, several sample preparation
techniques were evaluated.  The preparation methods are discussed in
terms of their potential utility for environmental assessment measurements.

                             INTRODUCTION

Elemental analysis plays a key role in all phases of an Environmental
Assessment (EA) Measurement Program.  In the IERL-RTP three-tiered
approach (2) to the EA of stationary sources, the first phase (Level 1)
is devoted to comprehensiveness.  The Level 1 study cannot rely on
currently known toxic species or on prior knowledge of the source for
selection of a set of parameters to be determined.  The constraint which
disallows these assumptions at the outset of the measurement program is
meant to preclude the possibility of missing potential hazards.  The
objective of the Level 1 elemental analysis scheme is to provide a
complete elemental survey of the source emissions in order to define the
hazard potential of inorganic emissions and to identify possible control
needs.

Subsequent phases of the assessment seek more specific determinations.
The elemental analysis requirements of the second phase, Level 2, are
very specific and are usually limited to a few elemental determinations.
Level 2 elemental data are needed to confirm and refine Level 1 results
that flag emission problems.  Elemental analyses in conjunction with
separation schemes can also be used for indirect compound speciation at
Level 2.  Elemental analysis may also be required in the final phase,
Level 3, which entails process and control monitoring to define the
temporal variation of source emissions.

A series of studies have recently been conducted to evaluate several
areas of the Level 1 elemental analysis scheme.  As part of this research,
the effects of organic matter on elemental determinations by spark source
mass spectrography (SSMS), the primary Level 1 elemental analysis tech-
nique^), were investigated.  In addition, several sample preparation
techniques were evaluated in terms of efficacious elimination of organic
matter without resultant elemental losses or contamination.  The utilities
of the methods as preparation techniques for atomic absorption spectrom-
etry  (AAS) measurements  (principally Hg) were also assessed.
                                   256

-------
A summary of the results of these investigations is presented herein.
While emphasis is placed on Level 1 methods, the preparation studies
have definite implications regarding Level 2 analyses.

                              EXPERIMENTAL

Test Sample Fabrication

The test samples used to evaluate the effects of organic matter on
Level 1 elemental analyses were prepared from spiked NBS SEM 1633 Fly
Ash and an organic mixture.  The fabrication procedure consisted of
three steps:  First, the standard fly ash was spiked by the addition of
solid compounds containing nine elements (U, Pb, Ce, I, Ag, Se, As, Sc,
and Cl).  Because previous experience had shown that inhomogeneity is
likely to result when a sample is diluted by more than 10:1, a multiple
dilution technique was utilized in preparing the fly ash starting
material.  In the next fabrication step, an organic mixture was prepared
consisting of 80% ascorbic acid, 10% benzoic acid and 10% carbon.  Each
of these materials was screened for elemental impurities and found to be
sufficiently pure for the present purpose.   The test samples were then
fabricated from the fly ash and organic mixtures.  Several sets of four
dilutions were made.  They are referred to throughout this paper as con-
taining 0%, 10%, 50% and 90% organic matter.  However, it should be noted
that the NBS fly ash proper does contain some residual organic matter.
Nevertheless, this sample is considered to be the baseline for 0% or
negligible organic content.  The theoretical concentration values for the
spiked NBS fly ash are given in Table 1.

Sample Preparation Methods

The four sample preparation methods investigated in this study are the
Parr oxygen bomb technique (PB), the low temperature plasma ashing method
(LTA), a hydrofluoric acid digestion bomb procedure (HFB), and a mineral
acid extraction (AE).   All four methods were applied to the fabricated
test samples and the PB and LTA methods were also applied to NBS SRM 1632
coal.
The PB procedures were performed in accordance with the Level 1 manual
(4).  The oxygen bomb apparatus was purchased from Parr Instrument Co.
and modified in house.  The modifications consisted of replacing the
standard electrodes with electrodes fabricated from 97% platinum and 3%
rhodium and fitting the bomb with a fused silica liner with cover.

The LTA preparations followed the instrument manufacturer's recommenda-
tions (5).  An LFE Model LTA-302 was used for all in-house preparations.
The instrument was operated at an rf power level of 300 watts with an 02
flow of 200 cc/min.  The ashing times varied from 6 to 12 hours.  All
samples were ashed to constant weight and the percent ash calculated.  The
percent ash values were used to adjust the analytical results to reflect
the original sample weight.
                                   257

-------
                                  TABLE 1

               Theoretical Values for NBS SRM 1633 Fly Ash
                        Including Spiked Elements
Element
u*
Th
Pb*
Tl
W
Ta
Hf
Lu
Yb
Tb
Eu
Sm
Ce*
La
Ba
Cs
I*
Sb
Cd
Ag*
Zr



Concentration
(ppm)
360
20
1160
3.2
3.7
1.5
6.4
0.81
5.7
1.5
2.0
10.0
2390
67
2200
7.0
79
5.6
1.2
1960
245



Element
Y
Sr
Rb
Br
Se*
As*
Zn
Cu
Ni
Co
Fe
Mn
Cr
V
Ti
Sc*
Ca
K
Cl*
Si
Al
Mg
Na
Be
Concentration
(ppm)
50
1380
102
9.8
257
980
176
104
80
34
5.0 (%)
404
103
191
6030
326
3.8 (%)
1.3 (%)
685
17 (%)
10 (%)
1.5 (%)
2610
9.7
* - Elements which were added (as compounds) to the fly ash.
                                    258

-------
The HFB procedure studied was proposed by Hartstein et al. (3).  This
procedure is similar to the classical HF bomb technique (1) routinely
employed in the preparation of mineral samples.  However, two steps have
been added which make the technique much more amenable to the present
needs.  First, a nitric acid digestion step is employed to aid removal
of organic matter prior to HF digestion which is intended to be a disso-
lution step.  After HF digestion the resulting sample is treated with
boric acid to further aid dissolution.  The procedure^ utilizes a teflon-
lined stainless steel bomb (also available from Parr Instrument Co.).

The AE procedure was performed according to the Level 1 acid extraction
method for solid samples and particulate filter samples to be subjected
to cold vapor mercury analysis (4).  This procedure specifies a 6-hour
extraction in a 4 to 1 mixture of HNO-j to HC1, respectively, followed by
filtration and dilution to 100 mL,

Elemental Analysis

SSMS analyses were performed on the four test samples by three different
laboratories.  The combination of replicate analyses and the number of
laboratories involved yields 24 possible data sets for the four samples
used in the evaluation of organic matter on SSMS measurements.  No
effort was made to specify the SSMS procedures to be used.  The choices
of instrument operating conditions, electrode preparation techniques,
standardization procedures, interpretation and quantitation methods, etc.,
were left completely up to the individual laboratories.  The laboratories
were selected to represent a cross-section of instrument types and pro-
cedural details.  SSMS analyses were also performed on NBS SRM 1632 coal
samples after preparation by the PB and the LTA techniques.  Selected
samples prepared by these techniques were analyzed by the different
participating laboratories.

All atomic absorption spectrometry (AAS) analyses were performed in-house
on a Perkin Elmer Model 460 equipped for both flame and flameless methods.
The cold vapor technique described in the Level 1 manual (4) was used
for all mercury determinations.  Other elemental determinations by AAS
employed either standard flame or graphite furnace (Model HGA 2100)
techniques as appropriate.

Selected elemental determinations were also made by Inductively Coupled
Argon Plasma Spectroscopy (ICAPS).  A Jarrell-Ash Model 1160 Plasma
AtomComp '™' equipped with 24 fixed wavelength channels and a Mark V N+l
channel was used for all ICAPS analyses.

                         RESULTS AND DISCUSSION

Organic Effects on SSMS

Organic constituents of a sample can potentially affect SSMS elemental
determinations in several ways.  The general areas in which organics
can cause problems include:
                                   259

-------
     • Electrode Preparation

     t Data Collection

     • Qualitative Interpretation

     • Quantitation

     • Detection

The nature of the effects and the degree to which the analysis may suffer
are dependent on the specific organic material present.  A consideration
of the vast number of different organic matrices that can be envisioned
clearly places constraints on the evaluation study at the outset.  It
would be an extraordinarily costly if not futile effort that endeavors
to precisely determine all effects produced by a host of matrices.  There-
fore, the effects that can be reasonably studied are limited by the choice
of test materials.

The test samples employed in this study allowed some specific areas to be
addressed.  The organic matrix chosen contributed only carbon, hydrogen
and oxygen to the test samples.  This matrix produces an abundance of
molecular spectra that are typical of those which might be produced by
organic matter in environmental samples.  Thus, the effects of molecular
fragments on qualitative interpretation and consequently on quantitation
could be observed.  Also, the organic matrix present in the test samples
allowed reasonable simulation of the sparking (burning) characteristics
of incompletely combusted matter and thereby allowed some evaluation of
the effects on data collection.  The only effect on detection expected
from the organic constituent of the test samples used is that of simple
dilution which alters the ultimate detectability but does not alter the
sensitivity.  The remaining area, electrode preparation effects, was not
addressed in this study.  Electrode preparation problems are immediately
obvious to the analyst; if suitable electrodes cannot be prepared, severe
matrix problems are indicated.

The 0% sample was used as the baseline for all effects considered.
Figure 1 shows the results obtained from all laboratories on the 0%
sample.  This histogram plots elements versus A (= log (^/X )) where
Xjn is the concentration reported and Xa is the actual concentration.
For example, all determinations in the range 0.0 to ± 0.10 are within
25% of their true values.  The vertical lines indicate deviations from
the actual values by factors of 2 (A = ± 0.30) and 3 (A = ± 0.48).  The
results reported by Laboratory B for the 0%, 10%, and 50% samples, which
are representative of the results reported by all three laboratories,
are shown in Figure 2.  As seen from Figure 2, there is little difference
in the results for these three samples.  Although the data for the 50%
organic sample indicate a slight positive bias, it does not exceed the
precision of the measurement.  Also, the results from the other labora-
tories do not substantiate any positive bias with increasing organic
content.
                                   260

-------

LABORATORY A

1





Mg
Yb K Sc
i Be I Lu U As
i i >i i il i i i |
•






.Ti
Mn
Ni
Br
TI
n
<.-.? -.6 -.5 -.4 -.3 -.2 -.1 0 .1


LABORATORY B
Al
Si
Ca
Ti
As
Se
Na Y
Bu Fe La I
i •>•> i Sc 1 Mg i U Cs i Lu iCe



^
K
V
Mn
-Co
Ni
Zr
Yb
Pb
{.V.
£-.7 -.6 -.5 -.4 -.3 -.2 -.1 0 .1
\
LABORATORY C






Ti .
Mn
Se
Fe As Sb
Be Sc I Ce Lu
1 !•>•>! ll 1 1 1 1


Co
Ni
"Br
Rb
Y
Zr
.Ba
La
Sm
Eu
U
i
tt
<-.7 -.6 -.5 -.4 -.3 -.2 -.1 0 .1
1
1
1
1
1
I V
1
1 Zr
1 Sb
Na 1 Ba
Co | La
Zn Ce
Se | Sm
Y Zu Pb
Cu Cd ' Tb Th
i i i










Cr
Cs
Hf Rb
1 •Nt 1 1
cc
.2 .3 .4 .5 .6 >_ .7
\
\
1
1
1
1
1
Cu
Zn '
Sm 1
Tb 1
Hf Rb i Sr Br
Th , Eu , Sb , TI












1 1 ->•) lcd 1
tv.
.2 .3 .4 .5 .6 >_ .7
I
1
I
1
1
Si ,
K '
Ca |
Cr .
Cu '
Zn Mg 1
Cd V |
Cs Sr I
Pb Tb Yb
Th Hf 1 TI
i i i













Na
1 ' >y '
(C
.2 .3 .4 .5 .6 > .7
                             LOG




FIGURE 1   Results For All Laboratories On The 0% Organic Sample



                                  261

-------
1
1
0% ORGANIC |
' ' Al
' 1 Si
| Ca
1 Ti
1 As
' 1 Se
| Na Y
Ba . Fe La I
Sc Mg U ' Cs Lu Ce
1 1 V» 1 II I I 1 I




K
V
Mn
.Co

Zr
Yb
Pb
1
U
i 1
1 I
1 1
1 1
1 1
1 1
Cu i 1
Zn |
Sm 1
Tb 1 '
Hf Rb - Sr Br 1
Th Eu Sb Tl 1 Cd
1 1 1 II 1 v» 1 1
ll
           -.6  -.5  -.4 -.3 -.2   -.1    0    .1
     .2
                  .4  .5   .6
10% ORGANIC
<-.7
1 1
1 1
1 1

| 1 Ti
Sc Mn
1 La .
] | Y Si Ce "
• Sb Ca Sm
1 Na Lu Fe Yb
Al I Mg! Pb I U
i i y» i il i i i 	 i

••

1
1
1
1
1

i
.V Rb
"As Ni
Se Tb Co
1
Zn 1
Eu Hf Cu Zr |
Th Tl Ba Cs Sr Br Cd
i i i i li i .« i i
tt CC
7 -.6 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6 >_. 7
ORGANIC i
1 i
I ,
i
1 1
1
' 1
i
, 1 Na
1 i Ca
1 ' Ce
I 1 Mg Sm "
I i As Lu
1 Sr Se Pb
1 Y 1 U La Th
,„..!. iii
r

1
1
l
1 I
1 i
1 1
Si
Al K
Mn V
1
I
Ti .
Sc Zn Cu Ni '
Fe I Ba
Rb Cs Eu
Sb Zr 1 Co Br
Hf Tl 1 Tb Cd
1 i i i ll 1^1 i
11 « (i
0
.1   .2  .3   .4  .5
                                                                  .6
.7       -.6   -.5   -.4  -.3 -.2  -.1
                               LOG
FIGURE 2   Laboratory B Results On The 0%, 10% and 50% Organic Samples
                                    262

-------
SSMS analysis of the 90% organic sample was problematic for all labora-
tories.  While electrode preparation problems were not reported, data
collection and photoplate analysis problems with this sample were
encountered.  Of the six samples submitted for analysis, only one data
set on the unaltered 90% sample was obtained.  Although this data set
did contain substantially more unacceptable values and showed consider-
able scatter, valid conclusions regarding organic interferences cannot
be drawn from the results of this single sample.  Excessive molecular
spectra were cited as a cause for concern with this sample.

The photoplates produced by all laboratories revealed increased molecular
spectra with increasing organic content.  Molecular spectra are typically
produced from organic constituents in the sample in fairly consistent
patterns or "runs".  A hydrocarbon run usually takes the form Cnt^-i^
(X=0, ± 1, ± 2, ± 3, etc.).  The spectral lines for X = 0 are the most
intense and the intensity rapidly decreases with increasing X.  Hydro-
carbon runs resulting from the organic constituent of our test samples
were easily discernible by all laboratories by visual inspection of the
photoplate as evidenced by the fact that all reports included comments
to this effect.  In fact, the observation of these spectra provided the
analysts with a guideline for sample preparation requirements.  In
general, if the molecular spectra were judged excessive by the individual
analysts, sample preparation steps were undertaken.

The data resulting from this set of experiments substantiate the validity
of the "molecular spectra guidelines" for determining sample preparation
requirements; however, this may not be the most cost-effective procedure
for deducing the necessity for sample preparation.

It should be noted that while the vast majority of the determinations
fell within a factor of 3 of the actual value, some did exceed this range.
The factor of 3 is used as a breakpoint for accuracy evaluations in view
of Level 1 objectives.  Because a factor of 3 allows no margin for error
in other phases of the measurement endeavor, deviations approaching a
factor of 3 should be considered marginally acceptable.  The data obtained
during this study revealed determinations outside the acceptable range
for all participating laboratories.  Although the unacceptable determina-
tions were consistent (some elements) for the different samples within the
same laboratory, an interlaboratory comparison reveals few consistent
problems.  It was observed that beryllium appears to be a consistent
problem; however, the sign (±A) of the error was not the same for all
laboratories.  These results indicate that specific procedural differences
are probably responsible for the observed problems; there is no evidence
suggesting that organics were responsible for the outlying determinations.

Sample Preparation Methods

In this phase of the study, several preparation methods were evaluated
in terms of 1) their abilities to eliminate organic constituents with-
out concomitant elemental losses, and 2) the comprehensiveness of their
dissolution capabilities.  Although the SSMS analysis does not require


                                   263

-------
sample dissolution, liquid samples are amenable to SSMS analysis.  Con-
versely, other elemental analysis techniques, especially the likely
choices for Level 2 analysis, do require solution samples.  If multi-
element analyses are necessary at Level 2, a comprehensive dissolution
technique for solid samples would be desirable.

The preparation methods studied were:

     t Parr Oxygen Bomb Combustion (PB)

     • Low Temperature Plasma Ashing (LTA)

     • Acid Digestion Bombing (HFB)

     • HN03/HC1 Extraction (AE)

The PB procedure was performed according to the Level 1 procedures manual
(4).   It should be noted that two slightly different PB procedures are
specified.  The necessity for two procedures arises because of the possi-
ble "mineral" residues after combustion.  While these residues must be
digested and removed prior to the cold vapor mercury analysis, they
should be retained and combined with the resulting solution prior to SSMS
analysis.  In order to avoid subsequent aliquotting problems, we recommend
that the entire sample (all solution and all residue) be used for the SSMS
analysis.  This eliminates inaccuracies or biases produced by aliquotting
a solid/solution mixture and reduces contamination introduced by volume
adjustment and subsequent aliquotting.

