United States Environmental Protection Agency

  Office of Water Regulations and Standards
        Industrial Technology Division
      NINTH ANNUAL ANALYTICAL SYMPOSIUM
              Norfolk,  Virginia

              March 19-20,  1986

-------
                                   FOREWARD
       The Industrial Technology Division of the USEPA Office of Water Regulations
and Standards sponsors EPA's Annual Analytical Symposium to provide a forum where
scientists and other interested parties can  present new ideas and  advances in method-
ology for the analysis of pollutants in the environment.

       Prior  symposia have focused primarily on  the determination  of  pollutants in
wastewater.  The Ninth Symposium  expanded  the  scope of  analytical testing  to the
determination of pollutants in wastes, soils, sediments, and other sample matrices using
wastewater and solid waste analytical methods.

       These Proceedings were  produced  by recording the talks at  the Symposium,
transcribing the recordings  to hardcopy form, providing the hardcopy to  the speakers
for review,  and incorporating the revised  text into  this document.  The questions  and
comments at the end of each talk were also provided to  each respective speaker for
review and editing.

       EPA trusts that  the information contained in these Proceedings will  aid  in the
analysis of  environmental  samples, and we look forward to the Tenth Symposium in
May of 1986, again planned for the OMNI Hotel in Norfolk.
                                             W. A. Telliard

-------
           NINTH ANNUAL ANALYTICAL SYMPOSIUM

       Office of Water Regulations and Standards
             Industrial Technology Division

                   March 19-20, 1986
                   Norfolk, Virginia
                         INDEX

                     March 19, 1986

Presentation/Speaker ,•.

WELCOME AND ' INTRODUCTION	
William A. Telliard, Chief
Energy and Mining Branch
USEPA, Industrial Technology Division
STATUS OF ANALYTICAL PROGRAMS	,
Thomas Fielding
USEPA, Industrial Technology Division
EPA1 S GROUNDWATER MONITORING TASK	,
FORCE FOR RCRA APPENDIX VIII ANALYTES
Paul Friedman, Ph.D.
USEPA, Office of Solid Waste

SYSTEMATIC APPROACH TO METHODS DEVELOPMENT,
FOR RCRA APPENDIX VIII ANALYTES
Sam Lucas
Battelle Memorial Institute
COMPARISON OF SW-846 AND 304(h) METHODS FOR.
ANALYSIS OF APPENDIX VIII ORGANIC COMPOUNDS
Dianna Kocurek
Tischler-Kocurek
25
53
80
METHOD PERFORMANCE CHARACTERISTICS FOR.
SELECTED RCRA APPENDIX VIII ANALYSIS
David N. Speis
Environmental Testing and
Certification Corporation
135

-------
                                                        2a
                           INDEX
                       March 19, 1986
Presentation/Speaker
DEVELOPMENT OF COMPREHENSIVE ANALYTICAL	  169
METHODS FOR THE DETERMINATION OF PESTICIDES
IN GROUNDWATER
Tina Engel
Battelle Memorial Institute

RAPID MULTI-ELEMENT SCREENING USING	  213
SEQUENTIAL ICP
Edward Handel, Ph.D.
Centec Analytical Services

NATIONAL CRITTER	  238

DRILLING FLUID TEST PROCEDURES: PARTICIPATION,	  239
DATA AND COMPARISON IMPLEMENTATION
Thomas W. Duke, Ph.D.
USEPA, Gulf Breeze Environmental Research Laboratory

STATISTICAL EVALUATION AND VALIDATION OF THE	  248
EPA DRILLING FLUID TOXICITY TEST PROCEDURE
Barrett P. Eynon
SRI International

ORGANIC CHEMICAL CHARACTERIZATION OF DIESEL	  310
AND MINERAL OILS USED IN DRILLING FLUIDS
John Brown
Battelle N.E.M.R.L.

CORRELATION OF BIOLOGICAL INDICATORS WITH	  342
CHEMICAL ANALYSIS DATA
Robert C. Barrick
Tetra Tech, Incorporated

TESTING CONSIDERATIONS FOR THE TOXICITY	 385
CHARACTERISTIC LEACHING PROCEDURE (TCLP)
Todd A. Kimmell
USEPA, Office of Solid Waste

-------
                                                         2b
                            INDEX

                       March  20,  1986

Presentation/Speaker                                  Page

DETERMINATION OF PRIORITY POLLUTANT	    422
PESTICIDES BY ISOTOPE DILUTION GC/MS
Robert Be inter, Ph.D.
S-Cubed

PLANS FOR AN IMPROVED ALGORITHM FOR	.	    458
FINDING GC PEAKS IN GC/MS DATA
John McGuire, Ph.D. „••
USEPA, Athens Environmental Research Laboratory

HYDROLYTIC TRANSFORMATION OF  RCRA	    495
APPENDIX VIII COMPOUNDS
N. Lee Wolfe
USEPA, Athens Environmental Research Laboratory

EVALUATION OF METHOD 200.1 DETERMINATION	   527
OF ACID SOLUBLE METALS
Theodore D. Martin
USEPA, Cincinnati Environmental Monitoring
and Support Laboratory

RESULTS OF THE TRW EPRI ANALYTICAL METHODS	   571
QUALIFICATION PROGRAM, PART 1
Raymond F. Maddalone, Ph.D.
TRW Space & Technology Group

CLOSING REMARKS	   599

Roster of Attendees	  600

-------

-------
                PROCEEDINGS
                          MR. TELLIARD:  Good morning.



My name is Bill Telliard, I'm from EPA, and I'm here



to help you.  I'd like to welcome you to the Ninth



Annual Office of Water's Analytical Program.



     You know, we've been doing this for nine years,



and I think a lot of the things that we've done here



have been epitomized by the conversations and the



opportunity to meet people.  I was listening to Bruce



Colby talk to Bob Beimer out in the lobby.  People ask



us what we do here, and I always explain to them it's



a bunch of analysts who get together and lie to each



other.  But, you know, Colby was explaining to Beimer



that he can get down to the picogram/phentogram level



and Bob pointed out to him that's been his problem in



his whole career; he couldn't think big.



     So, over the next day and a half, we'd like to



think big and talk about a number of different



issues.



     Our first speaker this morning is my colleague,



associate, and the guy who does all the work. Dr.



Tom Fielding, and he's going to tell you a little



bit about what the Industrial Technology Division is

-------
doing in 1986, and I'll put the real onus on him,



and why we're doing it in 1986.  Tom.

-------
                   THOMAS FIELDING

    UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
            INDUSTRIAL TECHNOLOGY DIVISION
      STATUS OF ANALYTICAL PROGRAMS WITHIN EPA'S
            INDUSTRIAL TECHNOLOGY DIVISION
                          MR. FIELDING:  Well, one

of the reasons we're doing things this year is we

have the money to do it.

     You had an opportunity, I hope everybody got

these, to pick up a couple of books.  One is the

1986 List of Lists, which is a compilation of every-

body's favorite list of compounds and, of course, as

you may have noticed, some of them are other people's

favorites.  So, the book is fairly extensive.  It's

not quite as extensive as it might look.  The List

of Lists' texts or charts list the pollutants alpha-

betically by name, and if it's got more than one

name, it's listed more than once.  So, it's not quite

as long as it looks.

     The other document that you should be able to

pick up in the back is this nice kelly green document

in honor of St. Patrick's Day, the ITD List of Analytes

for 1986.  You can guess, since we said 1986, that

it may change for 1987.   In fact, as we'll find out,

-------
it may well change in 1986.


     The List of Lists, this nice big book, includes


whole bunches of compounds, pollutants, analytes,


some of which are not amenable to the traditional


analyses we've done before; the GC/MS type of analysis


for organic compounds.  Some of the pollutants on the


List of Lists react with water, in some cases so fast


it's a puff of smoke.  In other cases, they hydrolyze


or react fast enough that we don't believe we can get


a sample back to the lab, extract, and have enough


left of that pollutant to actually analyze.
                          f

     Some of the pollutants are very thermally un-


stable and may not survive being injected into a gas


chromatograph.  Some are too polar to move, some of


them cannot be extracted, some have too high a


boiling point to be amenable to gas chromatographic


analysis.


     In these cases, probably a method such as liquid


chromatography would work, but at its present state


of development, we don't believe that liquid chroma-


tography is sufficiently developed to use as a


screening technique.  You pretty much have to know


what you're looking for to use it.  Others, we don't


know what is going to work, so those compounds did


not survive giong from the List of Lists to the List

-------
of Analytes.



     Some of the pollutants on the List of Lists, or



the original list from which we derived it, are



listed as NOS, Not Otherwise Specified.  For example,



chlorinated benzenes, NOS.  But as analysts, you need



to know what to monitor, what to analyze.  In that



case, we have selected examples of the Not Otherwise



Specified category of compounds which are most likely



to be found in the environment.  Quite often this is



the priority pollutants or similar type compounds.



And in some cases, we've added an additional example




because we did find that pollutant in the environ-



ment.



     Many of the compounds on the List of Lists are



inorganic chemicals; salts.  These will dissociate in



water, and you cannot actually monitor for that com-



pound.  What you can monitor for is the metal, or in



some cases for anions that are quite important, such



as cyanide, you could analyze for the anion.



     Some of the pollutants on the List of Lists, the



method that you have to use to monitor for it shows



what you can actually do.  For example, the pesticide



2,4,5-T, salts and esters.  The method derivatizes



whatever it was into the methyl ester of 2,4,5-T and



that's what you actually monitor.

-------
                                                    8




     We also felt that pollutants, if we're going to



monitor for them, we really ought to have a standard



so that we can be sure we actually saw what we think



we saw.  In some cases, we don't know where to find



standards, so those pollutants are not on the List of



Analytes.



     The List of Lists also includes the most recent



published list, the Volatile Toxic Compounds, published



by the agency late last year.  Those pollutants did



not survive in the List of Analytes because we have



not had time to see and examine them all and see



which ones really can be done.  So, if it was listed



only as a volatile toxic compound, it did not survive



to the List of Analytes.  These selections are generally



similar to those followed in deriving the groundwater



monitoring guidance by the Office of Solid Waste,



which our next speaker will discuss in more detail.



SLIDE 1



     At any rate, what we end up with, with all these



deletions, is shown in our first slide here.  This



is as of February 25th, 1986.  The date is on your



printouts, and it does imply we might in fact have



some changes, very few this year, but certainly next



year probably additional analytes.  Well, there are 377



analytes, using nine analysis types and 26 different

-------
methods or variants of methods.
SLIDE 2
     Now, our second slide outlines our projected
sampling program for the Industrial Technology Divi-
sion for this year.  The column labeled RCRA  indicates
that's the group that we're going to...the number of
samples we expect to analyze for the List of Analytes.
The difference, as you can see, is the Metals Industry
Branch, because several of their projects do not, they
know, involve industries which have very little organic
pollution and, therefore, they don't need to spend the
money to prove that materials aren't there.
     The Consumer Commodities Branch has been involved
with the domestic sewage study that we did last year,
as directed by Congress, and the reauthorization of
the Solid Waste Disposal Act.  This year they're
going to follow up on that study with more data, more
sampling and more industries.
     The Chemicals Industry Branch is primarily
interested in two new projects; one,  hazardous waste
treaters.   These are facilities that treat or dispose
of hazardous wastes and generate a wastewater which
is discharged.  If it's discharged,  it's under the
Clean Water Act and we're going to take a look at it.
The second major project there is  solvent and barrel

-------
                                                    10
reclaimers, which again may generate a wastewater
which must be discharged and we're going to take a
look at that.
     The Energy and Mining Branch, most of those
samples represent the ongoing study of the oil and
gas industry.
SLIDE 3
     The ITD RCRA List of Analytes and the List of
Lists includes the origin or the lists from which we
selected the pollutant.  These are shown on this
third slide.  Most of these are probably familiar to
all of us.  The priority pollutants, of course, are
the 126 priority pollutants required by the Consent
Decree and the Clean Water Act.
     Appendix C to the Consent Decree was a little,
short page of additional analytes that we said, Judge
and NRPC et al, we promise someday we will look at
these pollutants.  Some day has arrived.
     Paragraph 4(c) takes advantage of the fact that
GC/MS analysis collects data on nine track magnetic
tape for all pollutants in the sample, not just those
we thought we were interested in at the time.  The
paragraph of the Consent Decree told agency to examine
those tapes and identify all the pollutants we could,
massage that data to select those that are most

-------
                                                     11
 important  environmentally  because  they're  not  as  bio-
 degradable  and probably  are  pretty toxic.
     The 4(c) program  found  or  identified  over 1,500
 compounds  at  least once  in samples collected from
 1977 through  approximately 1980.   Various  massaging
 and confirmation techniques  and review  and analysis
 to determine  whether the pollutants are biodegradable
                                /
 or not, and whether they would be  toxic or not, has
 reduced it down to 55  pollutants or analytes for  our
 first bite at the Paragraph  4(c) compounds.
     The Hazardous Substance List, HSL  in  the  origin
 column of your books,  is the Hazardous Substance  List
 under the Superfund law.
     RCRA Appendix VIII, of course, is the  promulgated
 RCRA compounds.  The Michigan Petition is  the  petition
 received from the State of Michigan which  said,
 Agency, please add these compounds to your  list of
 hazardous materials under  the RCRA.  We have proposed
 to do that, but no final action has been taken at
 this date.
     In addition to our looking at the lists to see
which ones  we thought we could really analyze,  the
Office of Solid Waste also did the same thing for
 their RCRA Appendix VIII compounds, and in some cases,
 they found  that while the compound itself could not

-------
                                                    12




be analyzed in water, what it decayed to was.a very



good measure, or changed to, was a very good measure.



But the pollutant...analyte they are going to look



for, doesn't happen to appear on anybody else's list.



The four pollutants that we found this way were



osmium, tin, fluoride and sulfide, which don't appear



on anybody else's list. ^So to keep track of those,



we chose to use the name of the Chairman of the



Technical Committee which reviewed the Appendix VIII



compounds, Dr. Bob April.



SLIDE 4



     Now, the next slide will show the analysis types



that we are concerned about.  CS2 liberation is Method



630, promulgated last October as part of the pesticides



regulation under the Clean Water Act.  Basically, it



makes use of the fact that dithiocarbamate pesticides,



under appropriate conditions, hydrolyze to lead to



reduce or to liberate carbon disulfide, and you can



measure the amount of carbon disulfide released.



     It has two very good advantages for us; it's



fast, it's cheap.  The bad thing about it is you



don't know what pesticide actually produced the CS2.



The ITD List of Analytes has six analytes that will



be done by Method 630, but Method 630 lists 17 analytes,



and in fact, anything else that will yield carbon

-------
                                                    13
disulfide will give you the same result.
     Cold vapor AA, of course, is for mercury.  The
dioxin method is the promulgated Method 613.  This is
for the 2,3,1,8-tetrachlorodibenzodioxin.  For
obvious reasons we call it dioxin.
     One thing that we can do also is to do a more
general test for chlorinated dioxins and chlorinated
furans, and we will attempt these tests if requested
by project officers.
     Furnace AA, of course, is for metals.  Gas
chromatography with selective detectors is primarily
for pesticides and PCBs.  I'll get into that a little
more later on because there are a number of specific
tests that we would like to consolidate, if we can.
     Some of the pollutants in the various lists are
soluble, volatile organics that don't purge too well
under the standard room temperature purge, but we
think will work pretty well if you heat up the purging
device.  So those are identified by the term, hot
purge.  What we need to do shortly, and will be doing
shortly, is to test this hypothesis and find out how
well it works and what kind of results we actually
get for quality control and quality assurance purposes.
     Search is a reverse search to match retention
times and mass spectral data in your library against

-------
                                                    14
the pollutants found or identified...! won't even say
identified...in the tapes.  This would be done by the
laboratories themselves by adding the 120 or so
pollutants to their library.
     ICP is Inductively Coupled Plasma Emission
Spectroscopy.  Again, for obvious reasons, it's ICP.
Twenty elements are listed in the ITD List of Analytes,
but the method can be used for a rapid screening for
up to 70 or so analytes if necessary, for weird metals,
erbium, and like that.
     Finally, we have some pollutants that are better
handled by a classical wet chemical method.  These
five that are listed in the List of Analytes are
cyanide, chloride, fluoride, sulfide and nitrite.  We
would, of course, also expect to do analyses for some
of the other types of wet chemical pollutants, such as
biochemical oxygen demand, chemical oxygen demand,
total organic carbon, ammonia, or whatever else is
requested.
SLIDE 5
     This next slide gets back to the GC with selected
detectors problem, primarily for pesticides.  We
think, we hope, we can consolidate this list of
different methods to three, or perhaps four methods.
Method 608, 608.1 and 617 are for organohalide

-------
                                                    15



pesticides and PCBs using an electron capture detector.



We think, we hope, they can be consolidated to one



method, since it will mean just one preparation and



one chew.  This will be tested, and any pesticides



that we find can't be done by one method, will have to



have their own special method, will be dropped from



the routine list, although they will be available if



we think there's a real possibility of finding the



compound.



     The herbicides Method 615 is for the herbicides



2,4-D, 2,4,5-T, 2,4,5-TP, captan and dinex.  This



method derivatizes the pollutants to the methyl ester



and, therefore, cannot be consolidated with some other



method.



     We have four methods for organophosphorus



compounds, actually it's three methods, with two



different types of detectors, one using the flame



photometric detector, one using a nitrogen phosphorus



detector.  We hope these can be consolidated in some



fashion to reduce the number of separate preparations



and analyses, because each preparation costs more



money.



SLIDE 6



     Finally, this last slide will show or summarize



the GC/MS analytes, which basically is whatever is

-------
                                                    16




left on the ITD List of Analytes.  They are, of



course, the organic priority pollutants that we have



been analyzing by GC/MS for the past 10 years;



volatile acids, base neutrals; Appendix C organic




pollutants, which we know can be done by GC/MS;



Paragraph 4(c) pollutants, which we know can be done




by GC/MS because that's how we found them; and the



hazardous substance list, Appendix VIII analytes,




Michigan List analytes that can be done by GC/MS.



     Now, all of the data on GC/MS is still being



collected on nine track magnetic tape, and we do plan




to go through a tape screen similar to what we did



before for the Paragraph 4(c) program to see what



else might be there.  That will be done probably next



year because we've got a pretty heavy schedule for



this year.  To summarize that schedule again, 750



samples are planned, almost 600 of which we




expect to do a detailed study for the ITD List of



Analytes which contains some 377 pollutants.



     Are there any questions about this program for



this year?



                          MR. TELLIARD:  Questions?



                          MR. FIELDING:  No questions,




that's good.  We might be on schedule for a change.



                          MR. TELLIARD:  As Tom

-------
                                                    17
mentioned, we have a number of new industries we're
going to look at this year.  One of them is onshore
oil and gas, which we've just started.  Another one
is oil re-refiners, whatever they are.  In view of
the number of samples we have this year, we have to
do something that we haven't done for the last couple
of months; we're going to put some contracts on the
street.
     Most of you aren't interested in that, but I
saw this fellow in the back who had a box of pencils
saying Contract Lab Program.  I just thought I'd
point that out to you that yes, there is work.  Of
course, we're not on the scale of Stanley, but a
few dollars here and there helps.  So, yes, we plan
on going out with some contracts in the next couple
of months for GC/MS work, probably for metals, and
also for the pesticides, which will be GC.
     But it's kind of an interesting program in that
we've kind of fallen into a lull of the priority
pollutants and the joy of the priority pollutants,
which you've all grown to love and understand.  But
now, with the List of Lists, it's really neat because
we can change things at any moment and never tell
you, which makes me feel warm and fuzzy because I can
stay two steps ahead of Stanko.

-------
                                                    18
                          MR. TELLIARD:  Our next



speaker is a fellow who's trying to stay two steps



ahead of everybody.  He's with the Office of Solid



Waste.



     Dr. Paul is going to talk to us this morning



about Appendix VIII.

-------
                                   19
ITD ANALYTICAL PROGRAMS — ITD/RCRA ANALYTES

   OVERVIEW
       377 ANALYTES
       9 ANALYSIS TYPES
       26 METHODS OR VARIANTS OF METHODS

-------
                                   20
ITD ANALYTICAL PROGRAMS — INDUSTRIES SAMPLED
   INDUSTRY

   METALS
   CONSUMER COMMODITIES
   CHEMICALS INDUSTRIES
   ENERGY AND MINING
                               SAMPLES
TOTAL

250
150
200
150
RCRA

   75
   50
  200
  150

-------
                         21
ANALYTE LISTS
   PRIORITY POLLUTANT
   CONSENT DECREE
      APPENDIX C
      PARAGRAPHIC)
   HAZARDOUS SUBSTANCE
   RCRA APPENDIX VIII
   BOB APRIL
   MICHIGAN PETITION

-------
                                   22
ITD/RCRA ANALYTES — ANALYSIS TYPES
   CS2 LIBERATION (AS NABAM)
                    (6)
                   (1)
   COLD VAPOR AA (MERCURY)
   DIOXIN (METHOD 613)      (1)
   FURNACEAA     (5)
   GC WITH SELECTIVE DETECTORS     (80)
       ELECTRON CAPTURE     (44)
       FLAME PHOTOMETRIC     (21)
       NITROGEN PHOSPHORUS      (15)
   GCMS      (258)
       1624     (31)
       1625     (81)
       HOT PURGE      (26)
       SEARCH     (120)
   ICP     (20)
   TEM (ASBESTOS)     (1)
   WET
(5)

-------
                                 23
METHODS FOR PESTICIDES/HERBICIDES
   ANALYTES FOR CONSOLIDATION
   ORGANOHALIDE PESTICIDES/PCB's BY ECD
       608
       608.1
       617
   HALOGENATED HERBICIDES
       615
   ORGANOPHOSPHORUS PESTICIDES BY BEAD/FPD
       622
       701
   ORGANOPHOSPHORUS PESTICIDES BY NPD
       614
       622

-------
                                  24
GCMS ANALYTES
   ORGANIC PRIORITY POLLUTANTS
      VOLATILE
      ACID
      BASE/NEUTRAL
   APPENDIX C ORGANIC POLLUTANTS
   PARAGRAPH 4(c) POLLUTANTS
   HAZARDOUS SUBSTANCE LIST POLLUTANTS
   GCMS-ABLE APPENDIX VIII ANALYTES
   GCMS-ABLE MICHIGAN LIST ANALYTES

-------
                                                    25
                    PAUL FRIEDMAN

    UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                OFFICE OF SOLID WASTE
       EPA'S GROUNDWATER MONITORING TASK FORCE
           FOR RCRA APPENDIX VIII ANALYTES
                          DR. FRIEDMAN:  Thanks, Bill.

I guess this happens when you don't submit an abstract.

I'm only peripherally going to discuss Appendix VIII

in light.of the activities of the Groundwater Monitor-

ing Task Force which is involved in laws, as you'll

see as we go through this.

     The possibility of commercial hazardous waste

management facilities becoming the Superfund sites of

the future is one of the immediate concerns growing

out of the national awareness of the groundwater

resources.  This general concern is being focused on

several questions.  One; are these commercial .hazardous

wastes and site management facilities leaking or do

they have the potential to leak?  Are the present

wells and monitoring systems at these sites adequate

to detect leaking?  Are the analyses currently being

carried out to support permits and compliance moni-

toring acceptable and adequate?  As a corollary to

that, are the RCRA OSW procedures, SW-846 essentially,

-------
                                                     26
 adequate  to  analyze  samples  for  Appendix  VIII  analytes
 in groundwater?
 SLIDE 1
     If I could get  the  slide machine  on.   How about
 that?  Now,  if I can figure  out  how  to brighten it
 I'll be all  set.
     Well, the agency response to  this question, or  to
 this issue,  has been the establishment of a groundwater
 strategy group which provides overall  guidance on the
 generation of national groundwater policies.   To ad-
 dress the specific questions, the Groundwater  Monitor-
 ing Task Force was established;  those  questions con-
 cerning commercial waste management  facilities.
     The basic approach, on  a site by  site basis...these
 are the questions concerning leaking sites, wells and
monitoring systems,  the chemical analysis and  the
 analytical methods.   The basic approach on a site by
 site basis has been  the...well,  to take an analysis
of the site  history  in terms of  wastes that have been
accepted there and the past management practices, to
analyze the monitoring systems and hydrogeology in
terms of well placement, well construction and general
well conditions,  and  to survey the analytical chemistry,
both from the standpoint of the historical background
of the analytical chemistry in terms of the analyses

-------
                                                    27
done under RCRA permits, as well as the current
analytical effort being undertaken by the Groundwater
Monitoring Task Force.
     I'd like to concentrate my comments on the
analytical aspects and the attendant quality control
related to these aspects.
     Early on in the startup of the Groundwater
Monitoring Task Force a decision was made to employ
the resources of the Contract Laboratory Program for
its analysis capabilities.  The advantages of this
situation is that it provided a relatively fast
startup, basically in a short period of time, and the
possibility of lower costs to the agency.
     The disadvantages are that, and this is not
necessarily due to the Contract Laboratory Program
but to the nature of the way the Task Force was
structured, these analytical contracts had to be
of relatively short duration, between six and
nine months, and that because of that you are always,
with the possibility of getting new labs all the
time, always on the low portion of the learning
curve.
     The program itself is basically broken up into
two portions.  One was termed Phase I because it was
the first phase,  very very imaginative,  designed

-------
                                                    28
to look at the first seven sites out of about 65 total
sites to be investigated across the country with the
idea that this would be on altering curve at which
point there would be a reassessment and some mid
course corrections made for the rest of the program.
     The first seven facilities were examined and
analyzed by the Task Force using a combination of the
routinely used CLP analytical methods.  These include
the ICP methods for metals, the GC/MS volatiles and
semivolatile methods out of the CLP for the organic
analytes, and additional procedures coming out of
OSW, which included the pesticide procedures and
herbicide procedures that are referenced as 8080 and
8150.
     There were also a number of indicator analytes
which basically arise out of the monitoring require-
ments of RCRA, which require monitoring for analytes
such as TOC and TOX as, again, indicator analytes or
parameters which puts one into compliance monitoring
with respect to the regulations.  Those regulations
then force the owner/operator to analyze for Appendix
VIII compounds.
     So, the whole suite of analytes, with respect to
the indicator parameters were required.
     The analyses were carried out.  One lab was

-------
                                                    29
designated for inorganic indicator analytes and one
lab for organic analytes.  The organic compounds for
this effort were, at least for Phase I, were the HSL
compounds.  In fact, all the compounds were the HSL
compounds, the metals and the organic compounds.  It
was not the full suite of Appendix VIII compounds
which were at that time currently required for
groundwater monitoring.
     At the conclusion of Phase I, which were the
first seven sites, and coincidental with the first
analytical contract, additional compounds would be
added from the Appendix VIII list with more being all
those compounds which were analyzable by GC/MS and
for which the agency could secure standards from a
commercial source.
     As well as adding new compounds, the analytical
processes were also evaluated in terms of methods, in
terms of QC procedures and in terms of analytes.  In
accordance with this, the Quality Control Project
Plan, which was developed for the evaluation of the
Task Force data, consisted with the required standard
operating procedures and a set of DQO's.
     Just as a digression, the setting of DQO's
required a process which was not terribly compatible
with the Task Force operations.  First of all,  the

-------
                                                    30
numbers of samples per site were either unknown or, at
best, variable.  The relationships of wells up and
down gradient sometimes were not well known or well
understood, and the methods to be used to assess the
facilities in terms of leaking were combinations of
qualitative and quantitative .processes involving
hydrogeologic opinions and analytical data, as well
as searching of records.  The analytical methods
employed, and as an adjunct to that, the analytical
methods employed were already established as the
methods we had discussed.
     In the best of all possible worlds, the DQO,
or in a situation, usually the DQO dictates the
appropriate methods which allow those objectives to
be achieved.  In this situation, procedures used
placed constraints on the data quality objectives for
precision, for accuracy and detection limits.  So the
DQO's adopted were, therefore, dictated by these
methods, as opposed to the other way around.
     The QC program can then be described as intensive
for laboratory activities, combining a synthesis of
Contract Laboratory Program QC requirements with
additional Task Force requirements.  Because of the
the strategy of taking a sample from each well or
from as many wells as possible on a site, but only

-------
                                                     31
 one  sample,  the  use  of  all  data  generated  had  to be
 maximized.   This means  that the  limitations  of the
 data,  as  a result of the methods,  as  a  result  of the
 sampling, the matrix and other incidental  conditions,
 must be fully understood.
     The  Task Force  philosophy on  the data evaluation
 was  that  the evaluation process, the  actual  qualifying
 of the data, if  you  will, had to be independent  of the
 data interpretation,  which  meant understanding the
 data in light of the  site conditions.   These two
 processes had to be  independent  in terms of  the
 analysis of the waste management facility.
     This independent evaluation of data is  intended
 to describe and, as  I said, to qualify  the data, and
 hopefully will allow  the user to put  the data  in the
 proper context with  respect to the other information,
 the non-quantitative, the descriptive information
 about the site.   Therefore, the data  evaluation  then
 becomes not a guarantor of  the quality, as in some
 cases it's purported to be, but more  an estimator  of
 the quality.
     So, employing what might be called traditional
quality control  procedures, such as the usual labora-
tory-generated information, along with an intensive
program of performance evaluation sample program,  the

-------
                                                    32
quality control information and analytical results
are evaluated with respect to their usability.
     Thus, depending on sample conditions and analyti-
cal procedures, this evaluation occurs either on an
analyte by analyte basis, on a sample by sample basis
or a batch basis.  It's expected that this type of
evaluation will help to ^minimize the over- or under-
reliance on the analytical data.  And I say that in
terms of looking at the entire site information
developed.
     In addition to the quality control data developed,
such as control charts for accuracy, precision,
general chromatographic quality, information verifi-
cation of spectra and chromatograms, the data are
also checked for internal consistency.  That means
either the consistency or the correlation of various
sorts of analytes with other analytes.  For instance,
the TOX total organic halogens and  purgeable
organic halogens should not necessarily be colinear
with, but should, or hopefully should, correlate with
results gotten from GC/MS analyses  for chlorinated
organic compounds.  In a parallel fashion, phenolics
with acid fraction and semivolatiles, and that kind
of an internal check on the data.   There is, at
best, a tenuous  relation there.

-------
                                                    33
     Upon completion of the evaluation, the data are
released to the user who can then integrate the data
with respect to the analytical data and the hydro-
geologic data and the background information on the
site.  This basically gives a flow of the activities
which lead up to that released data to either NEIC
or the region; the idea of the site planning meeting
each year in which the background information is
obtained, the site visit that is basically the
sampling episode which sends the samples to the
laboratories.
     At that point a PE sample is introduced into the
loop, hopefully double-blind, but usually just blind.
This then goes to the data evaluation for the inde-
pendent evaluation of the quality control indicators
within the package itself.  A report on that, plus the
data, plus other contractor reports go to the Task
Force.
     The NEIC regional involvement is basically
they are the lead with the first seven sites, the
lead being taken by NEIC,  and thereafter the region.
So, these first sites are basically not only an evolu-
tionary stage for the Task Force, they're also a
training stage for the regions as well.
     The data user then generates the site report.

-------
                                                    34
Basically, the site report discusses the kinds of
things that obviously go on in the process of gener-
ating the site information; the site history again,
as indicated previously, analysis of the waste
management practices, indication of wastes deposited
at the site which could pose future problems to
groundwater, information^.on geological strata, the
impact of waste management practices on the geology
of the site, as well as the depth, direction of flow
and extent of the aquifer.
     Here the monitoring systems also get an evalua-
tion in terms of their adequacy with respect to
design and with respect to the actual facility's  own
sampling procedures.  The analytical chemistry data
are integrated here  in the sense that both  the
historical data with respect to whatever monitoring
has gone on, and the current data are compared and
contrasted with respect to the kinds of analytes  and
the levels found.
     One has to be careful here because basically
the Task Force  is  looking  at a snapshot with respect
to the analytical  data and trying to look at that and
compare that to say, averages for previous  results
can be deceiving because you could  be looking  at
temporal changes which distort the  picture  with

-------
                                                     35



respect to one particular time.  Basically, the



interpretation is in terms of the  implications with



respect to groundwater quality and site  integrity.



     The assessment of Phase I activities with respect



to this task indicate, for the methods,  some interesting



points in that the application of  routine contract



lab procedures to the groundwater  regime here requires



more guidance than is currently available in either



the Contract Lab Program or OSW methods.  Contractually



stipulated procedures by the so-called routine



analytical services did not allow  for the adjustments



required for a variety of samples or a variety of



conditions that were encountered.



     Phase II, which is basically the balance of the



program, will include a more elaborate and a more



detailed protocol within the contractual framework to



incorporate readily available techniques to overcome



matrix constraints.



     In terms of the management aspect of this, which



I think is a more interesting activity in terms of



the various aspects  that have to be pulled together,



the fast pace of activities in the Groundwater



Monitoring Task Force of analyzing anywhere from a



site every week to three weeks.



     The Laboratory  and Field Group and Data Evalua-

-------
                                                    36




tion Committee have to have a much more, greater



degree of communication.  Basically, this will stop



the institutionalization of procedural problems which



manifest themselves in the analytical result.  Basi-



cally, you have analytical and sampling artifacts



which become standard operating procedures, essentially,



unless they are dealt with in a real time basis.



     Another aspect is the pre-site planning which,



in terms of planning for an analytical sampling



episode, can be extended to the actual analysis



itself because it gives one a chance by understanding



the situation at the site to maximize the quality



control.  For instance, there may be particular



analytes that develop as a result of the site history



or particular wells which turn out to be critical,



and the procedures essentially at this point call for



a random selection of which samples should become a



spiked sample, which should become a duplicate, and



this allows the Task Force to select the particular



well or the critical sample or the critical analyte



with regard to that.



     Additional measures under consideration include



tailoring performance evaluation samples to the site



matrix or to particular analytes.  This is especially



relevant for the indicator parameters such as TOC,

-------
                                                    37
POC, TOX and various anion indicator parameters.  As
I mentioned before, Phase II will monitor for an
additional set of Appendix VIII analytes.
     This set of analytes most closely resembles the
guidance which is being put out by Bob April's office.
There is considerable overlap between the two lists.
The list was basically derived on a system of taking
what standards were available, determining which
compounds could be chromatographed and which compounds
would survive a very small partition study to see
that, in fact, they are extractable.  The same methods
as Phase I will be used, with the addition of a
method for...a screening method, essentially, for
dioxins and dibenzofurants, which in OSW parlance is
number 8280.
     While only a limited amount of data for Phase I
have been thoroughly scrutinized, a tentative con-
clusion, and it has to remain tentative, is that the
facility monitoring systems, in general, have to be
upgraded.  The data appear to be adequate though,
that are generated by the Task Force, to answer the
questions or objectives as outlined in terms of site
leakage and data adequacy.
     The procedures instituted in Phase II will enable
the delegation of the Task Force leadership from NEIC

-------
                                                    38
to the regions with what we expect to be a minimum



loss of consistency in the evaluation and interpreta-



tion process for the remainder of the 65 sites to be



surveyed.



     In conclusion, we expect to deliver a report to



Congress sometime in May concerning the first 10



sites, and we expect the Phase II corrections to be



implemented sometime within this by the end of this



month.  Thank you.



     Any questions?

-------
                                                    39
             QUESTION AND ANSWER SESSION
                          AUDIENCE PARTICIPANT: Paul,
in your studies, did you find situations where up-
gradient and downgradient POC changed significantly
but were not seen as TOC?
                          DR. FRIEDMAN:  Yes.  As
a matter of fact, there were inconsistencies in that
area.  Is that you, Bob?  Never mind.  There were
instances where TOX would be high, and you'd
expect the TOC to be high as well because these
are supposedly chlorinated organic compounds.  That
did not turn out to be the case, which could mean, if
one looks at the chloride, one might be seeing
interference there as well.
     As I said, some people tend to rely on these
quite heavily, but they have to be integrated into a
picture to negate artifacts that occur.
                          AUDIENCE PARTICIPANT:  How
about the specific circumstance that I mentioned?
                          DR. FRIEDMAN:  Do you want
to give me that again?
                          AUDIENCE PARTICIPANT:  Where
you saw significant changes in POC, but because of a
small fraction of TOC it might be, you wouldn't see it

-------
                                                    40




when comparing POC upgradient and downgradient.



                          DR. FRIEDMAN:  I don't have



a specific indication in mind.  I just know that there



is a high probability that could happen because of



the way those particular parameters...! don't like to



call them analytes...vary.



                          DR. GAIND:  My name is



Arun Gaind from Nanco.  You mentioned that there were



problems with the CLP protocol.  Can you elaborate



what areas you specifically found problems in and



what corrections you are going to put in place for



Phase II?



                          DR. FRIEDMAN:  Yes.  There



were two kinds of problems.  One was with what I call



an administrative side, which was basically calculation



of data.  It left out certain key aspects in terms of



the use of some of the quality control data which are



there to compensate for certain kinds of interferences



using the pre- and post spike, pre-digestion and



post-spiked digestion results, so that while those



were not calculated or compensated for in the result,



the data were captured.  Basically, what we have to



do is go back and correct some numbers.



     The other aspect is that there are obviously



tricks of the trade which were not built into the

-------
                                                     41



 analytical  protocol  for  instances  where  you  have



 certain  kinds  of  interferences  or  high backgrounds  of



 certain  kinds, which can be corrected for with  respect



 to certain  additions to the metals analysis  in



 particular.  I'm  referring mostly  to metals  now.



     Those  are the changes in the  analytical protocol



 that we're  going  to make, and that's basically  to



 reach groundwater/drinking water detection limits,



 or to attempt  to  reach drinking water detection limits,



 in a groundwater  matrix, in a groundwater matrix that



 has a high  concentration of cations and  anions.



     Another problem with respect  to that is the



 actual calculation of a useful detection limit which



 takes into  account those kinds of backgrounds and



 those kinds of interferences to give you something



 realistic with respect to your result.  And  that's



 still up in the air as far as I'm concerned.



                          MR.  SLOAN:  Tim Sloan,



 Rogers and Callcott Laboratory.   Did you see in any



 of your studies,   I know you spoke about a correlation



between POX and the,  did you see any that didn't



 show a definite correlation between the two?



                          DR.  FRIEDMAN:   Yes.  As I



said, I'd like to back away from the word correlation.



That's a bit strong.   I think  you see some relationship,

-------
                                                    42




or should see a relationship.  If you don't, people



tend to interpret that in a different way.  For



instance, you might say well, okay, I've got a high



TOC or TOX, let's use that, but I don't see any



volatile organics, I don't see any semivolatile



halogenated organic compounds, so one might infer



from that, and it's just an inference at best, that
                        *r •


you have acidic material that's chlorinated, which




might be the result of some oxidative process going



on, or the result of adding pollutants that are



in a pesticide process that are still in an acidic



form or a non-volatile form, or a form with low



volatility, again, that's an inference, and



it's not something you'd leap to and stand up and say




it's definitely that because of that.



     The real problem with these analytes, if you



will, is that they are used  in the reproprobate to




kick into, or variations are used to kick  into com-



pliance monitoring which used to require analysis or



determination of Appendix VIII, which is not going to



be the case any longer because the agency  has come




out with guidance that takes a subset of Appendix



VIII to monitor for, and that subset is basically



based on what is determinable by the conventional



methods for metals and the GC/MS and GC methods for

-------
                                                    43
organic analytes, with possibly a few exceptions.
                          MR. GEYER:  My name  is Tom
Geyer from Ford Motor Company.  Maybe just to  shed
some kind of light on the discrepancies that might be
seen between POX and POC, the inherent problem of
taking the samples, the collection method for  POC,
is not as fully encompassing for POC's, so if  your
increase in POX is due to a volatile organic,  say,
and you've collected your TOC's in bottles with head
space and acidified, you've lost your volatile POX.
                          DR. FRIEDMAN:  The volatile
TOX's, which are called PCX's for this particular
effort, were taken as volatile compounds.  But you're
right, there are a tremendous amount of uncontrolled
variables in sampling which contribute to all  these
inconsistencies.
                          MR. GEYER:  There's so much
chance for variance in the collection method for TOC
that I don't see how they can ever really compare.
                          DR. FRIEDMAN:  And that's
not what we call something that's going to be relied
on heavily.  They're indicator and really that's what
it means.
                          MR. GEYER:  Right.  I think
there could be more of a correlation if the collection

-------
                                                    44
method for TOG were to be in vials with no head
space.
                          DR. FRIEDMAN:  Could be.
                          MR. HANDEL:  Ted Handel
from Centec.  I'd just like to point out that we are
receiving the samples in the Phase II part of this
for the inorganics, and we're receiving the POX and
the POC samples with no head space, so that's not a
problem.
                          DR. FRIEDMAN:  Between the
sampling and the method itself there's probably enough
variability to limit its usefulness in that regard.
                          MR. TELLIARD:  I think that
was a comment,
comment.
                          DR. FRIEDMAN:  That was a
                          MR. TELLIARD:  Anyone else?
                          MR. TROIANO:  Jeff Troiano,
Ford Motor Company.  You referred to the abbreviated
Appendix VIII guidance for groundwater monitoring
purposes that's just come out.  Is it EPA's intention
with regard to Part B permits to make every permit
holder look at all those analytes at least once in
every well or will some consideration be given to the
likelihood of presence?

-------
                                                    45



                          DR. FRIEDMAN:  Perhaps



somebody else could address that in the audience.



I'll just say it's my understanding that that list, or



let's call it a subset of Appendix VIII, will replace



the requirement for compliance monitoring for all of



Appendix VIII in groundwater.  Bob?  Bob April.



                          MR. APRIL:  The regulation



will state that, with regard to Part B, that if your



site is suspected to be leaking, then you have to analyze



the clues for everything that's in Appendix VIII.  Now,



there is certain discretion in terms of what you have



to analyze for, but that discretion is in no way linked



to the probability of finding something at a site.



That may change down the road, but right now you can't



look at what was the waste at a site in terms of



the analytes.



                          MR. TROIANO:  Thank you.



                          MR. TELLIARD:  Thanks, Paul.



                          DR. FRIEDMAN:  Thank you.

-------
                                  46
  GROUND
  WATER
MONITORING
   TASK
  FORCE

-------
                                                     47
                   QUESTIONS
0    LEAKING SITES
0    WELLS & MONITORING SYSTEMS
0    CHEMICAL ANALYSIS
0    ANALYTICAL METHODS

-------
                                          48
           SITE
      INVESTIGATION
  SITE PLANNING MEETING
       SITE VISIT
      LABORATORIES
DATA EVALUATION COMMITTEE
       TASK FORCE
       NEIC/REGION
       SITE REPORT

-------
                 SITE STRATEGY
0    SITE PLANNING MEETING





0    SITE VISIT





0    SITE ANALYSIS

-------
                                                     50
           ELEMENTS OF A SITE REPORT
0    SITE HISTORY
0    HYDROGEOLOGICAL ANALYSIS
0    MONITORING SYSTEM ANALYSIS
0    ANALYTICAL CHEMISTRY
0    INTERPRETATION OF CURRENT ANALYTICAL DATA

-------
                                                     51
              ANALYTICAL STRATEGY
0    PHASE I





0    REASSESSMENT





0    PHASE II

-------
                                                    52
                          MR. TELLIARD:  Our next



speaker is Sam Lucas, and Sam is going to talk



on one of our favorite subjects today, which is



Appendix VIII.

-------
                                                                           53
                                   SAM LUCAS
                                   BATTELLE
                   SYSTEMATIC APPROACH TO METHODS DEVELOPMENT
                        FOR RCRA APPENDIX VIII ANALYTES
          Mr. Lucas:  In the preliminary program announcement Jim Longbottom
was scheduled to give this talk.  The title of the talk (Slide 1) hasn't
changed; certainly the content has.  Undoubtedly, Jim intended to address some
aspects of EPA's future efforts and/or philosophy for developing methods for
the RCRA Appendix VIII analytes.  I'm not equipped to talk about EPA policy.
So, I'm going to spend most of my time talking about some results of research
efforts that we've been working on with Jim recently.
          The research results presented involve two contracts (Slide 2).
Both of these are currently still in progress so that all of the results that
you're going to see here are not finalized.
          In Slide 3 are acknowledged some of the people that were working and
are now working on this research at Battelle.  Probably some of you recognize:
Marcus Cook and Afaf Wenski are managing the programs that I'm talking about.
Judy Gebhart is the principal investigator for the Method 8010, -15 and -20
validation work.  The others are the key people who actually generated the
results.
          Schematically, the experimental strategy for what we're doing is
shown in Slide 4.  Basically, we're trying to filter the organic Appendix VIII
analytes through a scheme like this.  First, we're looking at GC suitability,
and, to do this, we're simply testing to see whether the organic analytes in
the two lists can be chromatographed on the column for either Method 8240 for
volatile compounds or the column for Method 8270, semivolatile compounds.
          Compounds which can't be eluted and detected by mass spectrometry
will be targets for future work.  We expect some of the analytes will require
HPLC-MS.  Some of them undoubtedly will fall into method development involving
derivatization.  It is possible that some non-chromatographic methods will be
required.

-------
                                                                        54
          The step in this scheme that's near completion, and, therefore,  most
of what you're going to be seeing today, is the GC suitability testing to  see
which of the analytes can be chromatographed on these two systems.
          As this method development effort continues, there will  be more
detailed testing of Method 8240 and Method 8270 to demonstrate purge-trap-
desorb, PTD, recoveries and extraction recoveries, respectively.   Those method
validation studies undoubtedly will involve some field matrix testing.
Certainly a key issue on the two sides of this chart are the PTD  recoveries
for Method 8240 analytes and the recoveries through extraction and/or cleanup
for the semivolatile analytes of Method 8270.
          Straightforward extraction techniques will be tried for the 8270
analytes, and, if those don't work, further development in modifying the
extraction conditions using novel techniques as well as development of
improved cleanup techniques for field matrices will be the main focus for
future development on the semivolatile side of this chart.
          Following and/or concurrently with this effort will be  testing in
non-MS methods for the semivolatile and the volatile analytes.  Some of the
results presented below are from Methods 8010, 8015, and 8020 testing for
analytes which have already passed the chromatography test so that we're now
looking at non-MS methods with expanded analyte lists in method validation
testing.
          Ultimately, we expect that the tremendous quantity of methodology
that's going to be required for Appendix VIII analysis (and its future evolu-
tions) is going to require a more flexible approach than what we  seem to have
in place now.  Instead of a list of methods to which analytes are assigned, a
methodology option matrix of the type indicated by Slide 5 will provide analy-
sis methodology options based on the analytes of interest, the type of matrix,
and the specific data requirements.  This decision matrix might provide a  set
of analysis conditions that don't necessarily correspond to a specific method
as currently designated.  This approach will facilitate periodic  updates as
new information comes in about specific analytes in specific matrices,
especially regarding the aspects of extraction and cleanup where  one most
needs the flexibility that this type of matrix approach provides.

-------
                                                                           55
          For the GC-MS suitability testing conditions (Slide 6), we adopted
those of the CLP for both Methods 8240 and 8270.  Thus, I haven't listed any
of those conditions in detail.  I might mention that the column we're using
for the volatile analysis is the usual 1 percent SPlOOO/Carbopack B.  For
semivolatiles, we used a J&W DB5,.30 meter, 0.25 millimeter ID, 0.25 micron
film; a very commonly used column.
          Since this, work involves injection of standards prepared from rea-
gent materials, sample workup aspects such as PTD procedures are specifically
omitted.  For the Method 8270 work we used a Carlo Erba GC which has an
extremely competent splitless mode injector.
          The set of analytes that we were asked to address consisted of the
Appendix VIII set plus the Michigan List, which I'm sure most of you are
familiar with.  The priority pollutants were omitted since their GC charac-
teristics are well known.  Therefore, we combined the original  Appendix VIII
list, the Michigan List, subtracted out all of the priority pollutants and
came up with 328 compounds to address in the study (Slide 7).  We eliminated a_
priori 59 compounds either because there was no possibility of them eluting
from any GC column or because a derivitization-based method was indicated, or
because other circumstances indicated omission.  Some of the analytes aren't
commercially or otherwise available.  We were required to use compounds
supplied by EPA whenever possible since they were of certified purity.
Alternatively, we tried to obtain analytes from commercial sources and found
that, after excluding the 59, 22 compounds were not available from either
source.  Thus, 81 of these 328 analytes weren't addressed.  So, 247 is the
grand total of analytes that we have tried to put through a GC-MS system.
Fifty-four of those were tested as volatile compounds, i.e., Method 8240
candidates, and 197 of them were semivolatile compounds, i.e., Method 8270
candidates.  Four of the 197 semivolatile compounds were originally tested as
volatile compounds but didn't elute from that column and were, therefore,
tested again in the semivolatile.  Thus, 54 volatile compounds plus 197
semivolatile compounds minus 4 compounds tested in both sets equals 247 total
compounds tested.
          For the 54 volatile analytes, we had reasonable GC-MS results for
all but 15 of them (Slide 8).  In six cases we found low enough GC-MS response

-------
                                                                        56
factors that it was likely the analyte would fail  method validation testing.
Indeed, when we included those marginal analytes in the Method 8010, -15, -20
testing protocol, we found 19 compounds weren't recovered by PTD.   All  six of
these low response factor analytes failed the PTD testing.
          For the 197 semivolatile analytes (Slide 9), we were able to  detect
127 with adequate response factors.  An additional 42 analytes remain in sus-
pense awaiting further GC-MS data processing.  Possibly, up to half of them
may yet be qualified for the next testing stage.  Typically, the analytes in
suspense are ones for which reference mass spectra are not included in  the
EPA/NIH Mass Spectral Library.  The scope of this work did not allow genera-
tion of these mass spectra by probe introduction, so we hope to get that
information from the literature or by manual prediction.  There are 28 anal-
ytes for which reference mass spectra are available but which we have not been
able to find in the GC-MS data.  In every case, analytes that we couldn't
detect but that we thought we ought to be able to detect were analyzed in a
subsequent experiment at a higher level, so this number (28) is not expected
to be substantially decreased by further work.  Some of the analytes on this
list of failuresxare quite surprising.  Unfortunately, the scope of this
project doesn't allow us to focus on those failures to determine whether we
have a bad analytical reagent or whether we're mixing it with other analytes
that react with it, and so on, but certainly these surprising failures  will be
the subject of additional work in the  future.
          Slide 10 shows the 59 compounds that we removed, a priori.  Among
these compounds were 14 free and/or strong acids since a derivatization step
would probably be required for them.   Another 12 compounds, typically, anti-
biotics and other very large and/or extremely polar and/or extremely fragile
molecules, were simply not volatile enough to go through a GC.  We felt that
aldehydes would ultimately involve a derivatization method, probably with
HPLC, so they were eliminated also.  Another five compounds were simply redun-
dant with priority pollutants or other entries in the same list.
          Slide 11 shows the list of analytes for which no source could be
identified.
          Slide 12 shows the kinds of  information that will be  included in our
report.  This particular table summarizes the results for the volatile com-
pound set.  Of course, a similar one for the semivolatile compounds will be

-------
                                                                         57
generated.  Shown here are:  the chemical  abstract number,  the  RCRA index
number, the MS library number for the EPA/NIH/NSRDS Mass  Spectra  Library (the
1983 edition of some 39,000 compounds), the retention time, the GC  retention
relative to benzene-D5, the recommended quantification ion, significant ions
to be used in a specific, reverse search of GC-MS data for  the  analyte  and  the
relative abundance of those ions, and the response factor,  relative to
benzene-De for GC-MS detection.  The last column is our recommendation  of
whether the analyte should be included in further method  validation testing.
The Slide 12 table has been updated to reflect the Method 8010, 8015, and 8020
validation testing so that entries marked "NP" passed the GC-MS suitability
testing but failed PTD testing (see table footnotes).
          Slide 13 shows the purgeable compounds which are  not  recommended  for
further testing.  A number of these compounds either weren't detected at all
or had unacceptably low response factors.  Many of these  compounds  are  some-
what reactive.  Some of these compounds, shown separately in a  later slide,
were included in the Method 8010, -15, -20 validation study, and  they were  all
eliminated from that study also because of the poor PTD characteristics found.
          Slide 14 shows the set of 20 compounds which, based on  some  PTD work
in the 8010, -15, -20 study, we are recommending for further detailed evalua-
tion and/or addition to the analyte sets of the respective  Methods.
          Slides 15, 16, and 17 list the 127 semivolatile analytes  of the 197
tested in the GC-MS suitability study which can be recommended  for  further
testing,  i.e., extraction/solvent concentration recovery.  Seven  of the com-
pounds at the end of the table have rather low MS detection response factors.
Our work-up of the semi-volatile analyte data is still in progress, explaining
why some  of the recommended compounds do not have complete  entries  across all
columns.
          Slide 18 shows the 28 compounds for which we have library reference
mass spectra but still have not been found in our data.  All 28 of these anal-
ytes have been analyzed  in two separate mixes.  There are some  surprises here:
We didn't find rotenone, although in some prior work we were able,  after some
difficulty, to optimize  a  direct capillary GC method for it.   I'm not sure
why we're not seeing tricresyl phosphate unless the data acquisition was
terminated  before  its elution.  Certainly 2-naphthylamine should have been

-------
                                                                         58
found since we did find 1-naphthylamine.  We have also routinely analyzed
methoxychlor using similar conditions.  Since this research effort is still  in
progress, this number (28) of analytes which failed the GC testing may yet be
reduced, especially regarding the four compounds just noted as anomalously not
detected.
          In Slide 19 are listed the 42 compounds which we have not found in
the GC-MS data, but for which our GC-MS data examination is not complete.  Of
these 42 compounds, 31 do not have reference mass spectra in the EPA/NIH
library which makes EICP searching of the data more difficult.  Literature
spectra or manual prediction of the mass spectra will be performed for these
cases.  Essentially all of the work remaining to be done on this study will
address data reductions regarding these 42 analytes.
          Subsequent to the GC-MS suitability testing for the volatile
analytes, we undertook some non-MS detection method validation studies.
Slide 20 shows the 39 Method 8010 analytes plus another 13 from Appendix VIII.
This method uses a Hall electrolytic conductivity detector in the halogen-
specific mode.  The 13 Appendix VIII analytes are the halogenated compounds
which were detected in the GC-MS suitability testing and which we, therefore,
thought had some chance of surviving a validation study in the actual PTD
procedure.  None of those 13 compounds'are on any of the other analyte lists
that EPA methods addressed prior to Appendix VIII.  The results were that more
than half of the 13 Appendix VIII compounds had such low PTD recoveries that
testing was terminated at an early stage.  Of the 13 Appendix VIII compounds,
we weren't able to detect 7 by PTD because they are either too polar
(alcohols, nitriles, epoxides) or they are too chemically labile (especially
the chloroethers and bromoacetone).  In addition to the 7 Appendix VIII
analyte failures, another 4 analytes listed in Method 8010 failed PTD
analysis.  The 11 compounds which were not amenable to PTD analysis are shown
in Slide 21.  Of these compounds not detected by PTD, the tested level was up
to 2,000 nanograms in the purge vessel.  Initially, they were tested at 40
yg/L and, when they were still not detected, the  level was increased to 400
Vig/L and they still were not detected.  We had some PTD recovery of chloro-
ethyl vinyl ether, but not high enough or reproducible enough to justify
further testing.  This compound probably either decomposes in the purge

-------
                                                                         59
vessel, or, more likely, doesn't survive the trap desorption very well.
Pentachloroethane has been definitively shown to decompose during chro-
matography and/or during trap desorption.  When analyzed by septum injection,
pentachloroethane gives an unambiguous indication that it decomposes to
tetrachloroethylene and other compounds in the injector and then continues to
do so during chromatography on the SPlOOO/Carbopack column.
          On Slide 22 are shown the Method 8015 analytes for flame ionization
detection.  The,original Method 8015 analyte set is fairly abbreviated,  with
only 6 analytes.  Our GC-MS suitability testing came up with another 15
analytes that were successfully eluted from the chromatograph and detected
well by the mass spectrometer, and those 15 Appendix VIII compounds were
included in the Method 8015 validation study.  None of the these additional
15 analytes are halogenated or aromatic, although some of them probably  are
detectable with a photoionization detector, for example, the esters and
lactones.
          Slide 23 lists the compounds with low PTD recoveries under Method
8015 conditions.  As expected, most of these compounds are too water soluble
and too polar to be purged at room temperature.  The list contains nitriles,
alcohols, an amide, even a dinitrile.  Paraldehyde wasn't of any particular
concern since we expect there eventually will be better methods involving
derivatization for all of these volatile aldehydes.  Methyl mercaptan, sur-
prisingly, must have some stability problem under the PTD conditions speci-
fied.  We didn't want to give up on this compound and actually expended  a
disproportionate effort trying to get it to perform adequately.  Ethylene
oxide was recovered by PTD but the data indicated breakthrough and/or PTD
decomposition was occurring.  The real problem with this compound, however,  is
that it co-elutes with methanol and there is almost always going to be some
methanol interference.  Since the PTD spiking matrix is always methanol, a
blank PTD run almost always has an appreciable methanol peak.  We finally had
to give up on ethylene oxide testing due to this problem.
          On slide 24 are the analytes tested with Method 8020.  There are 10
analytes in the original method list.  These aromatic analytes are detected  by
photo-ionization.  Another four analytes were added to that list based on our
suitability testing.  Except for styrene, we didn't expect acceptable PTD

-------
                                                                         60
recovery for the four new analytes, and that was, in fact,  the case as
indicated in Slide 25.  Styrene was the only analyte that could be added to
the Method 8020 list.  Pyridine and thiophenol  exhibit some very severe
tailing effects on the GC column (5% SP-1200/1.75* Bentone 34 on Supelcoport).
Although 2-picoline did chromatograph decently, it simply wasn't recovered by
PTD.
          In Slide 26 is outlined some future related work that we expect to
be doing.  In the near term, we have outlined a new study to start looking at
the extraction efficiency of the semivolatile analytes that did chromatograph
under Method 8270 conditions and show detection possibilities using GC-MS.
That initial extractability study will be limited to a reagent water matrix.
          We are also planning a new program starting this spring on heated
PTD, especially for those Method 8015 analytes that apparently failed due to
low PTD recovery.  That work will involve MS detection and procedures similar
to Method 8240.
          Some of the problems that I cited along the way involve chromatog-
raphy.  It is well known that many purgable compounds will not elute from the
SPlOOO/Carbopack B column, so we are presently working on development of a
capillary GC PTD approach.  Our primary goal here is to try to come up with a
chromatography system that could also be used  for Method 8270.  There are
recent reports of using megabore "capillary columns which can handle the same
kind of flow as packed columns and give analysis times that are about the same
as  for packed columns.  These columns are clearly not applicable  for
Method 8270 and also  have a number of other drawbacks.  Therefore, we're
trying to go beyond  that  particular state of development in this  program.
           In the  long term, we're  anticipating some  HPLC-MS method development
for those analytes that can never  be expected  to elute from a  GC  column.
Another crucial area for  Appendix  VIII method  development  and  improvement will
involve the generation of a generic approach for controlling  extract cleanups
of complex  solid  waste matrices.   Hopefully, this research will enable  the
flexible  selection of a cleanup method expected  to  work  reliably  on specific
types  of  waste  matrices.

-------
                                                                         61
                          QUESTION AND ANSWER SESSION
MR. TELLIARD:
MR. LIN:
MR. TELLIARD:
Questions?

Denis- Lin from ETC.   Sam, sitting in the front row,  I  have
the advantage of seeing everything you've shown,  so...

I'll be accused of warning you of that in advance,  but I
didn't.
MR. LIN:
MR. LUCAS:
But for this group of audience, maybe I'd like to share our
experience in terms of having been involved with Appendix
analysis in the past two and three years.  I'm real  happy
that a place like Battelle is really doing the fundamental
work to back up what we have observed.  When all our
research have done foot licking fashion, the conclusions are
not different from what you have seen.  So I just want to
say that our observation has not been different from yours.

Thank you.
MR. TELLIARO:
Any other questions?  Sam, thank you.
At this time we're going to take a short break for coffee
and the tinkletorium or whatever has to be done, and try to
get back here in about 10, 15 minutes, please.  Thank, you.
 (WHEREUPON, a brief recess was taken.)

-------
                                                                                                                                             62
CO

 OJ
                                                    t-
                                                    o>
                                                    en
                                                    ra
                                                    c
                                                    10
                                                    ro
                                                    i. O
                                                    cnio
                                                    O r-~
                                                    S- ,-H
                                                   O-  I
                                                      ro
                                                    I  O
                                                    i   i
                                                      CO
                                                    CU IO
                                                   .^
                                                    O 4->
                                                    o  o
                                                   CJ  to
                                                       t-
                                                    10 4->
                                                    3  C
                                                    O  O
                                                    t- CJ
                                                     O
                                                                   i- CSJ
                                                                   rO O
                                                                   -c co
                                                                    • o
                                                                  LU CO
 ro  Cn
+J  C
 rO *r—
O  w
    t/1
t/>  01
SL  O
 I   O
CJ  i.
O D-
             ro C
            -C O
             (I) •.-
                                                                                   cr
                                                                                O '

                                                                                ฃ
                                                                                ro
                                                                               O-
                                                                                                       TD
                                                                                                        O
                                                                                                       4-)
                                                                                                        O)
                                                                                                        ro
                                                                                                       U-
              rz
              o
              
                                                                                   S-  ro
                                                                                   ,
                                                                                                          en c
                                                                                                          ro ra
                                                                                                       >, a. E
                                                                                                       !-  E ro
                                                                                                       S-  ro T3
                                                                                                        ra •!-
                                                    •i- 10 C
                                                    C 3 3
                                                    <: oo uj
              tn
              01
              >1
              ro
•5
               1
 O)
•a
CO
           ts
             . o:
           L)

           4-J
           ra

           CU
    C.
    OS
               o

               0)
               01
              o
                                    ,-2   "
                                    .  E  tn
•^O E

 E  (y Q


   "cu
                                      ro
                                     CO
                                         O
                                     1	 4->
                                     ra  ro
                                  E~ 4J  L.
                                  o c  o  0
                                  +J CU -O  u
                                  •(-> E  <0  ฃ;
                                  o c _J S
                                  J3 O    ฐ
                                  Ol S- +->   „
                                  C -r-  V. ._
                                  Q >  C3 J3

                                           ' ro
                                                                                                      csi
                                                                                                      CM
                                                                                                      CO
                                                                                                        o;
                                                                                                        o
                                                                                                                   CJ3
                        00
                        10
                                                                =  ฐ- :
                                                                LU  Q-
                                                  LU   • CO ,
                                                  
                                                       c
                                                   4->  CU
                                                    o  E
                                                    rO  C
                                                    S-  01
                                                                                                                                                 in
                                                    c
                                                    o
                                                   CJ •*

                                                   •a: j^
                                                   o- s_
                                                   LU O
                                                                                                                        o o
                                                                                                                        >—< CVJ
                                                                                                                        h- o
                                                                                                                        =t CO
                                                                                                                        o
                                                                                                                        •-< -a
                                                                                                                        _j c
                                                                                                                        •f. ra
                                                                                                                        >
                                                                                                                            uo
                                                                                                                        O I-H
                                                                                                                        o o
                                                                                                                        :c co
                                                                             >- o
                                                                             a: co
                                                                                                                                                I— oo
                                                                                                                                                ซ=C o
                                                                             o
                                                                             CD
                                                                                                                                                           LU ซt
                                                                                                                                                           _1 Q_
                                                                                                                                                           C3 LU
                                                                                                                                                           <-> err
                                                                                                                                                           CO O

-------
                                                                                                                      63
  O -f-
                                                                            in
                                                                           -o
                                                                            o
                                                                    +->
                                                                "O to  E -t->
                                                                -a CD  o  QJ



r—
•r—
03
U_










i
u
re
i.
+j
X
QJ

•o
Q)
•r—
(4—
• r-
-D
O
2-
1 — 1

t/^
E
0
•r-
4->
•r —
"C
C3
O
u

c
o
•I —
4->

1
r—
o
in
^-^
i_ .
QJ U
-E +J
•(-> QJ
O
rt
" in
bJ -M
d- E
to QJ
" — ^


QJ
>
O
S-
CL
E
•r—
*>^^
Q.
O

"QJ
>
QJ
Q
CM




1
O
s_
Q.

Q.
3
E
re
QJ

u










i/)
QJ
S_
zs
-o
QJ
O

                                                                                                    to
                                                                                                    c
                                                                                                    o

                                                                                                    +J
                                                                                                    Q-
                                                                                                    O

                                                                                                    s-
                                                                                                    O)

                                                                                                    +J
                                                                                                    o
                                                                                                    (D
                                                                                                    Z3
                                                C  C
                                                o  o
                                                           c/1
                                                          -O
                                                           O
                                                U  U
                                                QJ  re  o
                                            O -1—5  i_ .i— 4_}
                                            I—  E -4-> 4-)  QJ
                                            o. -i-  x  re  E
                                                    QJ i—

                                            QJ  u  o ^ c:
                                            4->  QJ  i- 4J  I
                                            re  s-  u  to  E
                                            QJ -r- -r- -i-  O
                                            •3^ a-5L cz-zz.
 OJ
•a
"O
•o
                                                                            to
                                                                       is> -a'
                                                                       S  O
                                                                         I  -E
                                                                        E 4->
                                                                        O  CD
                                                                        E  E

-------
oo
CM
rn
oo
r-.
ซ*
CM
                                                                  64

VI
^

CO
^f

S
4->
IO
r J


^ *o
r^ cn
OJ S
•o u
r— y
(/"} =

^***
>— t

>

X
•r~
•o
C
QJ
o.
o.









C
0
cn .i-
C OJ 4->
•i- cn u

0. i. i.
E 0 4J
10 4J X
CO CO LU

1 1 1


C
0
J->
to
M
'= >,
to cn
ID 0 0
OJ ฐ 1
.— O J=
J= 4->
co ^ ^
u
cu
C
01
CD
1
1
1

OJ
4->
^)

IO
cz
*c





r_
CO
C in
en c
•r- -r- i/,
1- E QJ
O 4-> •
* 	 4J ^
in in 1o
o) .1- > LU
t— C -r- -O
tO to S>~ E QJ
C Cn o O >> T3 .
to v -r- i_ i — 3
-C i_ It- , — , — r
O U O. to O
•1- •!— QJ ••— X
C SI  cu .— E to
0 ZJ C 4-> -r- E -!-> ^
^~ fO ro rO O O
t- O-J-' C > O 1— f
O 3 *i— to
"-• •— E S-
5- >— I r— -r- 4-> O
. Ol •-", O r- O
J3 5> Q. Qj c: >,
E s_
3 X >-, in in o
C -r- 4-> QJ Q> 4->
T3 -i- 4J 4-> -r-
i — c S- >, >> in
tO CU O c— i— O
+J CL-t- to to O.
O D. S- c C QJ
1— eE Q. ff. ซt 1_
• • •
= >,
0 J=
•r- O.
4-> tO
10 J.
N cn c
•r— O O
CJ. 4J 4-> •r-
3 rO to 4->
C > E 0
tO 'r— O CU
CU S- i. 4J
•— O) -C 0)
O Q CJ O

III

Q.
_1
CJ
1 ^
4J Jl! 10
cu t c
C X ซ" 10 0 0
41 ! .ซ 8 !
0 E .i ^z^J
- i B
5
n3
CJ
to




X cn
•r— 4_*
4- C
4-> CU
to E
s: cu
.^:
cu
to
to
o



^

in
0)
O 00 4J
Q) QJ IO
/) tO -p-
QJ T3 <1J T3
— T- t— C
TD '^ IO
in c +J u
cu cu to to
U r— O i— ^->
>, -^ o Q
ro to O -r- CM
C i— •* E 03
f. O CM CU
> OO in -o
— Q
ro in T3 in jc
u to o to 4->
O -C QJ
~ "O 4-* X3 ET
QJ QJ <1J
| ^ JC" | \ M
in i/> Q)
OJ •ป CU i —
' 4-J QJ 4J jQ
i— 10
in JD in .ป->
cu to cu o
•P QJ +J ro
>, cn >, s_
' r-? S- ^4J
tO 13 ro X
CO- C UJ
• •










_l
--. to
_l cn +J
3. C to
TJ
CM O
in oo
•ป t— i o
"o 4-1 z:
c to —<
i: J= -o
4J O C
OJ tO tO
€ OJ
oo
C ฃ
•ซ- X
•r- O
C E 0
O CM
•r- S.  OJ
O CL •
CU  , — . C3
C OJ T3
•i- 4-> CU O
>> 4J O
E i— 0 U3
rj to OJ cn
4-> C ••-)
Q- 10 C
QJ -r-
10 IT) C
4-> c cn E
O U O -i- OJ O
^f QJ 4-> O C 4-> f^-
CM l_ O C I/) CM
OO •!- C3 rO -i- >, CO
Q i— 1 	 U. t/>
•O T3
O O
J= • • • J=
! < , ^
OJ OJ












































OJ
c oo
cu o
">> 3. z:
j= --^ >-.
4J cn
cu c -o
O to
^— 10
>, 1 00
r— SZ
4-> 4->
tn . to o
O CO
E J= CM
— o ro
IO
c cu c
•ป- rtj
• CO
C X T-
o — c
'•" E c
O i. U.
OJ CU
•>-J Q. •ป
c: o
•i- 10 —~ O
cu -o
— J 4J CU CD
3. >,4J ID
i — U t— I
CM to O) ซd"
C "~3 E
ซ 10 C to OJ
tn oj CO s- 
_ ir- c: in
4-> $- O O
•i- O 4-> O i — to
r— t— O !- 4J
O- .C m ซ — 1 to tO
I/) O i-H 	 CJ -O


• • •







-------
                                  Slide 8

                            Volatile  Analyte  Results
                              (54  Analytes  Tested)
                                                                      65
GC-MS.Suitability Testing

          •  Analytes recommended for PTD testing

          •  Analytes not detected

             Total

          •  Analytes detected with low response factors

Method 8010. 8015 and 8020 Validation

          •  Analytes passing GC-MS suitability testing
             and also passing PTD testing

          •  Analytes passing GC-MS suitability testing
             but not recovered by PTD

             Total
     39

     li

     54

      6
     20



     li

     39
                                 Slide 9


                        Semi volatile  Analyte  Results
                               (197 Analytes)
          •  Analytes recommended for extractability
             testing
          •  Analytes requiring further evaluation
127
 42
          •   Analytes  not  detected
                                                               28

                                                              197

-------
                                                                                                                                                                       66
                                                                                   CD
                                                                                   Z
                                                                                  fa

                                                                                  CL.
u
z
Ii  Z
I.  O
I*  *-
I   to
I   to
I1.  1-1
i   o
II
   OS

   ฃ


   o
   to
ป-*  II
•J  II
t-l  II
      o
      z
X  '"
 I   11
o  ซ
U  H


ill
a
EC
E-
I-
V.  U
C  I
o
a.

ง
u
n
n
II  C4
II  U
II  Z
U  <
I*  (-
II  1/3
.   o;
i'  a
   to
   IV
                                                                                  o
                                                                                  CO

                                                                                  2
                                                                                  DS
                                                                                  E-
m
•<
E-
CO
Z
=>
      E-
      Z
      <
      E-
      =1

      O
      0.

      >-  K
      E—  i "
ta    i-  as
<    cs  <
E-    O  E-
to    >-  co
Z    DS  Z
a    a.  =>

>    B  >•
E
                                                         MM
                                                         o:a!
                                                         ==>
                                                         E-EH
                                                         KX
   < ce  u:          <    X  <       o u    ca
   o -3  ^          u        u       M S    _3
   M i-.  >- f- E-    i-    E-i-c       Dub.    1-1
CO E- E-  E- Z Z    E-    ZE-    UWMUH
a><<<<    >•    <>•    a o u a <
&* *^ -—  ป-3 a a    i-3    a  1-3    ^ so os ^* 1-3
BOOOZZ    O    ZO    BO.CLBO


-3>-ooci:ttiu>-oM
-<     H
                                                            CO     S
                                                            Z     ,-3
                                                            O     -a
                                                            U     O
                                                            zcu
                                                            "-1
                                                               Cfl  >•  ,
                                                            BJ -a  E- .-i
                                                            os ec  w ci
                                                            B •<  cs 1-1
                                                            E- E-  0
                                                            -<-
                                                            JZCBO
                                                            O3O.Z
                                                            Z
                                                            ca>.B:3:
                                                                         w z < I
                                                                         5?
                                                                         flu
                                                                         S
                                                                         o
                                                                         as
                                                                         CS
                                   Z
                                   =>
                                   O     CO
                                   a.     E-
                                   Z     Z
                                   0     M
                                   u     sป
                                          -3
                                   uo
                                   i—co
                                   Z
                                   <     u
                                   o     •-<
                                   B     z
                                   0     <
                                          C3
                                   CeCE
                                   OO
                                   U.
                                          z
                                                                     I—I 1*4 I—I C—  t—  US
                                                                     f- E E- Z  Z  0,

                                                                     i-a H 1-3 a  a  os
                                                                     O CJ O Z  Z  0.
                                                                                               o
                                                                                               co
                                                                                         >•
                                                                                         B
                                                                                                               CL
                                                                                                               3
                                                                                                               o
                                                                                                               as
                                                                                                               U

                                                                                                               o z
                                                                                                               Z M
                                                                                                               B ,j
                                                                                                               O ซ
                                                                                                               a. o
                                                                                                               Z OS
                                                                                                               OD.
                                                                                                               u
                                                                                                                  >•
                                                                                                               o t-
                                                                                                               h-i-i
                                                                                                               ZJ
                                                                                                               < I-H
                                                                                                               uซo
                                                                                                               cs <
                                                                                                               0 E-
                                                                                                                  CO
                                                                                                               CB
                                                                                                               OJ
                                                                                                               b,<
                                                                                                                  E- S
                                                                                                                  < u
                                                                                                                  03 O
                                                                                                               ca a.
                                                         >• a:  <
                                                         sea.w
                                                                                                            Z
                                                                                                            <

                                                                                                            >-
                                                                                                            os
                                                                                                            ซ:
                                                                                                            z
                                                                                                            ฃ
                                                                                                            M
                                                                                                            E-i
                                                                                                            J

                                                                                                            E-i
                                                                                                                                           o
                                                                                                                                           <
                                                                                                                                           CS
                                                                                                                                           u.

