5014
                              905R81118
         i

         2

         3

         4

         5

         6

         7

         8

         9

        10

        11

        12

        13

        14

        15

        16

        17

        18

        19

        20

        21

        22

        23

       24

        25
                         Br>'orE THE
              ENVIRONMENTAL PROTECT I OH AGENCY
                      AD HOC MEETING
         COMMON CONCERNS ON ANALYTICAL METHODS

                           * *  35
               r
                           Waterside  Mall
                           West Tower,  Room 908
                           401  M  Street, S.W.
                           VI a s h 1 n p. t o n ,  D . C .

                           January 5,  1981
          The  above entitled meeting convened, pursuant

to notice,  at  1:48 p.m., Mr. Courtney Riordan, presiding

official.

APPEARANCES:

          On behalf of the Environmental Protection  Agency
           COURTNEY RIORDAN, OMSQA

           RICHARD SCHECHTER,  Office of General Counsel

           ROBERT MEDZ, OMSQA/WWMMRD

           JAMES LICHTENBERG,  EMSL-CI

           JAMES LONGBOTTOM, EMSL-CI

           ROBERT L. BOOTH,  EMSL-CI
                                      U.S. Environmental Protection
           BILL LACY, OMSQA/WWMMRD   Region V, Library
                                      230 South Dearborn Street
           RON KITES, Consult ant/SABChicaSO. Illinois  60604

           JOHN LASETER,  Consuitant/SAB

                       NEAL R.  GROSS
                  COURT REPORTERS AND TRANSCRIDERS
                     1330 VERMONT  AVENUE, NW
Agency

-------
 1

 2

 3

 4

 5

 6
 9

 10

 11

 12

 13

 14

 15

 16

 17

 18

 19

 20

 21

 22

 23

24

 25
On
             P  of the Chi
                                                            •>r. I
   ROM  KAGEL,  Dow Choird r.al Co.

   WILLIAM H.  KROCHTA,  T?G Industries,  Inc.

   JAMES E. NORRIS, CIBA-GEIGY

   DAVID R. MARRS, SOIIIO

   DONALD 'j'. DAHU, Monsanto

   JUDY SHAW,  API

*'  JA14ES S. SMITH, Allied Chemical

   JANET MATEY,  CHA
     (202) 234-4433
                NEAL  R. GROSS
           COURT REPORTERS AND TRANSCRIBERS
              1330 VERMONT AVENUE, NW
               WASHINGTON, D.C. 20005

-------
 2

 3

 4

 5


 6

 7


 8

 9


 10

 11

 12


 13


 14

 15

 16

 17

 18

 19

 20

 21

 22


23

24

25
                         r R o c E E Ti T
            MR. KJORDAN:  On behalf cf  E?A,  I  would J 1 ke
j  to call this meeting to order.

            My name is Courtney Rlordan.   I  am Deputy

  Asr.lrtant Administrator for Monitoring  Systems  and

  Quality Assurance in the Office of Research and

  Development at EPA.

            The, 'purpose of this meeting is  for members
              <•
  of the research arm of EPA to meet with representatives

  from industry and acadenia to discuss ~,he  technical rneri;

  of the proposed 304(h) methods which were  initially

  promulgated as proposed in December of  1979 and fully

  commented on during the last year and are  now in the

  process of review for final promulgation.

            I would like to turn the meeting over briefly

  to Richard Schechter from the Office of General Counsel

  who will have a few comments regarding  some documentation

  that has been made available prior to the  meeting.  I

  would like to then open the meeting up  to  Ron Kagel to

  make some introductory comments as well.

            MR. SCHECHTER:  I am Richard  Schechter.  I .lust

  have a few procedural comments, one of  which is the

  comment period on this proposed rulemaking was closed

  and now we are for practical purposes reopening it for

  a short period of time to have this meeting and then

                        NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, HW
(202) 234-4433       -       WASHINGTON, D.C. 20005

-------
 1


 2


 3


 4


 5


 6

 7


 8


 9


10

11


12


13


14


15


16

17

18

19


20


21

22


23


24


25
  aft.orwardo  to see If there arc any public  comment:; to

  this  incot.11if,.

            The purpose of th^ meeting  is  discussion of

  the technjeal aspects of the 30'-I (h) proposed ruleinakj ng.

  We  would  lilce to limit the discussion to those  issues.

            Staff of the Agency did make available  a

  document  which indicates what our rulemaking may  be in

  the future,  but it is totally unofficial in  nature.  The
              <•
  only  official document concerning the rulemaking  is the

  proposed  ruieriaking which was promulgated  months  ago.

            The document with indicated test methods which

  was released by staff within the last week or so  can only i

  be  considered an informal and draft document and  not

  reflect any  official Agency position.

            Thank you.

            MR.  RIORDAN:  Ron?

            MR.  KAGEL:  I am Ron Kagel  with  Dow Chemical.

            Just to bring everybody up  to  speed on  this

  meeting,  let me go into a little background  on  how this

  occurred.

            Apparently Bob Finnegan of  Finnegan Instruments

  had some  discussions with Dick Dowd regarding the GCMS

  methodology versus the GC methodology.  Following those

  discussions, Dick Dowd talked to me to get our input.

            It appeared to me at that time that higher

                        NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433              WASHINGTON, D.C 20005

-------
 2


 3


 4
 9


10


11


12


13


14


15


16


17


18


19


20


21


22


23


24


25
 iiian.iiivinent  in  R',P/E?/\ war, unaware of Industry's

 and problems v.'lth the Cectjon 30'-! (h) nethodo] ogy.

 Consequently after that.  Court nc\v and I had  dj scur, r, ionr,.

 I outlined  our concerns and objections and offered  to

 arrange this meeting and  try to air some of  these  masters

 also involving CI/iA,  API and we hope the UV/AG  people  will

 join us shortly.


           I think our concerns fa]1 into three primary
             t
 categories  and I  would like to run through these quickly;

 the OC methodology we feel is not appropriate for

 compliance  monitoring of  a complex facility.   We have

 voiced these objections as early as 197^.  We did  so again

 in our comments last year.

           The  second point is there appears  to be  some

 confusion right nov.T on what constitutes a 30^4 (h) method.

           At the  Norfolk  conference in January of  1980,

 Dean Neptune,  Effluent Guidelines Division,  declared that

 the methods being used to propose effluent guidelines

 were equivalent to the 30Mh) methods.

           We have seen no evidence this is true.   There

 are methods other than the proposed 30^(h) methods  that

 were and  still are being used to generate the effluent

 guidelines.  There appears to be some rif beti^een

 Effluent  Guidelines, the R&D Branch and Enforcement.

 I wrote a letter to Bob' Schaffer last year suggesting

                        NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433       •       WASHINGTON, D.C. 20005

-------
  1


  2


  3


  4


  5


  6


  7


  8


  9


 10


 11


 12


 13


 14


 15


 16


 17


 18


 19


 20


 21


 22


 23


24


25
                                                       6


  that perhaps the  three  of ur-. could i-i.it down, Rf.-P,


  Enforcement and Effluent  fiuiriel j nes,  and reso.lve  some  of

  th^-;-;e pro Moms tof.ci her v.'ith CMA arid other industrial

  representatives.   1'erhaps v:e can net into that today;


  perhaps that is beyond  the scope of this meet.inp.   I


  think it is somethinf: that our:ht to be considered  in the

  future because 1  do  not believe we need more than  one


  set of 304 (h). methods floating around.
              <•

             CMA has  has consistently maintained that


  304(h) methodology  should be validated and that quality


  assurance/quality  control procedures should be followed.

