-------

-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
               OFFICE OF WATER
        INDUSTRIAL TECHNOLOGY DIVISION
     THIRTEENTH ANNUAL.EPA CONFERENCE ON
  ANALYSIS OF POLLUTANTS IN THE ENVIRONMENT
               MAY 9 & 10, 1990

-------

-------

             UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                            WASHINGTON. D.C. 20460
MEMORANDUM

TO:       Attendees
OFFICE OF
 WATER
FROM:     William A. Telliard, Chief
          Analytical Methods Staff
          Engineering and Analysis Division
SUBJECT:  The Thirteenth Annual Analytical Symposium
          Proceedings
    Enclosed please find a copy of the Proceedings of the
Thirteenth Annual Analytical Symposium sponsored by the USEPA
Office of Water, Industrial Technology Division (now known as
the Engineering and Analysis Division), held in Norfolk,
Virginia, May 9-10, 1990.  We believe you will find these
Proceedings to be an invaluable reference for the analytical
methods discussed during the symposium.

    Thank you for your continued interest and, support in the
Division's Analytical Symposium.

-------


-------
                        FOREWORD
     The Industrial Technology Division of the USEPA
Office of Water Regulations and Standards sponsors an
annual conference on the analysis of pollutants in the
environment.  This symposium is attended by laboratory
and sampling personnel, data users, and regulatory
agencies.  It is intended to provide a forum for the
discussion of current and proposed analytical methods
for the analysis of priority pollutants.

     These proceedings document the presentations and
discussions from the Thirteenth Annual Analytical
Symposium.  Topics this year encompassed discussions of
method validation, determination of volatile and semi-
volatile compounds, and the effects of new regulations
on laboratories.  A special panel discussion was held
by Federal, State, and industry representatives on
laboratory certification and reciprocity.

     This year's conference was very successful, with
over 270 people in attendance.  We look forward to the
Fourteenth Symposium in May 1991.

                                         W. A. Telliard

-------

-------
             THIRTEENTH ANNUAL EPA CONFERENCE ON
          ANALYSIS OF POLLUTANTS  IN THE ENVIRONMENT
          Office of Water Regulations  and Standards
               Industrial Technology Division

                       May 9-10, 1990
                      Norfolk,  Virginia

                      TABLE  OF  CONTENTS
Wednesday, May 9, 1990

Welcome and Introduction	
     William A. Telliard, Chief
     Analysis and Analytical Support Branch
     USEPA, Industrial Technology Division

Revision and Updates to EPA's List of Lists.
     James King
     USEPA Sample Control Center
     Viar & Company

A Toxicity-based Approach to Pollutant
Identification	
     D. R. Mount
     ENSR Corporation
 Page

,. .1
 ,42
Development of a HRGC/HRGC/LRMS System for
Determination of Chloro-dioxin/furan Isomers/
Congeners; Instrument Design and Performance	
     L. L. Lamparski
     Dow Chemical Company

Development of a MAGIC Interface for HPLC/FT-IR..
     James A. de Haseth
     Department of Chemistry
     University of Georgia

New Capillary Column for the Determination
of PCDDs and PCDFs	
     T. 0. Tiernan                           ^
     Wright State University

Single Laboratory Evaluation of EPA Method 200.8,
Determination of Trace Elements by Inductively
Coupled Plasma - Mass Spectrometry	
     Theodore D. Martin
     Chemistry Research Division
     USEPA-EMSL
 ,78
 119
,160
,209

-------
             THIRTEENTH ANNUAL EPA CONFERENCE ON
          ANALYSIS OF POLLUTANTS  IN THE  ENVIRONMENT
                    TABLE OF CONTENTS  -  2
ICP Performance for the Measurement of 14 Trace
Metals in Power Plant Waste Streams	,
     James K. Rice, P.E.
     James K. Rice Consulting
 Page


..235
Quantitation/Detection Limits for the Analysis
of Environmental Samples	
     W. G. Krochta
     PPG Industries, Inc.

Laboratory Determination of Diesel Oil
in Drilling Fluids	
     Warren Haltmar
     EPTD-Environmental Technology
     Texaco, Inc.

Validation of a Method for the Determination
of Diesel Oil in Drilling Fluids	
     M. T. Stephenson
     EPTD-Environmental Technology
     Texaco, Inc.

Preparation and Analysis of Air Emission
Samples	
     Larry D. Johnson
     Source Methods Standardization Branch
        Quality Assurance Division
     Atmospheric Research and Exposure Assessment
        Laboratory
     USEPA-RTP

Chesapeake Bay Program - Experience with Nutrient
Analytical Methods in the Estuarine
Environment	,
     Bettina Fletcher
     USEPA Region III CRL
  270
  301
  326
  352
  392
Thursday, May 10, 1990

Determination of Semi-volatile Organic Compounds
in River Water at the Part-per-quadrillion (ppq)
Level by High Resolution Gas Chromatography/High
Resolution Mass Spectrometry	,
     Yves Tondeur
     Triangle Laboratories
 ,438

-------
             THIRTEENTH ANNUAL EPA CONFERENCE ON
          ANALYSIS OF POLLUTANTS IN THE ENVIRONMENT

                    TABLE OF CONTENTS  - 3
EPA's RREL Sampling and Analysis Methods Database	
     Lawrence Keith
     Radian Corporation

Micro-extraction Isotope Dilution GC/MS Determination
of Volatile Organic Compounds	
     Bruce N. Colby
     Pacific Analytical

Liquid-solid Extraction for Determination of
Acid Herbicides in Drinking Water	
     Jim Eichelberger
     USEPA-EMSL-Cincinnati

Application of Multispectral Techniques to the
Identification of Aldehydes in a Combined
Sewer Overflow		
     Susan D. Richardson
     Environmental Research Laboratory
     USEPA-Athens

A Laboratory Robotic Method for the Automated
Determination of Total Suspended Solids in
Environmental Water Samples	
     Joe C. Raia
     Shell Development Company

Determination of Semi-volatile and Pesticide
Pollutants in Sewage Sludge by Soxhlet-Dean Stark
Extraction, High Performance Liquid Chromatography
Cleanup, Gas Chromatography with Selective
Detectors, and Isotope Dilution Gas Chromatography/
Mass Spectrometry	
     D. R. Rushneck
     ATI-Colorado
 Page

..465
 ,502
 ,546
 ,585
 ,616
  648
Federal Government Perspective on the Regulation
of Laboratories under the CLIA of 1988	
     Rhonda Whalen
     Office of Survey and Certification
     U.S. Department of Health and Human Services

Management of a Clinical Laboratory Certification.
     Dr. Gerald Hoeltge
     Cleveland Clinic Foundation
  673
  693

-------
             THIRTEENTH ANNUAL EPA CONFERENCE ON
         ANALYSIS  OF  POLLUTANTS  IN THE ENVIRONMENT
                    TABLE OF CONTENTS - 4
                                                       Page
Panel Discussion:  Laboratory Certification
and Reciprocity	717
     A. W. Tiedemann, Jr., Ph.D.
     Commonwealth of Virginia	718
     Mike Carter
     USEPA, OERR	723
     Arthur Perler
     USEPA, Office of Drinking Water	728
     Benjamin R. Tamp1in
     California Department of Health Services	.....757
     Margaret Prevost
     State of New York	763
     George Stanko
     Shell Development Company	767
     Question and Answers	774
Closing Remarks	784
List of Speakers	785
List of Attendees	789

-------
                   PROCEEDINGS



                              MR. TELLIARD:  Good morning.



My name is Bill Telliard.  I am from EPA, and I am here to



to help you.  Welcome to the 13th, our lucky 13th, annual



symposium on measurement of pollutants in the environment.



     A couple of notes on housekeeping things before we get



started.  We have a substitution at 9:00 o'clock.  Dr.



Fielding will not be presenting.  His paper has been dropped



due to a lack of interest on Dr. Fielding's part.  And for



tomorrow's talk on the coastal marine environment by Gordon



Wallace, Larry Keith will speaking in his place.



     This morning and throughout the next couple of days, we



are going to talk about a myriad of different problems in



environmental measurement.  Our people to our left are



County Court Reporters, Inc., who take down your every



single, dripping word that you might want to put forward.



If for some reason you don't want your words on the



record...and we all know what that means; it is like telling



your wife where you are...we will be happy to tell them to



stop recording and stop typing which they do real well.



Stopping is one of the best things they do.  And we will be



glad to take it off the record.



     Our first speaker this morning is Jim King from Viar &



Company...which reminds me of a story.  Has nothing to do



with Viar & Company, but I had to lead in with something.

-------
     This lawyer has a parrot on his shoulder, walking down
the street, walks into a bar, walks up to the bar, and the
parrot says give this guy a beer.  Bartender says wow, this
is really kind of a neat bird, you know.  So, he gives the
guy a beer, goes down the bar, and he turns to this other
guy and says that parrot is really kind of neat.  So, he
comes back up the bar and he turns to the lawyer and he
says, where did you get him?  And the parrot says
Washington, there's thousands of them there.
     Our first speaker is Jim King, and Jim is going to talk
about one of my favorite subjects which is the list of
lists.
     Jim is going to talk about a program that ITD has come
up with dealing with the use of a listing series to cover a
catalog of analytical methods and also those listed
compounds that appear on various agency lists throughout the
agency.
     Jim King from Viar and Company.

-------
                              MR. KING:  Thank you, Bill.



     Hi.  I am Jim King from Viar and Company.  We operate



the Sample Control Center for Bill Telliard and EPA's



Industrial Technology Division (ITD).



     In 1985, ITD project officers were faced with the



prospect of regulating a wide variety of compounds, not only



Priority Pollutants,but also this very strange list that



appeared in the 1984 Hazardous and Solid Waste Amendments



called the RCRA Appendix VIII list.  There were a number of



compounds on this list that were unanalyzeable by present



methods.  Bill Telliard decided to develop an automated



catalog of these analytes and ideas and thoughts about the



methods that might work for them.



     Over the years, this system has grown into a compendium



of all of the analytes of interest at the Agency contained



in regulatory, as well as, office-based lists.  When I say



regulated, I am referring to environmentally significant



regulations rather than product regulations such as the



FIFR/TSCA rules and others like that.



     Now, I'll demonstrate the current system.  I will talk



everyone through the system as we view it on the screen.  If



anyone has any questions, feel free to ask them at the end.



     This is the main menu for the system.  It does work



with a Microsoft compatible mouse.   When the device driver

-------
for the mouse if installed, you will see a highlighted bar
menu at the bottom of the screen.
     The system runs on an IBM PC or compatible.  It
requires about 256K of memory.  A hard disk is required.
The database is indexed by these seven primary keys and a
number of other alternate keys that once you become familiar
with the system, you can utilize.
     The primary key and the one on which we place the most
significance is the CAS number.  When entering data into the
system, we go through and thoroughly research the CAS number
either using Chemical Abstracts on line or two hard copy
references in order to verify it, because we have found a
number of regulations and other listings of compounds with
CAS numbers that just were not correct.
     The easiest way to view the data is by name, so we will
take a look at it by name.  Place the cursor in the lookup
field and search for a compound.   Here we have
acenaphthene, CAS Number 83329.  Here we have the CAS
number, the regulatory origin and sequence, reportable
quantities if there are any associated with the regulation,
the analyte name as listed in the regulation, the
International Union of Pure and Applied Chemists (IUPAC)
name, and then any synonyms would appear after the IUPAC
name.

-------
     The first screen is the original screen developed in



1985, so we had relatively few locations for apparatus,



method, and a use code, but at that time, we were very



interested in sources for standards.  So, here we have a



little catalog of sources of standards for true analytes.



     We also have some physical properties data such as



whether the compound hydrolyzes or decomposes in water,



extracts or purges from water, whether it is possible to do



by gas chromatography or whether you have to resort to



liquid chromatography.



     There is also a location for the EPA-NIH mass spectral



library page number.



     This is a relational data base, and if we place the



cursor in one of these fields, we can scroll through these



regions, and handle a relatively large number or regulations



per compound.



     If we .flip over to page 2, we see the methods screen.



This was developed in early 1988, and the objective here was



to catalog as many methods as were available for a



particular compound and some information about the method.



We have the origin of the method here, the apparatus



utilized, the method number if there is one, the suffix



which is really a use code (the Contract Laboratory Program



uses that for low soils, medium soils, water, et cetera).

-------
                              6



     We also have the detection limit appearing here,



expressed in the identical terms that it was expressed in



the method.  In other words, we did not convert everything



into some common detection limit or common unit.



     Some fields that we will be updating in 1990 include



precision expressed as standard deviation or coefficient of



variation and other methods information.  We are going to



take the same approach that we took with the detection



limits.  However precision is expressed in the method, that



is how it will appear here for that particular compound.



And also bias expressed as percent recovery.



     We are also talking about adding an additional screen,



and we have a prototype of that.  If you place the cursor on



a method number and then hit PF key 5, it will take you into



the Method Information Screen.  This screen is in prototype



form right now.  We are still experimenting with what sort



of data we want to display on this screen.



     We often receive comments from people who say "well,



why don't you just put the whole method in there". We are



trying to keep the focus of the system very clear so that it



doesn't become unwieldy and you need a monster 80386 PC to



run it on.  We have run this on an XT for a number of years



with no problem.  It works quite well.



     On the Methods Information Screen, we look up by method



number, the primary key.  We have the revision, the date of

-------
the method, the status, whether it is draft, final,

proposed, or promulgated, the number of labs that

participated in the validation study, the type of analysis,

the matrix, the apparatus used, the appropriate

concentration range, a citation for the method, and a short

abstract of the method.
                           i

     If we go back to the previous screen, one thing that is

real interesting, is that we can actually switch indices

here with the mouse just by clicking the right button on the

actual fields.  So, here we are in the name index, and we

are viewing it by name.  With a right click we will go over

to the method index, organization index, etc.

     Let's go back to the name index here.  The name index
       »
contains a little over 10,000 entries.  What appears in this

window is the compound name as it appears in the reg, the

UP AC name, and any synonyms.   So, there are actually just

under 1800 unambiguous analytes in the system,  150 methods

for their analysis, compiled from 26 regulatory lists.

     We can move from A to Z down through an index of 10,000

in less than one second.  Here we see zectran,  our first

entry in the Z's, but let's look up...and basically what we

have here is a byte by byte search of the data  base.  So, if

we enter an "L", it will bring up the actual first

occurrence of an "L" compound.

-------
                              8
     This particular index contains analytes names listed
multiple times, as they appear in the regulation, by IUPAC
name, and by synonym.  We can look up lasso, and we can also
look it up as alachlor or metachlor or by the IUPAC name
which I won't even try to pronounce.
     The 1990 revision of the system is scheduled to be
complete by the end of the current fiscal year, and we
expect at that time to publish a new List of Lists document,
and it is probably going to be a little unwieldy at 200
pages to reduce this bulk, we are looking at is distributing
the system, the system software and the data on diskette for
a break-even cost, outside of the agency.
     The diskette version is slated to start beta tests
within the next 60 days, and we would hope to send it out to
all current requesters.  There were a number of people at
the Water Pollution Control Federation conference last year
that requested this system, and at that time, we were hoping
to have it available sooner, but we have maintained their
names on file, and they will be getting a copy of the system
software in this first round of distribution.
     If anyone has any questions at this time or has any
favorite compounds that they would like to look up, feel
free to ask, and I will do my best to answer them.
                              MR. TELLIARD:  I should point
out that these people who wanted a copy of this wonderful

-------
system, we don't take green stamps, but we are going to use



something that we feel the price will be probably about $70



on the first issuance.



     We were talking about handling it through NTIS.  I



think, for the moment, we will handle it through Sample



Control Center.  I think it is faster, and I think the



product is better if we can get it out to the folks.



     We are putting it in all the  EPA Regional libraries,



permits office, and BSD laboratories, and Region III is the



beta test area, so we will put it in all the State offices



there to see how the thing works and what, if any, problems



we have with it.



     Since we have a working product, of course, the new



EMMC, Environmental Methods and Management Committee, has



put a work group together to figure a way that this can cost



more.  So, we are working on that now.  It is your



government in action:  if it is working fine, let's fix it.



     So, one of the areas that is totally lacking, as



pointed out by Jim, is the fact that we don't have a lot of



data on the precision and accuracy of the method, and we are



just going to extract that from the existing method write-



ups.  Some of that is good, some of that is bad, some of



that is indifferent, but we are going to put it in there at



least as a starting point and see what type of feedback we

-------
                             10
receive from the other users in the other laboratories that
may have an interest in this particular program.
     So, are there any questions?  When you come to the
microphone, please identify yourself and tell us who you
are.

-------
                                  11



                    QUESTION AND ANSWER SESSION







                                   MR. RICE:  Jim Rice. I am



 a consulting engineer.



     Bill, I have several questions.  First, what is the



data base management system you are using?



                                   MR. KING:  The software



is written in system J.  It is as relatively unknown system.



However, in 1985...



                                   MR. TELLIARD:  Obscure is



the word.



                                   MR. KING:  Yes, obscure.



                                   MR. RICE:  When you make



it available on diskette, what will come with it is the



license to the data base management system as well and the



software?



                                   MR. KING:  No, it will



not.  It will be run-time only module.  So, in other words,



we will send you only a couple files that are required to



run the system.



                                   MR. RICE:  Right.  Well,



that is what I meant.



     And now the other question, Bill, and a comment.  You



mentioned about putting precision...I think it says



precision and bias on those columns up there.

-------
                             12



                              MR. KING:  That is correct.



                              MR. RICE:  But I would make a



plea to you not to use and put up and, in effect, put into a



data base for wide distribution data that you know is not



worthy to reprint.  I think I would only put that which



is...that you consider really valid interlaboratory data on



a method that has been checked out by a large suite of



laboratories with it rather than put up junk, because other



people than this group which are technically astute and can



understand that get hold of that data, and then it causes



all kinds of mischief.



                              MR. TELLIARD:  Thank you, Jim.



I agree with you, but, you know, there is a box to be filled



in.



(Laughter.)



                              MR. TELLIARD:  It is kind of



like your census.  Irrespective of how many live in the



house, you have to fill in the box.



     We are aware of the problem that some of the data is



better than others, and I think for the first go-around, as



we mentioned, for the beta test, we are going to try to put



everything in and circulate it and have people pull it out



or revamp it based on their experience or their information.

-------
                             13



We haven't figured a way to flag what we call real good, not



so good, and mediocre, but I agree with you.  I think we



ought to do that, but we haven't addressed that issue yet.



                                   MR. FRAZIER:  My name is



Bill Frazier, City of High Point, Central Lab Services.



     In the first screen, you had information about chemical



and physical properties.  Is that also going to have



toxicology data available with it?



                                   MR. KING:  We have talked



about it, but the problem is there is such a limited data



base of toxicological information.



                                   MR. FRAZIER:  Is that a



future plan for it?



                                   MR. KING:  That is



currently in the plans.



     The agency has a data base called IRIS.  I believe



there are only about 200 entries in the data base FOR



mammalian toxicity..  So, that would be a very minor subset



unless we wanted to extend it to different salmonella type



toxicity tests or other toxicities that may or may not have



relevance on mammalian toxicity.



                                   MR. FRAZIER:  Thank you.



                                   MR. TELLIARD:  TSCA has a



similar list that they have generated for the Office of



Toxic Substances.  We are not trying to combine them, but

-------
                             14



where their data interfaces with ours, what we are thinking



of doing is having a reference arrow that says see block



something or other.



     This system will allow us to add additional screens



where we could add a toxicity screen or one of the other



things that we are interested in is a production screen.



How much of 1,2-diphenol bat shit is made, exported, used



internally, controlled and kind of from a permitting



standpoint when the guy is writing the permit know how much



he should be concerned about it.



     That is an option that is probably a couple of years



away if we wish to implement it, but right now, we are



trying to get it out that says it is on somebody's list,



somebody is concerned about it, and we are monitoring for



it.



     Now, when we say a list, it is not necessarily a



regulatory list.  It is a list whether you are in the Office



of Solid Waste or you are CERCLA or you are inter-toxic



where someone is going to have to make a measurement, either



in a permit, in a monitoring mode, or somewhere.  So, when



we say list, it is not the narrow list of just those



regulated compounds which, of course, is probably like 85



and we have, what, 1700 or something on here.



                              MR. KING:  Right.

-------
                             15



                                   MR. TELLIARD:  So, when



we say list, our analogy is those compounds which you as a



user or a producer or a manufacturer may have to monitor



for.



                                   MR. KING:  In fact, if we



look at the on-line help screen, you can see where are



origins are derived, and these are the regulatory origins



here.  We have various lists on there from the Appendix C



list, from the consent decree, California list of



pollutants, CERCLA 302, Clean Water Act 116, the Michigan



Petition list, the administrator's VTOX list which



ultimately became the SARA Section 313 list.



     Also, there are a couple of office-based lists on



there.  There is a list of fish tissue contaminants that are



monitored by the Office of Water Regulations and Standards



that we do have on here.



                                   MR. TELLIARD:  Good



morning, George.



                                   MR. STANKO:  George



Stanko, Shell Development Company.



     I think I may have a solution to Jim Rice's problem and



also would ask a question.



     Could we replace that precision field with something on



MSDS for these compounds, or was anything considered with



respect to MSDS?

-------
                             16



                              MR. KING:  No.



                              MR. STANKO:  Because a lot of



us have to handle these, ship these, and do that, and that



kind of information with this kind of a data base looks like



it would be very useful.



                              MR. TELLIARD:  Well, we



certainly have the capability of adding that, George.  Thank



you.



     Any other questions?



(No response.)



                              MR. TELLIARD:  Again, if you



will pick up this little blue card, we have had the format



and the obscure...obscure was the term?...obscure language



has been approved by the agency for circulation, and we are



going to try to get this thing out by mid summer to the



world.



     There is also an issue in fact that we may rewrite the



program into a different format, that is to say, a different



language based on what our computer folks tell us in RTP,



but right now, we are shooting to have the program available



by mid summer.  So, if you would like a copy. . .as I say, we



are estimating $70 for the cost.



                              MR. TELLIARD:  If you fill out



the little blue card, when we figure out what the real price

-------
                             17



is, we will send you a notice, and if you still want one, if



you send the check to my sister...



(Laughter.)



                                   MR. TELLIARD:  We will be



glad to send you a copy.



     We would also like as you get this copy any feedback



that you can send to us on ways of making it better or



changing it or reformatting it.  Since it is our first



effort at it, we would be glad to have some feedback from



you folks, particular the user community like you.  Also, as



was pointed out, you are probably more on top of this than a



lot of the other people who are going to be using it.  So,



any information you can send back to us to make it better



would certainly be appreciated.



     Thank you, Jim.



                                   MR. KING:  Thank you.

-------
THE ITD LIST OF LISTS SYSTEM
    "A Catalog ofAnalytes and Methods"
                                                CO
           Analytical Methods Staff
  USEPA Office of Water Regulations and Standards
         Industrial Technology Division
                                  OWRS

-------
       WHAT IS THE LIST OF LISTS?

The Office of Water Regulations and Standards Industrial
Technology Division has developed, maintained, and
distributed, since October of 1985, an automated
composite index of analytes listed by the Agency.  This
master database is known as the "List of Lists" (LISTS).

                   ANALYTES
                                       OWRS

-------
WHY WAS THE LIST OF LISTS CREATED?
PROLIFERATION OF LISTS AND ANALYTES
    By 1985, there were 15 Agency lists and over 1,000
    analytes. Analyte lists could no longer be managed
    individually.
TO HELP ITD DESIGN MONITORING PROGRAMS

    Tasked with promulgating industrial effluent limits for the
    Priority Pollutants,  RCRA Appendix VIII list of 385
    compounds and compound classes, as well as CERCLA's
    Hazardous Substances List, ITD created an automated
    composite list of pollutants of concern to EPA as a resource
    to support its monitoring programs.
to
o
                                          OWRS

-------
 WHY THE LIST OF LISTS  IS IMPORTANT
                      TODAY

METHODS STANDARDIZATION
    With the proliferation of field and laboratory measurement
    methods, it is necessary to establish degree of
    standardization among methods.

STUDY DESIGN
    EPA monitoring program designers need analyte and
    methods information to make correct choices:
       • Addressing cross-media contamination
       • Intersection of Regulatory lists
       • Selection of appropriate methods

SECTION 518 REPORT TO CONGRESS
    The 518 Report recommends "establishment of a
    computerized catalogue of the availability, applicability, and
    degree of standardization of methods currently in use in the
    Agency."

-------
           SYSTEM  DESCRIPTION

LISTS is a PC-based system developed using System J.

The List of Lists is distributed in both hardcopy and computer-
readable formats. Over 10,000 hardcopies and ASCII files
have been distributed since 1985.

Run-time (read-only) module created for system dstribution to
user community.

LISTS database contains information on:
   -1,716 regulated analytes
   - 26 statutorily-mandated and Office-based lists
   -150 analytical methods
                                           OWRS J

-------
              THE LISTS SYSTEM
 FEATURES

Menu Driven

Rapid Text
Search Using
Lookup Field
Indexed on Key
Fields

Simultaneous
Display of Key
Data Elements

Optional use of
pointing device
("mouse")
 CAPABILITIES

Search, Retrieve,
Display, and Print
Data Sorted by Key
Field

Database Add,
Delete,'and Modify

Output ASCII Text
Files in PC or
Mainframe Formats

Generate Standard
Reports by Nine
Key Fields
REQUIREMENTS

• IBM PC XT, AT, or
 Close Clone

• DOS Operating
 System

• System J Software

• Hard Disk and
 256K RAM
 (640K RAM
 Recommended)
                                           OWRS

-------
      SYSTEM CHANGE CONTROL

OWRS MAINTAINS CENTRALIZED CHANGE CONTROL OVER
 LISTS SYSTEM CONFIGURATION AND DATABASE FILES.
                  OWRS
                   ITD
                  Full System
                  Add/Delete/Modify
                   Capabilities
              USER COMMUNITY
              Run-Time Only Module
                Read-Only Capability
                                   OWRS

-------
 WHAT DATA ARE IN THE LISTS SYSTEM?
 ANALYTE DATA
Analyte Name (as
appears on Agency list)
i?^1
&:<
Common, trade,
synonym, and IUPAC
Names
CAS Number and Base
CAS Number
        $
Regulatory Origin/
Agency List
Analytical Method(s)

EPA/NIH Mass Spectral
File Reference

Physical Properties
Source for Standard
                      METHOD DATA
Method Number/
Identification
Custodial
Organization
Instrumentation /
Technique
(Apparatus)
Sample Matrix,
Fraction, and Level
(Suffix)
Detection Limit by
Analyte
Precision and Bias
by Analyte
                    LIST DATA
Regulatory
Origin or Name
of Agency List
Custodial
Organization
Reportable
Quantity
Analyte Names
(as appear on
Agency list)
to
Ul
                                            OWRS

-------
                 DATA INTEGRITY
  Integrity of LISTS system data is ensured by use of an
  unambiguous identifier (CAS number) as a primary key and
  stringent verification of each CAS number before entry into
  the system,                ;        ,   :   :
C. .*.. .v./v ... .".,'
  Each CAS number is checked to verify that it is an
  unambiguous identifier for a specific analyte.
    -  Check-sum algorithm applied to ensure that CAS
       number is valid.

    -  CAS number checked against two published sources or
       one published source and one on-line source to ensure
       that it accurately identifies the listed analyte. If
       necessary, the Agency office responsible for the list is
       contacted for confirmation.
                                            OWRS
to
en

-------
                   KEY FIELDS
LISTS data can be retrieved, displayed, and printed by
any of nine key fields:
 ~*«L.  ,      PRIMARY
 CAS Number    KEY
Analyte Name
Base CAS Number
Apparatus and Method
Method and Apparatus
Regulatory Origin/Agency List
and Analyte Sequence in List
                             Custodial Organization
                            Standard Source and Analyte
                            Name
Method, Suffix, Apparatus,
and Analyte Name
                                                            10
                                                            sj
                                           OWRS

-------
WHAT ANALYTICAL TOOLS ARE AVAILABLE TO
                  LISTS USERS?

 •  Data search on key fields
 •  Data retrieval and display sorted by selected index (key
   field)
 •  Look up data for specific variable while in selected index
 •  Print data sorted by selected index
      -if  !'#
 •  Output data from selected index for loading to other
   computers (create print file)
to
00
                                        OWRS

-------
      WHAT SYSTEM REPORTS CAN BE
                  GENERATED?

A user can print LISTS reports presenting data ordered by any of
the nine key fields.  Report data is presented in alphanumeric
order by data element(s) in the key field.
•*« CAS Number
> Analyte Name

!;/Pase
 * Apparatus anf) Method
 t Method andI Apparatus
alyte
''

                                                         to
                                                         ID
                                            n
                           ,
                          eustodlal Organisation
                                 Source and
                       -;t> Methodi^Suffbci
                        JandAnalyte;

-------
            SYSTEM ENHANCEMENTS

 Addition of Analyte-Specific Precision and Bias
 Information
 Data fields for precision and bias information have been created and
 entry of these data is ongoing.

> Addition of Methods Abstract Screen
 A methods abstract screen has been created to provide information on
 method standardization and validation. Entry of method abstract data is
 in process.

 Development of User Documentation
 A Draft User Manual has been written, which provides detailed
 information on how to install the LISTS system, retrieve and display
 data, and print standard reports. The Draft User Manual will be utilized
 in the upcoming beta test and user input will be incorporated in the final
 manual.

                                          = OWRS

-------
    EXAMPLE SCREENS AND
       ON-LINE HELP FILE

1.   Main Menu
2.   Data Screen Indexed by Name - Pages 1 and 2
3.   Methods Abstract Screen
4.   Excerpt from On-line Help File
               U)
OWRS

-------
                                       32
                                 MAIN MENU
OW LIST OF LISTS MENU
                        The OW LISTS database contains environmentally
                        significant analytes and information necessary
                        to develop methods for their regulation.
      -DISPLAY ANALYTE DATA=
 F2
 F3
 F4
 F5
 FG
 F7
 F8
= By CAS number
= By name (alphabetically)
= By base CAS number
= By apparatus and method
= By method and apparatus
s By origin and sequence
= By organization
 To PRINT ANALYTE DATA, use Alt
 vith any of these function keys.
                           ORGANIZATION TO DISPLAY/PRINT
                           PRINT FILE
                                3THER LIST OF LISTS FUNCTIONS'

                              F16 = Help for this menu
                          Alt-Fie = Rebuild database  indices
                                                =SYSTEM J ACCESS=
                                            Fl  = Exit menu  to  J()  prompt
                                                 Type I_HELP to  get  back
                                            F9  = System  J command  menu
Fl
   F2
F3
F4
F6
F7
F8
F9
HELP

-------
                                     33
                   DATA SCREEM INDEXED BY NAME - Page 1
 ;. :::> Lookup  field
 AROCLOR_1242
        KEYS
 ANTIMONY_TRICHLORIDE:
 ANTIMONY_TRIFLUORIDE;
 ANTIMONY_TRIOXIDE
 ANTIMYCIN_A
 ANTU
 AQUACIDE
:AQUA_FORTIS
 ARAMITE
 ARASAN
.ARGENTATE(1-),_DICYA
 AROCLORS
 AROCLOR_1816
 AROCLOR_1221         '<•
 AROCLOR_1232         :•>
 AROCLOR_1242         £
 AROCLOR_1248         '
 AROCLOR_1254
 AROCLOR_1268
 ARSENATES
=OWRS  ITD AASB
 CAS No
   Base
53469219
 1336363
 Created  89/26/87
 Updated  82/14/88
Names  and comments
PCB-1242
Aroclor  1242
OW LIST OF LISTS =
 ORIGIN  SEQUENCE
 CAL
 CER^
 CWA,_
 P-POLL iiae
 RCRA   i386-84
=86/13/98  87:55:83=
   At   8 of   8  Page
               one.
JQ=18  Ib        F6
SQ=18  Ib        gets
               page
               two.
      APPARATUS  METM SUFFIX
Organ iCGCEC  :&        :|i
zationGCEC   5:'        \
      GCMS   '?•        >;,
                     STD
                     CIN
                     LV
                Hyd/Dec
                Ext/Prg E
                GC poss Y
                LC poss
                    EPA/NIH page
 Fl=Exit,Up,Set    2-Beg,End,Reread    3=Up,Page,Next
                                                CL NL  SL
                                4=Down,Page,Next   18=More

-------
                                      34
                  DATA SCREEN INDEXED BY NAME - Page 2
   Lookup field
AROCLOR 1242
        KEYS
ANTIMONY TRICHLORIDE
ANTIMONY~TRIFLUORIDE
ANTIMONY_TRIOXIDE
ANTIMYCIN A
ANTU
AQUACIDE
AQUA FORTIS
ARAMITE
ARASAN
ARGENTATE(1-),_DICYA
AROCLORS
AROCLOR 1016
AROCLOR 1221
AROCLOR"1232
AROCLOR 1242
AROCLOR 1248
AROCLOR_1254
\ROCLOR 1260
ARSENATES
=OWRS ITD AASB — (
CAS No 53469219
Names and comments
PCB-1242
Aroclor 1242


ORG APPARATUS METH
ASTMftGCEC :.SD3534
CIN IGCEC ?=608
CIN %-GCMS ';625
CLP l-GCEC WEST
CLP iGCEC -:;PEST
CLP -GCEC ilPEST
ITD :--CGCEC &1618
ODW ;iGCEC -;505
OSW l;::GCEC ;:;8080
OSW J--GCMS :::-8250
USGS GCEC O-31G4
5W LIST OF LISTS —06/13/96 07:52:28=
Page two. F6 gets page one.



At n

SUFFIX DETECT LIMIT PREC BIAS of 11
fEDL=l ug/L
i;MDL = 6.065 ug/L
BN ii
LS ^CRQL=86 ug/kg
MS :?CRQL=1200 ug/kg
•W ?;CRQL = 6.5 ug/L

j|MDL = 0.31 ug/L
iPQL=56 ug/L
•:;PQL=106 xig/L
•EDL = 6t.«l ug/L 'I •
 5 =Method,Svap/Dup    6=Page,Ed   FlG=More   Shft-Fn = Fn  Help
                                                                   CL NL SL

-------
                                     35
                         METHODS ABSTRACT SCREEN
::;.;.-: Lookup field ^ ::C:
1618
        KEYS
1613
1618
1624
1625
=OWRS  ITD  AASB =
    Method 1618
                METHOD INFORMATION =06/13/90 08:12:44=}
                     L
                                     Created
                                     Updated
Revision    0        dated 05/01/89
  Status DRAFT       Draft, Final, Proposed,
  # labs    1 in validation      Promulgated
05/09/90
05/09/90
                                 PESTICIDES
                                 GC
                                 MULTI   Water, Soil, Sludge,
                                 LOW     Low, Med, High
  An'alytes
 Apparatus
    Matrix MULTI    Water,  Soil,  Sludge,  Biota,  Air
     Level
 Ci tat ion
 "Chlorinated  and  Phosphorous  Containing Pesticides  by
 GC  with Selective Detectors",  USEPA,  OWRS-ITD  WH552,
 Washington, DC 2646Q,  May 1989.
 Abstract
 The method  is designed to meet  the  survey  requirements
 of  the USEPA  ITD.  The method is  used to determine  the
 chlorinated- pesticdes,  PCB's,  phenoxyacid  herbicides,
 and phosphorous  containing pesticides amenable to GC.
 Fll/12 = Do  ...; depends on current field and Shft/Ctr1/Alt.
                                               CL  NL  SL
                                          l«=Reset

-------
                                       36
                      EXCERPT FROM ON-LINE HELP FILE

                     ITD LIST OF LISTS DATABASE
                Legend for Information in Data Fields.
I_RCRA.DSC
  02/27/89
        The Industrial Technology Division's List of Lists system is an
        automated catalog of analytes of environmental concern and methods
        for their analysis.

CAS NO  The Chemical Abstracts Service (CAS) Registry Number for the analyte.
        In certain instances, CAS has assigned a number to a compound class and
        this number is used.
        Note:  where CAS has not assigned a number to an analyte or class, a
        synthetic numbering system has been used.  This number begins with a
        digit(s) followed by an underscore or hyphen followed by three digits
        (e.g., 1_001) and assures that an analyte can be unambiguously identi-
        fied in relationship to the class from which it is derived.  The three
        digits following the underscore identify its position on the parent list
        and match the ORIGIN SEQUENCE number.  At present, the following leading
        digits are used (definitions of acronymns and abbreviations are listed
        under ORIGIN below):
                0- identifies the Drinking Water Priority List
                0_ identifies the RQ List
                1- identifies analytes on ITD's List
                1_ identifies the RCRA Appendix VIII List
                2- identifies the AIR List
                2_ identifies the RPAR List
                3- identifies the SWDA List
                3_ identifies the VTOX List
                4- identifies the SEC_313 List
                4_ identifies the OAGJSRB List
                5- identifies the SEC_112 List
               10- identifies the FTC List

        CAS Number Error Checking on the PC:  When a CAS number is entered in
        the CAS NO or BASE CAS NO field, the CAS error checking algorithm is run
        at the instant the cursor leaves the field.  If the CAS number is
        incorrect, the message "CKSUM" will appear immediately above the CAS
        number or below the BASE CAS number, indicating that an incorrect number
        has been entered.

BASE    The "base" CAS NO.  If the analyte is derived from a compound class
        (e.g., "Chlorobenzenes"; "Silver and Compounds, NOS"), and the analyte
        can be traced to the class, the CAS NO of the class from which the
        analyte is derived will appear in this field.

ORIGIN  An acronym or abbreviation for the list from which the•analyte is
        derived.  The following lists are used:
        AIR     Analytes of the air "List of 37"
        APP-C   Analytes listed in Appendix C of the Consent Decree.
        APRIL   Analytes added to the RCRA groundwater monitoring list by
                Bob April of EPA.
        CAL     California List pollutants  [40 CFR Part 268, Appendix III;
                52 FR 25791].

-------
                                     37
page 2            EXCERPT FROM ON-LINE HELP FILE (cont.)          I_RCRA,DSC
                                                                        02/27/89
        CER_302 CERCLA Reportable Quantities List [40 CFR 302.4,  Table 302.4].
        CWA_116 Hazardous substances under Section 311(b)(2)(A)  of the Federal
                Water Pollution Control Act [40 CFR 116,  Table 116.4A] and
                Reportable Quantities-HO-eFirliTT Table  117.3].
        CWS_DIS List of analytes for which Community Water Systems and non-
                transient, non-community water systems shall monitor at the
                discretion of the State [52 FR 25715, 08  Jul 87].
        CWS_REQ List of analytes for which Community Water Systems and non-
                transient, non-community water systems shall monitor [52 FR
                25715, 08 Jul 87].
        DWPL    Draft Priority List of Drinking Water Contaminants
                [52  FR 25720]
        FTC     ITD's list of Fish Tissue Contaminants
        ITD     Additional metals, classical analytes, and dioxins that the
                Industrial Technology Division monitors in its sampling and
                analysis programs.
        MICH    The  list of analytes proposed to be added to the RCRA Appendix
                VIII List by the Michigan Petition [49 FR 49793,  21 Dec 84].
        OAG_SRB Oil  and gas,  secondary recovery biocides:  biocides, slimicides,
                and  molluscides used on oil platforms.
        P-POLL  The  priority pollutant list [NRDC vs Train, 8 ERC 2120 (DDC
                1976)] as expanded to the 129 "Priority Pollutants", Appendix
                C Pollutants, and High Priority Paragraph 4(c) Pollutants.
                (The specific compounds on this combined list are given in
                Methods 1624, 1625, plus the original Priority Pollutant list
                of pesticides, metals, cyanide, and asbestos).
        PARA-4C The  list of 56 compounds detected in the 4(c)  study.
        PARA_4C The  remaining 367 compounds detected in the 4(c)  study.
        RCRA    RCRA Appendix VIII list [51 FR 28305, 06  Aug 86].
        RCRA_IX The  RCRA Appendix IX Groundwater Monitoring List [51 FR 26632,
                24 Jul 86].
        RPAR    "Rebutable Presumption Against Registration" - compounds EPA is
                considering removing from registration as pesticides.
        SARA110 Hazardous substances most commonly found at facilities on the
                CERCLA National Priorities List [52 FR 12866]  under Section 110
                of the Superfund Amendments and Reauthorization Act (SARA).
        SDWA    Safe Drinking Water Act Amendments of 1986 [House Report 99-575]
        SEC_112 Pollutants listed as hazardous under the Clean Air Act.
        SEC_313 The toxic chemicals subject to the provisions of Section 313 of
                the Emergency Planning and Community Right to Know Act of 1986
        TCL     Superfund Target Compound List (current as of August 1987).
        VTOX    Compounds on the "Acutely Toxic Chemicals" List in EPA's
                Chemical Emergency Preparedness Program [EPA OPTS-00066,
                November 1985; 52 FR 13378], mandated under Section 302 of
                the Superfund Amendments and Reauthorization Act (SARA).

SEQUENCE  The sequence number on the respective list.  In those instances in
        which the list was unnumbered, a sequential number starting with 001
        was assigned to each analyte, except for the HSL in which the number
        used by the QA Formaster data base was used  (because a sequential
        reference number would change every time the HSL is revised), and
        numbers beginning with Z to represent the atomic number of a given

-------
                                       38
page 3            EXCERPT FROM ON-LINE HELP FILE (cont.)

        element  (e.g., boron is Z05) .
                                                        I_RCRA.DSC
                                                          02/27/89
REGULATORY NOTES (Column heading not given; found next to "SEQUENCE" field)
        Location for encoding information pertinent to a given regulation.
        RQ      Reportable quantities under CERCLA and FWPCA (CWA).

NAMES AND COMMENTS  Various names for this analyte and other unrestricted com-
        ments.  Names are listed in approximate order of common usage.  Each
        distinct name starts on a new line and has no leading spaces.  Contin-
        uation lines for this name should have four leading spaces.  Blank lines
        and comments can be entered anywhere between names.  Comment lines
        should have two leading spaces.

        The I_JSXPORT program assumes the following about names:
        1.      The longest name is 168 characters — not counting leading
                spaces.
        2.      When making a continuation line back into a full name, a space
                should be inserted if the previous line ends in a letter, digit,
                colon, semicolon, close parenthesis, close bracket, or a comma
                preceded by a dash.
ORGANIZATION
        ASTM
        CIN

        CLP
        ITD
        ODW
        OSW
        STD

        USGS
The organization originating the method, as follows:
  American Society for Testing Materials
  EPA's Environmental Monitoring and Support Laboratory in
  Cincinnati, Ohio
  EPA's Office of Emergency Response Contract Laboratory Program
  EPA's Industrial Technology Division
  EPA's Office of Drinking Water
  EPA's Office of Solid Waste
  "Standard Methods" published by the American Public Health
  Association
  US Geological Survey Techniques of Water Resources Investiga-
  tions

                          The following apparata are encoded:
APPARATUS  As derived from the METHOD.
        BRIDGE  Conductivity bridge.
        CGCEC   Combination method using gas chromatography with electron
                capture detector.
        CGCFPD  Combination method using gas chromatography with flame
                photometric detector.
        COLOR   Colorimetric determination
        COUL    Coulometric detector
        CS2     Analysis of a carbamate by liberation of carbon disulfide.
        CVAA    Cold vapor Atomic Absorption Spectrometry
        DICHROM Dichromate oxidation
        EVAP    Evaporation
        FID     Flame ionization detector
        FILTER  Filtration
        FLAA    Flame atomic absorption spectrometry
        FURNAA  Furnace atomic absorption spectrometry
        GCAFD   Gas chromatography with alkali  flame detector
        GCEC    Gas chromatography with electron capture detector

-------
                                      39
>age 4             EXCERPT FROM ON-LINE HELP FILE (cont.)         I_RCRA.DSC
                                                                        02/27/89
        GCFID   Gas chromatography with  flame  ionization  detector
        GCFPD   Gas chromatography with  flame  photometric detector
        GCHRMS  Gas chromatography with  high resolution mass spectrometry
        GCHSD   Gas chromatography with  halogen specific  detector  (Hall,
                O.I.,  microcoulometric,  electrolytic conductivity)
        GCMS    Gas chromatography/mass  spectrometry
        GCNPD   Gas chromatography with  nitrogen-phosphorus  detector
        GCPID   Gas chromatography with  photoionization detector
        GRAV    Gravimetric
        HPLC    High performance liquid  chromatography
        HPLCUV  HPLC with an ultra-violet detector
        HYDAA   Hydride atomic absorption spectrometry
        ICP     Inductively coupled plasma spectrometry
        MICRODF Micro diffusion
        NEPHELO Nephelometer
        OXY-FID Oxidation/reduction followed by flame ionization detection
        OXY-IR  Oxidation followed by infra-red detection
        PHMETER pH meter
        RETORT  Oil platform operator's  apparatus for determining  oil content
                of a drilling fluid
        SCINT   Scintillation counter
        SPECTRO Spectrophotometer
        TITR    Titration
        WET     Analysis by a classical  wet method
        WINKLER Incubation in airtight bottle
METHOD  The Method number where it is known.
        associated with an ORGANIZATION.
SUFFIX
STD
                                      Methods and method numbers are
The suffix to the METHOD.  The suffix is specific to the sample
fraction, matrix, and level.  Suffixes are as follows:
Suffix  Frac    Matrix  Level
AW      Acid    Water
BN      B/N
BNW     B/N     Water
CHS     Combine Solids  High
HS              Solids  High
LS              Solids  Low
MS              Solids  Medium
S               Solids
W               Water
Sources for a standard for the analyte.
are defined as follows:
ALD     Aldrich, Milwaukee, WI
ALF     Alpha
ATH     Athens ERL
CIL     Cambridge Isotope Laboratories
GIN     EMSL Cincinnati
DUP     DuPont
EPA     EPA repository at RTP
EXX     Exxon
Acronyms and abbreviations

-------
                                       40
page 5
   EXCERPT FROM ON-LINE HELP FILE (cont.)
I_RCRA.DSC
  02/27/89
        LV      EMSL Las Vegas
        NAN     Nanogens, Watsonville,  CA
        NCI     National Cancer Institute
        PAB     Pfaltz & Bauer
        SCC     Sample Control Center
        SIG     Signal
        SUP     Supelco
        SYN     Must be synthesized in the lab
        ULT     Ultrex

HYD/DEC  The analyte hydrolyzes (H) as estimated by Athens-ERL,  or decomposes
        (D) as determined by Dr Beimer of S-CUBED.

EXT/PUR  The analyte can be extracted from water as determined by Athens-ERL
        or by S-CUBED.  E=can be extracted; P=can be purged? N=can neither be
        extracted nor purged.

GC POSS  The analyte can be gas chromatographed as determined by Athen-ERL or
        S-CUBED.

LC POSS  W Roy Day of Waters Associates believes that the analyte can be deter-
        mined by LC.

EPA NIH PAGE  The page number in the EPA/NIH mass spectral file where the
        reference mass spectrum for the analyte can be found.

*** Information Specific to Screen 2 ***
DETECTION LIMIT
        EDL
        MDL
        ML
        PQL
Estimated Detection Limit
Method Detection Limit [49 FR 43234, (Appendix B)]
Minimum Level-ITD's definition of the minimum level that must
give recognizable mass spectra and acceptable calibration
points (see footnote 2 to Table 2 of Method 1624, Revision B
[49 FR 43234].
Practical Quantitation Limit-EPA Office of Drinking Water
definition of the lowest level that can reliably achieved
within specified limits of precision and accuracy during
routine laboratory operating conditions [52 FR 25699], or
the Office of Solid Waste definition of EPA's current best
estimate of the practical sensitivity of the applicable method
for RCRA groundwater monitoring purposes [52 FR 25945].
NOTES  (Column heading not given; found next to "DETECTION LIMIT" field)
        Notes pertaining to a given method for the analyte.

-------
                                41
                                   MR. TELLIARD:  Our next



speaker is Dave Mount who is going to talk about the joys of



a toxicity-based approach to pollutant identification.

-------
                             42
                              MR. MOUNT:  Thank you, Bill.
     I want to start with two disclaimers which should make
you all nervous.  Could I have the first slide, please?
     For those of you who are astute, you will notice that I
have made a one-word change in my title from "A Toxicity-
based Approach" to "The Toxicity-based Approach."  The
reason I did this has to do with my first disclaimer.
Although I am presenting this information today, it is by no
means a lone effort on the part of our laboratory.  It
really represents the efforts of a great many laboratories.
In particular, the EPA lab in Duluth, Minnesota, was heavily
involved in the development of some of the techniques I will
discuss today.
     The other disclaimer I want to make is that I am not a
chemist.  I am an aquatic toxicologist, and I think that is
an interesting point to make prior to this talk.  As we have
entered the world of toxicity-based toxics identification,
it has become apparent the need for interdisciplinary
communication and effort in order to most effectively deal
with these problems, because the knowledge that is needed
really far exceeds the knowledge of any one person.
     I need to provide information as a toxicologist.
Analytical chemists and other types of chemists as well have
to provide me information from their disciplines.  Engineers

-------
                             43
and all kinds of people have to become involved in order to
make this effective.
     As many of you are aware, toxicity, as measured in
toxicity tests, is gaining fairly wide acceptance, at least
in some areas of environmental regulation.  The most notable
of these is the NPDES effluent discharge program, but there
are other areas in which toxicity as a unit of measure is
becoming favored as well. "The hazardous waste
classification system is using it some places.  EPA has
recently issued guidance on using biological assessment
techniques to look at hazardous waste sites.
     And there are a lot of reasons why.  There are several
advantages to using toxicity testing for environmental
monitoring.  One is that measuring toxicity addresses a lot
of questions that aren't addressed by general analytical
techniques, such as are questions of bioavailability.  A
chemical analysis can tell you if a contaminant is present,
but it can't tell you whether it is in a form that will
cause adverse effect to organisms out in the environment.
     Another consideration is matrix effects.  Of course,
all analytical chemists are familiar with matrix effects,
but the same things happen in toxicity testing.  A common
one we are probably all familiar with is the effect of
hardness on metal toxicity.  There are considerations
relating to the physical/chemical environment in which a

-------
                             44
contaminant exists in the environment that will heavily
influence its toxicity.
     Finally, there are interactions among toxicants.  We
can study a single compound to death and know exactly how
toxic it is in and of itself, but when it is present with
other contaminants, how toxic will it be?
     All of these considerations are addressed by toxicity
testing.  As some of us are fond of saying, an organism
knows more than we do.
     Another feature of toxicity testing is that it will
detect the presence of all toxic chemicals that are there,
if they are present in toxic amounts; and that beats any
analytical chemistry technique known.  Every last one is
detected if it is present in toxic amounts.  An interesting
sidebar to that quote is the toxic materials in toxic
amounts reference, which is in the Clean Water Act, and is
encompassed by toxicity testing as a monitoring approach.
     However, if we are to regulate on the basis of
toxicity, we have an obligation, to develop means to
determine sources of toxicity, so that control strategies
can be implemented.  This gets into the focus of what I am
going to talk about today.  This led to the buzz words, the
"toxicity reduction evaluation," or TRE as it is commonly
referred to.  In the case of effluents, a TRE is a
systematic evaluation, of both plant and effluent to

-------
                             45
identify sources of toxicity and, ultimately, to control
toxicity.
     Now, people have been trying to do things like this for
a long time, and the traditional approach has been
analytically-based,  using tools like a priority pollutant
analysis.  If you have something toxic, what do you do?  You
do a priority pollutant analysis and find out what is in
there.  The problem is that in order for this to be
effective, you have to be measuring whatever it is that it
toxic.  Another problem is that analytical approaches, at
least most of them, don't account for the questions I spoke
of earlier, bio-availability matrix effects, and
interactions among toxicants.  Hence, this analytical
approach generally falls short, in our experience, of
identifying toxicants in a complex matrix such as an
industrial effluent.
     The other approach that has been used is the
engineering and treatability approach.  Basically, you don't
worry about what is causing toxicity, you just find some
treatment strategy that will remove it.
     There are some real disadvantages to this approach
also.  First of all, you run a large risk of having your
treatment remove toxicity, but doing it in a way that you
didn't expect or that does not lie along the usual means of
that treatment technology.  Another is that if you don't

-------
                             46
know what it is and you are just going to treat it, things
like source control aren't an option.  If you don't know
what it is, you can't figure out how to stop putting it in.
     This is a very simple schematic example I will use to
talk about a couple concepts (Fig. 1).  Let's imagine that
the center box is a waste water treatment plant.    We have
two influent waste streams in our waste water treatment
plant.  If we were to measure the toxicity of Influent I, it
would be moderately toxic, and it would have a toxicity due
to compound A, although we don't know what compound A is at
this point.  Influent 2 would be extremely toxic.  It has
toxicity due to compound B.
     Both influents enter the waste water treatment plant.
Compound B is effectively treated.  Compound A is resistant
to treatment, and we end up with a final effluent that is
moderately toxic, and it is toxic due to compound A.
     Now, if we were to go and look for toxicity back up
through this system, we would say it looks like Influent 2
is the source of our toxicity when, in fact, it has nothing
to do with it.
     If we were to take the analytical approach, we would
analyze the final effluent.  Well, if compound A is a
priority pollutant or other commonly analyzed parameter, we
might pick it up.  On the other hand, I have had lots of
people come up to me with a list of tentatively identified

-------
                             47



GC/MS compounds and say, "are any of these toxic?"  And I



say, " yes, they are all toxic".  But the question is, how



much?  Furthermore, with all the other matrix effects and



interactions, you simply can't take a list like that,



compare it to toxicological benchmarks, and know whether or



not any of those compounds are causing your problem.



     So, the goal of the procedures that I am going to talk



about today is to work on that final effluent, and using



toxicity-based procedures, find the identity of compound A.



Once we know what compound A is, then we can take analytical



approaches, go back through the system, and find out exactly



where compound A comes from.   And with that all kinds of



control options become possible.   In addition to just



treating the final effluent for compound A, we can treat



just that influent line, we can modify the source, we can do



all kinds of things that we couldn't do before.



     So, the strategy for the toxicity-based approach is to



use simple separation chemistry techniques to separate toxic



components of the effluent from other components of the



effluent.-  In all cases, we use toxicity tests as our



analytical detector, rather than flame ionization or some



other standard analytical detector.  We use a toxicity test



to find out, for that separation technique, where did the



toxicity go?

-------
                             48
     This has led to the development of a subset of studies
under the TRE, the Toxicity Identification Evaluation or
TIE.  As you might have guessed from my discussion earlier,
the objective of the TIE is to relate observed toxicity to a
parameter that has application to engineering solutions.
     A lot of what I'll talk about today is heavily geared
toward effluent toxicology, because that is where the
majority of the toxicity limits have been placed.  In the
case of hazardous waste sites, a toxicity test can also be a
very useful tool for determining whether or not a hazardous
waste presents a potential toxicological effect.  But more
than just that, we then need to know what the toxic compound
is.  And if we know what it is, then we can go about setting
cleanup criteria.  So, there are all sorts of other
applications of this approach.
     The phased approach to the TIE was developed largely by
the EPA Duluth lab and consists of three phases.   I am not
going to go through each phase indepth, except to tell you
that they exist.  There are guidance manuals that have been
issued by EPA that explain each in great detail.  I am going
to talk specifically about some of the tests, some of the
information that you get out of and these tests, and how
this information is applied.
     The approach for a Phase I TIE may not seem very
unusual (Fig. 2).  It is a lot like a traditional

-------
                             49
engineering treatability approach.  We have a whole effluent
sample or any other sort of environmental sample, that can
be used in a toxicity test, a sediment, anything.  We split
it.  On the left side, we do a toxicity test on it.  Down
the right side, we do some sort of physical/chemical
manipulation, and as you will see in a minute, there are a
great many of them that are done in parallel.     We conduct
a toxicity test on the manipulated sample and then compare
the toxicity in both of the tests to determine whether or
not that physical or chemical manipulation had any effect on
the causative toxicant in the matrix.
     I hate to get into methodological detail, but I want to
make sure we are on the same wavelength as we go through
some example data.  These are some of the tests that are
used in this procedure.  (Fig. 3).  At the top, we have the
whole effluent test which is simply a standard toxicity
test.  The whole effluent or other sample, is diluted and
tested for toxicity along a concentration series to
determine just how toxic that material is.  This serves as
the comparison for all the other tests.
     There are two pH adjustment tests, pH 3 and pH 11.  The
sample is simply adjusted to that pH, left there for three
hours and then returned to the original pH.  Then the
toxicity of the manipulated samples is measured using
toxicity tests.

-------
                             50
     The pH adjustment tests primarily address compounds
that undergo some sort of physical transformation at those
pH's, an irreversible transformation like hydrolysis.  They
also serve as the procedural blank for other tests that use
pH adjustment.
     We also use aeration tests.  In these, we use aliquots
of sample at the initial pH, and also at acidic and basic
conditions.  These aliquots are aerated for 30 minutes,
returned it to the original pH, and then tested it for
toxicity.  These tests obviously address volatile or easily
oxidizable compounds.
     The filtration tests have a similar structure to the
aeration tests except that  we filter the sample instead of
aerating at those three pH levels.  This gets at materials
whose toxicity is associated with filterable solids or
toxicants whose solubility is affected at extremes of pH.
For example, a lot of heavy metals will precipitate at pH
11, and we can filter them out.   If they were the cause of
toxicity, the sample will lose its toxicity.
     Solid phase extraction is a very simple technique using
Sep-Pak CIS columns.   We run the effluent through the
column, and it removes a great many non-polar organic
compounds.  Like the aeration and filtration tests, this is
done at acidic, basic, and neutral conditions.

-------
                             51



     We use an EDTA test in which we add EDTA over a range



of concentrations to determine if that has any effect on



toxicity.  EDTA will reduce the bioavailability of many



heavy metals.  Therefore, if those are the source of



toxicity, we will see a reduction in toxicity when we add an



appropriate amount of EDTA.



     There is an oxidant reduction test using sodium



thiosulfate.  This test has the same structure as the EDTA



test, except use sodium thiosulfate to reduce residual



chlorine and other oxidants.



     Finally, there is a graduated pH test where we adjust



the pH to 6, 7, and 8, test the toxicity, and see if there



is a marked difference in toxicity at those different pH's.



Toxicity of many materials is greatly affected by pH.



Ammonia is a good and common example of a pH-sensitive



toxicant.



     So, what kind of information do we get out of this?  I



am going to show you some specific data in a moment, but



this is conceptually what we find out.  These are results



from a wire coating facility in the Northeast, showing the



information from a Phase I characterization. (Fig. 4).



     The solid phase extraction test, which is geared toward



non-polar organics, did not remove toxicity.  EDTA chelation



did remove toxicity.  With that, we begin to think metals



may be the cause in this case.  Aeration did nothing at any

-------
                             52
pH.  Acid -and neutral filtration did nothing.  However,
filtration at pH 11 did remove the toxicity.
     So, we can say at tliis point we feel pretty confident
we are zeroing in on heavy metals.  We did some more tests
to confirm that heavy .metals were,, in fact, the cause,  I am
not going to go through tlio.se, tout once we knew that it was
a metal, we turned to analytical approaches to determine
which metal it actually -was.
     These are actual data from a municipal wastewater
treatment plant effliuent !'(Fl
-------
                             53



and test the toxicity of the post-column effluent.  It is



not toxic.  We assume, therefore, that whatever is toxic is



on that column.



     Next we do a very simple elution of that column, using



a series of methanol concentrations.  We use methanol



because methanol is not very toxic to the organisms, so we



can do toxicity tests with samples that are in a methanol



matrix, as long as they are appropriately diluted.  Other



solvents like hexane, more common analytical solvents, are



too toxic to be of much use in this procedure.



     At any rate, we elute the column with a series of



methanol fractions, ranging from 25 to 100 percent.  Then we



test each one these eight functions for toxicity.



     The resulting 48-hour LC50  valves are shown here in



percent (Fig. 7).  The LC50,  again,  is the theoretical



concentration of that sample that would cause 50 percent



mortality of the test organisms.



     Of the eight methanol fractions, only two showed



toxicity, the  85 and 90 percent fractions.   The recovery



of this toxicity shows that we have the causative toxicant



out of the effluent matrix and have begun to separate what



is toxic from what is not toxic.



     We can extend these results by conducting additional



separation procedures.  And at every step along the way, we

-------
                             54
do toxicity tests to confirm that we still have the
causative toxicant isolated.
     We combine the toxic fractions, concentrate them on a
C18 column, inject the concentrate into the HPLC, and  split
those toxic fractions into 25 fractions.  For illustration
purposes, we'll assume that compounds distribute themselves
randomly among all functions.  We took the original sample,
split it in 8 fractions, then threw out 6 and kept 2.  Based
on this, we would be  down to a quarter of the original
compounds in the sample.  Then, we split the remaining
compounds into 25 more fractions.  In theory, then, each of
these fractions contain only 1 percent of the compounds that
were in the original sample, if you will forgive the
simplistic assumptions.
     After the HPLC separation, we test each of the 25
fractions for toxicity.   In this particular case, we found
that fractions 21 and 22 were toxic.  We then combined those
two fractions, concentrated them, and submitted them for
GC/MS analysis.  This analysis turned up diazinon in
sufficient concentration to explain the toxicity of the
sample.
     There are some analytical wrinkles that can arise from
this process.  First, the question being asked of the
analytical chemist is not "How much of compound X is in this
sample?"  Instead, we are asking "There is an unknown toxic

-------
                             55
compound in here; what is it?"  As you no doubt realize,
those two questions are quite different, and require
different analytical thinking.
     Another potential analytical difficulty is that not all
organic compounds are amenable to GC/MS analysis.  Knowing
whether this is the case is impossible, a priori, since we
don't know what it is we are analyzing for.  This kind of
problem also requires some innovative thinking, and
sometimes some innovative techniques.  HPLC-MS and super-
critical fluid extraction are some alternative procedures
that are showing some promise for these situations.    This
diagram illustrates some ongoing research that we are doing
in our laboratory (Fig. 8).   One of the big problems we run
into is heavy metal toxicity, because the bioavailability of
heavy metals varies widely.   Even if you measure total,
total recoverable, and dissolved concentrations, toxicity
will not necessarily relate to any one of those.  If you
have a concentration below the toxicological benchmark, you
can feel confident that it is not a source of toxicity.
But, for example, a great many effluents in this country
have potentially toxic amounts of zinc in them, yet they
aren't toxic.  The reason is that the zinc is not in a
bioavailable form.

-------
                             56
     To identify the heavy metal responsible for toxicity,
we need to give the analytical chemist a boost. What we have
done is to combine tests using two different metal ligands.
     Although the sodium thiosulfate test is intended to
remove toxicity due to oxidants, we discovered during the
course of one TIE that the sodium thiosulfate test would
remove copper toxicity.  When we went back and asked the
chemists about this, they weren't surprised, as sodium
thiosulfate will chelate copper.
     We decided to explore this further, by conducting both
EDTA and thiosulfate tests on a range of metals.
The results are presented here (Fig. 8) in a simple 2 by 2
contingency table, with toxicity removal by EDTA in the left
column, and no removal by EDTA in the right column.  The
same applies to results of the thiosulfate test, positive
results in the upper row and negative in the lower.  So, for
example, in the case of that original effluent, we noticed
that both  the EDTA test and the thiosulfate test took out
toxicity.  At the time we did that study, we didn't have
this table to look at.   But if we had, we could have
pointed to the yes/yes box in the upper left, and would have
suspected that the toxicant was either copper, cadmium, or
inorganic mercury.
     As another example, we recently completed some work on
a drain water from an historic mining operation.  It was

-------
                             57

quite toxic and had a myriad of heavy metals in it.  We

conducted both tests on the sample and found that EDTA  the

removed the toxicity and sodium thiosulfate didn't.  By

using this table, we then removed the range of metals that

might be responsible for toxicity.  So, even though there

were a tremendous number of heavy metals in the effluent, we

zeroed in on zinc as being the only metal in this category

that was present at potentially toxic concentrations.

     I have a final example TIE, to illustrate that it is

not always as simple as I have been showing,  this effluent

had an llC50  of  about 33 percent,  and solid phase extraction

removed this toxicity.

     In order to get you to discuss the fine points of this

example, I have to introduce one more term, the toxic unit.

A toxic unit is 100 percent divided by the LC-50 of the

solution.  An effluent with LC-50 of 50 percent, has 2 toxic

units.  One with an LC50 of 25 percent would have 4 toxic

units.  The number of toxic units also indicates the number

of times you need to dilute that sample to reach 50 percent

survival in a 100 percent sample.

     As I said, the toxicity of this effluent was removed by

solid phase extraction.  We then conducted the solid phase

(Fig. 9) extraction and elution that I showed you earlier,

with the following results.  The whole effluent had an LC50
                                                           <^
of 33 percent, corresponding to 3 toxic units.  The toxicity

-------
                             58
tests on the fractions showed toxicity in the 75, 80, 85,
and 90 percent fractions.  There is a concentration step
involved with the SPE dilution, which is 5x with the
methodology that we use.  Calculating toxic units and
dividing by 5 we have these toxic units listed on the right,
corrected to the whole effluent concentration.  Adding the
four toxic fractions together, we had 1.3 toxic units, but
we had removed 3 from the whole effluent.
     Well, this kind of disturbed us, because it meant
either  that something was still stuck on that column or
that we had very, very poor recovery for some reason.  In
this case, it is hard just to live with poor recovery,
because of the risk that it actually indicates  the presence
of a second toxicant.
     The clue to this problem came in another sample (Fig.
10).  We did a graduated pH test on  this other sample, and
found no survival at pH 6, 100 percent survival at pH 7, and
no survival again at pH 8.
     Now, there was ammonia in this sample, and ammonia is
more toxic at high pH, so we were expecting and could
explain the toxicity at pH 8.  But we couldn't explain the
toxicity at pH 6.  Because we had seen non-polar organic
toxicity in this effluent before, we did a solid phase
extraction on that sample and then did the graduated pH
test.  This manipulation took out the toxicity at pH 6.   We

-------
                             59



assumed/ therefore, that the toxicant responsible for the



toxicity at pH6 was now on the CIS column.  Since ammonia



isn't affected by the column, there was still toxicity at



pH8.



     Having found that out, we did the solid phase elution



(Fig. 11).'



     As background information, the effluent pH was in the



high 7's, while the pH of the dilution water that was



originally used for testing fractions was about 8.5.  When



we tested the toxicity of these fractions at pH 8, we saw no



toxicity at all.  But when we tested them at pH 6, we saw



toxicity in the 75 and 80 percent fractions, and the pieces



began to fall together.



     What had happened in the first sample was that the



effluent fractions were tested in dilution water with pH



just a few tenths higher than the whole effluent.  Because



this material appears to be more toxic at lower pH than at



higher pH, we could thereby account for the difference in



toxicity that we observed originally.



     With that, I would be happy to open it up for any



questions you might have.

-------
                             60



               QUESTION AND ANSWER SESSION



                                   MR. HACHIGIAN:  Lee



Hachigian of General Motors.



     Have you tried the guidance manual with regard to



chronic toxicity applying the same types of methodologies?



                                   MR. MOUNT:  Yes, we have.



For those of you who aren't familiar with toxicity testing,



there are two general types of toxicity tests, acute and



chronic.  I have been talking about acute toxicity, which



generally deals with survival.  Chronic toxicity a longer



term test, and deals not only with survival but with more



subtle and points.



     Yes, we have worked on applying the TIE methods to



chronic toxicity.  That is certainly an area of heavy



methods development right now.  EPA Duluth is working on it.



We are working with a couple of dischargers who are being



regulated on chronic toxicity, and have to find ways to



control it.



     It is definitely a trickier situation.  There are a lot



of shortcuts that can be taken with acute toxicity that are



more difficult with chronic toxicity, but some of the



methods still apply.  One of the problems is simply that the



chronic tolerance of organisms for substances like EDTA and



thiosulfate haven't been worked out thoroughly yet.  But



that information is coming, and I would imagine...well, I am

-------
                             61
not going to speak for when EPA will issue guidance, but
they will probably be technically ready to issue guidance
within the very near future.
     Yes?
                              MR. HOWE:  Stavros Howe with
the Molecular Ecology Institute.
     I think there is a whole suite of assays that currently
exist that would compliment your scheme there very well and
perhaps would be more sensitive in addressing some of the
chroniti toxicity issues.  Specifically, I am thinking about
stress protein assays, metal binding ligand assays, various
enzyme assays that would correlate much closer with the
mechanism of toxicity and give you a better handle on what
is causing the problem, basically.
     I think the screening approach would be far more rapid
and much more cost effective.
                              MR. MOUNT:  A lot of other
assays have been suggested.  There are a couple comments I
might make on the choice that is made.
     For those of you who are familiar with toxicity testing
as a monitoring tool, you will know that they cost from
several hundred to a thousand dollars each test.  When they
are done in the context of a TIE, however, they are much,
much less expensive, down on the order of $50 per test.  So,
the cost effectiveness issue starts to go away.

-------
                             62
     People get into this situation as the result of
compliance monitoring, which is whole organism monitoring.
Therefore, whole organism testing is a very appropriate
measure to use, because it is the measure by which the
problem is defined.
     The danger in using other assays is that they may
respond to things other than what is causing the toxicity to
the original organism, other things that are present in the
effluent.  For example, the Microtox assay has been
suggested for use, and it is an effective tool if it is
responding to the same material as is the original test
organism.
     If you use something other than the original organism,
you need to do some background work to establish for certain
that there is a correlation between the alternate assay and
the tests that is of regulatory concern.
     Finally, I'll admit that this is a regulatory policy
constraint on the scientific process, and what you are doing
the study for depends on how you might approach that
problem.
                                   MR. YOCKLOVICH:  Steve
Yocklovich from Burlington Research.
     While we use this type of analysis ourselves, I was
wondering what you would see if this was implemented
nationwide.  What kind of cost of analysis could you see if

-------
                             63
it were put into regulation, and how many labs in the
country are equipped to do it?
               MR. MOUNT:  Let me get all the parts of your
question here.  As far as implementation, it has been
implemented in many, many parts of the country.  In our
region, Region VIII, and in most of the EPA regions in the
East, it is in place, and we are working with several
clients who are doing it not as a proactive stance, but who
are in it as a result of their permit monitoring.
     In terms of cost, we have solved problems for a few
hundred dollars if it is real obvious.  It can also cost a
lot more than that.  But I would venture to say some of the
horror stories that are floating around about the potential
cost aren't realistic, if the studies are properly done.
There are stories running around about expenditures of
$200,000 or $250,000.  We certainly haven't run up any bills
like that, and I don't think that it is necessary.
     This is an emerging area, so there is varying expertise
out in the world regarding how effective labs are.
     In terms of equipment, there is really very little
equipment needed aside from the analytical capabilities. Of
course, what analytical equipment you need depends on what
direction you end up going, metals, non-polars, or some
other compound.  But in terms of the laboratory equipment to
do the separations and the toxicity tests, there really

-------
                             64



isn't much.  A couple thousand dollars in equipment is all



that is required.  Obviously, the analytical situation is



different.



     Within our company, we have the full range of



analytical equipment.  On-site, we have GC, HPLC, and AA.



If we run into GC/MS work, we, often contract with a



laboratory that happens to be right across the street from



us.



     Is there anything I didn't touch on you want to follow



up with?



               MR. YOCKLOVICH:  Thank you.



               MR. MOUNT:  Thank you.



               MR. TELLIARD:  Any more questions?



               MR. FRAZIER:  In the TRE/TIE approach, there



is not anything specifically addressing the diseases of



organisms.  Are you all aware of any body of knowledge of



this either affected by effluents or once you get them in



the laboratory?



               MR. MOUNT:  There are diseases of organisms



that definitely influence test results.  In terms of



diseases inherent to the laboratory, the laboratory



culturing and quality assurance practices should address the



presence of pathogens in the laboratory in general.



     As far as pathogens in the effluent, we have never had



any experience with that.  I do know of one example, but I

-------
                             65



haven't reviewed the data to know if I feel their



conclusions are accurate.  But I do know of one example



where a municipal plant had an experience where they were



convinced that a bacterium growing inside their sampler was



causing disease in the toxicity tests.



     And there are some things as a toxlcologist you can do



to spot that kind of problem.  For example, if you have



toxicity due to a pathogen^ you essentially have a



toxicant that can multiply.



     As a result, things like concentration response



relationships will likely fall apart.  Frequently, you will



see higher mortality at lower concentrations,, or other



irregularities that lead you to suspect that something else



may be involved.



     Another option If you suspect that, of cours-ev, is to



use ultrafiltration to try to physically sterilize the



sample, and remove at least bacterial pathogens from the



matrix.



               MR. TELLIARD:  Thank you, Dave.



               MR. MOUNT:  Thank you.

-------
                      66
   Influent #1

   Moderately toxic
   (Compound A)\
 Influent #2

 Extremely toxic
/(Compound B)
Figure 1
                     WWTP

               Compound A resistant
                 to treatment

               Compound B very
               amenable to treatment
                      Effluent
                  Moderately Toxic
                   (Compound A)

-------
                        67
        APPROACH FOR A PHASE I TIE
                      Whole Effluent Sample
                                        Physical/Chemical
                                         Manipulation
        Toxicity Testing
Toxicity Testing
                        Compare Toxicity
Figure 2

-------
                          68
      Phase I - Toxicity Characterization

     Physical/Chemical Manipulations

         Baseline Toxicity
         pH 3 Adjustment
         pH 11 Adjustment
         pH 3 Aeration
         Initial pH Aeration
         pH 11 Aeration
     -    pH 3 Filtration
     -    Initial pH Filtration
     -    pH 11 Rltration
         pH 3 Solid-Phase Extraction
         Initial pH Solid-Phase Extraction
         pH 9 Solid-Phase Extraction
         EDTA Chelation
         Oxidant Reduction
         Graduated pH

     Toxicity Testing

         Compare toxicity of manipulated effluent samples
         with toxicity of whole effluent
Figure 3

-------
                       69
                   Phase I TIE
Manipulation
Reduce Toxicity?
Yes         No
Solid-Phase Extraction
  (non-polar organics)
EDTA Chelation
Aeration (any pH)
Acid or Neutral Filtration
pH 11  Filtration
             X
            X
            X
Figure 4

-------
                        70
               Municipal WWTP
               Phase I  Results
    Manipulation
48-hour Survival (%)
   Whole Effluent
   pH 3 Adjustment
   pH 3 Filtration
   pH 3 Aeration
   pH 3 SPE
   Initial pH Filtration
   Initial pH Aeration
   Initial pH SPE
   pH 11  Adjustment
   pH 11  Filtration
   pH 11 Aeration
   pH 9 SPE
   EDTA Chelation
   Oxidant Reduction
       20
       60
       20
       100
       100
        0
       100
       100
       20
        0
       100
       100
       0-80
       0-20
Figure 5

-------
                            71
                        SPE Elution
                           1 Liter
                             of
                          Effluent
                            SPE
                           Column
25%

50%

75%

80%

85%

90%

95%

100%
          Test Each Methanol Fraction for Toxicity (5x)
Figure 6

-------
                      72
            SPE Elution Results
   Fraction
Survival (%)
     25%
     50%
     75%
     80%
     85%
     90%
     95%
     100%
    100
    100
    100
    100
     0
     0
    100
    100
Figure 7

-------
                       73
                Toxicity Removal by EDTA
                Yes                  No
M
o
S
CO
I
0)
DC  o
>» 2
'o
'x
o
 Copper Chloride
Cadmium Chloride
 Mercuric Chloride
            Zinc Chloride
         Manganese Chloride
             Lead Nitrate
           Nickel  Chloride
                                  Silver Chloride
                                 Sodium Selenate
                          Iron Chloride
                     Chromium [III] Chloride
                     Potassium Dichromate
                       Sodium m-Arsenite
                        Sodium Arsenate
                        Sodium Selenite
                       Aluminum Chloride
  Figure 8

-------
                         74
                Refinery Effluent
                  SPE Elution
Test
48-h LCcn Concentration Toxic Units
                     '50
Whole Effluent
25%
50%
75%
80%
85%
90%
95%
100%
33
>100
>100
71
55
62
71
>100
>100
1x
5x
5x
5x
5x
5x
5x
5x
5x
3
-
-
0.3
0.4
0.3
0.3
-
_
Total of Fractions
Figure 9
                         1.3

-------
                     75
               Refinery Effluent


       Graduated pH Test (24-hour survival)
           pH 6       pH 7      pH 8
            0%       100%        0%

       After Solid-Phase Extraction
           pH 6       pH 7      pH 8
           100%       100%        0%
Figure 10

-------
r
                             76
                     Refinery Effluent
                       SPE Button
           Fraction
 24-hour Survival (%)
pH 6          pH 8
25%
50%
75%
80%
85%
90%
95%
100%
100
100
0
40
100
100
100
100
100
100
100
100
100
100
100
100
       Figure 11

-------
                             77



                                   MR. TELLIARD:  Is the



coffee out back?  Yes?  Okay.  We are going to take a 15-



minute break, 15, that is, a 10 and a 5...a 15-minute break



to get a cup of coffee, come on back in, and you can listen



to Dow tell us about dirty, dangerous, and deadly dioxin.



(WHEREUPON, a brief recess was taken.)



                                   MR. TELLIARD:  Our next



speaker this morning is Les Lamparski from Dow Chemical.



The agency at the present time is spending $1.7 trigabucks



looking at the impact of dioxins and furans in the



environment, in particular, as it relates to the pulp and



paper industry but also in refining, centralized waste



treaters, and so forth.



     This morning, we have two people speaking on dioxin



analysis.  Les is probably one of the more formal speakers



in the sense that he has been at it longer than most of the



average bears dealing with the analysis of dioxin and



furans, and we are very fortunate to have him this morning.



     So, with no further ado, Les?

-------
                        78
                              MR. LAMPARSKI:  Thanks, Bill.
     Probably one of the few people who have, as Bill said,
who has been involved in this longer than I have is the
other speaker this morning. Tom Tiernan.  So, I haven't been
involved the longest.
     This morning, Terry and I were sitting in the bar
having a cup of coffee, and we decided that this talk was a
little bit too boring the way I had it originally been
configured.  So, we took the slides from two talks and
combined them, just sort of shuffled them up and put them
together.  So, I am going to try to go through these slides.
Don't look at the numbers at the bottom of the slides,
because they jump all over.  Basically, what I am trying to
do is describe the instrument and show some applications at
the same time, so it may be a little bit jumbled around.
     Our work in the measurement of chlorinated dioxins is
driven by the fact that our laboratory at Dow has
responsibility basically for the analysis of chlorinated
dioxins and furans for all of Dow U.S.A.  So, there are a
number of sites across the country that we are responsible
for answering to the various State and EPA regional
regulations.
     So, we have a wide variety of samples that have to be
analyzed for these types of compounds, and we are always

-------
                             79



looking for ways to make the analysis a little bit easier.



And by being easier, we mean faster and cheaper.



     Currently, chlorinated dioxins require quite a



sophisticated cleanup procedure prior to the measurement,



generally by GC mass spec.  These cleanup restrictions tend



to limit the size of studies that can be realistically



performed in an environmental survey.  If you have to spend



a week to clean up a series of 5 or 10 samples, you are



naturally not going to want to invest the time to analyze or



to conduct a survey that would contain 1000 samples.  So,



this is a very strong limitation that is placed on



environmental studies that are undertaken to analyze for



these compounds.



     Because of this, there is always possibility that any



data that is generated in an environmental study...the data



has been compromised in some fashion.  If the samples have



not been homogenized properly, if homogenization is the



method that is chosen to reduce the number of samples that



are analyzed, possibly we could collect 1000 samples and



combine them in sets of 10 in some fashion and only analyze



100.



     Well, if this combination procedure is not truly



indicative of the samples in the environment, we may have



compromised the integrity of the final data.

-------
                             80
     This type of approach may have a very serious effect on
any risk assessment that comes out of a study, whether it is
in a positive fashion and the risk is determined to be
greater than it is or if the risk is falsely determined to
be too low.  Either of those possibilities can cause a
problem.
     We decided that it might be an interesting approach for
an environmental study to focus on specific types of
compounds.  Rather than looking for all of the chlorinated
dioxins and chlorinated furans that are possible, all 210
isomers, let's look specifically at only a couple of the
key...let's call them the most toxic isomers.
     In this case, we could say 2,3,7,8-TCDD and 2,3,7,8-
TCDF.  I will show this structure on a later slide for those
of you who aren't familiar with them.
     By limiting the amount of analytes and limiting the
sensitivity that we are going to look for these compounds,
let's not go down to 1 part per quadrillion or whatever in a
soil sample.  Let's set a realistic limit.  If it is 10
parts per trillion or 1 part per billion, that limit will
determine how large a sample is going to be analyzed to
begin with.
     Naturally, if you have a smaller sample to begin with,
you can use a cleanup...the cleanup procedure will be much
more effective, generally, if it only has to be applied to

-------
                             81
10 mg of sample rather than to 100 grams.  So, it will allow
us to use more efficient methodology and simpler equipment.
     What kind of simpler equipment are we talking about?
We decided to investigate the use of 2-dimensional gas
chromatography followed by a low resolution mass
spectrometer as a detection system.  There were a number of
potential advantages that we saw when we decided to try to
construct this instrument, and they are shown here on this
slide.
     Basically, they allow us to automate the system, and
the changes that had to be made to the system were
compatible with the instrumentation that we currently had
in-house.  Naturally, if somebody has a low resolution mass
spectrometer and all of a sudden the EPA mandates that high
resolution mass spectrometry is the only way that a sample
can be analyzed, you are stuck with a $200,000 boat anchor.
     So, we were looking for ways to extend the capability
of low resolution mass spectrometry and, in fact, in some
cases, to try to use low resolution mass spectrometry as an
alternative to the high resolution mass spectrometers that
are currently in vogue.
     The system that we constructed shown here consists of a
5987 Hewlett-Packard low resolution quadruple mass
spectrometer.  It is a 7-year-old instrument, so it is a
well run instrument.

-------
                             82
     The 2-dimensional system is contained in a 5890 GC
shown here.  It was purchased from a company called
Analytical Controls, Incorporated.(ACI), and the 5890
contains a Dean switch and a cryogenic trap for collecting
the analytes.  It is configured presently with an FID and an
electron capture detector for monitoring pre-column
effluents.
     The whole system is controlled by a single controller
provided by ACI which monitors all of the components.  It
basically monitors all the components in the 5890.  The only
electronic link between the 5890 and the 5880, the GC/MS
portion of the instrument, is by a remote start cable.  That
is the only connection except for the analyte transfer line
which is that large black line going across the top of the
two instruments.
     The analyte transfer connection is an approximately 40-
inch heated line.  The transfer is through a fused silica
tube which can be changed to match the characteristics of
your analyte.  So, if you want to have a coated fused silica
column containing DB-5 or Supelco 2330 or whatever, you can
change that.  Presently, we are using DB-5 because it is
very rugged, and that is also the column that we have on the
pre-column instrument.

-------
                             83
     As you can see, the system is also automatable.  We
have an autosampler which can be and is routinely used, but
it can be removed for manual injections.
     This is the slide that we showed to management to try
to sell them on the subject.  It is very crude, and it is
totally non-scientific, but management bought it, so they
bought the instrument for us.  But basically what happens is
the sample is injected into the prep GC, goes through the
capillary column, a switch is effected, and the analyte can
either go to the detector or the cryo trap where it is
collected by the CO2-cooled trap,  and then  it can be heated,
transferred into the second chromatographic column, and
finally into the mass spectrometer for detection.
     As I said, that was for management.  What really
happens looks something like this.  This is the flow
schematic for a Dean switch that is used by the ACI
controller.  I am going to show you four of these slides
that are essentially the same, and what we are doing is
looking at where the analyte goes during the various stages
of the separation.
     In the initial pre-column separation,  the analyte is
injected into the preparatory capillary column, and
separation is achieved in this column.  Generally, the
analyte goes into the EC detector for detection and
monitoring of effluent.

-------
                             84
     The system contains two basic carrier gas systems.  The
flow controlled system controls the preparatory GC column,
and the pressure controlled system controls the Dean switch
which is the method of switching the analyte flow between
the monitor detector and the analytical column.
     So, at the appropriate time when the analyte is eluting
from the prep column, the Dean switch is switched.  Valve 3
is switched so that flow .goes in the opposite direction and
essentially pushes the analyifces onto the cryogenic trap
which is being cooled by t3ae C02.
     The analyte is trapped out for a given period of time,
and that is pre-determined by analyses of standards.  The
analyte can then be reinjected by stopping the cooling CO2
onto the cryogenic trap., and it is resistively heated in
roughly one minute from —SO to 250 degrees, and the analyte
is deposited into the heated zone through the transfer line
and onto the analytical column where separation can then be
made under the types of conditions that are best for the
analyte and the column that is used for the final
separation.
     At this time during the reinjection onto the analytical
column, the flow on the prep column has stopped.  So, if we
want to collect multiple fractions, it is possible to then
go back, start the flow again on the column, and collect

-------
                             85



another fraction for a second analyte or any number of



analytes.



     After the analyte is injected into the analytical



column, separation is made in the normal fashion, and the



flow on the preparatory column is reversed.  What this does



is elutes off the very heavy boiling garbage that normally



would stick on the front end of the column and cause



problems.



     This is a very real advantage for this type of system.



I will show you later that we are looking at crude extracts



of waste treatment sludge, let's say.  We have analyzed



through the single preparatory column which is a DB-5 column



over 800 injections.  We have made over 800 analyses on this



column without having to change the column, and part of that



advantage comes from the fact that the high boiling



components are eluted backwards off the column during this



phase of the analysis scheme.



