ISO/TS 14167:2003

TECHNICAL ISO/TS
SPECIFICATION 14167
First edition
2003-04-15
Gas analysis — General quality
assurance aspects in the use of
calibration gas mixtures — Guidelines
Analyse des gaz — Aspects généraux de l'assurance qualité dans
l'utilisation de mélanges de gaz pour étalonnage — Lignes directrices

Reference number
©
ISO 2003
PDF disclaimer
This PDF file may contain embedded typefaces. In accordance with Adobe's licensing policy, this file may be printed or viewed but
shall not be edited unless the typefaces which are embedded are licensed to and installed on the computer performing the editing. In
downloading this file, parties accept therein the responsibility of not infringing Adobe's licensing policy. The ISO Central Secretariat
accepts no liability in this area.
Adobe is a trademark of Adobe Systems Incorporated.
Details of the software products used to create this PDF file can be found in the General Info relative to the file; the PDF-creation
parameters were optimized for printing. Every care has been taken to ensure that the file is suitable for use by ISO member bodies. In
the unlikely event that a problem relating to it is found, please inform the Central Secretariat at the address given below.

©  ISO 2003
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized in any form or by any means,
electronic or mechanical, including photocopying and microfilm, without permission in writing from either ISO at the address below or
ISO's member body in the country of the requester.
ISO copyright office
Case postale 56 • CH-1211 Geneva 20
Tel. + 41 22 749 01 11
Fax + 41 22 749 09 47
E-mail copyright@iso.org
Web www.iso.org
Published in Switzerland
ii © ISO 2003 — All rights reserved

Contents Page
Foreword. iv
Introduction . ivi
1 Scope. 1
2 Terms and definitions. 1
3 Gas analysis using a comparison method. 4
4 Uncertainty. 4
4.1 General. 4
4.2 Step-by-step approach . 5
4.3 Direct approach. 6
5 Traceability . 7
5.1 General. 7
5.2 Calibration for response function determination . 8
5.3 Calibration for bias detection and correction . 8
5.4 Verification of individual results . 8
5.5 Traceability of reference values . 9
6 Validation and verification . 9
6.1 Validation of methods. 9
6.2 Verification of individual results . 10
7 Documentation . 10
Annex A (informative) Examples where traceable standards are available . 11
Annex B (informative) Example where traceable standards are not available. 16
Annex C (informative) Hierarchy of reference gas mixtures. 18
Bibliography . 22

Foreword
ISO (the International Organization for Standardization) is a worldwide federation of national standards bodies
(ISO member bodies). The work of preparing International Standards is normally carried out through ISO
technical committees. Each member body interested in a subject for which a technical committee has been
established has the right to be represented on that committee. International organizations, governmental and
non-governmental, in liaison with ISO, also take part in the work. ISO collaborates closely with the
International Electrotechnical Commission (IEC) on all matters of electrotechnical standardization.
International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 2.
The main task of technical committees is to prepare International Standards. Draft International Standards
adopted by the technical committees are circulated to the member bodies for voting. Publication as an
International Standard requires approval by at least 75 % of the member bodies casting a vote.
In other circumstances, particularly when there is an urgent market requirement for such documents, a
technical committee may decide to publish other types of normative document:
— an ISO Publicly Available Specification (ISO/PAS) represents an agreement between technical experts in
an ISO working group and is accepted for publication if it is approved by more than 50 % of the members
of the parent committee casting a vote;
— an ISO Technical Specification (ISO/TS) represents an agreement between the members of a technical
committee and is accepted for publication if it is approved by 2/3 of the members of the committee casting
a vote.
An ISO/PAS or ISO/TS is reviewed after three years in order to decide whether it will be confirmed for a
further three years, revised to become an International Standard, or withdrawn. If the ISO/PAS or ISO/TS is
confirmed, it is reviewed again after a further three years, at which time it must either be transformed into an
International Standard or be withdrawn.
Attention is drawn to the possibility that some of the elements of this document may be the subject of patent
rights. ISO shall not be held responsible for identifying any or all such patent rights.