                                                                                                                                           co
                                                                                                                                           <

                                                                                                                                           ca
                                                                                                                                           J
                                                                                                                                           ea
                                                                                                                                           <
                                                                                                                                  oa       EH
                                                                                                                                  i-3       co
                                                                                                                                  eo       z
                                                                                                                           E-
                                                                                                                           J
                                                                                                                           <:
                                                                                                              1-3        CO
                                                                                                              <  W  u <



                                                                                                              E-  E-  E-i i-3


                                                                                                              O  O  o p-a
                                                                                                                                               M M    W
                                                                                                                                               a. a.    a.
                                                                                                                                               OOO
                                                                                                                                               c> 5    S
                                                                                                                                               CO CO    CO
                                                                                                                                              •
                                                                                                                                               E- f-    E-
                                                                                                                                             •  Z Z
                                                                                                                                               ca ca
                                                                                                                                               CO CO
                                                                                                                                                                    CO
                                                                                                                                                                    K
                                                                                                                                              >.
                                                                                                                                              ffi
                                                                                                                                                          cs DS
                                                                                                                                                       KO.0.0.
                                                                                                                                                       ij
                                                                                                                                                       to a os     co
                                                                                                                                                       < fie s     B
                                                                                                                                                       E- H S     ฃ-'
                                                                                                                                                       co
                                                                                                                                                       ZttfECE
                                                                                                                                                       3OOO
                                                                                                                                                          U. to.     U.
                                                                                                                                                       >-
                                                                                                                 u  < < < sa <



                                                                                                                 E-  E- 0. 0. E- Cu

                                                                                                                 Nj  i-3 B OS i-3 CS
                                                                                                                 O  O 0. 0, O 0.
                                                                                                                                                              1-3
                                                                                                                                                              o
                                                                                                                                                                        M CO
                                                                                                                                                                        0. CU
                                                                                                                                                                     CO  => =>
                                                                                                                                                                     ooo
                                                                                                                                                                     o  os CK
                                                                                                                                                                     SOW
                                                                                                                                                                     E-i        W
                                                                                                                                                                     w  a a iJ
                                                                                                                                                                     Z  Z Z 1-1
                                                                                                                                                                        = 3 eo
                                                                                                                                                                     <  O O <
                                                                                                                                                                     o,  o. oป >J
                                                                                                                                                                     USE
                                                                                                                                                                        0 0 >•
                                                                                                                                                                     os  u u >J
                                                                                                                                                                     K        >J
                                                                                                                                                                     3;  o u <
                                                                                                                                                                     E,  M ,-, z
                                                                                                                                                                     0  Z Z OS M
                                                                                                                                                                        <•  u ts B co
                                                                                                                                                                     as  os K E- <
                                                                                                                                                                        O O    H
                                                                                                                                                                     now
                                                                                                                                                                     MKCBZZ
                                                                                                                                                                     NJOO<3
                                                                                                                                                                     l-ibb.
                                                                                                                                                                     a        >• >
U < < <  < W M
< M M u  u 1-3 i-a

< 0. B. E-  E- E- E-

o os os iJ  -a ซJ J
   0. 0. O  O O O
 I   '*  .  i   ;   i   i   i   i"  i   i   i   i   i' i   i  i


 ii      !   i   i   ,   t      i   :   i   i   i   i   i  i
inmocoNrrmcscc*-
                                                         o co
                                                         CM
                 I   !   '   I  I   I   I   I   I   I   I


                 I   I   I   I  I   I   I   I   I   I   I


                   co ~<        co m --i    .-i
                   o
                   N



                                H


                                co
                                                                     a
                                                                     o
                                                                                                         i   i   i   ;   i  i   i   •   i   i   i   i   i   i   i   ;   i   i   i    i   i


                                                                                                         i   i   i   i"  i  i   i   i   i   i   i   i   i   i   i   i   i   i   i    i   i  T "i  7  i  T "i
                                                                                                                                                                                                              H
                                                                                                                                                                                                               * M3 M
ES B O
ta o >•
B    X
•-a .J o
< >• z
H E- ซ
ce co E-
u u u
< < <
                      ca
                         u co
                           : H  co
co
a
>•
                        o
                        ~
                        B  a             _
                        U  >•    Q CJ b. i-t
                        OB    I— l-i i-3 CS    B
                        B        B Z E> O i-a iJ
                        BCO    O O CO J l-l <
                     W >•  Z    i-3tO    = O E-
                     ZB>-lCOBBCos
                     MCOI-ll-i    CO>HBZ<<
                     OZCOK>JZaEซu
                            in co t~  co en o i-i i
                                                                           cs
                                                                           o
                                                                        CO i-3

                                                                        CO U
                                                                        OS

                                                                        0. >
                                                                                  OS Q
                                                                                  < o
                                                               B
                                                               >
                                                               B
                                                               CO
                                                                                     o
                                                                                     H
                                •< e-

                                > B
                                                            m B i-i
                                                         Z M < Z
                                                         M a c> <
                                >• ca
                                Z O
                                                                                  a <       co
                                                                                   i           a
                                                                                  co u co    >•
                                                                               co i-i I-H a a B
                                                                               z oa Eซ >• M co
                                                                               I-H    co B u a co
                                                                                                           co
                                                                                                           co
                                                                                                         a B
                                                                                                                             O
                                                                                                                             09
                                                                                                                             CS

                                                                                                                             o
                                                                                                                             o
                                                                                                                       o
                                                                                                                       Z  CO
                                                                                                                       a  i-i

                                                                                                                      ,is
                                                                                                                      : co  os
                                                                                                                                             >•  z

                                                                                                                                          Z CO  B

                                                                                                                                          CO 0.  HI

                                                                                                                                          ซ CO

                                                                                                                                          B a  co
                                                                                                                        D_ CO Z EC
                                                                                                        w
                                                            E-ซ*cD>->-WCCt-ii-iceM<
                                                                                                                                              o M <
                                                                                                                                                                             O
                                                                                                                                                                             NI
                                                                                                                                                                                a co

                                                                                                                                                                                O <
                                                                                                                                                             o  NJ  >-* PS    o E-
                                                                                                                                           O  •-=
                                                                                                                                           cc  co
                                                                                                                                           o
                                                                                                                  *3woK<^>'a:c?CE!Q>HCOHjaiซCL.CQODSOWCUtOC*&a3
                                                                                         i  oz>*ozzuceui-Hซ>*c(3i-t>-u  -  - o o
                                                                                        voi^sso:ce>^Mป
                                                                                         -ZO-E-<.-JC>O'-3>-W<
-------
67
n
COMMERCIALLY AVAILABLE

OJ EH
^ o
r^ Z
oo
Ed
i-4
i-H
EH
•<
t-3
o
>
M
2
Ed
CO

o
z


Ed

M
EH
^5
,.3
0











II
II
II
II
II
II
II
II
II O
H Z
II
M CO
II •<
II o
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II K
II 0
H Z
II EH
II CO
H PS
II ฃ3
II CO
II
II
II
It
II
II
H
II
II
H
II
II
II
II
II
II
II
II
II
II
II
H K
II Ed
II CQ
II 32
II P
II Z
II
II
11
II
II
II
11
II
II
II
II
II
II
II
H t^ O5 CM O CO CO
II 1 1 1 1 1 1
II in O Ci 00 i— 1 CM
II i— 1 O) 1C 00 i-H CO
II 1 1 1 1 1 1
n 05 co ^ oo r- m
M CM o o: co r-t o
II r-l CM i-H 00 00 CM
II t- CO
II
II
11
II
II
II
II
II
II
II
II
II
II
II
II
ll Ed
II Z
II Ed
II N
II <
II I-H ^
II K Ed
II EH Z
II *~* i— i i-3
II r3 i-3 O
II W >- rH Z
II Z Z Z <
MOW < EH
ii z as -53
H rH O.
n S O
HOB;
II < O
II PS i-3
n as BB
II EH O
H Z 1
II < Q,
ll O *— '
ll OJ 1
II EH I-H
II M 1
II Z t-3
11 | &-
ii I-H as
II I EH
II hH U
II >ซ aS
ii as M
II EH O
II Ed 1
II ฃ CO
11 |
II CM CO
II
II
II
II
II
II
II
II i-l CM
II
II
II
II
II
II
II
>•
w as
1-3 EH
O W
N 3S
< 1
CQ CM
OS '-'
< CO
O r-i
O co
N Ed
Z Z
Ed Ed
CQ ซ-a
t-4 >•
a as
1 E-
CO Ed
-as
in t

*sj* ^*
. .
CO 5ฑ







CO <*







CQ
1
i— 1
1
S*^
O Ed
Z Z
rH Ed
3S BB
< EH
0 Z
CO <
0 OS
OS O
EH S
CH rH
Z tv
i-3 N
P- SE
EH Ed
P CQ
ซ 1
s_ ' 00
1








in co







^ rH 00 00 0
1 1 1 1 1
I-H co co CM m
m o m oo t~
i i i i i
in "ป 00 rH O
CM co in co t—
CM ^ co m m
CO CO









CM
Ed Ed
Z Z .
rH I-H O
O N Z
rH <
PS BB 0
o cu w
< < OS
>-v Z
O O CO
>—DS D
N O OS
Z tJ EH
Ed BB M
CQ U U







C— 00 CT3




















Ed i-3
-H 0
O N
N <
< M
rH 83
as EH
EH OS

OS b
S M








O rH
rH rH






CO O
*• 1 1
1 ^t O
C~ C55 ป*
in i i
1 O5 OT
rH CO >*
co I-H m
ป*
Ed
Z
rH
ฃ
<
>-3
>•
Z
rH
>
HH
>•
as
EH
Ed
Ed Ed Z;
rJ 0 O
O rH CO
N N O
< < OS
0 rH EH
rH aS rH
OS EH Z
rH M 1
Z Z Z







CM CO ^f
rH rH rH






16543-55-8
SPHORODITHIOATE 3288-58-2
4104-14-7
NE SULFIDE
66-75-1
1615-80-1
75-87-6
o
BE
Cu

1-3
W >•
z as
M EH
EH Ed
0 ฃ
O 1
rH CO
Z 1
OS tJ
0 >•
z as
O EH
CO Ed
o as

EH Q
rH 1
Z O
1
Z 0







in co
i— i i— i






rH
as
a.
CO
O
as
0.

r~*.
i-3
>ซ
Z
rH
O
I-H
•ฃ. BS
M Csl
EH <
Ed 1
O I""1
^ x_x
co co
O rH
as as
OH EH







t- 00
rH rH



i







W
z
rH
N
<
OS
a o
OS >•
< BB
H i-3
CO >ซ
3 as
as H
Ed r4
M- rH <
rH 0 OS
0 1 0
^ CM r-3
os -as








O5 0 rH
r-l CM CM






II
II
II
II
•11
II
CM II
i II
CM II
ซ# II
1 II
CM II
O5 II
CO II
II
11
11
II
II
II
II
11
II
II
II
II
It
II
II
II
II
II
II
II
11
11
II
II
II
II
II
II
II
II
11
11
II
II
II
II
II
II
II
II
II
II
Ed ll
Z ll
rH II
CO 11
OS ll
< II
h-5 II
> II
BB II
EH 11
Ed II
rH |l
0 II
II
II
II
11
II
11
II
CM II
CM II
II
II
II
II
II
II

-------
                                                                    68
             Slide  12
RESULTS OF 6C-HS SUITABILITY TESTING OF VOLATILE ANALYTES
NO. SU8SIANCC
I ACnONlTfUU
Z AILTL ALCOHOL
3 ALLYL CHLORIDE
4 iEHZTL CHLORIDE
S 8IS-U-CHLOROETHTL) SULFIDE
6 IIS(CHLOROHEIHU) ETHER
; JRMOACnONE
8 2-IUTANME PEROXIDE
9 MUFAHOHE
10 CARBON DISULFIDE
11 CHIMAI HYDRATE
12 2-CHL8ROETHAIWL
11 CHlOROHETHtL HETHYl ETHES
!4 CHLOROPREHE
J-CHLOROPROPIOHtlRILE
1 ,2-D16RW-;-CHLORl)PROPANE
1? DUROHOHฃTKAHฃ
18 IIICHLORODIFLUOROHETHANE
1ซ 1.4-6ICHLORO-2-8UIENE
SO i,3-DICaORO-2-PROPAHOL
21 1.2,3,4-D1EPOXY8UTAHE
22 U-eilOTlKYDRAZIHE
23 1,2-BtHETHttHlfDRAZINE
24 1,4-BlOXAHE
25 EPICaOROHYDRIH
26 EIHYLEK DISROHIDE
V EIHY1EHE OXIDE
28 ETHYLEHIHINE
2? EfHIt KETHACRYLAIE
JO HOACHLOROPROPEHE
31 H-12-HtDROXlETHYLIETHYLEHEININE
32 2-HfDROXtFROPIONIISILE
33 IS0IUML ALCOHOL
34 HAIOTOHJTRILE
X KETHACRU'SirrRILE
34 2-HEIHYLA2IIUDINE
3? REnซLHJD8AMHE
38 ISIHJL 10DIK
JO KETHtl ISSCTAHATE
40 2-HEtHYliACIOHITRILE
41 KEIKtl HfRCAPFAH
42 KEIHIL HEIHACRYLATE
4} rtNIACHLCROETHAHE
44 2-PICOLIHE
4$ PR0PARSYL ALCOHOL
44 b-PfiOPIOLACTOHE
4? PSHPiOHnSILE
48 ft-PROPTLAHM
& PY9IBIHE
50 STYREHE
SI 1,1,1,2-TETSACflLOflOETHAHE
52 TEIDANITROHETHAKE
S3 [HIOPKENOL
54 1,2,3-TRICKLOROPROPANE
(ป) EM/HIH-HSRDS/H8S Hass Spectral Data Base,
(W Retention indti relative to benzene-D6.
Area (quant ion)

Area (i/e 82, beiuene-D6)
(d) HB •• not detected.
CAS NO.
75-05-8
107-18-6
107-05-1
100-44-7
505-60-2
542-88-1
598-31-2
1338-23-4
78-93-3
75-15-0
75-87-6
107-07-3
107-30-2
126-99-8
542-76-7
"6-12-8
74-95-3
75-71-8
764-41-0
96-23-1
1464-53-5
57-14-7
540-73-8
123-91-1
106-89-8
106-93-4
75-21-8
151-56-4
97-63-2
1888-71-7
1072-52-2
78-97-7
78-83-1
109-77-3
126-98-7
75-55-8
60-34-4
74-88-4
624-83-9
75-86-5
74-93-1
80-62-6
76-01-7
109-06-8
107-19-7
57-57-8
107-12-0
107L10-B
110-86-1
100-42-5
630-20-6
509-14-8
108-98-5
96-18-4
RCRA NBS LIB.
HUHBtR NUHBER(a) RT
U003
POOS
U317
P028
P158

P017
U160


U034
P133
U046
U276
P027



U074

U085
U098
U099
U108


U115
P054
U118
U243
U289


U149
U152
P067
P068
U138
P064
P069

U162

U191
P102 •
U302
P101
UI94
U196
U323

P112
P104

4
69
313
3603

2309


223
310
6678
355
356
698
745
20858
11361
2963
3297
3829
584
90
92
705
316
13846
15
11
2341
22282
657
203
290
125
140
60
23
5939
58
532
28
1230
16044
843
49
214
44
81
349
1625
10335
15217
1903
6685
(HIM)
3.97
9.77
8.83
29.50
33.53
RECOHHENDED RECOHHENDED SEARCH IONS IN DECENDING ORDER . RESPONSE
RRI(b) uUAN ION OF IHPORTANCE (RELATIVE ABUNDANCE) FACTOR(c)
0.23
0.57
0.52
1.73
1.96
41
57
76
91
109
ND(d)
16.33
ND
12.20
7.47
15.77
12.93
ND
14.77
17.37
27.23
12.53
2.47
22.73
21.83
14.87
ND
ND
13.70
13.10
18.40
1.30
ND
23.53
ND
ND
18.97
13.80
19.60
12.37
' ND
ND
5.37
NO
ND
ND
19.77
24.83
23.20
10.77
13.00
8.53
23.00
18.57'
30.83 .
20.33
ND
ND
22.20
0.96

0.71
0.44
0.92
0.76

0.87
1.02
1.60
0.73
0.14
1.33
1.28
0.87


0.80
0.77
1.08
0.08

1.38


1.11
0.81
1.15
0.72


0.31



1.16
1.46
1.36
0.63
0.76
0.50
1.35
1.09
1.81
1.19


1.30
136

72
76
82
49

53
54 '
157
93
85
75
79
55


88
57
107
44

69


44
43
66
41


142



69
167
93
55
42
54
59
79
104
131


75
41 (100) 40 (53) 39 (20)
57:(100) 58 (20) 39 (30)
76 (30)' 41 (100) 39 (65)
91, (100) 126 (20) 65 (14)
109 '(100) 111 (35) 158 (20)

43 (100) 136 (7) 138 (7)

43 (100) 72 (23)
76 (100) 78 (10) 44 (12)
82 (90) 44 (100) 84 (70)
49 (12) 44 (100) 43 (60)

53 (100) 88 (60) 90 (20)
54 (95) 49 (100) 89 (20)
157 (87) 75 (100)155 (63)
93 (100) 174 (80) 95 (82)
85 (100) 87 (35) 101 (10)
75 (80) 53 (100) 77 (2JI
79 (100) 43 (85) 81 (33'
55 (100) 57 (15) 56 ('ป


88 (100)' 58 (70) 43 (301
57 (100) 49 (22) 62 (15)
107 (100) 109 (95) 93 (6)
44 (100) 43 (20) 42 (15)

69 (100) 41 (70) 99 (20)


44 (100) 43 (60) 42 (15)
43 (100) 41 (66) 42 (60),
66 (100) 39 (14) 65 (10)
41 (100) 67 (50) 39 (51)


142 (100) 127 (37) 141 (15)



69 (80) 41 (100)100 (30)
167 (77) 130 (79) 132 (77)
93 1100) 66 (36) 92 (27)
55 (100) 39 (25) 38 (15)
42 (100) 43 (32) 44 (5)
54 (100) 52 (17) 55 (16)
59 (100) 41 (47) 39 (25)
79 (100) 52 (73) 51 (33)
104 (100) 103 (50) 78 (42)
131 (100) 133 (97) 117 (70)


75 (100) 77 (30) 110 (33)


78 (10)
128 (5)
160 (12)

93 (6)



86 (35)
51 (9)

51 (23)
91 (8)
77 (30)
172 (40)
103 (7)
124 (10)
49 (27)



57 (23)
51 (8)
188 (3)


86 (15)


53 (51
' 74 (9)
38 (11.)
52 (22)






39 (47)
165 (57)
78 (17)
53 (10)

40 (7)

50 (25)
51 (25)
119 (68)


112 (20)






95 (6)



111 (33)
80 (5)




176 (36)

89 (35)









114 (5)





66 (20)







169 (37)






77 (20)
95 (33)


97 (17)
0.34
0.18
0.11
0.56
0.02

0.03

0.17
0.68
0.009
0.009

.11/.06
0.17
0.26
0.31
0.18
0.14
0.20
0.14


J.24
0.34
0.56
0.19

0.58


0.009
0.28
0.15
0.34


0.51



0.38
(note e) 0.12
0.43
0.005
0.08
0.26
0.01
0.37
0.67
0.31


99 (12) 0.62
STATUS
CODE(f)
NP
NP
OK
OK
LR.NP
ND
HP
ND
OK
OK
LR.NP
LR,NP
ND.NP
OK
NP
OK
OK
'OK
OK
NP
NP
ND
ND
NP
NP
OK
OK
ND
OK
ND
ND
LR
NP
NP
OK
ND
ND
OK
ND
ND
ND
OK
NP
OK
Lfl.NP
NP
NP
LR
OK
OK
OK
ND
ND
OK
1983 version, 39,763 lass spectra.

/ Aiount

/ Aiount


analyte

benzene-D6

tt) Ihf base peak at i/e 117 was not used due to an interferance at that
If! 01: apparantly analyfable by root teiperature PTD, uith or nithout
His not detected in 6C-HS data







lass uith a
HS detection






nearly







coeluting







internal







standard, chlorobenzene-D5.


































KPซ not recovered by PTD using root teiperature purge
.8: Ion response factor












-------
UJ  V_>
CO	
"X.  0=
CD  CD
                CDCDCDCDCDCDCDCD
                                                   CD CD CD  CD
                       —• C^ CO
                                                                                                                                                                               g
                                                                                                                                                                              •g
                                                                                                                            CD  CD  —H  CD —t —t

                    CD  —  CXI CO CM
Scr-  -ซ3-  NO  ro r-*. r-^
  M5  C^ CSI
            -M UTJ
                                                                                  CSJ CSJ lO CO
                                                                                                                    CK CT' -ซ•
                                                                                                             NO  '*O O--
                                                                                                             CD  •—< CD
                            —ซ  co co
                                                                                                                                                                               —

                                                                                                                                                                       S     ฃ
                                                                                                                                                        (/I      --H
                                                                              ฃ
                                                                              ฃ
                                                                              i
or  oe  or  -ซi  3r ^x
*—  CD  ป—      CD _J i
*-*  	i  i—i  _j  oe CD i

Cง  S  CD  *^  S ol!
•—  -i  z  =D  :c <=>
a_  H-  CD  CD  co oc :
CD  Z  —I  CD  •—' 0*.
OC  UJ  -^  CO  Q-  I
oป  <%.  ac  *—ซ  uj ja '
                               •<:
                               3E
                               -^
                               >~
                               cj
                               CD
                               co
                                               .    —  -  -                  —
                                               *— ซ  O=  CD  C_d -V 3C ULt •— i
                                               ac  c=>  "—"ป—•   - cs 3Z 3C
                                                                                                     1_LJ
                                                                                                     ar         ,    uj uj
                                                                                                     ^-  uj         ae :c
                                                                                                     uj  — j     uj ซ— • i— *
                                                                                                         >— ^     f~^ i^-j r^j
                                                                                                     — ^  OC  UJ t— I *C -ซt- — I
                                                                                                     __i  i—  ac x o= o= cr>
                                                                                                     >~  >—*ป—ซ CD ca ca ^
                                                                                                     3=3E(=0=S->-CD
                                                                                                     t—  CD  •— ' LU ^ ^ CJ  •
                                                                                                 ujujh—  Qea._j_j_j
                                                                                                 =c  To  ac  ac  ซ=ป  Csj -i-. o —i  co  c_5  ck. rr
                                                                                                 I—.—*]    l    |    -   - 0= 3T  I—•   I    I   I
                                                                                                                                                       T  ง
                                                                                                                *->  CD  *-ป  O  tO
                                                                                                                OL>  O>  OL>  O>  CJ
                                                                                                                -O  Q.  T3  w_  w



                                                                                                                ~O  CO  O  O  O

-------
                                                                                                                                                                  70
        in
        S3
        a"
        UJ


        3
                \  s.
                 j  5

                ii  -ซr
                 I  UJ
                 '  CO

                 i  CA

                 '.  &
fm^f  H      vo  F-O ro uo  ro  •••* LO uo  *—t  ro  NO NO uo  **H  ^y  1—4 ซซ^ CM  NO  ro

CD  It      CDCDCTJCDCDOCDCDCDCD^CDO'OOCDCSOCDCD







                                                                                CM




                                                                                ON
                                                                                ON




            CD*             "e?         —-     NO                        LO     "^ "ro
            CM             CM         uo      ro                        ro     —*  ro








            "uo" "uo" ^     ^     •—- LO  —-  CD  ro     —* or*-.     CD CD  CD  co




            —*  CD T>     CM     CO NO  ro  CM  —-4     CO  CO  CO     •ซ• f^  CM  ON  11


                                                                                        it






            CO—ซ CD —M  ON  CM MS ON  —i  UO  CD -ซ9- UO  ON  CM     P-*. up  CD  t^.




            """""* "-"fc ^"^ """*" "™^  """" *""""" "^"~ **"""  *!"?  """""" T"T ***^r Ss?  *^^  "^^ S5 S2  ^^ ^^










            "CD" "CD" — cs  CD*  CD" "CD" CD" "CD"  "CD"  CD CD" CD  —.  CD  CD —- —-  CD  CD
            CDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDrN.CDCD
            —  .—ซco—ซ—<  —*—ซ.-ซ—ซ—ซ—ซ—*—*  ON  r^. r^v. 10  r*.  ro











            ^H  —-( •—t CD  CD  CD —H —^ C3  CD  CD CD ป—ซ CD  —-•  CD •—* •—i  —<  —ซ





            ror^r^r^p^csCDror^rop^r^CDrocDCDroroCDro
            co  LO f*. ro  ro  ro •ซ*• uo •ซ*  LO  r^. -^- uo co  CNJ  CM ป"-. CM  CM  ro

            CD  CO* C"" uO  CM  ซ-H CO ro OJ  CM  -^ r^. ON  CO  ro  CM CM r-.  CM  CD
            ro^M.^4     ^^     —t CM     —ซ  —*     CM     CM  —ซ CNJ CM  CM  CM
CO
                                  NO ro CM  ON  -H     co
                                  ^-^     ~-*  uo         ro
                                                             CM  CM — t
                                                                   5  ro  t—^ co uo  LO
                                                          	      J-  CM  ON U"S CC>  ro
                                                          NO ro co  CM  CM co NO  ro
                                                          ro             fo CD NO  CD
                                                                                co  r-H ^-ซ     ~^-

                                                                                CD  ro 2^     CD
                                                                                    —t co ro  CD  co
                                                                                                          NO
                                  CNJ NO CM  co  co —i ro  ro  —i  uo ON uo  -ซ9-  vo NO ro  •—*  CM  co CD
                                  ^ CO NO  CO  ON CM ON  NO  Is".  O"* ON —4  •*!•  CD CD ON  -^-  ~^  ซ-H CM
                                   I   I    I    I    I   1   I    1    I   I   I    I    I    I   I   I    I    I   I   I
                                  CD CD CD  •ซ•  MD UO so  f*^.  UO  •ซ• NO UO  CD  f— ON CO  ^  NO  NO CD
                                  *-~> .—( co  i***  CM r*^ CD  ON  rx.  r^. CNJ [•ป•.  c?  cs CD r^*.  NO  ^^  o*ป ro
                                                                                                                  w      o
                                                  uj     g iฑ!  a^


                                                  OC  UJ CD S^  |p
                                                                            I—I  C=>  UJ
                                                                            u_  •—•  ea
                                           —
                                          UJ CD
                                          ac ป— H
                                                                         CM  t


                                                                         CD  1
                                                                         OS  <^>
                                                                         CD
                                                                             CD *—• H-
                                           CD ac oc CA  co  3= •—< CD  zc  ce ac  i

                                           CDaCCDSE—J6^^?"^^"0^-
                                           _j CD_OC CDS—  _jci>—  e=iearo—i
                                                                                                                  -ijr-*  i>  27^

                                                                                                                  ro  c:  cr
                                                                                                                  <=ป  
-------
                                                                                                                                                                                                          71
                                             CS! ^-< — ซ  u"3  CO  CD CSJ
                                                         --
                                                                                                                            CO  "•*• CO •
                                                                                                                             --
                                                                                                                                                          CD  —1  —i CD —t  CD  CD ro     CM  CD  —> —t CD  ro  —i

                                                                                                                                                          CDCDCDCDCDCDCDCD     CD  CD  CD* CD* CD  CD  CD
                                                                                                                                                              LO  CD     LO      CD CD
                                                                       tOC---W
                                                                       — .CSJ— ซ
                                                                                                                                   O->CSl
                                                                                                                                   _ -, CSI
                                                                                                                                                                         CM     CSJ
                                                                                                                                                                                                     >  I*-.  CM ro •ซ-

                                                                                                                                                                                                       CSI  CM —i CM
   ฃ 3
   CD LU
;  at' "~
ill

                                                                                                                                                             CD CD O11
                                                                                                                                                                         f*^ CD CM
                                                                                                                                                                         -*S-i-x,LO
                                                                                                                                                                         — i CM CM

                                         _   _

                                         COCZCD
                                                                     ,
                                                                   — -i  — ~t CSJ •— ซ  CSI
                                                                                              ปo
                                                                                             h*5
                                                                                                                             i cr--  r
                                                                                                                               CD
                                                                                                                                                          CS'^OLO  —
                                                                                                                                                              -   -   -
                                                                                                                                                                                      .
                                                                                                                                                                           , CM CD  o--
                                                                                                                                                                                           ro LO  r-v.  ro co o>  so

                                                                                                                                                                                           —.            ro —  ro

                                                                                                                                                                                           CD CD' •—• -—- -—- -—-  -—-
                                                                                                                                                                                           CD CD  LO  CD CD CO  O1-
                                                                                                                                                                                             .              ,
                                                                                                                                                                                             •  -V  CD  CM  C^ CSJ -C-
                                                                                                                                                                          ' C*- CO  CO
                                                                                                                                                                                                   o->  co  ro co
                                                                                                                                                                                                    *        "
                                                                                                                                                          — i — — 4      CM
co  co co o--  —i  r^. CD co      —•  CM o~- co  ro  co

—  —•—t     CM     —         CM         CMCMCSJ
:  g ^
      a
                                  - ^o ^c  co
                                  f C"~  — ซ  IO
                                      -
                                    CD—.  —  CDCD — CD*  CD*  CD  — CDCD—I  •
                                                                                                            ro      •ซ• r^ co  CM i-o csj
                                                                                                            —H      csj -w CD  -ซr co —
                                                                                                            •w      -ซ9- co LO  LO r^. co

                                                                                                            —'      CD CD CD  CD CD CD
                                           >  CD  CD  O- CO

                              —  — CD CD*  —  —  CD* CD
                                 -
                                 —     — .CSI     —.CM
                                                                   CD CSJ r-o  CD  r-^  CM     CCt  ro
                                                                                                                                                         CO CO  f-O  '
                                                                                                                                                         CMCM— tf
                                                                                                                                                                                       co co  r-*  CM i
                                                                                                            Si
                                                                                                                       sOCOCf-C^CSl
                                                                                                                                   a
                                                                                                                                                         CMCM—iCDLOLOCDLOCDO
                                                                                                                                                         —, —  —, CM — —  CM —i CM —4
                                                                                                                                                                                                  f*- CD  CM  CM I-O
                                                                                                                                                                                                  r^. *a  —ซ  CD LO
                                                                                                                                                                                                  LO os  o-  r-*. LO
                  CD CD •—.  CD
                                                                                                                                              CD  CD CD CD
                                                                                                                                                                    CDCDCDCD-OCDCDCDCDCDCDCDCD
                                                                              CDO--CD     Cf.  u*3 CO CSj
                                                                            .^--sC'Cr-     CS)  C*- -*J- xฃ.
                                                                              I^^C^-CC     LOCS--3-— H
                                                                              ro  cf. LO  •   csjc">-ซ-rT>
                                                                                  CS)         hOCSJCSI
                 ^LO-^r^-w-'CD'vO  —  r^.CMtOsOOs'^.COrO     CC-O-O-CDCO-^-r^,

                 —  —  -   -                                 CM —< —H  —^     CD ro  ro  —i  CM CD fo
                                                                                                                   co  —  ro-WNorocD^s-CMrocNj     rocoCD-*-CDo-co  —iNOONr^cMror-v,


                                                                                                                   ฃS^i?S;'~1S^vC'!i20^*:^^C'     CoCo^NOC^Scor-tJ^CMCDLONO —
                                                                                                                   CD  ON  ON  CO —-. CD  ON ro r*.  f*-.  CD     CD CM  ซ—i  ปO CM  LO  O-  fO     ro  CD CO U"3 CM
                                                                                                                   CM—      —. CM CM  —• CM i—i      —i     —'"ocsirocMCMCsjro     ,    rocsjcsiro
                                            =  = s
                             ro —ซ CD  CO  CD ^9- LO ซ—'  Csj  CM CM M5  NO  CO Nฃ> -^-  -^-  -^  s^ to r-^

                             S ^5 S  iS  1^ trt S ซ  "*     "^ "*      ^* L/^ crs  *~c'  ****  f*>* ^^ ^
                                                                                                                   co-NOCDr-.csiLO(— sor-*co
                                                                                                                    I    I   1   I    I    I    I   I    I    I
                                            NO  ปO
                                            CO  CD
                                                   CO CM  f"5

                                                   S s  -
                                                                      -sO
                                                                      CO  I
                                                                                                                         ..
                                                                             CD — —  ON  LO lO CD  fO  O- CSJ Os CJ"'  "-^  CO >O -

                                                                             - "^   •   S     R:  —     ev'~     -
                                                                                                                                              WO —, CM  i
                        >  -*T LO CM  NO  CD r^. CM  CfN  LO

                        i  -V- •ซ- CซJ  Cf-  C^ LO CO  CO  ON


                          CO •ซ• ro  —-t  CM CM LO  UO  LO
                        .  QS co NO  LO  so trt —*  LO  -v-
                          CM —1 tO             	1      CM
                         ^0. ar :
                         -r LU
                   :  CM CM -^  -w
                 —-i  CM ro -^  to  so r^. co  o~>  CD —ซ <
                                                                                                                                             B
                                                                                                                                         ฃ5^3
                                                                                                                                         3E -^  *—ป  3T
                                                                                                                                             S^  •—t  uj a:  a=  ac
                                                                                                                                             SC  r-sj  31 LU  LU  Uj
                                                                                                                                         •^ so  uj  :c''Z:  Z:  ac
                                                                                                                                         CD —I  03  Q- LU  UJ  UJ

                                                                                                                                         ฃฃฃฃง!!

                                                                                                                                         Ssiiiis
                                                                                                                                                                                              -
                                                                                                                                                                                              CO  3C= OS U- CD  CD
                                                                                                                     .  co  os CD — <  csj
                                                                                                                                         ro -w  LO  so  r^
                                                                                                                                                            CO
                                                                                                                                                                               LO  d  LU


-------
                                                                                                                                                                                                                          72
                                               *ป    SO  l^".  tO
                                                                                                  ซ—i  CM     CC CM ซ-H  OS  CD —' CD  CD  CD  CD
                                                                                                                                                                      cc=     CM  '
                                                                                                                                                                                                                      CO 3
                                    r*.     CD e*.
             ฃS  K
                      t      irt     ez>
                            lO  t/?  *O     iO CD CD  CD     CD  tO








                            NO  t/?  rC     to to c2  CM CM ป••;  -i

                 ฃ  ::
                                                                                                                       CD      CD CD LO  CD
                                                                                                                                          	  _ .  CD CD      >O
                                                                                                                                          CM  so —4 -W      NO

                                                          <  IO  LO  vO  CD fcO LO  CD









                                                          i  LO  VO  tO  CD CD tO  CD
                                                                                                                   CD tO      LO tO lO  I
                                                                                                                   —H <—i      SO NO CM  i
CM  —'     CO ^-  ฃ?••  SO tO SO  '^
so  -w     NO NO  NO  CD r*- *—i  •ซ•
                                                                                                                                                                            KtO
                                                                                                                                                                            —i
                                                                                                                                                                        —*  CM


                                                                                                                                                                        g"""~  "CD"
                                                                                                                                                                            •%o
                                                                CD tO      CD LO     LO






                                                                1=0 ~ฃ>      CM NO     —



                                                                CD LO      LO CD     CD




                                                                f-ป vO      I*O—*     Cf-
                                                                |sO ^T      O-- CM     CD
                      i      ซ,     KM     CM fO —t         —  •


                      I      ฃ3 —*.  CD     —- —-• O ^^ —— •—•  <
                                                                                                      LO      lO
                                                                                                                   —i  C>-.      CM
                                                                                                                   vO  so      IO
                                                                                                                                                           LO      tO  CD     tO  tO
                                                                                                                                                                                               CM  ro     iO
CD to  r^


1o "CD"  to
ro -w  ir>
                                                                                                                   ^o CM      c?* CM to CM  ซ—'
                                                                                                                   r-o LO      i— to co 10  CD
                                                                                                                                          CD  CD  CD CD tO CD  CD



                                                                                                                                                     CM CO -v  —'
                                                                                                                                                                                       f-^     Crv  sO
                                                                                                                                                                                       s     ss
                                                                                                                                                                            CM      C^-
                                                                                                                                                                                        —i     CM  CO
                            ฃg

                            2 —  —     C^ — —     ^ฐ '
           iO tO cy-      CD -~ Cfซ CM  CM —* CTป

er't     CD CD* CD*      ซ -^ CD CD*  CD —^ CD
                                                                         gg.ggg
                                                                             Sfs, i—  rป,  p-^  LO CM r^.
                                                                             CD CD  CD  CM  so r^ CM
                                                                                                    >•••-
                                                                                                                                    . CD CD  ^O  CD
                                                                                                                                                             CD  SO     CD
                                                                                                                   CO O^
                                                                                                                                                                                                   CO •    CM
                                                                                                                    C3 CD     CD  CD —ซ
                                                                                                                                                             —t—H     —.CM     —<  CM
                                                                                                                                                                        o-, r*, cc  co

                                                                                                                                                                        CSCDCDCD
                                                                                                                                                                                                   8
                                                                                                                                                                                                                      fO  CS CM
                                                                                                                                                                                                                      r^  — so
                                                                                                                                                                                               CMCO-W
O)
o
             sa
                           R S ง 3  S 5 R:     S  S E? S=  ฐ  S  3 งIS

                           fOrocM        rocM     CM  — — —  —*CMfoCM—*

                           O\ I^ซ UO O*  CO CT1- *O     CM  •ซ• sC> C^  ^  —'  '*O LO O~-
                           ro _ co to  1/5 co CM     CM  ^ r-^ +o  o*-  rซo  CM ^ co
                                                                                                                    CD r^o     CM  i/^ co i
                                                                                                                    CM-ซJ-     i—  NO cป-- *o •ซ-•*•  CM ec* (ป*>  CP-

                                                                                                                    —H (-^     LOiOCO^OrocDuO-WCO'^.
                                                                                                                                                                                                           *
                                                                                                                                                                                               CM MI  ho  CM
                                                     C=COCM>OCD     -ซ3-—ซLO—^
                                                     p-^ esj  10  ^-  r--.     ^g-  —*  ^  so
                                                     ^ _-*__ซ_4     -HCMCM—<


                                                                                                        —* C^ r>,  ^-     co NO
                                                                                                       .  O1* •so r^.  LO
                                                                                         —- —t CM  —<  ro ^- ^  '
                                                                                         o_ a_ rs  ca_  z^ :=> =3
                                                                                                                               CO  CO  CM I
                                                                                                                               CD  to  *O r
                                                                                                                               i—t  —4  CM
                                                                                                                               fO  LO  CO —<
                                                                                                                               M3  ซ*^  CO CO
                                                                                                                               CM  CM  CM 	1
                                                                                                                                                                          *,
                                                                                                                                                                     —ซ  C"d CM CM  —<  CM     CM
                                                                                                                                                                                                         I  CM     CM CM
                                                                                        i^^y^^H^^^K^9^M?^^^^^^^^p^^^^^^cI^JvOfT)Crj?
                                                                                                                                                                 LocOl^C^sD^CDioCD^^-iCM^-lpwO-VuO
                                                                                                           . —* r*o  —H  co ^ to 0-*  r~*  LO i/^ —•+  ~-~*  c^
                 s   :  s
                                                                       .s.OLOLOrOCDCD'^-'WCMCDONCMsOC^sC-WvOlOCDeO
                                                                       i CO —-•* OO  CM  O NO ^ lO  CD  CD 0s! C'"'  NO  C*~- tO CM LO  C^-  —
                                                                        I— NO h-7  Os  f^O —< —t     SO  —^         CO         —      ""•
                                                                        r^.     CM  so                                —•*
                                                                                                                                                          LO  •—< '^^ —t
                                                                                      I  —* LO CD  C3  CO
                                                                         	 _--..jr^cocMtocDi
                                                                         —or^ซ     —4  ov  -^
                                                                                CM      IปO         NO
                      ::  S S  i S ง

                                                                                                     *—      n: ze
                                                                                                     CD      Q_l—

                                                                                                                                                          s
                                                                                                                                                 ! LU UJ  5

                                                                                                                                                  ^ S  ^  LLJ ^
                                                                                                                                                                z

                                                                                                                                                                fei         LU

                                                                                                                                                              I  U-l         ซ—'

                                                                                                                                                              :S         5
                                                                                                                                                              OC CD —I
                                                                                                                        C_3  5C  O
                                                                                                                                                          g  % :
                                                                                                                                                          co  eo
                                                               to  >—  ac >— ,
                                                                                                                                                                                          S  -ซ  OC *t OS  C>
                                                                                                                                                                                          eo  re  LLJ ae CD  o=
                                                                                                                                                                                          CD  t—  o_ *o eo  u.
                                                                                                                                                                                          a:  nr  ป—• ee LU  -i
                                                                                                                                                                                          o_  a-  a. o- o=  to

-------
                                                                                                                                                                                                                                   73
                           CD CD CD CD  CD  CD
                                                                CO  sO  CSJ  PO  CO


                                                                —,  __j  CSJ  —*  Cs!


                                                                CD  CD  CD  CD  CD
                                         CSJ UO CF- SO CD '^ SO


                                         CD CD CD CD CD CD CD


                                         CD CD CD O CD CD CD





                                             CD CD         tO
                                       25  3
                                                                                                -^     u-> PO
                                                                            SS  •   ~
                                              LO  UT>      W>  IO  WD
                                              |>O  ro      -w  PO  cs;



                                              •*9-  Cf'      ^  >O  PO
                                              CS/  0--      •*!-  CO  CO
                                        O-  CD  —
                            CD  u->  C?  CD  uo"'
                            PO  CS)  —i  CC  r-.
                                                                CD CD ul CD
                                                                               . u7>      —i C*-  —<
                                                                                                     UT)  CD  CD  U1
                                                                                             CD  uT.  CD  CD  u-j  CD
                                                                                             CD  PO  CD  CD  ^C  CD
                                —  CXJ  CNi
                                                                                         r--J  CSI  O-   •ซ•  —<  i-O  CD
                                                                                         •*O-  CSI  PO  -^-  P"^  PO  I"*.
                                                                                 — i      CT-
                                                                                  "
                            — CD CD C2 O
                                                                CD  CD  CD  CD  C3
 CD
-o

                                                                      ^    .
                                                                 CO  CSJ  CO
                                                                                         ^-5  CD  CSJ  CO  PO  ^O  —ซ

                                                                                         O^  UT>  Cfi  SO  ^C*  •?••  S&
                                                                                         CD  CO  sd  CD  CD  •ซ•  ปO

                             ~-COCOLOCOsOsOCi--Lrt-



                             CD  C1"-  -O  CO  CD  CSi  r~~.      CSl  •
                             r^j  —i  csi  —'  csi  PO  csi      csr
                      •^_  CO CSJ CT' IO CP  -y- -^ i-^> CD -*O- CO




                     ["""t^PO     PO     CSJ     CSI CSj CSl —1
                                                 CO      CO  CD  PO  CO
                                                         fcO CSJ O->
                                                 ป-
                                                                 ฃ     'ฃ
                                                                 •<:     o_
                                                                                 cr>  CD  •— ซ
                                                                                                  _J  CD  UJ
                                                                                                  ZD  a:  —i
                           •  3^ —' LJ.
                           ,  S -r =
                             >— l_k. e*
                             oc —i o:
                                                                              •-  PO



                                                                          l-O  _J  L4-J *
-cjL-cj-ttzSVS^csi^ar^^ujuj-^
-   --c-ซcaeijj_iirt_jLij-—•  uj  •ซ  =c  e=ป  ^  ag  S"
     :o=CD  ,=.=><=    -u-ZfOtljceP--.!^^
                                                                                                                                  *-*  (U  O
                                                                                                                                          o>  a*
                                                                                                                                          in  >-
                                                                                                                                      *-  W  CD
                                                                                                                    -
                                                                                                    I  CSJ  CSJ  CSI  CSI

-------
CO
 a>
•a
II
II
II
II
II
II
II
II
II
II
II
CO II
HH II
CO II
S* II
Ha II
< II
Z H
•< II
II
CO II
E ii
I it
o n
O li
n
0 II
H II
II
Cd II
i-a II
05 II
< II
Z II
cd ii
E H
•< II
II
EH H
0 II
Z II
II
>< II
Ha II
E-< II
Z H
< M
0* II
< II
OH M
OH H
< H
II
CO II
O II
Z II
ฃ3 II
0 II
OH II
E ii
O II
O H
ll
Cd H
Ha II
M II
EH II
< II
Ha II
0 II
> II
M II
E n
Cd n
CO II
n
n
n
ll
ll
n
n
n
ii
n
n
n
n
n
n
n
n
n
n
n
u
^~> II
• CO II
CQ <— n
M ca ii
1-3 Cd II
CQ II
CO ฃ II
ฃ O II
SE ll
II
II
II
II
II
II
II
II
II
II
ca n
< CO 11
OS CQ II
O ฃ ll
OS ป II
Z II
n
u
n
u
n
u
u
u
u
u
• ii
O II
Z II
u
CO II
< II
U M
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
Cd II
U II
Z II
< II
EH ||
CO II
CQ II
C3 II
CO II
II
II
II
II
II
II
II
II
II
II
II
II
• II
0 II
Z II
                                in
                                co
                                       CM
                                               CO   '  PJ CM  >— 1
                                                                           COCM
                                                    COCOO09COCOO)COOrHOซ<

                                                    ^^ ฃ^ !••{ ^^f ^^ ^ft  p^ ฃ^  QQ Q^  C^J QO
                                                    i-H               CO  >-l CM  "-H CM  CO CM
                                    r-HCOC^'-HOCOi-HO
                                    OCOCMOCO'-ICO'—I
                                   CO     inCOCOCOt>t>C^Cnt-OOซS'CO'-IO
                                   O     CM'-Hi-HOr-H'-HrHOCMCMi-HOr-lCOO
                                  I   I    I   I    I   I   I   I   I   I    I    I   I    !   I    I   I    I    I   I    I   I    I   I    I   I    I   I


                                  I   I    I   I    I   I   I   I   I   I    I    I   I    I  I    I   I    I    I   I    I   I    I   I    I   I    I   I
                                 ^O f™H  ^*H ^f* C^ lA CO CO C^ CO ^tf*  C^ C^  ^""^ ^^  p""t lO C73 'd* 1^5 CO CO  ^^ C75  CO C^  00 ^H
                                 ^M CO  ^3 CD IO Oi CO CO C^ O5 tf5  1^5 ^^  C3 ^^  O5 tf5 U) 00 f^^ C5 00  00 00  Oj lซO  ^^ 00
                                 i-HrHOOCO        n >-t         t^rH            I-COCO"-!        CD"-H
                                           (ป.                          CO                                        CO  O5
                                                                               Cd
                                                                               z
                                                                               1-4
                                                                                  Cd
                                                                                  z
                                                   Cd
                                                   z
                                                   o
                                                   Cd
                                                   z
                                                   Cd
              z
              Cd
              35
              OH
               I
              CM
                         Cd
                         Z
                                                                    Cd
                                                                    ca
       <
       OS
       o
       >•
                                                      Cd
                                                      CH
<  < Q

0  K O

35  O
Ha  M O
>•  X M
Z  EH Z
Cd     HH
   < Q
    I   ^^
   H3 Z
   > <
   B a

   Cd O
   ฃ co
   HH O
   O K
    I   EH
   Z I-H
     >z
   Z  !

   CO  I
   HH O
   CQ OS  Cd
   Cd EH  Z
                                                                                          Cd
       M
       Z
       <
       B
   < EH
< Cd Cd
Cd K OH
03 S d
                                                                                             >•  B B  <
                                                                                             B  EH EH  CO
                                                                                             H  Cd Cd  OS
Cd
EH
•<
B
O.
CO
o

OH
O
OH
>-
OH
O
OS Cd Z  E HH   |   |    i  O
                                                                            O H-1   I  <
                                                                            i-a >•  z
                                                     0,  Z Z  Z M
                                                     Cd   I   I   I  B
                                              Z
                                              O
                                              Cx
Cd
EH

B
OH
CO
o
B
OH
                                 CQ
                                 On
                                 HH  EH i-a  O O B
                                                                              —   Cd"  Z
                                                                                        CO  HH
                                                                                        Cd  OS
                                                                                        OS  <
                                                                                        O  CK
1-^  r—i ^^  toi^ ^^  ^* -—<  — ^ —^  ^f fc •  fc •
Ezcdos  i   >-HH   -EHCOWW
                                                                                    1   |
                                                                                              I    I   I   I
                                              O  < Cd  B OS  OB
                                                                                                                          CO t>  00
                                                                                                                          CM CM  CM
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
M
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
                                                                                                                                                          74
                                                                                                                                           U
                                                                                                                                           CO
                                                                                                                                           tn
                                                                                                                                           m
                                                                                                                                           oo
                                                                                                                                           S

                                                                                                                                           CO
                                                                                                                                           CD
                                                                                                      03
                                                                                                      CO
                                                                                                       c
                                                                                                       o
                                                                                                      •H
                                                                                                       n
                                                                                                       u
                                                                                                      CO
                                                                                                      03
                                                                                                      CJ
                                                                                                       0)
                                                                                                       n
                                                                                                       CO
                                                                                                      CQ

                                                                                                       CO
                                                                                                      •P
                                                                                                       CO
                                                                                                      ca
 CO
 k,

 o
 o

CO

 CO
 n
 co
ฃ

co
o
OS
CO
z
 I
B
HH
Z
                                                                                                                                           OH
                                                                                                                                           Cd

-------
                                                                                  75
                                   Slide 19
                      SEMIVOLATILE  COMPOUNDS FOR WHICH FURTHER GC-MS DATA
                      REDUCTION  AND EVALUATION ARE REQUIRED
NO.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
SUBSTANCE
ACRYLAMIDE
3-AMINO-9-ETHYLCARBAZOLE
l-AMINO-2-METHYLANTHRAQUINONE
5-(AMINOMETHYL)-3-ISOXAZOLOL
ANILINE
AURAMINE
AZINOPHOS-ETHYL
4-CHLORO-l , 3-PHENYLENEDIAMINE
l-(2-CHLOROPHENYL)THIOUREA
p-CRESIDINE
CUPFERRON
CYCLOPHOSPHAMIDE
DEMETON
2,4-DIAMINOANISOLE SULFATE
1,2:5,6-DIBENZACRIDINE
DiBENZo(A, DPYRENE
DIETHYLSTILBESTROL
DIHYDROSAFROLE
DIISOPROPYL FLUOROPHOSPHATE
EPINEPHRINE
2-FLUOROACETAMIDE
HEXAETHYL TETRAPHOSPHATE
LASIOCARPINE
MALEIC HYDRAZIDE
METHYL METHANESULFONATE
METHYLTHIOURACIL
1 , 5-NAPHTHALENEDIAMINE
l-NAPHTHYL-2-THIOUREA
NICLOS AMIDE
NITROGEN MUSTARD
N-NITROSODIETHANOLAMINE
p-NITROSODIPHENYLAMINE
N-NITROSOMORPHOLINE
OXYDEMETON-METHYL
PHENAZOPYRIDINE HYDROCHLORIDE
1,2-PHENYLENEDIAMINE
1 , 3-PHENYLENEDIAMINE
1,3-PROPANE SULTONE
4,4'-THIODIANILINE
0,0,0-TRIETHYL PHOSPHOROTHIOATE
2,4,5-TRIMETHYLANILINE
1,3, 5-TRINITROBENZENE
CAS NO.
79-06-1
132-32-1
82-28-0
2763-96-4
62-53-3
492-80-8
2642-71-9
5131-60-2
5344-82-1
120-71-8
135-20-6
50-18-0
8065-48-3
39156-41-7
226-36-8
189-55-9
56-53-1
56312-13-1
55-91-4
51-43-4
640-19-7
757-58-4
303-34-4
123-33-1

56-04-2
2243-62-1
86-88-4
50-65-7
51-75-2
1116-54-7
156-10-5
59-89-2
301-12-2
136-40-3
95-54-5
108-45-2
. 1120-71-4
139-65-1
126-68-1
137-17-7
99-35-4
RCRA
NUMBER

U253
U265
P007

U014
P150
U305
P026
U262
U290
U058
P155
U307

U064
U086
U090
P043
P042
P057
P062
U143


U164
U298
P072
U321
P132
U173
U287

P157
U320


U193
U258

U259
U234
MS LIB.
NUMBER(a)
204



840

31487


5213









13317


34981
2041



16231



15632



1744

3126




a) EPA/NIH-NSRDS Mass Spectral Data Base,1983  version;  39,763 mass spectra

-------

                                                                                         CO
                                                                                         QJ

                                                                                         "o.
                                                                                               00
                                                                                               ai
                                                                                                                                 76
                       ai
                      •a
                               QJ
00 li-
CU I ro
4J  C
                               ro,	
                               c 1-
                               r—

          01 3
       r— E CO


       -C X r—


       ^ฐ *ซ
                                             ro
                                             S—
                                             O
                                             •—
                                             JT
                                             U
     I
 ro o
 j= s.
    o
 QJ •—
 o x:
 s- o
 O T-
r— "O
x: i
 o n
 i    -
CM .—<
    4J 4J •.-
    CU QJ !_
f—'  O O TTJ


5ฐฐ!*!
 QJ x: x: i. QJ
 E  o o o o
 O  I   I r-  ,0
 S- CM CM JET  o
 O  i   i  o  E
I—  CO 00 •!—  O
x: .1- .,- CL t.
O X) X) QJ _O
                                                                   QJ
                                        Q.
                                        O
                                        1-
                                        a.
                                        O
                                        S-
                                        o
                                                           E  S 4J
                                                           I  =J C
                                                          LO  4-> ro
                                                              O- 3
                                                              
                                                          o  >  s-
                                                          O     O
                                                          QJ _J
                                                          1-  E .—
                                                            •^-.  ro
                                                         6-S  Co  :3
                                                         O  ra  cr
                                                          •     0}
                                                         o o
                                                            O <4-
                                                                o
                                                                                       i — TD
                                                                                        ro  C
                                                          C CD
                                                          QJ tฑ
                                                          (/)
                                                          00 4->
                                                         LU re
                                                           CO
                                                           o
                                                           a.
                                                           e
                                                           o
                                                           o
                                                           QJ
                                                          73 ••
                                                              i.
                                                           O O
 OJ T3
 3  QJ
•O  >
    i.
 CO  QJ
4->  00
                                                                                                                QJ
                                                                                 O  O
                                                                                 O  t—
                                                                                O-  D_
                                                                                            QJ
                                                                                           XT
                                                                                           4-> Ql
                                                                                            QJ C
                                                                                              ro
          C OJ
          ••- o
          > S_
             o
jr  o
4J  ro
QJ 4->
o  c
S-  OJ
O  CL
       o

 CU
•a
       O
       co
       g
       o:
       o
       CO
       LU






00
QJ
4-* *• ' *
>>TTJ
i — QJ
ro z>
c c
•a: T-
4->
IO C
"3- 0
CO 0

CO












00
cu
4-^
>}
^—
ro
c.
^c

*>o
^"
CO

3
CO



















CU QJ
TTJ C
•r— CU
QJ S- 3
QJ i — O
M -e~ 4->
COO
QJ S_
XI i— O
0 >,.-
E N x:
o c o
1- QJ 1
X) X! CM

S_
QJ
X:
4-J
CU

^~
>1
^*
4->
QJ QJ
•a E QJ
•r— "O
S- 1 	 •!—
O ^> E
•— -c o
-C 4-> S-
O QJ XJ
E
i — O i —
>,$->,
-E: 0 -c
4—) r~~ 4—^
QJ -C QJ
E 0 E
1-
QJ
XI
4J
Q)

*• — *
r—
>)
CU QJ
O C
S- CU
O- N
O C
CO QJ
•r- X>
O O
S- S-
o o
J- f
o o
1 *l~~
CM -a
* — t
co co
•r- •*
X) ~->



QJ
C
03
XT
4->
QJ
E
0
S-
QJ 0
"O 3
•r— r—
t- *4—
0 ••-
r— -a
xz o
0 S-
o
^— ^*
>^ >ฃZ
C O
•1 — 'f—
> TD


QJ
C
ro

4->
Q)
ฃ
QJ^-— QJ
C >, C QJ
QJ X QJ T5
NJ O M >i
C -C C XI
QJ 4-> QJ QJ
XI QJ XI TT*
O O O i —
S_ I- S- ro
O O O 4->
x: x: x: o
o o o ro
•i- 1 -r- O
-a CM TTJ s-
1 	 I O
CM CO ^^ r~~
f< tf— ft e~
t— 1 X3 t— 1 O




CU
C
ro QJ
-ET C
QJ 4-> QJ
TTJ QJ i —
•r- E >,
s- o x:
QJ O S- 4->
"O ^~ O QJ
•r- X: =3 O
!- O r— S-
O M- 0
i— QJ O .—
XT C i- XI
O QJ O O
i — r— pt—
r— S-,X: T3
>>x: o i
x: 4-> •!- t-i
4-> QJ S- •>
QJ E 4-> ^-<


























QJ

QJ
r—
^^
f~
4-*
QJ
QJ O
ฃ1 1-
ro o
XT ^~
4J J=.
QJ O
O ••-
s- -o
0 I E
^- CM S-
x: •< o
O t— 1 ซ-
•r- 1 O
T3 00 S-
1 C O
r—troi —
• s- x:
•—I 4-> O































QJ
c
ro
QJ x:
C 4->
ro QJ
x: o
4-> QJ S-
QJ C o
o ro r—
s- x: x:
O 4-> O
•— QJ -r-
x: E i-
O O 4->
••- E 1
T3 O r~J
1 S- ซ
CM XI ซ— I
ft .r— ป>
1 — 1 "O ซ — 1





OO
c
o
*lป
4_)
•r-
T3
TQ

X
•^>
7rj
c
QJ
Q.
Q.
<:






QJ
•o
• r-
J_
O

j^;
O
ro
i-
4_>
QJ
4_>

C
O

c
ro
o

QJ

ro QJ
CL-O
o •<—

QJ QJ O-r-
i — C i — O Z3
- •""• QJ O S— OO
S*- QJ 4-> C O
4-> "O 3 ro QJ i — i —
•r- -i- x> a. c x: >,
c i — c E'orooxr
O QJO •!— OC-JS_xri4->
••-QJ-OC S- ^ICL4JCOQJ
CLTTJ •ซ— ro "OQJX)OOQJ 1 O
O ••— S 	 CjC'i— i-t-OOS-
s-"ao4->cx:o-ooos- EO
o •<— x: o i- i-ai ojxrxit— s_x:
!- Ol-O-OOCCJUJCXlO
o i — o o i — ro QJ •(—•ซ— o ซr— i
t— >>•— r— s-x: or— -OTD ro-ocM
x: x: >,x: o o E >i i i *-> i i
O4->i — Oi — •!— OXT^I-CO CCM co
1 QJ I — IX:CLS_4-> ซ ซปQJ *ซT—
COEroCMOQJXlQJ'— Ir-lCLt— 1X1

S- QJ QJ
QJ C C
x: ro ro

QJ QJ QJ 4-> 4->
QJ C QJ CT QJ QJ
CQJQJ cro i — o QJO
rocr- rox: >>s_ cs_
xrro>>Qj-C4-> co QJO
4JO-CLC4-JQJ -r-i — i — i —
QJOOQJQJOQJ>X: 5-.X:
ES-i-i — ES-4-> o x:o QJ
oo.Q->,ooror— ro 4->roojc:
s-ooxrs-r— s_>-,s- ajt-cro
os-s-4->ox:-ax:4-' O4->ojx
i — O O CU i — O >>4J QJ S- QJ IM QJ
x: r— r- o J= ••- J= QJ+J e o 4-> ex:
OXTXTS-Oi- Olt-r— IQJO
•<— OOOO4->r— i-CMOXICMXl S-
•a ••-••- f— E i roo -H-O -oo
o-o-ox: oog i_^.^-< o rocM s_r-
Ei los- -ox: -ES- -ox:
O CM CO -r- XJ t — 1 i — O ซ — • O 4-> < — ' i — O

X> .— < i— 1 +-! TD •— 1 O CM i— I X> 4-> r— I o i— I

-------
                                                                                    77


UJ
_J
CO
z:
UJ
jr
ซo:
i 	
0
•z.

co to
CXJ LO
| —
CU >-
-0 _l
•r- ec.
T— 21
CO -.
	 1
eฃ

•a:
Q
| —
CL

O



















*
>>
S c
o S
ll
Q
1 —
Q.













O)
">!,
r—
ro
C



Ln
Q O O Q O LT>
CD O .— < LO CO CTi CT>
CO








CU
4_i

JD *O ,—
>•} .C ซr—
X 0 S_
O CU U Ol CU 4->
CLi — i — C. i — Ol -i-
O) Ol •'— roOCU -i- i — C
T3-I—S-CU +-> c: i — s- -I— o
^"•j "O +J "TZ? r~~ O ra O 4-3 t- ^—
.c i T- -r- >, (o x c: •!— 4-> >>
0) ซ* C E m,— o re C -r- s-
"n> " O ra i- O *r- 4-S o 5T O
' — CO Ci — fa-i--O 3-i-O ro
fa • • o >5 O- Q. i jo Q. +J jr
i-CSJr— i-OO'St-OOO)4->
rO "roOt-S- '•coi-OCU
Q-— i E ro CLO..— i-r- CXra E

CU
•• s
—1 ra r-"
e 4-- o
*-. CO C
CD c ra
CO CO CU Ol
0-r- _0 E
•^r co.
>•> o ^:
• •* r— 4-^ 4->
C ra .^
o c: -o 5
•i— ro CU
•>-> s_ -o
O Q ra O)
CU 1 — CU 4->
••-3 Q- CL 3
C D. , —
•"— S- ro CU
_ o o
E ซ*- c o
3 ra
•ซ-> CU •>-> CU
ex,— a. -o
CU CL ra -r-
to E U • x
ro t- Q O
CO CO CU 1 —
3 E O- O)
CO _J f-
S- E ,- T3 0)
CU 1 >i S_ , —
^* LO jr ra >%
•(-> 2 J=
Ol O •(->
* s: 4-> LU
CU
IS)








2^
o
t— t
h-

un
o
CO
UJ
s:
0
U-

CO
UJ
1—
>-
— 1
c^
^"
^c












1— 1
1 — 1
•— ' CO
>• c
o
X -r-
•i — 4_>
TD -i-
c -a
cu -o
O ซ^"
Q.
ซ^








CO
Ol
4_3
>^
^—
ra
f~
^c

10
^3-
CO

3:
CO





CU
• — 4-> ro
O tO r—
f~ J— >^
0 >, S-
r— O 0) J- 0
O i — C CJ ra
-C ra O rO -C
o i— -!-> x: 4->
O t— O O •!-> CU
i — >-> C ra CU E
ra CD ro r— E
1- 4-> O i —
r— ro 3 ••— r— >•)
>•} r") r"> Q_ >^ j^
•— O O 0 -C 4->
i — S- CO S- 4-> CU
fO CL-r- CL CU E


Ol

o
O) •!->
c: 01
o ^
Ol •(->
"O CU r—
•r- ^ >,
M- S- *->
i — Ol i — 3
3 x: >,_o oi
Ol co 4-> x: o "O
-O -i— Ol -M CO >-)
•r- -o cu -i- -c:
Er— Ol
ra C >->i — i — T3
<— 0 -C >, >,t—
4_ S- 0) +J 4-> S-
O ro -i— OJ CU ro
ra O "O E E CL
01
C 01
ra i —
3 Ol t-
J3 i — +J C
>-j ปr— ปf— rO
01 X J_ C 4->
T3 O CU CU 4-ป O CL
•I— CL CU r— i — *r- •ป— rO
cu x oil — •!-••- c: CLO
CO-r--r-S_S-OOS-
rO "O i- 4-* 4-1 r— t- O)
X Ol 1 •(-> -I- -r- >, CL E
oeซi-"-ces-:>ป
•r-OI "COOOXi—
-Oi — coo---c:rao>>
1 >> • • 4-> CL O -C S- JT
Tt-JTCvJ CU Oi — •f-J'O4-ป
•>+-> "O5_roO)>)OI
.— la'T-'racLeE.cE
























-------
78









i.
ID O
CM 3
CU CU
"^3 i-
•r— 13
i— 4-J
Ll_









Ol
^~
**"
4-^
03
O
••-
CU
to
c
o
T3
3
to
O
C
cu
•r—
O
u-
cu
c
o
4-5 CU
0 4->
03 >,
e ss
CU LlJ 03
^_ &
O3
CU
zr

4-"^
o
cr

to
cu
1>>
03 Q
03 Q.

CU CU
r- S-
••- 3
1 ^ 1 %
03 03
0 CU
> O.
p—
tz
'— CU
O 4-5
q-
>> O
-a o
3 S_
to :>,
o
1- -o
O_ CU
t
Heated
recovet

•


c
cu
E
Q.
O
cu
CU
•o
o
C"
cu
01
•r— .
to
03
03
Q
Q.
C_3
CJ3
ฃ>
Q. ฃ=
O 
_1 •!-
Q- 3
rc to

V







D?
O
o
•o
o

4-> 01
CU CU
E 0
= S-
sl E
Ol
c -o
CU r-
cn cu
s -r—
O.
3 i.
1 0
C "-
CU 4->
i — cr
o cu
t=
40 .ex
0 O
03 r—
5- CU
X CU
LU -O

^


s_
0
Q-
o



2:
o
1 — 1
1—
^ฃ
<^**
1 — 1
_J
•a:
=>

o
CM
o
'Jf CO
CM
O
cu o
•a p=
•r- 1—
i — LU
CO S

<")-^
0
Lu.

CO
LU
K-
>-
_1
• C
o
X -r-
•i^ 4->
T3 -r-
C T3
CU "O
o.ซa:
Q.
<:










(/)
QJ

*^L
^^*
frt
c
^

10

oo
-
-*
t/J





cu
r~ ^~
CU -r- O
C i— CU C
•i- O C CU
•O O CU Q-
•1- •!- i- O
S- OL >,••-
>> 1 4J J=
CLCM 01 4->









CU CU CU
C C C
cu cu cu
ISI IM Nl
C C C
cu cu cu
qi o o -O qi
C= O O O C
CU i- S- S- CU
N O O O N
C i — i — i — C
CU JH J^ -C CU CU CU CU
cuj3Oooj3cuc:cc:
C O -I-T- T- C CU CU CU
CU S-T3T3T3i — CUi — i — • —
M O I I 1 >, 3 >> >j >-,
Ci — CMCOซd-JTr— X X X
CU J^ ซ *> •ป 4-) O 1 1 1
J3Oป— 
CU >>
-T-J 1 	
.;: 03
r— c:
CO ^^

o
CM
f""^
CO

•o
o
4-9
cu
s:
r— 3

03

Q 0
1— CU
Q- s- sd-
CU CO

01 JD cu cu
•t- 01 ฃZ
01 T3 O
>, CU 1 4->
r- S- 1C
03 CU CU
C > •— CO
•a: o o
O C Sซ
Q CU CU LO
1— i. JZ |-^
Q. CL .
4J O i —
O O V- ^
4-> c jr o
4-> CD
1 CM
1 -0 r—
C D_
CU 03 OO
•i- . CU 6-S
• — c un
O -r-
o -o c
•I- •!- O
0- S_
1 >,
CM O-



-------
                                                    79
                          MR. TELLIARD:  On our



second half of our morning session, our next speaker



is Diane Kocurek.  She's going to talk again about




our favorite subject right now, which is the 304(h)



and the SW-846 methodology as a' comparison.  Diane.

-------
                                                         80
    COMPARISON OF SW-846 AND 304(H) METHODS
FOR ANALYSIS OF APPENDIX VIII ORGANIC COMPOUNDS
           Chemical Manufacturers Association
                   Washington, D.C.
                         by

                Dianna S. Kocurek, P.E.
               Lial F. Tischler, Ph.D., P.E.
                   Tischler/Kocurek
                     116 East Main
               Round Rock, Texas 78664
                     Presented at

                     Ninth Annual
               U.S. EPA Symposium on the
         Analysis of Pollutants in the Environment
                    Norfolk, Virginia
                   March 19-20, 1986

-------
                                                                                     81
                                         ABSTRACT
      A  study was  designed to compare  analytical  performance of gas chromatography/mass
spectrometry (GC/MS) 304(h) methods 624 and 625 as alternatives to SW-846 GC/MS methods 8240 and
8270 for  the analysis of organic compounds in the Appendix VIII list in 40 CFR 261.  Three prominent
laboratories participated in the study. A relatively simple groundwater matrix was spiked with various
combinations and concentrations of 24 volatile (11  priority pollutants) and 24 semivolatile compounds (8
priority pollutants) from the Appendix VIII list.  Based on an evaluation  of precision, accuracy, and false
negative  and false  positive observations, the  use of the promulgated  304(h)  GC/MS methods for the
analysis of pollutants in groundwater is as  effective as using SW-846 methods. The list of compounds
amenable to GC/MS,  however, is somewhat shorter than the Appendix VIII list as demonstrated by the
large number of false negative observations reported by all three laboratories. GC/MS methodology is
more amenable to the analysis of Appendix  VIII compounds which are priority pollutants.  In general, false
negative observations are more likely to occur at the lower concentration levels.  The data user should be
aware of the limitations of the analytical methodology and interpret analytical data with caution due to the
occurrence of both  false negative and false positive observations.  Additional studies should include a
GC/MS (304(h)) method validation for those compounds on the Appendix VIII list of pollutants which can
be analyzed adequately by GC/MS methodology.

-------
                                                                              82
                COMPARISON OF SW-846  AND  304(H) METHODS
           FOR  ANALYSIS OF APPENDIX  VIM ORGANIC  COMPOUNDS
                                  INTRODUCTION
       The U.S. Environmental Protection Agency (EPA)  published  "Test Methods for
 Evaluating Solid Wastes, Physical/Chemical Methods" (SW-846) to serve as a methods manual for
 the sampling of ground water or leachate and analyses of compounds listed in Appendix VIII of 40
 CFR 261.  In October 1984,  EPA proposed to make SW-846 mandatory for all testing  and
 monitoring activities required under Subtitle C of the Resource Conservation and Recovery Act
 (RCRA).  EPA stated that SW-846 contained  the analytical  methods for all Appendix  VIII
 compounds, excluding exotics and water reactive compounds.

       A recent study by the  Chemical Manufacturers Association (CMA),  "Inter- and
 Intralaboratory Assessment  of Selected SW-846 Methods for  Analysis of Appendix  VIII
 Compounds in Ground Water"  (April 1985),  cited several reports which revealed that SW-846 was
 inadequate to provide  proper guidance to analytical laboratories due to lack of sufficient
 information and details, technical inaccuracies, and inconsistencies, and pointed out numerous
 problems associated with analysis compounds in the Appendix VIII list. The CMA study confirmed
 the findings of the other reports.  CMA concluded that SW-846 does not contain analytical
 methods for all Appendix VIII compounds, excluding exotics and water reactive compounds.

       The present study by CMA was an extension of the earlier CMA study.  One of the
 purposes of this study was to compare analytical performance of gas chromatography/mass
spectrometry (GC/MS) alternative methods to GC/MS methods of SW-846 for the analysis of
organic compounds in  the Appendix VIII list.  This study, as well as the previous  CMA study, was
designed to evaluate those compounds which are amenable to GC/MS methods. The alternative
methods selected for comparison were Clean Water Act Section 304(h) methods 624 (volatiles)
and 625 (semivolatiles).  In addition, this study was designed to compare the 304(h) and SW-846
methods  for priority pollutants in the Appendix VIII list versus other compounds  in the list.
Comparisons were based on precision, accuracy,  and false positive and false negative
observations.

-------
                                                                                83
                                   STUDY  DESIGN

 GROUND WATER MATRIX

        The water samples used in this study were ground water spiked with various organic
 compounds. The same ground water matrix which was used in the previously referenced CMA
 study was used for the spiked samples in this study. A description of the ground water as taken
 from the previous study is provided here for convenience.

        The ground water was collected from a monitoring well in the coastal plain region of Texas.
 Physical and chemical analyses, including organics, were conducted on the ground water.  The
 physical and inorganic chemical analyses are shown  in Tablel.  No organic compounds were
 reported as being present in the ground water as determined by GC/MS analysis. Therefore, for
 the purpose of calculating percent recovery (accuracy) in this study, the spiking value in the
 sample was taken to be equal to the true value.
SAMPLE PREPARATION

       A total of 24 volatile and 24 semivolatile compounds from the Appendix VIII list of 375
were selected for spiking into the ground water matrix.  Of the 24 volatile compounds, 11  were
priority pollutants. Of the 24 semivolatile compounds, 8 were priority pollutants.

       Four spiked samples were prepared  from the  ground water matrix.  The  spiking
arrangement was such that each compound was spiked  in  three of the four samples at  three
different concentration levels. The list of spike compounds and concentrations for each  sample
are presented in Table 2 for the volatiles and in Table 3 for the  semivolatiles.

       As  noted in Table  2, paraldehyde  is included  in the list  of volatile compounds.
Paraldehyde is included in the volatile group since EPA has listed this compound as amenable to
the purge and trap methodology.  Analytical experience has indicated, however, that paraldehyde
is not purgeable.
ANALYTICAL LABORATORIES

-------
                                                                                84
       Three prominent laboratories participated in this study, and were identified in this report
as Labs A, B, and C.  Lab A utilized SW-846 methods 8240 (GC/MS for volatiles) and 8270
(GC/MS for semivolatiles) for analysis of the spiked samples. Lab A also employed additional SW-
846 methods for all Appendix VIII organic compounds.  Labs B and C utilized 304(h) methods,
specifically method 624 (GC/MS for volatiles) and method 625 (GC/MS for semivolatiles). None of
the laboratories were asked to analyze for metals.

       The laboratories using the 304(h) methods were requested to look for other compounds
that may be present, but not priority pollutants.  Both laboratories offer this service to customers
on a  routine basis for a slight additional  cost.  The cost for analysis at Labs B and  C was
approximately one-fourth to one-third the cost of analysis at Lab A.
                            DATA  ANALYSIS METHODS

QUALITATIVE DATA

       Qualitative data are considered in this study to be data reported as less than some value,
below the method detection limit (BDL or BMDL), and nondetected.  These data were reported
differently by each of the three laboratories in this study. No data were reported as greater than
some value.
       Lab A reported qualitative data in its individual sample reports as nondetect (ND) or BMDL.
The laboratory report defined ND as the absence of any detectable concentration of a compound
and BMDL as a detectable concentration below the method detection limit (MDL).  Lab A's sample
summary report, however, for all four samples was in conflict with this terminology.  In the summary
reports, both NDs and BMDLs were reported as less than the MDL (for example, < 4.7 micrograms
per liter (ng/l)).  This resulted in  an initial misinterpretation of the "less than" summary data as
positive observations of these compounds, or BMDLs.  A review of the individual sample reports
corrected this misinterpretation and as a result, most of the less than values in the summary
reports were actually found to be NDs.

       Lab  B used below detection  limit (BDL) to report compounds which were  both
nondetectable and those which were detectable, but below the  method detection limit. Lab C
reported qualitative data as NDs for compounds not detected at the limit of quantitation (LOQ),
and detected (D) less than the LOQ (for example, D,<10 |o.g/l).
                                           3

-------
                                                                                 85
 DATA REJECTION
 Outliers
        Outlier tests are normally performed on a given sample for each individual compound. For
 example, one might test if any of the laboratories' results for a sample should be considered
 unusual enough that it would be outside of the expected range due to normal variance.

        Two different outlier tests were considered for this study, Student's t-test based on the
 interlaboratory mean and standard deviation, and Dixon's ranking test.  Since the study design
 was limited in the number of samples and participating laboratories, neither test was considered
 practical.   Therefore, no data in this study were rejected as outliers.

 Other Data Handling

        Lab C reported the presence of 2-chlorophenol and 3-chlorophenol where the samples
 were spiked with 2-chlorophenol and 4-chlorophenol. The 3-chlorophenol data were not rejected
 and it was assumed that this laboratory had "detected" 4-chlorophenol, but had misidentified the
 specific isomer.

        Lab C also reported results for 1,2-dichlorobenzene and 1,3-dichlorobenzene from both
 methods 624  and 625. Although these compounds are listed as semivolatiles, the method 624
 (volatiles)  results  were  used  because  they  had higher  accuracies  (recoveries).
 Methylmethacrylate, also listed as a semivolatile, was reported by Lab C from method 624. These
data were used since they were the only data for this compound from this laboratory.
PRECISION
       This study was not designed to include replicate analyses for the purpose of comparing
precision between the 304(h) and SW-846 methods. Replicate analyses were only reported by
Labs A and C as part  of their quality assurance/quality control programs.  These data were
insufficient for method precision comparison even though both methods were represented (Lab

-------
                                                                                 86
 A used SW-846, Lab C used 304(h)) since Lab A had replicate analyses on only one sample.
 Therefore, method precision was compared by an alternate means.

        Since the 304(h) methods in this study were methods 624 and 625, the overall
 interlaboratory precision equations published as interim final for these methods (49 FR 43234,
 October 26, 1984) were used to calculate the acceptable interlaboratory range in reported
 concentrations for a given spiked compound in each sample. It was then determined whether
 each laboratory's result was within the acceptable  range.  The number of times that a laboratory
 failed to meet this criterion was totaled and the three laboratories were compared over all samples
 and applicable compounds. The applicable compounds were the priority pollutants in this study
 since the precision equations are limited to this group for methods 624 and 625.

        Although this precision comparison is based on calculations for the 304(h) methods, it
 provides a very simple  qualitative assessment of whether the 304(h)  and SW-846 methods
 perform similarly. An example of the calculations performed is given below.

        Sample 1. benzene
        Equations are taken from Table 6,49 FR 43379.

