-------
a
r
o
5
0.01 0.1 1 10 100
Total Metal or SEM (umol/g)
1000
"3
o
0.01 0.1 1 10 100
Interstitial Water Toxic Units
1000
"5
o
25 50
SEM-AVS
Figure 3-18. Percentage mortality of saltwater and freshwater benthic species in 10-day toxicity tests in
spiked sediments (open symbols) and sediments from the field (closed symbols) (silver data from Berry et
al., 1999; all other data modified after Hansen et al., 1996a). Mortality is plotted as a function of (a) the
sum of the concentrations of cadmium, copper, lead, nickel and zinc in Mmoles metal per gram dry weight
sediment; (b) interstitial water toxic units; and (c) SEM/AVS ratio. Species tested include: the
oligochaete Lumbriculus variegatus, polychaetes (Capitella capitata and Neanthes arenaceodentata), the
harpacticoid (Amphiascus tenuiremis), amphipods (Ampelisca abdita and Hyalella aztecd) and the snail
(Helisoma sp.). Data below the SEM detection limit are plotted at SEM/AVS = 0.01. Data below the
detection limit of metals in interstitial water are plotted at IWTU = 0.01.
Draft for SAB 1-70
-------
experiments are very high relative to those often suspected to be of toxicological significance in
field sediments. This has sometimes been interpreted as a limitation of the use of SEM and AVS
to predict metal-induced toxicity. However, the range in AVS hi these sediments spiked with
metals is similar to sediments commonly occurring hi the field. The important point is that even
a sediment with only a moderate concentration of AVS has a considerable capacity for
sequestering metals as a metal sulfide, a form which is not bioavailable (Di Toro et al., 1990).
In contrast, the combined data from all available freshwater and saltwater spiked-sediment
experiments supports the use of IWTU to predict mortality of benthic species hi spiked sediment
toxicity tests (Figure 3-17b). Mortality hi these experiments was sediment independent when
plotted against IWTU. Sediments with IWTUs of < 0.5 were generally not toxic. Of the 96
sediments with IWTU < 0.5, 96.9% were not toxic, while 76.4% of the 89 sediments with
IWTU * 0.5 were toxic (Table 3-2). This close relationship between IWTU and sediment toxicity
hi sediments spiked with metals was also observed hi studies with field sediments contaminated
with metals (See Section 3.2.3 below), sediments spiked with nonionic organic chemicals (Adams
et al., 1985; Di Toro et al., 1991; 1985; Swartz et al., 1990) and field sediments contaminated
with nonionic organic chemicals (Hoke et al., 1994; Swartz et al., 1994).
Draft for SAB 1-71
-------
Table 3-2. Toxicity of sediments from freshwater (FW) and saltwater (SW) field locations,
spiked-sediment tests and combined field and spiked-sediment tests as a function of the difference
between the molar concentrations of SEM and AVS (SEM-AVS), interstitial water toxic units
(IWTUs) and both SEM-AVS and IWTUs (modified from Hansen et al., 1996a).
Study
Type Parameter
Lab-Spike SEM-AVS
(FW&SW)
IWTU
SEM-AVS, IWTU
Field SEM-AVS
(FW&SW)
IWTU
SEM-AVS, IWTU
All SEM-AVS
IWTU
SEM-AVS, IWTU
Percent of Sediments
Value
sO2
>&
< 0.5
^0.5
<. O2, < 0.5
X)3, * 0.5
sO2
X)3
< 0.5
*0.5
* O2, < 0.5
> O3, ;> 0.5
sO2
X)3
< 0.5
*0.5
s O2, < 0.5
> O3, * 0.5
n
101
95
96
89
83
78
57
79
79
53
49
45
158
174
175
142
132
123
Non-toxic1
98.0
26.3
96.9
23.6
97.6
14.1
98.2
59.5
98.7
45.3
100.0
33.3
98.1
42.0
97.7
31.7
98.5
21.1
Toxic1
2.0
73.7
3.1
76.4
2.4
85.9
1.8
40.5
1.3
54.7
0.0
66.7
1.9
58.0
2.3
68.3
1.5
78.9
1 Non-toxic sediments < 24 percent mortality. Toxic sediments > 24 percent mortality.
2 SEM-AVS sO is the same as an SEM/AVS ratio of * 1.0.
3 SEM-AVS >0 is the same as an SEM/AVS ratio of > 1.0.
Draft for SAB 1-72
-------
The interstitial water metal concentrations in all spiked-sediment studies were usually
below the limit of analytical detection in sediments with SEM/AVS ratios below 1.0 (Berry et al.,
1996). Above an SEM/AVS ratio of 1.0, the interstitial metals concentrations increased up to five
orders of magnitude with increasing SEM/AVS ratio. This several order of magnitude increase
in interstitial water metals concentration with only a factor of two or three increase in sediment
concentration is why mortality is most often complete in these sediments, and why the chemistry
of anaerobic sediments controls the toxicity of metals to organisms living in aerobic micro-
habitats. It also explains why the toxicity of different metals in sediments to different species is
so similar. Interstitial water metals were often below or near detection limits when SEM/AVS
ratios were only slightly above 1.0 indicating the presence of other metals binding phases in
sediments.
The combined data from all available saltwater and freshwater spiked sediment
experiments also supports the use of SEM/AVS ratios to predict sediment toxicity to benthic
species in spiked-sediment toxicity tests. All tests yield similar results when mortality is plotted
against SEM/AVS ratio (Figure 3-17c). Mortality in these experiments was sediment independent
when plotted on an SEM/AVS basis. With the combined data, 98.0% of the 101 metals-spiked
sediments with SEM/AVS ratios s 1.0 were not toxic, while 73.7% of the 95 sediments with
SEM/AVS ratios > 1.0 were toxic (Table 3-2).
These overall data show that when both SEM/AVS ratio and IWTU are used,
predictions of which sediments would be toxic were improved. Of the 83 sediments with
SEM/AVS ratios s 1.0 and IWTU < 0.5, 97.6% were not toxic, while 85.9% of the 78
sediments with SEM/AVS ratios > 1.0 and IWTU * 0.5 were toxic (Table 3-2). ( Note: Table
3-2 uses SEM-AVS instead of SEM/AVS ratios. An SEM-AVS of sO is the same as an
SEM/AVS ratio of <. 1.0. An SEM-AVS of > 0 is the same as an SEM/AVS ratio of > 1.0.) '
These results show that SEM/AVS and IWTU are accurate predictors of the absence
of mortality in sediment toxicity tests, however, predictions of which sediments might be toxic
are less accurate. The fact that a significant number of sediments (26.3%) tested had SEM/AVS
ratios of > 1.0, but were not toxic indicates that other binding phases, such as organic carbon
(Mahony et al., 1996), may also control bioavailability in anaerobic sediments. While the
SEM/AVS model of bioavailability accurately predicts which sediments will not be toxic, a model
which utilizes SEM/AVS ratios or (SEM-AVS) (Hansen et al., 1996a) and incorporates other
Draft for SAB 1-73
-------
binding phases might more accurately predict which sediments will be toxic (Di Toro et al., 1987;
Mahony etal., 1996).
Organism behavior may also explain why sediments with SEM/AVS ratios of > 1.0
were not toxic. Many of the sediments which had the highest SEM/AVS ratios in excess of 1.0
that produced little or no mortality were from experiments using the polychaete, Neanthes
arenaceodentata (see Pesch et al., 1995, Figure 8). This appeared to be related, in part, to the
ability of this polychaete to avoid burrowing into the test sediments, thereby limiting its exposure
to the elevated concentrations of metals in the interstitial water and sediments. This same
phenomenon may also explain the low mortality of snails, Heliosoma sp., in freshwater sediments
with high SEM/AVS ratios. These snails are epibenthic and also have the ability to avoid
contaminated sediments (G. Phipps, personal comm.). Increased mortality was always observed
in sediments with SEM/AVS ratios >5.9 in tests with the other five species.
Similarly, a significant number of sediments with * 0.5 IWTUs were not toxic. This
is likely the result of IW ligands which reduce the bioavailability and toxicity of dissolved metals,
sediment avoidance by polychaetes, or snails, or methodological problems in contamination-free
sampling of IW. Ankley et al. (1991) suggested that a toxicity correction for the hardness of the
IW is needed to compare toxicity in IW to that in water-only tests. Absence of a correction for
\
hardness might affect the accuracy of predicting metal-induced sediment toxicity using IWTUs in
freshwater. Further, a significant improvement in the accuracy of metal-induced toxicity
predictions using IWTUs might be achieved if DOC binding in the IW is taken into account.
Green et al. (1993) and Ankley et al. (1991) hypothesized that increased DOC in the IW reduced
the bioavailability of cadmium in sediment exposures, relative to the water-only exposures. Green
et al. (1993) found that the LC50 value for cadmium in an IW exposure without sediment was
more than twice that in a water-only exposure, and that the LC50 value for cadmium in IW
associated with sediments was more than three times that in a water-only exposure.
x
3.3.4 Field sediments
In addition to short-term laboratory experiments with spiked sediments, there have been
several published studies of laboratory toxicity tests with metal-contaminated sediments from the
field. Ankley et al. (1991) exposed L. variegatus and the amphipod Hyalella azteca to 17
sediment samples along a gradient of cadmium and nickel contamination from a
Draft for SAB I -74
-------
freshwater/estuarine site in Foundry Cove, NY. In 10-d toxicity tests, H. azteca mortality was
absent in all sediments where SEM (cadmium plus nickel) was less than A VS. Mortality was
greater than controls only in sediments with more SEM than AVS. Lumbriculus variegatus was
far less sensitive to the sediments than H. azteca, which correlates with the differential sensitivity
of the two species in water-only tests with cadmium and nickel.
In 10-day toxicity tests with the saltwater amphipod A. abdita in these same sediments
from Foundry Cove, Di Toro et al. (1992) observed that metals concentrations ranging from 0.1
to 28 yumoles SEM/g sediment were not toxic hi some sediments, whereas metals concentrations
ranging from 0.2 to 1000 /imoles SEM/g were lethal in other sediments. These results indicate
that the bioavailable fraction of metals in sediments varies from sediment to sediment. In contrast,
the authors also observed a clearly discernable mortality-concentration relationship when mortality
was related to the SEM/AVS molar ratio (i.e., there was no significant mortality where SEM/AVS
ratios were < 1.0, mortality increased hi sediments having SEM/AVS ratios 1.0-3.0, and there
was 100% mortality in sediments with ratios > 10). The sum of the interstitial water toxic units
(IWTU) for cadmium and nickel ranged from 0.08 to 43.5. Sediments with <; 0.5 IWTUs were
always non-toxic, those with >2.2 IWTUs were always toxic, and two of seven sediments with
intermediate IWTUs (0.5 to 2.2) were toxic. Molar concentrations of cadmium and nickel in the
interstitial water were similar. However, cadmium contributed over 95 percent to the sum of the
toxic units because cadmium is 67 tunes more toxic to A. abdita than nickel. The latter illustrates
the utility of interstitial water concentrations of individual metals hi assigning the probable cause
of mortality in benthic species (Hansen et al., 1996a).
In tests with the same sediments also from Foundry Cove, Pesch et al. (1995) observed
that six of the 17 sediments tested had SEM/AVS ratios < 1.0, interstitial water toxic units < 0.5,
and none of the six were toxic to the polychaete Neanthes arenaceodentata. Nor were 11
sediments that contained SEM/AVS ratios > 1.0 toxic. These results were not surprising because
only one sediment had > 0.5 IWTUs, and because N. arenaceodentata is not sensitive to cadmium
and nickel and can avoid sediments containing toxic concentrations of these metals.
Ankley et al. (1993) examined the significance of AVS as a binding phase for copper
hi freshwater sediments from two copper-impacted sites. Based upon interstitial water copper
concentrations hi the test sediments, the 10-d LC50 for //. azteca was 31 Mg/L; this compared
favorably with a measured LC50 of 28 //g/L hi a 10-d water-only test. Sediments having
Draft for SAB \-l'5
-------
SEM/AVS ratios < 1.0 were not toxic. They also observed no toxicity in several sediments with
markedly more SEM than AVS suggesting that copper was not biologically available in these
sediments. Absence of copper in interstitial water from these sediments corroborated this lack of
bioavailability. This observation suggested the presence of binding phases hi addition to AVS for
copper hi the test sediments. Recent studies suggest that an important source of the extra binding
capacity hi these sediments was organic carbon (Mahony et al., 1996; U.S. EPA, 1994a).
Hansen et al. (1996a) investigated the biological availability of sediment-associated
divalent metals to A. abdita and H. azteca hi sediments from five saltwater locations and one
freshwater location hi the United States, Canada and China using 10-day lethality tests. Sediment
toxicity was not related to dry weight metals concentrations. In the 49 sediments evaluated where
metals were the likely cause of toxicity (i.e., those with less SEM than AVS and those with less
than 0.5 IWTU), no toxicity was observed. One third of the 45 sediment samples with more SEM
than AVS and more than 0.5 IWTU were toxic.
Hansen et al. (1996a) made an observation that is important to the interpretation of the
toxicity of sediments from field locations, particularly those from industrial harbors. They
observed that if these sediments are toxic and SEM/AVS ratios are < 1.0, non-metals associated
toxicity should always be suspected even if metals concentrations are very high on a dry weight
basis. Further, they stated that the use of such data to reach the conclusion that this EqP approach
is not valid is incorrect. This is because when SEM/AVS ratios were less than 1.0, there was an
almost complete absence of toxicity in spiked sediments, and field sediments where metals were
the only known source of contamination and IWTUs for metals were <0.5. Metals
concentrations, when expressed on a sum of the IWTU basis, can therefore provide insight that,
hi part, may explain apparent anomalies between SEM/AVS ratios and the observed toxicity of
these sediments and sediments from other field sites. The joint use of both SEM/AVS ratios and
interstitial water concentrations are powerful tools for explaining the presence of toxicity when
SEM/AVS ratiosxare < 1.0, and the absence of toxicity when SEM/AVS ratios are > 1.0. Over
all saltwater and freshwater field sediments tested in the laboratory, 100% were not toxic when
SEM-AVS was <;0.0 and IWTUs were <0.5 and 66.7% were toxic when SEM-AVS was >0.0
and IWTUs were ;>0.5 (Table 3-2).
Draft for SAB \-l 6
-------
3.3.5 Field Sites and Spiked Sediments Combined
Figures 3-18, a,b, and c and Table 3-2 summarize available data from saltwater and
freshwater sediments spiked with individual metals or metal mixtures, saltwater field sites and
freshwater field sites on the utility of metals concentrations in sediments normalized by dry
weight, interstitial water toxic units (IWTUs) or SEM/AVS ratios to explain the bioavailability
and acute toxicity of metals in sediments. Data are from Hansen et al.(1996a) and Berry et al.,
1999. This analysis contains all available data from 10-day lethality tests where mortality,
IWTUs, and SEM/AVS ratios are known from experiments with sediments most certainly toxic
only because of metals. The relationship between benthic organism mortality and total dry weight
metals concentrations in spiked and field sediments is not useful to causally relate metal
concentrations to organism response (Figure 3-18a). The overlap among bulk metals
concentrations which cause no toxicity and those which are 100 percent lethal is almost four
orders of magnitude.
Data in Figure 3-18b show that over all tests, the toxicity of sediments whose
concentrations are normalized on an IWTU basis are typically consistent with the interstitial water
toxic unit concept; that is if IWTUs are <. 1.0, then sediments should be lethal to <. 50 percent of
the organisms exposed, and significant mortality probably should be absent at < 0.5 IWTU
(Figure 3-18c). Of the spiked and field sediments evaluated which had IWTUs < 0.5, 97.7
percent of 175 sediments were non-toxic (Table 3-2). For the 142 sediments having IWTUs *
0.5, 68.3 percent were toxic (Table 3-2). However, and as stated above, given the effect on
toxicity or bioavailability of the presence of other binding phases (e.g., DOC) in interstitial water,
water quality (hardness or salinity) and organism behavior, it is not surprising that many
sediments having IWTUs k 0.5 are not toxic.
Data in Figure 3-18c show that over all tests, organism response in sediments whose
concentrations are normalized on an SEM/AVS basis is consistent with metal-sulfide binding on
a mole to mole basis as first described by Di Toro et al. (1990), and later recommended for
assessing the bioavailability of metals in sediments by Ankley et al. (1994). Saltwater and
freshwater sediments spiked with metals and from field locations with SEM/AVS ratios £ 1.0
were uniformly (98.1 percent of 158 sediments) non-toxic (Figure 3-18b; Table 3-2). The majority
(58.0 percent) of 174 sediments having SEM/AVS ratios > 1.0 were toxic. On the other hand,
given the effect on toxicity or bioavailability of the presence of other sediment phases that also
Draft for SAB 1-77
-------
affect bioavailability (Di Toro et al., 1987; Mahony et al., 1996), it is not surprising that many
sediments having SEM/AVS ratios > 1.0 are not toxic.
Over all tests, the data in Figure 3-18a, b, and c indicate that the use of both IWTUs
and SEM/AVS ratios together did not improve the accuracy of predictions of sediments that were
non-toxic (98.5 percent of 132 sediments; Table 3-2). However, it is noteworthy that 78.9
percent of the 123 sediments with both SEM/AVS > 1.0 and IWTUs ;> 0.5 were toxic (Table 3-
2). Therefore, the approach of using SEM/AVS ratios, IWTUs, and especially both indicators
to identify sediments of concern is very useful.
The results of all available data demonstrate that using SEM, AVS and interstitial water
metals concentrations to predict the toxicity of cadmium, copper, lead, nickel, silver and zinc in
sediments is quite certain. This is very useful, because the vast majority of sediments found in
the environment in the U.S. have SEM/AVS ratios z 1.0. This suggests that there should be little
concern about metals in sediments (Wolfe et al., 1994; Hansen et al., 1996a; Leonard et al.,
1996a; Section 4 of this document) on a national basis, even though localized areas of biologically
significant metal contamination do exist. However, a very important consideration is that most
of these data are from field sites where sediment samples were collected in the summer. This is
the time of the year when the seasonal cycles of AVS produce the maximum metal-binding
potentials (Boothman and Helmstetter, 1992; Leonard et al., 1993). Hence, sampling at seasons
and conditions when AVS is at minimal values is a must hi establishing the true level of overall
concern about metals in sediments and in evaluations of specific sediments. Predicting which of
the sediments with SEM/AVS >1.0 will be toxic is presently less certain. Importantly, the
correct classification rate seen in these experiments (accuracy of predicting which sediments were
toxic was 58.0% using the SEM/AVS ratio alone, 68.3% using IWTUs and 78.9% using both
indicators) is high. An SEM/AVS ratio >1.0, particularly at multiple adjacent sites, should
trigger additional tiered assessments which might include characterization of the spatial (both
vertical and horizontal) and temporal distribution of chemical concentration (AVS and SEM) and
toxicity, measurements of interstitial water metal and toxicity identification evaluations (TIE's).
In this context, the SEM, AVS, IWTU approach should be viewed as only one of the many
sediment evaluation methodologies.
Because AVS can bind divalent metals hi proportion to their molar concentrations,
Hansen et al. (1996a) proposed the use of the difference between the molar concentrations of SEM
and AVS (SEM-AVS) rather than SEM/AVS ratios used previously. The molar difference
provides important insight into the extent of additional available binding capacity and the
Draft for SAB 1-78
-------
magnitude by which AVS binding has been exceeded (Figure 3-19). Further, absence of organism
response when AVS binding is exceeded can indicate the potential magnitude of importance of
other binding phases in controlling bioavailability. Figure 3-19 shows that for most non-toxic
saltwater and freshwater field sediments, one to 100 yumoles of additional metal would be required
to exceed the sulfide binding capacity (i.e., SEM-AVS = -1 to -100 /zmoles/g). In contrast, most
toxic field sediments contained 1.0 to 1000 jmioles of metal beyond the binding capacity of sulfide
alone. Data on non-toxic field sediments whose sulfide binding capacity is exceeded (SEM-AVS
is > 0.0 Atmoles/g) indicates that other sediment phases, in addition to AVS, have great
significance in controlling metal bioavailability. In comparison to SEM/AVS ratios, the use of
SEM-AVS differences is particularly informative where AVS concentrations are low, such as
those from Steilacoom Lake and the Keweenaw Watershed, where the SEM-AVS difference is
numerically low and SEM/AVS ratios are high (Ankley et al., 1993),
EPA believes that results from tests using sediments spiked with metals and sediments
from the field in locations where toxicity is metals-associated demonstrate the value in explaining
the biological availability of metals concentrations normalized by SEM/AVS ratio and IWTUs
instead of dry weight metals concentrations. Importantly, data from spiked sediment tests
strongly indicate that metals are not the cause of most of the toxicity observed in field sediments
when both SEM/AVS ratios are ^ 1.0 and IWTU are < 0.5 (Table 3-2). Expressing
Draft for SAB 1-79
-------
100
80-
# 60-
CO
40
O 0
o
o
iSr
O
ID
V A\K>
A O v A
r—-frnO® *—r-
1 10
SEM-AVS Oimnol/g dry wt)
-**«:*-
100
1000
Figure 3-19. Percentage mortality of amphipods, oligochaetes and polychaetes exposed to
sediments from three saltwater and four freshwater field locations as a function of the sum
of the molar concentrations of SEM minus the molar concentration of AVS (SEM-AVS)
(from Hansen et al., 1996a): The vertical dashed line at SEM-AVS = 0.0 indicates the
boundary between sulfide-bound unavailable metal and potentially available metal.
Draft for SAB 1-80
-------
concentrations of metals in sediments on an SEM-AVS basis provides important insight into
available additional binding capacity of sediments and the extent to which sulfide binding has been
exceeded. It, along with measurement of interstitial water concentrations of metals, can
potentially identify the specific metal causing toxicity. This can theoretically be accomplished by
subtracting the metals-specific molar concentrations in order of their sulfide solubility product
constants (Kjp). Predictions of sediments not likely to be toxic, based on use of SEM-AVS and
IWTUs for all data from freshwater or saltwater field sediment and spiked sediment tests are
extremely accurate (98.5 percent) using both parameters (Table 3-2). While the predictions of
sediments likely to be toxic are less accurate, the use of SEM-AVS is extremely useful in
identifying sediments of potential concern (Table 3-2). Hansen (1995) summarized data from
amphipod tests using freshwater and saltwater laboratory metals-spiked sediments and field
sediments where metals were a known problem by comparing the percentage of sediments that
were toxic to the SEM-AVS concentration (tests with polychaetes and gastropods were excluded
because these organisms avoid exposure). Seventy percent of the sediments in these amphipod
studies with an SEM-AVS concentration of ^0.76 //moles of excess SEM/g were toxic. The
corresponding values for 80, 90 and 100% of the sediments being toxic were 2.7, 16 and 115
jumoles of excess SEM/g, respectively.
Of course, SEM, AVS and IWTUs can only predict toxicity or the lack of toxicity due
to metals in sediments. They cannot be used alone to predict the toxicity of sediments
contaminated with toxic concentrations of other contaminants. However, SEM/AVS ratios have
been used in sediment assessments to rule out metals as probable causative agents of toxicity
(Wolfe et al., 1994). Also, the use of SEM and AVS to predict the biological availability and
toxicity of cadmium, copper, lead, nickel, silver and zinc is applicable only to anaerobic
sediments that contain AVS. In aerobic sediments binding factors other then AVS control
bioavailability (Di Toro et al., 1987; Tessier et al., 1993). Measurement of interstitial water
metal may be useful for evaluations of these and other metals in aerobic and anaerobic sediments
(Ankley et al., 1994). Even with these caveats, EPA believes that the use of SEM, AVS and
interstitial measurements in combination are superior to all other currently available sediment
evaluation procedures to causally assess the implications of these six metals associated with
.sediments (see discussion in Section 5 "Implementation" for further guidance).
3.4 PREDICTING METAL TOXICITY: LONG-TERM STUDIES
Taken as a whole, the short-term laboratory experiments with metal-spiked and field-
collected sediments present a strong argument for the ability to predict an absence of metal
Draft for SAB 1-81
-------
toxicity based upon sediment SEM:AVS relationships and/or interstitial water metal
concentrations. However, for this approach to serve as a valid basis for ESG derivation,
comparable predictive success must be demonstrated in long-term laboratory and field experiments
where chronic effects could be manifested (Luoma and Carter, 1993; Meyer et al., 1994). This
demonstration was the goal of experiments described by Hare et al. (1994), DeWitt et al. (1996),
Hansen et al. (1996b), Liber et al. (1996) and Sibley et al. (1996). An important experimental
modification to these long-term studies, as opposed to the short-term tests described in Section
3.3, was the collection of horizon-specific chemistry data. This is required because AVS
concentrations often increase, and SEM/AVS ratios decrease, with an increase in sediment depth
(Howard and Evans, 1993; Leonard et al., 1996a); hence, chemistry performed on homogenized
samples might not reflect the true exposure of benthic organisms dwelling in surficial sediments
(Luoma and Carter, 1993; Hare et al., 1994; Peterson et al., 1996).
3.4.1 Life-cycle toxicity tests
DeWitt et al. (1996) conducted an entire life-cycle toxicity test with the marine
amphipod Leptocheirus plwnulosus exposed for 28 d to cadmium-spiked estuarine sediments
(Table 3-3). The test began with newborn amphipods and measured effects on survival, growth
and reproduction relative to interstitial water and SEM/AVS normalization. Seven treatments of
Cd were tested: 0 (control), 0.34, 0.74, 1.31, 1.55, 2.23 and 4.82 molar SEMa/AVS ratios
(measured concentrations). Gradients in AVS concentration as a function of sediment depth were
greatest in the control treatment, decreased as the SEM^/AVS ratio increased, and became more
pronounced over tune. Depth gradients in SEMa/AVS were primarily due to the spatial and
temporal changes in AVS concentration, because SEMa concentrations changed very little with
tune or depth. Thus, hi most treatments SEMa/AVS ratios were higher at the top of sediment
cores than at the bottom. This is expected because the oxidation rate of iron sulfide in laboratory
experiments is very rapid (100% in 60 to 90 minutes) while for cadmium sulfide it is quite slow
(10% hi 300 hours) (Mahony et al., 1993). Interstitial cadmium concentrations increased hi a
dramatic step-function fashion hi treatments having SEM/AVS ratios *2.23; and were below the
96-h LC50 for this amphipod hi lesser treatments. There were no significant effects on survival,
growth or reproduction hi sediments
Draft for SAB 1-82
-------
Table 3-3. Summary of the results of full life-cycle and colonization toxicity tests conducted in the laboratory and field using
sediments spiked with individual metals and metal mixtures.
to
oo
Toxicity Test Metal
Life-cycle:
Leptocheints Cadmium
plumulosus
Chironomits Zinc
tentans
Colonization:
Laboratory- , Cadmium
saltwater
Field-saltwater Cadmium, copper, nickel,
zinc
Field- Cadmium
freshwater
Field- Zinc
freshwater
Duration Measured
NOECs
28 -3.4, -2.0,
0.78, 1.9
56 -4.3, -2.6,
-1.4, 6.4
118 -13.4
120 <0
-0.45, -0.25,
0.5b
368 -3.6, -3.5,
-2.9, -2.0
SEM-AVS*
OECs
8.9, 15.6
21.9,
32.2
8.0, 27.4
—
4.5b
1.0
Effect
Mortality 100%
Larval mortality 85-100%.
Weight and emergence reduced.
Fewer polychaetes, shifts in
community composition, fewer
species, bivalves absent,
tunicates increased.
No effects observed.
Reduced Chironomits numbers.
Bioaccumulation.
Occasional minor reductions in
Naididae oligochaetes.
Reference
DeWitt et
al., 1996
Sibley et
al., 1996
Hansen et
al., 1996
Boothman
et al., 1996
Hare et al.,
1994
Liber et al.,
1996
- AYS ditterences are used instead ot SfcM/AVS ratios to standardize across the studies referenced. An bt-M-AVS
<0 is the same as an SEM/AVS ratio of < 1.0. An SEM-AVS difference of >0 is the same as an SEM/AVS ratio of >1.0.
b Nominal concentrations.
-------
containing more AVS than cadmium (SEM/AVS ratios 0.34, 0.74, 1.31 and 1.55), in spite of the
fact that these samples contained from 183 to 1370 ^g cadmium/g sediment. All amphipods died
hi sediments having SEM/AVS ratios 2.23. These results are consistent with predictions of metal
bioavailability from acute tests with metal-spiked sediments (i.e., that sediments with SEM^/AVS
ratios ^ 1 are not toxic, interstitial water metal concentrations are related to organism response,
and sediments with SEMa/AVS ratios > 1 may be toxic).
9
Sibley et al. (1996) reported similar results from a 56-d life-cycle test conducted with
the freshwater midge Chironomus tentans exposed to zinc-spiked sediments (Table 3-3). The test
was initiated with newly hatched larvae and lasted through one complete generation during which
survival, growth, emergence and reproduction were monitored. In sediments where the molar
difference between SEM and AVS was <0, at dry wt zinc concentrations as high as 270 mg/kg,
concentrations of zinc hi the sediment interstitial water were low and no adverse effects were
observed for any of the biological endpoints measured. Conversely, when SEM-AVS exceeded
0, AVS and interstitial water concentrations of zinc increased with increasing treatments (being
highest hi surficial sediments), and reductions hi survival, growth, emergence and reproduction
were observed. Over the course of the study, the absolute concentration of zinc hi the interstitial
water hi these treatments decreased due to the increase hi sediment AVS and loss of zinc from
twice daily renewals of the overly ing water.
3.4.2 Colonization tests
Hansen et al. (1996b) conducted a 118-d benthic colonization experiment hi which
sediments were spiked to achieve nominal cadmium/AVS molar ratios of 0.0 (control), 0.1, 0.8
and 3.0 then held hi the laboratory hi a constant flow of unfiltered seawater (Table 3-3).
Oxidation of AVS hi the surficial 2.4 cm of the control treatment within two to four weeks
resulted hi sulfide profiles similar to those occurring hi sediments hi nearby Narragansett Bay, RI
(Boothman and Helmstetter, 1992). In the nominal 0.1 cadmium/AVS treatment, measured
SEMcd was always less than AVS, interstitial cadmium concentrations (<3-10 /xg/L) were less
than those likely to cause biological effects, and no significant biological effects were detected.
In the nominal 0.8 cadmium/AVS treatment, measured SEMo, commonly exceeded AVS hi the
surficial 2.4 cm of sediment, and interstitial cadmium concentrations (24-157 ,ug/L) were
sufficient to be of lexicological significance to highly sensitive species. In this treatment, shifts
hi the presence or absence over all taxa, and fewer macrobenthic polychaetes (Mediomastus
ambiseta, Streblospio benedicti and Podarke obscura) and unidentified meiofaunal nematodes,
were observed. In the nominal 3.0 cadmium/A VS treatment, concentrations of SEMcd were
Draft for SAB 1-84
-------
always greater than AVS throughout the sediment column. Interstitial cadmium ranged from
28,000 to 174,000 /*g/L. In addition to the effects observed in the nominal 0.8 cadmium/AVS
treatment, sediments in the 3.0 cadmium/AVS treatment were colonized by fewer macrobenthic
species, polychaete species and harpacticoids, had lower densities of diatoms, lacked bivalve
molluscs, anchexhibited other impacts. Over all treatments, the observed biological responses
were consistent with predicted possible adverse effects resulting from elevated SEM^/AVS ratios
in surficial sediments and interstitial water cadmium concentrations.
Boothman et al. (1996) conducted a field colonization experiment in which sediments
from Narragansett Bay, RI were spiked with an equi-molar mixture of cadmium, copper, nickel
and zinc at nominal SEM:AVS ratios of 0.1, 0.8 and 3.0, placed in boxes, and replaced in
Narragansett Bay (Table 3-3). AVS concentrations decreased with tune in surface (0-3 cm)
sediments in all treatments where SEM < AVS, but did not change in subsurface (6-10 cm)
sediments or in the entire sediment column hi the SEM > AVS treatment. SEM decreased with
time only where SEM exceeded AVS. The concentration of metals in interstitial water was below
detection limits when SEM was less than AVS. When SEM exceeded AVS, significant
concentrations of metals were present in interstitial water in order of their sulfide solubility
product constants. Interstitial water concentrations in these sediments decreased with time
exceeding the WQC hi interstitial water for 60 days for all metals, 85 days for cadmium and zinc,
and for the entire experiment (120 days) for zinc. Benthic faunal assemblages hi the spiked
sediment treatments were not different from the control treatment. Lack of biological response
was consistent with the vertical profiles of SEM and AVS. AVS was greater than SEM hi all
surface sediments, including the top 2 cm of the 3.0 SEM: AVS treatment, due to the oxidation
of AVS and loss of SEM. The authors speculated that interstitial metal was likely absent hi the
surficial sediments hi spite of data demonstrating the presence of significant measured
concentrations of interstitial metal. This is because the interstitial water hi the nominal 3.0
SEM/AVS treatment was sampled from sediment depths where SEM was hi excess. It is in
surficial sediments where settlement by saltwater benthic organisms first occurs. Also, there was
a storm event which allowed a thin layer of clean sediment to be deposited on top of the spiked
sediment (Boothman, USEPA, personal communication). These data demonstrate the importance
of sampling of sediments and interstitial water hi sediment horizons where benthic organisms are
active.
Hare et al. (1994) conducted an approximately 1-yr field colonization experiment hi
which uncontaminated freshwater sediments were spiked with cadmium and replaced in the
oligotrophic lake from which they originally had been collected (Table 3-3). Cadmium
Draft for SAB 1-85
-------
concentrations in interstitial waters were very low at cadmium:AVS molar ratios < 1.0, but
increased markedly at ratios > 1.0. They reported reductions in the abundance of only the
chironomid Chironomus salinarius in the nominal 10.0 SEM/AVS treatment. Cadmium was
accumulated by organisms from sediments with surficial SEM concentrations that exceeded those
of AVS. These sediments also contained elevated concentrations of cadmium in interstitial water.
Liber et al. (1996) performed a field colonization experiment using sediments having
4.46 fimole sulfide from a freshwater mesotrophic pond (Table 3-3). Sediments were spiked with
0.8, 1.5, 3.0, 6.0 and 12.0 ^mole zinc, replaced in the field, and chemically and biologically
sampled over 16 mo. There was a pronounced increase in AVS concentrations with increasing
zinc concentration; AVS was lowest in the surficial 0-2 cm of sediment with minor seasonal
variations. With the exception of the highest spiking concentration (ca., 700 mg/kg, dry wt),
AVS concentrations remained larger than those of SEM. Interstitial water zinc concentrations
were rarely detected in any treatment, and were never at concentrations that might pose a hazard
to benthic macroinvertebrates. The only observed difference in benthic community structure
across the treatments was a slight decrease hi the abundance of Naididae oligochaetes at the
highest spiking concentration. This absence of any noteworthy biological response was consistent
with the absence of interstitial water concentrations of biological concern. This was attributed to
the increase in concentrations of iron and manganese sulfides, produced during periods of
diagenesis, which were replaced by the more stable zinc sulfide which is less readily oxidized
during whiter months. In this experiment, and theoretically in nature, excesses of sediment metal
might be overcome over tune due to the diagenesis of organic material. In periods of minimal
diagenesis, the oxidation rates of metal sulfides, if sufficiently great, could release biologically
significant concentrations of the metal into interstitial waters. This phenomenon should occur
metal-by-metal in order of then- sulfide solubility product constants.
Draft for SAB 1-86
-------
SECTION 4
DERIVATION OF ESG FOR METALS
4.1 GENERAL INFORMATION
Section 4 of this document presents the technical basis for establishing ESG for
copper, cadmium, nickel, lead, silver and zinc. The basis of the overall approach is the use of
EqP theory linked to the concept of maintaining metal activity for the sediment interstitial
water system below effects levels. Extensive lexicological data from short-term and long-term
laboratory and field experiments, with both marine and freshwater sediments and a variety of
t
species, indicates that it is possible to reliably predict an absence of metal toxicity based upon
EqP theory. ESG for all six metals collectively can be derived using two procedures: (a) by
comparing the sum of their molar concentrations, measured as SEM, to the molar
concentration of AVS in sediments (AVS Guideline); or (b) by comparing the measured
interstitial water concentrations of the metals to WQC final chronic values (FCVs) (Interstitial
Water Guidelines). These approaches are described hi more detail below. A lack of
exceedence of ESG based upon any one of the two procedures indicates that metal toxicity
should not occur. Exceedence of either the AVS or Interstitial Water Guidelines is indicative
of a potential problem that would entail further evaluation.
At present, EPA believes that the technical basis for implementing these two
approaches is supportable. The Organic Carbon and Minimum Partitioning Approaches as
proposed to the SAB and hi Ankley et al. (1996) require additional research prior to their
implementation. Research issues for these latter two approaches include the development of
robust partitioning datasets for the six metals, as well as investigation of factors such as the
importance of other binding phases. The four approaches have been presented to and
reviewed by the Science Advisory Board of EPA (U.S EPA, 1994a; 1995a).
Additional research required to fully implement other approaches for deriving ESG
for these metals and to derive ESG for other metals includes the development of uncertainty
estimates associated with any approach; part of this would include their application to a variety
of field settings and sediment types. Research also is needed to establish the technical basis
for ESG for metals other than the six described herein, such as mercury, arsenic and
chromium. Finally, the ESG approaches are intended to protect benthic organisms from direct
toxicity associated with exposure to metal-contaminated sediments. They are not designed to
Draft for SAB \-W
-------
protect aquatic systems from metal release associated, for example, with sediment suspension,
or the transport of metals into the food web either from sediment ingestion or the ingestion of
contaminated benthos. This latter issue, hi particular, should be the focus of future research
given existing uncertainty in the prediction of bioaccumulation of metals by benthos (Ankley,
1996).
The following nomenclature is used in subsequent discussion of ESG derivation for
metals. The ESG for the metals are expressed in molar units because of the molar
stoichiometry of metal binding to AVS. Thus, solid phase constituents (AVS, SEM) are hi
moles/g dry wt. The interstitial water metal concentrations are expressed hi //moles/L, either
as dissolved concentrations [MJ or activities {M2+} (Stumm and Morgan, 1981). The
subscripted notation, Md, is used to distinguish dissolved aqueous phase molar concentrations
from solid phase molar concentrations with no subscript. For the combined concentration,
[SEMT], the units are moles of metal per volume of solid plus liquid phase (i.e., bulk). Note
also that when [SEMAg] is summed and/or compared to AVS, lh. the molar Ag concentration is
applied.
One final point should be made with respect to nomenclature. Use of the terms
non-toxic and having no effect mean only with respect to the six metals considered in this
paper. The toxicity of field collected sediments can be caused by other chemicals. Therefore,
avoiding exceedences of ESG for metals does not mean that the sediments are non-toxic. It
only ensures that the six metals being considered should not have an undesirable biological
effect. Moreover, as discussed in detail below, exceedence of the guidelines for the six metals
does not necessarily indicate that metals will cause toxicity. For these reasons, we strongly
recommend the use of toxicity tests, TIEs, chemical monitoring hi vertical, horizontal and
temporal scales, and other assessment methodologies as integral parts of any assessment
concerned with the effects of sediment-associated contaminants (Ankley et al., 1994).
4.2 SINGLE METAL SEDIMENT GUIDELINES
Except in rare instances, single metal guidelines are not usually applicable to field
situations since there is almost always more than one metal to be considered. As will become
subsequently clear, it would be technically indefensible to derive guidelines for one metal at a
tune because of the competitive nature of AVS binding. Nevertheless, it is illustrative to
present the logic for single metals as a prelude to the derivation of the multiple metal
Draft for SAB 1-88
-------
guidelines.
4.2.1 AVS Guidelines
It has been demonstrated that if the SEM of a sediment is less than or equal to the
[SEMMAVS] (4-1)
AVS then no toxic effects are seen. This is consistent with the results of a chemical
equilibrium model for the sediment - interstitial water system (Di Toro et al., 1992). The
resulting metal activity {M2*} can be related to the total SEM of the sediment and water, and
to the solubility products of the metal sulfide (K^) and iron sulfide (KfeS) . In particular, it is
true that at [SEM] <; [AVS] then:
KFeS
Because the ratio of metal sulfide to iron sulfide solubility products (KMS/KFeS) is very small
(< 10'5) even for the most soluble of the sulfides, the metal activity of the sediment is at least
five orders of magnitude smaller than the SEM (see Di Toro et al. (1992) for data sources and
references). This indicates that no biological effects would be expected. Therefore, the
condition [SEM] < [AVS] is a "no effect" ESQ.
The reason we use the term "no effect" is that for the condition [SEM] < [AVS] no
biological impacts are expected. However, for [SEM] > [AVS], which might seemingly be
considered a ESG violation, there are many documented instances where no biological impacts
occur (e.g., because organic carbon partitioning controls metal bioavailability in the interstitial
water, or the species of concern avoid or are insensitive to metals).
4.2.2 Interstitial Water Guidelines
• The condition [SEM] z [AVS] indicates that the metal activity of the sediment -
interstitial water system is low and, therefore, below lexicologically-significant
Draft for SAB 1-89
-------
concentrations. Another way of ensuring this is to place a condition on the interstitial water
activity directly. Suppose that we knew the metal activity, denoted by {FCV}, that
corresponded to the [FCV]. Then the ESG corresponding to this effect level is:
(4-3)
It is quite difficult, however, to measure and/or calculate metal activity in a solution phase, at
the low concentrations required, since it depends on the identities, concentrations and
thermodynamic affinities of other chemically reactive species that are present. Also, the WQC
are not expressed on an activity basis. An approximation to this condition is:
[Md]*[FCVd] (4-4)
where [FCVj is the FCV applied to total dissolved metal concentrations. That is, we require
that the total dissolved metal concentration hi the interstitial water [MJ be less than the FCV
applied as a dissolved guideline. Although this requirement ignores the effect of chemical
speciation on both sides of the equation - compare Equations (4-3) and (4-4) - it is the
approximation that is currently being suggested by EPA for the WQC for metals (Prothro,
1993). That is, the WQC should be applied to the total dissolved - rather than the total acid
recoverable - metal concentration (Table 4-1; U.S. EPA,1995b). Hence, if this second
condition is satisfied it is consistent with the level of protection afforded by the WQC.
In situations where the SEM exceeds the AVS ([SEM] > [AVS]), but the interstitial
water total dissolved metal is less than the final chronic value ([MJ < [FCVj), this sediment
would not violate the guidelines. These cases occur when significant binding to other phases
occurs. It should be noted that using the FCV for metals in freshwater samples requires that
the hardness of the interstitial water be measured since the WQC vary with hardness.
4.3 MULTIPLE METALS GUIDELINES
As described in the previous subsection, and from a practical standpoint it is
insufficient and inappropriate to consider each metal separately because of the interactive
Draft for SAB 1-90
-------
nature of metal-sulfide binding. This is of particular concern for the AVS guidelines.
Table 4-1. Water quality criteria (WQC) criteria continuous concentrations (CCC) based on
the dissolved concentration of metal*. These WQC CCC values are for use in the Interstitial
Water Guidelines approach for deriving sediment guidelines based on the dissolved metal
concentrations in interstitial water.
Metal
Cadmium
Copper*
Lead
Nickel
Silver
Zinc
" (U.S.EPA.
b D/-mn/-l/»H oil
Saltwater CCC, /zg/L
9.3
3.1
8.1
8.2
NAf
81
1995b).
1 orifrwiG tr\ tu/r\ cifrnifi/Mint •ftmiir/»c
Freshwater CCC, Atg/Lc
CF*1 [e<° 7852['n]-3 4901
0 960re(0'8545tln(hardness)]"1 465)1
0*701 iW'^SflnfliaixlnessXH 705)1
. / y L ic j
\) . yy i |_e j
NAf
0.986[e(a8473[ln(hardness)1+0-7614)]
For example, the freshwater CCC at a hardness of 50, 100, and 200 mg CaCO3/L are 0.56,
0.94, and 1.6 Mg cadmium/L; 6.2, 12, and 20,ug copper/L; 1.0, 2.5, and 6.1//g lead/L; 88,
160, and 280,ug nickel/L; and 58, 108, and 187 pg zinc/L.
CF= Conversion factor to calculate the dissolved CCC for cadmium from the total CCC for
cadmium: CF=1.101672-[(ln hardness)(0.041838)]
The saltwater CCC for copper is from the "Ambient Water Quality Criteria- Saltwater
Copper Addendum" (U.S. EPA, 1995c).
The silver criteria are currently under revision to reflect water quality factors that influence
the criteria such as hardness, DOC, chloride and pH among other factors. Since silver has
the smallest solubility product (see Table 2-2) and the greatest affinity for AVS, it would be
the last metal to be released from the AVS or the first metal to bind to the AVS. Therefore,
it is unlikely that silver would occur in the interstitial water. However in sediments
contaminated with silver the user should be aware of the limitations in the above criteria for
silver. AVS Guidelines can be applied although the Interstitial Water Guidelines can not.
If the AVS Guideline is exceeded (£SEM > AVS) and the sediment is contaminated with
silver, further testing and evaluations would be warranted to access toxicity.
Draft for SAB 1-91
-------
4.3.1 AVS Guidelines
The results of calculations using chemical equilibrium models indicate that metals act in an
competitive manner when binding to AVS. That is, the six metals: silver, copper, lead,
cadmium, zinc and nickel will bind to AVS and be converted to their respective sulfides in this
sequence (i.e., in the order of increasing solubility). Therefore, they must be considered
together. There cannot be a guideline for just nickel, for example, since all the other metals
may be present as metal sulfides and, therefore, to some extent, as AVS. If these other metals
are not measured as SEM, then the £SEM will be misleadingly small, and it may appear that
[£SEM] < [AVS] when in fact this would not be true if all the metals are considered
together. It should be noted that EPA currently restricts this discussion to the six metals listed
above; however, in situations where other sulfide forming metals (e.g., mercury) are present
at high concentrations, they also must be considered.
The equilibrium model prediction of metal activity is similar to the single metal example
when a mixture of the metals is present. If the molar sum of SEM for the six metals is less
than or equal to the AVS, that is:
E, [SEN1MAVS] ' (4.5)
then:
(4-6)
where [SEAMED! is the total SEM (/zmoles/L(bulk)) for the 1th metal. Thus the activity of each
metal, {M,}, is unaffected by the presence of the other sulfides. This can be understood as
follows. Suppose that the chemical system starts initially as iron and metal sulfide solids and
that the system proceeds to equilibrium by each solid dissolving to some extent. The iron
sulfide dissolves until the solubility product of iron sulfide is satisfied. This sets the sulfide
activity. Then each metal sulfide dissolves until reaching its solubility. Since so little of each
dissolve relative to the iron sulfide, the interstitial water chemistry is not appreciably changed.
Hence, the sulfide activity remains the same and the metal activity adjusts to meet each
solubility requirement. Therefore, each metal sulfide behaves independently of one another.
Draft for SAB 1-92
-------
The fact that they are only slightly soluble relative to iron sulfide is the cause of this behavior.
Thus, the AVS Guidelines are easily extended to the case of multiple metals.
4.3.2 Interstitial Water Guidelines
The application of the Interstitial Water Guideline to multiple metals is complicated, not by
the chemical interactions of the metals in the sediment - interstitial water system (as in the case
with the AVS Guideline), but rather because of their possible toxic interactions. Even if the
individual concentrations do not exceed the FCV of each metal (Fcvo)» the metals could exert
additive effects that might result in toxicity (Besieger et al., 1986; Spear and Fiandt, 1986;
Enserink et al., 1991; Kraak et al., 1994). Therefore, to address this potential additivity, the
interstitial water metal concentrations are converted to toxic units (TUs) and these are
summed. Since FCVs are used as the no effects concentrations these TUs are referred to as
interstitial water guidelines toxic units (IWGTUs). For freshwater sediments, the FCVs are
hardness dependent for all of the divalent metals under consideration, and thus, need to be
adjusted to the hardness of the interstitial water of the sediment being considered. Because
there are no FCVs for silver in freshwater or saltwater, this approach is not applicable to
sediments containing significant concentrations of silver (i.e. , SSEM > AVS). Since silver
has the smallest solubility product (see Table 2-2) and the greatest affinity for AVS, it would
be the last metal to be released from the AVS or the first metal to bind with AVS. Therefore
it is unlikely that silver would occur hi the interstitial water. For the i* metal with a total
dissolved concentration [Mlid], the IWGTU is:
A lack of exceedence of the ESG requires that the sum of the IWGTUs be less than or equal to
one:
Hence, the multiple metals guideline is quite similar to the single metal case (Equation 4-4),
except that it is expressed as summed IWGTUs.
Draft for SAB 1-93
-------
To summarize, the proposed ESG are as follows. The sediment passes the ESG for the six
metals if either of these conditions is satisfied:
(a) AVS Guideline:
[SEM.MAVS] (4.5)
where
£ ,[SEMJ = [SEMcJ + [SEMcJ + [SEMpt] + [SEMNl] + [SEMzJ + [l/2SEMAg]
(b) Interstitial Water Guideline:
where
[FCV.d] [FCV^d] [FCVc4d] [FCVpbd]
(.
If either of these two conditions are violated, this does not mean that the sediment is toxic.
For example, if the AVS in a sediment is non-detectable, then condition (a) will be violated.
However, if there is sufficient organic carbon sorption so that condition (b) is satisfied, then
the sediment would be deemed acceptable.
If both of these conditions are violated, or if the AVS guideline is violated and the sediment
is contaminated with silver, then there is reason to believe that the sediment may be
unacceptably contaminated by these metals. Further testing and evaluations would therefore be
useful in order to assess actual toxicity and its causal relationship to the six metals. These
may include acute and chronic tests with species that are sensitive to the metals suspected to be
causing the toxicity. Also, in situ community assessments, sediment TIEs and seasonal
Draft for SAB 1-94
-------
characterizations of the SEM, AVS, and interstitial water concentrations would be appropriate
(Ankleyetal., 1994).
4.4 ESG FOR METALS VS. ENVIRONMENTAL MONITORING DATABASES
The purpose of this Section is to compare ESG based on SEM-AVS or IWGTUs to
chemical monitoring data from freshwater and saltwater sediments in the United States. This
comparison of AVS-SEM and interstitial water concentrations can indicate the extent of metals
contamination in the United States. When toxicity or benthic organism community health data
are available hi conjunction with these concentrations it is possible to speculate as to potential
causes of the observed effects.
4.4.1 Data Analysis
Three sources were identified which contain both AVS and SEM databases; one also had
data on concentrations of metals hi interstitial water. Toxicity tests were also conducted on all
sediments from these sources. The databases are from the Environmental Monitoring and
Assessment Program (EMAP) (Leonard et al., 1996a), National Oceanic and Atmospheric
Administration, National Status and Trends Program (NOAA NS&T) (Wolfe et al., 1994;
Long et al., 1995; 1996) and from the Regional Environmental Monitoring and Assessment
Program (REMAP) (Adams et al., 1996).
Freshwater sediments:
The AVS and SEM concentrations hi the 1994 EMAP database from the Great Lakes were
analyzed by Leonard et al. (1996a). Forty-six sediment grab samples and nine core samples
were collected in the summer from forty-two locations in Lake Michigan. SEM, AVS, TOC,
interstitial water metals (when sufficient volumes were present), and 10-day sediment toxicity
to the midge, Chironomus tertians, and the amphipod, Hyallela azteca, were measured hi
sediments collected by the grab (Appendix C).
The AVS concentrations vs. SEM-AVS differences from Appendix C are plotted hi Figure
4-1. Grab sediment samples containing AVS concentrations below the detection limit of 0.05
AVS are plotted at that concentration. Forty-two of the 46 (91 percent) samples had
Draft for SAB 1-95
-------
10
^x
CO
-10
-20
LU
CO
-30
-40
0.01
10
UJ
CO
-5
-10
0.01
(a)
0.1 1 10
Acid Volatile Sulfide (umol S/g)
100
(b)
4> 4
0.1 1 10
Acid Volatile Sulfide (umol S/g)
100
Figure 4-1. SEM minus AVS values versus AVS concentrations in EMAP-Great Lakes
sediments from Lake Michigan. Data are from surficial grab samples only (this figure is
taken from Leonard et al., 1996, see data in Appendix C). The upper plot shows all
values, the lower plot has the ordinate limited to SEM minus AVS values between -10 and
+10.
Draft for SAB 1-96
-------
SEM-AVS differences greater than 0. Thirty-six of these had less than 1.0 //mol of £SEM
metal per gram sediment; and none had over 5.8 //moles/g of excess metal. In theory,
sediments with SEM concentrations in excess of that for AVS have the potential to be toxic
due to metals. However, the majority of exceedence occur in places where the AVS is very.
small and the amount of SEM is also very small. For these Lake Michigan sediments, a closer
look at both interstitial water metal and toxicity test results is needed. Measurement of the
concentrations of metals hi interstitial water can be used to determine if the excess metals are
bound to other sediment phases, therefore, prohibiting toxicity due to interstitial metal.
Interstitial water guidelines toxic units (IWGTU) can be calculated for each metal as the
interstitial water concentration divided by the final chronic value for that metal. Interstitial
water volumes were sufficient to measure metals concentrations in 20 of the samples. The sum
of the IWGTU for cadmium, copper, lead, nickel and zinc in these sediments was less than 0.4
(Leonard et al., 1996a). In 10-d toxicity tests using Chironomus tentans and Hyalella azteca,
no toxicity was observed 81 % of the 21 sediments not exceeding the ESG. They conclude that
for the toxic sediments that did not exceed the metals ESG, the observed toxicity is not likely
due to metals. Further, these sediments are unlikely to be contaminated by metals (Leonard et
al., 1996a). These data demonstrate the value of using both SEM-AVS and IWGTUs to
evaluate the risks of metals in sediments.
Saltwater sediments:
Saltwater data from a total of 398 sediment samples from five monitoring programs
representing the eastern coast of the United States from Chesapeake Bay to Massachusetts are
included in Figure 4-2. The EMAP Virginia Province database (U.S. EPA, 1996) consists, in
part, of 127 sediment samples collected from August to mid-September 1993 from randomly
selected locations in tidal rivers and small and large estuaries from the Chesapeake Bay to
Massachusetts (Strobel et al., 1995). The NOAA data is from Long Island Sound, Boston
Harbor and the Hudson River Estuary. Sediments were collected from 63 locations in the
coastal bays and harbors of the Long Island Sound in August, 1991 (Wolfe et al., 1994).
Sediment samples from 30 locations in Boston Harbor were collected in June and July 1993
(Long et al., 1996). Sediment samples from 38 locations in the Hudson River Estuary were
collected from March to May 1991 (Long et al., 1995). Sediment samples were collected in
the REMAP program from 140 locations from the New York/ New Jersey Harbor Estuary
System (Adams et al., 1996). All of the above sediment grab samples were from
approximately the top 2 cm of undisturbed sediment.
Draft for SAB 1-97
-------
50
w
•50
•100
-150
-200I
0.01
O EMAPVP
O REMAP NY-NJ HARBOR ESTUARY
O NOAAUBO.HR
a
o
0.1
1 10
AVS (umol/g)
100
1000
10.0
-5*
"5
UJ
5.0
on
-5.0
10.0
0
e
0 fjffr t
•
•
.01
0
0
0 0
a o o
0 4 t^Jpo^k^tiL "°" °
^^qjg^ O^Q
t^^ °
&
#•
0.1 1 10 100 10
AVS (umo!/g)
Figure 4-2. SEM minus AVS values versus AVS concentrations in EMAP-Estuaries
Virginian Province (U.S. EPA, 199t), REMAP-NY/NJ Harbor Estuary (Adams et al.,
1996) and the NOAANST Long Island Sound (Wolfe et al., 1994), Boston Harbor (Long
et al., 1996), and Hudson-Raritan Estuaries (Long et al., 1995); (see data in Appendix D).
The upper plot shows all values, and the lower plot has the ordinate limited to SEM minus
AVS values between -10 and +10.
Draft for SAB 1-98
-------
For saltwater sediments, molar concentration of AVS typically exceeds that for SEM (SEM-
AVS <0) for most of the samples across the entire range of AVS concentrations (Figure 4-2).
A total of 68 of the 398 saltwater sediments (17 percent) had an excess of metal, and only 4 of
the 68 (6 percent) had over 2 /^mol/g excess SEM. As AVS levels increase above this
concentration fewer and fewer sediments have SEM-AVS differences that are positive; none
occurred when AVS was >8.1 ^mol/g. Unlike the sediments from the freshwater EMAP
survey in Lake Michigan, interstitial water was not measured hi these saltwater sediments.
Only five of the 68 sediments (7 percent) having excess of up to 0.9 ^mol/g SEM were toxic
in 10-d sediment toxicity tests with the amphipod Ampelisca abdita, whereas 79 of 330 (24
percent) sediments having an excess of AVS were toxic. The data support the interpretation
that (1) toxicity was NOT metals-related in the 79 sediments where AVS was in excess over
SEM; (2) metals might have caused the toxicity in the five toxic sediments having an excess of
metal, but even hi the absence of measurements of interstitial water metals concentrations, we
speculate that metals toxicity is unlikely because there was only £0.9 /zmol/g excess SEM (the
molar concentration SEM most often exceeds that of AVS, in sediments having AVS
concentrations si /zmol/g); and (3) the absence of toxicity in sediments having an excess of
SEM of up to 4.4 i/mol/g indicates that significant metal-binding potential over that of AVS
existed in some sediments. Organic carbon concentrations of from 0.05% to 15.2% (average
1.9 percent) provides for some of this additional metal-binding.
The data above appear to suggest that in the United States direct toxicity caused by metals
in sediments is extremely rare. While this might be true, these data by themselves are
inconclusive and it would be inappropriate to use the data from the above studies to reach this
conclusion. All of the above studies were conducted hi the summer when the seasonal
biogeochemical cycling of sulfur should produce the highest concentrations of iron
monosulfide which should make direct metal-associated toxicity less likely than hi the
whiter/spring months. Accurate assessment of the extent of the direct ecological risks of
metals in sediments requires that sediment monitoring occur hi the months of minimum AVS
concentration; typically November to early May. These yet to be conducted studies must
monitor at a minimum SEM, AVS, and interstitial water metal and toxicity. The data
presented here are not intended to be used to draw conclusions about toxicity due to
resuspension or bioaccumulation.
Draft for SAB 1-99
-------
SECTION 5
IMPLEMENTATION
5.1 CONSIDERATIONS IN PREDICTING METAL TOXICITY
Results of the short- and long-term laboratory and field experiments conducted to date
using sediments spiked with individual metals and mixtures of metals represent convincing
support of the conclusion that an absence (but not necessarily a presence) of metal toxicity can
be reliably predicted based upon metalrsulfide relationships and/or interstitial water metal
concentrations. In contrast, much confusion exists in the use of this convincing evidence to
interpret the significance of metals concentrations in sediments from the field when toxicity
and benthic community structure measurements are available. In addition, the use of these
observations as a basis for predicting metal bioavailability, or deriving ESG, raises a number
of conceptual and practical issues related to sampling, analytical measurements and effects of
additional binding phases. Many of these were addressed by Ankley et al. (1994), though the
most salient to the proposed derivation of ESG are described below.
5.2 SAMPLING AND STORAGE
Accurate prediction of exposure of benthic organisms to metals is critically dependent
upon sampling appropriate sediment horizons at appropriate times. This is because of the
relatively high rates of AYS oxidation due to natural processes in sediments and the
requirement that oxidation must be avoided during sampling of sediments and interstitial
water. In fact it is this seemingly labile nature that has led some to question the practical
utility of using AYS as a basis for EqP-derived ESG for metals (Luoma and Carter, 1993;
Meyer et al., 1994). For example, there have been many observations of spatial (depth)
variations in AYS concentrations, most of which indicate that surficial AYS concentrations are
less than those in deeper sediments (Besser et al., 1996; Boothman and Helmstetter,1992;
Brumbaugh et al., 1994; Hansen et al., 1996b; Hare et al., 1994; Howard and Evans, 1993;
Leonard et al., 1996a; Liber et al., 1996). This likely is due to oxidation of AYS at the
sediment surface, a process that is enhanced by bioturbation (Peterson et al., 1996). In
addition to varying with depth, AYS can vary seasonally. For example, in systems where
overly ing water contains appreciable oxygen during cold weather months, AYS tends to
Draft for £181-100
-------
decrease, presumably due to a constant rate of oxidation of the AVS linked to a decrease in its
generation by sulfate-reducing bacteria (Herlihy and Mills, 1985; Howard and Evans, 1993;
Leonard et al., 1993). Because of potential temporal and spatial variability of AVS, it appears
that the way to avoid possible under-estimation of metal bioavailability is to sample the
biologically "active" zone of sediments at times when AVS might be expected to be present at
small concentrations. We recommend that at a minimum AVS and SEM measurements be
made using surficial (0-2.0 cm) sediments during the period from November to early May in
aerobic aquatic ecosystems. Minimum AVS concentrations may not always occur during cool-
weather seasons; for example, systems that become anaerobic during the winter can maintain
relatively large sediment AVS concentrations (Liber et al., 1996). Therefore, seasonal
measurements of AVS, SEM and interstitial metal concentrations may need to be determined.
Importantly, the biologically active zones of some benthic communities may be within only the
first few millimeters hi surface depth while in other communities the biologically active zones
may be up to a meter. So in order to ensure sufficient characterization, multiple sediment
horizons may require sampling of interstitial water, SEM and AVS to determine the potential
for exposure to metals.
The somewhat subjective aspects of these sampling recommendations have been of
concern. Recent research suggests that the transient nature of AVS may be over-stated relative
to predicting the fate of all metal-sulfide complexes in aquatic sediments. Observations from
the Duluth EPA laboratory made in the early 1990s indicated that AVS concentrations in
sediments contaminated by metals such as cadmium and zinc tended to be elevated over
concentrations typically expected in freshwater systems (G.T. Ankley, unpublished data). The
probable underlying basis for these observations did not become apparent, however, until a
recent series of spiking and metal-sulfide stability experiments. The field colonization study of
Liber et al. (1996) demonstrated a strong positive correlation between the amount of zinc
added to test sediments and the resultant concentration of AVS in the samples. In fact, the
initial design of their study attempted to produce test sediments with as much as five-tunes
more SEM^ (nomimal) than AVS; however, the highest measured SEMz,, /AVS ratio
achieved was only slightly larger than 1. Moreover, the expected surficial depletion and
seasonal variations hi AVS were unexpectedly low hi the zinc-spiked sediments. These
observations suggested that zinc sulfide, which comprised the bulk of AVS in the spiked
sediments, was more stable than the iron sulfide that presumably was the source of most of the
AVS hi the control sediments. The apparent stability of other metal sulfides versus iron
Draft for SAB 1-101
-------
sulfide also has been noted in laboratory spiking experiments with freshwater and saltwater
sediments (Boothman et al., 1996; De Witt et al., 1996; Hansen et al., 1996b; Leonard et al.,
1995; Peterson et al., 1996; Sibley et al., 1996).
In support of these observations, recent metal-sulfide oxidation experiments conducted
by Di Toro et al. (1996b) have confirmed that cadmium and zinc form more stable sulfide
solid phases than iron. If this is also true for sulfide complexes of copper, nickel and lead, the
issue of seasonal/spatial variations hi AVS becomes of less concern because most of the studies
evaluating variations in AVS have focused on iron sulfide (i.e., uncontaminated sediments).
Thus, further research concerning the differential stability of metal-sulfides, both from a
temporal and spatial perspective, is definitely warranted.
5.2.1 Sediments
At a minimum, sampling of the surficial 2.0 cm of sediment in between November and
early May is recommended. A sample depth of 2.0 cm is more appropriate for remediation
and monitoring. In some instances such as for dredging or where depths greater than 2 cm are
important than sample depths should be planned based on particular study needs. Sediments
can be sampled using dredges, grabs, or coring, but mixing of aerobic and anaerobic
sediments must be avoided because the trace metal speciation La the sediments will be altered
(See Bufflap and Allen, 1995 for detailed recommendations to limit sampling artifacts).
Coring is generally less disruptive, facilitates sampling of sediment horizons and limits
potential metal contamination and oxidation if sealed PVC core liners are used.
Sediments not immediately analyzed for AVS and SEM must be placed in sealed air-
tight glass jars and refrigerated or frozen. Generally, 50 ml or more of sediment should be
added to nearly fill the jar. If sediments are stored this way there will be little oxidation of
AVS even after several weeks. Sampling of the stored sediment from the middle of the jar
will further limit potential effects of oxidation on AVS. Sediments experiencing oxidation of
AVS during storage will become less black or grey if oxidized. Because the rate of metal-
sulfide oxidation is markedly less than that of iron sulfide, release of metal during storage is
unlikely.
Draft for SAB 1-102
-------
5.2.2 Interstitial Water
Several procedures are available to sample interstitial water in situ or ex situ. Carignan
et al. (1985) compared metals concentrations in interstitial water obtained by ex situ
centrifugation at 11,000 rpm followed by double filtration (0.45 //m and 0.2 or 0.03 /^m) and
in situ diffusion samplers with a 0.002 /^m interstitial size. For the metals of concern in this
guidelines document, concentrations of nickel and cadmium were equivalent using both
methods and concentrations of copper and zinc were higher and more variable using
centrifugation. They recommended the use of in situ dialysis for the study of trace
constituents hi sediments because of its inherent simplicity and the avoidance of artifacts that
can occur with the handling of sediments in the laboratory.
More recently Bufflap and Allen (1995) reviewed four procedures for the collection of
interstitial water for trace metals analysis. These included ex situ squeezing and centrifugation
and in situ dialysis and suction filtration. This paper should be read by those selecting a
interstitial water sampling method. They observed that each method has its own advantages
and disadvantages, and that each user must make their own choice given the inherent errors of
each method. Importantly, interstitial water must be extracted by centrifugation or squeezing
in an inert atmosphere until acidified because oxidation will alter metal speciation. Artifacts
may be caused by temperature changes in ex situ methods that may be overcome by
maintaining temperatures to those in situ. Contamination of interstitial water by fine particles
is important in all methods as differentiation of paniculate and dissolved metal is a function of
interstitial size. The use of 0.45 pm filtration, while an often accepted definition of
dissolved, may result hi laboratory to laboratory discrepancies. The use of suction filtration
devices is limited to coarser sediments, and they do not offer depth resolution. The use of
diffusion samplers is hampered by the tune required for equilibrium (7-14 days) and the need
for diver placement and retrieval in deep waters. Acidification of interstitial water obtained
by diffusion or from suction filtration must occur immediately to limit oxidation. Bufflap and
Allen (1995) conclude that in situ techniques have less potential for producing sampling
artifacts than ex situ procedures. They concluded that of the in situ procedures, suction
filtration has the best potential for producing artifact free interstitial water samples directly
from the environment. Of the ex situ procedures they concluded that centrifugation under a
nitrogen atmosphere followed immediately by filtration and acidification was the simplest
technique likely to result in an unbiased estimate of metal concentrations in interstitial water.
Draft for SAB 1-103
-------
At present, EPA recommends filtration of the surface water through 0.4 to 0.45 ^
polycarbonate filters to better define that fraction of aqueous metal associated with toxicity
(Prothro, 1993). Thurman equates the organic carbon retained on a 0.45 micrometer glass-
fiber filter to suspended organic carbon so that this filtration procedure under nitrogen
atmosphere followed immediately by acidification is acceptable for interstitial waters.
However in studies comparing collection and processing methods for trace metals, sorption to
filter membranes or filtering apparatus has been identified when losses occur (Ozretich and
Schults, 1998). Ozretich and Schults, 1998 have recently presented a method combining
longer centrifugation times with a unique single-step IW withdrawal procedure which has the
potential for minimizing metal losses by eliminating the need for filtration.
In contrast to the above recommendations, EPA recommends the use of dialysis
samplers to obtain samples of interstitial water for comparison of measured concentrations of
dissolved metals with WQC. This is primarily because diffusion samplers obtain interstitial
water with the proper in situ geochemistry thus limiting artifacts of ex situ sampling. Further,
EPA has found that in shallow waters where contamination of sediments is most likely,
placement of diffusion samplers is easily accomplished and extended equilibration times are
not a problem. Secondly, EPA recommends the use of centrifugation under nitrogen and
double 0.45/^m filtration using polycarbonate filters for obtaining interstitial water from
sediments in deeper aquatic systems. Probably most importantly, the extremely large database
comparing interstitial metals concentrations with organism responses from spiked and field
sediment experiments in the laboratory has demonstrated that, where the interstitial water toxic
unit concept predicted that metals concentrations in interstitial water should not be toxic,
toxicity was not observed when either dialysis samplers or centrifugation were used (Berry et
al., 1996; Hansen et al.f 1996a). Therefore, it is likely that when either methodology is used
to obtain interstitial water for comparison with WQC, if metals concentrations are below 1.0
IWGTU sediments should be acceptable for protection of benthic organisms.
5.3 ANALYTICAL MEASUREMENTS
An important aspect to deriving "global" ESG values is that the methods necessary to
implement the approach must be reasonably standardized or have been demonstrated to
produce results that are comparable to those of standard methodologies. From the standpoint
of the proposed metal ESG, a significant amount of research has gone into defining
Draft for SAB 1-104
-------
methodologies to obtain interstitial water and sediments (see Section 5-2 above), to extract
SEM and AVS from sediments, and to quantify AVS, SEM and the metals in interstitial water.
X
5.3.1 Acid Volatile Sulfide
The SEM/AVS extraction method recommended by EPA is that of Allen et al. (1993).
In terms of AVS quantification, a number of techniques have been successfully utilized
including gravimetric (Di Toro et al., 1990; Leonard et al., 1993), colorimetric (Corawell and
Morse, 1987), gas chromatography - photoionization detection (Casas and Crecelius, 1994;
Slotton and Reuter, 1995) and specific ion electrodes (Boothman and Helmstetter,1992;
Brouwer and Murphy, 1994; Brumbaugh et al., 1994; Leonard et al., 1996b). Allen et al.,
1993 report a limit of detection for 50% accuracy of 0.01 ^mol/g for a 10-g sediment sample
using the colorimetric method. Based on several studies Boothman reports a detection limit of
1 fj.tnol AVS which translates to 0.1 //mol/g dry weight for a 10 g sediment sample using the
ion specific electrode method (personal communication).
5.3.2 Simultaneously Extracted Metal
Simultaneously extracted metals are operationally-defined as metals extracted from
sediment into solution by the acid volatile sulfide extraction procedure. The "dissolved"
metals in this solution are also operationally defined as the metal species which pass through
filter material used to remove the residual sediment, and thus are defined by the interstitial size
of the filtration material used. Common convention defines "dissolved" as metal species
<0.45-/un in size. SEM concentrations measured in sediments are not significantly different,
however, using Whatman 1 filter paper alone (< ll-/xm nominal interstitial size) or in
combination with a 0.45-/*m filter (W. Boothman, unpublished data). SEM solutions
generated by the AVS procedure can be analyzed for metals, commonly including cadmium,
copper, lead, nickel, silver and zinc by routine atomic spectrochemical techniques appropriate
for environmental waters (e.g. inductively coupled plasma atomic emission or graphite furnace
atomic absorption spectrophotometry) (U.S. EPA, 1994b). Because of the need to determine
metals at relatively low concentrations, additional consideration must be given to preclude
contamination during collection, transport and analysis (U.S. EPA, 1995d,e,f,g).
Draft for SAB 1-105
-------
5.3.3 Interstitial Water Metal
Interstitial water can be analyzed for the metals cadmium, copper, lead, nickel, silver
and zinc by routine atomic spectrochemical techniques appropriate for environmental waters
(e.g. inductively coupled plasma atomic emission or graphite furnace atomic absorption
spectrophotometry) (U.S. EPA, 1994b). Because of the need to determine metals at
concentrations at or below the threshold of biological effects (i.e., WQC concentrations);
additional consideration must be given to preclude contamination during collection, transport
and analysis (U.S. EPA, 1995d,e,f,g). (See guidance on clean chemistry techniques in U.S.
EPA, 1994c.)'Generally, detection limits should be at <;0.1 IWGTU, because the toxic unit
contributions of each of the metals must be summed.
i
5.4 ADDITIONAL BINDING PHASES
Although AYS is an important binding phase for metals, there clearly are other
physico-chemical factors that influence metal partitioning in sediments. In aerobic systems, or
those with low productivity (i.e., where the absence of organic carbon limits sulfate
reduction), AYS plays little or no role in determining interstitial water concentrations of
metals. For example, Leonard et al. (1996a) found that a relatively large percentage of
surficial sediments from open areas in Lake Michigan did not contain detectable AYS. In fact
the great majority (42 of 46) of samples analyzed by Leonard et al. (1996a) contained less
AYS than SEM, yet interstitial water metal concentrations of cadmium, copper, nickel, lead
and zinc were consistently small or non-detectable. Even in sediments where concentrations of
AYS are significant, other partitioning phases may provide additional binding capacity for
SEM (e.g., Ankley et al., 1993; Calamono et al., 1990; Slotton and Reuter, 1995). In aerobic
sediments both organic carbon and iron and manganese oxides control interstitial water
concentrations of metals (Calamono et al,. 1990; Jenne, 1968; Luoma and Bryan, 1981;
Tessier et al., 1979). In anaerobic sediments, organic carbon appears to be an important
additional binding phase controlling metal partitioning, in particular for cadmium, copper and
lead (U.S. EPA, 1994a).
Even in substrates with very little metal binding capacity (e.g., chromatographic sand),
surface adsorption associated with cation exchange capacity will control interstitial water metal
concentrations to some degree (Hassan et al., 1996). Although an ideal ESG model for metals
Draft for SAB 1-1Q6
-------
would incorporate all possible metal binding phases, current knowledge concerning
partitioning/capacity of phases other than AVS is insufficient for practical application of a
multiple phase model for deriving ESG hi this sediment guidelines document.
5.5 PREDICTION OF THE RISKS OF METALS IN SEDIMENTS BASED ON EqP
It is important to repeat that conclusions about sediment toxicity based on SEM-AVS
concentrations pertain only to cadmium, copper, lead, nickel, silver and zinc. (1) When the
molar concentration of AVS exceeds that of SEM (negative SEM-AVS) sediment toxicity due
to these metals is unlikely and any observed toxicity is most likely from some other cause.
This is important because toxicity observed in sediments having an excess of AVS is often
incorrectly assumed to disprove the EqP metals theory. The correct conclusion is that some
factor other than metals caused the effect. This can be further substantiated if the toxic unit
concept is applied to metal concentrations measured hi interstitial water; the absence of
significant concentrations of metals coupled with the negative SEM-AVS are powerful
evidence that metals are an unlikely cause of the effect. (2) Sediments can only be toxic from
the metals cadmium, copper, lead, nickel, silver and zinc when the molar concentrations of
SEM exceed those of AVS (SEM-AVS differences are positive). Measurements of interstitial
water concentrations of metals are invaluable hi demonstrating that the sediments are toxic
because of metals, and these measurements will provide insights into the specific metal(s)
causing the observed toxicity. (3) It is not uncommon for toxicity to be absent in sediments
having concentrations of SEM that exceed those of AVS (SEM-AVS is positive). This is
because other metal binding phases hi sediments often reduce the concentrations of
bioavailable metal. (4) When sediments are toxic, and SEM-AVS is greater than 0.0, the
toxicity may or may not be metals-related. Often sediments having SEM-AVS of up to 10
Mmoles SEM/g are not toxic because the excess metals are associated with other binding
phases. Measurements of interstitial water concentrations of metals are invaluable hi
demonstrating an absence or presence of bioavailable metal.
Draft for SAB 1-107
-------
SECTION 6
GUIDELINES STATEMENT
The procedures described in this document indicate that, except possibly where a
locally important species is very sensitive, benthic organisms should be acceptably protected in
freshwater and saltwater sediments if any one or both of the following two conditions is
satisfied: (a) If the sum of the molar concentrations of SEM cadmium, copper, lead, nickel,
silver and zinc is less than or equal to the molar concentration of AYS or (b) the sum of the
dissolved interstitial water concentration of cadmium, copper, lead, nickel, silver and zinc
divided by their respective WQC is less than or equal to 1.0.
(a) AVS Guidelines:
i [SEN1MAVS] (4.5)
where
£ ,[SEMJ = [SEMcJ + [SEMcJ + [SEMpJ + [SEMNJ + [SEMzJ + [l/2SEMAg]
(b) Interstitial Water Guidelines
l <4-8>
where
[M.d]
,] [FCV^] [FCV^] [FCVpbd] [FCVN.d] [FCV^] [FCV^]
If any one of these two conditions are violated, this does not mean that the sediment violates the
ESG and is unacceptable. For example, if SEM exceeds AVS, or if the AVS in a sediment is non-
detectable, then condition (a) will be violated. However, if there is sufficient sorption to particles,
Draft for SAB 1-108
-------
or organic carbon or other binding phases so that condition (b) is satisfied, then the sediment
meets the guideline and benthic organisms are acceptably protected from metals-induced sediment
toxicity.
If both of these conditions are violated, or if the AVS Guideline is violated and the
sediment is contaminated with silver then there'is reason to believe that the sediment may be
unacceptably contaminated by these metals. Further testing and evaluations would, therefore, be
useful hi order to assess actual toxicity and its causal relationship to the five metals. These may
include acute and chronic tests with species that are sensitive to the metals suspected to be causing
the toxicity. Also, in situ community assessments, sediment TIEs and seasonal characterizations
of the SEM, AVS and interstitial water concentrations would be appropriate (Ankley et al., 1994).
The ESG approaches are intended to protect benthic organisms from direct toxicity
associated with exposure to metal-contaminated sediments. They are not designed to protect
aquatic systems from metal release associated, for example, with sediment suspension, or the
transport of metals into the food web either from sediment ingestion or the ingestion of
contaminated benthos. This latter issue, hi particular, should be the focus of future research given
existing uncertainty in the prediction of bioaccumulation of metals by benthos (Ankley, 1996).
It is repeated here that these guidelines apply only to the six metals discussed in this
document, copper cadmium, lead, nickel, zinc and silver. Procedures for sampling and analytical
methods for interstitial water and sediments are discussed hi Section 5, Implementation.
Draft for SAB 1-109
-------
SECTION 7
REFERENCES
Adams, W.J., R.A. Kimerle and R.G. Mosher. 1985. Aquatic safety assessment of chemicals
sorbed to sediments. In Aquatic Toxicology and Hazard Assessment: Seventh Symposium.
R.D. Cardwell, R. Purdy and R.C. Banner, Eds. STP 854. American Society for Testing
and Materials, Philadelphia, PA, USA, pp. 429-453.
Adams, D.A., J.S. O'Connor, and S.B. Weisberg. 1996. Draft Final Report Sediment Quality of
the NY/NJ Harbor System, An Investigation under the Regional Environmental Monitoring
and Assessment Program (R-EMAP). Monitoring and Assessment Branch, Division of
Environmental Science and Assessment. United States Environmental Protection Agency,
Region 2. Edison, New Jersey.
Allen, H.E., R.H. Hall and T.D. Brisbin. 1980. Metal speciation effects on aquatic toxicity.
Environ. Sci. Technol. 14:441-443.
Allen, H.E., G. Fu and B. Deng. 1993. Analysis of acid-volatile sulfide (AVS) and
simultaneously extracted metals (SEM) for the estimation of potential toxicity hi aquatic
sediments. Environ. Toxicol. Chem.l2:l-13.
Anderson, D.M. and F.M.M. Morel. 1978. Copper sensitivity of Gonyaulax tamarensis.
Limnol. Oceanogr. 23:283-295.
Anderson, L.D., D.B. Kent, et al. 1994. Batch experiments characterizing the reduction of
Cr(VI) using suboxic material from a mildly reducing sand and gravel aquifer. Environ.
Sci. Technol. 28:178-185.
Ankley, G.T., G.L. Phipps, E.N. Leonard, D.A. Benoit, V.R. Mattson, P.A. Kosian, A.M.
Cotter, J.R. Dierkes, D.J. Hansen and J.D. Mahony. 1991. Acid volatile sulfide as a
factor mediating cadmium and nickel bioavailability hi contaminated sediments. Environ.
Toxicol. Chem. 10:1299-1307.
Draft for SAB\-\ 10
-------
Ankley, G.T., V.R. Mattson, E.N. Leonard, C.W. West and J.L. Bennett. 1993. Predicting the
acute toxicity of copper in freshwater sediments: Evaluation of the role of acid-volatile
sulfide. Environ. Toxicol. Chem. 12:315-320.
Ankley, G.T., D.M. Di Toro, D.J. Hansen, J.C. Mahony, R.C. Swartz, R.A. Hoke, N.A.
Thomas, A.W. Garrison, H.E. Allen and C.S. Zarba. 1994. Assessing the potential
bioavailability of metals in sediments: a proposed approach. Environ. Management
18:331-337.
Ankley, G.T. 1996. Evaluation of metal/acid volatile sulfide relationships in the prediction of
metal bioaccumulation by bentbic macroinvertebrates. Environ. Toxicol. Chem. 15:2138-
2146.
Ankley, G. T., D.M. Di Toro, D.J. Hansen and WJ. Berry. 1996. Technical basis and proposal
for deriving sediment quality criteria for metals. Environ. Toxicol. Chem. 15:2056-2066.
Barrow, N.J. and A.S. Ellis. 1986a. Testing a mechanistic model, m. The effects of pH on
fluoride retention by a soil. J. Soil Sci. 37:287-295.
Barrow, N.J. and A.S. Ellis. 1986b. Testing a mechanistic model. TV. Describing the effects
of pH on zinc retention by a soil. J. Soil Sci. 37:295-302.
Barrow, N.J. and A.S. Ellis. 1986c. Testing a mechanistic model. V. The points of zero salt
effect for phosphate retention and for acid/alkali titration of a soil. J. Soil Sci. 37:303-
310.
Bergman, H.L. and E.J. Dorward-King. 1997. Reassessment of Metals Criteria for Aquatic Life
Protection. Society of Environmental Toxicology and Chemistry (SETAC) Press:
Pensacola, FL.
Berner, R.A. 1967.- Thermodynamic stability of sedimentary iron sulfides. Am. J. Sci. 265:773-
785.
Draft for £451-111
-------
Berry, W.J., D.J. Hansen, J.D. Mahony, D.L. Robson, D.M. Di Toro, B.P. Shipley, B. Rogers,
J.M. Corbin and W.S. Boothman. 1996. Predicting the toxicity of metals-spiked
laboratory sediments using acid-volatile sulfide and interstitial water normalization.
Environ. Toxicol. Chem. 15:2067-2079.
Berry, W. et al. 1999. Predicting the toxicity of sediments spiked with silver. Environ. Toxicol.
Chem. 18:40-48.
Berry, WJ. and W.S. Boothman. 1999. Results of toxicity tests with Ampelisca abdita exposed
to chromium-spiked sediments. Memorandum to the record. February 17, 1999.
Besser, J.M., C.G. Ingersoll and J.P. Giesy. 1996. Effects of spatial and temporal variability of
acid volatile sulfide on the bioavailability of copper and zinc in freshwater sediments.
Environ. Toxicol. Chem. 15:286-293.
Biesinger, K.E., G.M. Christensen and J.T. Fiandt. 1986. Effects of metal salt mixtures on
Daphnia magna reproduction. Ecotoxicol. Environ. Safety. 11:9-14.
Blake, R.C., D.M. Choate, et al. 1993. Chemical transformation of toxic metals by a
pseudomonas strain from a toxic waste site. Environ. Toxicol. Chem. 12:1365-1376.
Boothman, W.S. and A. Helmstetter. 1992. Vertical and seasonal variability of acid volatile
sulfides in marine sediments. Research Report to the U.S. EPA Environmental Monitoring
and Assessment Program, Narragansett, RI, USA.
Boothman, W., D. Hansen, W. Berry, D. Robson, A. Helmstetter and J. Corbin. 1996. Abstract:
Variation of acid volatile sulfides and metals and biological response in field-exposed
sediments. Society of Environmental Toxicology and Chemistry Annual Meeting.
Washington, D.C.
Boulegue, J., C.J. Lord in and T.M. Church. 1982. Sulfur speciation and associated trace metals
(Fe, Cu) in the interstitial waters of Great Marsh, Delaware. Geochim. Cosmochim. Acta.
46:453-464.
Draft for SAB 1-112
-------
Brouwer, H. and T.P. Murphy. 1994. Diffusion method for the determination of acid-volatile
sulfides (AVS) in sediment. Environ. Toxicol. Chem. 13:1273-1275.
Brumbaugh, W.G., C.G. Ingersoll, N.E. Kemble, T.W. May and J.L. Zajcek. 1994. Chemical
characterization of sediments and interstitial water from the upper Clark Fork River and
Milltown Reservoir, Montana. Environ. Toxicol. Chem. 13:1971-1983.
Buerge, I.J. and SJ. Hug. 1997. Kinetics and pH dependence of chromium (VI) reduction by
iron(n). Environ. Sci. Technol. 31:1426-1432.
Bufflap, W.E. and H.E. Allen. 1995. Sediment interstitial water collection methods: A review.
Water Res. 29: 65-177. Calder, E.R. Long and C.G. Ingersoll. 1996. Development and
evaluation of sediment quality guidelines for Florida coastal waters. Ecotoxicology, In
Press.
Calamono, W., W. Ahlf and U. Forstner. 1990. Exchange of heavy metals between sediment
components and water. In Metal Speciation hi the Environment. J.A.C. Broekart, S.
Gucer and F. Adams, Eds. Springer-Verlag, Berlin, Germany, pp. 503-522.
Call, D.J. et al. 1999. Silver toxicity to Chironomus tentans hi two freshwater sediments.
Environ. Toxicol. Chem. 18:30-39.
Campbell, P. G. C. 1995. Interactions between trace metals and aquatic organisms: a critique of the
free-ion activity model. Metal Speciation and Bioavailability in Aquatic Systems. John
Wiley & Sons.
Carignan, R. 1984. Interstitial water sampling by dialysis: methodological notes. Limnol.
Oceanogr. 20:667-670.
Carignan, R., F. Rapin and A. Tessier. 1985. Sediment interstitial water sampling for metal
analysis: a comparison of techniques. Geochim. Cosmochim. Acta. 49:2493-2497.
Carlson, A.R., G.L. Phipps, V.R. Mattson, P.A. Kosian and A.M. Cotter. 1991. The role of
acid volatile sulfide hi determining cadmium bioavailability in sediments. Environ.
Draft for SAB 1-113
-------
Toxicol. Chem. 10:1309-1319.
Casas, A.M. and E.A. Crecelius. 1994. Relationship between acid volatile sulfide and the
toxicity of zinc, lead and copper in ma rine sediments. Environ. Toxicol. Chem. 13:529-
536.
Chapman, G.A. 1987. Establishing sediment criteria for chemicals - regulatory perspective. In
Fate and Effects of Sediment-Bound Chemicals in Aquatic Systems. K.L. Dickson, A.W.
Maki and W.A. Brungs, Eds. Pergamon Press, New York. pp. 355-376.
Cornwell, J.C. and J.W. Morse. 1987. The characterization of iron sulfide minerals in anoxic
marine sediments. Mar. Chem. 22:193-206.
Cranston, R.E. and J.W. Murray. 1978. The determination of chromium species in natural
waters. Anal. Chim. Acta. 99:275-282.
Davies-Colley, R.J\, P.O. Nelson and K.J. Williamson. 1985. Sulfide control of cadmium and
copper concentrations in anaerobic estuarine sediments. Mar. Chem. 16:173-186.
Deng, B.L. and A.T. Stone. 1996. Surface-catalyzed chromium (VI) reduction: reactivity
comparisons of different organic reductants and different oxide surfaces. Environ. Sci.
Technol. 302484-2494.
De Witt, T.H., R.C. Swartz, DJ. Hansen, W. J. Berry and D. McGovern. 1996. Interstitial
metal and acid volatile sulfide predict the bioavailability of cadmium during a full life-
cycle sediment toxicity test using the estuarine amphipod, Leptocheirus plumulosus.
Environ. Toxicol. Chem. 15:2095-2101.
Di Toro, D.M., F. Harrison, E. Jenne, S. Karickoff, and W. Lick. 1987. Synopsis of Discussion
session 2: Environmental fate and compartmentalization. In Fate and Effects of Sediment-
Bound Chemicals in Aquatic Systems. K. L. Dickson, A.W. Maki and W.A. Brungs,
Eds., Pergamon, New York, NY, USA, pp. 136-147.
Di Toro, D.M., J.J. Mahony, D.J. Hansen, K.J. Scott, M.B. Hicks, S.M. Mayr and M.S.
Draft for SAB 1-114
-------
Redmond. 1990. Toxicity of cadmium in sediments: the role of acid volatile sulfide.
Environ. Toxicol. Chem. 9:1487-1502.
DiToro, D.M. et al. 1991. Technical basis for the equilibrium partitioning method for
establishing sediment quality criteria. 12.
DiToro, D.M. et al. 1991. Technical basis for the equilibrium partitioning method for
establishing sediment quality criteria. Environ. Toxicol. Chem. 11(12): 1541-1583.
Di Toro, D.M., C. Zarba, D.J. Hansen, R.C. Swartz, C.E. Cowan, H.E. Allen, N.A. Thomas,
P.R. Paquin and W.J. Berry. 1991. Technical basis for establishing sediment quality
criteria for nonionic organic chemicals using equilibrium partitioning. Environ. Toxicol.
Chem. 10:1541-1583.
Di Toro, D.M., 1992. Personal communication relating to Di Toro et al., 1990 on titration of
FeS by Cd2+.
Di Toro, D.M., J.D. Mahony, D.J. Hansen, K.J. Scott, A.R. Carlson and G.T. Ankley. 1992.
Acid volatile sulfide predicts the acute toxicity of cadmium and nickel in sediments.
Environ. Sci. Technol. 26:96-101.
Di Toro, D.M., J.D. Mahony, and A.M. Gonzalez. 1996a Particle oxidation model of synthetic
FeS and sediment acid volatile sulfide. Environ. Toxicol. Chem. 15:2156-2167.
Di Toro, D.M., J.D. Mahony, D.J. Hansen and W.J. Berry. 1996b. A model of the oxidation of
iron and cadmium sulfide in sediments. Environ. Toxicol. Chem. 15:2168-2186.
Durrani, P.J. and B. Durrant. 1970. Introduction to advanced inorganic chemistry. London,
Longman.
Dzombak, D.A. and F.M.M. Morel. 1990. Surface Complexation Modeling, Hydrous Ferric
Oxide. New York, NY, John Wiley & Sons.
Eary, L.E. and D. Rai. 1987. Kinetics of chromium (JH) oxidation to chromium (VI) by reaction
Draft for SAB 1-115
-------
with manganese dioxide. Environ. Sci. Tech. 21:1187-1193.
Eary, L.E. and D. Rai. 1989. Kinetics of chromate reduction by ferrous ions derived from
hematite and biotite at 25 °C. Am. J. Sci. 289:180-213.
•
Eary, L.E. and D. Rai. 1991. Chromate reduction by subsurface soils under acidic conditions.
Soil Sci. Soc. Am. J. 55:676-683.
Elovitz, M.S. and W. Fish. 1995. Redox interactions of Cr (VI) and substituted phenols:
products and mechanism. Environ. Sci. Technol. 29:1933-1943.
Emerson, S., L. Jacobs and B. Tebo. 1983. The behavior of trace metals in marine anoxic
waters: solubilities at the oxygen-hydrogen sulfide interface. In Trace Metals in Sea
Water. C. S. Wong.
Enserink, E.L., J.L. Maas-Diepeveen and C.S. van Leeuwen. 1991. Combined toxicity of
metals: an ecotoxicological evaluation. Water Res. 25:679-687.
Fendorf, S.E. and RJ. Zasoski. 1992. Chromium (HI) oxidation by delta-
MnO2.1.characterization. Environ. Sci. Technol. 26:79-85.
Fendorf, S.E., G.M. Lamble, et al. 1994. Mechanisms of chromium (HI) sorption on silica, l.cr
(HI) surface structure derived by extended x-ray absorption fine structure spectroscopy.
Environ. Sci. Technol. 28:284-289.
Fendorf, S.E. and D.L. Sparks. 1994. Mechanisms of chromium (HI) sorption on silica.2. effect
of reaction conditions. Environ. Sci. Technol. 28:290-297.
Fendorf, S.E. and G.C. Li. 1996. Kinetics of chromate reduction by ferrous iron. Environ. Sci.
Technol. 30:1614-1617.
Goldhauber, M.B. and I.R. Kaplan. 1974. The sulfur cycle. In The Sea, Vol. 5, Marine
Chemistry. E.D. Goldberg, Ed. J. Wiley & Sons, NY.,'USA, pp. 569-655.
Draft for SAB 1-116
-------
Green, A.S., G.T. Chandler and E.R. Blood. 1993. Aqueous-, interstitial water-, and sediment-
phase cadmium: toxicity relationships for a meiobenthic copepod. Environ. Toxicol.
Chem. 12:1497-1506.
Hansen, D.J. 1995. Memorandum to Catherine Fox. Subject: Assessment tools that can be used
for the National Sediment Inventory. February 28, 1994.
Hansen, D.J., W.J. Berry, J.D. Mahony, W.S. Boothman, D.L. Robson, G. T. Ankley, D. Ma,
Q Yan and C.E. Pesch. 1996a. Predicting toxicity of metal-contaminated field sediments
using interstitial water concentrations of metals and acid-volatile sulfide normalization.
Environ. Toxicol. Chem. 15:2080-2094.
Hansen, D.J., J.D. Mahony, W.J. Berry, S.J. Benyi, J.M. Corbin, S.D. Pratt and M.B. Able.
1996b. Chronic effect of cadmium hi sediments on colonization by benthic marine
organisms: an evaluation of the role of interstitial cadmium and acid-volatile sulfide in
biological availability. Environ. Toxicoi. Chem. 15:2126-2137.
Hare, L., R. Carignan and M.A. Huerta-Diaz. 1994. An experimental study of metal toxicity
and accumulation by benthic invertebrates: implications for the acid volatile sulfide (AVS)
model. Limnol. Oceanogr. 39:1653-1668.
Hassan, S.M., A.W. Garrison, H.E. Allen, G.T. Ankley and D.M. Di Toro. 1996. Using the
distribution coefficients of five trace metals in sandy sediments to select the lowest capacity
natural adsorbent. Environ. Toxicol. Chem. 15:2198-2208.
Herlihy, A.T. and A.L. Mills. 1985. Sulfate reduction in freshwater sediments receiving acid
mine drainage. Appl. Environ. Microbiol. 49:179-186.
Hesslein, R.H. 1976. An hi situ sampler for close interval interstitial water studies. Limnol.
Oceanogr. 30:221-227.
Hoke, R.A., G.T. AnkleyjA.M. Cotter, T. Goldenstein, P.A. Kosian, G.L. Phipps and P.M.
VanderMeiden. 1994. Evaluation of equilibrium partitioning theory for protecting acute
toxicity of field-collected sediments contaminated with DDT, DDE, and DDD to the
Draft for SAB 1-117
-------
amphipod Hyalella azteca. Environ. Toxicol. Chem. 13:157-166.
Howard, D.E. and R.D. Evans. 1993. Acid-volatile sulfide (AVS) in a seasonally anoxic
mesotrophic lake: seasonal and spatial changes in sediment AVS. Environ. Toxicol.
Chem. 12:1051-1058.
Hug, S.J., H.U. Laubscher, et al. 1997. Iron (HI) catalyzed photochemical reduction of
chromium (VI) by oxalate and citrate in aqueous solutions. Environ. Sci. Technol.
31:160-170.
Ingersoll, C.G., P.S. Haverland, E.L. Brunson, T.S. Canfield, F.S. Dwyer, C.E. Henke and
N.E. Kemble. 1996. Calculation of sediment effect concentrations. J. Great Lakes Res.
In Press.
Jenne, E.A. 1968. Controls on Mn, Fe, Co, Ni, Cu, and Zn concentrations in soils and water-
the significant role of hydrous Mn and Fe oxides. In Advances in Chemistry. American
Chemical Society, Washington, DC, USA, pp. 337-387.
Jenne, E.A. 1977. Trace element sorption by sediments and soil — sites and processes. In
Symposium on Molybdenum in the Environment, Vol. 2. W. Chappell and K. Petersen,
Eds. M.Dekker, Inc., New York. (pp. 425-553).
X1 i
Jenne, E.A., D.M. Di Toro, H.E. Allen and C.S. Zarba. 1986. An activity-based model for
developing sediment criteria for metals: a new approach. Chemicals in the environment.
International Conference, Lisbon, Portugal.
Johnson, C.A. and A.G. Xyla. 1991. The oxidation of chromium (HI) to chromium (VI) on the
surface of manganite (gamma-MnOOH). Geochim. Cosmochim. Acta. 55:2861-2866.
Kaczynski, S.E. and R.J. Kleber. 1994. Hydrophobic C18 bound organic complexes of chromium
and their potential impact on the geochemistry of chromium in natural waters. Environ.
Sci. Technol. 28:799-804. v
Kemble, N.E. et al. 1994. Toxicity of metal-contaminated sediments from the upper Clark Fork
Draft for SAB 1-118
-------
River, Montana, to aquatic invertebrates and fish in laboratory exposures. Environ.
Toxicol. Chem. 13:1985-1997.
Kemp, P.P. and R.C. Swartz. 1986. Acute toxicity of interstitial and particle-bound cadmium to
a marine infaunal amphipod. Mar. Environ. Res. 26:135-153.
Kraak, M.H.S., D. Lavy, H. Schoon, M. Toussaint, W.H.M. Peelers and N.M. van Straalen.
1994. Ecotoxicity of mixtures of metals to the zebra mussel Dreissena polymorpha.
Environ. Toxicol. Chem. 13:109-114.
Landers, A.H., M.B. David and M.J. Mitchel. 1983. Analysis of organic and inorganic sulfur
constituents in sediments, soils, and water. Int. J. Environ. Anal. Chem. 14:245-256.
Leonard, E.N., V.R. Mattson, D.A. Benoit, R.A. Hoke and G.T. Ankley. 1993. Seasonal
variation of acid volatile sulfide concentration in sediment cores from three northeastern
Minnesota lakes. Hydrobiologia. 271:87-95.
Leonard, E.N., V.R. Mattson and G.T. Ankley. 1995. Horizon-specific oxidation of acid
volatile sulfide hi relation to the toxicity of cadmium spiked into a freshwater sediment.
Arch. Environ. Contain. Toxicol. 28:78-84.
Leonard, E.N., G.T. Ankley and R.A. Hoke. 1996a. Evaluation of metals in marine and
freshwater surficial sediments from the Environmental Monitoring and Assessment
Program relative to proposed sediment quality criteria for metals. Environ. Toxicol.
Chem. 15:2221-2232.
Leonard, E.N., A.M. Cotter and G.T. Ankley. 1996b. Modified diffusion method for analysis
of acid volatile sulfides and simultaneously extracted metals in freshwater sediment.
Environ. Toxicol. Chem. In Press.
Liber, K., D. Call,,T. Markee, K. Schmude, M. Balcer, F. Whiteman and G.T. Ankley. 1996.
Effects of acid volatile sulfide on zinc bioavailability and toxicity to benthic
macroinvertebrates in spiked-sediment field experiments. Environ. Toxicol. Chem.
15:2113-2125.
Draft for SAB 1-119
-------
Long, E.R. and L.G. Morgan. 1991. The potential for biological effects of sediment-sorbed
contaminants tested in the national status and trends program. NOAA Technical
Memorandum NOS OMA 52.
Long, E.R. et al. 1995. Incidence of adverse biological effects within ranges of chemical
concentrations in marine and estuarine sediments. Environ. Management. 19(l):81-97.
Long, E.R., D.A. Wolfe, K.J. Scott, G.B. Thursby, E.A. Stem, C. Peven and T. Schwartz.
1995. Magnitude and Extent of Sediment Toxicity in the Hudson-Raritan Estuary.
National Oceanic and Atmospheric Administration Technical Memorandum NOS ORCA
88 NOAA/NOS Office of Ocean Resources, Conservation, and Assessment, Silver Spring,
MD USA.
Long, E.R., G.M. Sloane, R.S. Carr, K.J. Scott, G.B. Thursby and T.L. Wade. 1996. Sediment
Toxicity in Boston Harbor: Magnitude and Extent, and Relationships with Chemical
Toxicants. National Oceanic and Atmospheric Administration Technical Memorandum
NOS ORCA 88 NOAA/NOS Office of Ocean Resources, Conservation, and Assessment,
Silver Spring, MD USA. (REMAP)
Long, E.R., L.J. Field and D.D. MacDonald. 1998. Predicting toxicity in marine sediments
with numerical sediment quality guidelines. Environ. Toxicol. Chem. 17(4):714-727.
Luoma, S.N. 1989. Can we determine the biological availability of sediment-bound trace
elements? Hydrobiologia 176/177:379-396.
Luoma, S.N. and W. Bryan. 1981. A statistical assessment of the form of trace metals in
oxidized sediments employing chemical extractants. Sci. Total Environ. 17:165-196.
Luoma, S.N. and J.L. Carter. 1993. Understanding the toxicity of contaminants in sediments:
beyond the bioassay-based paradigm. Environ Toxicol. Chem. 12:793-796.
MacDonald, D.D., R.S. Carr, F.D. Calder, E.R. Long and C.G. Ingersol. 1996. Development
and evaluation of sediment quality guidelines for Florida coastal waters. Ecotoxicology
5:253-278.
Draft for SAB 1-120
-------
Mahony, J.D., D.M. Di Toro, R. Koch, W. Berry and D. Hansen. 1993. Vertical distribution of
AVS and SEM in bedded sediments, biological implications and the role of metal sulfide
oxidation kinetics. Abstract. Annual Meeting. Society of Environmental Toxicology and
Chemistry Annual Meeting, p.46.
Mahony, J.D., D.M. Di Toro, A.M. Gonzalez, M. Curto and M. Dilg. 1996. Partitioning of
metals to sediment organic carbon. Environ. Toxicol. Chem. 15:2187-2197.
Masschelyeyn, P.H., J.H. Pardue, R.D. DeLauna and W.H. Patrick, Jr. 1992. Chromium redox
chemistry in a lower Mississippi Valley bottomland. Environ. Sci. Technol. 26:1217-1226.
Mesuere, K. and W. Fish. 1992. Chromate and oxalate adsorption of goethite. 1 .calibration of
surface complexation models. Environ. Sci. Technol. 26:2357-2364.
Mesuere, K. and W. Fish. 1992. Chromate and oxalate adsorption of goethite.2.surface
complexation modeling of competitive adsorption. Environ. Sci. Technol. 26:2365-2370.
Meyer, J.S., W. Davison, B. Sundby, J.T. Oris, D.J. Lauren, U. Forsmer, J. Hong and D.G.
Crosby. 1994. The effects of variable redox potentials, pH and light on bioavailability in
dynamic water-sediment environments. In Bioavailability: Physi-cal, Chemical and
Biological Interactions.
Morse, J.W., F.J. Millero, J.C. Cornwell and D. Rickard. 1987. The chemistry of the hydrogen
sulfide and iron sulfide systems in natural waters. Earth Sci. Rev. 24:1-42.
Oakley, S.M., K.J. Williamson and P.O. Nelson. 1980. The geochemical partitioning and
bioavailability of trace metals in marine sediments. Water Res. Inst., Oregon State Univ.,
Corvallis, Oregon. 1-84.
Ohtake, H. and Hardoyo. 1992. New biological method for detoxification and removal of
hexavalent chromium. Water Sci. Technol. 25:295-402.
Ozretich, R.J. and D.W. Schults. 1998. A comparison of interstitial water isolation methods
demonstrates centrifugation with aspiration yields reduced losses of organic constituents.
Draft for SAB 1-121
-------
Chemosphere 36:603-615.
Pankow, J.F. 1979. The dissolution rates and mechanisms of tetragonal ferrous sulfide
(mackinawite) in anoxic aqueous systems. Ph.D thesis. California Inst. of Technology,
Pasadena, Calif. 1-146.
Patterson, R.R., S. Fendorf, et al. 1997. Reduction of hexavalent chromium by amorphous iron
sulfide. Environ. Sci. Technol. 31:2039-2044.
Persaud, D., R. Jacqumagi and A. Hayton. 1989. Development of provincial sediment quality
guidelines. Ontario Ministry of the Environment, Toronto, ON, Canada.
Pesch, C.E., D.J. Hansen, W.S. Boothman, W.J. Berry and J.D. Mahony. 1995. The role of
acid volatile sulfide and interstitial water metal concentrations in determining
bioavailability of cadmium and nickel from contaminated sediments to the marine
polychaete Neanthes arenaceodentata. EnvironT Toxicol. Chem. 14:129-14.
Peterson, G.S., G.T. Ankley and E.N. Leonard. 1996. Effect of bioturbation on metal-sulfide
oxidation in surficial freshwater sediments. Environ. Toxicol. Chem. 15:2147-2155.
Pettine, M. and F.J. Millero. 1990. Chromium speciation hi seawater: the probable role of
hydrogen peroxide. Limnol. Oceanogr. 730-736.
Pettine, M., I. Barra, et al. 1998. Effect of metals on the reduction of chromium (VI) with
hydrogen sulfide. Water Research. 32:2807-2813.
Phillips, H.O. and K.A. Kraus. 1965. Adsorption on inorganic materials VI. Reaction of
insoluble sulfides with metal ions hi aqueous media. J. Chromatography. 17:549-557.
Prothro, M.G. 1993. Office of Water Policy and Technical Guidance on Interpretation and
Implementation of Aquatic Life Metals Criteria. Letter from M.G. Prothro, Acting
Assistant Administrator for the Office of Water, to Water Management Division Directors.
U.S. EPA Regions I - X. October 1, 1993.
Draft for SAB 1-122
-------
Rai, D., B.M. Sass, et al. 1987. Chromium (III) hydrolysis constants and solubility of chromium
(ffl) hydroxide. Inorg. Chem. 26:345-349.
Richard, F.C. and A.C. Bourg. 1991. Aqueous geochemistry of chromium: a review. Wat.Res.
25(7):807-816.
Schmieman, E.A., J.N. Petersen, et al. 1997. Bacterial reduction of chromium. Appl. Biochem.
Biotech. 63-5:855-864.
Schoonen, M.A.A. and H.L. Barnes. 1988. An approximation of the second dissociation
constant for H2S. Geochim. Cosmochim. Acta. 52:649-654.
Schroeder, D.C. and G.F. Lee. 1975. Potential transformations of chromium in natural waters.
Water, Air and Soil Pollution. 4:355-365.
Sibley, P.K., G.T. Ankley, A.M. Cotter and E.N. Leonard. 1996. Predicting chronic toxicity of
sediments spiked with zinc: an evaluation of the acid volatile sulfide (AVS) model using a
life-cycle test with the midge Chironomus tentans. Environ. Toxicol. Chem. 15:2102-
2112.
Silvester, E., L. Charlet, et al. 1995. Mechanism of chromium (HI) oxidation by nabuserite. J.
Phys. Chem. 99:16662-16669.
Slotton, D.G. and I.E. Reuter. 1995. Heavy metals in intact and resuspended sediments of a
California reservoir, with emphasis on potential bioavailability of copper and zinc. Mar.
Freshwater Res. 46:257-265.
Smillie, R.H., K. Hunter, et al. 1981. Reduction of chromium (VI) by bacterially produced
hydrogen sulfide in a marine environment. Wat.Res. 15:1351-1354.
Spehar, R.L. and J.T. Fiandt. 1986. Acute and chronic effects of water quality criteria-based
metal mixtures on three aquatic species. Environ. Toxicol. Chem. 5:917-931.
Sposito, G, J.C.M. De Wit and R.H. Neal. 1988. Selenite adsorption on alluvial soils: III.
Draft for SAB 1-123
-------
Chemical modeling. Soil Sci. Soc. Am. J. 52:947-950.
Stephan, C.E., D.I. Mount, D.J. Hansen, J.H. Gentile, G.A. Chapman and W.A. Brungs. 1985.
Guidelines for deriving numerical national water quality criteria for the protection of
aquatic organisms and their uses. PB85-227049. National Technical Information Service,
Springfield, VA. 98pp.
Strobel, C.J., H.W. Buffum, S.J. Benyi, E.A. Petrocelli, D.R. Reifstick and D.J. Keith. 1995.
Statistical summary EMAP-Estuaries Virginian Province 1992-1993. U.S. EPA/620/R-94/-
26. U.S. Environmental Protection Agency, Narragansett, RI. 72 pp.
Stumm, W. and J.S. Morgan. 1996. Aquatic Surface Chemistry, Chemical Equilibria and Rates
in Natural Waters, Third Edition. John Wiley & Sons, Inc. New York.
Stumm, W. and J.S. Morgan. 1981. Aquatic Surface Chemistry - an Introduction Emphasizing
Chemical Equilibria in Natural Waters. John Wiley & Sons, New York, NY, USA.
Stumm, W. 1987. Aquatic Surface Chemistry, Chemical Processes at the Particle-Water
Interface. John Wiley & Sons, New York.
Sullivan, J., J. Ball, E. Brick, S Hausmann, G. Philarski and D. Sopcich. 1985. Report of the
technical subcommittee on determination of dredge material suitability for in-water
disposal. Wisconsin Department of Natural Resources, Madison, WI, USA.
Sunda, W.G. and R.R.L. Guillard. 1976. The relationship between cupric ion activity and the
toxicity of copper to phytoplankton. J. Mar. Res. 34:511-529.
Sunda, W. G. and J.M. Lewis. 1978. Effect of complexation by natural organic ligands on the
toxicity of copper to a unicellular alga, Monochrysis lutheri. Limnol. Oceanogr. 23:870-
876.
Sunda, W.G., D.W. Engel and R.M. Thuotte. 1978. Effect of chemical speciation of toxicity of
cadmium to grass shrimp, Palaemonetes pugio: importance to free cadmium ion. Environ.
Sci. Technol. 12:409-413.
Draft for SAB 1-124
-------
Swartz, R.C., G.R. Dittsworth, D.W. Schults and J.O. Lamberson. 1985. Sediment toxicity to a
marine infaunal amphipod: cadmium and its interaction with sewage sludge. Mar.
Environ. Res. 18:133-153.
Swartz, R.C., D.W. Schultz, T.H. DeWitt, G.R. Dittsworth and J.O. Lamberson. 1990. Toxicity
of fluoranthene in sediment to amphipods: a test of the equilibrium partitioning approach to
sediment quality criteria. Environ. Toxicol. Chem. 9:1071-1080.
Swartz, R.C., F.A. Cole, J. O. Lamberson, S.P. Ferraro, D.W. Schults, W.A. DeBen, Henry
Lee n and R.J. Ozretich. 1994. Sediment toxicity, contamination, and amphipod
abundance at a DDT- and dieldrin-contaminated site in San Francisco Bay. Environ.
Toxicol. Chem. 13:949-962.
Takacs, M.J. 1988. The oxidation of chromium by manganese-oxide: the nature and control of
the reaction. Michigan State.
Tessier, A., P.G.C. Campbell and M. Bisson. 1979. Sequential extraction procedure for the
speciation of paniculate trace metals. Anal. Chem. 51:844-851.
Tessier, A. and P.G.C. Campbell. 1987. Partitioning of trace metals hi sediments: relationships
with bioavailability. Hydrobiologia 149:43-52.
Tessier, A., Y. Couillard, P.G.C. Campbell and J.C. Auclair. 1993. Modeling Cd partitioning
in oxic lake sediments and Cd concentrations in the freshwater bivalve Andonata grandis.
Limnol. Oceanogr. 38: 1-17.
Thurman, E.M.1985. Organic Geochemistry of Natural Waters. Kluwer Academic Publishers
Group, Martinus Nijhoff/Dr. W. Junk Publishers, Boston, MA. pp. 497.
Troup, B.N. 1974. The interaction of iron with phosphate, carbonate, and sulfide hi Chesapeake
Bay interstitial waters: a thennodynamic interpretation. Ph.D thesis. Johns Hopkins
University, Baltimore, Maryland. 1-114.
U.S. Environmental Protection Agency. 1980. Ambient Water Quality Criteria for Silver -1980.
Draft for SAB 1-125
-------
Office of Water Regulations and Standards. EPA-440/5-80-071.
U.S. Environmental Protection Agency. 1984. Ambient Water Quality Criteria for Chromium -
1984. Office of Water Regulations and Standards. EPA-440/5-84-029.
U.S. Environmental Protection Agency. 1985a. Appendix B - Response to public comments on
"Guidelines for deriving numerical national water quality criteria for the protection of
aquatic organisms and their uses". July 19, 1985. Fed. Register. 50:30793-30796.
U.S. Environmental Protection Agency. 1985b. Ambient Water Quality Criteria for Cadmium -
1984. Office of Water Regulations and Standards. EPA 440/5-85-032.
U.S. Environmental Protection Agency. 1985c. Ambient Water Quality Criteria for Copper -
1984. Office of Water Regulations and Standards. EPA 440/5-85-031.
U.S. Environmental Protection Agency. 1985d. Ambient Water Quality Criteria for Lead -
1984. Office of Water Regulations and Standards. EPA 440/5-85-027.
U.S. Environmental Protection Agency. 1986. Ambient Water Quality Criteria for Nickel -
1986. Office of Water Regulations and Standards. EPA 440/5-85-004.
U.S. Environmental Protection Agency. 1987. Ambient Water Quality Criteria for Zinc -1987.
Office of Water Regulations and Standards. EPA 440/5-85-032.
U.S. Environmental Protection Agency. 1989. Sediment classification methods compendium.
Watershed Protection Division, U.S. EPA. 280pp.
U.S. Environmental Protection Agency. 1994a. Equilibrium partitioning approach to predicting
metal bioavailability in sediments and the derivation of sediment quality criteria for metals.
Vol 1. Briefing Report to the Science Advisory Board. EPA 822-D-94-002. Office of
Water, Office of Research and Development, Washington, DC, USA.
U.S. Environmental Protection Agency. 1994b. Methods for the determination of metals in
environmental samples. Supplement I. EPA-600-R-94-111. Office of Research and
Draft for SAB 1-126
-------
Development, Washington, DC.
U.S. Environmental Protection Agency. 1994c. Interim guidance on determination and use of
water-effect ratios for metals. EPA-823-B-94-001. Office of Water, Washington, DC,
USA.
U.S. Environmental Protection Agency. 1995a. An SAB report: review of the agencies approach
for developing sediment criteria for five metals. EPA-SAB-EPEC-95-020. Office of
Water, Washington, DC, USA.
U.S. Environmental Protection Agency. 1995b. Water quality criteria for dissolved metals.
Federal Register Notice, 60(86):22231-22237.
U.S. Environmental Protection Agency. 1995c. Ambient water quality criteria- saltwater copper
addendum. April 14, 1995. U.S. Environmental Protection Agency, Office of Water,
Office of Science and Technology, Washington, D.C.
U.S. Environmental Protection Agency. 1995d. Method 1669: Sampling ambient water for trace
metals at EPA Water Quality Criteria Levels. EPA-821-R-95-034. Office of Water,
Engineering and Analysis Division (4303), Washington, DC 20460.
U.S. Environmental Protection Agency. 1995e. Method 1638: Determination of trace metals in
ambient waters by inductively coupled plasma-mass spectrometry. EPA-821-R-95-031.
Office of Water, Engineering and Analysis Division (4303), Washington, DC 20460.
U.S. Environmental Protection Agency. 1995f. Method 1639: Determination of trace metals hi
ambient waters by stabilized temperature graphite furnace atomic absorption. EPA-821-R-
95-032. Office of Water, Engineering and Analysis Division (4303), Washington, DC
20460.
U.S. Environmental Protection Agency. 1995g. Method 1636: Determination of hexavalent
chromium by ion chromatography. EPA-821-R-95-029. Office of Water, Engineering and
Analysis Division (4303), Washington, DC 20460.
U.S. Environmental Protection Agency. 1996. EMAP-Estuaries Virginian Province Data 1990-
1993. Available from: EMAP Home Page WWW site, http://www.epa.gov/emap via the
INTERNET. Accessed 1997 Sept 17.
Draft for SAB 1-127
-------
U.S. Environmental Protection Agency. 1997a. The Incidence & Severity of Sediment
Contamination in Surface Waters of the United States. Vol 1: National Sediment Quality
Survey (EPA 823-R-97-006), Washington, D.C.
U.S. Environmental Protection Agency. 1997b. The Incidence & Severity of Sediment
Contamination in Surface Waters of the United States. Vol 2: Data Summaries for Areas
of Probable Concern (EPA 823-R-97-007), Washington, D.C.
U.S. Environmental Protection Agency. 1997c. The Incidence & Severity of Sediment
Contamination in Surface Waters of the United States. Vol 3: National Sediment
Contaminant Point Source Inventory (EPA 823-R-97-008), Washington, D.C.
U.S. Environmental Protection Agency. 1998a. Technical basis for establishing sediment quality
criteria for nonionic organic chemicals by using equilibrium partitioning. (In preparation)
«
U.S. Environmental Protection Agency. 1998b. Users guide for multi-program implementation of
sediment quality criteria. (In preparation)
s^
U.S. Environmental Protection Agency. 1998c. Guidelines for deriving site-specific sediment
quality criteria for the protection of benthic organisms. (In preparation)
Vaughan, D.J. and J.R. Craig. 1978. Mineral chemistry of metal sulfides. Cambridge, UK,
Cambridge University Press. ^-r ^. -v
Wang, W-X, S.B. Griscom and N.S. Fisher. 1997. Bioavailability of Cr(in) and Cr (VI) to
marine mussels from solute and paniculate pathways. Environ. Sci. Technol. 31:603-511.
Wang, Y.T. and H. Shen. 1997. Modelling Cr (VI) reduction by pure bacterial cultures. Wat.
Res. 31:727-732.
Wells, A.F. 1962. Structural inorganic chemistry. London, Oxford University Press.
Wittbrodt, P.R. and C.D. Palmer. 1995. Reduction of Cr (VI) in the presence of excess soil
fulvic acid. Environ. Sci. Technol. 29:255-263.
Wolfe, D.A., S.B. Bricker, E.R. Long, K.J. Scott and G.B. Thursby. 1994. Biological effects
of toxic contaminants in sediments from Long Island Sound and environs. National
Draft for SAB 1-128
-------
Oceanic and Atmospheric Administration Technical Memorandum NOS ORCA 80
NOAA/NOS Office of Ocean Resources Conservation and Assessment, Silver Spring, MD
USA.
Zamuda, C.D. and W.G. Sunda. 1982. Unavailability of dissolved copper to the American
oyster Crassostrea virginica: Importance of chemical speciation. Mar. Biol. 66:77-82.
Zouboulis, A.I., K.A. Kydros, et al. 1995. Removal of hexavalent chromium anions from
solutions by pyrite fines. Wat. Res. 29:1755-1760.
Draft for SAB 1-129
-------
APPENDIX A
Glossary of Abbreviations and Equations
Draft for SAB 1-130
-------
ACR Acute-Chronic Ratio
Ag Silver
AVS Acid Volatile Sulfide
ASTM American Society for Testing and Materials
Cd Cadmium
Cd Freely dissolved interstitial water concentration of contaminant
C Total interstitial water concentration of contaminant
Cs Concentration of contaminant in sediment
C^ Concentration of contaminant in sediment on an organic carbon basis
CCC Criteria Continuous Concentration
CFR Code of Federal Regulations
CLOGP Computer program for generating partition coefficients
CMC Criteria Maximum Concentration
CV Coefficient of Variation
CWA Clean Water Act
DOC Dissolved Organic Carbon
EDTA Ethlyene diamine tetra-acetic acid
EMAP Environmental Monitoring and Assessment Program
ESG Equilibrium Partitioning Sediment Guidelines
ESGoc Organic carbon-normalized Equilibrium Partitioning Sediment Guidelines
f^ Fraction of organic carbon in sediment
EqP ' Equilibrium partitioning
FAV Final Acute Value
FCV Final Chronic Value
{Fe2+} activity of Fe2+ (mol/L)
[Fe2+] concentration of Fe2+ (mol/L)
[FeS(s)] concentration of iron sulfide (mol/L)
[FeS(s)]j ' initial iron sulfide concentration in the sediment (mol/L)
FeS Iron monosulfide
FPV Final Plant Value
FRV Final Residue Value
GC/EC Gas Chromatography/Electron Capture
GC/MS Gas Chromatography/Mass Spectrometry
GFAA Graphite Furnace Atomic Absorption
Draft for SAB 1-131
-------
Kp
Kgp
LC50
IWGTU Interstitial Water Guidelines Toxic Unit
IWTU Interstitial Water Toxic Unit
KFeS solubility product for FeS(s) [(mol/L)2]
KMS solubility product for MS(s) [(mol/L)2]
Organic carbon: water partition coefficient
Octanol: water partition coefficient
Sediment: water partition coefficient
Solubility product constant
Concentration estimated to be lethal to SO percent of the test organisms within a
specified time period.
Liter
divalent metal activity (mol/L) —
concentration of M2+ (mol/L)
concentraton of added metal (mol/L)
concentration of solid-phase metal sulfide (mol/L)
Cubic meter
Microgram
Micrometer
//mole Micromole
mg Milligram
mg/1 Milligram per liter
ml Milliliter
mm Millimeter
NA Not Applicable, Not Available
ND Not Determined, Not Detected
ng Nanogram
Ni Nickel
NOAA National Oceanographic and Atmospheric Administration
NOEC No Observed Effect Concentration
NST National Status and Trends monitoring program
NTA Nitrilotriacetic acid
Pb Lead
pH Negative logarithm of the effective hydrogen ion concentration
OEC Observed Effect Concentration
[M2*]
[M]A
[MS(s)]
m3 or cum
//g
Draft for SAB 1-132
-------
POC
ppb
ppm
ppt
REMAP
{s2-}
[S2!
[SEM]
SAB
SD
SLC
SEM
SOP
STORET
TDS
TOC
TU
TVS
U.S. EPA
WQC
Zn
fcr
[SFe(aq)]
[SM(aq>]
[SS(aq>]
Particulate Organic Carbon
Parts per billion
Parts per million
Parts per trillion A
Regional Environmental Monitoring and Assessment Program
activity of S2" (mol/L)
concentration of S2" (mol/L)
simultaneously extracted metal concentration (/xmol/g)
simultaneously extracted Cd concentration (/xmol/g)
simultaneously extracted Cu concentration (/xmol/g)
simultaneously extracted Ni concentration (/xmol/g)
simultaneously extracted Pb concentration (/xmol/g)
simultaneously extracted Zn concentration (/xmol/g)
U.S. EPA Science Advisory Board
Standard Deviation
Screening Level Concentration
Simultaneously Extracted Metals
Standard Operating Procedure
EPA's computerized water quality data base
Total Dissolved Solids
Total Organic Carbon
Toxic Unit
Total Volatile Solids
United States Environmental Protection Agency
Water Quality Criteria
Zinc
(Fe2+}/[SFe(aq)]
{M2+}/[Sm(aq)]
activity coefficient of Fe2+
activity coefficient of M2+
activity coefficient of S2"
concentration of total dissolved Fe(n) (mol/L)
concentration of total dissolved M(H) (mol/L)
concentration of total dissolved S(H) (mol/L)
Draft for SAB 1-133
-------
APPENDIX B
Solubility Relationships for Metal Sulfides
Draft for SAB 1-134
-------
Consider the following situation: a quantity of FeS is titrated with a metal that
forms a more insoluble sulfide. We analyze the result using an equilibrium model of the M-
(II)-Fe(II)-S(-II) system. The mass action laws for the metal and iron sulfides are
YM"[M2+]Ys,-[S2-] = KMS (B-l)
v r\>f2+iv re 2-1 — IT /ti o\
T Fe^*l J • S L J ~' "^FeS \**~~/
where P^+], [Fe2+] and [S2"] are the molar concentrations; YM2»,1^2» and7sj- are the
activity coefficients; and K^ and KFeS are the sulfide solubility products. The mass balance
equations for total M(H), Fe(II) and S(-II) are
a-VfA/2*] + [A«(5)] = [M]A (B-3)
a"1 Fe^P762*] + [FeSCs)] = [FeS(s)]j
where
a'1 §J-[S2-] * [MS(s>] + [FeS(s>] = [FeSCs^j (B-5)
(B-6)
» = |Fea1/lEFe(aq)] (B-7)
Draft for SAB 1-135
-------
= [S2-]/[£S(aq)] (B-8)
are the ratios of the divalent species concentrations to the total dissolved M(n), Fe(n), and S(-
n) concentrations, [SM(aq)], [SFe(aq)L and [SS(aq)], respectively. [MS(s>] and [FeS(s)] are
the concentrations of solid-phase metal and iron sulfides at equilibrium. [FeS(s)]i is the initial
iron sulfide concentration in the sediment, and [M]A is the concentration of added metal.
The solution of these five equations can be obtained as follows. The mass balance
Equations B-3 and B-4 for M(II) and FE(n) can be solved for [MS(s>] and [FeS(s)] and
substituted in the mass balance Equation B-5 for S(H):
= [M]A (B-9)
The mass action Equations B-l and B-2 can be used to substitute for [Fe2"1"] and [M2+], which
results in a quadratic equation for [S2"]:
YS2-[S21
The positive root can be accurately approximated by:
(B-ll)
which results from ignoring the leading term in Equation B-10. This is legitimate because the
term in parentheses in Equation B-10 is small relative to [M]A due to the presence of the
sulfide solubility products. As a result, [S2"] is also small since it is in the denominator.
Draft for SAB 1-136
-------
Hence, the leading term in Equation B-10 must be small relative to [M]A and can safely be
ignored.
The metal activity can now be found from the solubility equilibrium Equation B-l:
so that
where
and
F\>r2*l - V L - V
* - K - K
KMS
Draft for SAB 1-137
-------
Equation A-13 can be expressed as
K,
FeS
(B-16)
The magnitude of the term in parentheses can be estimated as follows. The first term in the
denominator is always greater than or equal to 1, ^+>. 1, because it is the reciprocal of two
terms both of which are less than or equal to 1, Equation B-14. They are 0, since
all of'its terms are positive. Thus, the denominator of the expression in parentheses is always
greater than 1, Ppe2* + Vu^Msfcvts > 1- Therefore, the expression in parentheses is always
less than 1. Hence, the magnitude of the ratio of metal activity to total added metal is
bounded from above by ratio of the sulfide solubility products:
(Me2+}/[M]
(B-17)
This results applies if [FeS]; > [M]A so that excess [FeS(s)] is present.
If sufficient metal is added to exhaust the initial quantity of iron sulfide, then [FeS(s)]
= 0. Hence, the iron sulfide mass action equation (B-2) is invalid and the above equation no
longer applies. Instead, the only solid-phase sulfide is metal sulfide and
[MS] = [FeS];
(B-18)
so that, from the metal mass balance equation
Draft for SAB 1-138
-------
(B-19)
this completes the derivation of Equations 2-8 and 2-9.
Drc&forSAB 1-139
-------
APPENDIX C
Lake Michigan EMAP Sediment Monitoring Database
Draft for SAB 1-140
-------
Concentrations of SEM, AVS, TOC, and IWCTU for cadmium, copper, lead, nickel, and zinc in 46 surficial samples from Lake Michigan
Sample TOC
(»)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
0.18
4.63
3.36
4.89
0.92
4.37
5.27
0.08
4.27
2.11
1.89
0.41
2.87
3.68
0.28
0.07
3.51
0.40
1.73
0.69
2.51
1.17
0.13
1.03
0.63
0.30
0.29
0.21
0.11
0.05
0.27
4.95
0.54
6.75
0.18
0.15
0.56
0.10
0.06
2.68
0.16
1.80
1.29
0.05
0.14
0.57
SEM AVS
(umol/g) (umol/g)
0.53
3.46
2.78
3.55
0.14
2.82
1.20
0.17
1.47
0.25
1.12
0.74
1.17
1.56
1.32
0.17
0.75
0.97
1.74
0.70
0.19
0.59
0.21
0.62
0.13
0.15
0.25
0.12
0.20
0.04
0.85
1.17
0.44
1.37
0.26
0.06
0.17
0.22
0.06
5.83
0.16
0.56
1.02
0.06
0.16
0.66
0.03
0.35
0.06
0.05
0.03
1.13
0.13
0.03
4.49
" 0.03
0.03
0.07
0.18
0.03
0.44
0.05
0.08
0.03
0.15
0.03
0.05
0.03
0.03
0.03
0.20
0.03
0.03
0.03
0.06
0.03
0.03
1.66
0.12
0.09
0.03
0.05
0.05
0.12
0.03
0.03
0.07
0.03
2.25
0.03
0.05
0.03
SEM-
AVS
0.51
3.11
2.72
3.50
0.12
1.69
1.07
0.15
-3.02
0.23
1.10
0.67
0.99
1.54
0.88
0.12
0.67
0.95
1.59
0.68
0.14
0.57
0.19
0.60
-0.07
0.13
0.23
0.10
0.14
0.02
0.83
•0.49
0.32
1.28
0.24
0.01
0.12
0.10
0.04
5.81
0.09
0.54
-1.23
0.04
0.11 .
0.64
Cadmium
-
0.029
0.018
0.018
0.0002
0.024
0.029
0.115
0.050
-
-
0.0002
-
0.0002
0.0002
-
0.018
-
0.079
-
-
.
-
-
-
.
-
0.0002
0.0002
.
.
0.012
-
0.018
-
-
-
-
-
0.003
-
0.006
0.0002
-
-
-
Copper Lead
-
0.003
0.308
0.266
0.034
0.049
0.003
0.003
0.034
-
-
0.070
-
0.003
0.119
-
0.060
-
0.013
-
-
-
-
-
-
-
0.155
0.003
-
-
0.036
-
0.041
-
-
-
-
-
0.119
-
0.003
0.028
-
-
-
-
0.00004
0.002
0.0004
0.0008
0.0002
0.0001
0.001
0.0008
-
.
0.002
-
0.0004
0.0002
-
0.0008
.
0.0008
-
-
.
.
-
-
.
-
0.0001
0.0004
.
.
0.0004
.
0.0002
-
-
-
-
-
0.001
-
0.0006
0.002
-
-
IWCTU
Nickel
.
0.005
0.003
0.003
0.006
0.004
0.006
0.006
0.004
-
.
0.0005
.
0.006
0.004
-
0.008
.
0.010
.
-
.
.
.
.
.
-
0.011
0.007
.
\
0.002
.
0.017
-
-
-
-
.
0.0005
.
0.008
0.0005
-
-
-
% Survival
Zinc
.
0.003
0.029
0.006
0.032
0.020
0.020
0.055
0.026
-
.
0.001
.
. 0.015
0.050
-
0.058
.
0.020
-
.
.
.
-
.
.
-
0.0003
0.0003
.
.
0.020
.
0.012
-
-
-
-
.
0.020
.
0.015
0.044
-
-
-
Sum HyaleUa Chironomus
azteca Unions
.
0.040
0.360
0.293
0.073
0.097
0.058
0.180
0.115
-
.
0.074
.
0.025
0.173
-
0.145
.
0.123
-
-
.
.
-
-
.
-
0.167
0.011
.
.
0.070
.
0.088
-
-
-
-
-
0.144
.
0.033
0.075
-
-
-
92.5
90
92.5
100
0
97.5
92.5
95
95
77.5
97.5
-
97.5
96.5
90
100
100
95
97.5
97.5
75
97.5
57.5
72.5
95
.
35
75 '
80
97.5
97.5
97.5
100
95
95
95
-
60
97.5
90
62.5
75
100
82.5
-
70
40
90
90
97.5
90
100
100
87.5
100
87.5
100
.
97.5
92.5
87.5
100
100
100
97.5
97.5
92.5
100
65
57.5
90
.
35
72.5
82.5
100
97.5
95
100
90
100
92.5
-
55
100
95
65
95
55
72.5
-
67.5
Source: Columns for Sample, TOC, SEM, AVS, SEM-AVS and IWCTU taken directly from Leonard et al., 1996a. Column for lurvival
from personal communication with Leonard, 1998.
a AVS LOD=0.05 urn S/g
b Insufficient pore-water volume for metals analysis
t Cadmium LOD=0.01 ug/L (0.0002 IWCTU)
d Copper LOD =0.2 ug/L (0.0003 IWCTU)
e Lead LOD=0.1 ug/L (0.0001 IWCTU)
f Nickel LOD=0.5 ug/L (0.0005 IWCTU)
Draft for SAB 1-141
-------
APPENDIX D
Saltwater Sediment Monitoring Database
Draft for SAB 1-142
-------
APPENDIX D
Concentrations of SEM, AVS, Toxicity and TOC for EMAP, NOAA NS &T and REMAP Databases
STUDY1
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
SEM
umol/g
.289
1.500
.066
.134
.266
.266
1.292
.347
.750
.212
.497
.624
.032
.988
.604
.031
1.597
1.065
.189
.018
.079
.421
.798
.903
1.202
.159
.246
.687
.699
1.663
.083
.740
.878
.044
.910
.567
.734
2.171
3.423
.197
.162
2.803
.472
2.079
AVS
umol/
1.400
.742
.029
.028
3.740
1.080
1.230
.087
.948
.283
.490
13.400
.024
81.100
3.340
.331
72.400
8.480
6.460
.034
.976
3.210
68.000
3.150
67.700
3.310
4.870
2.420
.430
116.000
1.300
.976
1.220
.025
3.430
.621
25.000
5.610
138.000
.892
3.590
11.900
12.500
26.600
SEM-AVS
ugmol/g
-1.111
.758
.037
.106
-3.474
-.814
.062
.260
-.198
-.071
.007
-12.776
.008
-80.112
-2.736
-.300
-70.803
-7.415
-6.271
-.016
-.897
-2.789
-67.202
-2.247
-66.498
-3.151
-4.624
-1.733
.269
-114.337
-1.217
-.236
-.342
.019
-2.520
-.054
-24.266
-3.439
-134.577
-.695
-3.428
-9.097
-12.028
-24.521
SURVIVAL"
%
100.
98.
99.
103.
99.
102.
107.
102.
99.
108.
103.
113.
101.
101.
107.
98.
102.
93.
103.
99.
97.
111.
104.
99.
105.
104.
106.
93.
91.
100.
99.
101.
98.
106.
104.
104.
107.
102.
100.
107.
82.
101.
101.
94.
SIGNIFICANCE'
%
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
TO(
.60
2.68
.17
.14
.49
.56
1.80
.30
.95
.37
1.00
1.58
.11
3.36
1.38
.09
4.19
3.17
.32
.15
.14
.49
2.84
2.85
2.28
.51
.71
1.70
2.05
4.12
.14
2.30
2.84
.15
3.00
.76
2.21
2.57
4.14
.37
.81
2.36
2.77
3.18
Draft for SAB 1-143
-------
STUDV
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
SEM
umol/g
.445
2.228
.847
1.402
1.425
.263
2.936
.394
3.074
2.555
.452
.173
.578
.209
5.411
1.298
1.039
.960
7.369
1.380
4.259
8.229
3.535
2.543
2.124
.188
.229
1.820
3.468
1.622
.693
.294
.178
.223
.239
.801
.751
.299
.341
.205
2.415
.632
1.516
3.249
.462
.043
.050
AVS
umol/
.056
15.100
17.300
52.700
22.300
.079
29.600
.031
10.400
.402
.480
.201
.257
3.460
17.800
.228
.705
12.900
3.460
2.270
54.600
68.000
61.800
35.600
35.600
.836
.692
.227
14.600
6.080
1.200
.026
.074
.087
1.120
5.120
.090
.090
.174
.611
4.050
28.200
52.700
12.300
6.140
.024
.025
SEM-AVS
jsangife.
.389
-12.872
-16.453
-51.298
-20.875
.184
-26.664
.363
-7.326
2.153
-.028
-.028
.321
-3.251
-12.389
1.070
.334
-11.940
3.909
-.890
-50.341
-59.771
-58.265
-33.057
-33.476
-.648
-.463
1.593
-11.132
-4.458
-.507
.268
.104
.136
-.881
•4.319
.661
.209
.167
-.406
-1.635
-27.568
-51.184
-9.051
-5.678
.019
.025
SURVIVAL11
%
106.
103.
99.
109.
88.
84.
100.
87.
104.
96.
100.
98.
101.
96.
100.
100.
102.
94.
87.
97.
76.
43.
99.
33.
0.
108.
95.
104.
102.
102.
99.
95.
81.
104.
88.
92.
102.
104.
105.
95.
100.
88.
85.
103.
108.
100.
102.
SIGNIFICANCE" TOC
%
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
o
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.20
2.92
2.38
2.70
3.14
.27
4.15
.18
2.47
2.18
1.07
.22
.65
.36
2.78
.51
.30
1.91
1.86
.25
2.47
4.98
3.19
2.50
2.15
.35
.46
1.90
2.08
2.02
1.11
.38
.42
.43
.31
1.88
.66
.43
.99
.71
2.25
3.35
7.01
3.29
2.19
.18
.17
Draft for SAB 1-144
-------
STUDY*
EMAP-VA
EMAP-VA
EMAP-VA s
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
NOAA- LI
NOAA-LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA-LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA-LI
NOAA-LI
NOAA- LI
NOAA- LI
NOAA-LI
NOAA-LI
NOAA-LI
NOAA- LI
NOAA-LI
NOAA- LI
NOAA-LI
NOAA-LI
NOAA- LI
NOAA- LI
NOAA-LI
NOAA-LI
NOAA-LI
NOAA-LI
NOAA-LI
NOAA-LI
NOAA-LI
NOAA- LI
NOAA-LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA-LI
NOAA- LI
SEM
umol/g_
1.177
.624
..799
,.020
.088
2.220
.813
.851
.701
1.113
.601
1.505
.701
.717
2.163
.616
2.368
1.278
2.253
.865
.950
1.113
1.026
1.446
2.777
.211
2.665
2.813
1.235
2.198
3.624
3.594
1.342
2.462
.964
.332
2.311
.623
.896
.544
.641
.355
.222
2.262
1.307
1.963
2.785
AVS
umol/
3.460
6.210
29.700
.259
4.150
59.600
.381
.029
3.600
3.510
6.440
18.730
5.630
13.090
65.310
6.940
19.990
4.710
59.590
3.880
16.520
14.950
.850
12.480
29.720
.090
78.900
35.050
2.080
14.690
21.800
27.410
37.970
46.450
1.000
4.010
79.890
6.610
16.370
2.170
2.060
1.390
4.180
39.960
.380
51.820
61.020
SEM-AVS
ugmol/£_
-2.283
-5.586
-28.901
-.239
-4.062
-57.380
.432
.822
-2.899
-2.397
-5.839
-17.225
-4.930
-12.373
-63.147
-6.324
-17.622
-3.432
-57.337
-3.015
-15.570
-13.837
.176
-11.034
-26.943
.121
-76.235
-32.237
-.844
-12.492
-18.176
-23.816
-36.628
-43.988
-.036
-3.678
-77.579
-5.987
-15.475
-1.626
-1.419
-1.035
-3.958
-37.698
.927
-49.857
-58.235
SURVIVAL* S
% 5
100.
104.
100.
96.
100.
74.
93.
87.
100.
96.
96.
93.
93.
93.
92.
92.
91.
91.
91.
91.
90.
89.
88.
88.
87.
87.
87.
86.
84.
84.
83.
82.
82.
82.
81.
81.
81.
80.
80.
79.
79.
79.
77.
77.
76.
76.
76.
SIGNIFICANCE" '
fe
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1.
1.
1.
1.
1.
1.
1.
1.
TOC
1.83
2.25
4.10
.30
.25
2.18
.98
.57
.74
1.12
1.43
2.56
.77
2.05
3.22
.81
3.02
1.81
2.51
1.32
1.52
2.00
1.63
2.05
2.81
.54
3.33
3.83
1.58
2.80
2.48
2.59
1.85
3.18
1.60
1.29
3.69
.67
1.11
.27
1.56
.64
.45
2.67
1.56
3.46
3.81
Draft for SAB 1-145
-------
STUDY*
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- LI
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NNOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
SEM
umol/g.
4.333
1.927
.004
3.831
.808
1.783
2.622
.597
1.181
1.862
2.726
2.102
2.471
1.870
1.607
4.942
2.705
2.087
1.514
2.629
3.194
.872
1.080
.123
2.914
2.218
2.609
3.650
1.634
1.267
2.892
2.511
.661
2.458
1.872
.959
2.480
.784
.943
1.683
1.753
2.447
1.839
1.296
1.697
1.390
2.310
AVS
umol/
16.080
3.710
24.580
9.250
.960
40.630
61.840
1.090
3.730
50.390
62.760
33.630
7.220
17.120 ,
17.810
100.800
83.010
26.730
30.880
32.050
35.390
25.810
11.300
5.310
2.893
2.369
43.959
101.984
5.237
3.256
80.584
2.241
13.490
23.077
48.062
53.288
7.599
22.486
8.831
42.399
17.697
10.958
68.306
56.838
9.089
43.801
51.857
SEM-AVS
ugmol/g_
-11.747
-1.783
-24.576
-5.419
-.152
-38.847
-59.218
-.493
-2.549
-48.528
-60.034
-31.528
-4.749
-15.250
-16.203
-95.858
-80.305
-24.643
-29.366
-29.421
-32.196
-24.938
-10.220
-5.187
.021
-.151
-41.350
-98.334
-3.603
-1.989
-77.692
.270
-12.829
-20.619
•46.190
-52.329
-5.119
-21.702
-7.888
-40.716
-15.944
-8.511
-66.467
-55.542
-7.392
-42.411
-49.547
SURVIVAL*
%
75.
75.
74.
73.
71.
70.
70.
69.
68.
67.
67.
64.
63.
61.
59.
54.
53.
47.
42.
39.
37.
34.
16.
10.
8.
15.
26.
29.
36.
52.
83.
86.
87.
87.
89.
90.
90.
91.
91.
92.
94.
94.
95.
96.
97.
97.
97.
SIGNIFICANCE' TOC
%
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3.48
1.60
2.87
3.08
1.19
2.50
3.49
.76
.91
2.81
2.81
3.42
2.80
3.29
2.07
3.15
3.62
3.45
2.69
2.68
3.17
1.83
1.91
.22
3.05
2.89
3.74
1.83
1.72
1.53
6.98
2.12
1.00
3.15
3.25
2.39
4.45
1.88
1.78
3.41
1.41
4.45
2.54
3.05
2.68
3.27
3.35
Draft for SAB 1-146
-------
STUDY1
NOAA-BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA-HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA-HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA-HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA-HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA- HR
NOAA-HR
NOAA- BA
REMAP-BA
SEM
umol/g.
.399
2.481
1.736
.958
9.192
1.525
.678
5.037
4.202
1.174
1.855
3.092
2.997
2.581
2.869
5.442
2.618
5.061
2.376
6.998
4.480
4.662
5.896
3.103
1.662
3.512
.273
.335
1.664
2.674
5.532
4.029
4.614
3.379
4.240
4.303
5.209
4.801
4.697
2.600
1.013
1.527
.505
3.341
3.449
.270
.341
AVS
umol/
3.899
19.604
148.969
18.622
120.622
81.842 •
5.679
69.320
21.980
27.540
14.170
51.770
79.710
61.050
28.080
25.900
1.080
12.240
4.390
63.450
20.780
23.720
51.580
59.780
7.230
25.840
.050
.036
18.760
3.630
29.210
18.440
20.530
30.120
19.320
22.570
14.570
35.370
54.710
56.730
10.160
15.130
.630
43.920
37.860
.950
.156
SEM-AVS
ugmol/g_
-3.500
-17.123
-147.233
-17.664
-111.430
-80.317
-5.001
-64.283
-17.778
-26.366
-12.315
-48.678
-76.713
-58.469
-25.211
-20.458
1.538
-7.179
-2.014
-56.452
-16.300
-19.058
-45.684
-56.677
-5.568
-22.328
.223
.299
-17.096
-.956
-23.678
-14.411
-15.916
-26.741
-15.080
-18.267
-9.361
-30.569
-50.013
-54.130
-9.147
-13.603
-.125
-40.579
-34.411
-.680
.185
SURVIVAL" S
% i
99.
99.
99.
99.
100.
102.
103.
0.
41.
11.
18.
101.
112.
119.
81.
95.
109.
97.
108.
0.
20.
14.
2.
77.
19.
0.
91.
93.
69.
3.
96.
51.
91.
88.
101.
102.
101.
70.
38.
37.
29.
68.
105.
86.
76.
96.
84.
«GNIFICANCEC TO<
K
0
0
0
0
0
0
0
1.
1.
1.
1.
0
0
0
0
0
0
0
0
1.
1.
1.
1.
1.
1.
1.
0
0
1.
1.
0
1.
0
0
0
0
0
1.
1.
1.
1.
1.
0
0
1.
0
0
c
.80
3.31
2.94
1.77
4.61
2.96
1.45
5.02
3.47
1.88
4.44
3.86
3.09
2.86
2.50
2.20
2.67
2.98
2.49
1.98
2.98
3.19
4.78
3.99
2.61
4.44
.07
.07
.69
1.00
3.18
2.20
1.94
2.80
3.15
3.02
3.21
2.98
3.47
1.47
.77
.95
.25
2.55
3.63
.26
.06
Draft for SAB 1-147
-------
STUDY*
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA -
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
SEM AVS
umol/g, umol/
.888
.722
.362
2.138
3.008
.151
.115
.543
.103
.167
.073
.294
.120
.109
.185
.120
.347
.120
2.275
.344
.258
.119
.258
.494
.109
.266
.327
.230
2.026
14.550
3.332
3.763
.357
.524
.244
1.247
2.478
1.744
.131
.846
4.399
3.884
.673
3.150
.270
.162
2.880
12.971
4.948
.936
3.295
3.941
.555
.156
.156
.156
.932
.156
.156
.156
.156
.156
.156
.156
.156
16.592
.012
.343
.156
.156
.156
.156
.156
.393
6.400
47.793
389.857
243.322
201.687
10.923
3.974
4.502
48.130
47.376
.156
1.184
.927
116.954
237.650
21.769
43.975
4.491
.873
153.755
SEM-AVS SURVIVAL* SIGNIFICANCE" TOC
ugmol/g_ % %
-12.083
-4.226
-.574
-1.157
-.933
-.404
-.041
.387
-.053
-.765
-.083
.138
-.036
-.047
.029
-.036
.191
-.036
-14.317
.332
-.085
-.037
.102
.338
-.047
.110
-.066
-6.170
-45.767
-375.307
-239.990
-197.924
-10.566
-3.450
-4.258
-46.883
-44.898
1.588
-1.053
-.081
-112.555
-233.766
-21.096
-40.825
-4.221
-.711
-150.875
92.
85.
98.
95.
95.
96.
99.
94.
85.
97.
99.
91.
84.
92.
90.
88.
89.
81.
69.
91.
94.
84.
91.
86.
89.
86.
93.
83.
51.
0.
37.
79.
95.
98.
84.
91.
36.
69.
94.
73.
93.
89.
77.
91.
91.
98.
92.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1.
0
0
0
0
0
0
0
0
0
1.
1.
1.
1.
0
0
0
0
1.
1.
0
1.
0
0
1.
0
0
0
0
4.05
.40
.26
.43
.18
.15
.08
.07
.05
.16
.05
.34
.83
.92
4.48
.83
1.26
.62
1.81
3.85
.77
2.23
.88
2.10
4.07
1.06
.29
.19
.77
1.52
.83
.97
.26
.35
.27
.54
1.12
1.14
.21
1.58
6.55
8.45
4.11
5.47
.74
1.40
7.70
Draft for SAB 1-148
-------
STUDV
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
SEM
umol/g
.323
.413
.377
.099
1.100
.209
.213
.954
2.759
.711
1.915
2.186
2.480
.606
3.289
3.241
.616
1.506
2.485
1.894
3.149
.632
1.057
.638
1.087
3.711
2.990
8.894
1.277
3.925
5.632
6.809
7.645
4.012
3.905
.942
3.515
2.216
3.323
3.391
3.443
2.466
2.294
5.768
1.013
2.479
.554
AVS
umol/
1.684
3.056
3.056
.686
58.945
1.466
.780
1.542
6.498
10.240
12.596
17.605
23.523
~ 2.501
91.773
56.100
1.070
26.201
28.248
25.394
64.643
1.310
4.647
.218
.312
17.184
59.256
60.816
23.266
42.727
114.770
135.354
150.012
43.663
26.229
6.531
7.134
11.243
7.573
4.820
3.982
20.273
11.046
5.028
11.079
25.687
2.634
SEM-AVS
ugmol/g
-1.361
-2.643
-2.679
-.587
-57.845
-1.257
-.567 '
-.588
-3.739
-9.529
-10.681
-15.419
-21.043
-1.895
-88.484
-52.859
-.454
-24.695
-25.763
-23.500
-61.494
-.678
-3.590
.420
.775
-13.473
-56.266
-51.922
-21.989
-38.802
-109.138
-128.545
-142.367
-39.651
-22.324
-5.589
-3.619
-9.027
-4.250
-1.429
-.539
-17.807
-8.752
.740
-10.066
-23.208
-2.080
SURVIVAL" !
%
93.
94.
92.
93.
96.
93.
95.
83.
96.
97.
97.
95.
99.
98.
95.
97.
95.
96.
96.
93.
93.
87.
90.
92.
90.
88.
80.
85.
92.
90.
86.
91.
92.
86.
89.
84.
87.
86.
85.
83.
95.
82.
84.
75.
90.
83.
84.
SIGNIFICANCE'
%
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1.
0
0
0
TOC
.20
1.20
1.30
.75
3.86
.58
.69
.26
.45
.56
.21
.27
.32
.25
.77
1.14
.15
.95
.25
' .98
.90
1.51
- 2.44
3.52
7.36
3.99
5.24
3.63
3.18
3.85
4.29
4.36
6.04
3.73
3.93
.67
.75
1.22
1.25
1.05
.88
1.40
.95
1.77
.76
.99
.60
Draft for SAB 1-149
-------
STUDY"
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-UH
REMAP-UH
REMAP-UH
SEM
umol/g.
5.222
5.116
14.791
4.917
.398
4.855
3.290
5.822
9.167
6.214
.794
4.985
5.280
2.268
6.678
2.833
.333
.756
.582
1.012
1.596
.326
2.709
5.485
f
3.596
5.329
.337
.986
.856
5.364
1.706
.371
.193
.869
1.288
1.650
2.422
.512
4.198
5.081
6.095
8.471
3.370
1.198
2.127
1.360
1.197
AVS
umol/
22.617
7.352
109.780
.530
.218
9.606
10.105
51.460
93.563
42.415
2.651
43.663
1.934
6.300
17.559
45.222
22.315
1.216
.821
.567
.447
.156
3.120
14.666
19.503
4.321
2.901
.156
.156
39.700
23.515
4.210
.156
19.617
.593
.624
.156
.156
4.086
36.490
5.957
8.078
17.247
.156
12.446
1.790
3.373
SEM-AVS
ugmol/g_
-17.395
-2.236
-94.989
4.387
.180
-4.751
-6.815
-45.638
-84.396
-36.201
-1.857
-38.678
3.346
•4.032
-10.881
-42.389
-21.982
-.460
-.239
.445
1.149
.170
-.411
-9.181
-15.907
1.008
-2.564
.830
.700
-34.336
-21.809
-3.839
.037
-18.748
.695
1.026
2.266
.356
.112
-31.409
.138
.393
-13.877
1.042
-10.319
-.430
-2.176
SURVIVAL"
%
83.
9.
8.
89.
94.
83.
60.
41.
25.
68.
93.
53.
83.
16.
77.
54.
93.
92.
94.
94.
95.
93.
70.
92.
62.
91.
97.
96.
96.
91.
93.
91.
92.
85.
92.
91.
98.
93.
90.
89.
4.
91.
94.
94.
83.
99.
92.
SIGNIFICANCE' TOC
%
0
1.
1.
0
0
0
1.
1.
1.
1.
0
1.
0
1.
1.
1.
0
0
0
0
0
0
1.
0
1.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1.
0
0
0
0
0
0
1.48
1.45
9.15
3.10
2.42
2.62
5.70
2.22
6.48
3.24
2.36
3.90
6.10
1.99
15.20
2.02
1.23
.33
.30
.30
.17
.08
.42
2.29
.88
.97
.53
.12
.51
1.17
3.21
3.54
2.52
2.39
2.44
2.68
2.60
.42
2.63
2.08
3.03
5.30
3.91
1.03
3.43
1.26
5.85
Draft for SAB 1-150
-------
STUDY" SEM AVS
umol/g_ umol/
REMAP-UH 1.975 17.136
REMAP-UH 2.829 25.189
REMAP-UH 2.830 56.401
REMAP-UH 1.385 44.588
REMAP-UH 1.519 11.549
REMAP-UH 3.186 86.235
REMAP-UH 2.086 11.713
REMAP-UH 1.799 12.631
REMAP-UH .930 10.093
REMAP-UH .459 .156
REMAP-UH .889 2.623
REMAP-UH .833 2.464
REMAP-UH 1.317 15.563
REMAP-UH 2.480 32.123
REMAP-UH .626 9.949
REMAP-UH 1.500 5.427
REMAP-UH .723 1.341
REMAP-UH 4.158 13.504
REMAP-UH 2.241 27.788
REMAP-UH 2.907 29.285
REMAP-UH .852 1.591
REMAP-UH 2.294 53.955
REMAP-UH 2.995 33.995
REMAP-UH 2.981 44.910
REMAP-UH .677 10.323
a) Sources: EMAP-VA is U.S. EPA, 1996
NOAA-LI is Wolfe et al., 1994
NOAA-BO is Long et al., 1996
NOAA-HR is Long et al., 1995
REMAP is Adams et al., 1996
SEM-AVS SURVIVAL11 SIGNIFICANCE" TOC
U£inol/g_
-15.161
-22.360
-53.571
-43.203
-10.030
-83.049
-9.627
-10.832
-9.163
.303
-1.734
-1.631
-14.246
-29.643
-9.323
-3.927
-.618
-9.346
-25.547
-26.378
-.739
-51.661
-31.000
-41.929
-9.646
.
% - «
45.
84.
96.
88.
82.
93.
82.
37.
89.
98.
95.
86.
88.
87.
97.
89.
89.
96.
70.
95.
93.
15.
88.
94.
91.
%
1.
0
0
0
0
0
0
1.
0
0
0
0
0
0
0
0
0
0
1.
0
0
1.
0
0
0
2.33
.91
1.21
1.03
1.06
1.39
.79
1.06
.43
.13
.21
4.96
2.56
3.06
2.58
2.71
3.89
4.78
2.66
5.15
2.03
4.37
3.55
2.97
3.32
b) Conclusion of signifigance varies for three databases.
EMAP significance based on percent survival of control
NOAA significance based on percent survival kss than 80%
REMAP significance based on percent survival less than 80%
c) Significance: 0 - No significant toxichy
1 - Significant toxichy
Draft for SAB 1-151
-------
Sediment Assessment Presentation Materials
-------
The Technical Basis of the Use of Acid Volatile Sulfide (AYS) and Interstitial
Water Normalizations in Sediment Guidelines
by
Walter J. Berry
U.S.EPA, Atlantic Ecology Division, Narragansett RI
Dominic M. Di Toro
HydroQual, Inc., Mahwah NJ
David J. Hansen
Great Lakes Environmental Center, Traverse City MI
March 1999
INTRODUCTION
The U.S. EPA Science Advisory Board (SAB) has endorsed the use of the equilibrium
partitioning (EqP) approach as the foundation for the Agency's sediment guidelines for both
nonionic organics and divalent metals (U.S.EPA, 1990,1992,1995). The EqP approach was chosen
because it was the only available approach that could provide numerical guidelines, applicable across
sediments, which had a firm theoretical basis (Figure 1). The usefulness of the sediment guideline
for metals (then referred to as the Sediment Quality Criterion) had several limitations, however. One
limitation was that it only applied to five metals: cadmium, copper, lead, nickel, and zinc. Another
limitation was that it was a "one-way criterion", which is to say that it could predict that some
sediments would not be toxic, but could not predict that other sediments would be toxic.
This SAB review is intended to examine some modifications to the metals guidelines
previously reviewed by the SAB which extend the usefulness of the EqP approach to assessing the
potential effects of metals in sediments. These modifications extend the use of acid volatile sulfide
(AVS) and interstitial water normalization to include chromium and silver, and may eliminate the
"one-way" limitation by incorporating organic carbon into the assessment of sediments contaminated
with metals.
The technical basis of the use of acid volatile sulfide (AVS) and interstitial water
normalizations in sediment guidelines has been extensively reviewed elsewhere. Ankley et al.
(1994) described the AVS and interstitial water approach and Ankley et al. (1996) reviewed all of
the available data. Ankley et al. (1996) was the first of a large collection of papers relating to the
use of AVS and interstitial water normalization in sediment assessment which appeared in the same
2-2
-------
issue of Environmental Toxicology and Chemistry (December, 1996, Vol 15, No. 12). In this
introduction we will first review the published literature on EqP as it relates to metals. Then we will
summarize the available data from acute laboratory sediment toxicity tests with spiked sediments
and with field sediments which can be used to test the utility of AVS and IWTU normalizations in
predicting sediment toxicity. Then the available chronic data will be summarized. The data
pertaining to the bioaccumulation of metals in relation to AVS normalization will be reviewed.
Some considerations of the use of AVS normalization of field sediments will be discussed. Finally,
the guidelines statements will be presented.
LITERATURE REVIEW
Di Toro et al. (1990) showed that the toxicity of cadmium-spiked marine sediments was
linked to metals/AVS ratios and interstitial water (IW) metals concentrations. Since then several
studies using fresh and salt water sediments spiked with cadmium, copper, lead, nickel and zinc
(Berry et al., 1996; Casas and Crecelius, 1994; Green et al., 1993; Di Toro et al., 1992; Carlson et
al., 1991) have demonstrated the utility of these parameters in causally linking toxicity to metals in
sediments. Kemp and Swartz (1998) maintained constant IW concentrations in cadmium-spiked
sediments modified by varying quantities of organic carbon and found that mortality was correlated
with IW concentration but not total sediment concentration. The utility of AVS and IW
normalizations has also been demonstrated in studies conducted at field sites contaminated with
copper (Ankley et al., 1993) and a mixture of cadmium and nickel (Pesch et al., 1995; Di Toro et al.,
1992; Ankley et al., 1991). Two colonization experiments with cadmium-spiked sediments, one
conducted in a freshwater lake (Hare et al., 1994) and a seawater colonization test conducted in the
laboratory (Hansen et al., 1996a) also support the use of this approach for predicting the toxicity of
these metals in sediments. The success of this approach for predicting the bioavailability of these
metals in sediments is in direct contrast to the lack of success in using dry weight metals
concentrations for this purpose (Di Toro et al., 1990; Di Toro et al., 1992; Luoma, 1989).
TECHNICAL BASIS
The theoretical foundation for EqP-based AVS predictions of metal toxicity is that the
sulfides of cadmium, copper, nickel, lead, and zinc all have lower sulfide solubility product constants
than do the sulfides of iron and manganese, which are formed naturally in sediments as a product
of the bacterial oxidation of organic matter (Goldhaber and Kaplan, 1974). As a result, these metals
will displace manganese and iron whenever they are present together with manganese and iron
monosulfides (Di Toro et al., 1992). Because the solubility product constants of these sulfides are
small, sediments with an excess of AVS will have very low metal activity in the IW and no toxicity
due to these metals should be observed in the sediments.
The results of the studies cited above were consistent with the following predictions based
on EqP theory: 1) when sediments have an excess of AVS over metals, sediments will not be toxic,
and little or no metal will be present in the IW; and 2) when sediments have an excess of metals
over AVS, AVS binding potential will be exceeded and metals will be present in the IW or available
to bind with other sediment phases (i.e. total organic carbon) (Di Toro et al., 1990). Nontoxic
sediments with a small excess of metals over AVS may have low IW concentrations, less than those
2-3
-------
known to be toxic in water-only tests, suggesting the importance of additional metal binding phases
in sediments (Green et al., 1993; Ankley et al., 1993; Gonzales et al., 1992). The appropriate
fraction of metals to use for AVS normalization is referred to as simultaneously extracted metal
(SEM). This is the metal which is extracted in the cold acid used in the AVS procedure. This
fraction is appropriate because some metals form sulfides which are not fully labile in the short time
required for the AVS extraction (e.g. nickel and zinc) (Di Toro et al., 1992). If a more rigorous
extraction were used to increase the fraction of metal extracted which did not also capture the
additional sulfide extracted, then the sulfide associated with the additional metal release would not
be quantified. This would result in an erroneously high metal to AVS relationship (Di Toro et al.,
1992).
An analysis of a simple chemical equilibrium model for the system M(H), Fe (II), S(-D)
where M(E) are the divalent metals that form sulfides shows that
{M}/[M]T > KWK^
Where {M} is the activity of M in the IW, [M]T is the total metal concentration, KMS is the solubility
product for the metal sulfide, MS, and KFeS is the solubility product of FeS. The ratio KMS/K^ is
10'5 6, 10'6 °, 10-10 5,10'1' °, 10-18 6 for M = Ni, Zn, Cd, Pb, and Cu, respectively (Berry et al., 1996).
If the metals are present in excess of the sulfides (SEM-AVS >0.0), and there are no other sediment
phases capable of binding the metals (e.g., dissolved organic carbon (DOC) or total organic carbon
(TOC)), then metal will be present in the IW and the sediment may be toxic.
Toxicity predictions based on sulfide binding for sediments contaminated with mixtures of
metals which form insoluble sulfides would use the sum of the molar concentration of SEM for the
divalent metals present (i.e., Cd, Cu, Ni, Pb, Zn) for comparison with the molar concentration of
AVS in the sediment. If the sum of the SEM is greater than that of AVS, metals may occur in the
IW in sufficient concentrations to be toxic. If the toxicity of the cationic metals in IW is assumed
to be additive (Spehar and Fiandt, 1986), it should be possible to predict the toxicity of the sediments
in the same way as in the individual metal experiments, using the sum of the interstitial water toxic
units (IWTU). The divalent metals should appear in the IW in reverse order of the solubilities of
their sulfides (Di Toro et al., 1990). Thus, nickel should appear first in the IW in sediments with
SEM-AVS slightly greater than 0.0, followed by zinc, cadmium, lead, and copper as the
concentration of metals increases relative to that of AVS. This predicted outcome was observed
by Berry etal. (1996).
ACUTE TOXICITY DATA
Di Toro et al. (1990) found that toxicity of cadmium in sediments increased with increasing
concentration when the cadmium concentration was expressed on a dry weight basis, but that the
response was sediment specific (Figure 2). This meant that it was not possible to predict the toxicity
of cadmium across sediments using dry weight normalization. This sediment specificity
undoubtedly contributes to the scatter seen when mortality is plotted against dry weight
concentration in a data set which includes a number of spiked metals tests and field sites where
metals were thought to be the cause of observed toxicity (Figure 3 from Hansen et al., 1996b as is
presented in the Metals Mixtures ESG document). Note that in Figure 3 the sediment concentrations
2-4
-------
that caused little or no mortality overlap with those concentrations which caused 100% mortality for
three orders of magnitude, and there is no theoretical basis for predicting where toxicity might occur.
When mortality was plotted against the concentration of cadmium in the interstitial water,
however, Di Toro al. (1990) found that the results were similar across sediments (Figure 4). The
larger data set of lab and field data seen in Figure 5 further validate the use of IWTUs. In this data
set little or no toxicity was seen in sediments with less than 0.5 IWTU, while mortality increased
with increasing IWTU in sediments with greater than 0.5 IWTU.
Similarly, Di Toro et al. (1990) found that mortality was not sediment specific when
cadmium was expressed on an SEM/AVS basis (Figure 6). (SEM/AVS ratios were used in earlier
papers but have been supplanted by SEM-AVS differences, especially in field sediments; see Hansen
et al., 1996b and the Metals Mixtures ESG document). In the larger data base when there was an
excess of sulfide (SEM-AVS < 0.0) the sediments were not toxic, but when the metals were in
excess, many of the sediments were toxic (Figure 7). Note that many of the sediments with large
excesses of metals were not toxic. This is the reason that the SEM-AVS guideline is considered a
"one-way" guideline, which is more useful in the prediction of lack of toxicity than it is in the
prediction of toxicity. This is still very useful, because the vast majority of field sediments have an
excess of AVS. Furthermore, it should be pointed out that the ability of AVS normalization to
predict toxicity, although not as good as its ability to predict lack of toxicity, is comparable to the
ability of the empirically-derived methods to predict toxicity.
CHRONIC DATA
The data from the available life-cycle and colonization exposures also support the use of
AVS and interstitial water normalizations in the prediction of the lack of biological effects of metals
in sediments. Data are available for a number of metals from both freshwater and saltwater sediment
tests. These data are summarized in Table 1. Note that all sediments in which effects were observed
had an excess of metal over AVS.
BIOACCUMULATION
In a perfect world organisms would not take up metals from sediments with SEM - AVS <
0.0. Unfortunately, data exist which show that bioaccumulation of metals from sediments is often
more highly correlated with dry weight concentration than it is with an AVS-normalized
concentration. One such data set is shown in Figure 7. In some of the early experiments it was
thought that the apparent accumulation of metal in sediments where SEM-AVS < 0.0 may have been
an artifact of the use of very high and increasing concentrations of metals in spiked sediments.
However, a recent experiment, in which the metal concentration was held constant and the AVS was
varied, seems to show conclusively that this is not the case. In this experiment the accumulation of
metal was more closely related to dry weight concentration than AVS-normalized concentration
regardless of whether the metal concentration or AVS was varied (Lee et al., 1998). It would appear
then, that a guideline using AVS normalization would not be protective ofbioaccumulation of metals
from sediments. It must be noted, however, that although metal uptake has been shown from
sediments with an excess of AVS, no biological effects of these sediments have been shown. If there
2-5
-------
was an effect of metals uptake from these sediments on the lower levels of the food chain, then it
might have been expected to show up in the colonization and life-cycle tests described above.
Furthermore, there is little evidence of bioconcentration of these metals up the food chain.
Therefore, the fact that metals can be taken up from sediments with an excess of AVS means that
the guideline must be qualified, but this finding does not negate the usefulness of the guideline.
USE OF AVS WITH FIELD SEDIMENTS
There are some considerations relative to the use of AVS in predicting biological effects in
field sediments which should be noted here. 1) AVS can break down in aerobic storage, but it is
relatively stable, especially if frozen in glass. 2) AVS varies with season, being lowest in the winter
and early spring, and highest in the late summer. Since most sampling is done in the warmer months
there may be a concern about sampling when AVS is at a peak. If sediments which have a slight
excess of AVS over metals are found or expected sampling can be done in the winter or early spring.
3) AVS concentration varies with depth, being higher at depth and lowest at the surface. Thus,
attention must be paid to the depth of sampling. Sampling should be at the depth of concern (depth
of dredging, for example). The top 2.0 cm is recommended for evaluation of "in-place" sediments.
4) AVS and interstitial water normalization can provide insight into the causes of sediment toxicity
hi field sediments with mixtures of contaminants, but cannot by itself predict the toxicity of those
sediments because other toxicants may make these sediments toxic even if the sediments are not
toxic due to metals (Figure 9).
STATEMENT OF GUIDELINE
Based on the considerable evidence summarized above the following ESG for metals were
. proposed in the 1995 draft document:
Solid phase guideline:
S[SEM], - AVS <0.0
(Where S[SEM], = molar sum of cadmium, copper, lead, nickel, and zinc)
Interstitial water guideline:
E(PW], /[?€¥],< 1.0
(Where S([TW], = interstitial water concentration of cadmium, copper, lead, nickel, and zinc; and
[FCV], = Final Chronic Value, from water quality criteria)
If either of these guidelines are met, sediments should not be toxic due to metals (cadmium, copper,
lead, nickel and zinc).
2-6
-------
References
Ankley. G.T. 1996. Evaluation of metal/acid-volatile sulfide relationships in the prediction of metal
bioaccumulation by benthic macroinvertebrates. Environ. Toxicol. Chem. 15:2138-2146.
Ankley, G.T., V.R. Mattson, E.N. Leonard, C.W. West and J.L. Bennett. 1993. Predicting the acute
toxicity of copper in freshwater sediments: evaluation of the role of acid-volatile sulfide. Environ.
Toxicol. Chem. 12:315-320.
Ankley, G.T., G.L. Phipps, E.N. Leonard, D.A. Benoit, V.R. Mattson, P.A. Kosian, A.M. Cotter,
J.R. Dierkes, D.J. Hansen and J.D. Mahony. 1991. Acid-volatile sulfide as a factor mediating
cadmium and nickel bioavailability in contaminated sediments. Environ. Toxicol. Chem. 10:1299-
1307.
Ankley, G.T., N.A. Thomas, D.M. Di Tore, D.J. Hansen, J.D. Mahony, W. J. Berry, R.C. Swartz and
R. A. Hoke. 1994. Assessing potential bioavailability of metals in sediments: A proposed approach.
Environ. Mgt. 18:331-337.
Berry W.J., D.J. Hansen, J.D. Mahony, D.L. Robson, B.P. Shipley, B. Rogers and J.M. Corbin.
1996. Predicting the toxicity of metals-contaminated sediments using acid volatile sulfide and
interstitial water normalization in laboratory-spiked sediments. Environ. Toxicol. Chem. 15:2067-
2079.
Carlson, A.R., G.L. Phipps, V.R. Mattson, P.A. Kosian and A.M. Cotter. 1991. The role of acid-
volatile sulfide in determining cadmium bioavailability and toxicity in freshwater sediments.
Environ. Toxicol. Chem. 10:1309-1319.
