close

Вход

Забыли?

вход по аккаунту

?

3-540-27093-0 47

код для вставкиСкачать
RM certification is the whole process of
obtaining the property values and their uncertainties, which includes homogeneity testing,
stability testing, and RM characterization [5].
ISO Guide 35 [1] requires one to show that
the value of such a certified property does not
exhibit a systematic error specific to a method or to a laboratory. By widespread opinion,
correctness of analytical results is an obvious
prerequisite for the RM characterization in
contrast to stability and homogeneity studies
in which analytical bias is acceptable [5].
In the present paper we would like to
discuss primarily the situation when a systematic error in measurement (analytical)
results, specific to a measurement method
or to a laboratory developing an IHRM, can
be overcome. This is possible when a certification of an IHRM is based on a comparative approach providing a characterization
of the value carried by the IHRM in comparison to the value carried by the reference
material with higher metrological status and
sufficiently similar matrix (a CRM).
Papers published in this section do not necessarily reflect the opinion of the Editors,
the Editorial Board and the Publisher. Apart from exceptional circumstances,
they are not submitted to the usual referee procedure and go essentially unaltered.
Accred Qual Assur (2002) 7:122 124
DOI 10.1007/s00769-001-0436-9
' Springer-Verlag 2002
Ilya Kuselman
Alexander Weisman
Wolfhard Wegscheider
Traceable property
values of in-house
reference materials
Abstract The traceability of in-house reference materials (IHRM) is discussed. It is
shown that a systematic error in results of a
measured value, specific to a measurement
method or to a laboratory developing an
IHRM, can be overcome if a comparative
approach to IHRM characterization is used.
A traceability chain of the value carried by
the IHRM to the value carried by the reference material with higher metrological status
and sufficiently similar matrix (for example,
a certified reference material CRM according to ISO Guide 30) is helpful in such a
case. The chain is realized when the IHRM
samples are analysed simultaneously with
the CRM samples under the same conditions.
This and other traceability chains necessary
for the IHRM development are examined as
the measurement information sources.
Keywords Traceability • Reference
materials • Uncertainty estimation •
Certification • Comparative approach
Introduction
The traceability of the value carried by a
reference material (RM) should be demonstrated by the RM producer [1 3]. The producer shall provide the traceability of results
of its measurements to the national or international measurement standards. Where this
is not possible, the correlation of results
with the values of national or international
certified reference materials (CRMs) is required. Ideally, the values of the CRMs
should themselves be traceable [2].
More than 220 producers of CRMs
throughout the world produce today
12,000 20,000 materials with dif ferent matrixes, analytes and properties [4]. However, many testing (analytical) laboratories
cannot find suitable CRMs in the market
and develop in-house reference materials
(IHRMs) themselves. Often IHRMs are developed in a laboratory to conserve the corresponding expensive CRMs. For example,
a pharmaceutical company Chemagis
Ltd. produces 30 active pharmaceutical ingredients: steroids, benzodiazepines, antihistamines, hipolipidaemics, blood flow reactants, etc. Only for a few of them Mometasone Furoate, Fluticasone Propionate
and Dobutamine Hydrochloride are of ficial reference standards for assay supplied
by US, British and European Pharmacopoeias with prices of about $ 180 per unit
(50 200 mg). Thus, to support its customers Chemagis is forced to develop IHRMs
for assay as well as for impurities and related substances of each produced compound.
Therefore, certification of such IHRMs that
leads to traceable values is very important.
Comparative approach
This approach is based on the transmission
of the measurement information from a
corresponding CRM to the IHRM [6 9].
The following conditions are required for
the IHRM characterization:
