The Selection and Use of Reference Materials

8/13/2019 The Selection and Use of Reference Materials

1/11

EEE/RM/062rev3EL/07-06/02/977

THE SELECTION AND USE OF REFERENCE MATERIALS

A BASIC GUIDE FOR LABORATORIES AND ACCREDITATION BODIES

ABSTRACT

The aim of this paper is to provide a short, simple and user-friendly guide for laboratories, and accreditation

and certification bodies. It employs mainly ISO (3) and I! (") definitions and provides references to othermore comprehensive and e#pert te#ts. It is intended to provide help for the ine#perienced rather than the e#pert

and necessarily simplifies some topics. It includes discussion and clarification of some misunderstood issues.$lthough based on the re%uirements of chemical measurement, it is intended to be also of use in other areas of

measurement.

INTRODUCTION

There are a number of authoritative and detailed te#ts on various aspects of reference materials and these are

listed at the end of this paper, together &ith some internationally recognised ("-') definitions. This paper aimsto provide a simple guide on the use of reference materials (!s) &ithin a &ider %uality programme. eference

materials are an important tool in realising a number of aspects of measurement %uality and are used for methodvalidation, calibration, estimation of measurement uncertainty, training and for internal %uality control (*)

and e#ternal %uality assurance ($) (proficiency testing) purposes.

In the &ider sense, the validity of measurements can be assured &hen+

&or is carried out to a clearly defined customer re%uirement

validated methods and e%uipment are used

%ualified and competent staff undertae the &or

comparability &ith measurements made in other laboratories is assured (traceability and

measurement uncertainty)

independent evidence of performance is available (proficiency testing)

&ell defined * and $ procedures are employed, preferably involving third party accreditation

Often a measurement operation serves more than one %uality purpose and there can be overlap of function as

illustrated in igure ". ifferent types of reference materials are re%uired for different functions. or e#ample, acertified reference material &ould be desirable for method validation but a &oring level reference material

&ould be ade%uate for *.

!ore detailed guidance on the $ of chemical measurements, including cover of reference materials,

calibration, * and validation is provided in a /oint *IT$* 0urachem guide (1).

TYPES OF REFERENCE MATERIALS


2/11

!s are used to support measurements concerned &ith chemical composition, biological, clinical, physical,

engineering properties and miscellaneous areas such as taste and odour. They may be characterised for

2identity (e.g. chemical structure, fibre type, microbiological species etc.) or for 2property values (e.g. amountof specified chemical entity, hardness etc.). Some commonly encountered types of reference materials are as

follo&s+

". Pure substancescharacterised for chemical purity and4or trace impurities.

5. Standard sout!ons and "as #!$tures, oftenprepared gravimetrically from pure substances and used forcalibration purposes.

3. Matr!$ re%erence #ater!as, characterised for the composition of specified ma/or, minor or trace chemicalconstituents. Such materials may be prepared from matrices containing the components of interest, or by

preparing synthetic mi#tures.

'. P&'s!co(c&e#!ca re%erence #ater!ascharacterised for properties such as melting point, viscosity, and

optical density.

1. Re%erence ob)ects or arte%actscharacterised for functional properties such as taste, odour, octane number,

flash point and hardness. This type also includes microscopy specimens characterised for propertiesranging from fibre type to microbiological specimens.

CLASSIFICATION OF REFERENCE MATERIALS

T&o classes of materials are recognised by ISO, namely 2certified reference materials6 (*!s) and 2referencematerials (!s). *!s must by definition be traceable to an accurate realisation of the unit in &hich the

property values are e#pressed. 0ach property value must be accompanied by an uncertainty at a stated level ofconfidence. !s are materials &hose property values are sufficiently homogeneous and &ell established to be

used for the calibration of an apparatus, the assessment of a measurement method, or for assigning values to

materials.

or most chemical reference materials produced before the late "778s, the measurement uncertainty valuesgiven by the producers are unliely to have been estimated by the no& recommended ISO (9,:) procedure. The

actual uncertainty can be e#pected to be larger than stated by a factor of 5-3, &here only &ithin laboratory

precision measurements are used and by a smaller factor &here certification included a range of validatedmethods and a number of laboratories. $lso some materials sold as *!s have no stated traceability.