        S1      =0.25X-1.33
        where  S' is the overall precision (ug/l)
               X is the accuracy as calculated below
        X      = 0.93C  + 2.00
       where  C is the true concentration (u.g/1)(spike level assumed to equal true concentration)

       Spike concentration of benzene in Sample  1 was 35.6 ug/l.
       X     . 0.93 * 35.6 + 2.00
              = 35.108 p.g/1
       S1     =0.25*35.108-1.33
              = 7.447
       Assuming a 95 percent confidence  limit (t = 1.96), then the range in interlaboratory
concentrations is:
       R      =Xฑ1.96*S'
              = 35.108 ฑ1.96* 7.447
              = 20.5, 49.7

-------
                                                                                87
       With reported values of 22.2, 35.0, and 26.0 u.g/1 for Labs A, B, and C, respectively, each
laboratory was within the interlaboratory range.
ACCURACY

       Accuracy was calculated as a percent recovery of the spike concentration for each
compound in a sample.  Average percent recoveries for each compound and each laboratory
were calculated as simple arithmetic means. The standard deviations of the average recoveries for
grouped compounds for each laboratory were calculated by the standard method, assuming a
normal distribution on an arithmetic scale.
FALSE POSITIVE AND NEGATIVE OBSERVATIONS

       A false positive  observation is defined as a reported  positive analytical result for a
compound where the compound is not actually present in the sample.  Conversely, a false
negative observation is an analytical result that indicates that a compound is not present when in
actuality it is.

        In the evaluation of false positive and false negative observations in this study, analytical
results reported as ND were taken as zero concentrations and therefore a negative observation.
Where Lab A reported BMDL data, indicating a positive observation,  but below the method
detection limit, these  data were taken as positive observations.  Results for Lab C which were
reported as "detected", but less than some value, were handled as positive observations.  Since
Lab B reported data as BDL, and did not distinguish between nondetects and  results below the
method detection limit, it was assumed that the BDL data were  all NDs.  This assumption was
based on the  performance of the other two laboratories which reported more NDs than "detects"
when the compound was not actually present in the sample.
                           COMPARISON  OF METHOnS
PRECISION

-------
                                                                                 88
       The comparison of the 304(h) and SW-846 method interlaboratory precision was based
on the overall precision equations available for methods 624 and 625 (304(h) methods) since
replicate analyses in this study were too limited for this purpose. A discussion of these equations
is given in "Data Analysis Methods."

       The results of the precision performance comparison are  presented in Table 4.  Results
are only shown for the priority pollutants in this study since the methods 624 and 625 precision
equations have not been determined for the other  compounds in Appendix  VIM. No extreme
differences are seen in the overall performance of the three laboratories. Lab A (SW-846) was not
within acceptable concentration ranges a total of 7 times (out of 54), Lab B (304(h)) was not within
range 3 times, and Lab C (304(h)) was not within range 5 times.

       Based on this comparison, one can conclude  that the 304(h) methods perform at least as
well as SW-846 methods with respect to interlaboratory precision for analysis of priority pollutants.
ACCURACY
       The accuracy or percent recovery  of each compound spiked into the groundwater
samples was calculated for each sample and laboratory. A summary of these data as average
percent recovery for each compound and laboratory is shown in Tables 5 and 6.

       Of the 11 volatile  priority pollutants that were spiked into the samples, there is little
observed difference between results when the 304(h) or SW-846 methods are used, as shown
by the average percent recovery for all 11 compounds among the three laboratories in Table 5.
Based on the standard deviation for each laboratory, there is no statistical difference at the 5
percent significance level  between any of the laboratories.  This lack of difference  is further
illustrated in Figure 1, where the average percent recoveries among the laboratories do not show
any strong trends or relationships. Lab A (SW-846), however,  is shown to have the highest
standard deviation and the highest number of recoveries that exceed 100 percent (4 out of 11
compounds), a fact that causes its overall average percent recovery to be greater than the other
two laboratories using 304(h) methods.

       Most of the other volatile Appendix VIII compounds which are not priority pollutants were
not detected in any of the samples into which the compounds were spiked, shown by the many
zero recoveries in Table 5.  Only 3 of these 13 volatile compounds  which are not priority pollutants
                                           7

-------
                                                                                  89
were detected by at least one laboratory and the recovery results are mixed. Only Lab A (SW-846)
detected methyl ethyl ketone, and with an average recovery of 120.7 percent.  Labs B (304(h))
and  C  (304(h)) had fairly similar  recovery results (37.0 and  46.2 percent)  for 1,2,3-
trichloropropane, but experienced much poorer recovery than Lab A (SW-846) at 84.1 percent.
The recovery results for carbon disulfide were widely divergent among the laboratories (187.7,
325.6, and 31.2 percent).  Since the non-zero recovery data for the  other volatile Appendix VIII
compounds were so limited, statistical analyses and recovery plots were not made.

        The recovery data for semivolatile priority pollutants in Table 6 appear to show that Lab A
(SW-846) performed better than the  other two laboratories utilizing the 304(h) methods.  The
average percent recovery for Lab A (SW-846) was 78.8 percent as  compared to 67.1 and 70.2
percent for Labs B and C (304(h)), respectively. However, statistically, the performance of Lab A
(SW-846) is not shown to be different from the other two laboratories at the 5 percent significance
level. Again, as with the volatile priority pollutants, Lab A (SW-846) had the highest number of
recoveries greater than 100 percent (3 of 8 compounds spiked).

        The plotted data for the semivolatile priority pollutants, in Figure 2, do not show a strong
predominance of higher percent recovery for Lab A (SW-846).  The data in the figure also  show
that percent recovery for these 8 pollutants varies dramatically for each of the laboratories. Both of
these observations are in agreement with the lack of statistical difference among the laboratories.

        Only 9 of the 16 other semivolatiie Appendix VIII compounds were detected by at least
one laboratory. The comparison of percent recoveries in Table 6 for  other semivolatile Appendix
VIII compounds is similar to semivolatile priority pollutants.  Lab A (SW-846) had the highest
average recovery for these nine compounds (46.3 percent); Labs B and C (304(h)) had 33.9 and
32.6 percent recovery, respectively. Statistically, Lab A (SW-846) is not different from Labs B and
C (304(h)) at the 5 percent significance level. Figure 3 illustrates this relationship.

        In summary, Lab A, using SW-846 analytical methods,  had  the highest overall average
recoveries for the volatile and semivolatile groups for the priority pollutants and other Appendix
VIII compounds. This laboratory also  had the highest number of calculated recoveries exceeding
100  percent.  Despite the higher overall average recoveries attributable to the laboratory using
SW-846 methods, there is sufficient variation in the  individual compound recoveries  among all
laboratories that there is no significant difference in overall recoveries  among laboratories using
the 304(h) or SW-846 methods.
                                            8

-------
                                                                                  90
FALSE POSITIVE AND NEGATIVE OBSERVATIONS

       The false positive and negative observations occurring for volatile analyses are shown in
Table 7 for  individual compounds.  Among the three laboratories, only three false  positive
observations were reported, once by Lab B (304(h)) and twice by Lab A (SW-846).

       A large number of false negatives were observed, however, by all three laboratories.
False negative observations occurred for 3 of the 11 volatile priority pollutants, and for 11 of the
13 other Appendix  VIII volatile compounds.  Ten of these 13 other Appendix VIII compounds
were never detected by any laboratory. Lab A alone, using the SW-846 method, did not detect
ethylbenzene in any of the 3 samples spiked with the compound, but this same laboratory was the
only laboratory of the three which was able to detect methyl ethyl ketone in the spiked samples.

       The false positive and negative observations for the semivolatile compounds are shown
in Table 8 for individual compounds.  Lab B (304(h)) reported 2 false positive observations for
naphthalene and methoxychlor; Lab A (SW-846)  reported 1 false positive  observation for
methoxychlor.

       Fewer false negative observations were reported for the semivolatile compounds as
compared to the volatile compounds in Table 7. Lab C (304(h)) reported the fewest false negative
observations (25).  Labs A (SW-846) and B reported 30 and 34 false negative observations,
respectively.   Seven of the other Appendix VIII compounds were  never detected by any
laboratory.  Diphenylamine and  4-nitrophenol were only detected by Labs B and C.  Lab B was
never able to detect methoxychlor, but reported it  as a false positive in the one sample not spiked
with the compound.

       As mentioned in the section on data analysis methods, Lab C  (304(h)) reported the
presence of  2-chlorophenol and 3-chlorophenol  (the samples were spiked with 2-chlorophenol
and 4-chlorophenol).  It was assumed  for this study that the laboratory had "detected" 4-
chlorophenol, but had misidentified the specific isomer.  This assumption does not change the
overall conclusions of the false positive and negative observations evaluation.
       A summary of the false positive and negative observations is shown in Table 9 for the
volatile and semivolatile compounds. Separate results are also shown for priority pollutants versus
other Appendix VIII compounds. The total number of false analytical observations was nearly the
                                          9

-------
                                                                                   91
same for all three laboratories.  Lab A (SW-846) had slightly more and fewer false negative
observations for volatile priority pollutants and other Appendix VIII compounds, respectively, than
Labs B and C which used the 304(h)  methods.  Lab C had fewer false negative observations
reported for the semivolatiles (25), however, Lab B which used the same methods reported 34
false negative observations.  Lab A using the SW-846 methods reported 30 false negative
observations. Lab C was the only laboratory not reporting any false positive observations.

       As shown in Table 9, 72 is the maximum number of false negative observations which
could be reported by any one of the laboratories for either the volatile or semivolatile compound
groups in this study. Since 3 of the 4  samples were spiked with 24 volatile compounds and 24
semivolatile compounds, the maximum is 3 times 24, or 72, for each group.  To break it down
further, based on the number of priority pollutants and other Appendix VIII compounds in the
volatile and semivolatile groups which were spiked in the samples, the maximum number of
possible false negative observations are:  33 volatile priority pollutants, 24 semivolatile priority
pollutants, 39 volatile other Appendix VIII compounds, and 48 semivolatile other Appendix VIII
compounds.

       Based on the  above maximums, the range in false negative observations among the
laboratories was 9-18  percent of the total volatile priority pollutant analyses from all the spiked
samples, 0-13  percent for semivolatile  priority pollutants, 77-85 percent for other volatile
Appendix  VIII compounds,  and 52-71  percent for the other semivolatile  Appendix VIII
compounds.   For all priority pollutants,  the total false  negative observations represented 3
percent of the combined 432 analyses for all Appendix VIII compounds  (48)  in this study (48
compounds x 3 samples x 3 laboratories).  The  overall total false negative observations for other
Appendix VIJI compounds which were not priority pollutants was 42 percent.

       The number of possible false positive observations is theoretically unlimited.  That is, a
laboratory could report innumerable compounds which were not actually present in  a sample.
Therefore, a calculation similar to the percent false negative observations was not attempted for
the percent false positive observations.
       Based on the data in Tables 7 through 9, there is no significant difference in false positive
or negative observations among the laboratories and compounds. The same can be concluded
for the difference in the SW-846 and 304(h) methods. The data do indicate that many more false
negative observations occur for other Appendix VIII compounds than priority pollutants.  One
explanation is that the priority pollutants are more commonly analyzed for, operators are more

                                           10

-------
                                                                                92
familiar with these compounds' identification and quantification, and standards are more readily
available for definitive identification.  The data here are not conclusive on this point, however,
since the number of compounds used in this study was limited.

       The false negative observation data do suggest a concentration effect.  In Tables 7 and 8,
where a laboratory did not report false negative observations at all spike levels (indicated by 1 or 2
in the column), the false negative observations generally occurred at the lower spike levels. For
example, there are four instances in these tables where 1 or 2 false negative observations are
reported.  In three of these four cases, it was at the lowest spike levels that the false negative
observation occurred.
                    CONCLUSIONS  AND RECOMMENDATIONS

       Based on an evaluation of precision, accuracy, and false negative and false positive
observations, the three laboratories in this study, one using SW-846 GC/MS methods, and the
other two using 304(h) GC/MS methods (methods  624 and 625), performed equally well.
Therefore, the use of the promulgated 304(h) GC/MS methods for the analysis of pollutants in
groundwater is as effective as using SW-846 methods and the analytical costs are approximately
one-fourth to one-third the cost of analysis by SW-846 methods.

       GC/MS methodology is the current state-of-the-art  for analysis of pollutants  in
groundwater, however, the list of compounds amenable to GC/MS is somewhat shorter than the
Appendix VIII list. This is demonstated by the large number of false negative observations that
were reported by all three laboratories.   Overall, 45 percent of the analyses for the spiked
compounds were reported as false  negative observations.

       GC/MS methodology is more amenable to the analysis of Appendix VIII compounds which
are priority pollutants. This is shown by the frequency of false negative observations reported by
all three laboratories for the Appendix VIII compounds which are not priority pollutants. Forty-two
percent of the analyses for the non-priority pollutant Appendix VIII  compounds  were false
negative observations; only 3 percent false negative observations were reported for priority
pollutants.  Seventeen of the 48 compounds which were spiked into the samples were never
reported by any of the laboratories;  all seventeen compounds were other  Appendix VIII
compounds which were not priority pollutants.
                                           11

-------
                                                                                 93
       Data in the study suggest a concentration effect in the false negative observations.  In
general, false negative observations were more likely to occur at the lower spike levels.

       From a data user's point of view, analytical data must be interpreted with caution.  The user
should be aware of the limitations of the analytical methodology. For instance, an analytical result
of "not detected" may be really a false negative observation  and the compound may be actually
present in the sample.  There is also a problem with false positive observations.  Compounds not
present in the samples are actually being reported as being present. False positive observations
are likely to be  reported as very low concentrations (near the method detection limit), and
therefore would be a particularly difficult problem in compliance monitoring where limitations are
set near or equal to the method detection limit.

       Since the three laboratories  participating  in this study were aware of  its scope and
objectives, it can be assumed that each laboratory made every effort to maximize the quality of its
performance. Thus, it is reasonable to assume that the results of this comparison represent the
"best" performance achievable by  these analytical methods.

       Additional studies  should include  a GC/MS (304(h))  method validation for those
compounds on the Appendix VIII list which can be analyzed adequately by GC/MS methodology.
Reagent water as well as groundwater matrices should be tested.  GC/MS reference spectra
generated from  such a study should be considered for publication and  incorporated into  a
reference spectra library similar to the priority pollutant spectra library.  This would improve the
qualittative aspects for the  analysis of those compounds that are really amenable to GC/MS
analysis.
                                   REFERENCES

1. Stanko, G. H. and Fortini, P.E., "Inter- and Intralaboratory Assessment of Selected SW-846
Methods for Analysis of Appendix VIII Compounds in Ground Water," presented at the U.S. EPA
Symposium on the Analysis of  Pollutants in the Environment, Norfolk, Virginia, April 1985.
                                           12

-------
                                                                        94
                                   TABLE 1
                 PHYSICAL AND INORGANIC CHEMICAL PROPERTIES
                                    OF THE
                            GROUND WATER MATRIX
                          USED FOR SPIKED SAMPLES
             Property
Description
             Appearance
             pH
             Total suspended matter
             Total dissolved solids*
             Chloride, Cl
             Hardness, as CaCO3
Very turbid, sandy solids
    6.9
1060  mg/l
 580  mg/l
   70  mg/l
 318  mg/l
*Filtered through 0.45u. membrane
Analysis by GC/MS demonstrated that the ground water matrix was free of organics.

-------
                                    TABLE 2

                            SPIKE CONCENTRATIONS
                     IN SAMPLE SOLUTIONS (micrograms per liter)
                                   VOLATILES
                                                                           95
Compound
                                                  Sample Number
Priority Pollutants
acrolein
benzene
bromoform1
chlorobenzene
chloroform
1,1-dichloroethane
1 ,2-dichloroethane
1 ,2-dichloropropane
ethylbenzene
toluene
1,1,1-trichloroethane
Other Aopendix VIII Compounds
acetonitrile
bromoacetone
carbon disulfide
chloroacetaldehyde
1 ,2,3,4-diepoxybutane2
1 ,4-dioxane
hexachloropropene
iso-butyl alcohol3
malononitrile
methacrylonitrile
methyl ethyl ketone4
paraldehyde**
1 ,2,3-trichloropropane

it
35.6
*
24.9
*
62.4
*
47.8
78.4
16.4
12.8

36.4
22.7
69.6
36.6
18.0
*
35.7
72.3
59.8
16.4
31.1
30.0
*

64.4
*
16.2
*
65.3
15.6
62.5
*
15.7
32.9
76.6

12.2
11.3
* .
48.8
35.9
102.0
47.6
24.1
It
32.8
*
20.0
21.4

32.2
11.9
32.4
49.9
32.7
*
31.3
15.9
*
65.7
38.3

72.9
34.0
46.4
, *
*
68.2
*
*
39.9
49.2
15.6
40.0
32.2

48.3
59.4
64.7
12.5
16.3
31.2
15.6
31.8
47.0
*
*

*
*
61.9
24.4
53.9
34.1
23.8
48.2
19.9
*
62.2
*
53.6
*Not spiked in sample.
"Paraldehyde is included in the volatile group since EPA has listed this compound as amenable
to the purge and trap methodology. Analytical experience has indicated, however, that
paraldehyde is not purgeable.
1tribromomethane
2dl-1,3-butadiene diepoxide
32-methyl 1-propanol
42-butanone

-------
                                    TABLE 3

                             SPIKE CONCENTRATIONS
                     IN SAMPLE SOLUTIONS (micrograms per liter)
                                 SEMIVOLATILES
                                                                           96
 Compound
                                                 Sample Number
Priority Pollutants
2-chlorophenol
1,2-dichlorobenzene
1,3-dichlorobenzene
di-n-octylphthalate
naphthalene
4-n'rtrophenol
p-chloro-m-cresol1
2,4,6-trichlorophenol
Other Appendix VIII Compounds
aniline
benzenethiol2
4-chlorophenol
3-chloropropionitrile
di-epinephrine
diethylstilbesterol
3,3'-dimethyIbenzidine
di-n-propylnitrosoamine3
diphenylamine
ethylenethiourea4
hexachlorophene
methoxychlor
methylmethacrylate
methyl parathion
phenacetin5
thioacetamide

92.0
*
157.0
37.3
231.0
121.0
*
32.7

93.4
*
*
59.3
87.6
63.2
39.7
146.0
it
39.4
It
181.0
81.9
77.0
178.0
88.4

*
158.0
39.3
74.7
46.2
*
277.0
196.0

31.1
48.2
161.0
*
117.0
31.6
79.4
*
39.6
78.7
162.0
*
*
51.3
59.3
118.0

30.7
79.1
*
149.0
*
40.4
185.0
98.0

187.0
72.3
80.3
119.0
*
94.8
*
97.3
79.2
118.0
81.1
120.0
40.9
103.0
it
*

153.0
39 6
W%/ป \J
78 6
/ \J*\J
*
139 0
1 ww * \J
so a
\j\j t\j
92.3
*

*
121 n
1 f~ I • \J
40 2
"WปC.
29 7
fm\J t 1
58.4
*
1190
1 1 W . W
48 6
"w*U
1*58 n
1 \J\J* \J
*
122 0
1 CnC_> \J
1RO n
i \J\Jt \j
164 0
I \J^ • \J
*
1 IQ n
I 1 *7. \J
58.9
*Not spiked in sample.
M-chloro-S-methylphenol
2thiophenol
3n-nitrosodi-n-propylamine
42-imidazolidinethione
5p-acetophenetidide

-------
                                                                            97
                                    TABLE 4

                 PERFORMANCE COMPARISON OF LABORATORIES
       BY INTERLABORATORY PRECISION FROM METHODS 624 AND 625 (304(H))
                              PRIORITY POLLUTANTS
                          Intel-laboratory Range*(|og/l)     Number of Times Outside of Range
             Sample  1
                                                      Lab A  LabB  Lab C
                                                     (SW-846) (304(h)) (304(h))
Volatiles
acrolein
benzene
bromoform
chlorobenzene
chloroform
1,1-dichloroethane
1 ,2-dichloroethane
1 ,2-dichloropropane
ethylbenzene
toluene
1,1,1-trichloroethane







— not given for Method 624 — ...
21-50
**
17-37
**
44-87
It*
6-90
42-116
14-23
9-20
**
8-25
**
39-83
11-23
38-90
**
12-24
23-46
50-117
9-17
21-50
29-73
20-42
**
20-45
2-30
**
41-92
26-59
32-83
47-101
11-18
10-21
22-44
10-22
4-60
27-70
**
**
1
0
0
0
1
0
0
3
0
1
0
2
0
0
1
0
0
0
0
0
1
2
0
0
0
0
0
0
0
0
Semivolatiles
                  31-113
2-chlorophenol
1,2-dichlorobenzene    **
1,3-dichIorobenzene  27-248
di-n-octylphthalate
naphthalene
4-nitrophenol
p-chloro-m-cresol       **
2,4,6-trichlorophenol  13-46
                    5-51
                  74-280
                   4-142
66-187
 6-61
13-101
 16-57
  **

98-368
98-259
 9-39
 33-94
  **

33-225
  **

 0-50
65-247
47-131
                                                   Total 6
52-187
 16-48
13-124
  **

46-169
 0-96
31-125
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
"Calculated from equations in Table 6, 49 FR 43379
**Not spiked in sample
                                                   Total 1

                                                   Grand
                                                   Total 7

-------
                                                                              98
                                     TABLE 5

                            COMPARISON OF ACCURACY
              FOR LABORATORIES USING 304(H) AND SW-846 METHODS
                                    VOLATILES
                                                   Average Percent Recovery
                                          Lab A(SW-846)  Lab B(304(h))   Lab C(304(h))
Priority Pollutants
       acrolein
       benzene
       bromoform
       chlorobenzene
       chloroform
       1,1-dichloroethane
       1,2-dichloroethane
       1,2-dichloropropane
       ethylbenzene
       toluene
       1,1,1-trichloroethane
                            Average
  0.0
 64.2
101.3
107.9
 95.5
144.3
113.2
 87.8
  0.0
 89.4
 79.9
 80.3
                            Standard Deviation  44.6
Other Appendix VIII Compounds

       acetonitrile
       bro mo acetone
       carbon disulfide
       chloroacetaldehyde
       1,2,3,4-diepoxybutane
       1,4-dioxane
       hexachloropropene
       iso-butyl alcohol
       malononitrile
       methacrylonitrile
       methyl ethyl ketone
       paraldehyde
       1,2,3-trichlororopane
  0.0
  0.0
187.7
  0.0
  0.0
  0.0
  0.0
  0.0
  0.0
  0.0
120.7
  0.0
 84.1
  0.0
 97.8
 57.6
 81.5
 87.8
 64.6
105.6
 81.7
 93.2
 91.8
 83.0
 76.8
 29.0
  0.0
  0.0
325.6
  0.0
  0.0
  0.0
  0.0
  0.0
  0.0
  0.0
  0.0
  0.0
 37.0
71.4
48.5
40.7
82.2
92.9
89.7
98.7
81.4
82.2
80.2
96.1
78.5
18.6
 0.0
 0.0
31.2
 0.0
 0.0
 0.0
 0.0
 0.0
 0.0
 0.0
 0.0
 0.0
46.2

-------
                                                                               99
                                      TABLE 6

                            COMPARISON OF ACCURACY
               FOR LABORATORIES USING 304(H) AND SW-846 METHODS
                                  SEMIVOLATILES
                                                    Average Percent Recovery
                                          Lab A(SW-846)  Lab B(304(h))   Lab C(304(h))
Priority Pollutants
       2-chlorophenol
       1,2-dichlorobenzene
       1,3-dichlorobenzene
       di-n-octylphthalate
       naphthalene
       4-nitrophenol
       p-chloro-m-cresol
       2,4,6-trichlorophenol
                            Average
                            Standard Deviation
Other Appendix VIII Compounds

       aniline*
       benzenethiol
       4-chlorophenol*
       3-chloropropionitrile
       di-epinephrine
       diethylstilbesterol
       3,3'-dimethylbenzidine
       di-n-propylnitrosoamine*
       diphenylamine*
       ethylenethiourea
       hexachlorophene*
       methoxychlor*
       methylmethacrylate*
       methyl parathion*
       phenacetin*
       thioacetamide
                            Average*
                            Standard Deviation*
 86.2
101.1
101.4
 96.1
 56.3
  0.0
109.2
 80.0
 78.8
 35.9
 54.6
  0.0
  0.0
  0.0
  0.0
  0.0
  0.0
 86.6
  0.0
  0.0
 10.9
 59.7
 37.4
109.2
 58.5
  0.0
 46.3
 38.1
 70.7
 62.8
 53.5
 81.3
 70.2
 30.7
 97.5
 69.7
 67.0
 19.6
 40.0
  0.0
  0.0
  0.0
  0.0
  0.0
  0.0
 74.5
115.0
  0.0
  0.0
  0.0
  0.0
 41.8
 33.9
  0.0
 33.9
 40.2
 68.7
107.9
 89.5
  6.2
 47.4
 57.1
 87.7
 97.3
 70.2
 33.0
 18.7
  0.0
 47.9
  0.0
  0.0
  0.0
  0.0
 54.1
102.4
  0.0
  0.0
 21.1
 14.6
 10.4
 24.0
  0.0
 32.6
 31.3
       *Average and standard deviation of nine compounds noted.

-------
                                                                             100
                                      TABLE 7

                   FALSE POSITIVE AND NEGATIVE OBSERVATIONS
                          BY COMPOUND AND LABORATORY
                                     VOLATILES
                                                         Observations
                                                   False Negative*(False Positives)
                                                 Lab A
                                                 SW-846
                LabB
                304(h)
              LabC
              304(h)
 Priority Pollutants

 acrolein
 bromoform
 chloroform
 ethylbenzene
 toluene

   Total
3
0
(1)
3
(1)

6(2)
3
0
0
0
0
2
1
0
0
0
 Other Appendix VIII Compounds

 acetonitrile
 bromoacetone
 carbon disulfide
 chloroacetaldehyde
 1,2,3,4-diepoxybutane
 1,4-dioxane
 hexachloropropene
 iso-butyl alcohol
 malononitrile
 methacrylonitrile
 methyl ethyl ketone
 paraldehyde

   Total

 Total for All Compounds
3
3
0
3
3
3
3
3
3
3
0
3

30

36(2)
3
3
(1)
3
3
3
3
3
3
3
3
3

33(1)

36(1)
3
3
0
3
3
3
3
3
3
3
3
3

33

36
Summary
• Ten of the total 24 compounds were never detected by any laboratory. All 10 were other
Appendix VIII compounds.
• Ethylbenzene was never detected by Lab A.
• Methyl ethyl ketone was never detected by Labs B and C.
• Only 3 false positive observations were reported.
'Maximum number of false negative observations possible for any laboratory was three for each
compound since each compound was spiked in 3 of 4 samples analyzed by each laboratory.

-------
                                                                               101
                                      TABLE 8

                   FALSE POSITIVE AND NEGATIVE OBSERVATIONS
                          BY COMPOUND AND LABORATORY
                                   SEMIVOLATILES
                                                         Observations
                                                   False Negative*(False Positives)
                                                 Lab A
                                                 SW-846
           LabB
           304(h)
           LabC
           304(h)
 Priority Pollutants

 naphthalene
 4-nitrophenol

   Total
0
3
(1)
0

(1)
0
0
Other Appendix VIII Compounds

aniline
benzenethiol
4-chlorophenol
3-chloropropionitrile
di-epinephrine
diethylstilbesterol
3,3'-dimethylbenzidine
diphenylamine
ethylenethiourea
hexachlorophene
methoxychlor
methylmethacrylate
thioacetamide

   Total

Total for All Compounds
0
3
3
3
3
3
3
3
3
0
(1)
0
3

27(1)

30(1)
1
3
3
3
3
3
3
0
3
3
3(1)
3
3

34(1)

34(2)
0
3
0
3
3
3
3
0
3
3
0
1
3

25

25
Summary
• Seven of the total 24 compounds were never detected by any laboratory  All 7 were other
Appendix VIII compounds.
• Diphenylamine and 4-nitrophenol were never detected by Lab A.
• Methoxychlor was detected by Labs A and C. Lab B reported it as a false positive obervation
• Only 3 false positive observations were reported.
'Maximum number of false negative observations possible for any laboratory was three for each
compound since each compound was spiked in only 3 of 4 samples analyzed by each laboratory.

-------
                                                                                 102
                                      TABLE 9

                   FALSE POSITIVE AND NEGATIVE OBSERVATIONS
                                      SUMMARY
      Laboratory
                                                   Number of Observations
                                                False
                                              Positives
                                             (PP/App.VIII)
                            False
                          Negatives*
                          (PP/App.VIII)
Volatiles
     A(SW-846)
     B(304(h))
     C(304(h))
2/0
0/1
0/0
6/30
3/33
3/33
Semivolatiles

     A(SW-846)
     B(304(h))
     C(304(h))
0/1
1/1
0/0
3/27
0/34
0/25
PP - priority pollutants in Appendix VIII
App.VIII - other pollutants in Appendix VIII
*Maximum number is 72 (3 samples x 24 spiked compounds)possible false negative observations
for each laboratory (for voiatiles: 33-priority pollutants, 39-other Appendix Vlil compounds; for
semivolatiles: 24-priority pollutants, 48-other Appendix VIII compounds).

-------
CO
o
H
                               CD

                               00
                               CO
                               CD
O
ฃ2-
CD

CD
o
CO,

O
m
                                                                              C/3
                                                                             •*—>

                                                                              cd
                                                                             H—'
                                                                             JD

                                                                             "o
                                                                             DL
                                                                              O
                                                                             "k_
                                                                             Q_

                                                                              CD
                                                                                                                  O
                                                                                                                  o
                                                                                                                  o
                                                                                                                  CD
                                                                                                                 DC
                                                                                                                 •ป— >
                                                                                                                  c
                                                                                                                  CD
                                                                                                                  O

                                                                                                                  O
                                                                                                                 Q_

                                                                                                                  CD
                                                                                                                  CD
                                                                                                                  CO
                                                                                                                  i__
                                                                                                                  CD
                                                                                                                  O
                                                                                                                  O
                                                                                                                 JD
                                                                                                                  CO
                                                                                                                  0)
                                                                                                                  i_
                                                                                                                  13
                                                                                                                  CD

-------
o
H
CD


CO
 I



Q




CQ
O
CO,


CQ

CQ
O

CQ
                                                                         CO
                                                                                         CO
                                                                                     o
                                                                                                             I
                                                                                                              Cfl
                                                                                                              O
                                                                                                             Q_
                                                                                                              g
                                                                                                             "k—
                                                                                                             CL
                                                                                                             o

                                                                                                             1
                                                                                                             CD
                                                                                                             CO
                                                                                  CD

                                                                                  o
                                                                                  o
                                                                                  CD
                                                                                  DC
                                                                                  •+-'

                                                                                  CD
                                                                                  o

                                                                                  CD
                                                                                  Q_

                                                                                  CD
                                                                                  CD

                                                                                  2
                                                                                  CD
                                                                                                             >,

                                                                                                             O
                                                                                                             O
                                                                                                             JD
                                                                                                             CO
                                                                                                             CM
                                                                                                             CD
                                                                                                             O)

                                                                                                             LJL

-------
10
o
H
                            CO
                           T3
                            C
                            D
                            O
                            CL

                            o
                           O

^
CD
CO
1
^^
CD
_J




^;
JZ
Xj"
O
CO^
CO
m
_j
^^3



^^
ฃ,
^sh
o
co^
0
CQ
_J
|iง>i||






1
Q
O





CD
C
'E
03

"





c
CD
sz
CL
O
!c
o

CM
cu
.E
03
0
c/)
O
i_
•<— ;
|c
Q.
0
cL
cr
T3

CO



CD
r~
diphenylamine
hexachlorophel

^- LO





methoxychlor

CD


$
4^
CL*
methylmethac

"^



c
_o
i
o3
Q.
">^
•<— '
CD

CO
CD
CD
^
•0
i—
phenacetin
ther 7 compoui
o
CT>^
                                                                               w
o
                                                                            CD
                                                                                      o
                                                                                                           Q.
                                                                                                           Q.
                                                                                                           CD
                                                                                                          ^1
                                                                                                          •<— >
                                                                                                          O

                                                                                                           CD
                            O

                           1
                            CD
                           CO
                                                                                                           CD

                                                                                                           O
                                                                                                           O
                                                                                                           CD
                                                                                                           DC
                                                                                                           •+—•
                                                                                                           c:
                                                                                                           CD
                                                                                                           O

                                                                                                           CD
                                                                                                           Q_

                                                                                                           CD
                                                                                                           O)
                                                                                                           CO

                                                                                                           CD
                                                                                                            >~,
                                                                                                            o
                                                                                                           •g
                                                                                                            o
                                                                                                           JD
                            CO


                             CD
                             \—
                             13
                             D)
                            LL.

-------
                                        106
         INTRODUCTION
Follow-up to previous Chemical Manufacturers
Association (CMA) study

Objectives of present CMA study

•   Compare performance of alternative GC/MS
   methods (304(h)) to GC/MS methods of
   SW-846 for analysis of organics in
   Appendix VIII

•   Compare performance of both groups of
   methods for priority pollutants versus other
   organic compounds in Appendix VIII

•   Evaluate those compounds which are
   amenable to GC/MS methods

-------
                                              107
                STUDY DESIGN
Ground water matrix spiked with organic compounds
in various concentrations and combinations

4 spiked samples

24 volatile compounds

•  1 1 priority pollutants (1 1 .9 - 78.4 jig/L)
•  1 3 other Appendix VIII compounds (1 1 .3 - 1 02.0 ja,g/L)

24 semivolatile compounds
•  8 priority pollutants (30.7 - 277.0 |ig/L)
•  1 6 other Appendix Vlll compounds (29.7 - 1 87.0

3 prominent laboratories

•  1 used SW-846 methods
•  2 used 304(h) methods

-------
                                             108
               STUDY DESIGN
304(h) GC/MS methods
•  624 (volatiles)
•  625 (semivolatiles)
SW-846 GC/MS methods
•   8240 (volatiles)
•   8270 (semivolatiles)

-------
                                              109
                 PRECISION
Study not designed specifically for precision calculations

Qualitative comparison made based on precision equations
for methods 624 and 625

-------
                                                                      110
              PERFORMANCE  COMPARISON  OF LABORATORIES
  BY  INTERLABORATORY  PRECISION  FROM  METHODS  624 AND  625 (304(H))
                            PRIORITY  POLLUTANTS
                          Intel-laboratory Range*(ng/l)     Number of Times Outside of Range
             Sample  1
                                  Lab A  LabB  LabC
                                  (SW-846)  (304(h)) (304(h))
Volatiles

acrolein
benzene
bromoform
chlorobenzene
chloroform
1,1-dichloroethane
1,2-dichloroethane
1,2-dichloropropane
ethylbenzene
toluene
1,1,1 -trichloroethane
Semivolatiles

2-chIorophenol
1,2-dichlorobenzene
1,3-dichlorobenzene
di-n-octylphthalate
naphthalene
4-nitrophenol
p-chloro-m-cresol
2,4,6-trichlorophenol
     —• not given for Method 624 —
21-50     **      9-17    32-83
         8-25   21-50   47-101
17-37     **     29-73    11-18
                         10-21
                         22-44
                         10-22
                         4-60
                         27-70
**
44-87
**
6-90
42-116
14-23
9-20
39-83
11-23
38-90
**
12-24
23-46
50-117
20-42
**
20-45
2-30
**
41-92
26-59
31-113
**
27-248
5-51
74-280
4-142
**
13-46
**
66-187
6-61
13-101
16-57
**
98-368
98-259
9-39
33-94
**
33-225
**
0-50
65-247
47-131
52-187
16-48
13-124
**
46-169
0-96
31-125
**
'Calculated from equations in Table 6,49 FR 43379
"Not spiked in sample
      1
      0
      0
      0
      1
      0
      0
      3
      0
      1

Total 6
                                     0
                                     0
                                     0
                                     0
                                     0
                                     1
                                     0
                                     0

                               Total  1

                               Grand
                               Total  7
0
2
0
0
1
0
0
0
0
0
            0
            0
            0
            0
            0
            0
            0
            0
1
2
0
0
0
0
0
0
0
0
       0
       0
       0
       1
       1
       0
       0
       0
                                        16

-------
                                                 Ill
           ACCURACY - VQLATILES
Volatile priority pollutants

•  No statistical difference between labs at 5% level
•  Average recoveries (80.3, 76.8, 78.5 %)

Other Appendix VIII volatile compounds

•  Mixed recovery results on 3 compounds
•  Other 10 compounds were never detected by any lab

-------
                                                 112
         ACCURACY • SEMIVQLATILES

Semivolatile priority pollutants

•  No statistical difference between labs at 5% level
•  Average recoveries (78.8, 67.0, 70.2 %)

Other Appendix VIII semivolatile compounds

•  No statistical difference between labs at 5% level
•  7 compounds were never detected by any lab

-------
CD

CO
 I



ฃ2.



CD
O
ฃ0-

CQ

CD
O
CO,


O

CO
ซ .E
C CD
                        CO CD
                                         O
         O    O    O   O    O    O
                                                             03
                                                             •*—'
                                                             JD

                                                             "5
                                                             CL
                                                             g
                                                             "ป_
                                                             Q_
                                                             O
                                                                               >,

                                                                               o5

                                                                               O
                                                                               O
                                                                               CD
                                                                               cc
                                                                               (D
                                                                               O

                                                                               CD
                                                                               Q_

                                                                               CD
                                                                               O)
                                                                               03
                                                                               i_ •
                                                                               CD
                                                             o
                                                            JD
                                                             Oi
                                                             CD
                                                             i_
                                                             13
                                                             D)

                                                             LL

-------
H
H
                                                                          I
                                                                          03
                           CD

                           00


                           C/D,

                           <
                           m
o
CO,

m
m
o
ฃ2-
O
m
                                                                                   O
 o
a.
.ฃ•
 v_
 o
ol


1
 o

1
 CD
CO
                                                                                                          CD

                                                                                                          O
                                                                                                          O
                                                                                                          CD
                                                                                                         DC
                                                                                                         •*—•
                                                                                                          c
                                                                                                          CD
                                                                                                          O
                                                                                                          i_
                                                                                                          CD
                                                                                                         Q_

                                                                                                          CD
                                                                                                          CD
                                                                                                          C3

                                                                                                          CD
                                                                                                          >>

                                                                                                          O
                                                                                                          O
                                                                                                         JD
                                                                                                          03
                                                                         CvJ


                                                                          CD
                                                                          i_
                                                                          ZJ
                                                                          D)
                                                                         LL

-------
LO
H
                                         CO
                                         •a
                                         c
                                         D
                                         O
                                         Q.
                                         E
                                         o
                                         O
CD

00
                        CD
      O

      d

      CD

      CD
O
CO,

o

CD
                                                  . E  R-S ,	
   CX Q. cd  2 O CD Ctf
   O  O tr-  O >^ P Q.
n^ป-  ป-  >^-T= xฃ 	
^_g  Q. ง •ง o^>
                                   9^ o  .
                                   CD >^ CO
                                   ^ s ฎ
                                   co

      o
    r-tซ
alE
                                                                       c
         E 815
         8SS
                  c5
                                             o3
                          r5 ^ ^
                           *-' T3 J=
                        Q)  CD  CD
                        E  E  E
                                                                    CD
                                                                       _
                                                                       CD
                                               cMco^LncDh-oocn+r
                                                                          
                                         O
                                         CD
                                         O


                                         E
                                         CD
                                         CO
                                                                                                       Q)
                                                                                                       >
                                                                                                       O
                                                                                                       O
                                                                                                       CD
                                                                                                       DC
                                                                                                       •*— '
                                                                                                       c
                                                                                                       CD
                                                                                                       o

                                                                                                       CD
                                                                                                       CL

                                                                                                       CD
                                                                                                       D)
                                                                                                        CD
                                                                                                        O
                                                                                                        o
                                                                                                       _a
                                                                                                        CD
                                                                                                       u_

-------
                                                 J.16
FALSE POSITIVE AND NEGATIVE OBSERVATIONS
                VOLATILES
Only 3 false positive observations reported
•  2 - 304(h); 1 - SW-846

Large number of false negative observations by all
Slabs
•  3 of 11 priority pollutants
•  11 of 13 other Appendix VIII compounds

-------
                                                       J.17
FALSE POSITIVE AND NEGATIVE OBSERVATIONS
               SEMIVOLATILES
Only 3 false positive observations reported
•  2 - 304(h); 1 - SW-846

Large number of false negative observations by all
3 labs, but fewer than for volatiles
•  1 of 8 priority pollutants
•  13 of 16 other Appendix VIII compounds

-------
                                                                118
        FALSE POSITIVE  AND NEGATIVE  OBSERVATIONS
                          SUMMARY
        Laboratory
Volatiles
    A(SW-846)
    B(304(h))
    C(304(h))
                                       Number of Observations
                                       False
                                    Positives
                                   (PP/App.VIII)
                   False
                 Negatives*
                (PP/App.VIII)
2/0
0/1
0/0
6/30
3/33
3/33
Semivolatiles

    A(SW-846)
    B(304(h))
    C(304(h))
0/1
1/1
0/0
3/27
0/34
0/25
PP - priority pollutants in  Appendix VIII
App.VIlI -  other pollutants in Appendix VIII
                                21

-------
                                                    JJ.9
      FALSE NEGATIVE OBSERVATIONS
                 SUMMARY

VOLATILE
Priority pollutants
•   SW-846:   18%
•   304(h):    9%
Other Appendix VIII compounds
•   SW-846:   77%
•   304(h):    85%

SEMIVOLATILE
Priority pollutants
•   SW-846:   13%
•   304(h):    0%
Other Appendix VIII compounds
•   SW-846:   56%
•   304(h):    52-71%

VOLATILE and SEMIVOLATILE
•   Priority pollutants:     3% (15 out of 432 analyses)
•   Other Appendix VIII:    42% (182 out of 432 analyses)

-------
                                                      120
FALSE POSITIVE AND NEGATIVE OBSERVATIONS
                 SUMMARY
No significant difference among laboratories/methods

More false negative observations than false positive
observations

Less false negative observations for priority pollutants

False negative observations generally occur at lower
concentration levels

-------
                                                     121
   CONCLUSIONS AND RECOMMENDATIONS
GC/MS methods from SW-846 and 304(h) perform equally
well for the analysis of organic Appendix VIII compounds
in ground water

 Cost of 304(h) GC/MS methods are 1/4 to 1/3 cost for
SW-846 GC/MS methods

List of compounds amenable to GC/MS methodology is
shorter than Appendix VIII list (45% of spikes were
reported as false negative observations)

GC/MS methodology is a more amenable to analysis of
Appendix VIII compounds which are    priority pollutants
•.  17 of 48 compounds never reported by any lab
•  All 17 compounds were not priority pollutants

-------
                                                       122
   CONCLUSIONS AND RECOMMENDATIONS
False negative observations are more likely to occur at
lower concentrations

Data user should interpreted analytical with caution
•   False negative observations
•   False positive observations

Labs were aware of study scope and objectives
•   Analytical results represent "best" performance

Additional studies suggested
•   304(h) GC/MS method validation for Appendix VIII list
•   Reagent water and ground water
•   GC/MS reference spectra from study should be
   incorporated into library similar to that for priority
   pollutants

-------
                                                    123
             QUESTION AND ANSWER SESSION
                          MR. APRIL:  Could you




explain what some of the more striking differences




are between the 304(h) methods and the SW-846?  I was



especially interested by the large cost differential.




                          MS. KOCUREK:  Do you mean



differences in the analytical procedures or...




                          MR. APRIL:  Just the



differences in the procedures, especially the ones



that lead to the large cost




differential.




                          MS. KOCUREK:  I can't



really address the differences in the analytical




procedures because I'm not an analytical chemist.  I



merely looked at the data that resulted from using



these methods.  So I can't address analytical




methodologies, the specifics of them.



                          MR. LIN:  Denis Lin from



ETC.   Well, maybe I can answer Bob's question, too.




But let me first point out that ethyl benzene is a



priority pollutant.  It is not on the Appendix VIII



list, therefore I don't know what affect it will




have in your false positive/false negative or position



study because one lab did not report ethyl benzene.

-------
                                                    124
I kind of suspect that lab was not looking for it
because it is not an Appendix VIII compound.
                          MS. KOCUREK:  Yes, let me
take a look.
                          MR. LIN:  While Diane is
looking, Bob, the answer to your question is if it is
comparing apples to apples, in other words, if somebody
asks for 8240 as opposed to 624, I'm sure the price
differential is not as great as stated, or for that
matter, 625 as opposed to 8270.  However, if the
request is for all the methodology involved in so-
called Appendix VIII analysis, which include all the
GC method, all HPLC method, I'm sure the price
differential will add up very quickly.
     If you are looking for say the volatile
analysis, no, I'm sure there is some difference,
because Appendix VIII volatiles will go up to about
50, 55 compounds.
                          MR. APRIL:   ...then there's
no significant cost difference.
                          MR. LIN:  We can do a survey
here.  There's enough labs around.  The answer to
your question is if you are looking for the same
compounds, I don't think there should be substantial
difference.

-------
                                                    125
                          MS. KOCUREK:  Denis, I've
got an answer to your question.  The false negative
observations for ethyl benzene for one of the labs
was reported three out of the three spiked samples,
so, yes, it's possible that if it were not reported
because it was not an Appendix VIII compound that
could explain the three false negatives.
                          MR. LIN:  Yes.  It is very
obvious to me.  I checked the list, so I figured out
that probably is the reason.
                          MS. KOCUREK:  Yes, it
would make a slight difference.  It wouldn't make a
difference in the comparison between the methods
because overall, the performance wouldn't change that
much.
                          MR. STANKO:  George Stanko
from Shell.  I'd like to explain a little bit of
difference between Method 624, 625 and the 8240 and
8270.  The 624, 625, the 304(h) methods were designed
primarily to satisfy the requirements of the Clean
Water Act.  The list of priority pollutants is the
only compounds listed in these methods although the
method is amenable to other compounds.  In other
words, you could look for priority pollutants plus.
Laboratories who offer that service will charge you

-------
                                                     126


almost the same as the priority pollutant  list.   If


they don't find any, you don't get charged any more


for 624 or 625 than you would if they had been looking


for an effluent or priority pollutants alone.


     Laboratories who are using or supposedly using


8240 and 8270 are supposed to be looking for all


Appendix VIII compounds.  Now, one of the compounds


was just pointed out right now may be a priority


pollutant but it's not on the Appendix VIII list.  It
                             l

is amenable to the GC/MS methodology.


     When you request analysis for a groundwater  sample


for Appendix VIII compounds, whoever those laboratories


are who are capable of doing that, and I have some


doubts at this point in time, you will be charged


more for that sample and analysis on that sample.


Our experience has shown, when we have submitted


samples and requested Appendix VIII analysis which


include 8240 and 8270 plus some additional methods


which also don't work, the data you get back costs


you four times as much for the laboratory doing SW-


846 methods as the laboratories...if you sent the


same sample and requested 304(h) plus.


     The analytical methods here again,  the procedures


that you follow are the same.  However,  304(h)  will


not allow you to run a solid through the purge  and

-------
                                                    127




trap device.  Now, I don't know how it's done in SW-



846, but in theory, you can use SW-846, Methods 8240



and 8270 for solids as well as groundwater samples.




     Now, this paper dealt nothing beyond groundwater



samples.  But that is another difference in the two



methods.  There is more flexibility allowed in 8240




and 8270 with respect to pre-sample preparation,




cleanup techniques and all that, which are not allowed




in the 304(h) methods.



                          MR. TELLIARD:  George, do



you think that the cost of the analysis was just the




fact that Shell was sending out the samples?



                          MR. STANKO:  There was a



true positive...




                          MS. KOCUREK:  Are there



any other questions?



                          MR. TROIANO:  Jeff Troiano,




Ford Motor Company.  Could it be that this cost



differential, the magnitude that you indicated, is



that based on just using the costs associated with




using these three labs or a much larger population?



                          MS. KOCUREK:  I believe



it's just using the three labs.  George, you may want
to confirm that.
                          MR. STANKO:  There is a lot

-------
                                                    128
of cost information available for Method 624 and 625.
For this study, we found only one laboratory that
claimed to be able to run SW-846 for all Appendix
VIII.  We have a second laboratory that said they did
and they backed out.
                          MR. TROIANO:  Well, going
back to my question, is the difference in price
between 304(h) and SW-846 methods, is it based on
these three laboratories or a survey of many
laboratories?
                          MS. KOCUREK:  Well, I
think what George was saying is that the SW-846 cost
was based on one lab, and then the other ones were
based on known, I think, analytical costs for 624 and
625.
                          MR. TROIANO:  And the SW-846
costs were based on all Appendix VIII or just the
ones in this test?
                          MS. KOCUREK:  I think it
was derived from the cost estimate for this test.  Is
that correct, George?
                          MR. STANKO:  Basically, yes.
                          MR. ARLAUSKAS:   Joe
Arlauskas with Martin Marietta Environmental Systems.
I found something else out, and somebody can correct

-------
                                                    129
me on this/ but the published list of compounds under




the October Federal Register 624 method, which are




supposedly priority pollutants, actually includes




compounds; at least I think one or two that are not



priority pollutants.  And one of those, I think, are



the trichlorofluoro...




                          MS. KOCUREK:  Are you



saying October '85 or '84?




                          MR. ARLAUSKAS:  '84.



                          MS. KOCUREK:  '84?  I



don't remember that.




                          MR. ARLAUSKAS:  Well, if



you compare the priority pollutant list against those




compounds, six...and published in that...



                          MS. KOCUREK:  You're



talking about what was published in the Federal



Register?




                          MR. ARLAUSKAS:  As the



compounds to be analyzed under 624.



                          MS. KOCUREK:  Oh, okay.




I'm thinking of the tables that were presented under



the method for the, like precision and accuracy



equations.  I only recall the priority pollutants



being listed for that.



                          MR. ARLAUSKAS:  Okay.

-------
                                                    130
                          MR. TELLIARD:  But you have
to remember on that publication, it's rather spurious
and you really can't count on it.
                          MR. ARLAUSKAS:  But a lot
of laboratories will give you all that analysis as
they are priority pollutants, and if you're not
careful as a client...
                          MS. KOCUREK:  Right.
                          MR. ARLAUSKAS:  ...you're
going to get parameters that you don't have to look
for.
                          MS. KOCUREK:  That's
correct.  Yes.
                          MR. STITES:  Ron Stites
with Cenref Labs.  I don't know if I can shed any
light on this at all, but I can tell you the practice
we have in our laboratory which may show a difference
in pricing between the 624, 625 and then true Appendix
VIII analysis, which I don't think anybody can really
do.
     If we have a reguest to run Appendix VIII
compounds, we have to talk to the client, and the
first thing we ask is, okay, do you really want us to
standardize for all of the compounds that we can get
our hands on and actually see if we can get some

-------
                                                    131




recovery, which is a quasi-research project because



you really don't have methods you can really go to



and say for sure this is going to work, or do you



want us to just run basically 624, 625 and see if



anything else shows up on the TICs, Tentatively



Identified Compounds.  And there's a big cost



differential in those two approaches.



     Just the fact of purchasing all sorts of compounds



and injecting on your GC/MS and standardizing and



getting response factors, there's where the big cost



differential comes from.  I say again, I don't think



anybody can run Appendix VIII compounds, not really,



unless they just do it by definition and say we run



it this way, therefore we can do it.



                          MS. KOCUREK:  Thank you.



                          MR. TUROSKI:  Victor Turoski,



James River.  What were the three spiking levels that



you used?



                          MS. KOCUREK:  They were



all different for the different compounds.  That's



why I gave the range on that one overhead for the



different priority pollutants.  They were different



for each compound.



                          MR. TUROSKI:  I'm sorry, I



missed that.  What were they?

-------
                                                    132
                          MS. KOCUREK:  You want it
for every compound?
                          MR. TUROSKI:  General range,



                          MS. KOCUREK:  General



range?  Okay, just a second.  For the volatiles the



priority pollutants range was about 12 to...it looks



like 78 Mg/L.  For the other Appendix VIII compounds,



the volatiles, it was about 11.3 to 102 Mg/L.  For



the semivolatiles, priority pollutants, the spike



range was 30.7 to 277 Mg/L.  For the other Appendix



VIII compounds, semivolatiles, it was 29.7 to 187



Mg/L.



                          MR. PRESCOTT:  Dianna,



while I'm walking around here I have a



question.



                          MS. KOCUREK:  Yes, Bill.



                          MR. PRESCOTT:  Did all



three of the laboratories know what compounds to



look for, or were they told the compounds that have



been spiked are on the Appendix VIII list?



                          MS. KOCUREK:  They were



told to look for Appendix VIII compounds.  I don't



think they were given the list of compounds.



                          MR. STANKO:  The two



laboratories using 304(h) were told to look for

-------
                                                    133




priority pollutants plus anything else they could



identify; primarily the Appendix VIII compounds.



                          MS. KOCUREK:  Right.  Now,



let me add, on any of the false positive observations,



they were only limited to the list of spiked compounds.



We did not pull out any data beyond the list of 48



that we had.



                          MR. STANKO:  One other



comment on the spiking levels that were used.  They



were multiples, and they were chosen for a purpose not



at even numbers.  Previous studies we have done 50,




100 and 150.  This time we tried not to bias the



data.  We picked some numbers between 30 and 80 for



the 50, and 100 was between 80 and 120.  Whatever it



weighed out, that's what got put in.  The numbers are



the true values that were actually put in the samples



and they were odd intentionally.



                          MR. TELLIARD:  That figures.



Thank you very much, Dianna.



                          MS. KOCUREK:  Thank you.

-------
                                                    134
                          MR. TELLIARD:  Our next



speaker is Dave.  Do you want to come on up?

-------
                                                                  135
                             DAVID N.  SPEIS

          ENVIRONMENTAL TESTING AND CERTIFICATION CORPORATION
                 METHOD PERFORMANCE CHARACTERISTICS FOR
                  SELECTED RCRA APPENDIX VIII ANALYTES
                             MR. SPEIS:  Actually, I feel that

my presentation might better be entitled, What I Did On My Summer

Vacation in 1982, since it is the third presentation on Appendix VIII

analytes this morning.


     I'm happy to have the advantage of following Sam and Dianne since

I will be showing a different perspective for the performance of RCRA

Appendix VIII analytes in water.


     Our laboratory has been able to develop a quality assurance data

base, which has given us the capability of gathering precision and

accuracy data for some of the more unusual Appendix VIII analytes wbj.:h

we spiked into groundwater.  From that information, we can illustrate

performance expectations for SW-846.  I'll be sharing this information

with you later on this morning.


     Before I get started, I'd like to take care of a bit of

housekeeping and introduce my colleagues who worked with me on this

presentation.  Of course,  we have Denis Lin who provided significant

input into the initial classification of many of the compounds on the

-------
                                                                  136
Appendix VIII list.  He has also provided significant imput to EPA on



the recent short term guidance that was issued by the agency.  Today,



he's my chief slide flipper.





     Jim Bower is our data systems specialist at ETC.  Jim is the



person who designed and developed the QA data base that we use for all



our analytes.  His work enabled me to gather the information that I'm



going to be showing you a little bit later on.





    ETC was presented with the unique opportunity of being involved



with the development of the Appendix VIII analytical chemistry from the



ground up.  Our involvement began in July '82, when the agency first



published the list of Appendix VIII analytes.  We perceived a need for



the analysis of groundwater samples for these compounds.  Based on the



perceived need, we attempted to classify the compounds into analytical



methodology that we could perform for our clients.





     We began the large and difficult task of classifying the Append!::



VIII compounds by common physical and chemical properties so they ccvild



be lumped into what we call a survey method approach.  This approach if>



designed to analyze a large number of compounds with similar properties



using the fewest number of methods as possible.  We are all aware of



survey method approaches and the difficulties associated with



performing them.  As the number of compounds targeted by a specific



technique increases, you decrease your chances for successful analysis



for all those compounds targeted by the method.

-------
                                                                  JL37
     The primary objective in using  a survey method approach is to get



data quickly.  If you took a more singular approach, an approach that



used many methods for many compounds, obtaining data would be extremely



difficult, this approach was designed to produce data.





SLIDE 2






     In September '82, the agency facilitated the method development



process by incorporating SW-846 as the methods of choice for Appendix



VIII.  What we did at that point was to merge our list of classified



compounds into the SW-846 analytical scheme.  The results of that



merging process are shown in the second slide.





     There are six general techniques used for Appendix VIII analytes



in the SW-846 approach.  Volatile organics are performed by Method



8240, extractable organics using a modification of Method 8270.   The



modification is initial extraction of the sample at neutral pH or



ambient pH followed by basic extraction which is combined with the



neutral extract prior to analysis before going to acidic extraction of



the sample.   Pesticides and herbicides are analyzed by Methods 8040,



8140, and 8150; water soluble volatile organics by direct aqueous



injection;  polar and thermally labile organics by high performance



liquid chromatography using either direct aqueous injection or



liquid/liquid extraction.   Metals and metal complexes  were performed by



atomic absorption and inductively coupled Argon Plasma.

-------
                                                                  138
     The original Appendix VIII compound list started out with



approximately 375 compounds.  Included on the list were 51 compounds



that were metals or metallic compounds. Also included on the list were



27 compounds that were considered exotic.  By exotic, I mean they would



require a unique or specific method to determine their presence.



During the categorization process, we identified an additional 44



compounds that we thought were also exotic or unstable in water.  We



omitted those from our classification.  This left 253 compounds from



the original list for categorization into the SW-846 based analytical



scheme we have described.





     By far, the two largest compound classes were the volatile



organics, which had a total of 59 compounds including 28 priority



pollutants and thfe extractable compounds which consisted of 164



compounds.  As you can see and as I mentioned earlier, executing survey



methods for a group of compounds which numbered 164 is extremely



difficult.





    I'll be spending a bit more time on the extractable organics and



the volatile organics later on in the presentation, when I discuss th>a



information we have extracted from our data base.





SLIDE 3





    For our analytical scheme, we reduced the metals list,  which had



started out as 51 metals and metal compounds, to a total of 22 metals.

-------
                                                                  139
I'm going to be moving through these slides quickly since much of this



ground has previously been covered this morning.





SLIDE 4





     The extractable HPLC list in this classification numbers 20



compounds. .One of the more difficult tasks in Appendix VIII analysis



was obtaining these compounds.  The HPLC portion of the Appendix VIII



scheme, as most of you already know, has been excluded from the program



by the short term guidance.  This change was extremely pleasing to our



HPLC specialist.





SLIDE 5





    We also have a number of compounds that we classified for direct



aquesous injection HPLC.  There are 18 compounds on that list.   These



compounds have some water solubility and our early investigations



indicated that it would be very difficult to recover these compounds



from water using liquid/liquid extractions.





SLIDE 6





     We've classified 26 compounds on the list as pesticides which can



be analyzed using gas chromatography (GC) with an electron capture



detector.  Fifteen of those compounds are priority pollutant pesticides



and seven are Arochlor PCBs.   We classified an additional four

-------
                                                                  140
pesticides into this group which we determined would be amenable to



analysis by electron capture GC.





SLICE 7





Eight compounds are phosphorus-containing pesticides that can be



determined using GC with a flame photometric detector.





SLIDE 8





    The herbicides rated their own special classification.  We followed



the traditional approach for these compounds using liquid/liquid



extraction followed by diazomethane methane derivatization and electron



capture GC analysis.





SLIDE 9





     A large number of compounds are water soluble volatile organics



that we determined could be analyzed using direct injection GC/MS.



Last year we presented some information at this symposium that showed



that many of these compounds could be successfully analyzed using a



heated purge and trap technique.  Using that technique, we were able to



achieve much lower detection limits.  Since that time, however, we



haven't devoted sufficient attention to heated purge and trap



methodology.  We are therefore continuing to perform the analysis by



direct aqueous injection GC/MS.  A number of the compounds on DAI/VOA

-------
                                                                  141
list were not amenable to analysis using heated purge and trap, which



supports our reasons for using direct aqueous injection instead.





     The most difficult task that we faced after categorizing these



compounds into the different methods that we could use for their



analysis was to find standard reference materials.  We were not under



the same restrictions that Sam was.  We were not required to use the



EPA repository, and as a result we were free to go to any source that



we could find.





     Nonetheless, it was still extremely difficult to obtain many of



the compounds on the Appendix VIII list.  I'd estimate of the



approximately 250 compounds that we classified, we could not find



sources for about one third of them.





     Our next task, after procuring the standards, was to determine



whether we could actually analyze the compounds using the scheme we



proposed.  Our first step in this process was to perform instrument



detection limit experiments which was equivalent to determining whether



these compounds could be analyzed by GC/MS.   This work was performed



using standards only.  It was not performed using spikes,  and it was



done only for the Appendix VIII compounds, which were not on the



priority pollutants list.  We borrowed detection limit data that had



previously been developed using Methods 624 and 625 for those priority



pollutants that were also listed as Appendix VIII targets.  For those

-------
                                                                  142
specific priority pollutant  an instrument detection limit study was not
performed.

     Using the instrument detection limit data developed for the
Appendix VIII compounds, we  began spiking experiments to determine if
we could actually recover the compounds which had been assigned to
various categories using the analytical scheme designed for those
catagories.  These spiking experiments used the IDL values that we had
obtained in our initial studies as the starting point.  If we were not
successful with the initial  concentration we had used for spiking, we
increased the spiked concentration until we were either successful or
we decided further increase  was not justified because of previously
unknown technical concerns.

SLIDE 10

     After going through the process of obtaining standards and
performing our IDL and spiking experiments,  there were 28 compounds
remaining that we had categorized into analytical schemes that we were
experiencing extreme difficulties analyzing.  Whether these
difficulties were the result of chromatographic problems associated
with compound polarity or thermal lability has not been investigated.
However, I'm certain that some of these problems were extraction
problems as well,  and for lack of a better term,  I've classified these
28 compounds as the Appendix VIII mystery organics.

-------
                                                                  143
     After performing the spiking experiments, we determined we were



ready to offer an Appendix VIII analytical scheme to clients.  Our



system at ETC is somewhat different than most laboratories.  When we



analyze batches of samples, the batches can range in size anywhere from



one sample to 16 samples.  In that analysis we include a spiked blank



and a spiked matrix sample, which gives us two valuable pieces of QA



information.  We can get recovery data for both spiked blanks and



spiked matrices.





     About six months ago, our systems specialist, Jim Bower, designed



a QA data base which would allow us to automatically transfer processed



GC/MS data files from the GC/MS data system directly to an HP3000



computer.  In the 3000 we can manipulate the information in a wide



variety of modes to obtain statistical information on any facet of



GC/MS analysis.





     The construction of that data base gave us the opportunity to



compile some very valuable information on Appendix VIII analysis.  T^.at



information consists of precision and recovery data for Appendix VII'I



compounds that previously had not been accumulated on a large scale.



I'll be focusing on this data for the next few minutes.





SLIDE 11





    What I did initially was to quickly compare precision and accuracy



data for the priority pollutants which were common to the Appendix VIII

-------
                                                                  144
list that had been analyzed using Method 8240 and Method 624.  There's



some important pieces of information I'd like to point out to you on



this slide.  During the study period for the Method 624 analytes we



were able to obtain all our comparison data within a relatively short



period of time.  The reason for this is that our laboratory performs



many more priority pollutant analysis than we do Appendix VIII



analysis.





     The length of the study period needed to obtain 15 data points



from samples which had been analyzed for Appendix VIII compound was



about three months.  Excuse me, Denis is pointing out an error on my



slide here.  This date is not 2/24/85, it 2/24/86, so the study period



is actually one week instead of one year and one week.





     The spiking concentrations that we used for the priority



pollutants on Method 624 and Method 8240 were based on the detection



limits listed for the 600 methods.  Another item that, I also would



like to point out to you is that it appears as though we've got better



precision for Method 624.  The reason for that is when we intially



started collecting data on Method 8240, we were using a single internal



standard.  All method 624 analysis is performed using multiple internal



standards which yeilds better precision.





    If you notice,  there's a few compounds on the list that show



especially good precision, in fact,  better precision than Method 624,



these compounds happen to elute very close to the internal standard,

-------
                                                                  J.45
which accounts for the better precision.  As far as accuracy is



concerned, it's apparently equivalent but statistical equivalency has



not been determined for this data.





SLIDE 13





     The data that I think you're most interested in seeing is for the



Appendix VIII compounds that are not common to the priority pollutant



list.  Ordinarily when using an internal standard purge and trap



method, I'd expect to get good accuracy, or accuracy in the range of



about 80 to 120 percent recovery for all compounds.  For a few of these



compounds, I'm getting much lower accuracy than I expected.  I don't



really have a good explaination for that at this time.  The focus of



this presentation is not to discuss the specifics of why any one



compound failed within this analytical scheme.  It is geared to display



the type of data obtainable for compounds on the Appendix VIII list



using the analytical scheme I described to you earlier.





SLIDE 14





     We also did the same comparisons for the extractable priority



pollutants.  Again, I did a quick comparison of spiked blank data from



Method 625 with the data from the modified Appendix VIII scheme.   The



compounds on this slide are common to the priority pollutant and



Appendix VIII list.

-------
                                                                  146
     You will notice that the number of data points chosen for the



Appendix VIII analytical scheme is smaller than those for the priority



polllutants.  While preparing for this presentation, I found that



recovering  some of the older data from our data base was going to be



more difficult than I had originally thought.  The reason for this is



that the individual data points would have had to have been entered



manually into data base to be eligible for this study.  As a result I



have a smaller data base to deal with for the data from the



modification of Method 8270 I think, however, that we've got enough



information from the reduced size data base to illustrate trends.





     The precision of Method 625 and Modified 8270 seems to be almost



identical.  In some cases, the precision for the Appendix VIII scheme



comes out better for some compounds.  I'm trying to avoid making



statements  on statistical significance since I have not applied any



statistical significance tests to the precision data, but in a



qualitative sense we do find that a few compounds exhibit a little bit



better precision.  In some cases, we have also noted higher recovery



for the modified Appendix VIII extraction scheme.





SLIDE 15





    Again, we have applied the same comparisions to the matrix spikes.



This time, we see that better precision favors the Method 625 scheme,



which seems to make a statement on whether the precision differences we



are observing are actually significant or not.  I think from the

-------
                                                                  147
information I have displayed we can conclude that the SW-846 schemes



for both volatiles and extractables perform satisfactorily for priority



pollutant compounds.






     We also did a comparison of spiked blank data and matrix spike



data for the Appendix VIII compounds which are not on the priority



pollutants list.  I've selected a cross section of compounds that are



exclusively on the Appendix VIII list to demonstrate the types of



recovery and precision that we are seeing using these techniques.





    We see some unusual behavior that I think is important for each one



of you to consider.  In many cases, we're seeing large standard



deviations.  As a general rule, I'd say the non-polar compounds tend to



perform a little better within the scheme.  The polar compounds have a



tendency towards lower recoveries and higher precisions.  If you've



been forced into the Appendix VIII mode of monitoring for one reason or



another, I think it's extremely important that you know the value of



the data you're getting and whether you're going to be satisfied wibh



results for a specific compound where the percent relative standard



deviation for its concentration could be as much as 100 percent.





     Where does this information I've presented mean for the regulatory



community?  Many of the "exotic" compounds and poor performers have



been excluded by the short term guidance that EPA has recently issued.



This program is in its early adolescence.   I'm sure that most of  you



recall the evolutionary period that the 304(h)  methods went through

-------
                                                                  148
before they were accepted by all of us.  I think that the analytical



scheme used for the Appendix VIII program is going to go through the



same type of evolutionary period.





    I've displayed precision and recovery data for Appendix VIII



compounds that perform well within the scheme and we've also seen data



for compounds that perform poorly within the scheme and probably would



not be acceptable to any of you.  I'm going to leave you with this



final thought and that is, before we can use these methods in a



regulatory or monitoring setting we must be aware of the type of data



that we get from these methods and what type of performance in terms of



precision and accuracy we can expect for any particular analyte.  Thank



you.

-------
                                                                  149
                      QUESTION AND ANSWER SESSION
              MR. TELLIARD:  Any questions?





              MR. FOSTER:  Russ Foster from RAI.  What was your



rationale for separating the neutral extraction, the basic extraction,



and did it meet the objectives?






              MR. SPEIS:  Well, we thought that for some compounds



there was a chance that we might hydrolyze them by making the solution



basic before extraction.  So, to avoid that, we inserted a neutral



extraction followed by a basic extraction and an acid extraction.  As



it turns out, doing a basic and neutral extraction usually recovers all



the compounds that we had spiked.   There's very few compounds that end



up in the acidic fraction.  I think it's less than five, and it may be



as low as two when you employ this extraction procedure.





                  AUDIENCE PATICIPANT:  In doing this, we realize that



there's going to be problems with these new methods, and you've shovw



this statistically.  Well, that's fine and good,  but what if we're



analyzing for these analytes for a client, okay?  Five years from now



they decide that the methodology we used was substandard in their



finding, you know.

-------
                                                                  150
              MR. SPEIS:  This is exactly the point.

              MR. TELLIARD:  You're in deep trouble is what it is>

              MR. SPEIS:  How can you make a regulatory decision using
a method that gives you poor precision and accuracy in terms of 100
percent relative standard deviation?

              MR. TELLIARD:  No, it never stopped the agency in the
past and I don't think it will in the future.  I think, going back to
the question, that we've gone through...! mean, it's taken me almost
nine years to convince STanko that GC/MS would work on priority
pollutants, so you're not going to do this overnight.
              MR. SPEIS:  It's only been three and a half years for
this program.
              MR. RICE:  Jim Rice.  It is interesting to note though,
that the data you have put up is essentially a single lab's QC data.

              MR. SPEIS:  That's correct.

              MR. RICE:  All of which goes to say that if you were to
ever try at this stage of the game, an interlaboratory comparison would
be totally off the wall.

              MR.  TELLIARD:  Thanks so much, Dave.

-------
                                                                   J.5J.
     We're running  a little late, as usual.  For those  folks  who are

new to this facility,  for lunch there's a couple of restaurants in the

hotel.  There's  a number of restaurants next door  at  the Waterfront,

some fast food places.   You can go make yourself a chocolate  sundae,

put a lot of weight on,  whatever you want to do, but  let's  get back

here about 1:30, quarter to., two^ at .'the, latest.  .  . ,.
                       : •--'•< .'••'•••..v'" "..-  ' ,•-. - '•':"' "'''; • >  ""V • '  ' •

     Thank you so much for your attention.


(WHEREUPON, a lunch recess was taken.)

-------
                                       152
METHOD PERFORMANCE CHARACTERISTICS
                 FOR
   SELECTED APPENDIX VIII ANALYTES
            DAVID N. SPEIS
            DENIS C.K. UN
           JAMES N. BOWER
          Environmental Testing
       and Certification Corporation
           Edison, New Jersey

-------
                                         153
APPENDIX VIII ANALYTICAL SCHEME
o  VOLATILE ORGANICS:
       Method 8240, GC/MS

o  EXTRACTABLE ORGANICS:
       Modified Method 8270, GC/MS

o  PESTICIDES & HERBICIDES:
       Method 8080, 8140, and 8150, GC/EC, GC/FPD

O  WATER SOLUBLE VOLATILE:
       Direct Aqueous Injection, GC/MS

o  POLAR & THERMALLY LABILE ORGANICS:
       HPLC Direct Aqueous Injection

o  METALS & METAL COMPLEXES:
       Representative Metals by AA & ICAP

-------
                                 154
       METALS (22)
ALUMINUM
ANTIMONY
ARSENIC
BARIUM
BERYLLIUM
CADMIUM
CALCIUM
CHROMIUM
COPPER
IRON
LEAD
MERCURY
NICKEL
OSMIUM
POTASSIUM
SELENIUM
SILVER
SODIUM
STRONTIUM
THALLIUM
VANADIUM
ZINC

-------
                                           155
POLAR & THERMALLY LABILE ORGANIC COMPOUNDS
           EXTRACTABLE HPLC (20)
      WARFARIN
      MITOMYCIN C
      AZASERINE
      BENZIDINE
      CHLORAMBUCIL
      CITRUS RED NO 2
      DAUNOMYCIN
      3,3-DJCHLOROBENZIDINE
      3,4-Dl H YDROXY-ALPH-(M ETH YLAMINO)-
          METHYL BENZYL ALCOHOL
      3,3-DIMETHOXYBENZIDlNE
      3,3-DIMETHYLBENZIDINE
      1-N APHTH YL-2-THIOURE A
      m-PHENYLENEDซAMINE
      o-PHENYLENEDIAMINE
      p-PHENYLENDIAMINE
      STREPTOZOTOCIN
      THIOACETAMSDE
      TOLUENE-2.4-DIAMINE
      TRYPAN BLUE
      ALDICARB

-------
                                                   156
            POLAR AND THERMALLY LABILE
                ORGANIC COMPOUNDS
            Direct Aqueous Injection HPLC (18)
1-Acetyl-2-thiourea
Acrylamide
1-(o-Chiorophenyl) thiourea
Diethylstilbesterol/Ethyl carbamate
Ethyleneirnine
Ethylenethiourea
Maleic hydrazide
Malononitridine
Methomyl
2-Methylaziridine
Nicotinic acid
Nitroglycerine
N-Nitroso-N-ethylurea
N-Nitroso-N-methylurea
N-Phenythiourea
Reserpine
Thiourea

-------
                                     157
PESTICIDES & POLYCHLORINATED BIPHENYLS
                   by
         GAS CHROMATOGRAPHY

    ELECTRON CAPTURE DETECTOR (26)

   15 PRIORITY POLLUTANT PESTICIDES
   7 POLYCHLORINATED BIPHENYLS (Arochlors)
   CHLOROBENZILATE
   DIMETHOATE
   KEPONE
   METHOXYCHLOR

-------
                                 JL58
           PESTICIDES
               by
     GAS CHROMATOGRAPHY
FLAME PHOTOMETRIC DETECTOR (8)
    CARBOPHENOTH1ON
    THIONAZIN
    DISULFOTON
    METHYL PARATHION
    PARATH1ON
    PHORATE
    FAMPHUR
    TETRAETHYLPYROPHOSPHATE

-------
                                159
         HERBICIDES
             by
   GAS CHROMATOGRAPHY
ELECTRON CAPTURE DETECTOR
  2,4-DICHLOROPHENOXYACETIC ACID
  2,4,5-TRICHLOROPHENOXYACETIC ACID
  2,4,5-TP (SILVEX ACID)

-------
                                       160
    WATER SOLUBLE COMPOUNDS

   Direct Aqueous Injection GC/MS (15)
ALLYL ALCOHOL
CHLORAL
CHLOROACETALDEHYDE
2,3-DICHLOROPROPANOL
1,4-DIOXANE
ETHYL CYANIDE
ETHYLENE OXIDE
FLOUROACETIC ACID  *
GLYCIDYLALDEHYDE
ISOBUTYL ALCOHOL
METHACRYLONITRILE
METHANETHIOL
N-NITROSOPYRROUDINE
2-PROPYN-1-OL
PYRIDINE

-------
                                                161
    APPENDIX VIII MYSTERY  ORGANICS
                 Purgeable Organics (6)
Acetontrile
1-Chloro-2,3-epoxypropene
1,2:3,4-Diepoxy butane
  Paraldehyde
  Crotonaldehyde
  Methyl hydrazine
                Extractable Organics (22)
5-(AminomethyI)-3-isoxazoloI
Auramine
Dibenzo (aj) pyrene
DlisopropylfSuorophosphate
Thiofanox
a-a-Dimethylphenethylamine
Dimethyl phthalate
Dimethyl sulfate
Formic acid
Hydrazine
Maleic anhydride
Benzenethioi
Melphaian
2-Methyllactonitrile
N-Methyl-N-nitroso-
   quandine
Methylthiouracil
N-Nitrosodiethanolamine
Endothal
n-Propylamine
Saccharin
Thiuram
tris-(2,3-Dibromoproplyl)
    phosphate
                                                    io

-------
                                          162
   PRIORITY POLLUTANT VOLATILE ORGANICS
                 SPIKED BLANK
                       METHOD 624 METHOD 8240
METHYL CHLORIDE
VINYL CHLORIDE
ACRYLONITRILE
DICHLOROBROMOETHANE
CHLOROD1BROMETHANE
TR1CHLOROETHYLENE
1,1,2-TRICHLOROETHANE
BROMOFORM
106
 95
 94
 95
100
 87
1,1,2,2-TETRACHLOROETHANE1Q6

                        N = 19
     STUDY PERIOD 2/24/86 - 3/2/86
                              SD
                   SD
16
8.7
9.9
8.5
9.3
9.2
9.3
12
19


97 28
107 19
93 22
95 6.1
94 9.7
97 5.9
99 10
89 10
108 16
N = 15
11/14/85 - 2/17/86
                Spiked Concentrations
                     18 - 50 ug/I

-------
                                            163
     PRIORITY POLLUTANT VOLATILE ORGANICS
                SPIKED MATRIX

                      METHOD 624
METHYL CHLORIDE
VINYL CHLORIDE
ACRYLONITRILE
DICHLOROBROMOMETHANE
CHLORODIBROMETHANE
TRICHLOROETHYLENE
1,1,2-TRICHLOROETHANE
BROMOFORM
1,1,2,2-TETRACHLOROETHANE
X
103
101
99
101
96
95
102
87
115
SD
18
16
15
16
15
13
12
19
24
                         N = 19
METHOD 8240
  ~x    SD
 106   22
 106   18
 105   57
  99   13
  99   22
 114   35
 111    16
  92   27
 105   31

 N = 12
               Spiked Concentration 18-50 ug/I

-------
                                         164
       APPENDIX VIII VOLATILE ORGANICS
                 METHOD 8240
1.2-DIBROMO-3
   -CHLOROPROPANE
METHLY METHACRYLATE
1,2-DIBROMOMETHANE
1,2-DICHLOROPROPANE
2,3-DICHLOROPROPENE
METHYL ETHYL KETONE
DIBROMOMETHANE
FREON TF
BLANK SPIKE  MATRIX SPIKE
   x  SD       x   SD
   40   24       62  38
99
105
99
97
87
93
68
17
8.8
5.0
4.8
31
24
35
108
110
106
98
92
99
54
22
25
25
22
47
13
19
                          = 15
                 = 12
                          SPIKED CONCENTRATION
                                50 ug/l

-------
                                                 165
   PRIORITY POLLUTANT  EXTRACTABLES
               SPIKED BLANK
Hexachioroethane
Hexachlorobutadiene
2-Chloronapthaiene
N-Nitrosodi-n-propylamine
Butyl benzyl phthalate
2,4-Dinltrotoluene
Bis (2-Chloroethoxy) methane
Pyrene
Benzo (a) anthracene
Chrysene
Phenol
2,4-Dinitrophenol
P-Chloro~M-Cresol
Method 625
X
55
59
83
91
77
89
86
85
85
87
47
78

SD
17
17
18
20
17
17
15
18
14
15
19
25

Appendix VIII
X
57
68
121
90
94
110
100
85
96
95
62
110
75
SD
8.8
9.1
41
18
19
20
20
15
6.8
12
22
21
38
                                n=18
                   n=6
           Study Period
1/29/86 - 2/24/86   7/12/85 - 9/15/85
               Spiked Concentration  100 ug/l

-------
                                              166
    PRIORITY POLLUTANT EXTRACTABLES
                SPIKED MATRIX
Hexachloroethane
Hexachlorobutadiene
2-Chloronapthalene
N-Nitrosodi-n-propylamine
Butyl benzyl phthalate
2,4-Dinitrotoluene
Bis (2-Chloroethoxy) Methane
Pyrene
Benzo (a) anthrocene
Chrysene
Phenol
2,4-Dinitrophenol
P-Chloro-M-Cresol
                            Method 62S  Appendix VIII
                            X    SD   X     SD
62
65
84
94
78
89
89
89
84
86
50
73
19
15
13
21
14
23
12
24
11
15
21
37
48
59
86
83
99
99
83
77
87
88

99
31
27
54
21
17
32
27
32
18
19

32
                              n=18
n=6
         Spiked Concentration  100 ug/l

-------
                                                  167
    APPENDIX VIII EXTRACTABLE  ORGANICS
1,2,3,4-Tetracfolorobenzene
2,3,5,6-TetrachIorophenol
2,4-Dithiobiuret
2-AcetyIaminofluorene
2-sec-Butyl-4,6-dinitrophenol
3-Chloropropionitrile
Benzotrichloride
Benzyl chloride
Diallate
Dibenzo(a,j)acridine
Dichloromethylbenzene
Hexachloropropene
N-Nitroso-N-methylurethane
Pentachioronitrobenzene
Pentachlorophenol
o-Cresol
SPIKED BLANK
If
82
59
84
21
84
>l 67
38
4
66
111
140
36
40
30
114
65
84
SD
46
42
78
28
53
42
22
6
4
18
89
33
28
31
25
54
16
                                          SPIKED MATRIX

                                             ~x    SD
92  66
19  32
79  55
32  19
107
87
37
30
28
103
64
73
48
79
36
25
33
65
48
20
                              N = 5
 N = 6
                                 Spiked Concentration*
                                    100-300 ug/l

-------
                                                    168
                          MR. TELLIARD:  Starting.off



our afternoon session, we have Tina Engel to talk



about pesticides determination in groundwater.

