Full text of DOI:10.1002/etc.5620210905

Environmental Toxicology and Chemistry, Vol. 21, No. 9, pp. 1788–1795, 2002
Printed in the USA
0730-7268/02 $9.00
ϩ .00
USE OF ANODIC STRIPPING VOLTAMMETRY IN PREDICTING TOXICITY OF COPPER
IN RIVER WATER
Z
IJIAN WANG,* SHENGBIAO HUANG, and QING LIU
State Key Laboratory of Environmental Aquatic Chemistry, Research Center for Eco-Environmental Sciences, Shuangqing Road 18,
Haidian District, Beijing, 100085, China
(
Received 30 July 2001; Accepted 4 March 2002)
Abstract—The labile concentration and toxicity of Cu as influenced by alkalinity and different concentrations of ethylenediami-
netetraacetic acid (EDTA) and naturally derived fulvic acid (FA) were determined by bioassays carried out in the culture media
for Daphnia magna (D. magna). The labile concentration of Cu was obtained by differential pulse anodic stripping voltammetry
with a double-acidification method (DAM-DPASV). Changes in water alkalinity did not affect the labile concentration of Cu, but
increase in alkalinity did reduce the mortality of D. magna. In the presence of EDTA and FA, both labile concentration of Cu and
mortality were reduced. By excluding Cu-carbonate complexes from the labile concentration, a bioavailable concentration of Cu
([Cu*]) was obtained and was used to predict the acute toxicity of Cu on D. magna. For natural waters, the labile concentration
of Cu was measured by DAM-DPASV, and [Cu*] was calculated using MINTEQ A2 software (developed by the U.S. Environmental
Protection Agency) based on the anion composition of waters. This procedure was tested for waters and sediment elutriates sampled
from the Le An River (Jiangxi Province, China) that were severely polluted by the discharges from a copper mine. The results
showed that [Cu*] was a good indicator for Cu toxicity and could be used under field conditions.
Keywords—Copper speciation
Acute toxicity
Daphnia magna
Natural waters
INTRODUCTION
time that may improve the model’s predictability and our un-
derstanding of its strengths or limitations [10]. For example,
in testing the model’s ability to predict Cu toxicity, it was
assumed that the dissolved organic carbon (DOC) was 10%
(w/w) humic acid and that the conditional stability constant
of Cu with humic acid was not considered to be a site-specific
parameter [10].
Bioavailability and toxicity of trace metals to aquatic or-
ganisms depend on the physical and chemical speciation of
the metals [1]. Therefore, the speciation of a metal, rather than
its total concentration, is the key for understanding its effect
on the biota [2]. Previous studies on several metals and va-
rieties of aquatic organisms have demonstrated that the re-
sponse of organisms to metals could be modeled by using their
free-ion activity [3–5]. To rationalize these experimental re-
sults and to explain what was perceived as ‘‘the universal
importance of free-ion activities in determining the biological
effect of all cationic trace metal,’’ Morel [5] formulated the
free-ion activity model (FIAM) for metal-organism interaction.
The FIAM has generally been interpreted to imply that a con-
stant degree of biological effect will occur at a constant chem-
ical activity of the free ion of a metal for divalent transition-
metal cations. These studies suggest that the concentration of
free-ionic metal is the key factor in determining the toxicity.
It might be tempting to conclude that the free ion alone is
responsible for the observed toxicities, but this does not appear
to be supported by the previously published work [6]
Free-ion activity does not appear to be a good predictor of
toxicity under certain water-quality conditions (e.g., water
hardness), and it does not explain competition of the metal
ion in question with other cations [7]. The formation of a
biotic-ligand model (BLM), a recent surface-interaction model
of metal binding to fish gill, simultaneously incorporates metal
speciation with its toxicity and competition between the metal
ion under study and other cations [8,9]. Thus, adopting the
BLM model could help to predict acute toxicity of Cu to fish
on a mechanistic basis and under various conditions of water
hardness. However, several factors should be investigated over
Anodic stripping voltammetry (ASV) is a powerful tech-
nique for the study of trace element speciation, which can be
used to provide speciation information regarding the labile/
inert discrimination [11]. Fulvic acid (FA) binds with Cu,
thereby making it no longer labile in ASV measurements.
When Cu is bound to FA in natural aquatic environments, it
is no longer available for biota. This procedure thus provides
a good approximation of labile metal speciation. This leads to
the deposition process of labile metal on the surface of an
electrode approximating as closely as possible the kinetics of
its uptake processes by the cells of organisms [12,13]. The
correlation between ASV-labile concentrations of Cu and tox-
icity derived from bioassay may, therefore, be useful in pre-
dicting its toxicity. However, some aquatic components, such
as FA, can affect the stripping current not only through af-
fecting aqueous Cu species but also through adsorbing organic
components on the surface of an electrode (e.g., mercury drop).