1) IHRM and CRM are similar materials,
2) the difference in concentrations of the
IHRM and CRM matrix components,
which are not under characterization,
and corresponding properties of the materials (for example, solubility) does
not hinder the use of the same measurement (analytical) method for both the
IHRM and CRM,
3) the concentrations of a component under characterization in IHRM and CRM
do not differ by more than a factor of
two [8].
Test portions in pairs one of the IHRM
and one of the CRM are analysed by the
same analyst and method in the same laboratory and conditions, each pair practically
simultaneously. The concentration of the
IHRM component under characterization,
CIHRM, is compared with its certified
concentration, CCRM, in the CRM using
the differences in the analysis results
Ei=(CIHRM i C CRM i ) for all pairs:
i=1, 2, ..., n (n≥20). From this comparison
the characterized value is calculated as
CIHRM = CCRM + Eavg ,
where
Eavg = Σ Ei/n.
(1)
Obviously, even if CIHRM i and CCRM i have
an additive systematic error, Ei is free from
this error by definition. Additivity of bias is
a reasonable approximation for nearly iden-
270
I. Kuselman • A. Weisman • W. Wegscheider
Table 1 Results of the alcohol determination in the IHRM and CRM LGC5404
i
CIHRM i
CCRM i
Ei
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Average
Standard
deviation
4.75
4.80
4.78
4.77
4.76
4.77
4.76
4.76
4.75
4.76
4.75
4.80
4.79
4.80
4.78
4.76
4.75
4.78
4.75
4.76
4.769
0.018
5.03
5.04
5.03
5.00
5.00
5.03
5.02
5.03
5.00
4.99
5.02
5.01
5.03
5.02
5.04
5.02
4.99
5.02
5.03
5.01
5.018
0.015
0.28
0.24
0.25
0.23
0.24
0.26
0.26
0.27
-0.25
0.23
0.27
0.21
0.24
0.22
0.26
0.26
0.24
0.24
0.28
0.25
0.249
0.019
tical matrices: a multiplicative bias component is assumed negligible at similar concentrations of the analyte in the CRM and
IHRM as limited above. Therefore, Eavg and
CIHRM by Eq. (1) are also unbiased. So, the
characterization standard uncertainty [10] is
u2 (CIHRM) = [u2 (CCRM) + u2 (Eavg )]1/2
u2 (Eavg ) = Σ (Ei E avg )2 / (n 1) n,
(2)
(3)
and u(CCRM) is the standard uncertainty of
the value carried by the CRM. If the CRM
can be selected according to the criteria of
ILAC-G9 Guidelines [11] as satisfactory or
acceptable, i.e., if u(CIHRM)/u(CCRM)>4, the
uncertainty of the value carried by the
CRM is negligible. Otherwise it should be
taken into account by Eq. (2).
A numerical example of the calculations is shown below.
Characterization of an IHRM of ethanol in
water. A reference spirit CRM LGC5404
from LGC with the certified value of the
ethanol concentration in water of
5.00–0.03% by volume at the 95% level of
confidence is used for the alcohol determination in beer [12]. So, CCRM=5.00% and
u(CCRM)=0.03/2=0.015% by volume (2 is
the coverage factor). To prepare an IHRM
a suitable ethanol was diluted gravimetrically with water to produce a solution close
to 5% ethanol concentration. The solution
when thoroughly mixed was suggested to
be homogeneous. Twenty test portions in
pairs one of the IHRM and one of the
CRM LGC5404 were analysed in the
same conditions using a standard gas-chro-
It is important that the information on a
value carried by a CRM is obtained from
measurement (analytical) results in other
laboratories (not or not only in the laboratory-producer of the IHRM), probably using different methods. Therefore, if a suitable (adequate) CRM is not available, participation of a second laboratory in an
IHRM characterization is desirable, even
when an unbiased validated method is
used, to evaluate a possible bias specific to
the laboratory developing the IHRM.
Traceability chains of the method form in
this case a traceability scheme of the value
carried by the IHRM.
Fig. 1 Traceability chains of a value CIHRM
carried by an IHRM. Here m is the mass of
the IHRM and V is its volume
matographic method. Results of the analysis are presented in Table 1. The average
difference between the results for IHRM
and CRM pairs shown in Table 1 is
Eavg= 0.25% by volume. The value carried
by the IHRM according to Eq. (1) is
CIHRM=5.00+( 0.25)=4.75% by volume.
Note, that the average result of the CRM
analysis CCRM-avg = Σ CCRM-i /20=5.02%
is biased from the certified value CCRM=