-------
                                                    169
                      TINA ENGEL

             BATTELLE MEMORIAL INSTITUTE
   DEVELOPMENT OF COMPREHENSIVE ANALYTICAL METHODS
  FOR THE DETERMINATION OF PESTICIDES IN GROUNDWATER
                          MRS. ENGEL:  Considerable

attention has been given the problem of contamination

of groundwater resources.  In lieu of a National

Groundwater Survey to be implemented jointly by the

Office of Drinking Water and the Office of Pesticides

Programs, Battelle is conducting a study for the
development of comprehensive analytical methods for

pesticides in groundwater.

     The groundwater analysis methods development

study is funded jointly by the Office of Drinking

Water and the Office of Pesticides Programs through a

contract with the Environmental Measurement Support

Laboratory in Cincinnati, Ohio.
     The purpose of this study is to develop and

validate screening methods for pesticides and pesti-
cides metabolites in groundwater samples.  These
validated methods would be made available for use in

the analysis of groundwater samples collected for the

National Groundwater Survey.

-------
                                                     170
      Specific goals of the methods development study
 include the  ability to detect and quantify targeted
 pesticides and pesticide  metabolites at  sub part per
 billion levels in groundwater.   Many of  the pesticides
 were  included in the scope of the study  because of
 their proven or suspected toxic or carcinogenic
 properties.   In many cases,  pesticide toxicity
 information  was suspected, so low detection limits  in
 groundwater  were desirable as a precautionary  measure.
      During  methods development,  efforts were  con-
 tinuously made to consolidate and simplify analysis
 methods.  The potential scope of  a groundwater analysis
 survey is enormous.   Methods  are  designed to incor-
 porate as many pesticides as  possible, yet are stream-
 lined to simplify and ruggedize the methods.
      Survey  ground  water  samples  will be analyzed by
 different laboratories.   For  this  reason,  efforts
 were  made to ruggedize methods.   The  goal  was  to
 simplify or  ruggedize  the method  in order  to minimize
 variability  of  analysis results between  laboratories.
      After development and preliminary evaluation of
 the analysis methods,  the methods  will be  thoroughly
 validated by determining  analyte method  detection
 limits and method ranges  and by evaluating potential
matrix effects on method  results.  .Validation  of the

-------
                                                    171




groundwater analysis methods are now underway.



     The scope of the groundwater analysis methods



development effort was defined by the Office of



Drinking Water and Office of Pesticides Programs



to include the determination of approximately 90



priority pesticides as chosen by an EPA Analytes



Selection Task Group.  Additionally, inclusion of



approximately 80 non-priority pesticides to the



methods development effort.  Third, development of



methods to look for these approximately 170 pesticides



in groundwater matrix.



     Groundwater samples were not normally expected



to contain large levels or high levels of interfering



organics.  Although a confirmation technique or



techniques were included in methods development



scope, methods development efforts were not designed



to include a cleanup method.  Methods development




studies were conducted using reagent water matrices.



Validation efforts will include the use of groundwater



matrices.



     A set of selected criteria were used by the



Analyte Selection Group to formulate the list of



priority pesticides for the methods development



effort.  First, a list was compiled of known active



ingredients of pesticides currently or recently in

-------
                                                    172
use.  This master list was then limited to those
compounds demonstrating physical properties of interest
to the groundwater survey.  Properties evaluated in-
cluded adequate solubility of the pesticide In water;
adequate stability of the pesticide in water (hydro-
lysis data were considered when they were available);
volatility of the pesticide; distribution of the
pesticide between soil and water; mobility of the
pesticides through soil; and speciation of the pesti-
cide.  Identified pesticide metabolites and decompo-
sition products, especially those displaying un-
desirable toxilogical properties, were included as
priority analytes.
     The list of methods development analytes was
expanded to include approximately 80 non-priority
pesticides.  Criteria used for selection of non-
priority pesticides included any analyte included in
any EPA 600 series method referencing a priority
pesticide.  For instance, aldicarb was designated by
the analyte selection group as a priority pesticide.
Aldicarb is included as an analyte of EPA Method 531,
which is entitled, Measurement of N-Methyl Carbamoyl-
oximes and N-methyl Carbamates in Drinking Water by
Direct Aqueous Injection HPLC with Post Column
Derivatization.  All other Method 531 analytes were

-------
                                                    173




included in the scope of the groundwater survey as



non-priority analytes.



     The non-priority analytes were included in the




scope of the groundwater survey only if initial




methods development studies indicated no difficulties




in analysis or detection of the analyte.  Non-priority




analytes were included in analysis methods designed



for the determination of priority analytes in




groundwater.  Extraordinary efforts were not made to



include non-priority analytes in the survey.  However,



since the non-priority analytes were already included




in one of the EPA 600 series methods, few difficulties




were encountered with non-priority analytes during



the methods development study.




     Groundwater analysis methods were based on the



EPA 600 series methods.  Methods incorporated extraction



and analysis techniques used in the 600 series methods




including extraction of analytes from groundwater



using separatory funnels; concentration of resultant



organic extracts using Kuderna-Danish equipment and




techniques; and finally, analysis of concentrated



sample extracts by packed column gas chromatography



using electron capture or nitrogen-phosphorus detectors,



or by high performance liquid chromatography using



ultraviolet or post-column derivatization detectors.

-------
                                                    174
     EMSL 600 series methods used as a basis for
groundwater analysis methods development studies
included Methods 531, 608, 614, 615, 619, 622, 632,
633, 634, 643 and 645.
     Five analysis methods were developed for the
determination of the 170 priority and non-priority
analytes in groundwater.  The first method incorporates
separatory funnel extraction of the neutralized water
sample with triplicate 60 mL aliquots of methylene
chloride.  The organic extract is dried and concentrated
using Kuderna-Danish techniques.  The resultant
concentrated extract is analyzed by packed column gas
chromatography using a nitrogen-phosphorus detector.
     Method 1 can be used to determine approximately
75 nitrogen- or phosphorus-containing pesticides and
pesticide metabolites in groundwater.  Method 2 uses
the same extraction and concentration techniques as
Method 1.  The resultant extract is analyzed by packed
column gas chromatography using an electron capture
detector.  Method 2 can be used to determine
approximately 28 halogen-containing pesticides and
pesticide metabolites in groundwater.
     The extraction and analysis procedures used in
Analysis Method 3 are based on EPA Method 615.   Acids
and acid esters are extracted from the acidified water

-------
                                                    175
sample using separatory funnel extraction with 360 mL
aliquots of ethyl ether.  Acid esters are converted
to the parent acid by base hydrolysis.  The acids are
reextracted into ethyl ether and the extract is
concentrated using Kuderna-Danish techniques.
     After methylation with diazomethane, the
resultant extract is analyzed by packed column gas
chromatography using an electron capture detector.
Method 3 has been used to determine approximately 26
halogen-containing acidic pesticides and pesticide
metabolites in groundwater.
     In order to determine the analytes covered by
these first three GC methods it is necessary to use
three different packed columns; a non-polar column
packed with 1.5 percent OV-17/1.95 percent QF-1; a
moderately polar column packed with 3 percent SP-
2250; and a polar column packed with 5 percent Carbo-
wax 20M-TPA.
     The fourth method  incorporates separatory funnel
extraction of the neutralized water sample with
triplicate 60 mL aliquots of methylene chloride.  The
organic extract  is dried and concentrated using a
rotating evaporator.  The resultant extract  is analyzed
by reverse phase high performance liquid chromatography
using an ultraviolet detector.  Method 4 can be used

-------
                                                     176
 to determine  approximately  21  pesticides  and  pesticide
 metabolites in groundwater.
     Method 5 is a  direct injection HPLC  procedure
 based on EPA  Method 531.  The  aqueous sample  is
 injected directly onto a reverse phase HPLC column.
 Analytes are  hydrolyzed with 0.5 N sodium hydroxide
 at an elevated temperature  after elution  from the
 column.  The methylamine formed during hydrolysis is
 reacted with  2-mercaptoethanol and ortho-phthalaldehyde
 to form a highly fluorescent derivative which is
 detected using a fluorescence  detector.   Method 5 can
 be used to determine approximately 12 N-methyl
 carbamoyloximes and N-methyl carbamate pesticides and
 pesticide metabolites in groundwater.
     After initial methods development studies, two
 major modifications were made  in the analysis methods.
 The first modification was the substitution of a
 tumbling liquid/liguid extraction procedure for the
 sequential separatory funnel partitioning used in
most of the water analysis methods.  The  second
modification was the use of capillary columns instead
of packed columns in methods using gas chromatography
 for analyte identification and quantification.
     The tumbling extraction technique involves
tumbling the one liter aqueous sample with 300 mL of

-------
                                                    177


organic solvent end over end for one hour.  The sample

and extraction solvent are placed in a sealed 1.7 L

culture bottle and a mechanical device is used to


turn the bottle reproducibly for a predetermined

length of time.


     Substitution of the tumbling extraction technique

for the separatory funnel extraction technique was


based on mathematical calculations and later demon-

strated with real water samples.  The partition
                                                      t

coefficient of an analyte between water and organic

solvent can be easily calculated.  The aqueous sample


is equilibrated with solvent and the concentration of

the analyte in the organic phase can be determined.

     The analyte coefficient, Kdf can be calculated

from the original amount of the analyte in the aqueous

sample, Ao, the amount of the analyte in the solvent

after partitioning, As, and the volumes of the two


phases, Vw and Vs.  Calculation of Kd is simplified

when the equation is put in terms of analyte recovery

into the organic phase, R.  The partition coefficient

can then be calculated in terms of percent recovery

and phase volumes.  The assumption is made throughout

these calculations that analyte equilibrium between

the aqueous and organic phases is reached prior to

any recovery measurements.

-------
                                                    178
     The expected analyte percent recovery from a
single partitioning can be calculated from the
partition coefficient and the phase volume ratio, Vr,
which is the ratio of the volume of the organic phase
to the volume of the aqueous phase.
     The expected analyte percent recovery from
triplicate equilibrations with equal volumes of
organic solvent can just as easily be calculated from
the analyte partition coefficient and the volume ratio
from each equilibration.
     Using the prior recovery equations, expected or
ideal recoveries can be calculated for partition
coefficient values representing a wide range of
analytes.  This table contains calculated percent
recoveries for a fictitious group of analytes with
partition coefficient values varying from 1 to 80.
Calculations were made for a one liter aqueous sample,
assuming either triplicate separatory funnel extrac-
tions with 60 mL each of an organic solvent, or a
single tumbling partitioning with 300 mL of the
identical organic solvent.
     Lower partition coefficient values indicate an
analyte which is more water soluble and thus harder
to extract from the water into the organic solvent.
As expected, calculated partition recoveries are

-------
                                                    179



lower for analytes having lower partition coefficients.



However, in no case does the calculated recovery



using a single tumbling partition vary from the



separatory funnel results by more than seven percent.



     The experimental comparison of the extraction



techniques was conducted with many pesticides.  Some



results are demonstrated in this table.  These data



were generated from reagent water samples spiked with



the listed pesticides at the low parts per billion



level.  Methylene chloride was used as the organic



phase.  In most cases, comparable or superior recoveries




were observed for most compounds when the tumbling



extraction procedure was used.  In most cases,



reproducibility of the measurements, expressed as the



percent relative standard deviation, was greatly



reduced.  In other words, the procedure was more



repeatable.



     The use of tumbling extraction provides several



advantages over the use of separatory funnels.



Tumbling is less time and labor consuming.  Sample



processing is limited only by the equipment avail-



ability, specifically the number of available tumblers.



Use of separatory funnels requires a person to



physically shake the apparatus for the entire



equilibration period.

-------
                                                    180
     Tumbling results in increased reproducibility.
The bottle motion, and thus the mixing of the two
phases, is mechanically controlled.  Partitioning in
a separatory funnel is inherently dependent on the
operator.  Last, tumbling results in an increased
method ruggedness evidenced by lower variability of
method results.
     The second major methods modification was the
substitution of capillary column GC for the packed
column techniques originally evaluated.  A 30 meter
by 25 millimeter ID SPB-5 fused silica capillary
column was used instead of the three packed columns
originally needed for the three gas chromatography
analysis methods.
     Several advantages were obtained from using the
capillary GC column.  The modified methods were more
sensitive for the method analytes.  Use of capillary
columns resulted in sharper peaks leading to lower
analyte detection limits.
     All analytes could be determined using only one
column.  Using packed column would require analysis
of all sample extracts on all three packed columns.
Alternatively, one injection on a single column is
required when the capillary columns are used.
     Finally, capillary columns demonstrate superior

-------
                                                    181
resolution properties.  Lower analyte coelutions
allow for a more effective screening of the sample
extracts.
     The advantages of using capillary columns are
demonstrated in the next two slides.  This first
chromatogram was generated during the analysis of 13
nitrogen- and phosphorus-containing pesticides on a
non-polar packed column using a nitrogen-phosphorus
detector.  Addition of other pesticides to the mixture
would most likely cause coelution problems.
     This second chromatogram was generated during
the analysis of 34 pesticides on an SPB-5 fused silica
column.  Although this mixture contains approximately
three times as many pesticides as that shown in the
previous chromatogram, the capillary column gas
chromatogram is much less complicated and crowded
than the packed column gas chromatogram.
     In summary, five analysis methods have been
developed for the determination of approximately 170
pesticides in groundwater.  These methods are suitable
for the screening and quantification of trace
levels of these analytes.  The methods, originally
based on EPA 600 series methods, have been modified
to include tumbling extraction techniques and capillary
column gas chromatography where applicable.  These

-------
                                                    182
modifications have resulted in improved method
repeatability and ruggedness.  Further, method
ruggedization and methods validation studies are
ongoing.
     Thank you.

-------
                                                    183
             QUESTION AND ANSWER SESSION
                          MR. TELLIARD:  Questions?



                          MR. TUROSKI:  Victor




Turoski, James River Corporation.  Did you notice any



difference in limited detection when you went from




packed columns to capillary columns, since you can



obviously inject 20 microliters on a packed column




and maybe one on capillary?



                          MRS. ENGEL:  With the



detectors that we were looking at as far as the ECDf




no, because the detector really defines the linear



range that you have, and we found that you can only



really do about one and a half orders of magnitude




anyway with an BCD, which is what most of our work



was done with.



                          MR. TUROSKI:  What level




spikes were these?



                          MRS. ENGEL:  The work that



we did was usually 1:10 part per billion in the water




sample.  We are actually going to be validating



the method at sub part per billion.



                          MR. TUROSKI:  My concern  is



that you can see a lot less.  Your limited detection



should theoretically be less or your limited

-------
                                                     184



detection should be theoretically more on the packed



column.  If I were going for  limited detection, I'd



certainly use a packed column.



                          MRS. ENGEL:  But these are



screening methods and they're trying to look for



everything in one run.  They're trying to really



limit the costs of the analyses/ and if that's the



case, I don't see how you can use three different



packed columns and save any money.



                          MR. TUROSKI:  I understand.



                          MR. TELLIARD:  We just



don't want these labs to get  fat on us, you know what



I mean?  They're already just rolling.



                          MR. CALDWELL:  I'm Chan



Caldwell, International Technology.  Have you tried a



wider bore column, and do you expect that you'd get the



same type of results?  Certainly, in regard to this



gentleman, you would be able  to maintain higher



capacities in such a wider bore, say a .3 or .5 milli-



meter ID column.



                          MRS. ENGEL:  The wide bore



column certainly would give you some kind of inter-



mediate results between using the packed and the



narrower bore capillary.  No, we really haven't



looked into it.  I suppose it depends on what's more

-------
                                                    185




important, the higher resolution capabilities or




being able to inject more sample and have higher



capacities.




                          MR. MOSESMAN:  Neil




Mosesman, Supelco Incorporated.  Did you do




any optimization of the amount of solvent in



the extraction technique?  I see you're using




almost twice as much solvent in the tumbling technique



versus liquid, you know, separatory funnel extraction.




                          MRS. ENGEL:  We found that



when you do the mathematical calculations, that the




less solvent you use for the single partitioning or



equilibration, obviously the lower your recoveries are




going to be; 300 mL seemed to pretty closely mimic



what you could do with triplicate extractions with




smaller volumes of solvent.



                          MR. MOSESMAN:  But that




also means you've got twice as much solvent to



concentrate down.



                          MRS. ENGEL:  Yes.  That




didn't seem to pose any problems as far as using



Kd concentration techniques.  If we used more than



300 mL we really didn't get any added recovery




advantages, and I suppose it would be harder to



handle, Kd concentration-wise.

-------
                                                     186
                          MR. APRIL:  You had  two
HPLC methods at the end.  Are there any general
observations you can make about  the HPLC methods as
contrasted with the GC methods?
                          MRS. ENGEL:  The HPLC
methods were limited to those compounds that we could
not do by GC, the compounds that could be done by
Method 531, which is the post-column derivatization
detection technique, we left in  that method because
it is highly specific for those  compounds and  it gave
good detection limits.
     All compounds in Method 4 were screened by GC
and could not be done by GC.  We were instructed in
this study that gas chromatography was the preferred
method simply because the detectors that were available
for use with the GC were more specific.
                          MR. APRIL:  Can you comment
on the results you got from HPLC as contrasted with
the results for GC in terms of detection limits and
interferences?
                          MRS. ENGEL:  Detection
limits for the PCD procedure were generally in the
low part per billion level.  They did not get into
the sub part per billion level in water.  The UV
methods we could go very low.  In fact, in some cases

-------
                                                    187
you can go lower than the GC methods.  But it probably
will be more likely to have interference problems
when we actually start looking at real samples.
                          MR. WHITLOCK:  Stu Whitlock
from ESE.  Did you try any cleanup techniques on
these things?
                          MRS. ENGEL:  No, we didn't
try any cleanup techniques at all.  That was not part
of the study.  The general tack that's been taken
with these, or will be taken with these, samples is
that interferences are not really expected in most of
the samples, and there is going to be a backup analysis
technique or a confirmation technique.  In other
words, either using mass spectrometry for confirmation
purposes or another column.
                          DR. GAIND:  Arun Gaind from
Nanco.  Liquid solid extraction procedures are up and
coming thing in preparatory chemistry.  Are you
planning to try that in the future or no?
                          MRS. ENGEL:  Would you like
to fund the study?
                          DR. GAIND:  No, I am not
U.S. EPA, thank God.
                          MRS. ENGEL:  Anybody?  I
would love to, I would love to.  Definitely.  It has

-------
                                                    188
a lot of potential advantages if it can be demonstrated.
                          MR. SLOAN:  You said you
used methylene chloride as extracting solvent.  From
the BCD methods I know you had to change to something.
Was it hexane or iso-octane, and did you see any
difference between the two of them?
                          MRS. ENGEL:  We used methyl
di-butyl ether, and we found generally that
gives you can do a solvent substitution with it
and it gives better recoveries because you don't have
to go to as high a temperature in the Kd, to use the
Kd concentration.  And that's usually the solvent we
wind up in before we do any mass procedure.
                          MR. SLOAN:  What about the
effects on the people who work with it?
                          MRS. ENGEL:  We use hoods.
                          MR. TELLIARD:  Thank you,
Tina.  Thank you so much.
                          MRS. ENGEL:  Thank you.

-------
                                          I8y
DEVELOPMENT OF COMPREHENSIVE
   ANALYTICAL  METHOD FOR
 PESTICIDES IN GROUND WATER
             BY
 T.M, ENGEL AND J.S. WARNER

-------
                                             190
         STUDY SUPPORTED

               BY
  THE OFFICE OF DRINKING WATER
               AND
THE OFFICE OF PESTICIDES PROGRAMS

-------
                                                       191
                    STUDY PURPOSE
DEVELOP AND VALIDATE SCREENING METHODS FOR PESTICIDES
 IN GROUND WATER IN SUPPORT OF ODW'S NATIONAL GROUND
                    WATER SURVEY

-------
                                                    192
                 PROGRAM GOALS
0    DETECT  AND QUANTIFY  PESTICIDES  AT SUB  PARTS-
     PER-BILLION LEVELS IN GROUND WATER

0    CONSOLIDATE AND SIMPLIFY METHODS

0    RUGGEDIZE METHODS

0    VALIDATE METHODS

-------
                                                    193
                  STUDY SCOPE
0    DETERMINATION  OF  APPROXIMATELY  90  "PRIORITY"
     PESTICIDES

0    INCLUSION  OF   APPROXIMATELY  80  "NONPRIORITY"
     PESTICIDES

0    GROUND MATER MATRIX

-------
                                                    194
              PRIORITY PESTICIDE
              SELECTION  CRITERIA
0    KNOWN ACTIVE INGREDIENTS
0    PHYSICAL PROPERTIES
          WATER SOLUBILITY
          STABILITY (HYDROLYSIS)
          VOLATILITY
          SOIL/WATER DISTRIBUTION
          MOBILITY
          SPECIATION

-------
                                                    195
             NONPRIORITY PESTICIDE
              SELECTION CRITERIA
0    ANY ANALYTE INCLUDED  IN ANY EMSL-EPA 600-METHOD
     REFERENCING A PRIORITY PESTICIDE

0    EASY INCLUSION IN A DEVELOPED METHOD

-------
                                                     196
              METHODS DESCRIPTION
0    BASED ON EMSL/EPA 600-METHOD SERIES
          SEPARATORY FUNNEL EXTRACTIONS
          CONCENTRATION      USING      KUDERNA-DANISH
          EQUIPMENT

          ANALYSIS BY  PACKED COLUMN GC  (ECD  OR NPD)
          OR BY REVERSE PHASE HPLC (UV OR PCD)

-------
                                                                 197
      EMSL600-SERIES METHODS USED
         FOR PRIORITY ANALYTES
531,608, (608,1,608,2), 614, (614,1), 615,619,622
  (622,1), 632 (632,1), 633 (633,1), 634,643,645

-------
                                                          198
        GROUND WATER ANALYSIS METHODS
METHOD
 NO.
 1
 2
 3
 4
 5
 SAMPLE PREP CONDITIONS
SEPARATORY FUNNEL/NEUTRAL PH
SEPARATORY FUNNEL/NEUTRAL PH
SEPARATORY FUNNEL/ACID PH/
HYDROLYSIS/METHYLATION
(METHOD 615)
SEPARATORY FUNNEI7NEUTRAL PH
NONE
 ANALYSIS CONDITIONS
PACKED COLUMN GC-NPD*
PACKED COLUMN GC-ECD*
PACKED COLUMN GC-ECD*

HPLC-UV
DIRECT INJECTION HPLC-PCD
•3 PACKED COLUMNS USED

-------
                                                    199
               METHODS MODIFICATIONS
0    TUMBLING EXTRACTION
0    CAPILLARY COLUMN GC

-------
                                                        200
             TUMBLING EXTRACTION
TUMBLE 1-L SAMPLE AND 300 ML SOLVENT END-OVER-END
                 FOR 1 HOUR

-------
                                                     201
             CALCULATION OF PARTITION

                 COEFFICIENT  (Ko)
KD - Cs - As x Vw
     Cw   Aw x Vs
                   As x Vw
               (Ao - As) x Vs
R - 100_x As
       flo
KD =
     	VM

Vs x [(100/R) - 1]

-------
                CJJLAT.
                 MRIES
R = 100 x
(VR x KD)
            TVRXKD) + 1
                                                      202

-------
                                                      203

R = 100 x
 1-    (fv    1   .    .  V
        I [KD X VRj  + I  )
&                          ซ

-------
MATHEMATICAL COMPARISON
                                       204
        % RECOVERY
KD
1
5
10
20
40
80
SEPARATORY FUNNEL
16
54
76
91
97
99
TUMBLE
23
60
75
86
92
96

-------
                                                          205
                   EXPERIMENTAL COMPARISON
ANALYTE
AMETRYN
DISULFOTON SULFONE
METHYL PARATHION
PRONAMIDE
ALPHA-CHLORDANE
DIELDRIN
METHOXYCHLOR
% RECOVERY
StKAKAIUKY hUNNtL
88 ฑ 17
54 ฑ25
110 ฑ14
89 ฑ 15
68 ฑ 1
84 ฑ2
111 ฑ 11
(ฑ2 RSD)
I UMttLtk
97 ฑ 5
99 ฑ 1
99 ฑ8
82 ฑ6
80 ฑ 4
82 ฑ3
97 ฑ4

-------
                                                     206
                ADVANTAGES OF USING
                TUMBLING EXTRACTION
0    LESS TIME-CONSUMING
0    INCREASED REPRODUCIBILITY
0    INCREASED RUGGEDNESS

-------
                                                     207
                CAPILLARY COLUMN GC
SUBSTITUTION OF A 30 M x 25 MM ID SPB-5 FUSED SILICA
              COLUMN FOR PACKED COLUMNS

-------
                                                     208
                ADVANTAGES OF USING
                CAPILLARY  GC  COLUMNS
0    MORE SENSITIVE
0    ONLY ONE COLUMN REQUIRED FOR ALL ANALYTES
0    SUPERIOR RESOLUTION CAPABILITIES

-------
209

-------
210

-------
209

-------
210

-------
                                                     211
                      SUMMARY
0    FIVE    ANALYSIS    METHODS   DEVELOPED    FOR   THE
     DETERMINATION OF 170 PESTICIDES IN GROUND WATER

0    TUMBLING  EXTRACTION  AND  USE  OF  CAPILLARY  COLUMNS
     FOUND TO GIVE SUPERIOR RESULTS
0    FURTHER    METHOD    RUGGEDIZATION    AND    METHODS
     VALIDATION ONGOING

-------
                                                    212
                          MR. TELLIARD:  Before our next
speaker I have a couple of cleanup things that I'm
supposed to do.  Is Jim King here?  This morning's
speakers, those folks who have transparencies, if
you'll get them to Dr. King, he'll take them
downstairs and run a copy of them and give them
back to you before you leave today, Jim is in
the back corner, so that we have them for the
transcript.  Did I do right?  Good.  Thank you.
     Our next speaker is Ted Handel from Centec.
He's going to talk about some very, very expensive
data.  I paid for it.  I've never seen a platinum
ICP before but they said that's what it costs to
run them."  Ted's going to talk about some metals
analysis.

-------
                                                    213
                    EDWARD HANDEL




              CENTEC ANALYTICAL SERVICES









  RAPID MULTI-ELEMENT SCREENING USING SEQUENTIAL ICP
                          DR. HANDEL:  Thanks, Bill.



Basically, what I'd like to talk to you about today is



some preliminary work we've done with our new ICP.



Basically, the question that we're going to try and



answer today is what is SuperScan.  If you take a



look in all your manuals for your various ICPs, you



won't find the term SuperScan, so don't bother.  It's



something that we've just coined for lack of a three



letter word to call it by.



     Basically, SuperScan is a way of screening an



environmental sample for the presence of up to 73



different elements.  Since we chemists are fairly



curious types, we're always looking to see what's



actually in those samples.  Take Bill, for example.



The first thing he said when he heard about SuperScan



is, a little smile came to his face, "You can measure



all that, huh?" The second thing, he says, "What's



this?  Why can't we get that?"  And he says, "If we



can get it, let's find a way."  Well, we didn't, but



we might still work at it.  That element, by the way,

-------
                                                    214
is technitium.
SLIDE 1
     The 73 elements shown on this periodic table can
all be analyzed by an ICP and can be run on a sample
in a SuperScan analysis mode.  As you can see, the
ICP analysis is able to identify the presence of most
of the transition metals and most of the alkali
metals.  The rubidium and cesium, however, it should
be noted, suffer fairly high detection limits and may
not be useful in terms of a SuperScan.
     As a laboratory, we often find people that come
to our facility and they're bringing us a sample,
maybe something like this Tina had, and said, can you
tell me what's in it.  Well, when we had AA and we
were running our metals analysis by AA, we went and
counted our lamps, told them we could run this and it
was going to cost that.
     Now, we take a look at our chart and we say,
well, we can run up to 73 different things using our
new ICP.  If we were to run those four in quantitation,
it could cost guite a bit of money.  Yet, on the
other hand, we can run this in a scanning mode or a
screening mode and save the client an awful lot of
money and still get some useful information as to
what the composition or the contaminates in that

-------
                                                     215



sample might be.



     The SuperScan screening procedure is based on



the ability of high resolution sequential ICP



instruments to run such a wide range of elements



quickly and with relatively high selectivity.  I



don't want to spend a lot of time comparing the



relative benefits of sequential versus simultaneous



ICP analysis, for we all know that they both offer



their own benefits.  What I would like to do is point



out that the screening function offered by sequential



ICP is something that we can use in a valuable way.



SLIDE 2



     We purchased a high resolution duo-monochrometer



sequential ICP earlier this year.  We use .this



instrument heavily for the analysis of 19 of the 24



elements required for the inorganic Contract Lab



program.  Our chemists found the power and the speed



of their new toy to be fairly irresistible and they



experimented on the prospect of running all 73 elements



in that periodic table shown previously.



     In many cases, as they looked for standards and



stuff to compare, or to run on the machines, they



went through our chemical stockroom and they went



through the stockrooms of the various universities



around, and they were fairly unsuccessful in finding

-------
                                                    216
a lot of those elements.  We've since sent out orders
and we are receiving standards for the various elements
so that we can quantitate the detection limits on our
particular instrument and see how useful the information
is going to actually turn out to be.
     Our initial idea was to run a standard solution
or two in order to obtain a semi-quantitative result.
If we had continued this thinking we would have
probably never gotten anywhere.  It would probably
take a dozen or more standards containing compatible
groupings of compounds of the various 73 elements in
order to even begin to calibrate a run.  It could
take years just to figure out the chemistry and
probably several more years to convince anybody that
we had figured out the chemistry.
SLIDE 3
     The worst part is that when we look at the
periodic table here and look at those elements that
we can screen in the SuperScan mode, the chances of
finding some of the odder ones are pretty slim.  I
know I always am asking for trouble when I say you'll
never find some of those things in a sample, but the
odds are pretty much against you.  So if you're going
to spend a lot of time and money trying to quantitate
a result before you even have a suspicion that the

-------
                                                    217
element is in your sample, you're defeating your
benefit of screening.  So, we'd like to promote this
as a means to pre-screen a sample or to screen a
sample that's being run for other elements as well,
but to screen it to find some information about what
other things might be in it, and I'd like to show you
some data later on to point out how that might be of
benefit to us.
     The bottom line question that one would have to
ask himself is how would you like to be able to screen
a water sample for the presence of 73 elements in
less than five minutes and for less than, let's say,
50, 60 cents an element.  That's exactly what we're
talking about here.
     I'd like to briefly now discuss how the SuperScan
works and then end up with a few quick examples of
SuperScan runs that we've made.
     The elements shown in the periodic table emit
intense characteristic light wave lengths in the Uv
and visible part of the spectrum when they're subjected
to the high temperatures of plasma.  The intensity of
the emission is proportional to the amount of the
specific element in the sample.   In essence, all one
has to do is to carefully select 73 wave lengths which,
on the one hand have sufficient sensitivity to yield

-------
                                                    218
a useful detection limit, and on the other hand, are
not subject to significant interferences from nearby
emission lines of other elements.
     Our ICP has a very powerful software application
package.  It contains a data base of several thousands
of analytical emission wave lengths from which we can
choose our method file for the SuperScan run.  In
addition to that, we can set our windows, our viewing
windows, for each particular emission line quite
narrow, and this excludes the possibility of mis-
identifying peaks of nearby emission lines.
     Our instrument is a dual monochrometer system,
as I mentioned before.  One high resolution mono-
chrometer has a spectral resolution of less than
100th of a nanometer, and the other monochrometer
has a little lower resolution of less than 200ths of
a nanometer.  In creating a method file, one can use
the low resolution monochrometer for the easy elements
and set the high resolution monochrometer to measure
the more difficult ones.  The other nice feature is
that when you're running these analyses you want to
make sure that you've got sufficient speed because
speed is where you're providing yourself with the
profit.
     So each monochrometer,  when you're running a dual

-------
                                                    219
system, is controlled independently of each other so
that the second monochrometer is getting in position
while the first one is making the measurement on the
previous element.
     Another factor one has to control is the viewing
height.  There's an optimal region in the plasma for
each element where the signal to noise ratio is
maximized.  This factor can be preset for each element
when you're creating your method file, and during a
run then, the computer, under the control of your
method file for the SuperScan, rotates the grading to
a specific wavelength at a speed of 50 or so nanometers
per second, stops within a hundredth of a nanometer
on the wave length, with a very narrow window, makes
a measurement, and during that measurement time the
second monochrometer would be adjusting for the second
element out of the 73, and the process would continue.
Over a period of four or five minutes you'd have a
SuperScan run.
     The fee of SuperScan analysis is one of the
critical factors to keep in mind.  The ICP measurement
for a laboratory can make a laboratory a lot of money,
but not if it's tied up for half an hour or so making
a SuperScan measurement.  You can't afford to run a
screening test that costs hundreds of dollars.  So the

-------
                                                    220
screening test would have to be considerably less to
be of value to the client.
SLIDE 4
     Right now, I'd like to take a look at some of
our data that we've generated on our SuperScan runs.
This particular slide here is shown in graphical
forms in the following slides so you don't have to
worry about the numbers here.  What I'd like to point
out though, is that these elements here are some of
the 73 that were measured in a SuperScan run.  The
other ones were unimportant because they didn't show
up in the particular samples at all.  If you looked
at the periodic table, all of the elements on the
periodic table that were indicated in the first slide
were run.
     This sample here is a groundwater sample from
around a hazardous waste site, or is from a batch
of groundwater samples from around a hazardous waste
site.  We picked that as one of our samples and we
picked this from a different batch that's also a
groundwater sample from a hazardous waste site.  As
it turned out, we pulled this off the shelf as I was
leaving to get ready to go to this meeting here.  We
had these run and we had three samples run; this one,
this one and this one, in the SuperScan mode within

-------
                                                    221



20 minutes.



     These values here are actual ICP quantitative



results in this column here.  Some of the measurements,



like for arsenic and so forth, are run by furnace AA



for comparison.  Mercury, of course, is run by cold



vapor.  In this particular sample, we inadvertently



pulled it out and it turned out to probably be a



field blank, so it's kind of interesting to pick this



sample because as we look at the SuperScan result,



which is a ratio of the signal of the sample to a



digested blank...so a ratio of 1 says that there's



essentially no detection in the SuperScan mode...we



see that all these values here are essentially 1,



indicating that that is acting like a blank sample.



As we take a look at another sample that is an



actual sample that has something in it, we see that



that's not the case.



SLIDE 5



     Now, I'd like to go to the next slide, which is



graphical, and we'll show this a little clearer.



Now, this is the second sample that we SuperScanned



and this is a signal over background of about 1,100,



and that's for calcium.  I'll point out what elements



these are since they didn't come out very clearly.



Calcium came out very high.  Sodium was about 400 times

-------
                                                    222
background.  Strontium was about 700 times background
in the signal.
     Now, I'll point out also that these elements
here, all the way up to zinc, which is right here,
are normal elements that we run in the Contract Lab
Program.  The rest of them were additional elements
that showed something in this particular sample.
SLIDE 6
     This is the same slide where we've just clipped
the top off and expanded from 0 to 100 instead of 0
to 1,000.  It's really unimportant.  This is still
1,100, this is still 700, and so forth.  Values around
1 indicate that the sample showed no evidence of
that element being run in the SuperScan mode, however,
when we see aluminum here, we've got about two or
three times background.  The same for barium.  And in
the case of iron, we see we've got about 30 times the
background signal.
     When we get down to the lower end of this group
of metals here, or elements, we find that sulfur
shows up, erbium shows up at about three or four
times background.  Cerium, strontium, boron, cesium
and silicon and lithium show up.  Iodine and phos-
phorus didn't show up, but they'll show up in the last
sample, so I left them all on for comparison purposes.

-------
                                                     223
     Now, if we would have run this sample, as we did
run it for a client...when we ran this particular
sample, we only measured the first 24 elements, and
we used normal ICP and furnace methods for that and
we have quantitative results.  In each case, the
quantitative result  indicates that there was either
a high amount of material or, in a case where none
was detected, there was a good relationship between
the SuperScan being an indicator for these 24 elements
in this particular case.
     However, the client didn't know anything about
the last group of materials that we identified.  Some
of them may not be interesting.  But if the client
can see this picture, then the client can make a
decision.  He may say, what the heck is strontium
doing in there.  Strontium tends to show up in more
and more lists these days.  Strontium is not a normal
element that we'd run for the Contract Lab Program,
for example.
     Osmium might be another one that we would iden-
tify in terms of a SuperScan mode where it wouldn't
be normally asked for.  But a client can come back
now and take that same sample and request a quanti-
tative result for that particular element that showed
up in a SuperScan.

-------
                                                    224
SLIDE 7
     This slide is of a sample...I like to do things
in threes, so instead of running another hazardous
waste site sample we ran a cup of coffee.  Same 24
elements here.  We see that we get the normal hits;
we get a hit on calcium, we get a hit on magnesium
and manganese.  We also get a hit on potassium and
sodium, and a little bit of barium in there.  This
is not a quantitative result, it's just two times the
background level of a digested blank, or a blank
water sample in this particular case.
     We see there's sulfur in there at a level about
three times the background level of sulfur.  And we
see strontium again, so I can't tell you what kind of
coffee it is.  We've got silicon.  I can tell you if
you put cream in your coffee, you'll probably find
more of this stuff in there, but besides that, we
also found the iodine and the phosphorus in this.
All the other 73 elements were not found in that
particular sample.
     So, the question really becomes what kind of
usefulness can the SuperScan technique provide to us,
either the user or the laboratory.  And it has benefits
for both of us.  In the case of the laboratory, we've
got a number of different benefits that we can achieve

-------
                                                    225
by running a SuperScan.  One of the problems of
running emission spectroscopy is the fact that you
can have interferences from high amounts of different
elements in that sample.  Those interferences need to
be corrected for because you'll get a high value, for
example, for antimony and tin, if you've got a lot of
iron and calcium in there.
     Well, if you're running a hazardous waste sample
for antimony and tin, chances are you're not going to
normally run iron and calcium, and so the only way you
can determine that you've got an interference from
high amounts of iron and calcium is by doing spiking.
So that would be one thing.
     The second usefulness is when you get a client
that comes in your door, and every lab's got these.
Once a month or so we get somebody coming in and
saying, what can you tell me about this water, I
think I'm dying.  You don't want to sit there and
tell the poor person that you're not going to measure
it because they'll die when they get the bill, you
know.  So, it's nice to be able to have a little
method that you can say, for $50 or $40 or whatever
you're going to charge for a SuperScan, you can sit
there...
                          AUDIENCE PARTICIPANT:  You're

-------
                                                    226
undercharging.
                          MR. HANDEL:  Oh, okay, $100
or $200, we'll tell Bill that his water's all right,
or whoever.
     The other advantages are your end clients.  If
you've got a client that wants to know a certain
number of things for sure but is also out on a little
bit of a fishing expedition...maybe it's his own
backyard and he wants to see what he's got back
there...you can run a SuperScan in addition to the
things that he suspects to be back there that you
would normally run quantitative anyhow, as we did in
the first two examples.
     We're proceeding.  We haven't gotten real far on
this project yet.  We're proceeding on an as we can
basis.  It is interesting for us to do it, and we'd
be glad to talk to anybody about it if they're
interested in doing it on their own machines.
Thank you.

-------
                                                    227
             QUESTION AND ANSWER SESSION
                          MR. BIRRI:  John Birri,




EPA.  Have you thought about using this as a screening




technigue, possibly for TOC and sulfate?



                          MR. HANDEL:  Basically...




yes, I guess we could.  We have in fact done that a



couple of times where we had a high carbon answer.




The detection limit on carbon is not the greatest in



the world, but then when you're running a TOC anyhow,



you're talking about part per million level.  We have




correlated with our TOC runs that we've run for the



client after...for example, oil in his groundwater or



gasoline in his groundwater, we've actually picked



it up running this way.



                          MR. BIRRI:  At what level




do you think you could pick something up?



                          MR. HANDEL:  I think we



were talking in that particular case about 10 ppm.



                          MR. BIRRI:  That's not too



bad.  What about your detection levels in general?



What were you coming up with?



                          MR. HANDEL:  Okay, the




detection limits are going to be element-specific



on this.  They're going to also depend on what

-------
                                                    228



spectral lines you choose.  You're going to have



different response factors for those lines.  But in



general, a good ballpark range would be 50 to maybe



100 ppb.



     You can improve on that if you want to, but we



want to make sure that we're keeping in the spirit of



a screening technique where we're not going to make



it cost a fortune to run it by increasing the amount



of run times or running two or three element lines



instead of just a single one.  There's an awful lot



of things you can do to make it fancy, but it may



lose its value along the way.



                          MR. TELLIARD:  As he pointed



out, it isn't very difficult, it's just expensive.



                          MR. BIRRI:  My last question



is was this all done in vacuum?



                          MR. HANDEL:  Yes.



                          MR. BIRRI:  Thank you.



                          MR. TELLIARD:  Thanks, Ted.

-------
                                                 229
                   cu
 cd
 o
CO
 S-H
 cu
O)
                   o
                   O)

-------
                                                          230
CO
     CO
          ex:

-------
                                                                                                    231
                                                                         oo      CD
                                                                    o tป o      i-i o
                                                                    •*ป< CNI ป-i o   ^^ oo
                                                                                m CD
                                                                                c- co
                                                                     O
                                                                    •*-ป

                                                                    CO
                                                                     O O> i-< __ ป-< ..^ ป-H
                                                                     f>	ซ 	^0 ~~
                                                                             t CO CO 00


                                                                               I ป-H ID
                                                                    —.. O3 _,. _ rH ฃ• CO
                                                                    ^^   CO CO   r*



                                                                    C O
                                                                    i-( Ift CO CO CO i-H CD
o
     feซ

   -It

gill
= ฃ0-5

     J<
  ^ in 1-1
  ~c-'?1
  OS W '" !
  Cu Cu
                                                                                 cv

                                                                                 o
    CD


     E


    t-
    i-(

    •D
     C
     CO

     3
     o

                                                                                             u
                                                                                            E
     -I
     •*-* .c
     "O U

     s "ง
        CO

        E
                                                                     5
                                                                                  ; S>
o
53
a.
O
                                                                                            -a •-
                                                                                            -oo
                                                                                            < *•'
                                                                     < CQ O Q W b O
                                                                                                       O

                                                                                                       I
                                                                                                       o
                                                                                                      f—ป
                                                                                                      CU
                                                                                                       Q>  w
                                                                                                      ฃ  +->
                                                                                                      ^S
                                                                                                      ^ง
                                                                                                       2  Q,
                                                                                                       i  E
                                                                                                      .5  ซ>
                                                                                                      Qซ
                  ^
               eij

                o>


               43>
               b.

-------
232

Element
Al
Sb
As
Ba
Be
Cd
Ca
Cr
Co
Cu
Fe
Pb
Mg
Mn
Hg
Ni
K
Se
Ag
Na
Tl
Sn
U
Zn
S
Th
Er
Pr
Ce
Sr
B
Cs
Si
Li
I
P
50301

ICP SuperScan
<7B
<58
<4
<21
<2
<4
<825
<8
<18
<19
<3B
<5
<351
<5
<0.2
<27
<2460
<3
<9
<523
<7
<40
<18
<11












0.96
1.01
1.01
1
0.99
1.0B
0
0.99
0.97
0.99
1.11
0.9B
0.3
0.99
1.01
0.97
1.03
1.04
0.97
0.78
0.98
1.01
1
1
0.92
0.96
0.99
1.01
0.95
0.35
1.54
0.96
4.09
1.05
1.11
1.08
50238
Coffee
ICP SuperScan SuperScan
15700
196
140
107
<2
12
2300000
59
<18
<19
8590
13
483000
83
<1
<27
32600
5
<9
6830000
<10
42
<18
79












3.76
1.11
1.1
3.93
1.02
1.11
1156
1.01
1.04
0.96
31.83
0.96
76.84
3.2
1.08
1.05
3.11
1.11
0.99
387.97
1.11
1.02
1.05
1.27
47.98
1.33
3.79
1.8
18.27
675.19
25.94
3.57
91.64
3.01
1.01
0.98
1.03
0.94
1.06
1.99
1
1
11.97
1.02
0.98
0.99
1.41
0.98
9.45
6.47
0.98
0.98
32.5
1
0.99
8.63
1.06
0.99
1.03
1.2
2.89
1.03
1.04
1.01
1.18
2.13
0.57
0.98
5.48
1.02
2.71
4.42

-------
                                                                   233
 O n.
 Q |

(/] ป
 L. -a

 ฐ-
                                                  - 17)
                         1 \ \ \ \ \ \ \ \ \ \ \ \ \ \
                                   i\\\.\\\\\
                                                  - a.
                                                   N
-ft
- F
-I
       i\\\\\\\\\\\\\\\\\\\\\\\\
         I    i   I    i    I   i    I    i    i   i    i
                                                  - Z
                                                    I

                                                    ฃ

                                                    ฃ
                                                    3 .
                                                    8

                                                 h 5
                                              —  c>

-------
234
D B.
o ง
00 ป
L- TJ
13 ""•
Cfl S
I
I \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \\ \ \\
K
k\W\\ V
.\\\\\\\\\\\\\\\\\\ \ \\ \\\\
^"X""\\ \'
C
[
K\\\ \\x\\S\\
[
. \ \ v \ \ \ \ \ \ \ \ \ \ \. \ \. \ \ \ v x \. \. \ v
I
E
k\\\\\\\\\\\\\\\\\\\
[
I^K- \\\\\\
!
\\\\\ \\x\\\\\\\\\\\\\\\\\
I
I
f\
i i t i i i i i i
- D.
- "Jj
I I I I I I I I I I I I I 1 I 1 1 1 I 1 I I 1 I 1 1 1 1 1 1 1 1 1
fil SbA^BtaBrsCdCaCrCoCuFc PbMglvfriHgr-4i K SeAgNaTI 5n V Zn 5 Th CrPrOsSr B CsSi

-------
                                                235
                             l\\\- CL
                            r\ v\\f- is
                                   8
                                   ID
                                 E
                                 [
                                  - |fl
                                   8
                                 >ฃ
                                 E-ij
                                 E-f
                                 E-
                                 E-
                                 K-
                                   fi
                                   w
                                   F
                                   D
                                   l^
                                   ^
                                   z

                                   ฃ
                                    Dป

                                    !
                                    ฃ
                                    3
\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ X
                           KXXXX
                        IX \VV\\\
                                 E-
                                 E-
                     K\\X\\\\\
                                    o
                                    (g
                                    D
                                    (D
                                    ^
                                    ft
                                    3
i  i  i
           i  i  i
I  I  1  I   I  I  I

-------
                                                    236
                          MR. TELLIARD:  Last year we
broke with tradition and had some of the critter
people come in.  Our next speaker is a critter person.
Last year we had a meeting in San Diego on a public
hearing on the oil and gas regs, and Dr. Tom couldn't
make it because of no travel money, which I guess is
similar this year, so I had to give his talk for him.
Of course, being a real critter person like myself, I
talked about it, and in the process, one of the
questions was, why were we using mysids instead
of the California avocado or something.  And I said,
well, it's a national standard.  We have a national
method to compare national things.  And from the
front row I hear a voice go, "Oh, shit, a National
Critter."
     So, Tom is now the keeper of the National Critter.
He works down there in Gulf Breeze, Florida at our
laboratory, R&D lab, and they have bus loads of little
kids coming in to see the National Critter.  Of
course, it's a lot easier than using bald eagles on
these tests.
     But we thought about how we were going to really
bring this home, this fact that it's a National
Critter.  We have the bald eagle.  We thought of a
flag, a National Critter flag, but that really isn't

-------
                                                    237
going to sell.  So, what we did is we arranged with



the Postmaster General to develop a stamp with the



National Critter on it.  It will be 22 cents and it



will bring it out.  So, when you go back to the lab,



Tom, you can give that to Hank and say, just send



it in to the Postmaster General and you're all set to



go.



                          DR. DUKE:  I'll tell him



that, Bill.



                          MR. TELLIARD:  Dr. Tom.

-------
238

-------
                                                                      240
in the fluids), biocides, hydrocarbons for lubricity,  caustic,  and  other
material.  The exact formulation of drilling fluids discharged  depends
upon the substrate through which the drilling is taking place,  depth  of
the well, and particular functions of the drilling fluid required at  any
time.  Therefore, there is no "typical" drilling fluid.

Best Available Technology   .
     It may be helpful at this point to review some of the criteria for a
toxicity test for BAT determinations because the guidelines are somewhat
unique.  The test must be "generic" in the sense that BAT guidelines are
for general application as opposed to Section 403-c guidelines that pertain
to specific regional areas and can include requirements for indigenous
organisms and more extensive testing procedure.  The test organism for
BAT must be reasonably sensitive to drilling fluids yet strong enough to
be transported to remote  laboratory testing facilities  if necessary.
Test procedures  should be straight-forward yet  rigorous enough to yield
scientifically acceptable.data.  Also, a  substantial data base involving
the test animals and  drilling  fluids  should exist.
 Methods
      With  these criteria  in  mind,  the  mysid,  Mysidopsis  bahia, was selected
 as the test animal.   A rather extensive  data  base  is  available on the
 effects of laboratory-prepared as  well as  "discharged" drilling  fluids
 on mysids  (Duke and  Parrish, 1984).  Methods  for handling,  acclimating
 and sizing bioassay  organisms given  by Borthwick (1978)  and Nimmo et al.
 (1977) are followed  in the toxicity  tests.
      Our laboratory's responsibilities in  the comparative study  conducted

-------
                                                                      239
                     DRILLING FLUID TEST PROCEDURES:
            PARTICIPATION, DATA COMPARISON AND IMPLEMENTATION
                                    By
                        T.W. Duke and P.R. Parrish
Introduction
     Our laboratory (Environmental Research Laboratory, Gulf Breeze)  has
been involved in the development and implementation of toxicity.testing
methods for Best Available Technology (BAT) guidelines for discharges
from off-shore oil and gas platforms.  We have contributed to the
test method described by Petrazzuolo (1983), tested various drilling
fluids (Duke and Parrish, 1984) and participated in a comparative
laboratory study where sub-samples of a laboratory-prepared drilling
fluid were tested by 11 laboratories.
     The purpose of this talk is to provide background information on the
toxicity tests that serve as a basis for the statistical  evaluation.   A
statistical analysis of the comparative laboratory study will be presented
next by Drs. Eynon and Bailey.  I apologize to those of you that attended
last year's Symposium because some of this material necessarily will  be
repetitive.
Drilling Fluids
     Drilling fluids are used in the rotary drilling process for several
purposes, including transporting cuttings produced by the bit,  lubricating
the bit, coating the bore to prevent fluid loss, and reducing corrosion.
These fluids are a complex mixture of chemicals and clays and,  in order
to perform their functions, can contain barite, bentonite, lignite,
lignosulfonate (these components comprise about 90% of the materials used
                                    1

-------
                                                                      241
by the Office of Water Programs included storing the drilling  fluid
prepared by a commercial  company,  sub-sampling the original  stock  (a
commercially-prepared generic mud  (number 8)  with 3% mineral oil), sending
sub-samples to participating laboratories, establishing test concentrations
through range-finding tests, conducting definitive toxicity  tests,
and supplying to the test method to the other laboratories.  Our  results
were included in th* comparative study.

     A 96-hour concentration lethal to 50 percent of the test  population
(LC50) was used to express the toxicity of the suspended particulate
phase (SPP) of drilling fluid sample used in  the comparative study.
Details of the method were supplied to each participant and  followed
the procedure published in the Federal Register (1985).  In  general,  the
sub-sample to be tested is thoroughly mixed to a volumetric  mud-to-seawater
ratio of 1 to 9.  This slurry is mixed and the pH is adjusted, if necessary,
to within 0.2 units of seawater.  Then, the slurry is allowed  to  settle
for 1 hour.
     At the end of the settling period, the SPP is decanted (not siphoned)
into an appropriate container.  Decanting the SPP is one continuous action.
The decanted solution is defined to be 100 percent SPP.  The SPP is mixed
for 5 minutes  (it must not be preserved or stored) and a sample taken
so that the filterable and unfilterable residue of the SPP can be measured.
Also, dissolved oxygen (should be at least 4.9 ppm or 60% saturation)
and pH of SPP  are measured and adjusted if necessary.  Appropriate volumes

-------
                                                                      242
of 100 percent SPP are mixed with appropriate volumes  of seawater  to
obtain the desired SPP concentrations.   The control  is seawater  only.
The animals are randomly selected and placed in dishes to begin  the
96-hour toxicity tests.

     Definitive test concentrations are based on results of the  range-
finding test (in this instance, ERL-GB supplied range-funding values).
A minimum of five concentrations plus a negative and positive (reference
control) is required for the definitive test.  To estimate the LC50,  two
concentrations are be chosen that give (other than zero and 100  percent)
mortality above and below 50 percent.  Cups constructed of nylon-mesh
screen are inserted into every test dish prior to adding animals.   Twenty
organisms are exposed in each test dish and three replicates (total  of
60 animals) are used in each test concentration.  Individual animals  are
randomly assigned to treatments.  Throughout the test period, mysids  are
fed daily with apporximately 50 Artemia (brine shrimp) nauplii per mysid.
Dishes are covered and incubated in a appropriate test chamber;  test
mixtures are gently aerated throughout the test.  The test medium is
not replaced during the 96-hour test.

     At the end of 96 hours, all live animals are counted.  Death is the
end point, so the number of living animals is recorded.  Death is determined
by the lack of spontaneous movement.
     Data are analyzed according to Finney (1971) to obtain the probit
model estimate of the LC50 and the 95 percent fiducial (confidence) limits

-------
                                                                      243
for the LC50.  These estimates are obtained by using the logrithmic
transform of the concentration.
Results of Comparative Toxicity Test
     The results of the comparative toxicity test and a statistical
analyses will be presented in the next paper by Drs. Eynon and Bailey.
From a toxicologicaV point of view, the variability of acceptable results
from the various laboratories was within a reasonable range, i.e., the
range was within those limits reported for single chemical toxicity
tests conducted by several laboratories in similar studies.  In our
opinion, the manner in which the SPP was prepared by each laboratory was
probably the largest contributing factor to variation of results among
the test laboratories.  This conclusion is based on experience of our
staff as well as discussions with personnel from other laboratories.
There is no doubt that this BAT toxicity test can be improved, but it is our
opinion that it conforms to state-of-the-art testing methodology and is
appropriate to determine the acute effects of drilling fluids on mysids.

-------
                                                                      244
                             Literature Cited
Borthwick, Patrick W. 1978.  Methods for Acute Static Toxicity Tests with
     Mysid Shrimp (Mysidopsis bahia).  In:   Bioassay Procedures for the Ocean
     Disposal Permit Program, EPA-600/9-78-010:   ERL-GB.

Duke, T.W. and P.R. Parrish.  1984.  Results of the drilling fluids research
     program sponsored by the Gulf Breeze Environmental  Research Laboratory,
     1976-1984, and their application to hazard assessment.   EPA-600/4-
     84-055, Environmental Research Laboratory,  Gulf Breeze, FL.  94 pp
     plus appendices.

Federal Register Aug. 26, 1985.  40 CFR Part 435.  Oil  and Gas Extraction
     Point Source Category, Offshore Subcategory; Effluent Limitations
     Guidelines and New Source Performance Standards; Proposed Rule.
     Part II. pp 34631-34635
Finney, D.J.  1971.  Probit Analyses.  3pd Edition.  Cambridge University
     Press.

Nimmo, D.R., L.H. Bahner, R.A. Rigby, J.M. Sheppard and A.J. Wilson, Jr.
     1977.  Mysidopsis bahia:  an estuarine species suitable for life-
     cycle toxicity tests to determine the effects of a pollutant.  In:
     Aquatic Toxicology and Hazard Evaluation, F.L. Mayer and J.L. Hamelink,
     Eds., pp. 109-116.  ASTM STP 634, American Society for Testing and
     Materials, Philadelphia, PA.

Petrazzuolo, G.  1983.  Proposed methodology:  Drilling fluids toxicity test
     for offshore subcategory; oil and gas extraction industry.  Unpublished
     Report.  May 19, 1983. 45pp.

-------
                                                    245
             QUESTION AND ANSWER SESSION
                          MR. TELLIARD:  Questions?
                          AUDIENCE PARTICIPANT: Where
do you get a copy of this...
                          DR. DUKE:  This has been
sent out by Bill's group.
                          MR. TELLIARD:  It was in
the Federal Register notice on the proposed regulations
for the oil and gas industry, which date was August
28thf 1985.
                          AUDIENCE PARTICIPANT:  The
second question concerning  the method.  You talk
about DO requirements.  What is a DO requirement?  I
mean, I can look it up in a method, but what is a DO
requirement for this particular method, when you're
setting it up?
                          DR. DUKE:  We keep it at 60
percent...at least 60 percent of saturation.
                          DR. COWGILL:  Michela
Cowgill, Dow.  I have sort  of a technical question.
You had an LC50 for selenium with  something like 35
ppb?  Did I misread it from back here?
                          DR. DUKE:  No, the only
LCSO's I showed were drilling fluid LCSO's.

-------
                                                    246
                          DR. COWGILL:  Excuse me?
                          DR. DUKE:  Those were not
drilling fluid LCSO's.  The table with selenium, I
think, was an analytical analysis.
                          DR. COWGILL:  Of the?
                          DR. DUKE:  Of a series of
elements that were analyzed by various laboratories.
                          DR. COWGILL:  It wasn't an
LC50, I see.
50.
you.
Tom.
                          DR. DUKE:  It wasn't an LC-
                          DR. COWGILL:  Okay, thank
                          DR. TELLIARD:  Thanks, Dr,

-------
                                                    247




                          MR. TELLIARD:  Our next




speaker is Barry Eynon from SRI.  He's going to talk



about the statistics and the round-robin that was run.



     We ran 10 laboratories, as Tom pointed out.  This




was with a drilling fluid that was a simulated



drilling fluid.  It was hot-rolled and the labora-




tories ran these check samples.  The Gulf Breeze




laboratory ran the range finders so it gave them the




window, and then they ran the samples for the



drilling fluid.



     The drilling fluid was Number 8 MUCL with  two




percent mineral oil added.

-------
248

-------
249

-------
250

-------
                                   251
              PRnfFTIURF;
         WATER QUALITY
|fe*M>
siLvi ' .r"?!1"-,.1'. -
INCUBATE FOR 96 HOURS

COUNT LIVING WSIDS

CALCULATE 96-HOUR LC50
                              •'.-'. -.-'j -^m

                              .^t&isj&t

                               •••'*••'";•"ซ*?ป<'

-------
252

-------
253

-------
254

-------
                                                    255
Csi     CN1
                    CSI     OJ
csi
                    ," -  ..-i." •;',','". ---,;-'1-' -.'"• ..^;'-, "•'-.'' -f'.^^&ri
                - - •-   • -. - ./.;"--L.. -, • A, , _ j ;..,'-•''-•",-'V'.^. j.'"^.-^---^^';^-^

-------
                                                            256
                   V ,-,  -^:,•''..' :..•'•• .1"'"f;"|j"'•*•••','•' ••'>'• •••"f\'lj'i-v:,;>-f ^'•••'fffS1'''^


                  •".-•'•"'  •;'••. ••.'•".•'• w'-'V^,':'--:" • ••'••:•., .'.'.'.j'A '>"'i/"'''v:!:;'||i:&?*?
                  ""•   •- •'.,- ". ,"'-i-^  "'' •ซ'-/', ••,':i'.!,f A- '"-'iiliwJtt
                   '.; ,   •:  ". ••- ',•.'•'-,'"fe'v'r--,1'-'•A i,'",- 'l';'^i::  '","4&E%lf


                  SOURCES
                     FIUIDS
                        TEST  ANIMALS
 tEST MATERIAL  PREPARATION—  Phase se|


flEST MATERIAL  COMPOSITION


        CONDITIONS       .   ;          '    .   •:  •'

                                              V

-------
                            257
FLUIDS ACUTE TOXICITY

    — Source
    ~ Age
    — Condition

-------
                                                  258
i^"t^''vvi:"Oซ'•'-'" *"^"t?'.**:'.'„'•'' '•> ' '.".. • '•'
TK"W?\. •J^/^'.'JSi'rit*"-.*':,-^'. ':.•_',: ..•,.""'••:•' '•.•'
S^..;1..:'^/..-1."-"

tn*"W",.  * ,;.

                                          '
              Pl^Ifi  |pf E TOXlCilTY
                 CcMnpositipn,^      ^

                 Ass

                 Preparation
                 1.   Phase separation
                 2.   pH  adjustment
                 3.   Aeration

-------
                                                   259
  ^tosgfew*^;^iV' ''V^'v^x -r; '•'•, ,'... .'•  v ''
      ^.W^:
5H*<.V"' ^"T = - "-'-' '- ':•' . - •'
%8&.?':y'^-X'•'
feSg^IiM::,:-':.,
SOORCES OF VARIATION
H*-^^i-*'' .*,J •-••'• '; '- :•-
r^^S^'v/" --
       fplilNG FLUIDS ACUTE TOXICITY  TfST

               i            •'-.''.

       TEST CONDITIONS ~ Aeration


                          — Containers


                          — Other

-------
                                      260
        SOURCES OF VARIATION

DRILLING FLUIDS ACUTE TOXICITY TEST
                         ! ' '" :'.-*' •> - ,V i ;^ ^";:

TEST FACILITY  —Experience

               — Skill

-------
                                                     261
                   BARRETT P. EYNON

                  SRI INTERNATIONAL
        STATISTICAL EVALUATION AND VALIDATION
  OF THE EPA DRILLING FLUID TOXICITY TEST PROCEDURE
                          MR. EYNON:  Thanks, Bill.

Yes, I'm glad Tom was able to talk and show you

something about the shrimp and how that's all done.

That's something he's far more familiar with than I

am.  What I would like to talk about is the results

of this round-robin study that was run.

SLIDE 1

                          MR. EYNON:  The results

of the round-robin study that was performed about a

year ago.  That study had two objectives.  One was to

evaluate and analyze the performance of the method in

support of the effluent guideline preparations, and

also to aid in the selection of contract laboratories

for performances of these analyses, which that's been

done and analyses are ongoing.

     This is part of the development process of the

method, and the kinds of analyses we're looking at

today are being fed back to Dr. Duke to help him

develop the method further and provide extra informa-

-------
                                                    262
tion.  I guess we can go to the next slide.
SLIDE 2
     I've got a couple background slides here.  These
were drilling fluids.  Pretty much this is what we
said before.  These are discharged into the water and
that's why we mix the SPP down with water to determine
toxicity.  Let's flip on here.
                          MR. TELLIARD:  Barry, what's
the SPP?
SLIDE 3 & 4
                          MR. EYNON:  Oh, I'm sorry.
The particulate...suspended particulate phase of
the material.  Here's the reference to the method in
the Federal Register, Monday, 1985, and this method
was distributed to 10 laboratories who were participat-
ing in the study, along with a...we can go to the
next slide...along with a well mixed preparation of
drilling fluid Number 8.  They all got the same
sample.  Then they were asked to perform the method
using their procedures.  Where not defined by the
method, the labs were given the procedure and told to
perform the analysis.
     They also performed a reference toxicant.  They
took three to five day old mysid shrimp, and as Tom
said, we took 96 hour exposure at five dilutions of

-------
                                                    263
the particulate phase and a control at 0 percent.
The three dishes of 20 shrimp of the drilling fluid
at each concentration.  I will note that, as Tom
showed, the randomization procedure was used to make
sure that the shrimp were randomly distributed and
fairly distributed in the various dishes, and I think
that's an important competent of the procedure.
SLIDE 5
     Then we have our obligatory National Critter
slide here with shrimp on it*  Flip on.
SLIDE 6
     In conjunction with the study, there were also
six replicate runs done at Gulf Breeze of the same
drilling fluid by the same method so we can compare
inter- and intralaboratory variation.  We also
collected information on possible factors that could
be used to control the method or aid in controlling
for variability.
     Some of those were the amount of total suspended
solids...each laboratory was instructed to measure
that...the amount of acid used in the sample, the
initial pH of the sample, and some other factors as
possible to measure.
SLIDE 7
     The analytical form of the dose response function

-------
                                                    264
that we used in this analysis is the traditional
Probit function with an adjustment for spontaneous
response rate.  Sof just to refresh us on that, the
form is a S-shaped dose response curve.  Here we
applied it to the logarithms of the concentrations
and then the C+l-C is the correction for the spontaneous
response rate.  So the intercept at O...as dose goes
off to 0, is C.  These parameters were fitted by...
no, let's hold with that.  Yes, let's hold with the
slide for a second there.
     These parameters were fitted to the data for
each laboratory by maximal likelihood fitting techniques
to estimate the three parameters, C, Mu and Sigma.
The LC50 is then the exponential...because we're
working with log doses, the LC50 is the exponential of
the parameter Mu.  Mu is the center, the log of the
50 percent point of the curve; Sigma determines the
rate at which the curve increases with increasing
dose, the rate of mortality increase; and then C, as
we said, is the spontaneous response rate.
     The procedure that we wrote was written in the
statistical analysis system package as a macro program
using the procedures NLIN and MATRIX.  And if
anybody's interested in the analytical techniques we
used, I'd be happy to talk to them later.

-------
                                                    265




     We found that there were some limitations in



some of the published software.  For instance, SAS




Proc Probit had some convergence problems with some



of the data, not that the data was bad but I think



the program has a few limitations.  So we developed




our own analytical package to do this.




     We also looked at, and I won't say a whole lot,




but we also did look at some other methods of




calculating LC50, including the moving average method




and also an extension to the Probit called the Burrit



model, which has some extra parameters for shape of




the response curve.  Let's go to the next slide.



SLIDE 8




     The next three slides are just for your edification



to see some examples of the actual data and the fitted



dose response curves in several situations.  This is



the Gulf Breeze Laboratory data, and we have mud




concentration in percentage across the bottom.  This



is on a log scale, and then the proportion of shrimp



which died at each concentration as the vertical



scale.




     So, the data are the pluses at each concentration;



1, 2, 4, 8, 16 percent.  And then there's a control




dose, which is way over on the left hand side.  The



two plots are just...one plot is the raw response and

-------
                                                    266
the other plot is the adjusted response.  Here, the
control response was 0, so the two plots are actually
the same.
SLIDE 9
     The other two slides are just two of the labs.
This is Lab 3.  A similar curve, but notice the slope
is a little bit shallower.
SLIDE 10
     For Lab 4, we get a steeper slope response.
This kind of variability is characteristic of the
variability in the response curves on the same material
at different labs.  So, that's the kind of variability
we're looking at.  We'll come back to that a little
bit more.
SLIDE 11
     The first objective of the study that we looked
at was to evaluate the laboratories for contract
performance.  Several criteria were used.  We compared
the results for each lab with the reference lab, with
the Gulf Breeze lab.  Each calculation of the LC50
also obtains a confidence bound within the procedure,
and the width of that confidence bound is an indication
of the accuracy of the lab.  Then we also performed a
comparison of each lab with the group of all other
labs so that we were sure...if the reference lab was

-------
                                                    267
distinctly different, then we would capture that
comparison.
     Each of these factors was used to rank the labs,
and a combined ranking was put together which allowed
an overall ranking of the labs that was then given
back to EGD to aid them in the contract selection.
Okay, we've got the criteria and the rankings.
So that was the first step.  Now we want to go look
at variability and possible explanations.
SLIDE 12
     Here are the parameters of the fitted curves for
each of the 10 labs and for each of the six replicates.
The top plot is...one comment on the laboratory
ranking.  We found that 3 of the 10 labs did not
perform what we thought were acceptable runs.  The
results were too far away from the central lab.  One
of the labs consistently got over 80 percent control
mortality and so was rejected on that basis, and the
other two labs; one was very high mortality and one
was very low.
     For purposes of this study, because we felt that
with proper training those labs could be improved or
were not considered within the range of acceptable
variation, they were removed from basically the rest
of the study.

-------
                                                    268
     So, here we have the parameters plotted for the
labs that we were considering.  The Mu is the log of
the LC50, Sigma is the slope response, and you can
see that there is some variation between labs on the
top plot, and within labs on the bottom plot, and
that there's more, as we expect, there's more variation
between labs in the top plot.  Go to the next slide.
SLIDE 13
     We can lay out the actual results.  The parameters
for each of the six trials, trials labeled A through
P, are given in the table for the drilling fluid and
for the reference toxicant.  This table also shows
the achieved log likelihood, which is a measure of
fit, and the criterion by which we fit the parameters,
and that's always a negative number; 0 would be
perfect, perfect fit.  Differences between the log
likelihood can be used to measure and test the
effectiveness of different models.
     So, the table is designed to do three things;
first, to show the parameters; second, to test.  At
the bottom, we can test for whether there are signi-
ficant differences between the samples within labs,
and the Chi-Sq test, listed as Chi-Sg of intralab,
does show significant...there are significant varia-
tions within labs, even within the reference lab, on

-------
                                                    269
the different trials.  And that simply means there is
intralab variability above and beyond the natural
variability of the parameter estimates themselves.
     The second Chi-Sq is the same test where we
allowed a separate spontaneous response rate to be
estimated for each trial and gets a similar result.
And the bottom Chi-Sq is a test of whether the
spontaneous responses are separate for each trial.
This is-also significant.
     So, it appears that one feature to be considered
is that separate spontaneous response needs to be
considered for each trial.  this may be due to
different batches of shrimp or different other time
factors.  So, all runs should be allowed a separate
spontaneous response rate and then the LC50 is the
corrected value.
SLIDE 14
     The next table is...that was intralab.  This table
looks at interlab variations.  This doesn't, unfor-
tunately, have the parameter estimates on them, but
there were two objectives on this table.   One was to
test within each lab a goodness of fit result.  We
can also test versus an arbitrary response curve
that's not the Probit curve, each lab individually,
and we can see that on each lab there is  a not

-------
                                                    270
significant result for the test of goodness of fit,
which means the Probit model is doing a good job of
fitting the data.
     Then, when we get down to the bottom, we can
also test whether there were interlaboratory differ-
ences.  Given that we saw intra-, it would be surpris-
ing if there weren't.  We do confirm that there are
interlaboratory differences.  So, all of those sources
of variability are present in much the same way as
they are with an ordinary analytical function.  So we
have that.
SLIDE 15
     The next slide lays out...we can do almost an
analog to the analysis of variants table using the
likelihood functions, and we can split up the likelihood
into inter, intra, and several other factors, basically
which are all goodness of fit factors.  The inter- and
intralab are highly significant.  There are slightly
significant variations between the trials.  That's, I
think, all I want to say about that slide.
SLIDE 16
     To try and summarize the information that we can
get from the inter- and intralaboratory comparison, I
put together this table, which takes the LCSO's and
looks at what the average...geometric average LC50

-------
                                                    271
would be for both the inter- and intra-, and also



what a 95 percent variability multiplier, since these



are again on the log scale, what a 95 percent



variability multiplier would be for the amount of



variability we saw among these labs.



     For the drilling fluid, six concentrations times



60 shrimp.  Intralab...well, the minimum that this



could ever achieve would be the inter-experiment



variability that is just due to randomization and the



repetition of the process.  That's the best you're



ever going to be able to get, and that's the number




I've listed under experiment.  The best you could get



with six times 60 and the design series, about 1.2,



or about + or -20 percent.



     The actual that we see; intralab is more like a



multiplier of 1.9, and interlab we see a multiple of



about 2.6.  Tom says that in his experience this is



actually pretty good for biological testing and that



he's very pleased with the fact that the labs do a



pretty comparable job.  This is larger than one sees



in analytical testing situations and one has to keep



that in mind in thinking about using these types of



procedures.