     If we look very briefly at the types of control



sequences for various components of the system, basically,



the prep controller...that is the ACI controller that is the



dominant computer in the whole system...it looks for ready



signals from the various other components in the system, the



autoinjector, the prep GC, the mass spec computer, the mass



spec GC, and the mass spectrometer itself.

-------
                             86
     When it gets ready signals from all of those
components, it initiates the start of the run by telling the
autoinjector to make the injection.  Prep GC does the
separation, and, as you can see, over a period of time, all
of these control sequences are brought into play by the prep
computer.  All of this is done.  There is no operator
intervention at all.  Once you load the samples into the
autosampler rack, you can just walk away and forget about
it.
     If we look at some of the data that we have gotten out
of this system, this is 2,3,7,8-TCDD.  We have analyzed this
on the top trace by GC/MS, and this is 40 picograms injected
onto a DB-5 column running under conditions optimum for
detection of that component.  You see the retention time is
about 6.5 minutes.
     The bottom trace is the GC/GC/MS separation on a Lee SB
smectic column which we have chosen at this time to be
useful for determination of 2,3,7,8-TCDD isomer-
specifically.  We have yet to hear Tom's talk, so I can't
really say that we want to use that column, but I am sure
that is one that we are going to be investigating in the
future.
     I didn't want to steal any of his thunder, but I
couldn't resist.

-------
                             87
     As you can see here, the sensitivity of the two traces,
the signal to noise ratio for 40 picograms injected onto the
column are virtually identical.  Now, what this tells us is
that we are getting good transfer through the Dean switch
which is just a series of T's in the gas flow stream, and
also, we are not seeing very drastic effects, if any, of
adsorption onto various components in the switch itself.
     It is predominantly a fused silica system, but there
are places where, if the system is not put together
properly, there could be some stainless steel active sites
which would cause adsorption of TCDD, but we don't see any
of that as evidenced by this chromatograph.
     We then looked at the reproducibility of the system
compared to GC/MS, and using a manual injection without the
autosampler, you can see that the reproducibility over a
period of time...this was 9 runs that were run over a period
of 2 days... is certainly comparable to anything that can be
done by GC/MS for a similar type compound.
     Using the autoinjector where 10 runs were run over a
period of about 20 hours total time, you can see that the
reproducibility is essentially the same.  There is some
evidence that there was a little bit of gradual drift of the
instrument sensitivity which is causing this number to be
probably a little bit larger than it would be if it were a
very short-term experiment, but each analysis takes roughly

-------
                             88
38 minutes, and when we put in blank runs in between, it
took quite a long time to do that precision study.
     We looked at the linearity over a small range of
concentrations for TCDD again, and these are injections from
5 picograms on column to 200 picograms on the first column.
As you can see, the linearity is very good.
     If we increase the concentration range up by a factor
of 10 up to 2000 picograms on column, the linearity does
skew a little bit.  It is biased somewhat high at the higher
concentrations, but this is certainly usable for most
applications.
     If we examined what does a real sample looks like on
this system: here we have a complete electron capture gas
chromatogram of an extract of waste solid from a waste
treatment plant in our company.  This is the crude extract,
a benzene extract, of the sludge out of the treatment plant.
As you can see, by electron capture, there are a lot of
compounds eluting off of this column.
     This is a 30-meter DB-5 column programmed from 135 to
roughly 280.  The fraction of interest...this is supposed to
be highlighted in grey, and I don't know if you can see it
in the back of the room...but we are looking at a roughly
half-minute wide window between 9.5 and 10 minutes.
     A 1-nanogram response with TCDD is shown on the upper
right of the chromatogram.   So, as you can see, some of

-------
                             89
these components that are eluting off are very large
compared to the TCDD that we are supposedly looking for.
     We collected the 9.6-10.1 minute fraction and then put
it onto our analytical column, the Lee smectic column, and
what comes out by GC/MS...by the second GC/MS...is a single
peak with no interfering components.  The amount observed is
5.3 ppb.  The detection limit was set up in this study to be
of the order of .5 ppb.  The regulatory level for this
compound in this matrix was 1 ppb.
     So, we set up the analysis by using a 10 mg sample to
begin with.  We extracted a large sample, took an aloquat
corresponding to 10 mg and injected that directly into the
GC/GC/MS, and we were able to determine the TCDD as shown
there.
     The bottom trace is the internal standard that is added
prior to the analysis.  We use this to monitor recovery
through any cleanup procedures that could be used.  In this
case, we used the internal standard in a typical response
factor calculation.  The internal standard is used to
determine the amount of analyte that is present.
     What does all of this mean?  Why would anybody want to
invest the time?  This instrument costs roughly... the 5890
GC and the controllers cost about $50,000.  This is if you
have a mass spectrometer that you are willing to devote to

-------
                             90
this type of system.  Why would I want to spend $50,000 for
a single type of analysis like this?
     First of all, we should make the point that this isn't
a single type of analysis.  This is a very broad... there are
a lot of applications that are potential for this type of
instrument, not only in the field of trace analysis but also
in product analysis.  If you are looking for a very minor
component in a mixed product or a waste stream from a
product, there is a very good likelihood that this type of
instrument would be applicable for that type of analysis.
     But let's stick to the original topic, determination of
chlorinated dioxins and furans in environmental samples.  We
currently have an environmental study which is being
mandated by the State of Michigan in which we are going to
look at a variety of sediment samples from nearby the plant.
     At present, this study is slated to be 100 samples
large.  Now, if we were to only look for 2,3,7,8-TCDD and
TCDF by our current method which involves a significant
amount of sample preparation prior to GC/MS measurement, we
are looking at a time period of 24 weeks for two people to
do the analysis, and that breaks down to a cost of roughly
$2400 per sample.
     Now, I am sure a lot of you contract labs say that you
can analyze the samples for significantly less than that,

-------
                                  91



     but let's just look at the relative amounts that we re



talking about.



     Using the 2-dimensional LRMS system, we project that



these 100 samples can be done in 3 weeks.  This breaks down



into $240 per sample.  Well, obviously, it won't take too



many studies of this size to pay for that instrument many



times over.



     So, that is why we feel that it will be a useful tool



in an general analytical laboratory in the future.  As I



say...this is the commercial...Analytical Controls is the



company that sells this GC instrument and controller, and



they will be more than happy to take your money.



                                   MR. TELLIARD:  Thank you,



Les.  Any questions?

-------
                             92



               QUESTION AND ANSWER SESSION



                              MS. KLATT:  I am Kelly Klatt



with J&W Scientific.



     The question I do have is, you evidently showed first



the GC BCD slide, the extract, and you had evidently spiked



in the 2,3,7,8-TCDD.  Right?



                              MR. LAMPARSKI:  In the extract



of the sample?



                              MS. KLATT:  Right, right.



                              MR. LAMPARSKI:  Yes.



                              MS. KLATT:  And then you



showed the...



                              MR. LAMPARSKI:  No, that is



not spiked in.  This is naturally occurring TCDD.



                              MS. KLATT:  Okay, right.  And



then you showed this.  Was this sample that you extracted



this from, did it also have the other tetra isomers?



                              MR. LAMPARSKI:  Yes, it does.



                              MS. KLATT:  It does.  So, this



does separate all the other tetra isomers?



                              MR. LAMPARSKI:  Yes.  The Lee



column does separate 2,3,7,8-TCDD from the other TCDD's.  It



doesn't separate the furans.



                              MS. KLATT:  Oh, okay.

-------
                             93



                              MR. LAMPARSKI:  And, to be



truthful, when we started this study, Tom's column wasn't



available yet.



                              MS. KLATT:  Okay, thank you.



                              MR. TELLIARD:  Any other



questions?



                              DR. SCHULTZ:  Bill Schultz



from Eastern Kentucky University.



     I don't suppose you have a trace of what it looks like



with selected ion monitoring from a single column separation



that compares to your BCD?



                              MR. LAMPARSKI:  I don't have



it on an overhead, but I can show you a similar comparison.



                              DR. SCHULTZ:  It seems to me



that the selected ion would give you some separation.



                              MR. LAMPARSKI:  No, it is



worse than a picket fence.  It goes up, and it is just



humpy.



                              MR. TELLIARD:  A humpogram?



                              MR. LAMPARSKI:  Yes.  Well,



this is the saturated one.  It is humptane.



(Laughter.)



                              MR. TELLIARD:  Good morning,



John.

-------
                             94
                         MR. MCQUIRE:  John McQuire from
EPA.
     Have you, in the interest of trying to come up with low
cost...lower cost...low resolution mass spec investigated
any of the work that Don Hunt did a few years back, maybe
10, on negative oxygen CI?  He claimed that you could get
isomer-specific breakdowns and all sorts of things.  I was
wondering if you had done anything with that.
                         MR. LAMPARSKI:  We haven't done
anything with that, but in review of his work, the
separation is not really that clear cut.  You are looking at
fragment ions that are potentially formed in one case and
not in the other case.  If the separation is not such that
the appropriate isomer is separated, then you effect very
little advantage.
     Also, there are generally very real problems with
matrix effects for NCI analyses.  So, if you inject a
standard, you get one response, and if you inject a dirty
sample matrix, you will get something totally different.
     For that reason, we felt that the reliability of that
type of system was not what we were looking for.  Basically,
what we tried to do is, by selling the EPA the concept of
running fewer samples, what we also try to say is we are
going to run fewer samples, but we are going to give you
more reliable data on those fewer samples.

-------
95
                     So,  we are
                              MR. MCQUIRE:  Okay,



understood.  Thank you.



                              MR. LAMPARSKI:



trying to balance things out that way.



     I might add that we have run a number of comparison



studies between the 2-dimensional analysis and the GC/MS



analysis using our full 3-day-long cleanup procedure, and



the results are very comparable.  So, the agreement between



the two systems is very good.  The only difference is the



amount of time that it takes to do the analyses.



                              MR. MCQUIRE:  Okay, thank you.



                              MR. TELLIARD:  Anyone else?



     I certainly like that $240 price, Les.



                              MR. LAMPARSKI:  For you, it



will be a little bit different.



                              MR. TELLIARD:  Yes, Iknow, you



are going to deal.  Right.



     Thank you very much.

-------
                                          96
           DEVELOPMENT OF AN  HRGC-HRGC-LRMS  SYSTEM:
            INSTRUMENT DESIGN  AND  PERFORMANCE  DATA
                         L. L Lamparskl and T. J. Nestrick
                            The Dow Chemical Company
                         Michigan Research and Development
                  Analytical Sciences, Special Analysis, 1602 Building
                           Midland, Michigan 48674  USA
Analytical Controls, Inc., can supply a Hewlett Packard Model  5890 high  resolution gas
chromatograph that is appropriately  configured to  accomplish fully  automated, single-oven,
2-cfimensiona! gas chromatography. As a portion of an extended research program dedicated to
Improving the capabilities of conventional instrumentation, we have developed and assembled
the necessary components to permit the linkage of such a unit to a Hewlett Packard Model 5987A
high  resolution gas chrornatograph-fow resolution  mass spectrometer (HRGC-LRMS).  The
results of this project have produced an instrument that can routinely  conduct fully automated,
dual-oven, 2-dimensiona! HRGC separations in conjunction with LRMS identification and detection
for solutes eluting from the secondary analytical gas chromatograph.  A description of this new
analytical instrument and the techniques required to assemble  it from commercially available
resources will be the subject of our presentation.  Preliminary evaluation and testing of the
operational characteristics of the unit indicate that it may be a valuable tool for measuring a
variety of compounds at trace concentrations in extremely complex matrices.

-------
                     97
  Development of HRGC-HRGC-LRMS
Instrument Design & Perfomance Data
       Lester L. Lamparski & Terry J. Nestrick
              The Dow Chemical Company
            Michigan Research and Development
     Analytical Sciences, Special Analysis Group, 1602 Building
             Midland, Michigan 48674 USA
The Dow Chemical Company
EPA Conference, Norfolk, May 90  SLIDE # 1

-------
                           98
    Current Situation for CDD/CDF Surveys
    Prelim matrix preparations often required.
    Variable extraction procedures for different matrices,
    Specialized cleanup often required for analytes.
    Cleanup simplification = Sophisticated instruments.
    Expensive & time consuming projects.
The Dow Chemical Company
EPA Conference, Norfolk, Hay 90   SLIDE # 2

-------
                      99
       END RESULT for Current Situation
     Limited number of surveys actually performed.
     Limited number of samples examined in a survey.
     Often necessitates sample compositing.
     Sample representativeness often questionable.
     Statistics on resultant survey data less reliable.

    impact on Risk Assessment policies?
The Dow Chemical Company
EPA Conference, Norfolk, May 90   SLIDE # 3

-------
                           100
     New Approach for Regulatory Surveys
     Limit analytes (e.g., toxicological significance).
     Limit sensitivity (e.g., >10 PPT).
     Use larger sample cohorts to improve reliability.
     Use efficient methodology & simpler equipment.
The Dow Chemical Company
EPA Conference, Norfolk, May 90   SLIDE # 4

-------
                        101
               Why GC - GC - MS ?
      Affords increased chromatographic efficiency.
      Automation capability.
      Compatible with current GC-MS equipment.
      Extension of current cleanup capabilities.
      Potential means to reduce /eliminate cleanup
      Broad applicability.
TTie Dow Chemical Company
EPA Conference, Norfolk, May 90   SLIDE # 2

-------
                        102
                  POTENTIAL
       to extend capability of LRMS ?
         to HRMS for trace analysis ?
The Dow Chemical Company
EPA Conference, Norfolk, May 90   SLIDE # 3

-------
                             103
     GC- ANALYTICAL
       HP-5880 (LRMS)
       Heated Transfer Line
       Independent Oven
       Secondary Injector
                                      GC - PREPARATIVE
                                       HP-5890 (FID & EC)
                                       Deans Switching
                                       Cryogenic Trap
                                       ACI Controller
                                       HP-7673AAutoinjector
The Dow Chemical Company
EPA Conference, Norfolk, May 90   SLIDE $ 4

-------
                                       104
                                            Cryogenic Trap
                                                         Heated
                                                         Transfer
                                                         Line
            Preparatory Gas Chromatograph
                                Analytical Gas Chromatograph
The Dow Chemical Company
EPA Conference, Norfolk, May 90     SLIDE # 5

-------
                                105
                  Precolumn Separation
                 Injector
           Flow Controller
        Preparatory
        Capillary
        Column
      • Solutes Path
      EH Carrier Gas
      D No Flow
                     Helium
                      Inlet
                     =4*,,,
                                                  Pressure Regulator

                                                         CQz Inlet
  Analytical Capillary
      Column
The Dow Chemical Company
EPA Conference, Norfolk, May 90    SLIDE 16

-------
                                       106
                        Analyte Trapping
                  Injector
                               Flow Controller
        Preparatory
        Capillary
        Column
      • Solutes Path
      El Carrier Gas
      D No Flow
                                         Hejium
                                          Inlet
Cryogenic

  TraR   n
  «cool»
                       FID
EC
                                               y 9    Pressure Regulator

                                                            "COfe Inlet
                                Trap     LRMS
                               Effluent
                                     Heated
                                  Transfer Line
Analytical Capillary
     Column
The Dow Chemical Company
                   EPA Conference, Norfolk, May 90    SLIDE # 7

-------
                                 107
                     Analyte  Reinjection
                 Injector
           Flow Controller
        Preparatory
        Capillary
        Column
'••* ' l"l





*mt

Is 	

Cryogenic
-f
Trap
«heat»
FID c
//

l»

K
W/L—

^^Jf ^vj—
	 1
Heate<
[Transfer 1
^^
     • Solutes Path
     B Carrier Gas
     D No Flow
EC
                                                   V5
                                        V3
                      Helium
                       Inlet
                      Is^EUII
                                             v 9    Pressure Regulator

                                            jX]      4"" CQz Inlet
                                            NV3
                                                     Trap    LRMS
                                                    Effluent
 Analytical Capillary
     Column
The Dow Chemical Company
EPA Conference, Norfolk, May 90    SLIDE £ 8

-------

                                  108
Analytical Separation &  Precolumn Backflush
                Injector
                          Flow Controller
Preparatory
Capillary
Column
      • Solutes Path
      D Carrier Gas
      D No Flow
                            Effluent    v 2  MW o
                             Filter     Jf  NV2
                                               V5
                                     V3
                     Cryogenic
                       Trap
                      «heat»
                     FID
                                    Helium
                                     Inlet
                                                Pressure Regulator
                                 =X
                                 COz Inlet
                                     V4  NV3
                                                         LRMS
                           Effluent
                                 Heated
                               Transfer Line
EC
                               Analytical Capillary
                                   Column
The Daw Chemical Company
                EPA Conference, Norfofk, May 90    SLIDE * 9

-------
                            109
     GC-GC-MS Control  Sequence for TCDD
           0   3   6   9  12  15  18  21  24  27  30  33   36(min)
Prep Computer
Autoinjector
Prep GC
 Cryo valve
 Cryo trap
 Trap heat
 Backflush
 Foreflush
 Oven cool
 Oven equil
Anal Computer
Analytical GC
LRMS
          Run
          Ready
The Dow Chemical Company
EPA Conference, Norfolk, May 90   SLIDE # 10

-------
                            110
        Chromatographic comparison...
                 ci   ^   a  ^  a
 Ion Intensity
   "A 322
      0%J
       5.0
 6.0
 7.0
     100%
 Ion intensity
  •A 322
      0%J
         GC- GC-MS
       9.0
10.0
11.0
                        m/z = 322
                        40 pg injected
 8.0      9.0   (min)


      m/z = 322
      40 pg injected
12.0
13.0  (min)
77w? Dow Chemical Company
               EPA Conference, Norfolk, May 90   SLIDE # 12

-------
                        Ill
         Manual Injection GC - GC - MS
          TCDD Reproducibility Data for 40pg
     n (number of runs)
     Hg(avg response)
     o M (relative S.D.)
                          m/z 320
23489
 3.5%
              m/z 322
29188
 4.6%
The Dow Chemical Company
  EPA Conference, Norfolk, May 90   SLIDE # 13

-------
                           112
            Autoinjector GC - GC - MS
          TCDD Reproducibility Data for 40pg
      n (number of runs)
      xm (a vg response)
         (relative S.D.)
m/z 322
  10
 7784
  5.5%
m/z 324
  10
 3774
 3.8%
The Dow Chemical Company
  EPA Conference, Norfolk, May 90   SLIDE # 14

-------
                           113
             Autoinjector GC - GC - MS
        Detector Linearity for TCDD (normal range)
    Detector
    Response
       50000
       40000
       30000
       20000
       10000
                 /z 322
                                                 '/z 324
                   40     80     120    160     200
                     Picograms TCDD injected
The Dow Chemical Company
EPA Conference, Norfolk, May 90   SLIDE # 15

-------
                               114
              Autoinjector GC - GC - MS
        Detector Linearity for TCDD (extended range)
Detector
Response


  600 000


  500 000


  400 000


  300 000


  200 000


  100000
                 m
                                                   /z 322
                                                  m
                  '/z 324
                   400    800    1200   1600    2000
                      Picograms TCDD injected
The Dow Chemical Company
EPA Conference, Norfolk, May 90   SLIDE # 16

-------
                          115
        Prep HRGC (ECD) for Waste Solids
                                      Response for ~1 ng

                                      2378-TCDD injected
                                     Trapped

                                     analytes

                                     fraction
i
2
i
4
                  i • i
                     6
I '  I '  M 1
   8     10
 I ' 1  ' I  ' I ' I '  I
12     14     16(min)
TTre Dow Chemical Company
               EPA Conference, Norfolk, May 90   SLIDE # 11

-------
                             116
   Analytical HRGC-LRMS for Waste Solids
                  2378-TCDD = 5.3 PPB
        3000-
        2000-
        1000:
           0
Native
                           m
'/z 322
            8.0    9.0  10.0  11.0  12.0  13.0   14.0  (min)

        1604 Native
        1200
         800 -3
         400-
                            l/z 324
            8.0   9.0  10.0  11.0  12.0  13.0   14.0  (min)
       80 000 -
       60 000 -
       40 000:
       20 000 -i
PCJ
                           m
'/z 334
            8.0   9.0  10.0   11.0  12.0  13.0   14.0  (min)
ThA r>nw (Jhftmlnal Ciamnanv
                  EPA Conference. Norfolk. Mav 90   SLIDE # 12

-------
                     117
Why should I bother with HRGC-HRGC-LRMS?


Dow perspective:    100 samples

                  2378-TCDD & 2378-TCDF
Analytical Method    Analysis Time   -Cost per Sample
Full Cleanup

HRGC-HRGC-LRMS
~ 24 weeks
  "•.." f    \
 f''s'  '  '\
 ~ 3 weeks
*2,400.
OI)
The Daw Chemical Company
     EPA Conference, KotroK,U3y 30  SLIDE* 15

-------
r
                                        118



                                         MR. TELLIARD:  The next



           speaker is going to talk about MAGIC.  I always like that.



                Jim de Haseth is with the University of Georgia.  We



           heard about MAGIC four years ago when Dr. Browner talked



           about MAGIC as it related to particle beam, and today we are



           going to talk about HPLC/FT-IR MAGIC.  Anytime we get MAGIC,



           it is fun.

-------
                 119
DEVELOPMENT OF A MAGIC INTERFACE
            FOR HPLC/FT-IR
James A. de Haseth* and Richard F. Browner1

          Department of Chemistry
         School of Chemical Sciences
            University of Georgia
            Athens, GA 30602
            ^School of Chemistry
        Georgia Institute of Technology
            Atlanta, GA 30332

-------
                                        120
                                       ABSTRACT
       The interfacing of separation methods with selective detectors has been a long sought after



goal.  In some instances, such as gas chromatography (GC) interfaced with spectrometric methods




such as mass spectrometry (MS) or Fourier transform infrared (FT-IR) spectrometry, considerable




success has been achieved. Interfaces with liquid chromatography have been far less successful due




to the difficulty in eliminating the  mobile phase.  In the infrared, liquid mobile phases must be




eliminated as they are opaque at all wavelengths at even short pathlengths. The solutes are generally




in such low concentration that long pathlengths must be used to record their spectra in the presence




of the solvent. Nonetheless, the solvent opacity precludes this option. Solvent elimination has been




possible for high volatility nonpolar mobile phases, yet the majority of liquid chromatography is



accomplished with polar (reverse phase) solvents.  An  interface has  been developed that can




accommodate both normal and reverse phase solvent systems, as well as gradient elution systems.



This interface is the Monodisperse Aerosol Generation Interface for Combining (MAGIC) liquid



chromatography (LC) with Fourier transform infrared (FT-IR) spectrometry.  This method provides



a sensitive interface between  HPLC and FT-IR spectrometry.   Detection limits below 1 ng are



possible, and the use of buffered  solvent systems with the HPLC does not impair the performance



of the interface.

-------
                                         121
                                    INTRODUCTION








       The increased identifying power that results from coupling separations techniques with




selective detectors has been dramatic.  The value of such procedures arises because the process of




separation of a mixture into its components does not in itself identify the chemical or structural




nature of the compounds present.  Such vital information is only provided by a detector that is




capable of responding in some unique manner to the chemical nature of the separated compounds.




       Depending on  the  perspective of the researcher,  the development of  such  analytical




techniques, loosely known as "hybridized techniques", may be seen either as a means of improving



the quality of analytical detectors for chromatography, or of simplifying clean-up procedures for




sample  introduction  to the detector of choice.  In the recent past, when many of the selective




detectors available commercially were extremely expensive, the second approach was prevalent. The




advent of powerful but inexpensive detectors for mass spectrometry (MS) and Fourier transform



infrared (FT-IR) spectrometry, for example, has radically changed the situation.








                              GC VERSUS LC SEPARATIONS
       The separation and  identification  of volatile components of  materials is generally



 accomplished by interfacing gas chromatography (GC) with sophisticated and highly selective





                                             1

-------
                                        122



physical analysis detectors. The most notable physical analysis techniques used for this purpose are



mass spectrometry (MS) and more recently, Fourier transform infrared (FT-IR) spectrometry. Both



of these techniques provide information about an analyte's molecular structure; mass spectrometry



provides structural information based  on molecular fragmentation  patterns,  whereas FT-IR



spectrometry provides characteristic functional group and isomeric information. A key factor is that




the two techniques provide information that is very complementary in nature. By interfacing MS



or FT-IR spectrometry with GC, either individually (i.e. GC/MS or GC/FT-IR spectrometry), or




in combination (GC/FT-IR/MS), a wealth of information for the identification of gas chromato-



grapnic eluites results.  Appropriate combination of the data from both techniques can give rise to




a substantial increase in the identifying power available for unknown  compounds.  For example,




work with combined systems has shown that a much greater success rate is obtained when spectral




data from FT-IR and MS are used jointly, than when either technique is used alone (1).



       The types of molecules that may be successfully separated by gas chromatography, while




extensive, is nevertheless severely limited by the dual problems of compound volatility and thermal




lability.  While many procedures, such as derivitization, may help to circumvent these problems,




there is clearly a limit to the molecular weight and the polarity of molecules that may be separated



by gas chromatography.  The only practical alternatives currently available for separation of high




molecular weight and high polarity molecules are liquid chromatography (LC) and supercritical fluid




chromatography (SFC).  While SFC clearly has great potential for a large class of compounds of



intermediate molecular weight and low or intermediate polarity, effective separation procedures are



not yet available using SFC for either very  large molecules  (e.g. > 2000  daltons) or very polar



molecules.  Many compounds  of great biological and environmental importance are therefore not



presently amenable to SFC separation. SFC has an additional problem resulting from the limited

-------
                                          123





sample mass that can be injected on-column.  Consequently, while absolute detection limits may be



in the picogram range, concentration detection limits may be inadequate for trace analysis.



       Interfacing liquid chromatography with MS and FT-IR detectors is a much more demanding




problem than that experienced with interfacing GC with these  detectors.  This results from the



much greater difficulty  involved with removing the LC mobile  phase. With GC separations, the




mobile phase, which is often helium, presents little difficulty for the detector.  For example,  in




GC/MS analysis with packed columns an enrichment device, such as a jet separator, is used.  With




low-flow capillary columns, the pumping capacity of the mass spectrometer itself provides sufficient




enrichment.  By these means, the carrier gas is reduced to a level that does  not influence the




operation of the mass spectrometer.  In GC/FT-IR spectrometry, no mobile phase elimination is




necessary, as  the carrier gases  typically used are optically transparent  in this region of the




electromagnetic spectrum. On the other hand, liquid chromatographic mobile phases are neither so




readily eliminated nor are they optically transparent in the infrared region. As a consequence, there




are currently no commercially available or widely accepted methods for directly interfacing LC and



FT-IR spectrometry.




       Two basic research approaches to LC/FT-IR interfacing have been followed to date:  (1)




effluent flow-through systems and (2)  solvent elimination procedures.








         SOLVENT CONSIDERATIONS FOR DIRECT LC/FT-IR MEASUREMENTS
     If a liquid chromatographic eluite is sufficiently concentrated, its infrared spectrum can still




 be measured even in the presence  of LC solvent.  No solvent is totally infrared transparent,




 however, even in the regions where there is no apparent infrared absorption band.  The object of




 the LC/FT-IR spectrometric experiment is therefore to record the spectrum of the eluite and ignore

-------
                                         124

or eliminate  the spectrum of the solvent.  In infrared spectrometry, as well as in many other


absorption spectrometries, the spectrum of a  solute is recorded by ratioing a solution spectrum


against the spectrum of the pure solvent. This operation is valid only when the absorbance of the


solute is relatively high and the absorbance of  the solvent low. If the solvent absorbance becomes


sufficiently great, the spectrum  is opaque  in the region of the absorption band and no solute

                                 o/
absorbance can be measured. If the concentration of the solute is too low,  the absorbance due to


the solute will be in the noise level of the measurement, and again a spectrum cannot be recorded.


In normal practice, spectrometric absorbance can be controlled by pathlength. In an ideal situation


the solvent should absorb no more than approximately  37% of the infrared  radiation.  In this case


an optimum signal-to-noise ratio can be attained (2).


       Unfortunately typical liquid chromatographic solvents do not absorb uniformly across the


infrared region, and it is impossible to design an ideal cell for a complex absorber.  A compromise


must thus be made, and generally pathlengths  are chosen so that only a few bands are opaque and


most of the spectrum is transparent. The relatively transparent regions of the solvent spectrum form


windows in which the signal-to-noise ratio (SNR) is sufficiently high that a spectrum of the solute


can  be  recorded.  The pathlengths vary depending upon the solvent (3).  For example, a weak


infrared absorber such as carbon tetrachloride can be used with pathlengths of up to 1 mm. Very


strong absorbers, such as water,  permit only very short pathlengths, up to about 0.1  mm.  Other


solvents may fall into intermediate regions.


        In a  dynamic system, such as in LC/FT-IR spectrometry, the interface cell should be


designed for the maximum SNR of the eluite.  At a first approximation this is when the entire


analyte band is trapped momentarily in the interface cell. In this situation, the maximum possible


analyte mass is measured spectrometrically,  and so highest sensitivity will be achieved.   The


practical problem this approach gives rise to is that of ensuring the chromatographic resolution is

-------
                                          125





not compromised.  If the volume of the interface cell is too great, it is possible for two chroma-




tographically resolved eluites to be in the cell at same time. To avoid this problem, the volume of



the interface cell must be the full  width at  half height volume of the eluite peak  (4).   An




appropriate diameter for an LC/FT-ER cell is on the order of about 4 mm, based on optical and



chromatographic flow considerations. In a conventional HPLC experiment (column i.d. of 4.6 mm)




a typical narrow peak is on the order of 250 /*L.  An ideal cell of 4 mm diameter (12.5 mm2 area)




would then have a pathlength of 20 mm. This is more than an order of magnitude longer than the




maximum cell pathlength for the weakest infrared absorbing solvent.  Cells of larger diameter




provide an ideal shorter pathlength  at the cost  of sensitivity;  smaller diameter  cells  are more




sensitive because the ideal pathlength is longer, but the discrepancy between ideality and practice



is greater.




       LC/FT-ER spectrpmetry, therefore, presents the researcher with an interesting dichotomy:




cell pathlength must be long for maximum sensitivity,  but it must be short to provide adequate




solvent windows to record solute spectra.  As long as the chromatographic eluite is sufficiently



concentrated, then short pathlength cells can be employed successfully. Where solute concentrations




are  extremely low,  cell pathlengths  are  in practice  too short to record spectra  at useful



signal-to-noise ratios.








                          FLOW-THROUGH LC/FT-IR SYSTEMS
        The earliest LC/FT-ER spectrometry systems used flow-through cells (5-8). These prototype




 systems suffered from severe problems, and were not really practical.  Two related areas where




 flow-through cells have proved to be of value, though, are those of gel-permeation chromatography




 (GPC)  and size-exclusion chromatography (SEC).  In these chromatographic  techniques, where

-------
                                       126
components are separated on the basis of molecular size rather than by polarity or adsorption

characteristics, chlorinated solvents can be used, that makes GPC/FT-IR spectrometry into a

powerful technique.  Chlorinated solvents, for the most  part, do not absorb strongly across the

infrared spectrum except in the C-C1 stretch region and possibly in the C-H stretch region. These

regions are opaque even at short pathlengths. Vidrine (3) compiled a series of spectra of common

LC solvents at various pathlengths and showed that many of the solvents that are useful for GPC

can be used with reasonable pathlengths. Vidrine showed  GPC/FT-IR spectra of the separation of

poly(butyl acrylate) and polystyrene on a 100 A /i-Styragel column using THF (tetrahydrofuran) as
the solvent. Although THF is far from an ideal solvent because it is not transparent, some spectral

information was still recorded in the flow-through system. The regions 3000 to 2800 cm*1 and 1300

to 800 cm"1 are obliterated due to THF absorptions. Had a more polar solvent been used, however,

much less spectral information would have been recovered.

       Flow-through systems are not restricted solely to GPC, and this technology has been applied

to both normal-phase (NP-) and reverse-phase  (RP-)  HPLC.  Most of  the  solvents used in

NP-HPLC  are less polar, and  hence absorb  IR  radiation less strongly  than THF, but overall

performance is often not as good as in GPC/FT-IR spectrometry. In GPC, chromatographic sample

loadings can be quite high, on the order of mg, but typical HPLC 4.6 mm i.d. silica gel columns

have capacities of only 2 to 50 jig. The concentration of most NP-HPLC eluites  is on the order of
200 ppm.  These low sample loadings therefore require  long cell pathlengths in order to record

spectra with good SNR, but this contradicts the  requirement for short pathlengths dictated by the

solvent system.

       A series of fine studies involving flow cells with NP-HPLC has been completed under the

direction of Taylor (9-16). In the earliest of these studies  Johnson  and Taylor (9) compared the use

of semi-preparative, analytical and microbore columns for  NP-HPLC/FT-IR spectrometry. In order

-------
                                         127
to improve the transmission of the ceil and reduce the number of opaque regions, Freon-113 was
used as the mobile phase.  Even so, it was necessary to use a 0.2 mm pathlength cell to minimize
the number of opaque regions. These workers showed that microbore HPLC (/aHPLC) columns were
practical for flow-through cell FT-IR spectrometry. This was a consequence of the smaller volumes
of the microbore eluite peaks compared to the analytical or semi-preparative peaks.
       In an  analytical column the typical eluite peak volume is at least 250 pL, and a 0.2 mm x
4 mm diameter cell has a volume of only 2.5 pL.  This means that no more than 1% of the eluite
peak will  be  in the ceil at any given time. As the volume of the eluite peak decreases, a higher
percentage of the eiuite will be present in the cell; but this is offset by the reduced sample loading.
Nevertheless, the tradeoff is in  favor of the microbore column, hence, high SNR spectra can be
measured in  the transparent regions of the solvent.  In later studies, Brown and  Taylor studied
material of intermediate polarity in coal-derived process solvents, using analytical NP-HPLC
columns (10)  and also compared analytical and microbore columns (11).  Both studies used deuter-
ated chloroform as the mobile phase. CDC13 is transparent in the C-H stretch region of the infrared
spectrum, and hence has more useful windows than chloroform.
       Taylor and coworkers have investigated many other aspects of NP-HPLC/FT-IR spectro-
metry (12-15).  Coal derived products  have  been investigated:  on a polar amino cyano (PAC)
analytical column using a 98:2 CDCl^CH^CN mobile phase (12), and, on an amino (NH2) bonded
microbore column using  a 70:30 CDCl^CClj + 0.02%  triethylamine mobile phase.  Amateis and
Taylor (14) investigated the optimum experimental conditions for NP-HPLC/FT-IR spectrometry,
including spectrometric signal-averaging and flow cell pathlength.  Column overload and system
detection limits were also investigated. Brown, Amateis and Taylor (15) determined detection limits
for phenols and amines in NP-/iHPLC/FT-IR spectrometric systems. These detection limits ranged
from 300 ng  to 12 /»g, depending upon the eluite.  These figures reflect injected quantities, not the
quantities in the  ceil.  One interesting development was a cylindrical  bore  cell for  flow cell

-------
                                        128
/jHPLC/FT-ER spectrometry. This has been named a zero dead volume (ZDV) cell (16).  This cell
provided a, detection limit of less than SO ng for 2,6-ditert-butylphenol  measured at  the O-H
stretching frequency.
       This cell, which is cylindrical, was designed to give a wide dynamic range of solventsolute
ratios. Unfortunately, the absorbance across the spectrum is then, by necessity, nonlinear, although
some bands will exhibit linear behavior. Other workers have investigated flow cells for LC/FT-IR
spectrometry involving normal  phase systems (17-19). An unusual set of studies has involved
interfacing a high speed planet coil countercurrent chromatograph (CCC) to an FT-IR spectrometer
(19,20). The CCC is a liquid-liquid chromatograph that has no stationary support, and hence is able
to operate at  very high sample loadings.  The eluites are then present  in  the system at high
solute-to-solvent ratios, making  flow cell LC/FT-IR spectrometry feasible.  Regrettably, the
chromatographic efficiency of this system is poor.
       A few studies have  involved reverse phase systems (21-23).  These  systems have fewer
spectral windows than the normal phase  systems and analysis is severely  restricted.  Even when
deuterated mobile phases are used, the number of useful spectral windows is small. Although much
has been accomplished with flow-through LC/FT-IR spectrometric systems, these designs are hardly
applicable as a universal approach.  Reverse phase systems are almost totally excluded  from this
technology, and normal phase HPLC systems operate under the severe handicaps of poor solvent
transmission and very low eluite concentrations.

                            SOLVENT-ELIMINATION SYSTEMS
        As stated above, the problem with flow-through systems  is that the spectrum is always
 obscured by absorption bands of the mobile phase. To avoid this problem the mobile phase, or
                                             8

-------
                                          129

solvent, must be removed to leave  the  eluite for  spectrometric study.   The first attempts at
eliminating the solvent were applied to NP-HPLC systems and involved a two-step process (24,25).
The first step was to concentrate the eluite by about a factor of ten by evaporating some of the
solvent in a heated tube. The second step was to deposit the concentrated eluite onto a powdered
potassium chloride (KC1) substrate where the remainder of the solvent was evaporated. Spectra were
recorded by diffuse reflectance (DR) infrared spectrometry. DR spectrometry is well-suited to the
measurement of small quantities of a sample; but this two-step process  is time-consuming and
necessitates long delays  between elution and recording of the spectrum. This system has also been
adapted to microbore columns (26).
       One major disadvantage of the DR solvent elimination technique is that polar (reverse phase)
solvents will dissolve the K.C1 DR substrate.  A remedy for this shortcoming is to extract the eluite
into a nonpolar solvent for deposition onto the KC1 powder (27).  This has been achieved  in an
extraction coil by mixing methylene chloride with reverse phase HPLC effluents, then physically
separating the polar and nonpolar phases (aqueous and CH2C12) in a separation tee. The nonpolar
phase is passed to a concentrator for deposition on a DR K.C1 substrate, as in the normal phase
HPLC system. This has the disadvantage  that an additional, low extraction efficiency, step is added
to the process. An alternate method is to  spray the HPLC effluent onto an  insoluble substrate, such
as industrial-grade diamond  powder, using an ultrasonic nebulizer (28).
        In the case of an aqueous mobile phase, the water can be removed by reaction with 2,2'-di-
methoxypropane (DMP), that produces methanol and acetone when the reaction  is acid catalyzed
(29). The methanol, acetone, and excess DMP can be deposited on a DR  substrate along with the
eluite. All components, except the eluite, can then be evaporated readily. This is an elegant system,
but it is restricted to simple  mobile phases, such as  H2O:CH,OH and H2O:CHSCN.  If nonvolatile

-------
                                         130




ionic modifiers are added to the mobile phase, these modifiers can have concentrations exceeding




that of the eluite, and hence dominate the spectra.



       A different solvent elimination method is the buffer memory technique developed by Jinno




(30-33).  In this technique the effluent from a normal phase microbore HPLC column is deposited



onto a slowly moving KBr plate. The solvent is removed by evaporation, and spectra of the eluites




are recorded by transmission spectrometry. This technique, like the diffuse reflectance technique,



is essentially an off-line method that cannot be used easily with polar mobile phases.



       Other methods have been devised to remove the solvent from HPLC systems and deposit the



solute for infrared spectral interrogation. Gagel and Biemann (34,35) have developed a continuous




recording reflection-absorption interface.  The interface involves taking the effluent stream from



a chromatograph and mixing it with nitrogen gas. The nitrogen gas disperses the liquid effluent and




sprays it onto a reflective surface.  The nitrogen then evaporates the solvent and leaves the solute




as a residue on the surface.  The solutes are deposited on a spiral track  on a reflective surface.



Spectra of the solutes are obtained as the track is moved into an infrared beam and the spectra are




obtained as reflection-absorption spectra.  This system is capable of eliminating the mobile phase




when it contains up  to 55% water (with methanol), but in gradient elution systems the  temperature



of the nebulizing  gas had to be computer controlled to  evaporate the solvent as the mobile phase



composition changed.




        Griffiths  has  developed a  similar interface, except that the spectra are  collected as




transmission spectra with an infrared microscope (36). The nebulizer for this apparatus is similar




to the one reported by Gagel and Biemann, but the deposition is done onto an infrared transparent




substrate, such as zinc selenide (ZnSe).  The advantage to this system is that Griffiths  realized that



the ultimate detection limits will be achieved by concentrating the solute into a minimum cross-




sectional area, then focussing the infrared beam onto that area.  The general system has been






                                             10

-------
                                         131




described (37,38), but these references reported only normal phase systems. Griffiths has shown




one spectrum of methyl violet from  100% water using his HPLC nebulizer when the system is




evacuated as opposed to the normal ambient pressure system (39).




       All  of the above solvent-elimination methods suffer from the need for complex steps to




effect solvent removal or reverse phase systems cannot be accommodated on a routine basis. These




solvent elimination steps are sometimes time-consuming and are not always very efficient.  As with




the flow-through systems, these solvent-elimination techniques  work best with normal phase




systems.








                             MAGIC-LC/FT-IR INTERFACE
       The Monodisperse Aerosol  Generation  Interface for  Combining (MAGIC)  Liquid




Chromatography (LC) with Fourier Transform Infrared (FT-IR) spectrometry provides an alternate




interface that solves many of the problems  exhibited by the  above-mentioned interfaces.  The



MAGIC interface, in its generic form, is a transport based device that takes LC column effluent,




removes the solvent with high efficiency, and passes the desolvated solute to a detector.  As this




device  is a high  efficiency desolvating interface, it has the property most  desirable for an



LC/FT-IR spectrometric interface.




        MAGIC-LC has three primary components: (1) a monodisperse aerosol generator (MAG) (2)




 an atmospheric, ambient temperature, solvent evaporation chamber and (3) a momentum-based




 particle enrichment separator.  A schematic of the MAGIC-LC/FT-IR spectrometric interface is




 shown  in  Figure  1.  The MAGIC-LC device has already been described  in some detail in  the




 literature, in  its application to mass spectrometry (40-42). Basically, the MAG is a device that




 converts the effluent stream into highly uniform droplets. This is accomplished by pumping the






                                            11

-------
                                         132
effluent through a fine bore fused silica capillary tube, typically 25 pm in internal diameter. The
jet formed at the end of the fused silica tube is unstable and forms an aerosol of uniform droplets,
as originally modelled mathematically by Rayleigh (43). The droplets are dispersed by a gas sprayed
perpendicular to the jet, to prevent the droplets colliding and forming a polydisperse aerosol.  A
detailed schematic of the MAG has been given elsewhere (40).
       The dispersed droplets flow towards a nozzle at the opposite end of the desolvation chamber
to the MAG.  The pressure  in the desolvation chamber is near atmospheric and the pressure is
considerably reduced on the other side of the nozzle. The fact that the droplets are of uniform size,
and that they are rather small (on the order of 50 jum in  diameter), the rate of evaporation of the
solvent is high.  As the volume of the droplets decreases the relative surface area increases, and
hence the evaporation of the droplet is rapid.  The desolvation chamber serves the purpose of
evaporating the solvent and passing the remaining solute, in droplet form, into the momentum-based
separator.  Both solute droplets  and solvent  vapor pass into the separator, but only the solute
droplets have  sufficient momentum to pass through the  separator to a collection plate where  the
solute droplets  are deposited.    The momentum-based  separator has  been discussed in detail
previously (42).
        The collection plate  is  translated during deposition so that each  eluite is deposited in a
unique area on the plate.  In  this  way each eluite is separated spatially for infrared analysis.  Once
the HPLC separation is  complete,  the  deposition plate is removed and placed in  an  infrared
spectrometer. Typically a beam condenser is used to focus the beam onto  the eluite  deposition. A
spectrum is collected by transmission spectrometry as the collection plate is an infrared transparent
window, such as KBr or KC1.  An alternate approach is to use an infrared microscope instead of
a beam condenser. An infrared microscope will focus a beam to a much smaller area than a beam
condenser, and as long as an  eluite deposition can be constrained to a small area, for example
                                             12

-------
                                          133
100 ftm in diameter, the microscope will give higher sensitivity than the beam condenser.  The size
of the deposition is dependent upon the quantity of component injected and the configuration of
the MAGIC device.  Depositions of approximately 100 ftm have been achieved.
       The MAGIC  device as used for LC/FT-IR spectrometry differs somewhat from the device
used for LC/MS.  These differences are to accommodate changes in the aerosol to produce good
depositions for FT-IR spectrometry, as opposed to producing an easily vaporized solute stream for
mass spectrometry. The aerosol considerations for different spectrometric techniques are described
in detail elsewhere (44).
       Spectra have  been collected successfully using the MAGIC-LC/FT-IR spectrometric device
for a variety of solvent systems and at  varying concentrations of analytes (45,46).   Figure  2
illustrates a series of spectra of erythrosin B from different mobile phase compositions.  Figure 2a
is the erythrosin B from 100% methanol,  in 2b the mobile phase is  80:20 methanol:water, 2c it  is
60:40 methanol:water, and Figure 2d shows the spectrum when the mobile phase is 100% water. All
four spectra are good representations of erythrosin B and match the reference spectrum closely.
Figures 2a-c show some methanol trapped in the eluite depositions. This is residual methanol after
the solvent elimination process. In the original solutions of erythrosin B the solute concentration
was only a few  parts per thousand, and after desolvation the solvent concentration is only a few
percent of the solute concentration. Clearly the desolvation is  not 100% efficient for some solvents,
yet it is  close to that efficiency. It is interesting to note that  the spectra from 100% water mobile
phase  show no  trapped solvent. The reasons why some solvents are trapped are currently being
investigated in this laboratory? as well as methods to alleviate the problem. One immediate solution
is to remove the spectrum of the trapped  solvent by spectral subtraction.
        MAGIC-LC/FT-IR spectrometry can also work well with pure normal phase separations.
 It can be argued that 100% methanol is not a normal phase solvent.  Figure 3 shows the spectrum
                                             13

-------
                                         134




of anthracene as deposited from 100% hexane. There is no hexane residue apparent in the spectrum.



Figure 4 is a spectrum of caffeine from 100% water, a reverse phase solvent, and again there is no




residual solvent apparent.



       All the spectra shown thus far have been of rather large samples, on the order of a few tens



of micrograms per  injection.   This  is typically greater than standard  injections  for HPLC




separations, and detection limits in the tens of nanograms or lower are  desirable.  The  limiting



factor for the detection of eluites by the MAGIC-LC/FT-IR spectrometric  device is not in the




solvent elimination step, but in the optics for the spectrometry. Figure 5 illustrates a spectrum of




100 ng of methyl red (injected) and analyzed using an infrared microscope. The spectrum is clearly




identifiable, despite the interference of atmospheric water vapor.  The baseline is skewed due to




dispersion effects of the deposition; that is, the deposition is too thick. This spectrum was from an




earlier version of the MAGIC-LC/FT-IR device that  was rather  inefficient.  Only five to ten




percent of  the injected eluite was  recovered  at  the deposition plate.  Current models of the




MAGIC-LC/FT-IR device appear to have efficiencies of 30%, and further improvements are



anticipated. With microscopic optics instead of beam condensing optics, detection limits of less than



one nanogram (injected) can be projected.




       The interface functions well in the presence of  buffered solvents (47).  Even with buffer




concentrations as high as IF, spectra of the chromatographic eluites can be collected. Figure 6



shows the spectrum of caffeine deposited when the solvent contained IF KHP buffer.  Obviously



the spectrum of the caffeine contains extra bands, and these are attributable to the buffer. As the



buffer is involatile, it is deposited continuously.  A spectrum of the buffer  without the eluite is




measured and subtracted from the spectrum of  the mixture.  In the upper half of  Figure 7, a



spectrum of the caffeine minus the KHP buffer is shown. A reference spectrum of caffeine is
                                             14

-------
                                          135





shown in the lower half of the figure. The spectral differences are due to the protonation of the




caffeine.



       The only present shortcoming of the MAGIC-LC/FT-IR spectrometric device is that it is




currently an "off-line" technique. Largely the mechanics of making it an "on-line" technique are



only engineering changes,  and systems previously devised for LC/FT-IR interfaces could be




implemented (for example,  those given  in references 34-39).  Clearly, MAGIC-LC/FT-IR




spectrometry is  a method  that holds great  promise  for the interfacing  of  HPLC with FT-IR




spectromeftry.  This is the first device that  has been used successfully for both normal and reverse




phase separations, and also the first to work with buffered reverse phase  systems (47).  Both




KH3PO4 (potassium dihydrogen phosphate) and KHP (potassium hydrogen phthalate) buffers were




used successfully. Effluent flow rates of  up to 1 mL/m can be accommodated by the interface.




One of the most beneficial aspects of the MAGIC-LC/FT-IR device is that the solvent elimination




is done at ambient temperature; no  heating of the effluent stream or of the desolvating  gas is




necessary. Not only can changes in the mobile phase composition be handled without changing the



desolvation  parameters, thermally labile components can be analyzed with no  concern for thermal



degradation.








                                 ACKNOWLEDGMENTS








        This work would not have been possible without the collaboration of Professor Richard F.



Browner, who originally developed  the MAGIC device.  This collaboration has led  to a joint




research project under the  direction of J. A. de Haseth and R. F. Browner, and the support of the



National Institutes of Health, under grant  number 1 R01 GM40715-01 is greatly appreciated.
                                            15

-------
                                        136
LITERATURE CITED
1. Wilkins, C. L.; Giss, G. N.; White, R. L.; Brissey, G. M.; Onyiriuka, E. C. Anal. Chem., 1982, 54,
2260-2204.

2. This is a well-documented phenomenon, see, for example, Christian, G. D. Analytical Chemistry,
3rd Ed., Wiley, New York (1980), p. 400.

3. Vidrine, D. W. "Liquid Chromatography Detection Using FT-IR", in Fourier Transform Infrared
Spectroscopy: Applications to Chemical Systems, Vol. 2, Ferraro, J. R.; Basile, L. J., Eds., Academic
Press, New York (1970), Chapter 4.

4. Griffiths, P. R. AppL Spectrosc., 1977, 31, 284-288.

5. Kizer, K. L.; Mantz, A. W.; Bonar, L. C. Amer. Lab.,  1975, 7(5), 85-97.

6. Vidrine, D. W.; Mattson, D. R. AppL Spectrosc.. 1978, 32, 502-506.

7. Vidrine, D. W. /. Chromatogr. Sci.,  1979, 17, 477-482.

8. Shafer, K. H.; Lucas, S. V.; Jakobsen, R. J. /. Chromatogr. Sci.. 1979,  17, 464-470.
i.
9. Johnson, C. C.; Taylor, L. T. Anal. Chem.. 1983, 55, 436-441.

10. Brown, R. S.; Taylor, L. T. Anal. Chem., 1983, 55, 723-730.

11. Brown, R. S.; Taylor, L. T. Anal. Chem., 1983, 55, 1492-1497.

12. Amateis, P. G.; Taylor,  L. T. Chromatographia.  1984, 18, 175-182.

13. Amateis, P. G.; Taylor,  L. T. Anal. Chem., 1984, 56, 966-971.

14. Amateis, P. G4 Taylor,  L. T. LC Mag., 1984, 2, 854-857.

15. Brown, R. S.; Amateis,  P. G.; Taylor, L. T. Chromatographia,  1984, 18, 396-400.

16. Johnson, C. C; Taylor,  L. T. Anal. Chem., 1984, 56, 2642-2647.

17. Combellas, C.; Bayart, H.; Jasse, B.; Caude, M.; Rosset, R. J. Chromatogr., 1983, 259, 211-225.

 18. Combellas, C;B Bayart, H.; Jasse, B^  Rosset, R. Analusis, 1985, 11, 225-233.

 19. Romanach, R. J^ de Haseth, J. A^ Ito, Y. J. Liq. Chromatogr., 1985, 8, 2209-2219.
                                            16

-------
                                          137

20. Romanach, R. J.; de Haseth, J. A. /. Liq. Chromatogr., 1988, 11, 133-152.

21. Jinno, K^ Fujimoto, C.; Uematsu, G. Amer. Lab.,  1984, 16(2), 39*45.

22. Fujimoto, C.; Uematsu, G.; Jinno, K. Chromatographia, 1985, 20, 112-116.

23. Tartar, A.; Huvenne, J. P.; Gras, H.; Sergheraert, C. /. Chromatogr., 1984, 298, 521-524.

24. Kuehl, D.; Griffiths, P.R. /. Chromatogr. Set., 1979, 17, 471-476.

25. Kuehl, D. T.; Griffiths, P. R. Anal. Chem., 1980, 52, 1394-1399.

26. Conroy, C. M.; Griffiths, P. R.; Jinno, K. Anal. Chem., 1985, 57, 822-825.

27. Conroy, C. M.; Duff, P. J.; Griffiths, P. R.; Azarraga, L. V. Anal. Chem., 1984, 56, 2636-2642.

28. Castles, M. A.; Azarraga, L. V.; Carreira, L. A. Appl. Spectrosc., 1985, 40, 673-680.

29. Kalasinsky, K. S.; Smith, J. A. S.; Kalasinsky, V. F. Anal. Chem., 1985, 57, 1969-1974.

30. Jinno, K.; Fujimoto, C. /. High Res. Chromatogr.; Chromatogr. Common., 1981, 4, 532-533.

31. Jinno, K.; Fujimoto, C^ Hirata, Y. Appl. Spectrosc., 1982, 36, 67-69.

32. Jinno, K.; Fujimoto, C.; Ishii, D. /. Chromatogr., 1982, 239, 279-286.

33. Jinno, K. Spectrosc. Lett., 1981, 14, 659-663.

34. Gagel, J. J.; Biemann, K.  Anal. Chem., 1986, 58, 2184-2189.

35. Gagel, J. J.; Biemann, K.  Anal. Chem., 1987, 59, 1266-1272.

36. Griffiths, P. R.; Fraser, D. J. J. "HPLC/FT-IR with Continuous Solvent Elimination and Eluate
Deposition",  ACS 39th Annual Summer Symposium  on  Analytical  Chemistry,  "Chromato-
graphic/Spectroscopic Combinations", Salt Lake City, Utah, June  1986.

37. Griffiths, P. R.; Pentoney, S. L., Jr.; Pariente, G. L.; Norton, K. L. Mikrochim. Acta [Wein],
 1987, 3, 47-62.

38. Fraser, D. J. J4 Norton, K. L.; Griffiths, P. R. "HPLC/FT-IR Measurements by Transmission,
Reflection-Absorption, and  Diffuse Reflectance Microscopy", in Infrared Microspectroscopy:
Theory and Applications, Messerschmidt, R. G.; Harthcock, M. A., Eds., Marcel Dekker, New York
(1988), Chapter 14.

39.  Griffiths,  P.  R. "Recent  Advances  in FT-ER  Hyphenated Techniques", 40th Pittsburgh
Conference on  Analytical Chemistry and  Applied Spectroscopy,  Atlanta, Georgia, March 1989,
paper 623.
                                            17

-------
                                       138


40. Wffloughby, R. C; Browner. R. F. Anal. Chem.. 1984, 56, 2626-2631.


41. Browner, R. F.; Winkler, P. C^ Perkins, D. D.; Abbey, L. E. Microchem. J., 1986, 34, 15-24.


42. Winkler, P. C.; Perkins, D. D.; Williams, W. K.; Browner, R. F. Anal. Chem., 1988,60, 489-493.


43. Rayleigh, J. W. S. Proc. London Math. Sac.. 1878, 10, 4-17.


44. Browser, R. F. Microchem. J., 1989, 40, 4-29.


45. Robertson, R. M.; de Haseth, J. A.; Browner, R. F. Mikrochim. Acta [Weinj, 1988, 2, 199-202.

                                   •*
46. Robertson, R. M.; de Haseth, J. A.; Kirk, J. D.; Browner, R. F. Appl. Spectrosc., 1988,42,1365-
1368.


47. Robertson, R. M^ de Haseth, J. A.; Browner, R. F. Appl. Spectrosc., 1990, 44, 8-13.
                                            18

-------
                                         139
FIGURE CAPTIONS

1.  Schematic diagram of the MAGIC-LC/FT-ER. spectrometric interface showing the three major
components of the MAGIC device and the placement of the deposition plate for collection of the
eluites from an HPLC separation.

2.  MAGIC-LC/FT-IR spectra of SO p% injections of erythrosin B using various different mobile
phase compositions: (a) 100% methanol, (b) 80:20 methanokwater, (c) 60:40 methanokwater, and (d)
100% water.  (Reproduced from Ref. (46).)

3. Spectrum of deposited anthracene using 100% hexane as the mobile phase.

4. Caffeine as deposited from the MAGIC-LC/FT-IR device, when 100% water was used as the
mobile phase.

5. Spectrum of 100 ng of methyl red (injected) from a 100% methanol mobile phase.  (Reproduced
from Ref. (45).)

6. Spectrum of caffeine and IF K.HP buffer. (Reproduced from Ref. (47).)

7. Spectrum of caffeine (upper) after spectral subtraction of the KHP buffer spectrum. A reference
spectrum of caffeine is shown below.  (Reproduced from Ref. (47).)
                                           19

-------
He
DISPERSION
GAS
    L
                DESOLVATION CHAMBER
MOMENTUM
SEPARATOR
       LC EFFLUENT
D
             IR TRANSPARENT
             WINDOW FOR
             DEPOSITION OF
             SOLUTES

-------
o
GO
     D

          3500
30OO     250O     2OOO
     WAVENUMBER (CM-1)
15OO
1000

-------
  .4-
o
0:
O
  .2-
  0-
ANTHRACENE	100% HEXANE
                             JLJJJUi
                                LLJ
             3000        200O
                WAVENUMBER (CM-1)
                           1000
                                                       N)

-------
  .2-
LU
O
  .1 -
  .1 -
       CAFFEINE	100% WATER
                                                                     H
                                                                     *»
                                                                     W
                3000          2OOO
                    WAVENUMBER (CM-1)
1OOO

-------
97.23  97.98
     '/. TRANSMITTflNCE
98.73  99. *8 100.23 100.98 101 .73 102. «t8

-------
  .6-
  4-
O
O
GO
 .2-
  0
      80:20 MethanohWater

      Caffeine + KHP Buffer

      pH 4

      Collection time: 1.0 min
            3000
      2000

Wavenumber  (cm"1)
                                                             Ln
1000

-------
o

cd
,0

o OH
CO •"/»
  0
      Caffeine deposited from KHP

      at pH  = 4
      Caffeine reference deposited

      from 100% Methanol
            3000
      2000


Wavenumber (cnT1)
1000

-------
                               147
Slide 2
Slide 3
Slide 4
Slide 5
                   LC/FT-IR Spectrometry
                    The mobile phase is the problem
               In HPLC separations the compounds can have
                    concentrations as low as 100 ppm
         Dynamic Range

   FT-IR spectrometer sensitivity is limited
     by the dynamic range of the ADC


  Two Basic FT-IR Approaches
                       1. Use flow-through cells
                              (keep the solvent)
         2. Eliminate the solvent


Aerosol Beam Separator
 — Needed Properties —
 • High Solute Transport Efficiency
 • Effective Vapor Removal
 • Minimal Peak Broadening
 • No Memory Effects
 • No Adjustments

-------
                            148
Slide 10
                        MAGIC
             Monodisperse Aerosol Generation

                  Interface Combining...
Slide 22
MAGIC-LC/FT-IR SPECTROMETRY CHARACTERISTICS

                  • Elimination of all the solvent
                  • Normal phase and reverse phase operation
                  • Mobile phase can be 100% aqueous
                  • LC flow rates to 1 mL/min
                  • No heating of the effluent

-------
           149
I   »  I  »  I  »   I  »   I   »  I
                             0
                             c. °d

-------
                          150
 MdNOOJSPERSE AEROSOL  AEROSOL BEAM
      GENERATOR          SEPARATOR
DISPERSION /
   GAS
                       N1  S1 N2 S2

                         \  II  /
LC EFFLUENT
             OESOLVATION
              CHAMBER
                                      MS IONIZATION

                                         SOURCE

-------
                  % TRANSMITTANCE
97.33  97.96  98.73  99.*8 100.Z3 100.98 101.73 102.
 *F
                          TST

-------
             152
                             ffl
                             ^^^v









                             UJ
                                 O

                                 8
                                 O
                                 O
                                 o
                                 CM
                                 o
                                 K)
                                    U
                                    03
                            Ld

                            I
O
o
30NVUJSNVyi %
                      10

-------
      153
                 OQ
                                  o
                                  o
                                  (O
                                  o
                                  o
                                  00

                                  o
                                  o
                                  ^ »
                                  r-  o:

                                  o  OQ
                                  O  2
                                  ^  ^
                                 o
                                 o
                                 00
30NVJJJflSNVyi %
o
o
o
CM

-------
o
  B
                                              H
                                              Ul
2000 1800 1600  1400 1200 1000 800  600

          WAVENUMBER (CM-1)

-------
  .4-
O
o
en
,0
  ,2-
  o-
       80:20 Methanol:Water

       Caffeine + KHP Buffer

       pH 4

       Collection  time: 1.0 min
                                                               U1

                                                               Ol
             3000
      2000


Wavenurnber (crof1)
1000

-------

-------
                             157
               QUESTION AND ANSWER SESSION
                                   MR. TELLIARD:  Any
questions?
                                   DR. MSIMANGA:  My name is
Huggins Msimanga from Kennesaw State College in Marietta,
Georgia.
     It looks like to me you have to time...you have to
specify the amount of time to allow for the deposition of
the solute analyte.  Supposing two components come off from
the HPLC within a specified time of less than a minute, how
do you take that into account?  Suppose it is two different
analytes on the window.
                                   MR. DE HASETH:  Well, the
plate itself translates at a reasonable speed, the
deposition plate.  We can certainly separate two components
one minute apart.  That is a long time.  We can complete for
higher concentration components... by higher concentration,
we are talking a few hundred picograms.
     It has been demonstrated through direct deposition
GC/FT-IR which is a parallel technique to this that you can
get reasonable spectra a less than 200 picograms in 3
seconds of data collection time.   So, all you need is about
3 seconds of analysis, not a minute.
     The one I showed was through a beam condenser, and
there, yes, we did need a minute, because that was a much

-------
                             158
less efficient interface in terms of optics.  With this new
system, we can do the same analysis in a much shorter time.
     The depositions are confined to a radius of about 100
to 150 micrometers.  So, they are very small depositions.
We can focus indirectly on that and collect our spectra.
               MR. TELLIARD:  Thank you very much,.Jim.

-------
                             159



                                   MR. TELLIARD:  Our next



speaker is Tom Tiernan from Wright State University.  Dr.



Tom has been working on the joys of dioxin for the last few



years and spent the last couple of years working on a new



column which is going to save me a lot of money.  Right?



                                   DR. TIERNAN:  Probably.



                                   MR. TELLIARD:  Okay.  Tom



is going to talk about the analysis of dioxins and furans.

-------
                          160
                                   DR.  TIERNAN:  You heard
earlier  this  morning about  chlorinated dibenzodioxins and
chlorinated dibenzofurans from Les Lamparski.
          Most of  you  who do  environmental  analyses  .
and I know that  includes  many  of  you . .  . know that  there
are 75 polychlorinated dibenzodioxin (PCDD) isomers and 135
polychlorinated dibenzofurans (PCDF)  isomers.  No one,  to my
knowledge, is  even contemplating doing a routine type of
analysis for all 210 isomers of  these, but  in more recent
years, there has been a  focus  on determining the 2,3,7,8-
substituted isomers of the various  chlorinated groups of
PCDD   and  PCDF,   particularly   the  tetra-  through
heptachlorinated congeners.   The  reason for this  interest,
of course,  is  that these are thought  to be  the more toxic of
the PCDD/PCDF  congeners.    Clearly,  the seventeen 2,3,7,8-
substituted isomers  are  more readily  analyzed  than  all of
the 210 isomers.
          Even though  there are  methods  that purport to
measure  the  2,3,7,8-substituted  PCDD and  PCDF  isomers,
including some EPA  methods,  one of the problems with  these
methods is  that,  in order to achieve  optimum  reliability and
specificity for  the  quantitation  of  the 2,3,7,8-substituted
isomers the use  of  multiple  GC columns is required.    There
is no single GC column that will resolve all  of  the 2,3,7,8-
substituted dibenzodioxins  and  dibenzofurans  in  a  single
analysis.

-------
                           161
          We originally began the work which  I will describe
today with  the objective of  trying  to  come up with some
better columns  for  resolving and quantitating the 2,3,7,8-
substituted PCDD/PCDF  isomers.   At  the  outset,  we  focused
primarily on  just  the  tetrachlorinated isomers,  because we
were involved with the  U.S. EPA and the  paper  industry
(NCASI)  in a  program in  which the primary objective was to
measure  uniquely  and  specifically,   only  2,3,7,8-TCDD and
2,3,7,8-TCDF in sludges,  effluents,  pulp  and  other paper and
pulp mill wastes and products.
          The requirement in the latter case, of course, is
somewhat  simpler than to separate  all of  the  2,3,7,8-
isomers,  but,  still,  separating  even   those   two
tetrachlorinated isomers  absolutely  uniquely  is something of
a challenge.   At the time when we began this  work, there was
no single GC  column that would  separate both 2,3,7,8-TCDD
and 2,3,7,8-TCDF uniquely in  a single analysis.
          Of  course,  2,3,7,8-TCDD had  been  routinely
analyzed on a number of  different GC columns, specifically
and uniquely,  including the DB-5,  SP-2330 and other columns.
Also, there  was purported  to be  a unique  separation of
2,3,7,8-TCDF  from all  other  TCDFs  on the DB-225 GC  column.
Thus, determining both 2,3,7,8-TCDD  and 2,3,7,8-TCDF  in  a
given sample using these  columns  required two separate GC-MS
analyses.
          Much of the early work which we did with the U.S.
EPA  and  the  paper  industry  (NCASI)  in connection  with
developing a new

-------
                          162
column  for  determining both  2,3,7,8-TCDD  and 2,3,7,8-TCDF
focused on the concept of  using serially  connected  capillary
GC columns.   By that,  I mean using  two  different portions of
two different  columns  which are simply connected  together.
Typically, one does this using a dead volume fitting.
          There is, in fact, in the literature a good bit of
information about modeling the performance of these kinds of
capillary columns, as well  as columns which are coated with
mixtures  of different stationary  phases.   There  has  been
little  or no  modeling, however, which would  permit  one to
predict theoretically the  composition of  a single stationary
phase or  GC  column coating containing multiple functional
groups  which  would achieve the kinds of  separations of
interest here.
          So,  effectively, the initial  goal  of our work was
to come up with a modeling procedure that would permit us to
predict a stationary  phase  or  coating for a  capillary GC
column  which  would contain multiple functional  groups,  and
which would achieve the  unique separations  of these  two
isomers (2,3,7,8-TCDD and 2,3,7,8-TCDF).   The next phase of
this work was conducted  largely by our partner  in these
studies,  J  &  W  Scientific, and involved  synthesis  of  a
specific  polymer  coating  containing the  desired functional
groups in the predicted  amounts  (as  indicated  by  the
modeling results)  and  preparation  of a bonded capillary GC
column  coated with this  new stationary phase.   Finally,
testing of this  column to  determine its effectiveness  for
separating 2,3,7,8-TCDD and 2,3,7,8-TCDF was accomplished at
Wright State.
          I am going to describe to you today the outcome of
these studies.
          One  might well ask, what  is wrong with coupled

-------
                           163
GC columns if they achieve the desired separation?  Indeed,
we previously  developed a coupled column  that would,  in
fact,  separate 2 , 3,7,8-TCDF  uniquely.   In  addition,  we
eventually developed  a  three-section  capillary GC  column
combination,  that is,  portions of  three  capillary columns
coated with different stationery  phases  which were coupled
in  sequence,  that  would resolve  both  2,3,7,8-TCDD  and
2,3,7,8-TCDF uniquely  in a single  analysis.
          However,  it is difficult  to  convince  analysts  to
prepare  and  use  these  kinds  of  columns,  and  even  more
difficult to get a manufacturer to make them.  Such coupled
columns tend to be somewhat fragile, and the connections are
difficult  to  make  and the  junctions  tend  to  leak.
Chromatographers  are  therefore reluctant to make and use
such columns for routine analyses.
          These,  then, were  some of  the  reasons  for
attempting to develop a single-phase coated column, using a
specially tailored phase, which could be  manufactured and
supplied by a recognized gas chromatography  specialty firm.
          As we will see, what was required initially  in
terms  of  developing a  model  such  as that mentioned  was  to
measure  GC  retention  times,  capacity  factors,   and
temperature  dependences  for the gas  chromatographic
separations of  the tetrachlorinated  dibenzodioxin isomers
and the  tetrachlorinated dibenzofuran isomers using several
different capillary columns.
          Among the columns  that  were evaluated  here  were
DB-WAX,  SP-2250, DB-17,  DB-5,  SP-2200  and  DB-1.   The  data
derived from isothermal  measurements of the  parameters cited
were  then incorporated into  a .multidimensional  computer
model, which I

-------
                        164
am going to describe in rather general terms, to ultimately
predict  optimum combinations of  the various functional
groups which would be  required as  components of  the
specially synthesized phase in order  to obtain the desired
resolution.
          J & W Scientific used the modeling predictions to
synthesize the new polymer or  stationary  phase, and prepared
a bonded capillary GC column coated with  that phase.
          Ideally,  the first new column prepared would have
yielded the desired separations of  2,3,7,8-TCDD and 2,3,7,8-
TCDF.  In reality, of course,  scientific methods are rarely
that good,  and some iteration  of  the  modeling/synthesis
procedure proved to be necessary.
          One of the  reasons  that  the model for predicting
GC column performance did  not  work  quite  as  well as expected
initially is that, even when one synthesizes accurately the
polymer or  stationary phase  containing  the  predicted
functional groups in the predicted  amounts,  there are likely
to be  interactions between the  functional groups  on the
polymer that vary the activity and  the separation capability
of the column  coated with this  polymer.   We are  unable to
accurately predict these interactions.
          What we do, therefore,  is take the retention time
and related data derived from  the initially  prepared column,
feed these data  back into the model,  obtain new solutions,
and  come up with a second  prediction  of  the optimum
stationary phase composition.   If  necessary, this procedure
is repeated still again.

-------
                            165
          Now,  just  to give you  an indication  of  how
difficult  and  sensitive these kinds  of  isomer separations
are, I wanted to just show you some retention time data for
several  closely-eluting TCDF  isomers  which  was obtained on
two capillary  GC columns  at two  not  very  different
temperatures  (Slide 5) .   The columns here  are the  SP-2330
and DB-5

capillary  columns.   What you can  see if you  compare  the
retention times for  the  several isomers that are shown there
is that as you go from 220 to  250 degrees with both of these
columns,  you get some inversion in the order of  elution of
the isomers.
          Obviously,  if you look at  all 22 TCDDs  and  33
TCDFs,  this  becomes  even more severe.   You have frequent
mixing,  jumping,  and  shifting of positions in  terms  of  the
order of  elution.
          So, it is no easy feat to reliably predict where
these numerous  isomers are going to elute.
          The next slides  (Slides  6, 7  and 8}  show flow
charts which give you an idea  of  the  procedure used  in
applying  the model which I  have  mentioned.   I am going to
discuss in greater  detail,  a little  later, what  are  the
specific  critical  parameters here.
          The first thing we have to do,  of course, is write
the  software  and develop  the  model,  and  this  was
accomplished  for use with a PC type  computer.  The model  was
actually  written in Turbo Pascal for  an  IBM-compatible  386
PC with a 387 co-processor.   The model will generate  about
1000 to 2000  solutions, in terms of

-------
                         166
predicted retention  times,  in  a minute, and  in an hour  or
so, we  typically get the bulk of the  predictions we are
after.
          In applying the model,  we first enter the relevant
experimental  parameters  which  include   the  column  length,
radius, film thickness,  inlet and outlet pressures, and the
dead  times  at a minimum of  two temperatures.   These are
entered into an equation or formula from which we  ultimately
calculate capacity factors.  Initially, we  determine two
experimental parameters  or constants  which  are called  Kl and
K2.   Of course,  to  get  the  Kl  and K2 constants, we have  to
measure isothermal retention times for  the group  of  isomers
in which we are  interested at  different temperatures.  Data
obtained at a minimum of two  different temperatures are
required for each of the columns evaluated.
          We  then  insert the  retention time  data  obtained
for a given column  at two different  isothermal  temperatures
into  a  set  of simultaneous  linear equations  and  solve them
to determine the constants,  Kl  and K2 (Slide 9).
          Finally,  we utilize  these  constants in  the  third
phase of the model which gives  us the capability
to predict  the  effect on retention times of each  functional
group that  we wish  to incorporate  into  the  polymer backbone
that  becomes  the ultimate coating or stationary  phase for
the column.   Finally,  of course,  we synthesize  the  phase,
coat  the column with this phase and test the column.
          So here then,   is  the  capacity factor relationship
(Slide  9) ,  and you  can  see  from equation  number 3  at the

-------
                           167
bottom,  which you eventually work down  to ...  the  input
data,  of course, is  again retention  time, dead time,
isothermal temperature  .  .  .  one can  see  that  by measuring
the retention  time  parameters for  a  given  column at two
different T's,  one can insert these data  into  equation  (3)
and solve the two resulting simultaneous equations, and come
up with  the Kl and  K2  factors,  which  are  the  empirical
factors which we must obtain initially.
          Now,  these  constants  are  then  used  in the
retention time relationship  shown in Slide 10 (Equation  4A)
to come  up  with retention time  predictions at different
temperatures,  in other words, to predict  the times  at
which isomers  are going to elute  on a  given  column as  a
function of temperature.
          Once we have  that  data,  we then  go to  the  last
stage.   We  put  these retention times  into equation 5  shown
on  Slide 11  and,  again,  by  doing  this at  multiple
temperatures and solving  simultaneously, we come up with the
K-prime  factors which are  characteristic  of  the actual
functional groups, that is, the pure functional groups, methyl,
phenyl,  or whatever  groups we want to incorporate into  the
polymer.

          On Slide 11,  for example,  are  shown the  applicable
 equations for  two different columns  that we have  tested in
 this  study.   Column 1,  of course,  is  actually  the  DB-5
 column,  which  incorporates 95 percent methyl and  5 percent
 phenyl  components,  and Column 2 is  the  DB-17 column which
 incorporates about 50/50 methyl/phenyl components.

-------
                        168
          After we have  obtained  the  K prime factors, that
is, the functional group factors,  we get to the very end of
the procedure.   We put these back  into the retention time
relationship (Slide 12), and we determine, for  any given
combination of these several substituent  groups  in a
polymer, what the predicted retention times for the various
isomers to be separated are going  to be.
          So, in  other words, we  can now  predict  what a
totally  new   column   containing  entirely  different
combinations of  functional  groups  would yield  in terms  of
retention times for the PCDD/PCDF  isomers.
          Of course,  the experimental  verification  of the
results requires actual measurements using a GC or GC-MS
(Slide 13).   In this  case,  we are using  a low  to medium
resolution sector mass spectrometer, a Kratos MS-25.  It is
operated in the selected ion monitoring  mode, of course, for
monitoring TCDD and TCDF.   Here,  you see  the  sets of ions
that are  typically  monitored as indicators  of  both native
TCDF and the isotopically labeled TCDF and the same for
TCDD.
          The chromatograph used here  is  a Carlo-Erba, and
in the development phase  of  this  work,  we  used hydrogen as
the carrier  gas.   Subsequently, we have  demonstrated that
the predictions are also  applicable and  the same column
performance can be  achieved  when we  use  helium  as the
carrier gas.
          Throughout  this  study, retention times  were
measured  on  the various columns for at least three different
temperatures, typically 220, 235 and 250 degrees  C.
          Well,  let's  look  now at  some  data  obtained  in  the
initial  stages of this  work  where we  attempted  to  predict
the  combinations of functional groups  in a  single stationary

-------
                           169
phase column coating  which  would give us  an adequate
separation of  both  2,3,7,8-TCDD  and 2,3,7,8-TCDF  from all
other PCDD/PCDF isomers.
          Clearly,  you must  have  a separation of  at least
one  peak width  of  the 2,3,7,8  isomer  from  the nearest
eluting TCDD or TCDF congener which is  adjacent to it.   So,
here  you  see  (Slide 14)  some  of  the  predicted separations
which resulted from the modeling  calculations.   It appears
that  the  very  first combination of functional  groups shown
on Slide  14 does, in  fact,  yield  the  optimum separation in
terms of  peak  width separation for 2,3,7,8-TCDD  from the
other TCDD isomers, and this  combination  also  results  in a
quite good separation  of 2,3,7,8-TCDF.
          So,  initially,  we  indicated  to  J & W   that  they
should synthesize a polymer  that would  contain the
 40:25:25:10  proportions  of  the   methyl: phenyl: cyano : WAX
 functional  groups.   WAX simply denotes a  PEG  (polyethylene
 glycol)  functional group.
          J  & W Scientific synthesized a  polymer  substrate
 containing  these functional  groups in  what they  believed  to
 be the proportions specified  by  us,  and prepared a bonded
 column  coated  with  this polymer.   When we tested  this column
 for its  ability to separate  2,3,7,8-TCDD  and  TCDF, the
 observed column performance  was different  than  that
 expected.  When  the experimentally observed  retention  times
 were compared  with  the predictions of  the model,  it could  be
 seen that the column  was  acting  as if the composition  were
 that shown on the  second  line (Result  1st synthesis)  of
 Slide 15.   The  first line  on the slide  shows  the column
 composition predicted by the initial calculation,  which J  &

-------
                         170
W attempted to obtain.  The difference in the theoretically
expected and experimentally observed behavior of the column
was  undoubtedly  due  to  the  fact   that  there  were
unpredictable  functional  group interactions  in the  newly
synthesized stationary phase coating.
          We next inserted  the experimentally  observed
retention time  data  for the  initial  column back into  the
model and calculated a new  solution,  that  is,  an  iterative
prediction of the stationary phase composition.
          Before looking at the new solution,  let's just
look at what the  initial column did  in  terms of separating
the isomers (Slide 16).  Here  you  can see that the  2,3,7,8-
TCDD  is not at all well separated from adjacent TCDD
isomers, and exhibits  a  strong overlap  with 1,2,3,8-  and
1,4,6,9-TCDD.
          The separation of 2,3,7,8-TCDF  is considerably
better  (Slide  17) , but,  there is  a  considerable  valley
between 2,3,7,8- and 2,3,4,6-TCDF and the  other co-eluting
isomers.
          Well,   the  results  of  the  second  modeling
prediction  obtained  by reiterating  the  model calculations,
as described above,  are  shown  in Slide  18  on the  third and
fourth  lines.   The  second prediction  of  the  optimum
functional  group composition of the column stationary phase
which we selected was the 43:26:22:9  combination, and when a
polymer with this expected composition was  synthesized and a.
bonded  capillary GC column was coated  with  this polymer, the
performance of this column in  terms  of retention times of the
separated isomers, corresponded closely to  the composition which
was originally predicted.  The last  column  on Slide 18 (hori-
zontal) shows the calculatedcomposition to which the experi-
mental behavior corresponded.

-------
                           171
          This  is,  in  fact, the composition of the capillary
GC column that  has now become known as DB-DIOXIN and  is
currently being marketed by J & W  Scientific  under that
name.
          The separation  which this  column  achieves for
2,3,7,8-TCDD is  illustrated in Slide 19, and as can be seen,
this  separation  is  quite  good.    Slide  20  shows the
separation achieved with the DB-DIOXIN  column for 2,3,7,8-
TCDF, again, well within desirable bounds.
          You can see  from the  comparison shown  in Slide  21
that the retention time predictions derived  from  the  model
are quite close  to the experimentally  observed retention
times.   The  second vertical column here shows the expected
or calculated retention times for this set of isomers,  while
the third column shows the actual  experimental retention
times that were observed.  You  can  see that,  in  most cases,
these are extremely close to the predictions.
          The sensitivity of  the  DB-DIOXIN  column  is
excellent.    The  mass  chromatograms  shown in Slide  22, for
example, which  were obtained  for  a  2.5 pg  injection  of
2,3,7,8-TCDF, exhibit  about a 25:1 signal to noise ratio for
the responses at the ion masses which are indicators of both
native  TCDF and the isotopically-labeled TCDFs that would  be
monitored in a  normal analysis  run.   The  DB-DIOXIN column
exhibits similar sensitivity for 2,3,7,8-TCDD, as  shown  on
Slide 23.

-------
                          172

          We have now used the  DB-DIOXIN column  very
extensively for analysis of  extracts  obtained  from a large
variety of  sample matrices.   Slide 24  shows  some  of the
types of  samples  that have been analyzed  with  this column
during the  past  year or so,  during which time a  total of
about 500 samples  have been characterized for  2,3,7,8-TCDD
and TCDF.   In fact, the very same DB-DIOXIN column was used
for all of these analyses.
          You can  see  that the samples  analyzed  with this
column range from relatively clean biologicals and human
tissues to  very  dirty chemical wastes.   It  turns  out that
pulp and paper  samples  and wastes pose a  particular
challenge here,  partly because of  the  kinds and  types of
interferences.
          The DB-DIOXIN column also  exhibits good,  long-term
stability.  You can see from the mass chromatogram shown in
Slide 25 that even after some 600 analyses with the column,
the resolution  is essentially unchanged for the  2,3,7,8-
TCDD.  The  same long-term  performance  of this column is
observed for 2,3,7,8-TCDF,.as seen  from the results shown in
Slide 26.
          The conclusions from this study are summarized on
Slides 27 and 28.   The main conclusion is that the DB-DIOXIN
column is quite effective for the analyses of both 2,3,7,8-
TCDD and 2,3,7,8-TCDF in a single GC-MS  run.   Moreover, the
theoretical model developed  in  this study  appears  to be
capable  of reliably  predicting  the  combinations  of
functional  groups  in a polymeric  stationary phase coating
which will yield a capillary GC column capable  of achieving
the desired separation of  virtually any  closely eluting
compounds.

-------
                           173
          Of course,  we are  not only  interested in  the
separation of  tetrachlorinated  dioxins  and  dibenzofurans,
but are also interested in developing  capillary  GC columns
that will be  useful for separating  the  entire chlorinated
range of these compounds,  and in particular,  in uniquely
separating all of  the other  2,3,7,8-substituted  congeners.
We have done some preliminary work toward this end already.
          Obviously,  the  first  step  in  achieving this
objective is to see what isomers the DB-DIOXIN column which
we already have  in  hand is capable  of resolving.   We have
already determined that this  column  will, in fact,  separate
all  of the  2 , 3,7,8-substituted  congeners of  all  the
chlorinated  groups  from each  other.   However,  we  do not  yet
know that this  column will  separate all of the other isomers
of each  chlorinated  group from  the 2 , 3,7,8-substituted
isomers.
          Some  overlap  of  the  chlorinated  group  retention
time windows occurs  with the DB-DIOXIN column.  For example,
some of  the TCDFs  coelute with  some  of  the  PeCDFs,  as
indicated by Slide 29.  So the DB-DIOXIN column is somewhat
different than the  DB-5 column,  on which  the respective
chlorinated  congener  groups  are quite well  separated from
each other,  at least  under the  conditions  that are usually
applied,  and the  DB-5 column is therefore more  useful  for
the  total  chlorinated  congener group  type  of  analysis.
Nevertheless,  at this point,  it  appears  that  the DB-DIOXIN
column shows considerable promise for analyses of  just  the
2,3,7,8-substituted  isomers.

-------
r
                                      174
                      We are in the process of evaluating the separation
            capability of the DB-DIOXIN column for the entire 210 isomer
            PCDD/PCDP set.
                      The  last  slide   (Slide  30)   shows  you  the
            separations that are  achieved with this column  for  some  of
            the 2,3,7,8-penta CDDs and CDFs.
                      In conclusion, let me just  say  that  the DB-DIOXIN
            column is currently available from J & W Scientific,  and,  in
            their most recent catalog, which  has  just  been issued,  some
            very nice mass  chromatograms are shown which exhibit the
            separations obtained  for  all of the TCDF  and  TCDD isomers,
            using both helium and hydrogen carrier gases.
                      So, here  for  the  first  time, we  have  developed  a
            working, stable, commercially-available capillary  GC column
            that will permit one  to determine at  least 2,3,7,8-TCDD and
            2,3,7,8-TCDF in  a  single analysis  and  will,  as  Bill  said,
            possibly save the agency a little  money.
                                               MR. TELLIARD:   A lot of
            money,  Thank you,  Tom.
                                               DR. TIERNAN:  Thank you.

-------
                            175
                QUESTION AND ANSWER SESSION
                                   DR.  TELLIARD:   Are  there
any questions?
                                   MR.  LEWIS:   Michael  Lewis
of TC Analytics.
          What  was  the  optimum column  dimensions that  you
are using for your . .  .
                                   DR.  TIERNAN:   Sorry, what
was the question?
                                   MR.  LEWIS:   What was  the
optimum  column dimensions  that  you were  using for  your
development work?  And conditions.  Were they  standard?
                                   DR.  TIERNAN:   Oh,  yes.
You mean like temperature programs and  what  not?
                                   MR.  LEWIS:   A  30 meter or
a ...
                                   DR.  TIERNAN:   Oh,  well,
the original  coupled column  we  put together  was about  40
meters in length.   But  the  present DB-DIOXIN column can  be
anywhere from 30 to 60 meters.  It just depends on what  you
want to achieve here.  But these  are standard  length columns
as compared to current  technology, and  the temperature
programs one would use here  are very standard,  too.