To avoid interference by organic components on stripping cur-
rent, a double-acidification process is recommended [14]. By
assuming that strong organic complexes of Cu are neither la-
bile nor bioavailable, leading to acute toxicity [1], and that
only part of inorganic complexes are bioavailable or toxic, the
acute toxicity of Cu in natural water can be predicted. This
prediction will be based on measurements obtained by differ-
ential pulse ASV with the double-acidification method (DAM-
DPASV) and calculated by a MINTEQ A2 software (developed
by the U.S. Environmental Protection Agency) and will ex-
clude Cu-carbonate complexes from labile speciation of Cu.
* To whom correspondence may be addressed
(wangzj@mail.rcees.ac.cn).
1788

Predicting toxicity of copper in river water
Environ. Toxicol. Chem. 21, 2002
1789
Fig. 2. Toxicity of Cu to Daphnia magna as a function of alkalinity
(expressed as NaHCO3), fulvic acid, and dissolved organic carbon
(DOC). The median lethal concentrations (LC50s) are expressed in
terms of total Cu concentrations.
␮
m filter (Bedford, MA, USA) and collected in a precleaned
polyethylene bottle. The samples were stored at 4 C and used
Њ
within one week. Sediment samples were collected from the
same locations on the Le An River and at the same time as
the water samples. Sediment samples were stored at 4ЊC in
the dark and used within four weeks. Sediment elutriate was
prepared by extracting wet sediment with four volumes of
distilled water and shaking at 200 rpm for 12 h, followed by
isolation of the aqueous fraction by centrifugation at 2,500
g
for 30 min at 4 C. After sampling, concentration of DOC was
Њ
measured by high-temperature combustion method on a total
organic carbon analyzer (TOC; Phoenix 8000; Tekman Do-
krmann, Cincinnati, OH, USA). Major anions were measured
by ion chromatography (Series 4500i, Dionex, San Francisco,
CA, USA). The FA was extracted and separated from Le An
River sediment and purified according to the method developed
by Wang [18].
Fig. 1. Median lethal concentrations (LC50s) expressed as total Cu
concentrations (
ࡗ), free Cu concentrations (᭡), and bioavailable Cu
Stock solution of Cu was prepared in double-distilled, de-
ionized water (dddH2O) from analytical Cu sulfate and acid-
(
ⅷ
) as a function of ligand concentrations. Three different ligands
were used: (a) NaHCO3, (b) ethylenediaminetetraacetic acid (EDTA),
and (c) fulvic acid (FA) expressed as dissolved organic carbon (DOC).
Ϫ
ified with HNO3 (10 4 mol/L). The test solutions were prepared
by diluting the stock solution with reconstituted water con-
In the present study, we have developed the approach and a
model to predict the acute toxicity of Cu on Daphnia magna
.
MATERIALS AND METHODS
Sampling and test solution preparations
Nine samples were taken from the Le An River. The sam-
pling sites were chosen along the downstream from Haikou to
Panlong Temple. Site L04 was 20 km downstream from site
L01, where the largest open-cast Cu mine in China (Dexing
Copper mine) is located and whose discharges flow into the
Le An River at the convergence with the Dawu River. Site
L07 was 50 km downstream and at the convergence with the
Jihui River (39 km long), which is another metal-polluted river
with several small sulfide mines and smelters. Site L09 was
230 km downstream and in the delta area of Poyang Lake, the
largest freshwater lake in China. The metal pollution and the
ecological consequences in the Le An River have been reported
previously [15–17].
Fig. 3. Relationship between toxic probit (TP) and logarithmic con-
centration of bioavailable Cu ([Cu*]). Toxic data (
lected from Figures 1 and 2.
n ϭ 28) were col-
A 20-L water sample was filtered through a Millipore 0.45-

1790
Environ. Toxicol. Chem. 21, 2002
Z. Wang et al.
taining NaHCO3 (48 mg/L), MgSO4 (30 mg/L), CaSO4·2H2O
(30 mg/L), and KCl (2 mg/L) before use. The test solutions
of different alkalinity (expressed as [NaHCO3]) or containing
different amounts of ethylenediaminetetraacetic acid (EDTA),
and/or FA were prepared by adding appropriate amounts of
stock solutions of NaHCO3, Na-EDTA, FA, and Cu to the
reconstituted water. Concentration of FA (mg-C/L) was mea-
sured after filtering through 0.45-
TOC was measured. All test solutions were buffered with N-
[2-hydroxyethyl]-piperazine-N -[2-ethanesulfonic acid] (HE-
␮m membrane filters, and
Ј
PES; Sino-American Biotec, Beijing, China), and the pH was
adjusted with NaOH (0.5 mol/L) to 6.8 before use. In the test
solutions, the concentration of HEPES was 0.006 mol/L. It
has been demonstrated that HEPES, when present in concen-
trations of less than 10 mmol/L, forms weak complexes with
copper and does not interfere with copper speciation in solution
[19]. The test solution was allowed to equilibrate for at least
24 h at a constant temperature of 23 mean
Ϯ SEM 1ЊC before
DAM-DPASV measurements and bioassay.