     Similarly, on the reference toxicant we also see



that as you go from the intra-experiment to intralab

-------
                                                    272
to interlab there's also an increasing variability,
so this is an indication that differing lab standards
and procedures can have an effect on the result/ and
that one thing this leads us to is to ask the question
can we identify those qualities and propose better
details to be provided to the laboratories to help
them control these features.  So, this leads us to
ways to measure other information that will help us
evaluate the test.  If we can go to the next slide.
SLIDE 17
     We have proposed some covariate models that take
the Mu and Sigma, which in the first round are simply
parameters to be estimated, and now we express
those, we attempt to express those, as a functional
model dependent on other covariate information.  So
that, for instance, the LC50 or the log of the LC50
may be dependent upon some other factor we can measure,
and we can measure...by keeping track of that over
several trials, we can estimate that much like a
regression model.
     This is imbedded in the maximal likelihood fitting
scheme.  I don't think I'll say too much about that
except that it follows as a sort of a natural analog
to a regression type of analysis.  We get the same
kinds of results, we get confidence intervals on the

-------
                                                    273



parameters, and we can test the significance of the



covariates.  Next slide.



SLIDE 18



     The reasons for looking at this are threefold,



depending on the particular nature of the covariate.



In each of these cases we may be able to do something



to improve methodologies.



     If there are factors that can be both measured



and fixed, we can fix them if they're important.  And



that's a cost versus benefit type tradeoff.  If there



are procedures that are out there but we can't control



them and we can't measure them, we can randomize over



possible sources of variability to reduce their



effect.  If we can measure but not control certain



parameters, we can consider the possibility of numer-



ically adjusting the results of the analysis to



compensate for the variations of those between runs.



All three of those procedures can show...have some



possibility of improving the results and that's



what we're working on now.



     Just for instance...let's show the next slide.



SLIDE 19



     One way of thinking of this is to look at the



MU, which is the log of the LC50, as a function of



one or more covariates.  Here's a plot where we show

-------
                                                     274
the actual Mu's for the various labs plotted as a
function of the initial pH of the SPP of the pre-
pared dishes.  Now, notice, we're looking at a very
restricted range here.  The pH is controlled anyway,
and in fact, in this particular case, it's not clear
that it can be controlled more.  It's one of those
that has been controlled, and the question is, is
there residual effect of that variable.
     We can see a small, small trend here, possibly,
in the upper picture.  That's of Mu versus pH.  Then,
similarly, there may be a small negative trend in the
lower picture where the steepness of the slope also
depends on the pH.  These are possible potential
relationships we're still exploring.
SLIDE 20
     When we do this analysis, we can actually produce
tests of hypotheses and layouts similar again to the
regression or analysis of various types of approaches
where we test for the effects of factors.  Here, this
is a table where we have the log of the original TSS,
the amount of acid, and Mu Ref, which is the log LC50
from the reference toxicant, all three of which we
thought might be predicters for toxicity or might be
influences on toxicity or correlated with toxicity in
the sample.  Here we do find some significant results.

-------
                                                    275
Singly and in combination we get parameter estimates
for the various coefficients, and likelihood value
improves, and it improves it by a significant amount
in each case.  So, this says that these parameters are
having some effect and we can think about controlling
them.
     We've looked into them.  I think that one thing
we've decided is that TSS needs to be more precisely
measured in the various samples because it's got a
high possibility of being important, and that some of
the other factors are also potential.
SLIDE 21
     Let me skip these and show just the last slide,
which is just to come back to the Burrit model.  To
compare some of the models, again for goodness of fit,
we fitted the Burrit function as opposed to the
Probit.  The Probit uses the cumulative distribution
from the normal.  The Burrit uses a different F
function which has two other parameters which control
its shape.  We've made the two sets of parameters
comparable in that Mu and Sigma mean the same thing
in both models.   We can see that under the Probit,
the parameters are very similar under the Burrit.
CB and KB are the shaped parameters.  The KB likes
to be very large, which is maybe a Wibel model, which

-------
                                                    276
is different particular distribution, but way over on
the right hand side...well, I guess it's not.  The
point is that these improvements are not significant
in any of the particular cases, and so we are reassured
again that the Probit is probably a good model to
use.
     Let me stop there and entertain questions.

-------
                                                     277
             QUESTION AND ANSWER SESSION
                          MR. TELLIARD:  Questions?
                          MR. PLOST:  Charles Plost
with EPA.  As chemists, we're traditionally accustomed
to biological variability being the analyst, and so
we characterize our methods by running round-robins,
where typically we'll have two dozen laboratories as
opposed to 12.  We'll go to great pains to throw out
one of the laboratories, and really give ourselves gas
when we try to throw out two, and at the end of that
we'll feel that we know something about the precision
and accuracy of the methodology when applied by any
laboratory.  It seems that you've done something
that's somewhat akin to that.
     May I translate this one?
                          MR. PLOST:  Well, one, why
did you throw out three laboratories?  Two, can you
extrapolate from this as we've extrapolated with...to
how would the next laboratory perform?  Is this an
alternative way of running a round-robin?
                          MR. EYNON:  The judgment to
throw out; one laboratory was just clearly not in
compliance with the procedures.   It had very,  very
bad response.   The other two laboratories, one was

-------
                                                    278
the highest and one was the very lowest, and in a
best judgment situation, given that this is a
development study and not a validation study, the
best judgment was that those laboratories had failed
to perform the procedure as specified.
     If we were going to try and actually state the
precision of the method rather than evaluate it, then
I would worry about that because you'd have to
say you have to either have some check or some formal
test for outliers.  But the feeling was, and I think
Tom said to me, that what he wanted to do was get the
labs in and train them.  They were given the procedure
and told to perform it, and he feels that they can
make improvements, and I think have made improvements
since then.  He's been able to see improvements when
they learn the method and improve.
     So, I could only express a personal desire to
have another round-robin at some point in the future.
I have no idea whether such a thing is in the works.
                          MR. PLOST:  But would this
technique substitute for a round-robin?
                          MR. EYNON:  Well, this is
a round-robin.
                          MR. PLOST:  Sort of.
                          MR. TELLIARD:  Well, Charlie,

-------
                                                     279
I think that the fact that we used 10 laboratories...!
mean we validated some of the 600 series on that  same
number of labs, or less in some  instances.  On some  of
the methods we only had, we won't bring that up.   In
this case we did use 10.
     I think also the universe of available laboratories
compared to...remember when we started back with  1624
and 1625 and all that, we didn't have...George had
the lab in his garage that he was generating API
data and we had a couple of others.  Now we have  this
magnitude of laboratories.  I think the same thing is
true of the biological environment.  There aren't
that many labs doing this type of work.
     I was very surprised that We had 13 or 14 even
submit.  We were thinking more like six when we orig-
inally started to do it.  So, we were pretty pleased
with the study.  Of course, I don't know what's a
good critter test as...but I think the folks that do
critter things are fairly pleased with this and we
are going to try to move on with it.   If we're going
to do another validation,  that costs money.  I prefer
not to spend my own.   I  prefer to spend R&D's.  So,
that's where we're coming  down the road on it.
                          MR. RICE:   Jim Rice.  I'd
like to make a little comment on that.   You know, one

-------
                                                    280
of the problems that we have...we've debated the
issue of casting out data in every round I've ever
heard.
                          MR. RICE:  Unfortunately
what happens, once you cast it out...we hear your
reasons heref but they get lost.  The conclusions
that you draw on the basis of only 70 percent, in
effect, of the laboratories participating is what
tends to stay for a long time, and it's hard, people
make decisions on the basis of that sort of repro-
ducibility.  It would be well if you, in whatever
you publish out of this, if you perform both analyses,
but a much more rigorous interpretation on what you
cast out.
     I understand the point when you're trying to do
it for methods development.  That's perfectly
legitimate.  You really want to find out what the
core of the problem is and get some of the really bad
actors out, but when the data sits around, as this
does, and it becomes very important to others in
making decisions on whether to use it or what's
possible with it or not, I think you owe it to
everybody to put both analyses in.
                          MR. EYNON:  Sure.  I think
if you noticed on a couple of the slides, I had double

-------
                                                     281



line entries for the...well, I probably went right by



it, but some of the slides do have double entries



with all labs and with 5, 9 and 10 excluded.  I felt



the most interesting results are when you get rid of



the labs because the bad labs are going to be the



ones with the most influence, too, and if that's due



to reasons that you have not been able to look at or



the lab failing to follow the protocol at all, you



can totally mess up your results.  Because of the



high influence and stuff, I think that a much stronger



piece of information comes from looking at those labs



that appeared good.



     Certainly we have all the data and all the



results for all the labs in our reports, and we felt



that this was the appropriate subset to be looking at



to gain the most information.  I agree with you, as



you start to use it, things like this do sit around,



and it's a good point.



                          MR. WHITLOCK:  Stu Whitlock



from ESE.  Did you go visit the laboratories before



you sent them the samples, or did you just do it on



publication only?



                          MR. TELLIARD:  We didn't



visit the labs at all, Stu.   On the five lowest



bidders, like we do anytime we issue...this was an

-------
                                                    282
IFB that we're issuing...we visited those five
laboratories.  I don't know if those were any of the
ones we threw out.  Do we know?
                          MR. WHITLOCK:  Well, let me
ask a second question first and then you can answer
that.  When you looked at the types of animals used,
did you look at whether or not they purchased the
animals they used, whether they cultured their own
animals or what the situation was there?  That could be
a very large variability that probably should be
considered.
                          MR. EYNON:  That's one of
the factors we thought about.
                          MR. TELLIARD:  After we put
out that IFB, the price of mysids went up.  You
could buy lobster for that price.
                          MRS. DENAGY:  Susan DeNagy,
EPA.  I guess I can comment on your question.  The
laboratories just were not arbitrarily thrown out.
In the procedure, the whole testing was done because
of a contract, and we have to underline that.  It was
IFB, the laboratories, everything was predesignated
in the contract bid.  In the procedure, if you have
greater than 10 percent mortality, you're thrown out.
It's just...that's it.  So it's not arbitrarily thrown

-------
                                                    283
out, it's part of the procedure.
     There are other stated requirements for that IFB
that if you...if the data did not reach that peak of
quality, you were told ahead of time you're invalidated.
So, it's not as if you're good, you're bad.  Everything
was predesignated.  The testing was controlled in the
respect that Gulf Breeze purposely tested and gave
out the SPP concentrations ahead of time, which would
not normally be done.  So this was not a round-robin
in the true sense, but we were gaining a lot of
information and learning where variability would be
coming in from that point on.
                          MR. EYNON:  One way to
characterize the labs that were thrown out, and I
guess a point I didn't really mention, is that Gulf
Breeze...the procedure is actually a two step procedure
and contains a range finding step.  The laboratories
were told basically the results of a range finding
which had been done, because Gulf Breeze was familiar
with the mud.  They were told what concentrations to
use so that that source of variability would be
controlled.  One lab had a very bad control response,
and no question, they failed to meet the stated
specifications.  The other two labs, one was well
above and the other was well below the stated range

-------
                                                    284
of the concentrations, and for that reason alone, the
results at those labs are very unreliable, if for no
other reason.
     So, that's actually...we didn't say we're going
to throw out all labs who come in above and below the
range.  We looked at the data and we said this guy's
way high, this guy's way low and we think it's best
for our purposes that we remove those from the study.
                          MR. TELLIARD:  Dr. Tom.
                          DR. DUKE:  I just want to
comment real quickly about the animal question.  It's
a good question.  But for the purpose of this
experiment...! won't say round-robin, I'm just going
to say project...we wanted to let people use the
animals as they would if they got a sample.  Some
would culture them, some would send off for them, what-
ever.  That was just one of the variables that we saw.
     The other was seawater.  Some people used natural
seawater, filtered.  Some used sea salts that they put
together themselves, put their own seawater together.
There was a requirement of salinity, temperature, pH,
so forth and so on, but all of these things...and to
me it's...I still, as a toxicologist, I've seen
studies with single pure compounds, and about the best
we can do when they're sent nationwide like this with

-------
                                                    285
10 to 20 labs...I know of two studies where there's a
factor of four between the lows and the highs.  It's
sort of amazing to me that we came this close with a
complex drilling fluid under the conditions that we
did.  It's encouraging.  It leads us to say, well,
now we can look forward and do some other things with
it.
                          MR. TELLIARD:  Thanks, Barry.
It's time for a break.  Is Mr. Brown here from
Battelle?  Would he go to the registration desk at
break time?
     All right, 10 minutes or 15 minutes we'll get
back in here.  Thank you.
(WHEREUPON, a brief recess was taken.)

-------
                                     286
Statistical Evaluation and Analysis of the EP|
Drilling Fluid Toxicity Test Procedure

Barrett P. Eynon, SRI International

R. Clifton Bailey, EPA

Objectives:
   Evaluate and statistically analyze the
   performance of the Drilling Fluid Toxicity
   Test in support of the NSPS/BAT Effluent
   Guideline for the Offshore Segment of thi
   Oil and Gas Extraction Industry.

   Aid in selection of contract laboratories.
Participants1
   Industrial Technology Division and Analys
   and Evaluation Division of the EPA Office
   of Water Regulations and Standards, and
   EPA Gulf Breeze Laboratory.

-------
                      Background

     While developing I1SPS/BHT Effluent Limitations

  Guidelines for the Offshore segment of the oil and gas
                             w1                  ป

       extraction industry point source category,
                                                  287
are
     Fluids mere identified as toxic materials rohich
discharged into the aquatic environment surrounding
          offshore drilling operations.
   Drilling fluids provide lubrication and assist in the
                extraction of cuttings.
    II, ISSi

-------
                                                     288
       To regulate the discharge of drilling fluids

                   EPfl proposed a
              Drilling Fluids Toxicitg Test
            to be used as a compliance tool.

       The Industrial Technology Division (OWRS)
                in cooperation urith the
     Office of Research and Development Laboratorg
                          in
                  Gulf Breeze, Florida
                         and
       Rnalgsis and Evaluation Division (OWRS)
                        refined
                a laboratory procedure
                         for
        evaluation of the Drilling Fluid Toxicitg
                          in
       Support of the proposed DSPS/BflT Effluent
      Guideline for the Offshore Oil and Gas Industrg
                         and
               Published the test in the
  Federal Register. Vol. 50. Do. IBS, Dlondag, Rugust 28,
   1965, Oil and Gas Extraction Point Source Categorg,
Offshore Subcategorg; Effluent Limitations Guidelines and
   Dem Source Performance Standards; Proposed Rule.
     flppendix 3-Drilling Fluids Toxicitg Test. Pages
   34631-34635 Hfllso  contained in the Development
                     Document).
     I, Uli

-------
                                      289
Hethod:
   Toxicity analysis of drilling fluid by
   dilution of the suspended particulate phase
   (SPP) of prepared samples into dishes
   containing 3-5 day old mysid shrimp
   (Hysidopsi bahia).

   Shrimp mortality was evaluated after 96
   hours of exposure, at 5 dilutions (\%, 2%,
   4%, Q%, and }Q% SPP) plus a control.
   Mortality on a reference toxicant (soldium
   laurel sulfate) was also measured.

   Each  test concentration was applied to
   three dishes of 20 shrimp each. Shrimp
   were assigned to dishes according to a
   randomization procedure.

-------
                                                                                        290
                     antennule
                     antenna
                    aatcnnal ink
                                                                   donalproco
                                                                      natocyn
                                           8th thoncic limb .     pfeopods


                                              abdominal Mtmenis

                                                                     •leison

                                                                     •eodoood
                                          thoracic K^ntcnis    donaj process
xopod
       Figure 17. Lateral and  dorsal view of a typical mysid.(From Stuck  et al.,
                   1979).
                                     B
        Figure 18. Morphological  characteristics used  in  mysid  identification
                    (Mysidopsis bahia).  A,  antenna  1,  ventral male; B, antenna 2;
                    C,  telson; D,  right uropod, dorsal.   Scale lines 0.5 mm  in
                    length.  (From Molenock, 1969).
_
                                                105

-------
                                       2Q1
Data:
   10 test laboratories applied the protocol to
   an EPA provided sample of the same well-
   mixed, pretested batch of driling fluid.
   Dose levels were determined based on
   preliminary analyses at reference lab

   Results were compared with six replicate
   trials at the EPA laboratory.

   Additional information was collected on
   possible covariate  factors, including the
   total suspended solids (TSS) of the sample,
   the amount of acid used in the
   neutralization of the sample, the pH and
   dissolved oxygen (DO) of dishes each day.

-------
Calculation of mortality:
Probit model :
                                       292
   P(d) = C + (1-C) * ( (log(d) - ji ) /
   d : dose
   C : spontaneous response rate
   0 : probit function
   jj, 
-------
                                                         293
      DRILLING MUD  TOXICITY  RESULTS
RESPONSE
 1.0
 0.8
 0.6
 0.4
 0.2
 0.0
      CTRL
                   WITH FITTED PROBIT CURVE
                           IAB=0
12      4

 MUD CONCENTRATION {
            32
      DRILLING  MUD  TOXICITY RESULTS
 ADJRESP
  1.0
  0.8



  0.6



  0.4



  0.2



  0.0
               ADJUSTED FOR NATURAL RESPON9IVITV
                           LAB=0
      CTRL
1       2      4

  MUD CONCENTRATION (ซJ
8
32
       Figure IV-1  DRILLING MUD TOXICITY RESULTS:  REFERENCE LAB
                              19

-------
                                                         294
      DRILLING MUD TOXICITY RESULTS
                   WITH FITTED PROSIT CURVE
                          LAB=3
RESPONSE
 1.0
 0.8



 0.6



 0.4



 0.2



 0.0
      CTRL
1      i      4

  MUD CONCENTRATIONS)
16
32
      DRILLING MUD TOXICITT RESULTS
               ADJUSTED FOR NATURAL RESPONSIVITV
                          LAB=?
 ACURESP
  1.0
  0.8



  0.6



  0.4



  0.2



  0.0
      CTRL
                                                     32
                      MUD CONCENTRATION (f ]
         Figure IV-4  DRILLING MUD TOXICITY RESULTS:  LAB NO. 3
                             22

-------
      DRILLING  MUD  TOXICITY RESULTS
                                                        205
RESPONSE
 1.0
 0.8



 0.6



 0.4



 0.2



 0.0
      CTRL
                  WITH FITTED PROBIT CURVE
                          LAB=4
124

 MUD CONCENTRATION <
16
     DRILLING  MUD  TOXIC ITT RESULTS
 ADJRESP
  1.0
 0ฃ



 0.6



 0.4



 0.2



 0.0
      CTRL
              ADJUSTED FOR NATURAL RE9PONSIVITV
                          LAB=4
1      2      4

 MUD CONCENTRATION^)
16
32
         Figure IV-5  DRILLING MUD TOXICITY RESULTS: LAB NO. 4
                            23

-------
                 Toble
v-e
             summflRY OF BBBKS
                    FOB
             TEST LHBOBRTOBIE5
                     BY
                  CBITEBIR
                                              296
                  Criterion
Lab
*


1
2
3
4
5
6
7
8
9
10
Lab
vs.
Ref . Lab

3
1
5
7
*
2
4
8
6
9
Lab
vs.
Other
Labs
5
2
1
6
*
4
3
7
8
9


Width
Confidence
Interval
In-Mftl

3
1
7
7
6
2
4
8
9
10
Iff-Ogl

1
5
7
8
9
2
3.5
3.5
10
6
ji

3
6
7
1
10
5
4
2
9
8
0*

4
6
7
1
10
5
3
2
9
8
Not ranked

-------
 I
 m
          1JO-
         0.8-
OJ6-
         0.4-
         02-
                        Ttsl Labs 1,2,3,4,6.7.6
                                                                       297
         OJO
i
         1JO-
         OJB
         OA-
         02-
         OJO
                               Ml
                          EPA Gulf Breeze Lob
                               MU

-------
Tests of Intro-Laboratory Variability IProbit Model, Reference Lab]
Drilling Fluid
Trial C Hu Sigma LogLik
A 0.00 0.87 0.74 -110.55
B 0.08 1.51 0.42 -113.92
C 0.07 0.79 0.78 -133.16
D 0.15 1.47 0.48 -140.87
E 0.01 0.94 0.59 -99.67
F 0.02 0.88 0.82 -128.31
Total -726.48
All* 0.04 1.01 0.76 -764.58
Chi-Sq llntra-Lab] 76.21
Of 15
All** various 1.04 0.73 -757.68
Chi-Sq ilntra-lab] 62.40
df 10
Chi-Sq [common spont. resp.l 13.81
df 5
Reference Toxicant
C Hu Sipa LogLik
0.04 2.01 0.34 -29.50
0.00 2.08 0.09 -13.87
0.10 1.78 0.43 -43.34
0.00 1.80 0.68 -35.47
0.05 1.98 0.10 -24.34
0.04 1.97 0.40 -31.84
-178.36
0.04 1.92 0.41 -197.19
37.65
15
various 1.94 0.39 -189.98
23.23
10
14.42
C
ซ
 * with common spontaneous response rate
 ** with separate spontaneous response rate

-------
                                                                299
Pisbit Model and Summed Replicate Likelihoods

Diilling Fluid
Lab











Total - All
Tstal -75,9,
_ogLik(Model Loglik(Rep)
0
1
2
3
4
5
6
7
8
9
10

10
-110.55
- 1 08.02
-154.31
-174.60
-83.49
-52.53
-131:29
-126.80
-99.6 1
-87.52
- 1 1 0.50
-1239.22
-988.67
-100.57
-97.37
-144.72
-163.65
-73.59
-43.38
-125.96
-117.12
-88.14
-81.74
-100.96
. -1137.20
-911.12
Chi-Sq
19.95
21.31
19.19
21.90
19.79
1 8.30
1 0.66
1 9.36
22.94
11.56
19.08
204.04
155.10
Df
15
15
15
15
15
15
15
15
15
15
15
165
120
Signif
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
#
*
                                                              Notes
"fets of Interlaboratory Differences
CMC-All        -1558.17 vs Total Model  637.90
Fised O Al 1      -1713.35 vs Total Model  948.26
Opt C/5,9,10
FisedC/5,9JO
-1122.19 vs Total Model   267.04
-1161.00 vs Total Model   344.66
20   **
20   **   (Approx)

14   **
14   **   (Approx)

-------
likelihoodTflblefor Inter/I ntrc Laboratory Effe cts
MaterialDrillingFluiil
L8bs:0(A-F)l234678
                                                                   300
Model
Probit*
Probit/Lab*
Probit/Trial
PeiDose/Trial
PerRep/Trial
Complete
Parameters
15
29
39
78
234
4680
Up
-1790.51
-1635.80
-160460
-154L4Z
-1441.52
0.00
Effect
Inter-Lab
Intra-Lab
6-of-fit
Hetero-Rep
Residual

df
14
10
39
156
4446

Chi-Sq
309.42
62.40
112.35
213.81
2883.04

Z
55.83
11.72
8.31
3.27
-16.57

* Spon, Resp final

-------
                                            301
Inter  and Intra-Laboratory  Variation
            A/
 Jbunu-Cu?
             Co SH&mf
                s
              *>•ซซ
                          Sff)
                         2?
ฃ.0
9.9
            /ฃ?*-

-------
                                        302
Covariate models :
   Jl = JiO + jll Xj + J12 X2
log(tf) =
                            X2
   Maximum  likelihood fitting allows
   estimation of parameters, calculation of
   confidence intervals, and testing of
   significance of covariates.

-------
                                        303
Ue measure the variation to know what
can be expected. Ue attempt to explain
the variation so that we can improve
the test method.  Ue can improve the
test method by

   a)  fixing or specifying influential
       factors in the protornlr
       e.g. specify materials used,
           temperature and other
           condidtions,

   b]  introducing procedures that
       compensate for  sources nf
       variation, e.g. mixing,
       randomization,  blinding
   c) numerically adjusting the
      results to remove the
      influence, e.g. the correction for
       spontaneous response.

-------
     2-2



       3
     0-
   -1 -
    7.80
7.85
        LEGEND. LAP
SIGMA
  1.50 -d
  1.25 -
  1.00 -
 0.75 -
 0.50 -i
 0.25 -i
     7.80



LEGEND. LAB
 7.85
                                                                                         304
                           x
7.90
7.95
              4-4-4  0
              <:> c  o  A
              2226
8.00

 PH

y.  x x
A  ฃ A
8.05
                            1
                            5
                            9
8.10
                          * *  *   2
                          ซ *  *   6
                          A *  *  10
8.15
.620
                      D D D
                      V Y V
                        x
7.92
     4-4-4-  0
     C O O  4
     222  8
7.98      8.04

       PH

 XXX  1
          •  • i
          8.10
          8.16
          8.22
                           * * *   2
                           s* ป *   6
                           * * *  10
                        ODD
                        Y  Y Y
                                                                             7
                Figure  VI-4   PLOTS OF  TOXICITY PARAMETERS  VERSUS PH


                                              45

-------
                                                                   305
COVARIATES:  InTSS.ACID.HUrcf   UBS: 0123469
     C:
                 Mil  NPM1  InTM*   IPIIS
                 nu  sibnA  miss   ACID
                                                   L
           P
InTSS
ACID
HUM
MSS,ACID
 0.92
•1.19
 1.20
 0.07
•0.95
•1.19
 0.57
•0.39
1.02
0.83
0.99
0.98
0.82
0.83
0.98
0.83
                             0.23
                             0.22  -0.04
                             0.23
                                  -0.06
                             0.23  -0.09
      -1057.3
       -906.3
      -10315
 0.40  -1030.9
       -902.6
 0.01   -906.3
 0.24  -1028.0
-0.27   -897.8
ACID | InTSS
ACIDlnilref
MSSIACID
ISSlhUref
HUrefllnTSS
InTSSlACID.HUref
ACIDIMSS,
302.0   1
 51.5   1
 52.8   1
309.3   2
302,0   2
 58.5   2
318.9   3
  7.3   1
  5.7   1
257.8   1
249.2   1
  7.0   1
  0.0   I
260.4   I
  9.6   I
 16.9
                                                                  ).017
                                                                  1002

-------
                                                         306
IIUHiTB: MS.ro.HM      UK: 0123465
          fc FIXED
            HU
              P
\m
m
m\m,
                              i)
•1057.3
•897.7
 0.92  1.02
-1.82  IJ5 029
 1.14  1.11
 022  0.66                   0.33  0.18 -1028.4
•1.54  1.78 027 -0.08 -0.02 -0.02          -895.9
•2.10  1.32 0.38 -0.16          -0.27  0.48 -877.8
 0.58  0.94         -0.04 -0.04  021 -0.04 -1025.9
-1.37  0.49 0.34 -0.16 -0.05 0.11 -0.36  0.79
319.02
 56.32
 57.82
     4
     4
 62.84
375.06
  3.8  2 0.151
  5.02
266.62
331.02
  6.52
 39.82
31232
 5222
 1622
                           March 8, 1986

-------
                                                                  307
COVARIATES:  iVflDO.pH
                              LABS: 01254578    C: Fie
none
avgDO          0.73   0.86
avgpH         30.25   0.87
avgDO.avgpH   29.84   0.87
avgDOIavgpH
avgpHlavgDO

CDVARIATES:  avgDOJ
                                      -1159.3
  HYPOTHESIS   HI   SIGHAavgDOaflipH       L  CHI-SQ  DF   P
  3.4  I  0.064
145.9  I  0.000
147.8  2  0.000
  1.9  I  0.170
144.3  I  0.000
                               -3.64
                          0.06  -3.64
-1087.1
                              LABS: 01234678    C: OPTIflli
  HYPOTHESIS   HU   SIGHAiVflDOavopH       L  CHI-SQ  DF   P
                                                 11.8  I  0.001
none            1.50  0.66              -1122.2
avgDO          0.42  0.64  0.16        -1116.3
avgpH        29.64  0.78       -3.55   -1068.4   107.7  I  0.000
avgDO.avgpH  28.15  076  0.13 -3.47   -1064.8   114.7  2  0.
avgDOIai/gpH
avgpHlavgDO
                                                 7.1   I
                                                103.0  I  0.000
                               Morch 8, 1986

-------
                                                 308
TI
-^•H

i
 n
IV
             -tป
             ฃ
      9-
      "v^

      )(
 II


fi

 -I-
              CN
        '—    i
         fc=ป  a
^N


?

*
        z   n
           WB^iX
JP  ^1 1
--,    V
  '    l P  ซx:

 ป•--    ^
 r
ป
      w
                O


                >
                p5
                k>
                r\>
                ro
             P p p p p p p p p p p
             O O O O — Co O O O O O
             cyป o —* iQ —ป 10 04 10 cn -ES. o

                i
             ro o ro —* —"O —• —ปooc
                      — — o4O4O-urocnvQ->j
                   p —    o    —   — —
                    i       ii
                   —ป i   i  — —  i
                                          Co
                   ppppppppp p p
                   ooooocoooooo
                           sj —
                                O — — — O —
                      ro         —   — — *
                             Co
                             en — \o oo o to

                             — rocn — roro
                   O O O O O — * .— o O — —
                                   ii       ii
                             i   i  — —  i  •  —* —
                          K>
                                Osi — ro— '
                                — Co OJ ro
                          ro ro ro ro ro ro ro ro ro ro ro
                                                     m

                                                     n
                                                     o
                                                     -n
                                                     m
                                                  co
                                                     O
                                                     GT>
                                                i
                                               ro

-------
                                                    309
                          MR. TELLIARD:  Our next
speaker this afternoon is John Brown from Battelle.
He's going to describe some work that has been going
on with the Offshore Operators Committee, which is a
group within API.
     We have proposed regulations for offshore oil
and gas, and one of the things the industry has really
appreciated is, of course, we've always been nice to
them, the question of prohibition in the regula-
tions on the use of diesel as a lubricity agent in
the drilling muds.
     One of the problems that we have looked at in
this is how do you analyze drilling muds for number
two diesel fuel for the purposes, as you might want
to say, of enforcement...that's what we would want to
say...since it will be a condition in their permit,
which we've all grown to love and desire.
     These general permits are due out shortly in the
Gulf and there are two proposed in Alaska and one in
California.  So this work is rather important, not
to us, because after all, we can enforce without a
method, but to the industry at hand.  They have taken
somewhat of a different view.  John.

-------
                                                    310
                      JOHN BROWN

                  BATTELLE N.E.M.R.L
     ORGANIC CHEMICAL CHARACTERIZATION OP DIESEL
   AND MINERAL OILS USED AS DRILLING MUD ADDITIVES
                          MR. BROWN:  In recent

years, concern over the impact of toxic diesel oil

components on benthic and pelagic communities sur-

rounding offshore drilling platforms has promoted

government agencies to issue a ban on the ocean

discharge of diesel-containing muds.  In response,

the offshore oil industry has resorted to the use of

alternative lubricants, such as mineral oil, in

drilling muds destined for ocean disposal.  The

selection of mineral oil was based on a presumed

decreased toxicity due to its lower aromatic hydro-

carbon content relative to diesel oil.

     In anticipation of new federal regulations

banning the use of diesel oil in drilling muds,

Battelle New England was contracted by the Offshore

Operators Committee to initiate a multiphase study

aimed at developing an analytical method capable of

measuring the diesel content of drilling muds, as well

as distinguishing between individual diesel oils and

-------
                                                     311




mineral oils in mud formulations based on differences



in their organic composition.




     One method proposed by EPA to analytically




measure diesel oil in drilling muds, what we call the




Top 10 method, involves GC/FID analysis of a drilling



mud extract and subsequent quantification based on



the concentration of the 10 major peaks in the sample



chromatogram relative to the same 10 peaks in a




chromatogram of a reference diesel oil.



     This method appears to be analytically sound and




capable of yielding accurate quantitative results,




as is evidenced by the following slides of GC/MS



reconstructed ion chromatograms, or RIC's, of selected



mineral and diesel oils.  These RIC's are virtually



identical to chromatograms that would be obtained



from GC/FID.



SLIDE 1




     The first slide is an RIC of Mineral Oil A,



ortho-turphenyl is the added internal standard.



SLIDE 2




     This figure represents another mineral oil,



Mineral Oil C.



SLIDE 3




     This slide is representative of a California



diesel oil.

-------
                                                    312
SLIDE 4
     This one is representative of a low sulfur
diesel.
     As evidenced by the last four slides, there
appears to be compositional differences between
mineral and diesel oils which should be easily
distinguished by the Top 10 method.
SLIDE 5
     However, in certain cases, a potential ambiguity
exists, as shown by this slide of Mineral Oil B, on
top, and Alaskan diesel on bottom.  A comparison of
the chromatographic pattern exhibited by each oil
reveals that they are nearly identical, with the
exception of the relative abundance of the added
internal standards ortho-terphenyl.  In fact, our
results show that if the EPA Top 10 method is applied
in this instance, Mineral Oil B is mistaken for diesel
oil.
     These potential ambiguities emphasize the need
to develop tracers of specific additive types beyond
chromatographic pattern matching, which can serve as
definitive1 quantitative indicators of the presence of
mineral and diesel oils.
SLIDE 6
     Nine oils representative of those used in

-------
                                                    313
different offshore regions were analyzed by the best



available methods for the following targeted compound



classes: organic sulfur, total sulfur and total



nitrogen; sulfur-, nitrogen- and oxygen-containing



PAH, or PAC's, carboxylid acids, phenolic acids,



aldehydes and ketones, phenol and its alkyl homologs,



up to C4, individual aromatic hydrocarbons, and total



aromatic content.



SLIDE 7



     PAH analyses were performed using a modified



version of Standard Method D3239.  The results



presented here show the percent total aromatic



content of the nine oil samples.  The mineral oils



are found to have significantly lower aromatic con-



tent, with Mineral Oil A exhibiting the highest value



of 10.2 percent.



     Among the diesel oils, EPA Number 2 fuel oil had



the highest value, 35.6 percent, while the remainder



ranged from 11.7 to 29 percent.



     The total concentration of individual PAH and



their alkylated homologs, through C5, in general



mirror the total aromatic content of each respective



oil.  However, distinct differences were found among



the distributions within the individual oils.  The



mineral oils are found to have signficantly lower

-------
                                                    314




concentrations of benzene, naphthalene and their



alkyl homologs than the diesel oils.



     The differences both in total and individual



aromatic contents of the oils indicate the potential



to use these parameters in future analytical programs



designed to distinguish between mineral and diesel



oils in actual mud samples.



SLIDE 8



     Organic sulfur and total dibenzothiophenes were



determined by GC Hall BCD with dibenzothiophene peak



identifications being made by GC/MS.  The results



show the organic sulfur, total dibenzothiophene and



percent dibenzothiophene concentrations of the diesel



and mineral oils analyzed.  It is evident that the



mineral oils generally exhibit lower sulfur and



dibenzothiophene levels than the diesel oils, with



the exception of the low sulfur diesel.



SLIDE 9



     This figure represents Hall BCD chromatograms of



four of the oil samples analyzed.  Figures A and C,



corresponding to Mineral Oil A and California diesel,



are representative of the two extremes of organic



sulfur content, with Mineral Oil A essentially devoid



of peaks and the California diesel containing many



peaks, including a broad, unresolved hump.  Thianthrene

-------
                                                    315
is the added internal standard.
     The majority of the samples exhibited a distribu-
tion intermediate between these two extremes,
which are represented by Mineral Oil B and Gulf of
Mexico diesel.  Mineral Oil B.  Gulf of Mexico.
SLIDE 10
     Total sulfur was determined by the oxygen bomb
method, Standard Method D129 Modified, and total
nitrogen was determined by the chemi-luminescence
method.
     The sulfur contents of the three mineral oils is
low and differs by an order of magnitude from those
found for the diesels.  The mineral oils also exhibit
a lower nitrogen content than the diesels, although
the differences here are not as great.
     As with the aromatic hydrocarbon compositions,
the large differences in the sulfur contents of the
mineral and diesel oils make this a desirable para-
meter to use in future analytical programs, while
the use of total nitrogen would not be recommended.
SLIDE 11
     The phenol and alkylphenol concentrations of the
nine oil samples were determined by GC/MS analyses.
The important feature of this data is the presence of
phenolic compounds exclusively in the diesel oils.

-------
                                                    316
No phenolic compounds were detected in any of the
mineral oils analyzed.  This is a highly significant
finding which suggests that the alkylated phenols are
another suitable compound class for distinguishing
between diesel and mineral oils in actual drilling
mud formulations.
SLIDE 12
     This figure shows a mass chromatogram map of the
alkylphenols in two of the diesels analyzed.  The
distribution is analogous to those found in the
aromatic hydrocarbons in which the alkylated homologs
are prevalent over the parent compound in any one
series.  You can see, in this case, phenol is absent.
Cl, C2, C3 and C4 phenols occur in greater concentra-
tions, and the same pattern here.
     To sum up the results of Phase I, it was found
that quantitative differences between diesel and
mineral oils are evident in total aromatic, total
sulfur, organic sulfur contents, as well as in the
concentrations of individual PAH.  In addition, phenol
and its alkyl homologs have been identified as a
compound class which has the potential to, by its
presence or absence, denote the occurrence of diesel
oil in drilling mud formulations.
     The phenolic acids, aldehydes and ketones were
not found in measurable quantities in any of the oils,
while the carboxylic acids, the sulfur, nitrogen and

-------
                                                    317
not found in measurable quantities in any of the oils,
while the carboxylic acids, the sulfur, nitrogen and
oxygen PAC's, were detected in most of the oil samples,
but their compositional differences did not allow
differentiation between diesel and mineral oils.
     The purpose of Phase II of this study was to
evaluate the efficiency of two extraction techniques;
retort distillation and solvent extraction, in
isolating the diesel oil organic tracers from actual
drilling mud formulations.  In addition, we hope to
validate the analytical techniques from Phase I when
applied to drilling mud samples.
     The compound classes from Phase I which were
considered to have the greatest potential to
differentiate between diesel and mineral oils were
the alkylphenols and the individual PAH.  Total sulfur
and organic sulfur determinations were not chosen for
Phase II due to possible matrix interferences when
analyzing lignosulfonate muds.
SLIDE 15
     You probably can't see this in back.  I'll try
and walk everyone through it.  This figure represents
the analytical approach used in Phase II.  Briefly,
drilling mud samples were acidified to pH 1, mixed
with sodium sulfate after which the internal standards,

-------
                                                    318
orthoterphenyl and d-8-p-cresol were added.  The mud
formulation was then extracted at ambient temperature
on a shaker table, two times with methanol and two
times with a 9:1 methylene chloride methanol mixture.
     The combined organic extracts were partitioned
versus one normal hydrochloric acid, the organic
phase isolated, and the aqueous phase extracted again
with methylene chloride and ethyl ether.
     At this point, the methylene chloride ether
extract was dried over sodium sulfate and a five
percent aliquot removed for extractable weight
determination and subsequent aromatic hydrocarbon
analysis.
     The remaining extract was partitioned versus
base which was then acidified and extracted with
ethyl ether to isolate the alkylphenols.  Both the
phenols and the aromatic hydrocarbons were analyzed
by the same electron impact GC/MS procedure used in
Phase I.
     The retort distillates, containing the same
internal standards used in the solvent extractions,
were introduced into this analytical scheme just
prior to the acid water partitioning step and were
carried out through the remainder of the procedure in
a manner identical to the solvent extracts.

-------
                                                    319



     The retorts were introduced here and were carried



out identically through the rest of the scheme.



     The results of Phase II indicate that phenol



analysis by high temperature retort is flawed due to



artifact formation and should not be considered.  It



was found that considerable concentrations of phenols



were produced in the drilling mud matrix, even with a



modified low temperature retort distillation.



     It was determined that the solvent extraction



method be adopted for isolating phenols and hydrocarbons



from drilling muds since our preliminary analyses



have shown that this method yields accurate and



internally consistent data which compares favorably



with the composition determined previously in Phase I



for the neat Alaskan diesel.



     In the future, we plan to conduct further



development work to purify the phenol isolates from



the solvent extracts, which would permit direct GC/MS



analysis.  We've recommended that a derivatization



and subsequent absorption column chromatography be



explored, as initial results seem promising.



     In addition, a column chromatography step to



clean up aromatic hydrocarbon extracts prior to the



GC/MS analysis would be recommended since this would



permit better chromatography and subsequently improve

-------
                                                    320
our accuracy and precision.  We also plan to analyze
drilling muds with a mixture of additives, including
crude oil, and analyze drilling muds which have been
hot rolled at high temperature, and ultimately analyze
wild muds from offshore drilling platforms.
     We hope that the final product of this research
will be an analytically validated method for determin-
ing the presence of diesel oil in drilling muds.

-------
                                                    321
             QUESTION AND ANSWER SESSION
                          MR. TELLIARD:  Questions?



                          DR. FRIEDMAN:  Paul Friedman,




EPA.  Have you looked into IR as a method for



characterizing these waste oils or oils in drilling



muds?



                          MR. BROWN:  No, we haven't.




At the time of the study we didn't have IR capabilities.



                          DR. FRIEDMAN:  Thank you.




                          MR. TELLIARD:  No one else?




You're going to let him off this easy?

-------
                                          . 322
ORGANIC CHEMICAL CHARACTERIZATION
 OF DIESEL AND MINERAL OILS USED
    AS DRILLING MUD ADDITIVES
           A.G,  REQUEJO
            J.S.  BROUN
            P,D.  BOEHM

-------
                                                       323
                 OBJECTIVES OF PHASE  I
       TO DEVELP AN ANALYTICAL METHOD CAPABLE OF
  DIFFERENTIATING BETWEEN  INDIVIDUAL DIESEL AND MINERAL
OILS, BASED ON DIFFERENCES IN THEIR ORGANIC COMPOSITION,

-------

-------
325
           '
      fev-

-------

-------

-------
 MO-B-6-84-7
Alaska  Diesel
 GC/MS/OS  Ur:.CONSTRIJCTFin  ION  CHROMATOC.RAMS  (RIC)  OF  THE
 MtNF.RAL OIL  MO-B-f.-g';-? ANH THi:.  M.A.SKAN nif-SIiL.  NOTE;  THf!
 SIMILARITY IN  THI-  CHF^OMATOC.R APHIO  I'ATTF-KN  HXHIIMTfin  BY
 EACH.  o-TF.RPHFlNYL IS THE AHOI-in INTI'KNAL  STANOARO (IS).

-------

-------
                                                         330
 PERCENT AROMATIC CONTENT OF DIESEL AND MINERAL OIL ADDITIVES
SAMPLE
PERCENT
MINERAL OIL
MQ-A-6-84-3
MO-B-6-84-7
MO-C-6-84-19A
  10,2 ฑ0,7
   2,1
   3.2
DIESEL OIL
LOW S DIESEL
HIGH S DIESEL
GULF OF MEXICO DIESEL
ALASKA DIESEL
CALIFORNIA DIESEL
EPA-API NO. 2 FUEL OILA
  16,1
  29,0
  23.8
  11,7
  15.9
  35,6 + 3,9
ATRIPLICATE DETERMINATIONS

-------
                                                       . 331
        ESTIMATES OF ORGANIC SULFUR CONTENT AND TOTAL
    DIBENZOTHIOPHENES CONCENTRATIONS  (PARENT  COMPOUND TO
                     C3) USING GC/HECD,
SAMPLE
                                      DIBENZOTHIOPHENES    ฃ
MIN
ERAL OIL
                                         (UG/6 OIL)
                                                 i
        [II
             0 DIESEL
   -APT RO.~2~FUEL OIL
   rfBrra

-------

-------
cr\ oo "—*
O ox O
CM ^ CM
 I      I
<  c6 u

666
                    Jfi  
-------
                                       334

s
|T
PC
       U U U
      11 ii i \' uucor"
      t——tO  I ' IU7
                        5  KT
                      m
*^
25
mF"
P2
                          •n:
                           :C3rp
                                          m
                -o
                m
                o
                m
                CO
                                          70
                                          O
                                          CO

-------
ON OO ^
rf\ CM fS
oQ CO 6O
< ob u
666

-------
                                 334
LJLJLJLJOC
UOLJLJOC
en
oo
                    Cd
                    en
                    oo
                    mac
                en  men
                cr>
                    r—oo
                                    CO
        ro

-------
                                                                                        335
Gulf of Mexico Diesel
          p—cresol  |  |  m-*-p—cresol
                                           '&ฃ phenols
                                                                   .phenols
California Diesel
                  o- cresol
                                        C-2 phenols
                                                           "1288 SCAN
                                                            28:08 TINE'.
                                                                     C4 phenols
                                                           C^ phenols.       /'
      phenol (absent)-:
                                                            I:288-.-SCflN:-
                                                            28:88"TIME
        GG/MS/DS   MASS  CHROMATOGRAM  MAPS CORRESPONDING TO  THE
        ALKYLATED PHENOLS FOUND  IN THE GULF OF MEXICO  DIESEL AND
        THE CALIFORNIA DIESEL.      •;, /V'"x:  O- •  —" : :^--V.,-V-?.;';^

-------
                                                   . 336
               CONCLUSIONS  OF  PHASE  I

QUANTITATIVE  DIFFERENCES  BETWEEN  DIESEL  AND  MINERAL
OILS ARE  EVIDENT IN TOTAL AROMATIC,  TOTAL  SULFUR,  AND
ORGANIC   SULFUR   CONTENTS,   AS   WELL   AS   IN   THE
CONCENTRATIONS OF INDIVIDUAL  PAH  (BENZENE, NAPHTHALENE,
BIPHENYL,  FLUORENE  AND  PHENANTHRENE  ALKYL  HOMOLOGUE
SERIES).
THE IDENTIFICATION OF  ALKYL  PHENOLS AS A COMPOUND CLASS
WHICH  HAS  THE   POTENTIAL  TO  BY   (ITS   PRESENCE  OR
ABSENCE), DENOTE  THE  OCCURRENCE  OF  DIESEL OIL  IN MUD
FORMULATIONS.

-------
                                                   . 331
              OBJECTIVES OF PHASE  II
TO EVALUATE THE EFFICIENCY OF  TWO EXTRACTION TECHNIQUES
(RETORT   DISTILLATION   AND    SOLVENT  EXTRACTION)   IN
ISOLATING THE  ORGANIC TRACERS  (IDENTIFIED IN  PHASE  I)
OF DIESEL OILS FROM DRILLING MUDS.

-------
 -1 30g Drilling
      Mud
       |  I.  Acidify with !.*ml6NHCI
         2.  Mix with 50g Na2SO$ (3:1 wet weight:Na2SO<,)
         3.  Add internal standards
 Drilling Mud
         I. Extract with methanol (2x, 100 ml each)
         2. Extract wtth 9:1 methylene chloride:methanol (2x, 100 nil each)
         3. Centrifuge after each extraction (3000 rpm, 10 min)
                   Combined
              Methanol/Methylcn
               Chloride Extracts
                         1. Partition v. 100 ml IN HC1
                         2. Isolate organic phase
                         3. Extract aqueous phase with methylene chloride
                           (50 ml) and  diethyl  ether (50 ml)
  Combined
  Methylene
Chloride/Ether
   Extracts
        II. Remove 3% aliquot
        2. Dry over Na2SO$
           Ielectrobalance
       *. GC/MS
            Aromatic
          Hydrocarbons
                                        1.  Partition v. 100 ml
                                            IN NaOH
                                                   1. Acidify with 6N HC1
                                                   2. Extract 3x with diethyl ether (50 mi each)
I                      Combined
                        Ether
                            I. Dry over Na
                            2. Concentrate
                            3. CC/MS
                     Alkyl Phenols

-------
                                                        339
                   CONCLUSIONS OF PHASE II
1.   THE ANALYSIS OF  INDIVIDUAL PHENOLS BY  HIGH  TEMPERATURE
     RETORT DISTILLATION IS FLAWED DUE TO  ARTIFACT FORMATION
     AND SHOULD  NOT BE CONSIDERED,   PRELIMINARY  EVALUATIONS
     OF  THE  FEASIBILITY  OF   A   LOWER  TEMPERATURE  RETORT
     APPARATUS INDICATE THAT THESE ARE SIMILARLY SUBJECT TO
     ARTIFACT FORMATION,

2,   SOLVENT EXTRACTION SHOULD BE ADOPTED AS THE METHOD FOR
     ISOLATING  PHENOLS AND  HYDROCARBONS  FROM DRILLING  MUD
     SAMPLES,    PRELIMINARY  ANALYSES  HAVE  SHOWN  THAT  THIS
     METHOD  YIELDS ACCURATE,   INTERNALLY  CONSISTENT  DATA
     WHICH   COMPARES   FAVORABLY   WITH   THE   COMPOSITION
     DETERMINED PREVIOUSLY FOR THE NEAT ALASKAN DIESEL,

-------
                                                         340
                         FUTURE WORK
1,   CONDUCT FURTHER DEVELOPMENT WORK TO  SUFFICIENTLY PURIFY
     THE PHENOL  ISOLATES OBTAINED  FROM SOLVENT EXTRACTS  TO
     PERMIT  DIRECT  GC/MS  ANALYSIS.    WE  RECOMMENDED  THAT
     DERIVATIZATION   AND   SUBSEQUENT    ADSORPTION   COLUMN
     CHROMATOGRAPHY  OF  THE  PHENOL FRACTION  BE EXPLORED  AS
     INITIAL RESULTS SEEM QUITE PROMISING,

2,   INCORPORATE AN ADSORPTION  COLUMN CHROMATOGRAPHY STEP  TO
     CLEAN-UP  SOLVENT  EXTRACTS  PRIOR  TO  GC/MS ANALYSIS  OF
     AROMATIC  HYDROCARBONS.     THIS  WOULD   PERMIT  BETTER
     CHROMATOGRAPHY OF  THE HYDROCABRON ISOLATES BY REMOVING
     INTERFERING POLAR  MATERIAL. WHICH  IN TURN WOULD IMPROVE
     ACCURACY AND PRECISION.

3.   ANALYZE  DRILLING  MUD FORMULATIONS  WITH  A MIXTURE  OF
     ADDITIVES (INCLUDING CRUDE  OIL).   ANALYZE DRILLING MUDS
     WHICH  HAVE  BEEN  "HOT ROLLED"  AT HIGH TERMPERATURES, AND
     ULTIMATELY ANALYZE "WILD" DRILLING MUDS.

-------
                                                    341
                          MR. TELLIARD:  Our next



speaker is going to talk on critters and the joys




thereof, and how you can use critters to find out



what's really bad around the house.

-------
                                                     342
                  ROBERT C. BARRICK

                   TETRA TECH, INC.
         CORRELATION OF BIOLOGICAL INDICATORS
             WITH CHEMICAL ANALYSIS DATA
                          MR. BARRICK:  Thanks, Bill.

Actually, about a year ago at the last conference

Peggy Knight of Weyerhaeuser Technology Center

gave a summary of some results that they had

been working on using an isotope dilution technique

for very low level organic analyses of marine

sediments.

     Dale Rushneck and I thought I might give

you some presentation on what happened to those data,

which are part of a larger data set for the Commence-

ment Bay Superfund investigation.

     In that study, there were a number of chemical

biological measurements made.  We threw these into a

blender and tried to come out with some fingers

pointed in one particular direction or another.  What

I'm going to talk on today is a little bit about how

that came about, exactly what the requirements were,

and where we ended up.  I'd like to have the slides,

please.

-------
                                                    343
SLIDE 1
     The subject of my talk is specifically
correlation of biological indicators with
chemical analysis data.  These correlations are
not proof of cause/effect relationships, but they're
suggestive of associations that you might see with
chemical and biological data.
     I want to give credit to the Contract Laboratory
Program which was instrumental in providing the neces-
sary- funds to do the chemistry.  The developmental
work for the procedures that we used, for example,
the isotope dilution technique, was a joint effort
between Tetra Tech and California Analytical Labora-
tories.  Additional analyses for metals were done by
Rocky Mountain Analytical Laboratory.  Weyerhaeuser
Technology Center provided additional organic analyses
and other people in Tetra Tech assembled the biological
data.
     Basically, you've got a problem out in the
environment that can be measured.  You've got industry
and other sources putting out chemicals into the
environment.  We know that that's happening.  We've
measured the waste streams, and we can measure chemi-
cals in the environment.
     We also know, for example, that there are a

-------
                                                     344
lot of people that use the marine environment, as  in
this slide, fishermen.  These people can't go down
and see the chemicals.  What they do are see the
effects.
     Some of these effects can be measured.  For example,
in this slide are livers of English sole.  The liver
on the left has visible lesions.  These are actual
malignancies.  The liver on the right is a healthy
liver.
     These are the kind of things that the public
sees.  These are the kind of things that they react
to.
     The next slide is a characterization that came in
an editorial cartoon.  If you can't read in the back,
it talks about how the defendant is admitting that,
yes, she fed her husband Commencement Bay bottom fish
three nights in a row, which led to his demise,
presumably.
     So, this is what the public sees, this is what
they react to, and this is what they want to have
explained.
     Besides the health risk questions, we have other
kinds of problems in the environment; ecological
questions.  These organisms in the next slide are
benthic infauna, or to analytical chemists like me,

-------
                                                    345
these are clams, worms and other slimy things.  But
they react to chemicals, too, and we can measure the
effects, we can get some indices on effects, and we
can relate them to the chemistry.
     So, basically, in the Commencement Bay Superfund
study, we focused on the sediments.  In the sediments,
as seen in the next slide, we can measure the
chemistry on identical homogenates that are also
tested using bioassays.  We can also collect
samples at the same station, same time, for the
benthic infauna (or worms).
     We also looked at fish.  In fish we can look at
bioaccumulation of substances in muscle tissue
and livers, and also look at exactly what's going on
with the pathology of the livers.
     These are all different kinds of indicators.
They are relatively independent, even though they
might be measuring the same kind of thing, and we
need a way to link them together.
     Some of the things we're going to be talking
about later on have to do specifically with the
results of bioassays and infauna.  These bioassays
shown in this slide don't involve the National
Critter.  Instead, we've got the national cousin,
specifically the Rhepoxynius abronius amphipod, which

-------
                                                    346
is used to measure toxicity via mortality.
     We also used oyster larvae embryos as a check
on abnormality of organism development.  That's
considered more of a chronic test.  The abundance of
infauna is another biological indicator; the numbers
per square meter in a particular sample.  And
chemistry/ you guys know"all about that.
     This slide is an example of some kinds of relation-
ships we can see.  On the vertical axis is plotted the
abundance of Praxillella gracilis; that's a worm.
On the horizonal axis is the concentration of pyrene,
a polynuclear aromatic hydrocarbon.
     For the statisticians in the audience, I have
arbitrarily and capriciously drawn a line on this
slide.  I have arbitrarily and capriciously indicated
a point which appears to be a transition from a wide
ranging abundance to relatively low abundance at a
particular concentration of pyrene.  This kind of
relationship is suggestive of an association between
pyrene concentrations and the abundance of this
organism.  It doesn't prove it.  There may be other
chemicals out there producing the effect.  Typically,
in the environment, we're dealing with a chemical
soup and pyrene is just one of the constituents
of that soup.  We need a mechanism to unravel what

-------
                                                    347



things might be problems in a problem sediment, and



a means to actually identify a problem sediment.



Then, once we do that, we can start getting back to



the source and try and do some remedial action.



     The kind of responses you may see in a bioassay



is shown in this slide.  On the vertical axis we have



the response of the oyster larvae bioassay, the circles,



and the amphipod bioassay, the squares.  These



bioassays tend to show the same behavior.  Increasing



mortality and increasing abnormality appear to



occur with increasing concentration of 4-methyl-



phenol.



     There's a very good relationship here...and in



fact, the samples that make up this data set go off in a



gradient from an outfall that is located at the sample



point in the upper right hand corner of the slide.



There's a decreasing gradient away from the outfall.



So, we see decreasing biological effects, decreasing



chemical concentrations, and an apparent relationship



between the two with distance from the outfall.



     What we're left with is coming up with some idea



of how we sort all this information out.  We had



the requirements shown in the next slide.  First,



we needed to develop our relationships from field



data.  Laboratory cause/effect data are sparse or

-------
                                                    348
nonexistent in a lot of cases, and their
applicability to the field has not really been
proved.
     Second, we needed to provide chemical specific
values.  We couldn't just go take a bioassay and say
this is a problem, this industry caused it.  We
needed to have some way to demonstrate a problem;
these are the chemicals we suspect, these are the
chemicals that are associated with this particular
source.
     Third, we also had to integrate several bio-
logical indicators.  We wanted the system to be
driven by statistically significant effects.  In
other words, the effect is statistically different
from what we see in a control area or reference
area which is presumably not contaminated, or contami-
nated very little.
     Finally, we wanted to have something that was
supported by fairly strong evidence, non-contradictory
evidence.  We didn't want to have to stand up in
court and say, this is contaminated and it caused
this effect while at another site the same
concentration was observed and there was no effect.
In this instance, how can you say there's a problem.
So, we wanted it to be fairly noncontradictory so we

-------
                                                    349
would have strong evidence of a problem.
     The next slide is a schematic of what we worked
out that fits these different criteria.  The
upper axis displays the elevation above reference
for lead.   You don't need to pay too much attention
to that.  All it means is that as you go along the
axis, concentrations of lead at the study area are
getting higher and higher relative to concentrations
at our reference area.  On the bottom axis concen-
trations of lead are plotted.
     The chemical data are divided into two general
groups on this slide.  One group of stations had
statistically significant biological effects and the
remaining stations had no significant effects.
Stations indicated by the top bar on the slide had
no significant benthic depressions.  In other words,
the abundances of infauna were similar at these
stations to what we found in the reference area.
There are about 32 of these sites, and what is plotted
is the range in lead concentrations for those
samples.
     This center bar on the slide represents lead
concentrations at all the stations where we
found no significant toxicity by either of our
bioassays.  Over this range of lead concentrations

-------
                                                     350
there was, apparently, for these samples, not a problem
with toxicity.
     The bottom bar on the slide represents lead
concentrations at all of our stations where one kind
of effect or another was observed; either toxicity
or benthic depressions.  There was a range of concen-
trations where we had a fair amount of contradictory
evidence.  For example, we had a sample at 100 parts
per million lead that was toxic, but the lead
concentration in another sample was identical and we
didn't have an effect.  Using these data, we defined
an apparent benthic effect threshold and an apparent
toxicity threshold.  That's the data point, at 300
parts per million lead, for benthos, and 660 parts
per million lead for toxicity, above which we always
observed an effect, a statistically significant
biological effect.
     This is really a pretty simple concept, but
it allows you to say we've got a problem up in
this region.  By itself, lead doesn't explain an
awful lot of data points.  What we're presuming is in
the region below the apparent effect threshold for
lead, we've got stations that are contaminated by
other chemicals perhaps, and that those chemicals
might be causing the observed biological effect.

-------
                                                    351
Likewise, for stations above the apparent effect
threshold, we're not necessarily saying that it's
lead, there's simply an association.
     So what this technique will let you do is
look at all of your chemicals, plotting them
similarly to this slide, and see how many of your
impacted stations can be potentially explained.
In other words, you have some chemicals that are
above their apparent effect threshold at each
of your stations where you have effects.  That's
the real proof.
     What I want you to keep track of in this
slide for lead is the position of the impacted
stations indicated by arrows along the bottom bar.
The stations that are shown above the apparent
effects threshold for lead will jump to below the
apparent effects threshold for 4-methyl phenol in the
next slide.  The data points indicated below the
apparent effects threshold for lead will jump up to
above the apparent effect threshold for 4-methyl
phenol.
     The only difference between the two slides is
the chemical plotted.  Stations indicated by arrows
above the apparent effect threshold for lead had
very high lead concentrations and fairly low or

-------
                                                                   352
               relatively low 4-methyl phenol concentrations.  The
               stations indicated by arrows above the apparent
               effect threshold for 4-methyl phenol had low lead
               concentrations and fairly high 4-methyl phenol
               concentrations.
                    When we do this exercise for all the different
               chemicals, we come up with an apparent effect
               threshold for each chemical, based on the given
               data set.
                    The next slide presents some of these differ-
               ent threshold values for low molecular weight PAH
               (these are basically the two and three ring compounds)
               high molecular weight PAH (these are the four, five
               and six ring compounds); PCBs; and 4-methyl phenol.
               Apparent effect thresholds have been determined
               for many substances, I just put up a few for an
               example.
                    Using the amphipod bioassay, oyster larvae
               bioassay, Microtox bioassy, (this is a bacterial
               luminescence bioassay), and benthic infaunal
               depressions, with the scheme that I just showed in
               the last two slides, we come up with approximately
               five to six parts per million as our threshold for
               low molecular weight PAH.   There is an error in
               the slide, the columns labelled Microtox and
_

-------
                                                    353




benthic infauna should be reversed.



     Basically, around the part per million level



for most of these organic chemicals, most of the



biological effects are explained.  For arsenic, lead



and mercury...to put up some examples of metals...



the apparent effect threshold is around in the part



per million range for mercury and quite a bit higher



for arsenic and lead.



     Now, we can come up with these different values,



but do they really tell us anything?  Can we explain



our data?  Does it say anything about the biology?



     An evaluation of the approach is shown in the



next slide.  This is actually taking not just



Commencement Bay data, but we've recently expanded



this for seven data sets from all over Puget Sound,



approximately 200 to 250 samples.  For the amphipod



bioassay, for example, there were 150 samples that



went into this evaluation.  Using the different



apparent effect thresholds and the different



impacted stations that I showed you in the previous



slides, we were able to account for approximately 54



percent of the significant amphipod bioassays.



In other words, at 54 percent of the stations where



the amphipod bioassay was significant, we had a



chemical that was above its apparent effect threshold.

-------
                                                     354
     When we go down  to  the oyster  larvae bioassay,
100 percent of the  impacted stations were accounted
for/ Microtox 86 percent/ and benthic  infauna 82  .
percent.  So, in other words/ we had fairly high
predictive power.   We had chemicals that were above
their apparent effect threshold at most of our
stations where we had these kinds of biological
effects.  When we looked at only stations with
severe impacts...this is, for example, in bioassays
where we had greater than 50 percent mortality...
we accounted for 92 to 100 percent of  the impacted
stations.  The 92 percent shown in the slide, in
essence, is one station missed by the  technique.
     What this suggests  is that by using this tech-
nique, we have fairly high predictive  power for
identifying potential problem chemicals at biologically
impacted stations.  Now, whether these chemicals are
the actual problem or not we don't know. .But we do
know that they're associated with biological effects,
and it suggests that they are the ones that should be
examined further.
     As in any technique there are several sources
of uncertainty.   The next slide summarizes these
sources.   Primarily, the one we found to be most
important is the classification of bioeffects.

-------
                                                    355



There's a statistical error involved, which in our



data set was set experiment-wise at P = .05.  In other



words, we had a five percent probability of saying



that a station was impacted when it actually wasn't.



The statistical power of this analysis, which is the



percent of stations that you say are not impacted



but actually were, is another story.  It's more



complicated and I won't go into that now although



it was factored into our calculations.



     A second source of uncertainty is the range



between the value that sets the apparent effect




threshold, (that's the non-impacted station with



the highest concentration of the chemical) and the



next higher value, which by definition is always an



impacted station.  The AET may lie anywhere along



the concentration range between these two data values.



     A third source of uncertainty is the reguire-



ment for representative sampling.  If you haven't



sampled your environment very well you're not going



to be able to get enough data points to help sort out



the different possible combinations of effects



and chemical compositions.  If you don't have



representative sampling you're going to have a problem.



     Finally, there's uncertainty associated with



chemical analysis Variability, which we tried to

-------
                                                     356
control, for example, in the organic analyses with
the isotope dilution technique.  Generally, in our
most recent analyses, the first two sources of error
I discussed are the predominant sources.
     The next slide gives an example of the uncer-
tainty, for example, in benthic AET.  The concen-
tration range between the AET and the next station up
with biological effects is the upper confidence
range.  The lower confidence range is determined
as the 95 percent confidence interval based on
potential classification errors for nonimpacted
stations.  What that means, for example, is that if
we statistically misclassify the two nonimpacted
stations with the highest concentrations, the AET all
of a sudden drops to the value of the nonimpacted
station with the third highest concentration.
     In our application we simply estimated the
probability of misclassifying nonimpacted samples,
cranked that through, and derived the concentration
above which the AET lay with 95 percent probability.
Together, these uncertainty estimates give us a
confidence range for the AET.
     Last, then, I want to show you what you can
actually do with AET.  These values have several
potential .uses.   Problem chemical identification

-------
                                                    357
is one use.  In other words, identification of
what chemicals lie above their apparent effect
thresholds in areas that are impacted so you
can start allocating your resources for source
evaluation.  Problem area identification is another
use.  A lot of times we have chemical data and no
biological data.  With these chemical data we can
start making predictions on where effects may occur.
Problem areas can be defined by how far out you have
to go to get everything below their apparent effect
thresholds.
     Next, AET can be used as a screening tool for
biological sampling.  Sediment concentrations that
are above, for example, the lower limit of an AET,
may indicate the need for additional bioassay testing.
You might want to go out and do some other biological
tests as well.
     Also an important use of AET, and for analytical
chemists I think it's very important, as well as
for biologists, is helping to prioritize cause/effect
studies.  There's potentially decades and decades and
decades of laboratory work that might have to be done
to sort out all the potential effect relationships
between each chemical and environmental indicators.
With this empirical evidence we can start prioritizing

-------
                                                    358
chemicals/ to focus on at least the chemicals
with a high probability of being associated with
significant biological effects in the field.
     Finally, what we're having to do, at least in
the Puget Sound region, is to derive interim
regulatory and permitting guidance for sediments.
For example, in the Superfund cleanup program
in Commencement Bay, industries are likely going to
have to get sediment levels below apparent effect
thresholds.  AET also have potential application in
dredge material disposal.  Materials that are below a
certain level can be disposed of in Puget Sound
and materials that are above can't.
     A specific example of the use of AET in the
Commencement Bay area is shown in this last slide.
Using the apparent effect thresholds and biological
data we're able to prioritize our worst problem areas
shown in red.  Sediments in these areas contained
chemicals that exceeded-apparent effect thresholds,
and also exhibited multiple biological effects.
We used the apparent effect thresholds to define the
spatial extent of these problem areas, because we
didn't have biological data everywhere.  The
lowest priority problem areas contained sediments
above AET but had no biological data available

-------
                                                    359
for confirmation of the predicted impacts.



     In essence, we're able to use AET to rank and



prioritize all our stations and provide focus for



the continuing investigations on the Superfund



feasibility study.



     Thank you.

-------
                                                     360
             QUESTION AND ANSWER SESSION
                          MR. TELLIARD:  Questions?
                          MR. EYNON:  Barry Eynon,
SRI.  Just, I thought it was a very interesting study
a'nd I'll be very interested to see it.  Is it out yet?
                          MR. BARRICK:  The final
report for the Commencement Bay was issued in October
or November...October of last year.  It's available
from the Washington State Department of Ecology.
                          MR. EYNON:  Some areas that
I, as a statistician, think that you might even be
able to go further with the data analysis.  Let me
just mention them and see if you thought about them.
One is, what kind of correlation structure did the
presence or absence of the compounds that you were
looking at have?  Were they correlated and did you
account for that?
                          MR. BARRICK:  I want to
make sure I understand your...what correlation struc-
ture are we talking about?
                          MR. EYNON:  The correlation
between the presence of the analytes in different
samples.  In other words, if you had high lead, did
you also have high copper?

-------
                                                    361
                          MR. BARRICK:  Oh, okay.
Typically, when we were looking at these, when we
looked at...let's take a particular station, Station
X.  Only rarely when we had biological effects there
did we find a single chemical that was above its appar-
ent effect threshold.  That did sometimes happen, but
in all the areas where we had real significant effects,
I mean major, major things like almost...azoic
sediments for example, or 100 percent mortality...we
had a number of chemicals that were up there.  And
that's to be expected because chemicals aren't issued
out in a single little packet.  They go out with lots
of other analytes.  That's why we're very careful to
say this doesn't prove cause/effect.  It's simply an
association.
                          MR. EYNON:  Right.
                          MR. BARRICK:  What we are
able to see though, is that chemicals that are above
their apparent effect thresholds in several of these
areas coincidentally are ones that you would expect,
given the local source discharging right into that
area.  We've then got biological effects, we've got
chemistry, the chemistry suggests you're a potential
source, now you go ahead and figure it out.  In
deciding how to handle the identified problem

-------
                                                    362
chemicals, they may take care of whatever's going on.
     It could be that the real causative agent is not
being measured here.  But supposedly, if this technique
is working, it's correlating with this, and supposedly,
if you control for these you will hopefully control
for the causative agent.
                          MR. EYNON:  Okay.  Yes, let
me suggest that I hope you can think a little fur-
ther about ways to cross-adjust for the various
compounds.  I think that might make a...
                          MR. BARRICK:  Some of the
industries think we've thought too far.
                          MR. EYNON:  The second one
is along the same lines, that the different compounds
may have synergistic effects and stuff, too, obviously,
You've probably thought about that.
                          MR. BARRICK:  Yes.  One
thing I guess I would like to point out on that is
synergism is something that was brought up.  There's
no...none of the sediment quality values or sediment
criteria can really take that into account.  It's
very difficult to in a laboratory.  My personal
hypothesis on this is that if we had...if synergism
and other kinds of chemical/chemical interactions were
really driving this in different areas, and weren't

-------
                                                    363




being accounted for by this technique, then we would



expect to have very low success in picking out the




stations.  The fact that we do, suggests that maybe




synergism is going on but it is taken into account in



the setting of these values.



                          MR. EYNON:  Okay.  And my




third point was simply to note that your...it appeared




to me that when you compute your success rate in your




classification, that's taking the same data that you



used to define your classifications.  Am I correct?



                          MR. BARRICK:  Yes.




                          MR. EYNON:' So, is there



some thought to maybe cross-validating by subdividing



your data or going and getting new data to check




those success rates?



                          MR. BARRICK:  We've done



that...yes, we've tried to do that two ways.  One of




them is  that there's no guarantee in  the way the



method is set up that you will have  things fall out.



Simply by using those data, the significant effects



data, to help define these groups doesn't guarantee in



itself that you will have chemicals  falling above an



apparent effect threshold.



     So  the fact that we're going back and checking our



success  isn't really predestined.   In other words, we

-------
                                                    364
aren't going to have real good success simply because
we used those indicators in defining our two categories,
     But we have taken this on.  There's been some
samples, nine samples, collected in the San Francisco
Bay, and using these apparent effect thresholds, it
predicted fairly well.  In the majority of the cases
where we said given that chemical concentration we'd
expect an effect, there was.
     Right now, the status of this is it's being
expanded into a lot of areas in Puget Sound and we
will be adding later this fall a lot of samples to...
another whole data set with synoptic data, and we are
suggesting that we'll take these data and predict
effects in the other data set and then see how it
pans out.
                          MR. EYNON:  Great.  Thank
you.
                          MR. BARRICK:  Thank you.

-------
365

-------
366

-------
367

-------
368

-------
369

-------
370

-------
371

-------
372

-------
373

-------
   St-dS
          91-dS
            Zl-dS

              U-dS
                                      374
Ainviaon a
AinvwaoNav

-------
375

-------
376

-------
..377,

-------
378

-------
379

-------
380

-------
381

-------
382

-------
383

-------
                                                    384
                          MR. TELLIARD:  Our last




speaker for today is going to talk about the



proposed...are you going to talk about the proposed




method?   .                           .            .    ,



                          MR. KIMMELL:  Sort of.'



                          MR. TELLIARD:  Oh.  Almost




proposed?  Leaching test.

-------
                                                     385
      TESTING CONSIDERATIONS FOR THE

TOXICITY CHARACTERISTIC LEACHING PROCEDURE

                   (TCLP)
  NINTH  ANNUAL ANALYTICAL SYMPOSIUM ON THE

 ANALYSIS OF POLLUTANTS IN THE ENVIRONMENT
                     by

              TODD A. KIMMELL
          ENVIRONMENTAL SCIENTIST
      OFFICE OF SOLID WASTE  (WH-562B)
    U.S. ENVIRONMENTAL PROTECTION AGENCY
                MARCH 19,  1986

-------
                                                              386





                           ABSTRACT





                TESTING CONSIDERATIONS FOR THE



      TOXICITY CHARACTERISTIC LEACHING PROCEDURE (TCLP)





     As a result of EPA's efforts to expand the capability of



the Extraction Procedure (EP) leaching test to address organic



components, including volatiles, and also to address.some of



its operational problems, a new leaching test, known as the



Toxicity Characteristic Leaching Procedure (TCLP), has been



developed.  This test has been proposed for use in the Land



Disposal Restrictions Rule and in the expansion of the Extraction



Procedure  (EP) Toxicity Characteristic.  Both actions were



mandated by the Hazardous and Solid Waste Amendments of 1984.



     This paper describes the new leaching test with respect to



its development, evaluation and use.  Data are presented



regarding precision and ruggedness, and several operational



aspects of the procedure are described in detail.  Emphasis



is placed on a comparison of the existing EP to the TCLP, in



terms of additional operational and analytical requirements.



Finally, several aspects of the test are discussed which are



being examined for possible improvement.

-------
                                                               387
                      Table of Contents

I.  INTRODUCTION

     A. Objectives

II.  DEVELOPMENT OF THE TCLP

     A. Disposal Environment and Model
     C. Experimental Design
     D. Results

III.  OPERATIONAL ASPECTS

     A. pH Adjustment
     B. Liquid/Solid Separation
     C. Use of Extraction  Devices
     D. Other  Improvements

IV.  OTHER ASPECTS

     A. Particle Size  Reduction
     B. Treatment of Alkaline  Wastes
     C. Use of a Pre-Screen  Test
     D. Quality Assurance  Requirements

V.   OVERVIEW  OF THE TCLP

     A. EP Comparison
     B. Metals and  Semi-Volatiles
     C. Volatiles

V.   EVALUATION OF THE  TCLP

     A.  Precision
     B. Ruggedness
     C.  Collaborative  Study

 VI. CONCLUSIONS


-------
                                                               388
                TESTING CONSIDERATIONS FOR THE

      TOXICITY CHARACTERISTIC LEACHING PROCEDURE (TCLP)



Introduction

     In carrying out Section 3001 of the Resource Conservation


and Recovery Act (RCRA), EPA identified a number of hazardous

waste characteristics.  One of these characteristics, the

Extraction Procedure Toxicity Characteristic  (40 CFR 261.24),


involved a leaching test, known as the Extraction Procedure,

that is used in identifying wastes that pose  a hazard due to


their potential to leach significant concentrations of toxic
  *                               '                      .
compounds.  When the EP was promulgated, however, the Agency

recognized that, while  it is being used to address the leaching


of several organic pesticides, it was designed mainly to

model the leaching of metals.1  In addition,  EPA has also

become "aware that the EP suffers a number of  operational

problems, that, among other things, adversely affect its


precision.

     As  a result of  these shortcomings, a new leaching test,

known as the Toxicity Characteristic  Leaching Procedure  (TCLP),


has  been developed.  This test  addresses  the  leaching  of  inorganics

and  organics,  including volatiles,  solves the operational problems


of the EP,  and is  also  more  precise  than  the  EP.   On January 14,

1986, EPA proposed this test  for  use  in the  Land  Disposal

Restrictions  Rule  (51  PR 1602), and the  test  will soon be proposed

in an expanded Toxicity Characteristic under  RCRA.   Both  of these



                              -1-

-------
                                                               389
actions were mandated by the Hazardous and Solid Waste Amendments
of 1984.
     This paper describes the developmentr evaluation and use of
the TCLP.  A Background Document detailing the various aspects of
the TCLP is available in EPA's docket.2
Development of the TCLP
     In 1981, through an interagency agreement with the U.S.
Department of Energy's Oak Ridge National Laboratory  (ORNL), EPA
began a research program to develop the TCLP.  The experiments
used to develop the TCLP were set up to conform as closely as
possible with the sanitary waste codisposal model used to develop
the Ep.3'4
     Briefly, industrial solid wastes containing a variety of
organic and inorganic species (target compounds), were loaded
into large glass columns and leached for approximately 3
months with municipal waste leachate generated from adjacent
lysimeters.  The concentration of these target compounds in the
leachate was then measured over time and used to develop
lysimeter leachate target concentrations.  These target
concentrations were established based on both practical
considerations, and the need to represent a mid-to-long term
leaching interval.  This was important, as the purpose of the
leaching test is primarily to evaluate the leaching potential
of chronically toxic contaminants.^
     The same industrial solid wastes were then subjected to
several  laboratory leaching test procedures and leaching*
media, and the concentrations of the target compounds was

                             -2-

-------
                                                               390
determined.  A variety of statistical tests was then used to



compare the various media as to their ability to reproduce



the target concentrationsf and the leaching medium which best



simulated the lysimeter/column was selected for the TCLP.



This medium is a 0.1 N acetate buffer with a pH of approximately



5.0.