ISO/TS 14167 was prepared by Technical Committee ISO/TC 158, Analysis of gases.
iv © ISO 2003 — All rights reserved

Introduction
Gas analyses are performed on samples covering a wide range of compositions.
All gas analyses fall into one of two categories:
 those for which calibration gas mixtures exist with compositions that are traceable to reference gas
mixtures (e.g. primary or national standards);
 those for which the above do not exist.
In both cases, this Technical Specification provides the overall guidance on quality assurance aspects
required to achieve a result with a valid measurement uncertainty.
It is applicable only to calibration gas mixtures of gaseous, or totally vaporized, components which do not
react with each other or with the cylinder walls.
The user of this Technical Specification chooses, prior to any analyses, an appropriate measurement
procedure depending on the application of the final results of the analyses and the requirements for a
particular measurement uncertainty. This use may vary considerably, ranging from qualitative analysis to
accurate quantitative analysis for which evidence has to be provided that claimed measurement uncertainty
levels are met. Each type of measurement procedure involves a number of issues, which are considered
beforehand. Typically, for gas analysis these include:
a) sampling;
b) selection and use of calibration gas mixtures;
c) selection and validation of measurement method;
d) identification of uncertainty sources;
e) quantification of uncertainty contributions;
f) documentation.
For a given measurement procedure, the effect of the above issues, influencing the uncertainty of the final
measurement result, is calculated approximately. This may imply that several calculations have to be made in
advance, with different sets of values for the parameters involved, before the required level of uncertainty is
achieved. In practice, this calculation process is repeated until the desired target uncertainty is reached. This
process of defining target uncertainties is an effective way of finding correct solutions for specific
measurement procedures. The final analysis is then performed using this evaluated procedure.
To illustrate the use of this Technical Specification, two practical examples are given in Annex A.
Annex B gives information on the validation of reference gas mixtures in those cases where calibration gases
of widely acknowledged composition do not exist.
Annex C describes the traditional hierarchy of reference gas mixtures.
The references are given in the Bibliography.
TECHNICAL SPECIFICATION ISO/TS 14167:2003(E)

Gas analysis — General quality assurance aspects in the use of
calibration gas mixtures — Guidelines
1 Scope
This Technical Specification provides guidance on the quality aspects in the field of gas analysis that are
implemented in order to achieve a result with a valid measurement uncertainty.
It provides guidelines on quality aspects to be employed in gas analysis using calibration gas mixtures and
their subsequent validation and/or verification, and the testing of the analytical performance of gas analysis
instruments. These guidelines have the overall objective of defining procedures which will ensure that
measurements of gas composition are reliable, comparable and consistent between different organizations
and countries.
This Technical Specification explains, in particular, the concepts of measurement uncertainty and of
traceability as effective quality assurance tools for defining the measurement uncertainty of particular
measurement results. It also gives guidance on how to identify and estimate measurement uncertainty
components of the result, and how to combine these uncertainty components in order to obtain the overall
uncertainty.
2 Terms and definitions
For the purposes of this document, the following terms and definitions apply.
2.1
traceability
property of a result of a measurement or the value of a standard whereby it can be related to stated
references, usually national or international standards, through an unbroken chain of comparisons all having
stated uncertainties
NOTE 1 The unbroken chain of comparisons is called a “traceability chain”.
NOTE 2 A calibration gas mixture is traceable at best to a primary reference gas mixture.
[1]
[VIM ]
2.2
uncertainty (of measurement)
parameter, associated with the result of a measurement, that characterizes the dispersion of the values that
could reasonably be attributed to the measurand
NOTE 1 The parameter may be, for example, a standard deviation or a given multiple of it, or the half-width of an
interval having a stated level of confidence.
NOTE 2 Uncertainty of measurement comprises, in general, many components. Some of these components may be
evaluated from the statistical distribution of the results of series of measurements and can be characterized by
experimental standard deviations. The other components, which can also be characterized by standard deviations, are
evaluated from assumed probability distributions based on experience or other information.