The DAM-DPASV measurements were performed to obtain
the labile concentrations of Cu. Bioassays using D. magna
were performed to obtain the acute toxicity in the test solutions
and in river water samples and sediment elutriates.
Chemical measurements
The DAM-DPASV measurements were performed with a
Model 303A static mercury drop electrode assembly plus a
Model 303 stirrer coupled with a polarographic analyzer (Mod-
el 263; EG&G, Princeton, NJ, USA). The mercury electrode
used as the working electrode was washed with 10% nitric
acid and dddH2O twice. To avoid metal contamination, all
plastic and glassware were washed with 1:1 (w/w) nitric acid
followed by dddH2O. In all experiments, sodium nitrate (0.1
mol/L) was used to maintain a constant ionic strength of the
solutions. The specific setting used for the differential pulse
mode was as follows: scan rate, 10 mV/s; drop size, medium;
pulse height, 50 mV; step time, 0.3 s; scan width, 2 mV; initial
potential,
Ϫ1.0 V; and final potential, ϩ0.15 V versus Ag-
AgCl. A rotation rate of 4,000 rpm was used. During each step
of titration, the electrode surface was wiped clean with plain
filter paper and rinsed with dddH2O. Measurements were car-
ried out in 10 ml of working solution in a cell at 23
Then, the samples were degassed with oxygen-free N2 for 15
min. Samples were plated at 1.0 V for 300 s, followed by
Ϯ 1ЊC.
Ϫ
a 15-s quiescent period before the film was stripped by scan-
ning the potential in the positive direction, using the differ-
ential pulse mode, to a final potential of
ϩ0.15 V. The double-
acidification procedure involved a deposition at pH 6.8 for 20
s; then, before completion of the 300-s deposition and with
the cell circuit uninterrupted, 50 ␮l of HNO3 (5 mol/L) were
added through the standard addition port. After recording the
stripping voltammogram, a new run was initiated for the same
solution; this time, the pH during both deposition and stripping
runs was adjusted to 2.0. The percentage of labile Cu in the
total Cu could be obtained as
Labile-Cu (%)
ϭ 100[1 Ϫ (c1 Ϫ c2)/cT]
where c1 is the concentration of Cu measured with deposition
and stripping at pH 2, c2 is the deposition at pH 6.8 and
stripping at pH 2, and cT is the total Cu concentration. The
calibration was carried out at Cu concentrations ranging from
0.5 to 20
␮g/L in 10 increments in reconstituted water (NaNO3,
0.1 mol/L; HEPES, 0.06 mol/L). For each sample, triplicates

Predicting toxicity of copper in river water
Environ. Toxicol. Chem. 21, 2002
1791
measurements were taken, and the labile concentration of Cu
was calculated from the calibration curve. Precision for mea-
surements in DAM-DPASV was
viation for test solution containing Cu at 2
Ϯ
9% relative standard de-
g/L ( 6).
␮
n ϭ
The total concentrations of dissolved Cu, Pb, and Zn in
water samples were measured in acidified filtrates by a graphite
furnace atomic absorption spectrometer (Model 3100; Perkin-
Elmer, Norwalk, CT, USA). Concentrations of Cu, Pb, and Zn
in sediments were measured by graphite furnace atomic ab-
sorption spectrometer after microwave digestion [20].
Acute toxicity bioassays
Bioassays for the acute toxicity of Cu were performed on
D. magna. The static bioassay procedure described by the U.S.
Environmental Protection Agency guideline for test chemicals
was adapted with triplicate samples [21]. The test species D.
magna was obtained from the Institute of Hydrobiology (Wu-
han, China) and adapted in the laboratory for a period of one
month. Healthy D. magna, not more than 24 h old at the
beginning of the test, did not have to be fed during the test.
Ten animals were used for each concentration in the 50 ml of
test solution in a 100-ml beaker. The test temperature was 23
Ϯ 1ЊC. For the validity of the test, no more than 10% im-
mobility (or mortality) of the animals was accepted as the
control. The dissolved oxygen concentration at the end of the
test was maintained at 60% or greater air saturation.