             We submitted  a  validation protocol to the
I

  Effluent Guidelines  Division and the R&D Branch in


  December of 1975-  We also submitted definitions  for


!  detection  limits,  Units  of determination, et cetera,  to


I  R&D in 1976 and 1977.  We have been reinforcing these

  positions  over the  intervening years.

             The CMA positions that go back to 1975  v;ere

  the basis  for the  guidelines on environmental measurement-n


  that were  published  by  the ACS in December of 1980.  We

  are concerned that  EPA  R&D seems to be choosing a path


  that differs from that  taken by the recognized


  scientific community on these matters.


             We feel the 20  lab round robin is a very


  positive step.   Unfortunately we do not have the  data

                        NEAL R. GROSS
                   COURT REPORTERS AMD TRANSCRIBERS
                      1330 VERMONT AVENUE. NW
(202) 234-4433       -       WASHINGTON, D.C. 20005

-------
 4


 5


 6

 7


 8

 9


 10

 11


 12


 13


 14

 15


 16

 17

 18

 19


 20

 21

 22


23


24


25
 arui  tin'  benefit of the results of  tho;.,e  round robins ru,d

 we v.'L31  not  have- those until lon;r  after  the method;-, are

 proHul gated.   I think this is contrary  to good

 scientific  practice.   It also raises  and I think

 unnecessaryly so the  basis for legal  challenges 3 n the

 future.

           We  in C:IA and I know this is  true of ATI

 and  UWAG., have labored to resolve  some  of the issues wjth |
              < '
 EPA  at the  technical  level.  We are Still committed to

 resolve  then  at the technical level rather thai: at rhr-

 legal  level.

           I think today, if I understood Dick Dowd

 correctly,  vie have the ear of EFA  RT.-D management.  I

 think  we should take  advantage of  that  situation.

           KR. RIORDAI-I:  When you say  OCA is not

 appropriate  for compliance monitoring,  could you elaborate;
                                                    *
 on the basis  of that  statement so  we  can have an

 understanding of some of the technical  difficulties?

           I'iR. KAOEL:   In a complex manufacturing

 facility, and I will  restrict my remarks to a large

 facility, the nature of the waste  in  that facility

 changes, if not day by day, hour by hour.  This is

 primarily because in addition to continuous operations

 there  are batch operations and things come and thing5 B°

 rather quickly.

                        NEAL  R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE.  NW
(202) 234-4433               WASHINGTON, D.C. 20005

-------
  1


  2

  3

  4


  5


  6

  7


  8

  9


 10

 11

 12


 13


 14

 15


 16

 17

 18

 19


 20

 21

 22


 23

24


 25
                                                      8

            GO methodology JP  v-i'.v  sensitive.   It J:; not.

  ypc-c il'.l c .   Interfcronees from  both  pr;i ority po] "J utant s

  and non-priority pollutants  can obscure portions  of the   |
                                                             j
  chromatogramj can interfere  and lead  to erroneous  r'e::,ult ;•.;
                                                             i
            I believe Shell has  done  some extensive  studiers '•
                                                             I

  between  00 mann -spec and 0-C  comparing the methods  within

  their own labs, within the contractor lab.   Those  data

                                                             i
  were  submitted with the CMA  comments.   I believe  they werq

  also  submitted with the API  comments.   They were

  submitted independently by Shell  to the Agency.

            They very clearly  show  there  are problems with

  the GO methods in terms of interferences and erroneously

  high  results.

            MR. RIORDAN:  What were they  using for

  confirmation, were they using  GCMS?

            MR. KAGEL:  Yes.   I  think it  is true that most

  of the large companies will  use GC  mass spec for  doing

  this  monitoring rather than  0-0.   In a small location

  where there are a limited number  of products beinr, made,

  it very well may be the appropriate thing to use  GC,

  provided the GC methods are  appropriately GC mass  spec

  based, that is at the beginning,  all  the peaks in the

  chromatogram are identified  by GC mass spec and if there

  are any interferences or things that  go wrong during the

  course of a run, then that  is  backed  and reconfirmed by

                       NEAL R.  GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE. NW
202) 234-4433              WASHINGTON, D.C. 20005

-------
 1


 2


 3


 4


 5


 6


 7


 8


 9


 10


 11


 12


 13


 14


 15


 16


 17

 18


 19


 20


 21


 22


23


24


25
  GO  UCi Go Spf'O .


            That Is probably the  Mr;nest  problem v;c 300.

            MR.  RIORDAN:  The empirical  basis for t ho

  quesrJon of tho GO Is the She] 1  data base that war

  submitted during; the comment period?

            MR.  ]; A G EL :  T t j s ha r d  d at a .

            MR.  RIORDAN:  That is  the  set of data wo could

  look  toy to find Inconsistencies  with  the GC

  methodology?

            MR.  KAGEL:   Thar is correct.

            MR.  MARRS:   My name is  Dave  Marrs with SOHIO.

  We  submitted some data on the LC  method which showed

  not being able to differentiate  PllAs overlapping

  unresolved both priority pollutant   and non-priority

  pollutant iscners.   The sane is  true with the phenols.

            MR.  RIORDAN:  Did you  submit  that data?^

            MR.  MARRS:  Yes, through API.

            MR.  RIORDAN:  Is it my  understanding you

  support the GCMS method as the  method  to be used for

  compliance?

            MR.  KAGEL:  As circumstances dictate, yes. We

  would prefer to use GCMS.

            MR.  MARRS:  I think our objection is certainly

  GC  versus GCMS but none of the  methodology so far has

  been  validated to the point that it  would stand up

                        NEAL R. GROSS
                   COURT REPORTERS AND  TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433              WASHINGTON, D.C. 20005

-------
  1


  2


  3


  4


  5


  6

  7


  8


  9


 10

 11


 12


 13


 14


 15


 16

 17

 18


 19


 20


 21

 22


23


24


25
                                                      30

  through a concensus  method  writing association,  none of

  the methods as proposed  on  December 3ivi •

            MR. KAGCL:   I  can cite a couple of number r.

  out of the Shell data.   In  the  raw waste, GC mass spec

  versus GC, the Shell  C-C  mass spec ran 2100 mg/1.  Radian,

  an independent contractor,  round 1^00 ng/1.   Radian doins;;

  the same analysis by  GC  found 13,000 mg/1.  EPA doing

  the analysis through  MRI by GC  found 12,000 and 13,900.
              t '
  There Is quite a difference there, a factor of six.

            That is one of the things v;e have a problem wit!

  GC may falsely read  high.   If it is a compliance

  situation, you may be in a  compliance situation when you

  really are not because of the methodology.


            If the methodology were appropriately

  validated with precision and accuracy, I think we could

  eliminate some of .these  problems.

            MR. RIORDAU:   Those numbers were based on

  analysis of a single  sample?  Was that a split sample?

            MR. KAGEL:   I  will have to read up on that.

            MR. RIORDAN:   Yes?

            MR. NORRIS: Jim  Norris with Ciba-Geigy.  I

  think there is one point that needs to be clarified,

  in the event there is any confusion.  The CMA position

  is as an absolute analytical technique, we think GC

  mass spec ir> probably less  subject to the sorts of

                        NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE. NW
(202) 234-4433               WASHINGTON. D.C. 20005

-------
  1


  2

  3

  4


  5


  6

  7


  8

  9


 10

 11


 12


 13


 14

 15

 16

 17

 18

 19


 20

 21

 22


23

24


25
                                                      11


 interference  as  positive and negative that we  see  engaged

 with  GO.   There  should be no confusion that our position

 with  regard to  62^4 and 625, which were the methods

 published  in  the Federal Register on December  3, 1979,

 we  find  some  deficiencies in those methods. We are  not

 offering them as the standards of the world at this

 juncture".