NOTE 3 It is understood that the result of the measurement is the best estimate of the value of the measurand, and
that all components of uncertainty, including those arising from systematic effects, such as components associated with
corrections and reference standards, contribute to the dispersion.
NOTE 4 Uncertainty can be expressed as a standard uncertainty or, when multiplied by a coverage factor, as an
expanded uncertainty.
[2]
[Adapted from GUM ]
2.3
uncertainty of a certified value
estimate attached to a certified value of a quantity which characterizes the range of values within which the
“true value” is asserted to lie with a stated level of confidence
[3]
[ISO Guide 30:1992 ]
2.4
calibration gas mixture
gas mixture of sufficient stability and homogeneity whose composition is properly established for use in the
calibration of a measuring instrument or for the validation of a measurement or gas analytical method
NOTE Calibration gas mixtures are the analogues of measurement standards in physical metrology
[4]
[ISO 7504:2001 ]
2.5
reference gas mixture
calibration gas mixture whose composition is sufficiently well established and stable to be used as a reference
standard of composition from which other composition data are derived
NOTE Reference gas mixtures are the analogues of reference standards.
[4]
[ISO 7504:2001 ]
2.6
primary reference gas mixture
reference gas mixture which is designated, or generally accepted, as realizing a specific composition of the
highest metrological quality
NOTE 1 Primary reference gas mixtures are the analogues of primary standards.
NOTE 2 Normally the use of primary reference gas mixtures is confined to comparisons with other primary reference
gas mixtures of similar compositions and to secure secondary reference gas mixtures by comparison.
NOTE 3 Primary reference gas mixtures are sometimes designated as measurement standards by national metrology
institutes, and may then be known as primary standard gas mixtures
[4]
[ISO 7504:2001 ]
2.7
secondary reference gas mixture
reference gas mixture whose composition is assigned by comparison with a primary reference gas mixture of
similar composition, or with several such primary reference gas mixtures
NOTE 1 Secondary reference gas mixtures are the analogues of secondary standards.
NOTE 2 A secondary reference gas mixture may be used as a calibration gas mixture then having traceability to a
primary reference gas mixture.
[4]
[ISO 7504:2001 ]
2 © ISO 2003 — All rights reserved

2.8
stability
attribute of a gas mixture, stored or used under specified conditions, to maintain its composition within its
specified uncertainty limits for a specified period of time (maximum storage life) and over a specified range of
pressure and of temperature
NOTE It is appropriate to specify the uncertainty limits for each component of interest.
[4]
[ISO 7504:2001 ]
2.9
homogeneity
state of a gas mixture wherein all of its components are distributed uniformly throughout the volume occupied
by the gas mixture
NOTE Unless any other indication is given, it is normally to be assumed that the gas mixture is homogeneous in
composition in time and space within the gas mixture.
[4]
[ISO 7504:2001 ]
2.10
validation
confirmation, through the provision of objective evidence, that the requirements for a specific intended use or
application have been fulfilled
NOTE 1 In design and development, validation concerns the process of examining a product to determine conformity
with user requirements.
NOTE 2 Validation is normally performed on the final product under defined operating conditions. It may be necessary
in earlier stages.
NOTE 3 The term “validated” is used to designate the corresponding status.
NOTE 4 Multiple validations may be carried out if there are different intended uses.
NOTE 5 In gas composition analysis, validation refers to the confirmation that the method, as applied, is fit for the
intended purpose.
[4]
[Adapted from ISO 7504:2001 ]
2.11
verification
Confirmation, through the provision of objective evidence, that the specified requirements have been fulfilled
NOTE 1 In design and development, verification concerns the process of examining the result of a given activity to
determine the conformity with the stated requirements for that activity.
NOTE 2 The term “verified” is used to designate the corresponding status.
NOTE 3 In gas composition analysis, verification refers to an individual result that agrees with the result of an
independent method.
[4]
[ISO 7504:2001 ]
3 Gas analysis using a comparison method
The acceptability of individual measurement results is improved when rigorous quality control provides
evidence that the results are in agreement with results from:
a) measurements on similar samples by other laboratories;
b) previous measurements on similar samples by the same laboratory;
c) previous measurements on other samples.
