ISO/TS 13530:2009
(Main)Water quality — Guidance on analytical quality control for chemical and physicochemical water analysis
Water quality — Guidance on analytical quality control for chemical and physicochemical water analysis
ISO/TS 13530:2009 provides comprehensive guidance on within‑laboratory and between‑laboratory quality control for ensuring the production of results with a known level of accuracy in the analysis of waters. ISO/TS 13530:2009 is applicable to the chemical and physicochemical analysis of all types of waters. It is not intended for application to the analysis of sludges and sediments (although many of its general principles are applicable to such analysis) and it does not address the biological or microbiological examination of water. Whilst sampling is an important aspect, this is only briefly considered.
Qualité de l'eau — Lignes directrices pour le contrôle de qualité analytique pour l'analyse chimique et physicochimique de l'eau
L'ISO/TS 13530:2009 fournit des lignes directrices détaillées relatives au contrôle qualité intralaboratoire et interlaboratoires afin de s'assurer que les résultats de l'analyse des eaux sont obtenus avec un niveau d'exactitude connue. L'ISO/TS 13530:2009 est applicable à l'analyse chimique et physicochimique de tous les types d'eaux. Elle ne s'applique pas à l'analyse des boues et sédiments (même si la plupart de ses principes généraux s'appliquent à de telles analyses) et ne concerne pas l'examen biologique ou microbiologique de l'eau. Bien que l'échantillonnage soit un aspect important, celui-ci n'est que brièvement abordé.
Kakovost vode - Navodilo za kontrolo kakovosti za kemijske in fizikalno-kemijske analize vode
To tehnično poročilo podaja celostne smernice za nadzor laboratorijske in medlaboratorijske kakovosti za pridobivanje rezultatov analiz vode z znano ravnjo natančnosti. To tehnično poročilo velja za kemijske in fizikalno-kemijske analize vseh vrst vode. Ni namenjeno uporabi za analizo blat in sedimentov (čeprav številna splošna načela iz tega poročila veljajo za takšno analizo) ter ne obravnava biološkega ali mikrobiološkega preučevanja vode. Čeprav je vzorčenje pomemben vidik, je upoštevan le bežno. Kontrola kakovosti analize, kot je opisana v tem tehničnem poročilu, je namenjena uporabi analize vode, ki se izvaja v okviru programa zagotavljanja kakovosti. To tehnično poročilo ne obravnava podrobnih zahtev za zagotavljanje kakovosti analize vode, ki je opisano v smernicah EURACHEM/CITAC (2002) [20]. Priporočila iz tega tehničnega poročila so skladna z zahtevami uveljavljene dokumentacije glede zagotavljanja kakovosti (npr. ISO/IEC 17025). To tehnično poročilo velja za uporabo vseh analitičnih metod na področju uporabe, vendar je za podrobna priporočila v njem potrebna interpretacija in prilagoditev, ki omogoča obravnavanje določenih vrst analitov (npr. nespecifičnih analitov, kot so neraztopljene trdne snovi ali biokemijska potreba po kisiku, BPK). V primeru neskladnosti med priporočili iz tega tehničnega poročila in zahtevami standardnih metod analize prevladajo zahteve metode.
General Information
Relations
Standards Content (Sample)
SLOVENSKI STANDARD
01-september-2013
1DGRPHãþD
SIST ENV ISO 13530:2000
Kakovost vode - Navodilo za kontrolo kakovosti za kemijske in fizikalno-kemijske
analize vode
Water quality - Guidance on analytical quality control for chemical and physicochemical
water analysis
Qualité de l'eau - Lignes directrices pour le contrôle de qualité analytique pour l'analyse
chimique et physicochimique de l'eau
Ta slovenski standard je istoveten z: ISO/TS 13530:2009
ICS:
13.060.45 Preiskava vode na splošno Examination of water in
general
2003-01.Slovenski inštitut za standardizacijo. Razmnoževanje celote ali delov tega standarda ni dovoljeno.
TECHNICAL ISO/TS
SPECIFICATION 13530
First edition
2009-03-15
Water quality — Guidance on analytical
quality control for chemical and
physicochemical water analysis
Qualité de l'eau — Lignes directrices pour le contrôle de qualité
analytique pour l'analyse chimique et physicochimique de l'eau
Reference number
©
ISO 2009
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 2009
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 2009 – All rights reserved
Contents Page
Foreword. iv
1 Scope . 1
2 Normative references . 1
3 Terms and definitions. 2
3.1 Terms related to measurement methods . 2
3.2 Terms related to measurement results. 3
3.3 Terms related to uncertainty . 5
4 Performance characteristics of analytical systems . 5
4.1 Introduction . 5
4.2 Scope of the method . 6
4.3 Calibration . 6
4.4 Limit of detection, limit of quantification . 10
4.5 Interferences and matrix effects . 12
4.6 Accuracy (trueness and precision) and uncertainty of measurement. 14
4.7 Robustness . 14
4.8 Fitness for purpose . 15
5 Choosing analytical systems . 15
5.1 General considerations. 15
5.2 Practical considerations . 16
6 Intralaboratory quality control. 16
6.1 General. 16
6.2 Terms relating to within-laboratory quality control . 17
6.3 Control of accuracy . 17
6.4 Control of trueness. 18
6.5 Control of precision. 19
6.6 Principles of applying control charts . 21
6.7 Conclusions . 25
6.8 Control charts with fixed quality criterions (target control charts). 27
7 Quality control in sampling . 27
8 Interlaboratory quality control. 28
9 Quality control for lengthy analytical procedures or analysis undertaken infrequently or
at an ad hoc basis. 28
9.1 Quality control for lengthy analytical procedures. 28
9.2 Analysis undertaken infrequently or on an ad hoc basis. 29
Annex A (informative) Verification of the limit of detection and the limit of quantification . 30
Annex B (informative) The nature and sources of analytical errors . 32
Annex C (informative) Estimating the measurement uncertainty . 35
Annex D (informative) Example for performing quality control for lengthy analytical procedures. 37
Bibliography . 38
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 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 13530 was prepared by Technical Committee ISO/TC 147, Water quality, Subcommittee SC 2,
Physical, chemical and biochemical methods.
This first edition of ISO/TS 13530 cancels and replaces ISO/TR 13530:1997, which has been technically
revised.
iv © ISO 2009 – All rights reserved
TECHNICAL SPECIFICATION ISO/TS 13530:2009(E)
Water quality — Guidance on analytical quality control for
chemical and physicochemical water analysis
1 Scope
This Technical Specification provides comprehensive guidance on within-laboratory and between-laboratory
quality control for ensuring the production of results with a known level of accuracy in the analysis of waters.
This Technical Specification is applicable to the chemical and physicochemical analysis of all types of waters.
It is not intended for application to the analysis of sludges and sediments (although many of its general
principles are applicable to such analysis) and it does not address the biological or microbiological
examination of water. Whilst sampling is an important aspect, this is only briefly considered.
Analytical quality control, as described in this Technical Specification, is intended for application to water
analysis carried out within a quality-assurance programme. This Technical Specification does not address the
detailed requirements of quality assurance for water analysis, which can be found in the EURACHEM/CITAC
[20]
Guide (2002) .
The recommendations of this Technical Specification are in agreement with the requirements of established
quality-assurance documentation (e.g. ISO/IEC 17025).
This Technical Specification is applicable to the use of all analytical methods within its field of application,
although its detailed recommendations may require interpretation and adaptation to deal with certain types of
determinands (for example, non-specific determinands, such as suspended solids or biochemical oxygen
demand, BOD). In the event of any disparity between the recommendations of this Technical Specification
and the requirements of a standard method of analysis, the requirements of the method should prevail.
2 Normative references
The following referenced documents are indispensable for the application of this document. For dated
references, only the edition cited applies. For undated references, the latest edition of the referenced
document (including any amendments) applies.
ISO 3534-2:2006, Statistics — Vocabulary and symbols — Part 2: Applied statistics
ISO 5725 (all parts), Accuracy (trueness and precision) of measurement methods and results
ISO 8466-1, Water quality — Calibration and evaluation of analytical methods and estimation of performance
characteristics — Part 1: Statistical evaluation of the linear calibration function
ISO 8466-2, Water quality — Calibration and evaluation of analytical methods and estimation of performance
characteristics — Part 2: Calibration strategy for non-linear second-order calibration functions
ISO 13528:2005, Statistical methods for use in proficiency testing by interlaboratory comparisons
ISO/IEC 17025:2005, General requirements for the competence of testing and calibration laboratories
ISO Guide 35, Reference materials — General and statistical principles for certification
ISO/IEC Guide 43-1, Proficiency testing by interlaboratory comparisons — Part 1: Development and operation
of proficiency testing schemes
ISO/IEC Guide 43-2, Proficiency testing by interlaboratory comparisons — Part 2: Selection and use of
proficiency testing schemes by laboratory accreditation bodies
3 Terms and definitions
For the purposes of this document, the following terms and definitions apply.
3.1 Terms related to measurement methods
3.1.1
validation
confirmation by examination and the provision of objective evidence that the particular requirements for the
specific intended use are fulfilled
[ISO/IEC 17025:2005]
3.1.2
accuracy
closeness of agreement between a test result or measurement result and the true value
NOTE 1 In practice, the accepted reference value (3.2.6) is substituted for the true value.
NOTE 2 The term “accuracy”, when applied to a set of test or measurement results, involves a combination of random
components and a common systematic error or bias component.
NOTE 3 Accuracy refers to a combination of trueness and precision.
[ISO 3534-2:2006]
3.1.3
bias
difference between the expectation of a test result or measurement result and a true value
[ISO 3534-2:2006]
3.1.4
trueness
closeness of agreement between the expectation of a test result or a measurement result and a true value
NOTE 1 The measure of trueness is usually expressed in terms of bias.
NOTE 2 Trueness is sometimes referred to as “accuracy of the mean”. This usage is not recommended.
NOTE 3 In practice, the accepted reference value is substituted for the true value.
[ISO 3534-2:2006]
3.1.5
precision
closeness of agreement between independent test/measurement results obtained under stipulated conditions
NOTE 1 Precision depends only on the distribution of random errors and does not relate to the true value or the
specified value.
NOTE 2 The measure of precision is usually expressed in terms of imprecision and computed as a standard deviation
of the test results or measurement results. Less precision is reflected by a larger standard deviation.
2 © ISO 2009 – All rights reserved
NOTE 3 Quantitative measures of precision depend critically on the stipulated conditions. Repeatability conditions and
reproducibility conditions are particular sets of extreme stipulated conditions.
[ISO 3534-2:2006]
3.1.6
limit of detection
output signal or value above which it can be affirmed with a stated level of confidence, for example 95 %, that
a sample is different from a blank sample containing no determinand of interest
[ISO 6107-2:2006]
3.1.7
limit of quantification
stated multiple of the limit of detection, for example, two or three times the limit of detection, at a concentration
of the determinand that can reasonably be determined with an acceptable level of accuracy and precision
NOTE Limit of quantification can be calculated using an appropriate standard or sample, and may be obtained from
the lowest calibration point on the calibration curve (excluding the blank).
[ISO 6107-2:2006]
3.1.8
analytical run
group of measurements or observations carried out together, either simultaneously or sequentially, without
interruption, on the same instrument by the same analyst using the same reagents
NOTE 1 An analytical run may consist of more than one batch of analyses. During an analytical run, the accuracy and
precision of the measuring system is expected to be stable.
NOTE 2 Definition taken from Reference [33] in the Bibliography.
3.1.9
batch of analyses
group of measurements or observations of standards, samples and/or control solutions which have been
performed together in respect of all procedures, either simultaneously or sequentially, by the same analysts
using the same reagents, equipment and calibration
3.2 Terms related to measurement results
3.2.1
error of measurement
test result or measurement result minus the true value
NOTE 1 In practice, the accepted reference value is substituted for the true value.
NOTE 2 Error is the sum of random errors and systematic errors.
NOTE 3 Adapted from ISO 3534-2:2006.
3.2.2
systematic error of result
component of the error of result which, in the course of a number of test results or measurement results, for
the same characteristic or quantity, remains constant or varies in a predictable manner
NOTE Systematic errors and their causes can be known or unknown.
[ISO 3534-2:2006]
3.2.3
random error of result
component of the error of result which, in the course of a number of test results or measurement results, for
the same characteristic or quantity, varies in an unpredictable manner
NOTE It is not possible to correct for random error.