Operational Aspects


     As indicated previously, in moving from the EP to the TCLP,



EPA hoped to solve several operational problems that have been



associated with the EP.  First, the EP involves continual pH



adjustment, which is tedious and is also probably the single most



important element in the EP protocol contributing to variability.
  A


Using a pre-defined leaching medium, such as the acetate buffer,



eliminates this problem.



     Second, the EP involves liquid/solid separation using 0.45 urn



pressure filtration.  This separation has proved difficult for



some materials, such as certain types of oily wastes, which have



a tendancy to clog the filter.  This problem is serious, since



materials which do not pass through the filter are operationally



defined as solids, even if they physically appear to be a liquid.



     This problem is particularly  serious for oily wastes, since



oils have been known to frequently migrate to ground waters.  It



is  important for the liquid/solid  separation to treat, as liquids,



those materials which can behave as liquids in the environment.
                                                        •*


It  is also important to recognize, however, that some materials,



such as certain paint wastes, while they have liquid properties,



they will generally behave as solids in the environment  (i.e.,





                             -3-

-------
                                                              391
will not migrate).



     After investigating several options, EPA opted for the



continued use of pressure filtration, but has changed the filter



medium to a 0.6-0.8 um glass fiber filter (nominal pore size).



Use of these filters decreases filtration time,6 improves the



precision of the method, and is believed to also provide a more



adequate differentiation between those materials which behave



as liquids in the environment, and those which behave as solids.



     The third problem involves the use of extraction equipment.



The need to more precisely describe what constitutes acceptable



agitation and to adequately prevent volatilization of volatile



compounds during extraction was critical.



     The EP protocol provides a descriptive definition of what



constitutes acceptable agitation.  Two types of extraction



equipment are identified which were determined to meet this



definition? blade/stirrer, and rotary  (end-over-end) agitation.



EPA has  learned that this  lack of specificity in agitation



conditions is also  a major source of variability,7 and has centered



on the use of rotary agitation, and  further specified an



agitation rate of 30 Hh  2 rpm.  Rotary  agitation is recognized



as a reproducible means of agitation,  and has been incorporated



into several  leaching and  similar test methods.8



     Loss of  volatile contaminants can occur during  liquid/solid



separation, and during  extraction.   With the assistance of



laboratory equipment manufacturers,  EPA  has addressed this



problem through development  of the Zero-Headspace Extractor



 (ZHE).   This  device, pictured in Figure  1,  is designed to




                              -4-

-------
                                                               392
prevent volatilization when conducting the procedure.  The
operation of this device is discussed in a subsequent section
of this paper.
     Due to the need to have the ZHE compatable with common
laboratory equipment, such as off-gassing ovens and laboratory
sinks, and also the need to have a device that is easily handled
by analysts, a device smaller than the 2-liter internal volume
device EPA originally had in mind was necessary.  Balancing the
need to also accomodate as large a sample size as possible, a device
with a 500 ml internal volume is specified.  Due to the 500 ml
internal volume, however, the ZHE can only accommodate a maximum
sample size of 25 grams for a 100 percent solids sample.  For a
waste containing less than 100 percent solids, the sample size
used in the ZHE is a function of the percent solids of the waste.
     Although considerably more expensive than the bottles used in
the current EP, these devices are only required for volatile
components.  Bottles are used for assessing non-volatile components.
While EPA had originally intended the ZHE to be capable of addressing
all waste components, the volume limitations and other constraints
have led EPA to only allow its use when  dealing with volatiles.
     Other improvements designed to reduce complexity and enhance
precision have also been introduced into the TCLP.  For example,
agitation is conducted over  18 hours rather than the EP's 24 hours.
As another example, in transferring samples from container to
filtration apparatus to estractor etc.,  the TCLP procedure calls  for
determining the weight of any residual sample material left behind,
and subtracting this from the total sample size.

                              -5-

-------
                                                               393
Other Aspects

     There are also several other aspects of the TCLP which warrant

discussion.  These are in the areas of particle size reduction,
 =.
treatment of alkaline wastes, the use of a pre-screen test, and

the quality assurance requirements.

     Whereas the EP allowed using what is known as the Structural

Integrity Procedure (SIP) for reduction of monolithic wastes, the

TCLP requires particle size reduction to 9.5 mm for all wastes,

if necessary.  The SIP amounts to pounding monolithic wastes with

hammer-like blows, and was designed to simulate the action of

heavy landfill equipment, which can act to reduce monolithic

blocks into smaller pieces.  In addition to the action of heavy

landfill equipment, however, wastes are also reduced in size by

natural weathering forces, such as wet/dry and freeze/thaw cycles,

which can also act to reduce monolithic wastes.9  By not allowing

use of the SIP, the Agency assures that wastes that are solidified

into monolithic blocks will not leach even if broken down.

     Although the TCLP development work involved 11 wastes and

95 target compounds which leached from these wastes, alkaline

wastes were not adequately represented in the leaching experiments.

EPA believes that an increase in the leaching of inorganic and

some organic species may be observed as the ability of alkaline

wastes to resist the acidity encountered during leaching becomes

exhausted.  Data from the TCLP development program (on a moderately
                                                        ป-_

alkaline waste), and from subsequent studies on waste of moderate
                                                          •
to high alkalinity, demonstrated that the leaching rate of metals

was relatively constant or increasing slightly over liquid to



                             -6-

-------
                                                               394

solid ratios as high as 30 to I.10  The TCLP is carried out at a
20 to 1 liquid to solid ratio, during which most components from
non-alkaline wastes experience a decrease in leaching rate.  This
data demonstrated that the acetate buffer system may not adequately
account for the leaching of some metals from wastes of moderate to
high alkalinity.
     To address this problem, the TCLP specifies a second, stronger
leaching medium for the extraction of wastes of moderate to high
alkalinity.  To define this second medium, the basis behind the
EP's maximum amount of acetic acid, 2 milliequivalents of acid
per gram of waste, is used.1  The acetate buffer supplies only
0.7 milliequivalents of acid per gram of waste.
                              i
     A simple test of waste alkalinity is proposed as a means of
determining the appropriate leaching fluid to use.  This test
involves mixing the solid portion of the waste with water and
determining the pH.  If the pH  is ฃ  5.0, the buffer is used.  If
the pH- is  > 5.0,  a known amount of acid  is added to the slurry
(i.e., 0.7 acid milliequivalents per gram of waste), the mixture
is heated, and after cooling, the pH is  again measured.  If the
pH is <_  5.0, the  acetate buffer is used, and if the pH is  > 5.0,
the stronger leaching medium  is used.11
     While the TCLP  is  expected to be similar  to the EP in cost,
due to the TCLP's increased number of analytes, the overall cost
of analysis of the TCLP extract will be  considerably more  than
the EP extract.   To  reduce  these  costs,  EPA  is proposing to
establish  an optional  prescreen test  in  lieu of the TCLP*  This
prescreen  will  consist of  a total  analysis  of  the  waste using

                              -7-

-------
                                                               395
SW-846 methods,12 to determine if the waste contains sufficient
amounts of individual contaminants for the appropriate -regulatory
threshold to be exceeded, assuming all the compound leaches from
the waste.  If based on such an analysis, one can be certain that
the appropriate regulatory threshold could not be exceeded, then
the TCLP will not have to be performed.
     This option will be especially useful for those generators
who wish to demonstrate that their waste  is devoid of certain
contaminants.  For example, since fly or  bottom ashes resulting
from combustion processes are  unlikely to contain volatiles,
the prescreen might prove a less costly option to performing
the TCLP.                                       ••       .   '
     The quality assurance  (QA) requirements  of the  EP require one
blank  per sample batch, and the Method of Standard Addition  (MSA)
to be  run for all samples.  The Agency has been hearing  for  some
time now that the requirements with  respect  to the use of blanks
is unclear, and has  changed the  requirement  to  specify that  one
blank  should  be run  for every  10  extractions, and that the leaching
media  should  also  be subjected to a blank.
      EPA has  also  received  comment  that  requiring MSA, which is
very  expensive,  is  unnecessary for all situations.   The  Agency
agrees with this.   Accordingly,  MSA is only required when the
measured concentration of a contaminant is close  enough  to the
 threshold,  that matrix interferences could yield  a  wrong ^decision
 regarding the determination of hazard, or that severe matrix
 interferences are preventing the analytical method from accurately
 determining the concentration of the analyte.

                              -8-

-------
                                                                396
      EPA is also adding QA requirements dealing with acceptable
 sample holding times.   While specifying holding times for the  EP
 metals was not critical, since the TCLP involve,? quite a  few more
 organic analytes, sample holding times begin  to take on more
 importance.  The sample holding times  specified are  consistent
 with  those used in the Agency's Contract Laboratory  Program.
 Overview of the TCLP
      The major differences between the EP and the TCLP are
 summarized in Table 1.   The flow, diagrams of  the two methods,
 presented in Figures 2 and 3,  respectively, indicate that the
 TCLP  is a batch leaching test  similar  to the  EP in overall
 conduct.   It is obvious that the TCLP  relies  heavily on many
 procedural aspects of  the EP.   The main differences  are in the
 areas  of  leaching fluid,  filter type,  particle  size  reduction,
 extraction vessels,  agitation,  extraction time,  and  in  the QA
 requirements.   I have  discussed these  differences  in some  detail.
     Precedurally,  then,  the two protocols are  very  much  alike.
 For wastes containing  less than 0.5 %  solids, the waste, after
 filtering through the glass  fiber  filter,  is  defined as the TCLP
 extract.   For wastes containing  greater than  0.5  % solids, the
 liquid phase,  if  any,  is  separated from the solid phase by glass
 fiber filter filtration,  and stored for later analysis.  The
particle  size of  the solid phase is reduced (if necessary),
weighed,  and then extracted with an amount extraction fluid equal
 to 20 times the weight  of  the solid phase.  The extraction fluid
 employed  is  a function  of  the alkalinity of the waste,and "a special
extractor  vessel, the ZHE, is used when testing  for  volatiles.

                             -9-

-------
                                                               397
     Following an 18-hour agitation period, the liquid extract is
separated from the solid phase, again by filtration through a
glass fiber filter.  If compatible (e.g., precipitate or multiple
phases will not form on combination), the initial liquid phase
of the waste  (if any), is added to the liquid extract, and these
liquids are analyzed together.  If incompatible, the liquids
are analyzed separately, and the results are combined to yield
a volume-weighted average concentration.
Evaluation of the TCLP
     The TCLP has been subjected to a single laboratory precision
evaluation for both the bottle extractor (metals and semi-volatiles),
and for the ZHE  (volatiles).  Two wastes containing a variety of
contaminants were used  (i.e., an oily waste and an alkaline waste),
and these wastes were also spiked with various volatile compounds.
     As shown in Table 2, earlier research indicated that the TCLP
was more precise than the EP.4  Although the single laboratory
precision evaluation did not  include a side-by-side evaluation of
the EP, the results for the bottle extractor show the TCLP to be
of acceptable precision.13  Tables 3 and 4 present some of the
results for metals and organics, respectively.  For the most part,
the percent coefficient of variation  (CV)  between triplicate
extractions was  less  than 30  percent.  This  includes  the variability
contributed by sampling variability and analytical variability.
Although sampling  variability was minimized  to  the extent possible,
it is  reasonable to expect a  sampling variability contribution of
between 2 and 5  percent.  Analytical  variability was  in many cases
comparable  to, and in  some cases  exceeded  the total variability.

                              -10-

-------
                                                               398
This observation is significant as the analytical methods used
                                                       •~ ,
are well-accepted and in widespread use.
     The results of the precision evaluation for the ZHE,14 are
not as clearly interpreted.  There are several reasons  for this.
Firstr experience with the ZHE, especially by laboratory technicians
who were familiar with the work, was limited.  Second,  the evaluation
was conducted using several draft protocols, which differed in their
treatment of the ZHE.  Substantial changes were made between the
draft protocols due to experience gained with the device.  Third,
inadvertant errors were apparently made in following the protocol.
For example, whereas the protocol placed a maximum of  25 grams on
the amount of solid material the ZHE could accommodate, considerably
  *                                                     .
more solid material was extracted in the device, providing for a
variable liquid to solid ratio.  Fourth, due to extenuating
circumstances, two laboratories conducted the analytical work,
rather than the intended single laboratory.  It is apparent that
higher concentrations were obtained by one of the laboratories.
     As indicated above, these factors make the precision data
for the volatiles difficult to interpret.  The percent CV's for
the alkaline waste were mostly less than 60 or 70 percent, which is
fair given the nature of volatiles.  It is unrealistic  to expect
any leaching procedure to provide the same variability for volatiles.
as it does for metals and semi-volatile organics.  The precision
data for the oily waste- indicated more variability.  Some of this
                                                        9\,
can be attributed to severe laboratory contamination problems, as
well as to the oily character of the waste, which seems t'o have
dominated the extraction.  Due to the inconclusive nature of

                             -11-

-------
                                                               339
these results, EPA is in the process of another, more extensive
precision evaluation for the volatiles.                  ~
     EPA has also conducted a ruggedness evaluation for the
TCLP.  As with the precision evaluation, ruggedness was evaluated
for both the bottle extractors and the ZHE, and the same wastes
were used.  While the evaluation of the volatiles is currently
ongoing, the evaluation of the bottle extractors has been completed.1
Table 5 presents the parameters which were evaluated for ruggedness
using both types of equipment.
     The ruggedness evaluation for the bottle extractors demonstrated
that for the most part, the TCLP is fairly rugged.  This is
especially true for the semivolatile organics, which for the most
part, were unaffected by the parameters investigated.  For the
metals, the results suggest that two parameters may be critical.
As expected, the acidity of the extraction fluid directly affects
the extraction of metals.  Recognizing this, the TCLP emphasizes
accuracy in the preparation of the extraction fluids, by specifying
the exact recipes for these media, and also indicating that the pH
of these media should be accurate to within +_ 0.05 pH units.
     Bottle type  (i.e., borosilicate vs flint glass) also appears
to affect extract concentrations for the metals.  It appears that
using flint glass can result in significantly higher metal
concentrations.  While acid washing or an expanded use of blanks
may help to resolve this problem, specifying borosilicate over
other types of glass would resolve the problem entirely.  Due
to the  substantially higher cost of borosilicate versus other types
of glass  (i.e., from 3 to 5 times higher), this is one of the

                             -12-

-------
                                                               400
areas where further work may be warranted.
     In addition to single laboratory ruggedness and precision
evaluations, two independent collaborative evaluations are being
conducted, one by EPA and the other by the Electric Power Research
Institute (EPRI).  EPA's evaluation, in which a number of business
associations and individual companies are participating, involves
over 20 laboratories, five different wastes, and both types of
extraction e'quipment.  This study is currently ongoing.
     The EPRI study, which is nearing completion, is very similar
to an evaluation EPRI conducted on the EP.16  This study is limited
to the determination of inorganic species and deals with the bottle
extractors only.  As with the 1979 study, EPRI is investigating
the contribution of both variability in sampling, and variability
introduced through the analytical methods.  Unlike EPA's evaluation,
EPRI is also conducting EP extractions, and will be comparing the
variability of the two methods.
Conclusions
     The development of test methods, whether they be physical
test methods like the TCLP, or analytical methods, is an evolving
process.  As advances are made in technology and as additional data
is gathered and evaluated, test methods become more accurate,
precise, sensitive, rugged, and less costly.  The evolution
from the EP to the TCLP is a prime example of this.
     I have presented a -discussion of the development, evaluation
and use of the TCLP.  Work is continuing to help resolve remaining
problems, and I am sure that future advancements will be made
that will lead to further improvement of the procedure.

                             -13-

-------
                                                               401
Acknowledgements                 '                     "
     I would like to express my appreciation to C. Francis*-
M. Maskarinec, N. Rothman, T. Varouxis and the rest of those
many people who had a part in the development and evaluation
of the TCLP.  Special appreciation is in order for L. Williams
and D. Friedman.  Through the dedication and expertise of all
these people, the TCLP has become a reality.
                              -14-

-------
                                                               402
                      Tables and Figures
Tables
Table 1:  Comparison of EP and TCLP
Table 2:  ORNL Precision Data Comparing EP to TCLP
Table 3:  Precision Data For Metals
Table 4:  Precision data For Semi-volatiles
Table 5:  Parameters Investigated During TCLP Ruggedness Evaluation

Figures
Figure  1:   ZHE Device
Figure  2:   EP Flow Diagram
Figure  3:   TCLP  Flow Diagram
                              -15-

-------
                                                                             403
                                  Table  ฑ

               Conparison of the Extraction Procedure (EP)  and the
                Tbxicitv Characteristic Leaching Procedure  (TCLP)*
    Item

1) leaching Media
2) Liquid/solid
   separation
3 f Monolithic mat-
   erial/particle
   size reduction

4) Extraction
   Vessels
5) Agitation
6) Extraction time

7) Quality Control
   requirements
 EP

 0.5 N Acetic acid added
 to distilled deionized
 water to a pH of 5 with
 400 ml maximum addition.
 Continual pH adjustment.

 0.45 urn Filtration to
 75 psi in 10 psi
 increments.
 Unspecified filter, type.

 Use of Structural
 Integrity Procedure
 or grinding and milling.

 Unspecified design.
• Blade/stirrer vessel
 acceptable.
- Prose definition of
  acceptable agitation.


- 24 hours.

- Standard additions
  required.
- One blank per sample
  batch.
                              TCLP

                              0.1 N pH 2.9 acetic
                              acid solution for moderate
                              to high alkaline wastes and
                              0.1 N pH 4.9 acetate buffer
                              for other wastes.

                              0.6-0.8 urn Glass fiber filter
                              filtration to 50 psi.
- Grinding or milling only.
- Structural Integrity Procedure
  not used.

- Zero-headspace vessel
  reguired for volatiles.
- Bottles used for non-volatiles.
-Blade stirrer vessel not used.

- Rotary agitation only in
  an end-over-end fashion at
  30 ฑ2 rpm.

- 18 hours.

- Standard additions required
  in some cases.
- One blank per 10 extractions
  and every new batch of extract.
- Analysis specific to analyte.
 * All other attributes between the two tests are generally the same, although
  there are seme minor differences.  Note  also that while the EP only addresses
  those species for which National Interim Primary Drinking Water Standards
   (NIPDHS)  exist, the TCLP can be  applied  to other toxicants.

-------
                                                                     404
    TABLE V-.   RANKING OF LABORATORY METHODS BY HEAN  COEFFICIENT OF
           VARIATION (PHASES I and II COMBINED, 95  TARGETS)*
Chemical
class
Organic and
Inorganic
Inorganic
Organic
Ranklna of media

Acetate
27.9
12.6
35.8
bv mean coefficient of

Carbonic add
39.6
21.6
50.8
variation 1%)
Extraction
procedure
39.1
20.4
47.8
No significant differences were observed (P < 0.05) within any
single chemical class.

-------
TABLE 3:

METALS IN API SEPARATOR SLUDGE, BUFFER A

METALS ANALYSIS

UNITS {UB/KL)
405



1


























METAL

— ปL
SB
AS
8ft
BE
B
CD
CA
CR
CO
cu
FE
1C
PB
KG
KN
HO
NI
SE
AB
NA
SR
TL

V
ZN
ZR
1
i
APIE BUFFER A
i : i
: SMPL J SHPL s
t SSAI : SSA2 :
: i !
: :
1 17.3 i
! 1.22 !
: ND :
s 1.01 ;
s ND :
S' 9.22 !
ND :
: - IBB !
S 18.3 i
: 0.103 i
: ND :
I 276 1
J 3.78 !
I NO S
S 39.2 !
: 6.91 :
i 0.076 i
S 90.3 1
! ND !
i ND i
i 125 :
: 1.49 :
: 0.620 :
i ND i
! 0.060 !
'. 360 !
i ND :
^ 	 ! _
S
ts.9 :
0.117 I
NO :
1.02 :
ND !
10.0 i
. ND :
175 i
11.2 :
0.076 !
ND ;
233 1
4.01 I
ND I
38.7 !
6.95 !
0.087 I
76.3 !
ND ;
ND ;
136 :
1.54 S
ND i
ND 1
0.024 !
350 !
ND i
SMFL ! !
SSA3 '. MEAN I
: i
•
15.9 I
0.143 i
ND :
1.02 :
ND :
9.37 I
ND !
179 :
13.7 :
0.059 i
ND :
268 !
3.64 I
ND !
38.3 i
6.86 1
0.07 1
73.4 !
ND :
ND !
122 i
1.59 1
0.660 1
0.013 i
0.042 !
332 t
ND !
... J...
5
16.4 i
0.493 :
NA !
1.02 :
NA !
9.53 !
NA !
iBi :
14.4 ;
0.079 !
NA :
259 I
3.81 I
NA :
38.7 :
4.11 :
•.•78 :
ป.• :
NA :
NA :
128 :
1.54 :
0.64 1
NA I
0.042 i
347 !
NA :
t
SD t
i
•.808 :
f.429 i
M i
0.006 :
NA :
0.414 !
NA J
6.66 '
3.60 t
0.022 :
NA :
22.9 !
0.187 !
NA :
•.45i :
•,ซ5 :
•.009 i!
f.<04 !
MA I
NA 3
7.37 :
o.oso :
NA :
NA :
O.OIB :
14.2 ;
NA :
i
CV I
•
•
ซ
i
•.049
1.28
NA
0.006
NA
0.043 !
NA :
•.037 :
i.250 8
•.280 :
•A:
•.ess:
•.049 :
MA:
C.<012 S
Cu'Oo? :
•o.nm :
tO.115 3
INA :
INA :
C.05B 3
0.032 !
DA 6
NA :
0.429 !
0,041 :
NA t
1
1 i
'.METHOD !
1CV 1 BLANK A!
• i
i i
!
4.94 !
12B 1
NA !
0.568 !
NA I
4.34 S
MA i
3.69 •
25.0 t
28.0 S
•A 1
1.83 1
4.90 S
NA S
1.16 i
0.653 !
U.I J
11.3 1
NA !
NA i
5.77 I
3.25 !
HA 1
HA i
42.9 i
. 4.09 1
HA i
J
1
0.227 !
ND :
ND :
0.313 !
ND :
0.347 i
ND ;
1.40 :
no :
ND :
•.018 S
•.191 t
•.soo :
HO t
o.ue s
MD 1
ND :
u> :
ND :
ND :
3.25 !
ND :
•.530 !
ND ;
, ND :
0.560 !
ND :
' 1
1
1
1
1
1
i !
LAB ! DET. !
BLANK i LIMITS!
; i
i
O.OBO ;
ND :
NO :
ND !
ND :
ND !
0.030 :
0.050 i
ND i
ND :
ND i
o.no :
0.510 :
M) S
0.016 !
. WB 4
liD i
tffi :
KD :
MO i
0.168 !
liD I
0.230 :
NO :
0.036 :
0.167 i
ND :
. {
s
•
0.070 :
o.oso ;
0.004 :
0.040 :
o.oio :
o.oio ;
s
0.040 :
o.oto :
o.oos :
•

O.OBO :

o.ooi ;
o.oso ;
0.002 :
o.oso :
0.040 :

0.070 t
O.OiO 1
0.00? !
0.010 i

o.oso :
I
1
  ND  (NOT DETECTED)
  NA  (NOT APPLICABLE)
  BUFFER A  * PH 2.9
  BUFFER S  * PH 4.9

-------
TABLE
                                                                                              406
SEMI-VOLATILE ORGANIC COMPOUNDS IN STILL LIKE BOTTOM, BUFFER B

UNITS (UB/L)
    ACID/EASE NEUTRAL
       COMPOUNDS
 PHENOL
 ANILINE
52-HETHYLPHENDL
•M-HETHYLPHENOL
la^BIHETHYLPHENQL
{NAPHTHALENE
12-METHYLNAPHTHALENE
SDIBENZOFURAN   .
JACENAPHIHYLENE
,'FLUORENE  •
5PHENANTHRENE
{ANTHRACENE
5FLUDRANTHRENE
SPYRENE
JBISt2-ETHYLHE)WJPHTHALATE
IBEN2QlA)ANTHJyฃENE
JCHRYSENE      *   -
;DI-N-BUTYLPHTHALATE
j

1
SHPL i
SLBB1 I
j
•
17500 5
318 5
1820 5
7040 5
292 !
3480 i
254 !
170 i
635 5
141 !
234 i
32.4 J
24.9 5
15.2*5
ND :
ND :
ND 5
16.7*5
!


SHPL
SLBB2


21800
154
2340
9530
375
4300
340
213
804
171
266
39.8
27.3
14.5*
ND
ND
ND
14.5*

J
SHPL i
SLBB3 5
5
i
18600 5
230 5
1830 !
7250 5
296 5
3970 I
275 5
179 5
670 5
140 5
222 !
27.5 5
23.8 !
14.3*5
ND 5
ND t
ND !
23.8*!
:
STILL LIME
J
5
HEAN 1
i
•
,19300 5
234 1
2000 S
7940 i
321 i
3920 t
290 5
187 5
703 i
151 5
241 i
33.2 I
25.3 5
14.7*1
NA i
NA 5
NA 5
18.3*5
J
BOTTOM BUFFER B
S
i
SD !
1
. {
2230 5
82.1 !
297 5
1380 5
46.8 5
413 5
44.8 1
22.7 1
89.2 i
17.6 i
22.7 5
6.19 5
1.79 !
0.473 !
NA !
NA 5
NA :
4.86 !
	 J
I

CV


0.116
0.351
0.149
0.174
0.146
0.105
0.155
0.121
.0.127
0.117
0.095
0.186
0.071
0.032
NA
NA
NA
0.265

ID


11
35
14
17
14
10
15
12
12
11
9.
18
7.<
3.4
1
1
1
26
	 1 	
ND  (NOT DETECTED)
HA  (NOT APPLICABLE)
BUFFER A * FH 2.9
BUFFER B * PH 4.9

*  INSTRUMENT DETECTION LIMITS ARE DETERMINED BY HL'LTIPLYINB BY THREE THE
STAE'ARD DEVIATION OF THREE SUCCESSIVE INJECTIONS OF A LOW-LEVEL STANDARD
SOLUTION.  DETECTION LIMITS IN THE TABLE HAVE BEEN CORRECTED FOR THE
APPROPRIATE DILUTIGH FACTO* (10:1).  QUANTlTATION LIMITS  IN MANY CASES
CAN BE CONSIDERABLY HI&HER OR LOWER THAN THESE INSTRUMENT DETECTION LIMITS,
DEFENDING GN THE NATURE OF THE SAMPLE.


ICV
11.6
35.1
14.9
17.4
14.6
10.5
15.5
12.1
12.7
11.7
9.45
18.6
7.06
3.22
NA
NA
NA
26.5

DET.
LIKITS
9
2
9
15
3
13
5
17
5
26
12
4
9
17
14
9
2
30

-------
                                                                               407
                                     Table S

                Parameters Investigated During TCLP Ruggedness Evaluation
 Parameter

 1) Liquid/Solid ratio:

 2) Extraction time:

 3) Headspace:  ZHE:

                Bottles:

 4) Medium #1 acidity:
    (milliequivalents acid)

 5) Medium #2 acidity:
    (milliequivalents acid)

 6) Aliquots:
    (taking of aliquots
     directly form ZHE
     for analysis)

 7) Extractor vessel:
 8) Acid wash filters:
 9) Filter Type:
10) Pressurization of
    ZHE during agitation
    (psi)

11) ZHE extract collection
    devices
        TCLP
    Specification

        20

        18

       zero

     variable

        70


       200
 Allowed for ZHE
   in  sane cases
(See TCLP protocol
 in Section VCI)

   (See TCLP
   protocol in
   Section VEI)

    Required
    for metals

    0.6-0.8 urn
    glass fiber

    5-10
    Tedlar bag
    or Syringe
      ZHE
     Device

    19 - 21
    0-5%
   60 - 80
Common Equipment
   (Bottles)

    19-21

    16 - 20
                   20  - 60%
                  190 - 210
   Yes - No
Associated ZHE
Millipore ZHE
    Borosilicate
    Flint glass


    Yes - No
                   Glass fiber -
                   Polycarbonate
   0 - 20
Tedlar bag -
  Syringe

-------
                  Extract Outlet
                                                         408
                     •Filter-
            Waste/Extract ion  Fluid
                      Piston
                                                  Top
                                                  Plate
                                             Body
                                                  VI TON
                                                  o-rings
                                                  Bottom
                                                  Plate
                  Pressure Inlet
Figure ฑ :  Zero-headspaoe Extraction Vessel

-------
                                                              "409
Wet Waste Sample Representative Wet Wast
Contains < 0.5% ^ Waste Sample ^ Contain*
Nonfilu
Solids
(rafale * > 1UU brams r Nonfiltw
1 ^
^. Dry Waste Sample ^
Liquid Solid 	 ^_ ...
_ . •"•••ป Solid
Separation r^"ซ
Lie
Discard
^
uid Partic
> 9.5mm < 9.1
4
Sample Size
Reduction
i
e Sample
> 0.5%
able
L
I *0,id 	 	 „ Liquid Solid
^ - Separation
r
e Size Lie
>mm Monolithic
4
Structural
Integrity
Procedure
r *
^ Store

4
Solid 4-- Liquid Solid Separation
I ,
+ J
Discard
	 Liq
EPE
J
w •

\r
uid
,
u
Ktract
L
Methods
uid
r
it 4ฐC
-2
Figure  2: Extraction Procedure Flowchart.

-------
                                                                             .410
                          FIGURE  3 :  TCLP Flowchart
WET WASTE SAMPLE
CONTAINS < 0.5 %
NCN-FIUTERABIE
SOLIDS
        REPRESENTATIVE WASTE
               SAMPLE
                                 ERY WASTE
                            WET WASTE SAMPLE
                            CONTAINS > 0.5 %
                            NON-FILTERABLE
                            SOLICS
                                    I
 LIQUID/SOLID
  SEPARATION
  0.6-0.8 urn
 GIASS FIBRE
   FIUTERS
DISCARD
 SOLID
                                                SOLID
SOLID
                                            LIQUID/SOLID
                                            . SEPARATION
                                             0.6-0.8 urn
                                             GIASS FIBRE
                                               FILTERS
                                               LIQUID
                                               ~~~i
                                              STORE AT
                                                4ฐC
                     REDUCE PARTICLE SIZE IF >9.5 mm
                         OR SURFACE AREA <3.1 crni2
                             TCLP EXTRACTICN*
                                 OF SOLID
                           0-HEAD3PACE EXTRACTOR
                           REQUIRED FOR VOIATIIES
                                     1
                               LIQUID/SOLID
                                 SEPARATION
                              0.6-0.8 urn GIASS
                               FIBRE FILTERS
                                   DISCARD
                                    SOLID
                                    LIQUID
                                  TCLP EXTRACT
            TCLP EXTRACT
            ANALYTICAL
             METHODS
            	 TCLP EXTRACT	
 * The extraction fluid alloyed is a function of the alfcalinity of the
   phase of the waste.

-------
                                                                 411
                          References
(1)



(2)


(3)



(4)



(5)
 (6)



 (7)



 (8)




 (9)



(10)



(ID



(12)
U.S. EPA.  Background Document, Section 261.24, Character-
istic of Extraction Procedure Toxicity.  National Technical
Information Service (NTIS) PB 81 185-027.  May, 1980.

U.S. EPA.  Background Document, Section 261.24, Toxicity
Characteristic Leaching Procedure.  March 10, 1986.

Francis, C.W. et al.  Mobility of Toxic Compounds Prom
Hazardous Wastes.  National Technical Information Service
(NTIS) PB 85 117-034.  August, 1984.

Francis, C.W. and M. Maskarinec.  Field and Laboratory Studies
in Support of a Hazardous Waste Extraction Test.  Oak Ridge
National Laboratory Report No. 6247.  February, 1986.

Kimmell, T.A. and D. Friedman.  Model Assumptions and Rationale
Behind the Development of EP-III.  Fourth ASTM Hazardous and
Industrial Solid Waste Testing Symposium.  Proceedings.  J.K.
Petros, W.J. Lacy, and R.A. Conway, Eds.  ASTM Publication
Code Number  (PCN) 04-886000-16.   1985.

Energy Resources Co.  Filtration  of Various Wastes Using
Various  Filter Media.  U.S. EPA Contract 68-01-7075.  April,
1985.

Brown, O.K.  et al.  Mobility  of Organic Compounds From
Hazardous Wastes.  National Technical  Information Service
 (NTIS) PB 83 163-956

American Society For Testing  and  Materials  (ASTM).   Committee
D-34 Draft Method.  Method  for 24-Hour Batch  Type Distribution
Ratio  (DR) For Contaminant  Sorption  on Soils  and Sediments.
ASTM D34.02-022RO.  Philadelphia,  PA.  1985.

Hannack, P.   Personal  Communication  From P.  Hannack,  Canada
Ministry of  the  Environment,  Alberta Research Centre, to T.A.
Kimmell, U.S.  EPA Re:   TCLP.   October 28,  1985.

Francis, C.W.  and M. Maskarinec.   Leaching of Metals From
Alkaline Wastes  by Municipal  Waste Leachate.   Oak  Ridge
National Laboratory Report.   January,  1986.

 Energy Resources Co.   Extraction Fluid Study and Development
 of an Alkalinity Test  For The Toxicity Characteristic Leaching
 Procedure.   U.S. EPA Contract 68-01-7075.   February, 1986.

 U.S.  EPA.   Test Methods For Evaluating Solid Waste - Physical/
 Chemical Methods.   Second ed.  Government Printing Office
 (GPO)  055-002-81001-2.  EPA SW-846.   Washington D.C.  1982.
                              -16-

-------
                                                                  412
(13)   S-Cubed.   Precision Evaluation of the TCLP Protocol  For
      Non-Volatile Components.   U.S. EPA Contract 68-03-1958.
      January,  1986.

(14)   Francis,  C.W. and M. Maskarinec.   Precision Analysis For
      the Zero-Head Extractor.   Oak Ridge National Laboratory.
      January,  1986.

(15)   Energy Resources Co.  Evaluation of Bottle TCLP Draft
      Protocol.  U.S.  EPA Contract 68-01-7075.   February,  1986.

(16)   Electric Power Research Institute (EPRI).   Proposed  RCRA
      Extraction Procedure:  Reproducibility and Sensitivity.
      Palo Alto, CA.   November,  1979.
                              -17-

-------
                                                    413
             QUESTION AND ANSWER SESSION
                          MR. TELLIARD:  Questions?



                          AUDIENCE PARTICIPANT:  You




mentioned you got lower results for the borosilicate



glass.  That was for metallic ions, I presume?



                          MR. KIMMELL:  That's correct.



                          AUDIENCE PARTICIPANT:   So




isn't that probably due to the ion exchange capacity



of the glass?




                          MR. KIMMELL:  I don't think



we can conclude yet what it's due to.



                          MR. BIRRI:  John Birri,



EPA.  What's the cost of one of these units?




                          MR. KIMMELL:  The price



ranges, depending on which company you buy it from,




I think from $1,200 or $1,300 to $1,500 or $1,600.



                          MR. BIRRI:  I see.



                          MR. KIMMELL:  The primary



differences in the cost of running the method will




be in the additional analytes that we're going to be



asking for.



                          MR. IMBUR:  Bill Imbur, Law



Environmental Services.  I've just gone through this



process.  The Rotary extractor is about $2,100 and the

-------
                                                     414
Zero Headspace extractor  is  $1,500  on  the  current
market.
     I have a question  for you, Todd.  How soon do you
expect this to become official, the methodology?
                          MR. KIMMELL:  The method is
being used in two actions under RCRA.  First of all,
the method has been proposed for use in the Land
Disposal Restrictions Rule, and in  addition, we intend
to propose use of the method in an  expanded toxicity
characteristic under RCRA within a  month.   It could
be later than that, but it should be within a month.
I've been wrong on my predictions before.
                          DR. LICHTENSTEIN:  Harris
Lichtenstein, Spectrix, Houston.  I'm trying to
understand the concept of the volatile organic addition
to the EP toxicity test.  Is that how it should be
understood?  You have a new method  because  we want to
understand more about the VGA's?
                          MR. KIMMELL:  We  want to be
able to...first of all, the characteristic  of EP
toxicity is used in seeing if a waste poses as a
hazard due to its potential to leach primarily
inorganic compounds.  We've developed the zero
headspace extractor specifically to address
the leaching of volatile organic compounds.

-------
                                                     415
                           DR.  LIECHTENSTEIN:   Do you



have  any data  that would  allow one  to  design a mini-




extractor  in a VGA bottle...let me  just  think out



loud...to  where you basically...




                           MR.  KIMMELL:   We  considered



that  option.   Specifically, we considered taking




large VGA  vials and putting the waste  in there with



an amount  of extraction fluid.  I think  what it



amounted to was two grams  of the waste.




      I think most of you are aware  of  the problems



that everybody has with respect to  representative



sampling.  I think everybody agrees that 100  gram



sample size that the EP and the TCLP involved is...it



still doesn't  get you to what  you might  call  a repre-




sentative  sample.  Several samples  are generally



evaluated.  But going down to  a two gram sample




might pose additional representative sampling




problems.  That's why we opted  to develop the  zero



headspace extractor.




                          MR.  TROIANO:   Jeff Troiano,



Ford Motor Company.  A strict reading of the proposed



procedure would seem to indicate that you are  not to



use the ZHE extract for metals  analysis.




                          MR.  KIMMELL:   Yes, and



there's a good reason for that.  Due to  the 25 gram

-------
                                                    416
maximum sample size that the device can accommodate,
what you've got is an extract that's less than 500
milliliters.  EPA's analytical methods in SW-846
require a one liter sample size for just the analysis
of semivolatiles.
                          MR. TROIANO:  What if you
know from a total analysis you have no semivolatiles
of any consequence in your sample?
                          MR. KIMMELL:  I'm sorry,
can you repeat that?
                          MR. TROIANO:  You spoke
earlier of a screening procedure whereby you analyze
a sample on a total basis to determine whether or not
you even have to consider certain compounds.  Once
you've eliminated semivolatiles as being of concern,
you no longer have to worry about a volume constraint
for those parameters.
                          MR. KIMMELL:  That's true.
Again, the method is written, as was the EP, to apply
to all wastes, so again, if you've got a better
idea...we've exhausted a lot of ideas though the
development of this method, but if you've got a
better idea on how to deal with it, we'd be glad to
hear it.
                          MR. TROIANO:  Let me clarify

-------
                                                    417




what I've said.  You've got your sample.  You perform




a total analysis for metals, semivolatiles, volatiles,



and...




                          MR. KIMMELL:  I think I




understand what you're saying.  You're saying you



don't want to take a look at semivolatiles...



                          MR. TROIANO:  Anymore.




                          MR. KIMMELL:  You don't




want to take a look at them anymore, you just want




to evaluate metals.  Why can't you use the ZHE?



                          MR. TROIANO:  Metals and



volatiles.




                          MR. KIMMELL:  Yes, and



again I invite you to comment on that in comments...



                          MR. TROIANO:  We did.



                          MR. KIMMELL:  ...on the




proposed rule.  Then I'll be reading them.  But




there's also another reason why it's not a good



idea to use different devices.  We have found that



when you use different devices you get different



results.  To reduce the variability of the method, it



makes sense to have one device, one set of operating



conditions.  So that's another reason for going to




just the use of one device over another.



                          MR. TROIANO:  Would it be

-------
                                                     418
that much more difficult to validate the ZHE extracts
versus the non-ZHE extract to show that the results
are basically equivalent, if they are?
                          MR. KIMMELL:  In my opinion,
I think that you would find that the results are
different, for various reasons.  One thing, the TCLP,
as applied to metals and semivolatiles, uses conven-
tional gas pressure filtration to separate the liquid
from the solid phase of the waste.  The volatiles
procedure with the ZHE uses piston pressure.  Those
two methods of liquid/solid separation can produce
different amounts of liquid; in other words, a different
result for a percent solids for the same waste.
     So, I think there's a lot of give and take that
went along with the development of this method.  For
example, we originally wanted to use the ZHE for
everything, that's why we developed the first devices
with a two liter sample size.  But then we ran into
the problem that the device was so large that many
analysts, unless they're 6'6" and 240 pounds couldn't
deal with it.  You couldn'.t fit it in laboratory
sinks and you couldn't fit it in an off-gasing oven,
so we had to go down to a smaller size.
                          MR. TROIANO:   I just see a
tremendous amount of duplicative efforts, because I

-------
                                                    419

can see a lot of people might have wastes and...

                          MR. KIMMELL:  Yes.  In fact,

as part of our collaborative study...again, as I

said, we had a lot of input from various labs who

gave their time and their expertise to us, and one of

their comments from the very beginning was, you've

tried to do too much with the same procedure.  By

developing the zero headspace extractor to apply to

all components, you're not taking a look at what it's

really good for, and that's volatiles.

     So, basically, their comment that EPA is trying

to put too much into one device led us to separate

the volatiles from the metals and semivolatiles in

this procedure.

                          MR. TROIANO:  Thank you.

                          MR. TELLIARD:  We're running,

as usual, late, and I'm going to have to try to bring

this to closure because we only have an hour between

getting everybody out of here and getting back for

the murder mystery dofungus.

     Todd, thanks so very much for your efforts.

Nine o'clock tomorrow morning.  Nine.

(WHEREUPON,  the proceedings were continued to March 20,
1986.)

-------
                                                    420
                          MR. TELLIARD:  Could I
have this morning's speakers please come up and take
a seat and spread the fire, so to speak?
     A number of folks have asked about a list of
analytes as it relates to the Groundwater Monitoring
Strategy.  A copy of the new Appendix VIII guidance
can be had...had is a good term...through the RCRA
Hotline, which has a number which is 800-424-9346.
They will deliver it in a small brown paper bag.
800-424-9346.  There's also a notice of availability
in the February 14th Federal Register, but as we pointed
out earlier today, or yesterday, you really can't trust
anything printed in the Federal Register.
     There's one small change on the program.  We're
talking about...instead of most of those things we
have listed, we're going to talk about animal breeding
and habits.  Our last speaker from Duluth didn't make
it.  I don't know if he even knows where Norfolk is.
He's up there counting snowflakes.  Ray Maddalone
will be giving a talk on some metals analysis that
EPRI has been doing, which we heard a little bit
about last year.
     Next year, folks, is the Big Ten.  God willing
that I still be here...I would like someone to jot
some notes down.  Any ideas you have for speakers or

-------
                                                    421




any ideas you have for what we can do for our Tenth



Anniversary...I have a few...I would certainly



appreciate.  I would really like to have a nice show



next year...not that this isn't a nice show and the



last one wasn't a nice show...but I want to blow out



the stops.  So any ideas on speakers, papers or



volunteers for papers, I certainly would be very



receptive, and remember, it's only money, particularly



the folks in the Contract Lab Program, it's only



money.



     Our first speaker this morning is a man who



doesn't do metals.  And there's a reason.  Bob Beimer



is with S-Cubed.  He couldn't hold his job with TRW



so he's now with S-Cubed.  He is now going to talk a



little bit about one of our favorite subjects, isotope



dilution, as it relates to pesticides.  Now, he does



do pesticides even though he doesn't do metals.

-------
                                                     422
                    ROBERT BEIMER

                       S-CUBED
    DETERMINATION OF PRIORITY POLLUTANT PESTICIDES
              BY ISOTOPE DILUTION GC/MS
                          DR. BEIMER:   I'm really

happy to see such a nice turnout here this morning

after what most of you probably did to  yourselves

last night.  The guy I've got turning my slides really

hurt himself last night, so if he stumbles over them

a little bit we'll take that into consideration.

     I left some reports in the back.   I notice

they've all been picked up.  They were  supposedly on

sale.  So those of you who picked them  up, just leave

me $5 each and we'll take care of it.   I will sign

them if you'd like.

     A lot of what I'm going to present here today

is...

                          MR.  TELLIARD:  Benign.

                          DR.  BEIMER:  That too, but

there's a lot of data, and the report really goes

into more detail on listing a lot of this stuff.  I

don't like statistics and there's a lot of that in

the report as well.  So, if any of you would like a

-------
                                                    423




copy of the report, if you'd leave me one of your



business cards during the break I'll get it for you.



It is all right if I do that?



                          MR. TELLIARD:  Oh, yes.



                          DR. BEIMER:  Okay.  To give



you a little historical background of what we're



doing, the isotope dilution method was developed, oh...



Bruce, how long ago?  Five years?  Nine years?  Six



years ago.  One of the movers and shakers of that



development was Bruce Colby.



     What we're doing here is really just an extension



of that method, which has come to be known as Method



1624 and 1625.  Specifically, the 1625 portion of



that is the extractable part, the extractable priority



pollutant determination by isotope dilution.  We've



pretty well copied that for this determination of



pesticides.



     Going into this I think we all knew that the



procedure was going to work.  That wasn't really a



question.  It was how well and what kind of interfer-



ences might one expect from chlorinated priority



pollutant pesticides which, generally speaking, have



somewhat messy mass spectra.



     The procedures that we followed for this method



validation were provided by Dale Rushneck and the

-------
                                                                   424



               money to do this work was provided by Bill Telliard.



               Hence, that's why I'm here today talking about this.



               They're not going to stand and clap.



                    I would like to summarize just a little bit,  if



               I could, what we observed during the method, and then



               we'll get into some of the slides and I'll show you



               some of the numbers.



                    Certainly, the use of a labeled pesticide in



               calculating the concentration of an unlabeled pesticide



               provides for a significant reduction in the relative



               standard deviation of the measurement itself.  We



               know that isotope dilution corrects for recovery,  and



               certainly we observed that in the data.



                    Interferences, although there are some, are



               rather minimal.  I'll get into a little more detail



               on that.  The detection limits that we were able to



               achieve by the method approach 200 to 400 parts per



               trillion in relatively clean water samples.  There



               are some exceptions to that, which I will also talk



               about.



                    The method itself involves the extraction of two



               liter water sample with the pH held in the range of



               six to eight.   The sample is extracted with methylene



               chloride in a continuous liquid/liquid extractor



               concentrated to one milliliter by Kd.   Pass that
_

-------
                                                    425




extract then through an anhydrous sodium sulfate



column to dry it, reduce the volume to less than 200



microliters, add internal standard, bring the volume




back up to 200 microliters and analyze the sample by



GC/MS.



     If you could put the first slide up, Bruce.



SLIDE 1



     The analytical conditions for the determinations



that we're making.  The GC, we're using a DBS capillary




column, helium carrier gas, of course...mass spectro-



meters like helium...30 centimeters a second linear




velocity, using a splitless injector and a silanized



quartz liner net injector.



     For those of you who have done a lot of pesticide



determination, you know the importance of clean



injectors and properly cared for injectors, because a



lot of these pesticides get lost right there at the




front end, and you don't analyze them at the tail end



if you can't get them through the injector.



     This was operated in a splitless mode.  Column



temperature was 100 to 280 at eight degrees a minute



with a 10 minute hold at the column maximum.  We



found that this temperature program provided for ade-




quate separation of the various pesticide components



and it kept the analysis time to a reasonable minimum.

-------
                                                     426
      The mass spectrometer system we used was a
 Finnigan 1020;  typical  scan range, one second per scan,
 and  the  instrument  was  tuned to the DFTPP specifica-
 tions of Method 1625B.
      We  used  difluorobiphenyl as the internal standard.
 We initially  were going to use D-10-phenanthrene,  like
 everyone else,  but  there was a lot of interference
 problems with D-10-phenanthrene with the  various
 pesticide materials  that we were analyzing  for, and
 the  2,2-prime difluorobiphenyl eluted in  a  relatively
 clean portion of the  spectrum and showed  no inter-
 ference.  The internal  standard,  of course,  was used
 for  the  calculation of  recoveries of the  labeled
 compounds.  The unlabeled compounds were  calculated
 relative  to their labeled analogs.
      What we  established  during  the course  of  this
method development was  the  calibration range  of the
 instrument, calibration  linearity,  instrument  detection
 limit, analytical range  of  the  instrument,  analytical
detection limit, method  detection  limit, calibration
reproducibility, and the  precision  and recovery of
the method.  We also have a  limited  amount of data on
interference problems.
     The first step in doing an isotope dilution method
evaluation is to determine the masses that one is going

-------
                                                    427
to use for quantitation.  We chose a set of primary
masses for quantitation which exhibited the minimum
amount of interference between the labeled and
unlabeled analogs, which generally coelute, and with
the other pesticide materials that would be in the
standard mixtures.
SLIDE
     In addition to giving those various masses on
this slide, it also lists the priority pollutant
pesticides and the labeled analog which were used in
this determination.  Not all of the labeled pesticides,
obviously, were available.  Most of these were
deuterated compounds, although there were some carbon
13 labeled compounds as well.  We only did the quanti-
tative determination on these primary masses.
     I'm going to step back on my soapbox here a
little bit and make some suggestions.  When one is
dealing with a limited number of analytes like we're
dealing with here, it would probably not be a bad
idea to choose a secondary set of quantitation masses,
repeat the quantitation, and determine if there's any
significant difference between quantitation at two
different mass numbers in a given GC/MS run.  As far
as I could tell, this would probably only double the
cost of the analysis, but I think it's a wonderful

-------
                                                    428



idea.



     Basically, we've got some pretty sophisticated



computer systems out there tied to these mass



spectrometers and when we're dealing with a limited



set of analytes it would make sense to make this



secondary measurement to determine if there was



interference problems occurring at your primary



quantitation mass.



     Computers could make stupid decisions on these



as well as a human can.  Determine if the difference



between the two numbers is significant and if so,



perhaps a tertiary mass or even a fourth mass could



be used in that quantitation, driving the price up a



factor of four.  Then we could proceed to let the



computer make the decision as to whether or not there



was an interference problem, choose the two masses



which were interfered with the least...in other words,



they gave both the same number...and then that number



presented.



     We've gone to a lot of trouble in determining



the accuracy of the fit of the library and the quality



of which that library is fitted to the analyte spectrum



in a run, but we haven't gone to a lot of trouble in



determining whether or not the quantitation mass is



interfered with for quantitation purposes.

-------
                                                    429
SLIDE
     In these various samples, we also spiked chloro-
dane toxophene and PCB 1254.  The purpose of these
initially was to try to do the analysis of these
materials also by isotope dilution.  As it turned
out, it was more of an exercise in determining what
kind of interferences these mixture pesticides served
to be on the single component pesticides.
     For those of you who have analyzed, especially
toxiphene and chlorodane, you know that the chromatogram
looks like a forest and the mass spectrum is certainly
the  trees, because these isotope patterns are
throughout, and they certainly cause interferences.
SLIDE
     It was our determination that these interferences
were not significant as long as these concentrations
of the mixture pesticides were in the neighborhood of
a factor of  10, at least not more than a factor  of 10
greater than the single component pesticides.
SLIDE
     These are the observed  interferences that  I'm
talking about.  The top one  there says that hepta-
clor is a component of chlorodane.  That's obviously
an  interference we're not going  to be able to overcome.
If  the  chlorodane  concentration  is high  enough,  you

-------
                                                    430
will get, obviously, a positive interference with
heptaclor.
     The chlorodane interferes also with the labeled
and unlabeled forms of alpha-endosulfan and with the
labeled alterant.  Toxophene interferes with D4-endo-
sulfan as well as the unlabeled endosulfan, and PCB
1254 interferes with DDE and alpha-endosulfan...excuse
me, the labeled DDE alpha-endosulfan and labeled
alpha BHC.
     Again, these concentrations have to be significant
for these interferences to be a real problem, but if
you do detect significant quantities of the mixture
pesticides you're not going to be able to get good
quantitative determination using isotope dilution on
single component pesticides in that sample.
SLIDE
     The next step in the method validation was the
determination of the calibration range and the
estimated instrument detection limit.  To do this, we
have analyzed standards covering...well, including
concentrations of .1, .3, 1, 3, 10, 30, 100,  300 and
1,000 micrograms per milliliter.  In these mixtures,
the labeled pesticide concentrations were maintained
constant at 50 micrograms per milliliter and  the
internal standard at 100 micrograms per milliliter.

-------
                                                    431



At each of these concentrations where the pesticide



was detected, we calculated the relative response



factors, the relative standard deviation for the



relative response factors.  Again, all of this



data is presented in the report.  It covers a number



of pages.  Suffices to say that the estimated detection



limits were reasonable, as illustrated, a portion of



which on this slide.



     Observation-wise, we were unable to detect the



unlabeled pesticides at the concentration levels of



.3 and .1 micrograms per milliliter...not terribly



surprising...and that we had some real linearity



problems with a number of the pesticides at the 1 and



3 micrograms per milliliter range.  This is a range



in which a lot of absorption is occurring and we get



fall off in the relative response factor as the con-



centration approaches these levels.



     The estimated detection limit for the instrument



was determined at that point in which the area for



the extracted ion current profile for the quantitation



mass was 1000; 1000 is a fairly arbitrary number



based on the use of an Incos data system.  If one is



using a different type of mass spectrometer and data



system they would have to establish this number for



themselves.  But on the Incos, that level of 1000 for

-------
                                                     432
a peak area is fairly reproducible and you're not
getting a lot of statistical variation in that
measurement itself.  And we needed some arbitrary
limit to calculate the detection limit.  That's what
Dale chose and that's what we used.
     At the upper end, DDE, DDT and ODD all saturated
the mass spectrometer at the 1000 microgram per milli-
liter.  The rest of the components did not saturate
the mass spectrometer, but they certainly did overload
the GC column.  All of the peaks at that level showed
a decrease in resolution and began looking something
like bananas, and interference problems would certainly
increase dramatically because of the lack of separation
that one would achieve.  May I have the next slide,
Bruce.
SLIDE
     This is just a continuation of the estimated
instrument detection limits.  I left the other one up
there longer because the numbers were smaller.  These
numbers are somewhat larger.  Beta endosulfan there,
you see, is a 30 microgram per milliliter and endo-
sulfan sulphate at 24.  Obviously these materials are
not detected at the lower concentrations as well.
SLIDE
     The next determination that was done was the

-------
                                                    433




calculation of the method analytical range and the



estimated detection limit for the method itself.  To



do this we spiked seven 2-liter water samples with



pesticides of the concentrations of lf 3, 10, 30,



100, 300 and 1000.  Again, we dropped off the .1 and



.3 because we were unable to detect the pesticides at



those levels and spiked the rest of the levels.  The



labels were maintained at 50 and the internal standard



at 100.  These calculations are based on the con-



centration in the final extract after it had been



concentrated down to 200 microliters.



     Again, the calculation of estimated detection



limit is based on an extrapolation to an extracted



ion current profile of 1000 on the Incos data system.



SLIDE



     Basically, the estimated method detection limits



paralleled the estimated instrument detection limits




reasonably well.  We expected that the recovery would



be above 50 percent, and most of these numbers reflect



that.



     Observation-wise, again at the high concentration,



the 1000 level, we're saturating the GC column, or



overloading the GC column, and in many cases saturating



the mass spectrometer.  At the one microgram per



milliliter level, we're generally not detecting the

-------
                                                    434
pesticides.  So, again, we're trimming a little bit
on both ends in order to come up with a range over
which calculations can actually be made.
SLIDE
     The method detection limit was determined by
spiking another seven 2-liter water samples at a
concentration near that of the estimated detection
limit, which we have just determined in the previous
set of slides.  The labels at this point are spiked
into the water at a five microgram per liter con-
centration, and the detection limit, as presented
on this and the next slide for the various pesticides,
is a calculation of three times the standard deviation
of the measurements.  We're looking here at levels of
200 to 400 parts per trillion.
SLIDE
     The worst case for the endosulfan sulfate is
about two and a half parts per billion.
SLIDE
     The next step was a determination of a series of
five point calibration curves to determine calibration
linearity and calibration reproducibility.  We ran...I
don't have any slides on this.  We ran a series of
curves at concentrations of 25, 50, 100, 250 and 500
micrograms per milliliter, calculated the response

-------
                                                    435




factors of the various concentrations for individual



curves, relative standard deviation, and the between



curve relative standard deviations to determine that



indeed, the calibration was linear or reasonably linear,



very linear over this concentration range, and the



reproducibility of that calibration was within reason.



SLIDE



     The last set of slides that I have here provide



the precision and recovery data.  To achieve the



measure of precision in recovery we analyzed two sets



of four replicates each at- a concentration 20 times



the minimum detection limit for the unlabeled



pesticides.  Again, the labels were maintained at a



five microgram per liter concentration.



     Generally speaking, as you look down this slide,



the recoveries were reasonable.  Again, the labeled



compounds were calculated, their recoveries based on




internal standardization.  The unlabeled compounds,



which have labeled analogs, are calculated on the



basis of using the label isotope dilution method.



The column to the far right is the relative standard



deviation of the detected amount, and as you can see,



the relative standard deviation is always...is most



often less when you're dealing with an unlabeled



compound which has a labeled analog.

-------
                                                     436
     The  top  one  there, Alpha  BHC,  the  labeled  compound
has an RSD of 18,  and  the  unlabeled compound  an RSD  of
8, and this held  pretty true throughout the study.
We got a  significant decrease  in  the relative standard
deviation of  the  measurement and  the recoveries were
in the neighborhood of 100 percent  or greater.
SLIDE
     This just  follows through the  trend showing the
result.   And  I  think the last  slide is  the next one,
which continues on.
SLIDE
     The  recovery of the labeled  compunds in  those
materials which did not have labeled analogs  is also
reasonably good.   We're talking in  the  60 to  80
percent recovery  range.  At these levels that's not
too awfully bad.            ,
     I noticed that there were only  three of  us  that
didn't submit an  abstract as I was  reading that  thing.
They've got little stars by your names.  Paul Friedman,
myself and some guy named BREAK, and  so I did bring
some of the report.  And as I mentioned at the  start
of the meeting, I'd be happy to send  you one  if  you
just want to  leave me a card during  that break.
     Now, I'd be happy to answer any  questions  that you
might have.

-------
                                                    437
             QUESTION AND ANSWER SESSION
                          MR. TELLIARD:  Trying to
make up for the fact that you didn't submit an
abstract.  Mailing some of those silly reports purely
isn't adequate.  Dr. George has a question.
                         ' MR. STANKO:  George Stanko
from Shell Development.  Bob, on one of your slides
you showed that the recovery of the carbon 13 analog
of aldrin was 39 percent with a relative standard
deviation of 29, and that the analyte itself, aldrin,
the recovery was 117 percent with a standard deviation
of 15.  How can the isotope dilution procedure be
applied when you're not recovering your deuterated or
carbon 13 analyte at the same percentage as the
analyte itself?
                          MR. TELLIARD:  I think
that's a knit-picky question.
                          DR. BEIMER:  I tried to go
over that portion quickly, George.  You must have
picked up one of the reports, right?
                          MR. STANKO:  I didn't go
out last night, I stayed home.
                          DR. BEIMER:  What we're
dealing with here, George, is the...where are we?

-------
                                                     438
                          MR. STANKO:  The bottom
two.  That is the whole basis for  isotope dilution.
                          DR. BEIMER:  Yes.  The
labeled compound there, George, the recovery of the
labeled compound is calculated relative to an internal
standard.  The unlabeled recovery  is calculated
relative to the labeled compound,  so it will naturally
be higher.  It's corrected for that recovery.  We did
not go through and calculate the unlabeled recovery
as a function of internal standard.
     Did I make that clear?
                          MR. STANKO:  It was clear,
but that would mean that the slide would be confusing
to someone who tried to interpret  it the way I had
done.  What you're really trying to tell me is that
the recovery of the aldrin itself was approximately
39 percent in this case.
                          DR. BEIMER:  Yes, that's
correct.
                          MR. STANKO:  And you're
showing me how the recovery corrected data would
appear.
                          DR. BEIMER:  That's correct.
                          MR. STANKO:  I think I would
recommend that you alter this slide to point this out.

-------
                                                    439




I think some people are going to be confused later



on.



                          DR. BEIMER:  The idea here



was to point out that isotope dilution does indeed



correct for recovery of the materials, and I apologize



for the confusion that that might have presented.



The relative standard deviation in the final column,



though, is a measure of the absolute concentration as



determined, and I was looking more here at the



precision of the measurement than the actual recovery.



                          MR. MADDALONE:  Ray Maddalone,



TRW.  Bob, two questions.  One, when you did your MDL,



did you use a distilled water sample or did you use



a real water sample with low pesticide concentrations?



                          DR. BEIMER:  All of this was



done on as pure a laboratory water as we possibly could



obtain.  We ran blanks alongside the...lab blanks



alongside of the analyses., We have yet to test this



method on real world samples, and I think there should



be some legitimate concern about concentrating real



world samples to 200 microliters.  I'll put both hands



up in the air on that.  We're undoubtedly going to



have to use some GPC cleanup in order to achieve



these kind of detection limits, and our recoveries



will probably drop off accordingly.

-------
                                                     440
                          MR. MADDALONE:  The  second
question.  On the RSD's that you have  for the  labeled
compounds versus the unlabeled compounds, it looks
generally that the labeled compounds have a higher
relative standard deviation.  Is there any particular
reason for that?
                          DR. BEIMER:  Yes.  The
unlabeled compounds...the scatter in the data  is
reduced because they are calculated using the  labeled
compounds as an internal standard.  You're probably
confused the same way George is, and I apologize for
that.  I probably should have put this on two  separate
slides so that you could separate the two effects.
                          MR. MADDALONE:  So in
effect, the deuterated compounds are telling us what
the actual recovery and precision is of the method,
and then both of those data bits were used to  correct
the recovery and the relative standard deviation or...
                          DR. BEIMER:  Well, you
don't actually correct the standard deviation.
                          MR. MADDALONE:  Not  correct
it, I'm sorry.
                          DR. BEIMER:  The standard
deviation goes down because you're correcting  for the
variability of the method because the labeled  compound

-------
                                                    441
mimics the unlabeled compound in the extraction and
in the chromatography.
                          MR. LUCAS:  Sam Lucas from
Battelle.  You had a number of compounds you were
clearly disappointed in your detection limits on.  I
wonder if you had a chance to try to determine whether
the fault was extraction of those compounds, losses
in your evaporator cavity injector, losses on
chromatography, poor mass spec characteristics or
whatever might cause the high detection limits.
                          DR. BEIMER:  As you may
have noticed in the early slides, the detection limits
that we observed...the high detection limit compounds
were also those compounds that we got poor detection
with the mass spectrometer, so that we're not dealing
with a significant extraction problem, we're dealing
with a mass spectrometer problem.  And, unfortunately,
in some of those  cases, we're having to choose masses
because of potential interference problems which are
not the largest in the spectrum.  If you throw away,
let's say, you look at a 40 or 30 percent abundance
mass peak, you're throwing away that fraction of the
total sensitivity to that compound  in your instrument,
and we did that in some of these cases.  We're down
with a couple of  those compounds at mass abundances,

-------
                                                     442
mass peaks, that are  in  the  10  to  20  percent  range  in
order to not have  interference  problems.
                          MR. LUCAS:   Did you have  an
opportunity to  try cold  long column injection to  see if
any of your thermile  labile  compounds  could be improved
that way?
                          DR. BEIMER:  No, we didn't
try that.
                          MR. LUCAS:   You mentioned
that a proper deactivation of your injector liner was
extremely critical, and  that's  what made me think
that perhaps your method could  be substantially
improved with cold long  column  injection.
                          DR. BEIMER:  It's a  very
good possibility.
                          MR. BARRICK:  Bob Barrick,
Tetra Tech.  Do you have any...for any of these
labeled analogs, do you  have sufficient chromatographic
resolution such that  they might possibly be done by
GCCD?
                          DR. BEIMER:  I'm sorry, by
what?
                          MR. BARRICK:  Do you get
enough chromatographic resolution on any of the
labeled analogs such  that you might be able to take

-------
                                                    443




any of these and use just a few of those as recovery




standards using GCCD detection?




                          MR. TELLIARD:  Electron



capture.




                          DR. BEIMER:  Electron



capture.  Okay, I'm sorry.



                          MR. BARRICK:  Yes, for you



modern chemists, electron capture, yes.




                          DR. BEIMER:  Okay, I'm



sorry.  Generally speaking, the resolution on the




column is not sufficient to do that.  The peaks would



at least be overlapped.  I'm going to embarrass'



myself right here to tell you that I've not seen one



of the chromatograms.  I'm not sure what the degree



of separation is.



                          MR. BARRICK:  At least



you're honest there.




                          DR. BEIMER:  The other



fellow whose name's on the report actually did



the work.  He's the one who really deserves the credit



for what was done here.  His name is Lee Helms.



                          MR. TELLIARD:  Thank you,



Robert.

-------
                                               444
 ANALYTICAL  CONDITIONS
GO  - 30M, 0.25mm i.d. - DB-5
      He carrier at  30 cm/sec
      Splittless injector
      100ฐC -  280ฐC @ 8ฐC/min (10 min hold)

MS  - Finnigan OWA 1020
      Scanned  35-450 amu @ 1  sec
      Tuned for Method 1625B  DFTPP
   (2j2'-difluorobiphenyl used as internal  standard)

-------
                                             445
CHARACTERISTIC PESTICIDE MASSES
   Compound

   Aldrin

   a-BHC

   /3-BHC

   -y-BHC

   (5-BHC

   4,4'-DDD

   4,4'-DDE

   4,4'-DDT

   Dieldrin
Analog

13,
'C
 16
 16
d
d
 8
 8
13
  C
Primary
 m/z's

263/269

219/224



219/224



235/243

246/254

235/243

263/269

-------
                                                  446
Characteristic  Pesticide Masses (Continued)
   Compound

   a-Endosulfan

   /3-Endosulfan

   Endosulfan  sulfate

   Endrin

   Endrin  aldehyde

   Heptachlor

   Heptachlor  epoxide

   2,2c-Difluorobiphenyl (I.S.)
Analog
13
  C
Primary
 m/z?s

170/164

241
   .';  , !
272

81

67

160/164

81

190

-------
                                               447
               INTERFERENCES
• Heptachlor is  a component of chlordane
• Chlordane interferes with (d,) a-Endosulfan,
                    13
  a-Endosulfan and (  C.) Aldrin

• Toxaphene interferes with (d.) a-Endosulfan and
  a-Endosulfan
  PCB  1254  interferes with  (dg) 4,4'-DDE,
  a-Endosulfan and (dfi) a-BHC

-------
                                              448
 CONCENTRATIONS  AND  STANDARD
              DEVIATIONS FOR
         PESTICIDE  COMPOUNDS
                                           Relative
                               Average     Standard
                             Recovery (%)  Deviation
402 (d6)a-BHC
102 a-BHC
103 /2-BHC
404 (d6h-BHC
104 7-BHC
105 &-BHC
400 (13C4) Heptachlor
100 Heptachlor
289 (13C4) Aldrin
089 Aldrin
75
18
80
65
65
99
59
103
75
39
117
8
22
23
8
23
26
21
29
15

-------
                                                  449
CONCENTRATIONS AND STANDARD
DEVIATIONS FOR
PESTICIDE COMPOUNDS  (Continued)
101  Heptachlor epoxide

295  (d4) a-Endosulfan

095  a-Endosulfan

293  (d8)4,4<-DDE

093  4,4C-DDE

294  (d8)4,4'-DDD

094  4,4<-DDD

292  (d8)4,4'-DDT

092  4,4'-DDT
                                               Relative
                                  Average     Standard
                                Recovery  (%)  Deviation
75
20
72
134
56
120
61
119
162
139
23
6
32
5
37
8
46
10

-------
CONCENTRATIONS AND  STANDARD
DEVIATIONS FOR
PESTICIDE COMPOUNDS  (Continued)
                                                  450
                                               Relative
                                  Average     Standard
                                Recovery  (%)  Deviation
290  (13C4)Dieldrin

090  Dieldrin

098  Endrin

096  /3-Endosulfan

099  Endrin Aldehyde

097  Endosulfan sulfate

186  (d5)2c-Chlorobiphenyl

187  (d5)3c,4',56-Trichlorobiphenyl

188  (d6)3,3',4,4'-Tetrachloro-
        biphenyl
58
35
113
59
83
53
100
69
66
62
22
30
31
32
36
11
24
37

-------
                                         451
   ESTIMATED  INSTRUMENT
        DETECTION LIMITS
 FOR  PESTICIDE  COMPOUNDS
Compound
089 Aldrin
102 a-BHC
103 /3-BHC
104 7-BHC
105 &-BHC
094 4,4C-DDD
093 4,4'-DDE
092 4,4'-DDT
 Estimated Instrument
Detection Limit (#g/mL)
            5
            5
            5
            5
            5
            3
            3
            3

-------
                                                 452
ESTIMATED INSTRUMENT
DETECTION LIMITS
FOR  PESTICIDE COMPOUNDS  (Continued)
  Compound
  090 Dieldrin
  095 a-Endosulfan
  096 /?-Endosulfan
  097 Endosulfan Sulfate
  098 Endrin
  099 Endrin Aldehyde
  100 Heptachlor
  101 Heptachlor Epoxide
 Estimated  Instrument
Detection Limit f/zg/mL)
            10
            10
            30
            24
             5
             3
            10
             3

-------
                                           453
    ESTIMATED  EXTRACTION
     DETECTION  LIMITS  FOR
     PESTICIDE COMPOUNDS
Compound
089 Aldrin
102 a-BHC
103 /3-BHC
104 7-BHC
105 ฃ-BHC
094 4,4'-DDD
093 4,4'-DDE
092 4,4;-DDT
Estimated Extraction
  Detection Limit
       (2L  Sample)
         0.5
         0.5
         0.5
         0.5
         0.5
         0.3
         0.3
         0.3

-------
                                                 454
ESTIMATED EXTRACTION
DETECTION LIMITS FOR
PESTICIDE COMPOUNDS  (Continued)
  Compound
  090 Dieldrin
  095 a-Endosulfan
  096 /?-Endosulfan
  097 Endosulfan sulfate
  098 Endrin
  099 Endrin aldehyde
  100 Heptachlor
  101 Heptachlor epoxide
Estimated  Extraction
  Detection Limit
       (2L Sample)
          1
          1
          3
          2.4
          1
          0.5
          2
          0.3

-------
                                         455
  METHOD DETECTION  LIMIT
 FOR  PESTICIDE  COMPOUNDS
  BASED  ON INITIAL  VOLUME
 OF  2  L  FOR  WATER SAMPLE
Compound

102 a-BHC
103 /3-BHC
104 7-BHC
105 ฃ-BHC
100 Heptachlor
089 Aldrin
101 Heptachlor  epoxide
095 a-Endosulfan
Method Detection
   Limit
       0.21
       0.34
       0.19
       0.26
       0.47
       0.44
       0.16
       0.80

-------
                                                  456
METHOD DETECTION LIMIT
FOR  PESTICIDE  COMPOUNDS
BASED  ON INITIAL  VOLUME
OF 2 L  FOR  WATER SAMPLE (Continued)
Compound
093 4,4C-DDE
094 4,4C-DDD
092 4,4'-DDT
090 Dieldrin
098 Endrin
096 /?-Endosulfan
099 Endrin aldehyde
097 Endosulfan sulfate
Method Detection
    Limit  //g/L
        0.39
        0.18
        0.20
        0.24
        0.62
        2.14
        0.54
        2.45

-------
                                                    457
                          MR. TELLIARD:  Our next




speaker is John McGuire from our R&D lab in Athens.



Most of you know John as the father of Walt Shackel-




ford.  He's not that old; I just made that up.  John.

-------
                                                     458
                  JOHN  MCGUIRE,  PH.D.

    UNITED STATES ENVIRONMENTAL PROTECTION  AGENCY
                    ATHENS  E.R.L.
     PLANS FOR AN  IMPROVED ALGORITHM  FOR  FINDING
                GC PEAKS  IN GC/MS DATA
SLIDE 1

                           DR. MCQUIRE:  This  is

what it's all about.   I don't know how  legible that

slide really is,  it's  rather an old slide.  In fact,

it dates back to  the early days when EPA was  getting

set up.  Those are fish, all belly up,  and had to do

with why EPA was  established in the first place.  The

Clean Water Act helped to  focus on insults like this

to the environment, and since then, year by year,

there has been a  general improvement.

SLIDE 2

This is a question that we were asking  at the very

beginning of EPA...actually two questions...and, you

know, those questions are still being asked today.

So, I think they were very pertinent then, they are

very pertinent now, and have an awful lot to do with

analyzing samples connected with the RCRA program in

particular, and with all of the EPA programs  in

general.

-------
                                                    459
     In the early 70's, the infant EPA began to use
an extremely expensive detector for a gas chromato-
graph, but that extremely expensive detector did give
a fingerprint, a mass spectrum, of each GC peak as it
eluted from the column.
     Our Athens lab, together with research grantees
in various parts of the country, led the way in
making a spectra matching program that had been
developed by Klaus Bieman and his associates at MIT
available, first of all, to all of the other EPA
laboratories to make tentative identifications
much more quickly than could be done by manual
interpretation; from them, it was picked up and
spread to other users of System Industries equipment.
This approach had the unfortunate drawback, of
course, that sometimes people who didn't know what
they were doing used the computer's interpretation as
being gospel.
     We also realized  that the available spectral
collections were biased strongly towards petroleum
spectra and biological spectra; and we began a project
to collect spectra  of  pesticides and other environmental
pollutants, which later was picked up by Steve
Heller and Bill Milne  in Washington, and has now
become the EPA-NIH  data base that is the standard for

-------
                                                     460
all GC/MS users.
     In  '76, EPA entered  into a consent decree with
NRDC and other environmentalist groups concerned
about establishment of effluent guidelines for
industry.  This gave visibility to the Effluent
Guidelines Division which, as you all know, is now
ITD.
     Our Athens Research  Lab, together with the
Cincinnati Lab, helped the Effluent Guidelines
Division in setting up the priority pollutant list in
its final, analyzable form.  I'm sure you all are
aware of the difference between the original list and
the final one.  The original Consent Decree has been
modified several times.
     When the survey phase was well underway, parties
to the original consent decree agreed to a significant
amendment:  yesterday Tom Fielding made reference
to Paragraph 4(c) of the  consent decree, which agreed
that EPA would massage data collected by contract
labs as part of the EGD Effluent Study for non-
priority pollutants.  Walter Shackelford,  who at
the time was part of my group in Athens, has reported
to this meeting in the past on plans, progress and
some of the results of that study.   Since  his work
was what Tom referred to as the "first bite of

-------
                                                    461




Paragraph 4(c)" and this is the second, I want to




refresh your memory on what was involved.



     By the way, yesterday Telliard questioned the




repute of the Federal Register as a scientific journal,



so much of what I have said to date can be found in



Environmental Science and Technology.  I'm sure Bill




would approve of that reference since he is one of



the authors.




     Data from the contract labs' priority pollutant




runs were stored on mag tape and sent, ultimately at




least, to our Athens lab for processing.  Walter and



his contractors had set up a set of programs based on



the CLEANUP and HISLIB, Historical Library, programs



developed at Stanford, and the PBM program, Probabi- ,



lity Based Matching, developed at Cornell.



SLIDE 3




     The first stage of that set of programs,




RETRIEVE, was a simple one designed to read 800 bits



per inch tape from Finnigan, HP, or System Industries



data systems, produce a chromatogram of the data, and



store the data in a fixed format for future processing.



     The second, or BUILD stage, was an interactive



one wherein the chemists made sure that all pertinent



parameters, such as scan speed, internal standards,



mass range and chromatographic column, were entered

-------
                                                     462
into a data base for the particular GC/MS run.
     The next three programs located the internal
standards and assessed the GC peak profiles for all
standards.  These profiles were considered model
profiles for compounds eluting nearby.  I'll get
into that more in a moment.
SLIDE 4
     MODELS was followed by CLEANUP, a program that
extracted a typical mass spectrum reasonably free of
background for each GC-separated compound.  This
spectrum was then put through McLafferty's data
compression program to emphasize unique mass numbers
and processed through what we then felt was an extremely
large, 50,000 plus, collection of spectra.  It was,
in fact, the super set of the EPA/NIH library at that
t ime.
     Walter and his colleagues found that reliability
of the matches could be improved greatly...and that
means less "false positives"...by requiring fairly tight
retention time windows.  This relative retention time
window was calculated for the PBM hits in the same
part of the program that calculated concentration,
based on the known concentration of internal standards.
     If any of you remember Walter talking about
problems with the initial tape study several years

-------
                                                    463
ago, finding these internal standards at known



concentrations was one of the biggest problems we had



initially.  The d^o anthracene that was supposed to



be there as an internal standard simply wasn't there.