            If  I1'remember correctly, the Shell work  was
         f     »"
             < '
 done  by  Shell's  own GC mass spec technologies  and  not

 necessarily calling upon 62^ and 625.

            MR. RIORDAN:  In your comments you also  provide

 recommendations  on what represents validation  in your

 mind  for a method to be accepted.  Is that correct?

            MR. KAGEL:  That is correct.  Essentially it

 is  the same — this is the article published in

 December of last year, analytical chemistry, guidelines

 from  the ACS.

            What we proposed was a validation scheme

 very  simply 7-7-7, following the N-N-N type scheme  or

 seven is recommended, to determine recovery, precision

 and accuracy  and so forth.

            MR.  SCHECHTER:  Do we have a copy?

            MR.  KAGEL:  That was just published  in 1980.

 I  submitted  informal copies to the Effluent Guidelines

 Division.   I  think I submitted an informal copy to R&D

                       NEAL R.  GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE, NW
202) 234-4433       .       WASHINGTON, D.C. 20005

-------
 1

 2

 3

 4

 5

 6

 7

 8

 9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25
  Clnoinu-itl  while it was -.1,13 1  j n  draft.

            MR.  SCllECIiTEK:  Why  don't  we snhnlt 3t so

  It  v.'.Il]  be  part, of the  formal  record.

            MR.  KAGEL:  All night.   We have proposed in the j

  past  dealing with validation inputs  to that article.

  Dow input strongly.  C^ d:!d input also.

            MR.  MARKS:  In the December  3rd Federal

  Register, EPA-had an appendix  on  quality  assurance on

  62^4 and  6?|5.   One of the things we tried  to get across

  in  our  comments through API was we believed that method

  validation  is  distinct  from quality  assurance,  that

  method  validation is the responsibility of the  Agency,

  not the  responsibility  of the  user.

            The  way that  was structured  on  December 3^d

  is  they  had each individual user  validating the GCMS

  methodology.   There was no criteria  as to what  was an
                                                    4

  acceptable  validation.

            MR.  RIORDAM:  What is the  distinction between

  quality  assurance and validation?

            MR.  MARKS:  In our thinking, very often in

  a concensus method, for example,  which ASTM would

  publish, there would be with that method  performance

  criteria. For example,  there had  been established through

  interlaboratory testing that the  analyst  could meet and

  say I have done this right or  I have to do it again

                       NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE. NW
(202) 234-4433               WASHINGTON, D.C. 20005

-------
 2


 3


 4


 5


 6


 7


 8

 9


 10

 11


 12


 13


 14


 15


 16


 17


 18


 19


 20


 21

 22


 23


24


 25
j  because iujinetbl nf, is wronr-   In tho method  ur. published

  by  the Agency, that  has not been  provided.   The analyst

  is  to develop his own confidence  levels;  not only doer.

  that duplicate effort with everybody  doing it,  it coots

  about $50,000 per ]ab to do it.   Not  only  is it

  expensive but a bad laboratory would  have  a wider sM-

  of  control limits.  They would not  have any reason to know

  they were do^rtg things poorly and ought to clean their act

  up  a little bit.

            MR. KAGEL:  If you  follow the ACS article

  on  methods validation, the way they have it lined out,

  this is something you do once on  a  method  for a matrix.

  A method which is valid for my waste  may not be valid

  for your waste.  It may be matrix specific because of

  interferences.  It is something that  need  be done only

  once.  It is not an ongoing thing that is  done  every

  day or every week.

            The protocol is laid out.   The steps  are

  followed.  As long as the protocol  is not  changed, th>-M

  the validation should be verified.

            MR. RIORDAN:  Does  the  article say it would

  have to be done for every matrix  or is "matrix" defined

  as  an industrial effluent from one plant?   Would 3t be-

  good enough that a matrix represents  effluents from the

  industry?

                        NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433              WASHINGTON, D.C. 20005

-------
 3

 4

 5

 6

 7

 8

 9


 10

 11

 12


 13


 14

 15

 16

 17

 18

 19

 20

 21

 22


23

24


25
            MR.KAuEL:  I do not think  you  can be that

  broad w:ith the whole industry.   I  think  it  hac to be

  matrix specific.

            Part of the problem is in  the  chemical industry

  you will find that most of the effluents have high salt

  contents.   These,vary from plant to  plant.   This has a

  dramatic effect on the recovery  in a particular effluent .
                 :
  There ar;e  other things there.  I think it is well known
              <•
  that priority pollutants constitute  probably less than

  one percent of a total effluent, there are  other thinp;

  there that will interfere.

            It is matrix specific.

            MR.  RIORDAN:  Would validation in your terms

  have to be done on each plant effluent?   Would validation

  be  possible?

            MR.  KAGEL:  We differ  on this.
            MR.  MARKS:  Because the  effluents  are more
  uniform.
            MR.  KAGEL:  Ours are not.   I  can guarantee you

  that the effluent in Midland, Michigan  is not the same

  as the effluent of Dow/Texas and  so  forth.   It is

  dependent on the product mix of the' plant.

            MR.  LASETER:  You indicated earlier that will

  change based on the operations and  variations within the

  plant itself.

                       NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
1202) 234-4433              WASHINGTON, D.C. 20005

-------
 6

 7

 8

 9


 10

 11

 12


 13


 14

 15

 16

 17

 18

 19

 20

 21

22


23

24


25
            MR. KAfiEL: That  Is  corroci .

            MR. LASETER:   How would  .YOU  identify sono

  standard matrix, If you  increased  the  content, T do not

  know what would happen.

            MR. KAOEL:   I  do not  know if you can identify

  some standard matrix.  That Js  why I think what wo are

  really talking about is  a  program  to identify when that

  happens-/ whep'-you have the change.   I  know it is kJnd of

  stick-y.   It is a tough one to wrestle  with.

            MR. DAHIJ:  Don Dahn,  Monsanto.

            I think what we  are talking  about is a hierachy

  of method validation as part  of the QA/QC program.  When

';  you start off and you  are  developing a new set of

  methods for use under  304(h), that is  something you

  expect to have broad utility  on.   You  should check-

  under a fairly wide variety of  circumstances whether

  these methods are rugged.  If they are found to be rugged

  under a variety of circumstances,  then they are good

  methods to be published.

            There ought  to be another set of QA/QC

  procedures that I validate that method in my own

  laboratory that shows  I  can use that method competently

  as an individual laboratory on  some standard matrices

  and further,  I have to have some experiments that tell

  me that set of methods is  working under my own matrix.

                       NEAL  R. GROSS
                   COURT REPORTERS AND  TRANSCRIBERS
                     1330 VERMONT AVENUE, NW
(202) 234-4433             WASHINGTON, D.C. 20005

-------
  4


  5


  6


  7


  8


  9


 10


 11


 12


 13


 14


 15


 16


 17


 18


 19


 20


 21


 22


 23


24


25
                                                      16

            Under niy day to day operations,  I  have  to have

  some  QA/Qt procedures that show something has  not  gone

  off-,  that  I have a reasonable expectation that ruy  matrix

  is  being constant.

            All of these things have to be  balanced  against

  the cost.   If you.did too many method validation

  procedures all along the way, you would soon be  spendlnf

  more  on/method validation than on the experimentation.
              t '
            You have to have a hierachy within the 30*1 (h)


  methods  before they are published, they ought  to hav^

  a pretty rigorous method validation.  When  you get down

  to  what  an individual laboratory is doing using  that

  method day to day, it would be much less  on any  particu];..


  sample and therefore less costly.

