    SCOTT-MARRIN, INC.

6531 BOX SPRINGS BLVD. • RIVERSIDE, CA 92507-0725

PHONE (951)653-6780  •  FAX (951)653-2430  •    HYPERLINK
"mailto:jack@scottmarrin.com"  jack@scottmarrin.com 

January 30, 2005

Robert Wright,

Air Pollution Prevention and Control Division

US Environmental Protection Agency

Research Triangle Park, NC 27711

wright.bob@epa.gov

References:

1) “Request for Suggestions for Revisions to EPA Traceability Protocol
for Gaseous Calibration Standards”, Robert S. Wright, email dated
December 23, 2004.

2) “EPA Traceability Protocol for Assay and Certification of Gaseous
Calibration Standards”, Report Number EPA-600/R-97/121 dated September
1997

3)   HYPERLINK
"http://www.epa.gov/airmarkets/monitoring/calgasauditfailureratebyrefere
ncestandard.xls" 
http://www.epa.gov/airmarkets/monitoring/calgasauditfailureratebyreferen
cestandard.xls 

Dear Bob,

Thank you for the opportunity to respond to your Request for Suggestions
for Revisions to the EPA Traceability Protocol (Ref 1).  Our company,
Scott-Marrin, Inc. has been producing gas mixtures since 1974.  Our
mixtures have been audited by EPA beginning in the late 1970s and we
have been producing EPA Protocol mixtures since the middle 1980s when
the first EPA protocol documents were published.  We are a small company
competing against very large companies and the majority of our business
is producing EPA protocol mixtures.

My first comment addresses your proposal to revive the audit program by
contracting with NIST to perform the audits with the costs of this audit
program being borne by the EPA Protocol Gas producers themselves.  This
is a departure from all previous audits where the audits were performed
by third party contractors presumably selected on a competitive bid
basis and the costs were provided through the budget of the USEPA.  I
have two problems with your proposed approach.

First, although NIST is qualified to provide accurate analyses of gas
mixtures, their prices to provide analytical services and program
management are sure to be much higher than qualified private sector
companies presumably contracted on a competitive bid basis.  Although I
have no first hand knowledge of what EPA was paying for the previous EPA
Protocol Gas audits, I am very familiar with the prices of NIST SRMs and
the price to have NIST periodically recertify these SRMs when they
expire.  We consider NIST pricing to be very expensive, but necessary to
achieve accurate SRM gas standards traceable to NIST.  I believe that it
would be much more suitable for NIST to be the auditor of a
competitively selected private sector Protocol Gas auditor.  In any
case, a reliable, independent audit of the work of any EPA Protocol Gas
auditor, including NIST, should be a requirement.

Second, since our company is a producer of EPA Protocol Gas mixtures, it
is proposed that we directly pay for the audit services.  I am concerned
about the equity and affordability of this new approach.  This is
especially true since our company is one of the smallest companies
currently manufacturing and selling our own EPA Protocol Gas mixtures. 
Our manufacturing competitors have an annual gross income which is 20 to
1000 times our annual gross income.  I am very concerned about a small
company such as ours being required to pay a currently unknown large
amount of money relative to our net income to participate in the audit
program compared to the small amount relative to the net income of the
large specialty gas companies.  The relative sizes of our competitors
would unfairly penalize small companies such as ours if this audit
proposal is adopted.  Since Congress has presumably seen fit to
inadequately fund EPA for this activity, I am suggesting that alternate
funding sources, such as the much larger user community who are
currently regulated by EPA and who benefit from using EPA Protocol
Gases, be sought to pay for the audit function.  The audit program costs
spread over a much larger population will not burden any small group.

In general, I consider the EPA Protocol-97 document (Ref 2) to be an
excellent guide to producing quality gas mixtures which doesn’t need a
lot of modifications.  I also believe that the use of performance
specifications in the present EPA Protocol document is the best
approach, while leaving the exact details of how to specifically meet
the performance specifications up to the individual gas manufacturers.

Regarding the list of informal suggestions which you have supplied, I
have the following comments:

-include a procedure specifically for the use of Fourier Transform
Infrared Spectroscopy;

	Comment:  Does this suggestion imply that EPA is going to include a
recommended list of analyzers or methods to be used to analyze EPA
Protocol Gases?  Or is this instrument being singled out for inclusion
in the Protocol document because of its inherent operational complexity?
 If this is the case, then standardized operating procedures should be
included in the Protocol document.