Data on 96-h median lethal concentration (LC50) deter-
mined by the immobilization (or mortality) tests were used as
acute toxicity data. The LC50s and associated 90% confidence
intervals for each replicate toxicity bioassay were calculated
by trimmed Spearman-Karber analysis [22]. Toxic probit (TP)
was transformed from mortality according to the method of
probit unit [23]. Regression analysis was performed using Sig-
maplot௡ 4.0 (SPSS, Chicago, IL, USA).
Cu speciation
The concentrations of different Cu species present in test
solutions, other than that containing FA, could be calculated
from the spiked concentrations of Cu, the composition of the
reconstituted water, and the known stability constants of Cu
complexes using MINTEQ A2 (Ver 4.1). For water samples
and test solution containing FA, inorganic Cu speciation could
be calculated from the labile concentration of Cu, assuming
that Cu associated with strong organic ligands was nonlabile
in DPASV measurements and that Cu associated with weak
organic ligands was negligible in comparison to inorganic
complexes of similar stability constants. This hypothesis is
valid because DPASV has long been used to obtain the ap-
parent stability constants of Cu with FA, both by metal titration
and by proton titration [24].
RESULTS AND DISCUSSION
Deriving the expression for the relationship between Cu
speciation and acute toxicity
Influences of alkalinity and different concentrations of
EDTA and FA derived from the Le An River on the acute
toxicity of Cu to D. magna are shown in Figure 1. Addition
of NaHCO3, EDTA, and FA to test solutions inhibited signif-
icantly the Cu toxicity. Similar observations on the inhibition
effects of Cu to different organisms have been reported in
previous studies [25].
The toxicity of Cu to D. magna increased with spiked con-
centrations of Cu and decreased with spiked concentrations of

1792
Environ. Toxicol. Chem. 21, 2002
Z. Wang et al.
NaHCO3, EDTA, and FA. This suggests that the complexation
of Cu with HCO3 , EDTA, and FA reduced the free-ion con-
LC50s could be expressed as the total spiked concentration of
Ϫ
ϩ
Cu ([CuT]), concentration of free-ion Cu ([Cu2 ]), and [Cu*].
centration of Cu in the aqueous phase according to the FIAM
model, thereby reducing the potential of Cu uptake by the
organisms. Thus, Cu, when present in the complex form with
HCO3 , EDTA, and FA, could be considered as nonbioavail-
able or nontoxic.
The expressions are illustrated in Figure 1. The LC50 ex-
pressed as [CuT] is a function of alkalinity and organic com-
ϩ
ponents. However, LC50s expressed as either [Cu2 ] or [Cu*]
Ϫ
could give a constant value, independent of variation in al-
ϩ
kalinity and organic components. In fact, [Cu2 ] was propor-
When NaHCO3 was spiked into the reconstituted water (Fig.
tional to [Cu*] under constant alkalinity. The LC50 could be
ϩ
ϩ
ϩ
1a), the dominant species of Cu were Cu2 , CuOH ,
predicted by FIAM or BLM models regarding [Cu2 ]. How-
ϩ
ϩ
Cu(OH)2aq, CuCO3aq, CuHCO3
,
CuCl , CuCl2, and
ever, the difficulty in using FIAM or BLM models to predict
Cu toxicity in natural water is due to determination of the site-
specific apparent complexation constant for Cu-DOC or Cu-
FA complexes. Therefore, we could use [Cu*] to predict the
acute toxicity of Cu on D. magna. By excluding Cu-carbonate
complexes from the labile concentration, [Cu*] was obtained
and used to predict the acute toxicity of Cu on D. magna. By
performing a DAM-DPASV measurement on natural water
sample, the labile concentration of Cu was obtained, which
consists mainly of inorganic Cu species. The value of [Cu*]
could be obtained by exclusion of Cu-carbonate complexes
using MINTEQ A2 calculation based on measured composi-
tion of anions in the waters.
CuSO4aq. Under constant pH, the decrease in toxicity was due
ϩ
to the increase of CuCO3aq and CuHCO3 . The linear regres-
sion between LC50 (
␮g/L) and concentration of NaHCO3 (mg/
L) could be expressed as
LC50 ( g/L) 0.12[NaHCO3 (mg/L)]
r2
0.9796,
When EDTA was spiked into the reconstituted water (Fig.