-------
                          176
          I didn't mention it, but,  typically,  you  would be
programming the column from about  180  to 220 degrees  at
about 10 degrees C per minute and then holding at the higher
temperature.  We know, however,  that this  column will go as
high as 250 to 270 degrees C with no degradation.
                                   MR.  LEWIS;    Were  you
using attempt  to  look  at more the megabore standard
dimensions right now?
                                   DR. TIERNAN:   No, this is
not a megabore  column,  and we probably are not  going to do
that  at  this point.   No,  this  is  a standard  0.25  micron
dimension column.
                                   MR.  TELLIARD:   Anyone
else?
          Tom,  what  is  the run  time  when  you  use  the
hydrogen carrier roughly?
                                   DR. TIERNAN:  Actually,
you can  adjust  the head  pressure  of helium or  hydrogen to
achieve  the  same effective  separation with either carrier
gas, and  the  run time is on  the order of about  30 minutes
for analyses  of both  2,3,7,8-TCDD  and 2,3,7,8-TCDF  in  the
same run.
                                   MR. TELLIARD:   That  is
good.
                                   DR. TIERNAN:   Yes.
                                   MR. TELLIARD:  Thank  you
very much.
                                   DR. TIERNAN:   Thank you.

-------
   NEW CAPILLARY COLUMN FOR THE
DETERMINATION OF PCDD's AND PCDF's
      T.O. Tiernan, J.H. Garrett, and J.G. Solch
            Department of Chemistry
      and Toxic Contaminant Reseach Program
            Wright State University
                Dayton, Ohio

                   and

        R.M.A. Lautamo and R.R. Freeman
               J & W Scientific
              Folsom, California

-------
                    BACKGROUND

1. Separation of the 2378- substituted isomers of the DibenzoDioxins and
    DibenzoFurans provides a difficult challenge for capillary GC columns.
 2. Tetrachlorinated Isomers
      •Dioxins (TCDD)-22
      • Furans (TCDF)-38
 3.  Requirements:
      •Separate 2,3,7,8-TCDD from all other TCDDs
      •Separate 2,3,7,8-TCDF from all other TCDFs
4. While separation of 2,3,7,8-TCDD from all other TCDDs is routinely
    accomplished on several capillary GC columns (DB-5, SP-2330,
    SP-2340), no single liquid-phase coated column has been reported
    which separates 2,3,7,8-TCDF from all other TCDFs and
    2,3,7,8-TCDD from all other TCDDs in a single GC-MS analysis.
00

-------
                      BACKGROUND
5. Initial studies in our laboratory using retention time data obtained on
    seven columns (DB-5, DB-225, DB-1701, DB-WAX, SP-2250, SP2340,
    and SP2401) led to the development of two different columns
    consisting of coupled serially-connected sections of single-phase
    columns. A combination of two different phases (DB-5 and DB-225)
    successfully resolved 2,3,7,d-TCDF from all other TCDFs. A
    combination of three phases (DB-WAX, DB-225, and SP-2550)
    simultaneously resolved 2,3,7,8-TCDD from all other TCDDs and
     2,3,7,8-TCDF from all other TCDFs.
IO
6. Coupled columns, while capable of the required fisomer resolution, are
    somewhat difficult to handle and reproduce.

-------
    OBJECTIVES OF PRESENT STUDY
1. Measure GC retention times, capacity factors and temperature dependences
    of separations for all TCDD and TCDF isomers using a variety of capillary
    GC columns.

2. Utilize data obtained empirically in a multidimensional computer model to
    predict optimum combinations of various functional group phases whicji
    will yield complete resolution of both 2,3,7,8-TCDD and 2,3,7,8-TCDF.
00
o
3. Synthesize a single liquid phase polymer coating incorporating the
    functional groups selected in the relative proportions predicted by the
    model and construct a capillary column coated with this phase.

-------
OBJECTIVES OF PRESENT STUDY
4. Test the efficacy of the column for achieving
    resolution of 2,3,7,8,-TCDD and 2,3,7,8-TCDF
    in a single run.
5. Use data derived from initial column to obtain
    an iterative solution using the model, as necessary.
6. Synthesize new polymer coating based on revised
    model predictions, and construct new column.
7. Demonstrate complete resolution of 2,3,7,8-TCDD
    and 2,3,7,8-TCDF using finally developed column.
                                                        CO

-------
    EFFECT OF TEMPERATURE ON
      TCDF ISOMER LOCATIONS
ISOMER
SP2330 (220 C)
1,2,3,9
2,3,4,7
1,2,6,9
2,3,7,8
2,3,4,8
ISOMER
2,3,7,8
1,2,7,9
2,3,4,8
1,2,6,9
2,3,6,7
1,2,3,9
0.938
0.950
0.984
1.000
1.012
DB-5 (220 C)
1.000
1.001
1.004
1.064
1.080
1.081
SP2330 (250 C)
                                  0.965
                                  0.963
                                  0.996
                                  1.000
                                  1.009
                                DB-5 (250 C)
                                  1.000
                                  1.005
                                  1.001
                                  1.046
                                  1.030
                                  1.063
                                                00
                                                NJ
        Retention times relative to 2,3,7,8

-------
                                   183
/       Create        \
I    Component List     I
V   (Isomer Names)    /
  Enter Initial Data: ^

 Column length*radius.
    and film thickness
 Inlet and outlet
    pressures
 Dead volume times at
    a minimum of tuo
    temperatures
     Enter
  Temperature
    Program
      and
Retention Times
   Calculate
   Capacity
    Factors
flnother
column
 type?

-------
                              184
    fill Data  \Yes
  Collected on
Single functional
      group
    columns?
           No
     Select
   Isothermal
   Temperature
       and
Functional Group
       to
     resolve
       Solve
simultaneous linear
   equations to
   calculate neu
 capacity factors
  that represent
"pure" phase at the
     selected
   temperature-
                     Yes
      flnother
    "Functional
       Group
  or Temperature
        to
     resolve?
No
        Select:
          - functional
            groups
          - components
            to separate

-------
                                185
      Calculate
      Capacity
     Factors for
         New
       Column
       Enter:

 Temperature Program
         and
   Retention Times

   for New Column
 Enter Data for Neu
       Column:
Column length*radius»
   and film thickness
Inlet and outlet
   pressures
Dead volume times at
   a minimum of two
   temperatures
Predict New Column '•

 - Calculate % of
  each functional
 group required to
  produce maximum
   resolution of
selected components
    Synthesize
        new
      column
    New Column
   meet required
    resolution?
Success

-------
       EXPERIMENTAL CONDITIONS
                (FOR DEVELOPMENT)

Kratos MS-25 mass spectrometer
DS-90 Version 5.00 software
Selected-Ion monitoring mode
Ions monitored:
      TCDF m/z 304, 306, 312, 316, 318
      TCDD m/z 320, 322, 328, 332, 334

Carlo-Erba Mega 5300 series Gas Chromatograph
Splitless Injection
Hydrogen carrier gas  1.20 kg/square cm  and  2.40 kg/square cm
Inlet and exit carrier gas pressures measured to
+/- 0.07 kg/square cm
Retention times measured at 220 C, 235 C, 250 C
CO

-------
           Capacity Factor  Calculations
              Tr
              Td
              T
              k'
         k1 & k2
- Retention Time
- Dead Volume
- Temperature
- Capacity factor at a given Temperature
- Experimentally derived constants
         Tr • Td ( 1 + k' )

         Ln ( k' ) • Ln ( k1 ) + k2 / T
                                 (1)

                                 (2) a.
oo
       Substituting Equation 1 into Equation 2 yields:

         Ln ( ( Tr / Td ) - 1 ) - Ln ( k1 ) + k2 / T    (3)

       Under experimental conditions Equation 3 contains two
       unknowns ( k1 and k2 ). Values for k1 and k2 can be
       calculated by  solving simultaneous equations for data
       collected at two or more isothermal temperatures.
a. Akporhoor.Vent.and Taylor, J. Chrom.,405 (1987)

-------
        New  Retention  Time Calculations
             Tr
             Td
              T
              k'
         k1 & k2
- Retention Time
- Dead Voiume
- Temperature
- Capacity factor at a given Temperature
- Experimentally derived constants
         Ln ( k' ) - Ln ( k1) + k2 / T

       Equation 2 can be written as

         Tr • Td [ 1 + k1 exp ( k2 / T ) I
                                 (2) a.
                                (4) a.
       Since k1 and k2 were determined during the capacity
       factor calculation, Equation 4 permits calculation of
       predicted retention times.
                                                00
                                                00
a. Akporhoor.Vent.and Taylor, J. Chrom.,405 (1987)

-------
    Calculation of Functional Group Capacity Factors


           Existing column phases:
                        Column.1: 95% Methyl,  5% Phenyl
                        Column_2 : 50% Methyl, 50% Phenyl

           Column_1:

            Tr • Td ( 1 * 0.95 * k'.Methyl * 0.05 * k'_Phenyl )   (5)

           Column_2:

            Tr • Td ( 1 + 0.50 * k'_Methyl + 0.50 * k'_Phenyl )   (6)

           -  Equations 5 and 6 each contain two unknowns
              ( k'_ Methyl and k'.Phenyl ). Effective capacity
              factors for each functional group  can be calculated
              by simultaneously solving the above equations for
              data collected under the same experimental
              conditions.
H
CO
VO
I	

-------
   PREDICTING A  NEW COLUMN
Equation for retention time prediction for new combinations
of "n" functional groups:

    Tr - Td [ 1 + ( %_Qroup_1 * k'_Group_1 )
              + ( %_Group_2 * k'_Group_2 )
                  ... additional groups ...
                %_Group_n * k'_Group_n
   where %_Group_1 + %_Group_2 + ... + %_Group_n
1
After calculating the expected retention times, (Tr), for
each component in a specific combination of functional
groups, the isomer least resolved from the 2,3,7,8 isomer
is used for the resolution value for that phase combination.

The best solution must be selected from those combinations
giving a minimum resolution of one peak width.

Other factors that must be considered:
  _ nurnber of functional groups contained in the prediction
  - stability of the solution ( width of solution window)
                                                                vo
                                                                o

-------
FORWARD PREDICTION FOR DB-DIOXIN
     % Functional Group

Methyl   Phenyl  Cyano   Wax
 Separation
 (Peak Width)
7CDD   TCDF
40
40
40
45
40
40
45
45
45
45
25
25
25
25
30
25
20
30
2$
35
25
30
20
15
20
25
5
5
10
5
10
5
15
15
10
10
10
15
15
10
1.19
1.15
1.09
1.05
1.03
0.99
0.95
0.89
0.83
0.75
1.59
1.78
1.42
1.34
1.32
1.53
1.11
1.16
1.32
1.10

-------
  PREDICTIONS FOR DB-DIOXIN COLUMN
First Prediction

Result 1st Synthesis
                      % Functional Group

                   Methyl  Phenyl  Cyano  Wax
vo
to
40
45
25
32
25
13
10
10

-------
RESOLUTION OF TCDD ISOMERS - FIRST COLUMN
          1234,1236,1268
                           1269
                           1279
                                     1267
                                                   u>
                                          1
    23:50
25:50
27:50
29:50
    Wriglit State University and J & W Scientific

-------
RESOLUTION OF TCDF ISOMERS - FIRST COLUMN
    25:30
                 2346, 2348,1239
              2347
                                   3467
                                                   VO
                 I
                         I
27:30
29:30
31:30
    Wright State University and J & W Scientific

-------
PREDICTIONS FOR NEW DIOXIN/FURAN COLUMN
                  DB-DIOXIN
                        % Functional Group
 First Prediction
 Result 1 st Synthesis
 Second Prediction
 Final Synthesis
Methyl
  40
  45
  43
  44
Phenyl
  25
  32
  26
  28
Cyano
  25
  13
  22
  20
Wax
10
10
 9
 8
                                                vo

-------
RESOLUTION OF TCDD ISOMERS ON DB-DIOXIN COLUMN
                           2378
                      1246
                      1249
            1238
        1268 1234   1236
                  1469
                                                         VD
                                                         CTl
        1278
            27:00
28:00
29:00
30:00
          Wright State University and J & W Scientific

-------
RESOLUTION OF TCDF ISOMERS ON DB-DIOXIN COLUMN
                    2378
                                         2346
              2347
                                I
   30:00
31:00
32:00
                             1
33:00
       Wright State University and J & W Scientific

-------
I
         EXPECTED vs ACTUAL RT's FOR DB-DIOXIN
                        (TCDD REGION)
            Isomer
             1478
             1249
             1246
             1268
             1234
             1236
             2378
             1469
Expected Rt
  24:40
  24:44
  24:46
  25:00
  25:32
  25:28
  26:02
  26:14
Actual fit
 24:40
 24:38
 24:40
 25:22
 25:25
 25:26
 26:06
 26:14
ID
CO

-------
2.5 PG OF 2378 TCDF ON DB-DIOXIN
     m/z306
     m/z304
     m/z 241
     m/z 312
     m/z 316
     m/z 318
          30:00
30:26
30:51
  Wright State University and J & W Scientific

-------
2.5 PG OF 2378 TCDD ON DB-DIOXIN


                  2378


     m/z 322
     m/z 320
     m/z 257


     m/z 328
     m/z 334
     m/z 332
        27:03
                          to
                          o
                          o
27:13
28:44
   Wright State University and J & W Scientific

-------
 MATRICES ANALYZED ON DB-DIOXIN

• Biologicals:
   Beaver, fish, frogs, and mice
• Human Tissues:
   Blood and fat
• Air Samples:
   High volume ambient air and fly ash
• Other Environmental Samples:
   Soils, sludges, and waters
• Chemical Wastes
• Paper/Pulp Products and Wastes
           500 Samples of Above Types Were Analyzed
            Over a 12 Month Period Using This Column
to
o

-------
RESOLUTION OF 2,3,7,8-TCDD ON DB-DIOXIN AFTER 250
       STANDARDS AND 350 SAMPLE ANALYSES
                         2378
                    1246     1268 1238
                    1249    " 1237 1234
                                    1236
                                    1469
           1278
         25:47
26:37
27:28
28:19
29:10
                                      10
                                      o
       Wright State University and J & W Scientific

-------
RESOLUTION OF 2,3,7,8-TCDF ON DB-DIOXIN AFTER 250

      STANDARDS AND 350 SAMPLE ANALYSES
                 2378
              2347
                                 2346
1 1
29:10
I I
31:26
I I
31:16
I
32:07
                                                   to
                                                   o
       Wright State University and J & W Scientific

-------
             CONCLUSIONS
An analytical approach has been developed to solve
difficult separations of compounds. This approach
has been successfully utilized to produce a capillary GC
column capable of separating both 2,3,7,8-TCDD and
2,3,7,8-TCDF from all other TCDD and TCDF isomers in
a single GC-MS analysis.
to
o

-------
                CONCLUSIONS
The approach developed takes into account variances in:
       •temperature programming
       -head pressure
       -column length
       -film thickness
       -functional group interactions
10
o
01
The synthesis of a tailored bonded-phase column coating was
essential in order to obtain a column of sufficient durability
and reproducibility which would achieve the required separations.

-------
    RETENTION TIME WINDOWS FOR CONGENER GROUPS
               USING HELIUM CARRIER GAS
 OCDD

 OCDF

HpCDD

HpCDF

HxCDD

HxCDF

 PCDD

 PCDF

 TCDD

 TCDF
fj 2,3,7,8 Substituted
        20
               40            60
             Retention Time in Minutes
                                                   NJ
                                                   O
                                                   cn

-------
SEPARATION OF SOME PCDD & PCDF ISOMERS ON DB-DIOXIN
                 13468
         PCDPs
         m/z 338-340
         PCDD's
         m/z 356-354
           L
          32:50
                              12378
                   23478
                                           12389
                JL
                        12468/12479
                   12367 12389
                 12378
                      ..JL.
                            J
36:28
40:06
43:43
47:21
                                                             to
                                                             o
            Wright State University and J & W Scientific

-------
                                        208



                               AFTERNOON SESSION



                                         MR. TELLIARD:  Our next



           speaker is going to be Tom Fielding.



                                         MR. FIELDING:  Thank you.



                Bill asked me to speak for 15 minutes this morning,  and



           then he told me I was going to moderate the afternoon



           session, and I figured that was my 15 minutes right there.



           So,  without further ado/ let's start with our afternoon



           session.



                Our first speaker is Mr. Ted Martin of the Chemistry



           Research Division, USEPA-EMSL who will talk about single



           laboratory evaluation of EPA Method 200.8.



                Ted?
L

-------
                                      209
     SINGLE LABORATORY EVALUATION OF METHOD 200.8, DETERMINATION OF TRACE
           ELEMENTS BY INDUCTIVELY COUPLED PLASMA-MASS SPECTROMETRY
                    T.D.  MARTIN,  USEPA,  EMSL-CINCINNATI AND
                   S.E. LONG, TECHNOLOGY APPLICATIONS, INC.
     Inductively coupled plasma-mass spectrometry (ICP-MS) is currently
receiving attention as the newest spectrochemical technique for trace element
analyses of environmental samples.  It is a multielement technique that
provides extremely low limits of detection and freedom from the potentially
complex interelement spectral interferences of atomic emission spectrometry.
It is not without limitations, but when properly recognized and its strengths
utilized, it is an extremely powerful tool for environmental  analyses.
     In this presentation certain procedural features of Method 200.8 and
specific requirements for its use will be discussed.  Also,  presented will be
detection limit data, single laboratory data compared with inductively coupled
plasma-atomic emission spectrometry for the analysis of aqueous and solid
matrices and the status of a joint EPA/AOAC multi-laboratory study currently
being conducted.

-------
 SINGLE  LABORATORY  EVALUATION  OF METHOD 200.8


DETERMINATION  OF TRACE ELEMENTS IN WATERS AND


       WASTES  BY ICP MASS SPECTROMETRY
                  Theodore D. Martin

       Environmental Monitoring and Systems Laboratory


                        and


                    Stephen E. Long

               Technology Applications Inc.



       UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

                 CINCINNATI, OHIO 45268
                                                                to
                                                                H*
                                                                o

-------
                                       211
                 SINGLE LABORATORY EVALUATION OF METHOD 200.8
                   DETERMINATION OF TRACE ELEMENTS  IN WATERS
                      AND  WASTES BY ICP-MASS SPECTROMETRY
                                                   f

                   T.D.  MARTIN,  USEPA-EMSL AND S.E.LONG, TAI
     The information that will be presented in today's talk can be divided
into three topical or subject areas.  The first portion will be a brief
functional description of ICP-MS, the second part will be a discussion of
certain procedural requirements of Method 200.8 followed by comparative data
to inductively coupled plasma-atomic emission spectrometry (ICP-AES) and the
final portion will be a status report of the joint EPA/AOAC multi-laboratory
study currently being conducted.

     ICP-MS offers some unique and distinct advantages to.environmental
analyses.  Some of these are listed on the next slide (#2).  The first item of
particular importance is a linear range of 5 orders of magnitude, and up to 8,
if the instrument is equipped with some type of extended range capability.
This wide dynamic range is very useful for analysis of dilute solutions
especially when viewed along with the low, PPT, detection limits.  ICP-MS is
considered to be a simultaneous analysis technique.  This is achieved by
repeated, rapid, sequential scanning of the ionized sample over the mass range
5-250 amu.  This process provides multielement determinations of the entire
periodic table with the few noteworthy exceptions (C, N, 0, F, Si,  P and S).
By the very nature of the technique, isotope ratio information is available
with each analysis.  This feature gives the analyst the capability of doing
very accurate determinations by isotope dilution.  This same isotope
information supplies recognizable patterns, like fingerprints, which can
facilitate qualitative analyses.  Sample throughput is rapid (5 min. per
sample) while generally providing for some analytes detection limits that are
equal or lower than graphite furnace atomic absorption spectrometry (GFAA).

     The next slide (#3) is a functional schematic of the VG PlasmaQuad
spectrometer.  This is the instrument used in the single laboratory validation
study to be discussed later.  The slide is self-explanatory.   However,  it may
be helpful to trace the analyte route through the system.

     The sample solution is pumped to the nebulizer where the resulting
aerosol droplets of the analyte traverse the spray chamber and enter the
plasma through the injector tube of the torch assembly.   The energy of the
plasma desolvates, dissociates and ionizes the analyte.   A portion  of these
ions enter the expansion chamber through a small  hole in the sample cone
interface.  The ions are then drawn into the spectrometer through even a
smaller hole in the skimmer cone and are focused by the lens supply past the
photon stop into the quadrupole.  By virtue of their mass-to-charge ratios,
the ions sequentially pass through the charged rod assembly onto the detector.
The resulting signals from the detector are processed by the multichannel
analyzer and the data transferred to a dedicated computer.

     It is important to note the most critical area of the  instrument is the
expansion chamber or cone assembly.  It is in this area where the oxides and

-------
                                    212
polyatomic molecular ions are primarily formed and where analyte deposition
can occur at the cone orifice causing instrument drift and memory effects.

     The following slide (#4) shows the actual instrument.  The stand alone
unit on the right is the RF generator for plasma.  The torch box assembly and
spectrometer interface are enclosed behind the dark door in the center of the
instrument.  On the right are the gas controls for the plasma and on the left
are the controls for quadrupole, vacuum pumps and readout systems.  The rotary
vacuum pumps and required cooling system are behind the instrument and cannot
be seen on this slide.

     The next slide (#5) shows the torch lighted and with the torch box
assembly in the advanced or working position in front of the spectrometer
sample cone interface.

     The following slide (#6) shows the appearance of the sample cone
interface of the spectrometer after a period of operation.  The cone is made
of nickel and when new is very shiny.  The heat of plasma is the main cause
for change of appearance.  Two hours of operation can account for such a
change.

      A new sample cone is shown in the next slide (#7).  Although shielded by
the sample cone during operation, the heat also affects the skimmer cone shown
at the right.

     With all the advantages offered by ICP-MS there are still some limiting
factors.  On the next slide (#8) are listed the most common interferences that
affect the technique.

     Although the inorganic mass spectrum is far simpler and less complex than
atomic emission, it is not free of spectral interferences.  Spectral
interference occurs as an isobaric interference from either an isotope of
another element, that form singly or doubly charged ions of the same nominal
mass-to-charge ratio, or from polyatomic ions commonly formed in the plasma or
interface system from support gases and sample components.  The mass range
from 12 to 75 amu is the region most affected by polyatomic interferences.
The analytes of this region include the first series transition metals -
scandium through zinc.  Fortunately, the elemental and most of the common
polyatomic interferences have been identified.

     In Method 200.8 only two of the recommended analytical isotopes
experience isobaric elemental interferences.  Ruthenium interferes on Mo-98
and krypton on Se-82.  However, eight of the analytes can experience a
polyatomic interference.  These analytes are As, Cd, Cr, Cu, Mn, V and Zn with
a remote, but possible interference on Ag.  Again, the majority of these
analytes are first series transition metals and the polyatomic interferences
are well documented and noted in the method.  The third spectral interference,
abundance sensitivity, is a problem of resolution where the wing of a large
ion peak contributes to a small ion peak.  These occurrences can be minimized
by selecting the proper instrument operating conditions.

     The listed physical interferences also are best controlled by operational
measures such as: (1) using a peristaltic pump to control  and provide a steady
rate of sample introduction to the nebulizer, (2) the use of a mass flow

-------
                                       213
controller to provide accurate gas flow and uniform aerosol transport to  the
plasma, (3) limiting dissolved solids to 0.2 %  (w/v) to reduce the potential
clogging of the orifice of the interface cone assembly (4) the use of a water
cooled spray chamber to lower the water vapor content of the aerosol to reduce
oxide formation and (5) the use of internal standards to correct for matrix'
suppression.  Method 200.8 requires the use of  all these procedural measures
except the water cooled spray chamber which is  listed as a recommendation.

     Memory interference results when isotopes  of elements in a previous
sample contribute to the signals measured in a  new sample.  Memory effects  can
result from sample deposition on the sample and skimmer cones and from buildup
of the analytes in the spray chamber and sample uptake system.  Adequate
flushing of the system with a rinse blank reduces the effect.  Method 200.8
requires as a minimum, the use of a one minute  rinse time.  Also, the presence
of a memory effect may be assessed during analysis by noting a consecutive
drop in replicate integrations.

     On the next slide (#9) is a summary of Method 200.8 features.  The method
is applicable to both aqueous and solid matrices for the analysis of 20
analytes.  It should be noted that elements Ca, K, Mg, Na and Fe have not been
included in the method because they are normally present at relatively high
concentrations in environmental samples and can be determined accurately  and
more effectively by other techniques.  However, the elements thorium and
uranium are included because of the increasing  interest of these elements in
groundwater and because of the relatively good  sensitivity of these elements
by ICP-MS.

     The method provides similar acid digestion procedures for aqueous and
solid matrices.  This occurs when the remaining particulate material in an
evaporated aqueous sample is acid refluxed in the same manner as a solid
sample.  A combination of HN03 and HC1  acids is used in  both procedures.   The
resulting analyses are considered to be "total  recoverable" determinations.
     Prior to ICP-MS analysis, internal standards are added to an aliquot of
the digestate.  Method 200.8 requires the use of at least 3 internal
standards; however, the use of the 5 listed on  the slide is recommended and
was used in the single laboratory validation study.

     Outlined on the next two slides (#10 and #11) are the "total  recoverable"
sample preparation procedures for aqueous and solid samples, respectively.
Specifics included in the outlines are the size sample used, the acid
addition, the reflux time and dilution volumes.  Note that prior to ICP-MS
analysis an aliquot of the digestate is diluted further  to limit the dissolved
solids to 0.2% (w/v) and to control the chloride concentration to  a level  of
0.4% (v/v).  In each case, the analytical  determinations are multiplied by a
dilution factor to obtain the sample analyte concentrations.

     A combination acid "total recoverable" procedure was selected for sample
preparation because it requires less time,  is less labor intensive,  and past
studies have shown it to be equal  or superior to the "hard digestion"  (HNO,  +
H202) for solubilizing toxic and priority pollutant metals.  With the "total
recoverable" procedure insoluble oxides are less likely  to occur while the HC1
has been shown to stabilize Ag in  solution  and improves  the solubilization of
oxides and certain analytes such as Sb and  Cr.

-------
                                      214
     As a side note,  it may be  important to know, that in order to provide
uniform sample preparation for  the various spectroscopy techniques these same
digestion procedures  have been  incorporated into the recently revised  ICP-AES
Method 200.7 and the  new stabilized temperature graphite furnace atomic
absorption methodology, Method  200.9.
     On the next slide (#12) are shown the requirements of instrument
calibration.  After a 30 min warm-up period, a tuning solution containing 5
elements (Be, Co,  In, Mg and Pb) is used to check instrument resolution and
adjust mass calibration.  The Mg isotopes 24, 25, 26 are used to check
resolution for the low mass range while the Pb isotopes 206, 207, 208 are used
for the high mass range.  Mass  calibration is adjusted if it has shifted by
more than 0.1 amu from unit mass.  This solution is also used to check the
stability of the instrument.

     After tuning is  complete,  the instrument is calibrated using a
calibration blank and two composite standard solutions.  Before beginning
sample analysis, the  calibration is initially verified for all analytes by
analyzing a QC sample obtained  from an outside source.  The analytical result
must not exceed ±  10% of the established values of the QC sample.  The
calibration is then verified on a continuing basis by analyzing a standard
solution as surrogate sample at frequency of 10% with the determined value
being within the QC window of ± 10% of the true concentration.

     As mentioned before, at least three internal standards must be used and
their response monitored.  If their response falls outside the QC window of
+25% to -40% of the original response, the situation must be corrected.  This
may require only sample dilution or may entail termination of the analysis to
clean the sampling cone and/or retune the instrument.

     On the next slide (#13) are given the quality control  requirements of
Method 200.8.  For initial demonstration of performance,  method detection
limits or MOLs and the linear calibration range must be determined.
Procedures for these  two determinations are described in the method.   These
two determinations must be completed every six months or whenever there is a
significant change in background or instrument response (e.g. changing the
detector).

     For assessing laboratory performance,  a reagent blank and a laboratory
fortified blank are analyzed with each batch of samples.   If an analyte in the
reagent blank determination exceeds its MDL, laboratory or reagent
contamination should  be suspected.   Recovery of the analytes in the fortified
blank should be within the established control limits or 85 to 115%,  if
control limits have not been developed.  If recovery is outside the limits,
the analysis is out of control.  The analysis should be terminated and the
problem identified and corrected before continuing the analyses.

     To assess analyte recovery from the matrix,  10% of samples are fortified
with the same analyte concentration as that used in the fortified blank.   If
recovery from the sample matrix falls outside the designated range of the
fortified blank, a matrix affect should be suspected.

     The method detection limits for Method 200.8 are given on the next slide
(114).  The aqueous method detection limits are all  1 ng/L or less with the

-------
                                      215
exception of As, Se, V and Zn.  Method detection limits for solids are 1 mg/kg
or less with the exception of Se.  These, along with instrument detection
limits, are included as data tables in Method 200.8.

     The next slide (#15) shows the analytical scheme used in the single
laboratory validation study.  A matrix set of five waters and three solids
were used in the study.  The five waters consisted of drinking water, well
water, pond water, a sewage treatment primary effluent and an industrial
effluent.  The three solid samples were NITS sediment #1645, EPA
electroplating sludge #286, and EPA hazardous waste soil  #884.  For each
matrix, a total of five replicate sub-samples were analyzed in order to make
an estimate of precision.  Two further sets of duplicate sub-samples were
fortified with a multielement analyte mixture, one set a low concentration (10
Mg/L .to 50 M9/L for aqueous and  20 mg/kg for  solids), the other at high
concentrations (100 /xg/L to 200 /xg/L  for aqueous and 100 mg/kg for solids) in
order to assess element recoveries in a given matrix.  A total of nine samples
were therefore analyzed for each matrix with the total  number of samples 72.
The digestate or aliquots of the digestate were analyzed by both ICP-AES and
ICP-MS.

     On the next slide (#16) are given the instrument operating conditions
used in the single laboratory validation study.  These are the same operating
conditions that appear in Method 200.8 and Method 200.7,  respectively.  There
is an error in the ICP-AES listed conditions.  The time used for data
acquisition was a 16 sec integration period not 20 sec as shown.   Also, of the
ICP-MS operating conditions given, only the minimum of 3 replicate
integrations is required by Method 200.8.

     During the course of the single laboratory validation study,  spectral
interferences were experienced by both technniques in the analysis of certain
samples.  Those which were most significant are listed on the next slide
(#17).  The analysis of all samples by ICP-MS required correction of the
interference on As, V and Cr from the polyatomic ions of argon-chloride and
chlorine-oxide.  These corrections were anticipated and necessary because of
the sample preparation procedure used and the inherent concentration of
chloride in environmental samples.   Cu-65 was use for copper analysis of the
industrial effluent sample because the high sodium concentration  caused a
NaAr* interference on  Cu-63.   Also,  correction of MoO on  Cd-111 was necessary
because the industrial effluent contained approximately 1 mg/L Mo.

     The largest ICP-AES spectral interferences occurred in the analysis of
the sediment and sludge samples.   The interference was of an interelement
direct spectral overlap nature from high concentrations of Cr (#1645-28,000
mg/kg and #286-7500 mg/kg) and Fe (#1645-89,000 mg/kg).   The interferences
effects in these two samples were so severe that it prevented the analysis of
Th and Se in the sediment sample and caused the reporting of inaccurate data
for Sb in both the sediment and electroplating sludge samples.

     In the next slide (#18) are recovery data from the fortified primary
sewage treatment effluent.  Except for Ba and Th the recoveries are between
85-115%.  The low Ba recovery is attributed to the presence of sulfate.
However, the low Th recovery cannot be explained.  In the other four aqueous

-------
                                     216
samples, Th recoveries were acceptable.  The formation of insoluble phosphates
during sample processing has been suggested, but not confirmed.

     The next 5 slides (#19-23) are figures that contain ICP-MS/ICP-AES
comparison plots for the elements Cu, Mn, Ni, Ag, and V.  In each case the
center bar represented the mean concentrations of the five replicates and the
vertical bars represents two standard deviations of the replicates each side
of the mean.  The concentration of the analyte is noted on the y-axis, and the
sample matrix along the x-axis.  The ICP-MS data are noted by the wider bars.
These five elements were selected for presentation because of the available
data.  We will step through these slides for your observation without
additional discussion.

     These five figures will not appear in the proceedings of this meeting
because they are being published elsewhere.  However, these figures, along
with others, and a complete dicussion of the single laboratory validation
study has been prepared as a draft report by Dr. Stephen Long of Technology
Applications, Inc.  For those interested in obtaining a copy of the report see
me following this session or during the course of the conference.

     Given in this slide (#24) is a brief summary of the single-laboratory
validation study.  Approximately 3000 analytical determinations were
completed.  The matrix analyte background data were used to compare the two
spectroscopic techniques using a paired t test.  A null hypothesis approach
was used with the analyte means of the two techniques being compared at the 5%
level of significance.  For an analyte to be included in this comparison, a
requirement was set that data must be available for at least two of the
matrices.  For aqueous matrices, 11 analytes qualified; for the solid
matrices, 17 analytes qualified.  Only Se, Tl and U did not meet the
requirements.  These comparisons represented 860 determinations.  A
significant difference between the two techniques exists if calculated p value
is < the level of significance set at 0.05.  The lowest p value calculated was
0.08 for Zn in water.  Therefore, it would appear that the two techniques
provide data that are equivalent in quality for the range of concentrations
applicable to both techniques.

     The summation of the precision determined in the study was the following.
Intra-sample precision of the two techniques was similar being 1 to 2% RSD.
Although the overall precision appears to be sample limited, for the matrices
tested, the mean RSD was < 10% except for the hazardous waste soil.  Higher
RSDs for this sample were attributed to the inhomogenity of the sample
material.  Of the 298 spike duplicate determinations, only 8% had a relative
percent difference (RPD) that exceeded 10%.  If the hazardous waste soil
duplicates were not included in the calculation, the number would drop from 8%
to < 6%.

     Recovery of the fortified analytes was very good.  For the fortified
blank, recovery ranged from 94% to 107%.  For the fortified matrices < 10% of
the recoveries were outside the 85% to 115% limits.  As for specific analytes,
Sb gave good recovery in the aqueous matrices but low recoveries (55.4% -
81.2%) in the solid matrices.  Poor recoveries obtained for Ba in the sewage
effluent, hazardous waste soil and electroplating sludge have been attributed
to precipitation as the sulfate.  Low recoveries of Th in the sewage effluent
and hazardous waste soil cannot be explained at this time.  However, all low

-------
                                      217
recoveries appear to be matrix related and with the exception of these three
elements, recovery data indicates adequate method performance for the range of
elements and matrices study.
     On the last slide (#25) is a summary description of the joint USEPA/AOAC
multilaboratory ICP-MS study currently being conducted.  This project was
started in early December, 1989 after concurrence by AOAC that a joint effort
multilab study of Method 200.8 would be of benefit to the analytical community
doing environmental analyses.  The study design that you see here was
completed and agreed upon by both parties in early January, 1990.  Samples and
ampul concentrates were shipped to the 18 participating laboratories on March
22, 1990, and as of May 3, seven of the participants had already forwarded
data packages.

     In addition to determining sample background concentrations, sample
aliquots are to be fortified as Youden pairs with analyte concentrations
necessary to collect the data required.  For determinations in reagent, tap
and ground water this is to be done at three levels of concentration.  The
participants will use their own reagent water and a local drinking water for
sources of these two types of water.  For the ground water determinations each
laboratory has received for analysis one of the waters collected from five
different sources.  Besides the reagent, tap and ground water determinations,
eight of the laboratories will receive for analysis a wastewater digestate and
also will analyze a wastewater of their choice.

     The data from this study will be compiled by EPA and presented to the
AOAC Executive Board for approval along with a copy of the method in the AOAC
format.  The project and final report are scheduled to be completed by October
1990.  The results of the study are to be presented as a poster paper on
September 12 at the annual AOAC meeting in New Orleans, LA.

     For more detailed information concerning this study, please contact James
Longbottom of the Quality Assurance Division of EMSL-Cincinnati.

     This concludes my presentation.

-------
 SINGLE  LABORATORY  EVALUATION OF METHOD 200.8
DETERMINATION  OF TRACE ELEMENTS IN WATERS  AND
       WASTES  BY ICP MASS SPECTROMETRY


                  Theodore D. Martin
       Environmental Monitoring and Systems Laboratory
                         and
                    Stephen E. Long
               Technology  Applications Inc.

       UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                 CINCINNATI, OHIO 45268
to
f->
00

-------
                                                           SLIDE
          INDUCTIVELY COUPLED PLASMA
               MASS SPECTROMETRY
ADVANTAGES
1. LINEAR RANGE 5 - (8) ORDERS
2. PPT  DETECTION  LIMITS
3. 'SIMULTANEOUS"  DETERMINATION  OF PERIODIC TABLE
4. ISOTOPE RATIO INFORMATION
  ISOTOPE DILUTION CAPABILITY
5. FACILITY FOR QUALITATIVE ANALYSIS
6. RAPID SAMPLE THROUGHPUT
to

-------
                                                                  SLIDE * 3
VACUUM  tTA««t
                                                                         to
                                                                         to
                                                                         o

-------
                    221
                                           SLIDE #
        ICP-MS INTERFERENCES
SPECTRAL
  1.  ISOBARIC ELEMENTAL - Mo,  Se
  2.  POLYATOMIC IONS  -  8  analytes
  3.  ABUNDANCE SENSITIVITY - Resolution

PHYSICAL
  1.  VISCOSITY - Peristaltic pump
  2.  AEROSOL TRANSPORT  - Mass  flow  control
  3.  DISSOLVED SOLIDS - Limit 0.2% (w/v)
  4.  OXIDE FORMATION  -  Cooled spray chamber
  5.  SUPPRESSION - Internal standards

MEMORY  - ANALYTE  BUILDUP
  1.  SAMPLE UPTAKE SYSTEM
  2.  CONE DEPOSITS
  3.  DROP IN REPLICATE INTEGRATIONS

-------
I
                  METHOD 200.8 :  SUMMARY FEATURES
                                                                               SLIDE # 9
              SCOPE  & APPLICATION



              ELEMENT COVERAGE



              SAMPLE DIGESTION


              DISSOLVED SOLIDS


              INTERNAL STANDARDS
: Ground, Surface and Drinking Waters
  Wastewater,  Sludge and Solid Waste

:  Al,  Sb, As, Ba, Be,  Cd, Cr, Co, Cu,  Pb
   Mn, Mo, Ni, Se, Ag, Tl, Th, U, V,  Zn

:  Nitric,  Hydrochloric  acid

:  0.2% Limit

:  Sc, Y,  In, Tb, Bi
to
to
to

-------
                                                    SLIDE # 10
SAMPLE PREPARATION - TOTAL RECOVERABLE
1. AQUEOUS

 100ml sample  +  1ml  cone, nitric
                0.5ml cone. HCI

 Heat  to reduce volume  to  15ml. Reflux for  30min.
 Cool  and  dilute to 50ml.
                                       to
                                       M
                                       GJ
 Dilute  20ml
50ml and analyze.
 0.4% HCI, 0.8% nitric. Dilution factor 1.25.

-------
                                                       SLIDE I 11
SAMPLE PREPARATION - TOTAL RECOVERABLE

2.  SOLIDS

 1g sample + 4ml (1+1) nitric
             10ml (1+4)  HCI

 Reflux for 30min. Cool and dilute to 100ml.
 Centrifuge or allow to stand  overnight.
 Dilute 10ml
50ml and analyze.
 0.4% HCI,  0.4%  nitric.  Dilution factor 0.5
                                          to
                                          to

-------
                                                                  SLIDE #12
      METHOD 200.8 :  CALIBRATION
DEMONSTRATION  OF INITIAL CALIBRATION
• Tuning Solution  (mass  calibration)
• Quality Control  Sample QCS
CALIBRATION VERIFICATION
• Standard run as surrogate  sample at 10% frequency
  QC Window ±  10%  of true value
INTERNAL STANDARDIZATION
• Minimum of  Three Internal  Standards
  QC Window  + 25%  of Original Response
              - 40%
to
to
Ln

-------
                                                              SLIDE I 13
     METHOD 200.8 : QUALITY CONTROL
1. INITIAL DEMONSTRATION OF PERFORMANCE
  •  MDL,  Linear Calibration Range
2. ASSESSING LABORATORY PERFORMANCE
  *  Reagent Blank (LRB) - 1 per batch
  •  Laboratory Fortified Blank (LFB)  -  1 per batch
3. ASSESSING ANALYTE RECOVERY
  •  Laboratory Fortified Sample  Matrix  - 10% of samples
to
CO

-------
                                                              SLIDE # 14
TOTAL RECOVERABLE METHOD DETECTION LIMITS
  ELEMENT
    Al
    Sb
    As
    Ba
    Be
    Cd
    Cr
    Co
    Cu
    Pb
    Mn
    Mo
    Ni
    Se
    Ag
    Tl
    Th
    U
    V
    Zn
MASS
  27
 121
  75
 137
   9
 111
  52
  59
  63
 208
  55
  98
  60
  82
 107
 205
 232
 238
  51
  66
AQUEOUS
 (ug/l)

  1.0
  0.4
  1.4
  0.8
  0.3
  0.5
  0.9
  0.09
  0.5
  0.6
  0.1
  0.3
  0.5
  7.9
  0.1
  0.3
  0.1
  0.1
  2.5
  1.8
SOLIDS
(mg/kg)

 0.4
 0.2
 0.6
 0.4
 0.1
 0.2
 0.4
 0.04
 0.2
 0.3
 0.05
 0.1
 0.2
 3.2
 0.05
 0.1
 0.05
 0.05
 1.0
 0.7
10

-------
                                                         SLIDE | 15
  ICP-MS METHOD 200,8 SINGLE LAB VALIDATION
        WATERS
                  8x6
                                3x8
  ANALYZE ICP-MS
   METHOD 200,8
                      TECHNIQUE
                      COMPARISON
  PRECISION AND
ACCURACY STATEMENT
  SOLIDS
s
SPIKE

^*

REAGENT BLANK
LFB
6 x (3+2)
y
DIGEST
' SPLITS >
r ^
SPTKF

V
ANALYZE ICP-ES
 Total Samples 72
                                                              to
                                                              to
                                                              CO

-------
                                                                   SLIDE # 16
                  METHOD 200.8 VALIDATION
            INSTRUMENT OPERATING CONDITIONS
Instrument
Forward power
Coolant flow rate
Auxiliary flow rate
Nebulizer flow rate
Solution Uptake

Repl. Integrations
Data acquisition
   ICP-MS

VG PlasmaQuad Type I
1.35 kW
13.5 l/min.
0.6 l/min.
0.78 l/min.
0.6 ml/min.
320 us dwell
2048 channels
85 sweeps
    ICP-ES

Jarrell-Ash AC 1160
1.1 kW
19 l/min.
0.3 l/min.
0.63 l/min.
1.2 ml/min.
    integration
48 channels
to
to
10

-------
                                                        SLIDE I 17
SPECTRAL INTERFERENCES 200.8 ELEMENT SUITE
   ICP-MS
      1. Chloride on As-75, V-51, Cr-52
      2. Na(Ar) on Cu-63 (effluent)
      3. Mo(O) on Cd-111  (effluent)
•  ICP-ES
       1. Cr on Sb 206.8 nm (NIST 1645)
       2. Cr on Sb 206.8 nm (EPA 286)
       3. Cr on Th 283.7 nm (NIST 1645)
       4. Fe on Se 196.0 nm (NIST 1645)
                                                              to
                                                              U)
                                                              o

-------
SPIKE RECOVERIES - PRIMARY EFFLUENT
                                              SLIDE # 18
MENT BKGRD CONG. SPIKE RECOVEI

Sb
As
Ba
Be
Cd
Co
Pb
Ni
Se
Ag
Tl
Th
U
V
Zn
(ppb)
1.5
<1.4
202
<0.3
9.2
13.4
17.8
84.0
<7.9
10.9
<0.3
0.11
0.71
<2.5
163
(ppb)
10
50
50
10
10
10
10
10
50
50
10
10
10
50
50
(%)
95.7
104.2
79.2
110.5
101.2
95.1
95.7
88.4
112.0
97.1
97.5
15.4
109.4
90.9
85.8
                                                   NJ
                                                   U)

-------
                                                           SLIDE I 24
       VALIDATION STUDY SUMMARY
• ANALYTICAL DETERMINATIONS - 3000

• MATRIX COMPARISON ICP-MS /  ICP-AES
     1. Paired t test,     p •  0.05 for n > 2
     2. Aqueous (11)  -  Solids (17) 860 determinations
     3. Aqueous Zn, p • 0.08

• PRECISION
     1. Intra-sample RSD  1-2%
     2. Matrices, mean RSD <10%
     3. 8% RPD >10% -  298  determinations

• RECOVERY
     1. Fortified blanks  94-107%
     2. <10% Fortified matrices >85-115%
     3. Sb, Ba and Th
to
OJ
to

-------
                                                 SLIDE # 25
 JOINT USEPA/AOAC MULTILABORATORY STUDY

• STUDY PERIOD, DEC  1989  TO OCT 1990

            STUDY DESIGN
  WATER SOURCES
YOUDEN  PAIRS
  REAGENT
  TAP
  GROUND (5)
  WASTE (Digestate) *
  WASTE (of  choice)*
      3
      3
      3
      2
      1
BACKGROUND
   DATA
  PARTICIPATING LABORATORIES  18  -  (8)
to
CO
CO

-------
                             234
     MR. FIELDING:  Our next speaker is Jim Rice who will
talk about ICP performance for the measurement of 14 trace
metals in power plant waste streams.
     Jim?

-------
EPRI
                                235.
                        13th Annual EPA Conference on
                    Analysis of Pollutants in the Environment
                             Norfolk, Virginia
                              May 9-10,1990
            ICP PERFORMANCE FOR THE MEASUREMENT OF

                       FOURTEEN TRACE METALS

                   IN POWER PLANT WASTE STREAMS
                                   By
                         Raymond F. Maddalone, PhD
                                TRW, Inc.
                          Redondo Beach, California

                             James K. Rice, PE
                            Consulting Engineer
                             Olney, Maryland

                                  and

                             Winston Chow, PE
                        Electric Power Research Institute
                            Palo Alto, California

-------
                                   236
                    ICP PERFORMANCE FOR THE MEASUREMENT OF
                 14 TRACE METALS IN POWER PLANT WASTE STREAMS
                                        BY

                             Raymond F. Maddalone, PhD
                                     TRW, Inc.

                                 James K. Rice, PE
                                     Consultant

                                        and

                                 Winston Chow, PE
                           Electric Power Research Institute
                                     ABSTRACT
     A 26 laboratory study by the Electric Power Research Institute determined the
performance of an Inductively Coupled Plasma Atomic Emission Spectroscopic (ICP) method
for measuring 14 elements (Al, Ba, Be, B, Cd, Cr, Cu, Fe, Pb, Mn, Mo, Ni, V, Zn) in typical
power plant waste streams. The resulting single operator and overall precision were used to
compute limits of detection and quantitation for each of the elements in each matrix. These
limits of detection are compared to published EPA detection limits. In addition, an
algorithmic approach was used to compare ICP precision and recovery results from the
utility sample matrices to those obtained in reagent grade water samples. From this
analysis the presence or absence of chemical matrix effects were investigated.

-------
                                     237
BACKGROUND AND OBJECTIVES OF THE ANALYTICAL METHODS  QUALIFICATION PROJECT

The utility power Industry  1s required under the  Federal  Mater Pollution
Control  act to Monitor their discharges for a number of parameters.   As a
result of the 1976 Consent  Decree  (National Resources Defense Council ver-
sus Train), the EPA was  required to establish effluent limitation
guidelines, pre-treatment standards, and also new source  performance stan-
dards for 65 pollutant classes  (129 specific priority pollutants).   In an-
ticipation of a requirement that utility discharges  be monitored for some
or all of the priority pollutants, EPRI Project RP 1851-1 was initiated
both to collect concentration and  frequency data  about utility aqueous dis-
charges and to assemble  a set of sampling and analysis guidelines for the
species of interest in those discharges.  Utility chemists would then have
a firm basis to select methods  to  monitor the species of  Interest In those
discharges.

The data bases produced  in  Phase I of EPRI Project RP 1851-1, Sampling and
Analysis of Utility Pollutants  (SAUP), contain both  the average steam elec-
tric power plant discharge  concentrations for all conventional, nonconven-
tional,  and priority  pollutants (1) as well as a  comprehensive compilation
of precision and bias data  on the  methods used to monitor the pollutants
(2).  In developing this latter data base, it was found that the validation
data were often obtained for matrices and concentrations  not representative
of the steam electric Industry.  The lack of an Independent data base de-
rived from utility laboratory performance In typical  discharge matrices
placed individual utilities at  a disadvantage during siting and permit
negotiations.  At the same  time, there was a trend toward setting effluent
discharge limits at the  end of  pipe or edge of the mixing zone at the water
quality criteria concentration.  Many of these water quality criteria are
at or below the detection limit of current analysis  methods.  In order to
determine whether current monitoring methodology  Is  capable of detecting
pollutants at the EPA detection limit or the water quality criteria
concentration, EPRI instituted  a project under RP 1851-1  entitled Utility
Aqueous Discharge Monitoring -  Analytical Methods Qualification (AMQ).

The primary objective of the AMQ Project is to collect precision and bias
data for methods used to determine selected parameters and elements  in util-
ity discharge streams.   The required analytical data are  collected through
collaborative testing using representative utility laboratories.  The lim-
its of detection and  quantitation  derived from this  project thus provide a
realistic estimate of the capabilities of EPA-approved methods In the com-
pliance monitoring situation under utility process stream and laboratory op-
erating conditions.

The specific AMQ project goals  are to:

    •    Establish both  the single operator and overall precision that
         can be expected from utility laboratories utilizing the se-
         lected compliance  methods.

    •    Establish whether  any  statistically significant  bias exists
         in a given matrix  for  the methods validated.

-------
                                238
    •    Determine the expected  limits  of detection and quantitation
         of compliance methods when  used by a utility analyst for util-
         ity matrices.

Table I provides an overview of  the  entire AHQ project.  It 1s divided Into
four parts In order to Minimize  the  Impact on the participating Industry
laboratories and to permit modification of the test design to reflect chang-
ing environmental Issues or regulatory  requirements.  The elements and pa-
rameters In each part were selected  on  the basis of their Importance to the
utility Industry, regulatory Interest,  and a comparison of discharge and in-
take concentrations.  Under the  AMQ-III validation program, the subject of
this paper, Inductively Coupled  Plasma  (ICP) was validated for 14 metals in
6 different utility matrices.

AMQ-III TEST METHODOLOGY

The AHQ project is based on the  premise that a method qualification project
should use matrices representative of those encountered by the analyst in
routine work.  Furthermore, since the shipment and storage of samples is
part of the normal analysis procedure at most laboratories and comparison
of the results from different laboratories on split samples is often encoun-
tered in compliance monitoring,  the  test program should Include spiking the
matrices and sending aliquots to each participant.  A detailed review of
the AHQ test and analysis protocols  can be found in references (1,1).  A
summary of the test program protocols 1s provided in the following
paragraphs.

laboratory Selection

The participation of utility laboratories was solicited through the offices
of EPRI and the Utility Water Act Group (UWAG).  Twenty laboratories were
eventually enlisted to validate  the  AMQ-III elements.

Laboratories were contacted before the  test to determine which labs had the
experience or interest in determining the elements in the seawater
samples.  These laboratories participated In a pre-test study to assist in
selecting test concentrations, assessing sample stability, and providing
the operators with some experience with those difficult matrices.  The
QA/QC vial was used as an absolute,  Independent measure of the bias associ-
ated with an individual laboratory.

Matrices Tested

A key component of this validation effort was the use of typical utility
matrices.  To that end, sufficient quantities of river water, ash pond
overflow, seawater Intake, seawater  discharge (seawater with a proportional
amount of fireside wash added to simulate routine plant discharges), and
treated chemical metal cleaning  wastes  (TCHCW)  were obtained from utility
sources.  Following the overall  objective of simulating typical NPDES
monitoring, the test matrices were homogenized, spiked and then split.  In
addition to these samples, spiked reagent grade water and a QA/QC vial sup-
plied by the ERA (Environmental  Research Associates) were also sent to

-------
                                     239
participating laboratories.  All laboratories were required to analyze the
reagent grade water and QA/QC vial while the analysis of the TCMCW sample
was optional.

Test Concentration Selection

Depending on the background concentration, every effort was made to have
the lower test concentrations near the expected detection limit of the
method.  This requirement was difficult to meet in some of the matrices,
due to the high background.  The remaining test concentrations were select-
ed to stay within the published (£) estimate of the optimum analytical
range, if possible.  A total of 4 test concentrations (neat and three
spiked) were prepared for each matrix and element.

Sample Preparation

Samples were prepared (1,1) by filling a polyethylene 120 L churn (Figure
1) splitter with the acid (HN03) stabilized test matrix.  The volumes
were determined by weight and density measurements and were accurate to
±100 ml based on the calibrated accuracy of the scale.  After the volume
was determined, weighed amounts of single element 1000 ug/L certified AAS
standards were added to the churn splitter using Teflon beakers.  After the
addition of a spike, the churn was operated for 5 minutes before samples
were removed from a spigot at the bottom of the churn.  Each sample consist-
ed of a 500 ml aliquot in a precleaned (HC1, HN03, high purity water)
polyethylene bottle.

OA/OC Activities

A comprehensive QA/QC program was part of the sample preparation and dis-
bursement effort.  Internally, the sample spiking and disbursement activi-
ties were guided by a test plan whose adherence was audited by personnel
not associated with the test program.  In addition, test samples from each
matrix were selected at random and analyzed prior to shipment to ensure
their proper preparation.  Samples were also analyzed by GFAAS to determine
the stability with time.  Based on the seawater pre-test, it was found that
barium could not be stabilized in the seawater and TCMCW matrices at the
test concentrations.  Consequently, barium data from those matrices are not
included in the results.

Test Procedures and Reporting Requirements

Each participating laboratory was sent a detailed test protocol describing
the sample digestion and analytical procedures.  The test protocol con-
tained a copy of EPA Method 200.7 and detailed instructions on the report-
Ing requirements.  Each sample was analyzed in duplicate and the results re-
ported to the TRW test coordinator on standardized reporting forms or on a
hard-copy computer output from the ICP.   A QA/QC survey, covering equipment
used, background/experience of the operator, and laboratory practices, was
included in this package.

-------
                                240
Statistical Analyses

The design of the validation program was based on ASTH 02777-85 using repli-
cate analyses.  The data from the test program was reduced using a PC-based
statistical program, which implemented the data analysis and outlier test
protocols in ASTH 02777-85.  All  data received from the participants was
Initially screened for values greater than 5x and less than 1/5 of the ex-
pected true values.  Laboratories exceeding those limits were asked to
check their calculations and reports, but were not asked or permitted to
re-analyze their samples.  On this error-free data set, the statistical pro-
gran performed the lab ranking and Individual outlier testing to screen the
data.  The program then computed  the single operator tnd overall precision
and recovery at each test concentration tlong with Its linear tnd curvilin-
ear regression equation for precision and recovery.   Figure 2 1s in example
of overall precision versus true  value.

The true concentration was computed by taking the mean of the reported re-
sults for the lowest test level and then adding the known amounts of each
element added during the sample preparation effort.   In most cases for a
given element, the lowest test level  was the as-received matrix.  For se-
lected elements and matrices, small quantities of the element were added to
the background concentration to raise their lowest test concentration near
the expected ICP detection limit.   This approach was taken based on previ-
ous test experience to avoid data drop-out due to concentrations below an
element's detection limit.

CALCULATION OF THE LIHITS OF DETECTION AND QUANTITATION

Background

An important consideration in selecting an analytical  method 1s its limit
of detection (LOD).  Over the years,  several  definitions for limit of detec-
tion and its companion concept, limit of quantitation (LOQ), have evolved
(Table 2).  Under Phase I of RP1851-1,  a methodology for computing an LOD
was developed utilizing open literature precision data.  That approach will
be discussed and its problems noted.   An alternate approach utilizing the
greater amount of data that was derived from the AHQ validation studies
will be presented.

Definition of Limit of Detection

In reviewing the sources of LOD definitions (Table 2)  several  areas of
agreement are apparent:

    I.  The consensus is that the  LOD can be  defined as some factor
        times the standard deviation  of the blank.

    2.  The factor chosen depends  on  the risk level  of making false
        positive identification of zero values and false negative Iden-
        tification of nonzero detectable values.

    3.  The blank should be well characterized (>10  replicates).

-------
                                     241
    4.  The blank sample mist be processed exactly as the actual
        sample.

The key questions remain:  What  is  the value for the factor and what type
of standard deviation (single operator,  SQ,  or overall,  ST) should be
used?  Based on  the definitions  in  Table 2,  the consensus value for the
factor appears to be at or near  3.   The interpretation of this value in
terms of a confidence level depended on the point of view of the reader.
If one is only concerned about false positive identifications of zero
values, then the 3 sigma limit provides <1% risk of making a false positive
identifica- tion.  The actual risk  Involved depends on the number of
replicates used  to calculate it. Should one also be concerned with the er-
ror of not identifying detectable levels of a species, then the 3 sigma LOD
represents a 7X  chance for both  false positive and false negative
responses.  The  7X risk is for a large number of blank replicates, and the
risk increases with a decrease in the number of replicates.  It should be
noted that most  regulatory limits consider only the chance for false posi-
tive errors.  In those cases, 3  sigma represents a risk level of <1X of
falsely identifying a species as present when It is not.

The Environmental Monitoring and Support Laboratory (EMSL) has published an
approach to LOD  calculation stressing the complete analytical procedure,
which includes specifying a matrix, an analytical procedure (calibration
through the evaluation of results), and a particular instrument.  EMSL
proposes the concept of method detection limit (HDL) to describe processing
a sample through all the steps comprising an established analytical
procedure.  The  HDL is defined as "minimum concentration of a substance
that can be identified, measured, and reported with 99% confidence that the
analyte concentration is greater than zero"  (10.).  The HDL is determined
from replicate analyses of a sample in a given matrix containing the
analyte at concentrations 1 to 5 times the estimated HDL.  Two approaches
are given (£); one employs seven replicate measurements and is expressed
as:
   HDL
t(N-I, 1- a) * sc
(1)
where N » 7, « « 0.01 and S- is the standard deviation of the seven
replicates.  Substituting the appropriate one tail  "t" value, equation 1
becomes:

   HDL - 3.143 (Sc)                                                      (2)

An alternative approach tests the reasonableness of the HDL estimate by us-
ing an Iterative process (2 sets of seven replicates).  The HDL is then de-
fined as:

   HDL « 2.8681 x SpooT                                                 (3)

where SDOQI is the pooled Sc for the two sets of seven replicates and
2.681 13 the value for a single tail t/j* n 0.1)• The HDL approach has
been incorporated in promulgated (49 FR 43430,'Friday, October 26, 1984)
regulations.

-------
                                242
laplicit in the LOD discussions by all  the quoted authors except Rice was
the assumption that single operator precision data are used to calculate
the LOD.  The reasoning, as expressed by the ACS (S),  is that the analyst
needs to report his/her detection limit.   From a regulatory standpoint,  in-
terlaboratory (overall) precision data may be more applicable, since split
saaples are involved.  In our earlier studies (£), LODs were calculated  us-
ing both single operator, S0, and overall,  Sj, data.

An estimate of the S0 and Sj values at zero concentration can be ob-
tained from the y-Intercept of the precision versus mean test concentration
linear regression equations.  The drawback of this approach was twofold.
First, precision is not a linear function of concentration all the way to
zero (Figure 3).  In practice, the precision tends to  flatten to a value
that could be equated to the Instrument noise and effects due to a given
natrix.  A linear regression of precision data, especially with data points
far from zero, can underestimate the precision at zero and produce negative
intercepts.  Negative intercepts were often obtained in earlier studies
(£»!>£) and rendered the computation of an LOD impossible.

Recognizing this problem, the statistical  program used in the AMQ-I and  -II
validation efforts was modified to compute curvilinear regression equations
in the form:
   S « ab1
(4)
where T is the estimated true value at each  test level.   These curvilinear
equations were found to fit the data from the  test program much better than
the linear regression equations.

Calculation of the Limit of Detection

The first step in calculating the LOD 1s  to  use  the precision versus true
value and recovery regression equations;  they  are used to compute a preci-
sion versus mean concentration equation.   The  resulting equations are used
to compute the limit of detection in the  following manner.

Starting with the EPA definition of an HDL (1P_,  H), the regression equa-
tion for precision based on the mean concentration will  be substituted:
   K)L - t*S
(5)
    S - nx + b
(6)
where t is the students t,  S the standard deviation,  and x is the mean
concentration.  Therefore:
    HDL - t(mx +b)
(7)

-------
                                     243
In order to satisfy the EPA requirement that the precision be obtained at a
concentration 1 to 5 tines an estimated MDL, x will be set equal to the HDL
or
    HDL * t(m*MDL + b)
    LOD - tb/(l-tni)
(8)
(9)
where t 1s the "t" value for the degrees of freedom of the precision versus
Bean concentration regression (related to the number of libs for single op-
erator precision or the total number of data points for overall precision).
Equation 6 can be solved directly to obtain an LOO.  Using the same substi-
tution approach, the curvilinear equation Is:

    (LOD)/(bLOD) - ta                                                  (10)

Equation 10 cannot be solved directly; but, using an Iterative approach
found In a commercial software package (TKSolver Plus), a root can be
obtained.

Definition of Limit of Ouantltatlon

In our review (£) of the literature definitions for limit of quantltatlon,
several areas of agreement were noted:

    I.  The LOQ 1s equal to a factor times the standard deviation of
        the blank.

    2.  The factor 1s related to the expected/required precision at
        the LOQ.

As with the LOD calculation, establishing the factor and choosing the type
of standard deviation are the key Issues.

Compared to the LOD, the LOQ Is established 1n a less rigorous manner.  Cur-
rle (£) and the ACS (9.) define the factor as Inversely proportional to the
relative standard deviation (RSD) at the LOQ.  The ACS (£) has chosen ±10%
RSD as the precision expected/required at the LOQ.  The choice of ±10% Is
arbitrary and does not accurately reflect the day-to-day precision that an
analyst finds at ppb levels.  In TRW's review of the literature validation
data (£), RSDs were calculated at the primary drinking water standard and
at the average power plant discharge concentration using the overall preci-
sion linear regression equations for Flame AAS and GFAAS.  The average of
all the RSDs computed was 23.5%.  This value represents the best generally
attainable precision at levels which are encountered In real world
analyses.  Based on those data, expecting an RSD of ±20% at trace levels
seems reasonable and defensible and will be used 1n the evaluation of the
ANQ data.
                                   8

-------
                                244

Rice (12) has suggested that the LOQ  be defined as "the lowest true concen-
tration for which the relative overall  precision is 20%."  This definition
parallels the EPA's Practical Detection Limit.

One approach to fulfill that definition is to use the computed linear re-
gression equations for the precision  versus mean.  The regression equations
are in the form of a linear equation:

    S - mx + b                                                         (11)


Converting the S to relative standard deviation by dividing through by x:
    S/x « m + b/x


Setting the S/x - 0.2 and solving for x:
    x - b/(0.2-m)
(12)
(13)
or
    LOQ - b/(0.2-m)                                                    (14)

Using the curvilinear precision equation,  the resulting equation would be:

    bx/x - 0-2/a
or
    (bLO{*)/(LOQ) - 0.2/a
(15)
TKSolver Plus was also used to solve equation 15.
By taking this approach,  the actual  concentration which produces an RSD of
20% can be found.  For comparison, both  single operator and overall preci-
sion regression equations were used  to produce a value for the LOQ.  Nega-
tive LOQs are computed from the linear regression data when either the
slope exceeds 0.2 or when the intercept  is  negative and the slope is less
than 0.2.  In the former  case, a slope of >0.2 indicates that an RSD of
±20% was never obtained over the test range.

RESULTS

The linear and curvilinear precision (single  operator and overall) versus
mean concentration equations were computed  using the respective precision
versus true value and the recovery regression equations.  Using the method-
ology described, the linear and curvilinear LODs and LOQs by precision type
and matrix were computed.  Those preliminary  results can be found in Table
3 together with the EPA detection limit  and the WQC.

-------
                                     245
Table 3 shows the problems  associated with  using a linear regression equa-
tion to estimate LODs.  Most of the tine the  linear regression precision
equations produced negative LODs, probably  as the result of the test range
being too far away from the detection limit.   Even though every attempt was
made to have the lower  test concentrations  near the detection limit, the
background concentration  of some elements in  some matrices did not permit a
test range as low as we would  have liked.   The curvilinear precision regres-
sion equations were less  sensitive to this  problem and,  in most cases, were
able to produce a positive  LOO.  Another problem with using the linear re-
gression precision equations was negative slopes.  In this case, the data
is probably poorly correlated,  since is would be quite unlikely that the
precision (standard deviation  of mean) would  increase with decreasing
concentration.  Consequently,  no LODs were  computed from linear regression
precision equations having  a negative slope.

Similar problems with the calculation of an LOQ from the linear regression
precision equations were  also  seen.  If the slope was >0.2 (Indicating that
an RSD of 20% was not obtained over the test  concentrations) or the Inter-
cept was negative, LOQs could  not be computed from the linear regression
equations.  Once again, the curvilinear regression equations seemed to pro-
duce results much more  reliably than the linear equations (Table 3).

As part of the final review effort, the LODs  in Table 3 will be reviewed to
determine whether they  will meet EPA guidelines (H) for computing an MDL.
As such, the results in Table  3 should be considered preliminary pending
the formal review of the  RP 1851-1 Project  Advisory Committee.

Comparison of EPRI AHO-III  LOP and LOQ to the EPA Detection Limit

The preliminary single  operator and overall precision based EPRI LODs and
LOQs from the curvilinear regression equations were ratioed to the EPA de-
tection limit in Table  3.  To  further highlight this comparison, the ratios
were collated by matrix and plotted as a bar  chart.  Where there was no
data from the curvilinear regression equation,  the linear regression result
was used.  The overall  precision based LODs and LOQs are presented here
since they represent the  situation (comparison of multiple laboratory data
from split samples) found in most NPDES monitoring audits.  In all cases, a
ratio less than one indicates  that the EPRI LOD or LOQ was lower than the
EPA detection limit. A ratio greater than one indicates that the EPRI LOD
or LOQ was greater than the EPA detection limit.   For plotting and Interpre-
tation convenience, the y-axis  (ratio) Is a log scale.

Reagent Grade Water. Reagent grade water should be the  closest to the EPA
detection limits since  the  EPA  ICP detection  limits (£)  are the concentra-
tions which produced a  net  analytical signal  three times the standard devia-
tion of the background  beneath  the spectral line.   In application of this
definition, the resulting detection limits were based on single operator re-
agent grade water data.  However, except for  cadmium,  all  the EPRI computed
LODs are above EPA detection limit (Figure 4).   Host of  the EPRI LODs are
greater than a factor of  3  higher and, for beryllium,  Iron,  and zinc,  the
LODs were greater than  10 times higher than the  EPAs.  The EPRI LOQs fol-
lowed the same pattern; but, since they were  usually 2 to  4 times higher
than the LODs, most of  them were 7 times higher  than  the EPA detection
limit.
                                 10

-------
                                246
River Water.   Based on the chemical characterization  studies  done on each
matrix during the test program, river water could  be  considered the next
most complex  matrix. In this case (Figure 5), all  of the EPRI  LODs are
greater than  3 times the EPAs and five elements  (boron,  beryllium,
manganese,  nickel, and zinc) are 10 times larger than the EPA detection
Unit.  In  most cases, the EPRI LOQs were 7 times  higher than the EPA detec-
tion Unit.

Ash Pond Water.  The ash pond water sample was taken  at  the same plant as
the river water and consists of the overflow from  ash pond.   It had a
higher IDS  and sulfate content then the river water.   Both cadmium and chro-
niun were below the EPA detection limit (Figure  6).   The remaining elements
were 2 tines  higher than the EPA detection limit.   A  substantial  number of
those eleaents (barium, beryllium, boron, Iron,  Manganese, nickel,  and
vanadium) were 10 times greater than the EPA detection limit.  All  the EPRI
LOQs were greater than the EPA detection limit and most  (except cadmium)
were 8 or core times higher than the EPA detection limit.

Seawater Intake and D1 scharoe.  Both seawater matrices,  as expected,  pro-
duced similar results (Figures 7 and 8).  All EPRI  LODs  (except for copper
in the seawater intake) were 10 times higher than  the EPA detection limit.
Host of the element's LODs were 20 times greater than the EPA detection
limit.  Boron and cadmium exhibited LODs 100 times higher than  the EPA de-
tection limit in one or both of the matrices.  Comparison of  the EPRI LOQ
to the EPA  detection shows an even worse situation as boron,  beryllium,
iron, nickel, and zinc LOQs were 100 times higher  than the EPA  detection
Unit.

Treated Chemical Hetal Cleaning Wastes, fTCHCWK  The  TCMCW sample was simi-
lar in composition (high sulfate and chloride) to  the seawater  samples, but
with the added organic flocculating agents to remove  iron and copper.  The
LODs computed from the precision data taken from this matrix  are not as bad
as the seawater samples, but show only 3 elements  (cadmium, copper, and
vanadium) having an LOD less than 10 times the EPA detection  limit (Figure
9).  All elements having data were 4 times higher  than the EPA  detection
limit.  Only  the LOQ for aluminum was less than  10 times the  EPA detection
Unit.

Symnarv of  EPRI LOD Results.  Figure 10 shows the  percentage  of reporting
eleaents having an LOD between 0.1 to 1, 1 to 10,  10  to  100,  and 100 times
greater than  the EPA detection limit.  First, this  three-dimensional  bar
chart shows that, though different, the results  from  the reagent grade wa-
ter (RGW),  river water (RH), and ash pond overflow (APO) are  comparable.
On the other  hand, the seawater Intake and discharge  (SU-I and  SH-D)  and
the TCNCU have similar results.  Secondly, the performance, based on In-
creasing LOD, is much worse in the complex matrices.   Thirdly,  even though
the LODs in the RGW matrix were the lowest of the  matrices studied, almost
70% of them were between 2 to 9 times higher than  the EPA's and a signifi-
cant (23%)  were 10 tines higher than the EPA's.

At the minimum, these results show that detection  limits will vary by
matrix, and that fact should be considered when  a  discharge or  performance
Unit is set.
                                  11

-------
                                     247
Comparison of EPRI AMO-III LCDs  to  EPA Hater Quality Criteria

Water Quality Criteria (WQC)  represent another benchmark that utility labo-
ratories are being required to meet.   A similar comparison of EPRI LODs for
those elements having a proposed or approved WQC published In the Federal
Register was performed.  The  results  of those comparisons by matrix are
found in Figures 11 through 16.   The  WQC limits are generally higher than
the EPA detection limits and,  for most of the elements having WQCs, the
EPRI overall LOD was within a factor  of 1 to 10 of the WQC.  However, the
EPRI LOD for lead 1s consistently a factor of 10 or 100 greater than the
WQCs In all matrices.  In the APO,  SW-I/D, and TCHCW matrices, the EPRI LOD
1s 10 to 100 times the EPA WQC.   The  EPRI LOO for aluminum in the SW-D/I
and TCHCW matrices was consistently a factor of 10 or more than the WQC
limit.  Finally, only one element,  zinc,  consistently had an EPRI overall
LOD lower than the EPA WQC in all the matrices.

These data Indicate that most laboratories would have a difficult time de-
termining an element at Its WQC.

Matrix Effects

As part of the data analysis  effort,  the slopes of the precision and recov-
ery linear regression equations  for each discharge stream were compared to
the slopes for reagent grade  water.  Standard statistical tests were em-
ployed to determine whether the  differences seen were significant.  Table  4
summarizes the results of the analysis.   No significant trends were seen
based on significant differences between the slopes of the single operator
and overall precision, and recovery equations.  In fact, one would have ex-
pected more elements to have  had significant slope differences for matrices
such as the two seawater.  This  prediction did not turn out to be true.
Part of the problem with this sort  of statistical comparison 1s that the er-
ror bands associated with the slopes  are relatively large and get worse
with matrices such as seawater.   As a result, differences tend to be ob-
scured In the same matrices where one might predict there to be a matrix
effect.  Additional development  effort is required to devise a statistical
test to determine the presence of a matrix effect.

SUMMARY

The EPRI RP 1851-1 AHQ validation programs have developed a comprehensive
data base of precision and bias  data  using utility laboratories analyzing
typical utility discharge streams.  These data have shown that EPA detec-
tion limits based on single operator  tests in reagent grade water do not
produce similar detection limits.  In fact,  detection limits 10 to 100
times the EPA detection limits have been seen.  In light of these
differences, consideration should be  given to setting detection limits that
are more representative of the real world performance of NPDES monitoring
methods.
                                  12

-------
                                248
REFERENCES

1.  White, H.R., H.D.  Powers,  C.C.  Shlh,  and R.F.  Maddalone,  "Aqueous Dis-
    charges from Steam-Electric Power Plants:  Data Evaluation," EPRI
    CS-3741, November 1984.

2.  Maddalone, R.F., J.W.  Scott,  and H.D.  Powers,  "Aqueous Discharges from
    Steam-Electric Power Plants:  The Precision and Bias of Methods for
    Chemical Analysis," EPRI CS-3744, November 1984.

3.  Maddalone, R.F., J.W.  Scott,  and J.  Frank, "Round-Robin Study of Meth-
    ods for Trace Metal Analysis; Volume 1:   Atomic Absorption Spectroscopy
    - Part 1," EPRI CS-5910, Volume 1, August 1988.

4.  Maddalone, R.F., J.W.  Scott,  and J.  Frank, "Round-Robin Study of Meth-
    ods for Trace Metal Analysis; Volume 2:   Atomic Absorption Spectroscopy
    - Part 2," EPRI CS-5910, Volume 2, August 1988.

5.  U.S.E.P.A., "Methods for Chemical Analysis of Water and Wastes,"
    EPA-600/4-79-020,  March  1979  (updated March 1983).

6.  Curie, L.A., "Limits for Qualitative Detection and  Quantitative
    Determination:  Application to  Radiochemistry," Anal. Chem.. 40(3), 586
    (1968).

7.  Kaiser, H., "Quantitation 1n  Elemental Analysis (Part 2)," Anal. Chem.
    41(4) 26A (1970).

8.  Kaiser, H., 1. Anal. Chem.. 209. 1 (1965)

9.  Keith, L.H., et al.. "Guidelines for Data Acquisition and Data Quality
    Evaluation in Environmental Chemistry,"  Anal.  Chem.. 55(14), 2210
    (1983).

10.  Glaser, J.A.; D.C. Forest, G.D. McKee,  S.A.  Quane, and W.L. Budde,
     "Trace Analysis for Wastewaters," Environ. Set. Techno!. 15(12), 1426
     (1981).

11.  Rice, J.K., "Analytical Issues 1n Compliance Monitoring," Environ.
     Sci. Techno!.. 14(*12), 1455 (1980).

12.  Rice, O.K., Private communication,  February 1987.

13.  U.S.E.P.A., "Guidelines Establishing Test Procedures for the Analysis
     of Pollutants," Federal Register. 49. (209),  Friday, October 26, 1984.
                                  13

-------
                                      249
                                   Table 1

                    UTILITY AQUEOUS DISCHARGE MONITORING-
                   ANALYTICAL METHODS QUALIFICATION  (AHQ)
                            OVERALL PROJECT PLAN
    Pro.lect Title

Sampling and Analysis
of Utility Pollutants

Analytical Methods
Qualification

Analytical Methods
Qualification
Analytical Methods
Qualification
                                                 Test Program Scooe
Part
*
I
II
Roui

1
2
1
III
                   Parameters
As, Se
Ni, Pb, Cr, Cu

Cd
Hg
Fe, Zn

Al, B, Ba, Be,
Cd, Cr, Cu, Fe,
Mn, Mo, N1, Pb,
V, Zn
                    Method
GFAAS**
GFAAS

GFAAS
CVAAS
Flame AAS

ICP
*   Initial effort under EPRI Project RP 1851-1 which involved collection
    and analysis of data on discharge rates and data on the analytical pre-
    cision and bias for utility discharge species.

**  Approximately 6 seawater laboratories determined As and Se by GHAAS.

GFAAS • Graphite Furnace Atomic Absorption Spectrometry

GHAAS - Gaseous Hydride AAS

CVAAS - Cold Vapor AAS

ICP   - Inductively Coupled Argon Plasma Optical Emission Spectroscopy
                                      14

-------
                                                      Table 2

                                        LOD DEFINITION FROM VARIOUS SOURCES
(71
     Author;

     Currle (fi)
     ACS (9)
     R1ce
Title

Decision Limit


Detection Limit
Critical Level
  Definition
LC « 1.6450
                                             LQ • 3.290a
     Kaiser (2,8)     Limit of Detection     LOD - 3a B
Limit of Detection     LOD - 30
     EPA/EHSL (Ifl)    Method Detection       HDL - 3.143a R
                      Limit
LC * 2.576a g
                      Limit of Detection     LOD « 4.652a
      Mathematical                Degrees of
     Interpretation                Freedom

5% chance of reporting zero       Infinite
value as detected,

5% chance of reporting zero       Infinite
value as detected or not
reporting a real value as >0.

Approximately a 5% risk of        Infinite
reporting zero values as
detected for normal distri-
butions and as high as 11%
risk of reporting zero values
as detected for asymmetric or
broad distributions.

7% chance of reporting zero       Infinite
value as detected or not
reporting a real value as >0.

1% chance of reporting zero       6
value as detected.

0.5% risk of reporting zero       Infinite
value as detected or not
reporting a real value as >0.

1.0% risk of reporting zero       Infinite
value as detected or not
reporting a real value as >0.
                                                                                                                      to
                                                                                                                      en
                                                                                                                      o
     *  R1ce also suggests using overall precision data from Interlaboratory studies on real, spiked samples.

-------
                                                           Table  3


                                     Summary of AMQ-III  ICP LOD  and LOQ  Data
          IITIKHTEO TEIT
                                                            IT
  Tom

MMUft i
UKM

 '
                                                          PJKVItlHtM
              COMPUTED' I  I COMPUTED  11 COMPUTED I  I COMPUTED 11  UNIT

                                     5
443T1  T/*/H I Httti tlcMhii  IWttitlOMi jasSiiT
I OUOllTT I
(CRITERIA
                                                       IINCM
                                                                    CURVILIRCM
                                                                                    IIKEM
                                                                                                 CURVILIIItM
         M^

-------
                      Table 3
Summary of AHQ-III ICP LOO and LOQ Data (continued)

-------
                        Table 3
Summary of AMQ-III ICP LOD and LOQ Data  (continued)
             loatut/i) lice iu»/i)HioQ cut/I

-------
I
                                                                 Table  3
                                           Summary of AHQ-III ICP LOD and  LOQ  Data (continued)
    «O

-------
                                                                  Table 3
                                         Summary of AMQ-III  ICP LOD and LOQ Data  (continued)
— Mittr owllty CrlttrU fw frwfcMMr MUMIe.
mtPlM VTiMMl MMV Ml CVVBTIMn •? •!•• 9f
                                                                                                                                          N)
                                                                                                                                        (  ui

-------
                                256
                                   Table 4

                   ELEMENTS HAVING A MATRIX EFFECT  BASED ON
                    REGRESSION EQUATION SLOPE DIFFERENCES
River
Water
                          Single
                          Operator
AT, Cd
                 Overall
Recovery


   N1
Ash Pond
Overf1ow
                                 Be, Cd, Mo
Seawater-Intake
Treated Chemical
Metal Cleaning
Wastes
                   Ni
   Fe


 Al, Pb
                                     21

-------
f\>
                                                                                                                          to
                                                                                                                          Ul
                                                                                          Effor,

-------
                           258
  0.15
I
* 0.10
§
L
0
a
  0.06
  o.oo
     0
   RIv»r Water 	
   Ften Pond Overflow 	•
   Soawator Intake 	
   Soawater Dlecharae  	
 .  Reagent Grade Water
   TCMCW 	
.0     0.1     0.2     O.3     O.4     0.5
               True Concent rot I on. fna/L.
0.0     0.7
                       Figure 2.  Zn by ICP
         Overall Operator Precision vs True Concentration
                              23

-------
                vt
                c5
    MATRIX "NOISE"
INSTRUMENT "NOISE"
                                                                 NEAR ZERO LINEAR
                                                                 EXTRAPOLATION O
                                                                    to
                                                                    U1
                                                                    vo
LONG RANGE LINEAR

EXTRAPOLATION 0
                                                 TRUE CONCENTRATION
                      Figure 3.  Effect of Test Concentration on the Calculation of LODs.

-------
                                   260
I
E
s
                      I
                                                       LEGEND

                                                            LOO/EPA

                                                            LOO/EPA
         Al  B  Bo Be Cd Cr Cu F« Un Uo  N! Pb  V  Zn
                         Elements
                                         No data far •; Cd LOO/EPA ratio O.I.

       Figure 4.  Ratio of EPRI  LOD and LOQ to  EPA ICP Detection Limit for
                   Reagent Grade Water
   1*OO X
                                                       LEGEND

                                                           LOD/EPA

                                                           LOQ/EPA
        At  B Bo B« Cd  Cr  Cu  F«  Un Uo Nt Pb  V Zn
                         El«m«nta
                                                              tar ••.
       Figure 5.  Ratio of  EPRI  LOD and LOQ to EPA ICP Detection  Limit for
                   River Water
                                      25

-------
                                         261
                                                       LEGEND


                                                            LOO/CPA

                                                            LOO/CPA
         Al  B  Ba B« Cd Cr Cu F« Mn Ue  Nl  Pb  V  Zn
                         El«m«nts
      Figure  6.   Ratio of EPRI LOO and LOQ  to EPA ICP Detection Limit for
                  Ash  Pond Overflow
g
                                                      LEGEND


                                                          LOD/EPA

                                                          LOQ/EPA
        Al  B Bo B« Cd Cr Cu F« Mn Me  KJ Pb  V  Zn

                         El«m«nts

                                         N* «^« f*r •• UJO: Bo.V LOQ.
       Figure 7.  Ratio of EPRI LOD and LOQ  to EPA ICP Detection Limit for
                  Seawater Intake
                                     26

-------
                                    262



«OO ;
3 :
I :
-c
S- "*=
3 '
K
CJ
§
* :
I


_t •


































r

















f










, '



p

. ^
;
^
»
, j
^
/
^
^

>





••




'




-•





P
/
X





J

/








;

'









i
^

/


t
t

>
*
?





P
2
'
i
^
^
t
t
t
t






mn
t

?

\
t

4
j
t
,







'
X
+
*
f
f

s
t
0

t




















1
^
^
^
, (
. t

t
<
f
^
«
«
«l
t
t
t
















LEGEND
l^xfl LOO/EPA
^H LOO/EPA













         Al  B  Bo B» Cd Cr Cu  F« Mn Mo Ni  Pb  V Zn
                          Elements
                                           Na ««ta tar •• LOO; AI.««.r«.V LOO.

       Figure 8.  Ratio of EPRI LOD and LOQ to EPA ICP Detection Limit for
                   Seawater Discharge
   ratio
•i
                                                       LEGEND

                                                            LOO/EPA

                                                            LOO/EPA
         A!  B  Bo B« Cd Cr Cu F« Un Mo  NI l»b  V  Zn
                         Elements
                                      N« «ot
-------
                                  263
                                                                         swo
                                                                    SWI
                                    >100
    Range of LCD to B=H ICP detection Unit
Figure 10.   Percentage of Reporting Elements Having a Given  Ratio
                                 28

-------
                                    264
I

I  "
                                                       LEGEND

                                                           UOO/WQC

                                                           LOQ/WQC
         Al  B  Bo Bm Cd Cr Cu F« Un Uo Nt Pb  V Zn
                         Elements
                                               N« WOC tar
       Figure 11.  Ratio  of EPRI LOO and LOQ to  Water Quality  Criteria for
                    Reagent Grade Water
   two !•
*-+  too :
                                                       LEGEND

                                                       t£^3 LOD/WOC

                                                       •H LOQ/WQC
         Al  B Bo B« Cd Cr Cu F« Un Uo  N1  Pb  V  Zn
                          Element*
                                                N« WOC tar •,IXi.F«.Un.M«.V.

       Figure 12.   Ratio of EPRI LOO and  LOQ to Water Quality Criteria for
                    River Water
                                     29

-------
                                          265
1
                                                       LEGEND

                                                            LOO/WQC

                                                            LOQ/WQC
         Al  B  Bo B« Cd Cr Cu Fe Mn Uo Nl  Pb V  Zn
                         Elements
                                                N* WQC tar •*o.r«.Un.M».V.