Whenever links are made between measurement systems, these links have to be identified (traceability chain)
and its strength quantified (uncertainty). In the following clauses, these concepts are elaborated as well as the
quality control aspects of the steps in the measurement procedure.
Calibration gas mixtures should be prepared using either static or dynamic methods:
[5]
 Gravimetric preparation of fully gaseous gas mixtures is described in ISO 6142 .
[6]
 Static volumetric methods are described in ISO 6144 .
[7] to [15]
 The various parts of ISO 6145 deal with a great number of dynamic methods for preparing gas
mixtures. Mixtures prepared in accordance with any of these parts will have stated compositions and
uncertainty evaluations.
Mixtures prepared by the methods given should first be verified before they can actually be used as calibration
[16]
gas mixtures. ISO 6143 gives a comparison method that is in most cases suited for verification purposes.
[17]
ISO 6141 describes the contents of a certificate of a calibration gas mixture.
Figure 1 gives a scheme explaining gas analysis using a comparison method. The numbers in the boxes refer
to references in the Bibliography.

Figure 1 — Gas analysis scheme using a comparison method
4 Uncertainty
4.1 General
The main task of gas analysis is to determine the composition of gases, i.e. to measure the content of one or
several specified target components (analytes) in a matrix gas. This clause mainly gives guidance on how to
evaluate and express the uncertainty of results in gas composition analysis. However, the same principles and
analogous procedures may be used to evaluate and express the uncertainty of results in the preparation of
calibration gas mixtures or in other fields of gas analysis such as the direct measurement of physico-chemical
properties of gases.
4 © ISO 2003 — All rights reserved

The uncertainty of an analytical procedure is determined in accordance with two fundamentally different
strategies:
 the step-by-step approach (also called the bottom-up approach), where the uncertainty is calculated as a
combination of various components, relating to building blocks of the overall procedure or to significant
error sources;
 the direct approach (also called the top-down approach), where the uncertainty is determined directly by
investigating the spread of results obtained on gases of known composition.
The step-by-step approach (see 4.2) requires a thorough investigation of the analytical procedure and proper
assessment of component uncertainties. Its main tools are uncertainty budgets and uncertainty propagation,
based on an appropriate mathematical model of the measurement process or the preparation process.
The direct approach (see 4.3) requires appropriate reference gases or reference analytical methods. In
addition, the comparison measurements have to be performed under appropriate reproducibility conditions.
4.2 Step-by-step approach
This approach attempts to evaluate the uncertainty of the result of an analytical procedure by identifying the
uncertainty components of that procedure. Each of the uncertainty components are evaluated to determine
[2]
their contribution and are combined mathematically using the principles described in GUM . This approach
requires the user to produce a detailed uncertainty budget in a series of steps as follows:
a) Separate the analytical procedure into a series of well-designed steps, e.g. sample pre-treatment,
measurement, data evaluation.
b) For each step, identify and list all factors that may influence the performance of this step and
consequently give rise to uncertainty.
NOTE 1 It is understood that a correction is applied to every identified source of systematic error, leaving as a
contribution to the overall uncertainty budget the uncertainty of the correction.
NOTE 2 In a number of International Standards produced by ISO/TC 158, comprehensive lists of uncertainty sources
[5], [6], [7], [16]
and assessments of their relevance are given.
NOTE 3 To prove the validity of the results, traceability to acknowledged measurement standards should be
demonstrated where available.
c) Some of the identified uncertainties will have a large influence on the uncertainty of the method, and
others only a small influence. To determine this influence, an estimate is now made of the contribution of
each of the uncertainty sources to the final combined standard uncertainty. Each of the identified
uncertainties should now be classed as being either significant or insignificant when compared to the
combined standard uncertainty. Uncertainty sources classed as insignificant should be neglected.