The SSMS results for the Parr bombed 90% organic sample are shown in
Table 2.  The set of elements shown in Table 2 were used for all prepara-
tion technique evaluations when SSMS was the analytical finish.  With
some exceptions, these elements represent all of the well characterized
concentration data for NBS SRM 1633 and 1634.  The few exceptions were
elements determined to be problematic during the baseline study described
above and, therefore, are inappropriate for use here.

As seen from Table 2 the data resulting from the PB preparation satisfy
the Level 1 requirements.  It is clear from this data set that elemental
losses are not a problem with the PB method.  Inspection of the photo-
plates resulting from this analysis indicates that the organic matrix
was effectively eliminated.

We note that, ostensibly because of the mineral nature of the inorganic
constituents, residues did result on combustion of this sample.  The
data presented in Table 2 correspond to sample electrodes prepared from
the combination of the solid and liquid fractions.  The SSMS and AAS
results obtained on the solution only portion of this sample are shown
in Table 3 and Table 4, respectively.   The SSMS data shown for the two
laboratories consistently illustrate the elemental losses which occur
upon discarding the solid residue.  The AAS data, given in Table 4,
confirm the SSMS results.  The mean concentration values in Table 4
were derived by first normalizing the results for all (0%-90%) test

                                   264

-------
                             TABLE 2
        SSMS Results on Oxygen Bombed 90% Organic Sample
Element
U
Th
Pb
Yb
Eu
Sm
Ce
La
Ba
Cs
I
Sb
Cd
Zr
Br
Se
As
Zn
Cu
Ni
Co
Fe (%)
Mn
Cr
V
Ti
Ca
K
Si (%)
Al (%)
Mg
Na
Expected
Value
(ppm)
36.0
2.0
116
0.57
0.20
1.0
239
6.7
220
0.70
7.9
0.56
0.12
24.5
0.98
25.7
98
17.6
10.4
8.0
3.4
0.58
40.4
10.3
19.1
603
3800
1300
1.7
1.0
1500
261
Measured
Value Effect*
(ppm)
34 /
3.3 /
120 /
ND o
0.12 /
0,83 /
260 /
13 /
430 /
0.72 /
5.0 /
0.84 /
0,14 /
35 /
0.52 /
38 /
140 /
20 /
16 /
16 /
6.0 /
0.88 /
53 /
13 /
24 /
1000 /
4900 /
670 /
2.1 /
MC o
600 /
240 /
*/ indicates within a factor of 2; /± indicates high (+) or low (-)
 but within a factor of 3; + or - denote results unacceptably high (+)
 or low (-), respectively; o means no evaluation made.
                                  265

-------
                              TABLE 3

    SSMS Results On Oxygen Bomb Solution Of 90% Organic Samples
Element
U
Th
Pb
Yb
Eu
Sm
Ce
La
Ba
Cs
I
Sb
Cd
Zr
Br
Se
As
Zn
Cu
Ni
Co
Fe
Mn
Cr
V
Ti
Ca
K
Si (%)
Al (%)
Mg
Na
Expected
Value
(ppm)
36.0
2.0
116
0.57
0.20
1.0
239
6.7
220
0.70
7.9
0.56
0.12
24.5
0.98
25.7
98.0
17.6
10.4
8.0
3.4
5800
40.4
10.3
19.1
603
3800
1300
1.7
1.0
1500
261
LAB X
Measured
Value Effect*
(ppm)
6.9
0.4
21
ND o
ND o
0.30
38
0.63
44
ND o
0.30
0.30 /
0.06 /
1.9
0.50 /
4.1
69 /
8.8 /
4.3 /-
1.4
0.60
750
14
1.9
4.7
44
1400 /-
1800 /
0.21
0.31
3400 /+
280 /
LAB Y
Measured
Value
(ppm)
4,4
ND
23
ND
ND
ND
17
0.72
19
0.082
1.7
0.18
0.037
1.2
0.54
8.6
44
7.1
1.0
4.7
0.26
220
2.4
1.9
1.9
18
250
120
0.021
0.27
110
320

Effect*
—
o
-
0
o
o
-
-
-
-
-
-
-
-
/
-
/-
/-
-
/
-
-
-
-
-
-
-
-
-
-
-
^
*/ indicates within a factor of 2; /± indicates high (+) or low (-)
 but within a factor of 3; + or - denote results unacceptably high (+)
 or low (-), respectively; o means no evaluation made.

                                   266

-------
                               TABLE 4

        AAS Results (Normalized) on the Oxygen Bomb Solutions
                   Of All (0-90%) Organic Samples*
Element
Pb
Hg
Cd
Se
As
Zn
Cu
Ni
Mn
Cr
V
Expected
Value
Cppm)
1160
0.11
1.2
257
980
176
104
80
404
103
191
Mean
Value
(ppm)
444
0.12
2.1
106
1270
45
31
26
114
13
102
% RSD
19
_ 5%*
2,0
51
26
3.2
22
30
45
30
10
% D
-62
9
75
-59
29
-74
-70
-68
-72
-87
-47
 * Data were normalized to the 0% sample concentrations prior to
   statistical analysis.

** Hg concentrations in the organic diluted samples were not detectable.
                                   267

-------
samples to the 0% organic level and then subjecting the data to statistical
analysis.  The normalization to 0% organic content creates the factor of
ten difference in expected values (and mean values for AAS) seen in these
two tables.  The differences between the AAS and SSMS values observed are
well within the combined experimental errors of the two techniques.

The AAS data also indicate the utility of the PB method for mercury
determinations even when the inorganic matrix is, in general, refractory.
It is equally interesting that arsenic was well determined on the PB
solution by both techniques.  As noted in Table 4, no statistical analysis
could be performed on the mercury data.  Since the mercury concentration
of the 0% sample is very near the experienced detection limit, any organic
dilution resulted in non-detectable mercury values.  This result is
simply a consequence of the restricted sample size employed.  The sample
size limitation is the only drawback of the PB technique experienced
during this study.  While contamination effects were not specifically
addressed, none were noted.

The results obtained from two laboratories on the 90% sample prepared by
LTA are shown in Table 5.  By and large, these data satisfy Level 1
requirements, but some inconsistencies between the two laboratories are
noted.  The values reported for zinc and zirconium by one laboratory
are abnormally high and suggest ambient environmental contamination.  No
rationalization can be given for the low values reported for cerium and
titanium.  Inspection of the photoplates produced by each laboratory
reveals very clean essentially organic free spectra.  The results from
both laboratories indicate losses due to preparation for iodine and
bromine.

Table 6 compares the SSMS results on NBS SRM 1632 coal prepared by the PB
and the LTA methods.  As seen from these data, the two methods yield
comparable results.  Halogen losses are observed for both techniques with
this sample.  This differs from the previous results which indicated that
no losses were incurred via the PB preparation.  This discrepancy suggests
that matrix effects may be a problem with halogen determinations by SSMS.
The inconsistency observed for the silicon determination is not unexpected
since major component measurements are difficult to make by standard SSMS
techniques.  Inspection of the photoplates resulting from each preparation
of this sample indicated adequate organic removal for both techniques.
The LTA technique again produced the cleaner photoplate.

The SSMS data for the 50% sample prepared by HFB are presented in Table 7.
The 50% sample was selected for study by this procedure because it repre-
sents the maximum tolerable organic content (6) according to the manu-
facturer of the bomb.  The manufacturer's warning arises from the
explosion potential produced by samples containing primarily (>50%)
organic constituents.  It should be noted that this recommendation
applies to the standard HF procedure for which the bomb was designed.
The authors of the modified procedure  (3), which uses a nitric acid
digestion step to remove organics, did not address this issue specifically
but the samples prepared  (coal samples) did contain high organic content.


                                   268

-------
                            TABLE 5

           SSMS Results  on  LTA  of 90%  Organic  Samples
Element
U
Th
Pb
Yb
Eu
Sm
Ce
La
Ba
Cs
I
Sb
Cd
Zr
Br
Se
As
Zn
Cu
Ni
Co
Fe (%)
Mn
Cr
V
Ti
Ca
K
Si (%)
Al (%)
Mg
Na
Expected
Value
(ppm)
36.0
2.0
116.0
0.57
0.20
1.0
239.0
6.7
220.0
0.70
7.9
0.56
0.12
24.5
0.98
25.7
98.0
17.6
10.4
8.0
3.4
0.5
40.4
10.3
19.1
603.0
3800.0
1300.0
1.7
1.0
1500.0
261.0
LAB 1
Measured
Value Effect*
12.0 /-
2.0 /
124.0 /
1.2 /
0.30 /
1.4 /
16.0
3.5 /
380.0 /
1.1 /
> 5.0 o
1.4 /
4.8 +
77.0 +
> 1.4 o
21.0 /
32.0 /-
75.0 +
23.0 /
18.0 /+
11.0 /+
0.35 /
37.0 /
> 31.0 o
28.0 /
160.0
2000.0 /
1900.0 /
1.9 /
1.2 /
1500.0 /
240.0 /
LAB 2
Measured
Value Effect*
(ppm)
42.0 /
3.0 /
250.0 /+
1.8 /+
0.21 /
1.2 /
220.0 /
6.1 /
280.0 /
0.98 /
0.48
0.38 /
0.16 /
20 /
0.29
10.0 /-
33.0 /-
8.8 /
6.8 /
5.3 /
2.4 /
0.20 /-
32.0 /
6.7 /
18.0 /
580.0 /
3800.0 /
540.0 /-
0.96 /
MC o
600.0 /-
640.0 /+
^ indicates within a factor of 2; >/± indicates high (+) or low  (-)
but within a factor of 3; + or - denote results unacceptably high  (+)
or low (-), respectively; o means no evaluation made.
                                269

-------
                            TABLE 6
           SSHS Results On Prepared NBS SRM 1632 Coal
Element
U
Th
Pb
Yb
Eu
Sm
Ce
La
Ba
Cs
Sb
Cd
Br
Se
As
Zn
Cu
Ni
Co
Fe
Mn
Cr
V
Ti
Ca
K
Si
Al
Mg
Na



















(%)




(%)
(%)
(%)
(%)
(%)

Accepted
Value
(ppm)
1.4
3.0
30.0
0.7
0.33
1.7
19.5
10.7
352.0
1.4
3.9
0.19
19.3
2.9
5.9
37.0
18.0
15.0
5.7
0.87
40.0
20.2
35.0
1100.0
0.43
0.28
3.2
1.85
0.2
414.0
LTA
Measured
Value Effect*
(ppm)
1.8 /
5.4 /
21.0 /
ND o
ND o
ND o
12.0 /
9.6 /
270.0 /
1.1 /
3.0 /
0.11 /
3.1
3.5 /
13.0 /
53.0 /
45.0 /+
25.0 /
9.1 /
1.3 /
40.0 /
20.0 /
73.0 /
1500.0 /
0.71 /
0.21 /
12.0 +
MC o
0.20 /
1500.0 +
Oxygen
Measured
Value
(ppm)
0.56
2.7
56.0
ND
ND
3.1
16.0
13.0
430.0
0.74
3.4
0.073
2.0
2.3
8.6
43.0
32.0
33.0
6.1
0.90
53.0
26.0
24.0
1000.0
0.5
0.3
4.4
MC
0.063
480.0
Bomb
Effect*
/-
/
/
o
o
/
/
/
/
/
/
/-
-
/
/
/
/
/
/
/
/
/
/
/
/
/
/
o
-
y
*/ indicates within a factor of 2; /± indicates high (+) or low (-)
 but within a factor of 3; + or - denote results unacceptably high (+)
 or low (-), respectively; o means no evaluation made.
                               270

-------
                            TABLE 7

             SSMS Results On HF Bombed 50% Organic Sample
Element
U
Th
Pb
Yb
Eu
Sm
Ce
La
Ba
Cs
I
Sb
Cd
Zr
Br
Se
As
Zn
Cu
Ni
Co
Fe (%)
Mn
Cr
V
Ti
Ca (%)
K (%)
Si (%)
Al (%)
Mg (%)
Na
Expected
Value
(ppm)
180
10
580
2.9
1.0
5.0
1200
33.5
1100
3.5
39.5
2.8
0.60
123
4.9
129
490
88
52
40
17
2.9
202
51.5
95.5
3020
1.9
0.65
8.5
5.0
0.75
1300
Measured
Value Effect*
(ppm)
210 /
33 +
1200 /+
7.0 /+
2.5 /+
15 /+
2500 /+
65 /
3100 A
7.5 /+
7.5
2.9 /
0.38 /
290 /+
1.2
41
380 /
49 /
65 /
55 /
19 /
1.6 /
320 /
70 /
95 /
2200 /
1.8 /
0.31 /-
0.0039 /-
MC o
0.29 /-
750 /
*/ indicates within a factor of 2; /± indicates high (+) or low (-)
 but within a factor of 3; + or - denote results unacceptably high (+)
 or low (-), respectively; o means no evaluation made.
                                 271

-------
Advisedly, the present study did not attempt to pursue this limitation
of the technique.

In general, as seen from Table 7, the SSMS results for this preparation
are in reasonable agreement with the expected values.  The observed values
for thorium, iodine, bromine, selenium and silicon are notable dis-
crepancies.  The loss of silicon via the formation of volatile silicon
hexafluoride was expected; however, similar interactions apparently occur
between the HF and iodine, bromine, and selenium which consequently result
in losses of these elements as well.  While the value observed for
thorium is only slightly outside the acceptable range (well outside the
desired range), it emphasizes the general trend of positive bias observed
in the high atomic weight region for this sample.  No further investigation
of this trend was attempted in the present study.

The modified HFB procedure was also applied to Fluidized Bed Combustion
ash samples for AAS analyses in order to determine a materials mass balance.
Because the samples involved did not contain significant organic con-
stituents, the nitric acid step was omitted.  The boric acid step was
included since sample dissolution is mandatory for AAS analysis.  While
this study confirmed the dissolution capabilities of the modified HFB
procedure, some problems were noted.  Specifically, the use of boric acid
to aid dissolution introduces a severe interference to the graphite
furnace AAS determination of trace element constituents.  While the boric
acid has little effect on the standard flame AAS determination, the
graphite furnace AAS measurements required for ultra-trace constituents
are precluded by the effects of the boric acid step.

An acid extraction (AE) procedure that uses 4 parts nitric acid to 1 part
hydrochloric acid as the extraction medium was also evaluated.  In the
Level 1 scheme (4), this extraction procedure is employed to prepare all
primarily inorganic (non-combustible) samples for cold vapor mercury
analysis.  This method is also used to prepare particulate filter samples
for SSMS analysis in situations where the particulate cannot be removed
from the filter substrate.

The elemental analysis results obtained on NBS SKM 1633 fly ash prepared
by the AE procedure are shown in Table 8.  While most of the data shown
in Table 8 are ICAPS determinations, some of the data are AAS measure-
ments.  The AAS data were used to supplement the ICAPS results for
elements that were either not detected by ICAPS or the presence of inter-
ferences to the technique were obvious.  As seen from Table 8, the AE
procedure did extract substantial quantities of most elements.  With some
exceptions, quantities sufficient to meet Level 1 requirements were
extracted; however, the data clearly show a negative bias owing to
incomplete dissolution.  This bias from preparation, combined with the
limitations of the Level 1 SSMS procedure, could lead to unacceptable
SSMS results.  It should be noted that the AE in conjunction with cold
vapor AAS yields results for mercury that are adequate for Level 1 and
Level 2.  However, with a few exceptions, the AE method appears inadequate
preparation for Level 2 analyses.
                                   272

-------
                              TABLE 8
             Elemental Results on NBS SRM 1633 Fly Ash
                   Prepared By The AE Procedure*

Al <%)
Ba
Ca
Cd
Co
Cr
Cu
Fe (%)
Hg
K (%)
Mg (%)
Mn
Ni
Pb
Si (%)
Ti
V
Zn
Accepted
Value
(ppm)
12.7
2700
4.7
1.45
41.5
131
128
6.2
0.14
1.72
1.8
493
98
70
21
7400
214
210
Measured Level 1 Level 2
Value Requirements Requirements §
(ppm")
5.4 /-
2710 / /
2.6t /
2.0t / o
27 /
77 /
131 / /
3.7 /
O.llt / /
0.25t
0.92t /
283f /
46 /
22 -
0.0253
2760 /
214 / /
154 /
*Determined by ICAPS unless noted otherwise.
tDetermined by AAS.
§An accuracy of ± 25% is assumed for Level 2  requirements.
                                 273

-------
                              CONCLUSIONS

SSMS analyses were conducted by three laboratories on a set of test samples
containing various amounts of organic matter in an effort to evaluate the
effects of organics on the SSMS results.   These organic test samples and
selected NBS SRM materials were used to evaluate several sample prepara-
tion methods.  The following conclusions are drawn from the organic effects
study:

     • Organic components of a sample can adversely affect, and
       in some cases preclude, SSMS data collection.

     • Organic molecular spectra can hinder qualitative inter-
       pretation of the elemental spectra.

     • Molecular spectra produced by organic constituents are
       readily discernible by SSMS analysts.

     • Interference from the coincidence of organic and elemental
       spectra was not a problem.

     • For the test organic matrix used,  50% organic content was
       tolerable.

It is clear from the foregoing discussion that, in some cases, preparation
prior to SSMS analysis is necessary.  In addition, the delineation of
sample preparation methods suitable for Level 2 elemental analyses is
highly desired.  The following conclusions are drawn from the sample
preparation study:

     • The Parr Oxygen Bomb technique eliminates organic
       constituents adequately for SSMS analysis.   Although
       this procedure can result in a mixed phase sample, it
       appears to be the most comprehensive SSMS preparation
       technique when all residues can be collected.  While
       the method is inadequate for the dissolution of many
       elements, it is an excellent preparation method for
       cold vapor mercury analysis.  In terms of the Level 1
       elemental analysis scheme, the only significant draw-
       back to this method is the limited sample size that can
       be accommodated by the standard apparatus.