       Figure 13.  Ratio of EPRI LOD  and  LOQ to Water Quality Criteria for
                    Ash Pond Overflow
1
                                                       LEGEND

                                                           LOO/WQC

                                                           LOQ/WQC
         Al  B  Bo B« Cd O Cu F« Un Uo NI r*b  V Zn
                         Elements
                                  N* WQC tor B,B«,r*.Mn.Ua.V: n* *»ta tar M UOQ

       Figure 14.  Ratio  of EPRI LOD and LOQ to  Hater Quality Criteria for
                    Seawater Intake
                                     30

-------
                                   266
I  -!
                                                       LEGEND

                                                           LOD/WQC

                                                           LOQ/WQC
         AI  B  Bo B« Cd Cr Cu F« Un Uo NI Pb  V Zn
                         Elements
                                  N« WOC tor
                                                     . n* «•*« *w AI UX3
      Figure  15.   Ratio of EPRI LOD and  LOQ to Water Quality Criteria for
                   Seawater Discharge
                                                      LEGEND

                                                           LOD/WQC

                                                           LOQ/WQC
         Al  B  Bo  B«  Cd  Cr Cu r«  Un Uo Nt  Pb V  Zn
                         El«m«nts
                                    tt» WOO tar B.B«.r«JtfnAI«.VS n* tf^a tar ••
      Figure 16.  Ratio of EPRI LOD and LOQ to Water Quality Criteria  for
                   Treated Chemical Metal Cleaning Waste
                                     31

-------
                             267



                     QUESTION AND ANSWER SESSION



                                   MR. HACHIGIAN:  Lee



Hachigian of General Motors.



          I assume the intent was to use this data to work



with your regulatory agency in establishing limitations on



effluents from the utility industry.  Do you foresee that



during these negotiations, if you can call it that, you



would be required to do site-specific...the same type of



report site specific for each of the operations or you could



use this data that you have already determined?



                                   MR. RICE:  That is a very



good question.  Obviously, the collection of the data was



directly related to the entire compliance monitoring and



permit negotiation effort.



          Both things have happened.  In some cases, the



published information has been able to be used by a member



utility in its negotiations of a permit limit.



          In other instances, it has at least secured for



the utility the opportunity to run a site-specific MDL or to



obtain information on their recovery and reproducibility by



a round-robin with a number of corresponding labs.  We have



used the latter situation in connection with and MDL when



that was in question (no detectible discharge for a limit)



to get two or three additional qualified laboratories

-------
                             268



together with the utility laboratory in an approved round



robin simply on MDL.



                              MR. HACHIGIAN:  As a followup



question, has it been effective?



                              MR. RICE:  Pardon me?



                              MR. HACHIGIAN:  Have your



efforts been effective?



                              MR. RICE:  Yes, very much so.



          The entire method validation program has cost the



Electric Power Research Institute close to $1.5 million to



date including furnace AA and ICU on a range of metals and



matrices.  EPAL is now continuing with additional elements



that were not run earlier but are now requested by members



since they in turn are being requested by regulatory



agencies to monitor or to control parameters not previously



to subject of interest.



                              MR. HACHIGIAN:  Thank you.



                                        MR. FIELDING:  Any



other questions?



          (No response.)



                              MR. FIELDING:  Thank you.

-------
                             269
                                   MR. FIELDING:  Our next



speaker is Mr. Bill Krochta.  He will be speaking about the



quantitation detection limits in analysis of environmental



samples.

-------
                            270
    QUANTITATION/PETECTION LIMITS FOR THE  ANALYSIS
                 OF  ENVIRONMENTAL SAMPLES
W. G.  Krochta.  PPG  Industries:
	 L.  I.  Bone,  The  Dow Chemical
CMA;  T.  L. Dawson,  Union  Carbide
    Inc;  K.  J.  G.  Hillig,  BASF
               N.  E.  Prange,  B.  M.
Company;  B.  A.  Cuccherini,
Chemical   &  Plastics  Company,
Corporation;  R.  A.  Javick, FMC Corporation;
Hughes, Monsanto Company; F.  I. Saksa, CIBA-GEIGY Corporation; G.
H. Stanko,  Shell Development  Company

-------
                                     271
                        QUANTITATION/DETECTION LIMITS

                   FOR THE ANALYSIS OF ENVIRONMENTAL SAMPLES
I.
INTRODUCTION
Analytical technology continues its unrelenting pace to develop methodology to
lower the concentration limits at  which the analytes can be measured.  Picogram
(10~12 grams) quantities are  commonly  reported  as  new  detector systems for gas
and liquid  chromatography are developed.   Advances in mass  spectrometry are
leading  to  lower  levels  of  quantitation.    For  example,  ion  trap  mass
spectrometers and inductively coupled plasma-mass spectrometry (ICP-MS) are some
highly sensitive techniques, which are becoming more commonly used for organic
and elemental determinations respectively and capable of detecting subnanogram
(<10  gram)  quantities.  The  statement following depicts the situation that we
are encountering:
         "... the number of compounds detected in a sample of
         water is related to the detection level.  As the
         detection level decreases an order of magnitude, the
         number of compounds detected increased an order of
         magnitude.  Based on the number of compounds detected by
         current methods, one would expect to find every known
         compound at a concentration of 10-12 g/L or  higher."  -
         Dr. William T. Donaldson (EPA Athens Laboratory)
As  the  regulated  community  is  required  to  perform  within  the  level  of
increasingly restrictive compliance limits, the analytical chemist must emphasize
to  the  public  that  all  measurement  data  have  an  associated  uncertainty
interval(1).   This  information  becomes   critical  as  measurements  are  made
approaching the lowest analytical capability of a given procedure.

-------
                                   272
II.   ANALYTICAL LIMITS OF MEASUREMENT - DEFINITIONS

In their regulatory programs, the USEPA uses a variety of procedures to establish
limits of measurement.  In the following  slide definitions are given to present
an approach to define analytical capability.

LIMIT OF DETECTION (LOP) - Lowest concentration level that can be
determined to be statistically different from a blank(7).

METHOD DETECTION  LIMIT (HDL) - Minimum  concentration  of analyte that  can be
determined with 99% confidence that the true value is greater than zero(2,3,13).

INSTRUMENT DETECTION LIMIT flDU  - Smallest signal above background noise that
an instrument can detect reliably(7).

LIMIT OF QUANTITATJON fLOQ) - Concentration above which quantitative
results may be obtained with a specified degree of confidence(7).

PRACTICAL QUANTITATION LIMIT fPQL) - Lowest level that can be reliably
achieved within specified  limits  of precision  and accuracy during  routine
laboratory operation conditions(S).
III.  APPLICATION OF METHOD DETECTION LIMITS fHDLl SUBJECT TO MATRIX EFFECTS

The HDL is similar to the LOD except that the LOD is defined with a sample blank
whereas  the MDL  is defined  with  either  a  blank  or  in  each  matrix  being
analyzed(2,3).  In most cases, however, laboratories report MDLs  determined at
one point  in  time and  routinely based on reagent water.  They  do  not normally
perform the MDL evaluation on  the  different matrices analyzed  for regulations
development, compliance monitoring, or tested to determine permit requirements.

In their proposal  to set enforceable maximum contaminant levels (MCLs) for
volatile synthetic organic chemicals in drinking water(5), EPA explains that
the MDL could not be used as the basis for quantitative maximum contaminant
levels: "The specification of such a concentration is limited by  the fact that
MDLs are variables affected by  the performance  of a  given measurement system.
MDLs are not necessarily reproducible over time in a given  laboratory, even when
the same analytical procedures, instrumentation, and sample matrix are used."

-------
                                    273
 IV.  PRACTICAL OUANTITATION LIMITS (POL) AS A MEANS OF IDENTIFYING
      HEASUREABLE CONCENTRATIONS
Many observations for organic toxic pollutants are below the MDLs, thus creating
difficulties in developing effluent  limitations  guidelines  and permit limits.
In such instances where analytical  and effluent variability cannot be determined,
only those concentrations above quantifiable levels (17) should be considered.
It should  also be recognized that  there is a fundamental  difference between
detection  and  quantitation  limits.   Unfortunately these terms are  too often
misused as being synonymous.  EPA has developed a method for establishing such
quantifiable numerical limits for its proposed drinking water standards (50 FR
46902)  and for its  proposed  organic toxicity  characteristic  (51  FR 21652),
designated as  the practical quantitation limit  (PQL).   EPA has developed this
concept of a PQL for specific analytical methods and lists of chemicals.


A.    RECOMMENDED PRACTICAL OUANTITATION LIMITS COMPARED TO  METHOD DETECTION
      LIMITS

The EPA used  PQLs which are recommended as 10 times the MDL for selected volatile
organic chemicals when it proposed MCLs for drinking water.  The Agency states
that:  "setting the PQLs in a range between 5 and 10 times the MDL achieved by
the  best  laboratories  is  a fair expectation for  most state  and  commercial
laboratories"  (50 FR 46907).  At the PQLs chosen by EPA for this rulemaking, its
performance  evaluation  studies  indicate  that   80%  of the   EPA  and  State
laboratories in its water program evaluation studies could measure within ±40%
of the true concentration.  This was the basis for setting the PQL at 10 times
the MDL.   This is not a very high standard of performance  as admitted by the
Agency  in  the  preamble  to this proposed regulation.   Thus, even  at the PQLs,
over 20%  of  the "good" laboratories would not be expected to obtain results
within  ±40%  of the concentration of a  specific component.   At concentration
levels  below  PQL,   performance  of even  the   best   of "good"  laboratories
deterioriates  rapidly.
B.    PRACTICAL OUANTITATION  LIMITS IN REAL MATRIX SAMPLES  REFLECT EFFECT OF
      MATRIX INTERFERENCE

A recent presentation(12) described a study evaluating Method 8020, which is a
gas/liquid chromatpgraphy procedure in SW-846 "Test Methods for Evaluating Solid
Wastes, Physical Chemical Methods" for the determination of low concentrations
of toluene, benzene, and xylenes in real  matrix groundwater samples.  The round
robin study involved 20 commercial laboratories. Method 8020 lists  the practical
quantification limits for all  three compounds as 2.0 /*g/L.  The PQLs derived from
results achieved by the  laboratories  in  this  study are much higher.  The PQLs
at which 80% of the laboratories could achieve a recovery  within ±40% of a true
value from this study are 7.5 /»g/L for benzene, >20 /*g/L for toluene, and 18.5
/ig/L for  total  xylenes.  It is  clear  that the Method  8020 published PQLs are
seriously underestimated when applied to this groundwater matrix  and for these
20 laboratories.

-------
                                   274
The  Inability  of  these  laboratories  to perform within the method PQL criteria
should not be surprising even to EPA.  In the preamble to  the final rule on the
federal  primary drinking  water standards for eight volatile organic compounds
(VOCs),  EPA states that:

"PQLs for the  VOCs were based on the MDL and surrogate test data ... The PQLs
based on these laboratory data are considered a two step  removed surrogate for
actual laboratory performance,  first because they  are estimated from another
measurement  (the  MDL)  and  second,  because  they  are  derived  from laboratory
performance under ideal circumstances.   Therefore, they do not  actually
represent the results of normal  laboratory procedures, but are a model of what
normal procedures might achieve.  Specifically:

(1) Laboratories receive performance evaluation samples in which a limited number
of  concentrations are  analyzed  and the  samples  do not  have matrix
interferences as might actual samples;

(2) PQLs are based on EPA and State laboratory data which are considered to be
representative of the best laboratories, but not all laboratories; and

(3) Samples are analyzed  under  controlled ideal  testing  conditions  which may
not be representative of routine practices.

For these reasons, the  PQL  represents a relative  stringent target for routine
performance."  (52 Federal Register 25699).

More  specific to groundwater  samples,  EPA discussed  the significance  and
reliability of the PQLs  that are  included in Appendix IX of the  rule:  "The PQLs
listed were EPAs  best estimate of the practical  sensitivity of the applicable
method for RCRA groundwater monitoring purposes.   However,  some of the PQLs may
be unattainable because they are based on general  estimates  for the specific
substance.  Furthermore, due to  site-specific factors, these limits may not be
reached."  53 Federal Register 39721.

For  solid  wastes  the  matrix problem  has  also been  demonstrated to  be  very
significant(21).   Member  companies  of  the Hazardous  Waste  Treatment Council
obtained initial  information that showed  33 out  of the  91  Best Demonstrated
Available Technology (BOAT)  standards  promulgated for the First  and Second Third
Land Disposal  Restrictions were  set at levels below the PQLs.  As a follow-up,
a formalized inter!aboratory study using incinerator ash samples was performed.
In this  matrix a  range finder study was conducted  by  six member companies to
determine  appropriate  spiking  levels  to  determine  MDLs  for each  of  the
constituents.   As  part  of this study a matrix spike was  prepared  at the BOAT
level and determined.   The  results of the study showed that 65 percent of the
volatile constituents, 73 percent of the acid extractable constituents, and 23
percent of the base  neutral  extractable constituents  were not  detected at the
spike performed at the treatment standard limit.

The PQLs and also MDLs published by EPA for its analytical methods are
based on reagent water spiked with the compounds of interest, so they do not
represent limits achievable where matrix interferences exist, as with actual
samples.   EPA does identify in Method 8020 that the PQLs are highly matrix-
dependent and that they are listed only to provide guidance and may not always
be achievable.  Unfortunately, these caveats or warnings are likely to be
Ignored,  particularly by some regulatory agencies, when permit limits or other
regulatory levels are set.

-------
V.
                              275

PROPER TREATMENT OF THE DATA CAN AVOID MISREPRESENTATION OF THE FACTS
      A.
      RULES FOR THE USE OF SIGNIFICANT NUMBERS
Despite the wide  attention given to numbers for  quantitative  and qualitative
limits the improper use of rules for use of significant numbers goes virtually
unnoticed.  As measurements are required more and  more frequently to be made at
decreasing concentrations, the relative analytical variability and uncertainty
can increase substantially and the need  to understand and recognize significant
data is essential.  Horwitz  et al  (22)  reviewed data from over 50 independent
Association of Official Analytical Chemists (AOAC)  inter!aboratory collaborative
programs  covering  numerous AOAC drug and  pesticide studies.   The analytical
methods covered were chromatography,  atomic absorption spectrometry, absorption
spectrometry,   polarography,  and  biossay.    In Figure  1 the  %  variation  is
expressed as powers of 2 with the mean concentration expressed as powers of 10.
A convenient  reference point  is  that at  1 ppm the  variation  is 16%.   The %
variation was found to double for each decrease of concentration by 2 orders of
magnitude.   It is  important to  note that this  curve  is independent  of the
analyte or analytical  technique that was used to make the measurements.  These
relationships should also apply to environmental levels of measurement as well.

Analytical chemists must always emphasize to  the users  of the data  that the
single most important  characteristic of any result obtained from one or more
analytical measurements  is an  adequate  statement  of its  uncertainty interval.
Often in  legal judgments  there is  an attempt to dispense with  uncertainty and
try to obtain  unequivocal  statements; therefore,  an uncertainty interval must
be clearly defined in cases involving litigation and/or enforcement proceedings.
Otherwise, a value of  1.001  without  a specified uncertainty,  for example, may
be viewed as legally exceeding a permissible level of 1(7).

The analytical inclusion  of  only significant numbers  is  vital  to the accurate
interpretation of data.  Scientific personnel are not exempted from the tendency
to retain all  values,  no  matter how  divergent  or  suspect they  may be.  One of
the principles of handling the data of physical  and  chemical  measurements is
that  a  numerical  result by  itself  should give  an  approximate  idea  of the
precision of the value  as indicated  by  the  number of significant figures used
in expressing  the  value.   An inaccurate representation  of significant figures
may give  one  an impression  nearly as  erroneous as from  an  inaccurate value.
Misuse of significant  figures can cause reporting  violations  when indeed the
measured  value does not exceed the limit.  Adherence to proper expression of
significant numbers is especially important when permit limits  are  near the
limit of  quantitation  for  the procedure and  its relative  uncertainties are
1arge.

The  number of significant  figures  reported   as a  result of  a  scientific
measurement depends on  establishing previously the  relative precision with which
the measurement can be  made as  shown in Table 11(11).   In  considering the proper
use of  significant figures  for  regulatory   reporting,   it  is  imperative that
significant figures start at the laboratory bench and be adhered to by anyone
who  further treats  or  handles the  data.  Otherwise,  false  conclusions and
misunderstanding will  develop and possibly lead to serious consequences.

-------
                                    276

 B.     GUIDELINES  FOR  REPORTING DATA

 EPA  has recognized  that data measured  at or  near the detection  limit have
 considerably more uncertainty associated with them than when significant
 amounts are present(6).  In this discussion EPA acknowledges the recommendations
 by the American Chemical Society Report(7).  A graphical  illustration of the
 relationship of LOD and LOQ is shown in Figure 2(7).  The base scale  is  in units
 of standard deviation, which is assumed to be  the same  for all the measurements
 Involved.

 Confidence in the apparent analyte concentration increases as  the analyte signal
 increases above the  LOD.  The value for LOQ = IQa  is recommended,  where a is
 the standard deviation of the measurements.  Assuming a large  number  of samples,
 the LOQ then corresponds to  an  uncertainty of ±30% in the measured value (10<7
 ±3ff) at the 99% confidence level.  The  LOQ is most useful  for  defining the lower
 limit  of the useful range of measurement methodology.

 From these guidelines in Table III,  if  the measured value is less than the limit
 of detection, one should report "not detected" together with the value for the
 LOO.  When the measured value is larger than the LOD  but  smaller than the limit
 of quantification (LOQ),  report  "detected but not quantifiable"  together with
 the value for the LOQ.   If  the  measured value is greater than the LOQ, report
 the value and its uncertainty.
VI.   IMPACTING THE REGULATORY PROCESS

Data measured at or near the limit of detection may cause  serious difficulty for
the user  in developing valid  conclusions  from any  study.   Not only  can the
amount  of uncertainty  approach  and even equal  the  reported value,  but also
confirmation   of   the  species  reported  is   virtually  impossible  as  the
identification must depend solely on the selectivity of the methodology and
knowledge  of the  absence of  possible  interferents.   As  the  concentrations
increase to measurable  amounts these problems diminish.  As stated previously,
quantitative interpretation, decision-making, and regulatory actions should be
limited to data  at or  above  the  limit  of  quantitation(7).   The  following
discussion  graphically  illustrates  how analytical variability  can  impact the
regulatory process.
A.    GRAPHICAL ILLUSTRATIONS OF THE IMPACT OF
      ON COMPLIANCE LIMITS
ANALYTICAL VARIABILITY
Figures  3  through  7  were  developed  in  order to  visualize  the  impact  of
variability upon laboratory measurements of concentrations in plant effluents.
Figure 3 shows the general probability distribution function for random error
when the  measured  concentration is expressed  in a units.  This  curve  can  be
thought of as a frequency distribution when a large number of effluent samples
of  the same  concentration are  analyzed repetitively.    The  x-axis  is  the
concentration that a laboratory may measure; the y-axis is proportional  to the
frequency the  laboratory measures a given  concentration and is  expressed  in
probability units.  The y-axis data have  been normalized  so that the total area
under the curve gives a value of 1.000.  This curve applies to all analyses  in
which only random error occurs.

-------
                                    277
Several observations can be made  regarding  the  probability distribution shown
in Figure 3.

Observation #1:  Only a small percentage of the total  analyses may give the best
estimate of the true value.

Observation #2:  One-half  the measurements  are  above the mean  and one-half of
the measurements are below the mean.   Therefore,  if  the mean  is some effluent
trigger concentration above which a plant would be  violating  its permit, the
plant would be failing one-half the time, if these data were treated as having
QO uncertainty.

Observation #3:  The measured  concentrations shown  in  Figure  3,  99.7% of the
reported values would fall between plus or minus 3a of the mean concentration;
therefore, it  can be seen  that the  a of a determination is a very fundamental
property of a distribution which must be used  in evaluating data which contains
uncertainty.

B.    THE APPLICATION TO REGULATORY LIMITS

In  order  to  translate  this  general  probability  distribution to  real-world
examples,   Figures  4 through  7  were  generated  assuming  different  analytical
uncertainty  in  the random  errors.    All  figures   were generated  for  the
measurement of an  effluent sample containing 100 ng/L  of the  target analyte.
Figure 4 shows the distribution of measured concentrations when the analytical
uncertainty produces a value of  1/jg/L for &; Figure 5 shows the distribution of
measured concentrations when the analytical uncertainty produces a value of 10
^g/L for a\ Figure  6 shows  the distribution  of measured concentrations when the
analytical uncertainty produces  a  value of 30 v%/\- for a; and Figure 7 shows the
distribution of measured concentrations when the  analytical  uncertainty produces
a value of 100>g/L for a.  The probability distribution for the last case has
been truncated at 0 /*g/L since negative values of concentration  are meaningless.

These four cases show  clearly  the impact of  determinations which are carried
out  with  different  amounts  of  analytical  uncertainty.     Unfortunately,
regulations are written as if data were being  obtained with an uncertainty less
than that shown in  Figure  4.  Permits which  give a specific limit for a certain
compound,  fall into this category.  However, the  analytical data which are being
obtained by a  typical environmental laboratory for the analysis  of reagent water
are most likely analytical data obtained with the uncertainty shown in Figures
6 or 7.  Figure 6 describes most analytical  data obtained  using EPA Methods 624
and 625 when  the measured concentration  is ten times higher  than  the method
detection limit determined  in reagent water.  Figure 7 describes most analytical
data obtained  using EPA Methods 624 and 625 when the measured concentration is
equal to  the  method detection limit which can  be the  case if the  sample or
sample extract must be diluted due  to interfering substances.   The concern is
that  the  probability  distribution  summarized  in  Figure 6  is  used  by  the
Environmental   Protection  Agency to  characterize  data obtained by  analytical
laboratories for effluent  analyses.   However,  these data represent a best case,
since method  detection  limits for  Methods  624  and  625 are derived  from the
analysis of reagent water.  Reagent  water data  should not necessarily be used
to determine  the random error associated with  all  plant  effluents which may
contain relatively  high levels of inorganic salts, and  unregulated  organic
compounds  which may interfere with these methods.

-------
                                   278
An additional important conclusion from this set of figures is that these case
studies must be applied not only to analytical data obtained using the classical
EPA Methods  624 and 625,  but it also  applies  to  BOD,  COD,  toxicity, opacity,
etc.  Any time any measurement is being made which includes random error, this
measurement contains the same types of uncertainty as described above which is
certainly accentuated as the required concentration limits are decreased.
Therefore, regulations should not be  written with wording that  implies that
Figure 4  uncertainties exist when in  fact Figures 6 and 7  uncertainties are
typical  representations of  analytical uncertainties  associated  with  permit
violation data.

Evaluation of permit violations  cannot be properly made without first knowing
the analytical  uncertainty of the  determination  in the vicinity of the permit
trigger concentrations  and for the exact  matrix  under study.   This means that
a data which are published with EPA Methods for the analysis in spiked reagent
water  should  only be used  as  a guide.    For  application,  however,  this
information should be developed for each compound/parameter in each laboratory
performing these analyses.

C.    THE APPLICATION TO BACKGROUND CONTAMINATION

As regulated limits go  lower and lower, background contamination  becomes an ever
increasing problem.  However, this topic can  be treated  in much the same way as
were regulatory limits  discussed  above.   The same analytical  uncertainty must
also be applied to the analysis of method blanks.  In this case, method blanks
are not defined as replicate injections of a sample extract,  but are replicate
extraction  and  extract  analyses  when  representative  glassware,  solvents,
instrumentation, etc. are used in the analyses.  Using this same reasoning, an
analysis of the  mean concentration of the background contamination and the a for
that determination  gives one the information needed  to determine  whether  a
measured quantity  in  a plant effluent sample is actually different  from the
quantity present in the method blank.

For commonly occurring  background  contaminants  such  as methylene  chloride,
acetone, toluene, 2-butanone, and common phthalates no positive sample results
should  be  considered   real  which  are  within  10
-------
                                    279
VII.  RECOMMENDATIONS

There is a LOD or MDL which  can be determined for every analyte in every matrix
below which it is not possible to reliably ascertain that an analyte is present
or absent. There is also a concentration  range  above the LOD or MDL where it is
possible  to  qualitatively  establish  the  presence  of  an  analyte,  but  the
concentration cannot  be accurately and  reliably quantified.   It  is  also  not
practical to  determine precisely the  LOD  or  MDL  for  all  analytes,  in every
matrix, and at all laboratories.   All regulatory programs must recognize these
facts.   As a  practical solution to  this  problem,  every  method  should  have
published  practical  quantification  limits  (PQLs)  which  are  at  least media
(water/soil) specific.  Many of these PQLs have been published by media,  and for
most analytes these PQLs  are  representative of levels  that  can be achieved at
most commercial  laboratories.  However, there should  also be  procedures  for
determining matrix specific detection  and quantitation  limits.   Unfortunately
it is not possible to analyze a large enough universe of matrices to establish
generalized quantitation  limits  for  comparison with  regulatory levels.   An
approach must be established which will preserve the utility of published PQLs
as guidance, while recognizing the significant number of compliance limits which
are below their respective PQLs and thus require a variance procedure.

If  a laboratory  determines  that  it  can  not meet published detection  and
quantitation limits in  their  sample matrix, they should  be  allowed to measure
these  levels  using  established  procedures  which  include  mandated  QA/QC
requirements.    These  levels would then  be  used as reporting limits.   If  the
quantitation limit, so established,  is  above the regulatory level, the compound
would be considered to be  in compliance until such a time that a level above the
quantitation  limit is  measured.   This  assumption  of  compliance  would apply
whether  or  not  the  quantitation limit  were  a published PQL  or  a  measured
quantitation limit.  EPA would also determine the frequency that these published
PQLs would  be re-evaluated  pending method and  equipment  improvement.   In some
cases the Agency has suggested that  a facility  may petition for such a variance
(24).

We also recommend that the EPA establish uniformity  among the various regulatory
programs for  the  determination of the method  detection limit.    Although  the
definition  is essentially the same,  the number of replicates and blanks may be
different,  therefore, the calculation  is effected.   This  can further compound
the current state  of  confusion in understanding and applying quantitation  and
detection limits.  The  corresponding quantitation  limit  should be established
at five to ten times the MDL or substantially higher as the matrix would dictate
(19).  The use of such  factors,  however, must be used with extreme care as  the
method variability may  well be underestimated  by most  laboratories (17).   EPA
recognized this need  for  consistency in  its Report to  Congress in CWA Section
518.  It was  reported  that  analytical  methods are  sometimes  unnecessarily
different  for  similar sample  matrices,  target  analytes  and  data  quality
objectives.   The  Agency  should  move  to greater  method  uniformity  and  more
consistency in  the use of quantitation and   detection limits and  use these
concepts in regulatory  compliance situations.

-------
                                    280
VIII.
REFERENCES
 1. Rogers, L. B., et al., Eds., "Recommendation for Improving the Reliability
    and Acceptability of Analytical Chemical Data Used for Public Purposes",
    Chem. Eng. News, 1982, 60 (23) 44.
 2. Glaser, J. A., et al.,  "Environmental  Science  & Technology",  Vol. 15. No.
    12, pp 1426-1434 (1981).
 3. 40 Code of Federal Regulations 136, Appendix B, 1987.
 4. 40 Code of Federal Regulations 136, Appendix A, 1987
 5. 50 Federal Register 46902, November 13, 1985.
 6. 53 Federal Register 48849 December 2, 1988.
 7. Keith, L.H., et al,  Anal. Chem. 1983, 55, 2210-2218
 8. 53 Federal Register 48840 December 2, 1988.
 9. 53 Federal Register 48839 December 2, 1988.
10. Standard Methods for the Examination of Water and Waste Water, 15th ed.,
    pp 16-18, American Public Health Association, American Water Works
    Association, and Water Pollution Control Federation, (1980).
11.   Private Communication.
12.   Stanko, 6. H., and R. W. Hewitt, "Performance Evaluation of Contract
      Laboratories for Purgeable Organics",  Presented at:  12  Annual  EPA
      Conference on Analysis of Pollutants  in  the  Environment,  Norfolk,  VA,
      Hay 10-11, 1989.
13.   40 Code of Federal Regulations 136.2 (f) 1987
14.   USEPA Laboratory Data Validation  Functional  Guidelines  for  Evaluating
      Organic Analyses.   Prepared  for the  Hazardous  Site Evaluation  Div.,
      USEPA.  Washington,  DC.   Prepared by the  USEPA Data Review Work Group,
      July 1, 1986.
15.   USEPA Laboratory Data Validation  Functional  Guidelines  for  Evaluating
      Inorganic Analyses.  Prepared for  the  Hazardous Site Evaluation Div.,
      USEPA, Washington, D.C.  Prepared  by the USEPA Data Review  Work Group,
      July 1, 1988.
16.   Koors, S. J., "Environmental Law Reporter News and Analysis", May,   1989,
      p. 10213.
17.   Koors, S. J., "Environmental Law Reporter News and Analysis", May,   1989,
      p. 10219.

-------
                                    281
18.   USEPA  "Statistical  Analysis  of Ground  Water Monitoring  Data  at  RCRA
      Facilities,  Interim Final  Guidance"  Office  of  Solid  Waste Management
      Division, February, 1989, Section 8.

19.   "Test Methods for Evaluating Solid Wastes, Physical/Chemical Method"
      Third Edition, 8010-10, USEPA Office of Solid Waste, Revision I,
      December, 1987.

20.   Parr. J., K. Carl berg, and G. Ward, "Reporting of Low Level Data for
      U.S. Environmental Protection Agency Needs", Presented at:  Third
      Chemical Congress of North America Symposium in Honor of W. E. Harris,
      June 8, 1988.

21.   Method Detection Limits and Practical Quantitation Limits for Incinerator
      Ash Matrices-Interlaboratory  Study.    Prepared  for  the Office  of Solid
      Waste,  USEPA,  Washington,  D.C.    Prepared  by the  Analytical  Chemistry
      Committee, Hazardous Waste Treatment Council, December 22, 1989.

22.   Horwitz, W., Anal. Chem., 1982. 54 (1), 67A - 76A

23.   "Handbook  for  Analytical  Quality  Control  in  Water  and  Wastewater
      Laboratories" EPA-600/4-79-019, Chapter  7,  Environmental  Monitoring and
      Support Laboratory, USEPA Office of Research and Development, Cincinnati,
      Ohio.

24.   54 Federal Register 26603, June 23, 1989.

-------
                    282
          DR. WILLIAM T. DONALDSON
           (EPA ATHENS LABORATORY)
"... THE  NUMBER OF  COMPOUNDS  DETECTED IN  A
SAMPLE OF WATER IS RELATED TO THE  DETECTION
LEVEL.  AS  THE DETECTION LEVEL DECREASES  AN
ORDER OF  MAGNITUDE,  THE NUMBER OF  COMPOUNDS
DETECTED  INCREASED  AN  ORDER  OF MAGNITUDE.
BASED ON THE NUMBER  OF  COMPOUNDS  DETECTED  BY
CURRENT METHODS,  ONE  WOULD EXPECT TO  FIND
EVERY KNOWN COMPOUND AT A CONCENTRATION  OF
10-12 G/L OR HIGHER."

-------
                       283
            DEFINITION  OF

     ANALYTICAL CAPABILITY

LIMIT OF DETECTION (LOP) - Lowest concentration
level that can be determined to be statistically different
from a blank.

METHOP PETECTION LIMIT (MPL) - Minimum
concentration of analyte that can be determined with 99%
confidence that the true value is greater than zero.

INSTRUMENT PETECTION LIMIT (IPL) - Smallest
signal above background noise that an instrument can
detect reliably.

LIMIT OF QUANTITATION (LQQ)   - Concentration
above which quantitative results may be obtained with a
specified degree of confidence.

PRACTICAL QUANTITATION LIMIT (PQL) - Lowest
level that can be reliably achieved  within specified
limits of precision and accuracy during routine
laboratory operation conditions.

-------
                          284
COMPARISON  OF REPORTABLE  SIGNIFICANT FIGURES AS A
           FUNCTION  OF RELATIVE PRECISION
    Precision  f%")

  ±0.001  to ±0.01
  ±0.01   to ±0.1
  ±0.1   to ±1
  ±  1   to ±10
  ± 10   to ±30
Significant
 Figures

   5
   4
   3
   2
   1
                                             Example
Calculated Reported
 54.8149
 54.8149
 54.8149
 54.8149
 54.8149
54.815
54.81
54.8
55
5 x 101

-------
                            285
       ACS  GUIDELINES  FOR REPORTING DATA
Analyte Concentration
   in Units  of a
     t - sb)
Region of Reliability
     <3

      3
   3 to 10
     10
Region of questionable detection
(and therefore  unacceptable)
Limit of detection  (LOD)
Region of less-certain quantisation
Limit of quantisation (LOQ)
Region of quantisation

-------
                                        286
I

I
               Chemical
               Product
               Analysis
                                                         Drinking
                                                         Water
                                                         Limits
     -60
     FIGURE 1. CURVE RELATING VARIABILITY BETWEEN LABORATORIES AND CONCENTRATION.
                  Reprinted with permission from Analytical Chemistry, vol. 54
                  No. 1, January, 1982. Copyright 1982 American Chemical Society.

-------
            ±3a
                     Total  Signa
                             (St)
                              ±3CT
               Zero
       LOD
LOQ
Region of High
Uncertainty
   i   i   i
           Region of Less
           Certain
           Quantitation
           J	I	I	I	I	L
         Region of
         Quantitation
    1  I   I	LJ
     -4-3-2-1  0  1  2  3  4  5 6  7  8  9 10 11 12 13  14
Sb      i
     (Sb + 3a)
       (In units ofa)
                                         (Sb + 10cr)
 FIGURE 2, RELATIONSHIP OF LOD AND LOQ TO SIGNAL STRENGTH
 REPRINTED WITH PERMISSION FROM ANAL. CHEM. 1983. 55. 221 0-22 18.COPYRIGHT 1983. AMERICAN CHEMICAL SOCIETY.
to
CO
-J

-------
                             288
      Figure  3. Normal Curve  of Random Error
              (x in  sigma-units  from  mean)
 0 ^Probability-Units
 0.3
 0.2
 0.1
 0.0
        -6    -4-2
             Sigma-Units
                                0246
                                from the  Mean
   Rgure 4. Normal Curve of Random Error
     Mean = 100 ug/L; Sigma = 1 ug/L

 ^Probability-Units
0.3


0.2


0.1

O.Ol
     0     50     100     150
      Measured Concentration (ug/L)
                                    Figure 5. Normal Curve of Random Error
                                      Mean = 100 ug/L; Sigma = 10 ug/L
                                   ^ Probability-Units
                                  0.3


                                  0.2


                                  0.1

                                  0.0
                                      0    50   100   150   200
                                       Measured Concentration (ug/L)
   Figure 6. Normal Curve of Random Error
    Mean = 100 ug/L; Sigma = 30 ug/L
0^ Probability-Units
                                    Figure 7. Normal Curve of Random Error
                                      Mean = 100 ug/L; Sigma = 100 ug/L
0.3

0.2

0.1

O.Ol
     0     100     200    300
      Measured Concentration (ug/L)
                                       0     200    400    600
                                        Measured Concentration (ug/L)

-------
 Figure 3. Normal Curve of Random Error
      (x in sigma-units from mean)
Q ^Probability-Units
0.3
0.2
0.1
0.0
to
00
10
         -4-20    2   4    6
       Sigma-Units from the Mean

-------
  Figure 4. Normal Curve of Random Error
     Mean = 100 ug/L; Sigma = 1 ug/L
0 ^Probability-Units
0.3
0.2
0.1
0.0
L
                                           N)
                                           U3
                                           O
       0      50     100      150
      Measured Concentration (ug/L)

-------
 Figure 6. Normal Curve of Random Error
    Mean = 100 ug/L; Sigma = 30 ug/L
   Probability-Units
0.3
0.2
0.1
0.0
to
<£>
                              300
      Measured Concentration (ug/L)

-------
 Figure 5. Normal Curve of Random Error
    Mean = 100 ug/L; Sigma = 10 ug/L
0 /| Probability-Units
0.3
0.2
0.1
0.0
                                           to
                                           vo
                                           10
      0     50    100   150    200
      Measured Concentration (ug/L)

-------
 Figure 7. Normal Curve of Random Error
   Mean = 100 ug/L; Sigma = 100 ug/L
Q ^Probability-Units
0.3
0.2
0.1
0.
to
vo
CO
             200     400
      Measured Concentration (ug/L)

-------
                            294
(1)   Laboratories receive performance evaluation samples in which
     a limited number of concentrations are analyzed and the
     samples do not have matrix interferences as might actual
     samples;
(2)   PQLs are based on EPA and State laboratory data which are
     considered to be representative of the best laboratories,
     but not all  laboratories; and
(3)   Samples are analyzed under controlled ideal  testing
     conditions which may not be representative of routine
     practi ces.

     For these reasons,  the PQL represents a relative stringent
     target for routine performance.   (52 Federal  Register
     25699).

-------
                             295
                QUESTION AND ANSWER SESSION
                                   MR. ENWEZE:  I am Tony
Enweze from EBASCO.
     My question is/ having looked at all these different
values of method of detection limits, don't you think that
one of the primary reasons EPA is using what they have now
is something related to the risk level that has been
documented that particular contaminants could cause certain
kinds of diseases as certain concentration levels?  You see,
in order for them to maintain that every laboratory report a
value within a certain range, they have established a value
such that it will be compatible with that risk level for
that particular compound.
     Do you see that as a possible reason why they have
these random values for detection limits?
                                   MR. KROCHTA:  I am not
sure if I understand that question.
                                   MR. ENWEZE:  What I am
trying to focus on here is that detection limits sometimes
as established by EPA often are related to the risk level of
a particular contaminant or any compound in the environment.
It isn't necessarily that it might be coming from a
practical approach, in other words, how this can be analyzed
or how this can be found, but EPA values have been

-------
                             296
established to give the limit above which the toxicity
effect of a particular compound may impact a person who
comes into contact with it.
     So, do you think that those kinds of reasons is a major
impact on why EPA has established their limits to be such in
a random fashion?
                              MR. KROCHTA:  Well, I am not
sure if I understand the complete question.  I may be the
reasons, but our primary concern is that EPA will need to
establish for these compounds PQLs that we should use as
only guidance, because published PQLs are set primarily in
specific matrices that are not necessarily site specific.
We should only use them as a guidance as they may not be
achievable even if for risk assessment purposes or routine
practical applications.
     As I mentioned earlier, work was done evaluating solid
matrix effects, when they did a study on incinerator ash.
It was found that about 33 percent of the EPA designated
limits were actually below what could be measured in that
particular matrix, in that site matrix.
     Now, EPA has agreed that they do have a problem, and
they are making an effort to establish more realistic method
detection limits, but we have to recognize that they are
only to be used as a guidance.  Unfortunately, they are too
often used, when they are setting regulations, as a level

-------
                             297



that can be achieved, and that is not the case.  They cannot



be achieved.



     One more comment, in our own opeation, we had a



regulation limit set which was about 100 times below what we



could actually measure in that particular matrix.



                                   MR. ENWEZE:  This is



exactly what I am trying to get to.  Don't you then think



that for EPA to try to make it sort of arbitrary that



wouldn't that put the environment... the people at some kind



of risk to exposure to contaminants?  If the detection



limits...the results that are reported by labs often are



reported than less than the detection limit which is



considered from a risk point of view as not relevant, how



would you then control that factor of the fact that...



     Let me give you an example.  Use the method detection



limit of 5 ppb.  It is clearly established that at 5 ppb, a



particular contaminant can cause cancer to anybody who



inhales it at this certain dose.  So, how can you then



cont.^1 the public or the industries from polluting above



those limits that could cause the risk that we are all



worried about?



     That is where my concern comes in, that what you are



saying, leaving the public to decide where the method



detection limit will be based on the analytical approach

-------
                             298



might be risky to the public as a whole because of that



factor of risk that is associated with each compound.



                              MR. KROCHTA:  I can see your



point.  It is a problem.  However, if they are going to



establish a regulation, it should be at a level that can



actually be measured.  To be controlled, you have to be able



to measure it.



     You want to have a regulation that can actually protect



the public, but if you can't measure it, it is very



difficult to operate under it.



                              MR. ENWEZE:  So, do you think



then that the risk level that is established in risk



assessment books are probably not valid or those values are



reported with a method that was not capable of detecting



that level?



                              MR. KROCHTA:  I am not sure.



I don't think we can answer that question right here.



                              MR. ENWEZE:  Well, okay.  I



think I made my point that the values that we are talking



about in detection limits, it is a little bit hard to leave



the public exposed because the industry would like to see



EPA compromise at a certain level that they consider



practicable when in that quality it has been documented



through research that below that level that can be detected

-------
                             299



there is a high risk of possible diseases that could be



inflicted on the public.



     So, I find it difficult to let the system go to a



higher detection limit when considering the risk level that



might be encountered.



     Thank you.



                                   MR. FIELDING:  Any other



questions?



(No response.)



                                   MR. FIELDING:  All right.



Let's take about a 15-minute break and be back about 3:15 or



3:20.



(WHEREUPON, a brief recess was taken.)

-------
                             300
                              MR. TELLIARD:  Over the last
eon, the Office of Water Regulations and Standards has been
working on a regulation for the offshore oil and gas
industry.  We feel that this regulation will outlast all of
us.  It is the longest known regulation in history.  It has
taken so long to write, most of the panel members have died.
(Laughter.)
                              MR. TELLIARD:  The youngest
kid on the committee is 63, and he started out fresh out of
school.
     In this continuing story today, we have two presenters,
both of them from the man who wears the star at Texaco, and
we are going to talk about some applications and concerns
about analysis of drilling fluids and drilling muds and
cuttings.
     Warren who is our first one up, followed by Mike
Stephenson, are both involved in this program which we are
trying to implement in relation to both offshore oil and gas
and also, in the next year or 2 or 500, onshore oil and gas
regulations.
     So, Warren, would you like to lead off?

-------
                             301
                              MR. HALTMAR:  Thank you.
     Good afternoon.  Today, I will be talking about a
method for determining diesel oil in drilling fluids Method
1651, Revision A.  This is a proposed method for determining
the total oil and identifying and determining diesel oil in
drilling fluids which include muds and cuttings.
     Proposed methods, I always believe, are not actual
facts but are expected to undergo modifications.
     The detection limits for this method are 200 mg of
total oil and 100 mg of diesel oil in our samples.  This
amounts to, if are taking a 20 gram sample, 2 mg of diesel
oil.
     Again, this is a proposed method.  I got involved in
this last year in validating the GC portion of this method.
This year, I became a participant in a round-robin study to
determine how effective this method is.
     Today, I will talk about only the high points of the
method, retorting or distillation of the sample, the
gravimetric determination for total oil, and then the
capillary gas chromatography method for determining diesel
oil.
     Then I am going to talk about possibly a few
modifications to this method.  Some of them are drastic
modifications.  I really originally thought they might be a

-------
                             302
pie in the sky type thing, but maybe they are more down to
earth than I originally thought.
     Finally, I am going to bring you my conclusions about
the method.
     For those of you who are not familiar with drilling
fluids, the definition of a drilling fluid is it encompasses
all of the compositions used to aid the production and
removal of cuttings from a borehole.
     For our case, we are looking at a water-based mud, and
this is made up of three major components:  water, clays,
and additives.  The additives can be weight enhancers,
viscosity enhancers, viscosity reducers, fluid loss agents
and lubricating agents.  Diesel oil is not  supposed to be
added to these drilling fluids.
     Occasionally, a driller will add it because he has
problems.  The problem is he is getting a drill bit stuck in
a hole, he left his lubricating agents back on shore, he is
off shore, he only has a finite time to do something.  So,
he sees diesel oil is handy, and he mixes it in the drilling
fluid.
     This first step of the procedure is to retort the
sample.  Retorting is a simple distillation procedure, which
is semi-quantitative at best.  The device has been around
the oil industry for a long time and consist of a metal
sample cylinder, an aluminum condenser and a graduated

-------
                             303
cylinder to correct the distillate.  The drilling fluid
sample is retorted until liquid stops coming from the
sample.
     Once we have our distillate.  Total oil is  determined
gravimetrically.  We extract the distillate with methylene
chloride.  The methylene chloride removes the hydrocarbon
material, which includes total oil and diesel oil from the
distillate.
     The methylene chloride is removed using a Kuderna-
Danish concentrator to give a gravimetric determination of
total oil.  I was unsuccessful in removing all of the
methylene chloride from the sample using this device.  The
Kuderna-Danish Apparatus is a concentrator not a device to
bring to dryness.
     What I finally did with this apparatus is I took the
receiver off the bottom after I had removed as much
methylene chloride as I could.  I put it in a sand bath,
warmed it gently, and blew dry hydrogen on it.  I knew when
I did that I was probably going to lose some of the light
ends of the diesel oil, and, indeed, we did.
     One of the interesting things that I did with the
Kuderna-Danish is that once I removed all the methylene
chloride, I still had methylene chloride in my sample.  I
took a sample of that, injected it into my GC, and I got a
chromatogram very similar to what I was expecting from my

-------
                             304
diesel oil.  So, it really didn't look like the retort was
having no adverse effect on the diesel oil.
     For the gas chromatography, section of the procedure,
we redissolve the sample in methylene chloride and added an
internal standard, 1,2,5-trichlorobenzene.  We inject this
into our GC.  We essentially fingerprint the sample, to
identify the diesel oil by looking at the alkanes.  We are
looking at ten alkanes eluting after the internal standard.
     Gas chromatography conditions, we are using a DB-1
column or equivalent, 30 meters long, 0.25 mm inside
diameter with a film thickness of 1.0 micron.
     The method calls for a split ratio is 0 to 120.  When I
did the verification on the calibration method, I found out
when we use the high split ratios, we came up with the old
bugaboo of discriminating distillation.  If we reduced the
split ratio down to 40:1 or less, then eliminated the
discriminating distillation.
     Our detector and injector temperature is 275.  Our oven
temperatures are programmed from 90 to 250 degrees C at 5
degrees per minute.  This can change.  There is a criterion
in the procedure that calls for relative retention times to
elute at a particular time.  The internal standard must
elute at 8 minutes.
     I had some difficulties making my G.C. meet the
retention time requirements of the method.  Because of the

-------
                             305
many variables associated with G.C. retention times, many
manufacturers test mixes would be a better check for G.C.
and column operating efficiency.  This is especially true
when analyzing complex samples.
     Finally, a data system is needed that can store and
process data from the G.C.
     A typical chromatogram.of f2 diesel is this right here.
We have our internal standard.  We are looking at these 10
alkane peaks.  We start with C12.  The method does allow
that if we run into some interferences with these alkanes,
say, from a crude oil, that we can use some of the lesser
peaks, possibly some of the aromatics here.  The aromatics
are really the compounds that could be harmful to our
environment.
     The data we obtained on our calibration samples is
shown here.  The first three standard samples have relative
retention times and response factors in agreement with the
procedure.  The last sample is our lowest standard and
failed to meet the requirements for response factors.
     There is a caution here that I think everybody needs to
be when aware of using data systems.  Data systems  can have
problems picking the correct baseline for samples ranging
over a wide range of concentrations.  The expanded
chromatograms show an alkane region in and out of the

-------
                             306
calibration range.  Large differences do exist because of
baseline differences.
     Taking a look at the data generated from the method, it
was found that the light ends of the diesel oil were lost.
This occurs when the methylene chloride is evaporated to
dryness.  This data was obtained by spiking mud samples with
known amounts of diesel.  As can be seen, 40 to 50 percent
of the diesel oil is lost during the gravimetric analysis.
     Earlier, I stated that the retort device is a
semiquantitative device.  We ran a blank down here of just
water, and lo and behold, we find out that there are 61 mg
of diesel oil in that water blank.
     To see what effect this loss has on our chromatogram,
we took two samples, spiked with known weights of diesel
oil.  One of them was dissolved in methylene chloride; one
of them we carried through our retort procedure.  What we
want to see is differences in the chromatograms of the two
samples.
     Here we have our internal standard, and you can see
this chromatogram is very, very similar to what I showed you
earlier.
     Now, if we carry diesel through the retort, the
extraction, and the concentration, this is our diesel.  We
have lost all the light ends.  This is our internal
standard.  Before, on the other chromatogram, the maximum

-------
                             307
alkane was peaking about 17 minutes retention time.  Here,
we have pushed our maximum alkanes out to 20 to 24 minutes.
So, we have drastically altered the diesel oil.
     If we look at the aromatic fraction in here and compare
the two different samples/ we see that we have also lost
part of our aromatics.
     Some of the modifications that could improve the
method.  We need larger samples.  Two mg of diesel is a
rather small sample to work with.  The retort evidently is
retaining some of the diesel oil.  So, possibly, a glass
distillation device would be better.
     The split injection mode of the gas chromatograph
reduces sensitivity.  Performance and sensitivity could be
improved by reducing the split ratio or by splitless or on-
column injection.  If additional sensitivity is needed then
widebore or megabore columns could be used.
     The problems were encountered with one internal
standard, but in complex matrices it is advisable to use
multi-internal standards.
     Working through this method, it became apparent that
the internal standard should be added to the drilling fluid.
This would allow tracking the sample all the way through the
procedure, not just the G.C. portion.

-------
                             308
     The method uses only an FID detector.  Because the
aromatics are the compounds of concern, it is suggested that
a PID detector be used along with the FID detector.
     One of the major modifications would be to use FT-IR.
FT-IR's with the latest in software techniques, like
partial-least-squares could be used to determine solvent
(methylene chloride), crude oil and diesel oil.  This
procedure would still need a suitable extraction technique.
     Another possible method would be fluroescence.  This
would monitor only fluroescening compounds (aromatics) a
non-fluroescening solvent would be needed for the
extraction.
     Finally, a method that I am not too familiar with is
SFE/SFC.  This method appears to be ideal for this type of
analysis.  The method could be one continuous procedure
using on-line extraction.  The chromatography could give
chromatograms similar to the G.C. or it could give only
total saturates and total aromatic.
     In conclusion, I was unable to determine how effective
the retort separated the hydrocarbons from the drilling
fluids (gravimetric analysis).  Chromatograms of diesel oil
versus retorted diesel oil were similar.
     Methylene chloride is a problem with the gravimetric
analysis.   A little bit of error can cause a large error in
total oil and diesel oil.

-------
                             309



     The gas chromatography proved satisfactory with one



internal standard with low split ratios.



     If you all have any questions now, I will be glad to



entertain them.  Thank you.



               MR. TELLIARD:  Thank you,  Warren.  Thank you



very much.

-------
                                      310
      LABORATORY DETERMINATION OF DIESEL OIL IN DRILLING FLUIDS
                           BY:  WARREN C. HALTMAR
                                  ABSTRACT

Drilling muds and/or drill cuttings occasionally become contaminated with diesel oil, which
contains toxic compounds and could be harmful to the environment if discharged in an
improper manner. The Environmental Protection Agency (EPA) has proposed a method
(Method 1651) for determining total oil and diesel oil in drilling fluids (cuttings and muds).
This method involves the retorting, gravimetric determination of total oil, and the capillary
gas chromatography determination of diesel oil.

Detection  limits are  estimated to be 200 mg/kg for total oil and 100 mg/kg of diesel oil
in the drilling fluids.  This translates to 2 mg of total oil or 1 mg of diesel oil per 10 gram
sample. After retorting the distillate (water and oil) is extracted with methylene chloride
to separate oil from the water coming from the drilling fluid. The methylene chloride is
evaporated to give a gravimetric determination of total oil, which includes diesel oil.
The dried  extract (oil) is dissolved into methylene chloride with an internal standard and
analyzed by split injection capillary G.C..

This 5s a  proposed method and is  expected to experience modifications before final
acceptance. The merits and possible modifications of this method will be discussed.

-------
11:
                                               oo

-------
       METHOD 1651 REVISION A
^•ytjSisSv*;-' ,.•
sS&fyivtXi-:?!^'-:

fl^p^sfc
?tf<&""*iVJ?-'fV>i."lJX.<'<-.t '
 Ki^

-------
co

-------
                        DEFINITION
; ,-iv*, *;•< ,,?{:./iffi&%fy$?ffi
P^^^l^»i|p*iii
 llffi^simlM
WiseiiWl^roii^
• * ••v.i'if"1!-/„''^^'^T^X.•*,-..,' • ^, ' 1 '"":';" -'x'"..^.1'^" "^''^S,,^'1;*?:^ jil "'•;:1;1.'^.^;,1:?

-------
315

-------
OJ

-------
317

-------
G. C.  CALIBRATION DATA FOR #2 DIESEL
         PEAK




         ISTD



            1




            2



            3




            4




            5



            6




            7



            8




            9




           10
                  760 STD
                300 STD
                      150 STD
RELRET




    1



  2.64




  4.99



  7.48




  9.98



 12.41



 14.98




 17.01



 19.15




  21.2




 23.16
  RF








0.013




0.013



0.024




0.018




0.018



0.018




0.012



0.009




0.005




0.003
RELRET




    1



  2.62




  4.97



  7.45




  9.94




 12.36



 14.73




 16.97



 19.12




 21.18




 23.15
  RF








0.016




0.016



0.030




0.021




0.020



0.014




0.009



0.009




0.005




0.002
RELRET




    1



  2.62




  4.96



  7.43




  9.92




 12.35



 14.71




 16.96



 19.11




 21.18




 23.15
  RF








0.019




0.019



0.033




0.023




0.023



0.023




0.016



0.012




0.009




0.004
                               60 STD
RELRT




   1



 2.79




 5.13



 7.60




 10.09



 12.53



 14.88




 17.13



 19.26




 21.36



 23.34
  RF








0.370




0.393



0.253




0.407




0.458



0.512




0.792



0.988




1.467




2.768

-------
SAMPLE
  1
mg DIESEL OIL   mg DIESEL OIL
   ADDED      RECOVERED
                                      RECOVERY
408.7
384.3
403.0
404.0
0
205.4
230.7
228.8
262.6
61.2
50.2
60.0
56.8
65.0

                                                feri

-------
ie.ee  '
 9.00
 8.00
 7.00
 6.00
 5.00
 4.00
 3.00
 2.00
 1.00
 0.00
                      SAMPLE  D-l
                Jj
                                                    to
                                                    to
                                                    o
                    i	h
H	1	1	h
     0.00 4.00 8.0012.0016.0020.0024.0023.0032.003S..0040.00
                    RETENTION TIME (MINUTES)

-------
10.09
 9.00
 8.00
 7.00
 6.00
5.00
4.00
3.90
2.00
1.00
0.00
                    SAMPLE  D  STD
                                           kA*~l~_J.
                        1	1	1——i	1	1-—i
u>
to
    0.00 4.00  3.0012.0015.0020.0024.0023.0032.0036.0040.90
                    RETENTION TIME (MINUTES)

-------

-------
CO
CN
ro

-------




-------
                             325 "



                                   MR. TELLIARD:  In the



same drilling mode, Mike Stephenson is going to come up and



explain how nice the method is and how he really likes it.

-------
                             326



                              MR. STEPHENSON:  Warren got up



and told you something about the nuts and bolts of this



Method 1651.  What I want to talk about is the validation of



the method.



     In order to get to where we can talk about the



validation study that we are undergoing, I want to go



through a little bit of history that is involved with this



elusive Method 1651.



     The purpose of the method is to identify the presence



of diesel oil in drilling fluid and to quantify the amount



of diesel oil present.  Diesel oil is a highly aromatic



fuel, and aromatic compounds are known to be bad for the



environment.  One of the most famous ones is benzene.



Naphthalene is a known poison.  They use it in moth balls.



     So, we don't want to put these in the ocean.  The



agency saw the wisdom of doing that, and banned the



discharge of drilling muds that contain diesel in them.



Very logical.



     The next thing we had to do was say, how do we know



whether there is no diesel in it?  That means you have to



have a way of analyzing for it.



     So, back in...I think the first official proposal of



this method was in August of 1985.  It was unofficially



proposed some time before that, and we have been sort of at

-------
                             327



cross purposes with Bill on it since the first day he



mentioned it.



     One of the things that brought us to cross purposes was



the method that we were talking about using.  We were



analyzing a highly aromatic fuel by using aliphatic



compounds as the labels.  The agency's recommended



substitute for using diesel fuel was to use mineral oil



which has very low to no aromatic content.  Since it doesn't



have any aromatics, it has as lot of aliphatics which makes



it look almost the same as diesel.



     In fact, one of the studies that we conducted as an



industry showed that from the chromatograms, you couldn't



tell the difference between diesel and mineral oil in some



cases.



                                   MR. TELLIARD:  If you



were blind, you couldn't.  That is true.



(Laughter.)



                                   MR. TELLIARD:  There was



an incident.



                                   MR. STEPHENSONs  Okay,



there were two.



     After the method was originally proposed, one of the



things that we as an industry tried to convince Bill and the



rest of the agency, was that we needed to be able to use



diesel for periodic operational requirements.

-------
                             328



     It costs a lot of money to move these drilling rigs



around out there, so we like to park in one place and drill



two or three holes from one location.  In order to do that,



you do a lot of bending of your pipe and drilling off at



angles.  You can imagine, if you will, a bent shaft with



something that is inside it turning, that it ends up binding



periodically.



     In order to reduce that binding, historically, one of



the best lubricating agents that we knew as an industry was



diesel oil.  It would unstick our pipe, and we could get on



with drilling pretty effectively.



     If we put a slug of diesel in the mud to unstick our



stuck drilling pipe, then if we recovered all of that



diesel, we should end up with no diesel left in the mud.



And it turned out that we ended up having to recover a large



slug of mud.



     At any rate, I am not here to talk about the diesel



pill monitoring program.  That was a long winded program



that did a lot of analyses, and it is the source of a lot of



the operational history with Method 1651.



     That method came out and was published.  As an



industry, we had a few problems with it.  There were



typographical errors in a couple of the formulas and...



                              MR. TELLIARD:  Conclusions.

-------
                             329



                                   MR. STEPHENSON:  Yes,



conclusions.



     Also, one of the things that it didn't do was address,



from our viewpoint, the possibility of interferences.  If we



were using a recommended substitute such as a mineral oil in



our drilling fluid and we happened 'to get lucky and drill



through an oil bearing zone, then we ended up with a little



bit of oil in the drilling fluid.



     How can you tell that what you measure isn't diesel?



Or, how can you tell if there is any diesel there or how



much is there?  You end up with an interference problem from



the formation oil.



     If you drill a dry hole and don't find oil, it becomes



a pretty simple system, but if you find oil or you put a



little diesel oil and a little mineral oil in the mud, it



becomes very complicated.  And that issue is not resolved



yet.



     At any rate, we made our comments on the first



publication , and they came out with Revision A.  When they



came out with in October of 1988, we had 14 rigs that were



out in the Gulf of Mexico drilling; drilling not using



diesel and not using oil based muds.



     So, we went out and took some samples.  We took a



sample before they hit an oil bearing formation, a sample



after they hit an oil bearing formation, and a sample at the

-------
                             330
end of the well.  We were going to analyze all those samples
using Method 1651 to see how much total oil was in the mud
and if there was any problem with interferences with diesel.
We were trying to convince Bill that we needed to do
something else.
     When we started out, and we had difficulty with the
method.  Our contract laboratory spent about  $30,000 just
trying to do the calibration according to the method.  At
that point, we decided he could stop trying to calibrate.
     We had an emergency meeting with Bill and Dale
Rushneck and the contractor that was doing the work.  We got
a little advice from Dale on how to get on the right track,
and our contractor went back and tried some more.  We
finally decided to stop.
     At that point, I said, okay, we have a problem.  I knew
I had a good analytical chemist.  So, I called  Warren, and
I gave him the method.  I asked him to go down to the local
Texaco station, get some diesel, and do the calibration.
However, he couldn't calibrate following the letter of the
method, either.
     So, I said okay, that is wonderful.  Then I asked him
to go back and see what had to be done to make the method
calibrate.  So, he did.
     Then we had another meeting.  We are good at meetings.
We meet all the time.  We had another meeting, and we made

-------
                             331



some modifications to the method.  At that point we decided



that with those modifications, we needed to do a study by



several laboratories of this whole method.  The purpose was



to determine just exactly how good the method was and what



it was going to accomplish, and we weren't even going to



talk about such things as interferences at this time.  We



were just going to see if we could do straight analyses.



     So, what I am talking about in this validation study is



the purpose of the study, the laboratory samples, the



control of the study, and what results we intend to get.



     As you see, we had a joint study, industry and



government.  You can see Bill is the guy over there in the



blue coat and the white beard.



     The purpose of the study was to determine the



difficulty of the method.  One of my colleagues, Dan Caudle,



and I have maintained for a long time that this was a very



difficult method.  So, one of the things we wanted to find



out because Bill wanted to show that it really wasn't all



that difficult, was just what the degree of difficulty of it



was.



     We also wanted to find out what the reproducibility of



the method was.



     And finally, we wanted to get some rough idea of the



detection limits, because we weren't comfortable with the



limits that were set forth in the proposed method.

-------
                             332



     And Bill said, okay, I will let you do that.  I will



even donate one laboratory to do that.  Very magnanimous of



me.



                         MR. TELLIARD:  Nice guy, nice guy.



                         MR. STEPHENSON:  Yes, nice guy.  He



is here to help us.



     The laboratories that we are using for the study



include, of course, Warren Haltman of Texaco as one of the



laboratories.  Conoco Research Center up in Oklahoma is



another one of the laboratories.  Since we already have



laboratories that have good analytical capability, and since



we are not going to be reported on these results, and we are



just doing some research, we figured that we might as well



use them to do this little study.



     We did use an independent laboratory, CORE Laboratories



in Lafayette which used to be known as Weintritt Testing



Labs.  That was the laboratory that did the bulk of the



testing in the diesel pill monitoring program, and that was



the laboratory that was trying to do the calibration in the



14 rig study.



     Then, of course, we have Bill's laboratory, Dave



Thompkins at ETS Analytical Services.



     And then as a matter of interest, there is a new



technology that has arisen at Ruska Instruments.  They call



it thermal chromatography, but it is controlled distillation

-------
                             333
through a detector.  It is an interesting device, and he has
some data that shows good results.  In fact, it is so good/
Bill to uses it for another method.
     Since Dan Caudle from Conoco and I had been arguing
with Bill about this for several years, we decided that we
were a little prejudiced in what was going on; and since
Texaco and Conoco had provided two of the laboratories to do
the testing, we ought not be involved as being the control
officers.  That would mean we would know what the unknown
samples were, and we might prejudice our own people.
     So, we asked Joe Raia from Shell Development Company
act as the industry control officer.  Our good friend here
acts as the other one.
     The next topic is the results that we intend to
achieve.  Each of the laboratories, when they turn in their
analysis package, are supposed to turn in the chromatogram
of each run, the oil content from each retort, the diesel
content as a result of doing the GC runs and doing the
calculations as set forth in the method, and the QA and QC
requirements of the method.  That was the way we set up the
program with triplicates and so forth, to make sure we got
sufficient data.  This program provides 23 retorts and 23
G.C. runs.

-------
                             334
     At $250 a run, it costs the industry about $5750 for
CORE Laboratories.  I would hate to see what Warren has
spent on the testing.
     What about the data received to date?  Well, one
laboratory, CORE Labs, has turned in its data.  The others
are still working on it.  I saw Dave Thompkins the other
day, and ETS is still working on it.  Texaco still has not
finished all of their analyses.  Conoco hasn't finished
theirs.  We have heard from Ruska and they are still working
on it.
     It is taking a little longer to do this than we first
envisioned.   But, when we get all the data  in to Bill and
Joe, we are going to get back together again, probably at
Texaco, to discuss all of this data and what it means.
Hopefully, we will come to some way of either modifying the
method to make it suitable to all of us and something we can
live with or inventing some new method.  Who knows?
     At any rate, I want to thank Bill for inviting me to
come speak and for being our control officer.  I want to
thank Dan Caudle who dragged me, kicking and screaming, into
this debate; Joe Raia from Shell who worked with Warren a
great deal on some of our comments and on the recommended
revisions to this method; Warren who has done a lot of work
and consultation with me; the API for paying for Weintritt
Laboratories; and, of course, the EPA for paying for ETS.

-------
                             335



     That is basically all I wanted to say, except I wanted



to know if you had any questions, because we have a bunch.

-------
                             336
                 QUESTION AND ANSWER SESSION
                              MR. HOPPE:  My name is Eric
Hoppe.  I am with Battelle.
     I am kind of wondering if I am missing a point here.  I
am not sure I understand what the difficulty of the method
is.  I think there are a lot of commercial laboratories that
have done this for years and years, although I don't think
there has been any check as far as their recovery limits or
anything.
                              MR. TELLIARD:  We have not met
before the meeting, right?
                              MR. HOPPE:  No, we have not.
(Laughter.)
                              MR. TELLIARD:  And before you
went to Battelle, you worked in which part of the EPA?  Just
kidding.
                              MR. STEPHENSON:  Actually, the
basis of one of our sets of comments describing the
difficulty of this method was a Battelle report, just to
keep things in perspective.
     It turns out that we end up doing things like
evaporating to dryness in a Kuderna-Danish.  A Kuderna-
Danish, you know, is not a dryness device.  It is not
designed for that.  It doesn't do that well.  Hence, it

-------
                             337



doesn't do it.  And that is one of the problems we have with



the method.



     Another problem is the fact that the retort is not a



fully quantitative distillation procedure.  The recoveries



that we get here, the recoveries that Weintritt got in their



work were low.



     So, you know, there are some problems with it.



     One of the reasons we went to doing the retort...maybe



this will help clarify it and why it becomes such a problem



to us...we started out initially...and Bill had this idea



for doing this...and we wanted to have a method that we



could do out on this platform bobbing up and down on the



ocean.



                                   MR. TELLIARD:  You didn't



like my first idea.  I recommended we wouldn't have this



difficulty if they just used GC/MS.  They didn't think it



was a good idea to put a mass spectrometer on an oil rig.  I



thought they had a very narrow window to look through.  You



had room for a Coke machine, but you didn't have room for a



mass spectrometer.



(Laughter.)



                                   MR. STEPHENSON:  What can



I say?  It is a matter of priorities.  We didn't want to do



this in the first place.

-------
                             338
     And we may end up going back to GC/MS.  I can see it
coining.
                              MR. TELLIARD:  Five years.
     Go ahead.  It is your show.  I shouldn't be heckling
you from the side.
                              MR. STEPHENSON:  That is all
right.  I heckle you.
     So, we chose what would be a method we could do out on
this rig that is bobbing up and down out there in the ocean.
One of the things that the companies that sell us our
drilling muds and all the components have is this retort.
It was designed as a means of measuring the amount of solids
that were in the mud.  All they did was put a weighed mud
sample in this device, and they heated it up, and drove off
all the liquids.  Hence, they could weigh what was left,
because it was a tared flask in the beginning, and knew what
their solids content was.
     We decided we would pervert this method and use it as a
method of collecting the oil and water.  The water we didn't
care about, but the oil was important.  So, we tried to use
that as our distillation, our oil recovery method from the
mud, and it has its problems.
     The other thing is we wanted a simple GC method.  GC is
a pretty simple thing.  Graduate students use them;
undergraduate students use them.  I know of a trucking

-------
                             339



company that that is how they monitor what comes in on the



18-wheelers off the road.  They have some guy who didn't



finish high school yet who takes a sample off of this truck



and goes and runs it in a GC, and he interprets the data,



and they spend thousands of dollars based on his



interpretation.



     So, we figured a GC method was a good idea.



Howsomever, by the time we got through with all the QA, QC,



and everything else that we cranked into it to make it a



good, reliable EPA method, it was no longer a field method.



The calibration restrictions are also extremely tight.



     So, we are shifting the samples back to shore to be



analyzed anyway.  I think Bill is going to win.



     Did I cover it?  I still haven't answered it?



                                   MR. HOPPE:  I still don't



know what the problem is.



(Laughter.)



                                   MR. TELLIARD:  You can't



make up for the weaker players is the problem.



                                   MR. STEPHENSON:  We have



been trying to compensate for Bill, but...



     Maybe at the break, you can talk to Warren.  He has



been doing this, and he can maybe discuss some of the



details.  As I say, he is the nuts and bolts man.  I am the



guy who has to go meet with Bill.

-------
                             340



     Do I have any other questions?



(No response.)



                              MR. TELLIARD:  Thank you,



Mike.  Thanks a lot.

-------
  VALIDATION OF A METHOD
 FOR THE DETERMINATION OF
DIESEL  OIL  IN DRILLING FLUIDS
        M. T. STEPHENSON
          TEXACO INC.
             CHAiRMAN
        TECHNOLOGY WORK GROUP
  API OFFSHORE GUIDELINES STEERING COMMITTEE

-------
     PURPOSE OF METHOD
  IDENTIFY THE PRESENCE OF DIESEL OIL
  IN DRILLING FLUIDS

  QUANTIFY  THE DIESEL OIL PRESENT
                                    CO
TEXACO INC.

-------
   HISTORY OF THE METHOD
  • ORIGINAL PROPOSAL - AUG 26, 1985
  • DIESEL PILL MONITORING PROGRAM
  • API COMMENTS
  • REVISION A
  • 14 RIG STUDY
  • TEXACO  RESEARCH
to
£*
GO
TEXACO INC.

-------
       VALIDATION STUDY
            PURPOSE
            LABORATORIES
            SAMPLES
            CONTROL
            RESULTS
TEXACO INC.

-------
      PURPOSE OF STUDY
• DETERMINE DIFFICULTY OF METHOD


• DETERMINE REPRODUCIBILITY OF METHOD


• DETERMINE "ROUGH" DETECTION LIMITS
u>
*•
Ul
TEXACO INC.

-------
 PARTICIPATING LABORATORIES
   TEXACO E&P TECHNOLOGY DIVISION
   CONOCO RESEARCH CENTER
   CORE LABORATORIES - LAFAYETTE, LA,
   ETS ANALYTICAL SERVICES
   RUSKA INSTRUMENTS
TEXACO INC.

-------
        STUDY CONTROL
      INDUSTRY CONTROL OFFICER:
      J. C. RAIA


      AGENCY CONTROL OFFICER:
      W. A. TELLIARD
to
*»
-J
TEXACO IMC.

-------
        STUDY RESULTS
  • CHROMATOGRAMS OF EACH GC RUN

  • OIL CONTENT ANALYSES

  • DIESEL CONTENT ANALYSES

  • QA/QC

  • DATA RECEIVED TO DATE?
w
£*
00
TEXACO INC.

-------
      ACKNOWLEDGMENTS
   W. A. TELLIARD, USEPA
   D. D. CAUDLE, CONOCO INC.
   J. C. RAIA, SHELL DEVELOPMENT CO.
   W. C. HALTMAR, TEXACO INC.
   AMERICAN PETROLEUM INSTITUTE
   ENVIRONMENTAL PROTECTION AGENCY
CO
£*
VD
TEXACO INC.

-------
                          350
o
LU
                                                  o
                                                  z

                                                  o
                                                  X
                                                  UJ
                                                  I-

-------
                             351



                                   MR. TELLIARD:  Our next



speaker is from EPA and our lab at RTF, Larry Johnson who



has spoken with us before.  It has been a while since he has



been back.  Larry is going to talk to us on some air source



methods they have been working on at RTF.

-------
                             352
                              MR. JOHNSON:  As Bill
mentioned, we are the air methods people.  Of course, we
don't really talk all that much with the water people except
when we get water in our air and they get air toxics coming
out of the water.
     Some of the push for looking at air samples comes from
a couple of different directions.  There is  lot of interest
now because of the new Clean Air Act that is coming through,
air emissions from Superfund sites, air emissions from
lagoons.  The sewage sludge incineration regs are well into
coming out, and we have been looking at the emissions from
incinerators.
     What I am going to try to do today is just a  quick
look at some of the kinds of samples that you might see
coming into your laboratories related to air.
     Just like with water samples, air samples have three
major phases that have to be dealt with.  The sampling part
of the operation is greatly different, of course, than
taking samples out of water or sludge.
     The preparation steps, you will find as we get into
them, are going to look fairly familiar to you with just a
few differences.  You will be getting new kinds of samples,
but you are going to use old familiar techniques, although
on different matrices.

-------
                             353
     After the sample is extracted, the analysis methods are
really quite similar except in fairly rare cases.
     I am going to go through some of the sampling methods
just briefly to show you what kind of equipment these
samples are taken with and, therefore, where some of these
samples are coming from.  I am skewing this discussion quite
a bit towards stack emissions mostly because that is our
specialty.
     In the interest of time, I am going to talk mostly
about organics, and only the more frequently used sampling
trains.  There are a lot of other trains available,
especially if you start trying to look for a wide variety of
compounds.
     So, I am not going to make any pretense that this is
complete.  We are just going to hit the high spots,
basically.
     Most of the work has been done on stack emissions from
combustion sources.  We started out working on power plants.
We worked on a whole variety of combustion sources, and a
lot of our funding and, therefore, our focus in the last few
years has been hazardous waste incineration related.
     The new Clean Air Act is going to focus attention
combustion stacks, but also vents and similar sources.
     Where are these methods provided?  In fact, somebody
asked me a very good question out in the hall before the

-------
                             354
talk.  Is there a compendium somewhere of all of the air
methods?  The answer is not quite.  A list of the better
references will help.
     Most of the Federal Register methods that are used in
relation to the air part of our program are in 40 C.F.R 60
and 61,  There are air and stack methods in SW-846, and we
have a number of them in various stages of review and
clearance for inclusion in SW-846.  There are also other
sources.  We have a number of guidance documents and other
references.  In the hard copy of the paper, I will list a
few of those.
     Also, there is an operation called EMTIC, Emission
Measurements Technical Information Center, that has been set
up at RTF, and its function is primarily to supply methods
information to the States and to the regions.   The number
is (919) 541-1059.
     One of the methods I am going to talk about is Method
0010/ and that is an SW-846 method.
     We are also going to talk about 0030.  A lot of stack
samples and some ambient samples are taken in Tedlar bags.
Metal cylinders are very popular, and sometimes they are
great and sometimes they are terrible.  Usually, for stacks,
they are terrible, but we will talk a little bit about their
use for ambient sampling, too.  Likewise, sorbent tubes are
used for source and for ambient sampling.

-------
                             355
     A lot of the time we will spend on 0010.  This is a
very flexible method, so a lot of the samples will come in
from this hardware.  It is also one of the more complicated
methods, so you will get more subsamples to deal with.
     It is the same thing as Modified Method 5.   The term
Modified Method 5 became so ambiguous that we had to quit
calling it that.
     It is also the same train as the so-called Semi-VOST
which is less ambiguous but a poor name also.
     This method is used for compounds with boiling points
greater than 100 degrees C.  So, we are talking about what I
call the semi-volatiles and the non-volatiles.  It doesn't
necessarily correspond exactly to what you would call those
things, based on water analysis terminology.
                                       »'^
     This is a diagram of the Modified Method 5 train.  If
you haven't seen one of these stack sampling train before,
it looks like a glass blower's nightmare, but, believe me,
you need all of these pieces there to get the samples.
     From your standpoint, the important thing is what
samples will be generated from this device probe rinse
sample comes from this part of the train,  a sorbent or a
filter sample is also generated there.  The sorbent is
usually XAD-2.  A condensate sample is also produced.
     We will talk a little about how each of those sample
types is handled in the laboratory.

-------
                             356
     Solvent extraction is used to recover organic compounds
from the samples.  Usually, it is methylene chloride, almost
a universal solvent, which, sometimes, we do have to follow
up with another solvent.  In the vast majority of cases,
methylene chloride is the solvent of choice for all the
usual reasons.
     The probe rinse has two sub-samples to it, in effect.
The particulate material, and the solvent that the probe was
rinsed with.  The rinse is filtered and the two sub-samples
dealt with separately.  Usually, they are combined in with
some of the other subsamples.  The solvent is combined with
an extract, and the particulate is combined with the filter.
Depending on the information needed the combination might be
different.
     The sorbent and the filter are where most of the sample
is collected in this train.  If you have any luck at all, it
will be distributed between those two, and you should never
try to figure out what this means in terms of what was in
the vapor state and what wasn't in the vapor state in the
stack.   It is totally meaningless if you do try that, and
unfortunately, a lot of people do.
     Again, the samples are Soxhlet extracted.  We went out
of our way to misspell Soxhlet here.  I think about four of
senior scientist types looked these things over and didn't
see the mistake until I started packing the slides.    The

-------
                             357



sorbent which is usually XAD-2 is well-behaved compared some



solids like soils and sludges.  XAD-2, assuming it has been



cleaned and treated properly, is consistent from batch to



batch, and you know what to expect from it.



     The particulate is another matter.  It is a little bit



like a soil sample in that respect.  It may be very docile



for months and months and then turn around and bite you by



not giving back some of the material that you are interested



in recovering.



     Very much like your water sample, once you have



extracted the sample, you have to concentrate the extract,



and then you have to do an analysis.  We almost always go



with GC/MS or GC, and when we can't, we go with HPLC.  It is



very much like water or sludge analysis in that respect.



     The condensate has passed through the filter, and has



been through XAD-2, so most of the organics are stripped



out.  It is relatively clean.  It is not something you want



to drink, but compared to a dirty water sample with a lot of



organics in it, it is relatively clean, and you analyze this



material just to prove that nothing broke through.  It is



almost a QC sample.



     In some cases, if you have enough history, you can quit



running this sample.  Unfortunately, a lot of people have



quit when they shouldn't, and we are tightening up



requirements on running it.

-------
                             358
     A separatory funnel extraction is required very much
like  a water sample, and like with a water sample,
sometimes  another extraction with a second solvent may be
required.      Once the extract is obtained, it is
concentrated in a K-D and analyzed with one of the unusual
analytical methods.
     Method 0030 is the VOST  or VOST, depending.  The
pronunciation is like tomatoes or tomatoes.   It is designed
for compounds that boil between 30 and 100 degrees C.  We
have found that most of the compounds that boil between 100
C and 132 C will also work, but we don't allow the train to
be used above 132 C, because things up in that range can be
involved with the particulate, and the particulate is
discarded in this method.
     Here is the location of the particulate filter.  You
throw it away.  The sample stream passes through a cooler,
then into a tenax cartridge.  There will be condensate,
which is usually minimal.  This is a low volume train which
is why it will catch the volatiles.  There is a  back-up
tenax tube and charcoal which we included in sheer
desperation.  Most of us have wished we hadn't put it in
there for years now.  We are going to replace it one of
these days.
     Most of the sample typically collects on the front
tenax tube, but you have to analyze the condensate to prove

-------
                             359



it contains no compound of interest, and you always analyze



the second tenax tube, because it is not unusual for it



contains some of the compound of interest.



     You only have two subsamples here, two sets of sorbent



tubes.  In fact, a sampling run usually consists of



approximately six sorbent tubes.



     The condensate must be dealt with, even though some



people discard it when it should be analyzed.  Again, it is



more of a QC sample just to prove that nothing got through



the rest of the train.  It protects against breakthrough,



channeling, things like that.



     This is what the tenax tubes look like.  We actually



adopted the biggest ambient air sampling tubes that we could



find.  There is about 2 grams of tenax in the one, and the



other has about a gram of tenax and a gram of charcoal in



it.



     Method 5040 which, again, is a SW-846 method is used to



analyze these tubes.  It is just a heat desorption method



with a twist.  You heat desorb into a purge and trap



chamber, because these tubes are going to have maybe a



milliliter of water on them sometimes.   You get rid of the



water by running it into water.



     We have tried a lot of more elegant techniques, and so



far, none of them have worked very well for us.  That is not

-------
                             360
to say they wouldn't if we spent enough money to really make
them work.
     We have a draft Method 5041.  5040 is in the book, and
it is a packed column.  In fact, it is the old 624 method.
We didn't want to develop an analytical method.  We were
busy with the sampling method, so we just adopted the
analytical method.
     5041 is a megabore version of 5040, and it will give
you a lot better resolution.  It is under review, and a year
or less it will be in SW-846.
     The condensate is analyzed by purge and trap whenever
possible.  If you have something like acetonitrile that
doesn't purge and trap well, you may have to use direct
injection which I know all of you love.
     Tedlar bags is a viable way to sample, but one problem
with  these bags is that they are deceptively simple, and
there is a multitude of ways to use them wrong.  You use
them only for low boilers because you don't want things
condensing out in the bag.  It is necessary to run through a
water trap if there is any moisture in the stack.  You
usually do the analysis by GC/MS, and if there is any
condensate in the trap, you have to analyze that however you
can, usually purge and trap.
     Some of the sampling trains that are available for bags
right now don't have the condenser, but most of the time,

-------
                             361



you need that if there is any moisture in the stack.



Moisture can't be allowed in the bag, because if any



condensation occurs in the bag, the sample is invalid.



Metal cylinders are very popular.  The Summa polished



cylinders are used a lot for ambient air.  Method TO-14 uses



these devices.  The summa polishing is just a pacifization



technique that is applied to the cylinder.  Cryo-focused



GC/MS is usually used to do the analysis, sometimes just GC.



     Method 25 is an attempt to obtain a total organics



concentration.  A metal cylinder is used, and then an FID



analysis is performed after running the sample through some



treatment to convert it into methane.



     All of these metal cylinder techniques only really work



when you don't have high levels of acid gases like HC1 and



SO2 and other corrosive agents.  High moisture levels



interfere with the cryo-focusing analysis.  Most of the time



canisters don't work well for  stack sampling.



     Sorbent tubes, you can use those same tenax tubes to do



ambient air samples from around dumpsites and off of



lagoons.  The technique is a very similar to Method 0030.



You heat desorb, but don't usually run through the purge and



trap procedure, because the water isn't present..



     The QA/QC is very important for all of these methods,



and we are just realizing how  truly important it is.  We are

-------
                             362
being forced to tighten up on the sampling QC and on the
laboratory workups.
     Because we had clean sample matrices in a lot of cases,
we would typically demonstrate that the technique worked and
then only have minimal safeguards against analysts
performing poorly in the laboratory.  We have found that
that is not enough.  Poor performance has occurred all too
often.
     We have recent guidance on QA, especially related to
hazardous waste incineration.  I will put the references to
that in the written version of the paper.  We are going to
tighten many of the QC requirements and use more labeled
spikes.  There is going to be less combining of subsamples
allowed, and I think where we can, we are going to have to
utilize isotope dilution.
     Isotope dilution procedures are not yet developed for
stack samples, but if funding allows, we would like to work
on them.  We have typically fewer things to look for in a
given sample, so it is not quite as big a cost burden as if
you are looking for 200 or 300 things at once.
     Audit samples are available for a lot of these methods.
VOST audit samples are available at RTF.  They have to be
requested by an EPA person.  So, if you want to be audited
asks an EPA person to call RTF and request the audit
samples.

-------
                             363



     There are audit sample development and validation work



going on right now on the TO-14 canister method for use



around Superfund sites.  There are two different groups in a



real laboratory that are doing that work.  So, there will be



audit samples available.



     Are there any questions?



                                   MR. TELLIARD:  If there



are no questions, thanks a lot, Larry.  We appreciate it.

-------
PREPARATION AND ANALYSIS
          OF
u>
  AIR EMISSION SAMPLES

-------
SAMPLING
PREPARATION

ANALYSIS
CTi
Ul

-------
SAMPLING METHODS

STACK EMISSIONS

ORGANIC

MOST FREQUENT

COMBUSTION CONNECTION
CTl
cn

-------
WHERE?
40 CFR, PARTS 60 & 61

SW-846
OTHER SOURCES
                            co
                            CTl
                            -J

-------
METHOD 0010
METHOD 0030
TEDLAR BAGS
METAL CYLINDERS
                         u>
                         CT>
                         00
SORBENT TUBES

-------
METHOD 0010


MODIFIED METHOD 5
B.P. GREATER THAN 100 °C
                            u>
                            en

-------
                                                                                  Sortnnt
                                                                                  Trap
  Temperature
     Senior
   Probe -*Hr*
Reverie-Type
 Pilot Tub*
                                                                                      Cheek
                                                                                        Valve
                                           Thermometer

                                                Filter Hotrftr
                          Radrculation Pump
                                    Tharmomettrs
                                                               OJ
                                                               ^j
                                                               o
                                                                                            Vacuum Line
                                      DryGai
                                       Meter
Airtight
 Pump
                                        Modified Method 5 train.

-------
PROBE RINSE
FILTER
SORBENT



CONDENSATE
u>

-------
SOLVENT EXTRACTION
                          U)
                          ^J
                          NJ

-------
PROBE RINSE
PARTICU LATE/SOLVENT
                          OJ
                          •-J
                          tjj

-------
SORBENT
FILTER
SOHXLET EXTRACTION
                         GO

-------
SORBENT CONSISTENT MATRIX



PARTICULATE MAY BE INCONSISTENT
                                  (Jl

-------
EXTRACT CONCENTRATION
ANALYSIS
Co
•j
en

-------
CONDENSATE
CLEAN, PREDICTABLE



SEP. FUNNEL EXTRACTION
SECOND SOLVENT
                         CO

-------
EXTRACT CONCENTRATION
ANALYSIS
oo
-j
oo

-------
METHOD 0030


VOST


B.P. 30°CTO100°C


RANGE EXTENDED TO 132 °C
oo
-j

-------
                Heated Probe
 Glass Wool
 Particulafe
 Filter
     ft
    STACK
(or test System)
                           Isolation Valves
Carbon Filter
                          Thermocouple
                          Sorbent
                          Cartridge
                    Condensate
                   Trap Impinger
               Condenser
                Backup
                Sorbent
                Cartridge
                                                     Silica Gel
                                                                   Vacuum
                                                                  Indicator
9
xL-/-Jr=pq=.
 f    \   i
                                                       Exhaust
                              1    2^4
                                   Pump
              Dry Gas
               Meter
co
oo
o
                                Rotameter
                                     Volatile organic sampling train (VOST).