Casas, A.M. and E.A. Crecelius. 1994. Relationship between acid volatile sulfide and the toxicity
of zinc, lead and copper in marine sediments. Environ. Toxicol. Chem. 13:529-536.
Di Toro., D.M., J.D. Mahony, D.J. Hansen, K.J. Scott, A.R. Carlson and G.T. Ankley. 1992. Acid
volatile sulfide predicts the acute toxicity of cadmium and nickel in sediments. Environ. Sci.
Technol. 26:96-101.
Di Toro, D.M., J.D. Mahony, D.J. Hansen, K.J. Scott, M.B. Hicks, S.M. Mays and M.S. Redmond.
1990. Toxicity of cadmium in sediments: The role of acid volatile sulfide. Environ. Toxicol. Chem.
9:1489-1504.
Goldhaber, M. B. and I.R. Kaplan, 1974. The sulfur cycle. In E.D. Goldberg, ed., The Sea, Vol. 5-
Marine Chemistry. John Wiley and Sons, New York, NY, USA, pp. 569-655.
Gonzales, A.M., J.D. Mahony and D.M. Di Toro. 1992 The role of organic carbon in the toxicity
of anoxic sediments contaminated with copper and other metals: An experimental study. Abstracts.
13th Annual Meeting, Soc. Environ. Toxicol. Chem., Cincinnati, OH, USA, Nov. 8-12, 1992, p.
162.
2-7
-------
Green, A.S., G.T. Chandler, and E.R. Blood. 1993. Aqueous-, pore-water-, and sediment-phase
cadmium: toxicity relationships for a meiobenthic copepod. £HVZ>O/Z. Toxicol Chem. 12:1497-1506.
v
Hansen, D.J., J.D. Mahony, W. J. Berry, S. Benyi, J. Corbin, S. Pratt and M.B. Able. 1996a.
Chronic effect of cadmium in sediments on colonization by benthic marine organisms: An evaluation
of the role of interstitial cadmium and acid volatile sulfide in biological availability. Environ.
Toxicol. Chem 15:2136-2137.
Hansen, D.J., W. J. Berry, J.D. Mahony, W.S. Boothman, D.M. Di Toro, D.L. Robson, G.T. Ankley,
D. Ma, Q. Yan and C.E. Pesch. 1996b. Predicting the toxicity of metals-contaminated field
sediments using interstitial concentration of metals and acid volatile sulfide normalizations. Environ.
Toxicol. Chem 15:2080-2094.
Hansen, D.J., J.D. Mahony, W. J. Berry, S. Benyi, J. Corbin, S. Pratt and M.B. Able. 1996a.
Chronic effect of cadmium in sediments on colonization by benthic marine organisms: An evaluation
of the role of interstitial cadmium and acid volatile sulfide in biological availability. Environ.
Toxicol. Chem 15:2136-2137.
Hare, L., R. Carignan and M.A. Huerta-Diaz. 1994. A field experimental study of metal toxicity and
accumulation by benthic invertebrates; implications for the acid volatile sulfide (AVS) model.
Limnol. Oceanogr. 39:1653-1668.
Kemp, P.P., and R.C. Swartz. 1988. Acute toxicity of interstitial and particle-bound cadmium to
a marine infaunal amphipod. Mar. Environ. Res. 26:135-153.
Lee, B.-G., H.-S. Jeon, S.N. Luoma, J.-S. Yi, C.-H. Koh. 1998. Effects of AVS (Acid Volatile
Sulfide) on the bioaccumulation of Cd, Ni, and Zn in bivalves and polychaetes. Abstract: 19th
Annual Meeting of the Society of Environmental Toxicology and Chemistry. Charlotte, N.C.
Luoma, S.N. 1989. Can we determine the biological availability of sediment-bound trace elements?
Hydrobiologia. 176/177:379-396.
Pesch, C.E., D.J. Hansen, W.S. Boothman, WJ. Berry and J.D. Mahony. 1995. The role of acid-
volatile sulfide and interstitial water metal concentrations in determining bioavailability of cadmium
and nickel from contaminated sediments to the marine polychaete, Neanthes arenaceodentata.
Environ. Toxicol. Chem. 14:129-141.
Spehar, R.L. and J.T. Fiandt. 1986. Acute and chronic effects of water quality criteria-based metal
mixtures on three aquatic species. Environ. Toxicol. Chem. 5:917-931.
U.S. Environmental Protection Agency. 1990. An SAB evaluation of the equilibrium partitioning
(EqP) approach for assessing sediment quality. EPA-SAB-EPEC-90-006. Office of Water,
Washington, DC.
U.S. Environmental Protection Agency. 1992. An SAB report. Review of sediment criteria
2-8
-------
development methodology for non-ionic organic contaminants. EPA-SAB-EPEC-93-002. Office
of Water, Washington, DC.
U.S. Environmental Protection Agency. 1995. An SAB report: Review of the Agencies approach
to developing sediment quality criteria for five metals. EPA-SAB-EPEC-95-020. Office of Water,
Washington, DC.
2-9
-------
Figure Legends
Figure 1: Conceptual models of chemical exposure.
Figure 2: Amphipod mortality vs. dry weight cadmium concentration in three cadmium-spiked
sediments. From Di Toro et al. (1990).
Figure 3: Mortality vs. Total metal or SEM (dry weight) in field and lab sediments in which
metals were the probable cause of toxicity. From Metals Mixtures ESG document.
Figure 4: Amphipod mortality vs. interstitial water cadmium concentration in three cadmium-
spiked sediments. From Di Toro et al. (1990).
Figure 5: Mortality vs. interstitial water toxic units in field and lab sediments in which metals
were the probable cause of toxicity. From Metals Mixtures ESG document.
Figure 6: Amphipod mortality vs. SEM/AVS in three cadmium-spiked sediments. From Di Toro
etal. (1990).
Figure 7: Mortality vs. SEM - AVS in field and lab sediments in which metals were the probable
cause of toxicity. From Metals Mixtures ESG document.
Figure 8: Cadmium tissue concentration vs. SEM/AVS in polychaetes exposed to cadmium-
spiked sediments. From Pesch et al. (1995).
Figure 9: Mortality vs. SEM-AVS in field sediments contaminated with a mixture of toxicants.
2-10
-------
Table 1. Summary of the results of full life-cycle and colonization toxicity tests conducted in the laboratory and field using
sediments spiked with individual metals and metal mixtures. Note that for all effects concentrations SEM-AVS > 0.0 (Bold data).
Toxicity Test Metal
Life-cycle:
Leptocheirus Cadmium
plumulosus
Chironomus Zinc
tentans
Colonization:
Laboratory- Cadmium
saltwater
Field-saltwater Cadmium, copper, nickel,
zinc
Field- Cadmium
freshwater
Field- Zinc
freshwater
Duration SEM-AVS
(dayS) No Effect
28 -3.4, -2.0,
0.78, 1.9
56 -4.3, -2.6,
-1.4, 6.4
118 -13.4
120 <0
-0.45, -0.25,
0.5b
368 -3.6, -3.5, -
2.9, -2.0
a Cone.
Effects
8.9, 15.6
21.9,
32.2
8.0, 27.4
—
4.5b
1.0
Effect
Mortality 100%
Larval mortality 85-100%.
Weight and emergence reduced.
Fewer polychaetes, shifts in
community composition, fewer
species, bivalves absent,
tunicates increased.
No effects observed.
Reduced Chironomus numbers.
Bioaccumulation.
Occasional minor reductions in
Naididae oligochaetes.
Reference
DeWitt et
al., 1996
Sibley et
al., 1996
Hansen et
al., 1996
Boothman
et al., 1997
Hare et al.,
1994
Liber et
al., 1996
a SEM- AVS differences are used instead of SEM/AVS ratios to standardize across the studies referenced. An SEM-AVS difference
of <0 is the same as an SEM/AVS ratio of < 1.0. An SEM-AVS difference of >0 is the same as an SEM/AVS ratio of >1.0.
b Nominal concentrations.
(N
-------
Conceptual Models of Chemical Exposure
Sediment
Pore Water
(N
Water Column
Sediment
"Equilibrium Partitioning"
-------
-------
i i i INI i i i i i mi i i i i i mi
100
—. 80
O Spiked Sediment
_ O Field Sediment o
60
(0
r
o
-------
DC
O
MORTALITY vs INTERSTITIAL WATER CADMIUM
100
80
60
40
20
— WATER ONLY
EXPOSURE
A AMPEUSCA
•& RHEPOXYNIUS
0.00001
0.00100
0.10000 -10.00000 1000.00000
CADMIUM ACTIVITY (mg Cd2+/L)
Figure 2. Mortality versus interstitial water cadmium activity. Medians and interquartile ranges
for each decade of interstitial water activity. Water only exposure data torAmpelisca and Rhepoxy-
nius hudsord. The line is a joint fit to both water only data sets.
-------
I I III III I I I I 11 III I I I I 11 III
100
— 80
CO
t
o
O Spiked Sediment
O Field Sediment
60
O
o ooOoO OOOoQ o £(00 oo OOODO^JO
o
O O o
O
O
o O O
O
O o
«0
<)
<)
10
100
1000
Interstitial Water Toxic Units
-------
I-
i
OC
o
MORTALITY vs SEDIMENT CADMIUM
ACID VOLATILE SULFIDE NORMALIZATION
• U SOUND
• MIXTURE
<> NINIQRETPOND
0.01
0.10
1.00
10.00
100.00
SEDIMENT CADMIUM (umol Cd / umol AVS)
Figure 6. Mortality versus AVS normalized sediment cadmium for Lond Island Sound, Ninigret
Pond, and a 50/50 volume mixture. The sediment cadmium and AVS are the averages of the ini-
tial and final concentrations in the control vessels.
-------
r
o
100
O Spiked Sediment
O Reid Sediment
oo
r-H
cs
-25
25
50
75
100
125
SEM-AVS
-------
CADMIUM - SPIKED SEDIMENTS
joo
10
0
3
F 1
0)
T3
o
1 °-1
it
0.01
0.001
0.
«<
= o £ n o
^™ J^P^^— ^^^k ^^^^
^™ I3B ^^^^^n ^^"1
L m LJ Q
•••^ ~^^~" j
•^
•^
= a
HB
•I
. ^
«• ^l^^p &
•V ^^^^ ^^k
" ™ ^^^^p
= 4^ • us
^ ^ D LIS - no burrowing
E
-------
SEM - AVS normalization must be used with caution in
field Sediments with multiple contaminants:
Freshwater and Saltwater Sites
I^U
100
& 8°
1
3 60
§
^ 40
20
(5(
T
r
<&<& 0$> A *
A A k>
A A »
^
A *\H B
r
A
-------
2-21
-------
Predicting the Toxicity of Metals in Sediments
by
Dominic M. Di Toro n, David J. Hanson**, Joy A. McGrathf, and Walter J. Berry5
faydroQual, Inc. Mahwah, NJ
^Environmental Engineering and Science Dept. Manhattan College, Bronx NY
§US EPA National Health and Environmental Effects Research Laboratory, Narragansett, RI.
"Great Lake Environmental Center, Traverse City MI
March 1999
Introduction
The SEM-AVS method for evaluating the toxicity of metals [1,2] has proven to be quite successful
at predicting the lack of toxicity in spiked and field contaminated sediments [3,4]. However, it does
not appear to be able to predict very well the onset of toxicity in a sequence of sediments spiked with
increasing concentrations of metals, or for sets of field contaminated sediments. In fact, in a recent
article [5] it was claimed that the empirically derived methods [6] are better at making this
prediction. The purpose of the note is to introduce a modification of the SEM - AVS procedure that
significantly improves the prediction of mortality. Additionally, the inability of the empirically
derived methods to make this prediction is demonstrated.
Theory
The Equilibrium Partitioning (EqP) model gives the prescription for the development of sediment
concentrations that predict the toxicity or lack of toxicity in sediments. The sediment concentration
C'sed that corresponds to a measured LC50 in a water only exposure of the test organism is
2-22
-------
Csed~KpCw (1)
where C*^ is the sediment concentration ( g/g dry wt), ]£ (L/kg) is the partition coefficient between
pore water and sediment solids, and C*w is the LC50 concentration ( g/L)[7]. For application to
metals that react with AVS to form insoluble metal sulfides, this formula becomes [1]
CSed=AVS + KPCw (2)
where AVS is the sediment concentration of acid volatile sulfides. The formula simply states that
since AVS can bind the metal as essentially insoluble sulfides, the concentration of metal in a
sediment that will cause toxicity is at least as great as the AVS that is present. The sediment metal
concentration that should be used is the SEM concentration since any metal that is bound so strongly
that IN hydrochloric acid cannot dissolve it is not likely to be bioavailable [2]. Of course, this
argument is just speculation, which is why so much effort has been expended to demonstrate
experimentally that this is actually the case [2-4, 8]. Therefore eq.(2) becomes
SEM= AVS + KPCW (3)
The basis for the SEM/AVS method is to observe that if the second term in eq.(3) is neglected then
the critical concentration is
SEM = AVS (4)
or the criteria for toxicity or lack of toxicity is
SEM/AVS = 1 (5)
or, equivalently
SEM - AVS = 0 (6)
The failure of the either the ratio condition (eq.5) or the difference condition (eq.6) to predict toxicity
is due to the neglect of the partitioning term Kp C*w- Note that ignoring the term does not affect the
prediction of lack of toxicity since it makes the condition conservative (i.e. smaller concentrations
of SEM are at the boundary of toxicity and no toxicity).
2-23
-------
The key to improving the prediction of toxicity is to approximate the partitioning term rather than
ignoring it. In anaerobic sediments, the organic carbon fraction is an important partitioning phase
and partition coefficients for certain metals at certain pHs have been measured [9]. This suggests
that the partition coefficient Kp hi eq.(3) can be expressed using the organic carbon based partition
coefficient, KQC, together with the fraction organic carbon in the sediment, f^.
Kp = Kocfoc (7)
Using this expression hi eq.(3) yields
SEM= AVS + KocfocC* (8)
Then moving the known terms to the left-hand side of the equation yields
(SEM- AVS)/foc = KOCCW (9)
If we knew both KOC and C*w we could use eq.(9) to predict toxicity. The method evaluated below
uses (SEM — AVS)/ f^ as the predictor of toxicity and evaluates the critical concentrations (the
right hand side of eq.9) based on observed SEM, AVS, foe, and toxicity data.
Data Sources
Data from toxicity tests using both laboratory-spiked and field-collected sediments were compiled
from the literature. Three sources of laboratory spiked tests using marine sediments were included,
Casas and Crecelius [10] and Berry et al. [3, 11]. Data reported included total metals,
simultaneously extracted metals (SEM), acid volatile sulfide (AVS), organic carbon fraction (f^ and
10 day mortality. In the study by Casas and Crecelius [10], the toxicity of zinc, lead and copper
were tested on the marine polychaete Capitella capitata. In Berry et al. [3] the toxicity of cadmium,
2-24
-------
copper, lead, nickel, zinc and a mixture of four metals (cadmium, copper, nickel and zinc) were
tested on the marine amphipod Ampelisca abdita. In Berry et al. [11] the toxicity of silver to A.
abdita was tested. Three sources for metal contaminated field sediments were included, Hansen et
al. [4], Kemble et al. [12], and Call et al. [13]. Data reported included total metals, SEM, AVS,
fraction organic carbon and mortality. The metals included in total SEM were cadmium, copper,
nickel, lead and zinc. In Hansen et al. [4], data were reported for five saltwater and four freshwater
locations. For the freshwater stations, organic carbon and total metals data were not provided.
Organic carbon data for one location, Keweenaw Watershed, were obtained separately [Berry,
personal communication]. Ten day bioassays were conducted using A. abdita and a freshwater
amphipod, Hyalella azteca. In Kemble et al. [12] 14-day bioassays were conducted on Chironomus
riparius, a freshwater midge. Data included from Call et al. [13] were the control freshwater
sediment data with 10-day mortality to the midge, Chironomus tentans.
Methods
\,
Laboratory spiked and field-contaminated sediment data were grouped together and analyzed as one
data set. Mortality data were compared against the SEM-AVS difference and this difference
normalized to the fraction organic carbon, (SEM-AVSyf,,,.. For each comparison two bounds were
computed for the SEM - AVS comparison and the (SEM - AVS)/ f^.: a lower bound concentration
equivalent to a 95 percent chance that the mortality observed would be less than 50 percent and an
upper bound concentration equivalent to a 95 percent chance that the observed mortality would be
greater than 50 percent. The lower bound limit was computed by evaluating the fraction of correct
»
classification starting from the lowest abscissa value. When the fraction correct dropped to below
95%, the 95* percentile was interpolated. The same procedure was applied to obtain the upper
bound. Upper and lower bounds were calculated for both SEM-AVS difference and the organic
carbon normalized difference.
Using the same data set, individual total metal concentrations were divided by the effects range
median (ERM) values of Long et al. [6] for that metal or metals and the mean quotient value was
2-25
-------
determined. This was compared to observed mortality. Upper and lower bounds were calculated
using the same method as described above.
Results
Mortality in the laboratory spiked and field-contaminated sediment tests were organism independent
when plotted against the SEM-AVS difference (Figure 1, top panel). The lower and upper
boundaries determined for the SEM-AVS difference were 1.7 and 190 mol/g respectively and are
shown on the top panel of Figure 1. A line indicating 50 percent mortality is also shown. These
boundaries can be used as predictive limits. For SEM-AVS values lower than 1.7 mol/g, there is a
95 percent chance that the observed mortality will be less than 50 percent. Similarly, for SEM-AVS
values greater than 190 mol/g, there is a 95 percent chance that the observed mortality will be greater
than 50 percent. The uncertainty falls in the range of SEM-AVS values between the upper and lower
bounds, which in this case are approximately two orders of magnitude wide. When normalized to
the organic carbon content of the sediment, the range of uncertainty is narrowed to approximately
one order of magnitude (Figure 1, bottom panel). The lower and upper boundaries are 86 and 2000
mol/goc respectively.
This same analysis is shown on a metal specific basis for cadmium, copper, nickel, lead and zinc in
Figures 2 through 6 together with the upper and lower boundaries determined above. It appears that
for each metal, the boundaries could be adjusted slightly to encompass a narrower range in
uncertainty. However, these metals all appear to be behaving similarly in this analysis. In fact, this
analysis was tested using the four metal mixture experiment (cadmium, copper, lead and nickel) and
the area of uncertainty fell within the boundaries (Figure 7) suggesting that these metals behave
similarly and it is appropriate to analyze them together. ,
A metal which appears to be more toxic than the other five metals is silver (Figure 8). Mortality was
observed at much lower SEM-AVS values. Upper and lower boundaries could not be computed for
silver due to lack of data at low SEM-AVS values that had an absence of mortality.
2-26
-------
Long et al. [5] have suggested that sediment quality guidelines (SQGs) based on dry weight
normalizations were equally if not more accurate, in predicting the non-toxicity or toxicity of
sediment associated metals than AVS-normalized SEM concentrations. It was concluded that using
the mean ERM quotient (metal concentration/ERM) was more effective at predicting whether or not
a sample was toxic compared to SEM/AVS ratios. The data set from Hansen et al. [4] was used.
/
We reexamine this question using the complete data set described above, which includes spiked as
well as field contaminated sediments. Upper and lower bounds corresponding to 95% correct
predictions for the mean ERM quotient were determined as described above. The results are shown
in Figure 9. The lower and upper bounds were 0.24 and 270 respectively. The range of uncertainty
is three orders of magnitude. By contrast, the range of uncertainty for the SEM-AVS difference was
two orders of magnitude, indicating that the AVS normalized SEM concentrations were more
accurate at predicting the absence or presence of metal-associated toxicity. However, normalizing
these concentrations to the sediment organic carbon content lowered the range of uncertainty to close
to one order of magnitude, therefore providing the most reliable method of predicting toxicity or
non-toxicity of sediment associated metals.
Conclusions
The use of the organic carbon normalized SEM-AVS as a predictor of toxicity reduces the
uncertainty of the prediction from three orders of magnitude using average ERM ratios, and two
orders of magnitude using SEM-AVS, to one order of magnitude using (SEM-AVS)/ f^. There
appears to be no basis for the claim that average ERM ratios are preferable.
References
1. Di Toro, D.M., et al., Toxicity of Cadmium in Sediments: The Role of-Acid Volatile Sulfide.
Environ..Tox. Chem., 1990. 9: p. 1487-1502.
2-27
-------
2. Di Toro, D.M., et al, Acid volatile sulfide predicts the acute toxicity of cadmium and nickel
in sediments. Environ. Sci. Tech., 1991. 26(1): p. 96-101.
3. Berry, W. J., et al., Predicting the toxicity of metals-spiked laboratory sediments using acid
volatile sulfide and interstitial water normalizations. Environ. Tox. Chem., 1996.15(12): p. 2067-
2079.
4. Hansen, D.J., et al, Predicting the toxicity of metals-contaminated field sediments using
interstitial concentrations of metal and acid volatile sulfide normalizations. Environ. Tox. Chem.,
1996.15(12): p. 2080-2094.
5. Long, E.R., LJ. Field, and D.D. MacDonald, Predicting toxicity in marine sediments -with
numerical sediment quality guidelines. Environ. Toxicol. Chem., 1998.17(4): p. 714-727.
6. Long, E.R., et al, Incidence of adverse biological effects within ranges of chemical
concentrations in marine and estuarine sediments. Environ. Management, 1995.19(1): p. 81-97.
7. Di Toro, D.M., et al, Technical basis for the equilibrium partitioning method for
establishing sediment quality criteria. 1991.12.
8. Di Toro, D.M., et al, Technical Basis for the Equilibrium Partitioning Method for
Establishing Sediment Quality Criteria. Environmental Toxicology and Chemistry, 1991.11(12):
p. 1541-1583.
9. Mahony, J.D., et al, Partitioning of metals to sediment organic carbon. Environ. Toxicol.
Chem., 1996.15(12): p. 2187-2197.
10. Gasas, A.M. and E.A. Crecelius, Relationship between acid volatile sulfide and the toxicity
of zinc, lead and copper in marine sediments. Environ. Tox.icol. Chem, 1994.13: p. 529-536.
11. Berry, W., et al, Predicting the toxicity of sediments spiked with silver. Environ. Toxicol.
Chem., 1999.18: p. 40-48.
12. Kemble, N.E., et al., Toxicity of metal-contaminated sediments from the upper Clark Fork
River, Montana, to aquatic invertebrates and fish in laboratory exposures. Environ. Toxicol. Chem.,
1994.13: p. 1985-1997.
13. Call, D. J., et al, Silver toxicity to Chrionomus tentans in two freshwater sediments. Environ.
Toxicol. Chem., 1999.18: p. 30-39.
2-28
-------
120
100
80
C. 60
40
20
-20
1
o Field Amphlpod 1
• V Field Midge *, • •? • • f
• Lab Amphlpod o *
DLabPolychaete °
o
• c
• o
•*$o
1
1
1
•1
'.•: °
1
, v 1
. 1
°o !
0 1
1
1
1
1
-
-
120
100
80
£. 60
40
20
-20
0.1
0.001 0.01 0.1 1 10 100
SEM-AVS (umol/g)
1000 10000
1
1
i o on i
i i
i
d °
lo
I
I
I
__ |_ _
I
I °
V 1 *° V
1
• TT V D
o ^Z> o^
o 1 o
V °l £> go
0 , iglr
1
1
1
1
H»fk •»•«• •••,• •
0 *• _ O
_ •
0 1 •
0 « S
1 •
1
'l
*>__•_
:\ •
*.••
V 1
0| V
* *l
1
• 1
1
1
-
•
.
•
—
.
-
.
'
10
100
1000
10000 100000 1000000
SEM-AVS/foc (umol/goc)
Figure 1. Mortality versus [SEM]-[AVS] difference (top panel) and
this difference normalized to organic carbon (bottom panel) on a
sum total SEM basis.
2-29
-------
120
100
80
£, 60
CADMIUM
£
o
40
20
-20
o Field Amphipod
- V Field Midge
• Lab Amphipod
O Lab Polychaete
"
•
, V
0 1
i
i
i
•••
•
o
£
1
O
f °
-
0
U — — — —
_
'
"
o
0.001 0.01 0.1 1 10 100 1000 10000
SEM-AVS (umol/g)
lifll
100
80
60
40
20
0
-20
•
-
-
— — -I
V
O
0
I 0_ _ 4
s
o
o
_
• • • -
•
•
0 *
"
_
0.01 0.1 1 10 100 1000 10000 100000
SEM-AVS/foc (umol/goc)
Figure 2. Mortality versus [SEM]-[AVS] difference (top panel) and
this difference normalized to organic carbon (bottom panel) for cadmium.
2-30
-------
120
100
80
«S 60
40
20
-20
o Field Amphlpod
V Field Midge
• Lab Amphlpod
D Lab Polychaete
COPPER
• • «o n«n 4
o o
o
o
0.1
°o
0.001 0.01 0.1 1 10 100 1000 10000
SEM-AVS (umol/g)
12U
100
80
60
40
20
0
-20
.
-
_
O
V
V
0
o
1 '
O O D *d
o
o
o
0
- v- ° -J
o
v *l
0°
• •• • •
o
.
"
•
10 100 1000 10000 100000 1000000
SEM-AVS/foc (umol/goc)
Figure 3. Mortality versus [SEM]-[AVS] difference (top panel) and
this difference normalized to organic carbon (bottom panel) for copper.
2-31
-------
120
100
80
e, eo
40
20
NICKEL
£
s
o Field Amphipod
- V Field Midge
• Lab Amphipod
O Lab Polychaete
V
V
0
o
•
f * 1 '
o
Q
° 0
o
o
•
t
• •
.
o
l_ — —
.
0.1
0.001 0.01 0.1 1 10 100 1000 10000
SEM-AVS (umol/g)
120
100
80
60
40
20
0
'
-------
120
100
80
C. 60
40
20
LEAD
I
-20
r o Field Amphipod
- V Field Midge
• Lab Amphipod
Q Lab Polychaete
j
V
v «
o
CJ
V .Q
"
-
-
0.0001 0.001 0.01 0.1 1 10 100 1000
SEM-AVS (umol/g)
12U
100
80
60
40
20
0
-20
_
"
_ j
V
V
V
o
a
L _ _ J
D •
•t •
-
-
_ _ _ .
0 *
-
0.01 0.1 1 10 100 1000 10000 100000
SEM-AVS/foc (umol/goc)
Figure 5. Mortality versus [SEM]-[AVS] difference (top panel) and
this difference normalized to organic carbon (bottom panel) for lead.
2-33
-------
120
100
80
£, 60
ZINC
o
40
20
•20
o Field Amphipod
- V Field Midge
• Lab Amphipod
HI Lab Polychaete
m
V* v
v v
o
-------
120
100
80
Sil 60
1
!
5
40
20
-20
120
100
80
60
20
-20
MIXTURE - LAB AMPHIPOD TEST
0.001 0.01 0.1 1 10 100 1000 10000
SEM-AVS (umol/g)
.Ii
0.1 1 10 100 1000 10000 100000 1000000
SEM-AVS/foc (umol/goc)
Figure 7. Mortality versus [SEM]-[AVS] difference (top panel) and
this difference normalized to organic carbon (bottom panel) for an
amphipod bioassay test spiked with a four metal mixture.
2-35
-------
SILVER -LAB AMPHIPOD TEST
£
£»
1
o
120
100
80
60
40
20
0
-20
• • • «
.
_ |
.
•
• •
•
_ _ _
9
•
•
.
-
_
'
€
0.0001 0.001 0.01 0.1 1 10 100 1000
SEM-AVS (umol/g)
120
100
80
£. 60
*
20
-20
,li
0.01 0.1 1 10 100 1000 10000 100000
SEM-AVS/foc (umol/goc)
Figure 8. Mortality versus [SEM]-[AVS] difference (top panel) and
this difference normalized to organic carbon (bottom panel) for silver.
2-36
-------
I
1201 r-n
100
80
60
40
20
0.001 0.01 0.1 1 10 100 1000 10000
Average Metal/ERM
Figure 9. Mortality versus the average Total Metal/ERM quotient
for field data (o) and laboratory spiked data (v).
2-37
-------
Chromium Chemistry in Sediments
by
Dominic M. Di Torotj and John D. Mahony*
^HydroQual, Inc. Mahwah, NJ
^Environmental Engineering and Science Dept. Manhattan College, Bronx NY
March 1999
Introduction
The purpose of this section is to present a review of the aqueous chemistry of
chromium as it influences its toxicity in sediments. The speciation chemistry of chromium is
presented in Fig.l, and Eh - pH diagram [1]. The principle species that exist at the indicated
Eh (redox potential) and pH is listed. Chromium exists in two oxidation states in natural
systems. Oxidized Cr(VI) species are present as oxyanions CrO42" and HCrCV. The reduced
Cr(ni) species are Cr3* and the hydroxide complexes CrOH2+, Cr(OH)2+, Cr(OH)3° and
Cr(OH)4~. Although chromium sulfide minerals are known to exit, e.g. 0283(5) [2] and 0*384
(Brezinaite) [3] they hydrolyze to chromium hydroxide upon exposure to water [4].
Cr(in) concentrations are regulated by an insoluble hydroxide Cr(OH)3(s) in the pH
range from 6 to 11 (the shaded region hi Fig.l). The actual concentrations can be seen in
Fig.2 from [5]. Solubility limits the dissolved concentration to approximately 10"7 M in the
pH range of approximately 6-7 to 10. The water quality criterion for 0(111) is
approximately 10"6 M (74 g Cr/L) [6]. Therefore in the pH range of approximately 6-7 to
10 we expect no toxicity since the interstitial water toxic unit concentration is approximately
10"7 M/10"6 M = 0.1. By contrast to Cr(IH), the oxidized Cr(VI) species are soluble and are
also toxic. The EPA chronic water quality criteria for Cr(VI) are 11 and 50 gCr/L in
freshwater and saltwater, respectively.
2-38
-------
Reduction of Cr(VI) to CrflU)
The reduction of Cr(VI) to CrflU) can occur only in reducing environments (Fig.l).
The reduction can occur by bacterial action [7-10], by ferrous iron [11-13], by organic matter
[14-16], photochemically [17], in soils [18], in aquifer material [19], by hydrogen sulfide [20,
21], pyrite fines [22] and amorphous iron sulfide [23]. We have also examined the reduction
of chromate using amorphous iron sulfide in neutral solutions in support of the effort to
develop an EqP based sediment criteria for chromium by the EPA.
x
For the data presented in Fig.3, the stoichiometry of the reaction appears to be
CrO42' + FeS + 3H2O -> Cr(OH)3(s) + FeOOH(s) + S° + 2OH' (1)
i.e. one mole of chromate reduced per mole of FeS initially present. We have also seen
titrations that correspond to complete oxidation of the sulfide to sulfate when sediments are
>
used. The important fact, from the point of view of sediment criteria, is that if any FeS is
present in sediments, then all the chromate would have been reduced to chromium hydroxide
by the reduction reaction (eq. 1).
Oxidation of Cr(HI) to Cr(VI)
The reverse reaction: the oxidation of Cr(III) to Cr(VI), is of direct concern because
Cr(VI) is soluble and toxic. The oxidation rate of Gfni) to CrCVI) with oxygen as the only
oxidant is quite slow with half-lives of months reported (see [24] for a review). Oxidation
can occur rapidly (in minutes) with hydrogen peroxide as the oxidant [24] leading to the
suggestion that the photochemical production of H2O2 in surface waters can be producing
Cr(VT). For sediments, however, the likely oxidant is manganese dioxide, which has been
shown to oxidize Cr(IH) to Cr(VI) rapidly [25-29].
At acid pHs the stoichiometry of the reaction appears to be [30]
2-39
-------
2* 2+
3 MnO2(s) + 2 Cr(OH)* -> 3 Mn+ + 2HCrO4- (2)
which goes to completion very quickly (minutes to hours). We have modeled the reported
data for pH = 4.5 assuming that the oxidation rate is 1/2 order with respect-to Cr3+, which
suggests a surface mass transfer limitation, and saturation kinetics for MnO2.
The kinetic equation for the concentration, Cr, of Cr(ni) is
(3)
^>
.
it
where Cr and MnO2 are both functions of time, k is the reaction rate constant and KMnO2 is
the half saturation constant. The concentration of MnO2 is determined by mass balance based
on the stoichiometric eq.(2). The data for experiments at varying MnO2(s) concentrations are
shown in Fig.4. Both the Cr(VI) and Mn(H) produced by the oxidation are shown. The
Mn(D) concentrations are multiplied by 2/3 from eq.(2) so that their concentrations should be
the same as the Cr(VI). Other data are available for pH = 3-5 which indicates that the
oxidation rate increases with pH [29].
The situation at neutral pHs is less clear. Since Cr(in) forms in insoluble Cr(OH)3(s),
the issue is whether chromium hydroxide can be oxidized by MnO2(s). Takacs (1988) reports
a slow rate of oxidation from a single experiment. However Johnson and Xyla [27] report
that the rate is independent of pH. It is clear that further experimentation is required to settle
this important issue.
Sorption of Cr(III) and Cr(VI)
The sorption of Cr(in) and Cr(VI) are important reactions that limit the
bioavailability of chrome to organisms in both the water column and sediment. In addition,
complexed forms of chrome can diffuse from the pore water of the sediment to the overlying
2-40
-------
water. The sorption of Cr(VT) to hydrous iron oxide has been studied extensively and the
surface complexation model has been shown to apply [31-33].
The modeling of the sorption of Cr(in) is complicated by the precipitation kinetics of
Cr(OH)3. Studies of Cr(ni) sorption to silica have been carried out at low pHs [34, 35]. We
have analyzed a set of sorption data generated at neutral pH [36] with a partitioning model.
Varying amounts of DOC and suspended matter were added to solutions of Cr(in) at initial
concentrations of Cr(m>r = 1.0, 0.5, 0.2 and 0.1 mg/L. Dissolved Cr(m) and Cr(VI)
concentrations were measured. No Cr(VI) was detected, indicating that no significant
oxidation occurred. However, the concentration of dissolved Cr(in) varied systematically
with increasing DOC and suspended solids. The results of a linear partitioning model which
considers the species: Cr3"1", Cr=DOC, and Gr=SS, and Cr(OH)3(s) is shown in Fig.5.
The linear partitioning between Cr3"1" and Cr(OH)3(s) is modeling the initial stages of
precipitation of Cr(OH)3(s). Fig.SA (DOC = 9.4 mg/L) and Fig.SB (DOC = 24 mg/L)
present the data and the model results. The total dissolved chrome is plotted versus the
suspended solids (SS) concentration. Each line represents a different initial concentration of
Cr(DT)r, which is plotted, on the left edge of the plot as a star for visual reference. Fig.SC
presents the concentrations of species computed from the model for one case: Cr(Iir>r =1.0
mg/L and DOC = 24 mg/L.
Chromium Cycle
A summary of the chromium cycle in natural waters is presented in Fig.6 [1]. The cycling
from Cr(VT) to Cr(ffl) in sediments is illustrated, as is the oxidation of Cr(in) to Cr(VI) by
manganese dioxide MnOa- It is this latter reaction that is the only source of concern.
Preliminary experiments that we have performed indicate that the reaction is extremely slow,
and is limited by the solubility of Cr(OH)3(s). Therefore we do not expect that this is a
source of concern.
2-41
-------
The formation of Cr(OH)s(s) lowers that dissolved concentration at SS=0 below
Cr(ni)T. Note that if the normal solubility product formulation were applied, [Cr34] =
10'7 M and since [Cr=DOC] = KCTDOC [Cr^fDOC], the dissolved Cr(in) = [Cr3*] +
[Cr=DOC] would also be constant. The fact that dissolved chrome is varying is a
consequence of the system not having reached equilibrium with respect to Cr(OH)3(s). As
shown in Fig.SC, most of the dissolved chrome is complexed to DOC and, therefore, is
presumably, not bioavailable [37]. The results of this modeling exercise are sorption
constants of Cr(m) to DOC and SS.
2-42
-------
References
1. Richard, F.C. and A.C. Bourg, Aqueous Geochemistry of Chromium: A Review. Wat.
Res., 1991.25(7): p. 807-816.
2. Wells, A.F., Structural Inorganic Chemistry. 3rd ed. 1962, London: Oxford
University Press. 1055.
3. Vaughan, DJ. and J.R. Craig, Mineral Chemistry of Metal Sulfides. 1978, Cambridge,
UK: Cambridge Univerity Press.
4. Durrant, PJ. and B. Durrant, Introduction to Advanced Inorganic Chemistry. 1970,
London: Longman.
5. Rai, D., B.M. Sass, and D.A. Moore, Chromium(IU) hydrolysis constants and
solubility ofchromium(III) hydroxide. Inorg. Chem., 19,87,26: p. 345-349.
>,
6. EPA, Ambient Aquatic Life Water Quality Criteria for Chromium EPA 440/5-84-
029. 1984, WashingtonJDC, 20460: US Environmental Protection Agency. 99.
7. Ohtake, H. and Hardoyo, New Biological Method for Detoxification and Removal of
Hexavalent Chromium. Water Science and Technology, 1992. 25: p. 395-402.
8. Schmieman, E.A., et a/., Bacterial reduction of chromium. Applied Biochemistry
and Biotechnology, 1997. 63-5: p. 855-864.
9. Blake, R.C., et al., Chemical transformation of toxic metals by apseudomonas strain
from a toxic waste site. Environmental Toxicology and Chemistry, 1993. 12: p. 1365-1376.
10. Wang, Y.T. and H. Shen, Modelling cr(VJ) reduction by pure bacterial cultures.
Water Research, 1997. 31: p. 727-732.
11. Eary, L.E. and D. Rai, Kinetics ofchromate reduction by ferrous ions derived from
hematite and biotite at 25 C. Am. J. Sci., 1989. 289: p. 180-213.
12. Buerge, I.J. and SJ. Hug, Kinetics andph dependence of chromium(VI) reduction by
ironfll). Environmental Science & Technology, 1997. 31: p. 1426-1432.
u
13. Fendorf, S.E. and G.C. Li, Kinetics of chromate reduction by ferrous iron.
Environmental Science & Technology, 1996. 30: p.' 1614-1617.
14. Wittbrodt, P.R. and C.D. Palmer, Reduction ofCr(VI) in the presence of excess soil
fulvic acid. Environmental Science & Technology, 1995. 29: p. 255-263.
2-43
-------
15. Elovitz, M.S. and W. Fish, Redox interactions of cr(VI) and substituted phenols:
products and mechanism. Environmental Science & Technology, 1995. 29: p. 1933-1943.
16. Deng, B.L. and A.T. Stone, Surface-catalyzed chromium(VI) reduction: reactivity
comparisons of different organic reductants and different oxide surfaces. Environmental
Science & Technology, 1996. 30: p. 2484-2494.
17. Hug, S.J., H.U. Laubscher, and B.R. James, Iron(IH) catalyzed photochemical
reduction of chromium(VI) by oxalate and citrate in aqueous solutions. Environmental
Science & Technology, 1997. 31: p. 160-170.
18. ' Eary, L.E. and D. Rai, Chromate reduction by subsurface soils under acidic
conditions. Soil Sci. Soc. Am. J., 1991.55: p. 676-683.
19. Anderson, L.D., D.B. Kent, and J.A. Davis, Batch experiments characterizing the
reduction ofcr(VI) using suboxic material from a mildly reducing sand and gravel aquifer.
Environmental Science & Technology, 1994. 28: p. 178-185.
20. Smillie, R.H., K. Hunter, and M. Hunter, Reducion of Chromium(VI) by bacterially
produced hydrogen sulphide in a marine environment. WatRes., 1981.15: p. 1351-1354.
21. Pettine, M., et al., Effect of metals on the reduction of chromium (VI) with hydrogen
sulfide. Water Research, 1998. 32: p. 2807-2813.
22. Zouboulis, A.I., K.A. Kydros, and K.A. Matis, Removal of hexavalent chromium
anions from solutions by pyrite fines. Water Research, 1995. 29: p. 1755-1760.
23. Patterson, R.R., S. Fendorf, and M. Fendorf, Reduction of hexavalent chromium by
amorphous iron sulfide. Environmental Science & Technology, 1997. 31: p. 2039-2044.
24. Pettine, M. and F.J. Millero, Chromium speciation in seawater: The probable role of
hydrogen peroxide. Limnol. Oceanogr., 1990.: p. 730-736.
25. Schroeder, D.C. and G.F. Lee, Potential transformations of chromium in natural
waters. Water, Air, and Soil Pollution, 1975.4: p. 355-365.
26. Eary, L.E. and D. Rai, Kinetics of Chromium(IU) oxidation to Chromium(VI) by
reaction with Manganese Dioxide. Environ. Sci. Tech., 1987. 21: p. 1187-1193.
27. Johnson, C.A. and A.G. Xyla, The oxidation of chromium(III) to chrpmium(VI) on the
surface ofmanganite (gamma-MnOOH). Geochim. Cosmochim. Acta, 1991. 55: p. 2861-
2866.
28. Fendorf, S.E. and RJ. Zasoski, Chromium(III) Oxidation by delta-MnO2 .1.
Characterization. Environmental Science & Technology, 1992. 26: p. 79-85.
2-44
-------
29. Silvester, E., L. Charlet, and A. Manceau, Mechanism of Chromiwn(III) oxidation by
Na-Buserite. J. Phys. Chem., 1995. 99: p. 16662-16669.
30. Takacs, M.J., The Oxidation of Chromium by Manganese-Oxide: The nature and
Controls of the Reaction. 1988, Michigan State:
31. Dzombak, D.A. and F.M.M. Morel, Surface Complexation Modeling. Hydrous
Ferric Oxide. 1990, New York, NY: John Wiley & Sons. 1-393.
32. Mesuere, K. and W. Fish, Chromate and Oxalate Adsorption on Goethite .1.
Calibration of Surface Complexation Models. Environmental Science & Technology, 1992.
26: p. 2357-2364.
33. Mesuere, K. and W. Fish, Chromate and Oxalate Adsorption on Goethite .2. Surface
Complexation Modeling of Competitive Adsorption. Environmental Science & Technology,
1992. 26: p. 2365-2370.
34. Fendorf, S.E., et al., Mechanisms of chromium(III) sorption on silica .1. cr(III)
surface structure derived by extended x-ray absorption fine structure spectroscopy.
Environmental Science & Technology, 1994. 28: p. 284-289.
35. Fendorf, S.E. and D.L. Sparks, Mechanisms of chromium(III) sorption on silica .2.
effect of reaction conditions. Environmental Science & Technology, 1994. 28: p. 290-297.
36. Masscheleyn, P.H., et al, Chromium redox Chemistry in a Lower Mississippi Valley
Bottomland Hardwood Wetland. Environ. Sci. Tech., 1992.26: p. 1217-1226.
37. Campbell, P.G.C., Interactions between Trace Metals and Aquatic Organisms: A
Critique of the Free-ion Activity Model, in Metal Speciation and Bioavailability in Aquatic
Systems, A. Tessier and D.R. Turner, Editor. 1995, John Wiley & Sons:
2-45
-------
o
O)
-1
-2
-3
-4
-5
-6
-7
-8
-9
~160£~~4 6
pH
8 10 12 14
Fig.1. Stability fields for chromium species [Richard and Bourg, 1991]
WQC
h O 6-8d
V 18-22 d
O O
D 134d
va<>
Cr(OH)!
3 4 5 6 7 8 9 10 11 12 13 14 15
PH
Fig.2 (A) Stability fields for chromium species. (B) Solubility of chromium hydroxide
[Rai, Sass et al, 1987]
2-46
-------
2000
1500
Initial AVS = 475 nmol/L
500 1000 1500
Initial CiC>42~ (jimol/L)
2000
c
c
15
12
9
6
• Fig. 3. Titration of AVS with cbromate. The final chromate
concentration in solution is plotted vs the initial chromate concentration.
Line corresponds to a 1:1 stoichiometry for chromate reduction by FeS.
• O 4.6 uM
T V 9.1 uM
19 uM
1 I • I
• O 29 uM
T 290 uM
10 20 30 40 50
Time (min)
10 20 30 40 50
Time (min)
Fig. 4. Oxidation of Cr(III) by MnO2. Initial concentrations of MnO2 indicated in the legend. Open
symbols are Cr(VI) concentrations, closed symbols are 2/3 Mn(II) concentrations. The lines are
model computations.
2-47
-------
DOC = 9.4 mg/L
DOC = 24mg/L
Speciation
1.00
i.oo
0.01
0 200 400 600 BOO 1000
0.01
Cr-SS
0 200 400 600 800 1000
Suspended Sol kit (mj/L)
0 200 400 600 800 1000
Fig.5. Model of Cr(ni) sorption to suspended solids and DOC. Varying initial concentrations of CrQEh (see the
stars plotted on the y-axis). Plots of total dissolved chrome vs total suspended solids. Alternating filled and hatched
symbols represent the data for each Cr(ni)r. (A) DOC = 9.4 mg/L. (B) DOC = 24 mg/L. (C) An example of the
computed speciation for the DOC = 24 mg/L and CXm^ = 1.0 mg/L case.
input
CKVD
+ FeCn)or
org. matter
.— ,
\
\
weak
adsorption
Y
seating
1
diffusiop I
\ I
CrCVD
^
^
+MnQz
\
)
adsorption
or
precipitation
settling
CrOn^,
sedimentation
1 -
V
CKm)
• — 'diffusion, — •
-v
+ organic
matter
\
\x
Cr(m)-org
t
i
diffusion •
1
CrOID-org
+ dissolved/
organics
Fig.6. Chromium cycling in the aquatic environment [Richard and Bourg, 1991]
2-48
-------
2-49
-------
The Addition of Chromium to the Metals Mixtures ESG
by
Walter J. Berry
Warren S. Boothman
U.S.EPA, Atlantic Ecology Division, Narragansett RI
David J. Hansen
Great Lakes Environmental Center, Traverse City MI
March 1999
The usefulness of the measurement of acid volatile sulfide (AVS) as a tool to predict
biological effects is limited by the small number of metals for which it is currently applicable.
Therefore, it is of interest to expand the use of AVS to include predictions of biological effects for
as many metals as possible. One likely expansion candidate is chromium. AVS should be useful in
the prediction of biological effects from chromium in sediments because (1) chromium HI is
sparingly soluble and therefore nontoxic in typical freshwater and marine sediments that are
anaerobic, contain measurable concentrations of AVS, and have interstitial water pHs ranging from
about 6.5 to 11.5; (2) once formed, chromium HI does not oxidize to chromium VI; (3) soluble and
toxic chromium VI can only occur in sediments with no detectable AVS; and (4) only chromium VI
will occur dissolved in interstitial water. (See chromium section of the Metals Mixtures ESG
document.) The data available for assessing the usefulness of these predictions are discussed below.
Toxicity tests were conducted with the amphipod (Ampelisca abdita) exposed to sediment
spiked with chromium VI as potassium dichromate (Berry and Boothman, 1999). The authors
examined the lexicological implications of the reduction of toxic chromium VI to insoluble nontoxic
chromium HI (Kaczynski and Kleber, 1994) in anaerobic sediments. This reaction is
environmentally significant for the reasons described above. The results of these tests are reported
in detail here because the results are unpublished, and because so little other data are available from
experiments with sediments spiked with chromium.
Ten day toxicity tests with A. abdita were conducted by Berry and Boothman (1999) using
methods described by Berry et al. (1996). Sandy sediments from saltwater Ninigret Pond, RI
containing 1.7 //mole AVS/g and about 0.15% TOC were spiked to achieve nominal chromium
concentrations of 11 to 520 /^g/g. The pH of the spiking solution was adjusted with sodium
2-50
-------
hydroxide to 7.6 prior to addition to the sediment. Initial sediment interstitial water pH ranged
between 7.3 in the control and 8.2 in the highest concentration. A modification of the Cranston and
Murray (1978) method was used to determine concentrations of total chromium and chromium VI
in interstitial water. Chromium VI in simultaneously extracted metals (SEM) was determined using
the Wang et al. (1997) modification of the method of Cranston and Murray (1978).
In the sediments that were not toxic to A. abdita, all of the chromium in the SEM was present
as chromium HI; concentration ranged from 0.24 to 2.08 //moles/g. Sediments with 3.73 and 6.54
females total chromium/g were lethal (Table 1, Figure 1). In SEM of the two toxic sediments,
concentrations of chromium VI were 1.01 and 4.08 /^moles/g. Average measured total chromium
concentrations were not appreciably different from nominal concentrations. AVS concentrations in
nontoxic sediments ranged from 0.64 to 1.42 Aonoles/ g sediment, but AVS was not detected in toxic
sediments. The mean concentrations of AVS exceeded those of SEM in two of four nontoxic
chromium-spiked sediments (day 0-10), and three of four nontoxic sediments (after 10 days).
Chromium VI was absent from the interstitial waters of nontoxic sediments. IWTUs of chromium
VI explained observed toxicity: those sediments in which chromium was not detectable in the
interstitial water were not toxic, those with greater than one toxic unit were toxic.
The results of these tests were consistent with other chromium experiments performed with
amphipods at the U.S.EPA laboratory in Narragansett, RI. (W.Berry, unpublished data). Crffl was
not toxic in water-only tests in which the test solutions were buffered to ambient sea water pH, but
were toxic if the test solutions were not buffered (pH < 6.1, CrlH > 40,000 |ig/L). Similarly,
sediments spiked with pH-adjusted CrJUI solutions were not toxic at concentrations up to 3,000 ng/g
dry weight. (Sediments with 30,000 ng/g dry weight were toxic. These sediments were
approximately 1/3 chromium HI precipitate by volume.)
Further confirmatory research is still needed. Planned testing includes experiments with
sandy and muddy saltwater sediments spiked with chromium HI and an experiment with a muddy
saltwater sediment spiked with chromium VI. The published water-only research must also be
examined further. Most water-only tests with chromium HI indicate a lack of toxicity when testing
is done within normal interstitial water pH ranges at concentrations below chromium in solubility.
Some of these data, however, are difficult to evaluate and some may indicate the presence of toxicity
at normal pHs.
Taken together the results summarized above indicate that chromium in sediments is not
toxic if AVS is present and that chromium IWTUs can be used to identify nontoxic sediments.
Conversely, sediments containing measurable concentrations of chromium in SEM when AVS is not
present or when IWTUs are of toxicological significance might be toxic. This approach does not
apply to unique sediments having interstitial water pHs of less than about 6.5 or greater than about
11.5 that may occur in water bodies with low water column pHs or near acid mine drainage.
2-51
-------
References
Berry W.J. and W.S. Boothman. 1999. Results of toxicity tests with Ampehsca abdita exposed to
chromium-spiked sediments. Memorandum to the record. February 17, 1999.
Berry, W.J., D.J. Hansen, J.D. Mahony, D.L. Robson, D.M. Di Tore, B.P. Shipley, B. Rogers and
J.M. Corbin. 1996. Predicting the toxicity of metals-spiked laboratory sediments using acid-volatile
sulfide and interstitial water normalization. Environ. Toxicol. Chem.. 15:2067-2079.
Cranston, R.E. and J.W. Murray. 1978 The determination of chromium species in natural waters.
Anal. Chim. Acta. 99:275-282.
Kaczynski, S.E. and R.J. Kleber. 1994. Hydrophobic CIS bound organic complexes of chromium
and their potential impact on the geochemistry of chromium in natural waters. Environ. Sci. Technol.
28: 799-804.
Masscheleyn, P. H., J.H. Pardue, R.D. DeLauna and W.H. Patrick, Jr. 1992. Chromium redox
chemistry in a lower Mississippi Valley bottomland. Environ Sci Technol. 26: 1217-1226.
Wang, W-X, S.B. Griscom and N.S. Fisher. 1997. Bioavailability of Cr (III) and Cr (VI) to marine
mussels from solute and particulate pathways. Environ. Sci. Technol. 31:603-511.
2-52
-------
Figure Legend
Figure 1. Results of a toxicity test with the amphipod toxicity test with the amphipod Ampelisca
abdita exposed to sediments spiked with chromium VI. Top panel: Percent amphipod mortality vs
SEM-AVS. Middle panel: Percent amphipod mortality vs. CrVI interstitial water toxic units.
Bottom panel: Interstitial water toxic units (IWTU) vs. SEM-AVS.
2-53
-------
Table 1. -Results of a toxicity test with the amphipod Ampelisca abdita exposed to sediments spiked with chromium VI.
Treatment,
/-onoles Cr VI
added/yumol
AVS
Control
0.215
0.464
1.0
2.15
4.64
10
Nominal
chromium,
//g/gdry
wt.
-
11
24
52
110
240
520
, SEM, /tfnoleg/g
Total Cr
ND2
0.24
0.55
1.28
2.08
3.73
6.54
CrVI
ND2
ND2
ND2
ND2
ND2 ,
1.01
4.08
AVS,
;umole/g
1.42
1.31
1.30
0.92
0.64
ND
ND
SEM-
AVS
-1.42
-1.07
-0.75
0.36
1.44
3.73
6.54
Interstitial Water,
mg/L
Total Cr
ND2
ND2
ND2
ND2
?
215
725
CrVI
ND2
ND2
ND2
ND2
ND2
201
681
IWTUs1
sO.02
sO.02
*0.02
*0.02
sO.02
109
368
Percent
Mortality
0
0
0
5.0
2.5
72.5
97.5
1IWTU: Interstitial Water Toxic Units = Chromium concentration in interstitial water -5- 10-day LC50 for Ampelisca abdita
(1.85 mg/L).
2 ND= not detectable: sl.O /^g SEM chromium/g dry weight sediment; s0.04mg chromium/L; iO.l^mole AVS/g sediment.
cs
-------
Chromium Sediment Experiment
Mortality vs. SEM-AVS
(0
t:
o
I
0)
Q.
100 -
50 -
-5
—,_» — ,_
0 5
SEM-AVS
10
Mortality vs. CR6 IWTU
100 -
£ 50 H
-------
The Addition of Silver to the Metals Mixtures ESG
by '
Walter J. Deny
U.S.EPA, Atlantic Ecology Division, Narragansett RI
March 1999
Silver forms a highly insoluble sulfide. This fact makes it a natural candidate for inclusion
into a sediment guideline which uses acid volatile sulfide (AVS) normalization. The need for
normalization and the technical basis for the prediction of the effects of silver in sediments using
AVS and interstitial water normalizations has been recently described by Berry et al. (1999), and so
will only briefly be described here.
Previous experiments conducted with freshwater sediments spiked with silver have shown
that, when expressed on a dry weight basis, the toxicity of silver is sediment-specific and dependent
on the form of silver added (e.g. AgNO3, Ag2S: Figure Ib). Berry et al. (1999) assessed the
usefulness of silver interstitial water toxic units (TWTU) and acid volatile sulfide (AVS)
concentrations in predicting the biological effects of silver species across sediments, regardless of
the species of silver present. Two saltwater sediments were spiked with a series of concentrations
of silver. The amphipod, Ampelisca abdita, was then exposed to the sediments in ten-day toxicity
tests. Amphipod mortality was sediment-specific when expressed on a dry weight basis, but not
when based on IWTU or simultaneously extracted metal (SEM) - AVS (Figure 2). Sediments with
an excess of AVS relative to SEM had IWTU <0.5, and were generally not toxic. Sediments with
an excess of SEM relative to AVS had silver IWTU X3.5, but no measurable AVS, and were
generally toxic. Sediments with measurable AVS were not toxic. Re-analysis of the previously
published data from the freshwater sediments spiked with silver showed mortality to be correlated
with nominal SEM-AVS and with silver IWTU (Figures la and figure 3). Taken together, these
results support the use of AVS and silver IWTUs in predicting the toxicity of silver in sediments.
Silver is slightly different from the other metals originally included in the Metals Mixtures
ESG (cadmium, copper, lead, nickel, and zinc) because it is essentially monovalent in nature. For
this reason one half of the silver concentration is what is used for comparison to AVS. Another
difference is that silver forms a sulfide that is not soluble in the normal AVS extraction. These
differences affect the way that AVS normalization is used for silver.
Consider the case of a sediment that is contaminated with silver and other sulfide-formmg
metals. The possible contingencies are summarized hi Table 1. A mole of sulfide will be bound for
2-56
-------
every two moles of silver (the 1:2 ratio is due to the fact that silver is monovalent) present because
the sulfide solubility product of silver is lower than that of other metals (Lide, 1995). If sulfide
exceeds the sum of the [Ag]/2 plus the other SEM metals (cadmium, copper, lead, nickel, and zinc),
measurable AVS will exceed measurable SEM; therefore, these metals should not be present in
lexicologically significant concentrations hi the interstitial water, and the sediment should not-be
acutely toxic due to these metals. If the sum of the [Ag]/2 plus the other SEM metals exceeds the
sulfide in the sediment, measurable SEM will exceed measurable AVS, and the sediment may be
acutely toxic due to these metals. Sediments which do not have lexicologically significant amounts
of these metals in the interstitial water should not be toxic due to these metals, even if SEM exceeds
AVS in the sediment. However, any sediment in which SEM exceeds AVS should be looked at
carefully. Sediments which do have toxicologically significant amounts of metals in the interstitial
water are certainly potentially toxic due to these metals (Berry et al., 1996).
Theoretically, almost all of the silver hi the sediment will be bound to sulfide in any sediment
hi which there is an excess of sulfide over [Ag]/2. This is because of the extremely low solubility
of silver sulfide (Lide, 1995). However, if the sum of the [Ag]/2 plus the other SEM metals exceeds
the sulfide in the sediment, some of the other SEM metals may be present hi the interstitial water and
the sediment may be toxic due to these metals. In these sediments the silver may be contributing to
the overall toxicity of metals in the sediment, by tying up sediment sulfide which might otherwise
bind the other SEM metals. This is apparently what happened in some of the Call et al. (1999)
sediments, in which zinc and copper were released into interstitial water due to the addition of silver
to the sediment.
The release of metals into the interstitial water in relation to sulfide solubility is not peculiar
to silver. Berry et al. (1996) found that metals appeared in the interstitial water in the order of their
K^, with copper appearing last in their experiments. What is unusual about silver, however, is that
the solubility of silver sulfide hi the AVS extraction is so low that any sediment with an excess of
[Ag]/2 over sulfide will have no measurable AVS present. Thus, any sediment with measurable
AVS should not have silver hi the interstitial water, and should not be acutely toxic because of silver.
It is important to remember that the data presented here apply primarily to acute mortality,
and may not address all effects due to chronic exposure or bioaccumulation (Luoma et al., 1995).
For example, Hook and Fisher (1997) demonstrated hi preliminary experiments that silver may
affect copepod reproduction at environmental concentrations. However, taken together, the
freshwater and saltwater sediment results indicate that silver:AVS relationships and IWTU can
provide insight into the role of silver in the possible toxicity of sediments. From the point of view
of AVS and SEM measurements, these results would indicate that silver can be included along with
cadmium, copper, lead, nickel, and zinc in sediment assessment. If the sum of the SEM for these
metals is less than AVS in a sediment, the sediment should not be acutely toxic due to these metals.
Furthermore, even in sediments which have an excess of metal over sulfide, as long as there is
measurable AVS any observed acute mortality should not be due directly to silver in the interstitial
water.
2-57
-------
References
Berry, W.J., DJ. Hansen, J.D. Mahony, D.L. Robson, D.M. Di Toro, B.P. Shipley, B. Rogers and
J.M. Corbin. 1996. Predicting the toxicity of metals-spiked laboratory sediments using acid-volatile
sulfide and interstitial water normalization. Environ. Toxicol. Chem. 15:2067-2079.
Berry, W.J., M. G. Cantwell, P.A. Edwards, J.S. Serbst, DJ. Hansen 1999. Predicting toxicity of
sediments spiked with silver. Environ. Toxicol. Chem. 18:40-48.
Call DJ, Markee TP, Brooke LT, Polkinghorne CN, Geiger DL. 1999. Bioavailability and toxicity
of silver to Chironomus tentans in water and sediments. Environ Toxicol Chem. 18:30-39.
Hook S, Fisher N. 1997. Sublethal response of zooplankton to silver: the importance of exposure
route. Abstract. Eighteenth Annual Meeting of the Society of Environmental Toxicology and
Chemistry. San Francisco, Ca. November 16-20,1997.
Lide, D.R.., ed. 1995. CRC Handbook of Chemistry and Physics, 76th ed CRC, Boca Raton, FL,
USA, pp 8-85.
Luoma SN, Ho YB, Bryan GW. 1995. Fate, bioavailability and toxicity of silver in estuarine
environments. Mar Poll Bull 31:44-54.
Rodgers JA Jr, Deaver E, Rodgers PL. 1997. Partitioning and effects of silver in amended
freshwater sediments. Ecotoxicol Environ Saf 37: 1-9.
2-58
-------
Figure Legends
Figure 1. Percentage mortality of the amphipodAmpelisca abdita as a function of dry weight silver
concentration (A), ([Ag]/2) - AVS (B), interstitial water toxic units (IWTU) (C), and measured AVS
(D) in two saltwater sediments spiked with silver. Nin = Ninigret Pond sediment. Pojac = Pojac
Point sediment. Sediments below the dashed line at 24% mortality are not considered toxic.
Vertical dashed lines at SEM-AVS = 0 (B) and IWTU = 0.5 (C) indicate predicted break points in
toxicity. Data points believed to be the result of interstitial water ammonia are included but
highlighted and not connected by lines in A. AVS detection limit is indicated as "ND" in D. (After
Berry et al., 1999)
Figure 2. Percentage mortality of the amphipod Hyalella azteca as a function of nominal ([Ag]/2)-
AVS (A) and dry weight silver concentration (B) in four freshwater sediments spiked with silver.
(From Rodgers et al., 1996). Sediments below the dashed line at 24% mortality are not considered
toxic. A vertical dashed line at SEM-AVS = 0 (A) indicates the predicted break point in toxicity.
Figure 3. Percentage mortality of the midge Chironomus tentans as a function of nominal
([Ag]/2) - AVS (A), IWTU (B) and dry weight silver concentration (C) in a freshwater sediment
spiked with silver. (From Call et al., 1999). Sediments below the dashed line at 24% mortality are
not considered toxic. Vertical dashed lines at SEM-AVS = 0 (A) and IWTU = 0.5 (B) indicate
predicted break points in toxicity.
2-59
-------
Table 1. Contingency table for predicting toxicity due to metals from measurements of silver,
other SEM metals, and AVS in laboratory-spiked and field sediments. (From Berry et al., 1999)
Nominal Metal and AVS in
Sediment
Measured Metal and AVS
in Sediment
Prediction of Acute
Toxicity
Silver Only
[Ag]/2 < AVS
[Ag]/2 > AVS
AVS > Detection limit
and
(SEM-AVS) < 0.0
AVS < Detection limit
and
(SEM-AVS) > 0.0
Sediment not acutely toxic
due to silver. No metals
detectable in interstitial
water.
Sediment may be acutely
toxic due to silver (but not if
IWTU < 0.5).
Metals Mixtures
([Ag]/2+[Cd]+[Cu]+[Ni]
+[Pb]+[Zn]) < AVS
([Ag]/2+[Cd]+[Cu]+[Ni]
+[Pb]+[Zn]) > AVS
but
[Ag]/2 < AVS
([Ag]/2+[Cd]+[Cu]+[Ni]
+[Pb]+[Zn]) < AVS
and
[Ag]/2>AVS
AVS > Detection limit
and
(SEM-AVS) < 0.0
AVS > Detection limit
and
(SEM-AVS) > 0.0
AVS < Detection limit
and
(SEM-AVS) > 0.0
Sediment not acutely toxic
due to these metals. No
metals detectable in
interstitial water.
Sediment may be acutely
toxic due to these metals, but
not silver directly (but not if
sum of IWTU < 0.5).
Sediment may be acutely
toxic due to silver and /or the
other metals (but not if sum
of IWTU < 0.5).
2-60
-------
A
&
•S
o
2
xP
s-
inn
1UU
80-
60-
40-
20-
n.
• Sed1
• Sed3
ASedG
• Sed7
••-
._ .
J A
N
f
II
*
I
j'
i 1 - 1- I .
-2
-1
[Ag/2]-AVS (^Moles/g)
a-
1
O
&
+&
80-
60-
40-
— *— Sed1
-^-Sed3
— A— Sed6
— x— Sed7
/
r
/
0.1 1 10
Dry wt silver
100
1000
2-61
-------
10 100 1000 10000 100000
Silver (ng/g Dry Wt)
"n
t
O
S
0
80-
60-
40-
20-
n.
4. Pojad
X Pojac3
+ ,*
1 ^
I
I
I
1
4
x*<
-30 • -20
20
[Ag/2]-AVS
0.01
0.1 1
IWTU (LC50 = 20
10
40
c
*co
|
5S
100 -,
80-
40-
20-
0
^Nin
4. Pojad
X PojacS
X
(\?) \f)/ A f*
I *
I
I
A 4_
+ 1 x r A
T » ! 1 *J
100
uu-
80-
60-
40-
20-
n,
ANin
4. Pojad
X PojacS
T ND
I
1 JL •
$ • * *
0.01
0.1 1
AVS (iiMoles/g)
10
100
2-62
-------
100
80-
75 60-
40-
o
20-
-10 0 10 20 30 40 50
[Ag/2] - AVS (|aMoles/g)
B 100
80 f
o
40f
20
0.001 0.01 0.1 1 10
Interstitial Water Toxic Units (IWTU)
v*
£*
O
*
• uu -
80-
60-
40-
20-
n.
•
"•
i
*
•
• *
0.01 0.1 1 10 100 1000 10000
Silver (Ug/g Dry Wt)
2-63
-------
IMPORTANT NOTE
Section 3 (A Model of the Acute Toxicity of Metals) of the Integrated Approach to Assessing the
Unavailability and Toxicity of Metals in Surface Waters and Sediments will be mailed to reviewers
on March 15,1999 as an addendum to this document.
-------
DRAFT
- DRAFT-
A BIOTIC LIGAND MODEL OF THE ACUTE TOXICITY OF METALS
in. APPLICATION TO FISH AND DAPHNIA EXPOSURE TO SILVER
by
PAUL PAQUINA, DOMINIC Di Tono*3, ROBERT SANTOREC,
DEVANSHI TRTVEDI* AND BENJAMIN WuA
A HydroQual Inc., 1 Lethbridge Plaza, Mahwah, NJ 07430
B Environmental Engineering Department., Manhattan College
4513 Manhattan College Parkway, Bronx, NY 10471
c HydroQual Inc., 4914 West Genesee Street, Suite 119, Camillus, NY 13031
March 23, 1999
-------
Abstract- A Biotic Ligand Model of the Acute Toxicity of Metals. III. Application to Fish and
Daphnia Exposure to Silver. When silver is discharged to a water body, speciation and complexation
reactions control its distribution among various organic and inorganic complexes and as ionic silver. It is
the distribution of the silver among these forms, as well as the other water quality characteristics of the
system, that control its bioavailabihty. The bioavailabihry can be evaluated in the context of a biotic
ligand modeling framework. In this mechanistically based framework, the biotic ligand, the organism
tissue at the site of action, is represented in the same way as any other ligand in solution. It has a
characteristic binding site density and a conditional stability constant for each species that it reacts with.
The biotic ligand model simultaneously accounts for the speciation and complexation of dissolved silver
and competitive binding of silver and other cations at the site of action. The organism LC50 corresponds
to the point where silver accumulation at the biotic ligand reaches a critical level. This paper describes
the version of the biotic ligand model developed for silver, as applied to several data sets for fish and
invertebrates. The model is first used to analyze fish gill silver accumulation data over ranges of
hardness, DOC, chloride, pH and alkalinity. It is then used to account for the variation in acute toxicity
of silver to both fish and invertebrates in bioassays where hardness, DOC and chloride levels were
varied systematically. The biotic ligand model has a number of potential applications including use in
prediction of water effect ratios (WERs) from site water chemistry, development of water quality criteria
for metals, and more generally, for ecological risk assessment purposes.
Key Words- Silver Biotic ligand model Bioavailability Speciation Toxicity
3-60
-------
INTRODUCTION
It is generally accepted that total metal concentration is not a good measure of exposure that can be
directly related to biological effects, in either the water column or the sediment. This acceptance arises
from repeated demonstrations that consideration of the form of a chemical stressor and its
bioavailabiliry are pre-requisites for the successful prediction of effects (Di Toro et al., 1991; Ankley et
al., 1994 and 1996; Allen and Hansen, 1996). The importance of considering bioavailability in
assessing ecological impacts has been recognized by both regulatory authorities and the scientific
community (Prothro, 1993; Renner, 1997; Bergman and Dorward-King, 1997). As a result, much
research has been carried out during recent years to obtain an improved understanding of
bioavailability and the mechanisms of toxicity, and in developing modeling frameworks that are
appropriate for use in assessing the environmental fate and effects of metals, including silver, in aquatic
systems.
This last in a series of three papers describes the biotic ligand model (BLM), a generalized framework
for assessing the bioavailability and acute toxicity of metals in aquatic systems, as developed and
applied for silver. A version of the model is presented that is developed to predict the acute toxicity of
silver to both fish and Daphnia. The model framework as structured for silver is first described. It is
then used to analyze silver accumulation data from experiments performed with fish in laboratory and
natural waters. This analysis will illustrate how the model predicts changes in silver accumulation at
the gill, the site of action of acute toxicity in fish, that result from changes hi water quality
characteristics. The framework is then applied to several bioassay data sets to refine the parameter
values used in the model to account for the effects of variation in water quality (e.g., hardness, DOC,
chloride, pH and alkalinity) on the acute toxicity of silver to fish. Finally, the application of the model
to an invertebrate, Daphnia magna, is presented.
3-61
-------
BACKGROUND
Description of Model
The conceptual framework for the biotic ligand model is an adaptation of the Gill Surface Interaction
Model (GSIM) originally proposed by Pagenkopf (1983) and applied by him to several metals. A
detailed description of the generalized biotic ligand model framework, consisting of chemical speciation
and acute toxicity sub-models, and the technical basis of the model is described elsewhere (Di Toro et
al., 1999). Briefly, the model is based on the premise that toxicity is not simply related to the total metal
concentration in solution. Rather, metal speciation and complexation reactions and interaction of the
metal and other cations at the site of action of toxicity must be considered as well. The role of
complexation is critical, since formation of both organic and inorganic metal complexes renders a
significant fraction of the total metal non-bioavailable.
The version of the biotic ligand model developed for silver is illustrated on Figure 1. The dissolved
silver exists in solution as the free silver ion (Ag+) and as a variety of organic and inorganic metal
complexes (Figure 1, left side). The free silver ion typically represents a relatively minor fraction of the
total silver in solution. However, it is this free metal species that is hypothesized to control the degree of
biological effects in the free ion activity model (FIAM) of toxicity, either by direct interaction at the site
of action, or indirectly, via its role in the formation of other silver complexes (Morel, 1983; Campbell,
1995). These other complexes form as a result of reactions of free silver with the other organic and
inorganic ligands present in solution, and collectively, they represent the predominant form of silver in
solution. With limited exceptions (e.g., it will be shown that AgCl appears to be toxic to fathead
minnows), they are typically not considered to be bioavailable.
The metal speciation computations needed for the chemical speciation sub-model could be performed
using any of a number of alternative chemical equilibrium models. For example, the CHEMICAL
Equilibria in Soils and Solutions model, CHESS (Santore and Dnscoll, 1995), or EPA's MINTEQA2
model (Allison et al., 1991; Brown and Allison, 1987) or any one of a number of programs (Westall et
al., 1976) are available for use. Though Pagenkopf (1983) recognized the ameliorating effect of organic
matter on toxicity, the effect of organic matter was neglected in this early attempt to model metal-gill
3-62
-------
interactions, since it was to be applied to data obtained using water that was low in organic matter
content. However, dissolved organic matter is frequently judged to be important, and in such instances,
the Wmdermere Humic Aqueous Model (WHAM; Tipping, 1994 and 1996) offers a relatively detailed
framework for evaluating metal-organic matter interactions. Given that it has been tested against an
impressive number of published data sets and with a wide variety of ambient waters, the Environmental
Chemistry Workgroup at the Pellston Workshop on Reassessment of Metals Criteria for Aquatic Life
Protection, recommended that Model V, WHAM Version 1.0 (Tipping, 1994), be used to represent
metal-organic matter interactions (Kramer et al., 1997). The chemical speciation sub-model employed
herein represents a synthesis of the advanced representation of metal-organic matter interactions of
WHAM with the chemical equilibrium framework of CHESS. However, since the requisite WHAM
stability constants have not been previously determined for silver, and a suitably designed experimental
data set is not currently available for use in performing this evaluation, it was necessary to perform a
preliminary evaluation of the requisite inputs as part of the calibration process to be described.
The remaining component of the biotic hgand model is the acute toxicity sub-model. For fish, the site of
action of acute toxicity is the gill (McDonald et al., 1989). Silver toxicity is caused by binding of silver
at physiologically active, functionally important sites on the surface membrane of the gill. This
interaction of free silver and other bioavailable silver species at the gill results in an iono-regulatory
disturbance that manifests itself in the form of an acutely toxic effect. Specifically, in the case of
freshwater fish such as rainbow trout, silver toxicity results from disruption of the iono-regulatory
processes that control the active transport of ions such as Na*, Cl" and Ca2+ across the gill (Morgan et al.,
1996; Wood et al., 1996a and 1999). Janes and Playle (1995) proposed a conceptual model that could be
used to predict the degree of silver binding to gills of rainbow trout (Figure 1, right side) and conducted
studies to investigate the factors that control the degree of accumulation. Given that the acute toxicity of
silver to fish was understood to be related to accumulation of silver at the gill, this conceptual approach
seemed well suited for use in a model that could be used to predict the acute toxicity of silver to fish. All
that was needed was a way to relate the degree of accumulation to the organism response. As described
herein, this remaining step is accomplished by application of the model to bioassay test results.
In the context of the biotic hgand model, binding of the free metal ion at the site of action - the
membrane at the gill surface in the case of fish - is analogous to formation of a metal complex in
3-63
-------
solution, where the tissue at the site of action is viewed as a "biotic hgand." The metal-biotic hgand
interaction is represented in the same way as any other reaction of a metal with an organic or inorganic
hgand. Both the binding site density of the biotic hgand, which is analogous to the concentration of an
inorganic hgand in solution, and the conditional stability constant for the metal-biotic hgand complex
and for any other cations that interact with the biotic hgand, must be specified.
Experimental results for use in developing initial estimates of model parameter values needed to predict
metal accumulation on the total fish gill, including sites that may or may not be active physiologically,
are currently available for a number of metals, including copper and cadmium (Playle et al., 1993a and
1993b) and silver (Janes and Playle, 1995). The results of these tests provide an indication of the levels
of complexing hgands and competing cations that will affect accumulation of metal at the gill. With
respect to copper, direct evidence also exists that demonstrates a relationship between the short term gill
uptake of copper and acute toxicity (MacRae, 1994; reviewed by Santore et al., 1999). The situation is
less well defined for silver, however, with recent experimental results showing that the level of silver
accumulation on the gill vanes over time and that static equilibrium conditions are not necessarily
achieved (Wood et al., 1999). Moreover, only a relatively small fraction of the measured total
accumulation of silver at the gill is associated with binding to physiologically active sites (Wood et al.,
1999). As a result of these factors it has been difficult, at least in the case of silver, to establish empirical
evidence of a direct relationship between total gill silver accumulation and acute effects. Given the
difficulties associated with establishing a definitive relationship between silver accumulation and
toxicity, this is still an area of active research. It is useful, therefore, to review the results of some recent
studies prior to a review of the model being developed for silver.
Relationship of Silver Accumulation to Acute Toxicity
Several studies have shown that when juvenile fathead minnows and rainbow trout are exposed to
copper, there is a relatively rapid increase above background levels of copper bound to the gill (Playle et
al., 1992; MacRae, 1994). This short term initial increase, which takes place over a time scale of a few
hours to a day, has been shown to be related to survival (MacRae, 1994; Hollis et al., 1997). Based on
these data, Santore et al. (1999) estimated that an increase in gill copper accumulation of 10 nmol/gw
above background gill copper levels, was associated with 50% mortality in juvenile rainbow trout.
3-64
-------
During the initial stages of development of the biotic ligand model for silver it was assumed that the
uptake of silver by fish gills followed a similar pattern to that of copper, with a rapid short term initial
uptake of silver at the surface membrane of the gill, and that this initial uptake reached a plateau. Janes
and Playle (1995) showed that this short term accumulation of silver at the gill was associated with
adverse effects to fish. They showed that the sodium efflux from the gills of juvenile rainbow trout (1-3
grams) increased as the short term (2- to 3-hour) gill loading of silver increased from about 1 to 15 nmol
Ag/g wet tissue (nmol/gw), when the fish were exposed to silver concentrations spanning the range of
LC50 levels analyzed herein (about 6-40 ug/L) in experiments carried out in low chloride waters.
Several investigators (McGeer and Wood, 1998; Bury et al., 1999a; Wood et al., 1999), in tests with
rainbow trout, reported decreases in Na+ influx to the gill and/or reduced ATPase activity as a function
of predicted free silver, while the relationship of these effects with gill silver levels measured in a
separate set of experiments, following the same experimental protocol, was not clearly defined.
However, the exposure concentrations used in these tests, about 3-4 ug/L, were significantly lower than
the range of concentrations used by Janes and Playle (up to about 50 ug/L of silver) to develop a
relationship between Na efflux from the gill and gill silver burden. The silver concentrations they used
were also about one half of representative rainbow trout LCSOs measured in unamended laboratory
waters (Davies et al., 1978; Bury et al., 1999b). Since the acute toxicity data to be analyzed with the
biotic ligand model includes LC50 results in the range of about 6 to 40 ug/L, the results of Janes and
Playle (1995) are more relevant to consider for purposes of the analyses to be described herein.
Recently reported measurements of the silver loading on juvenile rainbow trout gills over a 48-hour
penod show that the gill loading vanes markedly over time and with the exposure level of silver used
(Wood et al., 1999). The results suggest that an equilibrium gill silver level is not achieved. Hence,
there may not be a unique equilibrium gill silver level associated with a given response. In another long
term 48-hour study with 320 gram rainbow trout, the measured Na+/K+ ATPase inhibition response did
not vary with gill silver as chloride was added (McGeer and Wood, 1998; Wood et al., 1999), but the
same response did vary with gill silver in a shorter 6-hour duration study using 8.8 gram juvenile
rainbow trout (Bury et al., 1999a). Thus, it is still possible that the short term (a few hours to a day)
accumulation of silver on the gills of juvenile rainbow trout can be related to acute effects associated
3-65
-------
with a more extended (e.g., 72 to 96 hours or longer) exposure period, as it has been with copper
(MacRae, 1994; Hollis et al., 1997), Hg (Playle, 1998a) and Cd (Hollis et al., 1997). Although not yet
confirmed by an independent set of tests, the limited results with silver suggest that the short term
accumulation of silver may be directly related to binding at physiologically active sites at the gill
surface, while for longer term exposures, the accumulation of silver occurs at other relatively inert sites
as well, thereby obscuring a direct relationship between longer term gill Ag level and effects.
From a practical perspective, what is actually needed is to be able to predict changes in the toxic
response that results from changes in the complexation of the metal and competitive effects on binding at
physiologically active cells at the site of action, the biotic ligand. The reported difficulties in achieving a
consistent correlation of effects with total gill silver may reflect the relatively minor contribution of
silver at the active gill sites to the total gill silver load. As currently understood it is the chloride cells,
those cells that transport the Ag+ across the gill membrane to a point where it can exert its inhibitory
effect on the Na+/K+ ATPase activity, that correspond to the physiologically active sites on the gill
(Bury et al., 1999a), and thus to the biotic ligand. It has been estimated that these cells represent only 10
percent or less of the total gill cell population (Bury et al., 1999a). Unfortunately, it is not possible to
measure the concentration of silver associated with these cells directly, independent of other silver that is
bound to the gill. Even so, if the total gill silver varies in a similar way to the silver bound to these
functional sites, and/or it can be related to an acute toxicity end point, then a direct measure of the silver
at these specific sites may not be required.
In the interest of expediency, given the preceding difficulties, the approach followed in developing the
BLM is to predict the acute toxicity of metals by use of a parameter that corresponds to a predicted
equilibrium biotic ligand concentration. This approach provides a generalized computational framework
for evaluating the effect on metal toxicity of both the formation of metal complexes and competitive
binding at a biotic ligand that represents the true site of action of toxicity. The degree to which these
effects (e.g., the "hardness correction") differ from one metal or organism to another can be evaluated by
calibrating the model directly to toxicity data. At the same time, as the understanding of the underlying
mechanism of toxicity is elucidated in the future, it is expected that this modeling framework will be
amenable to future refinements.
3-66
-------
Given the difficulties associated with directly relating gill silver accumulation levels to toxicity, the
calibration of the biotic hgand model for fish, as presented herein, is based on both the laboratory silver
accumulation test results of Janes and Playle (1995) and acute toxicity data as well. Development of the
model using toxicity data is critical, since the ultimate use of the model will be to predict toxicity. Since
silver accumulation data that are suitable for use in the development of the BLM for Daphnia magna are
not currently available, when the model is developed for Daphnia, it is calibrated to toxicity data alone.
This may ultimately be the only practical approach for organisms the size of Daphnia, since the site of
action is uncertain and sampling and analysis of the relevant body parts, other than perhaps the whole
body, would be exceedingly difficult if not impossible to do.
As applied herein, the primary use of the BLM is to predict the organism LC50, the dissolved
concentration corresponding to 50 percent mortality, as a function of water quality characteristics. It is
emphasized that the critical biotic ligand accumulation level used in the model, the lethal accumulation
at 50% mortality (LA50) that is associated with the dissolved LC50 to be predicted, is viewed as being a
nominal concentration. It does not necessarily correspond to, nor is it intended to be equivalent to, a
steady state equilibrium concentration that can actually be measured at the physiologically active sites of
the biotic ligand. Rather, the LA50 is simply a quantitative benchmark that can be correlated to the
dissolved silver LC50 in water. This approach is adopted, rather than simply relying on a prediction of
free silver concentration, to expedite the prediction of toxicity with consideration given to alternative
mitigating factors. That is, it provides a way to evaluate not only the concentrations of free metal and
other metal complexes in solution, but the net effect of competitive binding of the metal and other
cations in solution at the physiologically active sites of the biotic ligand as well. The degree to which
mitigating effects such as competitive binding between the metal of interest and other cations vanes with
organism type and metal is addressed in the context of the model calibration process.
3-67
-------
Applicability of the Biotic Ligand Model
The current EPA water quality catena (WQC) for copper (USEPA, 1985) and silver (USEPA, 1980) are
functions of hardness, but are independent of other water quality characteristics. Since it is known that
other water quality characteristics such as pH and organic carbon often affect toxicity as well, it is
frequently necessary to perform extensive bioassay testing to develop water effect ratios (WERs) that can
be used to establish site-specific WQC for metals. The model framework described herein provides a
computational alternative to bioassay testing that can be used to explicitly evaluate these effects.
The biotic ligand model is currently being developed for use in predicting silver LCSOs on the basis of
site-specific water quality characteristics. The relative sensitivities of the organisms to which the model
has been applied thus far are shown on Figure 2 (left panel). This figure shows the cumulative
distribution of Genus Mean Acute Values (GMAVs) upon which the 1987 draft water quality criterion
(WQC) for silver was based (USEPA, 1987). The data from the 1987 draft WQC document are used for
this illustration because they include more data than were available in 1980, when the current acute
criterion was developed (USEPA, 1980). Each data point represents the mean LC50 for a single genus,
and the points are ranked from the most sensitive organisms (low LCSOs) to the least sensitive organisms
(high LCSOs). The BLM analyses to be presented for fish will use data for rainbow trout and fathead
minnows, species that have similar intermediate levels of sensitivity to silver (near the 50 percentile on
the GMAV probability distribution; USEPA, 1980 and 1987). BLM results will also be presented for
Daphnia magna, a relatively sensitive invertebrate having an LC50 that is close to the current acute
water quality criterion for silver, a Criterion Maximum Concentration (CMC) of about 1 ug/L.
\
It should be recognized that the concentration ranges to which the BLM will be applied herein are
significantly greater than typically observed ambient levels of silver in aquatic systems (Figure 2, right
panel). Development of the BLM at these levels is appropriate for several reasons. First, it is envisioned
that the BLM will provide a computational alternative to measuring Water Effect Ratios (WERs) for use
in establishing site-specific criteria. Use of the BLM for prediction of fathead minnow LCSOs is
consistent with the current routine use of fathead minnows in determination of WERs for silver. Further,
EPA guidelines recommend that the WER be developed using an organism that is sensitive at a level
close to the WQC (USEPA, 1994), and development of the model for Daphnia is consistent with this
3-68
-------
recommendation. These published guidelines also stipulate that the endpomt of a toxicity determination
must not be below the Criterion Maximum Concentration (CMC) or Criterion Continuous Concentration
(CCC) to which the WER is to be applied. Since the freshwater acute WQC for silver, the CMC, is about
1 ug/L, development of the BLM for application at lower concentrations would not be appropriate for
development of a site-specific freshwater acute criterion. Although it is envisioned that the applicability
of the biotic ligand model will eventually be extended to lower chronic effect levels, the model is
currently only applicable for use in predicting acute toxicity.
The biotic ligand model framework described herein has thus far been applied to copper (Santore et al.,
1999) and, as presented below, to silver. The principal features that distinguish this model from the
earlier conceptual frameworks that have been proposed to predict metal toxicity are that (1) it is a
working program incorporating a state of the art representation of metal-organic matter complexation,
and (2) it can be used to predict site-specific acute toxicity levels for fish and invertebrates. Studies are
being planned or are in progress to develop the requisite data to evaluate the BLM parameters for other
metals, for both fish and invertebrates. Once these results are available, it is envisioned that the
conceptual framework can be extended to other metals, and in its more general form, to organisms other
than fish.
BIOTIC LIGAND MODEL CALIBRATION
The biotic ligand model for silver will be applied to a variety of data sets for purposes of obtaining a
preliminary calibration. As a starting point, the model was first calibrated using total gill silver
accumulation data generated by Janes and Playle (1995) in the hopes that the resulting parameter values
would to some degree reflect the characteristics of the active sites corresponding to the biotic ligand
itself. The model was then applied to aquatic toxicity data sets to refine the preliminary parameter
values that were determined. The results to be presented will be based on the parameter values that were
determined as a result of following this overall approach.
Analysis of Silver Accumulation Data
»
The first application of the biotic ligand model was to rainbow trout gill silver accumulation data from
3-69
-------
exposure to silver in laboratory water spiked with varying levels of complexmg ligands and competing
cations (Janes and Playle, 1995). This set of data is of particular interest because it demonstrates the
protective effects of both silver complexation (by the addition of DOC and chloride) and of competition
of Ag+ with Ca2* and Na+ for binding sites on the gill. The analysis will also illustrate how the model can
be used to simulate these effects. The model is then applied to gill silver accumulation measurements in
natural waters that were spiked with silver to show that the model can be used to predict similar effects
over a range of natural water characteristics.
Janes and Playle (1995) conducted an elegant set of experiments to demonstrate the effects of
complexation and competition on the accumulation of silver on the gills of juvenile rainbow trout, and to
obtain data to evaluate the required gill site densities and gill binding constants needed in a chemical
equilibrium model that could be used to predict gill silver levels. The first step in the procedure was to
conduct tests where rainbow trout were exposed for two to three hours to 0.07 uM of silver nitrate in the
presence of increasing levels of thiosulfate. The effect of the thiosulfate additions on silver
accumulation on the gills was monitored. The accumulation results are shown on Figure 3. The first bar
shows the background level of gill silver m the control is less than 1 nmol Ag/gram wet weight of gill
(nmol/gw). The second bar shows the increase in gill silver to about 6 nmol/gw when the silver is added
in the absence of thiosulfate, and the remaining bars show the effect on gill silver with increasing
amounts of thiosulfate in the test chamber. Note that at levels in excess of 0.5 uM of thiosulfate, there is
a progressive decrease m gill silver levels. This decrease was interpreted to indicate that the thiosulfate
was forming a complex with the added silver and out-competing the gill for Ag* from binding to the gill.
Since silver thiosulfate (AgS2O3~) is a very strong complex (Log K = 8.8, a value considered to be well
established) this seemed reasonable. Also, since it required about 29 times more thiosulfate than silver
N
to prevent the accumulation of Ag+ on the trout gills, it was reasoned that the Ag-gill conditional
equilibrium constant must be greater than 8.8. Pursuant to several simplifying assumptions it was then
shown algebraically that the Ag-gill equilibrium constant was <10.3. Based on further computations
with MINEQL+, they determined that the Ag-gill Log K = 10.0.
The preceding approach yielded a gill-Ag binding constant that appears to be quite large, at least with
regard to constants typically associated with binding to organic acid functional groups (Sikora and
Stevenson, 1994; Varshal et al., 1995). Alternatively, if the Ag is binding to reduced sulfur groups, then
3-70
-------
it may be reasonable. However, if this were the case, then the Log K for Cu-gill binding (7.4; Playle et
al., 1993b and Santore et al., 1999) would be expected to be higher as well, for a similar reason. To the
degree the gill Ag is associated with the silver thiosulfate complex, this would lead to an overestimation
of the Ag+ that is bound to the gill There is some recent evidence that shows that elevated gill silver
levels can result from exposure offish to Ag-thiosulfate under conditions where acute toxicity to
rainbow trout is not observed and where most of the silver is complexed as Ag-thiosulfate. For example,
in longer term exposures Hogstrand et al. (1996) reported gill silver levels of approximately 30 nmol/gw
at 98 mg/L Ag after 7 days and Wood et al. (1996b) reported 29 nmol/gw at 30 mg/L Ag after 6 days.
Although the total silver concentrations in these silver thiosulfate tests were relatively high, the free Ag
concentration was calculated to be <0.003 ug/L by Wood et al. (1996b), a level that appears inconsistent
with such high gill silver levels. For example, in a parallel set of experiments with 10 ug/L Ag added as
silver nitrate, the gill silver was 11.5 nmol/gw at a calculated free silver concentration that was 3 orders
of magnitude higher, at 3.8 ug/L (Wood et al., 1996a). The inclusion of Ag-thiosulfate m the gill silver
measurements by Janes and Playle (1995) would have significant implications to the analysis because the
Ag-gill constant evaluated from the thiosulfate experiments was subsequently used in the evaluation of
the remaining conditional equilibrium constants, including a Ag-DOC equilibrium constant of Log K =
9.0 - 9.2.
Given the possibility that the Ag-thiosulfate was binding to the gills in these experiments, the BLM was
calibrated without use of the thiosulfate data. The calibration approach used herein also differed from
the approach followed by Janes and Playle (1995) because the representation of metal-organic matter
interactions developed for WHAM was applied. As a preliminary step, the model was applied to Ag-
DOC sorption data obtained with hurmc and frilvic acids (Sikora and Stevenson, 1988). The initial
metal-proton exchange constants evaluated from this analysis were about 2.6 for humic samples and 2.2
for fulvic samples (Figure 4 shows representative results), with a range of values associated with the
different samples analyzed. These values were then subsequently refined over the course of the BLM
calibration, with the final values set at 2.0 and 1.4 for humic and fulvic acids respectively. Table 1
summarizes the base case model input parameters that resulted from the model calibration. Departures
from these values will be noted as appropriate in the discussion of the calibration results that follow.
Given the high concentrations of silver used in the experiments by Sikora and Stevenson (1988), and
recognizing that measured proton exchange constants are typically inversely related to the level of metal
3-71
-------
tested, these initial estimates were considered to set a lower bound on the metal-proton exchange
constants to use for the silver accumulation test conditions. This effect occurs because of variations in
the strength of the organic matter binding sites. At low silver levels, binding occurs only at the strong
sites, while as the silver concentration increases, more and more weak sites form complexes with the
silver and the effective equilibrium constant decreases. The analysis of the data indicated that when
these initial estimates of proton exchange constants were used, the mitigating effect of DOC on toxicity
was underestimated (as expected), and the values were adjusted to 2.0 and 1.4 for Ag-humic and Ag-
fulvic proton exchange constants, respectively. (Note that as these values are expressed here, this
adjustment increases the degree of DOC complexation.)
The calibration of the biotic ligand model to the rainbow trout gill accumulation data (Janes and Playle,
1995) is summarized on Figure 5. These graphs compare measured juvenile rainbow trout gill Ag levels
following 2-3 hour exposures to silver (the open bar, the control and the single hatched bars for silver
treatments) with the gill-Ag levels computed by the BLM (solid bar = Ag+ binding to the gill and the
cross hatched section indicates AgCl binding to the gill). The water chemistry used in these
computations is reported elsewhere (Janes and Playle, 1995) and the constants that were developed from
the calibration are listed in Table 1 and discussed throughout the remainder of this description of the
calibration of the BLM. The upper left panel shows gill silver accumulation levels as a function of the
total dissolved Ag that was added. A gill-Ag Log K of 7.3 was used in conjunction with a gill site
density of 35 nmol/gw to achieve these results. Generally good agreement is obtained, with both
measured and predicted gill silver levels increasing as the silver concentration increases from less than
the detection limit (the control) to about 54 ug/L (0.5 uM). The data suggest a leveling off of the gill
silver concentration at the three highest exposure concentrations, suggesting that the available gill sites
are saturated at about 15 nmol/gw. The slight over prediction of the gill silver concentration that is
evident at the higher exposure levels occurs because a gill site density of 35 nmol/gw is assigned.
Although use of a lower gill site density would provide a better fit of these data (Janes and Playle, 1995,
used an average of 13 nmol/gw), an effort was made to maintain consistency across all data sets analyzed,
and a higher gill site density is needed to predict the observed gill silver levels in the natural water
experiments, reviewed subsequently.
The upper nght panel of Figure 5 shows gill silver results obtained with increasing amounts of sodium
3-72
-------
added to water containing about 11.9 ug/L of silver. This set of data illustrates how a competing cation
can reduce the level of silver on the gill. The first two pairs of bars compare model and data in the
control water, with and without silver addition. With the addition of silver, the gill silver level increases
to about 10 nmol/gw and model and data (measured) are in good agreement. Increasing the sodium
concentration by about a factor of 4, from 400 to 1600 uM (about 9 to 37 mg/L) has a negligible effect
on the results. However, when the sodium concentration is increased by another factor of 10, to 16,000
uM (370 mg/L), gill silver levels decrease to about 4 nmol/gw. The model, with the gill-Na binding
constant set at Log K = 2.3 responds accordingly. The utility of this data set is that it provides a way to
evaluate the gill-Na Log K by defining the concentration range over which this change in gill silver level
occurs. It should be understood, however, that the values set for these gill binding constants are to some
degree dependent on each other, so the overall calibration process is somewhat less straightforward than
might at first appear.
Another example of a competing cation that was tested is calcium. Janes and Playle (1995) reported that
gill silver levels remained in the range of 5 to 7 nmol/gw at calcium levels as high as 10.6 mM (about 265
mg/L). These results indicate that calcium does not compete well with Ag for binding sites on the gill.
As calibrated (gill-Ca Log K = 2.3), BLM predictions for gill silver are within this range up to about
2000 uM, and then decrease slightly to about 3.5 nmol/gw at 10.6 mM of calcium.
The lower left panel of Figure 5 presents gill Ag results versus chloride concentration. Until recently, it
was understood that chloride mitigates the toxicity of silver by forming the AgCl complex, thereby
reducing the free silver concentration that was considered to be the bioavailable species. Although AgCl
was previously thought to be non-toxic, recent data, to be reviewed when the toxicity test results are
discussed, suggests that it may actually bind to the gills offish and be toxic to some species (Enckson et
al., 1998; Bury et al., 1999b). Accumulation of gill-AgCl, in addition to gill-Ag+, is thus included in the
BLM. The model results indicate these components separately as the solid (gill-Ag) and cross-hatched
regions (gill-AgCl) of the bars. The model and data both show excellent agreement up to 1500 uM of
chloride, with total gill silver levels of about 10 nmol/gw. When the chloride is increased an additional 6-
fold, however, the observed and predicted total gill Ag decreases to about 4 to 5 nmol/gw. This decrease
occurs because most of the silver in solution is now AgCl (the free silver is reduced significantly at this
level of chloride), and the gill-AgCl Log K of 6.7 is less than the gill-Ag Log K of 7.3, so less Ag binds
3-73
-------
to the gill. It is also of interest to note that the predicted contribution of AgCl to total gill-Ag increases
with increasing chloride levels, reflecting the shift in the distribution of silver species in the water.
One of the primary difficulties encountered in the analysis of the gill accumulation data in the laboratory
water experiments was that the water had a relatively high DOC concentration, reported to be m the
range of 1.7 to 2.4 mg/L. This was significantly higher than the 0.3 mg/L of DOC in the laboratory RO
water prior to its use in these experiments. The source of this increase in DOC was attributed to mucosal
secretions and excreta from the fish (Playle, 1998b), with the effect possibly enhanced by the relatively
high fish biomass/volume (6-18 grams offish/liter) used in these short term static tests. When the BLM
was initially applied to this data set, the gill silver was consistently over-predicted. Decreasing the Ag-
gill equilibrium constant was not a viable solution since it was desirable to maintain consistency in the
parameter values used in the model for all data sets, and this change was inconsistent with the
comparisons to the other data sets, particularly the toxicity data sets. Since this increase in DOC was
apparently limited to the gill accumulation experiments, perhaps because fish biomass /volume used in
the toxicity experiments was much lower (about 1 to 2 g/L), it was decided that adjusting the
characteristics of DOC from this source was reasonable for calibration purposes. Additionally, although
information on the complexation capacity of this type of DOC was not available, it would be reasonable
to expect that it could have characteristics that differed from those of humic and fulvic acids of a
terrestrial origin. In light of these considerations, the site density of this incremental DOC from the fish
was assigned to be approximately a factor of 4 times the site density of humic acid in order to fit these
gill accumulation data. It was also assumed that the source water background DOC was 10% humic acid
and 90% fulvic acid for this data set and for the other data sets to be analyzed as well. These adjustments
were included in the BLM results discussed previously for Figure 5 (i.e., the Ag, Na and Cl addition
tests).
Janes and Playle (1995) also conducted studies where the DOC of the water was increased from the
control water DOC of 2.4 mg/L, to about 24 mg/L at the highest treatment level (Figure 5, lower nght
panel). The silver concentration was 0.17 uM (18.3 ug/L) in these tests. The data and model results are
compared on the lower nght panel of Figure 5. The results provide a clear illustration of how formation
of a silver-organic matter complex reduces Ag accumulation at the gill by reducing the free silver
concentration in solution.
3-74
-------
The DOC that was added in these experiments was isolated from Luther Marsh, Ontario, by tangential
flow ultrafiltration. No other characteristics of this source of organic matter were reported. When the
model was applied using the standard set of humic and fulvic acid characteristics developed for WHAM
for the added DOC, gill silver was overestimated. It was therefore necessary to increase the binding site
density of the Luther Marsh DOC by about a factor of seven to obtain the indicated results for the DOC
experiments. The need to increase the Luther Marsh DOC site density is consistent with results of
analyses with mercury and with mixtures of metals reported by Playle (1998a) where it was necessary to
increase the binding site density of Luther Marsh DOC by as much as a factor often. It would be
preferable if the DOC characteristics did not need to be adjusted for these different sets of experiments,
and these changes were made with due consideration given to this. However, they were necessitated by
changes in the experimental conditions that were unique to these tests. Fortunately, when the model is
applied to the toxicity data sets, the conditions were such that the contribution of DOC from the test
organisms did not seem to be a significant factor (static renewal test procedures and/or a much lower
mass offish per unit volume of water were used). Also, the DOC that was added, Aldnch humic acid, is
commonly used and relatively well characterized, so a consistent set of DOC related inputs could be
assigned in these critical analyses.
Finally, with regard to pH, Janes and Playle (1995) reported that pH was not a significant factor over the
range of pH levels tested. Although the detailed data for the pH experiments were not presented, the
measured gill-Ag was reported to be about 6 to 6.5 nmol/gw over a pH range of 4.5 to 6.5 and a silver
concentration of 0.06 uM (6.5 ug/L) . The gill proton binding constant used in the analyses of Figure 5 is
Log K = 4.3. The modeling analyses of the datasets to be presented herein (Figures 5 -10) were not very
sensitive to this parameter over a range of Log K values of 4.3 to 5.4, with the higher value currently
used in the copper BLM model (Santore et al., 1999). The msensitivity of the results occurs because
most tests were conducted at pH > 6.2. A gill proton binding constant of Log K = 4.3 is used here
because it is more consistent with the low sensitivity of the reported results at pH levels as low as 4.5.
Note that with the silver DOM model used in the BLM, pH changes will also affect the results through
proton-DOC interactions. As pH increases, DOC binding sites de-protonate and this results in an
increase in Ag:DOC binding. Over the pH range of 6.5 to 8.5, this appears to result in a decrease in
calculated free silver that is qualitatively inconsistent with the limited data considered to date (Sikora
3-75
-------
and Stevenson, 1988) As a result, for purposes of the current version of the model, the inputs are
specified to avoid this effect at pH levels above 6.5. Although this is an area where future refinements to
the model are envisioned, these refinements must await the completion of a carefully designed Ag:DOM
adsorption experiment for use in this task.
The BLM has also been applied in the analysis of a similar set of gill silver accumulation experiments
conducted by Janes and Playle (1995) using a variety of natural water samples. Here, the DOC of the
laboratory control waters was characterized as before, but the natural waters were modeled using
standard WHAM-based BLM inputs for DOC. That is, the percent of the DOC that is characterized as
being humic (10%) and fulvic (90%) material is specified. The remaining Log K values are as
summarized previously in Table 1. The measured gill silver accumulation and predicted values are
compared on Figure 6. Exposure waters 1 and 2 are the lab water controls, with and without silver
added, and the remaining exposure waters (3 - 8) are from surface waters in southern Ontario. The water
quality characteristics in these waters exhibited significant variability: DOC = 4.4 - 8.5 mg/L, pH = 7.4 -
8.2, Ca = 500-2100 uM, Na = 46 - 1045 uM and Cl = 32 - 1150 uM. The measurements, the open
(control) and filled (silver treatments) are shown with the 95% confidence interval. The model results
are indicated as ranges corresponding to the predicted gill Ag with and without gill-AgCl included. The
model performs quite well in simulating the range of variations that were measured using these test
waters. This is important, since the ultimate use of the model will be to predict acute effect levels for
silver in natural waters.
It is worth reiterating at this point m the discussion that the total gill silver concentrations presented
above and predicted with the BLM should not be construed as representing the silver at physiologically
active sites that control the lono-regulatory response of the organism. The purpose for the analysis of
\
these data was to obtain an indication of the magnitude of the biotic ligand model input parameters, the
gill binding site density and the gill binding equilibrium constants, that are associated with protective
effects. Since the short term gill silver uptake studies reviewed thus far were not carried out for purposes
of eliciting an acutely toxic response and none was measured, the data cannot be related to effects. The
purpose of the next set of analyses of acute toxicity studies will be to make this connection.
Analysis of Acute Toxicity Data
3-76
-------
The short term accumulation studies of Janes and Playle (1995) were never directly related to an acute
toxicity end point. Further, only a fraction of the measured gill silver is binding to physiologically active
sites, and this accumulation level appears to vary significantly over time and with the exposure regime
(Wood et al., 1999). Thus, the requisite data are not available at this time to empirically relate a
particular accumulation level to an acutely toxic effect. The biotic hgand model will therefore be applied
to the analysis of LC50 data from bioassays to establish this connection between the predicted level of
silver accumulation at the biotic hgand and toxicity (i.e., the LC50). The practical test of the utility of
the model will then be to determine if it can be used to predict toxicity levels of silver over a range of
water quality conditions.
The remaining data sets are analyzed to define the relationship between a predicted biotic hgand
accumulation level and a toxic effect concentration for a range of water quality characteristics. Since gill
silver levels were not measured in the toxicity tests, the data were analyzed with the underlying
assumption that a single biotic hgand accumulation level of silver is associated with 50 percent mortality
for a particular age or species offish, or organism type. The analysis was originally conducted in
parallel with the analysis of the accumulation data, so the model parameter values presented previously
are employed here as well. The model will be used to explain how the variation in each of these water
quality parameters affects the bioavailability and hence the toxicity of silver.
The first set of data to be analyzed was a comparative study of the acute toxicity of silver to juvenile
rainbow trout and fathead minnows (Bury et al., 1999b). As discussed previously, these fish species are
similar in their sensitivity to silver. Figure 7 presents what are considered to be interesting and
previously unexpected results for the effect of chloride. The upper panel presents measured LC50
\
concentrations for fathead minnows (open bars) and rainbow trout (filled bars) as a function of
increasing chloride concentration (50, 250, 800 and 1500 uM). The LC50 for rainbow trout increases
from about 7 to 25 ug/L over this range in chloride levels, but, surprisingly when first reviewed, the
fathead minnow data are nearly constant, with only a marginal increase in LC50 with chloride
concentration. The increase in rainbow trout LC50 conforms with the notion that bioavailability is
reduced at increasing chloride levels, due to the formation of AgCl and the decrease in free silver. The
fathead minnow results, on the other hand, suggest that AgCl is in fact toxic to this fish species. Enckson
et al. (1998) reported similar results for fathead minnows. Although the data were somewhat limited, a
3-77
-------
slight increase in toxicity was actually observed with increasing chloride concentration.
The middle panel of Figure 7 shows the predicted free ion concentration at each of the LCSOs. For
rainbow trout, the free ion concentration is approximately constant, but for fathead minnows, since the
LC50 silver concentration is approximately constant, the free ion concentration decreases with increasing
chloride. Hence, these data suggest that for fathead minnows at least, free ion concentration is not a
good uniformly applicable predictor of toxicity.
On the bottom panel of Figure 7, the predicted biotic ligand concentration that exists in association with
the LC50 is shown. For rainbow trout, the gill-Ag is relatively uniform across the test conditions, at
about 12 to 18 nmol/gw, with a slightly downward trend with increasing chloride concentration. For the
fathead minnow, the bar indicates the two components of Ag on the gill, the lower portion corresponding
to bound Ag+ and the upper portion to bound AgCl. Since AgCl appears toxic to fathead minnows, and it
has been reported to bind to the gill (e.g., Wood et al., 1999), it is assumed that it contributes to toxicity
as well. Considered in this way, the total gill silver on fathead minnows follows a similar pattern to that
of rainbow trout (without AgCl included in the gill loading, although it is considered to bind to the gill as
well), with both exhibiting relative uniformity across the range of treatments (about 15 +/- 3 nmol/gw) in
comparison to the water LCSOs shown on the upper panel. Note that this level of silver at the gill that is
associated with the water LC50 will ultimately be used to set the model parameter that will be used to
predict toxic effect levels.
Figure 8 shows analogous results from this same set of experiments, but in this case the water quality
variable is DOC, added as Aldrich hurmc acid. Here, the LC50 increases for both the fathead minnows
and rainbow trout, as would be expected. That is, since the added DOC forms an organic complex with
the silver, thereby reducing the free silver, more silver must be added to exert the same level of effect
(50% mortality). The resulting calculated free silver concentrations at the silver LC50 (middle panel) are
somewhat variable, but considerably more consistent than the LC50 concentrations. Similarly, the
calculated gill silver concentrations are relatively consistent for both fish species and across all DOC
treatments, at about 15 to 20 nmol/gw.
The final set of experiments by Bury et al. (1999b) considered the effect of calcium (i.e., a cation that
3-78
-------
contributes to hardness) on toxicity of silver (Figure 9). Here, the LC50 concentrations increased
slightly as calcium was increased from 50 to 2000 uM, but not as much as was observed with the
addition of DOC, or Cl as well in the case of rainbow trout. The calculated free silver follows a similar
pattern, increasing with increasing calcium, since the calcium does not affect silver speciation and the
calculated free silver is based on specification of the silver LC50 as the total silver concentration in the
water Finally, the calculated gill silver level (lower panel) is about 17 to 18 nmol/gw across all of the
calcium treatment levels. This accumulation level is also relatively consistent with the gill silver levels
presented on Figures 7 and 8 as well.
A second independent set of experiments is available where the effect of DOC (added as Aldnch humic
acid) and chloride additions on the toxicity of silver to 28 day old fathead minnows was measured. The
results are shown on Figure 10. The upper panel shows the three concentrations of DOC (nominally, 0, 5
and 10 mg/L) in each of 4 chloride treatments (3, 20, 40 and 60 mg/L or about 100 to 2100 uM, similar
to the range tested by Bury et al., 1999b). The measured LC50 concentrations are shown on the middle
panel. In this case, within each chloride treatment, there was a slight increase in the silver LC50 with
increasing DOC, though less than in the previous set of experiments. In some cases the increase is
masked by the fact that the measured DOC levels were actually quite similar (e.g., the two highest DOC
treatments at 40 mg/L of chloride were approximately 9 and 10 mg/L, and the silver LCSOs were about
equal). Overall, the LC50 concentrations tended to be higher than in the previous set of experiments,
ranging from about 20 to 35 ug/L. This trend is qualitatively consistent with the fact that the reported
unamended DOC concentrations in the studies by Bury et al (1999b) were lower than the levels in these
studies (about 0.3 mg/L, versus about 1 to 3 mg/L). As previously observed for fathead minnows,
however, the LC50 did not increase markedly as the chloride concentration was increased. Finally, the
predicted gill silver levels, when bound AgCl is included, are approximately uniform across all
treatments, at about 15 to 18 nmol/gw. The low DOC controls tend to be slightly higher within each
chlonde treatment group, resulting in a slight decrease in gill silver with increasing DOC. This suggests
that the model may be overstating the mitigating effect of DOC for this set of experiments. However,
considering all of the data sets analyzed collectively, this unaccounted for variation is acceptable.
USE OF BIOTIC LIGAND MODEL TO PREDICT THE ACUTE TOXICITY OF SILVER
3-79
-------
At this point a set of model parameters has been evaluated that performs a reasonable job of predicting
the accumulation of silver on fish gills over a range of water quality characteristics (Figures 5 and 6)
The model was then used to evaluate the short term accumulation of gill silver that is associated with the
96-hour silver LC50 concentration for fathead minnows and rainbow trout, again, over a range of water
quality characteristics (Figures 7 - 10). Based on these results, the predicted gill silver concentration
associated with the silver LC50 in water was relatively consistent, typically in the range of 15 to 20
nmol/gw. It remains now to make use of these results to predict LCSOs.
The approach for predicting LCSOs is as follows. A critical total gill silver lethal accumulation level at
50 percent mortality, an LA50 of 17 nmol/gw_ is assumed to be associated with the dissolved silver LC50
concentration in water. This gill silver concentration includes the sum of gill-Ag and gill-AgCl for
fathead minnows and only gill-Ag for rainbow trout. A "numerical titration" is next performed with the
BLM. That is, the water chemistry for the sample of interest is specified as an input to the BLM and
silver is incrementally added to this water. At each step, the gill Ag concentration is predicted. The
dissolved silver concentration corresponding to the point where the gill silver reaches the critical gill
concentration (LA50 =17 nmol/gw) is the predicted LC50. A more detailed description of this procedure
and an example of how it is implemented for copper is available in the companion papers (Di Toro et al.,
1999;Santoreetal, 1999).
The preceding approach has been applied to the LC50 data for fathead minnows and rainbow trout that
were analyzed herein. The results of the predicted LCSOs are summarized on Figure 11, which shows the
predicted LC50 versus the measured LC50 for silver. The diagonal line is the line of perfect agreement
(predicted LC50 = observed LC50) and the dashed line corresponds to plus or minus a factor of two
\
times the measured LC50. Fathead minnow data are indicated with filled squares and rainbow trout data
with open squares. As shown, the predictions are typically within a factor of 1.5 of the measured LCSOs,
and more generally within a factor of two of the measured values, over a range of LCSOs of about 6 to 35
ug/L. Although some refinement to the BLM calibration could be made, this agreement was considered
quite good, given that LC50 test results will very often vary by a factor of two. Thus, some of the
deviations from the line of perfect agreement may be attributed to the LC50 measurements themselves.
As discussed previously, it would be useful to develop the BLM for an organism that is sensitive near the
3-80
-------
acute water quality criteria for silver. This has been done for Daphma magna using a data set similar to
the fathead minnow data set of Figure 10 (Bills et al., 1997). The DOC, chloride and calcium levels were
varied in this experiment, with the primary factor that mitigated toxicity being the addition of DOC
(Aldnch humic acid). Briefly, the same model input parameters were used to predict the biotic ligand
silver accumulation levels at the LC50 concentration. A critical biotic ligand silver concentration was
then determined based on these results (LA50 = 2.3 nmol/gw). The numerical titration was then
performed with this value specified as an input to the BLM and the LCSOs were predicted. The results of
this analysis are presented in conjunction with the preceding results for fish (Figure 11) on Figure 12.
The Daphma magna results are indicated by the triangle plot symbol. The measured LCSOs ranged over
approximately one order of magnitude, from approximately 0.5 to 5 ug/L. As with the results for fish,
the BLM predictions of LCSOs are typically within a factor of 1.5 of the line of perfect agreement, and
consistently within a factor of two (within the dashed lines).
SUMMARY
In summary, the biotic ligand model for silver seems to work quite well. The calibrated model generally
reproduces the observed gill silver levels over a range of laboratory test conditions, and is also capable of
tracking gill silver levels in tests with natural waters having a significant range in water quality
characteristics as well. As parameterized in these calculations the Ag-Cl component of the predicted
total gill silver loading (Ag-gill + AgCl-gill) was relatively small, but it did improve agreement with the
measured gill silver data for rainbow trout. The model yielded relatively uniform 3-hour gill silver levels
of about 15 to 20 nmol/gw over a range of LCSOs, where AgCl was assumed to be toxic to fathead
minnows and relatively benign to rainbow trout over the range of DOC, chloride and hardness levels
tested. These results indicate that the proposed model framework can explicitly account for variation in
silver toxicity to both fish and daphnia that result not only from changes in hardness, but from site-
specific variations in DOC and chloride as well.
The results presented herein are considered to provide a demonstration of the viability of this general
approach for using the biotic ligand model. To date, the predictions for LCSOs have only been made for
the bioassay data analyzed herein. The model will need to be verified by application to one or more
independent sets of data as they become available. There may be situations where additional organic or
3-81
-------
inorganic hgands that form complexes with silver (e.g., sulfide) will be present at high enough levels to
be significant. In such instances a conservative result will be obtained (i.e., the BLM predicted LC50
will be lower than a corresponding measured result).
The analysis illustrates how consideration of metal speciation and bioavailabihty provides an improved
understanding of how environmental exposure levels of metals are related to acute effects. The modeling
framework that was described has a number of potentially useful applications, including: (1) setting
discharge permit limits, (2) development of site specific WQC via a modified Water Effects Ratio
(WER) procedure, (3) evaluation of fate and toxicity of metals for use in ecological nsk assessments, and
(4) development of updated water quality catena for metals where, in addition to hardness, the refined
WQC might also be a function of TOC, DOC, pH and other variables which affect the speciation,
complexation and toxicity of metals m aquatic systems.
Recently reported results (McGeer and Wood, 1998; Bury et al., 1999a & b; Wood et al., 1999) indicate
that the toxicity of silver to fish may be more closely related to the free ion concentration of silver than
to gill silver levels. Competitive ion effects seem to be relatively unimportant. Although considered
preliminary at this time, this finding is consistent with the results presented herein that the effect of
calcium on silver toxicity to fish is relatively minor in comparison to its effect on the toxicity of other
metals, such as copper. Even so, the continued use of the BLM framework for silver is warranted for
several reasons. First, the BLM framework includes a detailed speciation evaluation, and hence the
requisite exposure assessment based on free silver can still be addressed in the context of the BLM
modeling framework. Second, it remains to be determined if the competitive ion effects that are
included in the biotic ligand model framework for silver are required for other organisms, even if not
required for fish. Finally, the BLM framework provides a computational framework for assessing the
bioavailabihty and toxicity of metals in general. It therefore makes sense to include silver within what is
considered to be a unified framework for use in making these evaluations.
Acknowledgment- This work was completed with the financial support of the Eastman Kodak Company
and the Silver Council.
3-82
-------
REFERENCES
Allen, H.E. and D.L. Hansen, 1996. "The Importance of Trace Metal Speciation to Water Quality
Criteria," Water Environment Research, 68:42-54.
Allison, J.D., D.S Brown, K.J. Novo-Gradac, March 1991. "MINTEQA2/PRODEFA2, A Geochemical
Assessment Model for Environmental Systems Version 3.0, Users Manual," EPA/600/3-91/021, USEPA
ERL ORD, Athens, GA.
Ankley, G.T., W.J. Berry, L.D. De Rosa, D.M. Di Toro, DJ. Hansen and M. Reiley, December 1994.
"Briefing Report to the EPA Science Advisory Board on the Equilibrium Partitioning Approach to
Predicting Metal Bioavailability in Sediments and the Derivation of Sediment Quality Criteria for
Metals," USEPA, Office of Water and Office of Research and Development.
Ankley, G.T., D.M. Di Toro, DJ. Hansen, WJ. Berry, December, 1996. "Technical Basis and Proposal
for Denvmg Sediment Quality Criteria for Metals, Environmental Toxicology and Chemistry, 15:2056-
2066.
Bergman, H. L. and E. J. Dorward-Kmg, 1997. Reassessment of Metals Criteria for Aquatic Life
Protection. Pensacola, FL, SETAC Press.
Bills, T., SJ. Klame, T.W. LaPoint, G. Cobb, B. Forsythe and M. Wenholz, June 1997. "Influence of
Water Quality Parameters on Acute Silver Nitrate Toxicity to Daphnia magna," Final report No.
TTWET-9506.1, submitted to the Silver Council c/o the National Association of Photographic
Manufacturers by the Department of Environmental Toxicology, The Institute of Wildlife and
Environmental Toxicology (TTWET), Clemson University, Pendleton, SC.
Brown, David S. and Jerry D. Allison, February, 1987, "MINTEQA1 Equilibrium Metal Speciation
Model: A User's Manual," prepared for the USEPA Environmental Research Laboratory, Office of
Research and Development, Athens, Georgia.
3-83
-------
Bury, N.R., J.C. McGeer and C.M. Wood, 1999a. "Effects of Altering Freshwater Chemistry on
Physiological Responses of Rainbow Trout to Silver Exposure," Environmental Toxicology and
Chemistry, Vol. 18, No. 1, pp. 49-55.
Bury, N.R., F Galvez and C.M. Wood, 1999b. "Effects of Chloride, Calcium and Dissolved Organic
Carbon on Silver Toxicity: Comparison Between Rainbow Trout and Fathead Minnows," Environmental
Toxicology and Chemistry, Vol. 18, No. 1, pp. 56-62.
Campbell, P.G.C., 1995. "Interactions Between Trace Metals and Aquatic Organisms: A Critique of the
Free-ion Activity Model," Metal Speciation and Bioavailability in Aquatic Systems, A. Tessier and D.R.
Turner, eds., IUPAC, John Wiley and Sons.
Davies, P.H., J.P. Goettl, Jr., J.R. Smley, 1978. "Toxicity of Silver to Rainbow Trout (Salmo gairdnen)."
Water Research. Vol. 12, pp. 113-117.
Di Toro, D.M., C.S. Zarba, D.J. Hansen, W.J. Berry, R.C. Schwartz, C.E. Cowan, S.P. Pavlou, H.E.
Allen, N.A. Thomas and P.R. Paquin, 1991. "Technical Basis for Establishing Sediment Quality Criteria
for Nomonic Organic Chemicals by Using Equilibrium Partitioning," Environmental Toxicology and
Chemistry, Volume 10, No. 12, pp. 1541-1583.
Di Toro, D.M., H.A. Allen, H.L. Bergman, J.S. Meyer, P.R. Paquin and R.S. Santore, 1999. "A Biotic
Ligand Model of the Acute Toxicity of Metals. I. Technical Basis," manuscript in preparation.
Enckson, R.J., L. Brooke, M.D. Kahl, F.V. Venter, S.L. Hartmg, T.P. Markee and R.L. Spehar, April
1998. "Effects of Laboratory Test Conditions on the Toxicity of Silver to Aquatic Organisms,"
Environmental Toxicology and Chemistry, Vol. 17, No. 4, pp. 572-578.
Hogstrand, C., F. Galvez and C.M. Wood, 1996. "Toxicity, Silver Accumulation and Metallothionein
Induction in Freshwater Rainbow Trout During Exposure to Different Silver Salts," in Environmental
Toxicology and Chemistry, 15:1102-1108.
3-84
-------
Holhs, L , L. Muench and R.C. Playle, 1997. "Influence of Dissolved Organic Matter on Copper
Binding, and Calcium on Cadmium Binding, by Gills of Rainbow Trout," Journal of Fish Biology,
50:703-720.
Janes, N., and R.C. Playle, 1995. "Modeling Silver Binding to Gills of Rainbow Trout (Oncorhynchus
mykiss}, Environmental Toxicology and Chemistry, Vol. 14, No. 11, 1847-1858.
Kramer, J.R, H.E. Allen, W. Davison, K.L. Godtfredsen, J.S. Meyer, E.M. Perdue, E. Tipping, D. van de
Meent and J.C. Westall, 1997. "Chemical Speciation and Metal Toxiciry in Surface Freshwaters," in
Reassessment of Metals Criteria for Aquatic Life Protection: Priorities for Research and
Implementation, Bergman, H.L., and E.J. Dorward-King, eds., Proceedings of the Pellston Workshop on
Reassessment of Metals Criteria for Aquatic Life Protection, February 10-14, 1996, Pensacola, Florida,
SETAC Press.
MacRae, R.K., December, 1994. "The Copper Binding Affinity of Rainbow Trout (Oncorhynchus
mykiss) and Brook Trout (Salvelinus fontinahs) Gills," a thesis submitted to the Department of Zoology
and Physiology and The Graduate School of the University of Wyoming in partial fulfillment of the
requirements for the degree of Master of Science in Zoology and Physiology.
McDonald, D.G., J.P. Reader and T.R.K. Dalziel, 1989. "The Combined Effects of pH and Trace Metals
on Fish lonoregulation," in: Acid Toxicity and Aquatic Animals, Edited by R. Moms, E.W. Taylor, D.J.A
Brown and J.A. Brown, Soc. Exp. Biol. Seminar Series, Vol. 34, Cambridge University Press,
Cambridge, U.K., pp. 221-242.
\
McGeer, J: C. and C. M. Wood, August 4,1998. "Protective Effects of Water Cl on Physiological
Responses to Waterbome Silver in Rainbow Trout," final manuscript accepted by Canadian Journal of
Fisheries and Aquatic Sciences.
Morel, P.M., 1983. "Complexation: Trace Metals and Microorganisms," in Chapter 6 of Principles of
Aquatic Chemistry, Wiley Interscience, New York, pp. 301-308.
3-85
-------
Morgan, I.J., R.P. Henry, CM. Wood, August 1996. "The Mechanism of Acute Silver Nitrate Toxicity
in Freshwater Rainbow Trout (Oncorhynchus mykiss) is Inhibition of Gill Na+ and Cl" Transport." (Draft
of manuscript published in Aquatic Toxicology)
Pagenkopf, O.K., 1983. "Gill Surface Interaction Model for Trace-metal Toxicity to Fishes: Role of
Complexation, pH, and Water Hardness," Environmental Science and Technology, 17:342-347.
Playle, R.C., R.W. Gensemer and D.G. Dixon, 1992. "Copper Accumulation on Gills of Fathead
Minnows: Influence of Water Hardness, Complexation and pH on the Gill Micro-environment,"
Environmental Toxicology and Chemistry, 11:381-391.
Playle, R C., D.G. Dixon and K. Burnison, 1993a. "Copper and Cadmium Binding to Fish Gills:
Modification by Dissolved Organic Carbon and Synthetic Ligands," Can. J. Fish. Aquat. Sci., 50: 2667-
2677.
Playle, R.C., D.G. Dixon and K. Bumison, 1993b. "Copper and Cadmium Binding to Fish Gills:
Estimates of Metal-Gill Stability Constants and Modeling of Metal Accumulation," Can. J. Fish. Aquat.
Sci., 50: 2678-2687.
Playle, R.C., 1998a. "Modelling Metal Interactions at Fish Gills," The Science of the Total Environment,
219, pp. 147-163.
Playle, R., 1998b. Personal communication to Paul Paquin.
Prothro, M.G., 1993. "Office of Water Policy and Technical Guidance on Interpretation and
Implementation of Aquatic Life Metals Criteria," Memorandum from USEPA Acting Assistant
Administrator for Water to Water Management Division Directors and Environmental Services Division
Directors, Regions I-X.
Renner, R., 1997. "Rethinking Water Quality Standards for Metals Toxicity," Environmental Science
and Technology, 31:466-468.
3-86
-------
Santore, R.C and C T. Dnscoll, 1995 "The CHESS Model for Calculating Chemical Equilibria in Soils
and Solutions," Chemical Equilibrium and Reaction Models, SSSA Special Publication 42, The Soil
Society of America, American Society of Agronomy.
Santore, R.S., D.M. Di Toro, P.R. Paqum, 1999. "A Biotic Ligand Model of the Acute Toxicity of
Metals. II. Application to Fish and Daphnia Exposure to Copper," manuscript in preparation.
Sikora, FJ. and F.J. Stevenson, 1988. "Silver Complexation by Humic Substances: Complexation
Constants and Nature of Reactive Sites," Geoderma, 42:353-363.
Tipping, E., 1994. "WHAM-A Chemical Equilibrium Model and Computer Code for Waters,
Sediments, and Soils Incorporating a Discrete Site/Electrostatic Model of Ion-Binding by Humic
Substances," Computers and Geosciences, Volume 20, No. 6, pp. 973-1023.
Tipping, E., 1996. Modeling the Interactions of Metals with Dissolved Organic Matter in Surface
Waters," In Press.
USEPA, October 1980. "Ambient Water Quality Criteria for Silver," Office of Water Regulations and
Standards, Catena and Standards Division, Washington, DC. EPA 440/5-80-071; PB81-117822.
USEPA, January 1985. "Ambient Water Quality Criteria for Copper -1984," Office of Water
Regulations and Standards, Criteria and Standards Division, Washington, DC.
\
USEPA, September 24, 1987. "Ambient Aquatic Life Water Quality Criteria for Silver." - Draft.
USEPA, August 1994. "Water Quality Standards Handbook: Second Edition, Appendix L, Interim
Guidance on Determination and Use of Water Effect Ratios for Metals," USEPA, Office of Water, EPA-
823-B-94-005a.
Varshal, G.M., T.K. Velyukhanova, N.N. Baranova, I.Y. Koshcheyeva, T.V. Shumskaya, Y.V. Khohn,
1995 "Geochemical Significance of Complexing of Silver(I) with Humus Acids," Geochemistry
3-87
-------
International, 32:1-9.
Westell, J C., J.L. Zachary and F. Morel, 1976 MINEQL, A Computer Program for the Calculation of
Chemical Equilibrium Composition of Aqueous Systems. Technical Note 18. Department of Civil Engr.
MIT. Cambridge MA.
Wood, C.M., C. Hogstrand, F. Galvez, and R.S. Munger, 1996a. "The Physiology of Waterborne Silver
Toxicity in Freshwater Rainbow Trout (Oncorhynchus mykiss) 1. The Effects of Ionic AgV Aquatic
Toxicology, Vol. 35, pp. 93-109.
Wood, C.M., C. Hogstrand, F. Galvez, and R.S. Munger, 1996b. "The Physiology of Waterborne Silver
Toxiciry in Freshwater Rainbow Trout (Oncorhynchus mykiss) 2. The Effects of Silver Thiosulfate,"
Aquatic Toxicology, Vol. 35, pp. 111-125.
Wood, C.M., R.C. Playle and C. Hogstrand, 1999. "Physiology and Modeling of the Mechanisms of
Silver Uptake and Toxicity in Fish," Environmental Toxicology and Chemistry, 16:1, 71-83.
3-88
-------
Table 1. Biotic Ligand Model Parameter Values for Silver
3-89
-------
LIST OF FIGURES
Figure 1 Schematic of Biotic Ligand Model for Silver
Figure 2. Probability distribution of Genus Mean Acute Values (GMAVs) used to establish the
1987 draft water quality criteria for silver (left side) and probability distribution of
measurements of freshwater silver concentrations downstream of POTWs (right side)
Figure 3. Gill silver accumulation on the gills of 1 to 3 gram rainbow trout exposed for 2 to 3
hours to 0.07 uM Ag and varying levels of thiosulfate (data from Janes and Playle,
1995).
Figure 4 Calibration of WHAM to Ag-DOC complexation data of Sikora and Stevenson (1988).
Figure 5 Comparison of data to biotic hgand model results for gill silver accumulation varying
with silver concentration, sodium, chloride and DOC (data from Janes and Playle, 1995).
Figure 6 Comparison of data^Janes and Playle, 1995) to biotic hgand model results for gill silver
accumulation resulting from exposure to silver in control waters (1 and 2) and 6 natural
waters (3-8) (data from Janes and Playle, 1995).
Figure 7 Effect of calcium on 96-hour silver LC50 for fathead minnows and rainbow trout,
calculated free silver concentration, and calculated gill silver concentration (data from
Buryetal., 1999b).
Figure 8 Effect of DOC on 96-hour silver LC50 for fathead minnows and rainbow trout,
calculated free silver concentration, and calculated gill silver concentration (data from
Bury et al., 1999b).
Figure 9 Effect of chloride on 96-hour silver LC50 for fathead minnows and rainbow trout,
calculated free silver concentration, and calculated gill silver concentration (data from
Buryetal., 1999b).
Figure 10 . DOC levels in treatments at 4 chloride levels, variation in 96-hour silver LC50 for
fathead minnows by treatment, and calculated gill silver concentration, (data from Bills
etal., 1997).
Figure 11 Biotic hgand model predictions of silver LC50 for fathead minnows and rainbow trout
versus measured silver LC50. Biotic hgand LA50 =17 nmol/gw. AgCl binds to gills but
is only toxic to fathead minnows. Diagonal solid line is line of perfect agreement and
dashed lines are within a factor of+/- 2 of solid line.
Figure 12 Biotic hgand model predictions of silver LC50 for fathead minnows, rainbow trout and
Daphnia magna versus measured silver LC50. Biotic hgand LA50 = 17 nmol/gw for fish
and 2.3 nmol/gw for D. magna. AgCl binds to gills but is only toxic to fathead minnows.
Diagonal solid line is line of perfect agreement and dashed lines are within a factor of
+/- 2 of solid line.
3-90
-------
CONCEPTUAL DIAGRAM OF BIOTIC LIGAND MODEL OF
ACUTE TOXICITY OF SILVER (After Pagenkopt, 1983)
Organic
Complexes
Inorganic
Complexes
e.g.: Ag - Chlorides —
Ag - Hydroxides
Ag - Thiols
Ag - Sulfides
Gill Surface
(biotic ligand)
Active Metal
Sites
\
FIGURE 1
-------
U)
vb
N)
SILVER - FRESH WATER SPECIES DOWNSTREAM SILVER MEASUREMENTS
SILVER GENUS MEAN ACUTE VALUE, GMAV (ug/L)
1UUUO
tooo
100
10
1
0.1
0.01
0.001
J.0001
0
-
1
.1
GENUS:
1 - Daphnia
2 • Leptophlebia
3 - Ceriodaphnla x xGMAVs
4 - Gammarus
5 - Crangonyx
X
X ~
xxx*x
_ '• _xxx
D.pufex: 345
M *
Daphnia* x
D. magna
mum i IIIIIIH t i I I t i i mill ii i muni
1 1020 50 8090 99
PERCENTILE
rri
•=
-
-
:
j
99.9
10000E
1000
100
10
O
UJ
55
0.1
0.01
0.001 =
0.0001
:l 1IIIIIIT
: O
1 A
: v
' 11 Illllll 1 — III) — r~
Total Ag
Ag < 0.45 um
Ag < 0.20 um
Ag<0.10um
Detection Limit
(DL = 0.01 ng/L)
"i mini r i — mum r=
- - o E
0°
! ^'"' \
'- */ ^ =
:
;
-
i i iitim
O _» fcJ^
o +,3^
o • v
•v
«< «r
-------
Q>
o
10
Legend:
Q Control Unit
H Means with 95% C.I.
8
4
I
I
I
0.00 0.00 0.05 0.10 0.50 1.00 2.00
Thiosulfate (uM)
FIGURE 3
3-93
-------
0.9
0.8 -
0.7 -
* FA1 (fulvic sample)
• HAS (humic sample)
BLM calibration to Fulvic Acids
BLM Calibration to Humic Acids
u>
vb
•S?
"o
E
E.
CO
•o
o
CO
0.6
0.5 -
0.4 -
0.3-•
0.2-
0.1
345
Free Ag (mole/liter x 1E-4)
FIGURE 4
-------
0.0 4.3 5.4 23.7 42.1 53.9
Total Ag (ug/L)
400 400 1600 16000
Sodium (uM)
u>
0)
o
_ 20
|> 15
1 10
o> 5
5 o
Total Ag = 18.3 ug/L
300 300 400 1500 11300
Chloride (uM)
2.4 3.0 5.0 7.010.914.624.2
DOC (mg C/L)
Legend: Q Control Unit, 0 Measured (Means with 95% Cl), @ GilkAgCI, BLM Calculated, | GilhAg, BLM Calculated
Silver Binding to Gills of Rainbow Trout - Laboratory Experiments
Data: Janes and Playle, 1995
FIGURE 5
-------
O\
0>
o
Silver Binding to Gills of Rainbow Trout in Natural Waters
Data: Janes and Playle, 1995
50
^ 40
30
— 20
D)
10
; • Measured, Means with 95% C.I.
0 Control Unit
J Calculated, Ranges
1
i
ti ;' ;i
O.x . . ' .
9
1" T
1 Tl ii
i
1 1 ' 1
•
•
•
™
-
-
•
Exposure Water
8
FIGURE 6
-------
- I I Fathead Minnows
TO—.
15
<0 O>
"53
£ ™ 10
OQ
CD
a_
LL.
50
250
800
50
T3 | 40
1^ 30
OS 20
li 10
5
I I Fathead Minnows: GilhAgCI + GilhAg
Rainbow Trout: GillrAg only
50
250 800
Chloride (uM)
1500
1500
Effect of Chloride on Silver 96-hour LC50
Data: Bury, Galvez and Wood, 1998
FIGURE 7
3-97
-------
~O
co 10
+•»
o
H 5
- CHI Fathead Minnows
Rainbow Trout
0.3
1.6
20
"Sli 15
<*-» •—.
CO O)
"55
•MS- 10
o<
c_
§2 5
0
50
•of 40
IS O) rtrt
3| 3°
"<5 *=
OS 20
S O)
If 10
5 o
0.3
1.6
CD Fathead Minnows: GHI:AgCI + GillrAg
YZ& Rainbow Trout: GillrAg only
5.8
5.8
0.3 1.6 5.8
DOC (mg/L)
Effect of DOC on Silver 96-hour LC50
Data: Bury, Galvez and Wood, 1998
FIGURE 8
3-98
-------
0) 3
v. •—••
8s
C0<
OQ
—I T
OQ==
30
25
20
15
10
5
- I I Fathead Minnows
Rainbow Trout
50
500
2000
20
«j 15
1o*o)
"53
£ ™ 10
"go
OS 20
^ O)
10
0
I—I Fathead Minnows: GIlhAgCI + GIllrAg
Rainbow Trout: GilkAg only
50 500
Calcium (uM)
2000
Effect of Calcium on Silver 96-hour LC50
Data: Bury, Galvez and Wood, 1998
FIGURE 9
3-99
-------
o
o
CD
O
O
Q
O)
ro
(0
•s
0)
Qtfl
O
15
10
5
0
50
40
30
20
10
0
50
40
30
20
10
0
- Free
-Complexed
- GillrAg
Chloride = 3 mg/L
20 mg/L
40 mg/L
60 mg/L
Effect of DOC and Chloride on Silver 96-hour LC50 of 28-day Fathead Minnows
(H = 50 mg/L as CaCOS, Data: TIWET, Clemson University, 1997)
FIGURE 10
-------
Predicted versus Observed Silver LC50
100
U)
i
V—^
o
D)
O
UD
O
o
'"&
Qi
CL
CQ
1 D ' ',
• Fathead Minnows
D Rainbow Trout
10
10
Measured LC50 (ug/L)
100
FIGURE 11
-------
Predicted versus Observed Silver LC50
u>
(—•
o
100
o
.5
o
in
o
_i
•o
0)
+*
O
'•B
0>
CU
DQ
I 1 1 I I I I
• Fathead Minnows
D Rainbow Trout
A Daphnia magna
1 1—I I I I
I I I
I I I I I I
1 10
Measured LC50 (ug/L)
100
E 12
-------
Surface Water Assessment Presentation Materials
-------
Site Specific Water Quality Criteria for Metals
Herbert E. Allen
Department of Civil and Environmental Engineering
University of Delaware
Newark, DEI 9716
Water Quality Criteria (WQC) are derived from bioassay tests in which the toxicant is
added to the water. Because organisms have different sensitivities to toxicants, the LC50
values (lethal concentration to 50% of the organisms exposed) will vary greatly as shown
in Slide 1 for the freshwater criteria. Saltwater criteria are similarly derived. Criteria are
developed for protection of 95% of the species which is achieved by statistical
calculation. This provides the Final Acute Value (FAV). The freshwater criteria are
functions of the hardness, which is acknowledged to affect the toxicity. However, in
addition to hardness a number of other water quality parameters, including solids, pH,
dissolved organic matter (DOM), and inorganic ligands (Slide 2) also affect the toxicity
of metals. Presently, WQC are implemented on a dissolved basis. Thus, partitioning of
metals to suspended solids need not be considered.
Site-specific variation of water quality parameters from those present in the waters in
which the bioassays used to develop the criteria often results in the criteria not being
predictive of the actual effects. This is due to the effects of water chemistry on the metal
speciation which affect the bioavailability of the metal. The Water Effect Ratio (WER),
in which the ratio of toxicity test results for a site water compared to that for a reference
water are used to correct the criteria, has been used as a means to account for the effects
of metal speciation on a site specific basis (Slide 3). Water Effect Ratios have been
determined for a number of metals (including Cd, Cr, Cu, Pb, Ni, Ag, and Zn) and for
other constituents, such as ammonia. As shown in Slide 4, at many sites WERs are very
large, even exceeding a value of 10. This means that the WQC are over protective
relative to the intended level of protection. At some sites values less than one have been
reported for WERs indicating that the WQC are under protective. The value of the WER
is not a constant for a given water, but depends on the organism tested. More senitive
organisms (lower LC50 values in Slide 1) give higher WERs (Slide 4).
The principles involved in the WER are shown in Slide 5 for the acute toxicity of copper
in a laboratory reference and a site water. In the bioassay test various concentrations of
copper are added to both the reference and the site waters (x-axis). Because these two
waters have different concentrations of DOM present, the metal speciation (y-axis)
differs between the two waters. If the pH, hardness and other water quality parameters
are kept constant, the values on the y-axis are a good predictor of bioavailability. The
importance of this can be followed for the three organisms shown. The WQC for
polychaetes is 120 (ig/L. I have assumed that toxicity in the reference water is also 120
H-g/L. Following the solid arrow from this value on the x-axis to the y-axis gives the
equivalent bioavailable copper concentration in the reference water. Clearly, for the
same biological effect to occur in the site water, the same concentration of bioavailable
4-2
-------
metal must be present irrespective of the total metal concentration. This is shown by the
dashed arrow which extends from the y-axis to the solid line for the site water and thence
to the x-axis. The value intercepted on the x-axis is the toxicity for copper in the site
water. The ratio of the x-axis values for the site water to the reference water is the WER.
Inspection of Slide 5 indicates that the WER will not be the same for the 3 organisms
considered. This is more clearly shown in Slide 6 which presents the same data in bar
graph form for these three organisms plus two that are more sensitive. Lines for the more
sensitive organisms cannot be easily plotted in Slide 5. The sensitivity of the organisms
increases in Slide 6 from left (mysid) to right (mussel) as indicated in smaller
concentrations of copper for the LC50 values in both the reference water and the site
water. However, the non-linearity of the curves in Slide 5 result in inconsistent WERs
between the five species considered in Slide 6. The WER values dramatically increase
for the more sensitive organisms.
From these considerations of chemical speciation, it is clear that development of WQC
based on bioavailable metal would obviate the need for determining a WER. The biotic
ligand model shown in Slide 7, which will be discussed in the next presentation, couples
the environmental chemistry of metals to the response of the organism.
4-3
-------
Copper - Fresh Water Species
"5
o
CO
Q)
cn
3
o
O
100000
10000
1000
100
10
Adjusted to Hardness = 50 mg/L
Acute: 1 -hr C < 9.23 [H/50]0'9422
Chronic: 4-d C < 6.54 [H/50]0'8545
Exceedance Frequency
< once in 3 years
1. Northern Squawfish
2. Daphnia
3. Ceriodaphnia
- 4. Gammarus
Q<
&— FAV =
CMC =
FCV =
18.48^g/L
FAV/2 = 9.23 M-g
FAV/ACR = 6.54
u.g/lJ
0.1
10 30 50 70
Probability
90
99 99.9
HEA Slide 1
4-4
-------
Water Quality Parameters
Affecting Metal Toxicity
• Hardness (Ca and Mg)
• Solids
• pH
• Dissolved Organic Matter (DOM)
• Inorganic Ligands
HEA Slide 2
4-5
-------
Water Effect Ratio
Site-Specific WQC = NWQC x WER
= NWQC x site-water LC50
reference—water LC50
HEA Slide 3
4-6
-------
Water Effects Ratios
LOCATION
METAL
SPECIES
t., Q.m., S.s., P.p., A
. D. m.
WERs
Most to Least
Sensitive
St. Louis R., MN
Nemadji R., WI
Little Pokegama R., WI
Selser's Cr., LA
Cuyahoga R., OH
Lehigh River, PA (4 months)
Cadmium
Cadmium
Cadmium
Cadmium
Cadmium
Cadmium
_, _.
A, Dm.
A, D.m.
P-sp.
C.d., P.p.
C.d., P.
Boggy Cr., OK
Leon Cr., TX
Chromium
Chromium
C
A
Naugatuck R., CT
St. Louis R., MN •
Nemadji R., WI
Little Pokegama R., WI
Quinnipiac River, CT •
Columbis R. WA
Lehigh River, PA (4 months)
Wissahickon Creek, PA
Copper
Copper
Copper
Copper
Copper
Copper
Copper
Copper
A
A
D.p.
C.d.
C.d., P.p.
P.]
P.p., Q.m,
St. Louis R., MN
Norwalk River, CT
Lehigh River, PA (4 months)
Lead
Lead
Lead
A, D.m., Q.m^P.p.
D.m.,. Q.m.
g.d..P.p,
Norwalk R., CT
Boggy Cr., OK
St. Louis R., MN
Naugatuck R., CT
Cuyahoga R., OH
Lehigh River, PA (4 months)
Zinc
Zinc
Zinc
Zinc
Zinc
Zinc
D.rn^Q.m.
C
Dm,, Q.m^A.P.p.
Cd., P.p..
P.p.
Cd., P.p..
11,4.4,5.0,4.3,0.8
3.6, 2.8
1.8, 1.0
3.0
1.0, 1.7
1.0, - 9.1
1.2
2.7
1.0, 1.1
15, 9.2, 6.3, 3.3
4.3
3.1
2.5 - 36.6 (18 values)
1.2 - 3.8 (5 values)
3.5 -17.1
1.6 -7.6 (8 values)
4.1, 5.7, 3.4, 1.9
4.1, 3.7
1.0 - 25.6
Lehieh River, PA (4 months)
2.2, 1.5
1.2
2.9, 1.1, 0.8, 0.7
0.9, 0.7
0.8
1.0-7.6
A=Amphipod; C=Caddisfly; C. d. =Ceriodaphnia dubia: P_.m. =Daphnia magna: D.p. =Daphnia pulex:
Q.m.=Qnchorhvnchus mykiss: P.p.=Pimephales promelas:P.sp.=Palaemonetes sp; S.s. =Simocephalus
serrulatus: S. t. =Salmo trutta
HEASlide4
4-7
-------
Effect of Metal Complexation on Toxicity
U.
U.
Total Copper (mol/L)
q
c>
in T-
co i^
-------
Effect of Species Sensitivity on WER
500
4004-
CD
Q.
O.
O
O
Jl
O
IO
O
300-
2004-
10044
Reference Water
Site Water
B WER
CC
LLI
Mysid Polychaete Copeptod Oyster Mussel
HEA Slide 6
4-9
-------
AQUATIC BIOTIC LIGAND MODEL
CONCEPTUAL DIAGRAM
(after Pagenkopf, 1983)
Complexes
r
I Inorganic
1 Complexes
e.g. Cu - Hydroxides
Cu - Carbonates
GUI Surface
(biotic ligand)
Metal Binding
Site
HEA Slide 7
Tipping, 1994 (WHAM) Playle, 1993
4-10
-------
DRAFT
Biotic-ligand Model: Future Directions
Other freshwater species
Other metals
Marine species
Chronic toxicity
Explanatory power
Joseph S. Meyer
Department of Zoology and Physiology
University of Wyoming
Laramie, WY 82071
Presented to:
Ecological Processes and Effects Committee
Science Advisory Board
U.S. Environmental Protection Agency
Washington, DC
7ApriM999
4-11
-------
General Hypotheses
to
in
O
Metal
cone.
Total
Free ion
Tissue
burden
o
in
O
Metal
cone.
Total
Free ion
Tissue
burden
Organic matter
Hardness
-------
FRESHWATER SPECIES
Rainbow trout (Oncorhynchus mykiss)
Fathead minnow (Pimephales promelas)
Tubificid worm (Lumbriculus variegatus)
Am phi pod (Hyallela azteca)
Daphnid (Daphnia magna)
Algae
4-13
-------
Lumbriculus variegatus Exposed to Cu
^ ^
51
0
10,
1 J
0.1 -
<0
00
0.01
pH6.5
[C"tota|l
{Cu2+}
[Cubody]
10,
1 r
0.1 ->
0.01
pH7.5
-1
Hardness (meq L" )
-1
Hardness (meq L")
-------
Hyallela azteca Exposed to Copper
^ ^
fel
10° -
0
o
10'1 -
-------
METALS
Ag* Playle,
McMaster U.
Cd Campbell,
Playle,
McMaster U.
fish toxicity, gill binding &
physiology
fish toxicity & gill binding;
algal toxicity and uptake
Co Playle,
Cu* Borgmann,
Playle,
McMaster U.,
U. Delaware,
U. Wyoming
Hg Borgmann
Mi'
gill binding
fish toxicity, gill binding &
physiology;
invertebrate toxicity & uptake
Pb Borgmann
Tl Borgmann
Zn* Allen,
Borgmann,
Campbell,
McMaster U.
invertebrate toxicity & uptake
U. Wyoming fish toxicity & gill binding
invertebrate toxicity & uptake
invertebrate toxicity & uptake
fish toxicity, gill binding &
physiology;
invertebrate toxicity & uptake
algal toxicity & uptake
* = metals directly tested for biotic-ligand model
4-16
-------
Fathead Minnows Exposed to Nickel
(Pimephales promelas)
1 H
O o
m
5
pH7.5
I
2
I
6
-1
Hardness (meq L")
4-17
-------
Freshwater Amphipod Exposed to Cadmium
(Hyalella azteca)
[Cdtotal]
GO
100 r
80
80
40
20
iDbttfed
EDTA
I
100
80
60
40
20
EDTA Osfflted
Humte
' : Sediment
0.1 1 10 100
Cadmium In water (up/U
Fio. 1. Percent survival after 4 wk (sediment experiments) or 6 wk
(all experiments without sediments) as a function of measured con-
centrations of Cd in tap water (•. Tap), 20 mg humic acid/L (•,
Humic), 0.5 \iM EDTA (•, EDTA), 90% distilled water (A, Dis-
tilled), or 10% sediment by volume (broken lines: •, experiment A;
•, experiment B).
032
31B
316
Cadmium In Hyatete
Fio. 3. Percent survival after 4-6 wk as a function of Cd accumulated.
Symbols as in Fig. 1.
[from Borgmann et al. 1991, Ca. J. Fish. Aquat. Sci. 48:1055-1060]
-------
Rainbow Trout Exposed to Zinc
(Oncorhynchus my kiss)
96 h LC
ft
3
2200 -
2000 -
1800 -
1600-
1400 -
200 -
150 -
100 -
50 -
n .
•%
r^
T
•
•
i
Control 2 mM Na+ 1 mM Mg2+ 1 mM Ca'
[from Alsop and Wood 1999, manuscript]
4-19
-------
MARINE SPECIES
Tidepool sculpin (Oligocottus maculosus)
European eel (Anguilla anguilla)*
Grass shrimp (Palaemonetes pugio)
Aigae
4-20
-------
Estuarine Diatom Exposed to Copper
(Thalassiosira pseudonana)
N)
I
•S
5
I
[CUtotal]
pH 8.1-8.2
x O.SmMtrit
OIO
A 2.0
930
050
2.0
V no tris
pH7.7 , i
OlOmMtrlt ,f --- 1
HJOmMtrls /
.
4
I
&6
~7.4
-toyCur
Figure 2. Growth rate of done 3H vs the negative log of the total copper concentration in
M-f/2 seawater media containing 0.10 mM tris at pH 7.7 to 8.7. Results are from experi-
ments 1-5. Error ban represent the standard deviation for least squares linear regression of
growth curves.
(0
pCu
pH7.7
O 10 mM tris
pH8.7
10 mM tris
11
Figure 3. Growth rate of clone 3H vs pCu in M-f/2 seawater culture media containing 1-10
mM tris for the pH range 7.7 to 8.7. Individual points are from experiments 1-5. The pCo
error ban represent variation in pCu caused by changes in culture pH.
[from Sunda and Guiliard 1976, J. Mar. Res. 34:511-529]
-------
Grass Shrimp Exposed to Cadmium
(Palaemonetes pugio)
[Cdtotal]
100 •
NJ
>
i
5
u
&
Figure 3. Effect of NTA on short-term toxlclty of CdCI2 to P. puglo at
5.2%» salinity. Results from experiment 3
100
- .0
fc
2 60
i
u
& 40
20
[Cd2+]
6.0
NTA (M)
A 0
• 1 X 10 *
• 3 X IP '
1 X 10 4
7.0
Figure 5. Relationship between 4-day survival of P. puglo and measured
negative log of free cadmium Ion concentration (p[Cd2+])
Media have salinity of 5.2ft and contain from 0 to 1X. 10~4 M NTA. Results from
experiment 3.
[from Sunda et al. 1978, Environ. Sci. Technol. 12:409-413]
-------
Grass Shrimp Exposed to Cadmium
(Palaemonetes pugio)
100
M
u
30
ts)
u>
[Cdtotal]
Salinity (X.)
A A 4.1+0.4
• • t.4+0.2
T T 16.310.3
• •20.0+0.3
*—.+ 28.9+0.6
•
CdT
Figure 2. Effect of total cadmium concentration on 4-day survival of
P. pug/o at salinities of 4-29%o
Cdr Is added molar concentration of CdCI2. Results from experiments 1-3
100
«o
M «0
z
ui
U
40
20
Salinity (
A 4.1
• «.4
TI6.3
• 20.0
• 21.9
/
4.0
7.0
Figure 4. Relationship between 4-day survival of P. puglo and negative
log of measured free cadmium ion concentration (p[Cd2+])
p[Cd2+] values are means of Initial and final measurements. Media contain no
NT A. Results from experiments 1-3
[from Sunda et al. 1978, Environ. Sci. Technol. 12:409-413]
-------
Ion and Water Movement in Freshwater and Marine Fish
FIGURE 9.5 A marine teleost Is osmotically more dilute
than the water in which It lives. Because of the higher
osmotic concentration In the medium, Jhe fish con-
stantly loses water (top diagram), primarily across the
thin gill membranes. Additional water is lost In the
urine. To compensate for the water loss, the marine
FIGURE 9.6 A fresh-water teleost Is osmotically more
concentrated than the medium and therefore suffers a
steady osmotic Influx of water, mainly through the gills
(top diagram). The excess water is eliminated as urine.
Loss of solutes through the gills and in the urine is com-
pensated for primarily through active uptake in the gills
(double arrow, bottom diagram).
MARINE TELEOSTS
Water
Drinking
Solutes
Salts from
sea water,
Na*,CI~
(gill secretion)
(In urine)
FRESH-WATER TELEOSTS
Water
Solutes
Food
Solutes in urine
[from Schmidt-Nielsen 1979, Animal Physiology, 2nd Edition, Cambridge Univ. Press]
-------
Chronic Toxicity to Freshwater Species
[from Cu Criteria Document (USEPA1985)]
IUUU -
*3
(0
o 100-
E
o
o
ij 10-
o
<
1 ,
IUUU -
i
3
9
o> 100-
****
• 0
• • 2
I (0
> 10 -
•• .2
• ^ • g
5 1
0
* • » *
*/•* • *
10 100 1000 10000
-1'
Acute value (|.ig Cu L )
10 100 1000 10000
Acute value (jag Cu L1)
-------
Chronic Toxicity
Fathead minnow
Daphnia pulex
N>
3 1000 -
o
0)
J3
(0
o
CO
o
o
'E
o
£
O
100-
10
Acute
Chronic
10
100
1000
-1
Hardness (mg L as CaCCL)
80 -,
Pond
Medium
Soft
Humic acid (mg L )
[data from USEPA (1985)]
[data from Winner (1985)]
-------
Why is the slope of ln(LC50) vs.
In(hardness) approximately 1.0?
10000 r,
—> 1000 J
3
o
CD
n
ID
O
100 -J
10 J
1
1
slope
10000
-1
Hardness (mg L"1 as CaCO.)
4-27
-------
Winderinere Humic Aqueous Model (WHAM)
Presentation Materials
4-28
-------
H
Humic
Acid
>COOH
H OH
OH
Stevenson, F. J. (1982).
Humus Chemistry.
H,CO\/\H
T ICOOH
O H
Aliphatic
dicarboxylic
acid
._-:'-"•.,, Hydroxyacid
Aromatic
dicarboxylic
acid
CH2—CH2-|CO2Hi
Aliphatic
acid
Aromatic
acid
ThurmaaE. M. (1985).
Organic geochemistry of natural waters.
4-29
-------
WHAM REACTIONS
Proton Binding
(A) Carboxylic groups: COOH
(B) Phenolic groups: /~V_QH
RA-H -K RA + H+
RB-H •> R + H+
Mass Action for 1 site
exp{2wZ}
Electrostatic Effect
w=|p]log..(I)
4-30
-------
Distribution of Proton Binding Sites
A-types sites: K1 to K4 B-types sites: K4 to K8
nA = Site Density nB = Site Density
pKA = Midpoint pK PKB = Midpoint pK
ApKA = Spread in pK ApKB = Spread in pK
COOH pheno!ic-OH
1.5
S0.5H
0)
•*-»
CO
0 +-
K1 K2 K3 K4
• • • • nA/4
APKA
K5 K6 K7 K8
nA/8 . • • •
. APKB ^
PKB
0.00 5.00 10.00 15.00
pK
4-31
-------
Proton Binding Calibration Results
nA pKA pKB ApKA ApKB log10(-P)
to
-------
Metal Binding Reactions
A and B Sites
RA-H + M2+ -* RA-M+
H
RB-H + M
2+
RB-M+
pK'sforAandBSites: KM1 to KM8
KMi = KMHA ' Ki i = 1.-.
KMi = KMHB/Ki '
PKMHB = 1-38 PKMHA + 2-57
Bidentate
RA-H
RB-H
+ M
2+
RA-H
- RB-H/
KM12 = (KMA KMB)
2H+
Sites (1 ,2),(1 ,4),(1 ,6),(1 ,8),(2,3),(2,5),
(314),(3,6),(3,8),(4,5),(4,7)
4-33
-------
WHAM CALIBRATION
pN
2.5
3.0
3.5
4.0
4.5
5.0
10
8 6
p[Cu]
pN
2.5
3.0
3.5
4.0
4.5
5.0
11
9753
p[Cu]
pN
2.5
3.0
3.5
4.0
4.5
5.0
Cd
654
PtCd]
pN
2.5 r
3.5
4.5
5.5
10 8 6 4
p[Cu]
PN 4
10 8 6 4
P[Cu]
2.5
pN 3.5
4.5
6.0 5.0 4.0
p[Pb]
4-34
-------
Binding Site Occupancy
1.0
08
0.6
0.4
•| 0.2
Q
ii n O
r pM = 9
•
I
\
i
tffa
pM = 7
•
.
i
:
_ \ ^
To
0>
«••
1 °-5
o 0.4
CO
0.3
0.2
0.1
0.0
pM = 5
•
- i
• J ^ JUL
r pM = 3
-
i
J
PW m mi ,-, •
ra — V^//////////A m
12345678 123456789101112 D 12345678 123456789101112 D
MONO- BIDENTATE MONO- BIDENTATE
DENTATE DENTATE
4-35
-------
Data Fitting Results.
PKMA
Ca
Table 7. Tipping. E. and M. Hurley (1992).
Geochim. Cosmochim. Acta 56: 3627-3641.
4-36
-------
Integrated Approach to Assessing the Bioavailability
and Toxicity of Metals in Surface Waters and
Sediments
Appendices
Presented to the EPA Science Advisory Board
April 6-7, 1999
U.S. Environmental Protection Agency
Office of Water
Office of Research and Development
Washington, D.C.
-------
APPENDIX A
-------
Research Papers on the Unavailability and Toxicity of Metals in Sediments1
Key Papers (Roughly in order of importance)
Ankley, G.T., D.M. Di Toro, DJ. Hansen, J.D. Mahoney, W. J. Berry, R.C. Swartz, R. A. Hoke, N.A.
Thomas, A.W. Garrison, H.E. Allen and C.S. Zarba. 1994. Assessing potential bioavailability of
metals in sediments: A proposed approach. Environ. Management. 18:331-337.
Ankley, G.T., D.M. Di Toro, DJ. Hansen and WJ. Berry. 1996. Technical basis and proposal for
deriving sediment quality criteria for metals. Environ. Toxicol. Chem. 15:2056-2066.
Di Toro, D.M., J.D. Mahoney, DJ. Hansen, K J. Scott, M.B. Hicks^ S.M. Mayr and M.S. Redmond.
1990. Toxicity of cadmium in sediments: The role of acid volatile sulfide. Environ. Toxicol. Chem.
9:1487-1502.
Hansen, D J., W J. Berry, J.D. Mahoney, W.S. Boothman, D.M. Di Toro, D.L. Robson, G.T. Ankley,
D. Ma, Q. Yan and C.E. Pesch. 1996. Predicting the toxicity of metal-contaminated field sediments
using interstitial concentration of metals and acid-volatile sulfide normalizations. Environ. Toxicol.
Chem. 15:2080-2094.
Berry, W J., D. J. Hansen, J.D. Mahoney, D.L. Robson, D.M. Di Toro, B.P. Shipley, B. Rogers, J.M.
Corbin and W.S. Boothman. 1996. Predicting the toxicity of metal-spiked laboratory sediments
using acid-volatile sulfide and interstitial water normalizations. Environ. Toxicol. Chem. 15:2067-
2079.
Call, DJ., C.N. Polkinghome, T.P. Markee, L.T. Brooke, D.L. Geiger, J.W. Gorsuch and K.A.
Robillard. 1999. Silver toxicity to Chironomus tentans in two freshwater sediments. Environ.
Toxicol. Chem. 18:30-39.
Berry, W J., M.G. Cantwell, P. A. Edwards, J.R. Serbst and D J. Hansen. 1999. Predicting toxicity
of sediments spiked with silver. Environ. Toxicol. Chem. 18:40-48.
Di Toro, D.M., J.D. Mahoney, DJ. Hansen, KJ. Scott, A.R. Carlson and G.T. Ankley. 1992. Acid
Volatile Sulfide predicts the acute toxicity of cadmium and nickel in sediments. Envrion. Sci.
Technol. 26:96-101.
Other Relevant Papers
Ankley, G.T. 1996. Evaluation of metal/acid-volatile sulfide relationships in the prediction of metal
bioaccumulation by benthic macroinvertebrates. Environ. Toxicol. Chem. 15:2138-2146.
'Articles listed as "Key Papers" are reprinted in this volume. Additional relevant papers
have been listed, but not reprinted.
-------
DeWitt, T.H., R.C. Swartz, D J. Hansen, D. McGovern and W.J. Berry. 1996. Bioavailability and
chronic toxicity of cadmium in sediment to the estuarine amphipod Leptocheirus plumulosns.
Environ. Toxicol. Chem. 15:2095-2101.
Di Toro, D.M., J.D. Mahoney and A.M. Gonzalez. 1996. Particle oxidation model of synthetic FeS
and sediment acid-volatile sulfide. Environ. Toxicol. Chem. 15:2156-2167.
Di Toro, D.M., J.D. Mahoney, DJ. Hansen and W.J. Berry. 1996. A model of the oxidation of iron
and cadmium sulfide in sediments. Environ. Toxicol. Chem. 15:2168-2186.
Gonzalez, A.M. 1996. A laboratory-formulated sediment incorporating synthetic acid-volatile
sulfide. Environ. Toxicol. Chem. 15:2209-2220.
Hansen, DJ., J.D. Mahoney, W.J. Berry, SJ. Benyi, J.M. Corbin, S.D. Pratt, D.M. Di Toro andM.B.
Abel. 1996. Chronic effect of cadmium in sediments on colonization by benthic marine organisms:
An evaluation of the role of interstitial cadmium and acid-volatile sulfide in biological availability.
Environ. Toxicol. Chem. 15:2126-2137.
Hassan, S.M., A.W. Garrison, H.E. Allen, D.M. Di Toro and G.T. Ankley. 1996. Estimation of
partition coefficients for five trace metals in sandy sediments and application to sediment quality
criteria. Environ. Toxicol. Chem. 15:2198-2208.
Leonard, E.N., G.T. Ankley and R.H. Hoke. Evaluation of metals in marine and freshwater surficial
sediments from the environmental monitoring and assessment program relative to proposed sediment
quality criteria for metals. Environ. Toxicol. Chem. 15:2221-2232.
Liber, K., DJ. Call, T.P. Markee, K.L. Schmude, M.D. Balcer, F.W. Whiteman and G.T. Ankley.
1996. Effects of acid-volatile sulfide on zinc bioavailability and toxicity to benthic
macroinvertebrates: A spiked-sediment field experiment. Environ. Toxicol. Chem. 15:2113-2125.
Mahoney, J.D., D.M. DiToro, A.M. Gonzalez, M. Curto, M. Dilg, L.D. De Rosa and L.A. Sparrow.
1996. Partitioning of metals to sediment organic carbon. Environ. Toxicol. Chem. 15:2187-2197.
Peterson, G.S., G.T. Ankley and E.N. Leonard. 1996. Effect of bioturbation on metal-sulfide
oxidation in surficial freshwater sediments. Environ. Toxicol. Chem. 15:2147-2155.
Sibley, P.K., G.T. Ankley, A.M. Cotter and E.N. Leonard. 1996. Predicting chronic toxicity of
sediments spiked with zinc: An evaluation of the acid-volatile sulfide model using a life-cycle test
with the midge Chironomus tentans. Environ. Toxicol. Chem. 15:2102-2112.
-------
Assessing Potential Bioavailability of Metals in
Sediments: A Proposed Approach
GERALD T. ANKLEY*
NELSON A. THOMAS
US Environmental Protection Agency
6201 Congdon Blvd., Ouluth. Minnesota 55804, USA
DOMINIC M. Dl TORO
Hydroqual, Inc
Mahwah, New Jersey 07430, USA
DAVID J. HANSEN
US Environmental Protection Agency
27 Tarzwell Or, Narragansett, Rhode Island 02882, USA
JOHN D. MAHONY
Chemistry Department. Manhattan College
Bronx. New York 10471, USA
WALTER J. BERRY
SAIC Corporation
27 Tarzwell Dr., Narragansett. Rhode Island 02882, USA
RICHARD C. SWARTZ
US Environmental Protection Agency
Hatfield Manne Science Center
Marine Science Drive, Newport, Oregon 97365, USA
ROBERT A. HOKE
SAIC Corporation
411 Hackensack Ave ,
Hackensack, New Jersey 07601, USA
A. WAYNE GARRISON
US Environmental Protection Agency
College Station Rd., Athens, Georgia 30605, USA
HERBERT E. ALLEN
Department of Civil Engineering, University of Delaware
Newark, Delaware 19716, USA
CHRISTOPHER S. ZARBA
US Environmental Protection Agency
401 M Street S W.. Washington, DC 20460. USA
ABSTRACT / Due to anthropogenic inputs, elevated
concentrations of metals frequently occur in aquatic
sediments. In order to make defensible estimates of the
potential risk of metals in sediments and/or develop sediment
quality criteria for metals, it is essential to identify that fraction
of the total metal in the sediments that is bioavailable.
Studies with a variety of benthic invertebrates indicate that
interstitial (pore) water concentrations of metals correspond
very well with the bioavailabilrty of metals in test sediments.
Many factors may influence pore water concentrations of
metals, however, in anaerobic sediments a key phase
controlling partitioning of several cationic metals (cadmium,
nickel, lead, zinc, copper) into pore water is acid volatile
sulfide (AVS) In this paper, we present an overview of the
technical basis for predicting bioavailability of cationic metals
to benthic organisms based on pore water metal
concentrations and metal-AVS relationships. Included are
discussions of the advantages and limitations of metal
bioavailability predictions based on these parameters.
relative both to site-specific assessments and the
development of sediment quality criteria.
Introduction and Technical Background
The occurrence of elevated concentrations of met-
als in aquatic sediments is common, and resource
managers often must decide whether these concentra-
tions may result in adverse ecosystem impacts, and
KEY WORDS Sediment, Metal, Bioavailability, Toxicrty, Sediment
quality criteria
*To whom correspondence should be addressed
whether remedial action is required. Unfortunately,
this can be difficult because it is impossible to utilize
total concentrations of metals in sediments to predict
the occurrence of environmental impacts. Differences
in physicochemical properties among sediment types
result in variations in biologically available metal at
any given total metal concentration (for review, see
Luoma 1989). Some have attempted to address this by
predicting metal bioavailability in sediments through
the use of differential extraction and fractionation
schemes (e.g., Tessier and Campbell 1987). The ma-
Environmental Management Vol 18, No 3, pp 331-337
© 1994 Spnnger-Verlag New York Inc
-------
332
G. T. Ankley and others
jor shortcoming with this type of approach, however,
is that often it has been difficult to relate observed
chemistry results to the end point of concern, that is,
biological exposure and response.
Early studies by Swartz and others (1985) demon-
strated a correlation between sediment interstitial
(pore) water concentrations of cadmium and the
acute toxicity of cadmium-spiked sediments to the
marine amphipod Rhepoxynius abronius. This observa-
tion suggested that, based on acute toxicity, cadmium
in pore water represented the bioavailable fraction of
the total cadmium in sediment. This observation also
was consistent with the results of studies demonstrat-
ing that the toxicity of nonionic organic chemicals to
epibenthic and benthic species could be predicted by
comparison of chemical concentrations in interstitial
water to toxicity data generated in water-only expo-
sures (e.g., Adams and others 1985, Swartz and others
1990). These observations served as the basis for the
development of the equilibrium partitioning approach
to deriving sediment quality criteria (SQC) for non-
ionic organic compounds (Di Toro and others 1991).
In the case of nonionic organic chemicals, organic
carbon is the major partitioning phase controlling
pore water concentrations and bioavailability in sedi-
ments; however, until recently, comparable partition-
ing phases had not been as well defined for metals.
In spiking experiments by Di Toro and others
(1990), it was demonstrated that the acute toxicity of
cadmium to amphipods (Ampelisca abdita, R. hudsoni,
R. abronius) in marine sediments could be predicted
based upon acid volatile sulfide (AVS) content of the
sediments. AVS is defined as that fraction of sulfide in
sediments extracted by cold HC1, and it exists in natu-
ral sediments primarily as iron sulfide complexes
commonly referred to as mackinawite and greigite
(Berner 1967, Goldhaber and Kaplan 1974). A num-
ber of cationic metals of environmental concern (zinc,
lead, copper, nickel, cadmium) will displace iron from
the monosulfide and thereby may be rendered biolog-
ically unavailable. Thus, in the experiments by Di
Toro and others (1990), when the molar ratio of cad-
mium to AVS exceeded one (i.e., the AVS binding
pool was exhausted), interstitial water concentrations
of free cadmium increased dramatically, and there
was a corresponding increase in amphipod mortality.
Subsequent spiking experiments conducted by
Carlson and others (1991) used AVS concentrations
to accurately predict cadmium toxicity in freshwater
sediments. In those experiments, acute toxicity to oli-
gochaetes (Lumbriculus variegatus) and snails (Helisoma
sp.) was observed only when the molar cadmium/AVS
ratio exceeded one.
Based on these studies, Di Toro and others (1992)
proposed that AVS should be a suitable normalization
phase for predicting the lack of acute toxicity of cad-
mium, nickel, lead, zinc, and copper in both marine
and freshwater sediments. Di Toro and others (1992)
also described the utilization of simultaneously ex-
tracted metal (SEM) for normalization to AVS. That
is, it would be inappropriate to use total sediment
metals for normalization to AVS; metals should be
measured in the same acid extraction fraction as that
used for liberation of the AVS from the test sedi-
ments. With the harsher extractions commonly used
to measure "total" metals in sediments, a significant
portion of the metals liberated are from the mineral-
ogical matrix of the sediment, and hence are of little
significance from the standpoint of biological avail-
ability and effects.