     Finally, HISLIB checked to see if the tentative



identifications fit the historical relative retention



time for that particular compound.  If it did, a HIT



file was updated to show a new identification of that



particular compound.  If it did not, a MISS list was



consulted to see if that same unidentified spectrum



had been found in that same relative retention time




window for other samples before.  If it had been, the



MISS file was updated.  If it had not, or if either



the PBM score or the relative retention time window



was a 'close but no cigar1 type, the spectrum was



output for an analyst's decision as to which file it



would go into.  This, briefly, was the program Athens



and EGD had in place during the "first bite" phase.



SLIDE 5



     A summary of what was found is given in the next



slide.  I don't want to take time to dwell on any of



the points here except the number of spectra not



identified.  That 2,500 represents spectra that



occurred more than once and were felt by the chemists



who did the final decision making to be reasonable

-------
                                                     464
 appearing  spectra,  but were  not identified.   There
 were many  other  spectra that were not identified that
 were judged  by those  same  chemists to be  trash and
 they were  not included in  the MISS list.
     As  I  have already indicated,  we  superimposed the
 relative retention  time windows on the PBM searches,
 and we tightened  those windows up to  eliminate false
 positives.   Well, any  analyst knows if you tighten
 things up  to eliminate false  positives you create
 false negatives.  Going through a  portion of  our  MISS
 list by hand, it  seems quite  evident  that a substan-
 tial portion of that list, perhaps  as  high as a quarter
 of it, is  indeed  false negatives.
     As well as the results and the summary,  certain
 other conclusions came  out of  this  "first bite".   These
 included the obvious facts that with some caution,
 capillary column data  was more  apt to  give
 chromatographically resolved peaks than packed columns.
And that more than one  or two  internal standards were
needed.
     As a result, contractors for the data collection
of the "second bite" were instructed to use experimental
improvements.  Now those improvements, basically, so '
far as  my particular concern at this point,  come down
to these.   Fused  silica capillary columns  were used

-------
                                                    465




for all but the VGA's, and multiple internal standards



were used.  It was also felt, based on our results in




that first study, that significant changes in the




algorithm would be needed to do justice to capillary



column data.



SLIDE 6




     Yesterday, Bill Telliard alluded to his having



more money.  Well, the implication when one says he




has more money is that one had less money in the



past.  We, down at Athens, have had that problem, too.




So, at the end of the "first bite" tape study, we put



our production tape program in mothballs for lack of




money and saw one after another of our contract



programmers and contract chemists who had gotten the



Athens system operating in the first place, leave for



bigger and better things.



SLIDE 7




     Accordingly, our first priority on setting up



for the "second bite" has been to restaff for the study.



SLIDE 8




     A contract programmer with numerous phone



consultations with Walt Shackelford, has brought the



old programs back on line on the PDF 11/70 used in the



earlier work.  He has made the programs portable by



redoing all machine language I/O calls into FORTRAN 77,

-------
                                                     466
and  is now  in  the  process  of  transferring  the  system
to the VAX  785  to  give  both the  faster  speed and  the
double precision arithmetic that we  found  we needed
in the first study.
     We're  now  having the  contractor recruit a data
chemist to  work on the  project;  if anybody knows
one, let me know.  We have placed an order for the
May, 1986,  release of the Wiley Library.   This will
contain over 100,000 spectra.  We plan  to  merge that
with our present collection, which is the  current EPA/
NIH collection  plus our original collection, by
eliminating all identical spectra.
     Incidentally, along with the tape  program, we
also have been  working with ITD in other ways.  We've
been analyzing  selected samples for the Domestic
Sewage Study, and  as a  result of that we recommended
that 30 organics from Appendix VIII be  added to 1625B,
and I believe from one of the booklets  that Bill has
handed out, they have been.
     Back on the main theme though, we've  identified...
and that was no mean task if any of you have tried
improving someone  else's software...those  areas that
need to be updated for multiple internal standards,
and are currently beginning to address them.
SLIDE 9

-------
                                                    467
     When one remembers that the early priority
pollutant days for semivolatiles had only a single
internal standard, and compares that to the abundance
of, for example, the cocktails that are added
for isotope dilution, it seems obvious that program
modifications must be made to best find and utilize
the information that will be available in the newer
data.  This slide shows a single spike set, but all
told, there are around 75 spikes that have been used
at one point or another in this phase of the program.
     In work evaluating a different program, the
Master Analytical Scheme that was developed by the
Athens lab and a contractor, we found that in capillary
column GC/MS with internal standards it was much
more important to compare the analytes to the standard
eluting closest to them than it was to compare them
to similar chemical compounds.  And that's fortunate
since the spiked sets...again using the same spiked set
that's here...tend to be reasonably well spaced over
the entire chromatogram.  This gives convenient
references for almost all the analytes that we're apt
to find.
SLIDE 10
     I said earlier we had felt, based on the original
tape study, that significant modifications of the

-------
                                                     468



 algorithms  would  be  needed to apply the programs to



 cap  column  data.   I'd  intended to go into these



 changes  in  depth  today;  however,  as we  have  dug deeper



 and  deeper  into the  old  programs  in the past three



 months,  we  found  that  that may not be necessary.  I



 believe  right  now our  job  will be much  simpler.   Rather



 than constructing new  peak recognition  algorithms,  I



 now  feel that  tuning existing system parameters  will



 do,  and  this work has  just started.



 SLIDE 11



     The need  for the  tuning  is shown in  the  next



 slide.   Here we have real  time traces of  a continuous



 monitor  of  GC  peaks eluting:   first, the  early part



 of a capillary column  run;  next,  a little bit later;



 then, a  little bit later;  and,  finally  up towards



 the  point where one is starting to get  into  an



 isothermal  portion of  the  run.  The peak width at the



 base of  this peak, which is at the beginning  of  the



 run, is  eight  and a half seconds.   That is,



 incidentally,  a single compound.   As RT increases,



 peak width  decreases:  four and a  half  seconds on



 the  second  peak,  three seconds on  the next, and  two



 and  a quarter  seconds on the  last.



     Now, consider what happens if  your mass  spectro-



meter is scanning at one second per  scan.  If you do

-------
                                                    469
your scans on the first peak, you're going to get
eight and a half scans across it.  If the scans
sync with the eluting GC peak, your mass spectrometer
is going to come up with a very nice RIC that looks
like a single peak.
     If the sync is slightly different, the mass
spectrometer is going to think that it sees two
peaks.  Similarly, further in the chromatogram
an improper sync will cause the system to miss the
top part of the peak.
     For that reason we feel we're going to have to
do a little bit of massaging.  The GC peak shown
here is indeed a valid peak shape for the early part
of the run.  We are going to have to tune our
parameters so that the system can sense that.
During the programmed portion of the GC run, we've
no problems.  At the second isothermal stage, we're
going to have to come up with a means of extrapolating.
SLIDE 12
     The confusion of the tape processing programs is
illustrated by the next few slides.  This one is a
reconstructed chromatogram as it conventionally
appears.   (For those who use Finnigan terminology, a
RIC.)
SLIDE 13

-------
                                                     470
      This  slide is of the actual data points that
 went  into  that pretty scan of Slide 12.   Note the
 scans on the  ascending (front)  edge of the "939"
 peak, followed by the large gap on the descending
 side  where  there  were no  data being taken.   At the
 "1021" peak,  we've a  nice set of data points going
 up  the peak and then  nothing at all until we get  down
 to  scan  1022,  where the chromatogram comes  back up
 again.   Well,  again,  we have to tune our  parameters
 so  that  the system can recognize that.
      I'd like  you to  take a look at the two GC peaks
 near  scans  902  and 1143.   The leftmost arrow is
 pointing to the front edge of the  small GC  peak near
 scan  902; the  next arrow  points  to the trailing edge,
 but the  top of  the particular GC peak  is  actually
 between  them.   In  other words,  I do not have  an arrow
 pointing to it.   Similarly,  near scan 1142,  I've an
 arrow pointing  to  the  front  and  an  arrow  pointing
 to the rear, but  none  pointing  to  the top.   The
probability of  finding scans on  the  sides of  peaks
rather than the top is real  and  these are possibili-
ties that we could run into  in actual data  scans.
     Now, I'm going to show  you  the  four spectra that
I have "arrowed".   The four  consist of two pairs of
spectra,  and the  important thing here is to notice in

-------
                                                    471




the first pair (Slides 14 and 15) the relative




abundance of 131 to 232.  It's around 30 percent in




Slide 14, nearly twice that in Slide 15.  The next




one (Slide 16)...look at 198 to 442...it's around



two to one in Slide 16 and about six to one in



Slide 17.  In other words, and very much to be




expected, an instrument that scans from low mass to




high mass will give more high mass bias on the



ascending side as the concentration in the ion source




is building up, and more low mass bias as it's




dropping off.




     The best spectrum for the GC-peak, the one that



should most closely correspond to that in the reference




spectrum collection, should be at the top of the peak.



For example, our friend DFTPP gives a spectrum at the




top of the...gives this spectrum (Slide 18) at the



top of the peak, and a library match (Slide 19) shows




that agreement with the reference is very good.  You



see the disagreement is plotted here, and if we look



at a library output (Slide 20), it has a purity of



846 and a fit...almost a perfect fit.  Now, that is



again taken at the top of the peak.



     So, the point is we have a bias towards the high




mass side on the ascending side of the GC peak and



the low mass side on the descending side:  it is

-------
                                                     472
 this  bias  in  the  spectra  that  we  plan  to  investigate
 as a  means of estimating  where the  top of  a GC  peak
 would have been if  our mass  scan  happened  to miss  it.
      We will  estimate linear correction factors to
 permit extrapolating the  data  points and  compare
 quantitative  results with those obtained  by the
 existing programs.  We haven't done any of it yet.
 I hope to  include a discussion of this approach in
 the overall report  of the "second bite" at next
 year's meeting.
      In concluding, I want to  thank Dr. Walter
 Shackelford and David Cline, who  are now of RTF but
 who were originally at the Athens lab,  who set  up  the
 original programs with help  from  Computer Science
 Corporation, Al Thruston  of  our group  in Athens who
 was the one responsible for  analyzing  the domestic
 sewage samples; Bruce Bartell  of Computer Science
 Corporation, who succeeded in  getting  the computer
 system operational  again  and making substantial
 improvements in the present  coding.  I must also
 recognize  the Stanford and Cornell researchers  who
started the two main portions  of the system.
     Thank you.  Are there any questions?  Bill, take
over.
                          MR. TELLIARD:  Any questions?

-------
                                                    473
You're going to let him off?  People must really be



hurting.   Thank you, John.

-------
474

-------
475

-------
476

-------
                                                 477
CD   O
                                                
-------
478

-------
479

-------
480

-------
481

-------
                                                   482
                                                oo
                                                CD
                                                <_5   CD
                                                
                                                os:   
-------
483

-------
484

-------
485

-------
                                                                                       486
   S
-------
                                                                                          487
O    31= UU
LU 10
Cl_ TO •• ซ•
1-0 V UJ  .
   CSJ i.O o C3

-------
                                                                                488
   _
LU 
-------
                                                                                           489
     .
    .. IX<    UJ
   XSi C-J    O
O    *= UJ
LU U~ซ
Q_ CO  — ป• CU
ui x LU  • .sr
   10 _J ".O LU
Lfl    -'-  •""• '
'.'"I   , .
a: u~j •

-------
490

-------
                                                                            491

-------
492

-------
   _j in
LU.< ID
                                                                           493
                                                  -" t  N  ro -o oo
                                                  u. DO>t-  ru n ru
                                                  DC
                                                   =n-C  O  CO 03-0
                                                      '  ~     O O
                                                  •>-< CO  CM  HI 01 CM
                                                   f-
                                                   3             ,
                                                     co  n  co N co
                                                  •p ru  o  co o co
V}  dj ID
O .0  -
                               LU I  _j
                               a.  i   >
                               ea LU  a. lu LU
                                  Q  - 2 z:
                                  Q   -,  T
                                  _i  _!  DL

                               O Z  UJ  Z LU
                             c t> ru  o  m ru
                                                     Q.-Z  ru tn
   o -o   .  ru  z
     i o a  a   .
        .   .  - .  in
 iซ   . O O  03  O
r-< in ru
                                                   t- CD co  ru 
-------
                                                    494
                          MR. TELLIARD:  Our next



speaker is Lee Wolfe from the Athens Laboratory and



he's going to talk a little bit about the joys of



hydrolysis, or the lack thereof.

-------
                                                    495
                     N. LEE WOLFE

    UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
              ENVIRONMENTAL RESEARCH LAB
                   ATHENS, GEORGIA
             HYDROLYTIC TRANSFORMATION OF
             RCRA APPENDIX VIII COMPOUNDS
                          MR. WOLFE:  The objective

of this talk is to acquaint you with a report that

assesses the hydrolysis of many of the compounds that

have been discussed here yesterday and today.  The

hydrolysis data were prepared by EPA1s Office of

Research and Development for the Office of Solid

Waste (Slide 1).  The tabulation of data provides

the bfest available data base, I think, on hydrolysis

for these compounds.  The kinetic data contained in

this report should be useful in evaluating, designing

sampling procedures, sample storage and sample

analyses.

     Specifically, I want to address parameters that

are in the report: hydrolysis rate constants, half-

lives, key to the data set, partial summary of the

reactivity, and some data-generating efforts that

are underway at the Athens Environmental Research

Laboratory.

     A while back, at an ACS meeting I bumped

-------
                                                     496
 into an old professor.   I  said,  "Well,  Dick,  what



 are  you doing  now;  what systems  are you working on?"



 He said,  "I just  happen to have  a  slide in my



 briefcase." He pulled  out the slide (Slide 2)  and



 said,  "I've got this  exo-bicylco-2, 2-hexyl-2-yl



 tosylate  system.  These are the  products we're



 getting."  I said,  "Well,  Dick,  you haven't changed



 much,  you're still  looking for complex  solutions to



 simple problems."



     He said,  "Well, what are you doing?"  I said,



 "I've  got this really neat system.   We've  been  working



 with it for quite a while  now and  I just happen to



 have a slide (Slide 3).  This is a pretty  exciting



 system."  "Well*," he said,  "it doesn't look  to me  like



 you've changed much either;  you're still looking  for



 simple solutions to complex  problems."



     I use  this example  to make the point,  that  in the



 last several years or so it  has been shown, that with



 a great deal of confidence one can  extrapolate



 laboratory  hydrolysis data to the  field, and that's



 pretty much  the basis for  this report and the work



 that we have done for the Office of Solid Waste.



     The  report is titled, Screening of  Hydrolytic



Reactivity of OSW Chemicals.  The  report, as I said,



was prepared for the Office of Solid Waste to support

-------
                                                    497
their efforts in modeling, fate and transport in
groundwaters, and certainly the data that were tabulated
is applicable to the work that's being done in
environmental analytical chemistry.
     The list departs, I guess, from the Appendix VIII
list in that there are only 362 compounds.  I'm not
really sure of the origin.  It has things like Fast
Track, California List, and then thirds.  My
understanding was that originally it was the
Appendix VIII.
     We've only looked at data on hydrolysis for the
organic compounds.  There are some organo-metallic
compounds on the list, and what data was available, we
have tabulated and evaluated.  We have not made
any attempt to do anything with the inorganics such
as hydrolysis of organic ligands.
     What can this report do for us?  We know that
hydrolysis reactions or hydrolytic reactions can
result in products that are unigue to the list.  For
example, benzyl chloride, one of the compounds on the
list, undergoes very rapid hydrolysis (Slide 4).
It has a half-life of a couple hours in water and
gives benzyl alcohol as a product, which would say
that you're wasting your time if you look for benzyl
chloride in a water sample, but if you want to know

-------
                                                     498
 if  it  was  there,  the  presence  of  benzyl  chloride
 might  be an  indication.
     Also, products that  are already  on  the  list,  for
 example, 1,1,1-trichloroethane, readily  undergoes
 reaction in  water, in the presence  of base,  to  give
 1,1-dichloroethene.   1,1,1-trichloroethane is on the
 OSW list,  and  also 1,1-dichloroethene is on  the list.
 One doesn't  have  to look  very  far to  find examples
 of  compounds that are not on the  list but will  undergo
 hydrolysis reactions  to produce compounds that  are on
 the list.
     Well, let's  be a purist for  just a  second  and go
 back and define hydrolysis  (Slide 5).  The definitions
 do  become  impdrtant sometimes.  I define  hydrolysis
 as  a reaction  in  which a  bond  between two atoms  of a
 molecule is  cleaved and a new  bond  is formed with  the
 oxygen atom  of a  water molecule.  These  reactions  are
 often, though not necessarily, mediated  by acid  or
 base.  Certainly  a good example of  this  is the
 hydrolysis of 2,4-D acid esters.  The  acid esters
 undergo alkaline hydrolysis to give the acid, and  in
many cases,  the hydrolysis of  compounds result  in
more polar compounds,  that are difficult  to extract
 from water and often more difficult to analyze  for.
     Another term that's thrown around a  lot, a

-------
                                                    499
much broader term, is hydrolytic degradation, a term
that's used to denote a chemical reaction that occurs
in water.  For example, the reaction you just saw,
1,1,1-trichloroethane going to 1,1-dichloroethene
is really an elimination reaction, but it's generally
just grouped as hydrolytic degradation.
     Many hydrolysis reactions are affected by pH,
and I put this slide up (Slide 6) a plot of PH
versus half-life to demonstrate some of the effects
that you might run into.  You see three types of
effects.  Looking at D, an acid catalyzed hydrolysis
reaction, you see that as the pH increases, the
half-life for this reaction increases.  The thing to
note Is that for each pH unit increase, the half-life
for this reaction decreases by one order of magnitude.
     The other type of reaction you have is a base
catalyzed reaction shown by E, whereas the pH
increases, the half-life decreases.
     The third type of reaction, which is a neutral
hydrolysis reaction is indicated by B.  We see
that this reaction is pH independent, that a
change  in pH of, as shown here, six orders of
magnitude, does not change the hydrolysis half-life.
     The report addresses this in the following way
(Slide  7).  What you're interested in, in assessing

-------
                                                     500
 whether or not your compound is going to hydrolyze,
 is the observed rate constant.   The observed rate
 constant is made upf in many cases, of three
 contributions: acid hydrolysis, neutral hydrolysis
 and alkaline hydrolysis.  The report contains the
 second-order acid catalyzed hydrolysis rate constants,
 the first-order neutral hydrolysis rate constants,
"and the second-order alkaline hydrolysis rate
 constants.
      What you need to calculate K-observed is the
 hydrogen ion concentration or the hydroxide ion
 concentration, and those numbers are available from
 the pH of the solution.
      K-observe'd is a fairly difficult number to live
 with.   It's  a fairly abstract number, and to most
 people it doesn't have  much meaning.   A much better
 way to get a feel for these numbers is to convert
 them to half-lives (Slide  8).   Half-life is generally
 defined as the time  reguired for the  initial
 concentration  of  the  reactant to be reduced to  one-
 half  the  initial  concentration.   In this  expression,
 remember  that  one-half  life, T  one-half  is  egual  to
 the  constant  .693  divided  by K-observed,  recalling
 that  K-observed  is really  the summation  of  the  three
 hydrolysis processes.

-------
                                                    501
     Because of the large amount of data and diversity
of reactivity of the compounds, varying from a few
seconds to hundreds, or even thousands of years, we
divided the report up into 15 fields (Slide 9).
The first two fields are the constituent code and
the chemical abstract register numbers.  The constitu-
ent code is a code that was provided to us by OSW.
The chemical abstract's register number is important
to us because that's an unambiguous way of identifying
all the compounds.
     The rank, field three, is really a scheduling
priority.  That's the order in which OSW is going to
be making decisions on whether or not these compounds
can bte dumped into landfills, and this was the order
in which they wanted us to do our work.
     Compounds, field four, were named as received.
When you get down to fields five and six, this is
where we start to get into the meat of the report.
What we established was whether or not the compounds
hydrolyze or not, and if they do hydrolyze, will
their hydrolysis half-lives be less than a year or
greater than a year.
     Field number five  indicates that there is no
hydrolyzable functional group on the compounds.  For
example, a compound like chrysene, an aromatic hydro-

-------
                                                     502
carbon, has no  functional groups  that  can  hydrolyze.
     Field number  six, non-labile functional  groups,
is something like  chlorobenzene.   Chlorobenzene  does
have a carbon-chlorine bond  that  can hydrolyze,  but
it does not hydrolyze at a significant rate in water
even under extreme reaction  conditions, so we indicated
those compounds as non-labile functional groups.
     Fields seven  and eight  indicate whether  the
compounds have  a half-life of less than a  year or a
half-life greater  than a year.  Fields 9,  10  and 11
indicate where  the data came from.  Field  9 is
experimental values.  These were  experimental values
that came either out of the Athens laboratory or
literature.  'Sen indicates estimated values.  These
are values that can be estimated  through various
free energy relationships.  In some cases  you can
estimate the values almost as well as you  can measure
them.  Then we have expert judgment, field 11.  For
some of the compounds there was not any hydrolysis
data available, so we used a panel of four experts
and their best estimates.
     The numbers in fields 12, 13 and 14 are the
acid hydrolysis rate constant, base hydrolysis rate
constant and the neutral hydrolysis rate constant
respectively.   We also set up an additional field,

-------
                                                    503
a kind of catch-all field that included compounds that
are not stable, or compounds which we felt there was
not much known about their reactivity and we couldn't
even make good expert judgments, and gave these high
priority for measurement.
     This slide (Slide 10) is page six out of the
report, and I know it's difficult or impossible to
read.  I've put i.t up just to show you the format of
the report.  What you see are the 15 fields across
the top with Code, CAS number, rank, compound and
then the data.
     For example, dimethylsulfate has a half-life of
less than a year, it is a measured value and it is a
neutrtal hydrolysis.  The hydrolysis rate constant of
0.6 translates into hydrolysis half-life of about one
hour.
     Some of the compounds are so unstable in water  •
that measured values for the rate constants are not
available.  They're just so reactive that you can't
measure them by conventional means  (Slide 11).  This
is a list of some of these compounds.  There are
eight compounds that undergo very fast hydrolysis
reactions; half-lives on the orders of seconds or
less.
     There are other compounds that do not undergo

-------
                                                     504
 hydrolysis reactions but do undergo oxidation-
 reduction reactions, and we identified these as such.
 The two peroxides are very strong oxidizing agents
 and very reactive.   The three hydrazines are very
 strong reducing agents, very reactive, and will
 not exist in most natural waters.
      Also, we identified seven aldehydes that will
 likely exist in water as the hydrates.  The aldehydes
 react with water to  form the hydrate.   It's a
 reversible reaction  and by adjusting  the pH,  you  can
 sometimes force a reaction back to the aldehyde.   But
 it  does present problems in extracting these  compounds
 out of water.
      I've pullซed out some  of the  data  to give you an
 idea  of  how  reactive some  of the  compounds  are
 (Slide 12).   These are  neutral  hydrolysis  half-lives
 and are  pH independent.  These  compounds will
 hydrolyze with  these  half-lives at pH  4, the  same
 as  they  will  at  pH 10.   We  see  that the  half-lives
 vary  anywhere from the  order of seconds down  to about
 90 days.  This does  indicate, I think, the  reactivity
of some  of these compounds.
     Now, many of the other compounds  react very
fast, or may not react very fast, depending on the
pH of the water.  Some of the compounds will be very

-------
                                                    505
stable at pH 5, but at pH 9 their half-lives might
well be on the order of days, and vice versa.
Compounds such as the epoxides, which are very stable
at alkaline pH's, react very fast at acidic pH's.
     Does the report include data on hydrolysis for
all the compounds on the OSW list?  The answer is I'm
afraid not.  But we have a new branch at the Athens
Laboratory, the Measurements Branch, that's being
headed up by Mr. Bill Donaldson, that is in the
business of measuring hydrolysis rate constants
(Slide 13).  Not only are they going to provide
hydrolysis rate constants, they're going to address
such things as precision and accuracy and cost-
effectiveness in attaining these numbers.
     In some cases, they'll be doing product iden-
tification.  They have the capability to do that.
Also, they are going to publish some standard
measurement procedures, and along with these standard
measurement procedures they will include some standard
reference compounds.  I think they're going to have a
standard reference compounds that can be used to test
your measurements procedures for an acid catalyzed
reaction, a neutral hydrolysis reaction and an alkaline
hydrolysis reaction.  They will then tabulate the data
and keep statistical data on accuracy and precision.

-------
                                                     506
      Last but not least,  they are also going to be
 developing a data base.   This data base will contain
 what's  called process  reactivity constants.   It will
 include hydrolysis rate  constants, octanol-water
 partition coefficients,  biodegradation rate
 constants and other data  that's  deemed necessary.
      One additional area  that I  wanted to bring in
 is  that OSW is concerned  about fate and transport  of
 organics in groundwater.  We  had done  some work, and
 they  are supporting us to do  additional work  on soil
 mediated hydrolysis reactions.   I  just want  to  point
 out the state  of  the art  at this point (Slide 14).
 In a  sediment  water system or a  soil water system  you
 have  two phase's,  the solid phase and the aqueous
 phase.   We  can say  with a great  degree of  confidence
 that  how fast  the compounds react  in the aqueous
 phase of  this  two-phase system,  depends  primarily  on
 the structure  of  the compound  and  the  pH of the
 water.   The solid particles do not  affect  the
 hydrolysis  rate constant  in the  aqueous  phase.
      In  the same two-phase system,  part  of the
 compound  is sorbed  to the solid  phase,  and we know
now that alkaline hydrolysis  is  retarded in the solid
associated phase, relative to reaction  in the water
phase.  Neutral hydrolysis is unaltered  in the solid

-------
                                                    507




phase, and acid mediated hydrolysis is actually




accelerated in the solid phase; again, when everything




is referenced back to reactivity in water.  Some of



our ongoing efforts in the Chemistry Branch in the



Athens Laboratory will be taking this hypothesis and



expanding the data base and the theoretical basis for




these reactions.



     Well, I'd like to summarize with this slide




(Slide 15).  This is really just a summary of all the



fields in the report.  As I said, there are 362



compounds on this list of which we've looked only at




the organics.  There are quite a few metals.  The



thing I think is important is that there are 49



compounds that have no hydrolyzable functional group,




there are 79 compounds that have no labile functional




group.  So, these two groups are not going to



hydrolyze, so they present no problem in sampling, in




sample storage or in sample analysis.



      In field seven, we see 85 compounds that have



hydrolysis half-lives of less than a year, and for



these compounds, their half-lives will be dependent



on pH.  We see in field eight, that 85 of the



compounds have half-lives of greater than a year and




should not present any problem in sample analysis and



sample storage.

-------
                                                    508
     That pretty much sums up what I have to say
about hydrolysis.  I'd be glad to answer any
questions.

-------
                                                    509
             QUESTION AND ANSWER SESSION
                          MR. TELLIARD:  Any questions?



                          MR. LUCAS:  Sam Lucas of




Battelle.  I think your first or second slide showed a



report to EPA in 1985.  Can you tell us something




about the availability of that report?



                          MR. WOLFE:  Yes.  That




report is available through the Athens Laboratory at




the present time.



                          MR. LUCAS:  From you




directly?



                          MR. WOLFE:  Yes.



     ซ                     MR. LUCAS:  Thank you.




                          MR. FOSTER:  Russ Foster




from RAI.  What sort of laboratory  technique was used



to prepare the samples for hydrolysis study?



                          MR. WOLFE:  Well, maybe I




didn't make that point quite clear.  For  20 or  25 of



these compounds, the values came out of the Athens




Laboratory, and some of them are literature values.



The methods of measuring  are quite  varied.  There is



really no standard technique.  All  you really need to



do is be able to measure  concentration as a function




of time.  I think most of these compounds are done

-------
                                                     510
 easier with HPLC or GLC.  Hopefully, the Measurements
 Branch, as it gets set up, is going to come up with
 some standard procedures, some protocols that can be
 used to obtain these numbers.
                           MR. WISE:  Hugh Wise, ITD.
 The  appearance of your bis (chloromethyl) ether
 struck me  odd when I first came with the agency and
 saw  the priority pollutant list for the first time;
 that a compound that hydrolyzes at  that rate would
 ever appear on a list of compounds  we were to look
 for  in wastewater samples.
      Have  these other half-lives of things that
 hydrolyze  very rapidly been  factored in to eliminating
 compounds  front the  List of Lists and so on?
                          MR.  TELLIARD:   The  List  of
 Lists,  yes.   When we  put together the ITD List  and
 the  List of  Lists,  we  used a  much smaller window.   We
 used 12  to  18  hours as our window,  realizing  that  you
 can't  run back  to the  lab that  fast  and  analyze  it.
 So if we aren't going  to see  it  in  12  to  18 hours...
we care, but we don't  care as much as we  would.
     Thank you very much, Lee.   Ladies and gentlemen,
we have a break for coffee and the tinkeltorium, and
then get back in here, please.  Thank you.
 (WHEREUPON, a brief recess was taken.)

-------
                                         511
Possible Mechanistic  Routes for the  Solvo;lysis

  of 
-------
512

-------
CO
iH
m
                 O
                    <
                    O
                 oo
                  >-
                    o
                  zo
                  Ld
                  LJ ^r

                     UJ
cฃ
o
o_
LJ
Cฃ
   00

-------
o:
o
LJ
I—
Ld
Q_
          UJ
          o
UJ
a:

o
          o
          Q
                         O
                          -
           M
           Z
           Ld
           m
                              o
                          Ld
                                 CM
                               o
                 Q
                             _
                             O
                               A
Ld

ง    ">
                                        CM
                                   <  IT
                 o    o
                 Q
                 O
                         a.
                                               Ld
                                     Ld
                       -T1   o
                       O    =!
                                     O

                                     Q
                                               Ld
                                               Ld
           O    i
             hO   o
           _L_    jv;
           O    i-

-------
                                                        515
                           DEFINITIONS






HYDROLYSIS -- A REACTION IN WHICH A BOND BETWEEN TWO ATOMS OF A




MOLECULE IS CLEAVED AND A NEW BOND IS FORMED WITH THE OXYGEN ATOM




OF A WATER MOLECULE.  THESE REACTIONS ARE OFTEN,, ALTHOUGH NOT




NECESSARILY,, MEDIATED BY ACID OR BASE.
HYDROLYTIC DEGRADATION -- A BROADER TERM  USED  TO  DENOTE  ANY




CHEMICAL REACTION THAT OCCURS  IN WATER.

-------
                                        516
      Possible pH Rate Profiles for the
  Reaction of Organic Compounds in Water
o
CD
(f)
cn
o
                                             ~5
8x10
8xlO"3
                                               IJLJ
8x10
                                            -1 7
     •=3:
                                         8x10
   1

-------
CO
I	
o
C_J


LU
h-
DC


CO
>-
 O
 DC
 a
 n:
           X
           o
            X
            O
 II
  _Q
  O
_^:
        0
        0
       JZ
   c
   o
   •4—
   to
   c
   o
   o

   0
   TJ
   0
   N
                   D
V)

o
 C
 D
 CO
 C
 O
 o

 0

 o
 V-

.to
*to
          D
 o

TJ
                        I     I      I
                        I     I      I
                                       c
                                       0
                                       CO
                                       c
                                       o
                                       o

                                       CD
                                                 CD
                                      TJ
                                       0
                                                  31
                                                  O
                                              c
                                              o
                                                        o
c
o
to
c
0
0

-t-
0

TJ
0
ฃ
0
CO
jQ
O
D
O
TJ
*O
D
^_
0
TJ
O
1
TJ
C
0
o
0
to
-4—
c
0
O
C
o
o


.—
c
0
0
TJ
^~
-E-
_jr-
-t—
0
C

0
TJ
0
JL
CO
H-

i_
0
to
D
JQ
5^
0
TJ
i^.
|
I
TJ
C
O
o
0
to

0
cone
c
o
• —
0
TJ
*X
O
~o
>



-------
                                                         518
Half-life is defined as the time required for the initial
concentraiton of a reactant to be reduced to one-half the
initial concentration.
The general half-life expression for a acid or base
catalyzed hydrolysis assuming pseudo-first-order reaction
kinetics is:
                             0.693
kH[H+]
                                   kOR[-OH]

-------
                                      519
KEY TO HYDROLYTIC DATA
FIELD
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
ABREVIATION
CODE
CAS NO.
RANK
COMPOUNDS
NHYF
NLFG
LESS
MORE
MES
EST
EXP
ACID
BASE
NEU
MIS
                   REPRESENTS




        CONSTITUENT CODE




        CHEMICAL ABSTRACTS REGISTRY NUMBER




        SCHEDULING PRIORITY




        NAMES AS RECEIVED




        NO HYDROLYZABLE FUNCTIONAL GROUP




        NON-LABILE FUNCTIONAL GROUP




        HALF-LIFE LESS THAN  A YEAR




        HALF-LIFE GREATER  THAN  A  YEAR




        EXPERIMENTAL  VALUES




        ESTIMATED VALUES




        EXPERT  JUDGEMENT




        ACID HYDROLYSIS  RATE CONSTANT




        BASE HYDROLYSIS  RATE CONSTANT




        NEUTRAL HYDROLYSIS RATE CONSTANT




        OTHER REACTIONS

-------
Page
No. 6


520
05/22/85

'CODE
U053
U061
U063
U064
U066
IW67
U074
IM86
U089
U103
Ul'}8
UII5
111:4
U133
U13?
U151
U154
U155
U1S7
U153

CAS NO
123-73-9
50-29-3
53-70-3
189-55-9
96-12-8
106-93-4
764-41-0
1615-80-1
56-53-1
77-78-1
123-91-1
75-21-8
110-00-9
302-01-2
193-39-5
7439-97-6
67-56-1
91-90-5
56-49-5
101-14-4


RANK
1ST THIRD
1ST THIRD
1ST THIRD
1ST
1ST
1ST
1ST
1ST
1ST
1ST
1ST
1ST
1ST
1ST
1ST
1ST
1ST
1ST
THIRD
THIRD
THIRD
THIRD
THIRD
THIRD
THIRD
THIRD
THIRD
THIRD
THIRD
THIRD
THIRD
THIRD
THIRD
1ST THIRD
1ST THIRD
OS*
CDHPซ NHYF NLFS LESS HQRE HEAS EST EXP ACID BASE NEU HIS
CROTONALDEHYDE YES 2
DDT YES YES .35 6.8E-6
DIBEHZO (A, H) ANTHRACENE YES
1,2,7,8DIBENZOPYRENE YES
l,2-DIBROซQ-3-CHLOROPROPA YES YES
NE
ETHYLENE DIBRDHIDE YES YES 3.7E-5
l,4-DICHLORO-2-BUTENE YES ' YES 1.7E-3
N,N!-DIETHYLHYDRAZINE YES 3
DIETHYLSTILBESTRQL YES
DLIETHYL SULFATE YES YES .6
1,4-DIQXANE YES
ETHYLENE OXIDE YES YES 33.5 2.4E-3
FURAN YES
HYDRAZINE- YES 3
INDENO(1,2,3-CD)PYRENE YES
HERCURY 5
HETHANOL YES
HETHAPYRILENE YES
3-HETHYLCHOLANTHRENE YES
4,4-HETHYLENE-BIS-(2-CHLQ YES YES lE->i
                              ROANILINE5
U171  79-46-9     1ST THIRD   2-NITROPROPANE
YES

-------
                                                521
  COMPOUNDS UNSTABLE IN WATER






           HYDROLYSIS




CHLORIDE CYANIDE




PHOSGENE




PHOSPHINE




ACETYL CHLORIDE




CARBONYL FLUORIDE




CYANOGEN BROMIDE




CHLOROMETHYL METHYL  ETHER




BIS-(CHLOROMETHYL) ETHER






             REDOX




METHYL  ETHYL KETONE  PEROXIDE




N,N'-DIETHYLHYDRAZINE




1,1-DIMETHYLHYDRAZINE




1,2-DIMETHYLHYDRAZINE




61 ,a-DIMETHYLBENZHYDROPEROXIDE






             HYDRATES




7  ALDEHYDES WILL EXIST AS THE




HYDRATE IN MOST NATURAL WATER




SAMPLES.

-------
                                                         522
         HALF-LIVES FOR THE NEUTRAL HYDROLYSIS
                 OF SELECTED COMPOUNDS
           COMPOUND

  BROMO ACETONE

  BENZENE SULFONYL CHLORIDE

  CARBONYL FLUORIDE

  METHYL CHLOROCARBONATE

_3 -CHLOROMETHYL ETHER

  METHYL BROMIDE

  DIMETHYL SULFATE

  ETHYLENE OXIDE

  DIMETHYL CARBONYL CHLORIDE

  BENZOTRICHLORIDE

  ETHYL METHANE SULFONATE

  BENZYL CHLORIDE

  METHYL ISOCYANATE

  BENZAL CHLORIDE

  1,3-DICHLOROPRENE

  2>2'-BIOXIRANE

  MALEIC ANHYDRIDE

  PHTHALIC  ANHYDRIDE
   HALF-LIFE
 4 MIN



35 MIN

24 SEC

20 DAYS

 1 HR

12 DAYS

 4 MIN

 3 MIN

19 HR

 9 HR

 8 MIN

 7 MIN

11 DAYS

29 DAYS

26 SEC

 3 MIN

-------
CO
(M
CO
CO
_J
o<
Q -o
in
o
DL. JH
O uj
~" — * ~X^
UJ —
""*"*>ป> \/
Ld
Q
m
o
z:
DD
(/)
h-
z:
Ld
Ld
ID
CO

Ld


CO
h-
<
h~
CO
O
o
Ld
H-
<ฃ
CE
Ld
Q
^>
0

Q_

Z
O
h-
o
Lu
h-
z
Ld
Q
h-
O
ID
Q
O
(y
Q_

Ul
Ld
Q
• *
Ld
O
O
Q_
H-
z:
Ld
Ld
Cฃ
CO

-------
                                                  524
      SOIL-WATER MEDIATED  HYDROLYSIS  REACTIONS

 IN WATER


      HOW FAST THE COMPOUNDS REACT DEPENDS ON THE
      STRUCTURE OF THE COMPOUND AND THE pH OF THE WATER

 IN SOILS


      ALKALINE HYDROLYSIS  IS RETARDED IN THE SOLID
      ASSOCIATED PHASE

      NEUTRAL HYDROLYSIS IS UNALTERED IN THE SOLID
      ASSOCIATED PHASE

      ACID MEDIATED HYDROLYSIS IS ACCELERATED IN THE
      SOLID ASSOCIATED PHASE

WHEN REFERENCED BACK TO REACTIVITY IN WATER

KINETIC STUDIES FOR 10 ADDITIONAL COMPOUNDS WILL
CARRIED OUT TO FURTHER SUPPORT THIS MODEL

THEORETICAL BASIS

      CONSISTENT WITH SIMILAR SYSTEMS
      TRANSITION STATE CHEMISTRY

-------
                                                                    525
       Field







 1  CONST CODE





 2  CAS NO





 3  RANK




 4  COMPOUND





 5  NHYF





 6  NLFG





 7  LESS




 8  MORE




 9  MEASURED





10   ESTIMATED





11   EXPERT




12   ACID HYDRO





13   BASE HYDRO




14  NEUT HYDRO.





15   MIS




    1





    2




    3




    4
25




8




10




9
                Count







                362




                359




                362




                362




                49




                79




                85




                85




                44




                9




                115




                15




                58




                64
               50

-------
                                                    526
                          MR. TELLIARD:  We'd like to
get the session started.
     Our next speaker is Ted Martin from our Cincinnati
lab/ who is going to talk a little about some metals
analysis, and then followed by Ray Maddalone from
TRW, who is also going to be talking about the EPRI
metals study that has been going on for the last
couple of years on the round-robin for the Edison
Electric Institute.  Ted.

-------
                                                                        527
                              THEODORE D.  MARTIN

                UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
               ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
                 EVALUATION OF METHOD 200.1 DETERMINATION OF
                             ACID SOLUBLE METALS
                              (ABRIDGED REPORT)

                                 INTRODUCTION
    This report describes a practical  and controlled evaluation of the
"acid-soluble" metal methodology,  Method 200.1 DETERMINATION OF ACID SOLUBLE
METALS.  The method is basically a sample extraction/preparation procedure
with reference to the determinative step being the 200 series methods,
Method for Metals - Methods for Chemical Analysis of Hater and Wastes,
EPA-600/4-79-020, March 1983.For the purposes of this report the
discussion of Method 200.1 is limited almost entirely to the different
aspects of sample preservation and extraction.

    The term "acid soluble" metal, along with the concentration requirements
as related to water quality criteria as a possible enforceable standard
pursuant to Section 304 (a) (1) of the Clean Water Act [33 U.S.C.
1314 (a) (1)] is given in the Federal  Register (FR), July 29, 1985 [50 FR
30784].  Included in this FR reference is the National Technical Information
Service (NTIS) listing of publication order numbers for the criteria
documents for each of the six metals of concern:  arsenic (As), cadmium
(Cd), chromium (Cr), copper (Cu),  lead (Pb), and mercury (Hg).  Four
additional metals:  aluminum (Al), nickel (Ni), silver (Ag), and zinc (Zn)
were evaluated in this report since regulations concerning them are being
considered.                                                  ,

    This FR states that for ambient water the Agency believes that the
measurement of "acid-soluble" metal will provide a more scientifically
correct basis upon which to establish criteria.  The latest criteria [50 FR
30784] has been  developed on this basis.  The concept of "acid-soluble"
metal  is a measurement believed to be less rigorous than "total" or
"total-recoverable" metal analyses.  Although it is only a partial
determination of the metal content in a sample, it is intended to provide an
estimate of the  available portion associated with suspended  and particulate
material in addition to the dissolved metal fraction.  It should be noted
that for those metals with stable oxidation states of reported different
toxicity, the Agency has included required measurement of "acid-soluble"
metal  speciation as a part of the water quality criteria.  At present only
arsenic and chromium are affected, but other metals may be added at a later
date.  Also,  the FR states that this methodology should be applicable to
analysis of saltwater matrix as well as freshwater.  The present version

-------
                                                                        528
 (1.4)  of Method 200.1  does not attempt  to  provide  for  these  two  specific
 needs,  but  is  a generalized approach  and first  offering methodology to the
 analysis of "acid-soluble" metals.  The purpose of this report is to discuss
 the ruggedness of Method  200.1 and  provide a  standard  method that may be
 used to evaluate the toxicity of  "acid-soluble"  metals to  aquatic life.  A
 final  draft copy of the proposed  method with  acknowledgment given to
 reviewers and  others who  have provided  significant contributions is
 available from the Environmental  Monitoring and  Support Laboratory -
 Cincinnati, Ohio  45268..

                                  CONCLUSIONS

     The findings of this  evaluation support the  opinion that the measurement
 of  "acid-soluble"  metal is less rigorous than other established methods used
 to  measure  metal  concentration in multi-phase samples.  Since the
 measurement includes both  the  dissolved metal fraction as well as an
 extracted portion  of metal  from the particulate  material,  it is by
 definition  a new parameter and is affected by the  operational technique
 used.   The  concentration of metal extracted will vary  with pH and is not
 uniform between  metals.  An example is  an  81 percent increase in lead
 concentration  at pH 1.5 over that extracted at pH  2.0, while for cadmium
 there is  less  than  a 7 percent increase for the  same change in pH.   The
 technique is precise when  the  sample  is acidified  to a specific pH range of
 narrow  limits  (1.75 ฑ 0.1)  and remains  at that pH  during the extraction
 period.   The pH  selected for extraction of lead  and chromium is most
 critical.   The concentration of metal  extracted  also is affected by the
 holding period after pH adjustment.   However, with the exception of chromium
 the  evaluation data indicates the rate of extraction at pH 1.75 and 22ฐC
 decreases considerably beyond  96 hours  (4 days)   and there is usually less
 than a  20 percent  increase  in concentration between 16 and 96 hours of
 extraction.   Although the  amount of particulate material  in the sample,  up
 to 2.5  grams/liter, did not appear to  affect extraction efficiency,  other
 factors that do  affect the concentration measured  are acidification at the
 time of sampleocollection to preserve  the dissolved metal  fraction,  and
 transport at 4ฐC to reduce chemical activity.  Of the ten metals  evaluated
 the  data  indicate that the solution chemistry of silver and mercury are  not
 suited to "acid-soluble"  metal measurement.

                                RECOMMENDATIONS

    From the compiled data and observations made during this evaluation  of
Method 200.1, certain changes are suggested and  further work recommended
before Method 200.1 is adopted as acceptable methodology for freshwater
"acid-soluble"  metals analyses.  These recommendations  are as follows:

1.  Continue to analyze silver and mercury  as "total"  or  "total recoverable"
    metals and  set limits  on that basis. The vapor pressure of
    organic-mercury compounds will most  probably result in partial  losses
    during filtration under vacuum.  Also,  silver chloride and other
    compounds of silver and mercury are  relatively  insoluble in the

-------
                                                                        529
    specified acid conditions,  subject to the dynamics of sample matrix and
    in some cases either lost or not recovered as a result of using  the
    "acid-soluble" criteria.

2.  Develop an improved method for mercury analysis with a lower detection
    limit.  The water quality criteria limit set for "acid-soluble"  mercury
    is more than an order of  magnitude below the detection limit of  the
    present cold vapor mercury methodology.

3.  Evaluate and verify that  interlaboratory precision and percent recovery
    data for "acid-soluble" graphite atomization atomic absorption analyses
    will meet the criteria limits listed in Table 1.

4.  Verify that the determinations for "acid-soluble" metals analyses as
    described in Method 200.1, Version 1.4 can be experimentally correlated
    to known aquatic toxicity.

5.  Develop separation methods for the analysis of stable oxidation  states
    of arsenic, chromium and  selenium and the organometallic species of
    arsenic, mercury, selenium and tin as required.

                              EXPERIMENTAL  DESIGN

    In planning the experimental work for the evaluation of Method 200.1,
the allowable concentration given in 50 FR 30784 was considered.  For most
metals the concentration limit for both freshwater and saltwater is very low
and a level most  suited to analysis by graphite atomization atomic
absorption.  Also, the allowable "acid-soluble" concentration of cadmium
(Cd), trivalent chromium (Cr+3), copper (Cu), and lead (Pb) in freshwater
vary with the determined concentrations of hardness as calcium carbonate.
Listed in Table 1  are examples of the expected limits for different types of
water.  For the testing and evaluation of the procedure, an assumption was
made that the extraction characteristics of these metals would be similar at
higher concentrations as well as at the expected limits.  This assumption
permitted the use  of simultaneous inductively coupled-plasma atomic emission
spectrometry  (ICP) in the determinative step.

    The sample used for the evaluation was a composited sample of Ohio river
water spiked with  particulate material and an acidfied-aqueous spike of
silver  (Ag),  arsenic (As)  and mercury  (Hg).  The particulate material was
uniform and  homogeneously mixed with a moisture content of approximately 50
percent.  The particulate material, taken from an industrial sludge stock,
was added to the  aliquot of river water to a concentration of 5
grams/liter.  The addition of 5 grams/liter yielded a test solution
containing  2.5 grams per liter of particulate material.  All test solutions
were  prepared  in  this manner, except those used  for the generation of
precision  and percent recovery data.   In this case the sample solutions were
prepared  at  various  levels of particulate concentration.

-------
                                                                          530
           TABLE 1.  EXAMPLES OF ACID SOLUBLE METAL CRITERIA LIMITS*
Average Concentration, pg/L
Metal
Arsenic"1"3
Arsenic"1"5
Cadmi urn
50 mg CaCOs/L
100 mg CaC03/L
200 mg CaCOa/L
Chromium"1"3
50 mg CaC03/L
100 mg CaCOs/L
200 mg CaCOs/L
Chromium*6
Copper
50 mg CaC03/L
100 mg CaC03/L
200 mg CaCOs/L
Lead
50 mg CaCo3/L
100 mg CaCOs/L
200 mg CaC03/L
Mercury
Freshwater
4 Day
190

••^
0.66
1.1
2.0

120
210
370
11
•_ป
6.5
12
21
wm —
1.3
3.2
7.7
0.012
1 Hr
360

__
1.8
3.9
8.6

980
1700
3100
16
__
9.2
18
34
vimrM
34
83
200
2.4
Saltwater
4 Day 1 Hr
36 69

9.3 43
— __
_ _ __
— —

__ __
__ ซ._
— —
50 1100
2.9

_ซ __
— —
5.6 140
M_ 	
_— __
	 	
0.025 2.1
* FR, Volume 50, Number 145, July 29, 1985

-------
                                                                         531
    Before using the sludge particulate material in this study, it was
analyzed by ICP in the EMSL-Cincinnati laboratory and by the USEPA Region 4,
Analytical Support Branch.  In the Cincinnati Laboratory, a 2 gram aliquot
of wet sludge was prepared for analysis using a combination nitric-
hydrochloric acid reflux, while the Region 4 laboratory used a 1 gram
aliquot of dried sludge and the nitric acid-hydrogen peroxide digestion
(Method 3050) given in SW-846, Test Methods for Evaluating Solid Waste.  The
results, reported on a dry weight basis of those determinations constitute a
"total" analysis and the comparative data are given in Table 2.  It should
be noted that over the period of time needed to complete the evaluation, the
stock sludge with exposure to the atmosphere from repeated weighings lost
moisture causing a gradual increase in metal concentration on a wet weight
basis.  To validate the precision and percent recovery data at the time of
analysis, four additional replicate aliquots were analyzed just prior tothat
aspect of the evaluation.  These concentrations are reported as "total"
values in  Table 15 under "total acid reflux" and expressed in mg/L for
comparison to the "total recoverable" data.

    Since "acid-soluble" metal analysis includes both the dissolved metal
fraction as well as an acid extracted portion of metal from the particulate
material, it is by definition a new parameter and affected by the
operational technique used.  Although it is the intent that the measurement
be less rigorous than "total recoverable" analysis, it must be rugged for
the determination to be meaningful.  The following list of variables were
considered and tests were conducted to determine the significance of each:

1.  The conditions of filtration including type of filter material, cleaning
    procedure, prefiltration, and filter prewash requirements;

2.  The effect of varying pH and whether allowable range limits would be
    required around the selected pH;

3.  The effort of varying the time period of extraction following
    acidification;

4.  The need for acid preservation at the time of sample collection and
    whether specified transport conditions to the laboratory would be
    required;

5.  The effect varying amounts of particulate material would have on
    extraction efficiency.

    The test solutions for each experimental phase were prepared in acid
cleaned cubitainers.  The amount of particulate material needed was weighed
into the cubitainers and the appropriate volume of river water added.  If a
soluble spike was to be added to the solution, it was done immediately prior
to acidification.  The volume of (1+1) nitric acid added to each type of
solution for pH adjustment was predetermined and verified after addition.
Following extraction and filtration the pH was again measured to verify that
the pH was maintained and constant throughout the extraction period.

-------
                                                                           532
         TABLE 2.  INDUSTRIAL SLUDGE DATA - ACID REFLUX TOTAL ANALYSES
Metal
Ag
Al
As
Cd
Cr
Cu
Hg
Ni
Pb
Zn
Concentration, wg/gram*

U.S. EPA Region 4 EMSL - Cincinnati
Anal. Support Branch Phys. & Chem. Methods Branch
6
3800
19(a)
58
7000
840
3(b)
360
450(c)
3400
10
3300
21
67
7000
920
2
370
530
3600
(a) Determined by graphite furnace atomic absorption.
(b) Determined by cold vapor atomic absorption.
(c) Determined by flame atomic absorption.
*Reported on a dry weight basis.

-------
                                                                           533
    The concentrations of all  spiking solutions and laboratory control
standards used in this evaluation were verified with a quality control  check
sample obtainted from the EMSL-Cincinnati, Quality Assurance Branch.   In
addition to control standards, method blanks,  and river water control
samples were prepared and processed with the test solutions.

    In the final phase of the evaluation, precision, and percent recovery
data were generated by varying the amount of particulate material in  the
test solution.  In addition to the river water control, seven replicate
solutions were prepared for each of three different levels of particulate
material concentration.  To calculate percent recovery, the average river
water control value was subtracted from the average test solution
concentration and compared to the expected concentration as determined from
the "total recoverable" analyses of the same samples.
                            RESULTS AND DISCUSSION
Conditions of Filtration
    In the first phase of the evaluation various types of filters were
examined.  This examination included both the 0.45 vm filters and the
prefilters that were considered necessary to the procedure.  It was
important to determine that the specified filtering material could withstand
the mild acid (0.2% v/v HNOs) used for extraction and not contaminate the
samp]le with trace impurities.  To determine if filter contamination would
be a problem, each filter was evaluated separately.  Also, to check the
resistance to acid, a more concentrated acid blank (1..535 v/v HNOs) was
used to extract the filter.  A 10 ml aliquot of the acid blank was recycled
through the filter three times while allowing one minute of contact with the
filter each time before vacuum was applied.  This aliquot was then analyzed
by ICP to determine if any metals of interest were extracted.  The types of
filters tested were the following:

Course Prefilters

1.  Gelman glass fiber filter:  Type AE
2.  Gelman 5 ym PVC membrance:  Type VN-1

Fine Prefilter

1.  Gelman PVC arcylic copolymer:  Type DM-800

Fine Filters

1.  Mi 11ipore mixed esters of cellulose:  Type  HAWP-047
2.  Gelman PVC acrylic copolymer:  Type DM-450
3.  Gelman PP/PTFE  (Teflon):  Type TF-450

    The  filtering  apparatus  used  throughout  this  evaluation was  a
polysulfone  Gelman  47 mm Magnetic Filter Funnel #4201  and  500 ml suction
flask.   The  magnetic  lock between the  funnel  housing  and base was  most

-------
                                                                           534
 convenient and easy to use.   The manufacturer notes  that  polysulfone  is
 resistant to nitric and hydrochloric  acids,  but  not  to chromic  nor sulfuric
 acids.   Cleaning of the funnel,  as  with  the  suction  flask, was  effectively
 accomplished using a detergent wash,  rinsing with water followed by a dilute
 nitric  acid rince and copious amounts of deionized distilled water.  Also,
 rinsing the apparatus with copious  amounts of deonized distilled water
 between samples appears to be sufficient cleansing to avoid cross
 contamination as long as a previous sample does  not  contain an  oily phase.
 A  key factor in keeping contamination to a minimum was to never allow the
 filtering apparatus or receiving labware to  dry  without being properly
 rinsed  or cleaned before reuse.   Since the rubber stopped on the funnel stem
 comes in contact with the suction flask  over which the filtrate is poured,
 to eliminate contamination,  it was  wrapped 1 inch PTFE laboratory tape.

    Of  the fine filters (0.45 ym) tested only the Gelman PP/PTFE type proved
 unacceptable.   This filter would not  wet with sample contact alone, but
 required initial  wetting with methanol before aqueous filtration would
 occur.   Since adding an extra step  to the procedure  seemed impractical, the
 use of  the teflon filter is  not  recommended.

    The other two types of fine  filters,  the PVC acrylic copolymer and the
 mixed esters  of cellulose, both  proved acceptable.  The cellulose filter was
 able  to withstand the acid because  of the relatively dilute acid used and
 limited contact time.   The extract  analyses  from both type of filters gave
 values  near the acid  blank and below  the instrumental detection limits.

    Although  the  PVC  acrylic  copolymer fine  prefilter (0.8 ym)  also proved
 acceptable giving similar analytical  results  to that of a 0.45  pm filters,
 the same was  not  true  for the coarse  prefilter.  The extract analysis of the
 PVC type VM-1  prefilter indicated that significant contamination of aluminum
 (Al)  and zinc  (Zn) may occur.  Concentrations of approximately 0.5 mg/L of
 both  elements  were found  in the  extract.  Also, over 0.3 mg/L of Al  was
 detected in the extract  of the glass fiber type AE filter.  Since Al  and Zn
 may eventually be added  to the list of "acid-soluble" metals, the need for a
 coarse  prefilter  was  reconsidered.

    In  the preliminary work of this evaluation the three filter system was
 observed  to be  very bulky and in some cases did not allow a good seal
 between  the funnel housing and base.  This resulted in leaks and affected
 the rate  of filtration. The actual need for coarse prefiltration to
 eliminate  clogging of the 0.45 ym filter  is most probably minimal.   On this
 basis and  because possible contamination and problems of leaking during
filtration, the use of the coarse prefilter was removed  from the procedure.

    To  check the  possibility that metals may be leached  from the cubitainer
during  transport, seven cleaned  cubitainers were subjected to a five  percent
nitric  acid leach for 28 days.  At the end of the period,  the acidified
deionized  distilled water extract was concentrated by evaporation  and
analyzed.  These results verified that any metals that may be leached  from
the cubitainer were at a concentration insignificant  to  ICP analyses.   The
determined concentration of all  metals in the leaching solution were  below
the ICP   instrumental detection limits.

-------
                                                                            535
    Although the combination 0.8 ym prefilter and 0.45 pin fine filter
revealed no contamination from the filtering material, a 50 ml sample
prewash of the filtering apparatus was used to remove traces of the
deionized water rinse and to condition the membrane filters.  The
combination of the 0.8 wm prefilter with either of the 0.45 pm filters was
equally effective in filtering the prepared samples used in this
evaluation.  The time of filtration for an acidified sample following sample
prewash was less than 2 minutes for a 200 ml aliquot decanted from a sample
that had settled no longer than one-half hour.

Effect of Varying pH

    Once the various aspects of the filtering process were determined, the
effect of varying pH was investigated.  For determining the effect of pH on
the extraction of "acid-soluble" metals, six levels of hydrogen ion
concentration (pH 1.0, 1.5, 1.75, 2.0, 3.3 and 7.0) were selected.  The wide
pH range and points on either side of 1.5 and 2.0 were selected in response
to concerns for investigating the least rigorous extraction while providing
a practical and rugged method.  The intent of those researchers who
suggested the concept of "acid-soluble" metals was to provide a measurement
where the sample would be acidified to a low enough pH for the sufficient
time to dissolve the carbonates, hydroxides and metal precipitates without
leaching or dissolving substantial quantities of the metals occluded in
minerals, clays and sand or strongly sorbed to particulate material.
Originally pH 4 was selected for extraction of "acid-soluble" metals,
however, it was changed to pH range 1.5 to 2.0 because of the buffering
capacity of the carbonate system of ambient waters.  The adjustment of
samples to the higher pH of 4, in practice, is more time consuming and
difficult to accomplish since this is the lower boundary of the
carbonate-bicarbonate buffering system..

    The concentrations of "acid-soluble" metals extracted at each selected
pH level are given in Table 3.  Four replicate samples were prepared for
each pH.  Each set of samples was prepared in the same manner by spiking
river water with particulate material (5 grams/liter) and an acidified-
aqueous spike of Ag, As and Hg to total concentrations of 0.050, 0.15 and
0.1 mg/L, respectively.  All samples were mixed and allowed to extract for
16 hours at 22ฐC.  In reviewing the data, it  is apparent that precision
between replicates is very good with the highest relative standard deviation
at concentrations greater than 5x the instrumental detection limit being 6.5
percent for Al at pH  1.5.  Also,  it is obvious from the data that varying
the pH produces trends  in the measured concentration  of all metals with some
being more drastic than others.   This effect  is best  appreciated from the
graphic  illustration  of Table 6 data shown in Figures 1 through 4.

-------
in
                                           CO
                              tvj
                                       CM
                CM
             If) in
             10 O
               •   •
             CM O
                •H
                                                                 co o
                                                                   •   •
                                                                 CM O
                                                                                 in
                                                                              co oo
                                                                              CM O
                                                                                 •H
                                          Ch


                                      CO O
                                        •  *

                                      i-H O
                                          •H
                                                   LO Cn
                                                   If) O
                                                   O O
         o
         0
         CM
         CM
Cn LD
to i-H
r-. o
CM CO
r^ CM
LO O
               •f
                                                                 Cn CO
                                                                 LO i-H
                         LO CM
                         i-l CM
                         CO O
             i-H CO
             in o
             o o
                                                      0
                                                                                                       O  I
          re

         00
         LO
                                       LO CO
                                       ซd- O
                                       en o
                                         •  •
                                       o o
             i-H CM
             en o
                                                    o o
             CM LO
             O i-l
             CO O
              •   •
             O O
                -H
                         in co
                         CM CM
                         CO O
                           •  •
                         o o
             r-. in
             o o
             LD o
              •   •
             o o
             CO ซ3-
             r*- o
             o o
                                                                                                       o o
                                       LO ซd-
                                       CO O
                                       o o
                                         •  •
                                       o o
                                           •H
             CM CO
             CO O
             O O
               •   •
             o o
                •H
             cn co
             CO O
             o o
              •   •
             o o
                -H
                         LO CM
                         co o
                         o o
                           •  •
                         o o
                             -H
             (Ft CO
             LO O
             o o
              •   •
             o o
                •H
             in in
             CO O
             o o
              •   •
             o o
         o
         UD
   CJ1
O t-H
co o
  •   •
CM O
   •H
   CD
•* 1^
CM O
  •   •
CM O
   •H
             o o
              •   •
             CM o
                •H
   CM
cn 10
O O
  •   •
CM O
   -H
                                                                                              CO
                                                                                           i-H CO
                                                                                           LO O
             CM CM
             CM O
             O O
              •   •
             O O
                •H
          Q.
                                           in
                                       CO t-l
                                           •ป
                CO
             o t->.
             
                             O)
         UJ

         CO
         00
                                           in
                                 .      to ^j-
                              r-       in o
                                           ฐ
   LO
ซ* CM
cn CM
  •   •
CO O
                          LO cn
                          LO o
                                                                 co
                                                                    0
                                      in o
                                                                              co
                                      LO ^~
                                      o o
                                                                                           CM
         co

         UJ
cr> 1-4
o o
o o
o o
o o
             LO i-l
             o o
             o o
CO t-H
o o
O O
CO i-H
o o
O O
  •   •
o o
                                                                                                                       -o
                                                                                                                        at
                                                                                                                        4-ป
                                                                                                                        cu
                              (O
                              00
                                                 in
                                        (O
                                        a>
                                           00
                                                        
                                                                                              00
                                                    (O
                                                3:  a>
                                                                                                           00

-------
                                                                             537
ui
CJ
                                                               1.75
                                   SAMPLE EXTRACTION, pH
                   FIGURE 1. fletal Extraction as a Function of Saraole oH -
                             Aluminum (Al), Zinc (Zn) and Copper      '

-------
                                                                           538
i
UJ
o


I   0.4
                 7.0
              FIGURE 2.
5.0
3.3
2.0  I 1.5 1.0

   1.75
                               SAMPLE EXTRACTION, pH
                                          ^a Function of Sample  oH  -
                               (Nv), Lead (Pb) and Cadmium (Cd)   '

-------
                                                                             539
o
        20.0
        18.0.
        16.0
        14.0
        12.0
        10.0
LU
O

5
o
8.0
         6.0
          4.0
          2.0
                     7.0
                            5.0
—i	ป—i—i	1	
 3.3        2.0  I 1.5   1.0
                                                                1.75
                                      SAMPLE EXTRACTION, pH
                      FIGURE 3. Metal Extraction as a Function of Sample pH
                               Chromium (Cr)

-------
                                                                                540
UJ

1
o
         0.20
         0.18
          0.16-
          0.14
          0.12
          0.10
0.08
          0.06
          0.04
          0.02-
                     7.0
                           5.0
                            : M'.•.,!
3.3
                                  SAMPLE EXTRACTION, pH
                                                                        o  As
                                                                    a
2.0 I 1.5  1.0
   1.75
                     FIGURE 4. Stability of Soluble Spike as a Function of Sample oH-
                              Arsenic  (As), Mercury  (Hg) and Silver (Ag)

-------
                                                                              541
    In Figure 1 the extraction trend exhibited for the Cu and Zn are very
similar.  Concentration increases as the pH is lowered, but for pH 2 and
below there is little change in the amount of metal extracted.  While for
Al, the other metal in Figure 1, the concentration extracted continues to
increase as the pH is lowered, but still only amounts to a 5 percent
increase in concentration between pH 2.0 and 1.5.

    In Figure 2, Cd displays the least effect from pH change while nickel
(Ni) shows a 10 percent increase in concentration between pH 2.0 and 1.5.
However, when the change in Ni concentration is compared to that of Pb, it
is relatively minor.  The increase in Pb concentration extracted between pH
2.0 and 1.5 is 81 percent.  Of all the metals investigated the extraction of
Pb is most affected by slight changes in acid below pH 3.

    In Figure 3, the effect on Cr extraction is  illustrated by itself
because of the high Cr concentration in the prepared sample.  As can be
seen, the effect of varying pH on Cr is also extreme being second only to
Pb.  The change in Cr concentration extracted between  pH 2.0  arid 1.5 is 41
percent.

    In  Figure 4 the three metals As, Hg and Ag are displayed.  For As there
is a  slight increase  in concentration,  but since the As  concentration in the
particulate material  is very  low, a conclusive statement on  extraction  can
not be  made.   However, it is  apparent that the As soluble  spike  is not  lost
as  is that of  Hg.  For Hg,  as the pH is lowered, its solubility  decreases.
Speculation on  this occurrence is that  the lowering of pH  solubilized  a
matrix  constituent of the particulate material which reacted with the
soluble Hg spike forming  insoluble mercury compound.   The  loss  of the Ag
spike is  attributed to the  precipitation  of  silver chloride.   For silver,
there is  a slight  increase  in concentration  as pH is  lowered, but less  than
 20% of  the possible 0.05  mg/L is solubilized  at  pH'1.0.   It  should  be  noted
from  the  data given  in Table  15 that both the Hg and  Ag  spikes  are  recovered
 using the "total  recoverable" analyses.

     Since both Pb  and Cr demonstrate extreme  changes  in the concentration  of
 metal extracted for slight  changes in  pH, more exacting pH limits were
 incorporated into  version 1.3 of Method 200.1.   The  midpoint (pH 1.75)  of
 the "acid-soluble" pH range (1.5 to 2.0)  was selected with a required
 tolerance limit of ฑ0.1  pH  units.  The midpoint  was  selected because the
 dissolved metals can be acid preserved by eliminating the carbonate
 buffering capacity of ambient waters with little likelihood that the sample
 pH will actually reach or go below 1.75.   Also,  this pH provided less
 rigorous data while also being rugged.

 Effect of Extraction Time

     Having decided on pH 1.75 ฑ0.1 for the extraction of "acid-soluble"
 metals, the length of contact or extraction time was investigated.   Samples
 were prepared as described earlier, adjusted to  pH 1.75, mixed, and held at
 22ฐC for various periods of time (1, 4, 16, 96,   168 and 600 hours).  The

-------
                                                                              542
 concentrations  of  "acid-soluble" metals extracted for each time period are
 given  in Table  4 and  illustrated in Figures 5 through 8.

    With the  exception  of Ag  and Hg, the data in Table 4 indicate a gradual
 increase in concentration with  increased extraction time.  For example there
 is  for all metals,  except Cr, less than a 10 percent increase in
 concentration between the 4 and 16 hour extraction time.  Also, only Al, Cr
 and Zn show an  approximate 20 percent increase in concentration between 16
 and 96 hour (4  day) extractions, and except for Cr and Zn, there is little
 increase in concentration beyond four days.  From these findings and for the
 convenience of  receiving samples and processing them in the laboratory, a 16
 hour extraction has been included in Method 200.1.

    Since extended  holding of the sample after collection and acid
 preservation  may increase the concentration of "acid-soluble" metals, a
 limit  of 3 days for sample collection and transport at 4ฐC, has been
 incorporated  into the method.   It is assumed that once the sample is
 received into the laboratory and equilibrated to room temperature,
 processing, including pH adjustment and the 16 hour extraction period will
 begin.  The data in Table 4 should be evidence that a time period of two to
 three  days needed for sample collection and transport should not greatly
 affect  an "acid-soluble" metal determination using Method 200.1.

 Sample  Preservation and Transport

    In  the present USEPA metal methods, except for Cr+6, all samples are
 acid preserved  at the time of collection and shipped to the laboratory in
 the most convenient manner.   Actual conditions of transport and shipment
 time are not  considered critical since the metals are either already in
 solution or will be subjected to an acid digestion or solubilization prior
 to  analysis.  However, in determining "acid-soluble" metals sample
 preservation  and transport conditions should be of prime concern,  since the
 dissolved fraction must be preserved while the extraction of metals  from the
 particulate material must be minimized during transport.  These two
 procedural  aspects of the method are the least controllable variables.

    To determine the significance of both acid preservation and sample
 transport conditions on the determination of "acid-soluble" metals,  a series
 of tests were devised in an attempt to reflect the various types of
 environmental  conditions that might occur.   To simulate transport, the
 samples were  divided into three groups and  held in storage for three days,
 each group at a different level  of temperature (4ฐC,  22ฐC and 49ฐC). To
 simulate acid preservation,  half of the samples were acidified to  pH 1.75
 before being  placed in storage,  while the other half were maintained at a
neutral pH until the start of sample processing.   In  these tests,  following
mock transport,  the extraction period used  was 16 hours at pH 1.75 and
 22 C.    (Note:    It should be realized when reviewing  the data from  these
tests  that the acidified samples actually experienced the equivalent of a 4
day or 96 hour extraction).

-------
3
^ LO
3 O LO









O
o
CM
CM
| —


LO
[•-.
rH
n:
D-

1

UJ

i — i
I—
Z.
O
1 —
o

I—
X
UJ
CD
z.
1— )

=ฃ
>

1—
HH
3

oo
	 i
=i
UJ

UJ
	 i
CO

oo

f — \
H-l
o



•
ซ=!•
1 1 1
1
CO

1 —














CM O
c • •
NJ CM O


O CM
CM rH
t~\ ^- CO
Q_ • .
O O



O LO
LO CM

y • •
O O
•H

rH CM
CO O
Crt O O

O O
•H
en rH
E CM r^
3 en o
ซ o • •
I "ฐ
•p
re
S- CM
.p ซ* LO
C S- LO rH
co O • •
o •* o
!= -H
o
o

r^ LO
CM O
-O rH O
o • •



t-x LO
rH O
tO rH O

ฐ'ฐ*



en
ซ3- CM
1 — CM rH
^c * •
oo o
-H


o o
rH O
Oil O O
^f 1 * *
O O
-H



c •
o >
•i- CO

O CO S- C •
re E 3 re -o
S- •!- O CO -P
-PI — n: s: oo
X
UJ rH

CO rH O
ซ3- LO r^ LO LOLO r^ LO OLO
OOO OOO COO COO rHrH

CMO CMO CMO CMO "3- O
-H -H •ซ ฃ
to
CO
cnvo cnco COCM ซ* co LOCO re
CMLO LOrH ^- LO o-* cnoo o
^- o ซ* o ซ* o LOO . ^- o, . -r~
OO OO OO OOOO Q-
•ซ •" •" -i

CO
rHrH CMLO OCn l^f-s COOO CO
CnOO OrH CTlCM Cr>rH OOOO S
i^ o coo coo coo en o -t->
. . • . • . • . •• CO
OO OO OO OO OO -Q
•H -H -H <„
3
re
Looocnoo • • • >
LO O OO O Q O Q
oooo • • • . S>
.. • • z. \ z. \ z. \ en
oooo z
•H -H c? .
OO ป ^- -stCM LOCO El
OOO OOO ^O ซ*O LOO -i— LO
rHO rHO rHO rHO rHO _JCM

OO OO OO OO OO C~O
•H -H -H ฐ *=
•r- re •
-l-> C
o >> o
co re -i—
LD LO CM ซ* - — • LO OO rH rH -P "O -(->
OO rHO CM OO rHO COlre
rH O rH O *— ' rH O rH O C2 *3- -t-
• • •• | •• •• >
OO OO OO OO i — S-CO
•fl -H -H ซ 0 T3
c -a o>
CO "OS- r-
CT QJ fQ f~}
en rH LO 3 s- TS re
CMcn Locn OCPI LOCTI oorH s_rec'—
LO i— i LOO ซ*o LOO OrH -pa.re-r-
• • •• • • • • •• toco-pre
OOO OOO ^l"O ซ!l" O LOO Ci-tO>
•H -tt • •ซ i— i o- re
M—
V to O -P
CO 0
OrH LOrH .'• ซ T3" — COC
rHO OO Q, Q O CUQ-O
oo oo • • "•• • -pErere
.. • . z' I 2:1 2:1 o re i— 4->
oooo cotocxre
m •ซ -P _ -o
co o c:
*-. *—. Q 3 ••- 0
. — . • -— ป --^ • rH tO +J •!-
• +J > i— 1 . — . tO > 	 >, +-> C C
> jz & — • vt to>>cotoreco o>>coco
co tocnQ to>) co s-re Q s-Qccn z i — > to
 ป_-co o Q •— ป --^
LOO LO LO O ซrHCM
^J. rH^^ CTl rH LO Z. 	 	 •

-------
                                                                                    544
UJ
M.
O
 7.0.
 6.5-
 6.0 -
 5.5-
 5.0
 4.5
 4.0
 3.54
 3.0
   •
 2.5
 2.0
 1.5.
 1.0 •
0.5
0
            1116
             4
                 96
168
                                                                     600
                               EXTRACTION HOLDING TIME,  HOURS
                FIGURE 5. Metal  Extraction as  a  Function of Sample Holding Time
                          Aluminum (Al),  Zinc  (Zn)  and Copper (Cu)

-------
                                                                                  545
O
i—i
ง
UJ
o
O
o
1.4
1.3
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
         Pb
                                                                           -ฉ
            l] 16     96      168~
                                                                      600
                              EXTRACTION HOLDING TIME, HOURS
             FIGURE 6. Metal Extraction as a Function of Sample Holding Time
                       Nickel (Ni),  Lead (Pb)  and Cadmium (Cd)

-------
                                                                                      546
UJ
o
                                                                           6QO
                                 EXTRACTION HOLDING TIME,  HOURS
               FIGURE  7.  Hetal  Extraction as a Function of Saroole Holding Time
                         Chromium (Cr)                                   •

-------
                                                                           547




^
^
o
ป— 1
t:
CONCENT*/




0.26
0.24
0.22
0.20
0.18
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0







0
ฉ ฉ
Hg o
\
\
'A'-^V, T
             96
168
                                                                 600
                      EXTRACTION HOLDING TIHE, HOURS
FIGURE 8. Stability of Soluble Spike as a Function of Saronle Holding  Time
          Arsenic (As), Mercury (Hg) and Silver (Ag)      '

-------
                                                                             548
    To test the stability of the dissolved metal fraction under the
described conditions, one set of samples was prepared by spiking Ohio river
water with a soluble spike containing all ten metals.  To test the
extraction from particulate material alone, a second set of samples was
prepared where river water was spiked with weighed aliquots of the sludge.
To test the two fractions together, a third set of samples was prepared
where the river water was spiked to contain both the soluble spike and
sludge particulate material.  In each set, duplicate samples were prepared
for each condition described, and as in previous experiments, the
particulate material was spiked into river water to a concentration of 5
grams/liter.

    Because the data from these tests are extensive and are best reviewed by
comparing the results from each of the selected conditions, the data for
each metal are given in individual tables.  Included in the "acid-soluble"
data of the prepared samples are the analyses of a river water control and
laboratory control standard.  Also included in the tables for comparison are
the results of direct analysis of the spiking solution diluted in two
different acid matrices, as well as the "total recoverable" analyses of the
laboratory control standard and the river water containing the soluble
spike.  The mean value listed for each analysis is the average of the
duplicates with the range value being their difference.  The data for all
ten metals are contained in Tables 5 through 14.  Each metal will be
discussed in the same order as the tables.

Aluminum - Table 5.

    The analyses data of the laboratory control standard show very good
agreement between the "acid-soluble" and "total recoverable" analyses.  The
data also agree with the spiking solution which gave the same analyzed value
in both acid matrices.

    As might be expected, the "acid-soluble" data for the acid preserved
samples gave higher extracted concentrations than the non-preserved
samples.  In both groups, the range values appear to be less than 10 percent
of the mean indicating relatively good precision with the acid preserved
samples showing somewhat better agreement.

    The acid preserved samples stored at 49ฐC yielded considerably higher
extracted concentrations than those stored at 4ฐC.  Increases in
concentration ranged from 20 to 50 percent.  Since there is a similar
increase in all three types of sample-spike combinations, the increase
appears to be due to Al extracted from the river water.  A comparison of the
river water control to total recoverable analyses of the river water plus
soluble spike support this statement, since only 20 percent of the Al in
river water was made available using Method 200.1.  In the non-preserved
samples, the Al concentration showed aoslight decrease for those samples
stored at the elevated temperature (49ฐC).