            MR.  KAGEL:  I think this varies depending on


  the size of the location.  We have been using  a  set of

  methods  for some time now and they are validated 'for our

  waste.   Even though we have batch operations up  and dov;;<,

  the changes are so slight that it does not  really

  interfere With the methods.  When it does,  we  know it

  does  and we go back to the GC mass spec and find out


  what  is  going on.

            In a smaller operation which is all  batch,


  that  can be a real problem.

            The initial method is GC mass spec  based.

                       NEAL R.  GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE, NW
202) 234-4433              WASHINGTON, D.C. 20005

-------
  1


  2

  3

  4

  5

  6

  7

  8

  9


 10

 11

 12


 13


 14

 15

 16

 17

 18

 19

 20

 21

 22

 23

24


 25
  represented here  use  UC  as ;-">  roiA'ino nion i t or' nc  ciev.ico,

  either for process  conlri.i]. or ef.Huent monitor! nr;?


            MR. D.A1IN:   Certainly.

            MR. R'lORDAN:   Have  you had. experience  in  ternis

  of data that shov.rs  thJs  is less  sensitive than some

  other methods or  goes  out  of  control more often?    Is  tliJ

  because /the extremes  are fairly  constant in terms of
              f
  chemical composition  that  you are able to use them?


            MR. KAGEL:   Sensitivity is not a problem.   It

  is specificity.
i

j            MR. DAHN:   I think  we  could produce a  wide

  variety of examples where  GC  as  a method worked  very

  effectively at measuring an effluent for a particular

  compound in a particular wastestrean.  I think v*e could

  also dig up examples  where problems occur with using  GC
                                                    %
  which would be less likely ^-to occur with GC mass spec.

            The point I vrauld like to get across is you

  really have to make sure you  have the experiments that

  tell you whether  or not  a method is valid before you  can

  use it for compliance monitoring.

            MR. RIORDAN:  You are  saying basically that


  EPA should do that?

            MR. DAHN:   It  seems to me that EPA's

  responsibility is to  validate the method and show it  is

                        NEAL R.  GROSS
                    COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT  AVENUE, NW
(202) 234-4433             WASHINGTON,  D.C. 20005

-------
  1


  2

  3

  4


  5


  6

  7


  8

  9


 10

 11

 12


 13


 14

 15

 16

 17

 18

 19

 20

 21

 22


23

24


25
                                                          18
  i'ur,ned and it works under a v;j de  variety of
  circumstance;--.   It is the individual  laboratory's

  responsibility to show they  can  use  that  method

  competently and it is also theJr responsibility to have

  a record of some quality control and quality assurance

  data that shows the method has been  applied correctly

  throughout the history of the use of that method.

            MR.'-RIORDAN:  In your  comments  did you

  recommend what might be a reasonable validation scheme

  for a 304(h) type broad utility  method?   Is that in the

  CMA article?

            MR. KAGEL:  When did we submit  that? December,

  1975 was the first time I sent it in and  I sent it in

  again in January, February and March.

            I was trying to get a  response  from someone

  in EPA and they never did get that.  I was forced to write

  a rather harsh letter a year later about  where the

  response was and we got a response about  a year after

  that.

            They were submitted some time  ago.  I do not

  know whether we made reference to that in our comments

  specifically.  We did make reference to  the letters

  we had sent off to Dr. Ballinger on  detection limits

  and limits on determination.

            MR. DAHN:  In the  CMA  comments, there are

                       NEAL R.  GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE. NW
(202) 234-4433             WASHINGTON, D.C. 20005

-------
    1


    2

    3

    4"
  '  7


*   8.


    9


   10


   11


   12


   13


   14


   15


   16


   17


   18


   19


  20


  21


  22


  23


  24


  25
                                                      19


  comments on the 20 lab round robin tests as  being a


  beneficial part of' such a. validation scheme.


            MR.  LASETEH:  Ire we saying low resolution, ir

    ,* b
  th"   primarily v;hat we ,sre speaking of or are  you taking
        t

        .'J/p,h resolution with selective detectors  to begin
        *"•»      J .-

 •to- T.1n_hnizo f-bme.of the interferences using  GC  as a tool


  or using >j'Cr _.j the generic sense.   I am having a problem
     *"     Do the people who are addressing this  from


  industry primarily use low resolution type detection


  systems?

            MR.  NORRIS:  1 have seen good work done  with


  such  systems,  as long as the matrix remains fairly


  constant and you have 
-------
  3

  4
  8


  9


 10

 11


 12


 13


 14


 15


 16


 17

 18


 19


 20


 21


 22


 23


24


 25
            HT\. KAHEL:   I  think  ve  sro talking primarily


  about pacl:cd column.   Some  of  the method:- have specific

  detectors, others are  just  general FID.   I think we wou"

  h c\y-: tog e t 1 n t o s p e c i f i c s .

            MR. MARKS:   I  think  that point is more

  general; regardless of the  method, the homework has

  to be done before it hits the  streets.

          ,-  MR. t KROCHTA:   I think we need  to be more
              r
  specific because we are  using  a combination of which

  you would consider high  resolution and slow resolution

  columns.  For example, using less capital area columns
i

I  which are considered high resolution and some

  applications, for example,  P'.iAs,  the phenols, versus


  maybe low resolution,  on the volatiles.

            We are using mass spec  in addition to the

  high resolution tolerance.

            I might add  we are talking about validation.

  I do not think this is unique  with CMA or API.  There

  was just a recent article  in "Environmental Science and

  Technology" and this article referred to the fact that

  they were asking EPA for validation of 304(h) methods.


  It was about a month ago,

            MR. RIORDAN:  Basically what you are saying


  is you do have some data based upon the Shell

  experience which shows —  I guess that was low

                       NEAL R. GROSS
                   COURT REPORTERS AND TUANSCRIDERS
                     1330 VERMONT AVENUE, NV/
202) 234-4433              WASHINGTON, D.C. 20005

-------
 . o
 8


 9


 10

 11


 12


 13


 14


 15


 16


 17


 18


 19


 20


 21

 22


23


24
  resolution f,as --  provides  resu.li.r, v.'hicb are

  inconr.iKtent v.'ith  that  of GCMS re^i;lt^.  I guess  the-

\  CCKio Is considered  to be  in that instant the more

  accurate measurement.   Is that right?

            MR. KAGEL:  These were the 600 series net hod-.

  that were app'l :i ed • for tolulene.

            MR. RIORDAK:  That is  basically the data  to

  look at .-in terms of Inconsistency at this point.

            MR. KAGEL:  That  is the best data available

  at this point in tine.

            MR. RIORDAN:  In  terms of method validation,

  you have a general  position which says that EPA should

  not promulgate a  304(h) method for general application

  for identifying and measuring pollutants and effluent

  unless it has gone  through  a fairly vigorous validation

  procedure.  You define  "validation" as being the

  application of the  method to several different waste

  streams under inter-laboratory colaborative conditions.

            MR. SCHECHTER:   Basically a reflection  of the

  same position that  was  in the comments.

            MR. KAGEL:  Yes.

            MR. DAHI-I:  My feeling is you should be  able

  to give some direction  to the user of that method as

  to whether that method  is applicable to a given

  circumstance by giving  him a set of experiments he should

                       NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433             WASHINGTON, D.C. 20005

-------
   1



   2



   3



   4



   5



   6



  7



  8



  9



 10



 11



 12



 13



 14



 15



 16



 17



 18



 19



 20



 21



 22



23



24



25
  do before  ur-jnrr, the method.   It may turn out  a  specific



  method  is  not  applicable but 3 f it is, then lie  can  use 1



  Thar, was the point  Dave made earlier with the fact  that



  there was  not  any performance criteria included in  the



  62M and G25 section.