[ISO 3534-2:2006]
3.2.4
true value
value which characterizes a quantity or quantitative characteristic perfectly defined in the conditions which
exist when that quantity or quantitative characteristic is considered
NOTE The true value of a quantity or quantitative characteristic is a theoretical concept and, in general, cannot be
known exactly.
[ISO 3534-2:2006]
3.2.5
conventional true value
value of a quantity or quantitative characteristic which, for a given purpose, may be substituted for a true value
NOTE A conventional true value is, in general, regarded as being sufficiently close to the true value for the difference
to be insignificant for the given purpose.
[ISO 3534-2:2006]
3.2.6
accepted reference value
value that serves as an agreed-upon reference for comparison
NOTE The accepted reference value is derived as:
a) a theoretical or established value, based on scientific principles;
b) an assigned or certified value, based on experimental work of some national or international organization;
c) a consensus or certified value, based on collaborative experimental work under the auspices of a scientific or
technical group;
d) the expectation, i.e. the mean of a specified set of measurements, when a), b) and c) are not available.
[ISO 3534-2:2006]
3.2.7
certified reference material
reference material, accompanied by a certificate, one or more of whose property values are certified by a
procedure which establishes traceability to an accurate realization of the unit in which the property values are
expressed, and for which each certified value is accompanied by an uncertainty at a stated level of confidence
NOTE Definition taken from Reference [36] in the Bibliography.
3.2.8
metrological traceability
property of a measurement result relating the result to a stated metrological reference through an unbroken
chain of calibrations of a measuring system or comparisons, each contributing to the stated measurement
uncertainty
[ISO/IEC Guide 99:2007]
4 © ISO 2009 – All rights reserved
3.3 Terms related to uncertainty
3.3.1
uncertainty of measurement
non-negative parameter characterizing the dispersion of the quantity values being attributed to a measurand,
based on the information used
[ISO/IEC Guide 99:2007]
3.3.2
standard uncertainty
u(x )
i
uncertainty of the result x of a measurement expressed as a standard deviation
i
[ISO/IEC Guide 98-3:2008]
3.3.3
combined standard uncertainty
u (y)
c
standard measurement uncertainty that is obtained using the individual standard measurement uncertainties
associated with the input quantities in a measurement model
[ISO/IEC Guide 99:2007]
3.3.4
expanded uncertainty
U
product of a combined standard measurement uncertainty and a factor larger than the number one
NOTE 1 The factor depends upon the type of probability distribution of the output quantity in a measurement model
and on the selected coverage probability.
NOTE 2 The term “factor” in this definition refers to a coverage factor.
[ISO/IEC Guide 99:2007]
3.3.5
coverage factor
k
numerical factor used as a multiplier of the combined standard uncertainty in order to obtain an expanded
uncertainty
NOTE A coverage factor is typically in the range from 2 to 3.
[ISO/IEC Guide 98-3:2008]
4 Performance characteristics of analytical systems
4.1 Introduction
ISO/IEC 17025 requires the validation of methods; the validation process is described in detail in EURACHEM
[18]
Guide (1998) . Primary validation is part of the development of a new analytical method, and is performed
during the standardization of the method. Important points to be considered for the primary validation of an
analytical method are the following:
⎯ scope of the method;
⎯ calibration;
⎯ limit of detection/limit of quantification;
⎯ interferences;
⎯ estimation of accuracy (trueness and precision);
⎯ uncertainty of measurement;
⎯ robustness;
⎯ fitness for purpose.
According to ISO/IEC 17025, the laboratory shall confirm that it can correctly operate standard methods
before applying these methods. This procedure is called secondary validation, where emphasis is laid on
calibration and interferences, as well as the laboratory limit of quantitation and measurement uncertainty. In
4.2 to 4.8, the topics calibration and quantification, matrix effects and measurement uncertainty are dealt with
especially.
4.2 Scope of the method
A clear definition should be given of the forms of the substance that are determined by the procedure and also,
when necessary to avoid ambiguity, those forms that are not capable of determination. At this point, it is worth
emphasizing that the analyst's selection of an analytical method should meet the user's definition of the
determinand. Non-specific determinands need the use of rigorously stipulated analytical methods in order to
obtain reliable and comparable results.
Many substances exist in water in a variety of forms or “species”, and many analytical systems provide a
differential response to the various forms. For example, when a separation of “dissolved” and “particulate”
material is required, special care is necessary to define precisely the nature and pore-size of the filter to be
used.
A precise description of the types and natures of samples is important before the analytical system can be
chosen. The precautions to be taken when a sample is analysed will depend to a high degree on the sample.
The analyst needs information that is as complete as possible on sample types, concentration levels and
possible interferences. The scope should contain a clear statement of the types of sample and sample
matrices for which the procedure is suitable. If necessary, a statement should also be made of important
sample types and matrices for which the procedure is not suitable.
The range of application corresponds to the lowest and highest concentrations for which tests of precision and
bias have been carried out using the system without modification. Where an extension can be used to enable
the examination of samples containing concentrations greater than the upper limit, such as by analysis after
dilution, then it should be regarded as a different procedure but whose performance characteristics may be
inferred from the values quoted for the original.
The concentration range of interest can have a marked effect on the choice of analytical technique; of primary
concern is the smallest concentration of interest.
4.3 Calibration
4.3.1 Basics of calibration and quantification
A calibration function is determined from information values y , obtained by measuring given standard
i
concentrations x . The resulting standard deviation of the method s or confidence interval is a performance
i xo
characteristic of this calibration, not a performance characteristic of the quantification. Quantification of the
analyte content of samples using a calibration function is based on interpolation. The most important
performance characteristic for quantification is the estimation of measurement uncertainty (see 4.6) where the
uncertainty of calibration is one of the contributions.
6 © ISO 2009 – All rights reserved
Depending on the kind of the functional relationship between the analyte concentration and the measurement
response, different mathematical or statistical tools can be used. Mathematical and statistical models are
subject to different assumptions; therefore, not only is the performance of the analytical equipment essential
but also the kind of mathematical approach. The models should help the analyst to find a clear and reliable
functional relationship for calibration. The models should not limit the capabilities of the analytical equipment.
Therefore, the selection of the calibration model should be undertaken carefully, taking into consideration the
fitness for purpose and the measurement uncertainty required.
⎯ Model of linear regression (ISO 8466-1): the classical model for calibration with the limitations of normal
distribution of the responses and a homogeneity of variances over the working range. Normally, this
homogeneity of variances is only given in a working range of one or at maximum two decades of
concentration. This fact limits the applicability.
⎯ Model of non-linear second-order calibration functions (ISO 8466-2): after discovering a significant non-
linearity. This model has the same limitations as linear regression: normal distribution of the responses
and homogeneity of variances over the working range.
⎯ Model of weighted regression, weighting the information value with the reciprocal variance, calibration
over more than a decade of concentration range is possible.
⎯ Calibration over more than one decade of concentration with at least two concentrations, e.g. for
inductively coupled plasma/mass spectrometry (ICP/MS). Linearity must have been checked beforehand
with a minimum of five concentrations, e.g. by graphical presentation.
NOTE Detailed instructions on calibration procedures are given in References [26] and [28] in the Bibliography.
All these kinds of calibration should be tested for their contribution to measurement uncertainty. For example,
analyse N times, as a minimum in triplicate, an independent standard or a certified reference material within
the concentration range of interest, and calculate the results in accordance with the applied calibration method.
The standard uncertainty component, s , representing the deviation of the single results, ρ , from the reference
p i
value, ρ, is calculated as follows:
N
ρρ−
()
∑ i
i=1
s = (1)
p
N−1
where
s is the standard uncertainty component from calibration;
p
N is the number of replicate measurements.
N should be chosen with respect to the confidence required. s should be compared with the respective target
p
measurement uncertainty to set the tolerable amount.
4.3.2 Calibration strategies
4.3.2.1 Basic or instrument calibration
This type of calibration is carried out without matrix components and without sample preparation; for
calibration, pure standard solutions are used. This is inexpensive and directly suited for quantification, if matrix
components do not change the slope and the intercept of the calibration function significantly. Normally this
calibration is more precise than a calibration including all sample preparation steps.
4.3.2.2 Calibration with matrix material
This kind of calibration is comparable with basic calibration (4.3.2.1). No sample preparation steps are carried
out. Instead of the pure solutions of the standards, solutions of the standard substances in analyte-free matrix
or artificial matrix are used.
4.3.2.3 Calibration over the total procedure
This type of calibration includes all sample preparation steps, and the calibration solutions are prepared with a
representative matrix material. It is suitable to show whether matrix components or sample preparation steps
do or do not change the slope and intercept of the calibration function compared to the basic calibration
function. Frequently, it is not easy to find a representative matrix material to use it as a basis of calibration
solutions. In water analysis, this means that blank free natural water representing the possible matrices of
other natural waters has to be found.
If the matrix components significantly change intercept and slope of the calibration function compared to the
basic calibration, it is possible to use this calibration over the total procedure for quantification.
4.3.3 Internal standardization
4.3.3.1 General aspects
The use of internal standardization for the quantification of concentrations minimizes possible errors made
both during injection and by sample losses during sample pre-treatment steps, and also differences in the final
sample extract volumes and changes in recoveries caused by matrix effects. Substances used as internal
standards should have the following properties.
⎯ Chemical-physical properties should be the same concerning the error-prone procedure steps which
should be corrected. If the total procedure should be controlled by the internal standardization, isotopic
labelled compounds are recommended; if only final volumes or detection should be controlled, other
substances, which are representative concerning these steps, can be used.
⎯ There should not be any measurement interferences with the internal standards.
⎯ No occurrence of the internal standard, neither in real samples nor as blanks, which cannot be avoided.
⎯ Concentration of the added internal standard: in the dynamic range of the method, preferably in the same
concentration range as the analytes.
⎯ They should have similar intensities of responses as the analytes.
It is recommended to carry out the internal calibration as a basic calibration, because all multiplicative matrix
effects are corrected by the internal standard if it has the properties listed above. Otherwise, an internal
calibration over the total procedure has the same disadvantages as described in 4.3.2.3, often there is a lack
of representative matrix material. As with all the other possibilities for correcting matrix effects, internal
standardization cannot overcome additive matrix effects as well.
4.3.3.2 Calibration with internal standards
Calculation of the calibration function is usually available as an option in the quantification programs of most
manufacturers’ data analysis software.
Adjust the concentrations according to the sensitivity of the equipment used and the range of determinations
required. Evaluate the linear range and, subsequently, set up a calibration. Establish the linear function as a
basic calibration of the pairs of values yy and ρ ρ of the measured series using Equation (2):
iis,i iis,i
8 © ISO 2009 – All rights reserved
y ρ
ii
=+ab (2)
ii
y ρ
is,i is,i
where
y is the measured response of substance i; the unit depends on the evaluation, e.g. area value;
i
ρ is the mass concentration of substance i (external standard) in the working standard solution, e.g. in
i
nanograms per millilitre, ng/ml;
a is the slope of the calibration function of substance i; the unit depends on the evaluation, e.g. area
i
value millilitres per nanogram, area value·ml/ng.
b is the ordinate intercept of the calibration curve; the unit depends on the evaluation, e.g. area value;
i
y is the measured response of the internal standard for the substance i; the unit depends on the
is,i
evaluation, e.g. area value;
ρ is the mass concentration of the internal standard, for the substance i, e.g. in nanograms per
is,i
millilitre, ng/ml.
Calibration can be done as linear regression or as a two point calibration over more than one decade after the
previous linearity check.
4.3.3.3 Quantification with internal standards
Add a known amount of the internal standards to the sample prior to sample preparation. Adjust this amount
of the internal standards in such a manner that the mass concentration ρ in the final volume, e.g. of the
is,i
extract, is nearly the same in the prepared samples as in the calibration solutions. Use the same solvent
composition for the standard solutions and the samples.
Calculate the mass concentration ρ of the substance using Equation (3).
i,sample
y
i,sample
− b
i
y ρ
mm
is,i,sample is,i i,sample extract is,i
ρ=×= × (3)
i,sample
aV ρ V
i sample is,i,sample extract sample
where
y is the measured response, e.g. peak area, of substance i in the sample extract;
i,sample
y is the measured response, e.g. peak area, of the internal standard, for substance i, of the
is,i,sample
sample;
ρ is the mass concentration of substance i in the sample extract, e.g. in nanograms per
i,sample extract
millilitre, ng/ml; usually calculated by the software;
ρ is the mass concentration of the internal standard in the sample extract, for substance i,
is,i,sample extract
e.g. in nanograms per millilitre, ng/ml; usually reported by the software;
ρ is the mass concentration of substance i in the water sample, e.g. in micrograms per litre,
i,sample
µg/l;
m is the mass of the added internal standard substance, e.g. in micrograms, µg;
is,i
V is the sample volume, in litres, l;
sample
a is defined in Equation (2);
i
b is defined in Equation (2).
i
4.3.3.4 Determination of recoveries of the internal standards
It is necessary to control the recovery of the internal standards for the total procedure for each sample.