The follo&ing classes of reference materials are also encountered+

;rimary reference material

Secondary reference materials ecreasing ). ;an (7) has also published a useful paper onthis topic.

$n internationally recognised %uality approval system to assure the competence of ! producers (based on

ISO ?uide 3' and reference "8) is under consideration &ithin the International @aboratory $ccreditation *o

operation, but it &ill tae some years for any system to mae a significant impact on the maret place.


3/11

Measure#ent Met&od Traceab!!t'

;rimary method SI

!ethod of no&n bias SI4International standard

Independent method(s) esults of specified methods

Interlaboratory comparison esults of specified methods

$ combination of certification procedures is sometimes employed, such as a consensus value derived from an

interlaboratory comparison &here primary methods &ere used. In the absence of formally stated traceability it&ill be necessary for the user to mae /udgements about implicit traceability, based on the certification data

available in reports and the technical literature. It is important to ensure that chemical interferences and matri#effects are ade%uately addressed in arriving at both the certified value and its uncertainty.


4/11

0stimation of bias (the difference bet&een the measured value and the true value) is one of the most difficult

elements of method validation, but appropriate !s can provide valuable information, &ithin the limits of the

uncertainty of the !s certified value(s) and the uncertainty of the method being validated. $lthough traceablecertified values are highly desirable, the estimation of bias differences bet&een t&o or more methods can be

established by use of less rigorously certified !s. *learly the !s must be &ithin the scope of the method interms of matri# type, analyte concentration etc. and ideally a number of !s covering the full range of the

method should be tested. Dhere minor modifications to a &ell-established method are being evaluated then lessrigorous bias studies can be employed.

eplicate measurement of the !, covering the full range of variables permitted by the method being validated

can be used to estimate the uncertainty associated &ith any bias, &hich should normally be corrected for ("1).

The uncertainty associated &ith an ! should be no greater than one third of that of the sample measurement.

*er!%!cat!on o% t&e Correct Use o% a Met&od

The successful application of a valid method depends on its correct use, both &ith regard to operator sill andsuitability of e%uipment, reagents and standards. !s can be used for training, for checing infre%uently used

methods and for trouble shooting &hen une#pected results are obtained.

Ca!brat!on

=ormally a pure substance ! is used for calibration of the measurement stage of a method. Other components

of the test method, such as sample digestion, separation and derivatisation are, of course, not covered and lossof analyte, contamination and interferences and their associated uncertainties must be addressed as part of the

validation of the method. The uncertainty associated &ith ! purity &ill contribute to the total uncertainty ofthe measurement. or e#ample, an ! certified as 77.7E pure, &ith an e#panded uncertainty < (F5) of 8."E

&ill contribute an uncertainty component of 8."E to the overall measurement uncertainty budget. In the case oftrace analysis, this level of uncertainty &ill rarely be important but for assay &or, it can be e#pected to be

significant.

Some other methods, such as G analysis, use matri# !s for calibration of the complete analytical process.

In addition to a close matri# match, the analyte form must be the same in the samples and !s, and theanalytical concentrations of the !s must span that of the samples ("5).

ISO ?uide 35 and reference : provide additional useful information.

+ua!t' Contro and +ua!t' Assurance ,+C-+A.

!s should be characterised &ith respect to homogeneity, stability, and the certified property value(s). or in-

house *, ho&ever, the latter re%uirement can be rela#ed, but ade%uate homogeneity and stability are essential.Similar re%uirements apply to samples used to establish ho& &ell or badly measurements made in different

laboratories agree. In the case of proficiency testing, homogeneity is essential and sample stability &ithin thetime-scale of the e#ercise must be assessed and controlled. $lthough desirable, the cost of certifying the

property values of proficiency testing samples often prohibits this being done and consensus mean values areoften used instead. $s a conse%uence, there often remains some doubt concerning the reliability of assigned

values used in proficiency testing schemes. This is because, although the consensus mean of a set of data hasvalue, 2the ma/ority is not necessarily correct and as a conse%uence the values carry some undisclosed element

of uncertainty. The interpretation of proficiency testing data thus needs to be carried out &ith caution.