     • The Low Temperature Plasma Ashing technique is superior
       to any technique studied for the elimination of organic
       matter.  However, when nominally "volatile" elements are
       of interest, the method is questionable; halogen losses
       represent a fundamental problem with the method.  A major
       advantage of the method is the ability to accommodate
       large sample sizes.  Disadvantages to the method include-
       its vulnerability to ambient environmental contamination
       and the long ashing times required for some samples.

                                   274

-------
     t The Modified Hydrofluoric Acid Digestion Bomb technique
       is the most comprehensive dissolution technique studied.
       With the addition of a fuming nitric acid step, organic
       removal can also be accomplished.  Concentration steps
       achieved by solution evaporation will result in the loss
       of silicon; halogen losses also occur.  While the boric
       acid step aids sample dissolution, the resulting matrix
       interferes with graphite furnace AAS measurements.

     f The Level 1 Acid Extraction technique is adequate for
       the preparation of particulate samples for cold vapor
       mercury analysis.  Some other elements can also be
       successfully extracted by this method; however, the
       method has limited application potential for Level 2
       analysis.

In summary, the Level 1 elemental analysis scheme was found to be reliable
and to provide the kind of information desired from the Level 1 study.
Problems were noted with some SSMS determinations and additional work in
this area is recommended.  The preparation of particulate filter samples
for SSMS analysis by the AE procedure may bias the resulting data.

The HFB method shows great promise as a Level 2 preparation procedure
for analytical techniques requiring solution samples.  Additional work
in this area is needed.

                              REFERENCES

1.  Bernas, B., "A New Method for Decomposition and Comprehensive
    Analysis of Silicates by Atomic Absorption Spectrometry,"
    Anal. Chem., 40, 1682 (1968).

2.  Dorsey, J., C. Lochmuller, L.  Johnson, and R. Statnick,
    "Environmental Assessment Sampling and Analysis:  Phased
    Approach and Techniques for Level 1," EPA 600/2-77-15 (1977).

3.  Hartstein, A.M., R.W. Freedman, and D.W. Platter, "Novel Wet
    Digestion Procedure for Trace Metal Analysis of Coal by Atomic
    Absorption," Anal. Chem., 45,  611 (1973).

4.  Lentzen, D.E., D.E, Wagoner, E.D. Estes, and W.F. Gutknecht,
    "IERL-RTP Procedures Manual: Level 1 Environmental Assessment
    (Second Edition)," EPA 600/7-78-201 (1978).

5.  LFE Corporation, "Technical Bulletins for Applications,"
    LFE Bulletin, 8101, (1974).

6.  Parr Instrument Company, "Technical Information for the 4745
    Acid Digestion Bomb," Parr Bulletin, 1100, (1974).
                                   275

-------
           LEVEL 2 INORGANIC SAMPLING AND ANALYSIS METHODOLOGY
                        APPLIED TO FGD SYSTEMS

                            R.  F.  Maddalone
                               TRW DSSG
                               ABSTRACT

During the past 3 years TRW has been involved in the development of Level
2 inorganic sampling and analysis methodologies.  A comprehensive analysis
scheme was developed to meet both the trace element analysis and the compound
identification needs of IERL/PMB.

Under several EPA programs TRW has applied this analysis scheme, which
uses such techniques as TGA/DSC, optical microscopy, SEM-EDX, AAS, ICP,
FTIR, XRD, ESCA, and SIMS, to samples from SASS and Method 5 type trains
and impactor systems.  This paper will present a discussion of the applica-
tion of several of these inorganic sampling and analysis methods to par-
ticulate samples from oil-and coal-fired boilers equipped with FGD systems.
Data will be presented on the trace element enrichment across the scrubber,
surface characterization of size classified particulate matter, and com-
pound identification using FTIR.

                            INTRODUCTION

The development of more stringent air quality standards is changing the
inorganic analysis requirements for environmental samples.  There is a
growing trend in government regulations and academic interest to go beyond
trace elemental analysis to compound identification.  Unfortunately, no
single inorganic method can identify all compounds in all samples.  Added
to this problem is the fact that surface and bulk composition might be
widely different.  In order to meet these requirements and characterize the
inorganic portion of any environmental sample, specific data needs must be
defined and then appropriate analytical tools integrated into an analysis
plan to answer toxicological, engineering or process-related questions.

The samples discussed in this paper were taken as part of the comprehensive
assessment test program (EPA 68-02-2197) conducted by TRW at the coal- and
oil-fired industrial boiler at the Firestone Tire and Rubber (Pottstown,
Pennsylvania) plant and a coal-fired power plant.  Both sites had FGD
systems.  The Firestone boiler was equipped with a prototype 10 MW dual-
alkali system, where the power plant had a full scale (100 MW) commercial
limestone system.  At both sites, Method 5, SASS and impactors were used
to sample the flue gas streams.  These samples in turn were analyzed by
both Level 1 and Level 2 procefures.  This paper discusses selected results
from the  test program to illustrate the use of state-of-the-art analytical
techniques.  As a result of these analyses, data was obtained on such
topics as SMSS accuracy, trace element enrichment across an FGD, particle
size and  elemental distribution, surface composition and compound iden-
tification.
                                   276

-------
                     TECHNIQUES  FOR ELEMENTAL  ANALYSIS

Acquiring basic information on elemental emission concentrations from com-
bustion sources is becoming a major part of any site characterization.
Methods of analysis must be able to meet the Level 2 accuracy and sensi-
tivity requirements while being cost effective.  The dependable Atomic
Absorption Spectroscopy (AAS) is, and will be, the mainstay of future
elemental analysis.  However, for more cost effective analyses, multi-
element techniques will be required.  Classical techniques like DC or AC
arc emission are still used to provide generally qualitative elemental
information for specific matrices.  The Industrial Environmental Research
Laboratory at RTF has recommended Spark Source Mass Spectrometry (SSMS) as
its Level 1 elemental analysis method.  While it is sensitive (ppb) and has a
wide elemental range, its accuracy as employed in the Level 1 (1) procedures
is approximately a factor of 2-3.  Unfortunately, uncertainties in SSMS data
can be much greater than those quoted limits.   Figure 1 shows a histogram of
the ratio of SSMS to AAS values for 17 elements (A£, As, Ca, Cd,  Co, Cr, Cu,
Fe, Hg, Mn, Ni, Pb, Sb,  Se, Si, V, Zn), analyzed by SSMS and graphite rod
atomizer AAS.  The samples were obtained at the inlet and the outlet of the
coal-fired power plant limestone wet scrubber.


When  SSMS  had  a  "greater  than"  number,  the upper  limit  listed was  used  to
determine  the  ratio  to  the AAS  value  (A).  Real number  to  real number
comparisons  are  shown as  (0).   Solid  circles  or  triangles  are data from
the inlet  while  the  open  circles  or triangles  represent  data from  the out-
let gas stream samples.   While  in the  qualitative  sense  the SSMS data ful-
filled its role, the quantitative comparison was not as  good.  Overall,
only  55% of  the  comparisons were  within a factor of 3.   Real number com-
parisons showed  that 85%  of  the SSMS values were within  a  factor of 3.
Inlet  SSMS results were within  a  factor of three 69% of  the time compared
to 44% for outlet values.

Comparison of  the 17 elements analyzed by both methods  shows good  agreement
(values within a factor of two) only  for inlet and outlet  concentrations of
Co, Ni, Sb,  Se,  and V.  Order of  magnitude agreement was obtained  for Ca,
Cd, Be, Fe, Mn,  Pb,  Sr, and  Zn.   SSMS  analysis of  other  elements,  such  as
A£, As, Cu and Cr, correspond to  the AAS values only to  within approximately
two orders of  magnitude.

Within the last  5 years a new method,  Inductively  Coupled  Plasma Optical
Emission Spectroscopy,  has attracted  a.  lot of  attention  by commercial
manufacturers  and the EPA.   ICAP,  the most commonly used abbreviation for
Inductively  Coupled A.rgon Plasma,  is  essentially a flame emission  tech-
nique where  atoms are thermally excited and emit light  at  discrete wave-
lengths.   The  atomization and excitation source  is an argon plasma torch
whose  temperature is on the  order of  9000°K.   In practice,  samples in the
form  of a  liquid are aspirated  into the argon  plasma torch, and  the emitted
light  is separated by a curved  grating and quantitated  by  preset photo-
multiplier tubes.  Figure 2  shows a block diagram  of a  plasma torch and
spectrometer.

Current commercial instrumentation employs up  to 60 photomultiplier tubes
so that 60 preselected  elements can be measured  simultaneously.  For the
quantitative analysis,  standard solutions and  calibration  curves have
                                    277

-------

O£
%
5
^i.
D

Z g
j^ KX-
^ II 1
X 03
ti
UJ ^
5 5
% <
0 g
u u
< o





































J_ <«<<««*
s a
sl
u n
S z
—I UJ
SO < 4 «
4



OOO*«*«««
III II 1 1 III

O O O O « •<
"
• -4
_






—



™
•
_
0 <

^HB




O
•
rx
A
0
•
«0
0
«n
o
•
•*
o
" 0
0 P
• "^
CM <
02
o «^
"" " ^ 5
^.
"0 to
0 5
8 «

0

•o
CM
0
o
CM
0
(X

•
o
N-
•
0
V



c
0)

H
o
c3
H
t>0
O
•u
CO
•H
43
e
o
w
•H
Vi
cd
a,
S
o
o
rf)
\Jj
$
^
CO
S
CO
CO
                                   o
278

-------
 to be made.   Since ICAP is a solution technique, solids have to be dissolved
 in acid and/or undergo a fusion and acid digestion to solubilize them.

 ICAP has a limit of detection on the order of 10~2 yg ml"1.  Table 1 lists
 some detection limits reported for some elements by ICAP compared to flame
 and furnace AAS.  Sensitivity limits compare favorably with AAS flame but
 are higher than furnace AAS.  However ICAP offers three orders of magnitude
 linear range for most elements, and the advantage of routine analysis for
 30-40 elements simultaneously.
                                   POLYCHROWTOR
                                                CALCULATING
                                                DEVICE AND
                                                PRIN7ER
       FIGURE 2   A block  diagram  of  a  plasma  torch  and  spectrometer
The EPA has extensively evaluated  ICAP  (3)  for use  in  drinking  and  waste-
water applications and late  this summer  the EPA  proposed  ICAP as  an
alternate method of analysis  for effluent guideline work.  Besides
demonstrating good accuracy  and sensitivity, cost per  analysis  for  40
elements ranges from $30 to  $80, depending  on the vendor.  Vendor avail-
ability is good and can be expected to improve as the  new generation of
less expensive ($60K) ICAP's  reach the market.

Illustrative of the use of elemental data from ICAP was the work  that  TRW
performed as part of a comprehensive assessment  program sponsored by
IERL/RTP (4).  During part of this program  an evaluation  was made of the
Firestone Tire and Rubber coal- and oil-burning  industrial boiler,  which
was partially controlled by a prototype  double alkali  venturi scrubber
with a cyclonic mist elimination FGD system.  The test program  included
sampling at the inlet and outlet gas streams of  the scrubber as well as
selected liquid or slurry streams.  Both coal and oil  firings were  con-
ducted so that a comparison of the fuels could be made.

The elemental data obtained from the gas sampling tests showed  that all
of the trace elements studied were removed  to some degree, as would be
expected from the mass removal data which showed 99% and  84% efficiency,
respectively,  for coal and oil firings.
                                   279

-------
                                  TABLE 1

                Comparative Detection Limits (ppb) (vg/L) (2)
    Element               ICP             Flame AA            Furnace AA
A£
As
B
Cd
Co
Cr
Cu
Fe
Mh
Mo
Ni
P
Pb
Pt
Se
Si
Ti
U
V
Zn
10
15
2
1
2
2
2
1
0.5
5
5
30
15
20
15
10
1
75
2
1
20
100
1000
1
5
3
2
5
3
10
8
105
10
50
100
60
50
7000
20
0.6
0.004
0.060

0.008
0.03
0.005
0.008
0.003
0.004
0.06
0.02
3.0
0.03
0.45
0.10
0.10
0.30

0.15
0.0007
We were interested to determine whether the scrubber affected the composition
of the particulate matter.  To study this question the individual elemental
concentrations (ng/J) were ratioed to a selected element.  The element
selected was aluminum since it was part of the basic fly ash matrix and
would not be expected to be partitioned among the different particle sizes.
The ratio of an element to aluminum was termed a concentration factor.
The next step was to ratio the outlet concentration factor to the inlet
concentration factor (enrichment ratio).  An enrichment ratio (ER) greater
than one indicated that the outlet particulate matter was enriched with
respect to that element while an ER < 1 indicated a depletion of that
element in the matrix.   The ER's were converted into percentage enrichment
[(ER-1)100%] and displayed as a histogram (Figures 3 and 4).

It was found that As, Cr, Mn, Ni, Sb, Se, V, and Zn were enriched in the
outlet particulate matter for both the coal and oil firings.  Questions
arose as to why these elements were enriched and why the overall range
of the percentage enrichment was different between the coal-and oil-fired
tests.

One clue to these differences was the particle size distribution of the
inlet and outlet  particulate matter (Table 2).  In both cases the major-
ity of particles that were emitted by the scrubber were < 1 ym while the
                                   280

-------
 \s

 8
X
                                                                                               (U
                                                                                               to
                                                                                              TJ
                                                                                               (U
                                                                                               J-l
                                                                                              •H
                                                                                               o
                                                                                               u
                                                                                               w
                                                                                               0)
                                                                                      —      u
                                                                                               cfl
                                                                                               G.
                                                                                               3
                                                                                               O
                                                                                               4J
                                                                                               C
                                                                                               0)

                                                                                              I
                                                                                               o
                                                                                              •H
                                                                                               M

                                                                                               §
                                                                                               0)
                                                                                               O
                                                                                               cfl
                                                                                              W
                                                                                              &

                                                                                              g
                                    281

-------
0
•—
1 z '
T
x— — «v
0
d
o

^
z
LU
U
z
o
u
1—
II 1

D
O
>^
c£
O
d
2
Z
O
^^
i
LU
U
0
u

	
LU
Z
^s

II
t—
z
tj^J
X
u
2
z
LU
v^
































% \ r5






























o






























6
N








^
















>.



M^MMMWBM







Z




o



01
O
U.
L.
oo


3
^••MMMM

^


C
S
f
CO
U





8
K? oJ
"*" 03
A n)
O
~ 13

-------
majority of the inlet particles were > 3 urn.  The largest difference
in the two particle size distributions occurred during the coal-firing
where the largest spread in the ER's was also found.
                                 TABLE 2

      Scrubber Inlet and Outlet Particulate Size Distribution
 Aerodynamic                               Weight %
 Diameter Size         Coal Firing (Test 201-1)   Oil Firing (Test 202-1)
 Range, Microns        Scrubber        Scrubber   Scrubber      Scrubber
                         Inlet           Outlet     Inlet         Outlet
< 1
1-3
3-10
> 10
0.0017
0.041
2.24
97.7
62
30
7
1
20
1
74
5
83
12
5
0
 Others (5,  6, 7) have noted these differences when large particles were
 preferentially removed by a control device.  They attributed the enrich-
 ment of selected elements to their increased concentration in the smaller
 particle size.  One way to confirm this hypothesis was to collect a sized
 particle sample and analyze each size fraction.  This was what we did at
 the coal-fired power plant FGD system.   In this case, the FGD system was a
 full sized (^ 100 MW) limestone scrubber with a venture/absorber combina-
 tion using Chevron mist eliminators.

 A variety of impactors are available for sampling flue gas streams.  An
 MRI 1502 impactor was selected, since the particles could be collected on
 a greased (Apiezon L) substrate (Kapton film) placed on the collection
 stage.  This substrate with the particles could be removed for elemental
 analysis without contamination from the 316 SS collection plate.  The samples
 collected in this fashion were analyzed for trace elements using Particle
 Induced X-Ray Emission (PIXE).  In Particle Induced X-Ray Emission the
 sample is prepared as a thin film and bombarded with a beam of protons or
 alpha particles which cause the characteristic X-ray emission (fluores-
 cence)  from all the elements in the sample.  The X-rays are detected by
 an energy-sensitive semiconductor device and the data are processed by
 an on-line computer.  Figure 5 shows a diagram (8) of the PIXE unit at
 the Crocker Nuclear Laboratory at the University of California, Davis.
 PIXE can detect elements Na through U,  but the sensitivity (nominally
 ppm) depends on the element and matrix.  In practice a piece of the
 impactor substrate containing one impaction spot is mounted in a photo-
 graphic slide and sent for analysis.

 Impactor samples were taken at the outlet of the FGD and the impactor data
 was reduced using the procedure (8,9) developed by Southern Research
 Institute (SoRI) for IERL/RTP.  In this approach [dM/d(log E>5())] the data

                                    283

-------
                                                              en
                                                              >-,
                                                              CO

                                                             w
                                                             X
                                                              CO
                                                             •H
                                                              to
                                                             •H
                                                              a
                                                              n
                                                              o
                                                             4-1
                                                             •H
                                                             rH
                                                              cd
                                                             CJ

                                                             4-1
                                                              O
                                                             J-J
                                                             •H
                                                              W
                                                              M
                                                              0)

                                                             •5
                                                              a
                                                             U">

                                                             W



                                                             O
284

-------
is normalized so that a smooth curve can be drawn through the limited
number of data points obtained from the impactor.  The normalized mass
data placed in the dM/d (log DC,-)) format is shown in Figure 6.  The data
for Run 135 was corrected for a negative filter weight using the weight
percentage of the filter from Run 136.  The filter data point for 135 is
shown in brackets, but even without this correction the data agrees rather
well.  The slight offset in the cut points is due to the difference in the
impactor flow rate during the two runs.  In both cases the mass data
showed a bimodal distribution with an apparent maximum around 0.4 to
0.5 ym.

