Metals
217
Moreover, the variability of the mass of the fIlter after drying is higher than for new
fIlters. That means that a part of the fIltration variability is not due to the sample
heterogeneity but is introduced in calculating the net mass by the allowance made for
fIlter tare. The bias due to this error could be as great as the true fIltration error, or
even greater as indicated by a comparison with the much better reproducibility
obtained by calculating the fIltration error on the fIltrate volume (Table 9.4).
The sonifier method is affected by another error, namely the efficiency of the
sonification treatment in removing the particulate matter from the filter: this error
was quantified by weighing ten filters before and after sonification. Because only a
tiny increase of under 1 % of the initial fIlter net mass was observed at the end of the
treatment, even assuming that this was all due to particles left on the fIlter itself, about
91 % of the 1.39 mg collected has clearly been removed. This is very likely to produce
an undervaluation of the efficiency of the method because, as previously seen, after
wetting and drying (Table 9.3) the filter mass is somewhat higher than for new fIlters.
Moreover, microscopic observation of the filters after the sonification treatment
indicates that practically all of the particles were removed.
Table 9.3
Filtration reproducibility (on a mass basis) (from [625])
Sample
Mean mass / mg
RSD,%
n
Candidate reference material
suspensionsPlatihypnidium riparioides
6.82
19
10
Lagarosiphon major
7.89
23
6
Freshwater plankton
10.12
9
8
11.05
10
8
Freshwater particulate matterLake Maggiore (pre-filtered
on a 50-lUll screen)
4.41
30
12
Lake Maggiore (pre-filtered
on 10.00-1Ull glass-fibre
pre-filters)
1.84
8
8
Filters -
New filters (0.4 lUll)
14.81
7
20
Acid washed and dried filters
15.35
11
20
Table 9.4
Filtration reproducibility (on a volume basis) (from [625])
Sample
Mean filtrate
RSD,%
n
volume/ml
Candidate reference material suspension -
Platihypnidium riparioides
490
5
10
Freshwater particulate matter -
Lake Maggiore, 1.0-1Ull filters (pre-filtered on a 50-lUll
screen)
720
7
14
Lake Maggiore, 0,4-1Ull filters (pre-filtered on a 50-lUll
screen)
600
6
12
Lake Maggiore, 0,4-1Ull filters (pre-filtered on 10.0-1Ull
glass-fibre pre-filters)
1380
5
3
217
Moreover, the variability of the mass of the fIlter after drying is higher than for new
fIlters. That means that a part of the fIltration variability is not due to the sample
heterogeneity but is introduced in calculating the net mass by the allowance made for
fIlter tare. The bias due to this error could be as great as the true fIltration error, or
even greater as indicated by a comparison with the much better reproducibility
obtained by calculating the fIltration error on the fIltrate volume (Table 9.4).
The sonifier method is affected by another error, namely the efficiency of the
sonification treatment in removing the particulate matter from the filter: this error
was quantified by weighing ten filters before and after sonification. Because only a
tiny increase of under 1 % of the initial fIlter net mass was observed at the end of the
treatment, even assuming that this was all due to particles left on the fIlter itself, about
91 % of the 1.39 mg collected has clearly been removed. This is very likely to produce
an undervaluation of the efficiency of the method because, as previously seen, after
wetting and drying (Table 9.3) the filter mass is somewhat higher than for new fIlters.
Moreover, microscopic observation of the filters after the sonification treatment
indicates that practically all of the particles were removed.
Table 9.3
Filtration reproducibility (on a mass basis) (from [625])
Sample
Mean mass / mg
RSD,%
n
Candidate reference material
suspensionsPlatihypnidium riparioides
6.82
19
10
Lagarosiphon major
7.89
23
6
Freshwater plankton
10.12
9
8
11.05
10
8
Freshwater particulate matterLake Maggiore (pre-filtered
on a 50-lUll screen)
4.41
30
12
Lake Maggiore (pre-filtered
on 10.00-1Ull glass-fibre
pre-filters)
1.84
8
8
Filters -
New filters (0.4 lUll)
14.81
7
20
Acid washed and dried filters
15.35
11
20
Table 9.4
Filtration reproducibility (on a volume basis) (from [625])
Sample
Mean filtrate
RSD,%
n
volume/ml
Candidate reference material suspension -
Platihypnidium riparioides
490
5
10
Freshwater particulate matter -
Lake Maggiore, 1.0-1Ull filters (pre-filtered on a 50-lUll
screen)
720
7
14
Lake Maggiore, 0,4-1Ull filters (pre-filtered on a 50-lUll
screen)
600
6
12
Lake Maggiore, 0,4-1Ull filters (pre-filtered on 10.0-1Ull
glass-fibre pre-filters)
1380
5
3