             MR.  MARRS:   Especially as you pet down  at the



  ppb level, uncertainty becomes a major factor.   I think-



  there q,ught to be requirements — or we think at  API,
              r


  there ought to be some kind of requirements in  the



  method  to  require the analyst to report not only  his



  results but the uncertainty associated with that  result.



  Twenty-six and 25 ppb may be the same number.



             MR.  KAGEL:   Or 26 and 35 ppb or 26 and  70



  may be  the same number.



             In the letter we submitted some time  ago,



  defining detection  limits, we spelled that out  very

                                                    «

  carefully  in terms  of signal to noise, to make  an



  instrument independent.   At that tine we said a



  qualitative result  should be 2.5 times signal to  noise.



             The  ACS has adopted a slightly different



  position since then,  that was their original position.



  They now say  it should be three times signal  to noise.



  That is a  qualitative number. Yes, it is there, no, it



  is not  there.   That is a detection limit.



             A  quantitative number should be ten times



                       NEAL R.  GROSS

                   COURT REPORTERS AND TRANSCRIBERS

                     1330 VERMONT AVENUE, NW

202) 234-4433              WASHINGTON, D.C. 20005

-------
  3


  4



  5



  6


  7



  8



  9



 10


 11


 12



 13



 14


 15



 16


 17


 18


 19



 20


 21


 22



23



24



25
  s:i [\na.l to noise.   That  jr. :/. ill cons.1si.ent with  t.h^  ACS


j  position.  Any nusijbor botv:een 2.5 and 10 should  not  K;


  ta]:en ar. a quant itr.tive numhey, number that ha?  moan:ing,


  but should be tal:en  as  an indication it is there and  in   !


  order to get meaning for that number, it needs to bo
I

  repeatt-d, to be verified.


[            That is  part  of the problem working down by


!  detection lipiit.   We see numbers coming out.  We saw  this


  in effluent guidelines'  screening phase where the methods


  were said to have  a  detection limit of 10 ppb and yet


  we saw results coming in that two ppb, one-tenth of a ppb,


  and. I do not know  what  those  numbers mean.


            MR. RIORDAN:   You are saying the method should


  have a realistic detection limit associated with it?


            !3R. XAGEL:  The method should have a detection


  limit, yes, a realistic detection limit associated with


  it.  I think what  Dave  was saying is there should be


  realistic precision  and accuracy associated with the


  result, such that  result will achieve the goals.


            MR. MARRS:  I think it is a real problem for


  non-analysts to appreciate what uncertainty is.


  Engineers do not and lawyers  sure do not.


            MR. RIORDAN:   Can you really say this  data


  is good with 90 percent comfidence?


            MR. MARRS: There are two things; it really


                        NEAL  R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE. NW
202)  234-4433               WASHINGTON. D.C. 20005

-------
 3


 4


 5


 6


 7


 8


 9


 10


 11


 12


 13


 14


 15


 16


 17


 18


 19


 20


 21


 22


23


24


25
|  dopcn'lL- on IMV: tl>> data  is  usod.


j            MR. K^EL:  V.'t-  hnve  run into problems whoro

I
j  dat-a out of a data bar,_-  of  20  pcrinlr,,  eight of thoso

I
J  point.;:, vere rejected and  a  standard of deviation V.TJS


  calculated, it was called a two sigmn  95 percent


  confidence level-.  That  sparked quite  a debate at an


  open meeting.


            MR.- RIORDAM:   In  your discussions with Dick,
              «•

  are these basically the points that were covered?  Did


  you cover anything in addition?


            MR. KAGEL:  I think  those were the basic


  points I covered in our very short  discussion.  I think


  there are other points that people  may want to bring out.


            I wanted to bring up one  other point and that


  was recovery.  You po through  the exercise nov: vi^h


  method 62^ and 625 for determining  the recovery.  Then


  recovery is not used.  The  number is reported and the


  recovery is reported along  v;ith the number.  1 guers T


  am a little uncertain as  to how one makes a decision as


  to whether "that is a good number of not.


            Even though it  work  not to our benefit to have


  recovery correct numbers, I think there is some


  argument that recovery correct numbers are probably


  better than just raw numbers.


            If the number  is  reported with the recovery


                        NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433              WASHINGTON, D.C. 20005

-------
  1


  2

  3

  4


  5


  6

  7


  8

  9


 10

 11


 12


 13


 14

 15


 16

 17

 18

 19


 20


 21

 22


 23

24


25
  njicl j ^ 1 p;et 38 wjth  a  recovery of -1'; percent, uv  38

  with a recovery of  8r5 percent ,  I do not roniemhrr any

  r,u:l dance telliri)T re what  is  a  f;ood nuvib'-'r or bad number.

            I think that  really  promoter, jn the lon{;; run

  bad science.  People  start reporting numbers wit])  bad

  recoveries instead  of workinr.  things over to get good

  recoveries.

            We -would  like to see  all recoveries above 85
              <•

  percent.   That is not always practical.

            Some allowance  ought  to be made for this

  recovery.   Right now  it is kind of use of the yardstick.
                              «
  I believe the ACS guidelines do call for recovery

  correction.  This is part of the validation procedure.

            MR. LASETER:  Even the recovery will change

  as you start to aoproach  the lov: end all the way tbrourh

  regardless of the ultimate.
                                                    *
            MR. KAGEL:  The recovery has to be run at

  different levels, if  that is v;hat you mean.  You will

  trace out a heteroscadastic  curve on your recovery and

  follow the error curve  almost  perfectly.  At very  low

  concentrations, your  recovery  starts going bananas both


  ways, plus and minus.

            Some of the data vie  have seen in a

  preliminary round robin would  indicate recoveries  jump

  all the way from 20 percent  up to 200 percent.

                        NEAL R. GROSS
                   COURT  REPORTERS AND  TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433             WASHINGTON, D.C. 20005

-------
 9


10

11

12


13


14

15


16

17

18

19


20


21

22


23


24


25
            MR .  HARK3:  Vh:\t ] eve] ?

            MR.  KA(;;•',!,:  T believe  that was  ]00 npb.

            MS.  SHAY.':   Judy Shaw v::i th AFT.   We are doinr,

 a project  right now with the ETA nuideljnes  people at

 tv.'O  oil  refineries.  ' We arc using the  stable label method.

 VJe are puttinr in  about 20 stable label compounds.  VJhat

 we are going to do with the data is looking  at  the

 stable Label spikes, it gives you some indication  of the

 accuracy of 6?^ and 625.  We are ,1ust  beginning to work

 with that  data.

            One  of the things is the recovery  from one

 sample to  another  is different for a particular compound.

 It is also different between compounds.   You might have

 in one sample  60 percent recovery of benzene and 20

 percent  of tolulene and in the next sample it might

 flip flop  and be the opposite.

            If you look at average recoveries, they  are

 great but  the  range is terrible.

            VJe are running the extract.ibles at 100 ppb

 and  the  volatiles  at 20. That data could  be  very useful.

 We only  did it up  to four weeks.  The  guidelines people

 did  it  for two months, sampling  every  other  day effluent

 and  influent.   We  can also look  at whether or not

 recovery is a lot  different between the effluent which

 is a pretty messy sample and the influent for

                        NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433               WASHINGTON, D.C. 20005

-------
 3

 4


 5


 6
 9


 10


 11


 12


 13


 14


 15


 16


 17


 18


 19


 20


 21


 22


 23


24


25
  re-cover !'.•:•. which  is  vretty cie-Mri.

