United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
Research Triangle Park, NC 27711
Research and Development
EPA/600/S4-86/045 Apr. 1987
SEPA Project Summary
NBS/EPA Certified Reference
Material Performance Audit
Program: Status Report 2
Robert S. Wright, W. Gary Eaton, and Clifford E. Decker
A traceability procedure has been
established that allows specialty gas
producers to prepare gaseous pollutant
Certified Reference Materials (CRMs).
The accuracy, stability, and homoge-
neity of the CRMs approach those of
NBS Standard Reference Materials.
Part of this procedure is an independent
quality assurance audit of each CRM
batch by an EPA-coordinated auditor.
As of October 1986. 36 candidate
batches had been audited. Generally,
the agreement between the producer's
analytical results and the audit results is
within plus or minus 0.5 percent relative.
Additional data suggest that CRMs are
stable during their two-year certification
periods.
This Pro/ect Summary was developed
by EPA's Environmental Monitoring
Systems Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that Is fully docu-
mented In a separate report ol the same
title (see Project Report ordering In-
formation at back).
Introduction
The legal implications of many environ-
mental measurements require that ac-
curate and stable standards be used to
calibrate the monitoring instrumentation.
In the past, EPA regulations have required
that gaseous pollutant calibration stan-
dards lor ambient air quality measure-
ments be traceable to National Bureau of
Standards (NBS) Standard Reference
Materials (SRMs). Similarly, calibration
standards for mobile source emission
measurements had to be traceable to
either SRMs or other approved gas stan-
dards. The use of SRMs in these and
other applications created a demand
which occasionally exceeded NBS's pro-
duction capacity.
In response to this demand for high
accuracy standards, NBS and EPA estab-
lished a traceability procedure (EPA Pub-
lication No. EPA-600/7-81-010) by which
specialty gas producers can prepare
standards whose accuracy approaches
that of SRMs. These standards are called
NBS/EPA Certified Reference Materials
(CRMs). EPA regulations now allow CRMs
to be used as equivalents of SRMs for
ambient air quality surveillance, for con-
tinuous emission monitoring systems, for
new motor vehicle certification, and for
vehicle inspection and maintenance
programs.
The traceability procedure contains
appendices that describe the preparation
and analysis of CRMs which contain car-
bon monoxide in nitrogen or air, nitric
oxide in nitrogen, sulfur dioxide in nitro-
gen, and propane in nitrogen or air. The
procedure also contains an appendix for
recertifying CRM batches. An appendix
for preparing and analyzing CRMs that
contain oxygen in nitrogen is being
developed.
The purchaser of a CRM is assured
that its accuracy is high because of four
features of the CRM program:
1. The general technical procedure for
producing and analyzing CRMs has
been specified by NBS;
2. CRMs are directly traceable to SRMs
which are used as reference stan-
dards during the analysis of CRMs;
3. The producer's analytical results are
verified by measurements which are
performed by an independent audi-
tor; and
-------
4. Both the producer's and the auditor's
analytical results must be reviewed
and approved by NBS before the
CRMs may be sold.
Since the CRM program's inception in
November 1980, the Research Triangle
Institute (RTI) has been the EPA-coordi-
nated auditor. The Project Report sum-
marizes the CRM program in general and
the audit program specifically, including
the analytical instrumentation and the
procedures used in the audits. The results
of the CRM audits are summarized and
compared with the producers' analytical
results. Additional information is pre-
sented concerning CRM stability and
availability.
Procedure
CRMs must be produced in batches of
at least 10 cylinders having identical
concentrations. The traceability procedure
specifies that the CRM producer must
analyze the batch on two different oc-
casions that are separated by at least a
one-month interval. Further, an SRM
must be used as the reference standard
during these analyses. These analyses
must demonstrate that the batch is
homogeneous and stable. Following the
analyses, the producer sends a report of
the analytical results to NBS and a letter
to the auditor stating that a CRM batch is
available for audit.
The purpose of the EPA performance
audit program is to provide NBS with an
independent set of analytical results for
each CRM batch. NBS uses these audit
results to verify the producer's analytical
results. The audit is conducted on two
random samples from the batch. If there
is good agreement between the audit
results and the producer's results, NBS
can conclude that the producer has ac-
curately measured the two samples and,
by inference, the remainder of the batch.