This is possible by comparing the mean response from the internal standard in the calibration solutions with
the response obtained from the prepared sample. To achieve this, it is essential, that the final volume of the
prepared sample is known for calculating the theoretical final concentration of the internal standard in the
prepared sample. In routine use, the theoretical concentration of the internal standards in the prepared
samples is the same as in the calibration standard solutions, therefore no additional effort results.
Alternatively, a second internal standard, e.g. an injection standard can be used for the calculation of
recoveries of internal standards. This is a useful procedure if the final volumes after the sample preparation
vary, as frequently happens after enrichment to very small final volumes. This second internal standard is
added to the calibration solution and to the final prepared sample prior to the sample measurement. The final
concentration shall be the same for the calibration solution and the prepared sample with a theoretical final
volume. Now it is possible to obtain recoveries directly by comparing the responses of the internal and the
second internal (injection) standards obtained in the calibration with those obtained from prepared samples,
e.g. extracts.
yy×
is,i,sample is,inj,calibration
A=×100 (4)
is,i,sample
yy×
is,i,calibration is,inj,sample
where
A is the recovery of the internal standard, for substance i, in percent, %;
is,i,sample
y is the measured response, e.g. peak area, of the internal standard, for substance i, in the
is,i,sample
sample;
y is the measured response, e.g. peak area, of the internal standard, for substance i, in the
is,i,calibration
calibration solution;
y is the measured response, e.g. peak area, of the injection internal standard, in the sample;
is,inj,sample
y is the measured response, e.g. peak area, of the injection internal standard, in the
is,inj,calibration
calibration solution.
A control chart plot of the recoveries of the internal standards or another kind of documentation should
indicate that the recovery rates lie within defined control limits.
4.4 Limit of detection, limit of quantification
4.4.1 General aspects
In broad terms, the limit of detection is the smallest amount or concentration of an analyte in the test sample
that can be reliably distinguished from zero or blank (Reference [31] in the Bibliography). For some physical
parameters, e.g. pH and redox potential, the concept of limit of detection does not apply and no attempt
should be made to determine it for these parameters.
There is much diversity in the way in which the limit of detection and limit of quantification of an analytical
system is estimated. Most approaches are based on multiplication of the within-batch standard deviation of
results of typical matrix blanks or low-level material or the multiplication of the standard deviation of the
method, s , by a factor. These statistical inferences depend on the assumption of normality, which is at least
xo
10 © ISO 2009 – All rights reserved
questionable at low concentrations. Notwithstanding this, in method validation, a simple definition, leading to a
quickly implemented estimation of the detection limit, can be applied.
The different ways of estimating the limit of detection and limit of quantification which are described in the
following subclauses are optional. With the recommended minimum degrees of freedom, the value of the limit
of detection is quite uncertain, and may easily be in error by a factor of 2. Where more accurate estimates are
required, more complex calculations should be applied. For special cases, see ISO 11843, Parts 1 to 4.
The essential step after the estimation of the limit of detection and limit of quantification is the verification. The
analyst needs to prove that he is able to detect and, respectively, quantify the analyte at the estimated limits in
the respective matrix. If the criteria given in 4.4.6 are fulfilled, the estimated limits, e.g. the limit of detection
according to 4.4.2, are verified. After verification, the limit of detection and limit of quantification of different
laboratories can be compared, e.g. with quality targets.
4.4.2 Limit of detection based on standard deviation of results of blank samples
The limit of detection can be estimated as:
xs=+3 x (5)
LD 0 Bl
where
x is the limit of detection;
LD
s is the standard deviation of the outlier-free results of a matrix blank sample;
x is the mean concentration of the matrix blank.
Bl
If the applied method of calculation of results comprises the subtraction of a matrix blank, the second term, x ,
Bl
has to be ignored.
The precision estimate, s , should be based on at least 10 independent complete determinations of analyte
concentration in a typical matrix blank or low-level material, with no censoring of zero or negative results. For
that number of determinations, the factor of 3 corresponds to a significance level of α = 0,01.
4.4.3 Limit of detection based on the standard deviation of the method
For analytical methods which have a linear calibration function determined alternatively to the way described
in 4.4.2, the limit of detection can be estimated as:
x = 4s (6)
LD xo
where
x is the limit of detection;
LD
s is the standard deviation of the method (from calibration).
xo
4.4.4 Limit of detection based on baseline noise
For methods which show a baseline noise, the limit of detection, x , can be estimated as the concentration of
LD
the analyte at a signal/noise ratio S/N = 3 after blank correction (if possible).
4.4.5 Limit of quantification
The limit of quantification represents a concentration of the determinand that can reasonably be determined
with an acceptable level of accuracy. Usually it is arbitrarily taken as a fixed multiple of the detection limit.
For method validation, the limit of quantification, x , can be estimated as:
LQ
xx= 3 (7)
LQ LD
The factor k = 3 corresponds to a relative result uncertainty of approximately 33 %.
4.4.6 Verification of the limit of detection and the limit of quantification in the matrix
Verification of the limit of quantification is vitally important if routine samples frequently show analyte
concentrations near the limit of quantification. In that case, investigations for verification of the limit of
quantification should be performed regularly in routine analysis.
For verification of the limit of detection and limit of quantification, spiked blank matrix samples at these
concentration levels and blank matrix samples shall be analysed in the same manner as real samples, i.e.
under within-laboratory reproducibility conditions.
If the mean response of the samples spiked at the limit of detection level is greater than the maximum blank
value, the limit of detection is verified.
If the uncertainty of results for the samples spiked at the limit of quantification level is smaller than or equal to
the relative precision corresponding to the factor k, the limit of quantification is verified.
NOTE At the limit of quantification, the uncertainty component bias can be neglected, as the blank is included in the
estimation of the limit of detection and limit of quantification.
Examples for the verification of the limit of detection and the limit of quantification are given in Annex A.
4.4.7 Reporting limit
The reporting limit is a specific concentration at or above the limit of quantification that is reported to the client
with a certain degree of confidence. It is often defined on a project-specific basis. If the reporting limit is set
below the limit of quantification by the client, method modification is required.
4.5 Interferences and matrix effects
4.5.1 General considerations on matrix effects
An important source of random and systematic error in results is the presence of constituents of a sample
other than the determinand that cause an enhancement or
...
TECHNICAL ISO/TS
SPECIFICATION 13530
First edition
2009-03-15
Water quality — Guidance on analytical
quality control for chemical and
physicochemical water analysis
Qualité de l'eau — Lignes directrices pour le contrôle de qualité
analytique pour l'analyse chimique et physicochimique de l'eau
Reference number
©
ISO 2009
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 2009
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 2009 – All rights reserved
Contents Page
Foreword. iv
1 Scope . 1
2 Normative references . 1
3 Terms and definitions. 2
3.1 Terms related to measurement methods . 2
3.2 Terms related to measurement results. 3
3.3 Terms related to uncertainty . 5
4 Performance characteristics of analytical systems . 5
4.1 Introduction . 5
4.2 Scope of the method . 6
4.3 Calibration . 6
4.4 Limit of detection, limit of quantification . 10
4.5 Interferences and matrix effects . 12
4.6 Accuracy (trueness and precision) and uncertainty of measurement. 14
4.7 Robustness . 14
4.8 Fitness for purpose . 15
5 Choosing analytical systems . 15
5.1 General considerations. 15
5.2 Practical considerations . 16
6 Intralaboratory quality control. 16
6.1 General. 16
6.2 Terms relating to within-laboratory quality control . 17
6.3 Control of accuracy . 17
6.4 Control of trueness. 18
6.5 Control of precision. 19
6.6 Principles of applying control charts . 21
6.7 Conclusions . 25
6.8 Control charts with fixed quality criterions (target control charts). 27
7 Quality control in sampling . 27
8 Interlaboratory quality control. 28
9 Quality control for lengthy analytical procedures or analysis undertaken infrequently or
at an ad hoc basis. 28
9.1 Quality control for lengthy analytical procedures. 28
9.2 Analysis undertaken infrequently or on an ad hoc basis. 29
Annex A (informative) Verification of the limit of detection and the limit of quantification . 30
Annex B (informative) The nature and sources of analytical errors . 32
Annex C (informative) Estimating the measurement uncertainty . 35
Annex D (informative) Example for performing quality control for lengthy analytical procedures. 37
Bibliography . 38
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 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 13530 was prepared by Technical Committee ISO/TC 147, Water quality, Subcommittee SC 2,
Physical, chemical and biochemical methods.
This first edition of ISO/TS 13530 cancels and replaces ISO/TR 13530:1997, which has been technically
revised.
iv © ISO 2009 – All rights reserved
TECHNICAL SPECIFICATION ISO/TS 13530:2009(E)
Water quality — Guidance on analytical quality control for
chemical and physicochemical water analysis
1 Scope
This Technical Specification provides comprehensive guidance on within-laboratory and between-laboratory
quality control for ensuring the production of results with a known level of accuracy in the analysis of waters.
This Technical Specification is applicable to the chemical and physicochemical analysis of all types of waters.
It is not intended for application to the analysis of sludges and sediments (although many of its general
principles are applicable to such analysis) and it does not address the biological or microbiological
examination of water. Whilst sampling is an important aspect, this is only briefly considered.
Analytical quality control, as described in this Technical Specification, is intended for application to water
analysis carried out within a quality-assurance programme. This Technical Specification does not address the
detailed requirements of quality assurance for water analysis, which can be found in the EURACHEM/CITAC
[20]
Guide (2002) .
The recommendations of this Technical Specification are in agreement with the requirements of established
quality-assurance documentation (e.g. ISO/IEC 17025).
This Technical Specification is applicable to the use of all analytical methods within its field of application,
although its detailed recommendations may require interpretation and adaptation to deal with certain types of
determinands (for example, non-specific determinands, such as suspended solids or biochemical oxygen
demand, BOD). In the event of any disparity between the recommendations of this Technical Specification
and the requirements of a standard method of analysis, the requirements of the method should prevail.
2 Normative references
The following referenced documents are indispensable for the application of this document. For dated
references, only the edition cited applies. For undated references, the latest edition of the referenced
document (including any amendments) applies.
ISO 3534-2:2006, Statistics — Vocabulary and symbols — Part 2: Applied statistics
ISO 5725 (all parts), Accuracy (trueness and precision) of measurement methods and results
ISO 8466-1, Water quality — Calibration and evaluation of analytical methods and estimation of performance
characteristics — Part 1: Statistical evaluation of the linear calibration function
ISO 8466-2, Water quality — Calibration and evaluation of analytical methods and estimation of performance
characteristics — Part 2: Calibration strategy for non-linear second-order calibration functions
ISO 13528:2005, Statistical methods for use in proficiency testing by interlaboratory comparisons
ISO/IEC 17025:2005, General requirements for the competence of testing and calibration laboratories
ISO Guide 35, Reference materials — General and statistical principles for certification
ISO/IEC Guide 43-1, Proficiency testing by interlaboratory comparisons — Part 1: Development and operation
of proficiency testing schemes
ISO/IEC Guide 43-2, Proficiency testing by interlaboratory comparisons — Part 2: Selection and use of
proficiency testing schemes by laboratory accreditation bodies
3 Terms and definitions
For the purposes of this document, the following terms and definitions apply.
3.1 Terms related to measurement methods
3.1.1
validation
confirmation by examination and the provision of objective evidence that the particular requirements for the
specific intended use are fulfilled
[ISO/IEC 17025:2005]
3.1.2
accuracy
closeness of agreement between a test result or measurement result and the true value
NOTE 1 In practice, the accepted reference value (3.2.6) is substituted for the true value.
NOTE 2 The term “accuracy”, when applied to a set of test or measurement results, involves a combination of random
components and a common systematic error or bias component.
NOTE 3 Accuracy refers to a combination of trueness and precision.
[ISO 3534-2:2006]
3.1.3
bias
difference between the expectation of a test result or measurement result and a true value
[ISO 3534-2:2006]
3.1.4
trueness
closeness of agreement between the expectation of a test result or a measurement result and a true value
NOTE 1 The measure of trueness is usually expressed in terms of bias.