Ankley and others (1991) further investigated the
use of AVS normalization for prediction of the bio-
availability of metals in field-collected sediments con-
taminated by multiple metals. Sediments from the up-
per end of a marine tidal estuary contaminated with
cadmium and nickel were tested for toxicity to fresh-
water amphipods (HyaUUa azteca) and oligochaetes (L.
variegatus). Toxicity of the sediments to H. azteca, a
relatively sensitive species, could be predicted based
on SEM (nickel plus cadmium)/AVS ratios in the sedi-
ments. Elevated amphipod mortality was observed
when the molar SEM/AVS ratio exceeded one, while
toxicity was not seen at ratios of less than one. Oli-
gochaetes, which are relatively tolerant of metals,
were less sensitive than the amphipods to the test sed-
iments; however, bioaccumulation of metals by the
worms was correlated with the sediment SEM/AVS
ratios. Ankley and others (1991) also noted a correla-
tion between sediment toxicity and pore water cad-
mium and nickel concentrations.
The same sediments evaluated by Ankley and oth-
ers (1991) also were tested in salt water using marine
amphipods (A. abdita). In these experiments, toxicity
was not observed at SEM/AVS ratios less than one;
however, toxicity also was not observed in several sam-
ples with ratios greater than one (unpublished data).
Evaluation of pore water concentrations of the nickel
and cadmium, however, did provide an accurate pre-
diction of the occurrence of toxicity. Regardless of the
SEM/AVS ratio, toxicity was observed only in those
sediment samples with pore water concentrations of
cadmium and nickel that exceeded the joint water-
only toxicity of the two metals to A. abdita. These data
suggested the presence of binding phases in addition
to AVS for cadmium and nickel in the test sediments.
Further laboratory spiking of marine sediments
-------
Metal Bioavailability in Sediments
333
with cadmium, zinc, nickel, and copper, either singly
or in combination, demonstrated that acute toxicity to
A. abdita could be predicted based upon SEM/AVS
ratios in the test sediments (Hansen and others 1990;
Berry and others 1991). Toxicity was observed when
SEM/AVS ratios exceeded one, while at ratios less
than one, survival of the amphipods was comparable
to that in the control sediments. Pore water concentra-
tions of the cadmium, zinc, nickel, and copper also
could be used to accurately predict the occurrence
and extent of toxicity to the amphipods, that is, mor-
tality was observed only at pore water metal concen-
trations exceeding those causing mortality to the
amphipods in water-only exposures. In the spiking
experiments with multiple metals, it also was noted
that, in the sediments with SEM/AVS ratios greater
than one, relative pore water concentrations of the
test metals were inversely proportional to their bind-
ing constants for the sulfide.
Ankley and others (1993a) evaluated AVS as a nor-
malization phase for determining copper bioavailabil-
ity in sediments from two freshwater sites contami-
nated with the metal: Steilacoom Lake, Washington,
and the Keweenaw Waterway, Michigan. SEM
(copper)/AVS ratios in the test sediments were found
to overpredict copper bioavailability in these studies.
Acute toxicity to H. axteca did not occur in several test
sediments with extremely high copper/AVS ratios,
nor were pore water copper concentrations elevated
in these sediments. This suggested the presence of an
additional binding phase(s) for copper. Sediment ti-
tration experiments, similar to those described by Di
Toro and others (1990) for cadmium, subsequently
confirmed the presence of. strong binding phases,
other than AVS, for copper in anaerobic sediments
(Mahony and others 1991). This additional binding
capacity was correlated with the organic carbon con-
tent of the test sediments. Significantly, although
copper/A VS ratios could not be used to predict the
occurrence of acute toxicity of copper to amphipods,
measurement of pore water copper concentrations
and comparison of these data to water-only copper
toxicity data'for H. azteca enabled accurate predic-
tion of the presence and extent of sediment toxicity
(Ankley and others 1993a).
The majority of studies examining SEM/AVS ra-
tios and/or pore water metal concentrations as expo-
sure models has been limited to short-term tests with
lethality as an end point. One exception was a recent
study by Ankley and others (1993b) which evaluated
the bioaccumulation of copper, lead, zinc, cadmium,
nickel, and chromium by L. vanegatus held for an
extended time in three different sediment samples
from the lower Fox River, Wisconsin. These sedi-
ments had elevated concentrations of all six metals;
however, based upon SEM/AVS ratios or pore water
metal concentrations, metals in the sediments were
predicted to be of minimal biological availability to the
oligochaetes. After 30 days of exposure, metal con-
centrations in L. vanegatus held in the sediments were
similar to concentrations in worms held in clean wa-
ter. This indicates that it may be possible to use metal/
AVS relationships to predict metal exposure in long-
term as well as short-term studies with benthic
invertebrates.
AVS concentrations in aquatic sediments are inti-
mately related to a number of biogeochemical cycles
and thus may vary greatly. In addition to varying
among sites, AVS concentrations at a particular site
change with depth and with season. Therefore, in the
absence of alternative binding phases, it is possible
that metals in some systems may be bioavailable dur-
ing some portions of the year but not during others.
To address this, studies were initiated to better define
the seasonal variability in AVS concentrations among
sites and at sediment depths. For example, one study
has monitored sediment AVS concentrations hi three
northern Minnesota lakes for approximately two
years (Leonard and others 1993). Although the three
lakes were relatively similar, both geographically and
hydrologically, there were marked variations among
them in sediment AVS concentrations. Furthermore,
AVS concentrations varied with depth, with decreas-
ing amounts of AVS at greater depths. Most impor-
tantly there were marked seasonal variations in AVS,
in some instances almost two orders of magnitude,
with the minimum concentrations generally occurring
in late winter and the maximum concentrations ob-
served in late spring and early summer. These sea-
sonal variations were much more pronounced in surf-
icial sediments than in sediments at greater depths.
Limitations to Prediction of Metal Toxicity
in Sediments
Based on the above studies it appears that evalua-
tion of pore water metal concentrations and/or SEM/
AVS ratios can lend insights concerning metal bio-
availability in sediments. We feel that the two
techniques are complementary and should be used in
conjunction with one another as one approach to pro-
viding assessments of the potential ecological impacts
of metals in sediments. Both techniques have specific
problems and common limitations. These problems
and limitations are discussed below.
-------
334
G. T. Ankley and others
It appears that comparison of metal concentrations
in pore water to water-only toxicity data can be used to
predict not only the presence, but also the extent, of
metal toxicity in sediments. The ability to actually
quantify bioavailable metals in sediments is attractive
for a number of reasons. For example, quantification
of bioavailable metal facilitates the evaluation of dif-
ferences in relative species sensitivity and thus enables
the identification of species at risk. This is not possible
with SEM/AVS ratios because when SEM/AVS ratios
exceed one, due to other possible binding phases, it is
impossible to predict actual pore water concentrations
of metals (e.g., Ankley and others 1991, 1993a). An-
other advantage to monitoring pore water metal con-
centrations is that they should be useful for predicting
the toxicity of metals, such as chromium, not reactive
with AVS. Finally, because AVS is readily oxidized, it
is not an important binding phase for metals in aero-
bic sediments; however, the bioavailable fraction of
metals should still be predictable based upon pore
water concentrations.
There also are disadvantages to using pore water
concentrations to predict metal bioavailability. First,
because pore water is operationally defined (i.e., there
is no standard method for isolation), there is a valid
concern that among-Iaboratory variations in prepara-
tion may result in significant differences in metal con-
centrations found in pore water. A1 second disadvan-
tage to using pore water metal concentrations to
predict toxicity is that, if one accepts the paradigm
that pore water is indeed a major route of contami-
nant exposure for epibenthic and benthic inverte-
brates, it may be difficult to account for the effects of
the pore water matrix (e.g., dissolved organic carbon,
hardness, salinity) on metal complexation and bio-
availability. This is, of course, also an issue of ongoing
concern in the area of water quality criteria issued by
the US Environmental Protection Agency. A final po-
tential complication is that it is necessary to have a
water-only effects data base for comparative purposes
for the metal and species of concern.
Our studies have clearly demonstrated that AVS
can be a key factor influencing interstitial water con-
centrations and bioavailability of metals in sediments.
In no instance have we seen metal toxicity at SEM/
AVS ratios less than one, and ratios greater than one
often have been predictive of the presence (but not
extent) of metal toxicity. The use of SEM/AVS con-
centrations alleviates the need for water-only effects
data in an assessment since no bioavailability is ex-
pected at SEM/AVS ratios less than one. A further
advantage to measuring sediment SEM and AVS in
sediments is that it gives an indication as to the relative
size of the pool of both components. This is not possi-
ble through monitoring pore water metal concentra-
tions; pore water metal concentrations should be low
in sediments with SEM/AVS ratios ranging from ex-
tremely low values to, theoretically, a ratio of 0.99.
Yet, sediments with relatively high ratios would be of
more potential concern than those with low ratios; in
the absence of other metal binding phases, slight in-
creases in SEM or decreases in AVS could cause the
SEM/AVS ratio to exceed one and, thereby, result in
toxicity.
There are a number of limitations inherent in us-
ing either pore water or SEM/AVS ratios to predict
metal bioavailability. First, because AVS varies season-
ally in a system-specific manner, it is desirable that
SEM/AVS ratios and pore water metals be measured
over time, or at least when AVS is expected to be
minimal (e.g., late winter in our studies). A single
sampling is only a snap-shot of what occurs through
the course of the year. Furthermore, even under re-
ducing conditions, biological activity in surficial sedi-
ments may serve to effectively oxidize AVS. At
present, there is little understanding of the role of
AVS in deeper sediments relative to metal partition-
ing at the sediment surface, where most biological
activity and exposure occurs. This is significant be-
cause the AVS pool in deeper sediments appears to
remain relatively constant, as opposed to AVS in surf-
icial sediments.. It may be, for example, that as surfi-
cial sediments are depleted of AVS, metals will
subsequently bind to AVS in deeper sediments. Alter-
natively, as AVS concentrations are depleted in sur-
face sediments, other binding phases for metals may
become important in determining bioavailability.
Neither pore water metal concentrations nor SEM/
AVS ratios can be used to assess potential metal bio-
availability in situations where sediments are expected
to be altered and become aerobic through physical
disturbance (e.g., storms, boat traffic, dredging). In
fact, in these cases, it may be appropriate to "exhaust"
sediment AVS (e.g., by aerating the sample) before
attempting to evaluate the presence of bioavailable
metals, possibly through evaluation of pore water
metal concentrations.
An important limitation in using either pore water
metal concentrations or SEM/AVS ratios to evaluate
metal bioavailability is that these approaches have not
been thoroughly validated for predicting chronic tox-
icity. With the exception of the 30-day bioaccumula-
tion study described by Ankley and others (1993b)
and a recently completed marine colonization experi-
ment, all the laboratory exposures conducted thus far
have been relatively short and generally used only
-------
Metal Bioavailability in Sediments
335
lethality as an end point. Thus, because of a current
lack of understanding of the dynamics of metals and
AVS during long-term exposures, we do not sanction
the use of either pore water metal concentrations or
SEM/AVS ratios for predicting the absence or pres-
ence of chronic toxicity.
A final limitation to using either approach for pre-
dicting even acute toxicity to epibenthic or benthic
species is that neither model has been field validated.
To date, the most exhaustive studies have been con-
ducted in a laboratory setting. Preliminary field stud-
ies focused on changes in benthic community struc-
ture or bioaccumulation of metals by benthos have
been consistent with predictions based upon a pore
water exposure model (Campbell and Tessier 1991,
L. Hare, University of Quebec, personal communica-
tion); however, further work in this area is needed.
Therefore, care should be taken in interpretation of
SEM/AVS ratios and pore water metal concentrations
in the field relative to predicting the presence or ab-
sence of shifts in benthic community structure.
Recommendations for Predicting Metal
Bioavailability in Sediments
Based both on the technical considerations and
limitations described above, we present the following
recommendations/caveats for assessing the potential
bioavailability of metals in sediments.
1. Both SEM/AVS ratios and pore water metal con-
centrations should be measured in sediment as-
sessments focused on defining metal bioavailabil-
ity. A standard method for the extraction and
measurement of SEM and AVS has been de-
scribed by Allen and others (1993). For the stud-
ies described above, pore water was isolated using
either of two different techniques: dialysis cham-
bers (peepers) or centrifugation (e.g., see Di Toro
and others 1990, Ankley and others 1991). Other
pore water isolation techniques also may be use-
ful; however, we have had little experience with
•them.
2. If AVS is used as a normalization phase, it should
be used only for cadmium, nickel, lead, zinc, and
copper, and only for these metals when simulta-
neously extracted with the AVS. Molar concentra-
tions of the metals then can be summed to gener-
ate SEM/AVS ratios (Di Toro and others 1992)
Theoretically, however, it is possible to utilize
pore water measurements of metals, other than
the five listed above, to evaluate their potential
bioavailability (e.g., chromium; Ankley and oth-
ers 1993b).
3. It is strongly recommended that cadmium, nickel,
copper, lead, and zinc all be measured when eval-
uating SEM/AVS ratios and pore water metal con-
centrations, at least in initial test samples. This is
because although all five of these metals have a
higher affinity than iron for sulfide in monosul-
fide complexes, individually they also have vary-
ing affinities (solubility products) for the sulfide.
Thus, for example, cadmium will displace nickel
from sulfide, and if excess sulfide is not available,
nickel will be released to the pore water. If only
pore water metals were measured, or only nickel
was measured in the solid phase, the analyst
would erroneously conclude that nickel was die
only problem in the sediments, when in fact, ele-
vated concentrations of cadmium also were
present. In order to have a complete understand-
ing as to why a particular metal is present at ele-
vated concentrations in pore water, it is necessary
to know the molar concentrations of all the SEM.
This is particularly true when considering the fact
that metal concentrations often covary in contam-
inated aquatic sediments, that is, rarely is only one
metal of concern.
4. In fully aerobic sediments (e.g., sand), AVS con-
centrations should not be used to attempt to pre-
dict the bioavailability of metals in sediments.
Theoretically, however, it should be possible to
infer bioavailability based on pore water metal
concentrations. Moreover, significant progress is
being made in identifying alternative normaliza-
tion phases for metals in aerobic sediments (Camp-
bell and Tessier 1991, Mahony and others 1991,
Tessier and others 1993).
5. Only a limited amount of research has been con-
ducted to assess the utility of SEM/AVS ratios or
pore water concentrations for predicting metal
bioavailability in long-term exposures. Given un-
certainties in kinetics of metal and AVS interac-
tions in temporal cycles, and a lack of information
of the importance of other metal binding phases
relative to these cycles, extrapolations of the ex-
posure model to long-term situations should be
made with care. Further information also is
needed concerning the nature of the microhabitat
of invertebrates relative to long-term changes in
metal bioavailability in sediments.
6. As with any chemical-specific monitoring meth-
od, the analyst should be aware that: (a) not all
chemicals of possible lexicological concern can be
measured in environmental samples; and (b) in
most instances, it is difficult to account for possi-
ble lexicological interactions among measured
-------
336
G T. Ankley and others
toxicants. For both these reasons, we strongly rec-
ommend that toxicity tests be an integral part of
any assessment concerned with the effects of sed-
iment contaminants.
Summary
Significant progress has been made in understand-
ing factors influencing the bioavailability of metals in
sediments, particularly from the standpoint of acute
toxicity. Evaluation of pore water metal concentra-
tions, as well as AVS-metal interactions, has enabled a
much better understanding of the dynamics of metal
bioavailability in sediments. However, researchers uti-
lizing SEM/AVS ratios and/or pore water metal con-
centrations to evaluate metal bioavailability in sedi-
ments should be aware of possible drawbacks and
limitations to these approaches. For example, the con-
sequences of spatial (depth) and/or seasonal variations
in AVS relative to metal bioavailability need to be
better defined, and further research is required to
determine if metal bioavailability predictions made
using these approaches are valid for predicting
chronic toxicity and/or impacts in situ. Research in all
these areas is ongoing, and a summary of our results,
including a theoretical basis for developing SQC for
metals, will receive a formal critical review by the Sci-
ence Advisory Board of the US Environmental Pro-
tection Agency.
Acknowledgments
We thank all of our colleagues at the Duluth and
Narragansett EPA laboratories and Manhattan Col-
lege for input to various technical aspects of this work.
Literature Cited
Adams, W. J., R. A. Kimerle, and R. G. Mosher. 1985
Aquatic safety assessment of chemicals sorbed to sedi-
ments. Pages 429-453 in R. D. Cardwell, R. Purdy and
R. C. Bahner (eds.), Aquatic toxicology and hazard assess-
ment: Seventh symposium. STP 854. American Society
for Testing and Materials, Philadelphia, Pennsylvania
Allen, H. E., G. Fu, and B. Deng. 1993. Analysis of acid
volatile sulfide (AVS) and simultaneously extracted metals
(SEM) for the estimation of potential toxicity in aquatic
sediments. Environmental Toxicology and Chemistry 12
1441-1453.
Ankley, G. T., G. L. Phipps, E N. Leonard, D. A. Benoil,
V. R. Mattson, P. A. Kosian, A. M. Cotter, J. R. Dierkes, D
J. Hansen, and J. D. Mahony. 1991. Acid volatile sulfide as
a factor mediating cadmium and nickel bioavailabiliu in
contaminated sediments. Environmental Toxicology and
Chemistry 10.1299-1307.
Ankley, G. T., V. R. Mattson, E. N. Leonard, C. W. West,
and J. L. Bennett. 1993a. Predicting the acute toxicity of
copper in freshwater sediments: Evaluation of the role of
acid volatile sulfide. Environmental Toxicology and Chemistry
12:315-320.
Ankley, G. T., E. N. Leonard, and V. R. Mattson. 1993b.
Prediction of the bioaccumulation of metals from contam-
inated sediments by the oligochaete, Lumbriailus vanega-
tus. Water Research (in press).
Berner, R. A. 1967. Thermodynamic stability of sedimen-
tary iron sulfides. American Journal of Science 265:773-785.
Berry, W. J., D. J. Hansen, J. D. Mahony, D. L. Robson, and
J. M. Corbin. 1991. The role of acid volatile sulfide in
controlling the toxicity of a metals mixture in sediment
Page 91 in Abstracts of the 12th Annual Meeting of the
Society of Environmental Toxicology and Chemistry.
Novermber 1991. Seattle, Washington.
Campbell, P. G. C., and A. Tessier. 1991. Biological avail-
ability of metals in sediments: analytical approaches
Pages 161-174 m J.-P. Vernet (ed.), Heavy metals in the
environment. Elsevier, New York.
Carlson, A. R., G. L. Phipps, V. R. Mattson, P. A. Kosian,
and A. M. Cotter. 1991. The role of acid volatile sulfide in
determining cadmium bioavailability and toxicity in fresh-
water sediments. Environmental Toxicology and Chemistry
10:1309-1319.
Di Toro, D. M., J. D. Mahony, D. J. Hansen, K. J. Scott,
M. B. Hicks, S. M. Mayr, and M. S. Redmond. 1990. Tox-
icity of cadmium in sediments: The role of acid volatile
sulfide. Environmental Toxicology and Chemistry 9:1487-
1502.
Di Toro, D. M., C. Zarba, D. J. Hansen, R. C. Swartz, C. E.
Cowan, H. E. Allen, N. A. Thomas, P. R. Paquin, and
W. J. Berry. 1991. Technical basis for establishing sedi-
ment quality criteria for non-ionic organic chemicals us-
ing equilibrium partitioning. Environmental Toxicology and
Chemistry 10:1541-1583.
Di Toro, D. M., J. D. Mahony, D. J. Hansen, K. J. Scott, A. R.
Carlson, and G. T. Ankley. 1992. Acid volatile sulfide
predicts the acute toxicity of cadmium and nickel in sedi-
ments. Environmental Science and Technology 26:96—101.
Goldhaber, M. B. and I. R. Kaplan. 1974. The sulfur cycle.
Pages 569-655 tn E. D. Goldberg (ed.), Sea, Vol. 5—Ma-
rine chemistry. John Wiley & Sons, New York.
Hansen, D. J., W. J. Berry, J. D. Mahony, A R. Carlson,
K M. McKenna, and D. L. Robson. 1990. Acid volatile
sulfide controls divalent metal toxicity in sediments. Page
88 m Abstracts of the 11th Annual meeting of the Society
of Environmental Toxicology and Chemistry, Washing-
ton, DC.
Leonard, E. N., V. R. Mattson, D. A. Benoit, R. A. Hoke,
and G. T. Ankley. 1993. Seasonal variation of acid volatile
sulfide in sediments from three northeastern Minnesota
lakes. Hydrobiologia (in press).
Luoma, S. N. 1989. Can we determine the biological avail-
ability of sediment-bound trace elements? Hydrobiologia
176/177:379-396.
Mahony, J. D., D. M. Di Toro, A. M. Gonzalez, D J. Hansen,
-------
Metal Bioavailability in Sediments
337
W. J. Berry, and G. T. Ankley. 1991. A sediment compo-
nent in addition to acid volatile sulfide that may further
control toxicity of metals. Page 91 m Abstracts of the 12th
Annual Meeting of the Society of Environmental Toxicol-
ogy and Chemistry, Seattle, Washington.
Swartz, R. C., G. R. Ditsworth, D. W. Schults, and J. O.
Lamberson. 1985. Sediment toxicity to a marine infaunal
amphipod: Cadmium and its interaction with sewage
sludge. Marine Environmental Research 18:133-153.
Swartz, R. C., D. W. Schults, T. H. DeWitt, G. R. Ditsworth,
and J. O. Lamberson. 1990. Toxicity of fluoranthene in
sediment to marine amphipods: A test of the equilibrium
partitioning approach to sediment quality criteria. Envi-
ronmental Toxicology and Chemistry 9:1071-1080.
Tessier, A., and P. G. C. Campbell. 1987. Partitioning of
trace metals in sediments: Relationships with bioavailabil-
ity. Hydrobiologia 149:43-52.
Tessier, A., Y. Couillard, P. G. C. Campbell, and J. C. Au-
clair. 1993. Modeling cadmium partitioning in oxic lake
sediments and Cd concentrations in the freshwater bi-
valve Anodonta grandis (Mollusca, Pelecypoda). Limnology
and Oceanography 38:1—17.
-------
Environmental Toxicology and Chemistry, Vol 15, No 12, pp 2056-2066, 1996
C 1996 SETAC
Printed in the USA
0730-7268/96 $6.00 + 00
TECHNICAL BASIS AND PROPOSAL FOR DERIVING SEDIMENT
QUALITY CRITERIA FOR METALS
GERALD T. ANKLEY,*! DOMINIC M. Di ToRo4 DAVID J. HANSEN§ and WALTER J. BERRY§
tU.S. Environmental Protection Agency, Mid-continent Ecology Division, 6201 Congdon Boulevard, Duluth, Minnesota 55804
JHydroqual, Inc., One Lethbndge Plaza, Mahwah, New Jersey 07030, USA
Environmental Engineering Department, Manhattan College, Bronx, New York 10471, USA
§U.S Environmental Protection Agency, Atlantic Ecology Division, 27 Tarzwell Drive, Narragansett, Rhode Island 02882
(Received 3 March 1996; Accepted 18 June 1996)
Abstract—In developing sediment quality criteria (SQC) for metals, it is essential that bioavailability be a prime consideration.
Different studies have shown that while dry weight metal concentrations in sediments are not predictive of bioavailability, metal
concentrations in interstitial (pore) water are correlated with observed biological effects. A key partitioning phase controlling
cationic metal activity and toxicity in the sediment-interstitial water system is acid-volatile sulfide (AVS). Acid-volatile sulfide
binds, on a mole-to-mole basis, a number of cationic metals of environmental concern (cadmium, copper, nickel, lead, zinc) forming
insoluble sulfide complexes with minimal biological availability. Short-term (10-d) laboratory studies with a variety of marine and
freshwater benthic organisms have demonstrated that when AVS concentrations in spiked or field-collected sediments exceed those
of metals simultaneously extracted with the AVS, interstitial water metal concentrations remain below those predicted to cause
effects, and toxicity does not occur. Similar observations have been made in life-cycle laboratory toxicity tests with amphipods
and chironomids in marine and freshwater sediments spiked with cadmium and zinc, respectively In addition, field colonization
experiments, varying in length from several months- to more than 1 year, with cadmium- or zinc-spiked freshwater and marine
sediments, have demonstrated a lack of biological effects when there is sufficient AVS to limit interstitial water metal concentrations.
These studies on metal bioavailability and toxicity in sediments serve as the basis for proposed SQC for the metals cadmium,
copper, nickel, lead, and zinc Specifically, four approaches for denving criteria are described (a) comparison of molar AVS
concentrations to the summed molar concentration of the five metals simultaneously extracted with the AVS; (b) measurement of
interstitial water metal concentrations and calculation of summed interstitial water criteria toxic units (IWCTU) for the five metals,
based upon final chronic values from water quality criteria documents; (c) calculation of summed IWCTU based upon sediment
AVS concentrations and metal-specific partitioning of the metals to organic carbon, and (d) calculation of summed IWCTU based
upon partitioning of the metals to a minimum binding phase sorbent (chromatographic sand). For a number of reasons, SQC derived
using these approaches generally should be considered "no effect" values, i.e., with these techniques it is possible to predict when
sediment metals will not be toxic, but not necessarily when metal toxicity will be manifested. Currently, approaches (a) and (b)
are the most useful in terms of predicting metal bioavailability and deriving SQC. Further research is required, however, to fully
implement approaches (c) and (d). Additional research also is required to thoroughly understand processes controlling bioaccu-
mulation of metals from sediments by benthic organisms, as well as accumulation of metals by pelagic species that ingest metal-
contaminated benthos.
Keywords—Sediment Metal Criteria Programmic input Interstitial water Acid-volatile sulfide
BACKGROUND
Due to their widespread release and persistent nature, metals
such as silver, cadmium, copper, mercury, nickel, lead, and
zinc are commonly elevated in aquatic sediments. Thus, there
have been various proposals for denving sediment criteria or
standards for protecting benthic communities from metal tox-
icity. Many such attempts have featured measurement of total
sediment metals followed by comparison to background metal
concentrations, or in some cases an effects-based endpoint [1-
5]. An important limitation to these types of approaches is that
causality is difficult to establish because values are based on
total rather than bioavailable metal concentrations, i.e., for any
given total metal concentration, adverse toxicological effects
may or may not occur, depending upon physicochemical char-
acteristics of the sediment of concern [6-8]. The phenomenon
of differential bioavailability across sediment type also has
* To whom correspondence may be addressed.
This document has received an official EPA technical review; how-
ever, the content does not reflect official EPA policy. Mention of trade
names does not indicate endorsement by EPA or the federal govern-
ment
been noted for other classes of contaminants, including non
ionic organic chemicals such as pesticides [9]. The sediment
specific nature of contaminant bioavailability clearly repre-
sents a challenge for regulatory agencies, such as the U.S
Environmental Protection Agency (EPA), attempting to derive
technically defensible sediment quality criteria (SQC) with
broad applicability.
As a prelude to the development of national SQC by EPA,
a workshop was held in 1984, at which experts in the disci-
plines of environmental toxicology and chemistry met to re-
view possible approaches for dealing with the issue of differ-
ential contaminant bioavailability across sediments [10]. The
results of various studies concerning contaminant bioavail-
ability in sediments were presented, including one evaluating
the organochlorine pesticide Kepone®. An intriguing obser-
vation from this study was that, while Kepone toxicity was
not predictable based upon total sediment dry weight concen-
trations, effects of the pesticide were strongly correlated with
its interstitial (pore-)water concentrations. That is, irrespective
of sediment type, toxicity-response relationships based upon
pore-water concentrations were similar [11]. In the case of
.2056
-------
Sediment quality criteria for metals
Environ. Toxicol. Cfiem 15, 1996
2057
Kepone, it appeared that pore-water concentrations were con-
' trolled by partitioning of the organochlonne pesticide to or-
ganic carbon in the test sediments. This was an important
determinant in the recommendation that equilibrium partition-
ing (EqP) be pursued as an approach for deriving SQC [10].
Initially, EPA focused upon using the EqP approach to de-
velop and validate SQC for nomonic organic chemicals; the
technical basis of the paradigm currently used for denying
criteria for these types of contaminants is described in detail
by Di Toro et al. [9]. The two basic elements of the model are
(1) prediction of bioavailabihty based upon organic carbon
content of the sediment and the octanol-water partition co-
efficient (Kov) of the nonionic organic contaminant of concern,
and (2) comparison of the resultant organic carbon "normal-
ized" concentration to effects data from existing water quality
criteria (WQC) for the contaminant. The EqP approach for
predicting bioavailability of nonionic organics has received
extensive peer review, including two assessments by the EPA
Science Advisory Board, an independent panel of internation-
ally recognized scientists, who recommended the method as
the best currently available for deriving SQC [12,13]. To date,
EqP has been used as the basis for developing draft national
SQC for five nonionic pesticides and polycyclic aromatic hy-
drocarbons [14-18] and also has served as a basis for the
derivation of "Tier 2" sediment quality guidelines for a na-
tional sediment inventory [19].
More recently, EPA has evaluated the potential utility of the
EqP approach for deriving SQC for metals. Initial studies sup-
porting the concept were described by Swartz et al [20] and
Kemp and Swartz [21] who found that toxicity of cadmium
to amphipods in marine sediments could be accurately pre-
dicted based upon pore-water concentrations of the metal.
However, as opposed to the situation for nonionic organic
chemicals and organic carbon, those sediment partitioning
phases controlling interstitial water concentrations of metals
initially were not readily apparent.
The purpose of this paper is to present a proposed approach
for deriving SQC for the metals cadmium, copper, nickel, lead,
and zinc based upon a series of relatively recent studies ex-
amining the bioavailability and toxicity of metals in sediment.
As a prelude to describing the derivation of these SQC, we
first present a brief overview of the technical basis of the
approach.
PREDICTING METAL TOXICITY: SHORT-TERM STUDIES
Considerable research has focused upon elucidating sedi-
ment partitioning phases controlling metal bioavailability, of-
ten through the use of elaborate sequential extraction proce-
dures to identify physicochemical fractions with which metals
are associated [22,23]. Through these types of techniques, it
has been established that key binding phases for metals in
sediments include iron and manganese oxides and organic car-
bon. However, an important shortcoming with these approach-
es was that much of the work was done with sediments that
had intentionally, or unintentionally, been oxidized through
procedures such as drying. Thus, the techniques were appro-
priate only for examining metal bioavailability in oxidized
sediments. This resulted in an underestimate of the importance
of metal-sulfide binding in the anaerobic horizons character-
istic of most natural in-place sediments. A number of catiomc
metals form stable complexes with sulfide generated in sedi-
ments by sulfate-reducing bactena [24-28]. From a lexico-
logical standpoint this could be a particularly critical reaction
because predictions based upon chemical equilibria [27] sug-
gest that a number of metals of environmental concern (e.g.,
cadmium, copper, mercury, nickel, lead, silver, zinc) form rel-
atively insoluble sulfides that should not be present in pore
water, and hence, biologically unavailable.
Di Toro et al. [8] investigated the significance of sulfide
partitioning in controlling metal bioavailability in marine sed-
iments spiked with cadmium. In those experiments, the op-
erational definition of Cornwell and Morse [29] was used to
identify that fraction of amorphous sulfide, commonly termed
acid-volatile sulfide (AVS), available to interact with cadmium
in the sediments. Specifically, the AVS was defined as the
sulfide liberated from wet sediment by treatment with 1 N HC1
acid. Di Toro et al. [8] found that when the molar concentration
of AVS in the test sediments was larger than that of the cad-
mium (i.e., when the cadmium-to-AVS ratio was less than 1,
or the cadmium-to-AVS difference was less than 0), interstitial
water concentrations of the metal were small and no toxicity
was observed in 10-d tests with the amphipods Rhepoxynius
hudsoni or Ampelisca abdita. Studies by Carlson et al. [30]
with cadmium-spiked freshwater sediments yielded similar re-
sults; when there was more AVS than metal, significant toxicity
was not observed in 10-d tests with oligochaetes (Lumbriculus
variegatus) or snails (Helisoma sp.). Based upon these initial
studies, Di Toro et al. [31] suggested that it may be feasible
to derive metal SQC by direct comparison of molar AVS con-
centrations to the molar sum of the concentrations of catiomc
metals (specifically, cadmium, copper, nickel, lead, zinc) ex-
tracted with the AVS (summed simultaneously extracted met-
als; XSEM).
Casas and Crecelius [32] further explored this possibility
by conducting 10-d toxicity' tests with the marine polychaete
Capitella capitata exposed to sediments spiked with zinc, lead,
or copper. As was true in earlier studies, elevated pore-water
metal concentrations and toxicity were observed only when
SEM concentrations exceeded those of AVS. Green et al. [33]
reported results of another spiking experiment supporting the
general EqP approach to deriving SQC for metals. In their
study metal-sulfide partitioning was not directly evaluated, but
it was found that toxicity of cadmium-spiked marine sediments
to the meiobenthic copepod Amphiascus tenuiremis was pre-
dictable based upon interstitial water (but not sediment dry
weight) cadmium concentrations. Further spiking experiments
by Pesch et al. [34] demonstrated that 10-d survival of the
marine polychaete Neanthes arenceodentata was comparable
to controls in cadmium- or nickel-spiked sediments with more
AVS than SEM. An important observation in their study was
that significant mortality did not always occur in sediments
with more SEM than AVS. This appeared to be related, in
part, to the ability of the polychaete to sense elevated metal
concentrations and avoid burrowing into the test sediments,
thereby limiting their exposure to metals in the interstitial
water.
Berry et al. [35] described experiments in which A. abdita
was exposed for 10 d to sediments spiked either singly, or in
combination, with cadmium, copper, nickel, lead and zinc. As
in previous studies, significant toxicity to the amphipod did
not occur when AVS concentrations exceeded those of SEM.
Berry et al. [35] also analyzed their data by comparing ob-
served mortality to interstitial water metal concentrations ex-
pressed as toxic units (IWTU):
IWTU = [Md]/LC50
(D
-------
2058 Environ Toxicol. Chem 15, 1996
G.T. Ankley et al.
where [Md] is the dissolved metal concentration in the inter-
stitial water, and the LC50 is the concentration of the metal
causing 50% mortality of the test species in a water-only test.
If pore-water exposure in a sediment test is indeed equivalent
to that in a water-only test, then 1IWTU should result in 50%
mortality of the test animals. Berry et al. [35] reported that
significant (>24%) mortality occurred in only 2.6% of sedi-
ments with less than 0.5 IWTU, while samples with greater
than 0.5 IWTU were toxic 91.9% of the time. Berry et al. [35]
also made an important observation relative to pore-water met-
al chemistry in their mixed-metals test. Chemical equilibrium
calculations suggest that the relative affinity of metals for AVS
should be copper > lead > cadmium > zinc > nickel [27,31].
Hence, the appearance of the metals in pore water as AVS is
"exhausted" should occur in an inverse order, e.g., zinc would
replace nickel in a monosulfide complex and nickel would be
liberated to the pore water, etc. This trend was, in fact, ob-
served by Berry et al [35].
In addition to short-term laboratory experiments with spiked
sediments, there have been a number of laboratory toxicity
tests with metal-contaminated sediments from the field. Ankley
et al. [36] exposed L variegatus and the amphipod Hyalella
azteca to 17 sediment samples from a freshwater/estuarine site
along a gradient of cadmium and nickel contamination. In 10-d
toxicity tests, H. azteca mortality was greater than controls
only in sediments with more SEM (cadmium plus nickel) than
AVS. Lumbriculus variegatus was far less sensitive to the
sediments than H. azteca, which correlates with the differential
sensitivity of the two species in water-only tests with cadmium
and nickel. Ankley et al. [37] examined the significance of
AVS as a binding phase for copper in freshwater sediments
from two copper-impacted sites. Based upon pore-water copper
concentrations in the test sediments, the 10-d LC50 for H.
azteca was 31 u,g/L; this corresponded very well with a mea-
sured LC50 of 28 n-g/L for amphipod in a 10-d water-only
test. However, Ankley et al. [37] also found that survival of
H. azteca was comparable to control values in several sedi-
ments with markedly more SEM than AVS suggesting that, in
these samples, SEM in excess of AVS was not biologically
available. Measurement of pore-water concentrations in the
samples corroborated this lack of bioavailabihty; when sur-
vival was comparable to controls, pore-water copper was non-
detectable. This observation suggested the presence of binding
phases in addition to AVS for copper in the test sediments.
Recent studies suggest that an important source of the extra
binding capacity in these sediments was organic carbon [38].
Hansen et al. [39] summarized the results of 10-d laboratory
toxicity tests with amphipods, oligochaetes, and polychaetes
using metal-contaminated field sediments from five different
marine sites and four freshwater sites, including the three de-
scribed above. They compared the toxicity data both to SEM
(summed for mixed metal sites): AVS relationships and pore-
water metal concentrations (expressed as IWTU). In the 49
sediments evaluated where metals were the likely cause of
toxicity, 100% with less SEM than AVS and less than 0.5
IWTU did not exhibit significant toxicity to the test species.
Prediction of the occurrence of metal toxicity was less certain
than prediction of an absence of toxicity, 66.7% of the 45
samples with more SEM than AVS and more than 0.5 IWTU
were toxic.
PREDICTING METAL TOXICITY: LONG-TERM STUDIES
Taken as a whole, the short-term laboratory experiments
with metal-spiked and field-collected sediments present a
strong argument for the ability to predict an absence of metal
toxicity based upon sediment SEM-to-AVS relationships and/
or interstitial water metal concentrations. However, for this
approach to serve as a valid basis for SQC derivation, com-
parable predictive success must be demonstrated in long-term
laboratory and field experiments where chronic effects could
be manifested [40,41]. This demonstration was the goal of
experiments described by Hare et al. [42], DeWitt et al. [43],
Hansen et al. [44], Liber et al. [45], and Sibley et al. [46]. An
important experimental modification to these long-term stud-
ies, as opposed to the short-term tests described above, was
the collection of horizon-specific chemistry data. This is re-
quired because AVS concentrations often vary inversely with
depth [47,48]; hence, chemistry performed on homogenized
samples might not reflect the true exposure of benthic organ-
isms dwelling in surficial sediments [40,42,49].
DeWitt et al. [43] conducted a life-cycle test with the marine
amphipod Leptocheirus plumulosus exposed to cadmium-
spiked sediments for 28 d. There were no significant effects
on survival, growth, or reproduction in sediments containing
more AVS than cadmium, in spite of the fact that these samples
contained up to 363 mg cadmium/kg on a dry weight basis.
Sibley et al. [46] reported similar results from a life-cycle
(56-d) test conducted with the freshwater midge Chironomus
tentans exposed to zinc-spiked sediments. In that experiment,
significant effects on survival, growth, emergence, and repro-
duction were only observed in sediments with SEM in excess
of AVS. Performance of the midge was comparable to controls
in sediments with more AVS than SEM, at dry weight zinc
concentrations as high as 270 mg/kg.
Hansen et al. [44] conducted a 118-d colonization experi-
ment in which cadmium-spiked sediments were held in the
laboratory in a constant flow of raw seawater. Qualitative and
quantitative analysis of colonization of the test sediments re-
vealed no discemable difference from controls when surficial
AVS concentrations were greater than those of SEM. In those
instances where SEM was greater than AVS, pore-water cad-
mium concentrations were elevated, and significant alterations
in the types of benthic species present, and their richness and
abundance were observed. Among-treatment differences in
benthic community structure based on interstitial water con-
centrations of cadmium were consistent with known sensitiv-
ities of the species in water-only tests with cadmium.
For approx. 1 year Hare et al. [42] conducted field colo-
nization experiment in which uncontaminated freshwater sed-
iments were spiked with cadmium and replaced in the oligo-
trophic lake from which they originally had been collected.
They reported little or no effects on abundance or biomass of
major benthic taxa in any of the treatments. However, they
did report a significant accumulation of cadmium by organisms
from sediments with surficial SEM concentrations greater than
those of AVS. These sediments also contained elevated con-
centrations of cadmium in interstitial water. Liber et al. [45]
reported the results of colonization experiment conducted us-
ing a design similar to that of Hare et al. [42]. Sediments from
a freshwater mesotrophic pond were spiked with zinc, replaced
in the field, and sampled over 16 months for collection of
biological and analytical data. With the exception of the high-
est spiking concentration (ca. 700 mg/kg, dry weight), AVS
concentrations remained larger than those of SEM, and pore-
water zinc concentrations were generally nondetectable. The
only observed difference in benthic community structure
-------
Sediment quality criteria for metals
Environ. Toxicol Cheat. 15, 1996 2059
across the treatments was a slight decrease in the abundance
of Naididae oligochaetes at the highest spiking concentration.
CONSIDERATIONS IN PREDICTING METAL TOXICTTY
In summary, results of the long-term laboratory and field
experiments conducted to date represent convincing support
of the conclusions reached in the initial short-term laboratory
experiments, i.e., an absence (but not necessarily a presence)
of metal toxicity can be reliably predicted based upon metal-
sulfide relationships and/or pore-water metal concentrations.
However, to use these observations as a basis for predicting
metal bioavailability or deriving SQC, a number of conceptual
and practical issues, related to analytical measurements, sam-
pling, additional binding phases, and metal "activity" must
be considered. Many of these were addressed by Ankley et al.
[50]; the most salient to our proposed derivation of SQC are
described below.
Analytical measurements
An important aspect to deriving "global" SQC values is that
the methods necessary to implement the approach be reasonably
standard. From the standpoint of the metal SQC proposed below,
a significant amount of research has gone into defining the
extraction of SEM and AVS, the measurement of AVS, and the
extraction of pore water. The basic SEM/AVS extraction method
recommended by EPA is that of Allen et al. [51]. In terms of
AVS measurement, a number of techniques nave been success-
fully utilized including gravimetric approaches [8,52], colon-
metric assays [29], gas chromatography-photoionization detec-
tion [32,53], and specific ion electrodes [54-57]. Bufflap and
Allen [58] recently reviewed approaches for isolating interstitial
water for metal analysis. They concluded that centrirugation
under a nitrogen atmosphere was the simplest technique likely
to result in an unbiased estimate of metal concentrations in pore
water. At present, the EPA recommends filtration of the isolated
pore water through 0.4 to 0.45-u, polycarbonate niters to better
define that fraction of aqueous metal associated with toxicity
[59]. An alternative approach to sampling pore water for de-
termination of dissolved metals that has proven to be quite
successful is in situ dialysis [35,39,42,45].
Sampling
As alluded to above, accurate prediction of exposure of ben-
thic organisms to metals is critically dependent upon sampling
appropriate sediment horizons at appropriate times. This is be-
cause of the relatively labile nature of AVS due to oxidation.
In fact it is this seemingly labile nature that has led some to
question the practical utility of using AVS as a basis for EqP-
derived metal SQC [40,41]. For example, there have been many
observations of spatial (depth) variations in AVS concentrations,
most of which indicate that surficial concentrations are smaller
than those in deeper sediments [42,44,45,47,48,54,57,60 ]. This
likely is due to oxidation of AVS at the sediment surface, a
process that is enhanced by bioturbation [49]. In addition to
varying with depth, AVS can vary seasonally. For example, in
systems where overlying water contains appreciable oxygen dur-
ing cold weather months, AVS tends to decrease, presumably
due to a constant rate of oxidation of the AVS linked to a
decrease in its generation by sulfate-reducing bacteria
[47,52,61]. Based on these studies, it appears the way to avoid
possible underestimation of metal bioavailability is to sample
the biologically "active" zone of sediments (e.g., 0-2 cm) at
times when AVS might be expected to be present at small con-
centrations (e.g., during the winter in aerobic systems).
The somewhat subjective aspects of these sampling rec-
ommendations have been of concern to us. However, recent
research suggests that the transient nature of AVS may be
overstated relative to predicting the fate of all metal-sulfide
complexes in aquatic sediments. Observations from the Duluth
EPA laboratory made in the early 1990s indicated that AVS
concentrations in sediments contaminated by metals such as
cadmium and zinc tended to be elevated over concentrations
typically expected in freshwater systems (G.T. Ankley, un-
published data). The probable underlying basis for these ob-
servations did not become apparent, however, until a recent
series of spiking and metal-sulfide stability experiments. The
field colonization study of Liber et al. [45] demonstrated a
strong positive correlation between the amount of zinc added
to test sediments and the resultant concentration of AVS in
the samples. In fact, the initial design of their study attempted
to produce test sediments with as much as five times more
SEM (zinc) than AVS; however, the highest SEM-to-AVS ratio
achieved was only slightly larger than 1. Moreover, the ex-
pected surficial depletion and seasonal variations in AVS were
unexpectedly low in the zinc-spiked sediments [45]. These
observations suggested that zinc sulfide, which comprised the
bulk of AVS in the spiked sediments, was more stable than
the iron sulfide that presumably was the source of most of the
AVS in the control sediments. The apparent stability of both
zinc sulfide and cadmium sulfide versus iron sulfide also has
been noted in laboratory spiking experiments with freshwater
sediments [46,49,62].
In support of these observations, recent metal-sulfide oxi-
dation experiments conducted by Di Toro et al. [63] have
confirmed that cadmium and zinc form more stable sulfide
solid phases than iron. If this is also true for sulfide complexes
of copper, nickel, and lead, the issue of seasonal/spatial vari-
ations in AVS becomes of less concern because most of the
studies evaluating variations in AVS have focused on iron
sulfide (i.e., uncontaminated sediments). Thus, further research
concerning the differential stability of metal-sulfides, both
from a temporal and spatial perspective, is definitely war-
ranted.
Additional binding phases
Although AVS is an important binding phase for metals,
clearly other physicochemical factors influence metal parti-
tioning in sediments. In aerobic systems, or those with low
productivity (i.e., where the absence of organic carbon limits
sulfate reduction) AVS plays little or no role in determining
interstitial water concentrations of metals. For example, Leon-
ard et al. [48] found that a relatively large percentage of sur-
ficial sediments from open areas in Lake Michigan did not
contain detectable AVS. In fact the great majority (42 of 46)
of samples analyzed by Leonard et al. [48] contained less AVS
than SEM, yet pore-water metal concentrations of cadmium,
copper, nickel, lead, and zinc were consistently small or non-
detectable. Even in sediments where concentrations of AVS
are significant, other partitioning phases may provide addi-
tional binding capacity for SEM [37,53,64]. In anaerobic sed-
iments, organic carbon appears to be an additional binding
phase controlling metal partitioning, in particular for cadmium,
copper, and lead [38], while in aerobic sediments both organic
carbon and iron and manganese oxides undoubtedly play a role
in controlling interstitial water concentrations of metals
-------
2060 Environ Toxicol Chem 15, 1996
G T. Ankley et »"•
[22,23,64,65]. Even in substrates with very little metal binding
capacity (e.g., chromatographic sand), surface adsorption as-
sociated with cation exchange capacity will control pore-water
metal concentrations to some degree [66] Although an ideal
SQC model for metals would incorporate all possible metal
binding phases, current knowledge concerning partitioning/
capacity of phases other than AVS is insufficient for practical
application of this model. However, in the proposed metal SQC
presented below, we describe two approaches through which
phases other than AVS could be incorporated into criteria der-
ivation.
Metal activity
A substantial number of water-only experiments suggest that
biological effects often are correlated to the divalent metal
activity, {M2+}. Metal activity is the molar divalent metal
concentration [M2*] corrected for the shielding effect of anions
that are electrostatically attracted to the atoms in high ionic
strength (e.g., concentrated) solutions. The law of mass action
for dissolved chemical species is expressed in terms of metal
activities rather than metal concentrations. This procedure cor-
rects for ionic strength effects. The claim is not that the only
bioavailable form is M2* (for example MOH+ may also be
bioavailable) but that dissolved organic carbon and certain
other ligand-complexed fractions generally are of less impor-
tance in terms of toxicological significance.
There are a number of examples of correlation of biological
responses with metal activity. The acute toxicity of cadmium
to grass shrimp was determined at various concentrations of
chloride and nitriloacetic acid (NTA), both of which form
cadmium complexes [67]. In that study, the response was quite
different at different concentrations of chloride, indexed by
salinity, and NTA. However, when the concentration/response
relationships were evaluated with respect to Cd2* activity in
the solution, the response curves collapsed into a single re-
lationship. Comparable results have been reported for exper-
iments with Gonyaulax tamarensis exposed to varying con-
centrations of copper and the inorganic ligand EDTA; in these
studies, concentration/response relationships correlated with
Cu2+ activity [68]. Chronic toxicity of zinc, with phytoplank-
ton growth as the endpoint, also has been examined [69]; in
this case as NTA was added, the toxicity of zinc to Microcystis
decreased. However, the toxicity data were correlated with free
zinc activity. Similar results for diatoms exposed to copper
and the complexing ligand Tris also have been found [70].
Variations in Tris concentrations and pH produced markedly
different growth rates, which all could be related to the Cu2*
activity. A similar set of results with copper and algae were
described by Sunda and Lewis [71] with dissolved organic
carbon from river water as the complexing ligand. Bioavail-
ability, as a function of free metal activity, also has been ex-
amined from the standpoint of metal accumulation. For ex-
ample, uptake of copper by oysters was correlated, not to total
copper concentration, but to Cu2+ activity [72].
There are, however, examples of where metal toxicity seem-
ingly is not well correlated with metal activity, suggesting that
other metal species or variables also can be important [73].
Nevertheless, even in these instances, factors that reduce metal
activity limit toxicity. What can be drawn from this is that any
approach to denving no effect SQC for metals should focus
on minimization of metal activity. In doing so, there is no
implication that M2+ is the only form of metal bioavailable,
only that if metal activity is low, no toxicity should occur.
PROPOSED METAL SQC
Based upon the technical information summarized above,
we propose the following approach to denving metal SQC
These values are intended to be the best estimate of sediment
metal concentrations that should not cause toxicity to benthic
organisms. The SQC are for five metals: cadmium, copper,
nickel, lead, and zinc. They are derived from EqP-based es-
timates of metal concentrations associated with a lack of ad-
verse biological effects. This approach has been presented to
and reviewed by the Science Advisory Board of EPA [38,74].
The SQC for all five metals collectively can be derived using
four procedures: (a) comparing the sum of their molar con
centrations to the molar concentration of AVS in sediments
(AVS criteria); (b) comparing the measured interstitial water
concentrations of the metals to WQC final chronic values
(FCVs) (interstitial water criteria); (c) using organic carbon
based partition coefficients,, in addition to the AVS and SEM
relationships, to compute interstitial water concentrations, fol
lowed by comparison to FCVs (AVS and organic carbon en
teria); or (d) using minimum partition coefficients (e.g., gen
crated from chromatographic sand [66]) to compute sediment
concentrations that would not result in interstitial water ex
ceeding metal FCVs (minimum partitioning cnteria). These
approaches are described in more detail below. When an SQC
is exceeded, based upon any one of the four procedures, metal
toxicity should not occur. Failing all of the approaches is in-
dicative of a potential problem that would entail further eval-
uation. At present, we believe that the technical basis for im-
plementing approaches (a) and (b) is supportable. As discussed
below, however, additional research is required to implement
procedures (c) and (d).
The following nomenclature is used in subsequent discus
sion of SQC derivation for metals. The SQC for the metals
are expressed in molar units because of the molar stoichi
ometry of metal binding to AVS. Thus, solid-phase constitc
ents (AVS, SEM) are in u.mol/g dry weight. The interstitial
water metal concentrations are expressed in u,mol/L, either as
dissolved concentrations [MJ or activities {M2*} [75]. The
partition coefficients are in L pore water/g dry weight, con-
sistent with the above solid- and aqueous-phase concentration
units. The subscripted notation, Md, is used to distinguish dis
solved aqueous-phase molar concentrations from solid-phas;-,
molar concentrations with no subscript. For the combined con
centration, [SEMT], the units are moles of metal per volum-->
of solid plus liquid phase (i.e., bulk).
One final point should be made with respect to nomencls
ture. When we use the terms nontoxic or having no effect, w;
mean only with respect to the five metals considered in this
paper. The toxicity of field-collected sediments can be caused
by other chemicals. Therefore, avoiding exceeding the SQC
for metals does not mean that the sediments are nontoxic It
only ensures that the five metals being considered should not
have an undesirable biological effect. Moreover, as discussed
in detail below, exceeding the criteria for the five metals does
not necessarily indicate that metals will cause toxicity. For
these reasons, we strongly recommend toxicity tests as an
integral part of any assessment concerned with the effects of
sediment-associated contaminants [50].
SINGLE METAL SEDIMENT QUALITY CRITERIA
Single metal criteria are not usually applicable to field sit-
uations because there is almost always more than one metal
to be considered. In any case, as will become subsequently
-------
Sediment quality criteria for metals
Environ. Toxicol Chem. 15, 1996 2061
clear, it would be technically indefensible to derive criteria for
one metal at a time because of the competitive nature of AVS
binding. Nevertheless, it is illustrative to present the logic for
single metals as a prelude to the derivation of the multiple
metal criteria.
AVS criteria
It has been demonstrated that if the SEM of a sediment is
less than or equal to the AVS
[SEM] £ [AVS] (2)
then no toxic effects are seen. This is consistent with the results
of a chemical equilibrium model for the sediment-interstitial
water system [31]. The resulting metal activity {M2*} can be
related to the total SEM of the sediment and water, and to the
solubility products of the metal sulfide (KMS) and iron sulfide
). In particular it is true that at [SEM] =s [AVS] then
[SEMT]
(3)
Because the ratio of metal-sulfide to iron-sulfide solubility
products (KMS/KFCS) is very small (<10~5) even for the most
soluble of the sulfides, the metal activity of the sediment is at
least five orders of magnitude smaller than the SEM (see Di
Toro et al. [31] for data sources and references). This indicates
that no biological effects would be seen if this sediment were
tested. Therefore the condition [SEM] s [AVS] is a no effect
SQC.
The reason we use the term "no effect" is that for the
condition [SEM] ^ [AVS] no biological impacts are expected.
However, for [SEM] > [AVS], which might seemingly be
considered an SQC violation, there are many documented in-
stances where no biological impacts occur, e.g., because or-
ganic carbon partitioning controls metal bioavailability, or spe-
cies of concern are insensitive to metals.
Again, because of potential temporal and spatial variability
of AVS, we recommend that AVS and SEM measurements be
made using surficial sediments during periods when AVS con-
centrations are smallest. Note, that this will not always be
during cool-weather seasons; for example, systems that be-
come anaerobic during the winter can maintain relatively large
sediment AVS concentrations [45].
Interstitial water criteria
The condition [SEM] s [AVS] indicates that the metal ac-
tivity of the sediment-interstitial water system is low and,
therefore, below lexicologically significant concentrations.
Another way of ensuring this is to place a condition on the
interstitial water activity directly. Suppose that we knew the
metal activity, denoted by {FCV}, that corresponded to the
[FCV]. Then the SQC corresponding to this effect level is
{FCV}
(4)
It is quite difficult, however, to measure and/or calculate metal
activity in a solution phase at the low concentrations required
because it depends on the identities, concentrations, and ther-
modynamic affinities of other chemically reactive species that
are present. Also the WQC are not expressed on an activity
basis. An approximation to this condition is
[Md] < [FCVJ
(5)
where [FCVJ is the FCV applied to total dissolved concen-
trations. That is, we require that the total dissolved metal con-
centration in the pore water [MJ be less than the FCV applied
as a dissolved criterion. Although this requirement ignores the
effect of chemical speciation on both sides of the equation—
compare Equations 4 and 5—it is the approximation that is
currently being suggested by EPA for the WQC for metals
[59]. That is, the WQC should be applied to the total dis-
solved—rather than the total acid recoverable—metal concen-
tration. Hence, if this second condition is satisfied it is con-
sistent with the level of protection afforded by the WQC.
In situations where the SEM exceeds the AVS ([SEM] >
[AVS]), but the interstitial water total dissolved metal is less
than the final chronic value ([MJ £ [FCVJ), this sediment
would not violate the catena. These cases occur when sig-
nificant binding to other phases occurs. It should be noted that
using the FCV for metals in freshwater samples requires that
the hardness of the interstitial water be measured because the
WQC vary with hardness.
AVS and organic carbon criteria
For sediments with an appreciable AVS concentration rel-
ative to SEM, the SQC requirement that [SEM] s [AVS] is
a useful comparison. However, if the AVS concentration is
small, then this comparison is of little value. Similarly, as
described above, even in situations where significant AVS oc-
curs in sediments, other sorption phases may limit the metal
activity when SEM exceeds the AVS.
One well-established example of an important metal-binding
phase in sediment is organic carbon. It has been demonstrated
that a relationship exists between the SEM that is in excess
of the AVS and the interstitial water metal activity {M2*}
[SEM] - [AVS] =
(6)
where KLOC is the partition coefficient between organic carbon
and interstitial water of the sediment on a metal activity basis,
and/oc is the weight fraction of organic carbon of the sediment
[76]. If it is required that the pore-water metal activity be at
the FCV, then the SQC for SEM in the single metal case would
be
[AVS] + AT..oc/oc{FCV}
(7)
If the activity is replaced with the total dissolved FCV, the
resultant SQC is
= [AVS] +
(8)
where AT4oc is the partition coefficient between organic carbon
and interstitial water on the basis of dissolved metal. Note that
the metal-specific organic carbon-based partition coefficients
vary with respect to pH [76], so pH of the interstitial water
must be determined. In addition, because the FCV for the five
metals is hardness dependent in freshwater, an estimate or
measurement of pore-water hardness is required.
This is the third approach from which an SQC can be de-
rived. For sediments where organic carbon provides all binding
capacity in addition to AVS, it is interpreted the same as SQC
for nomomc organics [9]. That is, exceeding the criterion
would imply that unacceptable biological impacts would occur.
Because the analysis of sediment binding data and the esti-
mation of the K^ attributes all the binding to organic carbon,
using these constants would imply that this criterion is the
boundary between no effect and effect. It is likely, however,
that the assumption that organic carbon is the only important
phase in addition to AVS often will not be correct. In these
cases, it becomes a no effect criterion. Of course, using this
-------
2062 Environ. Toxicol. Chem. 15, 1996
G.T. Ankley et al.
as an effect criterion assumes that applying the FCV on a total
dissolved basis is appropriate. If, in fact, a significant fraction
of the interstitial water metal is not bioavailable, then again,
this criterion would be a no effect value.
Minimum partitioning criteria
It would be useful to have solid-phase criteria that would
effectively screen sediments for which metal concentrations
are low enough so that no problem is anticipated. In developing
this approach we examined substrates for which the partition
coefficients were likely to be quite low, such as clean chrc-
matographic sand [66]. From these experiments it was possible
to establish minimum partition coefficients (K^^ which could
be applied to any sediment. Based on this, no effect SQC would
be:
^[FCVJ
(9)
Unlike Equation 8, no AVS term is included because it is to
be applied only to sediments having no detectable AVS. If
AVS is quantifiable, then Equation 2 •would apply. In deriving
Kcunta values appropriate to a given sediment, measurement or
estimation of interstitial water pH is required, as metal binding
is highly pH dependent [66]. Also required is pore- water hard-
ness for appropriate adjustment of FCVs for freshwater sam-
ples.
Multiple metals criteria
As described in the previous section, from a practical stand-
point it is insufficient and inappropriate to consider each metal
separately. Because of the interactive nature of metal-sulfide
binding, this is of particular concern for the AVS criteria.
AVS criteria
The results of calculations using chemical equilibrium mod-
els indicate that metals act in a competitive manner when
binding to AVS. The five metals copper, lead, cadmium, zinc,
and nickel will bind to AVS and be converted to their re-
spective sulfides in this sequence, i.e., in the order of increasing
solubility. Therefore, they must be considered together. There
cannot be a criterion for just nickel, for example, because all
the other metals may be present as metal sulfides and, there-
fore, to some extent as AVS. If these other metals are not
measured as SEM, then the 2SEM will be misleadingly small,
and it may appear that [XSEM] < [AVS] when in fact this
would not be true if all the metals are considered together. It
should be noted that we currently restrict this discussion to
the five metals listed above; however, in situations where other
sulfide-forming metals (e.g., mercury, silver) are present at
high concentrations, they also must be considered.
The equilibrium model prediction of the metal activity is
similar to the single metal example when a mixture of the
metals is present. If the molar sum of SEM for the five metals
is less than or equal to the AVS, i.e.,
then
, [SEMJ s [AVS]
[SEM,T]
(10)
(ID
where [SEM,,T] is the total SEM (txmol/L (bulk)) for the ith
metal. Thus the activity of each metal, {M,}, is unaffected by
the presence of the other sulfides. This can be understood as
follows. Suppose that the chemical system starts initially as
iron and metal sulfide solids and that the system proceeds to
equilibrium by each solid dissolving to some extent. The iron
sulfide dissolves until the solubility product of iron sulfide is
satisfied. This sets the sulfide activity. Then each metal sulfide
dissolves until reaching its solubility. Because so little of each
dissolve relative to the iron sulfide, the interstitial water chem-
istry is not appreciably changed. Hence, the sulfide activity
remains the same and the metal activity adjusts'to meet each
solubility requirement. Therefore, each metal sulfide behaves
independently of one another. The fact that they are only slight-
ly soluble relative to iron sulfide is the cause of this behavior.
Thus, the AVS criteria is easily extended to the case of multiple
metals.
Interstitial water criteria
The application of the interstitial water criteria to multiple
metals is complicated, not by the chemical interactions of the
metals in the sediment-interstitial water system (as in the case
with the AVS criteria) but rather because of their possible toxic
interactions. Even if the individual concentrations do not ex-
ceed the FCV of each metal (FCV,), the metals could exert
additive effects that might result in toxicity [77-80]. Therefore,
to address this potential additivity, the interstitial water metal
concentrations are converted to toxic units (TUs) and these
are summed. Because FCVs are used as the effects concen-
trations, these TUs are referred to as interstitial water criteria
toxic units (IWCTUs). For freshwater sediments, the FCVs are
hardness dependent for all of the metals under consideration
and, thus, need to be adjusted to the hardness of the pore water
of the sediment being considered. For the ith metal with a total
dissolved concentration [My], the IWCTU is
Failure to exceed this SQC requires that the sum ,of the
IWCTUs be less than or equal to one
(13)
Hence, the multiple metals criterion is quite similar to the
single metal case (Eqn. 5) except that it is expressed as summed
TUs.
AVS and organic carbon criteria
The case in which the sediment organic carbon is considered
in addition to AVS is more complicated. Consider, first, a single
metal. As discussed above, the relationship between the in-
terstitial water concentration and sediment concentration for
the ith metal is given by the equation
[SEMJ = [AVS] + Kd.ocj/ocIHJ (14)
where K^oty is the metal-specific partition coefficient between
sediment organic carbon and interstitial water, and [MJ is the
total dissolved interstitial water metal concentration. For this
case, where the interstitial water concentration is predicted
using the SEM in excess of the AVS, and the partition coef-
ficient between the excess SEM and the interstitial water
[SEMJ - [AVS] _
*d.OC.i/OC
(15)
In order to apply this equation to the case of multiple metals,
it is first necessary to identify and quantify the metals that are
not entirely present as metal sulfides. The best way to do this
-------
Sediment quality criteria for metals
Environ Toxicol. Chem 15, 1996 2063
is to establish which metals are present as the metal sulfides
and in what quantity. The procedure is to assign the AVS to
the metals in the sequence of their solubility products from
the lowest to the highest: SEMo,, SEMn,, SEMo,, SEMz,,, and
SEMN>. That is, the AVS-complexed metals would be copper,
followed by lead, followed by cadmium, etc., until the AVS
is exhausted. The remaining SEM is that amount present in
excess of the AVS.
To be specific, let A[SEMJ be the excess SEM for each of
the ith metals. The least soluble metal sulfide (of the five metals
being considered in this analysis) is copper sulfide. Thus if
the simultaneously extracted copper is less than the AVS
([SEMcJ < [AVS]), then essentially all of it must be present
as copper sulfide with no additional SEMo, present, such that
AfSEMcJ = 0. The remaining AVS is A[AVS] = [AVS] -
[SEMcJ.
This computation is repeated next for lead because lead
sulfide is the next least soluble sulfide. Suppose, unlike copper,
the simultaneously extracted lead is not less than the remaining
AVS ([SEMpJ > A[AVS]). Hence, only a portion of the si-
multaneously extracted lead is present as lead sulfide and the
remaining SEM, which is denoted as AtSEMpJ, is the differ-
ence between the remaining AVS, A[AVS] and the simulta-
neously extracted lead: AfSEMpJ = [SEMpJ - A[AVS]. Thus,
a portion of the lead is present as lead sulfide, and the re-
mainder is excess SEM. Because the AVS has been exhausted
by the lead in this example, the remaining three metals would
all be present as excess SEM such that: A[SEM]Cd = [SEMoJ;
ArSEMlz, = [SEMzJ; and A[SEM]Nl = [SEMNl].
For each of these metals, interstitial water concentrations
can be determined from the appropriate partition coefficients
[Mud]
A [SEM,]
(16)
This equation is analogous to Equation 14 for the single metal
example. Note that if A[SEMJ = 0, then so is the interstitial
water metal concentration. The IWCTUs are computed using
this equation for the interstitial water concentrations
A [SEM,]
[FCV,J
(17)
where Equation IS is used to compute the interstitial water
concentrations. Note that the organic carbon-based partition
coefficients vary with respect to pH, so pH of the interstitial
water must be determined, together with the hardness if nec-
essary.
If this SQC is not exceeded, the computed total IWCTU
concentration must be less or equal to one
A [SEM,]
1
(18)
Thus, this criterion is the IWCTU value, Equation 13, with
the interstitial water concentrations calculated from the excess
SEM for each metal and the appropriate organic carbon par-
tition coefficients.
Minimum partitioning criteria
The no-effect criterion using the minimum partition coef-
ficients (ATdjmnJ) is analogous to that using the organic carbon-
based coefficients
[SEM.]
Again, minimum binding phase coefficients vary with pH, so
pH of the pore water must be determined, along with hardness
for adjustment of freshwater FCVs. Because the minimum
partition coefficients are being used, this criterion corresponds
to the upper bound estimate of the IWCTUs.
To summarize, the proposed SQC are as follows. The sed-
iment passes the SQC if any one of these conditions is satisfied:
(a) AVS criteria
[SEMJ =£ [AVS]
(b) Interstitial water criteria
. ITCV J
(c) AVS and organic carbon criteria
v A[SEMJ
. *d.oc,/oc[FCV,d]
(d) Minimum partitioning criteria
v tSEM,] ^
(10)
(13)
(18)
(19)
(19)
If any one of these conditions is violated, this does not mean
that the sediment is toxic. For example, if the AVS in a sed-
iment is nondetectable, then'condition (a) will be violated.
However, if there is sufficient organic carbon sorption so that
either condition (b) or (c) is satisfied, then the sediment would
be deemed acceptable.
If all of these conditions are violated, then,there is reason
to believe that the sediment may be unacceptably contaminated
by these metals. Further testing and evaluations would there-
fore be useful in order to assess actual toxicity and its causal
relationship to the five metals. These may include acute and
chronic tests with species that are sensitive to the metals sus-
pected to be causing the toxicity. Also, in situ community
assessments, sediment toxicity identification evaluations, and
seasonal characterizations of the SEM, AVS, and interstitial
water concentrations would be appropriate [50]'.
SUMMARY AND RECOMMENDATIONS
This paper summarizes the technical basis for predicting the
bioavailability of metals in sediment and a proposal for es-
tablishing SQC for copper, cadmium, nickel, lead, and zinc.
The basis of the overall approach is the use of EqP theory
linked to the concept of maintaining metal activity in the sed-
iment-interstitial water system below effects levels. Extensive
lexicological data from short-term and long-term laboratory
and field experiments, with both marine and freshwater sed-
iments, and a variety of species indicates that it is possible to
predict reliably an absence of metal toxicity based upon EqP
theory. At present, two of the four proposed components for
deriving metal SQC, the AVS and interstitial water approaches
(a and b), are technically defensible; work remains to establish
the applicability of the AVS and organic carbon and minimum
partitioning methods (c and d) [74] Research issues for these
latter two approaches include the development of robust par-
titioning data sets for the five metals, as well as investigation
of factors such as metal competition for binding sites. With
the possible exception of the organic carbon approach, all these
criteria are intended as "no effects" rather than "effects"
values. Even so, they should be useful for dealing with the
majority of sediments to which they are applied [48,66].
-------
2064 Environ Toxicol Chem 15, 1996
G.T. Ankley et al.
Additional research required to implement fully the pro-
posed SQC includes the development of uncertainty estimates
associated with any of the four approaches; part of this would
include their application to a variety of field settings and sed-
iment types. Research also is needed to establish the technical
basis for SQC for metals other than the five described herein,
such as mercury, silver, arsenic, and chromium. Finally, the
SQC approaches described in this paper are intended to protect
benthic organisms from direct toxicity associated with expo-
sure to metal-contaminated sediments. They are not designed
to protect aquatic systems from metal release associated, for
example, with sediment suspension, or the transport of metals
into the food web either from sediment ingestion or the in-
gestion of contaminated benthos. This latter issue, in particular,
should be the focus of future research given existing uncer-
tainty in the prediction of bioaccumulation of metals by ben-
thos [81]
Acknowledgement—This work was supported by the EPA Office of
Research and Development and Office of Water. The support of Mary
Reiley and Chns Zarba is gratefully acknowledged. Technical con-
tributions and assistance were provided by numerous scientists at the
Duluth and Narragansett EPA laboratories, and cooperating univer-
sities, including Manhattan College, University of Wisconsin-Supe-
rior, University of Rhode Island and the University of Minnesota.
David Mount, Russell Enckson, Chns Ingersoll, and Nelson Thomas
provided valuable comments on an earlier version of this manuscript.
Marlene Johnson provided considerable assistance in manuscript prep-
aration.
REFERENCES
1. Sullivan, J., J. Ball, E. Brick, S. Hausmann, G. Pilarskl and
D. Sopclch. 1985. Report of the technical subcommittee on de-
termination of dredge material suitability for in-water disposal
Wisconsin Department of Natural Resources, Madison, WI, USA.
2. Persaud, D., R. Jacqumagi and A. Hayton. 1989. Development
of provincial sediment quality guidelines. Ontario Ministry of the
Environment, Toronto, ON, Canada.
3. Long, E.R. and L.G. Morgan. 1991. The potential for biological
effects of sediment-sorbed contaminants tested in the national
status and trends program NOAA Technical Memorandum NOS
OMA 52. National Oceanic and Atmospheric Administration, Se-
attle, WA. USA.
4. MacDonald, D.D., R.S. Carr, FJX Colder, E.R. Long and C.G.
Ingersoll. 1996. Development and evaluation of sediment quality
guidelines for Florida coastal waters Ecotoxicology 5-253-278.
5. Ingersoll, C.G., P.S. Haverland, E.L. Branson, T.S. Canfleld,
F.S. Dwyer, C.E. Henke and N.E. Kemble. 1996. Calculation
of sediment effect concentrations J. Great Lakes Res. 22-602-
623.
6 Tessier, A. and P.G.C. Campbell. 1987. Partitioning of trace
metals in sediments- Relationships with bioavailability. Hydro-
biologia 149:43-52.
7. Luoma, S.N. 1989. Can we determine the biological availability
of sediment-bound trace elements? Hydrobiologia 176/177.379-
396.
8. Di Toro, DJVI., JJ>. Mahony, DJ. Hansen, KJ. Scott, M.B.
Hicks, S.M. Mayr and M.S. Redmond. 1990. Toxicity of cad-
mium in sediments: The role of acid volatile sulfide Environ.
Toxicol. Chem. 9 1487-1502
9. Di Toro, D.M., C. Zarba, DJ. Hansen, R.C. Swartz, C.E.
Cowan, H E. Allen, N .A. Thomas, P.R. Paquin and W J. Berry.
1991 Technical basis for establishing sediment quality criteria
for non-ionic organic chemicals using equilibrium partitioning.
Environ. Toxicol Chem 10:1541-1583
10. Battelle. 1984. Sediment quality catena development workshop
Office of Water Regulations and Standards, Criteria and Standards
Division, Washington, DC.
11 Adams, W J., R-A. Kimerle and R.G. Mosher. 1985 Aquatic
safety assessment of chemicals sorbed to sediments. In R.D Card-
well, R. Purdy and R.C. Bahner, eds.. Aquatic Toxicology and
Hazard Assessment Seventh Symposium STP 854 American
Society for Testing and Materials, Philadelphia, PA, USA, pp.
429-453.
12. VS. Environmental Protection Agency. 1990 An SAB Report
Evaluation of the equilibrium partitioning (EqP) approach for
assessing sediment quality. EPA-SAB-EPEC-90-006 Office of
Water, Washington, DC.
13. UJS. Environmental Protection Agency. 1992 An SAB Report.
Review of sediment criteria development methodology for non-
ionic organic contaminants. EPA-SAB-EPEC-93-002 Office of
Water, Washington, DC.
14. US. Environmental Protection Agency. 1993. Sediment quality
, criteria for the protection of benthic organisms: Acenaphthene.
EPA 822-R-93-013. Office of Water, Washington, DC
15. VJS. Environmental Protection Agency. 1993. Sediment quality
cntena for the protection of benthic organisms: Endnn. EPA 822-
R-93-016. Office of Water, Washington, DC.
16. VS. Environmental Protection Agency. 1993. Sediment quality
criteria for the protection of benthic organisms: Dieldrin EPA
822-R-93-015. Office of Water, Washington, DC.
17. U.S. Environmental Protection Agency. 1993. Sediment quality
criteria for the protection of benthic organisms Phenanthrene
EPA 822-R-93-014 Office of Water, Washington, DC.
18 U.S. Environmental Protection Agency. 1993 Sediment quality
criteria for the protection of benthic organisms: Fluoranthene.
EPA 822-R-93-012 Office of Water, Washington, DC
19. U.S. Environmental Protection Agency. 1996 The national sed-
iment quality survey. A report to Congress on the extent and
seventy of sediment contamination in surface waters of the United
States EPA 823-D-96-002. Office of Water, Washington, DC.
20. Swartz, R.C., G.R. Ditsworth, D.W. Schults and J.O. Lam-
berson. 1985. Sediment toxicity to a marine infaunal amphipod.
Cadmium and its interaction with sewage sludge. Mar. Environ.
Res 18-133-153.
21. Kemp, P.F. and R.C. Swartz. 1986 Acute toxicity of interstitial
and particle-bound cadmium to a marine infaunal amphipod Mar.
Environ. Res 26:135-153.
22. Tessier, A., P.G.C. Campbell and M. Bisson. 1979. Sequential
extraction procedure for the speciation of paniculate trace metals
Anal. Chem, 51:844-851.
23. Luoma, S.N. and W. Bryan. 1981. A statistical assessment of
the form of trace metals in oxidized sediments employing chem-
ical extractants. Sci. Total Environ. 17:165-196.
24. Goldhaber, M.B. and I.R. Kaplan. 1974 The sulfur cycle In
E.D. Goldberg, ed.. The Sea. Vol. 5—Marine Chemistry John
Wiley & Sons, NY, USA, pp. 569-655.
25. Berner, R.A. 1980 Early Diagenesis- A Theoretical Approach
Princeton University Press, Princeton, NJ, USA.
26. Davies-Colley, RJ., P.O. Nelson and KJ. Williamson. 1985.
Sulfide control of cadmium and copper concentrations in anaer-
obic estuanne sediments. Mar Chem. 16:173-186
27. Emerson, S., L.^ Jacobs and B. Tebo. 1983. The behavior of
trace metals in marine anoxic waters: Solubilities at the oxygen-
hydrogen sulfide interface. In C.S. Wong, E. Boyle, K.W. Bruland
and J.D. Burton, eds.. Trace Metals in Sea Water. Plenum Press,
NY, USA, pp. 579-608.
28. Morse, J.W., F J. Mlllero, J.C. Conmell and D. Rlckard. 1987.
The chemistry of the hydrogen sulfide and iron sulfide systems
in natural waters Earth Sci. Rev. 24:1-42.
29. Cornwell, J.C. and J.W. Morse. 1987. The characterization of
iron sulfide minerals in anoxic marine sediments. Mar. Chem.
22:193-206.
30. Carlson, A.R., G.L. Phipps, V.R. Mattson, P.A. Kosian and
A.M. Cotter. 1991. The role of acid volatile sulfide in deter-
mining cadmium bioavailability in sediments Environ Toxicol.
Chem. 10-1309-1319.
31. Di Toro, DJvI., J.D. Mahony, DJ. Hansen, KJ. Scott, A.R.
Carlson and G.T. Ankley. 1992 Acid volatile sulfide predicts
the acute toxicity of cadmium and nickel in sediments. Environ
Sci. Technol 26.96-101.
32. Casas, A.M. and E.A. Crecelius. 1994. Relationship between
acid volatile sulfide and the toxicity of zinc, lead and copper in
marine sediments. Environ Toxicol Chem, 13:529-436.
33 Green, A.S., G.T. Chandler and E.R. Blood. 1993 Aqueous-,
pore-water-, and sediment-phase cadmium Toxicity relationships
for a meiobenthic copepod Environ Toxicol Chem 12 1497-
1506.
34. Pesch, C.E., DJ. Hansen, W.S. Boothman, WJ. Berry and
-------
Sediment quality criteria for metals
Environ. Toxicol Chem. 15, 1996 2065
J.D. Mahony. 1995. The role of acid volatile sulfide and inter-
stitial water metal concentrations in determining bioavailability
of cadmium and nickel from contaminated sediments to the ma-
rine polychaete Neanthes arenaceodentata. Environ Toxicol.
Chem. 14.129-141.
35 Berry, WJ., DJ. Hansen, J.D. Mahony, D.L. Robson, D.M.
Di Toro, B.P. Shipley, B. Rogers, J.M. Corbln and W.S. Booth-
man. 1996. Predicting the toxicity of metal-spiked laboratory
sediments using acid-volatile sulfide and interstitial water nor-
malizations. Environ, Toxicol. Chem. 15:2067-2079.
36. AnkJey, G.T., G.L. Phipps, E.N. Leonard, DA. Benoit, V.R.
Mattson, PA. Kosian, A.M. Cotter, J.R. Dierkes, DJ. Hansen
and JJ>. Mahony. 1991. Acid volatile sulfide as a factor medi-
ating cadmium and nickel bioavailability in contaminated sedi-
ments Environ. Toxicol. Chem. 10:1299-1307.
37. Ankley, G.T., V.R. Mattson, E.N. Leonard, C.W. West and
JJL,. Bennett. 1993. Predicting the acute toxicity of copper in
freshwater sediments. Evaluation of the role of acid-volatile sul-
fide. Environ. Toxicol. Chem. 12:315-320.
38. U.S. Environmental Protection Agency. 1994. Equilibrium par-
titioning approach to predicting metal bioavailability in sediments
and the derivation of sediment quality criteria for metals, Vol.
1—Briefing Report to the Science Advisory Board. EPA 822-D-
94-002. Office of Water, Office of Research and Development,
Washington, DC.
39. Hansen, DJ., WJ. Berry, JJ>. Mahony, W.S. Boothman, D.M.
Di Toro, D.L. Robson, G.T. Ankley, D. Ma, Q. Van and C.E.
Pesch. 1996. Predicting the toxicity of metal-contaminated field
sediments using interstitial concentration of metals and acid-vol-
atile sulfide normalizations. Environ. Toxicol. Chem. 15:2080-
2094.
40. Luoma, S.N. and J.L. Carter. 1993. Understanding the toxicity
of contaminants in sediments: Beyond the bioassay-based para-
digm. Environ. Toxicol Chem. 12:793-796.
41. Meyer, JJS., W. Davison, B. Sundby, J.T. Oris, DJ. Lauren,
U. Ffirstner, J. Hong and D.G. Crosby. 1994 The effects of
variable redox potentials, pH and light on bioavailability in dy-
namic water-sediment environments In J.L. Hamehnk, P.F Lan-
drum, H.L. Bergman and W.H. Benson, eds., Bioavailability:
Physical, Chemical and Biological Interactions. CRC, Boca Ra-
ton. PL, USA, pp. 155-170.
42. Hare, L., R. Carignan and MA. Huerta-Diaz. 1994. An ex-
perimental study of metal toxicity and accumulation by benthic
invertebrates: Implications for the acid volatile sulfide (AVS)
model. Ltmnol. Oceanogr. 39:1653-1668.
43. DeWitt, T.H., R.C. Swartz, DJ. Hansen, D. McGovern and
W. J. Berry. 1996. Bioavailability and chronic toxicity of cad-
mium in sediment to the estuarine amphipod, Leptocheirus plu-
mutosus. Environ. Toxicol. Chem. 15:2095-2101.
44. Hansen, D J., J.D. Mahony, W J. Berry, S J. Benyi, J.M. Cor-
bin, S.D. Pratt, D.M. Di Toro and M.B. Able. 1996 Chronic
effect of cadmium in sediments on colonization by benthic marine
organisms- An evaluation of the role of interstitial cadmium and
acid-volatile sulfide in biological availability. Environ. Toxicol.
Chem. 15:2126-2137.
45. Liber, K., D J. Call, T J». Markee, K.L. Schmude, M.D. Balcer,
F.W. Whlteman and G.T. Ankley. 1996. Effects of acid-volatile
sulfide on zinc bioavailability and toxicity to benthic macrom- '
vertebrates: A spiked-sediment field experiment Environ. Toxi-
col. Chem. 15:2113-2125.
46. Sibley, P.K., G.T. Ankley, A.M. Cotter and E.N. Leonard.
1996. Predicting chronic toxicity of sediments spiked with zinc:
An evaluation of the acid-volatile sulfide (AVS) model using a
life-cycle test with the midge Chironomus tentans. Environ. Tox-
icol Chem. 15:2102-2112.
47. Howard, D.E. and R.D. Evans. 1993 Acid-volatile sulfide (AVS)
in a seasonally anoxic mesotrophic lake. Seasonal and spatial
changes in sediment AVS. Environ. Toxicol. Chem. 12-1051-
1058.
48. Leonard, E.N., G.T. Ankley and RA. Hoke. 1996 Evaluation
of metals in marine and freshwater surficial sediments from the
Environmental Monitoring and Assessment Program relative to
proposed sediment quality criteria for metals. Environ. Toxicol.
Chem. 15:2221-2232.
49 Peterson, G.S., G.T. Ankley and E.N. Leonard. 1996. Effect
of bioturbation on metal-sulfide oxidation in surficial freshwater
sediments Environ Toxicol. Chem. 15:2147-2155
50 Ankley, G.T., D.M. Di Toro, D J. Hansen, J.D. Mahony, W J.
Berry, R.C. Swartz, RA. Hoke, NA. Thomas, A.W. Garrison,
H.E. Allen and CJJ. Zarba. 1994. Assessing the potential bio-
availability of metals in sediments. A proposed approach Envi-
ron. Manage 18:331-337.
51. Allen, H.E., G. Fu and B. Deng. 1993. Analysis of acid-volatile
sulfide (AVS) and simultaneously extracted metals (SEM) for the
estimation of potential toxicity in aquatic sediments. Environ.
Toxicol. Chem. 12-1-13.
52. Leonard, E.N., V.R. Mattson, DA. Benoit, RA. Hoke and G.T.
Ankley. 1993. Seasonal variation of acid volatile sulfide con-
centration in sediment cores from three northeastern Minnesota
lakes. Hydrobiologia 271:87-95.
S3. Slotton, D.G. and J.E. Renter. 1995. Heavy metals in intact and
resuspended sediments of a California reservoir, with emphasis
on potential bioavailability of copper and zinc. Mar. Freshwater
Res. 46:257-265
54. Boothman, W.S. and A. Helmstetter. 1992. Vertical and sea-
sonal variability of acid volatile sulfides in marine sediments.
Research Report. U.S. Environmental Protection Agency, Nar-
ragansett, RI
55. Brouwer, H. and T.P. Murphy. 1994 Diffusion method for the
determination of acid-volatile sulfides (AVS) in sediment En-
viron. Toxicol Chem. 13:1273-1275
56. Leonard, E.N., A.M. Cotter and G.T. Ankley. 1996. Modified
diffusion method for analysis of acid volatile sulfides and si-
multaneously extracted metals in freshwater sediment Environ.
Toxicol. Chem. 15:1479-1481
57. Brumbaugh, W.G., C.G. Ingersoll, N.E. Kemble, T.W. May
and J.L. Zajcek. 1994. Chemical characterization of sediments
and pore water from the upper Clark Fork River and Milltown
Reservoir, Montana. Environ. Toxicol. Chem. 13:1971-1983.
58. Bufflap, WJE. and H.E. Allen. 1995. Sediment pore water col-
lection methods: A review. Water Res 29-165-177.
59. Prothro, M.G. 1993. Office of Water Policy and Technical Guid-
ance on Interpretation and Implementation of Aquatic Life Metals
Criteria. Letter from M.G. Prothro, Acting Assistant Administra-
tor for the Office of Water, to Water Management Division Di-
rectors, U.S. EPA Regions I-X. Office of Water, Washington, DC.
60. Besser, J.M., C.G. Ingersoll and J.P. Glesy. 1996. Effects of
spatial and temporal variability of acid volatile sulfide on the
bioavailability of copper and zinc in freshwater sediments. En-
viron. Toxicol Chem. 15:286-293.
61. Herlihy, A.T. and A.L. Mills. 1985. Sulfate reduction in fresh-
water sediments receiving acid mine drainage. Appl. Environ.
Microbiol. 49:179-186.
62. Leonard, E.N., V.R. Mattson and G.T. Ankley. 1995. Horizon-
specific oxidation of acid volatile sulfide in relation to the toxicity
of cadmium spiked into a freshwater sediment. Arch. Environ.
Contam. Toxicol. 28.78-84.
63. Di Toro, D.M., J.D. Mahony, DJ. Hansen and WJ. Berry.
1996. A model of the oxidation of iron and cadmium sulfide in
sediments. Environ. Toxicol. Chem. 15:2168-2186.
64. Calamono, W., W. Ahlf and U. Fdrstner. 1990. Exchange of
heavy metals between sediment components and water In J.A.C.
Broekart, S. Gucer and F. Adams, eds. Metal Speciation in the
Environment. Springer-Verlag, Berlin, Germany, pp 503-522
65. Jenne, EA. 1968 Controls on Mn, Fe, Co, Ni, Cu, and Zn
concentrations in soils and water—The significant role of hydrous
Mn and Fe oxides. In Advances in Chemistry. American Chemical
Society, Washington, DC. USA, pp. 337-387.
66. Hassan, S.M., A.W. Garrison, H.E. Allen, D.M. Di Toro and
G.T. Ankley. 1996. Estimation of partition coefficients for five
trace metals in sandy sediments and application to sediment qual-
ity criteria. Environ. Toxicol. Chem 15 2198-2208.
67. Sunda, W.G., D.W. Engel and R.M. Thuotte. 1978. Effect of
chemical speciation of toxicity of cadmium to grass shnmp, Pa-
laemonetes pugio- Importance to free cadmium ion. Environ. Sci.
Technol 12:409-413.
68. Anderson, D.M. and F.M.M. Morel. 1978. Copper sensitivity
of Gonyaulax tamarensis. Limnol. Oceanogr. 23:283-295.
69. Allen, H.E., R.H. Hall and T.D. Brisbin. 1980. Metal speciatioi
effects on aquatic toxicity Environ Sci. Technol 14.441-443
70. Sunda, W.G. and R.R.L. Gulllard. 1976 The relationship (x
tween cupnc ion activity and the toxicity of copper to phyt>
plankton. J. Mar. Res. 34 511-529.
-------
2066 Environ. Toxicol. Chem 15, 1996
G.T. Ankley et al
71 Sunda, W.G. and J.M. Lewis. 1978. Effect of complexation by
natural organic ligands on the toxicity of copper to a unicellular
alga, Monochrysis luthen Limnol. Oceanogr. 23.870-876.
72. Zamuda, C.D. and W.G. Sunda. 1982. Bioavailabihty of dis-
solved copper to the American oyster Crassostrea virgimca- Im-
portance of chemical speciation. Mar. Biol. 66:77-82.
73. Erickson, RJ., DA. Benoit, V.R. Mattson, H.P. Nelson and
E.N. Leonard. 1996. The effects of water chemistry on the tox-
icity of copper to fathead minnows. Environ. Toxicol Chem. 15:
181-193.
74 U.S. Environmental Protection Agency. 1995. An SAB report:
Review of the agencies approach for developing sediment criteria
for five metals. EPA-SAB-EPEC-95-020. Office of Water, Wash-
ington, DC.
75. Stumm, W. and JJ5. Morgan. 1981. Aquatic Chemistry—An
Introduction Emphasizing Chemical Equilibria in Natural Wa-
ters. John Wiley & Sons, New York, NY, USA
76 Mahony, J.D., D.M. Di Toro, A.M. Gonzalez, M. Curto, M.
Dllg, L.D. De Rosa and L.A. Sparrow. 1996. Partitioning of
metals to sediment organic carbon. Environ. Toxicol Chem IS.
2187-2197.
77. Biesinger, K.E., G.M. Christensen and J.T. Fiandt 1986 Ef-
fects of metal salt mixtures on Daphnia magna reproduction
Ecotoxicol. Environ. Saf. 11:9-14
78. Spehar, R.L. and J.T. Fiandt. 1986. Acute and chronic effects
of water quality cntena-based metal mixtures on three aquatic
species. Environ. Toxicol. Chem. 5.917-931.
79. Enserink, E.L., J.L. Maas-Diepeveen and C.S. van Leeuwen.
1991. Combined toxicity of metals: An ecotoxicological evalu-
ation. Water Res. 25:679-687.
80. Kraak, M.HJ5., D. Lavy, H. Schoon, M. Toussalnt, W.H.M.
Peelers and N.M. van Straalen. 1994 Ecotoxicity of mixtures
of metals to the zebra mussel Dreissena polymorpha Environ.
Toxicol. Chem. 13:109-114.
81. Ankley, G.T. 1996. Evaluation of metal/acid-volatile sulfide re-
lationships in the prediction of metal bioaccumulation by benthic
macroinvertebrates. Environ. Toxicol. Chem. 15:2138-2146
-------
Environmental Toxicology and Chemistry, Vo\ 9, pp 1487-1502,1990
Printed in the USA Pcrgamon Press pic
0730-7268/90 $3 00 -f 00
Copyright © 1990 SET AC
TOXICITY OF CADMIUM IN SEDIMENTS:
THE ROLE OF ACID VOLATILE SULFIDE
DOMINIC M. Di TORO,* JOHN D. MAHONY, DAVID J. HANSEN,
K. JOHN SCOTT, MICHAEL B. HICKS, SUZANNE M. MAYR,
and MICHELE S. REDMOND
•Environmental Engineering and Science Program, Manhattan College, Bronx, New York 10471
(Received 20 October 1989; Accepted 2T February 1990)
i
Abstract—The toxicity of chemicals in sediments is influenced by the extent that chemicals bind to
the sediment. It is shown that acid volatile sulfide (AVS) is the sediment phase that determines the LC50
for cadmium in the marine sediments tested. Although it is well known that metals can form insolu-
ble sulfides, it apparently has not been recognized that AVS is a reactive pool of solid phase sulfide
that is available to bind with metals. Amphipod sediment toxicity tests were conducted in the labo-
ratory and the observed amphipod LCSOs on a normalized cadmium concentration basis, [Cd]/[AVS],
is the same for sediments with over an order of magnitude difference in dry weight normalized cad-
mium LCSOs.
Because other toxic metals also form insoluble sulfides, it is likely that AVS is important in de-
termining their toxicity in sediments as well. Most freshwater and marine sediments contain sufficient
acid volatile sulfide for this phase to be the predominant determinant of toxicity. The other sorption
phases are expected to be important only for low AVS sediments, for example, fully oxidized sedi-
ments. From the point of view of sediment quality criteria the other sorption phases would be impor-
tant for metals with large partition coefficients and large chronic water quality criteria.
Keywords—Sediment quality criteria Metal bioavailability Iron sulfide
INTRODUCTION
The toxicity of chemicals in sediments is influ-
enced by the extent that chemicals bind to the sed-
iment. This modifies the chemical potential to
which the organisms are subjected. As a conse-
quence, different sediments will exhibit different
degrees of toxicity for the same total quantity of
chemical. These differences have been reconciled
by relating organism response to the chemical con-
centration in the interstitial water of the sediments
[1-7]. The relevant sediment properties, therefore,
are those that influence the distribution of chemi-
cal between the solid and aqueous phases.
The varying toxicity of nonionic organic chem-
icals in different sediments is related to the organic
carbon content of the sediments [1,6-8]. This is
*To whom correspondence may be addressed.
The address of J.D. Mahony, M.B. Hicks and S.M.
Mayr is Chemistry Department, Manhattan College,
Bronx, NY 10471.
The address of D.J. Hansen is EPA Environmental
Research Laboratory, Narragansett, RI02592.
The address of K.J. Scott and M.S. Redmond is Sci-
ence Applications International Corp., Narragansett, RI
02592.
due to the importance of sediment organic carbon
in determining the extent of sorption of nonionic
organic chemicals to sediments. The analogous
sediment properties for metals would be the phases
that influence partitioning behavior. It has been
suggested that oxides of iron and manganese as
well as organic carbon may be relevant in deter-
mining the toxicity of metals hi sediments [9].
The purpose of this paper is to establish the pri-
macy of the acid volatile sulfide (AVS) phase—the
solid phase sediment sulfides that are soluble in
cold acid—in determining the toxicity of cadmium
in sediments. The results of toxicity tests using
three sediments with differing AVS concentrations
indicate that the LCSOs are different on a dry
weight basis. However, the interstitial water and
water only LCSOs are similar. The problem is to
determine what sediment parameter is controlling
the cadmium activity. ,
Experimental cadmium titrations of iron sulfide
and natural sediments indicate that cadmium can
react with the solid phase AVS to form cadmium
sulfide precipitate. If the quantity of AVS hi a sed-
iment exceeds the quantity of added cadmium, the
concentration of cadmium in the interstitial water
1487
-------
1488
D. M. Dl TORO ET AL.
is nondetectable and no mortality is observed. Be-
cause the AVS is sufficiently reactive, the added
cadmium precipitates as cadmium sulfide which is
insoluble so that the interstitial water cadmium
concentration is low. In addition, CdS itself appar-
ently is not bioavailable. As long as excess AVS is
present no free cadmium exists. However, if the
added cadmium exceeds the AVS, free cadmium is
measured in the interstitial water and amphipod
mortality occurs. The presentation that follows
gives the experimental evidence that leads to these
conclusions.
MATERIALS AND METHODS
Organism collection, holding and
exposure system design
Ampelisca oW/to were collected from tidal flat
sediments in the Pettaquamscutt (Narrow) River,
a small estuary flowing into Narragansett Bay,
Rhode Island; transferred to the laboratory within
one-half hour and separated using a 0.5 mm mesh
screen. Rhepoxynius hudsoni—a new test species-
were collected in shallow water at Ninigret Pond,
Rhode Island. Ampelisca is a tube-dwelling amphi-
pod living in the aerobic surface sediments or in an
oxic tube which penetrates anoxic sediments. Rhe-
poxynius is a free-burrower in the oxic zone. Sub-
adult animals were separated from the sediment
using a 1 mm mesh screen in the field, transported
to the laboratory within 1 h, screened again and
transferred to holding containers. The amphipods
were acclimated in presieved uncontaminated col-
lection site sediment and flowing filtered 20°C sea-
water. During acclimation, the Ampelisca were
fed, ad libitum, the laboratory cultured diatom
Phaeodactyium tricomutum. R. hudsoni could ob-
tain food from sediments but received no sup-
plemental diet.
A 96-h static renewal toxicity test was con-
ducted with the two marine amphipod test species
to determine their response to cadmium in a seawa-
ter-only exposure. The water-only tests were con-
ducted in 900 ml glass chambers at 20°C with no
aeration.
The sediment toxicity tests generally followed
American Society for Testing and Materials
(ASTM) recommendations [10] with changes to ac-
commodate experimental design requirements of
this experiment. For all sediment experiments,
flowing filtered seawater (10 volume replace-
ments/d) and aeration ensured acceptable dis-
solved oxygen concentration and cadmium-free
overlying water, which was confirmed by measure-
ments. In all experiments-lighting was continuous
so that the amphipods would not leave their tubes.
The interstitial and overlying water was sampled
using a diffusion sampler ("peeper") [11,12], de-
signed to fit within the gallon jar exposure cham-
ber. It was constructed of Plexiglas G grade
unshrunk cast acrylic sheet: 15.2 x 7.6 x 5.1 cm
deep with six rows of three 1.9 cm diameter, 3.8
cm deep holes, each with a volume of about 5 ml.
The samples from the three horizontal cavities are
combined to yield a sample volume of 15 ml,
which is required for the electrode measurement.
The open side of the peeper is covered by a sheet
of 1 micron polycarbonate membrane (Nucleo-
pore, Pleasanton, CA), followed by a 0.076 cm
tow density polyethylene gasket and a 1.3 cm Plexi-
glas cover plate, both of which have the same hole
pattern as the body and secured with PVC-1 cap
screws and nuts. Equilibration time for cadmium
in the peeper cavities was determined to be less
than 1 d in water-only experiments in which seawa-
ter was allowed to diffuse into the cavities and time
course measurements were made.
Three 10-d tests were conducted exposing the
amphipod Ampelisca abdita to control and cad-
mium-spiked Long Island Sound sediment in gal-
lon jars, using one liter of sediment and two liters
of overlying water. In two of these tests the jars
also contained the diffusion samplers.
The final toxicity test was conducted using three
sediments in 900 ml exposure vessels with 200 ml
of sediment (3.5 cm depth) and 600 ml of overly-
ing seawater. Amphipods were exposed for 10 d to
control and cadmium-spiked sediments. Interstitial
water samples were taken by centrifuging the sed-
iment from the chemical control vessels at the end
of the experiment.
Sediment acid volatile sulfide
The principal sediment property of concern in
these experiments was the acid volatile sulfide
(AVS) concentration. It is the solid phase sulfide in
the sediment that is soluble .in cold acid (HCl). The
measurement technique is to convert the sulfides to
H2S(aq), purge it with nitrogen gas and trap it
[13]. A 500 ml Erlenmeyer flask reaction vessel
fitted with a three-hole stopper is followed by three
sequentially connected 250 ml Erlenmeyer flask
trapping vessels. The first is a chloride trap with
200 ml of pH 4 buffer (0.05 M potassium hydrogen
phthlate) to prevent chloride carry over. The sec-
ond and third traps contain 200 ml of a 0.1 M sil-
ver nitrate solution for trapping H2S as Ag2S
-------
Acid volatile sulfide and cadmium toxicity
1489
precipitate. The four flasks are connected with air-
tight appropriately shaped glass and Tygon tubing.
To prevent oxidation of H2S the nitrogen gas
flows through an oxygen-scrubbing system consist-
ing of a vanadous chloride solution in the first
scrubbing tower and seawater in the second tower.
Vanadous chloride is prepared using 4 g of ammo-
nium metavanadate boiled with SO ml of concen-
trated hydrochloric acid and diluted to 500 ml.
Amalgamated zinc, prepared by taking about IS g
of zinc, covering it with deionized water and add-
ing three drops of concentrated hydrochloric acid
before adding a small amount of mercury to com-
plete the amalgamation, is then added to the vana-
dous chloride solution.
The sediment sample (10-15 g of wet sediment)
or standard to be analyzed is placed in the reaction
vessel after the entire system has been purged with
nitrogen for about 1 b. The system is again purged
for 5 to 10 min, and deaerated 6 u hydrochloric
acid is added from a thistle tube to achieve a O.S u
final concentration in the vessel. The system is run
at room temperature for 1 h which has been found
by experiment to be sufficient to complete the ex-
traction. The nitrogen gas flows at a bubble rate of
about four per second. The sample vessel is swirled
every 5 or 10 min. At completion, all hydrogen sul-
fide produced has been converted to silver sulfide
in the first silver nitrate trap and no precipitate is
found in the second trap. The suspension in the
first silver nitrate trap is passed through a 1.2 mi-
cron GF fiber filter, dried at 102°C and weighed.
Standards prepared from appropriate quantities
of iron(II) sulfate and sodium sulfide (the latter be-
ing added from a solution standardized against
lead perchlorate), typically gave yields of 95 to
103%. Silver sulfide precipitates were usually in the
range 20 to 30 mg. When a blank was run (sample
without acid), about 0.9 mg silver sulfide was ob-
tained. When the acid was run without a sample,
about 0.6 mg silver chloride was obtained. This
corresponds to a detection limit of ~0.5 /unol/g.
Sediment characterization and
spiking procedure
Sediments of three different acid-volatile sulfide
concentrations were used in the toxicity tests. The
Long Island Sound sediment, with a high AVS
concentration, was collected from an uncontami-
nated site in central Long Island Sound (40°7.95'N
and 72°52.7'W) with a Smith-Mclntyre grab sam-
pler, returned to the laboratory, press sieved wet
through a 2 mm mesh stainless steel screen, ho-
mogenized and stored at 4°C. A. abdita has been
tested many times in this sediment and both its sur-
vival and reproduction are satisfactory [14]. The
Ninigret Pond sediment was a low AVS sand col-
lected from the Rhepoxynius collection site. The
upper few inches of sediment were collected with
a shovel, returned to the laboratory, sieved wet
through a 2 mm stainless steel screen, rinsed sev-
eral times to remove high-organic fine particles,
homogenized and stoned at 4°C. The third sedi-
ment was a 50/50 (v/v) mixture of Long Island
Sound and Ninigret Pond sediments. Sediment
samples were also obtained from the saline region
of the Hudson River and from Black Rock Harbor
on the north shore of Long Island Sound. These
were used only in the sediment titration procedure.
Sediments were spiked by adding 1.0 liter of wet
sediment to 2.0 liter of 20°C filtered seawater
where a weighted amount of cadmium chloride
had been dissolved. The mixture was stirred with
a nylon spatula, capped and placed on a Ro-Tap
sieve shaker for 5 min to ensure complete mixing
and held at ambient temperature (~15°Q in a wa-
ter bath for 7 d (three-sediment experiment) or 3 to
5 d (one-sediment experiment) with gentle aeration
of the water column. The overlying seawater and
a thin layer of cadmium sulfide precipitate that had
formed on the surface of the sediment were re-
moved. In the three-sediment toxicity test, the test
sediments were then homogenized, and 200 ml
were transferred to each of three replicate exposure
containers. For the one-sediment experiments, sed-
iments were not homogenized. Clean filtered sea-
water was added to replace the water siphoned out,
and the peepers, if used, were inserted. In all cases,
containers were then placed in a water bath with
air and seawater delivery.
Toxicity tests
The tests were initiated by adding the sediments
to the exposure containers, inserting the peepers
and waiting 1 d (except third one-sediment exper-
iment: 6 d). The amphipods were separated from
holding containers using a O.S mm stainless steel
screen and distributed sequentially in 100 ml plas-
tic beakers until 30 (three-sediment experiment), SO
(one-sediment experiments) or 25 (water-only ex-
periment) Ampelisca or 20 Rhepoxynius were ac-
cumulated. After sorting and eliminating dead
animals, the delivery in the exposure system was
halted, and one beaker of amphipods was added
randomly to each duplicate exposure container in
each treatment. Rhepoxynius were tested in Nini-
-------
1490
D. M. Dl TORO ET AL.
gret Pond sediments, and Ampelisca in Long Is-
land Sound and the sediment mixture. Additional
containers for each treatment were chemical con-
trols and received no amphipods. Temperature and
salinity of delivered seawater averaged approxi-
mately 20°C and 31 g/L during the exposure pe-
riod for the three-sediment test and the three
one-sediment tests, respectively.
After termination, the contents of each expo-
sure container, but not the chemical controls, were
sieved through a 0.5 mm screen. For Ampelisca,
material retained on the screen was preserved in
5% buffered formalin with Rose Bengal stain for
later sorting. For Rhepoxynius, material retained
on the screen was examined immediately. In both
cases, recovered amphipods were counted, and
missing individuals were counted as mortalities.
Cadmium determinations and (((rations
The AVS and solid phase cadmium were mea-
sured in the chemical control vessels at both the
stan and end of the experiments. The average
of these concentrations is used in the subsequent
analysis. At the termination of the experiments the
cadmium ion concentrations from the peepers
(one-sediment experiments), or from the centrifu-
gate (three-sediment experiment) and during the ti-
trations described below, were measured as Cd2+
activity using an Orion 94-48 cadmium ion selective
electrode and a double junction reference electrode
(Orion 90-02, Cambridge, MA). The electrode was
standardized with a serial dilution of a 1 g/L cad-
mium solution that was also used as the titrant.
Cadmium concentrations are reported as cadmium
activity, mg Cd2+/L. Independent experiments in
seawater indicate that the cadmium activity is 5.0%
of the total dissolved cadmium concentration. Sed-
iment cadmium was determined using a cold con-
centrated nitric acid (16 M, 5 ml) digestion of 10 ml
wet sediment followed by a peroxide oxidation
(10 ml 30%) and evaporation to dryness. The res-
idue was reconstituted to 20 ml using 0.1 M nitric
acid and the cadmium measured using flame atomic
absorption.
Cadmium titrations of FeS suspensions (pre-
pared in the same manner as the AVS standards)
and sediments were performed using sample sizes
of 5 to 10 g dry wt. added to 50 ml seawater which
was constantly stirred. Cadmium chloride was
added and dissolved cadmium was monitored using
the electrode. Oxygen-free conditions were main-
tained using a nitrogen atmosphere provided by a
glove box or by constantly bubbling nitrogen
through the covered titration vessel. In the sedi-
ment titrations where electrode response was slow,
the readings were taken after the response had
stabilized to less than 0.3 mV/min as recommended
by the manufacturer (Orion 94-48 Instruction
Manual p. 22).
CADMIUM TOXICITY AND INTERSTITIAL
WATER CORRELATIONS
Dry weight normalization
The toxicity of cadmium to Ampelisca in Long
Island Sound sediment for the one-sediment exper-
iments are shown in Figure 1 A. The results of the
three-sediment experiment using Rhepoxynius hud-
soni in Ninigret Pond sediment; and Ampelisca in
both Long Island Sound sediment and an equal
parts mixture of the two sediments, is shown in
Figure IB. Mean control mortalities were 5.0, 1.7
and 16.7%, respectively. The Spearman-Karber
median LC50 estimates and 95% confidence lim-
its [15] are listed in Table 1. The curves are log-
logistic concentration response functions fit to the
data simultaneously using the same slope param-
eter. They are included as an aid in visualizing the
data. The LC50s range from 290 pg/g to 2,850
pg/g on a sediment dry weight basis..As shown be-
low these two organisms have nearly the same 96-
h cadmium activity LCSOs in water-only exposures:
17.0 pg Cd2+/L for Rhepoxynius and 32.0 /tg
Cd2+/L for Ampelisca. Therefore the differences
in the cadmium toxicity are likely to be attributable
to sediment properties affecting bioavailability. In
addition, Swartz et al. [2] reported the LC50 for
cadmium to the amphipod Rhepoxynius abronius
in a water-only exposure to be 1.6 mg Cd/L, which
would be a cadmium activity of approximately
81.0 ne, Cd2+/L and a sediment toxicity of 6.9
/tg/g in a Yaquina Bay sediment. Thus, the sensi-
tivity of these three test species in water differs by
less than a factor of 5 whereas the LC50s in four
sediments differ by greater than a factor of 400
(Table 1). An explanation for the variation in
LCSOs in sediments would be useful.
Correlation to interstitial water concentration
Sediments with differing toxicities on a per unit
sediment dry weight basis have been shown to have
similar toxicity based on the interstitial water con-
centrations [1-7]. In addition, the evidence sug-
gests [16] that biological response correlates to
chemical activity, in particular to the divalent metal
activity, (Me2*) [17-19]. Figure 2 presents a com-
parison of the observed mortality to the observed
-------
Acid volatile sulfide and cadmium toxicity
1491
MORTALITY vs SEDIMENT CADMIUM
DRY WEIGHT NORMALIZATION
>•
c
§
100
80
6O
40
2O
(A) INITIAL EXPERIMENTS
• U SOUND
O
100
-7 8°
>• 60
40
20
1
O
(B) JOINT EXPERIMENT
• U SOUND
• MIXTURE
O NINKSRET POND
10
100
1000
1OOOO
1OOOOO
SEDIMENT CADMIUM (ug Cd/gm dry wt)
Fig. 1. (A) Toxicity test results for Long Island Sound sediments (Ampelisca). (B) Toxicity test results for Ninigret
Pond (Rhepoxynius hudsoni). Long Island Sound and the SO/50 (v/v) mixture of the two sediments (Ampelisca).
Cadmium concentrations on a sediment dry weight basis.
interstitial water cadmium activity, measured with
the specific ion electrode, for the three sediments
examined in the paper. The water-only response
data for Ampelisca and R. hudsoni are included
for comparison although they represent a shorter
duration exposure. The curve represents the pooled
water-only data. The interstitial water concentra-
tion data from the sediment exposures are some-
what scattered. However, if the data are grouped
by decades, then the medians (50th percentile) and
interquartile ranges (25th to 75th percentiles) par-
allel the water-only exposure results as shown in
Figure 2. These results conform to previous obser-
vations that the concentration response curves for
sediment exposures, which are different on a sed-
iment cadmium dry weight basis (Fig. 1), are com-
parable on an interstitial water basis (Table 1).
Sediment cadmium vs. interstitial water
The prediction of the toxicity of cadmium in
sediments requires that the relationship between
sediment cadmium concentration and interstitial
water concentration be established. A plot of solid
phase vs. aqueous phase cadmium concentra-
tions—which is referred to as an isotherm plot
when used for the analysis of sorption data—is
-------
1492
D. M. Dl TORO ET Al.
Table 1. LCSOs for cadmium toxicity tests
Experiment
Water only exposure (4-d exposure) (us Cd2+/L)
Ampelisca abdita
Rhepoxynius hudsoni
Rhepoxynhu abronha
Interstitial water (10-d exposure) (/tg Cd2*/L)
A. abdita and R. hudsoni*
Sediments— dry weight normalization (10-d exposure) (/ig Cd/g)
Long Island Sound (A. abdita)b
Long Island Sound (A. abdita)'
Mixture (A. abdita)'
Ninigret Pond (R. hudsoni)'
Yaquina Bay (R. abronius)a
Sediments— AVS normalization (10-d exposure) (pmol Cd/pmol
AVS)
Long Island Sound (A. abdita)b
Long Island Sound (A. abdita)'
Mixture (A. abdita)'
Ninigret Pond (R. hudsoni)'
LC50
17.0
32.0
81.0
22.0
2,580.0
2,850.0
1,070.0
290.0
6.9
1.54
1.70
2.19
1.97
95% Confidence
limits
15.0 , 19.0
27.0 , 37.0
71.0 . 91.0
3.0 , 130.0
2,310.0 , 2,880.0
2,400.0 , 3.390.0
870.0 , 1,310.0
240.0 , 360.0
5.6 . 7.9
1.38, 1.72
1.44. 2.02
1.79. 2.68
1.60, 2.44
•Computed for the medians of the grouped data in Figure 2.
bOne-sediment Long Island Sound sediment experiments (Fig. 1A).
Three-sediment experiment (Fig. IB).
"Swartz et al. [2]. Cd2+ concentrations estimated from total Cd concentrations using measured ratio of [Cd2*]/
[Cd] = 0.050. J
shown in Figure 3. The data can be envisioned as
a titration in which cadmium is added incremen-
tally to the sediment and the resulting aqueous and
solid phase cadmium concentrations are measured.
Initially, the solid phase concentration increases
but the aqueous phase concentration remains be-
low the detection limit of the cadmium electrode.
A critical sediment concentration is reached at
MORTALITY V8 INTERSTITIAL WATER CADMIUM
?
K
i
100
60
40
20
— WATER ONLY
EXPOSURE
A AMPEUSCA
•ft RMEPOXYMUS
O.OOO01 0.0010O 0.10000 1O.OOOOO 1OOO.OOOOO
CADMIUM ACTIVITY (mg Cd2+/L)
Fig. 2. Mortality versus interstitial water cadmium activity. Medians and interquartile ranges for each decade of in-
terstitial water activity. Water only exposure data for Ampelisca and Rhepoxynius hudsoni. The line is a joint.fit to
both water only data sets.
-------
Acid volatile sulfide and cadmium toxicity
1493
SEDIMENT vs INTERSTITIAL WATER CADMIUM
1000OO
n 10000
5
u
<0
1OOO
100
• U SOUND
< LESS 1MAN DETECTION
MECmTATON | TRANSITION
SORPTION
10
O.0001
0.0100 1.OOOO 10O.OOOO
CADMIUM ACTIVITY (mg Cd2+/L)
Fig. 3. Sediment cadmium versus interstitial water cadmium activity for Long Island Sound sediment.
the "transition" region where the aqueous con-
centration increases sharply. In this region the in-
crease in aqueous concentration is over two orders
of magnitude whereas the sediment concentra-
tion remains nearly constant. As cadmium con-
tinues to be added, the data appear to follow a
linear trend which is characteristic of a sorption
reaction.
From a lexicological point of view the transi-
tion region is the critical part of the relationship
between solid and aqueous phase cadmium concen-
trations. The rapid increase in interstitial water
cadmium activity from nonlethal levels below 1.0
/ig Cd2VL, passing the water only LCSOs: 17 and
32 fi% Cd2VL for Ampelisca and Rhepoxynius, to
concentrations in excess of 100 pg Cd2+/L marks
the transition between nontoxic and toxic concen-
trations hi the interstitial water.
The solid phase-aqueous phase relationship at
the lower sediment cadmium concentrations is un-
clear because the aqueous concentrations are below
detection. However, the data do not appear to con-
form to a straight tine sorption isotherm that would
be inferred by extrapolation from the high concen-
tration data because detectable dissolved concen-
trations would have been present at the lower
sediment cadmium concentrations. The data pre-
sented below suggest that a cadmium sulfide precip-
itation reaction is maintaining the aqueous phase
concentration at below detectable values in the re-
gion of low sediment cadmium concentrations.
METAL SULFIDES AND CADMIUM TTTRATIONS
The importance of sulfide in the control of
metal concentrations in the interstitial water of ma-
rine sediments is well-documented [13,20-22].
Metal sulfides are insoluble and the equilibrium in-
terstitial water metal concentrations in their pres-
ence are small. It is possible that the interstitial
water sulfide concentration in the sediments used
for these toxicity tests was initially high enough so
that as cadmium was added to the sediment, cad-
mium sulfide was precipitating following the
reaction:
(1)
However direct measurements of the interstitial
water sulfide activity, (S2~|, with a sulfide elec-
trode failed to detect any free sulfide in the unspiked
sediments. This was a puzzling result because a
bright yellow cadmium sulfide precipitate was
forming as cadmium was added to the sediment.
The lack of significant quantity of dissolved sul-
fide in the interstitial water and the evident forma-
tion of solid phase cadmium sulfide suggested the
following possibility. The majority of the sulfide
in sediments is in the form of solid phase iron sul-
-------
1494
D. M. Di TORO ET AL.
fides [13]. Perhaps the source of the sulfide is this
solid phase sulfide initially present. As cadmium is
added to the sediment it causes the solid phase iron
sulfide to dissolve releasing sulfide which is avail-
able for the formation of cadmium sulfide. The
plausibility of this mechanism is examined below.
Solubility relationships and
displacement reactions
The majority of sulfide in sediments is in the
form of iron monosulfides (mackinawite and
greigite) and iron bisulfide (pyrite) of which the
former are the most reactive. These sulfides can be
partitioned into three broad classes that reflect the
techniques used for quantification [13,23,24]. The
most labile fraction, acid volatile sulfide (AVS), is
associated with the more soluble iron and manga-
nese monosulfides. The more resistant sulfide min-
eral phase, iron pyrite, is not soluble in the cold
acid extraction used to measure AVS. Neither is the
third compartment, organic sulfide associated with
the organic matter in sediments [25].
Iron monosulfide, FeS(s), is in equilibrium with
aqueous phase sulfide by the reaction:
(2)
If cadmium is added to the aqueous phase, the re-
sult is:
As the cadmium concentration increases, [Cd2+] x
[S2~] will exceed the solubility product of cad-
mium sulfide and CdS(s) will start to form. Be-
cause cadmium sulfide is more insoluble than iron
monosulfide, FeS(s) should start to dissolve in re-
sponse to the lowered sulfide concentration in the
interstitial water. The overall reaction is:
FeS(s) -»CdS(s) + Fe:
,2+
(4)
FeS(s) ~
Fe
2+
(3)
The iron in FeS(s) is displaced by cadmium to
form soluble iron and solid cadmium sulfide,
CdS(s). A theoretical analysis of the Cd(II)-Fe(II}-
S(II) system, presented in Appendix I, supports
this conclusion. The relevant parameter, which can
be termed the metal sulfide solubility parameter for
any metal, Me, is orMe.j+ATMeS. It is the product of
"Me2* = IS Me(aq)]/[Me2+], the ratio of total dis-
solved Me to the divalent species concentration;
and KMeS = [Me2+][S2~], the metal sulfide solubil-
ity product (Table 2). The sulfide solubility param-
eters determine the behavior of FeS(s) and any
MeS(s) as the metal is added to the sediment.
For example because the cadmium sulfide solubil-
ity parameter is less than the iron sulfide solubility
parameter, cadmium will form a sulfide at the ex-
pense of the iron sulfide which will dissolve. Note
that all the metals listed in Table 2 below the
dashed line are predicted to dissolve FeS and MnS.
Titration results—amorphous FeS
The calculations presented in Table 2 reflect the
chemical composition expected at thermodynamic
Table 2. Metal sulfide solubility and ratio of total dissolved
to free cation metal concentration
Metal
sulfide
MnS
FeS(am)
FeS
NiS
ZnS
CdS
PbS
CuS
HgS
log**..
-0.40
-3.05
-3.64
-9.23
-9.64
-14.10
-14.67
-22.19
-38.50
log^
-19.15
-21.80
-22.39
-27.98
-28.39
-32.85
-33.42
-40.94
-57.25
loi
pH = 7.6
0.13
0.10
0.10
0.11
0.12
1.50
1.12
0.50
15.10
5«
pH = 8.2
0.13
• 0.12
0.12
0.17
0.14
1.50
1.32
0.92
15.10
Average
-19.02
-21.69
-22.28
-27.84
-28.26
-31.35
-32.20
-40.23
-42.15
Solubility products, KVtl, for the reaction: Me2+ HS~ « MeS(s) + H+ for CdS (Greenockite),
FeS(amorphous) and Mackinawite, MnS (Alabandite) and NiS (Millerite), from [21]. Solubility
products for CuS (Covellite), HgS (Metacinnabar), PbS (Galena) and ZnS (Wurtzitej and pK2 =
18.57 for the reaction HS~ ** H+ + S2~, from [34]. *„ is for the reaction: Me2+ + S5~ « MeS(s)
is computed from log Kv 2 and pK2. Ratios of total tofree metal concentrations: a = [£Me(aoJ] /
[Me2* 1, from 135] at T = 5°C. \og(aKv) - loga + log*,,. All logs are log,0.
-------
Acid volatile sulflde and cadmium toxicity
1495
equilibrium. However many solid phase reactions
are not at equilibrium with respect to either the
aqueous phase of other solid phases because of the
slow kinetics involved in the necessary transforma-
tions. Therefore, a direct test of the extent to which
this reaction' takes place was performed.
A quantity of freshly precipitated iron sulfide
was titrated by adding dissolved cadmium. The re-
sulting aqueous cadmium activity, measured with
the cadmium electrode vs. the ratio of cadmium
added, [Cd]A, to the amount of FeS initially pres-
ent, [FeS(s)]j, is shown in Figure 4. The lines con-
necting the data points are an aid to visualizing the
data. The electrode potentials (left) correspond to
a low cadmium concentration during the initial
portion of the titration. Then a sharp upward in-
flection occurs near [Cd]A <= [FeS(s)]j indicating
that all the iron sulfide has dissolved to form CdS.
Any additional cadmium added appears as free
cadmium. The plot of dissolved cadmium vs. cad-
mium added (right) illustrates the rapid increase
in dissolved cadmium that occurs near [Cd]A/
[FeS(s)], = 1. A similar experiment has been per-
formed for amorphous MnS with comparable re-
sults. These displacement reactions among metal
sulfides have been observed by other investigators
[26]. The reaction was also postulated [27] to ex-
plain an experimental result involving copper and
FeS.
These experiments demonstrate that solid phase
amorphous iron and manganese sulfide can be dis-
solved by adding cadmium. As a consequence it is
a source of available sulfide which must be taken
into account in evaluating the relationship be-
tween solid phase and aqueous phase cadmium in
sediments.
Titration results— sediments
A similar titration procedure has been used to
evaluate the behavior of sediments taken from four
different marine environments: the Long Island
Sound and Ninigret Pond sediments used in the
toxicity tests; and sediments from Black Rock Har-
bor and the Hudson River (Fig. 5). The binding
capacity for cadmium, [Cd]B, is estimated by ex-
trapolating a straight line fit to the dissolved cad-
mium data. The equation is:
= max[0,maCd]A-[Cd]B)J (5)
where [SCd(aq)] is the total dissolved cadmium,
[Cd]A is the cadmium added, [Cd]B is the bound
cadmium, and m is the slope of the straight line.
The sediments exhibit different binding capacities
for cadmium, listed in Table 3, ranging from ap-
proximately 1 /tmol/g to more than 100 pmol/g.
The possibility that acid volatile sulfide is a
duvet measure of the solid phase sulfide that reacts
with cadmium can be examined in Table 3, which
lists the sediment binding capacity for cadmium
and the measured AVS for each sediment. The sed-
iment cadmium binding capacity appears to be
somewhat less than the initial AVS for the sedi-
ments tested. However a comparison between the
initial AVS of the sediments and that remaining af-
ter the cadmium titration is completed, Table 3,
CADMIUM TITRATION OF IRON SULFIDE
ui
CL -100
•v
O
-ISO
CO
0.8
1.0
tJO
oo
O.O
O.5
2.O
CADMIUM ADDED (umol Cd/umol FeS)
Fig. 4. Cadmium titrations of amorphous FeS. Abscissa is cadmium added normalized by FeS initially present. Or-
dinate is cadmium electrode response (left panel) and total dissolved cadmium (right panel)- The lines connecting the
data points are an aid to visualizing the data.
-------
1496
D. M. Di TORO ET AL.
CADMIUM TITRATION OF SEDIMENTS
DRY WEIGHT NORMALIZATION
o>
£ OJB
a
<
o
s
o
w
0.6
O.4
O.2
* BUHAR8OH
• UCOUNO
• HUDSON (OVER
O NWMRETPONO
O.1
1.0
10.0
100.0 1OOO.O
CADMIUM ADDED (unto! Cd/gm dry wt)
Fig. S. Cadmium titration of sediments: Black Rock Harbor, Long Island Sound, Hudson River, Ninigret Pond. Cad-
mium added per unit dry weight of sediment versus total dissolved cadmium.
suggests that some AYS is lost during titration. In
any case the covariation of sediment binding ca-
pacity and AVS is clear in the data in Table 3. This
suggests that AVS is the proper quantification of
the solid phase sulfides that can be dissolved by
SEDIMENT TOXICITY AND AVS
NORMALIZATION
The three-sediment toxicity experiment illus-
trated in Figure IB was designed to test the utility
of AVS as a predictor of the cadmium binding ca-
pacity of sediments and therefore a predictor of the
concentration of cadmium that would cause sedi-
ment toxicity. The results are shown in Figure 6 in
which the sediment cadmium is normalized by the
AVS for that sediment. The averages of the initial
and final values are used for AVS. Mortality occurs
at the point where the sediment cadmium begins to
exceed the sediment AVS on a molar basis. Total
mortality occurs at [Cd]/[AVS] > 3. The estimated
LCSOs for the three sediment experiment are 1.7 to
2.2 fimol Cd/pmol AVS (Table 1). There are no
significant (95%) differences for each pair of
LCSOs as shown by a f test.
The critical point is that the sediment AVS can
be used to normalize the sediment cadmium con-
centration in the same way that sediment organic
carbon is used to normalize nonionic organic
chemicals. The reason that both methods work is
that they properly account for the chemical activ-
ity of the chemical in both the aqueous and sedi-
Table 3. Cadmium binding capacity and AVS of sediments
Sediment
Black Rock Harbor
Hudson River
Long Island Sound0
Mixture0
Ninigret Pondc
Initial AVS
(Minol/g)'
175.0 (41.0 )
12.6 ( 2.8 )
15.9 ( 3.3 )
5.45 ( - )
2.34 ( 0.73) '
Final AVS
(/imol/g)b
13.9 (6.43)
3.23(1.18)
0.28 (0.12)
Cd binding
capacity
(pmol/g)
114. (12.1 )
8.58 ( 2.95)
4.57 ( 2.52)
1.12 ( 0.42)
•Average (standard deviation) AVS of repeated measurements of the stock.
bAverage (standard deviation) AVS after the sediment toxicity experiment.
'From the three sediment experiment.
-------
Acid volatile sulfide and cadmium toxicity
1497
MORTALITY vs SEDIMENT CADMIUM
ACID VOLATILE SULFIDE NORMALIZATION
1OO
-. 80 -
Zj
K
i
0.10 1.OO 1O.OO 1OO.OO
SEDIMENT CADMIUM (umol Cd / umol AVS)
Fig. 6. Mortality versus AVS normalized sediment cadmium for Long Island Sound. Ninigret Pond, and a 50/50 (v/v)
mixture. The sediment cadmium and AVS are the averages of the initial and final concentrations in the control vessels.
• U SOUND
• MIXTURE
O MNKMET POND
meat phases [8]. Below 1 /unol Cd/pmol AVS the
cadmium is all precipitated as CdS(s) and the ac-
tivity of cadmium is low. Above 1 pmol Cd//unol
AVS there exists free cadmium in the interstitial
water, sorbed cadmium in the sediment phase, as
well as CdS(s). The activity of cadmium in the sys-
tem is now high enough to cause mortality. The
reason is that the additional cadmium added in ex-
cess of 1 faaol Cd/pmol AVS is large enough to ex-
ceed the activity of cadmium in the system that
causes mortality even in the presence of some sorp-
tion phases (see Fig. 3).
The normalization of metal concentration by
AVS is, of course, invalid if the AVS is zero. This
would be the case for a fully oxidized sediment.
For sediments with trace amounts of AVS it is
likely that other phases would be important as
well. The results of these experiments indicate that
the lower limit of applicability is an AVS of ap-
proximately 1 pmol/g (Table 3) and possibly lower.
IMPLICATIONS FOR METAL TOXICITY
IN SEDIMENTS
The first order importance of AVS in determin-
ing the toxicity of cadmium in sediments has im-
portant implications. These are discussed below.
A VS in freshwater sediments
Acid volatile sulfide is commonly found in ma-
rine sediments (Table 4). Remarkably, it is also a
common constituent of freshwater sediments. Its
The precipitation of iron sulfide [23]:
presence can be rationalized as follows. Sulfide is
produced by the diagenesis of paniculate organic
carbon, represented as CH2O, with sulfate as the
electron acceptor [24]:
(10)
01)
forms iron monosulfide which is the majority of
the AVS. It might be expected that AVS is signifi-
cant only in marine sediments because the concen-
tration of sulfate in seawater is 28 mM = 2,700 mg
SO4/L. By contrast typical river water sulfate
concentration is 0.12 mM - 11.5 mg SO4/L [28].
However, sedimentary organic matter is present in
either locale and the sulfate in fresh water appears
to be sufficient to produce a significant quantity of
AVS. Surprisingly large values (0.31-112 pmol/g)
are found for sediments from the Great Lakes,
rivers and other freshwater lakes (Table 4). There-
fore the AVS concentration must be considered
when addressing cadmium and other metal toxic-
ity in freshwater sediments.
Application to other metals
The other potentially toxic metals all form
metal sulfide precipitates that are more insoluble
than iron sulfide (Table 2). The iron and manga-
-------
1498
D. M. Di TOKO ET AL.
Table 4. AVS in freshwater and marine sediments
Temperature'
Location (°Q
Fresh water sediments
Everglades peat basin
Lake Mendota
Lake Ontario
Lake Erie
Marine sediments
Long Island Sound
NWC
NWC
NWC
DEEP-1
FOAM-1
Sapelo Island
Mud Flat
Mud Flat
Tidal Ck.
Tidal Ck.
—
—
—
(W)
3.0
13.2
19.0
18.5
20.0
(W)
(S)
(W)
(S)
AVS (/tmol/g)
depth interval
(0-1 cm)
—
—
11.6
15.0
0.0
0.60
0.097
0.62
7.50
1.88
3.44
9.69
5.94
(0-10 cm)
0.31- 1.3
8.7 -112.0
27,1
7.5
, 8.35
10.5
10.3
17.4
13.3
14.6
43.2
28.4
31.9
Reference
[36]
137]
138]
[39]
[30]
'
[31]
•(W) = Winter; (S) = Summer.
nese sulfides have \og(aKv) > -22.3 (above the
dotted line) whereas the remaining sulfides have
log(ojrv) < -22.3 (below the dotted line). The
implication is that the results found for cadmium
are applicable to these other metals as well since,
at equilibrium, they can displace iron and man-
ganese sulfide to form a more insoluble sulfide
precipitate.
In particular it is likely that for any metal the
LCSO, and probably the chronic effect concentra-
tion, is at least 1 pmol/pmol AVS. Given the high
concentrations of AVS in most sediments, the ef-
fect concentrations for these metals are likely to be
large concentrations. Consider a sediment with an
AVS of ,1 junol/g, a value well below the normal
range (Tables 3 and 4). For the metals listed in
Table 2 the effect concentrations corresponding to
1 pmol/taaol AVS are: Ni (58.7 /ig/g); Cu (63.6
pg/g); Zn (65.4 /ig/g); Hg (201 /ig/g) and Pb (207
pg/g). A more common AVS concentration of 10
funol/g would increase these effect concentrations
10-fold.
Application to mixtures
An additional conjecture based on an effects
concentration of 1 pmol/pmol AVS for the metals
listed in Table 2 is that the molar AVS normalized
toxicity of metals is additive. Because all these di-
valent metals have lower sulfide solubility param-
eters than FeS, they would all exist as metal
sulfides if their molar sum is less than the AVS.
For this case no metal toxicity would be expected
and:
IAVS]
<1
(6)
where [MeTh is the total cold acid extractable
metal concentration in the sediment. On the other
hand if their molar sum is greater than the AVS
concentration then a portion of the metals with the
lowest sulfide solubility parameters would exist as
free metal and presumably exert a toxicity. For this
case the following would be true:
StMeT],
[AVS]
>1
(7)
But these two equations are the formulas used to
determine the extent of metal toxicity in sediments
assuming additive behavior and neglecting the ef-
fect of partitioning. Whether the normalized sum
is less than or greater than one discriminates be-
tween nontoxic and toxic sediments. The additivity
does not come from the nature of the mechanism
that causes toxicity. Rather it results from the
-------
Acid volatile sulfide and cadmium toxicity
1499
equal ability of these divalent metals to form metal
sulfides with the same stoichiometric ratio of Me
andS.
This discussion is predicated on the assumption
that all the metal sulfides behave similarly to cad-
mium sulfide. In addition, it has been assumed that
only acid soluble metals are reactive enough to af-
fect the free metal activity. At present no experi-
mental data to support either of these conjectures
exists so that this discussion is purely speculative.
AVS and sediment quality criteria
Because AVS can bind cadmium and presumably
other metals and thereby eliminate then- toxicity,
AVS will obviously play a role in the determination
of sediment quality criteria for metals. For fully
oxidized sediments with little or no AVS, AVS nor-
malization would not be appropriate. Partitioning
would be controlled by other sediment phases such
as iron and manganese oxides and organic carbon
[9].
An estimate of when partitioning to other
phases can be important can be made using the
proposed sediment quality criteria formula [8]:
CSQC —
(8)
where CSQC G*g/g) is the sediment quality criteria,
Kp is the partition coefficient (L/g), and CWQC
(/ig/L) is the chronic water quality criteria. For the
case where there is only one metal competing for
the AVS, the molar equivalent of the AVS would
not be bioavailable. Therefore it should be added
to the allowable concentration so that:
(9)
where [CSQC] is the molar sediment quality crite-
ria Otinol/g), ATP is the partition coefficient (L/g),
and [CWQC] istne molar chronic water quality cri-
teria (/tmol/L). The range for freshwater chronic
criteria for the metals named in Table 2 (hard-
ness = 100 mg/L) is 0.0001 to 1.6 jimol/L. The
marine criteria for the same metals are 0.0001 to
0.88 /tmol/L [29]. The importance of partitioning
can be judged by comparing the product Kp x
[CWQC] to the AVS concentration.
Consider an AVS concentration of 1 junol/g. If
the partition coefficient is Kf = 1 L/g then a metal
with a criteria concentration of 1 /imol/L would
have its sediment quality criteria doubled because
of the partitioning. For Kp = 10 L/g the criteria
concentration at which partitioning doubles the
sediment quality criteria drops to 0.1 /miol/L.
Therefore, the effect of partitioning only becomes
significant for relatively low AVS concentrations
(-1 /*mol AVS/g) and for the metals with larger
partition coefficients and chronic water quality cri-
teria concentrations.
Vertical and temporal AVS profiles
The normal method for sediment preparation in
spiked sediment toxicity tests is to produce a uni-
form distribution of chemical and sediment by
careful mixing. For these systems the AVS is uni-
formly distributed and the concentration to be used
for normalization is unambiguous.
However, the distribution of AVS in intact sed-
iment cores exhibits both vertical and temporal
variation over the annual cycle [30,31] (Table 4).
There is a seasonal variation in the surface concen-
tration of AVS at the Long Island Sound NWC
station. All stations exhibit a strong vertical gradi-
ent between the surface 1 cm and the average of
the top 10 cm. It appears that intact cores should
be used for sediment toxicity testing if metal tox-
icity is suspected. Indigenous predators such as
Nephtys incisa [32] should be eliminated, however,
perhaps by asphyxiation. This method has the ad-
vantage of not affecting the AVS concentration be-
cause anaerobic conditions are maintained.
The vertical and temporal variation in AVS
makes it more difficult to decide what AVS con-
centration should be used in evaluating the poten-
tial toxicity of metals in natural sediments. This is
in contrast to the distribution of sediment organic
carbon which is more spatially uniform and tem-
porally stable. Clearly, further work is required to
understand the effect of the spatial and temporal
variability of AVS on metal toxicity in intact
sediments.
Sediment sampling and interstitial
water generation
Ferrous sulfide oxidizes rapidly in aerobic envi-
ronments. For suspensions, oxidation is virtually
complete within a few hours [33]. A decline hi AVS
of up to 50% was noted for sediments that were
held for a long period in apparently airtight con-
tainers or that were exposed to air. It is clear,
therefore, that care should be taken to keep sedi-
ments free from exposure to oxygen before AVS
measurements or toxicity testing.
The use of elutriates as a surrogate for intersti-
tial water is also suspect because oxidation of metal
sulfides and release of soluble metals can occur.
Procedures for producing large volumes of "pore"
water by equilibrating suspensions of sediments
-------
1500
D. M. Di TORO FT Al-
be checked for the extent of AVS oxidation
that occurs. •
concentrations are equivalent. The details will be
reported subsequently.
CONCLUSIONS
It has been shown that AVS is the proper nor-
malization parameter for cadmium toxicity in sed-
iments. The amphipod LCSOs on an AVS normal-
ized basis, [Cd]/[AVS], is the same for sediments
with over an order of magnitude difference in dry
weight normalized cadmium LCSOs. The correla-
tion between mortality and interstitial water metal
activity has also been confirmed. Although the fact
that metals can form insoluble sulfides is well-
known, it apparently has not been recognized that
FeS and MnS, quantified as AVS, is a reactive pool
of solid phase sulfide that is available to bind with
metals which have sulfide solubility parameters
smaller than FeS, for example, nickel, zinc, cad-
mium, lead, copper and mercury.
Titrations of amorphous FeS and MnS with
cadmium demonstrate that the displacement reac-
tion, Equation 4, occurs. Further, titrations of sed-
iments with cadmium indicates that an abrupt
increase of dissolved cadmium occurs when the
added cadmium exceeds the measured AVS. How-
ever, these data are not as certain because AVS ap-
pears to be lost during the titration and the
relationship is only approximate (Table 3). Never-
theless, the AVS normalized toxicity data (Fig. 6)
demonstrate that the normalization is quantitative.
Surprisingly, the AVS of freshwater sediments
is in the same range as marine sediments. There-
fore, AVS should also be the proper normalization
for these sediments. The normalization is invalid if
the AVS is zero, such as for a fully oxidized sedi-
ment. For sediments with trace amounts of AVS it
is likely that other phases would be important as
well. The experiments reported in this paper indi-
cate that the lower limit of applicability is AVS -
1 fimol/g and possibly lower. The other sorption
phases are expected to be important only for sed-
iments with smaller AVS concentrations and for
metals with large partition coefficients and large
chronic water quality criteria.
We have since determined that using the metal
concentration in the sediment which is simulta-
neously liberated by the AVS extraction, rather
than the total metal concentration of the sediment,
is the correct procedure for other metals (Ni, Cu,
Zn). A sample of the solution remaining in the re-
action vessel after the AVS procedure is complete
is filtered and analyzed for the metals concentra-
tions. For the Cd experiments reported above the
total metal and simultaneously extracted metal
Acknowledgement—This research was sponsored by an
EPA Cooperative Agreement CR812824-01 between
Manhattan College and EPA Environmental Research
Laboratory, Narragansctt, RI. The assistance and encour-
agement of Christopher Zarba, EPA Criteria and Stan-
dards Division; Herbert Allen, Drexel University and our
research assistants at Manhattan College: Indra Sweeney,
Paul Morgan, Clare Sydlik, Luisa Milevoj and Christine
Begley and at the EPA Narragansett Laboratory: De-
borah Robson and Kathleen McKenna (SAIQ. are grate-
fully acknowledged.
REFERENCES
1. Adams, WJ., R.A. Kimetfe and R.G. Mosher. 1985.
Aquatic safety assessment of chemicals sorbed to
sediments. In R.D. Cardwell, R. Purdy and R.C.
Banner, eds.. Aquatic Toxicology and Hazard As-
sessment: Seventh Symposium. American Society for
Testing and Materials, Philadelphia, PA, pp. 429-
453.
2. Swartz, R.C., G.R. Ditswortb, D. W. Scbults and
J.O. Lamberson. 1985. Sediment toxicity to a marine
infaunal amphipod: Cadmium and its interaction
with sewage sludge. Mar. Environ. Res. 18:133-153.
3. Muir, D.C.G., G.P. Rawn, B.E. Townsend, W.L.
Lockhart and R. Greenhalgh. 1985. Bioconcentration
of cypermethrin, deltamethrin, fenvalerate, and per-
methrin by Chironomus tentans larvae in sediment
and water. Environ. Toxicol. Chem. 4:51-61.
4. Adams, WJ. 1987. Bioavailability of neutral lipo-
philic organic chemicals contained in sediments: A
review. In K.L. Dickson. A.W. Maki and W.A.
Brungs, eds., Fate and Effects of Sediment-Bound
Chemicals in Aquatic Systems. Pergamon Press,
Elmsford, NY, pp. 219-244.
5. Kemp, P.P. and R.C. Swartz. 1988. Acute toxicity of
interstitial and particle-bound cadmium to a marine
infaunal amphipod. Mar. Environ. Res. 26:135-153.
6. Nebeker, A. and G. Schoytema. 1988. DDT/Endrin
Results. Evaluation of Carbon Normalization The-
ory. U.S. Environmental Protection Agency Envi-
ronmental Research Laboratory, Corvallis, OR.
7. Swartz, R.C., D.W. Schults, T.H. DeWitt, G.R.
Ditsworth and J.O. Lamberson. 1990. Toxicity of
fluoranthene in sediment to marine amphipods: A
test of the equilibrium partitioning approach to sed-
iment quality criteria. Environ. Toxicol. Chem. 9:
1071-1080.
8. Di Tore, D.M., C. Zarba, DJ. Hansen, W. Berry,
R.C. Swartz, C.E. Cowan, S.P. Pavlou, H.E. Allen,
N.A. Thomas and P.R. Paquin. 1991. Technical basis
for establishing sediment quality criteria for nonionic
organic chemicals using equilibrium partitioning. En-
viron. Toxicol. Chem. (in press).
9. U.S. Environmental Protection Agency. 1989. Brief-
ing report to the EPA science advisory board on the
equilibrium partitioning approach to generating sed-
iment quality criteria EPA 440/5-89-002. Office of
Water Regulations and Standards, Criteria and Stan-
dards Division, Washington, DC.
10 American Society for Testing and Materials. 1989.
-------
Acid volatile sulfide and cadmium toxicity
1501
Proposed new standard guide for conducting solid
phase 10-day static sediment toxicity tests with ma-
rine and estuarine amphipods. Draft No. 5. Philadel-
phia, PA.
11. Hesslein, R.H. 1976. An in situ sampler for close in-
terval pore water studies. Ltmnol. Oceanogr. 21:912-
914.
12. Carignan, R., F. Rapin and A. Tessier. 1985. Sedi-
ment porewater sampling for metal analysis: A com-
parison of techniques. Geochim. Cosmochim. Acta
49:2493-2497.
13. Morse, J.W., FJ. Millero, J.C. CornweU and D.
Rkkard. 1987. The chemistry of the hydrogen sulfide
and iron sulfide systems in natural waters. Earth Sci.
Rev. 24:1-42.
14. Scott, K J. and M.S. Redmond. 1989. The effects of
a contaminated dredged material on laboratory pop-
ulations of the tubicolous amphipod, Ampelisca ab-
dita. In U.M. Cowgill and L.R. Williams, eds.,
Aquatic Toxicology and Hazard Assessment: Twelfth
Symposium. American Society for Testing and Ma-
terials, Philadelphia, PA, pp. 289-303.
IS. Hamilton, M.A., R.C. Russo and R.V. Tburston.
1977. Trimmed Spearman-Karber method of estimat-
ing median lethal concentrations in toxicity bioas-
says. Environ. Sci. Technol. 11:714-719. Correction
12:417 (1978).
16. Borgmann, U. 1983. Metal spetiation and toxicity of
free metal ions to aquatic biota. In J.O. Nriagu, ed.,
Aquatic Toxicology. John Wiley & Sons, New York,
NY, pp. 47-72.
17. Sunda, W. and R.R.L. Gnillard. 1976. The relation-
ship between cupric ion activity and the toxicity of
copper to phytoplankton. J. Mar. Res. 34:511-529.
18. Sunda, W.G., D.W. Engej and R.M. Thuotte. 1978.
Effect of chemical speciation of toxicity of cadmium
to grass shrimp, Palaemonetes pugio: Importance of
free cadmium ion. Environ. Sci. Technol. 12:409-
413.
19. Sunda, W.G. and J.M. Lewis. 1978. Effect of cpm-
plexation by natural organic ligands on the toxicity
of copper to a unicellular alga, Monochrysis hitheri.
Limnol. Oceanogr. 23:870-876.
20. Boulegue, J., C J. Lord III and T.M. Church. 1982.
Sulfur speciation and associated trace metals (Fe, Cu)
in the pore waters of Great Marsh, Delaware. Geo-
chim. Cosmochim. Acta 46:453-464.
21. Emerson, S., L. Jacobs and B. Tebo. 1983. The be-
havior of trace metals in marine anoxic waters: Sol-
ubilities at the oxygen-hydrogen sulfide interface. In
C.S. Wong, E. Boyle, K.W. Bruland and J.D. Bur-
ton, eds., Trace Metals in Sea Water. Plenum Press,
New York, NY, pp. 579-608.
22. Davks-Colley, R J., P.O. Nelson and KJ. William-
son. 1985. Sulfide control of cadmium and copper
concentrations in anaerobic estuarine sediments.
Mar. Chem. 16:173-186.
23. Berner, R.A. 1967. Thermodynamic stability of sed-
imentary iron sulfides. Am. J. Sci. 265:773-785.
24. Goldhaber, M.B. and I.R. Kaplan. 1974. The sulfur
cycle. In E.D. Goldberg, ed., The Sea, Vol. 5-
Marine Chemistry. John Wiley & Sons, New York,
NY, pp. 569-655.
25. Landers, D.H., M.B. David and MJ. Mitcheu. 1983.
Analysis of organic and inorganic sulfur constituents
in sediments, soils and water. Int. J. Environ. Anal.
Chem 14:245-256.
26. Phillips, H.O. and K.A. Kraus. 1965. Adsorption on
inorganic materials VI. Reaction of insoluble sulfides
with metal ions in aqueous media. J. Chromatogr.
17:549-557.
27. Pankow, J.F. 1979. The dissolution rates and mech-
anisms of tetragonal ferrous sulfide (Mackinawite) in
anoxic aqueous systems. Ph.D. thesis. California In-
stitute of Technology, Pasadena, CA.
28. Stumm, W. and JJ. Morgan. 1970. Aquatic Chem-
istry. Wiley-Interscience, New York, NY.
29. U.S. Environmental Protection Agency. 1986. Qual-
ity criteria for water. EPA 440/5-86-001. Office of
Water, Washington, DC.
30. Alter, R.C. 1980. Diagenetic processes near the sed-
iment-water interface of Long Island Sound. I. De-
composition and nutrient element geochemistry (S,
N, P). In B. Saltzman, ed., Estuarine Physics and
Chemistry: Studies in Long Island Sound. Academic
Press, New York, NY, pp. 238-350.
31. Reaves, C. 1984. The migration of iron and sulfur
during the early diagenesis of marine sediments.
Ph.D. thesis. Yale University, New Haven, CT.
32. Redmond, M.S. and KJ. Scott. 1989. Amphipod
predation by the inf aunal polychaete, Nephtys incisa.
Estuaries 12:205-207.
33. Nelson, M.B. 1978. Kinetics and mechanisms of the
oxidation of ferrous sulfide. Ph.D. thesis. Stanford
University, Palo Alto, CA.
34. Schoonen, M.A.A. and H.L. Barnes. 1988. An ap-
proximation of the second dissociation constant for
HjS. Geochim. Cosmochim. Acta 52:649-654.
35. Byrne, R.H., L.R. Knrnp and KJ. Cantrdl. 1988.
The influence of temperature and pH on trace metal
speciation in seawater. Mar. Chem. 25:163-181.
36. Altschuler, Z.S., M.M. Sdmepfe, C.C. SBber and
F.O. Simon. 1983. Sulfur diagenesis in Everglades
peat and origin of pyrite in coal. Science 221:221-
227.
37. Nriagu, J.O. 1968. Sulfur metabolism and sedimen-
tary environment: Lake Mendota, Wisconsin. Lim-
nol. Oceanogr. 13:430^439.
38. Nriagu, J.O. and R.D. Coker. 1976. Emission of sul-
fur from Lake Ontario sediments. Limnol. Ocean-
ogr. 21:485-489.
39. Matisoff, G., J.B. Fisher and P.L. McCall. 1981. Ki-
netics of nutrient and metal release from decompos-
ing lake sediments. Geochim. Cosmochim. Acta
45:2333-2347.
APPENDIX I
SOLUBILITY RELATIONSHIPS
FOR METAL SULFIDES
The behavior of iron sulfide during a titration with cad-
mium can be analyzed using a simplified equilibrium
model of the Cd(II)-Fe(II)-S(II) system 128]. The mass ac-
tion laws for the sulfide solubilities are:
-red** lCd2+hs2- [S2-] =Kcas (12)
>Fe2+ [Fe2+hs2- [S2-] = KFtS (13)
where [Cd2+], [Fe2*] and [S2~] are the molar concentra-
tions; red2*, 7Fe2*. and 7s2- are the activity coeffi-
cients; and KFeS and KCAS are the sulfide solubility
-------
1502
D. M. Dl TORO ET AL.
products. The mass balance equations for total cadmium, nr.c^i rc.c/..\i
iron(II) and sulfide are: ^(s)]. [FeS(s)], - —-^ ^ + ^
lCdS(s)] = [Cd]A (14)
[Fe2+] + [FeS(s)] = [FeS(s)], (15)
«s2- IS2'] + lCdS(s)] + [FeS(s)] = IFeS(s)]i (16)
where «Q,2* = [SCd(aq)]/[Cd2+],
[Fe2*J and as2- = [SS(aq)J/[S2-] are the ratios of the
total dissolved Cd, Fe(II) and S(II) to the divalent species
concentrations, respectively. (CdS(s)J and fFeS(s)J are the
concentrations of solid phase cadmium and iron sulfide;
[FeS(s)]i is the initial iron sulfide in the sediment and
[Cd]A is the added cadmium.
The solution of these equations begins with substitut-
ing Equations 14 and IS into Equation 16. Noting that
«s2-tS2-] = ES(aq)J « [Cd}A, which states that the to-
tal dissolved sulfide in the interstitial water is much less
than the cadmium added, it follows that:
•ys2-[S2-]<
KfeS/yetn-
XcdS/Tcd2*
[Cd]A
(17)
Then substituting Equations 12, 13 and 17 into Equa-
tions 14 and IS yields the concentrations of solid phase
sulfides:
[CdS(s)J - [CdJA 1 -
**Cd2"*" ^CdS |
Ctcd^"*"^CdS ^" OfFe^^^FeS/
(18)
[Cd]A
(19)
where it has been assumed that the activity coefficients
for Cd2+ and Fe2* are equal, 7ca2+ " 7F«2+> because
they are both divalent cations.
The relative magnitudes of otFc2+ KFeS and <*ca2+ x
KCOS determines the behavior of [FeS(s)] and (CdS(s)] as
cadmium is added to the sediment. For this reason they
are termed sulfide solubility parameters. Table 2 presents
reported values. Because the cadmium solubility param-
eter is much less than the iron sulfide solubility param-
eter, that is, ocdj+ KCOS < <*Fc2* *RS. Equations 18 and
19 become:
and:
[CdS(s)J«[CdAJ (20)
lFeS(s)] • [FeS(s)], - [CdA] (21)
Therefore, as cadmium is added to this system cadmjmp
sulfide forms at the expense of iron sulfide. The overall
reaction is:
Cd2+ -I- FeS(s) -> CdS(s) +
(22)
Note that
[FeS(s)]i; (CdS(s)]
form.
^cas > a**2* ^Fes then [FeS(s)] »
•* 0 and no cadmium sulfide would
-------
Environmental Toxicology and Chemistry, Vol IS, No 12, pp 2080-2094, 19%'
Printed m the USA
0730-7268/96 $6.00 + .00
PREDICTING THE TOXICITY OF METAL-CONTAMINATED FIELD SEDIMENTS USING
INTERSTITIAL CONCENTRATION OF METALS AND
ACID-VOLATILE SULFIDE NORMALIZATIONS
D.J. HANSEN,*t WJ. BERRY,*f J.D. MAHONY.t W.S. BOOTHMAN.f D.M. Dl TORO,t§ D.L. ROBSON,||
G.T. ANKLEY,* D. MA,tt Q .YANft and C.E. PESCH|
tU.S. Environmental Protection Agency, National Health and Environmental Effects Laboratory, Atlantic Ecology Division,
27 Tarzwell Drive, Narragansctt, Rhode Island 02882
JDepartment of Environmental Engineering, Manhattan College, Bronx, New York 10471, USA
SHydroQual, Inc., 1 Lethbridge Plaza, Mahwah, New Jersey 07430, USA
((Rhode Island Department of Environmental Management, Providence, Rhode Island 02903, USA
#U.S. Environmental Protection Agency, Mid-Continent Ecology Division, 6201 Congdon Boulevard,
Duluth, Minnesota 55804
ttPRC Oceanic Administration, Institute of Marine Environmental Protection, Dalian, People's Republic of China
(Received 18 September 1995; Accepted 20 May 19%)
Abstract—We investigated the utility of interstitial water concentrations of metals and simultaneously extracted metal/acid-volatile
sulfide (SEM/AVS) ratios to explain the biological availability of sediment-associated divalent metals to benthic organisms exposed
in the laboratory to sediments from five saltwater and four freshwater locations in the United States, Canada, and China. The
amphipod Ampeliica abdita or the polychaete Neanthes arenaceodentata were exposed to 70 sediments from the five saltwater
locations, and the amphipod Hyalella azteca or the oligochaete Lumbricv.hu variegatiu were exposed to 55 sediments from four
freshwater locations in 10-d lethality tests. Sediment toxicity was not related to dry weight metals concentrations. Almost complete
absence of toxicity in spiked sediments and field sediments where metals were the only known source of contamination and where
interstitial water toxic units (IWTUs) were <0.5 indicates that toxicity associated with sediments having SEM/AVS ratios <1.0
from two saltwater locations in industrial harbors was not metals-related as these sediments contained <0.5 IWTU. Metals-associated
toxicity was absent in 100% of sediments from the remaining three saltwater field locations, where metals were the only known
source of contamination and SEM/AVS ratios were £1.0. Two-thirds of 45 sediments from seven saltwater and freshwater field
locations having both IWTUs >0.5 and SEM/AVS ratios >1.0 were toxic. Toxicity was observed less often when SEM/AVS ratios
>1.0 (39%) or IWTUs >0.5 (55%) were used alone. The difference between the molar concentrations of SEM and AVS (SEM -
AVS) can provide important insight into the extent of additional available binding capacity, the magnitude by which AVS binding
has been exceeded, and, when organism response is considered, the potential magnitude of importance of other metal binding
phases. For these reasons, SEM - AVS should be used instead of SEM/AVS ratios as a measure of metals availability. Over all
published experiments with both metal-spiked and field sediments, SEM - AVS and IWTUs accurately (99.2%) identified absence
of sediment toxicity and with less accuracy (79.1%) identified the presence of toxicity.
Keywords—Sediment Acid-volatile sulfide Metals Toxicity Interstitial water
t INTRODUCTION
Accurate prediction of the bioavailability of metals in sed-
iments requires mechanistic knowledge of the role of sediment
geochemistry in sediment/pore water partitioning, metal form,
and organism-mediated exposure pathways [1]. In oxic sedi-
ments, metals availability for bioaccumulation is related to
organic carbon binding and complexation to iron oxyhydroxide
[2]. Other geochemical processes also control metals avail-
ability and form in oxic sediments [3-5].
In anoxic sediments, the availability of divalent metals to
organisms living in nearby oxic surficial sediments or tubes
has been related to acid-volatile sulfide (AVS), principally iron
monosulfide, binding [6] and organic carbon partitioning [7].
Simultaneously extracted metal (SEM), the metal extracted by
the AVS analytical method, not total metal, is the best estimate
* To whom correspondence may be addressed
Contribution 1575, U.S. Environmental Protection Agency (EPA),
National Health and Environmental Effects Research Laboratory, At-
lantic Ecology Division, Narragansett, Rhode Island.
of potentially bioavailable metal concentrations for compari-
son to AVS [8]. Acid-volatile sulfide, hence availability of
metals, can vary by season and sediment depth in response to
sulfur cycles, which are related to temperature and productivity
[9]. Sediments spiked with cadmium, copper, lead, nickel, zinc,
or mixtures of these divalent metals have been shown to con-
tain little interstitial metal and to be nontoxic to saltwater or
freshwater snails, oligochaetes, polychaetes, or amphipods
when molar concentrations of AVS exceed molar concentra-
tions of SEM (SEM/AVS ratio sl.O) [6,8,10-13]. Toxicity
was often, but not always, observed at SEM/AVS >1.0. Fur-
thermore, field-collected saltwater [13] and freshwater [14-
15] sediments with SEM/AVS ratios sl.O also were nontoxic.
Most laboratory-spiked and field-collected sediments were
toxic when SEM/AVS ratios exceeded 1.0. Absence of toxicity
in spiked and field sediments was associated with interstitial
metal concentrations less than those known to be toxic to tested
species in water-only tests. Toxic sediments contained inter-
stitial metal concentrations similar to concentrations toxic in
water-only tests.
2080
-------
SEM/AVS ratio and interstitial metals: Field sediment toxicity prediction
Environ. Toxicol. Chem. 15, 1996 2081
The objective of this article is to further demonstrate the
utility of interstitial water concentrations of metals and sedi-
ment concentrations normalized on the basis of SEM/AVS
ratios to explain the bioavailability of sediment-associated
metals to benthic organisms. The first part presents previously
unpublished data on the relationship between total metal con-
centrations, interstitial, metal concentrations, and SEM/AVS
ratios and toxicity to the saltwater amphipod Ampelisca abdita
(which was exposed to sediments from five marine sites located
in Maryland, Massachusetts, and New York, USA; New Bruns-
wick, Canada; and Liaoning Province, China) together with
previously published results using sediments from New York
with the saltwater polychaete Neanthes arcnaceodentata [13].
Next, data are presented on these relationships with the fresh-
water amphipod Hyalella azteca and the oligochaete Lum-
briculus variegatus, which were exposed to sediments from
four freshwater field locations. Data from locations in New
York, Michigan, and Washington, USA, have been published
previously [8,14-15]; those from Missouri, USA, are new. All
are herein analyzed collectively. Finally, this article combines
results from all experiments using field-collected saltwater and
freshwater sediments with those from all available laboratory
spiked-sediment tests using a variety of saltwater and fresh-
water species [10].
METHODS
Saltwater field sites
Sediment collection, storage, and handling. Sediments were
collected by plastic scoop, shovel, Ponar grab, or modified
Van Veen grab from Jinzhou Bay, China (September 1992);
Belledune Harbor, New Brunswick, Canada (August 1990);
Bear Creek, Maryland (February 1992); a tidal marsh near
Fairhaven, Massachusetts (March 1991); and Foundry Cove,
New York (August 1989) (Fig. 1). Samples consisting of ap-
prox. 5 to 10 cm of surficial sediment were homogenized, and
aliquots removed for total metal, total organic carbon, and
grain size analyses. Sediments were transported under ice and
stored at 4°C in sealed glass jars with limited headspace con-
tabling nitrogen until use. Prior to conducting toxicity tests,
sediments were rehomogenized, taking care to limit oxidation
of metal sulfides.
Toxicity tests. The 10-d lethality tests with the amphipod A.
abdita generally followed methodologies described by the
American Society for Testing and Materials [16], Di Tore et al.
[6], and Berry et al. [10]. Those with the polychaete N. aren-
aceodentata are described by Pesch et al. [13]. Amphipod ex-
posure chambers consisted of 900-ml glass canning jars, with
a 1.3-cm-diameter overflow hole covered with 400-|un Nitex*
mesh. Each chamber contained 200 ml of sediment and 600 ml
of seawater. Polychaete chambers consisted of 600-ml beakers
containing 200 ml of sediment One day before the start of the
test, sediment from each station was placed into each of four
(two chemistry, days 0 and 10. and two biology) replicate ex-
posure chambers. For each experiment with sediments from the
five saltwater locations, one or more treatments consisted of
four replicate chambers containing sediment from an uncontam-
inated reference station in central Long Island Sound, New York,
Narragansett Bay, Rhode Island, USA, or an uncontaminated
sediment from a location near the study site. Sediments from
all stations at Foundry Cove and stations 1 to 10 at the Mas-
sachusetts salt marsh site had interstitial salinities less than those
tolerated by A. abdita or N. arenaceodentata; therefore, sedi-
ments were mixed with brine to obtain 26 to 32%o interstitial
salinities prior to testing. Diffusion samplers (peepers) were
placed in both biology replicates and the day 10 chemistry
replicate to sample interstitial water. Peepers consisted of 5-ml I
polyethylene vials (21 mm high, 20 mm in diameter) covered
with a 1-um polycarbonate membrane and filled with 30%o
salinity water [10]. A plastic strap around the peeper extended
above the sediment to facilitate recovery. To provide continuous
renewal of overlying water, filtered seawater (20°C, 28 to 34%o
salinity) flowed through each replicate chamber at approx. 30
volume additions/d.
Each exposure began with random placement of 20 amphi-
pods or 15 polychaetes in the day 10 chemistry replicate and
in the two biology replicates for each treatment Sediment from
the day 0 chemistry replicate was homogenized, and aliquots
removed and frozen for AVS. SEM, and bulk metal analyses.
Experimental chambers were checked daily for dead animals
and water flow. Overlying water was sampled at the beginning
of every test and at least once thereafter, with samples acidified
and stored in vials as described above. On day 10, peepers
were removed from each sediment, and the water sample acid-
ified and stored. Sediment from the day 10 chemistry replicate
was homogenized, and aliquots removed for AVS and SEM
analyses. Sediments from the biology replicates were sieved
through a 0.5-mm mesh screen to quantify dead and surviving
organisms. Samples with more than 10% of the amphipods
missing were recounted by a second person. Missing amphi-
pods were assumed to be dead. For illustrative purposes, sed-
iments were classified as toxic if mortality was greater than
24%, as proposed by Meams et al. [17] from results of sed-
iment tests with the amphipod Rhepoxynius abronius. Sedi-
ments having £24% mortality were considered nontoxic.
Chemical analyses. Sediment samples were analyzed for
AVS by the cold-acid purge-and-trap technique described by
Allen et al. [18], Cornwell and Morse [19], and Boothman and
Helmstetter [20]. Simultaneously extracted metal and bulk
metals analyses were performed using inductively coupled
plasma emission spectrometry (ICP). For analyses of bulk met-
als, the metals were extracted from freeze-dried sediments by
ultrasonic agitation with 2 M of cold nitric acid (50 ml/5 g
wet sediment) at 60°C overnight followed by centrifugation.
Total metals analyses of sample blanks and recoveries of
known metal additions demonstrated 85 to 100% recoveries
from sediments, 85 to 115% recoveries from sample extracts,
and an absence of contamination in our analytical procedures.
The SEM concentration reported is the sum of cadmium, cop-
per, lead, nickel, and zinc on a micromole per gram dry sed-
iment basis. Concentrations of all metals in sediments ex-
ceeded analytical detection limits.
Interstitial water from peepers and overlying water were
analyzed using ICP or graphite furnace atomic absorption spec-
troscopy. Immediately prior to sediment sampling, peepers
were removed and rinsed to remove sediments. The water
contained within the peepers was removed by pipette and
placed in a 7-ml polyethylene vial and acidified with 50 IL! of
concentrated (pH s 1.0) nitric acid. Detection limits varied as
a function of sample size and method of analysis. Concentra-
tions in water are reported as the sum of the interstitial water
toxic units (TWTUs) of detectable metal, i.e., the sum of mi-
crograms of metal per liter in interstitial water divided by the
10-d LC50 in water-only tests (in u.g/L) for each of the five
metals, where the 10-d LC50 for A. abdita is 36.0 u>g Cd/L,
20.5 jig Cu/L, 3,020 jig Pb/L. 2.400 (tg Ni/L, and 343 (ig
Zn/L [10] and for N. arenaceodentata is 3,670 u.g Cd/L and
-------
2082 Environ Toxicol Chem 15, 1996
D J. Hansen et al
Om S00n< 1000m
Patapsco River
West
Foundry
Foundry Cove, NY
Battery
Plant
2 Source
16
Foundry
Brook
100m
Fig. 1. Location of field sites and stations sampled in Jmzhou Bay, Belledune Harbor, Bear Creek, Foundry Cove, and a salt marsh in Massachusetts.
16,090 >JLg Ni/L [13]. Thus, if interstitial water is the principal
source of metals toxicity and availability of metals is the same
in water of water-only tests and interstitial water in sediment
tests, 50% mortality would be expected with sediments having
IWTUs of 1 0 In this article we use IWTUs of 0.5 to indicate
sediments unlikely to cause significant mortality because on
average water-only LCO and LC50 values differ by a factor
of approx. 2. This factor is reasonable because in saltwater
tests mortality was always absent in sediments spiked with
metals when IWTUs were <0.5 [10]. For illustration, a con-
centration of 0.01 IWTU is used to indicate interstitial water
samples that contain no detectable metal.
Freshwater field sites
Methods used to collect, store, and handle freshwater sed-
iments and to test sediments with the amphipod H. azteca have
-------
SEM/AVS ratio and interstitial metals- Field sediment toxicity prediction
Environ Toxicol Chem 15, 1996 2083
been described for samples from Steilacoom Lake, Washing-
ton, and Keweenaw Watershed, Michigan, by Ankley et al.
[15] and for H. azteca and the oligochaete L. variegatus with
sediments from Foundry Cove, New York, by Ankley et al.
[14]. These same procedures were also used with sediments
from Turkey Creek, Missouri General biological and chemical
procedures, as well as die conceptual experimental design,
were essentially the same for saltwater .and freshwater tests,
except that for freshwater tests bulk metals analyses were not
performed and interstitial water was extracted by centrifuga-
tion instead of diffusion samplers.
RESULTS
Saltwater field sites
Description of field sites and toxicity test results. Jinzhou
Bay is located in the northeastern quadrant of the Bohai Sea,
China (Fig. 1). It has an area of about 150 km2, including 62
km2 of tideflats, with an average depth of 3.5 m [21]. A zinc
smelter located near the mouth of the Wuh River is the largest
source of metals to the bay, although other industrial dis-
charges are significant contributors. Sediments for this study
were collected from seven locations along a 30-km transect
from the mouth of the river to the northeastern portion of the
bay. Total concentrations of divalent metals in sediments col-
lected ranged from 261 to 21,641 u.g/g dry weight (Table 1).
Zinc constituted between 78.5 and 86.5% of the total. Sedi-
ments from stations 1 and 2 also contained low concentrations
of polycyclic aromatic hydrocarbons (PAHs) (<12 u-g/g for
individual PAHs), polychlorinated biphenyls (PCBs) (<0.03
(xg/g for individual congeners), and chlorinated pesticides
(<0.03 (jig/g for any individual pesticide). Concentrations of
total organic carbon (TOC) ranged from 0.11 to 11.5%; AVS,
from 3.0 to 126 jimol/g; SEM, from 2.9 to 374 u.mol/g; and
SEM/AVS ratios, from 0.51 to 8.36. The sum of the IWTUs
for the five divalent metals ranged from no metal detected
(<0.01) to 0.58. The four sediments with the highest metals
concentrations were toxic (>24% mortality) to A, abdita.
However, only the most contaminated sediment contained
>0.5 IWTU of metals and had an SEM/AVS ratio >1.0, which
suggests that metals may not be the principal cause of the
toxicity observed in the other three toxic sediments (Figs 2
and 3). (The vertical dashed lines located on these figures and
the figures that follow at an SEM/AVS ratio of 1.0 or an IWTU
of 0.5 indicate concentrations below which mortality is not
expected. The horizontal dashed line at 24% mortality indi-
cates the approximate mortality at which statistically signifi-
cant effects are expected.)
Belledune Harbor, which receives outfalls from a lead smelt-
er and fertilizer plant, is located in the southwestern portion
of Chaleur Bay, New Brunswick, Canada (Fig. 1). Harbor
sediments are particularly enriched, relative to adjacent areas,
in concentrations of cadmium, lead, and zinc; other metals are
somewhat elevated [22]. The closure of the lobster fishery due
to the elevation of cadmium concentrations in algae, snails,
mussels, scallops, barnacles, crabs, and lobsters has been of
particular concern [23,24]. Sediments for our study were col-
lected by Ponar grab from 10 stations, seven inside and three
outside the harbor Total concentrations of divalent metals in
these sediments ranged from 277 to 2,200 jig/g dry weight,
with 74.7 to 93.5% of the total consisting of lead and zinc
(Table 1). Concentrations of TOC ranged from 0 73 to 1.62%;
AVS, from 5.5 to 102 u,mol/g; SEM, from 1.9 to 18.4 u.mol/g;
and SEM/AVS ratios, from 0.17 to 0.36. The sum of the IWTUs
ranged from <0.01 to 0.58. None of the sediments were toxic
(£24% mortality) to A. abdita, as would be predicted on the
basis of SEM/AVS ratios and IWTUs (Figs 2 and 3)
Foundry Cove, New York, is located on the upper tidal reach
(salinities 0 to 6%o) of the Hudson River immediately south
of Cold Spnng, New York (Fig. 1). A battery plant was the
principal source of the approximately equimolar concentra-
. tions of cadmium and nickel in the sediments, smaller amounts
of cobalt were also discharged [25]. Sediments for our study
were collected by shovel or Ponar grab from 16 stations in
East Foundry Cove. Total concentrations of divalent metals in
our sediments ranged from 169 to 71,140 ng/g dry weight,
with cadmium plus nickel accounting for up to 99.0% of the
metal measured in the most contaminated sediments (Table 1).
Concentrations of TOC ranged from 0.55 to 16.4%, with many
sediments consisting principally of partially decayed marsh
vegetation. Concentrations of AVS ranged from 0.40 to 64.6
(j-mol/g; SEM, from 0.20 to 779 (unol/g; and SEM/AVS ratios,
from 0.04 to 139. The sum of the IWTUs for cadmium and
nickel ranged from 0.08 to 43.5. Molar concentrations of cad-
mium and nickel in the interstitial water were similar. However,
cadmium contributed over 95% to the sum of the IWTUs be-
cause the 10-d LC50 for nickel to A. abdita (2,400 u,g/L) was
67 times that of cadmium (36.0 u.g/L). Sediments with the high-
est dry weight metals concentrations (8,600 to 71,200 u-g/g)
were generally toxic (>24% mortality) to A. abdita. In contrast,
others with similar concentrations (£13,800 u.g/g) were not
toxic (£24% mortality). Sediments with SEM/AVS ratios £1.0
were always nontoxic, whereas only five of 11 sediments with
SEM/AVS ratios >1.0 were toxic (Fig. 2). Sediments with £0.5
IWTU were always nontoxic, those with >2.2 IWTUs were
always toxic, and two of seven sediments with intermediate
IWTUs (5*0.5 to 2.2) were toxic (Fig. 3). Data on chemical
concentrations and N. arenaceodentata mortality in tests with
Foundry Cove sediments are not included in Table 1 because
they have been presented elsewhere by Pesch et al. [13]. Six
of 17 sediments tested with this polychaete had SEM/AVS ratios
<1.0, 16 of 17 sediments had IWTUs <0.5, and none of the
sediments were toxic (Figs. 2 and 3). Absence of toxicity to N.
arenaceodentata in the 11 sediments that contained SEM/AVS
ratios >1.0, five of which were toxic to amphipods, is not
surprising. This polychaete is more tolerant to cadmium and
nickel than the amphipod and can avoid sediments containing
toxic concentrations of these metals [13].
