Field Site.The field site of Glattfelden is located in the lower
Glatt Valley, Switzerland. In this region, the river Glatt
infiltrates over a distance of about 5 km into a quaternary
aquifer composed of layers of gravel and sand containing
very little organic carbon (<0.1%) (15). At Glattfelden, the
river Glatt contains about 20% effluent from several waste-water treatment plants, which are the main sources of
pollutants for the river. The water from the river infiltrates
into the uppermost part of the aquifer with an average flow
velocity of 4-5md-1(23).
The site is equipped with observation wells that allow
sampling at different depths and distances from the river.
The wells are lined with hard PVC pipes (diameter 5 cm).
Figure 1 shows the situation of the site with the wells used
in this study. The aquifer consists of three layers. The top
layer is freshly infiltrated river water, the middle layer is water
that was infiltrated upstream of the test site, and the lower
layer consists of groundwater that is not affected by infiltra-tion.
Sample Collection.Samples were collected from the river
and from the groundwater wells in March and July 1995.
Samples from the river were taken by a 4900 portable
contaminant sampler from Manning Products. The sampler
was cleaned by pumping for at least 20 cycles 0.01 M HNO3
through the tube. The bottles were soaked for 48 h in 0.01
M HNO3and rinsed with Nanopure water. The stainless steel
end of the tube was removed, and the end was fixed on a
stone to hold the inflow under water. Prior to the sampling,
river water was pumped through the tube for at least five
cycles. For the sampling of the Glatt, every 15 min a sample
of 150 mL was taken, and six samples were collected together
in one bottle (1.5-h composite sample). A total of 39 samples
was taken. Groundwater samples from well GW1 were taken
by a second sampler (same type as for the river). The
autosampler tube was introduced together with a submerged
pump (Whale) into the well. The pump was operated with
a low flow (0.5-1 L/min) during the whole sampling. This
lead to a steady flow of fresh water toward the well. The
autosampler took a sample every 30 min, and six samples
were collected in a bottle (3-h composite sample). A total of
57 samples was taken.
Samples from the wells GW2, GW3, and GW4 and the deep
groundwater (deepGW) were taken by submerged pumps
(Whale). The pumps were cleaned in the lab by pumping
0.01 M HNO3 for at least 30 min. Prior to the sampling,
groundwater was pumped during 20-30 min to remove old
water from the pipe. Samples for metal concentration
determination by ICP-MS were taken directly at the outflow
of the pump and filtered (0.45ím, Acrodisc, Gelman Science)
on the site into precleaned PE tubes (with added HNO3 to
give a final concentration of 0.1 M). The filters had been
soaked for a few days with 0.01 M HNO3. The first 10-20 mL
of the filtrate was discarded. Twelve samples from GW2 and
GW3 were taken, four from GW4 and three from deepGW. pH
and temperature were measured in the field. The samples
for EDTA and metal speciation were transported to the lab
every 24 h, under exclusion of light to avoid photolysis of
Fe(III)-EDTA.
Analyses. Major and trace metal concentrations were
determined by ICP-MS. Blanks with the field procedure for
Zn, Cu, and Ni were 210
-8, 310-9 , and 3 10
-9 M, respectively. The blank for Pb was<210
-10 M. Fe was measured by graphite furnace AAS, alkalinity was measured
by titration with acid, O2was measured by Winkler titration,
and chloride was measured by argentometry. NTA was
measured by GC after derivatization to the propylester; DOC
was measured with a Shimadzu TOC-analyzer.
EDTA and Fe-EDTA were measured as described in
Nowack et al. (24). One aliquot of the sample was analyzed
for total EDTA; another aliquot was irradiated for 10 min
under a mercury lamp (850 Wm
-2) to degrade the Fe-EDTA.
The samples from July were exposed for2htothefull sun
if possible. This time is sufficient to photolyse all the Fe-(III)-EDTA (7). The other EDTA species are not degraded by
light. Both aliquots were analyzed for EDTA. Fe-EDTA was
calculated by the difference from total EDTA and photostable
EDTA.