NOTE 2 Trueness is sometimes referred to as “accuracy of the mean”. This usage is not recommended.
NOTE 3 In practice, the accepted reference value is substituted for the true value.
[ISO 3534-2:2006]
3.1.5
precision
closeness of agreement between independent test/measurement results obtained under stipulated conditions
NOTE 1 Precision depends only on the distribution of random errors and does not relate to the true value or the
specified value.
NOTE 2 The measure of precision is usually expressed in terms of imprecision and computed as a standard deviation
of the test results or measurement results. Less precision is reflected by a larger standard deviation.
2 © ISO 2009 – All rights reserved
NOTE 3 Quantitative measures of precision depend critically on the stipulated conditions. Repeatability conditions and
reproducibility conditions are particular sets of extreme stipulated conditions.
[ISO 3534-2:2006]
3.1.6
limit of detection
output signal or value above which it can be affirmed with a stated level of confidence, for example 95 %, that
a sample is different from a blank sample containing no determinand of interest
[ISO 6107-2:2006]
3.1.7
limit of quantification
stated multiple of the limit of detection, for example, two or three times the limit of detection, at a concentration
of the determinand that can reasonably be determined with an acceptable level of accuracy and precision
NOTE Limit of quantification can be calculated using an appropriate standard or sample, and may be obtained from
the lowest calibration point on the calibration curve (excluding the blank).
[ISO 6107-2:2006]
3.1.8
analytical run
group of measurements or observations carried out together, either simultaneously or sequentially, without
interruption, on the same instrument by the same analyst using the same reagents
NOTE 1 An analytical run may consist of more than one batch of analyses. During an analytical run, the accuracy and
precision of the measuring system is expected to be stable.
NOTE 2 Definition taken from Reference [33] in the Bibliography.
3.1.9
batch of analyses
group of measurements or observations of standards, samples and/or control solutions which have been
performed together in respect of all procedures, either simultaneously or sequentially, by the same analysts
using the same reagents, equipment and calibration
3.2 Terms related to measurement results
3.2.1
error of measurement
test result or measurement result minus the true value
NOTE 1 In practice, the accepted reference value is substituted for the true value.
NOTE 2 Error is the sum of random errors and systematic errors.
NOTE 3 Adapted from ISO 3534-2:2006.
3.2.2
systematic error of result
component of the error of result which, in the course of a number of test results or measurement results, for
the same characteristic or quantity, remains constant or varies in a predictable manner
NOTE Systematic errors and their causes can be known or unknown.
[ISO 3534-2:2006]
3.2.3
random error of result
component of the error of result which, in the course of a number of test results or measurement results, for
the same characteristic or quantity, varies in an unpredictable manner
NOTE It is not possible to correct for random error.
[ISO 3534-2:2006]
3.2.4
true value
value which characterizes a quantity or quantitative characteristic perfectly defined in the conditions which
exist when that quantity or quantitative characteristic is considered
NOTE The true value of a quantity or quantitative characteristic is a theoretical concept and, in general, cannot be
known exactly.
[ISO 3534-2:2006]
3.2.5
conventional true value
value of a quantity or quantitative characteristic which, for a given purpose, may be substituted for a true value
NOTE A conventional true value is, in general, regarded as being sufficiently close to the true value for the difference
to be insignificant for the given purpose.
[ISO 3534-2:2006]
3.2.6
accepted reference value
value that serves as an agreed-upon reference for comparison
NOTE The accepted reference value is derived as:
a) a theoretical or established value, based on scientific principles;
b) an assigned or certified value, based on experimental work of some national or international organization;
c) a consensus or certified value, based on collaborative experimental work under the auspices of a scientific or
technical group;
d) the expectation, i.e. the mean of a specified set of measurements, when a), b) and c) are not available.
[ISO 3534-2:2006]
3.2.7
certified reference material
reference material, accompanied by a certificate, one or more of whose property values are certified by a
procedure which establishes traceability to an accurate realization of the unit in which the property values are
expressed, and for which each certified value is accompanied by an uncertainty at a stated level of confidence
NOTE Definition taken from Reference [36] in the Bibliography.
3.2.8
metrological traceability
property of a measurement result relating the result to a stated metrological reference through an unbroken
chain of calibrations of a measuring system or comparisons, each contributing to the stated measurement
uncertainty
[ISO/IEC Guide 99:2007]
4 © ISO 2009 – All rights reserved
3.3 Terms related to uncertainty
3.3.1
uncertainty of measurement
non-negative parameter characterizing the dispersion of the quantity values being attributed to a measurand,
based on the information used
[ISO/IEC Guide 99:2007]
3.3.2
standard uncertainty
u(x )
i
uncertainty of the result x of a measurement expressed as a standard deviation
i
[ISO/IEC Guide 98-3:2008]
3.3.3
combined standard uncertainty
u (y)
c
standard measurement uncertainty that is obtained using the individual standard measurement uncertainties
associated with the input quantities in a measurement model
[ISO/IEC Guide 99:2007]
3.3.4
expanded uncertainty
U
product of a combined standard measurement uncertainty and a factor larger than the number one
NOTE 1 The factor depends upon the type of probability distribution of the output quantity in a measurement model
and on the selected coverage probability.
NOTE 2 The term “factor” in this definition refers to a coverage factor.
[ISO/IEC Guide 99:2007]
3.3.5
coverage factor
k
numerical factor used as a multiplier of the combined standard uncertainty in order to obtain an expanded
uncertainty
NOTE A coverage factor is typically in the range from 2 to 3.
[ISO/IEC Guide 98-3:2008]
4 Performance characteristics of analytical systems
4.1 Introduction
ISO/IEC 17025 requires the validation of methods; the validation process is described in detail in EURACHEM
[18]
Guide (1998) . Primary validation is part of the development of a new analytical method, and is performed
during the standardization of the method. Important points to be considered for the primary validation of an
analytical method are the following:
⎯ scope of the method;
⎯ calibration;
⎯ limit of detection/limit of quantification;
⎯ interferences;
⎯ estimation of accuracy (trueness and precision);
⎯ uncertainty of measurement;
⎯ robustness;
⎯ fitness for purpose.
According to ISO/IEC 17025, the laboratory shall confirm that it can correctly operate standard methods
before applying these methods. This procedure is called secondary validation, where emphasis is laid on
calibration and interferences, as well as the laboratory limit of quantitation and measurement uncertainty. In
4.2 to 4.8, the topics calibration and quantification, matrix effects and measurement uncertainty are dealt with
especially.
4.2 Scope of the method
A clear definition should be given of the forms of the substance that are determined by the procedure and also,
when necessary to avoid ambiguity, those forms that are not capable of determination. At this point, it is worth
emphasizing that the analyst's selection of an analytical method should meet the user's definition of the
determinand. Non-specific determinands need the use of rigorously stipulated analytical methods in order to
obtain reliable and comparable results.
Many substances exist in water in a variety of forms or “species”, and many analytical systems provide a
differential response to the various forms. For example, when a separation of “dissolved” and “particulate”
material is required, special care is necessary to define precisely the nature and pore-size of the filter to be
used.
A precise description of the types and natures of samples is important before the analytical system can be
chosen. The precautions to be taken when a sample is analysed will depend to a high degree on the sample.
The analyst needs information that is as complete as possible on sample types, concentration levels and
possible interferences. The scope should contain a clear statement of the types of sample and sample
matrices for which the procedure is suitable. If necessary, a statement should also be made of important
sample types and matrices for which the procedure is not suitable.
The range of application corresponds to the lowest and highest concentrations for which tests of precision and
bias have been carried out using the system without modification. Where an extension can be used to enable
the examination of samples containing concentrations greater than the upper limit, such as by analysis after
dilution, then it should be regarded as a different procedure but whose performance characteristics may be
inferred from the values quoted for the original.
The concentration range of interest can have a marked effect on the choice of analytical technique; of primary
concern is the smallest concentration of interest.
4.3 Calibration
4.3.1 Basics of calibration and quantification
A calibration function is determined from information values y , obtained by measuring given standard
i
concentrations x . The resulting standard deviation of the method s or confidence interval is a performance
i xo
characteristic of this calibration, not a performance characteristic of the quantification. Quantification of the
analyte content of samples using a calibration function is based on interpolation. The most important
performance characteristic for quantification is the estimation of measurement uncertainty (see 4.6) where the
uncertainty of calibration is one of the contributions.
6 © ISO 2009 – All rights reserved
Depending on the kind of the functional relationship between the analyte concentration and the measurement
response, different mathematical or statistical tools can be used. Mathematical and statistical models are
subject to different assumptions; therefore, not only is the performance of the analytical equipment essential
but also the kind of mathematical approach. The models should help the analyst to find a clear and reliable
functional relationship for calibration. The models should not limit the capabilities of the analytical equipment.
Therefore, the selection of the calibration model should be undertaken carefully, taking into consideration the
fitness for purpose and the measurement uncertainty required.
⎯ Model of linear regression (ISO 8466-1): the classical model for calibration with the limitations of normal
distribution of the responses and a homogeneity of variances over the working range. Normally, this
homogeneity of variances is only given in a working range of one or at maximum two decades of
concentration. This fact limits the applicability.
⎯ Model of non-linear second-order calibration functions (ISO 8466-2): after discovering a significant non-
linearity. This model has the same limitations as linear regression: normal distribution of the responses
and homogeneity of variances over the working range.
⎯ Model of weighted regression, weighting the information value with the reciprocal variance, calibration
over more than a decade of concentration range is possible.
⎯ Calibration over more than one decade of concentration with at least two concentrations, e.g. for
inductively coupled plasma/mass spectrometry (ICP/MS). Linearity must have been checked beforehand
with a minimum of five concentrations, e.g. by graphical presentation.
NOTE Detailed instructions on calibration procedures are given in References [26] and [28] in the Bibliography.
All these kinds of calibration should be tested for their contribution to measurement uncertainty. For example,
analyse N times, as a minimum in triplicate, an independent standard or a certified reference material within
the concentration range of interest, and calculate the results in accordance with the applied calibration method.
The standard uncertainty component, s , representing the deviation of the single results, ρ , from the reference
p i
value, ρ, is calculated as follows:
N
ρρ−
()
∑ i
i=1
s = (1)
p
N−1
where
s is the standard uncertainty component from calibration;
p
N is the number of replicate measurements.
N should be chosen with respect to the confidence required. s should be compared with the respective target
p
measurement uncertainty to set the tolerable amount.
4.3.2 Calibration strategies
4.3.2.1 Basic or instrument calibration
This type of calibration is carried out without matrix components and without sample preparation; for
calibration, pure standard solutions are used. This is inexpensive and directly suited for quantification, if matrix
components do not change the slope and the intercept of the calibration function significantly. Normally this
calibration is more precise than a calibration including all sample preparation steps.
4.3.2.2 Calibration with matrix material
This kind of calibration is comparable with basic calibration (4.3.2.1). No sample preparation steps are carried
out. Instead of the pure solutions of the standards, solutions of the standard substances in analyte-free matrix
or artificial matrix are used.
4.3.2.3 Calibration over the total procedure
This type of calibration includes all sample preparation steps, and the calibration solutions are prepared with a
representative matrix material. It is suitable to show whether matrix components or sample preparation steps
do or do not change the slope and intercept of the calibration function compared to the basic calibration
function. Frequently, it is not easy to find a representative matrix material to use it as a basis of calibration
solutions. In water analysis, this means that blank free natural water representing the possible matrices of
other natural waters has to be found.
If the matrix components significantly change intercept and slope of the calibration function compared to the
basic calibration, it is possible to use this calibration over the total procedure for quantification.
4.3.3 Internal standardization
4.3.3.1 General aspects
The use of internal standardization for the quantification of concentrations minimizes possible errors made
both during injection and by sample losses during sample pre-treatment steps, and also differences in the final
sample extract volumes and changes in recoveries caused by matrix effects. Substances used as internal
standards should have the following properties.
⎯ Chemical-physical properties should be the same concerning the error-prone procedure steps which
should be corrected. If the total procedure should be controlled by the internal standardization, isotopic
labelled compounds are recommended; if only final volumes or detection should be controlled, other
substances, which are representative concerning these steps, can be used.
⎯ There should not be any measurement interferences with the internal standards.
⎯ No occurrence of the internal standard, neither in real samples nor as blanks, which cannot be avoided.