␮
ϭ
ϩ
12.22
(
ϭ
n
ϭ
6)
(1)
ϩ
ϩ
1b), the dominant species of Cu were Cu2 , CuOH ,
ϩ
ϩ
Cu(OH)2aq, CuCO3aq, CuHCO3 , CuCl , CuCl2, CuSO4aq,
and Cu-EDTA complexes. When the mole concentration of
EDTA was greater than that of Cu, Cu speciation other than
Cu-EDTA could be negligible. Under constant pH and alka-
linity, the decrease in toxicity was likely due to the increase
in Cu-EDTA concentration. The linear regression between
The available toxicity data (except in cases where mortality
was
Ͻ10% or Ͼ90%) obtained from laboratory experiments
with different concentrations of NaHCO3, EDTA, and FA were
transformed into TP, and the relationship between TP and
[Cu*] was obtained by linear regression analysis (Fig. 3). The
regression equation could be used in predicting Cu toxicity in
natural water, where mortality can hardly be exactly 50%:
LC50 (␮g/L) and concentration of EDTA (mg/L) could be
expressed as
LC50 (
␮
g/L)
ϭ
234.63[EDTA (mg/L)]
r2
0.9926,
When FA was spiked into the reconstituted water (Fig. 1c),
ϩ
12.85
TP
ϭ
2.86 log[Cu*] (
␮
g/L)
ϩ
1.75
28)
(
ϭ
n
ϭ
6)
(2)
(
r2
ϭ
0.4695,
n
ϭ
(5)
the dominant species of Cu were Cu2 , CuOH , Cu(OH)2aq,
ϩ
ϩ
ϩ
ϩ
CuCO3aq, CuHCO3 , CuCl , CuCl2, CuSO4aq, and Cu-FA
complexes. The decrease in toxicity was due to the increase
in Cu-FA concentration. The linear regression between LC50
Measuring water-quality parameters and acute toxicity
The pH, concentrations of cations and anions, concentra-
tions of heavy metal pollutants, and concentration of DOC
(mg-C/L) in Le An River waters are shown in Tables 1 and
(␮g/L) and concentration of FA (mg/L) could be expressed as
LC50 (
␮
g/L)
ϭ
16.52[FA (mg/L)]
ϩ
17.81
2. The copper concentration was highest at site L04 (73.2
␮g/
(
r2
ϭ
0.9919,
n
ϭ
4)
(3)
L) but lower at site L06 (16.6 g/L). Previous investigation
␮
indicated that, among heavy metals, aqueous concentrations
of Cu, Pb, and Zn during the dry season [15] and of Cu during
the rainy season [26] exceeded the environmental quality stan-
dards for surface water (GB 3838-88, China).
The regional backgrounds of metals in Le An River sedi-
ments were 34, 117, and 45 mg/kg for Pb, Zn, and Cu, re-
spectively. Accordingly, the Le An River sediments were
heavily polluted by heavy metals (Table 3). Metals that had
sediment concentrations significantly greater than their back-
ground levels were Zn, Pb, and especially, Cu. At site L04,
the concentration of Cu reached 2,878 mg/kg (the highest lev-
el), and at site L07, the concentration of Cu was 1,012 mg/
kg.
In most of the natural freshwaters, Cu was mainly bound
to organic substances [12]. Because of the presence of naturally
derived organic matter, the DPASV labile concentrations of
Cu could not be directly measured in water samples from
several other locations (data not shown). The DAM was ap-
plied to eliminate the effects of complexing agents on the ASV
process and of surfactants on both deposition and stripping
processes in natural water [14]. The labile concentration of Cu
was obtained from the measured stripping current in DAM-
DPASV. From the labile concentration, the bioavailable con-
centrations of Cu could be obtained.
For test solutions in which both NaHCO3 and FA were
present, LC50s under different concentrations of NaHCO3 and
FA were obtained, and the results are shown in Figure 2. By
multiregression analysis, the relationship between LC50 and
concentrations of NaHCO3 and FA could be expressed as
LC50 (
␮
g/L)
ϭ
17.55[FA (mg/L)]
0.21[NaHCO3 (mg/L)]
ϩ
ϩ 11.79 (4)
The intercept in Equation 1 represents the bioavailable con-
centration of Cu that causes the LC50 for D. magna, which
should be 12.22
␮g/L. Equations 2 and 3 indicate that the
LC50s for Cu in reconstituted water containing 48 mg/L of
NaHCO3 in absence of EDTA or FA should be 12.85 and 17.81
␮
g/L, respectively. If the influence of carbonate is excluded,
the bioavailable concentration of Cu that causes the LC50 for
D. magna should be 7.09 and 12.05 g/L from Equations 2
␮
and 3, respectively. In Equation 4, the LC50 or bioavailable
concentration of Cu in the absence of alkalinity and FA should
be 11.79 ␮g/L, which was quite similar to those in Equations
1 and 3. In addition, from Equation 4, the influence of FA on
toxicity is 83-fold greater than that of alkalinity.