-------
SORBENT TUBES
CONDENSATE
                        u>
                        00

-------
SORBENT TUBES

ANALYZE BY METHOD 5040
DRAFT METHOD 5041
                         U)
                         00
                         to

-------
CONDENSATE

PURGE AND TRAP
                        OJ
                        00
                        oo

-------
TEDLAR BAGS
LOW BOILERS



ANALYSIS BY GC/MS
ANALYZE CONDENSATE
00

-------
                            385
                  COOLER
                CONDENSER
CRITICAL
ORIFICE
                         CONDENSATE
                          COLLECTOR
                                ACTIVATED
                                CHARCOAL
                                 FILTER
                  EXHAUST
 FROM
PROCESS
                                            TO
                                           AMBIENT
              PUMP
     BAG SAMPLE COLLECTION APPARATUS

-------
METAL CYLINDERS
METHOD 25
FID ANALYSIS

METHOD TO14

CRYOFOCUS GC/MS
CO
00

-------
SORBENT TUBES


SIMILAR TO METHOD 0030
HEAT DESORB
Co
oo
•-j

-------
QA/QC

VALIDATION LIMITED

RECENT GUIDANCE

MORE LABELED SPIKES
ISOTOPE DILUTION?
oo
00

-------
AUDIT SAMPLES
                          oo
                          00
                          10

-------
QUESTIONS?
                         vo
                         o

-------
                             391



                                   MR. TELLIARD:  Our last



speaker today...which means this will also be on the



final...is going to talk about the Chesapeake Bay and the



joys thereof.  Tina Fletcher from Region III and the Bay



Office is going to talk about the nutrient measurements that



the folks have been spending a great deal of time on in the



last couple of years.

-------
                               392
      The Chesapeake Bay Program:   Experience with Nutrient
         Analytical Methods in the Estuarine Environment

Five  years into  the mainstem monitoring program  for nutrient,
demand and physical parameters the program  is  involved  in a re-
evaluation of the project design and implementation.  This test of
our assumptions has involved an analysis of the data available for
the detection  of trends  in  nutrient levels which have  a direct
bearing on the availability of oxygen and thus the survival of our
living resources  in the Chesapeake Bay  waters.   Concurrent with
this  trend  estimation  is  a  power  analysis  to determine  the
theoretical sensitivity of the current database with respect to its
temporal and spatial  coverage  and the detection limits available
from the existing field processing and analytical methodologies.
Were the initial  assumptions which set us out on this first five
years of perhaps a twenty year journey correct?  But I am getting
ahead of my story.
Here in Norfolk, we are close to the mouth of this the largest and
most productive estuary in the  continental  U.S.   The Bay  as we
presumptuously call her stretches 195 miles along a mainstem with
7,000  miles of  shoreline 18  trillion gallons of  water and an
average depth of less than 30 feet.   Formed  10,000  years ago as
melting glaciers flooded the Susguehanna River Valley she is home
to  200 species of fish  along with 2,700 other plant  and animal
species,  12 million  people  and their hundreds of  thousands of
pleasure boats.  The  resource  is too valuable to let go and we at

-------
                             393
the   Chesapeake   Bay  Program,   a  consensus   organization  of
representatives from  federal,  state and  local  governments along
with  citizen  groups,  and  universities from  Maryland,  Virginia,
Pennsylvania and the District of Columbia  take that responsibility
very seriously.

The   drainage  basin for the Chesapeake spans 4,500 square miles
from the states of Maryland, Virginia, Pennsylvania,  Delaware, West
Virginia,  New York and  New  Jersey.    Data  are  received  from
environmental organizations in each of  these regions.  The quality
of these data  must  be understood  so that  they may be effectively
used in the management decisions.

From before 1950 through the 1970s many state and university groups
conducted research oriented studies some  of which operated over a
protracted  period  but  most  of   which were  both  spatially and
temporally very limited.  Between  1978  and 1983  the  Chesapeake Bay
Program as coordinated by the Environmental Protection Agency was
in its  research  phase where it sought to characterize the Bay's
water quality, sediment quality and resources.
This period was  devoted to determining what  data were available
from historical sources and to identifying the issues at hand for
the  Chesapeake.    The  infamous  "Appendix  F-  A Monitoring and
Research  Strategy to  Meet Management  Objectives1*, defined the
Chesapeake Bay Water Quality Monitoring Program Objectives as:

-------
                               394
               Characterizing  existing  water  quality
               Baywide
               Determining trends in water quality that
               might develop  in  response to management
               actions   or   additional   sources   of
               pollution
               Integrating  the   analyses  of  various
               monitoring components with a view toward
               achieving    a    more     comprehensive
               understanding of processes affecting
               water quality and  the linkage with living
               resources
In 1984 the Chesapeake Bay Program entered the Implementation Phase
in which the designs of the Research Phase took form.  Within the
committee structure of the Chesapeake Bay Program, the Monitoring
Subcommittee attempted to translate the research questions into an
EPA  funded mainstern monitoring  program for physical,  nutrient,
demand and biological parameters.   The Mainstem Monitoring Program
was based upon the segmentation scheme developed by researchers to
encapsulate the variable portions of the Chesapeake's central trunk
and near tributaries.  The MMP was designed as a 50 station network
which would be sampled 20 times per year.  As with any monitoring
effort,  the  scope of  the  program was  heavily  linked to  the
budgetary constraints.   The extent to which  existing laboratory
capabilities and configurations could be utilized would be pivotal
in moderating the costs of the  program.  Data from the tributaries
would be  taken from the monitoring programs  within  each  of  the
states vastly increasing the reach and complexity of the database.

-------
                             395
The MMP measures more than 25 forms of nutrient, demand, physical
and biological parameters.   The Mainstem Monitoring Program is not
a regulatory program and as  such  there is  no EPA mandated set of
methods.  The  EPA water chemistry manual,  "Methods  for Chemical
Analysis of  Water and Wastes" describes  most of its  methods as
suitable for the  analysis of saline  waters.   However,  the degree
to which  this has  been evaluated  should not  be overestimated.
There were no  Agency guidelines for the  selection  of  analytical
methods.  The Chesapeake Bay  Program was on its own to identify its
needs and to determine the analytical methods which would satisfy
those requirements.

With strong input from EPA and those organizations which would be
implementing the analytical work,  it  was decided by the Monitoring
Subcommittee that EPA water methods and procedures would be adopted
for this program.  There was  a  sense that this  would provide a
continuity with the historical database and would provide a linkage
with the tributary data which was  generated by the principal state
laboratories which were driven by these same EPA methods required
under the National Pollution Discharge Elimination  System.,  the
workhorse of those labs at that time.
While the Chesapeake Bay Program is a consensus organization, this
consensus is not  instantly  achieved.   A series of laboratory and
field  consensus workshops  were held  to  thrash out  the program
design beyond the work done in the Monitoring Subcommittee.  A

-------
                               396
strong objection  to the analysis of these  estuarine waters with
TKN as the measure of total nitrogen was made by Chris D'Elia and
the University of  Maryland representing a strong  linkage with
standard oceanographic methods.  Intense challenge was brought to
the analysis  of  the crucial particulate phase of  nutrients  by
difference  between  the   total   and  dissolved  determinations.
Detection limits were said to be unacceptably high.  However, the
analytical  program  was in place  and  it  was determined  that
"comparability"   would have to  be demonstrated before  a change
would be made.
Change,  the enemy  of trend  analysis.    Nonetheless, change  is
inevitable in any monitoring program.  A learning curve is present
initially  and then  with  each  new  analyst,  each  new piece  of
equipment and so forth.   Trend analysts may argue that consistency
is more important than "truth".   However,  new instrumentation is
inevitable  and  should  be  embraced.    Improvement  in  system
efficiencies  which lead to  lower  detection  limits are  always
desirable as long as the cost equation is not shifted.  It is the
role  of  the  formal  comprehensive  quality  assurance program  to
provide the documentation and management of that change so that the
needs of the data users can still be met.

-------
                             397

In 1986, two  years into the monitoring program,  special studies
outside the normal monitoring program were  initiated to accommodate
the recommendations for change while maintaining a linkage to the
on-going monitoring program.   These  special studies  addressed
recommendations  that   freezing  be  allowed  as  a  method  of
preservation and that  .45 micron  membrane filters be  replaced by
.8 micron effective pore size glass fiber filters.

Further,   special   studies  were  initiated  to  evaluate  the
comparability between  the  EPA methods and those  coming from the
oceanographic protocols.  It  was  determined  that the recommended
changes would benefit the goals of the Mainstern Monitoring Program
for a variety of reasons from improved detection limits, greater
analytical throughput, improved recoveries and even the lessening
of hazardous waste materials.   Field analytical procedures did not
change,  however,  many changes were adopted  in the  laboratory
analytical methods.
As a  reflection of the Appendix  F concerns and with the strong
realization that there was no natural barrier at the tributaries,
a Baywide Quality Assurance  Program was initiated  in 1986.  Work
was begun to bring the tributary monitoring programs in line with
changes  underway  in the  Mainstern  Monitoring  Program.    As  a
beginning a  Coordinated Split  Sampling Program was  designed to
identify the  variability associated  with the processes  used to
collect, process, analyze and report the data which are stored in
the Chesapeake Bay Program database.

-------
                               398
The Coordinated Split Sampling Program utilizes four organizational
components to  collect and distribute split samples to  a varied
group of federal, state and university laboratories who routinely
produce data from the mainstem and tributaries.  The attempt is to
establish the comparability between organizations so that the data
will be of known quality and apparent  differences can be evaluated
for  possible  organizational  artifacts.    Triplicate  samples,
spiking, duplicate  samples and standard reference  materials all
facilitate  the   evaluation   of   sources  of  variability.     An
established  statistical protocol  is  used  to  evaluate  the  data
produced in this split sample program and rapidly turn it back to
the organizations for evaluation and implementation of changes as
indicated.
Some of  the data which have come  out of the  Coordinated Split
Sampling Program  clearly indicate the need for  change.   In the
case of Total Dissolved Phosphorus,  it was quite apparent that the
university laboratories who are responsible for the analysis of the
Mainstem Monitoring Program were reporting data which were nearly
an order of magnitude  lower  than the principal state laboratory.
This is not surprising considering the bulk of the sample load for
that laboratory is from point source and hazardous waste sampling
locations.   State management used the data  from  the Coordinated
Split  Sampling  Program  to  leverage  changes   in  laboratory
arrangements which would  allow the  production  of  data which were
not a function of the-analyzing laboratory.

-------
                             399
Conversely, it has been heartening to see how tightly bunched the
data are for field and lab precision for TDP overall.  Silica split
sample results are hard to separate lab to  lab.  Each of the other
parameters  has its  own characteristics  and trends  point to  a
tightening of limits.  Organizations are using these data to track
and implement changes in their field and laboratory operations.

Split sample data provide  a  measure of the comparability between
the different sampling and analytical organizations which have been
crucial  in  the   evaluations  of  the  trend.    The role  of  the
Coordinated  Split  Sampling  Program  and  the  Baywide  Quality
Assurance  Programs  can  only be  expected  to expand  as the  CBP
attempts  to utilize more  effectively  the  data from  myriads of
historical and on-going  monitoring programs  throughout the basin
as identified in the Chesapeake Bay Basin Monitoring Program Atlas.
The Living Resources Monitoring Plan seeks  to forge a link between
habitat requirements of the biota and water quality.  The quality
of data throughout all  of  these programs must be known to permit
their use in the decision making process.
The Chesapeake Bay Monitoring Program represents an evolving data
collection network and a process of providing information necessary
for management of the Chesapeake Bay's Resources.  If the support
of management decisions is the reason for data collection, then a
database  is  only as  good  as its match  with the needs  of those
decision makers.   Five years into the monitoring program an

-------
                               400
intense process of re-evaluation is in process to be certain that
the data produced to date are of known quality.   To determine if
the monitoring program is able to  detect  a trend if one develops
is the task at hand.  Will the existing monitoring program be able
to track  the 40  % reduction  in nutrients by  the year  2000  as
stipulated by the Governors in the Chesapeake Bay Agreement?  The
users of  this  database  need to be identified  and their concerns
prioritized.  Then, they must stipulate explicitly the quality of
data they require to meet the needs of the sophisticated computer
models or to satisfy the monitoring of the  Living Resources Habitat
Criteria.

The configuration which is represented by the Mainstern Monitoring
Program  in  the  future will  be  a combination  of the  original
perspectives of the Research Phase with clear direction from the
users of this evolving database.  Analytical systems will be honed
to produce data of the quality required by the program,  not just
by the goal of "how low can you go?"
What have we learned?  In the area of monitoring, consensus may be
an impossible form of management but it is essential to the success
of the multi-jurisdictional  nature of  estuarine monitoring.   The
success of a monitoring program is directly related to the clarity
of the  needs  of  the  decision  makers  who will use these  data.
Program objectives and the specific questions  must  be understood
and  translated into  a  network  design.    Trends resultant  from
management actions will rarely be seen in two to five years.

-------
                             401
Natural variability in rainfall and other cycles has a much longer
period.  Monitoring  of this type is there for the  long haul and
must  have  a  long  term  stable  source of  funding  to  provide
consistent implementation of a program of the quality required.

A  comprehensive quality  assurance program  is  crucial for  the
success of any  monitoring program.  A coordinated split sampling
program  is  necessary  to   establish  the  comparability  between
multiple organizations involved in the monitoring  effort.   Data
must be statistically  evaluated and returned to the participants
for action.  The technical  consensus  of experts must be utilized
in the design and implementation of the monitoring program.

Data management cannot be  ignored.  A  well  designed,  thoroughly
documented and consistently  implemented data management system must
be in place  to facilitate the accurate and efficient acquisition
of data and to effectively make the  data available to the user
communities.   Close attention  must  be paid to the handling of
historical data sets where the quality is often unknown.
Data analysis must not be  an afterthought.   Rather, it should be
leading the  process.   Data which  sit on the shelf  often are of
inappropriate  quality since  they have  not had  the  benefit of
feedback from the data analyst.   The strongest organizations are
clearly correlated with those operations who routinely utilize the
data generated  by  the Hainstem Monitoring  Program  for their own
purposes.

-------
                               402
The  Chesapeake Bay Program  experience with  nutrient analytical
methods has been or is being repeated in some  form in each of more
than a dozen estuarine programs nationwide.  The isolation of the
technical communities of the Agency's estuarine programs is clear.
Regions II  and III have been  involved in a year long  effort to
identify  the  analytical  needs  of  the  estuarine  and  marine
communities.   Last week we held with  support from  the  Office of
Marine and  Estuarine Protection the first annual  Estuarine and
Marine Analytical  Methods  Workshop in an  attempt to  establish a
network of technical experts in the estuarine and marine matrices.
The Estuarine and Marine Analytical Methods Committee was formed.
Analytical needs  were coordinated under four workgroups  who are
charged with the compilation of candidate methods, as well as the
evaluation of existing methods  and  SRMs.   Round-robins which can
be  organized within  the community  will be  encouraged  and the
interest and  support on the part of the Office  of  Research and
Development is clear. It is an idea whose time has come as we seek
to promote each estuarine program with through  the experience of
the other in this most challenging matrix.

-------
                             403
I would like to offer special thanks to Peter Bergstrom and Rodney
Buckingham of  CSC  and Rich Batiuk of the  Chesapeake Bay Liaison
Office for their assistance in the preparation of these materials.
                             OVERHEAD

-------
I
         THE CHESAPEAKE BAY PROGRAM
      Experience with Nutrient Analytical Methods
           in the Estuarine Environment

-------
        405
CHESAPEAKE BAY BASIN
      NY
      PA
  WV
MD
                   NJ

-------
                           406
 CHRONOLOGY OF THE CHESAPEAKE BAY
            MONITORING PROGRAM
1950s-1970s    Independent data collection efforts, often more research
             oriented and short-term in nature
1978-1983     Chesapeake Bay Program Research Phase
             •  Characterization of the Bay's water quality, sediment
               quality and resources
             •  Compilation of historical water quality data
             •  Chesapeake Bay Synoptic Survey
             •  Chesapeake Bay Segmentation Scheme
             •  "Appendix F - A Monitoring and Research Strategy to
               Meet Management Objectives"

-------
                        407
 CHRONOLOGY OF THE CHESAPEAKE BAY
       MONITORING PROGRAM  (cont'd)
1984 - Present  Chesapeake Bay Program implementation Phase
1984
1985
Monitoring Subcommittee established
Chesapeake Bay Water Quality /Biological Resource
Monitoring Program implemented in July 1984
Technical Monitoring Program Consensus Workshops
Expansion of definition of the Chesapeake Bay
Monitoring Program to include non-tidal data collection
programs
First Biennial State of the Chesapeake Bay Report
published
Increased emphasis on coordinated data management

-------
                          408
 CHESAPEAKE BAY PROGRAM BASIN SEGMENTS
                              CB-l
                                  ET-1
                     WT-1.
                     WT-2
                   WT-;
                 WT-4
                WT-5
                                    ET-2
    RET -2'
Figure 2.
                                            ET-10
                                           EE-3
          TF-5'
             RET-
                                       B-8

-------
                      409
   MARYLAND CHESAPEAKE BAY MAINSTEM
   WATER QUALITY MONITORING STATIONS
        Washington D.C.
  SCALE 11682,463
0  15   30   45
    VTT.1S

-------
                   410
VIRGINIA CHESAPEAKE BAY MAINSTEM
WATER QUALITY MONITORING STATIONS
 SCALE 1571650
 K
 10    20

-------
                      411
    MARYLAND CHESAPEAKE BAY WATER
QUALITY MONITORING PROGRAM: TRIBUTARY
       CHEMICAL /PHYSICAL STATIONS
                Baltimore
      Washington D.C. MWT8-]

            pxroeos^   4yviwT8.2
                     MWT8.3 V
            PXT0402
        XFB2470

      XFB1433    XED94!
               XDE533

                  XDE279?
  15  30   45
   rnr.re

-------
                     412
VIRGINIA CHESAPEAKE BAY TIDAL TRIBUTARY
   WATER QUALITY MONITORING STATIONS
   10    20
     IflLES

-------
                          413
   WATER QUALITY PARAMETERS MEASURED IN
           TIDAL MONITORING PROGRAMS
Temperature
Dissolved Oxygen
Specific Conductivity
Salinity
pH
Secci Depth
Total Phosphorus
Total Dissolved Phosphorus
Paniculate Phosphorus
Dissolved Inorganic Phosphorus
Dissolved Organic Phosphorus
Total Nitrogen
Total Dissolved Nitrogen
Paniculate Organic Nitrogen
Dissolved Inorganic Nitrogen
Dissolved Organic Nitrogen
Nitrate
Nitrite
Ammonia
Total Organic Carbon
Dissolved Organic Carbon
Paniculate Organic Carbon
Dissolved Silica
Total Suspended Solids
Chlorophyll a
Phaeophytin

-------
                            414
 CHRONOLOGY OF THE CHESAPEAKE BAY
       MONITORING PROGRAM  (cont'd)
1986
Comprehensive nutrient analysis and sample storage
comparison studies funded by the Monitoring
Subcommittee

Baywide Quality Assurance Program initiated
1987
Second Biennial State of the Chesapeake Bay Report
Published

Expansion of program coordination to include
fisheries/shellfish resource monitoring

1987 Chesapeake Bay Agreement signed by the states
and the EPA
1988
Sediment processes monitoring expanded to support
Time Variable Chesapeake Bay Water Quality Model
development

Baywide Coordinated Split Sample Program implemented

-------
          LABORATORY ANALYTICAL METHODS
VARIABLE
EPA METHOD
GBP METHOD
Silica, dissolved  EPA 370.1 (AA)
TOC           EPA 415.1 (Infrared)
DOC           EPA 415.1 (Infrared)
Particulate Carbon TOC - DOC = PC
TSS

TKN
TON

PN
Standard Methods
(Gravimetric)
EPA 351.2
TKN(f) + NO3 + NO2

TKN (W) - TKN (F)
EPA 370.1 (AA)
PC + DOC
EPA 415.1 (Infrared)
 High Temp Oxidat.
  Combustion
Standard Methods
 (Gravimetric)
Not Analyzed
Alk. Pers. Djg.
  EPA 353.2
High Temp Oxidat.
  Combustion

-------
          LABORATORY ANALYTICAL METHODS
VARIABLE
EPA METHOD
GBP METHOD
Ammonia

NO2 + NO3

NO2
TP
TOP

OrthoP
PP
EPA 350.1 (AA)
  Automated Phenate
EPA 353.2 (AA)
  Cad. Red, Diaz.
EPA 353.2 (AA)
EPA 365.1
TP (w) - TP (f)

EPA 365.1 (AA)
TP (w) - TP (f)
Chlorophyll "a"   Spectrophot.
Phaeophytin "a"  Spectrophot.
EPA 350.1 (AA)
 Automated Phenate
EPA 353.2 (AA)
   Cad. Red, Diaz.
EPA 353.2 (AA)
TOP + PP
Alk. Pers. Dig.
  EPA 365.4 (AA)
 EPA 365.1 (AA)
 High Temp Oxid.
   Dig., EPA 160.2-1
Spectrophot.
Spectrophot.

-------
               417
                FIELD
           VARIABLES AND
        ANALYTICAL METHODS
      Variable

Temperature
Depth
Dissolved Oxygen

Conductivity
PH
Seech i Depth

Salinity
Analytical Method

 CTD
 CTD
  DO Sensor
   w/ Winkler Cal.
  CTD
  pH Probe
  20 cm disk
   after Welch
  Calculated from
    Conductivity

-------
I
Figure 6. UiesapeaKe Bay coorumaica apiii Dumpie n ug
                                                                                                                               RavfetonNo.3
                                                                                                                               3/19/90
                                                                                                                               Page 19 of 2i
              Key:

               VWCB - Field/Program Agency
               DCLS - Analytical Laboratory
               t    | - Component lead agency
                                            SRBC
                                            PADER
PADER
PAOER
                                                               Tributaries/Fall-Lint
                                                                  Component
                                                                 (•latton varlisl
                                                                               OWML
                                                                               OWML
                                                       USGS Towaon
                                                       USGS Denver
                                              OCDCRA
                                              EPACRL
                        MDE
                        MDDHMH
                                              Potomac
                                                River
                                             Component
                                              (PMS-101
                                                            _«	_J_i._\--
                                                                                          Malnstem/
                                                                                            Tidal
                                                                                          Tributaries
                                                                                         Component
                                                                                           (CB5.31
                             Virginia
                            Malnstem/
                            Tributaries
                            Component
                             (TFS.S)
                                                                                                                                          CO

-------
                                               419

                        Figure 4. Schematic of the Operational Flow of Analyses,
                                 Coordinated Split Sample Program
       LARGE
       VESSEL

       (See Figure 1
       for dispensing
       order)
                Normal Laboratory
                  Quality Control
                    Procedures
Replicate Analyses**
        Analyze for
     Routine Parameters
                                                            Split Sample
                                                            Revision No.
                                                            3/21/90
                                                            Page 12 of 26
                                               Triplicate Aliquots
                                             (sent to each laboratory)*
                                                    *__••
                       Analyze for
                   Routine Parameters
                           Analyze for
                       Routine Parameters
                                                                *(in-matrix estimate
                                                                  of field precision)
                                                                  Spike Sample
                                                                              ***
   Analyze for
Routine Parameters
               ** (in-matrix estimate
                  of lab precision)
     Analyze for
  Percent Recovery

***(in-matrix estimate
      of accuracy)
                  Deionizcd/distillcd
                    water dilution
                                              EPA Standard
                                            Reference Material
                                                  Matrix water
                                                     dilution
                                                  (saline only)
                     Analyze for
                   SRM Parameter
                                                   Analyze for
                                                 SRM Parameter

-------
            Splii
                  sample results at four labs, TDP
                   oif tfc?fa aliquot* for Total Daaaolvad PhotphoMli, aw/l
                                aa*pn* iton« quwttrly at station CM.3
0,07
0
z
A
N 0.0*
0.08
0.01
0.00
     B
        	
                ««*'
                                                                              to
                                                                              o
  Aw^^^V'
                                        WMMW


                                 SAMM.XM9 OATt
                                                         JULtt
NOVM
                    Lcttart •hew analysis laboratory (A,  B. C, 0)

-------
               CO *0  *i
                                                       Mont
oeau
MAON
trwr
MNVT
  i
ttwnr
ZtOlO    iMIW
  1*1
            >00'0
                                                                                      10'0
                                                                                      80*o a
                                                                                           t
                                                                                           i
                                                                                      CO'O \

                                                                                           A
                                                                                      >o*o J
                                                                                           0

                                                                                           0
                                                                                      SO'O i
                                                                                           s

                                                                                      •O'OJ

-------
utra
                                          'nan
     SWWP WdiS ttNOT MUVN £1310 £t9W £tAVN itUJ 91100 WinT 9tW
       1.1,1.1,1  .	I   tl.j.l.l.lr
                                                  8idB6
                                                 >   i  ,
CN
CN
                                                                OOO'O


                                                                WO'O

                                                                ovo-o

                                                                9IO'0

                                                                OfO'O

                                                                WO
                                                                                          a
                                                                                       '0 J
                   ddL 'V

-------
                        423
COMPONENTS OF THE CHESAPEAKE BAY
           MONITORING PROGRAM
    Mainstem Chesapeake Bay Water Quality Monitoring Program

    Maryland, Virginia and District of Columbia Tributary Water Quality
    Monitoring Programs

    Citizen Monitoring Program

    Fall Line (River Input) Monitoring Programs

    Baywide Phytoplankton and Zooplankton Monitoring Programs

    Baywide Benthic Monitoring Programs

    Fisheries Independent Shellfish Monitoring Programs

    Fisheries Independent Beach Seine and Trawl Survey Programs

-------
                       424
COMPONENTS OF THE CHESAPEAKE BAY
     MONITORING PROGRAM  (cont'd)
   Submerged Aquatic Vegetation Aerial and Ground Survey Program

   Sediment Toxicant Monitoring Programs

   Finfish and Shellfish Tissue Monitoring Programs

   Baywide Waterfowl Survey Programs
   Sediment Processes Monitoring Programs

-------
                   425
  OTHER CHESAPEAKE BAY BASIN
      MONITORING PROGRAMS
State Ambient Water Quality Monitoring Networks

State Shellfish Bacteriological Monitoring Programs

Utilities Monitoring Programs

USGS NASQUAN Program
USGS Streamflow Gauging Network
State Groundwater Observation Network
Nonpoint Source Watershed Monitoring Programs

-------
                   426
 OTHER CHESAPEAKE BAY BASIN
 MONITORING PROGRAMS (cont'd)
Point Source Compliance Monitoring


Radiological Monitoring Programs


NOAA NWS Meteorological Data Networks


State Air Quality Monitoring Networks


State Acid Depostion Networks
NOAA Tidal Height Network

-------
                     427
CHESAPEAKE BAY LIVING RESOURCE
MONITORING PROGRAM OBJECTIVES
 Document the current status of living resources and their habitats in
 Chesapeake Bay
 Track the abundance and distribution of living resources and the
 quality of their habitats over time
 Examine correlations and relationships between water quality, habitat
 quality and the abundance, distribution and integrity of living resource
 populations

-------
THE CHESAPEAKE BAY
MONITORING PROGRAM
   An Evolving Data Collection Network
   and Process Providing Information
   Necessary for Management of the
     Chesapeake Bay's Resources
                                    to
                                    00

-------
                            429
             MONITORING PROGRAM
         DESIGN AND IMPLEMENTATION
Chesapeake Bay Monitoring Program Experience:
  •  Institutionalized the program through an effective committee/technical
     workgroup structure
  •  Planned adequately for water quality network design but not for living
     resource components; monitoring of toxics still not fully addressed
Recommendations:
     Establish a multi-jurisdictional monitoring committee
     Clearly state the program objectives using them in developing Data
     Quality Objectives and the network design
     Continually seek long-term,  stable funding sources
     Integrate existing monitoring programs into the design of a
     coordinated monitoring program
     Consider future modeling needs during network design

-------
                               430
                 DATA MANAGEMENT
Chesapeake Bay Monitoring Program Experience:
   •   Did not adequately plan up front for data management needs
   *   Centralized computer data base networked with other data bases
   •   Working with numerous data submitting organizations demanded
      specific data submission formats and data management
      requirements
Recommendations:
  •   Plan adequate resources for data management prior to monitoring
      program implementation
  •   Seek consensus on and require adherence to specific data
      submission requirements
  •   Clearly state objectives for data base development up front and
      adhere to them when structuring the data base
  •   Target acquisition of key historical data sets early on
  •   Establish procedures for quality assurance of all data entered onto a
      common data base

-------
                              431
                QUALITY ASSURANCE
Chesapeake Bay Monitoring Program Experience:

   •   More than 15 laboratories analyzing water quality samples alone
      eventually contributing to the centralized computer data base

   •   Significant effort required to ensure use of comparable sample
      collection and analysis methods across component programs

   •   Routine analysis and interpretation of QA data critical to future quality
      of data
Recommendations:

  •   Establish quality assurance as an integral part of all monitoring
      program components (QAOs, audits, QA plans)

  •   Set up a coordinated split sample program between analytical
      laboratories

  •   Seek technical consensus on sample collection and analysis
      procedures in view of program objectives and implementation of
      resultant consensus

-------
                               432
   DATA ANALYSIS AND INTERPRETATION
Chesapeake Bay Monitoring Program Experience:

   •   Insufficient resources were devoted to data analysis

   *   Direct connections between information resulting from the program
      and management decisions were limited until adoption of the
      Baywide Nutrient Reduction Strategy and setting of Baywide
      restoration goals

   •   Establishment of consensus on data analysis priorities up front and
      sharing of data management and data analysis resources between
      agencies was necessary
Recommendations:

  •   Dedicate resources for analysis and interpretation of monitoring data

  •   Establish a tiered reprting system to force routine analysis and
      synthesis of data targeted toward various levels of agency managers
      and the public

  *   Create a dependence on using results from the monitoring program
      for management decision-making

-------
                  433
   ANALYTICAL METHODOLOGY

Chesapeake Bay Program Monitoring Experience:
* Adopted EPA Water and Wastewater Analytical
   Methods
* Hostage to continuity with the "historical database"
* CBP special studies to establish comparability
* Some data usability compromised
Recommendations:
 * Develop performance data for estuarine
   and marine analytical methods through
   round robin and validation studies
 * Evaluate and develop SRMs
 * Develop a network among technical
    estuarine/marine professionals

-------
     ESTUARINE
        AND
      MARINE
                          OJ
ANALYTICAL METHODS
     COMMITTEE

-------
WORKGROUPS
en

-------

ORGANICS
NUTRIENTS
 METALS
BIOLOGICS
             CO
             O1

-------
                             437



                                   MR. FIELDING:  Good



morning.



     I have one announcement to make.   EMSL-Cincinnati is



soliciting labs for two interlaboratory method validation



studies.  The first is a joint EPA/AOAC effort to validate



NFS Method 6 for ethylene thiourea in water by GC/NPD.  The



second is a joint ASTM/EPA study announced in the Federal



Register for Cr6 ion chromatography.



     If anyone is interested in participating in either of



these two method studies, information will be provided at



the table in the rear of the hall.



     I hoped everybody survived last evening's excursion on



the HMS Sinkfast.  If you didn't, let us know.



     We have several interesting papers today.  We would



like to start off with a paper by Yves Tondeur of Triangle



Laboratories who will discuss the determination of semi-



volatile organic compounds in river water at the ppq level



by high resolution GC and high res mass spec.



     Yves?

-------
                             438
                              MR. TONDEUR:  Good morning.
     Sometime last fall, the Canadian Government, which is
engaged in a study of pollutants in the Niagara River,
established a list of some 23 organic chemicals for which
methodologies capable of achieving parts per quadrillion
detection limits were necessary, and the U.S. Environmental
Protection Agency, in the spirit Of international
cooperation, has offered to help and subcontracted the work
to our lab.
     The objectives (Fig. 1) of the study were to develop an
analytical procedure that would be able to detect and
quantify 21 target compounds in river water sample extracts
at the parts per quadrillion level.  Originally, the list
was for 23 compounds, and we had to reduce it to 21,
primarily because one of the compounds, the trichlorotoluene
could not be obtained as a standard at the time of the
study, and the other one was a series of chemicals called
toxaphene for which we didn't think we could develop a
procedure at the parts per quadrillion level within the
required 21-day turnaround.
     So, during that short period of time, we were involved
in the evaluation of short and long-term reproducibilities
of the response factors for the various analytes relative to
internal standards and also somehow evaluate the procedure

-------
                             439



on simulated matrix spike or laboratory control spike



extracts.



     The method itself calls for isotope dilution mass



spectrometry during which a group of 16, labeled standards,



mostly deuterium labeled, were used.  The mass spectrometer



uses a detector which is operating in the selected ion



monitoring mode and with a resolving power of 10,000 for



selectivity.



     The samples are introduced inside the system through a



capillary GC column.  The samples were supposed to be



obtained from the extraction of a 40-liter sample size and



analyzed following a concentration step down to 25



microliters.  In other words, a million-fold concentration,



and that was the requirement set for us.



     The target analytes (Fig. 2) represent a broad range of



compounds ranging from compounds such as acenaphtene,



benzidines, a group of halogenated aliphatics and aromatics



such as chloroethylether or bromophenylphenylether and also



chlorinated benzene and brominated benzene.  We have also a



series of nitrosamines (n-nitrosodimethyl, dipropyl, and



diphenylamines), DEHP and di-n-octylphthalate, along with



some phenolic compounds such as tetrachlorophenol and



dinitrophenol.



     So, the very first thing we did was to prepare a set of



calibration solutions whose composition is summarized on

-------
                             440



Figure 3.  Two types of compounds:  1)  A series of 21



unlabeled analytes representing the target compounds, and



2) A group of 16 labeled internal standards.



     The analyte concentrations along these five calibration



solutions vary from 25 picograms per microliter in



concentration all the way to 300 picograms per microliter,



while the concentrations of the internal standards remain



constant at either 100 picograms per microliter or 200



picograms per microliters, depending on the compound.



     Now/ if we consider a sample size of 40 liters and a



final extract volume (if achievable) of 25 microliters, then



the concentration domain represented by this set of



calibration solutions covers approximately the 15 ppq level



up to about 200 ppq level, which was part of the requirement



of the study.



     Since I am going to be showing quite a few tables this



morning, I will just show a few selected current ion



profiles obtained on as little as 5 picograms of selected



compounds such as the chloroethylether for which a 5-



picogram injection gives you a very reasonable and



respectable signal-to-noise ratio (Fig. 4).  The same can be



said for the chlorinated and brominated benzene ( 5-



picograms of material injected)  (Fig. 5).



     Acenaphthene and its internal standard, DiO-



acenaphthene (Fig. 4). eluting slightly before  or  the

-------
                             441
chlorophenylphenylether with the M and M+2 responses/ the
M+2 becoming kind of a noisy thing at the 5 picogram level.
     Another final example is the one obtained on
tetrabromobenzene.  Figure 6 represents 5-picograms of
material injected on the GC column for which a noisy signal
is obtained.
     I am not saying here that all the target compounds
could be analyzed at the 5 picogram injection level.  These
examples are the only ones for which a reasonable signal
could be obtained.  They reveal that the methodology itself
could very well exceed the required detection limits
requested by the Canadians.
     So, an aliquot from each of these calibration solutions
were then injected once and the response factors of the
various analytes calculated for each of the calibration
points.  The means of those five points were calculated
along with the relative standard deviations of the means.
(Fig. 7).
     As you can see from this summary table, with the
exception of acenaphthene, the standard deviation remained
below the 20 percent mark, and most of the time, it is even
less than 10 percent, which is quite remarkable considering
t the levels at which we are working.  The results are in
agreement with expectations from the use of isotope dilution
techniques.

-------
                             442
     There are a few analytes for which we were not able to
measure a response factor.  Benzidine and dinitrophenol:
Those two compounds were never seen at those levels in the
chromatogram.  We suspect that the benzidine, being a basic
compound, and nitrophenol, an acidic compound, reacted
together and, therefore, make their detection very
difficult.
     We also believe that at those very low concentrations,
picogram per microliter range, that the reaction is much
more noticeable.  We are able to see benzidine and
nitrophenol simultaneously in the solution when their
concentration in the solution exceeded the nanogram per
microliter level.
     Other compounds such as the DEHP for which we were not
able to generate a response factor because of a problem with
the mass spec operator who did not cover the right retention
time window.  This was, in fact, corrected afterwards during
the matrix spike studies.
     Nitrosodiphenylamine constituting another exception.
Since we use a splitless injection technique, we observed
almost quantitative conversion of nitrosodiphenylamine into
diphenylamine.
     As far as the relative retention time for these
analytes, the mean relative retention times are summarized
in Figure 8. They were determined using the internal

-------
                             443



standards that were used during the computation of the



response factors.



     As you can see, over the 5-point calibration curve,



these standard deviations are extremely small which makes it



evident that isotope dilution will provide pretty reliable



results as far as qualitative characterization of the



analytes in the samples.



     We then looked into the continuing calibration (con-



cal), that is, after a 48-hour and 4-day periods following



the initial calibration.  Figure 9 contains the summary



results obtained after 48 hours.



     The middle point of the calibration curve was used to



determine the relative response factors for each of the



analytes, we then compared those response factors to the



mean that were obtained during the initial calibration (I-



cal), and then calculated the percent difference between the



con-cal and the I-cal response factors.



     As you can see, for most of the analytes for which we



could measure a response factor, the percent difference



remained within 20 percent of the I-cal, which is quite



remarkable.



     Then we also looked at the response factors obtained



for these various analytes four days following the initial



calibration (Fig.  9).  Four days may not be a lot, but when

-------
                             444
you have a 21-day period to develop and evaluate a method, 4
days is a significant amount of time.
     So, again, the daily response factors calculated here
from the middle point of the calibration curve were compared
to the I-cal response factors and then the deviation from
the I-cal summarized in the last column.  And with the
exception of acenaphthene, dichlorobenzidine and
diphenylamine, the deviation remained within 20 percent of
the I-cal.
     Even though some of these analytes exceeded 30 percent
deviation, we believe that the initial calibration can be
used for routine analysis at least after 4 days after
establishment.
     We then evaluated the overall procedure of high-
resolution GC and high resolution mass spectrometry on what
we called matrix spike analysis (a laboratory control
spike).  The sample is supposed to be obtained from the
extraction of 40 liters of water, then concentrated down to
1 ml methylene chloride and sent out to the lab in vials of
1 ml.  The lab is then required to add the internal
standards to that 1 ml solution before blowing it down to 25
microliters.
     So, what we have done since we were not able to obtain
actual river samples, we simulated a matrix spike by taking
1 ml of dichloromethane and spiking known quantities of the

-------
                             445
unlabeled compounds shown here along with the internal
standards.  The spike level is expressed here in parts per
quadrillion and varies from some 30 ppq to 63 ppq/ and for a
couple of analytes, the concentration of the spike is 125
ppq (Fig- ii)- •
     The second column shows the result obtained from the
triplicate set of matrix spike studies.  These are the mean
and the relative standard deviation of the mean.
     The last column, which is called here percent accuracy,
is in fact is the recovery (amount found in the sample
relative to the amount spiked in the sample), and offers an
idea on how accurate the procedure is.
     There are a few compounds, acenaphtene and the
phthalates along with tetrabromobenzene, tribromobenzene and
diphenylamine for which these percent accuracies were very
high, exceeding what we would call an acceptable limit of
150 percent.  For the remaining compounds, the percent
accuracies were close to the 100 range.
     With the exception of the phylates for which an obvious
explanation can be offered simply because of background
contamination, we cannot come with any reasonable
explanation for the acenaphthene observation.  For the
others, one can always assume that tetrabromobenzene and
tribromobenzene,  are not actually measured using isotope

-------
                             446
dilutiont need their measurement are more susceptible to
instabilities during the GC/MS run.
     In conclusion, the procedure developed here of high-
resolution GC and high-resolution mass spectrometry could
definitely offer parts per quadrillion detection limits on
most of the target compounds, that is, provided that 40
liters of extract can be blown down to 25 microliters
without any form of a cleanup.
     Some of the limitations we encountered were primarily
associated with chemical reactivity between some of the
components but also background contamination.
     As far as the precision and accuracy of the method (as
measured on matrix spike studies for 18 analytes), the
precision, was in the neighborhood of 15 percent (mean),
ranging from 3.5 to 50 percent, which is kind of remarkable
considering the low levels at which we were working.
     Finally, the accuracy gained here from matrix spike
studies gave us a mean of 157 percent, ranging from 86 to
374.  Of course, that mean is biased high because of a
couple of extremely high data points originating from
tetrabromobenzene and the phlalates.
     Finally, we recommend strongly that:  1) Separate
analyses be done for the acidic compounds.   It is not a very
good idea to mix the acid and the base neutral fractions of
the extraction;   2).  The actual testing of river water

-------
                             447



sample be evaluated directly so that one can really



determine the applicability of the method at the parts per



quadrillion level; and, 3) finally, as we go to lower and



lower levels, special precautions will have to be taken to



minimize or control the contribution from background such as



phlalates.  Otherwise, the method will not be applicable to



analytes for which we have some background contribution.



     As far as the determination of nitrosodiphenylamine and



diphenylamine goes, they will have to be analyzed separately



at this time perhaps by using other GC conditions (we found



that nitrosodiphenylamine was degrading into diphenylamine



using cold on-column injection.



     Thank you for your attention, and if you have any



questions, I would be glad to answer them.

-------
questions?
                             448
                   QUESTION AND ANSWER SESSION
                         MR. FIELDING:  Are there any
                         MR. STANKO:  George Stanko, Shell
Development Company.
     I would like to point out that one of the definitions
of practical quantitation limit is that 80 percent of the
labs can get within plus or minus 40 percent of the true
value.  I don't think you got there.
     Thank you.
                         MR. FIELDING:  Any other questions?
(No response.)
                         MR. FIELDING:  Thank you, Yves.

-------
                                 449
DETERMINATION OF  SEMI-VOLATILE  ORGANIC



   COMPOUNDS  IN  RIVER  \VATER  AT  THE



   PART-PER-QUADRILLION  (PPQ>  LEVEL BY



  HIGH-RESOLUTION GAS  CHROMATOGRAPHY



                             /



   HIGH-RESOLUTION  MASS  SPECTROMETRY
    Yves  Tonclexar. Mick.  Cbxi  & Don

-------
r
                                        450
                                 OBJECTIVES
           1.   Develop  Methodology  CZ1  target  compounds,



               river  -water sample  extract,  ppq  level)
                Demonstrate
                                an
I-Cal  with. Reproducible
               Response  Factors
                Demonstrate Con-Cal  \Arith   Reproducible



                 Response  Factors






                Demonstrate Matrix Spilte  Recoveries  after



                 Spils-ing  and  Concentration Steps

-------
                                  451
                        (METHOD!
    Isotope  -  Dilution   Mass    Spectrometry

Fifteen   Deuterated    or    Carh>on-13   Labeled

        Selected    Ion    Monitoring    M S

                 Electron.   lonization

                 Positive   Ion.   Nlode

          Resolving   Power    of   1O.OOO
                   for   Selectivity

       Capillary GC Column    C6O-m  DB-S)
    to   :In.sure   Adequate    GC   Resolution
   Extraction   to
be   from.   4O-L   River
 "Water
Concentration of the  Final  Extract  to  25 uL

  Greater  than   a IVIillion-Fold Concentration

-------
                             452
        ANALYTES   of   INTEREST
Acenaphthene



Benzidine



Bis(2-chloroethyl)ether



Bis(2-ethylhexyl)phthalate



4-Bromophenylphenylether




4-Chlorophenylphenylether




3,3'-Dichlorobenzidine



Di-n-Octylphthalate



N-Nitrosodimethylamine



N-Nitrosodipropylamine



N-Nitrosodiphenylamine




Diphenylamine



1,2-Diphenylhydrazine  (or decomposition product azobenzene)




1,2,3,4-Tetrachlorobenzene



1,2,3,5-Tetrachlorobenzene



1,2,4,5-Tetrachlorobenzene



2,4,5-Trichlorotoluene



2,3,4,5-Tetrachlorophenol




2,4-Dinitrophenol



1,3-Dibromobenzene




1,3,5-Tribromobenzene



1,2,4,5-Tetrabromobenzene

-------
                            453
INITIAL    CALIBRATION   SOLUTIONS
    Compound
              Solution Number
 Concentration (pg/uL)s
 12345
Unlabeled Analytes

Acenaphthene
Benzidine
Bis (2-chloroethyl)ether
Bis(2-ethylhexyl)phthalate
4-Bromophenylphenylether
4-Chlorophenylphenylether
3,3*-Dichlorobenzidine
Di-n-Octylphthalate
N-Nitrosodimethylamine
N-Nitrosodipropylamine
N-Nitrosodiphenylamine
Diphenylamine
1, 2-Diphenylhydrazine
1,2,3, 4-Tetrachlorobenzene
1,2,3,5-Tetrachlorobenzene
1,2,4,5-Tetrachlorobenzene
2 , 4 , 5-Trichlorotoluene
2,3,4,5-Tetrachlorophenol
2, 4-Dinitrophenol
1, 3-Dibromobenzene
1,3, 5-Tribromobenzene
1,2,4, 5-Tetrabromobenzen.e

Internal Standards
 !0
 8-
 8-
 4-
 5-
 5-
 6-
 4-
d,
1-
-Acenaphthene
Benzidine
Bis (2-chloroethyl)ether
Bis(2-ethylhexyl)phthalate
4-Bromophenylphenylether
4-Chlorophenylphenylether
3,3 ' -Dichlorobenzidine
Di-n-Octylphthalate
N-Nitrosodimethylamine
-N-Nitrosodipropylamine
N-Nitrosodiphenylamine
-Diphenylamine
-1, 2-Diphenylhydrazine
6-l,2, 4,5-Tetrachlorobenzene
2, 4-Dinitrophenol
6-1, 4-Dibromobenzene
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
100
200
100
100
100
100
200
100
100
200
100
100
100
100
200
100
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
100
200
100
100
100
100
200
100
100
200
100
100
100
100
200
TOO
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
       100
      100
      200
      100
      100
      100
      100
      200
      100
      100
      200
      100
      100
      100
      100
      200
      100
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                   200
                  100
                  200
                  100
                  100
                  100
                  100
                  200
                  100
                  100
                  200
                  100
                  100
                  100
                  100
                  200
                  100
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                        300
                                                            100
                                                            200
                                                            100
                                                            100
                                                            100
                                                            100
                                                            200
                                                            100
                                                            100
                                                            200
                                                            100
                                                            100
                                                            100
                                                            100
                                                            200
                                                            100
(a) Based  on  a 40-L sample size,  25  uL final extract volume,  25
pg/ul represents 15.6 ppq.  (Calibration Range:  15.6 ppq  to  187.5
ppq->

-------
ro  a*, en co  *r;
co  co co co  °~
i    i   t   i    P**
                     ro  oj
                     X
                         CO — -(
                            ri
 ro  tn
 53. rn
 4 I'  u
   5'  I
    CO
 =1  CO
 o   ••

•a  cT

 en pt-
    cu
   U3
 cu
 tn
 in
 co m
 co «=>

 ro co
             en

             co
                                     CD
                                     CO
                                     ro
                                     CO
co


co
             •vj

             ca
             co
                                     co
                                     co
                                     co

                                     ro
                                     CO


                                     -O-
                                     co
                                      CO
                                      co
                    ro  *ix co oo ~_*
                    co  co co co S2
                    lit!   f"
                                         —i  to
                                         ro  cu
I
O
              o
              H
              O
              t)

              (D
              a
              a1
              ID
              a
                                                   a1
                                                   re
                                                                     =J  CO
                                                                     n>  to
                                                                     o  m
                                                                     o  co
                                                                     =3  CO
                                                                        to
                                                                     cnr;-


                                                                     c?  «
                                                                        cu
                                                                       10
                                                                        ro
                                                                     SK
                                                                     OJ
                                                                     in  vj
                                                                     in  co
                                                                     j*. co
                                                                        CO
                                                                     co m
                                                                     cu «o
                                                                     u>»
                                                                     •— • • to
                                                                       1C
                                                                        ui
                                                           o
                                                           —t
                                                           to
                                                           cs>


en
-n.
CO

CO
CD
CO

CT3
ro
CO

t— *
CO
-o.
CO
*—*
CD
CO


j
»O
ro
CO


^
•vj
.0»
CO
CO
CO
CO
CO
ro
CO

CO
_r>-
co
CO
CO
CO CD
i i t t i^
—t to — < *—
ro o> i — en_
2-fL 5? £
ro
CO
5" CO CD
ro' to
— -__ a s-eg _
r^^ o s'co ??.
, » to ro
' t co
O
u CO ^TT
pj ~7 ~3 " *
•o ° ^- £
a1 -o o
£ ^n^
a1 cu «—
re ua vj
» =."•«•
(0 CU • CD
in ^J
in CD
i
•—* ro _
co en -^j
-*^ CO .. -
co ro
•C^ lft_J ^"^
^^
•— • to
ur . 	
T. ^
S s"
CO
CD
CD
CO
sc ro'
o co
3
CO
CD
co to
to ^
CO
ga -».


•&_
re

fc—> >— *
=] CO
ro" to
o m
f-*- ~O
r-' t
^- _^_ O CO
.^ ~~
n to
re CTJ r~*
M ~l ~>
a o ••
m c •*=.
•d "° °.
rt O^
a1 n>
re =>K
M cu
-" in vj
re in CD
t
»— ' ro
en en
J>. co
> to
co m
XJ K3
CO
OJ tO

*T.
1 	
>

•x.
p
at


J3^
«JL>
cn

                                                                                                                                       Cn

-------
  BE.
  28.
  e
                                              455
                                        Bis (2-chloroethyl) ether
                                                                                451
           i;4B      2=38      3=28       4=18       5=68
    TLIB       1B-SEP-B9     Sir=Voltage  7B-25BSEC Sys=  TLI
    Sample 1    Injection  1    Group 1   Rass 95.J
    Text--
                                                          5=58      6=48      7=38
1BL
88.
68.
48.
28.
 8
                                          Bis { 2-chloroethyl) ether
          1;48       2=38       3=28       4=16      5=68      5=58       6=46       7=38
   TLIB      1B-SEP-B3     Sir 'Voltage  7B-25BSEQ Sys= TLI
   Sanple  1   Injection 1    Group 4   Rass 215.
   Text=
                                       Tetrachlorobenzene*
46.
28.
B
                                      II!
                                                                 Nor*=
                                                                               858
 12=28    12=48    13=88    13=28    13=46    14=BB    14=28    14=48    15=88    15=28
TLIB      1B-SEP-B9     SirVoltage  7B-25BSEC Sgs=  TLI
Sanple 1   Injection  1    Group 3   Rass 235.BB53
Text'
                                   1,3-Dibromobenzene
                                                               Kora=
                                                                           33B
    3=48      1B--6B      18=28     18^48     11=88      11=28     11=48      12=88

-------
                                     456
    TLI6       JB-SEP-BS     Sir Voltage  7B-25BSEO Sys=  TU
    Sanple 1   Injection  1    Group 6   Rass 391.BB6S
    Text'
IBB.

88.
 EBJ
                   1,2,4,5-Tetrabromobenzene

                                                            13
   19*28    2B-BB    28=4B    21=28    22=88    22=48    23=28
   TUB      1B-SEP-B9     Slr-'Voltage  7B-25BSEO Sys= TU
   Sample 1   Injection 1    Group E   Kass-393.6B4S
   Texts.
                                           24=BB    24=46     25=2B
IBB.
                                              Hor«=
                     1,2,A,5-Tetrabromobenzene
13=28    2B'{
2B=4B     21=28    22'
                                             22--4B    23=28    24=88   24=4B    25=28

-------
                           457
INITIAL   CALIBRATION   RESULTS I

Analyte
Acenaphthene e
Benzidine
Bis (2-chloroethyl ) ether
Bis (2-ethylhexyl) phthalate
4-Bromopheny Ipheny lether
4-Chlorophenylphenylether e
3,3* -Dichlorobenzidine
Di-n-Octylphthalate e
N-Nitrosodimethylamine
N-Nitrosodipropylamine
N-Nitrosodiphenylamine
Diphenylamine . **e
1 , 2-Diphenylhydrazine e
1,2,3, 4-Tetrachlorobenzene
1,2,3, 5-Tetrachlorobenzene
1,2,4, 5-Tetrachlorobenzene
2,3,4, 5-Tetrachlorophenol
2, 4-Dinitrophenol
1 , 3-Dibromobenzene
1,3, 5-Tribromobenzene
1,2,4, 5-Tetrabromobenzene
Mean RRF
1.69
r j a
0.84
b
0.70
1.31
1.44
0.89
0.79
0.85
c
1.04
1.28
1.04
0.91
1.15
0.04
rja
0.36
0.11
0.10
RRF
Z RSD
25
_ _
4

8
6
12
5
4
6

12
18
5
5
6
1
--
2
0.3
0.8
(a)  rj -  rejected  data  point  (fails  to  meet  acceptable initial
     calibration  criteria).

(b)  The  signal  corresponding  to DEHP was  not detected  due to
     improper  selection  of  retention  time window.

(c)  N-Nitrosodiphenylamine  is  quantitatively  converted  into
     diphenylamine  inside the GC  injector and on the GC column.

(d)   Contains  a contribution  from N-Nitrosodiphenylamine.

(e)   Corrected for  isotope  impurities  using equations 6-9.

-------
                                    458
 INITIAL
CALIBRATION   RESULTS!
Analyte
Acenaphthene
Benzidine
Bis(2-chloroethyl)ether
Bis(2-ethylhexyl) phthalate
4-Bromophenylphenylether
4-Chlorophenylphenylether
3,3' -Dichlorobcnzidine
Di-n-Octylphthalate
N-Nitrosodimethylamine
N-Nitrosodipropylamine
N-Nitrosodiphenylamine
Diphenylamine
1,2-Diphenylhydrazine
1,2,3,4-Tetrachlorobenzene
1,2,3,5-Tetrachlorobenzene
1,2,4,S-Tetrachlorobenzene
2,3,4,5-Tetrachlorophenol
2,4-Dinitrophenol
1,3-Dibromobenzene
1,3,5-Tribromobenzene
1,2,4,5-Tetrabromobenzene
;an RRT
1.005
1.013
1.002
1.002
1.001
1.001
1.012
1.015
1.002
1.003
1.054
0.998
1.001
1.228
0.832
1.101
1.034
RRT
Z RSD
0.00
0.09
0.03
0.00
0.04
0.00
0.55
0.11
0.00
0.00
0.00
0.00
0.00
0.07
0 .01
0.05
0.01

-------
CONTINUING
                            459
   CALIBRATION
RESULTS
(48 hours  following  the  initial  calibration)

Analy te

Acenaphthene *
Benzidine
Bis{2-chloroethyl) ether
Bis(2-ethylhexyl)phthalate
4-Bromophenylpheny lether
4-Chloropheny Ipheny lether *
3,3'-Dichlorobenzidine
Di-n-Octylphthalate ^
N-Nit rosodime thy lamine
N-Nit rosodipropy lamine
N-Nit rosodipheny lamine
Diphenylamine *"
1 , 2-Diphenylhydrazine *
1,2,3,4-Tetrachlorobz
1 , 2 , 3 , 5-Tetrachlorobz
1,2,4,5-Tetrachlorobz
2,3,4,5-Tetrachlorophenol
2 , 4-Dinitrophenol
1 , 3-Dibromobenzene
1,3, 5-Tribromobenzene
1,2,4, 5 - Tetrab romo benzene

RRF
i-cal
1.69

0.84
b
0.70
1.31
1.43
0.89
0.79
0.85

1.04
1.28
1.04
0.91
1.15
0.04

0.36
0.11
0.10e

RRF
con-cal
1.15
a
0.87
5.08
0.73
1.49
1.60
0.77
0.83
0.77
c
1.12
1.43
1.04
1.02
0.99
0.05
a
0.43
0.12
0.11

Percent
Difference
-21.4

3.5
--
4.6
14.3
11.5
-13.5
4.4
-9.6

7.1
11.7
0.1
11.5
-14.3
20.4

20.9
3.7
6.2
(a)  Benzidine  and  2,4-dinitrophenol  are not  detected  at lov
    levels.
(b)  No  response
    calibration.
factor  is  available  for DEHP  in the  initial
(c)  Thermal  decomposition.

(d)    Contribution    from
    nitrosodiphenylamine.
              the    decomposition   of
                     N-
(e)  Ion-abundance  ratio was  outside  the  expected  QC  limit.

(f)   Corrected  for isotope  impurities  using  equations  6-9.

-------
                                   460

I CONTINUING    CALIBRATION    RESULTS"
 (4 days following the initial calibration)

Analyte
Acenaphthene *
Benzidine
Bis (2-chloroethyl) ether
Bis(2-ethylhexyl)phthalate
4-Bromophenylphenylether
4-Chlorophenylphenylether ^
3,3' -Dichlorobenzidine
Di-n-Octylphthalate f
H-Hitro so dime thy lamine
N-Nitrosodipropy lamine
N-Nitrosodipheny lamine
Diphenylamine df
1,2-Diphenylhydrazine ^
1,2,3, 4-Tetrachlorobz
1,2, 3,5-Xetrachlorobz
1,2,4, 5-Tetrachlorobz
2,3,4, 5-Tetrachlorophenol
2, 4-Dinitrophenol
1,3-Dibromobenzene
1,3, 5-Tribromobenzene
1,2,4, 5-Tetrabromobenzene
RRF
i-cal
1.69

0.84
b
0.70
1.31
1.43
0.89
0.79
0.85

1.04
1.28
1.04
0.91
1.15
0.04

0.36
0.11
0.10e
RRF
con-cal
0.93
a
0.92
4.60
0.77
1.49
1.90
0.94
0.76
0.79
c
1.40
1.11
1.04
1.03
1.12
0.05
' a
0.35
0.11
0.09
Percent
Difference
-35.9

9.0
— _
10.6
14.2
32.7
-4.8
-4.2
-6.5

33.9
-13.0
-0.4
12.4
-2.9
20.7

-2.1
0.0
-6.2
(a) Benzidine  and  2,4-dinitrophenol  are not  detected  at low
    levels.

(b) No  response  factor  is  available  for DEHP in  the  initial
    calibration.
(c)  Thermal decomposition.

(d)    Contribution    from
    nitrosodiphenylamine.
the
decompos ition
of
N-
(e)  Ion-abundance  ratio was  outside the expected QC limit.

(f)   Corrected  for isotope impurities using equations 6-9.

-------
                     461
MATRIX    SPIKE
ANALYSES

Analyte
Acenaphthene
Benzidine
Bis (2-chloroethyl)ether
Bis (2-ethylhexyl)phthalate
4-Bromophenylphenylether
4-Chlorophenylphenylether
3,3* -Dichlorobenzidine
Di-n-Octylphthalate
N-Nitrosodimethylamine
N-Nitrosodipropy'lamine
N-Nitrosodiphenylamine
Diphenylamine
1 , 2-Diphenylhydrazine
1,2,3, 4-Tetrachlorobenzene
1,2,3, 5-Tetrachlorobenzene
1,2,4, 5-Tetrachlorobenzene
2,3,4, 5-Tetrachlorophenol
2, 4-Dinitrophenol-
1 , 3-Dibromobenzene
1,3, 5-Tribromobenzene
1,2,4, 5-Tetrabromobenzene
Spike
Level
(_ppq)
31.
•
31.
31.
62.
31.
125
31.
62.
62.
.
62.
31.
31.
31.
31.
125
-
31.
31.
31.
3

3
3
5
3

3
5
5

5
3
3
3
3


3
3
3
Found
Cone .
(ppq)
63.
»
31.
113.
81.
80.
108
67.
66.
58.
.
352
29
40
36
34
293
-
31
53
117
4

4
7
8
2

1
9
9


. 0
.2
.5
.9


.0
.0

RSD
2
23
.
3
_
13
13
7
17
16
10
_
40
9
7
3
4
14
.
7
3
51
Accuracy
X
.1

.4

.8
.4
.8
.2
.2
.1

.8
.9
.6
.9
.8
.1

.5
.6
.1
203
»
100
364
131
257
86.
215
107
94.
-
563
92.
129
117
112
235
-
99.
169
374






2


3


7





3



-------
                          462
                     I CONCLUSIONS'!
1.    PPQ   Detection  Limits   Achievable   for
      Most   of   the   Target   Analytes


2.    Chemical   Reactivity     Limitations


3.    Back.groun.dl   Contamination


4.    Precision,  i

             IS  Analytes    (MS  Studies)

                      Meant  14.37.

              Range:  347.    to    51.1%


S.    Accuracy  »
                      Analytes

                  Mean:   1S77.

        Range:   S6.27.    to
3747.

-------
                               463
                  RECOMMENDATIONS
1.   Separate Analysis  for  Acidic  Substances

2.   Testing  on Actual River "Water  Samples

3.   Special  Precautions  to  Control  Background
           Contamination

4.   Different  Approach.  Necessary for
                 Reporting  Separately

                N-Nitrosodiph.enylam.ine

                            &

                     Diphenylamine

-------
                             464



                              MR. FIELDING:  There is a



substitution next.  Gordon Wallace from MIT was unable to



get here, so we are substituting a very nice paper by Larry



Keith of Radian Corporation who will talk about his sampling



and analysis methods data base.

-------
                             465
                                   MR. KEITH:  Thanks, Tom.
     Good morning.  The sample and analysis methods data
base I am going to talk about a little bit this morning
originated in Cincinnati at EPA.   The overall objective is
very similar to Jim King (Viar) and Bill Telliard's (EPA)
List of Lists.  In some places, the data are similar and in
other places it is different.
     Jim King and I have talked about these programs and our
conclusion is that, basically, these are complimentary
programs.  They use different searching techniques, and
there are some significant differences between them.  Why
are we interested in databases of methods and analytes?  The
answer is because the world has become a jungle of analytes
and methods, and it is growing all the time.  I can remember
back not too long ago when life was a lot simpler, and there
were relatively few pollutants that people looked for.  In
the 1970s there were some lists of pesticides, and we didn't
have very many methods.  We had GC with electron capture
detectors, but we didn't have mass spectrometry back in the
late 1960s and early 1970s.
     However, since that time, people have become interested
in more and more different kinds of organic compounds, and,
with that interest, we environmental analytical chemists
have become craftier and smarter and developed many more
techniques.

-------
                             466
     The result is a proliferation of methods that we have
today.
     A compounding factor is that many analytes have
multiple methods that can be used for their analysis.  So,
the question is, what method should be used?
     There are some obvious things, of course, that  method
selection well depends upon  (Figure 1).  First of all, one
must consider the available instrumentation.  If the method
of choice uses a GC/MS method and your lab doesn't have a
GC/MS, you can't use that method.  The same analogy may be
made using an I-cap method for elemental analysis.  If you
only have atomic absorption spectrometers  available, then
you cannot use I-cap methods.
     Another important factor to consider is the
environmental matrix.  A third factor to consider is whether
or not there are interferences present.  Different methods
and techniques may introduce or be prone to different
interferences.  So, sometimes you may need to select the
best technique for your purposes by considering what the
interferences may be in the samples.
     A fourth factor to consider is that various methods
have different detection levels.  When the need for the
greatest sensitivity is paramount, this may be the deciding
factor in selecting a method.

-------
                             467



     Not very often, but sometimes, you may also run into



the problem of maximum holding time.  Some of the methods



have different holding times and in these cases this may



influence method selection.



     Some examples will illustrate the influences that the



above factors can have. Take, for example, 1,2-



dichlorobenzene.  If you want to do the analysis by GC with



a packed column and a photoionization detector, there are at



least three different methods that use that particular



technique (Figure 2).



     However, if you wish to use an electroconductivity



detector, then there are at least three other GC methods



that use packed columns with that detector.



     On the other hand, if you wish to use a capillary



column instead of a packed column with an



electroconductivity  detector, then you may use Method



502.2.  But, if you wish to use an electron capture detector



with a capillary column, then Method 8120 is applicable.



     Finally, if you wish to use a mass spectrometer for a



detector then there are at least six other methods with



packed or capillary columns.



     Consider the same example influenced by its



environmental matrix (Fig. 3).  If you want to analyze for



1,2-dichlorobenzene in drinking water, EPA's 500 series of



methods will apply.  If you want to analyze for 1,2-

-------
                             468
dichlorobenzene in wastewater, then EPA's 600 series of
methods will apply.  If you want to analyze for 1,2-
dichlorobenzene  in sludges, soils, groundwater, and various
other matrices, then EPA's 8000 series of methods will
apply.
     Each of these methods differ in required
instrumentation, sample preparation, detection levels, etc.
     Next consider the problem of interferences with method
selection.  If a non-selective detector (e.g. an electron
capture detector) is used then phthalate esters may
interfere with the analysis.  Or taking elemental analyses
using I-cap as another example, some elements will cause
interferences with that technique that would not exist if
atomic absorption methods were used and vice versa.
     Thus, knowledge of potential interferences  can also be
very important in making method selections.
     Next, consider the influence of detection levels on
method selection (Fig. 5).  Using I-cap to analyze for total
chromium, one can select EPA Methods 200.7 and 6010 which
are essentially identical; they both have a method detection
level of 7 ug/L.
     I-cap instrumentation is expensive and not all
laboratories have such instruments.  However, most
laboratories do have atomic absorption (AA) instruments.
Direct aspiration AA methods have a  detection level of  50

-------
                             469
ug/L.  On the other hand, if you use a graphite furnace
technique (EPA Method 7191), then the detection level for
total chromium is 1 ug/L.  In this example graphite furnace
AA provides the most sensitive detection and also a less
expensive instrumentation than the I-cap.
     The above example is just one of hundreds of variations
where method detection levels may influence method
selection.
     Thus, the overall problem becomes how do you find and
select the most appropriate method for sampling and analysis
without having to be an expert on all of these or without
reading through mountains of EPA reports, the Federal
Register and the literature?
     Although environmental chemists may be familiar with
this detailed information, the people who have to work with
environmental chemists (i.e. engineers, regulators, etc.) do
not often have the knowledge to make the best decisions
involving sampling and analytical methods.   How can we help
them?
     One solution is to provide method summaries from a
sampling and analysis database that can be searched by the
criteria of interest in order to find applicable method
summaries (Fig. 6).  This is essentially what we have done.
In 1987 and 1988, Bill Mueller and David Smith started
compiling a sampling and analysis database at EPA 's Risk

-------
                             470



Reduction Engineering Lab in Cincinnati, Ohio.   (Fig.  7).



They began compiling this database primarily for engineers



and contractors who worked with EPA so that the best methods



for sampling and analysis of a large variety of pollutants



in many matrices could be selected.



     The records were compiled using d-Base III and, as the



database got larger and larger, the searches got slower and



slower.



     At the ACS meeting in Los Angeles in September of 1988,



and suggested that we use a software program written in C



language that we had developed at Radian.  It is a " free-



text" searching program that may be used to search ASCII



text files.



     In October of 1988, I received an ASCII dump of the d-



Base files.  In December, 1988, we did some trials with the



free-text searching program and made minor changes.



     Then in March of 1989, I sent final copies of the



program to EPA and, during the next nine months, we



essentially doubled the size of the database and used a



revised format.  Finally, in March of 1990, the program was



sent to Lewis Publishers for publication in the summer of



1990.



     A private publishing firm was used so that the



publication would get wide distribution and be made  easily



accessible to everybody.

-------
                             471



     This publication is the first-of-its-kind, electronic



reference book (a book on diskettes) with environmental



sampling and analysis methods (Fig. 9).



     Each method and analyte summary is self-standing and it



is about one page long.  What I mean by self-standing is



that all of the information is there to allow you to make a



selection for a method and an analyte combination without



having to go to any other reference.



     There are 150 methods, and 660 Method-analyte



summaries.  The EPA List of Lists program has about 1700



method-analyte summaries.



     One of the main differences between EPA's Sampling and



Analysis Methods Database and EPA's List of Lists is that,



the former has fewer analytes, but contains more information



and in a different format List of Lists program.       The



Sampling and Analysis Methods Database is completely menu



driven, so it is very easy to run.



     The program performs and/or searches.



     It has a hypertext "feel" to its use.   It is not a



hypertext program, but it has a hypertext "feel" to it.  The



menu has a highlight bar, which is moved to the analyte-



method summary of interest.  When the "Enter" key is



pressed, the selected summary is instantly displayed.

-------
                             472
     Once the files have been searched and the appropriate
summaries have been found any of them may be displayed
easily and very rapidly in any order desired.
     Once a summary is displayed, you can page up and down
and browse through it. The length of each summary is about
two or three computer screens.
     Any or all of the summaries may be printed either to a
disk file or to your printer.
     Within the text on the screen, the key words that were
used for searching are reversed video, so they are
highlighted.  This immediately locates in the text the key
words that were of interest. The highlighted keywords do not
appear any differently from the other text when printed.
Key word highlighting is only used with video screen
presentations.
     The publication comes complete with an illustrated
tutorial.  In addition, it has a printed manual.
     The publication is provided in three volumes (Fig. 9).
Volume I covers industrial chemicals.  It has three
diskettes of data, plus Systems Diskette.  Within Volume I,
there are files on chlorinated aliphatic volatile organics;
all the other halogenated volatile organics (e.g.,
aromatics r bromo, chloro and fluoro compounds); non-
halogenated volatile organics; and selected semi-volatile
compounds.

-------
                             473
     The second volume is on one diskette.  It contains
method-analyte summaries of pesticides, herbicides, PCBs,
and polychlorinated dioxins and furans.
     The third volume also is on one single diskette.  It
contains one file with elemental methods and another file
with water quality parameters (e.g. BOD, suspended solids,
pH, etc.).
     The information was grouped into three volumes because
some people may only be interested in selected groups of
analytes and, therefore, no want or need all of the data.
     Figure 10 shows the menu for the searching program.  It
is very simple.  When you select "Search" the program
searches the active file.
     "File" is an option that allows you to move from one
file to another without going back through the menu.
When you select the option "menu", you are returned to the
main menu.
     When you select the "Print" option, you may print
either to diskette files or to paper.
     The "criteria" option is used to enter key words that
are to be used to search the files selectively.  Five lines
of search criteria may be entered.  (Fig. 11).    If
criteria are placed on each line, they are additive during a
search.  Thus, if "sludge" is on line 1 and "water" is
placed on line 2, the program will only find methods that

-------
                             474
have both sludge and water as key words somewhere in the
summary.  In this example, both sludge and water would have
to be present for a method summary to be selected.
     As a second example, if you want to do  "or" searches,
where one or another key word is in a summary (for example,
aromatic or aliphatic), then the piping character is used to
separate those key words on the same line.  Thus, if
"aromatic:aliphatic" was placed on line 1, "water" on line 2
and  "sludge" on line 3, then the searching program would
only find those method summaries that had the words
"aromatic" or "aliphatic" and "water" and "sludge" in their
text.
     Each summary contains the primary name of a chemical,
its CAS number and the applicable EPA method number (Fig.
12).  It also contains a brief description of the EPA
method; a short one-paragraph description of the method.  In
addition, each summary contains the applicable matrices that
the method may be used to analyze.  A very important section
of each summary includes common interferences and solutions
for removing those  interferences if they are known.
     Another part of each summary lists the instrumentation
required — not just the major instrumentation required but,
very importantly with respect to the organics, the gas
chromatographic columns that are needed.

-------
                             475



     The sections involving quality control include the



published concentration range and the published method



detection level (MDL) for the method analyte combination.



In addition, practical quantitation limit factors are listed



when those are available for the RCRA methods.



     The method-analyte precision and accuracy data includes



another part of the QC information summarizes sampling and



preservation instructions, including the maximum holding



time (M.H.T.).  Quality control requirements for performing



the method are also summarized in this section.  The final



information presented lists the EPA reference source.



     An example of the data is illustrated with Aroclor 1242



using Method 8080.  This is shown in Figures 13-15.



     There are also some limitations with this publication



(Fig. 16).  First, not all of EPA's method-analyte summaries



are included yet — specialized analytes, (e.g.



organophosphorous pesticides, and many of the semi-volatile



compounds with GC/MS methods) are not yet summarized.



     These may be added later if there is sufficient



interest in this publication format.



     Second, the publication is not available in the



Macintosh format.  However, people with no computers or who



have Macintosh computers will be able to get a printed



version by the end of 1990..  Of course, that can't be



searched.

-------
                             476
     Lastly, you can't perform the analyses using only the
summaries.  There is not sufficient detailed information
included in the summaries to perform the analyses, but that
is not the purpose of this publication.  Its purpose is to
enable one to make an informed decision as to which EPA
method to select for any given situation.

-------
                                  477



                     QUESTION AND ANSWER SESSION



                                   MR. FIELDING:  Does



anybody have any questions?



                                   MR. PRONGER:  Greg



Pronger from National Environmental Testing.



     I was wondering if this program or the List of Lists



programs allows the user to append compounds or methods to



it.



                                   DR. KEITH:  I will have



to let Jim King from Viar Corporation answer with respect to



the List of Lists. Jim, will you answer that?



                                   MR. KING:  The List of



Lists in the form that we use actually allows us to enter



the information.  A number of EPA program offices also can



enter information if they have programming capabilities with



System J language.  However, the average person could not



add information to this program.



                                   DR. KEITH:  With EPA'S



Sampling and Analysis Methods Database, you would be able to



add compounds or methods to it because, essentially, it is



in two pieces.  The searching program is a run-time program



that can't be changed, but that searching program goes out



and loads an ASCII file into the computer's memory.  One can



easily append information to the ASCII files so, yes, it



would be possible.

-------
                        478
                              MR. PRONGER:  Thank you.



                              DR. KEITH:  Any other



questions?  Yes?



                              MR. LEWIS:  Mike Lewis.



     How much is the software?



                              DR. KEITH:  Well, I hate to



get into costs in a technical meeting.  Let me just say that



it is...



                              MR. LEWIS:  Free?



                              DR. KEITH:  No, it is not



free, of course, because it is produced by a commercial



publisher, so they hope to make a profit on it.  One never



knows whether they will or not.  Volume 1, the industrial



chemicals, will be in the same range of cost as the List of



Lists will be.  It will be under $100.  The other volumes



will be even cheaper than that.



     I brought some information sheets, and for those of you



who may be interested further in this publication, I will



put the information sheets back on the back table. If you



are interested, you can pick them up, and they will give you



more details.



     Any other questions?  Yes?



                              MR. GILLENWATERS:  Bill



Gillenwaters, Newport News Shipbuilding.

-------
                                  479
Any thought about including the spectra to be tagged in with



it when you call the compound?



                                   DR. KEITH:  No, there is



so much trouble to do it with words, that spectra would just



be beyond the intended scope. I will never put spectra in



it.  It is just too damn much trouble and not worth it.



                                   MR. NEIN:  John Nein from



the Naval Supply Center.



     Have you considered correlating any non-EPA methods



such as ASTM, USGS standard methods or AOAC into this



database, or do you feel that would just add to the



confusion?



                                   DR. KEITH:  Yes.



Actually, I have thought about doing that, and, again, it



depends on how much interest there is in this type of thing.



Of course, this data base only has EPA methods in it,



because it was compiled by EPA chemists in Cincinnati.  So,



that is what their interest was.



     But there are many other methods, the ASTM methods and



Standard Methods, etc., all of which could be quite readily,



with a slight massive amount of work, be added to this



database.



     So, I have thought about it.  Really, this publication



is an experiment.  The List of Lists is also an experiment,

-------
                             480
and the key is how helpful these databases will be to the
technical community.  Are people really going to find these
databases to be useful?  If so, then we are on the right
track, and, of course, we can add other data to them.  If
people don't need these types of database summaries,  then
there is no point in adding to them in the future.
     Looking at it from my own point of view, since I don't
know the details of all these methods, I find it to be
pretty useful.  In fact, let me digress and tell you the
first time I ever used EPA's Sampling and Analysis Methods
Database.
     A lady called me to inquire about inorganic methods.  I
am an organic chemist; therefore, I don't know the details
of the many inorganic methods. She told me that she had a
laboratory...and it wasn't Radian Corporation...perform some
analyses for chromium, and the results were inconsistent.
     So, I had my computer on, and I quickly accessed the
EPA inorganic method summaries.  I searched all the chromium
methods while I was talking to her.  Of course, she didn't
know that I had this information available at my fingertips.
I asked what method the lab used and she told me.  I quickly
displayed the summary of that method and I observed it was a
method for direct aspiration atomic absorption for chromium.
However, when I asked about the type of matrix, the samples
were in, she told me that they were solid samples.

-------
                             481
     The interferences section of the summary warns that
high solids content can cause significant interferences.
So, I told her that the lab used the wrong method.
     When I searched for chromium and for solids/ the
computer program found a different method that used graphite
furnace atomic absorption.  Thus, the lab did use atomic
absorption spectroscopy, but it didn't use a graphite
furnace.  It used the  direct aspiration technique and
suffered from interferences.
     The client was appreciative of this information and
said they were going to resample and, this time, send them
to Radian.  So I said okay, that sounds good.
(Laughter.)
     So, the client thought that I knew whereof I was
speaking about elemental analyses, and, as I have told you,
being an organic chemist, I didn't know anything about it.
But that was the first time that I personally used EPA's
Sampling and Analysis Methods Database to try to help
someone find correct methodology.
     Yes?
                                   MR. RICE:  Jim Rice.
     I have a question.  I noticed that you stated that all
of the methods contained in the database were EPA methods
since that was the original objective.
                                   DR. KEITH:  Yes.

-------
                             482



                              MR. RICE:  My questions



involves information that you show for method detection



limit and for precision and accuracy and performance, in



effect, of the methods in various matrices.  Have you



limited yourself to just using the information furnished by



EPA, or have you gone further than that to other major study



results?



                              DR. KEITH:  No, the data has



been limited to just the information that was published in



the EPA methods.



                              MR. PRONGER:  If a user would



buy one of the early versions, is there going to be some way



for him to get an update if there would be further updates



without having to rebuy the entire package?



                              DR. KEITH:  I really don't



know, but let me philosophize.  The publisher publishes



books, and their concept is, instead of treating this



publication like a software program and selling it for know,



a lot of money, to instead treat it like a book and make it



cheap so that everybody will buy it.



     Therefore, I suspect that like when a new edition of a



book comes out, you rebuy the book, right?  And I suspect



that that is probably what they will have to do because the



price is so low.  That is what they will probably have to do



to recover their advertising and production costs.

-------
                             483



     So, my guess is that if there are updates, they will



probably be not at the full price but not free either —



probably at some reduced price in between.  But that is just



a guess.  That would be the logical way to do it if I were



the publisher.



     Okay, thank you very much.



                                   MR. FIELDING:  Thank you,



Larry.

-------
                           484

                    SUGGESTED READING
1.   Johnson, L.D., "Detecting Waste Combustion Emissions,"
     Environmental Science and Technology, 20, 223 (1986).

2.   Johnson, L.D., James, R.H.,  "Sampling and Analysis of
     Hazardous Wastes" in Standard Handbook for Hazardous
     Waste Treatment and Disposal, H.M. Freeman, Editor,
     McGraw Hill, New York, New York, 1988.

3.   Test Methods for Evaluating Solid Waste, Physical/
     Chemical Methods. SW-846 Manual, 3rd ed. Document No.
     955-001-0000001. Available from Superintendent of
     Documents, U.S. Government Printing Office, Washington,
     DC, November 1986.

4.   J.H. Margeson, J.E. Knoll, M.R. Midgett, D.E. Wagoner, J,
     Rice and J.B. Homolya, An evaluation of the semi-VOST
     method for determining emissions from hazardous waste
     incinerators, J. Air Pollut. Control Assoc., 37(9)
     (1987)1067.

5.   R.G Fuerst, T.J. Logan, M.R. Midgett and J. Prohaska,
     Validation studies of the protocol for the volatile
     organic sampling train, J. Air Pollut. Control Assoc.,
     37 (4)(1987) 388.

6.   Johnson, L.D., "Trial Burns:  Methods Perspective,"
     Journal of Hazardous Materials, 22, 143, (1989).

7.   "Handbook, Hazardous Waste Incineration Measurements
     Guidance Manual", EPA-625 / 16-891021, June 1989.

8.   "Handbook, Quality Assurance/Quality  Control (QA/QC)
     Procedures for Hazardous Waste Incineration, EPA-
     625/6-89023, January 1990.

-------
Selection of best method depends on
    • Analytical instrumentation;
    « Environmental matrix;
    • Interferences present;
    • Detection levels; and
    • Maximum holding time.
00
(Jl

-------
      Analytical Instrumentation

Example: 1,2-dichlorobenzene
• GC, packed column, PID
  -  Methods 503.1, 602, 8020
• GC, packed column, HECD
  -  Methods 502.1 601, 8010
• GC, capillary column, HECD
  -  Method 502.2
• GC, packed column, ECD
  -  Method 8120
• GC/MS, packed and capillary columns
  -  Methods 524.1, 524.2, 624, 625, 1624, 1625
CO
en

-------
                           Matrix
Example: 1,2-dichlorobenzene
• Drinking water
  -  Methods 502.1, 502.2, 503.1, 524.1, 524.2
• Waste waters
  -  Methods 601, 602, 624, 1624, 625, 1625
• Soils, sludges, groundwater, wastes
  -  Methods 8010, 8020, 8120
Methods differ in sample preparation,
instrumentation, detection limits, holding times, etc.
00

-------
         Interferences Present

Example: RGBs, pesticides, chlorinated organics

• Methods using nonspecific detectors  have
  problems when phthalate esters, other oxygen-
  and sulfur-containing compounds are present.
  -  Method 8120, GC with ECD
00
00

-------
           Detection Levels

Example: total chromium
• ICP
  -  Methods 200.7 and 6010 : MDL = 7 ug/L
• AA direct aspiration
  -  Method  7190 : MDL = 50 ug/L
• AA graphite furnace
  -  Method  7191 : MDL = 1 ug/L
00
10

-------

Problem:  How to find and select most appropriate
          method for sampling and analysis
          without being an expert or reading
          massive volumes of information
          scattered throughout Federal Register,
          EPA reports, and the literature.
Solution:  A Sampling and Analysis Database in a
          convenient widely accessible format.
vo
o

-------
              Background
1987-1988:
     (EPA
Oct.
Nov.
Dec.
Mar.
Dec.
Mar.
    1988:
     1988:
     1989:
     1989:
     1989:
     1990:
 William Mueller and David Smith
-Cl RREL) compile S & A Base.
Larry Keith obtains ASCII dump.
 Revise presentation format.
 Trial with "free text" searching
 Provide EPA with copies
Double size of database in new format
 New version sent to Lewis Publishers
vo
Jul. 1990: Sampling and Analysis Methods
         Database published

-------
               Features
First of its kind electronic reference book on
sampling and analysis methods;
Each method/analyte summary "self standing";
150 methods;
650 method/analyte summaries;
Menu-driven program;
Performs "and/or" searches of multiple key
words;
Hypertext "feel" to program use;
Page up/down browsing of each summary;
Print any or all summaries  to file or paper;
Highlights the searched text in each summary;
Tutorial  with illustrations on disk; and
Printed manual.
to

-------
Volume I - Industrial Chemicals (3 diskettes)
  • Chlorinated Aliphatic Volatile Organics
  • Other Halogenated Volatile Organics
  • Nonhalogenated Volatile Organics
  • Semivolatile Organics
Volume II - Pesticides, Herbicides, PCBs, Dioxins
          and Furans (1 diskette)
Volume III - Elements and Water Quality Parameters
          (1 diskette)
10
CO

-------
           Menu Selections
Search
Criteria
File
Print
Menu
                                            VD

-------
You can search for five different criteria by entering in the 5 lines
below.
The criteria placed on each line are additive - sludge on line 1 and
water on line 2 will only find entries that have  BOTH sludge AND
water in them.

To search for two or more criteria separate them on a line by PIPING
(|). Thus, aromatic|aliphatic on a line will find EITHER aromatic OR
aliphatic.
Partial words may be searched: vol will find volatile, volume, volt, etc.
Press Function Key F10 when finished; then select SEARCH from
the main menu.
1.
2.
3.
4.
5.
vo
Ln

-------
Each one page summary contains:
  Primary name and CAS number
  EPA Method No. and its description
  Applicable matrices
  Interferences and solutions (if known)
  Instrumentation  required
  Concentration range and MDL
  Practical quantitation limit factors
  Precision and accuracy
  Sampling and preservation instructions
  Maximum holding time
  Quality control requirements
  EPA reference source
10

-------
PRIMARY NAME: Aroclor 1242 (PCB-1242)        Method 8080
TITLE: Organochlorine Pesticides & PCBs
MATRIX: Groundwater, soils, sludges water miscible liquid wastes,
         and non-water miscible wastes
CAS #: 53469-21-9
APPLICATION: This method is used for the analysis of 19 pesticides
              and 7 Aroclor (PCB) mixtures. Samples are extracted,
              concentrated and analyzed using direct injection of
              both neat and diluted organic liquid into a gas
              chromatograph (GC).
INTERFERENCES: Solvents, reagents and glassware may introduce
                  artifacts. Other interferences may come from
                  coextracted compounds from samples.
                  Phthalate esters are common interferences
                  when using an electron capture detector (ECD)
                  so all plastics must be strictly avoided.
                  Exhaustive cleanup of reagents and glassware
                  may be required to eliminate phthalate
                  contamination.  Use of a halogen specific
                  microcoulometric or electrolytic conductivity
                  detector will eliminate phthalate interference.
10
-J

-------

INSTRUMENTATION: GC capable of on-column injections and an
ECD or a halogen specific detector (HSD). Column 1:1.8 meter by
4 mm with 1.5% SP-2250 / 1.95% SP-2401 on Supelcoport.
Column 2: 1.8 meter by 4 mm with 3% OV-1  on Supelcoport.
RANGE: 8.5 to 400 ug/L    MDL: 0.065 ug/L (in reagent water)
PRACTICAL QUANTITATION LIMIT FACTORS FOR MULTIPLYING
TIMES  FID MDL VALUE:
                                           Multiplication
	Matrix	Factor
Groundwater                                       10
Low-level soil by sonication with GPC cleanup          670
High-level soil and sludge by sonication             10,000
Non-water miscible waste                        100,000
PRECISION: 0.21X + 1.52 ug/L (overall precision)
ACCURACY: 0.93C + 0.70 ug/L (as recovery)
ID
00

-------
SAMPLING METHOD: Use 8 oz. widemouth glass bottles with Teflon
lined caps for concentrated waste samples, soils, sediments and
sludges. Use 1 or 2 1/2 gallon amber glass bottles with Teflon
lined caps for liquid (water) samples.
STABILITY: Cool soil, sediment, sludge and liquid samples to
4 deg. C. If residual chlorine is present in liquid samples add 3 mL of
10% sodium thiosulfate per gallon of sample and cool to 4 deg. C.
M.H.T. = 14 days for concentrated waste, soil, sediment or sludge.
M.H.T. = 7 days for liquid samples. All extracts must be analyzed
within 40 days.
QUALITY CONTROL: A quality control check sample concentrate
containing each analyte of interest is required. The  QC check
sample concentrate may be prepared from pure standard materials
or purchased as certified solutions. Use appropriate trip, matrix,
control site, method, reagent and solvent blanks. Internal, surrogate
and five concentration level calibration standards are  used. The
quality control check sample concentrate should contain Aroclor
1242 at 50 ug/mL in acetone.
REFERENCE: Method 8080, SW-846, 3rd ed., Nov  1986.