Bear Creek is a tributary of the Patapsco River just east of
Baltimore, Maryland (Fig 1). Sediments from this portion of
Baltimore Harbor are known to be toxic and contain high
concentrations of metals, PAHs, PCBs, and other substances
[26,27] from many municipal and industrial sources. Sedi-
ments used in our study were collected from 14 stations using
a modified Van Veen grab. Total concentrations of divalent
metals in sediments we tested ranged from 44.2 to 2,210 u,g/g
dry weight, with zinc accounting for approx. 75% of the total
concentration (Table 1). Concentrations of TOC ranged from
0.13 to 7.38%; silt and clay, from 4 to 99%; AVS, from 0.40
to 304 n-mol/g; SEM, from 0.64 to 31.0 u,rn°l/g, and SEM/
AVS ratios, from 0 10 to 16.7. Seven of the 14 sediments from
Bear Creek were toxic to A. abdita; these included seven of
the nine sediments with the highest dry weight metals con-
centrations (12.5 to 30.6 u,mol/g). Sediments that were non-
toxic contained metals concentrations from 0.6 to 21.0 u.mol/g.
Both toxic and nontoxic sediments had £0.03 IWTU of metal
(Fig. 3). Given the absence of detectable interstitial water met-
-------
2084
Environ Toxicol Chem 15, 1996
D J Hansen et al
Table I Summary of sediment charactenstics, metals concentrations, and amphipod mortality in sediments from Jinzhou Bay, Belledune
Harbor, Foundry Cove, Bear Creek, Baltimore Harbor, and a salt marsh in Massachusetts
%
Station TOC
Jinzhou Bay
1 11.5
2 2.0
3 0.37
4 050
5 0.26
6 0.11
7 017
• __
Belledune Harbor
1 098
2 .29
3 08
4 .62
5 20
6 12
7 10
8 0.73
9 0.87
10 092
0.99
Foundry Cove
1 102
2 5.20
3 ' 130
4 8 13
5 9.37
6 5.03
7 079
8 136
9 5.82
10 10.9
1 1 0.55
12 164
13 14.6
14 7.18
15 4.76
16 145
0.88
Bear Creek
1 '7.1
2 7.38
3 5.75
4 5.47
5 6 15
6 3.32
7 0.13
8 519
9 4.4
10 0.17
11 4.61
12 0.16
13 419
14 3.14
389
%
Silt/-
clay
_
_
_
_
_
_
_
-
37
68
76
68
52
42
44
23
31
38
94
-
-
_
-
-
-
-
_
-
-
-
-
-
-
_
-
94
99
97
97
96
80
58
7
97
93
5
98
4
94
91
87
Total divalent metals, M-g/g
Cd
182
151
9.1
41
2.6
84
48
01
97
11
15
13
8.0
68
74
1.2
1.9
20
09
38,900
5,920
5,940
9,520
13,100
5,500
66
522
6,230
19
35
163
88
363
21
10
04
88
10
5.4
48
34
48
00
58
42
02
26
0.2
17
13
0.8
Cu
1,200
295
39
47
64
12
13
13
51
101
104
86
58
49
«8
15
19
22
46
143
87
81
106
116
101
28
67
74
31
18
58
44
104
93
26
56
206
228
265
207
151
191
32
254
241
9.4
140
41
139
97
32
Ni
19
24
10
17
86
94
7.4
23
33
38
40
39
37
32
31
29
28
33
21
31,500
5,180
4,160
3,700
7,670
2,340
60
386
3,500
39
45
137
92
227
29
18
26
54
62
60
56
38
50
20
49
53
4.9
51
2.8
47
39
28
Pb
3,240
608
113
142
22
17
18
32
498
955
1,140
906
539
463
689
94
131
149
32
194
157
93
135
156
357
10
98
113
27
61
88
48
127
177
48
35
195
212
209
175
162
173
30
250
274
9.4
162
74
128
88
13
Zn
17,000
4,440
737
1,320
221
300
239
67
401
783
896
874
581
503
843
137
178
192
122
403
297
278
313
356
303
80
219
246
101
65
231
142
317
234
124
160
1,576
1,700
1,140
958
566
1,000
36
1,110
978
69
617
. 43
459
346
141
2|i.mol
297
77.3
12.7
223
38
5 1
41
1.4
100
189
21.7
19.9
13 1
11.3
179
3.3
42
46
3.1
893
147
130
155
256
96.7
33
161
121
30
24
86
55
142
6.6
2.9
40
29.3
318
237
197
12.5
20.1
0.6
23 1
21.0
1.3
13.3
1.4
10.6
79
3.5
SEM
funol/g
380
647
11.0
18.8
3.04
3.94
297
1.42
6.73
17.5
184
16.6
11.2
10.3
17.0
2.00
3.20
189
272
779
93.5
105
136
168
523
0.67
864
86.5
1.27
0.42
191
1.94
54
0.56
0.20
0.22
29.3
31.0
20.4
176
17.3
16.7
0.64
23.2
19.5
127
119
074
9.96
679
285
AVS
limol/g
447
126
17.8
366
3.02
542
356
122
27.2
803
102
966
474
385
561
5.54
167
11.3
16.8
5.60
18.0
12.2
26.8
64.6
124
0.44
202
24.7
2.62
0.41
040
0.69
37 1
13.1
1.38
12.0
268
304
76.1
70.1
45.3
46.6
0.40
146
892
0.45
50.0
0.40
720
0.40
9.75
SEM/AVS
ratio
85
052
062
051
10
073
0.84
Oil
025
022
0.18
017
024
0.27
0.30
0.36
019
0.17
0.16
139
5.18
8.6
5 10
2.60
4.20
1 55
0.43
3.51
048
102
483
2.80
014
0.04
0.14
0.02
Oil
0 10
0.27
0.25
038
0.36
160
0 16
0.22
282
0.24
187
1.38
167
0.29
IWTU
058
<0.01
<001
0.17
003
<001
<001
<001
003
006
018
0.07
0.01
0.04
058
<001
0.06
<001
<001
435
9.21
3.33
217
1.81
109
042
045
148
030
030
142
044
1.12
1.60
0.08
0.02
003
0.03
0.03
0.02
003
003
003
002
003
0.03
003
002
003
003
003
%
Mor-
tality
100
90
38
30
2
8
2
2
8
12
12
12
15
12
12
8
18
5
10
82
32
52
20
80
35
18
18
, 15
8
18
12
15
20
18
20
10
82
95
85
85
68
40
5
92
12
5
2
5
18
2
0
Salt marsh, Massachusetts
1 3.22
2 156
3 036
4 048
5 200
6 228
7 249
8 255
9 061
10 198
11 354
172
38
1.9
25
615
209
236
485
37
7.8
609
21
06
01
01
2.0
58
16
10
01
04
77
147
144
128
32
300
155
823
1,423
152
359
851
25
49
9.2
59
36
108
113
72
12
29
31
409
67
88
28
96
61
143
304
63
137
128
517
398
66
33
463
1,100
930
1,480
235
455
2,300
12.6
9.51
3.60
1.25
129
214
298
47.8
651
13.8
499
8.51
642
132
094
969
181
251
318
301
708
233
864
142
0.44
059
419
124
166
69.1
0.50
11.50
141
0 10
5.00
304
1.60
002
146
1.51
046
60
062
1.65
069
151
053
070
043
0.59
041
037
049
041
046
13
40
13
8
18
0
10
23
5
15
5
-------
SEM/AVS rauo and interstitial metals Field sediment toxicity prediction
Table 1. Continued
Environ. Toxicol. Chem. 15, 1996 2085
Station
12
13
14
15
16
17
18
19
20
21
22
•
•
•
a
%
TOC
_
4.39
1.13
2.74
2.45
0.51
1.18
3.13
0.63
2.03
0.13
1.24
1.47
0.99
0.99
%
Silt/ —
clay
164
26
106
33.2
123
4.5
12.9
27.5
5.1
174
1.5
17.2
10.9
94
94
Total divalent metals, p.g/g
Cd
1.7
2.3
0.7
0.0
2.6
0.6
0.8
0.4
0.5
0.5
0.2
4.0
0.5
0.1
0.1
Cu
508
572
236
639
348
184
179
47.5
93
89
30
8.2
15
56
61
Ni
42
43
20
28
28
12
12
11
79
9.5
4.2
38
40
25
28
Pb
148
192
63
226
97
39
52
. 28
20
29
6.2
81
83
35
39
Zn
535
629
272
685
410
172
188
86
103
152
42
25
28
142
164
Sjtmol
17.6
20.3
8.53
22.1
12.7
5.92
6.17
2.40
3.27
4.03
1.21
0.62
0.79
3.6
4.1
-. -SEM
jimol/g
11.5
14.0
6.12
16.2
724
4.59
3.74
172
210
2.55
0.73
040
050
2.84
2.55
AVS SEM/AVS
limol/g ratio
85.4
19.1
18.6
2.35
21.8
5.86
38.9
117
18.0
18.0
3.24
124
400
15.5
161
013
073
0.33
6.92
033
0.78
0.10
0.15
0.11
0.14
022
0.03
0.12
018
0.16
IWTU
0.67
1.38
058
047
046
045
0.42
035
0.37
0.40
043
031
031
0.42
0.21
%
Mor-
tality
18
10
5
8
5
13
10
8
5
3
3
10
2
5
0
• Reference sediments from Long Island Sound, lower Narragansett Bay, or a clean site nearby.
al, it is not surprising that SEM/AVS ratios for sediments from
Bear Creek were not related to sediment toxicity; i.e., five
sediments having SEM/AVS ratios >1.0 were not toxic, and
seven of the sediments having SEM/AVS ratios <1.0 were
toxic (Fig. 2). Most toxic sediments released visible oil sheens
when stirred, suggesting that PAHs may ultimately prove to
be a source of the observed sediment toxicity. These obser-
vations support the conclusion that toxicity observed in Bear
Creek sediments was not metal-associated.
JffgHOU (A.O.)
80
60
40
2CV
0.1
A
A
A
A 4i
31 01 1
10 100 101
FOUNDRY COVE (A.cO
80
(SO-
40
20
n
: A A
A
. %
A* i \**
BEUEDUNEfAO)
SO
40
0
^
31 01 1 10 100 ID
FOUNDRY COVE (Ro)
0.01 01 1 10 100 1000
80
60
40
20
0
0
•
• 4 X : A
31 0 1 1 10
A
100 101
The salt marsh, containing a small tidal creek less than 500
m long (Fig. 1), is near Fairhaven, Massachusetts, on the west-
ern side of Buzzards Bay. The creek is divided by a hurricane
barrier into an upper section of low salinity and a lower section
with higher salinity. A metal products manufacturer was the
principal source of metals in the sediments. Sediments for our
study were collected by plastic scoops from 23 locations, 10
from the upper side of the hurricane barrier and 13 from the
I
JINZHOU (A.o j
BELLEDUNE (AaJ
80
60
40;
?o
n'
01
A
A
31 01 1 10 1C
80
60
40
20
n1
n o
.
A ^ A A
91 01 1 10 10
FOUNDRY COVE
FOUNDRY COVE (Ho)
80
60
40
2d
0
A A
A
*. . A
Aft JA
31 01 1 10 1(
IUU
80
60
40
20
M o
A
1x1*1 -
01 01 1 10 1C
BEAR CREEK (AA)
80
60
40-
20
n
AA
A
A '
A A
V 'A A
001 01
10 100 1000
SEM/AVS
100
80-
60-
40
20-
SALT MARSH (A.o.t
001 01
10 100 1000
SEM/AVS
Fig. 2 Percent mortality of the amphipod Ampehsca abdita (A a)
and the polychaete Neanthes arenaceodentata (Na)asa function of
SEM/AVS ratios in sediments from Jinzhou Bay, Belledune Harbor,
Foundry Cove, Bear Creek, and a salt marsh in Massachusetts.
f
SALT MARSH (A.Q.l
80
60
40-
20
0
A |
A*
A
A
4 :
31 01 1 10 1(
100
80
60
40
20
» 0
' A
A*/ *
31 01 1
10 1C
Interstitial Water Toxic Units
Interstltcd Water Toxic Units
Fig. 3 Percent mortality of the amphipod Ampehsca abdita (A.a.)
and the polychaete Neanthes arenaceodentata (N.a.) as a function of
IWTUs of cadmium, copper, lead, nickel, and zinc in sediments from
Jinzhou Bay, Belledune Harbor, Foundry Cove, Bear Creek, and a
salt marsh in Massachusetts.
-------
2086
Environ Toxicol Chem 15, 1996
D J Hansen et al
Table 2 Summary of sediment characteristics, metal concentrations, and amphipod (Hyalella azteca
= H.a ) or ohgochaete (Lumbriculus variegatus = L.V) mortality in freshwater sediments from
Steilacoom Lake, Keweenaw Watershed, Turkey Creek, and Foundry Cove
SEM
Station Day 0
Steilacoom Lake
1 0.89
2 3.05
3 1.93
4 284
5 125
6 0.60
7 0.66
8 1.33
9 195
10 1.27
11 2.90
c
(junol/g)1
Day 10
1.36
200
185
3.78
356-
202
1 11
2.68
3.91
2.01
2.05
-
AVS(|
Day 0
4.01
2.89
0.30
194
2.06
4.16
1.48
1.60
0.39
2.17
0.65
-
M-mol/g)
Day 10
5.65
102
<0.02
0.92
0.29
0.81
1.44
<005
<003
1.60
0.41
-
Average
SEM/AVS
023
151
>49.5
279
645
132
0.61
>27.2
>67.5
093
411
-
IWTU"
Ha (L.v)
<022
<022
<0.22
<022
<0.22
<0.22
<022
<022
<0.22
<0.22
<0.22
-
% Mortality
Ha. (L.v)
0
5
0
10
5
15
0
0
10
5
0
0
Keweenaw Watershed
1
2
3
4
5
6
7
8
9
10
11
c _
Turkey Creek
vl
2
3
4
5
6
7
c _
Foundry Cove
1 789
2 66.4
3 43.8
4 92.2
5 50.1
6 176
7 029
8 9.23
9 92.9
10 0.31
11 038
12 3.02
13 2.02
14 3.34
15 1.78
16 0.01
4.68
264
626
15.1
19.6
5.65
8.49
174
281
0.36
10.8
-
672
515
854
476
501
82.1
945
' -
703
115
915
106
74.8
210
0.50
14.0
58.5
0.31
0.52
2.20
1.79
7.86
0.44
0.05
_
-
-
-
-
-
_
-
-
_
-
-
_
-
-
-
-
-
-
-
3.12
9.39
15.3
10.4
7.59
46.9
0.09
5.12
6.73
0.92
0.39
1.92
1.18
200
9.07
0.94
11.6
<0.006
003
0.08
006
0.01
0.12
0.01
0.46
0.09
0.02
-
28.1
52.2
30.1
48.4
44.2
382
782
-
5.65
13.8
136
19.0
9.83
31.2
0.10
5.20
14.7
1.17
0.16
6.15
0.64
10.0
9.92
0.49
0.40
>4,440
2,090
189
332
565
707
17,400
61 1
4.0
674
-
2.39
099
284
098
1 13
215
1.21
-
189
7.69
478
7.24
7.11
5.25
4.11
2.25
8.90
032
2.11
0.97
225
0.31
0.12
005
041
1 19
9.97
4.52
2.81
171
5.45
196
319
0.52
3.10
-
0.73
0.44
1 11
038
183
1.31
0.49
-
188 (0.50)
A\5 (1.54)
946 (2.44)
7.29 (0.61)
4.58 (0.18)
1193(0.55)
- (3 29)
77.3 (649)
3 16(153)
2.43 (0.27)
- (0.54)
32.7 (1.35)
110 (1.13)
2 03 (3 02)
0 40 (0.32)
- (0.38)
20
55
100
100
85
75
95
100
95
35
90
10
20
15
45
0
5
20
45
5
100 (87)
100(0)
100(0)
100(0)
80(0)
100(0)
100(0)
40(0)
100 (24)
0(0)
100(0)
60(-)
80(0)
20(0)
0(0)
0(0)
* Simultaneously extracted metal is SEM copper for Steilecoom Lake and Keweenaw Watershed, SEM
zinc for Turkey Creek, and SEM cadmium plus nickel for Foundry Cove
" Interstitial water toxic units are calculated using 10-d water-only LCSOs for H. azteca of 2 8 ng/L for
cadmium, 31 jtg/L for copper, 780 p-g/L for nickel, and 436 (jig/L (hardness 330 mg/L) for zinc, and
for L. variegatus of 158 (ig/L for cadmium and 12,200 (ig/L for nickel.
c Reference sediments from uncontammated West Bearskin Lake, Minnesota, USA.
lower section. Total concentrations of divalent metals in these
sediments ranged from 82.6 to 3,320 jtg/g dry weight (Table
1). Zinc and copper were the principal metals on a dry weight
basis in these sediments. Concentrations of TOC ranged from
0.13 to 4.39%; silt and clay, from 1.5 to 61.5%, AVS, from
0.44 to 419 (tmol/g, SEM, from 0.73 to 31.8 jimol/g; and
SEM/AVS ratios, from 0.10 to 6.90. Only one of 23 sediments
from the salt marsh was toxic to A. abdita (Figs. 2 and 3).
-------
SEM/AVS ratio and interstitial metals Field sediment toxicity prediction
Environ Toxicol Chem 15, 1996 2087
100
80
60
40
20
10
100
Copper (ug/L)
1,000
Fig 4. Toxicity of copper to Hyalella azteca versus copper concen-
trations in a water-only exposure (open symbols) and interstitial water
in sediment exposures using Keweenaw Watershed sediments (closed
symbols) (Modified from Ankley et al. [15] )
The SEM/AVS ratio for the toxic sediment was 4.49, and the
IWTU was 1.51. All other sediments had SEM/AVS ratios
£1.0 and IWTUs £0.64 and were nontoxic.
Freshwater field sites
Description of field sites and toxicity test results. High con-
centrations of copper in sediments from Steilacoom Lake, Wash-
ington, originated principally from attempts to control aquatic
vegetation using copper sulfate Copper SEM concentrations in
sediments from 11 locations tested ranged from 0.60 to 3.91
fimol/g (38 to 248 pig/g); AVS, from <0.02 to 5 65 junol/g;
and SEM/AVS ratios, from 0.23 to >67.5 (Table 2) [15]. Eight
of the 11 sediments tested had SEM/AVS ratios > 1.0. No
copper was detected in interstitial water (IWTU < 0.22), and
no sediments were toxic to H. azteca Absence of toxicity in
sediments having SEM/AVS ratios >1.0 and the lack of de-
tectable copper in the interstitial water are likely consequences
of the presence of other sediment binding phases [15].
In contrast, 10 of 11 sediments from Keweenaw Watershed,
Michigan, were toxic to H. azteca [15]. Mining-derived copper
concentrations in sediments ranged from 0.36 to 174 |imol/g
(22.9 to 11,000 jtg/g); AVS, from <0.006 to 11.6 u.mol/g; and
SEM/AVS ratios, from 0.4 to 17,400 (Table 2). The one sed-
iment not toxic to amphipods had 0.41 toxic units of copper
in interstitial water and an SEM/AVS ratio of 0.40. Toxic
sediments had 0.52 to 19.6 IWTUs of copper and SEM/AVS
ratios a4.0. Acid-volatile sulfide concentrations in the 10 toxic
sediments were extremely low (<0.01 to 0.46 u,mol/g) with
comparatively high copper concentrations (0.36 to 1.74 u,mol/
g); nine SEM/AVS ratios were a61. Amphipod mortality in
response to copper concentrations in water-only tests was sim-
ilar to amphipod mortality as a function of interstitial water
copper concentration in sediment tests (Fig 4 [15]). The 10-d
LC50 (95% confidence limits) for amphipods exposed to cop-
per in water-only tests did not differ from the LC50 on the
basis of interstitial dissolved copper concentrations and am-
phipod mortality from tests with Keweenaw sediments, 31 (28
to 35) versus 28 (21 to 38) jtg/L [15].
Sediments from Turkey Creek, Missouri, contained high and
relatively uniform concentrations of zinc (47.6 to 94.5 u,mol/g,
3,110 to 6,180 ^g/g) and AVS (28.1 to 78.2 junotfg, Table 2)
originating from stnp mine tailings. Therefore, SEM/AVS ra-
tios (0.98 to 2.84) and IWTUs (0.44 to 1.83) varied little in
the seven sediments tested. The two sediments having SEM/
AVS ratios £1.0 were nontoxic and had ==0 44 IWTU of zinc.
The SEM/AVS ratios of the five remaining sediments ranged
from 1.13 to 2 84, and IWTUs ranged from 0 49 to 1.83. Two
of these sediments were toxic.
Sediments from Foundry Cove, New York, tested with salt-
water A abditaandN arenaceodentata were also tested using
the freshwater amphipod H. azteca and the ohgochaete L.
variegatus by Ankley et al. [14]. Sediments contained ap-
proximately equimolar concentrations of cadmium and nickel,
with the sum of the SEM concentrations of these metals from
freshwater tests ranging from <0.01 to 789 (unol/g; AVS, from
0.09 to 46.9 u,mol/g; and SEM/AVS ratios, from 0.05 to 189
(Table 2). Four of five sediments with SEM/AVS ratios £1.0
were not toxic to amphipods, while all sediments having SEM/
AVS ratios >1.0 were toxic to amphipods. Only the two sed-
iments with the highest SEM/AVS ratios (8.90 and 189) were
toxic to the oligochaete; 14 of 16 sediments were not toxic to
oligochaetes. Sediments with IWTUs &3.16 were toxic to am-
phipods; when 0.40 to 2.43 IWTUs were present, no toxicity
was observed. Interstitial molar concentrations of nickel al-
most always exceeded those of cadmium by one to three orders
of magnitude (data not shown). However, cadmium was most
likely the cause of both amphipod and oligochaete mortalities
because cadmium is over 250 times more toxic than nickel to
H. azteca, as evidenced by the 10-d water-only LC50 of 2.8
(ig/L for cadmium and 780 (ig/L for nickel. Similarly, cad-
mium is about 80 times more toxic to L. variegatus than nickel,*
with 10-d water-only LCSOs of 158 ftg/L for cadmium and
12,200 M-g/L for nickel. Cadmium contributed from 88.6 to
99.9% of the total IWTUs of metals
DISCUSSION
Saltwater field sites
Bulk metals concentrations in saltwater sediments cannot be
used to causally relate metals concentrations to the acute re-
sponse of amphipods and polychaetes (Fig. 5). Mortality of
amphipods in 70 sediments from five saltwater locations or
polychaetes in 16 sediments from Foundry Cove was not re-
lated to the sum of the molar concentrations of cadmium,
copper, lead, nickel, and zinc on a dry weight of sediment
basis. Sediments having dry weight metals concentrations from
9.50 to 885 u,mol/g from 17 stations in Jinzhou Bay, Bear
Creek, Foundry Cove, and the marsh in Massachusetts were
toxic (>24% mortality). In contrast, dry weight metals con-
centrations from 0.20 to 885 u,mol/g were nontoxic (£24%
mortality), an overlap of two to three orders of magnitude in
metals concentration.
Normalizing metals concentrations in these sediments using
SEM/AVS ratios, without insight into mortality caused by
co-occurring toxic substances, also does not permit accurate
causal predictions of metal toxicity in sediments from the field
(Fig. 6). Of the 59 sediments with SEM/AVS ratios £1.0 (Table
1) from the five locations, 49 (83%) were not toxic, and 10
(17%) were toxic. These 10 toxic sediments were from Imzhou
Bay and Bear Creek, both highly industrial locations. Of the
37 sediments with SEM/AVS ratios >1.0, only seven were
toxic. Absence of toxicity when SEM/AVS ratios are > 1.0 has
commonly been observed However, when SEM/AVS ratios
are £1.0, toxicity has been observed in only one of 92 sedi-
ments spiked with metals [10] and one of 15 sediments from
freshwater field sites [14,15] (Table 3). For these two sediments
the true SEM/AVS ratios may have been >1 0, as concentra-
tions were within the precision expected in AVS and SEM
-------
2088 Environ Toxicol Chem. 15, 1996
lOOi
D J Hansen et al
80
60
40
20
«A
10
100
1000
Bulk Metal (umol/g dry wt)
Fig 5 Percent mortality of the amphipod Ampehsca abdita (A.a ) and the polychaete Neanthes arenaceodentata (Ma.) in sediments from
saltwater locations in a salt marsh (A = A.a.), Belledune Harbor (4 = A.a.), Bear Creek (• = A a ), Foundry Cove (• = A.a., * = Na ), and
Jinzhou Bay (+ = A.a ) as a function of the sum of the concentrations of cadmium, copper, lead, nickel, and zinc in micromoles divalent metal
per gram dry weight sediment
analyses and they had 0.4 and 1.4 IWTUs of metal. Given the
fact that field sediments from highly industrialized locations
contain many substances other than metals and are often toxic,
nonmetals-associated toxicity should always be suspected. If
toxic sediments have SEM/AVS ratios sl.O, we might suspect
the cause to not be metals; with SEM/AVS ratios > 1.0, toxicity
may or may not be related to metals.
Metals concentrations expressed on a sum of the IWTU basis
(Fig. 7) can provide insight that in part may explain apparent
anomalies between SEM/AVS ratios and the observed toxicity
of these sediments. In spite of the presence of very high dry
weight metals concentrations, 56 of 70 sediments had <0.5
IWTU of metal. Of the 10 toxic sediments having SEM/AVS
ratios <1.0, none had >0.5 IWTU of metal. This suggests that
metals are unlikely to be the cause of the toxicity. Seven of
these sediments (most of which released oil when agitated)
were from Bear Creek, and the rest were from Jinzhou Bay.
The absence of toxicity in many sediments having SEM/AVS
ratios >1.0 is understandable because most (66.7%, or 12 of
18) of these nontoxic sediments had <0.5 IWTU of metal. Of
the seven toxic sediments having SEM/AVS ratios >1.0 (one
each from Jinzhou Bay and the salt marsh and five from Foun-
dry Cove) all had >0.5 IWTU of metals. Furthermore, inter-
stitial metal concentrations are likely to overestimate the con-
centration of available metal because of differences in metal
form, greater binding to dissolved organic carbon or hgands
in interstitial water [28], release of bound metal during sam-
pling or analytical procedures [18], or organism avoidance of
metal exposure [13].
We believe it is inappropriate to include in this article data
from locations having sediments with toxicities almost cer-
tainly not due to metals. This decision is additionally justified
100
80
6 60
0.01
100
1000
SEM/AVS
Fig 6 Percent mortality of the amphipod Ampehsca abdita (A a ) and the polychaete Neanthes arenaceodentata (N.a ) in sediments from
saltwater field locations as a function of the ratio of the sum of the molar concentrations of cadmium, copper, lead, nickel, and zinc simultaneously
extracted (SEM) with AVS to the molar concentration of AVS (SEM/AVS ratio) (See Fig 5 legend for definitions of symbols.)
-------
SEM/AVS ratio and interstitial metals- Field sediment toxicity prediction
Environ Toxicol Chem. 15, 1996
2089
Table 3. Accuracy of prediction of the toxicity of sediments from using saltwater and freshwater
field locations, spiked-sediment tests, and combined field and spiked sediment tests
as a function of SEM/AVS ratios, IWTUs, and both SEM/AVS and IWTUs
Study
type
Saltwater field
Freshwater field
Laboratory spike
(freshwater and
saltwater)
All
% of Sediments
Parameter
SEM/AVS
IWTU
SEM/AVS, IWTU
SEM/AVS
IWTU
SEM/AVS, rWTU
SEM/AVS
IWTU
SEM/AVS, IWTU
SEM/AVS
IWTU
SEM/AVS, IWTU
Value
£1.0
>1.0
<0.5
2=0.5
s=l 0, <0.5
> 1.0,2:0.5
£1.0
>1.0
<0.5
>0.5
£1.0, <0.5
->1.0, s=0.5
£1.0
>1.0
<05
s:05
£1.0, <0.5
>1.0, £0.5
£1.0
>1.0
<0.5
3:0.5
£1.0, <0.5
> 1.0, 2:05
n
42
31
59
15
39
11
15
48
20
38
10
34
92
83
88
76
76
65
149
162
167
129
120
110
Nontoxic1
100.0
80.6
100.0
533
1000
45.5
93.3
47.9
950
421
100.0
294
98.9
26.5
98.9
22.4
987
12.3
98.7
43.2
98.8
31.8
99.2
20.9
Toxic*
0.0
19.4
00
467
0.0
54.5
6.7
52 1
50
57.9
0.0
70.6
1.1
735
1.1
77.6
1.3
87.7
13
56.8
1.2
68.2
08
79.1
• Nontoxic sediments, <24% mortality, toxic sediments, >24% mortality.
because in experiments with metal-spiked sediments [10], only
one of 76 sediments having rWTUs <0.5 and SEM/AVS ratios
< 1.0 were toxic. Therefore, data from Bear Creek and Jinzhou
Bay are not included in the text, figures, and tables that follow.
These data were included above to demonstrate the value of
both SEM/AVS ratios and IWTUs in discriminating between
metals-associated and nonmetals-associated toxicity in sedi-
ments and to demonstrate the point that toxicity in field sed-
iments is common even when SEM/AVS ratios are < 1.0 but
that when this occurs it is unlikely metals-related. For the data
from saltwater field locations in Belledune Harbor, the salt
marsh and Foundry Cove, all 42 sediments with SEM/AVS
ratios =£1.0 were not toxic (Fig. 8, Table 3). Of the 31 sedi-
ments that had SEM/AVS ratios >1.0, only six sediments
(from Foundry Cove and the salt marsh) were toxic, and all
had >0.50 IWTU (Table 3). Of the 25 nontoxic sediments with
SEM/AVS ratios >1.0, 71.4% (10 of 14) of the sediments
tested with amphipods and 90.9% (10 of 11) of the sediments
100
80
60
40
20l
0.01
-*-
0.1 1
Interstitial Water Toxic Unit
10
100
Pig. 7 Percent mortality of amphipod Ampehsca abdita (A.a ) and the polychaete Neanthes arenaceodentata (N.a ) as a function of IWTUs of
metals in sediments from saltwater field locations. The IWTUs are the sum of metal-specific interstitial water concentrations per 10-d LC50 for
cadmium, copper, lead, nickel, and zinc Interstitial water concentrations with nondetectable metal are plotted at 0.01 IWTU. (See Fig 5 legend
for definitions of symbols )
-------
2090 Environ Toxicol Chem 15, 1996
lOOi
'o
o
80--
60-
40-
20-
D.J Hansen et al
D.01
0.1
10
100
1000
SEM/AVS
Fig 8 Percent mortality of the amphipod Ampelisca abdita (A.a.) and the polychaete Neanthes arenaceodentata (N.a ) in sediments from three
saltwater field locations in a salt marsh (A = A.a ), Belledune Harbor (• = A.a.), and Foundry Cove (• = A a., * = N.a ) as a function of
the SEM/AVS ratio
tested with polychaetes had <0.5 IWTU, in part explaining
the absence of toxicity.
Saltwater and freshwater field sites combined
Metals concentrations in sediment interstitial water from
freshwater sites suggest that metals contributed to the observed
mortalities of amphipods and oligochaetes (Fig. 9). Therefore,
all available freshwater data are included in Figures 10 through
15. For sediments with IWTUs =s0.5, 57.9% of 38 sediments
were toxic, 20 of 26 for amphipods and 2 of 12 for oligochaetes
(Table 3).
The pattern of organism response to metals normalized on
an SEM/AVS basis is similar for saltwater (Fig. 8) and fresh-
water (Fig. 10) sediments. Therefore, data in both figures were
pooled (Fig. 11) to illustrate the overall utility of the SEM/
AVS normalization to explain metals availability in field sed-
iments. The absence of toxicity in all but one sediment having
SEM/AVS ratios £1.0 from all field sediments, from locations
where metals are the principal contaminant, is important given
that total divalent metals concentrations for saltwater sedi-
ments or SEM for freshwater sediments ranged from 43.8 to
13,800 ftg/g for tests with saltwater or freshwater amphipods,
from 170 to 71,200 u,g/g for polychaetes, and from 170 to
76,800 u,g/g for oligochaetes (Tables 1 and 2). Fifty-six of 57
(98.2%) of these freshwater and saltwater sediments having
SEM/AVS ratios £1.0 were not toxic to sensitive organisms.
In the toxic sediment with SEM/AVS <1.0, the SEM/AVS
ratio was 0.97 (Table 3, Fig. 11). For field sediments having
SEM/AVS ratios >1.0, 31 of 79 (39.2%) were toxic (Table
3) Therefore, we believe that SEM/AVS ratios sl.O can ac-
curately predict field sediments likely to not be acutely toxic
due to metals. Use of an SEM/AVS ratio > 1.0 alone to predict
£•
D
O
IUU-
80
60
40
20
<
<
<
rv
LJ \J LJVJ-I UJ LJ
0 O
*
D D
O
D
O
A A
D
O
... ........ ... .-Jr. .......................
0 A AD
> A
I . — <*>
-------
SEM/AVS ratio and interstitial metals Field sediment toxicity prediction
lOO
80
g 60}
"S 40
2
20-
0-
0.01 0.1
Environ Toxicol Chem 15, 1996 2091
DO ^
O O
ao o
tv!"A-A-/y-iMr — rt-/s — , , .
10
100 1000 10000 100000
SEM/AVS
Fig. 10 Percent mortality of the amphipod Hyalella azteca (Ha) and the ohgochaete Lumbnculus vanegatus (L.v) in sediments from four
freshwater locations as a function of the SEM/AVS ratio. (See Fig 9 legend for definitions of symbols.)
sediment toxicity is useful but less accurate than predicting
absence of toxicity (Table 3), as would be expected on the
basis of partitioning theory and organism/sediment interac-
tions. In both pxic and anoxic transition zones occupied by
organisms, other sediment binding phases, metal form, and
avoidance behavior of organisms can limit metal availability,
exposure, and toxicity.
Field sites and spiked sediments combined
The utility of metals concentrations normalized by dry
weight, IWTUs, or SEM/AVS ratios to explain the bioavail-
ability of divalent metals and permit prediction of sediment
toxicity is summarized in Figures 12 through 14 and Table 3.
The figures and table are compilations of all available data
from 10-d lethality tests for which mortality, IWTUs, and
SEM/AVS ratios are known from experiments with sediments
most certainly toxic only because of metals. They include
sediments from saltwater field sites, freshwater field sites
(summarized in this article), or spiked with individual metals
or metal mixtures (summarized by Berry et al. [10]). The re-
lationship between benthic organism mortality in 10-d sedi-
ment lethality tests and bulk metals concentrations in spiked
and field sediments is not useful to causally relate metal con-
centrations to organism response (Fig 12). The overlap among
bulk metals concentrations which cause no toxicity and those
which are 100% lethal is almost four orders of magnitude.
Sediments having <0.01 jjumol of metal/g dry weight are all
reference or control sediments.
The toxicities observed when sediment concentrations are
normalized on an IWTU basis are typically consistent with the
toxic unit concept. That is, if IWTUs are £1.0, sediments
should be lethal to £50% of the organisms exposed; significant
mortality probably should be absent at <0.5 IWTU (Fig. 13).
Of the spiked and field sediments evaluated which had IWTUs
<0.5, 98 8% of 168 sediments were nontoxic (Table 3). For
all sediments having IWTUs >0.5, 68.2% of 129 sediments
iuu-
80-
g 60-
*i
0
B 40-
S
20
0-
0.
0
A • Jfc°$A JM
00
•*• '
a
e
A
0 A
.._-....£.-.----- -------
• A * Z*^> ' ^» *^ A ^
M |LW>WLJ V U"
31 0.1 1
10 100 1000 IOC
SEM/AVS
Fig 11. Percent mortality of the amphipods Ampelisca abdita (A.a ) and Hyalella azteca (H a ); ohgochaete Lumbnculus vanegatus (L.v). and
polychaete Neanthes arenaceodentata (N a.) exposed to sediments from saltwater locations (solid symbols) as a function of the SEM/AVS ratio.
(See Fig 8 and 9 legends for definitions of symbols )
-------
2092 Environ Toxicol Chem. 15, 1996
D.J Hansen et al
80
to
20
* £. . •***>•**IJ. *
• • • » • A • m^m
m* • • ••
0-1—'
0001
001 01
10
100
1000
Total Metal or SEM (pmol/g)
Fig 12 Percent mortality as a function of total dry weight metals
concentrations of the oligochaete Lumbriculus variegalus (L.v ), pol-
ychaetes Capitella capitata (C.c.) and Neanthes arenaceodentata
(N.a ), harpacticoid Amphiascus tenuiremis (A.t), amphipods Am-
pelisca abdita (A a) and Hyalella azteca (H.a.); and snails Helisoma
sp (H.sp.) exposed to sediments from saltwater field locations, fresh-
water field locations, and sediments spiked with individual metals or
mixtures
were toxic (Table 3). Given the effect on toxicity or bioavail-
ability of the presence of dissolved organic carbon or ligand-
associated metal in interstitial water, water quality (hardness
or salinity), and organism behavior, it is not surprising that
many sediments having IWTUs >O.S are not toxic.
Organism response in sediments whose concentrations are
normalized on an SEM/AVS basis is consistent with metal-
sulfide binding on a mole to mole basis as first described by
Di Toro et al [6] and in recommendations for assessing the
bioavailabihty of metals in sediments proposed by Ankley et
al. [29]. Sediments spiked with metals and field sediments from
saltwater and freshwater locations with SEM/AVS ratios sl.O
were uniformly (98.7% of 149 sediments) nontoxic (Fig 14,
Table 3). The majority (56.8%) of 162 sediments having SEM/
AVS ratios >1.0 were toxic. Use of both IWTUs and SEM/
AVS ratios did not improve the accuracy of predictions of
sediments that were nontoxic (99.2% of 120 sediments; Table
3). However, it is noteworthy that toxic sediments were pre-
dicted with 79.1% accuracy in 1 10 sediments when both SEM/
AVS >1.0 and IWTUs SO 5 were used jointly as decision
parameters (Table 3). This approach is, therefore, very useful
in identifying sediments of concern
Because AVS can bind divalent metals on a mole to mole
basis, and other metals in proportion to their molar concen-
trations, normalizing concentrations of SEM in sediments from
the field as the difference of SEM - AVS instead of the con-
ventional SEM/AVS ratio can provide important insight into
the extent of available additional sulfide binding capacity or
the extent to which AVS binding has been exceeded (Fig. 15).
Furthermore, absence of organism response when AVS binding
is exceeded can indicate the potential magnitude of importance
that other binding phases may have in controlling bioavail-
abihty. This insight into the additional binding capacity of
AVS and other sediment phases and the magnitude of ex-
ceedance of binding are important advantages for normaliza-
tion of the concentration of metals in sediments on an AVS
basis over that of interstitial water concentration. For most
nontoxic saltwater and freshwater field sediments we have
tested, 1 to 100 u.mol of additional metal would be required
to exceed the sulfide binding capacity; i.e., SEM — AVS =
g
0.01
10
100
1000
10000
Interstitial Water Toxic Unit
Fig 13. Percent mortality as a function of IWTUs of metals for saltwater and freshwater benthic species, including the oligochaete Lumbriculus
variegatus (L.v ); polychaetes Capitella capitata (C c) and Neanthes arenaceodentata (N a ), harpacticoid Amphiascus tenuiremis (A t), am-
phipods Ampehsca abdita (A.a ) and Hyalella azteca (H.a ), and snails Helisoma sp (H.sp.) exposed to sediments from saltwater field locations
(salt marsh [A = A.O.], Belledune [* = A.a ], and Foundry Cove [• = A a., •*• = N a ]), freshwater field locations (Foundry Cove [D = Ha.,
if = L.V.], Steilacoom Lake [•$ = Ha], Keweenaw Watershed [O = H.a ], and Turkey Creek [A = H.a ]), and sediments spiked with individual
metals or mixtures (Saltwater Cd, O = A.a., Cu, © = A.a , Ni, © = A.a ; Pb, O = A.a., Zn, © = A a , mix, O = A a , Cu, O = Cc , Pb, O
= C c.; Zn, © = C c , Cd, © = Mo., Ni, O = N.a., Cd, •*• = A t Freshwater Cd, (D = H.s; Cd, © = L.V.) Interstitial water concentrations
with nondetectable metals are plotted at 0 01 IWTU
-------
SEM/AVS ratio and interstitial metals Field sediment toxicity prediction
100
Environ Toxicol Chem 15, 1996 2093
80
g
1000
100CX)
10
SEM/AVS
Fig. 14. Percent mortality as a function of the SEM/AVS ratio for saltwater and freshwater benthic species including the oligochaete Lumbriculus
variegatus (L.V.); polychaetes Capitella capitata (C c) and Neanthes arenaceodentata (N a.), amphipods Ampehsca abdita (A.a.) and Hyalella
azteca (H.a ); and snails Helisoma sp. (H.s.) exposed to sediments from saltwater field locations, freshwater field locations, and sediments spiked
with individual metals or mixtures. (See Fig 13 legend for definitions of symbols )
-1 to -100 n-mol/g. In contrast, most toxic field sediments
contained 1.0 to 1,000 }i,mol of metals beyond the binding
capacity of sulfide alone. Data on nontoxic field sediments
whose sulfide binding capacity is exceeded (SEM — AYS is
>0.0 |unol/g) provide the best indication of magnitude and
importance of nonsulfidic binding phases. This is particularly
true for locations such as Steilacoom Lake and the Keweenaw
Watershed, where AVS concentrations were low, resulting in
high SEM/AVS ratios with little difference between SEM con-
centrations and sulfide binding potentials (SEM - AVS is
numerically low, whereas SEM/AVS ratios are high). Other
field sediments we tested frequently contained 1.0 to 1,000
u,mol of metal over that bound by sulfide yet were not toxic.
This indicates that the role of other sediment phases in metals
bioavailability has great significance. Therefore, accurate pre-
diction of sediments likely to cause toxicity will require es-
timates of partition coefficients and binding strengths of these
sediment phases.
g
^
*
IUO-
80-
'60-
•
40-
20-
-
H
-1
A A
8
a a • "
A
a
A H
•
id
A • fl
» A*m AA £ ' A .T>^ 4 A *
4ff* TA Jt^^ OvA»\ ++ ^
DO -10 -i o i 10 100 ia
SEM-AVS (pmol/g dry wt)
Pig. IS. Percent mortality of amphipods Ampehsca abdita (A a.) and Hyalella azteca (H a); oligochaete Lumbriculus variegatus (L.v); and
polychaetes Capitella capitata (C c.) and Neanthes arenaceodentata (Na.) exposed to sediments from three saltwater and four freshwater field
locations as a function of the sum of the molar concentrations of SEM minus the molar concentration of AVS (SEM — AVS) The vertical
dashed line at SEM - AVS = 00 indicates the boundary between sulfide-bound unavailable metal and potentially available metal (See Fig. 8
and 9 legends for definitions of symbols.)
-------
2094 Environ Toxicol Chem 15, 1996
D.J. Hansen et al
SUMMARY
We believe that results from tests using sediments spiked with
metals and sediments from the field in locations where toxicity
is metals-associated demonstrate the value of sediment concen-
trations normalized by SEM/AVS ratio and IWTUs, instead of
dry weight metals concentrations, in explaining the biological
availability of metals. Importantly, data from spiked sediment
tests strongly indicate that metals are not the cause of the toxicity
observed in field sediments when both SEM/AVS ratios are
<1.0 and IWTUs are <0.5. Expressing concentrations of metals
in sediments on an SEM - AVS basis provides important insight
into available additional binding capacity of sediments, the ex-
tent to which sulfide binding has been exceeded, and potentially
the specific metal that may be causing toxicity, if the metals-
specific molar concentrations are subtracted in order of their
sulfide solubility products. Predictions of sediments not likely
to be toxic based on SEM/AVS ratios and IWTUs for all data
from freshwater or saltwater field sediment and spiked sediment
tests are extremely accurate (93.3 to 100%) using either or both
parameters. While predictions of sediments likely to be toxic
are less accurate (19.4 to 79.1%), this approach is extremely
useful in identifying sediments of potential concern. Several
sources of uncertainty related to sediment geochemistry, the
kinetics of binding and release, and organism-sediment inter-
action need further research.
Acknowledgement—This document has been reviewed in accordance
with EPA policy and approved for publication. The contents of this
publication do not necessarily reflect the views of the EPA. Mention
of trade names or commercial products does not constitute endorse-
ment or recommendation for use
REFERENCES
1 Luoma, SM. 1989. Can we determine the biological availability
of sediment-bound trace elements? Hydrobiologia 176/177:379-
396.
2. Tessier, A., Y. CouiUard, P.G.C. Campbell and J.C. Auclair.
1993. Modeling Cd partitioning in oxic lake sediments and Cd
concentrations in the freshwater bivalve Anodonta grandis. Lim-
nol. Oceanogr. 38:1-17.
3. Di Toro, D., F. Harrison, E. Jenne, W. Lick and S. Karickhoff.
1987. Synopsis of discussion session 2. Environmental fate and
compartmentahzations. In K.L. Dickson, A.W. Maki and WA.
Brungs, eds., Fate and Effects of Sediment-Bound Chemicals in
Aquatic Systems, Pergamon, Elmsford, NY, USA, pp. 136-147.
4. Bryan, G.W. and WJ. Langston. 1992 Bioavailability, accu-
mulation and effects of heavy metals in sediments with special
reference to United Kingdom estuaries1 A review. Environ. Pollut.
76:89-131
5. Tessier, A. and P.G.C. Campbell. 1987 Partitioning of trace
metals in sediments. Relationships with bioavailabihty. Hydro-
biologia 149.43-52.
6. Di Toro, D.M., J.D. Mahony, DJ. Hansen, KJ. Scott, M.B.
Hicks, S.M. Mays and M.S. Redmond. 1990. Toxicity of cad-
mium in sediments- The role of acid volatile sulfides. Environ.
JToxicol Chem. 9.1487-1502.
7. TOahony, J.D., D.M. Di Toro, A.M. Gonzalez, DJ. Hansen,
WJ. Berry and G.T. Ankley. 1991. A sediment component in
addition to acid volatile sulfide that may further control toxicity
of metals. Abstracts, 12th Annual Meeting, Society of Environ-
mental Toxicology and Chemistry, Cincinnati, OH, USA, No-
vember 8-12, p. 91.
8 Di Toro, D.M., J.D. Mahony, DJ. Hansen, KJ. Scott, A.R.
Carlson and G.T. Ankley. 1992. Acid volatile sulfide predicts
the acute toxicity of cadmium and nickel in sediments. Environ.
Sci Technol 26.96-101.
9 Jorgensen, B.B. 1977. The sulfur cycle of a coastal marine sed-
iment (Lunfjorden, Denmark). Umnol Oceanogr. 22:814-832
10. Berry, WJ., DJ. Hansen, J.D. Mahony, D.L. Robson, D.M.
DI Toro, B.P. Shipley, B. Rogers, J.M. Corbln and W.S. Booth-
man. 1996 Predicting the toxicity of metal-spiked laboratory
sediments using acid-volatile sulfide and interstitial water nor-
malizations Environ Toxicol. Chem 15:2067-2079.
11. Carlson, A.R., G.L. Phipps, V.R. Mattson, P.A. Kosian and
A.M. Cotter. 1991. The role of acid volatile sulfide in deter-
mining cadmium bioavailabihty and toxicity in freshwater sedi-
ments. Environ. Toxicol Chem 10.1309-1319
12. Casas, A.M. and E.A. Crecelius. 1994 The relationship between
acid volatile sulfide and the toxicity of zinc, lead, and copper in
marine sediments. Environ Toxicol Chem 13 529-536
13. Pesch, C.E., DJ. Hansen, W.S. Boothman, WJ. Berry and
J.D. Mahony. 1994 The role of acid volatile sulfide and inter-
stitial water metal concentrations in determining bioavailabihty
of cadmium and nickel from contaminated sediments to the poly-
chaete Neanthes arenaceodentata. Environ. Toxicol. Chem. 14
129-141.
14. Ankley, G.T., et al. 1991. Acid volatile sulfide as a factor me-
diating cadmium and nickel bioavailabihty in contaminated sed-
iments. Environ. Toxicol. Chem. 10:1299-1307.
15. Ankley, G.T., V.R. Mattson, E.N. Leonard, C.W. West and
J.L. Bennett 1993. Predicting the acute toxicity of copper in
freshwater sediments. Evaluation of the role of acid volatile sul-
fide. Environ Toxicol. Chem 12:315-320.
16. American Society for Testing and Materials. 1995. Guide for
conducting 10-day static sediment toxicity tests with marine and
estuanne amphipods E 1367-92 In Annual Book ofASTM Stan-
dards, Vol. 11.05. Philadelphia, PA, USA, pp. 767-792.
17 Mearns, AJ., R.C. Swartz, J.M. Cummins, P.A. Dinnel, P.
Plesha and P.M. Chapman. 1986. Inter-laboratory comparison
of a sediment toxicity test using the marine amphipod, Rhepox-
ynius abronlus. Mar. Environ. Res. 19:13-37.
18. Allen, H.E., F. Gongmin and B. Deng. 1993. Analysis of acid
volatile sulfide (AVS) and simultaneously extracted metals (SEM)
for estimation of potential toxicity in aquatic sediments Environ
Toxicol. Chem 12.1441-1453.
19. Cornwall, J.C. and J.W. Morse. 1987. The characterization of
iron sulfide minerals in recent anoxic marine sediments. Mar.
Chem. 22-55-69.
20. Boothman, W.S. and A. Helmstetter. 1992 Vertical and seasonal
variability of acid volatile sulfides in marine sediments. EPA 600/
X-93/036. U.S. Environmental Protection Agency, Narragansett,
RI.
21. Ma, D. and F. Zhang. 1988. Lead, zinc and cadmium distribu-
tions in different geochemical phases of surface sediments from
Jinzhou Bay. Acta Scientiae Circumstantiae 8:49-55.
22. Samant, HLS., K.G. Doe and O.C. Vaidya. 1990. An integrated
chemical and biological study of the bioavailabihty of metals in
sediments from two contaminated harbors in New Brunswick.
Can Sci. Total Environ. 96:253-268.
23. Uthe, J.F., C.L. Chou and D. Robinson. 1980 Cadmium in Amer-
ican lobster (Homarus amencanus) from the area of Belledune
Harbour. Cadmium pollution of Belledune Harbor, New Brunswick,
Canada. Can. Tech. Rep. Fish. Aquat. Sci 963:65-71.
24. Ray, S., D.W. McLeese, C J). Metcalfe, L.E. Burridge and B.A.
Waiwood. 1980. Distribution of cadmium in marine biota in the
vicinity of Belledune Cadmium pollution of Belledune Harbor,
New Brunswick, Canada. Can. Tech -Rep Fish. Aquat. Sci. 963:
11-34
25. Bower, P.M., HJ. Simpson, S.C. Williams and Y.H. Li. 1978
Heavy metals in the sediments of Foundry Cove, Cold Spnng,
NY. Environ Sci. Technol. 12:683-687.
26. Bieri, R., et al. 1982 Toxic substances. In E.G. Macalaster, D.A
Barker and M. Kasper, eds , Chesapeake Bay Program Technical
Studies: A Synthesis. U.S Environmental Protection Agency,
Washington, DC, pp. 263-375.
27. Pinkney, A.E., G. Gowda and E. Rzemien. 1991. Sediment
toxicity testing of the Baltimore Harbor and C & D Canal ap-
proach channels with the amphipod, Leptocheirus plumulosus.
Department of Natural Resources, Annapolis, MD, USA
28 Green, A.S., G.T. Chandler and E.R. Blood. 1993. Aqueous-,
pore-water-, and sediment-phase cadmium: Toxicity relationships
for a meiobenthic copepod Environ. Toxicol Chem 12 1497-
1506
29. Ankley, G.T., N.A. Thomas, D.M. Di Toro, D J. Hansen, J.D.
Mahony, WJ. Berry, R.C. Swartz and R.A. Hoke. 1994. As-
sessing potential bioavailabihty of metals in sediments A pro-
posed approach. Environ Manage 18:331-337.
-------
Environmental Toxicology and Chemistry, Vol IS, No 12, pp 2067-2079, 1996
Printed in the USA
0730-7268/96 $6.00 + 00
PREDICTING THE TOXICITY OF METAL-SPIKED LABORATORY SEDIMENTS
USING ACID- VOLATILE SULFIDE AND INTERSTITIAL
WATER NORMALIZATIONS
W.J. BERRY,*t D.J. HANSEN.t J.P. MAHONY.t D.L. RoBSON,§ D.M. Di ToRO,|| B.P. SHIPLEY,*
B. RoGERS.tt J.M. CoRBmtt and W.S. BooTHMANt
tU.S. Environmental Protection Agency, National Health and Environmental Effects Research Laboratory, Atlantic Ecology Division,
27 Tarzwell Dnve, Narragansett, Rhode Island 02882
^Department of Environmental Engineering, Manhattan College, Riverdale, New York, New York 10471, USA
§ Rhode Island Department of Environmental Management, 291 Promenade Street, Providence, Rhode Island 02908, USA
PydroQual, 1 Lethbridge Plaza, Mahwah, New Jersey 07430, USA
#Springborn Laboratories, 790 Main Street, Wareham, Massachusetts 02571, USA
ttScience Applications International Corporation, 165 Dean Knauss Drive, Narragansett, Rhode Island 02882, USA
tfTexas Natural Resource Conservation Commission, Water Quality Standards, Box 13087, Austin, Texas 78711-3087, USA
(Received 18 September 1995; Accepted 2 July 1996)
Abstract — Numerous studies have shown that dry weight concentrations of metals in sediments cannot be used to predict toxicity
across sediments. However, several studies/ using sediments from both freshwater and saltwater have shown that interstitial water
concentration or normalizations involving acid-volatile sulfide (AVS) can be used to predict toxicity in sediments contaminated
with cadmium, copper, nickel, lead, or zinc across a wide range of sediment types. Six separate experiments were conducted in
which two or three sediments of varying AVS concentration were spiked with a series of concentrations of cadmium, copper, lead,
nickel, or zinc or a mixture of four of these metals. The amphipod Ampelisca abdita was then exposed to the sediments in 10-d
toxicity tests. Amphipod mortality was sediment dependent when plotted against dry weight metals concentration but was not
sediment dependent when plotted against simultaneously extracted metal (SEM)/AVS or interstitial water toxic units (IWTUs).
Sediments with SEM/AVS ratios <1.0 were seldom (2.3%) toxic (i.e., caused >24% mortality), while sediments with SEM/AVS
ratios >1.0 were frequently (80%) toxic. Similarly, sediments with <0.5 IWTU were seldom toxic (3 0%), while sediments with
>0.5 IWTU were toxic 94.4% of the time. These results, coupled with results from related studies, demonstrate that an understanding
of the fundamental chemical reactions which control the availability of cadmium, copper, lead, nickel, and zinc in sediments can
be used to explain observed biological responses. We believe that using SEM/AVS ratios and IWTUs allows for more accurate
predictions of acute mortality, with better causal linkage to metal concentration, than is possible with sediment evaluation tools
which rely on dry weight metal concentrations
Keywords — Acid-volatile sulfide Interstitial water Metals Sediments Toxicity
INTRODUCTION
Di Toro et al. [1] showed that the toxicity of cadmium-
spiked marine sediments was linked to metals/acid-volatile
sulfide (AVS) ratios and interstitial water (IW) metals con-
centrations. Since then several studies using freshwater and
saltwater sediments spiked with cadmium, copper, lead, nickel,
and zinc [2-5] have demonstrated the utility of these param-
eters in causally linking toxicity to metals in sediments. Kemp
and Swartz [6] maintained constant IW concentrations in cad-
mium-spiked sediments amended with varying quantities of
organic carbon and found that mortality was correlated with
IW concentration but not total sediment concentration. The
utility of metals/AVS ratios and IW concentrations has also
been demonstrated in studies conducted at field sites contam-
inated with copper [7] and a mixture of cadmium and nickel
[3,8,9]. Two colonization experiments with cadmium-spiked
sediments, one conducted in a freshwater lake [10] and a sea-
water colonization test conducted in the laboratory [11], also
support the use of this approach for predicting the toxicity of
* To whom correspondence may be addressed
Contribution 1622, U.S. Environmental Protection Agency, Na-
tional Health and Ecological Effects Research Laboratory, Narragan-
sett, Rhode Island.
these metals in sediments. The results of these studies have
been reviewed by Ankley et al. [12], who proposed practical
sediment assessment methodologies using simultaneously ex-
tracted metal (SEM)/AVS ratios and IW metals concentrations
to evaluate cadmium, copper, nickel, lead, and zinc contami-
nation in sediments. The success of this approach for predicting
the bioavailability of these metals in sediments is in direct
contrast to the lack of success in using dry weight metals
concentrations for this purpose [1,3,13].
The theoretical foundation for equilibrium partitioning
(EqP) theory-based SEM/AVS predictions of metal toxicity is
that the sulfides of cadmium, copper, nickel, lead, and zinc all
have lower sulfide solubility product constants than do the
sulfides of iron and manganese, which are formed naturally in
sediments as a product of the bacterial oxidation of organic
matter [14]. As a result, these metals will displace manganese
and iron whenever they are present together with manganese
and iron monosulfides [3]. Because the solubility product con-
stants of these sulfides are small, sediments with an excess of
AVS will have very low metal activity in the IW, and no
toxicity due to these metals should be observed in the sedi-
ments.
The results of the studies cited above were consistent with
the following predictions based on EqP theory: (1) when sed-
2067
-------
2068 Environ. Toxicol Chem. 15, 1996
iments have a metals/AVS ratio of <1.0, sediments will not
be toxic, and little or no metal will be present hi the IW; and
(2) when sediments have a metals/AVS ratio of >1.0, AVS
binding potential will be exceeded, and metals will be present
in the IW or available to bind with other sediment phases (i.e.,
total organic carbon) [1]. Nontoxic sediments having metals/
AVS ratios >1.0 may have low IW concentrations, less than
those known to be toxic in water-only tests, suggesting the
importance of additional metal binding phases in sediments
[4,7,15]. The appropriate fraction of metals to use for AVS
normalization is referred to as simultaneously extracted metal
(SEM). This is the metal which is extracted in the cold acid
used in the AVS procedure. This fraction is appropriate be-
cause some metals form sulfides which are not fully labile in
the short time required for the AVS extraction (e.g., Ni and
Zn) [3]. If a more rigorous extraction were used to increase
the fraction of metal extracted which did not also capture the
additional sulfide extracted, then the sulfide associated with
the additional metal release would not be quantified. This
would result in an erroneously high metal/AVS ratio [3].
An analysis of a simple chemical equilibrium model for the
system M(II), Fe(n), S(-II), where M(D() are the divalent met-
als that form sulfides, shows that
where
the activity of M in the IW,
[Af]T = the total metal concentration,
^MS - the solubility product for the metal sulfide (MS), and
K^ = the solubility product of FeS.
The ratio JW*F*s is 10-", 10-60, 10-'", 10-"°, and lO"186
for M = nickel, zinc, cadmium, lead, and copper, respectively
[3]. If the metals are present in excess of the sulfides (SEM/
AVS >1.0) and there are no other sediment phases capable of
binding the metals, e.g., dissolved organic carbon (DOC) or
total organic carbon (TOC), metal will be present in the IW,
and the sediment may be toxic.
Toxicity predictions based on sulfide binding for sediments
contaminated with mixtures of metals which form insoluble
sulfides would use the sum of the molar concentration of SEM
for the divalent metals present (i.e., Cd, Cu, Ni, Pb, and Zn)
for comparison with the molar concentration of AVS in the
sediment. If the sum of the SEM is greater than that of AVS,
metals may occur in the IW in sufficient concentrations to be
toxic. If the toxicity of the cationic metals in IW is assumed
to be additive [16], it should be possible to predict the toxicity
of the sediments in the same way as in the individual metal
experiments, using the sum of the interstitial water toxic units
(IWTUs). The divalent metals should appear in the IW hi
reverse order of the solubilities of their sulfides [1]. Thus,
nickel should appear first in the IW in sediments with SEM/
AVS ratios slightly >1.0, followed by zinc, cadmium, lead,
and copper as the concentration of metals increases relative
to that of AVS.
In this article we summarize the available data from acute
laboratory sediment toxicity tests with spiked sediments which
can be used to test the utility of SEM/AVS ratios and IWTUs
in predicting sediment toxicity. First, the methodology and
results from a series of toxicity tests using saltwater sediments
spiked with cadmium, copper, lead, nickel, or zinc and one
test using an equimolar mixture of cadmium, copper, nickel,
and zinc are examined in detail. These tests are highlighted
W.J Berry et al.
because they serve as an example from a single laboratory
(U.S. Environmental Protection Agency, National Health and
Environmental Effects Research Laboratory, Narragansett, RI)
of results from a senes of sediment spiking experiments with
metals which followed a consistent methodology performed
with a relatively sensitive species, the amphipod Ampelisca
abdita. Ampelisca abdita is an estuarine, tube-building, in-
faunal amphipod commonly used in sediment toxicity testing
[17]. Data from the A. abdita tests with nickel and cadmium
have been published elsewhere [1,3], but this is the first time
the data from tests using this species with copper, lead, and
zinc have been reported. This is also the first published report
of a toxicity test in which sediments spiked with a mixture of
metals have been used to assess the usefulness of AVS and
IW metals concentrations in the prediction of sediment tox-
icity. Published results from spiked sediment tests in which
SEM/AVS and/or IW metals concentrations were measured,
including those using polychaetes [5,9] and copepods [4] in
saltwater sediments and oligochaetes and snails [2] in fresh-
water sediments, are also discussed.
MATERIALS AND METHODS: AMPHIPOD TESTS
Organism collection and acclimation
Ampelisca abdita were collected from tidal flats in the Pet-
taquamscutt (Narrow) River, a small estuary flowing into Nar-
ragansett Bay, Rhode Island. Surface sediment containing the
amphipods was either sieved hi the field or transferred to the
laboratory within Vi h and then sieved through a 0.5-mm mesh
screen. In the laboratory amphipods and amphipod tubes were
vigorously sieved hi a tub of seawater, then the sieve was
quickly lowered into the water, and the amphipods were col-
lected from the water surface. The amphipods were maintained
for 3 to 7 d in the laboratory in sieved collection-site sediment
and flowing filtered seawater hi 4-L glass jars and acclimated
to the test temperature at the rate of 2 to 4°C/d. During ac-
climation amphipods were fed the laboratory-cultured diatom
Phaeodactylum tricornutum ad libitum.
One sediment from Ninigret Pond, Charlestown, Rhode Is-
land, used in the cadmium experiment, was tested using the
amphipod Rhepoxynius hudsoni. Rhepoxynius hudsoni was
collected using collection and acclimation methods similar to
those used for A. abdita, except that after collection R. hudsoni
was washed directly from the sieve into sorting dishes.
Water-only tests
Ten-day static renewal tests were conducted with A. abdita
to determine water-only LC50s for cadmium, copper, nickel,
lead, and zinc in seawater. Animals were exposed, unfed, to
five concentrations of metal and a control, with two replicates
per concentration. Amphipods were exposed in 900-ml glass
canning jars that contained 800 ml of water. Acclimated am-
phipods were sieved from the holding jars, sequentially dis-
tributed to 100-ml plastic cups (10 amphipods/cup), then ran-
domly added to the exposure chambers. Seventy-five to 100%
of the water in each replicate was renewed every other day,
depending on the experiment. Water was sampled at least once
during the test (usually twice, once near the beginning and
once near the end of the test) to determine the concentration
of metal. In some experiments, aliquots from the two replicates
were pooled prior to analysis. Exposure chambers were cov-
ered with black plastic. The exposure chambers were checked
daily, and amphipods which appeared dead were removed and
-------
Predicting the toxicity of metal-spiked sediments
examined under a dissecting microscope. Live animals were
returned to the test, and dead animals were recorded and dis-
carded.
Spiked sediment tests
Amphipods were exposed to control and metal-spiked sed-
iments in 10-d tests with continuous renewal of overlying wa-
ter. In all experiments two sediments of different AVS con-
centration were used: one from Ninigret Pond (AVS = 1.18—
2.25 (J-M S/g) and one from Long Island Sound (AVS = 9.72-
19.9 u.M S/g) (Table 1). In the cadmium test, a 1:1 mixture
of these two sediments was also used (AVS = 4.34 fiM S/g).
The nominal treatments used in most experiments, expressed
as the molar ratio of metal to AVS, were 0.0 (control), 0.1,
0.3, 1, 3, 10, and 30 (Table 1). There were four replicates per
treatment in each test: two "biological" replicates were used
to assess mortality, and two "chemical" replicates were used
for metal and AVS analyses of the sediment at test initiation
and termination. Twenty (30 in the cadmium test) amphipods
were added to each biological replicate and the day 10 chem-
ical replicate at the start of the test.
The Long Island Sound sediment was collected from an
uncontaminated site in central Long Island Sound (40°7.95'N,
72°52.7'W) with a Smith—Mclntyre grab sampler, returned to
the laboratory, press-sieved wet through a 2-mm-mesh, stain-
less-steel screen, homogenized, and stored at 4°C. Two sep-
arate collections-of Long Island Sound sediment (LIS1 and
LIS2) were made. The TOC for LIS1 was 0.88%, and the grain
size composition was 5% sand, 71% silt, and 24% clay. Grain
size data are not available for LIS2, but the TOC was 0.99%.
Sediment was also collected in Ninigret Pond. The upper 5
to 10 cm of sediment was collected with a shovel, returned to
the laboratory, sieved wet through a 2-mm stainless-steel
screen, rinsed several times to remove high-organic fine par-
ticles, homogenized, and stored at 4°C. Sediment was collected
from two different sites in Ninigret Pond (NIN1 and NIN2),
but both sediments had TOC values of 0.15% and were 100%
sand (NIN2 was composed mostly of sand which would pass
through a 0.5-mm sieve, while most of the NIN1 sediment
was retained on a 0.5-mm sieve).
Sediments were spiked with metal chloride or nitrate salts
in 1-gallon glass jars. Methods differed slightly from experi-
ment to experiment, but typically 1,200 ml of wet sediment
was added to 2,000 ml of 20°C seawater that contained the
desired weight of metal salt. The spiked sediments were stirred
with a nylon stirrer attached to an electric drill until homo-
geneous, then the overlying air in each jar was replaced with
nitrogen, and the jars were capped and rolled for 1 h. Jars of
sediment were held at 20°C for 8 to 10 d before the start of
the test. Excess water and any precipitate was siphoned off
the sediment surface, and the sediment rehomogenized before
it was added to the exposure chambers.
The exposure chambers were 900-ml glass canning jars,
each with a 1.3-cm-diameter overflow hole (covered with 400-
H-m Nitex® mesh) 11.7 cm from the bottom of the jar. Each
jar contained 200 ml of sediment and approx. 600 ml of sea-
water over the sediment. Each jar was covered with an 8-cm-
diameter glass Carolina dish with a 17-mm-diameter hole for
the seawater delivery tube and air line, which consisted of a
2-ml glass pipette.
Diffusion samplers (i.e., peepers) were constructed from
polyethylene vials (21 mm high, 20 mm diameter, 5 ml ca-
pacity) with 1-fun polycarbonate membranes in the cap [9].
Environ. Toxicol. Chem. 15, 1996 2069
For the mixed metals experiment each diffusion sampler con-
sisted of two vials attached back-to-back to double the volume
of IW collected. The peeper design used in the cadmium ex-
periment was described by Di Tore et al. [1].
Sand-filtered Narragansett Bay water, heated to 20 ± 1°C,
with a mean salinity of 30 ppt (28 to 34 ppt) was used in the
experiments. Seawater flowed into each exposure chamber
from a distribution system consisting of chambers with self-
priming siphons and splitter chambers. Flow rate for each ex-
posure chamber was approx. 28 to 35 volume additions/d (ex-
cept in the cadmium expenment, in which it was approx. 10
volume additions/d). Exposure chambers were placed in 20°C
water baths to maintain temperature. The exposure chambers
were kept under constant light to help keep the amphipods in
the sediment, thus maximizing exposure.
The test was started by placing a diffusion sampler in each
exposure chamber and adding 200 ml of sediment to just cover
the diffusion sampler. Seawater was allowed to flow through
the chambers for 1 d. Amphipods were removed from the
holding containers as described above, distributed sequentially
to 100-ml plastic cups until there were 20 amphipods/cup (30
in the cadmium experiment), then one cup of amphipods was
added randomly to each exposure chamber. The seawater de-
livery system was turned off for 1 h, and any amphipods that
had not burrowed into the sediment in that time were replaced,
except in those replicates where there was an obvious dose
response (i.e., where there were a greater than average number
of unburrowed individuals in all replicates).
Samples of sediment were taken from the day 0 chemistry
replicates for metals and AVS analyses. All but about 1 cm
of overlying water was removed from each day 0 chemistry
replicate with a vacuum pump and pipette tip. The sediment
and a small amount of remaining seawater was homogenized
with a stainless-steel spatula. Approximately half the sedunenl
was placed in an acid-stnpped polyethylene jar for total metals
analysis, while the remainder was placed in a 100-ml poly-
ethylene specimen cup for AVS and SEM analysis. Each jar
was capped, and the samples held in the dark at 4°C until
analysis.
The experimental chambers were checked daily, and am-
phipods which appeared dead were removed and examined
under a dissecting microscope. Live animals were returned to
the test, and dead animals were recorded and discarded. The
volume of water delivered to each exposure container was
measured before each test, and the total flow rate to the system
was measured and adjusted daily. Temperature of the water
bath and salinity of the incoming seawater were measured
daily. The overlying water in each biological replicate was
sampled for metals concentrations at least once near the be-
ginning and once near the end of each test. In some tests the
samples from the two replicates were pooled. Each overlying
water sample was placed in an acid-stnpped, 7-ml polyeth-
ylene vial and acidified with 50 ul of concentrated nitric acid
(pH < 1).
At the end of the test, the diffusion samplers were carefully
removed from each replicate. Any sediment remaining on the
cap or membrane portion of the sampler was rinsed off using
clean seawater. The membrane was then punctured with an
acid-stripped, 5-ml disposable pipette Up, and the'contents of
the sampler removed by pipette. The IW collected from each
diffusion sampler was added to an acid-stnpped, 7-ml poly-
ethylene vial, acidified with 50 (jil of concentrated nitric acid
(pH < 1), and stored for metals analysis. The sediment from
-------
2070 Environ. Toxicol Chem. 15, 1996
W.J. Berry et al
the chemistry replicates was sampled for metals and AYS con-
tent as described for the day 0 chemistry replicates. The con-
tents of each biology replicate were sieved through a 0.5-mm
screen. Material retained on the sieve was examined imme-
diately or preserved with Rose Bengal stain for later sorting.
Amphipods were counted, and any missing animals were as-
sumed to have died and decomposed. Any replicate in which
10% or more amphipods were not found was recounted by
another investigator as a quality assurance check.
Chemical analyses
Sediment samples were analyzed for AVS by a cold-acid
purge-and-trap technique described by Di Toro et al. [1,3].
Simultaneously extracted metal analyses for the copper, nickel,
and zinc experiments were performed using graphite furnace
atomic absorption spectrometry (AAS). Simultaneously ex-
tracted metal analyses for the lead and the mixed metals ex-
periments were performed using inductively coupled plasma
emission spectrometry (ICP). Simultaneously extracted metal
was not measured in the cadmium experiment because the
importance of SEM versus total metal was not understood at
that time. However, cadmium does not form sulfides which
are insoluble in the AVS procedure [3], so acid-extractable
and SEM cadmium concentrations are interchangeable. In the
mixed-metals experiment the acid-extractable and SEM cad-
mium concentrations were similar (W. Berry, unpublished
data). Simultaneously extracted metal for only the metal(s)
under study was measured in the individual chemical exper-
iments. However, the sum of'the SEM for all the cationic
metals averaged only 3.2 jtM/g for unspiked Long Island
Sound sediment and 0.081 n-M/g for unspiked Ninigret Pond
sediment and was thus of little importance in the SEM/AVS
ratio for the level of metal spiking used.
To allow for comparisons with other metal toxicity studies,
total (dry weight) metals analyses were also performed. For
this analysis, metals were extracted from freeze-dried sedi-
ments by ultrasonic agitation with 2 M cold nitric acid (50 ml
to 5 g wet sediment). The supernatant containing the extracted
metals was separated from the sediment residue by centrifu-
gation and analyzed by ICP.
The IW samples from the copper experiment were analyzed
for total copper using graphite furnace AAS. The IW samples
from the cadmium experiment were analyzed for Cd2+ using
a cadmium ion-specific electrode. Total cadmium was esti-
mated by multiplying the measured Cd2+ by 20, which is the
ratio of total cadmium to Cd2+ in seawater [1]. All other in-
terstitial and overlying waters were analyzed for trace metals
by ICP. The IW samples from the mixed metals experiment
were diluted fivefold with 2 M HNO3 in order to, provide
sufficient solution for analysis.
Data handling
Ten-day LC50 values for the water-only tests were calculated
by the trimmed Spearman-Karber method [18]. Detection lim-
its were calculated for all chemical analyses on the basis of
instrument detection limits and sample size. In those instances
where a mean concentration is a summation of measured data
and data below the limit of detection, one-half the detection
limit was used for those values below the limit of detection.
Means for which there are no measured values above the de-
tection limit are indicated by "n.d." in the appropriate tables
and graphs.
For illustrative purposes, sediments which caused >24%
mortality were classified as toxic. Mearns et al. [19] found
that sediments which caused <24% mortality in tests with the
amphipod R. abronius were not consistently classified as toxic.
This criterion is similar to the "80% of control survival" cri-
terion used in the Environmental Monitoring and Assessment
Program (EMAP) program for sediment tests with A. abdita
[20]. Sediments having £ 24% mortality were classified as
nontoxic.
Many of the interstitial and overlying water concentrations
in this article are expressed as IWTUs. An IWTU is the mea-
sured interstitial water concentration divided by the water-only
LC50 concentration for that particular compound for the test
organism. For example, a sediment with an IW concentration
equal to the water-only LC50 concentration for the test or-
ganism would have 1.0 IWTU. When more than one toxic
metal is present, IWTUs are calculated as the sum of the toxic
units of the individual metals, e.g.,
IWTUoj^N, = (IW concn.
+ (IW concn.
Thus, if IW is the principal source of metal toxicity, and avail-
ability of metals is the same in water of water-only tests and
IW in sediment tests, 50% mortality would be expected with
sediments having IWTUs of 1.0. In this article we use IWTUs
of <0.5 to indicate sediments unlikely to cause significant
mortality because, on average, water-only LCO and LC50 val-
ues differ by a factor of approx. 2 [21] and because the data
in our experiments support this value as a break point between
toxic and nontoxic sediments. Calculation of IWTUs was based
solely on detectable metal concentrations.
RESULTS: SALTWATER AMPHIPOD TESTS
Water-only tests
The LC50 values from the water-only tests are summarized
in Table 2. No 10-d water-only LC50 value for cadmium was
available for R. hudsoni (it is very difficult to conduct 10-d
water-only tests with amphipods from this genus because of
problems with excessive control mortality), so the 10-d LC50
value for A. abdita (36 \ig/L) was substituted for the calcu-
lation of toxic units for this species. This substitution is rea-
sonable because these amphipods have similar sensitivities,
i.e., the 4-d Cd LCSOs for R. hudsoni (640 jig/L) and A. abdita
(340 u,gfL) differed by less than a factor of 2 [1].
Sediment chemistry
Day 0 versus day 10 chemistry values. Acid volatile sulfide,
SEM, and dry weight metals sediment chemistry measure-
ments varied somewhat from day 0 to day 10, but the variation
was generally within 20% and did not show a definite time-,
concentration-, or metal-dependent pattern. Therefore, all
AVS, SEM, and dry weight metals sediment chemistry data
are reported as the mean of day 0 and day 10 value (Day 0
and day 10 values are available on request.)
Interstitial water metal versus SEM/AVS ratio. The IW
metal concentrations in all experiments were usually below
the limit of detection in sediments with SEM/AVS ratios below
1.0 (Fig. 1 and Table 1). In the cadmium and mixed metals
experiments, the IW metals concentrations, while appearing
large, are unknown because of the large detection limits in
these experiments. In the cadmium experiment, one-half the
detection limit of the cadmium electrode was 1.33 u,mol/L.
For the mixed metals experiment, the sum of one-half of the
detection limits of the four metals spiked in this test was 1.54
-------
'redictmg the toxicity of metal-spiked sediments
Environ. Toxicol Chem 15, 1996 2071
Table 1. Summary of AVS, metal concentrations and amphipod (A abdita) mortality in 10-d toxicity
tests with sediments spiked with cadmium, copper, lead, nickel, zinc, and a mixture of
cadmium, copper, nickel, and zinc
Metal
Cadmium0
Copper
Lead
'
Nickel
-
Zinc
Sedi-
ment
LIS
.
NIN
Mix
LIS
NIN
LIS
NIN
LIS
NIN
LIS
Nominal
SEM/AVS
Control
0.1 X
0.3 x
IX
3X
10X
Control
01X
0.3 x
IX
3X
10X
Control
0.1 x
0.3 X
IX
3X
10X
Control
0.1 X
0.3 x
IX
3X
10X
30X
Control
0.1 x
0.3 x
IX
3X
10x
30X
Control
0.1 x
0.3 X
IX
3X
10X
30x
Control
01X
0.3 X
IX
3X
lOx
30X
Control
0.1 X
0.3 X
IX
3x
10X
30X
100X
Control
Olx
0.3 x
IX
3x
10X
30X
100X
Control
0.1 x
0.3 X
SEM
metal*
(nmol/g)
0.00
1.57
4.85
16.7
51.7
177
0.00
0.15
0.64
2.57
5.90
24.3
0.00
0.30
1.75
9.64
20.7
48.4
0.27
1.00
1.57
11.6
47.0
176
306
0.00
0.05
0.09
0.43
2.08
5.40
10.4
0.23
1.25
4.14
14.5
28.3
67.9
78.2
0.02
0.20
0.60
1.70
7.10
16.6
20.2
0.70
1.12
2.90
9.80
26.4
70.8
266
573
0.17
0.35
0.24
0.54
1.21
2.62
7.62
22.9
1.20
2.80
5.50
AVS
Ounol/g)
14.9
14.9
14.9
14.9
14.9
14.9
1.31
1.31
1.31
1.31
1.31
1.31
4.34
4.34
4.34
4.34
4.34
4.34
13.3
12.2
4.44
1.21
1.94
1.67
1.84
1.22
1.42
1.08
0.635
0.323
0.345'
0.63
19.9
18.6
12.8
16.4
14.9
15.5
14.2
1.20
1.92
2.23
3.10
5.75
4.08
3.37
12.6
120
106
6.17
3.89
4.01
2.70
1.90
1.93
1.88
1.84
1.02
0:60
065
0.90
0.50
11.2
11.7
134
SEM/AVS
ratio
0.00
0.10
0.33
1 12 i
3.50
11.9
0.00
0.12
0.50
1.95
4.34
18.5
0.00
0.10
0.40
2.22
4.80
11.2
002
0.10
0.35
9.60
24.3
105
166
0.00
0.04
0.08
0.68
6.46
15.67
16.58
0.01
0.07
0.32
0.89
1.90
4.38
5.49
0.02
0.09
0.26
0.60
124
4.08
5.97
0.05
0.10
0.30
1.60
6.80
17.7
99.1
303
010
020
013
0.53
202
4.02
8.72
49.3
0.10
0.24
041
IWTU"
<8.3
<8.3
<8.3
> <8.3
2,410
13,400
<83
<8.3
<83
<8.3
264
81.1
<83
<8.3
<8.3
<8.3
967
3,280
0.10
0.10
0.13
0.87
264
1,320
1.980
0.10
030
0.14
0.39
2.08
415
7,920
<0.016
<0.016
<0.016
0.20
0.30
1.10
18.4
<0.02
0.02
<0.02
0.03
0.21
2.50
43.3
<0.02
<0.02
<0.02
<0.02
9.44
164
844
2,980
<0.02
<0.02
<0.02
006
1.50
55.2
272
763
0.12
0.10
010
Mortality
1.65
8.35
16.7
10
100
88.4
5
12.5
12.5
40
95
100
16.7
11.7
234
467
100
85
12.5
7.5
17.5
100
100
100
100
22.5
5
15
30
100
100
100
10
5
12.5
7.5
22.5
,42.5
100
10
17.5
15
5
17.5
55
92.5
0
7.5
25
10
95
100
100
100
5
7.5
2.5
7.5
2.5
97.5
100
97.5
15
7.5
17.5
-------
2072 Environ. Toxicol. Chem. 15, 1996
WJ. Berry et al.
Table 1. Continued.
Metal
SEM
Sedi- Nominal metal' AVS SEM/AVS
ment SEM/AVS (p-mol/g) ((imol/g) ratio rWTU"
Mortality
IX
3X
10x
30X
NIN Control
0.1 X
0.3 X
IX
3X
10X
30X
Mixture LIS Control
(Cd, Cu. Ni, Zn) O.lx
0.3 X
IX
3X
10X
NIN Control
0.1 X
0.3 x
IX
3X
10X
20.3
74.3
155
140
0.01
0.30
0.70
1.50
2.00
4.13
8.82
3.40
4.60
8.85
19.6
53.8
116
0.10
0.21
0.30
0.62
2.44
3.00
15.1
18.2
15.0
14.0
2.25
2.45
3.00
2.73
1.82
1.31
1.94
9.72
11.2
6.93
3.35
1.72
0.21
1.99
1.34
1.43
0.55
0.90
0.09
1.34
4.09
10.3
9.96.
0.00
0.11
0.23
0.54
1.09
3.15
4.54
0.35
0.41
1.30
5.90
31.3
538
0.04
0.15
0.20
1.13
2.82
33.1
0.02
10.2
345
8.280
<0.014
<0.014
<0.014
0.03
23.3
755
2,910
<3.2
<3.2
<3.2
1.91
142
16,200
<3.2
<3.2
<3.2
5.53
23.7
2,010
15
77.5
100
100
5
12.5
12.5
5
35
95
100
5
2.5
2.5
15
100
100
5
17.5
5
22.5
30
100
• Simultaneously extracted metal concentrations reported are only for the metal(s) spiked in the particular
experiment.
" Interstitial water toxic units were calculated using the water-only LCSOs in Table 2.
c Simultaneously extracted metal concentrations were not available for the cadmium experiment, so bulk
measurements for cadmium were substituted.
jimol/L. Above an SEM/AVS ratio of 1.0 the IW concentration
increased up to five orders of magnitude with increasing SEM/
AVS ratio (Table 1). In each experiment there were usually
one or more sediments with SEM/AVS ratios slightly >1.0
with IW concentrations below or near detection limits. The
presence of low IW metals concentrations in sediments with
SEM/AVS ratios >1.0 may in part be due to analytical vari-
ability but is also due to the presence of other binding phases.
In some sediments spiked with copper, nickel, and a mixture
of metals, AVS decreased with increasing metals concentration
(Table 1), presumably due to the formation of copper and
nickel sulfides not completely soluble in the short time required
for the AVS extraction. This apparent decrease in AVS is an-
other example of the importance of using SEM (as opposed
to total metal) in the calculation of metals/AVS ratios.
The relationship between IW metal concentration and SEM/
AVS ratio in the mixed metals experiment was similar to that
in the individual metal experiments when the molar concen-
trations of all the metals are summed (Fig. 1). Further insight
into the partitioning of the metals in the IW from the mixed
metals experiments can be gained by plotting the IW concen-
trations for each individual metal in this experiment (Fig. 2).
Table 2. Ten-day water-only LC50 values for A. abdita
Metal
LC50
95% Confidence limits
Cadmium
Copper
Lead
Nickel
Zinc
36.0
20.5
3.020
2,400
343
Not reliable
16.5-25.5
1,980-4,610
2,050-2,820
291-405
In the Long Island Sound sediment, all four metals were below
the limit of detection in treatments with SEM/AVS ratios of
1.25 or lower (Fig. 2). As the SEM/AVS ratio of the treatments
increased, detectable concentrations of metal began to appear.
The most soluble sulfide (Ni) appeared first and at the highest
IW concentration. As SEM/AVS ratios increased, the other
metals appeared in the order of their solubility product con-
stants. The metal with the least soluble sulfide (Cu) appeared
last and at the lowest concentration. The relationship between
1
5
IS
&
JO
1
1
g
£
H
£
IUUUUUU'
100000
10000
1000
100
10
V
0.1-
nm-
*
. • **
.V A *
V* .
•%*, v
A* **
— (**^ +»^ill mf fm
'—* itJPiMcm+ * A
^ • ** A \m Jl% "
0.001
0.01
01
10
100
1000
SEM/AVS
Fig 1. Interstitial water metals concentration (jiM/L) as a function
of SEM/AVS ratio Data from the mixed metals experiment represent
the molar sum of cadmium, copper, nickel, and zinc. All experiments
combined. Data below the IW detection limit are plotted at one-half
the detection limit, indicated by arrows. All data in the copper ex-
periment were above the limit of detection. Data below the SEM
detection limit are plotted at SEM/AVS = 0.001 • = Cd, A = Cu;
* = Ni; * = Pb; • = Zn, and += mixed metals.
-------
Predicting the toxicity of metal-spiked sediments
D 10000
§
£ 1000
f 100
001
0.01
1000
001
1000
SEM/AVS
Fig 2. Individual IW metals (jiM/L) concentration in the mixed met-
als experiment as a function of SEM/AVS ratio The top panel rep-
resents data from the Long Island Sound sediment, the bottom panel
data from the Ninigret Pond sediment Data below the IW detection
limit are plotted at one-half the detection limit, indicated by arrows.
• = Cd, A = Cu; * = Ni; and • = Zn
IW concentration and SEM/AVS ratio in the Ninigret Pond
sediments was similar to that in the Long Island Sound sed-
iments (Fig. 2). In the sediment treatments with SEM/AVS
ratios of < 1.0 there was no detectable metal in the IW. In one
sediment with an SEM/AVS ratio slightly >1.0 (1.12) there
were small, but measurable, zinc and cadmium concentrations
in the IW. In the sediment treatment with the next higher SEM/
AVS ratio, there was measurable nickel, zinc, and cadmium,
with the metal concentrations decreasing in that order. Only
in the sediment with the highest SEM/AVS ratio was mea-
surable copper found in the IW.
Sediment toxicity
The mortality of amphipods as a function of total metals
concentrations followed a similar pattern in each of the five
individual metal and the mixed metals toxicity tests. Mortality
appeared sediment dependent when plotted on a total metals
basis (Fig. 3). Mortality increased with increasing metals con-
centration (u.g/g dry weight) for each sediment, but in each
experiment there were treatments in low AVS sediments (Ni-
nigret Pond) which caused 100% mortality at dry weight met-
als concentrations which did not cause appreciable mortality
in treatments from the high AVS sediment (Long Island
Sound). Thus, although mortality is concentration dependent
for each sediment, the concentration-response curves do not
overlap. Therefore, it is not possible to predict amphipod mor-
tality in different sediments on the basis of total metals con-
centrations alone (Figs. 3 and 4a).
Environ Toxicol Chem 15, 1996 207 ^
Mortality did not appear to be metal specific when plotted
on a molar dry weight metal basis (Fig. 4a). Some factor in
the sediment appeared to affect the toxicity of all five metals
similarly because within a sediment the results for all five
metals were very similar (Fig. 4b).
Mortality in the individual and mixed metals experiments
was sediment independent when plotted on an SEM/AVS ratio
basis (Fig. 5). Sediments with an SEM/AVS ratio <1.0 did
not cause mortality significantly different from the control (i.e.,
>24%). In sediments with an SEM/AVS ratio >1.0, mortality
increased with increasing SEM/AVS ratio, although in each
experiment there were usually one or two sediments with SEM/
AVS ratios slightly >1.0, and in one instance 5.8 (Fig. 5f),
which did not cause significant mortality. This indicates that
there are other binding phases in the sediment. Thus, it is
possible to predict with accuracy which sediments will be
nontoxic (cause <2A% amphipod mortality) and, with slightly
less accuracy, which sediments will be toxic (cause >24%
amphipod mortality). When the results of all of the experiments
are plotted together the mortality of amphipods as a function
of SEM/AVS ratio appears to be metal and sediment indepen-
dent for the five individual metals and for a mixture of metals
(Fig. 6).
Mortality was also not sediment or metal specific when plot-
ted against IWTUs. Little amphipod mortality occurred in sed
iments with IWTUs of <0.5 (Fig. 7). (The data from those
sediments in which IW cadmium was not detected in the cad
mium experiment and the data from the mixed metals expei
iment in which IW metal was not detected are not include'
in Figs. 7 and 9 and Table 3 because the detection limits in
these experiments were greater than a toxic unit, making these
data uninterpretable.) Amphipod mortality was higher in sed
iments with >0.5 IWTU and increased with increasing IWTU
value. As was the case when mortality was plotted againsi
SEM/AVS and ratios exceeded 1.0, there were several sedi
ments with IWTU values >0.5 which did not cause mortality
This was especially true in the range of IWTU values >0.5
but <10.0 (Fig. 7). This may be due in part to the variability
inherent in water-only tests but may also indicate that not al?
of the IW metal is bioavailable. Thus, for both SEM/AVS ratio,'
and IWTUs, sediments not likely to cause amphipod mortality
can be predicted with near certainty, but predicting which sed-
iments are likely to cause amphipod mortality is less accurate.
The lack of metal specificity and the results of the mixed metals
experiment ( Fig. 7 and Table 1) indicate that the sum of the
IWTUs can be used to make predictions about amphipod mor
tality of any combination of metals tested in these experiments.
When the results of the individual metals and the mixed
metals tests are taken together, 97.7% of the 43 sediments with
SEM/AVS ratios <1.0 were not toxic (i.e., caused mortality
<24%). Of the 45 sediments with SEM/AVS ratios >1.0,
80.0% were toxic (i.e., caused mortality £24%). Ninety-seven
percent of the 33 sediments with an IWTU value of <0.5 were
not toxic, while 94.4% of the 36 sediments with an IWTU
value of >0.5 were toxic. When both SEM/AVS ratio and
IWTU were combined, the predictive ability was not improved
over the use of IWTU alone. Of 29 sediments with SEM/AVS
ratios <1.0 and IWTUs <0.5, 96.6% were nontoxic, while
94.4% of 36 sediments with SEM/AVS ratios > 1.0 and IWTUs
>0.5 were toxic (Table 3).
DISCUSSION
Our results demonstrate that it is not possible to predict the
toxicity of sediments spiked with metals using the total metal
-------
2074 Environ. Toxtcol Chem. 15, 1996
WJ. Berry et al.
100
80
60
40
20
100
80
60
40
20
0.01 0.1 1 10 100 1000 0.01 0.1 1 10 100 1000
Dry Wt Metal (Mmd/g). Cd Dry Wt Metal Gjmol/g): Cu
100
80
>.
f6 60
^ «>
20
n
c
•**"»••*%
I — t*"
;
•
i
•
•
§
i
•
100
80
60
40
20
0.01 0.1 1 10 100 1000 0.01 0.1 1 10 100 1000
Dry Wt. Metal (Mmol/Q):N Dry Wt. Metal (pmol/a). Pb
100
80
60
40
20
0.01 0.1 1 10 100
Dry Wt. Metal (Mmol/g): Zn
100
80
60
40
20
1000
0.01 0.1 1 10 100 1000
Dry Wt. Metal (pmol/g). Cd + Cu + K + Zn
Fig. 3. Percentage mortality of A. abdita (R. hudsoni in the Ninigret Pond sediment in the cadmium experiment) as a function of the sum of
the concentrations of cadmium, copper, lead, nickel, and zinc (p.M metal/g dry weight sediment) in three sediments. Long Island Sound (LIS),
Ninigret Pond (NIN), and a SO/SO mixture of these two sediments (Mix). Each panel represents data from a separate experiment. A = Long
Island Sound sediment; • = Ninigret Pond sediment; and • = mix.
concentration on a dry weight basis. The relationship between
mortality and total metal concentrations in our tests was sed-
iment specific. In contrast, the relationships between mortality
and SEM/AVS ratio and between mortality and IWTUs were
demonstrated to be sediment independent in our studies with
A. abdita. This supports the use of SEM/AVS ratios and
IWTUs to predict organism response.
The dry weight metals concentrations required to cause
acute mortality in these experiments are very high relative to
those often suspected to be of lexicological significance in
field sediments, and this has sometimes been interpreted as a
limitation of the use of SEM and AVS to predict toxicity. The
sediments we used for the spiking, however, were not unusu-
ally high in AVS. The AVS in the Ninigret Pond sediment
(AVS = 1.18-2.25 |iM/g dry weight) is representative of a
clean sand, and the Long Island Sound sediment (AVS = 9.72-
19.9 u,M/g dry weight) is representative of a typical muddy
sediment. Wolfe et al. [22] reported AVS values ranging from
0.1 to 101 |xM/g dry weight in a survey of sediments from
Long Island Sound and environs. The important point is that
even a sediment with only a moderate concentration of AVS
has a considerable capacity for sequestering metal in a form
which is not bioavailable [23].
Most of the metal-spiked sediments we tested caused either
little or no mortality or nearly complete mortality (Table 1 and
Fig. 6). We believe that this "all or nothing" result is prin-
cipally a consequence of the sharp increase in IW metal con-
centrations when the SEM/AVS ratio exceeds 1.0 and sulfide
no longer is a significant binding phase [1]. This sharp increase
is most obvious when the data are reported on an experiment-
by-experiment basis (Table 1). When sufficient sulfide is pres-
ent in the sediment, partitioning is controlled by precipitation
-------
Predicting the toxicity of metal-spiked sediments
Environ. Toxicol Chem. 15, 1996 2075
100
80
g 60
a* 40
20
n
a
• Cd
A cu
* Nl
* Pb
• Zn
+ MX
'J
*A **A *. ****• *
: •
•
•
. •
A +"
> * • .•« 4m
• ** • l«? 4 •
+* * * *JP* *
0.01 0.1 1 10 100 100C
DryWt. Metal (pmol/8)
100 J
80 J
60
20
jtAJAA
*
•••'
0.01 0.1 1 10
DtyWt. Metal (pmol/g)
100
1000
Fig. 4. Percentage mortality of A. abdita (R. hudsoni in the Nmigret
Pond sediment in the cadmium experiment) as a function of the sum
of the concentrations of cadmium, copper, lead, nickel, and zinc (pM
metal/g dry weight sediment). All experiments combined. Upper panel
plots data by metal; • = Cd; A = Cu; * = Ni; * = Pb, • = Zn,
and + = mixed metals. Lower panel plots data by sediment. A =
Long Island Sound sediment; • - Ninigret Pond sediment; and • =
mix.
of the metal sulfide [1], and little or no metal is present in the
IW [IS]. When the binding capacity of the sulfide is exhausted,
partitioning is controlled by sorption [1], and in most of our
test sediments the IW concentrations of metal increased sharp-
ly enough that nearly "100% mortality resulted. The effect is
similar to the "throwing of a switch" at SEM/AVS ratios > 1.0.
This multiple order of magnitude increase in IW concentration
with a factor of 2 or 3 increase in sediment concentration
explains why the chemistry of the anaerobic sediments controls
the toxicity of metals to organisms living in aerobic sediment
microhabitats (e.g., the amphipods living in their burrows in
our experiments). It also explains why the toxicity of different
metals in the sediments in our experiment was similar on an
SEM/AVS basis (Fig. 6) even though their toxicities differ
markedly in water-only toxicity tests (Table 2). Finally, the
observed sequence in appearance of nickel, zinc, cadmium,
and copper into sediment IW is expected because this is the
order of the sulfide solubility products: nickel = -27.98, zinc
= -28.39, cadmium = -32.85, and copper = -40.94. There-
fore, metals availability and organism response observed in
our experiments were controlled by fundamental chemical re
actions [1,3]. Knowledge of these reactions allows prediction
of biological effects and can also be useful in explaining which
metal(s) might be causing sediment toxicity.
Available data from other spiked sediment experiments also
support our observations on the use of SEM/AVS ratios and
IWTUs to predict mortality in sediment toxicity tests. Our data
and data from freshwater tests using oligochaetes and snails
exposed to sediments spiked with cadmium [2] and saltwater
tests using polychaetes exposed to sediments spiked with cad-
mium, copper, lead, nickel, or zinc [5,9] and copepods exposed
to sediments spiked with cadmium [4] all yield similar results
when mortality is plotted against SEM/AVS ratio (Fig. 8).
These data describe tests with six freshwater and saltwater
species and sediments from seven sites, with AVS concentra-
tions ranging from 1.9 to 65.7 |unol/g dry weight and TOC
ranging from 0.15 to 10.6%. Mortality in these experiments
was sediment independent when plotted on an SEM/AVS ratio
basis. With the combined data, 98.9% of the 92 metals-spiked
sediments with SEM/AVS ratios <1.0 were nontoxic, while
73.5% of the 83 sediments with SEM/AVS ratios >1.0 were
toxic.
The presence of additional binding factors may account for
the fact that not all sediments with SEM/AVS ratios >1 0
caused increased mortality. In addition, organism behavior in
a toxicity test can control exposure and limit the impact of
metals hi sediments. Many of the sediments which had the
highest SEM/AVS ratios in excess of 1.0 that produced little
or no mortality were from experiments using the polychaete
Neanthes arenaceodentata (Fig. 8). The polychaetes did not
burrow hi most of these sediments and presumably were not
fully exposed to the metals in the sediment (see Fig. 3 in Pesch
et al. [9]). This same phenomenon may also explain the low
mortality of snails, Heliosoma sp., in freshwater sediments
with high SEM/AVS ratios (Fig. 8). These snails are epibenthir
and also have the ability to avoid contaminated sediments (G
Phipps, personal commication). Increased mortality was al
ways observed in sediments with SEM/AVS ratios >5.9 ir?
tests with the other four species tested.
The combined data from all available freshwater and salt
water tests also follow the same pattern as our saltwater am
phipod data when plotted on an IWTU basis (Fig. 9). Mortality
was not sediment specific when it was plotted against IWTU
and sediments with IWTUs of «C0.5 were generally nontoxic
Of the 88 sediments with IWTUs <0.5,98.9% were nontoxic,
while 77.6% of the 76 sediments with IWTUs >0.5 were toxic
When SEM/AVS ratio and IWTU were combined, our ability
to predict which sediments would be toxic was improved. Of
the 71 sediments with SEM/AVS ratios < 1.0 and IWTUs <0.5,
98.6% were nontoxic, while 87.7% of the 65 sediments with
SEM/AVS ratios > 1.0 and IWTUs >0.5 were toxic (Table 3).
This close relationship between IWTUs and sediment toxicity
was also observed in other studies with metal-contaminated
field sediments [11,24] as well as in studies with nonionic
organic chemicals both in the field [25,26] and in the labo-
ratory [27-29].
One limitation to the overall spiked sediment data we sum-
marize above is that all tests used acute exposures, and the
interpretation of these results may not be applicable to chronic
exposures. Bioaccumulation of metals was also not measured,
except in two cases [2,9]. The applicability of AVS and IWTU
normalizations to chronic exposures and bioaccumulation in
benthic organisms are discussed elsewhere [10,11,30-32]. Fur-
-------
2076 Environ. Toxicol Chem 15, 1996
WJ. Berry et al.
100
80
60
40
20;,
0001 0.01 0.1 1 10 100 1000
SEM/AVS-Cd
100
60
60
40
20
•A-A AA
-*-
0.001 0.01 0.1 1 10
SEM/AVS: Cu
100 1000
100
60
60
40
20
.A A
'*
0.001 0.01 0.1 1 10
SEM/AVS: Ml
100 1000
100
80
60
40
20-1
-4-
0.001 0.01 0.1 1 10 100 1000
SEM/AVS: Pb
100
60
f 60-1
40
20
0.001 0.01
-+-
0.1 1 10
SEM/AVS. Zn
100 1000
100
60
60
40
20
•4
0.001 0.01 0.1 1 10 100 1000
SEM/AVS.Cd + Cu+NI + Zn
Fig. 5. Percentage mortality of the amphipod A. abdaa (R. hudsoni in the Ninigret Pond sediment in the cadmium experiment) as a function of
the ratio of the sum of the molar concentrations of cadmium, copper, lead, nickel, and zinc simultaneously extracted with AVS to the molar
concentration of AVS (SEM/AVS) in three sediments: Long Island Sound (LIS), Ninigret Pond (NIN), and a 50/50 mixture of these two sediments
(Mix). Each panel represents data from a separate experiment. Data below the SEM detection limit are plotted at SEM/AVS = 0.001. A = Long
Island Sound sediment; • = Ninigret Pond sediment; and • = mix.
thermore, these data are from tests conducted in the laboratory
with homogenized spiked sediments. Therefore, conclusions
from these studies must be carefully applied when conducting
risk assessments with field sediments.
While our results also show that SEM/AVS and IWTU are
accurate predictors of the absence of mortality in a sediment
toxicity test, predictions of which sediments might be toxic
are less accurate. The fact that a significant number of sedi-
ments (20%) tested had SEM/AVS ratios of > 1.0 but did not
cause increased mortality indicates that other binding phases,
such as organic carbon [33], may also control bioavailability
in anaerobic sediments. While the SEM/AVS model of bio-
availability accurately predicts which sediments will not be
toxic, a model which utilizes SEM/AVS ratios or SEM - AVS
[23] but incorporates other binding phases might more accu-
rately predict which sediments will be toxic [34].
Similarly, a significant number of sediments with IWTUs
>O.S were nontoxic. This is likely the result of IW ligands,
which reduce the bioavailability and toxicity of dissolved met-
als, sediment avoidance by polychaetes or snails, or meth-
odological problems in contamination-free sampling of IW.
Ankley et al. [8] suggested that differences between the hard-
ness of the IW and that of the water in the water-only tests
might affect the accuracy of prediction of sediment toxicity
using IWTUs in freshwater unless the IWTUs are hardness
corrected. Furthermore, Green et al. [4] and Ankley et al. [8]
hypothesized that increased DOC in the IW reduced the bio-
availability of cadmium in their sediment exposures relative
to the water-only exposures. Green et al. [4] found that the
LCSO value for cadmium in an IW-only exposure was more
than twice that in a water-only exposure and that the LC50
value for cadmium in IW associated with sediments was more
than three times that in a water-only exposure. A significant
improvement in the accuracy of toxicity predictions using
IWTUs might be achieved if DOC binding in the IW is taken
into account.
-------
Predicting the toxicity of metal-spiked sediments
Environ Toxicol Chem. 15, 1996 2077
100
80
1 *
a? 40
20
oS
t
./ A*/f.
*A«***'At
01 0.01 0.1
•«»* **+* * * * +
•
.
IF*
;+*
1 10 100 101
"~
DO
SEM / AVS
Fig. 6. Percentage mortality of A. abdita (R. hudsoni in the Nimgret
Pond sediment in the cadmium experiment) as a function of the ratio
of the sum of the molar concentrations of cadmium, copper, lead,
nickel, and zinc simultaneously extracted with AVS to the molar con-
centration of AVS (SEM/AVS). All experiments combined Data be-
low the SEM detection limit are plotted at SEM/AVS = 0.001 • -
Cd; A = Cu; * = Ni; * = Pb; • = Zn; and + = mixed metals.
The results from our experiments and those published by
others as summarized in this article demonstrate that using
SEM, AVS, and IW metals concentrations to predict which
sediments that contain cadmium, copper, lead, nickel, and zinc
will not be toxic is quite certain. This is very useful, because
the vast majority of sediments found in the environment have
SEM/AVS ratios <1.0 [22,23,35]. An important consideration
here is that most existing data from field sites are from sed-
iment samples collected in the summer, when the seasonal
cycles of AVS produce the maximum binding potentials
[36,37]. Hence, sampling at seasons and conditions when AVS
is at Fnipi^al values is a must in evaluations of specific sed-
iments and in establishing the true level of overall concern
about metals in sediments. Predicting which of the'remaining
sediments (those with SEM/AVS ratios >1.0) will be toxic is
presently less certain, although the correct classification rate
100-
80
* 40
20
* * * +• **• **+* a 4»
* * . f
'
*
* +
*******
61)1 o.i i 10 100 looo loooo looooo
Interstitial Water Toxic Units
Fig. 7. Percentage mortality of A. abdita (R. hudsoni in the Nimgret
Pond sediment in the cadmium experiment) as a function of IWTUs.
In the individual metal experiments IWTU equals the IW concentra-
tion of the individual metal/A, abdita LC50 for that metal. The IWTU
for the mixed metals experiment is the sum of (IW Cd concnVCd
LC50 for A. abdita) + (IW Cu concnyCu LC50 for A. abdita) + (IW
Ni concnVNi LC50 for A. abdita) + (IW Zn concnJZn LC50 for A.
abdita). All experiments combined. Data below the detection limit
are plotted at IWTU = 0.01. • = Cd; A = Cu; * = Ni; * = Pb;
• — Zn; and += mixed metals
.g-
•g
2
»e
10O
80
60
40
20,
<
O1
00
«*r»^»%****
" o":
^
« = L.V.). copper (A = A.a.; A = C.c.), lead (4 = A.a.;
0 = C.c.), nickel (* = A.a.; -to = Ma.), zinc (• = A.a.; a = C.c ),
or a mixture of metals (+ = A.a.) as a function of the SEM/AVS
ratio Data below the SEM detection limit are plotted at SEM/AVS
= 0.001.
seen in these experiments (better than 70% of sediments pre-
dicted to be toxic were toxic) is probably better than that of
any other available method. An SEM/AVS ratio >1.0 should
trigger additional tiered assessments, which might include tox-
icity tests, measurements of IW metal, toxicity identification
evaluations (TIE), and characterization of the spatial (both
vertical and horizontal) and temporal distribution of chemical
concentration and toxicity. In this context, the SEM, AVS,
IWTU approach should be viewed as only one of many sed-
iment evaluation methodologies.
Of course, SEM, AVS, and IWTUs can only predict toxicity
or the lack of toxicity due to metals in sediments. They cannot
be used alone to predict the toxicity of sediments contaminated
with toxic concentrations of other contaminants, e.g., poly-
cyclic aromatic hydrocarbons. However, SEM/AVS ratios have
been used in sediment assessments to rule out metals as prob
able causative agents of toxicity [22]. Also, the SEM/AVS
approach to predicting the biological availability and toxicity
0.01
1 10 100 1000
mtefstttio! Water Toxic Units
10000 100000
Fig. 9 Percentage mortality of saltwater and freshwater benthic spe-
cies, including oligochaetes (L. variegatus, L.V.), polychaetes (C. cap-
itata, C.c., and M. arenaceodentata, N.a.), amphipods (A. abdita,
A.a.), harpacucoids (Amphiascus tenuiremis, A.t.), and snails (Heli-
soma sp., H.s.) exposed to sediments spiked with cadmium (® = A.t.,
see Fig 8 legend for definitions of other symbols), copper, lead, nickel,
zinc, or a mixture of metals as a function of IWTUs Data below the
detection limit are plotted at IWTU = 0.01.
-------
2078 Environ Toxicol. Chem. 15. 1996
W.J. Berry et al.
. Table 3 Accuracy of predictions of the toxicity of sediments from spiked sediment tests with
saltwater amphipods and combined freshwater and saltwater spiked sediment tests as
a function of SEM/AVS ratios, IWTUs, and both SEM/AVS and IWTUs
% Sediment
Study type
Parameter
Value
Nontoxic Toxic
Amphipods
(saltwater)
All species (fresh-
water and salt-
water)
SEM/AVS
IWTU
SEM/AVS, IWTU
SEM/AVS
IWTU
SEM/AVS, IWTU
£1.0
>1.0
<0.5
2=0.5
£1.0, <0 5
>1.0,a0.5
£1.0
>1.0
<0.5
20.5
£1.0,<0.5
>%1.0, S0.5
43
45
33
36
29
36
92
83
88
76
71
65
97.7
20.0
97.0
5.6
96.6
5.6
98.9
26.5
98.9
22.4
98.6
12.3
2.3
800
3.0
944
3.4
94.4
1.1
73.5
1.1
77.6
1.4
87.7
of cadmium, copper, lead, nickel, and zinc is applicable only
to anaerobic sediments that contain AVS, because in aerobic
sediments binding factors other then AVS control bioavail-
ability [38,39]. Measurement of IW metal may be useful for
evaluations of these metals in aerobic sediments and other
metals in sediments in general [12]. Even with theses caveats,
we believe that the use of SEM, AVS, and interstitial mea-
surements in combination are superior to all other currently
available sediment evaluation procedures to causally assess
the implications of these five metals associated with sediments.
Acknowledgement—We thank S. Benyi, A. Kuhn, and C. Schlekat
for graphics and database support and R. Burgess, R. Hoke, K. Scott,
and two anonymous reviewers for comments that improved the manu-
script W. Berry, J. Corbin, D. Robson, B. Rogers, and B. Shipley
were supported under U.S. Environmental Protection Agency (EPA)
Contract 68-C1-0005 to Science Applications International Corpo-
ration. D.M. Di Toro and I.D. Mahony were supported under U.S.
EPA Cooperative Agreement R812824010 with Manhattan College.
This document has been reviewed in accordance with U.S. EPA, Na-
tional Health and Environmental Effects Research Laboratory, policy
and approved for publication. The contents of this publication do not
necessarily reflect the views of the U.S. EPA. Mention of trade names
or commercial products does not constitute endorsement or recom-
mendation for use.
REFERENCES
1. Di Toro, D.M., JJ>. Mahony, DJ. Hansen, KJ. Scott, MA.
Hicks, S.M. Mays and MS. Redmond. 1990. Toxicity of cad-
mium in sediments: The role of acid volatile sulfide. Environ.
Toxicol. Chem. 9:1489-1504.
2 Carlson, A.R, G.L. Phipps, V.R. Mattson, PA. Kosian and
A.M. Cotter. 1991. The role of acid-volatile sulfide in deter-
mining cadmium bioavailability and toxicity in freshwater sedi-
ments. Environ. Toxicol. Chem. 10:1309-1319.
3. Di Toro, D.M., JJ>. Mahony, DJ. Hansen, KJ. Scott, A.R.
Carlson and G.T. Ankley. 1992. Acid volatile sulfide predicts
the acute toxicity of cadmium and nickel in sediments. Environ.
Sci. Technol. 26:96-101.
4. Green, A.S., G.T. Chandler and E.R. Blood. 1993. Aqueous-,
pore-water-, and sediment-phase cadmium: Toxicity relationships
for a meiobenthic copepod. Environ. Toxicol. Chem. 12:1497-
1506.
5. Casas, A.M. and E.A. Crecelius. 1994. Relationship between
acid volatile sulfide and the toxicity of zinc, lead, and copper in
marine sediments. Environ. Toxicol. Chem. 13:529-536.
6 Kemp, P.F. and R.C. Swartz. 1988. Acute toxicity of interstitial
and particle-bound cadmium to a marine infaunal amphipod. Mar.
Environ. Res. 26:135-153.
7. Ankley, G.T., V.R. Mattson, E.N. Leonard, C.W. West and
J.L. Bennett. 1993. Predicting the acute toxicity of copper in
freshwater sediments: Evaluation of the role of acid-volatile sul-
fide. Environ. Toxicol. Chem. 12:315-320.
8. Ankley, G.T., et al. 1991. Acid-volatile sulfide as a factor me-
diating cadmium and nickel bioavailability in contaminated sed-
iments. Environ Toxicol. Chem. 10.1299-1307.
9. Pesch, C.E., DJ. Hansen, W.S. Boothman, WJ. Berry and
' J.D. Mahony. 1995. The role of acid-volatile sulfide and inter-
stitial water metal concentrations in determining bioavailability
of cadmium and nickel from contaminated sediments to the ma-
rine polychaete, Neanthes arenaceodentata. Environ. Toxicol.
Chem. 14:129-141.
10. Hare, L., R. Carignan and NLA. Huerta-Diaz. 1994. A field
experimental study of metal toxicity and accumulation by benthic
invertebrates: Implications for the acid volatile sulfide (AVS)
model. Limnol. Oceanogr. 39:1653-1668.
11. Hansen, DJ., J.D. Mahony, WJ. Berry, SJ. Benyi, JJVI. Cor-
bin, SJ>. Pratt, D.M. Di Toro and M.B. Able. 1996. Chronic
effect of cadmium in sediments on colonization by benthic marine
organisms: An evaluation of the role of interstitial cadmium and
acid-volatile sulfide in biological availability. Environ. Toxicol.
Chem. 15:2126-2137.
12. Ankley, G.T., NA. Thomas, D.M. Di Toro, DJ. Hansen, JJ).
Mahony, WJ. Berry, R.C. Swartz and RA. Hoke. 1994. As-
sessing potential bioavailability of metals in sediments: A pro-
posed approach. Environ. Manage. 18:331-337.
13. Luoma, S.N. 1989. Can we determine the biological availability
of sediment-bound trace elements? Hydrobiologia 176/177:379-
396.
14. Goldhaber, M.B. and I.R. Kaplan. 1974. The sulfur cycle. In
E.D. Goldberg, ed.. The Sea, Vol. 5—Marine Chemistry John
Wiley & Sons, New York, NY, USA, pp. 569-655.
IS. Gonzales, A.M., J.D. Mahony and D.M. Di Toro. 1992. The
role of organic carbon in the toxicity of anoxic sediments con-
taminated with copper and other metals: An experimental study.
Abstracts, 13th Annual Meeting, Society of Environmental Tox-
icology and Chemistry, Cincinnati, OH, USA, November 8-12,
p. 162.
16. Spehar, RX,. and J.T. Flandt. 1986. Acute and chronic effects
of water quality criteria-based metal mixtures on three aquatic
species Environ. Toxicol. Chem. 5:917-931.
17. Scott, KJ. and MS. Redmond. 1989. The effects of a contam-
inated dredged material on laboratory populations of the tubi-
culous amphipod, Ampelisca abdita. In U.M. Cowgill and L R
Williams, eds.. Aquatic Toxicology and Hazard Assessment, Vol-
ume 12. STP 1027. American Society for Testing and Materials,
Philadelphia, PA, USA, pp. 289-303.
18. Hamilton, MA., R.C. Rosso and R.V. Thurston. 1977.
Trimmed Spearman-Karber method for estimating median lethal
concentrations in toxicity bioassays. Environ. Sci. Technol 11:
714-719, Correction 12.417 (1978).
19. Mearns, AJ., R.C. Swartz, J.M. Cummins, PA. Dinnel, P.
Plesha and P.M. Chapman. 1986. Inter-laboratory comparison
of a sediment toxicity test using the marine amphipod, Rhepox-
ynius abromus. Mar. Environ. Res. 19.13-37.
-------
Predicting the toxicity of metal-spiked sediments
Environ Toxicol. Chem 15, 1996 2079
20 U.S. Environmental Protection Agency. 1994. Statistical sum-
mary, EMAP—Estuaries, Virginian Province—1991. EPA/620/
R-94/005. Technical Report Office of Research and Develop-
ment, Washington, DC
21 Stephen, C.E., D.I. Mount, DJ. Hansen, J.H. Gentile, GA.
Chapman and WA. B rungs. 1985. Guidelines for denying nu-
merical national water quality criteria for the protection of aquatic
organisms and their uses PB8S-227049 National Technical In-
formation Service, Springfield, VA, USA.
22 Wolfe, D.S., S.B. Bricker, E.R. Long, KJ. Scott and G.B.
Thursby. 1994. Biological effects of toxic contaminants in sed-
iments from Long Island Sound and environs. Technical Mem-
orandum NOS ORCA 80. National Oceanic and Atmospheric
Administration, Silver Spring, MD, USA
23 Hansen, DJ., et al. 1996. Predicting the toxicity of metal-con-
taminated field sediments using interstitial concentration of metals
and acid-volatile sulfide normalizations. Environ. Toxicol. Chem.
15:2080-2094.
24. Swartz, R.C., G.R. Dittsworth, D.W. Sennits and J.O. Lam-
berson. 1985. Sediment toxicity to a marine infaunal amphipod:
Cadmium and its interaction with sewage sludge. Mar. Environ.
Res. 18:133-153
25. Hoke, R.A., G.T. Ankley, A.M. Cotter, T. Goldstein, PA. Ko-
sian, G.L. Phlpps and F. VanderMeiden. 1994. Evaluation of
equilibrium partitioning theory for predicting acute toxicity of
field-collected sediments contaminated with DDT, DDE, and
DDD to the amphipod, Hyalella azteca. Environ Toxicol. Chem.
13.157-166.
26 Swartz, R.C., FA. Cole, J.O. Lamberson, S.P. Ferraro, D.W.
Schults, WA. DeBen, Henry Lee H and RJ. Ozretich. 1994.
Sediment toxicity, contamination, and amphipod abundance at a
DDT- and dieldrin-contaminated site in San Francisco Bay. En-
viron. Toxicol. Chem. 13:949-962.
27 Di Tore, D.M., et al. 1991. Technical basis for establishing sed-
iment quality criteria for non-ionic organic chemicals using equi-
librium partitioning Environ. Toxicol. Chem. 10:1541-1583.
28 Adams, WJ., RA. Kimerle and R.G. Mosher. 1985. Aquatic
safety assessment of chemicals sorbed to sediments. In R.D. Card-
well, R. Purdy and R.C. Banner, eds., Aquatic Toxicology and
Hazard Assessment: Seventh Symposium. STP 854. American
Society for Testing and Materials, Philadelphia, PA, USA, pp.
429-453.
29. Swartz, R.C., D.W. Schults, T.H. DeWitt, G.R. Dittsworth and
J.O. Lamberson. 1990. Toxicity of fluoranthene in sediment to
marine amphipods: A test of the equilibrium partitioning approach
to sediment quality criteria. Environ Toxicol. Chem. 9.1071-
1080.
30. Ankley, G.T. 1996. Evaluation of metal/acid-volatile sulfide re-
lationships in the prediction of metal bioaccumulation by benthic
macroinvertebrates Environ. Toxicol Chem. 15:2138-2146.
31. Sibley, P.K., G.T. Ankley, A.M. Cotter and E.N. Leonard.
1996 Predicting chronic toxicity of sediments spiked with zinc:
An evaluation of the acid-volatile sulfide model using a life-cycle
test with the midge Chironomus tentans Environ. Toxicol. Chem.
15:2102-2112.
32. DeWitt, T.H., R.C. Swartz, DJ. Hansen, D. McGovern and
WJ. Berry. 1996. Bioavailabihty and chronic toxicity of cad-
mium in sediment to the estuanne amphipod, Leptocheirus plu-
mulosus. Environ Toxicol. Chem. 15:2095-2101.
33. Mahony, J.D., D.M. Di Toro, A.M. Gonzalez, WJ. Berry and
G.T. Ankley. 1991. A sediment component in addition to acid
volatile sulfide that may further control toxicity of metals. Ab-
stracts, 12th Annual Meeting, Society of Environmental Toxi-
cology and Chemistry, Seattle, WA, USA, November 3-7, p. 91.
34 Mahony, JJ>., D.M. Di Toro, A.M. Gonzalez, M. Curto, M.
Dilg, L.D. De Rosa and LA. Sparrow. 1996. Partitioning of
metals to sediment organic carbon. Environ. Toxicol. Chem 15:
2187-2197.
35. Leonard, E.N., G.T. Ankley and RA. Hoke. 1996. Evaluation
of metals in marine and freshwater surficial sediments from the
Environmental Monitoring and Assessment Program relative to
proposed sediment quality criteria for metals. Environ. Toxicol.
Chem. 15:2221-2232.
36. Boothman, W.S. and A. Helmstetter. 1992. Vertical and seasonal
variability of acid volatile sulfides in marine sediments. EPA 600/
X-93/036. U.S. Environmental Protection Agency, Narragansett,
RI.
37. Leonard, E.N., V.R. Mattson, DA. Benoit, RA. Hoke and G.T.
Ankley. 1993. Seasonal variation of acid volatile sulfide in sed-
iment cores from three northeastern Minnesota lakes. Hydro-
biologia 271:87-95.
38. Di Toro, D.M., F. Harrison, E. Jenne, S. KarickofT and W.
Lick. 1987. Synopsis of discussion session 2. Environmental fate
and compartmentalization. In K.L. Dickson, A.W. Maki and W. A.
B rungs, eds.. Fate and Effects of Sediment-Bound Chemicals in
Aquatic Systems Pergamon, New York, NY, USA, pp. 136-147.
39. Tessier, A., Y. Couillard, P.G.C. Campbell and J.C. Audair.
1993. Modeling Cd partitioning in oxic lake sediments and Cd
concentrations in the freshwater bivalve Andonata grandis. Lim-
nol. Oceanogr. 38:1-17.
-------
Environmental Toxicology and Chemistry, Vol 18, No 1, pp 30-39,1999
© 1999 SETAC
Printed in the USA
0730-7268/99 $9 00 + 00
Annual Review
SILVER TOXICITY TO CH1RONOMUS TENTANS IN TWO FRESHWATER SEDIMENTS
DANIEL J. CALL,*! CHRISTINE N. POLKINGHORNE.! THOMAS P. MARKEE,! LARRY T. BROOKED
DIANNE L. GEiGER.t JOSEPH W. GoRSUCH,^: and KENNETH A. ROBILLARD}:
fLake Superior Research Institute, University of Wisconsin-Superior, Superior, Wisconsin 54880, USA
^Eastman Kodak Company, Rochester, New York, New York 14652 USA
(Received 27 May 1998, Accepted 24 July 1998)
Abstract—Sediment collected from two freshwater lakes, West Bearskin Lake (Cook, MN, USA) and Bond Lake (Douglas, WI,
USA), was characterized for grain size, total organic carbon, (TOC), acid-volatile sulfides (AVS), simultaneously extracted metals
(SEM), and iron (Fe) Both sediments had low levels of TOC (< 1.0 percent) and AVS (< 1.1 u,mol/g) West Bearskin Lake sediment
contained more small (silt and clay) particles than Bond Lake, which was 95% sand West Bearskin Lake also had higher SEM
(S, SEM = 0815 vs 0 074 fimol/g) and had an Fe content that was approximately 30-fold greater than that of Bond Lake These
sediments were amended with AgNO, in a series of concentrations, some of which were intended to exceed the total silver (Ag)-
bindmg capacity of the sediments, allowing for the appearance of dissolved Ag in pore water (PW) Sediment toxicity tests were
then designed such that the AgNO, amendment levels would result in PW concentrations that bracketed the 10-d concentration
causing 50% lethality (LC50) for dissolved Ag of 0 057 mg/L, as determined in a toxicity test in water alone (i e , without sediment
present). The 10-d LC50 values for Chironomus tent cms, based upon nominal additions of Ag to the sediments, were 2 75 and
1 17 g Ag per kilogram dry sediment for West Bearskin and Bond Lake sediments, respectively An LC50 value based upon
dissolved Ag in the PW was determined only for Bond Lake sediment and was approximately 275 times greater than the water-
only LC50 value This indicated that a high proportion of the dissolved fraction was not readily bioavailable to cause lethality A
reduction in PW pH and the displacement of other metals from sediment into PW with Ag additions to the sediment likely contributed
to the observed mortalities and weight losses, particularly at the higher exposure levels The concentrations of Ag in these sediments
that resulted in biological effects are considerably higher than levels reported in the environment.
Keywords—Silver Toxicity Freshwater Sediments Chironomus tentans
INTRODUCTION
It has been estimated that one fourth to two fifths of the
silver (Ag) released to the environment enters the aquatic com-
ponent [1,2]. Silver is a highly reactive element that may exist
in the aquatic environment in a number of chemical forms.
However, only the free ionic form of Ag (Ag*) is reportedly
toxic to freshwater organisms [3-8]. This free ionic form is
present only at very low concentrations due to the fact that it
readily reacts with different inorganic and organic ligands that
are present in the freshwater environment. Prior to the use of
ultraclean analytical techniques, measured concentrations of
dissolved Ag in freshwater samples in the United States were
usually reported as <0.2 u,g/L [9] Studies in which ultraclean
sampling and analytical procedures were used have reported
even lower concentrations (<0.001 u,g/L) in water [10].
Most of the Ag-hgand complexes that form in freshwater
[11] settle to the bottoms of streams or lakes to become a part
of the sediment [12,13]. Concentrations of Ag in sediment
from highly industrialized areas range from 1 to 150 mg/kg,
whereas background concentrations in nonurban areas are usu-
ally <0.1 mg/kg [9] The surface sediments act as a biogeo-
chemical reactor, controlling the return of metals to the over-
lying water or their retention in the underlying sediment re-
pository [14]. In the sediment, the dynamics of Ag speciation
are dependent upon a number of factors, including the geo-
logical characteristics of the sediment, the composition of nat-
* To whom correspondence may be addressed
(dcall@staffuwsuper.edu) Contribution 106, Lake Superior Research
Institute
ural ligands in the overlying and pore water (PW), seasonal
changes that influence biological and chemical cycling (e.g.,
pH, redox status, acid-volatile sulfideTAVS] content, organic
carbon content), and physical influences upon the sediment
(e.g., bioturbation, storm events, and dredging activities). Be-
cause it is the ionic form of the metal in PW that is of toxi-
cological concern to benthic organisms [15-17], it is important
to obtain an understanding of the interactions between Ag and
sediments in determining the concentrations of dissolved Ag
that might exist in a bioavailable and toxic form in PW This
paper is one in a series of papers on Ag in this issue. It ex-
amines the current state of knowledge regarding Ag in fresh-
water sediments and reports our research results on the toxicity
of Ag-spiked sediments toward larvae of the freshwater Dip-
teran insect Chironomus tentans. Berry et al. [18] have applied
the data from this study in evaluating PW toxicity and si-
multaneously extracted metals (SEM)/AVS relationships
MATERIALS AND METHODS
Lake sediment collection, handling, and properties
Two lakes were selected that were known to contain sed-
iments of relatively low total organic carbon (TOC) and AVS
levels Low TOC and AVS sediments were selected for study
based upon the rationale that these sediments would represent
worst case conditions for then: capacities to bind Ag regarding
the TOC and AVS binding phases Sediment was collected
from West Bearskin Lake (Cook, MN, USA) and Bond Lake
(Douglas, WI, USA) with Eckman and Ponar samplers It was
placed into polyethylene containers, and refrigerated (~4°C)
30
-------
Silver toxicity in freshwater sediments
Environ Toxicol Chem 18, 1999 31
Table 1 Physicochemical properties of test sediments
Sediment
Property
West Bearskin Lake
Bond Lake
Particle size (%)
Gravel
Sand
Silt
Clay
Dry weight (%)
Total organic carbon (%) ^
Acid-volatile sulfide ((unol/g)1
Simultaneously extracted metals (SEM, |Amol/g)b
Cadium
Copper
Lead
Nickel
Silver
Zinc
2SEM (u,mol/g)c
Iron ((xmol/g)d
n = 1
00
54.2
37.6
82
59 7 (n = 16)
087 (n = 1)
<0 1 (n = 8)
n = 2
0.003
0.085
0.042
0.121
0.004
0.560
0815(n = 2)
165 (n = 1)
n = 1
13
953
34
00
76 1 (n= 10)
022(n = 1)
<0.1 (n = 15)
n = 4
0.001
0.015
0.007
0013
0006
0035
0.077 (n = 4)
5.8 (n = 7)
• Determined dunng toxicity tests
b Mean as determined from control sediment dunng toxicity tests
c SEM = simultaneously extracted metals.
d Determined by SEM extraction procedure
until mixed. Shortly after collection, the sediment was thor-
oughly homogenized in a stainless steel drum with a com-
mercial drill equipped with a mortar paddle, apportioned back
into the polyethylene buckets, and refrigerated until used in
tests. The sediment was stored for 10 and 4 to 5 months prior
to the initiation of Ag equilibration and toxicity tests for West
Bearskin and Bond Lakes, respectively. Sediment samples
were remixed prior to use in each type of sediment test. Sam-
ples were analyzed for particle size distribution, dry weight,
and TOC by a contract laboratory Acid-volatile sulfide, SEM
and iron (Fe) content were also determined (Table 1), for which
methods are described below.
Silver toxicity test in water
A 10-d water-only toxicity test was conducted with 11-d-
old larvae of C. tentans exposed to AgNO3. The exposure
chambers were 300-ml high-form beakers containing two
mesh-covered holes (12-mm diameter) in the walls of the beak-
ers to allow for an outflow of water into the aquarium in which
they were placed. Ten larvae were exposed in duplicate to five
exposure concentrations, plus a control. The nominal total Ag
concentrations were 0, 12.5, 25, 50, 100, and 200 u,g/L. Mean
measured dissolved Ag concentrations of <5.0, <5.0, <5.0,
13.0, 66.0, and 155 p,g/L for the control and five treatments
were provided by pumping test solutions directly from stock
containers with a multihead peristaltic pump (Cole-Farmer In-
strument, Chicago, IL, USA) to the bottom of each exposure
beaker at the rate of approximately 1.0 ml/min. The AgNO3
stock solutions were prepared in laboratory water, which was
dechlorinated municipal water that had been passed through
an ion-exchange resin column for removal of metal ions.
A thin layer of quartz sand was added to each beaker pnor
to the addition of larvae to reduce stress. Each test chamber
was supplied with 6.0 mg of Tetrafin® slurry daily for the
larvae to feed upon. Daily measurements were made at each
exposure concentration of temperature, dissolved oxygen, and
pH in alternating replicates Hardness, alkalinity, and conduc-
tivity were measured twice dunng the test near the beginning
and end in control, low, medium, and high treatments. The
overall means for test temperature, dissolved oxygen, and pH
were 23.0 ± 0.2°C, 7.7 ± 0.4 mg/L, and 7.4 ± 0.1 units,
respectively (n = 78). Mean values for hardness (n = 8),
alkalinity (n = 8), and conductivity (n = 24) were 52.1 ± 4.8
mg/L, 45.0 ± 1.4 mg/L, and 147 ± 5.41 u,mhos/cm, respec-
tively. Total, dissolved, and free ionic Ag were measured in
all replicates on days 0 and 10 of the exposure and in alternate
replicates on days 3 and 6. Total Ag was the amount measured
by atomic absorption (AA) spectroscopy in unfiltered acidified
samples. Dissolved Ag was procedurally defined as the Ag
that passed through a 0.2-^m polyethersulfone membrane
(Gelman Sciences, Ann Arbor, MI, USA) and was measured
by AA spectroscopy. Free Ag* was the form that passed
through the 0.2-u.m membrane and was measured by ion-spe-
cific electrode.
Silver binding capacity experiments
Preliminary spiking experiments without animals present
were conducted to determine the binding capacity of each
sediment for Ag. AgNO3 was added to the sediment in a series
of concentrations, and dissolved Ag concentrations were mea-
sured in the PW. It was assumed that dissolved Ag would not
be measurable in the PW until the sediment binding capacity
was reached, at which point it would appear in the PW in
increasing concentrations with increasing AgNO3 spiking lev-
els. The objective was to ascertain the spiking levels needed
to yield concentrations of dissolved Ag in the PW that would
bracket the 10-d dissolved Ag LC50 value for C. tentans as
determined in the water-only toxicity test. These experiments
were conducted by spiking AgNQ3 solutions into homogenized
sediment in 2-L fluorinated polyethylene containers. The
spiked sediment was stirred with a motorized stainless steel
stirrer for 5 mm under a nitrogen atmosphere. An acrylic plas-
tic peeper containing two 6.0-ml chambers covered with a 0 2-
u,m pore size polyethersulfone membrane and containing de-
gassed, deiomzed water was inserted into the sediment, and
the container was flushed with nitrogen, covered with a screw
lid, and placed into a refrigerator at 4°C. The peepers were
withdrawn from the sediment at weekly intervals, with a new
-------
32 Environ Toxicol Chem 18, 1999
D.J Call et al
peeper inserted each time to replace the withdrawn peeper. The
water within the peeper chambers was withdrawn with an Ep-
pendorf pipette, placed into Teflon* bottles cleaned with am-
monium hydroxide and nitric acid, acidified to pH ^2 with
concentrated nitric acid (trace metals grade), and analyzed for
Ag and other metals. Concentrations of Ag in the peeper cham-
ber water (i.e., sediment PW) were procedurally defined as
dissolved, having passed through the 0.2-u.m pore size mem-
brane. The spiked sediment samples were analyzed over 2 to
% weeks to determine if an equilibrium had been established
based upon constant concentrations of dissolved Ag in the
peepers on successive sampling dates.
Silver tbxicity tests in sediment
West Bearskin Lake. Based upon results from the water-
only toxicity test and the Ag binding capacity study, West
Bearskin Lake sediment was spiked with AgNO3 at nominal
concentrations of 1.7, 2.2, 3.0, 4.0, 5.4, 7.2, and 9.6 g Ag per
kilogram dry sediment. The test was performed in a modified
system based upon that of Benoit et al. [19], with test con-
ditions as described in standard procedures [20]. Sediment
control and sand performance controls were also tested, with
mean survival rates of 80.0 and 93.3%, respectively. In the
sand performance controls, the mean dry weights averaged 0.8
mg per individual, indicating adequate growth for test ac-
ceptability [20]. Exceptions to standard procedures were that
the system provided for renewal of overlying water to the
aquaria at a rate equivalent to approximately four volume ad-
ditions daily rather than two, and only three biology replicates
were used rather than four. Screened exposure beakers (300
ml) were used for both biological and chemical testing at each
exposure level, with four beakers used for chemistry mea-
surements. Each beaker received approximately 100 ml of ei-
ther AgNOj-amended sediment, control sediment, or quartz
sand (used as a test performance control). The chemistry beak-
ers each received a two-chambered peeper covered with a 0.2-
ujn polyethersulfone membrane, positioned to a depth within
the sediment such that the lower chamber was buried below
the surface of the sediment to collect PW and the upper cham-
ber was in the overlying water. The beakers containing sedi-
ment were placed into the aquaria (30.5 X 15.0 X 8.0-10.0
cm, length X width X water depth) of the toxicity test system
7 d prior to the introduction of test animals, with renewal of
overlying water occurring during this time period. This al-
lowed for elimination of any excess nitrate ion concentration
in the overlying water. Temperature, dissolved oxygen, and pH
were measured in all test chambers on days 0 and 10 and in
at least one replicate per treatment on intermediate days (n =
199). The overall respective test means (±SD) were 23.3
(0.6)°C, 7.5 (0.7) mg/L, and 7.3 (0.4) units.
Ten second-third instar C. tertians larvae (10-11 d old)
were randomly added to each replicate, including the chemistry
beakers, 39 d after Ag addition to the sediment. Each replicate
was fed the equivalent of 6.0 mg of dry solids daily, as in the
water-only test. Chemistry beakers were sampled on days 0,
5, and 10 of the toxicity test. A peeper was removed from one
replicate at each exposure concentration, and the peeper cham-
ber water was analyzed for pH and Ag. A subsample of the
peeper water was also analyzed for concentrations of zinc (Zn),
nickel (Ni), lead (Pb), copper (Cu), total ammonia and non-
purgeable dissolved organic carbon. Cadmium (Cd) was not
analyzed due to the very low concentration in the SEM extract.
The overlying water was also sampled and measured for Ag
and other metals.
On day 10, the biology test beakers were sieved. Survivors
were counted and placed into preweighed pans for each rep-
licate and dried in an oven at 105°C for ~24 h. Survival,
individual dry weight, and total biomass per replicate were
statistically analyzed.
Bond Lake. Two 10-d toxicity tests were performed with
Bond Lake sediment, due to the fact that the PW-dissolved Ag
concentrations in the first test were less than expected based
on the initial binding study, and the exposures were not lethal
to C. tentans. The AgNO3 additions to the bulk sediment in
the first test resulted in nominal concentrations of 0.005,0.01,
0.03, 0.08, 0.20, and 0.50 g Ag per kilogram dry sediment.
The AgNOj additions in the second test yielded nominal con-
centrations of 0.03, 0.08, 0.20, 0.50, 1.2, and 3.1 g Ag pei
kilogram dry sediment. Each test also contained sediment and
sand performance controls, with respective mean survival rates
of 93.3 and 100% for test 1 and 96.7 and 83.3% for test 2.
The tests were performed as described for West Bearskin Lake.
Larvae (2nd/3rd instar) were added to the sediment 22 d after
the AgNO3 additions in both tests. For test 1, the overall means
(±SD) for temperature, dissolved oxygen, and pH (n = 176)
were 23.1 (0.3) "C, 7.6 (0.8) mg/L, and 7.4 (0.2) units, re-
spectively. For test 2, the respective means (n = 176) were
23.1 (1.0) °C, 7.2 (1.0) mg/L, and 7.4 (0.2) units, respectively.
In Bond Lake, PW was analyzed for Ag, Zn, and iron (Fe)
Nickel, Pb, Cu, and Cd were not analyzed due to their presence
only at very low concentrations in the SEM extract.
Analytical chemistry. Total and dissolved Ag concentra-
tions were measured using a Varian SpectrAA200 AA spec-
trophotometer. Samples were acidified to a pH £2 using trace
metal grade nitric acid. Flame AA was utilized in all cases,
with the exception of the first Bond Lake toxicity test, where
graphite furnace AA was used to measure the dissolved Ag
concentrations due to the fact that the concentrations were
below the flame detection limit. Free ionic Ag+ was measured
using an ion-specific electrode (Orion 9616BN). For total, dis-
solved and free ionic Ag measurements, a calibration curve
was developed from a minimum of four standards bracketing
sample concentrations. If necessary, samples were diluted with
deionized water to bring them within the range of the stan-
dards. Measurements of additional dissolved metals in peeper
samples were analyzed using flame AA (Vanan SpectrAA200).
The metals that were quantified included Cu, Zn, Fe, Ni, and
Pb.
Nonpurgeable dissolved organic carbon was analyzed using
a Shimadzu TOC 5050A organic carbon analyzer according
to U.S. Environmental Protection Agency (U.S. EPA) method
415.1 [21]. Samples (0.5 ml) were collected from peepers, 0.5
ml of deionized water was added to provide an adequate sam-
ple volume, and the samples were preserved with 5 uJ of 2
M HC1. Upon analysis, each sample was sparged with air for
1 min to eliminate any inorganic carbon. The concentrations
of organic carbon were then quantified relative to a standard
curve developed from four standards.
Acid-volatile sulfide and SEM analyses were conducted
using the HC1 gas train method [22]. A measured mass of
frozen sediment was placed into a reaction flask with 100 ml
of deionized water and purged with nitrogen, and 20 ml of
6M HC1 was added. Nitrogen gas was bubbled into the reaction
flask and then passed through a series of glass impinger bottles.
The first impinger bottle was filled with KHP solution and the
-------
Silver toxicity in freshwater sediments
final two were filled with 0.1M AgNO3. Sulfide present in a
sample was converted to hydrogen sulfide, which passed
through the first impmger bottle and reacted with the AgNO3
in the second to form insoluble Ag2S precipitate. The Ag,S
precipitate was filtered and weighed to determine AVS. The
HC1 extract containing the SEMs was filtered to separate the
extract from the sediment. This extract was analyzed using
flame AA for Ag, Cd, Cu, Ni, Pb, Zn, and Fe.
Ammonia was measured both in the overlying water and
PW collected from peeper samples as well as overlying H2O
collected with an Eppendorf pipettor. In both cases, samples
were treated with 10 M NaOH just prior to analysis. Overlying
water samples were analyzed using an ion-specific electrode
(Orion 9512), and peeper samples were analyzed using an
ammonia microelectrode (MI-740 Microelectrodes, London-
derry, NH, USA).
Statistical Analyses. Estimates of concentration causing
50% lethality (LC50) were obtained using the trimmed Spear-
man-Karber program (version 1.5) from a U.S. EPA statistical
software package [23]. Larval survival, individual dry weight,
and total biomass per replicate were analyzed by one-way
analysis of variance, and a comparison of control and treatment
means was performed by Bonferroni's t-test, as provided in a
TOXSTAT* software package (University of Wyoming, Lar-
amie, WY, USA).
RESULTS AND DISCUSSION
Silver toxicity in water
Ten-day LC50 values (95% confidence intervals) for C.
tentans larvae exposed to Ag in the absence of sediment were
0.063 (0.036-0.108), 0.057 (0.030-0.108), and 0.035 (0.017-
0.075) mg/L for total Ag, dissolved Ag, and free Ag+, re-
spectively. Dry weights of surviving larvae were less than half
of control larval weights at the two highest exposures where
the mean free Ag* concentrations were 0.041 and 0.110 mg/
L. Based upon significant (p = 0.05) reductions of 82.4 and
95.1% in dry weight at the two highest exposures of 0.066
and 0.155 mg/L dissolved Ag, respectively, but not at 0.013
mg/L, the no observable effect concentration (NOEC) and low-
est observable effect concentration (LOEC) range was 0.013
to 0.066 mg/L of dissolved Ag. Dry weight was only slightly
reduced (11.1%) at 0.013 mg/L. The observed sensitivity of
C. tentans to Ag in our laboratory water is lower than the 10-
d LC50 of 0.259 mg/L obtained by Rodgers et al. [24]. They
reported a 10-d NOEC value of 0.125 mg/L of Ag. The lab-
oratory water used in our study may have had fewer ligands
to bind the Ag+ ion than the pond water used by Rodgers et
al. [24], and some of the larger colloids that may have passed
through their 0.45-u.m filter may have been retained by our
0.2-u.m filter. However, our LC50 value was greater man a 48-
h LC50 value of 0.010 mg/L of total Ag for third instar larvae
of this species in an unfed test in which the Ag concentrations
were not measured [25].
Silver binding capacity of sediments
The preliminary Ag equilibration test results (Table 2) in-
dicated that PW concentrations in AgNO3-amended West Bear-
skin Lake sediment had come into equilibrium within 15 d.
Pore-water Ag concentrations were generally similar on days
15, 22, and 29. However, the day 8 concentrations at the three
highest Ag.additions were much higher than later measure-
ments. From this study, it was determined that a minimal equil-
Environ. Toxicol. Chem. 18, 1999 33
Table 2. Concentrations of "dissolved" silver (Ag) in pore water from
peeper chambers buned within West Bearskin Lake sediment in
preliminary equilibration studies
Nominal
sediment
Ag
(g Ag/kg)
Control
1.7
2.2
3.0
4.0
5.4
7.2
9.6
Dissolved Ag concentration (mg/L)
Day 8
<0.005
<0.005
<0.005
0.005
<0.005
1.03
4.09
13.9
Day 15
<0.005
0.010
0.007
0.010
0.012
0.019
0022
0087
Day 22
<0.005
<0.005
0.005
0.008
0.021
0.025
0.041
0.080
Day 29
<0.005
0006
<0.005
0.006
0.010
0.013
0.017
0.092
ibration period of 15 days was required for West Bearskin
Lake sediment. It was deduced that this equilibration time
would also be adequate for Bond Lake sediment, with fewer
Ag binding sites in the form of AVS, TOC, and total surface
area of particles in this sediment. Silver equilibration periods
of 39 and 22 d were used for West Bearskin and Bond Lake
sediments, respectively, before animals were added in the sed-
iment toxicity tests. These sediments had been stored under
refrigeration for up to 10 months prior to use and likely un-
derwent some chemical changes during this storage period.
Plots of the PW Ag concentrations during both the equilibra-
tion and toxicity testing periods are shown in Figure 1.
The concentrations of Ag and other heavy metals analyzed
in the peeper chamber water during the tests at the various
AgNOj spiking levels (Tables 3 and 4) show that West Bearskin
Lake sediment had a greater binding capacity for Ag* than
Bond Lake sediment. In West Bearskin Lake sediment, no
dissolved Ag was measured in the PW at spiking levels up to
2.2 g Ag per kilogram dry sediment At 3.0 g Ag per kilogram
dry sediment, dissolved Ag in the PW reached a maximum,
as did toxicity. It is unclear as to why this intermediate con-
centration had the highest PW concentration of Ag. Silver
displaced the other metals as the spiking levels increased, with
particularly high concentrations of Zn and Ni released into the
PW. The pH decreased with increased additions of AgNO3.
In Bond Lake sediment, no appreciable concentrations of
dissolved Ag were measured in the peeper chambers at spiking
levels up to 0.08 g Ag per kilogram dry sediment. However,
at spiking levels of 0.20 g Ag per kilogram dry sediment and
higher, dissolved Ag concentrations increased with increasing
AgNO3 additions. A'dramatic increase in dissolved PW Ag
occurred between'the 0.20 and 0.50 g Ag per kilogram sedi-
ment spiking levels. Concentrations of Zn in PW also increased
as with West Bearskin Lake sediment and, at comparable spik-
ing levels, were displaced to the PW in greater concentrations
in Bond Lake sediments than in West Bearskin Lake sediments.
The concentrations of Fe in PW decreased dramatically be-
tween the spiking levels of 0.50 and 1.2 g Ag per kilogram
sediment. This may indicate that more Fe oxides had reacted
with the additional Ag+ to form either a precipitate or a large
(i.e., >0.2 uin) soluble Ag-Fe oxide complex that could not
pass through the 0.2-u.m peeper membrane.
These data suggest that West Bearskin Lake sediment can
effectively bind 2.2 g Ag per kilogram dry sediment before
any dissolved Ag occurs in the PW. By comparison, Bond
Lake sediment can effectively bind only 0.08 g Ag per kilo-
gram dry sediment. This difference in binding capacity may
-------
34 Environ. Toxicol. Chem 18, 1999
A West Bearskin Lake
D.J. Call et al.
B. Bond Lake
0001
-•-Control
-o- 1 7 g Ag/kg Diy Sod
-T- 2 .2 g Ag/kg Dry Sad
-V- 30gAg/kgDiySed
-•- 4 0 g Ag/kg Dry Sod
-o-S 4 g Ag/kg Dry Sod
-•- 7.2 g Ag/kg Dry Sad
-0-B6 g Ag/kg Dry sed
•0 Control
-0-0 03 g Ag/kg OiySad
-V- 0 20 g Ag/kg Dry 8*d
-•- 0 60 g Ag/kg Dry 8«d
-0-12gAgftaDiy8«l
-*-31gAgfligOiy8
-------
Silver toxicity in freshwater sediments
Environ. Toxicol. Chem. 18, 1999 35
Table 4. Mean overlying water (OW) and pore-water (PW) pH and dissolved metal concentrations and survival, organism dry weight, and total
biomass of Chironomus tentans larvae in two toxicity tests with Bond Lake sediment amended with silver (Ag) as AgNO,
Nominal
dry
sediment
concn.
(g Ag/kg)
Dissolved metal (mg/L)1
OW/PW
pH
Ag
Zinc
Iron
Mean (SD)
% survival
Mean (SD)
organism dry
weight (mg)
Mean (SD)
replicate total
biomass (mg)
Test 1
Control
0.005
0.01
0.03
0.08
0.20
0.50
OW
PW
OW
PW
OW
PW
OW
PW
OW
PW
OW
PW
OW
PW
7.65
6.82
7.62
6.82
7.62
6.82
7.51
6.84
7.68
6.91
7.52
6.90
7.50
6.21
0.0008
0.0002
0.0002
0.0006
0.0002
0.0010
0.0010
0.0027
0.0061
0.0060
0.016
0.034
0.013
0.127
0017
<0.005
<0.005
<0.005
<0.005
0.009
<0.005
<0.005
<0005
<0.005
<0.005
<0.005
0.012
0007
1.012
17.3
<0.065
13.4
<0.065
14.2
<0.065
13.9
<0.065
9.04
<0.065
5.16
<0.065
1.20
93.3
100
100
100
83.3
86.7
83.3
(11.5)
(0.0)
(0.0)
(0.0)
(153)
(5.77)
(5.77)
2.66
2.29
1.99
1.92
2.53
1.98
1.85
(0.40)
(0.26)
(0.11)
(0.44)
(0.56)
(0.16)
(0.26)b
22.3
20.4
17.9
16.4
18.4
17.2
15.3
(6.95)
(1.33)
(1-21)
(2.89)
(2.53)
(2.46)
(1.19Y
Test 2
Control
0.03
0.08
0.20
0.50
1.2
3.1
OW
PW
OW
PW
OW
PW'
OW
PW
OW
PW
OW
PW
OW
PW
7.56
6.89
7.50
6.95
7.50
6.89
7.49
, 6.90
7.66
6.32
7.46
4.41
6.89
3.57
<0.005
<0.005
<0.005
<0.005
<0.005
0.033
0.009
0.019
0.017
1.044
0.156
7.979
34.61
1,389
<0.005
<0.005
<0.005
0.006
0.019
<0.005
<0.005
<0.005
0.077
0.025
<0.005
0.311
0.031
1.49
<0.065
12.1
0.113
9.64
0.275
8.74
0.167
6.42
0.065
4.71
<0.065
1.15
<0.065
0.634
96.7
100
93.3
93.3
(5.77)
(0.00)
(5.77)
(5.77)
90.0 (10.0)
53.3
0.0
(11.5)"
(0.0)"
2.96
3.03
2.98
2.56
2.29
(0.16)
(0.16)
(0.11)
(0.55)
(0.10)'
0.36 (0.037)"
-
_d
27.6
29.2
25.8
22.2
20.5
1.89
(3.97)
(1.38)
(2.17)
(2.12)
(1.51)"
(0.3 iiy
_d
• Dissolved is procedurally defined as having passed through a 0.2-|un pore size membrane filter.
b Significantly different from survival or dry weight of control animals in the same test.
c Significantly reduced from total biomass of control animals with test 1 and 2 data pooled (p & 0.05).
c No data as there Were no survivors.
Silver toxicity in sediments
West Bearskin Lake. During the toxicity tests with C. ten-
tans, no measurable Ag was present in the PW in the two
lowest exposures of 1.7 and 2.2 g Ag per kilogram dry sed-
iment (Table 3). At the 1.7-g Ag per kilogram exposure, PW
concentrations of Zn, Ni, Cu, and Pb were either at or below
analytical detection limits. At 2.2 g Ag per kilogram dry sed-
iment, PW concentrations of Zn, Ni, and Pb were detectable,
but Cu remained below detection. In fact, Cu was measurable
only at the two highest exposures. Dissolved Ag concentrations
behaved erratically in the PW at exposures of 3.0 g Ag per
kilogram dry sediment and higher. A reddish orange-colored
precipitate, presumably oxides of iron, was visible at spiking
levels of 4.0 g Ag per kilogram dry sediment and higher.
Amorphous iron oxides formed at the higher AgNO3 spiking
levels may have scavenged the freely dissolved Ag to varying
degrees, thereby affecting dissolved Ag concentrations.
The biological responses of C. tentans showed that the
greatest reduction in survival occurred at the 3.0 g Ag per
kilogram sediment spiking level, which also had the highest
mean measured PW dissolved Ag concentration of 0.299 mg/
L. Mean survival was improved at the higher exposures but
remained below 50% in all treatments higher than 3.0 g Ag
per kilogram dry sediment. Mean organism dry weights were
most greatly reduced between the exposure levels of 2.2 and
5.4 g Ag per kilogram dry sediment; however, only the 4.0-
g/kg exposure resulted in a significant (p = 0.05) reduction
(Table 3). The total replicate dry weight biomasses of surviving
larvae were significantly (p £ 0.05) decreased at all Ag amend-
ment levels except for the lowest. This endpoint integrates
both survival and individual organism dry weight. The PW-
dissolved Ag concentrations were below detection (<0.005
mg/L) at the two lowest spiking levels, so it would appear that
Ag was not directly responsible for the observed biomass re-
duction at the 2.2 g Ag per kilogram sediment amendment
level. The NOEC-LOEC range based upon total biomass and
nominal bulk sediment additions of Ag was 1.7 to 2.2 g Ag
per kilogram dry sediment. An LC50 value of 2.75 g Ag per
kilogram dry sediment was estimated based on nominal bulk
sediment additions. Due to the erratic dissolved Ag concen-
trations in the PW, it was not possible to estimate a PW LC50
value for dissolved Ag.
-------
36 Environ Toxicol Chem 18, 1999
D.J. Call et al
Table 5 Summary of biological responses of Chironomus tentans larvae exposed to silver (Ag) in water
and sediments
Biological responses
Test medium
Laboratory water
West Bearskin
Lake sediment
Bond Lake
sediment
Matrix1
PW
S
PW
S
10-d LC50"
0.057
— c
2.75
15.1
1.17
NOEC"
0.013
d
1.7
0.034
0.20
LOEC"
0.066
d
22
0127
0.50
Most sensitive
endpomt
Dry weight
Total biomass
Dry weight and to-
tal biomass
• PW = pore water; S = sediment.
b LC50 = concentration causing 50% lethality; NOEC = no observable effect concentration; LOEC =
lowest observable effect concentration. Concentrations are for dissolved silver in milligrams per liter
and nominal additions of Ag to bulk sediment in grams Ag per kilogram dry sediment.
G Not calculable due to nature of dose-response data.
" Not determined as the PW Ag concentrations at both the NOEC and LOEC were below detection
(<0.005 mg/L)
The displacement of other metals likely contributed to ob-
served mortalities at higher AgNO3 amendment levels. Con-
centrations of Zn, Ni, and Pb increased in the PW as AgNO3
spiking level increased. Due to the displacement of Zn, Ni,
Cu, and Pb in the sediment by Ag*, it is likely that Ag+ in
combination with these metals, as well as decreasing pH, con-
tributed to observed reductions in survival at spiking levels
of S3.0 g Ag per kilogram dry sediment. The 10-d LC50 value
for Zn with C. tentans is 1.12 mg/L [28]. Therefore, concen-
trations of Zn alone would have contributed 0.021, 0.124,
0.390,1.10, and 3.52 acute toxicity toxic units (TUs) at spiking
levels of 3.0, 4.0, 5.4, 7.2, and 9.6 g Ag per kilogram dry
sediment, respectively. Ten-day LC50 values for C. tentans
were not available for the other metals that were measured.
Nickel and Pb would also have contributed some fractional
TUs, whereas Cu would have contributed'fractional TUs only
at the two highest spiking levels. The observed displacement
of Zn and Cu followed the order to be expected in terms of
their respective metal sulfide solubility products [29]. This was
also previously observed in laboratory sediments spiked with
several divalent metals [15]. Acid-volatile sulfide concentra-
tions were very low, starting at approximately 1.1 u,mol/g
before the test started and decreasing to <0.1 |unol/g during
the toxicity test in the control, low, medium, and high spiking
levels. Therefore, it would appear that the dissolution of these
metals from their associations with other ligands may follow
the same general pattern as with sulfide.
Bond Lake. In two tests with Bond Lake sediment, no ap-
preciable dissolved Ag appeared in the PW until a spiking
level of 0.08 g Ag per kilogram dry sediment was reached.
Zinc concentrations in PW first appeared at 0.50 g Ag per
kilogram sediment and increased with increasing Ag amend-
ments up to a high mean concentration of 1.49 mg/L of Zn at
the highest exposure. Iron concentrations in PW remained be-
tween 12.1 and 17.3 mg/L in the controls and the lower spiked
sediments but decreased noticeably at an amendment level of
0.08 g Ag per kilogram dry sediment in test 1 and 0.50 g Ag
per kilogram dry sediment in test 2. Pore-water Pb concen-
trations in test 2 started to increase at the 0.20-g Ag per ki-
logram spiking level, attaining a high concentration of 0.126
mg/L at the highest Ag amendment level (3.1 g Ag per ki-
logram dry sediment). Copper concentrations were elevated to
0.046 and 0.067 mg/L in the PW at the two highest exposures
in test 2.
Chironomus tentans larvae were unaffected in their sur-
vival at mean dissolved PW concentrations up through 1.044
mg Ag per liter in the two Bond Lake sediment tests, corre-
sponding to a nominal sediment amendment level of 0.50 g
Ag per kilogram dry sediment (Table 4). Survival was sig-
nificantly reduced at the 1.2 g Ag per kilogram dry sediment
amendment level. Mean organism dry weight was reduced only
at the exposures of &0.50 g Ag per kilogram dry sediment.
The mean total biomass of survivors was significantly less (p
£ 0.05) than controls, with a decrease of 28.3% for the two
tests combined at the 0.50-g Ag per kilogram dry sediment
level. Total biomass was significantly reduced to a very low
level at the exposure of 1.2 g Ag per kilogram dry sediment,
due to the combined significant reductions in both survival
and organism dry weight. A sharp change occurred between
the amendment levels of 0.20 and 0.50 g Ag per kilogram dry
sediment in the capacity of the sediment to bind Ag+. Survival
was similar at these two exposures, but dry weight and biomass
were reduced, indicating that these were more sensitive end-
points than survival. The NOEC-LOEC range based upon dry
weight and total biomass was 0.20 to 0.50 g Ag per kilogram
dry sediment and 0.034 to 0.127 mg/L of dissolved Ag in the
PW. At a mean PW concentration of 7.979 mg dissolved Ag
per liter, 46.7% of the larvae were dead at day 10, and biomass
of survivors was significantly reduced to only 6.8% of the
controls. Mortality was 100% at the highest exposure of 3.1
g Ag per kilogram dry sediment. As in West Bearskin Lake
sediment at high Ag amendment levels, elevated levels of other
metals and reduced pH also likely contributed to the observed
mortalities at these high Ag exposures. Values of LC50 for
PW and sediment were estimated at 15.1 mg dissolved Ag per
liter and 1.2 g Ag per kilogram dry sediment, respectively
(Table 5). Compared with the 10-d water-only LC50 value of
0.057 mg/L for dissolved Ag, this PW LC50 value is 275 times
greater. It would appear that a very high percentage, perhaps
more than 99%, of the Ag that passed through the 0.2-jim
membrane filter was not readily bioavailable to result in acute
lethality as in the water-only toxicity test. However, the PW
LOEC based upon total biomass is only about twice as high
as the PW LC50 value. This may indicate that the process by
which Ag becomes bioavailable to produce mortalities is slow
and that Ag more readily affects growth than survival.
Comparison with other tests with freshwater benthos. This
study used the nitrate salt, which has been shown to be the
-------
Silver toxicity in freshwater sediments
most toxic salt of Ag [24,26,30]. Studies with other salts of
Ag have resulted in either no toxicity or reduced toxicity to-
ward freshwater benthic organisms. Amendment of sediments
with the highly insoluble silver sulfide (Ag2S) at concentrations
as high as 0.753 g/kg dry sediment did not significantly affect
survival of the amphipod Hyalella azteca [31]. Hyalella az-
teca was found to be the most highly sensitive of nine species
-to Ag in water-only acute toxicity tests [32]. This suggests
that either none or only minute amounts of Ag* were released
from Ag2S-spiked sediments. A bioaccumulation study with
the freshwater oligochaete Lumbriculus variegatus and AgzS-
amended sediment at approximately 0.444 g Ag per kilogram
dry sediment also demonstrated a lack of toxicity and a low
accumulation factor of 0.18 [33], further indicating the relative
unavailability of Ag* in PW of sediments amended with the
Ag2S salt. Ten-day LC50 values for H. azteca were greater
than the highest concentrations of sediment amendment levels
tested when Ag was introduced as AgCl (>2.56 g/kg) or Ag2
(S203)n (>0.569-1.12 g/kg) [26].
It is not believed that NH3 contributed highly to the ob-
served effects. Chironomus tentans can tolerate considerable
NH3. with a 10-d LC50 value in the range of 530 to 700 mg/
L of total ammonia [34]. The highest single measured NH3
concentration in this study of 100 mg/L occurred in a control
sediment. Ammonia concentrations did not correlate with Ag
concentrations. However, pH may have added to the observed
adverse responses in some cases. The genus Chironomus can
tolerate reduced pH, as it has been found to occur in nature
in waters with pH as low as 4.4 [35]. Laboratory studies have
also demonstrated that other members of the Chironomus ge-
nus can survive acute exposure to pH 4 [36,37]; however,
survival was reduced nearly 50%. Exposure of animals to a
sudden large difference in pH, as would have occurred at the
higher exposures of this study, may have contributed to the
observed toxicities at these higher exposures.
The endpoint of mean organism dry weight did not show
a consistent decline with increasing Ag concentration in either
of the sediments in this study. However, in the West Bearskin
Lake exposures of 3.0, 4.0, and 5.4 g Ag per kilogram dry
sediment, the mean organism dry weights were reduced to
0.65, 0.76, and 0.61 mg, respectively. Previous studies have
demonstrated correlations between reduced growth of C. ten-
tans larvae and reduced success of both emergence to adults
and reproduction [38,39]. These studies also indicated that a
dry weight of about 0.5 to 0.6 mg was the minimal weight for
successful emergence and reproduction. Future studies of Ag
toxicity to C. tentans that extend over the full life-cycle will
be useful in interpreting .the effects of reduced weight upon
reproductive success of this species.
Freshwater sediment binding phases for silver
Both sediments selected for study had low AVS and TOG
levels to minimize the binding capacity for Ag. Acid-volatile
sulfide and TOC are considered to be two major metal-binding
phases in anoxic sediments [16,18,40,41]. With the low AVS
levels during the toxicity tests (<0.10 jtmol/g), the presence
of approximately 0.020 mg/L of Ag* in the PW (0.20 fimol/
ml) would have yielded a SEM[Ag/2] - AVS difference of
&0.0. This could have potentially resulted in some toxicity,
according to equilibrium partitioning theory [42]. Due to the
low AVS content of our test sediments, SEM[Ag/2] - AVS
was >0.0 in all cases where significant toxicity was observed
However, mortalities were less than 100% in most cases, and
Environ. Toxicol Chem. 18, 1999 37
toxicity was not attributable to Ag* alone because significant
quantities of displaced Zn (plus Ni in West Bearskin Lake)
were also present when sufficient AgNO3 was incorporated
into the sediment to yield appreciable quantities of dissolved
Ag in the PW.
This study has shown that two sediments, both with rela-
tively low AVS and TOC levels, varied considerably in their
capacities to bind Ag. Relatively high spiking levels of AgNO3
resulted in the displacement of other metals from the sediment
into the PW, with Zn being displaced in greatest concentration
and Cu the least. The capacity for these sediments to either
bind Ag* or to react with Ag* to form a nontoxic dissolved
species appears to be largely dependent upon certain physi-
cochemical characteristics of these oxic test sediments in ad-
dition to AVS and TOC. Acid-volatile sulfide and TOC com-
bined likely could not have bound all of the Ag* at several of
the exposures in this study. For example, at the West Bearskin
Lake amendment level of 2.2 g Ag per kilogram dry sediment,
a total Ag concentration of approximately 20.4 |unol/g of dry
sediment was added. One mole of AVS can bind 2 mol of
Ag*, and the partition coefficient (K^ for organic matter and
Ag is approximately 1.82 X 10s L/kg^ [39]. This means that
West Bearskin Lake sediment with an AVS level of <0.1 \imoU
g could bind a maximum of 0.2 junol of Ag by AVS. From
the work of Mahony et al. [40,41], it was estimated that ap-
proximately 10 to 20 u,mol of Ag* per gram dry sediment
could be bound by TOC hi West Bearskin Lake sediment con-
taining 1% TOC, dependent upon pH [42]. However, if fresh-
water sediments are anoxic and have AVS levels in excess of
one half the molar concentration of Ag, then binding by AVS
should apply in rendering Ag* unavailable to biota, as in the
cases of the divalent metals [18]. The binding of Ag by the
sediments of this study may have been somewhat reduced
relative to that of in-place bedded sediments from the field.
Manipulations involved in performing the tests likely reduced
the AVS of the test sediments. Also, cold storage has been
shown to result in a decrease in TOC content during the first
6 months of storage [43].
The observation of high rates of survival of C. tentans at
high dissolved concentrations of Ag within the peeper cham-
bers (e.g., up to ~300 times the water-only LC50 value) in-
dicates that most of the dissolved Ag was not freely dissolved
ionic Ag* and was not in a form that was readily available to
cause acute lethality. Possibilities for association with Ag*
include complexes of Fe or manganese oxides or colloids of
organic matter <0.2 |un in size. Amorphous Fe oxides are
known to be a major sink for metals in oxic sediments [27,44]
and were likely important ligands for these particular sedi-
ments under the oxic conditions of the laboratory tests. Binding
site densities may be in the range of 1 to 15-u.m sites per gram
sediment for amorphous iron oxides [45]. As more knowledge
is gained on binding of metals by Fe oxides in freshwater
environments, it may be possible to develop more accurate
predictions of metal bioavailability, as suggested by Tessier et
al. [46]. Ambient sediment quality criteria might be developed
based upon concentrations of AVS plus organic carbon for
anoxic waters and organic carbon plus amorphous Fe oxides.
for oxygenated waters. Based upon the recent work of Berry
et al. [18], it appears that AVS can be used to accurately predict
when sediments containing Ag plus the divalent metals will
not be toxic. A similar approach may be desirable to account
for the additional binding phases.
-------
38 Environ. Toxicol Chem 18, 1999
D.J. Call et al.
CONCLUSIONS
In summary. West Bearskin Lake sediment had a greater
capacity for binding Ag than Bond Lake sediment, likely due
in part to its greater proportion of fine-grained particles and
higher TOC and Fe concentrations. The AVS concentration
was also initially higher in West Bearskin Lake sediment, but
both sediments had levels below detection during the perfor-
mance of sediment toxicity tests. The 10-d LC50 values for
C. tentans, based upon nominal additions of Ag to the sedi-
ments, were 2.75 and 1.17 g Ag per kilogram dry sediment
for West Bearskin and Bond Lake sediments, respectively (Ta-
ble 5). Although a PW LC50 value was not calculable for
sediment from West Bearskin Lake amended with Ag, a PW
LC50 value of 15.1 mg/L for dissolved Ag was determined
for Bond Lake sediment. This was 275 times greater than,the
10-d water-only LC50 value of 0.057 mg/L, indicating that
most of the dissolved fraction was not readily bioavailable to
cause lethality. Ranges of NOEC-LOEC for West Bearskin
and Bond Lake sediments were 1.7 to 2.2 and 0.20 to 0.50 g
Ag per kilogram dry sediment, respectively, based upon total
biomass per replicate or mean organism dry weight. The ad-
dition of Ag to both sediments resulted in the displacement of
other metals (i.e., Zn, Ni, Cu, and Pb) from the sediment to
the PW and reduced the PW pH values. These changes likely
contributed to the observed toxicity, particularly at the higher
Ag amendment levels. The concentrations of Ag in the sedi-
ments that caused significant adverse effects in C. tentans were
well above concentrations reported in the environment.
Acknowledgement—The authors thank Gary Ankley, Dave Mount,
Russ Brickson, Walter Berry, Dominic Di Toro and Tom Purcell for
their thoughtful input in the planning of this study. We are grateful
to Brenda Maldonado, Heidi Saillard, lody Spnngsteele, Kathryn
Roche, and Matthew TenEyck, for their work as student research
assistants, and to our secretary, Joyce Barnes, for manuscript prep-
aration. Support for this study was provided by a grant from the
Photographic and Imaging Manufacturers Association.
REFERENCES
1. ScowK, etal. 1981. Exposure risk assessment for silver. Technical
Report. US Environmental Protection Agency, Washington, DC.
2. U.S. Department of Health and Human Services. 1990. Toxico-
logical profile for silver. TP-90-24. Atlanta, GA.
3. Cooper CF, Jolly WC. 1970. Ecological effects of silver iodide
and other weather modification agents' A review. Water Resour
Res 6:88-98.
4. Terhaar CJ, Ewell WS, Dziuba SP, Fassett DW. 1972. Toxicity of
photographic processing chemicals to fish. Photogr Sci Eng 16:
370-377.
5. Terhaar CJ, Ewell WS, Dziuba SP, White WW, Murphy PJ. 1977.
A laboratory model for evaluating the behavior of heavy metals
in an aquatic environment. Water Res 11:101-110.
6. Nebeker AV 1982. Evaluation of a Daphma magna renewal life
cycle test method with silver and endosulfon. Water Res 16:739-
744.
7. Nebeker AV, McAuhffe CK, Mshar R, Stevens DG. 1983. Tox-
icity of silver to steelhead and rainbow trout, fathead minnow
and Daphnia magna Environ Toxicol Chem 2 95-104.
8. Hogstrand C, Wood CM 1998 Toward a better understanding of
the bioavailabihty, physiology and toxicity of silver in fish: Im-
plications for water quality criteria. Environ Toxicol Chem 17:
547-561.
9. Eisler R. 1996 Silver hazards to fish, wildlife and invertebrates-
A synoptic review. Contaminant Hazard Reviews Report 32. U.S.
Department of the Interior, Patuxent Wildlife Research Center,
Laurel, MD
10 Shafer MM 1995 Sampling and analytical techniques for silver
in natural waters. Proceedings, 3rd International Conference,
Transport, Fate and Effects of Silver in the Environment, Wash-
ington, DC, USA, August 6-9, pp 99-108.