In contrast to the mass data, the elemental data obtained from the PIXE
analysis showed a maxima at slightly different particle sizes (Figure 7).
Other elements were detected, but only the elements with PIXE data for
all stages were presented.  Calcium, Pb, Zn, and Si had a maximum at 20 ym
while S was the only element to show a maximum at 0.5 ym.  Iron showed its
concentration maximum over the mass minimum.  The lower particle size
maximum for the elements corresponds closely to the total mass value
maximum (0.4 to 0.5 ym).  To determine if the elements exhibited a con-
centration increase with decreasing particle size, the elemental concen-
trations were ratioed to Si which, because it is a major component of fly
ash, should not have exhibited a concentration change over the particle
size range.  Figure 8 shows the result of this effort.  Sulfur showed a
consistently increasing concentration with decreasing particle size.  The
high concentration of sulfur on the filter might be due to H_SO, aerosols
being collected since the impactor was run at stack temperatures (+ 55°C),
which would not prevent the collection of H«SO. aerosols.  Calcium showed
a weak positive trend while Pb and Fe showed a moderately strong trend
to increasing concentration at a smaller particle size.  Zinc showed the
strongest trend of the metals paralleling sulfur throughout impactor
stages, but dropping off on the filter.

These data are similar to other researchers (8,9,10) who found that elements
like Zn, Pb, and S have increasing concentrations with decreasing particle
size.  As a result, those elements which tended to concentrate in the
smaller particles would show a concentration increase in the outlet particle
matter, since the FGD preferentially emits the small particles.  This result
would be seen as an enrichment ratio greater than one and would be most
apparent when the inlet particle distribution is skewed toward the large
particles.  In the Firestone Coal-firing, as discussed earlier, greater than
95% of the inlet particles were above 3 ym (Table 2).  As a result of this
reversal of the particle size distribution, the histogram (Figure 3) showed
a large increase in the enrichment ratios for selected elements.

The same reasoning would predict that if the inlet particle size dis-
tribution was spread evenly or skewed to the smaller (< 3 ym) particles,
then enrichment ratios would tend to be near one.  In the oil-fired Firestone
tests, the inlet particle distribution was weighted toward the smaller
particles, and the total mass loading was low.  Consequently, removal of the
large particles (> 3 ym) did not have as great an effect on the enrichment
ratios.  This fact was illustrated by Figure 4 which showed that most of the
elements exhibited little or no concentration enrichment.

                                    285

-------
   1000
    TOO
I
O
O


I
                                                             o RUN 136

                                                             A RUN 135
                               1.0                     10.0


                            GEOMETRIC AERODYNAMIC DIAMETER,
100.0
               FIGURE 6   Coal-fired power  plant outlet  impactor study
                                       286

-------
   10,000
    1,000
                   8
                   o
               0       a
                   A
               O

               A
                                  0
                                  a
                                • -Co
                                O -S
                                A.Ft
                                O-Zn
                                0-F.
                                ©•Si
                        1.0                10.0
                      GCQMETRIC AERODYNAMIC DIAMETER, f*
                                                          100.0
FIGURE 7    Elemental  concentration by particle  size
             for  fly ash from the  outlet of  an FGD
     100
   Z
   8
                   s   .
                                 g
                               • * Ca/Si
                               ° - S/Si
                               A - Fb/Si
                               O - Zn/Si
                               0 - F«/Si
                        1.0                 10
                      GEOMETRIC AERODYNAMIC DIAMETER,
                                                           100
    FIGURE  8
Element  to Si  concentration ratio by
particle  size
             287

-------
                             SURFACE ANALYSIS

Models have been proposed to explain the apparent concentration of certain
trace elements in the smaller particles.  Simply stated these models
involve the volatilization of elements during combustion, followed by con-
densation over the magnesium aluminum silicate matrix of the fly ash.   (The
explanation is not quite this simple - see Kaakinen(5), Smith(6) and
Ondov(6) for a more complete explanation).  One of the primary tools for
surface research is Electron Spectroscopy for Chemical Analysis (ESCA).

In ESCA, low energy X-rays illuminate the surface of the sample.  Absorp-
tion of these X-rays result in prompt emission of photoelectrons from
atomic orbitals whose binding energy is less than the energy of X-ray
photons.  The total energy of the incident X-ray photon to the first approx-
imation is equal to the energy required to remove the electron from the
atom (binding energy).  Additional energy is required for the recoil energy
of the electron and for the work function of the spectrometer, but both
represent less than 1% of the incident energy.  Energy in excess of the
binding energy appears as the kinetic energy of the electron when it leaves
the atom, and is measured by an electron spectrometer.  Subtracting the
kinetic energy of the electron from the energy of the incident X-ray leaves
the electron binding energy, which is only slightly modified by the work
function.  In commercial electron spectrometers, this subtraction is car-
ried out electronically so that a spectrum of binding energy versus electron
counts is produced.  Figure 9 shows a typical ESCA spectrum taken of fly
ash collected at the outlet of a Fluidized Bed Combustor (FBC).

ESCA is a surface sensitive technique because the photoelectrons cannot
travel great distances in solids without undergoing scattering and loss of
energy.  It is only those which originate near the surface that leave the
sample with their full compliment of energy and appear as part of a peak
in the ESCA spectrum.  Escape depth varies as a function of electron
kinetic energy and the type of material being observed.  It is normally in
a range 5-15 A for metals, 15-20 A for inorganic compounds, and 50-100 A
for organic materials.  Combining this surface analysis capability with Ar+
sputtering allows the chemist to remove individual layers and analyze suc-
cessive surface elemental concentrations.

Chemical information available in ESCA spectra beyond the elemental com-
position is due to the "chemical shift" effect.  Metals exhibit a shift to
higher binding energies as they move to higher oxidation states.  In this
fashion high resolution ESCA can tell the difference between species like
C or C03= and S=, S03= and S04=.  Figure 10 is an actual high resolution
ESCA spectrum at 100 A depth of 27 ym particles from the Battelle FBC
unit.  The surface spectrum only showed S+6, but after sputtering, a clear
layer or particle of S~2 was uncovered.  The surface layer of pure S0^=
was probably from the air oxidation of  the  sulfide compound.

To illustrate the depth profile capability of ESCA, Figures 11 and 12 sum-
marize the ESCA analysis performed on particulate samples from the La Cygne
power plant tests.  The samples were obtained from the inlet and outlet of
                                   288

-------
                                                                        a.
                                                                        6
                                                                        s-i
                                                                        o
                                                                        O
                                                                        O

                                                                       T)
                                                                        a)
                                                                       ,0

                                                                       TJ
                                                                        
-------
     I
    UJ
    I—

  «>5
  — 1/1
                                                                     C
                                                                     •H
                                                                     O


                                                                     I
                                                                     O
                                                                     O
                                                                     3  cu
                                                                     M-l tH
                                                                     rH  &
                                                                     3  S
                                                                     w  td
                                                                         w
€3  ^3


4-) rH
o m
cu

co  o
   4J

w  e
    o
G  O
o
•H T3
4-1  CU
                                                                      O 13
                                                                      co  0)
                                                                      CU  N
                                                                      )-i -H
                                                                        T3
                                                                     fi -H
                                                                      60 3
                                                                     •H ^H
                                                                     ffi M-l
290

-------
                                            •Co (OUT)
                                            O Co (IN)
                                            • s (OUT)
                                            D S (IN)
                                            A C (OUT)
                                            A C(IN)
                      69% OF INLET PART
                      17% OF OUTLET PART
                      97.4% REMOVAL OF
                         200
                          o
                          A
300
400
500
FIGURE  11   ESCA depth  profile analysis of X3 pm participate
           matter before and after an FGD
                          291

-------
                            8
                            I
                            8
                            CO
                            8
                            8
                                                CO
                                                r-l
 u
 cd
 A.

 0


CO
V
 O

 Vl Q
•H O
 CO Pn
 CO 
-------
the limestone FGD using an EPA Method 5  train with a cyclone  and  filter.
The data for Ca, C, and S are presented  as atom percentages,  normalized  to
100%, and ratioed to the A& concentration of 500 A.  These  are  relative
values, not absolute concentrations, since not all elements present  in the
particles are included.  Ratioing  the elemental concentrations  to the A£ con-
centration at 500 A accentuates any surface concentration and tends  to
normalize any analysis errors.

Relative concentrations for the Method 5 cyclone (particles >3  ym) showed
that the sulfur/aluminum ratios for both the inlet and outlet particles
were high at the particle surface and decreased with depth.   This type of
data would be expected if a volatile compound of S were adsorbed  or  con-
densed on the particles.  Note that the sulfur concentration  tended  to level
out above zero, indicating that the coated particle contained sulfur.
Later XRD analysis confirmed the presence of CaSO^ in the inlet and  outlet
cyclone catches.  The sulfur rich  surface layer was probably  formed  by
^2^04 condensing on the particles.  An opposite effect was  seen for
carbon; the C/A£ ratio for the cyclone samples increased with depth,
which might indicate that unburnt  coal particles were present in  the inlet
and outlet >3 ym particles.  The Ca/A& for the inlet and outlet cyclone
and inlet filter samples were very similar, both showed a slight  depletion
at the surface and roughly equivalent concentrations.  The outlet filter
showed a higher concentration of Ca, which, after 100 A, corresponded
nicely with the S concentration.

The relative carbon concentration  of the inlet filter «3 ym) particles
was much lower than the cyclone fraction and showed a sharp surface  depend
ence.  Interestingly, the C/A£ ratio went to zero after 400 A for the
inlet samples, but in the outlet filter  samples remained fairly constant
after the initial drop, which might indicate the presence of  carbon  con-
taining particles.  In this case it is possible that <3 ym particles of
unburnt coal (concentrated in the  outlet filter by removal of the large
particles by the FGD) or CaCC>3 (a  scrubber reaction product)  might be pre-
sent.  The S/A£ ratio indicated a  surface layer of sulfur rich  material
similar to that found in the cyclone catch.

The impactor samples from the outlet of  the FGD system at the La Cygne
were also analyzed by ESCA to determine  the relative surface  and depth pro-
file concentrations as a function of particle size.  An interesting  rela-
tionship was found when the Ca/S ratio was computed by particle size and
at various depths (Table 3).  The  first  two stages exhibited  mixed trends
that might be due to a mixture of  fly ash and a Ca-S compound coated with
H2S04.   The partictes in the 8.5 and 3.5 ym size fraction appear  to  be
primarily a Ca-S compound with little or no coating of 112804.   The
particles in the 1.8 ym fraction exhibits a medium coating of H^SC^,.  on a
Ca-S particle.  Finally the last three stages show a thick coating of
H2S04 and may represent H2SC>4 droplets formed when the H2S04  vapor enter-
ing the FGD was rapidly cooled in a stream of fine particles.   This  inter-
pretation would also explain the PIXE data which showed increasing sulfur
concentrations with decreasing particle size (Figure 8).
                                     293

-------
                                  TABLE 3

       ESCA Data - S/Ca Ratio by Particle Size (um) and Depth (A)

Depth (A)
Surface
100-500
500
Average Geometric Aerodynamic Diameter (ym)
53.9
2.78
4.55
-
20.2
3.23
2.50
-
8.5
-
0.83
1.16
3.5
1.35
1.28
1.00
1.8
5.56
1.52
0.93
0.9
7.69
7.14
4.76
0.51
-
6.25
5.00
0.29
-
11.11
2.94
    (-)  ESCA spectra not  run
                    COMPOUND IDENTIFICATION PROCEDURES

It is possible that future emission regulations might control specific elements
rather than gross physical parameters like TSP.  The level at which these
elements will be regulated will be selected in part by the toxicity of the
element emitted.  Because the specific compound in which the element
resides affects the toxicity of the element, information on the type of
compound being emitted must be included along with the total output of the
element.

While the inorganic analytical chemist has a variety of sources of infor-
mation, he does not have a universal identification technique like GC/MS.
Using accurate elemental and anion concentrations, a list of candidate com-
pounds can be developed to reduce the number of potential compounds and
estimate the expected concentrations.  Prior knowledge of the expected con-
centration for a specific compound will aid in the selection of the appro-
priate identification technique.  The primary compound identification tech-
niques for the inorganic analytical chemist are IR spectroscopy and X-ray
Diffraction (XRD).  Infrared spectroscopy has been a traditional supple-
ment to XRD for inorganic compound identifications.  Since XRD can only
identify crystalline materials, IR spectroscopy provides information on
many compounds which cannot be seen by XRD due to their amorphous nature.

Within the last ten years a revolution in IR spectroscopy has occurred
with the advent of Fourier Transform Infrared  (FTIR) instruments  (H) •
Both conventional and FTIR spectrometers produce a spectrum of intensity
versus wavelength, but there are drastic differences in how the spectrum
are produced.  The key component of a conventional  IR  is a monochromator
which sorts the polychromatic radiation into discrete  bundles of nearly
monochromatic radiation.  In FTIR a monochromator is replaced by a Michel-
son interferometer  (shown in Figure 13).  The  light beam is split and
falls on both a moving and  stationary mirror.  When the two beams are
recombined, they will either constructively or destructively interfere
with each other since a path difference has been added to one beam.  The
signal generated will be a  cosine wave, whose  strength will be  the sum of
all the cosine oscillations caused by the optical frequencies in  the source.
The interferogram  (Figure 14) generated is  related back to the  original
spectrum by a cosine Fourier transform pair.   In practice this  is a com-
plex operation and  requires a dedicated minicomputer.
                                    294

-------
   h
J  Z
3  u
§  9
r  o
z  r
                        Ul
   -CD
                                                60
                                                O
                                                l-i
                                                 0)
                                                 4J
                                                 C
                                                w
                                                Pi
                                                g
£
£ *
Is
£?





* X1
X
., <
s
/
' J
/
x a
fe
x ^
/ Q
1
X


F


. r
t 4
•^ tit
IU (O





       <


       c -
il
Ir
    a:
    O   uJ

    §   I
    Z   ^
                                                 (U
                                                 g
                                                 o
                                                 CO
                                                 a
                                                 •1-1
                                                 14-1
                                                 O


                                                 I
                                                 t-l
                                                 N)
                                                 ca
                                                 • H
                                                 O
                                                            g
       295

-------
The advantages (12) of FTIR outweigh its initial high cost through improved
sensitivity, resolution, and data manipulation.  Improved sensitivity is
obtained as a direct result of the increased energy throughput (Jacqunot's
advantage) available in an FTIR due to the elimination of slits.  An inter-
ferometer has a circular aperture of 50 mm in diameter compared to slit
area of M mm^ for conventional IR.

In FTIR a complete spectrum is taken in approximately 1 second which is
about the time it takes for a conventional IR to scan one wavelength.  From
this, one can see that for equivalent experiments the FTIR system is N times
faster than a dispersive spectrometer (where N is the number of resolution
elements in the spectrum) or that the FTIR has N^/2 more analytical sen-
sitivity than the despersive spectrometer for the same measurement time.
This multiplex or Fellgett advantage can be exploited by co-adding succes-
sive interferograms to increase the signal-to-noise ratio.  Since the signal
increases with the number scans and the noise with the square root of the
number of scans, the signal-to-noise ratio increases with the square root
of the number of scans.

The most important advantage for FTIR is that the spectrum is present in a
digital form and can be processed easily using the minicomputer.  In this
way, spectra can be co-added, subtracted, ratioed, or stored for future
use.  This allows a comparison of different spectra and permits small dif-
ferences in the spectra to be easily detected.

When sampling at the outlet of an FGD most often the mass loading is so
low that the total particulate mass collected is <200 mg.  This small
amount of material is also impacted onto the glassfiber filter making
sample removal for analysis next to impossible.  TRW has developed FTIR
procedures to circumvent these problems.

