EPA Protocol Revisions – Traceability for Gaseous Calibration
Standards

Include a procedure specifically for the use of Fourier Transform
infrared Spectroscopy:

The current document’s purpose is to provide traceability and ensure
accuracy of the analysis. It is not designed to provide detailed
instructions on doing the analysis on a specific analyzer. If FTIR were
to be added, then various GC detectors and other process analyzers would
need to be added making the document cumbersome and limiting the new
technology that could be used.

Improve statistical procedures for uncertainty calculations and revise
the Excel spreadsheet:

We see no real reason for changing the statistical calculations and the
Excel spreadsheet. In fact, there are numerous other acceptable methods
for calculating total uncertainty that could be used. 

Include mercury calibration standards in the protocol: 

There appears to be no reason to specifically add Mercury to Table 2-3
or create a special section for mercury or any other compound. However a
footnote for Table 2-3 that makes it clear that the components listed
are not the only ones that can be made as EPA Protocols. As long as new
traceable standards exist, and the protocol can be followed, then a
component can be called an EPA Protocol. 

Eliminate (or retain) gas manufacturers intermediate standards (GMISs):

The purpose of the EPA protocol is to ensure accuracy and traceability.
Changes that should be considered should help improve these two aspects.
GMISs are standards that are unverified by EPA or NIST and as such are
suspect. In fact they are made using a similar methodology to making EPA
Protocol mixes, so if a company cannot accurately make protocols, then
it is likely that they will have significant errors in their GMISs that
will carry over to samples that they analyze. Even using the best of
techniques, the uncertainty of the GMIS has to be greater than that of
the SRMs, NTRMs, or PRMs due to the propagation of error. As a result of
these two issues, using GMISs inherently is less accurate and invites
problems. Using the verified metrology standards is the best guarantee
for eliminating or minimizing the uncertainty contribution from the
reference standards.

Lengthen the certification periods for EPA Protocol Gases:

The current expiration dates are valid from the perspective of stability
and use.  Most users will run out of gas in about 18 months, so the 24
and 36 months expiration periods for stable components are sufficient.
In regard to reactive components at low concentrations, increasing the
expiration dates runs the risk of failure due to instability, which was
not identified during the 2 required triad analyses.  The rate of
reaction typically slows as the concentration of the reactive component
decreases, leading to the potential for component instability.  The
requirement to recertify these mixtures after the initial 6-month period
is designed to prevent  long-term stability problems in the field.

Change the cylinder pressure limitations for EPA Protocol Gases:

We see this as a two-part issue. We believe the limit of 150 psig, below
which the Protocol cannot be used, should remain unchanged. We have seen
instances where wall effects in the cylinder resulted in concentration
changes at low pressure.

The second part of the issue relates to the 500 psig limit for
recertification of mixtures. We agree with lowering that limit, but
certainly no lower than 150 psig. There are instances, where a Protocol
has expired while in use, and being able to recertify it helps the end
user with the state or federal agency that monitors that facility.

Include a procedure for analyzing and certifying zero gases:

There are no trace standards currently available from NIST to analyze
most of the contaminants of interest. These would have to be developed
in conjunction with any specific procedure for analyzing and certifying
these types of gases. Currently, some manufacturers do not specifically
analyze the zero gases for individual impurities, and just “assume”
that they are clean, and of course, this is not always the case. This
may be more involved than initially thought and perhaps should be an
addendum to the protocol. Otherwise we are in favor of this.

Include gas dilution systems in the protocol:  

The G2 protocol includes mention of dilution. If the suggestion is to
add specifics on what specific types of systems/ devices can be used and
how they are to be used, we feel this would be too restrictive and
beyond the scope of the intent of the protocol.

Tighten documentation requirements: 

Field verification of conformance of a calibration gas mixture to the
protocol is an important compliance tool.

In order to facilitate calibration gas verification, a full disclosure
certification should include: 

Cylinder identification number

Certified minor component concentration(s)

Balance gas

Cylinder pressure at time of certification

Assay / certification date

Certification expiration date

Reference standard identification (standard number, cylinder number,
minor component concentration, expiration date)

Statement that the assay / certification conforms to EPA Protocol
guidelines

Analytical methodology used for certification

Laboratory identification (name, address, telephone number)

Record of previous EPA protocol gas certification

Analytical accuracy for each certified minor component

Statement that the certified concentration has been corrected for
interference, if applicable.

Analyzer identification (make, model, serial number, measurement
principle, and last multi-point calibration date)

All triad analyzer readings used to obtain the certified minor component
concentration(s)

Correlation coefficient and calibration curve equation used to calculate
the certified minor component concentration(s).

Specific Total Oxides of Nitrogen reference value for all Nitric Oxide
protocol gas cylinders.

Having all this information will allow end users as well as auditors/
inspectors to more readily identify where expired standards were used,
falsification of the ZRT Triad analysis, inappropriate analyzers used
for analysis without addressing interference issues, as well as
independent verification of the mathematics used to calculate the
concentrations against a curve.

 Develop audit participant identification number for regulatory data
reporting purposes:

There is no value in regard to strengthening the purpose and intent of
the existing protocol, as the manufacturer’s contact information is
already on the Certificate of analysis.

Require ISO 17025 accreditation for producers of environmental
calibration gases:

ISO 17025 is an extremely expensive certification and does not guarantee
that a better protocol product will be manufactured. In fact, a gas
manufacturer, holding certification to the ISO 17025 standard,
registered at least 1 failure in the recent blind audit. The high cost
of accreditation to the ISO 17025 standard would create an economic
inequity among gas manufacturers, creating a competitive advantage
favoring large manufacturers.

Technically, one of the core concepts of the ISO 17025 standard is the
manner by which uncertainty is propagated.  In order to assert
accreditation of a laboratory to the ISO 17025 standard as a
requirement, the protocol would need to be changed so the calculation of
uncertainty would match the conventions dictated by ISO 17025.  As
previously stated (see item 2), we do not view this as a valuable change
to the existing protocol.

Include provision for preparation and analysis of batches of EPA
Protocol gases:

Batch analysis of protocol mixes does not move us in the direction of
making better product. It will reduce costs but expose the customer to
possible problems that can and will occur since homogeneity is not
guaranteed. NIST has seen problems of homogeneity with batch preparation
of SRMs, and those same problems occur when preparing Protocols. In
fact, the problems will be greater since many protocols are now
multicomponent mixtures, and homogeneity is more difficult than with a
single component Protocol.

As an aside, some producers analyze NOx in NO Protocol mixes on a batch
basis, which is erroneous. Scott recommends that EPA require that NOx
values that are assigned to NO protocol gas mixtures be actual values
and not batch values since end users need to know the actual NOx value
for each cylinder of NO Protocol.

 Require on-site visits of specialty gas producers by EPA
representatives:

This would provide no value to the end users. A producer would do the
“right” things if and when an audit would occur, and the costs would
be prohibitive. Better to spend the resources of EPA auditing end users,
and if problems are found there, test those protocols that are
questionable.

 Provide for technical assistance/outreach to specialty gas producers by
NIST     representatives:

This should not be part of the protocol. Any gas company can approach
NIST with questions regarding product production, whether it is related
to protocols or not. At the same time, having NIST and/or EPA auditing
gas producers opens up the possibility of the auditors passing on
technical suggestions for improvement that originate in one producer and
would provide an advantage to a competitor.

 PAGE   3 

 PAGE   1 