11 Andren AW, Wingert-Runge B, Sedlak D. 1995 The occurrence
and behavior of silver in natural waters. Andren A, Bober T, eds,
Proceedings, 3rd International Conference, Transport, Fate and
Effects of Silver in the Environment, Washington, DC, USA,
August 6-9, pp 7-8.
12 Purcell TW, Peters JJ. 1998. Sources of silver in the environment.
Environ Toxicol Chem .17.539-546.
13. Shafer MM, Overdier JT, Armstrong DE. 1998. Removal, parti-
tioning, and fate of silver and other metals in wastewater treatment
plants and effluent-receiving streams. Environ Toxicol Chem 17-
630-641.
14. Van Cappellen P, Wang Y. 1995. Metal cycling in surface sedi-
ments: modeling the interplay of transport and reaction. In Allen
HE, ed, Metal Contaminated Aquatic Sediments. Ann Arbor
Press, Ann Arbor, MI, USA, pp 2-64.
15. Berry WJ, Hansen DJ, Mahony JD, Robson DL, Shipley PB,
Rogers B, Corbin JM. 1996. Predicting the toxicity of metals-
contaminated sediments using acid volatile sulfide and interstitial
water normalization in laboratory-spiked sediments. Environ Tox-
icol Chem 15:2067-2079.
16. Ankley GT, Di Toro DM, Hansen DJ, Berry WJ. 1996 Technical
basis and proposal for deriving sediment quality criteria for met-
als. Environ Toxicol Chem 15.2056-2066.
17. Hansen DJ, Berry WJ, Mahony JD, Boothman WS, Robson DL,
Ankley GT. 1996. Predicting the toxicity of metals-contaminated
field sediments using interstitial concentration of metal and acid
volatile sulfide normalizations. Environ Toxicol Chem 15:2080-
2094.
18. Berry WJ, Cantwell MG, Edwards PA, Serbst JR, Hansen DJ.
1999. Predicting toxicity of sediments spiked with silver. Environ
Toxicol Chem 18:40-48.
19. Benoit DA, Phipps GL, Ankley GT. 1993. A, sediment testing
intermittent renewal system for automated renewal of overlying
water in toxicity tests with contaminated sediments. Water Res
16:457-464.
20. U.S. Environmental Protection Agency. 1994. Methods for mea-
suring the toxicity and bioaccumulation of sediment-associated
contaminants with freshwater invertebrates. EPA/600/R-94/024.
Duluth, MN.
21. U.S. Environmental Protection Agency. 1979. Method 415.1. Or-
ganic carbon, total (combustion or oxidation). EPA-600/4-79-020.
Cincinnati, OH.
22. Allen HE, Gongmin F, Boothman W, Di Toro D, Mahony JD.
1991. Determination of acid volatile sulfides and simultaneously
extractable metals in sediment. In Draft Analytical Method for
the Determination of Acid Volatile Sulfide in Sediment. U.S.
Environmental Protection Agency, Washington, DC.
23. Lazorchak JM. 1994. Computer program containing latest ver-
sions of Dunnett's procedure, probit analysis, the trimmed Spear-
man-Karber method, and the linear interpolation method. U.S.
Environmental Protection Agency, Cincinnati, OH
24. Rodgers JH Jr. Deaver E, Suedel BC, Rogers PL. 1997. Com-
parative aqueous toxicity of silver compounds: Laboratory studies
with freshwater species Bull Environ Contam Toxicol 58:851-
858.
25. Khangarot BS, Ray PK. 1989. Sensitivity of nudge larvae of
Chironomus tentans Fabricius (Diptera Chironomidae) to heavy
metals. Bull Environ Contam Toxicol 42:325-330.
26. Rodgers JH Jr, Deaver E, Rogers PL 1997. Partitioning and ef-
fects of silver in amended freshwater sediments. Ecotoxicol En-
viron Saf 37:1-9.
27. Newman MC, Jagoe CH 1994. Ligands and the bioavailabihty
of metals in aquatic environments. In Hamehnk JL, Landrum PF,
Bergman HV, Benson WH, eds, Bioavailabihty Physical, Chem-
ical and Biological Interactions. Lewis, Boca Raton, PL, USA,
pp 39-61.
28. Phipps GL, Mattson VR, Ankley GT. 1995. Relative sensitivity
of three freshwater benthic macroinvertebrates to ten contami-
nants. Arch Environ Contam Toxicol 28.281-286
29. Dean JA, ed. 1985. Lange's Handbook of Chemistry, 3rd ed
McGraw-Hill, New York, NY, USA.
30. Hogstrand C, Galvez F, Wood CM. 1996. Toxicity, silver accu-
mulation and metallothione in induction in freshwater rainbow
trout dunng exposure to different silver salts. Environ Toxicol
Chem 15.1102-1108
31 Hirsch MP. 1998 Toxicity of silver sulfide-spiked sediments to
-------
Silver toxicity in freshwater sediments
the freshwater amphipod Hyalella atteca. Environ Toxicol Chem
17:601-604.
32. Diamond JM, Mackler DG, Collins M, Gruber D. 1990. Den-
vation of a freshwater silver criteria for the New River, Virginia,
using representative species. Environ Toxicol Chem 9:1425-
1434.
33. Hirsch MR 1998b. Bioaccumulation of silver from laboratory-
spiked sediments in theohgochaete Lumbriculus variegatus. En-
viron Toxicol Chem 17:605-609
34. Whiteman FW, Ankley GT, Kahl MD, Rau DM, Balcer MD. 1996.
Evaluation of interstitial water as a route of exposure for ammonia
in sediment tests with benthic macroinvertebrates. Environ Tox-
icol Chem 15:794-801.
35. Krantzberg G, Stokes PM. 1988. The importance of surface ad-
sorption and pH in metal accumulation by chironomids. Environ
Toxicol Chem 7:653-670.
36. Fisher SW, Lohner TW. 1986. Studies on the environmental fate
of carbaryl as a function of pH. Arch Environ Contain Toxicol
15:661-667.
37. Rousch JM, Simmons TW, Kerens BL, Smith BP. 1997. Relative
acute effects of low pH and high iron on the hatching and survival
of the water mite (Arrenurus manubriator) and the aquatic insect
(Chironomus riparius). Environ Toxicol Chem 16:2144-2150.
38. Liber K, Call DJ, Dawson TD, Whiteman FW, Dillon TM. 1996.
Effects of Chironomus tentans larval growth retardation on adult
emergence and ovipositing success: Implications for interpreting
freshwater sediment bioassays. Hydrobiologia 323:155-167.
39. Sibley PK, Benoit DA, Ankley GT. 1997. The significance of
growth in Chironomus tentans sediment toxicity tests: Relation-
Environ. Toxicol. Chem. 18, 1999 39
ship to reproduction and demographic endpomts Environ Toxicol
Chem 16:336-345.
40 Mahony JD, Di Toro DM, Gonzalez AM, Curto M, Dilg M,
DeRosa LD, Sparrow LA. 1996. Partitioning of metals to sedi-
ment organic carbon. Environ Toxicol Chem 15-2187-2197.
41. Mahony JD, Di Toro DM, McLeod PB, Page JJ, Santos LN,
Trivedi DK. 1999. The role of AVS and organic carbon in par-
titioning of silver to sediments. Proceedings. 5th International
Conference, Transport, Fate and Effects of Silver in the Envi-
ronment, Hamilton, ON, Canada, September 28-October 1,1998,
(in press).
42 Di Toro DM, et al. 1991. Technical basis for establishing sediment
quality criteria for non-ionic organic chemicals using equilibrium
partitioning Environ Toxicol Chem 10:1541-1583.
43. Sijm RTHM, Haller M, Schrap SM. 1977. Influence of storage
on sediment characteristics and of drying sediment on sorption
coefficients of organic contaminants. Bull Environ Contain Tox-
icol 58.961-968.
44. Tessier A, Couillard Y, Campbell PGC, Auclair JC. 1993. Mod-
eling Cd partitioning in oxic lake sediments and Cd concentrations
in the freshwater bivalve Anadonta grandis. Limnol Oceanogr
38:1-17.
45. Gagnon C, Fisher NS. 1997. The bioavailability of sediment-
bound Cd, Co and Ag to the mussel Mytilus edulis. Can J Fish
Aquat Sci 54:147-156.
46. Tessier A, Campbell PGC, Auclair JC, Bisson M. 1984. Rela-
tionship between the partitioning of trace metals in sediments and
their accumulation in the tissues of the freshwater mollusc Elliptio
complanata in a mining area. Can J Fish Aquat Sci 41:1463-
1472.
-------
Environmental Toxicology and Chemistry, Vol 18, No 1, pp 40-48, 1999
Printed in the USA
0730-7268/99 $9 00 + 00
Annual Review
PREDICTING TOXICITY OF SEDIMENTS SPIKED WITH SILVER
WALTER J. BERRY,* MARK G. CANTWELL, PHILIP A. EDWARDS, JONATHAN R. SERBST, and DAVID J. HANSEN
U.S Environmental Protection Agency, National Health and Environmental Effects Research Laboratory,
Atlantic Ecology Division, 27 Tarzwell Drive, Narragansett, Rhode Island 02882
(Received 18 May 199S; Accepted 2A July 1998)
Abstract—Previous experiments conducted with freshwater sediments spiked with silver have shown that, when expressed on a
dry weight basis, the toxicity of silver is sediment-specific and dependent on the form of silver added (e.g., AgNO3, Ag2S). This
study was conducted to assess the usefulness of silver interstitial water toxic units (IWTU) and acid volatile sulfide (AVS)
concentrations in predicting the biological effects of silver species across sediments, regardless of the species of silver present.
Two saltwater sediments were spiked with a senes of concentrations of silver. The amphipod, Ampelisca abdita, was then exposed
to the sediments in ten-day toxicity tests. Amphipod mortality was sediment-specific when expressed on a dry weight basis, but
not when based on IWTU or simultaneously extracted metal (SEM)-AVS. Sediments with an excess of AVS relative to SEM had
IWTU O.S, but no
measurable AVS, and were generally toxic. Sediments with measurable AVS were not toxic. Reanalysis of the previously published
data from the freshwater sediments spiked with silver showed mortality to be correlated with nominal SEM-AVS and with silver
IWTU. Taken together, these results support the use of AVS and silver IWTUs in predicting the toxicity of silver in sediments.
Keywords—Acid volatile sulfide Interstitial water Silver Sediments Toxicity
INTRODUCTION
Silver is a highly toxic metal, with a wide distribution in
the environment. Sediments are a dominant sink for silver
released into aquatic environments [1]. Given that silver is one
of the most strongly bioaccumulated of the elements [2], this
combination of factors would seem to indicate that silver in
sediments poses a large environmental risk. However, little is
known about the biological effects or bioavailability of sedi-
ment-associated silver in the field [2]. This is because silver
is generally found in low concentrations as one component of
a mixture of contaminants and because speciation is extremely
important in silver toxicity [1].
In the absence of compelling field data, one of the best
ways to gain insight into the toxicity and bioavailability of a
metal in sediment is to perform spiked-sediment toxicity tests
[3]. Relatively few studies have been performed with fresh-
water sediments spiked with silver, and none have been per-
formed with marine sediments. Hirsch [4,5] mixed silver sul-
fide powder into freshwater sediments and found that even at
very high concentrations of silver there was no evidence of
toxicity to the amphipod Hyalella azteca [4] or of silver uptake
in the oligochaete Lumbriculus variegatus [5]. Call et al. [6]
found increased mortality to the midge Chironomus tentans
with increasing silver nitrate addition. Rodgers et al. [7], in
experiments with four freshwater sediments, found that the
form of silver influenced the toxicity of silver to H. azteca.
Sediments spiked with silver thiosulfate or silver chloride were
not toxic. Those spiked with silver nitrate were toxic. Although
toxicity increased with increasing silver concentration in both
studies of sediments spiked with silver nitrate, the toxicity of
silver was sediment specific when expressed on a dry weight
basis [6,7].
* To whom correspondence may be addressed
(berry.walter@epa.gov) Contribution 1988 National Health and En-
vironmental Effects Research Laboratory, Narragansett, RI, USA.
Numerous studies have highlighted the utility of interstitial
water (IW) metals concentrations and metalsracid-volatile sul-
fide (AVS) relationships in explaining the bioavailability of
sediment-associated metals to benthic organisms in both the
laboratory and the field [3,8,9]. In all of these studies, the
toxicity of metals was sediment specific when expressed on a
dry weight basis but not sediment specific when expressed on
an IW or AVS-normalized basis. When the IW contained less
than 0.5 toxic units of metal, or when AVS exceeded metal in
the sediment (when expressed on a molar basis as "simulta-
neously extracted metal," or SEM, the metal extracted in the
AVS procedure), the sediments did not cause acute mortality.
Conversely, when the IW contained more than 0.5 toxic units
of metal, and/or when SEM exceeded AVS in the sediment,
sediments generally did cause acute mortality.
Acid-volatile sulfide has been used primarily to predict the
bioavailability in sediments of divalent cationic metals that
form sulfides: cadmium, copper, lead, nickel, and zinc but
should also be useful for other metals that form sulfides, such
as silver. Importantly, silver is different from the divalent tran-
sition metals because (1) silver is (mostly) monovalent, so
each mole of sulfide binds two moles of silver. For this reason,
the concentration of silver divided by two ([Ag]/2) is compared
with AVS. (2) In saltwater sediments the formation of highly
insoluble silver chloride during spiking complicates spiking
and chemical analysis. (3) Equilibration times after spiking
can be very long, up to 90 d or more. (4) Silver sulfide and
silver chloride are not extracted with the conventional AVS
extraction.
In this study two saltwater sediments were spiked with a
series of concentrations of silver nitrate, such that in the low
treatments AVS exceeded [Ag]/2 and in the higher treatments
[Ag]/2 exceeded AVS. One sediment was coarse grained with
low AVS, and the other was fine grained with higher AVS.
The amphipod Ampelisca abdita was. then exposed to the sed-
-------
Predicting the toxicity of silver-spiked sediments
iments in 10-d toxicity tests to determine the utility of AVS
and IW silver concentrations to predict acute mortality in sil-
ver-spiked sediments across a range of sediment types. The
purpose of this paper is threefold: to present data on a new
study with saltwater sediments spiked with silver, to sum-
marize the available studies on freshwater sediments spiked
with silver, and to combine information from all of these stud-
ies to evaluate the overall utility of AVS and IW silver in
predicting the toxicity and bioavailability of silver in sedi-
ments.
MATERIALS AND METHODS
Organism and sediment collection
Ampelisca abdita were collected following the methods
described in Berry et al. [3] (Ampelisca abdita is an estuarine,
tube-building, infaunal amphipod commonly used in sediment
toxicity testing [10]). Two sediments of differing AVS con-
centration were used: one from Pojac Point (initial AVS =
23.9-26.9 u,mol S/g) and one from Ninigret Pond (initial AVS
= 2.1 (unol S/g). The Pojac Point sediment was collected from
an uncontaminated site in northeastern Narragansett Bay
(41°38'54"N and 71°23'41"W) with a Van Veen grab sampler,
returned to the laboratory, press sieved wet through a 2-mm
mesh stainless steel screen, homogenized, and stored at 4°C.
This sediment had 1.3% total organic carbon (TOC) and 5%
sand and 95% silt/clay. The Ninigret Pond (Charlestown, RI,
USA) sediment was collected with a shovel. The upper 5 to
10 cm of sediment was returned to the laboratory, sieved wet
through a 1.0-mm stainless steel screen, rinsed several times
to remove high-organic fine particles, homogenized, and stored
at 4°C. This sediment had 0.2% TOC and was made up of
100% sand.
Water-only tests
Ten-day static renewal tests were conducted with A. abdita
to determine water-only concentrations causing 50% lethality ,
(LCSOs) for silver in sand-filtered Narragansett Bay seawater
(31 ± 1 ppt, 19 ± 1"C) following the methods described in
Berry et al. [3], except that some water samples from the tests
were filtered before they were analyzed. Unfed amphipods
were exposed to five concentrations of silver nitrate (Fisher*
certified ACS reagent, Fisher Scientific, Fairlawn, NJ) and a
control, with ten amphipods per replicate and two replicates
per concentration.
Spiked sediment tests
Sediment spiking and equilibration. Sediments were spiked
with a solution of silver nitrate salts as described in Berry et
al. [3]. The spiking solution was prepared by dissolving silver
nitrate into 18 Mft deionized water while maintaining the pH
at 8.1 ± 0.2. Each jar was purged with nitrogen, capped, and
rolled for 15 rnin/d at 4°C during the equilibration period.
The slow kinetics of silver sulfide formation necessitated
the sampling of sediments at 10- to 14-day intervals for AVS
concentration during the up to 106-d equilibration period to
ascertain when equilibrium was reached. At each sampling
period the sediment in each jar from selected (or all) treatments
was lightly homogenized, and 30-ml samples of sediment were
removed for AVS analysis. Before capping, the jar was again
purged with nitrogen and returned to the cooler. Jars from the
other treatments were not opened until test initiation
Sediment toxicity test method. Exposure and sampling
methods for sediment, IW, and overlying water followed those
Environ Toxicol Chem. 18, 1999 41
used in Berry et al. [3]. Amphipods were exposed to control
and metal-spiked sediments in 10-d tests with continuous re-
newal of sand-filtered Narragansett Bay seawater (31 ± 1 ppt,
19 ± 1°C). The nominal treatments spanned the range from
sediments that were nominally spiked with less [Ag]/2 than
initial AVS to sediments which had an excess of [Ag]/2 over
AVS (See Table 2 for concentrations). For each sediment there
were six treatments plus a control. Each treatment included
two biological repbcates to assess mortality and two chemical
replicates for interstitial water, metal, and AVS analyses of the
sediment at test initiation and termination. Diffusion samplers
were placed in each biological replicate and one chemical
replicate when the sediments were added to the test chambers.
Twenty amphipods were added to each of these replicates at
the start of the test (the next day). The day 0 chemistry rep-
licates did not get amphipods or diffusion samplers.
One sediment toxicity test was "conducted with sediment
from Ninigret Pond. The sediment toxicity test with Pojac
Point sediment was repeated (tests are referred to as Pojac 1
and PojacS) because there was unexplained toxicity in one of
the intermediate concentrations in the first sediment toxicity
test with this sediment.
Interstitial water silver concentrations were taken from the
diffusion samplers on day 10. Bulk metal, AVS, and SEM
samples were taken on day 10. The bulk metal samples from
Pojacl and 2 were not analyzed.
Chemical analyses
Water analysis. Interstitial water (from diffusion samplers)
and overlying water were analyzed for silver using graphite
furnace atomic absorption spectroscopy (GFAA). Detection
limits for silver by GFAA were 1 ng/L.
Sediment analysis. Sediment samples were analyzed for
AVS by the cold-acid purge and trap technique [11,12] using
nitric acid in place of HC1. (Nitric acid was used in place of
HC1 because of the low solubility of silver chloride. Side-by-
side testing with HC1 showed little difference in the measure-
ment of AVS or SEM if nitric acid or HC1 was used; M.
Cantwell, unpublished data.) Simultaneously extracted metal
and bulk metals analyses were performed using inductively
coupled plasma atomic emission spectrometry (ICP-AES). For
analyses of bulk metals excluding silver, the metals were ex-
tracted from freeze-dried sediments by microwave digestion
with HF/HNOj/HCl acids followed by filtration. Bulk silver
in the sediments was extracted by microwave digestion in 20
ml of HC1 followed by a postdigestion addition of 5 ml of
NH^OH. Total metals analyses of sample blanks and recoveries
of known metal additions demonstrated 85 to 100% recoveries
from sediments, 85 to 115% recoveries from sample extracts,
and an absence of contamination in our analytical procedures.
The SEM concentration reported is the sum of cadmium, cop-
per, lead, nickel, and zinc concentrations and one half of the
silver concentration on a micromole per gram dry sediment
basis. Concentrations of all metals in sediments exceeded an-
alytical detection limits.
Ammonia. Samples were taken to quantify the ammonia
concentrations in the IW and test the hypothesis that ammonia
was the cause of the unexplained mortality in an intermediate
treatment in the experiments using Pojac Point sediments. Dif-
fusion samplers were placed in sediment left over from the
third Pojac Point experiment, 1 month after the experiment
was run. Ammonia was measured using an Orion ammonia
probe (Boston, MA), and pH was measured using a Ross com-
-------
42 Environ. Toxicol Chem. 18, 1999
r*
Table 1. Mortality of Ampehsca abdita exposed to silver nitrate in two 10-d water-only tests*
W.J. Berry et al
Test 1
Test 2
Nominal
silver (ftg/L)
Control
6.5
11
18
30
50
84
140
Nominal
LCSO = 44 u«/L
% Mortality
15
10
20
0
50
80
60
80
Nominal
silver (ji-g/L)
Control
6.5
11
18
30
50
84
140
Nominal
LC50 = 27 tigfL
Mean measured
silver ((ig/L)
Control
3.57
7.05
12.8
22.6
39.9
76.8
128
Measured
LC50 = 20 ng/L
% Mortality
0
35
15
50
45
80
60
55
• LC50 = concentration causing 50% lethality.
bination pH electrode. The ammonia probe was calibrated us-
ing a four-point standard curve. Un-ionized ammonia concen-
trations were estimated using the Hampson model [13].
Data handling
Ten-day LC50 values for the water-only tests were calcu-
lated by the trimmed Spearman-Karber method [14]. Detection
limits were calculated for all chemical analyses based on in-
strument detection limits and sample size. In those instances
where a mean concentration is a summation of measured data
and data below the limit of detection, one half the detection
limit was used for those values below the limit of detection.
Means for which there are no measured values above the de-
tection limit are indicated as ND in the appropriate tables and
graphs.
For illustrative purposes, sediments that caused >24% mor-
tality were classified as toxic. Mearns et al. [15] found that
sediments that caused <24% mortality in tests with the am-
phipod Rhepoxynius abronius were not consistently classified
as toxic. This criterion is similar to the "80% of control sur-
vival" criterion used hi the U.S. Environmental Protection
Agency Environmental Monitoring and Assessment Program
(EMAP) for sediment tests with A. abdita [16]. A horizontal
dashed line at 24% mortality has been included on the appro-
priate figures for reference.
Because silver is essentially monovalent, 2 mol of silver
are required to bind with 1 mol of sulfide. For this reason it
is appropriate to use one half of the silver concentration, rep-
resented as silver/2 or [Ag]/2, in comparisons with AVS con-
centrations. This stoichiometry would suggest that any sedi-
ment with an excess of sulfide relative to one half of the silver
concentration, such that ([Ag]/2) - AVS < 0.0, should not be
acutely toxic because of silver. Therefore, in this study, SEM
is defined as the sum of one half of the concentration of silver
plus the sum of the concentrations of cadmium, copper, lead,
nickel, and zinc in the SEM extract. A dashed vertical line at
SEM - AVS = 0.0 is included on the appropriate figures for
reference.
Many of the IW concentrations in this paper are expressed
as toxic units. A toxic unit is the measured water concentration
divided by the water-only LC50 concentration for that partic-
ular compound and test organism. For example, a sediment
with an IW concentration equal to the water-only LC50 con-
centration for the test organism would have 1.0 IW toxic units
(IWTU). When more than one toxic metal is present, IWTUs
are calculated as the sum of the toxic units of the individual
metals, e.g., IWTU^+z,, = (TW cone AgJ/LCSO^ + (IW cone
ZnVLCSOz,,). Thus, if IW is the principal source of metal tox-
icity, and availability of metals is the same in water-only tests
and IW hi sediment tests, 50% mortality would be expected
with sediments having 1.0 IWTU. A dashed vertical line at
IWTU = 0.5 is included on the figures in this paper to indicate
sediments unlikely to cause significant mortality. This value
was selected because on the average, water-only LCO and
LC50 values differ by approximately a factor of two [17] and
because the data hi our earlier experiments with other metals
[3] support this value as a break point between toxic and
nontoxic sediments. Only silver IWTUs are reported because
the concentrations of other metals in the IW from these ex-
periments were lexicologically insignificant (Berry, unpub-
lished data). Calculation of IWTUs was based solely on de-
tectable metal concentrations.
RESULTS AND DISCUSSION
Water-only tests
Two 10-day, water-only definitive tests were conducted (Ta
ble 1). Chemistry samples from the first definitive were lost.
so the test was repeated. Silver concentrations from the filterec
versus total and prer versus postrenewal samples from the
second test were nearly identical (Berry, unpublished data) sc
only the mean prerenewal, total silver concentrations are re
ported. Based on nominal concentrations, the LCSOs from th<
first and second definitive tests were 44 and 27 u.g/L, respec
lively. The measured 10-day LC50 from the second test of 20
jtg/L was used to calculate silver IWTUs in the sediment tests
with A. abdita.
Sediment tests
Sediment equilibration. Acid-volatile sulfide concentra-
tions were measured to determine the time to equilibration. It
took weeks or months for AVS concentrations in the sediment
to equilibrate after spiking, much longer than required for
cationic metals used in other experiments [3]. This is presum-
ably because the dissolved silver nitrate is rapidly converted
to solid silver chloride when mixed with seawater. This silver
chloride must first dissolve before silver sulfide can form. In
anoxic sediments, almost all of the silver will end up as silver
sulfide, unless the supply of sulfide is exhausted, because the
solubility product constant of silver sulfide is so low, relative
to that of other possible complexes [18]. (See [2] for a dis-
cussion of silver complexation in oxic sediment.) Silver sulfide
-------
Predicting the toxicity of silver-spiked sediments
100i
0.01
106
Fig 1. Acid-volatile sulfide (AVS) as a function of time after spiking
the sediment used in the Pojac 3 experiment. Treatments are labeled
with nominal ([Ag]A2) - AVS concentrations in micromoles per gram
dry weight sediment.
is not soluble in the AVS extraction, so the measured AVS
concentration decreases as the sulfide is formed. The pattern
of change in AVS with time during the Pojac 3 equilibration
(Fig. 1) was typical of all the spikings. The AVS concentration
in the control held near the prespiking level throughout the
equilibration period. In low treatments there was an initial
decrease in AVS, but AVS was relatively constant thereafter.
In intermediate concentrations the AVS started at a reduced
level and then decreased below detectable levels. In the higher
Environ Toxicol Chem. 18, 1999 43
concentrations, AVS was below detection at the first sampling.
Sediment toxicity tests were initiated when AVS was no longer
detectable in the treatments with decreasing AVS concentra-
tion. Sediment equilibration times for Pojac 1 and 3 and Ni-
nigret were 91, 112, and 48 d, respectively. The equilibration
time of the Ninigret sediment may have been shorter because
of its lower initial AVS concentration.
Overlying water. Silver was not detected in the overlying
water in all but the two highest treatments of each of the
sediment experiments (Table 2). Even in those treatments in
which silver was detected in the overlying water, concentra-
tions were generally a factor of five to ten lower than in the
pore water. For this reason, and because there was 100% mor-
tality in the [Ag]/2 - AVS = 52.6 treatment in the Pojac 3
sediment, which did not have detectable silver in the overlying
water (detection limits were well below the 10-d LCSO con-
centration), it seems unlikely that the overlying water had a
controlling effect on mortality.
Simultaneously extracted metal. Silver concentrations in
the SEM solution were very low in all sediment tests, even at
extremely high silver concentrations (Table 2). There was some
evidence of release of other metals into the SEM as a result
of silver addition (Berry, unpublished data), but none of these
releases appeared to be of any lexicological significance.
Amphipod mortality in relation to sediment and pore-wa-
ter silver. Mortality and available chemistry data for the
spiked-sediment experiments are summarized in Table 2. Mor-
talityxoncentration relationships were sediment-specific when
Table 2. Results of 10-d sediment toxicity tests with Ampelisca abdita exposed to silver-spiked sediments from Pojac Point and Ninigret Pond*
Treatment
(nominal
[Ag]/2 - AVS)
(u.mol/g)
Mean Mean Mean Mean
measured measured measured Mean overlying Mean bulk
AVS SEM SEM - AVS pore-water water measured
(umol/g) (M-mol/g) (junol/g) IWTU IWTU (|ig/g)
% Mortality
Pojac 1
Control
-21.5
-16.1
-5.38
16.1
59.2
242
242
27.2
17.1
7.4
ND
ND
ND
ND
—
1.28
0.84
1.24
1.8
2.3
2.4
2
—
-25.9
-16.2
-616
18
2.3
2.4
1.98
—
0.1
0.15
015
0.45
7.45
67.6
81.8
—
ND — "
ND —
ND —
ND * —
ND —
5.67 —
11.7 —
_ _
5
7.5
2.5
71
2.4
93
100
100
Pojac 3
Control
-14.3
-4.78
14.3
52.6
129
454
18
4.9
ND
ND
ND
ND
ND
1.01
0.87
0.83
1.26
1.6
164
29.2
-17
-4.03
0.83
1.26
1.6
164
29.2
0.05
0.27
1.28
4.84
87.6
43.4
347
ND
ND
ND
ND
ND
8.34
7.45
2.6
2,060
3,210
6.550
10,900
23,300
69,800
10
7.5
100
125
100
100
100
Ninigret
Control
-1.68
-126
-0.42
1.26
4.62
18.9
0.7
06
0.9
0.2
ND
ND
ND
0.04
004
0.04
0.04
0.04
012
0.05
-0.66
-0.56
-0.86
-016
0.04
012
005
01
0.15
0.58
052
4.04
61.9
89.6
ND
ND
ND
ND
ND
2.02
6.48
26
82.7
146
302
621
1,320
4,260
15
30
15
27.5
17.5
100
100
• AVS = acid-volatile sulfide; SEM
b Data not available
• simultaneously extracted metal, IWTU = interstitial water toxic unit.
-------
44 Environ Toxicol Chem 18, 1999
W J Berry et al
10
100 1000 10000
Silver (ug/g Dry Wt)
100000
80-
|-
o 40-
S? 20-
0.
..
±Nm
+ Pojacl
XPojac3
« ,*
I •*
I
**T*
-30 -20
20
40
[Ag/2]-AVS
»
n
o
cS
80-
60.
40-
20.
0.
4 Pojacl
XPojac3
X
I ^
I
|
A 4_
tt X f , ***,
001 01 1 10
IWTU (LC50 = 20 ng/L)
100
I1
o
»o
0s
80-
60-
40-
20-
0
— ^_
A '(Tin — ^
+ Pojacl (£)
• x Pojac3 T
I
f • B
A X^ *^^
01 01 1 10 10
AVS (uMoles/g)
Fig 2 Percentage mortality of the amphipod Ampelisca abdita as a
function of dry weight silver (Ag) concentration (A), ([Ag]/2) — acid-
volatile sulfide (AVS, B), interstitial water toxic units (IWTU, C),
and measured AVS (D) in two saltwater sediments spiked with Ag
Nin = Nimgret Pond sediment, Pojac = Pojac Point sediment Sed-
iments below the dashed line at 24% mortality are not considered
toxic 'Vertical dashed lines at simultaneously extracted metal (SEM)
- AVS = 0 (B) and IWTU = 0 5 (C) indicate predicted break points
in toxicity Data points believed to be the result of interstitial water
ammonia are included but highlighted and not connected by lines in
A AVS detection limit is indicated as ND in D
amphipod mortality was plotted against dry weight silver (Fig.
2A) Although mortality increased with added silver in both
Nmigret and Pojac Point sediments, more silver was required
on a dry weight basis in Pojac Point sediment for the sediment
to be toxic In both Nmigret Pond and Pojac Point sediments,
those with an excess of AVS relative to SEM (i e., SEM -
AVS < 0) had IWTU <0 5 and were generally not toxic (Fig
2B and C) Sediments with an excess of SEM relative to AVS
(i.e., SEM — AVS > 0) had measurable silver present but no
measurable AVS. In these sediments silver IWTUs > 0.5 and
were generally toxic (Fig 2B and C) No sediments in which
measurable AVS was present were toxic (Fig 2D).
The results of the sediment toxicity tests with Pojac Point
sediment indicate a source of toxicity other than silver may
be present. This is primarily because the observed toxicity at
the intermediate treatments does not correlate with nominal or
measured silver concentrations We suspected ammonia tox-
icity based on a preliminary test with Pojac sediment (Pojac
2), which suggested that this toxicity was removed if clean
water was run over the sediment for 10 d (Table 2). Purging
with clean water is routinely used in dredging programs when
sediments have pore-water ammonia levels above a specific
concentration [19] Measurement of ammonia in sediment left
over from the Pojac 3 expenment supported this suspicion,
un-ionized ammonia was highest in the treatment with the
anomalous toxicity (Berry, unpublished data). The calculated
concentration of un-ionized ammonia was almost one half of
the LC50 value of 0 82 reported for this species at a similar
pH [20] It is also possible that the presence of silver in the
IW may have exacerbated the toxicity of ammonia in the IW
[21]. The data points showing this anomalous mortality in
Pojac 1 and 3 have been included in the figures that follow
for completeness. However, to better illustrate the patterns of
mortality attributable to silver in the absence of ammonia,
these data points are not connected by lines in Figure 2A and
have been highlighted in Figure 2B to D.
Analysis of data from freshwater silver-spiked sediment
tests. Rodgers et al [7] and Call et al [6] measured AVS con-
centrations before spiking but did not present plots of mortality
in relation to AVS. Rodgers et al [7] (E Deaver, personal
communication) demonstrated that the concentration response
was not sediment specific when normalized to ([Ag]/2) - AVS
(Fig. 3A). Silver nitrate-spiked sediments in which there was
an excess of AVS relative to added silver, i.e., ([Ag]/2) - AVS
<0, were not toxic to H azteca; sediments in which there was
an excess of silver, i.e., ([Ag]/2) — AVS >0, were generally
toxic (Fig. 3A). (For the freshwater studies, for which SEM
is not available, a dashed line at nominal ([Ag]/2) - AVS =
0 is included on the appropriate figures for reference.) Mor
tality was sediment specific when expressed as dry weight (Fig
3B).
As was shown with A. abdita in the saltwater expenments
and H. azteca in the expenments of Rodgers et al. [7], sedi-
ments with an excess of AVS over silver were not toxic to C.
tentans [6] (D. Call, personal communication) However, most
of the sediments with an excess of silver were toxic (Fig. 4A)
Call et al [6] also measured IW metal concentrations and found
very little silver in the IW, even in sediments in which silver
greatly exceeded AVS. Although mortality was generally low
when IWTUs were less than 0.5, and increased when IWTU
were above 0.5, there appears to be an almost log-linear re-
lationship between mortality and IWTU (Fig 4B) This is in
contrast to the pattern seen in the saltwater data (Fig 3C) and
that seen in most other studies [3] For several of the treat-
ments, the contribution of other metals to the combined IWTUs
was significant, with zinc and copper adding up to 4 9 rWTUs
This was presumably because the added silver caused disso-
lution of zinc and copper sulfide in these sediments, releasing
them into the pore water [3,11]. This underscores the impor-
tance of using the sum of the IWTUs of all metals together
-------
Predicting the toxicity of silver-spiked sediments
A
i
i
Environ Toxicol Chem 18, 1999 45
UUi
80
60
40-
20-
3
• Sed1
• Sed3
ASed6
• Sed7
-2
i
i
j
j
!«•
P *JL i ii
-101234
[Ag/2]-AVS (uMoles/g)
01
1 10
Dry wt silver (ug/g)
100
1000
Fig. 3 Percentage mortality of the amphipod Hyalella azteca as a
function of nominal ([Ag]/2) — acid-volatile sulfide (AVS, A) and
dry weight silver (Ag) concentration (B) in four freshwater sediments
spiked with Ag. Sediments below the dashed line at 24% mortality
are not considered toxic A vertical dashed line at simultaneously
extracted metal (SEM) - AVS = 0 (A) indicates the predicted break
point in toxicity.
Mortality of C. tentans from Call et al. [6] (D. Call, personal
communication) is plotted against dry weight concentration in
Figure 4C. Significant mortality was seen at concentrations
greater than approximately 1,000 u.g/g. That the dry weight
data (Fig. 4C) seems to be as predictive of toxicity as AVS
and IWTU normalized data (Fig. 4A and B, respectively) is
due, at least in part, to the fact that only one sediment was
tested by Call et al. [6].
Laboratory results in relation to sediment quality guide-
line concentrations and concentrations found in the field. One
way to put the spiked-sediment mortality data reported in this
paper in context is to compare them with commonly used
empirically derived sediment quality guideline concentrations.
The threshold effect level (TEL) is used as a concentration
below which toxicity would not be expected in a marine field
sediment; the probable effect level (PEL) is used as a con-
centration above which there is a higher probability of adverse
biological effect in a marine field sediment. The available dry
weight results of the saltwater experiments (Fig. 2A) can be
compared, for example, with the TEL for silver (0 733 u,g/g)
and the PEL for silver (1.77 u,g/g) [22]. There were no acute
effects in many of the exposures at concentrations which, on
a dry weight basis, might be predicted to be toxic in the field,
even when the PEL was exceeded by several orders of mag-
nitude. The apparent effects threshold (AET, 4.5 u.g/g [23]) is
a value that, if exceeded, should always be toxic to Hyalella.
The dry weight data from the freshwater experiments (Figs
3B and 4C) can be compared with a draft Hyalella AET for
silver. Compared with the freshwater mortality data in Figures
o
100
80
60
40
20
____ I ___ —
• I ^
-10 0 10 20 30 40 50
[Ag/2]-AVS(nMoles/g)
£
?
0
5?
80
60
40
20
ft
.
I
•
•
B
m-r r -
0.001 0.01 01 .1 10
Interstitial Water Toxic Units (IWTU)
C 100
80-
if
=5 604
40-
20-
0 01 0.1
10 100 1000 10000
Silver (Ug/g Dry Wt)
Fig 4 Percentage mortality of the midge Chironomus tentans as a
function of nominal ([Ag]/2) - acid-volatile sulfide (AVS, A), inter-
stitial water toxicity unit (IWTU, B), and dry weight Ag concentration
(C) in a freshwater sediment spiked with Ag (Call et al., unpublished
data) Sediments below the dashed line at 24% mortality are not con-
sidered toxic Vertical dashed lines at SEM - AVS = 0 (A) and
IWTU = 0.5 (B) indicate predicted break points in toxicity.
3B and 4C, this level appears to be lower than all but one of
the sediments in the study of Rodgers et al [7], which had^
very low AVS concentrations. These data indicate that thesdH
sediment quality guidelines may be useful as screening level
concentrations for the effects of silver in sediments, although
they may be highly overprotective in some sediments. There-
fore, dry weight measures should not be used to indicate that
a toxic effect is due specifically to silver in a sediment in
which they are exceeded If a chemical concentration exceeds
an empirically derived value, it does not necessarily mean the
-------
46 Environ Toxicol Chem 18, 1999
W.J Berry et al
Table 3 Contingency table for predicting toxicity due to metals from measurements of silver (Ag),
other SEM metals, and AVS in laboratory-spiked and field sediments
Nominal metal
and AVS in sediment
Measured metal and
AVS in sediment
Prediction of acute toxicity
Ag only
[Ag]/2 < AVS
[Ag]/2 > AVS
AVS > detection limit
and (SEM - AVS)
< 00
AVS < detection limit
and (SEM - AVS)
>00
Sediment not acutely toxic due to Ag,
no metals detectable in interstitial
water
Sediment may be acutely toxic due to
Ag (but not if IWTU < 0 5)
Metals mixtures
([Ag]/2 + [Cd] + [Cu] + [Ni]
+ [Pb] + [Zn]) < AVS
([Ag]/2 + [Cd] + [Cu] + [Ni]
+ [Pb] + [Zn]) > AVS but
[Ag]/2 < AVS
([Ag]/2 + [Cd] + [Cu] + [Ni]
+ [Pb] + [Zn]) < AVS and
[Ag]/2 > AVS
AVS > detection limit
and (SEM - AVS)
<00
AVS > detection limit
and (SEM - AVS)
> 00
AVS < detection limit
and (SEM - AVS)
>0.0
Sediment not acutely toxic due to these
metals, no metals detectable in inter-
stitial water
Sediment may be acutely toxic due to
these metals but not silver directly
(but not if sum of IWTU < 0 5)
Sediment may be acutely toxic due to
Ag and/or the other metals (but not if
sum of IWTU < 0.5)
• SEM = simultaneously extracted metal, AVS = acid-volatile sulfide, IWTU = interstitial water toxic
unit, Cd = cadmium, Cu = copper, Ni = nickel, Pb = lead, Zn = zinc
chemical caused the effect. Rather, these field-derived values
are concentrations that have been associated with an effect
Another way to put the dry weight toxicity data in context
is to compare it with silver concentrations found in field sed-
iments. The National Sediment Inventory (NSI) [24] includes
data from both fresh and saltwater sediments of the United
States. In the NSI database, approximately 90% of the sedi-
ments collected in the field had less than 1.0 jxg/g dry weight
silver, whereas approximately 99% of these sediments had less
than 10.0 (ig/g dry weight silver (J. Keating, personal com-
munication) Bearing in mind that several of the sediments
used in the studies shown in Figures 2A, 3B, and 4C had low
AVS concentrations, the results of the spiked-sediment tests
would indicate that most field sediments do not contain enough
silver to cause an acute effect if silver was acting alone. How-
ever, silver may also contribute to metals toxicity by binding
with available sulfide, freeing up other metals in the way that
spiking with silver increased the bioavailability of zinc and
copper in the study of Call et al [6]
Bioaccumulation and chrome toxicity. Hirsch [5] did not
find any accumulation of silver in obgochaetes when they were
exposed to sediments spiked with high concentrations of silver
sulfide. This is consistent with the lack of biological avail-
ability of other metal sulfides [8]. However, several studies
have shown divalent metals are sometimes accumulated from
sediments with an excess of AVS over metal [see 25]. Also,
some benthic organisms can accumulate metals from the over-
lying water [26], and silver can also be accumulated from food
[27]. Further, it should be noted that, with the exception of
the [5] bioaccumulation study, all of the studies discussed in
this paper have examined acute mortality. It has been shown
that the same methods used to predict acute toxicity can be
used to predict biological effects in chronic sediment tests with
some metals [28-30], but this has yet to be shown with silver.
Predicting the toxicity of silver in field sediments. Acute
toxicity predictions in sediments with various concentrations
of silver and the other SEM metals in relation to AVS are
presented in Table 3. Consider first the unlikely sediment in
which silver is the only toxic metal present. If AVS exceeds
silver in this sediment, the AVS should be above detection
limits and will exceed silver in the SEM. Silver should not be
present in the IW, and this sediment should not be toxic because
of silver If silver exceeds the sulfide in the sediments, there
will be no detectable AVS in the sediment, and even a small
amount of silver in the SEM will exceed the measured AVS
This sediment may be acutely toxic due to silver. (It may, of
course, be toxic due to something else.) It should be noted
that the lack of detectable AVS in a sediment does not imply
that all of the available sulfide is bound by silver. The sediment
could be completely oxic or the AVS could be bound by an-
other metal that is not entirely acid soluble, such as copper
[3]. If IW silver measurements are available, further refinement
of the prediction of toxicity is possible. Sediments that do not
have lexicologically significant amounts of silver in the IW
should not be toxic due to silver, even if silver exceeds AVS
in the sediment. However, any sediment in which silver ex-
ceeds AVS should be looked at carefully. Sediments that do
have lexicologically significant amounts of silver in the IW
are certainly potentially toxic due to silver (Fig. 2C).
Next consider the much more likely case of a sediment that
is contaminated with silver and other sulfide-forrrung metals.
A mole of sulfide will be bound for every 2 mol of silver (the
1.2 ratio is due to the fact that silver is monovalent) present
because the sulfide solubility product of silver is lower than
that of other metals [18] If sulfide exceeds the sum of the
[Ag]/2 plus the other SEM metals (cadmium, copper, lead,
nickel, and zinc), measurable AVS will exceed measurable
SEM, these metals should not be present in lexicologically
significant concentrations in the IW, and the sediment should
not be acutely toxic due to these metals. If the sum of the
[Ag]/2 plus the other SEM metals exceeds the sulfide in the
sediment, measurable SEM will exceed measurable AVS, and
the sediment may be acutely toxic due to ihese metals. As was
Irue with the silver-only case described previously, sediments
-------
Predicting the toxicity of silver-spiked sediments
Environ Toxicol Chem. 18, 1999 47
that do not have lexicologically significant amounts of these
metals in the IW should not be toxic due to these metals, even
if SEM exceeds AVS in the sediment. However, any sediment
in which SEM exceeds AVS should be looked at carefully.
Sediments that do have lexicologically significant amounts of
metals in the IW are certainly potentially toxic due to these
metals [3].
Theoretically, almost all of the silver in the sediment will
be bound to sulfide in any sediment in which there is an excess
of sulfide over [Ag]/2. This is because of the extremely low
solubility of silver sulfide [18] However, if the sum of the
[Ag]/2 plus the other SEM metals exceeds the sulfide in the
sediment, some of the other SEM metals may be present in
the IW, and the sediment may be toxic due to these metals In
these sediments, the silver may be contributing to the overall
toxicity of metals in the sediment by tying up sediment sulfide
that might otherwise bind the other SEM metals. This is ap-
parently what happened in some of the sediments of Call et
al [6], in which zinc and copper were released into IW due
to the addition of silver to the sediment.
The release of metals into the IW in relation to sulfide
solubility is not peculiar to silver. Berry et al. [3] found that
metals appeared in the IW in the order of their Kv, with copper
appearing Jast in their experiments. What is unusual about
silver is that the solubility of silver sulfide in the AVS ex-
traction is so low that any sediment with an excess of [Ag]/2
over sulfide will have no measurable AVS present (Table 2).
Thus, any sediment with measurable AVS should not have
silver in the IW and should not be acutely toxic because of
silver.
CONCLUSION
It is important to remember that the data presented in this
paper apply primarily to acute mortality and may not address
all effects due to chronic exposure or bioaccumulation [2]. For
example, Hook and Fisher [31] demonstrated in preliminary
experiments that silver may effect copepod reproduction at
environmental concentrations. However, taken together, the
freshwater and saltwater sediment results indicate that silver:
AVS relationships and IWTU can provide insight into the role
of silver in the possible toxicity of sediments From the point
of view of AVS and SEM measurements, these results would
indicate that silver can be included along with cadmium, cop-
per, lead, nickel, and zinc in sediment assessment. If the sum
of the SEM for these metals is less than AVS in a sediment,
the sediment should not be acutely toxic due to these metals.
Furthermore, even in sediments that have an excess of metal
over sulfide, as long as there is measurable AVS, any observed
acute mortality should not be due directly to silver in the pore
water
Acknowledgement—We thank R Burgess, C. Ingersoll, R Pruell, W.
Boothman, and two anonymous reviewers for comments that im-
proved the manuscript The manuscript also profited from conversa-
tions about experimental design and data interpretation with D Di
Toro and J Mahony D Call and E Deaver provided raw data from
their expenments and graciously consented to having them replotted
This document has been reviewed in accordance with U.S. Environ-
mental Protection Agency, National Health and Environmental Re-
search Laboratory, Atlantic Ecology Division policies and approved
for publication. The contents of this publication do not necessarily
reflect the views of the U S Environmental Protection Agency Men-
tion of trade names or commercial products does not constitute en-
dorsement or recommendation for use
REFERENCES
1 Purcell TW, Peters J 1998 Sources of silver in the environment
Environ Toxicol Chem 17:539-546.
2 Luoma SN, Ho YB, Bryan OW 1995 Fate, bioavailabihty and
toxicity of silver in estuanne environments Mar, Pollut Bull 31
44-54
3 Berry WJ, Hansen DJ, Mahony JD, Robson DL, Shipley BP,
Rogers B, Corbin JM 1996. Predicting the toxicity of metals-
contaminated sediments using acid volatile sulfide and interstitial
water normalization in laboratory-spiked sediments Environ Tox-
icol Chem 15-2067-2079
4. Hirsch MP 1998 Toxicity of silver sulfide-spiked sediments to
the freshwater amphipod (Hyalella azteca) Environ Toxicol
Chem 17 601-604
5 Hirsch MP 1998. Bioaccumulation of silver from laboratory-
spiked sediments in the ohgochaete (Lumbriculus variegatus)
Environ Toxicol Chem 17:605-609
6 Call DJ, Polkinghome CN, Markee TP, Brooke LT, Geiger DL, Gor-
such JW, Robillard KA 1999 Silver toxicity to Chironomus tentans
in two freshwater sediments. Environ Toxicol Chem 18:30-39.
7. Rodgers JA Jr, Deaver E, Rodgers PL 1997 Partitioning and
effects of silver in amended freshwater sediments Ecotoxicol
Environ Saf 37 1-9
8 Ankley GT, Di Toro DM, Hansen DJ, Berry WJ, Mahony JD.
1996. Technical basis and proposal for deriving sediment qual-
ity criteria for metals Environ Toxicol Chem 15-2056-2066.
9. Hansen DJ, Berry WJ, Mahony JD, Boothman WS, Robson DL,
Ankley GT. 1996 Predicting the toxicity of metals-contaminated
field sediments using interstitial concentration of metal and acid
volatile sulfide normalizations Environ Toxicol Chem 15:2080-
2094.
10. Scott KJ, Redmond MS. 1989 The effects of a contaminated
dredged material on laboratory populations of the tubiculous am-
phipod, Ampelisca abdita In Cowgill UM, Williams LR,
Aquatic Toxicology and Hazard Assessment, Vol 12. STP 1027.
American Society for Testing and Materials, Philadelphia, PA, pp
289-303.
11. Di Toro DM, Mahony JD, Hansen DJ, Scott KJ, Hicks MB, Mays
SM, Redmond MS 1990. Toxicity of cadmium in sediments: Role
of acid volatile sulfide Environ Toxicol Chem 9-1489-1504
12. Di Toro DM, Mahony JD, Hansen DJ, Scott KJ, Carlson AR,
Ankley GT. 1992. Acid volatile sulfide predicts the acute toxicity
of cadmium and nickel in sediments. Environ Sci Technol 26:
96-101.
13. HampsonBL 1977 Relationship between total ammonia and free
ammonia in terrestrial and ocean water Cons Inter Explor Mer
37:117-122
14. Hamilton MA, Russo RC, Thurston RV. 1977. Trimmed Spear-
man-Karber method for estimating median lethal concentrations
in toxicity bioassays Environ Sci Technol 11 714-719 Correc-
tion 12 414 (1978)
15. Mearns AJ, Swartz RC, Cummins JM, Dinnel PA, Plesha P, Chap-
man PM. 1986. Inter-laboratory comparison of a sediment toxicity
test using the marine amphipod, Rhepoxymus abronius. Mar En-
viron Res 19:13-37
16. U.S Environmental Protection Agency 1994. Statistical sum-
mary, EMAP-estuanes, Virginian Province—1991 EPA/620/R-
94/005 Technical Report Washington, DC
17. Stephan CE, Mount DI, Hansen DJ, Gentile JH, Chapman GA,
Brungs WA. 1985. Guidelines for denvmg numerical national
water quality criteria for the protection of aquatic organisms and
their uses PB85-227049. National Technical Information Service,
Springfield, VA, USA
18. Lide DR, ed 1995. CRC Handbook of Chemistry and Physics,
76th ed CRC, Boca Raton, FL, USA, pp 8-85
19 US Environmental Protection Agency 1994 Methods for as-
sessing the toxicity of sediment-associated contaminants with
tuanne and marine amphipods EPA/600/R-94/025 Technical Re
port Washington, DC
20 Scemtific Applications International Corporation 1993 Data re-
port on comparative sensitivity of four amphipod species to aque-
ous ammonia. Report 1071 Narragansett, RI, USA
21 Shaw JR, Wood CM, Birge WJ, Hogstrand C 1998 Toxicity of
silver to the marine teleost (Ohgocottus maculosus) Effects of
salinity and ammonia Environ Toxicol Chem 17 594-600
22 State of Florida 1994 Approach to the Assessment of Sediment
-------
48 Environ Toxicol Chem 18, 1999
W J Berry et al
Quality in Florida Coastal Waters, Vol 1 Florida Department
of Environmental Regulation, Tallahassee, FL, USA
23 Cubbage J, Batts D, Bnedenbach S 1997 Creation and analysis
of freshwater sediment quality values in Washington State Pub-
lication 97-323a Washington State Department of Ecology,
Olympia, WA, USA
24 U S Environmental Protection Agency 1997 The incidence
and severity of sediment contamination in the surface waters
of the United States EPA 823/R-97/006 Washington, DC
25 Ankley GT 1996 Evaluation of metal/acid-volatile sulfide rela-
tionships in the prediction of metal bioaccumulation by benthic
macromvertebrates Environ Toxicol Chem 15 2138-2146
26 Warren L, Tessier A, Hare L 1998. Modeling sediment accu-
mulation by benthic invertebrates in situ Relative contributions
of sediment and overlying water compartments to organism cad-
mium concentrations Limnol Oceanogr (in press)
27 Fisher N, Wang W-X 1998 Trophic transfer of silver to marine
herbivores A review of the recent studies Environ Toxicol Chem
17.562-571
28 DeWitt TH, Swartz RC, Hansen DJ, McGovem D, Berry WJ
1996 Interstitial metal and acid volatile sulfide predict the bio-
availabihty of cadmium dunng a full life cycle sediment toxicity
test using the estuanne amphipod, Leptocheirus plumulosus En-
viron Toxicol Chem 15 2095-2101
29 Hansen DJ, Mahony JD, Berry WJ, Benyi S, Corbm J, Pratt
S, Able MB 1996 Chronic effect of cadmium in sediments
on colonization by benthic marine organisms An evaluation
of the role of interstitial cadmium and acid volatile sulfide in
biological availability Environ Toxicol Chem 15 2126-2137
30 Sibley PK, Ankley GT, Cotter AM, Leonard EN 1996 Predicting
chronic toxicity of sediments spiked with zinc An evaluation of the
acid volatile sulfide (AVS) model using a life-cycle test with the
midge Chironomus tentons Environ Toxicol Chem 15 2102-2112
31 Hook S, Fisher N. 1997. Sublethal response of zooplankton to
silver, the importance of exposure route Abstracts, 18th Annual
Meeting of the Society of Environmental Toxicology and Chem-
istry San Francisco, CA, November 16-20, p 51.
-------
Environ. Sd. Technot 1992, 26. 96-101
Acid Volatile Sulfide Predicts the Acute Toxicity of Cadmium and Nickel in
Sediments
Dominic M. Di Toco.*-' John D. Mahony,* David J. Hanson.5 K. John Scott,1 Anthony R. Carlson/ and
Gerald T. Ankleyx
Environmental Engineering and Science Program and Chemistry Department, Manhattan College, Bronx, New York 10471,
EPA Environmental Research Laboratory, Narragansett, Rhode Island 02592. Science Applications International Corporation,
Narragansett, Rhode Island 02592, and EPA Environmental Research Laboratory, Duluth, Minnesota 55804
• Laboratory toxicity tests using amphipods, oligochaetes,
and snails with spiked freshwater and marine sediments
and with contaminated sediments collected from an EPA
Superfund site demonstrate that no significant mortality
occurs relative to controls if the molar concentration of acid
volatile sulfide (AVS) in the sediment is greater than the
molar concentration of simultaneously extracted cadmium
and/or nickel Although it is well-known that these metals
can form insoluble sulfides, it apparently has not been
realized that AVS is a reactive pool of solid-phase sulfide
that is available to bind metals and render that portion
unavailable and nontoxic to biota. Thus, the AVS con-
centration of a sediment establishes the boundary below
which these metals cease to exhibit an acute toxicity in
freshwater and marine sediments.
Introduction
Predicting the bioavailability and toxicity of metals in
aquatic sediments is a critical component in the develop-
ment of sediment quality criteria (1). The use of total
sediment metal concentration Cimol/g dry weight) as a
measure of the bioavailability concentration is not sup-
ported by available data (2). Different sediments exhibit
different degrees of toxicity for the same total quantity
of a metal. These differences have been reconciled by
relating organism response to the chemical concentration
in the interstitial water of the sediments (3-5). In addition,
a substantial number of experiments using water-only
exposures point to the fact that biological effects can be
correlated to the divalent metal activity {M2*) (6, 7). This
suggests that the bioavailability of metals in sediments is
related to the chemical activity of the metal in the sedi-
ment-interstitial water system. Hence, the sediment
properties which determine the metal activity in the sed-
iment-interstitial water system also determine the fraction
of the metal that is bioavailable and potentially toxic.
Unless explicitly stated, when we refer to metals in this
paper we mean cadmium and/or nickel. For sediments
and metals tested to date, metal activity in the sedi-
ment-interstitial water system, as measured by acute
toxicity to benthic organisms, is strongly influenced by the
sulfide and metal concentrations that are extracted from
the sediment using cold hydrochloric -acid. This sulfide
fraction is conventionally referred to as the acid volatile
sulfide or AVS (8) The metal concentration that is si-
multaneously extracted we term the simultaneously ex-
tracted metal or SEM. The significance of performing the
sulfide and metal extraction under equivalent conditions
'Environmental Engineering and Science Program, Manhattan
College
' Chemistry Department, Manhattan College
'EPA Environmental Research Laboratory, Narragansett, RI
1 Science Applications International Corp.
-*• EPA Environmental Research Laboratory, Duluth, MN.
96 Environ Sd. Techno!., Vol. 26, No 1, 1992
is discussed below. For [SEM]/[AVS] < 1, no acute tox-
icity (mortality >50%) has been observed in any sediment
for any benthic test organism. For [SEM]/[AVS] > 1, the
mortality of sensitive species (e.g., amphipods) increases
in the range of 1.5-2.5 /tmol of SEM/pmol of AVS.
This observation is important because acid volatile
sulfide is found in most freshwater and marine sediments.
It is found in sediments with sandy and gravelly textures
that do not resemble the anoxic sulfidic sediments that are
more commonly associated with the presence of sulfide.
Concentrations range from <0.1 to >50 pmol of AVS/g
(9-13).
Experimental Procedures
The results of four separate experiments are presented
in this paper. The detailed experimental procedures are
described elsewhere (14-17). The sediment toxicity tests
generally followed ASTM recommendations (18). Flowing
water (~10 volume replacements/day) and aeration en-
sured acceptable dissolved oxygen concentration. The tests
were conducted using 200-900-mL exposure vessels with
200 mL of sediment (3.5-cm depth) and 600 mL of over-
lying water. The animals were exposed for 10 days to
control and test sediments.
Sediments were spiked by adding 1.0 L of wet sediment
to 2.0 L of dilution water into which a weighted amount
of cadmium or nickel chloride had been dissolved. The
tests were initiated by adding the sediments to the expo-
sure containers, waiting from 1 to 6 days, and adding the
animals. After termination, the contents of each exposure
container were sieved and counted. Missing individuals
were counted as mortalities. Parallel exposures were used
for chemical measurements.
The acid volatile sulfide (AVS) concentration is the
solid-phase sulfide that is soluble in room-temperature 0.5
M HC1 in 1 h. The measurement technique is to convert
the sulfides to HzSfeq), purge it with oxygen-free nitrogen
gas (four bubbles/6), and trap it in a gas-tight assembly
(14,19). The reaction vessel is followed by a pH 4 chloride
trap (0.05 M potassium hydrogen phthalate) and two
sulfide traps (0.1 M silver nitrate) for trapping HjS as AgjS
precipitate. For 10-15 g of wet sediment the detection
limit is ~0.5 pmol/g The simultaneously extracted metal
(SEM) was measured in a filtered aliquot by conventional
atomic absorption methods. The AVS and SEM were
measured at the start of the experiment when animals were
added and at the termination in the parallel exposure
vessels.
The initial experiment (14) exposed two marine am-
phipods (Ampelisca abdita and Rhepoxynius hudsoni) to
three uncontaminated marine sediments: a fine-grained
sediment with a relatively high AVS sediment (15 Minor
of AVS/g) from central Long Island Sound, NY; a sandy
sediment with a relatively low AVS (U$ jjmol of AVS/g)
from a salt water pond in Ninigret, RI; .^ft-an equivolume
mixture of the two sediments (4.3 MMf^f AVS/g). For
~~v
0013-936X/92/0926-0096$03.00/0 © 1991 American Chemical Society
-------
this experiment, the total acid extractable cadmium was
measured separately. As shown below, this is equivalent
to the SEM measurements for cadmium. For the re-
maining experiments, the SEM concentrations were
measured directly.
The second experiment (IS) simultaneously exposed two
freshwater organisms, a snail, Helisoma sp., and an oli-
gochaete, Lumbriculus uariegatus, to cadmium added to
three uncontaminated freshwater sediments from Pe-
quaywan Lake, MN (42 ftmo\ of AVS/g), East River, Wl
(8.8 Mmol of AVS/g), and West Bearskin Lake, MN (3.6
Mmol of AVS/g). The third experiment (16) exposed A.
abdita to nickel added to central Long Island Sound and
Ninigret Pond sediments.
The final experiment (17) exposed the freshwater am-
phipod, Hyalella azteca, to 17 sediment samples taken
from Foundry Cove, a small (213 ha) predominantly
freshwater cove in the upper reach of the tidal portion of
the Hudson River, NY. These sediments were contami-
nated with cadmium and nickel from a battery manufac-
turing facility (20,21). The sediments spanned the range
from fine-grained sediments, highly enriched in organic
carbon, to gravelly composites with low organic carbon
concentrations. The cadmium and nickel concentrations
are approximately equimolar throughout the range of
sediment concentrations present, 0.3-1000 taool of SEM/g.
AVS ranged from 0.1 to 47 itmol/g, and [SEM]/[AVS]
ranged from 0.1 to > 100 with several in the critical range
of <1-10.
Chemical Basis
The chemical basis for the primacy of the sediment
sulfide phase for metal binding is that, at equilibrium, the
sulfide ion successfully competes with any other commonly
present dissolved or particle-associated ligand for the metal
ion to form insoluble metal sulfides (22,23). Moreover,
as shown below, the solubility products of metal sulfides
are so small that the metal activity in the sediment-in-
terstitial water system is well below the metal activity that
causes mortality of the exposed organisms.
What is surprising is that this prediction, which is based
on the dissolved and solid-phase chemical species distri-
bution at thermodynamic equilibrium, appears to be
achieved quite rapidly, on the time scale of minutes to
hours (14). If the sulfides in the sediment-interstitial water
system were all aqueous species, then this rapid reactivity
would be expected. However, the largest reservoir of
sulfides in sediments is solid-phase iron sulfide and pos-
sibly manganese sulfide (8,19,24), and not all the sediment
sulfide phases are equally reactive. They are operationally
separated into three compartments: AVS, which are ex-
tracted with cold hydrochloric acid and are primarily iron
monosulfides, FeS(s); the more stable iron pyrite, FeSate),
which is not extracted by this procedure (8,24); and or-
ganic sulfides that are associated with the organic matter
in sediments and are also not extracted (25).
We have demonstrated (14) that when cadmium is
added to a sediment sample, the cadmium will rapidly
displace the iron in FeS(s) to form cadmium sulfide pre-
cipitate, CdS(s):
FeS(s) -» CdS(s) + Fe2+
(1)
The consequence of this replacement reaction can be seen
using an analysis of the M(n)-Fe(II)-S(II) system with
both MS(s) and FeS(s) present. M(II) represents any
metal that forms a sulfide that is more insoluble than FeS.
If the added metal, [M]A, is less than the AVS present in
the sediment then, as shown in the Appendix, the ratio of
metal activity to total metal in the sediment-interstitial
Table I. Metal Sulfide Solubility Products'
metal
sulfide
FeS
NiS
ZnS
CdS
PbS
CuS
HgS
log
K^
-364
-9.23
-9.64
-14.10
-14.67
-2219
-38.50
log
Kv
-2239
-27.98
-2839
-32.85
-3342
-40.94
-57.25
log
(KMS/KF.S)
-5.59
-6.00
-1046
-11.03
-18.55
-3486
• Solubility products, Kv3, for the reaction M2+ + HS~ — MS(s)
+ H* for CdS (gieenockite), FeS (mackmawite), and NiS (miller-
ite) from ref 23. Solubility products for CuS (covellite), HgS (me-
tacumabar), PbS (galena), and ZnS (wurtzite), and pK, = 18.57 for
the reaction HS" ** H* + S*~ from ref 26. Kv for the reaction NP+
+ S2" »• MS(s) is computed from log Kvl and pK,.
water system is less then the ratio of the MS to FeS sol-
ubility products:
This is a general result that is independent of the details
of the interstitial water chemistry. In particular it is in-
dependent of the Fe2* activity. Of course the actual value
of the ratio {M2*}/ [M]A depends on aqueous speciation, as
indicated by eq 1. However, the ratio is still less than the
ratio of the sulfide solubility products.
The sulfide solubility products and the ratios are listed
in Table I. The ratio of cadmium activity to total cad-
mium is less than 10"las. For nickel the ratio is less than
lO*6-6. By inference this reduction in metal activity will
occur for any other metal that forms a sulfide that is sig-
nificantly more insoluble than iron monosulfide. The
ratios for the other metals in Table I, Zn, Pb, Cu, and Hg,
indicate that metal activity for these metals will be very
small in the presence of excess AVS.
If, on the other hand, reactive metal is present in excess
of the AVS then, as shown in the Appendix, the metal
activity is
If no other strong complexing ligand is present, CM** ~ 1,
the metal activity will approximate the metal in excess of
the AVS, and the sediment is likely to exhibit toxicity to
sensitive organisms. We conclude that the concentration
of AVS determines the boundary between low metal ac-
tivity and potentially high metal activity.
Extraction
In order to quantify the metal concentration in the
sediment that will be compared to the AVS, we have used
the concentration that is simultaneously liberated by the
AVS extraction, rather than the total metal concentration
of the sediment. The necessity for this procedure is dem-
onstrated by the data in Figure 1, which compares the AVS
extracted at various concentrations of metal added to
sediments.
For [metal]/[AVS] < 1, the AVS is primarily FeS, which
is soluble in the AVS extraction. For [metal]/[AVS] > 1
the FeS is converted to a metal sulfide via the displace-
ment reaction, eq 1. For cadmium, the sulfide extracted
is essentially constant throughout the range of cadmium
additions, indicating that CdS is completely soluble hi the
AVS extraction. By contrast, the decrease in extracted
sulfide for [Ni]/[AVS] > 1 indicates that nickel eulfide is
not completely soluble in the extraction.
If a more efficient procedure were used to increase the
fraction of metal extracted which did not also capture the
Environ. SO. Techno)., Vol. 26, No. 1, 1992 97
-------
Cadmium
Nickel
10O.OO
O)
"5
w
10.00
1.00
East River
FeS
• T - 0 d
• T - 10 d
CdS
100.00
10.00
O.01 0.1O 1.OO 10.OO 1OO.OO
Cd/AVS
1.00
LI Sound • T - o d
FeS
A
*:
• T - 10 d
' NIS
1
1
!M • ^
i •
i * *
i *
Ni/AVS (umol/nmol)
Figure 1. Add volatte sulfide extracted versus the cadmium and nickel AVS ratios. Data from the start (f = 0 day) and end (f = 10 day) of
the exposure period. The dotted foe deimits the regions where the AVS Is either predominantly FeS or either CdS or MS.
additional sulfide extracted, then the sulfide associated
with the additional metal release would not be quantified.
This would result in an erroneously large metal to AVS
ratio. Hence, we use the AVS as the measure of reactive
sulfide in the system and the simultaneously extracted
metal, SEM, as the measure of reactive metal Since these
quantities are operationally defined, it remains to dem-
onstrate their utility experimentally.
Results
The four experiments described above have been per-
formed to test the utility of SEM to AVS ratios for pre-
dicting bioavailability and toxicity of metals to benthic
organisms. Three were performed with spiked sediments;
the fourth employed contaminated sediments from an
EPA Superfund site. The results are presented in Figure
2. The data for each experiment in which only the sed-
iment type is varied are superimposed with the exception
of the first experiment, where two amphipods, A. abdita
and R. hudsoni, with similar sensitivity (14) were used.
The lines connect the results for the same sediment The
lack of a unique relationship between extracted metal
concentration on a dry weight basis and organism mortality
is reflected by the different mortality-concentration re-
lationship for each sediment in the spiked metal experi-
ments.
The scatter in observed mortality for the Foundry Cove
sediments is striking. Metal concentrations from 0.1 to
28 pmol of SEM/g were not toxic in some sediments,
whereas 0.2-1000 pmol of SEM/g were lethal in other
sediments. These results reaffirm the observation that the
bioavailable fraction of metals in sediments varies from
sediment to sediment, and in this case, it varies dramat-
ically.
By contrast, a clearly discemable mortality-concentra-
tion relationship is observed in Figure 3, where mortality
is related to the SEM to AVS molar ratio. The chemical
theory predicts that mortality should begin to occur at
[SEM]/[AVS] = 1 (eqs 2 and 3). No mortality in excess
of 50% is observed for sediments with [SEM]/[AVS] <
1. For sediments with [SEM]/[AVS] > 1-3, the mortality
increases dramatically. For sediments with [SEM]/[AVS]
> 10,80-100% of individuals from all the tested species
died.
Conclusions
These data suggest the following conclusions. If the ratio
[SEM]/[AVS] = 1 is used to discriminate toxic from
nontoxic sediments (greater or less than 50% mortality,
respectively), then for the 117 experiments performed,
51% are correctly classified as nontoxic (bottom left
quadrant in Figure 3) and 42% are correctly classified as
toxic (top right quadrant). That 7% that are misclassified
as toxic (bottom right quadrant) follows from the as-
sumption that metal activity will invariably be high enough
to cause toxicity if [SEM]/[AVS] > 1, eq 3. It is possible
that other ligands, associated with sediment sorption, for
example, are reducing the metal activity below that which
is lethal to the text organisms. Also, less sensitive organ-
isms can tolerate the increased metal activity even if
[SEM]/[AVS] > 1. For organisms that are present when
[SEM]/[AVS] > 1, preliminary data suggest that the ex-
tent to which metals bioaccumulate is strongly influenced
by the AVS concentration (17, 28).
If a more restrictive interpretation is adopted and the
criterion [SEM]/[AVS] < 1 is used only to predict when
a sediment is not acutely toxic, then all experiments are
correctly classified. We assume that the toxic Foundry
Cove sediment with [SEM]/[AVS] = 0.98 is indistin-
guishable from a unity ratio. Hence our data indicate that
sediments with [SEM]/[AVS] < 1, perhaps [SEM]/[AVS]
< 0.9 as a safety factor, do not cause greater than 50%
mortality in all the sediment toxicity tests performed to
date. This is directly attributable to the excess AVS in
the sediment, which assures that the metal activity in the
sediment-interstitial water system is below the lethal metal
activity for the organisms tested.
It is possible that these results are due to a covariation
of AVS with other sediment properties, for example, or-
ganic carbon or iron content, which are actually controlling
the metal bioavailability. However, the fact that the
boundary occurs at [SEM]/[AVS] = 1, as predicted by eq
2, which is based on the supposition that AVS is the
controlling sediment property, strongly argues that the
relationship is casual, rather than correlative.
It should be noted that if the AVS concentration is
effectively zero, as it would be in fully aerobic sediments,
then other sediment properties would control the metal
activity. This does not contradict the assertion that the
[SEM]/[AVS] molar ratio of less than 1 predicts an ab-
sence of toxicity since in this case the molar ratio would
be very large. The prediction is not much help but it is
still correct. However, even a small AVS concentration,
[AVS] ~ 0.1 Mmol/g, can sequester a significant quantity
of metal and should be taken into account in determining
the potential for metal toxicity for these sediments.
98 Environ Set. Techno!. Vd 26. No 1. 1992
-------
AMPHIPOO - Cd - SW
AMPHIPOD - Ni - SW
100
ao
60
4O
20
o
0.1 1.O 10.0 1OO.O 10OO.O
O.1
1.0 1O.O 1OO.O 1OOO.O
SNAIL - Cd - FW
OUGOCHAETE - Cd - FW
rr
o
(O
<
o
c
o
100
80
60
4O
20
O
o-cr
0.1
1.0
10.0 100.0 10OO.O
FOUNDRY COVE
AMPHIPOD - Cd+Ni - FW
100
80
80
4O
20
O
•A* * •**** *
* *
*
*
*
• * *
100
80
60
40
20
O
O.1
1.0
1O.O 10O.O 10OO.O
LEGEND
SNAIL. OUGOCHAETE
• • PEOUAVWAN IK - Cd
• • EAST RIVER - Cd
O O WEST BEARSKIN LK - Cd
-AMPHIPOO-
•*• FOUNDRY COVE - Cd * NI
• U SOUND - Cd
• MIXTURE - Cd
O NMM3RET PD - Cd
d> U SOUND - HI
O NMIGRET PD - NI
O.I
1.O
10.0 100.0 1000.0
SEDIMENT (iimol SEM/gm dry wt)
Flgur«2. Organism mortality versus metal concentration. Animals and metals as indicated for saawater (SW) and freshwater (FW) exposures.
The Ines connect results from the same sedknenL For the Foundry Cove sedments. the metal concentration Is the molar sum of the simultaneously
extracted cadmium and nickel.
100
so
«o
4O
20
o
••
O.O1 O.1O 1.OO 1O.OO
SEDIMENT (nmol SEMXitmol AVS)
1OO.OO
Figure 3. Organism mortally versus the molar ratio of SEM to AVS
of the sediment The sedknent metal and AVS are the averages of
the initial and final measured concentrations. For the Foundry Cove
sediments, the metal concentration b the molar sum of the sbnuKa-
neously extracted cadmium and nickel. Symbols defined In Figure 2.
Additionally, it is important to realize that "anaerobic*
and "aerobic" sediments are not precise classifications.
Sediments are characterized by an aerobic layer underlaid
by an anaerobic layer. The sediments employed in these
experiments did not inhibit the survival of the obligate
aerobic organisms used in the control exposures—in par-
ticular the tube-building amphipods. Most benthic aerobic
organisms survive in sediments that are underlaid with
completely anaerobic sediments, which are characterized
by significant AVS concentrations. Our experiments
suggest that the presence of AVS in the anaerobic layer
is sufficient to reduce the metal activity to which the an-
imals are exposed.
However, we have not examined the extent to which an
aerobic layer depth is sufficient to mitigate the influence
of the lower layer AVS. Thus it is possible to imagine a
situation where AVS at depth (>10 cm) might not control
metal activity in a completely aerobic overlying layer— the
top 10 cm for example—where the animals are exposed.
It seems likely that even in this situation the presence of
a sink of metals at depth would reduce the activity in the
entire sediment to below toxic levels. The reasoning b that
the diffusional transport of metal in the interstitial water
would bring metals from a presumably higher concentra-
tion in the aerobic layer interstitial water to the lower
concentration in the anaerobic layer. This would even-
tually deplete the aerobic layer of metals and establish a
uniformly low metal activity in the interstitial water-
sediment system. Thus, even in this case, we would expect
that excess AVS would predict the absence of acute tox-
icity. However, this is yet to be demonstrated experi-
mentally.
We believe that the date presented in this paper dem-
onstrate that, for the first time, it is possible to predict
Environ. SO. Techno!., Vol. 26, No. 1, 1992 99
-------
when a sediment will not be acutely toxic due to cadmium
and/or nickel contamination and, by implication, to all
toxic metals that form sulfides that are significantly less
soluble than FeS. The criterion is that the molar sum of
simultaneously extracted Cd, Cu, Hg, Ni, Pb, and Zn is
less than the molar acid volatile sulfide concentration:
+ [SEM]Nl +
[SEM]Pb + [SEMWtAVS] < 1 (4)
It should be noted that in order to apply this relation-
ship it is necessary to measure all the toxic simultaneously
extracted metals that are present in amounts that con-
tribute significantly to the molar SEM sum, typically Cd,
Cu, Ni, Pb, and Zn. Failing to do this could lead to an
incorrect prediction of the lack of acute toxicity, i.e., the
sum of the measured SEMs to AVS ratio is less than 1,
when in fact the unmeasured metal could increase the
SEM concentration so that the ratio exceeds 1 and toxicity
would be possible. Thus, the acute toxicities of these
metals are interrelated, with each contributing to the AVS
that is bound by toxic metal. If excess AVS remains, no
acute toxicity is expected. If excess metal remains, then
the most soluble, Ni and Zn, will appear as free metal and
toxicity is possible.
Acknowledgments
The assistance and encouragement of the following
people is gratefully acknowledged: Christopher Zarba,
Criteria and Standards Division, U.S. EPA; Richard
Swartz, EPA Research Laboratory, Newport, OR; Mark
Springer and Mark Huston, Region n, US. EPA; Herbert
Allen, University of Delaware; Robert Thomann, Man-
hattan College; and, most of all, our research assistants at
Manhattan College, M. Hicks, S. Mayr, I. Sweeney, P.
Morgan, C. Sydlik, L. Milevoj, and C. Begley; at the EPA
Narragansett Laboratory, W. Berry, M. Redmond, D.
Robson, and K. McKenna (SAIC); and at the EPA Duluth
Laboratory, G. Phipps, V. Mattson, E. Leonard, P. Kosian
(AScI), and A. Cotter (AScI).
Appendix
Solubility Relationships for Metal Sulfides. Con-
sider the following situation: a quantity of FeS is titrated
with a metal that forms a more insoluble sulfide. We
analyze the result using an equilibrium model of the M-
(II)-Fe(H)-S(II) system. The mass action laws for the
metal and iron sulfides are
"I = KMS (5)
'-] = Kf# (6)
where [M2*], [Fe2*], and [S2"] are the molar concentrations;
TM". TFe1*. and 7s1- are the activity coefficients; and KMS
and KfeS are the sulfide solubility products. The mass
balance equations for total M(II), Fe(II), and S(II) are
a-V[M2+] + [MS(s)] = [M]A (7)
a-'F,«[Fe2+] -f [FeS(s)J = [FeS(s)], (8)
a-'s'-fS2-] + [MS(s)] + [FeS(s)] = [FeS(s)], (9)
where
aw* = [M2+]/[ZM(aq)J (10)
aF«« = [Fe2+]/[ZFe(aq)] (11)
«s>- = [S2-]/[£S(aq)] (12)
are the ratios of the divalent species concentrations to the
total dissolved M(H), Fe(H), and S(II) concentrations,
[£M(aq)], [£Fe(aq)],and [ES(aq)], respectively. [MS(s)J
and [FeS(s)] are the concentrations of solid-phase metal
and iron sulfides at equilibrium. [FeS(s)], is the initial iron
sulfide concentration in the sediment, and [M]A is the
concentration of added metal.
The solution of these five equation can be obtained as
follows. The mass balance eqs 7 and 8 for M(TJ) and Fe(II)
can be solved for [MS(s)] and [FeS(s)] and substituted in
the mass balance eq 9 for S(II):
-a-VfS2-] + crWFe2*] + a'VlM^J = [M]A (13)
The mass action eqs 5 and 6 can be used to substitute for
[Fe2+] and [M2*], which results in a quadratic equation
for [S2-]:
YM** /
[M]A (14)
The positive root can be accurately approximated by
TM»*TS»-
r
[M]A (15)
which results from ignoring the leading term in eq 14. This
is legitimate because the term in parentheses in eq 14 is
small relative to [M]A due to the presence of the sulfide
solubility products. As a result, [S2~] is also small since
it is in the denominator. Hence, the leading term in eq
14 must be small relative to [M]A and can safely be ignored.
The metal activity can now be found from the solubility
equilibrium eq 5:
VMS
[M]A
so that
where
and
Equation 17 can be expressed as
[M],
KftS\ a j- a KMS
\Pr 1, because it is
the reciprocal of two terms both of which are less than or
equal to 1, eq 18. They are «Fei* < 1, which is the ratio
of the divalent to total aqueous concentration, and -yFej+
< 1, which is an activity coefficient. The second term in
the denominator cannot be negative, 0Mj+.KMS/,KFeS > 0,
since all of its terms are positive. Thus, the denominator
of the expression in parentheses is always greater than 1,
0Fe»+ + PMH-KMS/KP& > 1- Therefore, the expression in
parentheses is always less than 1. Hence, the magnitude
100 Environ Sd Techno!. Vot 26. No 1, 1992
-------
of the ratio of metal activity to total added metal is
bounded from above by the ratio of the sulfide solubility
products:
|Me2+|/[M]
(21)
This result applies if [FeS], > [M]A so that excess (FeS(s)]
is present.
If sufficient metal is added to exhaust the initial quan-
tity of iron sulfide, then [FeS(s)] = 0. Hence, the iron
sulfide mass action equation (6) is invalid and the above
equation no longer applies. Instead, the only solid-phase
sulfide is metal sulfide and
[MS] = [FeS], (22)
so that, from the metal mass balance equation
IM2*} = 7M*»«M~UM]A - [FeS(s)],) (23)
This completes the derivation of eqs 2 and 3.
Glossary
[AVS]
|Fe2+)
[Fe2*]
[FeS(s)]
[FeS(8)]i
[M]A
[MS(s)]
is2-)
is2-]
[SEM]
[SEMJc,,
[SEM]H,
[SEMJNi
[SEM]Pb
[SEMJa,
acid volatile sulfide concentration (jtmol/g)
activity of Fe2+ (mol/L)
concentration of Fe2* (mol/L)
concentration of iron sulfide (mol/L)
initial iron sulfide concentration in the sedi-
ment (mol/L)
solubility product for FeS(s) [(mol/L)2]
solubility product for MS(s) [(mol/L)2]
divalent metal activity (mol/L)
concentration of M2* (mol/L)
concentration of added metal (mol/L)
concentration of solid-phase metal sulfide
(mol/L)
activity of S2" (mol/L)
concentration of S2" (mol/L)
simultaneously extracted metal concentration
0«nol/g)
simultaneously extracted Cd concentration
G«nol/g)
simultaneously extracted Cu concentration
Gtmol/g)
simultaneously extracted Hg concentration
(jtmol/g)
simultaneously extracted Ni concentration
Oimol/g)
simultaneously extracted Pb concentration
Oimol/g)
simultaneously extracted Zn concentration
(pmol/g)
{M2+}/[EM(aq)]
«s*-
[£M(aq)]
tLS(aq)]
activity coefficient of Fe2+
activity coefficient of M2*
activity coefficient of S2"
concentration of total dissolved Fe(II)
(mol/L)
concentration of total dissolved M(II) (mol/
L)
concentration of total dissolved SCO) (mol/L)
Registry No. Cd, 7440-43-9; Ni, 7440-02-0; SM8496-25-8;
ZnS, 1314-98-3; PbS. 1314-87-0; CuS, 1317-4(M; HgS, 1344-48-5.
Literature Cited
(1) Di Toro, D. M.; Zarba, C. S.; Hansen, D. J.; Swartz, R. C.;
Cowan, C. E.; Pavlou, S. P.; AUen, H. E.; Thomas, N. A.;
Paquin, P R; Berry, W. Environ. Toxicol. Chem., in press.
(2) Luoma, S. N. Sci. Total'Environ. 1983,28,1.
(3) Adams, W. J., Kimerle, R. A.; Mosher, R G. In Aquatic
Toxicology and Hazard Assessment: Seventh Symposium;
Cardwell, R. D., Purdy, R., Bahner, R. C., Eds.; American
Society for Testing and Materials, Philadelphia, PA, 1985,
p429.
(4) Swartz, RC.;Ditsworth,G. R;Schults,D.W.;Lambereon,
J. 0. Mar. Environ Res. 1985, 78, 133.
(5) Kemp,P.P.;Swartz,RC.Mar.Environ.Res 1988,26,135
(6) Sunda, W.; Guillard, R R. L. J. Mar. Res. 1976, 34, 511
(7) Borgmann, U. In Aquatic Toxicology, Nriagu, J O., Ed.;
J. Wiley: New York, 1983; p 47.
(8) Beroer, R. A. Am. J. Sci. 1967, 265, 773.
(9) AUer.RC. In Estuarine Physics and Chemistry. Studies
in Long Island Sound; Saltzman, B., Ed.; Academic Press:
New York, 1980; p 238.
(10) Reaves, C. PhJ). Thesis, Yale University, New Haven, CT,
1984.
(11) Nriagu, J. O. Limnol. Oceanogr. 1963,13, 430.
(12) Nriagu, J. O.; Coker, R D. Limnol. Oceanogr. 1976,21,485.
(13) Matisoff, G.; Fisher, J. B.; McCall, P. L. Geochim. Cos-
mochim. Acta 1981, 45, 2333.
(14) Di Toro, D. M.; Mahony, J. D.; Hansen, D. J.; Scott, K. J.;
Hicks, M B.; Mayr, S. M.; Redmond, M. S. Environ.
Toxicol. Chem. 1990,9,1487.
(15) Carlson, A. R; Phipps, G. L.; Mattson, V. R; Kosian, P.;
Cotter, A. ERL—Duluth Report No. 2471, EPA Environ-
mental Research Laboratory, Duluth, MN, 1990.
(16) Hansen, D.J.; Scott, K.J.ERL—Narragansett Report EPA
Environmental Research Laboratory, Narragansett, RI,
1990.
(17) Ankley,G.T4PhiPPe,G.L.; Kosian, P.; Cotter, A.; Mattson,
V. R; Mahony, J. D. Environ. Toxicol. Chem., submitted.
(18) American Society for Testing and Materials. Proposed New
Standard Guide for Conducting Solid Phase 10-Day Static
Sediment Toxicity Tests with Marine and Estuarine Am-
phipods. Draft No. 5,1989, American Society for Testing
and Materials, Philadelphia, PA.
(19) Morse, J. W.; Millero, F. J.; Cornwell, J. C.; Rickard, D.
Earth Sci. Rev. 1987,24,1.
(20) Hazen, R E4 Kneip.T.J. In Cadmium in the Environment,
Part I: Ecological cycling; Nriagu, J. O., Ed.; J. Wiley: New
York, 1980; p 400.
(21) Kmitson, A. B.; Klerks, P. L.; Levinton, J. S. Environ.
Pollut. 1987, 45, 291.
(22) Boulegue, J. In Trace Metals in Sea Water, Wong, C. S.,
Boyle, E., Bruland, K. W., Burton, J. D., Eds.; Plenum
Press: New York, 1983; p 563.
(23) Emerson, S.; Jacobs, L.; Tebo, B. In Trace Metals in Sea
Water, Wong, C. S., Boyle, E., Bruland, K. W., Burton, J.
D., Eds.; Plenum Press: New York, 1983; p 579.
(24) Comwell, J. C.; Morse, J. W. Mar. Chem. 1987, 22,193.
(25) Landers, D. H,-David, M.B^MitcheD,M.J./nt.J.£>iM)w».
Anal. Chem. 1983,14, 245.
(26) Schoonen, M. A. A.; Barnes, H. L. Geochim, Cosmochim.
Acta 1988,52, 649.
(27) Byrne, R H.; Kump, L. R; Cantrell, K. J. Mar. Chem. 1988,
25,163.
(28) Strip]in,B.D.Skagway Harbor Field Investigation. Tetra
Tech, Inc., Bellevue WA, 1990.
< Received for review December 4, 1990. Revised manuscript
received March 12,1991. Accepted July 19,1991. This research
was supported by an EPA Cooperative Agreement between
Manhattan College and EPA Environmental Research Labora-
tory, Narragansett, RI. The Manhattan College participation
in the Foundry Cove investigation was supported by the National
Institutes of Environmental Health Sciences, Superfund Haz-
ardous Substances Basic Research Program, Environmental
Medicine, New York University Medical Center.
Environ. Sd. Techno!., Vol. 26, No. 1, 1992 101
-------
APPENDIX B
-------
Research Papers on the Unavailability and Toxicity of Metals in Surface Waters2
Key Papers *
Allen, H. E. and D. J. Hansen. 1996. The importance of trace metal speciation to water quality
criteria. Water Environment Research. 68(l):42-54.
Campbell, P.G.C., 1995. "Interactions Between Trace Metals and Aquatic Organisms: A Critique
of the Free-ion Activity Model," Metal Speciation and Unavailability in Aquatic Systems, A. Tessier
and D.R. Turner, eds., IUPAC, John Wiley and Sons.
Janes, N. and R.C. Playle. 1995. Modeling silver binding to gills of rainbow trout (Oncorhynchus
mykiss). Environ. Toxicol. Chem. 14:1847-1858.
Meyer, J.S., R.C. Santore, J.P. Bobbitt, L.D. DeBrey, C.J. Boese, P.R. Paquin, H.E. Allen, H.L.
Bergman and D.M. DiToro. 1999. Binding of nickel and copper to fish gills predicts toxicity when
water hardness varies, but free-ion activity does not. Environ. Sci. Technol. (in Press).
Meyer, J.S. 1999. A mechanistic explanation for the ln(LC50) vs ln(Hardness) adjustment equation
for metals. Envrion. Sci. Technol. (in Press).
Pagenkopf, O.K. 1983. Gill surface interaction model for trace-metal toxicity to fishes: Role of
complexation, pH, and water hardness. Envrion. Sci. Technol. 17:342-347.
Playle, R.C., D.G. Dixon and K. Burnison. 1993a. Copper and cadmium binding to fish gills:
modification by dissolved organic carbon and synthetic ligands. Can. J. Fish. Aquat. Sci. 50:2667-
2677.
Playle, R.C., D.G. Dixon and K. Burnison. 1993b. Copper and cadmium binding to fish gills:
estimates of metal-gill stability constants and modeling of metal accumulation. Can. J. Fish. Aquat.
Sci. 50:2678-2687.
Tipping, E. 1994. WHAM--A chemical equilibrium model and computer code for waters,
sediments, and soils incorporating a discrete site/electrostratic model of ion-binding by humic
Substances. Computers and Geosciences. 20:973-1023.
Other Relevant Papers
Allison, J.D., D.S. Brown, K.J. Novo-Gradac. March 1991. MINTEQA2/PRODEFA2, A
Geochemical Assessment Model for Environmental Systems: Version 3.0, Users Manual.
EPA/600/3-91/021, USEPA ERL ORD, Athens, GA.
2Articles listed as "Key Papers" are reprinted in this volume. Additional relevant papers
have been listed, but not reprinted.
-------
Bury, N.R., F. Galvez and C.M. Wood. 1998. Effects of chloride, calcium and dissolved organic
carbon on silver toxicity: Comparison between rainbow trout and fathead minnows. Environ.
Toxicol. Chem. 18:56-62.
Felmy, A.R., D.C. Girvin and E.A. Jenne, 1984. MINTEQ: A Computer Program for Calculating
Aqueous Geochemical Equilibria. USEPA Environmental Research Laboratory, Office of Research
and Development, Athens, Georgia.
Gorsuch, J. ans S. Klaine, Editors. 1999. Annual Review Issue: Silver Toxicity. Environ. Toxicol.
Chem. 18:1-108.
MacRae, R.K., D.E. Smith, N. Swoboda-Colberg, J.S. Meyer and H.L. Bergman. 1999. Copper
binding affinity of rainbow trout (Oncorhynchus mykiss) and brook trout (Salvelinusfontinalis) gills.
Environ. Toxicol. Chem. (in Press).
Playle, R.C., R.W. Gensemer and D.G. Dixon, 1992. Copper accumulation on gills of fathead
minnows: Influence of water hardness, complexation and pH on the gill micro-environment.
Environ. Toxicol Chem. 11:381-391.
Playle, R.C. 1999. Physiological and toxicological effects of metals at gills of freshwater fish.
Environ. Toxicol. Chem. (in Press).
Santore, R.C. and C.T. Driscoll. 1995. The CHESS Model for calculating chemical equilibria in
soils and solutions. Chemical Equilibrium and Reaction Models, SSSA Special Publication 42, The
Soil Society of America, American Society of Agronomy.
Schecher, W.D. and D.C. McAvoy. 1992. MINEQL+: A software environment for chemical
equilibrium modeling. Computer Environ. Urban Systems. 16:65-76.
Wood, C.M., R.C. Playle and C. Hogstrand. 1999. Physiology and modeling of the mechanisms
of silver uptake and toxicity in fish. Environ. Toxicol. Chem. 16:71-83.
-------
The importance of trace metal speciation
to water quality criteria
Herbert E. Allen, David J. Hansen
ABSTRACT: Because the bioavailability of a trace metal, and conse-
quently its toxicity, is dependent on the physical and chemical form of
the metal, we have presented a detailed assessment of how speciation of
copper would be expected to affect its toxicity. Principles of chemical
speciation are applied to demonstrate that inorganic forms will be in
constant proportion to each other and to free copper ion during the
course of the titration of a sample of natural water with copper or in the
various treatments in a toxicity test conducted at constant pH and al-
kalinity. Binding of copper to dissolved organic matter or to suspended
paniculate matter may render the copper nonbioavailable. We have
considered a simple complexation model to describe the complexation
of copper to soluble ligands. Naturally occurring dissolved organic matter
is present at concentrations only slightly greater than that of copper.
Consequently, titration of water with copper results in a nonlinear re-
lationship between the concentration of copper present as free copper
ion plus inorganic copper species. The effects of stability constant of the
complex, concentration of ligand, and the total copper concentration
are evaluated. We have related bioavailable copper to the concentration
of free copper ion plus inorganic copper complexes, which is valid if the
pH and alkalinity of the waters used to develop a criteria are not different
On the basis of limited field data for the complexation of copper in
Narragansctt Bay water, we do not expect that, significant differences in
water quality criteria (WQQ would result if the criteria were to be based
on free copper ion plus inorganic copper complexes rather than total
copper concentrations. We examined the effect of speciation of copper
in different waters as related to empirical or theoretically calculated water
effect ratios (WER). We show that, on the basis of sound chemical prin-
ciples, it would be expected that the. most sensitive organisms would
have the greatest WER. This prediction is confirmed by the empirical
observations available. For insensitive organisms, knowledge of the con-
centration of ligand is sufficient to reasonably predict the WER. However,
for the more sensitive organisms that give higher WERs, it is necessary
to measure or calculate the speciation of copper to predict the WER.
Use of predicted WERs may replace use of empirically derived WERs
as is now part of regulatory guidance for derivation of site-specific WQC,
if correspondence has been demonstrated. Water Environ. Res.. 68,42
(1996).
KEYWORDS: copper, criteria, metals, model, speciation, waterquality.
Bioavailability of a trace metal to aquatic organisms, and the
toxicity of the metal is dependent on the physical and chemical
forms of the metal (Luoma, 1983; O'Donnel el al., 1985). Benson
el al. (1994) have stressed the importance of understanding the
speciation of a metal in an aquatic system to the prediction and
interpretation of toxicity. Important factors to be considered in
the speciation of a metal include oxidation state, precipitation
and sorption, complexation, and the formation of organometallic
compounds. Kelly (1988) has written ". . . the speciation of a
metal, rather than its total concentration, is the key to under-
standing its effect on the biota." Allen (1993a,b) has reviewed
the principles of speciation with particular emphasis toward
complexation and has shown how speciation may be used in
water quality criteria development
A predominance of studies have demonstrated that organism
response is correlated with free metal ion. The effect of speciation
on the bioavailability of copper has been studied most extensively
(Hodson et al.. 1979). Studies of copper toxicity to fish have
demonstrated that toxicity is not related to the total copper con-
centration, but rather to the concentration of the free copper
ion (Pagenkopf et al.. 1974). The observed toxicity of copper to
rainbow trout was strongly correlated to the concentration of
free copper ion (Brown etal.,l 974). Howarth and Sprague (1978)
and Chakoufnakos et al. (1979) concluded that the toxicity
of copper was dependent on the concentration of the free cop-
per ion. •
Although it might be tempting to conclude that the free metal
ion alone is responsible for toxicities observed, this appears to
be not supportable. Although organically bound copper appears
to be nontoxic, there is some debate over the toxicity of inorganic
hydroxy and carbonate complexes. The most toxic inorganic
forms are Cu2* and CuOH*; however, Magnuson et al. (1979)
reported that [Cuj(OH JJ2* was also toxic in some cases. Cowan
et al. (1986) performed a statistical analysis of the copper toxicity
literature and concluded that hydroxide species, but not car-
bonate species, contribute to the aquatic toxicity of copper.
Meador (1991) studied the toxicity of ionic copper to Daphnia
magna in an experiment in which the total copper, pH, and
naturally derived dissolved organic carbon were allowed to vary.
The same toxicity was achieved with less ionic copper as the pH
increased, which he explained by a competition of H+ for Cu2*
at a receptor site on the cell surface.
Many studies have related toxicity to the concentration of
free metal ion. Judgment must be used in the evaluation of
studies conducted in constant or varying pH media. In constant
pH systems, the concentration of the free metal ion and the
monohydroxy complex (e.g., Cu2* and CuOH*) covary. In mul-
tiple tests, when the pH of the system is changed, species other
than the hydroxy and carbonate complexes are affected. Com-
plexation by naturally occurring organic matter will change in
a nonlinear fashion.
The concentration of organic ligands and the strength of their
binding of copper has been evaluated by a number of analytical
techniques (Neubecker and Allen, 1983). These chemical anal-
yses, together with computation of inorganic speciation, allow
computation of the free metal ion concentration.
Water quality criteria and standards have been based on total
recoverable metal or on soluble metal concentrations (U.S. EPA,
1992). Recently, the VS. EPA (Prothro, 1993) has recommended
the use of dissolved metal to set and measure compliance with
water quality standards, because "dissolved metal more closely
approximates the bioavailable fraction of metal in the water col-
umn than does total recoverable metal." Water quality criteria
(WQC) do not take soluble speciation of metal into account
Water Environment Research, Volume 68. Number 1
-------
«nen ana nansen
For example, for cadmium, copper, lead, nickel, and zinc, hard-
ness alone has been used to adjust fresh water WQC; no ad-
justment is used for saltwater WQC (U.S. EPA, 1986).
Acute sensitivities of saltwater animals to copper in laboratory
exposures in which an average of 83% of the copper was dissolved
are reported to range from 5.8 ng/L (9.1 X 10"* mol/L) for the
blue mussel to 600 pg/L (9.4 X 10~* mol/L) for the green crab
(U.S. EPA, 1985). A large fraction of the copper in natural water
exists in complexed and biologically unavailable forms. In these
natural waters, WQC derived from laboratory tests with lesser
complexation capacities may be overprotective.
The purpose of this paper is to demonstrate how changes in
bioavailable metal, within the soluble metal pool, will result from
changes in water quality, increasing concentration of metal added
to a water, and the dilution of an effluent The water effect ratio
(WER), in which the ratio of toxicity test results for a site water
compared with that for a reference water are used to correct the
criteria, has been advocated as a means to account for the effects
of metal speciation on a site specific basis (Stephan et al., 1985;
U.S. EPA, 1992; Prothro, 1993). We discuss how the addition
of metals to laboratory water versus natural surface waters will
affect speciation; thus the WER determined for different organ-
isms at a site. This discussion will emphasize why selection of
appropriately sensitive species or tests is of particular importance
if the WER is to actually account for metal's availability in site
water. From these considerations of chemical speciation, it is
clear that development of WQC based on bioavailable metal
would obviate the need for determining a WER.
For clarity of presentation, we have chosen not to consider
paniculate metal. However, a sorption model to account for
partitioning of metals to solids could have been included. The
WQC are based on soluble metals (Prothro, 1993). The sorption
of metal to solids removes the metals from the soluble pool and
reduces their toxicity.
Although we specifically discuss the effects of speciation on
the bioavailability of copper, the same principles apply to a
number of other cationic metals, including cadmium, chromium,
cobalt, lead, mercury, nickel, silver, and zinc. We have chosen
copper because the national water quality criteria in marine wa-
ters is very low, 2.9 fig/L, and approaches background. Copper
tends to form strong complexes with naturally occurring organic
matter, and there have been a number of measurements of the
effects of water chemistry on the speciation and toxicity of
copper.
Principles of Chemical Speciation
Metal ions, such as Cu2*, can form complexes with a number
of inorganic ligands, such as OH~, HCCV, NH,, and CT, and
with organic ligands, such as glycine (Stumm and Morgan, 1981;
Pankow, 1991). Because the stability constants for these reactions
are well known, the distribution of species can be easily com-
puted. The total concentration of soluble metal in the system is
the sum of the free metal ion and of the metal contained in the
complexes. Stability constants for the reaction of metals with
naturally occurring organic matter, such as humic acid, are be-
ginning to be developed, allowing the distribution of species in
a system containing naturally occurring organic matter to be
predicted.
Typical concentrations of inorganic ligands such as chloride
are in the millimolar range. Thus, they tend to be in large stoi-
chiometric excess relative to the micromolar or lower concen-
tration of trace metals such as copper. The concentrations of
hydroxide and carbonate are small, but they remain approxi-
mately constant because of the buffering of the system. Conse-
quently, the addition of copper to a surface water sample for a
titration or in a toxicity test decreases the inorganic ligand con-
centration by a small and insignificant amount This results in
a constant ratio between free copper ion and the total concen-
tration of inorganic copper species. The principal soluble inor-
ganic complexes of copper in surface water are expected to be
CuOH*. Cu(OH)j0, CuCO,°, and Cu(CO,),2-. The reactions foi
the formation of these complexes and their stability constants
(Stumm and Morgan, 1981) are as follows:
Cu2* + OH' ** CuOH*; KOH.I = 10" (1)
Cu2* + 20H- *T Cu(OH)20; KOHJ - 10I2J (2)
Cu2* + CO,2" n CuCO,°; KCO..I » 10s" (3)
Cu2* + 2CO,2- ** Cu(CO,)22-; KCQ..I = 10IW" . (4)
The fraction of free copper ion, Cu2*, in equilibrium with
free copper ion plus all inorganic copper species, is ac»»* which
can be calculated from the concentrations of hydroxide ion,
[OH~], and carbonate, [COj2'], using the expression (Stumm
and Morgan, 1981):
1
1 + KOH..IOH-] +
+ Kco,.,[CO,2-] +
(5)
The concentration of carbonate ions can be calculated from
the total alkalinity, [tot-alk], and the pH or from the concen-
tration of dissolved inorganic carbon, CT, and pH. The rela-
tionship between total alkalinity and pH is (Stumm and Morgan,
1981)
[tot-alk] •= (a, + 2a2yCr + [OH'] - [H*] (6)
where a{ and a2 are the fractions of CT present as bicarbonate
and carbonate, respectively, where at and o2 are given by the
expressions
and
«i
«2
K..IH*]
[H*]2 + K.,[H*J -I-
[H*]2 + K.,[H*]
(7)
(8)
K.I and K^ in Equations 7 and 8 are the first and second acid
dissociation constants for carbonic acid (Stumm and Morgan,
1981) as represented by the following expressions
HjCO, ** H* + HCO,-; K., - IfT" (9)
and
HCOr ** H* + COj2-;
10
'""
(10)
For a typical set of conditions, pH •= 7.8 and total alkalinity
» 2 X 1(T3 equiv/L (100 mg/L as CaCOj), at = 0.966, a2
= 0.00306, and CT = 2.057 X 10~J mol/L. The concentration
of carbonate is
ICO,2-] = «2CT = (0.00306X2.057 X 10"J)
= 6.29 X 10-* mol/liter (11)
January/February 1996
43
-------
Alien and Hansen
Equation 5 then becomes
+ (10677X10-"°)+(10iaoiX10-s-20)2
or 2.4% of the free copper ion plus inorganic copper complexes
is free copper ion and the predominant inorganic complex is
CuCOj0. It should be noted that for a fixed pH and total alka-
linity, the fraction of the free copper ion plus inorganic copper
species that will be present as the free copper ion (acu>*) is un-
affected by the total copper concentration or the concentration
of copper present in organic complexes or adsorbed onto par-
ticulate matter. Likewise, the fraction of the copper that will be
present in any of the inorganic complexes, e.g., CuCOj0, will
also remain constant as the total concentration of copper varies.
The concentration of copper binding sites present in organic
ligands extends from submicroequivalents per liter concentra-
tions in sea water to tens of microcquivalents per liter in rivers.
The organic ligands in natural waters are composed of humic
substances and other compounds that usually are not specifically
determined. The ligand concentration is commonly considered
to be related to organic matter or organic carbon by an unspec-
ified and variable factor. Typically, the concentration of organic
ligands is in excess of the concentration of copper naturally pres-
ent in a sample. However, addition of metal for a titration or
to establish certain test concentrations to be used in toxicity
testing, will result in a stoichiometric excess of metal relative to
the concentration of the organic ligand. As will be shown, as
opposed to the inorganic complexation discussed above, the
fraction of the copper that will be complexed with the organic
matter will vary as a function of the concentration of copper
present in the system.
Naturally occurring organic matter contains a large number
of ligands differing in concentration and stability constants. A
number of different approaches have been taken to describe their
binding of protons and metal ions (Perdue and Lytle, 1983; Fish
et al.. 1986; Ephraim and Marinsky, 1986; Cabaniss and Shu-
man, 1 988; Tipping el al, 1990). These approaches include dis-
crete site models, models with a continuum of binding sites of
varying pK, and models that incorporate electrostatic interac-
tions.
Multisite binding models provide a better description of copper
titration data than does a single site model. Because of its sim-
plicity, a single site model was used for the simulations presented
in this paper. A single ligand model adequately illustrates the
effect of changing the ratio of copper to ligand on the inorganic
copper concentration, although it will not adequately charac-
terize metal binding to natural organic matter.
1 The complexation of copper by an organic ligand, L, can be
represented by the expression
Cu2* + L°- *»
K'<
[Cu2+][L-]
(13)
where K' is the thermodynamic formation constant for the for-
mation of the copper complex. Studies of copper binding in
natural waters often use electrochemical methods, such as anodic
stripping voltammetry or fixed-potential amperometry, for the
analysis of copper. These techniques respond to inorganic, but
not organic complexes, of copper in natural waters. Conse-
quently, the reaction in Equation 13 can be rewritten in terms
of free copper ion plus inorganic copper complexes, Cu,^. Be-
cause the charge on the organic ligand is generally not known,
but is constant in a constant pH system, Equation 14 will be
written without charges on the chemical species:
+ L tf CuL; K •
[CuL]
(14)
where K is the conditional formation constant for the formation
of CuL. The value of K is dependent on pH, ionic strength, and
on the concentrations of ions that can influence the above re-
action through participating in side reactions with either copper
or the organic ligand, for example, Zn2* and COj2". The sim-
ulations presented in this paper are based on Equation 14 in
which complexation is described by free copper ion plus inor-
ganic copper complexes.
The total copper concentration, [Cu]T, is given by the mass
balance on copper
[Cu]T - [CuL] +
(15)
Likewise, the total ligand concentration, [L]T, is given by the
mass balance on ligand:
[L]T=[CuL]
(16)
where the total ligand concentration is frequently expressed as
the analytically determined complexation capacity. Equations
15 and 16 can be solved for [CullwJ and [L], respectively. These
values substituted into the equilibrium constant expression in
Equation 14 yield the expression
K
[CuL]
«Cu)r-[CuL]X[L]T-[CuL])
(17)
for the average stability constant over the range of copper ion
added during the titration. The value of (Cu]T is the sum of thr
concentration of copper in the initial water sample plus the con-
centration of copper added during the titration. The values of
K and [L]T are determined by evaluation of the titration curve
using the method of Ruzic (1982).
Effect of Complexation Parameters on Copper
Speclation
The concentration of inorganic copper in a sample is a funr
tion of the concentration of total copper, total ligand, and in-
stability constant for the reaction. This is shown by substitution
of Equations IS and 16 into Equation 14:
(18)
The concentration of free copper ion plus inorganic copper
complexes is given solving this quadratic equation. The fraction
of the copper present as free copper ion plus inorganic copper
complexes, oo^,, is
(19)
+ [CuL]
Literature values for ligand concentrations and conditional
stability constants, as defined by Equation 14, for natural waters
are given in Table 1. Most of these values are for relatively un-
contaminated seawater. On the basis of these values, we have
used stability constants ranging from 1.0 X 107 to 1.0 X 109 and
ligand concentrations ranging from UOO X 10~7 to 3.50 X 10~*
44
Water Environment Research, Volume 68. Number 1
-------
Allen and Hansen
Table 1—LJgand concentrations and conditional stability constants for the complexation of copper In natural waters.
Ugand concentration
Study area
Adriatic Sea
North Atlantic
North Atlantic
Southern Calif omia
Delaware Bay
Narragansett Bay
New York Harbor
North East Pacific
North Sea
Lake Ontario
Scheldt Estuary
Vancouver Island
logK
7.5
8.3 to 10.0
7.8 to 8.6
7.8 to 8 6
7.6 to 62
73
6.8 to 8.4
8.5
7.3 to 8.4
8.6
7.2 to 7.7
6 6 to 7.8
mol/L
13I01.5X10-7
023 to 4.9 X 10-'
2.0 107.2X10-*
03 to 9.3X10'7
1 .Sfitol. 62X10-*
158X10'7
3.46 to 5.49 X 10*'
7.6 X 10-*
0.7 to 2.0 X 10"'
34X10-'
1.08 to 2.99 X10'7
• 1.0 to 3.0X10-'
jig Cu binding
capadty/L
856 to 9.53
1.46 to 31.1
1.27 to 4.58
1 9 to 59.1
0.99 to 1.03
877
22.0 to 34.9
0.48
4 44 to 12.7
21.6
6 86 to 19.0
6 35 to 19.1
Reference
Plavsic et el (1982)
Buckley and van den Berg (1986)
Kramer (1986)
Smaefa/ (1980)
Skrabal et al (1992)
Skrabal and Allen (1993)
Skrabal and Allen (1993)
Coale and Bruland (1988)
Kramer and Duinker (1984)
Florence (1986)
Kramer and Duinker (1984)
Robinson and Brown (1991)
moI/L (6.35 to 222.4 ng copper binding capacity per liter) to
conduct simulations of copper complexation. Total copper con-
centrations were 2.90, 29.0, and 145.0 ftg/L (4.56 X 10'*, 4.56
X I0~7, and 2.28 X 10~* mol/L), which correspond to the na-
tional water quality criteria for copper in marine water, and 10
and 50 times that value.
The effect of ligand concentration on the binding of copper
for a stability constant of 2.0 X 10s is shown in Figure 1. Each
of the simulations corresponds to a dilution of a receiving water
at constant pH and alkalinity and then addition of copper to
give the indicated total copper concentration. We have not con-
sidered precipitation of Cu(OH)j (s) or CuO (s) in these or the
other simulations presented in this paper; at concentrations
greater than that leading to formation of a solid phase at a par-
ticular pH, the system will be metastable with respect to the
formation of precipitated Cu(OH)2 (s) or CuO (s). As the ligand
concentration decreases from 3.50 X 10"* (222.4 /tg copper
binding capacity per liter), the concentration of free copper ion
plus inorganic copper complexes increases (Figure la). The con-
centration of free copper plus inorganic complexes increases as
the concentration of ligand is decreased. As the concentration
of ligand is decreased to a concentration similar to that of the
total copper present, the concentration of free plus inorganic
copper dramatically increases. Consequently, the curves for the
different total copper concentrations do not parallel each other.
The fraction of copper that is free copper ton plus inorganic
copper species depends not only on the ligand concentration,
but also on the concentration of the total copper present in the
sample (Figure 1 b). The remaining copper is present as the com-
plex with the organic ligand. For example, at a total ligand con-
centration of 5.00 X 10~7 mol/L (31.8 pg copper binding capacity
per liter), the fraction of organically complexed copper decreases
from approximately 99% for a total copper concentration of
4.56 X 10"' mol/L (2.90 jtg/L) to approximately 94% for a total
copper concentration of 4.56 X 10~7 mol/L (29.0 pg/L), to just
over 22% for a total copper concentration of 2.28 X 10"* mol/
L (145.0 jig/L).
The dependency of copper speciation on the stability constant
of a ligand present at a concentration of 5.00 X 10~7 mol/L (31.8
pg copper binding capacity per liter) is shown in Figure 2. The
highest concentration of copper, 2.28 X 10~* mol/L (145.0 jig/
10* -I
10*-
10-'-
10*-
10*-
10'*-
m-"
<
i
X
*
\
""
<
1
,
1
f*""--
""*"*-. -^ _^ _
....!....
3 P
o «-
....)....
- (1)
:
'_jm
.--' 0)
i
, t ;
s i i
• 1 1 i
a 1
P- 1
s ;
• 1 1 •
-_j
D 1
^
3
O 1
"635 g
'633 §•
8
Q.-T.
0635 OB)
•00635 ™
'0 0063S
rO 000635,2
—
n
•i
Ugand (moVL)
5 S 2 2 2
n 8 S M »
Ugand (jig Cu binding capadty/L)
1.000
0.100
0.010
X
(3)
(2)
(b)
(1)
0.001
° 5 2 2 i i I i
Ugand (no Cu binding capadty/L)
Figure 1—Free plus inorganic copper complexes as a
function of ligand concentration for a stability constant,
of 2.0 X 10* and total copper concentrations of (1) 145.0
/tg/L (2.28 X 10~* mol/L), (2) 29.0 pg/L (4.56 X 10~r mol/
L), and (3) 2.00 ng/L (4.56 X 10~* mol/L). Free plus Inor-
ganic copper (a) concentration and (b) fraction of total
copper.
January/February 1996
45
-------
Allen and Hansen
ID4-
10-T
10*-
<
(
*•-._
\
4
3
3 <
•,-T-T-
b '•
F- 1
». «
— i — i — i — |
(1)
_ _
. . ._j
3 1
— *
> «
(2)
(3>
T T- I-
b
•• *
3 1
.
i
-,-,, , :
b
P- <
-635
.fi 15
-0635
~0 0635
>
F»
».
1.000-
$5 o.ioo-
o fr
,J|.H 0.010-
J
0001-
i
i\
A
\
\
\
(b)
* * » . ^
*"--.._
"""
(D
~-L3>.
(2)
•-•- .-.
— ~:
o
CM
S
Rgure 2—Free plus inorganic copper complexes as a
function of stability constant for a ligand concentration
of 5.00 X 10~T mol/L and for total copper concentrations
of (1) 145.0 pg/L (2.28 X 10~* mol/L), (2) 29.0 pg/L (4.56
X 10~T mol/L), and (3) 2.90 ng/L (4.56 X10"" mol/L). Free
plus inorganic copper (a) concentration and (b) fraction
of total copper.
L), is a little more than 4 times that of the ligand concentration
and the concentration office copper ion plus inorganic copper
complexes is almost invariant with respect to changes in the
stability constant from 1.0 X 10T to 1.0 X 10'. At a total copper
concentration equal to or less than the concentration of the li-
gand, the concentration of free copper ion plus inorganic copper
species (Figure 2a) and the fraction of the copper that is free
copper ion plus inorganic copper species (Figure 2b) decrease
with increasing values of the stability constant
An environmentally important question regards the dilution
of a sample containing copper and a ligand. This is the situation
when a waste is discharged to the environment and dilution
occurs within a mixing zone, where the dilution water contains
no copper or ligand. We have considered a waste effluent with
a total copper concentration of 2.50 X 10~7 mol/L (15.9 /ig/L)
and a ligand having a stability constant of 2.0 X 10s (Figure 3).
In one case, the ligand concentration in the effluent is equal to
that of the copper. In the second case, the concentration of the
ligand in the effluent is 5 times that of the copper. The latter
case would be that expected for a waste effluent from a biological
wastewater treatment facility. In the case of the effluent with
equimolar ligand and copper, the concentration of free copper
ion plus inorganic copper complexes is greater than that which
would be expected from dilution alone (Figure 3a). This is a
result of the effect of dilution on the dissociation of the copper
complex. When the effluent containing the equimolar metal and
ligand is diluted to one-half its initial concentration, the percent
of free copper ion plus inorganic copper complexes in the effluent
will increase from 13% to 18% (Figure 3b). For the sample that
contained a fivefold excess of ligand relative to copper, the cor-
responding increase in inorganic copper would be from 0.5% to
1.0%, resulting in the free copper ion plus inorganic copper
complexes doubling during the dilution. However, it should be
noted that the concentration of free copper ion plus inorganic
copper complexes in this case would be very low.
Typical titration curves for copper additions to a ligand are
shown in Figure 4a. This figure considers the titration of the
same ligand at four different concentrations ranging from 1.00
X 10'7 to 2.00 X 10~* mol/L (6.35 to 127.1 fig copper binding
capacity per liter). It would appear that these curves differ only
in the position of the inflection point The inorganic copper
concentration has been plotted on a logarithmic scale in Figure
4b. This clearly shows the large differences in the concentration
of free copper ion plus inorganic copper complexes at a constant
small total copper concentration. It is obvious that titrations will
I
I
OS
g 4010*-j
•§. 3^10*-
o 3010*-
S. 2510*-
Oo 2010*-
c*^ 1^ 10*-
o 10 10*-
+ 6010*-
H
£ °-o H
; -*.
'*•->
(2)
•^
.
(a)
0)
' *'' ' 1
V
\
\
,
'2^4
l-o.oo
o
c
+
100 80 60 40 20
Effluent (%)
60-
— — —
. - —
(b)
•^ ^
^»
(2)
(
/
/
:
;
1 <
100 60 60 40 20 0
Effluent (%)
Rgure 3—Free plus inorganic copper complexes as a
function of dilution of effluent containing 2.50 X 10~7 mol/
L (15.9 pg/L) total copper. Stability constant for copper
complexation is 2.0 X 10* and ligand concentration In the
effluent Is (1) 2.50 X 10~T mol/L (15.9 pg copper binding
capacity/L) and (2) 1.25 X 10~* mol/L (79.4 nQ copper
binding capacity per liter). Free plus Inorganic copper:
(a) concentration and (b) percent of total copper.
46
Water Environment Research. Volume 68. Number 1
-------
Alien and Hansen
1 M1°*
|- 3.0 10*-
o 1.0 10*-
1 00
u. 0.0 H
<
i
X
a c
0) X
„ *
3 <
(a)
X
,-rx
(21'
.-'
, 1
(3).-
r * r"M
3 1
L P — '
X
.''
X
, 1
j
/
f *
, '
D 1
X
x-
• 190.6 g- m
9 Q. A
CL C O
h O«i •* o 3L
|" 06
_ ... p> Q.
•"f 8-
- 01 a 4. O A
o o
nn ff 7s
1 1 "" "• a
2 S g> o
aqqmqnoiq Ofc
•t
. "" ,•
' 0).
'[
m ,
(4^
__
'
t
t t
^
(b)
, t
-~-^.-
'
, ?
=;-.-»•-:'
, \
oas 5;
t
-R QC S
O C)
4-
i
ft
q«-^^^^^»-
o o o 10 o w o
1O r^ »-' CM CM CO
Total Copper (mol/L)
Total Copper (ng/L)
t»n-»-
s
i.-^ 10*-
!>
? in-«-
£ 10-
+
£ to1 H
(D/
/
/ \
« * ^
/'
, ' -*
X"
x
,^ —
, 1
r*" * i * ' ' '
--r'"
«..--
";J\'~-
i
> t. -t
v . * '
— --
(e)
, %
(3) .
I4i-
, -J
..— !
....!
- :
, t
-0.635
>
OOOOOOOO
»- CM n » «> o> r»
Total Copper (mol/L)
q g ,»,.» cq ^ n
Total Copper (jig/L)
Figure 4—Free plus Inorganic copper complexes as a
function of total copper concentration for ligand with sta-
bility constant for copper complexation of 1.0 X10* and
ligand concentrations of (1) 1.00 X 10~T mol/L (6.35 (tg
Cu binding capacity per liter), (2) 5.00 X 10~T mol/L (31.8
jig Cu binding capacity per liter), (3) 1.00 X 1Q-* mol/L
(63.5 /ig Cu binding capacity per liter), and (4) 2.00 X10~*
0.0 5.010'7 1.010-6 1.510"6 2.010"6 2.5
Total Copper (mol/L)
63.5 95.3 127.1 158.9
31.8
Total Copper (|ig/L)
Rgure 5 — Free plus inorganic copper complexes as a
percentage of total copper concentration as a function
of total copper concentration for a ligand concentration
of 2.00 X 10~* mol/L (127.1 fig Cu binding capacity per
liter) and stability constant of 1.0 X 10'. ,
have different free copper ion plus inorganic copper complexes
(and consequently free copper) concentrations at all total copper
concentrations if the waters being titrated differ in ligand con-
centration. The fraction of copper that is present in the free
copper ion plus inorganic copper species does vary if the total
copper concentrations is less than the concentration of the ligand
present in the system (Figure 5). Thus, for small additions of
copper to a sample, the increase of the concentration of copper
present as free copper ion plus inorganic copper complexes is
not proportional to the increase in the concentration of total
copper present. Of course, for copper additions in excess of the
concentration of ligand present, the increase in free copper ion
plus inorganic copper complexes begins to more closely mirror
the increase in total copper.
In a like manner, two waters having the same total concen-
tration ofligands, but with different equilibrium constants, will
have different concentrations of free copper ion plus inorganic
copper complexes at all total copper concentrations. Conse-
quently, neither free copper ion concentration nor the concen-
tration of free copper ion plus inorganic copper complexes is
directly proportional to the total copper concentration at different
concentrations of total copper during the titration of a water, at
the same concentration of total copper for samples of a water
that have been diluted to different extents, or for different water
samples containing the same total concentration of copper.
The speciation of copper is highly dependent on the pH of
the environment In addition to the effect that pH has on the
distribution of inorganic copper species, complexation with or-
mol/L (127.1 ng Cu binding capacity per liter). Free plus
Inorganic copper (a) linear axis, (b) exponential axis, and
(c) exponential axis and low concentrations of total cop-
per.
January/February 1996
47
-------
Allen and Hansen
ganic ligands is affected by changes in the pH of the system.
Protons compete with copper ions for binding sites on carboxylic
acids and other organic compounds. Titration curves for the
addition of copper to a secondary sewage effluent at pH values
ranging from 6.5 to 8.5 are presented in Figure 6a. Complexation
parameters for the wastewater effluent are given in Table 2. As
the pH rises, both the stability constant and the complexation
capacity increase. As the hydrogen ion concentration increases
by one order of magnitude (decrease of 1 pH unit), the concen-
tration of free plus inorganic copper complexes can increase by
more than an order of magnitude (Figure 6b).
Caution must be exercised in the use of experimentally de-
termined complexation capacities to express ligand concentra-
tions. Perdue (1988) has shown that the complexation capacity,
evaluated by the change in slope of the titration curve, is a func-
tion of the stability constant of the ligand in addition to the
concentration of the ligand.
8.0 10"-
7.0 10 -
6.0 10'*-
S.O 10*-
4.0 10"*-
3.0 10*-
-
— d
—
x.4'''
pHB
....
«i
|>H7J5
pHTO
pH8,5
,f
,'
'
,'
^
r~
/
/
t
r
,
'
.
y
r
/
t
t
X
t
s
•.".'.J....I
"DUO.*
•444 B
• 3177
- 954 9
- 190 6
• 1271
-0.0
o q
T-' c\i
o o o q q o q
V to *-
Total Copper (mol/L)
i I i i i i
Total Copper (ug/L*)
5 3
&> Q
iu —
4A-*.
in-7-
1H-*.
/f".
1
: x
;/
r> 1
'"'''
X
X*
3 T
x
3 1
J-..
X
(b)
3 1
- -
..-•
3 1
'"'<""
1
^
3 T
-
-
3 1
TJ
X
-:
•i'.—"
3H6
3H7
\
• --pH8jO
-ft
3 t t 1
|
.« nc &
°-oa S'^
-0635 •B'3
3
q
IO
qqq
«OO>»-
Total Copper (mol/L)
q « •- •+ a> n
° S Si g 5 t 5 5 o E S
"7-CMnn * m E <6
Total Copper (jig/L)
Rgure 6 — Concentration of free plus inorganic copper
complexes in a secondary sewage effluent titrated at dif-
ferent values of pH. Free plus inorganic copper (a) linear
axis and (b) exponential axis. Complexation parameters
are given in Table 2. Precipitation of CufOH)? (s) or CuO
(s) are not considered In these calculations.
Table 2—Parameters for copper complexation by sec-
ondary wastewater effluent (Grosser, 1980). Titration
simulations are shown in Figure 6.
Ligand concentration
PH
6.5
7.0
7.5
8.0
8.5
Stability constant
5.38 X 10s
1.61 X 10"
1.82 X 10s
1.15 X107
605X10'
mol/L
1.26 X 10"«
2.28 X 10-*
2.31 X 10"*
3.67 X ID"6
7.30 X 10-«
ng Cu binding
eapacity/L
80.1
144^
1468
233.2
465.8
Effect of Copper Speculation on Biological Response
Toxicity tests are conducted by the addition of several different
concentrations of metal to the test water to expose organisms
in the water for a fixed period of time after which the biological
response, frequently death, is evaluated. As indicated earlier, the
organic complexes of copper are not biologically active. It is
clear that to relate the biological response to the concentration
of copper present in a water it is necessary to develop a rela-
tionship between total copper and the concentration of biolog-
ically active copper. In the case of a bioassay, the concentrations
of free copper ion and of each of the inorganic copper complexes
are proportional to the concentration of free copper ion plus
inorganic copper complexes if all test chambers have the same
pH and alkalinity. Consequently, it is important to evaluate the
change in free copper ion plus inorganic copper species fhat
results from the addition of copper to a test water. Such
relationships are the titration curves that have been showr* In
Figure 4.
We have used this basic approach to evaluate the effect at
consideration of chemical speciation would have on water qu;- '>'>
criteria. We based our calculations on the results of a titras :-n
of a sample of water that was collected in Narragansett Bay i?t
the Winter of 1993. For that sample, we determined a stab' ''• •
constant of 727 X 107 and a ligand concentration of 1.38 X ,> '
(8.77 jig Cu binding capacity per liter) using anodic strips ^>i
results reported in the U.S. EPA's criteria document for coppu
(1985) were determined in Narragansett Bay water having th-,se
complexation properties. The four biological species and ihe
genus mean acute values that were used to establish the criterion
are soft-shell clam (39.00 pg/L), oyster (14.92 jtg/L), summer
flounder (13.93 pg/L), and blue mussel (5.80 pg/L).
We computed the concentration of free copper ion plus in-
organic copper species that would have been present at the genus
mean acute values for the four most sensitive genera in salt
water in the water quality criteria document We assumed that
the complexation parameters of the water used for the bioassay
tests for each of the four species were those for the Narragansett
Bay sample that was collected in the winter of 1993. Although
the complexation of copper is small, it is an important factor at
low copper concentrations. Using these complexation parame-
ters, the fraction of copper that is not bound to organic matter
at a total copper concentration of 9.1 X 10"' mol/L (5.8 pg/L),
which is the reported acute toxicity for the blue mussel, is
only 17.9%.
48
Water Environment Research, Volume 68. Number 1
-------
Mien andHansen
We have followed the procedure to establish the WQC for
copper in marine waters (Stephan et al, 1985) and, using the
computed concentrations of free copper ion plus inorganic cop-
per species, we have also computed a WQC for copper based
on bioavailable copper. The acute toxicities for each of the four
species are plotted versus both the reported total copper con-
centrations and the computed concentrations of free copper ion
plus that of the inorganic copper complexes (Figure 7). Using
the reported total copper concentrations, the final acute value
for copper in salt water is 5.832 jig/L (9.178 X 10"1 mol/L).
Because larval molluscs are most sensitive to copper, the criteria
maximum concentration (CMC) and criteria continuous con-
centration (CCQ are both derived by dividing the final acute
value by 2.0 and thus are the same 2.916 j*g/L (4.589 X 10~*
mol/L) concentration. Using the chemical speciation for the
Narragansett Bay water and assuming that all of the copper in
the toxicity tests was soluble, the final acute value based on
bioavailable copper would be 1.857 X 10~*mol/L(1.181 pg/L).
This corresponds to a back computed total copper concentration
of 9.786 X 10~* mol/L (6.219 jig/L). Using this value the criterion
maximum concentration would be (9.786 X 10"* mol/L)/2 =
4.893 X 10~* mol/L (3.110 /*g/L), which for copper would also
be the final chronic value. This analysis indicates that the current
water quality criterion for copper in salt water based on total
copper is virtually the same as would be a criterion based on
the concentration of free copper ion plus inorganic copper com-
plexes. If the computed criterion based on free plus inorganic
copper species is more reflective of the potential for copper tox-
icity than is the criterion based on total copper, the total copper
criterion is too conservative by a factor of 3.110/2.916 •= 1.07
for the intended degree of aquatic life protection. There is little
difference between the criterion based on total copper and the
criterion that has been computed for free copper ion plus in-
organic copper complexes.
Although there was little difference in the criteria that we
computed for copper based on total copper or on free copper
ion plus inorganic complexes, the difference may be greater if
each investigator had determined ligand concentrations and sta-
bility constants. We assumed that the same ligand concentration
100.0
3" 10.0-r
8) 6.210
.3 5.832
t.ieo
0.1
+
+
- e- - Total Copper (ug/L)
- -o - - Inorganic Copper (jigfl.)
+
5 10 15 20
Percentage Rank of Species
25
Figure 7—Determination of final acute value for copper
in marine water based on total copper and inorganic (free
plus inorganic complexes) copper. See text for discus-
sion.
and stability constant forcomplexation could be applied to each
of the toxicity bioassays. If we could have used the correct copper
speciation for each of the bioassays, the difference between the
copper criteria based on total copper and the criteria based on
free copper ion plus inorganic copper complexes could have
been greater or less than the differences found for our simulation.
t
Implication of Metal Speciation to Waste Load
Allocation and Discharge Permits
It is necessary to mathematically relate the concentration of
pollutant in a discharge to that in the receiving water environ-
ment at the point of required compliance with a water quality
standard to establish an acceptable concentration of the pollutant
that may be discharged. In the case of multiple discharges, the
waste load allocation process is used to establish the concentra-
tion of pollutant that can be permitted to be discharged from
each of the discharges. If the pollutant is a conservative substance,
only appropriate dilution computations based on the hydrody-
namics of the system are needed. If the pollutant is nonconser-
vative, for instance, through degradation, then the change in the
amount of material must also be taken into account. Metals
such as copper present a special case of a nonconservative pol-
lutant in that the form of the metal may change.
Present EPA guidance (Prothro, 1993) recommends that dis-
solved metals be used for water quality standards but that total
recoverable metals be regulated in effluents. This paper has dis-
cussed that only a portion of the dissolved metal is biologically
available. Irrespective of whether water quality standards are
based on dissolved metal or on biologically available metal in
the receiving water, there is a need to relate the concentration
of total recoverable metal in the effluent to the concentration
of dissolved or bioavailable metal in the receiving water at the
point of determination of compliance. This requires consider-
ation of a two-step dynamic process to account for the redistri-
bution of metal forms below the point of effluent discharge.
The total recoverable metal in an effluent can be made up of
paniculate forms including sorbed metal and metal incorporated
into the matrix of the particle, inorganic and organic complexes,
and free metal ions. Upon mixing of the effluent with the re-
ceiving water, the chemical environment will be different from
that of the effluent, except perhaps in an effluent-dominated
stream. Important changes would include change in pH and the
type and concentration of particles. In the first step of the re-
distribution process we could envisage the release of metal from
complexes ordesorption of metals from particles in the effluent
This would be likely if the pH of the receiving water was lower
than that of the effluent The rates of release of bound metal
will vary depending on the type of binding site. Concurrent with
this release, dissolved and free metal ion concentrations may be
lowered by reaction with ligands and paniculate matter in the
receiving water. These reactions of metal can be slow (hours to
days) (Allen et al.. 1982), and equilibration of released metal
ions with organic ligands may not be complete within the mixing
zone. Obviously, receiving waters with high concentrations of
organic matter and particles will reduce the concentration of
dissolved and bioavailable metal being released from bound
forms in the effluent
In contrast, in some environments the concentration of dis-
solved and of bioavailable metal will increase below the point
of effluent discharge, whereas in others it will decrease. The fac-
tors that would tend to increase dissolved metal below the point
January/February 1996
49
-------
Alton and Hansen
of effluent discharge include pH of the receiving water less than
that of the effluent, high concentration of soluble or of paniculate
metal in the discharge, and low concentration of organic matter
and of particles in the receiving water.
The above discussion provides a conceptual framework in
which to develop the data necessary to include metal speciation
in waste load allocation and discharge permits.
Effect of Copper Speciation on Water-Effect Ratio
Determination of a WER has been suggested as a means to
account for the effects of differences in chemical speciation on
the toxicity of copper and other metals in natural waters (U.S.
EPA, 1992;Trothro, 1993). In this procedure, toxicity tests are
conducted in a site water and in a reference (laboratory) water.
Reference water tests are used as surrogates for the laboratory
tests that were used to derive national criteria. The ratio of the
toxicities is used as a multiplier to adjust the national water
quality criteria (NWQQ to account for differences in bioavail-
ability, as measured by toxicity tests, that would be applicable
to that site. For example:
Site - Specific WQC = NWQC X WER
site - water LC50
NWQCX
reference - water LC50
(20)
Brungs el al (1992) have summarized the results of a number
of determinations of water effect ratios for different toxicants.
Their reported water effect ratios for copper are presented in
Table 3. It should be noted that, for a single location, the WER
usually increases as the organism sensitivity to copper increases.
This observation is consistent with the changes in metal specia-
tion as the concentration of total metal increases that we dis-
cussed earlier, and it must be considered in designing WER
studies.
To further illustrate the importance of species sensitivity to
the WER, we calculated WERs using acute values from five
species having different sensitivities to copper and copper spe-
ciation for hypothetical reference and site waters. The acute val-
ues used in this calculation were 1S9.8 pg/L for mysid, 120 jig/
L for polychaete, 39.97 0g/L for a copepod, 14.92 /*g/L for the
oyster, and 5.8 pg/L for the mussel, as obtained from the criteria
document Further, Narragansett Bay water with a stability con-
stant of 7.27 X 107 and a ligand concentration of 1.38 X I0~7
mol/L (8.77 fig Cu binding capacity per liter) was assumed to
have been the water for which these acute toxicity values were
determined. We then computed the concentrations of free copper
ion plus inorganic copper complexes that would have been ap-
propriate to each species-specific acute value if laboratory tests
would have been conducted in Narragansett Bay water (Table
4). Using these concentrations of free plus inorganic copper,
total copper concentrations required to produce these acute val-
ues were calculated for each species and the hypothetical labo-
ratory reference water and the four hypothetical site waters, each
with its own stability constant and complexation capacity (Table
5). The procedure used to compute the acute toxicity values in
the hypothetical laboratory reference and the site waters is shown
in Figure 8.
The toxicity of copper to each of the five organisms selected
is approximately the same in both the Narragansett Bay water
and in the hypothetical reference site water (Figure 9a). In the
four sites being evaluated, less total copper is needed to produce
50% mortality (the LC50 concentration) than is needed for the
Reference Site water even though the same concentration of free
copper ion plus inorganic copper complexes is needed for a given
species. In all cases the water effect ratio increases with the sen-
sitivity of the organism, expressed as the total copper concen-
tration causing acute toxicity (Figure 9b). The water effect ratios
for Site 2 are greater than those for Site 1 and the water effect
ratios for Site 4 are greater than are those for Site 3.
It should be noted that for the three least sensitive species
(mysid, polychaete and copepod) the copper LC50 values for
Sites 1 and 3 and for Sites 2 and 4 are almost identical (Figure
9a). Likewise, the WER values for Sites 1 and 3 and for Sites 2
and 4 are virtually identical (Figure 9b). The LC50 total copper
concentration required to be present in these four waters at the
LC50 value for these three organisms is greater than the con-
centration of organic ligand available for complexation of the
added copper. In this situation the toxicity of metal is primarily
controlled by the excess of metal relative to the ligand present
and is only secondarily related to the stability constant of the
Table 3—Water effect ratio study results for copper.
SH0
• From Brungs et al.. 1992.
"FromThursby. 1993.
Specie*
Laboratory water acute value, ng Cu/L
Water ertect ratio
Naugatuck River, Conn.*
St Lous River, Minn.*
Nemadp River, Wis.*
Little Pokegama River, Wts.*
New York Harbor, N.Y."
Ceriodaphnia dubla
Pimephales promotes
Daphnia magna
Amphipod
Pimephales prome/as
Onchoehynchus mytoss
Amphipod
Amphipod
Champia patwla
Mytilusedulis
Arbacia punctulata
Mulinia lateraTis
Mysidopsis bahia
17
85
17
25
84
120
90
90
11
26
49
35
308
1.1
1.0
15
92
< 6.3
3.3
4.3
3.1
32
1.8
1.8
1.6
1.5
50
Water Environment Research, Volume 68. Number 1
-------
Allen and Hansen
Table 4—Toxlcity of copper to organisms considered in simulated water effects ratio study.
Genus mean acute value
Total copper*
Organism
mol/L
Free plus Inorganic copper
complexes*
mol/L
Mysid
Pdychaete
Copepod
Oyster
Mussel
159.8
120
39.97
. 14.92
5.8
2.51 X 10-«
1. 89X10-"
6.29 X 10-'
2.37 X ID'7
9.13 X 10-6
151.3
' 110.8
31.5
7.18
1.04
2.38 X 10-«
1.74 X 10-6
4.95 X 1(T7
1.13 X10"7
1. 63X10-*
• Data from U S. EPA (1985).
* Computed assuming an tests had been conducted in Narragansett Bay water having a stability constant of 7.27 X 10T and a Ggand concentration of
1.38 X 10"' mol/L (8.77 itg Cu binding capacity per Dter).
complex. For example, for mysid the difference between the
LCSO value in microgram copper per liter and the ligand con-
centration (expressed in Cu binding capacity per liter) for Nar-
ragansett Bay, Reference Site, Site I, Site 2, Site 3, and Site 4
are 151.2, 151.2, 150.7, 148.7, 151.2, and 150.8 /tg Cu/L, re-
spectively.
However, for the most sensitive organism, the blue mussel,
the concentration of copper at the LC50 concentration for all
six locations is less than the organic ligand concentration at the
sites. The excess of ligand relative to the LC50 copper concen-
tration is not the same for all waters. Both the stability constant
for the complexation reaction and the value of the concentration
of the ligand are important in regulating the concentration of
bioavailable copper. For the blue mussel, the LCSO total copper
concentration increases from Site 1 to Site 2 and from Site 3 to
Site 4, reflecting the importance of the concentration of ligand
in determining the bioavailability of the copper as was the case
for mysid. However, for the blue mussel the LCSO values for
Site 1 and 3 and for Sites 2 and 4 are not the same. Because the
concentration of bioavailable copper is a function of the ligand
concentration and the stability constant of the complex as well
as the total copper concentration, the WER is different for or-
ganisms of different sensitivity. This change in WER as a function
of species sensitivity is why the WER procedure (U.S. EPA,
1994) recommends selection of species or life stages whose sen-
sitivity is equal to or slightly greater than the CMC or CCC
concentrations.
As we have noted, the effect of the organic ligand concentration
is particularly noteworthy. The increase in WER going from the
insensitive mysid to the highly sensitive mussel is much greater
at the higher than at the lower ligand concentration. Sites 1 and
2 and Sites 3 and 4 simulate the behavior expected for dilution,
such as could happen for an effluent discharge to a pristine re-
ceiving water, or in bioassay studies of effluents conducted by
dilution of the effluent
An increase in stability constant at the same ligand concen-
tration, simulated by Sites 1 and 3 and Sites 2 and 4, shows that
the WER is virtually unaffected for insensitive organisms. In
this situation, practically all the copper would be present as free
copper ion plus inorganic complexes. However, for the most
sensitive organisms, in those cases where the concentration of
ligand is in stoichiometric excess relative to the total copper
concentration, the stronger ligand results in a higher proportion
of the copper being bound. Consequently, there is a large dif-
ference in the WER. For example, the WERs for mysid for Sites
2 and 4 are 2.46 and 2.47, respectively, the WERs for mussel
for Sites 2 and 4 are 10.9 and 22.7, respectively.
If WERs are to accurately adjust WQC to reflect metals avail-
ability at specific sites, testing must use species sensitive at or
just above criteria concentrations. Use of insensitive species or
tests may result in low WERs and conservative site-specific cri-
teria. Conversely, use of species or tests sensitive below the cri-
terion may result in a large WER and a site-specific criterion
that is not sufficiently protective. This is to be avoided. For cop-
per, acute tests with larval mussels or other molluscs that are
most sensitive will result in the highest values of the WER. Be-
cause acute and chronic criteria for some other metals differ
markedly, use of chronic tests sensitive at the criteria continuous
concentration should result in larger WERs more appropriately
protective at this chronic criteria concentration.
Table 5—Copper complexation parameters for water effect ratio simulation. Results are shown in Figure 9.
Sample
Ugand concentration
Stability constant
mol/L
fig Cu binding capadty/L
Narragansett Bay
Reference Site
Srtel
Site 2
Site3
Site 4
7.27 X 107
9.00 X 107
3.50 X 1Cr*
3.50 X 107
200X10"
200X10*
1.38X10'7
2.00 X 10'7
1.00 xicr*
4.00 X ID"6
1.00 X 10-8
4 00 X 10-«
8.77
12.71
6355
254.18
63.55
25418
January/February 1996
51
-------
Total Copper (mol/L)
2 £
CD
8
O) K CO
ar» s
^ CD
cj co co
Total Copper (ug/L)
Figure 8—Relationship between free plus inorganic cop-
per complexes and total copper for the water at the site
used to develop the water quality criteria (dashed line)
and Site 4 water for the assessment of water effect ratio
(solid line). The relationship Is shown for the toxiclty of
copper for the three least sensitive, organisms used in
the evaluation as discussed In the text.
As a consequence .of the effect of the ligand concentration
and stability constant on metal spcciation, and consequently on
the bioavailability of copper, it can be concluded that the organic
matter content of a natural water (in the absence of paniculate
matter) should be predictive of the WER for nonsensitive or-
ganisms such as mysid or polychaete. However, for more sen-
sitive species, such as the blue mussel, knowledge of the organic
matter content alone will not allow the bioavailability of copper
to be predicted. For sensitive organisms, a more complete de-
scription of metal complexation will be required to predict bio-
availability. This will necessitate either measurement of specific
chemical species or the determination of both the concentration
of ligands and the stability constant for their complexation with
copper so that the concentration of bioavailable copper can be
calculated. If the quantity and/or quality of organic matter
changes seasonally, one WER may not be appropriate for the
entire year. Measurements or estimates of bioavailable copper
concentrations and a copper criterion based on bioavailable
metal, if appropriately determined, would be preferred over the
present site-specific approach.
Conduct of WER studies as specified in U.S. EPA (1994) and
simultaneous measurement of specific chemical species or sol-
uble and paniculate ligand concentrations and their stability
constants will permit comparisons of WERs derived toxicolog-
ically and by chemical theory. If they are similar, WERs derived
using chemical theory may be most appropriate because they
can be calculated for the exact criteria concentration and are
less expensive.
Summary
We have provided a detailed description of the complexation
chemistry of copper in natural waters and its importance to
copper speciation and to bioavailability, in particular its rela-
tionship to water quality criteria and to water effect ratios. In a
water at a constant pH and total alkalinity and at thermodynamic
equilibrium, the concentrations of free copper ion and of all
inorganic species covary as the total copper concentration is
changed. The bioavailability and toxicity of copper in a system
at constant pH and alkalinity will be proportional to the con-
centration of free copper ion and of inorganic copper complexes.
Because dissolved organic matter is present at low concentrations
in natural waters and forms strong complexes with copper, the
concentration of copper complexes with dissolved organic matter
changes in a nonlinear manner during the course of the titration
of a sample of natural water with copper. Because of this non-
linearity, which is most pronounced at concentrations of copper
500
400
1
a.
a.
300
200
100-
(a)
Pi
1 /
S
s
s
s
s
s
s
s
s
s
t\
t\
t\
t ^
t *(
1
-1
Q Narragansett Bay
M Reference Site
B Site 1
• Site 2
m Site 3
D Site 4
^J
"
rJ
-1
J
S.
CD
I
Figure 9—Simulated results for the development of a site
specific criterion for copper. Stability constants for Sites
1 and 2 are 3.5 X 10T and 2.0 X 10* for Sites 3 and 4.
Ligand concentrations are 1.00 X 10~* mol/L (63.5 jig Cu
binding capacity per liter) for Sites 1 and 3 and 4.00 X
10~* mol/L (254.2 ng Cu binding capacity per liter) for
Sites 2 and 4. (a) Total copper LC50 concentrations and
(b) water effect ratios.
52
Water Environment Research, Volume 68. Number 1
-------
Allen and Hansen
less than the equivalent concentration of ligand, the concentra-
tion of free copper ion plus inorganic copper complexes will not
be the same for different waters having the same total copper
concentration. Furthermore, the fraction of the copper present
as the free metal ion plus inorganic copper complexes does not
change proportionally to the concentration of ligand or to the
stability constant of the complex.
Discharged wastewater typically contains copper and an excess
of ligand. As this is diluted in the receiving stream, an increased
fraction of the copper is present in the form office copper ion
plus inorganic copper complexes, but because of the concurrent
dilution, the concentration of free copper ion plus inorganic
copper complexes decreases. If the pH of the receiving water is
not the same as that of the effluent, the conditional stability
constant will change as will the ligand concentration. The ability
of an effluent to complex copper may decrease by more than
one order of magnitude when the pH is lowered by one unit
It appears that if water quality criteria for copper in marine
water were to be based on free copper ion plus inorganic com-
plexes rather than total dissolved copper, the criterion would
change only slightly. Therefore, the importance of developing a
national criterion based on bioavailable metal is not that it will
markedly change the criterion, but that national and site-specific
criteria should not differ if both are expressed in terms of the
bioavailable form.
If receiving water standards are in terms of dissolved or bio-
available metal, it is necessary to predict the rate and extent of
transformation of metal from total recoverable metal to the reg-
ulated form below the point of discharge to enable appropriate
waste load allocation and issue discharge permits that are neither
over nor under protective.
On the basis of chemical speciation as a function of total
copper concentration, water effect ratios are expected to increase
as more sensitive organisms are used in the evaluation. This
prediction is borne out by experimental results. For insensitive
organisms, the concentration of ligand present is less than the
copper added to achieve the toxic response. The toxic response
is primarily determined by the difference in concentration be-
tween copper and the ligand. However, when the most sensitive
organisms are used in the test, the toxic response is obtained for
concentrations of metal less than the concentration of the ligand.
The toxicaty of the metal depends on both the concentration of
ligand in the sample and the stability constant for the complex-
ation reaction.
One goal of this simulation was to demonstrate the potential
practical importance of theoretical calculations of WERs as
checks against their empirically obtained values as presently re-
quired in EPA guidance. If appropriate data are derived, in-
cluding measurement of specific chemical species or ligand con-
centrations and stability constants, as part of WER studies (US.
EPA, 1994) it may be possible to demonstrate the practical utility
of WERs calculated directly from chemical properties of site
water. This could permit less expensive and more accurate de-
terminations of site-specific adjustments to national WQC that
are appropriately protective. The WER approach to adjust na-
tional WQC should be considered temporary with the ultimate
goal of understanding and accounting for all metal-containing
phases in water that account for the true biologically available
metal in waters that represent all surface water types. Through
this basic understanding, WQC for metals and other substances
can be expressed on a biologically available concentration basis.
Acknowledgments
Credits. The authors thank Wilson Jardim, Univcrsidade Es-
tadual de Campinas; Bo Shi, University of Delaware; E. Michael
Perdue, Georgia Institute of Technology, Dominic Di Toro,
HydroQual; Warren Boothman and Charles E Stephan, US.
EPA; Joseph S. Meyer, University of Wyoming; and two anon-
ymous reviewers for valuable comments. This document has
been reviewed according to EPA laboratory requirements, and
its content does not necessarily reflect agency policy.
Authors. Herbert E. Allen is a professor at the Department
of Civil and Environmental Engineering, University of Delaware,
Newark. David J. Hansen is a research aquatic biologist at the
US. Environmental Protection Agency, Environmental Re-
search Laboratory, Narragansett, Rhode Island. Correspondence
should be addressed to Herbert E Men, Department of Civil
and Environmental Engineering, University of Delaware, New-
ark, DE 19716.
Submitted for publication April 1, 1994; revised manuscript
submitted February 14,1995; accepted for publication March 8.
1995. Deadline for discussions of this paper is May 15, 1996.
Discussions should be submitted to the Executive Editor. The
authors will be invited to prepare a single Closure for all discus-
sions received before that date.
References
Allen, H. E (I993a) The Significance of Trace Metal Speciation for
Water, Sediment and Soil Quality Standards. Sci. Total Environ.,
134,23.
Allen, H. E (1993b) Importance of Metal Speciation to Toxicity. Proc.
Water Environ. Fed Preconfererux Seminar Aquatic Life Criteria
for Metals. Anaheim, CA, 55.
Allen, H. E, Boonlayangoor, C, and Noll, K. E (1982) Changes in
Phyacochemical Forms of Lead and Cadmium Added to Freshwater.
Environ. Internal.. 7,337.
Benson, W. H., Alberts. J., Allen. H. E, Hunt, C D, and Newman,
M. C (1994) Synopsis of Discussion Session on Bioavaflabflity of
Inorganic Elements: In A Mechanistic Understanding of BioavaD-
atxlity: Physical, Chemical, and Biological Interactions. J. L. Ha-
melink, W. H. Benson, H. L. Bergman, P. F. Landnim (Eds.), Lewis,
Boca Raton. FUt, 69-77.
Brown. V. M., Shaw, T. L, and Shurben, D. G. (1974) Aspects of Water
Quality and the Touchy of Copper to Rainbow Trout Water Res.
(G.B.).8,797.
Brungs, W. A, Holdennan, T. S., and Southenan, M. T. (1992) Synopsis
of Water-Effect Ratios for Heavy Metals as Derived for Site-Specific
Water Quality Criteria. EPA Contract 6S-CO0070. U.S. EPA,
Washington. D.C
Buckley, P. J. M., and van den Berg, C M. G. (1986) Copper Com-
ptexation Profiles in the Atlantic Ocean, A Comparative Study Using
Electrochemical and Ion Exchange Techniques. Mar. Chem.. 19,
281.
Cabaniss, S. E, and Shuman. M. S. (1988) Copper Binding by Dissolved
Organic Matter. L Suwannee River Fulvic Acid. Geochlm. Cos-
mochim. Ada. 52,185.
Chakoumakos, C, Russo, R. C, and Thurston, R. V. (1979) Toxicity
of Copper to Cutthroat Trout (Salmo darkif) under Different Con-
ditions of Alkalinity, pH, and Hardness. Environ. Sci. Technol.. 13,
213.
Coale, K. H., and Bruland, K. W. (1988) Copper Complexation in the
Northeast Pacific. Umnol. Oceanogr.. 33,1084.
Cowan, C E, Jenne, E A, and Kinnison, R. R. (1986) Methodology
for Determining the Relationship between Toxicity and the Aqueous
Speciation of a Metal Aquatic Toxicity and Environmental Fate:
Ninth Volume. ASTM STP 921.T. M. Poston and R. Purdy (Eds.)
American Society for Testing and Materials, Philadelphia, Pa., 463.
January/February 1996
53
-------
Allen and Hansen
Grosser. P. (1980) Investigation of Copper-Natural Ligand Complexes
by MnOj Adsorption. M.S. thesis, Illinois Institute of Technology,
Chicago, III
Ephraim, J., and Marinsky, J. A. (1986) A Unified Physicochemical
Description of the Protonation and Metal Ion Complexation Equi-
libria of Natural Organic Acids (Humic and Fulvic Acids). 3. Influ-
ence of Polyelectrolyte Properties and Functional Heterogeneity on
the Copper Ion Binding Equilibria in an Armadale Horizons Bh
Fulvic Acid Sample. Environ. Sci. Techno!., 20,367.
Fish, W., Dzombak, D. A., and Morel, F. M. M. (1986) Metal-humate
Interactions. II. Application and Comparison of Models. Environ.
Sci. Techno!.. 20,676.
Florence, T. M. (1986) Electrochemical Approaches to Trace Element
Spetiation in Waters: A Review. Analyst. Ill,489.
Hodson, P. V., Borgmann, U., and Shear, H. (1979) Touchy of Copper
to Aquatic Biota: In Copper in the Environment. Part II. Health
Effects. J. O. Nriagu (Ed), Wiley-Interscience, New York, N.Y.
Howarth, R. S., and Sprague, J. B. (1978) Copper Lethality to Rainbow
Trout in Waters of Various Hardness and pH. Water Res. (G.B.),
12,455.
Kelly, M. (1988) Mining and the Freshwater Environment Elsevier,
Essex, UK-
Kramer, C. J. M. (1986) Apparent Copper Complexation Capacity and
Conditional Stability Constants in North Atlantic Waters. Mar.
Chem. 18,335.
Kramer, C. J. M., and Duinker, J. C. (1984) Measurement of Copper
Complexation by Naturally Occurring Ligands. In Complexation
of Trace Metals in Natural Waters. C J. M. Kramer and J. C.
Duinker (Eds.) Nijhoff/Junk, The Hague, The Netherlands, 217.
Luoma, S. N. (1983) Bioavailability of Trace Metals to Aquatic Organ-
isms—A Review. Sci. Total Environ.. 28,1.
Magnuson, W. R., Harriss, D. 1C, Sun, M. S., Taylor, D. K., and Glass,
G. E (1979) Relationships of Activities of Metal-Ligand Species to
Aquatic Toxicity: In Chemical Modeling in Aqueous Systems: Spe-
ciation, Sorption, Solubility and Kinetics. E. A. Jenne (Ed.) Symp.
Ser. 93. American Chemical Society, Washington, D.C
Meador, J. P. (1991) The Interaction of pH, Dissolved Organic Carbon,
and Total Copper in the Determination of Ionic Copper and Tox-
icity. Aquat. Toxicoi. 19,13.
Neubecker, T. A., and Allen, H. E. (1983) The Measurement of Com-
plexation Capacity and Conditional Stability Constants for Ligands
in Natural Waters—A Review. Water Res. (G.B.), 17,1.
OTtonnel, J. R., Kaplan, B. M., and Allen, H. E (1985) Bioavailability
of Trace Metals in Natural Waters: In Aquatic Toxicology and Haz-
ard Assessment Seventh Symposium, American Society for Testing
and Materials, Philadelphia, Pa.
Pagenkopf, G. K., Russo, R. C., and Thurston, R. V. (1974) Effect of
Complexation on Toxicity of Copper to Fishes. / Fish. Res. Bd.
Con.. 31,462.
Pankow, J. (1991) Aquatic Chemistry Concepts. Lewis, Chelsea, Mich.
Perdue, E. M. (1988) Measurements of Binding Site Concentrations in
Humic Substances: In Metal Speciation—Theory, Analysis and
Application. J. R. Kramer and H. E. Allen (Eds.) Lewis, Chelsea,
Mich., 135.
Perdue, E M., and Lytle, C. R. (1983) Distribution Model for Binding
of Protons and Metal Ions by Humic Substances. Environ. Sci.
Techno!.. 17,654.
Plavsic, M.. Krznaric, D., and Branica, M. (1982) Determination of the
Apparent Copper Complexing Capacity of Sea Water by Anodic
Stripping Voltammetry. Mar. Chem.. 11,17.
Protbro, M. G. (1993) Memorandum: Office of Water Policy and Tech-
nical Guidance on Interpretation and Implementation of Aquatic
Life Metals Criteria. U.S. EPA, Washington, D.C.
Robinson, M. G., and Brown, L. N. (1991) Copper Complexation during
a Bloom of Gymnodinium sanguineum Hirasaka (Dinophyceae)
Measured by ASV. Mar. Chem.. 33,105.
Ruzic. 1. (1982) Theoretical Aspects of the Direct Titration of Natural
Waters and Its Information Yield for Trace Metal Speciation. Anal.
Chim-Acta, 140,99.
Skrabal, S. A., and Allen. H. E. (1993) Unpublished results.
Skrabal. S. A., Luther, G. W., and Allen, H. E (1992) Chemistry and
Bioavailability of Copper in Indian River Bay, Delaware. Final report
to Electric Power Partners.
Srna, R. F., Garrett, K. S., Miller, S. M., and Thum, A. B. (1980) Copper
Complexation Capacity of Marine Water Samples from Southern
California. Environ. Sci. Techno!., 14,1482.
Stephan, C E, Mount, D. L. Hansen, D. J., Gentile, J. H., Chapman,
G. A., and Brungs, W. A. (1985) Guidelines for Deriving Numerical
National Water Quality Criteria for the Protection of Aquatic Or-
ganisms and Their Uses. VS. EPA, Office of Research and Devel-
opment PB85-227049. NTIS, Springfield, Va.
Stumm, W., and Morgan, J. J. (1981) Aquatic Chemistry. John Wiley
and Sons, New York, N.Y.
Thursby, G. (1993) Personal communication.
Tipping, E, Reddy, M. M., and Hurley, M. A. (1990) Modeling Eleo
trostatic and Heterogeneity Effects on Proton Dissociation from
Humic Substances. Environ. Sci. Techno!.. 24,1700.
U.S. EPA (1985) Ambient Water Quality Criteria for Copper—1984,
EPA 440/5-84-031. US. EPA. Washington, D.C
US. EPA (1986) Quality Criteria for Water 1986. EPA 440/5-86401
VS. EPA, Washington, D.C
VS. EPA (1992) Interim Guidance on Interpretation and Implemen-
tation of Aquatic Life Criteria for Metals. U.S. EPA. Washington,
D.C
VS. EPA (1994) Interim Guidance on Determination and Use of Water
Effect Ratios for Metals. EPA 823-B-94-001. VS. EPA, Washington,
D.C.
van den Berg, C. M. G., Wong. P. T. S., and Chau, Y. K. (1979) Mea
surement of Complexing Materials Excreted from Algae and Theii
Ability to Ameliorate Copper Toxicity. J. Fish. Res. Bd. Can.. 36
901.
54
Water Environment Research, Volume 68. Number 1
-------
2 Interactions between Trace
Metals and Aquatic Organisms:
A Critique of the Free-ion
Activity Model
Peter G. C Campbell
IHRS-Eau. Quibee, Canada
1 Introduction 45
1.1 Scope 45
1.2 Metal-Organism Interactions (Notions) 46
2 Derivation of the Free-ion Activity Modd (HAM) 48
2.1 Historical Development 48
2 J Formulation of the Free-ion Activity Model 53
3 Critical Review of the Literature 56
3.1 Results Conforming to the Free-ion Activity Modd 56
3.2 Apparent Exceptions to the Free-ion Activity Modd
(Defined Media) 62
3.2.1 Lipophflic Complexes 62
L 3.2.2 Inorganic Ligands 65
3.2.3 Defined Organic Ligands Forming Hydrophflic Metal
Complexes 71
3.2.4 Problematic Examples 77
3 J Tests of the Free-ion Activity Modd in Systems Containing
Dissolved Organic Matter 79
3.3.1 Examples Conforming to the Free-ion Activity Modd 79
3 J2 Examples of Enhanced Toxidty in the Presence of
Dissolved Organic Matter 82
3.3.3 Examples of Enhanced Protection in the Presence of
Dissolved Organic Matter 90
4 Conclusions 91
Acknowledgment 95
Glossary 95
References 97
1 INTRODUCTION
1.1 SCOPE
ID considering the interactions of trace metals with aquatic biota, one can
identify three levels of concern: (1) metal speciation in the external
environment; (2) metal interactions with the biological membrane separating
Utul SpttiaUm mi KtxnoUMSty h Aquatic Sjtttnu. Edited by A. Tetjkr «nd D. R. Tomer
019951UPAC Published by JohaWDejr A Sons Ltd
-------
46 METAL SPEdATION AND BIOAVAILABIUTY IN AQUATIC SYSTEMS
the organism from its environment; (3) metal partitioning with the organism,
and the attendant biological effects.1. Each or these levels is dealt with
separately in this volume (e.g. speciation in Chapters 4 through 7, transport
, across biological membranes in Chapter 1, and intracellular partitioning of
metals in Chapter 10). This Chapter draws on this information and explores
the links between the speciation of metals in the external environment and their
bioavailability.
Much qualitative evidence exists to the effect that the total aqueous
concentration of a metal is not a good predictor of its 'bioavailability'. i.e., the
metal's speciation will greatly affect its availability to aquatic organisms. The
aim of this chapter is to examine this evidence from a quantitative point of view
and to determine to what extent the free-ion activity model (FIAM) can
explain the available data. Particular attention is accorded to apparent
'exceptions' to the FIAM. to test the limits of its possible application to the
natural environment
Given the rather broad subject'area to be covered, it seems advisable to
define its scope at an early stage. The metals of concern are those that exist in
natural waters as dissolved cations (e.g., A), Cd, Cu, Fe, Mn, Ni, Pb, Zn); we
do not deal with organometallic spedes (see Chapter 3), nor with paniculate
metals. The emphasis is on controlled laboratory studies, often carried out in
defined systems with unicellular organisms, but we do attempt to extrapolate to
multicellular organisms and to natural waters.
1.2 METAL-ORGANISM INTERACTIONS (NOTIONS)
The nature of the abiotic/biotic interface is described in detail by Simkiss and
Taylor (Chapter 1). On approaching the surface of a living organism, a, metal
will normally first encounter a protective polysaccharide or glycoprotein layer
(e.g., the cell wall for microorganisms or higher plants; mucus for animal cells).
The macromolecules making up this external layer contain a variety of simple
functional groups, usually nonchelating, dominated by ^containing donor
groups. At circumneutral pH values many of these groups will be ionized,
affording a matrix of negatively charged sites through which the metal must
migrate. Moving inward the metal win eventually meet the plasma membrane.
Though the detailed composition of such membranes may well vary
considerably from one species to another, or indeed from one tissue to
another2 (also see Chapter 1), their general properties remain similar. For our
purposes, the important features of the plasma membrane barrier are its overall
hydrophobia, phospholipidic character; the presence of proteins, some of
which may traverse the lipid bilayer; and the existence of transport proteins
and/or ion channels that facilitate the movement of ions across the membrane.
The incoming metal will thus encounter a wide range of potential binding sites,
which may be usefully divided into two types: physiologically inert sites, where
47
^-xA/vxrw^vxr*
-------
48 METAL SPECtATION AND BIOAVAtLABtUTY IN AQUATIC SYSTEMS
the metal may 'collect* without obviously perturbing normal cell function, and
physiologically active sites, where the metal affects cell metabolism. In the
latter case metal binding may affect cell metabolism directly (e.g., if the binding
site corresponds to a membrane-bound enzyme) or indirectly (if the bound
metal is subsequently transported across the plasma membrane into the cell).
Once within the cell, the metal may interact with a great variety of intracellular
sites, with obvious metabolic consequences.
Within this general model (see Figure 1), the interaction of a metal with an
aquatic organism involves the following steps: (1) advection or diffusion of the
metal from the bulk solution to the biological surface; (2) diffusion of the metal
through the outer 'protective layer'; (3) sorption/surface complexation of the
metal at passive binding sites within the protective layer, or at sites on the outer
surface of the plasma membrane; (4) uptake or 'intemalization* of the metal
(transport across the plasma membrane, see Chapter 1). The biological end
points considered include metal bioaccumulation per se (sorption, uptake), as
well as metal effects on such processes as photosynthesis, respiration, motility,
growth, and reproduction; the goal is to predict the 'bioavailability' of a metal
as a function of its speriation in the bulk solution. Certain aspects of this topic
were considered in earlier review articles.'4
2 DERIVATION OF THE FREE-ION ACTIVITY MODEL (FIAM)
2.1 HISTORICAL DEVELOPMENT
In much of the early (pre-1975) research on metal-organism interactions, the
emphasis was on the target organism and the influence of biological variables
(e.g., life stage, nutrition, age, etc.) rather than on the exposure regime (metal
speciation, pH, [Ca], alkalinity, ionic strength, etc.). Many of the experiments
were simply performed by adding the metal to a standard growth medium, with
no real appreciation of how this medium might affect the speciation of the
metal of interest. There was, however, a qualitative recognition of the
importance of metal 'complexation', gained largely from experiments in which
metals were added to the normal growth medium, with or without various
metal-binding ligands. In an example typical of this qualitative approach Lewis
el at.1 studied the effect of Cu on the calanoid copepod Euchaeta japonica.
Their experiments were carried out in seawater to which a variety of potential
complexing agents were added (ascorbic acid, 0-alanine, sewage effluent, soil
extract, humic acid, phytoplankton extract); the protection afforded by these
additions was attributed to their complexing capacity, which was expressed as
'EDTA-equivalents* (ethylenediaminetetraacetic acid).
A major impetus for change occurred in the early 1970$ with the propagation
of computer programs designed to perform complex chemical equilibrium
calculations. Equilibrium modeling of natural waters was begun in the 1960s,1
' P. O. C. CAMPBELL
49
but the potential of this approach for the study of metal-organism interactions
was not immediately appreciated. Improved access to computing facilities and
the popularization of these programs in the environmental literature9 opened
up the possibility of performing bioassays in defined synthetic media,
containing only known ligands. A subtle shift of emphasis ensued, from the
target organism to the chemistry of the exposure medium; bioassays were
performed in defined media, with synthetic ligands having known stability
constants. In such systems metal speciation in the exposure media could be
calculated and manipulated, subject, of course, to the limitations of the
stability constant database10 (see Chapter 4).
The advantages of using algae as bioassay organisms are well documented
(e.g., small size, access to large populations, ease of culture, simple inorganic
culture media, rapid growth rate; see Walsh" and Chapter 11). Indeed, many
of the early experiments in chemically defined exposure media were carried out
with unicellular algae; metal accumulation by the algae, or the effects of metal
additions on algae growth (stimulation or inhibition), were monitored as an
indication of the metal 'bioavailability'. Examples of such experiments, each
performed with Zn as the test metal, are described below.
In the first example, chosen to demonstrate the role of Zn as an essential
micronutrient, Anderson et a/.11 examined the effects of Zn on the growth of
Thalassiosira wcissflogii, a coastal diatom commonly used in laboratory assays.
The alga was innoculated into AQUIL,13-14 a synthetic seawater medium
specifically designed for studying metal-algae interactions. The free-Zn activity
was manipulated by increasing the total Zn concentration at each of three
different EDTA concentrations (1Q-4, IO-4-', and 10~s M EDTA). To avoid
possible complications due to the interaction of the added EDTA with other
trace metals, notably Cu and Mn, the free-ion activities of these metals were
kept constant by appropriate manipulations of [Cu]r and [Mn]r for each
EDTA concentration. For a given EDTA concentration algal growth increased
as Zn was added (Figure 2A), indicating that the algae were indeed Zn limited;
note that distinct response curves were obtained for each EDTA level. On
replotting the same growth data as a function of the free-Zn ion activity,
however, all three experiments showed the same dependence of the growth rate,
H, on [Zn2*] (Figure 2B). Clearly, Zn ion activity rather than total Zn
concentration determines the growth rate of T. weissflogii under Zn limiting
conditions. If Zn levels were increased above about 10~" M Zn2*, the
maximum growth rate of about 2.8 doublings per day was maintained until the
threshold for Zn toxicity was reached (10~» to 10-'° M Znz*).12
In the second example, illustrating this inhibitory effect of Zn on algal
growth, Allen el a/.is studied the effects of added ligands on the toxicity of Zn
to a blue-green alga, Microcystis aeruginosa. Experiments were carried out in a
synthetic medium at a constant Zn concentration (4.8 x 10~7 M) in the
presence of a variety of complexing agents (EDTA; nitrilotriacetate, NTA;
-------
50 METAL SPEC1ATION AND BIOAVA1LABIUTY IN AQUATIC SYSTEMS
3.0
.T* 2.5
•o
o>
5 2.0
-8
1.5
3.0
2.0-
1.5
B
o
A
0*0
13
12
pZnT
11
10
pZn2*
Flfore 2. Variation in the growth rate of the marine diatom, Thalaalotlra wtluflogtt, as
a function of (A) total Zn. (Zn]r. or (B) free-Zn concentration, (Zn2*]. Three different
EDTA concentrations were used: 10-' M (O); 10'° M (£). and I0~4 M (D).
(Modified from Anderson. M. A, el «/.. Nature, 276, 70 (1978)]
oxydisuccinic acids. ODS; carboxymethyloxysuccinic acid. CMOS;
carboxymethyltartronic acid. CMTA). To minimize possible metal-metal
interactions, only Fe was added to the AAP16 (see Glossary) growth medium as
a 1:1 Fe(III)-EDTA chelate; no other trace metals were added. In the absence
of added completing agents, growth of the test alga was completely inhibited.
Addition of the dictators reduced the toxkaty of Zn, but the concentrations
needed to restore growth differed for each ligand: EDTA, 2.5 x 10~7 M; NTA,
4x 10-7 M; ODS, l.Ox 1Q-J M; CMOS, l.Ox I()-J M; CMTA, l.Ox 10-« M.
The growth data for all experiments could be rationalized by plotting the
logarithm of the number of cells (day 5) vs the calculated free-Zn ion
concentration. A linear relationship was obtained (Figure 3), even though the
conditional stability constants of the five dictators varied by nine orders of
magnitude and their concentrations by three orders of magnitude. As in the
case for Zn stimulation of algal growth (see above), the biological effect of Zn
on the test alga proved to be directly dependent on the concentration of the
free-metal ion rather than the total metal concentration.