(Insignificant uncertainties may be classed as those which contribute less than 10 % to the combined
standard uncertainty.)
d) For the uncertainty sources classed as significant, design a mathematical model describing the final result
of the procedure, y, as a function of the input parameters (significant contributors) x (i = 1, 2, …, N), e.g.
i
sample flow, pressure and temperature, recovery rates, measured instrument response, response
function parameters, conversion factors and corrections:
y = f (x , x , …, x ) (1)
1 2 N
e) For each significant uncertainty source, evaluate the contribution u (y) to the uncertainty of the final result
i
y as a standard uncertainty, that is as a standard deviation, either of a series of repeated measurement
results (Type A evaluation) or of a hypothetical distribution expressing the available information about the
respective quantity (Type B evaluation). The contribution to the total uncertainty due to x is u (y), and is
i i
determined as a product of a sensitivity coefficient c taken from the model in accordance with step d) and
i
the standard uncertainty u(x ) of the respective input quantity:
i
uy() ≡c u(x ) (2)
ii i
∂f
where c =
i
∂x
i
f) Consider possible correlations between different uncertainty contributions u (y) and u (y), determined in
i j
step e), e.g. due to use of the same measurement standard. If the correlation is determined as being
significant, estimate the corresponding covariance u (y):
ij
uy() =ccu(x,x ) (3)
ij i j i j
g) Calculate the combined standard uncertainty u(y) of the final result y as the root of the sum of squares
of the component standard uncertainties and their possible covariances:
NN −1N
22 2
uy()=+cu (x ) 2 ccu(x,x ) (4)
ci i ijij
∑∑∑
ii==11j=i+1
4.3 Direct approach
In this approach, the analytical procedure is applied to appropriate reference samples, and the results are
compared with the reference values attributed to the reference samples. Alternatively, the procedure under
investigation and an established reference method are applied in parallel to appropriate samples, and the
results are compared. This comparison serves a double purpose:
a) By comparing the mean of repeated measurements with the corresponding reference value, any
significant analytical bias is detected. In addition, an appropriate bias correction should be derived.
b) From the spread of results of repeated measurements and the uncertainty of the reference values, the
uncertainty of the results of the analytical procedure, including a bias correction, is calculated. It is
emphasized that a proper evaluation of the variability is necessary.
A more detailed description of this comparison and its evaluation are given in 5.2.
This approach is applied either by an individual laboratory or in an interlaboratory study. In the first case, it is
essentially a calibration study. Then it is important to ensure that:
 the reference values and their uncertainties are well established;
 the calibration samples are representative of the range of samples to be analysed;
 the variability of the measurement conditions in the calibration study covers the variability in the intended
applications.
In the second case of an interlaboratory study, the procedures described in the ISO 5725 series of
[18] to [23]
International Standards should be followed.
If neither suitable reference samples nor suitable reference methods are available, the comparison may be
based on consensus values instead of reference values, as a substitute but by no means as an equivalent.
6 © ISO 2003 — All rights reserved

5 Traceability
5.1 General
Traceability is universally recognized as one of the basic prerequisites for comparability of the results of
measurements made in different laboratories and in different countries, and of the conclusions drawn from
these results. When comparing different measurement results for the same quantity, three basic requirements
have to be fulfilled:
a) the uncertainties of the measurement results have to be known;
b) the units used to express the measurement results (including the uncertainties) have to be the same, or
at least convertible;
c) the measures used to express the uncertainties have to be the same, or at least convertible.
Evidently, requirements b) and c) call for standardization. For the units of physical quantities, the
standardization problem was solved by establishing the International System of Units (SI System). Concerning
the standardization of uncertainty measures, the “standard uncertainty” proposed by the Guide to the
[2]
expression of uncertainty in measurement (GUM) has the potential of worldwide acceptance.
The remaining requirement a) is the most demanding one. In short, known uncertainty means that in the
uncertainty budget of the measurement no major contribution has been overlooked, and that the quantification
and combination of uncertainty contributions has been done correctly. Traceability, where applicable,
simultaneously solves all these problems by comparison between measured values and corresponding
reference values. Thus, where appropriate reference standards are available, traceability is a direct and
effective way of demonstrating accuracy.