⎯ Concentration of the added internal standard: in the dynamic range of the method, preferably in the same
concentration range as the analytes.
⎯ They should have similar intensities of responses as the analytes.
It is recommended to carry out the internal calibration as a basic calibration, because all multiplicative matrix
effects are corrected by the internal standard if it has the properties listed above. Otherwise, an internal
calibration over the total procedure has the same disadvantages as described in 4.3.2.3, often there is a lack
of representative matrix material. As with all the other possibilities for correcting matrix effects, internal
standardization cannot overcome additive matrix effects as well.
4.3.3.2 Calibration with internal standards
Calculation of the calibration function is usually available as an option in the quantification programs of most
manufacturers’ data analysis software.
Adjust the concentrations according to the sensitivity of the equipment used and the range of determinations
required. Evaluate the linear range and, subsequently, set up a calibration. Establish the linear function as a
basic calibration of the pairs of values yy and ρ ρ of the measured series using Equation (2):
iis,i iis,i
8 © ISO 2009 – All rights reserved
y ρ
ii
=+ab (2)
ii
y ρ
is,i is,i
where
y is the measured response of substance i; the unit depends on the evaluation, e.g. area value;
i
ρ is the mass concentration of substance i (external standard) in the working standard solution, e.g. in
i
nanograms per millilitre, ng/ml;
a is the slope of the calibration function of substance i; the unit depends on the evaluation, e.g. area
i
value millilitres per nanogram, area value·ml/ng.
b is the ordinate intercept of the calibration curve; the unit depends on the evaluation, e.g. area value;
i
y is the measured response of the internal standard for the substance i; the unit depends on the
is,i
evaluation, e.g. area value;
ρ is the mass concentration of the internal standard, for the substance i, e.g. in nanograms per
is,i
millilitre, ng/ml.
Calibration can be done as linear regression or as a two point calibration over more than one decade after the
previous linearity check.
4.3.3.3 Quantification with internal standards
Add a known amount of the internal standards to the sample prior to sample preparation. Adjust this amount
of the internal standards in such a manner that the mass concentration ρ in the final volume, e.g. of the
is,i
extract, is nearly the same in the prepared samples as in the calibration solutions. Use the same solvent
composition for the standard solutions and the samples.
Calculate the mass concentration ρ of the substance using Equation (3).
i,sample
y
i,sample
− b
i
y ρ
mm
is,i,sample is,i i,sample extract is,i
ρ=×= × (3)
i,sample
aV ρ V
i sample is,i,sample extract sample
where
y is the measured response, e.g. peak area, of substance i in the sample extract;
i,sample
y is the measured response, e.g. peak area, of the internal standard, for substance i, of the
is,i,sample
sample;
ρ is the mass concentration of substance i in the sample extract, e.g. in nanograms per
i,sample extract
millilitre, ng/ml; usually calculated by the software;
ρ is the mass concentration of the internal standard in the sample extract, for substance i,
is,i,sample extract
e.g. in nanograms per millilitre, ng/ml; usually reported by the software;
ρ is the mass concentration of substance i in the water sample, e.g. in micrograms per litre,
i,sample
µg/l;
m is the mass of the added internal standard substance, e.g. in micrograms, µg;
is,i
V is the sample volume, in litres, l;
sample
a is defined in Equation (2);
i
b is defined in Equation (2).
i
4.3.3.4 Determination of recoveries of the internal standards
It is necessary to control the recovery of the internal standards for the total procedure for each sample.
This is possible by comparing the mean response from the internal standard in the calibration solutions with
the response obtained from the prepared sample. To achieve this, it is essential, that the final volume of the
prepared sample is known for calculating the theoretical final concentration of the internal standard in the
prepared sample. In routine use, the theoretical concentration of the internal standards in the prepared
samples is the same as in the calibration standard solutions, therefore no additional effort results.
Alternatively, a second internal standard, e.g. an injection standard can be used for the calculation of
recoveries of internal standards. This is a useful procedure if the final volumes after the sample preparation
vary, as frequently happens after enrichment to very small final volumes. This second internal standard is
added to the calibration solution and to the final prepared sample prior to the sample measurement. The final
concentration shall be the same for the calibration solution and the prepared sample with a theoretical final
volume. Now it is possible to obtain recoveries directly by comparing the responses of the internal and the
second internal (injection) standards obtained in the calibration with those obtained from prepared samples,
e.g. extracts.
yy×
is,i,sample is,inj,calibration
A=×100 (4)
is,i,sample
yy×
is,i,calibration is,inj,sample
where
A is the recovery of the internal standard, for substance i, in percent, %;
is,i,sample
y is the measured response, e.g. peak area, of the internal standard, for substance i, in the
is,i,sample
sample;
y is the measured response, e.g. peak area, of the internal standard, for substance i, in the
is,i,calibration
calibration solution;
y is the measured response, e.g. peak area, of the injection internal standard, in the sample;
is,inj,sample
y is the measured response, e.g. peak area, of the injection internal standard, in the
is,inj,calibration
calibration solution.
A control chart plot of the recoveries of the internal standards or another kind of documentation should
indicate that the recovery rates lie within defined control limits.
4.4 Limit of detection, limit of quantification
4.4.1 General aspects
In broad terms, the limit of detection is the smallest amount or concentration of an analyte in the test sample
that can be reliably distinguished from zero or blank (Reference [31] in the Bibliography). For some physical
parameters, e.g. pH and redox potential, the concept of limit of detection does not apply and no attempt
should be made to determine it for these parameters.
There is much diversity in the way in which the limit of detection and limit of quantification of an analytical
system is estimated. Most approaches are based on multiplication of the within-batch standard deviation of
results of typical matrix blanks or low-level material or the multiplication of the standard deviation of the
method, s , by a factor. These statistical inferences depend on the assumption of normality, which is at least
xo
10 © ISO 2009 – All rights reserved
questionable at low concentrations. Notwithstanding this, in method validation, a simple definition, leading to a
quickly implemented estimation of the detection limit, can be applied.
The different ways of estimating the limit of detection and limit of quantification which are described in the
following subclauses are optional. With the recommended minimum degrees of freedom, the value of the limit
of detection is quite uncertain, and may easily be in error by a factor of 2. Where more accurate estimates are
required, more complex calculations should be applied. For special cases, see ISO 11843, Parts 1 to 4.
The essential step after the estimation of the limit of detection and limit of quantification is the verification. The
analyst needs to prove that he is able to detect and, respectively, quantify the analyte at the estimated limits in
the respective matrix. If the criteria given in 4.4.6 are fulfilled, the estimated limits, e.g. the limit of detection
according to 4.4.2, are verified. After verification, the limit of detection and limit of quantification of different
laboratories can be compared, e.g. with quality targets.
4.4.2 Limit of detection based on standard deviation of results of blank samples
The limit of detection can be estimated as:
xs=+3 x (5)
LD 0 Bl
where
x is the limit of detection;
LD
s is the standard deviation of the outlier-free results of a matrix blank sample;
x is the mean concentration of the matrix blank.
Bl
If the applied method of calculation of results comprises the subtraction of a matrix blank, the second term, x ,
Bl
has to be ignored.
The precision estimate, s , should be based on at least 10 independent complete determinations of analyte
concentration in a typical matrix blank or low-level material, with no censoring of zero or negative results. For
that number of determinations, the factor of 3 corresponds to a significance level of α = 0,01.
4.4.3 Limit of detection based on the standard deviation of the method
For analytical methods which have a linear calibration function determined alternatively to the way described
in 4.4.2, the limit of detection can be estimated as:
x = 4s (6)
LD xo
where
x is the limit of detection;
LD
s is the standard deviation of the method (from calibration).
xo
4.4.4 Limit of detection based on baseline noise
For methods which show a baseline noise, the limit of detection, x , can be estimated as the concentration of
LD
the analyte at a signal/noise ratio S/N = 3 after blank correction (if possible).
4.4.5 Limit of quantification
The limit of quantification represents a concentration of the determinand that can reasonably be determined
with an acceptable level of accuracy. Usually it is arbitrarily taken as a fixed multiple of the detection limit.
For method validation, the limit of quantification, x , can be estimated as:
LQ
xx= 3 (7)
LQ LD
The factor k = 3 corresponds to a relative result uncertainty of approximately 33 %.
4.4.6 Verification of the limit of detection and the limit of quantification in the matrix
Verification of the limit of quantification is vitally important if routine samples frequently show analyte
concentrations near the limit of quantification. In that case, investigations for verification of the limit of
quantification should be performed regularly in routine analysis.
For verification of the limit of detection and limit of quantification, spiked blank matrix samples at these
concentration levels and blank matrix samples shall be analysed in the same manner as real samples, i.e.
under within-laboratory reproducibility conditions.
If the mean response of the samples spiked at the limit of detection level is greater than the maximum blank
value, the limit of detection is verified.
If the uncertainty of results for the samples spiked at the limit of quantification level is smaller than or equal to
the relative precision corresponding to the factor k, the limit of quantification is verified.
NOTE At the limit of quantification, the uncertainty component bias can be neglected, as the blank is included in the
estimation of the limit of detection and limit of quantification.
Examples for the verification of the limit of detection and the limit of quantification are given in Annex A.
4.4.7 Reporting limit
The reporting limit is a specific concentration at or above the limit of quantification that is reported to the client
with a certain degree of confidence. It is often defined on a project-specific basis. If the reporting limit is set
below the limit of quantification by the client, method modification is required.
4.5 Interferences and matrix effects
4.5.1 General considerations on matrix effects
An important source of random and systematic error in results is the presence of constituents of a sample
other than the determinand that cause an enhancement or a suppression of the analytical response, so-called
matrix effects.
Matrix effects have an essential influence on the measurement uncertainty. It is possible to correct systematic
matrix effects, e.g. to correct the multiplicative matrix effects by standard addition, internal standardization
(4.3.3) or a calibration with matrix material (4.3.2.2). In the case of additional matrix effects, the source should
be determined. The main sources are poor selectivity of the analytical method and blanks. A control of blanks
often is no problem, the enhancement of selectivity is mostly connected with higher expenses on the
instrumental equipment. After examining the sources of additional matrix effects, sometimes a subtraction is
possible (e.g. blanks).
Most analytical techniques produce accurate results with standard solutions at the optimal concentration.
4.5.2 Quantification for samples showing matrix effects when using basic calibration
If the matrix cannot be defined, as often happens in water analysis, it is not possible to test whether the slope
and intercept will be changed significantly. In these cases, it is possible to use the basic calibration, but in
combination with additional tests on matrix reference material or recovery examinations on real samples.
...
SPÉCIFICATION ISO/TS
TECHNIQUE 13530
Première édition
2009-03-15
Qualité de l'eau — Lignes directrices
pour le contrôle de qualité analytique
pour l'analyse chimique et
physicochimique de l'eau
Water quality — Guidance on analytical quality control for chemical and
physicochemical water analysis
Numéro de référence
©
ISO 2009
PDF – Exonération de responsabilité
Le présent fichier PDF peut contenir des polices de caractères intégrées. Conformément aux conditions de licence d'Adobe, ce fichier
peut être imprimé ou visualisé, mais ne doit pas être modifié à moins que l'ordinateur employé à cet effet ne bénéficie d'une licence
autorisant l'utilisation de ces polices et que celles-ci y soient installées. Lors du téléchargement de ce fichier, les parties concernées
acceptent de fait la responsabilité de ne pas enfreindre les conditions de licence d'Adobe. Le Secrétariat central de l'ISO décline toute
responsabilité en la matière.
Adobe est une marque déposée d'Adobe Systems Incorporated.
Les détails relatifs aux produits logiciels utilisés pour la création du présent fichier PDF sont disponibles dans la rubrique General Info
du fichier; les paramètres de création PDF ont été optimisés pour l'impression. Toutes les mesures ont été prises pour garantir
l'exploitation de ce fichier par les comités membres de l'ISO. Dans le cas peu probable où surviendrait un problème d'utilisation,
veuillez en informer le Secrétariat central à l'adresse donnée ci-dessous.
DOCUMENT PROTÉGÉ PAR COPYRIGHT
© ISO 2009
Droits de reproduction réservés. Sauf prescription différente, aucune partie de cette publication ne peut être reproduite ni utilisée sous
quelque forme que ce soit et par aucun procédé, électronique ou mécanique, y compris la photocopie et les microfilms, sans l'accord écrit
de l'ISO à l'adresse ci-après ou du comité membre de l'ISO dans le pays du demandeur.
