118
3 Traceability: Reference Materials
as suitable as a sample standard prepared from real material; in fact, the
former may even be quite unfit for the intended purpose (see Box 3.10).
__ Box 3.10
The "history" behind the preparation of a CRM is as important as are its certified value and
associated uncertainty. The form in which the analytes are present in a sample standard is
decisive. There are two commercially available sample standards of soil for the determination
of traces of polychlorinated biphenyls (PCBs) that are endorsed and supplied by two renown
organizations; their certified values are at the same concentration level (in the microgram-per-kilogram range).when aliquots of the two sample standards are subjected to supercritical fiuid extraction under different experimental conditions (pressure, temperature,
presence or absence of methanol as modifier, etc.). the results are rather disparate. Thus, one
of the sample standards provides extraction yields of 99% under all experimental conditions
tested, whereas the other barely reaches 25- 30%. What is the origin of the differences
between these two analytical references, deemed excellent by their issuers? The answer is
quite simple: the two materials were prepared in a different way. In one, a PCB mixture was
spiked to clean soil; matrix-analyte interactions were thus minimal and extraction of the
analytes was very easy as a result. In the other, samples were collected from soil under the
polluting effects of an industrial area, so the PCBs were strongly bound to the matrix and thus
very difficult to remove.
Using one or the other standard as reference will make a significant difference. In fact, only
the sample prepared from polluted soil is fit for the purpose. How many analyses and assessments will clearly have been spurious if the sample obtained by spiking clean soil is used as
the sample standard?
(2) Analytical measurement standards should be varied enough to meet the
wide range of needs posed by CMPs, which are discussed in the next section.
(3) The way a standard is to be stored and used should be carefully described in
its accompanying documentation.
(4) Analytical standards should be affordable. Currently available materials of
this type are largely fine chemicals and thus of a high added value that
ultimately results in an also high price to the end user. In fact, even though
the starting material (e.g. cement, sea water, plant material) is usually inexpensive, its processing and providing certified values (and their uncertainties) for one or more quantities entails extensive, painstaking work
that results in the typically high added value of fine chemicals.
(5) The reference properties of primary standards should be well established
and, preferably, certified - particularly the accuracy and uncertainty in the
numerical features (quantities) that are to be used as references.
(6) Primary analytical standards should be stable, i. e. their properties should not
vary with handling or time (e.g. through interaction with atmospheric agents).
There are two types of stability in this context that relate to the time elapsed
from preparation (stated as an expiration date) and from use (viz. from the
moment the container is opened). Some standards can only be used once.
(7) Primary analytical standards should obviously be homogeneous and their
homogeneity assured.
3 Traceability: Reference Materials
as suitable as a sample standard prepared from real material; in fact, the
former may even be quite unfit for the intended purpose (see Box 3.10).
__ Box 3.10
The "history" behind the preparation of a CRM is as important as are its certified value and
associated uncertainty. The form in which the analytes are present in a sample standard is
decisive. There are two commercially available sample standards of soil for the determination
of traces of polychlorinated biphenyls (PCBs) that are endorsed and supplied by two renown
organizations; their certified values are at the same concentration level (in the microgram-per-kilogram range).when aliquots of the two sample standards are subjected to supercritical fiuid extraction under different experimental conditions (pressure, temperature,
presence or absence of methanol as modifier, etc.). the results are rather disparate. Thus, one
of the sample standards provides extraction yields of 99% under all experimental conditions
tested, whereas the other barely reaches 25- 30%. What is the origin of the differences
between these two analytical references, deemed excellent by their issuers? The answer is
quite simple: the two materials were prepared in a different way. In one, a PCB mixture was
spiked to clean soil; matrix-analyte interactions were thus minimal and extraction of the
analytes was very easy as a result. In the other, samples were collected from soil under the
polluting effects of an industrial area, so the PCBs were strongly bound to the matrix and thus
very difficult to remove.
Using one or the other standard as reference will make a significant difference. In fact, only
the sample prepared from polluted soil is fit for the purpose. How many analyses and assessments will clearly have been spurious if the sample obtained by spiking clean soil is used as
the sample standard?
(2) Analytical measurement standards should be varied enough to meet the
wide range of needs posed by CMPs, which are discussed in the next section.
(3) The way a standard is to be stored and used should be carefully described in
its accompanying documentation.
(4) Analytical standards should be affordable. Currently available materials of
this type are largely fine chemicals and thus of a high added value that
ultimately results in an also high price to the end user. In fact, even though
the starting material (e.g. cement, sea water, plant material) is usually inexpensive, its processing and providing certified values (and their uncertainties) for one or more quantities entails extensive, painstaking work
that results in the typically high added value of fine chemicals.
(5) The reference properties of primary standards should be well established
and, preferably, certified - particularly the accuracy and uncertainty in the
numerical features (quantities) that are to be used as references.
(6) Primary analytical standards should be stable, i. e. their properties should not
vary with handling or time (e.g. through interaction with atmospheric agents).
There are two types of stability in this context that relate to the time elapsed
from preparation (stated as an expiration date) and from use (viz. from the
moment the container is opened). Some standards can only be used once.
(7) Primary analytical standards should obviously be homogeneous and their
homogeneity assured.