A traceability chain is established by comparison between measured values and corresponding reference
values. In practice, there are different strategies for making use of such comparisons: calibration and
verification. In short, in calibration the comparison results are used to derive measurement results from raw
data, while in verification measurement results are tested for agreement with the results of an independent
reference method.
In calibration, different methods are used depending on the type of task. For example, calibration may refer to
a measuring system or to a measurement standard. The main case considered in this document is calibration
of an analytical system. Here, the purpose of calibration can either be determination and correction of bias or
determination of the relationship between measured response and analyte contents.
The traceability strategies considered in this document fall into three main categories:
 detection and correction of analytical bias;
 determination of response characteristics;
 examining agreement with the results of an independent method.
For each of these strategies, different procedures are available and the following options are available:
 single-level or multi-level comparison;
 interpolation or regression;
 uncertainty propagation or statistical analysis of residual scattering.
5.2 Calibration for response function determination
When determining the relationship between instrumental response and the target quantity — typically the
content of a specified analyte in a specified matrix gas — the response is measured using standards of known
analyte content, covering the measurement range. From the comparison between the measured responses
and the reference values, parameters of the response curve (e.g. slope and intercept of a straight line) are
derived, including the uncertainties of these parameters. Using this data, the analyte content of an unknown
sample is estimated from its measured response. In addition, the uncertainty of the estimated analyte content
is calculated from the uncertainty of the measured response and the uncertainties of the response curve
parameters.
The main requirements for this procedure are as follows:
 the calibration samples have to be representative of the range of samples to be analysed;
 the variability in the measurement conditions in the calibration study has to cover the variability in the
intended applications;
 the response curve has to be compatible with the calibration data, including their uncertainties.
By using this procedure, the uncertainty of the analytical method is traceable to reference values and includes
the uncertainties attributed to the calibration standards.
[16]
ISO 6143 specifies a method for deriving the composition, including uncertainty, of calibration gases from
a set of reference gases, using the approach outlined in this clause.
5.3 Calibration for bias detection and correction
To determine the uncertainty of the results of a selected analytical method, the method is applied to
appropriate calibration samples. The results are compared with the reference values attributed to the
calibration samples. Alternatively, the method selected and a reference method, i.e. a method whose
uncertainty is known, are applied in parallel to appropriate samples. The results of the method selected are
compared with those of the reference method. This comparison is performed and evaluated in a series of
three steps as follows:
a) The measurement results and the corresponding reference values are compared, and the differences
between corresponding values are tested for significance against the expanded uncertainty using an
appropriate coverage factor, e.g. Student’s t for 95 % confidence, and the relevant number of degrees of
freedom.
b) If significant bias was found in step a), a bias correction for the specified measurement range is derived.
For a restricted measurement range, this is a bias factor or an additive correction term based on a single-
level comparison. For extended measurement ranges, correction curves or correction functions are
required based on multi-level comparisons.
c) The uncertainty of the results of future measurements, including bias correction, is calculated by the
combination of two contributions: the uncertainty of the uncorrected measurement result and the
uncertainty of the correction. The uncertainty of the correction is estimated from the calibration data,
either by uncertainty propagation or by statistical analysis of the residual scattering.
The main requirements for this type of calibration are described in 4.3.
5.4 Verification of individual results
For the purpose of this Technical Specification, an individual result is the value of a specified measurand
obtained using a specified gas, e.g. the content of a specified analyte in a gas sample, determined either by
gravimetry or analysis, or the value of a physico-chemical property determined by direct measurement.
8 © ISO 2003 — All rights reserved

Verification of an individual result means examination whether, and confirmation that, the result agrees with
the result from an independent method (see Clause 6). Verification can provide traceability if the independent
method used is a reference method, that is if its uncertainty is known.