            U'P arc-  r^'eing t-hingr, from 20, if you  put  in

  "JOO,  you got betvreen 20 to 1'30 or more back.  There is in.

  way to ])]••:-dlet what  you are going to get.

            MR. LASETER:   Is it reproducible with the sam^

  split sample?  A-re you  criticising the method or the

  variability in the samples you are xu-ing?

            MS.- SHAW:   You have one sample that you are

  analyzing for everything else.

            MR. LASETER:   If you reanalyze that same

  sample, is your scatter because you have taken  a second


  Influent sample a few minutes later?

            MS. SHAW:   No.   The scatter is day to day

  within a sample,  if  you are getting 20 percent  recovery

  on benzene and 80 percent or; tolulene and you are not

  correcting, you are  not getting good numbers.

            MR. LASETER:   If you run that same identical

  sample again.

            1-13. SHAW:   We have only done that a little bit


  so we do not know.

            MR. LASETER:   You have to satisfy that question


  on scatter before you work on scatter sample to sample.

  That is why  I am  not sure what you are saying.

            MR. KAGEL: We have a similar program with

  the Effluent Guidelines Division.  We are  studying  five

                        NEAL R.  GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330  VERMONT AVtNUE, NW
(202) 234-4433              WASHINGTON. D.C. 20005

-------
  1


  2

  3

  4


  5


  6

  7


  8

  9


 10

 11


 12


 13


 14

 15


 16

 17

 18

 19


 20

 21

 22


 23


24


 25
                                                      ?8

  plantr,} the Influent  and  effluent and looking at data.

  We are not using  stable label.   We are using method 624

  and 625.  We will  have several  data points per plant

  where there will  be an EFA number, a company number

  and an independent contractor number,  not on every plant

  but on many of them.

            MR. RIORDAN:  If there are no more comments, 1

  will take under advisement with the people in ORD and
              <•
  the Agency the comments.   I assure you we will give them

  full consideration when we get  around to reviewing the

  package for final  rulemaking.

            A transcript of this  meeting will be made

  available to you  as well  as the. public.   Can we make

  it available to them  so they can edit the transcript?

!            MR..SCHECHTER:   Anyone can obtain a copy of

  'the transcript from the Reporter.

            MR. RIORDAN:  Before  we go public with'the

  transcript, is it  possible to give the participants a

  chance to edit their  comments?

            MR. SCHECHTER:   I do  not believe so.  It is

  not a sensitive transcript that would be used in a

  judicial proceeding.

            What we are going to  do is make the transcript

  available and open the comment  period for a 20 day

  period to allow any  interested  party to comment on the

                        NEAL R. GROSS
                   COURT  REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE. NW
(202) 234-4433             WASHINGTON, D.C. 20005

-------
  1


  2

  3

  4


  5


  6

  7


  8

  9


 10

 11


 12


 13


 14

 15


 16

 17

 18

 19


 20

 21

 22


23


24


25
I
  transcript or the meeting.   In  addition  we are r;u.inr;

  to  make the package of rnethr.ur.  that  we ;>re working on,

  the package made available to thus group,  avail able, aloir

  with the transcript.

            If there are any comments  about  the transcript

  of  the meeting, ^including corrections, they can  be made

  during that comment period.

            MR., 'KAGEL:  Can we get an  update on the
              t
  Cincinnati folks, as to where the methods  are,  what

  has happened to the methods?  Dr. ".leas,  v:hat IF  the

  status of the round robins?  What is the status  of the

  definition of "detection Units"?

            i-IR. SCHECKTER:  Vie cannot  at this particular

  meeting discuss anything beyond the  proposal.

            MR. RIORDAH:  We are  restricted  to discussing

  what was made public in December and are unable  to

  discuss with any single parties or independent  parties

  what we have been doing in thc  meantime.

            MS. MATEY:  You are unable to  discuss  the

  changes in the methods since the last time they  were

  proposed, although we have a copy of the draft?

            MR. SCHECHTER:  I think the situation to be

  precise is only the Administrator can promulgate a

  regulation.  As a consequence,  the package that you have

  IB  a package that staff is working on.   It does not

                       NEAL R.  GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE, NW
202) 234-4433             WASHINGTON, D.C.  20005

-------
  1


  2


  3

  4


  5


  6

  7


  8


  9


 10

 11


 12


 13


 14


 15


 16


 17

 18


 19


 20


 21

 22


23


24


25
  represent any off]rial  position on the port of the  Agency.

  ~Jt may or may not  be  the .methods that wo actually

  promulgate.'.

            VJe are open to disc-uss 1 n g t h at packap;e v; j t h

  that in nind.  The Agency  has  not taken any official

  position that it- Is going  to  promulgate that group  of

  methods.

            Wit-h that in  mind,  we would be willing to


  discuss it.

            MS. RAT FA7:  Dr.  Mcdz is familiar with the

  changes and what has  happened  since they were initially
                              •
  proposed.  If he or someone else could explain to us

  what the changes are  in this  draft —

            MR. SCHECHTER:   I do not think we can do  that,

  That really goes beyond the scope of this discussion.

            MR. RIORDAN:  We can receive comments on  the

  draft from you.

            MS. MATEY:  "We just  got 'hem.

            MR. SCHECHTER:   Since there is going to be

  a 20 day comnent period, perhaps we should wait and

  let you look at the document  and then submit any

  written comments.

            I think  it  is a  difficult procedural

  situation.  Ordinarily  we  would not be in this

  position but it is a  very  technical rulemaking.

                       NEAL  R. GROSS
                  COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVtNUE. NW
202) 234-4433              WASHINGTON, D.C. 20005
A lot

-------
  1


  2


  3

  4


  5


  6

  7


  8

  9


 10

 11


 12


 13


 14

 15


 16

 17

 18

 19


 20

 21

 22


23


24


25
                                                      31

  of It 1 r. purely a matter  of  chemistry.   Obvious] y the

  Agency wants to get thv most  input  from the regulated

  com-MUiiity as possible.

            On the other hand,  the  situation precisely is

  other than the proposed rulernaking,  the Agency lias not

  made a proposal.  As  such, the  Agency really cannot

  comment on what other possible  rule::iakings we might bo

  considering.
        s
            MR.  RTORDAN:  I think you  have given comments

  and I think they require  the  attention  of the policy

  makers within the Agency  to make  a  decision.

            MR.  KAGEL:  We  raised the  question of the 20

  lab round robin.  Could v;e go into  that now?

            MR.  RIORBAN:  The  20  lab  round robin, in

  terms of the earlier  comments or  in  terms of the

  results?

            MR.  SCHECHTER:  We  have  no  objection to'your

  discussing it.  On the other  hand,  anything that was

  discussed in the comments will  be considered by the
  Agency.
            MS. SHAW:  That  was  not in the comments because
  the results were not published.

            MR. KAGEL:   That  would be my major comment,

  the results were not published.   We did open the door

  to say we ought to  have  those results before the methods

                       NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
[207) 234-4433             WASHINGTON, D.C. 20005

-------
  1


  2

  3

  4


  5


  6

  7


  8

  9


 10

 11


 12


 13


 14

 15


 16

 17

 18

 19

 20

 21

 22


 23

24


25
  :uv promulr.at od.   ''or  what  rood it diet, \vo rer,.-'rvod  the

  right to reopen  tin?  conment period ourrcJves.