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
Version française parue en 2010
Publié en Suisse
ii © ISO 2009 – Tous droits réservés
Sommaire Page
Avant-propos .iv
1 Domaine d'application .1
2 Références normatives.1
3 Termes et définitions .2
3.1 Termes liés aux méthodes de mesurage .2
3.2 Termes liés aux résultats de mesure .3
3.3 Termes liés à l'incertitude .5
4 Caractéristiques de performance des systèmes analytiques.6
4.1 Introduction.6
4.2 Domaine d'application de la méthode.6
4.3 Étalonnage .7
4.4 Limite de détection et limite de quantification .11
4.5 Interférences et effets de matrice .13
4.6 Exactitude (justesse et fidélité) et incertitude de mesure.14
4.7 Robustesse .15
4.8 Aptitude à l'emploi.16
5 Choix des systèmes analytiques .16
5.1 Considérations générales.16
5.2 Considérations d'ordre pratique.17
6 Contrôle qualité intralaboratoire.17
6.1 Généralités .17
6.2 Termes relatifs au contrôle qualité intralaboratoire .17
6.3 Contrôle de l'exactitude.18
6.4 Contrôle de la justesse .19
6.5 Contrôle de la fidélité .20
6.6 Principes d'application des cartes de contrôle.22
6.7 Conclusions .26
6.8 Cartes de contrôle ayant des critères de qualité fixés (cartes de contrôle cibles) .28
7 Contrôle qualité de l'échantillonnage.29
8 Contrôle qualité interlaboratoires.29
9 Contrôle qualité pour des modes opératoires analytiques longs ou analyses effectuées
peu fréquemment ou ponctuellement .30
9.1 Contrôle qualité pour des modes opératoires analytiques longs.30
9.2 Analyses effectuées peu fréquemment ou ponctuellement .30
Annexe A (informative) Vérification de la limite de détection et de la limite de quantification .32
Annexe B (informative) Nature et sources des erreurs analytiques.34
Annexe C (informative) Estimation de l'incertitude de mesure.37
Annexe D (informative) Exemple de contrôle qualité pour des modes opératoires analytiques
longs .39
Bibliographie.40
Avant-propos
L'ISO (Organisation internationale de normalisation) est une fédération mondiale d'organismes nationaux de
normalisation (comités membres de l'ISO). L'élaboration des Normes internationales est en général confiée
aux comités techniques de l'ISO. Chaque comité membre intéressé par une étude a le droit de faire partie du
comité technique créé à cet effet. Les organisations internationales, gouvernementales et non
gouvernementales, en liaison avec l'ISO participent également aux travaux. L'ISO collabore étroitement avec
la Commission électrotechnique internationale (CEI) en ce qui concerne la normalisation électrotechnique.
Les Normes internationales sont rédigées conformément aux règles données dans les Directives ISO/CEI,
Partie 2.
La tâche principale des comités techniques est d'élaborer les Normes internationales. Les projets de Normes
internationales adoptés par les comités techniques sont soumis aux comités membres pour vote. Leur
publication comme Normes internationales requiert l'approbation de 75 % au moins des comités membres
votants.
Dans d'autres circonstances, en particulier lorsqu'il existe une demande urgente du marché, un comité
technique peut décider de publier d'autres types de documents:
— une Spécification publiquement disponible ISO (ISO/PAS) représente un accord entre les experts dans
un groupe de travail ISO et est acceptée pour publication si elle est approuvée par plus de 50 % des
membres votants du comité dont relève le groupe de travail;
— une Spécification technique ISO (ISO/TS) représente un accord entre les membres d'un comité technique
et est acceptée pour publication si elle est approuvée par 2/3 des membres votants du comité.
Une ISO/PAS ou ISO/TS fait l'objet d'un examen après trois ans afin de décider si elle est confirmée pour trois
nouvelles années, révisée pour devenir une Norme internationale, ou annulée. Lorsqu'une ISO/PAS ou
ISO/TS a été confirmée, elle fait l'objet d'un nouvel examen après trois ans qui décidera soit de sa
transformation en Norme internationale soit de son annulation.
L'attention est appelée sur le fait que certains des éléments du présent document peuvent faire l'objet de
droits de propriété intellectuelle ou de droits analogues. L'ISO ne saurait être tenue pour responsable de ne
pas avoir identifié de tels droits de propriété et averti de leur existence.
L'ISO/TS 13530 a été élaborée par le comité technique ISO/TC 147, Qualité de l'eau, sous-comité SC 2,
Méthodes physiques, chimiques et biochimiques.
Cette première édition de l'ISO/TS 13530 annule et remplace l'ISO/TR 13530:1997, qui a fait l'objet d'une
révision technique.
iv © ISO 2009 – Tous droits réservés
SPÉCIFICATION TECHNIQUE ISO/TS 13530:2009(F)
Qualité de l'eau — Lignes directrices pour le contrôle de qualité
analytique pour l'analyse chimique et physicochimique de l'eau
1 Domaine d'application
La présente Spécification technique fournit des lignes directrices détaillées relatives au contrôle qualité
intralaboratoire et interlaboratoires afin de s'assurer que les résultats de l'analyse des eaux sont obtenus avec
un niveau d'exactitude connue.
La présente Spécification technique est applicable à l'analyse chimique et physicochimique de tous les types
d'eaux. Elle ne s'applique pas à l'analyse des boues et sédiments (même si la plupart de ses principes
généraux s'appliquent à de telles analyses) et ne concerne pas l'examen biologique ou microbiologique de
l'eau. Bien que l'échantillonnage soit un aspect important, celui-ci n'est que brièvement abordé.
Le contrôle qualité analytique décrit dans la présente Spécification technique s'applique à l'analyse de l'eau
effectuée dans le cadre d'un programme d'assurance de la qualité. La présente Spécification technique ne
traite pas des exigences détaillées d'assurance qualité pour l'analyse des eaux, consultables dans le guide
[20]
EURACHEM/CITAC (2002) .
Les recommandations de la présente Spécification technique sont en accord avec les exigences relatives à la
documentation d'assurance qualité établie (par exemple l'ISO/CEI 17025).
La présente Spécification technique s'applique à l'utilisation de toute méthode analytique employée dans son
domaine d'application, même s'il est admis que ses recommandations détaillées peuvent nécessiter une
interprétation et une adaptation afin de traiter de certains types de paramètres (par exemple paramètres non
spécifiques tels que les matières en suspension ou la demande biochimique en oxygène, DBO). En cas de
divergence entre les recommandations de la présente Spécification technique et les exigences d'une méthode
d'analyse normalisée, il convient de donner la priorité aux exigences de la méthode concernée.
2 Références normatives
Les documents de référence suivants sont indispensables pour l'application du présent document. Pour les
références datées, seule l'édition citée s'applique. Pour les références non datées, la dernière édition du
document de référence s'applique (y compris les éventuels amendements).
ISO 3534-2:2006, Statistique — Vocabulaire et symboles — Partie 2: Statistique appliquée
ISO 5725 (toutes les parties), Exactitude (justesse et fidélité) des résultats et méthodes de mesure
ISO 8466-1, Qualité de l'eau — Étalonnage et évaluation des méthodes d'analyse et estimation des
caractères de performance — Partie 1: Évaluation statistique de la fonction linéaire d'étalonnage
ISO 8466-2, Qualité de l'eau — Étalonnage et évaluation des méthodes d'analyse et estimation des
caractères de performance — Partie 2: Stratégie d'étalonnage pour fonctions d'étalonnage non linéaires du
second degré
ISO 13528:2005, Méthodes statistiques utilisées dans les essais d'aptitude par comparaisons
interlaboratoires
ISO/CEI 17025:2005, Exigences générales concernant la compétence des laboratoires d'étalonnages et
d'essais
Guide ISO 35, Matériaux de référence — Principes généraux et statistiques pour la certification
Guide ISO/CEI 43-1, Essais d'aptitude des laboratoires par intercomparaison — Partie 1: Développement et
mise en œuvre de systèmes d'essais d'aptitude
Guide ISO/CEI 43-2, Essais d'aptitude des laboratoires par intercomparaison — Partie 2: Sélection et
utilisation de systèmes d'essais d'aptitude par des organismes d'accréditation de laboratoires
3 Termes et définitions
Pour les besoins du présent document, les termes et définitions suivants s'appliquent.
3.1 Termes liés aux méthodes de mesurage
3.1.1
validation
confirmation par examen et apport de preuves objectives du fait que les exigences particulières en vue d'une
utilisation prévue déterminée sont remplies
[ISO/CEI 17025:2005]
3.1.2
exactitude
étroitesse de l'accord entre un résultat d'essai ou résultat de mesure et la valeur vraie
NOTE 1 Dans la pratique, la valeur de référence acceptée (3.2.6) remplace la valeur vraie.
NOTE 2 Le terme «exactitude», appliqué à un ensemble de résultats d'essai ou de mesure, implique une combinaison
de composés aléatoires et d'une erreur systématique commune ou d'une composante de biais.
NOTE 3 L'exactitude fait référence à une combinaison de justesse et de fidélité.
[ISO 3534-2:2006]
3.1.3
biais
différence entre l'espérance mathématique d'un résultat d'essai ou résultat de mesure et une valeur vraie
[ISO 3534-2:2006]
3.1.4
justesse
étroitesse de l'accord entre l'espérance mathématique d'un résultat d'essai ou d'un résultat de mesure et une
valeur vraie
NOTE 1 La mesure de la justesse est généralement exprimée en termes de biais.
NOTE 2 La justesse a été également appelée «exactitude de la moyenne». Cet usage n'est pas recommandé.
NOTE 3 Dans la pratique, la valeur de référence acceptée remplace la valeur vraie.
[ISO 3534-2:2006]
2 © ISO 2009 – Tous droits réservés
3.1.5
fidélité
étroitesse d'accord entre des résultats d'essai/de mesure indépendants obtenus sous des conditions stipulées
NOTE 1 La fidélité dépend uniquement de la distribution des erreurs aléatoires et n'a aucune relation avec la valeur
vraie ou la valeur spécifiée.
NOTE 2 La mesure de la fidélité est généralement exprimée en termes d'infidélité et est calculée à partir de l'écart-type
des résultats d'essai ou des résultats de mesure. Une fidélité faible est reflétée par un grand écart-type.
NOTE 3 Les mesures quantitatives de la fidélité dépendent de façon critique des conditions stipulées. Les conditions
de répétabilité et de reproductibilité sont des ensembles particuliers de conditions extrêmes stipulées.
[ISO 3534-2:2006]
3.1.6
limite de détection
valeur ou signal de sortie au-delà desquels on peut affirmer avec un certain niveau de confiance, par exemple
95 %, qu'un échantillon est différent d'un blanc ne contenant pas d'élément à déterminer
[ISO 6107-2:2006]
3.1.7
limite de quantification
valeur ou signal de sortie calculé à partir de la limite de détection, par exemple deux ou trois fois la limite de
détection à une concentration de l'élément à déterminer qui puisse raisonnablement être établie avec un
niveau acceptable de justesse et de fidélité
NOTE La limite de quantification peut être obtenue à l'aide d'un échantillon ou d'un étalon approprié comme étant le
plus petit point d'étalonnage sur la courbe d'étalonnage (à l'exclusion du blanc).
[ISO 6107-2:2006]
3.1.8
regroupement analytique
groupe de mesurages ou d'observations réalisés ensemble, soit simultanément, soit séquentiellement sans
interruption sur le même instrument par le même analyste, au moyen des mêmes réactifs
NOTE 1 Un regroupement analytique peut comporter plus d'une série d'analyses. Au cours d'un regroupement
analytique, il est attendu que l'exactitude et la fidélité du système de mesurage soient stables.
NOTE 2 Définition tirée de la Référence [33].
3.1.9
série d'analyses
groupe de mesurages ou d'observations d'étalons, d'échantillons et/ou de solutions témoins, qui ont été
effectués ensemble selon tous les modes opératoires définis, soit simultanément, soit séquentiellement, par
les mêmes analystes, au moyen des mêmes réactifs, du même équipement et de la même fonction
d'étalonnage
3.2 Termes liés aux résultats de mesure
3.2.1
erreur de mesure
résultat d'essai ou résultat de mesure moins la valeur vraie
NOTE 1 Dans la pratique, la valeur de référence acceptée remplace la valeur vraie.
NOTE 2 L'erreur est la somme des erreurs aléatoires et des erreurs systématiques.
NOTE 3 Adapté de l'ISO 3534-2:2006.