5.5 Traceability of reference values
For the purpose of this Technical Specification, reference values are values of gas mixture composition or
physico-chemical gas properties provided by reference standards. Reference standards have to be prepared
by a method with an established uncertainty, and verified by independent analysis where possible. Reference
standards meeting these requirements are called reference gas mixtures (see 2.5, 2.6 and 2.7). Other
calibration gas mixtures have to be made traceable to appropriate reference gas mixtures using the protocol
[16]
specified in ISO 6143 .
[24]
Annex C shows a hierarchy of calibration gas mixtures, based on ISO 14111 .
As part of the criteria for calibration gas mixtures being acceptable as reference standards, the methods of
analysis have to be thoroughly understood. A thoroughly understood method means that the analysis
methodology has a fully established and robust uncertainty budget and, in addition, the user has to supply
adequate evidence showing implementation of the uncertainty budget.
6 Validation and verification
6.1 Validation of methods
Method validation, as defined in this Technical Specification, consists of four steps:
a) specification of the analytical task;
b) determination of the method's performance characteristics;
c) comparing specifications against characteristics;
d) confirming fitness, or stating lack of fitness, for the intended use.
The main task in method validation is determining the method's performance characteristics. The requirement
is to specify a given analytical method with an application range with defined uncertainty, for example:
 Specificity and selectivity are intended to specify the range of matrix gases for which the uncertainty
budget is valid. For other matrix gases, a correction and/or an additional uncertainty component is
necessary.
 Robustness and ruggedness are intended to specify a range of operating conditions for which the
uncertainty budget is valid. For other operating conditions, a correction and/or an additional uncertainty
component are necessary.
 Linearity is intended to specify that range of analyte content for which a linear response function applies.
Beyond this range, a correction and/or an additional uncertainty component are necessary.
 Reproducibility aims at establishing a “top-down” estimate of the uncertainty of an analytical method,
including interlaboratory bias as an uncertainty component.
Validation may also include other issues such as sample size, time and costs. These items are not addressed
in this Technical Specification.
Method validation has to be distinguished from instrument qualification because the latter refers to proper
functioning rather than fitness for a specified purpose.
6.2 Verification of individual results
Verification of an individual result, as defined in this Technical Specification, consists of four steps. In this
clause, a general procedure for verification of an assigned analyte content is specified. The verification of
other results follows the same principles:
a) Establish the assigned analyte content x and its standard uncertainty u .
1 1
b) Establish the result of the independent analysis x and its standard uncertainty u .
2 2
c) Examine the agreement between these results in accordance with an appropriate statistical criterion, e.g.
by comparing x – x against the expanded uncertainty U for this difference, given by:
1 2
Uk=+u u (5)
using a suitable coverage factor k. This test is most conclusive when u and u are of the same order of
1 2
magnitude.
d) Confirm agreement or state disagreement.
In principle, the result of a verification step is a yes/no decision. However, it may be profitable to combine
agreeing results because the mean of a set of independent measurements can lead to a lower uncertainty.
In the case of a positive test result, the assigned analyte content of the calibration gas, including its
uncertainty, is confirmed by comparison with an independent analysis. If this analysis was made using an
established reference method, this amounts to establishing a link to a reference value (see 5.4).
Using appropriate statistics, the procedure outlined above may be extended to a multi-level comparison.
Results of determination of gas composition may also be verified by comparing physical-property values
calculated from an assigned composition (e.g. density) with the results of direct measurements. In this case,
the sensitivity, i.e. the degree of correlation between variations in composition and variations in property
values, has to be taken into account.
7 Documentation
For each analytical task, a uniquely identified report should be made and filed in a retrievable way. This report
should contain fully documented data on the following items:
a) specifications of the analytical assignment;
b) identification of the sample;
c) the method of sampling used;
d) the analytical method used;
e) the analytical procedure, including data on the calibration gases;
f) an assessment of all possible contributions to the uncertainty;
g) the calibration model used;
h) the calculation of the final analytical result, including the uncertainty.
[25]
Detailed information on documentation can be found in a number of standards, e.g. ISO/IEC 17025 and
[17]
ISO 6141 .