The traceability procedure has several
features that help to assure that the audit
results are representative, accurate, and
precise. The audit is representative be-
cause the two audit samples are chosen
at random by the auditor, who is unaware
of the producer's results. The procedure
specifies that an SRM must be used as
the reference standard for the audit. In
addition to assuring the accuracy of the
audit, this specification allows the audit
results to be validly compared with the
producer's results because both analyses
use SRMs as their reference standards.
The precision of the audit may be quanti-
fied because the auditor must make 10
replicate measurements of the audit
samples. The audit must be repeated if
standard deviation of these measure-
ments exceeds 0.5 percent relative.
The audit procedure being used by RTI
is adapted from that used by NBS in its
analysis of SRMs. The audit samples, the
reference SRM, and a zero gas are
plumbed by a solenoid valve manifold to
the analytical instrumentation. The gases
are not diluted for the analysis. A timer
controls the solenoid valves and produces
a measurement cycle of zero gas-SRM-
CRM (1) - CRM (2), which is repeated at
least 10 times. An instrument zero value
and an instrument sensitivity are deter-
mined for each CRM measurement by
interpolation between the two adjacent
zero gas measurements and the two ad-
jacent SRM measurements. These cali-
bration values are used to calculate
concentrations for the audit samples. In
addition to the audit measurements, three
SRMs of differing concentrations and the
zero gas are measured at least three
times to determine the linearity of the
instrumentation.
Following the audit, a written report of
the audit results is sent to NBS for review.
All audit and linearity check measure-
ments are included in the report. The
report also includes information concern-
ing the analytical instrumentation and
procedures, the reference standards, and
a statistical analysis of the data.
NBS reviews the two sets of results.
The producer's results must demonstrate
that the batch is homogeneous and stable.
The audit results must be in good agree-
ment with the producer's results. If NBS
is convinced that a batch has satisfied
the requirements for CRMs, both the
producer and EPA will be notified that
the batch may be sold as CRMs. In the
event that NBS is not convinced of the
stability, homogeneity, or accuracy of the
batch, the producer or the auditor may be
required to perform additional analyses.
Alternatively, audit samples may be sent
to NBS for analysis. In this manner, NBS
remains the final arbiter concerning the
stability, homogeneity, and accuracy of
each CRM batch and ensures its credibility
as a SRM substitute.
Results and Discussion
As of October 1986, audits have been
performed on 36 candidate CRM batches
containing carbon monoxide in nitrogen
or air, nitric oxide in nitrogen, and propane
in nitrogen. For all 36 batches, acceptable
agreement was found between the pro-
ducer's results and the audit results. Two
of these batches could not be certified
because of problems not associated with
the audit (i.e., batch instability and the
producer's analytical precision). The fre-
quency distribution of the absolute value
of the agreement between the producer's
and the audit results is given in Figure 1.
In general, the absolute value of the
agreement has been within 0.5 percent
relative. The mean absolute value for the
72 audit samples is 0.18 percent relative,
and the maximum absolute value is 0.60
percent relative. This level of agreement
is within the limits set for the program
and demonstrates that specialty gas pro-
ducers can prepare and analyze CRMs
with a high degree of accuracy.
The coefficient of variation (CV) for the
10 or more measurements in each audit
can be used as an index of precision for
the audit program. The CV varies with
the instrument used for the audit mea-
surements. Non-dispersive infrared
analyzers were used for carbon monoxide
CRM audits at concentrations of 1 percent
or less. The 44 audit samples in this
concentration range had a mean CV of
0.05 percent relative and a maximum CV
of 0.14 percent relative. A gas chromato-
graph with a thermal conductivity detec-
tor was used for carbon monoxide CRM
audits at concentrations of 2 percent or
greater. The 10 audit samples in this
concentration range had a mean CV of
0.35 percent relative and a maximum CV
of 0.56 percent relative. A chemilumine-
scence analyzer was used for nitric oxide
CRM audits. The 10 nitric oxide audit
samples had a mean CV of 0.09 percent
relative and a maximum CV of 0.19 per-
cent relative. A gas chromatograph with
a flame ionization detector was used for
propane CRM audits. The four propane
audit samples had a mean CV of 0.19
percent relative and a maximum of 0.27
percent relative. The CVs for all of the
audits are within the limits specified by
the NBS/EPA traceability document, al-
though the precision for the high con-
centration carbon monoxide audits is not
as good as is desirable.
