Chapter 2
66
was made. The precision and the accuracy of our analyses were assessed by measuring
replicates and the Certified Reference Materials (CRM) BCR-414, respectively.
Table 2.2: Determined concentrations (mg.kg
-1
) and resulting recoveries (%) of the Certified Reference
Material plankton BCR-414, compared to certified or indicative (*) values, analyzed by HR-ICP-MS.
External calibration curves were ran at the beginning, middle and end of the run. Data
processing included drift correction through
115
In monitoring, Cd-Mo oxides correction, and
blank correction.
Major elements
The measurements of major elements (Ca, K, Mg, Na, Sr) were
performed using an ICP-AES
(ULTIMA 2) at the PSO using the archive 3% HNO
3
solutions that were prepared for the totally
digested trace element concentrations. The precision and the accuracy of our analyses were
assessed by measuring replicates and the Certified Reference Materials (CRM) BCR-414.
Ba
Ca
K
Mg
Na
P
Sr
Indicative value (µg.g
-1
)
30
65955
6850
12840
220
Determined concentration (n=2; µg.g
-1
) 28 ± 1 51133 ± 1057 5887 ± 165 2053 ± 78 6601 ± 13 10910 ± 275 188 ± 2
Recovery (%)
95
78
86
85
85
Table 2.3: Determined concentrations (µg.g
-1
) and resulting recoveries (%) of the Certified Reference
Material plankton BCR-414, compared to indicative values, analyzed by ICP-AES.
Certified/indicative* value (mg/kg)
Measured value (n=5; mg/kg)
Recovery (%)
Cd
0.383 ± 0.014
0.385 ± 0.020
101%
Pb
3.97 ± 0.19
2.95 ± 0.11
74%
Al*
2154 ± 803 *
3112 ± 464
144%
P*
12840 ± 4978 *
15541 ± 3732
121%
Ti*
105 ± 25 *
95 ± 8
90%
V
8 ± 0.2
9 ± 1
111%
Cr
23.8 ± 1.2
24.8 ± 3.07
104%
Mn
299 ± 13
301 ± 42
101%
Fe*
1850 ± 190 *
1860 ± 230
101%
Co*
1.43 ± 0.06 *
1.44 ± 0.21
101%
Ni
18.8 ± 0.8
17.8 ± 2.91
95%
Cu
29.5 ± 1.3
29.7 ± 3.8
101%
Zn
111.6 ± 2.5
112.4 ± 18.8
101%
Mo*
1.35 ± 0.2 *
1.37 ± 0.28
101%
Ba*
29.5 ± 3.9 *
32.7 ± 5.15
111%
Chapter 2
67
It was then possible to estimate the calcium carbonate concentration, assuming that all Ca
comes from the carbonate (CaCO
3
)
and the seawater, following this equation:
[Ca]
CaCO3
= Ca – [(Ca/Na)
SW
× Na] (Equation 2.1)
where
[Ca]
CaCO3
is the estimated concentration of calcium carbonate in µmol.L
-1
;
Ca and
Na
are respectively the calcium and sodium concentrations analyzed in µmol.L
-1
;
(Ca/Na)
SW
is
equal to 0.022 mol.L
-1
and represents the mean ratio determined in seawater by Copin-
Montégut (1996).
The CaCO
3
concentrations ranged from 0.002 ± 0.0001 µmol.L
-1
to 0.139 ± 0.0003 µmol.L
-1
in
the large size fraction and from 0.006 ± 0.0002 nmol.L
-1
to 1.54 ± 0.01 µmol.L
-1
in the small
size fraction. The median errors accounted for 1.2% (n=51) and 0.5% (n=76) of the
concentration for the
large and small size fractions, respectively.
Biogenic Silica
Following the method of Ragueneau et al. (2005), BSi was estimated after alkaline digestions
allowing the correction of the LSi that can represent up to 90% of the total particulate silica in
coastal waters. All experiments using clean polymethylpentene and polypropylene (VWR)
centrifuge tubes. One Supor punch (1-53 µm particle size), or a 25 mm diameter Nuclepore
filter previously Beta counted (>53 µm particle size), was placed in a centrifuge tube and
digested with 8 mL of a 0.2 M NaOH solution (pellets for analysis, Merck) at 95°C during 1h in
a hot bath. After cooling, the pH was neutralized by adding 2 mL of 1M HCl (Analar Normapur,
Merck). Tubes were then centrifuged at ambient temperature for 10 min at 4000 rpm (Thermo
Scientific Multifuge 3S+/3SR+) and the supernatant was separated from the remaining
suspended material. Finally, 1 mL of supernatant was diluted 10 times with Milli-Q water for
the silicic acid analyses, and similarly, 1 mL of supernatant was diluted 10 times with Milli-Q
water for the aluminium analyses. At the end of the first digestion, all the BSi should be
converted into silicic acid. However, it is possible that a part of LSi is dissolved during this first
digestion involving a bias for the determination of BSi concentration. This bias is resolved with
Chapter 2
68
a second digestion, equivalent to the first one, allowing the determination of a ratio (Si/Al)
characteristic of the minerals. The corrected BSi concentration is then deduced using the
following equation:
BSi = [Si
1
] – [Al
1
] × (Si/Al)
2
(Equation 2.2)
where
BSi is the final concentration of biogenic silica in µmol.L
-1
;
[Si
1
] and
[Al
1
] are the
concentrations of silicic acid and aluminium determined during the first digestion in µmol.L
-1
;
(Si/Al)
2
is the ratio of the concentrations of silicic acid and aluminium determined during the
second digestion.
After each alkaline digestion of the filter, silicic acid concentrations were determined following
the automated acid/molybdate colorimetric method (Aminot and Kérouel, 2007), using an
AutoAnalyzer3 Bran&Luebbe (detection limit: 0.07 µmol.L
-1
). Dissolved aluminium
concentrations were determined by a fluorimetric method (fluorimeter MITHRAS LB 940)
detecting a complex Aluminium-Lumogallion developed by Hydes and Liss (1976; detection
limit: 0.05 µmol.L
-1
). The median precision of the values obtained during 4 different
instrumental sessions reached 0.25% in >53 µm
particles, and 7.1% in 1-53 µm.
µmol
BSi 1-53 µm
BSi >53 µm
range (n=60)
0.02 - 2.2
0.02 - 48
blank (n=9)
0.001
0.02
detection limit
0.003
0.002
Table 2.4: Ranges, blanks and detection limits of BSi (µmol) in large and small particles collected during
GEOVIDE.
The BSi concentrations ranged from 0.0009 ± 0.0006 µmol.L
-1
to 1.414 ± 0.004 µmol.L
-1
in the
large size fraction and from 0.005 ± 0.002 µmol.L
-1
to 0.871 ± 0.012 µmol.L
-1
in the small size
fraction. The median errors accounted for 1.8% (n=57) and 9.1% (n=55) of the concentration
for the
large and small size fractions, respectively.