-------
             Limitations:

"Specialized" analytes and many semivolatiles
using GC/MS methods are not included yet. May
be added if user interest is sufficient.
Not available with Macintosh formats. Printed
version, Compendium of EPA's Sampling and
Analysis Methods will help people with no
computers or only Macintosh computers, but it
can't be searched.
Can't perform analyses from summary
information - only select the best method.
tn
o
o

-------
                             501
                                   MR. FIELDING:  Let's take
about 15 minutes for some coffee.  Can we get back at about
10:20?
(WHEREUPON, a brief recess was taken.)
                                   MR. FIELDING:  Before we
start, I would like to remind people if they have a
question, give your name and the company name so the young
ladies on the side there can get it correctly.  "Voice from
the audience" doesn't look quite as good as it might, and
you don't get credit for your comments.
     Our next speaker is Bruce Colby of Pacific Analytical
who will talk about microextraction isotope dilution GC/MS
determination of volatile organic compounds.
     Bruce?

-------
                        502
                              MR. COLBY:  Good morning.
     Since we are running about half an hour behind
schedule, I am going to stretch my talk out so that we can
catch up.  I knew that would excite you all.
     I am going to talk this morning about a technique that
has been around for a long, long time.  It was used, at
least, I recall people using it probably about 10 years ago
when the organic chemicals industry was being screened by
the Effluent Guidelines Division at the time.  It is called
microextraction, and we, I guess, kind of resurrected it
recently to help solve a problem that we encountered where
no other solution seemed viable.
     The problem we encountered was a series of samples that
contained relatively high levels of some ketones, relatively
high being in the 1 percent or just under 1 percent type
level, and we were looking for the volatile organic species
in the samples.  If we tried purge and trap on these, the
ketones would absolutely wipe out any chance of making a
reasonable determination of the other targets.
     The options we felt we had with respect to dealing with
this problem were to approach the situation by two possible
routes.  One was direct injection of the aqueous sample, and
the other was to resurrect this microextraction technology.
     Well, we decided before we were to strike off on things
that we should establish a few goals for ourselves.  They

-------
                             503



are pretty standard sorts of goals for what, in effect,



would be a method development effort, not a big one, just a



little one.



     We have to determine these volatile organics in the



presence of what I call semi-purgable organics, fairly polar



but nevertheless volatile species.  The ketones fall in that



category.  Phenol, we know, falls in that category.  If you



have a sample with a tremendous amount of phenol in it, it



will get in the purge and trap device and cause trouble for



days to come.



     And we wanted to have a system that was sufficiently



precise, accurate, and with enough sensitivity that it would



be close to the purge and trap technology.



     Direct injection, we pretty much decided, wouldn't



provide us with enough sensitivity.  Purge and trap allows



one to put the equivalent of a 4 or a 5 ml water sample in



the instrument.  The water does not go in but all of the



volatile organics in the water presumably do.



     Clearly, you can't inject 5 ml.  A few microliters is



more like it.  So, this is roughly a factor of 1000 in loss



of sensitivity with direct injection.



     Microextraction falls somewhere in between, and it



provides a couple of interesting things.  One is that it



provides a mechanism to generate a little bit of additional



chemical selectivity.  By properly selecting the solvents,

-------
                             504
we can discriminate against things that are more polar and
keep them out of our extracting solvent and leave them in
the water.  That was the big key for what we were up to.
     The other thing is it is relatively simple/ but when
you are trying to solve a problem that is tough, even if you
have to go at it a tough way, that seems okay.
     The problem with microextraction historically has been
that it is not really as sensitive as purge and trap and it
doesn't really have the accuracy and precision that you
expect to get with purge and trap or direct injection, for
that matter.
     The sensitivity problem we thought we had a decent shot
at by using one of the newer GC/MS instruments that is on
the market.  They seem to have gotten much more sensitive in
the last year or so.  We used a VG Trio 1 on this job.  I
suspect that there are several other instruments out there
that could do the job just as well.
     Our older equipment, we know, cannot do the job.  So,
it really is confined to the newer GC/MS instruments.
     We thought we could beat the old problems of the
accuracy and precision that are associated with
microextraction by going to isotope dilution.  The problems
with accuracy and precision, incidentally, are associated
with the fact that there can be some pretty significant
matrix effects.  We heard something about matrix effects

-------
                             505



yesterday, and isotope dilution does help us get around



these to a reasonable extent.



     The method I have outlined in my first  slide.  We take



35 ml of the water sample, add the labeled compounds to it.



We are using a 40 ml VOA vial, we call them.



     We then add 1 ml of hexadecane, cap the vial, shake it



up, let the layers separate, pull off 1 ul of the



hexadecane, shoot it onto a 30 m DB-624 megabore column that



is directly coupled to the GC/MS ion source through a



restrictor, use a reasonable sort of temperature program,



acquire full scan mass spec data.



     Hexadecane elutes after all of the target analytes.



So, at that point, we shut down the filament and let the



solvent through, getting the targets out before the solvent



rather than after it which is the more conventional way.



     The target analytes are identified and quantified,



basically, according to the methodology described in 1624.



     Well, a blank looks something like this which I guess,



might look a little bit horrifying, but these peaks are



actually very, very small, as you will see in the next slide



or two.



     We see the air come through.  Principally, it seems to



be argon, the C02 staying behind.   This blank is a water



sample that we know to be clean extracted with the



hexadecane and just shot.

-------
                             506
     We see some silicon compounds.  The silicon compounds
are not reproducible.  We suspect they probably are
associated with the analyst not properly cleaning or
handling the teflon caps on the vials.  We don't know that
for a fact, but that is what we suspect, and we haven't been
back to verify it.  They don't cause a problem in the actual
analytical work, so we haven't made a big issue of it.
     There appears to be a little bit of hexane in the
hexadecane as well.
     A chromatogram for a standard.  This is the high point
on the calibration curve, 10 ng/ul injection.  It gives us
some pretty nice looking peaks.  I cut the chromatogram off
just about at the point where we shut the instrument down
and the hexadecane starts to come out.
     Low concentration standard.  Now you can just start to
see the silicon compounds in there.  So, you can see that
even though they looked really huge in the previous one,
when we get down to injecting 100 picogram per analyte, that
is still a pretty small quantity of that silicon compound.
     Looking for targets in the sample, we use a software
that comes with the instrument.  This is looking at a 1 ng
spike into a water sample.  Here you are looking at 1,2-
dichloropropane.  Just picked one at random.

-------
                             507
     The top trace is the total ion chromatogram.  You see a
little bit of a blip in that.  Again, it is only 1 ng, so
quite a small quantity.
     Quantitation mass area is shown in the center drawing.
The bottom one I won't bother with.
     So, the targeting that we used was all automatic.  We
didn't go back and edit any of the data.  We did take a look
at it, but whatever was there we accepted, because we just
didn't have enough time to fool around with it, basically.
     Spectra, again, looking at that dichloropropane that we
had before.  It is an unbackground subtracted spectrum or
raw data spectrum on top.  You have a lot of peaks from
background within the equipment.
     If you do a background subtraction, you get the center
kind of a spectrum.  It certainly has all the major peaks.
It even picks up the minor isotope peaks.  There is a set at
97 and one at 112.  There are a few other peaks that show up
as well.
     The bottom spectrum is a spectrum from the NBS library.
     So, there seems to be plenty of sensitivity for what we
are up to.
     What we do now is calibrate the instrument at a very
low level compared with conventional GC/MS technologies.  We
are starting at a high level of 10 ng/ul and going down to
100 picograms per microliter using a 5 point calibration.

-------
                             508
That is roughly equivalent to the Method 525 sort of
calibration system.  I think we are going to get a little
bit more information on that from another speaker later on.
     Basically, it is a drinking water method.  Here, we are
using the kind of instrument sensitivity you would use for
that drinking water method, but we are applying it to
volatile analytes in a wastewater, but the key is the
sensitivity.
     The calibration compounds are spiked into a water
sample.  They are prepared in methanol, then extracted from
the water sample with the hexadecane, and we do our
calibration of the target analytes, again, by isotope
dilution, and the labeled compounds we do by a mean response
factor calculation.
     Target analyte calibration curves were generally very
good.  What I have plotted here for about...I think it is 23
compounds... is the correlation coefficient that we get for
that 5-point calibration curve.
     The two on the far right actually were correlation
coefficients of 1.0000.  I don't know what the fifth decimal
place would have with it, but that was pretty impressive.
     We have a couple that don't look quite as good on the
far left.  Those turn out to be tetrachloromethane and
toluene.  I went back and checked those, and it turns out
that each of those had a bad measurement, and we just left

-------
                             509



it in the data because we were taking the philosophy we



would do it all automated.



     So, we have a couple of calibration curves that are a



little bit suspect.  They, nevertheless, are pretty good



calibration curves.



     We wanted to look at calibration a little bit more



carefully, because we are pushing down into a very low



concentration regime with this method.  So, we also look at



the data with respect to higher order regression fits to the



line.  This is an allowable part of the Method 525



technology.



     In order to compare the quality of calibration curves



in these higher order fits, we can't really use correlation



coefficient.  We have to go to some other statistic to



evaluate the quality of the curve.



     What we use is a statistic called the mean residual.



Basically, what we do is take the calibration data and



generate a regression line of some type through the



calibration data, in this case, a linear fit to the data.



We then plug those calibration data back into that



regression line and calculate concentrations.



     Now, we know the concentrations, because they are our



standards.  We made them up, and we are now plugging the



data back into the calibration curve, and we get some

-------
                             510
concentrations that are slightly different than the true
values.
     It is the difference between the true value and the
determined value as having plugged it back into the curve
that is the residual, and what I have plotted here is that
correlation coefficient from the previous slide on top
versus the mean residual.
     The mean residual also has the advantage that it can be
expressed in terms of a measurable quantity, in this case,
ng/ul.
     For the isotope dilution measured compounds, we have a
mean residual of approximately 1.28 ng/ul.  So, our low end
of the calibration curve, on the average, is probably going
to be roughly plus or minus 100 percent.  So, we are pushing
it kind of hard down there.
     Now, that average includes those two bad actors up
there, and if we take those two out, then the mean residual
is quite a bit nicer.
     Well, we looked at a second order calibration curve in
addition to the linear one, and we got what I would call a
pretty substantial improvement in the statistic known as the
mean residual.  You can see that on here.  Essentially, all
of the second order fit residuals were lower than the linear
regression residuals.

-------
                             511
     It is tempting, in that sense, to want to use a second
order calibration curve, but we will look at some
implications of that later on that may change your minds.
     It turns out that we had a couple of compounds in there
that we calibrated by internal standard because we didn't
have the labeled analogs for them.  I just throw them up
here because we did make the measurements on them, and we do
have some information on them.  The mean residual on the
internal standard compounds is really fairly similar to that
of the linear regression isotope dilution values with the
exception that they literally had to be done by a second
order fit, and I suspect that has something to do with
matrix effects associated with spiking of methanol into the
samples and the quantities of solvents used.
     We also had to quantitate the labeled analogs in order
to get some kind of an evaluation of recovery.  So, I have
again plotted the mean residual for the labeled analog
measurements here.  Again, these were done by mean response
factor, and the mean residual on this is on the order of
about 0.13 ng/ul.
     In a quick summary on the calibration, isotope dilution
looked pretty good by linear regression.  Internal standard
data looks like it is going to have to be done by secord
regression, second order fit.  And labeled analogs, as they
are in there at a constant amount and we only want...our

-------
                             512
interest in them was fairly basic and single point
calibration.  Our mean response factor was all we were after
there.
     The next thing after the instrument is calibrated is to
back off and say okay...I almost hate to bring up the
term...what are the method detection limits?  In this
calculation, we used the equation that is in the Federal
Register, basically, the student's t value times the
standard deviation.  So, it is roughly three times the
standard deviation of...I guess we used eight replicate
measurements, and these were prepared and measured
separately.
     When we look at these, now we can see the difference
between linear regression and second order calibrations.   It
is pretty clear that there is not much difference between
the two, and if you try to sort out where the lines go on
the screen, it is just about next to impossible, but you can
only see what, in effect, is one set of data.
     So, even though the calibration curve looked better
with the second order fit, when we start applying it to
actual measurements of spiked samples, we don't see that
improvement.  I don't think that we are at a stage where we
can definitely say that a second order fit isn't needed,  but
the implication is at this point that a linear fit will
probably be okay.

-------
                             513



     The few compounds we did by internal standard.  The



mean detection limits there were about roughly a ballpark



factor of 2 to 3 higher than they were by isotope dilution.



     And we could calculate a detection limit for the



labeled analogs.  I don't know quite what it means, but at



least it gives you some sense for the reproducibility of the



data at the level at which they were spiked.  Again, they



are roughly the same as the internal standard measured



numbers.



     The method detection limits I broke out in a summary



slide here with the gases and the non-gases separately.  It



turns out that the compounds that are generally considered



gases, chloromethane, vinylchloride, and so forth, they were



on the method detection limit slides, those were the



compounds on the left.  They are harder to handle.  They are



less reproducible, and using this method detection limit



calculation, they then have a higher detection limit.



     It appears that the method detection limit for most of



the volatile compounds, i.e., the non-gas compounds, is



about 6.4 ug/1 according to the calculation.  The gases are



substantially worse.  I think we probably could improve that



some by fine-tuning our sample handling a little bit and



spiking.  For what we were involved in, they were not



terribly relevant, however.

-------
                             514
     The internal standard and labeled analog values, again,
are very similar to one another.
     I put the standard deviation up there so you could see
how much variation there was involved in the mean method
detection limits that I put up there.  The isotope dilution
values are reasonably tightly packed around that 6.4 value.
So, most of them were behaving quite well.
     We also looked at the recoveries of the spiked
compounds.  Keep in mind now that isotope dilution has an
inherent recovery correction associated with it.  We did the
isotope dilution measured compounds using both linear and
second order polynomial fits on the calibration curve.  As
we saw with the detection limit numbers, the way we
calibrate for isotope dilution doesn't appear to be very
significant when it comes to measuring values in the
samples.
     The percent recoveries we would expect to average out
around 100 percent for the non-gases.  They were about 109
percent using a linear fit and about 108 percent using a
second order fit.  So, there is effectively no real
difference.
     For the gases, they weren't quite as good.  There
appeared to be a positive bias that may have been associated
with the order in which the compounds were spiked into the
sample.  I am not sure.

-------
                             515
     Internal standard values, again, as we are spiking our
standards into a water and extracting them out, there is an
inherent recovery correction going on here, and, again, we
would expect to see the recoveries showing up at around 100
percent.  There appear to be some biases in this, three
compounds being quite high and three quite low, but I don't
really personally believe there is enough data there to say
much.  Certainly, there is a lot more spread in these
recoveries than there was in the isotope dilution values.
     The labeled analogs, like the internal standard
measured compounds, had recoveries over a fairly wide range
from roughly 50 to 250, but what we are interested in the
labeled analogs is can we get a decent signal from it, and
if we can, we are generally quite happy.  If we can't detect
the labeled analog, then we are pretty darn sure we can't
detect the naturally abundant compound and we know we have a
problem with the analysis.
     So, we are basically at this point interested in seeing
the labeled analog, not so much concerned about whether we
got 47 or 82 percent recovery.
     Percent recoveries, in summary, we are looking at
isotope dilution non-gases again, about 109 percent, pretty
tightly grouped.  Standard deviation of 7 percent.

-------
                             516
     Isotope dilution, gases, again, there appears to be
some positive bias, but there is quite a bit of variation in
that, a 16 percent standard deviation.
     Internal standards, even more variation.
Interestingly, quite similar to the labeled analogs and some
bias with the labeled analogs, again, possibly having to do
with the order in which things are spiked in.
     In summary, we have a method that for most of the
volatile organic compounds that are not highly polar, we
have pretty decent detection limits with this technology
using the new GC/MS instrumentation that is available to us.
For the gaseous compounds, not quite as good.  We know they
are tougher to handle.
     Recoveries are in the vicinity of 100 percent.  So, it
starts to look like a fairly viable technique for dealing
with some kinds of samples.  We believe the technology could
be pushed into applications where other kinds of complex or
interference situations are an issue, and we expect to be
trying a little bit with the base neutral and acid fractions
in the future.
     I might add that the base neutral and acid fractions
may have an additional benefit, and that is that by working
with relatively small samples and, in particular, smaller
quantities of solvents, we are involved a pollution
reduction problem which is, for some of us in the analytical

-------
                             517



community, becoming more and more of an issue, the



chlorinated solvents.  We know where they go, and the



regulators are figuring it out, and I think EPA is probably



coming under a little pressure on all this, too.



     That takes me to the end of the data that I had.  I



think I am doing a little bit in helping catch up, but I



would be more than happy to try to answer a few questions.

-------
                             518



                QUESTION AND ANSWER SESSION



                              MR. FIELDING:  Does anybody



have any questions?



                              MR. MCCARTY:  Harry McCarty,



Viar and Company.



     Bruce/ presumably, your calibration data, since you are



spiking your standards into the reagent water, you are



spiking varying volumes to get the different concentrations,



did you look at making up an equivalent volume of methanol



so that all of them end up with the same amount of methanol?



Because that may explain some of the problems.



     We have done a couple of studies, and it seems to help



some of the calibration data.  I would assume with isotope



dilution, it would just make it even better.



                              MR. COLBY:  I don't think it



would affect the isotope dilution data at all.  I believe it



would probably have some effect on the internal standard



compounds, but they were not of real interest to us in this.



They were in there just because they happened to be there,



and we looked at them as a consequence, but yes, I am quite



sure you are right.  That would be an improvement.



                              MR. MCCARTY:  It is a



relatively simple thing, you know, when people are doing it



in practice, and it may make a significant difference there.



                              MR. COLBY:  Quite true.

-------
                                 519
          Anybody else awake?
(No response.)







thank you.







Bruce.
MR. COLBY:  If not, okay,
MR. FIELDING:  Thank you,

-------
                    520
MICROEXTRACTION ISOTOPE DILUTION GC/MS
   DETERMINATION OF VOLATILE ORGANIC
               COMPOUNDS
                    by
    Dante J. Bencivengo and Bruce N. Colby

               PacificAnalytical
                    and

              James S. Smith

                 Trillium, Inc.

-------
                      521
             EXPERIMENTAL GOALS
• Determine volatile organics in the presence of large
  quantities of semi-purgabie organics.

• Provide detection limits equivalent to purge  & trap
  GC/MS.

• Provide accuracy and precision equivalent to purge &
  trap GC/MS.

-------
                        522
               MICROEXTRACTION
Advantages:



        • Chemical selectivity



        • Simplicity



Disadvantages:



        • Sensitivity



        • Accuracy



        • Reproducibility

-------
                       523
               METHOD SUMMARY
• Spike 35 mL sample with Method 1624 labeled com-
  pounds and internal standards.

• Add 1 mL hexadecane, shake and wait for layers to
  separate.

• Inject 1 uL onto a 30 m DB-624 Mega-bore column
  directly coupled to the GC/MS ion source through a 1
  m 0.25 mm restrictor.

• Program from 40 to 180 °C at 8 °C/min with initial and
  final holds of 3 and 20 min respectively.

• Acquire full scan data until just before hexadecane is
  eluted (filament turned off while solvent is eluted).

• Identify targets and quantitate via isotope dilution.

-------
006    008    00*.    009     80S
                                              .   age .,..,. •??•.......
                                                                        8
in
      U.S
                                         T
                                              98 • ₯
                                                               uoBay
                                                                        SIX
                                                                        881
                                                 OUM :
                                                                        Z0TT
                      lAIVUOOlVlAIOHHO HldllAIVS >!N\na

-------
                STANDARD CHROMATOGRAM (10 ng/uL)
1702
06-flpr-90
    Pacific Analytical
USTP 10.0
UG»4 TRI01
 IC04U06F

 100
XFS
                                  13.71-  41411
                                                         '" 900'.'.'' ie'aa
                                                                     Ol

-------
                STANDARD CHROMATOGRAM (0.1 ng/uL)
1251

0G-flpr-90
                 Pacific Analytical

      Satiple; USTDJI.M	
     VG»4 TRI01
 IC04M86B
100
XFS
        1.49
  0
 Sen
                                                 11.13
                                                       470865
                                                          *IC
                                                           ll
                                                         61
                                                              to
100
200   300    400
500   600
700
900   1000

-------
07-Apr-98
                   TYPICAL TARGETING (1ng/uL)
            Pacific Analytical
Sanple; Matrix Spike
                 SCN 560
                 VS1945MS                    Pred :569
                 MQAID 30 1,2-Dichloropropane

                 100 -         8-
                 XFS
            565
570
575
                                                        UGft4 TRI01
                                                                           U1
                                                                           to
                                                                           -J

-------
TYPICAL MASS SPECTRUM AND LIB SEARCH
0008

Pacific Analytical VGft4 TRI01
07-flpr-90 Sanple: Hatrix Spike
VS1945MS 569 (8.553)
100

XFS-
0
A

39
/III
• • • • i • • • •
0


NBS 102 Hits : 102 Searched
11328
4144 57 63
/
. 49 \
\ 1 1 1 1 1 1 1 1 1 1 1 1
1 I'l 1 1 1 1 i 1 IH 1 1 1 II
VS1945NS'569
100


XFS
0
ttl


39 41
38 | 1 ,,«si
\ 4,9/


. 1 1
7
?67


63
62
\

^



F
l
QQ

,
6016


76
75Nlfi, 90 97. . 1*2 130 . 148
F:707 207V:PRbPAIIE, i,2-i>iCHLfaRb~ ' ' 	
100


XFS-
0.


39xy
^.ll I/2 4I9/1
M/Z30 40 50
63
62
\

H



r
255



..\lh,. . . .?,?/ ..if2..
60 ' 70 ' 80 '90 ' 100 ' 110 ' 120 ' 130 ' 140 ' 150
                                             U)
                                             to
                                             00

-------
                     529
          INSTRUMENT CALIBRATION
Calibrate at 0.1, 0.5,1,5 and 10 ng/uL.

Prepare working standards in methanol.

Spike into water and extract into hexadecane.

Calibrate Labeled Analogs using mean response fac-
tors.

Calibrate Target Analytes by isotope dilution.

-------
           Correlation Coefficient
     CO
     •v]
0
CD
^1
Ol
P
CD
CO
0
CD
CO
CJ1
CD
CO
CD
CD
CJ1
o
o
en
o
o
3
o
c
3
(/)
o
-J
o
13
m
D
                                      ^^••1

                                      6
                                      z
                                      o
                                                 r;
                                                 CD
                                                 33
                                                 H
                                                 o
                     oes

-------
  ISOTOPE DILUTION CALIBRATION
1.005
                  Correlation Coefficient
                        Mean Residual
   O)
   c
                                       0-3
                                             tn
                                             CO
                                          C
                                          C3
           1  I  I  I I  I  I  1  1 I  I  1  I I  I  T
0
                 Compound

-------
ISOTOPE DILUTION CALIBRATION
       Linear Regression
               ii  i i  i i  i  i
                                    Ul
                                    W
                                    to
             Compound

-------
 INTERNAL STANDARD CALIBRATION
0,25
           (2nd Order Fit)
                                   U1
                                   w
                                   w
              Compound

-------
LABELED ANALOG CALIBRATION
                                en
                                CO
           Compound

-------
                         535
             CALIBRATION SUMMARY
Category
Calibration
Mean Residual
Isotope Dilution



Internal Standard



Labeled Analogs
Linear Regression        0.128 ng/uL



Second Order Regression  0.172 ng/uL



Mean Response Factor    0.129 ng/uL

-------
 ISOTOPE DILUTION DETECTION LIMITS
5"
*O>
o
                            Linear Regression
                            2nd Order Fit
                                          00
                                          en
                  Compound

-------
INTERNAL STANDARD DETECTION LIMITS
-J
^)
o
100
 90
 80
 70
 60
 50
 40
 30
 20
 10
 0
                                      Ul
                                      00
                                      -J
                Compound

-------
LABELED ANALOG DETECTION LIMITS
 100
 90
 80
 70
 60
 50
 40
 30
 20
 10
en
(A)
00
         1	1	1	1	1	1	1
               Compound

-------
               539
METHOD DETECTION LIMITS SUMMARY
Category
Isotope Dulution, non-gases
Isotope dilution, gases
Internal Standard
Labeled Analogs
MDL (ug/L)
6.4
38.4
12.1
12.0
SD
2.7
30.9
13.2
10.2

-------
    ISOTOPE DILUTION RECOVERIES
O
o
0)
DC
•i-*
c

-------
  INTERNAL STANDARD RECOVERIES
0)

O
O

-------
  LABELED ANALOG RECOVERIES
250
                                  ui
                                  *>
                                  to
             Compound

-------
              543
PERCENT RECOVERY SUMMARY
Category
Isotope Dulution, non-gases
Isotope dilution, gases
Internal Standard
Labeled Analogs
%R
109
119
96
118
SD
7
16
44
44

-------
    544
METHOD SUMMARY
Category
Isotope Dulution, non-gases
Isotope dilution, gases
Internal Standard
Labeled Analogs
MDL (ug/L)
6.4
38.4
12.1
12.0
%R
109
119
96
118

-------
                             545



                                   MR. FIELDING:  Our next



speaking will be Mr. Jim Eichelberger of EMSL-Cincinnati who



will discuss the liquid-solid extraction for determination



of acid herbicides in drinking water.

-------
                             546
                              MR. EICHELBERGER:  Good
morning.
     Actually, what I am going to try to do today is give
you a status report of two projects we have going on in the
laboratory in Cincinnati, and one of them is the subject on
the agenda.  The second is a project we have going on to
develop a cleanup procedure for sludge extracts.
     The reason they are status reports is because neither
one of them is completed yet, and we hope to have them
completed later this year.
     The first project I would like to talk about a is
drinking water project that is an attempt to use disk
technology, solid phase disk technology, for the extraction
of the chlorinated acids, the herbicides, from  groundwater
and finished drinking water.  This project is being
conducted by Jimmie Hodgeson in our laboratory in EMSL-
Cincinnati, and he couldn't be here.
     For those of you who would like some more details on
this work, you might want to take down his phone number.  I
know he will be happy to talk to you.
     Actually, what we are trying to do is develop a
procedure that is better than what we have on the books
right now for doing chlorinated herbicides.   That procedure
is Method 515.1 which is currently in the drinking water
methods manual that was published in December of 1988.

-------
                             547



     A summary of that method is the following:  it uses a 1



liter sample in which the analyst adjusts the pH to 12 with



sodium hydroxide and shakes it for an hour to hydrolyze the



derivatives.  Then he does a methylene chloride extraction,



which is a solvent wash, to remove all possible



interferences, then readjusts the pH to 2 with sulfuric



acid.  He then does the conventional three serial



extractions with ethyl ether, concentrates that extract, and



does a solvent exchange to methyl tertiary butylether in



methanol.  Methylation is achieved with diazomethane, and



the separation and determination are done by capillary



column GC with an electron capture detector.



     Folks in the lab tell me that this procedure for a set



of five to seven samples takes approximately eight to ten



hours.



     So, what we have is a fairly cumbersome method that



uses a liquid-liquid extraction with an undesirable solvent,



and is time consuming.  So, we are trying to eliminate some



of the undesirable features of Method 515.1.



     This is the set of analytes for 515.1, and we are using



that same set in our study to come up with a better method.



You have your garden variety phenoxyacetic acids, some



phenols, and some other types of compounds, all being



herbicides.

-------
                             548



     Next, the disk technology.  The disks are manufactured



by the 3-M Corporation.  Currently, there are only two types



of disks available commercially, and they contain the C8 and



the C18 adsorbents.



     This is one of the sets of equipment that you can use



to employ this technology.  I don't think it is a  good idea



to grab the disk with your fingers, but you could put it on



with a tweezers and seal the funnel to keep the apparatus



from leaking, and the entire thing sits on a suction flask



to which you apply a certain amount of vacuum, and put your



samples through the disk.  It's a fairly simple,



straightforward technology.



     I showed this slide last year, but for those of you who



weren't here, this is a picture of one of those disks.  The



is a Ci8 disk, if  I am not mistaken shown from the side.



You are looking at the thickness of it.



     What you see are the silica particles coated with the



GIB and interspersed in a teflon matrix.  I don't know what



the degree of magnification is, but it is a whopper.  I like



to show this slide.



     This is another picture of a disk from the side showing



the adsorptive capacity.  We used Disperse Dye #1 as a



surrogate.  It demonstrates the disk technology and how



compounds are going to behave.  The red layer there is



adsorbed right at the very top of the disk.  So, you can see

-------
                             549



the concentration right there on the top.  It doesn't



penetrate down into the disk far at all.



     So, after some experimentation...and bear in mind that



this isn't final yet, this is where we were about three



weeks ago...we have decided to use a 100 ml sample. We tried



liter samples, but we saw some breakthrough.  So, the ideal,



and probably the volume that we will decide on, is between a



liter and 100 ml.  I think it is going to be around 0.5



liter.



     You adjust the pH of the sample to 2, but before you



put the sample onto the disk, you have to pretreat the disk.



This is standard operating procedure for solid phase



extractions.



     First, you elute the disk with 20 ml methyl tertiary



butylether.  That is going to be the final eluting solvent.



So, the attempt here is to remove any interferences that



possibly could end up in the final elution.



     Then you pull a little air through it for a few minutes



and put some methanol through which is supposed to activate



the disk, and then introduce some reagent water before



putting the sample through.



     Right now, they are using a throughput time of 5



minutes for 100 ml.  The 3-M guys tell us that they



routinely put 100 ml a minute through the disk and find no



detrimental effects using that quick throughput.

-------
                             550
     Then you air dry the disk after the sample has passed
through.  That really doesn't remove all the residual water,
but it removes a great portion of it.  Then you elute with
the eluting solvent which is 10 percent methanol in methyl
tertiary butylether.
     You dry that over a little sodium sulfate just to
remove the residual water, and methylate with gaseous
diazomethane.  The separation is done on a GC capillary
column, and the determination with an electron capture
detector.
     Now, the sample prep time here...by the way, that line
is supposed to be at the bottom of the slide, but my Harvard
Graphics expertise isn't good.
     The sample prep time for the same size set of samples
is now cut to two hours.  So, we think we have a fairly
simple, straightforward procedure with little concentration
and  no solvent exchange.  You methylate in the eluting
solvent, and it looks as though we have a simple,
straightforward procedure.
     This is a little graph I put together on some results
testing the C8 and the C18 disks.   The third bar on the right
of each group will be discussed in a second or two.  You can
see the Ci8, in general,  seems to be a better matrix for the
adsorption.  The C8 gave a little bit lower recoveries for
the compounds on the x-axis than the C18 did.

-------
                             551
     We noticed when we were doing  the initial work that
when we used metal funnels, the old millipore type systems
that we had around the lab for a long time, and some of the
metal had peeled off, we would get reduced recoveries, and
we would find a little green precipitate in the bottom of
the eluate.
     It must be that the low pH is reacting with those metal
funnels.  So, in our final report, we are going to recommend
that glass funnels and glass filter holders be used, and not
the metal which will eliminate that problem.
     Again, we tried the old salting out trick.  We used a
15 percent ammonium sulfate added to each 100 ml, and we
studied the compounds on the x-axis at 5 ug/1.  That is not
very focused, but the lighter bars on the left of each group
are the salted results, and the grey or striped or whatever
they are on the right of each group are the unsalted.
     You will notice, except for one compound, dichlorprop,
that the unsalted results are always higher.  In solid phase
extraction work, that is common, and the only thing I can
think of is the compounds are salted out of the water before
they get to the disk.  They are plated out on the glassware
and whatever else they come in contact with.  So, salting
out in solid phase technology just doesn't work.
     This is a list of some comparisons of recoveries and
relative standard deviations of some of the compounds that

-------
                             552



are analytes in Method 515.1.  The column on the left of the



recovery section is from the disks, and on the right is from



liquid-liquid extraction.



     You can see there are still a few problems. Bentazon



still is lower on the disk.  Jimmie assures me that these



aren't insurmountable problems and we will overcome all



this.



     In some cases, we are better off with the disk than we



are with liquid-liquid extraction.  For the DCPA, we have



come quite a ways, and dalapon is another problem with the



disk.  So, we are working on those problems right now, and



we should have them resolved when the final report is



published.



     In the relative standard deviation portion of that



slide, the disks look much, much better than the liquid-



liquid extraction.  So, it looks like the disks will work



for these compounds.  The jury is still out on a few of



them, but I think we can resolve the problems.



     I know there are some of you out there interested in



looking at alternate forms of methylation.  We have been



looking for the last year or two, at some of the different



methods, and they are listed here.



     I have data today on the one that looked most



successful out of those four.  The borontrifluoride-methanol

-------
                             553



is just not an efficient methylating agent for the number of



compounds that we have in the method.



     Sulfuric acid-methanol looked promising, and I will



show you some results of that in a minute.



     The bottom two we tried, and the results were pretty



miserable.  So, we really haven't come up with anything



great in those three areas.



     Now, the sulfuric acid...and, by the way, this was done



by taking a small amount of methanol to which a 5 percent by



volume sulfuric acid solution had been added, spiking in



these compounds, and heating the solution to 70 degrees C



for two hours.  Then the solution was diluted with reagent



water a microextraction was done with methyl tertiary



butylether.



     For these compounds, this the good news.  Everything



worked really well here.  The phenoxy acids seemed to



esterify real well with this form of methylation, sometimes



better than the diazomethane.  This is the bad news.  The



phenols are terrible, and some of the other compounds just



don't methylate under those conditions.



     So, we still haven't found a suitable replacement for



diazomethane for all the compounds on the list.



     This project should be completed sometime this summer.



Jimmie is going to write up the results.  We will publish it



in one of the journals, and we will put a method together

-------
                             554
that should be available to the Office of Drinking Water
maybe in the fall.
     The second topic I wanted to talk about a little bit is
a cleanup procedure that we have been working on for sludge
extracts.  Back in September of last year, we delivered a
method to the Office of Water for extracting priority
pollutants from municipal sludges, mostly digested sludges
and filter cakes.
     The method uses 200 ml of sample.  We were trying to
come up with an extraction procedure that would allow us to
handle highly contaminated sludge samples so we didn't have
to dilute them, and end up with method detection limits in
the parts per million range.
     So, our method consisted of taking 200 ml of sample or
a 10 g equivalent of dry solids, centrifuging that sample,
separating the liquid from the solids, and doing a
continuous liquid-liquid extraction on the liquid and a
sonication sample preparation on the solids.
     The sonication consisted of three steps, first using
methanol followed by 1:1 methanol:methylene chloride, and
the third was the methylene chloride step.
     Then the methanol was separated from the solution.  All
the analytes were contained in the methylene chloride, and
we tried to concentrate that down to 10 ml.

-------
                             555
     Well, what we did  was create a bit of a monster.  We
created a procedure that extracted the analytes more
efficiently, but it also extracted all the unwanted gunk
that is found in a sludge sample just as efficiently. So, we
immediately realized that we had a problem and that we
needed an efficient cleanup procedure if this extraction
procedure were ever going to work.  So, we did some work
with silica gel adsorption, gel permeation chromatography
(GPC), and solid phase cartridges.
     The two that showed the most promise were the silica
gel and the GPC.  Since those techniques were already being
used in analytical laboratories that were doing sludge
samples, we pursued those two approaches.
     We tried to determine if varying the deactivation state
of the silica gel could improve the efficiency of cleanup of
sludge extracts.  We varied the deactivation state up to 30
percent water, and found that 10 percent deactivation was
about the optimum condition for silica gel cleanup of sludge
extracts.
     This slide just demonstrates on three different sludges
the amount of cleanup we can get with silica gel.  The three
sludges were collected in Cincinnati on three different days
to show you the variability in the sludges from one source.
The second column shows the mg/ml of residue that contained
in the sludges.  After the silica gel cleanup, which

-------
                             556
consisted of 65 g of 10 percent deactivated silica gel we
eluted with 425 ml of methylene chloride, you can see the
percent cleanup in the column on the far right.
     The percent cleanup is the difference between column 3
and column 2 divided by column 2.  So, in the case of sludge
A, it was almost not worth the effort.  In sludge C, it did
a fairly decent cleanup job.
     So, we figured that, at this point, we had a technique
that could possibly, when used with another technique, give
us a decent cleanup.
     We tried coupling this with gel permeation
chromatography.  We studied it with the silica gel process
prior to the GPC, and the GPC prior to the silica gel.
Results showed no difference.  It makes no difference which
one you do first.  You end up with the same amount of total
cleanup, about 30 percent, which wasn't enough to give us
the ability to concentrate the sludge extract to get decent
method detection limits.
     Then we studied a number of different bio-beads.  We
had done all the original work with SX-3.  We thought maybe
SX-2 or SX-4 would be more efficient in the GPC process.  It
turns out that no matter how we studied it, silica gel first
or second, the bio-beads gave us no improvement in overall
recovery.

-------
                             557



     By the way, the GPC that we had been using up to now



was done with all methylene chloride as the mobile phase.



     Due to the nonpolar nature of the types of materials



that we wanted to remove from the sludge extract, we thought



that it might be advantageous to add a less polar solvent.



     So, we experimented with different amounts of normal



hexane in the mobile phase in the methylene chloride, and



this slide shows you what we found.  The bar to the left of



each group is the gel permeation chromatography alone, and



the shaded bar is the silica gel plus the GPC result.



     As we increased the hexane in the methylene chloride,



our cleanup was increasing dramatically.  If we added 90



percent hexane to methylene chloride, we were getting almost



100 percent removal of all the gunk in the sludge extract.



However, that presented problems.



     In the GPC, when you add that much hexane, you have



memories in the column, and you have to rinse the column



before putting the next sample on, and this is too time



consuming.  So, we had to back off, and the work that we are



doing right now is with hexane in methylene chloride at



about 65 to 75 percent.



     We put a standard set of analytes together, and



evaluated the method...which is typical in a method



development...backwards.  We looked at the analytical



determinative step first, GC/mass spec.  Then we looked at

-------
                             558
how efficiently we were K-D'ing all the samples, the GPC
recoveries, and the silica gel/GPC recoveries.  Then we did
the entire procedure with a fortified sludge extract.
     That extract, before we cleaned it up, contained almost
78 nig/ml of residue.  After the silica gel and the GPC, it
was down to 22.9 mg which is a 71 percent cleanup.
     This enabled us to concentrate our extracts down to 1
ml with no trouble.  The silica gel step that we used here
was 60 g of 10 percent deactivated silica gel, and we eluted
with 425 ml of methylene chloride.  The GPC was done with
the SX-3 bio-beads, and eluted with  75 percent hexane:25
percent methylene chloride.
     These are some of the results we got for what we felt
would be a representative batch of analytes.   We were
pleasantly surprised with the amount of recovery that we got
for most compounds.
     There is a problem with isophorone and another with
benzyl alcohol.  It turns out the problem is coming from the
silica gel step.  We think we can solve that problem.  So,
presently we are evaluating this procedure on a large number
of analytes and with a number of different types of sludges.
We hope to have this method completed by late summer, and
available for anybody who would like to try it.
     There is one question which might come to mind.  What
happens to the GPC calibration when you add this much

-------
                             559



hexane?  Well, this picture here is a typical UV trace of



what happens when you have pure methylene chloride as the



mobile phase.



     The analytical envelope starts right before



dioctylphthalate and ends right before sulfur.   When you



put 40 percent hexane in the methylene chloride, it doesn't



change at all.  It stays exactly the same.  The sulfur is



still a useful peak.



     But when you increase the amount of hexane to 60



percent hexane and 40 percent methylene chloride, the



analytic envelope spreads out, and the sulfur is



incorporated into the extract.  We haven't found any



problems with this so far.  The sludges that we have been



looking at haven't had much of a sulfur background.



However, it could be a problem, and we are working on that.



     So, both of those products should be done by fall,



optimistically, late summer and will be available in



publications and methods.



     I would be glad to answer any questions anybody may



have.

-------
                             560



                 QUESTION AND ANSWER SESSION



                              DR. WILLIAMS:  Dan Williams,



Kennesaw State College.



     Have you looked at alternate techniques such as



something like silation as opposed to methylation, something



like hexamethyldisilizane or whatever?



                              MR. EICHELBERGER:  I think



Jimmie has been doing some work on that, but we don't have



any results yet.  Yes, that is a good point.



                              MR. LEWIS:  Do you see any



changes in your recovery between the difference in your



detection source when you GPC versus RI or fluorescence



detection, or did you limit it to UV?



                              MR. EICHELBERGER:  I think all



they used was UV.  I never was really involved in either one



of these projects, but I think they used UV solely.



                              MR. FALLICK:  I am Gary



Fallick from Waters.



     A question for you on the first procedure.  Have you



also looked at conventional SPE cartridges as an alternative



to the disks?



                              MR. EICHELBERGER:  We haven't



yet, but it is a good point.  We are going to do that.



                              DR. ARMSTRONG:  I am David



Armstrong from Southern Research Institute.

-------
                             561



     Have you looked at any other solvents but hexane?  Let



me give you a little more information.



     When I was at S-CUBED, we did a study on hazardous



wastes using butylchloride.  Are you familiar with that?  I



don't know if you remember any of that work.  It turned out



we were using about...I think it was 50 percent



butylchloride, and we were getting much better recoveries in



GPC of the analytes out of hazardous wastes.  So, that is an



option you might consider.



                                   MR. EICHELBERGER:  How



about your ability to remove the extraneous interfering



materials?



                                   DR. ARMSTRONG:  It worked



pretty well.  Again, this isn't exactly the matrix you are



looking at, but it seemed to work pretty well.



                                   MR. EICHELBERGER:  That



is a good point, too.  This procedure might just work for



sludges.  Hopefully, it will work for all sludges, but at



this point, we don't know.  It is tough to try out all



sludges.



                                   MR. YOCKLOVICH:  I am



Steve Yocklovich from Burlington Research.



     Concerning 515, there was a noticeable salt effect, and



I imagine there is a big effect when there is a large amount

-------
                             562
of some of the analytes.  Will that be clearly stated in the
interferences section of the method?
     The reason I ask that is regulators in our State are
starting to use waste water methods for groundwaters and
everything.
                              MR. EICHELBERGER:  I never
thought about that.  When you are in the methods production
mode, you just can't do everything that everybody expects,
but, yes, we could put that in there.  I mean, we could
recommend no salting out, because it is detrimental to our
recoveries.
                              MR. YOCKLOVICH:  Well, salting
out and also, you know, just a clear statement in the
interferences that the people that are writing regulations
can see that it is inappropriate in certain situations.
                              MR. EICHELBERGER:  Well, you
see, at this point, this technology has not been proven for
anything but waters with no particulate matter in them.  And
if somebody wants to try to use this technology for samples
other than drinking water, they might run into big problems.
So, I don't think States are going to be able to do that.
They are not going to be able to make you use a method that
won't work.
     Of course, I don't know.  States make you do a lot of
things.

-------
                             563
(Laughter.)







any other questions?



(No response.)







Jim.
MR. FIELDING:  Are there
MR. FIELDING:  Thank you,

-------
  EXTRACTION OF CHLORINATED ACIDS FROM
GROUND WATER AND FINISHED DRINKING WATER
          USING  DISK TECHNOLOGY
       Researcher: Dr. Jimmie Hodgeson
  Environmental Monitoring Systems Laboratory
                 Cincinnati
Ul
               (513) 569-7311

-------
LIST  OF ANALYTES
        Acifluorfen
         Bentazon
        Chloramben
          2,4-D
          2,4-DB
         Dalapon
         Dicamba
  3,5-Dichlorobenzoic acid
        Dichlorprop
         DCPA-AM
         Dinoseb
     5-Hydroxydicamba
       4-Nitrophenol
     Pentachlorophenol
         Picloram
          2,4,5-T
          Silvex
Ul
CT>
Ul

-------
     METHOD 515.KCURRENTLY)
     1 Liter sample pH 12 with NaOH
   Shake 1 hour to hydrolyze derivatives
       Solvent wash with CH2 CI2
       Adjust pH 2 with H2SO4 and
    3 serial extractions with ethyl ether
          KD add MTBE + MeOH
       Methylate with diazomethane
  Determine with capillary column GC-EC
Sample prep time: 5-7 samples in 8-10 hours
Ul
01

-------
  METHOD  515.1  USING DISK
  LIQUID-SOLID EXTRACTION
        100 ml sample pH < 2

      Pretreat disk-20 mL MTBE
               -air dry 5  min.
              -20 mL methanol
             20 mL reagent water

    Sample throughput time  = 4 min.
          at 5 in. Hg vacuum

           Air dry 10 min.

Elute with two 2.5 mL 10% MeOH in MTBE

     Dry over anhydrous  Na2SO4
Sample prep time: 5-7 samples in 2 hours
  Methyiate with gaseous diazomethane

  Determine with cap column GC/ECD
U1
en

-------
      DISK LSE PRECISION AND ACCURACY
          100 ML 5 UG/L   C-18 DISK
ANALYTE
Acifiuoufen
Bentazon
Chloramben
2,4-D
2,4-DB
Dalapon
Dicamba
3,5-DiClbenzoic acid
2,4,5-T
Dichlorprop
DCPA-AM
Dinoseb
Silvex
% RECOVERY   %RSD
Disk
96
56
65
89
115
21
75
114
85
66
79
35
88
LLE
90
90
55
94
87
100
79
88
73
90
23
74
77
Disk
7
8
1
9
8
9
8
9
7
8
8
8
1
LLE
4
22
4
13
15
20
4
6
4
12
74
10
6
                               Ul
                               <7»
                               00

-------
     ALTERNATIVE METHODS OF
          METHYLATION

• BF3 - METHANOL

• SULFURIC ACID - METHANOL

• AMBERLITE RESINS - METHANOL

• IN SITU ALKYLATION ON SOLID SORBENT
 WITH NUCLEOPHILE - EXAMPLE
 CH3I + RCOOH(ADSORBED) - ROCH3 + HI
ui
cr>

-------
          SULFURIC ACID METHYLATION
Analyte

Dalapon
35DCBA
Dichlorprop
2,4-D
Chloramben
Silvex
2,4,5-T
2,4-DB
% Recovery

  117.4
  114.0
  110.8
   86.2
   84.2
   97.2
   87.0
   97.4

-------
      100 ml Fortified Reagent Water - 5ug/L
    SULFURIC ACID-METHANOL METHYLATION
Analyte

Dicamba
Acifluorofen
Pentachlorophenol
DCPA-AM
Dinoseb
Bentazon
% Recovery

   2.0
  37.0
   0.0
   0.0
   0.0
   0.0
Ul

-------
      100 ml Fortified Reagent Water - 5ug/L
         AG1-X8 Resin Bed, 0.11 mL Volume
Analyte

Acifluorfen
Bentazon
Chloramben
2,4-D
2,4 -DB
Dalapon
Dicamba
3,5-DCBA
Dichlorprop
Dinoseb
4-Nitrophenol
Pentachlorophenol
Picloram
Silvex
2,4,5-T
% Recovery

  80
  70
  53
  90
  98
  68
  75
  81
  72
  18
  51
  83
  66
  78
  77
%RSD

 7
 4
16
13
 5
13
 6
19
28
19
 1
12
11
 3
 2
to

-------
A METHOD FOR THE DETERMINATION OF ORGANICS
             IN MUNICIPAL SLUDGES

        200 ml_ sample(10 g eq. dry solids)
                   Centrifuge
                  Liquid-CLLE
             Solids-sonication-MeOH
                     - 1:1 MeOH:CH2CI2
                         -CH2CI2
         Concentrate extract to 10 mL ???

               Attempted cleanup
                    Silica gel
                     GPC
              Solid phase cartridges
Ul
»J
OJ

-------
     DEACTIVATED(10%) SILICA GEL CLEANUP
          OF CRUDE SLUDGE EXTRACTS
Sample
  B
Crude Ext.
 mg/mL

   64.4
   82.4
              59.0
Clean Ext.
  mg/mL

  59.9
  62.2
  61.9
  60.8
  71.6
  66.7
  70.0
  66.8
  47.0
  47.6
  42.0
  44.9
  47.5
% Cleanup


    7
    4
    4
    6
    13
    19
    15
    19
    20
    21
    29
    24
    19
                                                    in
  65 g SG 425 mL MeCI- 10mL

-------
      Comparison of DSGF/GPC and GPC/DSGF with a
         Single Sludge C Crude Extract - 59 mg/mL
     DSG/GPC

1
2
3

1
2
3
mg/mL
DSG
47.0
46.7
42.0
GPC
39.7
41.0
35.8
 Total Cleanup
1        33
2        31
3        39
              res
                             1
                             2
                             3


                             1
                             2
                             3
    GPC/DSG
          mg/mL
            GPC
             42.7
             42.0
             42.0

            DSG
             38.0
             37.6
             38.3
res
 Total  Cleanup
1           36
2           36
3           35
              
-------
         Cleanup of Sludge Extracts Using Silica
               Gel and Different Bio-Beads
Silica gel - SX-3
Silica gel - SX-2
Silica gel - SX-4
SX-3 -  Silica  gel
SX-2 -  Silica  gel
SX-4 -  Silica  gel
No  improvement in  GPC or overall cleanup was observed
by  either of  the alternate stationary phases, SX-2
or  SX-4
                               01
Methylene chloride used as the mobile phase

-------
        Average Recovery and RSD From Fortified
           Sludge Extract - 100ug/mL - 5 Reps
Analyte
2-CI phenol
2-Nitrophenol
Aniline
Naphthalene
Isophorone
2-Flourobiphenyl
Acenaphthylene
Diethyl phthalate
N-Nitroso DPA
Benzyl alcohol
Dieldrin
4,4'-DDT
Benzo(a)pyrene
Benzo(b)fluoranthene
2,6-Dinitrotoluene
Hexachlorobenzene
%rec

 75.2
 74.6
 23.6
 73.8
  4.9
 93.2
 83.2
 60.6
 92.6
  5.6
 88.5
 70.2
 89.0
 87.8
 80.6
 78.5
RSD

 9.6
 7.2
25.0
 8.4
36.7
 1.6
 3.6
 5.6
 8.9
32.1
14.0
 9.1
 6.9
 1.8
18.3
 5.3
GPC: 60g SX3 75% Hex 25% MeCI

-------
      SAMPLE AND ANALYTICAL CONDITIONS
Each analytical  step individually evaluated
GC/MS KD  GPC  DSGF/GPC  FORTIFIED  EXTRACT
Residue before  cleanup       77.9  mg/mL
Residue after GPC cleanup    23.6  mg/mL (66%)
Residue after DSGF/GPC cleanup  22.9 mg/mL (71%)
All extracts KDed to  1 ml before analysis
Ul
^]
CO
DSGF: 60g  10% deactivated silica  gel
     425  mL MeCI elution  solvent
GPC: SX-3  bio beads
     75% hexane  25% MeCI mobile phase

-------
                         579
             GPC CHROMATOGRAMS   (UV)

                  Mobile Phase  - MeCl2
              Stationary Phase  - SX-3
                                           Analyte
                                           Cocktail
                                           GPC Calibration
                                           Solution
                 139 inL
                                   225 mL
                         Analytical
                         Fraction
A.  Polystyrene
B.  Corn Oil
C.  Dioctylphthalate
D.  Sulfur
                           MW = 280,000

-------
                              580
             SLUDGE CLEANUV BY OFC
                                                    Analvtes
                                        243 mL
                        Analytical
                          Fraction
                                                    GPC STD
                                                    Sample
Figure
GPC Chromatograms of Analytes, GPC Standard and Sample

Stationary Phase » SX3
Mobile Phase • 40% Hexane and 60% Methylene Chloride v/v
+ • Start of Collection
* » Stop of Collection
A - Polystyrene        C - Oioctylphthalate
B » Corn Oil           n . Sulfur

-------
                      581
                 S&UDQC CLEANUP BT OFC
                                   Analytes
                    Analytical
                     Fraction
Figure      GPC  Chromatograms  of  Analytes, GPC Standard and Sample

            Stationary Phase » SX3
            Mobile Phase -  60« Hexane and 40% Methylene Chloride V/V
            •*>  •  Start of Collection
            *  •  Stop of Collection
            A  »  Polystyrene        C •  Oioctylphthalate
            B  *  Corn Oil           D •  Sulfur

-------
 COMPOUND EXTRACTION FROM REAGENT WATER

                  USING DISKS
120
100
   % RECOVERY
     24D
                                                    t_n
                                                    oo
                                                    to
Chloramben
Silvex    245T

 COMPOUND
24DB
Bentazon
               C18
            C8
          C8 METAL

-------
         Salted vs. Unsalted 15% (NH4)2SO4
          Reagent Water  C18 Disks  5ug/L
   Mean ug/L recovered
     Dalapon  Chloramben Dinoseb   Dichlop  Bentazon   245-T
                     Salted
Unsalted
                                                           01
                                                           GO
                                                           GO
100 mL sample

-------
r
                                       584
                                        MR. FIELDING:  Our next
          speaker is Susan Richardson of  Environmental Research
          Laboratory, USEPA at Athens/ who will talk about application
          of multispectral techniques to  the identification of
          aldehydes in a combined sewer overflow.

-------
                             585



                                   MS. RICHARDSON:  My talk



today will cover the application of multispectral techniques



which are a combination of mass spectral and infrared



techniques for the identification of organic compounds in



environmental samples.  Specifically, I will focus on the



identification of straight-chain aldehydes in a combined



sewer overflow sample.



     This work has recently appeared in an EPA report, and I



have listed the report number here for anyone who is



interested in reading more about this work.  I also have



some flyers on a back table with more information about how



to obtain this report.



     The use of multispectral techniques for identifying



organic compounds in environmental samples has been a major



part of what I have been involved with at EPA's



Environmental Research Laboratory in Athens, Georgia.



     The multispectral techniques that were used for this



specific study, the identification of the aldehydes, are



shown in the red blocks.  Each technique incorporated the



use of gas chromatography with it.



     The techniques that we used for this study were, first



of all, low resolution electron-impact mass spectrometry,



EI-MS; high resolution EI-MS; low resolution chemical



ionization mass spectrometry^ CI-MS; high resolution CI-MS;



and Fourier transform infrared spectroscopy.

-------
                             586
     The techniques shown below that you can probably barely
see in the green block are fast atom bombardment mass
spectrometry and thermospray LC/mass spectrometry.  These
techniques are currently available on our instrumentation,
but they weren't used for the study that I will talk about
today.
     The techniques shown at the very bottom are currently
being developed, and we would like to incorporate those
techniques into our multispectral techniques analysis
program as they are developed.  Those are supercritical
fluid chromatography coupled with nuclear magnetic resonance
spectroscopy and supercritical fluid chromatography coupled
with infrared spectroscopy.
     The instrument that was used for the mass spectral work
is shown on this slide, and it was a VG 70SEQ High
Resolution Hybrid mass spectrometer.  It has a double
focusing sector composed of electrostatic analyzer plates
and an. electromagnet.  This double focusing sector allows us
to do the high resolution work that I will talk about later.
It also has, in tandem, a quadrupole analyzer which allows
us to do MS/MS work as well.
     I want to mention that I was primarily involved with
the mass spectral work, as was John McGuire and Al Thruston
who are also listed on your program.  Tim Collette was

-------
                             587



responsible for the infrared work that I will talk about



later.



     The purpose of this slide is to emphasize just how few



compounds are typically targeted by common analysis methods.



I have shown here specifically EPA Method 1625, which I am



sure most of you are familiar with.  It uses low resolution



electron impact mass spectrometry to target less than 200



analytes.



     Just this year, Chemical Abstract Services has gone



over the 10 million mark for the total number of chemicals



it has given a registry number to.  So, it is clear that we



are typically- only targeting a very small percentage of the



total amount of compounds that could be found in



environmental samples.



     Part of my laboratory's mission since the Consent



Decree of 1976 has been to analyze environmental samples for



compounds other than those targeted by current EPA methods.



     There are three main obstacles to obtaining



identifications of organic compounds in environmental



samples using common analysis methods that generally involve



an extraction of the sample into organic solvent followed by



low resolution GC/MS using El conditions and, finally,



library data base matching.



     First of all, since the sample is generally extracted



into an organic solvent such as methylene chloride, there

-------
                             588
will be compounds that will not be extracted and will be
missed at this stage in the identification process.
Compounds that are highly water soluble would fall into this
category.
     Of those compounds that are extracted, there will be
those for which GC/MS is not applicable.  High molecular
weight compounds or thermally labile compounds would fall
into this category.
     Of the compounds which are GC/MS applicable, there will
be those that won't have spectra in the library data base,
inhibiting a structural assignment by library data base
matching.
     Part of my work is involved in using multispectral
techniques to, first of all, identify those compounds which
do not show spectra in the library data bases and then enter
that spectrum into a library data base to make the next
identification a little easier.  We also use multispectral
techniques to add confidence in the structural assignment
for compounds which may have an entry in a library data
base.
     Many times, particular compounds will show similar
library fits from library data base matchings.  I am sure
all of you are aware of that, and there is not always a
clear choice available.  So, using these multispectral
•techniques which complement each other, we are able to

-------
                             589



obtain very precise identifications on compounds which would



not have been possible just using low resolution electron-



impact mass spectrometry and library data base matching.



     As I mentioned earlier, this sample was taken from a



combined sewer overflow, and it was extracted into methylene



chloride.  Bill Telliard of ITD was responsible for the



program which provided this sample for us, and Jim King of



the Sample Control Center was responsible for storing the



sample for us.



     This slide shows the GC/MS chromatogram obtained under



low resolution El conditions.  Among the many compounds that



we identified in this sample, the peaks labeled 1 through 8



represent the eight straight-chain aldehydes that we



identified in the sample.



     There were four saturated aldehydes:  n-hexanal, n-



heptanal, n-nonanal, and n-octanal; and four unsaturated



aldehydes:  2-heptenal, 2-octenal, 2-decenal, and 2-



undecenal.  The low resolution El data led us to believe



that we could possibly have these assignments that I have



shown you.  However, in this case, there were many similar



library fits, and there was no clear choice available from



library data base matching.



     So, at this point in the identification process, these



assignments were only tentative.

-------
                             590
     Also/ for those four unsaturated aldehydes, there were
typically only two possible isomer choices in the library.
For instance, for the compound  2-heptenal, 2-heptenal was in
the library and 4-heptenal was  in the library, but other
possible isomer choices such as 3-heptenal, 5-heptenal, and
6-heptenal were not.
     So, we could not be sure of a correct structural
assignment from library data base matching.
     Part of the reason that the low resolution El data was
so inconclusive was that there was very little molecular
weight information present in the El spectra.  I have listed
here the relative abundances of the molecular ions obtained
under El conditions for the aldehydes listed at the left.
You can see that most of these aldehydes show extremely
small molecular ions.
     Shown here are n-hexanal at 0.1 relative abundance; for
n-heptanal, 0.2; n-octanal, 0.1; for 2-octenal, 0.3.  The
one exception was 2-heptenal which did show a molecular ion
at a reasonable relative abundance.  The three aldehydes
listed at the bottom, n-nonanal, 2-decenal, and 2-undecenal,
did not show a molecular ion at all.
     So, in general, the low resolution El data was void of
molecular weight information.
     This slide shows an example of one of the El spectra
obtained, in this case, for nonanal which has a molecular

-------
                             591



weight of 142.  You can see that there are no ions in this



molecular weight region.



     Generally, the highest mass ion present for most of



these aldehydes in the El spectra was due to the loss of



water from the parent compound, shown here at m/z 124 for



nonanal.



     In order to determine the molecular ion for these



aldehydes, first of all, low resolution chemical ionzation



mass spectrometry was used.  The reason low resolution was



used before high resolution was to prevent the interference



of a calibration compound such as pfk that would have been



necessary for the high resolution experiment.



     So, we first of all determined to a nominal mass what



the molecular ion was for each of these compounds using low



resolution CI, and then we used high resolution CI to



determine the accurate mass.



     I have listed here the observed masses obtained



experimentally for the (M+H) ions under high resolution CI



conditions.  You can see that these observed masses do



compare very favorably with the calculated masses based on



the assignments given at the left.



     So, this high resolution CI data did support our



structural assignments.  However, we haven't yet shown



clearly where the location of the carbon-carbon double bond

-------
                             592
is for the unsaturated aldehydes, 2-heptenal, 2-octenal/ 2-
decenal, and 2-undecenal.
     In order to determine, the  location of that carbon-
carbon double bond, we used high resolution electron impact
mass spectrometry.  I am choosing to use an example here, in
this case, 2-octenal, to show you how we used high
resolution El mass spectrometry to determine the location of
the double bond, but we used the same procedure for each of
the other three unsaturated aldehydes as well.
     The premise that we used was this:  carbon-carbon
double bonds will rarely fragment across the carbon-carbon
double bond under El conditions.  So, if we could show we
had a fragment ion owing to cleavage at each of these other
carbon-carbon bonds along the hydrocarbon chain and if we
see an absence of an ion owing  to cleavage at this
particular location, then we could determine that this must
be the location of the carbon-carbon double bond.
     The reason that high resolution was necessary for this
work over the corresponding low resolution experiment was
that under low resolution El conditions, many of the
fragment ions obtained could be represented by more than one
possible empirical formula and, thus, by more than one
particular site of cleavage.  So, in order to determine that
site of cleavage accurately, we did need high resolution El
mass spectrometry.

-------
                             593



     This slide shows an example of that.  Using low



resolution El mass spectrometry conditions, we obtained an



ion at m/z 83 for the compound that I showed you on the



previous slide, 2-octenal.  This ion at 83 could be



represented either by C6HU which corresponds to a particular



site of cleavage along the hydrocarbon chain and also a



possible location for the carbon-carbon double bond, or that



ion at 83 could be represented by C5H7O corresponding to yet



another particular site of cleavage along the hydrocarbon



chain and another possible location for the carbon-carbon



double bond.



     High resolution electron impact mass spectrometry



provided us with the accurate mass that determined that the



CgHyO assignment was the correct assignment for the ion at



83.  Thus, we can assign that specific site of cleavage at



the location shown here.



     We used the same methodology for each of the other



fragment ions obtained under high resolution El conditions,



and we showed that we did have cleavage of each of these



carbon-carbon bonds along the hydrocarbon chain, but we did



see an absence of an ion owing to cleavage at this location.



So, we concluded that the double bond must be located at



this particular position, which is at the carbon-2 position.



     Incidentally, this particular location for the carbon-



carbon double bond allows it to be in conjugation with the

-------
                             594
aldehyde carbonyl group, making it a more thermodynamically
stable compound, something you might expect for a compound
found in the environment.
     Keep in mind that this double bond is in conjugation
with that aldehyde carbonyl group.  That will become
important in the next slide.
     This slide shows the infrared spectrum of 2-octenal.
Again, I am choosing to use 2-octenal to show you how we
used infrared spectroscopy to identify all of these
aldehydes.  I use this as one particular example.
     Infrared spectroscopy was a nice complement to mass
spectrometry, because it, first of all, confirmed some
structural assignments that we had made using mass
spectrometry, and it also provided new information that mass
spectrometry was not able to provide.
     First of all, the peaks shown here which are due to the
stretching fundamental of the hydrogen atom attached to the
carbonyl carbon and also due to a corresponding overtone
peak, those peaks, along with the peak due to the stretching
of the carbonyl shown here at 1715 cm"1 are clear evidence
for an aldehyde group
     So, first of all, IR spectroscopy confirmed the
existence of the aldehyde group for all eight straight-chain
aldehydes we identified.  Secondly, infrared spectroscopy

-------
                             595



confirmed the existence of the carbon-carbon double bond for



the four unsaturated aldehydes.



     The peaks shown here at 1634 cm"1 is clear evidence of



a carbon-carbon double bond.



     The third piece of information that infrared



spectroscopy provided was the location of that carbon-carbon



double bond along the hydrocarbon chain for those



unsaturated aldehydes.  If this carbon-carbon double bond



had not been in conjugation with that carbonyl group, this



particular frequency shown here at 1634 cm"1 would have been



shifted to a much higher frequency, to about 1650 cm"1.



Also, the carbonyl frequency would have been shifted to



about 1742 cm"1.



     So, because these frequencies are located where they



are at 1634 and 1715 cm"1, that is clear evidence for



conjugation of that carbon-carbon double bond with the



aldehyde carbonyl group, thus confirming that the double



bond is located at carbon-2, as high resolution electron-



impact mass spectrometry had suggested.



     Finally, infrared spectroscopy allowed an exact



isomeric determination for the four unsaturated aldehydes,



that is, whether we had a cis or a trans double bond.  This



was something that mass spectrometry was not able to



provide.

-------
                             596



     The peak shown here at 976 cm"1 is clear evidence for a



trans double bond.  So, we did determine that we had a trans



double bond for each of those four unsaturated aldehydes.



     To summarize, using the multispectral techniques, which



were low and high resolution electron-impact mass



spectrometry, low and high resolution chemical ionization



mass spectrometry, and Fourier transform infrared



spectroscopy, we were able to precisely identify these



compounds that would not have been able to be identified



this way using the common analysis methods of low resolution



electron-impact mass spectrometry and library data base



matching.



     First of all, low resolution chemical ionization mass



spectrometry allowed us to accurately determine the



molecular ion since it was either very small or wasn't



present at all in the El spectra for most of these



aldehydes.



     Secondly, high resolution CI-MS provided us the



accurate mass which allowed us to determine the exact



empirical formula of the molecular ion.



     Thirdly, infrared spectroscopy confirmed the existence



of the aldehyde group for all eight aldehydes.



     Fourthly, both high resolution El mass spectrometry and



infrared spectroscopy together determined the position of



the carbon-carbon double bond for the unsaturated aldehydes.

-------
                             597



     Finally, infrared spectroscopy allowed an exact



isomeric determination, that is, whether we had a cis or a



trans double bond.  We did determine that we had a trans



double bond, and this, again, was something that mass



spectrometry was not able to provide.



     You can see how nicely these techniques complemented



each other and how they worked very well together to allow



us to make these precise identifications.



     I would like to conclude by saying that the observation



of these straight-chain aldehydes in a combined sewer



overflow sample was very unexpected.  What we have typically



found in those types of samples are primarily fatty acids



and fatty acid methyl esters.  So, the observation of these



aldehydes in that particular type of sample was unusual from



that standpoint.



     It was also unusual from the standpoint that there are



very few reports of similar straight-chain aldehydes in the



literature where they were found in the environment at all.



So, we did feel that this finding was significant.



     There are many other types of aldehydes that are



commonly observed such as benzaldehyde, acetaldehyde, and



formaldehyde, but, evidently, straight-chain aldehydes are



not that common.



     I do want to mention, however, that I have been



recently told that there have been some tentative

-------
                             598



identifications of some similar straight-chain aldehydes in



Superfund samples, but these identifications are only



tentative at this point.



     I would like to thank you for your attention.  If there



are any questions, I would be happy to try to answer them.



                              MR. FIELDING:  Does anybody



have any questions?



(No response.)



                              MR. FIELDING:  If not, we



thank you very much, Susan.

-------
Application of Multispectral Techniques
  To  the Identification of Aldehydes

   In  a Combined Sewer Overflow
                                                Ul
                                                u>
                                                IO
         EPA/600/4-90/002
    NTIS No.  PB 90 160 995/AS

-------
Low Resolution
   EI-MS
                     High Resolution
                        EI-MS
             MULTISPECTRAL TECHNIQUES
                          V
High Resolution
    CI-MS
Fast Atom Bombardment
         MS
  Thermospray LC/MS
                    SFC/NMR S SFC/IR
                        Low Resolution
                           CI-MS
                                      O
                                      O
   Infrared
Spectroscopy

-------
                                                 ELECTROMAGNET
OUTER E.SA PLATE
                                                                                                    OUAORUPOL6
                                                                                                   COLLISION CELL
QUAORUPOLE
 ANALYSER
                           1st FIELD FREE REGION GAS CELL

                           SOURCE SLIT (FIXED)
                                                                                                              MULTIPLIER STACK
                                                                                                                                01
                                                                                                                                O
               SOURCE
     GC inlet
     Solid probe inlet
     FAB/FD inlet
     Thermospray LC/MS inlet

-------
            Method 1625
       v

Target Analytes
    (<200)
                    Low Resolution
                    EI-MS
                                               O
                                               to
     V
Other Chemicals

-------
                   Aqueous Sample
                  Solvent  Extraction
           Extracted
      1. Not Extracted
                                                        01
                                                        o
                                                        CO
2. GC/MS not

  Applicable
GC/MS Applicable
            Spectrum in File
          3. Spectrum not in File

-------
100
 95J
 90J
 85J
 80J
 75J
 70 j
 65J
 60J
 55j
 50J
 45J
 40J
 35J
 30J
 25J
 20 j
 15j
 10 j
  5J
  Oj
     300
    5:16
                                   1.  n-hexanal
                                   2.  n-heptana!
                                   3.  2-heptenal
                                   4.  n-octanal
                                          5.  2-octenal
                                          6.  n-nonanal
                                          7.  2-decenal
                                          8.  2-undecenal
 400
7:02
500
8:47
 600
700
860
10:32    12:17    14:03
 960
15:48
 1000
17:33
SCM
BXttfi
                                                                                 01

-------
     Electron Impact Relative Abundance
             Of Molecular Ions
Compound
n-hexanal
n-heptanal
2-heptenal
n-octanal
2-octenal
n-nonanal
2-decenal
2-undecenal
Relative Abundance
     0,1
     0.2
     10
     0,1
     0.3
01
©

-------
10 <5_
95j
90J
85j
80J
75J
70j
651
60j

55J
50J
45J
40J
35J

30 j
25.
20J
15-i
10 j
:
5j
0:
4
4.























0
3
















15





1
' sfo'
57






















1
'e'o' ' ' ''









71


8











o e'o












2


9








' ' '9V '








9






5







|

NONANAL F
:-
i
'•_
Molecular Weight =* 142 :
•
No Molecular Ion Present :
-
8 ;•

|-
'-
'-
'.


„
-
'r
114 r

124 "
, 1-
1 , :
160 iid 120 iid 140 isd i£d 170 M/Z
en
o
cr>

-------
High Resolution Cl Accurate Masses
Compound
n-hexanal
n-heptanal
2-heptenal
n-octanal
2-octenal
n-nonanal
2-decenal
2-undecenal
Empirical
Formula
C6.H13.O
C7.H15.0
C7.H13.0
C8.H17.0
C8.H15.0
C9.H19.0
C10.H19.0
C11.H21.O
Observed
Mass
101.096
115.111
113.097
129.130
127.113
143.143
155.144
169.158
Calculated
Mass
101.097
115.112
113.097
129.128
127.112
143.144
155.144
169.159

-------
            608
            0=0
 CO



 CD
•*-*

 O

O
 i

CM


t

t

t

t

t

o
II
o
CM
O
i

-------
 Low Resolution EI-MS
     Ion at m/z 83
                                         o
                                         vo
C6.H11   or
C5.H7.O
       By  High Resolution EI-MS

-------
               610
 05
 C
 CD
•*—»
 O
O
 i
CM
X
o
 II
X
o
  CM
x
o
 'CM
cof
oo 1

o
CM
X
O
1
CM
                 O

                   CO
                 X
                 O

-------
   6-
o
o


X

CD
O
c
o
_Q

O
CO
   0-
                    2-Octena
                   (O)C-H
 1715


,C=O




 conjugated





 1634


 C=C
                                                    trans
      3500
3000      2500      2000

     Wavenumbers (cm—I)
 1500
1000

-------
                       Summary
O  Accurate Determination
   Of Molecular Ion

O  Exact Empirical Formula
   Of Molecular Ion

O  Confirmation of
   Aldehyde Group

O  Position of Carbon-Carbon
   Double Bond

O  Exact Isomeric Determination
    (Cis or Trans Double  Bond)
   Low Resolution
   CI-MS
^  High Resolution
   CI-MS
                         to
•>  Infrared Spectroscopy
   High Resolution EI-MS
   Infrared Spectroscopy
 ->  Infrared Spectroscopy

-------
                             613



                                   MR. FIELDING:  It is time



for lunch.  Can we try to get back about 1:00 o'clock?  We



have a full afternoonf and I know some people have to leave



on early flights.  We will see you at 1:00 o'clock.

-------
                        614
             AFTERNOON SESSION
                              MR. TELLIARD:  We would like
to get going for this afternoon's session, please.
     One small announcement.  We routinely take the
proceedings down and put it out in a book called
"Proceedings."  As I alluded to before, last year's
proceedings are sitting waiting to be printed.  I called my
management, I called Washington, and they said yes, it is
sitting here.  The word is that we won't be able to print it
until September when we get new money.  So, when it is
available, I will mail it to you, he said.
     That also means that these proceedings that we are
taking down now will hopefully be out in September when we
get the new budget.  New money comes in in a big truck and
they dump it in the office.  We roll around in it for a day
or two and then spend it.
     So, I am sorry about this.  It is kind of embarrassing,
but the Office of Water took about a $1 million intramural
cut this year.  Apparently, we made somebody unhappy up on
the Hill, and now they have made us unhappy.  It works.
     So, because of that intramural cut, we have not been
able to print almost anything.  So, on those cards you
filled out, on some of those documents, you will get a
little notice that says "in printing" which means there is

-------
                             615
no money right now, and we will get those in September.
Those that are available we will mail out to you.
     So, our first speaker this afternoon is Joe Raia.  Joe
is going to speak about one of his favorite persons, a
robot, and the joys of doing suspended solids using a
robotic method.

-------
                             616
                              MR. RAIA:  Good afternoon.
     At Shell Development Company at the Westhollow Research
Center, I have the Environmental Wastewater Analysis group,
and in this laboratory we process a large number of samples
for the environmental research at this facility.  One of the
goals and focuses of our group is to try to automate methods
where it makes sense to do so in order to perform these
analyses as cost effectively as possible.
     The total suspended solids method was one of these
methods that we thought was a real good candidate for
automation.  It was highly repetitive, we did relatively
large numbers per year, on the order of about 2000, and it
really wasn't that much fun to do.
     We got together with our systems development
group...and the co-author up there on the slide, Al
Telfer...and began thinking about how we could automate this
procedure.  The result of that effort is what I will be
talking about this afternoon.
     Automation in the laboratory is really nothing new.  It
has been around for a number of years.  We have had
laboratory information management systems for identifying
the sample when it comes into the laboratory and to do
sample tracking.
     We can skip over the sample preparation part shown here
on the slide for a minute, and looking at the analysis side,

-------
                             617



there are all sorts of automated instrumentation with



autosamplers, continuous flow type systems,and



instrumentation which are run by microprocessors.  Then, in



the results and reports section of the analysis, we know



computerized data reduction has been around for years.



     One 'important step in this whole process has been



sample preparation.  About, I guess, 1982 or so, a company



out of Hopkinton, Massachusetts, the Zymark Corporation,



took a look at that and decided that it was a spot where we



really needed to do some work in terms of automation and



time savings efforts.  So, they have come up with their



Zymate system for primarily the sample preparation part of



the analysis procedure.



     The TSS analysis robotic syste'm that I am talking about



here in this presentation is one that measures total



suspended solids in water.  We developed the method to



follow the standard EPA protocol, Method 160.2, which



essentially is a filtration of the sample through a 1 micron



filter and then followed up by a gravimetric finish after



removal of the water at 105 degrees in an oven.



     The procedure will also allow us to measure the



volatile residue according to EPA Method 160.4 with what we



call a minimum of operator intervention.  As I go through



and describe the lab robot...and many of you here have



already seen them around for quite a few years, you will see

-------
                             618
that the hands of the robot would not be able to stick the
sample into a 550 degree furnace that is used for the
volatile residue part of the test.      So, we intervene at
that point and place the samples in a rack and manually
insert them into that type of a furnace.
     Now, in the beginning of our story here, this is a
photograph of the way we used to do suspended solids in the
lab.  We typically had our setup to do eight samples at a
time, batchwise.  At that time, we tried to automate at
least the end part of the procedure by linking the balance
to a computer so that you could start the analysis, get your
tare weight, store that weight in the data system, then
perform the filtration, the drying, the weighing,
reweighing, and then from the data system memory of the
initial and final weight, the computer would calculate the
suspended solids that were captured on the filter.
     This slide shows the system that we developed to
perform that analysis.  This is the Zymark Zymate system
that we purchased from Zymark Corporation, and this is their
controller to automatically operate the instrumentation.
     It is not shown here, but we have linked this system to
an IBM PC for report generation.  I have another schematic
that I will put up here in a second to show you the layout
of the system.

-------
                             619



     This whole apparatus is on about a 7 foot by 4 foot



table top, and the electronics for performing part of the



analysis are housed below here.



     The sample starts out in this sample rack.  Typically,



we use 100 ml of sample.  This is the oven for drying at 105



degrees C, a dessicator for conditioning before weighing in



the balance, and these items here are the stands to hold the



Gooch crucible for the filtration.



     To simulate our original earlier days with this method,



we stayed with a batch of 8 sample at a time, although we



can actually do 16 in two batches of 8.



     This next slide shows schematically what the photograph



is trying to show there.



     The feedback sensor is an item that I want to point



out.  This turned out to be the real challenge in terms of



trying to automate this procedure; the reason being, " how



to determine when the filtration had finished".  In order to



do that, we ended up using a capacitance type sensing device



which allowed you to essentially tell the liquid level in



the Gooch crucible.



     The oven and dessicator were built in our shops at



Shell in Westhollow, and they are pneumatically operated.



The balance is a Mettler balance with a pneumatic door, and



I have already mentioned the data system.

-------
                             620



     This next slide shows the step by step operation that



the robot does to perform the test.  I am going to read



through here as we go.  The sequence is as follows:



     The robot gets the Gooch crucible with filter from the



dessicator and moves it to the balance where the tare weight



is obtained.  Next, the robot takes them to the rinse



station where a small volume of water is placed on the



filter.  This helps seal the filter to the Gooch crucible.



     The robot then places them in the filter station and



backs away.  The vacuum is turned on to that filter station,



and the capacitance electronics system takes a reading of



the base capacitance.  That is the base point that will be



compared to determine when the sample is finished filtering.



     This is accomplished using the switched outputs



available to the power and event controllers as well as the



A/D input.  The output of the capacitance electronics is a 4



to 20 milliampere signal.  This is converted to a volatage



at the PEC switches using a 50 ohm resistor.  This voltage



reading is stored in the robot controller's memory for



further use.



     The robot then gets the proper sample flask from the



storage rack and inverts the flask into the crucible.  The



robot leaves the flask in a holder on top of the filter



station with the neck of the flask below the top of the



crucible.

-------
                             621



     This allows us to filter up to 100 ml of sample into a



40 ml Gooch. crucible.  You can kind of think of this as a



chicken feeder principle where the Gooch will not overflow



because of the hydrostatic pressure that won't allow the



rest of the 100 ml volume to fill until some of it filters



through.



     The robot backs away, and a second capacitance



measurement is made.  Now, if the filtration is complete,



this second reading will approach the base reading and the



vacuum will automatically be shut off for that filter



station.  If not, the robot proceeds with loading the next



sample and rechecks all loaded stations for capacitance



until filtration is complete.



     The rest of the procedure operates in exactly the same



way, and we rinse three times according to the protocol.



     The sample can sit for some time without affecting



performance as long as the vacuum is off.  We found that if



the filter gets too dry, some samples seem to get a



waterproof like coating that may hinder a rinse procedure



that follows.  Conversely, the filter must not be too wet,



particularly when the crucible is placed in the oven.



     For this reason, the program pauses after the final



rinse and before the crucible is loaded into the oven while



vacuum is applied to all occupied filter stations.

-------
                             622
     After drying at 105 degrees, the crucible is moved to
the dessicator to condition and then weigh.  Constant weight
is verified by redrying the crucible for a half an hour and
then reweighing.
     If the volatiles are measured, the crucible and dry
solids are moved by the analyst to a 550 degree furnace, and
after 45 minutes, the crucibles are allowed to cool to
ambient conditions and then placed in a dessicator before
being weighed as before.
     What I have here is a series of photographs to show
some of the steps that the robot is taking to carry out the
procedure that I have just described.  This  one shows
getting the crucible out of the dessicator at the beginning
of the analysis before placing it in the balance.
     The door of the balance is activated and pneumatically
opened, and the crucible is placed on the pan.  The arm
moves back, and the door is closed, and a weighing is made.
The average of 10 weighings is done.
     The arm goes back to get the crucible and moves it to
the filtration station.  The arm then goes to get the sample
of water that is in a 100 ml volumetric flask and moves it
to pour the sample into the Gooch crucible.
     Programming that step was a little tricky to make sure
that no water was dripping during the pouring process so
that we didn't have any loss of sample.  You can see how the

-------
                             623



filtration stations were designed to hold the flask in place



inverted as the water passed from the flask into the Gooch



crucible, and this is the place where the 100 ml of sample



is poured into a 40 ml volumetric Gooch crucible without



overflowing.



     The neck of the flask is just below the top part of the



Gooch crucible.  The filter has already certainly been put



in place early on in the procedure.



     Then the arm goes to get the flask and moves it over to



a rinse station where the rest of the sample is



quantitatively transferred back onto the filter.



     This shot was taken with a pen light put on the robotic



arm and a time delay photograph to show all the different



motions that are involved in carrying out this procedure by



the robot to do eight samples.



     This next slide shows the linearity with sample volume.



It turns out that you have to select samples that are real



high in solids.  You use a reduced volume of sample for



those type samples, and we wanted to see what sort of



linearity we were getting down in the lower sample volume



range and felt comfortable that it was linear down in that



range.



     This slide shows comparison data using the robotic and



manual methods for a couple of samples that were generated



in our research in the Environmental Science Department, in

-------
                             624
benchscale biotreaters, at two different TSS levels.  The
mean standard deviation and coefficient of variation at
these two levels were essentially about the same.
     We feel that we don't see any bias in the robotic
versus the manual method.  At the 100 mg/1 level, there
seemed to be a lower robotic recovery for this particular
set of data than for the manual, but that did not turn out
to be the case in later results comparison work.
     This data here shows percent recovery using samples
provided to us by the Analytical Products Group out of
Belpre, Ohio.  This vendor will supply proficiency testing
type samples to various laboratories who participate.
     So, we had a comparison here of about 40 different labs
around the country who were also getting the TSS proficiency
sample, and we were able to compare it by the robotic
procedure and the manual method performed in our laboratory,
and compare it to the manual procedure in these other
laboratories... and we felt satisfied with this recovery
comparison.
     The benefits of lab automation, as I conclude here, are
first the obvious one of unattended operation.  The sample
can be run at night or on weekends.
     Reduced time for analysis.  We have been operating this
system for about three years in our laboratory and have been
able to demonstrate a cost savings over that period of time.

-------
                             625
     Typically, you can get improved precision.  With the
TSS system however, I think the nature of the sampling
itself is such that we really don't see any benefit of
improved precision over the manual method if the manual
method is done real carefully.
     Finally, there is what we call gained "new" time for
R&D that we end up with for doing more interesting and
challenging tasks in the environmental area.  A lot of work
needs to be done.
     I believe that is it.  I will take any questions, if
you have any.

-------
                             626



                  QUESTION AND ANSWER SESSION



                              MR. TELLIARD:  I have a



question.



     When you take the sample and you have a lot of solids,



do you shake it and then take 100 ml of that, or do you just



get it from...when you get to your 100 ml volume? What do



you start with/ 2 liters or a liter that comes out of a



sampler or something like that?



                              MR. RAIA:  That is right.  So,



you take that sample, and you shake it initially to get that



aliquot.



                              MR. TELLIARD:  And then you



put that in the volumetric and transfer it to the system?



                              MR. RAIA:  Right.



                              MR. TELLIARD:  Any problems



with solids like sticking to the walls?



                              MR. RAIA:  We haven't seen



any, but that is something to be concerned about.