10 © ISO 2003 — All rights reserved

Annex A
(informative)
Examples where traceable standards are available
A.1 Assuring the quality of gas analysis where traceable standards are available
This example seeks to use the information given in this Technical Specification to help the user have a clearer
understanding of the steps involved in formulating a valid quality assurance method for gas analysis. This
example shows how the whole quality assurance process is iterative until the method is shown to be “fit for
purpose”.
In this example, the steps involved in producing a valid quality assurance procedure for the determination of
–3 −3
carbon monoxide in nitrogen over the range from 1 × 10 mol/mol to 5 × 10 mol/mol are covered.
Requirements:
 the analysis is to be traceable to internationally accepted standards of gas composition;
 a target expanded relative measurement uncertainty of 5 % (k = 2);
 a report at an appropriate level is to be provided.
The initial step is to confirm that traceable standards are available for the analysis and the uncertainty of the
standard is appropriate for the analysis. In this case, the measurement uncertainty requirement is 5 %. This
allows procurement of a traceable reference gas mixture from an accredited commercial supplier. If for
example the uncertainty requirement was more stringent, then a reference gas mixture with a lower
uncertainty would be required, and as a consequence the procurement would have to be made from higher in
the hierarchical chain (see Annex C).
The laboratory charged to perform this analysis has two instruments which are capable of performing the
analysis. The first uses non-dispersive infrared (NDIR), and the second involves a gas-chromatographic
method using a thermal-conductivity detector. A comparison of instrument performance is required to ensure
the most appropriate instrument is used for the analysis.
The next step involves sampling considerations. The laboratory's past experience has shown that the
materials used for construction of its sampling equipment is suitable for this analyte over the range
–3 –3
1 × 10 mol/mol to 5 × 10 mol/mol. The sampling systems of the analytical instruments have been tested
for both vacuum and pressure integrity, and the cleanliness of the system has been maintained through
regular heated purging and evacuation. Carbon monoxide is not considered as a reactive gas and, as a
consequence, purge times will not have to be excessive. However, in the case of more reactive gases such as
HCl, much longer purge and evacuation times will be required to ensure the sampling system has been
sufficiently conditioned prior to analysis.
The laboratory has chosen to use the NDIR instrument rather than the GC method after considering the
following performance aspects:
a) Sensitivity and stability: Both instruments were shown to have comparable sensitivities and stabilities.
b) Instrument response function: The linearity of both instruments was assessed and it was concluded
that the GC response function was more linear than that of the NDIR.
c) Repeatability: Repeat analyses of the sample showed the standard deviation of the NDIR to be smaller
than of the GC by a factor of 6. The standard deviation of the GC represented a standard uncertainty of
3 %, which alone is unacceptable considering the pre-set expanded-uncertainty requirement of 5 %.
d) Selectivity: The selectivity of the NDIR was confirmed through comparative analysis with the GC. This
comparison was necessary as the NDIR is sensitive to other infrared active species which may be
present in the sample.
Although the instrument linearity for the GC was better than that of the NDIR, the instrument repeatability was
the overriding factor in the selection of the analytical instrument. The calibration protocol was established for
the NDIR instrument and consisted of a multi-point calibration over the range of interest. The analysis function
[16]
and the uncertainty of the analysis function was determined using ISO 6143 software . The ISO 6143
[5]
software incorporates both the reference gas uncertainty as outlined in ISO 6142 , and the instrument
performance uncertainty.
In this case, the standard uncertainty arising from the method validation amounts to 2 %. In principle, this
method would be “fit for purpose” since the basic requirement is 5 % in terms of expanded uncertainty
A.2 Uncertainty evaluation for a dynamic blending system
A.2.1 General
[12]
The aim of such a system is to prepare gas mixtures using a dynamic gravimetric method (ISO 6145-7 ). In
this particular example, the gas mass flow is important rather than the gas mass. The example describes the
preparation of a four-component mixture consisting of:
−2
0,2 × 10 mol/mol propane;
−2
3,5 × 10 mol/mol carbon monoxide;
−2
14,0 × 10 mol/mol carbon dioxide;
the balance being nitrogen.
The system used
...