In our model, [Cu*] in test solution consists of inorganic
ϩ
species other than CuCO3aq and CuHCO3 . By definition,

Predicting toxicity of copper in river water
Environ. Toxicol. Chem. 21, 2002
1793
Table 3. Concentrations (mg/kg dry wt) of Cu, Pb, and Zn in sediments from different sampling sites in the Le An River (Jiangxi Province, China)
Sampling site
Element
L01
L02
L04
L05
L06
L07
L08
L09
Cu
Pb
Zn
36
44
117
215
81
226
2,878
42
221
2,173
44
201
1,788
29
202
1,012
208
878
733
94
664
523
68
451
Results of the acute toxicity bioassays for the Le An River
waters are shown in Table 4. The river water was slightly toxic
at site L04 (mortality, 6.7%) and extremely toxic at site L06
(mortality, 100%). Mortality predicted by the bioavailable con-
centration of Cu using Equation 5 was quite consistent with
that measured using bioassays for Le An River waters (Table
4). At sites L01 to L05 and L07 to L09, toxicities could be
neither measured nor predicted, because the mortality of less
than 10% was considered to be insignificant.
Because the highest concentration of Cu was observed at
site L04, not at site L06, some other factors should affect the
toxicity to D. magna. Among these factors, concentration of
DOC was the highest at site L04 (3 mg/L), whereas it was
undetectable at site L06. Consequently, the highest labile con-
soluble and bioavailable concentrations of Cu occurred at site
L07, and both measured and predicted toxicities indicated a
mortality of 100% at this site. A mortality of 87% was also
observed in sediment elutriate from site L06, where only mod-
erate mortality (19%) was predicted based on the measured
[Cu*].
The concentration of Cu in sediment from site L07 was
lower than that from site L06 (Table 3), whereas the total
dissolved and labile concentration of Cu from site L07 was
higher. This indicates that Cu bound to sediment at site L07
was more mobile and bioavailable and that the sediment tox-
icity could hardly be predicted by its total metal concentration.
This suggests that both mobility and bioavailability be taken
into account.
centration of Cu (14.7
␮
g/L) was observed at site L06, not at
To obtain the LC50 of sediment elutriate from site L07,
elutriate was sequentially diluted for bioassay. The results of
toxic bioassay and DAM-DPASV measurement for 50% se-
quential dilution of sediment elutriate from site L07 are shown
in Table 6. The data show both measured and predicted mor-
talities of D. magna with regard to [Cu*]. As can be seen, the
mortality of D. magna in the diluted sediment elutriate could
be well predicted by assuming that Cu was the toxicant.
From Table 6, it is quite interesting to note that the LC50
site L04. The influence of DOC appears to be responsible for
the reduced toxicity at site L04. A mortality of 100% occurred
at site L06, where a mortality of 45% was predicted based on
the toxicity of Cu. The observed toxicity at site L06 could
also be caused by the joint toxic effect of Cu and Zn. The
toxic equivalent of Zn at site L06 was 2.6, as calculated from
the measured concentration of Zn divided by its LC50 (540
␮
ϭ
g/L in reconstituted water). The toxic equivalent of Cu (LC50
13.6 g/L in reconstituted water) was 1.98, which was
␮
for [Cu*] in sediment elutriate was approximately 11.64
L, which is quite close to the laboratory-observed [Cu*] that
caused the LC50 in reconstituted water (13.6 g/L). This in-
dicates that the LC50 expressed as [Cu*] to D. magna in
freshwater was approximately 12 g/L, independent of water-
quality variations. Similarly, the lowest-observed-effective
concentration should be 2.3 g/L.
␮g/
similar to that of Zn. The influences of other metals on toxicity
could be negligible.
␮
Sediments from different sampling sites were used to obtain
sediment elutriates, and elutriates were used for acute bioassay
and DAM-DPASV measurements. The experiment could sim-
ulate the sediment toxicity under aerobic conditions. The total
dissolved and labile concentration of Cu in sediment elutriate
from the Le An River and the observed toxicity to D. magna
are shown in Table 5. The toxicity to D. magna was predicted
from the [Cu*] (Table 5). As shown in Table 5, both the highest
␮
␮
CONCLUSIONS
Alkalinity, EDTA, and FA have significant influences on the
acute toxicity of Cu to D. magna. These influences could be
Table 4. Concentrations of different Cu species and predicted/observed mortality on Daphnia magna in the Le An River waters
(Jiangxi Province, China)
Sampling
sites
[CuT]
g/L)a
[Culabile
]
[Cu*]
g/L)c
Measured
Predicted
(
␮
(
␮
g/L)b
(
␮
mortality (%)d
mortality (%)e
L01
L02
L04
L05
L06
L07
L08
L09
5.3
5.3
73.2
5.1
16.6
6.8
6.8
NDf
2.6
ND
ND
ND
ND
0.6
11.5
3.8
ND
ND
Ϯ
ND
Ϯ
ND
ND
ND
0
0
0
5
45
8
Ϯ
0.7
6.7
3.5
18
1.3
Ϯ
Ϯ
Ϯ
Ϯ
Ϯ
1.3
2.5
3.2
1.9
0.6
14.7
5.1
100
5.8
4.5
3.2
3
4
5.1
3.8
a The concentration of total Cu ([CuT]) was measured by graphite furnace atomic absorption spectrometry.
