Anions in seawater - Determination of Anions in Natural and Treated Waters - page 406

the sample, 0.50mm for the acidic molybdate solution, and 0.76mm for the oxalic acid
was used. The reductant (0.50mm) was mixed with a portion of the sample pumped by a
0.76mm tube.
The sample stream could be directed either through a short reaction coil of 1mm id
PTFE tubing or through a long coil of 1.6mm id Tygon tubing after the confluence of the
sample and the molybdate streams. The long coil, which could be heated, was used to
increase the sensitivity when necessary. The line between the confluence with oxalic acid
and the injector was 25cm of 1mm id. The line between the injector and the detector was
25cm of 0.8mm id. PTFE when stannous chloride was used as the reductant. Longer post-
injector reaction coils had to be used with ascorbic acid as the reductant because of the
slower reaction kinetics. The post-injector coil could also be heated to accelerate the
reduction rate with ascorbic acid.
The reductant was injected by a rotary injection valve (Rheodyne Model 5041) with a
pneumatic actuator. The injection loop was shortened to give a volume of 20µL. It was
necessary to match the salinity of the reductant solution to that of the acidified sample.
Otherwise false blank peaks occurred due to the difference in refractive index of the
carrier stream and the injected solution [205]. Matching of the refractive indexes was best
accomplished by mixing the reductant solution with a portion of the sample (Fig. 3.19).
A light emitting diode photometer [205] as used as the detector. The flow cell had a
path length of 2cm and an id of 1.5mm. A high output infrared light emitting diode and a
silicon phototransistor (Radio Shack 276-143 and 276-130) were used as light source
and detector, respectively. The maximum output of the light emitting diode was measured
to be at 886nm with a bandwidth of 37nm. The molybdenum blue complex formed by
reduction with ascorbic acid has an absorption maximum at 820nm. The amplified
voltage of the phototransistor, which is proportional to the transmittance, was recorded on
a strip chart recorder. The light emitting diode could be switched off to set 0%
transmittance.
The low concentration of reactive silicate in seawater requires a high sensitivity from
the analytical method. Thompson et al. [197] as a consequence of a study of the effect of
reaction time, temperature and type of reducing agent were able to optimise conditions
for obtaining good sensitivity.
In the determination of silicate, a detection limit of 0.5µmol L −1 was achieved with a
relative precision of better than 1% at concentrations above 10µmole L −1 Si .
The analysis rate is 80 per h when a continuous stream of seawater is analysed. Thirty
discrete samples can be analysed in duplicate per h. The refractive index interference was
eliminated. This allowed a slow detection limit to be obtained for a wide range of
salinities. A lower detection limit of 0.1µM Si can be obtained with longer reaction coils
and a lower sampling rate. There is a small salt effect and little interference from
phosphate. The method was tested on an oceanographic cruise from Monterey Bay
(California) to San Francisco Bay).
The method using stannous chloride was tested aboard the R/V Cayuse on a cruise
from Monterey Bay to San Francisco Bay. The analysis rate was 60 per h. The quasi-
continuous record of the reactive silicate