The general approach was to run each filter sample, standards on filters,
and blank filters using an Attenuated Total Reflectance (ATR) cell.  The
standards were prepared using the apparatus in Figure 15.  A known amount
of the standard was placed at the bottom of the jar and a pre-weighed fil-
ter (^8 cm dia.) was loaded into the open faced filter holder.  After the
apparatus was assembled, the vacuum pump was turned on to initiate the air
flow and to entrain the particles of the standard material.  After a period
of time, the filter was removed and re-weighed.  The mg/cm^ loading was
determined and a 10 x 50 mm section was removed from the filter for FTIR-
ATR analysis.  Figure 16 shows the FTIR spectra obtained for a CaSC>4
standard made in this fashion.  The broad bands at 975,, 775 and 450 cm~^
are due to the glassfiber filter and would interfere with later interpreta-
tions.  Using the subtraction capability of FTIR, the blank glassfiber
filter spectra was subtracted from the standard spectra.  The resulting
spectra (Figure 17) is equivalent to standard reference spectra for CaSC>4,
yet, having been run as the samples would be, it reflects any anomalies
introduced by the ATR procedure.  A series of spectra were run for CaSO^,
CaS03 and H2S04 at various surface loadings and stored in the FTIR's com-
puter for later spectral interpretation.
                                    296

-------
                        TO
           LAB AIR       VACUUM     LAB AIR
              I         PUMP
                                                RUBBER
                                                STOPPER
                                               FILTER
                                               OPEN FACE
                                               FILTER HOLDER
                                                WIDE MOUTH
                                               JAR
                                                STANDARD
FIGURE 15   Powder dispersion apparatus  used to prepare
           FTIR standards on glassfiber filters
                          297

-------
                                                                        CO
                                                                        tc
                                                                        uu
                                                                        Ul
                                                                        I
                                                                      §
                                                                      CO
si
S     8
888
8     ri     8
3DNVHIWSNVai
8
id
                                                             8    g
                                                                                  OJ
                                                                                  4-1
                                                                                  CD
                                                                                  w
                                                                                  cfl
                                                        O

                                                       13
                                                        M
                                                        cfl
                                                       T3
                                                        a
                                                        «
                                                        4-1
                                                        w
                                                       o
                                                       tn
                                                        cfl
                                                       u
                                                                                  u
                                                                                  QJ
                                                        W
                                                        Pi

                                                        O
                                        298

-------
                                                                                      (A
                                                                                      CC
                                                                                      Ul
                                                                                      CO
                                                                                      S
                                                                                      z
                                                                                      ui

                                                                                      i
I
8      S
£      8
I
88
8      S     5
 30NVJJ.IWSNVU1
88888
                                                                                              ,fl 0)
                                                                                              •U >
                                                                                              •H O
                                                                                              £ s
                                                                                       03 T3
                                                                                      TJ  C

                                                                                       cd  o
                                                                                       4-1  S~i
                                                                                       en  60

                                                                                       •* o
                                                                                      O  nj
                                                                                      CO  60

                                                                                      O  M
                                                                                          OJ
                                                                                      LH 4->
                                                                                       O iH
                                                                                         •H
                                                                                       0 MH

                                                                                       >-i  S-i
                                                                                       4-1  0)
                                                                                       O rd
                                                                                       OJ -H
                                                                                       CXM-I
                                                                                       en
                                                                                          en
                                                                                      pi  to
                                                                                                 60
                                                                                              w
                                                                                              Pi
                                                                                              p>
                                                                                              o
                                             299

-------
Sections of filter samples taken at the inlet and outlet of a limestone
FGD were run and corrected for the glassfiber filter background.   Since
we were interested in the differences between the inlet and outlet parti-
culate matter, the inlet FTIR filter spectra was subtracted from  the out-
let filter spectra.  The original inlet and outlet spectra as well as  the
difference spectra are displayed in Figure 18.  Any bands which appear in
the outlet spectra are either unique to the outlet spectra or represent
compounds which are present at higher levels in the outlet particulate
matter.  Once the difference spectra was generated, known percentages  of
the standard's (CaSCfy, CaSC>3 • 1/2^2° and l^SO^.) spectra were added to the
difference spectra.   By electronically "spiking" the difference  spectra
with the standard, unknown compounds were identified by noting which bands
increased in absorbance.  For example, Figure 19 shows the difference  spec-
trum and the addition spectrum after 3% of the 0.44 rag/cm2 CaSO^  standard
was added.  Note the increased absorbances at 1150, 1095, 670 and  620  cm"*
and compare those bands to the original difference spectrum.  Clearly  CaSO,
was present.  After the identification of CaS04 in the outlet material,
the next step was to quantitate the amount of CaSC>4 in the filter.

The original idea was to use the different surface loading standards to
develop a calibration curve.  It was found that the extention coefficient
decreased with increasing surface loading.  At first it was thought that
the amount of CaSO^. in each filter was incorrect.  However, after  review-
ing the results of some quantitative XKD work, it was decided that the
weights were correct.  The reason for the change in the extinction coeffi-
cient was related to the amount of sample which the FTIR beam sees in  an
ATR experiment.  At low surface loadings nearly a monolayer of sample  is
present, and since the beam interacts with those particles in contact with
the ATR plates, essentially all of the sample was "seen" by the beam.  As
the surface layers were increased due to increased loading, the actual
amount of sample in contact with the ATR plates decreased.  Consequently,
as the surface loading increases, the extinction coefficient decreases.

In order to circumvent this problem, the difference spectrum was  electroni-
cally "spiked" with varying (31, 3, 5 10%) amounts of the 0.13 mg/cm^  and
0.44 mg/cm^ CaSC>4 standard's spectra.  A standard additive curve was gener-
ated for each of the CaSO^ standard's surface loading.

The results are shown in Figure 20.  Clearly, different response  factors
were encountered, but in the final analysis the calculated values  for  CaSC>4
of 2.3 mg/cm^ and 3.1 mg/cm^ were within 15% of each other.

This work with FTIR illustrates the type of qualitative  and  quantitative
experiments which can be performed.  Additional work is needed to  develop
a library of spectra of scrubber reaction products.  These spectra would
be stored on tape or disc for ready access.  In fact this entire  approach
could be extended to selected hazardous or regulated substances to pro-
vide rapid, accurate and sensitive analysis for those compounds in environ-
mental samples.
                                   300

-------
                                                                       CU
                                                                      4-1
                                                                       cfl
                                                                       o   en
                                                                      •H   3
                                                                      4J   d
                                                                       5-1  -H
                                                                       rd   S
                                                                       a
                                                                          4-1
                                                                      O   0)
                                                                      O  H

                                                                           3
                                                                      4->   O
                                                                       0)  -^


                                                                       3   3
                                                                       O   V-)
                                                                          4-1
                                                                      "O   O
                                                                       d   
                                                                      pi
                                                                      H T)
                                                                      H  Pi
                                                                      fn  rt
                                                                       w
                                                                       Pi
                                                                       g
301

-------
                 «M
                                         CM
                                                      CO
                                                      cc
                                                      UJ
                                                      CD
                                                      UJ

                                                      I
IIS
 •        •       •

    13NVBUO68V
§      I       g
                                                                O
                                                                0)
                                                                cx
                                                                en

                                                                01
                                                              0) O
                                                             A fi
                                                              4-> QJ
                                                                M
                                                             14-1 Hi
                                                              O M-l
                                                                «
                              o oj
                             •H &
                             •P 4-1
                             •H
                             net O
                             13 4-1
                              a
                                T3
                              0) !-i
                             & flj
                             4J T3
                                q
                              oo to
                              C 4-1
                             •H M
                              !5
                              O  ^1
                             j: o
                              en co
                                03
                              tfl o
                                                              o S
                                                             *
                                                             H
                                                             W
                                                             Pi

                                                             g
                302

-------
        Ill
          30     20
                                                      30      40
                                 HG/CM2
FIGURE 20   Standard  addition curve for two different  CaSO   standards
                                                           4
                                303

-------
                                SUMMARY

The analysis of environmental samples is a complex task requiring the inte-
gration of various techniques.  Starting with elemental analysis, informa-
tion on the overall emission and composition of emitted particualte matter
can be determined.  It was found that certain elements (As, Cr, Mn, Ni, Sb,
Se, V, and Zn) exhibit a concentration increase in the particulate matter
emitted from the Firestone FGD system compared to the material entering a
FGD under both oil and coal-firings.  Impactor studies combined with ele-
mental analysis have shown that this enrichment is due to the preferential
emission of small particles «3 ym) which contain increased concentrations
of these trace elements.  Surface analysis using ESCA showed that particles
emitted from the La Cygne FaD system were coated with sulfur and carbon.

The materials emitted from an FGD can be identified using FT1R.  FTIR pro-
vides all of the information of conventional IR and adds the ability to
add, subtract, or ratio spectra to enhance differences or improve the sen-
sitivity.  Using CaSC>4 standards dispersed on glassfiber filters, FTIR
spectra were obtained and used to electronically spike the sample spectra.
CaSO^ was identified in the outlet filter and quantitated.  Sensitives
on the order of 0.1 yg/cm^ can be expected for FTIR.
                                REFERENCES
 1.  Lentzen, D. E., D. E. Wagoner, E. D. Estes, and W. F. Gutknecht,
     "IERL-RTP Procedures Manual:  Level 1 Environmental Assessment (Second
     Edition),11 EPA-600/7-78-201 (1979).

 2.  Kahn, H. L., S. B. Smith, and R. G. Schleecher, "Background and
     Development in Plasma Emission Spectroscopy," Amer. Lab, 11 (8), 65
     (1979).

 3.  Ronan, R., "Simultaneous Analysis of Liquid Samples for Metals by Induc-
     tively Compled Argon Plasma Atomic Emission Spectroscopy (ICAP-AES),"
     U.S. E.P.A., Region 5, Central Regional Laboratory, Chicago, Illinois.

 4.  Leavitt, C., K. Arledge, C. Shih, R. Orsini, W. Hamersma, R. Maddalone,
     R. Beimer, G. Richard, and M. Yamada, "Environmental Assessment of Coal
     and Oil Firing in a Controlled Industrial Boiler:  Volume II, Compara-
     tive Assessment," EPA-600/7-78-164b (1978).

 5.  Kaakinen, J. W., R. M. Jorden, M. H. Lawasani, and R. E. West, "Trace
     Element Behavior in Coal Fired Power Plant,"  Environ. Sci.  Tech,  9_ (9),
     862 (1975).

 6.  Smith, R. D., J. A. Campbell, and K. K. Nielson, "Concentration Depen-
     dence Upon Particle Size of Volatilized Elements in Flyash," Environ
     Sci. Tech, 13 (5), 553 (1979).
                                   304

-------
  7.   Ondov,  J. M.,  R.  C.  Razani,  and A. H.  Bierman,  "Elemental Emissions
      from  a  Coal  Fired Power  Plant.  Comparison  of a Venturi Wet  Scrubber
      System  with  a  Cold  Side  Electrostatic  Precipitator," Environ.  Sci.
      Tech, 13  (5),  598 (1979).

  8.   Ensor,  D. S.,  B.  S.  Jackson,  S. Calvert, C. Lake, D. Wallen, R. Nilan,
      K. Campbell, T. Cahill,  and  R. Flocchini, "Evaluation  of a Particle
      Scrubber  on  a  Coal-Fired Utility," EPA-600/2-75-074 (1975).

  9.   Harris, D. B., "Procedures for Cascade Impactor Calibration  and Opera-
      tion  in Process Streams," EPA-600/2-77-004  (1977).

 10.   Johnson,  J. W., G.  I. Clinard, L. G. Felix., and J. D. McCain, "A Com-
      puter Based  Cascade  Impactor  Data Reduction System," EPA-600/7-78-042
      (1978).

 11.   Koenig, J. L., "Application  of Fourier Transform Infrared Spectroscopy
      to Chemical  Systems," Appl.  Spectro.,  29_ (4), 293 (1974).

 12.   Griffiths, P.  R., H. J.  Sloane, and R. W. Hannah, "Interferometers Vs
      Monochromators:   Separating  the Optical and Digital Advantages," Appl.
      Spectro., 31 (6), 485 (1977).
                            ACKNOWLEDGEMENTS

The work described in this paper was done under several EPA and EPRI Contracts.
The original development of the Level 2 approaches was completed under EPA
Contract 68-02-2165, Mr. Frank Briden project officer.  The field testing at
Firestone and La Cygne was conducted under EPA Contract No. 68-02-2179, Mr.
Wade Ponder project officer.  The FTIR spectra were run under EPRI, Project
Number RP 1410-3, Mr. Dick Rhudy EPRI project manager.  The author also wishes
to thank Dr. Robert Julian of the Nicolete Instrument Corporation for running
the FTIR spectra.
                                  305

-------
             FTIR:   A TOOL FOR BOTH ORGANIC AND INORGANIC
            ANALYSES IN ENVIRONMENTAL ASSESSMENT PROGRAMS

                   R.  L.  Barbour and R.  J.  Jakobsen
                    Battelle Columbus Laboratories
                               ABSTRACT

Fourier Transform infrared spectroscopy (FTIR) can characterize both
organic and inorganic species in a given sample.   Application of this
technique to environmental assessment programs and various sample hand-
ling methods are discussed.  Two developing areas, inorganic compound
speciation and on-the-fly GC/FTIR, are discussed in relation to their
potential contributions to the field of environmental assessment.

                             INTRODUCTION

Fourier Transform infrared spectroscopy (FTIR) is a versatile tool in
the field of environmental assessment because it is the only method by
which both organic and inorganic compounds can be analyzed.  With this
capability, FTIR can be utilized in either overview-type programs such
as Level 1 or can characterize samples compound by compound.

                          INORGANIC ANALYSIS

The ability to identify inorganic compounds in environmental samples
places FTIR in a position of considerable importance.  Currently,
most inorganic analyses of samples for environmental assessment provide
only elemental information, or at best, anion and cation concentrations.
In rating the potential hazard of an effluent, this type of information
is not sufficient; the form which an element or ion takes, to a great
extent, determines the health effects it may have.  This problem may
be directly addressed by the use of FTIR for inorganic species infor-
mation.

Figure 1 shows the spectrum of an ambient air particulate; in this case,
one was collected by the National Bureau of Standards as their standard
urban sample.  This sample underwent no sample preparation whatsoever
aside from grinding and mixing with potassium bromide (KBr) in order
to produce a KBr pellet.  In addition, only about 1 mg of sample is
necessary to produce a spectrum.  Interpretation of the spectrum reveals
the presence of the following:

    bound water   (3300-3600 cm"1)      carbonate      (1410-1450 cm"1)
                  (1600-1650 cm"1)
    ammonium      (2800-3300 cm"1)      sulfate        (1090-1150 cnT1)
    hydrocarbon   (2800-3000 cm"1)      silicate       (1000-1080 cm"1)
    nitrate       (1380-1390 cirr1)      metal oxides   ( 400-600  cm'1)

Thus, in fewer than fifteen minutes from receipt of sample through inter-
pretation of the spectrum, a good deal of information concerning the
sample has been obtained.
                                    306

-------
Often programs endeavor to compare effluents.   A single spectrum of
the samples In question will aid such a comparison.   Figure 2 shows the
spectra of four ambient air particulates, from as-received samples,
obtained in four cities.

Baton Rouge (the top spectrum) consists of almost pure ammonium sulfate.
Upland (the bottom spectrum), on the other hand, has a low sulfate con-
centration (band near 1100 curl), ancj tne sulfate is not ammonium sul-
fate.  In this sample, a relatively large amount of nitrate (1385 cm"1)
is present.  Birmingham and Charleston both show significant levels of
sulfates (^1100 cm"1) and silicates (^1040 cm"1) although in different
ratios and varying types of sulfates.   These two samples also contain
carbonate (^1430 cm"1) and nitrate (1335 cm"1).

If comparative data are desired, a comparison of the spectra of the
unfractionated samples is a very good place to start.  If, however,
more specific inorganic compound information is needed, we have had
good success using a simple water extraction to separate species for
analysis.  Figure 3 shows the spectrum of the water insoluble portion
of the NBS sample.

This spectrum indicates predominately silicates (^1040 cm"1), probably
complexed with metal oxides whose presence is indicated by bands below
600 cm"1.  In addition, from spectra of this fraction, we can obtain
a relative measure of the crystallinity of the silicate by the intensity
and splitting of the band near 800 cm"1.  Silicate and metal oxide species
are seen in virtually all water insoluble fractions of environmental
samples.

The real utility of this water extraction is in removing the above
mentioned species so that we can isolate the water solubles, which
typically contain such species as nitrates, sulfates and carbonates.
In particular, the spectrum of the NBS sample (see Figure 4) indicates
OH", NH^"1", C03= and N03~ functionalities.  From this spectrum we can
positively identify CaSO^ as the major component of this fraction by the
position of the S-0 stretching vibrations (1100-1150 cm"1) and by the
more characteristic position of the S-0 bending bands between 600 and
700 cm"1.  We also know from the spectrum that CaSO^ is present as the
dihydrate by the shape and position of the OH stretching and bending
bands near 3500 and 1600 cm"1.

Another sample handling method that we have found helpful is the splitting
of a sample or fraction and heating one-half at 200°C for 3 hours, leaving
the other half unbaked.  This procedure both "cleans up" the spectrum by
removing volatiles and extraneous water and also for many compounds will
cause them to assume a reproducible hydration state.  The spectral changes,
which occur with hydration state changes, tend to be reproducible and
quite informative.  In Figure 5 are two spectra:  the lower is the
methanol extract residue of the as-received NBS; the upper is the methanol
extract residue of a portion of the NBS sample which had been baked as
described above.  The methanol extraction was originally performed in
order to isolate the more polar organics.  While organic material is
certainly present in the unbaked sample (bands near 2900 cm  ), the
number and broadness of bands below 1700 cm"1 suggests a substantial
inorganic content and sorting out inorganic versus organic species in
                                   307

-------
this setting is difficult.   The spectrum of baked residue shows almost
pure ammonium sulfate.   This accounts for all of the sulfate but only
part of the nitrate and parts of the other two major bands.   Now, with
knowledge of the polarity of the solvent, the remaining bands can be
assigned to short chain carboxylic acid and ammonium nitrate.  Ammonium
chloride may also be present.

In the past, it was thought that interactions of sample with potassium
bromide made the pellet technique a poor one.  While irjteractions do
occur with some substances, we have found that the KBr pellet produces
very good results with most inorganics and that in fact, on our instru-
ment, frequency reproducibility is excellent (1).

To illustrate this reproducibility, we had several people prepare pellets
of CaSO/ in KBr over a two-week period and ran spectra on two FTIR systems.
We chose KBr pellets for this experiment because this is how we most
commonly run environmental samples.  The results are summarized in Table 1.
In all three of the band positions shown, reproducibility was within
2 cm-1.

Using this precision, we have created our own computerized spectral search
system.  Sulfates, carbonates and nitrates have been run as received as
well as after dissolution in water and heating.  That is, the reference
compounds are treated in the same manner as the samples will be.  Three
parameters are entered into the system for each band in both reference
and sample spectra:  band position (frequency), band intensity and band
width at half height.  In addition, each reference band is assigned a
priority, thus utilizing the skill of the spectroscopist.

