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
-------
------- |