In the final example, taken from research in our own laboratory on Zn-
insemitivc algae.17 we examined the adsorption of Zn on the algal surface and
its transport across the cell membrane. The test algae (Chlumydomonas
Yariabilis, Scenetlexnnu subsplcattu) were maintained in AAP medium in the
absence of added Zn. Cells for the Zn uptake experiments were harvested from
a semicontinuous culture system by centrifugatton and resuspended in a
defined medium containing a constant total Zn concentration (15/iM) and
P. O. C. CAMPBELL
51
10'
10'
'J
s
to1
. o
0 0.2 0.4 0.«
[Zn2*] Ounol«L')
-Figure 3. Yield of the blue-green alga, Mieroeyats aeruginosa, after 5 days' growth as a
function of the free-Zn concentration [Zn2*]. Various different ligands were used to
adjust [Zn2*]: EDTA (D), NTA (A). ODS (x), CMOS (O), control (+). [Modified
from Allen. H. E. tt al.. Environ. Set. Teehnol., 14. 441 (1980).]
increasing concentrations of EDTA (0 to 12.9;iM). Short exposure times were
used (normally lOmin) to minimize the influence of the algae on the exposure
medium. After exposure, the algae were collected and Zn accumulatio
determined. The amount of Zn adsorbed onto the algal surface, Zn.,
operationally defined by extraction with EDTA (10~4 M); for both algae it
increased as a function of the free-Zn concentration remaining in the exposure
medium (Figure 4). and could be described by Langmuir-typc adsorption
isotherms. The amount of transported or cellular Zn, Zn,, operationally
defined as the Zn remaining with the algal cells after EDTA extraction, was a
linear function of free-ionic Zn over the-concentration range of 2 to 14 pM
(Figure 4). Once again, the activity of the free-metal ion rather than the
concentration of total metal determines Zn bioavailability to S. svbsplcatus and
C yariabilis.
Over the period 1976 to 1983, laboratory experiments similar to those
described above were carried out with a variety of aquatic organisms (c.g.,
algae, bacteria, macroinvertebrates, fish; see Section 3.1). A convincing body of
evidence was developed to support the tenet that the biological response
elicited by a dissolved metal is usually a function of the free-metal ion
concentration, M'^HjO),. The free-ion concentration is determined not only
-------
52
P. O. C. CAMPBELL
53
by the total dissolved metal concentration, but also by the concentration and
nature of the ligands present in solution. To rationalize these experimental
observations and explain what was perceived as 'the universal importance of
free-metal ion activities in determining the uptake, nutrition and toxicity of all
cationic trace metals'. Morel11 formulated the free-ion-activity model (FIAM)
for metal-organism interactions.
•a
-------
54 METAL SPEC1AT1ON AND B1OAVAILABILITY IN AQUATIC SYSTEMS *
By rearranging Equation (2), one obtains
<7)
' P. O. C. CAMPBELL
55
Substitution of Equation (7) into (6) then yields
{M-X^elO-Kj-K.-rX-cellXM'*) (8)
which shows the same dependency on Mz* as Equation (4).
'Possible mechanistic links between the formation of a surface complex.
M-X-cell, and the initiation of a biological effect have been suggested by
several workers.l*-22*24 If X-cell represents a physiologically active site at the
cell surface, then the binding of metal M might induce a direct biological
response (e.g., fish gills21). Alternatively, if X-cell corresponds to a transport
site that allows metal M to traverse the cell membrane and enter the cytosol,
then binding at the surface site would simply precede transport into the cell;
i.e., the actual reaction of M with the metabolically sensitive site would occur
intracellularly, following transport" In a variation of this scenario, -X-cell
might correspond to a transport site normally used by an essential
micronutrient; binding at the ceU surface site by metal M would then inhibit
the supply of the essential element and induce nutrient deficiency (e.g.,
phytoplankton: Mn/Cu, Fe/Cd).22-24
A number of key assumptions underlie the FIAM, some obvious and others
rather more subtle:
1. the plasma membrane is the primary site for metal interactions with living
organisms;
2. this interaction with the plasma membrane can be described as a surface
complexation reaction, forming M-X-cell (Equations (3) or (5); charges on
ligand not shown for simplicity);
3. metal transport in solution, towards the membrane, and the subsequent
surface complexation reaction occur rapidly, such that a (pseudo-)
equilibrium is established between metal species in the bulk solution and
those at the biological surface (rapid*, meaning faster than metal uptake,
faster than the expression of the biological response);
4. the biological response, whether it be metal uptake, nutrition, or toxicity, is
strictly dependent on the concentration of the M-X-cell surface complex,
{M-X-cell};
5. in the range of metal concentrations of toxicological interest, the
concentration of free sites, {-X-cell}. remains virutally constant and
variations in {M-X-cell} follow those of [M1*] in solution (Equations (4) or
(8);tnd
6. during exposure to the metal of interest, the nature of the biological surface
remains constant (i.e., the metal does not induce any changes in the nature
of the plasma membrane).
Assumptions (2) and (3) merit particular consideration, as they influence our
discussion of possible 'exceptions' to the FIAM (Sections 3.2 and 3.3).
Regarding assumption (2), metals are presumed to exist in the exposure
solution as hydrophilic species; lipophilic metal species present in solution that
might traverse the plasma membrane without first forming a surface complex
are not considered. In the original formulation of the model the interaction of
metal M with the plasma membrane was described as a ligand exchange
reaction (Equations (2) or (4)). such that in the resulting surface complex metal
M was bound only to X-cell (M-X-cell) and water molecules. If one allows one
or more of the ligands originally bound to M to remain in the coordination
sphere of the metal.
ML +- X-cell £ L-M-X-ccl! (9)
i.e., formation of a mixed-ligand complex, then the concentration of the surface
complex L-M-X-cell. is proportional to (ML] rather than to [M1*] alone.
{L-M-X-cell} - K4{~X-ceIl}(ML]
(10)
Putting aside this possible complication for the time being, let us now
consider the implications of assumption (3) (fast transport and adsorption/
desorption kinetics). There are frequent references in the literature to the free-
metal ion as the 'toxic' or 'bioavailable' species.20-21'2T However, if it is assumed
that the cell surface is in equilibrium with the various metal species in the bulk
solution, and that this equilibrium precedes the expression of the biological
response, it follows that the identity of the mctul form(s) reacting with the cell
surface is of no biological significance—no single species in solution can be
considered more (or less) available than another. Though this point was made
quite explicitly by Morel,"-2* who referred to the 'profound and widespread
misconception that hydrated metal species is the active one', it has often been
overlooked. In a system at equilibrium, the free-metal ion activity reflects the
chemical reactivity of the metal. It is this reactivity that determines the extent
of the metal's reactions with surface cellular sites, and hence its
'bioavailability*.
The surface equilibrium assumption also implies that other (cationic) species,
notably the hardness cations, Ca2+ and Mg2*, and the hydrogen ion, H*, may
compete with the metal of interest, M, for binding at the surface complexation
site. Such antagonistic interactions would tend to reduce the equilibrium
-------
56 METAL SPEC1ATION AND BIOAVAILABILITY IN AQUATIC SYSTEMS
concentration of the surface complex. M-X-cell. and thus diminish the
biological response. Many of the early experiments that led to the
formulation of the FIAM were performed on marine organisms in high
salinity media at slight alkaline pH. The variations in (Ca2'), [Mg2*], or (H*J
in such media were of little concern. In poorly buffered freshwater systems,
however, variations of these parameters must be taken into account.
Finally, if the surface equilibrium assumption proves untenable, e.g., if the
reaction of M1* or ML at the cell surface (Equations (3) or (5)) is slower than
the rate ofintemalization of the metal (i.e., k<1[Mt*]{-X-cell} X-cell},
see Figure I),29*90 or if metal transport to the cell surface proves to be the rate*
limiting step, then the FIAM will no longer apply. The biological response will
be determined not by thermodynamic equilibrium considerations but by the
kinetics of the complexation reaction or the rate of metal transport in the
reaction layer outside the cell."-32
3 CRITICAL REVIEW OF THE LITERATURE
3.1 RESULTS CONFORMING TO THE FREE-ION
ACTIVITY MODEL
As alluded to in Section 2.1, a considerable body of experimental evidence has
accumulated in support of the FIAM. Examples of such experiments are
compiled in Table 1. These examples were selected to cover as wide a range of
metals and organisms as possible; however, coverage for a given metal and a
particular type of organism is not comprehensive. The studies on marine and
freshwater phytoplankton outnumber all others in Table I. This distribution
presumably reflects the sensitivity of algne to trace metuls" and the relative
ease with which they can be studied in the laboratory (defined inorganic growth
media; rapid response). References to invertebrates are also frequent, with a
distinct bias toward marine species. In contrast, most of the flsh studies were
performed on freshwater species.
Table 1 includes a selection of papers based on the following criteria:
• studies with living organisms or viable tissue;
• controlled exposure conditions, notably pH, and 'known' metal speciation in
the exposure medium; i.e., experiments performed in defined media for
which equilibrium speciation calculations could be performed with
confidence, or experiments performed at free-metal ion concentrations
amenable to detection by ion-selective electrodes (ISE) or other techniques;
and
• where possible, short exposure times and/or frequent renewal of the
exposure medium, in order to minimize the influence of the organisms on
metal speciation in solution.
57
•2
i
R
S5 « 5
i
> ^*
- s
•a* aj
5^ ^
53;
S— ri
-------
Table 1. (continued)
Species
Marine Bacteria
1. Marine isolate (Gram
negative)
2. Open ocean microbial
community
3. Estuarine microbial
community
Marine lor ertebrafes
1. Paleomonetes pugio
(grass shrimp)
2. Crassoslfta vlrginlca
(oyster)
3. Heanthu
arenaceodenlata
(annelid)
4. C. ttrginica
(oyster)
5. Rhilhropanopeus
karritii
(crab larvae)
6. JLharrisii
(crab larvae)
7. Anemia sp.
(brine shrimp)
8. A. franciscana
(brine shrimp)
Freshwater Algae
1. CUorella vulgaris
(chlorophyte)
Oocystis marssonii
(chlorophyte)
2. C. vulgaris
(chlorophyte)
3. Scenedesmus
quadricauda
(chlorophyte)
4. 5. quadricauda
(chlorophyte)
Ankistrodesmus
falcatus (chlorophyte)
5. Chlamydomonas
wiabilis (chlorophyte)
Scenedesmus
subspicatus
(chlorophyte)
6. C. rariabilis
(chlorophyte)
7. S. quadricauda
(chlorophyte)
8. Chlamydomonas
reMardii
(chlorophyte)
Metal (pM'»)
Cu
(12-8)
Co
(12-8J)
Co
(«-«)
Cd
(7.2-5)
Cd
(7.4-5.9)
Cd
(12-8)
Ca
(11-8.7)
Ca
(13.4-9.7)
Ca
(12.4-7.9)
Ca
(9-7.6)
Ca
(6.5-4.2)
CD
(16.7-15.5)
Ca
(12.2-10.3)
Cu
(12.8-6.8)
Zn
(10-5.4)
Ni
(8.5-5)
Zn
(5.7-1.9)
Zn
(6-5.5)
Mn
(6-4.3)
Fe
(7-5.7)
Cd
(6.3-2.9)
Cu
(10.4-4.6)
Cu
(8.2-6)
Medium
UV-treated filtered
seawater + NTA
Seawater + NTA
Estuarine water + NTA
Diluted seawater ±
NTA
Variable NTA or Cl-
DQuted seawater
Variable Q-
Fillered seawater
-»• EDTA
Artificial seawater. or
filtered seawater.
+NTA '
Diluted seawater
+ NTA
Seawater + NTA
Artificial seawater
+ EDTA. citrate, or
acetate
Artificial seawater
Variable pH (6-8)
Inorganic growth
medium + EDTA
Inorganic growth
medium + NTA. MES
Inorganic growth
medium (FRAQUIL)
Inorganic growth
medium (AAP)
Inorganic medium
(AAP-Fe) + EDTA
Inorganic medium
(AAP-trace
metals) + EDTA
Variable pH (5-7)
Inorganic medium
Variable pH (5-8)
Inorganic medium
Variable pH (4-9)
Exposure time
2.5 h
5h
1-3 h
4d
4d
3 weeks
14d
—
—
90 min
3h
4d
20d
chemostat
4d
14d
lOmin
10-20 min
<
2h
10 min
Response
Incorporadon "C-glucose
(uptake |4C vs pCu7*)
Incorporadon JH-amino acids
(% control vs pGr**)
Incorporadon JH-amino acids
(turnover time vs pCu2*)
Mortab'ty (% survival vs
pCd")
Uptake Cd (log [Cd]_t vs
pCd1*)
Uptake Cd (log [CdUvs
PCd'-) *•
Uptake Cu QCul,,, vs pCu1*)
Uptake Cu (cytosoGc Ca vs
pCu1)
Growth inhibition
Uptake Cu (c>losolic Cu vs
pCtt1*)
Growth inhibition
Mortality
Uptake Cu (accumulation
rate vs ICu2*])
Uptake Cu
Growth stimulation
Growth inhibition
(CeO numbers vs pCu1*)
Growth inhibition
{jt vspC^'.pZn1*)
Growth inhibition
(u vi pNr')
Uptake Zn
flZaJ^ vs [ZnJ*D
QZn]^, vs (Zn]*D
Uptake Zn, M n(IO. FeGO
(MU vs f>lj*l
(Mlrf VS [M2*j
Uptake PO4, NH«. NO,
(inhibition)
ECM vs pH
Adsorption Cu vs [Cu1*]
\t
Rcf
42
43
44
45
46
34.47
48
49
50
51
35
52
53
20
54
17
55.56
57
27
58
{cuauanJ}
-------
60
Jt
I
ft
S
3
2 S
« - 3
2? 8rgr
•g
.1
a
I
2
f-l l 2B I"?, SL
> Cd>Zn> > all
others; references to trivalent metals of environmental importance (e.g.. A1J*
or Fe1*) are conspicuously absent.
~ Various endpoints were studied in the bioassays, covering a spectrum from
highly specific (e.g., metal accumulation perse: surface adsorption, absorption,
subcellular distribution) to more integrative measures (e.g., growth, motility,
mortality). In presenting their biological response data many authors resorted
to the use of logic scales, both for the free-ion concentration in the exposure
medium and sometimes for the biological response itself. Rarely is this choice
explained; the log-log plots are presented and the correlation between log
'response' and pM*+ noted, but the mechanistic implications of having
changed from a direct (arithmetic) relation to such a log-log relation are
glossed over.
Finally, the results obtained in freshwater can be contrasted with those
observed for marine organisms. From a close examination of the studies
performed on freshwater organisms under conditions in which the pH and/or
the water hardness was varied (Table I, freshwater algae: studies 6, 7, 8; fish:
studies 1 to S), it is clear that knowledge of the free-ion activity alone is
insufficient to predict the biological response. One must also consider potential
competition for the metal binding site, ~X-cell, by the hydrogen ion, H*. and
by the hardness cations, CuJ+ und Mg2*. These effects were explicitly included
in Morel's original formulation of the FIAM.™ but because most of the early
work was done in seawater, variations of (H *] and [CaJ*J were of relatively
little concern (for a notable recent exception, sec Blust et a/."). The influence of
the hydrogen ion on metal-organism interactions was reviewed by Campbell
and Stokes.16
•An added benefit of the use of meul buffer* h Out one jenertlly worki it quilt high total metal
concentration*, tuch that the effect* of any inadvertent contamination of the expoiure media are
-------
62 METAL SPEC1AT10N AND BIOAVAILABILITY IN AQUATIC SYSTEMS
3.2 APPARENT EXCEPTIONS TO THE FREE-ION MODEL
(DEFINED MEDIA)
Documented examples of experiments in which changes in metal bioavailability
do not conform to the FIAM were grouped into four categories, based largely
on the type of ligand involved:
1. organic ligands forming lipophilic complexes with the metal (Table 2);
2. inorganic anions (Table 3);
3. low molecular weight organic ligands forming hydrophilic complexes with
the metal (Table 4); and
4. miscellaneous examples defying rational explanation by the present author
(Table 6).
These classes are discussed below. Note that examples from less-well-defined
systems, i.e.. those containing natural dissolved organic matter (DOM), are
considered in Section 3.3.
3.2.1 Lipophilic Complexes
An early example of enhanced metal uptake in the presence of ligands forming
HpophiUc metal complexes was provided by Poldoski,70 who studied the short-
term accumulation (2-days) of Cd by Daphnla magna in the presence of a range
of inorganic and organic ligands. Metal concentrations were maintained at
sublethal levels and the free Cd ion concentrations were either measured (ISE)
or calculated. For most of the ligands (e.g., pyrophosphate, NTA, humic acid)
the observed steady-state levels of Cd in Daphnla followed the free Cd2*
concentration in solution (Figure 5), but for diethyldithiocarbamate (DDQ the
observed accumulation was far higher than predicted (up to 50% higher than
in the ligand-free control, even though the free Cd2* concentration was
reduced by ca. 10"). Diethyldithiocarbamate forms a very stable, neutral MLj
complex with Cd.
CH,CHj
CHjCH,
CH2CH,
CH2CH,
and it seems likely that the enhanced availability of Cd in the presence of DDC
reflects the ability of this complex to cross the plasma membrane.
Since the original report by Poldoski, several other workers have reported
enhanced metal uptake with analogous RiC(S)S" ligands, notably with
xanthates:
P. O. C CAMPBELL
63
I"
1"
re
i§ 0.5
0.0
10
IS 25 30
pCd2*
Figure 5. Uptake of Cd by the freshwater cladoceran, Daphnla magna, after 4 days in
the presence of different ligands forming hydrophilic complexes (closed circles), and in
the presence of DOC (open circles). Bioaccumulation ratio - ratio of Cd concentration
in D. magna in the presence of the ligand: Cd concentration in D. magna in the ligand-
free control culture. [Data from Poldoski, J. E. Environ. Sei. Ttchnol., 13, 701 (1979).]
R-O-C
II
These compounds are used in the mining industry as flotation agents and their
effects on metal availability were extensively studied by Swedish workers
(Table 2, studies 4 and 3). Experiments were performed both on perfused fish
gills (which allows the 'measurement of metal influx and retention) and on
living fish. In a recent study Block and Wicklund Glynn71 combined
determinations of Cd lipophilicity (octanol-water partitioning .of Cd
complexes) with measurements of l09Cd uptake by two fish species. They
determined the partitioning of Cd into octanol as a function of the xanthate
concentration and showed that the partition coefficient increased markedly at
xanthate concentrations greater than ca. 10~6 M (Figure 6A). Approximately'
the same concentration threshold was observed in the metal uptake
experiments (Figure 6B). Similar enhancements of metal uptake in the
presence of xanthates were reported for 4]Ni,ra M)Hg,71 and Pb74 (cited in
Block and Wicklund Glynn71).
-------
64
' P. 0. C. CAMPBELL
65
8
K
I
8
s
a
o
e*
3.
II
2 i
CM
Ok
•S
* *>;
3» I
03T3 ej?
itl 3rr!
"••a 0 u".".
VI iff
D^ 5 D =>
I1U
g
r'l ki
•as^-Big-l
l§-§55+5^;
£ £
d
S 'c » ?
s att
M w
eslas
isriisTii
'o iS 3*9 8 9
§•3.3 g-9-fl
S^lSss
•S42-S^a
IIIIII
8 3
Other examples of enhanced metal uptake/toxicity were reported for 8-
hydroxy-quinoline (oxine) and related ligands by Florence and co-workers
(Table 2, studies I and 3).
In this case only Cu appears to have been studied. Target species included a
marine alga75-7* and a marine invertebrate.77 In an experimental approach
resembling that described for Poldoski70 above, these workers screened a range
of ligands, usually at a fixed total Cu concentration, and compared the
biological response observed in the presence of the ligand with that observed in
ligand-free control. For ligands forming hydrophilic complexes, the expected
response (reduced toxicity) was observed, whereas for ligands forming lipid-
soluble complexes (oxine, ethylxanthate) Cu toxicity was enhanced (Figure 7).
In the case of oxine (Structure III), introduction of a sulfonate group in the 5-
position (Structure IV) renders the Cu complex hydrophilic, and the enhanced
toxicity disappears. Similar results have recently been reported for Cu, Cd and
Pb, in careful experiments performed on the marine diatom Thalasstosira
welssftogtl.w
In all the above examples, one may legitimately question the environmental
pertinence of the results, in that the ligand concentrations needed to provoke
enhanced metal uptake rates normally exceed those likely to be found in the
environment. However, the results may still be environmentally relevant (e.g.,
behavior of low molecular weight, lipophilic metal complexes) if such
complexes can be shown to exist in natural waters (see Chapter 6).
3.2.2 Inorganic Ligands (Table 3)
In experiments in which inorganic anions played an important role in
determining metal speciation. two examples were identified in which metal
availability apparently diverged from the FIAM prediction: one involving a
soft, monovalent cation (Ag*) and the other a hard, trivalent cation (A1J*). In
the first example, Engel and co-workers44 studied the influence of salinity on
the short-term (4-day) uptake of "°Ag by the grass shrimp. Palaemonetes
puglo. Their experimental design was similar to that used in earlier
experiments41 on the influence of salinity on Cd uptake (see Table 1. marine
invertebrates, study 1). As was the case with Cd, silver uptake decreased with
increasing salinity; however, unlike the Cd results, accumulation of Ag after
4days was not a function of the free-metal ion (Ag*), but was best related to
the concentration of the neutral AgCI° complex (Figure 8). The authors suggest
-------
6^WlETAL SPECIATION AND BIOAVAILABILITY IN AQUATIC SYSTENfT
(D
i
to
8
f
a.
c?
-7
log [xanthate] (mokL*1)
0»
"o
•o
O
7
e
s
4
3
2
1
0
1 |l Minnow
g3 Rainbow trout «
,
KEX
*
dh
. *;
O «
•
•
I«
«•
•
r
^
nfrf
iJUl PtvP7j44
5
{
B
^
5
j!
^
^
fi
c r.7 .e •£ .4 c r-7 -e -5 -
^
4
log [xanthate] (mol«L")
log [xanthate] (mol»L°)
P. O. C. CAMPBELL
67
100
o
6 20
pCuT
Figure 7. Toxicity of Cu to the marine unphipod. Atlorchtsta compressa, in the
absence (O) or in the presence (•) of oxine (O.S. I, and 2pM. as indicated next to the
data points). [Modified from AhsanuUah. M. and Florence, T. M.. Mar. Blot., 84, 41
(1984)]
that the correlation with AgCl° reflects the lipophilicity of the neutrally
charged complex and an enhanced rate of uptake across cell membranes.
Although the octanol-water partitioning of Ag(I) species as a function of
salinity was not determined, this explanation seems reasonable, and from a
mechanistic point of view this example probably belongs in Table 2 (lipophilic
complexes).
The second example, however, is clearly not a case of induced lipophilicity.
We exposed juvenile Atlantic salmon (Salmo salar) to AI at pH 4.S or 4.9, in
the presence of increasing fluoride concentrations. As anticipated, fluoride
eomplexation attenuated both the mortality of Atlantic salmon in the presence
Figure 6. (oppotlte) (A) The octanol/water partition coefficient of Cd (log) as a function
of the xanthate concentration. [CdJr-0.22nM; KEX-potassium ethyl xanthate (O);
KAX-potassium amyl xanthate (•). (B) Cd content of gill tissue from Eurasian dace
(minnow) and rainbow trout exposed to lolCd, as a function of the KEX concentration.
(Q Cd content of gill tissue from Eurasian dace (minnow) and rainbow trout exposed to
lo*Cd, as a function of the KAX concentration. The control (indicated by V). did not
contain either xanthate. Error ban correspond to ±SD (n-10). Starred values were
significantly different from the control (/><0.05). The dashed lines correspond to the
partition coefficient for Cd in an octanol/water system, i.e.. from (A), with the scale for
the partition values transformed to the scale for maximum and minimum uptake.
[Modified from Block, M. and Wickland Glynn, A.. Environ. Toxlcol. Chem.. 11, 873
(1992)1
-------
68
P. O. C. CAMPBELL
69
a
•a
u.
^
3
§
1
8
•3
OS
!
S'
I
•
o
e.10-1
I
HMO*
^4
•%!*
V4
4 o
, ,
o 12 11 10 98879
w pAg* pAgCI8 pAgCP2
Figure 8. Accumulation of Ag by the grass shrimp. Falaemonetes puglo, after 4 days at
Tour different salinities (O"*PPt; D"8ppl; A" 16ppt; O-32ppt). Linear log-log
regressions of (Ag]^^,- vs the concentrations of different chemical species of Ag had the
following r values: [Ag*], 0.72; (Agd°], 0.90; [AgClfJ, 0.51. (Modified from Engel.
D. W. el al., in Biological Monitoring of Marine Pollutants (New York: Academic Press.
1981), p. I27.J
of Al and the accumulation of Al at the gill surface, but not to the extent
predicted by the FIAM.71 At a given fluoride concentration, the survival
percentage did vary as a function of the free Al3* concentration, but the
response was not unique; i.e., it varied across experiments. Separate response
plots were obtained for each fluoride concentration (see Figure 9). For a given
AIJ* concentration, mortality was greater in the presence of fluoro-Al
complexes than in their absence; it follows that the biological response
depends not only on Al**, but also upon other Al species.
Multiple regression analyses were performed with percent mortality (7-day)
as the dependent variable and the various inorganic Al species (A1J*. A1F2*,
etc.) as predictors. The best prediction of fish mortality was obtained as a
weighted function of the species A1J* and A1F1* (r>-0.93 for smolt, 0.83 for
parr), a result that is clearly inconsistent with the FIAM. To explain the
toxicity results (i.e., the need to include A1F2* in the predictive model).
Wilkinson et al.n proposed that a mixed ligand fluoro-Al-gill complex formed
at the gill surface and contributed to the toxic response:
A1H + F- +" L-gill » F-Al-L-gill
Note that mathematically the concentration of this species, {F-Al-L-gill}, is
proportional to the concentration of [A1F2*] in solution (Equation (IS), where
-------
70 METAL SPECIATION AND B1OAVAILABIUTY IN AQUATIC SYSTEMS | * P. O. C. CAMPBELL
71
"0.0 1.0 2.0 3.0 4.0 6.0 6.0 7.0
lAI*] OunoUL'1)
Figure 9. Mortality of juvenile Atlantic salmon, Salmo solar, after 7 days as a function
of the free A1J* concentrations. Open symbols represent experiments in which no
fluoride was added. Closed triangles (A) correspond to 7/iM added fluoride, dosed
circles (•) correspond to 18 jiM added fluoride. (Modified from Wilkinson. K. J. tt at..
Can. J. Fish An suggest that
ternary surface complexes may be present not only as transient reaction
intermediates, but also as equilibrium species that contribute to the biological
response. Note that for such L-M-X-ccll species to 'contribute to the biological
response', one of two conditions must be satisfied: either ~X-ccll must
represent a metal-sensitive site at the origin of the biological response (i.e.. the
metal exerts its biological effect without crossing the plasma membrane), or
L-M must be transported intact across the plasma membrane (possible
examples of the latter behavior are discussed in the next section). If the L-M
bond breaks before M is transported in to the cell, leaving L behind, then the
dependence on {L-M-X-cell} and on [ML] disappears, and metal
bioavailability should conform to the FIAM.
333 Defined organic Uganda forming hydrophlUc metal complexes (Table 4)
Reports of enhanced metal availability in the presence of low molecular weight.
hydrophilic ligands are rare. Normally the complexation of a metal by such a
ligand is expected to decrease its bioavailability. and indeed in all the
documented examples such a decrease is noted. However, what distinguishes
each of the four cases described in Table 4 is the observation that the 'residual'
bioavailability of the metal in the presence of the ligand is greater than would
have been predicted on the basis of the free-metal ion concentration at
equilibrium. The experimental designs of the four studies are similar: a dose-
response curve is first determined in the absence of the ligand, and then the
experiment is repeated in the presence of the ligand. The free-metal ion
concentration in the presence of the ligand is either measured or calculated and
used to predict the biological response; in each case the metal proved to be
more 'available' than predicted from the dose-response curve.
Most of these apparent exceptions to the FIAM involve low molecular
weight metabolites as the ligands, e.g., amino acids or citrate. Daly et a/.61-62
determined the toxicity of Cu to the Australian freshwater shrimp. Paratya
mutraltensis. In the presence of increasing concentrations of NT A the total
concentration of Cu needed to kill the shrimp increased, but when expressed in
terms of the free Cu2* ion, the concentration required to kill 50% of the test
population (i.e., the LCjo(Cu2*) values) remained relatively constant over the
-------
72
P. O.C. CAMPBELL
73
4
I
i
3
8
a
1
„ c? S1
53 32 3?
2 ** c*. —
o ^* -*
8
I
u.a
0.4
^ °-3
•^ 0.2
i, 0>1
r* 0.0
N3
O.
A ' ' ' /
jfys s '
•
•
0.5 1.0 1.5
[Cu]T (fimol»L°)
•o
V
J2 0-5
0.3
0.2
0.1
0.0
B
0.5
1.0
1.5
lCu
Figure 10. (A) LCn values (OT the freshwater shrimp. Paraiya matralleiuls, exposed to
Cu in the presence of varied concentrations of NTA (D • 10'7 M NTA; • - 10~M M;
O -10-" M; • -10-* M; A -Cu only control, without NTA). (B) As above, but in
the presence of glycine (QJO-*M glycine; •-10-'" M; Q-10-"' M; »-10-»;
A- io-«' M; A-Cu control, without glycine). Results expressed in terms of [Cu'*J
" (Y-axis) and total Cu [CuJr (X-axis) for different ligund concentrations. [Modified from
Daly. H. R. el a/.. Environ. Toxleol. Chem.. 9. 997 (1990)]
range of NTA concentrations tested, as predicted by the FIAM (Figure 10A).
In the presence of glycine. however, the LCM(Cu2 *) values decreased fourfold
as the ligand concentration increased from 10"* to 2x tO~} M (Figure 10B).
The authors concluded thul the Cu-glycine* complex was 'mildly toxic*, but
they did not suggest a mechanistic basis for this toxicity.
Even more remarkable results were reported by Borgmann and Ralph.14 who
studied the complexation and toxicity of Cu to Daphnia magna in synthetic
media containing millimolar concentrations of the amino acids glycine.
0-alanine. or glutamic acid. In their experiments the ECw(Cu24) values
decreased 100-fold when the glycine concentration was increased from 0 to
8 x 10~4 M; similar but less marked decreases were observed in the presence of
0-alanine (20-fold decrease at [L}r-4x 10~J M) and glutamic acid (10-fold
-------
74 METAL SPECIATION AND BtOAVAlLABlUTY IN AQUATIC SYSTEMS
decrease at (L}r-2x I0~4 M). Borgmann and Ralph concluded that 'copper
toxicity to Daphnta is not simply a function of free-copper ion concentrations',
and suggested that Cu amino acid complexes somehow contributed to the
overall toxicity. la parallel experiments carried out on guppies (Poeellla
rettcula) in the presence of /f-alanine (4xlO~J M), the same authors
demonstrated that the LCjo (Cu2*) values were four times lower than in the
inorganic medium.
Similar data, indicating enhanced metal bioavailability in the presence of
citrate or ethylenediamine, were reported for an alga (Selenastntm
caprtcomutum) (Table 4, study 3). In this study Guy and Kean*7 determined
the effects of Cu on algal growth in the presence of a variety of hydrophilic
ligands. For the majority of the tested ligands (EDTA. NTA. triethylene-
tetramine, and bicine), algal growth ceased at the same pCu2* (s8.0), as
predicted by the FIAM. However, solutions containing 5/iM citrate or
ethylenediamine became algicidal at lower free Cu2* concentrations, i.e., at
pCu2* 9.8 and 8.65, respectively. In a second series of experiments the authors
determined algicidal Cu concentrations at different citrate or ethylenediamine
concentrations, and calculated Cu speciation for each algicidal solution.
Copper in the presence of citrate became algicidal at different pCu2* values (all
> 8), but at a constant pCuOHcit2" of 5.95. Similarly in Cu-ethylenediamine
solutions algal growth ceased at a pCu-en of 6.5 (Table 5). The authors
concluded that Cu-en2* and CuOHcit2' were 'toxic species'.
Similar anomalous metal behavior was demonstrated by Part and
Wikmark4* in their study of the effects of citrate on the flux of l09Cd across
perfused rainbow trout gills (Table 4, study 4). The uptake of Cd was first
determined in experiments run over a range of Cd concentrations in a synthetic
inorganic medium; free Cd2* concentrations were measured (ISE) and a dose-
response curve established Gog Cd flux vs pCd2*). Experiments were then run
in the presence either of EDTA (8/iM) or citrate (1 and 10 mM). In the first
case the influx of Cd was close to that expected for the free Cd2* concentration
calculated to be in equilibrium with EDTA (Figure 11). In the presence of
citrate, however, the flux was far greater than predicted from [Cd2*] (160 to
1000 times; see Figure 11); Cd accumulation in the gill tissue in these
experiments was also much greater than anticipated.
While the environmental relevance of these examples is debatable, in that the
ligand concentrations at which the enhanced metal bioavailability is observed
are far greater than normal environmental levels (e.g., millimolar rather than
submicromolar), they do provide a challenging test of the RAM. Possible
explanations of the 'anomalous* metal behavior include inadvertent changes in
the exposure medium as the ligand concentrations are manipulated (notably a
decrease in the free Ca2* levels), 'accidental* metal transport (the ligand is
animilated as a metaMigand complex), and the involvement of ternary surface
complexes (L-M-X-membrane) in the biological response.
' P. O. C CAMPBELL
•14
,T»-13
"S -12
o
§ -10
•08
0.00
75
8|iMEOTA
$
10mM citrate
ImMdtrata
Figure 11. Transfer of Cd [expressed in nmol/h (100 g fish)} through perfused rainbow
trout gills M • function of [Cd14] in the ventilatory water. The regression Hoe drawn
through the open circles (O) was determined in the absence of organic ligands. Transfer
values determined in the presence of 1 mM citrate (A.) or 10 mM citrate (•) lie below
this regression line, indicating that Cd influx is higher than predicted from the free Cd2 *
concentration. [Modified from Pan. P. and Wikmark, G.. Aquat. Toxicol., 5. 277
(I984).J
Table 5. Bioassays of Cu toxicity towards the green alga, Selenasinan eapricornutum. in
the presence of various ligands; Cu speciation as calculated for algicidal solutions
Ligand Cone
Ligand (jiM)
Ethylenediamine
Bicine
Q irate
2.5
5.0
10.0
2.5
5.0
10.0
5.0
10.0
pCu1*
8.54
8.65
9.11
8.0
8.0
8.0
9.8
9.9-
pCuL
6.53
6.49
6.53
6.29
5.87
5.43
5.98
5.95
pCuL,0
6.00
5.67
5.35
7.51
6.65
6.02
_
—
Dtu from O*y. R. D. ind Kan. A. H, Wtur Xn.. 14. S9I. I9SO.
Changes In the exposure medium
In none of the experiments summarized in Table 4 were efforts made to
maintain free Ca2* concentrations at a constant level. Citrate interacts strongly
with calcium (log K&
-------
76 METAL SPECIATION AND BIOAVAILABILITY IN AQUATIC SYSTEMS *
employed in both the alga! and fish gill studies equilibrium calculations
indicate that >90% of the Ca should be bound to citrate. Decreased
protection from Ca2* at high citrate concentrations could result in "enhanced
Cd or Cu availability. Indeed, in separate experiments Part and co-workers6*
manipulated (Ca2*] in the absence of citrate and demonstrated that a decrease
in [Ca2*] from 0.7 to OmM led to a 2.4-fold increase in the Cd flux. However.
this enhancement is still far less than that observed in the presence of citrate
(160 to 1000 times). Decreased competition from Ca2* thus provides only a
partial explanation for the anomalous behavior of Cu and Cd in the presence
of citrate, as reported in Table 4. Changes in free Ca2* concentrations cannot
be invoked to explain the enhanced availability of Cu in the presence of amino
adds or ethylenediamine.*6-*7 Complexation of Ca by these ligands is weak;
even at the millimolar concentrations employed no appreciable changes in
pCa2* are expected. Given the very high ligand concentrations used, however,
inadvertent contamination of the exposure media may have occurred (e.g., with
other trace metals).
Accidental transport
Given the central metabolic role played by amino acids and citrate, and the •
possible presence of specific membrane-bound transport systems designed to
facilitate the assimilation of these metabolites, it is tempting to speculate that
metals might traverse the biological membrane in the presence of amino acids
or citrate by 'accidental' transport4 In such cases, it is assumed that the
membrane-bound permease (P-cell) cannot distinguish between the free ligand
and the metal-ligand complex:
P. O. C. CAMPBELL
77
M-L -« M-L-P and'L-stereoisomcrs
would be expected to have different effects on metal bioavailability if
•accidental' transport occurred. There is an obvious parallel here with iron
nutrition in bacteria and certain algae (Chapter 1). wherein Fe bound to the
hydroxamic add moiety of siderophores is transported into the cell.
Formation of ternary surface complexes
As discussed in detail in Section 3.2.2. if mixed ligand complexes of the type
{L-M-X-cell} form at the biological surface (where -X-cell is a metal-sensitive
site), then the 'biological response* will vary as a function of the concentration
of ML in the exposure medium (see Equations (11) to (IS)). If both {M-X-cell}
and {L-M-X-cell} are present at the cell surface, the biological response will be
influenced by variations of both [M«*] and [ML]; if the response is additive,
then the metal will appear to'be more bioavailable than predicted on the basis
of the FIAM. The enhanced algiddal properties of Cu2* in the presence of
ethylenediamine and citrate*7 may thus be due to the contribution of such
surface complexes to the observed lexicological response. For certain algae it
has indeed been suggested that Cu exerts its algicidal action at the cell surface
without entering the cell." Similar reasoning may apply to the experiments
run with Daphnia magnet in the presence of amino acids, although data
suggesting that Cu can affect Daphnia without actually entering the animal are
lacking.
3b2.4 Problematic Examples (Table 6)
In compiling the list of documented 'exceptions' to the FIAM, two studies were
identified in which metal availability apparently increased in the presence of
hydrophilic chelators. relative to control exposures in the absence of added
ligand. Because metal spedation results for these experiments are unavailable,
and because they were long-term studies (3 weeks), they do not meet the criteria
for inclusion in this review. However, both studies are frequently died in the
literature (uncritically!), and thus it may be useful to point out their
shortcomings.
George and Coombs90 measured the long-term (21-day) uptake of ll5mCd by
the blue mussel, Mytiltu edulis. Experiments were first run in filtered seawater
in the absence of added ligands, and then in the presence of 'excess* ligands
(concentrations unspecified). Four ligands were investigated (humic add,
alginate. pectinate, and EDTA). In all four cases the authors reported that Cd
accumulation increased twofold as compared to the trials in filtered seawater.
Even more mysteriously, all four ligands gave the same response.
In the second example, Laube et a/.91 studied the effects of Cu. Cd. and Pb
on a blue-green alga, Anabaena sp., in long-term (20-day) laboratory
experiments. The bioassays were run in algal growth media that already
contained appreciable quantities of citrate and iron, to which equimolar
concentrations of the trace metal and NTA (10~* to 10-J M) were added. The
authors reported that 'NTA did not reduce, but in some cases even enhanced,
the toxicity of these metals for Anabaena'; no metal spedation data were
reported, however. The metal and NTA concentrations were not manipulated <
separately, but rather were increased together logarithmically at 'equimolar'
concentrations. The increased toxicity observed at moderate to high NTA
levels may have been caused by the presence of appreciable concentrations of
the free-metal ion in the'culture media, due to competition with other metals
-------
78
P. C. C CAMPBELL
79
edi
3
*>
I
s s
llJlfL
a
-------
80
P. C. C CAMPBELL
81
1
£
2L
8
i_
o
C
•s.
.s
I
1
5
O
1.4
»o U
• 1.0
§0.8
•)2 0.8
lo.2
0.0
• 10*A river water ~
• o 30% river water f
_* M%rlver water /
/
• X'.
e.o 7.0 8.0 9.0
pCu2*
Figure 12. Variation in the growth rate of the unicellular alga. Monoehrysa lutheri, as a
function of the free cupric ion activity in media containing 10% (•), 30% (O). «nd
90% (A) river water. [Modified from Sunda, W. G. and Lewis. J. A. M.. Umnol.
Ocemogr.. 23.870 (1978)]
euryhaline alga96 and a second with a marine bacterium.42 In the first study
Sunda and Lewis96 studied the effect of added Cu(II) on the growth rate of
Monochrysis lutheri. The standard growth medium contained TRIS buffer
(I mM) and 5% seawater. Filtered river water containing a high concentration
of DOM (22 mg L~' dissolved organic carbon, DOC) was added to the growth
medium in different proportions (10, 30, and 90% river water) in order to vary
the degree of Cu complexation. Frce-Cu ion concentrations, as measured
potentiometrically (Cu ISE), varied from 10-"-9 to 10"6-9 M (3.5 to 0.4% of
[Cu}r). The biological response followed the variations in [Cu2*] closely
(Figure 12), just as predicted by the FIAM. The authors concluded that 'the
decrease in copper toxicity with increasing complexation could be explained
quantitatively in terms of a dependency of toxicity on the concentration of the
free cupric ion*.
In a study of similar design, Sunda and Gillespie42 determined the response
of a bacterial isolate to added Cu in a low salinity medium containing different
proportions of DOM-rich river water. Experiments were also performed in
synthetic media containing varied quantities of commercially available humic
substances. In both cases the biological response (inhibition of cellular
incorporation of l4C-glucose) followed the measured variations in [Cu2*], as
determined potentiometrically (Figure 13).
A final example conforming to-the FIAM was reported by Meador,95 who,
determined the acute toxicity of Cu(II) to neonate Daphnia magna in buffered
synthetic media containing different proportions of a solution containing
natural DOM. This DOM was derived from 'microcosm media which had
experienced algal and daphnid blooms' and contained selling L~' DOC
-------
82 METAL SPEC1AT1ON AND BIOAVA1LAB1LITY IN AQUATIC SYSTEMS
g f.OTO
Q.
.& 8.000
a>
•S 5-0
-------
84
P. G. C. CAMPBELL
85
X
I
1
j3
.92
?Sa »-
SSbVJU
plifl
«|I8|
Hlt1IJt1
sS-3^S-5J!?S
The authon suggest that pond water alone could not have been the source or
toxicity because the test organism occurred naturally in the pond and no
mortality was observed in the Cu-free control experiments. These experimental
results obviously conflict with the predictions of the FI AM, according to which
LC5o(M*+) values should not vary from one exposure medium to another (cf.
Section 3.3.1). The 6-h bioaccumulation data are consistent with the bioassay
results: for similar concentrations of free Cu2* (3.74 and 4x 10-' M), Cu
accumulation was approximately twofold highe{ in the organisms exposed to
pond water (7.4±1.5x 10~J mol mg~t Cu) than in those held in well water
(3.6 ± 2.2 x 10~s mol rag-1 Cu). The authors conclude that although natural
DOM reduced the overall toxicity of Cu to S. serrvlatus (presumably by
complexing Cu and reducing the free-ion concentration), it also exerted a more
subtle effect and 'facilitated* or 'enhanced* the bioavailability of the'residual
Cu2*.
Similarly provocative results were reported by Borgmann and Charlton,97
who determined the toxicity of Cu to Daphnia magna in artificial media and in
samples of natural surface waters (Lake Ontario; Hamilton Harbour). Free
Cu2* concentrations were measured potentiometrically (ISE) and also
calculated with an ingenious bioassay technique;16 agreement between the
two techniques was good. Both the ISE and bioassay data indicated that free
cupric ion concentrations needed to provoke the same toxic response differed
in different test waters; the ECjo(Cu2*) values measured in the natural waters
were markedly lower (ca. fourfold) than in the synthetic test medium. The
authon suggest that Cu complexes with naturally occurring organic ligands
could have contributed to the observed toxicity. They conclude that free-metal
concentrations 'do not provide a good measure of copper toxicity towards
Daphnia in natural waters'. In defense of the FIAM it should be noted that the
natural water samples were analyzed for the major anions and cations, but not
for trace metals or organic micropollutants. The aggravated response to Cu2*
in the natural water samples may have reflected an additive interaction between
Cu and other unidentified 'stressors* present in the natural waters but absent
from the inorganic test solutions.
Anomalously high metal toxicity in natural waters was also reported by
Laegreid et a/.91 in their study of the seasonal variation of Cd toxicity toward
the freshwater alga, Sctcnastnm capricornutum, in two Norwegian lakes with
different levels of humic substances. Water samples were collected from a small
dystrophic bog lake (Lake Lille Bakketjern: pH 4.4,18 to 25 mg L~' DOC) and
a eutrophic, moderately humus-influenced lake (Lake Gjersjeen: pH 7 to 10,6
to 8mg L~' DOC). Samples were filtered and, in the case of the acidic bog
lake, the pH wfts raised to 7.5. After equilibration with clean air to adjust the
Pcoi the lake waters were enriched with inorganic nutrients and spiked with Cd
(3.8 x 10-* to 3.6 x 10~* M total Cd). An ion-selective electrode was used to
estimate free Cd2* levels. Note, however, that the spiked metal concentrations
-------
86 METAL SPEC1AT10N AND BIOAVAILAB1LITY IN AQUATIC SYSTEMS
gco
« 80
7 ft 5
pCd1*
Figure 15. Seasonal variation in the toxicity of Cd to the unicellular green alga.
Selenastrum capricornutum, as a function of the free-Cd ion activity. Toxicity is
expressed as percent inhibition of photosynthetic l4COj uptake by the test alga.
Bioassays were run in filtered, pH-adjusted water from Lake Bakketfem; samples were
collected in March (•). May (A). June (O). «nd September (•) 1981. [Modified from
Laegreid. M. tl al.. Environ. Set Technol.. 17. 357 (1983)]
were close to or below the expected lower limit of the operating range of the
electrode (see below). After a further 24 h equilibration, exponentially growing
cells of the test alga were transferred to the test solution. Finally, after 24 h of
Cd exposure, subsamples were withdrawn from the exposure flasks and
incubated with "CO* The toxic response was expressed as the reduction in I4C
uptake relative to a Cd-free control.
Bioassay results for pH-adjusted samples from the dystrophic bog
conformed to FIAM predictions. The total Cd concentration causing 50%
inhibition of MC uptake varied twofold (7.2 to 15xlO~T M). However,, if
percent inhibition was plotted as a function of the free Cd2* concentration, the
points from four separate samples (March, May, June, and September) fell
nicely on a single curve (Figure IS), and a single ECjo(Cd2*) value could be
extracted (10~M M).
In contrast, samples from the eutrophic Lake Gjersjeen showed a
pronounced seasonal variation in Cd toxicity. Because ISE measurements
proved unreliable for samples from this lake, the bioassay results were
expressed in terms of total Cd. Plots of ECjo(CdT) values vs time showed
marked seasonal variations; over three successive annual cycles, Cd was
substantially more toxic during summer months (Figure 16). Note that the
ECxXCd-r) values obtained for the summer samples (a 10~7 M) are not only
much lower than those determined during the nonsummer months, but are also
eight to ten times lower than the values determined in synthetic algal culture
medium and expressed as Cd2* (FRAQUIL: EC*(Cd2*)- 8 to 10 x 10'7 M).
The authors conclude that 'the assumed connection between the free-metal ion
activity and toxicity is not valid', and suggest that certain low molecular weight
organic ligands may increase metal uptake and toxicity. Once again, however.
P. O. C CAMPBELL
100
87
~* 10
o
I
0.1
o
UJ
0.01
MJ JASON 0|JFMAMJJASONO|JFMAMJJ A3 ON
1980 1981 1982
Figure 16. Seasonal variation in the toxicity of Cd to the unicellular green alga.
Stltnastrum caprtcornutum, as a function of the free-Cd ion activity. The EC* values for
inhibition of photosynthetic "COj uptake are expressed in terms of total Cd (/iM).
Bioassays were run in filtered water from Lake Gjersjeen. The shaded area represents
the estimated concentration of free Cd2* causing 50% inhibition in synthetic lake water.
[Modified from Laegreid. M. tl al.. Environ. Set. Technol.. 17. 357 (1983)]
we must qualify these results with the caveat that the anomalous summer
samples were not exhaustively characterized. The enhanced toxicity of Cd2* in
these samples may have reflected an additive interaction between Cd and other
unidentified 'stressors* present in the summer months but absent during the
other seasons. The distinct seasonal periodicity suggests that these factors may
have been of biological origin (e.g., algal exudates).
In addition to the three studies described above, there are several other
reports of 'enhanced* metal bioavailability in the presence of natural
DOM.*-'02 Technically, these studies did not qualify for inclusion in Table 8
because the results were reported in terms of the total metal concentration,
with no attempt to determine metal speciation in the exposure media.
Nevertheless, given the surprising nature of the observations, it seems
appropriate to attempt a brief summary.
One of the earliest reports of anomalous metal behavior in the presence of
natural DOM is that of Giesy and collaborators.9' who fractionated the
natural organic matter from a humic pond water using ultrafiltration and then .
determined the effects of these fractions on Cd toxicity to a freshwater
cladoceran, Simocephalus serrulaiia. Note that the test organism and the humic
pond are identical to those studied by Giesy el al.n in their study on Cu
toxicity (Table 8). LCj^Cdj) values were determined in artesian well water
-------
88 METAL SPECIATION AND BIOAVAILABIL1TY IN AQUATIC SYSTEMS
(low DOC), in whole pond water, and in well water containing the individual
DOM fractions at concentrations simitar to those at which they occurred
naturally in the pond water. Pond water, and the individual organic Tractions
added to the well water, bound added Cd, reducing the amount or Tree Cd*f as
determined potentiometrically (Cd ISE). As expected, whole pond water and
the three DOM fractions of higher molecular weight reduced Cd toxicity as
compared to controls run in well water. In contrast, the low molecular weight
fraction (nominal molecular weight < 500) caused a slight increase in Cd
toxicity; the LCso(CdT) decreased from 6.2 to 3.1 x 10~* M in the presence of
this DOM fraction. Note that mortality in controls containing the low
molecular weight fraction (but no Cd) was 20% (cf. 0 to 10% for the other
controls), suggesting that some natural components of this fraction contributed
to the observed toxicity. The authors suggest that 'it is possible that organics
alter Cd toxicity by a primary or secondary interaction with organisms as well
as Cd binding*.
Anomalous results were reported also by Winner and co-workers10*-'02 in
their extensive study of the effects of a commercial humic acid (HA, Aldrich
Chemical Co.) on the toxicity and bioaccumulation of Cd, Cu, and Zn. Two
test species, Daphnla magna and D. pulex, were used; measured biological
responses included mortality (48-h exposure, acute toxicity), long-term
survival, and abortion rates (42-day exposure, chronic toxicity). Exposure
variables included the humic acid concentration and water hardness.
Unfortunately, no attempts were made to determine metal speciation in the
exposure media and all results are necessarily reported in terms of 'nominal'
(i.e., added) metal concentrations.
For Cu and Zn. the toxicity results were qualitatively in agreement with the
FIAM. LQotMr) values increased in the presence of humic acid, and long-
term survival was prolonged (Figure 17A, B). In the case of Cd, however, the
results were counter-intuitive. Doth the ncute and the chronic toxicities of Cd
increased in the presence of 0.75 or 1.5mg L-' HA (Figure I7Q.100 In a
follow-up study, carried out at the same two humic acid concentrations but
three different levels of water hardness (58. 116. and 230mg L-' CaCOj).
Winner101 monitored both the long-term survival of the daphnids and their
abortion rate (% abortions/brood). The anomalous effect of humic acid on the
Figure 17. (opposite) Survivorship curves for daphnids exposed to Cd, Cu, or Zn at
"different humic acid concentrations in synthetic media having an alkalinity and a
hardness of lOOmg L"1 as CtCOj. (A) Exposure of Daphnla pulex to Cu (0.47 pM)
at humic acid concentrations of 0 (A A. 0.38 (•—•), 0.75 (O O), and 1.5
(A—A)nig L-'. (B) Exposure of D. magna to Zn (1.9pM) at humic acid
concentration* of 0 (A- - -A) »nd 1-5 (A—A)mg L'1. (C) Exposure of D. jmltx to
Cd (O.I8^M) at humic acid concentrations of 0 (A A). 0.75 (O O), and 1.5
(A—A)nig L-'. JModified from Winner, R. W., Aouat. Toxteal.. 5. 267 (1984) and
from Winner, R.W. and Gau«. J. D., Aouat. Toxleol., 9,149 (I986)J
P. O. C. CAMPBELL
89
,00rt—t \! *
75 - "• v—-•"
w
c 50
0>
u
e
w:
1.5mg«L-' HA
A(Cu)
• 0 mg»L-1 HA \
b.o
0.38, Vo^mg^HA
--i-L'1 HA 8
0 5 10 15 20 25 30 35 40 45 50
exposure (days)
100
1
1
H 50
0)
0
§.
O
~t \ 1.5mg«LMHA-*-
;U
; 1 i
A— 1 --^ -A
i N, i i i i i^
"^ A
A\
AA_
B(Zn)
i •
•?
•d
-
0
exposure (days)
too
'JE
^ 50
O
u
^V-Q
> **• b*-^o
1.5mg.L" HA %*\V"—*_0 mg-L' HA
\ \ ^ ^ ^
*-6/*0.75 mg-L1 HA 'A-A
0 5 10 15 20 25 30 35 40 45 50
exposure (days)
-------
90 METAL SPECIATION AND BlOAVAtLABILITY IN AQUATIC SYSTEMS
chronic toxicity of Cd was again evident; threshold or 'no-effect*
concentrations decreased in the presence of HA. the effect being more
evident in the harder waters. No mechanism was suggested to explain these
results.
333 Examples of Enhanced Protection in the Presence
of Dissolved Organic Matter (Table 8)
In the preceding section we have compiled and critically reviewed a series of
•exceptions' to the FIAM, in which metal bioavailability in the presence of
DOM is greater than it should be. This final section considers a solitary
example in which the opposite trend is observed, i.e., one in which metal
bioavailability in the presence of DOM is less than predicted from the
biological response curve (response vs [M*+D. determined in the absence of
DOM.
In experiments similar in design to those described in Section 3.2.3, Daly et
a/.94 determined the toxicity of Cu to the Australian freshwater shrimp.
Paratya aiutrallensis, in three DOM-rich natural waters. Laboratory tap water
and distilled water were mixed with each natural water in order to obtain
similar water qualities for each experiment (DOC levels of 7 to 8 and 12 to
13mg L-'; pH 7; hardness 14 to 17mg L~' CaCOj). As expected, the natural
organic matter reduced the overall toxicity of the added Cu. However, when
the LC» values were expressed in terms of the free Cu1* ion (ISE
measurements), they were consistently higher in the presence of the natural
DOM than in the inorganic (tap water) medium; i.e., for a given Cu2*
concentration, copper was lens toxic than anticipated.
Exponin Medium* LC*(Cu>4)(10-'M)
Tap water
Redwater Creek
1:7
1:3
Errinundra Dun
1:3
1:1
Inkpot
I*
1:4
2.5*0.3
3.3 ±1.1
4.4*0.9
5.2 ±1.3
8.5 ±1.4
11.1 ±2.2
ll.6±1.4
•DiMkm factor I put Mhml water lo V parti dbtlltaJ * Up wiur.
Interpretation of these experiments is not straightforward. Taken at face value,
the results suggest that natural DOM exerts both an indirect effect (involving
complexation of Cu in solution and reduction of the free metal ion
concentration), and a direct beneficial effect on the test organism (possibly
P. O. C CAMPBELL
91
involving the interaction of DOM at the biological surface). As noted by Daly
et at., however, use of ion selective electrodes in natural waters is often
complicated by unstable electrode potentials and relatively long response times
(see Chapter 5). In the present study the cupric ion ISE was calibrated in
ultraviolet-irradiated water, free of organic matter, before being used to
measure Cu2* activities in the diluted natural waters. Daly et al. speculate that
the electrode may have responded in some manner to organic matter present in
the natural waters, leading to an overestimate of the true cupric ion activity.
4 CONCLUSIONS
Within the general context of metal-organism interactions, this chapter has
explored links between metal spcciation in the external environment and metal
'bioavailability', using the FIAM as a convenient paradigm (the main
assumptions that underpin the FIAM are summarized in Table 9). The
literature on metal-organism interactions has been examined from a
quantitative point of view to determine to what extent the FIAM can
explain the available data. Particular attention has been accorded to apparent
'exceptions' to the FIAM, as these cases can be taken to define the limits of the
possible application of the model to the natural environment.
The three major conclusions of this chapter are summarized below, together
with suggestions of possible research opportunities.
STUDIES PERFORMED IN THE ABSENCE OF NATURAL DOM
Various responses were studied in these experiments, ranging from highly
specific (e.g., metal accumulation per se: surface adsorption, absorption,
subcellular distribution) to more holistic (e.g., growth, motility. mortality). In
experiments performed at constant pH and hardness, and in the presence of
synthetic ligands forming hydrophilic metal complexes, the biological response
consistently varies as a function of the concentration of the free-metal ion, as
predicted by the FIAM (52/59 cases: Tables 1. 3. and 4).
Documented examples of experiments in which changes in metal
bioavailability do not conform to the FIAM are relatively few in number.
These 'exceptions' fall into three main groups: (I) experiments with ligands'
forming hydrophobic metal complexes; (2) studies with ligands forming
hydrophilic metal complexes, in which the pH and/or water hardness was
varied; and (3) studies with ligands forming hydrophilic metal complexes,
usually of low molecular weight and often metabolites.
-------
92 METAL SPEC1ATION AND BIOAVAILABILITY IN AQUATIC SYSTEMS
Table 9. Assumptions inherent in the FIAM* and their implications
Condition
I. Plasma membrane is the primary site
for metal interactions with living
organisms; interaction with the
plasma membrane can be described
as a surface complexation reaction,
forming MOC-cell.
2. Metal transport in solution, toward
the membrane, and the subsequent
surface complexation reaction occur
rapidly ("rapid*-faster than metal
uptake, faster than the expression of
the biological response).
3. Biological response, whether it be
metal uptake, nutrition, or toxtcity. is
directly dependent on the
concentration of the M-X-cell surface
complex; in the range of metal
concentrations of lexicological
interest, variations in (M-X-cell)
follow those of IM'*J in solution.
4. During exposure to the metal of
interest, nature of the biological
surface remains constant
Implication
Negligible diffusion of the free-metal ion,
M1*, or its charged (hydrophilic)
complexes, ML*, occurs across the
plasma membrane barrier.
(Pseudo-)equilibrium established between
metal species in solution and those at the
biological surface; corollary- identity of
the metal form(s) reacting with the cell
surface is of no biological significance; no
one species in solution can be considered
more (or less) available than another.
FIAM more likely to apply to metals with
rapid complexation kinetics (e.g., Cd. Cu,
Mn, Zn) than to those exhibiting slower
reaction rates (e.g., Fe(HI). Cr(Ill), Al.
,Ni).
Hardness cations. Ca1* and Mg1*, and the
hydrogen ion, H*, will compete with
metal M for binding at the surface
complexation site.
FIAM will not apply to metals that form
stable ternary surface complexes, (L-M-
X-cell); if such complexes do form, and if
they contribute to the biological
response, then simple dependence of the
response on [M") is no longer observed.
Concentrations of free binding sites, {'X-
cell}, remains reasonably constant over
the biologically relevant range of metal
concentrations.
FIAM to more likely to apply short-term
(acute) exposures than to long-term
(chronic) experiments.
'Although we refer to the Free-ion HtMiy model, ttmott without exception the experimenul dm comta of
mtul mftrmm/fMi. In effect, it wovU tppetr thit tny fanpreciilen Introctaccd by rwjteetlnj activity coefficient
viriiUon* within • given experiment it neitifiMt in eomptrton with the Inherent bwlofletl vuiability.
• Group I. The literature review yielded several examples of enhanced metal
availability in the presence of ligands forming lipophilic metal complexes
(Table 2). Enhanced availability in such circumstances is consistent with the
principles of membrane transport developed in Chapter 1. It should be noted
that the ligand concentrations needed to provoke enhanced metal uptake
rates normally exceed those likely to be found in the environment. The
P. O. C. CAMPBELL
93
results may still be environmentally relevant as examples of the behavior of
low molecular weight, lipophilic metal complexes, if such complexes can be
shown to exist in natural waters. Are the dncumcnted exceptions of academic
Interest only, or ore they valid analogues for naturally occurring lipophilic
complexes?
• Group 2. From studies performed on freshwater organisms under conditions
in which the pH and/or the water hardness varies, it is clear that knowledge
of the free-metal ion concentration alone is insufficient to predict the
biological response. Potential competition for the metal binding site, ~X-
cell. by the hydrogen ion, H*, and by the hardness cations. Ca2* and Mg2*,
must be considered. Evidence for such competition remains largely
qualitative, however. Experiments are needed to test the hypothesis that the
effects ofpH and hardness are due solely to competition for binding sites at the
biological surface.
• Group 3. Without exception, the complexation of a metal by low molecular
weight, hydrophilic ligand leads to a decrease in its bioavailability. However,
in a limited number of cases (Table 4) the 'residual' bioavailability of the
metal in the presence of the ligand is greater than would have been predicted
on the basis of the free-metal ion concentration at equilibrium. Many of
these apparent exceptions to the FIAM involve low molecular weight
metabolites as the ligands, e.g., amino acids or citric acid. These examples
provide a challenging test for the FIAM. although as was the case for Group
1, their environmental relevance is debatable; the ligand concentrations at
which enhanced metal bioavailability is observed are far greater than normal
environmental levels (e.g., millimolar rather than submicromolar
concentrations). Possible explanations for the 'anomalous* metal behavior
include inadvertent changes in the exposure medium as the ligand
concentrations are manipulated, most notably, a decrease in the free Ca2*
levels; 'accidental' metal transport, in which the ligand is assimilated as a
metal-Iigand complex and the metal simply 'accompanies* the ligand; and the
involvement of ternary surface complexes (L-M-X-membrane) in the
biological response. These explanations should be tested experimentally.
STUDIES PERFORMED IN THE PRESENCE OF NATURAL DOM
Probably the most striking observation concerns the scarcity of studies suitable
for testing the applicability of the FIAM in the presence of DOM. There are
numerous reports in the literature of the effects of DOM on metal
bioavailability. but virtually all of these studies are qualitative in nature (i.e.,'
metal speciation is undefined). The few quantitative studies (Tables 7 and 8) are
more or less evenly divided between examples that conform to the predictions
of the FIAM and examples that appear to be in contradiction; i.e., unlike the
studies performed in the absence of DOM. no consensus is evident for the
-------
W METAL SPEC1ATION AND BIOAVAILABILITY IN AQUATIC SYSTEMS
experiments run in the presence of DOM. The applicability of the Ft AM to
natural voters remains to be demonstrated.
•
ENVIRONMENTAL ANALYTICAL CHEMISTRY NEEDS
Copper has been studied far more extensively than the other metals. Citations
decrease in order Cu>Cd>Zn» all others (Tables 1-4.7, and 8); references
to trivalent metals of environmental importance (e.g., Al1* or Fe**) are
conspicuously absent dearly, the availability of a cupric ISE that can be used
at environmentally relevant metal concentrations has greatly influenced
research in this area. Future progress would be greatly aided by the
development of analytical methods capable of measuring the free-ton
concentrations of metals other than Cu at levels
-------
96 METAL SPECIATION AND BIOAVAILABIL1TY IN AQUATIC SYSTEMS
CMTA: carboxymcthyltartronic acid
DDC: diethytdithiocarbamate
DOC: dissolved organic carbon
DOM: dissolved organic matter
ECyf. (metal) concentration needed to provoke a 50% response in a bioassay
experiment (e.g., 50% growth inhibition)
EDTA: ethylenediaminetetraacetatc
FIAM: free-ion activity model
FRAQVJL: defined culture medium for freshwater algae
HA: humic acid
ISE: ion-selective electrode
KEX: potassium ethylxanthate
KAX: potassium amylxanthate
kf. rate constant for the reaction of a metal with a surface binding site, -X-cell,
to form the surface complex M-X-ccll
kj: rate constant for the dissociation of the surface complex, M-X-ccll
ky,,: rate constant for the internalization of a metal in cases in which the surface
complex, M-X-cell, corresponds to a site for transport of the metal across the
plasma membrane
L: ligand (in solution; charges not shown)
LCyf (metal) concentration needed to kill 50% of the test population in a
bioassay experiment
LCjofM'*): LCjo, expressed in terms of the free-metal ion concentration
LCX(MT)\ LCjo, expressed in terms of the total metal ion concentration
Mt: metal adsorbed at (algal) cell surface
Mt\ intracellular metal
M'+\ free-metal ion (-M(HjO)S+)
pM'*: the negative logarithm of the free-metal ion concentration, expressed in
molarity(molL-')
[MJ-f. total concentration of metal M (in solution)
ML: aqueous complex, involving metal *M' and ligand *L' (charges not shown)
MES: 2-(W-morpholino)-ethancsulfonic acid
ft MR: nuclear magnetic resonance spectrometry
NTA: nitrilotriacetate
ODS: oxydisucctnic acid
TRIS: /r/j-hydroxymethylaminomcthane
X-cell: ligand (at cell surface)
-X-cetl: free (unbound) ligand at cell surface
M-X-cell: surface metal complex, involving metal 'M', and surface ligand
'X-cell* (charges not shown)
P-cell: membrane-bound permease
IK specific growth rate (usually for phytoplankton)
( ]: concentration of aqueous species
{ }: concentration of surface species
P. G. C. CAMPBELL
REFERENCES
97
I. Sunda, W. 0. (1991). Trace metal interactions with marine phytoplankton. Biot.
Oceamw.,6,411.
2. Jain, M. K. (1988). Introduction lo Biological Mtmhranes, John Wiley & Sons, New
York, Chap. 2.
3. Williami R. J. P. (1981). Physico-chemical aspects of inorganic element transport
through membranes, Philos. Trans. R. Soc. London B. 294, 57.
4. Luotna, S. N. (1983) Bioavailability of trace metals to aquatic organisms—a
review, Scl. Tot. Environ., 28, 1.
5. Turner, D. R. (1986). Biological availability of trace elements. In Blogeochemical
Processes at the Land-Sea Boundary, ed. Lasserre. P. and Martin. J. M., Elsevier.
Amsterdam, p. 191.
6. Hunt, D. T. E. (1987). Trace Metal Speclollon and Toxicily to Aquatic Organisms—
A Review. Environ Rep. TR 247, Water Research Centre, Marlow, Bucks. UK.
7. Lewis. A. G., Whitfield, P. H.. and Ramnarine, A. (1972). Some paniculate and
soluble agents affecting the relationship between melal toxicity and organism
survival in the culanoid copcpod Kurhaeta Japonirii. Mar. Hial., 17. 215.
8. Sillen, L. G. (1962). High-speed computers as a supplement to graphical methods. I.
Functional behavior of the error square sum, Ada Chem. Scand., 16, 159.
9. Morel, F., Morgan, J. J. (1972). A numerical method for computing equilibria in
aqueous chemical systems. Knvlron. Scl. Techno!., 6. 58.
10. Nordstrom. D. K., Plummcr, L. N., Wigley. T. M. L., Wolery, T. J.. Ball. W..
Jcnne. E. A., and Bassett. R. L. E. A. (1979). A comparison of computerized
chemical models for equilibrium calculations in aqueous systems. In Chemical
Modeling in Aqueous Systems—Spcciation. Sorption. Solubility and Kinetics, ed.
Jenne, E. A., ACS Symp. Ser. No. 93. American Chemical Society. Washington,
DC, p. 857.
II. Walsh, G. E. (1988). Principles of toxicity testing with marine unicellular algae,
Environ. Toxlcol. Chem., 7, 979.
12. Anderson, M. A.. Morel. P. M. M.. and Guillard. R. R. L. (1978). Growth
limitation of a coastal diatom by low zinc ion activity. Nature, 276, 70.
13. Morel, F. M. M., Reuter, J. G., Anderson. D. M., and Guillard. R. R. L. (1979).
AQUIL: a chemically defined phytoplankton culture medium for trace metal
studies. J. Phycol., 15. 135.
14. Price. N. M.. Harrison. G. I., Hering, J. G.. Hudson. R. J.. Nirel. P. M. V. Palenik.
B., and Morel. F. M. M. (1991). Preparation and chemistry or the artificial algal
culture medium AQUIL, Blot. Oceanogr., 6, 443.
15. Allen. H. E.. Hall. R. H., and Brisbin. T. D. (1980). Metal speciation. Effects on
aquatic loxicity. Environ. Set. Techno!., 14,441.
16. Payne. A. G. (1975). Responses of the three test algae of the algal assay procedure:
bottle test. Water Res., 9,437.
' 17. Bates. S. S.. Tessier. A.. Campbell. P. G. C, and Hurtle, J. (1982). Zinc adsorption
and transport by Chlamydomonas variabllis and Scenedesmus subsplcatus
(Chlorophyceae) grown in semicontinuous culture, /. Phycol., 18, 521.
18. Morel, F. M. M. (1983). Principles of Aquatic Chemistry. Wiley-lntencicnce, New
York, 301.
19. Nelson. P. 0., Chung, A. K... and Hudson, M. C. (1981). Factors affecting the fate
of heavy metals in the activated sludge process, J. Water Pollut. Contr. Fed., 53.
1323.
20. Petersen, R. (1982). Influence of copper and zinc on the growth of a freshwater alga
-------
98 METAL SPECIAT10N AND BIOAVAILABIUTY IN AQUATIC SYSTEMS
Scenedesmus qvadrlcaudo—lhe significance of chemical speciation. Environ. Set.
Technol.. 16.443.
21. Morel, F. M. M., and Morel-Laurens, N. M. L (1983). Trace metals and
phytoplankton in the oceans; facts and speculations. In Trace Metals In Sea Water,
ed. Wong, C. S., Boyle. E., Bruland. K. W., Burton, J. D., and Goldberg, E D..
Plenum Press. New York. 841.
22. Sunda, W. C., and Huntsman, S. A. (1983). Effect of competitive interactions
between manganese and copper.on cellular manganese and growth in estuarine and
oceanic species of the diatom Thalassloslra. Umnol. Oceanogr., 28.924.
23. Pagenkopf. C. K. (1983). Gill surface interaction model for trace-metal toxicity to
fishes: role of complexation, pH and water hardness. Environ. Set. Techno!., 17,347.
24. Harrison. G. I., and Morel, F. M. M. (1983). Antagonism between cadmium and
iron in the marine diatom Thalassloslra vtetujlogtt, J. Phycol., 19,495.
25. Gavis, J. R.. Guillard, R. L. L.. and Woodward. B. (1981). Cupric ion activity and
the growth of phytoplankton clones isolated from different marine environments, J.
Mar. Ret., 39, 315.
26. Turner, D. R. (1984). Relationships between biological availability and chemical
measurements. In Circulation of Metals In the Environment, Metal Ions in Biological
Systems. Vol. 18, ed. Siege!. H. Marcel Dekker. New York, Chap. 5.
27. Peterson, H. G., and Healcy. F. P. (198S). Comparative pH dependent metal
inhibition of nutrient uptake by Scenedesmu quadrlcauda (Chlorophyceae). J.
Phycol., 21, 217.
28. Morel, F. M. M. (1986). Trace metals—phytoplankton interactions: an overview.
In Biogeochemtcal Processes at the Land~Sea Boundary, ed. Ltssem, P., and
Martin. J. M. Elsevier. Amsterdam, 177.
29. Hudson, R. J. M.. and Morel. F. M. M. (1991). Iron transport in marine
phytoplankton: kinetics of cellular and medium coordination reactions, Umnol.
Oceanogr., 35,1001
30. Hudson, R. J. M.. and Morel. F. M. M. (1993). Trace metal transport by marine
microorganisms: implications of metal coordination kinetics, Deep-Sea Ret., 40,
129.
31. Morel, F. M. M., and Hering, J. G. (1993). Principles of Aquatic Chemistry—2nd
Edition, Wtley-Interseience. New York.
32 Tessier. A.. Buffle, J., and Campbell. P. G. C. (1994). Uptake of trace metals by
aquatic organisms. In Chemical and Biological Regulation of Aquatic Systems, ed.
Buffle. J.. and De Vitro, R., Lewis. Boca Raton, FL, Chap. 6.
33. Davies. A. G. (1970). Iron, chelation and the growth of marine phytoplankton. I.
Growth kinetics and chlorophyll production in cultures of the euryhaline flagellate
Dunallella tertlolecta under iron-limiting conditions, J. Mar. Blol. Assoc. UK, 50,
65.
34. Mason. A. Z.. Jenkins. K. D.. and Sullivan. P. A. (1988). Mechanisms of trace metal
accumulation in the polychaete Neanthes arenaceodentata, J. Mar. Blol. Assoc. UK,
68. 61
35. Blust, R.. Fontaine. A., and Dclcdr, W. (1991). Effect of hydrogen ions and
inorganic complexing on .the uptake of copper by the brine shrimp Anemia
franclscana. Mar. Ecol. Prog. Ser* 76, 273.
36. Campbell, P. G. C. and Stokes, P. M. (1985). Acidification and toxicity of metals to
aquatic biota. Can. J. Fish. Aquat. Scl., 42,2034.
37. Sunda, W. G., and Huntsman, S. A. (1992). Feedback interactions between zinc and
phytoplankton in seawater, Umnol. Oceanogr., 37,25.
P. G. C. CAMPBELL
99
38. Sunda. W. G.. and Guillard, R. R. L. (1976). The relationship between cupric ion
activity and the toxicity of copper to phytoplankton. J. Mar. Res.. 34. 511.
39. Anderson, D. M., and Morel, F. M. M. (1978). Copper sensitivity of Gonyaulax
tamarensis. Umnol. Oceanogr., 23, 283.
40. Sunda, W. G.. and Huntsman. S. A. (1986). Relationships among growth rate.
cellular manganese concentrations and manganese transport kinetics in estuarine
and oceanic species of the diatom Thalasslora, J. Phycol., 22. 259.
41. Canterford. G. S.. and Canterford, D. R. (1980). Toxicity of heavy metals to the
marine diatom Dltylum brightweltil (West) Grunow: correlation between toxicity
and metal speciation, J. Mar. Blol. Assoc. UK. 60. 227.
42. Sunda, W. G., and Gillcspie. P. A. (1979). The response of a marine bacterium to
cupric ion and its use to estimate cupric ion activity in seawater. J. Mar. Res., 37,
761.
43. Sunda, W. G.. and Ferguson, R. L. (1983). Sensitivity of natural bacterial
communities to additions of copper and to cupric ion activity: a bioassay of Cu
complexation in seawater. In Trace Metals In Sea Water, ed. Wong. C. S., Boyle. E.,
Bruland. K. W., Burton. J. D.. and Goldberg, E. D. Plenum Press. New York.
p. 871.
44. Jonas, R. B. (1989). Acute copper and cupric ion toxicity in an estuarine microbial
community. Appl. Environ. Microblol., 55.43.
45. Sunda, W. G.. Engel, D. W.. and Thuotte. R. M. (1978). Effect of chemical
speciation on toxicity of cadmium to grass shrimp, Palaemonetes puglo: importance
of free Cd ion. Environ. ScL Technol., 12.409.
46. Engel, D. W., Sunda, W. G., and Fowler, B. A. (1981). Factors affecting trace metal
uptake and toxicity to estuarine organisms. I. Environmental parameters. In
Biological Monitoring of Marine Pollutants, ed. Vemberg, F. J., Calabrese, A..
Thurberg, F. P.. and Vemberg, W. B., Academic Press, New York. 127.
47. Jenkins, K. D., and Saunders, B. D. (1986). Relationship between free cadmium ion
activity in seawater, cadmium accumulation and subcellular distribution, and
growth in polychaetes. Environ. Health Pert peel., 65. 205.
48. Zunuda, C. D.. and Sunda. W. C. (1982). Bioavailability of dissolved Cu to the
American oyster Crassostrea vlrglnlcd. I. Importance of chemical speciation. Mar.
Blot., 66,77.
49. Sanders, B. M., Jenkins. K. D.. Sunda. W. G.. and Costlow. J. D. (1983). Free
cupric ion activity in seawaten effects on metallothionein and growth in crab larvae.
Science. 222. 53.
50. Sanders, B. M.; and Jenkins, K. D. (1984). Relationship between free cupric ion
concentrations in sea water and copper metabolism and growth in crab larvae. Blol.
Bull., 167,704.
51. Blust, R.. Verheyen. E.. Doumen. C, and Decleir. W. (1986). Effect of
complexation by organic ligands on the bioavailability of Cu to the brine shrimp,
Aquat. 7V>x/c0/..8.2II.
52. Manahan, S. E. and Smith, M. J. (1973). Copper micronutrient requirement for
algae. Environ. Scl. Technol., 7,829.
53. Hall, J., Healey, F. P.. and Robinson, G. G. C. (1989). The interaction of chronic
copper toxicity with nutrient limitation in chemostat cultures of Chlorella, Aquat.
Toxlcol., 14. IS.
54. Spencer, D. F., and Nichols, L. H. (1983). Free nickel ion inhibits growth of two
species of green algae, Environ. Polhit., A-31.97.
55. Harrison. G. I., Campbell. P. G. C., and Tessier. A. (1986). Effects of pH changes in
-------
1 00 METAL SPEC1 ATION AND BIOAVAILABIUTY IN AQUATIC SYSTEMS
zinc uptake by CMomydomonas variabllis grown in batch culture. Can. J. Fish.
Aquat. Set., 43, 687.
56. Schenck. R.C., Teuier, A., and Campbell. P. C. C. (1988). The elfect of pH on iron
and manganese uptake by a green alga, Umnot. Octanogr.. 33. 538.
57. Peterson. H. G.f Healey. F. P., and Wagemann. R. (1984). Metal toxidty to algae: a
highly pH dependent phenomenon. Cm. J. Fish. Aquat. Set.. 41. 974.
58. Xue, H.-B., Stumm. W., and Sigg. L. (1988). The binding of heavy metals to algal
surfaces Water Ret., 22. 917.
59. Andrew. R. W.. Biesinger, K. E.. and Class. C. B. (1977). liffects of inorganic
complcxing on the toxicity of copper to Daphnla magna, Water Re*., II, 309.
60. Magnuson. V. R.. Harris*. D. 1C, Sun. M. S., and Taylor. D. K. (1979).
Relationships of active metal-ligand species to aquatic toxicity. In Chemical
Modeling In Aqueous Systems— Speclatlon, Sorptlon. Solubility and Kinetics, cd.
Jenne. E. A., ACS Sympos. Ser. No. 93. American Chemical Society. Washington,
DC. 635.
61. Daly, H. R., Campbell, I. C.. and Hart, B. T. (1990). Copper toxicity to Paralya
atutralicnsls. L Influence of nitrilotriacetic add and glydne. Environ. Toxlcol.
Chem^ 9. 997.
62. Daly. H. R., Campbell. I. C, and Hart. B. T (1990). Copper toxicity to Paratya
auittrallfHsls. II. Influence of bicarbonate and ionic mrcnitih. Knvlrim. Taxied.
Chrm., V. 1007.
63. Waiwood, K. <»., and Hcamish, V. W. 1 1. (197K). lillecls of c«>p|*cr. pi I and hardness
on the critical swimming performance of rainbow trout (Salmo galrdntrl
Richardson). Water Res.. 12. 611.
64. Howirth. R. S.. and Sprague. J. B. (1978). Copper lethality to rainbow trout in
waters of various hardness and pH. Water Res.. 12, 455.
65. Calamari, D., Marchetti, R.. and Vailati. C. (1980). Influence of water hardness on
Cd toxicity to Salmo gairdnert Rich, Water Res., 14, 1421.
66. Bradley. R. W.. Sprague, J. B. (1985). The influence of pit. water hardness and
alkalinity on the acute lethality of zinc to rainbow trout (Salmo g alrdnerl), Can. J.
Fish. Aquat. Scl., 42. 731.
67. Cusimano, R. P., Brakke, D.. and Chapman, O. A. (1986). Effects of add water on
the toxidties of cadmium, copper, and zinc to trout. Can. J. Fish. Aquat. Scl., 43,
1497.
68. Part, P.. and Wikmark. G. (1984). The influence of some complexing agents (EDTA
and citrate) on the uptake of cadmium in perfused rainbow trout gills, Aquat.
Toxlcol.. 5, 277.
69. Part. P., Svanberg. O., and Kiessling, A. (1985). The availability of cadmium to
perfused rainbow trout gills in different water qualities. Water Res^ 19, 427.
70. Poldoski. J. E. (1979). Cadmium bio-accumulation assays— their relationship to
various ionic equilibria in Lake Superior water. Environ. Scl. Techno!., 13, 701.
7 1 . Block. M., and Wicklund Glynn. A. (1992). Influence of xanthates on the uptake of
'°Td by euriulan dace (Pht>\lnu* phnvlnus) and ralnhiiw trout (Onmrhynchiu
myklss), Environ. Toxlcol, Chem., H. 873.
72. Goitofrcy, J., Doff, K.. Jatim, S., and Tjalve, II. (I9KH). lUTect of potassium
ethylxanthate and sodium diethyldithiocarbamale on the accumulation and
disposition of nickel in the brown trout (Salmo tnitta), Phamacol* Toxleol,, 63, 45.
73. Borg, K. Gottofrey. J., and Tjalve. H. (198X). Effects of some dictating agents on
the uptake and distribution of MHg** in the brown trout (Sat/no tnitta). studies on
ethyl- and isopropylxanthate, dicthyl- and diisopropyldiUiiophosphate, dimethyl-
and diethytdithiocarbamate and pyridinethione, Arch. Toxlcol., 62, 387.
P. G. C CAMPBELL
101
74. Borg, H.. Karbson. A.-M.. and Lithner. G. (1976). Effects of the Flouting Agent
Isopropyl Xanthate on Lead Uptake in Trout (Salmo trutta)—Results of
Experimental Studies, SNV PM 754, National Swedish Environmental Protection
Agency. Solna (in Swedish; English abstract; dted in Reference 71).
75. Florence, T. M., Lumsden, B. G.. and Fardy, J. J. (1983). Evaluation of some
physico-chemical techniques for the determination of the fraction of dissolved
copper toxic to the marine diatom, Nitischia closterium. Anal, Chim. Ada, 151,281.
76. Florence. T. M.. and Stauber. J. L. (1986). Toxicity of copper complexes to the
marine diatom Nltuchla cloiterlum, Aaual. Toxlcol., 11. 26.
77. Ahsanullah, M., and Florence, T. M. (1984). Toxicity of copper to the marine
amphipod Allorchestes compressa in the presence of water- and lipid-soluble
ligands. Mar. Blot.. 84.41.
78. Wilkinson. K. J.. Campbell. P. G. C. and Couture. P. (1990). Effect of fluoride
complexation on aluminium toxicity towards juvenile Atlantic salmon (Salmo
solar). Can. J. Fish. Aauat. Set., 47. 1446. -
79. Sanjuan. B.. and Michard. G. (1987). Al hydroxide solubility in aqueous solutions
containing fluoride ions, Ceochlm. Cosmochim. Ada, 51.1823.
80. Plankey, B. J.. and Patterson. H. H. (1988). Effect of fulvic add on the kinetics of
aluminium fluoride complexation in acidic waters, Environ. Scl. Techno!., 22,1454.
HI. Murklund, II. Sjobcrg, S., and Ohman, L O. (19X6). liquilibrium and structural
•liMiick of silicon (IV) ami aluminium (III) in atiucoun solution. XIV. Speciution anil
equilibria in the aluminium(III)-lacl!c add-OH system, Acta Chem. Scand., A40,
367.
82. Wilkinson. K. J., Bertsch, P. M.. Jagoe, C. H.. and Campbell. P. G. C (1993).
Surface complexation of Al on isolated fish gill cells. Environ. Scl. Technot., 27.
1132.
83. Randall. D.. Lin. H.. and Wright, P. A. (1991). Gill water flow and the chemistry of
the boundary layer. Physht Zool., 64.26.
84. Playle. R. C. and Wood. C. M. (1989). Water chemistry changes in the gill micro-
environment of rainbow trout experimental observations and theory. /. Comp.
Physlol., 159.1.
85. Wilkinson, K. J.. and Campbell. P. G. C. (1993). Aluminium bioconccntration by
juvenile Atlantic salmon (Salmo solar) in addic media. Environ. Toxlcol. Chem.. 12.
2083.
86. Borgmann, U.. and Ralph. K. M. (1983). Complexation and toxicity of copper and
the free metal bioassay technique. Water Res.. 17. 1697.
87. Guy, R. D., and Kean. A. R. (1980). Algae as a chemical speciation monitor. I. A
comparison of algal growth and computer calculated speciation. Water Res., 14.
891.
88. Harrison, S. J., Lepp, N. W., and Phipps, D. A. (1984). Uptake of copper by excised
roots. IV. Copper uptake from complexed sources, 7.. PJtanzenphyslol. Bd., 113,
445.
89. Stecmann Nielsen, K., Kamp-Nklsen, L., and Wium-Anderscn, S. (1969). The effect
of deleterious concentrations of copper on the photosynthesis of Chlortlla
pyrenoldosa. Physlol. Plant., 22,112).
90. George, S. G.. and Coombs, T. L. (1977). Effects of chdating agents on the uptake
and accumulation of cadmium by Mytllus edulls. Mar. /Mo/., 39, 261.
91. Laube. V. M.. McKenzie. C N.. and Kushner. D. J. (1980). Strategies of response
to copper, cadmium and lead by a blue-green and a green alga. Can. J. Mlcrobiol.
26,1300.
-------
102 METAL SPECIATION AND BIOAVAILABIUTY IN AQUATIC SYSTEMS
92. Block, M, and Put. P. (1986). Increased availability of cadmium to perfused
rainbow trout (Satnv galraneH, Rich.) gfflj in the presence of the complexiog
agents dietbyldithkicarbamate, ethyl xanthate and isopropyl xanthate. Aquat.
93. Giesy, J. P., Newell. A, and Leversee. O. J. Q9S3). Copper speciation in soft. acid.
hunuc waters: effects on copper bioaocumulation by and toxicity to Simoeephalta
ttrrulatus (DaphnidaeX ScL Tot. Environ^ 28. 23.
94. Daly. H. R, Jones. M . J, Hart. B. T., and Campbell. I. C (1990). Copper toxicity
to Faratya tnatraSeult. Ilr. Influence of dissolved organic matter. Emlron.
Toxicol. Chem.,9. 1013.
95. Meador, 1. P. (1991). The interaction of pH. dissolved organic carbon, and total
copper in the determination of ionic copper and toxicity. Aquat. ToxlcoL, 19. 13.
96. Sunda. W. G., and Lewis, J. A. M. (1978). Effect of complexation by natural
organic ligands on the toxicity of copper to a unicellular alga. Monochrytis hitHert,
LbnnoL Oceanogr., 23. 870.
97. Borgmann. U., and Caarlton. C C. (1984). Copper complexation and toxicity to
Daphnta in natural waters. /. Great Lakes Rei^ 10, 393.
98. Laegrcid, Mn Abtad. f ., Klavcness. Dn and Sdp, H. M. (1983). Seasonal variation
ofcadmium toxicity toward the alga Selenasinun eaprlcornutum PriaU in two lakes
with different humus content. Environ. ScL Techno!^ 17. 3S7.
99. Giesy. J. P.. Leversee. G. J, and WOUams. D. R. (1977). Effects of naturally
occurring aquatic organic fractions on cadmium toxicity to Simotephahu
temilatus (Daphnidae) and Cambusia qffinls (PoecOiidae), Water Ret^ 11. 1013.
100. Winner. R. W. (I9M). The toxidty and bioaccumulation of cadmium and copper
as affected by humk acid, Aquat. Toxicot., 5, 267.
101. Winner. R. W. (1986). Interactive effects of water hardness and hunuc acid on the
chronic toxicity ofcadmium to Daphnia putts, Aquat. Toxicol., 8, 281.
102. Winner. R. W, and Gauss. J. D. (1986). Relationship between chronic toxicity and
bioaccumulation of copper, cadmium and zinc as affected by water hardness and
hunuc acid. Aquat. Toxicol., 8. 149.
103. Buffle, J., Altmann. R. S, FileDa, M, and Tcssier. A. <1990). Complexation .by
natural heterogeneous compounds: site occupation distribution functions, a
normalized description of metal complexation, Geochlm. Cosmochlm. Ada, 54.
1S3S.
104. Myers. V. B, Iverson. R. L, and Harriss. R. C (1975). The effect of salinity and
dissolved organic matter on surface charge characteristic of some euryhaline
phytoptankton. J. Exp. Mar. Blol. Eeot^ 17. 59.
105. Prakash. A., and Rashid, M. A. (1968). Influence of humic substances on the
growth of marine phytoplaokton. DinoflageDates, Umnot. Oceanogr., 13. 598.
106. Hargeby. A., and Petersen, R. C. (1988). Effects of low pH and humus on the
survivorship, growth and feeding ofCammana pulex (L.) (Amphipoda), Frtsh*at.
Rei.. 19, 235.
107. Phinney. J. T. and Bndand. K. W. (1994). Uptake of Upophilic organic Cu. Cd and
Pb complexes in the coastal diatom Thaiaulotira wtlufiogU. Environ. Scl. Technol.,
28. 1781.
-------
Pergamon
Environmental Toxicology and Chemistry, Vol. 14, No. 11, pp 1847-18S8, 1995
Copyright © 1995 SETAC
Printed in the USA
0730-7268/95 $9.50 + .00
0730-7268(95)00146-8
MODELING SILVER BINDING TO GILLS OF RAINBOW TROUT
(ONCORHYNCHUS MYKISS)
NANCY JANES and RICHARD C. PLAYLE*
Department of Biology, Wilfrid Laurier University, Waterloo, Ontario, N2L 3C5 Canada
(Received 9 February 1995; Accepted 27 April 1995)
Abstract-Rainbow trout (Oncorhynchus my kiss, 1-3 g) were exposed to -0.1 jtM silver (Ag) (-11 /tg-L~' Ag) for 2 to 3 h
in synthetic soft water (Ca, Na -300 pM, pH 6.5-7.5) to which was added Ca. Na, H*. dissolved organic carbon (DOC). Cl.
or thiosulfate (SjOj). Gills were extracted and gill Ag concentrations were measured using graphite-furnace atomic absorption
spectrophotometry. The concentrations of cations (Ca, Na, H+) and complexing agents (DOC, Cl, SjOj) needed to keep Ag
off the gills were used to calculate conditional equilibrium binding constants (K) at the gills. Log K for Ag-gill binding was
10.0. with approximately 1.3 nmol Ag binding sites per fish. All experimentally determined log/rvalues were entered into an
aquatic chemistry equilibrium model, MINEQL+, to predict Ag binding at trout gills. For a series of natural waters, model-
predicted gfll Ag concentrations correlated well with observed gill Ag concentrations, with one exception, very hard city of Waterloo
tapwater. This exception may indicate a kinetic constraint on the thermodynamic basis of the model.
Keywords-Fish Gills Silver Binding Model
INTRODUCTION
Silver (Ag) is one of the most toxic metals to fish, with
96-h LCSOs of free Ag between about 0.05 and 0.30 /iM (5.3-
32 pg'L'1 [1-4]). There is a clear influence of water hard-
ness and Ag speciation on Ag toxicity, with less toxicity of
Ag in harder water (e.g., high Ca concentrations [1]) and
when Ag is bound by ligands such as Cl or thiosulfate [3].
Thiosulfate is a fixer used in developing photographic film,
and silver thiosulfate is the form of Ag that has the most po-
tential to enter the environment in photoprocessing effluents.
Although the expense of Ag ensures that efforts are made to
retain it during industrial processes, it has been estimated that
about 25% of the total annual consumption of Ag has the
possibility of entering the environment during product dis-
posal, and that 10% of this waste enters water as unrecycled
photographic Ag [5]. Although Ag is not often found in high
concentrations in most natural water sites, heavily contam-
inated sites may have surface-water concentrations as high
as about 0.35 pM [5].
Waterborne Ag will interact first with the gills of a fish.
Fish gills constitute about half a fish's total surface area and
are a sensitive and critical region of contact with water [6].
Fish gills function as the main location for gas transfer, ni-
trogenous waste excretion, acid-base regulation, and ion up-
take to counter passive diffusive losses of ions [?]• The gill
epithelium contains negatively charged binding sites, which
are due to phosphate, carboxyl, amino, and sulfate groups,
among others [8]. Metals and cations are attracted to these
negative binding sites on the external surface of the gill epi-
thelium, and competition for gift binding sites is likely to oc-
cur between Ag and other cations such as Ca. Complexation
•To whom correspondence may be addressed.
by ligands such as thiosulfate, dissolved organic carbon, and
Cl will decrease free Ag available to interact at the binding
sites. Thus, the influence of water hardness and speciation
on Ag toxicity to fish, and metal toxicity in general, can be
explained through competition and complexation [9,10],
The primary objectives of our study were to determine the
binding ability of Ag to sites on fish gills, to determine in-
teractions of other cations at the Ag binding sites, to de-
termine interactions of complexing agents with Ag, and to
model the system for predictive purposes. The method we
used consisted of estimating conditional equilibrium bind-
ing constants (K) from results obtained using different li-
gands and cations, and entering these binding constants into
an aquatic chemical equilibrium model, MINEQL+ [11].
Our method is similar to ligand-exchange methods used to
establish conditional stability constants of metal complexes
of organic ligands in seawater [12,13] and has been used be-
fore to model Cu and Cd interactions at fish gills [14,15].
MATERIALS AND METHODS
Experimental
Small rainbow trout (Oncorhynchus mykiss, 1 -3 g) were
purchased from Rainbow Springs Hatchery, Thamesford,
Ontario. Fish were acclimated to synthetic soft water for at
least 1 week before each experiment. Soft water was pro-
duced by reverse osmosis (Culligan Series E reverse-osmosis
system) and had a composition of 20 to 320 /iM Ca, 150 to
400 jiM Na, 100 to 290 pM Cl, and pH 6.5 to 7.5, depend-
ing on the efficiency of the reverse-osmosis system. Fish were-
fed Martin's starter food during acclimation but not during
the experiments.
For an experiment, six fish were randomly assigned to
each aerated polyethylene container containing 1 L of soft
water, modified by the addition of Ca, Na, H+, dissolved
1847
-------
1848
N. JANES AND R.C. PLAYLE
organic carbon (DOC), Cl, thiosulfate, Ag, or Cd (at 12 to
17°C, see below). Water samples were taken for later analy-
sis, and water pH was measured and adjusted to between pH
6.5 and 7.5 with dilute KOH or H2SO4 if necessary. Fish
were exposed to the water for 2 h, after which the fish were
randomly sampled, stunned with a blow to the head, and the
gills from both sides extracted using stainless steel forceps.
Extracted gills were rinsed in 100 ml of reverse-osmosis wa-
ter for 10 s, shaken to remove excess water, then placed in
1.5-ml polypropylene microcentrifuge tubes. In a normal ex-
periment, sampling time meant that fish were exposed to their
respective solutions for between 2 and 3 h.
Wet weight of the extracted gills was measured with a Met-
tler AE163 balance after their transfer to new microcentri-
fuge tubes. Following gill weighing, 1 H HNO} (Ultrex II
Ultapure Baker analyzed reagent) was added to each micro-
centrifuge tube at five times the gffl wet weight. Gills plus acid
were heated for 3 h at 80°C. After heating, the digested gills
were agitated, allowed to settle, and 100 pi of supernatant
was added to 900 pi of Mill! Q deionized water (Millipore)
in new microcentrifuge tubes, and left for later analysis.
Analysis of gill and water Ag concentrations was done
using a graphite-furnace atomic absorption spectrophotom-
eter (AAS; Varian AA-1275 with GTA-95 atomizer). Ten-
microliter samples were injected, N2 gas was used, and
operating conditions were 5 s at -75 °C, a short (30 s) dry-
ing time at -90°C, 12 s at - 120°C, and 4 s at 2,000°C, dur-
ing which Ag concentrations were read. Water and gill Cd
concentrations were measured using conditions documented
by Varian, but also with a 30-s drying time at -90°C. Gill
and water Ag and Cd samples were run against standards pre-
pared from Fisher-certified Ag and Cd reference solutions.
Water samples collected at the beginning and end of the
experiments were analyzed for Na and Ca using a Perkin-
Elmer model 3100 AAS; LaCl3 was added to the Ca samples
to reduce Na interference. Water Cl was measured using the
mercuric thiocyanate colorimetric method (Sigma reagents).
Dissolved organic carbon was measured using a Beckman
915B total carbon analyzer. Water pH was measured with a
PHM82 Radiometer pH meter and GK2401C combination
electrode.
Chemicals used were usually made as 1 mM stock solu-
tions. Silver was added as AgNO3 (AnalaR, BDH, Poole,
England), thiosulfate as NaAOj-SHjO (AnalaR, BDH), Cl
as KC1, Na as NaOH, H* as H2SO4, and Ca as Ca(OH)j.
The DOC was isolated from the south side of Luther Marsh
(43°54'N, 80°24'W, near Grand Valley, Ontario) by tangen-
tial flow ultrafiltration (Millipore, Bedford, MA), and was
supplied to us by Kent Burnison, National Water Research
Institute, Environment Canada, Burlington, Ontario.
A longer-term thiosulfate and Ag experiment was run, in
which 16 to 18 fish per 16-L bucket were held for a week
in 13°C aerated soft water. Fish were sampled periodically
for gill Ag, and water was sampled for water ions and Ag
concentrations. Half the water volume was replaced daily,
and each treatment was run in duplicate.
To test the Ag-gill model, which was based on the syn-
thetic softwater experiments, rainbow trout were exposed
2 to 3 h in Ag-supplemented water collected from various
locations in southern Ontario. Fresh water was collected in
May 1994 from four locations: Lake Ontario at the south pier
of the entrance to Hamilton Harbour, Hamilton (43°18'N,
79047' w), "Dundas pond" on the east side of the main access
road to Dundas Valley Conservation Area off Highway 99
(43°14'N, 80°00'W), Gait (Mill) Creek above the dams
at Shades Mills Conservation Area, Cambridge (43°23'N,
80°15'W), and Grand River at old highway 8, Kitchener
(43°25'N, 80°24'W). We also used water from Salmon and
Jack lakes (44°43'N, 78°03'W, north of Peterborough), which
had been stored since 1991 [15]. Finally, we used Waterloo
city tapwater entering our laboratory, passed through acti-
vated charcoal to remove chlorine.
In an effort to differentiate colloidal Ca from Ca in solu-
tion, we filtered three natural waters through 0.45- or 0.2-/tm
cellulose nitrate filters (Sartorius). Fifteen milliliters of MilliQ
water were passed through each filter under vacuum as a
rinse, then 15 ml of sample, then another 15 ml of sample,
which was collected. A Millipore graduated funnel assembly
was used; all water was at 16°C.
Statistical analysis of the gill Ag data consisted of un-
paired t tests or one-way analysis of variance (ANOVA) fol-
lowed by the Student-Newman-Keuls method of pairwise
multiple comparisons, as appropriate [16]. Significant dif-
ferences are reported at the p < 0.05, p < 0.01, and p <
0.001 levels (*, **, ***, respectively). Error bars in the fig-
ures represent the 95% confidence interval (C.I.) about the
mean gill Ag concentration, usually for six fish.
Conditional gill stability constants
and computer modeling
The main objectives of this study were to determine the
bidding ability of Ag to the gills of rainbow trout, to deter-
mine the interactions of other cations at Ag binding sites and
the complexation of Ag by ligands hi water, and to model
these interactions in a computer program. Powerful aquatic
chemistry programs exist that use thermodynamic stability
constants; we chose MINEQL+ (version 2.0), a chemical
equilibrium program for personal computers, mainly be-
cause of its ease of use [11]. Some equilibrium constants from
MINEQL+ are given in the Appendix. Results from the ex-
posures of fish to Ag were used to calculate stability constants
for insertion into the MINEQL* program.
To calculate conditional gin stability constants we consid-
ered two scenarios. The first scenario is the presence of a sol-
ute such as Na at a concentration that does not keep Ag off
the gills (no inhibition of Ag binding; Eqn. 1).
[Na] •
< [Ag] •
(1)
where [Na] = sodium concentration in the water, K = sta-
bility constant, Na-gillAg = sodium binding at silver bind-
ing sites on the gills, [Ag] = concentration of free silver in
solution (as calculated by MINEQL+), and Ag-gillAg = sil-
ver binding at silver binding sites on the gills.
-------
Modeling silver binding to gills of rainbow trout
1849
The second scenario is a solute such as Na at a concen-
tration that keeps all Ag off the gills (complete inhibition of
Ag binding; Eqn. 2).
(2)
Details of the derivation of these two equations can be found
in Playle et al. [IS]; a similar derivation can be found in
Midorikawa and Tanoue [12]. For Equations 1 and 2,
MINEQL* was used to calculate free Ag from water chem-
istry and the total concentration of Ag. This calculation takes
into account side reactions involving Ag (e.g. [13]).
Once conditional stability constants for the gills were es-
timated (see Results), they were entered into the MINEQL+
program. For each simulation the concentrations of Ca, Na,
Cl, DOC, Ag, and NO3 (=[Ag]> in the test water were en-
tered into the program. Water pH was fixed to the measured
value (i.e., pH changes at the gill surface were ignored; see
Playle et al. [IS]). This approach simplifies the model because
buffering capacity of the water does not need to be deter-
mined, and is appropriate because sensitivity analysis using
MINEQL+ indicated that Ag speciation is very insensitive to
changes in water pH. Measured water temperature was also
entered into the model, and the system was considered open
to the atmosphere. The concentration of HCO J in the water
was assumed to be the same as that of Ca; this simplification
was tested using the more accurate Henderson-Hasselbalch
equation. In any case, sensitivity testing of the model indi-
cated that Ag binding to gills is unaffected by HCOf.
RESULTS
Initial experiments in synthetic soft water showed that Ag
accumulation on gills of rainbow trout was significant at ex-
0.00 0.04 0.05 0.22 0.39 0.60
silver concentration (//M)
Fig. 1. Silver accumulation on gills of 1- to 3-g rainbow trout ex-
posed to 0.04 to 0.50 /iM Ag for 2 to 3 h in synthetic soft water. The
clear bar represents no Ag added (control fish); striped bars repre-
sent Ag added. Asterisks indicate significant differences from con-
trol mean (*p < 0.05, **p < 0.01, ***p < 0.001). Error bars represent
95% C.I. about the means, with six fish per bar.
posures of 0.05 /iM and greater (Fig. 1; 2- to 3-h exposures,
Ca = 27 (Ml, Na = 152 /
-------
1850
N. JANES AND R.C. PLAVLE
14
•o 12
<
o
c
I
"o>
50
100
150
time (h)
Fig. 3. Silver aocumulaion on gills of 1- to 3-g rainbow trout exposed to Ag for 6 d. Fish exposed to 0.06 pM Ag in the absence of thiosulfate
(•) showed significantly elevated gill Ag concentrations (asterisks; t test) compared to fish exposed to 0.07 pM Ag plus 2.S pM thiosulfate
(V). The cross at 147 h indicates significant gill Ag accumulation (p < 0.05,1 test) in the Ag-plus-thiosulfate treatment compared to control fish.
equilibrium program [11]. This program uses equilibrium
constants; some of the relevant constants from the program
are given in log K format in the Appendix. These values are
nearly identical to values used in other chemical programs
such as MINTEQA2/PRODEFA2 [17], and they are close
to the equilibrium constants given in Morel and Hering [10].
The silver thiosulfate complex (AgSjOJ) has a log A"
value of 8.80 (see the Appendix). It took about 29 tunes more
thiosulfate than Ag to keep Ag off trout gills (Fig. 2), so the
log Ag-gill conditional equilibrium constant must be >8.80.
The range of the Ag-gill conditional equilibrium constant can
be estimated using Equations 1 and 2, substituting J^Oj for
Na and AAC-SJO, for *N«-«niA«- Using Equation 1 (no protec-
tive effect of 1 pM thiosulfate), log A^.^niAg must be >10.0
(SjOj = 1 iM, logA'Ag.SjO, = 8.8, free Ag = 0.06
Using Equation 2 (complete protective effect),
must be < 10.3 (S2O3 = 2 jtM, log AXg^o, = 8.8, free Ag =
0.06 itM). Silver binding sites at the gill were inserted into
the MINEQL+ program as "gUlAg" (null component), then
defined as a complex with 1:1 Ag:gfllAg binding. From our
thiosulfate data (Fig. 2), we estimated 0.61 nmol binding sites
per fish (6.1 nmol Ag/g wet tissue for 0.07 pM Ag exposure,
no thiosulfate; average gill basket of our fish -0.1 g) and
entered 0.61E—9 for the molar concentration of the gillAg
component.
Log AA^JIIA, values of between 10.0 and 10.3 were tried,
and the MINEQL+ program output values were compared
to the thiosulfate results of Figure 2. Best fit to the experi-
mental data occurred with log AA,.,^ = 10.0 and 0.61 nmol
binding sites per fish. Predicted values for these conditions,
using water chemistry available from the experiments (Ca =
18 itM, Na = 174 /iM, Cl = 100 /*M, pH 6.5,17°Q, are given
in Figure 4. The addition of thiosulfate to the system kept
Ag off the gills only at relatively high thiosulfate:silver mo-
lar ratios. The correlation coefficient between observed and
MINEQL+-predicted gill Ag (using log AXg^iAg = 10.0) was
r = 0.924 (p < 0.01).
In contrast, free Ag* concentrations in the water, as cal-
culated by MINEQL+, decrease rapidly as the concentration
of thiosulfate is increased. If gill Ag concentrations were di-
rectly related to free Ag+ then observed gill Ag and gill Ag
predicted from free Ag* concentrations would track one an-
other, but the fit is poor (Fig. 4) and the correlation coeffi-
cient is not significant (r=0.525, p > 0.05). In other words,
although free Ag"1" is reduced by the addition of thiosulfate
to the water OogAAt.SzOj = 8.8; see Appendix), trout gills
successfully outcompete thiosulfate for Ag (log AAC-SWAS =
10.0), so gill Ag concentrations do not directly follow free
Ag* concentrations.
Complexation ofAg and competition
for Ag binding sites
Complexation of Ag by Cl and dissolved organic carbon
(DOQ and competition by Na, Ca, and hydrogen ions (H*)
for Ag binding sites on the gills were investigated to better
define and model Ag binding at fish gills. The addition of
11.3 mM of Cl kept Ag off the gills during a 0.11/tM Ag ex-
posure for 2 to 3 h in synthetic soft water, whereas 1.5 mM
of Cl did not (Fig. 5; Ca = 155 pM, Na = 396 ^M, 12°C,
pH 7.5). Chloride forms a number of weak complexes with
Ag, depending on the ratio of Cl:Ag (see Appendix). Using
Equation 1, log AA^O was estimated at <5.7 Gog AX,.,^ =
10.0, free Ag = 0.07 pM, Cl = 1.5 mM). Using Equation 2,
1°£ •KA(-CI was estimated at >4.8 Gog ^Ag-guiAt = 10.0, free
Ag = 0.07 i/M, Cl = 11.3 mM). Predicted gill Ag concentra-
tions given in Figure 5 will be dealt with in the next section.
There was a general decrease in gill Ag as DOC concen-
tration increased, and DOC concentrations of 24.2 mg C/L
kept Ag off the gills of trout (Fig. 6; 2- to 3-h exposures,
-------
Modeling silver binding to gills of rainbow trout
1851
10
I 6
0.0
0.5 1.0
thiosutfate concentration
1.5
2.0
Fig. 4. Observed and MINEQL "^-predicted gill Ag concentrations for the thiosulfate data of Figure 2. Exposure was 0.07 pM Ag and 0 to
2 pM thiosulfate. in synthetic soft water, for 2 to 3 h. Error bars about the observed data (•) are the 95% C.I. Two predictions of gill Ag
concentrations were made using MINEQL"1". The first prediction used log ft/mm*. = 10.0 and 0.61 nmol binding sites per fish (O); for clar-
ity, these symbols are shifted to the right in the figure. The second prediction of gill Ag (T) was based directly on free Ag* in the water, as
calculated using MINEQL*. Gill Ag predicted using log A'A.^UA, = 10.0 was a much better fit to the observed data. The open circle in the
bottom left corner represents background gill Ag. See text for more details.
Ag = 0.17 /*M, Ca = 315 pM, Na = 386 pM, Cl = 265 pM,
14°C, pH 6.8). Using Equation 1 and an estimated value of
0.05 fiM Ag binding sites per mg C DOC (from [15]), it was
estimated that logAXi.rxx: < 9.2 (log A^^IIA. = 10.0, free
Ag=0.12^M. DOC =14.6 mgC/L=0.73 AtM). Using Equa-
tion 2, it was estimated that log AAC.DOC > 9-° Gog ^A«-fniA»=
10.0, free Ag = 0.12 pM, DOC = 24.2 mg C/L = 1.21 pM).
Predicted gill Ag values will be discussed in the next section.
As an illustration of competition for Ag binding sites at the
gills, 16 mM Na kept Ag off trout gills, whereas 1.6 mM Na
did not (Fig. 7). Exposures were to 0.11 pM Ag in synthetic
soft water, for 2 to 3 h (Ca = 155 fM, Cl = 265 jtM. 12°C,
pH 7.5). From Equation 1 it was estimated that log AN.^UA,
was <5.7 (log A'Aj.jniA, = 10-°« free A* = OM MM« Na =
1.6 mM). From Equation 2, log AN*,^ > 4.7 flog AX,^DIA» =
10.0, free Ag = 0.08 /iM, Na = 16.0 mM). Predicted gill Ag
will be dealt with in the next section.
Competition for Ag binding sites by H* was investigated
0.3 0.4 1.5
chloride concentration (mM)
11.3
Fig. 5. Observed and predicted gill Ag concentrations in the pres-
ence of 0.3 to 11.3 mM Cl. The dear bar represents gill Ag concen-
trations from six control fish; striped bars represent fish gills from
0.11 j*M Ag exposures (six fish per bar). Error bars are 95% C.I.
*•* = significantly greater than control concentrations (p < 0.001).
Predicted gill Ag concentrations (hatched bars) were calculated us-
ing the MINEQL+ program, with the insertion of the final param-
eters given in Table 1 (see text for details).
3.0 5.0 7.0 10.9 14.6
DOC concentration (mg C/L)
24.2
Fig. 6. Observed and predicted gill Ag concentrations in the pres-
ence of 3.0 to 24.2 mg C/L DOC. Striped bars represent gill Ag con-
centrations from fish exposed to 0.17 pM Ag, six fish per bar. *, **,
*** = significantly greater than control concentrations (*p < 0.05,
**/> < 0.01, ••*/> < 0.001, respectively). See caption of Figure 5 for
more details.
-------
1852
N. JANES AND R.C. PLAYLE
Table 1. Input data used in the Ag-gill model
Complex
Initial log K
Final log A"
0.4 0.4 1.6 16.0
sodium concentration (mM)
Fig. 7. Observed and predicted gill Ag concentrations in the pres-
ence of 0.4 to 16.0 mM Na. Striped bars represent gill Ag concen-
trations from fish exposed to 0.11 pM Ag, six fish per bar. See
caption of Figure 5 for more details.
using a Ag concentration of 0.06 pM Ag at pH 6.8,5.5, and
4.5 (2- to 3-h exposure, Ca = 260 pM, Na = 388 fiM, Cl =
250 pM, 12°Q. There was no inhibition of Ag binding to the
gills (all 6.0 to 6.5 nmol Ag/g wet tissue) at even the lowest
pH used. From Equation 1, the log H"*-gillAg conditional
equilibrium constant was estimated at <7.1 (log A'Ac.0UA( =
10.0, free Ag = 0.04 pM, pH 4.5 = 32 pM H+).
Concentrations of Ca up to 10.6 mM did not reduce Ag
binding to trout gills (5 to 7 nmol Ag/g wet tissue, 2- to 3-h
exposure to 0.05 ^M Ag, Na = 399 pM, Cl = 265 jtM, 12°C,
pH 7.1). From these results and Equation 1, logA'c>.gQ1A(
was estimated to be <4.5 (log AA^UA, = 10.0, free Ag =
0.03 /iM. Ca = 10.6 mM).
Final modeling of Ag-gill interactions
From the thiosulfate data (Fig. 2), the Cl and DOC com-
plexation data (Figs. 5 and 6), and the Na (Fig. 7), H+, and
Ca competition data, we estimated conditional equilibrium
constants, the number of Ag-gill binding sites, and the num-
ber of Ag binding sites on DOC (summarized in Table 1).
Using the constraints of these initial estimates, we introduced
the log AT values one by one into the MINEQL+ program
and best-fit predicted gill Ag concentrations to the observed
values by trial and error.
The first fitting exercise was already described for thio-
sulfate and is illustrated in Figure 4, and log A^^QIA, was
optimized to 10.0 using 0.61 nmol Ag binding sites per fish.
The DOC data were best predicted (Fig. 6) using 35 nmol
binding sites per mg C/L, log AAc.DOC of 9.0, and setting
1.0 nmol binding sites Ag per fish. This higher value for bind-
ing sites was used to scale the predicted values to the observed
values in Figure 6, because the introduction of DOC (back-
ground = 2.4 mg C/L) into the model reduced the amount
of Ag available to bind with the gillAg sites: The correlation
Ag-gillAg
Na-gillAg
H-gillAg
Ca-gillAg
Ag-DOC
H-DOC
10.0-10.3
4.7-5.7
<7.1
<4.5
9.0-9.2
4.0
10.0
4.7
5.9
3.3
9.0
4.0
Binding sites
gillAg: 0.6 to 1.9 nmol binding sites per fish (varied for scaling)
DOC: 35 nmol binding sites per mg C/L (initial = 50)
Initial conditional equilibrium constants (log K) were calculated from
the experimental data using Equations 1 and 2 (see text). Final log K
values were chosen by best-fitting the initial log K values to the ex-
perimental data.
of this fit was 0.946 (p < 0.01). Varying the number of gill
binding sites used in the model by a factor of two or three
has little effect on the correlation between observed and pre-
dicted values (varying binding constants has a much greater
effect, because they are log values), and this was done to fit
predicted values to the vertical scale of the figures. We had
no data for H+ interactions with DOC, and we used the
value previously used in Playle'et al. [IS] of log A"H-DOC =
4.0 (Table 1).
The best fit to the Na data (Fig. 7) was log KNst.liUAe =
4.7, with the number of Ag-gill binding sites 1.2 nmol per fish
(r = 0.973, p < 0.05). Competition by H"*" for Ag binding
sites was more difficult to fit, because we did not have a com-
petitive effect at pH 4.5, and to take an experiment with fish
'much below pH 4.5 is unrealistic. We assumed a full com-
petitive effect at pH 3.5, so the final log A" value was set to
5.9, with 1.1 nmol Ag-gillAg binding sites per fish. In a sim-
ilar manner, full protective effect of Ca was assumed to occur
at 150 mM Ca (with 1.4 nmol Ag binding sites per fish), and
final logATanuiA, was estimated to be 3.3 (Table 1). We
could not physically keep Ca in solution in our system at con-
centrations >11 mM.
The Cl fitting (Fig. 5) gave an indication of the power of
our modeling approach. Unlike Ag-DOC interactions, for
which there are no equilibrium constants in MINEQL*, and
the Na, H+, and Ca competition at Ag binding sites, for
which equilibrium constants were calculated for the first time
by us, the MINEQL"1" program already contained Ag-Cl
equilibrium constants. Thus, no fitting was required, aside
from setting log AA,.,^ = 10.0 and setting the number of
gill binding sites to 1.25 nmol/fish. In the previous section,
it was estimated from our gill Ag data that log K/^^ was
>4.8 and <5.7. Running the MINEQL"1" simulation indi-
cated that with 1.5 mM Cl and 0.11 pM Ag concentra-
tions only AgCl was formed (log K=3.3; see Appendix), and
Ag was still bound to the gills (Fig. 5). At 11.3 mM Cl and
0.11 /*M Ag, MINEQL+ calculations indicated that AgClJ
was formed (log A~Af-a2 = 5-3)> and Ag was kept off the ^
(Fig. 5). Log 5.3 is approximately midway between log 4.8
and log 5.7, estimated from the gill Ag data of Figure 5
-------
Modeling silver binding to gills of rainbow trout
1853
4.0
Ag—9.0—DOC'
water
Ag—8.8—S2Oa
Ag—3.3-CI
Ag-5.3-CU
Ca-3.2—CO, H*
Na
Fig. 8. Illustration of the interactions of the most important param-
eters in the Ag-gill model. Complexation of Ag in the water may
reduce Ag binding to trout gills, depending on the relative concen-
trations of the ligands and the relative magnitude of their conditional
equilibrium constants (log K values from Table 1 are indicated in
the figure). Competition with Ag at a single Ag-gill binding site
(asterisk) is also dependent on relative binding strengths and cation
concentrations.
= log 5.5). The Cl data were a relatively independent
check of the fitting process, and the good fit between pre-
dicted and observed gill Ag (r - 0.956, p < 0.05) nicely il-
lustrates the power of the MINEQL+ aquatic chemistry
program.
A schematic illustration of Ag binding to the gill, cation
interference with that binding, and Ag and H+ binding to
DOC is given in Figure 8. The log/rvalues for these reac-
tions are from Table 1; some of the more important equilib-
rium constants from MINEQL* (Appendix) are also given
in the water portion of the diagram. In essence, complexation
of Ag (e.g., by DOC) can occur in water, rendering Ag un-
available to bind at the gills, and cation competition with Ag
(e.g., by Ca) occurs at the binding sites on the gills themselves.
Once Ag is on the gills, it can exert a toxic effect at the
gill itself or after active uptake or passive diffusion into
the gills, then into the blood of the fish. An indication of
the toxic effect of Ag is ion loss from rainbow trout into the
surrounding water (e.g., [7]). In two of our experiments
(Figs. 1 and 2) the concentration of Na in the synthetic soft
water was low enough (152 and 174 /tM, respectively) that
we were able to measure Na loss from the fish to the water
over 2 h (six 1- to 3-g trout per liter of water). There is an
Na diffusion gradient from inside the fish into the water
(plasma Na — 150 mM [7]). Sodium loss (net flux per unit
weight = Jnci [7]) plotted against gill Ag yielded a correlation
2.5
f 2-
1.5
1-0
b
«, 0.5
0.0
5 10
gill Ag (nmol Ag/g wet tissue)
15
Fig. 9. Sodium losses from fish to the test water, on a pmol/g/h ba-
sis, plotted against mean gfll Ag concentrations. Each point was de-
termined from six, 1- to 3-g rainbow trout held in 1 L of test water
for 2 h. The regression line (r = 0.778, n = l2,p< 0.01) does not
include the outlying point indicated by the open circle. See text for
more details.
coefficient of r = 0.560 (p < 0.05) if all 13 data points are
considered. Omitting one outlier (presumably due to sample
contamination) yielded a more significant correlation of r =
0.778 (p < 0.01; Fig. 9). These results give some indication
that the acute effect of Ag at trout gills is an ionoregulatory
disturbance (see Discussion).
Model testing
Once the model was constructed by inserting the final val-
ues given in Table 1 into the MINEQL+ program, it was
necessary to test how well the model predicted Ag accumu-
lation on fish gills. We ran a thiosulfate experiment in Lake
Ontario water to see how much thiosulfate was needed to
protect against Ag accumulation in natural, moderately hard
water, to compare with published thiosulfate protection
against Ag toxicity [3]. In addition, we ran Ag experiments
in a variety of waters, to test the predictive capabilities of the
model. All predictions of gill Ag concentrations were made
using water chemistry given in Table 2, before viewing the
gill Ag results from die experiments.
In water from Lake Ontario; only 0.5 j*M of thiosulfate
was needed to keep 0.09 pM Ag off trout gills (Fig. 10), and
predicted gill Ag matched observed values well (r = 0.954,
p < 0.001). For scaling predicted to observed results, the
number of gill binding sites was set to 1.9 nmol per fish.
Water chemistry for the Lake Ontario water used in the
model prediction is given in Table 2.
Observed and predicted gill Ag concentrations of fish held
2 to 3 h in Ag-supplemented natural waters also correlated
well, with one notable exception. The exception was city of
Waterloo tapwater, hi which high gill Ag concentrations were
-------
1854
N. JANES AND R.C. PLAYLE
Table 2. Input exposure water chemistry data used in model testing
Measured concentrations
Ag
Water
Reverse osmosis
Gait Creek
Salmon Lake
Grand River
Jack Lake
Dundas Pond
Lake Ontario
Waterloo tapwater
PH
7.9
8.2
7.5
8.2
7.4
8.1
7.8
8.3
• inji.
(mg C/L)
1.7
7.9
4.4
8.5
8.3
4.7
4.8
3.3
Ca
140
1,882
822
2,094
506
2,029
1,058
2,860
Na
1,090
505
46
1,045
78
489
545
980
a
784
664
32
1,146
32
505
602
1,249
Background
0.01
0.01
0.01
0.01
0.00
0.01
0.00
0.02
Ag added
0.06
0.10
0.09
0.10
0.08
0.09
0.12
0.06
All values entered into the model are the mean of two or three measurements, except for background Ag where
n— I. Exposure temperature was 15°C.
seen, whereas the model predicted nearly background levels
(Fig. 11). The model predicted low gill Ag because of the high
Ca, Na, and Cl concentrations hi the water, plus the mod-
erate DOC concentration (Table 2). The relationship between
observed and predicted gill Ag is dearly seen hi Figure 12,
where the straight line through observed and predicted gill
Ag, excluding Waterloo tapwater, has a significant correla-
tion coefficient (r = 0.8%, p < 0.01). The correlation includ-
ing Waterloo tapwater is not significant (r=0.543, p > 0.05).
Likely reasons for the poor prediction of the Waterloo
tapwater results will be given hi more detail in the Discussion
section, but we suspected colloidal Ca as a factor. A filter-
ing experiment was run to try separating colloidal from dis-
solved Ca. Filtering with 0.45- or 0.2-pm filters gave similar
results, so they are averaged together. Jack Lake water gained
4 pM Ca and 6 jiM Na upon filtering, Gait Creek water
gained 8 pM Ca and lost 6 /*M Na, while Waterloo tapwater
lost 72 /tM Ca and lost 8 /iM Na upon filtering. Small losses
or gains of Ca and Na after filtering probably represent an-
0.00 0.00 005 0.10 0.60 1.00 2.00
thtosuHate concentration (pM)
Fig. 10. Gill Ag concentrations of rainbow trout exposed to 0.09/iM
Ag for 2 to 3 h in Lake Ontario water in the presence of 0 to 2 /iM
thiosulfate (15°C). Clear bar = no Ag; striped bars = 0.09 jiM Ag
exposure, with 95% C.I. indicated. *** = significantly greater than
control values at p< 0.001. Hatched bars = model-predicted gill Ag.
alytical variability, but the relatively large loss of Ca from
filtered Waterloo tapwater may indicate the presence of col-
loidal Ca.
In addition to testing the model using a variety of waters,
we also added Ag and cadmium (Cd) together in a series of
exposures to determine if the presence of another metal
would affect Ag binding to trout gills. Cadmium was chosen
because it binds to fish gills with a similar binding constant
to that of Ag (log^cd-soicd = 8.6), and its accumulation on
gills at -0.1 pM exposure concentrations is as easily mea-
sured as is Ag [15].
The presence of Cd did not significantly affect gill Ag con-
centrations, except for a tendency for slightly lower gill Ag
concentrations (Table 3). Thus, the gill model predicts giu7
Ag equally well whether the exposure was to Ag alone (r =
0.896 (n = 8),p<0.01) or was in association with low con-
centrations of Cd (r = 0.886 (it = B),p< 0.01; including .
Waterloo tapwater, r = 0.344 (n = 9), p > 0.05). In these!
treatments, Ag and Cd concentrations were 0.09 ± 0.02 and
0.11 ± 0.01 pM, respectively (mean ± I so (n = 8)). Back-
ground Cd hi control exposures was 0.004 /tM (n = 2). The
significant difference between the controls (Table 3) was
likely an artifact of the choice of water used in the control
exposures: In the Ag-plus-Cd experiments the controls (no
metals added) were from Salmon Lake and Lake Ontario
(six fish from each), whereas the Ag controls were from all
waters (four fish from each) except Salmon and Jack lakes.
For completeness, gill Cd concentrations are included in
Table 3: Of most interest is the Waterloo tapwater result,
where Cd was kept off trout gills.
DISCUSSION
We were able to calculate the equilibrium binding constant
for Ag binding to rainbow trout gills (log AA,.gmAg = 10.0),
with about 1.3 nmol Ag binding sites per fish. The binding
constant was determined from complexation experiments
with thiosulfate (log A^jOj = 8.8; Figs. 2 and 4). Know-
ing the Ag-gill binding constant, Ag complexation by Cl
(Fig. 5) and by dissolved organic carbon (DOC; Fig. 6),
-------
Modeling silver binding to gills of rainbow trout
1855
R.O. GaltCr Salmon Grand R. Jack Oundas Ontario Waterloo
Fig. 11. Observed and predicted gill Ag concentrations for a series of lake waters to which 0.06 to 0.12 pM Ag was added. To scale the pre-
dicted values to the observed values, the predicted value for the reverse osmosis (R.O.) water was set to 1007* of the observed value. Means
with 95% C.I. are indicated. *, •*, **• = significantly greater than control values at *p < 0.05, **p < 0.01, and ***p < 0.001, respectively.
Controls are the mean of 24 fish held in all waters except Salmon Lake and Lake Ontario, four fish each. See Table 2 for full names and chemistry
of the waters.
and competition for Ag binding sites by Na (Fig. 7), Ca, and
H+, we were able to calculate other equilibrium binding con-
stants. These log K values were inserted into the MINEQL*
aquatic chemical equilibrium model [11] to model Ag-gill in-
teractions. This model (summarized in Fig. 8) was used to
predict Ag accumulation on gills of trout held in Lake On-
tario water supplemented with Ag and tniosulfate (Fig. 10),
and to predict gill Ag of trout held in Ag-supplemented nat-
ural waters (Fig. 11). Variables needed for prediction of gill
Ag were water pH and molar concentrations of Ag, DOC,
Ca, Na, and Cl (Table 2).
Although the model predicted gill Ag concentrations well
most of the time, it did not predict the high gill Ag concen-
trations in the Waterloo tapwater (Figs. 11 and 12). Water-
loo tapwater is very hard and is a mixture of well water and
Grand River water. This exception to the otherwise good pre-
dictions by the model may indicate a kinetic constraint on
the thermodynamic model. If the high concentrations of Ca,
Na, and Cl, and moderate amounts of DOC (Table 2), are
not available or are only slowly available to interact at Ag
binding sites at the gills or with Ag in the water, then their
protective effects will be overestimated. For example, Ca ex-
isting as colloidal CaCO3 would not necessarily be readily
available to interact at the gills. Our filtering experiment in-
dicated that some Ca may be in large enough colloids to be
retained by 0.45- and 0.2-fim filters, but most of the Ca,
whether colloidal or free in solution, passed through the fil-
ters. Colloidal particles passing through filters is a concern
when trying to speciate elements [18].
The possibility of colloidal Ca being slow to interact at
fisH gills and thus slow to protect against Ag deposition on
fish gills is analogous to organic-metal colloids controlling
Table 3. Statistical comparison (unpaired t test) between gill Ag concentrations
from the Ag and the Ag-plus-Cd-supplemented waters
Gill Ag (nmol Ag/g wet tissue)
Water
Ag exposure Ag + Cd exposure
/test
Gill Cd (nmol Cd/g wet tissue)
Ag + Cd exposure ANOVA
Controls
Reverse osmosis
Gait Creek
Salmon Lake
Grand River
Jack Lake
Dundas Pond
Lake Ontario
Waterloo tapwater
1.7 ± 1.1 (24)
11.2 ±3.1 (4)
6.8 ± 3.6 (8)
21.6 ± 9.3 (8)
7.8 ± 3.6 (8)
17.8 ± 8.6 (8)
9.9 ± 5.3 (8)
10.8 ± 2.3 (8)
23.8 ±1.4 (4)
2.S ± 0.8 (12)
8.5 ± 3.5 (6)
5.7 ± 1.3 (6)
16.4 ± 3.0 (6)
6.8 ± 1.3 (6)
14.2 ± 3.5 (5)
8.5 ± 2.0 (6)
9.5 ± 3.2 (6)
24.2 ± 5.4 (6)
p-0.03
p = 0.26
p = 0.49
p = OM
p = 0.57
p = 0.39
p = 0.55
p = 0.39
p = 0.91
•>
NS
NS
NS
NS
NS
NS
NS
NS
3.3 ± 0.5 (12)
7.1 ±1.6 (6)
4.8 ± 0.5 (6)
6.1 ± 1.5 (6)
4.3 ± 0.1 (6)
6.6 ± 1.4 (6)
5.2 ±0.7 (6)
5.5 ±0.9 (6)
4.1 ± 0.5 (6)
_
•**
•
•**
NS
***
•»
•*•
NS
Mean ± 1 so (it). The correlation between gill Ag concentrations of the two exposures was r = 0.967 (9), significant
at p < 0.001. Gill Cd concentrations for the Ag-plus-Cd exposure are also given (mean ± 1 so (n)). In this case the
statistical test (one-way ANOVA followed by Student-Newman-Keuls test) compared gill Cd of exposed fish with
control values.
*p < 0.05. **/> < 0.01, ***p < 0.001.
-------
1856
N. JANES AND R.C. PLAYLE
30
20
TJ
£
3
a
10
10
predicted gill Ag
20
Fig. 12. Comparison of observed and predicted gill Ag concentra-
tions (in nmol Ag/g wet tissue) presented in Figure 11. The open cir-
cle represents the Waterloo tapwater result, which was not included
in the linear regression. See text for details.
trace-metal speciation in seawater [18]. The "onion" model
[18] suggests the existence of colloids composed of layers, so
that only the outer layer of a colloid is available to interact
in the water. It would therefore take time for metals on in-
ner layers of the colloid to react with a ligand, for example.
Calcium in Waterloo tapwater does not necessarily need to
be in colloidal "onion" form to reduce its effectiveness at
keeping Ag off the gills; it just needs to be in some form such
as colloidal CaCOj, which may only slowly dissociate. Sim-
ilarly, Na, Cl, and DOC in the water may not be in available
forms to be protective. Thus, a kinetic constraint on the
thermodynamic model may exist for some waters such as
hard well water.
Why, then, did Waterloo tapwater keep Cd off trout gills*
(Table 3), but did not keep Ag off the gills? The answer may
lie in the relative Cd:Ca binding at Cd binding sites at the
gill, compared to Ag:Ca binding at Ag binding sites. Log
*ca-tiiicd = 8-6, and log Kc+jacd = 5.01151, a difference of
log 3.6 (-4,000 times). In contrast, log A^UA, = 10.0, and
log*c«-tuiA« = 3.3 (Table 1; Fig. 8), a difference of log 6.7
(-5 million times). Another way of viewing this question is
that 1 mM Ca kept O.OS pM Cd off the gills (a 20,000:1 ra-
tio; [IS]), whereas 10.6 mM Ca did not keep O.OS pM Ag off
the gills (a 212,000:1 ratio; see Results). Thus, Ca in a slowly
dissociating colloidal form would not reduce the protective
effect of Ca against Cd accumulation on gills as much as the
protective effect of Ca against Ag, because there would likely
be enough Ca in the free, fast-reacting form to interact at Cd
binding sites on the gills to keep Cd off the gills.
Although our methods do not identify specific binding
sites on the gills for Ag, the strength of Ag binding at the
gill (log A'Ag.guiA, = 10.0) suggests covalent binding to, for
example, sulfhydryl groups. Silver has a low ionic index (0.8)
and a high covalent index (2.9; see [8,19]), which also in-
dicates covalent rather than ionic binding (class B metal
cation; [20]). Cadmium also binds relatively strongly to gills
flog ^cd-»mcd = 8.6; [IS]). Cadmium has both a high ionic
index (4.1) and high covalent index (2.9), so Cd may form
ionic interactions with, for example, carboxyl groups [8], or
may also bind covalently with sulfhydryl groups. In our ex-
periment Cd did not interfere with Ag binding to trout gills
(Table 3), which suggests that these metals bind at different
sites on the gills. Cadmium interferes with Ca uptake at fish
gills [21], whereas Ag interferes with Na and Cl balance in
fish [22], again suggesting different binding sites.
Taken individually, unnaturally high concentrations of
DOC, Cl, Na, Ca, and hydrogen ion were required to keep
Ag off trout gills. Dissolved organic carbon at > 14.6 mg C/L
kept all 0.17 /iM Ag off the gills (Fig. 6), while our lakewater
values were £8.5 mg C/L (Table 2). Previously, ^4.8 mg
C/L DOC was needed to keep about 0.27 pM Cu off fish gills
[14]. With similar DOC-metal binding of about log K=9 for
each metal, these results corroborate the high affinity for Ag
at the gills (logA^u^ = 10.0) compared to Cu binding
at the gills (log Acn-fincn = 7.4; [IS]). The log AA«-DOC and
log ACU.DQC values are in the weak range, as defined for Cu-
humic acid interactions in seawater [13].
Chloride alone did not keep Ag off trout gills at rea-
sonable concentrations: >1.5 mM Cl was needed to keep
0.11 j«M Ag off the gills (Fig. 5). Average North Ameri-
can freshwater contains about 230 pM Cl ([10]; also see our
Table 2). LeBlanc et al. [3] found no toxicity of Ag in soft
water at a Cl:Ag molar ratio of 16,000:1. In our experiments,
103,000:1 Cl:Ag kept all Ag off the gills, and 14,000:1 Cl:Ag
did not keep Ag off the gills (Fig. S). The 16,000:1 Cl:Ag
ratio for no toxicity [3] is intermediate between our no pro-
tection and complete protection ratios. However, LeBlanc
et al. [3] added Cl as NaCl, so would have had some addi-
tional protective effect due to Na (see below).
Sodium, like Cl, had very little effect on Ag accumula-
tion on trout gills until very high concentrations were reached
(16 mM; Fig. 7). Average freshwater has about 390 fM Na
([10]; also see our Table 2). Calcium was expected to com-
pete with Ag for gill binding sites, but did not do so at up
to 10.6 mM. Average freshwater has about 500 /»M Ca ([10];
our Table 2). Similarly, H* did not successfully compete
with Ag even at pH 4.S. Calcium and H* therefore interact
at Ag binding sites at gills very weakly (Fig. 8).
Taken together, complexing agents such as DOC, Cl, and
thiosulfate, and competing solutes such as Na and Ca. ex-
erted cumulative protective-effects against Ag accumulation
on trout gills. For example, in the Ag and thiosulfate expo-
sures in Lake Ontario water, only S.6 times more thiosulfate
than Ag was needed to keep Ag off the gills (Fig. 10), as
opposed to the 29:1 S^.'Ag ratio in synthetic soft water
(Fig. 2). Less thiosulfate was needed to keep Ag off trout
gills in Lake Ontario water because of the higher DOC, Ca,
Na, and Cl concentrations in Lake Ontario water (Table 2)
compared to synthetic softwater (see Results for the first thio-
sulfate experiments). The ratio of thiosulfate to Ag in Lake
Ontario water (5.6:1) is remarkably close to that from Le-
-------
Modeling silver binding to gills of rainbow trout
1857
Blanc et al. [3], where a measured 5.7:1 SjOjiAg molar ra-
tio eliminated Ag toxicity to fathead minnows.
Cumulative protective effects of water chemistry were also
seen in our experiments designed to test the Ag-gill inter-
action model. Waters of lowest DOC and ion content (e.g.,
Salmon Lake) showed highest trout gill Ag accumulations
(both measured and predicted), and waters of highest DOC
and ion content (e.g., Grand River) showed lowest gill Ag
concentrations (Fig. 11; Table 2). The notable exception, as
discussed earlier, was Waterloo tapwater.
Sodium efflux from trout to the exposure water was gen-
erally high when gill Ag concentrations were high (Fig. 9).
Metals stimulate ion effluxes by affecting the permeability
of tight junctions of gill epithelial cells, allowing an ion such
as Na to passively diffuse down its electrochemical gradient
[23]. Ion fluxes to water are a potential noninvasive tool for
assessing metal effects on fish [7]. Our results iMlcate a Ag-
induced impairment of ionoregulation in trout, which agrees
with preliminary results of Wood et al. [22]. In their study,
adult rainbow trout showed large losses of plasma Na and
Cl (from about 140 to 100 mmol/L over 6 d) in response
to -0.09 ftM Ag exposure in Lake Ontario water [22]. It is
not known whether the losses were due to increased efflux
or to decreased active uptake of Na and Cl.
In our 6-d exposure to Ag, trout exposed to Ag in the ab-
sence of thiosulfate showed high initial Ag accumulations on
the gills, followed by a decline in gill Ag (Fig. 3). This accu-
mulation pattern has been seen for other metals such as alu-
minum [24], and likely represents initial metal accumulation
followed by sloughing of the metal with mucus. Mucus se-
cretion at the gills is a standard acute response of fish to a
toxicant [25]. Note that we had no fish mortalities in our
6-d experiment despite the fact that the Ag concentration
(0.06 nM) in soft water without thiosulfate was within the
range of LCSOs for soft water (e.g., [1]).
The small but significant accumulation of Ag on gills of
fish exposed 147 h to Ag plus thiosulfate (Fig. 3) may be a
result of enough time for the very low concentration of free
Ag"1" to interact at the gills, likely through a disjunctive path-
way of ligand exchange [26]. Alternatively, this Ag accumu-
Jation may represent diffusion of the AgS2O3 complex into
the gills. Although acute accumulation of metals on the gills
is due to free ions, not complexed metal [14], diffusion of
complexed metal may become important over the longer term.
In conclusion, we have been able to determine Ag-gill
equilibrium binding constants, and we have inserted these
values into an aquatic chemistry equilibrium model to pre-
dict Ag interactions at trout gills. The model considers com-
plexation of Ag in the water surrounding a fish and cation
competition for Ag binding sites on the gills. In essence, we
have inserted biological components into a powerful aquatic
chemistry program. This approach will ultimately allow bet-
ter understanding and prediction of interactions of metals
such as Ag with sensitive biological membranes.
Acknowledgement-We thank Kent Burnison for supplying and
analyzing the DOC used in this study. We also thank Lydia HoUis
and Diane Stanley-Horn for their capable laboratory assistance.
Evelyn Playle assisted with lake and pond water collection. This re-
search was supported by internal research grants from Wilfrid Lau-
rier University and by a research grant from the Natural Sciences and
Engineering Research Council of Canada.
REFERENCES
1. Davits, P.H., J.P. Goetd, Jr. and J.R. Sinley. 1978. Toxicity of
silver to rainbow trout (Salmo gairdneri). Water Res. 12:113-
117.
2. Nebeker, A.V., C.K. McAnliffe, R. Mshar and D.G. Stevens.
1983. Toxicity of silver to steelhead and rainbow trout, fait,
head minnows and Daphnla magna. Environ. Toxicol. Chem.
2:95-104.
3. LeBUnc,GA.,J.D.MastODe,A.P.Ptradice,BXWQson,H.B.
Lockhart, Jr. and K.A. RobOIard. 1984. The influence of spe-
ciation on the toxicity of silver to fathead minnow (Pimepha-
les promelas). Environ. Toxicol. Chem. 3:37-46.
4. Diamond, J.M., D.G. Madder, M. Coffins and D. Gruber. 1990.
Derivation of a freshwater silver criteria for the New River, Vir-
ginia, using representative species. Environ. Toxicol. Chem.
9:1425-1434.
5. Taylor, M.C., A. Demayo and S. Reeder. 1980. Guidelines for
surface water quality, Vol. 1—Inorganic chemical substances,
silver. Environment Canada, Inland Waters Directorate Water
Quality Branch, Ottawa, Ontario.
6. Lauren, DJ. 1991. The fish gill: A sensitive target for water-
borne pollutants. In M.A. Mayes and M.G. Barron, eds.,
Aquatic Toxicology and Risk Assessment: Fourteenth Volume.
STP 1124. American Society for Testing and Materials, Phila-
delphia, PA, pp. 223-244.
7. Wood, C.M. 1992. Flux measurements as indices of H+ and
metal effects on freshwater fish. Aquat. Toxicol. 22:239-264.
8. Reid, S.D. and D.G. McDonald. 1991. Metal binding activity
of the gills of rainbow trout (Oncorhynchus my kiss). Can. J.
fish. Aquat. Sci. 48:1061-1068.
9. Pagenkopf, G.K. 1983. Gill surface interaction model for trace-
metal toxicity to fishes: Role of complexation, pH, and water
hardness. Environ. Sci. Technol. 17:342-347.
10. Morel, F.M.M. and J.G. Hering. 1993. Principles and Applica-
tions of Aquatic Chemistry. John Wiley & Sons, New York, NY.
11. Schecher, W.D. and D.C McAvoy. 1992. MINEQL+: A soft-
ware environment for chemical equilibrium modeling. Comput-
ers, Environment and Urban Systems 16:65-76.
12. Midorikawa, T. and E. Tanone. 1994. Detection of a strong
ligand for copper in sea water and determination of its stabil-
ity constant. Anal. Chim. Acta 284:605-619.
13. Hirose, K. 1994. Conditional stability constants of metal com-
plexes of organic ligands in sea water: Past and present, and a
simple coordination chemistry model. Anal. Chim. Acta 284:
621-634.
14. Playle, R.C, D.G. Dixon and K. Burnison. 1993. Copper and
cadmium binding to fish gills: Modification by dissolved organic
carbon and synthetic ligands. Can. J. Fish. Aquat. Sci. SO:
2667-2677.
15. Playle, R.C., D.G. Dixon and K. Burnison. 1993. Copper and
cadmium binding to fish gflls: Estimates of metal-gill stability
constants and modelling of metal accumulation. Can. J. Fish.
Aquat. Sci. 50:2678-2687.
16. SlgmaSUt. 1992. SigmaStat Statistical Software, Version 1.02.
Jandel Scientific, San Rafael, CA.
17. Allison, J.D., D.S. Brown and KJ. Novo-Gradac. 1991.
MINTEQA2/PRODEFA2, a geochemical assessment model for
environmental systems: Version 3.0. User's manual. U.S. En-
vironmental Protection Agency, Washington, DC.
18. Mackey, DJ. and A. Zlrino. 1994. Comments on trace metal
speciation in seawater or do "onions" grow in the sea? Anal.
Chim. Acta 284:635-647.
19. Nieboer, E. and D.H.S. Richardson. 1980. The replacement of
the nondescript term 'heavy metals' by a biologically and chem-
-------
1858
N. JANES AND R.C. PIAYIE
ically significant classification of metal ions. Environ. Pollut.
18:3-26.
20. Sturom, W. and J J. Morgan. 1970. Aquatic Chemistry. John
Wiley & Sons, New York, NY.
21 Verbost, P.M., J. Van Roolj, G. Flik, R.A.C. Lock and S.E.
Wenddaar Bonga. 1989. The movement of cadmium through
freshwater trout branchial epithelium and its interference with
calcium transport. /. Exp. Biol. 145:185-197.
22. Wood, C.M., S. Monger and C. Hogstrand. 1993. The physio-
logical mechanisms of toxkity of sflver and other metals to fresh-
water fish. Proceedings, First International Conference on
Transport, Fate, and Effects of Silver in the Environment, Mad-
ison, Wl. August 8-10, pp. 89-92.
23. Evans, D.H. 1987. The fish gill: Site of action and model for
toxic effects of environmental pollutants. Environ. Health Per-
spect. 71:47-58.
24. McDonald, D.G., CM. Wood, R.G. Rhem, M.E. Mueller, D.R.
Mount and H.L. Bergman. 1991. Nature and time course of ac-
climation to aluminum in juvenile brook trout (Salvelinusfon-
tinalis): I. Physiology. Can. J. Fish. Aquat. Sci. 48:2006r2015.
25. Florence, T.M., G.M. Morrison and J.L. Sunber. 1992. Deter-
mination of trace element speciation and the role of speciation
in aquatic toxicity. Sci. Total Environ. 125:1-13.
26. Hering, J.G. and F.M.M. Moid. 1990. Kinetics of trace metal
complexation: Ugand-exchange reactions. Environ. Sci. Tech-
nol. 24:242-252.
APPENDIX
Selected equilibrium constants (K) from the MINEQL+ com-
puter program (version 2.0) (SjO3 = thiosulfate).
Complex
log/:
AgSjOJ
Ag(SjOj)j~
AgCl
Agcir
AgaJ-
AgCl*~
AgNO,
NaSjOJ"
NaCO,~
NaHCOj
CaSp,
CaCO,
CaHCO,*
HSjOf
HCOj"
H2C03
8.80
13.60
3.27
5.27
5.29
5.51
-0.29
0.60
1.27
10.08
1.90
3.15
11.33
1.70
10.33
16.68
-------
In press in Environmental
Science and Technology
vi
r
Binding of Nickel and Copper to
Fish Gills Predicts Toxicity When
Water Hardness Varies, But Free-Ion
Activity Does Not
JOSEPH S. MEYER.' '
ROBERT c. SANTQRE'* IOE p unRmjr/n
:ARRY D. lH-BTU-rt-JcoNME j. BOESEK)
PAUL R. PAQUIN,' HERBERT E. ALLEN,'
HAROLD L. BERGMAN.* AND
DOMINIC M. DITORO*
Department of Zoology and Physiology, University of
Wyoming, Laramie, Wyoming 82071-3166. HydroQual inc..
1 Lethbridge Plaza. Mahwah. New Jersey 07430. and
Department of Civil and Enilronmental Engineering,
University of Delaware. Newark. Delaware 19716-3120
Based on a biotic-ligand model (BLM), we hypothesized
that the concentration of a transition metal bound to fish gills
([Mgu]) will be a constant predictor of mortality, whereas
a free-ion activity model is generally interpreted to imply that
the chemical activity of the aquo ("free") ion of the
metal will be a constant predictor of mortality. In laboratory
tests, measured (Ni^) and calculated ICUgJ were
constant predictors of acute toxicity of Ni and Cu to
fathead minnows (Pimephales promelas) when water
hardness varied up to 10-fold, whereas total aqueous
concentrations and free-ion activities of Ni and Cu were
not Thus, the BLM. which simultaneously accounts for (a)
metal speciation in the exposure water and (b) competitive
binding of transition-metal ions and other cations to
biotic ligands predicts acute toxicity better than does free-
ion activity of Ni or Cu. Adopting a biotic-ligand modeling
approach could help establish a mere defensible,
mechanistic basis for regulating aqueous discharges of
metals.
lotroriictioi
For several decades, water hardness (the sum of the nor-
malities of the divalent cations in solution) has been known
to affect toxicity of transition metals to aquatic organisms.
with mortality at a specified total dissolved metal concentra-
tion decreasing as hardness increases (J). The U.S. Envi-
ronmental Protection Agency (USEPA) method to compen-
sate for the hardness effect in site-specific discharge permits
has been to calculate a hardness-adjusted LC50 (median
lethal concentration at a specified exposure time) using a
regression equation of the form
ln(LC50) = a-]n(H; T b
(1)
where H = water hardness and a and b are regression
coefficients fined to toxJciry data (e.g., refs 2.3}. However.
• Corresponding author phcne (307 766-2017-tax (307.766-
5625: e-mail: meyeo@uwyo.edu.
1 University of Wyoming.
« HydroQual. Inc.
» University of Delaware.
hardness is the only water-quality parameter for which such
a procedure is used. Other water-quality parameters that
modify metal toxicity (e.g.. pH, alkalinity, dissolved organic
carbon (DOC), total suspended solids (TSS)) can be accounted
for only by using costly acd time-consuming, site-specific
toxicity testing and tortuous calculations that still have no
underlying mechanistic bads (e.g., water-effects ratios (4)).
To provide a more mechanistic basis forpredicting toxicity
of metals to aquatic biota, the free-ion-activity model (FIAM)
was proposed (5. 6). Based mainly on research in which
toxldty correlated with free-ion activity as the concentrations
of organic ligands and dissolved organic matter were varied
in exposure solutions (Tables 1 and 7 in ref 6). the FIAM has
generally (but incorrectly) been interpreted to imply that a
constant degree of biological effect (e.g.. mortality) win occur
at a constant chemical activity of the aquo C free") ion of a
metal ({M2'} for divalent transition-metal cations such as
Cd2-. Cu2". NT2". Pb2-. andZn2-). independent of other water-
quality parameters. However, free-ion activity does not
appear to be a good predictor of toxicity across some water-
quality conditions (6), including when water hardness varies.
Building on that concept a recently resurrected surface-
interaction model (7) of metal binding to fish gills not only
incorporates an equilibrium between {M2'} and the con-
centration of metal accumulated at binding sites on the giU
surface (fl) but also incorporates competition of M2' with
other cations (e.g., Ca2', H~] for binding at the receptor sites.
Historically. Zitko and Carson (9) suggested the cation-
competition concept prior to Pagenkopf s (7) formal pre-
sentation of a chemical-spedation model that incorporated
a biotic ligand. And although Morel (5. p 303) stated in his
explanation of the FIAM, *_ the free metal ion activity is the
parameter that determines the physiological effect of die
metal ~.", he later specifically mentioned (a) the importance
of cation competition for binding sites on the surfaces of
biotic ligands (5. p 307) and (b) the relationship between
measurable physiological effects and the degree of cora-
plexation of metals with reactive ligands on (or in) organisms
(5, p 303). According to this type of model which we refer
to as a biotic-ligand mode! (BLM) to distinguish it from the
current general interpretation of the FIAM. free-ion activity
is anecessary but not sufficient component to describe metal
accumulation and, presumably, toxicity. Thus, bioavaflabiliry
of metals can be decreased in two ways: (1) by decreasing
{M2-} and. thus, decreasing the potential for M to bind to
receptor sites (e.g.. as increasing [DOC] does through
equilibrium partitioning aaong the dissolved ligands) or (2)
by increasing the concentrations of competing cations and,
thus, decreasing the amount of M bound to receptor sites
(e.g.. as increasing {Ca2'} does).
We hypothesized that if the BLM is correct, the amount
of a specified metal accumulated on fish gills (JMgaJ. expressed
as mol M-g tissue*1) will be a constant predictor of mortality.
independent of other water-quality parameters (except when
{H-} becomes high enough to cause add toxicity). Borgmann
(7) and Pagenkopf (7) alluded to but did not directly test this
concept for fish exposed to transition metals, although it
later was tested with Atlantic salmon (Salmo solar] exposed
to Al and Zn as pH varied '10). Herein we present the first
published evidence that {M2'} is not a constant predictor of
metal toxicity to fish as water hardness increases, but !M,aJ
is.
Experimental Section
Ni Toxicity Tests. We exposed subadult (1-6 g) fathead
minnows (FHM; PimephcLes promelas) to NiSO4 in four 96-
10 102Ves9M71Sq CCC J18 00
Published on Web 00/00,0000
± auut Ame-can CtwniaJ Scatty
PAGE EST 3.3
VOL XX. NC a. MOCK , ENVIRON SZl & TECHNCU • A
-------
h. continuous-Dow toxidty tests. In each test, a control and
11 serially diluted concentrations of Si (25% decrease in
[NitouiJ at each serial dilution) were tested in a mixture of
well water and reverse-osmosis-treated, deionized (RO-DQ
well water at the University of Wyoming's Red Buttes
Environmental Biology Laboratory. Exposure waters con-
tained a different concentration of Ca and a different series
of N1SO4 concentrations in each test, but temperature, pH.
alkalinity, and [DOC1 remained constant Water-quality
parameters (average or range of values) were as follows:
temperature. 20 °C pH. 73 (range = 7.2-7.5); alkalinity. 0.5
mN: DOC. <03 mg CM.'1; and major ions (mN) ICa2' - 03-
4.8 (varied among tests by adding 0. 1.0.2.6. or 4.5 mN of
Cadj to the base mixture of well water and RO-DI water to
adjust hardness); Q~. 0.02-4.7 (varied among tests as
addition of CaCt varied): K: 0.03; NV. 0.04; Mg2*. 021:
NOT. 0.02; SO4*~, 0.03-6.7 (varied within and among tests
as addition of N'iSO, varied)).
For each Ni concentration and the control, 20 FHM were
placed in one aquarium, and 10 or 15 FHM were placed in
a second aquarium. Fish in the first aquarium were monitored
for survival through 96 h. and the [NinaJ LC50 was calculated
by linear regression of the logit transformation of mortality
(ln(m/(l-m)). where m is the mortality proportion) on In-
([Xinud). The LC50 was the [Mom) at which the predicted
logh(mortality) equalled zero.
Fish in the second aquarium were removed in groups of
5 at 24 h. and their gills were excised, rinsed in control water
for 10 s. blotted dry. wet-weighed, and digested at 86 C for
6-8 h in an equrvolume mixture of 70% trace-metal-grade
HNOj (prepared in ultrapure water) and tfeOj (30%). The gill
digestates were then diluted to a known volume and analyzed
for [Niffl] by flame atomic absorption spectroscopy (MS).
Aqueous (Niu^J also was analyzed by flame AAS. Quality
control for Ni analyses included repeated injections of
samples (typically s5% relative standard deviation among
Ni concentrations determined in replicate injections) and
periodic analysis of certified reference standards (a prior set
of samples was rerun if analyzed vahie differed from the
reference standard's certified value by >20%). We sampled
gills at 24 h because results of preliminary studies indicated
accumulation of Ni by FHM gills was relatively rapid and
consistent with a one-compartment uptake-depuration
model in which the amount of Ni accumulated at 24 h was
~85% of its predicted asymptotic value. Thus, sufficient Ni
associated with (i.e.. accumulated on or in) the gills before
the majority of FHM began dying in exposure concentrations
that bracketed the LC30. Median lethal accumulations (LASOs.
analogous to the term LC50) of Ni on FHM gills were
calculated by regressing logit mortality) on Inl.'NigoJ).
Based on (a) the measured exposure-water quality and
(b) published stability constants (but with log K values
adjusted to zero ionic strength using the Davies equation
(11)) for complexes of the major anions with the major cations
and Ni (12). we calculated [NP-J and {Mi*-} at the (NiauJ
LC50 for each toxicity test using the geochemical spedation
program MINTEQA2 (13).
Ca Toxidty Calculations. To further test our hypothesis,
we analyzed published acute-toxicity data for FHM exposed
to CuSO4 at various water hardnesses. Erickson et aL (14)
conducted 12 96-h static-nonrenewal toxidty tests in which
larval FHM were exposed to CuSO4, and [Ca2-] and [Mg2']
were adjusted to vary water hardness among the tests (tests
•none' and 'add 2 mN CaSCV in set S2 and all 10 tests in
set S3 in their Table 2. all of which were conducted at pH
7.8). We calculated {Cu2"} and [Cu^i] at the [Cuc»i«J LC50
in each test by entenng ICUi^oi^J, other inorganic water-
quality parameters (IS], and IDOCI 0 8 mg-L"1) into the
geochemical specianon program CHESS (16). For these
calculations, the CHESS model was used to simulate (a) Cu-
2.
II
II
in <*£
Uo
[NigJ
Hardness (mN)
RGURE1. Measured 9S-hLC5fe (Median ledulcoMeatrations) for
INwl [NP-L and {HP*} in exposure witers and USDs (modiu
lethal •emulation. express* oo • wret weight
-------
5
0.2,
o 0.0
«•«. o.u* -
.s >
9^
*~i '» 0.01 -
fit
,3 0.00 .
C.
..* - : • —
-P
Ca (mN)
R6URE2. (a) Measured 96-hLCSh(medfee lethal cettentratioBs)
for f-fc^i.nl in exposure waters, (b) calculated 96-b LCSOs for
{Co1-}, and (c) calcalated LASfe (mediaa letfaal accumulation,
expressed on a wet weight (ww) of tissee basis) for F^yJ for
fathead minnows (FHM) exposed to CaSOi at various water
baldnesses (M). LCSOs are Ci coacentratioe flD or cbeaucal activity
({}) averaged over the 96-b exposare; lASOs are calcalated tissae
harden of to associated with HOI gills at 2-3 b (but corresponding
to 50% mortality at 96 b). Least-satires linear regression slopes are
(a) U08 [P< OJ001L (b) OJM17 (T < 0001L and (c) 0000710 (P =
0308).
explained as foDows. First, the decreasing slope at higher
[Ca] might result from competition between Car' and Cu2'
binding to the -1 rng-L'1 of DOC that was present in the
Lake Superior water in their toxicity tests (14). As a resuh of
this cation competition. {Cu2*} would have increased as ICa]
increased at a given [Civ&nmd]—thus, tending to increase
FHM mortality (i.e., decreasing the LC30 below the vahie
expected in the absence of DOM). Winner (18) reported such
a hardness-mediated increase in toxitiry of Cu in me presence
of hurnic acid, and Penttinen et aL (19: reported a loss of the
protective effect of DOM as water hardness increased when
Daphnia magna were exposed to Cd. Second. Cu disrupts
ionoregulation in fish (20). Because gills may be more
permeable to body ions at low [Ca] than they are at higher
[Cal (2Q). the lower than expected Oidi»oM LCSOs at low
ICaj might result from diminished, direct physiological
protection by gill-bound Ca—a process that is not accounted
for by our BLM. This might also explain why the LASOs of
[Nifjul (Figure 1) and (CUfodcu (Figure 2c) appear to increase
slightly at low hardness and approach plateaus at high
hardness. Third. Erickson et aL (14) only plotted results from
their toxicity test set 52 in their Figure 3. and the nonlinearity
of their Figure 3 was influenced strongly by the two extreme
LC50 values. However, when we included two additional data
points from their toxicity ten set S2 with their set S3, the
trend appears to be more linear at the high [Ca. end of our
curve (Figure 2a) than it does at the high hardness end of
their curve. Finally, combined with ute uncerainty about
the LC50 estimates 'see 95% confidence intervals in Table
2 and Figure 3 in (741). the apparent curvilinear relationship
at low [Ca, in Figure 2a might even be an artifact
We conclude that accumulations of Ni and Cu on FHM
gills are approximately constant predictors of toxicity when
the concentration of Ca (the major competitor with Ni and
Cu for binding to the gill) in exposure waters increases.
whereas the free-ion activities of Ni and Cu in exposure waters
are not constant predictors cf toxftiry. This also appears to
be valid when pH is varied among Cu toxicity tests, although
{Cu2-} is just as good a predictor of toxicity as [Cu^oc is
when JDOC! is varied (2/). And, emphasizing the importance
of cations other than Ca2" as competitors with some metals.
K- (but net Ca1*) ameliorates acute and chronic toxichy of
thallium (TI) to the amphipod Hyaklla azuca, and body
burden ofTlis an approximately constant predictor of chronic
lethality and growth effects across a range of aqueous K~
concentrations (221. Therefore, the BLM. which simulta-
neously accounts for (a) metal spedation in me exposure
water and .*b) competitive binding of transition-metal ions
and other cations to biotic Hgands predicts acute toxicity
better than does free-ion activity of Cu, Ni. and TL
Adopting a biotic-ligand modeling approach could help
advance the regulation of aqueous discharges of metals
beyond the current phenomenological approach and es-
tablish a more defensible, mechanistic basis. In the future.
regulatory limits might be based on accumulation of metals
on biotic figands (e.g.. fish gffls, soft tissues of invertebrates.
algal ceflsj measured in the field or predicted in dynamic
simulation models that would estimate the number of daily
exceedences of a regulatory Emit at a site downstream from
a metal discharge (21}. However, H~ (JO) and Caf~ (20) bound
to fish gills perform important, direct physiological functions
(e.g, altering membrane permeabuity and ion transport)
beyond just competing with transition-metal cations for
binding sites. Such beneficial functions have not yet been
incorponued into a BLM. Thus, biotic-Egand modeling could
be used to complement (but not totally replace) toxicity
testing. b>- providing much greater temporal coverage than
is currently feasible within the financial constraints and time
limitations of standard fish, invertebrate, and algal toxicity
tests.
Ackiiwiei'ineals
This research was funded by a cooperative agreement
between the U.S. Environmental Protection Agency and the
University of Delaware; a gran from the International Copper
Association: and a grant tc J.S.M. rand others) from the
National Science Foundation's EPSCoR Prcgram. Ellen
Axtmann assisted with the geochemical speciation calcula-
tions. Comments from three anonymous reviewers improved
the manuscript
litentire Chel
(1) Borgnann. U. In Aquatic Toxicology Nriagu, }. O., Ed.: John
\VOey ind Sons: New Yon. 1983; pp 47-72-
'2) Ambient Water Quality Crxria for .•iickel - !>&. U.S. Envi-
ronmental Protection Agenrr. Washi£gton.DC1986:EPAWO/
5-66-U4.
'3) Ambient Water Quality Criteria /or Copper - 1984 U.S.
Emnrcnmental Protection Agency. Wishington. DC. 1985; EPA
440/5-*4-031.
••I) traenm Guidance on Dete~ninatton and Use cf Water EFeca
Rano V.Wrtofc. U-S. Environmental Pwtection.^ency: \\ash-
ingtcf. DC 1994- EPA/823 B-94/001.
•5) MorttF Pnncip& of Aquae Chemizry. John Wiley and Sons:
New '.ork. 1983
VOL u. NO «. «JH/EN.'RON SC. kTECHNCL«C
-------
(6) Campbell P. G. C In Metal Speciatton and Bloanf-labUiiy in
Aquatic Sjitemr. Testier. /L. Turner. D. R-Eds^ John Wiley and
Sons: Oiichester. 1995; pp 45-102.
C7) PagenkopL G. K. Entiron. Set Ttchnol 1983. 17,312-347.
(8) Playle. R. C; Dixon. D.G.- Burn-son, K. Cm./. Flsh-Aquat. Sci
1993. 50. 2678-2667.
(9) Zitko. V.: Canon. W. G. Oimophm 1976. 5. 299-303.
(10) Roy.R-R.:CampbeaP.G.CAjaaj:r<»jrtot 1995.33.155-176.
(11) Sertiz. S. NU Allison. |. D^ Periue. E M: Allen. R E: Brown.
D. S. Wtowr Acs. 1996.30.1930-1933.
(12) .VIST Critically Selected Stability Constonu of Metal Complexes
Database. Version 5.0. fflST Sundard Reference Database 4&
National InstituteofStandards indTechnology: Gafchersburg.
\rfj 1998
(13) Allison, J. D.: Blown. D. S.; Navo-Gradac. K. I. MCfTEQM
PRODEFA2. A Geochetnical Assessment Model for Environmental
Systems: Version 3.0 Uter't Manual. US. Environmental Pro-
tection Agency: Washington. DC 1991; EPA/600/3-91/021.
(14) Erickson. R. I.: Benoit. D. A.: Manson, V. IU Neboa H. P.. Jr.:
Leonard. E N. Environ. Taxuci. Chem. 1996. 15.131-193.
(15) Erickson. K J J Benoit. D. A.: MKaon.V.KAProtar)T*Taxlciijr
Factors Model for Siu-spetific Copper Water Qualm Criteria;
US. Environmental Protection Agency: Ouluth. MM. 1987
(revised 1996).
(16) Santore. R. C; DriscoO. C T. In Chemical Equl&rium and
Reaction Models: Loeppeit R. Schwab. A. P, Goldbtfg, S.. Edi:
American Society cf Agronozzy: Madisco. \V1.1953: pp 357-
J/3.
(17) Tipping, E Comput Ceosci 1994.20. 9*3-1023.
(18) IVInner. R. W. Water Res. 1983. 19.449-455.
(19) Penrtinen. S.; Kostaao. A.: Kuikonen. ]. V. K. Envinn. ToxiosL
Chem. 1998.17.24£6-2503.
120; McDonald. D. & Reader. I. P.- DalzieL T. R. K. In Acid Toiiarr
and Aquatic Animate Morris. -L R.. Ta>iar. E W.. Brown. D. j.
.V. Brovsu). A, Frls: Cambnc^e Unvieniy Press- Cambridge.
UK. 1989: pp 221-242.
CD DfToro.D.M.;AUen.H.E:Berp3an.aL:Mahoro'.).D4Me>tr.
I. S.; Paquin. P. JU Santor*. R. A. Chemisuy of copper
bioavaOabOiry: a model of acute uwdcity to fish (nanuscript.
02) Borgmann. U.: Chraai. V.; Nonrood, W. P.- Lechner I. Environ.
Poltut. 1998. 99. 105-114.
Received for review ]uty 15, 1398. Revised manuscript re-
ceived December 29,1398. Accented January 4,1S99.
ES980715Q
O • ENVIRON. SC & TECHNCU / VOL a. NO u. uu
PAGEEST 33
-------
In press in Environmental
Science and Technology
A Mechanistic Explanation for the
ln(LC50) vs In(Hardness) Adjustment
Equation for Metals
JOSEPH S. MEYER*
Department of Zoology and Physiology, University of
Wyoming, Laramie, Wyoming 82071-3166
I demonstrate that a combination of (a) competitive
binding of transition-metal cations, hardness cations, and
protons to transition-metal-binding sites on fish gills
and (b) aqueous complexation of transition-metal cations
by HCOs~ and CDs2" explains why the regression slopes of
In(LCSO) vs In(nardness) for five divalent transition metals
(Cd, Cu, Ni. Pb, and Zn) are ~1, where LC50 is the
median lethal concentration. For these calculations, I
assumed the amount of the transition metal bound to the
fish gill at 50% mortality is constant (i.e., independent of water
quality). Although the slopes theoretically should vary
between 0 and 2 (at extremely low and high hardness.
respectively), a slope of ~1 is expected at midrange hardness
(~2Q-2QO mg-L"1 as CaCOj) if alkalinity covaries with
hardness—a common condition in most laboratory toxicity
tests. But if alkalinity is held constant while hardness is
varied, a slope of ~0.5 is expected at midrange hardness.
Although predictions of LCSOs using regressions of In-
(LC50) vs In(hardness) might be acceptable for regulating
discharges of transition metals to waters in the midrange
of hardness, extrapolations beyond this range might drastically
overpredict metal toxicity.
Introduction
Toxicity of transition metals to aquatic biota varies as a
function of several water-quality parameters (e.g., hardness,
pH, alkalinity, concentration of dissolved and paniculate
organic matter) (7,2). The U.S. Environmental Protection
Agency (USEPA) water quality criteria for Cd, Cu, Ni, Pb, and
Zn (3- 7) specifically account for die increase in LC50 (median
lethal concentration} as water hardness increases, using
regression equations written in the following general form
ln(LC50) = a-ln(haidness) + b (1)
where a = slope, b = ordinate intercept LC50 is in units of
/ig-L-', and hardness is in units of mg-L"1 as CaCQj. The
form of this equation has never been mechanistically justified;
instead, eq 1 merely provides a convenient, empirical fit to
the data. Surprisingly, though, the slopes for all five transition
metals are ~1 (range - 0.76-1.27; Table 1).
In this paper, I propose an explanation for these apparently
coincidental relationships between LC50 and water hardness.
This argument is based on the concept of competitive binding
of cations to fish gills (8,9).
Model
1 present the model in terms of H~, Ca2*, and a generic.
divalent transition-metal ion M2' binding to fish gills.
• Corresponding author phone (307)766-2017; lac (307)766-5625;
e-mail: meyeij@uwyo.edu.
TABU 1. Coefficients for the Equations Predicting Acceptable
Concentrations (pg-L-'i of Transition Metals is i Function of
Water Hardness (rng-l-' is CiCOj) for Freshwater Biota
Exposed to Cd. Co. lli. Ph. and Zir5
metal slope* intercept* dependent variable re!
Cd 1.128 -3.828 criterion max. concn (3]
Cu 0.9422 -1.464 criterion max. concn (4)
Ni 0.76 4.02 final acute concn (5)
Pb 1.273 -1.460 criterion max. concn (6)
Zn 0.8473 0.8604 criterion max concn (7)
• The coefficient • in eq 1. »The coefficient b in eq 1. 'The general
fonnc4trttpc«dtaorequationbdepeno*«viriable«exp{
+ intercept).
However, it also applies to Mg2* instead of (or in conjunction
with) Ca*- as the hardness modifier of toxicity. The model
also may apply to other biotic ligands such as soft tissue of
invertebrates.
Equilibrium equations for binding of Ca2* and M2* to
sites at which M binds on a fish gill can be written as (9)
and
(MegUlM]
' IM'+HegfllMI
ICasgfllMl
(2)
(3)
where XM^DM = stability constant for M2* binding to
M-binding sites on the gill (L-mol-'). Kca^tu = stability
constant for competitive binding of Ca2* to M-binding sites
on the gill (L-mol-1). IM2*] = aqueous concentration of M2*
(mol-L'1). [Ca2*J«=aqueous concentration of Ca2* (mol-L-1).
(egillMl = concentration of unoccupied M-binding sites on
the gill (mol-L"1), [MagulM] = concentration of MegfllM
complexes (mol-L-1), and (Ca»gillMl = concentration of
CaegillM complexes (mol-L-1).
For simplicity. I have ignored activity coefficients in these
equations. Moreover, I have expressed the concentrations of
the binding sites and the MagfflM and CasgillM complexes
as if they are uniformly distributed in the exposure water.
This implies that the fish gills are in equilibrium with the
entire volume of water in the exposure chamber.