                              MR. TELLIARD:  Any other



questions as I am up here hogging the microphone?



                              MR. COLLAMORE:  Martin



Collamore, City of Tacoma.



     How do you handle larger sample sizes than 100 ml?



                              MR. RAIA:  We don't.  We don't



run a sample size larger than 100 ml.

-------
                             627



                                   MR. COLLAMORE:  But you



can handle lower samples?  You can handle smaller samples?



                                   MR. RAIA:  Yes, we can



handle smaller samples.  We do handle samples more



concentrated by diluting down if we need to, but in terms of



total volume, we have only used up to 100 ml.  You are



limited to some extend there by the weight of that flask



with the water.  We really have not gone to larger sample



sizes, and I don't know what maximum volume we could end up



with, but this particular setup has been able to handle the



type of samples that we typically deal with.



     Some of our manufacturing locations also have a similar



type of robotic system for doing TSS, and they are using a



100 ml sample volume also.



                                   MR. TELLIARD:  Thank you



very much, Joe.

-------
                         628
A LABORATORY ROBOTIC METHOD FOR THE AUTOMATED DETERMINATION

  OF TOTAL SUSPENDED SOLIDS IN ENVIRONMENTAL WATER SAMPLES


                             by
                  Joe C. Raia and AT Telfer
                  Shell Development Company
                       Houston, Texas
                  For Presentation at the

               The Annual U.S. EPA Conference
                             on
          Analysis of Pollutants in the Environment
                     Norfolk, Virginia
                       May 9-10, 1990

-------
                                      629


       A LABORATORY ROBOTIC METHOD FOR THE AUTOMATED DETERMINATION

         OF TOTAL SUSPENDED SOLIDS IN ENVIRONMENTAL WATER SAMPLES


                                  by
                       Joe C.  Raia and Al  Telfer
                       Shell  Development Company
                            Houston, Texas
                                 ABSTRACT
This paper presents a new laboratory robotic procedure which automates the
standard method  for  the  determination of Total Suspended  Solids  (TSS)  in
environmental water  and  wastewater  samples.   The method  also  determines
Volatile  Total   Suspended  Solids  (VTSS)  with  a  minimum  of  operator
intervention.  The automated equipment, robotic procedure,  and results are
presented.   The  benefits  of  automation  in   the  environmental  analysis
laboratory are discussed.
                           ACKNOWLEDGEMENT
The authors wish to acknowledge P. J. Drymala and  R. A. Balderas of Shell
Development Company for their assistance in this project.

-------
                                630

         LABORATORY ROBOTIC METHOD FOR THE AUTOMATED DETERMINATION

         OF TOTAL SUSPENDED SOLIDS IN ENVIRONMENTAL WATER SAMPLES

                                  by

                      Joe C. Raia and Al Telfer
                      Shell Development Company
                           Houston, Texas
                              INTRODUCTION
Automation has been  used  in  the analytical  chemistry laboratory  for many
years.  This  has  primarily  included  microprocessor  controlled analytical
instrumentation with dedicated  autosamplers,  continuous  flow systems, and
computerized  data collection,  calculation,  and  report  generation.   In
recent  years, laboratory automation  has  been  extended  by  the use  of
robotics, combined with programmable computers, to new tasks which include
sample preparation and even entire analytical determinations.

This article presents a laboratory robotic system for the
determination of total suspended solids (TSS) in environmental
water samples.  The method adheres to the standard protocol
that is specified for the manual procedure (US EPA Method 160.2) (1).
It also performs Volatile Total Suspended Solids (VTSS) with a minimum
of operator intervention.  The robotic TSS procedure required the
development of new robot-friendly modules and sensors not yet commercially
available. These components and the automated analysis
method are described.  Results are presented which compare the robotic and
manual procedures.  The method has been validated and is
currently in routine operation in our environmental analysis laboratories.
a)  A portion of this paper is to be published in American
    Environmental Laboratory

-------
                                      631
                THE TSS ANALYSIS LABORATORY ROBOTIC SYSTEM
The TSS  analysis  laboratory robotic  system  and associated  equipment  are
shown  in  Figures  1  and  2.  The  system  consists  of  the following  major
components:
1.   A laboratory robot and controller (Zymark Corporation,  Zymate I
      upgraded to a Zymate II system)
2.   Three Zymate Power and Event Controllers (PEC)
3.   A printer for the Zymate controller
4.   A computer (IBM-XT) and printer with programms for communication
     with the Zymate, data calculation and report generation
5.   A  balance  (Mettler AE163)  with electronic  interface  and  a  remote
     pneumatic controlled door
6.   An Oven, thermocouple controlled with pneumatic controlled door
7.   A Desiccator with a remote pneumatic controlled door
8.   Eight filtration stations
9.   Capacitance sensing devices for eight positions
10.  A pump and reservoir system to supply rinse water
11.  A rinse dispenser station with pneumatic valve
12.  A storage rack to hold 16 polypropylene, volumetric flasks
     (100ml) used for the samples
13.  A  custom built  table  top  support  frame with  castors and  leveling
     jacks to hold the robot table and hardware
14.  Various actuators, sensors, and controls required by the system

-------
                                632
One of the challenges  encountered with  the  design  of the robot system was
the problem  of  how  to sense  that  filtration  had been  completed.   This
problem  also exists  for  sensing  rinse completion.  These problems  were
solved by detecting the  amount of water  in the  Gooch crucible  using  a
commercially  available capacitance  electronic  package  usually used  for
level  sensing in  tanks and  vessels.  The holder for  the Gooch  crucible in
the  filter  ^station   is  made  of  two  stainless   steel  parts.  They  are
electrically  separated by  a  rubber  gasket,  and  form two plates  of  a
capacitance  measurment system.   If  a  small  amount  of water  is  present
within the Gooch  crucible,  a  significant change in the capacitance output
signal occurs.

In operation, the sequence  of events  is as follows:  The  robot gets  the
Gooch crucible with filter from the desiccator and moves it to the balance
where the tare  weight  is  obtained.    Next,  the robot  takes  them to  the
rinse station, where  a small  (2-3 ml) quantity of water  is placed on  the
filter.    This helps  seal  the  filter to  the  crucible.   The   robot  then
places them  into  the filter  station,  and  backs  away.    The vacuum  is
turned on  to that filter  station, and  the  capacitance electronics system
takes a  reading  of the base  capacitance.  This  is accomplished  using  the
switched outputs  available at  the Power and event  controllers,  as  well  as
the A/D  input. The output of  the  capacitance  electronics  system is a 4-20
milliampere  signal.  This  is  converted  to  a  voltage at the PEC  switches
using a  50  ohm   resistor.  This  voltage reading  is  stored in the  robot
controllers memory for further use.   The robot then gets the proper sample
flask from the storage rack and inverts the flask  into the crucible.   The
robot leaves the  flask in a holder on  top of the filter station,  with  the
neck of  the  flask below the top of the  rucible.  This allows  up to 100 ml
of sample  to be  filtered using  a crucible of  about 40 ml capacity.   It
will  not  overflow since the neck  of the flask  is  sealed  below the liquid
surface.    The robot  backs  away  and  a second  capacitance measurment  is
made.  If the filtration is complete,  this second reading will  approach the
base reading, and the  vacuum will be  shut off for  that filter  station.  If
not,   the robot  proceeds  with loading  the  next sample and rechecks  all
loaded   stations  until  filtration  is  complete.    The  rinse  procedure
operates  in  exactly  the  same  way.   The  samples  can  sit  for some  time
without  affecting performance as long as the vacuum is off.  If the filter
gets too dry, some samples seem to get a water-proof like coating that may
hinder rinse procedure.   Conversely,  the  filter must not  be too  wet,
particularly when the  crucible  is placed in the oven.  For this reason  the
program  pauses  after the final  rinse and  before  the  crucible  is loaded
into the oven,  while vacuum  is applied to  all  occupied  filter stations.
After drying at  105  C,   the  crucible is moved  to  the desiccator  to
condition  and then weighed. Constant  weight  is verified  by redrying  the
crucible for a half hour and  reweighing. If volatile suspended solids  are
measured, the crucibles and dried solids are moved by the analyst to a 550
C furnace.  After 45 minutes,  the crucibles are allowed to cool ambiently,
and then placed in the desiccator before being weighed as before.

The Zymate II system  is  interfaced  with an IBM-XT-PC  to  perform  the  data
calculation and report generation  of the TSS and VTSS results.

-------
                                      633
                          RESULTS
The robotic procedure has been validated for unattended operation, and for
precision  and  accuracy.    Comparison   data  of  the  robotic  and  manual
procedures are shown in Tables 1 and 2.   Results  in  Table 1  show that the
precision of  both  methods is good  and  the values are  comparable for the
biotreater aeration basin samples tested.  Any precision  gains offered by
robotics  are  likely  obscured by  variability in  the  sampling  of  these
wastewaters.  The robotic procedure resulted in a  7% lower mean TSS value
than did  the manual  method for  the  aeration basin samples  tested.   No
consistent bias in the data, however,  has been found for either procedure.
Results  in  Table   2  compare  the  recovery  of  the robotic  and  manual
procedures for TSS  in  standard samples prepared at  various  concentration
levels for  round-robin  testing.   Data  for  an  in-house  control  are also
given.   The  robotic  and  manual  procedures  both  showed   comparably good
recovery for these  round-robin samples. The  robotic  method values were in
the 94% -  107% range, and the manual  method values were  in  the 83% - 97%
range.   The mean recovery of the fifty labs which participated in this APG
sample set were in the   91% - 95% range.   Linearity of TSS with volume of
sample filtered  is  shown in  Figure 3  for the  robot and  manual  methods.
Typically, the volume of  sample taken for analysis  is such that it can be
filtered in a reasonable time period and yet be enough to yield sufficient
solids  for accurate weighing.

Analyst time  required for  the robotic procedure  is less  than one-third
that for  the manual  procedure.  The capital  investment required  for the
robotic TSS  method  has  been  recovered  in  about  the  first 200  days  of
operation with  the current  sample  throughput demand.    The  analyst has
welcomed the automated robotic TSS procedure.   The responsibility
for operating the robotic instrumentation is  viewed  as  a  more interesting
task than the manual procedure, and  gained  "new"  time  can now be focused
on more creative method development and special  problem solving challenges
in the environmental analysis area.

REFERENCES

1) Methods  for  Chemical Analysis of  Water and Wastes,  EPA-600/4-79-020,
Revised March 1983.

-------
                             634
TABLE 1. COMPARISON OF TSS RESULTS BY  ROBOTIC AND MANUAL METHODS

SAMPLE: BENCH SCALE
METHOD: ROBOTIC TSS
MG/L
10670.
10130.
11631.
11219.
10820.
9981.
10761.
10200.
MEAN 10676.
S.D. 566.
C.V.(%) 5.3

BIOTREATOR
MANUAL TSS
MG/L
9670.
10590.
10250.
11130.
10540.
10320.
11530.
10640.
10584.
562.
5.3

AERATION
ROBOTIC TSS
MG/L
122.4
115.2
125.6
123.2
118.4
122.8
116.8
141.2
123.2
8.1
6.6

BASIN
MANUAL TSS
MG/L
139.2
152.0
127.6
130.0
128.8
128.4
126.4
124.8
132.2
9.1
6.9

-------
                                     635
TABLE 2.  RECOVERY RESULTS FOR TSS BY ROBOTIC AND MANUAL METHODS


SAMPLE

*
APG # 1
*
APG # 2
*
APG # 3
*
APG # 4
*
APG # 5
,*
APG # 6
**
KAOLIN CONTROL


TRUE TSS
MG/L

36.5

52.7

77.0

287.0

316.8

498.6

100.


ROBOTIC
% RECOVERY

107.

94.9

93.5

107.

98.7

94.9

104.

***
MANUAL MANUAL APG
% RECOVERY MEAN % REC.

83.0 92.5

96.8 90.7

92.2 90.7

94.8 94.2

97.2 94.6

96.3 94.2

95.6
*   Proficiency Environmental Testing Program Samples: prepared and
    provided by Analytical Products Group(APG), Inc., Belpre, Ohio

**  In-house control sample

*** Mean Recovery of 50 labs participating in the APG Program

-------
             636
1.
2.
3.
4.
5.
      Figure 1.  Robotic TSS Equipment

   Zymark Robotic Arm
   Balance
   Desiccator
   Oven
   Flask Rack
6.  Filtration Stations with Capacitance Sensors
7.  Zymate Controller
8.  Capacitance Sensors Electronics
9.  Pneumatic Valves for Vacuum and Pressure
                                         09941

-------
                                       637
5
'5
!•

i
                        2nd Hand   .  .(%
                     (Special Fingers)^1
                                            D
               n
              o
                 Rinse
                Station
                                        D
                                Filter Stations (8)
                                Robot
                            Microprocessor
                    Sample Input
              Printer
                                 Figure 2. TSS Robot
                  24


                  20


                  16
          Volume,
             ml
 D Y Actual
—Y Predicted
                                   20
                      40
              mg Weighed
60
                           Figure 3. TSS Linearity with Sample Volume
                           mg Weighed vs. Volume, Manual and Robot

-------
A  Laboratory  Robotic Method for
 the Automated Determination of
    Total Suspended Solids in
  Environmental Water Samples
                                       CA)
                                       00
         Joe C. Raia and Al Telfer

          She!! Development Company
           esthollow Research Center
             Houston Texas

-------
               in  the
Identification
Preparation
  Analysis
  Results
  Report
     LIMS
   Robotics
 Instrumentation
Microprocessors
 Computerized
Data Reduction
U)
IO

-------
     The TSS Analysis
Laboratory Robotic System


 • Measures  Total Suspended
   Solids in Water

 • Follows the Standard Procedure
   (EPA Method 160.2)

 • Measures  Volatile Residue
   (EPA Method 160.4)
en
£>
o

-------
641

-------
     Robotic   TSS  Procedure
 Precondition
   System
  Remove
Crucible from
 Desiccator
  Weigh
Dry Crucible
(Initial Weight)
 Moisten Filter
 in Crucible
 Place Crucible
   in Stand
     \
 invert Flask
  in Stand
   Detect
Conclusion of
Rinse Water
 (3 Times)
Return Empty
  Flask to
Reapply Vacuum
 to Crucible to
 Remove Any
                                                   to
Return Crucible
  to Oven
  and Heat
Cool Crucible
in Desiccator
Re=-Weigh
 Crucible
 Heat Crucible
 Again,Cool-jn:
Desiccator, and
 Weigh Again
   Calculate
   Difference
 Between Initial
and Final Weights

-------
Comparison  of TSS Results  by
 Robotic and Manual Methods
 Method
 Mean

 S.D.
Bench Scale Bfotreater
Aerator 61786
Robotic TSS,
mg/l
10670
10130
11631
11219
10820
9981
10761
10200
10676
566
5.3
Manual TSS,
mg/l
9670
10590
10250
11130
10540
10320
11530
10640
10584
562
5.3
Aerator Basin
Robotic TSS,
mg/l
122.4
115.2
125.6
123.2
118.4
122.8
116.8
141.2
123.2
8.1
6.6
Manual TSS,
mg/l
139.2
152.0
127.6
130.0
128.8
128.4
126.4
124.8
132.2
9.1
6.9
                               CO

-------
    Recovery  Results for  TSS  by
    Robotic  and  Manual  Methods
APG #1a)
APG #2a)
APG #3a)
APG #4a)
APG #5a)
APG #6a)
             True TSS,   Robotic WRC   Manual WRC   Manual APG
              rnq/1     % Recovery   % Recovery  Mean % Rec0'
36.5
52.7
77.0
287.0
316.8
498.6
100.0
107.0
94.9
93.5
107.0
98.7
94.9
104.0
83.0
96.8
92.2
94.8
97.2
96.3
95.6
92.5
90.7
90.7
94.2
94.6
94.2

a) Proficiency Environmental Testing Program Samples: prepared and provided by Analytical
  Products Group (APG), Inc., Belpre, Ohio.
b) In-house control sample.
c) Mean recovery of 50 labs participating in the APG Program.

-------
TSS  Linearity  with Sample Volume
mg Weighed vs. Volume, Manual and Robot
      24
      20-
        0
           Y Actual

           Y Predicted
                                       a\
                                       *»
                                       ui
    40
mg Weighed

-------
   Benefits of
Lab Automation
Unattended Operation
Reduced Time/Analysis
Improved Precision
Gained "New" Time for R&D
*>.

-------
                             647
                                   MR. TELLIARD:  Our next



speaker was supposed to be Gary Jackson, but I found out



they had some real technical work to do at the lab, so



instead of him, they sent Dale.  Dale is going to talk about



some pesticide analysis using the Dean Stark extractor and



isotope dilution GC/MS.

-------
                               648
Determination of Semivolatile Pollutants in Sewage Sludge by
Soxhlet/Dean-Stark Extraction, High Performance Liquid
Chromatography Cleanup, and Isotope Dilution Gas Chromatography/
Mass Spectrometry
Gary B. Jackson and D.R. Rushneck, Analytical Technologies, Inc.,
225 Commerce Drive, Fort Collins  CO  80524, and John Tessari,
Colorado Epidemiological Pesticide Study Center, Colorado State
University, Fort Collins  CO  80523.
ABSTRACT
    This paper gives the results of the determination of semi-
volatile pollutants in sewage sludge using Soxhlet/ Dean-Stark
(SDS) extraction, preparative scale high performance liquid
Chromatography (HPLC) cleanup, and isotope dilution gas
chromatography/mass spectrometry (GCMS).
    Large scale  (70 gram) sludge samples were extracted by SDS,
continuous liquid/liquid, soxhlet, and ultrasonic extraction
techniques.  All of these techniques resulted in large amounts of
interferences being co-extracted with the compounds of interest.
    Results showed that SDS extraction is more efficient at
extracting the compounds of interest than the other techniques,
but it is also more efficient at extracting interfering compounds.
    HPLC improves cleanup slightly, but not sufficiently to make
this technique desirable as a standard cleanup procedure.

-------
                               649
    The conclusion from these tests is that large-scale sewage
sludge samples will require further, alternative cleanup
techniques before reliable measurements of the semi-volatile
pollutants at the part-per-billion level in the solid phase of
the sludge can be made.
BACKGROUND
    EPA has been attempting to measure the semi-volatile and
pesticide pollutants in sewage sludge at the part-per-billion
(ppb) and sub-ppb level since the early 1970's.  Those familiar
with the determination of these analytes in sewage sludge will
expound at length on how difficult these determinations are,
mainly because of the large concentrations and large number of
interfering substances in the sludge.  A variety of different
approaches have been tried in attempts to improve the precision
and accuracy of measurements, and to lower the detection limit for
the pollutants of interest.
    The use of isotope dilution gas chromatography/mass
spectrometry (GCMS) and gel permeation chromatography (GPC) have
been helpful in improving the measurement of pollutants in sludge.
Further improvements in cleanup are limited by the desire to
measure analytes with a wide variety of chemical species.  Thus,
the normal cleanup techniques that may be effective for a small
group of similar compounds (e.g., sulfuric .acid cleanup of sludge
extracts for determination of polychlorinated biphenyls) cannot be
used for cleanup of normal and polynuclear aromatic hydrocarbons,
amines, phenols, ethers, and other species because of the
destruction or removal of many of these species by the cleanup
                           page -2-

-------
                                650
technique.
    Against this background, this study attempted to employ
Soxhlet/Dean-Stark (SDS) extraction and preparative scale high
performance liquid chromatography in addition to GPC and isotope
dilution GCMS as techniques for selective extraction and for large
scale cleanup of sludge extracts.  The study objectives are shown
in Figure 1.
SLUDGE
    The sewage sludge used for all measurements was filter cake
obtained from a nearby publicly owned treatment works (POTW).  The
characteristics of this sludge are shown in Figure 2.
    The sludge contained 14 percent solids.  EPA Method 1625, the
isotope dilution method for determination of semi-volatile
pollutants, employs either continuous liquid/liquid extraction
(CLLE) or ultrasonic (sonic) extraction, depending on the solids
content of the sample.  If the percent solids is less than one,
the sample is extracted directly.  If the percent solids is in the
range of 10 - 30, the sample is diluted with reagent water to one
percent solids (10 grams in one liter) and extracted using CLLE.
If the percent solids is 30 percent or greater, the sample is
extracted using the sonic technique.
    In order to test SDS and other extraction techniques, a mass
of 10 grams of solids was chosen.  At 14 percent solids, the
weight of sludge sample was therefore 70 grams.  The 10 gram mass
of solids is at the break point between the use of CLLE and sonic
extractions; i.e., this mass would be required if the sludge
contained 30 percent solids.
                           page -3-

-------
                               651
    Method 1625 requires the use of a "dilute aliquot" when
interferences are know or suspected.  For the dilute aliquot, the
sample is diluted by a factor of 10 and this diluted sample is
then analyzed.  The dilute aliquot was not employed in this study
because it is well known that one gram of sludge solids can be
successfully analyzed by Method 1625 as written.

OVERVIEW OF EXTRACTION/CLEANUP
    In order to determine the effectiveness of SDS extraction, the
SDS technique was compared to several other techniques, as shown
in Figure l.  The solvent systems used are given in the figure.
    An overview of the extraction and cleanup processes are shown
in Figure 3.  For each extraction technique, an aliquot of sludge
was extracted after spiking with the stable isotopically labeled
compounds, a second aliquot was extracted after spiking with the
labeled compounds and pollutants, and a third aliquot was
extracted unspiked.  The spiking level was 100 ng for each
component.
    Extracts were concentrated to 10 mL final volume and processed
through the GPC.  In this process, 50 percent of the extract is
recovered.  Each 5 mL GPC eluate was concentrated to 0.5 mL
(to compensate the 50 percent loss).  The 0.5 mL concentrated
eluate was split in two.  One of these splits was spiked with
internal standard and a one microliter aliquot was injected into
the GCMS.  This aliquot permitted comparison of the extraction
techniques.
    The remaining halves of the concentrated GPC eluates from the
SDS extraction technique were subjected to HPLC cleanup.  The
                           page -4-

-------
                                652
concentrated GPC eluate from the unspiked sludge was spiked with
the pollutants and labeled compounds so that losses associated
with the sludge matrix during HPLC cleanup could be quantified.
In addition, a standard containing the pollutants and labeled
compounds was processed through the HPLC to measure losses in the
absence of the sludge matrix.
    After HPLC cleanup, the HPLC eluates were re-concentrated to
250 uL, the internal standard was added, and one microliter
aliquots were injected into the GCMS.

SOXHLET/DEAN-STARK EXTRACTION
    SDS extraction employs a moisture trap in combination with a
Soxhlet extractor, as shown in Figure 4.  The successful
application of this technique to determination of chlorinated
dioxins and furans has been reported (reference 1), and the
technique is part of EPA Method 1613 for dioxin/furan
measurements.
    The apparatus uses a lighter than water solvent that forms an
azeotrope with water for the extraction.  Most commonly, the
solvents employed have been benzene and toluene.  The azeotrope is
condensed and falls into the moisture trap where the water settles
to the bottom of the trap.  The water can be drained and measured
either gravimetrically or volumetrically.  The percent moisture in
the sample can then be calculated based on the sample weight.
    In applying the SDS extraction technique, we believed that the
use of a slightly polar solvent (benzene) would reduce the
quantity of interferences co-extracted from the sludge, yet would
permit extraction of the components of interest.
                           page -5-

-------
                               653
HPLC CLEANUP
    This cleanup employed a preparative scale column with the
characteristics shown in Figure 5.  The column was operated
isocratically with acetonitrile as the eluting solvent in a Waters
HPLC system with an ultra-violet  (UV) detector operated at 254 nm.
RESULTS
Chromatocrrams
    Figures 6-10 show reconstructed ion current  (RIC)
chromatograms comparing the various extraction techniques.  These
chromatograms are scaled to the largest peak.  The fine structure
in these chromatograms give an indication of how much material is
being extracted and how readily the components of interest can be
measured.
    The small, well defined peaks at the beginning of these
chromatograms are the labeled compounds and can be used to compare
the total amounts of material extracted, and to indicate the
degree of difficulty in locating the labeled and native compounds
in the sample matrix.
    Chromatograms of the base/neutral and acid extracts are
analyzed separately in Method 1625.  As shown in Figures 6 and 7,
keeping these extracts separate provides an improvement over the
bulk simultaneous extraction of the base/neutral and acid
analytes, as evidenced by Figures 8-10.
    The chromatogram in Figure 10 gives an indication that much
more material is extracted using the SDS extractor than with the
other extraction techniques.  Nearly all of the analytes are
                           page -6-

-------
                                654
masked by the large amount of material extracted from the sample
matrix.  The two large peaks in this chromatogram, and at or near
the same retention time in all of the chromatograms are
hexadecanoic and octadecanoic acid.

Comparison of Extraction Methods
    Figure 11 compares the concentrations of pollutants detected
by the various extraction methods.  In reducing the GCMS data, a
formidable problem is presented in attempting to identify the
pollutants.  Spectra are heavily contaminated by interfering
compounds, and the quantitation m/z is frequently inflated by
a contribution from one or more of these compounds.  It must be
reiterated that the sample size used for these analyses is larger
than normally attempted, so the lack of complete success in
rigorously determining the concentration and identity of the
pollutants of interest is attributable to this large sample size.
    Further, the seeming lack of success of identifying pollutants
from the SDS extraction can be attributed to the large amounts of
interfering materials co-extracted from the matrix.

HPLC Cleanup
    Figures 12 and 13 allow comparison of the HPLC chromatogram of
the 100 ug/mL pollutant standard with that of the spiked SDS
extract.  The similarities in the chromatograms exist because the
UV detector responds primarily to the polynuclear aromatic
compounds in these solutions and a lack of naturally occurring
polynuclear aromatics in the sludge.  The detector is attenuated
to bring the chromatograms on scale.  Even so, saturation of the
                           page -7-

-------
                               655
detector occurs in the beginning of the HPLC run for the SDS
extract.
    Figure 14 compares the number of compounds detected before and
after HPLC cleanup with that of a standard containing all 74
labeled and 82 native compounds.  As can be seen, HPLC cleanup
improves the number of compounds detected in the SDS extract, but
HPLC cleanup also removes some of the analytes of interest.  The
analytes removed consist mainly of the normal hydrocarbons that
elute on the tail of the chromatogram (see Figures 12 and 13).

CONCLUSIONS
    From the test results presented in this report, it can be
concluded that:
(1)  Soxhlet/Dean-Stark extraction is more efficient at extracting
substances from the sludge matrix.  Unfortunately, the greater
numbers and amounts of substances extracted mask the compounds of
interest.
(2)  HPLC provides a slight improvement in cleanup of sludge
extracts.  Unfortunately, some analytes of interest are lost in
this cleanup.
(3)  Large amounts of sludge (10 grams of solids) cannot be
effectively extracted and cleaned up by the techniques tested in
this work.  Based on experience with other sludge samples, smaller
amounts of sludge (1-3 grams of solids) can be effectively,
extracted and cleaned up using continuous liquid/liquid extraction
and gel permeation chromatography.
                           page -8-

-------
                               656
REFERENCE
Lamparski, L.L., and Nestrick, T.J., "Novel Extraction Device for
the Determination of Chlorinated Dibenzo-p-dioxins  (PCDDs) and
Dibenzofurans  (PCDFs) in Matrices Containing Water", Chemosphere.
19:27-31, 1989.
                             page -9-

-------
SDS EXTRACTION AND HPLC CLEANUP OF SEWAGE SLUDGE

 OBJECTIVES
   COMPARE SDS EXTRACTION WITH OTHER METHODS
   ASSESS CLEANUP PROVIDED BY HPLC
   ACHIEVE LOW DETECTION LIMITS IN SLUDGE
 EXTRACTION METHODS COMPARED
     METHOD
   SOXHLET/DEAN-STARK
   SOXHLET
   CONTINUOUS
   SONICATION
SOLVENT
BENZENE
CH2CL2
CH2CL2
CH2CL2/ACETONE
               FIGURE 1
               STUDY OBJECTIVES

-------
SDS/HPLC

 SLUDGE USED

   TYPE: FILTER CAKE

   PERCENT SOLIDS: 14

   SAMPLE SIZE: 70 GRAMS
      14 % SOLIDS x 70 g = 10 g SOLIDS
      10 g SOLIDS IS CLLE/SONIC BREAKPOINT IN 1625
en
oo
        FIGURE 2
        SLUDGE CHARACTERISTICS

-------
OVERVIEW FLOW CHART
WEIGH

SPIKE

EXTRACT

CONC

GPC

SPLIT	

HPLC

CONC
                    70 grams

                    unsp/label/native
                    100 ngeach
                     lOmLFV

                     5 mL

                   > INT STD—> GCMS
                    250 ti
                    250
                                                      CTk
                                                      (Jl
          INT STD—> GCMS
FIGURE 3
OVERVIEW OF EXTRACTION AND CLEANUP

-------
                                                 en
                                                 o>
                                                 o
FIGURE 4

SOXHLET/DEAN-STARK EXTRACTOR

-------
HPLC CLEANUP

 COLUMN
   TYPE: ^BONDAPAK (WATERS)
   DIMENSIONS: 2.5X10 CM
   MATERIAL:
      CIS
      125 A; 10 /i
   MOUNTING DEVICE: RADIAL MODULE

 SOLVENT & FLOW RATE
   ACETONITRILE
   ISOCRATIC @ 5 ML/MIN
           FIGURE 5

-------
                        CLLE ACID EXTRACT AFTER GPC
      SAMPLE: CLLE ACID - LABELED
      CONOS.:•
      RANGE: G  1,2888  LABEL: N 0. 4.0  QUAN: A 0, 1.0 J  0 BASE: U 20*  3
100.0-1
                     FIGURE 6
 RIC
                            A
                                                                                                     CTV
                                                                                                     CTl
                                                                                                     to
         509
         8:20
1000
16:40
1500
25:00
2000
33:20
2500
41:40
SCAN
TIME

-------
                        CLLE B/N EXTRACT AFTER GPC
       SAMPLE: CLLE B/N - LABELED
       CONDS.:,
       RANGE: G   1,2888  LABEL: N 0, 4.8  QUAN: A  Q> 1.0 J 0  BASE: U 20,  3
100.9-1
 RIC
         500
         8:20
1000
16:40
1500
25:00
2000
33s 20
2500
41s 40
                                                                                                         cn
                                                                                                         CO
SCAN
TIME

-------
                      SOXHLET EXTRACT AFTER GPC
      SAMPLE: SOX - LABELED
      CONDS.:
      RANGE: G  1/2880  LABEL: N 0, 4.0  QUAH: A 0/ 1.0 J 0  BASE: U 20/  3
100,0-,
 RIC _
         500
         8:20
1000
16:40
2000
33:20
                                                                                                      en
                                                                                                      en
2500
41:49
SCAN
TIME

-------
                          SONIC EXTRACT AFTER GPC
      SAMPLE: SON - LABELED
      CONDS.;
      RANGE: G  1,2889  LABEL: N 0/4.6  QUAN: A V, 1.8 J  8 BASE: U 28/  3
100.0-1
 RIC
                                                                                                      (Jl
                                                            2000
                                                            33:20
2500
41:40
SCAN
TIME

-------
                     SDS EXTRACT AFTER GPC
      SAMPLE! SOS - LABELED
      COHDS, r
      RANGE J G   1,2880  LABEL; N  0/4,0  QUAN: A 0/ 1.0 J  0  BASE: U 20;  3
100,01
 R1C .
                                            1500
                                            25:00
2000
33:20
2500
41:40
SCAN
TIME

-------
COMPOUNDS DETECTED
PHENOL
n-CIO
1,3-DICHLROBZ
p-CYMENE
NAPHTHALENE
n-C12
BIPHENYL
ACENAPHTHENE
n-C16
DIBUTYLPHTH
n-C20
BIS(2ETIIEX)PH
n-C30
PHENANTHRENE -
DIOCTYLPHTH   -
CONC IN SOLIDS (me/be)
ACD
1
— •
-
—
—
-
-
-
—
-
- -
140
31
B/N
-
1
1
3
1
-
-
-
5
3
1
12
3
SOX
7
43
4
24
2
2
26
1
32
0.5
57
150
8
SDS
-
7
1
8
2
2
16
0.1
3
-
14
-.
—
SON
—
24
1
6
1
1
4
1
9
-
0.5
1500
—
SDS
-
-
-
—
2
-
24
-
—
0.7
9
112
—
                   (7)
2
3
      FIGURE 11
      RESULTS OF EXTRACTION METHODS

-------
                   668
  HPLC OF SDS EXTRACT (AFTER GPC)
 Ul

 H
u.
c
              FIGURE 12

-------
                  669
 HPLC OF 100 pg/mL NATIVE STANDARD
             CD
Ul
H
           FIGURE 13

-------
SDS + HPLC RESULTS
            SDS EXTRACTED
            EXTRACT  AFTER HPLC
                            SPIKED
            LBL NAT LBL NAT LBL NAT
NUMBER*
AVE % REC
29  42   41   43   52   58
13  103  31   97   81   106
*OF 74 LABELED & 82 NATIVE
                                           en
                                           -j
                                           o
        FIGURE 14
        EFFECT OF HPLC CLEANUP

-------
                             671
                                   MR. TELLIARD:  Are there
any questions?
(No response.)
                                   MR. TELLIARD:  No
questions?
                                   MR. RUSHNECK:  No
questions.  Everyone can go home and do sludge now.
Wonderful
                                   MR. TELLIARD:  Thank you,
Dale.  Thank you very much.

-------
                             672
                              MR. TELLIARD:  A couple of
notes.  The following speakers are going to be talking about
one of our favoritest subjects, laboratory certification and
reciprocity.  We have beat this horse so many times it is
wounded and dying, but we are going to do it again.  We are
going to do it until we get it right.
     The agency has done something between our last
discussion on the subject and today in that a committee was
formed with is called the Methods Management Committee which
is part of the agency that is looking at monitoring and
methods across all media and in all program offices.
     One of the ad hoc committees that has been formed under
that is a committee to look at the issue of certification.
Whether the agency is going to do it...as you know/ part of
our agency does do that, Drinking Water...but what the other
program offices are going to do and what form it might look
like and take.  So, you will hear some of that kicked around
today, I hope.
     So, our next speaker will be Rhonda Whalen of the
Department of Health and Human Services who will talk about
the Federal Government perspective on the regulation of
laboratories under CLIA of 1988.  She will explain that.

-------
                             673
                                   MS. WHALEN:  Good
afternoon.
I am Rhonda Whalen.  I am with the Health Care Financing
Administration with the Health Standards and Quality Bureau,
and we are charged with the responsibility of implementing
the Clinical Laboratory Improvement Amendments of 1988.
     Basically, the law that was enacted October 31/ 1988
pertains to all entities that test human specimens.  To set
this in context, I am going to give you a little background
information on what we currently do, what our current
framework is, and then what it is we are going to be doing
in the future.
     At present, the Health Care Finance Administration
regulates laboratories under two federal programs, the
Medicare/Medicaid programs and the Clinical Laboratories
Improvement Act of 1967.  Medicare and Medicaid are
reimbursement or payment programs, Medicare for
beneficiaries over 65 and Medicaid for recipients through
the State program.  The Clinical Laboratories Improvement
Act of 1967 regulates laboratories that test specimens that
cross State lines.
     At present, we have approximately 4900 Medicare-
approved independent laboratories, and We have approximately
6600 Medicare-approved hospital laboratories.  Of the 6600
Medicare-approved hospital laboratories, about 5200 are

-------
                             674
accredited by either the Joint Commission on Accreditation
of Health Care Organizations or the American Osteopathic
Association.
     What this means is that the Federal Government does not
directly inspect these hospitals.  Rather, we deem the
accrediting organization's inspection programs, and the way
we conduct inspections is that we have agreements with all
50 States, and the States do the direct inspections, and the
Health Care Financing Administration has regional offices
that oversee the State operations.
     There are about 2900 laboratories subject to the
Clinical Laboratories Improvement Act of 1967, that is, they
test human specimens in interstate commerce.
     So, the total universe of regulated laboratories by the
Health Care Financing Administration is approximately
12,000.  As I mentioned, for half of those, we deem the
accreditation programs of JCAA or the American Osteopathic
Association to inspect the hospital laboratories.
     Beginning in 1987, a series of newspaper and magazine
articles were published on the quality of laboratory
testing, and, simultaneously, television programs were aired
concerning the number of laboratories that were not subject
to either Federal or State regulations.
     Congress held hearings in 1988 and heard testimony from
victims of faulty laboratory testing.  Specific concerns

-------
                             675
were raised about the validity of cholesterol screening and
the accuracy of pap smear results.
     Now, in that environment, we currently are responsible
for regulating laboratories based on location.  I mentioned
independent laboratories, hospital laboratories.  We also
have laboratory services in nursing homes and dialysis
facilities, et cetera.
     In order to run the regulatory program for
laboratories, we had interagency agreements with the Public
Health Service.  We have a memorandum of understanding with
the Centers for Disease Control for provision of scientific
and technical expertise on questions related to advances in
instrumentation, new technology, proficiency testing, and
cytology services.  We also have a memorandum of
understanding between the Health Care Financing
Administration and the Food and Drug Administration for the
provision of technical assistance concerning blood bank
services.
     So, at present, we work very closely with the Public
Health Service and both CDC and FDA.  This is based on long-
standing arrangements, because prior to 1979, the Centers
for Disease Control had responsibility for regulating the
interstate laboratories.  In 1979, the Health Care Financing
Administration became responsible for licensing and

-------
                             676
inspection of laboratories that test specimens in interstate
commerce.
     In 1980, the Health Care Financing Administration
assumed responsibility for the inspection of approximately
4500 registered blood establishments that also participate
in Medicare.  These transfusion facilities that are
primarily located in the private hospitals either collect or
transfuse whole blood or packed cells or collect other blood
components in emergency situations.
     Both of these arrangements were generated on the
premise that we were trying to reduce inspections.   The
Department of Health and Human Services had simultaneous
programs and, per FDA, CDC, and the Health Care Financing
Administration, was conducting inspections on an annual
basis.  So, what we were attempting to do was reduce the
numbers of inspections.
     Therefore, we proceeded in negotiations with agreements
with both the Centers for Disease Control and FDA which took
over the inspection responsibility, leaving the Centers for
Disease Control and FDA with the responsibility for
providing HCVA with the technical expertise necessary to run
those two programs while we continued to carry on our
primary responsibility which is to inspect facilities for
approval for payment under Medicare and Medicaid.

-------
                             677
     Now, there is an historical basis which was occurring
during this timeframe prior to the enactment of CLIA.  One
piece of legislation was the Omnibus Budget Reconciliation
Act of 1987.  This was Congress' first attempt to say that
the laboratories located in physicians' offices that we
currently reimburse for services provided to Medicare
beneficiaries should be subject to regulations.
     Essentially, what Congress said was that a physician's
office laboratory that performs more than 5000 tests per
year should be subject to regulation, and they charged the
Department with regulating those laboratories.
     In the meantime, this omnibus law that we are charged
with implementing was enacted on October 31, 1988, and it is
bigger than physician office laboratories.  It says that all
entities, certainly including physician office laboratories,
that perform laboratory testing would be subject to federal
regulations.
     The main impact of the law is that it affects all
testing entities, dramatically increasing the Health Care
Financing Administration's role from regulating the current
12,000 laboratories to include an estimated 100,000
physician office labs and perhaps another 200,000 to 300,000
or 400,000 laboratories.
     A unique aspect of the Clinical Laboratory Improvement
Amendments of 1988 is that the law specifies the regulation

-------
                             678
of laboratories according to testing performed.  I mentioned
a few minutes ago that we are responsible for regulating
laboratories based on location.  It is the current basis of
our regulation.
     This law says we should regulate based on the testing
performed.  It certainly is a reasonable basis for
establishing regulation, but it is somewhat difficult
because we have no data.  We have no information on adverse
impact of incorrect laboratory results, and we are being
asked to establish regulations on the basis of the
complexity of laboratory testing, meaning that the if the
testing is more complex, we should have more strenuous
regulations.
     Therefore, if a particular test requires more
interpretation, more objectivity, more judgment, then that
type of regulation or that type of testing should be more
heavily regulated, and we are without data to establish that
type of regulatory program.
     Now, one thing that Congress did in December of last
year is it withdrew the OBRA '87 law which said we would
regulate physician office labs that do over 5000 tests.
This was extremely helpful to us, because, number one, it
was this OBRA '87 was in conflict with CLIA '88.  CLIA '88
says you regulate all entities.  It doesn't specify
physician office labs, and it doesn't talk about volumes of

-------
                             679
tests performed in a laboratory.  It talks about the kinds
of tests performed.
     So, we were faced with a problem that regulates by
location, being asked to implement a law that designates
physician office labs to be regulated and specifies volume
simultaneously with a law that says no, you should regulate
on the basis of testing.  So, we were in somewhat of a
dilemma.
     Congress helped us out by withdrawing the Omnibus
Budget Reconciliation Act of 1987 in December of 1989.
Therefore, we have one omnibus law that we must implement.
     The other thing that the Omnibus Budget Reconciliation
Act of 1989 did is tie in Medicare payments with CLIA.  What
that means is that in order for a laboratory to be paid
under Medicare, that laboratory will have to be certified
under CLIA.
     That has always been true for Medicare-approved
independent labs or hospitals or nursing homes but has not
been true for physician office labs.  Presently, physician
office labs are paid simply on the basis of the physician's
name and receiving a billing number.  The physician's lab
does not have to meet any standards.
     So, CLIA '88 changes that and says that, number one,
those labs will be regulated, and then the Omnibus Budget
Reconciliation Act of 1989 came along and said not only were

-------
                             680
those labs to be regulate, but in order to be paid, those
laboratories will have to have a certificate.
     Now, as I mentioned, at present, we are regulating
12,000 labs under Medicare and have the potential of
regulating over 300,000 labs under CLIA.  Therefore, the
CLIA program will be the largest program and Medicare will
then be a smaller subset of the CLIA program.
     During this time period in which Congress was hearing
testimony on the faulty laboratory testing practices, the
Health Care Financing Administration, in conjunction with
the Centers for Disease Control and the Food and Drug
Administration, was attempting to revise the current
regulations that we have for laboratories.  We had a Notice
of Proposed Rulemaking which was published August 5, 1989,
and we stated our proposed regulations to revise our current
requirements for personnel, quality control, record keeping,
proficiency testing, and we were attempting to establish new
quality assurance requirements.
     The closing date for the comments was November 3, 1988.
CLIA was enacted October 31, 1988.  So, we had somewhat of a
dilemma, that dilemma being would we throw out our proposed
rule and ignore the over 1600 comments we received and just
start with the implementation of CLIA  '88.
     We determined that what we had proposed as a revision
to our current requirements was in keeping with the

-------
                             681



requirements that we would have to implement under CLIA '88.



Therefore, we considered the 1600 comments, we analyzed



them, and we responded to them in the preamble to the



regulations that we published in final on March 14, 1990.



     Those regulations will be effective September 10, 1990.



They will affect those laboratories currently regulated



under Medicare and those laboratories that test specimens in



interstate commerce.  What this rule does is it establishes



the basis or the framework for the regulations that we will



implement under CLIA '88.



     This regulation published in final on March 14



establishes uniform proficiency testing requirements, a



grading system, criteria for approving proficiency testing



programs, the types of samples that should be sent by an



approved proficiency testing program to participating



laboratories, the kinds of challenges that should be



included, and the frequency of the testing events.



     We revised our quality control requirements to be



current with new methodology and changes in technology.  We



established quality assurance requirements that would



encompass the outcome measurements of laboratory quality,



and the self-implementing provisions of CLIA '88, that is,



the language in the statute that is so clear it doesn't



require interpretation or rulemaking we put in this rule.



We just lifted out the statutory language, and we put those

-------
                             682
requirements into the March 14 rule so that we could go
forward and implement those provisions of CLIA '88 that are
self-implementing.
     Now, we determined that the Clinical Improvement
Amendments of 1988 are so far reaching and so extensive that
we could not propose one rule to implement CLIA '88.
Therefore, we have five rules all of which will require
comment period and evaluation of the comments and then
publication of a final rule.
     The first rule is the most controversial rule, and I
called this morning to see where we were.  There is a strong
possibility that that rule is going to be published the end
of next week.  We have just about achieved the Office of
Management and Budget clearance needed for publication in
the Federal Register.
     This rule will establish standards based on the
complexity of test performed.  It will propose a list of
tests for waiver or exemption from the requirements, that
is, those tests that are so simple, accurate, and low risk
and pose no reasonable risk of harm if performed
incorrectly, then if laboratories do only those tests, the
laboratory would be exempt from meeting Federal standards.
     We are also going to propose personnel requirements
based on complexity of testing.  This is also quite
controversial since heretofore we have not done that, and

-------
                             683
anytime you specify personnel requirements, that always
generates a lot of comments.  That rule will have a 90-day
comment period, and one of the most difficult things about
this whole activity is that we don't know the number of labs
we need to regulate, we don't know the testing performed in
those laboratories, we do not know the types of personnel
employed in those laboratories.
     So, that brings me to the second rule that we are going
to attempt to implement.  Probably in June it will probably
be published as a proposed rule and will have a 60-day
comment period.
     That rule is an interim procedure for provisional
registration of laboratories.  That is where we will sign
laboratories up.  They will pay a registration fee, and we
will find out how many laboratories there are in the United
States, what kinds of testing the laboratories conduct, and
the types of individuals employed in these laboratories.
     Also, this is a self-implementing law, that is,
Congress is not appropriating any budget, any funds, to run
this program.  Therefore, we have to charge the laboratories
a fee which we determined it would be most appropriate to
charge each laboratory the portion of the amount necessary
to determine compliance.  That is, if the lab is smaller,
less complex, fewer tests are performed, that laboratory
would pay a smaller fee.  Ultimately, what we are proposing

-------
                             684
to do in this regulation is to  set forth our fee schedule
methodology.
     That rule will have a 60-day comment period/ and that
is probably the rule that we will implement first.  It is
not quite as controversial and, obviously, we do need to get
the laboratories registered, and we do need to start
collecting money in order to maintain the program.
     Now, the third rule that we need to publish as a
proposed rule is the criteria for recognition of
accreditation programs.  The law specifies that we can
recognize private non-profit organizations' accreditation
programs and State programs with equivalent standards.
     So, what we want to do is  propose the criteria for
recognition of those programs,  the kinds of information we
are going to request, the kinds of data that we are going to
need on an ongoing basis, and then as soon as we establish
our standards, then we can accept applications from
interested accreditation programs and State programs.  That
rule is probably going to be published later on this summer.
     The fourth rule is the proposed regulations that would
implement the intermediate sanction and adverse action
procedures, that is, the penalties that will accrue to a
laboratory that does not meet the requirements.  There are
monetary sanctions and various penalties that laboratories

-------
                             685
are subject to that do not meet the requirements.  That rule
will also be published later on in the summer.
     The important thing about the hearings attached to CLIA
are that in Medicare, if a laboratory does not meet the
requirements, we terminate that laboratory's approval at
that time.  The laboratory then is allowed an opportunity
for a hearing.
     Under CLIA '88, since any adverse action against a
laboratory, if it went to conclusion, would prohibit that
laboratory from conducting testing in this country.  If you
lose your certificate, then you can't conduct testing in the
United States.  The laboratory is allowed an appeal process
prior to any action on the certification.
     So, we have to propose the hearing process, the types
of evidence that will be allowed.  It particular pertains to
the judicial proceedings.
     Along with that rule, there is a particular statutory
provision that says that the department will make lists of
these laboratories available to the public that are subject
to intermediate sanctions, Medicare approval terminations,
and loss of certification.  So, that information is going to
be available to the public.
     The fifth and final rule will set forth the
responsibilities and functions of the State survey agencies
that we employ to do the direct surveys for us and also the

-------
                             686
 responsibilities of the Health Care Financing Administration
 in carrying out the activities of CLIA '88.
      There continues to be a great deal of interest from
 individual laboratories, laboratorians, physicians,
 laboratory equipment manufacturers, and the media in the
 implementation of CLIA.  For four nights in February,
 Channel 4 TV news in Washington,  B.C.  aired a program called
 "Deadly Mistakes:  Promises Broken" which continued to focus
 on individuals who have died or were irrevocably harmed due
 to incorrect laboratory results.   There is a continued
 emphasis by the media and a focus on the need for
 implementing these requirements.   We have felt a good deal
 of pressure to get these regulations out, the comment period
 allowed, and then publish these rules in final to start
 regulating laboratories under CLIA.
      Congress held hearings on March 7, and our
 administrator testified, and the thrust of the hearings,
 basically, from the Senate was, why don't you have these
 regulations published?  Why aren't you regulating
. laboratories under CLIA?  We enacted these laws because we
 wanted to protect the American public, and we don't see the
 Department responding.
      So, essentially, we are about ready to at least be able
 to say that we are doing something, but while the media and
 Congress are concerned that we have not acted, the soon to

-------
                             687
be regulated entity is very concerned that we are moving too
fast.
     There are five studies that are mandated under CLIA '88
all of which would be very useful in establishing
regulations.  Those studies were supposed to be out in May.
They are not completed.  That is the charge of the Public
Health Service.  In fact, they are in their very embryonic
stage.
     So, in the meantime, we are going to establish
regulations in the absence of the information that would be
gained from these studies.
     Thank you.
                                   MR. TELLIARD:  Thank you.

-------
                             688
                   QUESTION AND ANSWER SESSION
                              MR. TELLIARD:  Are there any
questions?
                              MS. ORDONA:  I have a
question.  My name is Alicia Ordona.  I am from the
Commonwealth of Virginia, Division of Consolidated
Laboratory Services.
     My next question is, is the State laboratory going to
be included in this regulation?
                              MS. WHALEN:  Is the State
laboratory what?
                              MS. ORDONA:  Going to be
included under this regulation.
                              MS. WHALEN:  If the State
laboratory does tests on human specimens, yes.
                              MS. ORDONA:  Really?
                              MS. WHALEN:  Yes, and all
Federal laboratories
                              MS. ORDONA:  Thank you.
                              MR. RUSHNECK:  Rhonda, Dale
Rushneck of ATI.
     I am curious about the fee schedule methodology.  I
know you don't have to have the rule out until late summer,
but can you give us your thinking?  Is it going to be based
on size dollar volume of the laboratory, on the number of

-------
                             689
tests performed, is it going to be split and a fee
associated with each test?  What is the current thinking?
                                   MS. WHALEN:  What we did,
we really did look at what we currently...our current costs
related to the inspection of laboratories and the
determination of laboratory compliance.  It is somewhat
difficult right now, because we are budgeted an overall
amount, and we negotiate with every State survey agency a
different hourly rate, because the States are our
contractors, so to speak.
     The part that makes it complicated is that our data
right now is somewhat limited because it is by State, not by
laboratory, and it is also based on our current regulatory
program which we are attempted to change and certainly
strengthen the requirements and the amount of time that will
be necessary to determine compliance.
     So, with all of that, then I am going to tell you that
we used our current information and attempted to establish
fees.  It is not a particular secret, but the fees will be
biennial, every two years, because that is specified by law.
If the laboratory is eligible for a certificate of waiver,
it is a nominal fee, and the nominal fee that we came up
with for the proposed rule is around $150.
     The certificate fee, that is, the administrative cost
for issuing the piece of paper, will be between $200 and

-------
                             690
$300 for a certificate, a provisional certificate, or an
accreditation certificate.  Now, the fees that we are
proposing for determining compliance...and remember again
this is for every two-year period...runs from around $800
to...it seems like it is up around  $2000.
     Now, we have set what we believe to be our base costs.
If it turns out that it costs us more to determine a
laboratory's compliance, meaning that the laboratory was not
ready for an inspection, had a lot  of deficiencies, we had
to go back, we had to conduct a follow-up inspection, the
laboratory will get a second bill.
     So, the first bill is going to be based on the extent
of services and the volume of services, and that will set a
base cost, and it is the minimum fee.  If we incur
additional costs, then the laboratory will get a second
bill.
     And as we gain more information on our costs relative
to determining compliance under CLIA '88, we will upgrade
our fees or downgrade if it turns out that we should be
.charging less, although I kind of doubt that, because CLIA
'88 is going to require many more things.  Therefore, I
think that fees will go up, not down.
                              MR. TELLIARD:  All right,
thank you very much.
     Baldev, did you have a question?

-------
                             691
                                   MR. BATHIJA:  A quick
question.  Will you be allowed to keep that fee you collect,
or does it revert back to the...
                                   MS. WHALEN:  Federal
Treasury, and then we get budgeted out of that money, we
hope.
                                   MR. TELLIARD:  I would
like to have the panelists come up during the break that we
are about to have so that you can go get your Coke and Pepsi
and come back in, and if the panelists would come up and
take a chair up here so that you are sitting target for the
rest of the audience, if you would do that right after the
break, I would appreciate it.
     Thank you.  Please get back in here in 15 minutes.
(WHEREUPON, a brief recess was taken.)

-------
                             692
                              MR. TELLIARD:  Our next
speaker is Gerald Hoeltge who is going to carry on with the
program of talking about program and laboratory
certifications.  He comes from our great town, Cleveland,
and is connected with the Cleveland Clinic.
     Gerald?

-------
                             693
                                   DR. HOELTGE:  Thank you
very much.
     In the next few minutes, I want to present the College
of American Pathologists' laboratory accreditation program
as an example of what the private sector can accomplish in
terms of laboratory certification.
     This particular program began about 1965 quite
informally when a group of pathologists got together and
realized that there was no good mechanism for
interlaboratory inspection directed towards improvement.
They agreed to inspect each other's laboratories to offer
constructive criticism and helpful suggestions.  That
activity has grown over the years to the point at which now
it has a fairly substantial structure.  We accredit now just
under 4100 clinical laboratories.
     I want to talk about some of the characteristics of the
program, to give you an idea of the philosophy behind it so
that the operational aspects, will be a little more
understandable.
     First of all, this is a voluntary program.  Now, when I
say voluntary, I am talking about the inspectors for the
program.  They are all practicing laboratorians.  They are
volunteering their time and expertise to the program.
They are recompensed only for meals and mileage.  It is also
voluntary for the laboratory directors who choose to be

-------
                             694
involved in this program.  I say that with a little bit of
caution because, in fact, many directors feel compelled to
get their laboratories accredited.  It comes from several
different directions.  Ms. Whalen talked about CLIA '67.
Those laboratories which are in interstate commerce have to
be licensed.
     The inspection agency for such laboratories, of course,
is the Centers for Disease Control or one of its contract
organizations.  However, a laboratory which is interstate
licensed and accredited by the College can apply to the
Health Care Financing Administration for a waiver from
regular CDC inspections, and that means that the laboratory
has one fewer inspection to suffer through on each cycle.
     Also, the Joint Commission on Accreditation of
Healthcare Organization recognizes CAP program for hospital
laboratories.  As Ms. Whalen pointed out, the Joint
Commission is one of two private agencies that can qualify a
provider for Medicare reimbursement.  So, that means that if
a hospital laboratory is certified by the College, that
certification will be accepted by the Joint Commission.  If
the laboratory using that Joint Commission accreditation to
qualify for Medicare reimbursement, then the CAP program is,
qualifying the laboratory for Medicare reimbursement.  Some
laboratories find that to be a compelling reason.

-------
                             695
     Thirdly, we are seeing increasing numbers of private
carriers who are offering  provider contracts for
competitive bidding, and they are including CAP laboratory
accreditation among the bidding specifications.   Points are
given to those laboratories that have chosen to be measured
by the accreditation criteria.  That has brought a lot of
commercial laboratories into the program who had never
expressed an interest before.
     So, those external characteristics aside, it is a
voluntary program, and it certainly is voluntary on the part
of the inspectors.
     Secondly, it is a peer review program.  The inspectors
are all practicing laboratorians, and the management of the
program, the commissioners for laboratory accreditation, are
all practicing laboratorians.  I run the blood bank and the
transfusion service at the Cleveland Clinic Foundation.
About an hour or two of my day, however, is devoted to CAP
activities, and we manage the accreditation for about 450
laboratories in the Ohio, Indiana, Michigan, and Ontario
area.
     Thirdly, the inspection covers the entire laboratory.
We will not accept an application from a facility that wants
only part of the laboratory inspected.  We are going to look
at the entire facility, even those areas for which we have
no technical expertise (such as in vitro fertilization or

-------
                             696
environmental water quality).  But we can look at any
laboratory at least in terms of safety requirements.
     Fourthly, it is based upon the Standards for Laboratory
Accreditation.  I will talk about that a little bit more in
a moment.
     Fifth/ it combines on-site inspection with proficiency
testing.  Both of them go together.  Neither one suffices
for accreditation by itself.
     Now, the Standards.  This is a document that is the
basis for our program.  It has about a five-year review
cycle.  It is written  generally to address the general
case.  It is the only part of the program that the Board of
Governors of the College reviews very carefully.  (All the
operational aspects are managed by the Commissions on
Laboratory Accreditation).
     There are only five standards, each one of which is
about a paragraph long, and then there is explanatory
material that goes along with each.  I will mention a few of
the highlights.
     The first is that it vests the responsibility for the
management of the laboratory with the director.  We have
very specific personnel requirements for the director.
     A director must be a physician or a doctoral level
clinical scientists, an individual whose education and
laboratory training is appropriate for the span of

-------
                             697
disciplines that are represented in that particular
laboratory.
     And the director must discharge a whole series of
responsibilities that you would find appropriate for any
director to discharge:  ensuring that there are sufficient
numbers of people and the proper equipment and the quality
control and the safety requirements and the educational
aspects of laboratory medicine.
     What we find is that in a number of commercial
laboratories, for example, or in hospital laboratories where
the medical director is not an employee of the laboratory
but is a consultant, the director does not have sufficient
authority to discharge those particular responsibilities.
Those laboratories do not meet the Standards.
     We also have personnel standards for employees other
than the director.  They are not as stringently defined.
They are, in fact, performance standards.
     Quality assurance is an important standard.  We make a
distinction between quality assurance and quality control.
Quality control we define as those intralaboratory practices
that contribute to the accuracy and timeliness of the
clinical laboratory data.  Quality assurance includes
quality control, but it also includes the pre-analytic
issues and the post-analytic issues all of which together

-------
                             698
will impinge upon accuracy and  timeliness in the healthcare
system.
     There  is  a  standard  that describes the need for
adequate  facilities  and operational procedures.  The
laboratory  must  participate in  the laboratory improvement
programs  of the  College,  that is  to say, on-site inspection
and proficiency  testing.
     That is a picture of the cover of the book in its
current edition..
     Now, our  organization.  At the top is the Board of
Governors of the College, and there are four councils.  The
council to  which the laboratory accreditation reports is the
Council on  Clinical  Pathology.  There are two commissions
within that council.  The Commission on Scientific Affairs
includes  16 scientific resource committees each of which is
devoted to  a specific clinical  laboratory discipline.
     A second  commission  is the Commission on Laboratory
Accreditation.   The  Commission  includes 14 regional
commissioners  and 4  additional  special commissioners.  To
•support us  are 61 State commissioners.  Some States, such as
New York, California, Ohio,  and Pennsylvania, have more than
one State commissioner, and the State commissioners have
approximately  1,800  inspectors  working for them.
     Now, there  are  about 17,000  pathologists in the
country.  About  11,000 belong to  the CAP.  So, that means

-------
                             699
that somewhere around 10 percent of all pathologists or
about 15 percent of our membership serve as inspectors for
the program.
     Our inspection cycle is a two-year cycle.  It begins
with the date of the first successful inspection.
     In alternate years, the laboratory will undergo an on-
site inspection and a self-inspection.  The self-inspection
is a very serious inspection.  The director uses exactly the
same tools as an on-site inspector would use.  The self-
inspection is followed by the same post-inspection data
entry routines to record the results of that inspection.
The accuracy of the data is verified by the next year's on-
site inspector.  If the laboratory does complete the self-
inspection with appropriate thoroughness, it is a very
egregious deficiency indeed.
     About 120 days before the lapse of accreditation, the
laboratory director is sent a reapplication packet of
materials.  He or she has 30 days to fill that out and to
return it to the program office in Northfield, Illinois.
That gives us about 90 days, then, to find an inspector, for
that inspector to get together an inspection team, and for
                    t,
the inspection to be scheduled and completed.
     I can't emphasize enough how important it is for all
deadlines to be met for this whole process to stay on
schedule.  Host of the problems that we have had running

-------
                             700
this program have been due to delays, and we have had to
adopt a "get tough" policy to stay on the inspection
schedule.
     The application packet that the director completes
includes an extensive questionnaire.  Specific information
is requested.  A list of all equipment, all of the tests
that the laboratory is doing, and the volume of those tests
have to be appended.
     We ask the director to insert description of the
quality control and the quality assurance programs that are
operative in that laboratory.
     We will get a floor plan of the facility, and we ask
for a diagram table of organization that depicts the
reporting relationships.  The latter is particularly
important in trying to identify sites in which there is a
titular director only.  The credentials of the key
personnel, (the department heads, section heads, and the
technical supervisors) are included.
     To that information, the Central Office in Northfield
adds the results from the last two inspections, (last year's
self-inspection and the previous year's on-site inspection),
                                     o
all the computer commentary that was generated from those
inspections, all relevant correspondence and, a statement of
surveys participation.

-------
                             701
     By "surveys", we are talking about the proficiency
testing program of the College.  The laboratory must
participate in this program.  It is interesting to note that
about 10 or 20 percent of all of the  subscribers to our
proficiency testing program, never send in their results for
evaluation.  So, it is important that the inspector know to
which surveys the laboratory has enrolled in and does have
data for even if that data was not submitted to the program
office for evaluation.
     The whole packet, is shipped to the inspector.  We will
try to choose an inspector who is a peer of the laboratory
director.  That means that we will try to identify an
individual from a facility of a similar size who can relate
to the kinds of problems and patient care issues that the
laboratory director is facing, somebody whose scope of
practice is comparable.  We will not, for example, send a
pathologist who limits his practice to cytopathology to a
clinical laboratory that is involved only in toxicology and
therapeutic drug monitoring.
     One walks a fine line, however, because some of the
peer review considerations make it difficult to preserve
objectivity, especially when one must worry about
competition.  For example, to inspect a proprietary
laboratory, peer review considerations would suggest that
the most appropriate inspector would come from another

-------
                             702
proprietary laboratory:  but, to do so might require
selecting an individual from a very long distance away so
that competitive interests will no interfere with
objectivity.
     Once the director has applied for accreditation, all
the materials have been received, and the inspector has been
assigned.  It is now up to the inspector to schedule an
inspection.  He or she will do so at a date that is mutually
convenient with the director.
     If it is a very large laboratory, the inspector will
want to bring a team of individuals.  A typical team is
about 5 or 6 people.  There may be a chemist, a
microbiologist, a couple of pathologists, and perhaps a
computer expert.
     We have very specific rules for the qualifications of
the inspector, but the selection of the inspection team
members is entirely at the discretion of the inspector.
This allows as much flexibility in the program as possible.
 The individuals who come are appropriate to the special
needs of that particular laboratory.  At a very big facility
like Mayo Medical Laboratories or Metpath, we will have to
take 12 or 15 people to do the laboratory well.  Most of us
who do inspections try to get the whole thing done in one
day simply to minimize the amount of time that we take away

-------
                             703
from our own practice.  So, to do so often requires
generally a fairly large team.
     Then the inspector completes the checklists and returns
them to the Centra Office.   We have a checklist of
questions that right now numbers around 2200 items, and they
are enclosed in about 15 different booklets each of which
focuses upon a different discipline within the laboratory.
     The checklists are what give our inspection its
structure.  They standardize the program from region to
region, from State to State, and from inspection to
inspection.
     Each of the questions have a value, and we talk about
Phase I and Phase II items.  A Phase I item is the less
serious of the two.  In some ways, you can think of it as a
recommendation.  A Phase II item is a requirement for
accreditation.  In fact, the laboratory must document
correction of any Phase II deficiency before accreditation
will be conferred.
     We do sometimes deny accreditation on the basis of
Phase I deficiencies if there is a very large number of them
.or if they have been recurrent and the director has shown no
interest in correcting them.
     Each item can be answered as "yes", as "no", or as "not
applicable".  It is the "no" items that drive the next part
of the process:  the generation of a checklist commentary.

-------
                             704
Each question has a commentary.  The commentary, which  is
generally quite a bit longer than the question, explains the
meaning of the question.  Appropriate journal references are
included.
     The specific set of checklist commentaries then is sent
back to the director who must respond in writing to each
deficiency, documenting  correction of the Phase II
deficiencies.  The director has 30 days to prepare this
reply.    It is reviewed then by the regional commissioner
who may decide that there is sufficient information there to
render a decision or may ask for supplementary information.
Once there is sufficient information, the recommendation can
be to accredit the laboratory, or the recommendation may be
to deny accreditation.
     The Commission as a whole meets three times a year.
Our first order of business is always to accredit all those
laboratories for which accreditation has been recommended.
Then we spend the rest of our time talking about the
individual facilities for which  denial has been recommended
by the regional commissioner.
     There is another document that is very important to the
process.  That is the inspector's supplemental report.  This
is a narrative, confidential document.  The inspector can
use this to amplify upon any items in the checklist.  And if
the inspector senses conflict between, let's say, the

-------
                             705
director and the medical staff or between the laboratory and
the hospital administrator, the issue should be described in
the inspector's supplemental report.
     The director will not see the information in this
particular report, as a general rule.  The regional
commissioner may  extract information from the report for
inclusion into a personal letter that goes to the director.
Especially when the commission is considering denial of
accreditation, the feelings that are expressed in the
supplemental report can be very important in shaping the
course of the discussion.
     Now, let us discuss proficiency testing.  (We call that
the "surveys program" or the CAP Interlaboratory Comparison
Program).  At the present time, there are over 90 different
packages that are offered, including clinical chemistry,
hematology, microbiology, toxicology therapeutic drug
monitoring, histocompatibility testing, cytogenetics, and
forensic pathology.  New ones are being produced in
molecular biology and in andrology.
     Each one includes a set of analytic tests.   There are
more than 300 analytes available.  An accredited laboratory
must participate in the surveys program according to its
repertory and complexity.  A large laboratory will have to
subscribe to many surveys, and a sophisticated laboratory

-------
                             706
will have to choose the comprehensive versions of certain
surveys.
     We had mentioned the resource committees briefly before
as being within the Commission on Scientific Affairs.  You
will remember that they are independent of the Commission on
Laboratory Accreditation.  The resource committees write the
survey specifications, and they evaluate the results.  Each
reported result can be scored as acceptable or not
acceptable.
     The unfavorable results are printed in a report that
goes to the regional commissioner once every three months,
called the surveys results exception reports.  In it we look
for evidence of a possible systematic problem.  All
laboratories will have certain problems from time to time,
but if it is a systematic problem, it will be recurrent.
     We will often ask the laboratory director to check into
a particular analytic test.  In fact, I send out about 100
such letters every three months.
     You can see how important it is that the laboratory
participate appropriately, because, in fact, the commission
is using the PT program as its mechanism for monitoring the
performance of laboratories between on-site inspections.
     Now, each of these resources committees is composed of
nationally recognized experts.  They are the scientific base
for the College, and all sorts of scientific questions are

-------
                             707
referred to them.  For purposes of this discussion, they
write the survey specifications and evaluate the results,
and they also develop the technical portions of the
checklists that we are using out there in the field, a very
important contribution to the program.
     We have an appeals process.  The appeals process, as
you might expect, will follow denial of accreditation.  The
director has 30 days to file the appeal.  The appeal board,
in fact, is the Commission on Laboratory Accreditation as a
whole.  The director is invited to come to the next meeting
of the commission to plead his or her case.
     If the commission affirms its original decision, then
that particular appeal can be extended to the Board of
Governors of the College, which is the final appeal board.
     What are the qualifications for our inspectors?  The
inspector must be a Fellow of the College.  Now, since
fellowship in the College is limited to board-certified
pathologists, that means that the inspector will be a board-
certified pathologist.
     It is also required that the inspector be affiliated
with an accredited laboratory, and that he or she has
undergone appropriate training.
     Now, the training is, in most cases on-the-job.  The
experienced inspectors will take a prospective inspector
along on two or three inspections to learn some of the

-------
                             708
operational aspects, but we are really not depending upon
the inspector for operational support.  We are depending
upon the inspector for expertise in evaluating a laboratory.
     We also conduct new-inspector workshops where we train
inspectors from scratch.  I almost did not make this
particular meeting, because I was originally scheduled to
host one inspector's workshop in Cincinnati this evening.
One of my fellow commissioners very graciously is covering
for me at that particular event.
     We also have update workshops that we hold periodically
throughout the country to refresh and revitalize  the
inspectors who have been doing these inspections for many
years.
     Like any big program, it has a set of policies and
procedures that codify all of the operational rules.  Having
such policies in writing, ensures fairness and uniformity.
     Our policies and procedures are on a three-year review
cycle.
     Computer support is essential to the program.  I cannot
emphasize enough how important computers are to us in this.
For example, we must hold to the inspection cycle.  If I had
to keep track of all of these laboratories in my own region,
I would fall way behind.  That would not serve the
inspection process well at all.  The computers keep the
process moving.

-------
                             709
     The whole surveys program is run on computer.  Most of
the survey results are analyzed by computer.  We talked
about the survey results exception report by which the
survey results are monitored by the commission.  What I did
not mention in that in that the report is not a simple
listing of the unacceptable results.    There is a fairly
sophisticated algorithm that selects for us those
laboratories that appear to have a higher probability of a
systematic bias.
     Of course, all of our. inspector lists and other
demographics are kept on computer.  Each of the regional
offices has a terminal to access the Chicago computer.  The
checklists and the commentaries are all on computer in  a
desktop publishing type of a format.   This allows us to
update these rules every 6 to 12 months to keep them current
with technologic change.  That can be a real challenge as
new instrumentation and methodology comes  onto the market.
     Lastly, let me just make a couple of comments about
finances.. This is a self-supporting program.  The revenue
comes from the annual subscription fees that the
laboratories pay.  Now, the surveys program also generates
revenue for the College, but we also incur significant
expenses in the manufacture of the PT materials.
     The expenses that are generated by these revenues
primarily are for personnel support.   We have about 75 full-

-------
                             710
time employees in the program office in Chicago who are
supporting the lab improvement programs of the College.
There are also part-time staff, such as the secretaries in
the regional office.
     There are direct expenses that one incurs with each
inspection (travel,  lodging, meals, and such), and those
are all charged to the program.
     What is not charged to the program is the time of the
individuals who are working on its behalf, and this is
substantial.  We have some solo practitioners, for example,
who work as inspectors, and when they go out once a year or
twice a year to do an inspection, they, must  hire a locum
tenens pathologist to come in and cover their practice for
them.  They pick up those costs themselves.
     So, our inspectors and our team members are not paid,
at least not by the College.  They may very well, of course,
be paid by the sponsoring organization that recognizes that
these people bring back to their laboratories, more new
information than they can confer onto the inspected
.laboratory.  So, it is a two-way laboratory improvement
program.
     The voluntary contribution to the program really cannot
be overestimated.  It is immense when you include 1800
inspectors, 75 commissioners, approximately 200 scientists

-------
                             711
that are on these resource committees, and then all the
officers of the College.  All are volunteers.
     So, in those few words, I hope I have conveyed kind of
an overview of the whole program.  In the panel discussion,
I will be happy to follow up on any items that you care to.
     Thank you very much for your attention.

-------
                          712
              Laboratory Accreditation Program
 College of American Pathologists
 325 Waukegan Road Northfield. Illinois 6OO93-275O 7O8 /446-88OO

Characteristics
voluntary
peer-review
must cover entire laboratory
based on Standards of Laboratory Accreditation
combines inspection with proficiency testing

Standards  for  Laboratory  Accreditation
• Vests the responsibility in the Director
• Personnel standards
• Quality assurance
• Adequacy of facilities and procedures
• Participation hi CAP Laboratory Improvement Programs
Organization
• Council on Clinical Pathology
  • Commission on Scientific Affairs
    16 Resource Committees
  • Commission on Laboratory Accreditation
    * 13 Regional Commissioners
      • 60 State Commissioners
         • ~4,100 laboratories
         • -1,800 inspectors

-------
                          713
Inspection  Cycle
• Begins with date of first successful inspection
• On-site and self-inspections in alternate years
• Reapplication begins 120 days before date of expiration
• All deadlines must be met to stay on schedule!


Application  materials
• questionnaire
• list of all equipment and scope of analytic testing
• statement of quality control and quality assurance methods
• floor plan
• diagram of table of organization
• credentials of key personnel

Application  packet  also  includes
• All results from the last two inspections
• Statement of Surveys enrollment
• Relevant interim correspondence
Choosing  an  inspector
• Peer-review considerations
  • facility of similar size
  • similar scope of practice
• Objectivity must be maintained
  • non-competitive practice

-------
                        714
Process
• Director applies for accreditation
• Application materials are reviewed for completeness
• Inspector is assigned by state commissioner
• Application packet is sent to the Inspector
• Inspection is scheduled, arranged, and conducted by the
 Inspector and his/her team
• Inspection materials are returned to the Program Office
Process  (ccnt'cH
• Checklist Commentary is printed and mailed to the Director
• Director responds to each deficiency in writing with
 appropriate documentation
* Reply is reviewed by Regional Commissioner
• Accreditation may be recommended or not recommended
• Commission votes on recommendation
 Inspector's  Supplemental  Report
 • confidential document
 • may amplify individual checklist items
 • appropriate location for sensitive information

-------
                         715
Surveys  Program  (CAP  Interlaboratorv
Comparison  Program
> 90 proficiency testing surveys
> 200 analytes
Participation is required for laboratory accreditation.


Monitoring  of  Proficiency  Testing
• Resource Committee evaluate results
• quarterly reports go to Regional Commissioners (Surveys
 Results Exceptions Reports)
• importance of appropriate participation
Resource  Committees
• Each disciplinary committee is composed of nationally
 recognized experts.
• Write survey specifications
• Develop checklists
Inspector Training
• Required for all new inspectors
• Recommended periodically for all inspectors
• Conducted throughout country

-------
                         716
Policies  and  Procedures
• codification of all operational rules
* special importance in a distributive program
• three-year review cycle

Appeals
• Follows denial of accreditation
• Director has 30 days to file
* Appeal heard by
  • entire Commission on Laboratory Accreditation
  * Board of Governors
Computer  support
• inspection cycle
• Surveys monitoring
• inspector lists
• updating checklists and commentary
Financial
• self-supporting
• revenues come from
  • annual subscription
  * Surveys program
• expenses include
  • full-time staff in Program Office
  • part-time staff in regional offices
  direct expenses incurred in performing inspections
• Inspectors and team members are not paid.The voluntary
 contribution  to the Program is immense!

-------
                             717

                                   MR. TELLIARD:  What we
would like to do now is there are five panelists.  The sixth
one is at National Airport in the fog.
     Each of the panelists is going to make a brief
presentation.  Some of them you have heard before.  Then
when they are done with each presentation/ we will open it
up to the panel.
     So, I guess Al is the first one.

-------
                             718
               PANEL PRESENTATIONS
                              DR. TIEDEMANN:   I am Albert
Tiedemann.  I am the director of the Virginia Division of
Consolidated Laboratory Services which is a central State
laboratory system, providing most of the State Laboratory
Services.  Except for Highways and Transportation, none of
the agencies have their own labs.  We are their
laboratories.
     We feel we are in the middle where we can recognize
both sides of problems in approving laboratories.  We are
subject to receiving approval from several Federal agencies.
We are also responsible for some different types of programs
for which we must give approval within the State,  drinking
water being the biggest.  But these programs include
venereal disease, driving under the influence of alcohol and
drugs, and commercial blood banks.  So, we feel we see both
sides of the picture.
     Since we are primarily in an environmental meeting
here.  I'm talking about the drinking water program.
     States actually are EPA surrogates if they have
primacy, that is, they conduct the laboratory approvals in
the States for EPA.  I am a little disappointed to look at
the attendance list.  There must be only about half a dozen
people from the States here today, because, basically, as

-------
                             719
acting for EPA, we are sort of under the gun.  The
laboratories in the States look to us for information as
well as the piece of paper that says you are approved.
     In Virginia, internal Virginia, we divide the
laboratories into two classes.  One is the government-owned
laboratories, and most of these are water works.  The large
ones which produce more than 3 million gallons a day have to
have at least an on-site approved microbiology lab.   We do
not charge fees for government laboratory approvals.  That
is funded by the general fund.
     The others are the commercial profit-making
laboratories for which we are not funded by taxpayers'
money.  Therefore, we do charge a fee.  We break the fee
into four categories:  microbiology, inorganic chemistry,
organic chemistry, and radiology.  Depending on how many
categories the lab wants to be approved for, the fees will
range from $200 to $680 every three years.  It is a three-
year cycle.
     In running the approval program, the basic requirement
is we have to be at least as stringent as EPA.  But within
that overall guideline, we try to take an attitude or
approach of being flexible, practical, and realistic.  That
is, we don't try to mandate a lot of very specifics.  Do
what you have to do with some flexibility in the way you do
it.  Keep good records, but we don't tell you how.

-------
                             720
     Our on-site visits analysts do not have to come back
and find something in that visit that they can report as a
deficiency.  I am very pleased when I can sign a report that
says no problems, no deficiencies, no recommendations, lab
is doing excellent work.
     You do find deficiencies, of course.  We class them
major and minor, and we base an approval for the lab overall
on an evaluation of all of its pluses and minuses, its past
performance, and the results it has obtained on proficiency
evaluation samples.  Basically, we want to know if we
approve a lab that that lab has shown it can produce valid
data.  That is the bottom line.
     Reciprocity.  We do participate by giving reciprocity
to out of State laboratories.  We have three conditions.
The first is they must show us that they have a need for
certification because of the business they have in the
State.  We found that too many labs just wanted something on
their letterhead, and it takes a lot of time to keep the
records.  So, they have to have business in Virginia.
     Secondly, they have to provide us with the
documentation of their approval by EPA or by another primacy
State.
     Third, because the number seems to keep
increasing...several years ago, we had to hire some part-
time clerical people to help with paperwork and had other

-------
                             721
costs...they pay an annual fee of $100 per category, that
is, the four categories I mentioned.  I don't think there
are probably many that pay more than $200 or, at the most,
$300.
     Because of the various time cycles in other states, we
do require that reciprocal certification be renewed each
year.
     We have some problems with terminology.  We talk about
certifying labs, but we are not really certifying labs.  If
we do, we are in trouble,, because if we take the word
certification in its essence and say that lab is certified,
that means we are guaranteeing that that lab's work is
always accurate, we are assuring that it is absolutely
correct.
     So, what we are really doing is saying we approve the
lab as being able to do good work or we are accrediting the
lab.  So, really, this term certification needs to be
examined, because it could possibly, the way the lawyers get
into things these days, come back to haunt the certifier.
     Let me mention a few problems.  I've noted that there
are only a few state professionals here.  I've gotten data
from a number of states.
     A key problem is information that EPA is considering
discontinuing supplying Performance Evaluation (PE), quality
control, and standard reference samples.  I say

-------
                             722
"considering: because program people in other states and
others are concerned over the lack of positive, direct
information from EPA Washington.   We are receiving only
rumors or bits and pieces of information from EMSL, the
regions, and/or third-parties.
     The term used is privatization.  EPA has yet to define
officially what they mean by privatization.  We've received
no plan, no details.
     We hear that EPA will stop providing these samples.  We
hear laboratories will have to purchase the samples.  Will
we be able to buy the samples from any vendor?  Who will
certify the quality?
     One rumor says that EPA will designate a sole source
vendor.  Many state procurement laws require competitive
bidding.
     One third-party source, which seems to be a valid
source, stated that in addition to being a sole source
vendor, that vendor would pay EPA a royalty on each sample.
Therefore, EPA would reduce costs and also generate revenue.
     What is the true story?  We don't know.  But someone in
authority in EPA Washington needs to provide direct, factual
information to the states.  The states then need to transmit
this information to the laboratories for which the states
have approval responsibility.

-------
                             723
     A typical problem of failure to coordinate that exists
today.  If you are in the drinking water area/ you are
probably well aware by now that EPA several months ago
approved the new COALERT procedure for microbiology testing
for coliform.  We can't approve anybody to use that
procedure.  We can't use it ourselves, because EPA says the
method is approved.  But EPA hasn't come out and said what
are the laboratory criteria, what does your laboratory have
to have in the way of everything from QA to detailed
procedures before EPA will accept the data.
     This is like the other things that pop up.  There is a
new requirement, a new parameter, something new to be
regulated, but the analytical procedure doesn't exist, at
least not in an approved form.
     Last July the Journal of Environmental Laboratories
published a series on laboratory certification.  Bill says
this Certification problem keeps coming up.  It keeps coming
up because the problems are still there.  They are not
really changing.

                                   MR. CARTER:  I am Mike
Carter.  I work for EPA in the Superfund program.  More
specifically, I am one of the managers of the contract lab
program.

-------
                             724
     A lot of you know more about that program than you
might want to.  For those of you who are somewhat unfamiliar
with it, we are the primary source of analytical data for
the Superfund program.  That means we are responsible for
the analysis of 80,000 to 100,000 samples a year.  Our
budget is, in round numbers, $45 million per year for
analytical services.  That includes both quality assurance
activities and the actual payment for the analyses.
     We are frequently referred to as a de facto
certification program, and I suppose that is an apt term.
Before contracts are awarded to laboratories, there is an
accreditation step.  Normally, we do not award a contract to
a laboratory without an on-site visit and a finding that
they have adequate personnel, facilities, procedures, et
cetera.
     I said normally.  The Small Business Administration can
choose to issue a certificate of competency which we are
obligated to accept.  Our experience says that, in many
cases, the laboratory was actually given a disservice when
it got that certificate, because they do have lots of
trouble performing.
     During the course of the contract, we have a great deal
of oversight.  That includes, normally, four performance
evaluation samples per year.  They are intended to be double

-------
                             725
blinds.  Most people recognize them right away.  So, at
best, they are single blinds.
     And we do have a certification activity.  The former
speaker's concern about nomenclature we happen to share.  We
accredit a laboratory, but then we exert fairly extensive
and intensive efforts in product certification.  By that I
mean virtually ever data package is inspected at least once.
     We a great deal of in-house activity.  We have computer
programs that are checking for compliance with our
deliverable requirements.  We utilize close to $3 million a
year in computing services just maintaining data bases and
cranking through these inspections.
     We would like to see and hope to move toward more self-
certification in which we do less and the laboratories do
more so that we get out of kind of vicious circle where we
get data, find some defects, notify the lab, give them a
chance to rectify it.  This is all taking time and effort.
In the lab's case, efforts cost.
     Another critical part of any credible accreditation
program was mentioned earlier by Ms. Whalen, and that is for
a program to be...and Dr. Hoeltge...for a program to be
credible, there has to be some provision for a
deaccreditation or a decertification.  We have historically
had administrative type actions that were, in essence, at
least an interim deaccreditation.  We would stop sending

-------
                             726
samples to laboratories that were having certain problems,
technical or through-put.
     Some of those decertification efforts have recently
become a whole lot more critical, sensitive, attention-
getting.  At the moment, we have three different audits
underway by EPA's Office of Inspector General and somewhere
in excess of ten investigations underway by the Inspector
General's office.
     There was an article in a journal that I had never
before.  Most of you have probably never heard of it before,
either.  It is called the Legal Times.  The week of April
23rd of this year, it was on the front page.  It wasn't at
the top of the page, but it was on the front page.  The
headline said, "Superfund Effort Jeopardized by Suspect
Data."  It goes ahead to quote our Assistant Inspector
General's investigation that says at least ten contractors
are under investigation for potential fraud.
     There have been three formal actions taken to date with
only one of them being final.  The final action was
regarding Roy F. Weston Company.   Basically, they signed a
settlement agreement with the government in which they
agreed to pay $750,000, and one of their labs was then
voluntarily suspended for 4 to 12 months from doing any work
for Superfund.

-------
                             727
     I must note here that the company has stated that they
did not and do not admit any wrongdoing.  It was strictly a
settlement.
     Subsequent to that, we had two laboratories, companies,
that were suspended under suspension and debarment
activities which are part of contract regulations.  A
suspension means that that entity is not allowed to compete
from any Federal Government work from any Federal agency,
and it probably extends to any cooperative agreement with a
State in which Federal money has been passed on to a State.
     An important little consideration here.  Suspension is
intended to be an interim action and may lead to debarment
which is a longer-term, even more final type determination.
The other important note is that the suspension does not in
any way hold up the investigations that also precede it.
     We now have inspectors general come to a lot of our
program meetings just to make sure everybody knows the
environment we are operating in.  The fact of the matter is
that if data are falsified or changed, there is the
potential that three felonies have been committed.  The
falsified data that is submitted to the government is
considered a false statement.  To then send a bill to the
government in which you are asking for payment for these
falsified data is a false claim.  And this one is a nice

-------
                             728
catch:  if you send in your invoice or you get your check
back through the mail, that is mail fraud.
     Each one of those felonies is punishable, as they tell
us, by up to five years in federal prison and a $10/000
fine.  So, falsify some data and, potentially, somebody is
looking at 15 years in jail and $30,000.
     The IG does tell us that indictments for felony charges
are in the works and probably will be applied to both
companies and individuals.  So, we are in an entirely
different world in terms of decertification.  That is a
fairly stringent decertification of a company or an
individual.
     So, we are still considered a de facto certification
program, and I guess it would be fair to say that a lot of
elements that are part of the certification program are
active in our program.  Certainly, it is a whole new world
for us and for a lot of people that have been working for
us.