            MR. DAHN:  You  mentioned this is a highly

  technical issue.   I  fuess the concern I have is we may

  have a set of methods that worl: fairly wol 1 in

  professional hands and which professional people  can

|  recognize when they  are not. -.-.'Or !;i ng and do POIIK-I '  Jng

I  to modify it and get the  right method.
I        '     t '
            I am concerned  when we get to an enforcement

  situation, the nethods have to be  rather rigoron.-'y

  followed in order  to comply with the conditions of the

  permit and we may  get  erroneous results ty blindly

  applying the methods which  result-  in arg-.,:::ents that

  have enforcement penalties, rather than just technical

  discussions attached to them.

            If at  all  possible, we would like to have

  as many of those technical  uncertainties ironed out

  rather than be a result of  a dispute betvreen a permit 13

  agency and one of  our  plants.

            MR. RJORDAN: Yes.

            MR. KAGEL:   We  would like to know what  the

  status of the 20 lab round  robins was, in terms of

  promulgation of  the  methods.  Are we going to  get a

  chance to see that data prior to promulgation?

            MR. RIORDAN: That is one of the things  we  are

                       NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVFHUE. NW
202) 234-4433              WASHINGTON, D.C. 20005

-------
 2

 3

 4

 5

 6

 7

 8

 9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25
n;o.i)-if.; to ha\v  to  consider w.ith th."Jt dec: 1 :ri <>n .

          JIK.  MARK.0,:   That would bo usefu]  from  our

standpoint to  mice  a  good represontat J ve  hope- fully

constructive technica]  comment, if we had a data base to

look at.  Otherwise wo are just shooting  from,  the hip.

          MR.  SCHECHTER:   It would be accurate to say

that the Agency's position at the moment  is no further

Information on this proposed rulemaking is  going to be
            t
released.  The Agency could conceivably change its

position.  At  this  point, that would be our official

position.

          MR.  KROCHTA:   The feeling is not  only  should

that data be available to us but it should  be  available

to EPA for evaluating the methods.

          MR.  RIORDATI:   That 1? the decision we  are goln

to make, you are  basically asking that vie do quite  a bit

more work on validation prior to promulgation  of the

methods.  V/e have to  decide v:hat position to adopt.

          MR.  KAOEL:   1 think a good share  of  the round

robins are well along the way now.

          MR.  MARRS:   The draft documents were available

last Spring.

          MR.  RIORDAN:  Let's take a brief recess and
see if there are any  further questions-.

           (Whereupon,  a short recess was held at

to 2: 52 p.m. )
                                                        p.m.
    (202) 234-4433
                       NEAL R. GROSS
                  COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE. NW
                      WASHINGTON. D.C. 20005

-------
 1


 2

 3

 4

 5

 6

 7

 8

 9


10

11

12


13


14

15

16

17

18

19

20


21

22

23

24


25
           MR.  R10KDAN:    Lut' ;- reconvene.  I wonOer  If

 we can summarise  and  also give everyone the opportunity

 to ask any additional  question:-,,  make any additional

 statements they  feel  would be  productive to the meeting.

           I might  start  off by trying to summarise the

 situation as  I understand it.   My understanding is

 basically the  position the representatives at this

 meeting have  adopted  is  that EPA should not promulgate
              «•
 the 30^(h) methods  until we have an adequate validation

 system, as you have defined it,  prior to putting That OUT

 for general use in  industry, for measuring chemicals

 and effluents.

           We  should carry out  a  validation system v:hich

 meets criteria.   You  have provided us with one

 recommended set of  criteria for  determining an adequate

 validation scheme.

           MR.  MARKS:   The Agency has its own also.

           MR.  RIORDAN:   You are  talking about the

 equivalency scheme?

           MR.  MARKS:   Yes.

           MR.  KAGEL:   Are you referring to the good  lab

 practices?  What  is it,  Handbook of Quality Assurance?

 I think it even proposes the 7-7-7 type validation

 scheme.  That  is  referred to i the ACS.article.  There

 are some guidelines already that exist within the

                       NEAL  R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE, NW
(202) 234-4433              WASHINGTON, D.C. 20005

-------
 1


 2


 3


 4


 5


 6


 7


 8


 9


 10


 11


 12


 13


 14


 15


 16


 17


 18


 19


 20


 21


 22


23


24


25
  Agency •

            MR. 1'IARRS:  V.'ithin  the  scientific communi t y.

            MR. KAGEL:  1 think that  adds  a  lot.


            MR. RI OR DAN:  VJould you not  agree validation


  is a flexible term, its design depends upon the use to


  which you are go-ing to put the data  and  the time at


  which you are going to be measuring the  data.


        >'  MR. >. KAGEL:  Definitely.   I think that is

  pointed out clearly in the ACS article.   If you go


  straight on an N-N-N type validation scheme using


  Walpole and Myers, you would  calculate the number of

  samples that you would need to properly  validate a method,

  you can come up with numbers  like 20-20-20 or  30-30-30


  which would be economically infeasible to  do.

            MR. RIORDAK:  The other point  I  would like to


  clarify is that you have made the suggestion that GC

  may not be an appropriate method  for measuring

  constituents and effluents.   I think you recognize it

  could be used, depending upon the wast^stream.    It


  could be applicable for use in the  GC.

            MR. MARRS:  My understanding of the  20 lab


  round robin studies is that data  should  give us a


  handle on that.

            MR. RIORDAN:  As far as the- GCMS 62^/625,

  your position is that also needs  validation prior to

                       NEAL R.  GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433             WASHINGTON. D.C. 20005

-------
 1

 2

 3

 4

 5

 6

 7

 8

 9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25
                                                     36

promulgation,  even  though it has been used under  the

conditions that  represent- adequate validation.  Is  that

correct?

          MR.  KAGEL:   That Is correct.

          MR.  DAHN:   I think you can go a little  bit

beyond that and  say  you have to do enough experiments to

determine what the  limitations are, of that methodology

and mal^e sure1. we do  not use that for permit purposes

in situations  where  it does not make sense to use it.

          MR.  RIORDAN:   Specifically, you recommend we

not promulgate methods until we have the results  of the
                             »
20 lab collaborative  tests available and evaluated, is

that another point?

          MR.  KAGEL:   Yes.

          MR. -RIORDA!;:   You also made the point that

on recovery for  the  extraction methods, it should be
                                                  *
built Into the estimate of the concentration.
          Another  point that was made was In an attemp
to build in
                    inty in the methods.
          MR. MARKS:   Precision and accuracy.

          MR. RIORDAN:   Precision and accuracy  for

a method, would  it  cover your uncertainty?

          MR. MARKS:    Yes, and they should  a] no  be
reported.
    (202) 234-4433
          MR.  RIORDAN:  As part of the QA?

                      NEAL  R. GROSS
                  COURT REPORTERS AMD TRANSCRIBERS
                     1330 VERMONT AVENUE. NW
                      WASHINGTON, D.C. 20005

-------
  2


  3


  4


  5


  6


  7


  8


  9


 10


 11


 12


 13


 14


 15


 16


 17


 18


 19


 20


 21


 22


 23


24


 25
                                                          37
                MR. MARRS:   Ye:
            MR. RIORDAN:  We  should  report  them initially?

            MR. MARRS:  Anyone who reports  a  number should.

            MR. RIORDAH:  Do  you  do  that  routinely in

  your measurements you carry out?

            MR. MARRS:  We ought  to  but we  do not.

            MR. KAGEL:  It will vary depending on the

  application.  ..Sometimes it  is done;  sometimes it is not.