Combining eqs 2 and 3.1 express [M**l as a function of
ICa2-):
(M
2+1
[CaH
,2+1
(4)
Assume (MegillM) at 50% mortality is constant (i.e..
independent of hardness, as demonstrated recently for Cu
and Ni (1O) and first proposed as a general concept by (1)
and (S)). Equation 4 can be rewritten as
ICa'1
(5)
t
where LCSQgg^isthe concentration of the aquo ion of M that
causes 50% mortality, and f is a constant (- (Xc*»0W
^M^avO'IMBgillMljonraxt^). Combining eq 5 with the equi-
librium spedation relationships in the Appendix, the LC50
of a divalent transition metal to fish can be approximated
algebraically as a function of pH, water hardness, alkalinity,
A*
-
10.102VesS807Uy CCC $1800
Published on Web 00/00/0000
C nxx American Chemtc*! Society
PAGE EST: 4.8
VOL. xx. NO KX. «u / ENVIRON SO & TECHNOU • A
-------
TABU 2. Approximations to Eq 7 it Variois Alkalinity and Hardness Conditions
baldness alkalinity approximation to tq 7*
constant
axplanatlon
ln(LC50) a 0.0-ln([Cal*l) + ln(« At low [Ca2*L [Ca2*]/lCasgillMI (= iyu.9w*iEB""viii
is approximately constant because the concentration of
unoccupied M-binding sites on the gill ([«glllM])
is approximately constant (i.e., little Ca binds to the gill,
causing little change in IMsgillMJ, [HsgillM] and
IsgillM)); thus. IndCa2*}) a InflCasgillM]) -t- In(constant).
equal to hardness In(LCSO) CM 0.0-ln([Caz*l) + ln(« Same as above, and Alk (alkalinity) approaches zero as |Ca2*l
approaches zero.
constant In(LCSO) a 1.0-ln([Ca2*]) + ln(M If [MsgillM) is constant at 50% mortality (independent
of tCa2*]), {CasglllM] will approach* COMWInt
(= [BgMIMumi] - [MsgillML- eq A-5) as Ca fills the
remaining binding sites on the gill (i.e., as H* is
outcompeted and no unoccupied M-binding sites remain).
high equal to hardness ln(LC50) a 2.0-ln([Ca**]) + ln(« Same as above, but Alk = hardness results in LC50 «
(Ca2*p «.«., ln(LC50) = ln(lCa**J) + ln(AJW +
In(constant) - ln([Ca2*]) + ln(ICa**J) + In(constant)).
• It to a composite constant that in each equation, represents the products of the individual fa and all other terms (or ratios of terms) in eq
7 that are approximately constant when hardness is extremely low or high.
low
low
high
the amount of Ca2* bound to the transition-metal-binding
sites on the fish gill and the amount of unoccupied metal-
binding sites, as follows (see derivation of eq A-4 in the
Appendix)
(Ca'+l
[CasgillM) \* ' ID-""
1+-
r + JT'-Alk) (6)
where LCSCKjmaftis the total dissolved metal concentration
at 50% moruuify, Alk is the alkalinity of the exposure water
(eq-L.-1). and k" and V" are constants. Taking the natural
logarithm of both sides results in a relationship that begins
to resemble eq 1:
tofJQr**!) - InUCaggillM])
IO-PH + r».AlkJ
ln(ifcO (7)
Because kT'-AJk «1 + IT/IO-*1 at low alkalinity and f"-AIk
» 1 + ir/lQ-f" at high alkalinity, eq 7 simplifies and attains
the same form as eq 1 when hardness is very low (e.g., <~1
mg-L-1 as CaCQj) or very high (e.g., >~1000 mg-L"1 as
CaCOj). At low hardness (either when alkalinity is constant
or when alkalinity equals hardness), the slope approaches 0
(Le.,ln(LCSO)aO.O-ln([Caz+])-Hn()k).wherefcisacomposite
constant: Table 2); at high hardness the slope approaches
either 1 or 2, depending on whether alkalinityis held constant
(slope approaches 1; i.e.. ln(LC50) a 1-0-lndCa5*]) + ln(ft);
Table 2) or alkalinity equals hardness (slope approaches 2;
i.e., bidJCSO a 2.0-lnUCa2-]) + ln(*); Table 2).
Example
To demonstrate the effects of hardness, alkalinity and pH on
transition-metal toxidty, I present results of simulations for
Cu binding to fathead minnow (FHM; Pimephalespromelas)
gills. These calculations incorporate all of the Cu complexes
and activity coefficients ignored for heuristic purposes in
the previous section and in the Appendix Although it would
be more mechanistically appropriate to express LCSOs of Cu
in units of^M and hardness and alkalinity in units of meq-L'1,
I have retained the more traditional units of/tg-L'1 for LCSOs
and ing-Ir1 as CaCO) for hardness and alkalinity to allow
easier comparison to the historic database on metals toricity.
Using toxidty data reported for FHM larvae by Erickson
et al. (11). I first calculated the amount of Cu bound to the
FHM gills at 50% mortality using the geochemical spetiation
program MINTEQA2 (12). For this and the subsequent
10000,
,T* 1000
slope
O
100
1000 10000
-1
Hardness (mg-L as CaCO,)
FIGURE 1. Predicted LCSOs of CdMi to fathead minnows as hardness
varies at pH 7. assuming that the amount of Ca bound to the gills
is constant at 50% Mortality (U, independent of water qualityJ. C«
bound to the gills was held constant at 365% of the 10~* M te-
binding she*. The solid curve b predicted LCSOs when alkalinity
equals hardness; the dashed cum is predicted LCSOs when alkalinity
is held constant at SO mg-L-' as CaCO>
calculations in this example, I assumed the density of Cu-
binding sites was 10~* M (Le., ~1 g of fish per L of exposure
water). Although this choice of binding-site density is
arbitrary, the calculations were insensitive to changes in
binding-site density ats!0-*M. Based on Figure 1 in Erickson
et al (11). I assumed the 96-h LC50 of Cu in How-through
exposures is 0.2/iM (12.7/ig-L'1) at pH 7.0 and alkalinity and
hardness equal to ~50 mg-L"1 as CaCOj. Additionally, I
replaced the default stability constants in M1NTEQA2 with
the values in Version 5.0 of the National Institute of Standards
and Technology's electronic database U3) (but with log K
values adjusted to zero ionic strength using the Davies
equation (14)) and used conditional stability constants
reported for binding of Cu2-, Ca2*, and H* to FHM gills (log
K= 7.4,3.4, and 5.4, respectively (9)). Because (a) those gill-
binding constants are not precise estimates and (b) the ionic
strength of the exposure solutions in which they were
determined was only -0.0001 M (Le.. using the Davies
equation (14), <4% correction would be needed to adjust for
nonzero ionic strength), I did not correct the gill-binding
(2
• • ENVIRON. SO. & TECHNOL. / VOL. xx. NO. xx. xxxi
-------
10000
~. 1000
o
S> 100
§
o
100 1000 10000
Hardness (mg-L*1 as CaCO3)
FIGURE 2. Predicted LCSOs of Co** to fathead minnows as pH
varies from 5 to 9. based to the assumptions that (a) the amount
of Cu bound to me gills at 50% mortality is 365% of the 10~* M
Co-binding sites and (b) alkalinity equals hardness.
constants to zero ionic strength. Under these conditions at
25 °C 36.5% of the Cu-binding sites on the gill were occupied
by Cu at the LC50. The remaining sites were occupied by Ca
(32.1%) or H (0.7%) or were unoccupied (30.7%).
Then I calculated the total aqueous concentration of Cu
(Cunoi) needed to achieve the same amount of Cu binding
at other hardnesses and alkalinJties but still at pH 7.0. If the
amount of Cu bound to fish gills is constant at 50% mortality,
this Cum* is the predicted LC50.1 used two scenarios for
these simulations: (a) with alkalinity equal to hardness and
(b) with alkalinity held constant at 50 mg-L"1 as CaCOj.
However, because the stability constants I used for gffl binding
are conditional (Le., not true thermodynamic constants
because they win tend to vary as a function of pH. hardness,
etc.). my calculations of percentage occupancy of gill binding
sites and. hence, LCSOs of Cuom are only approximations.
As shown theoretically in the previous section, the
predicted LC50 of Cu*** increased as hardness increased
(Figure 1). The slope of bi(LC50) vs Inftiardness) approached
zero at low hardness for both scenarios (alkalinity equal to
hardness or alkalinity constant); whereas the slope ap-
proached 1.0 at high hardness when alkalinity was held
constant, and the slope approached 2.0 at high hardness
when alkalinity equaled hardness. When pH was varied, the
predicted LC50 at a specified hardness decreased consider-
ably aspH decreased £rom9to6(Figure 2). However.between
pH 5 and 6 the predicted LC50 decreased only slightly at
high hardness, and it increased at low hardness.
Disnssioi
Based on the assumptions that (a) the amount of a divalent
transition-metal cation bound to a fish gill at 50% mortality
is constant (Le., independent of water quality) and (b) protons
and hardness cations compete with the transition-metal
cation for binding at these sites, it is not surprising that all
of the slopes of the ln(LC50) vs biQiardness) regressions for
Cd, Cu. Ni, Pb, and Zn reported in the USEPA water quality
criteria documents (3-7) are -1. Although the slopes
theoretically could range from 0 to 2 (at extremely low and
high hardness, respectively; Figures 1 and 2), the following
rules-of-thumb apply:
1. If hardness and alkalinity are extremely low (e.g., <~1
mg-L"1 as CaCOj), metal complexation with HCCv
10000 i
1000-
i
I
100 1
10
0.1
'A
10
100
1000 10000
Hardness (mg-L*1 as CaCOg)
FIGURE 3. Data ned la (4) to generate the slope of OJ4 for the
regression of ln(LC50) on In(hardaess) for Ca. The solid carve
represents the simulated LCSOs calculated in this paper; the dashed
line is a slope of 1. Species abbreviations are DM - Dipbaiatnigaa,
DP = Diptiaia policaria, CS = Chinook salmon. CTT = cutthroat
trout RBT = rainbow trout FHM = fathead minnow, and BG =
bluegilL
is negligible and competition by Ca2* for M-binding sites is
minimal because of the large percentage of unoccupied
binding sites on the gill. Thus, an incremental increase in
hardness and alkalinity has almost no effect on metal binding
to the fish gill, and the slope of ln(LC50) vs InOiardness) is
2. But if hardness and alkalinity are extremely high (e.g.,
> 1000 mg-L'1 as CaCOj, a large percentage of the metal in
solution is complexed with HOV or C
-------
pHs and alkalinities. and (c) the range of hardnesses is too
narrow to clearly demonstrate the curvilinear relationship.
Although hardness and alkalinity are approximately equal
in many surface waters (including those used in many unddty
tests), some transition-metal toxidty tests are conducted at
"constant alkallnit$while hardness is varied (or vice versa)
to demonstrate the effect of hardness (or alkalinity) on metal
toxitity. In the midrange of hardnesses, the slope of the In-
(LC50) vs in(hardness) regression for divalent transition
metals should be ~0.5 when alkalinity is maintained constant
Supporting this prediction, I found a slope of 0.50 (r2 = 0.85,
P = 0.0002) when I regressed ln(LC50) on In(hardness) for
the 10 data points in Figure 3 of Erickson et al. (11) (96-h
LCSOs for FHM larvae exposed to Cu; hardness ranged from
37 to 134 mg-L-1 as CaCOj).
A similar effect of alkalinity on transition-metal toxidty
should occur as hardness is held constant Thus, it is not
surprising that Erickson et al. (//) (their Figure 1) only
observed ~30-40% increases in LC50 of Cumd when they
increased alkalinity from -45 to ~150 mg-L"' as CaCCb at
pH between 7 and 9 and a hardness of ~45 mg-L'1 as CaCOj
(J5). Such small effects of alkalinity on LC50 would be
predicted if the slope of the ln(LC50) vs In(alkalinity)
regression in this alkalinity range was ~0 J—a realistic value
because hardness was held constant At higher alkalinities,
Erickson et al. (11) in theory would have seen a greater
percentage increase in LC50 for the same percentage change
in alkalinity, whereas at low alkalinities, they would have
seen almost no effect of alkalinity on the LC50. Knowledge
of alkalinity and hardness is cmdal to accurately predict the
LCSOs of transition metals using bi(LC50) vs Infhardness)
regressions.
Erickson et aL (11) (their Figure 1) also showed that LC50
of Cuncrf increased as pH Increased between 6 and 9. This
agrees with the gill-binding theory (Figure 2), because
competitive binding of protons to the gOl at pH 2:6 is minimal;
thus, the only effect of increasing pH is to increase the COj*~
concentration (at constant alkalinity) and the amount of Cu
complexed with that ligand. However, at pH <6 proton
binding to the gill can be significant because (H*] becomes
high enough (> 10~« M) for protons to compete with Ca*+
and Cu*"* (i.e., [H*]-KHmjKU approaches [Ca2*|-Ao^Bc»and
ICu^J-Ko-fiicu). Therefore, at low alkalinity and hardness,
the LC50 of Cutocd at pH 5 theoretically should be higher than
at pH 8:6 (Figure 2), but at high alkalinity and hardness, the
reduction in complexation of Cu by HCQr and C0j*~ caused
by the tower pH should help offset the increase in the LC50
of Gin* that otherwise might occur. This prediction does
not take into account the potential onset of add toxidty at
pH <6 that will be caused by accumulation of protons on the
gill, a process that would tend to decrease the Cuum
concentration at which 50% mortality occurs when the
toxicants act jointly.
I have presented examples for Cu because of the extensive
LC50 and gill-binding data available. However, similar
hardness-related trends should occur with Cd, Ni, Pb, and
Zn. In general, I predict the slopes of ln(LC50) vs m(hardness)
for these four transition metals will tend to be lower than
they are for Cu in the same hardness range (if alkalinity
covaries with hardness), because Cd, Ni, Pb, and Zn have
lower affinities for COs2* U3). But if alkalinity Is held constant.
I predict the slopes for all five transition metals will be
approximately the same in the same hardness range. The
pH-related trends for toxidty of Cd. Ni, Pb, and Zn to FHM
are less pronounced (Pb) or are the opposite (Cd, Ni, and Zn)
of the trend for Cu (16), in part because the chemical
spedation of Cd, Ni, Pb, and Zn is much less affected by
changes in pH than is the spedation of Cu (1).
In conclusion, the LC50 of a transition metal at a specified
hardness can be predicted with relatively minor error in the
hardness range 20-200 mg-L'1 as CaCOj by knowing the
LC50 at a different hardness and assuming (a) a slope of -1
for the relationship between In(LCSO) and In(hardness) and
(b) alkalinity covaries with hardness. Such predictions might
be acceptable for regulating discharges of transition metals
to surface waters. But extrapolations to hardnesses outside
this range might drastically overpredict the toxidty of the
metal (e.g., extrapolating from an LC50 determined at a
hardness of -30 mg-L'1 as CaCQ) to point A or B in Figure
3). Conversely, starting with an LC50 determined at a relatively
low or high hardness, extrapolations to midrange hardnesses
might drastically underpredia the toxidty of the metal. Such
errors might be magnified 'if the alkalinity does not change
proportional to the hardness in these extrapolations. Rather
than relying on amechanistic bi(LC50) vs In(hardness)
regression equations, a better approach might be to calculate
uptake of transition metals using a biotic-ligand model and
then predict toxidty from an empirical relationship between
mortality and the amount of accumulated metal.
Acknowledgments
This research was funded by a subcontract to the University
of Wyoming under a grant from the International Copper
Association to HydroQual, Inc., a subcontract to the Uni-
versity of Wyoming under a cooperative agreement between
the U.S. Environmental Protection Agency and the University
of Delaware, and a National Science Foundation EPSCoR
grant to the University of Wyoming. EOen Axtmann assisted
with the geochemical spedation calculations. Comments
from Williams Adams, Herbert Allen, Paul Paquin, and two
anonymous reviewers improved the manuscript
Appenfu
Mass Balance on Metal In Solution. To construct a tractable
mass balance for a divalent transition-metal cation to which
fish are exposed in a toxidty test, I make the following three
assumptions:
1. No partides or organic ligands are present.
2. Complexes of divalent transition-metal cations with
more than one HCQr. COj2". or OH~ (e.g., CuftttXhh0.
CutCOjh2-, Cu(OH) A or Cu(OH)r) are negligible. Although
this is not correct at high pH and alkalinity, the complexes
can be ignored for heuristic purposes.
3. At acutely lethal concentrations under realistic biomass
loadings (e.g., <10 g fish*L~'), the amount of the transition
metal that complexes with the fish gills will be a negligible
percentage of the total amount of dissolved metal remaining
in the water. For example, MacRae et aL (1 7) estimated 0.03
/imol Cu-binding sltes-g-' wet weight of rainbow trout
(Oncorhynchus myktss) gUL If the gQl constitutes <5% of the
wet weight of a rainbow trout, then the < 10 g fish-L"1 will
be able to bind < 1 pg Cu-L"1. Because not all of the dissolved
Cu will be present as Cu*" (the spedes that presumably is
in equilibrium with binding sites on the gill) except at pH s6
and because a large excess of Cu would be needed to saturate
the Cu-binding sites on a fish gill flog KCH.&CU = 7.25-7.5
(9, 1 7)). the amount of Cu bound to fish gills at acutely lethal
concentrations (LCSOs usually » 1 /ig-L~') will be a relatively
small percentage of the total Cu in the exposure water.
Thus, an approximate mass balance on the total amount
of the transition metal in the exposure system is
|M2*1 + [MOHl
(A-l)
Ignoring activity coefficients, the governing equilibrium
equations (with stability constants for the metal-ligand
complexes listed in (13)) are
D • ENVIRON SCI. & TECHNOL. / VOL xx. NO n. uaa
-------
10
-u
(HC03"1
[H+HC032-J "
IMOrf]
10
,10.33
|M2+HOH-]
1MHC03*1
IM2>(HC03-J
[MC03°1
[M21-[C032-]
Additionally, assuming the aqueous system is buffered
only by the hydroxyi and carbonate/bicarbonate systems,
the alkalinity (Alk, expressed as eq-Ir1) is (18):
Alk=P + IHCOj'] +
Substituting these relationships into eq A-l produces
By combining eqs 5 (which expresses the LC50 of M2* as a
function of [Ca**l and [CandllM)) and A-2, the LC50 of die
— cowl dlssotvea metal UA3JW.J) can °e expressed as a
function of (a) the competitrvTouiding of cations at the gfll
surface and (b) the equilibrium of the free-metal ion with
the aqueous inorganic figands, as follows:
^ -- x.
Equation A-3 can be simplified for heuristic purposes to the
following approximation
[C**
lf,nn
[CaEgillMJ
(l + -^5
V IQ-P"
(A-4)
where Jfand f" are constants.
Mass Balance on GDI Binding Sites. Assuming H*. Cal\
and M2* are the only cations of importance that bind to the
M-binding sites on a fish gill (although Mg2* could be
substituted for or act in combination with Ca2+), an ap-
proximate mass balance for the M-binding sites on the gill
is
IMeglllM) + [CasgillM] + [HsgfllM] +
NgiUM]
or
[CasgillM) = (sgiUMttttll - ((MsgillMJ + {HsgillM] +
(sgillM]) (A-5)
literature Cited
(1) Borgmann, U. In Aquatic Toxicology, Nriagu, I. O., Ed; John
Wiley and Sons: New York. 1983; pp 47-72.
(2) Campbell, P. a C In Metal SpedaOon and BloavailabWty in
Aquatic S/stemr.Tessiet, A., Turner, D. R., Eds.; John Wfley and
Sons: Chichester, 1995; pp 45-102.
(3) Ambient Water Quality Criteria for Cadmium - 1984. U.S.
Environmental Protection Agency: Washington, DC 1985; EPA
440/5-84-032.
(4) Ambient Water Quality Criteria for Copper - 1984. U.S.
Environmental Protection Agency: Washington, DC, 1985; EPA
440/5-84-031.
(5) Ambient Water Quality Criteria far Nickel - 1986. U.S. Envi-
ronmental Protection Agency: Washington. DC 1986; EPA 440/
5-86-004.
(6) Ambient Water Quality Criteria far Lead - 1984. US. Environ-
mental Protection Agency. Washington. DC 1985; EPA 440/
5-84-027.
(7) Ambient Water Quality Criteria for Zinc - 1987. VS. Environ-
mental Protection Agency: Washington, DC 1987; EPA 440/
5-87-003.
(8) Pagenkopt G. K. Environ. Set TechnoL 1983, 17, 342-347.
(9) Piayie,R.C;Dixon,D.G.;BiimlsoaK.Ciin./.Rtft.A7uatSci
1993, 50. 2678-2687.
(10) Meyer, ). S*- Santore, R. C: Bobbin, I. P.; DeBrey, L. D^ Boese.
C J.; Paquin. P. H.; Allen, H. E.: Beipnan. H. L; DfToro. D. M.
Environ. Set TechnoL ~~
(1 1) Erickson, R. I.; Benoit, D. A; Mattson, V. R.; Nelson, R P., Jr^
Leonard, £ N. Environ. TaxicoL Chem. 1996, IS, 181-193.
(12) Allison, I. 04 Brown. D. S.; Novo-Gradac. K. I. MDTTBQA2I
PRODEFA2.A GcochemicalAssetsment Model for Environmental
Systems: Venlon 3.0 Uteri Manual. U.S. Environmental Pro-
tection Agency: Washington. DC 1991; EPA/600/3-91/021.
(13) NET Critically Selected Stability Constant! of Metal Complexet
Database. Version 5.0. NET Standard Reference Database 46;
NationallnstmiteofStandanbandTechnology: Gaithersburg,
MD, 1998.
(14) Serkiz, S. M.; Allison. I D.; Perdue, E M; Alien. R E: Brown,
D. S. Water Ret. 1996, 30, 1930-1933.
(15) EricksoaR-J.; Benoit, D. A; Mattson, V.R. A Prototype rociciO1
Factors Model far Site-specific Copper Water Quality Criteria;
OS. Environmental Protection Agency. Duluth, MN, 1987
(revised 1996).
(16) Schubauer-Berigan,M.IC;Dierkes,I.R.;Monson,P.D.;Ankley,
G. T. Environ. ToxicoL Chem. 1993, 12, 1261-1266.
(17) MacRae, R. K.; Smith, D. E; Swoboda-Colberg, N.; Meyer, I. S.;
Bergman, RL Environ, roiicol. Chem. Accepted for publication.
(18) PanicaiM.l.f. Aquatic Chemistry Conceptx Lewis: Chelsea, ML
' 1991; p 171.
Received for review July IS, 1998. Revised manuscript re-
ceived November 30,1998. Accepted December 15, 1998.
ES980714Y
PAGE EST. 44 VOL. xx. NO «. moot I ENVIRON. SCI & TECHNOL • E
-------
Gill Surface Interaction Model for Trace-Metal Toxicity to Fishes: Role of
Complexation, pH, and Water Hardness
Gordon K. Pagenkopf
Department of Chemistry, Montana State University, Bozeman, Montana 59717
• A model has been developed to account for the varia-
bility in trace-metal toxicity to fishes at different values
of alkalinity, hardness, and pH. The model utilizes
trace-metal speciation, gill surface interaction, and com-
petitive inhibition to predict effective toxicant concen-
tration (ETC). Copper, cadmium, lead, and zinc bioassay
data have been utilized.
A review of the many research projects that have in-
vestigated trace-metal toxicity to fishes provides at least
three general conclusions: (1) for a particular trace metal,
some chemical species appear to be more toxic than others;
(2) the presence of elevated concentrations of the hardness
cations ions, Ca2+ and Mg2"1", reduces trace-metal toxicity,
(3) LC50 concentrations vary from metal to metal. These
are not the only generalities of course, but they do provide
a basis for the development of a model that can account
for changes in toxicity as a function of pH, complexation
capacity, and hardness of the test waters. Currently there
is disagreement regarding the relative importance of water
hardness and trace-metal complexation (1-4). This pape>
presents a model that combines both factors. What follow.-
is an identification of the chemical principles that arv
believed necessary to couple trace-metal toxicity to pH,
hardness, and trace-metal complexation. The identified
principles are utilized to formulate quantitative relation
ships, and finally, predicted variation in trace-metal tox-
icity is compared to that observed in laboratory testa,
Basis for Gill Surface Interaction Model (GSIM)
The following are set forth as basic to the development
of GSIM:
(1) For acute toxicity to fish, trace metals alter the gill
function, and the fish die as a result of respiratory im-'
pairment.
(2) Of the trace-metal species present in a test water,
some are significantly more toxic than others.
(3) The gill surfaces are capable of forming complexes
with the metal species and hydrogen ion present in the test
waters.
342 Environ. Sci Technol., Vol. 17, No. 6, 1983
0013-936X/83/0917-0342S01 50/0 © 1983 American Chemical Society
-------
(4) The rates of metal exchange between the gill surfaces
and test waters are fast when compared to the time re-
quired for a bioassay test.
(5) The gill surfaces have a finite interaction capacity
per unit weight.
(6) Competitive inhibition exists between the hardness
metals and the toxicants, which include the trace metals
and hydrogen ion.
A variety of experimental observations, both chemical
and biological, will be utilized to substantiate the model
Model Development
Excessive mucous secretions are often observed when
fish are stressed by elevated concentrations of trace metals
(5-7) and hydrogen ion (8). In addition trace metals may
be concentrated in the gill tissues (9,10) with the mech-
anism of toxicity apparently being related to salt and water
balances within the gill tissues (11,12). The physiology
associated with the toxic action of trace metals is extremely
complicated, and it is not the intent of this paper to discuss
or even speculate as to the mode of action. This model
utilizes competitive equilibria to predict changes in the
chemical activity of metal species associated with gill
surfaces. This associated ion activity correspondingly in-
fluences the physiological function of the gills.
Gill membranes consisting of phospholipids could pro-
vide a surface of a net negative charge and the sites nec-
essary for the formation of Lewis acid-base complexes with
the metal ions and hydrogen ion. The interaction may be
classified as surface complexation (adsorption) or ab-
sorption as long as exchange is rapid and it is reversible.
A schematic representation is shown in eq 1 with copper
Cu2* + ^=S"~ *=* =SCu~"'1'2 (1)
as the Lewis acid. The surface is represented by ssS"",
which designates a group of Lewis base sites that collec-
tively is capable of forming surface complexes with the
metal species. This approach permits application of
chemical principles that have been successfully utilized in
the interpretation of more precisely defined and controlled
chemical systems. The surface complex is designated by
EEESCu"*"1"2, and the equilibrium constant for the interaction
is KCV Utilizing the condition that complexation reactions
of this type are rapid, an equilibrium expression may be
written
KCU - {s=SCu~fl+2)/[Cu2+]{3sS't-} (2)
where braces and brackets designate concentration of
moles per kilogram and moles per liter, respectively.
Rearrangement of eq 2 provides
which identifies a linear relationship between the con-
centration of the surface complex and the concentration
of Cu2+ in solution, provided the concentration of sS"~
remains constant A relationship of this type is in agree-
ment with experimental observation where an increase in
total toxicant concentrations decreases survival time (13).
Implicit is the fact that a small fraction of the complexa-
tion sites is occupied by Cu2+, and thus |s=S"~| remains
essentially constant within the experimental uncertainties
of the bioassay test.
Trace-metal speciation in natural water systems is de-
pendent upon pH and which complexing Uganda are
present. Most bioassay test waters contain minimal
amounts of organic material, and thus these complexes will
not be considered. A majority of the complexes involve
hydroxide and inorganic carbon. Each of these species
constitutes some finite fraction of the total trace metal m
solution. For Cu2+ this fraction, designated by aCu», is
defined as
<*<:„«* = [Cu2+]/[CuT] (4)
where CuT is total copper in solution. Procedures for the
evaluation of a values have been presented elsewhere (14).
Combination of eq 3 and 4 provides
which establishes the concentration of the surface complex
in terms of speciation.
Bioassay tests indicate that other species may be toxic
and need to be included (2). For copper, these include
CuOH*. Cu(OH)2.aq, and possibily Cu2OH22+. All are
capable of forming surface complexes with the strength
of the interaction being species dependent Generalization
of eq 5 provides
KCa,^S'1«c«f[CuT]
(6)
where ^SCu.) represents the concentrations of the surface
complex for the ith copper species. Similarly KQU, and OQ.
are the surface complexation equilibrium constants and
the fractions of total copper present as the tth species. An
equation similar to eq 6 is applicable for the other trace
metals.
There are reports (1, 3, 4) indicating that an increase
in water hardness increases fish resistance to trace metals.
Test water concentrations for the common hardness
metals, calcium and magnesium, range from approximately
10 to 1000 mg/L as CaCO3. The corresponding concen-
trations would be lO-MO"2 M. For these metals to exhibit
a pronounced protective effect, their concentration gen-
erally has to be greater than 10~3 M. These metals form
few complexes with the ligands found in most test waters,
and therefore the equated metal concentration is essen-
tially equal to the total metal concentration.
The GSIM model interprets the protective action of the
hardness cations as a competition between these metal ions
and the toxic species, in essence competitive inhibition.
The total number of surface interaction sites is given by
eq 7, where s=S*- are the free sites, ^S(M)-"*2 are those
I"' + SES(M)"+2 + ssS(H)-"+1 + s=S(TM) (7)
occupied by the hardness cations, sS(H)~"+1 are those
occupied by hydrogen ion, and =S(TM) are those occupied
by the trace-metal species. Charges for sN3(TM) are
omitted, and since Ca2+ and Mg2+ exhibit similar chem-
istries, they are not differentiated.
For test waters of pH 6 or greater there appears to be
little hydrogen ion dependence, and with concentrations
of hardness cations many orders of magnitude greater than
the concentrations of the toxic trace-metal species, it is
assumed that (=S"- + ==S(M)-"+2) » (^(H)""*1 + =S-
(TM)). The equilibrium expression for the hardness metal
interaction is
Substitution into the simplified form of eq 7 with rear-
rangement provides
GIF
(9)
which is designated as the competitive interaction factor,
GIF. Substitution into eq 6 provides a relationship suitable
for the interpretation of the copper toxicity data:
j = Kc,laCu,[CuT]|sST|/(l + KM[M2+]) (10)
Environ Set Technol , Vol 17, No 6 1983 343
-------
Table I. Species Concentrations, Hardness, and ETC
Values for Rainbow Tiout (/ )
Mg/L
test
no.
1
2
3
4
5
6
7
8
9
10
11
pH
6.0
6.0
6.0
7.0
7.0
8.0
8.0
8.0
9.0
9.0
9.0
Cu,«
Mg/L
22.2
39.5
82.2
32.5
137
16.2
138.6
83.1
14.8
35.2
16.0
hard-
ness,
mg/L as
CaCOj
32
101
371
101
298
31
371
360
30
98
364
CIF*
0.44
0.17
0.08
0.17
0.06
0.44
0.05
0.05
0.44
0.17
0.05
ETC,C
9.77
6.72
6.58
5.50
8.22
7.13
6.93
4.16
6.51
5.98
0.8
av 6.21 i 2.29
" Cu= [Cu2*] + [CuOH*] = [Cu(OH),-aq]. * CIF= I/
(1 + KMIM*']),KU = 5X 10s. cETC=CIF([Cus*] +
[CuOH*] + [Cu(OHyaql).
Defining |^SCu,|/((KCaj){^ST|) as the effective toxicant
concentration, ETC, eq 10 becomes
ETC =
CIF[toxic species]
~ (11)
Application of GSIM
Copper. Chemical speciation has successfully accounted
for variation in toxicity of some metals (2, 4,15-17). In
these studies the total metal concentrations is often much
greater than the concentration of the most toxic species.
Equation 11 is capable of accounting for the speciation;
however, a value for KM has to be assigned before quan-
titative interpretation of the hardness can be presented.
An empirical estimate of 5 X 103 comes from observed
copper toxicity. The concentration of the hardness metals
has to be greater than 10~3 M before a pronounced effect
is observed.
The data presented by Miller and Mackay (3) provides
a way to estimate this constant They determined copper
toxicity as a function of hardness at constant pH and
complexation capacity. As a consequence CuT may be
compared directly. From these comparisons KM is calcu-
lated to be (7 ± 5) x 103.
Equation 11 indicates that ETC should have a constant
value for a given test species and constant time of exposure.
For copper
ETCcu = CIF([Cu**l + [CuOH+] + [Cu(OH)raq]) (12)
Data presented by Howarth and Sprague (1) include a
change in species distribution as well as hardness (see
Table I). The mean ETCo, value for rainbow trout is 6.21
Mg/L with a standard deviation of 2.29.
An extensive study utilizing cuttthroat trout provides
another system for the application of GSIM (4). Three
hardness and three alkalinity concentrations were em-
ployed, resulting in nine combinations. The toxic species
are considered to be Cu2+, CuOH+, and CufOH^aq, and
the fraction of the interaction surface available is regulated
by the hardness metal ion concentrations. The values of
ETCcu are listed in Table IL For comparison, the average
96-h ETCCu for rainbow trout is 6.21 Mg/L, whereas the
comparable value for cutthroat is 2.72 /ig/L.
Copper species distribution and ETC values have been
calculated for the Miller and Mackay data (3), and the
Table II. Predicted ETC Values for Cutthroat Trout (4y
hardness/
alkalinity,
mg/L as Cu,6
ETC,
1.65
2.44
3.04
2.02
5.54
2.98
0.81
3.75
2.24
av 2.72 * 1.33
° Data from ref 4, 96-h LC50. b Cu = Cu2* + CuOH* +
Cu(OH),-aq, species distribution calculated by using for
CuOH* log K, = 6.48 and for Cu(OH). log p. = 11.78 (18),
96-hLC50.
CaCO,
205/178
205/77.9
160/26.0
70/174
70/70
74.3/22.7
18/183
18/78.3
26.4/20.1
pH
7.73
761
7.53
8.54
740
7.57
8.07
8.32
7.64
GIF
0.089
0.089
0.11
0.22
0.22
0.21
0.53
.0.53
0.43
Mg/L
18.5
27.4
27.4
; 9.2
25.2
14.2
1.53
7.09
5.21
Table HI. Predicted ETC Values for Rainbow Trout and
Fathead Minnows
hardness/
alkalinity,
mg/L as
CaCO,
PH
CIF
Cu,
Mg/L
ETC, Mg/L
ref
rainbow trout
12/10"
99/10°
49/28°
98/28"
12/51°
97/51°
300/205*
198/161*
31.4/15*
360/150°
20/9°
7.1
7.0
7.3
7.2
7.4
7.3
7.35
7.9
7.2
8.2
7.5
0.62
0.17
0.36
0.17
0.62
0.17
0.0625
fathead
0.092
0.39
0.053
0.50
12.7
38.0
17.0
30.5
3.6
21.7
av
62.2
minnow
45.9
39.7
70.5
13.2
av
7.87
6.52
6.10
5.18
2.22
3.69
5.26 i 1.89
3.88
4.22
15.5
3.7
6.6
7.53 ± 5.48
3
3
3
3
3
3
19
20
21
22
22
" 15-day LC50 values. 6 96-h LC50 values.
results are summarized in Table m. Values for the 15-day
test are somewhat less than those observed for the 96-h
test with rainblows, 5.28 vs. 6.21 Mg/L- The order of re-
sistance of the test fish to copper appears to be fathead
minnows > rainbow trout > cutthroat trout.
The data presented in Tables I-QI include a variation
in pH from 6 to 9, an alkalinity variation from 10 to 205
mg/L as CaC03, a hardness variation from 12 to 371 mg/L
as CaCO3, and a variation hi total copper by more than
a factor of 100. Application of the GSIM to these data has
identified an effective toxicant concentration for each test
animal. The variability in the predicted ETC values is
equal to or less than the observed experimental variability.
This model is based on the premise that trace-metal
species bound to the gill surfaces cause impairment of
physiological functions. The amount of trace metal bound
is regulated by the chemical composition of the test waters.
Specifically a competition exists between the hardness
metals and the toxic species for interaction sites.
Zinc. The coordination chemistry of zinc is similar to
copper in many respects; however, the thermodynamic
stability constants are generally not as large. Zinc spec-
iation for a number of bioassay studies has been completed,
and a correlation exists between the sum of the concen-
344 Environ Sci. Technol, Vol 17, No 6, 1983
-------
Table IV. Zinc Toxicity to Brook Trout and
Rainbow Trout (23)
fish
size
mean
wgt, g
hardness/
alkalinity,
mg/L as
CaCO,
PH
ZnT,a
mg/L
Zn,"
mg/L
GIF
ETC,
mg/L
brook trout
30
3.0
3.9
3.9
19.0
19.0
46.8/41 8
177.6/170.2
47.0/42.8
179.0/170.1
44.4/42.5
169.7/43.0
7.63
7.41
7.58
7.17
7.38
7.31
1.55
6.14
2.12
6.98
2.42
4.98,
1.42
4.83
1.94
5.91
2.27
4.71
030
0.10
0.30
0.10
0.31
0.11
0.43
0.48
0.58
0.59
0.70
0.44
rainbow trout
3.9
3.9
4.9
4.9
28.4
28.4
46.8/41.8
177.6/170.2
47.0/42.8
179.0/170.1
44.4/42.5
169.7/43.0
7.63
7.41
7.58
7.17
7.38
7.31
0.370
2.51
0.517
2.96
0.756
1.91
0.339
1.96
0.475
2.56
0.712
1.81
0.30
0.10
0.30
0.10
0.31
0.11
0.10
0.20
0.14
0.26
0.22
0.19
0 96-h LC50 values. 6 Zn = Zn1* + ZnOH*
Table V.
hard-
ness,
mg/L as
CaCO,
20
80
320
320
Cadmium Toxicity to Rainbow Trout (25 f
Cdj,
mg/L
0.091
0.358
3.69
0.677
Cd,6
mg/L
0.083
0.262
1.66
0.618
GIF
0.50
0.20
0.058
0.058
ETC, mg/L
0.042
0.052
0.096
0.036
av 0.056 ±
0.027
0 pH 7.2, 48-h LC50.
Cd(OH),-aq.
6 Cd = Cd" -i- CdOH* +
trations of Zn2* and ZnOH* and the observed toxicity (1 7).
Utilization of the procedures outlined for copper provide
CtFdZn2*] + [ZnOH*]) (13)
A recent report by Holcombe and Andrew (23) presents
additional zinc toxicity data for brook trout and rainbow
trout. This study was designed to test the influence of
hardness. The conditions were well regulated, and the
results are presented in Table IV.
Lead. A chronic bioassay study by Davies et ai (24) has
established that the maximum acceptable toxicant con-
centration (MATC) for rainbow trout in hard water should
range from 18.2 to 21.7 Mg/L dissolved lead. In soft water
the MATC ranges from 4.1 to 7.6 Mg/L- Species distri-
butions for the test waters indicate that soluble lead was
dominated by Pb2* and PbOH* . The hardness of the test
waters was 353 and 28 mg/L as CaCO3, respectively.
Calculated CIF values are 0.054 and 0.42, which indicates
that soluble lead should be 7.8 times more effective in soft
water. The MATC ratio is 4.3 with means and ranges from
2.4 to 7.7 by use of the extreme values. A majority of the
lead studies are of static design and there is sizable dif-
ference between total lead and soluble lead. -
Cadmium. A study designed to test the.influence of
hardness on cadmium toxicity to rainbow trout (25) has
been reported. There is major variation in hardness,
20-320 mg/L as CaCO3, and the alkalinity of the waters
was not reported. Test water conditions were reported
earlier (19), and these values were utilized to calculate the
speciation. The results are summarized in Table V.
Given the range of hardness 20-320 mg/L and the range
of total cadmium 0.091-3.69 mg/L (a factor of 40), the
agreement is considered to be very good.
Combination of Metals. The interpretation of metal
toxicity resulting from the mixture of metals is of sizable
current interest and will probably become more critical as
time goes on. The terminology put forth by Sprague (26,
27) is useful for discussion purposes. When the toxicity
of a mixture corresponds to the sum of the fractions of the
single components, the effect is referred to as "additive".
When the effect is greater than or less than, a "more-
than-additive" or "less-than-additive" effect is assigned.
The GSIM assumes an additive effect This results from
the linear or near linear relationship between concentration
of the surface species and the corresponding concentration
of these species in solution. To apply the model, the toxic
unit concept (27) needs to be utilized. This may be ac-
complished through ratios of the ETC values for the re-
spective metals.
A paper by Eaton (28) reports 72- and 96-h LC50 data
for mixtures of Cu, Cd, and Zn. The data are summarized
in Table VL In a parallel study (28) it was observed that
5030 Mg/L total zinc was required for the 96-h LC50. The
pH, alkalinity, and hardness are the same, and thus 0.41
toxic units are assigned to zinc. This value is equal to the
ratio of the zinc required in the mixed-metal study to that
required when zinc was the only toxicant (2050/5030).
Another study from the same laboratory (20) reported a
96-h LC50 total copper value of 430 Mg/L. From these
values 0.36 toxic units are assigned to copper (154/430).
The contribution from cadmium is difficult to assign since
the acute study for cadmium and fathead minnows re-
ported the presence of sizable amounts of insoluble cad-
mium salts (29). Using the lowest total dissolved cadmium
concentration, 1400 Mg/L, a hardness of 201 mg/L as
CaCO,, a pH of 7.7, and an alkalinity of 161 mg/L as
CaC03, 118 Mg/L is predicted for the Cd ETC value.
Coupled with the value in Table VI, 0.22 toxic units are
assigned to cadmium (26.3/118). Summation of the re-
spective contributions provides a value near unity and may
be significant. There is sizable uncertainly in the cad-
mium contribution, however.
Other trace-metal mixture studies (30, 31) indicate an
additive response. There are insufficient chemical data
available to do chemical speciation, and thus ETC cannot
be calculated. In the cases where pH, alkalinity, and
hardness remain fairly constant, total concetrations may
be utilized as a measure of fish response.
Table VI. Toxicity of Cu, Cd, and Zn Mixture to Fathead Minnows" (28)
" 96-h
Cu(OH),
metal concn,
metal Mg/L (total)
Cu 154
Cd .320
Zn 2050
results, alkalinity = 154 mg/L
aq.Zn = Zn2* + ZnOH'.Cd
TM,& Mg/L
9.52
299
1517
as CaCO,, hardness =
= Cd1* + CdOH*
CIF
0.088
0.088
0.088
207 mg/L as CaCO,,
ETC. Mg/L
0.84
26.3
133
pH 7 7. 0
TU
0 36
-0.22
0.41
sum ~0.99
Cu = Cu2' + CuOH* +
Environ Sci Technol, Vol 17, No 6, 1983 345
-------
A sequence of papers (32-34) has reported the toxicity
of mixtures to guppies. In these experiments copper and
nickel appear additive, whereas copper and zinc appear
more than additive. For the zinc studies (32), a 96-h LC50
value of 6.76 mg/L is reported. This value exceeds the
calculated zinc solubility of 3.2 mg/L. A lower concen-
tration of zinc would correspondingly decrease the degree
of apparent enhanced toxicity due to the mixture of copper
and zinc.
Hydrogen Ion. It is difficult to precisely define the
tolerance of fish toward elevated hydrogen ion concen-
trations since it appears to be dependent upon many
variables. One report indicates that pH values near 5 may
represent a lower tolerance limit, however (35). In addi-
tion, alteration of gill surface tissues is observed at pH
values below 5.2 (36). The concentration ranges where the
respective metals exert an influence, [Ca] » 10~3 M; [Cu]
a* lO-MO-7 M; [Zn] <* 10* M; [Pb] « KT7 M; and [Cd]
tst 10~* M, provides an indication of the stability of the
metal-surface interaction. The stability constants are, in
general, the inverse of concentration where the median
effect is observed. The values are in the range of the
stability constants observed when these metals are com-
plexed by multidentate carboxylate or phosphate ligands.
Acid dissociation constants for carboxylic acids are often
near 10*6, and thus the observation that hydrogen ion
becomes toxic in the pH 5 region is not unexpected.
Weak acid species distribution changes rapidly as the
pH varies in the range of the acid pKt. At the lower pH
values the concentration ^^(H)"""1"1 in eq 7 is not small
when compared to the other components and thus can
contribute to the observed toxicity. Hydrogen ion will
compete for the interaction sites, and as a consequence
trace metals should not be as toxic in the more acidic
regions. Correspondingly, fish should be able to tolerate
more acid in hard water.
Summary
A model that includes chemical speciation and gill
surface interaction has been developed to account for the
variation of trace-metal toxicity to fishes under varying
conditions of pH, hardness, and test water complexing
capacity. The net result is a series of equations that relates
water chemistry to observed fish toxicity. Trace-metal
toxicity is considered to be additive. The protective action
of the hardness metals is due to their sucessful competition
with the trace-metal species for gill surface interaction
sites. These interactions are more extensive for the di-
valent metal ions than for the other common cations, Na+
and K*, which do not exhibit the same protective action.
The concept of effective toxicant concentration has been
formulated. For a given metal and test fish, ETC exhibits
a variation comparable to experimental reproducibility
even though hardness, pH, alkalinity, and total metal
concentration vary by several orders of magnitude. The
idea of competition between the metal ions has been
suggested previously (37) but not formulated this way.
Applicability of GSIM. GSIM is easily applied to
water suspected to contain sufficient quantities of trace
metals to cause acute toxicity. The data required, pH,
alkalinity, hardness, and total trace-metal content, are
usually available. The first step involves calculating the
species distribution. Second is an evaluation of GIF and
subsequently ETC. The latter is then compared to labo-
ratory observation. For natural waters that contain ad-
ditional complexing agents such as humic and fulvic acids,
a reduction in toxicity is to be expected since the trace-
metal complexes with these agents, appear to be nontoxic
(38). The trace-metals species adsorbed by suspended
particulate matter are also predicted to be nontoxic.
Acknowledgments
Appreciation is extended to G. R. Phillips and R V.
Thurston for resource" materials and to the reviewers for
many helpful comments.
Appendix
The following is a listing of the stability constants used
to calculate the species distributions. Calculations utilized
COMICS (39) and UMDEQ (40). HCO3', log K « 10.30;
CO2-aq, log K = 16.76; CuC03, log K = 6.75; Cu(CQJJ-,
log ft = 9.92; CuHOy, log K = 2.0; CuOH*. log K - 6.48;
Cu(OH)2-aq, log ft = 11.78; Cu2(OH)22+, log K - 17.7;
CaCOj, log K = 3.15; CaHC03+, log K = 1.0; CaOH+, log
K - 1.3; ZnCOj, log 1C - 5.0; ZnOH+, log K - 6.31; Zn-
(OH)raq, log ft - 11.19; CdC03, log K = 4.0; CdOH+, log
K - 4.61; Cd(OH)2-aq, log /J2 = 8.92.
Registry No. Cu, 7440-50-8; Cd, 7440-43-9; Rb, 7439-92-1; Zn,
7440-66-6.
Literature Cited
(1) Howarth, R 84 Sprague, J. B. Water Res. 1978,12,455-462.
(2) Pagenkopf, G. K.; Russo, R C.; Thurston, R V. J. Fish.
Res. Board Can. 1974, 31, 462-465.
(3) Miller, T. G.; Mackay, W. C. Water Res. 1980,14,129-133.
(4) Chakoumakos, C.; Russo, R C.; Thurston, R. V. Environ.
Sci. Technol. 1979,13, 213-219.
(5) Eisler, R; Gardner, G. R J. Fish. Biol. 1973,5,131-14?
(6) Eisler, R.J. Fish. BioL 1974,6, 601-2.
(7) Skidmore, J. P.; Tovell, P. W. A. Water Res. 1972, 6s,
217-230.
(8) Daye, P. G.; Garaide, E. T. Can. J. Zool. 1976,54,2140-55.
(9) Phillips, G. R; Russo, R C. "Metal Bioaccumulation iu
Fishes and Aquatic Invertebrates: A Literature Review";
EPA-600/3-78-103, UJS. Environmental Protection Agency
Duluth, MN, 1978.
(10) Brungs, W. A.; Leonard, E. R; McKim, J. M. J. Fish Res.
Board Can. 1973, 30, 583-86.
(11) Packer, R K; Dunson, W. A. J. Exp. Zool. 1970,174,65-72,
(12) Lore, H. W.; McPherson, B. P. J. Fish Res. Board Can. 197$,
33, 2023-30.
(13) EPA-440/9-76-023, U.S. Environmental Protection Agency,
Washington, DC, 1976.
(14) Pagenkopf, G. K. "Introduction to Natural Watte
Chemistry"; Marcel Dekker. New York, 1978.
(15) Andrew, R W.; Biesinger, K. E.; Glass, G. E. Water Res
1977,11, 309-15.
(16) Sunda, W. G.; Engel, D. W.; Thuotte, R. M. Environ. Set
Technol. 1978,12, 409-413.
(17) Pagenkopf, G. K. In "Zinc in the Environment, Part 2:
Health Effects"; Nriagu,J.O., Ed.; Wiley-Interscience: New
York, 1980; pp 353-361.
(18) Sunda, W. G.; Hanson, P. J. In "Chemical Modeling in
Aqueous Systems"; Jeanne, E. A., Ed.; American Chemical
Society: Washington, DC, ACS Symp. Ser. No. 93.
(19) Calamari, D.; Marchetti, R Water Res. 1973, 7,1453-64.
(20) Mount, D. I. Water Res. 1968, 2, 215-223.
(21) Mount, D. I.; Stephen, C. E. J. Fish Res. Board Can. 1969,
26, 2449-2457.
(22) Pickering, Q. H.; Henderson, C. Air Water Pollut. Int. J
1966, 10, 453-463.
(23) Holcombe, G. W.; Andrew, R W. EPA-600/3-78-094, U.S.
Environmental Protection Agency, Washington, DC, 1978.
(24) Davies, P. H.; Goettl, J. P., Jr.; Sinley, J. R.; Smith, N. F.
Water Res. 1976,10,199-206.
(25) Calamari, D.; Marchetti, R; Vailatis, G. Water Res. 1980,
14, 1421-1426.
(26) Sprague, J. B. Water Res. 1969, 3, 793-821.
(27) Sprague, J. B. Water Res. 1970, 4, 3-32.
(28) Eaton, J. G. Water Res. 1973, 7,1723-1736.
(29) Pickering, Q. H.; Cast, M. H. J. Fish Res. Board Can. 1972,
29, 1099-1106.
346 Environ Sci Technol . Vol 17. No 6, 1983
-------
Environ. Set. Techno!. 1983, 17. 347-352
(30) Brown. V. M. Water Res. 1968, 2, 723-733.
(31) Brown, V. M., Dalton, R. A. J. Fish Biol. 1970, 2, 211-216.
(32) Anderson, P D., Weber, L. J. Toxicol. Appl. Pharmacol.
1975. 33, 471-483.
(33) Anderson, P. D.; Weber, L. J. Proc. Int. Con/. Heavy Met.
1973, 2, 933-953.
(34) Muska, C. P.; Weber, L. J. EPA 600/3-77-085, U.S. Envi-
ronmental Protection Agency, Washington, DC, 1977, p
71-87.
(35) Jones, J. R. E. "Fish and River Pollution"; Butterworths:
London, 1964; 107-116.
(36) Daye, P. G.; Garside, E. T. Can J. Zool 1976,54,2140-55.
(37) Zitko, V Proceedings of Toxicity to Biota of Metal Forms
in Natural Water, International Joint Commission, 1976,
pp9-32
(38) Zitko, V., Carson, W. V.; Carson, W. G. Bull. Environ.
Contam. Toxicol. 1973, 10, 265.
(39) Perrin, D. D.; Sayce, I. G. Talanta 1967, 14, 833-42.
(40) Harnss, D. K.; Ingle, S. E.; Magnuson, V. R.; Taylor, D. K.
REDEQL-UMD, Department of Chemistry, University of
Minnesota—Duluth, 1982.
Received for review November 25, 1981 Revised manuscript
received October 25,1982. Accepted February 17, 1983.
-------
Copper and Cadmium Binding to Fish Gills: Modification by
Dissolved Organic Carbon and Synthetic Ligands
Richard C. Playle1 and D. George Dixon
Department of Biology, University of Waterloo, Waterloo, ON N2L 3C1, Canada
and Kent Burnison
National Water Research Institute, Environment Canada, Burlington, ON L7R 4A6, Canada
Playle, R.C., D.G. Dixon, and K. Burnison. 1993. Copper and cadmium binding to fish gills: modification by
dissolved organic carbon and synthetic ligands. Can. J. Fish. AquaL Sci. 50:2667-2677.
Adult fathead minnows (Pimephales promelas) were exposed to 17 jig Cu • L"1 or 6 |tg Cd • L"1 for 2 to 3 h in
synthetic softwater solutions at pH 6.2 containing either naturally-occurring, freeze-dried dissolved organiccacbon—7
(DOC) or synthetic ligands such as EDTA. After exposures, gills were assayed for bound Cu or CdyAs a first (
approximation, lake_Qf origin or molecular size fraction of DOC did not influence Cu binding to gills, while DOC j
concentration did.JDOC concentrations £4.8 mg • L~' prevented Cu from accumulating on fathead gillsJTif'trieV
"relatively low concentrations used, neither Cu nor Cd interfered with binding of the other metal on gills, j
^suggesting different gill binding sites.fCadmium accumulation on gills was more sensitive to increased
'concentrations of Ca and H* than was Cu. Surprisingly, Cd bound to gills to the same or greater extent than
did Cu: for synthetic ligands, Cd binds less well than Cu. This result corroborates previously published
observations that Cd, unlike Cu, is taken up at gills through high affinity Ca channels. Accumulation of Cd
on fish gills was never associated with 14C-labeTled EDTA or 14C-citrate, indicating that free metal interacts
with the gill while metal-ligand complexes usually do-not.
Des adultes du Tete-de-boule (Pimephales promelas) ont ete exposes a Cu en concentration de 17 ng • L"1 ou du
Cd en concentration de 6 |ig • L~1 pendant 2 ou 3 h dans des solutions d'eau douce synth&ique a pH 6,2 qui
contenaient du carbone organique dissous (COD) lyophilise et qu'on trouve naturellement, ou des ligands
synthetiques comme I'EDTA. La periode d'exposition terminee, le Cu ou le Cd fixe sur les branchies, e'tait litre. En
premiere approximation, le lac d'origine ou la taille molgculaire du COD n'agissait pas sur la fixation du Cu aux
branchies, alors que la concentration du COD avail un effet A une concentration & 4,8 mg • L"1, le Cu ne
s'accumulait pas sur les branchies des Tetes-de-boules. Aux assez faibles concentrations testers, ni le Cu ni le Cd
interfe>ait avec la fixation des autres me'taux sur les branchies; cela donne a penser qu'il existe diffe>enls siles de
fixation sur les branchies. Plus que le Cu, ('accumulation de cadmium sur les branchies e'tait davantage sensible
aux hausses de concentration du Ca et du H*. Fait surprenant, le Cd se fixait aux branchies autant ou meme plus '
que le Cu : avec les ligands synthetiques, le Cd se fixe moins bien que le Cu. Ce resultat vienl confirmer des
observations dont il a de\h et6 fail mention, a I'effet que le Cd, contrairemenl au Cu, est absorbe au niveau des
branchies par des voies a tres forte affinite' pour le Ca. L'accumulation de Cd sur les branchies des poissons n'a
jamais &e associ£e a I'EDTA marque au 14C ou au citrate marque au 14C; cela indique que les atomes metalliques
libres passenl en interaction avec les tissus des branchies alors que les complexes de metaux-ligands ne le font
habiluellemenl pas.
Received September 18, 1992
Accepted May 18, 1993
(JB634)
Recu le IBseptembre 1992
Accept^ le 18 mai 1993
Waterborne metals generally show their greatest
toxicity to aquatic organisms in soft water of low
alkalinity, low pH, and low dissolved organic carbon
(Sprague 1987; Flemming and Trevors 1989; Spry and Wiener
1991). This relationship can be explained by the usual
conceptual model of metal toxicity, which considers the free ion
to be the toxic form of the metal (Morel 1983; Pagenkopf 1983).
In essence, waterborne free metal ions must adsorb to gills before
they can either exert their toxic effect at the gill surface directly,
or pass through the gills on their way to internal sites of toxic
'Author to whom correspondence should be addressed. Current
address: Department of Biology, Wilfrid Laurier University,
75 University Ave. West, ON N2L 3C5, Canada.
Can. J. Fish. Aquat Sci., Vol. 50. 1993°-
action (Simkiss and Taylor 1989). Any process that prevents
initial adsorption on the gill surface by reducing either the
ambient free metal ion concentration, or the number of surface
binding sites on the gill, will reduce toxicity of -waterbome
metals.
A number of conditions can affect the degree to which metals
bind to gill surfaces. For instance, hardness cations such as Ca2+
and Mg2+ compete for metal ion binding sites, and metal
complexing agents such as dissolved organic carbon (DOC) may
reduce the free metal available to interact with an organism
(Morel 1983; Pagenkopf 1983). Low water pH may also affect
metal binding through H* competition for binding sites, and by
affecting metal speciation (Campbell and Stokes 1985). In
2667
-------
addition, H+ can displace metals from DOC, increasing available
free metal ions (Cabaniss and Shuman I988a). A number of
recent studies using fish have shown that toxicity or
accumulation of both Cu and Cd varies according to competition
and complexation (for Cu: Cusimano et al. 1986; Hutchinson and
Sprague 1986, 1987; Playle et al. 1992; for Cd: Part and
Wikmark 1984; Cusimano et al. 1986; Wicklund and Runn 1988;
Pratap et al. 1989; Bentley 1991).
In fish, waterborae Cu and Cd affect ion uptake through
interactions at the gill. Copper (12-50 fig • L'1) affects ion
balance by reducing Na* and Cl~ influx at gills (Lauren and
McDonald 1985), mainly through effects on Na+-K*-ATPase
(Lauren and McDonald 1987). High Cu concentrations
(200 u,g • L~') increase gill permeability, as shown by increased
ion efflux, probably because of displacement of Ca from inter-
cellular tight junctions (Lauren and McDonald 1985). Cadmium
also affects ionoregulation, but through interference with Ca
uptake at the gill, largely through effects on a high affinity
Ca2*-ATPase (Verbost et al. 1989). At Cd concentrations
<112 jig • L~' in hard water, Na* influx is unaffected (Verbost
et al. 1989). Because gills are the target organ of Cu and Cd, any
study of the mitigating effects of water hardness or DOC should
measure changes in the actual amount of metal accumulating on
the gills.
In a previous study (Playle et al. 1992) we characterized the
effects of H+ and Ca competition, and EDTA complexation, on
Cu binding to gills of fish held in soft water. Our first objective
of the present paper was to expand the previous work to include
Cd, and to examine the effects of naturally occurring DOC on
both Cu and Cd binding to gills. Our second objective was to
estimate metal-gill stability constants for Cu and Cd, by
comparison to ligands of known metal binding ability. Our
method of using complexing ligands to determine metal-gill
stability constants is based on the premise that complexed metal
does not react at fish gills, while free metal does. We verified this
assumption using Cd and l4C-labelled EDTA and MC-labelled
citric acid. Detailed calculation of the metal-gill binding
constants, their incorporation in a model, and testing of the
model with fish exposed to Cu and Cd in natural soft water, are
subjects covered in the companion paper (Playle et al. 1993).
Materials and Methods
Experimental Protocol
^ Adult fathead minnows (Pimephales prvmelas Rafinesque)
were raised from eggs in hard ground water at the University of
Waterloo (Ca = 2860 u,M,= 290mg-L-' as CaCO3). At least
2 wk before each experiment fish were transferred to 20-L
containers with aerated synthetic soft water, that was produced
by adding NaCl and CaCl2-H2O (to reach final Na and Ca
concentrations of -50 u,M) to water from a mixed bed deionizer
(Anderson Water Systems Ltd., Dundas, ON). Mean ± 1S.E.M
(n) water conditions during acclimation were: Na, 163 ± 10
(26) jtM; Ca, 143 ± 9 (26) jiM; pH, 6.23 ± 0.07 (26);
temperature, 18.5 ± 0.4 (24)°C; Cu, 2.5 ± 0.3 (22) u,g • L-'; and
Cd, <0.1 u,g • L~' (17). Fish were fed daily with frozen brine
shrimp.
For each experiment, 3-6 acclimated fish (0.5-4 g) were
placed randomly in 1 or 2 L of aerated soft water in polyethylene
containers, to which was added CaQ2 • 2H2O (Mallinckrodt;
Paris, Kentucky), 1 N H2SO4, freeze dried natural dissolved
organic carbon (DOC), or various ligands, depending on the
26t
experiment (see below). Background Cu was 2.1 ± 0
(32) u,g • L-'; with added Cu (as CuSO4- 5H2O; AnalaR,
Toronto, ON) the concentration was 17.4 ± 0.3 (114) u,g • L~'
(=0.27 uM). Backgrounded was <0.1 (33) u,g • L'1; with added
Cd (as CdCl2 • 2±H2O; Baker Chemical Co., Phillipsburg, NJ) the
concentration was 5.5 ± 0.1 (63) jtg • L"1 (=0.05 (iM). Although
these Cu and Cd concentrations are -10X higher than those
usually found in soft acidified surface waters (Hutchinson and
Sprague 1986; Johnson 1991;Spry and Wiener 1991), they gave
measurable metal accumulation on gills within 2-3 h of exposure
(see Results). For one metal mixture experiment, zinc was added
asZnSO4- 7H2O (AnalaR, BDH) to reach 19.1 ± 1.1 (8) fig • L~'
(=0.29 jiM); background Zn was 2.8 ± 0.5 (10) ng • L''. Unless
modified, exposure pH was 6.24 ± 0.01 (159), Na was 66 ± 1
(156) jtM, Ca was 34 ± 1 (134) jiM, and temperature was
19.2 ± 0.3 (21)°C. Exposure time was 2-3 h, which has been
shown to be enough time to allow metal accumulation on gills
to stabilize (Playle et al. 1992; see Results). Treatment containers
were randomized, and each treatment was usually run twice.
Dissolved organic carbon (DOC) was isolated from surface
waters of several lakes near Peterborough, Ontario, namely:
Salmon L., Sharps Bay and Williams Bay of Jack L., Anstruther
L., Loucks L., and Wolf L. (all approximately N44°4jr,
W78°13'). DOC from the water of Luther Marsh, near Guelph,
Ont. (N43°5T, W80°26') was also isolated. Water was collected-
in 20-L stainless steel containers in 1988. Tangential flow
ultrafiltration (Millipore, Bedford, MA) was used to isolate DOC
into three fractions: 1000-30 000 Da, 30 000 Da to 0.2 u,m, ancjj
>0.2 u,m. The DOC fractions were then freeze dried and stored
at room temperature in polyethylene containers. The DOC
fractions were converted to the soluble sodium form in Milli Q
(Millipore) water in the presence of AG50W-X8 cation exchange
resin (Na form, BioRad, Richmond, CA). After 20 h the solutions
were filtered through 0.2 u,m polycarbonate filters (Nuclepore
Corp., Pleasanton, CA). The concentrated filtrate was added
directly to the aerated synthetic soft water used in the metal
exposure experiments. Water DOC concentrations were
measured, after acidification and sparging to remove inorganic
carbon, using a Beckman model 915B total carbon analyzer. We
used DOC concentrations of <8 mg • L~' (= <8 mg C • L~'), that
are representative of soft lake waters (Spry and Wiener 1991).
In experiments to determine metal-gill binding constants, we
used the ligands ethylenediaminetetraacetic acid (EDTA;
disodium salt; Baker analyzed reagent), nitrilotriacetic acid
(NTA; disodium salt, Sigma Chem. Co., St. Louis, MO), ethyl-
enediamine (EN; Sigma), citric acid (AnalaR, BDH), glutamic
acid (Baker), oxalic acid (AnalaR, BDH), and salicylic acid
(Fisher). These ligands were chosen to ensure a wide range of
metal complexation strengths. Ligands were prepared as 1 mM
stock solutions in deionized water. Calculations of free and
ligand-bound metal concentrations were made using MINEQL+
(version 2.1; Schecher 1991) on an IBM-compatible computer,
using measured pH (fixed) and water chemical concentrations.
Equilibrium stability constants at zero ionic strength (K0, given
as K in this paper) were converted to values appropriate to thej
ionic strength of our water (/ ~ 1 X 10~4) by the compute
program. Equilibrium constants were from the MINEQL+
program, and are summarized in Table 1.
At the end of each experiment the fish were killed with a blow
to the head, both sets of gills (= gill baskets) were excised, and
the gills were rinsed for 10 s in 100 mL synthetic soft water. The
gills were then weighed (mean wet weight = 0.07 g) and digested
Can J. Fish Aquat. Sci., Vol 50. 1993
-------
TABLE 1. Log of equilibrium constants (1:1 binding to ligands) from the MINEQL+ computer program.
EN = ethylenediamine.
EDTA
NTA
EN
Citrate Oxalate Glutamate Salicylate
Cu-ligand
Cd-ligand
Ca-ligand
H-ligand
18.8
16.3
12.5
10.0
13.1
9.4
8.2
10.3
10.5
5.6
0.1
10.0
7.3
5.3
4.7
6.3
5.1
3.4
3.6
4.2
8.3
4.8-
2.5
10.0
10.6
6.7
—
13.4
for 8 h at 80°C in 5X their weight of 1 N H2SO4. Gill digests
were vortexed, centrifuged, and the supematent diluted 20X
with deionized water. For the NTA-Cd and citrate-Cd
experiments, run 6 mo after all other experiments, 1 N H2SO4
added to gill samples was 10X the gill weight, with a later 10X
dilution with deionized water. For all experiments, gill and water
Cu and Cd concentrations were measured by graphite furnace
atomic absorption spectroscopy (AAS; Varian AA-1275 with
GTA-95 atomizer) using 10-jJuL injection volumes, N2 gas, and
standard operating conditions documented by Varian. Water Na
and Ca concentrations were measured using a Perkin Elmer 4000
AAS, Zn concentrations using graphite furnace AAS, and pH by
a Radiometer PHM82 meter with GK2401C combination
electrode.
We compared Cu and Cd concentrations on gills of fish
exposed to modified soft water with those of all fish exposed to
metals in standard soft water conditions (grey horizontal bands
in Figures) and with those of all unexposed fish (dashed
horizontal band in Figures). Comparisons of gill metal concen-
trations were made using ANOVA followed by the Tukey-
Kramer procedure. Data are plotted in the Figures as means
±95% CI.
14C-Ligand Experiments
Experiments were run to determine the contribution of free
versus ligand-bound metal to the metal load of fish gills. If only
free metal binds to gills, then the metal alone would be detected
in gill samples. If metal complexed by a radiolabelled ligand
binds with gills, then both the metal and radiolabel would be
detected. We used MC-EDTA (tetrapotassium salt) and
l,5-l4C-citrate for these experiments (Sigma; specific activity
11.8 and 60 mCi • mmol"1, respectively). Cadmium was chosen
as the test metal because less is needed for detectable
accumulation hi fathead gills, compared to Cu, using graphite
furnace AAS.
The I4C experiments were run in a similar manner to those
outlined above. Two or 3 minnows were placed in 1 L of
synthetic unaerated soft water for 2-3 h. The exposure water
contained 0 or O.OS (imol Cd (6 u,g • L~'), 0 or 0.1 u,mol
'«C-EDTA, or 0 to 250 jtmol mixtures of 14C-citrate and "cold"
citrate. We aimed for ~107 disintegrations per minute (DPM) of
MC • L"1, to yield reasonable counts in the gill samples. Gills
were excised as before, weighed, and digested for 8 h at 80°C in
10X their weight of 1 N H2SO4. Digests were vortexed and
100 jiL was diluted 10X with deionized water before Cd
analysis by graphite furnace AAS. Remaining solutions
(0.4-1 mL) were added to 16 mL of Cytoscint ES fluor (ICN
Biomedicals Inc., Irvine, CA) in 20-mL plastic scintillation vials
(Fisher). The vials were shaken, and then counted in a Beckman
LS 1701 liquid scintillation counter with automatic quench
correction. The I4C label in water was determined by counting a
1-mL water sample in 16 mL of fluor.
To test detection of I4C in the gill samples^five gill samples
giving background counts were spiked with 10-200 jtL of
~2 X 10* DPM l4C-citrate • L~' solution. In addition, l4C-labelled
anthracene (Sigma), a lipophilic polyaromatic hydrocarbon, was
used to measure the uptake of a l4C-compound known to cross
biological membranes. Ten microlitres of 0.1 jtCi
14C-anthracene • \iL~l (=0.07 ujnol) in dimethyl sulphoxide
carrier (DMSO; Omnisolv, BDH) was added to 1 L of synthetic
soft water, to yield ~2 X 106 DPM • L~'. Three minnows were
held in this water for 2 to 3 h. Ten microlitres of DMSO in 1 L
was tested with l4C-citrate (-4 X 10« DPM • L~') to determine
if DMSO contributes to citrate entry into gills. Gill and water
samples were counted as before.
Because citric acid is a substrate in the tricarboxylic acid cycle,
it was possible that l4C-citrate bound by fish gills could be
metabolized to I4CO2 and expired. This possibility was tested
using 250-mL solutions in sealed jars with and without
l4C-citrate, Cd, and fish. Water in the jars was acidified after 2 h
to pH 3.7-4.0 with 1 N H2SO4 to liberate CO2 from the test
waters. Containers, with pleated pieces of filter paper soaked in
1 mL of ,10% KOH, were suspended inside the jars to trap
evolved CO2. Filter papers plus the KOH were added to 16 mL
of fluor, and 200 jxL of glacial acetic acid was added to each
sample to reduce chemolumescence. Samples were counted after
storage in the dark overnight
Results
Time Course and Effects of H+ and Ca on the Accumulation
of Cd on Minnow Gills
Cadmium accumulation on fathead minnow gills was linear to
about 2 h (r= 0.92, P < 0.001), after which accumulation
levelled off (Fig. 1). Two- to 3-h exposures were therefore
appropriate for assessing initial accumulation of Cd on minnow
gills. The Cd accumulation pattern was the same whether or not
0.1 fiM EDTA was present at twice the Cd concentration,
however no Cd accumulated on gills in 2 h if 0.25 jiM EDTA (at
5X the Cd concentration) was present (Fig. 1).
Increasing water Ca concentration from 35 to 95 u,M did not
significantly reduce Cd accumulation, but Ca concentrations of
1055 and 2000 \iM fully prevented Cd accumulation on gills
(Fig. 2). In Cd exposures run at pH 4.83 ± 0.08 (7) with no
additional Ca, no Cd accumulated on gills (Fig. 2).
Effects of Natural Dissolved Organic Carbon on Metal
Binding to Gills
To determine if either the source of dissolved organic carbon
(DOC), or the size fraction of DOC, affected Cu accumulation
on gills, we tested three size fractions of DOC from five lakes
and Luther Marsh. As preliminary data using Luther
Marsh DOC suggested that 5 mg DOC-L-' prevented Cu
Can 1. Fish. Agum. Sci.. Vol 50. 1993
2669
-------
0.8
4.0
Fta. 1. Time course of Cd accumulation on gills of fathead minnows exposed to 6 jig Cd • L~' (0.05 »J.M) in soft water(»). EDTA (0.1
. (A) did not affect Cd accumulation, whereas 0.25 jiM EDTA did (•). O = fish not exposed to Cd. • and O were only continued to 2.5 h. A
straight line was fit to the • and A data to 2.5 h. Dashed horizontal lines represent die 95% CI of gill Cd of all fish not exposed to Cd (n = 38).
The grey horizontal band represents the 95% confidence interval of gills of all fish exposed to 6 fig Cd • L"1 for 2-3 h in synthetic soft
water (n = 45). Background and Cd-exposed gills had significantly different concentrations of Cd (P < 0.001).
pH 6.3
pH 4.8
FIG. 2. Effects of Ca and H* on Cd accumulation on gills of minnows
exposed to 6 jig Cd • L~' for 2 to 3 h in soft water (pH 6.2). ss1,055 uM
Ca prevented Cd accumulation on gills. No Cd accumulated on gills of
fish held at pH 4.8 (no added Ca). Vertical lines represent the 95%
confidence interval (from left to right, n = 5,6,6,5). For all figures, *,
**,*** = significantly below metal-exposed gills (grey horizontal band)
and +, ++, +-H- = significantly above background gill metal
concentrations (dashed horizontal lines) for P < 0.05, < 0.01, < 0.001,
respectively.
accumulation on minnow gills (data not shown), we tried to test
that amount of each DOC fraction, although not all size fractions
for all lakes were examined.
Accumulation of Cu on gills offish exposed to 17 u,g Or L~' in
the presence of DOC indicated that DOC concentration was the
main factor in decreasing Cu binding to gills (Fig. 3). All gill Cu
concentrations were reduced to background at DOC
S4.8 mg- L"1, regardless of DOC size source or fraction, and
were significantly reduced relative to gills from fish exposed to
Cu in the absence of DOC (P < 0.001). Below 4.8 mg- L~" little
or no protective effect of DOC was observed, except for two
values at 3.7 mg-L~'. These two results were from the
- DOC <2
,-.lJ.
3 2
F
F;
4
J
- -
M
. i •
6 8
DOC (rng-L'1)
3.3. Gill Cu concentrations of fathead minnows exposed for 2-3 h in
soft water (pH 6.2) to 17 jtg Cu -L"1 and various concentrations of
dissolved organic carbon (DOC). Three size fractions of freeze-dried
DOC from various lakes were used: • = 1000-30000 Da, • =
>30 000 Da but <0.2 pm, and A = >0.2 u,m. Gill Cu was reduced
significantly at all DOC concentrations £=4.8 mg • L~'. The data fit a
straight line: gill Cu-= -0.26 • [DOC] + 2.52 (r = -0.77, P < 0.001).
Vertical lines represent the 95% confidence interval (n = 5 to 6 fish gills).
Dashed horizontal lines represent the 95% CI of background gill Cu
(n = 67). The grey horizontal band represents the 95% CI of all
Cu-exposed fish gills (n = 74) in the absence of added DOC (DOC
<2 mg • L~'). Background and Cu-exposed gill Cu concentrations were
significantly different (P < 0.001).
1000-30 000 Da DOC fraction from Salmon L. and Williams
Bay; neither was significantly different from background gill Cu,
and both were significantly lower than Cu-exposed gills in
absence of added DOC (P < 0.001). These two values sug
that lower molecular weight DOC fractions complex Cu to a
greater degree than do higher molecular weight fractions.
Otherwise it appears that, as a first approximation, DOC
concentration, not source or size fraction, is the dominant factor
in determining metal accumulation on fish gills.
Next, we looked, at the effects of DOC on Cu and Cd
2670
Can J. Fish. Aquat. Sci. Vol SO. 1993
-------
3 -
2 -
5 mg-L1 DOC
O>
a.
Cu Cu. Cd, Zn Cd Cu. Cd, Zn
Fie. 4. (A) Gill Cu concentrations of fathead minnows held for 2-3 h in soft water (pH 6.2)
with either 17 u.g Cu • L~L or 17,6, and 19 u,g Cu, Cd, and Zn • L"1, in the presence
(stripes) or absence of DOC. Five milligrams DOC per litre reduced gill Cu to background;
(B) gill Cd concentrations of fathead minnows held for 2-3 h in soft water (pH 6.2) with
either6u.gCd • L"1 or 17,6,and 19 ng Cu, Cd, and Zn • L~' ,in the presence or absence
of DOC. Five milligrams DOC per litre (stripes) did not significantly reduce Cd
accumulation on the gills.
accumulation on minnow gills, either present as individual
metals or as a mixture of Cu, Cd, and Zn. Once again,
5 tag DOC- L-' (1000-30 000 Da Luther Marsh) prevented
Cu accumulation on fathead gills (Fig. 4A). The simultaneous
presence of 6 jig Cd- L"1 and 19 jig Zn- L"1 did not affect Cu
deposition, in either the absence or presence of DOC. This
result indicates either strong binding of Cu to the gills
(relative to Cd and Zn), excess metal binding sites on the gills,
or different binding sites for each metal. Background gill Zn
concentrations were so high (20-40 u,g Zn- g wet tissue"1) that
the 19 M-8 Zn* L'1 exposures did not significantly alter gill Zn
concentrations (data not presented).
Cadmium accumulation was unaffected by simultaneous
exposure to Cu and Zn (Fig. 4B), again indicating either strong
Cd binding, or excess or different metal binding sites on gills.
With 5 mg DOC- L"1 there was no significant reduction in the
amount of Cd on the gills (Fig. 4B), in marked contrast to the
results for Cu (Fig. 4A). Later experiments to determine the
concentration of DOC needed to prevent Cd accumulation on
gills, yielded 0.29 ± 0.15 (6) and 0.30 ± 0.08 (6) u,g Cd • g wet
tissue'1 ( ±95% CI) for 5.1 and 7.7 mg DOC -L'1 of the
1000-30 000 Da Luther Marsh fraction, respectively. Both these
concentrations were significantly below values from Cd-
exposed fish in the absence of DOC (P< 0.001, <0.01,
respectively), and were not significantly above background.
Complexation of Cu and Cd Using Synthetic Ligands
We used synthetic ligands to estimate conditional metal-gill
equilibrium constants (K) for Cu and Cd. The approach is based
on equilibrium constants for metal-ligand complexes in soft
water, such as those published in Morel (1983) and used in
computer programs like MINEQL* (Schecher 1991). For
example, the Cu-EDTA complex (log K = 18.8) forms about
300X more strongly than does Cd-EDTA (log K = 16.3), and
about 0.5 X WX more strongly than does the Cu-NTA complex
(log K = 13.1; log K values from MINEQL*, summarized in
Table 1). Copper binds to NTA about 5000X more strongly than
does Cd (log KCM-NTA = 9.4). Note that in all cases Cu binds to
these ligands better than does Cd. By exposing fish to metals and
ligands at known concentrations in water of defined
composition, and measuring metal accumulation on gills, metal-
gill stability constants can be estimated.
Copper
To estimate Cu-gill stability constants, we exposed fathead
minnows to 17 u,gCu-L~' (0.27 p,M) in soft water in the
presence of various ligands. 0.25 u,M EDTA prevented Cu from
accumulating on gills (Fig. 5), indicating that enough Cu was
bound in the Cu-EDTA complex to render Cu unavailable to bind
to the gills. MINEQL* calculations yielded 0.019 |jiM free Cu,
with the remaining Cu in the Cu-EDTA form. Likewise, 0.25 u,M
NTA prevented Cu accumulation (Fig. 5). Only about 0.030 u,M
Cu was calculated to be in the free form in the presence of
0.25 uM NTA (MINEQL* calculation). 2.5 uM NTA also reduced
gill Cu to background (0.61 ± 0.16 (4)); only about 0.0002 u,M
Cu was calculated to be free at this NTA concentration.
Interpretation of results using ligands that bind Cu more
weakly than EDTA and NTA is complicated by the fact that Ca
and H* compete with Cu for these ligands (e.g. log K values are
similar, Table 1). Thus, not all the added ligand is available to
bind Cu. For example, only 0.04 u,M of the 0.25 u,M ethyl-
enediamine was in the CuEN form (-16%), the rest of the ethyl-
enediamine being in the H~ or H2' form (MINEQL* calculation).
In contrast, >96% of the 0.25 u.M EDTA or NTA was available
to bind Cu (above). Ethylenediamine, citric, and oxalic acid
(25 u,M) were needed to prevent Cu accumulation on gills
(Fig. 5), for which 0.006,0.007, and 0.044 u,M, respectively, of
the 0.27 u,M Cu in solution, was calculated to be free. In addition
to these data, we had conducted a series of experiments with
glutamic and salicylic acids, in which 25 u,M glutamate reduced
gill Cu accumulation slightly (but not significantly; to 1.84 ± 0.54
(6)), and 250 p.M salicylic acid was needed to reduce gill Cu to
1.31 ± 0.28 (5), not significantly above background.
Can. J. Fah. Aquat. Sci.. Vol. 50.1993
261 \
-------
1
3
5
8
3 -
1 -.
FlG. S. Amount of Cu on the gills of fathead minnows exposed to Cu
(0.27 »iM = 17 »tg • L-') and ligands for 2-3 h in soft water at pH 6.2.0.25 ftM EDTA
and NTA prevented Cu accumulation on the gills, whereas higher concentrations of the
other ligands were needed to do so. From left to right, n = 10,4,9,9,9,6,5,5,6,5.5.
The horizontal bands designated by dashed lines or grey shading represent, respectively, the
95% CI for background gill Cu, and Cu-exposed gills in the absence of ligands.
Cadmium
The results for Cd binding to gills in the presence of synthetic
ligands were surprising. Although Cd binds less well to synthetic'
ligands than does Cu (Table 1), Cd bound to minnow gills as well
or better than did Cu. Between 2 and SX more EDTA than Cd
was needed to prevent Cd accumulation on gills (Fig. 6; see also
the Cd time course experiment. Fig. 1). This much EDTA would
leave between 0.001 and 0.005 (iM of the 0.05 |iM total Cd in
solution as free metal (MINEQL+ calculation). Extremely low
concentrations of free Cd resulted in significant Cd
accumulation on gills, in contrast to the 0.006 to 0.04 p,M free
Cu that was not enough free Cu to result in increased gill Cu
concentrations (above).
Both Ca and H* complex with NTA and citric acid, reducing
the amount of ligand available to bind Cd. NTA (5 u-M) and
2500 |iM citric acid significantly reduced Cd accumulation on
minnow gills (Fig. 7). Free Cd under these conditions was
calculated to be 0.03 and 0.0002 jiM, respectively. The citric
acid results agree with the Cd-EDTA results: virtually all Cd in
solution must be complexed, otherwise measurable Cd accu-
mulation occurred. The Cd-NTAresults agree with the Cu-oxalic
acid results, in that 0.03-0.04 u,M of free metal was not enough
to result in metal accumulation on the gills. Considering all the
Cd and Cu data on a molar basis, there needs to be the same or
less free Cd than Cu if no metal is to accumulate on the gills.
That is, the affinity of Cd for the gills is the same or greater than
that of Cu.
14C-Ligand Experiments
To this point the experimental design and our interpretation of
the results were based on the assumption that free metals interact
with fish gills, while metal-ligand complexes do not. Copper-
EDTA and Cu-NTA complexes apparently do not bind to fish
gills (Fig. 5), but Cu accumulation in the presence of low
concentrations of ethylenediamine, citric acid, or oxalic acid
could be interpreted as Cu-ligand accumulation on or in gills.
Similarly, elevated Cd accumulation on gills for the lower ligand
concentrations (Fig. 6, 7) could have been due to Cd-ligand
accumulation on or in gills. To ensure the correctness of our
interpretations, we wanted to verify that the free metal, not the
metal-ligand complex, was responsible for elevated gill metal
concentrations in our experiments. Numbers offish used in these
experiments were small, due to the expense of the radiolabelled
ligands.
Three fathead minnows exposed to I4C-EDTA (0.1 ujnol
EDTA = 2.4 X106 DPM) in 1 L of soft water (no Cd) had
background gill Cd values of 0.11-0.13 u.g Cd-g-', anc!
36-48 DPM per gill sample (Fig. 8), the same level of
radioactivity as fish exposed to Cd without radiolabel
(29-43 DPM per gill sample). Thus, 14C-EDTA did not
accumulate on minnow gills. In Cd plus I4C-EDTA exposures,
gills accumulated Cd but there was no increase in gill DPM
(Fig. 8). MINEQL+-calculated Cd-EDTA was 90% of total Cd.
Assuming 1:1 .Cd:EDTA binding, the expected counts if 90% of
the Cd accumulation was due to Cd-EDTA would be 500-
1 800 DPM (gill samples = 0.07 g wet weight each; allowing a
generous (50%) margin for counting inefficiency and loss of
sample). To ensure that these seemingly large counts would be
detected, we added 10-200 u,L of 0.1 fiM l4C-citrate
(= 2.45 X 106 DPM • L-') to gill samples from control ftsh. Gill
DPM versus radiolabel added was a linear relationship
(r = 0.999, n = 5). A 10-ul, spike (1 X 10'6 junol citrate)
yielded 22 DPM above background, and a 20-u,L spike yielded
44 DPM above background. An additional 20-40 DPM (1.5-2 X
background counts) would therefore easily be detected in our
system.
l4C-labelled anthracene, a iipophilic polyaromatic
hydrocarbon, was also used to test detection of I4C. About
2 X 106 DPM of l4C-anthracene (0.07 u,mol) was added to 1 L
of soft water, and three fish were exposed to this solution for
2-3 h. These fish had gill counts of 370-550 DPM (Fig. 9),
whereas two control fish had gill counts of 39 and 46 DPM. Two
fish exposed to the dimethyl sulphoxide (DMSO) carrier plu
0.1 jimol MC-citrate (= 4.2 X 10* DPM • L-') had gill counts
49 and 58 DPM (Fig. 9); anthracene's entry was therefore due to
its own properties, not to DMSO. The uC-anthracene experiment
also indicates that 500-1 800 DPM (I4C-EDTA experiment)
would easily have been detected in our system.
Cadmium accumulation on gills was also never associated
with l4C-citrate (Fig. 10). Nine fish exposed to 0.1 u,M of
2672
Can J Fish Aqual Set., Vol. 50,1993
-------
0)
w
JO
»5
I
s
O)
O)
0.8
0.6
0.4
0.2
nu/i3»)fti
ICd = 0.05 |iM
FIG. 6. Amount of Cd on gills of fathead minnows exposed to Cd
(6 u,g • L~l = 0.05 n.M) and EDTA for 2-3 h in soft water at pH 6.2. It
took between 2 and 5 X more EDTA than Cd to prevent Cd accumulation
on the gills. From left to right, n = 13, 4, 5,4. The horizontal bands
designated by dashed lines or grey shading represent, respectively, the
95% CI for background gill Cd, and Cd-exposed gills in the absence of
ligands.
l4C-citrate had background gill Cd concentrations and a mean
gill radioactivity value of 58 DPM. Nine fish exposed to Cd alone
had gill Cd concentrations of 0.31 ng Cd • g~' (wet tissue) and
mean gill DPM values of 44. Three fish held in 0.1 ftM
MC-citrate plus 0.05 fiM Cd showed Cd accumulation on gills,
but no l4C-citrate entry. However, MINEQL* calculations
showed 98% free Cd under these conditions, so binding of free
Cd, as opposed to Cd-citrate, would be expected. Similarly, there
was no evidence of l4C-citrate on gills offish held in 5 (iM citrate
with 0.1 u,M l4C-citrate: again, plenty of free Cd would be
available to bind at the gills (80% of total Cd) under these
conditions.
With Cd plus 25 pJvl cold and labelled citrate (16.35 X 10*
DPM in 1 L) no counts due to I4C were detected in the gill
samples (Fig. 10). Here, 57% of the Cd would be bound by citrate
(MINEQL+ calculations). Under these conditions, again with a
generous provision for counting inefficiency, an additional 18,
21, and 31 DPM would have been expected in the samples if 57%
of the Cd accumulation was due to Cd-citrate entry. These
increases would have been detected in our system, but the gill
samples were within the background range. Finally, 250 u,M
citrate (19.44 X 106 DPM in 1 L) did not prevent Cd accumu-
lation on the gills, in accord with results presented in Fig. 7. Here,
95% of the Cd was calculated to be in the Cd-citrate form, and
9,6, and 6 DPM above background would have been expected
if 95% of the Cd accumulation on the gills had been due to
Cd-citrate. These values are at the detection limit of this method,
but the gill DPM values were within background. In total, there
was no evidence that l4C-citrate accumulates on gills.
There was, however, the possibility that l4C-citrate was taken
up at the gills and metabolized to I4CO2 in the tricarboxylic acid
cycle. This scenario was tested by trapping 14CO2 in 10% KOH
Can. J Fish Aquat Sci, Vol. 50. 1993
FIG. 7. Amount of Cd on gills of fathead minnows exposed to Cd
(6 ng • L~' = 0.05 |iM) and NTA or citric acid (2-3 h in soft water,
pH 6.2). NTA (5 fiM) and 2500 ftM citric acid were necessary to prevent
Cd accumulation on gills. These experiments were done -6 mo after
those presented in previous Figures, which shifted Cd values for
background (dotted horizontal lines, n = 12) and Cd-exposed gills in the
absence of added ligands (grey horizontal band, n = 12). From left to
right, n = 5,5,5,9,5,5.
in sealed jars containing 250 mL of a 0.1 u,M l4C-citrate solution
(6.7 X 106 DPM • L-'). The KOH sample from three jars con-
taining one fish each yielded 17 270 ± 1 140 DPM (mean
± 1 SEM). With 6 jig Cd • L-' in solution, KOH from jars con-
taining fish yielded 13 310 ± 1180 DPM. KOH from two sealed
jars (no fish) with magnetic stir bars set to mimic agitation of
water due to minnows swimming gave 18210 and 23 820 DPM,
KOH from a single, unstirred sealed jar (no fish), yielded
3 240 DPM, and KOH alone gave 80 DPM. I4C caught in the
KOH traps was unrelated to the presence of fish, and appeared
to be due to breakdown or volatilization of the l4C-citrate,
encouraged by agitation of the water either by fish or by stir bars.
It is therefore unlikely that metabolic breakdown of 14C-citrate
bound to gills was responsible for low counts of gills from the
MC-citrate plus Cd experiments.
Discussion
The most surprising result of our work was the strong binding
of Cd to fathead minnow gills relative to that of Cu. Synthetic
ligand experiments (Fig. 5,6,7) indicated that the concentration
of free Cd required for significant accumulation of Cd on gills —
using graphite furnace technology — was the same or less than
the concentration of free Cu required for significant Cu binding
If adsorption on the gill surface was the sole factor determining
binding, and was similar to binding to other ligands, then Cd
would be expected to bind about 10-100X less well to gills than
would Cu (Table 1). The probable reason for the difference is
uptake of Cd through high affinity Ca channels, in addition to
surface binding, whereas Cu uptake is likely through general
surface binding plus possible uptake through low affinity ion
channels. Both Cd and Ca cations have an ionic radius of-1.2 A,
and Verbost et al. (1989) showed that any treatment that
decreased gill Ca accumulation had a similar effect on Cd
accumulation, strongly suggesting that both cations enter gills
by the same pathway.
Reid and McDonald (1991) also found that Cd bound to fish
gills better than did Cu. Radiotracers bound to excised rainbow
trout gills in the order La* > Ca2+ = Cd2+ > Cu2+. Lanthanum was
expected to bind best, due primarily to its triple charge; they
explained the weaker than expected binding of Cu to gills
through the coordination chemistry of Cu2* (Reid and McDonald
2673
-------
5
CO
o>
0.6
0.4
o
f 0.2
5
= gill Cd
-i 300
4C-EDTA
C-EDTA
FIG. 8. Gill Cd (grey bars) and disintegrations per minute (DPM; clear bars) of gill samples
from fish exposed for 2-3 h in soft water to 0.1 u.M 14C-EDTA, Cd (6 u,g • L~l = 0.05 fiM),
or 0.1 fiM |4C-EDTA plus Cd. No additional counts due to 14C-EDTA were detected. Three
fish per treatment, with th'e order of gill Cd and DPM preserved left to right in each treatment
600 r
«
"o. 400
200
Q.
O
control "C-citrate 14C-anthracene
+ DMSO + DMSO
Fie. 9. Disintegrations per minute (DPM) of fathead minnow gills from
two fish exposed to soft water only (control), two fish exposed to
0.1 n-M l4C-citrate plus dimethyl sulphoxide (DMSO), and three fish
exposed to 0.07 u.M 14C-anthracene plus DMSO. Accumulation of
14C-!abelled anthracene in or on gills was easily measured by our
technique.
1991). From our viewpoint, the explanation needed is why Ca
and Cd bound better to the gills than did Cu. As presented above,
active uptake of Ca and Cd at high affinity sites can explain then-
results as well as ours. Radiotracers such as 109Cd can be used at
very low concentrations of total metal (e.g., 0.003 uM Cd, about
a 10X lower concentration than we used), but at such low Cd
concentrations exposure times necessary for adequate gill counts
need to be about one day (Wicklund Glynn et al. 1992). There-
fore, graphite furnace AAS is a good method to measure metal
uptake at gills, because it allows both low metal exposure
concentrations and short exposure duration.
Significant morphological changes in fish gills occur at
>10ngCd •!*-', even in synthetic soft water (Karlsson-
Norrgren et al. 1985). Low concentrations of Cd do not lead to
general increases in gill permeability to ions (increased leakiness
due to displacement of Ca from cell membranes; Verbost et al.
1989). Therefore, the ameliorative effects of Ca against low
concentrations of Cd are probably due to specific Ca-Cd
antagonism, not to a general reduction in gill permeability
2674 i
(Pratap et al. 1989; Verbost et al. 1989; Fig. 2). Reduced Ca
influx (Reid and McDonald 1988) can be explained by inhibition
of basolateral Ca2+-ATPase. This inhibition results in increased
intracellular Ca, with a resultant reduction in Ca entry at cell
surfaces (Verbost et al. 1989), and may eventually lead to cell death.
Our dissolved organic carbon (DOC) results indicate, as a first
approximation, that source and size fraction of DOC do not
determine the protective effect of DOC against Cu binding to,
gills. If DOC concentration was S4.8 mg-Lr1, Cu was
prevented from binding to minnow gills (Fig. 3). This result
agrees with those of Cabaniss and Shuman (1988b), who found
that source may be less important in explaining variations in Cu
binding by dissolved organic matter than are chemical factors
such as pH, alkalinity, and ionic strength. Cadmium binds to
DOC about 10X less well than does Cu (Morel 1983; Alberts
and Giesy 1983; Fig. 4). Estimates of Cu-DOC binding vary
considerably, with log K values ranging from 6 to 11 for water
with pH > 6 (Morel 1983). Our own estimates of Cu and Cd
binding constants to DOC are made in our companion paper
(Playleetal. 1993).
Our experiments were run for 2-3 h, long enough for Cd and
Cu accumulation on gills to stabilize (Fig. 1; Playle et al. 1992).
Measured metal accumulation is probably a combination of
surface-adsorbed metal, and subsequent metal transfer across
gill surfaces. This process of adsorption followed by entry is
assumed to occur for metal uptake by algae (Morel et al. 1991)
and is likely the process at fish gills. Verbost et al. (1989) used
an EDTA rinse to remove surface-bound l09Cd and 45Ca, and
found that both tracers were associated with the inside of the gill
epithelium. No differentiation between surface-bound and
interior metal was made in our study. Our premise was that any
process affecting initial adsorption at the gills — mainly
competition or metal complexation—would reduce initial metal
adsorption and therefore accumulation and toxicity. Results of
Skowronski et al. (1992) indicated that similar competition andj
complexation effects (by K* and Cl~, respectively) are seen iij[
either 5- or 90-min Cd exposures, even though approximately
20% more Cd accumulated in their cyanobacterium during the
longer exposure.
Metal binding proteins such as metallothionein may afford
some protection against Cu over time, but production of
metallothionein was likely not a factor in our short experiments.
Can J Fish. Aquat. Sci.. Vol 50. 1993
-------
flHI DPM
-, 300
200
gill DPM
100
250
0.1 j/M 0.05//M 0.1 5 25
"C-citrate Cd //M cold plus "C-citrate; 0.05 pM Cd
FIG. 10. Gill Cd (grey bars) and disintegrations per minute (DPM, clear bars) of gill samples from fish exposed for 2-3 h in soft water
to combinations of '4C-citrate, "cold" citrate, and Cd (6 ng • L"1 = 0.05 uM). Wider bars at left represent mean values ±95% CI
(n = 9); otherwise each bar represents one gill sample, with order preserved left to right in each treatment There were no additional
counts due to 14C-citrate.
In any case, the gill is the first organ affected by waterbome Cu;
induction of metallothionein-like proteins will only have
protective effects once gills have regained normal regulatory
function (Lauren and McDonald 1985). Longer-term studies
would indicate if acclimation to Cu and Cd occurs through
changes in metal-gill binding characteristics (Reid et al. 1991),
changes that could lead to reduced metal concentrations on the
gills (McDonald et al. 1991).
Interpretations of our results are based on the premise that
some ligands can outcompete gills for metals, whereas other
ligands that form weak metal complexes will not. Implicit in our
analyses and calculations is the assumption that metal-ligand
complexes do not bind to gills. Otherwise, measured gill metal
concentrations may actually represent metal-ligand accumu-
lation on gills. There is reasonable agreement that EDTA and
NTA do not pass through cell membranes easily (Jackson and
Morgan 1978; Harrison et al. 1984; Part and Wikmark 1984; Nor
and Cheng 1986; Daly et al. 1990; Block and Part 1992), due
both to their negative charges and relatively large size. Our
present data also gave no indication of I4C-EDTA entry into fish
gills (Fig. 8), although entry of l4C-anthracene was easily
measured (Fig. 9).
The argument becomes less clear when ligands such as citric
acid are used. Citric acid is a small molecule and is a substrate
in the tricarboxylic acid cycle. Studies with a number of
organisms have shown that Cu and Cd accumulation or toxicity
is not reduced as much as expected when small, weak ligands
such as citric acid or glycine are used to complex the metals.
Competition between ligand and organism for the metal ion, as
well as entry (and toxicity) of the small metal-ligand complexes
have been discussed in most of these studies, with conclusions
favouring ligand competition (Zitko and Carson 1976;
Knezovich et al. 1981; Harrison et al. 1984), metal-ligand entry
(Guy and Ross Kean 1980; Borgmann and Ralph 1983; Part and
Wikmark 1984), or a combination of both (Giesy et al. 1977;
Poldoski 1979; Laegreid et al. 1983; Nor and Cheng 1986; Daly
et al. 1990; Femandez-Pinas et al. 1991). Block and Part (1986)
and Block et al. (1991) even showed that l09Cd transfer through
perfused gills was enhanced when complexed with xanthate, due
to the lipid-soluble nature of the complex formed. Some
confusion has undoubtedly occurred because ion selective
electrodes, used frequently to measure free metal ions, do not
take into account metal that is only weakly bound to a particular
Can. J. Fish. Aquat. Set.. Vol. 50, 1993
ligand and would therefore act as a supply of free metal to
another ligand (e.g., gill) with a higher binding constant
(Florence 1977; Giesy et al. 1977).
In the absence of any knowledge of the relative strengths of
metal-gill and metal-ligand complexes, competition between
the gill and ligands for free metal, or uptake of small
metal-ligand complexes, are both valid arguments for metal
accumulation or toxicity in the presence of weak ligands.
However, once it is determined how little free metal is needed
for significant metal accumulation on gills (this study), and the
strengths of Cd-gill and Cu-gill binding are calculated (Playle et
al. 1993), the simpler interpretation of metal accumulation in the
presence of weak ligands becomes competition between gills
and ligands for free metal ions. For example, the Cd and citrate
work of PSrt and Wikmark (1984) suggested that the Cd-citrate
complex entered perfused rainbow trout gills, but by their
calculations 0.01-0.04 fiM Cd (of a total concentration of
9.5 p.M) was still free in solution—likely enough free metal to
explain their results through entry of free Cd alone. In our
experiments, just 0.0001-0.03 \iM free Cd was needed to result
in significant accumulation of Cd on gills. As another example,
this time for freshwater shrimp, the reported toxicity of the
Cu-glycine complex and lack of toxicity of the Cu-(glycine)2
complex (Daly et al. 1990) can be explained by the stability
constants of the two complexes. When the concentration of
glycine is high enough, the Cu-(glycine)j complex (log K= 16.0;
MINEQL*) probably out-competes the freshwater shrimp for Cu,
whereas at lower concentrations the weaker Cu-glycine complex
may not Gog K = 8.6).
Our present work gave no indication of l4C-citrate entry along
with Cd (Fig. 10), suggesting that only free metal interacts at the
gills. 14C-citrate itself did not enter the gills, even in the presence
of DMSO (Fig. 9), a strong surfactant, and there was no
indication that l4C-citrate was assimilated and then lost as I4CO2.
Although metal-ligand entry could be occurring at a slow rate
by diffusion, this route of entry does not need to be invoked to
explain metal accumulation on gills. Depending on the relative
strengths of metal-ligand and metal-gill complexes, and ligand
interactions with cations such as H* and Ca2*, different concen-
trations of ligand will be needed to prevent metals from binding
to fish gills (Fig. 5,6,7). Metal-gill and metal-ligand stability
constants, along with ligand concentrations, should therefore be
useful for predicting metal accumulation on gills, and hence
2675
-------
bioavailability and toxicity.
In summary, measurement of actual metal accumulation on
minnow gills has allowed us to determine that Cd binds to gills
as well or better than does Cu, probably because of Cd uptake in
high affinity Ca channels. This result emphasizes that physio-
logical processes at the gill must be considered when predicting
metal accumulation on fish gills. Dissolved organic carbon
protects against Cu binding to gills better than it does against Cd
binding, which is related to generally greater affinity of Cu to
.synthetic ligands, and the strong binding of Cd to fish gills. Thls^
information, and similar information for other metals, should be
useful in predictive models for metal accumulation on gills, and_
.therefore should be useful to predict metal toxicity to fishTCaTcu^"
lation of metal-gill stability constants from our synthetic ligand
results, and their use in a model to calculate metal binding to gills
offish exposed to Cu and Cd in natural lake water, is the subject
of our companion paper (Playle et al. 1993).
Acknowledgements
«
We thank Drs. Chris Wood and Gord McDonald, McMaster Uni-
versity, Hamilton, Ontario, for the use of their graphite furnace for metal
analyses, and Dr. Bruce Greenberg, University of Waterloo, for supply-
ing die 14C-anthracene. This research was funded by an Operating Grant
(No. 8155) from the Natural Sciences and Engineering Research
Council of Canada to D.G. Dixon, by a Research Grant from the Dorset
Research Centre, Ontario Ministry of the Environment, to D.G. Dixon,
and by an E.B. Eastbum Postdoctoral Fellowship to R. Playle.
References
ALBERTS, JJ., AND J.P. GIESY. 1983. Conditional stability constants of trace
metals and naturally occurring humic materials: Application in equilibrium
models and verification with field data, p. 333-348. In R.F. Christman and
E.T. Gjessing [ed] Aquatic and terrestrial humic materials. Ann Arbor
Science, Ann Arbor, MI.
BENTLEY, PJ. 1991. Accumulation of cadmium by channel catfish (Ictalurus
punctatus): influx from environmental solutions. Comp. Biochem. Physiol.
99C: 527-529.
BLOCK, M., AND P. PART. 1986. Increased availability of cadmium to perfused
rainbow trout (Salmo gairdneri. Rich.) gills in the presence of the
complexing agents diethyl dithiocarbamate, ethyl xanthate and isopropyl
xanthate. Aquat. Toxicol. 8: 295-302.
BLOCK, M., AND P. PART. 1992. Uptake of l09Cd by cultured gill epithelial cells
from rainbow trout (Oncorhynchus mykiss). Aquat. Toxicol. 23:137-151.
BLOCK, M.. A. WICKLUND GLYNN. AND P. PART. 1991. Xanthate effects on
cadmium uptake and intracellular distribution in rainbow trout
(Oncorhynchus mykiss) gills. Aquat. Toxicol 20: 267-283.
BOROMANN. U., AND KM. RALPH. 1983. Complexation and toxicity of copper
and the free metal bioassay technique. Water Res. 17:1697-1703.
CABANISS, S.E., AND M.S. SHUMAN. 1988a. Copper binding by dissolved organic
matter. L Suwannee River fulvic acid equilibria. Geochim. Cosmochim.
Acta 52:185-193.
CABANISS, S.E., AND M.S. SHUMAN. 1988b. Copper binding by dissolved organic
matter: II Variation in type and source of organic matter. Geochim.
Cosmochim. Acta 52:195-200.
CAMPBELL, P.G.C., AND P.M. STOKES. 1985. Acidification and toxicity of metals
to aquatic biota. Can. J Fish. Aquat. Sci. 42: 2034-2049.
CUSMANO, R.F., D.F. BRAKKE, AND G A CHAPMAN. 1986. Effects of pH on the
toxicities of cadmium, copper, and zinc to steelhead trout (Salmo gairdneri).
Can. J. Fish. Aquat. Sci. 43: 1497-1503.
', DALY, H.R., I.C. CAMPBELL, AND B.T. HART. 1990. Copper toxicity to Paratya
auitrahensis. I. Influence of nitrilotriacetic acid and glycine. Environ.
Toxicol. Chem. 9. 997-1006.
FERNADEZ-PINAS, F., P. MATEO, AND I. BONILLA. 1991. Binding of cadmium by
cyanobacterial growth media: free ion concentration as a toxicity index to
the cyanobacterium Nostoc UAM 208. Arch. Environ. Contam Toxicol.
21 425-431.
FLEMMINC, C.A., AND J.T. TREVORS. 1989. Copper toxicity and chemistry in the
environment- a review. Water Air Soil Pollut. 44: 143-158
FLORENCE, T.M 1977. Trace metal species in fresh waters. Water Res 11,4
681-687. •
GIESY, J.P. JR., GJ. LEVERSEE, AND D.R. WDJJAMS. 1977. Effects of naturally
occurring aquatic organic fractions on cadmium toxicity to Simocephalus
serrulatus (Daphnidae) and Cambusia affinis (Poeciliidae). Water Res 11 •
1013-1020.
GUY. R.D., AND A. Ross KEAN. 1980. Algae as a chemical speciation monitor -I.
A comparison of algal growth and computer calculated speciation Water
Res. 14:891-899.
HARRISON, SJ., N.W. LEPP. AND D.A. PHDTS. 1984. Uptake of copper by excised
roots IV. Copper uptake from complexed sources. Z. Pflanzenphysiol 113:
445-450.
HUTCHINSON, N J., AND J.B. SPRAOUE. 1986. Toxicity of trace metal mixtures to
American flagfish (Jordanella floridae) in soft, acidic water and impli-
cations for cultural acidification. Can. J. Fish. Aquat Sci. 43:647-655.
HUTCHINSON, N J., AND J.B. SPRAGUE. 1987. Reduced lethality of Al, Zn and Cu
•/mixtures to American flagfish by complexation with humic substances in
acidified soft waters. Environ. Toxicol. Chem. 6:755-765.
JACKSON, G.A., AND JJ. MORGAN. 1978. Trace metal-chelator interactions and
/phytoplankton growth in seawater media: theoretical analysis and com-
* panson with reported observations. Limnol. Oceanogr. 23:268-282.
JOHNSON, M.G. 1991. Temporal trends in metal concentrations in nvers, coastal-
zone waters, and sediments of northern and eastern Georgian Bay. J. Great
Lakes Res. 17:241-254.
KARLSSON-NORRGREN, L». P. RUNN, C. HAUX, AND L. FORLTN. 1985. Cadmium-
induced changes in gill morphology of zebrafish, Brachydanio rerio
(Hamilton-Buchanan), and rainbow trout, Salmo gairdnen Richardson. J.
FishBiol.27:81-95.
KNEZOVICH, If.. F.L, HARRISON, AND J.S. TUCKER. 1981. The influence of
organic chelators on the toxicity of copper to embryos of the Pacific oyster,
Crassostrea gigas. Arch. Environ. Contam. Toxicol. 10:241-249.
LAEORETO. M.. J. ALSTAD. D. KLAVENESS. AND H.M. SETP. 1983. Seasonal vari-
ation of cadmium toxicity toward the alga Salenastrum capncomutum
Pnntz in two lakes of different humus content Environ. Sci. Technol 17:
357-361.
LAUREN, DJ., AND D.G. MCDONALD. 1985. Effects of copper on branchial iono-
regulation in the rainbow trout, Salmo gairdnen Richardson. J. Comp.
Physiol. B 155:635-644.
LAUREN, D J., AND D.G. MCDONALD. 1987. Acclimation to copper by rainbow
trout, Salmo gairdnen. biochemistry. Can. J. Fish. Aquat Sci. 44:105-111.
MCDONALD, D.G., CM. WOOD, R.G. RHEM, M.E. MUELLER. D.R. MOUNT. AND
H.L. BERGMAN. 1991. Nature and time course of acclimation to aluminum
in juvenile brook trout (Salvelinus fontinalis). I. Physiology. Can. J. Fish.
Aquat Sci. 48: 2006-2015
MOREL, F.M.M. 1983. Principles of aquatic chemistry. John Wiley and Sons,
Toronto, ON. 446 p.
MOREL, F.M.M., RJ.M. HUDSON, AND N.M. PRICE. 1991. Limitation of produc-
tivity by trace metals in the sea. Limnol. Oceanogr. 36:1742-1755.
NOR, Y.M., AND H.H. CHENG. 1986. Chemical speciation and bioavailability of
copper, uptake and accumulation by Eichomia Environ. Toxicol. Chem.
5:941-947.
IPAGENKOPF, G.K. 1983. Gill surface interaction model for trace-metal toxicity
to fishes: role of complexation, pH, and water hardness. Environ. Sci.
Technol. 17: 342-347.
PART, P., AND G. WIKMARK. 1984. The influence of some complexing agents
(EOTA and citrate) on the uptake of cadmium in perfused rainbow trout
gills. Aquat. Toxicol. 5:277-289.
PLAYLE, RC, R.W. GENSEMER, AND D.G. DrxoN. 1992. Copper accumulation
on gills of fathead minnows: influence of water hardness, complexation and
, pH of the gill micro-environment Environ. Toxicol. Chem 11:381-391.
PLAYLE, R.C., D.G. DIXON, AND K. BURNISON. 1993. Copper and cadmium
binding to fish gills: estimates of metal-gill stability constants and
modelling of metal accumulation. Can. J. Fish. Aquat. Sci. 50.2678-2687.
POLDOSKJ, J£. 1979. Cadmium bioaccumulation assays. Their relationship to
various ionic equilibria in Lake Superior water. Environ. Sci. Technol. 13.
701-706.
PRATAP, H.B., H. Fu, R.A.C. LOCK, AND S.E WENDELAAR BONGA. 1989. Effec^
of waterborne and dietary cadmium on plasma ions of the teleost OreoW
chromis mossambicus in relation to water calcium levels Arch Environ.
Contam. Toxicol 18: 568-575.
REID, S.D., AND D.G. MCDONALD 1988. Effects of cadmium, copper, and low
pH on ion fluxes in the rainbow trout, Salmo gairdneri. Can J.Fish Aquat
Sci 45 244-253.
REID, S D., AND D.G MCDONALD. 1991. Metal binding activity of the gills of
rainbow trout (Oncorhynchus mykiss). Can J Fish Aquat Sci. 48:
1061-1068
2676
Can J. Fish Aquat Sci, Vol 50, 1993
-------
REID, S.D., D.G. MCDONALD, AND R.R. RHEM. 1991. Acclimation to sublethal
aluminum: modifications of metal-gill surface interactions of juvenile
rainbow trout (Oncorhynchus mykiss). Can. J. Fish. AquaL Sci. 48:
1996-2005.
SCHECKER. W. 1991. MINEQL* version 2.1. Environmental Research Software.
SMKISS, K., AND M.G. TAYLOR..1989. Metal fluxes across the membranes of
aquatic organisms. Rev. AquaL Sci. 1:173-188.
SKOWRONSKL T., S. SZUBINSKA, M. JAKUBOWSKL AND B. PAWLDC. 1992. Cad-
mium availability to the cyanobacterium Synechocystis aquatilis in
solutions containing chloride. Environ. Pollm. 76:163-167.
SPRAGUE.J.B.1987.Effectsofcadmiumonfreshwaterfish.p. 136-169./nJ.O.'
Nriagu and JJ3. Sprague [ed] Cadmium in the environmenL John Wiley
and Sons, Toronto, ON.
SPRY. DJ., AND J.G. WIENER. 1991. Metal bioavailability and toxicity to fish in
low-alkalinity lakes: A critical review. Environ. Pollut 71:243-304.
VERBOST, P.M., J. VAN Roou, G. Fine. RA.C. LOCK, AND SJE. WENDELAAR
BONCA. 1989. The movement of cadmium through freshwater trout
branchial epithelium and its interference with calcium transport. J. Exp.
Biol. 145:185-197.
WIOOUND, A., AND P. RUNN. 1988. Calcium effects on cadmium uptake, redistri-
bution, and elimination in minnows, Phoxinus phoxinus9 acclimated to
different calcium concentrations. AquaL Toxicol. 13:109-122.
WK3OUND GLYNN, A., C HAUX, AND C HOGSTRAND. 1992. Chronic toxicity and
metabolism of Cd and Zn in juvenile minnows (Phoxinus phoxinus) exposed
to a Cd and Zn mixture. Can. J. Fish. AquaL Sci. 49:2070-2079.
Zmco, V., AND W.G. CARSON. 1976. A mechanism of the effects of water hard-
ness on the lethality of heavy metals to fish. Chemosphere 5:299-303.
Can. J. Fish. Aquat. Sci.. VoL SO, 1993
2677
-------
Copper and Cadmium Binding to Fish Gills: Estimates of
Metal-Gill Stability Constants and Modelling of Metal
Accumulation
Richard C. Playle1 and D. George Dixon
Department of Biology, University of Waterloo, Waterloo, ON N2L 3C1, Canada
and Kent Burnison
National Water Research Institute, Environment Canada, Burlington, ON L7R 4A6, Canada
Playle, R.C., D.C. Dixon, and K. Burnison. 1993. Copper and cadmium binding to fish gills: estimates of metal-gill
stability constants and modelling of metal accumulation. Can. J. Fish. Aquat Sci. 50:2678-2687.
Fathead minnows (Pimephales promelas) were exposed to 17 jtg Cu • L"1 or 6 ng Cd • L"1 in synthetic soft water
in the presence of competing ligands. Measured gill metal concentrations correlated with free metal ion
concentrations, not with total metal. Langmuir isotherms were used to calculate conditional metal-gill equilibrium
constants and the number of binding sites for each metal. Log Kcu-gm was estimated to be 7.4 and the number of
Cu binding sites on a set of gills (70 mg, wet weight) was -2X10~8 mol (-30 nmol • g wet weight"1). Log Kco^n
was -8.6, and the number of Cd binding sites on minnow gills was -2X10~10 mol (-2 nmol • g wet weight"').
Stability constants for H+ and Ca interactions at metal-gill binding sites and for metal interactions with dissolved
organic carbon (DOC) were estimated using these metal-gill constants. All stability constants were entered into
the M1NEQL+ aquatic chemistry program, to predict metal accumulation on fish gills using metal, DOC, and Ca
concentrations, and water pH. Calculated metal accumulation on gills correlated well with measured gill metal
concentrations and with LCso values. Our approach of inserting biological data into an aquatic chemistry program
is useful for modelling and predicting metal accumulation on gills and therefore toxicity to fish.
Des Tetes-de-boule (Pimephales promelas) ont et£ exposes a du Cu ou du Cd en concentration de 17 |ig • L"1 ou
6 jig • L~', respectivement, dans une eau douce synth&ique, en presence de ligands concurrents. Les
concentrations mesurees de metal sur les branchies etaient en correlation avec celles des ions metalliques libres,
mais pas avec celles du metal total. Des isothermes de Langmuir ont servi a calculer les constantes conditionnelles
d'equilibre mgtal-branchies ainsi que le nombre de sites de fixation dans le cas de chaque metal. Le log Kcu-bnncWe
a et£ evalue a 7,4 et le nombre de sites de fixation du Cu sur une paire de branchies (70 mg poids frais) etait
d'environ 2 X 10~8 mol (-30 nmol • g poids frais-1). Le log KbuvancMe a et£ evalu<§ a environ 8,6, et le nombre de
sites de fixation de Cd sur les branchies etait d'environ 2X10~10 mol (-2 nmol • g poids frais"1). Les constantes
de stability des interactions H* et Ca sur les sites de fixation des metaux avec le carbone organique dissous (COD),
ont ete evaluees a partir de ces constantes metaux-branchies. Toutes les constantes de stability ont et<§ versees
dans le programme de chimie de I'eau M1NEOL+ dont on s'est servi pour prevoir i'accumulation des metaux sur
les branchies a partir des concentrations des metaux, du COD et du Ca ainsi qu'a partir du pH de I'eau.
L'accumulation calculee de metaux sur les branchies etait en bonne correlation avec la concentration de metaux
mesuree sur les branchies ainsi qu'avec les valeurs de CLso. Notre methode d'insertion de donnees biologiques
dans un programe de chimie de I'eau constitue une facon utile de modeliser et de preVoir I'accumulation des
metaux sur les branchies, done de la toxicit£ de ces metaux pour les poissons.
Received September 18, 1992
Accepted May 18, 1993
(JB635)
Recu lei 8 septembre 1992
Accept le 18 mai 1993
It has long been realized that metal-organism stability
constants would be useful in interpreting and predicting the
toxicity of waterbome metals to aquatic organisms
(Biesinger and Christensen 1972; Zitko and Carson 1976). To
date, conditional stability constants for Cu have been determined
for algae (Xue and Sigg 1990) and for complexing agents
released by algae (Van Den Berg et al. 1979; Xue and Sigg 1990)
and zooplankton (Fish and Morel 1983). These Cu stability
'Author to whom correspondence should be addressed. Current
address: Department of Biology, Wilfrid Laurier University,
75 University Ave. West, Waterloo, ON N2L 3C5, Canada.
2678
constants fall in the range 107-10n, indicating that 50% of the
Cu binding sites would be occupied a^ aqueous Cu concen-
trations of 10~7-10~n M.
For fish, Reid and McDonald (1991) determined metal-gill
stability constants for La, Cu, Cd, and Ca, using excised gills of
rainbow trout Their values, lO^-lO3-5, were determined using
radiolabelled metals in mM metal solutions. These values are
low compared to those for algae, most likely because of the
relatively high metal concentrations used. The conditional nature
of stability constants makes them dependent on the metal con-
centrations used during their determination: high metal
Con. J. Fish. Aquat ScL Vol 50.1993 A
-------
concentrations yield low stability constants because even low
affinity sites are able to bind excess metals. At low metal
concentrations, only high affinity sites are able to bind the
limited amount of metal, yielding higher stability constants.
Strength of metal binding is not the sole determinant of metal
interactions with aquatic organisms. Dissolved organic carbon
(DOC), Ca concentration, and pH are important modifiers of
metal bioaccumulation and toxicity in freshwater organisms
(Van et al. 1990; Meador 1991; Playle et al. 1992, 1993). Dis-
solved organic carbon can complex metals, rendering them
unavailable to interact with an organism, and calcium competes
with metals for binding sites, as does H*. In addition, water pH
determines the proportion of total, unbound metal that is in the
cationic form, generally thought to be the toxic species of metals
(Morel 1983; Pagenkopf 1983).
The objective of the present study was to determine metal-gill
stability constants for low concentrations of Cu and Cd, and to
use these constants, plus metal-DOC, Ca-gill, and H-gill
constants, in a computer model to predict metal accumulation on
fish gills. This approach has been used for algae (Xue and Sigg
1990) but has not yet been applied to fish. Metal accumulation
on minnow gills in the presence of complexing ligands such as
EDTA were determined in our preceding paper (Playle et al.
1993). We used the lowest concentrations of Cu and Cd (17 and
6 u,g • L~', respectively) that were feasible to yield measurable
metal deposition on fish gills, using graphite furnace atomic
absorption spectroscopy. The gill accumulation data plus cal-
culated free metal concentrations in the water were used to
determine the metal-gill stability constants. Free metal con-
centrations were calculated using MINEQL* (Schecher 1991),
an aquatic chemistry computer program based on equilibrium
constants. Finally, our estimates of the metal-gill stability con-
stants were assessed by comparing calculated metal accumu-
lation on gills with measured gill metal concentrations of
minnows exposed to Cu and Cd in water from a series of Ontario
lakes.
Materials and Methods
Ligand Experiments
Detailed descriptions of the ligand experiments are given in
the preceding paper (Playle et al. 1993) and the results are
summarized in Tables 1 and 2 of the present paper. In brief,
acclimated adult fathead minnows (Pimephales promelas) were
exposed to 17 ngCu • IT1 (0.27 |iM)or6 (JigCd- L'1 (0.05 p.M)
in synthetic soft water (Na, Ca -50 jiM) at pH 6.2, 19°C. Fish
were killed at the end of the 2-3-h exposures, their whole gill'
baskets were removed, weighed, then digested in 1 N H2SO4 for
8 h at 80°C. The supernatant of the gill digests was diluted with
deionized water, then analyzed for Cu and Cd using graphite
furnace atomic absorption spectroscopy. Experiments were run
in the presence or absence of freeze-dried dissolved organic
carbon (DOC, 0-8 mg- L~') or the synthetic ligands EDTA,
nitrilotriacetic acid (NTA), ethylenediamine (EN), and citric,
glutamic, oxalic, and salicylic acids (0-2500
Calculation of Gill Stability Constants
We used Langmuir isotherms to calculate Cu- and Cd-gill
stability constants (KCa.fM, #cd-g,ii). In order to make these calcu-
lations, estimates of free Cu2+ and Cd2* concentrations in the
synthetic soft water, with and without added ligands, were made
Can. J Fish. Aquat Sci , Vol SO, 1993
using the MINEQL* program. Input values were 0.27 jiM Cu,
0.05 jiM Cd, 70 p,M Na, 35 nM Ca, fixed pH 6.2, 19°C. and
concentrations of the added ligands (0-2500 pM). Solution
ionic strength was calculated by MINEQL* (I~ 1 X 10~4), and
metal-ligand log K values were supplied by the computer
program (Table 1,2). We ignored changes in water pH near the
gills (Playle and Wood 1989; Lin and Randall 1990) because
both Cu and Cd exist in the cation form at pH < 7; small pH
changes near pH 6 were therefore inconsequential with respect
to speciation of these two metals.
Because we had information on Ca and H* interactions at Cu
and Cd gill-binding sites (Playle et al. 1992, 1993), it was
possible to estimate Kc^n and AT^i once #a>-giii and AQ,^,, were
determined. These estimates were made in the following manner.
Metal-ligand equilibrium constants are defined as the
concentration of the metal-ligand complex divided by the
product of free metal and free ligand, once the reaction is
complete. For example, Cd and gill interactions can be defined
as
(i)
[Cd-gill]
[Cd] • [gill]
Calcium interactions at the same sites on the gill can be defined
by
[Ca-gill]
-
en
(2)
Rearranging Equation 1, and substituting for [gill] in Equation 2,
we get
_ [Ca-gill]
-
[Ca] .
This equation can be used to calculate K^.^ if we know either
how many gill sites, or the proportion of sites, that are occupied
by Ca and Cd. If half the gill sites are occupied by Ca, and half
by Cd, then Equation 3 simplifies to
(4)
[Ca]
In the situation where the same amount of Cd accumulates on
the gills in the presence of Ca as in its absence (i.e., the amount
of Ca in solution does not significantly out-compete Cd for the
gill binding sites) then
(5) Kt
Ca-g.ll
ICa]
If the Ca in solution prevents all Cd from accumulating on the
gills, then
(6)
[Ca]
Hydrogen ion interactions at metal binding sites can be cal-
culated in the same manner, by substituting H* for Ca. Copper
2679
-------
TABLE 1. Gill Cu concentrations (above background) of fathead minnows held for 2-3 h in soft water
containing 0.27 fiM total Cu, plus ligands. Summarized from Playle et al. (1992, 1993). EN =
ethylenediamine.
Ligand FreeCu2* GillCu
OiM) (fiM; MINEQL* calculation) (fimol • g~ ' wet tissue)
EDTA
NTA
EN
Citrate
Oxalate
Glutamate
Salicylate
No ligands
No ligands,
0.15 '
0.25
0.30
0.25
2.5
0.25
2.5
25
0.25
25
25
0.25
2.5
25
0.25
25
25
0.25
2.5
25
100
250
no added Cu
0.112
0.019
0.0002
0.030
0.0002
0.214
0.080
0.006
0.230
0.122
0.007
0.246
0.199
0.044
0.247
0.210
0.101
0.252
0.249
0.225
0.176
0.127
0.253
0.028
0.019
0.000
0.000
0.000
0.000
0.036
0.017
0.013
0.027
0.020
0.008
0.018
0.024
0.008
0.038
0.034
0.020
0.021
0.024
0.027
0.020
0.012
0.026
0.000
logtfCu-ligand
(from MINEQL*)
18.8
13.1
10.5
7.3
5.1
8.3
10.6
—
-
—
TABLE 2. Gill Cd concentrations (above background) of fathead minnows held for 2-3 h in soft water
containing 0.05 uJM total Cd, plus ligands. Summarized from Playle et al. (1993). EN=ethylenediamine.
Ligand FreeCd2* >
(H-M) (nM; MINEQL* calculation)
EDTA 0.05
0.10
0.25
0.5
5.0
NTA 0.1
0.5
EN 0.05
Citrate 0.1
250
2500
Oxalate 0.05
Glutamate 0.05
No ligands
No ligands,
no added Cd
14.4
4.83
1.41
0.64
0.05
48.7
46.4
49.3
49.1
2.60
0.24
49.3
49.3
49.3
0.99
GillCd logKCd-ligand
(nmol -g 'wet tissue) (from MINEQL4)
2.0
2.4
0.4
0.0
0.2
2.3
1.6
25
2.7
1.2
0.4
2.1
2.2
2.0
0.0
16.3
9.4
5.6
53
3.4
4.8
—
—
can be substituted for Cd. Implicit in these calculations is the
assumption that there is enough metal and time for the gill sites
to become saturated. In practice, the metal concentrations and
exposure times used in our experiments adequately fulfilled
these assumptions (Playle et al. 1992,1993).
2680
Lakewater Exposures and Toxicity Tests
All stability constants were entered into the MINEQL* aquatic
chemistry program, to calculate metal accumulation on fish gills.
To test the model, we wanted to measure metal accumulation on
Can. J. Fish Aquat. Sci., Vol SO. 1993
-------
4 r
3
1
3
u
1
O.06 _
•i
c
0.04 J
0.02
0.1 0.2
freeCu" (pM)
0.3
FIG. 1. Plot of measured gill Cu against free Cu2* concentrations. Gill
Cu concentrations were from fathead minnows exposed 2-3 h in
synthetic soft water with 17 jtg Cu • L~' (0.27 uM) and 0-250 oM of
synthetic ligands (data from Playle et al. 1992, 1993). Free Cu2* was
calculated using MINEQL* and known concentrations of ligands and
ions (Table 1). Gill Cu varied directly with free Cu2* (r= 0.858,
P < 0.001). The equation of the line was y = 6.23x+0.86 (y in jig Cu • g
tissue"1) ory = 0.098*+ 0.014 (y in u,mol' g tissue"1).
gills when minnows—were exposed to 17 (ig Cu • L"' and
6 jig Cd • L"1 in natural soft water, as opposed to synthetic soft
water. In addition, we wanted to correlate metal accumulation
on gills with metal toxicity.
Water was collected from five lakes NE of Peterborough,
Ontario, in June, 1991. The lakes were: Salmon L., Brooks Bay
and Sharps Bay of Jack L., Anstruther L., Loucks L., and Wolf L.
(all approximately N44°42', W78°13')- Twenty-Lstainless steel
containers (Spartenburg Steel Products, Spartenburg, SC) were
used to collect the surface water. These containers were then
pressurized with N2 gas and the water was filtered through
1.0 u,m glass fibre filters (A/E, 142 mm, Gelman Sciences Inc.,
Ann Arbor, MI) to remove algae. The filtered water was stored
in 40-L stainless steel containers (Simgo Inc., Toronto, ON),
brought back to Waterloo, transferred to 4-L polyethylene con-
tainers, and stored at 4°C. These water samples were used for
combined Cu and Cd metal exposures to adult fathead minnows,
conducted in the same manner as described for the synthetic
softwater experiments. A synthetic soft water plus a 10-uM
(total) mixture of glutamic, citric, and oxalic acids was also
used (3.3 u,M of each), to mimic a 5 mg DOC • L"1 solution
(Campbell and Stokes 1985).
Toxicity tests using a geometric series of concentrations of the
Cu and Cd mixture were run in the lake waters, using juvenile
fathead minnows. These fish were <1 mo old, and had been held
in soft water for 1 wk before the experiment. During acclimation
they were fed live brine shrimp and fine powdered fish food
twice daily. Ninety-six h LCSOs were determined using duplicate
exposures of five fish per 100-mL solution in un-aerated poly-
ethylene urine cups. Eighteen juveniles (blotted dry) weighed
individually yielded an average weight of 1.6 ± 0.5 mg. LC50
values were calculated using a trimmed Spearman-Karber analysis.
Results
Calculation of Metal-Gill Stability Constants
Copper
Previously, we measured Cu accumulation on gills of fathead
minnows exposed to 17 jig Cu • L"' (0.27 u,M) in synthetic soft
water, in the presence or absence of synthetic ligands (Playle
etal. 1992,1993). Because we knew the water chemistry and the
Can J. Fish. Aquat Set.. Vol SO, 1993
0.4
3
1
3
1
0.2
20
40
60
fraeCd" (nM)
FIG. 2. Plot of measured gill Cd against free Cd2* concentrations. Gill
Cd concentrations were from fathead minnows exposed 2-3 h in
synthetic soft water with 6 u,g Cd • L"1 (0.05 jtM) and 0-2500 jiM of
ligands (data from Playle et al. 1993). Free Cd2* was calculated using
MINEQL* and known concentrations of ligands and ions (Table 2). Gill
Cd varied directly with free Cd2* (r = 0.732, P < 0.01). The equation of
the line was y = 0.004* + 0.163 (y in u,g Cd • g tissue"1) ory = 0.03 lr
+ 1.451 (y in nmol -g tissue"').
f
ligand concentrations, we were able to calculate free Cu concen-
trations using the MINEQL* program. Gill Cu concentrations
varied directly with free Cu2*, the major component of free Cu:
gill Cu versus free Cu2* gave a straight line (Fig. 1; r= 0.858,
P < 0.001). Gill Cu concentrations did not correlate with total
Cu, which was constant at 0.27 jiM.
To calculate log K^^n and the number of Cu binding sites on
the gills, a Langmuir plot of Cu adsorption was constructed. Free
Cu2* (micromolar) divided by gill Cu (micromoles Cu per gram
wet tissue, minus background gill Cu, Table 1) was plotted
against free Cu2* (micromolar). The equation for the line was
y = 35.38* + 1.28. The inverse of the slope of the line is the
number of gill binding sites, = 0.03 u,mol • g wet tissue'1. The
inverse of the intercept = AT- (binding site number), therefore
K = 26 L • nmol"1 = 26 X 106 L • mol'1, and log KC^II - 7.4.
Average total gill weight for the minnows was 0.07 g (wet tissue;
Playle et al. 1993), therefore total binding sites for Cu on the gills
of our fathead minnows was about 2X 10~9 mol.
Cadmium
Accumulation of Cd on gills of fathead minnows exposed to
6 u,g Cd • L"1 (0.05 jiM) in the presence or absence of synthetic
ligands (from Playle et al. 1993) was plotted against free Cd2*
concentrations calculated using MINEQL* (Fig. 2). Although
there were not as many data points as there were for Cu, gill Cd
still varied directly with free Cd2* (r = 0.732, P < 0.01). It was
difficult to obtain intermediate values for the curve, because so
little Cd was used in the experiments (e.g., 0-50 nM is just
one-fifth the scale for Cu in Fig. 1).
The log Kcd-gin and the number of binding sites for Cd were
calculated using a Langmuir plot. Gill Cd concentrations above
background (Table 2) were used for these calculations. Th
Langmuir isotherm, calculated in nmoles, yielded the l
y = 0.44* +1.06. The number of gill binding sites was the inverse
of the slope, = 2.27 nmol • g wet tissue"1. From the inverse of
the intercept, K = 0.42 L • nmol"1 = 0.42 X 10' L • mol"', and
log tfcd-gui = 8.6. Average gill weight was 0.07 g wet tissue; thus
the number of gill binding sites for Cd was about 2 X 10~'° mol
per minnow.
2681
-------
TABLE 3. Input data for the metal-gill interaction model. Initial log K values were as calculated in the text.
Final log /rvalues were determined using MINEQL+ to best fit, within the constraints of the initial log K
values, calculated gill metal with measured gill metal concentrations for the synthetic soft water system.
Binding sites
Complex
Initial log K
Final log K
Cu-gill
2 X 10"9
mol per fish
Cd-gill
2X 10" IU mol per fish
DOC
1 mg • L~
1 = 5 X 10~8 mol
Cu-gillcu
H^-gillcu
Ca-gillcu
Cd-gillcd
H*-gillcd
Ca-gillcd
Cu-DOC
Cd-DOC
H*-DOC
7.4
<5.6
<3.5
8.6
5.0
8.4-9.4
6.4-8.4
4.0
7.4
5.4
3.4
8.6
6.7
5.0
9.1
7.4
4.0
Calculation of Ca, H*, and DOC Stability Constants
Calcium and H* can compete with Cu and Cd for binding sites
on gills. By using the log Km^^a values calculated above, and
data presented in Playle et al. (1992,1993), the stability constants
for Ca and H* were estimated.
Copper accumulation on minnow gills was not affected at
pH 4.8 (15.8 n-M H*) compared to pH 6.3 (0.5 p,M H*; Playle
et al. 1992). Inserting ATo,^, = 10™, free Cu = 2.53 X 10'7 M
(MINEQL* calculation), and H*=15.8X10~6 M into
Equation 5 yields log KH.^n < 5.6. Copper accumulation was not
affected by 2 000 jtM Ca (Playle et al. 1992): using Equation 5,
Cu = 2.53 X ID'7 M, Ca=2 X 10'3 M, and ATo,^, = 10™, yields
log Kc^a < 3.5.
Cadmium accumulation on the gills was reduced at pH 4.8
(Playle et al. 1993). Inserting Kc^n = 1086, free Cd =
4.93 X 10'* M (MINEQL* calculation), and H* =
15.8 X 10~6 M into Equation 6 yields log *7H-jiii > 6.1. For Ca,
95 u,M did not reduce Cd accumulation on gills, whereas
1 050 jiM Ca did (Playle et al. 1993). Using Equations 5 and 6,
respectively, log KC+JU < 5.3, but >4.3, the mean of which is
log Kc*fi\ = 5.0.
Five to 6 mg DOC • L"1 prevented Cu accumulation on gills,
by complexing all the 0.27 \iM Co in solution (Playle et al.
1993). Thus, there were at least 0.05 junol binding sites per
mg DOC • L~'. These metal binding sites must have had a higher
affinity for Cu than did the Cu binding sites on the gill, or the
DOC would not have been able to prevent Cu accumulation on
the gills. The log £O,-DOC value f°r the 0.05 |unol of high affinity
binding sites per mg • L~' DOC must therefore be about 1-2 log
units higher than the log ATo,^, = 7.4 value. Our initial estimate
of the stability constant was log KCWJOC = 8.4-9.4, which is in
the mid-range of published values (Van Den Berg and Kramer
1979;Morel 1983).Note that direct attack of the Cu-gill complex
by DOC is unlikely (adjunctive pathway of ligand exchange; see
Hering and Morel 1990). More likely is the disjunctive pathway,
that first involves dissociation of Cu from the Cu-gill complex,
the free Cu then being able to react with DOC in solution.
Cadmium has consistently been reported to bind <10X as well
to DOC than does Cu (Alberts and Giesy 1983; Tuschall and
Brezonik 1983). Thus, an initial estimate of the log KC^DOC value
of 10-100X less than for Cu would be 6.4-8.4. The log K for
humic substances is 3 to 4 (DeWit et al. 1991), therefore we set
log KH&K - 4. We had no information on the interactions of Ca
with DOC.
2682
Model Development
Our purpose in determining metal-gill and other stability
constants was to use these values in a computer model to predict
metal accumulation on fish gills, and thereby predict metal
toxicity. In essence, bringing biology into an aquatic chemistry
program. The model we chose to use was MINEQL* (Schecher
1991), an aquatic chemistry program based on log K stability
constants.
Our initial estimates of log K stability constants and the
number of metal binding sites are summarized in Table 3. To
input these data into the MINEQL* program, three null com-
ponents were defined (gillcu, gillcd, and DOC). We simulated
two different sites for Cu and Cd binding on the gills (gillcu,
gillcd) on the basis of our previous results (Playle et al. 1993)
that indicated, for the low metal concentrations we used, no
competition between Cu and Cd for gill binding sites. Metal, Ca,
and H* interactions with gillcu, gillcd, and DOC were defined in
the program as 1:1 reactions (Table 3).
We used Na = 70 jiM, Ca = 35 ftM, Cd = 0.05 pM, Cu =
0.27 u,M, fixed pH 6.2, and 19°C for water chemistry parameters
of the synthetic soft water..For log K values for which we just
had upper or lower values, or a range, we chose final log K values
after optimizing to best fit the observed data in Playle et al. (1992,
1993). In essence, the model was able to mimic the greater effect
of H* and Ca on Cd binding to gills, and the greater effect of
DOC on Cu binding to gills. Of note in Table 3 are the final
log ATmeui-Doc values, that we optimized to log KCU-DOC = 9.1 (in
the model, Cu binds to DOC SOX better than it binds to the gill),
and log KCMXX = 7.4 (in the model, Cd binds to DOC 50X less
well than does Cu, and Cd binds to DOC 16X less well than Cd
binds to gills). We chose 1 mg • L~' DOC = 0.05 fimole binding
sites, to allow enough binding sites for all the Cu and Cd in
solution at 6 mg DOC • L"1 (e.g., no competition by the metals
for the sites). In general, as long as a ligand is in excess relative
to metals in solution, metals do not compete with each other
(Morel 1983).
Metal Accumulation and Toxicity to Fish Exposed to Cu and
Cd in Soft Lake Water
Of the natural lake waters that we used in the Cu and Cd
exposures, Salmon L., and Brooks and Sharps bays contained
more Ca, more DOC, and had higher pH than did Anstriither,
Loucks, and Wolf lakes (Table 4). The synthetic soft water plus
a 10 ftM (total) addition of glutamic, citric, and oxalic acids had
water chemistry similar to that of synthetic soft water without
Can. J Fish. Aquat. Sci.. Vol 50. 1993
-------
TABLE 4. Lake water and synthetic soft water characteristics, and juvenile fathead minnow 96 h LCSOs. Water characteristics were measured on
water used for the metal deposition and LC50 experiments, except for conductivity, which was from a 1988 survey. SSW = synthetic soft waterj
+ ligands = plus 10 jtM (total) glutamic, citric, and oxalic acids. DOC = mg • L , Ca and Na = uM, conductivity (Cond.) = p,mho • cm~', and all
metals = u.g • L . Mean ± SEM (n) if n > 2. For LC50, values in parentheses = 95% CI.
Water
Salmon Lake
Brooks Bay
Sharps Bay
Anstruther Lake
Loucks Lake
Wolf Lake
SSW + ligands
SSW
pH
8.05±0.01
(12)
7.80±0.01
(12)
7.90±0.02
(12)
7.10±0.03
(12)
7.04±0.01
(12)
6.91 ±0.03
(12)
6.60±0.03
(8)
6.69±0.01
(20)
DOC
6.4±0.1
(3)
9.8
(2)
7.9
(2)
7.2±0.6
(3)
6.6
(2)
6.0±0.2
(3)
1.8
(2)
25
(2)
',
Ca
695±20
(4)
445±10
(4)
590±20
(4)
128±2
(4)
92±2
(4)
94±1
(5)
36±1
(3)
37±1
(H)
Na
34±6
(4)
76±9
(4)
90±32
(4)
39±5
(4)
37±4
(4)
48±I2
(4)
87±23
(3)
91±11
(H)
Cond.
85
—
78
24
22
19
—
—
Cu
2.3
3.1
6.6
4.1
1.4
0.8
—
2.1 ±0.3
(32)
Cd
0.1
0.1
<0.1
0.1
0.3
03
—
<0.1
(33)
Al
10.7
12.8
65
12.8
13.3
9.6
—
4.4
Fe
<50
<50
<50
<50
<50
<50
_ _
<50
Zn
12.0
4.0
25
1.5
4.2
1.6
__
2.8±05
(10)
96hLC50
Cu Cd
>126 >32
102 25.3
(78-133) (19.1-33.4)
115 27.8
(85-156) (19.8-39.0)
50.0 10.4
(31.9-54.4) (7.8-14.0)
27.8 6.8
(18.4-^2.0) (4.2-11.1)
31.2 6.9
(21.9-44.6) (4.0-11.9)
10.3 1.1
(6.8-15.6) (0.4-3.0)
9.4 1.0
(7.6-11.5) (0.6-1.7)
added ligands. Minnow exposures in both these synthetic soft
waters were at pH -6.7, to match the pH of the softer of the
natural waters.
Copper and Cd 96-h LC50 values for juvenile fathead
minnows exposed in the various waters are given in Table 4. In
general, LCSOs for the Cu and Cd mixtures were lowest (metals
were most toxic) in softer water. Metals were most toxic in the
synthetic soft water, and were only slightly less toxic in the
presence of the 10 jiM ligand mixture. Metals were least toxic
in water from Salmon L., with too few mortalities to accurately
determine the LCSO (Table 4).
Cadmium accumulation on gills of adult minnows exposed to
6 M-g Cd • L~' and 17 jig Cu • L~' was highest in Wolf L. water
(significantly above background, and not significantly different
from metal-exposed fish in synthetic soft water alone; Fig. 3),
and lowest in Salmon L. water (not significantly above back-
ground). Fish exposed to the lake waters without added metals
did not have gill Cd concentrations significantly above back-
ground (0.18 ± 0.02 u,g Cd • g wet tissue'1; ±95% CI, n = 17).
Fish exposed to Cd and Cu in synthetic soft water plus 10 u,M
(total) glutamic, citric, and oxalic acids had 0.64 ± 0.23 (6) u,g
Cd • g wet tissue'1, and fish exposed to the metals in synthetic
sqft water without added ligands had gill Cd concentrations of
0.54 ± 0.05 (9) u,g Cd • g. These exposures were run at pH 6.6
and 6.7, respectively; their gill Cd concentrations were not
significantly different from those of fish exposed in synthetic
soft water at pH 6.2.
Copper accumulation on gills of adult minnows was never
above background in fish held in lake waters supplemented with
Cu and Cd (Fig. 4), probably because DOC exceeded 5 mg • L"1
in all lake waters (see Playle et al. 1993). In lake water without
added metals, gill Cu was 0.80 ± 0.08 (17), also not significantly
above background. Minnows exposed to Cu and Cd in synthetic
soft water plus the 10 M-M ligand mixture (pH 6.6) had gill Cu
concentrations of 2.58 ±0.18 (6) jig Cd • g wet tissue"1, while
fish exposed to the metals in synthetic soft water without added
ligands (pH 6.7) had gill Cu concentrations of 2.16 ± 0.30(9) (ig
Cu • g"1. Neither concentration was significantly different from
pH
1//M)
DOC (mg-f)
B
SH
W
Lake
FIG. 3. Gill Cd concentrations of fathead minnows exposed 2-3 h in Cd-
and Cu-supplemented lake water. Cadmium deposition on gills was
highest in the softest water. S=Salmon L.. B=Brooks Bay, SH = Sharps
Bay, A=Anstruther L., L= Loucks L.. and W=Wolf L. Vertical lines on
bars = 95% CI; n = 6 for each bar. Dashed horizontal lines represent the
95% CI of gill Cd for minnows held in synthetic soft water in the absence
of added metals (n = 38). The grey horizontal band represents the 95%
CI of gill Cd for fish held 2-3 h in synthetic soft water with
6 jig Cd • L"1 (n - 45). *, **, *** = significantly below metal-exposed
gills (grey band), and +, -H-, -m- = significantly above background gill
metal concentration (dashed lines) for P< 0.05, <0.01, <0.001,
respectively.
Can. J. Fish Aguat. Sci.. Vol SO. 1993
gills of minnows exposed to Cu in synthetic soft water at pH 6.2.
Model Testing
Final values for the log K stability constants (Table 3) were
used to assess the usefulness of the model in predicting gill metal
concentrations and metal toxicity. In the model we used the
2683
-------
PH
Ca(f/M)
(mg-L")
Lake
FIG. 4. Gill Cu concentrations of fathead minnows exposed 2-3 h in Cd-
and Cu-supplemented lake water. There was no significant accumu-
lation of Cu on the gills, likely because DOC was a 6 rag • L~l in each
lake water. Dashed horizontal lines represent the 95% CI of gill Cu for
minnows held in synthetic soft water alone (n = 38). The grey horizontal,
band represents the 95% CI of gill Cu for fish held 2-3 h in synthetic
soft water with 17 ng Cu • L~' (n = 45). Lake names and other details
are given in caption of Fig. 3.
FIG. 5. Measured and modelled gill Cd, for fathead minnows exposed
2-3 h to 6 p.g Cd • L~' and 17 jig Cu • L~' in natural soft waters and
synthetic soft water (SSW) with or without 10 j»M added ligands (at
pH 6.7). Lakes are defined in the caption of Fig. 3. Gill Cd is expressed
as a percentage of the mean gill Cd concentration (100%) for fish held
in synthetic soft water plus Cd at pH 6.2 (grey bar, ±95% CI).
Background gill Cd = 0% (+95% CI, dashed line). Calculated gill Cd
correlated well with measured gill Cd (see text for details). Measured
gill Cd values are given with their 95% CI.
DOC, Ca, and Na values (molar), and pH (fixed) for the eight
waters listed in Table 4. The model mimicked a system open to
atmospheric CO2. Although we did not measure CO32~ in the lake
waters, for the model we assumed its concentration was equal to
that of Ca (Table 4), which yields approximately the same values
as if the more rigorous Henderson-Hasselbalch equation were
used to calculate HCO3~. For the comparison of measured and
calculated gill Cd and Cu, measured values from Rg. 3 and 4
were converted to percentages, with background gill metal = 0%,
and gill metal concentrations of fish exposed to metals in
synthetic soft water (at pH 6.2) = 100%. The calculated gill metal
concentration in synthetic soft water (pH 6.2, assuming back-
ground DOC = 1 mg • L~') was used as the model 100% value.
2684
Lake Uganda
FIG. 6. Measured and modelled gill Cu for fathead minnows exposed
2-3 h to 17 fig Cu • L"1 and 6 fig Cd • L~' in natural soft waters and
in synthetic soft water (SSW) with or without added ligands. Lakes are
defined in the caption of Fig. 3. Gill Cu is expressed as a percentage of
the mean gill Cu concentration (100%) for fish held in synthetic soft
water plus Cu (grey bars). Background gill Cu = 0% (dashed line).
Calculated and measured gill Cu correlated well (see text for details).
Modelled gill Cd agreed well with our measured values
(Rg. 5), with the largest discrepancy for Sharps Bay. The cor-
relation coefficient between measured and modelled gill Cd was
0.9008 (P < 0.01). Measured gill Cd versus the juvenile fathead
minnow LCSOs for Cd (Table 4) yielded a correlation coefficient
of -0.897 (P < 0.01). Modelled gill Cd versus the LC50 values
yielded a better correlation (r = -0.960; P < 0.001). These data
indicate that model-calculated gill Cd concentrations could rea-
sonably be used to predict acute Cd toxicity to fish.
Measured and modelled gill Cu also agreed well (Rg. 6,
r=0.926, P < 0.001), in spite of the fact that none of the gill Cu
concentrations from the natural lakewaters was above back
ground (Rg. 4). Measured gill Cu versus the log of the juvenile
fathead minnow LCSOs for Cu yielded r = -0.830 (P < 0.05; gill
Cu vs. linear LCSOs was not significant, r = -0.662, P > 0.05)
Modelled gill Cu had a slightly better correlation with the log
LCSOs (r = -0.895, P < 0.01; the correlation with linear LC50s>
was r = -0.727, P < 0.05). As was the case with Cd, modelled
gill Cu concentration was a good indicator of acute Cu toxicity.
Discussion
In the present study, we have better defined interactions of
Cu and Cd with fish gills, as influenced by dissolved organic
carbon (DOC), pH, and Ca. We used Langmuir isotherms to
estimate conditional metal-gill stability constants. These were
log Xcu^iii = 7.4 and log tfca-giii= 8.6. There were about 10X more
of the weaker Cu sites on the gills (~2 X 10~9 mol per fish) than
Cd sites (~2 X lO'10 mol per fish).
Interactions of Cu and Cd with the gill and with DOC, and
competing reactions with Ca and FT, are illustrated conceptually
in Fig. 7, along with their log K values. Cadmium binds to the
gills better than does Cu, and is more affected by competition
from Ca and H*. These interactions are likely a consequence of
active Cd uptake through high affinity Ca channels in addition
to general surface binding. Cadmium reduces Ca uptake, mainly
through effects on basolateral Ca2+-ATPase activity (Verbost
et al. 1989). The data of Reid and McDonald (1988) support the
existence of higher affinity binding sites for Cd compared to Cu.
Calcium competition reduces Cd uptake at gills, and therefore
Can. J. Fish Aquat Set. Vol. 50,1993
-------
the subsequent distribution to the rest of the fish body (Wicklund
and Runn 1988). Transfer of metals through the basolateral
membrane is probably by diffusion (Verbost et al. 1989).
Copper, on the other hand, appears to bind to weaker sites on
the gills, and possibly also enters through ion channels (Fig. 7),
but binds more strongly to DOC than does Cd (log XCU-DOC=9.1;
log AGHJOC = 7.4). The binding order for DOC reflects relatively
strong binding of Cu to ligands. The effect of low concentrations
of Cu at the gills is an inhibition of Na* influx, mainly through
effects on Na+-K+-ATPase (Lauren and McDonald 1987).
Relatively more inhibition of this effect was seen by carbonate
complexation than by Ca (Laurdn and McDonald 1986), which
agrees with our results of less protective effects of Ca on Cu
accumulation, compared to effects on gill Cd (Playle et al. 1992,
1993). Note that in the conceptual model illustrated in Fig. 7 the
gill mucus layer is ignored. Gill mucus probably binds metals
and slows metal access to the gill (Part and Lock 1983).
However, once a metal exceeds the complexing and sloughing
ability of mucus, the metal will become available to react at the
gill epihelium; this is the situation illustrated in Fig. 7.
There are few log Kmeui-giii values with which to compare our
values. Reid and McDonald (1991) found log K values of about
3.5,3.0,3.0, and 2.4 for La, Cd, Ca, and Cu, respectively. These
values are low, probably because they used mM concentrations
of metals in their experiments instead of environmental,
sub-micromolar concentrations. Log AT values are sensitive to the
metal concentration at which the experiments are run, and it is
difficult to extrapolate from a stability constant obtained at high
metakligand ratio to a stability constant for a low metakligand
ratio (Perdue and Lytle 1983). However, Reid and McDonald
(1991) found that Cd bound to gills better than did Cu, which
agrees with our results. Xue and Sigg (1990) determined
log Kca^gte = 8.8 (at pH 6.0), which is ~ 10X higher than ours for
fish at about the same pH. These workers were successful in model-
ling Cu binding to algae using a computer program based on
log K values, their input log Kcu^pc values, and the number of
binding sites (2 jxmol • g dry algae"') they determined for algae.
We inserted our metal-gill, H+-gill, Ca-gill, and metal-DOC
stability constants (Fig. 7; Table 3) into MINEQL* (Schecher
1991). The model adequately reflected measured metal accumu-
lation on the gills (Fig. 5,6). In addition to Cu and Cd bound to
gills, the MINEQL+ output included Ca and H+ bound to gill sites
(an indication of competition for the sites), unoccupied gill sites,
free metal in solution (an indication of the pool of metal available
to bind at the sites), and metal bound to DOC (representing the
extent of metal complexation). This simple model considered
only 1:1 metal-ligand interactions, and ignored Ca interactions
at metal binding sites on DOC. At the low metal concentrations
we used, 1:1 binding would be expected: Cabaniss and Shuman
(1988a) showed that Cu-DOC interactions were >90% 1:1
binding. Not incorporating Ca-DOC interactions probably
would have had little effect on the model, because both Cu and
Cd would bind much more strongly to DOC than would Ca
(Kerndorff and Schnitzer 1980; Cabaniss and Shuman 1988a;
Daly et al. 1990). It should be noted that our stability constants
represent average conditional stability constants, akin to con-
tinuous multiligand binding models (e.g., Perdue and Lytle 1983;
Grimm et al. 1991) or "quasiparticle" models (Sposito 1981).
Stability constants change with solution chemistry (pH, ionic
strength) as well as with metal and ligand concentration (Perdue
and Lytle 1983; Cabaniss and Shuman 1988a). Van Den Berg
and Kramer (1979) and Morel (1983) determined that there is
Ca
Cd**
9.1
Cu
/
Cu—7.4
DOC
4.0
H
Cd
H*
7.4
X
DOC
X
4.0
Na* \
Ca**
FIG. 7. Conceptual illustration of Cu, Cd, Ca, and H* interactions at fish
gills and with dissolved organic carbon (DOC). The numbers are the log
K conditional stability constants for the interactions. In water (left), Cu
and Cd may become complexed by DOC, and also face competition
from Ca and H+ for negative binding sites on gill surfaces. As well as
binding to the general gill surface, Cu may enter the gill through low
affinity channels. Copper disrupts Na* uptake at the basolateral Na
pump. The stability constant for Cd is higher than for Cu: if surface
binding was solely responsible for metal reactions at the gills, Cd would
be expected to bind less strongly than Cd. Uptake of Cd through high
affinity Ca channels is likely the explanation of the higher-than-^
expected /fcd-gm. Also illustrated is the inhibition of the basolateral Ca'
pump by Cd.
about a one log unit decrease in stability constants as pH
decreases by one unit. However, we used the approach of Grimm
et al. (1991) and Cabaniss and Shuman (1988a) where changes
in stability constants are considered to be due solely to H+
competition. That is, the use of one metal-gill stability constant,
with another H+ stability constant for competition, as opposed to
different metal-gill constants for each pH. In spite of the
qualifications and simplifications used in the model, relative
amounts of Cu and Cd deposition on the gills were reasonably
well predicted (Fig. 5, 6). Copper was complexed better by DOC
than was Cd, and Cd binding to the gills was affected more by
Ca than was Cu.
We calculated approximately 0.05 iuno] metal binding sites
per mg DOC. Reported number of Cu binding sites for DOC vary
from about 10 u,mol per mg DOC (Cabaniss and Shuman 1988a)
to 0.2 jtmol per mg DOC (Hering and Morel 1990). Our low
number of binding sites is indicative of the low metal concen-
trations we used in our experiments. The sites were relatively
strong metal binding sites (log KCU-DOC = 9.1, log KGI-DOC = 7.4),
which may represent metals binding with carboxyl groups (Taga
et al. 1991). Sunda and Hanson (1979) determined log K^ DOC =
9.0 for river water at pH 5.95, very similar to our Cu result.
Hering and Morel (1990) used 0.05' \iM Cu concentrations
their experiments, and found slightly stronger Cu binding
humic acid (log K^™** = 10.1). Holm and Curtiss (1990)
used 0.05, 0.2, and 10 u,M Cu in their experiments, and found
log K= 10. 1, 8.5, and 5.5, respectively, for natural organic matter
in ground water, demonstrating the dependence of stability
constants on the metal concentration used in their determination.
Grimm et al. (1991) determined (mean) log KO.-DOC = 4.2-4.9, a
Can J Fish. Aquat Sci., Vol 50. 1993
2685
-------
much lower stability constant than those listed above, but they
used experimental Cu concentrations S5 u.M. Holm and Curtiss
(1990) determined that complexes with natural organic matter
should dominate Cu speciation at the low Cu concentrations they
used. In agreement, our measured and modelled gill Cu values
(Fig. 6) were mainly influenced by DOC concentration.
Cadmium has been shown to bind less well than Cu to DOC
and fulvic and humic acids (e.g., Florence 1977; Kemdorff and
Schnitzer 1980; Alberts and Giesy 1983; Lund et al. 1990; Sahu
and Banerjee 1990). This binding order is related to general
metal-ligand binding strength. If Cu binds better to DOC and
humic acids, then Cu toxicity would be expected to decrease in
response to organic material more than does Cd toxicity (e.g.,
Winner 1985,1986 for Cu, Cd toxicity to Daphniapulex). In our
work, freeze-dried DOC protected against Cu deposition on
minnow gills at concentrations £4.8 mg-L~' (Playle et al.
1993), and insignificant Cu accumulation was observed in
natural lake waters containing £6 mg DOC • L~' (Fig. 6),
whereas Cd still bound to fish gills in the presence of & 6 mg
DOC • L~' (Fig. 5). Because of these binding results, we suggest
that toxicity of the metal mixtures in soft waters (Table 4) was
due mainly to Cd.
In the model we ignored differences in DOC source or com-
position, because concentration of DOC adequately determined
gill Cu concentrations (Playle et al. 1993). Although DOC from
different sources can bind or detoxify metals to varying degrees
(e.g., Nor and Cheng 1986; Sahu and Banerjee 1990), DOC from
similar sources (i.e. same size of watershed and vegetation types)
generally binds metals in a similar manner (Oliver et al. 1983;
Cabaniss and Shuman 1988b). Source of DOC may not be as
important as pH, alkalinity, or ionic strength in determining
metal binding to DOC (Cabaniss and Shuman 19885), and these
workers suggested that experimental effort should be expended
in defining these effects, not on documenting variations in DOC
binding properties with season or location.
Synthetic soft water plus 10 p-M (total) citric, glutamic, and
oxalic acids did not reduce Cu or Cd accumulation on fathead
minnow gills, nor did it alter metal toxicity (Table 4). This
mixture was meant to be a synthetic DOC analogue, similar to
but simplified from those used by Sposito (1981) and by
Campbell and Stokes (198S). Campbell and Stokes (1985) found
that a 22 (iM DOC analogue complexed -70% of a 6 u,g
Cu • L~' solution (pH 6). With 3X that amount of Cu in our
experiments, and half the ligand concentration, only about 12%
of the Cu would be expected to be complexed, not enough to
reduce Cu accumulation on gills. However, the model did
calculate some reduction in free Cu (to about 60% of total;
Fig. 6). It is not surprising that the 10 pM (total) ligand solution
did not reduce Cu or Cd bound to gills or reduce metal toxicity,
and it does not adequately reflect 5 mg DOC • L"' in our system.
The model should be expanded to include other metals. It may
be possible to calculate stability constants from previously
published ligand work, but some metal-gill stability constants
have already been published. For example, Wilkinson et al.
(1990) used salmonid data to estimate log KAH>n - 6.5. Con-
ditional stability constants for organic material and Al are
available: log ArA,.Doc = 6.2 (Urban et al. 1990), log K^™** =
7.8 (Shuman 1992). Constants are also available for other metals
such as Mn (log /^M-DOC = 3.8; Urban et al. 1990), Fe
(log KF..DOC = 9.4; Urban et al. 1990), and Co (log Kc^vmKmA =
5-6; Van Loon et al. 1992). Weaker complexing metals such as
Ni and Zn would likely bind to DOC about 10 X less well than -
does Cu (Morel 1983). Lin and Benjamin (1992) indicated some
of the confounding effects of ligand addition to a multiple-metal
system. A ligand may complex a metal that is otherwise bound
to a (gill) binding site, which would free that site (e.g., increase
free gill sites) to which another metal may bind. Models such as
ours will be useful to predict complex interactions of this type.
In an expanded model, competition for gill binding sites by Ca
and H* need to be considered for each metal using appropriate
log K values. Metals will be less toxic in hard water, and very
acidic conditions will keep metals off gills. Of course, H* itself
can be disruptive to gills. Simulations using our model indicated
that H* would protect against metal binding at gills before it
would displace metal from DOC. By the time enough metal was
displaced from DOC for the metal to interact at the gills (pH < 4),
H* itself would likely become toxic. This point illustrates the
usefulness of the model in integrating multiple effects such as
speciation, competition, and complexation, a situation where
models relying on correlation (Yan et al. 1990; Meador 1991)
are often inappropriate.
Output from the present model is toxicant bound to the gills,
the target organ, which then must be translated into toxicity to
the fish. This last step requires the most study. Together, the
model output would give an indication of all types of competition
for gill binding sites, of which Ca interactions can be considered
beneficial, and most others as potentially toxic. In our lake water
experiments, measured and calculated,gill'Cu and Cd concen-
trations did correlate with toxicity, but it is not known how much
metal accumulation is necessary to affect fish survival on a
long-term basis. Presumably, a constant accumulation (gill
"dose") will be toxic, such as was found for whole-body Cd
accumulation in Hyalella azteca (Borgmann et al. 1991).
In summary, measurement of metal deposition on minnow
gills has allowed us to estimate conditional stability constants of
Cu and Cd binding to gills, the number of metal binding sites on
the gills, plus interactions with Ca, H*. and DOC. This
information, and similar information for other metals, will be •
useful in predictive models for metal accumulation on gills, and
will therefore be useful to predict metal toxicity to fish.
Acknowledgements
We thank Drs. Chris Wood and Gord McDonald, McMaster Uni-
versity, Hamilton, Ontario, for the use of their graphite furnace for metal
analyses. We thank Dr. David Lean for help collecting lake water
samples. This research was funded by an Operating Grant (No. 8155)
from the Natural Sciences and Engineering Research Council of Canada
to D.G. Dixon, by a Research Grant from the Dorset Research Centre,
Ontario Ministry of the Environment, to D.G. Dixon, and by an E.B.
Eastbum Postdoctoral Fellowship to R. Playle.
References
ALBERTS, JJ., AND J.P. GIESY. 1983. Conditional stability constants of trace
metals and naturally occurring humic materials: Application in equilibrium
models and verification with field data, p. 333-348. In R F. Christman and
E.T. Gjessing [ed ] Aquatic and terrestrial humic materials Ann Arbor
Science, Ann Arbor, MI.
BIESINGER, K.E., AND G.M CHWSTENSEN. 1972 Effects of various metals on
survival, growth, reproduction, and metabolism ofDaphnia magna J. Fish,
Res Board Can 29. 1691-1700.
' BORGMANN,U..W.P NORWOOD, ANDI.M.BABIRAD 1991.Relationshipbetween
chronic toxicity and bioaccumulation of cadmium in Hyalella azteca. Can.
J Fish. Aquat. Sci 48: 1055-1060.
CABANISS, S E, AND M.S. SHUMAN. 1988a Copper binding by dissolved organic
matter I Suwannee River fulvic acid equilibria. Geochim. Cosmochim
Acta52. 185-193.
2686
Can J. Fish. Aquat. Sci. Vol 50. 1993
-------
CABANISS. S.E.. AND M.S. SHUMAN. 1988b. Copper binding by dissolved organic
matter: n. Variation in type and source of organic matter. Geochim.
Cosmochim. Acta 52:195-200.
CAMPBELL. P.G.C.. AND PM. STOKES. 1985. Acidification and toxicity of metals
to aquatic biota. Can. J Fish. Aquat ScL 42: 2034-2049.
DALY, HJt. MJ. JONES, B.T. HART. AND LC. CAMPBELL. 1990 Copper toxicity
to Paratya australiensis: JJJ. Influence of dissolved organic matter. Environ.
Toxicol. Chem. 9:1013-1018.
DEWTT,J.CM..MM.NEDERLOF,WJI VANRIEMSDUK.ANDL.K.KOOPAL. 1991.
Determination of H* and metal ion affinity distributions for humic
substances. Water Air Soil PoUuL 57-58:339-349.
FISH, W., AND F.M.M MOREL. 1983. Characterization of organic copper-
complexing agents released by Daphnia magnet. Can. J. Fish. AquaL ScL
40:1270-1277.
FLORENCE, T.M. 1977. Trace metal species in fresh waters. Water Research 11:
681-687.
GRIMM. DM., L.V. AZARRAGA, LA. CARRHRA, AND W. SUSETYO. 1991. Con-
tinuous multiligand distribution model used to predict the stability constant
of Cu(TD metal complexation with humic material from fluorescence
quenching data. Environ. Sci. Technol. 25: 1427-1431.
BERING, J.G., AND FMM. MOREL. 1990. Kinetics of trace metal complexation:
ligand-exchange reactions. Environ. Sci. Technol. 24:242-252.
HOLM. T.R., AND CD. CURTISS m. 1990. Copper complexation by natural
organic matter in ground water, p. 508-518. In D.C. Melchior and RJ-
Bassett [ed.] Chemical modeling of aqueous systems EL American Chemical
Society, Washington, DC.
KERNDORFF, H., AND M. SCHNITZER. 1980. Sorption of metals on humic acid.
Geochim. Cosmochim. Acta 44:1701-1708.
LAUREN, DJ, AND DO. MCDONALD. 1986. Influence of water hardness, pH, and
alkalinity on the mechanisms of copper toxicity in juvenile rainbow trout,
Salmo gairdneri. Can. J. Fish. AquaL Sci. 43: 1488-1496.
LAUREN, D J.. AND D.G. MCDONALD. 1987. Acclimation to copper by rainbow
trout. Salmo gairdneri: biochemistry. Can. J. Fish. AquaL Sci. 44:105-111.
LIN, C.-F.,_AND MM. BENJAMIN. 1992. The effects of strongly complexing
ligands on the adsorptive partitioning of metal ions. Water Res. 26:
397-407.
LIN. H., AND DJ. RANDALL. 1990. The effect of varying water pH on the
acidification of expired water in rainbow trout J. Exp. BioL 149:149-160.
LUND, W_ LA. HELBAK, AND RM. SHIP. 1990. Studies of the complexation
properties of aquatic bumic material by differential pulse polarography. ScL
Total Environ. 92:269-281.
MEADOR, J.P. 1991. The interaction of pH. dissolved organic carbon, and total
copper in the determination of ionic copper and toxicity. AquaL Toxicol.
19:13-32.
MOREL. FMM. 1983. Principles of aquatic chemistry. John Wiley and Sons,
Toronto. ON. 446 p.
NOR, Y.M., AND H.H. CHENG. 1986. Chemical speciation and bioavailability of
copper uptake and accumulation by Eichorma. Environ. Toxicol. Chem.
5: 941-947.
OLIVER, B.C., EM. THURMAN, AND RJ-. MALCOLM. 1983. The contribution of
lumic substances to the acidity of colored natural waters. Geochim.
Cosmochim. Acta 47:2031-2035.
PAGENKOPF, O.K. 1983. Gill surface interaction model for trace-metal toxicity
to fishes: role of complexation, pH, and water hardness. Environ. Sci.
Technol. 17: 342-347.
PART. P., AND R.A.C. LOCK. 1983. Diffusion of calcium, cadmium and mercury
in a mucous solution from rainbow trout. Comp. Biochem. Physiol. 76C:
259-263.
PERDUE, EM., AND C.R. LYTLE. 1983. A critical examination of metal-ligand
complexation models: application to defined multiligand mixtures,
p. 295-313. In R.F. Christman and Egil T. Gjessing [ed.] Aquatic and
terrestrial humic materials. Ann Arbor Science, Ann Arbor, ML
PLAYLE, R.C., AND CM WOOD. 1989. Water chemistry changes in the gill micro-
environment of rainbow trout: experimental observations and theory. J.
Comp. Physiol. B 159.527-537.
PLAYLE, R.C., R.W. GENSEMER, AND D.G. DKON. 1992. Copper accumulation
on gills of fathead minnows: influence of water hardness, complexation and
pH of the gill micro-environment Environ. Toxicol. Chem. 11:381-391.
PLAYLE, R.C, D.G. DIXON, AND K. BURNISON. 1993. Copper and cadmium
binding to fish gills: modification by dissolved organic carbon and synthetic
ligands. Can. J. Fish. Aquat ScL 50: 2667-2677.
REID, SD., AND D.G. MCDONALD. 1988. Effects of cadmium, copper, and low
pH on ion fluxes in the rainbow trout. Salmo gairdneri. Can. J. Fish. AquaL
ScL 45: 244-253.
REID, SD., AND D.G. MCDONALD. 1991. Metal binding activity of the gills of
rainbow trout (Oncorhynchiu mykiss). Can. J. Fish. Aquat Sci. 48:
1061-1068.
SAHU, S., AND D.K. BANERIEE. 1990. Complexation properties of typical soil and
peat humic acids with copper(Q) and cadmiumCTI). Int J. Environ. Anal.
Chem. 42:35-44.
SCHECHER,W. 1991. MINEQL* version 2.1. Environmental Research Software.
SHUMAN, M.S. 1992. Dissociation pathways and species distribution of alumi-
num bound to an aquatic fabric acid. Environ. Sci. Technol. 26:593-598.
SFOSITO, G. 1981. Trace metals in contaminated waters. Environ. ScL Technol.
15:396-403.
SUNDA, W.G.. AND PJ. HANSON. 1979. Chemical speciation of copper in river
water effect of total copper, pH, carbonate, and dissolved organic matter,
p. 147-180. In EA. Jenne [ed.) Chemical modeling in aqueous systems.
ACS Symposium Series 93, American Chemical Society, Washington, DC
TAGA, M., S. TANAKA, AND M. FUKUSHMA. 1991. Evaluation of copper(II)-
binding ability of humic acids in peat using sulphopropyl-sephadex C-25
cation exchanger. Anal. Chim. Acta 244:281-287.
TUSCHALL. J Jt JR.. AND PI. BREZONK. 1983. Complexation of heavy metals by
aquatic humus: a comparative study of five analytical methods, p. 275-294.
In RJ-. Christman and ET. Gjessing [ed] Aquatic and terrestrial humic
materials. Ann Arbor Science, Ann Arbor, ML
URBAN, NJt, E. GORHAM, JJC. UNDERWOOD, F.B. MARTIN. AND J.G. OGDEN m
1990 Geochemical processes controlling concentrations of ALFe, and Mn
in Nova Scotia lakes. LimnoL Oceanogr. 35:1516-1534.
VAN DEN BERG. CM.G., AND JJL KRAMER. 1979. Conditional stability constants
for copper ions with ligands in natural waters, p. 115-132. In EA. Jenne'
,[ed.) Chemical modeling in aqueous systems. American Chemical Society,
Washington, DC.
VAN DEN BERG, CM.G., P.T.S. WONG, AND Y.K. CHAU. 1979. Measurement of
complexing materials excreted from algae and their ability to ameliorate
copper toxicity. J. Fish. Res. Board Can. 36:901-905.
VAN LOON, LJt, S. GRANAGHER, AND H. HARDUF. 1992. Equilibrium dialysis-
ligand exchange: a novel method for determining conditional stability
constants of radionuclide-humic acid complexes. Anal. Chim. Acta 268:
235-246.
VERBOST. PM., J. VAN Roou, G. FUK. RA.C LOCK. AND S.E. WENDELAAR
BONGA. 1989. The movement of cadmium through freshwater trout
branchial epithelium and its interference with calcium transport J. Exp.
Biol. 145: 185-197.