Figure 6 shows the results of the spectral search for the spectrum of the
unbaked water soluble fraction which we saw previously.  Note that the
compound given the highest ranking (that is, considered most likely to
be present) is CaSO^ which we have already identified as the major com-
ponent of this fraction.  Note that all compounds are designated as
either U or B denoting either the unbaked or baked condition.

The real power of this spectral search is shown in looking at more minor
components.  For instance, ZnSO^ is listed in the eighth most likely
position, but is it present in the sample or not?  First, we note that
in the unbaked sample, the computer sees unbaked ZnSO^.  Below, the
bands used to achieve the match are summarized.  On the computer match,
this looks very good, but referring back to the actual spectrum  (Figure 4),
we see that in the complex mixture of S-0 stretching and S-0 bending
bands the match is hardly conclusive.  For corroborating evidence, we
look at the results of the spectral search on the spectrum of the baked
water soluble fraction  (Figure 7).  If ZnS04 is present, a match should
now be made with ZnSO^ in the baked condition and indeed this match is
seen in Figure 7.  The match is more convincing because band positions
of the baked sample have shifted as compared to the unbaked sample  (as
shown in the lower section of Figure 7), and still match.  The presence
of ZnS04 is further supported by elemental analysis indicating both Zn
and SO^" in this fraction.  Thus, the computer assisted spectral inter-
pretation program can greatly speed the interpretation of the spectra.

                                    308

-------
We have also found that for many inorganics pelleting in KBr produces
spectral bands whose intensities are linear with concentration.  Figure 8
plots three bands of L^CC^ in intensity versus concentration in KBr.
All three band intensities were found to be quite linear with concentration,
certainly well within our gravimetric error.  Not only does this raise
the possibility of quantitation of inorganics at the source, but the
ease of obtaining data is truly impressive; all pellets were prepared,
spectra obtained and the graph produced by the computer in the period
of about one hour.

                           ORGANIC ANALYSIS

So far, only inorganic compounds have been considered, yet organics are
an important part of the sample and are more traditionally the province
of infrared spectroscopy.  Easily extracted from the bulk of the sample,
organics can be analyzed separately, leaving the inorganics for separate
treatment.  After extraction, a fast, efficient method of obtaining
class information is the Level 1 liquid chromatography separation of the
sample into seven fractions of increasing polarity, after which infrared
spectra are obtained of each fraction.  One of the classic problems of
this method, however, is the difficulty in obtaining clean separations;
that is, there can be a great deal of overlap between fractions, making
interpretation of spectra more difficult.  An example of this is seen in
Figure 9.  The top spectrum is fraction 3 from a typical Level 1 SASS
train sample.  There is aliphatic hydrocarbon seen in conjunction with
what could be two carbonyls and some aromaticity, but it is impossible
to assign which bands go with which, or even make more than a rough
guess at how many compounds are present.  Fraction 2, seen at the bottom
of the figure, can easily be identified as an ester.  Comparing the two
spectra, fraction 3 appears to possibly contain fraction 2.  A subfraction
(the middle spectrum) reveals that not only is the ester a constituent
of fraction 3, but that on digital removal of the ester, a quinone can
quite easily be identified as the other compound present in this fraction.

As with inorganics, a cleaner separation is needed for compound identifi-
cations to be made.  In the case of organics, the low concentrations of
materials may cause problems due to detection limits.  Isolation of these
minor components may also be difficult, however, two techniques of promise
are HPLC/FTIR and GC/FTIR.  We are developing GC/FTIR in our laboratories
and GC/MS is already widely used in environmental assessment.  Since
both of these techniques are useful in some settings, we decided to com-
pare the two (2).  The three points to be considered are:

     (1)  sensitiv.ity
     (2)  speed and ease of operation
     (3)  amount of information

Figure 10 shows a gas chromatogram of a wastewater sample run on a packed
column with a flame ionization detector (FID).  Although this particular
work was done on packed columns, we have used both wall-coated and
support-coated capillary columns on line with the FTIR system.  Each
number represents an infrared spectrum obtained at that position.
During the most intense peak (aside from the solvent peak), three spectra
were taken:  numbers 2, 3 and 4, corresponding to the leading edge, center
and trailing edge of the peak.   Figure 11 shows these three spectra.
                                    309

-------
The spectra are seen to be of quite good quality,  and they are quite dis-
tinct, representing the para, meta and ortho isomers of chloronitrobenzene.
This ability of FTIR to isolate isomers, which GC/MS cannot do, has impor-
tant applications in environmental assessment, as  quite often only one
of several isomers is toxic.   In addition, IR can  provide functional
group information which is useful even if the exact molecule cannot be
identified.  This also is a useful advantage of GC/IR over GC/MS.

Table 2 summarizes results obtained for this sample by the two techniques:
GC/FTIR and GC/MS.  Obviously, the two are complementary; on the one
hand, GC/IR can give isomer information that MS cannot.  On the other
hand, GC/MS can elucidate compound structure where  GC/IR gives only func-
tional group and class type information such as dibutyl phthalate versus
alkyl phthalate.  Some of the GC/FTIR identifications are marked as ten-
tative.  This is because, as yet, the library of vapor phase infrared
spectra is limited to several thousand compounds and identifications must
often be made using condensed phase reference spectra which may be
drastically different from spectra in the vapor phase.  Here GC/MS has
a great advantage, at least temporarily.  Their reference library exceeds
that of GC/IR by more than a factor of 10 and computer software available
for GC/MS is currently more advanced that that available for GC/FTIR.

Returning to the three points of comparison, we can see that

      (1)  the quality of information from the two  techniques is
          very complementary in nature,

      (2)  GC/MS speed and ease of operation is much better than
          that of GC/FTIR for the time being, and

      (3)  GC/MS theoretically has a greater potential sensitivity.
          However, in complex samples, sensitivity is often matrix
          dependent and GC/FTIR under certain conditions will have
          equal or better sensitivity than GC/MS.

                                SUMMARY

In conclusion, the versatility of FTIR makes possible a variety of sample
handling techniques.  Information concerning both  inorganics and organics
can range from simple class determination to compound by compound charac-
terization of the sample.  Methods currently in developmental stages,
such as the computer assisted spectral interpretation program and GC/FTIR,
promise rapid expansion of application of FTIR to  the field of environ-
mental assessment.

                              REFERENCES

1.   Gendreau, R. M. and R. Burton, "The KBr Pellet:  A Useful Technique
     for Obtaining Infrared Spectra of Inorganic Species," Applied Spec-
     troscopy, 33, 581, November  6, 1979.

2.   Shafer, K. H.,  S. V. Lucas and R. J. Jakobsen, "Application of the
     Combined Analytical Techniques of HPLC/FTIR, GC/FTIR and GC/MS to
     the Analysis of Real Samples," J. Chrom. Sci. , 17_, 464, August 1979.
                                    310

-------


                                   O
                                   C
                                   O
                                   CN
                                   O
                                   O
                                   C
                                   C,
                                   CO
                                   (N
CO
g
                                           1
                                                     a,
                                                     6
                                                     cfl
                                                     en
                                                     Cfl
                                                     CL
                                                     C
                                                     
-------
Baton Rouge
   (a)
Birmingham
    (b)
 Upland
   (d)
Charleston
    (c)
              1600
1400
                                                             800
                        1200       1000

                         Wavenumbers

FIGURE 2  Spectra of four ambient  air  particulate  samples

                            312
600

-------
                                         c
                                         c
                                         OC
                                         o
                                         c
                                         CM
                                         c
                                         c
                                         c
                                         c
                                         CM
                                         o
                                         c
                                        c
                                        c
                                        OC
                                        CN
                                        c
                                        c
                                        CM
                                        o
                                        c
CO
o:
LU
                                                  £
        U-J
        o

        CJ
        o
        •H
        4J
        J-)
        o
        CX

        0)
        T-f  
-------

                                          .  I...
- -I
  tz.  ._-
                                                                                     o
                                                                                     o
                                                                                     00
                                                                                     o
                                                                                     c
                                                                                     o
                                                                                     c
                                                                                     c
                                                                                     o

                                                                                     03
                                                                                     CXI
                                                                                     O
                                                                                     o
                                                                                     o
                                                                                     o
                                                                                               00
                                                                                               OL
                                                                                               LU
                                                                                               LU
                                                                                                         C
                                                                                                         o
                                                                                                        •H
                                                                                                         O
                                                                                                         G.
                                                                                                            01
                                                                                                         o>  r-i
                                                                                                        rH  &
                                                                                                        ^  E
                                                                                                         3  HJ
                                                                                                        i-l  TO
                                                                                                         O
                                                                                                         tn  H
w  a
 to  u
 5 -H

 0)  U
X  to
4-1  P.


 O -H


 §  *
 1J  G
4J  CD
 O -H
 01 J3
 a  e
 en  ai

(Yj  PJ
H  nj
                                                                                                        Cd


                                                                                                        O
                                                   314

-------
                                                                          cd    •


                                                                     rd   01   a)

                                                                     4-1   CS  -H
                                                                     d  J3   01
                                                                     0)       u
                                                                     •T-4.   0)   01
                                                                     ,0  iH   !•<
                                                                     E   e*
                                                                     cd   E   CD
                                                                          cd   cd
                                                                          en
                                                                     Cd
                                                                              ^^
                                                                     tfl       ^
                                                                    ^  
-------
                     MATCHING FILE--UR34UUU
               THE BEST MATCHES WERE AS FOLLOWS
REFERENCE FILE

  CaS04-U
  FeNH4(S04)2
  A1NH4(S04)2-B
  NH4N03
  (NH4)2S04


  Fe2(S04)3-B
  ZnSO^-U
  MnS04-U
  CaS04-B


REFERENCE FREQ.
ZnSO^-U
  1142
   613
                      SCORE

                       417
                       281
                       279
                       210
                       202
                       165
                       154
                       152
                       149
                       140
                   SAMPLE FREQ.
1138
 602
                BASE SCORE
MATCHES
225
126
102
104
66
59
73
62
92
64



7
3
4
4
2
3
4
2
5
5
REF. INTEN.
3.0
3.0
RANK

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
                                           REF. PRIOR.
              4.0
              2.0
                           #34 HO Soluble Unbaked
             * B=Baked sample, U=Unbaked sample
      FIGURE 6  Results of the spectral search for the spectrum
                of the unbaked water soluble fraction.
                                   316

-------
         TABLE 1

Spectra of CaSO^ Pellets
     Prepared in KBr
CaSO^
Date Run
4
4
4
4
4
5
5
5
4
4
4
4
4
4
/10
/I
/I
/I
12
12
/3
/A
/I
12
12
3
5
7
3



7
0
3
/24
725
725
Instrument
FTS-
FTS-
FTS-
FTS-
FTS-
FTS-
FTS-
FTS-
10
10
10
10
10
10
10
10
FTS-14
FTS-
FTS-
14
14
FTS-14
FTS-
FTS-
14
14
• 2H20
Band
1
1
1
1
1
1
1
1
1
1
1
1
1
1
622.
266.
624.
621.
621.
621.
622.
621.
621.
621.
621.
621.
621.
621.
1
2
2
2
3
3
3
2
3
3
3
3
3
3
3
Band
668
666
668
669
668
668
668
668
668
668
668
668
668
668
2
.4
.5
.4
.3
.4
.0
.0
.4
.4
.4
.4
.4
.4
.4
Band
601
599
599
598
600
600
600
600
600
599
602
601
602
602
3
.8
.9
.9
.9
.9
. 9
.9
.9
.9
.9
.8
.8
.8
.8
           317

-------
                         MATCHING FILE—URA21
                   THE BEST MATCHES WERE AS FOLLOWS
REFEREECE FILE
   SCORE
BASE SCORE
MATCHES
CaS04-B*
MgS04-U
MnS04-U
MgS04-B
ZnS04-U
ZnS04-B
Alk(S04)2-U
Na2S04-U
285
222
202
192
189
180
179
172
REFERENCE FREQ.

   ZaS04-B

    1142
    1101
    1101
     631
SAMP. FREQ.
   1152
   1115
   1098
    629
146
93
110
75
92
80
79
68
REF. INTEN.
3.0
3.0
3.0
2.0
7
4
5
5
5
4
3
3





    RANK

     1
     2
     3
     4
     5
     6
     7
     8

REF. PRIOR.
                               3.0
                               4.0
                               4.0
                               3.0
                              #21 H20 Soluble baked
         *  B=Baked sample, U=Unbaked sample
           FIGURE 7  Results of the spectral search of a baked
                     water soluble fraction.
                                     318

-------
C/3
H
M

Z



w
o
CO
PQ
CO
z
w
H
z
o
z
to
       1.2.
1.0
                     BATTELLE COLUMBUS LABS


                           L12C03
                                   CONCENTRATION IN MG/GM
                   *  BAND AT 1435 cm
                                     -1
                   +  BAND AT 864.0 cm
                                      -1
                                               X  BAND AT 1088.0 cm'
                                                                   -1
                    FIGURE  8   Intensity versus  concentration of

                               Li2C03  in KBr.
                                      319

-------
 __ .FRACTION 3
       FRACTION 3 i
        ; MINUS ;   '  '
i_L_i.  FRACTION 2
        i  i   .
3000
                   2000
                WAVENUMBERS
FIGURE 9  IR spectra of Level 1 LC  fractions
                      320

-------
cc

cc
UJ
o
IT
O
O
UJ
cc
   100 i-
   90
   80 5
   70
   60
50 -
   40
   30 -
   20 -
   10 -
                 (2,3,4)
         2468   10  12  14  16  18  20  22  24  26 26

                        MINUTES


     FIGURE 10   Chromatogram  of GC  separation
                         321

-------
                                                             c
                                                             o
                                                             4-1
                                                             3
                                                             M
                                                             G
                                                             •H
                                                             C  cd
                                                             •H  (11
                                                             cd  CX

                                                             n  cj
                                                             cti  CL,
                                                             S-i  tfl
                                                             4-1  M
                                                             O  60
                                                             01  O
                                                             CX  4J
                                                             cn  cfl

                                                             pi  o
                                                             M  S-l
                                                                f
                                                             cu  a
                                                             >
                                                             •H  Cfl
                                                             cn  cd
                                                             to  bO
                                                             OJ
                                                             O  CO
                                                             o
                                                             3 U-i
                                                             CO  O
                                                             w
                                                             O
NOISSIWSNVMi -a
                  322

-------
                     CO
                     C
                     O
                    •H
                    4J
                     cd
                     CJ
                    •H
                    4-1
                    •H
                    4-1
                     C
                     0)
                    T)
         M
        •H
                               0)
                               C
                               QJ
                               N
                               C
                               0)
                               43
                               O
                               S-i
                               4-1
                               •H
                               0
                               O
                               r-l
                               O
                               rH
                               43
                               O


                               4-1
                               O
                               a)

                               o
                               CO


1)
C
(1)
N
C
0)
43
O
Vi
4->
•H
C
O
t-l
O
r-f
^
CJ
•H
O


0)
C
•H
r-l
•H
C
rt
0
u
4-1
•H
C
O
S-i
o
rH
43
O
•H
Q

0)
C
0)
N
a
CU
43
0
M
4J
•H
C
O
n
0
rH
43
O
•H
VI
H
yl phthalate
N
ff
0)
43
01 rH
4-1 >,
CO 4-1
rH 3
cd 43
43
4-1 M
43 0
P.
rH
rH ^
>-. N
•U f3
3 .
c
01
r-
ex
•H
!-j
H
ion of abietic ac
4-1
ct)
a
•H
IH
•H
4J
C
Q)
T3
•H

0)
>
•H
4-1
cd
4-1
ti
0)
H
         rd
         C3
CN

w
hJ
PQ
<
H
a
cd
        o

        4-1
         o

         W
         4J
        rH
         3
         CO
         0)
GC/FTIR Identifications
0)
J-^
0)
N
a
01
43
O
M
O
iH
43
O
•H
Q
1
CU
zene
c
0)
o
O
r-l
4-1
•H
C
O
r-l
0
rH
43
U
1
H
zene
pj
0)
43
0
M
4-1
•H
a
O
M
0
rH
r4.
O
1
a.
zene
c
01
43
0
Vi
4J
•H
C
O
^
o
rH
43
U
|
0
obenzene
vi
4J
•H
C
O
>-i
O
rH
^r-;
CJ
•H
O
1
~^-
*K
01
fication of 2,4~dichloro-6-nitroaniline
•H
4J
C3
0)
T3
•H

O>
[>
•H
4-1
ct)
4-1
C
0)
H
fication of either trichloronitrobenzene
•H
4-1
C
0)
tl
•H

0)
p>
•H
4J
cd
4-1
C
0)
H
onitroaniline
V-J
0
rH
f*.
a
•H
rrj
1
CO
A
CN

^
O




0)
4J
cd

cd
43
4-1
rr^
p,

rH
r^~l
^
, — |
<3
                                                                          CO
                                                                          43
                                                                          O.
                                                                          CO
                                                                          O
                                                                          43
                                                                          a
                                                                          rH
                                                                          >,

                                                                          0)
                                                                          43
                                                                          a,
                                                                          •H
                                                                          vi
of
on
                                                                                   en
                                                                                   o
                                                                                  •H
                                                                                   CD
                                                                                   cd
                                                                                   4-1
                                                                                   C
                                                                                   0)
                                                                                   H
                                                       323

-------
                 MEASUREMENT METHODS FOR SOX AND NOX IN
                          THE PRESENCE OF NH3

                      D.  S.  Chase and B. M.  Myatt
                            GCA Corporation
                          Technology Division

Ammonia reacts with SOV and NOY to form various ammonium sulfur and
                      A.       A.
nitrogen compounds.  These compounds can precipitate in standard sampling
systems and obstruct the flow of sample gas yielding invalid results.
Moreover, most of the analytical procedures used in conjunction with
SOX, NOX and NH^ Standard Method determinations are adversely affected
when all three of these components are present in the gas stream.  This
paper presents the results of a series of laboratory studies in which
modifications to both the sampling systems and the analytical procedures
were made in order to adequately measure these parameters.