3.2.2
erreur systématique de résultat
composante de l'erreur d'un résultat qui, lors d'un certain nombre de résultats d'essai ou résultats de mesure,
pour la même caractéristique ou grandeur, demeure constante ou varie de façon prévisible
NOTE Les erreurs systématiques et leurs causes peuvent être connues ou inconnues.
[ISO 3534-2:2006]
3.2.3
erreur aléatoire de résultat
composante de l'erreur d'un résultat qui, lors d'un certain nombre de résultats d'essai ou résultats de mesure,
pour la même caractéristique ou grandeur, varie de façon imprévisible
NOTE Il n'est pas possible de corriger une erreur aléatoire.
[ISO 3534-2:2006]
3.2.4
valeur vraie
valeur qui caractérise une grandeur ou une caractéristique quantitative parfaitement définie dans les
conditions qui existent lorsque cette grandeur ou caractéristique quantitative est considérée
NOTE La valeur vraie d'une grandeur ou d'une caractéristique quantitative est une notion théorique et, en général,
ne peut pas être connue exactement.
[ISO 3534-2:2006]
3.2.5
valeur conventionnellement vraie
valeur d'une grandeur ou d'une caractéristique quantitative qui peut être substituée à une valeur vraie dans un
but déterminé
NOTE Une valeur conventionnellement vraie est, en général, considérée comme suffisamment proche de la valeur
vraie pour que la différence puisse être non significative pour le but donné.
[ISO 3534-2:2006]
3.2.6
valeur de référence acceptée
valeur qui sert de référence consensuelle pour une comparaison
NOTE La valeur de référence acceptée résulte:
a) d'une valeur théorique ou établie, fondée sur des principes scientifiques;
b) d'une valeur assignée ou certifiée, fondée sur les travaux d'une organisation nationale ou internationale;
c) d'une valeur de consensus ou certifiée, fondée sur un travail expérimental en collaboration et placé sous les auspices
d'un groupe scientifique ou technique;
d) de l'espérance, c'est-à-dire la moyenne de la population spécifiée de mesures, dans les cas où a), b) et c) ne sont
pas applicables.
[ISO 3534-2:2006]
4 © ISO 2009 – Tous droits réservés
3.2.7
matériau de référence certifié
matériau de référence, accompagné d'un certificat, dont une ou plusieurs valeurs des propriétés sont
certifiées par un mode opératoire qui établit son raccordement à une réalisation exacte de l'unité dans
laquelle les valeurs de propriété sont exprimées, et pour lequel chaque valeur certifiée est accompagnée
d'une incertitude à un niveau de confiance indiqué
NOTE Définition tirée de la Référence [36].
3.2.8
traçabilité métrologique
propriété d'un résultat de mesure selon laquelle ce résultat peut être relié à une référence par l'intermédiaire
d'une chaîne ininterrompue et documentée d'étalonnages dont chacun contribue à l'incertitude de mesure
[Guide ISO/CEI 99:2007]
3.3 Termes liés à l'incertitude
3.3.1
incertitude de mesure
paramètre non négatif qui caractérise la dispersion des valeurs attribuées à un mesurande, à partir des
informations utilisées
[Guide ISO/CEI 99:2007]
3.3.2
incertitude-type
u(x )
i
incertitude du résultat x d'une mesure exprimée en tant qu'écart-type
i
[Guide ISO/CEI 98-3:2008]
3.3.3
incertitude-type composée
u (y)
c
incertitude-type obtenue en utilisant les incertitudes-types individuelles associées aux grandeurs d'entrée
dans un modèle de mesure
[Guide ISO/CEI 99:2007]
3.3.4
incertitude élargie
U
produit d'une incertitude-type composée et d'un facteur supérieur au nombre un
NOTE 1 Le facteur dépend du type de la loi de probabilité de la grandeur de sortie dans un modèle de mesure et de la
probabilité de couverture choisie.
NOTE 2 Le facteur qui intervient dans la définition est un facteur d'élargissement.
[Guide ISO/CEI 99:2007]
3.3.5
facteur d'élargissement
k
facteur numérique utilisé comme multiplicateur de l'incertitude-type composée pour obtenir l'incertitude élargie
NOTE Un facteur d'élargissement est généralement compris entre 2 et 3.
[Guide ISO/CEI 98-3:2008]
4 Caractéristiques de performance des systèmes analytiques
4.1 Introduction
L'ISO/CEI 17025 exige la validation des méthodes; le processus de validation est décrit en détail dans le
[18]
guide EURACHEM (1998) . Le développement d'une nouvelle méthode analytique comprend une étape de
validation primaire, qui est réalisée au cours de la normalisation de la méthode. Les points importants à
prendre en compte lors de la validation primaire d'une méthode analytique sont les suivants:
⎯ domaine d'application de la méthode;
⎯ étalonnage;
⎯ limite de détection/limite de quantification;
⎯ interférences;
⎯ estimation de l'exactitude (justesse et fidélité);
⎯ incertitude de mesure;
⎯ robustesse;
⎯ aptitude à l'emploi.
Conformément à l'ISO/CEI 17025, le laboratoire doit confirmer qu'il peut utiliser correctement les méthodes
normalisées avant de les appliquer. Ce mode opératoire est appelé validation secondaire, et met l'accent sur
l'étalonnage et les interférences, ainsi que sur la limite de quantification et l'incertitude de mesure du
laboratoire. L'étalonnage et la quantification, les effets de matrice et l'incertitude de mesure sont abordés de
façon spécifique de 4.2 à 4.8.
4.2 Domaine d'application de la méthode
Il convient de donner une définition claire des formes des substances qui sont déterminées par la méthode et
également, pour éviter d'éventuelles ambiguïtés, des formes qui ne peuvent être déterminées. Il est bon de
souligner ici qu'il est recommandé que l'analyste choisisse une méthode analytique répondant à la définition
que l'utilisateur donne du paramètre à déterminer. Des paramètres non spécifiques nécessitent l'utilisation de
méthodes analytiques rigoureusement décrites en vue d'obtenir des résultats fiables et comparables.
De nombreuses substances existent dans l'eau sous diverses formes ou «espèces» et de nombreux
systèmes analytiques fournissent une réponse différente à ces diverses formes. Par exemple, lorsqu'il est
nécessaire de séparer les fractions «dissoutes» et «particulaires», il est important de définir avec précision la
nature ainsi que la porosité du filtre à utiliser.
Il est important de disposer d'une description précise du type et de la nature des échantillons avant de choisir
le système analytique. Les précautions à prendre lors de l'analyse de l'échantillon dépendent dans une large
mesure de l'échantillon. Il est nécessaire que l'analyste dispose d'informations aussi complètes que possible
sur les types d'échantillons, les niveaux de concentration ainsi que les interférences éventuelles. Il convient
que le domaine d'application contienne une déclaration claire des types d'échantillons et des matrices
d'échantillons pour lesquels la méthode est appropriée. Si besoin, il convient de signaler également les
principaux types d'échantillons et matrices pour lesquels la méthode ne convient pas.
Le domaine d'application correspond aux concentrations la plus faible et la plus élevée pour lesquelles des
essais de fidélité et de biais ont été effectués au moyen du système sans modification. Lorsqu'une extension
peut être utilisée pour permettre l'examen d'échantillons contenant des concentrations plus importantes que la
limite supérieure, par exemple par analyse après dilution, il convient de considérer cet examen comme une
méthode différente, mais dont les caractéristiques de performance peuvent être déduites à partir des valeurs
mentionnées pour le domaine initial.
La gamme de concentration concernée peut avoir un effet marqué sur le choix de la technique analytique; la
plus faible concentration concernée joue un rôle primordial.
6 © ISO 2009 – Tous droits réservés
4.3 Étalonnage
4.3.1 Bases de l'étalonnage et de la quantification
Une fonction d'étalonnage est déterminée à partir des valeurs d'information y, obtenues en mesurant des
i
concentrations étalons données x. L'écart-type de la méthode s qui en résulte ou l'intervalle de confiance
i xo
est une caractéristique de performance de cet étalonnage, et non une caractéristique de performance de la
quantification. La quantification de la teneur en analyte des échantillons à l'aide d'une fonction d'étalonnage
est fondée sur l'interpolation. La caractéristique de performance la plus importante pour la quantification est
l'estimation de l'incertitude de mesure (voir 4.6), dont l'une des composantes est l'incertitude d'étalonnage.
En fonction du type de relation fonctionnelle entre la concentration en analyte et la réponse de mesure,
différents outils mathématiques ou statistiques peuvent être utilisés. Les modèles mathématiques et
statistiques sont soumis à différentes hypothèses; par conséquent, si les performances de l'équipement
analytique sont essentielles, le type d'approche mathématique l'est également. Il convient que les modèles
aident l'analyste à trouver une relation fonctionnelle claire et fiable pour l'étalonnage. Il convient que les
modèles ne limitent pas les capacités de l'équipement analytique. En conséquence, il convient de consacrer
une attention particulière au choix du modèle d'étalonnage, en tenant compte de l'aptitude à l'emploi et de
l'incertitude de mesure requise.
⎯ Le modèle de régression linéaire (ISO 8466-1): le modèle classique pour l'étalonnage, avec pour limites
la distribution normale des réponses et une homogénéité des variances sur le domaine d'analyse.
Normalement, cette homogénéité des variances est obtenue uniquement dans un domaine d'analyse
d'une ou au maximum de deux décades de concentration. L'applicabilité s'en trouve donc limitée.
⎯ Le modèle de fonctions d'étalonnage non linéaires du second degré (ISO 8466-2): après la mise en
évidence d'une absence significative de linéarité. Ce modèle présente les mêmes limites que la
régression linéaire: distribution normale des réponses et homogénéité des variances sur le domaine
d'analyse.
⎯ Le modèle de régression pondérée: la valeur d'information étant pondérée de la variance réciproque, il
est possible d'effectuer un étalonnage sur plus d'une décade de gamme de concentration.
⎯ L'étalonnage sur plus d'une décade de concentration avec au moins deux concentrations, par exemple
pour la spectrométrie de masse à plasma inductif (ICP/MS). Il faut vérifier la linéarité au préalable avec
au moins cinq concentrations, par exemple au moyen d'une représentation graphique.
NOTE Des instructions détaillées sur les modes opératoires d'étalonnage sont données dans les Références [26]
et [28].
Il convient d'évaluer la contribution de tous ces types d'étalonnage à l'incertitude de mesure. Par exemple,
analyser à N reprises, au moins trois réplicats, un étalon indépendant ou un matériau de référence certifié
dans la gamme de concentration concernée, et calculer les résultats conformément à la méthode
d'étalonnage appliquée. La composante de l'incertitude-type, s , représentant l'écart des résultats individuels,
p
ρ , par rapport à la valeur de référence, ρ, est calculée comme suit:
i
N
ρρ−
()
∑ i
i=1
s = (1)
p
N −1
où
s est la composante de l'incertitude-type de l'étalonnage;
p
N est le nombre de mesures répétées.
Il convient de choisir N par rapport au niveau de confiance requis. Il est recommandé de comparer s à
p
l'incertitude cible de la mesure correspondante pour établir la contribution acceptable.
4.3.2 Stratégies d'étalonnage
4.3.2.1 Étalonnage de base ou des instruments
Ce type d'étalonnage est effectué sans composés apportés par la matrice et sans préparation de l'échantillon.
L'étalonnage utilise des solutions étalons pures. Cette méthode est peu onéreuse et directement adaptée à la
quantification, si les composés apportés par la matrice ne modifient pas significativement la pente et le point
d'interception de la fonction d'étalonnage. Normalement cet étalonnage est plus précis qu'un étalonnage
comprenant toutes les étapes de préparation des échantillons.
4.3.2.2 Étalonnage avec matrice
Ce type d'étalonnage est comparable à l'étalonnage de base (4.3.2.1). Il ne comporte aucune étape de
préparation des échantillons. Des solutions de substances étalons dans une matrice sans analyte ou une
matrice artificielle sont utilisées à la place des solutions étalons pures.
4.3.2.3 Étalonnage sur l'ensemble du mode opératoire
Ce type d'étalonnage comprend toutes les étapes de préparation des échantillons et les solutions
d'étalonnage sont préparées avec un matériau représentatif de la matrice. Il est adapté pour vérifier si les
composés de la matrice ou les étapes de préparation des échantillons modifient ou non la pente et le point
d'intersection de la fonction d'étalonnage par rapport à la fonction d'étalonnage de base. Il est souvent difficile
de trouver un matériau représentatif de la matrice pouvant servir de base pour les solutions étalons. Dans le
contexte de l'analyse des eaux, cela signifie qu'il faut trouver l'eau naturelle exempte de blanc qui représente
les matrices possibles des autres eaux naturelles.