ASSESSMENT OF THE SUITABILITY OF REFERENCE MATERIALS

$s previously indicated the ey %uality parameter is the uncertainty associated &ith the certified value and the

reliability of the uncertainty estimate.


5/11

$ protocol for assessing the suitability of !s is detailed in igure 5 and discussed belo&. The user must

assess the appropriateness and fitness for purpose of any ! based on the customer and analytical

re%uirements. actors to be considered include the follo&ing+

". The suitability of a reference material depends on the details of the analytical specification. !atri# effects

and other factors such as concentration range can be more important than the uncertainty of the certifiedvalue as detailed in (""). The factors to consider include+

!easurand including analyte

!easurement range (concentration)

!atri# match and potential interferences

Sample sie

Homogeneity and stability

!easurement uncertainty

alue assignment procedures (measurement and statistical)

5. The validity of the certification and uncertainty data, including conformance of ey procedures &ith ISO

?uide 31 and other ISO re%uirements (9,:).

3. Trac record of both the producer and the material. or e#ample, &hen a ! in use has been sub/ected to

an interlaboratory comparison, cross-checed by use of different methods, or there is e#perience of use in anumber of laboratories over a period of years.

'. $vailability of a certificate and report conforming to ISO ?uide 3".

1. emonstrated conformance of the production of the reference materials &ith %uality standards such as ISO

?uide 3' or I@$* re%uirements ("8), or compliance of the measurement of property values &ith ISO4I0*":851 ("9) re%uirements.

$ll or some of the re%uirements may be specified in the customer and analytical specification, but often it &ill

be necessary for the analyst to use professional /udgement. inally, %uality does not necessarily e%uate to small

uncertainty and fitness for purpose criteria need to be used.

Cert!%!cates and Su//ort!n" Re/orts

Ideally, a certificate complying &ith ISO ?uide 3" and a report covering the characterisation, certification and

statistical analysis procedures, complying &ith ISO ?uide 31, &ill be available. Ho&ever, many !s,particularly older materials and materials not specifically produced as !s, may not fully comply &ith ISO

?uides 3" and 31. $lternative, e%uivalent information in &hatever form it is available, that provides credibleevidence of compliance can be considered acceptable. 0#amples include the follo&ing+ technical reports, trade

specifications, papers in /ournals or reports of scientific meetings and correspondence &ith suppliers.

Assess#ent o% t&e Su!tab!!t' o% Re%erence Mater!as

@aboratories must be able to e#plain and /ustify the basis of selection of all !s and of course any decision not

to use a !. In the absence of specific information it is not possible to assess the %uality of an !. The rigour&ith &hich an assessment needs to be conducted depends on the criticality of the measurement, the level of the

technical re%uirement and the e#pected influence of the particular ! on the validity of the measurement. Only&here the choice of ! can be e#pected to significantly affect measurement results is a formal suitability

assessment re%uired.

IN(HOUSE PREPARATION OF REFERENCE MATERIALS

High %uality !s are demanding and costly to produce and if materials are available from other sources it is

not normally cost effective for laboratories to mae their o&n. Ho&ever should this be necessary, there are

guides available ("5,"3) to help the non-specialist laboratory prepare their o&n !s. Some of the ey issuesthat need to be considered are+ selection of materials (appropriateness, native material versus spies, materialpreparation etc.), homogeneity testing, preparation and pacaging (homogeneity, contamination, stability etc.),

Page 5 of 10


6/11

stability testing, certification studies, uncertainty estimation, documentation and $, certification approval,

storage, distribution and relifing.

DEFINITIONS

Ca!brat!on ,0.1 Set of operations that establish, under specified conditions, the relationship bet&een the values

of %uantities indicated by a measurement instrument or measuring system or values represented by a materialmeasure or a reference material, and the corresponding values realised by standards.

Cert!%!ed Re%erence Mater!a ,0.1eference material, accompanied by a certificate, one or more of &hoseproperty values are certified by a procedure &hich establishes traceability to an accurate realisation of the unit

in &hich the property values are e#pressed, and for &hich each certified value is accompanied by an uncertainty

at a stated level of confidence.