  I will tell you the place where it is done  most

  frequently, a result is reported near the detection


  limit.  Then a detection limit  is  reported.
                              *
            MR. DAHN:  I think the reader has a right to

  expect if he sees a significant figure  placed in an

  analytical result without an uncertainty  attached to it,

  that the uncertainty is some small number.   Frequently

  we have been reporting numbers  where that is not true
                                                    %
  in a lot of cases, where we are a  factor  of five or ten

  off.

            MR. KAGEL:   normally it is assumed to be

  plus or minus ten percent.  I think that  is a normal

  rule of thumb.  Normally, it might not  be that much.

            A simple thing like significant figures can


  be very misleading unless you have precision and

  accuracy associated with the number; .62.5,  if it is

  plus or minus ^0 percent, the .5 does happen.

                       NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE, NW
202} 234-4433              WASHINGTON, D.C.  20005

-------
  1

  2

  3

  4
 10


 11


 12


 13


 14


 15


 16


 17


 18


 19


 20


 21


 22


 23


24


 25
            There 1 r, amn.hei1  thing  T.  think we have overt'os1!'.

  and that is roporr.ing 'thing?  at the detection limit, the

  use of less th'in.  It is a  very ambiguous term if you

  do  not detect anything and  you record it as being

  not detected, giving the detection  limit in parenthesec.

            There are people  that still will report less

  than the detection limit.   That could mean it was there

  but it was less than the detection  limit or it could
        >

  mean it was not there at all.

            I think we need some strides in resolving the,"-'

  difficulties.

            MR. KITES:  When  you are  talking about

  precision, accuracy, sampling errors is always an.ong

  the highest.  How would that  get  into the procedures?

            HR. KAQEL:  There is some reference in the

  ACS paper.

            MR. NORRIS:  I think Ron  you have hit on what

  is  an important issue.  Vie  were talking about taking

  the methods which EPA proposes to promulgate and apply

  them to a compliance situation and  then we have as of

  now undefined, variations in the  analytics, in the

  sampling technology.  These could be added or subtracted

            We are talking about from the industrial

  perspective, a live or death situation.  Somebody could

  go to jail.  Often times the person who is in the field

                       NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433             WASHINGTON, D.C. 20005

-------
 3


 4


 5


 6

 7


 8


 9


 10

 11


 12


 13


 14


 15


 16


 17

 18


 19


 20


 21

 22


 23


24


 25
                                                      39

  represent:; ng the Agency or  tho  state  rogu] atory body n,:;i v

  not. be rn.'are o f the se 'uncei l a J nties .   To him,  it  is blacl

  and white;  you are in compliance  or you are not.

            II is this which  goes to the heart of our

  concern.

            MS.  SHAW:  After  it has been promulgated, I

  think it  has to be clear what the variations are,

  because.it  is. going to be useful  for  other  things  and
              t
  lots of quality decisions and decisions that are  going

  to involve  a lot of expenditures.  You are  going  to hnvp

  states that are going to be  doing state water  quality

  standards and state laboratories measuring  ambient water

  quality and deciding whether or not they have  to  have

  a state water quality standard.

            They are doing waste  load allocations.   They

  are going to do all this with methods which right  now

  are reproducible to plus or  minus five times.

            MR.  RIORDA!::  I agree,  the  method should be

  appropriate to the ends to  which  it Is going to be put.

            MS.  SHAW:  Once they  have been promulgated, wo

  do not think the methods are sufficient right now.

            MR.  RIORDAH:  All  right.  The validation

  procedure you recommend would be  something that we

  adopt from the literature or from some variation of the

  20 lab collaborative tests  or some  other good practice

                       NEAL R. GROSS
                   COURT REPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE, NW
(202) 234-4433             WASHINGTON, D.C. 20005

-------
  1


  2

  3

  4

  5

  6

  7

  8

  9


 10

 11

 12


 13


 14

 15

 16

 17

 18

 19

 20

 21

 22


23

24

25
  bur  something more than you have  seen so far In the

  promulgation we have made.  Thyt  Is  the sum and substance

  that  I have gotten out of the meeting.

            MR.  MARRS:  Certainly AST.M would be another

  source.

            MS.  SHAW: There has not  been  a round robjn

  on  621) and 625. There is really no inter-laboratory data.

            MR.. 'KAGEL:  I was told  there  was but that was
        >

  only  a rumor.

            MS.  SHAV;:  Nothing was  mentioned when they

  were  promulgating it.

            MR.  MARRS:  I have a couple of things.   A

  transcript of this meeting is going  to  sound pretty

  negative and I think all of us realize  we are working

  with  state-of-the-art methods.  It is just going to take

  a  lot of work to take a state-of-the-art method and get

  it  to a point where It can be used routinely at another

  plant.  We have to do cur homework right.

            I guess that is our position.

            Speaking for API, we would certainly be

  willing to participate and help in any  way by providing

  laboratory services or whatever.

            MR.  KROCHTA:  We mentioned this 20 lab

  round robin study.  The five plant study has a lot of

  built In sampling procedures to give you interlaboratory

                       NEAL R.  GROSS
                   COURT RCPORTERS AND TRANSCRIBERS
                     1330 VERMONT AVENUE, NW
202) 234-4433              WASHINGTON, D.C. 20005

-------
  1


  2


  3


  4


  5


  6


  7


  8


  9


 10


 11


 12


 13


 14


 15


 16


 17


 18


 19


 20


 21


 22


 23


24
|  prec:i nion as well  as  intral aborntory prec 1 si on.  The

  data corning from that V,t ndy  should be extremely useful ,

  not only to us but  to EFA as well.

            I am not  sure;  we  have three plants  oomp]eted.

            MR. KAGEL:  Two to go.  That should  end up

  this month or ne?ct  month.    VJe are anticipating a

  termination date of Hay,  1981.

          ,  MR.. KROCHTA:   Do you have accuracy as well
              *
  as precision on data  obtained from that?

            MR. KAGEL:  V.'e  have done replicas.   V'e have

  done spikes. We will  have inter- arid intralaboratory

  variability along  with  it on a limited basis.  There

  will be a lot of data available and we will make those

i  available.

            MR. KROCHTA:   It should take into consideration


  the sampling problem  as well, in splitting samples.

            MR. RIORDAN:   Does anyone from EPA have any

  questions or any points to be clarified?

            MR. MEDZ:   Most o^ the things that have been

  said have been said in  the official comments and

  received.  They will  be taken into consideration.

            MR. RIORDAM:   That is the purpose of the

  meeting, to  simply reaffirm the fact that we are taking

  these into consideration and the decisions made will

  reflect deliberation  on the points made at this meeting.

                        NEAL R.  GROSS
                   COURT REPORTERS  AND TRANSCRIBERS
                      1330 VERMONT AVENUE, NW
(202) 234-4433              WASHINGTON, D.C. 20005

-------
 6

 7

 8

 9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25
                  We  appr-ec i ate your oomlnr,.   Thank you.

                   (Whoreunor, 'the  me"! 1 rip, war, adjourned at
       3:10 p.m.)
     (202) 234-4433
     NEAL R. GROSS
COURT REPORTERS AMD TRANSCRIBERS
   1330 VERMONT AVENUE, NW
    WASHINGTON, D.C. 20005

-------
          -  • r>
                                                            I
        "  Street
mois  60604

-------
                    C E R T I F
A
.H
          This  is to certify that  the attached proceedings

in the aforecaptioned matter were  held on January 5,  19Cf

and that this  is a true and accurate record thereof  and

that this is the" original transcriot thereof.
                                         R.
                         MEAL R.  GROSS
                    COURT REPORTERS AND TRANSCRIBERS
                       1330 VERMONT  AVENUE. NW

-------

-------