- improve statistical procedures for uncertainty calculations and revise
the Excel spreadsheet;

	Comment:  The EPA Protocol-97 statistical procedures for uncertainty
calculations are very sophisticated and although I am not a statistician
nor do I have a statistician on my staff, the calculations have always
appeared to be inappropriate when considering the very limited number
data points, namely 3, which they are being applied to.  In their
defense, the EPA Protocol-97 calculations give a desirable result and
they are useful in finding problems, however, whether they truly
represent the accuracy of the analyses is questionable.  Perhaps this is
why you are considering a change in the calculations?

- include mercury calibration standards in the protocol;

	Comment:  We don’t have any experience with these standards; however,
if there is a need, they might as well be included.

- eliminate (or retain) gas manufacturers intermediate standards
(GMISs);

Comment:  The elimination of the use of the GMIS is the MOST TROUBLING
SUGGESTION on your list.  As shown on your summary spreadsheet of the
most recent EPA Protocol Gas audit (Ref 3.), the use of a GMIS does not
result in failures of mixtures to meet the EPA accuracy specifications. 
In fact, one might conclude from your spreadsheet that the use of SRMs
and NTRMs leads to higher failure rates.  I’m not suggesting this is
an accurate interpretation; however, it certainly is an accurate
interpretation that the GMIS is NOT an impediment to achieving accurate
analyses.  In addition, my review of past EPA gas audits dating back to
the early 1980s shows that our company, Scott-Marrin, Inc. is one of the
only if not the only manufacturer that has provided gases which have all
had an accuracy of better than the EPA 2% specification.  Since we have
always used the GMIS or its predecessors as intercomparison standards
when performing our EPA Protocol analyses, our accuracy achievement is
due to our focus on using accurately analyzed GMIS reference standards
and carefully following the published EPA procedures.

In addition, eliminating the use of the GMIS would require the use of
either an SRM or an NTRM.  Since the SRMs contain too small a volume of
gas to use as daily intercomparison standards, the NTRMs will be the
sole reference standards to use for EPA Protocol Gas production.  Our
company is too small to produce NTRMs; therefore, we would be forced to
purchase very expensive NTRMs from our competitors.  This assumes, of
course, that ALL NTRMs are available for purchase any time they are
required.  This availability is nowhere near guaranteed and based on
past experience is probably not going to happen.  A review of the
internet sites and catalogs of all known producers of NTRMs shows no
listing of available NTRM products, let alone prices for these products.
 A recent telephone conversation with a large producer of NTRMs
confirmed the lack of a guaranteed availability for commercial customers
of these products.  In addition, the availability of every NTRM as a
companion to an existing SRM is also not assured.  Since the use of the
GMIS does not pose a technical problem as evidenced by your recent audit
data, I am very suspicious that the elimination of the GMIS is a
suggestion by one or more of the NTRM producers to achieve a competitive
advantage over other manufacturers.  One final point; even though NTRMs
are declared to be equivalent to SRMs, they are to my knowledge,
secondary to SRMs.  The GMIS, if it is compared directly to an SRM, is
also secondary to an SRM.  As a point of information, our company
performs additional work on the GMISs which we produce.  For example, we
initially analyze each GMIS 4 or more times over a 3 month period and we
analyze each GMIS every year rather the required 2 years.  We feel these
additional analyses increase the accuracy and reliability of the GMIS
and, by the way, the annual analyses exceed the analysis requirements of
NTRMs.  Please do not eliminate the GMIS.

- lengthen the certification periods for EPA Protocol Gases;

	Comment:  These comments are the result of our experience performing
EPA Protocol Gas recertifications over many years.  They also err on the
conservative side.  I apologize for not having supporting documentation
to submit.

a) The recertification interval for CH4, O2, CO2, and C3H8 can certainly
be greater than 3 years.  Although these gas mixtures are usually used
up in 3 years, these mixtures will certainly be stable for 4 to 6 years
and probably much longer.

b) A 36 month interval for CO from 8 to 50 ppm seems to be a prudent. 
We have never seen CO change at higher concentrations in aluminum
cylinders, so 4-6 years is acceptable above some threshold around 20 to
50 ppm.

c) The 4 ppm Nitric Oxide concentration range for 24 months specified
in the EPA Protocol-97 document is too low a threshold.  The lower
threshold for Nitric Oxide mixtures should be more like 20 ppm to
warrant a 24 month recertification interval.  For concentrations in the
range of 4 to 20 ppm, a range of 12 months would be more appropriate. 
For less than 4 ppm Nitric Oxide, a 6 month initial interval followed by
a 12 month interval if successfully demonstrating concentration
stability seems appropriate.