During the development of SRMs, NBS
found that the concentrations of carbon
monoxide, nitric oxide, sulfur dioxide, and
propane mixtures contained in aluminum
cylinders were stable for periods of two
or more years. Accordingly, the original
traceability procedure for CRMs specified
that a CRM batch may be certified for two
years following the batch's second
analysis. If no significant concentration
changes were found between the first
and second analyses, then it was highly
probable that the batch would remain
stable during the two-year certification
period. However, a batch could not be
-------
given an unlimited certification period.
Concentration trends which would not
affect the certified concentration over a
two-year period may become significant
after more than two years.
As CRMs were beginning to be pro-
duced, NBS and EPA decided to investi-
gate whether the CRMs' concentrations
would remain stable through their two-
year certification period. As a result, an
informal long-term stability study was
conducted with the cooperation of the
CRM producers. During the winter of
1983, nine selected audit sampleswere
reanalyzed. The time between the original
audit and the long-term stability study
reanalysis ranged from 11 to 25 months.
A histogram of the concentration change
which was observed between these two
measurements is given in Figure 2. Most
of the concentration changes were
statistically significant. However, the
magnitude of these changes was small in
comparison to the overall CRM uncer-
tainty. The largest variation observed was
0.29 percent relative after 25 months,
which is well within the total uncertainty
of 1 percent relative for the sample's
certified concentration. One cannot make
broad statements concerning the long-
term stability of all CRMs based on mea-
surements of a few samples; neverthe-
less, these data are encouraging and
suggest that CRMs remain stable during
their certification periods.
The traceability procedure contains an
appendix which describes how a producer
may recertify a CRM batch after its two-
year certification period has expired.
Basically, the recertification involves the
reanalysis of all unsold members of the
batch, using the same analytical techni-
que that was used in the original analysis.
The reanalysis must demonstrate that
the CRM's concentration has not changed
significantly since the first analysis. A
performance audit is not needed for
recertifications. As of October 1986, seven
CRM batches have been recertified.
Among the CRMs that were reanalyzed,
only one showed signs that it might be
unstable.
Conclusions
This Project Summary deals with the
performance audit program for CRMs.
The audit results have been compared
with the producer's results for each of 36
candidate CRM batches. In general, the
agreement between the two sets of re-
sults has been within plus or minus 0.5
percent relative. This comparison demon-
strates that specialty gas producers can
prepare and analyze CRMs with an ac-
curacy approaching that of SRMs. Limited
data have been presented concerning the
long-term stability of CRMs which sug-
gests that they are stable during their
certification periods.
0.1 0.2 0.3 0.4 0.5 0.6
Absolute Value of Percent Difference (percent relative)
0.7
Figure 1.
Frequency distribution for the percent difference between producer and audit
results for CRM audit samples
Percent Difference =100 f (Audit Result - Producer Result) |
L (Audit Result) J
|
-0.4 -0.2 0.0 +0.2
Concentration Change (percent relative)
+0.4
Figure 2.
Frequency distribution for the concentration change between the first and second
analysis of selected CRM audit samples
-------
Robert S. Wright. W. Gary Eaton, and Clifford E. Decker are with Research
Triangle Institute, Research Triangle Park. NC 27709.
DarrylJ. von Lehmden is the EPA Project Officer (see below).
The complete report, entitled "NBS/EPA Certified Reference Material
Performance Audit Program: Status Report 82," (Order No. PB 87-140 4637
AS; Cost: $ 11.95, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Umt«d States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
::':; C.2 Z:;
Official Business
Penalty for Private Use $300
EPA/600/S4-86/045
0000329 PS
U $ ENVIR PROTECTION AGENCY
REGION 5 LIBRARY
230 S DEARBORN STREET
CHICAGO It 60604
i.II..ii....!!..It..-!..!.i..n
------- |