5.00% for 0.02% by volume. It is not influencing the value of CIHRM as far as CCRM avg
is not used directly in the CIHRM calculation. The standard uncertainty in Eavg calculated by Eq. (3) is u(Eavg)=0.004% by
volume. So, the characterization standard
uncertainty by Eq. (2) is u(CIHRM)=
[0.0152+0.0042]1/2=0.016% by volume.
Conclusions
It is shown that a systematic error in results of a measured value, specific to a
measurement method or to a laboratory
developing an IHRM, can be overcome if
a comparative approach to IHRM characterization is used. A traceability chain
from the value carried by the IHRM to the
value carried by the reference material
with higher metrological status and sufficiently similar matrix is helpful in such
cases. The chain is realized when the
IHRM samples are analysed simultaneously with the CRM samples under the same
conditions.
If a suitable (adequate) CRM is not
available, participation of a second laboratory in an IHRM characterization is desirable, even when an unbiased validated
method is used, to assess the laboratory bias. Traceability chains of the method form
in this case a traceability scheme of the
value carried by the IHRM.
Traceability chains
If the first steps of a procedure as described above are weighing and dissolving,
a concentration of the IHRM component
under characterization CIHRM (the value
carried by the IHRM) expressed in % by
volume has three traceability chains:
1) of the IHRM mass to the SI kg,
2) of the IHRM volume after dissolving,
also to the SI kg, since calibration of
volumetric flasks is performed gravimetrically, and
3) of the IHRM carried value comparison
to the CRM carried value.
In such a case CIHRM can be shown as a
sun with three beams [13] that are metrological pyramids (traceability chains with
minimal uncertainty in the top of the pyramid
and maximal in the bottom): see Fig. 1. As a
rule, uncertainties in traceability chains to the
SI kg are negligible in comparison with the
ones from the own analytical process. So, the
problem is just how narrow the bottom of the
third pyramid is, i.e., how much information
is lost in the chain IHRM CRM.
References
1. ISO Guide 35 (1989) Certification of
Reference Materials General and Statistical Principles. 2nd ed, Geneva
2. ILAC-G12 (2000) Guidelines for the
Requirements for the Competence of
Reference Materials Producers
3. De Bievre P (2000) Accred Qual Assur
5:224 230
4. Zschunke A (ed) (2000) Reference Materials in Analytical Chemistry. A Guide
for Selection and Use. Springer, Berlin,
p 211
5. Van der Veen AMH, Linsinger TPJ,
Schimmel H, Lamberty A, Pauwels J
(2001) Accred Qual Assur 6:290 294
6. Shaevich AB (1971) Measurement and
Standardization of Chemical Composition of Substances. Publishing House
of Standards, Moscow, 280 p (in Russ.)
7. Shaevich AB (1987) Reference Materials for Analytical Purposes. Publishing
House Chemistry , Moscow ,
pp 147 149 (in Russ.)
Traceable property values of in-house reference materials
8. Pliner YuL (ed) (1988) Metrological
Assurance of Materials Composition in
Ferrous Metallurgy. Handbook. 2nd ed,
Publishing House Metallur gy , Moscow, pp 157 163 (in Russ.)
9. Pliner YuL, Kuzmin IM (1989) Metrological Problems in Analytical Control
of Metallic Products Quality. Publishing House Metallur gy , Moscow ,
pp 157 165 (in Russ.)
10. EURACHEM/CITAC Guide (2000)
Quantifying Uncertainty in Analytical
Measurement. 2nd ed, Teddington
11. ILAC-G9 (1996) Guidelines for the Selection and Use of Certified Reference
Materials
12. Kuselman I, Anisimov B, Goldfeld I
(2001) Accred Qual Assur 6:107 1 10
13. Kuselman I (2000) Proceedings of The
Millennium International Conference
of The Israel Society for Quality. Vol I,
Jerusalem, pp 124 129
MEETING REPORT
I. Kuselman (✉)
The National Physical Laboratory
of Israel (INPL), Givat Ram,
Jerusalem 91904, Israel
e-mail: [email protected]
Tel.: +972-2-6536534
Fax: +972-2-6520797
A. Weisman
Chemagis Ltd., P.O.Box 9091,
Tel-Aviv 61090, Israel
W. Wegscheider
Department of General
and Analytical Chemistry,
University of Leoben,
Franz-Josef Strasse 18, 8700 Leoben,
Austria
271
Документ
Категория
Без категории
Просмотров
2
Размер файла
113 Кб
Теги
540, 27093
1/--страниц
Пожаловаться на содержимое документа