Various levels of ammonia were injected into a sample stream containing
typical stock concentrations of SOX, NOX and water vapor.  This sample
stream was used to determine the effects of NH3 on the Methods.  Sampling
train modifications including precipitate traps, NH3 and SO^ scrubbers
were implemented.  Alternate analytical procedures compatible with these
complex stream were developed.
                                   324

-------
              EVALUATION OF STABLE LABELED COMPOUNDS AS
              INTERNAL STANDARDS FOR QUANTITATIVE GC/MS

                              B. N. Colby
                     Systems, Science and Software

Quantitation of organic compounds by GC/MS using stable labeled compounds
as internal standards is a widely accepted method of preference in many
research areas.  The reasons for this are all related to the in situ
recovery compensation which is built into each analysis and the resulting
high accuracy which this provides.  These analyses have typically been
empirically optimized for a single compound of interest and, although
the technique is simple, attempts to expand the approach for multiple
component applications have been very limited.  The reasons generally
expressed for not expanding this technology to the multiple component
analysis generally encountered in process measurement studies for environ-
mental assessments, revolve around potential high costs and poor avail-
ability of labeled materials, possible problems with label exchange,
difficulty in identifying proper experimental parameters, variable label
purity of commercially available compounds and propagation of errors
during data reduction.  Each of these potential problem areas has been
investigated in detail for a selected group of compounds, the organic
priority pollutants, and the results indicate that an expansion of this
methodology to certain types of environmental assessment measurements
could lead to improved analytical results at reduced cost.  The results
of these studies will be given in this presentation.
                                  325

-------
                   LEVEL 1 ENVIRONMENTAL ASSESSMENT:
                       FLUIDIZED-BED COMBUSTION

                      R. R.  Hall, P. F.  Fennelly,
                      G. T.  Hunt and R.  J.  Kindya
                            GCA Corporation
                          Technology Division

Environmental assessment samples were collected at the EPA-sponsored
Exxon Miniplant, a pressurized fluidized-bed combustor, in May 1979.
Level 1 chemical and biological tests have been completed and results
are presented.

Improved particulate control has almost eliminated potential trace
element emission problems.  Level 2 tests that will focus on positive
mutagenicity results in the Ames Salmonella Microsome assay are outlined.
                                  326

-------
               EVALUATION OF LEVEL 1 ANALYSIS PROCEDURES

                  R. K. M. Jayanty, W. F. Gutknecht,
                  A.  Gaskill,  Jr.  and D.  E.  Lentzen
                      Research Triangle Institute

The current environmental assessment  (EA) program of the Industrial
Environmental Research Laboratory, Research Triangle Park (IERL-RTP),
is designed to yield data that will result in the identification of
sources of recognized pollutants and other substances of potential
environmental concern.  As a part of the IERL-RTP Quality Assurance
Program for assessing data quality, two components of the EPA process—
the analytical methods and their application—were evaluated.  Eight
contractors were supplied with four simulated organic and inorganic
source assessment samples for analysis.   The four audit samples were
(1) a five-component synthetic organic mixture chosen to evaluate the
Level 1 organic measurement techniques of IR, LRMS, TCO, GRAV, and LC;
(2) a commercial dye mixture selected to test the LC scheme and to be
tested as a general procedure for assessing the performance of the
Level 1 LC column; (3) a modified fly ash sample selected to test the
SSMS and cold vapor atomic absorption techniques; and (4) a clean blank
XAD-2 resin selected to check the completeness of Parr bomb ashing of
the resin and contamination of the sample during the ashing process.

The specific objectives of this audit program and the results reported
by the contractors will be presented.  Problems experienced by the par-
ticipating contractors for both organic and inorganic analyses using
Level 1 analytical techniques will be discussed.  In general, the prob-
lems identified with organic analyses are data reduction errors, loss
of volatile substances, incomplete drying for GRAV analysis, and con-
tamination or misinterpretation of spectra.  Several problems identified
with SSMS, including incorrect reporting of results, quantitation of
detection limit values, and low recoveries for Be and Cd, will be dis-
cussed.  The accuracy, precision, and overall acceptability of the
Level 1 analysis procedures will be addressed and recommendations for
improving the data quality will be presented.
                                   327

-------
              ANOMALOUS HIGH TOTAL CYANIDE RESULTS DUE TO
           NITRITE IN BIOTREATER EFFLUENTS:  THE KEY TO THE
        "CYANIDE-GENERATION SYNDROME" IN BIOTREATER TECHNOLOGY

                     R. A. Johnson, J. C. Rapean
                           and T.  P.  Hanson
                         Shell Development Co.

The Total Cyanide Method, involving the distillation of HCN from dilute
sulfuric acid, is notorious for yielding erratic results.  Even in
laboratories with the method "in control", sporadic excursions occur in
an unexplained way.  This behavior is especially troublesome in the
determination of cyanide in feeds to and effluents from industrial waste-
water biotreaters.  These Total Cyanide results take on an extra dimen-
sion of importance when they are involved in NPDES reporting.

In biotreater technology, a "cyanide-generation syndrome" is sometimes
observed in which the Total Cyanide measured in the effluent greatly
exceeds that in the feed and results in NPDES permit violations.  The
explanation of the sporadic apparent generation of cyanide in certain
biotreaters has weighed heavily on the concerned analytical laboratories
and biotreater operators.

The key to this problem is now being found in the anomalous high results
for Total Cyanide due to the reaction of nitrous acid with certain organic
compounds to produce HCN.  The nitrous acid is variably produced by
microbial nitrification in biotreaters.  The variations in nitrite in
the samples parallel the variations in the Total Cyanide found.  This
analytical anomaly is avoided by the reduction of nitrous acid with
sulfamic acid prior to the distillation step.  This effect was observed
and described by S. Kanno and co-workers in Japanese journals in 1974
and 1975^- and, in English, in Chemosphere^ more recently.  However, this
anomaly has apparently remained unobserved in this country until recently.
Acting on Kanno's work, Shell Development Company modified its Total
Cyanide Method for wastewater and was able to solve the mystery of the
"Cyanide-Generation Syndrome" in a biotreater at one of the Shell plants.

Independently, Jeremiah Casey and co-workers of Air Products Co.,3 redis-
covered the effect and solved a similar problem.

This poster session is provided to spread the understanding of this
solution to a perplexing problem in the critical area of cyanide discharges
in treated wastewaters.
 1.  S. Kanno et al., J. Hyg. Chem. Japan, 21  (1) 1  (1975).

 2.  S. Kanno et al., Chemosphere, No. 8, 657-662 (1978).

 3.  J. Casey et al., to be published in Analytical  Chem. in 1980.

                                   328

-------
                      DESIGN MODIFICATIONS TO THE
                   SOURCE ASSESSMENT SAMPLING SYSTEM

                     R.  L. Campbell, K. W. Mason,
                     W.  R. Parker and T. J. Wagner
                       PEDCo Environmental, Inc.

Recently, there has been an interest in improving the Source Assessment
Sampling System (SASS) from a user and performance standpoint.  Based
on earlier designs, a complete SASS train was designed and built with
major modifications incorporated.  These major modifications include a
larger filter holder, separation of the filter and cyclones into separate
ovens, a complete glass organic module, using a refrigeration unit to
control the sample gas temperature, and larger impingers equipped with
all Teflon stems and connections.  Field use of the train indicates
significant improvements in operation.  However,  no comparative analytical
data are available at this time.  Recommendations for further design
considerations are made.
                                    329

-------
          DIAGNOSIS OF METAL SPECIATION IN AQUEOUS SOLUTIONS*
                                                Jf A
                       B. McDuffie and P.  Figura
              State University of New York at Binghamton

A method for differentiating soluble trace metal species on the basis
of relative lability has been developed for solutions of Cd, Cu, Pb and
Zn utilizing anodic stripping voltammetry (ASV) and Ca-Chelex resin in
successive column and batch procedures.  Species are classified as being
"very labile", "moderately labile", "slowly labile" or "inert", depending
on the characteristic time scale of the measuring technique.  Although
ASV has limitations, most heavy metals can be put into categories based
on Chelex uptake.  A range of dissociation rate constants can be estimated
for metal complexes in each category.  Preconcentration of the various
fractions leads to precise determinations even at ppb levels.  The method
has been applied to water from the St. Lawrence and Susquehanna Rivers,
to a Hudson River estuary sample, and to secondary sewage effluent.  In
spite of the variety of samples, reasonable distinctive patterns of spe-
ciation were observed.  Cadmium and zinc occurred almost entirely in
the "very labile" and "moderately labile" fractions.  Copper was found
primarily in the "moderately labile" and "slowly labile" fractions, no
"very labile" copper being found.  Lead showed some "very labile" species,
20-70% "slowly labile" material, and a significant "inert" fraction in
several cases.  Factors related to application of this method to process
measurements will be explored.
*
 Work supported in part by USDI - OWRT Matching Grant B-070-NY.
**
  Present Address:  Stauffer Research Center, Dobbs Ferry, NY.
                                  330

-------
                    SOLID SORBENTS FOR AIR SAMPLING

             J. F. Piecewicz, J. C. Harris and P. L. Levins
                        Arthur D. Little, Inc.

This paper gives results of an ongoing experimental program to evaluate
the breakthrough characteristics of sorbent resins for sampling of organic
vapors using an elution analysis chromatographic technique.  The effects
of water vapor and C02, at levels typical of gaseous effluents from
combustion processes, on retention of nonpolar and polar species on
two commonly used sorbents (XAD-2 and Tenax-GC) have been studied.
Effects on XAD-2 were small but the volumetric capacity of Tenax-GC
was substantially decreased.   Other sorbents were characterized for
potential use in vapor sampling systems:  coconut-based charcoal, petro-
leum-based charcoal, silica gel, Ambersorb XE-340 and XE-347, and XAD-8.
Retention volumes for XAD-8 and silica gel were roughly comparable to
those of XAD-2 and Tenax-GC;  those of the charcoals and the Ambersorbs
were 2 to 4 orders of magnitude higher.  Recovery of sorbate from charcoals
is known to be poor in some cases, however, and recovery from the Ambersorbs
is suspect but under investigation.
                                   331

-------
                  SCREENING - BIOLOGICAL AND CHEMICAL
                             DATA ANALYSIS

                   N. H. Sexton and L. I. Southerland
                      Research Triangle Institute

The Industrial Environmental Research Laboratory (IERL) has developed
a scheme for the assessment of the environmental hazard of effluents
and emissions from industrial sources.  This assessment scheme is com-
posed of a screening step (Level 1), a detailed, quantitative analysis
step (Level 2), and a long-term monitoring step (Level 3).  This presen-
tation deals with Level 1 which incorporates a battery of bioassays
developed in a cooperative program with EPA Health and Ecological Labora-
tories, and a group of chemical analyses.  These bioassays examine the
health and ecological effects of varying levels of chemical by-products
from four Level 1 studies.  Three Level 1 pilot studies involved a coal
gasifier, a fluidized-bed combustor, and a number of different aqueous
textile plants.  The fourth study involved a variety of industrial sources.

The basic approach for statistical analysis was to quantify, on a con-
tinuous or discrete  (1,2,...) scale, the possible biological responses,
and then to correlate the biological results with the amount of chemical
present and correlate biological measures with other biological measures.
                                   332

-------
              ANALYSIS OF COAL LIQUID SUBFRACTIONS WHICH
                 EXHIBIT MICROBIAL MUTAGENIC ACTIVITY

                     B.  W. Wilson, R. A.  Pelroy,
                   M.  R. Peterson and W.  C.  Weimer
                 Battelle Pacific Northwest  Laboratory

The chemical composition and biological activity  of a coal liquefaction
material was compared before and after catalytic  hydrogenation.   Chemical
analysis has employed solvent fractionation, gas  chromatography/mass
spectrometry and high pressure liquid chromatography and has been com-
plemented by Ames assay for microbial mutagenic activity.  In contrast
to the predominately aromatic nature of the  principal components of the
coal oil, the hydrogenated products contain  mainly hydroaromatic species
and a significantly smaller fraction of basic and tar components.  The
microbial mutagenic activity of moderately and severely hydrotreated
projects was reduced substantially to 2.8 percent and 0.8 percent,
respectively, when normalized to that of  the coal oil feed.   These
changes in biological activity have been related  directly to the content
of primary aromatic amines in the individual coal liquefaction materials.
                                   333

-------
            QUANTITATION OF POLYCYCLIC AROMATIC HYDROCARBONS
         IN COMPLEX MIXTURES BY HIGH RESOLUTION GLASS CAPILLARY
   GAS CHROMATOGRAPHY/MASS SPECTROMETRY USING SELECTED ION MONITORING

                     G. A. Gibbon and C. M.  White
                       U.S. Department of Energy
                  Pittsburgh Energy Technology Center

Quantitation of individual polycyclic aromatic hydrocarbon isomers in
complex mixtures is a difficult task.  Previous quantitation schemes
have involved time consuming, manpower intensive preseparation of samples
to produce an aromatic fraction which was subsequently quantitatively
analyzed by capillary gas chromatography or  capillary GC/MS using external
or internal standards.  These methods require the use of standards not
present in the original sample and a knowledge of the relative response
factors of the compounds of interest and the internal standards.   One
frequently neglected problem with this approach is that these relative
response factors can change as a function of sample injection technique
and the adsorptive characteristics of the chromatographic column.  Thus,
these methods often lead to quantitative results of poor precision and
accuracy.

We will describe the application of the method of standard additions to
quantitative PAH analysis by high resolution glass capillary gas chroma-
tography /mass spectrometry and selection ion monitoring in samples where
no preseparation has been performed.  This technique provides quantitative
values for PAH of high precision and accuracy even at the ppb concentration.
                                  334

-------
             ANALYTICAL RESULTS OF A PCB TEST INCINERATION

                      C. D. Wolbach, W. F. Fitch,
                        N. Flynn and B. Markoja
                          Acurex Corporation

Results of a test incineration of PCB-containing waste material are
presented.  A discussion of the analytical problems in quantitating
PCB-containing samples is given with emphasis of the DCB methodology
of armour.  A listing of other identified species of environmental
significance is offered.
                                   335

-------
               MANUSCRIPT AND POSTER SESSION AUTHOR INDEX

Manuscript Authors                                               Page
    Auyong, M	199
    Baldwin, R. P	169
    Harbour, R. L	305
    Bause, D. E	207
    Bodek, 1	155
    Bombaugh, K. J	    74
    Brusick, D. J.  	226
    Cairns, Jr., J	183
    Gate, Jr., J. L	199
    Chrisp,  C. E	241
    Cohen, M. J	    41
    D'Silva, A. P	117
    Dubay, G. R	134
    Farnum,  B. W	    87
    Farnum,  S. A	    87
    Fassel,  V. A	117
    Fisher,  G. L	241
    Gammage, R. B.     	119
    Gaskill, Jr., A	    55
    Giddings, J. M	104
    Guerin,  M. R	     1
    Gutknecht, W. F	    55
    Harris,  J. C	    41
    Jakobsen, R. J	305
    Kolber,  A	    17
    Knudson, C. L	    87
    Lee,  K.  W	    74
    Lewis, D.  S	    74
    Maddalone, R. F	275
    Martinez,  P. R	119
    McGregor,  K. T	255
                                   336

-------
Manuscript Authors (Cont'd)
                                                                 Page
    Menzies, K. T	155
    Natusch, D. F.  S	240
    Nichols, D. G	    17
    Petty, J. D	_	134
    Price, J. F	169
    Rueppel, D. W	199
    Schaeffer, D. J	143
    Siria, J	169
    Smith, L. M	134
    Stalling, D.  L	134
    Thompson, K.  W	183
    Tigwell, D. C	143
    Vo-Dinh, T	119
    Wilkie, M. B	    17
    Williams, C.  H	    74
    Wilson, F. D	241
    Wolff, T. J	    17
    Yang, Y	117
Poster Session Authors
    Campbell,  R. L	328
    Chase, D.  S.    	323
    Colby, B.  N	324
    Fennelly,  P. F	325
    Figura,  P	329
    Gaskill, Jr.,.A	326
    Gibbon,  G. A	333
    Gutknecht, W. F	326
    Hall, R. R	325
    Hanson,  T. P	327
    Harris,  J. C	330
                                   337

-------
Poster Session Authors (Cont'd)
   Hunt, G. T	325
   Jayanty, R. K. M.	326
   Johnson, R. A	327
   Kindya, R. J	325
   Lentzen, D. E	326
   Levins, P. L	330
   Mason, K. W	328
   McDuffie, B	329
   Myatt, B. M	323
   Parker, W. R	328
   Pelroy, R. A	332
   Peterson, M. R	332
   Piecewicz, J. F	330
   Rapean, J. C	327
   Sexton, N. H	331
   Southerland, L. 1	331
   Wagner, T. J	328
   Weimer, W. C	332
   White, C. M	333
   Wilson, B. W	332
                                   338

-------
                          APPENDIX

                   SYMPOSIUM PARTICIPANTS


Auyong, Marjorie
     Lawrence Livermore Laboratory, PO Box 5505 L-385, Livermore,
     California  94550     Phone:  415-422-5174

Baldwin, Richard P.
     Department of Chemistry, University of Louisville, Louisville,
     Kentucky  40208     Phone:  502-588-5892