                              MR. PERLER:  I am Arthur
Perler from the Environmental Protection Agency Office of
Drinking Water.  I suppose it is the folks that bring you
laboratory certification for drinking water.  I don't know
if that is popular or not in this room.  I have never spoken
to this particular crowd before.

-------
                             729
     I just want to review briefly the authorities and the
status of our program and some of the directions we are
thinking about moving in I feel you all might be interested
in.
     Our basic program of laboratory certification derives
from two elements of the Safe Drinking Water Act.  The first
one you see, we feel, gives us the basis for establishing
analytical methods and all things related to doing the
analyses concerning drinking water compliance, compliance
under the Safe Drinking Water Act, and the key words there
are, of course, are including quality control and testing
procedures to ensure compliance.
     Recently with the passage of the Lead Contamination
Control Act, an additional kicker was added that we must, in
a way, provide a program to ensure that the laboratories
testing for lead are providing reasonably good results.
     From that flows certain national primary drinking water
regulations which apply to the public water supplies which
have to provide the compliance analyses and provide those
analyses either in their own labs or through commercial
laboratories or State laboratories and certain requirements
also on State offices for the approval of those
laboratories.
     EPA's lab certification program and requirements are
basically, as most of you know, first of all, specified in

-------
                              730
 the laboratory certification manual the third edition of
 which is at the printers right now and would have been
 available at this meeting/ but I think it is just maybe a
 week or two behind schedule.  Some of you have advance
 copies of that.
      This manual is used as the official policy of EPA
 within its own part of the decentralized program that we
 have for operating laboratory certification.  That is that
 EMSL-Cincinnati, the EPA's headquarters laboratory program,
 certifies EPA regional laboratories.  EPA regional
 laboratories then certify a principal State lab in each
 State, and then it is up to those State labs either to do
 all the analyses or to certify the commercial laboratories.
      So, EPA is not in the business of certifying
 laboratories.  For the business that we are in which is
 certifying regional labs and principal State labs, we use
 the lab certification manual as our official policy, and the
 States, we understand, for the most part pick up most
 elements of that manual as they certify commercial labs or
. the labs with which they do business or do business in the
 States.
      State requirements then are pretty much again use of
 the lab cert manual, utilization of the EPA PE program, and
 then on-site evaluations either through State auditors or
 third party auditors.

-------
                             731
     Let me say just quickly in response to what Al said
about PE samples, I will add to the rumor mill and tell you
what I know about it.  On the privatization or the program
under which people may have to pay for a more limited number
of PE samples, it is not expected to apply, I don't believe,
to PE samples in its initial stages.  What is under
consideration now are basically the QC samples and the check
samples that EPA provides, and I don't believe that PE
samples in what is known as the water supply programs will
be covered for quite a while.  There is just too much
backlash against that, and I believe that the program that
we will be looking at won't be a sole source or a program
that drives up the cost.
     Let me point out also that there is nothing in the
drinking water regulations that requires that PE samples be
provided by EMSL, and there are a number of private entities
and some States that have contracted with those private
entities to provide samples except, let me say, for the
trihalomethane regulations.  The rest of the national
primary drinking water regulations allow you to use a
privately generated PE sample.
     Our biggest problem and really what I wanted to focus
on, although time is running fast, are the problems as we
see them with laboratory certification.  First of all, since
it is a decentralized program, there are divergent State

-------
                             732
requirements which we know now lead to a difference in
acceptability and a difference in reliability in the results
generated by laboratories in different States.
     I don't want to call it a difference in quality,
because I am sure everybody is out there trying to generate
the highest quality data, but because we all in this room
understand reliability in terms of, let's just say,
precision and accuracy, there is a difference in that
because of the different ways that States implement their
programs.
     These differences have led to limited reciprocity,
differing fee structures, we believe higher costs, and
limited availability of labs.  That bothers us somewhat,
especially since the number of drinking water regulations,
the number of required analyses, and the number of drinking
water systems required to perform them is going up and will
be growing exponentially over the next 10 or 15 years.
     The PE and audit requirement portions of the program
provide us with a little heartburn also.  Simply performing
on a PE sample acceptably once a year is no measure of how a
laboratory does on routine samples.
     I point your attention to an article in the April-May,
I guess it is, issue of Environmental Lab, the cover story
by Stan Blacker, cover article, which represents a position
which EPA and, I believe, the Office of Drinking Water may

-------
                             733



be moving towards.  In that article, he points out that



laboratories that perform perfectly acceptably on PE samples



and that even have reasonably good precision and bias still



have relatively high probability of making a false positive



or a false negative compliance determination because of the



interval that precision implies and the statistics that are



involved.



     And the result of that is exactly what Blacker says.



We believe that especially when environmental results are



close to standards, there axe probably an unacceptably high



frequency of compliance determination errors, either false



positives or false negatives being made.



     Now, we were criticized also fairly severely for



allowing a program to proceed which does have such limited



reciprocity and without any ability for the private sector



to prosecute the same kind of a program.  So, we have had on



the books a number of years, although it has been largely



ignored by most States, policies which do encourage



reciprocity and third party, and I am here to try to do that



some more.  I will take the opportunity to do that whenever



I can.



     The benefits of that, of course, are that there will be



shared capacity, and many States have deficient numbers of



laboratories.  There will be increased competition.  We



think it will reduce costs, especially since we are getting

-------
                             734
 into  regulating things  which have  quite expensive analytical
 methods  associated with them.
      And,  also,  according to the possibility of reciprocity
 and some other program  changes  which  I don't want to address
 today, we  feel that  laboratories will be able to adopt new
 methodologies  more quickly.  £et me say that it is not that
 the laboratories will be able to adopt them; EPA will accept
 the adoption of  laboratories of newer methodologies more
 quickly.
      These are not meant to  be  alternatives.  These are just
 various  things that  we  are considering and  some
 possibilities  of changes in  the program.
      We  would  hope that any  policy that we  would put forward
 would allow more uniform State  requirements...would end up
 in more  uniform State requirements in order that we have
 more  uniform reliability and probability of making correct
 compliance decisions in the  drinking  water  program.
      We  would  encourage those States  that have...and I
 understand there are not many States  here to listen...that
'many  States that have changed their statutes and their
 regulations to provide  programs which are less amenable to
 reciprocity to change those.
      We  are considering a complete overhaul of the system,
 as I  said.  We expect to come out  with proposed regulations
 which would more likely provide evidence that laboratories

-------
                             735
are performing within acceptable standards on a routine
basis.  Again, I would direct your attention to the general
revisions that Stan Blacker discussed in that article,
revisions that would attach to each piece of data the
appropriate quality information, and those kinds of changes,
by the way, could very well end up with environmental
regulations, especially under the Safe Drinking Water Act,
no longer even specifying the required analytical
methodology, that kind of requirement being unnecessary
within that type of program.
     Now, that certainly is a program that only assesses
quantitative analysis, and there are qualitative aspects of
that program as well as on-site inspections and audits which
I think we will maintain.
     As Bill and others have talked about, the agency has
gone into high gear on its EMMC which is an attempt to unify
quality assurance and quality control in all of the programs
and an attempt to unify, where possibly, analytical
methodologies, and that is something that I hope you will
all stay aware of, and as different solutions are proposed
or put somehow into either the Register or put forward in
meetings like this that you will comment on those.
     Let me skip to the last bullet and say that all of .
these changes we are trying to move as quickly as possible
on and we look for data and support from the laboratory

-------
                             736
community, because I think that the last thing that we both
want to happen, both within the Office of Drinking Water
Programs and from, I think, the laboratories' side is some
direct Congressional action which would alter the program,
because, actually, we don't want to end up where we feel
that we are being forced to make changes.  The last changes
that were forced on us that way resulted in drinking water
regulations which we think probably went a little bit too
far.
     So, that is the end of what I have prepared, and I
think we have questions at the end.

-------
                     737
    THE 13th ANNUAL EPA CONFERENCE
                     on
ANALYSIS OF POLLUTANTS IN THE ENVIRONMENT
   Laboratory Certification & Reciprocity
                Arthur H. Perler
                Baldev L. Bathija
            Office of Drinking Water

-------
                738
LABORATORY CERTIFICATION & RECIPROCITY
 DRINKING WATER LABORATORY  CERTIFICATION
   •  STATUTORY  REQUIREMENTS
   •  CURRENT SDWA REGULATIONS
 LABORATORY  CERTIFICATION REQUIREMENTS
   •  EPA
   •  STATE
 ADVANTAGES OF  INTERSTATE LAB CERTIFICATION
 DIFFICULTIES WITH RECIPROCITY
 POSSIBLE  SOLUTIONS
                                     NRFK-002

-------
                           739
 CURRENT STATUTORY REQUIREMENTS
SAFE  DRINKING WATER ACT  -  Section 1401(1)(D)

    "  .... criteria and procedures  to assure a supply
    of drinking water  which  dependably complies with
    maximum contaminant levels,- including quality
    control  and testing procedures to insure
    compliance with such levels ..."

LEAD  CONTAMINATION  CONTROL ACT  -  Section 4

    "EPA  shall  assure that programs for certification  of
    testing  laboratories which test drinking water supplies
    for lead contamination certify  only those laboratories
    which provide reliable accurate testing."
                                                NRFK-003

-------
                      740
     CURRENT  SDWA  REGULATIONS
• Section  142.10(b)(4)

    "Assurance of the availability to the State  of
     laboratory facilities  certified by the Administrator
     and capable  of  performing  analytical measurements
     of  all  contaminants specified by the State primary
     drinking  water regulations."

• Section  141.28 -  Approved Laboratories

 (a)  "For the purpose of determining Compliance  ....
      samples  may be considered only  if they  have
      been  analyzed  by  a  laboratory approved  by
      the State ...."
                                                NRFK-004

-------
                         741
CURRENT SDWA  REGULATIONS  -  Cont.
 •  Section 142.10(b)(3)(i)

 "The  establishment  and maintenance  of  a  State
  program  for the certification of laboratories
  conducting  analytical measurements  of  drinking
  water contaminants ..."

 "The  requirements  of  this paragraph  may  be
  waived  by  the  Administrator for  any State
  where all analytical measurements .... are
  conducted  at laboratories  operated  by  the
  State and  certified by the Agency."
                                            NRFK-005

-------
                    742
  EPA LAB  CERTIFICATION REQUIREMENTS
•  Specified in  Lab Cert Manual
     • Third  Edition - Published - June  1990
•  Manual  is  Official Policy for:
     •  EMSL Certification  of  EPA Regional Labs
     •  Regional Certification of Primary State  Labs
•  Manual  is  Guidance for:
    •  State  Certification of Local Labs
                                            NRFK-006

-------
                     743
STATE LAB CERTIFICATION REQUIREMENTS

  • EPA Lab Cert Manual
  • PE Samples
      • EPA, State,  Commercial
  • On-Site  Evaluations
      • State Auditors
      • Third-Party Auditors
                                       NRFK-007

-------
                   744
INTERSTATE LABORATORY CERTIFICATION
 GAO Report Criticized Lack of Reciprocity Among
   States  Resulting in High  Costs  of  Multi-State
   Certification

 ODW  Policy Encourages  Reciprocity  Among States

 ODW  Policy Encourages  Use of Third-Party Auditors

 Benefits  of Inter-State Certification

    •  Shared Capacity
    •  Increased  Competition
    •  Reduced  Costs
    •  Faster Adaptation of Latest Methodologies
                                            NRFK-008

-------
                        745
   INTERSTATE LABORATORY CERTIFICATION
              (Perceived Hindrances)
• State's Rights -  DW Primacy  Regs
• Acceptable On-Site  Evaluations
    •  Qualifications of  Other State Auditors
    •  Qualifications of  Third-Party  Auditors
• Restraint  of Trade
    •  Selection  of  Third-Party  Auditors
* Differing State Laws  -  Due Process
                                           NRFK-009

-------
                 746

   RECIPROCITY AMONG STATES
         (Limited Information Available)

• Reciprocity Possible
      AL, GA, IL, Ml, MN, MO, MT, NO, ND,
      NV, NM, Rl, SC,  TN,  TX, WA, Wl, WV
• Reciprocity Not Permitted
      AR, CA, FL, LA,  KS,  KY, MA, ME,
      NH, NJ, OH, OK, OR, PA,  UT
                                        NRFK-010

-------
                          747
            POSSIBLE SOLUTIONS
•  Uniform State  Requirements  - ODW Survey  of States
•  Third-Party  On-Site Evaluations  Sent to States  -
   State Lab Cert Officer Decides  if  On-Site by State
   Team is  Necessary
•  Encourage States  to Change Statutes
•  National  Federal Regulations
•  Complete Overhaul of  ODW System
      • New QAMS Approach to QA/QC
•  EPA Environmental Monitoring Management Council  (EMMC)
•  Consortium of Quality Environmental Data (CQED)
•  Congressional  Action
                                                 NRFK-011

-------
    INDUSTRY PERSPECTIVE
     FOR A NATIONAL PLAN
FOR LABORATORY ACCREDITATION
              by
         GEORGE STANKO
                                 00
           J]
        SHELL DEVELOPMENT CO.
          HOUSTON, TEXAS

-------
    CONCERNS WITH CURRENT SYSTEM
             OF ACCREDITATION
 NUMEROUS STATE ACCREDITATION PROGRAMS
A.   COSTLY FOR LABORATORIES
B.   CUSTOMER PAYS THE PRICE
C.   REDUNDANCY BECAUSE OF NUMEROUS PROGRAMS
D.   LABORATORIES SUBJECTED TO FAR TOO MANY AUDITS
E.   IMPROVEMENTS IN DATA QUALITY ARE MARGINAL AT BEST
F.   GENERAL LACK OF EFFECTIVE PERFORMANCE EVALUATION
    PROGRAMS
G.   LABORATORY CUSTOMERS HAVE LITTLE UNDERSTANDING
    FOR CURRENT ACCREDITATION  PROGRAMS
H.   EXPERIENCES WITH SOME ACCREDITED LABORATORIES
    WAS DISAPPOINTING IN SPITE OF ACCREDITATION
<£>

-------
  CONCERNS WITH CURRENT SYSTEMS
           OF ACCREDITATION
INDUSTRY (CUSTOMERS) EXCLUDED FROM
CURRENT ACCREDITATION PROGRAMS
   A.   ACCREDITATION PROGRAMS FUNDED BY INDUSTRY
   B.   NO CONTACT WITH INDUSTRY
       1.  PROBLEMS
       2.  HOW PROGRAMS COULD BE IMPROVED.
       3.  EFFECTIVENESS OF PROGRAMS.
01
o

-------
  WHO SUPPORTS NATIONAL ACCREDITATION OF
        ENVIRONMENTAL LABORATORIES?
1,   UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
2,   AMERICAN COUNCIL OF INDEPENDENT LABORATORIES
3,   MOST COMMERCIAL LABORATORIES
4,   INDUSTRY:
    A, CHEMICAL MANUFACTURERS ASSOCIATION
    B, AMERICAN PETROLEUM INSTITUTE
    C OTHERS
U1

-------
 ELEMENTS OF STATE CERTIFICATION PROGRAMS
1,   MANDATED BY STATE LEGISLATION
2,   REQUEST FROM LABORATORIES FOR CERTIFICATION
3,   CERTIFICATION REQUIRES FEES FROM LABORATORIES
4.   ON-SITE INSPECTION BY  STATE AUDITORS
5,   DEMONSTRATION OF LABORATORY PERFORMANCE WITH
    EPA WP/WS PE SAMPLES
6,   MUST MEET CRITERIA FOR PE SAMPLES
7,   PERIODICALLY  RECERTIFIED
8,   SUBJECT TO STATE ENFORCEMENT ACTION
9,   SOME PROVISIONS  FOR RECIPROCITY BUT
    SELDOM PRACTICED
UI
to

-------
 MAJOR ROADBLOCKS TO NATIONAL ACCREDITATION
        OF ENVIRONMENTAL LABORATORIES

 ,   STATES RIGHTS
    A,   CERTIFICATION AUTHORITY
    B,   ENFORCEMENT ACTION
    C,   STATE PROGRAMS (PEOPLE)
2,   RECIPROCITY
    A,   ON-SITE VISIT/AUDITING
                                            U)
    B,
CERTIFICATION  FEES
    C,   CERTIFICATION

-------
SOLUTION TO CURRENT ROADBLOCK FOR NATIONAL
   ENVIRONMENTAL LABORATORY ACCREDITATION

1, FORGET ABOUT EC92 AS THE DRIVING FORCE,
2, CONSIDER ONLY ENVIRONMENTAL LABORATORY ACCREDITATION
3, IDENTIFY THOSE ELEMENTS OF CURRENT PROGRAMS THAT ARE
   ONEROUS
    A, CERTIFICATION FEES
    B. MULTIPLE ON-SITE VISITS/AUDITS
    C, INCONSISTENCIES BETWEEN STATE PROGRAMS
    D, GENERAL LACK OF RECIPROCITY
    E, OTHERS
01

-------
SOLUTION TO CURRENT ROADBLOCK FOR NATIONAL
   ENVIRONMENTAL LABORATORY  ACCREDITATION
4,    FORM A NATIONAL COALITION  FOR ENVIRONMENTAL
     LABORATORY ACCREDITATION
     A, REPRESENTATIVES FROM USEPA, STATE AGENCIES,
       ACIL, INDUSTRY
     B, PREPARE A LABORATORY ACCREDITATION  GUIDANCE
       DOCUMENT BASED ON ASTM/ISO PRACTICES
          a. ASSESSOR CHECKLIST
          b, QUALIFICATIONS OF  ASSESSORS
          c, PROTOCOL FOR ACCREDITATION RECOGNIZED AND
           ACCEPTED BY ALL
     C. LET STATES RETAIN THE RIGHTS FOR PROGRAMS
          a. COLLECT CERTIFICATION FEES
          b, RIGHT TO VERIFY ASSESSORS REPORTS AND
           ACCREDITATION
          c, RIGHT OF ENFORCEMENT ACTION
          d, RIGHT FOR ADDITIONAL PERFORMANCE EVALUATION
          e, MODIFY STATE LEGISLATION TO ALLOW FOR
           ALL THE ABOVE
Ut
in

-------
             OTHER CONSIDERATIONS
2,
    IF WE DON'T DEVELOP A WORKABLE PROGRAM,
    CONGRESS WILL BE FORCED INTO MANDATING
    ONE STATES  WILL HAVE TO ACCEPT  AND WE
    MAY NOT LIKE,
BEFORE NATIONAL ENVIRONMENTAL  LABORATORY
ACCREDITATION IS POSSIBLE THERE IS A NEED
FOR STANDARDIZATION OF METHODOLOGY AND
ANALYTICAL PRACTICES ACROSS ENVIRONMENTAL
PROGRAMS,
                                                      CTt

-------
                             757



                                   MR. TAMPLIN:  We have



been asked to keep this short,  so no slides, no overheads.



We will try to keep you all bright-eyed and bushy-tailed.



     I am Ben Tamplin from the  State of California,



Department of Health Services,  Division of Laboratories.  I



am Chief of the Sanitation and  Radiation Laboratory.  We,



also, are a centralized laboratory service, although there



are a few laboratories in other agencies.



     We have a long history of  environmental laboratory



accreditation going back to the late 1940s when the Porter-



Cologne Water Quality Control Act was passed.  In 1951,



drinking water laboratories came under the purview of the



Health Department, and this laboratory approval program grew



at the request of Regional Water Quality Control Boards to



include wastewater laboratories as well.  Wastewater



laboratories were dropped in 1981, because the State Water



Resources Control Board could no longer afford to support



their part of the program.



     In 1985, Hazardous materials laboratories were required



to become accredited in a separate program.   A third



program exists in California, but it is really a program of



registration by the California  Department of Food and



Agriculture for laboratories doing pesticide analyses on raw



produce.

-------
                             758
     California began consolidating these programs several
years ago with the passage of two bills in our legislature,
one which brought together drinking water laboratories,
wastewater laboratories at the request of the State Water
Resources Control Board, and hazardous materials
laboratories.  Just last year, a bill was passed that
brought the Department of Food and Agriculture's
registration program under our purview, but full regulation
of these laboratories won't really occur until about 1992.
     The old drinking water and the hazardous materials
laboratories' regulations stay in effect in California until
the Environmental Laboratory Accreditation Program
regulations are in place.  We expect this to occur by the
end of the year, but recognize that it may require more
time.
     It should be no surprise that this combined program
follows the systems that were used by the drinking water and
hazardous materials programs, that is, that the
certification was granted upon approval of an application,
upon approval of a quality assurance plan, upon successful
participation in a performance evaluation study, and
following an uneventful site evaluation.
     One other thing:  there is the payment of a fee, and
that check is the very first step in the accreditation

-------
                             759
process.  When one sends in an application, the check comes
right off the top and in it goes to the cashier.
     Currently, the basic fee in California is $913.28.  For
each field of testing a fee of $411.34 is added.  These fees
are adjusted on July 1 or shortly after July 1 of each year
by the Department of Finance.  It is, in a sense, a COLA.
     Incidentally, we levy this fee every year.
     The main problem before us right now is completion of
our regulations package.  We are getting help here...and we
think we have done the right thing from an ad hoc committee
that is made up of representatives of the American Council
of Independent Laboratories, the Association of California
Testing Laboratories, the California Association of Public
Health Laboratory Directors, the California-Nevada Section
of the American Water Works Association, and the California
Water Pollution Control Association.
     We are ready for public hearings now, and we expect
these regulations will be in effect by January 1, 1991.
     The lack of regulations has forced us to use existing
regulations and has resulted in a very large backlog of
inspections.  We have to address this backlog in the next
three months.  We can give interim accreditation without a
site evaluation so long as we get the money.
     But we need to simplify our process and consolidate our
fields of testing.  We presently have 23 fields of testing.

-------
                             760
That is absolutely ridiculous.  We used to have the same 4
that Virginia has, but in the writing of our statute, people
got carried away, and we now have, as an example, separate
fields of testing for organics by GC/MS for drinking water,
for wastewater, and for hazardous materials.  Now, that is
really ludicrous, because, as you know, once you get the
lumps out of wastewater, the analysis is just as easy as
that for drinking water.
     We have to find a way to simplify accreditation,to get
back to where we were.  That may be difficult.
     We also feel we have to have an effective system of
reciprocity.  The requirements for analyses in California
are soon going to outstrip the State's laboratory resources.
We need to accredit out of California laboratories, non-
California laboratories.
     We need to accredit them in order to provide a resource
for some of the programs that are in the wings.  We have an
omnibus pesticide bill, a pesticides in food bill, which
will add perhaps 40 people to our laboratory staff looking
for pesticide residues in raw and processed foodstuffs.  I
don't know how many analyses this is going to take State-
wide, but our share will be immense.
     Also, we have something going in California now called
the Big Green Initiative.  It is the environmental plank of
our Attorney General's candidacy for Governor.  It is an

-------
                             761
all-encompassing environmental protection act.  It covers
food safety and pesticides, agricultural worker safety,
greenhouse gas reduction, ozone layer protection, commercial
and residential tree planting, estuarine and ocean water
protection, water quality protection, and marine resources
and human health standards.
     We don't have enough laboratories in California to do
the monitoring that will be required under this initiative.
We have an initiative system in California in which, if you
don't like what the legislature is doing, you may go to the
ballot box through the initiative process.  This particular
initiative came into the Secretary of State's office with
two times the required number of signatures.  It is
estimated in the polls to be running 4 to 1 in favor.  It is
expected to pass and we will need laboratory resources, so
we will have to go outside.
     Also, some of our California laboratories are divisions
of companies with laboratories in other States, and some of
them have different capabilities.  At the present time,
there is no way they can trade samples back and forth to get
the biggest bang for their buck.   That is to say, both
laboratories must be approved in California before the
movement of samples back and forth can occur, and that is
all wrong.  We are not supposed to be restraining trade.

-------
                             762
     Many California laboratories are accredited in other
States, some in 12 or more States, and I am certain there is
a very high cost involved for them to comply with all the
regulations.  It is particularly costly in the area of
proficiency testing.  Here we really need an effective
centralized system of performance evaluation.
     Our current regulations allow us to recognize other
programs, and we have told other laboratories if you are in
the EPA program and you are analyzing Water Supply or Water
Pollution Evaluation Samples or both, that is fine.  We will
accept those data.  So far our legislature has gone along
with this approach.
     The reason we have recognized EPA PE Studies is that we
really don't want to make our own PE samples.  We don't want
to create an empire.  We think it is wasteful when a program
like EPA's already exists, a program that really does work
pretty well.  We do criticize a lot, but we also recognize
that this performance evaluation program is a good one, and
it needs to stay in place.
     We are ready in California to join other States and EPA
in establishing some sort of nationwide centralized PE
system.  We have authority to charge laboratories for these
samples, and even if we didn't, we could finance it
ourselves out of the fees that we charge for accreditation.

-------
                             763



     We are happy that EPA has reversed at least part of its



stand on the phase-out of some of its PE activities, and we



call upon you at EPA to take a leadership role and help us



establish a national program.  If you want to come out to



California to discuss it, we would be happy to have you.



Just bring rain.







                                   MS. PREVOST:  I am



Margaret Prevost from the New York State Health Department,



the Environmental Laboratory Approval Program which is the



program that certifies all environmental labs in New York



State and out of the State.



     I think you just heard the West Coast.  You are going



to hear the East Coast.



     In our program, we certify labs in four categories,



drinking water, wastewater, solid and hazardous waste, and



air emissions.  We have a little under 850 certified labs.



Close to 300 of those are out of State labs.



     We make our own proficiency samples.  We have never



used EPA's, probably for a lot of reasons.  One of them is



that our program really started, although it had a



history...we certified drinking labs since the early



1950s...but the ELAP program to make it short, as it is



called, became official in late 1985, and about at that



time, there was already a little bit...I wasn't with the

-------
                             764
program then, so excuse me if I am making a mistake...but
there was concern about that perhaps EPA was going to cease
making samples.  So, we do make our own proficiency samples.
     A laboratory in our program is inspected once a year
and proficiency tested twice a year for what they are
certified for.
     I spent most of the day trying to get here.  You had
already started when I got here, so I don't know what the
newest rumor is about the PT samples and maybe it has all
been clarified now, but we received a number of calls from
other States when the announcement first came that the
wastewater samples were no longer going to be available.  We
are able to provide them to other States if that is
necessary.
     It sounds like now that isn't going to be necessary,
but because we have the whole system in place and it has
been in place for five years.
     I think we could go toe to toe with California trying
to explain our fee system.  If there are any New York labs
here/ they would agree with me.
     We had one fee system.  It changed as of April 1,  the
beginning of our permit year.  We now have a base fee of
$500 plus an analyte fee for every analyte you are certified
for plus a volume fee paid on the volume of your analysis of
those analytes you are certified for on New York State

-------
                             765
samples only.  The legislature came up with this one; we
didn't.
     But none of our laboratories are in revolt yet, because
the New York State budget has not passed.  It is now 41 days
overdue, and I can't send out any bills until they pass the
budget.
     I think one of our concerns...and we hear it from our
laboratories.  We have very active laboratory associations
in New York.  One of the things is we are very aware that a
large lab that is certified in many States spends a great
deal of time and a great deal of money.  Actually, they were
less concerned about our fees...they didn't like our fees,
but they were less concerned about that than the idea of the
time and money that is spent on doing proficiency testing,
inspections, people come to inspect.
     They would like that resolved.  We understand that.  In
fact,  the legislature just passed a bill saying that we are
going to have an advisory board, a 7-member advisory board,
and its mission, among other things, but its first mission
is to discuss reciprocity and come up six months after it is
formed with a report to the legislature.
     One of the problems...and I think this is worthy of
discussion...is the fact that as a northeastern State where
we have a lot of pollution...and I think it is true that
different regions have different problems... our legislature,

-------
                             766
our Governor, and our Commissioner of Health have concerns.
So, how it is done on a national certification program or
the PT samples.
     For instance, with EPA, it is 7 VOCs.   In our State, a
lab has to be certified for 52 VOCs.
     So, I don't think it is that easy to put together
something that would be universal, because then you would
still have these subprograms, because you know how States
are.  You never go backwards.  You only add.  We don't have
the Big Green coming down which sounds awful, but they never
stop.  I mean, you just keep testing for more and more.
     It is the argument that we have to be this stringent
and we go along with anything that is as stringent or more
stringent.  I am not"quite: aware of any other State...there
probably is...that does annual inspections, but we do, and I
don't think the Governor or the Commissioner of Health is
going to back down on that.
     So, there are a lot of issues, but we see the need for
some type of reciprocity.
     I think it is very important for your contiguous States
especially, even if it is almost sort of a regional.  So
many labs are on borders; large labs go right over.
     I don't know the answer.  I do think that if one of the
things is the proficiency test, if that can be addressed.

-------
                             767



My concern is on having some sort of uniform proficiency



test.



     What do you do about the individual State requirements



like ours where you have to do 52 VOCs?  How is that going



to be addressed?  How will that prevent the labs from having



just two levels, three levels, four levels?  Other States



probably have specific things they want tested.  I don't



know how that will be addressed.



     Thank you.







                                   MR. STANKO:  I would like



to present to you the industry perspective towards national



laboratory certification.  We will go through some of the



concerns we in industry have.



     I am not going to elaborate on any one of these



specific things.  The main thing is we have concerns with



the numerous State accreditation programs.  Other people



also have concerns, and these are some of the items that



lead to that concern:  costly for laboratories; the customer



pays the price.



     We are still bottom line people.  Essentially, all the



certification fees that States are charging, the laboratory



user has to pay it, and we are the laboratory user.



     Another major concern that we have is that industry is



left out of the entire system completely.  We, industry, are

-------
                             768
the customers of these contract labs more or less getting
data for regulatory purposes, and nobody from the State
agencies ever asked us/ are there problems with the program,
how could the program be improved, and are the programs
effective?  In other words, has it done anything for data
quality?
     The first question I am going to ask is who really
supports national accreditation for environmental
laboratories?  I think the United States Environmental
Protection Agency does and the American Council of
Independent Laboratories does.  I think most of the contract
laboratories do and certainly industry does.  The Chemical
Manufacturers Association the American Petroleum Institute
have gone on record as supporting national accreditation of
laboratories, and there are others.
     On item number 5, I needed 50 question marks, but they
wouldn't quite fit.  But in fairness to the States, I think
you have to look at what are the elements of all the State
programs, and this could be difficult to summarize.  There
is no way anyone can go through all 50 certification
programs and do a good job of it.
     What I have done is I have isolated some of the things
that are common to all of them.  Most of them are mandated
by some sort of State legislation.  They also require
request from the laboratories for the certification.  In

-------
                             769
other words, the State requires that the laboratories ask
for certification in that State so that certain requirements
have to be met.
     Number 3 you have heard a lot about.  Certification
requires fees from the laboratories.  No new taxes, huh?
     On-site inspection by State auditors, demonstration of
laboratory performance with the EPA WP and WS samples, must
meet criteria for PE samples, periodic recertification, as
we have heard, and they are subject to State enforcement.
     There are some provisions in all of these certification
programs for the big R word, reciprocity, but it is seldom
practiced.  If it is practiced at all, most of the State
programs rely on WP and WS samples, and that is about where
the reciprocity is.
     Now I would like to show you major roadblocks to
national accreditation of laboratories.   Number 1 is State
rights.  Once you give States the right to do something, you
can never get it back.  They have this certification
authority.  Someone has granted that authority, and now it
is difficult to take it back.
     A lot of States feel very strongly on enforcement
action, and here I agree that the States should enforce this
type of a program.
     Number 3, the State programs.   Most people think that
if you have a national accreditation program, all the State

-------
                             770
people involved in their own certification programs lose
their jobs.
     The big R word.  On-site visiting and auditing.  Every
State feels that they have a need to send their auditors to
that laboratory to inspect it.  Some of the laboratories
tell me that half of their staff is involved in nothing but
working with auditors.  The other half do the samples.
     We have heard a little bit about certification fees to
five significant figures at one point, and then who grants
certification?
     These are some of the major roadblocks, and now I am
going to do something unusual.  Instead of being critical
and criticizing which Bill Telliard tells me I am good at, I
am going to try to see if I can give some solutions to our
current roadblocks.
     Many of us thought that EC92 would be the driving force
for a national accreditation of laboratories.  Many of us
thought that the environmental labs that we deal with could
go in on the coattails of those product testing laboratories
that are going to be affected or impacted by the EC92
regulations.
     The first thing we have to do is forget EC92.  Just as
clinical laboratories are different from environmental
laboratories, the product testing laboratories associated
with EC92 are entirely different from the environmental

-------
                             771



laboratories.  That is one of the major flaws we have to



address.



     Consider only environmental laboratory accreditation.



That goes along with divorcing ourselves from EC92.  Groups



like CQED and IAETL have tied themselves with EC92.  I think



they are going to have to step back and divorce themselves



from EC92.



     What can we do to identify those elements of the



current program that are really onerous to us?   We don't



like certification fees, but just like I said, no new taxes.



Have you ever seen anybody give anything back yet?  I don't



think we stand a chance on that.



     Multiple site visits and audits.  I think this is one



area where if we have reciprocity we can really do away with



some of this nonsense of having 50 States going to audit



each single laboratory.  I think this is one area where we



can have reciprocity.



     There are lots of inconsistencies in the State



programs.  Here again, I think we have to look to the



national EPA to try to come up with some kind of a guidance



so that we can have consistency between the State programs.



     There is general lack of reciprocity.  We have heard



that R word enough times from a number of speakers.  I don't



have the solution to that.

-------
                             772
     I think we need to form a national coalition for
environmental laboratory accreditation and environmental
laboratory accreditation only.  I think we need
representatives from the USEPA, the State agencies, ACIL,
and the industry as well.
     I think we need to prepare a laboratory accreditation
guidance document based on ASTM and ISO guidelines.  They
are out there.  We don't have to make these up.  They are
available.
     We need to come up with an assessor checklist that is
uniform and acceptable to all State agencies as well as the
Federal EPA.  We need some sort of guidance and some sort of
qualifications for the assessors so that no matter which
assessor is used, all States and the Federal EPA and
industry and the laboratories will recognize the assessor as
being qualified and being able to conduct these audits.
     I think we need a protocol for accreditation recognized
by all and accepted, and this is where I think we look to
the national EPA to look at the State programs and see what
we have in common and come together with something that all
the States can accept.
     The politics involved says we are going to let the
States retain the rights for some of these programs.  The
collection of certification fees, I don't see any answer to
that.  The right to verify assessor reports and

-------
                             773



accreditation.  Here, again, I think the States should have



that right.



     I think the States ought to maintain the right for



enforcement action, and I think the States ought to have the



right for additional performance evaluations if they are not



willing to accept WP and WS or, like the last speaker said,



they have their own program.  Shell has its own program.  I



think the States have to be recognized and that they can do



something additional or accept WP and WS.



     The last thing, I think, is to modify the State



legislation to allow for all of the above.



     This slide has nothing to do with the actual



accreditation.  There are other considerations.  If we don't



develop a workable program, Congress will be forced into



mandating one that the States will have to accept and we may



not like at all.  I also am giving a pitch for the EPA EMMC.



We also need to address the 518 report issue.  Before a



national environmental laboratory accreditation program is



possible, I think there is a need for standardization of



methodology.



     Thank you.

-------
                             774



                QUESTION AND ANSWER SESSION



                              MR. TELLIARD:  Thank you,



George.



     Now, for any of you who would like to ask questions or



make comments, all of the panel is available at this time.



Would anybody like to start?



                              MS. ORDONA:  I am Alicia



Ordona from Virginia, and I want to address this one to the



California representative.  What is your main objection to



reciprocity?  I mean, at least New York has said about VOCs,



but what is the State of California's main objection to



reciprocity in certification for drinking water?



                              MR. TAMPLIN:  We have no



objection.



                              MS. ORDONA:  You don't?



                              MR. TAMPLIN:  No.  We have the



authority in the statute.  We don't have a mechanism in the



regulations yet.  That will probably be about the first of



the year.



                              MS. ORDONA:  In that case, I



think I am going to recommend to my boss that we are not



going to write reciprocal certification for laboratories



from States that don't write reciprocal certifications from



Virginia.

-------
                             775



                                   MR. LEVY:  Nathan Levy



with A&E Testing in Baton Rouge.



     George, thank you.  I wish that everybody in the agency



and in the State governments had your feelings.  I think



most of us contract labs do have your feelings, and I guess



my question becomes to anybody who wants to take it.



     Why not?  Why can't we do it?  Why can't we take one



person from every State and a few from the EPA and lock them



in a room and tell them they can't come out until they have



a program?  Why not?



                                   MR. FARRELL:  Well, that



gentleman just stole my thunder.  I am Jack Farrell from



Enseco.



     George, I would also like to extend my thank you.  I



don't think it could have been said better.



     I don't really have a question.  I have a couple of



comments I would like to make.



     The first one is, very few of the programs that were



talked about and very few of the programs that I know of,



excluding the CLQ, deal with more than a capability



discussion of what a laboratory can do.  We need to add



ongoing performance to whatever program or programs come up.



     The second comment I would like to juake is while we



have half a dozen approval programs up there, there has to



be at least 100 of them out there.  There is also starting

-------
                             776



to be a lot of different groups that are focusing on this



accreditation issue, and I applaud that.



     The request that I have is...and, Bill, maybe you can



help or somebody can help...let's bring it all together so



that it is one group that is addressing this issue and



coming up with one consensus standard.  If there is any way



that IATA or CQED or anybody can help do that, that is what



they are there for.



     Fifty groups looking at fifty different certification



programs is only going to cost us a lot more.



                              MR. TELLIARD:  Anyone else?



                              MR. HOLT:  I am Phil Holt from



Occidental Chemical.



     I would like to preface my remarks by saying I am in



favor of lab accreditation.  I think it has accomplished a



lot, but as a matter of curiosity...and I will address this



to New York since that is a program I am familiar with...I



would like to know what we are accomplishing with the



ongoing programs.



     We started five years ago.  What percentage of



laboratories did we weed out as being not acceptable five



years ago, and on an ongoing basis, are we continuing to



weed out bad laboratories, or has it kind of leveled out now



and only good laboratories are participating?

-------
                             777
                                   MS. PREVOST:  Very
honestly, only because I got a question a week ago from the
New York Times or I wouldn't really be up on this, you are
right.  Everybody is suddenly getting interested in this.
Just out of the blue, the New York Times wants to do a
series on environmental laboratories.  They haven't done it
yet, and I wonder how they will manipulate what information
I gave them.  It is always a matter of concern.
     Since the program won't be give years old, actually,
until November of 1990, 342 labs have...records have not
actually been kept of why 342 labs that were in the program
are no longer in.  I think I should preface this by saying
before this, except for the drinking water labs which was a
voluntary program...
     By the way, New York State's laboratory ELAP program is
a voluntary program in that wonderful way they like to
define voluntary.  The fact is that no laboratory can do
work for the State of New York or any of its political
subdivisions which means anything that is paid for by public
monies must be done in a certified lab, but if you want to
limit yourself to the private sector or just to drinking
water in private homeowners' wells, you don't have to be
certified.  Obviously, most people are certified because the
public sector money is very important.

-------
                             778
     Now, this was a group of laboratories that had never
been regulated before, the environmental labs in New York
State.  Actually, the mission of the program at the time was
to ensure there were competent labs for environmental
analysis.  It wasn't really to take labs in and throw them
out.
The whole point of it was to sort of put on a layer of
standards or regulations and then...and I do believe they
have been made somewhat more stringent over the last four
and a half years...the mission really is to try to bring
laboratories in compliance.  We don't really like to throw
them out.
     What they do do is many labs have withdrawn.  We
certify by specific analysis in New York what you want.  If
you fail your proficiency test enough times, you know,
finally some laboratories actually shrink down to have so
few analyses that they are certified for that they withdraw.
     But the real mission of the program was not to throw a
lot of labs out.  It was to bring them up to a level of
competency and, yes, it has leveled out.  It has leveled
out.  There are still labs that become withdrawn, as we like
to call it.  We withdraw them from the program or they
choose to be withdrawn, but certainly from the earlier, say,
the first two years, it has leveled out.

-------
                             779



     Of course, you cannot say that every lab in our program



is or very well...you do a proficiency sample.  They are the



best sample works that have been done.  I mean, you know,



you only have so much.



     We do have unannounced inspections in New York, so



nobody can kind of gear up for our inspections.



                                   MR. CARTER:  I would like



to make kind of a response to what you said and pick up on



something Jack Farrell said.



     In our program, we do accreditation, and I guess the



best way to say this is good labs have bad days.  In some



internal meetings that we have had to discuss some of these



implications of accreditation, one thing I would say that



most of us that have thought about it within EPA have a



concern about is that the purchasing public, industry, or



whomever really should understand that it is up to them to



verify that this capable laboratory performed on their



particular work.



     Like I say, the best labs and the best people have bad



days.  So, being accredited or certified does not assure



that the individual piece of data you got is exactly what



you wanted.



                                   MR. HOLT:  I would like



to ask one other piece of information of New York.  Since



things have sort of leveled out, is it time to start

-------
                             780
thinking about maybe annual testing instead of biannual or
if you are participating both in drinking water and
wastewater quarterly testing?
                              MS. PREVOST:  Did you say
quarterly testing?
                              MR. HOLT:  Yes.  The drinking
water and the wastewater were on opposite three-month
cycles, so you were getting samples in every quarter.
                              MS. PREVOST:  Oh, yes, right.
No, there has been absolutely no move by the Commissioner to
go to that type of...to reduce the number of proficiency
tests.
     Actually, if you will notice, we keep adding things
that we proficiency test for.  That seems to be, very
honestly, the way the direction goes in New York.  I am
being honest.
                              MR. TELLIARD:  Anybody else?
                              MR. PRONGER:  I am Greg
Pronger of National Environmental Testing.
     The first thing I would like to say is that there seems
to be an irony in the situation.  For the first time, the
USEPA has somebody coming up to them asking them to be
regulated, basically, and the USEPA is having huge problems
trying to figure out how to do this.  It is probably the
first time in history.

-------
                             781
     Normally, everybody is complaining about the EPA coming
in and regulating them.  Here is somebody looking to be
regulated and having trouble doing that.
     What seems to me one of the problems with getting
anything implemented is the level of discussion going into
it.  If something would be promulgated just to get a program
started, if the COP instead of...a problem that a lab has,
if you perform on a COP sample but your bid isn't low
enough, you technically aren't part of the program.
     If there were a two-tier level where you could pass a
performance sample, be a part of the program and not receive
a sample, but if your bid is not high enough not become part
of it and have to pay for that, the sites visits and such,
that would at least be a step in the direction where it is a
mechanism already established for going out checking on
laboratories, taking care of that.
     If you could then say you are a so-called COP lab and
just not receiving samples, it would address some of the
problems that the environmental labs are seeing, because the
COP program kind of initiated the problem in that the
laboratory was caught in the situation that you could pass
every PE sample, keep bidding on them, getting PEs, and
never become part of the program because your dollar bid is
too high.

-------
                             782
     Then, if a client would call and say no, are you a COP
lab, you would have to say no.  Even though you could pass
the samples forever, pass any type of inspection, you never
get any place like that.
     If something like that were started just to initiate
that and then build from that on trying to get reciprocity
from the States, trying to get some reciprocity between the
different programs, it would be a building block, and I
don't think anything will occur with this until there is
just some type of initial action taken.
     Thank you.
                              MR. CARTER:  One of the
problems with the de facto COP certification is that it is
not just a PE and accreditation system.  The fact that there
is this ongoing, literally daily, inspection of deliverables
means that we don't just do the quarterly PE samples.  In
essence, there is, like I say, almost a daily assessment of
lab performance, and we make decisions on lab eligibility
for samples under our program on a weekly basis.
     So, performance on periodic PE samples is a relatively
small part of the program.
     We have problems right now with requests from media,
various places, under Freedom of Information Act for PE
sample results.  What do you do about the lab that has
passed the most recent one but they failed the last four?

-------
                             783



Do we say yes, they passed the last one and they have done



one right out of four?  Is that really an accurate



assessment of the laboratory?



     The other problem is funding and authorization.  We



don't have any authority to sell our PE samples.  We are



prohibited by statute from giving away anything from



Superfund.



     So, as I say, there is a kind of an inherent problem in



our passing on our de facto certification outside our



program, and we find ourselves in a position where, due to



the various scrutinies we are under, we can't really cut it



back and accept other accreditations.



     I guess the best thing I can say is we understand the



problem.



     Recently, the Office of Waste Programs Enforcement



indicated they were considering making part of their consent



agreements with responsible parties a requirement that they



use COP laboratories.  We assured them that was improper and



at most what they could require is adherence to COP



protocols and COP reporting requirements.



     So, as far as I know, we have prevented them from



further contributing to the problems inherent in a COP de



facto certification.



                                   MR. TELLIARD:  I would



like to thank the panel.

-------
                          784

                              MR. TELLIARD:  I want to thank
you all for coining.  I would like to thank Jan Sears in the
blue dress in the back who makes all this happen every year.
Jan, would you like to stand up?
     I would like to thank the court reporters whose
proceedings we can't print until September, but we will.
     And I would like to thank you all for coming.  I hope
you have had a good meeting.  I hope to see you back here
next year, probably same time, same station.  Anybody who
has any ideas on subject panels or papers, please give me a
call or drop me a line.  We are always looking for new
material.
     Thank you so much for coming and thanks for your
attention.

-------
                                      785


                     13th  ANNUAL  EPA CONFERENCE  ON  ANLYSIS
                       OF POLLUTANTS IN THE ENVIRONMENT

                               LIST OF SPEAKERS
Mike Carter
USEPA, OERR
401 M Street, SW (WH-548A)
Washington, DC  20460
202-382-7909
Bruce N. Colby
Pacific Analytical
1989 B. Palomar Oaks Way
Carlsbad, CA  92009
619-931-1766
James A. deHaseth
Department of Chemistry
University of Georgia
Athens, GA  30602
404-542-1968
Jim Eichelberger
USEPA, EMSL
26 W. St. Clair Street
Cincinnati, OH  45268
513-569-7278
Thomas E. Fielding, Ph.D.
USEPA, ITD
401 M Street, S.W.
Washington, DC  20460
202-382-7156
Bettina Fletcher
USEPA, Region III CRL
839 Bestgate Road
Annapolis, MD  21401
301-224-2740
Warren Haltmar
Sr. Chemist
Texaco, Inc.
5901 S. Rice
Bellaire, TX  77401
713-432-2279
Dr. Gerald Hoeltge
Cleveland Clinic Foundation
9500 Euclid Avenue
Cleveland, OH  44195
216-444-2830
Larry D. Johnson, Ph.D.
Research Chemist
Source Methods Stand. Br. (MD-77A)
USEPA
Research Triangle Park, NC  27711
919-541-7943
Lawrence H. Keith
Senior Program Manager
Radian Corporation
P.O. Box 201088
Austin, TX  78720
512-454-4797
Jim King
Sample Control Center
Viar and Company
300 North Lee Street, Suite 200
Alexandria, VA  22314
703-557-5040
W. G. Krochta
PPG Industries
440 College Park Drive
Monroeville, PA  15146
412-325-5183

-------
                                     786
L. L. Lamparski
Midland Applied Science & Tech. Lab
DOW Chemical Company
Midland, MI  48640
517-636-2352
Theodore D. Martin
Chemistry Research Division
USEPA, EMSL
26 W. St. Clair Street
Cincinnati, OH  45268
513-569-8423
D. R. Mount
ENSR Corporation
1716 Heath Parkway
Fort Collins, CO  80524
303-493-8878
Arthur Perler
USEPA, ODW
401 M Street, SW (WH-550D)
Washington, DC  20460
202-382-3022
Margaret Prevost
Environmental Lab Approval Prg,
New York State Health-Dept.
P. 0. Box 509, ELAP-Room 299B
Albany, NY  12201-0509
518-474-8519
Joe C. Raia
Sr. Res. Chemist
Shell Development Co.
P.O. Box 1380
Houston, TX  77251-1380
713-493-7693
James K. Rice
President
James K. Rice Chartered
17415 Batchellors Forest Rd,
Olney, MD  20832
301-774-2210
Dr. Susan Richardson
USEPA
College Station Road
Athens, GA  30613
404-546-3199
Dale R. Rushneck
ATI - Colorado
225 Commerce Dr.
Ft. Collins, CO  80524
303-490-1511
George H. Stanko
Sr. Staff Res. Chemist
Shell Development Company
P.O. Box 1380
Houston, TX  77251-1380
713-493-7702
M. T. Stephenson
EPTD-Environmental
Texaco, Inc.
P. 0. Box 425
Bellaire, TX  77401
713-432-3329
Benjamin R. TampTin, Ph.D.
Chief, Sanitation &, Radiation Lab
California Dept. of Health Serv.
2151 Berkeley Way, Room 465
Berkeley, CA  94704
415-540-2201
William A. Telliard
Chief, Analytical Methods Staff
USEPA, ITD
401 M Street, SW (WH-552)
Washington, DC  20460
202-382-7131
A.W. Tiedemann, Jr., Ph.D.
Division Director
VA Div. of Consolidated Labs
1 N. 14th Street
Richmond, VA  23219
804-786-7905

-------
                                      787
Dr. Thomas 0. Tiernan
Director, Toxic Contaim. Res. Prg.
Wright State University
175 Brehra Lab, 3640 Col. Glenn Hwy.
Dayton, OH  45435
513-873-2202
Dr. Yves Tondeur
Triangle Laboratories
P. 0. Box 13485
Research Triangle Park, NC
919-544-5729
27709
Rhonda Whalen
Office of Survey & Certification
U.S. Dept. of Health & Human Serv.
2D2 Meadows East, 632 Security Blvd
Baltimore, MD  21207
301-966-6801

-------
788

-------
                                       789
                    13th ANNUAL EPA CONFERENCE ON ANALYSIS
                       OF POLLUTANTS IN THE ENVIRONMENT

                               LIST OF  ATTENDEES
Greig Aitken
County Court Reporters, Inc.
124 Cork Street
Winchester, VA  22601
703-667-0600
David J. Armstrong, Ph.D.
Southern Research Institute
P.O. Box 55305
Birmingham, AL  35255-5305
205-581-2000
Steve Azar
Supv. Env. Engineer
Atlantic Div. NFEC
Bldg. IAA Code 1811
Norfolk, VA  23511
804-445-1929
Clifford J. Baker
Laboratory Director
Continental Analytical Serv. Inc.
1304 Glendale Road
Salina, KS  67401
913-827-1273
Donald P. Ballard
Water Treatment Plant Leader
Naval Supply Center
Fuel Department, Code 700
Norfolk, VA  23512
804-484-6430
Louis B. Barber
Chief Chemist
Public Utilities
1400 Brander St.
Richmond, VA  23224
804-780-5338
Michael E. Barber
Laboratory Manager
Core Laboratories
1300 South Potomac St., Suite 130
Aurora, CO  80012
303-751-1780
Thomas Barber
Group Leader, Analytical Chemistry
CIBA-GEIGY Corp.
410 Swing Rd.
Greensboro, NC  27409
919-632-7297
John Barr
Lab Manager
City of Indianapolis, DPW
2700 South Belmont Ave.
Indianapolis, IN  46221
317-633-5429
Baldev Bathija
USEPA, ODW
401 M Street, SW (WH-550D)
Washington, DC  20460
202-382-3039
Bruce Bauman
American Petroleum Institute
1220 L St. NW
Washington, DC  20005
202-682-8345
Laura 0. Beach
Chemist
Acurex Corporation
4915 Prospectus Drive
Durham, NC  27713
919-541-1014

-------
                                     790
Mary Ann Becker
Chemist
USEPA, Region I
60 Westview Street
Lexington, MA  02173
617-860-4630
Robert G. Beimer
Program Manager
S-CUBED
3398 Carmel Mt. Rd.
San Diego, CA  92121
619-453-0060
Daniel W. Berisko
Vice President
Weerts Energy Associates
P.O. Box 3227
Johnstown, PA  15904
814-535-2992
Mark R. Bero
CLP Manager
IEA, Inc.
1901 N. Harrison Avenue
Gary, NC  27513
919-677-0090
Russ Bisping
Quality Assurance Office Code 130
Norfolk Naval Shipyard
Portsmouth, VA  23709-5000
804-396-9305
Daniel Bolt
Environmental Products Manager
Cambridge Isotope Laboratories Inc.
20 Commerce Way
Woburn, MA  01801
617-938-0067
Zvi Blank, Ph.D.
CHMM
E.C.R.A. Laboratories, Inc.
273 Franklin Rd.
Randolph, NJ  07869
201-361-4252

Howard Boorse
QA/QC Manager
Pacific Environmental Laboratory
9405 S.W. Nimbus Ave.
Beaverton, OR  97005
503-644-0660
Wanda Boyd
Louisiana State University
P.O. Box 20931
Baton Rouge, LA  70893
504-388-8521
Joel Bradley
President
Cambridge Isotope Laboratories Inc.
20 Commerce Way
Woburn, MA  01801
617-938-0067
Chris Bremer
Supervisor, Chromatography
Twin City Testing Corp.
662 Cromwell Ave.
St. Paul, MN  55043
612-641-9489
Anthony Bright
Laboratory Certification Coord.
OWRB
P.O. Box 53585
Oklahoma City, OK  73152
405-271-2580
Nancy Broyles
Chemist
Union Carbide Chem.  & Plas. Co.  Inc
P.O. Box 8361
South Charleston, WV 25303
304-747-4707
Douglas M. Brubeck
Analytical Chemist
Environmental Laboratories, Inc.
9211 Burge Ave.
Richmond, VA  23237
804-271-3440

-------
                                      791
Eugene A. Burns
Maxwell, S-CUBED Division
P.O. Box 1620
La Jolla, CA  92038
619-453-0060
Linda Carter
Sr. Sanitary Chemist
New York State Dept. of Health
Wadsworth Cntr. Empire State Plaza
Albany, NY  12201
518-474-0404
William H. Chambers
NEA, Inc.
10950 S.W. 5th Street, Suite 260
Beaverton, OR  97222
503-643-4661
David Chang, Ph.D.
Manager
Burlington Research, Inc.
P.O. Box 2481
Burlington, NC  27215
919-584-5564
Paul H. Chen, Ph.D.
Staff Scientist
Environmental Science & Eng., Inc.
P.O. Box 1703
Gainesville, FL  32602
904-332-3318
Elizabeth Chisholm
Lab Manager
ECO LOGIC
143 Dennis St., Rockwood
Ontario, Canada     NOB2KO
519-856-9591
Joe Chou
Organic Chemist
IL State Geol. Surv.
615 E. Peabody Dr.
Champaign, IL  61820
217-244-2744
Ellen W. Cobb
Analytical Chemist
Union Camp Corp.
P.O. Box 178
Franklin, VA  23851
804-569-4885
Sterley B. Cole
Supelco, Inc.
Supelco Park
Bellefonte, PA  16823
814-359-3441
Martin K. Collamore
Lab Supvr., Tech. Support
City of Tacoma, Public Utilities
2201 Portland Ave.
Tacoma, WA  98421
206-591-5582
Lee Collier, Ph.D.
Paracel Laboratories Ltd.
2319 St. Laurent Blvd. Unit 100
Ottowa, ON Canada     K1G 4K6
613-731-9577
Linda Crawford
BRAVN Environmental Laboratory
6800 S. TH-169, P.O. Box 35108
Minneapolis, MN  55435
612-992-4811
Robert E. Creekmur, Jr.
Analytical Chemist
Froehling & Robertson, Inc.
3015 Dumbarton Road
Richmond, VA  23228
804-264-2701
Mark Crews
Viar & Co./Sample Control Center
300 North Lee Street, Suite 200
Alexandria, VA  22314
703-557-5040

-------
                                     792
Jack Crissman
Supelco, Inc.
Supelco Park
Bellefonte, PA
814-359-3441
16823
Michael D. Crouch
President
ETC/TOXICON
3213 Monterrey Blvd.
Baton Rouge, LA  70814
504-925-5012
Brenda A. Cuccherini, Ph.D.
Associate Dir., Environmental Div.
Chemical Manufacturers Assoc.
2501 M Street, N.W.
Washington, DC  20031
202-887-1174
                        Amelia DaCruz
                        Chemist
                        Solutions Laboratories
                        814-H Greenbrier Circle
                        Chesapeake,  VA  23320
                        804-420-0467
Linda Darington
Organics Lab Manager
General Engineering Laboratories
P.O. Box 30712
Charleston, SC  29417
803-556-8171
                        Sandra Daussin
                        Viar & Co./Sample Control Center
                        300 North Lee Street,  Suite 200
                        Alexandria, VA  22314
                        703-557-5040
Rajesh Dave
Laboratory Director
Environmental Science & Engineering
217 Long Hill Crossroads
Shelton, CT  06484
203-926-9081
                        James E. Davis
                        Roy F. Weston, Inc. ESAT
                        Bldg. 209 Woodbridge Ave.
                        Edison, NJ  08837
                        201-548-1024
Ton Dawson
Group Leader R/D
Union Carbide Chem. & Plas. Co. Inc
P.O. Box 8361, Bldg. 770, Rm. 144
South Charleston, WV  25303
304-747-5711
                        Sue Deegan
                        ETS Analytical Services
                        2160 Industrial Drive
                        Salem, VA  24153
                        703-387-3995
Ivan B. DeLoatch
Chemist
USEPA Office of Drinking Water
401 M Street, SW
Washington, DC  20460
202-382-2273
                        Jessie Deluna
                        Chemist
                        Hampton Roads Sanitation District
                        1436 Air Rail Avenue
                        Virginia Beach, VA  23455
                        804-460-2261
Violeta Deluna
Water Chemist II
City of Norfolk/Dept.
6040 Waterworks Road
Norfolk, VA  23502
804-441-5678
      of Utilities
Kathy J. Dien Hillig
Manager, Ecology Analytical Serv.
BASF Corp.
1419 Biddle Ave.
Wyandotte, MI  48192
313-246-6334

-------
                                      793
Pam Dilsizian
Department of Labs & Research
Westchester County
Hammond House Rd.
Valhalla, NY  10595
914-524-5575
Jeffrey A. Dodd
S-CUBED, Div. of Maxwell Labs,
1800 Diagonal Rd., Suite 420
Alexandria, VA  22314
703-838-0220
         Inc.
Dr. Willard Douglas
Manager, Environmental Labs
Sverdrup Corp.
Bldg. 24-23
Stennis Spce Ctr, MS  39529
601-688-3155
Rolla M. Dyer
Professor of Chemistry
University of Southern Indiana
8600 University Blvd.
Evansville, IN  47712
812-464-1712
Dwight Easty
Laboratory Manager
James River Corp.
904 N.W. Drake St.
Camas, WA  98607
206-834-8318
Ronald J. Edgar
Chemist
Spokane County APCA
W. 1101 College Ave.
Spokane, WA  99201
509-456-4727
Room 230
Kenneth W. Edge11
Section Chief
The Bionetics Corporation
16 Triangle Park Drive
Cincinnati, OH  45246
513-771-0448
Nariman El Fino
Lead Organic Chemist
Froehling & Robertson
3015 Dumbarton Rd.
Richmond, VA  23228
804-264-2701
Chuck Emnett
General Lab Supervisor
Aptus Environmental Services
21750 Cedar Ave. South
Lakeville, MN  55044
612-469-3475
Anthony N. Enweze
EBASCO
2111 Wilson Blvd., Suite 1000
Arlington, VA  22201
David Evans
Quality Assurance Office Code 130
Norfolk Naval Shipyard
Portsmouth, VA  23709-5000
804-396-9305
Dave Fada
Metro-Seattle
322 W. Ewing St.
Seattle, WA  98119
206-684-2303
Gary Fa Hick
Waters
34 Maple Street
Milford, MA  01757
508-478-2000
Mike Filigenzi
Senior Scientist
Enseco-Cal Labs
2544 Industrial Blvd.
West Sacramento, CA  95691
916-361-6168

-------
                                     794
Edgar E. Folk, IV
Technical Officer
IEA, Inc.
1901 N. Harrison Ave.
Gary, NC  27513
919-677-0090
             Jim Forbes
             Lab Director
             Law Environmental,  Inc.
             112 Town Park Drive
             Kennesaw,  GA  30144
             404-421-3310
Peter Fowlie
Chief, Laboratory Division
Wastewater Technology Centre
P.O. Box 5050, Burlington
Ontario, Canada     L7R4A6
416-336-4633
             Drew Francis
             Chemist
             Hampton Road Sanitation District
             1436 Air Rail Avenue
             Virginia Beach,  VA  23455
             804-460-2261
William D. Frazier
Analytical Chemist
City of High Point, Central Lab
P.O. Box 230
High Point, NC  27261
919-883-3410
             Candace D.  Friday
             QA/QC Manager
             Keystone Lab - Houston
             3911 Fondren
             Houston, TX  77063
             713-266-6800
Robert E. Fuchs
President
Environmental Consultants, Inc.
391 Newman Avenue
Clarksvflle, IN  47130
812-282-8481
             Harry Gearhart
             Conoco
             P.O.  Box 1267
             Ponca City,  OK  74603
             405-767-4315
Peter Georges
Marketing Director
Environmental Science & Engineering
217 Long Hill Crossroads
She!ton, CT  06484
203-926-9081
             Noshi Gerges
             Philadelphia Naval  SY
             Philadelphia Bldg.  121
             Philadelphia, PA  19112
             215-897-3284
William E. Gillenwaters
Chemist
Newport News Shipbuilding
Dept. 031, 4101 Washington Ave.
Newport News, VA  23607
804-688-2475
             Ray Graves
             Senior Chemist
             GNB, Inc.
             P.O. Box 2165
             Columbus, GA  36867
             404-689-1701
Keith Greene
Chemist
American Analytical Labs,
840 South Main Street
Akron, OH  44311
216-535-1300
Inc.
John P. Gute
Lab. Supervisor Method, Res. & QA
L.A. County Sanitation District
1965 S. Workman Mill Road
Whitties, CA  90604
213-699-8903

-------
                                      795
Lee Hachigian
Manager, Water Pollution Control
General Motors Corp.
30400 Mound Road
Warren, MI  48090-9015
313-947-1656
                    Clarence Halie
                    PACE, Inc.
                    1710 Douglas Dr.  North
                    Minneapolis, MN  55422
                    612-525-3404
Guy J. Hall
President
Environmental Testing Services, Inc
816 Norview Ave.
Norfolk, VA  23509
804-853-1715
                    Philip Ham!in
                    ITT Rayonier Inc.
                    409 E. Harvard
                    She!ton, WA  98584
                    206-427-8232
Bryant Harrison
Acurex Corporation
4915 Prospectus Drive
Durham, NC  27713
919-544-4535
                    Lee Helms
                    Curtis & Tompkins
                    1250 S. Boyle Ave.
                    Los Angeles, CA  90023
                    213-269-7421
Rob Henry
VG Instruments
14513 Spotswood Furnace Road
Fredericksburg, VA  22401
703-786-5153
                    Michael Herbert
                    Baxters Health Care
                    Rt.#120 & Wilson Road
                    Round Lake, IL  60073
Geoff Hinshelwood
Organic Chemist
Jennings Lab
1118 Cypress Ave.
Virginia Beach, VA
804-425-1498
23451
Paula Hogg
Chemist
Hampton Roads Sanitation District
1436 Air Rail Avenue
Virginia Beach, VA  23455
804-460-2261
Dr. Philip Holt
Occidental Chemical
2801 Long Road
Grand Island, NY  14072
716-773-8538
                    Ben Honaker
                    Chemist
                    USEPA, ITD
                    401 M Street, SW (WH-552)
                    Washington, DC  20460
                    202-382-2272
Henry H. Hook
Deputy Director
Naval Supply Center
Fuel Department, Code 700
Norfolk, VA  23512
804-484-6140
                    E.W. Hoppe
                    Sr. Research Scientist
                    Battelle NW Labs
                    P.O. Box 999, Mail Stop P7-22
                    Richland, WA  99352
                    509-376-2126

-------
                                     796
Robert M. Houser, Ph.D.
Technical Director
TCT - St. Louis
1908 Innerbelt Bus. Ctr. Dr.
St. Louis, MO  63114
314-426-0880
Dr. Lyman H. Howe, III
Research Chemist
Tennessee Valley Authority
150 401 Chestnut St. (CC IN 150A-C)
Chattanooga, TN  37402-2801
615-751-3711
Stavros R. Howe
Lab Manager
Molecular Ecology Institute
1250 Bellflower Boulevard
Long Beach, CA  90840
213-985-4019
George D. Howe11
Supv. Chemist
Naval Supply Center
Fuel Department, Code 700
Norfolk, VA  23512
804-444-2761
Dr. Francis Y. Huang
Environmental Science & Eng., Inc.
11665 Li 1burn Park Road
St. Louis, MO  63040
314-567-4600
Greg Hudson
Oldover Corp.
P.O. Box 228
Ashland, VA  23005
804-550-2644
Frank Hund
USEPA
438 N. Armistead St. #304
Alexandria, VA  22312
202-382-7182
Mary M. Husted
Environmental Manager
Husted & Associates
P.O. Box 5256
High Point, NC  27262
919-869-3097
Nang Huynh
Laboratory Manager
National Laboratories Inc.
3210 Claremont Ave.
Evansville, IN  47712
812-422-4119
Richard Javick
Research Associate
FMC Corporation
Box 8
Princeton, NJ  08543
609-520-3639
Ellen E. Jenkins
Section Manager
DataChem Laboratories
960 W. LeVoy Drive
Salt Lake City, UT  84123
801-266-7700
Debra Johnson
Chemist
Aptus Environmental Services
21750 Cedar Ave. South
Lakeville, MN  55044
612-469-3475
Dr. Phanibhushan B. Joshipura
Supervisory Chemist
Naval Supply Center
Fuel Department, Code 700
Norfolk, VA  23512
804-484-6430
Jeffrey T. Keever
Research Analytical Chemist
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC  27709
919-541-7460

-------
                                      797
R. Michael Kennedy
Lab Supervisor
City of Rock Hill, Env. Man. Lab
P.O. Box 11706
Rock Hill, SC  29731-1706
803-329-5504
A. M. Kettner
Mobil Oil Corporation
P. 0. Box 1027
Princeton, NJ  08543-1027
Mary S. Khali 1
Instrument Chemist III
MWRD of Greater Chicago
550 S. Meacham Road
Schaumburg, IL  60193
708-529-7700
Alan D. King
Director of Env. Services
Sherry Laboratories
2203 S. Madison Street
Muncier IN  47302
317-747-9000
Dewey R. Klahn
Environmental Science Corp.
1910 Mays Chapel Road
Mt. Juliet, TN  37122
615-758-5858
Kelly Klatt
Technical Support Chemist
J & W Scientific
91 Blue Ravine Road
Poison, CA  95630
916-985-7888
Herman J. Kresse
Lab Director
M.B.A. Labs
340 South 66th Street
Houston, TX  77011
713-928-2701
Mark Kromis
Vice President
Bionomics Laboratory, Inc.
4310 E. Anderson Road
Orlando, FL  32812
407-851-2560
Beth Kummling
ECO LOGIC
143 Dennis St., Rockwood
Ontario, Canada     NOB 2KO
519-856-9591
An Lai
Chemist
City of Garland, Duck Crk. WWTP Lab
750 Duck Creek Way
Sunnyvale, TX  75182
214-226-7626
James D. Lamb, Jr.
Management Analyst
Naval Supply Center
Fuel Department, Code 700
Norfolk, VA  23512
804-444-5322
Rebecca A. LaRue
Assistant Professor of Chemistry
Cooper Union, Advan. of Sci. & Art
Sch. of Engr., 51 Astor Place
New York, NY  10003
212-353-4372
Peter A. Law
Laboratory Manager
Tighe & Bond, Inc.
53 Southampton Road
Westfield, MA  01085
413-562-1600
Nathan Levy
President
A & E Testing, Inc.
1717 Seabord Drive
Baton Rouge, LA  70810
504-769-1930

-------
                                     798
James W. Lewis
Laboratory Projects Manager
Bionetics Analytical Laboratories
18 Research Drive
Hampton, VA  23666
804-547-8935
Michael D. Lewis
T.C. Analytic, Inc.
1200 Boissevain Ave.
Norfolk, VA  23507
804-627-0400
Ron Lewis
Quality Assurance Office Code 130
Norfolk Naval Shipyard-
Portsmouth, VA  23709-5000
804-396-9305
Dr. Albert A. Liabastre, Ph.D
USAEHA - South
2489 King Arthur Circle
Atlanta, GA  30345
404-752-2826
James Longbottom
Research Chemist
USEPA, EMSL-Cin.
26 W. Martin Luther King Dr. MS-525
Cincinnati, OH  45268
513-569-7325
Lazaro Lopez
Asst. Director
Suburban Labs
4140 Lite Drive
Hillside, IL  60162
708-544-3260
Dr. Raymond J. Lovett
Professor
West Virginia University
Department of Chemistry
Morgantown, WV  26506
304-293-3068
Norman Low
Chemist
Hewlett Packard
1601 California Ave.
Palo Alto, CA  94304
415-857-7381
Curtis Lueckenhoff
Chemist
Missouri DNR
2010 Missouri Blvd.
Jefferson City, MO  65101
314-751-7930
A.J. Malanowicz
Mobil Oil Corporation
P.O. Box 1027
Princeton, NJ  08543-1027
Douglas B. ManigoId
Supervisor Chemist
U.S. Geological Survey, WRD
5293-B Ward Road
Arvada, CO  80002
303-236-5345
Mark Marcus
Director, Analytical Programs
Chemical Waste Management
150 W. 13th Street
Riverdale, IL  60627
708-841-8360
Michael J. Martin
Research Manager
BASF Corp.
1419 Biddle
Wyandotte, MI  48192
313-246-6878
Dr. Thomas D. Mathews
S.C. Wildlife & Marine Res.
P.O. Box 12559
Charleston, SC  29412-2559
803-762-5083
Dept.

-------
                                      799
Harry McCarty
Viar & Co./Sample Control Center
300 North Lee Street, Suite 200
Alexandria, VA  22314
703-557-5040
                        Frank McCullough
                        Applications Chemist
                        ABC Labs
                        P.O. Box 1097
                        Columbia,  MO  65201
                        314-474-8579
Cheryl McGuire
Chemist
Solutions Laboratories
814-H Greenbrier Circle
Chesapeake, VA  23320
804-420-0467
                        Neal  A.A.  McNeill
                        Chemist
                        Newport  News Shipbuilding
                        Dept. 031, 4101 Washington Ave.
                        Newport  News, VA  23607
                        804-380-7744
Carol Meyer
NY State Dept Health
Wadsworth Cntr. for Labs & Research
Albany, NY  12201-0509
518-486-5670
                        Ann G.  Miller
                        S-CUBED,  Div. of Maxwell Labs, Inc.
                        1800 Diagonal Rd.,  Suite 420
                        Alexandria,  VA  22314
                        703-838-0220
Dr. Deborah S. Miller
Union Carbide Chem. & Plas. Co. Inc
P.O. Box 8361, Bldg. 770, Rm. 318
South Charleston, WV  25303
304-747-4463
                        Harold W. Miller
                        Atlantic Div., NFEC
                        Bldg. IAA Code 1811
                        Norfolk, VA  23511
                        804-445-1929
Ray Mindrup
Supelco Inc.
Supelco Park
Beliefonte, PA
814-359-5414
16823
Karen L. Mixon
GC Supervisor/Chemist
Analytical Technologies, Inc.
560 Naches Ave. SW, Suite 101
Renton, WA  98055
206-228-8335
Jack Morgan, Jr.
Lab Manager
E.I. DuPont
P.O. Box 27001
Richmond, VA  23261
804-383-2968
                        Dr. Huggins Z. Msimanga
                        Kennesow State College
                        Chemistry Department
                        Marietta, GA  30061
                        404-423-5088
John R. Nein
Chemist
Naval Supply Center
Fuel Department, Code 700
Norfolk, VA  23512
804-444-2761
                        Gordon Nelson
                        Quality Assurance Office Code 130
                        Norfolk Naval Shipyard
                        Portsmouth, VA  23709-5000
                        804-396-9305

-------
                                     800
Becky Newman
County Court Reporters,
124 Cork Street
Winchester, VA  22601
703-667-0600
Inc.
Henry B. Ojeniyi
President
JENLABS, Env. & Con. Serv. Co.,
P.O. Box 116
Westville, NJ  08093
609-848-7227
                                                Inc
Beth 01sen
Naval Supply Center
Fuel Department, Code 700
Norfolk,. VA  23512
804-444-5137
                Alicia P. Ordona
                QA Analyst
                D6S-DCLS
                One North 14 Street
                Richmond, VA  23219
                804-786-3411
Martha C. Orr
Chief Chemist
HRSD, North Shore Lab
101 City Farm Road
Newport News, VA  23602
804-874-1287
                Michael  Palmer
                Organic  Chemistry Manager
                PACE,  Inc
                5460 Beaumont Center Blvd.
                Tampa, FL  33634
                813-884-8268
Wen Pan
Chemist
Sherry Laboratories
2203 S. Madison Street
Muncie, IN  47302
317-747-9000
                Trikam R.  Patel
                Associate  Chemist -  I
                Ney York City DEP
                Adm. Bldg.  Rm. 316 Wards  Island
                New York City, NY 10035
                212-860-3636
Michael N. Petterelli
OBG Laboratories, Inc.
5000 Brittonfield Prkwy, Suite 300
Syracuse, NY  13221
315-437-0200
                William F.  Pfeiffer
                Ginosko Laboratories,  Inc.
                17875 Cherokee
                Harpster, OH   43323
                614-496-4571
Eugene Pier
Business Development Manager
Dohrmann/Rosemount
3240 Scott Blvd.
Santa Clara, CA  95054
408-727-6000
                Alfredo Pierri
                Week  Laboratories  Inc.
                14859 E.  Clark  Ave.
                Industry,  CA 91745
                818-336-2139
Marvin Piworn'
Lab Manager
Hazardous Waste Center
One E. Hazelwood Drive
Champaign, IL  61820
217-333-8724
                Roy W.  Plunkett,. Jr.
                Analytical  Chemist  Supervisor
                Commonwealth  of  VA,  DGS/DCLS
                1  N.  14th St., Room 264
                Richmond, VA   23219
                804-225-4007

-------
                                      801
Gregory E. Pronger
Technical Director, Organic Labs
NET Midwest
850 W. Bartlett Road
Bartlett, IL  60103
708-289-7333
Bob Pullano
Quality Control
General Engineering; Laboratories
P.O. Box 30712
Charleston, SC  29417
803-556-8171
Gil Radolovich
Section Head
Midwest Research Institute
425 Volker Boulevard
Kansas City, MO  54110
816-753-7600
Katharine M. Raynor
Director, Quality Assurance Div.
Naval Supply Center
Fuel Department, Code 700
Norfolk, VA  23512
804-444-2761
Leah Reed
Sr. Analytical Chemist
Viar & Company
300 North Lee St., Suite 200
Alexandria, VA  22314
703-684-5678
Brenda Reeves
Conference Planner
ERCE
11260 Roger Bacon Drive
Reston, VA  22090
703-471-5550
Dennis G. Revel!
USEPA - Athens
College Station Road
Athens, GA  30613
404-546-3387
Lynn Riddick
Viar & Co./Sample Control Center
300 North Lee Street, Suite 200
Alexandria, VA  22314
703-557-5040
Debbie Rindfleisch
Sr. Lab Analyst
Newport News Waterworks
3629 George Wash. Memorial Hwy.
Newport News, VA  23602
804-867-9171
Ed Ritter
NJ Institute of Technology
323 King Blvd.
Newark, NJ  07102
201-596-5605
Roxanne M. Robinson
Scientific Officer
American Assoc. for Lab. Accredit.
656 Quince Orchard Rd. #704
Gaithersburg, MD  20878
301-670-1377
Dr. Peter D. Robison
Group Leader, Environ. Analysis
Texaco, Inc.
P.O. Box 509
Beacon, NY  12508
914-838-7692
David Roques
Research Associate
L.S.U. Institute for Env. Studies
Louisiana State Univ., 42 Atkinson
Baton Rouge, LA  70803
504-388-8521
Ann Rosecrance
Project Manager
ICF Tecnology
9300 Lee Highway
Fairfax, VA  22031
703-218-2587

-------
                                     802
Robert N. Rosenfeld
TBD Analysis, Inc.
2261 Federal Ave.
Los Angeles, CA  90064
213-478-4050
Dr. James R. Roth
Laboratory Manager
Alpha Analytical Labs
8 Walkup Drive
Westhorough, MA  01581
508-898-9220
Richard Rozene
Manager, Business Development
ABB Environmental Services
261 Commercial St., Box 7050
Portland, ME  04112
207-874-2400
Anna M. Rule
Chief Laboratory Division
Hampton Roads Sanitation Division
P.O. Box 5000
Virginia Beach, VA  23455
804-460-2261
Joseph H. Rule
Associate Professor
Old Dominion University
Geological Sciences
Norfolk, VA  23529
804-683-4301
Dr. Eric G.S. Rundberg
Deputy Projects Mngr. & QA Officer
Bionetics Corp.
18 Research Drive
Hampton, VA  23666
804-865-0880
Mark Rusler
Bionomics Laboratory, Inc.
4310 E. Anderson Road
Orlando, FL  32812
407-851-2560
Jeffrey V. Ryan
Chemist
Acurex Corporation
4915 Prospectus Drive
Durham, NC  27713
919.544.4535
William Schnute
Finnigan MAT
355 River Oaks Parkway
San Jose, CA  95134
408-433-4800
Alan Schoffman
Vice President
U.S. Testing Company
1415 Park Avenue
Hoboken, NJ  07030
201-792-2400
Dr. William D. Schulz
Dept. of Chemistry
Eastern KY University
Moore 337, EKU
Richnond, KY  40475
606-622-1456
Janice Sears
Conference Planner
ERCE
11260 Roger Bacon Drive
Reston, VA  22090
703-471-5550
Steven M. Shatkin
Organic Chemist
E.S. Babcock & Sons Inc.
P.O. Box 432
Riverside, CA  92502
714-684-1881
Edwin F. Shaw, Jr.
Division Operations Manager
Bionetics Corp.
18 Research Drive
Hampton, VA  23666
804-865-0880

-------
                                      803
Peter Shen
President
Quality Assurance Laboratory
6555 Nancy Ridge Drive, Suite 300
San Diego, CA  92121
619-566-1060
                                        Lawson E.  Sherman
                                        Project Chemist
                                        Texaco, Inc.
                                        P.O.  Box 509
                                        Beacon, NY  12508
                                        914-838-7531
Kate Simmons
Laboratory Director
Tighe & Bond, Inc.
53 Southampton Rd.
West-field, MA  01085
413-562-1600
                                        Husein Sitabkhan
                                        Lab Director
                                        ASAP Technical Services, Inc.
                                        19701 South Miles Road
                                        Warrensville,  OH  44128
                                        216-663-0808
Dorothy S. Small
President
Solutions Laboratories
814-H Greenbrier Circle
Chesapeake, VA  23320
804-420-0467
                                        Ronald B. Smart
                                        Professor
                                        West Virginia Universtiy
                                        Department of Chemistry
                                        Morgantown, WV  26506
                                        304-293-3068
Michael Smith
Environmental Supervisor
SD Dept. Health Lab.
500 East Capitol
Pierre, SD  57501
605-773-3368
                                        Nancy Souter
                                        Chemist/Project Manager
                                        Twin City Testing Corporation
                                        662 Cromwell Avenue
                                        St. Paul, MN  55114
                                        612-649-5517
Margaret E. Wickham St. Germain
Mass Spectrometrist
Midwest Research Institute
425 Volker Blvd.
Kansas City, MO  64110
816-753-7600
                                        Sally S. Stafford, Ph.D.
                                        Senior Applications Chemist
                                        Hewlett Packard
                                        P.O. Box 900
                                        Avondale, PA  19311-0900
                                        215-268-2281
Eric Steindl
Chemical Standards Chemist
Restek Corporation
110 Benner Circle
Bellefonte, PA  16823
814-353-1300
                                        Roger E. Stewart
                                        Technical Consultant
                                        Webb Technical Group, Inc.
                                        4320 Delta Lake Drive
                                        Raleigh, NC  27612
                                        919-787-9171
G. Edward Stigall
Technical Section Chief
USEPA Chesapeake Bay Liaison Off.
410 Severn Ave. Suite 109-110
Annapolis, MD  21403
301-266-6873
                                        Melanie Stoner
                                        Program Administrator
                                        ENSECO, Inc.
                                        7440 Lincoln Way
                                        Garden Grove, CA  92641
                                        714-898-6370

-------
                                     804
Kathleen Stralka
Statistician
SAIC
8400 Westpark Drive
McLean, VA  22102
703-734-2553
                  Mark Strangler
                  Chemist
                  Fairfax Cty. Health Dept.
                  10777 Main Street
                  Fairfax, VA  22030
                  703-246-3218
Dr. Chih-Wu Su
R & D Center
US Coast Guard
Avery Point
Groton, CT  06340
203-441-2720
                  Charles Sueper
                  Scientist V
                  Twin City Testing
                  662 Cromwell Avenue
                  St. Paul, MN  55114
                  612-649-5520
Roy Sutton
Development Chemist
Compuchem Inc.
P.O. Box 12652
Res. Triangle Pk, NC
919-248-6468
27709
Joseph Szlachciuk
Environmental Tech.
Texas Instruments, Inc.
34 Forest Street
Attleboro, MA  62703
508-699-1343
Jerry Thoroa
MAS Technology Corporation
110 South Hill St.
South Bend, IN  46617
219-233-3272
                  Frank H. Thorn
                  Sr. Laboratory Technician
                  Newport News Shipbuilding
                  Dept. 031; 4101 Washington Ave.
                  Newport News, VA  23607
                  804-688-4181
Samuel To
USEPA
401 M Street, S.W.  EN-338
Washington, DC  20460
202-475-8322
                  James C. Todaro
                  Laboratory Director
                  Water Control Laboratory
                  106 South St.
                  Hapkinton, MA  01748
                  508-435-6824
Susanne F. Tomajko
BP Research
4440 Warrensville Road
Cleveland, OH  44128
216-581-5939
                  David Tompkins
                  President
                  ETS Analytical Services
                  2160 Industrial Drive
                  Salem, VA  24153
                  703-387-3995
Allan Tordini
V.P., Technical Services
NET
220 Lake Drive East, Suite 301
Cherry Hill, NJ  08002
609-779-3373
                  Dr. David S. Trimble
                  Analytical Chemist
                  Union Camp Corp.
                  P.O. Box 178
                  Franklin, VA  23851
                  804-569-4596

-------
                                      805
Felicitas Trinidad
Hoffmann La Roche
340 Kingsland St.
Nutley, NJ  07110
201-235-3131
Jonathan S. Tschritte
Environmental Chemist
DuPont Company
P.O. Box 27001
Richmond, VA  23261
804-320-5398
F. Joseph Unangst
Laboratory Director, Vice President
Galson Laboratories
6601 Kirkville Road
East Syracuse, NY  13057
315-432-0506
Peter Unger
Vice President
American Assoc. for Lab. Accredit.
656 Quince Orchard Road #704
Gaithersburg, MD  20878
301-670-1377
Joe Viar
Chairman
Viar & Co., Inc.
300 North Lee Street, Suite 200
Alexandria, VA  22314
703-557-5040
Joseph S. Vital is
Chemical Engineer
U.S. EPA, OWRS, ITD
401 M Street, S.W., E-908
Washington, DC  20460
202-382-7172
Dr. Dallas Wait
Gradient Corporation
44 Brattle St.
Cambridge, MA  02138
617-576-1555
Tonie M. Wallace
President
County Court Reporters, Inc.
124 Cork Street
Winchester, VA  22601
703-667-0600
Gary Walters
Principal Scientist
ENSECO-RMAL
4955 Yarrow Street
Arvada, CO  80002
303-421-6611
Randy D. Ward
Senior Chemist
Environmental Science Corp.
1910 Mays Chapel Rd.
Mt. Juliet, TN  37122
615-758-5858
Dr. Walter C. Weimer
Battelle NW Labs
P.O. Box 999, Mail Stop P7-22
Richland, WA  99352
509-376-3995
Stuart Whitlock
Martel Laboratory Services Inc.
1025 Cromwell Bridge Rd.
Baltimore, MD  21204
301-825-7790
Robert Wichser
Manager, Chemical Services
Froehling & Robertson, Inc.
3015 Dumbarton Road
Richmond, VA  23228
804-264-2701
Dr. Daniel J. Williams
Association Prof, of Chemistry
Kennesaw State College
P.O. Box 444, Dept. of Chemistry
Marietta, GA  30061
404-423-6174

-------
                                     806
Allison Wilson
Chief Chemist
Hampton Roads Sanitation District
P.O. Box 5000
Virginia Beach, VA  23455
804-460-2261
Hugh Wise
USEPA, ITD
401 M Street, SW (WH-552)
Washington, DC  20460
Hark Yancey
Res. Scientist
Battelle
505 King Ave.
Columbus, OH  43201
614-424-4654
Thomas Yawaraski
Laboratory Manager
University of Michigan
181 Engineering IA
Ann Arbor, MI  48109
313-763-5686
Steve Yocklovich
Burlington Research, Inc.
P.O. Box 2481
Burlington, NC  27215
919-584-5564
Dr. Demetri Zadelis
LA Co. Sanit. Districts
1965 S. Workman Mill Rd.
Whittier, CA  90601

-------