    For Al the need for acid preservation is not confirmed, but sample
transport and storage below 22ฐC is required until the time of sample
processing.  This is most practically accomplished by cooling with ice to
4ฐC.

-------
                                                            549
TABLE 5.  THREE DAY SAMPLE HOLDING PERIOD AT
          VARIOUS TEMPERATURES - ALUMINUM
Concentration,
Acid Soluble Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Ohio River + Spike (49ฐC)
Ohio River + Sludge ( 4ฐC)
Ohio River + Sludge (22ฐC)
Ohio River + Sludge (49ฐC)
River + Sludge + Spike ( 4ฐC)
River + Sludge + Spike (22ฐC)
River + Sludge + Spike (49ฐC)
Total Recoverable Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Direct Analysis of Spike Solutions
Theoretical Spike Concentration
Analyzed Spike Concentration
Diluted in 0.2% (v/v) HN03 (pH 1
Diluted in 1.0* (v/v) HN03 + 5%
mg/L

Acid Preserved Not Acid Preserved
Mean Range Mean Range
0.406
2.08
2.17
2.55
4.23
6.23
6.60
8.25
6.70
7.02
8.45
-0.410
10.2
10.8


.5)
(v/v) HC1
0.008
0.02
0.04
0.01
0.19
0.03
0.09
0.03
0.08
0.24
0.14
0.003
0.05
0.81
Value
0.400

0.404
0.403
_ —
1.58
1.89
1.87
1.64
5.42
5.32
5.21
5.96
5.71
5.38

8.51
11.6



	
0.02
0.04
0.04
0.02
0.16
0.16
0.44
0.19
0.17
0.23
._
0.06
0.03




-------
                                                                             550
Arsenic - Table 6

    From prior acid-reflux analyses the determined As concentration in the
sludge is very low. In fact, if all were extracted by the "acid-soluble"
procedure, the concentration would be only 0.056 mg/L.  From the data given
in Table 6 it would appear that 50 percent of the As is extracted,  but the
poor precision between duplicates and low concentration of the analysis does
not permit a conclusive statement.

    The highest recovery of the soluble As spike is only 87 percent.   Acid
preservation does not appear to be necessary because it does not stabilize
the dissolved fraction, since storage at the elevated temperature (49ฐC) in
both situations indicates additional losses.  This consistently low recovery
cannot be explained, but the "total recoverable" analyses indicate  the loss
to be as a precipitate and not a loss to the walls of the cubitainer.

    For As the need for acid preservation is not confirmed, but sample
transport and storage below 22ฐC is required to reduce chemical activity and
loss of the dissolved fraction.

    An important note to the As analysis is the 75 percent recovery of the
laboratory control standard.  During evaluation the As analysis of  the
control standard was erratic.  This behavior was found to be a phenomenon of
the ICP analysis and was attributed to the difference in the concentration
of the acid matrix between the extracting acid solution and the calibration
standard.  (See the direct analysis of the spike solution.)  It is  assumed
this problem did not occur in prepared river samples because of the
naturally occurring chloride and higher dissolved solids concentration.

Cadmium - Table 7

    The data for Cd given in Table 7 indicate that there is little
difference between the acid preserved and non-preserved samples and that
temperature control during transport is not critical.  Although there is a
slight increase (7%) in the Cd extracted from the sludge material during
acid preservation, Cd was the least affected by simulated acid preservation
and the mock transport conditions.
Chromium - Table 8

    The data given in Table 8 for Cr indicate that both acid preservation
and temperature control are critical.   Without'acid preservation,  the
dissolved fraction is partially lost (13%)  as is shown in the non-preserved
river water plus spike data.  Also, this loss (37%) is accelerated  as  the
temperature is increased to 49ฐC.  Although there is an increase in the
extracted "acid-soluble" concentration because of acid preservation, the
data indicate the increase is limited to less than 16 percent when  the
samples are stored .at 4ฐC over the three day holding period.

    The conclusion for Cr "acid-soluble" analysis is that for valid
determinations both acid preservation and storage below 22ฐC are recommended.

-------
                                                                             551
                 TABLE  6.  THREE  DAY  SAMPLE  HOLDING  PERIOD  AT
                          VARIOUS  TEMPERATURES  - ARSENIC
Concentration, mg/L
Acid Soluble Analyses
Lab. Control Spike Std.
Ohio River
Ohio River + Spike
Ohio River + Spike
Ohio River + Spike
Ohio River + Sludge
Ohio River + Sludge
Ohio River + Sludge
River + Sludge + Spike
River + Sludge + Spike
River + Sludge + Spike
Acid Preserved
Mean Range
(22ฐC)
(22ฐC)
( 4ฐC)
(22 C)
(49ฐC)
( 4"0C)
(22 C)
(49UC)
( 4ฐC)
(22 C)
(49ฐC)
0.296
0.016
0.360
0.350
0.315
0.020
0.032
0.026
0.366
0.362
0.339
0.021
0.023
0.004
0.004
0.009
0.011
0.011
0.000
0.013
0.031
0.004
Not Acid Preserved
Mean Range
	
N.D.
0.353
0.342
0.318
N.D.
0.016
0.022
0.356
0.334
0.283
	
	
0.007
0.007
0.004
	
0.003
0.004
0.027
0.012
0.002
Total Recoverable Analyses
Lab.
Ohio
Ohio
Control
River +
River +
Spike Std.
Spike
Spike
(22ฐ
( 4ฐ
(22
C)
C)
C)
0
0
0
.421
.435
.433
0
0
0
.013
.006
.004

0
0
	
.428
.431
	
0.000
0.002
Direct Analysis of Spike Solutions
Value
Theoretical Spike Concentration

Analyzed Spike Concentration

  Diluted  in 0.2% (v/v) HN03  (pH 1.5)
  Diluted  in 1.0% (v/v) HNOs  + 5%  (v/v) HC1
0.400
0.302
0.409
 N.D. -  Not  Detected < 0.016 mg/L

-------
                                                             552
TABLE 7.  THREE DAY SAMPLE HOLDING PERIOD AT
           VARIOUS TEMPERATURES - CADMIUM
Concentration,
Acid Soluble Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Ohio River + Spike (49ฐC)
Ohio River + Sludge ( 4ฐC)
Ohio River + Sludge (22ฐC)
Ohio River + Sludge (49ฐC)
River + Sludge + Spike ( 4ฐC)
River + Sludge + Spike (22"C)
River + Sludge + Spike (49ฐC)
Total Recoverable Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Direct Analysis of Spike Solutions
Theoretical Spike Concentration
Analyzed Spike Concentration
Diluted in 0.2% (v/v) HMOs (pH 1.5)
Diluted in 1.0% (v/v) HNOa + 5% (v/v)
mg/L

Acid Preserved Not Acid Preserved
Mean Range Mean Range
0.163
0.001
0.165
0.162
0.164
0.150
0.150
0.159
0.310
0.301
0.307
0.160
0.160
0.158


HC1
0.001
0.000
0.002
0.000
0.000
0.005
0.004
0.002
0.004
0.006
0.003
0.001
0.001
0.001
Value
0.160

0.166
0.162
___
0.001
0.160
0.162
0.156
0.140
0.142
0.143
0.296
0.291
0.286

0.155
0.158



___
0.001
0.004
0.001
0.001
0.001
0.001
0.001
0.013
0.009
0.009

0.000
0.000




-------
TABLE 8.  THREE DAY SAMPLE HOLDING PERIOD AT
           VARIOUS TEMPERATURES - CHROMIUM
                                                             553
Concentration,
Acid Soluble Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Ohio River + Spike (49ฐC)
Ohio River + Sludge ( 4ฐC)
Ohio River + Sludge (22ฐC)
Ohio River + Sludge (49ฐC)
River + Sludge + Spike ( 4ฐC)
River + Sludge + Spike (22ฐC)
River + Sludge + Spike (49ฐC)
Total Recoverable Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Direct Analysis of Spike Solutions
Theoretical Spike Concentration
Analyzed Spike Concentration
Diluted in 0.2% (v/v) HN03 (pH 1.5)
Diluted in 1.0% (v/v) HN03 + 5% (v/v
mg/L

Acid Preserved Not Acid Preserved
Mean Range Mean Range
0.167
0.007
0.173
0.173
0.177
6.81
6.97
11.1
7.01
7.85
10.1
0.164
0.181
0.179


r) HC1
0.001
0.003
0.009
0.006
0.006
0.25
0.55
0.24
0.45
0.25
0.16
0.002
0.001
0.001
Value
0.160

0.166
0.166
	
0.005
0.149
0.130
0.108
5.88
5.34
5.28
6.46
5.84
5.29
...
0.173
0.182



	
0.001
0.007
0.008
0.001
0.26
0.26
0.38
0.03
0.09
0.10
	
0.001
0.002




-------
                                                                             554
Copper - Table 9

    The data for Cu given in Table 9 show good agreement between
duplicates.  The data indicate there is little difference, less than 10
percent, between the acid preserved and non-preserved samples, when
temperature is controlled to 4ฐC.

    In the non-preserved river water plus spike samples stored at 49ฐC there
is a loss (13%) in the dissolved fraction of the samples.  However,, when all
the non-preserved sample data are reviewed, a conclusion of loss because of
elevated temperature is not supported.  Also, a comparison of the
"acid-soluble" data to the "total-recoverable" data for the non-preserved
river water plus spike samples indicate not all Cu is made available for
analysis using Method 200.1, as is the case in acid preserved samples.  The
complete extraction of Cu is attributed to the longer acid contact time
(96 hours) of acid preservation (see Table 4).

    These data indicate that for Cu neither acid preservation nor controlled
temperature during transport and storage are required.

Lead - Table 10

    The data for Pb given in Table 10 show good agreement between the
laboratory control standard and the other standard solutions analyzed.
Comparison of the "acid-soluble" concentration to the "total recoverable"
analyses of the river water plus spike indicate no loss of the soluble spike
or dissolved fraction.

    The Pb concentration in the acid preserved river water plus sludge is
very similar to that in the non-preserved sample with acceptable precision
between duplicates.  As expected, if the temperature is increased to 49ฐC
during storage, the Pb extracted from the particulate material in acid
preserved samples increases.  However, this is not true for the
non-preserved sample.  In fact, although not conclusive, there appears to be
some indicationoof loss in the non-preserved samples with elevated storage
temperature (49ฐC).

    The data in Table 10 indicate for consistent and valid "acid-soluble" Pb
analyses, sample transport and storage at 4ฐC is recommended.  With the
temperature controlled at 4ฐC the increase in "acid-soluble" Pb
concentration from acid preservation over the three day holding period is
limited to less than 5 percent.

Mercury - Table 11

    The data for Hg given in Table 11 are evidence that Hg as an
"acid-soluble" metal can not easily be analyzed.  Even the acid-preserved
soluble spike at 4ฐC was not stable for the three day holding period giving
a recovery of less than 80 percent.  Also, when the soluble spike was
combined in the river water with the sludge particulate material, recovery
was further reduced to less than 50 percent.  In this evaluation only the
"total  recoverable" analyses of Hg appeared to be acceptable.

-------
TABLE 9.  THREE DAY SAMPLE HOLDING PERIOD AT



           VARIOUS TEMPERATURES - COPPER
                                                            555
Concentration,
Acid Soluble Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Ohio River + Spike (49ฐC)
Ohio River + Sludge ( 4ฐC)
Ohio River + Sludge (22ฐC)
Ohio River + Sludge (49ฐC)
River + Sludge + Spike ( 4ฐC)
River + Sludge + Spike (22ฐC)
River + Sludge + Spike (49ฐC)
Total Recoverable Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Direct Analysis of Spike Solutions
Theoretical Spike Concentration
Analyzed Spike Concentration
Diluted in 0.2% (v/v) HNOa (pH 1.5)
Diluted in 1.0% (v/v) HMOs + 5% (v/v
mg/L
Acid Preserved Not Acid
Mean Range Mean
0.396
0.010
0.414
0.420
0.415
2.28
2.30
2.34
2.73
2.70
2.72
0.384
0.420
0.415


) HC1
0.001
0.001
0.002
0.006
0.004
0.03
0.08
0.03
0.05
0.10
0.01
0.006
0.002
0.001
Value
0.400

0.392
0.393
___
0.008
0.394
0.393
0.341
2.11
2.16
2.08
2.57
2.54
2.40

0.418
0.432




Preserved
Range
	
0.001
0.005
0.001
0.001
0.02
0.04
0.05
0.08
0.11
0.09

0.020
0.017




-------
                                                             556
TABLE 10.  THREE DAY SAMPLE HOLDING PERIOD AT
           VARIOUS TEMPERATURES - LEAD
Concentration,
Acid Soluble Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Ohio River + Spike (49ฐC)
Ohio River + Sludge ( 4ฐC)
Ohio River + Sludge (22ฐC)
Ohio River + Sludge (49ฐC)
River + Sludge + Spike ( 4ฐC)
River + Sludge + Spike (22ฐC)
River + Sludge + Spike (49ฐC)
Total Recoverable Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Direct Analysis of Spike Solutions
Theoretical Spike Concentration
Analyzed Spike Concentration
Diluted' in 0.2% (v/v) HN03 (pH 1.5)
Diluted in 1.0% (v/v) HN03 + 5% (v/v!
mg/L

Acid Preserved Not Acid Preserved
Mean Range Mean Range
0.405
0.020
0.431
0.422
0.422
0.441
0.434
0.520
0.785
0.776
0.841
0.389
0.439
0.434


) HC1
0.002
0.016
0.017
0.027
0.033
0.044
0.021
0.038
0.020
0.018
0.005
0.012
0.021
0.013
Value
0.400

0.413
0.406
____
0.025
0.424
0.427
0.393
0.440
0.393
0.405
0.751
0.759
0.646

0.420
0.434



.MM,,^
0.019
0.033
0.022
0.002
0.042
0.013
0.005
0.054
0.013
0.017

0.000
0.012




-------
                                                                            557
                TABLE 11.  THREE DAY SAMPLE HOLDING PERIOD AT
                           VARIOUS TEMPERATURES - MERCURY
Acid Soluble Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Ohio River + Spike (49ฐC)
Ohio River + Sludge ( 4ฐC)
Ohio River + Sludge (22ฐC)
Ohio River + Sludge (49ฐC)
River + Sludge + Spike ( 4ฐC)
River + Sludge + Spike (22ฐC)
River + Sludge + Spike (49ฐC)
Total Recoverable Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Direct Analysis of Spike Solutions
Theoretical Spike Concentration
Analyzed Spike Concentration
Diluted in 0.2% (v/v) HN03 (pH 1.5)
Diluted in 1.0% (v/v) HN03 + 5% (v/v)

Acid
Mean
0.163
N.D.
0.126
0.116
0.083
N.D.
N.D.
N.D.
0.067
0.038
N.D.
0.163
0.169
0.165


HC1
Concentration,
mg/L
Preserved Not Acid
Range Mean
0.005
—
0.007
0.003
0.002
—
0.020
0.018
0.004
0.002
0.004
Value
0.160

0.272
0.164
u^ulll 	 ,
N.D.
0.070
0.044
0.015
N.D.
N.D.
N.D.
0.036
0.030
N.D.

0.157
0.145




Preserved
Range
t JMVU
—
0.001
0.001
0.003
—
0.003
0.002

0.006
0.004


•
N.D. - Not Detected <  0.015 mg/L

-------
                                                                           558
    An additional observation on the analysis of Hg from this evaluation is
that there is a significantly different response in the ICP analysis of the
spiking solution diluted in the two different acid matrices.  In the dilute
nitric there is a 60 percent increase in the mercury signal.  Over a period
of 2 to 3 days reanalysis showed the increase to Tessen, and if then mixed
with hydrochloric acid, the response equaled that of the calibration
standard.  Unfortunately, if Hg spiking solutions are freshly prepared in
0.2 percent nitric acid and then immediately mixed with hydrochloric acid
the response remains elevated.  This situation appears to have a chemistry
and reaction rate not understood and merits additional study.

Nickel - Table 12

    The data for Ni given in Table 12 are straightforward, reliable and are
similar to that of Cu.  Only non-acidopreserved river water with a soluble
spike at the elevated temperature (49ฐC) showed a decrease in
concentration.  The effect of acid preservation on samples stored at 4ฐC was
less than a seven percent increase over non-preserved samples stored at the
same temperature.  Therefore, for the determination of "acid-soluble" nickel
acid preservation is recommended.

Silver - Table 13

    As in the case of Hg, the data given in Table 13 for Ag indicate it also
is not suited to "acid-soluble'J analysis.  When spiked into river water,
acid preserved and stored at 4ฐC, Ag recovery approximated only 80 percent.
When combined with the sludge particulate material recovery dropped to 30
percent.  Only the "total recoverable" analyses of Ag gave consistent and
reliable data.

Zinc - Table 14

    The data given in Table 14 for Zn exhibit similar characteristics to
data of other metals.  There is good agreement between the laboratory
control standard and the direct analysis of the spiking solutions.  The
difference between "acid-soluble" analyses of acid preserved samples and
non-preserved samples stored 4ฐC is less than 10 percent.   The soluble spike
in the non-preserved samples stored at 49ฐC is partially lost, while acid
preserved samples stored at the same temperature show an increase.
Therefore, the data given in Table 14 for "acid-soluble" Zn analyses
indicate samples should be acid preserved and stored at 4ฐC until time of
processing.

    The following is a summation regarding sample preservation for
"acid-soluble" analyses:

    1.   Ag and Hg are not suited to "acid-soluble" analysis.  Acid
         preservation does not prevent loss from the dissolved fraction.
         Additionally, these metals are not extracted with dilute acid and
         recoveries are poor unless analysis is completed  as "total" or
         "total recoverable".

    2.   Although dissolved As will remain in solution without acid
         preservation, losses will  occur even with acid preservation if

-------
                                                           559
TABLE 12.  THREE DAY SAMPLE HOLDING PERIOD AT
           VARIOUS TEMPERATURES - NICKEL
Concentration,
Acid Soluble Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Ohio River + Spike (49ฐC)
Ohio River + Sludge ( 4ฐC)
Ohio River + Sludge (22ฐC)
Ohio River + Sludge (49ฐC)
River + Sludge + Spike ( 4ฐC)
River + Sludge + Spike (22ฐC)
River + Sludge + Spike (49ฐC)
Total Recoverable Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Direct Analysis of Spike Solutions
Theoretical Spike Concentration
Analyzed Spike Concentration
Diluted in 0.2% (v/v) HN03 (pH 1.5)
Diluted in 1.0% (v/v) HNOs + 5% (v/v
mg/L
Acid Preserved Not Acid
Mean Range Mean
0.409
0.019
0.434
0.431
0.438
0.932
0.929
0.964
1.35
1.32
1.35
0.404
0.454
0.440


) HC1
0.000
0.003
0.001
0.002
0.003
0.020
0.029
0.014
0.03
0.03
0.02
0.000
0.021
0.002
Value
0.400

0.414
0.406
—
0.016
0.409
0.388
0.315
0.872
0.865
0.857
1.27
1.26
1.18

0.428
0.490




Preserved
Range
___
0.000
0.016
0.011
0.004
0.008
0.020
0.020
0.06
0.00
0.01

0.001
0.019




-------
                                                                           560
                TABLE 13.  THREE DAY SAMPLE HOLDING PERIOD AT
                           VARIOUS TEMPERATURES - SILVER
Concentration, mg/L
Acid Soluble Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22eC)
Ohio River + Spike (49ฐC)
Ohio River + Sludge ( 4ฐC)
Ohio River + Sludge (22ฐC)
Ohio River + Sludge (49ฐC)
River + Sludge + Spike ( 4ฐC)
River + Sludge + Spike (22ฐC)
River + Sludge + Spike (49ฐC)
Total Recoverable Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Direct Analysis of Spike Solutions
Theoretical Spike Concentration
Analyzed Spike Concentration
Diluted in 0.2% (v/v) HN03 (pH 1.5)
Diluted in 1.0% (v/v) HNOs + 5% (v/v)
Acid Preserved
Mean Range
0.042
N.D.
0.035
0.033
0.030
N.D.
0.003
0.005
0.014
0.012
0.005
0.042
0.046
0.046


HC1
0.000
	
0.004
0.002
0.001
0.001
0.000
0.000
0.003
0.001
0.002
0.000
0.000



Not Acid Preserved
Mean Range
ซ_.
0.003
0.023
0.020
0.011
N.D.
0.003
N.D.
0.005
0.005
0.004

0.044
0.044
Value
0.043

0.043
0.043
___
0.000
0.003
0.000
0.002
0.001
0.001
0.002
0.002

0.000
0.001



N.D. - Not Detected < 0.003 mg/L

-------
                                                           561
TABLE 14.  THREE DAY SAMPLE HOLDING PERIOD AT
           VARIOUS TEMPERATURES - ZINC
Concentration,
Acid Soluble Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Ohio River + Spike (49ฐC)
Ohio River + Sludge ( 4ฐC)
Ohio River + Sludge (22ฐC)
Ohio River + Sludge (49ฐC)
River + Sludge + Spike ( 4ฐC)
River + Sludge + Spike (22ฐC)
River + Sludge + Spike (49ฐC)
Total Recoverable Analyses
Lab. Control Spike Std. (22ฐC)
Ohio River + Spike ( 4ฐC)
Ohio River + Spike (22ฐC)
Direct Analysis of Spike Solutions
Theoretical Spike Concentration
Analyzed Spike Concentration
Diluted in 0.2$ (v/v) HNOs (pH 1.5)
Diluted in 1.0% (v/v) HNOs + 5% (v/v
mg/L

Acid Preserved Not Acid Preserved
Mean Range Mean Range
0.161
0.099
0.261
0.263
0.283
2.86
2.95
3.62
2.99
3.03
3.82
0.153
0.316
0.314


) HC1
0.002
0.003
0.002
0.000
0.004
0.07
0.05
0.06
0.07
0.05
0.10
0.002
0.007
0.001
i
Value
0.160

0.163
0.159
• 	
0.088
0.243
0.238
0.205
2.65
2.66
2.64
2.81
2.80
2.69

0.301
0.355



	
0.001
0.007
0.001
0.001
0.02
0.01
0.03
0.05
0.04
0.09

0.005
0.034




-------
                                                                            562
          samples  are  transported  or  stored  at elevated temperatures  (49ฐC).

     3.    Of  the elements  investigated Cd does not appear to be adversely
          affected by  varying  storage temperatures or lack of acid
          preservation.

     4.    The remaining  six metals  (Al, Cr,  Cu, Ni, Pb and Zn) all had losses
          from the dissolved fraction, when  samples are not acid preserved
          and the  storage  temperature is increased (49ฐC).  Only Cr showed a
          loss (13%) for a non-preserved sample stored at 4ฐC for three
          days.  For samples that were acid  preserved and stored at 4ฐC, the
          increase in  concentration of Cu, Ni, Pb and Zn was less than 10
          percent  while  for Al and Cr, it was near 15 percent.

     Since the dissolved metal fraction is most available to aquatic  life and
should have  the greatest  toxicological impact, acid preservation is
recommended,  especially for Cr.  To  reduce  extractability and chemical
reactivity during transport,  samples should be stored at 4ฐC until
processing can begin.   These  two provisions have been incorporated into
Method 200.1  Acid preservation is accomplished by adding 2 ml of (1+1)
nitric acid  to approximately  800 ml  of sample at the time of collection.  It
has  been  determined that  this amount of acid will not lower deionized
distilled water below pH  1.75.  Samples are to be held at 4ฐC and processing
must begin within three days  after sample collection.

Effect of Particulate Material
    In the final phase of the evaluation the effect that varying amounts of
particulate material have on extraction efficiency was investigated.  To
accomplish this, four different levels of the sludge particulate material
were added to river water and preserved as instructed in version 1.4 of
Method 200.1.  To some samples, soluble spikes of Hg, Ag and As were added
because the concentration in the sludge material was too low for analysis.
Seven replicates were prepared for each level as well as for the river water
control.  To calculate percent recovery of the "acid-soluble" analyses,
average concentration found in the river water control was subtracted from
the average value determined for each level.  These net values were then
compared to the expected net concentration as determined from the "total
recoverable" analyses of the same samples.  The analytical data from these
determinations are given in Tables 15 through 17.  Also included are "total"
or estimated "total" analyses data as calculated from the acid-reflux
digestions.

    A review of the standard deviation data for "acid-soluble" analyses
given in all three tables shows the precision between the replicates to be
very good.  Although the extraction efficiency is not the same for each
metal, comparison of the mean values for each metal across the various spike
levels indicates extraction is uniform for varying amounts of particulate
material.  In fact, the "acid-soluble" metal extraction was more uniform
across the prepared samples tested than the "total recoverable" analyses.

-------
                                                                           563
             TABLE  15.   ACID  SOLUBLE METAL PRECISION AND PERCENT
                        RECOVERY DATA -2.5 grams/L


Concentration,
mg/L

Acid Soluble*
Metal
Ag
AT
As
Cd
Cr
Cu
Hg
Ni
Pb
Zn
Total
Acid Reflux
0.015
5.13
0.028
0.098
10.1
1.26
N.D.(b)
0.525
0.810
5.46
Total
Recoverable
0.012
4.83
0.031
0.085
7.48
1.17
0.114(c)
0.466
0.601
3.43
Mean
0.003
2.18
0.007U)
0.074
3.18
1.11
0.029
0.457
0.311
1.35
Std. % Total
Dev. Recoverable
ฑ0.002
ฑ0.047
ฑ0.001
ฑ0.002
ฑ0.085
ฑ0.025
ฑ0.004
ฑ0.008
ฑ0.010
ฑ0.018
25%
45%
23%
87%
43%
95%
25%
98%
52%
39%
* Data are from 7 replicate samples

(a)  Arsenic filtrate concentrated 4x before analysis
     instrumental detection limit = 0.003 mg/L.

(b)  N.D. - Not detected < 0.015 mg/L.

(c)  A soluble mercury spike of 0.125 mg/L was added to the sample
     before processing.

-------
                                                                           564
             TABLE 16.  ACID SOLUBLE METAL PRECISION AND PERCENT
                        RECOVERY DATA - 1.0 grams/L
Concentration, mg/L
Metal
Ag
AT
As
Cd
Cr
Cu
Hg
Ni
Pb
Zn
Est. Total
Acid Reflux
0.031U)
2.05
0.323(b)
0.039
4.04
0.505
0.063(c)
0.211
0.324
2.19
Acid Soluble*
Total
Recoverable
0.026
1.74
0.328
0.036
3.28
0.502
0.065
0.194
0.268
1.50
Mean
0.014
0.870
0.313
0.030
1.21
0.442
0.021
0.191
0.137
0.548
Std. % Total
Dev. Recoverable
ฑ0.001
ฑ0.045
ฑ0.007
ฑ0.002
ฑ0.070
ฑ0.019
ฑ0.002
ฑ0.008
ฑ0.008
ฑ0.017
54%
50%
95%
83%
37%
88%
32%
98%
51%
37%
* Data are from 7 replicate samples
(a)  Silver value includes a soluble spike of 0.025 mg/L.
(b)  Arsenic value includes a soluble spike of 0.313 mg/L
(c)  Mercury value includes a soluble spike of 0.063 mg/L.

-------
                                                                                   565
                     TABLE  17.   ACID  SOLUBLE  METAL  PRECISION AND  PERCENT
                                RECOVERY  DATA - 0.25  grams/L
Concentration, mg/L
Acid Soluble*
Metal
Ag
Al
As
Cd
Cr
Cu
Hg
Ni
Pb
Zn
Est. Total
Acid Reflux
0.052(a)
0.513
0.063(b)
0.010
1.01
0.126
0.025(c)
0.053
0.081
0.546
Total
Recoverable
0.050
0.421-
0.061
0.010
0.669
0.130
0.024
0.051
0.067
0.298
Mean
0.024
0.261
0.065
0.009
0.326
0.115
N.D.(d)
0.062
0.042
0.150
Std.
Dev.
ฑ0.001
ฑ0.019
ฑ0.003
ฑ0.001
ฑ0.032
ฑ0.006
	
ฑ0.006
ฑ0.007
ฑ0.010
% Total
Recoverable
48%
62%
103%
90%
49%
88%
	
132%
63%
50%
        * Data are from 7 replicate samples.
        (a) Silver value includes a soluble spike of 0.050 mg/L
        (b) Arsenic value includes a soluble spike of 0.063 mg/L
        (c) Mercury value includes a soluble spike of 0.025 mg/L
        (d) N.D. - Not Detected < 0.015 mg/L
_

-------
                                                                            566
Only Pb  showed  a  gradual  increase  (10%)  in extraction efficiency as the
amount of particulate material was decreased.

    As expected,  recovery of the soluble As spike was near 100 percent with
good precision.   However, recoveries of the soluble Hg and Ag spikes were
not readily  improved because of the lower concentration of particulate
material.  In each case where the concentration was measurable, precision
was good but recovery of  Hg and Ag was limited to near 30 and 50 percent,
respectively.

    Of the metals investigated that are suitable to the "acid-soluble"
analysis, seven can be grouped into two categories.  One group (Cd, Cu and
Ni) are easily extracted  and give recoveries near 90 percent.  For the other
group (Al, Cr, Pb and Zn) extraction is more affected by pH, temperature and
extraction time,  and give recoveries nearer 50 percent.  When experimental
conditions are well controlled the "acid-soluble" extraction of all of these
metals'is precise, uniformly consistent, but different. .These differences
cannot be attributed to the amount of particulate material present, but
rather to the chemical nature of these metals.  Factors that affect metal
solubility using Method 200.1 are the ease that insoluble compounds are
dissociated with the lowering of pH or the stability of a metal complex
under acidic conditions.  These factors are affected by the sample matrix,
anion constituents, and when "acid-soluble" metal concentrations are
compared to the "total-recoverable" analyses, the comparative data may vary
from sample to sample.  Although these differences are expected, the
solubilized metal  concentration in each case when using Method 200.1 is by
definition the "acid-soluble" metal concentration.

-------
                                                    567
             QUESTION AND ANSWER SESSION
                          MR. TELLIARD:  Any questions?



                          DR. GAIND:  Arun Gaind from




Nanco Labs.  The silver and mercury may have gotten



out of solution because the river water had high level




of chloride.



                          MR. MARTIN:  Yes.




                          DR. GAIND:  So, the procedure




now is going to institutionalize Ohio River water



use for this procedure?  No?  I mean...



                          MR. MARTIN:  I'm sorry, I




didn't understand you.



                          DR. GAIND:  You used...to




verify the method you used Ohio River water.



                          MR. MARTIN:  Yes.



                          DR. GAIND:  Right.  Okay,




but the procedure now when it becomes used for other



conditions, it will use DI water?



                          MR. MARTIN:  No, I don't




believe so.  It's going to be an ambient water



measurement.



                          DR. GAIND:  Okay.



                          MR. MADDALONE:  As part




of this EPRI program during this current round,

-------
                                                    568



we are doing the analysis of water for mercury, and



we talked with Gary McKee, Chief of the Inorganic



Analyses Section at EMSL-Cincinnati, and originally,



in the Federal Register a couple years ago, they



talked about using dichromate to stabilize mercury



samples.  It turned out that use of dichromate and



gold and all those other substances to stablilize



mercury, were really related to stabilized mercury in



distilled water, and we confirmed this in our labs.



     We spiked some distilled water and our ash point



overflow water.  Ash pond overflow samples were fine.



They didn't show any drop in the mercury concentration



with the normal acidification scheme, but the distilled



water showed a very significant drop in the amount of



mercury present.  So for our distilled water samples



in our program we had to put dichromate in them to



stabilize them, and that did stabilize them over a one



month period.

-------
569

-------
570

-------
                                                     571
               RAYMOND F. MADDALONE
   RESULTS OF THE TRW EPRI ANALYTICAL  METHODS
          QUALIFICATION  PROGRAM,  PART  I
                      Dr. MADDALONE:  I'm going to
discuss today  the  work that TRW  is doing  for the
Electric  Power Research  Institute.    The  program
that  we  have   from  the  Electric  Power  Research
Institute is divided into three distinct areas.
     The    first    area,    literature   studies,
identified  the  problems  and  scope  the  future
work.   As  part of  that  effort, we  collated and
performed an analysis  of all the  available data on
the  concentrations  of  parameters  and  elements in
the  aqueous  discharge  streams  from power plants.
A  report  summarizing  that  data  was written.   The
report  contains data  from open  literature studies,
and   a   summary  of   100  of   the  1980   2C   NPDES
permits.  All  of this  discharge  data was  placed in
a  computerized data  base.
      The  next  activity conducted  was to  acquire
precision and  bias data from  a  number  of sources,
including ASTM, EPA/Effluent  Guidelines  Division,
and   UWAG.    Those  data  were  then  collated  and
 reviewed   to   help   us  determine  what  were  the
 literature  values   for limits   of  detection  and
 limits   of    quantitation.      A  second   report
 summarized  those  findings.

-------
                                                    572
     A  third  report  was  written  summarizing  the
results  of  a  literature  review  of  sampling  and
analysis  methods   for  the  13 priority  pollutant
metals.   This  report  assessed  the  state  of  the
art,   provided   background    information   on   the
interferences associated with current methods,  and
looked at new methods or procedures.
SLIDE 1
     This  slide  summarizes  the  reports  that  we
have  already  published  or to  publish under  the
program.   As  you  can  see,  this  first  set  here
above the dotted  line  is  primarily the literature
work  that  we  have  done.   Below the  line we  have
the reports that  will  come out  at  the end of  each
one  of  the  field  studies.     Through   the  good
offices  of  EPA, we obtained copies  of  the  DMR/QA
data  sets   from the  first four  years.   We  are
currently analyzing those  results and will publish
a report
     The  QA/QC  guideline  document  is  directed
towards  the  utility   industry   to   provide   the
utility  chemist  information  on how  to  go  about
setting  up  a  quality  assurance  program.   We  think
that's  probably one of the more important  things
to  do,  because  in the  utility  industry,  you  have
laboratories that  range from the central  lab,  with

-------
                                                     573
the highly  qualified individuals,  to  basically a
boiler water chemistry lab that's being used to do
some environmental analyses at the  smaller plants.
     The  real  focus  of the program,  and  what  I'm
going  to be  talking about  today,  is  the method
validation  studies.    The literature  studies   led
the way  and  told us  what trace metals we  ought to
study.    We  then   set-up  a  validation  program
consisting of  four parts  called  Analytical Methods
Qualification  (AMQ).  Let me show you how that is
configured.
 SLIDE  2
      This   is  the  design  of   the   field   study
 program.   To  date,  we  have completed  Parts  I  and
 II.    Part  I  consisted  of two rounds.   During  the
 first   round  we  did   arsenic   and   selenium   by
 Graphite   Furnace  AAS,   and  in  round  2  chromium,
 copper,   nickel   and   lead  were  determined   by
 Graphite  Furnace AAS.
      We performed  these  validation  efforts in  a
 number of  different matrices,  including a  river
 water;  an    ash   pond   overflow  stream,   which
 consisted of  the  river  water  source  plus leachate
 from  the  ash pond.   We  had  a  seawater intake  and
 seawater discharge.  We  used a reagent grade water
 sample  to   provide  a   sample   with  no  matrix

-------
                                                     574
interferences while  a sample of  treated  chemical
metal cleaning wastes  provided  a  very challenging
sample for  analysis.   EPA/EMSL  provided  us with a
number of their QA/QC ampules that we used to have
an   absolute  measure   of  the   bias   of   each
laboratory.
     What  I  will  be  presenting  today   are  the
results  from the  first  part,   AMQ-I.   We  had  42
laboratories   participate   in   the   round-robin
study.   Approximately  30  of   them  were   what  we
called our freshwater laboratories.  Those are the
labs  that  analyzed  the   river water,   ash  pond
overflow,  reagent grade  water, the  QA/QC ampule
and  the  treated  chemical  metal   cleaning  waste.
The   seawater  labs   substituted   for   the   two
freshwater   streams,   a   seawater   intake  and  a
seawater  discharge.    The  data that  I'll present
today will be from the freshwater  laboratories.
     The  data  that  was   used  to  calculate  the
limits of  detection  was  processed using  standard
ASTM  precision  and bias statistics.   That is, we
used D-2777-85.   In that process you're allowed to
rank  the labs and  then  go  through  a standard t-
test  (1% double  tail).   You are  permitted to use
two  iterations of  the t-test which in  this case is
a  one percent double-tailed  t-test.  The  precision
data  obtained fram data analysis   using D-2777 was

-------
                                                     575
regressed versus the mean test concentration.   The
resulting y-intercept  (standard  deviation  at  zero
concentration) was used to compute an LOD.
     EPA/EMSL has  published  a  procedure  for  the
Method Detection  Limit  in ES&T a  number  of  years
ago for  organics.   It  has now been promulgated in
the  Federal  Register  as  part  of  the  GC  organic
methods.      The  EPA/EMSL  MDL   procedure  first
determines that the analyte concentration that you
are  using  is  between  one  and   five  times  the
estimated  limit of  detection  of the method.   If
so,  you   can then  use  that   sample directly  to
calculate  a  method  detection  limit.     If   it's
greater  than   five  times  the  method  detection
limit, but less than 10 times  the  detection limit,
you  can  still  use  the sample,  though  they  would
probably prefer for you  to find  a  sample with  a
lower  concentration.
      The MDL calculation  is based  on a  standard t-
value.   The  sample  is analyzed  seven  times  by  a
single operator,  and the  t-value  for a  one percent
single   tail  with  six  degrees  of   freedom  is
found.   The  standard deviation of the  mean of  the
 replicate  analyses is  multiplied  by that t-value
to calculate the method  detection limit.  The way
that  we  did the  MDL  calculation  under  the  EPRI
 program   utilized  the   number of   labs  and   data

-------
                                                     576
points obtained  for  each  matrix and  element.   We
used the  number  of  laboratories or the  number  of
data points, depending  on  whether  we  were dealing
with  overall  or  the  single  operator  MDLs  to
provide the degrees of freedom.
SLIDE
     As  I  mentioned,  we  also  calculated  an  LOD
using   the   precision   data   from  the   D-2777
analysis.  Each one of the samples  that we sent to
our  participants  consisted  of  four concentration
levels,  the  background  plus  three  spikes.   Using
that  data, we  calculated  a  precision  regression
equation   and  then  extrapolated  it  to  the  zero
concentration,  which  would give you the precision
at   zero   concentration   (i.e.,   the  standard
deviation  of  a  blank).   By using  that extrapolated
value,   you   can   then  apply  the  standard  ACS
definition of three  times  the  standard deviation
of  the  blank  to determine  the limit of  detection.
      All  the data  presented  today  was calculated
using   method  detection   limit.    The  limits   of
detection  based  on  the  regression equation data
are fairly similar to these  data, and  in  fact they
agreed  quite  wel1 .

-------
                                                     577
SLIDE
     This table here presents the results from the
freshwater  streams  (i.e.,   freshwater,  ash  pond
overflow,  and  reagent  grade water  samples).   The
data were very close and there is no reason not to
collate the LOD estimates from those data.  On the
other   hand,   I'll  be  showing   you   the  MDL's
calculated   from   the   treated   chemical   metal
cleaning  waste  and those   obviously  were  quite
different  from  the  freshwater1 s.   In  all  cases
they  showed  much  higher   limits  of   detection,
perhaps  indicating  a matrix  problem.
     The  data summarized in  this slide are for the
six  elements from  the  AMQ-I;  this  is the single
operator   precision  data.    We  have  the overall
precision based  limits  of detection  also.  The EPA
quoted  values  are the  values quoted by  the EPA  in
the  "Methods  of  Chemical  Analysis  for  Water and
Waste".   All the  values  on  this  table  are  in  parts
per  bi 1 1 i on .
      The EPRI  survey LOD value  is  the simple  mean
 of  the  LOD  data  that  we  collated  under  our
 literature  survey.    This  LOD  includes,  again,
 values   from  ASTM, US6S,  EPA  from  the  MCAW,  and
 EPA/EGD  "Metals  Methods Validation"  study.    The
 plus or  minus  gives you  some  indication of  what
 the range was of the values reported.
                        8

-------
                                                     578
     The  final  column  here, EPRI  field  study,  is
the results  that  we obtained from  the  EPRI  AMQ-I
study.   Those  results  were obtained using utility
laboratories  and   real   world  samples  that  were
spiked,  split,  acidified,  and  sent  out to  the
participating  laboratories.  We  tried  to simulate
as close  as  possible  the situation that occurs in
the normal NPDES sampling  exercise.
     For  the single operator data, the arsenic and
copper,   nickel,   lead  and  selenium   MDLs   are
slightly  higher than  the EPA quoted value and are
                                                  \
pretty  much  at the same level as  we  found  in the
literature  survey.   Now remember,  this  is  single
operator  data.
SLIDE
     Let's  graphically  take  a look at this single
operator  precision  data  for  graphite  furnace.
Remember  all  these LODs  are  based on the EPA MDL
calculation.   For  the freshwater matrices, you can
see  here  that  the  green  are the quoted EPA values,
and  the  orange  and  the  red  bars,  are  the  EPRI
field  study.   Again, there's  fairly good  agreement
with  the  survey  data,   but  the  calculated  MDLs
generally are  higher than  the  EPA  quoted  LODs.   In
the  1974  version  of  the MCAW,  the  EPA  quoted
values  for graphite furnace  were  on  the order  of

-------
                                                     579
0.3  ppb.    In  the  latest  revision (1983)  of  the
MCAW,  they  are  now 1.5,  which  is  a  little  bit
closer to what  we're seeing  in  terms  of  the real
world data.

SLIDE
     This slide  shows  the  overall  precision data
for  the  freshwater  streams.   Now you  can  see that
we're  starting  to  get,  both  in the  EPRI  survey
data and  in  the  EPRI field study data, a  slightly
higher    and,   in   come   cases,   much   .higher
calculations of the  limit  of  detection.  Also, you
can  see  that there  is  some  range associated with
the  values that we  have.   Arsenic and  selenium are
two  of   the  more  difficult  elements to   do  by
graphite  furnace.   In the  case  of some  of our more
difficult matrices, such  as  the seawater and the
treated   chemical   metal  cleaning  waste,   all  the
participants  complained   about   the   difficulties
that they were having.
     The  people  performing  this  work  were very
dedicated.   Even  though  we  told  them to  perform
the  analysis  according  to  exact  procedures that
are  provided by  the MCAW,  several of  participants
said,  "we were having  a great  deal   of difficulty
getting   this   element   to   recover   properly  or
getting   the instrument  to work,  so  we did it by
                       10

-------
                                                     580
Flame Atomic  Absorption or  we  did  it  by  Gaseous
Hydride  AAS."    It's   an  indication  that  these
people really want  to  do well,  but in the  case of
our study, it really didn't help us at all  because
we only wanted to evaluate Graphite Furnace AAS.

SLIDE
     Let's  show what  the  overall  precision  MDLs
look  like  graphically.  We  start to  see  a great
deal  of  divergence  in terms of  what's  quoted by
the  EPA   as  an  LOD (which  is  probably  really  a
single operator/single laboratory value)  compared
to  the overall  precision-based  limit of detection
that  we  found,  both from our field study  and from
our  EPRI  survey.  An overall  precision-based limit
of  detection is  really a better  measure  of what
goes  on   in  the  real  world, because  what you're
talking   about   is  one  laboratory  comparing   its
results  versus  another laboratory.   In the  case of
NPDES  sampling,  it  could be  the  utility  versus  the
EPA or EPA  contractor laboratory that might  have
taken  a  sample.    So  you  really  want  to   use  the
overall  precision data, in that  case, to  evaluate
the results  of  the  two laboratories.
 SLIDE
      Let's  go  to  the   treated   metal   chemical
                        11

-------
                                                     581
cleaning waste  data.   Here are  the  data  from the
treated chemical metal  cleaning  waste.   Treatment
started by adding various polymers and floculating
agents  to  the  waterside  boiler  wash.    It  was
pumped  into  a  receiving  pond,  allowed  to  stand
there  for  several   days,  and  then  it was  pumped
through  a  filter.    Analysis  samples  were  taken
until  the  typical  one ppm copper  and  iron  values
were  obtained,  and  then  that  solution  was  then
directly  discharged  into the   receiving  waters.
The  sample analyzed  in  the  EPRI  program  was the
sample  that  was  treated to  meet  NPDES  permit
di scharge  limits.
     You  can  see  here  from  the  field  study data
for  this  matrix that  we have much  higher  single
operator  values than  the EPA quoted  or  the EPRI
survey  LODs.    Furthermore,  they  are  much  higher
than  the  freshwater  values  that  I had previously
di spl ayed.
     Participants  had a  great  deal  of difficulty
with copper  and nickel  in the TCMCW samples.  The
MDL  for selenium,  3.5 ppb, is low, but I think  it
was  low because we had  a  stability problem with
the   selenium   in   the   treated  chemical    metal
cleaning   waste.    We  didn't  find  much  of  the
selenium,  and   the  precision  then was fairly good
because they were  dealing with low  concentrations.
                       12

-------
                                                     582
SLIDE
     Let's  take  a  look  at these  single  operator
data  graphically.   Note  that  there's  a  break in
the  scale.   The  yellow correspond  to  the  copper
and nickel  results, which  you  can  see have  a much
higher  MDLs in  this  matrix.    This  was  a   very
difficult matrix.   Even  though it meets discharge
requirements,  it  still probably  has  some  portion
of  the  floculating  agents, it  has  a high  calcium
value,  and  a  very  high  chloride   concentration.
So,  it's  a  very  nasty matrix, but  it's  a  matrix
that  the  utility  industry  has  to monitor  under  the
permit  laws.

SLIDE
      Here's the  table  for the LODs  based  on  the
overall  precision from the treated  chemical  metal
cleaning    wastes.      Again,   the   scale    break
correspond  to  the  bar colors.   You'll note  that
the computed  MDLs  for  copper and   nickel  show-up
very  high   and  definitely  higher  for  selenium,
arsenic,  chromium,  .and lead based  on the  data that
we obtained from  the  field validation effort.  You
can  see  that  the  literature  survey data  fits  in
 about the same level  as  the field  study data.
                        13

-------
                                                     583
     Let me  summarize  now just where  we  stand in
this program.   We  have,  as I  mentioned, completed
the two  rounds  associated with AMQ-I.   We expect
to  finish  the   report  in the   summer  1986  and
probably have a report out,  published by EPRI, by
the end  of the year.   AMQ-II has  been completed
and we've  just  received  the  last  bit of data two
weeks,  ago.   We  have   loaded  the  data onto  our
computer system, but we haven't started to analyze
it  yet.   The  AMQ-II  data  report  will  be out a bit
1ater  in the year.
     For   the   two   remaining  parts   of   the  AMQ
project, we want to  do a  study using  ICP for 10 to
15  elements.    We  also  want  to do total suspended
solids   (non-filterable   residue)   and   oil    &
grease.     Those   two   parameters,  which   are  in
everybody's  permit,  have  little  precision   data
associated with them.   Bias  data is very hard to
come  by because the ability  to  make  up a stable
 sample for oil  &  grease  and  total  suspended  solid
 is  very  difficult.
      So,  the EPRI AMQ  program is continuing.  We
 are publishing  reports   on our  validation effort
 and   reports   are  available   from   our   survey
 studies.   If  you  have  any questions,  I'll  be  happy
 to answer them  at  this time.
                    MR. TELLIARD:   Any questions?
                        14

-------
                                                      584
Thank you very much.
                        15

-------
585

-------
586

-------
587

-------
588

-------
589

-------
590

-------
591

-------
                                                             592
CD ~
 _ w CD
    Ere




O O

-------
591

-------

-------
593

-------
594

-------
595

-------
                  oo u.
                  C3 C3 ^ C3
                  in O Lf) C3
                  ro co cvj esi
O o

-------
595

-------
                  ซAU.  g
CD  ^
ป*- o
 —  03
 E  2

O  o
J0 C3 O  O
 LO C3 ITS  ^
 n ro csi  cvi

-------
597

-------
598

-------
                                                    599



                          MR. TELLIARD:  Thank you,



ladies and gentlemen.  I'd like to thank the County



Court Reporters.  They're down to one.  I'd like to



thank you for your attention, your attendance, your



patience for all the slides that you couldn't read,



and hope to see you all back here next year.



     We'll try to get the proceedings out before the



next meeting.  Thank you very much for coming.



(WHEREUPON, the proceedings were concluded at 11:45 a.m.)

-------
                                                                600
                      ROSTER OF ATTENDEES
JAMES R. ALEXANDER
SUPERVISORY CHEMIST
NORFOLK NAVAL SHIPYARD
CODE 134.2  BLDG. 184
PORTSMOUTH VA 23709
804-396-3373
                                        B. A. ALLEN
                                        MANAGER, ENVIRONMENTAL SERV.
                                        DOW CHEMICAL COMPANY
                                        P. 0. BOX BB, ENVIRONMENTAL SERVK
                                        FREEPORT TX 77541
                                        409-238-2415
MIKE ALSOP
DEVELOPMENT SCIENTIST
UNION CARBIDE CORP.
P.O. BOX 8361
S. CHARLESTON WV 25303
304-747-5467
                                        BOB APRIL
                                        ENVIRONMENTAL SCIENTIST
                                        USEPA-ITD
                                        401 M STREET, SW, (WH-565E)
                                        WASHINGTON DC 20460
                                        202-382-4654
JOE ARLAUSKAS
MARTIN MARIETTA ENVR. SYSTEM
9200 RUMSEY ROAD
COLUMBIA MD 21045
301-964-9200
                                        JOHN J. AUSTIN
                                        CHEMIST
                                        USEPA-REGION III
                                        839 BESTGATE ROAD
                                        ANNAPOLIS MD 21401
                                        301-224-2740
CLIFTON BAILEY
OURS
USEPA-ITD
401 M STREET, SW,  (WH-552)
WASHINGTON DC 20460
202-382-5411
                                        ROBERT C. BARRICK
                                        PRINCIPAL CHEMIST
                                        TETRA TECH, INC.
                                        11820 NORTHUP WAY, SUITE 100
                                        BELLEVUE WA 98005
                                        206-822-9596
ROBERT BEIMER
MANAGER, CHEMISTRY PROGRAM
S-CUBED
P.O. BOX 1620
LA JOLLA CA 92038
619-453-0060
                                        JOHN BIRRI
                                        QA CHEMIST
                                        USEPA-REGION II
                                        WOODBRIDGE AVE.
                                        EDISON NJ 08837
                                        201-321-6722
DAVE BOTTRELL
USEPA-EMSC-LV QAD
P.O. BOX 15027
LAS VEGAS NV 89104
702-798-2142
                                        C. DON BOWYER
                                        CHEMICAL LAB CERT. OFFICER
                                        NJDEP-OSR-QA (CN 409)
                                        436 E. STATE ST.
                                        TRENTON NJ 08625
                                        609-292-3950

-------
                                                               601
 STEVE  BRADY
 CHEMIST
 EASTMAN KODAK COMPANY
 KODAK  PARK.  BLDG.  34
 ROCHESTER NY 14650
 716-722-3430
 BILL BRAMSTEDT,  Ph.D
 SDS BIOTECH
 BOX 348
 PAINESVILLE OH 44077
 216-357-3350
MICHAEL BRENNAN
SALES ADMINISTRATOR
ENVIRITE ANALYTICAL SERVICES
OLD WATERBURY ROAD
THOMASTON CT 06787
203-283-8235
WADE  BROUGHTON
CHEMIST II
CITY  OP ATLANTA
2640  JONESBORO ROAD,  SE
ATLANTA GA  30316
404-627-1222
JOHN BROWN
MARINE CHEMISTRY RESEARCHER
BATTELLE N.E.M.R.L.
397 WASHINGTON ST.f BOX AH
DUXBURY MA 02332
617-934-5682
EARNESTINE BUTTONS
CHEMIST
CITY OF ATLANTA
2640 JONESBORO RD. SE
ATLANTA GA 30315
404-627-1222
CHAN CALDWELL
SUPV. OF GAS CHROMATOGRAPHY
PRINCETON AQUA SCIENCE/IT
165 FIELDCREST AVE., CN7809
EDISON NJ 08818
201-225-2000
GENE CHIU
LABORATORY DIRECTOR
TP ANALYTICAL LABS
P.O. BOX 919
CONYERS GA 30207
404-922-8000
WINSTON CHOW
PROJECT MANAGER
ELECTRIC POWER RESEARCH INST
3412 HILLVIEW AVE.
PALO ALTO CA 94304
415-855-2868
GORDON CHU
CAL LAB EAST, INC.
2240 DABNEY ROAD
RICHMOND VA 23230
804-359-1900
YEUH CHUANG
MARTIN MARIETTA ENVR. SYSTEM
9200 RUMSEY ROAD
COLUMBIA MD 21045
301-964-9200
BRUCE N. COLBY
PRESIDENT
PACIFIC ANALYTICAL, INC.
1989 B PALOMAR OAKS WAY
CARLSBAD CA 92008
619-931-1766

-------
                                                              602
DR. U. M. COWGILL
ASSOC. ENVR. CONSULTANT
DOW CHEMICAL
2030 WHDOW CENTER
MIDLAND MI 48674
517-636-1735
MARK J. CROCHET
ANALYTICAL CHEMIST
BASF CORP.
P.O. BOX 457
GEISMAR LA 70734
504-387-0631 EXT.2424
MICHAEL D. CROUCH
PRESIDENT
TOXICON LABS, INC.
3213 MONTERREY BLVD.
BATON ROUGE LA 70814
504-925-5012
JOHANNA M. CULVER
CHEMIST
NORFOLK NAVAL SHIPYARD
CODE 134.2  BLDG. 184
PORTSMOUTH VA 23709
804-396-3373
SEYED DASTGHEYB
GE/MS-SUPERVISOR
UNITED STATES TESTING CO.
1415 PARK AVE.
HOBOKEN NJ 07030
201-792-2400
WILLIAM DAVIS
MARKETING MANAGER
FINNIGAN CORP.
355 RIVER OAKS PARKWAY
SAN JOSE CA 95134
408-433-4800
SUSAN deNAGY
PROJECT OFFICER
USEPA-ITD
401 M STREET, SW  (WH-552)
WASHINGTON DC 20460
202-382-7141
KATHY J. DIEN HILLIG
RESEARCH STAFF
BASF CORP.
1419 BIDDLE AVE.
WYANDETTE MI 48192
313-246-6334
DR. THOMAS W. DUKE
SR. RESEARCH SCIENTIST
USEPA-REGION IV
ENVR. RESEARCH LAB, SABINE ISLAND
GULF BREEZE FL 32561
904-932-5311
TINA ENGEL
RESEARCH SCIENTIST
BATTELLE MEMORIAL INSTITUTE
505 KING AVENUE
COLUMBUS OH 43201
614-424-4149
BARRETT P. EYNON
STATISTICIAN
SRI  INTERNATIONAL
333  RAVENSWOOD AVE.
MENLO PARK CA 94025
415-859-5239
PAUL FARROW
VG MASSLAB
TUDOR ROAD, ALTRINGHAM
CHESIRE CD WAI4  5RZ
061-941-3552

-------
                                                              603
THOMAS FIELDING
USEPA-ITD
401 M STREET, SW, (WH-552)
WASHINGTON DC 20460
202-382-7156
RUSSELL D. FOSTER, JR.
TECHNICAL DIRECTOR
RESOURCE ANALYSTS, INC.
P.O. BOX 778
HAMPTON NH 03842
603-926-7777
DR. PAUL FRIEDMAN
USEPA OFFICE OF SOLID WASTE
401 M STREET, SW,  (WH-562-B)
WASHINGTON DC 20460
202-382-7700
DR. ARUN GAIND
EXECUTIVE VICE PRESIDENT
NANCO LABS INC.
P.O. BOX 10
HOPEWELL JCT. NY 12533
914-221-2485
DANIEL A. GARTIEZ
VESTEC CORPORATION
THOMAS J. GEYER
ENVIRONMENTAL LAB SPVR.
FORD MOTOR COMPANY
15201 CENTURY DRIVE
DEARBORN MI 48047
313-845-1648
WILLIAM E. GORE
AMERICAN CYANAMID
P.O. BOX 60
STAMFORD CT 06904
203-655-8873
KATHY GROSS
CAL LAB EAST, INC.
2240 DABNEY ROAD
RICHMOND VA 23230
804-359-1900
ROBERT E. HACKLEY
CHEMIST
BURLINGTON NORTHERN R.R.
1612 N. LEXINGTON
SPRINGFIELD MO  65802
417-864-3177
DR. EDWARD HANDEL
CHAIRMAN OF THE BOARD
CENTEC ANALYTICAL SERVICES
2160 INDUSTRIAL DRIVE
SALEM VA 24153
703-387-3995
WILLIAM C. HARRIS
LABORATORY SUPERVISOR
UNION CAMP CORP.
P.O. BOX 178
FRANKLIN VA 23851
804-569-4263
JOHN C. HENDRICKS
CHEMIST
AMERICAN ELEC. POWER &
1 RIVERSIDE PLAZA
COLUMBUS OH 43215
614-223-1238
SERV.

-------
                                                              604
DIANA HOLLEY
SR. ENVIRONMENTAL ENGINEER
ONION CARBIDE-INSTITUTE PLT.
P.O. BOX 2831
CHARLESTON WV 25330
304-747-1885
MIN 6. BU
GROUP LEADER, IND. ANALYSIS
ROLLINS ENVR. SERVICES
2027 BATTLEGROUND
DEER PARK TX 77536
713-479-6001
WILLIAM P. HUFF
TECHNICAL DIRECTOR
ANDREW S. McCREATH & SON
301 WEST FREEMASON ST.
NORFOLK VA 23510
804-622-6136
WILLIAM IMBUR
MANAGER, TESTING & MEASURE.
LOW ENVIRONMENTAL SERVICES
2749 DECK ROAD, SE
MARIETTA GA 30067
404-952-9005
PETER ISSACSON
CHEMIST
VIAR AND COMPANY
300 N. LEE ST., SUITE 200
ALEXANDRIA VA 22314
703-683-0885
RICHARD A. JAVICK
SR. RESEARCH CHEMIST
FMC CORPORATION
P.O. BOX 8
PRINCETON NJ 08540
CHARLOTTE W. K1MBROUGH
CHEMIST
MARTIN MARIETTA ENERGY SYST.
P.O. BOX Y
OAK RIDGE TN 37831
615-574-3715
TODD KIMMELL
USEPA-OFFICE OF SOLID WASTE
401 M STREET, SW, (WH-562-B)
WASHINGTON DC 20460
202-382-4795
JAMES R. KING
SAMPLE CONTROL CTR. COORD.
VIAR AND COMPANY
300 NORTH LEE ST., SUITE 200
ALEXANDRIA VA 22314
703-683-0885
JOHN A. KIRCHSTEIN
TECHNICAL SERVICES ENGINEER
E.I. DuPONT
ENGINEERING TEST CENTER
WILMINGTON DE 19898-7104
302-366-3608
DIANNA KOCUREK
PARTNER
TISCHELER-KOCUREK
116 E. MAIN ST.
ROUND ROCK TX 78664
512-244-9058
MARCIA KUEHL
DONOHUE & ASSOCIATES, INC.
4738 N. 40TH
SHEBOYGAN WI 53081
414-458-8711

-------
                                                             605
SUJITH KUMAR
TECHNICAL DIRECTOR
INTERNATIONAL TECH. CORP.
5103 OLD WILLIAM BERN HWY.
EXPORT PA 15632
412-731-8806
ROBERT LANNAN
ATTORNEY
ROBINSON & MCELWEE
P.O. BOX 1791
CHARLESTON WV 25326
304-344-5800
PATIRCIA LASKA
CHEMIST
USEPA-EMSL-LAS VEGAS
944 E. HARMON AVE.
LAS VEGAS NV 89109
702-798-2671
LINDA LAUGHLIN
LAB COORDINATOR
MARTIN MARIETTA ENVR SYSTEMS
9200 RUMSEY ROAD
COLUMBIA MD 21045
301-964-9200
EDWARD LAWLER
SENIOR CHEMIST
CAMBRIDGE ANALYTICAL ASSOC.
1106 COMMONWEALTH AVE.
BOSTON MA 02215
617-232-2207
SIM D. LESSLEY
ASSOCIATE DIRECTOR
UBTL, INC.
520 WAKARA WAY
SALT LAKE CITY UT 84108
801-584-3232
CONNIE LESTER
CHEMIST
CITY OF ATLANTA
2440 BOLTON RD. NW
ATLANTA GA 30318
404-351-6120 EXT.284
HAROLD LESTER
MARKETING DIRECTOR
POLICY PLANNING & EVAL., INC
8521 LESSBURG PIKE, SUITE 310
VIENNA VA 22180
703-893-0315
NATHAN LEVY
PRESIDENT
A&E TESTING, INC.
1717 SEABORD, SUITE 3,03
BATON ROUGE LA 70810
504-769-1930
JAMES W. LEWIS
LABORATORY MANAGER
BIONETICS CORP.
20 RESEARCH DRIVE
HAMPTON VA 23666
804-864-0680
HARRIS A. LICHTENSTEIN, Ph.D
PRESIDENT
SPECTRIX CORPORATION
3911 FONDREN, SUITE 100
HOUSTON TX 77063-5821
713-266-6800
DENIS LIN
ETC CORP.
P.O. BOX 7808
EDISON NJ 08818-7808
201-225-6759

-------
                                                              606
TOM A. LOWE
MGR. ANALY. & SURFACE CHEM,
KAISER ALUMINUM
P.O. BOX 877
PLEASANTON CA 94566
415-847-4626
JEFFREY C. LOWRY
IND. CS/MS SUPERVISOR
ROCKY MT. ANALYTICAL LAB
5530 MARSHALL ST.
ARVADA CO 80002
303-421-6611
SAMUEL V. LUCAS
BATTELLE COLUMBUS LABS
505 KING AVE.
COLUMBUS OH 43201
614-424-4178
LARRY LaFLEUR
ORGANIC PROGRAMS MANAGER
NCASI
720 S.W. 4TH STREET
CORVALLIS OR 97333
503-754-2015
CHRISTINE MACKO
ENVIRONEMTNAL ENGINEER
SAIC
8400 WESTPARK DIRVE
MCLEAN VA 22102
703-734-3106
RAYMOND F. MADDALONE, Ph.D
PROGRAM MANAGER
TRW SPACE & TECHNOLOGY GRP.
ONE SPACE PARK
REDONDO BEACH CA 90278
213-536-2451
GORDON DEAN MARBURY
SENIOR CHEMIST
TRIANGLE LABORATORIES, INC.
4915-F PROSPECTUS DRIVE
DURHAM NC 27713
919-544-5729
DR. MARK F. MARCUS
DIRECTOR, ANALYTICAL CHEM.
CHEMICAL WASTE MGMT.
150 W. 137TH STREET
RIVERDALE IL 60627
312-841-8360
THEODORE D. MARTIN
CHEMIST
USEPA/EMSL-CINTI
26 W. ST. CLAIR
CINCINNATI OH 45208
513-569-7312
ROBERT METTER
LABORATORY SERVICES MANAGER
BECKMAN INDUSTRIAL CORP.
4141 PALM ST.
FULLERTON CA 92635
714-447-2326
DR. DEBORAH S. MILLER
UNION CARBIDE CORP.
P.O. BOX 8361, B770-318
S. CHARLESTON WV 25314
304-747-4463
RAYMOND F. MINDRUP
MKT. DEVELOPMENT SPECIALIST
SUPELCO, INC.
SUPELCO PARK
BELLEFONTE PA 16823-0048
814-359-3441

-------
                                                              607
SANDY MOSER
COUNTY COURT REPORTERS, INC.
30 SOUTH CAMERON ST.
WINCHESTER VA 22601
703-667-0600
NEIL H. MOSESMAN
RESEARCH CHEMIST
SUPELCO INC.
SUPELCO PARK
BELLEFONTE PA 16823
814-359-3441
DR. JOHN J. MOUSA
MANAGER OF CHEMISTRY DIV.
ENVR. SCIENCE & ENG., INC.
P.O. BOX ESE
GAINESVILLE FL 32602-3053
904-332-3318
JOHN S. MURRAK
EMS LABS
7901 W. MORRIS ST.
INDIANAPOLIS IN 46231
317-298-4146
ALEXANDER C. McBRIDE
CHIEF,WATER QUALITY ANALYSIS
USEPA-ITD
401 M ST., SW.,  (WH-553)
WASHINGTON DC 20460
202-382-7046
KEVIN McGONNAGHY
COMPUCHEM LABORATORIES
P.O. BOX 12652
RESEARCH TRIANGLE PARK NC 27709
919-549-8263
DONNA L. McGOVERN
CHEMIST
NORFOLK NAVAL SHIPYARD
CODE 134.2  BLDG. 184
PORTSMOUTH VA 23709
804-396-3373
DR. JOHN McGUIRE
USEPA-ATHENS E.R.L.
COLLEGE STATION RD.
ATHENS GA 30613
404-546-3185
GEORGE ODELL
MANAGER/ORGANICS LAB
NANCO LABS INC.
P.O. BOX 10
HOPEWILL JCT. NY 12533
914-221-2485
VIJAYAMOHAN K. PALAT
LAB MANAGER
CHEM CLEAR INC.
992 OLD EAGLE SCHOOL RD.
WAYNE PA 19087
215-687-8990
JERRY L. PARR
DIRECTOR OF TECH. SERVICES
ROCKY MT. ANALYTICAL LAB
5530 MARSHALL ST.
ARVADA CO 80002
303-421-6611
GARY PETRAZZUOLO
TECHNICAL RESOURCES, INC.
3202 MONROE ST. SUITE 300
ROCKVILLE MD 20852
301-231-5250

-------
                                                              608
CHARLES PLEST
SENIOR CHEMIST
USEPA-OFFICE OF ACID RAIN
401 M.. STREET, SW
WASHINGTON DC 20406
202-382-5796
RICHARD POSNER
VICE PRESIDENT
UNITED STATES  TESTING CO.
1415 PARK AVE.
HOBOKEN NJ  07030
201-792-2400
DAVID H. POWELL, Ph.D
MANAGER, GC/MS
ENVR. SCIENCE & ENG., INC.
P.O. BOX ESE
GAINESVILLE FL 32602-3053
904-332-3318
WILLIAM B. PRESCOTT
CONSULTANT
724 HAWTHORNE AVENUE
BOUND BROOK NJ  08805
201-469-1198
JAMES K. RICE
JAMES K. RICE CHARTERED
17415 BATCHELLORS FOREST RD
OLNEY MD 20832
301-774-2210
ROSS K. ROBESON
DIRECTOR, ENVR. MARKETING
COMPUCHEM LABORATORIES
P.O. BOX 12652
RESEARCH TRIANGLE PARK NC 2770!
919-549-8263
ANN ROSECRANCE
ORGANIC LABORATORY DIRECTOR
JTC ENVR. CONSULTANTS, INC.
4 RESEARCH PLACE
ROCKVILLE MD 20850
301-921-9790
DARCEY ROSENBLATT
TECHNICAL RESOURCES, INC.
3202 MONROE ST. SUITE 300
ROCKVILLE MD 20852
301-231-5250
CURTIS ROSS
CENTRAL REGIONAL LAB DIR.
USEPA-REGION V
536 SO. CLARK ST.
CHICAGO IL 60604
312-353-8370
DALE R. RUSHNECK
INTERFACE INC.
P.O. BOX 297
FT. COLLINS CO 80522-0297
303-223-2013
EDWARD H. SANDERS, Ph.D
TECHNICAL MANAGER
UBTL, INC
520 WAKARA WAY
SALT LAKE CITY UT 84108
801-584-3232
ROBERT B. SCHAFFER
VICE PRESIDENT
CENTEC CORP.
11260 ROGER BACON DR.
RESTON VA 22090
703-471-6300

-------
                                                              609
ANDREW SCHKUTA
MANAGER, MASS SPECT. SERVICE
CAMBRIDGE ANALYTICAL ASSOC.
1106 COMMONWEALTH AVE.
BOSTON MA 02215
617-232-2207
RITA MARIE SCHMON-STASIK
SUPV. OF GAS CHROMATOGRAPHY
PRINCETON AQUA SCIENCE/IT
165 FIELDCREST AVE., CN7809
EDISON NJ 08818
201-225-2000
JUDITH W. SCOTT
CHEMISTRY DEPARTMENT
TRW
ONE SPACE PARK, 01/2161
REDONDO BEACH CA 90278
213-535-6886
JANICE L. SEARS
PROJECTS ADMINISTRATOR
CENTEC CORPORATION
11260 ROGER BACON DRIVE
RESTON VA 22090
703-471-6300 EXT.260
ANNETTE SIMON
CHEMIST
OCCIDENTAL CHEMICALS
LONG ROAD
GRAND ISLAND NY 14072
716-773-8655
MARGARET S. SLEEVI
CAL LAB EAST, INC.
2240 DABNEY ROAD
RICHMOND VA 23230
804-359-1900
TIMOTHY O. SLOAN
SR. ORGANIC CHEMIST
ROGERS & CALLCOTT ENG.
718 LOWNDES HILL RD.
GREENVILLE SC 29607
803-232-1556
JAMES S. SMITH, Ph.D
CHEMIST
WALTER B SATTERTHWAITE ASSOC
720 N. FIVE POINTS ROAD
WEST CHESTER PA 19380
215-692-5770
STEPHEN B. SMITH
VICE PRESIDENT, R&D
ENVIRITE ANALYTICAL SERVICES
OLD WATERBURY ROAD
THOMASTON CT 06787
203-283-8235
DAVID N. SPEIS
MANAGER, CHROMATOGRAPHY
ETC CORP.
P.O. BOX 7808
EDISON NJ 08818-7808
201-225-6759
DAVID E. SPLICHAL
CHEMIST
WILSON LABORATORIES
525 NORTH 8TH ST.
SALINA KS 67401
913-325-7186
GEORGE H. STANKO
STAFF RES. CHEMIST
SHELL DEVELOPMENT CO.
P.O. BOX 1380
HOUSTON TX 72251-1380
713-493-7702

-------
                                                               610
WILLIAM STEINHAUER
PRINCIPAL  RESEARCH  SCIENTIST
BATTELLE MEMORIAL LABORATORY
397 WASHINGTON  STREET
DUXBURY MA 02332
617-934-5682
 JAMES  STEPHENSON
 CAL  LAB  EAST,  INC.
 2240 DABNEY  ROAD
 RICHMOND VA  23230
 804-359-1900
RON C. STITES
GENERAL MANAGER
CENREF LABS
P.O. BOX 68
BRIGHTON CO 80601
303-659-0497
BRUCE O.  STRASSER
RESEARCH  SCIENTIST,  R&D
UNION CAMP CORP.,  R&D
P.O. BOX  3301
PRINCETON NJ 08543-3301
609-891-1200
DONALD SUMLIN
LABORTORY DIRECTORY
CITY OF ATLANTA
2640 JONESBORO RD., SE
ATLANTA GA 30316
404-627-1222
ROBERT SWAIN
MARKETING ENGINEER
FINNIGAN CORP.
355 RIVER OAKS PARKWAY
SAN JOSE CA 95134
408-433-4800
WILLIAM A. TELLIARD
CHIEF, ENERGY & MINING BRANC
USEPA-ITD
401 M STREET, SW,  (WH-552)
WASHINGTON DC 20460
202-382-7131
DR. P. MICHAEL TERLECKY
VICE PRESIDENT
FRONTIER TECHNICAL ASSOC.
8675 SHERIDAN DRIVE
BUFFALO NY 14221
716-634-2293
KURT THEURER
ALLIED CORP.
BOX 1021R-CRL
MORRISTOWN NJ 07960
201-455-2141
KATHLEEN E. THRUN
ARTHUR D. LITTLE, INC.
ACORN PARK
CAMBRIDGE MA 02140
617-864-5770 EXT.2311
LIAL F. TISCHLER
PARTNER
TISCHLER/KOCUREK
116 EAST MAIN ST.
ROUND ROCK TX 78664
512-244-9058
SAMUEL TO
USEPA-ITD
401 M STRETT, SW, (WH-552)
WASHINGTON DC 20460
201-475-8322

-------
                                                                611
 DAVID TOMPKINS
 CENTEC ANALYTICAL SERVICES
 2160 INDUSTRIAL DRIVE
 SALEM VA 24153
 703-387-3995
 ALLAN TORDINI
 ASST.  VICE  PRESIDENT
 UNITED STATES  TESTING  CO.
 1415  PARK AVE.
 HOBOKEN NJ  07030
 201-792-2400
 DONALD P. TREES
 SAMPLE CONTROL CTR. MANAGER
 VIAR AND COMPANY
 300 N. LEE ST., SUITE 200
 ALEXANDRIA VA 22314
 703-683-0885
 JEFFEREY TROIANO
 SR. ENVIRONMENTAL ENGINEER
 FORD MOTOR COMPANY
 48229 MALLARD
 DEARBORN MI 48047
 313-322-3890
 VICTOR TUROSKI
 JAMES RIVER CORPORATION
 1915 MARATHON AVENUE
 NEENAH WI 54956
 414-729-8168
 DR. JACK TUSCHALL
 SR. PROJECT SCIENTIST
 NORTHROP SERVICES INC.
 P.O. BOX 12313, 2 TRIANGLE DR.
 RESEARCH TRIANGLE PARK NJ 27709
 919-549-0611
 JOSEPH VIAR,  JR.
 PRESIDENT
 VIAR AND COMPANY
 300 N. LEE  ST., SUITE 200
 ALEXANDRIA  VA 22314
 703-683-0885
 FRANK VINCENT
 JAMES RIVER CORPORATION
 1915 MARATHON AVENUE
 NEENAH WI 54956
 414-729-8168
DAVID L. VINCI
SR. ENVR. CHEMIST
WESTCHESTER CTY. LAB DEPT.
HAMMOND HOUSE ROAD
VALHALLA NY 10595
914-347-3155
JOSEPH VITALIS
CHEMICAL ENGINEER
USEPA-ITD
401 M ST., SW  (WH-552)
WASHINGTON DC 20460
202-382-7172
DR. DALLAS WAIT
VICE PRESIDENT
ERCO/A DIVISION OF ENSECO
205 ALEWIFE BROOK PARKWAY
CAMBRIDGE MA 02138
617-661-3111
TONIE WALLACE
PRESIDENT
COUNTY COURT REPORTERS, INC
30 SOUTH CAMERON ST.
WINCHESTER VA 22601
703-667-0600

-------
                                                               612
BRUCE K. WALLIN, Ph.D
TECHNICAL DIRECTOR
E.G. JORDON CO.
562 CONGRESS ST., P.O. BOX 7050
PORTLAND ME 04112
207-775-5401
STAN WEST
CHIEF CHEMIST
SHERRY LABS
2203 S. MADISON ST.,P.O. BOX 2i
MONCIE IN 47302
317-747-9000
STUART A. WHITLOCK
ASSOCIATE VICE PRESIDENT
ENVR. SCIENCE & ENG., INC.
P.O. BOX ESE
GAINESVILLE FL 32609
904-376-0056
BRUCE E. WILKES
PRESIDENT
ENVR. ANALYTICAL CONSULTING
5176 CYRSTAL DRIVE
CROSS LANES WV 25313
304-776-2730
TOM WILSON
SENIOR CHEMIST
IT CORPORATION
5815 MIDDLESROOK PIKE
KNOXVILLE TN 37921
615-588-6401
HUGH WISE
ENVIRONMENTAL SCIENTIST
USEPA-ITD
401 M. ST., SW,  (WH-552)
WASHINGTON DC 20460
202-382-7177
N. LEE WOLFE
USEPA-ATHENS E.R.L.
COLLEGE STATION ROAD
ATHENS GA 30613
404-546-3185
MARK W. WOOD
SR. LABORATORY SUPERVISOR
PPG INDUSTRIES, INC.
P. 0. BOX 1000
LAKE CHARLES LA 70602
318-491-4450
ROBERT WOODS
GC/MS SUPERVISOR
ANALYTICAL TECH., INC.
225 WEST  30TH ST.
NATIONAL  CITY CA  92050
619-477-4173
LAUREN M. YELLE
MASS SPECTRONMETRIST
ARTHUR D. LITTLE, INC.
15 ACORN PARK
CAMBRIDGE MA 02140
617-864-5770 EXT.2586

-------