Barbour, Rachael L.
     Battelle Columbus Laboratories, 505 King Avenue, Columbus,
     Ohio  43201     Phone:  614-424-5078

Barrett, William J.
     Southern Research Institute, 2000 Ninth Avenue South,
     Birmingham, Alabama  35205     Phone:  205-323-6592

Bause, Daniel
     GCA/Technology Division, Burlington Road, Bedford, Massachusetts
     01730     Phone:  617-275-5444

Beimer, R. G.
     TRW, Inc., One Space Park 01-2020, Redondo Beach, California
     90278     Phone:  215-535-1458

Benson, Janet M.
     Lovelace Biomedical & Environmental Research Institute, PO Box 5890
     Albuquerque, New Mexico  87115     Phone:  505-844-2207

Borman, Stuart A.
     American Chemical Society, Room 702, 1155 16th Street NW,
     Washington, DC 20036     Phone:  202-872-4584

Bramlett, Royce N.
     Applied Biology, Inc., 641 DeKalb Industry Way, Decatur,
     Georgia  30033     Phone:  404-296-3900

Branscome, Marvin R.
     Research Triangle Institute, PO Box 12194, Research Triangle
     Park, North Carolina  27709     Phone:  919-541-5862

Briden, Frank
     U.S. Environmental Protection Agency, IERL/PMB, Mail Drop #62,
     Research Triangle Park, North Carolina  27711
     Phone:  919-541-2557

Brusick, David J.
     Litton Bionetics, Inc., 5516 Nicholson Lane, Kensington, Maryland
     20795     Phone:  301-881-5600

                              339

-------
Burns, Eugene A.
     Systems, Science & Software,  PO Box 1620,  La Jolla,  California
     92038     Phone:  714-453-0060

Cairns,  John
     Virginia Polytechnical Institute and State University,
     Blacksburg, Virginia  24061     Phone:  703-961-5539

Chait, Edward M.
     E.I. du Pont de Nemours & Co., Photo Products Department,
     AI&BP Division, Concord Plaza, Quillen Building,  Wilmington,
     Delaware  19898     Phone:  302-772-5609

Clark, Bruce R.
     Oak Ridge National Laboratory, PO Box X, Oak Ridge,  Tennessee
     37830    Phone:  615-574-4861

Colby, Bruce N.
     Systems, Science & Software,  PO Box 1620,  La Jolla,  California
     92038     Phone 714-453-0060

Cooke, Marcus
     Battelle Columbus Laboratories, 505 King Avenue,  Columbus,
     Ohio  43201     Phone:  614-424-5024

Cowen, Stanton J.
     Meteorology Research, Inc., 464 W. Woodbury Road, Box 637,
     Altadena, California  91001     Phone:  213-791-1901

Craig, Alfred Boyce
     U.S. Environmental Protection Agency, IERL/PMB, Mail Drop #62,
     Research Triangle Park, North Carolina  27711     Phone:  919-541-
     2509

Crumrine, Kenneth Z.
     Versar, Inc.,  6621 Electronic Drive, Springfield, Virginia
     22151     Phone:  703-750-3000

Cullum, Terry A.
     Shrader Analytical Laboratory, 3450 Lovett Avenue, Detroit,
     Michigan  48210     Phone:  313-894-4440

Denton, Mark S.
     Oak Ridge National Laboratory, PO Box X, 4SOON, Room B8,
     Oak Ridge, Tennessee  37830     Phone:  615-574-6837

Dippel, William A.
     E.I. du Pont de Nemours & Co., Experimental Station Laboratory -
     Building 336/Room 130, Wilmington, Delaware  19898     Phone:  302-
     772-4106
                                  340

-------
Dorsey, James
     U.S. Environmental Protection Agency, IERL/PMB, Mail Drop #62,
     Research Triangle Park, North Carolina  27711     Phone:  919-
     541-2557

Drewitz, Karen
     Arthur D. Little, Inc., Acorn Park, Cambridge, Massachusetts
     02140     Phone:  617-864-5770

Duke, Kenneth M.
     Battelle Columbus Laboratories, 505 King Avenue, Columbus,
     Ohio  43201     Phone:  614-424-7574

Duvoid, Daniel
     Ministry of Environment, 14 Bd. Gal Leduc,  92521 Neuilly,
     France    Phone:  758 12 12

Ellis, Stephen P.
     Southern Company Services, Inc., PO Box 2625, Birmingham,
     Alabama  35202     Phone:  205-870-6918

Fanning, Leah Z.
     Lawrence Berkeley Laboratory, Building 70,  Room 223, One Cyclotron
     Road, Berkeley, California  94720     Phone:  415-486-4767

Farnum, Bruce W.
     Grand Forks Energy Technology Center, U.S.  Department of Energy,
     Box 8213 University Station, Grand Forks, North Dakota  58202
     Phone:  701-795-8159

Fassel, Velmer A.
     Ames Laboratory, Iowa State University, Ames, Iowa  50010

Fisher, Gerald F.
     University of California, Davis, Davis, California  95616
     Phone:  916-752-7774

Fitchett, Arthur W.
     Dionex Corporation, 104 Alnick Court, Durham, North Carolina
     27712     Phone:  919-383-2539

Ford, Judith S.
     U.S. Environmental Protection Agency, IERL/PMB, Mail Drop #62,
     Research Triangle Park, North Carolina  27711     Phone:  919-
     541-2557

Foster, Robert B.
     E.G. & G. Bionomics, 790 Main Street, Wareham, Massachusetts
     02571     Phone:  617-295-2550

Gaskill, Alvia
     Research Triangle Institute, PO Box 12194,  Research Triangle
     Park, North Carolina  27709     Phone:  919-541-6743
                                341

-------
Giddings, Jeffrey
     Oak Ridge National Laboratory, PO Box X, Oak Ridge, Tennessee
     37830     Phone:  615-574-7337

Goodman, Randall G.
     University of North Carolina/Chapel Hill, Department ESE,
     School of Public Health, Chapel Hill, North Carolina  27514
     Phone:  919-933-2151

Gray, D. Anthony
     Syracuse Research Corporation, Merrill Lane, Syracuse, New
     York  13210     Phone:  315-425-5100

Guerin, Michael R.
     Oak Ridge National Laboratory, PO Box X, Oak Ridge, Tennessee
     37830     Phone:  615-574-4862

Gutknecht, William F.
     Research Triangle Institute, PO Box 12194, Research Triangle
     Park, North Carolina  27709     Phone:  919-541-6883

Hall, Robert R.
     GCA/Technology Division, Burlington Road, Bedford, Massachusetts
     01730     Phone:  617-275-9000

Hamersma, J. Warren
     TRW Systems & Energy, One Space Park, Building R4-2142,
     Redondo Beach, California  90254     Phone:  213-535-1544

Harris, Judith C.
     Arthur D. Little, Inc., Acorn Park, Cambridge, Massachusetts
     02140     Phone:  617-864-5770

Heiden, Richard W.
     Owens/Corning, Technical Center, Granville, Ohio  43023
     Phone:  614-587-7283

Higginbotham, E. B.
     Acurex Corporation, 485 Clyde Avenue, Mountain View, California
     94042     Phone:  415-964-3200

Howe, Lyman H.
     TVA Laboratory Services, 150-401 Chestnut Street, Chattanooga,
     Tennessee  37401     Phone:  615-755-3137

Jacobs, M. L.
     GT & E, Instrumentation and Analysis Division, 490 Orchard Street,
     Golden, Colorado  80401     Phone:  303-526-1057

Jakobsen, Robert J.
     Battelle Columbus Laboratories, 505 King Avenue, Columbus, Ohio
     43201     Phone:  614-424-5617


                                342

-------
Jayanty, R. K. M.
     Research Triangle Institute, PO Box 12194, Research Triangle
     Park, North Carolina  27709     Phone:  919-541-5934

Johnson, Larry D.
     U.S. Environmental Protection Agency, IERL/PMB, Mail Drop #62,
     Research Triangle Park, North Carolina  27711     Phone:  919-
     541-2557

John son, Irving
     Argonne National Laboratory, 9700 South Cass Avenue, Argonne,
     Illinois  60439     Phone:  312-972-4384

Johnson, Ralph A.
     Shell Development Corporation, PO Box 1380, Houston, Texas
     77001     Phone:  713-493-7263

Jones, Peter W.
     Environmental Assessment Department, Electric Power Research
     Institute, 3412 Hillview Avenue, PO Box 10412, Palo Alto,
     California  94303     Phone:  415-855-2737

Jung, Arthur D.
     Versar, Inc., 6621 Electronic Drive, Springfield, Virginia
     22151     Phone:  703-750-3000

Kidd, George
     Battelle Columbus Laboratories, 505 King Avenue, Columbus,
     Ohio  43201     Phone:  614-424-6317

Kline, Eugene A.
     Tennessee Technological University, Department of Chemistry,
     PO Box 5055, Cookeville, Tennessee  38501     Phone:  601-528-3423

Kolber, Alan
     Research Triangle Institute, PO Box 12194, Research Triangle
     Park, North Carolina  27709     Phone:  919-541-6521

Kolnsberg, Henry J.
     TRC Environmental Consultants, Inc., 125 Silas Deane Highway,
     Wethersfield, Connecticut  06109     Phone:  203-563-1431

Koppenaal, David W.
     Institute for Mining and Minerals Research, PO Box 13015,
     Lexington, Kentucky  40583     Phone:  606-252-5535

Lee, Kenneth W.
     Radian Corporation, 8500 Shoal Creek Blvd., Austin, Texas
     78758     Phone:  512-454-4797

Lentzen, Donald E.
     Research Triangle Institute, PO Box 12194, Research Triangle
     Park, North Carolina  27709     Phone:  919-541-6745
                                 343

-------
Levins, Philip L.
     Arthur D. Little, Inc., Acorn Park, Massachusetts  02140
     Phone:  617-864-5770

MacLeod, William D.
     U.S. Department of Commerce, NOAA, 2725 Montlake Blvd.  E.,
     Seattle, Washington  98112     Phone:  206-442-4240

Maddalone, Ray F.
     TRW, Inc., DSSG, One Space Park 01/2171, Redondo Beach,
     California  90278     Phone:  213-536-2447

Manissero, Claudio E.
     Interox America, PO Box 1000, Deer Park, Texas  77536
     Phone:  713-479-2381

Mason, Howard
     Acurex Corporation, 485 Clyde Avenue, Mountain View, California
     94042     Phone:  415-964-3200

Mason, Wade
     PEDCo Environmental, Inc., 11499 Chester Road, Cincinnati,  Ohio
     45246     Phone:  513-782-4800

McCaskill, Kenneth B.
     U.S. Department of Energy, METC, PO Box 880, Morgantown, West
     Virginia  26505     Phone:  304-599-7749

McDuffie, Bruce
     Laboratory for Trace Methods & Environmental Analysis,
     Department of Chemistry, State University of New York at Binghamton
     Binghamton, New York  13901
     (1980 Sabbatical Leave - U.S. Environmental Protection Agency,
     Environmental Research Laboratory, College Station Road, Athens,
     Georgia  30605     Phone:  404-546-3592)

McElroy, Francis C.
     EXXON Research & Engineering, PO Box 121, Linden, New Jersey
     07036     Phone:  201-474-3954

McGregor, Kenneth
     GCA/Technology Division, Burlington Road, Bedford, Massachusetts
     01730     Phone:  617-275-5444

Menzies, Kenneth T.
     Arthur D. Little, Inc., Acorn Park, Cambridge, Massachusetts
     02140     Phone:  617-864-5770

Merrill, Jr., Raymond G.
     U.S. Environmental Protection Agency, IERL/PMB, Mail Drop #62,
     Research Triangle Park, North Carolina  27711     Phone:  919-
     541-2557


                                  344

-------
Miller, Herbert C.
     Southern Research Institute, 2000 Ninth Avenue South, Birmingham,
     Alabama  35205     Phone:  205-323-6592

Mills, Paul
     U.S. Environmental Protection Agency, 26 W. St. Clair,
     Cincinnati, Ohio  45268     Phone:  573-684-4216

Mumford, Judy
     U.S. Environmental Protection Agency, Mail Drop #68, Health
     Effect Research Laboratory, Research Triangle Park, North
     Carolina  27711

Murawczyk, Carlos
     United Engineers & Constructors, 30 South 17th Street,
     Philadelphia, Pennsylvania  19101    Phone:  215-422-4410

Myatt, Barbara M.
     GCA/Technology Division, Burlington Road, Bedford, Massachusetts
     01730     Phone:  617-275-5444

Natusch, David F. S.
     Colorado State University, Fort Collins, Colorado  80523
     Phone:  303-491-5391

Olson, Dean L.
     University of North Carolina/Chapel Hill, 212 Craige Hall,
     Chapel Hill, North Carolina  27514     Phone:  919-933-2153

Panzer, Jerome
     EXXON Research & Engineering Company, PO Box 51, Linden, New
     Jersey  07036     Phone:  201-474-3127

Parcher, Jon F.
     University of Mississippi, Chemistry Department, University,
     Mississippi  38677

Piecewicz, John F.
     Arthur D. Little, Inc., Acorn Park, Cambridge, Massachusetts
     02140     Phone:  617-864-5770

Pope, Rodney
     Engineering-Science, 57 Executive Park South, Suite 590,
     Atlanta, Georgia  30329     Phone:  404-325-0770

Quilliam,  Michael A.
     McMaster University, Department of Chemistry, Hamilton, Ontario
     CANADA  L8S4M1     Phone:  416-525-9140

Samant, H. S.
     Environmental Protection Service, 16th Floor, Bank of Montreal
     Tower, 5151 George Street, Halifax, Nova Scotia, CANADA  B3J 1M5
     Phone:  902-426-6237

                               345

-------
Sasseville, Dennis
     Normandeau Associates,  Inc.,  25 Nashua Road,  Bedford,  New
     Hampshire  03102     Phone:  603-472-5191

Sferra, P. R.
     U.S. Environmental Protection Agency,  IERL/CINN,  Ridge Facility,
     26 West St. Clair, Cincinnati, Ohio  45268     Phone:   513-684-4252

Sharkey, A. G.
     Pittsburgh Energy Research Center,  4800 Forbes Center, Pittsburgh,
     Pennsylvania  15213     Phone:  412-675-5000

Siczek, Aldona A.
     Argonne National Laboratory,  9700 South Cass  Avenue, Argonne,
     Illinois 60439     Phone:   312-972-4378

Sides, Gary D.
     Southern Research Institute,  2000 Ninth Avenue South,  Birmingham,
     Alabama  35205     Phone:   205-323-6592

Sorlin, Debra J.
     Arthur D, Little, Inc., Acorn Park, Cambridge, Massachusetts
     02140    Phone:  617-864-5770

Southerland, Leslie I.
     Research Triangle Institute,  PO Box 12194, Research Triangle
     Park, North Carolina  27709     Phone:  919-541-6742

Stalling, David L.
     U.S. Department of the  Interior, Fish and Wildlife Service,
     Columbia National Fisheries Research Laboratory,  Route 1,
     Columbia, Missouri  65201      Phone:  314-442-2271

Steiber, Raymond S.
     U.S. Environmental Protection Agency,  IERL/IPD/PMB, Mail Drop  #62,
     Research Triangle Park, North Carolina  27711     Phone:  919-541-
     2288

Stiles, David A.
     Acadia University, Wolfville, Nova Scotia, CANADA BoPlXo
     Phone:  902-542-2201

Theis, Thomas L.
     University of Notre Dame,  Department of Civil Engineering,
     Notre Dame, Indiana  46556     Phone:   219-283-6247

Tigwell, David C.
     Illinois Environmental  Protection Agency, 401 E.  Colorado,
     Urbana, Illinois  61801     Phone:   217-384-5398

Trzcinski, Kristofer
     Studsvik Energiteknik AB,  S-611 82 NYKOPING,  SWEDEN
     Phone:  155/80 000

                                  346

-------
Tsoukalas, S.
     U.S. Forest Service, PO Box 5106, Macon, Georgia  31208
     Phone:  912-744-0236

Varga, Gideon M.
     EXXON Research & Engineering, PO Box 45, Linden, New Jersey
     07036     Phone:  201-474-3311

Vo-Dinh, Tuan
     Oak Ridge National Laboratory, PO Box X, Oak Ridge, Tennessee
     37830     Phone:  615-574-6249

Wagner, Thomas J.
     PEDCo Environmental, Inc., 11499 Chester Road, Cincinnati,
     Ohio  45246     Phone:  513-782-4634

Wagoner, Denny E.
     212 Beachers Brook Lane, Gary, North Carolina  27511
     Phone:  919-467-3355

Weimer, Walter C.
     Battelle Pacific Northwest Laboratory, 329 Building/300 Area,
     Richland, Virginia  99352     Phone:  509-942-3995

White, Curt
     U.S. Department of Energy, Pittsburgh Energy Technology Center,
     4800 Forbes Avenue, Pittsburgh, Pennsylvania  15213
     Phone:  412-675-5000

Wilson, Judith A.
     TRW, Inc., One Space Park, Redondo Beach, California  90278
     Phone:  213-535-1233

Winslow, Frank J.
     Monsanto Research Corporation, Station B., Box 8, Dayton, Ohio
     45407     Phone:  513-268-3411

Wolbach, C. D.
     Acurex Corporation, 485 Clyde Avenue, Mountain View, California
     94042     Phone:  415-964-3200

Zee, Carol
     TRW, Inc., One Space Park, Redondo Beach, California 90278
     Phone:  213-535-1233
                               347
                                                *USGPO: 1982—559-092/3384

-------