500 ppm SO2 @ 36 months should remain unchanged.

e) The current interval of 4+ ppm H2S@ 12 months should remain
unchanged.  If higher concentration H2S SRMs are made available then a
24 month interval should be considered for concentrations exceeding 50
to 100 ppm.

- change the cylinder pressure limitations for EPA Protocol Gases;

	Comment:  The current recertification pressure limit of 500 psig is
about right.  Below this concentration, the recertification is not cost
effective anyway.  The minimum use pressure limit of 150 psig could be
lowered to 100 psig without problem.  Most gas mixtures in aluminum
cylinders are going to be stable to less than 25 psig, however, some
‘reactive’ gas mixtures may not be.

- include a procedure for analyzing and certifying zero gases;

	Comment:  This would be very desirable.  Several different standardized
grades should be specified to include the range of EPA Protocol Gas
applications.

- include gas dilution systems in the protocol;

	Comment:  This too would be very desirable.  Performance specifications
and calibration standards and methods are required.

- tighten documentation requirements;

	Comment:  I have no idea what this is referring to.  The documentation
requirements specified in the EPA Protocol-97 Gas document seem
appropriate and complete.

- develop audit participant identification number for regulatory data
reporting purposes;

	Comment:  OK, if this really has a useful purpose.

- require ISO 17025 accreditation for producers of environmental
calibration gases;

	Comment:  We are not currently ISO certified nor do we have any
expertise with ISO certification.  However, in looking into the ISO
17025 accreditation standard, it would seem to duplicate some of the
existing EPA Protocol-97 Gas document requirements.  I believe the EPA
Protocol Gas document is a sufficient controlling standard that can be
used to achieve the goal of producing accurate EPA Protocol Gas
standards.  Therefore, I don’t think it is necessary or desirable to
be ISO 17025 accredited to achieve the EPA Protocol Gas program goals. 
I also don’t think it should be required, since it would be costly and
favor large companies. 

- include provision for preparation and analysis of batches of EPA
Protocol Gases;

	Comment:  To my knowledge there is no provision against preparing EPA
Protocol Gases in batches.  Therefore, in the absence of specifics, this
suggestion implies that a procedure for the batch analysis of EPA
Protocol Gases be considered.  This would seem to defeat the main tenet
of EPA Protocol Gases, namely that of maintaining quality by
individually certifying each and every mixture to eliminate errors.

- require on-site visits of specialty gas producers by EPA
representatives;

	Comment:  OK.  My only objection would be the cost to perform this
level of audit.  Perhaps this should be considered for those specialty
gas producers who fail an audit.

- provide for technical assistance/outreach to specialty gas producers
by NIST representatives;

	Comment:  Isn’t his already available?

I only have one additional recommendation.  NIST should be encouraged to
produce the following additional SRMs to support the EPA Protocol Gas
program.  50 and 100 ppm Hydrogen Sulfide in Nitrogen, 250 ppm Sulfur
Dioxide in Nitrogen, 5 ppm Carbon Monoxide in Air, and 50, 100 and 1000
ppm Methane in Air or Nitrogen are not currently available and their
availability would improve the EPA Protocol Gas program.



d

e

f

g

q

r



b

c

ê

ñ

=r

…

†

 t form is very good and could only be a little better with some
additions and a little tweaking.  In summary, the unspecified costs of
your audit proposals and the elimination of the GMIS are very bad ideas
which I feel would seriously impact our ability to produce EPA Protocol
Gases and thereby threaten the existence of our company.

Please feel free to discuss any of the above comments with me. I can be
contacted by phone at 951-653-6780 or by email at jack@ scottmarrin.com.

Best regards,

Jack Marrin

President