b The labile concentration of Cu ([Culabile]) was measured by a differential pulse anodic stripping voltammetry with a double-acidification method.
c The bioavailable concentration of Cu ([Cu*]) was calculated from [Culabile] and concentration of anions in water using MINTEQ A2 (developed
by the U. S. Environmental Protection Agency).
d The mortality was measured on D. magna
e The mortality was predicted using Equation 5 (see text).
f ND
not detected.
(n ϭ 3).
ϭ

1794
Environ. Toxicol. Chem. 21, 2002
Z. Wang et al.
Table 5. Concentrations of total Cu ([CuT]) and bioavailable Cu ([Cu*]) and predicted/measured
mortality on Daphnia magna in sediment elutriates down the Le An River (Jiangxi Province, China)
Sampling
sites
Predicted
Measured
[CuT] (
␮
g/L)a
[Cu*] (
ND
␮
g/L)b
mortality (%)c
mortality (%)d
L01
L04
L06
L07
L08
L09
NDe
3.8
7.7
249.8
7.1
0
1
19
100
8
ND
2.6
6.4
Ϯ
Ϯ
Ϯ
Ϯ
Ϯ
1.3
1.3
19.2
1.3
3.2
7
87
100
27
7
Ϯ
Ϯ
Ϯ
Ϯ
Ϯ
3
12
21
7
152.9
4.5
9.6
5.8
13
5
a Measured by graphite furnace atomic absorption spectrometry.
b Calculated from [Culabile] and measured alkalinity in elutriate (data not shown).
c The mortality was measured on D. magna
3).
d The mortality was predicted using Equation 5 (see text).
e ND
not detected.
(n ϭ
ϭ
REFERENCES
excluded by defining the bioavailable concentration of Cu, or
[Cu*], which could be obtained by a DAM-DPASV measure-
ment and a calibration to exclude Cu-carbonate complexes from
labile Cu species. The relationship between [Cu*] and toxicity
to D. magna could be formulated through laboratory simulation
experiments in the presence of NaHCO3, EDTA, and FA. This
relationship could be then be applied to predict the acute toxicity
of Cu in the natural water and sediment elutriates from a river
that has been severely polluted by a copper mine. The bio-
available concentration of Cu that causes 50% of mortality on
D. magna, both in the laboratory experiments and under field
1. Allen HE, Hansen DJ. 1996. The importance of trace metal spe-
ciation to water quality criteria. Water Environ Res 68:283–295.
2. Allen HE. 1993. The significance of trace speciation for water,
sediment and soil quality standards. Sci Total Environ 134:23–
34.
3. Campbell PGC. 1995. Interactions between trace metals and
aquatic organisms. In Tessier A, Turner DR, eds, Metal Speciation
and Bioavailability in Aquatic System. John Wiley, New York,
NY, USA, pp 45–102.
4. Morel FMM, Hering J. 1993. Principles and Applications of
Aquatic Chemistry. John Wiley, New York, NY, USA.
5. Morel FMM. 1983. Principles of Aquatic Chemistry. John Wiley,
New York, NY, USA. p 301.
6. Erickson RJ, Benoit DA, Mattson VR. 1996. The effects of water
chemistry on the toxicity of copper to fathead minnows. Environ
Toxicol Chem 15:181–193.
7. Playle RC, Gensemer RW, Dixon DG. 1992. Copper accumulation
on gills of fathead minnows: Influence of water hardness, com-
plexation and pH on the gill microenvironment. Environ Toxicol
Chem 11:381–391.
8. U.S. Environmental Protection Agency. 1999. Integrated ap-
proach to assessing the bioavailability and toxicity of metals in
surface waters and sediments. EPA-822-E-99-001. Technical Re-
port. Office of Water, Washington, DC.
conditions, was approximately 12 ␮g/L.
The proposed procedure provides an alternative to the
FIAM model. Furthermore, the bioavailable concentration of
Cu, or [Cu*], was used to replace the free-ion activity in the
prediction of toxicity. The information presented in this paper
provides a means to measure the bioavailable concentration
of Cu in natural water by combining DAM-DPASV measure-
ment and a model calculation to exclude the influences of Cu-
carbonate complexes.
9. Mayer JS, Santore RC, Bobbitt JP. 1999. Binding of nickel and
copper to fish gills predicts toxicity when water hardness varies,
but free ion activity does not. Environ Sci Technol 33:913–916.