WICKLUND. A., AND P. RUNN. 1988. Calcium effects on cadmium uptake, redistri-
bution, and elimination in minnows, Phoxinus phoxinus, acclimated to
different calcium concentrations. Aquat Toxicol. 13:109-122.
WILKINSON, KJ., P.G.C. CAMPBELL, AND P. COUTURE. 1990. Effect of fluoride
complexation on aluminum toxicity towards juvenile Atlantic salmon
(Salmo solar). Can. J. Fish. Aquat Sci. 47:1446-1452.
WINNER, R.W. 1985. Bioaccumulation and toxicity of copperas affected by inter-
actions between humic acid and water hardness. Water Res. 19:449-455.
WINNER, R.W 1986. Interactive effects of water hardness and humic acid on the
chronic toxicity of cadmium to Daphnia pulex. Aquat Toxicol. 8:281-293.
XUE, H.-B., AND L, SIGG. 1990. Binding of Cu(D) to algae in a metal buffer.
Water Res. 24:1129-1136.
YAM, NJ>., GI_ MAOOE, AND PJ. DILLON. 1990. Cadmium concentrations of
crustacean zooplankton of acidified and nonacidified Canadian shield lakes.
Environ. Sci. Technol. 24:1367-1372.
Zmco, V., AND W.G. CARSON. 1976 A mechanism of the effects of water
hardness on the lethality of heavy metals to fish. Chemosphere 5:299-303.
Can. J. Fish. Aquat. Set., Vol 50, 1993
2687
-------
Pergamon
Computers A Ctotcieaca Vol. 20. No. 6. pp. 973-102}, 1994
Copyright O 1994 Elsevier Science Lid
0098-3004(94)E0016-M Printed in Great Briuin. All rights reserved
0098-3004/94 $6700 +0.00
WHAM—A CHEMICAL EQUILIBRIUM MODEL AND
COMPUTER CODE FOR WATERS, SEDIMENTS,
AND SOILS INCORPORATING A DISCRETE
SITE/ELECTROSTATIC MODEL OF ION-BINDING
BY HUMIC SUBSTANCES
/
E. TIPPING
Institute of Freshwater Ecology. The < Ferry House, Ambleside. Cumbria LA22 OLP, UJC
(Received and accepted I December 1993)
Abstract—WHAM (Windermere Humic Aqueous Model) is designed to calculate equilibrium chemical
spedation in surface and ground waters, sediments, and soils. The model is suitable especially Tor problems
where the chemical spedation is dominated by organic matter (humic substances). WHAM combines
Humic Ion-Binding Model V with a simple inorganic spedation code for aqueous solutions. Precipitation
of aluminum and iron oxides, cation-exchange on an idealized day mineral, and adsorption-desorption
reactions of fulvic add also are taken into account. The importance of ion accumulation in the diffuse
layers surrounding the humic molecules is emphasized. Modd calculations are performed with a BASIC
computer code running on a Personal Computer.
Key Words: Chemical equilibrium, Chemical spedation, Humic substances, Ion-binding, Sediments, Soils,
Waters.
INTRODUCTION
Humic substances are recognized to play an import-
ant part in the chemical spedation of waters, sedi-
ments, and soils, but describing their interactions
with protons, metal ions, and other species under
environmental conditions has proved difficult in the
absence of an appropriate spedation code. Until
recently, the main attempt to do so has been with the
program GEOCHEM (Mattigod and Sposito, 1979),
rcpresentingdissolved humic matterbva collection of
low molecular weight organic compounds with
known .binding properties. Although useful as a guide
to the interactions, this approach is not entirely
satisfactory, because it ipiores the tnaeraiqnie nature
ofjulvic andJhumig adds. More recently, a "com-
petitive Gaussian" model for fulvic add has been
introduced into the MINTEQ code (Allison and
Perdue. 1994), .but the available database for metal
interactions is'limited (Susetyo and others, 1991). No
codes seem to be available to deal comprehensively
with solid-phase organic matter.
The Windermere Humic Aqueous Model
(WHAM), presented here, can be applied to equi-
librium speciation problems involving waters, sedi-
ments, and soils, with humic substances present in
dissolved or paniculate form. It follows from the
development and partial validation of Humic Ion-
Binding Model V (Tipping and Hurley, 1992; Tipping
1993a. 1993b, 1993c). Although Model V could be
. combined with existing inorganic spedation codes in
order to solve chemical equilibrium problems, a new
organic-inorganic combination that takes into
account the solving demands of Model V in the
overall scheme was developed, rather than attempting
to incorporate Model V into an existing code. This is
important especially for concentrated systems (sedi-
ments and soils) where the accumulation of tons in
the diffuse layer surrounding the humic molecules
takes on major importance. Moreover, a purpose-
built model should be better-suited for incorporation
into dynamic models of natural aquatic, soil, and
catchment systems.
MODEL DESCRIPTION
WHAM is a combination of several submodels.
These are Humic Ion-Binding Model V (Tipping and
Hurley, 1992; Tipping, 1993a, 1993b). and models of
inorganic solution chemistry, precipitation of alumi-
num and iron oxyhydroxides, cation-exchange on a
representative clay, and the adsorption-desorption
reactions of fulvic adds (Tipping and Woof, 1990,
1991). The interrelationships among these submodels
are depicted in Figure 1.
Model V: specific binding by humic substances
Humic compounds are represented by hypothetical
size-homogeneous, rigid, molecules, which carry pro-
ton-dissociating groups that can bind metal ions
973
-------
974
E. TIPPING
either singly or as bidentate pairs. The interactions
are described in terms of intrinsic equilibrium con-
stants and electrostatic terms. The former refer to the
(usually hypothetical) situation where the humic sub-
stances have zero electrical charge. The latter take
into account the influence on binding of the variable
humic charge; binding strength is enhanced-when the
metal species and the humic molecule carry opposite
charges, and diminished when the charges, have the
same sign. The electrostatic terms take the general
form e ~1™r where K> is the electrostatic interaction
' factor (dependent upon ionic strength), z is the charge
on the combining ion in question, and Z is the humic
charge. For example, the dissociation of a proton by
the following reaction
RAH'^RA'-' + H* m (1)
is characterized by the variable equilibrium quotient,
K(Z):
(2)
RAH*
T
DISCRETE SITES H A DIFFUSE LAYER
and J[-7UAV, balancing
lion ^l L\\t\ r3 counlerions
binding
dissociation
-2wzZ
e
Donnan
AI(OH)3
Fe(OH);
K
so
K
so.
BULK SOLUTION
A + B + C ^=^ ABC
Donnan
(K.ri)
CLAY
counlerions
balancing
ZCLAY
/ , , A Donnan
' ~2wzZ ' ,„ *
iMntr 6 \^se\)
binding and
dissociation
balancing
counlerions
DISCRETE SITES FA DIFFUSE IAYER
SOIL/SEDIMENT
SOLIDS
Figure 1. Functional relationships in WHAM. FA » fulvic acid, HA «= humic acid. Species in bulk
solution are in equilibrium with fulvic and humic discrete sites, and also with counterion species in diffuse
layers. In version of model for waters (WHAM-W), aluminum and iron (oxy)hydroxides and day cation
exchanger are omitted. In version for soils and sediments (WHAM-S), some of fulvic acid can be present
in both solid and dissolved forms. Humic and fulvic acids, and clay cation-exchanger, together with their
diffuse layers, have net zero charge, as does bulk solution.
-------
Equilibrium chemical spcciation by WHAM
975
Table I. Default parameter values Tor fulvic acid (FA) and humic acid
(HA) used in WHAM-W Version 1.0 and WHAM-S Version 1.0.
diffuse layer. See text for details.
"A
«i
PKA
-PKA
ApKA
•ApK*i>
f •
J*
P^MHt
KZ
/«.
y
f
K,
mol type A groups per g
mol type B groups per g
central pK for type A groups
central pK Tor type groups
range of type A pK's^ ft 7
range of type B pK's
defines w by w — /*logw/
proximity factor
molecular weight
molecular radius (run)
in terms of pKiou
charge factor for DL volume
factor for max DL volume
distribution parameter
adsorption parameter
adsorption parameter
FA
4.73 x 10-'
2J7 x 10-'
3.26
9.64
3.34
5.52
-103
0.4
1500
0.8
3.96pKnHn.
10'
0.25
1
10*
1
HA
3.29 x 10-'
1.65 x 10-'
4.02
8.55
1.78
3.43
-374
OJ
15.000
1.72
3pIW-3
101
0.25
—
—
—
where A^ is the intrinsic constant Because w is
positive and Z nearly always negative, Equation (2)
shows that the greater (more negative) is the humic
charge, the less does reaction (1) proceed, because of
the tendency of the humic charge to hinder proton
dissociation.
The proton-binding groups of the humic sub-
stances are heterogeneous, having a range of intrinsic
pK values. Two types of acid group are
distinguished, denoted by A and B. Within each type
there are four different groups, present in equal
amounts, the pK values of which are described in
terms of a median value, pKA or pK,, and a factor,
ApKA or ApK(, that defines the range of the values.
For example, the four type A groups have pK values
given by
PK«(0 = PKA +
(3)
where / «= 1,2,3, or 4. The humic content of type
B sites (n,) is fixed at one-half the content of the
type A (nA) sites. This rather strict arrangement of
proton-binding groups is selected with a view to
simplifying the description of the binding of other
ions.
Competitive metal binding takes place at single
proton-binding sites (monodentate), and at bidentate
sites formed by pairs of proton-dissociating sites. The
extent to which bidentate binding can take place is
constrained by the proximity factor (/",»), which
defines, on the basis of molecular geometry, the
likelihood of pairs of proton-binding groups being
close enough to form bidentate sites. Thus a proxim-
ity factor of 0.4 means that 40% of the proton-
binding groups are sufficiently close to pair up.
Because there are 8 proton-dissociating sites, 36
different bidentate sites can be formed theoretically,
but in order to speed computations, only 12 represen-
tative pairings — present in equal abundances — are
allowed in the model. Monodentate metal binding is
described with intrinsic equilibrium constants for the
metal-proton exchange reaction:
/WHz+Ml* = /ZXMtz*'-|)+H*. (4)
The negative logarithms of these intrinsic constants
are denoted by pKMHA and pKMHa for type A and
type B sites respectively. The smaller is the value of
pKMH, the stronger is the binding of the metal species
in question. The appropriate pK values are added
together to obtain the value for bidentate binding.
The model permits the binding of the first hydrolysis
product (e.g. CuOH* in the situation of CuJ* ) as well
as the parent species, the pK^i values for the first
hydrolysis product being assumed equal to those of
the parent. This assumption has not been tested fully,
but has been shown to account adequately for copper
binding results (Tipping and Hurley, 1992; Tipping.
1993a). It has been discovered that pK^u is
sufficiently well-correlated with pKutu that only one
of them needs to be specified to describe the binding
of a given species. The selected parameter is pKMHA,
because-this has been determined to exert the most
influence when fitting published data. sets. Values of
PKMHA h*ve been derived from literature data for
aluminum, alkaline earths, heavy metals, europium,
and actinides (Tipping and Hurley. 1992; Tipping
1993a, 1993b, I993c), and also have been estimated
from linear free energy relationships (Tipping,
I993c). In all, parameter values for 36 parent species
are currently available.
To implement Model V, it is necessary to know the
following parameters, which characterize the proton-
binding characteristics of the humic substance in
question; nA, na, pKA, pK», ApKA, ApKB, P (empiri-
cal parameter defining w, by w «= P log,,/, where / is
ionic strength), and/,,. Sets of parameter values have
been published for different humic materials (Tipping
and Hurley, 1992; Tipping, I993a, 1993b) and sets of
"best-average" default values have been arrived at for
fulvic and humic acids (Table 1). Generally, different
lf»-
CAGCO »/«-*•
-------
976
E. TIPPIHG
samples of the same type of humic material (i.e. fulvic
or humic acid) give similar sets of these Model V
parameters (Tipping and Hurley, 1992; Tipping,
1993b). In addition, a pKMHA value is required for
each species that can bind at the discrete sites, as
explained in the previous paragraph; default pKHHA
values are included in the SSED Database (see
Appendix 1).
Model V: nonspecific binding by counterion accumu-
lation
The modification of binding strength at specific
sites by electrostatic effects (discussed previously) is
the result of the accumulation of an excess of counter-
ions in a diffuse layer adjacent to the molecular
surface, and this nonspecific process can be con-
sidered to contribute to the total binding. For waters,
such accumulation is important in using Model V to
account for observed competition effects involving
species, such as alkaline earths, that show weak
specific binding (Tipping, 1993a). For more concen-
trated systems such as sediments and soils, where
humic concentrations can be high (tens of grams per
liter), counterion accumulation thus may account for
the greatest part of some species. Conventional elec-
trostatic theory (Gouy-Chapman, Debye-Huckel)
for charged interfaces in aqueous solutions regards
the diffuse layer as being, in principle, infinite in
extent, and ionic distributions are described using
Boltzmann statistics (Tanford, 1961; Hiemenz, 1977).
This description is the basis for the modifying terms
mentioned in the previous section, of the form e ~2nZ.
To compute counterion accumulation (and co-ion
deficit) with this theory, it is necessary to integrate for
the entire volume of the diffuse layer, or, more
practically, for some truncated version thereof (Bolt,
1982). By this approach, the modification of specific
binding and counterion accumulation would be
described with the same model. However, the inte-
grations required are demanding computationally,
and the theory does not take readily into account
interference among the diffuse layers of neighboring
particles, which must be significant in concentrated
systems such as sediments and soils. In the situation
of humic substances, heterogeneity among molecules
adds an extra problem in applying the fundamental
theory. For these reasons, it is considered justified to
adopt a much simplified description of counterion
accumulation in Model V and WHAM.
•' In Model V, the diffuse layer is regarded as a zone
of defined thickness around the humic molecules and
average concentrations of counterions within that
zone are considered. Co-ions are excluded com-
pletely. There is no mathematical connection with the
term modifying interaction at discrete sites. An
approximate thickness of the diffuse layer is taken to
be the ionic-strcngth-dependent Debye-Huckel par-
ameter, K (Hiemenz, 1977). With the thickness
defined, the volumes of the diffuse layers are calcu-
lated from geometrical formulae, assuming the humic
molecules to be spheres. It is necessary to assume
reasonable values for the humic radius and molecular
weight. The volume is given by
where NA, is Avogadro's number, M is humic mol-
ecular weight, and r is the radius of the humic
molecule. With the diffuse layer volume so-defined,
ionic distributions are calculated easily, using Don-
nan expressions [see Eqs. (8) and (9)] and requiring
that the total counterion charge within the diffuse
layer exactly balances the charge resulting from the
humic ionizable groups, modified by specific binding
of protons and metal ions. Also, the bulk solution is
charge-balanced perfectly (Fig. 1).
For dilute systems, especially at higher ionic
strengths, the diffuse layer volumes are small in
comparison with the total. However, this is not the
situation for more concentrated systems at lower
ionic strengths. For example, the diffuse layer volume
of a solution of Smg/1 fulvic acid with an ionic
strength of 0.1 M is calculated to I • 4 x 10~5 liters
per liter of total water. In contrast, iu an organic soil
with a water content of 75%, the humic acid concen-
tration might be 50 g per liter of total water, and the
ionic strength would be low, say 0.001 M. This would
correspond to a diffuse layer volume of ca 10 liters
per liter, an obvious impossibility. Model V attempts .
to handle this problem with an empirical factor,/^, /<>
that restricts the diffuse layer volume. Diffuse layer )
volumes calculated with Equation (5) are designated
maximum volumes, PDHUF and PP-— f for fulvic and
humic acid respectively, and the actual volumes VDf
and FDH are calculated from
'OF1
rDHuF
/DL + foauF +
(6)
The value of fa is set to a number between 0 and 1,
that is the asymptotic maximum total diffuse layer
volume, On the basis of preliminary results with acid
soils, a default value of 0.25 is used for/DL, but this
is not an optimized value. To suppress restrictions on
diffuse layer volume, fa can be set to infinity.
A further difficulty arising with the Model V diffuse
layer is the small requirement for counterions at low
net humic charge, which can lead to the paradoxial
situation in which counterion concentrations are cal-
culated to be lower than bulk solution values. This
has been noted for aluminum-rich soils, where the
specific binding of A1J* can make thejiumic charge
go to near zero. This seems to follow from a violation
of the implicit assumption that the humic charge is
spread evenly over the molecular surface. To take
into account such "discretization" of humic charge
an adjusting factor, AT2, is introduced to force the
diffuse layer volume to diminish as charge decreases.
-------
Equilibrium chemical specialion by WHAM
977
Thus the maximum diffuse layer volumes are
modified as follows;
'Drn.1
(7)
where ZBO< is the modulus of Z. Experience in
' modeling acid soils has suggested a value of 10' Tor
Kz, but as in the situation of/M.. this is not
considered an optimized value.
Although overlapping of diffuse layers is con*
sidered in WHAM, this is restricted to the diffuse
layers of like molecules, that is fulvic acid diffuse
layers overlap only with each other, and those of
humic acid behave likewise. For concentrated sys-
tems, dearly this is artificial, but in view of the
simplifications and approximations involved in
attempting to describe the diffuse layers, it does not
seem justified to elaborate the model further. On a
more practical level, keeping the diffuse layers
separate simplifies the description of fulvic acid
adsorption-desorption reactions.
In published work to date on Model V, and its
predecessor Model IV, counterfoil accumulation has
been assumed to depend only upon countcrion
charge. However, work in progress with soils suggests
that in order to explain major cation distributions
(Na*. Mgu, Ca1*, etc.) it may be necessary to
introduce some selectivity. Although this has yet to
be established firmly, the possibility is allowed in
attempting to present as comprehensive as possible a
picture of the WHAM models. In the nonselective
Donnan model, diffuse layer concentrations of
counterfoil species / are related to solution
concentrations as follows:
«o(0
(8)
where ZmxiCO is the modulus of the charge on species
i, and R is the ratio required for the sum of the
counterion charges to balance the humic charge Z.
Selectivity is introduced by writing;
Cp(0
(9)
where K^(i) is the selectivity coefficient. Only a few
values have been assigned, again from fitting with
organic acid soils, and the values selected are not
different greatly Jrom unity.
The description of ti\e diffuse layer within Model
V is a crude attempt to obtain a practical model,
and consequently there are several uncertainties
associated with it, as follows:
(a) The relative amounts of accumulation of differ-
ent counterfoils depend upon the diffuse layer
volume; the smaller the volume, the more are
ions of higher charge favored. Therefore uncer-
tainty in this volume translates to uncertainty
in nonspecific binding.
(b) Although reasonable results can be obtained
with the model for suspensions of soil solids,
experimentally it is difficult to make measure-
ments at the high effective concentrations
encountered in real soils. Extrapolation to such
conditions relies on the maintenance of the
assumed overlap of diffuse layers, as described
by Equation (6).
(c) The overlapping of diffuse layers in concen-
trated systems must alter the potential field at
, the humic surface, and thereby change the ionic
distribution that is assumed implicitly in apply-
ing the modifying term e '*"* to account for
the electrostatic influence on binding at specific
sites (see previous description). This is ignored
in the model.
It is clear that each of these uncertainties gets more
serious as concentration is increased, and predictions <^.
about ionic distributions in real soils and sediments
are only first approximations. Nonetheless, correct
trends are likely to be captured.
Parameters required for diffuse layer calculations
are humic molecular weights and radii./DL, Kz, each
of which is approximated in the present versions of
WHAM (Table 1). If selectivity among counterfoils is
considered, as in the soil/sediment version of the
model then values of K^(i) for each counterion
species are required also (see Appendix 1).
Fixed charge cation exchanger (clay)
This is included to allow some account to be taken
of the presence of soil and sediment clays. The same
approach as for the humic diffuse layer is taken,
except that the day is assumed to have a flat surface
for the purposes of calculating diffuse layer volume,
and selectivity is possible. The total diffuse layer
volume is constrained along with those of the FA and
HA, by extending Equation (6). Default values for
the cation exchange capacity and surface area of the
clay are 10"4equivg~' and I00mag~' respectively,
these being representative of values given by
Talibudeen (1981). In the present version of the
soil/sediment database (see Appendix 1), all the
selectivity coefficients for the day are set to 1 (cations
and neutral spedes) or 0 (anions).
Precipitation of aluminum and iron(lll) oxyhydroxides
WHAM allows these two solid phases to form.
Straightforward solubility products are used, and are
compared with the appropriate ion activity products,
to dedde whether precipitates are present The solid
phases are not considered to have active surfaces, that
is they cannot bind ions, nor is there any explidt
reaction with humic substances. In using the model,
these points must be borne in mind—thus WHAM in'
this version is best for situations in which the solid/
-------
978
E. TIPPING
for these, because different forms may be present,
with different solubilities. Values typical of amor-
phous precipitates are included in the database (see
Appendix 1), but these can be altered for specific
problems if necessary.
The inclusion of aluminum and iron(III) oxy-
hydroxides in the model reflects the environmental
systems, acid waters, and soils, for which the model
first was designed. Future applications may require
account to be taken of the formation and dissolution
of other inorganic phases such as carbonates and
sulfides. Descriptions of these reactions could be
incorporated readily into the code.
Sorption offulvic acid by soil and sediment solids
Tipping and Woof (1990, 1991) showed that net
electrical charge is important in die interactions of
fulvic acid with soil solids, and employed a modeling
approach that included a simple description of the
hydrophobicity of the fulvic molecules. This allows
the distribution of fulvic acid between the aqueous
and solid phases to be estimated. The fulvic acid
molecules are considered to consist of a series of ten
subtractions, each having the same ion-binding
properties, but differing in hydrophobicity. Thus,
desorption depends on whether the net fulvic charge
is sufficient to overcome the hydrophobicity. The
model has three parameters. The abundance distn-
bution of the ten fulvic fractions is described with 7,
whereas the adsorption of each fraction is character-
ized by (S and K*. Default parameter values are
included in the SSED database (see Appendix 1);
these have been determined to be typical for organic
matter release from acid organic soils, but their use
for higher pH systems or sediments has not been
investigated. In running the model, it is possible to
override the adsorption-desorption model, and
simply specify the concentration of dissolved fulvic
acid.
Chemical speclatlon In the solution phase
This concerns the humic-free solution phase, in
which only inorganic species, and low molecular
weight organics if required, are recognized. Complex-
ation reactions are formulated in terms of up to three
master species, as in the PHREEQE code developed
by Parkhurst and others (1980). Activity coefficients
are calculated with the extended Debye-Huckel
equation. Equilibrium constants and enthalpies for
the reactions considered are listed in the databases
for WHAM (see Appendix 1). These values were
obtained primarily from the compilations of Baes and
Mesmer (1976), Smith and Marteli (1977), Nord-
strom and others (1990), Read and others (1991) and
Mattigod and Sposito (1979). Some values also were
taken from Duffle (1988), Maes, De Brabandcre, and
Cremers (1988), Sunda and Hanson (1979) and
Tipping, Woof, and Hurley (1991). Users can assem-
ble their own databases if required. When only a few
reactions are of interest, it is advantageous to use a
small database in order to reduce execution times.
SOLVING THE WHAM EQUATIONS
There are two versions of WHAM; WHAM-W for
waters, and WHAM-S for soils and sediments. The
solving algorithms differ slightly, mainly because of
the different convergence characteristics required for
dilute and concentrated systems. The essence of the
problem is to distribute known total concentrations
of master species among the different chemical forms
considered. The master species are mostly parent
cationic and anionic species (H*. Na*. Al1*, d',
POJ-, etc.). The criteria for solution are mass
balance, charge balance (except for WHAM-W in
situations where pH is fixed), and correct net charge
on fulvic and humic acids. In outline, the procedure
is as follows:
(I) Set initial trial values of ionic strength, master
species' activities, fulvic, and humic net
charges.
(2) Calculate activities and concentrations of all
inorganic species.
(3) Calculate binding to fulvic and humic acid,
and clay cation-exchanger (WHAM-S only).
Calculate net fulvic and humic charge from
amounts specifically bound.
(4) Calculate new ionic strength.
(S) Perform mass balances on all components,
calculate the charge ratio in the humic-free
solution phase (unless pH is fixed), and com-
pare trial and calculated values offulvic and
humic charge. If all these meet the criteria for
convergence, and if ionic strength has reached
a constant value, proceed to step (9), (10), or
(11), depending on the status of the precipi-
tation reactions. If the criteria are not met,
proceed to step (6).
(6) Improve activity values by the continued-
fraction approximation, using ratios of calcu-
lated and known total concentrations. The
activity of H* is improved on the basis of
charge imbalance. '
(7) Improve ZFA and ZHA by interpolating
between the previous trials and the calculated
values from (3).
(8) Return to step (2).
(9) If the solubility of Fe(OH), has been
exceeded, return to step (2), and repeat the
procedure, except that the activity of Fe**
now is obtained from the activity of H* and
the solubility product. (WHAM-S only.)
(10) Repeat (9) for Al(OH), precipitation.
(WHAM-S only.)
<-)> (11) Calculate the distribution of FA between solid
and aqueous phases (WHAM-S only).
Computer codes in Turbo BASIC to perform the
described tasks are given in Appendices 2 and 3. Code
-------
Equilibrium chemical specialion by WHAM
979
verification has been achieved by checking individual
segments against hand calculations, by comparing
WHAM outputs with those from a version, of
PHREEQE (Parkhurst, Thorstenson, and Plummer,
1980) into which Model V has been incorporated
(M. Crawford, British Geol. Survey, pers. comm.),
and by comparing WHAM outputs for inorganic
speciation problems with outputs from WATEQ2
(Ball, Jenne, and Nordstrom, 1979). The robustness
of the algorithms has been investigated by performing
computations on 60 problems representative of
natural waters, sediments, and soils, and no signifi-
cant difficulties have been encountered. The usual
source of failure to execute (program crash or no
convergence) is an inappropriate selection of starting
pH, and this is rectified easily by editing the input file.
ILLUSTRATIVE CALCULATIONS
Soft freshwater containing trace metals (Appendix 4)
This is a fixed pH calculation and so the final
charge balance is not unity. When the model is used
in this mode, care has to be taken to supply approxi-
mately correct values for concentrations of major
electrolyte ions, in order to ensure reasonable values
of ionic strength. Because the system is dilute with
respect to humic substances, the humic-free solution
phase accounts for most of the total volume (993%).
As is usually the situation, the humic molecules have
negative charges, and these are balanced in the diffuse
layers almost entirely by sodium, magnesium, and
calcium. However, such accumulation accounts for
only a few percent of these major cations. The two
trace metals, Ni and Cu, show contrasting behaviors,
with less than 50% of the Ni being bound by humic
substances, but 99.8% of the copper. Because these
metals have appreciably greater affinities for the
humic substances than do Na, Mg, or Ca, their
binding is mainly by specific complexation; diffuse
layer accumulation is insignificant Nonetheless, the
tabulated values of » (moles of a species bound per
gram of humic substances) show that, by virtue of
their much higher solution activities, the alkaline
earths occupy more binding sites than the trace
metals.
Sediment containing trace metals (Appendix 5)
The sediment is assumed to be 90% by weight
water, and to contain appreciable organic matter and
clay. The volume of solution phase is significantly less
than the total volume, the combined diffuse layer
volumes accounting for 17% of the total water
volume. Despite the greater mass of clay present, the
organic matter dominates the calculated chemical
speciation. Sodium is recognized mainly (63%) in the
solution phase, but the alkaline earths are bound
predominantly to the solids, because of counterfoil
accumulation and binding at specific sites. The large
amount of Fe(III) in the system is calculated to be
present mainly (73%) as an oxyhydroxide precipitate,
with nearly all the rest specifically bound by humic
acid. Nickel is mostly sediment-bound, or bound to
dissolved fulvic acid, so that only 0.4% is present as
free Ni14. Copper has the same son of distribution,
but only 2 x 10~5% is calculated to be present as
the free divalent cation. The model output includes
Kt values, expressed in terms of the Model V
"architecture", that is the total amount of master
species bound per gram of solids divided by the
total concentration in the aqueous phase, which
includes that bound to dissolved fulvic acid. The
apparent /^ values are what would be calculated by
assuming all the water in the system to be bulk
solution, that is they are values that might be
calculated conventionally from experimental results.
Add soil containing radioelements (Appendix 6)
The soil is assumed to be 75% by weight water.
Because of the high concentrations of humic sub-
stances and the low ionic strength, nearly 25% of the
water is calculated to be present in the diffuse layers.
In this situation, the concentration of dissolved fulvic
acid has been calculated with the model, and deter-
mined to be 46mg/l, a reasonable value for an
organic soil solution. Sodium is predominantly in
solution, but this is not true of the other monovalent
cation, Cs, because of the different selectivity
coefficients for counterion accumulation (see Appen-
dix 1). Aluminum, frequently the dominant inorganic
element in acid soils, is bound specifically to the
humic substances; only 0.02% is free in solution. The
Kt values for the radioelements Co, Sr, and Cs are
determined mainly by nonspecific counterion
accumulation; higher pH values would be needed for
specific binding by humic substances to contribute
significantly to the distributions of Co and Sr,
whereas the model does not permit specific binding of
Cs by humics. However, even in organic-rich soils
there may be enough frayed-edge illite to bind a large
proportion of trace amounts of Cs. WHAM might be
able to handle such situations with a small amount of
the clay cation exchanger, having a high selectivity
for Cs. The K« value for americium reflects to a large
extent the distribution of fulvic acid between the solid
and solution phase, because strong complexing by
both solid-phase and dissolved humics means that the
concentration of free AmJ+ is only 0.01% of the total.
EXECUTION TIMES
A 386 machine operating at 25MHz, with a
numerical coprocessor, took 21, 264, and 80 sec
respectively to run the problems. Speeds can be
increased considerably if the database size is reduced,
because this allows smaller arrays to be employed.
However, to achieve maximum speeds, such a
reduction requires the renumbering of species, as it is
the identifying number of the last species, not the
number of species per se that determines array size.
-------
980
E. TIPPING
As an example, when the sediment problem discussed
previously was run with the identifier Tor the last
required species changed to 160, instead of the value
of 400 used with the default database, the execution
time was reduced by 70%.
A planned application of WHAM is in dynamic
ecosystem models, where it will be used in order to
compute chemical speciation on, for example, daily
timesteps. Under these circumstances, the model can
be made to run more quickly by storing values of
variables from one timestep to use as starting trial
values in the next, because step-to-step changes in the
values usually will be small.
AVAILABILITY OF WHAM
Executable versions of the programs, guidance
notes for users, and details of database documen-
tation are available. A charge will be made to cover
materials, postage, and handling. The author should
be contacted for further information.
Acknowledgments—This work was funded by the UK
Ministry of Agriculture, Fisheries, and Food. I am grateful
to Lorenzo Giusti (University of Lancaster, U.K.) for
helpful comments during the development of the programs.
REFERENCES
Allison, J. A., and Perdue. E M.. 1994. Modeling
melal-humic interactions with MINTEQA2: Proc. 6th
Intern. Meeting. Intern. Humic Substances Soc. (Ban,
Italy), in press.
Baes, C F^ and Meaner, R. E, 1976. The hydrolysis of
cations: John Wiley & Sons, New York, 489 p.
Ball, J. W., Jenne, E. A., and Nordstrom. D. K., 1979.
WATEQ2—a computerized chemical model for trace
and major element speciation and mineral equilibria of
natural waters. In Jenne, E A., ed.. Chemical modeling
in aqueous systems: Am. Chera. Soc. Symp. Ser.. v. 93,
p. 815-835.
Bolt, G. H., 1982, The ion distribution in the diffuse double
layer, in Boll, G. H.. ed.. Soil chemistry B. Physico-
chemical models: Elsevier, Amsterdam, p. 1-26.
Duffle, J.. 1988. Compilation reactions in aquatic systems:
an analytical approach: Ellis Horwood. Chichester,
692 p.
Hiemenz, P. C, 1977. Principles of colloid and surface
chemistry: Dekker. New York. 516 p.
Maes, A.. De Brabandere, J., and Cremen. An 1988, A
modified method for the measurement of the stability of
europium humic acid complexes in alkaline conditions:
Radiochira. Ada. v. 44/45, p. 51-57.
Mattigod, S. V.. and Sposito. G., 1979. Chemical modeling
of trace metal equilibria in contaminated soil solutions
using the computer program GEOCHEM, in Jenne,
E A., ed.. Chemical modeling in aqueous systems: Am.
, Chem. Soc. Symp., Washington, D. C, p. 837-856.
Nordstrom, D. K., Plummer, N. C, Langmuir. D., Busen-
berg, E, May, H. M., Jones, B. F., and Parkhurst, D. L,
1990. Revised chemical equilibrium data for major
water-mineral interactions and their limitations, In
Mekhior, D. C, and Bassett, R. L, eds.. Chemical
modeling of aqueous systems II: Am. Chem. Soc. Symp.,
Washington, D. C, p. 398-413.
Parkhurst. D. L.. Thontensen. D. C. and Plummer. N. L.,
1980. PHREEQE—« computer program for geochemi-
cal calculations: U.S. Geol. Survey Water Res. Invest,
RepL 80-96. 193 p.
Read, D., Fabriol. R^ Jamet, P., Tweed. C. and SeUin, P.,
1991, Application and validation of predictive computer
programs describing the chemistry of radionuclides
in the geosphere—CHEMVAL project: Comm. Eur.
Coramun. Rept. M0979.003, Brussels, 83 p.
Smith, R. M, and Martell. A. E, Critical stability con-
stants. Vol. 4: Inorganic complexes: Plenum Press, New
York. 257 p.
Sunda. W. G., and Hanson, P. J, J979, Chemical speciation
of copper in river water, in Jenne, E A., ed.. Chemical
modeling in aqueous systems: Am. Chem Soc. Symp.,
Washington. D. C, p. 147-180.
Susetyo, W., Carrara, L. A., Azanaga, L. V., and Grimm,
D. M.. 1991. Fluorescence techniques for metal-humic
interactions: Fres. Zeit Anal. Chem.. v. 339, no. 9.
p. 624-635.
Talibudeen, O.. 1981, Cation exchange in soils. In Green-
land. D. J., and Hayes. M. H. B., eds.. The chemistry of
soil processes: John Wiley & Sons, Chichester.
p. 115-177.
Tanfbrd, C, 1961, Physical chemistry of macromolecules:
John Wiley & Sons, New York, 710 p.
Tipping, E, I993a, Modeling the competition between
alkaline earth cations and trace metal species for binding
by humic substances: Env. Sci. Tech., v. 27, no. 3.
p. 520-529.
Tipping, E, 1993b, Modelling ion-binding by humic acids:
Coll. Surf., v. 73. p. 117-131.
Tipping. E. 1993c, Modelling the binding of europium and
the actinides by humic substances: Radiochim. Ada,
v. 62. p. 141-152.
Tipping, E, and Hurley, M. A., 1992, A unifying model
of cation binding by humic substances: Geochim. Cos-
mochim. Ada. v. 56. no. 10, p. 3627-3641.
Tipping, E, and Woof. C., 1990, Humic substances in acid
organic soils: modelling their release to the soil solution
in terms of humic charge: Jour. Soil. Sci., v.4l, no. 4,
p. 573-583.
-^Tipping, E, and Woof, C, 1991. The distribution of humic
substances between the solid and aqueous phases of acid
organic soils; a description based on humic heterogeneity-
and charge-dependent sorption equilibria: Jour. Soil
Sci.. v.42. no. 3, p. 437-448.
. Tipping. E. Woof. C. and Hurley, M. A. 1991. Humic
substances in acid surface waters; modelling aluminum
binding, contribution to ionic charge-balance, and con-
trol of pH: Water Res., v. 25. no. 4. p. 425-435.
to:
APPENDIX 1
WHAM
Database SSED, version 1.0 (complete), and database WATER, version 1.0 (first lines). The first 10 lines in SSED refer
(1) database name; (2) default parameters for humic acid (see text and Table 1); (3) default parameters for fulvic acid (see
text and Table 1); (4) clay cation exchange capacity (equiv g-1) and specific surface area (mj g"'); (5) factor/DL in Equation
(6) that limits the volume of the diffuse layer; (6) thermodynamic constants for AI(OH),; (7) thermodynamic constants for
Fe(OH),; (8) factor K. in Equation (7) to diminish the diffuse layer at low net humic charge; (9) constants 7. /) and K, used
in the fulvic acid desorplion model (see text): (10) number of following lines, each characterizing an individual chemical
-------
Equilibrium chemical spcciation by WHAM 981
species. In the WATER database, only five of these lines are required, corresponding to (I), (2), (5). (8). and (10) in SSED.
In SSED a data line for individual chemical species is interpreted as follows; species number, species name, charge,
composition (up to three master species), stoichiometry (three values). IogMKand AH for the formation of the species Of
master species, logN£ = 999), pKMHA values for humic and fulvic acids, selectivity coefficient for humic substances, and
selectivity coefficient for the clay cation-exchanger. In the WATER database, fines for the individual species are the same
as in SSED; except that the selectivity coefficients are absent The log,,/:and pK values refer to 25'C, and AH. the standard
enthalpy of reaction (kcalmol'1), is assumed to be constant with respect to temperature.
WHAM DATABASE SSED VERSION 1.0 (SSED10.DBS)
EA parameters, 3.29E-3, 4.02, 8.55, 1.78, 3.43, -374, 0.5, 1.72E-9, 15000
FA parameters, 4.73E-3, 3.26, 9.64, 3.34, 5.52, -103, 0.4, 8B-10, 1500
Clay parameters, 1B-4, 100
Double layer overlap factor, 0.25
Log KSO(25) 6 DeltaB for A1(OH)3, 9.0, -25000
Log KSO(25) 6 DeltaH for Fe(OH)3, 3.0, -25000
Constant to control DDL at low ZED, 1B3 - _ ' , . -. ^
Constants for FA desorption, 1.0,1.0,164 • ~ frS*.-*" jj).t*l\\'
No. of*data lines, 197
\\
2,BB,
3,Na,
4, Kg,
5,A1,
6, 1C,
7,Ca,
8,CrIII,
10, Fell,
12* CO,
13, Hi,
14,CU,
15, Zn,
16, Sr,
17, Cd,
18, Cs,
19, Ba,
20, Eg,
21, Pb,
22,002,
23,01V,
24,PuIII,
25,PuIV,
26,Pu02,
27, Th,
28, Am,
29,NH4,
30, Cm,
51,OB,
52,C1,
53,N03,
54,S04,
55, COS,
56,F,
57.P04,
101,EC03,
102,E2CO3,
104, EF,
105,EPO4. .
106,B2P04,'
107.H3P04,
113,BeOB,
114, Be (OH) 2,
115, Be (OH) 3,
116, Be (OH) 4,
117,BeS04,
118, BeF,
124,HgEC03,
125,MgC03,
126,MgS04,
127,MgHP04,
133.A10H,
1,
2,
1,
2,
3,
1,
2,
3,
2,
2,
3,
2,
2,
2,
2,
2,
2,
1,
2,
2,
2,
2,
4,
3,
4,
2,
4,
3,
1,
3,
-1,
-1,
-1,
-2,
-2,
-1,
-3,
-1,
0.
0,
-2,
-1,
0,
1,
0,
-1,
-2,
0,
1,
1,
o,
o,
o,
2,
1,0,0,
2O O
6,0,0,
8 0,0,
29,0,0,
30,0,0,
51,0,0,
52,0,0,
53,0,0,
54,0,0,
55,0,0,
56,0,0,
57,0,0,
1,55,0,
1,55,0,
1.56,0,
1.57,0,
1,57,0,
1.57,0,
2,51,0,
2,51,0.
2.51.0.
2,51,0,
2,54,0,
2,56,0,
4,1,55,
4.55,0,
4,54,0.
4,1,57,
5,51,0,
1,0,0,
10 0
1,0,0,
10 0
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,1,0,
2,1,0,
1,1,0,
1,1,0,
2,1,0,
3,1*0,
1,1,0,
1,2,0,
1,3,0,
1,4,0,
1,1,0,
1,1,0,
1,1,1,
1,1,0,
1,1,0,
1.1.1,
1.1,0,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
10.329,
16.681,
3.18,
12.35,
19.55,
21.70,
8.6,
14.35,
18.75,
18.59,
1.95,
5.2,
11.40,
2.98,
2.37,
15.26,
9.01,
ft.
0,
0,
0,
o,
o.
0,
0,
0,
o,
o,
0,
o,
o,
o,
o,
o,
o,
0,
0,
o,
o.
o,
o,.
o,
o,
o,
o,
0,
0,
o,
0,
o,
0,
o,
-3.561
-5.738
3.18,
-3.5,
-4.3,
-2.4,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
-2.77,
2.71,
4.55,
-0.5,
-1.54,
999, 999,
999, 999,
3.3, 2.2,
1.3, 0.4,
999, 999,
3.2, 2.2,
3.4, 1.7,
2.1, 1.3,
0.8, -0.2,
2.7, 1.7,
2.7, 1.4,
1.5, 0.8,
2.3, 1.3,
2.8, 2.3,
2.7, 1.5,
999, 999,
3.6, 2.6,
0.2, -0.3,
1.7, 0.9,
1.3, 0.9,
0.0, -0.7,
1.7, 0.8.
0.0,-0.7,
1.6, 0.5,
0.6,-0.4,
1.2, 0.3,
999, 999,
2.0, 1.6,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
,999, 999,
,999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
1.7, 0.4,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
1.3, 0.4,
1,
0.25,
0.75,
0.5,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1.5,
1,
1.5,
1.5,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
If
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
0.5,
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
1
1
0
0
1
1
1
0
0
1
1
1
1
1
1
1
-------
2 E. TIPPING
134,Al(OH)2, 1, 5,51,0, 1.2,0, 17.87, -3.68, 999, 999, 1, 1
135,Al(OH)4, -1. 5,51,0, 1,4,0. 33.84, -10.86, 999. 999, 1, 0
136,A1S04, 1. 5,54,0, 1,1,0, 3.20, 2.3, 999, 999, 1, l
137,A1P, 2, 5,56,0, 1,1,0, 6.98, 0.7. 999, 999, 1, 1
138,A1P2, 1. 5,56,0, 1,2,0, 12.60, 2.0, 999, 999, 1. 1
139,A1F3, 0, 5,56,0, 1,3.0, 16.65. 2.2. 999, 999, 1, 1
145,CaHC03, 1, 7,1,55, 1,1,1, 11.44, -0.87. 999. 999, 1, 1
146,CaC03, 0, 7,55,0, 1,1,0, 3.22, 3.55. 999, 999, 1, 1
147,C*S04, 0, 7,54,0. 1,1,0, 2.30, 1.65. 999. 999, 1. 1
148,C«HP04, 0, 7,1,57, 1,1,1, 15.09, -0.5, 999, 999, 1, 1
154,CrOH, 2, 8,51,0, 1,1,0, 10.0, 0.0, 0.5. 0.1, 1, 1
155,Cr(OH)2, 1, 8,51,0. 1.2.0, 18.3, 0.0, 999, 999. 1, 1
156,Cr(OH)3, 0, 8,51,0, 1,3,0, 24.0, 0.0, 999, 999, 1, 1
157,Cr(OH)4, -1, 8,51,0, 1,4,0, 28.6, 0.0, 999, 999, 1,
158,CrS04, 1, 8,54,0, 1,1,0, 2.0, 2.0, 999, 999, 1,
159.CTF. 2, 8,56,0,. 1,1,0, S.2, 2.0, 999, 999, 1,
165,KnOH, 1, 9,51,0, 1,1,0, 3.41, ' 1.04, 3.4, 1.7, 1,
166,KnS04, 0, 9,54,0, 1,1,0, 2.25, 3.37, 999, 999, 1,
167,MnC03, 0, 9,55,0, 1,1,0, 4.52, 0.0, 999, 999, 1,
168,MnCl, 1. 9,52,0, 1,1,0, 0.61, 0.0, 999, 999, 1,
169.MDHP04, 0, 9,1,57, 1,1,1. 15.93, 0.0, 999, 999, 1,
170,KnHC03, 1, 9,1,55, 1,1,1, 12.28, 0.0, 999, 999, 1,
175.FOOH, 1, 10.51,0, 1,1,0, 4.5, -0.2, 2.1, 1.3, 1,
176,F«S04, 0, 10.54,0, 1,1.0. 2.25, 3.23, 999, 999. 1,
177.F0C03. 0, 10,55,0. 1,1.0, 5.31, 0.0, 999, 999, 1, 1
178,F«HP04, 0, 10,1,57. 1,1,1, 15.95. 0.0, 999, 999, 1, 1
179,F«C1, 1, 10,52,0, 1,1,0, 0.14, 0.0, 999, 999, 1, 1
1BO,F«HC03, 1, 10,1,55, 1,1,1, 12.33, 0.0, 999, 999, 1, 1
185,FeOH, 2, 11,51,0, 1,1,0, 11.81, -2.96, 0.8,-0.2, 1, 1
186,Fe(OH)2, 1, 11,51,0, 1,2,0, 22.33, -9.62, 999, 999, 1, 1
187,F«(OH)3, 0, 11,51,0, 1,3,0, 29.44, -15.29, 999, 999, 1, 1
186,F0(OH)4, -1, 11,51,0, 1,4,0, 34.4, -21.55, 999, 999, 1, 0
190,F«S04, 1, 11,54,0, 1,1,0, 4.04, 3.91, 999, 999, 1, 1
191,F«F, 2, 11,56,0, 1,1,0, 6.2, 2.7, 999, 999, 1, 1
192,F«F2, 1, 11,56,0, 1,2,0, 10.8, 4.8, 999, 999, 1, 1
193,Feffl?04, 1, 11,1,57, 1,1,1, 21.55, 0, 999, 999, 1, 1
194,F«C1, 2, 11,52,0, 1,1,0, 1.48, 5.6, 999, 999, 1, 1
195,F«C12, 1, 11,52,0, 1,2,0, 2.13, 0.0, 999, 999. 1, 1
201.COOH, 1, 12,51,0, 1,1,0, 4.35, 0.0, 2.7, 1.7, 1, 1
202.Co(OH)2, 0, 12.51,0, 1,2,0, 9.2, 0.0. 999. 999. 1, 1
203,CoS04, 0, 12.54,0, 1,1,0, 2.36, 1.4. 999. 999. 1. 1
204,COC03, 0. 12,55,0, 1,1,0, 5.53, 0.0, 999, 999. 1, 1
205,CoCl, 1, 12,52,0, 1,1,0, 0.3, 0.5, 999, 999, 1, 1
206,CoHC03, 1, 12,1,55, 1,1,1, 13.22, 0.0. 999. 999. 1, 1
211,NiOH, 1, 13,51,0, 1,1,0, 4.14, 0.0, 2.7, 1.4. 1, 1
212,Ni(OH)2, 0, 13,51,0, 1,2,0, 9.0, 0.0, 999, 999, 1, 1
213.H1S04, 0, 13,54,0, 1,1,0, 2.32, 1.5, 999, 999, 1, 1
214,NiC03, 0, 13,55,0, 1,1,0, 5.78, 0.0. 999, 999. 1, 1
215,HiCl, 1, 13,52,0, 1,1,0, 0.4, 0.5, 999, 999, 1, 1
216.H1HC03. 1, 13,1,55, 1,1,1, 13.41, 0.0. 999, 999, 1, 1
221,CuOB, 1, 14,51,0, 1,1,0, 6.48, 0.0, 1.5, 0.8, 1, 1
222,Cu(OH)2, 0, 14,51,0, 1,2,0, 11.78, 0.0. 999, 999, 1, 1
223.CUS04. 0, 14,54,0, 1,1,0, 2.36, 2.1, 999, 999, 1, 1
224.CUC03. 0, 14,55,0, 1,1,0, 6.75, 0.0, 999. 999. 1, 1
225,CU(C03)2. -2, 14,55,0, 1,2,0, 9.92, 0.0, 999. 999, 1. 0
226,CUC1, 1, 14,52.0. 1.1.0, 0.4. 1.6. 999, 999. 1. 1
227.CUHC03. 1. 14,1,55, 1,1,1, 14.62, 0.0, 999, 999. 1, ' 1
232,ZnOB, 1, 15,51,0, 1,1,0, 5.04, 0.0, 2.3, 1.3, 1, 1.
233,Zn(OH)2, 0, 15,51,0, 1,2,0, 11.1, 0.0, 999, 999, 1, 1
234,ZnS04, 0, 15,54,0, 1,1,0, 2.38, l.S, 999, 999. 1. 1
23S.ZnC03, 0, 15,55,0, 1,1,0, 4.76, 0.0. 999. 999, 1, • 1-
236,ZnCl, 1, 15,52,0, 1,1,0, 0.4, 1.3, 999, 999, 1, 1
237,ZnHC03, 1, 15,1,55, 1,1,1, 13.12, 0.0, 999, 999, 1, 1
242,SrS04, 0, 16,54,0, 1,1,0, 2.29, 2.08, 999, 999, 1, 1
243,SrC03, 0, 16,55,0, 1,1,0, 2.81, 5.22, 999, 999, 1, 1
244,SrHC03, 1, 16,1,55, 1,1,1, 11.51, 2.49, 999, 999, 1, 1
249,CdOH, 1, 17.51,0, 1,1,0. 3.92, 0.0, 2.7, 1.5, 1, 1
250,Cd(OH)2, 0, 17.51,0. 1,2,0, 7.65, 0.0. 999, 999. 1, 1
251.CdS04. 0, 17,54,0. 1,1,0, 2.46, 2.3, 999, 999, 1, 1
252,CdCl, 1, 17,52,0, 1,1,0, 1.98, 0.3, 999, 999, 1, 1
2S3,CdC12, 0, 17,52,0, 1,2,0, 2.6, 0.9, 999, 999, 1, 1
259,BaS04, 0, 19,54,0, 1,1,0, 2.7, 0.0, 999, 999, 1, 1
-------
Equilibrium chemical spcciation by WHAM 983
260,BaCO3,
261,BaHCO3,
266,HgOH,
267,Hg(OH)2,
268,Hg(OH}3,
*)CQ VetRGA
Zo!f,I»Sl»u*»
270,HgCl,
271,HgC12,
272,HgCl3,
273,HgC14,
279,PbOH,
280.Pb(OH)2.
281,Pb(OH)3,
282,PbS04,
283,PbCO3,
284,Pb(C03)2,
285,PbCl,
286,PbC12,
292,U020H,
293,002 (OH) 2.
294,U02(OH)3,
0,
1,
1,
0,
-1.
0_
*
1,
0,
-1.
-2,
1,
0,
-It
0,
0,
-2,
1,
0,
1,
o.
-1.
295,(U02)2(OH)2,2,
296, (U02)3(OH)5,1,
297.U02S04.
298,002003,
299,U02(C03)2,
305, nOB,
306,U(OH)2,
307,U(OH)3,
308,U(OH)4,
309,U(OH)5,
310, UP,
311,tJF2,
312,USO«;
313, XJC1,
317,PuOH,
318, Pu (OH) 2,
319.PUC03,
320, Pu (COS) 2,
321,Pu804,
322, PUC1,
328,PuOH,
329, Pu (OH) 2,
330,Fu(OH)3,
331, Pu (OH) 4,
332,PuS04,
333,Pu(S04)2,
334, PuCl,
335,PuC03,
336.Pu(C03)2,
342,Pu020H,
343,Ptl02(OH)2,
344,PU02(OH)3,
345,Pu02C03,
o.
o.
-2,
3,
2,
1.
0,
-1,
3,
2,
2,
3,
2,
1.
1,
-1.
1.
2,
3,
2,
1.
0,
2,
0,
3,
2,
0,
1.
o.
-1.
o.
346,Pu02(C03)2,-2,
347,PU02C1,
348,PU02S04,
354,ThOH,
355,Th(OH)2,
356,Th(OH)3,
3S7,Th(OH)4,
358, ThP,
3S9,ThF2,
360,ThC03,
361,ThCl,
367,AmOH,
368,Am(OH)2.
369,Am(OH)3,
370,Am(OH)4,
371,AfflC03,
372,Am(C03)2,
1.
0,
3,
2,
1,
0,
3,
2,
2,
3,
2,
1.
0,
-1.
1.
-1.
19,55,0,
19,1,55,
20,51,0,
20,51,0,
20,51,0,
20,52,0,
20,52,0,
20,52,0,
20,52,0,
21,51,0,
21,51,0,
21,51,0,
21,54,0,
21,55,0,
21,55,0,
21,52,0,
21,52,0.
22,51,0,
22,51,0,
22,51,0,
22,51,0,
22,51,0,
22,54,0,
22,55,0,
22,55,0,
23,51,0,
23,51,0,
23,51,0,
23,51,0,
23,51,0,
23,56,0,
23,56,0,
23,54,0,
23,52,0,
24,51,0,
24,51,0,
24,55,0,
24,55,0,
24,54,0,
24,52,0,
25,51,0,
25,51,0,
25,51,0,
25,51,0,
25,54,0,
25,54,0,
25,52,0,
25,55,0,
25,55,0,
26,51,0,
26,51,0,
26,51,0,
26,55,0,
26,55,0,
26,52,0,
26,52,0,
27,51,0,
27,51.0,
27,51,0,
27,51,0,
27,56,0,
27,56,0,
27,55,0,
27,52,0,
28,51,0,
28.51,0,
28,51,0,
28,51,0,
28,55,0,
28,55,0,
1,1,0,
1,1,1,
1,1,0,
1,2,0,
1,3,0,
1,1,0,
1,2,0,
1,3,0,
1,4,0,
1,1,0,
1.2,0,
1,3,0,
1,1,0,
1,1,0,
1,2,0,
1,1,0,
1.2.0,
1,1,0,
1,2,0,
1,3,0,
2,2,0,
3,5,0,
1,1,0,
1,1,0,
1,2,0,
1,1,0,
1,2,0,
1,3,0,
1,4,0,
1,5,0,
1,1,0,
1,2,0,
1,1,0,
1,1,0,
1,1,0,
1.2,0.
1,1,0,
1,2,0,
1,1,0,
1,1,0,
1,1,0,
1,2,0,
1,3,0,
1,4,0,
1,1,0,
1,2,0,
1,1,0,
1,1,0,
1,2,0,
1,1,0,
1,2,0,
1,3,0,
1,1,0,
1,2,0,
1,1,0,
1,2,0,
1,1,0,
1,2.0.
1,3,0,
1,4,0,
1,1,0,
1,2,0,
1,1,0,
1,1,0,
1,1,0,
1,2,0.
1,3,0,
1,4,0,
1,1.0,
1,2,0,
2.71,
11.31,
10.60,
21.83,
20.9,
7.21,
13.98,
15.06,
15.42,
6.29,
10.88,
13.94,
2.75,
7.2,
10.5,
1.59,
1.8.
8.8.
16.1.
21.0.
22.4.
54.3.
3.0,
9.4,
16.4,
13.3.
24.7.
34.2.
41.7,
47.3,
8.65,
14.47,
6.11,
1.16.
6.8,
12.12,
9.6,
12.9,
3.34,
1.12,
13.30,
28.14,
39.11,
46.81,
6.42,
10.78,
1.22,
19.14,
33.12,
8.3,
16.85,
20.88,
12.0,
14.9,
0.63,
4.68,
11.66,
21.64.
30.3,
40.1.
8.44.
15.06.
11.03.
1.18.
6.73,
12.04,
17.84,
16.0,
7.42,
11.86,
3.55,
2.00,
-5.8,
-16.2,
-20.0,
0.0,
-4.8,
-12.8,
-15.0,
-14.9,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
4.4,
0.0,
-2.6,
0.0,
0.0,
-17.1,
-42.7,
0.0.
1.1.
3.5,
-7.0,
-14.0,
-21.0,
-28.7,
-39.2.
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
o.tr.
0.0.
0.0,
0.0,
o.o,'-
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
-7.4,
-12.75,
-19.67.
-34.7,
0.0,
0.0,
0.0,
0.0,
. o.o.
0.0,
0.0,
0.0,
0.0,
0.0,
999, 999,
999, 999,
0.2.-0.3,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999.
1.7, 0.9,
999, 999,
999, 999,
999. 999,
999, 999,
999. 999,
999, 999,
999, 999,
1.3, 0.9,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999.
999, 999,
999, 999,
0.0.-0.7,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999.
999, 999.
999, 999,
999, 999,
1.7, 0.8,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
0.0.-0.7,
999, 999,
993, 999,
.999, 999,
999, 999.
999, 999,
999, 999,
999, 999,
999, 999.
1.6, 0.5,
999, 999,
999, 999.
999, 999,
999, 999,
' 999, 999,
999, 999,
0.6.-0.4,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
1.2, 0.3,
999. 999,
999, 999,
999, 999,
•999, 999,
999, 999,
1,
1,
1,
1,
1,
1.
"•
1,
1,
1,
1,
1,
1.
• 1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
- 1,
1,
1,
. 1,
1,
1,
1,
1.
1.
1.
1,
1,
1,
1,
1,
1.
1.
' 1,
1,
1,
1
1
1
1
0
j.
1
1
0
0
1
1
0
1
1
0
1
1
1
1
0
1
1
1
1
0
1
1
1
1
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
0
. 1
0
1
1
1
1
1
1
1
1
1
1
1
1
1
0
1
0
END OF DATA
-------
984
E. TIPPING
WHAM DATABASE WATER VERSION 1.0 (HATER10.DBS)
HA parameters, 3.29B-3, 4.02, 8.55, 1.78, 3.43, -374,
FA parameter*, 4.73B-3, 3.26, 9.64, 3.34, 5.52, -103, 0.4,
Double layer overlap factor, 0.25
Constant to control DDL at low ZED, 1E3
No. .of data lines, 214
1,H, 1, 1,0,0, .1,0,0, 999, 0, 999, 999
2,Be, 2, 2,0,0, 1,0,0, 999, 0, 1.7, 0.4
3,Na, 1, 3,0,0, 1,0,0, 999, 0, 999, 999
0.5, 1.72E-9, 15000
8E-10, 1500
APPENDIX 2
WHAM-W Program Listing
'iffiiiiiiiiiiiiiiiiiiiiiiiffiiiiffiiiiiiiffiiiiiiiiiiiiiiiiiiiiiiiiiiiffiffftiiiiiffii
'ft* WHAM-W VERSION 1.0 f|
'iff Computer code in TURBO BASIC §|
'il Speciation of natural water* containing humic substance* ft
'iff This version produced on IS April 1993 by B.Tipping, IKE, Windennere iff
'iffiffiiiffiiiiiffiiiiiiiiiiiffiiiiiftiffiiiffiiiiiffiffiiiSftffiiiiffiiiiiiftiiffiffffiiitfffffi
CIiS
PRINT »**•*****••**•*"•*•***"
PRINT •** W indemere **
H umic
A queous
M odel
•*
•*
•*
*••
H ater* **
**•
Version 1.0 **"
PRINT «***•*•**•***•****••**«
PRINT
PRINT
PRINT
PRINT ••*
PRINT "**
PRINT ••*
PRINT «**
PRINT •••
Equilibrium speciation of waters"
Variable or fixed pH"
Variable or fixed pCO2-
HS present or absent"
E.Tipping IFE April 1993"
PRINTtINPUT -NAME OF SOURCE FILE [OMIT .DAT TRAILER] S ",SF$
PRINT
OPEN SF$ + ".DAT" FOR INPUT AS «1
INPUT il,S$,S$
INPUT il,S$,DBS$
INPUT il,8$,NSP*
' file headers
' database identifier
' array sixe required
DIKension and clear arrays, then read in chemical constants
GOSUB WWSR1
««•••••*•••**•••••*•••«••*•***•
' Continue to read in data file
INPUT il,8$,PRECISION
INPUT il,S$,TBMPK
INPUT iI,S$,TB(l),TH(2)
INPUT, il, S$, PHFH$
INPUT il,S$,PHSTART
INPUT il,6$,PC02
INPUT il,S$,NCOMPX
FOR I«t • 1 TO NCOMPX
INPUT fl,S$,X\,TCONC
N$(X%) - S$
T(X%) B TCONC
NEXT I\
precision
temp K
total cones (HA,FA)
pH fixed / variable
starting or fixed pH
pC02
no. of master species
' total cones
CLOSE
-------
Equilibrium chemical spcciation by WHAM
985
' Perform calculations
COSUB WWSR2
' Calculations finished ; signal and make output file
SOUND 400,5 S SOUND 500,5 t SOUND 700,8 ' sound output
OPEN SF$ + ".OUT* FOR OUTPUT AS §2
COSUB WWSR8
CLOSE f 2
PRINT
PRINT "CALCULATION FINISHED t CONSULT ** ";SF$;".OUT ** FOR DETAILED OUTPUT'
STOP
END
WWSRlt ' callod by main program ff
•ft DIMensions arrays if
»ff reads in data from data base ff
•ff sets up Modal V ff
'888888888888888888888886*88888888888888818888888888888888888881888888888888888
• DIHension arrays ; inorganic chemistry
DIM N*(NSP\)
DIM N$(NSP*)
DIM T(NSP*)
DIM TCALC(NSP\)
DIM A(NSP*)
DIM C(NSPX)
DIM CHMNSP*)
DIM M1MNSP*)
DIM M2\(NSP\)
DIM M3X(NSP\)
DIM S1*(NSP\)
DIM S2MNSP*)
DIM S3*(NSP\)
DIM LK(NSP\)
DIM DH(NSP*)
DIM GAMMA(S)
numerical identification of species
nominal identification of species
total concentrations of components
total calculated concentrations of components
activities of solution species
concentations of solution species
charges on solution species
master species identifiers
master species identifiers
master.species identifiers
stoichiometries
stoichiometries
stoichiometries
equilibrium constants
enthalpies
activity coefficients
DIMension humic arrays ; 1 • HA, 2 • FA
DIM SP(2,8)
DIM SM(2,12)
DIM KH(2,8)
DIM PKH(2,6)
DIM KHH(2,8,1ISP*)
DIM SITE*(12,2)
DIM PKMHM2,NSP*)
DIM PKMHB(2,NSP\)
DIM FP(8)
DIM BIBTERM(NSPX)
DIM HONBTERM(NSPSt)
DIM BISNU(12,NSP%)
DIM MONSNO(8,NSP%)
DIM BINU(NSP^)
DIM MONNU(NSP86)
DIM NU(2,NSP*)
DIM CHC(2,NSP5t)
DIM CDDL(2,NSP%)
proton-binding sites not forming bidentate sites
bidentate metal-binding sites
K's for proton-binding
PR's for proton-binding
K's for metal-proton exchange
proton sites making bidentate sites
metal-proton exchange constants
metal-proton exchange constants
dissociation factors for proton-binding
terms calc'd in finding bidentate thota's
terms calc'd in finding monodentate theta's
values of NU at each bidentate site
values of NU at each monodentate site
overall NU for each metal/bidentate sites
overall NU for each metal/monodentate sites
overall NU for each complexed species
cone of complexed species per litre
cones in HS DDL'o
-------
986
E. TIPPING
DIM BIZ(12)
DIM MONZ(8)
DIM BINDTEST*(NSP*)
n«t charge at each bidentate site
not charge at «ach monodentate Bite
allows binding of a species to be tested
' Input of constants for inorganic apeciation and HS complication
OPEN DBS$ + ".DBS" FOR INPUT AS 12
' database identifier
' HA properties
INPUT §2,8$
IHPUT «2,S$,NCOOH(1) ,PKHA.(1) .PKHB(l) .DPKHA(l) ,_
DPKHB(l),P(l),FI>R(l),IlAX>I08(l),lfOIMT(l)
INPUT t2,S$,NCOOH(2),PKHA(2),PKHB(2),DPKHA(2),_
DPKHB(2),P(2),FPR(2),RADIU8(2),MOZMT(2) ', PA properties
INPUT f2,S$,DLF
INPUT f2,S$,KZBD
INPUT f2,S$,NODATA*
• double layer overlap factor
' Db vol at low Z
' number of species
FOR I* m 1 TO NODATA*
INPUT «,XVN$(X\),CH\(X*),M1*(X*),M2MX\),M3MX*),S1MX*),S2*(XX),S3MXX),
LK(XX),nH(X*),PKMHA(l,XX),PKMHA(2,X*) ~
IF PKMHM1.XX) < 999 THEN BINDTEST*(XM « 1
NEXT I*
CLOSE 82
' identifies species that
' don't bind to HS
>***•***••****«****•***•***********
' Set derived constants for Model V
FOR
PKH(X*,7)
NEXT Xfc
FOR I* •
FOR J\ •
• l.TO 2
- PKHA(Xt)
- PKHA(X*>
« PKHA(X\)
• PKHA(Xfc)
• PKBB(X\)
• PKHB(X\)
- PKHB(XX)
« PKHB(X\)
(OPKHA(XX)/2)
(DPKHA(X%)/C)
(OPKHA(X%)/£)
(DPKHA(X\)/2)
(DPKHB(XX)/2)
(OPKHB(X\)/6)
(DPKHB(XX)/2)
1 TO 2
1 TO 8
NEXT
NEXT
' individual site pK's
' convert pK's to K's
' Identify the proton-binding sites that combine to make the bidentate sites
SITE\(1,1)
SITB\(2rl)
SITB\(3,1)
8ITEX(4,1)
SITE\(S/1)
SITE\(7/1)
SITE%(8,1)
SITE%(12,1)
8ITEX(1.2)
8ITBV(2,2)
8ITB«(3r2)
SITB\(8,2)
SITB%(10f2)
SITB%(11,2)
SITE%(12,2)
' Set site concentrations
FOR I* - 1 TO 4
SP(1,I«O « (1 - FPR(D) * NCOOH(l) / 4
SP(2,IM - (1 - FPR(2» * NCOOH(2) / 4
NEXT 1%
-------
Equilibrium chemical spctiation by WHAM
957
FOR I* « 5 TO 8
SP(1,IX) « (1 - PPR(l)) * NCOOH(l) / 6
SP(2,I*) « (1 - FPR(2» * NCOOH(2) / 8
NEXT 1%
FOR IX « 1 TO 12
SH(1,X*) « FPR(l) * NCOOH(l) / 16
SH(2,IM - FPR(2) * NCOOH(2) / 16
NEXT X*
' set natal -proton exchange constant* for individual sites
FOR I* • 1 TO NSP*
' species not defined
IF N$(IX)
IF PKMHA(1,XX)
"" THEN JMBRIM.
999 THEN
FOR OX
PKKB
1 TO 4
PKNH
KHH(2,JX,XX)
NEXT OX
PKKHA(1,X*)
• 10M-MMH)
• PKMHA(2,IX)
« 10»(-PKMH)
FOR JX • 5 TO 8
FKMH « (3«PKHHA(1,IX)) - 3
KMH(1,JX.IX) « 10M-PKMH)
PKMH « 3.9£*PKMHA(2,X*)
KMH(2,3*,X\) « 10»(-PKMH)
NEXT J%
GOTO WWSR1L2
' 3*A - 3 conversion ; HA
' 3.96*A conversion ; FA
• If cone to here, then no binding of species I\ can occur
HWSRlLlt
FOR J\ m 1 TO 8
KMH(1,J\,I*) m 0
KHH(2,J\.I«C) « 0
NEXT J\
WMSR1L21
NEXT X\
RETURN
WWSR2t
'it
'ft
'if
'if
•if
'ii
'ff
called fron main program - calls KWSR3 and WWSR4
sets initial trial values of master species'
initialises all activities and concns by calling HWSR4
controls pH improvement if pH not fixed
controls level of precision
calls WWSR3 to calc speciation at a given pB
tests for correct pH with CHRATIO if pH not fixed
sets up screen to report on progress of calculation
ft
if
if
ff
if
ff
ff
ff
'iiiifffifffffffffffftfffffiffiffffiffffffffffffffffffffffffffffffffffffffifiii
' Set initial trial values
IS « 0.1 '
ZED(l).- -1E-4
ZED(2) - -1E-4
ionic strength
RATIO (1)
RATIO (2)
10
10
PH « PHSTART
A(l) « 10*{-PH)
FOR X* - 2 TO SO
A(X%) B T(X*>
• Z for HA
' Z for FA
' RATIO for HA
' RATIO for FA
' fixed or starting pH
• activity of H+
• species il is H+
' cationic master species
-------
988
E. TIPPING
NEXT X*
A(52)
A<53)
A(S4)
A(56)
A(57)
T{52)
T(53)
T(54)
T<56) * 1E-4
T{57) * IE-IS
IF PC02 « 999 THEN A{55) « 1E-6 * T(55)
Cl
NO3
sot
F
P04
C03,2-
Suamary output to screen, and headers
LOCATE 13
PRINT "
LOCATE 13
PRINT
PRINT
IF TH(1) - 0 AND TH(2) • 0 THEN PRINT
IF 0
IF PRFIX* • "NO"
IF PHFIX$ - "YES*
IF PCO2 - 999
IF FCO2 < 999
THEN PRINT
THEN PRINT
THEN PRINT
THEN PRINT
THEN PRINT
SOURCE FILE ";SF$
PRECISION (*) "|_
USING "f.if*** "fl.0001*PRECISION
ABSENT"
PRESENT"
PH VARIABLE"
PH FIXED"
PC02 VARIABLE"
PC02 FIXED"
HUMIC SUBSTANCES
HUMIC SUBSTANCES
PRINT
PRINT "ITER*;TAB(15);"PH";TAB(30);"IS";TAB(45);"CHRATIO";TAB(£2);"
*****
' Begin calculations
NUMITfe • 0
GRIT » 1E-4 i COSUB WWSR4
IF PHFIX$ - "NO" THEN WWSR2L1
' itoration countar
• initialization
' routine if pH variable
' routine for
IF PHFIX$ « "YES" THEN CRTEST$ • "PHASE2"
IF PHFIX$ • "YES" THEN CRIT « (PRECISION/100) *2 ' fixed pH calcn
IF PRFIX$ • "YES" THEN COSUB WWSR3
IF PHPIX$•« "TES" THEN WWSR2L4
' fixed pH calca done; return
/•A********************************************
' UB« pH adjusting routine, solving for each pH
WWSR2L11
CRTEST$ • "PHASE1"
CRIT • 1E-4 t COSUB WWSR3
DPHF -"0.2
IF CHRATIO < 1 THEN WWSR2L2
IF CHRATIO > 1 THEN WHSR2L3
' find initial CHRATIO etc
• initial pH adjust factor
' Come to here If CHRATIO is < 1
WWSR2L2 :
DPH * DPHF / CHRATIO
IF DPH < 0.01 THEN CRIT • (PRECISION/ 100) *2
' pH increment
' adjust precision
-------
Equilibrium chemical specialion by WHAM
989
IP DPH < 0.0001 THEN CRTEST$ - "PHXSE2"
PH - PH - DPH
GOSUB WWSR3
IP CRTEST$ • "PHXSE2" XND_
ABS(CKRXTIO - 1) < SQR(CRIT) THEN WWSR2M
IP CHRXTIO > 1 THEM DPHF • DPHP/3
IP CHRXTIO > 1 THEN WWSR2L3
GOTO WWSR2L2
' refining
' increment pH
' calculate speciation
' finished
' change pH adjust factor
' go to > routine
' continue to increment pH
' Come to hare if CHRXTIO is > 1
WWSR2L3t
DPH • DPHP * CHRXTIO
IP DPH < 0.01 THEN CRIT « (PRECISION/100)*2
IP DPH < 0.0001 THEN CRTBST$ - BPHXSE2B
PH « PH + DPH
COSOB WWSR3
IP CRTBST$ m "PHXSE2" XND_
ABS (CHRXTIO - 1) < EQR(CRIT) THEN WWSR2M
IF CHRXTIO < 1 THEN DPHP - DPHP/3
IP CHRXTIO < 1 THEN WMSR2L2
GOTO WWSR2L3
WWSR2L4I
RETURN
pH increment
adjust precision
refining
increment pH
calculate speciation
finished
change pH adjust factor
go to < routine
continue to increment pH
•fffffftiiifffffffffififiiffffffffffffffffffffffffffifffffffffffffffffffifffffif
WWSR3t ' called by MWSR3 - calls WWSR4 ff
'ff controls improvement of activities (not H+) and ZFA,ZHX ff
•ff calls WWSR4 to do mass and charge calculations ff
'ff in PHXSE 1 returns when CHRXTIO has become nearly-constant ff
'ff in PHXSE 2 returns when mass and charge balances O.K. ff
'ff reports'progress of calculation to screen ff
'fffffffftfffffffffiffififfffffffifffffffffffffffffffffffffffffffffffffffffffffff
' activity of H+
' resetting values
10*(-PH)
XASTCHRXTIO • 0 » LXST1CHRXTIO
99m9mm9mm9mmmmmmm»tt9mmm»9mm^t^mmmmmmmmmmmm mmmmmmttmmmmmmm mmmmmmmmmmmmmmmmmmm
' Begin iterative cycle ; come back to here until convergence is achieved
WWSR3I.lt
• NOHITX + 1
' update iteration counter
««•*•***•*•••**•«•••*•••*•
' Calculate ionic strength
IS « 0
preparing to sum
POR X* • 1 TO NSP*
IP C(XK) > 0 THEN IS « IS + (0.5 * (C(Xfc) * CH%(X\) • CH*(X%)))
NEXT XX
IP IS > 100 THEN IS m 100
' avoids initial high IS
I*********************
' Improve trial values
FOR X«s = 2 TO SO
IF T(X%> " 0 THEN WWSR3L2
CONCFACTOR «= T(X%)/TCALC(X%)
-------
990 E. TIPPING
A (XX) • A(XX) • SQR(CONCPACTOR) ' cation activities
WWSR3L2:
NEXT XX •
FOR XX « 52 TO 100
IP T(XX) - 0 THEN WWSR3L3
CONCFACTOR • T(XX)/TCALC(XX)
A (XX) - A(XX) * CONCFACTOR ' anion activities
WWSR3L3J
NEXT XX
• Improve ZED values only whan itaration no. is multiple of 2 (see HWSR5)
IF TH(1) •> 0 AND TESTITER • TESTITBRX THEN.
ZED(l) - ZEO(l) 4 ((ZCALC(1)-ZED(1))/S)
IF TH(2) > 0 AND TESTITER - TESTITER* THEN.
ZED(2) - ZED(2) 4 ((ZCALC(2)-ZED(2))/5)
LASTICHRATIO « LASTCHRATIO ' record values to avoid
LASTCKRATIO - CHRATIO ' finding complete *oln. in
' firit phase - see below
' Calculate total concns of master apecies
GOSVB WWSR4 .
••••A****************************
' output current status to screen
LOCATE 20
PRINT NOMIT\;TAB(13)/USING "ft*.«»";PH;TAB(28)|USINC "f .««»***-;IS;TAB(44);
USING "*».««t";CHRATIO;TAB(60);DPH
' Test whether CHRATIO has bean found to acceptable precision in first phase
IF CRTEST$ - •PRASE1" AND.
(LASTCHRATIO / CHRATIO) > 0.9 AND.
(LASTCHRATIO / CHRATIO) < 1.1 AND.
(LASTICHRATIO / CHRATIO) > 0.9 AND.
(LASTICHRATIO / CHRATIO) < 1.1 THEN WWSR3L6 ' return
IF CRTEST$ « "PHASE1" AND.
LAST1CRRATIO - CHRATIO THEN HHSR3L6 ' oscillating - return
' Test cations, anions, Z's for convergence - note that final test
' is CHRATIO, dona in pH-adjusting routine
FOR X* - 2 TO 50 • cations
IF T(X*) « 0 THEN WWSR3M
CONCERR . (2*(T(X\) - TCALC(XX))/(T(X\) 4 TCALC(XX)))*2 ' error term
IF CONCERR > GRIT THEN WWSR3L1 ' re-iterate
VWSR3MI
NEXT XX
FOR X* • 52 TO 100 ' anions
IF T(X\) m 0 THEN WWSR3L5
CONCERR . <2*(T(XX) - TCAIiC(X\))/(T(X%) + TCALC(XX)))»2 ' error term
IF CONCERR > GRIT THEN KWSR3L1 ' re-iterate
WWSRSIiSt
NEXT X\
IF (TH{1) 4 TH(2)) - 0 THEN HHSR3L6 ' no humics * return
IF ZERR(l) > GRIT THEN WWSR3L1 » re-iterate
IF ZERR(2) > GRIT THEN KWSR3L1 ' re-iterate
WWSR3L6:
RETURN
-------
Equilibrium chemical spea'ation by WHAM
991
WWSR4*
•«
'Ǥ
'ft
' called by WWSR2 and WWSR3 - calls WWSRS
calls WWSRS to calc activities, cones of complexes and
amounts bound by FA and HA
sums to get total cale'd conens
*8
88
• gg calcs +ve and -ve charge and CHRATXO gg
'888888888888888888888888888888888888888888868888888888888888888888888888888888
• Calc activities and concna of inorganic complexes, amnts bound to FA 6 HA
GOSOB WWSRS
« jx> summations to obtain total calculated concns of master species
FOR XX « 1 TO 100
IF A(XX) « 0 THEN TCALC(XX) « 0
IF A(XX> « 0 THEM WWSR8L1
IF XX - 1 THEM WWSR8L1
IF XX -51 THEM WWSR8L1
TCAtiC(XX) - 0
FOR T* « 1 TO NSPX
IF MIX(TX) - XX THEN TCALC(XX) m
(VOLSOL
(CHCd.TCX)
IF
(DVOL(l)
(DVOL(2)
• XX THEN
(VOLSOL
(CHC(1,YX)
(CHC(2,YX)
(DVOL(l)
(DVOL(2)
IF
(VOLSOL
(CHC(1,TX)
(CHC(2,YX)
(DVOL(l)
(DVOL(2)
COX)
sixorxn_
S1X(YXJ}_
CDDL(1,TX)
CDDL(2,YX)
TCALC(XX)_
«
SIX(YX)}.
SIX(YX))
C(IX)
CDDL(1,XX)
CDDL(2,TX)
• S2X(XX)).
* 82X(YX)).
* S2X(TX)>
m XX THEN TCALC(XX) « TCALC(XX)_
C(TX)
S3X(TX»_
CDDL(2,Tf\)
* 63X(XX)).
S3X(TX».
S3X(TX))
NEXT TX
WWSRSLll
NEXT XX
preparing to sum
' solution
' HS bound (1)
' HS bound (2)
' DDL (1)
' DDL (2)
' solution
' HS bound (1)
' HS bound (2)
' DDL (1)
' DDL (2)
' solution
• HS bound (1)
' HS bound (2)
' DDL (1)
' DDL (2)
•*****•*•••*••«••••*••***••••**•«**«•*•*•••
• Calculate +ve and -ve charges and CHRATXO
POSCH
NEGCH
' preparing to sum
FOR XX » 1 TO NSPX
IF CHX(XX) > 0 THEN POSCH • POSCH 4 (C(XX) * CBX(XX))
IF CHX(XX) < 0 THEN NEGCH « NEGCH - (C(XX) * CHX(XX))
NEXT XX
CHRATXO
RETURN
(POSCH/NECCH)
CAGtOHM-G
-------
992
E TIPPING
•S88888S8Sese888Se8eSS«e88fif888S8e8888SI8S8f888teil88fflS88S8t8*8ll8S8«8*8SStl8
WWSRS: • called by WWSR4 - calls KWSR6 and KWSR7 §*
•ftft calcs OB- activity, act coeffs, CO3,2- if pCO2 fixed f§
'iff calcs activities, cone* of inorganic complexes |§
•«» every 2nd itertn, call* WWSR6 to gat HCT'« and Z's fox HS §|
«ft call* KWSR7 to gat binding by DIi accumulation (FA,HA) • ft
'ft8ft888888888ft8888ee8888e8888ft88888e88f8flf888i8litei88lf888tl888888iee8888*t8
' Calculate activity of OH- froa A(l), tamp, daltaH
LKW « - 14 4- (2935* (0.003354 - (1/TEMPK)))
A(S1) « 10*(LKW) / A(l)
,•«*•«*•******•***•*********•****•*••***********••***•***••••
• Calculate activity coefficients using extended Dabye-Huckal
GAKHA(O)
ACTA
ACTS
ACTC
ACTD
GAUHA(l)
GAHMA(2)
GAMHA(3)
GAMMA (4)
1
0.270 4-
0.330
- ACTA
ACTB
10* (
10* ( 4
10* ( 9
10* (16
(8E-4*TEMPK)
SQR (IS)
SQR (IS)
ACTC/(1 4- ( 3
ACTC/(1 + ( 6
ACTC/d + ( 9
ACTC/(1 «• (12
"
* ACTD)))
* ACTD)))
• ACTD)))
• ACTD)))
' const A in D-H
' const B in D-H
' act coaff M+/-
' act coaff M2+Y-
' act coaff M3+Y-
' act coaff M4+Y-
,***•****•«*•****•****•***•***•***********«***«
' Calculate concentrations of inorganic spacias
' First do the carbonate system
IF PC02 m 999 THEN WWSRSL1
ENTHTERM - 220 * (-0.962) * (0.003354 - (1/TEHPK))
A(55) m CC02 / 10*(18.149 4- ENTHTERM) / A(l) / A(l)
pCO2 not specified
activity of CO32-
KWSR5L1J
FOR X* - 101 TO NSP*
IF S1X(X*) - 0 THEN WWSR5L2
IF S1\(X\) > 0 AND A(H1\(X\)) • 0 THEN MWSR5L2
IF S2\(X«i) > 0 AND A(H2\(X«c)) - 0 THEN WWSR5L2
IF S3\(Z%) > 0 AND A(H3X(X^)) « 0 THEN VWSRSL2
ENTHTERM - 220 * DH(») * (0.003354 - (1/TEMPK))
LOCACT - U:(Z\) 4- ENTHTERM
XiOQACT m LOOACT + (El\(») * LOO10(A(M1\(X\) ) ) )
LOOACT - LOCACT + (82\(»() . * IiOO10(A(H2%(ZV) )))
IF S3MX*) > 0 THEN_
IiOGACT - LOGACT +
•A(X\) « 10* (LOGACT)
WWSR5L2»
NEXT X\
* IiOO10(A(M3X(X*) ) ) )
' complex not defined
' no calculation if
' contributing species
' absent
' Calculate concns from activities
FOR X* « 1 TO NSP*
CHARGE* - ABS(CH*(X\))
C(X*) - A(X\) / GAMKA(CHARGE*)
NEXT X*
-------
Equilibrium chemical spcciation by WHAM 993
»•*•«•***••••**»**•**••********•****•****•***•••••••
' Calculate concentrations of bumic-bound components
' Chock if tbe iteration no. (NOKIT*) is a multiple of 2 » if not, return
' Noto that TBSTITER, TESTITER* are also used in subroutine WWSR3
TESTITER m NOMIT*/2
TBSTITER* - INT(NOKXT*/2)
IP TESTITER > TESTITER* THEM NHSRSIiS » no oalc thi« iter; return
FOR BS* « 1 TO 2
IF TB(HS\) « 0 THEN WWSR5L3 • no calcn if BS absant
GOSUB KWSR6 ' calculate* ZCALC'B and NU's
' Calculate concns of epacifically bound cpacie* per litre from Ha and [HS]
FOR X* - 1 TO NSP*
CHC(HS\,X\) - OT(HS*,X*) * TH(HS\)
NEXT X*
WWSR5L3t
NEXT HS\
' Calculate maximum volume• of HA and FA diffuse layarc
FOR HS* « 1 TO 2
IF TH(HS\) • 0 THEN MWSR5L4 • no calcn if HS absent
VTERM1 m RADIUS (HS*) * (3.04E-10 / SQR(IS))
VTERM2 « (VTERM1*3) - ((RADIUS(HS*) )*3)
VTERM3 « 4.19 * VTERH2
DDLVOL(HS*) m «E23 * VTBRM3 * (1000 / MOLWT(HS*)) ' litre«/gHS
DVOLMAX(HS*) « DDLVOL(HS*) * TH(HS*J
• Adjust diffuse layer volume for low ZED
ZTERH « KZED * ABS(ZED(HS*) )
DVOLMAX(HS*) « DVOLKAX(HS*) * ZTERH / (1 4 ZTERM) • max Vol, litres/litre
MWSRSLis
NEXT HS*
' Calculate tbe actual diffuse layer volumes, using DLF
DBNOK • 1 + ((DVOLKAX(l) * DVOUSAX(2)) / DIiF)
FOR HS* « 1 TO 2
DVOXitHS*) « DVOUtAX(HS*) / DENOM
NEXT HS*
yOLSOL « 1 - DVOL(l) - DVOXi(2)
FOR BS* • 1 TO 2
IFTH(HS*) > 0 THEN COSUB WWSR7 ' Calc DDL concns
NEXT BS*
VWSRSLS:
RETURN
-------
994
E TIPPING
WWSR6* ' called by WWSR5
*gg calcs NO (specific binding) and Z for FA, HA
if
ff
W • P(HS\) * LOGIC(IS)
e'static intaraction factor
FOR IX - 1 TO 8
TEMPVAL • KH(HS*,X*)*BXP(2*W*i:BD(HSX))/A(l)
FP(I*) • 1/(1 4 TBMPVAI.)
NEXT I*
protonation factors
,«*********»**•**••*****••**
« Binding at bidentate sites
FOR J% - 1 TO 12
SYTB1* - SITB\(0\,1)
SYTE2* • SITE*(0*,2)
' do each site in turn
' identifier proton Bites that
' aake up the bidentate sites
SOMBITERM - 1
FOR KV - 1 TO NSPX
IF BINDTEST*(K*) » 0 THEN KHSR6Z.1
TEMFVMi » 2*W*ZED(HS\)*(2 - CB\(K%))
TEMPVXL - BXP(TEMFVAXt)
TEMPVM. - KMH{HS%,6YTEl%,K%)«KMH(HS\
BIBTERH(K^) • TEMFVXI.*FP(8TTE1\)*F?(8TTE2«()/(A(1) A2)
SOMBITERM « COMBITERM + BIBTERU(KV)
WWSR6I.lt
NEXT Kfc
SOMBITEBTA « 0
BIMETCH • 0
FOR KX • 1 TO NSP\
IF BINDTESTMKM - 0 THEN MH8R6L2
BITBETA • BIBTERMOCK)/SOKBITERM
SOMBITHETA • SOMBITHBTA 4- BITEETA
BISNO(J\,K«t) « BITBETA*0M(HS\,JX)
BIMETCH - BIMETCH * (BITHETA*C8*(K*)}
WWSR6L2t
NEXT Kk i
' preparing to sum
' theta's and charges
PROT2CH - 2*FP(STTE1\)*FP(STCTE2M * (1 - SOMBITHBTA)
PROT1CH1 m FP(srTEl^)*(l - FP(8TTE2\)) * (1 - SOMBITHBTA)
PROT1CH2 - FP(SYTE2M*(1 - PP(STTE1\)) • (1 - SOMBITHBTA)
BINETCH - BIMBTCH 4 PROT2CH 4 PROT1CH1 4 PROT1CH2 - 2
BIZ(J\) - BINETCH • 8M(HS\,J\)
NEXT 0*
2 bound B4
H4 bound at site 1
H4 bound at site 2
net charge
this site's charge
' Now calculate total amounts bound, and total net charge (bidentate sites)
FOR K* • 1 TO NSP*
IF BINDTESTMKfc) - 0 THEN WWSR6L3
BINO(K*) m 0
FOR OX - 1 TO 12
BINO(tfk) • BINO(KX) 4 BISNO(J\,K*)
NEXT JX
WWSR6L3i
NEXT K%
-------
Equilibrium chemical speciation by WHAM 995
BXZCALC « 0
FOR 3\ • 1 TO 12
8XZCALC m BXZCALC 4- BXZ(0*)
NEXT 3*
' Binding at monodantata *ita*
FOR J\ « 1 TO 8
SUKMONTERM • 1
FOR K* « 1 to NSP*
IF BINDTESTMK*) - 0 THEN WWSR6M
TBMPVAI, • 2«W*ZBD{HS*)*(1 - CH*(KM)
TEMPVM. « BXP(TEMPVMi)
TEMPVAL - KMR(HSX,J*,K*)*A(KX)*TKMFVAL
MONBTERM(KX) « TEMPVM.*FP(0\)/X(1>
SOKMONTERM - SOMMONTERM + MONBTERM(K\)
WWSR6L4I
NEXT r\ ''
SUKMONTHETA • 0 ' px«paring.to BUB
MONMETCH • 0 ' thata'c and charges
FOR K\ - 1 TO NSP\
IF BXNDTEST*(Kfc) - 0 THEN WWSR6L5
KONTHETA - KONBTBRM(K«()/8tniMONTERK
SUKMONTHETA • SUMMONTHETA 4 KONTHETA
MONSNa(J\,K\) - MONTHETA«SP(HS\,OX)
MONMETCH • MONMETCH «f (MONTHBTA*CH\(r*))
WWSR6L5J
NEXT K%
PROT1CH « FP(J\) * (1 - SUMMONTHETA) • H+ bound
MONNETCH - MONMBTCH •«• PROT1CH - 1 • nat chargo
MONZ(JV) - MONNETCH * SP(HS\,J\)
NEXT* J\
' Now calculata total amount* bound, and total nat eharga (nonodantata citas)
FOR K«t - 1 TO NSP*
IF BINDTEST«s(K\) • 0 THEN KWSR6I.6
MONNU(lCTt) - 0
FOR O\ « 1 TO 8
MONNO(K*) « MONND(K%) 4- KONSNa(OX,K*)
NEXT 0\
MHSR€Ii6t
NEXT let
KONZCALC « 0
FOR 0\ » 1 TO 8
MONZCALC « MONZCALC 4 MONZ(JX)
NEXT J*
' Overall summation ; bidantate 4 nonodantata
FOR K* • 1 TO NSP\
IF BXNDTEST%(K*) « 0 THEN WWSR6I.7
NO(HS%,K%) - BINU(K%) 4
VWSR6L7:
NEXT K%
-------
9% E. TIPPING
• Calculated value of Z, and Z error term
ZCALC(HSX) - BIZCALC 4- KONZCALC
ZERR(HSX) « (2MZED(HS*)-ZCALC(HSX))/(ZED(HSX)+ZCALC(HSX)))*2
RETURN
WWSR7t ' called by WH8RS fff
'ft calcs binding by DL accumulation, for FA and HA ff
•f*fi868S«8«S*888S8S«8S88t888ltt«fmi8ftfftff88t6tffeSff888t8*888i8«88tt88i8f88
•' First clear the DDL array for this HS%
FOR J\ - 1 TO NSPX
CDDL(HSX,JX) • 0
NEXT OX
TOTCHN - ZED(KSX) * TH(HSX)
TDCONC • - TOTCHN / DVOL(HSX)
IF TDCONC < 0 THEN WWSR7L1
' total charge to be
' neutralised, per litre
' total cone of counterionc
' per litre of diffuse layer
' anions attracted
' Come to here if hunics hare a net negative charge (cations attracted)
WWSR7L2t
TDCONCCALC « 0
FOR X* - 1 TO NSPX
IF CHX(X\) < 0 THEM CDDL(HS\,X\) - 0
IF CB\(X\) < 0 THEN KWSR7M
CDDXi(HS*,X*) • C(X\) * (RATIO(HSX)*(CH\(XX)))
TDCONCCALC • TDCONCCALC + (CDDL(HS*,X\)*CH\(X*))
WWSR7L3i
NEXT XX
TDCONCERR • (2 * (TDCONC - TDCONCCALC) / (TDCONC + TDCONCCALC)) *2
IF TDCONCERR < CRIT THEN WWSR7L5
' Adjust ratio and re-try
RATIO (HSX) « ((RATIO (H8\) * TDCONC / TDCONCCALC) + RATIO (HS\))/2
GOTO WWSR7L2
' Come to here if humics have a net positive charge (anions attracted)
WWSR7LJ.«
TDCONCCALC « 0
FOR XX - 1 TO NSP*
IF CHX(XX) > 0 THEN CDDL(KS\,XX) » 0
IF CHX(XIS) > 0 THEN WWSR7L4
CDDL(H8\,X\) m C(XX) * (RATIO (HS\) » (-CH\(X\) ))
TDCONCCALC • TDCONCCALC + (CDDL(HS\,XX)*CHMXX) )
NEXT XX
TDCONCERR • (2 * (TDCONC - TDCONCCALC) / (TDCONC + TDCONCCALC) )*2
IF TDCONCERR < CRIT THEN WWSR7L5
-------
Equilibrium,chemical speciation by WHAM 997
' Adjust' ratio and re-try
RATIO(HSX) - ((RATIO(HSX) * TDCONC / TDCONCCALC) + RATIO(HSX))/2
GOTO WWSR7L1
WWSR7L5i
RETURN
'•88i8S88888e8888tSS*8888t*l8888e8S88t88l88if88888888f8888S8688SS«88888888tS8S8
WWSRBs ' callsd by nain program - call* WWSRS fg
»f& nak«B output file f|
•f8eS888888888*8e88888IS888S8*8SS888886SI8*S88SS8888688888«8S88888888888S8888St
DIVIDERS • _
FF « 1.00001
PRINT *2, DIVIDERS
PRINT ft2, _
format factor ; avoids aaioy output
•••*••*•*•
DIVIDERS
OUTPUT FILE FROM WHAM-W VERSION 1.0
*.*******••
PRINT «2.
PRINT «2,
PRINT *2,
PRINT «2,
IF PHFIXS
IF PHFIXS
IF PC02 m
IF PCO2 •
PRINT §2,
PRINT «2,
PRINT ft2,
PRINT *2,
PRINT »2,
PRINT §2, "INPUT DATA11
PRINT *2,
PRINT ff2,
PRINT f2,
PRINT f2,
PRINT *2,
•SOURCE FILE •;TAB(25);SF$
•DATABASE -;TAB(25);DBS$
« -NO" THEN PRINT *2,"PHB;TAB(25);"VARIABLE-
« "YES" THEN PRINT f2,"PH";TAB(2S);"FIXED"
999 THEN PRINT 82,«PCO2";TAB(25);"VARIABLE"
999 THEN PRINT «2,«PCO2";TAB(25);"FIXED"
•PRECISION* ";TAB(25)/USING •*.§»!**A*-|FF*PRBCISION
•STARTING PH ";TAB(25)jUSINO "f.fif****"jFF*PHSTART
DIVIDERS
"TEMPK"; TAB(25);USINO "t.fI«****";FF*TEMPK
•TOT HA"jTAB(25)/USING "•.ifi****"/FF*TH(l)
•TOT FA"/TAB(25)/USING -f .t«****-;FF*TH(2)
"PC02"/ TAB(25)/USING •|.iti**»*"/FF»PCO2
PRINT §2,
PRINT «2, "MASTER SPECIES"/TAB(25)/"TOTAL CONC"
FOR XX • 1 TO 100
IF Xt » 1 THEN WWSR8L2 -,
IF XX « 51 THEN WWSR8L2
IF T(XX) •'0 THEN WWSR8L2
PRINT «2,X\/TAB(5)/N$(XX)/TAB(25)/USING "i.ftf»***"/FF*T(XX)
WWSR8L2J
NEXT XX
PRINT §2, •
PRINT f2, DIVIDERS
PRINT §2,
PRINT 12, •RESULTS"
PRINT §2,
PRINT §2,
PRINT §2,
PRINT 12,
PRINT *2,
/NDKITX
•NO. OF ITERATIONS'/TAB(22) /USING ••§§§! "
"PH-| TAB(2S)/USIHO «f .
•IONIC STRENGTH"/ TAB (25) /USING "i.f If **»*-/FF*IS
•CHARGE RATIO*; TAB(2S);USINO "f.f lt****";FF*POSCH/»BGCH
PRINT §2, "CHARGE DIFFERENCE";TAB(24)/USING -+t.itiA***"/FF*POSCH-NEGCH
IF TH(1) > 0 THEN PRINT §2,
IF TH(2) > 0 THEN PRINT 12,
IF TH(1) > 0 THEN PRINT 12,
IF TH(2) > 0 THEN PRINT «2,
PRINT 82,
PRINT 82,"WATER VOLUMES"
PRINT #2,"FRACTION HA-DDL •
PRINT 82,"FRACTION FA-DDL •
PRINT #2,"FRACTION SOLUTION «;TAB(25)/USING
PRINT #2,
PRINT 82,"CARBONATE ALKALINITY "/TAB(24)/USING "+#.#ftSA*A*"/FF*_
(C(101) + (2*C(5S)) * C(51) - C(l)) * VOLSOL
•ZBD-HA"/TAB(24)/USING «+t.f«****";FF*ZBD(1)
«ZED-FA"/TAB(24)/USING "+i.f»|**»*"/FF*ZBD(2)
•RXTIO-HA«/TAB(25)/USING "i.«l***A"/FF*RATIO(l)
"RATIO-FA"/TAB(25)/USING "ft.iff****";FF«RATIO(2)
;TAB(2Sl/USING "t.-f#»****«/FF*DVOL(l)
;TAB(25)/USING "i.«##****"/FF*DVOL(2)
•*.*##AA**«/FF*VOLSOL
-------
998
E- TIPPING
DIVIDERS
DIVIDERS
"CONCENTRATIONS AND ACTIVITIES-
PRINT #2,
PRINT 82,
PRINT 82,
PRINT 82, -FRACTIONAL DISTRIBUTIONS OF MASTER SPECIES"
PRINT 82,
PRINT *2, "MASTER SPECIES" ;TAB{23);"S";TAB(31);"DHA»>TAB(40) ;"DFA«/_
TAB(49);"KA";TAB(S8); "FA-
PRINT 82,
FOR XX - 1 TO 100
IF XX - 1 THEN WWSR8L3
IF XX • 51 THEN WWSR8L3
IF T(XX) « 0 THEN WWSR8L3
GOSUB WWSR9 ,
PRINT 82,XX;TAB(7);N$ (XX) ;TAB(21) /USING "8.881 -;FRACS;TAB(30) /FRACDHA;
TAB(39) /FRACDFA/TAB(48) /FRACHA;TAB<57) ;FRACFA
WWSR8L3I
NEXT XX
PRINT §2,
PRINT 82,
PRINT 82,
PRINT 82,
PRINT 82,
PRINT 82, TAB (25) t "TOTAL HATER" /_
TAB(Si); "SOLUTION PHASE"
PRINT 82, "SPECIES"/TAB(22) ; " [FREE] -;TAB(33) ; • [ORGANIC] ";
TAB (52) / -CONC"/TAB{€3) / "ACTIVITY-
PRINT 82,
FOR XX « 1 TO NSPX
IF C(XX) - 0 THEN WWSR8L4
PRINT 82,XX/TAB<7);N$(XX)/TAB(21)/USING "8.888*»**-/FF*VOLSOL*C(XX) /TAB(33) /
FF*«DVOL(1)*CDDL(1,XX)) + (DVOL(2)*CDDL(2,XX) ) + CHC(1,X*) + CHC(2,XX));
TAB(50);FF*C(XX);TAB(62);FF*A(XX)
WWSR8L4:
NEXT XX
PRINT 82,
PRINT 82, DIVIDERS
PRINT 82,
PRINT 82,
PRINT 82,
PRINT 82,
PRINT 82, TAB(22);"DNa-HA«;TAB(34);"DNa-FA";TAB(46);"Na-HA»;TAB(58);"NU-FA"
PRINT 82,
FOR XX « 1 TO NSPX
IF TR(1) - 0 THEN WWSR8L5
DNO1 « DVOL(1)«CDDL(1,XX)/TH(1)
WWSRBLSt
IF TH(2) - 0 THEN WWSR8L6
DNU2 • DVOL(2)*CDDL(2,XX)/TH(2)
MHSRBL6t
IF DN01 - 0 AND DNU2 - 0 AND NU(1,X\) « 0 AND NU(2,XX) « 0 THEN KWSR8L7
PRINT «2,XX;TAB(7);N$(XX);TAB(21);USINO "«.888****"»FF*DNai|TAB(33)|_
FF«DNU2;TAB(45);FF«NU(1,XX);TAB(57)|FF«NO(2,XX)
WWSR8L7t
NEXT XX
PRINT 82,
PRINT 82, DIVIDERS
RETURN.
"VALUES OF NU (MOL BOUND / O HUMIC SUBSTANCES)"
"NO IS SPECIFIC BINDING j DNU IS DIFFUSE LATER-
'88888888888888888888888ft»888888ft8ft888f8if888t*8ft88*ftt88ft8fftft888888888888888888
VJWSR9t ' callod by WWSR8 88
'88 calcs distribution of chosen master species 88
'888818888888888888888888888888888888888888888888888888888888888888888888888888
TOTALS « 0
TOTALDKA •> 0
TOTALDFA = 0
TOTALHA « 0
TOTALFA » 0
-------
Equilibrium chemical speciation by WHAM
999
FOR TX " 1 TO NSPX
IF MIX(YX)
IF HIX(YX)
IF MIX(TCX)
IF MIX(YX)
IF HIX(YX)
IF M2X(YX)
IF M2X(YX)
IF M2X(YX)
IF M2X(YX)
IF M2X(YX)
IF M3X(YX)
IF M3X(YX)
IF M3X(YX)
IF M3X(*X)
IF M3X(YX)
XX THEN TOTALS
XX THEN TOTALDHA
XX THEN TOTALDFA
XX THEN TOTALHA
XX THEN TOTALFA
XX THEN TOTALS
XX THEN TOTALDHA
XX THEN TOTALDFA
XX THEN TOTALHA
XX THEN TOTALFA
XX THEN TOTALS
XX THEN TOTALDHA
XX THEN TOTALDFA
XX THEN TOTALHA
XX THEN TOTALFA
TOTALS 4
TOTALDHA 4
TOTALDFA 4
TOTALHA 4
TOTALFA 4
TOTALS 4
TOTALDHA 4
TOTALDFA 4
TOTALHA 4
TOTALFA 4
TOTALS 4
TOTALDHA 4
TOTALDFA 4
TOTALHA 4
TOTALFA 4
(VOLSOL
(DVOL(l)
(DVOL(2)
(CHC(1,YX)
(CHC(2,YX)
(VOLSOL
(DVOL(l)
(DVOL(2)
(CHC(1,YX)
(CHC(2,YX)
(VOLSOL
(DVOL(l)
(DVOL(2)
(CHC(1,YX)
(CHC(2,TtX)
C(Tflt)
CDDL(1,YX)
CDDL(2,YX)
SIX(YX) )
SIX(YX))
C(YX)
CDDL(1,X\)
CDDL(2,YX)
S2X(TX))
S2X(Y\))
C(Y%)
CDDL(1,YX)
CDDL(2,YX)
83\(YX) )
S3X(YX))
* SIX(YX))
* SIX(YX))
* SIX(YX))
* S2X(YX))
* S2X(TtX))
* S2X(YX))
• S3X(YX))
* S3X(YX))
NEXT XX
FRACS « TOTALS / TCALC(XX)
FRACDHA • TOTALDHA / TCALC(XX)
FRACDFA « TOTALDFA / TCALC(XX)
FRACHA - TOTALHA / TCALC(XX)
FRACFA • TOTALFA / TCALC(XX)
RBTORN
APPENDIX 3
WHAM-S Program Listing
'•* KKAH-S VERSION 1.0 if
'iff Computer coda in TURBO BASIC ' ft
'ft Sp«ci*tion of huaic-rlch codlmantc and •oil* if
•ff This version produced on. 13 April 1993 by E. Tipping, IFE, Hindenaara if
'iifffffffffffffffffiiifffftftfffftfitftfifiiiflffiiiiffiiiiiiiiffffiiifffffffffffff
CLS
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
•*«*•****•*•«••*******•
"** W indermere **
•** H umic **
•** A queous **
«** H odel **
•*• _ ••
• **
S oils
edlaents
*•
**
• ** *' **
«** Version 1.0 *•
•***«*«***•*«*«***••**
Equilibrium speciation of soils/sediments"
Busies (FA and/or HA) present"
Variable pH"
Fixed or variable pCO2"
Precipitation of A1(OH)3 6 F«(OH)3"
Clay cation-exchanger"
Selectivity for ion-exchange"
FA adsorption/desorption" •
E.Tipping IFE April 1993"
PRINT:INPUT "NAME OF SOURCE FILE [OMIT
PRINT
OPEN SF$ + ".DAT" FOR INPUT AS il
INPUT «1,S$,S$
INPUT #1,S$,DBS$
INPUT #1,S$,NSPX
.DAT TRAILER] s ",EF$
file header
database identifier
array size required
-------
1000
E. TIPPING
^«www«wwww*>»»« — ~ — — — — —
DIMension and door arrays, then read in chemical constants
GOSUB HSSR1
• Continue to read in data file
INPUT il,S$,PRECISION
INPUT il,S$,TBMPK
INPUT tl,S$,CSOLID
INPUT il,S$,TH(l),TH(2)
INPUT fl,S$,TCLAr
INPUT tl,S$,FAAQST
INPUT il,S$,PC02
INPUT tl,S$,NCOMP*
FOR I* - 1 TO HCOMPX
INPUT *1,S$,X*,TCONC
N$(X*> « 8$
T(X*) - TCONC
NEXT I*
CLOSE Cl
precision
temp K
•oil cone g/1
total cones (HA, FA)
total clay cone g/1
[FA] in aqueous phase
PC02
no. of naater species
total cones
>***•****•***•********
• Perform calculations
GOSUB WSSR2
IF FAXQST • 999 THEN GOSUB WSSR9
IF FAXQST < 999 THEN FAXQ « FA&QST
' calculate FA aorption
' aq [FA] provided
rA***********************************A***************
' calculations finished ; signal and make output file
' sound output
SOUND 400,5 s SOUND 500,5 s SOUND 700,8
OPEN SF$ + -.OUT" FOR OUTPUT AS §2
GOSUB WSSR10
CLOSE *2
PRINT
PRINT •CALCULATION FINISHED S CONSULT ** •»SF$J«.OUT *• FOR DETAILED OUTPUT"
STOP
END
HSSRls
'it
'§8
'**
called by &*in program
DIMensions arrays,
reads in data from data base
sets up Modal V
it
II
if
If
DIMension arrays ; inorganic chemistry
DIM
DIM N$(NSP\)
DIM T(NSP\)
DIM TCALC(NSP\)
DIM A(NSP%)
DIM C(NSP*)
DIM CH%(N3P\)
DIM M1MNSP*)
DIM M2%(NSP\)
DIM M3«t(NSP»6)
numerical identification of species
nominal identification of species
total concentrations of components
total calculated concentrations of components
activities of solution species
concentations of solution species
charges on solution species
master species identifiers
master species identifiers
master species identifiers
-------
Equilibrium chemical spcciation by WHAM
1001
DIM S1MNSP*)
DIM S2«t(NSP*)
DIM S3*(NSPX)
DIM tK(NSP*)
DIM DH(NSPX)
DIM CAMMA(4)
stoichiometries
stoichionietries
stoichiometries
equilibrium constants
enthalpies
activity coefficients
DIMonsion humic arrays ; 1 « HA, 2 « FA
DIM SP(2,0)
DIM SM(2,12)
DIM KH(2,8)
DIM PKH(2,8)
DIM KMB(2,8,NSP*)
DIM SITE*(12,2)
DIM PKMSA(2,NSP\)
DIM PKMEB(2,NSP*)
DIM BITBRTA(12,NSP*)
DIM MONTBETA(8,NSP*)
DIM FP(8)
DIM BIBTERM(NSP*)
DIM MONBTERM(NSP*)
DIM BISNU(12,NSP*)
DIM MONSNU(6,NSP*)
DIM BINU(NSP*)
DIM MONNU(NSP*)
DIM N(7(2,NSP*)
DIM CEC(2,NSP*)
DIM CDDI.(2,NSP*)
DIM BIZ(12)
DIM MONZ(8)
DIM BINDTBSTMNSP*}
DIM KSELHS(NSPX)
DIM FA(10)
proton-binding sites not forming bidentate sites
bidentate metal-bindings
K'm for proton-binding
PR's for proton-binding
K's for metal-proton exchange
proton sites making bidentate sites
metal-proton exchange constants
metal-proton exchange constants
theta's for bidentate sites
theta's for monodentate sites
dissociation factors for proton-binding
terms calc'd in finding bidentate theta's
terms calc'd in finding monodentate theta's
values of NU at each bidentate site
values of NO at each monodentate site
overall NU for each metal/bidentate sites"
overall NO for each metal/monodentate sites
overall NO for each complexed species
cone of coaplexed species per litre
cones in ES DDIi's
net charge at each bidentate site
net charge at each monodentate site
allows binding of a species to be tested
selectivity coefficients for HA/FA DDIi's
FA fractions for sorption model
' DIMension humic arrays for clay mineral ion-exchanger
DIM KCLXY(NSPX)
DIM CLAYDDMNSP*)
' selectivity coefficient
' cone of species in clay DDL, per litre
'Input of constants for inorganic .speciation and ES complexation
OPEN DBS$ 4- ".DBS" FOR INPUT AS §2
IKPOT *2,S$
INPUT ft2,S$,NCOOH(l),PKHA(l),PKKB(l),DPKKA(l),.
DPKHB(l),P(l),FPR(l),ItADIUS(l),MOIiHT(l) '
INPUT «2,S$,HCOOH(2),PKBA(2),PKHB(2),DPKHA(2),
DPKHB(2),P(2),FPR(2),RADIUS(2),MOUfT(2)
INPUT i2,S$,EQCIAY,SACLAY
INPUT t2,S$,DLF
INPUT •2,S$,LKSOAI>25,DHAI.
INPUT •2,S$,LKSOFE25,DHFE
INPUT t2,E$,KZED
INPUT *2,S$,'GAMMA,KO,BETA
INPUT 82,S$,NODATA\
' database identifier
' HA properties
FA properties
clay properties
double layer overlap factor
A1(OB)3 ppt
Fe(OH)3 ppt
DL vol at low Z
FA sorption constants
number of species
FOR IX • 1 TO NODATA*
INPUT «2,X\,N$(X\) ,CH\(X»() ,Ml\(X«t) ,M2*(X*) ,M3«S(X»t) ._
S1\(X\),S2*(X%),83%(X\),LK(X\),DH(XX),
PKMHA(1,X%),PKMHA(2,X%),KSELHS(X%),
IF PKMHA(1,X%) < 999 TEEN BINDTEST%(X%)
NEXT I*
identifies species that
don't bind to HS
CLOSE #2
-------
1002
E. TIPPING
' Sot derived constants for Modal V
FOR XX - 1 TO 2
PKH(XX,1)
PKH(XX,2)
PKH(XX,3)
PKH(XX,4)
PKH(XX,5)
PKH(XX,«)
PKH(XX,7)
PKH(XX,8).
PKHA(XX) - (DPKHA(XX)/2)
PKHX(XX) - (DPKHA(XX) /6)
PKHA(XX) +
-------
Equilibrium chemical spetiation by WHAM
1003
FOR JX » 5 TO 8
PKHH • (3*PKMHA(1,IX)} - 3
KMH(1,3X,IX) m 10M-PKMH)
PKHH - 3.96*PKHHA(2,IX)
KMH(2,JX,IX) « 10M-PKMH)
NEXT OX
GOTO WSSR1L2
' If come to hero, than no binding of species IX can occur
WSSRlLlt
FOR OX - 1 TO 8
XMH(1,OX,IX) « 0
KMH(2,JX,IX) - 0
NEXT JX
WSSR1L2*
NEXT IX
RETURN
' 3*A - 3 conversion ; HA
• 3.96*A conversion ; FA
HSSR2t
'if
'tft
•iff .
'iff
•ii
•ii
' called from main program - call* HSSR3 and HSSR4
sets initial trial value* of master species
initialises all activities and concns by calling WSSR4
control! level of precision
call* WSSR3 to calc speciation
checks for pptn of AKOH)? and Fe(OB)3
sets up screen to report on progress of calculation
if
ff
f§
fff
ff
fi
ii
• Set initial trial values
IS « 0.1
ZED(l) - -1B-4
ZED(2) - -1B-4
RATIO (1) « 10
RATIO(2) m 10
RATIOCIAY • 10
A(l) - 10*(-4)
FOR XX - 2 TO 50
A(XX) - 1E-6 * T(XX)
NEXT XV
A(52) . T(S2)
A(53) - T(53)
A(S4) - T(S4)
A(56) » T(S€) « 1B-4
A(S7) " T(S7) * IE-IS
IF PCO2 « 999 THEN A(5S) - 1E-6 * T(5S)
' ionic strength
' Z for HA
' Z for FA
' RATIO for HA
' RATIO for FA
' RATIO for clay
' starting H+ activity
' species il is H
' cationic master species
' Cl
' N03
' 8O4
' F
' P04
' C03,2-
>*«****************«**«****••«******•**
• summary output to screen, and headers
LOCATE 14
PRINT "
LOCATE 14
PRINT « SOURCE FILE : ";SF$
PRINT "PRECISION (X) : ";OSINC "i.ii»***";1.0001«PRECISION
IF PC02 - 999 THEN PRINT « ' PC02 : VARIABLE"
IF PCO2 < 999 THEN PRINT " PCO2 : FIXED"
-------
(004 E TlpriNC
PRINT
PRINT •ITER";TAB(12);BPB-;TAB(24);"IS«;TAB(36);«CHRATIO";TAB(48)J
«FE PPT";TAB(60);«AL PPT";TAB(72); "ADJEXP"
>*********•*********
' Begin calculations
NUMIT* - 0 ' iteration counter
ALPPT$ « "NO" ' first speciate with
FEPPT$ « "NO" ' no pptn allowed
GRIT - 1B-4 s COSUB WSSR4 ' initialization
GRIT - (PRECISION/100) *2 * COSUB WSSR3 ' full speciation
• Check for precipitation of Fe(OB)3 and/or XI (OB) 3
LKSOFE . LKSOFE25 + (0.219 * DHFE * ((1/298) - U/TEMPK)))
IF T(ll) - 0 THEN WSSR2L1
LIAPFE - L0010(A(11)) + (3*PH)
IF LIAPFE >m LKSOFE THEN FEPPT$ • "YES"
IF LIAPFE >« LKSOFE THEN COSUB WSSR3 ' recalculate with Fa (OB) 3
HSSR2L1S
LKSOAL - LKSOAL25 + (0.219 * DBAL * ((1/298) - (1/TEHPK)))
IF T(5) - 0 THEN HSSR2L2
LIAPAL m LOG10(A(5)) + (3*PH)
IF LIAPAL >- LKSOAL THEN ALPPT$ • "YES"
IF LIAPAL >- LKSOAL THEN COSUB WSSR3 ' recalculate with A1(OH)3
WSSR2L2*
RETURN
HSSR3t ' called by WSSR3 - calls WSSR4 . ft
••8 controls improvement of activities and ZFA,ZBA il
'8ft calls WSSR4 to do nass and charge calculations , ' it
•ftt tests for aass and charge balances ftf
'•ft tests for convergence of ZFA, ZBA, ZCLAY ftft
'«» reports progress of calculation to screen ft!
'ftftffffftftftftftftft«ftiftfftftftftftiiftftftitftftffl«ftftftftftftftftffftffSftft«ftftftftft«ftfftftftftfftftiftftftffftftftffffftftftl
' Begin iterative cycle ; return to here until get convergence
WSSR3L1S
NOMIT* - NUMIT\ +1 ' update iteration counter
>*«*••****••*•*•«•***•*•**
' Calculate ionic strength .
IS • 0 ' preparing to sum
FOR X* - 1 TO NSP\
IF C(X\) > 0 THEN IS - IS + (0.5 • (C(») * CH\(X\) * CHX(X*)))
NEXT X*S
IF IS > 100 THEN IS - 100 ' avoids initial high IS
' Adjustment term ADJEXP; decreases with NUMIT* 6 with ionic strength
LADJEXP e ADJEXP t IF LADJEXP « 0 THEN LADJEXP « 0.5
ADJT1 « SQR(IS) : IF ADJT1 < 0.03 THEN ADJT1 « 0.03
ADJT2 = NUMIT^ / (3E4 * ADOT1) t IF ADJT2 > 0.48 THEN ADJT2 - 0.48
ADJEXP m 0.5 - ADOT2 : IF ADJEXP > LADJEXP THEN ADJEXP - LADJEXP
-------
Equilibrium chemical speciation by WHAM 1005
' improve A(l) (H+ activity) - begin after 10 iterations
A1FACTOR - (POSCH / ((POSCH + NECCH)/2))
IF NOMITX > 10 THEN A(l) « A(l) / (AlFACTOR'ADJEXP)
PH « -LOG10(A(D)
«•«***•*•***•••**•«*•***•*•*
* Improve cations and anions
FOR XX » 2 TO 50 ' cation*
IF T(XX) « 0 THEN WSSR3L2
IF XX « 5 AND ALPPT$ - "YES". , ' check for A1(OB)3 pptn .
THEN A(5) • (10*(LK£OAL)) • (A(l)*3) ' Al controlled by Al (OB) 3
IF XX - S AND ALPPT$ - "YES- THEN WSSR3I.2
IF XX • 11 AND FEPPT$ • "YES"_ • check Cor Fe(OH)3 pptn
THEN A(ll) » (10*(LKSOFE)) * (A(l)*3) ' F* controlled by Fe (OH) 3
IF XX • 11 AND FEPPT$ - "YES" THEN WSSR3L2
CONCFACTOR « T (XX) /TCALC (XX)
A(XX) - A (XX) * CONCFACTOR / (CHRATIO*ADOEXP)
WSSR3L2:
NEXT XX
FOR XX • 52 TO 100 ' anions
IF T(XX) - 0 THEN WSSR3L3
CONCFACTOR • T(XX) /TCAtiC(XX)
A(XX) « A(XX) * CONCFACTOR
HSSR3L3:
NEXT XX
«••«*•«***•*********
' Improve ZED values
IF TH(1) > 0 THEN ZED(l) - ZED(l) + ( (ZCALC(l)-ZED(l) )/30)
IF TH(2) > 0 THEN ZED(2) « ZED (2) + ( (ZCALC(2)-ZED(2) )/30)
•
******************************************
' Calculate total concns of master *pecies '
COS0B WSSR4
' Output current status to screen
XF FEPPT$ • "NO" THEN FEPPTX • 0
IF FEPPT$ « "YES" THEN FEPPTX • 1
IF AI«PPT$ - "NO" THEN ALPPTX • 0
IF ALPPT$ • "TES" THEN ALPPTX « 1
LOCATE 19
PRINT NOMITX7TAB(10)}USING "««.*«";PH;TAB(22) ;OSINO "t.f If *»»*";IS;TAB(35) ;
OSINO "«.«««|CHRATIO;TAB{50) /USING "I«;FEPPTX;TAB(62) ;AI.PPTX;
TAB(72);TTfiING "*.««";ADJEXP
Test cations, anions, charge-balance and Z'a for convergence
FOR XX « 2 TO 50 , cations
IF XX « 5 AND ALPPT$ « "YES" THEN WSSR3M • Al ppt
IF XX « 11 AND FEPPT$ « "YES" THEN WSSR3W, • p« ppt
IF T(XX) - 0 THEN WSSR3L4
CONCERR = (2*(T(XX) - TCALC(XX))/(T(XX) + TCALC(XX)))*2
-------
1006
E. TIPPING
IF CONCERR > CRIT THEN WSSR3L1
HSSR3L4s
NEXT XV
re-iterate
FOR XV • 52 TO 100 ' anions
IF T(XV) - 0 THEN WSSR3L5
CONCERR - (2MT(XV) - TCALC(XV) )/(T(XV) 4 TCALC(XV)))*2
IF CONCERR > CRIT THEN HSSR3L1 ' re-iterate
WSSR3L5t
NEXT XV
IF (CHRATIO - 1)*2 > CRIT THEN WSSR3L1
IF ZERR(l) > CRIT THEN WSSR3L1
' re-iterate
' re-iterate
IF ZERR(2) > CRIT THEN WSSR3L1
re-iterate
RETURN
WSSR4*
*ff
'ft •
'ft
'ft
called by HSSR2 and WSBR3 - calls WSSR5
calls WSSR5 to calc activities, cones of complexes and
amounts bound by FA/ HA. and CULT
sums to get total calc'd concns .
calcs 4ve and -ve charge and CHRATIO
ff
if
ft
ff
ff
'ffffffffffffftfffffffffffffffffffffffffffffffffffffffffffffftfffffffffffffttffftt
' Calc acts 6 concns of inorganic complexes, amnts bound to FA, HA, clay
GOSUB WSSR5
' Do summations to obtain total calculated concns of master species
FOR XV - 1 TO 100
IF A(XV) - 0 THEN TCALC(XV) - 0
IF A (XV) • 0 THEN WSSR4L1
IF XV « 1 THEN HSSR4L1
IF XV • 51 THEN WSSR4L1
TCAIiC(XV) - 0
FOR TCV « 1 TO NSPV
IF HIV(YV) - XV THEN TCI
+ (VOLSOL
- 4 (CHC(1,YV)
+ (CHC(2,-YV)
* (DVOL(l)
* (DVOL(2)
+ (DVOLCLAY
IF K2V(YV) - XV THEN T<
4 (VOLSOL
4 (CHC(1,YV)
4 (CHC(2,YV)
4 (DVOL(l)
4 (DVOL(2)
4 (DVOLCLAY
LLC(XV) • TCALC(XV)
C(YV) * SIV(YV))
S1V(YV»_
S1V(YV))_
CDDL(1,YV) * SIV(YV))
CDDL(2,YV) * SIV(YV))
CLAYDDL(YV) • SIV(YV))
^LC(XV) « TCALC(XV)
C(YV) • S2V(YV))
S2V(YV))_
S2V(YV))_
CDDL(l.YV) • S2V(YV))
CDDL(2,YV) * S2V(YV))
CLAYDDL(YV) • S2V(YV) )
' preparing to sum
solution
HS bound (1)
HS bound (2)
DDL (1)
DDL (2)
DDL clay
solution
HS bound
HS bound
DDL (1)
DDL (2)
DDL clay
(1)
(2)
-------
Equilibrium chemical speciation by WHAM
1007
IF
NEXT T\
HSSR4Xilt
NEXT X*
» X* THEN
(VOLSOL
(CHC(1,Y\)
(CKC(2,Hl)
(DVOIi(l)
(DVOI>(2)
(DVOIiCLAY
C(Y*)
CDDZ.(1,YX)
CDDX.(2,Y*)
' solution
' HS bound (1)
' HS bound (2)
' DDL (1)
' DDIi (2)
' DDL clay
' Calculate 4ve and -ve chargea and CHRATIO
POSCH - 0 s NEGCH - 0
FOR XX • 1 TO NSP*
IF CH*(X*) > 0 THEN POSCH » POSCH 4 (C(XX) '
IF CHMXM < 0 THEN NEGCH * NEGCH - (C(Xfc) * CH\(X*))
NEXT XV
CHRATIO - (POSCH/NEGCH)
RETURN
' preparing to aum
WSSRS*
*ff»
•ft
'«
'«
•9tt
called by N88R4 - calls WSSR6, WSSR7 and WSSR8
calca OH- activity* act coeffa, COS, 2- if pCO2 fixed
calca activities, conca of inorganic complexes
calls HSSR6 to get NU's and Z'a for HS
calls HSSR7 to get binding by DL accumulation (FA, HA)
calls WSSR8 to get binding by DZi accumulation (CIAT)
ffi
it
ft
if
if
it
8«ie*ff88e8e8t88888f88f8t888t888t8tl8fl888f88ttftf88888ft«88l888fff88l8teftM8«888
Calculate activity of OH- from A(l), temp, deltaB
« - 14 4 (2935* (0.003354 - (1/TEHPK)))
A(51) • 10*(LKW)
'•A**********************************************************
' Calculate activity coefficienta uaing extended Debye-Huckel
GAHMA(O)
ACTA
ACTB
ACTC
ACTD
GAHKA(l)
GAKKA(2)
CAKKAO)
GAMMA (4)
1
0.270 4
0.330
- ACTA
ACTB
10* (
10* ( 4
10* ( 9
10* (16
(8E-4*TEMPK)
SQR (IS)
SQR (18)
ACTC/U 4(3
ACTC/(1 4 ( 6
ACTC/d 4 f 9
ACTC/(1 4 (12
* ACTD)))
• ACTD)))
• ACTD)))
* ACTD)))
const A in D-H
const B in D-H
' act coeff M4/-
' act coeff M24/-
' act coeff K34/-
' act coeff H44/-
•**««************•****•*****••***•**«*****«****
' Calculate concentrations of inorganic apeciea
' First do the carbonate aystern
IF PCO2 - 999 THEN WSSR5L1
ENTHTERM - 220 * (-0.962) * (0.003354 - (1/TEMPK))
A(5S) m PC02 / 10*(18.149 4 ENTHTERM) / A(l) / A(l)
' pCO2 not specified
• activity of CO32-
CACtO »,»-H
-------
1008 E- TIPPING
WSSR5L1:
FOR X* - 101 TO NSP*
IF SI* (XX) - 0 THEN WSSR5L2
IF SIX(XX) > 0 AND A(M1*(X*) ) « 0 THEN HSSR5L2
IF S2*(X*) > 0 AND A(H2*(X*) ) • 0 THEN WSSR5L2
IF S3*(XX) > 0 AND A(K3X(XX) ) - 0 THEN HS8R5L2
BNTHTERM - 220 * DH(XX) * (0.003354 - (I/TOOK))
LOGACT m LK(X*) + BNTHTERM
LOGACT • LOGACT + (S1*(X*) * LOGIC (A (Ml* (»))))
LOGACT - LOGACT 4- (S2*(X*) * LOO10(A(M2*(X*)}) )
IF S3MXX) > 0 THEN.
LOGACT - LOGACT 4- (S3*(X*) • LOO10(A(M3*(X*) ) ) )
A(XX) » 10* (LOGACT)
/
WSSR5L2*
NEXT X*
' coaplax not dafinad
' no calculation if
' contributing •pocias
' abiant
' Calculate concns from «ctiviti«»
FOR X* • 1 TO NSPV
CHARGE* - ABS(CH*(X*))
C(X%) - A(X\) / CAUUA(CHARGE\)
NEXT Xt
' Calculate concentrations of humic-bound component*
FOR HS* - 1 TO 2
IF TH(HS*) • 0 THEN HSSR5L3
GOSUB WSSR6
' no calcn if HS ab>«nt
' calcc ZCALC'B and NO'*
' Calculate ooncns of specifically bound species per litre from NO and [HS]
FOR X* • 1 TO NSP*
CHC(HS\,X\) « NO(HS\,X\) * TH(H8\)
NEXT XX
W8SRSL3»
NEXT HSX
' Calculate maximua volunei of HA and PA diffuse layer*
FOR HS* - 1 TO 2
IF TH(HS*) - 0 THEN WSSR5L4 • no calcn if HS absent
VTERM1 . RADIUS(HS*) + (3.04E-10 / SQR(IS))
VTERK2 « (VTERM1*3) - ( (RADIUS (HS*) ) »3)
VTERH3 • 4.19 • VTBRH2
DDLVOL(HS*) « 6E23 * VTERM3 • (1000 / MOLWT(HS*))
DVOLUAX(HS*) • DDLVOL(RS*) * TH(HS*)
' litres/gHS
' Adjust diffuse layer volume for low ZED
ZTERM - KZED * ABS (ZED(HS*) )
DVOLMAX(KS*) « DVOLHAX(HS*) • ZTERM / (1 4 ZTERM)
' max vol, litres/litre
-------
Equilibrium chemical speciation by WHAM
1009
NSSRSL4:
NEXT HS\
• calculate maximum volume of clay mineral diffuse layer
IF TCLAY m 0 THEN NSSR5L5
DVOLMAXCLAY - 1E3 * TCIAY * SACLAY * (3.04E-10/gQR(IS)) ' litres/litre
WSSRSLSt
* Calculate the actual diffuse layer volumes, using DLP
DENOM « 1 + ((DVOLKAX(l) + DVOLMAX(2) 4 DVOLMAXCLAY) / DLF)
FOR HS\ « 1 TO 2
DVOL(HSX) - DVOLMAX(HS*) / DENOM
NEXT HS\
DVOLCLAY « DVOLMAXCLAY / DENOM
VOLSOL ml- DVOL(l) - DVOL(2) - DVOLCIAY
FOR HSX m 1 TO 2
IF TH(HS*) > 0 THEN GOSUB WSSR7
NEXT H8\
IF TCIAY > 0 THEN GOSOB WS8R8
RETURN '
' calculate HS DDL concns
' calculate clay DDL concn
WSSR6: • called by WSSR5 .
calcs N0 (specific binding) and Z for FA, fin
if
ii
H m P(HS\) * LOGIO(IS)
FOR I\ « 1 TO B
TEMFVAL - KH(HS\,I\)*EXP(2*W*ZED(HS\))/A(1)
FP(I*) - 1/(1 + TEMFVAL)
NEXT I*
' e'static interaction factor
protonation factors
' Binding at bidentate sites
FOR ax - 1 TO 12
SYTE1X - SITE*(3*,1}
SYTE2* - SITE*(JX,2)
SUMBITERM - 1
' do each site in turn
' identifies proton sites that
' Bake up the bidentate sites
FOR 1C* - 1 TO NSP*
IF BINDTESTMK*) • 0 THEN WSSR6L1
TEMFVAL « 2*W*2ED(HS\)«(2 - CH\(K\))
TEMFVAL « EXP(TEMFVAL)
TEMFVAL « KMH(HS\,SYTEH(/K\)*KMH(HSl6,SYTE2\/l?t)*A(K«t)*TEMPVAL
BIBTERM(Itt) « TEMPVAL*PP(SYTE1\)*FP(EYTE2%)/(A(1)A2)
SXJMBITERM m SDMBITERM + BIBTERM(K%) '
WSSR6L1:
NEXT KX
SDMBITHETA «: 0
BZMETCH «= 0
' preparing to sum
' theta'e and charges
-------
1010
E. TIPPING
FOR KX - 1 TO NSPX
IP BINDTESTX(KX) - 0 THEN HSSR6L2
BITHETA • BIBTBRM(KX)/SOMBITERM
STJMBITHETA • SOMBITHBTA 4- BITHBTA
BISNO(OX,KX) - BITHBTA«SM(HSX.OX)
BIHBTCH - BIMBTCH + (BITHBTA*CHX{KX))
WSSR6L2*
NEXT 1C*
PROT2CH • 2*FP(SYTB1X)*FP{STTB2X) * (1 - SUMBITHBTA) • 2 bound H+
PROT1CH1 • FP(SrTBl*)*U - FP(8YTB2X)) • (1 - SUMBITHBTA) ' H« bound at «it« 1
PROT1CH2 - FP(srTE2\)*(l - PP(SYTBIK)) • (1 - STJMBITHETA) ' B> bound at mit» 2
BINBTCH - BIKBTCB + PROT2CH 4- PROT1CH1 -f PROT1CH2 - 2 ' n«t chargo
BIZ(J%) - BINBTCH * SM(HSX,0\) ' thi« «it«'« eharga
NEXT O\
' Now calculate total amount* bound, and total not chargo (bidantat* sit«>)
FOR K% » 1 TO NSP\
IF BINDTESTMKt) " 0 THEN WSSR6L3
BINO(KX) - 0
FOR JX « 1 TO 12
BINO(IC\) » BINO(K«6) + BISNU(a\,K*)
NEXT J*
WSSR6Ii3t
NEXT 1C*
BIZCALC - 0
FOR OX - 1 TO 12
BIZCALC • BIZCALC + BIZ(JX)
NEXT J*
>*****•******•**««•****«******
' Binding at aonodantat* >it«i
FOR OX - 1 TO 8
SUHMONTERM • 1
FOR 1C* • 1 to RSP*
IF BINDTESTMKX) • 0 THEN WSSR6L4
TEMPVAL • 2*H*ZED(H8\)*(1 - CB*(K*) )
TEWPVAL » EXP(TEMPVAL)
TEMPVAL . KMH(HS*,0\,r*)*A(lC*)*TEMPVAL
MONBTBRM(K*) • TEMPVAL*FP(0*)/A(1)
8DHMONTBRH • SUMMONTERM + KOKBTERM(K*)
WSSR6L4t
NEXT K*
*
6OMMONTHETA • 0' • pr«pazing to cum
HONMBTCH • 0 » th«ta'« and charges
FOR K* « 1 TO N8P*
IF BINDTB3T*(1C*) • 0 THEN WSSR6L5
MONTHETA « MONBTERM(KX)/SXJMMOOTERM
SUMMONTHETA « SUMMONTHETA 4- MONTHETA
HONSHU (OX, KX) - KONTHETA*SP(RSX,OX)
MONMETCH - KONMETCH 4 (MONTHETA*CHX(KX) )
WSSR6L5:
NEXT KX
-------
Equilibrium chemical speciation by WHAM
1011
PROT1CH • PP(JX) * (1 - SUMMONTHBTA)
MONNETCH • KONMETCH + PROTlCH - 1
KONZ(JX) - MONNETCH * 8P(HSX,JX)
/ H+ bound
' net charge
NEXT OX
' Now calculate total amount* bound, and total net charge (monodentate sites)
FOR KX - 1 TO NSPX
IF BINDTBSTX(KX) " 0 TEEN WSSR6W
HONNU(KX) " 0
FOR J\ - 1 TO 8
HONND(KX) » MONNa(KX) 4- KONSNO(OX,KX)
MKXT JX
WSSR6L€«
NEXT K*
KONZCALC • 0
FOR OX - 1 TO 8
MONZOOiC « MONZCMiC 4- KONZ(OX)
NEXT JX
,>*«****•***••••**********•**•**••****•*****•*
• Ovarall aummation ; bidantata 4- aonodantata
FOR KX - 1 TO NSPX
IF BINOTBSTX(RX) • 0 THEN HSSR6L7
NO(HSX,KX) « BINa(KX) * MONNU(KX)
WSSR6L7*
NEXT KX
• Calculated valua of Z, and Z error tazm
ZCAXiC(HSX) « BIZCALC + MONZCALC
ZERR(HSX) . (2*(ZED(ESX)-ZCAZiC(B8X))/(ZED(E8»-fZCALC(HSX)))A2
RETURN
H88R7S ' called by KSSR5 ft
•ft calca binding by VL accumulation, for FA and HA ft
'«eifMitiiftttftiiiiiiiiiifMtiiiiiiiiitiitiiiiiiiitieii»tttftsteiiitiittiiii
* Firit clear the DDL array for thi« HSX
FOR OX • 1 TO NSPX
CDDL(H8X,JX) • 0
NEXT JX
TOTCHN - ZED(H8X) • TH(H8X)
TDCONC - - TOTCHN / DVOL(HSX)
IF TDCONC < 0 THEN WSSR7L1
• total charge to be
' neutralised, per litre
• total cone of counterions
' per litre of diffuse layer
anions attracted
' Come to here if humics have a net negative charge (cations attracted)
WSSR7L2:
TDCONCCALC B 0
-------
|OI2 E. TIPPING
FOR XX « 1 TO NSPX
IP CH*(XX) < 0 THEM CDDL(HSX,XX) • 0
IF CHX(XX) < 0 THEN WSSR7L3
CDDL(HSX,XX) - C(XX) • KSBLHS(XX) * (RATIO(HSX)*(CHX(XX)))
TDCONCCALC « TDCONCCALC + (CDDMHSX,XX) *CHX(XX))
WSSR7L3S
NEXT XX
TDCONCERR • (2 * (TDCONC - TDCONCCALC) / (TDCONC 4 TDCONCCALC))*2
IF TDCONCBRR < GRIT THEN WSSR7L5
' Adjust ratio and re-try
RATIO(HSX) '• ((RATIO(HSX) * TDCONC / TDCONCCALC) + RATIO(HSX))/2
GOTO KSSR7L2
««•«••••••••••**••*••••***•**********••***•***•**••**•*•••••••*•«*••*•
' Coma to bar* it bumics have a net positive charge (anions attracted)
WSSR7L1*
TDCONCCALC • 0
FOR XX - 1 TO NSPX
IF CHX(X\) > 0 THEN CDDL(HSX,XX) - 0
IF CH\(XX) > 0 THEN WSSR7L4
CDDL(HSX,XX) • C(X*) * KSELHS(XX) * (RATIO(HSX)*(-CHX(XX)»
TDCONCCALC - TDCONCCALC + (CDDL(HSX,XX)*CH\(XX))
WSSR7L4S
NEXT XX
TDCONCERR - (2 * (TDCONC - TDCONCCALC) / (TDCONC + TDCONCCALC))*2
IF TDCONCERR < GRIT THEN WSSR7L5
' Adjust ratio and re-try
RATIO(HS*) « ((RATIO(HSX) • TDCONC / TDCONCCALC) + RATIO(ESX))/2
GOTO WSSR7L1
WSSR7L5:
RETURN
HSSRSt ' called by W88R5 f|
'<• calcs binding by DL accumulation, tor CLAY if
' First clear the DDL array
FOR JX - 1 TO NSPX
CLATDDL(OK) - 0
NEXT OX
TOTCKN • - EQCLAY « TCLAY ' • total charge to be
'' neutralised, per litre
TDCONC « - TOTCKN / DVOLCIAY * total cone of counterions
' per litre of diffuse layer
WSSRBLli
TDCONCCALC • 0
FOR XX - 1 TO NSPX
CLAYDDL(XX) « C(XX) • KCLAY(XX) * (RATIOCLAY* (CHX(XX) ) )
TDCONCCALC • TDCONCCALC +
-------
Equilibrium chemical spetialion by WHAM (013
•Adjust ratio and re-try
RATIOCLAY • ((RATIOCLAY • TDCONC / TDCONCCALC) + RATXOCIAY)/2
ROTO MSSR8L1
GOTO WSSR8L1
KSSR8L2t
RETURN
'Cf««lltlS*»ft»ltlllttlltltlllttlitliltlllllllIfllitlftliifSI««8ICIISIftl88lllli
WSSR9i ' called by main program §§
««§ calcs distrbtn of FA between solid and aqueous phases' if
'CCSStSfCltltltlllllfitlllltllltflflllltltHftlttiiiflillltiltlfCCIMIIIflllllli
CFA « TH(2) / CSOLID ' 0FA par
,«•*•**«*«***«*•«**•***•**•*****•*•*
' Calculate distribution tarn for FA
DXSTSXCKA - 0
FOR I* - 1 TO 10
DISTSICMA « DZSTSZGHA + (I«t*OAMtt)
NEXT I*
<***********»**•**•****•**•»**********•*
'Calculate aquaou* cones of FA fractions
FOR I* « 1 TO 10
XFA - (0.1 • NCOOH<2) * 1%)
XFA - BETA * (XFA - ABS(ZBD(2)))
ZFA - 1 + (KO * EXP(XFA) * CflOLIO)
YFA - (X\*GAMMA) * CSOMD • CFA / DZ8T8ZOMA
FA(I%) - TFA / XFA-. '
NEXT I*
«**•******«******«****«*«**«•«
• sum to find total aquaous FA
FAAQ - 0
FOR I\ - 1 TO 10
FAAQ - FAAQ 4 FA(IH)
NEXT X%
StETORN
WSSRlOt ' < called by aain program - calls W88R11 and WS8R12 ft
'ft aakas output file ' f|
DIVIDER^ » _
•**«*•****•••«****«**«**««**•«•*****«•*««****•**•***••••*•***»•*******•••
FF « 1.00001 ' format factor ; avoids messy output
PRINT i2, DIVIDERS
PRINT f2,_
•***••***• OOTPDT FILE FROM WBAH-8 VERSION 1.0 *••*••*•••
PRINT «2, DIVIDER?
PRINT #2, "
PRINT «2, "SOURCE FILE ";TAB(2S) ;SF$
PRINT 82, "DATABASE ";TAB(25) ;DBS$
-------
1014
E. TlPHNC
IF PC02 « 999 THEN PRINT i2, "PCO2";TAB(25); •VARIABLE-
IP PC02 - 999 THEN PRINT f 2, «PCO2";TAB(25) ^FIXED-
PRINT 82, "PRECISION * ";TAB(25)/USING "8.888AAAA";FF«PRECISION
PRINT 82,
PRINT #2, PIVIOER$
PRINT 82,
"INPOT DATA"
PRINT #2,
PRINT 92,
PRINT 82, "TBMPK";
PRINT 82, "CSOLID";
PRINT 92. "TOT HA";
PRINT 92. "TOT FA";
».888AAAA";FF«TEMPK
».888AAAA";FF*CSOLID
f.888A**A"7FF*TH(l)
».8«AAAA";FF«TH(2)
.888AAAA";FF*TCLAT
TAB(25)/USING
TAB(25)/USING
TAB(25);USINO
TAB(25)/USING
PRINT 92, "TOT CLAY";TAB(2S);USINO
IF FAAQST « 999 THEN PRINT §2, "FAAQ-;TAB(25);" UNKNOWN"
IF FAAQST < 999 THEN PRINT 92, "FAAQ";TAB(25);USINO "8.888AAAA";FF*FAAQST
PRINT 92, "PC02"; TAB(25);0SINO "8.888AAAA";FF*PCO2
PRINT 92. '
PRINT 92. "MASTER SPECIES";TAB(25);"TOTAL CONG"
FOR XX « 1 TO 100
IF XX - 1 THEN WSSRIOIil
IF XX - 51 THEN HSSRlOZil
IF T(XX) « 0 THEN WSSR10L1
PRINT 82,XX;TAB(5);N$(XX>;TAB(25);USINO "8.888AAAA";FF*T(XX)
WSSR10L1:
NEXT XX
PRINT 92.
PRINT «2,DIVIDERS
PRINT 92.
PRINT 92, "RESULTS"
PRINT 92,
PRINT 82, "NO. OF ITERATIONS";TAB{23)/USING "88888"/NUMITX
PRINT 92, "PH"; TAB(25)/USING "8.888AAAA*;FF*PH
PRINT 92, "IONIC STRENGTH"; TAB(25)/USING "».fii****"/FF*IS
PRINT 92. "CHARGE RATIO"; TAB(25)/USINC "*.«8****"/FF*POSCH/NECCH
PRINT 92. "CHARGE DIFFERENCE";TAB(24)/USING "+t.tf8****-/FF*POSCH-KEGCH
IF FAAQST •> 999 THEN PRINT 82, "FAAQ (CALC)" ;TAB(25) /USING "8.888*A**-/FF*FAAQ
IF FAAQST < 999 THEN PRINT 82, "FAAQ (FIXED) "/TAB(25)/USING "8.888****"/FF*FAAQ
IF TH(1) > 0 THEN PRINT 82, "ZED-HA-/TAB(24)/USING "+8.888****"/FF*ZED(l)
IF TH(2) > 0 THEN PRINT 82, "ZED-FA"/TAB(24)/USING "+8.888A*AA"/FF*ZED(2)
> 0 THEN PRINT 82, "RATIO-HA" /TAB(25)/USING "8.8i8AAAA"/FF*RATIO(D
> 0 THEN PRINT 82, "RATIO-FA" /TAB(25)/USING "8.888AAAA";FF«RATIO(2)
»RATIO-CLAY-/TAB(25)/USING "8.888AAAA"rFF*RATIOCIAY
IF TH(1)
IF TH(2)
IF TCIiAY > 0 THEN PRINT 82,
PRINT 82,
PRINT 82,"KATER VOLUMES"
PRINT 82,"FRACTION HA-DDL
PRINT 82,"FRACTION FA-DDL
/TAB(25)/USING "8.888AAAA";FF*DVOL(D
/TAB(25)/USING "8.88*AA**"/FF*DVOL(2)
PRINT 82, "FRACTION CLAY-DDL "/TAB(25)/USING "8.888AAAA";FF*DVOLCLAY
PRINT 82,"FRACTION SOLUTION "/TAB(25)/USING "8.888AAAA";FF*VOLSOL
PRINT 82,
IF FBPPT$ - "YES" THEN PRINT 82,"[FE(OH)3 PPT]";TAB(25)/_
USING "8.888AAAA";*(11> - TCALC(ll)
IF FEPPT$ « "YES" THEN PRINT 82,"* TOT FE PPTD";TAB(24)/USING "888.88";
100*(T(11) - TCALC(11))/T(11)
IF T(ll) > 0 AND FEPPT$ • "NO" THEN PRINT 82,"FE(OH)3 SATN INDEX";TAB(24)/
USING «+8.888AAAA";LIAPFE - LKSOFE
IF ALPPT$ • "YES" THEN PRINT 82,"IAL(OH)3 PPT1«/TAB(2S);
USING "8.888AAAA";T(5) - TCALC(S)
IF ALPPT$ - "YES" THEN PRINT 82,"X TOT AL PPTD«;TAB(24)/USING "888.88";
100* (T(5) - TCAIiC(S))/T(5)
IF T(5) > 0 AND ALPPT$ • "NO" THEN PRINT 82,"AL(OH)3 SATN INDEX";TAB(24);_
USING "«8.888AAAA";LIAPAL - LKSOAL
PRINT 82,
PRINT 82, DIVIDER$
PRINT 82,
PRINT 82, "FRACTIONAL DISTRIBUTION OF MASTER SPECIES"
PRINT 82,
PRINT 82, "MASTER SPECIES-/TAB(23) /"S"/TAB(31);«DHA";TAB(40);"DFA";_
TAB (49)/"HA"/TAB(58)/-FA"/TAB(66)/"CLAY-
PRINT 82,
FOR X% m 1 TO 100
IF X% « 1 THEN HSSR10L2
IF XX « 51 THEN VJSSR10L2
-------
Equilibrium chemical spetialion by WHAM
1015
IF T(XX) - 0 THEN HSSR10L2
COSUB WSSRII
PRINT *2,XXjTAB(7)jN${XX);TAB(21);USING "#.«*8";FRACS;TAB(30);FRACDHA;_
TAB(39);FRACDFA;TAB(48) ;FRACHAfTAB(57) ;FRACFA;TAB(66) ;FRACDCLAY
WSSR10L2>
NEXT XX
PRINT 82,
PRINT 92, DIVIDERS
PRINT i2,
PRINT #2, "CONCENTRATIONS AND ACTIVITIES OF SPECIES IN SOLUTION PHASE"
PRINT #2,
PRINT *2,«SPBCIBS"»TAB(23);»CONC";TAB(3e);"ACTIVITT«
PRINT §2,
FOR X* m I TO NSPX
ZF C(XX) « 0 THEN WSSR10L3
PRINT «,XX;TAB(7)»N$(XX);TAB(21)|USINO «t.«l****"/C{XX);TAB(38) jA(XX)
HSSR10L3S
NEXT XX
PRINT i2,
PRINT f2, DIVIDER?
PRINT §2,
PRINT #2, "TOTAL AQUEOUS-PRASE CONCENTRATIONS AND KD VALUES"
PRINT f2,
PRINT 82, "SPECIES"»TAB(22); "AQCONC";TAB(42) ; "KD-;TAB(57) ; "KD-APP"
PRINT *2,
FOR X* » 1 TO 100
IF T(Xt) - 0 THEN WSSR10L4
IF XX - 1 THEN HSSR10L5
IF XX - 51 THEN WSSR10L5
GOSUB WSSRII
GOSUB WSSR12
WSSRIOLSt
PRIHT,«2,XX;TAB(7)»N$(X»0;TAB(21)jUSINO «f.«»**«*";AQCONCi
TAB{39)/USING "+fr.*«»**A";lE3*KD;TAB(S5);1E3*KDAPP
MSSRIOMs
NEXT XX
PRINT #2,
PRINT f2, DIVIDERS
PRINT §2,
RETURN
WSSRII» ' called by KSSR10 • ft
'ft calcs dlatribution of choc«n mast«r •p«ci«« if
'88l8e«88888888888888ie«e888888ie8888888888t8e88te8888888888ft888888888888e88888
TOTALS
TOTALDHA
TOTALDFA
TOTALDC
TOTALHA
TOTALFA
FOR *X « 1 TO NSPX
IF MIX(TX) '
IF KIX(TX)
IF HIX(TCX)
IF MIX(TX)
IF HlX(TfX)
IF KIX(YX)
IF M2X(YX)
IF M2X(7X)
IF K2X(YX)
IF M2X(yX)
IF H2X(YX)
IF M2X(yX)
XX THEN TOTALS •
XX THEN TOTALDHA
XX THEN TOTALDFA
XX THEN TOTALDC
XX THEN TOTALHA
XX THEN TOTALFA
XX THEN TOTALS
XX THEN TOTALDHA
XX THEN TOTALDFA
XX THEN TOTALDC
XX THEN TOTALHA
XX THEN TOTALFA
• TOTALS + (VOLSOL *
TOTALDHA 4 (DVOL(l)
TOTALDFA * (DVOL(2)
TOTALDC * (DVOLCLAT
TOTALHA + (CHC(l,tX)
TOTALFA 4 (CHC(2,YX)
TOTALS 4 (VOLSOL
TOTALDHA 4 (DVOL(l)
TOTALDFA 4 (DVOL(2)
TOTALDC 4 (DVOLCLAY
TOTALHA 4 (CHC(1,YX)
TOTALFA 4 (CHC(2,YX)
• C(*X) • SIX(TX))
CDDL(1,YX) * SIX(YX))
CDDIi(2fTX) * SIX(TX))
CLAYDDL(YX) • SIX(TX))
SIX(YX))
SIX(YX))
C(YX) * S2X(rX))
CDDL(1,YX) • S2X(YX))
CDDL(2,YX) * S2X(rX))
CLAYDDLCXX) * S2X(rX»
S2X(YX))
S2X(YX))
-------
1016
E. TIPPING
IF M3MYX)
IF M3MYX)
IF U3\(YM
IF M3*(Y*)
IF 113* (YX)
IF M3X(Y*)
X* THEN TOTALS
X* THEN TOTALDHA
X* THEN TOTALDFA
X* THEN TOTALDC
X* THEN TOTALHA
X* THEN TOTALFA
TOTALS +
TOTALDHA +
TOTALDFA +
TOTALDC *
TOTALHA +
TOTALFA +
(VOLSOL
(DVOL(l)
(DVOL(2)
(DVOLCLAY
(CHC(1,Y*)
(CHC(2,YX)
C(YX)
CDDL(1,Y*)
CDDL(2,Y\)
CLAYDDL(Y%)
S3*(Y\))
S3\(Y%))
S3\(Y\))
83*(Y*))
S3%(Y1t»
63\(Y\))
HEX! Y«t
FRACS
FRACDHA
FRACDFA
FRACOCLAY
FRACHA
FRACFA
TOTALS
TOTALDHA
TOTALDFA
TOTALDC
TOTALHA
TOTALFA
/ TCALC(XM
/ TCALC(XM
/ TCALC(XX)
/ TCALC(XX)
/ TCALC(Vt)
/ TCALC(rt)
RETURN
WSSR12J ' c&llad by KSSR10
'ft -^alc* total aguaou* eoo.com of aaatar •p«cias f|
'•ft calcB Kd valuaa . f|
'•8e88i8«888888«888l8f88«ll88S888888ff8888«8888l8«8888888S8888«»8it8Sff88l888t8«l
' Caleulata aquaous voluaa and total aolaa prasant
IF TH{2) > 0 THEN VOLAQ « VOLSOL / (1 - (DVOL(2)*FAAQ/TH(2)))
IF TH(2) • 0 THEN VOLAQ « VOLSOL
IF TH(2) > 0 THEN TOTAQ « TOTALS + ((TOTALFA + TOTALDFA)*FAAQ*VOLAQ/TH(2))
IF TH(2) « 0 THEN TOTAQ • TOTALS
AQCONC « TOTAQ / VOLAQ
TOTSOLID - T(X\) - TOTAQ
KD - (TOTSOLID / CSOLID) / AQCONC
' Apparent KD ; calculated Cor ALL vat«r baing in S (DDL ignored)
TOTSOLIDAPP • T(XV) - AQCONC
KDAPP « (TOTSOLIDAPP / CSOLID) / AQCONC
RETURN
'ifftftftiffftflftftiSftftlSfiffftliiftftftiftlfftftSftiftftftftSiiiilfftiiftSftiitftiffifliftiftillltifli
••ftftiftftftitffttfttt«ftftftftittftiittfttftftfftftftfttttfttftifftfttftHftlfttifttiltttftiiillftlliftt
APPENDIX 4
WHAM-W input and output files, water example. In the input file the master species' concentrations are in moles per
liter of total water in the system, and the concentrations of huraic and fulvic acids are in g I"1. The partial pressure of CO)
(pCO2) is in atmospheres. In the output file, concentrations of individual species, ionic strength, charge difference, and
carbonate alkalinity, are given in moll'1, concentrations of humic and fulvic acids in gl~*. The terms ZED-HA and
ZED-FA refer to the net charges on the fulvic and humic acid in equiv g'1. The terms RATIO-HA and RATIO-FA refer
to the variable R in Equation (8). The water volumes are in liters.
HATER ,WATERT1
Database ,HATER10
Database size ,400
Precision* ,0.01
TempK ,283
THA TFA g/1 ,lE-3,9E-3
-------
Equilibrium chemical speciation by WHAM
1017
pH fixed
PHSTART
pC02
No. mast.
Na
Kg
Ca
Ni
CU
Cl
S04
•P.
,YES
,7
,0.00035
,7
,3,0.0002
,4,0.0001
,7,0.0003
,13,18-fi
,14,18-6
,52,0.0005
,54,0.0002
•**•*•••**•••••****••*«•*•**•••*•*•****•••*•*••***•*•**••••*••*••*•••••
********* OUTPUT FILE PROM WHAM-W VERSION 1.0 •••*«••*•
A**********************************************************************
SOURCE FILE
DATABASE
PH
PRECISION *
STARTING PH
WATBRTl
KATER10
FIXED
l.OOOE-02
7.000E+00
INPUT DATA
TEHPK
TOT HA
TOT FA
PC02.
MASTER SPECIES
3 Na
4 Vg
7 Ca
13 Hi
14 Cu
52 Cl
54 804
2.830E+02
l.OOOE-03
9.000E-03
3.500E-04
TOTAL CONG
2.000E-04
l.OOOE-04
3.000E-04
l.OOOB-Ofi
l.OOOE-06
5.000E-04
2.000E-04
RESULTS
NO. OF ITERATIONS
PH
IONIC STRENGTH
CHARGE RATIO
CHARGE DIFFERENCE
ZED-HA
ZED-FA
RATIO-BA
RATIO-FA
MATER VOLUMES
FRACTION HA-DDL
FRACTION FA-DDL
FRACTION SOLUTION
48
7.000E+00
1.S18E-03
0.985B+00
•1.401B-05
-1.235E-03
-2.746B-03
4.4958*00
2.071E*00
7.731E-05
6.841B-03
0.993B400
CARBONATE ALKALINITY 46.954E-05
FRACTIONAL-DISTRIBUTIONS OF MASTER SPECIES
MASTER SPECIES S DBA DFA
3 Na
4 Kg
7 Ca
0.986
0.947
0.947
0.000
0.001
0.001
0.014
0.028
0.027
HA
0.000
0.002
0.002
FA
0.000
0.022
0.022
-------
1018
E. TIPPING
13 Ni
14 Cu
52 Cl
54 804
0.538
0.002
1.000
1.000
0.001
0.000
0.000
0.000
0.015
0.000
0.000
0.000
0.010
0.429
0.000
0.000
0.436
0.569
0.000
0.000
CONCENTRATIONS AND ACTIVITIES
SPECIES
TOTAL KATBR
[FREE] [ORGANIC)
1
3
4
7
13
14
51
52
54
55
101
102
124
125
126
145
146
147
-211
212
213
214
215
216
221
222
223
224
225
226
227
H
Ha
Kg
Ca
Hi
Cu
OH
Cl
SO4
COS
HC03
H2C03
MgHCO3
M0C03
MgSO4
CaHCO3
CaCO3
CaSO4
NlOH
Ni(OH)2
N1SO4
N1C03
NiCl
N1HCO3
CUOH
CU(OH}2
CuSO4
CUCO3
Cu(C03)2
CuCl
CuHCOS
1.037E-07
1.971E-04
9.268B-05
2.773B-04
4.938B-07
1.070E-09
3. OBOE-OB
5.000E-04
' 1.914B-04
2.666B-08
6.956B-05
1.827E-05
6.030E-08
1.334E-09
1.988E-06
1.664E-07
6.426E-09
6.591E-06
1.792E-10
3.696E-13
1.246E-08
5.739B-09
5.05BB-10
2.555E-08
8.495B-11
4.825E-13
2.803B-11
1.161E-10
4.608B-15
9.917E-13
8.980E-10
1.515E-09
2.881B-06
5.253E-06
1.589E-05
4.612B-07
9.848B-07
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
1.273B-07
8.813E-10
9.293E-12
1.385E-08
2.432E-09
4.476E-11
4.592E-08
3.925E-11
2.574B-15
8.680E-11
3.998E-11
7.393E-12
3.735E-10
1.302E-08
3.361B-15
1.953B-13
8.084B-13
O.OOOE+00
1. 4498-14
1.312B-11
SOLUTION PHASE
CONC ACTIVITY
1.044E-07
1.985E-04
9.333E-05
2.792B-04
4.973B-07
1.078E-09
3.102E-08
5.035E-04
1.927E-04
2.685B-08
7.004B-05
1.840E-OS
6.072E-08
1.343E-09
2.002E-06
1.676E-07
6.471E-09
6.637B-06
1.805E-10
3.721E-13
1.255E-08
5.779B-09
5.093B-10
2.573E-08
8. 5548-11
4.859E-13
2.B23E-11
1.169E-10
4.640E-15
9.986B-13
9.043E-10
l.OOOE-07
1.902E-04
7.910B-05
2. 3678-04
4.21SB-07
9.133B-10
2.971B-08
4.824E-04
1.633E-04
2.275E-08
6.710E-05
1.840E-05
5.817E-08
1.343E-09
2.002B-06
1.605B-07
6.471E-09
6.637E-06
1.729B-10
3.721E-13
1.255E-08
5.779B-09
4.880E-10
2.465B-08
8.195B-11
4.859E-13
2.823B-11
1.169B-10
3.933B-15
'9.567E-13
8.663B-10
VALUES OF HO (MOL BOUND / O RUMIC SUBSTANCES)
NO IS SPECIFIC BINDING } DNU IS DIFFUSE LAYER
1
3
4
7
13
14
102
124
125
126
145
146
147
211
212
213
214
215
216
221
222
B
Na
Hg
Ca
Ni
Cu
E2C03
MgRCOS
HgCO3
MgS04
CaHC03
CaC03
CaS04
NiOH
Ni(OH)2
NiSO4
N1CO3
NiCl
NiHC03
CuOH
Cu(OH)2
DMJ-HA
DNU-FA
HU-RA
NU-FA
3.627B-08
6.898B-05
1.458E-04
4.362B-04
7.768E-07
1.683B-09
1.422B-06
2.110B-08
1.039B-10
1.548B-07
5.823B-08
S.002E-10
5.131B-07
6.271E-11
2.877E-14
9.700E-10
4.467E-10
1.770E-10
8.942E-09
2.973E-11
3.756E-14
1.6438-07
3.12SE-04
3.043B-04
9.103B-04
1.621E-06
3.513B-09
1.398B-OS
9.558B-08
1.021E-09
1.S22E-06
2.638B-07
4.918B-09
5.045E-06
2.841B-10
2.829E-13
9.537E-09
4.392E-09
8.018E-10
4.050E-08
1.347E-10
3.693E-13
O.OOOE+00
O.OOOE+00
1. €068-04
€.5798-04
9. 5038-06
4.2588-04
0.0008+00
0.0008+00
0.0008+00
0.0008+00
0.0008+00
0.0008+00
0.0008+00
2.8868-10
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
2.829E-06
O.OOOE+00
O.OOOE+00
O.OOOE+00
2. 4548-04
7. 3418-04
4.848E-OS
6.211B-05
0. 0008+00
0.0008+00
O.OOOE+00
0.0008+00
0.0008+00
0.0008+00
O.OOOE+00
4.0388-09
OJOOOB+OO
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
1.132E-06
O.OOOE+00
-------
Equilibrium chemical spca'alion by WHAM
1019
223 CuSO4
224 CUC03
226 CuCl
227 CUHC03
2.182E-12
9.034E-12
3.470E-13
3.142E-10
2.145E-11
8.882E-11
1.S72E-12
1.423E-09
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
APPENDIX 5
WHAM-S input and output files, sediment example. See explanation of Appendix 4. In addition, note that CSOLID and
TCLAY are the concentrations of total soilds and total cation-exchanger day in gl~'.
SEDIMENT
Database
Array «lz«
Praciiion*
T«npK
CEOLID
THA TFA
TCIAT
FAAQ
pCO2
No. maat *p.
Na
Mg
Ca
F0III
Hi
Ctt
Cl
804
,SBDT1
,SSED10
,400
,0.1
,278
,100
,5,0.1
,20
,0.020
,0.001
,8
,3,€E-4
,4,4E-3
,7,5E-3
,ll,SE-3
,13,lB-5
,14,1E-S
,52,5E-4
,54,lE-4
**•**•***
OUTPUT FILE FROM WBAV-8 VERSION 1.0
•*•***•**
SOURCE FILE
DATABASE
PRECISION \-
SEDT1
SSBO10
0.100E+00.
INPUT DATA
TEMPK
CSOLID
TOT HA
TOT FA
TOT CLAY
FAAQ
PC02
MASTER SPECIES
3 Na
4 Mg
7 Ca
11 Falll
13 Ni
14 Cu
52 Cl
54 S04 '
2.780E+02
1.0008*02
S.OOOE+00
0.100E+00
2.000B+01
2.000E-02
l.OOOE-03
TOTAL CONC
6.000E-04
4.000E-03
5.000E-03
5.000E-03
l.OOOE-05
l.OOOE-05
5.000E-04
l.OOOE-04
RESULTS
NO. OF ITERATIONS
PH
IONIC STRENGTH
367
7.016E+00
1.475E-03
-------
1020
E. TIPPING
CHARGE RATIO
CHARGE DIFFERENCE
FAAQ (FIXED)
ZED-HA
ZED-FA
RATIO-HA
RATIO-FA
RATIO-CLAY
WATER VOLUMES -
FRACTION HA-DDL.
FRACTION FA-DDL
FRACTION CLAY-DDL
FRACTION SOLUTION
[FB(OH)3 PPT]
* TOT FB PPTD
0.999B+00
-1.044B-06
2.000E-02
-1.320B-03
-2.668B-03
9.4648*00
4.305B+00
2.5098*01
0.1388*00
2.6178-02
5.135E-03
0.8318+00
3.6628-03
73.25
••*•••«**•** •***••**•• •**•*•*•••*•*****•**•**•*••***«**•*••*«**•••*•*••
FRACTIONAL DISTRIBUTION OF MASTER SPECIES
MASTER SPECIES 8 DBA DFA HA FA
3 Na 0.633 0.248 0.021 0.000 0.000
4 M0 0.032 0.354 0.014 0.465 0.011
7 Ca 0.025 0.363 0.014 0.494 0.009
11 Felll 0.000 0.000 0.000 0.973 0.027
13 Ni 0.005 0.065 0.003 0.882 0.028
14 Co 0.000 0.000 0.000 1.000 0.000
52 Cl 1.000 0.000 0.000 ' 0.000 0.000
54 -SO4 0.994 0.005 0.001 0.000 0.000
CLAY
0.098
0.124
0.095
000
017
000
000
0.000
«******************«***«**«**•**•*••*•*••«**«*•***•«***«*«**•****•*****
TRATIONS AND ACTIVITIES OF SPECIES IN SOLUTION PHASE
SPECIES
1
3
4
7
11
13
14
51
52
54
55
101
102
124
125
126
145
146
147
185
186
187
188
190
194
195
211
212
213
214
215
216
221
H
Na
Mg
Ca
Falll
Ni
Cu
OH
Cl
804
COS
HC03
H2C03
M0HCO3
MgCO3
MgSO4
CaHC03
CaCO3
CaS04
FaOH
Fa (OH) 2
Fa (OH) 3
Fa (OH) 4
FaS04
FaCl
FaC12
NlOH
Ni(OH)2
N1S04
N1C03
NiCl
N1HC03
CuOH
CONG
1.006B-07
4.568E-04
1.S3UE-04
1.470B-04
2.674B-17
5.1998-08
1.9828-12
2.0918-08
6.0168-04
1.158B-04
7.968B-08
2.254E-04
6.13OB-OS
3.1368-07
6.0368-09
1.717B-06
2.614B-07
9.0868-09
2.007E-06
4.124B-13
5.516B-10
2.747B-10
1.133E-12
1.310E-17
1.933E-19
8.808B-22
1.276E-11
1.778E-14
7.567E-10
1.806E-09
6.283B-11
7.747E-09
1.065E-13
ACT]
9.643B-08
4.380B-04
1.302E-04
1.251E-04
1.881E-17
4.422B-08
1.686B-12
2.005E-08
5.769E-04
9.853B-05
6.776B-08
2.162B-04
6.130E-05
3.0078-07
6.036B-09
1.7178-06
2.5068-07
9.0868-09
2.007B-06
3.507E-13
5.289B-10
2.747B-10
1.086B-12
1.256B-17
1.645B-19
8.4478-22
1.224E-11
1.778E-14
7.567E-10
1.806E-09
6.02SB-11
7.429E-09
1.021E-13
-------
Equilibrium chemical spccialion by WHAM
1021
222 Cu(OH)2
223 CUS04
224 CUCO3
225 CU(C03)2
226 CuCl
227 CUHCO3
4.085E-16
2.938E-14
6.425B-13
7.5728-17
2.092B-1S
4.790E-12
4.085E-16
2.938E-14
6.425B-13
6.4418-17
2.006B-15
4.593E-12
TOTAL AQUEOUS-PHASE CONCENTRATIONS AND KD VALUES
SPECIES AQCONC KD
3 Ma
4 110
7 Ca
11 F«III
13 Ni
14 Cu
52 Cl
54 S04
4.570E-04
1.7468-04
1.7168-04
7.2088-06
1.2288-07
4.1448-10
5.9858-04
1.1908-04
44.7758*00
+2.2078+02
•2.83084-02
+6.9288*03
+8.0608+02
+2.4138+05
-2.9188-06
+5 .'1518-02
KD-APP
+3.1298+00
+2.1918+02
+2.8138+02
+6.9278+03
+8.0438+02
+2.4138+05
-1.6458+00
-1.5948+00
APPENDIX 6
WHAM-S input «nd output files, son example. See explanation to Appendices 4 end 5.
SOIL
Databaaa
Array «iza
Pr«ci0ion\
TanpK
CSOLID
TEA TFA
TCLAY
FAAQ
PCO2
No. naat «p.
Ma
Kg
Al
Ca
Felll
CO
Sr
Cm
An
Cl
8O4
,SOILT1
,S8BD10
,400.
,0.1
,278
,300.
.25.5
,0
,999
,0.01
,11
,3,28-4
,4,18-3
,5,38-2
,7,18-3
,11,28-3
,12,18-6
,16,18-6
,18,18-6
,28,18-6
,52,28-4
,54,48-5
******•*•
OUTPUT FILE FROM HBAK-8 VERSION 1.0
•*****••*
SOURCE FILE
DATABASE
PRECISION \
SOILT1
SSED10
0.1008+00
INPUT DATA
TEMPK
CSOLID
TOT HA
TOT FA
2.780E+02
3.000E+02
2.500E+01
5.000E+00
-------
1022
E. TIPPING
TOT CIAY
FAAQ
PC02
MASTER SPECIES
3 Na
4 Mg
5 Al
7 Ca
11 FaZXI
12 Co
16 Sx
18 C«
28 Am
52 Cl
54 604
O.OOOE+00
UNKNOWN
l.OOOB-02
TOTAL CONC
2.000E-04
l.OOOE-03
3.000E-02
l.OOOE-03
.OOOB-03
.OOOE-06
.OOOE-06
.OOOE-06
1.OOOE-06
2.000E-04
4.000E-05
2.
1.
1.
1.
RESULTS
NO. OF ITERATIONS 111
PH 4.199E+00
IONIC STRENGTH 5.162B-04
CHARGE RATIO l.OOOE-t-00
CHARGE DIFFERENCE +4.937B-08
—C.FAAQ (CALC) 4.622B-02
ZBD-BA -1.770B-04
ZED-FA -1.040B-03
RATIO-HA 1.799B+01
RATIO-FA 9.773B+00
WATER VOLUMES
FRACTION HA-DDL 3.710E-02
FRACTION FA-DDL 0.209B+00
FRACTION CLAY-DDL O.OOOE+00
FRACTION SOLUTION 0.754E+00
FE(OH)3 SATN INDEX -1.393E+00
AL(OB)3 SATN INDEX -2.861E+00
FRACTIONAL DISTRIBUTION OF MASTER SPECIES
MASTER SPECIES 6 DBA DFA
3 Na
4 Mg
5 Al
7 Ca
11 F«III
12 Co
16 Sr
18 C«
28 An
52 Cl
54 804
0.527
0.029
0.000
0.022
0.000
0.021
0.015
0.157
0.000
1.000
0.970
0.117
0.349
0.033
0.354
0.000
0.324
0.357
0.208
0.030
0.000
0.007
0.356
0.579
0.030
0.588
0.000
0.537
0.592
0.635
0.027
0.000
0.023
HA
0.000
0.010
0.769
0.010
0.486
0.042
0.023
0.000
0.771
0.000
0.000
FA
0.000
0.032
0.168
0.025
0.514
0.076
0.013
0.000
0.172
0.000
0.000
CLAY
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
CONCENTRATIONS AND ACTIVITIES OF SPECIES IN SOLUTION PHASE
SPECIES CONC ACTIVITY
1
3
4
5
7
11
12
H
Na
Mg
Al
Ca
Felll
CO
6.486E-05
1.398E-04
3.873E-05
9.094E-06
2.9S1E-05
2.667E-10
2.695E-08
6.325E-05"
1.363E-04
3.510E-05
7.319E-06
2.674E-05
2.147E-10
2.442E-OB
-------