Pr!#ar' #et&od ,2.1$ primary method is a method having the highest metrological %ualities, &hoseoperation can be completely described, and understood and for &hich a complete uncertainty statement can be

&ritten do&n in terms of SI units. $ primary direct method measures the value of an unno&n &ithoutreference to a standard of the same %uantity. $ primary ratio method measures the ratio of an unno&n to a

standard of the same %uantityA its operation must be completely described by a measurement e%uation. Themethods identified as having the potential to be primary methods are+ isotope dilution mass spectrometry+

gravimetry, covering gravimetric mi#tures and Jgravimetric analysisKA titrimetryA coulometryA determination offreeing point depressionA differential scanning calorimetry and nuclear magnetic resonance spectroscopy.

Other methods such as chromatography, &hich has e#tensive applications in organic chemical analysis, havealso been proposed.

Re%erence #ater!a ,RM. ,0.1!aterial or substance one or more of &hose property values are sufficientlyhomogeneous and &ell established to be used for the calibration of an apparatus, the assessment of a

measurement method, or for assigning values to materials.

Traceab!!t' ,0.1 ;roperty of a result of a measurement or the value of a standard &hereby it can be related to

stated references, usually national or international standards, through an unbroen chain of comparisons allhaving stated uncertainties.

Uncerta!nt' o% #easure#ent ,0.1;arameter associated &ith the result of a measurement that characterises thedispersion of the values that could reasonably be attributed to the measurand.

*a!dat!on ,3.1 *onfirmation by e#amination and provision of ob/ective evidence that the particularre%uirements of a specified intended use are fulfilled. !ethod validation has not been formally defined but a

?uide on the topic is available (')

4EY REFERENCE MATERIALSPUBLICATIONSThe follo&ing guides have been prepared by ISO 0!*O, the international committee that leads on reference

material matters+

ISO ?uide 38+"775 Terms and definitions used in connection &ith reference materials ISO ?uide 3"+5888 *ontents of certificates of reference materials

ISO ?uide 35+"77: *alibration of chemical analysis and use of certified reference materials

ISO ?uide 33+5888 473 (ec "77')+ ?uidelines for the production and

certification of B* reference materials

Page 6 of 10


7/11

=IST ;ublication 598-"88 ("773)+ Standard eference !aterials - Handboo for S!


8/11

;repared by Bernard Ling on behalf of and &ith assistance from the 000-! &oring group.

Page 8 of 10


9/11

F!"ure 01 O5era/ bet6een %unct!ons assoc!ated 6!t& Measure#ent Traceab!!t' and Ana't!ca

+ua!t'

F!"ure 21 Assess#ent o% t&e su!tab!!t' o% a Re%erence Mater!a ,/a"e 7.

Page 9 of 10

*a!dat!on Traceabe

ca!brat!on

+C8+A

Measure#entUncerta!nt'8

Traceab!!t'

*a!d #easure#ent


10/11

Page 10 of 10

De%!ne ana't!ca

re9u!re#ent ,Note 0.

Seect cand!date RM and

obta!n su//ort!n"

!n%or#at!on ,Note 2.

REFERENCE MATERIAL SUITABLE

L!#!tat!ons but RM !s

best a5a!abe and

#eets #!n!#u#

re9u!re#ents

RM NOT SUITABLE: SEE4 ALTERNATI*E

OR DO;N GRADE RE+UIREMENT

Yes

Yes

Yes

Yes

No

No

No

No

Re/orted RM c&arac(ter!st!cs %u' #atc& ana't!ca re9u!re#ent

Su//ort!n" e5!dence concern!n" 9ua!t'

!s sat!s%actor' ,Note 3.

L!#!tat!ons but RM !s

best a5a!abe and

#eets #!n!#u#

re9u!re#ents


11/11

=otes to ig. 5+

=ote "+ The analytical re%uirements specification should include details concerning the measurand,concentration, traceability, measurement uncertainty, etc

=oite 5+ Ley characterists should be available in the ! *ertificate. $dditional information, for e#ample

details of the method(s) used for value assignment and the full measurement uncertainty budget should also beavailable in the *ertificate or in a supporting report.

=ote 3+ See p' and p1

October 5885

The Selection and Use of Reference Materials

Documents

Transcript of The Selection and Use of Reference Materials