10. U.S. Environmental Protection Agency. 2000. Review of the bi-
otic ligand model of the acute toxicity of metals. EPA-SAB-
EPEC-00-006. Final Report. The Ecological Processes and Effects
Committee of the Science Advisory Board, Washington, DC.
11. Florence TM. 1986. Electrochemical approaches to trace element
speciation in waters: A review. Analyst 111:489–505.
12. Morrison GM, Florence TM, Stauber JL. 1990. The effects of
complexing agents and surfactants on the deposition and stripping
processes in DPASV of metals. Electroanalysis 2:9–14.
13. Florence TM, Morrison GM, Stauber JL. 1992. Determination of
trace element speciation and the role of speciation in aquatic
toxicity. Sci Total Environ 125:1–13.
Acknowledgement—This study was supported by the International
Copper Association (TP0602) and partially supported by the Chinese
Academy of Sciences (KZCX2-410). The authors would like to ac-
knowledge the valuable support and useful discussion from Herbert
Allen from the University of Delaware, Newark, Delaware, USA.
Thanks are also due to Shahamat U. Khan of George Mason Uni-
versity, Fairfax, VA, USA, for carefully revising the manuscript.
Table 6. Concentrations of bioavailable Cu ([Cu*]) and predicted/
measured mortality on Daphnia magna in the sequential dilutions of
elutriate from site L07 in Le An River (Jiangxi Province, China)
14. Florence TM. 1992. Trace element speciation by anodic stripping
voltammetry. Analyst 117:551–553.
15. He MC, Wang Z, Tang HX. 1998. The chemical, toxicological
and ecological studies in assessing the heavy metal pollution in
Le An River, China. Water Res 32:510–518.
16. He MC, Wang Z, Tang HX. 1997. Spatial and temporal pattern
of acidity and heavy metals in predicting the potential for eco-
logical impact on the Le An river polluted by acid mine drainage.
Sci Total Environ 206:67–77.
17. Wang Z. 1999. Major environmental and ecotoxicological pro-
cesses of heavy metals in Le An River polluted by discharges
from mining activities. Chin J Environ Sci 11:322–327.
18. Wang Z. 1987. Interaction between heavy metals and humic acid-
quarsi- particle surface complexation model. Acta Scientiae Cir-
cumstantiae (China) 7:313–318.
Dilution
proportion
Predicted
Measured
[Cu*] (␮
g/L)a
mortality (%)b mortality (%)c
1/1
1/2
1/4
1/8
1/16
1/32
152.9
98.6
49.3
23.7
9.6
Ϯ
19.2
14.7
8.3
5.8
4.5
100
100
95
75
44
100
100
100
67
Ϯ
Ϯ
Ϯ
Ϯ
Ϯ
Ϯ
21
12
34
10
8
Ϯ
Ϯ
Ϯ
Ϯ
Ϯ
47
6.4
2.6
18
20
6
a Calculated from labile Cu and measured alkalinity in elutriate (data
not shown).
b The mortality was measured on D. magna
(
n
ϭ
3).
c The mortality was predicted using Equation 5 (see text).

Predicting toxicity of copper in river water
Environ. Toxicol. Chem. 21, 2002
1795
19. Vasconcelos MT, Azenha MA, Lage O, 1996. Electrochemical
evidence of surfactant activity of the Hepes pH buffer which may
have implication on trace metal availability to culture in vitro.
Anal Biochem 241:248–253.
20. Wen XH, Wu LZ, Zhang Y. 1997. Optimized microwave prepa-
ration procedure for the elemental analysis of aquatic sediment.
Fresenius J Anal Chem 357:1111–1115.
in toxicity bioassays. Environ Sci Technol 11:714–719, Correc-
tion 12:417 (1978).
23. Zhou YX, Zhang ZS. 1989. Methods For Aquatic Bioassays.
Agricultural Publishing House, Beijing, China.
24. Wang Z, Stumm W. 1987. Heavy metal complexation by surfaces
and humic acids: A brief discourse on assessment by acidimetric
titration. Neth J Agric Sci 35:231–240.
21. U.S. Environmental Protection Agency. 1975. Methods for acute
toxicity tests with fish, macroinvertebrates, and amphibians. EPA-
660/3-75/009. Washington, DC.
25. Deaver E, Rodgers JH. 1996. Measuring bioavailable copper us-
ing anodic stripping voltammetry. Environ Toxicol Chem 15:
1925–1930.
22. Hamilton MA, Russo RC, Thurston RV. 1977. Trimmed Spear-
man-Karber method for estimating median lethal concentrations
26. Wang Z, Ma M, Du Q. 1994. Toxicity assessment by photobac-
terium phosphoreum. China J Environ Sci 5:159–164.