Determination of EDTA was performed by HPLC. All EDTA
species were transformed to Fe(III)-EDTA by heating for 3
h with Fe(NO3)3at pH 3.3. The complex is separated with
tetrabutylammonium bromide as counterion on a Lichrocard
250-4 C18 column with formate buffer as the eluent (pH 3.3,
8% acetonitrile) and is detected by UV at 258 nm.
Ni-EDTA was measured semiquantitatively as described
in Nowack et al. (24). The amount of Ni-EDTA is calculated
by the difference of total photostable EDTA and “fast reacting”
EDTA. Fast reacting EDTA is determined as follows: The
irradiated sample is measured without addition of Fe(III) and
heating. EDTA or metal-EDTA species flowing through the
HPLC column can react with Fe(III), adsorbed on free SiO2
groups on the column, forming Fe(III)-EDTA. Because Ni-EDTA reacts only slowly with Fe(III), it does not form Fe-(III)-EDTA during elution. Ni-EDTA does not absorb at
258 nm in contrast to Fe(III)-EDTA and is not detected.
Speciation of Cu and Zn. Cu speciation has been
determined by catechol ligand exchange and DPCSV (dif-ferential pulse cathodic stripping voltammetry), and Zn
speciation has been determined by EDTA ligand exchange
and DPASV (differential pulse anodic stripping voltammetry),
as described elsewhere in detail (25-27). DPCSV and DPASV
measurements for the speciation of Cu and Zn were carried
out within 1 week of the sample collection. The filtered
samples were stored in the dark at 4°C until use.
Briefly, the method for Cu speciation is based on ligand
exchange of added catechol with natural ligands, which are
bound to Cu; the Cu catechol complexes formed are
determined specifically by DPCSV. [Cu
2+] and the complex-ation parameters are determined from equilibrium calcula-tions with the added catechol. All the stability constants were
taken from Martell and Smith (28), except the ones of the
natural ligands. DPCSV sensitivity had to be calibrated for
each individual water sample from the portion of the titration
curve at high concentrations of Cu 2+. To obtain a Cu titration
curve of a water sample, we spiked a series of subsamples
with different Cu concentrations at buffered pH 7.8-8.0. The
series was allowed to equilibrate at 20(2°C overnight. The
next day, DPCSV was performed with a hanging mercury drop
electrode, an Ag/AgCl reference, and a graphite counter
electrode held in a Metrohm 647 VA stand combined with a
646 VA processor. Catechol in optimal concentrations (0.2-1
mM for the groundwater) was added to the samples during
DPCSV measurements.
The direct results of these Cu titrations are values of [Cu
2+] and Ki[Li], a complexing coefficient that corresponds to the
product of the stability constants and natural ligand con-centrations. Although natural waters contain a wide range
of different ligands with different stability constants, for the
sake of comparison, a two-ligand model was generally used
to estimate conditional stability constants Ki and total ligand
concentrations T[Li], using the FITEQL program (29).
Zinc speciation was calculated from the labile Zn measured
by DPASV at different concentrations of added EDTA, which
has exchanged with natural ligands for Zn. [Zn-EDTA] is
nonlabile and is measured as the difference between initially
labile Zn and measured labile Zn after the addition of EDTA.
In order to calibrate DPASV sensitivity of labile Zn, we also
titrated water samples with standard zinc ion solution. DPASV
measurements of labile Zn were performed with the same
apparatus as for DPCSV of Cu determination.
When the samples are titrated with EDTA in the same
concentration range as initially labile Zn (10-200 nM), the
added EDTA competes with inorganic and labile organic
complexes for Zn. Therefore, complexing coefficients of weak
(labile) organic complexes are obtained from the measured
labile Zn concentration in the presence of EDTA. Using this
coefficient, the concentrations of free zinc ions and inorganic
and weak organic complexes in original water are evaluated
from the mass balance of initially labile Zn. The difference
between total dissolved Zn and labile Zn is considered as
strong organic complexes. Assuming that Zn competes with
Cu for the same strong ligands, we obtained stability constants
of Zn complexes with strong ligands based on exchange
constants estimated from total dissolved and free ion
concentrations of Cu and Zn, and labile Zn concentrations
in the same samples (30).
การแปล กรุณารอสักครู่..