Si les composés de la matrice modifient significativement le point d'interception et la pente de la fonction
d'étalonnage par rapport à l'étalonnage de base, il est possible d'utiliser cet étalonnage sur l'ensemble du
mode opératoire d'essai pour la quantification.
4.3.3 Étalonnage interne
4.3.3.1 Aspects généraux
Le recours à l'étalonnage interne pour la quantification des concentrations minimise les erreurs possibles à la
fois pendant l'injection et suite aux pertes pendant les étapes de prétraitement des échantillons, ainsi que les
différences de volumes d'extrait final de l'échantillon et les modifications des rendements d'extraction
provoquées par les effets de matrice. Il convient que les substances utilisées comme étalons internes
présentent les propriétés suivantes.
⎯ Il convient que les propriétés physicochimiques soient les mêmes vis-à-vis des étapes du mode
opératoire concernées par l'erreur qu'il convient de corriger. S'il convient que le mode opératoire soit
contrôlé dans sa totalité par l'étalonnage interne, des composés marqués par des isotopes sont
recommandés; s'il convient de contrôler uniquement les volumes finaux ou la détection, il est possible
d'utiliser d'autres substances, représentatives de ces étapes.
⎯ Il convient qu'il n'y ait aucune interférence de mesure avec les étalons internes.
⎯ Aucune présence de l'étalon interne, ni dans les échantillons réels ni dans les blancs, qui ne puisse être
évitée.
⎯ Concentration de l'étalon interne ajouté: dans la gamme dynamique de la méthode, de préférence dans
la même gamme de concentration que les analytes.
⎯ Il convient qu'ils présentent les mêmes intensités de réponses que les analytes.
8 © ISO 2009 – Tous droits réservés
Il est recommandé d'effectuer l'étalonnage interne comme un étalonnage de base, car tous les effets de matrice
multiplicatifs sont corrigés par l'étalon interne s'il présente les propriétés susmentionnées. Sinon, un étalonnage
interne sur l'ensemble du mode opératoire présente les mêmes inconvénients que ceux décrits en 4.3.2.3, les
matériaux représentatifs de la matrice font souvent défaut. Comme toutes les autres possibilités permettant de
corriger les effets de matrice, l'étalonnage interne ne peut pas surmonter les effets de matrice additifs.
4.3.3.2 Étalonnage avec des étalons internes
Le calcul de la fonction d'étalonnage est généralement proposé en option dans les programmes de
quantification de la plupart des logiciels d'analyse de données des fabricants.
Ajuster les concentrations conformément à la sensibilité de l'équipement utilisé et à la gamme de mesure
requise. Évaluer le domaine de linéarité puis préparer un étalonnage. Établir la fonction linéaire par
étalonnage de base des couples de valeurs yy et ρρ des séries mesurées à l'aide de l'Équation (2):
iis,i iis,i
y ρ
ii
=+ab (2)
ii
y ρ
is,i is,i
où
y est la réponse mesurée de la substance i; l'unité dépend de la méthode de mesure, par exemple
i
valeur de la surface;
ρ est la concentration massique de la substance i (étalon externe) dans la solution étalon d'analyse,
i
par exemple en nanogrammes par millilitre, ng/ml;
a est la pente de la fonction d'étalonnage de la substance i; l'unité dépend de l'évaluation, par
i
exemple millilitres de valeur de la surface par nanogramme, valeur de la surface·ml/ng;
b est l'ordonnée à l'origine de la fonction d'étalonnage; l'unité dépend de la méthode de mesure, par
i
exemple valeur de la surface;
y est la réponse mesurée de l'étalon interne pour la substance i; l'unité dépend de la méthode de
is,i
mesure, par exemple valeur de la surface;
ρ est la concentration massique de l'étalon interne, pour la substance i, par exemple en
is,i
nanogrammes par millilitre, ng/ml.
L'étalonnage peut être réalisé sous forme de régression linéaire ou d'étalonnage sur deux points sur plus
d'une décade après la vérification précédemment décrite de la linéarité.
4.3.3.3 Quantification avec des étalons internes
Ajouter une quantité connue des étalons internes à l'échantillon avant la préparation de l'échantillon. Ajuster
cette quantité d'étalons internes pour que la concentration massique ρ dans le volume final, par exemple de
is,i
l'extrait, soit pratiquement identique dans les échantillons préparés et dans les solutions d'étalonnage. Utiliser
la même composition de solvant pour les solutions étalons et les échantillons.
Calculer la concentration massique ρ de la substance à l'aide de l'Équation (3).
i,sample
y
i,sample
− b
i
y mmρ
is,i,sample i,sample extract
is,i is,i
ρ=× = × (3)
i,sample
aV V
ρ
i sample is,i,sample extract sample
où
y est la réponse mesurée, par exemple l'aire de pic, de la substance i dans l'extrait de
i,sample
l'échantillon;
y est la réponse mesurée, par exemple l'aire de pic, de l'étalon interne, pour la
is,i,sample
substance i, de l'échantillon;
ρ est la concentration massique de la substance i dans l'extrait de l'échantillon, en
i,sample extract
nanogrammes par millilitre, ng/ml; généralement calculée par le logiciel;
ρ est la concentration massique de l'étalon interne dans l'extrait de l'échantillon, pour la
is,i,sample extract
substance i, par exemple en nanogrammes par millilitre, ng/ml; généralement donnée
par le logiciel;
ρ est la concentration massique de la substance i dans l'échantillon d'eau, par exemple en
i,sample
microgrammes par litre, µg/l;
m est la masse de substance d'étalon interne ajoutée, par exemple en microgrammes, µg;
is,i
V est le volume de l'échantillon, en litres, l;
sample
a est définie dans l'Équation (2);
i
b est définie dans l'Équation (2).
i
4.3.3.4 Détermination des rendements d'extraction des étalons internes
Il est nécessaire de contrôler le rendement d'extraction des étalons internes pour l'ensemble du mode
opératoire utilisé pour chaque échantillon.
Pour ce faire, il est possible de comparer la réponse moyenne de l'étalon interne dans les solutions
d'étalonnage à la réponse obtenue à partir de l'échantillon préparé. Pour y parvenir, il est essentiel de
connaître le volume final de l'échantillon préparé pour calculer la concentration finale théorique de l'étalon
interne dans l'échantillon préparé. En routine, la concentration théorique des étalons internes dans les
échantillons préparés est la même que dans les solutions mères pour étalonnage, par conséquent il n'en
résulte aucun effort supplémentaire.
Sinon, un second étalon interne, par exemple un étalon d'injection, peut être utilisé pour calculer les
rendements d'extraction des étalons internes. Ce mode opératoire est utile en cas de variation des volumes
finaux après la préparation des échantillons, comme cela se produit fréquemment après l'enrichissement de
très petits volumes finaux. Ce second étalon interne est ajouté à la solution d'étalonnage et à l'échantillon
préparé au stade final avant la mesure de l'échantillon. La concentration finale doit être la même pour la
solution d'étalonnage et l'échantillon préparé à un volume final théorique. Il est alors possible d'accéder aux
rendements d'extraction directement en comparant les réponses de l'étalon interne et du second étalon
interne (injection) obtenues dans l'étalonnage à celles obtenues à partir des échantillons préparés, par
exemple les extraits.
yy×
is,i,sample is,inj,calibration
A=×100 (4)
is,i,sample
yy×
is,i,calibration is,inj,sample
où
A est le rendement d'extraction de l'étalon interne, pour la substance i, en pourcentage, %;
is,i,sample
y est la réponse mesurée, par exemple l'aire de pic, de l'étalon interne, pour la substance i,
is,i,sample
dans l'échantillon;
y est la réponse mesurée, par exemple l'aire de pic, de l'étalon interne, pour la substance i,
is,i,calibration
dans la solution d'étalonnage;
y est la réponse mesurée, par exemple l'aire de pic, de l'étalon interne d'injection, dans
is,inj,sample
l'échantillon;
10 © ISO 2009 – Tous droits réservés
y est la réponse mesurée, par exemple l'aire de pic, de l'étalon interne d'injection, dans la
is,inj,calibration
solution d'étalonnage.
Il convient qu'une carte de contrôle des rendements d'extraction des étalons internes ou un autre type de
document indique que les taux de récupération se trouvent dans les limites de contrôle définies.
4.4 Limite de détection et limite de quantification
4.4.1 Généralités
De manière générale, la limite de détection est la plus petite quantité ou concentration en analyte dans
l'échantillon d'essai qui peut être distinguée de manière fiable de zéro ou du blanc (Référence [31]). Pour
certains paramètres physiques, par exemple le pH et le potentiel d'oxydation-réduction, le concept de limite
de détection ne s'applique pas et il convient dans ce cas de ne pas essayer de le déterminer.
Il existe un grand nombre de moyens d'estimer la limite de détection et la limite de quantification d'un système
analytique. La plupart des approches consistent à multiplier l'écart-type intra-série des résultats de blancs de
matrice ou de matériau de faible niveau typiques ou l'écart-type de la méthode s par un facteur. Ces
xo
inférences statistiques se fondent sur l'hypothèse de normalité, laquelle est discutable au moins aux faibles
concentrations. Malgré tout, dans le cadre de la validation d'une méthode, une définition simple conduisant à
une estimation rapidement mise en œuvre de la limite de détection peut être appliquée.
Les différents moyens d'estimer la limite de détection et la limite de quantification décrits dans les
paragraphes suivants sont facultatifs. Avec les degrés de liberté minimaux recommandés, la valeur de la
limite de détection est assez incertaine et peut facilement être erronée d'un facteur 2. Lorsque des
estimations plus exactes sont requises, il convient d'appliquer des calculs plus complexes. Pour des cas
spéciaux, voir l'ISO 11843, Parties 1 à 4.
L'étape essentielle après l'estimation de la limite de détection et de la limite de quantification est la vérification.
Il est nécessaire que l'analyste prouve qu'il est capable de détecter et, respectivement, de quantifier l'analyte
aux limites estimées dans la matrice considérée. Si les critères donnés en 4.4.6 sont remplis, les limites
estimées, par exemple les limites de détection conformément à 4.4.2, sont vérifiées. Après vérification, la
limite de détection et la limite de quantification de différents laboratoires peuvent être comparées, par
exemple à des valeurs cibles de qualité.
4.4.2 Limite de détection basée sur l'écart-type des résultats des blancs
La limite de détection peut être estimée comme suit:
xs=+3 x (5)
LD 0 BI
où
x est la limite de détection;
LD
s est l'écart-type des résultats d'un échantillon assimilé à un blanc de matrice après élimination des
valeurs aberrantes;
x est la concentration moyenne du blanc de la matrice.
Bl
Si la méthode de calcul des résultats appliquée comprend la soustraction d'un blanc de matrice, il faut ignorer
le second terme, x .
Bl
Il convient de baser l'estimation de la fidélité s sur au moins 10 déterminations complètes indépendantes de
la concentration en analyte dans un blanc de matrice type ou un matériau de faible niveau, sans rejet de
résultats nuls ou négatifs. Pour ce nombre de déterminations, le facteur 3 correspond à un niveau de
confiance de α = 0,01.
4.4.3 Limite de détection basée sur l'écart-type de la méthode
Pour les méthodes ayant une fonction d'étalonnage linéaire déterminée d'une autre manière que celle décrite
en 4.4.2, la limite de détection peut être estimée comme suit:
xs= 4 (6)
LD xo
où
x est la limite de détection;
LD
s est l'écart-type de la méthode (à partir de l'étalonnage).
xo
4.4.4 Limite de détection basée sur le bruit de fond
Pour les méthodes qui présentent un bruit de fond, la limite de détection, x , peut être estimée comme la
LD
concentration en analyte à un rapport signal/bruit S/N = 3 après la correction du blanc (si possible).
4.4.5 Limite de quantification
La limite de quantification représente une concentration de l'analyte pouvant être raisonnablement déterminée
avec un niveau d'exactitude acceptable. Généralement, celui-ci est arbitrairement choisi comme un multiple
fixe d
...












Questions, Comments and Discussion
Ask us and Technical Secretary will try to provide an answer. You can facilitate discussion about the standard in here.
Loading comments...