United States
Environmental Protection
Agency
Environmental Research
Laboratory
Duluth MN 55804
Research and Development
EPA/600/S3-86/001 Mar. 1986
v>EPA Project Summary
Control of Cadmium
Carbonate Precipitation
Interferences During the
Dialysis of Cadmium in High
Bicarbonate Alkalinity
Aquatic-Life Bioassay Waters
John E. Poldoski
The precipitation of cadmium car-
bonate during the dialysis of cadmium
in a high bicarbonate alkalinity natural
water, was linked to a significant
source of error when determining
dialyzate cadmium concentrations. The
relative standard deviation was re-
duced by approximately four-fold when
this precipitation was controlled by
adding a particular preparation of
humic acid to the dialysis bag filling
solution. Linear regression correlation
coefficients for sample-by-sample
comparisons between resultant dialysis
values and corresponding free cad-
mium values, obtained by cadmium ion
selective electrode, were 0.90 or
greater for concentrations in the range
of 2^
-------
Experimental Section
Apparatus
Experiments were conducted using
two different molecular weight cut-off
(MWCO) dialysis bags(Spectrum Medical
Industries, SMI), specifically the 1,000
MWCO (#132634) and 12,000-14,000
MWCO (#132700) sizes clamped with
SMI closures.
The flow-through mini-diluters were
operated as part of an aquatic-life related
study in this laboratory and details of this
work, including functioning of the mini-
diluters are described elsewhere (5,6).
Atomic absorption spectrometric
(AAS) measurements of total cadmium
and other elements were made with a
Perkin Elmer model 5000 atomic absorp-
tion spectrophotometer equipped with
deuterium arc background correction, a
model HGA-500 graphite furnace, a
model AS-40 autosampler, a model AS-
50 auto-sampler, and a model 56 strip
chart recorder.
Free cadmium ion and pH measure-
ments were made with an Orion model
801 pH meter equipped with a 500 ml
FEP teflon cell, an Orion model 9448A
cadmium ion selective electrode (ISE), an
Orion double junction reference elec-
trode (#900200), an Orion model 605
electrode switch, and a Perkin Elmer
model 165 strip chart recorder.
Calculations were performed with a
Texas Instruments model T159 pro-
grammable calculator equipped with the
statistics solid-state software module.
Chemicals and Reagents
Unless indicated otherwise, all chem-
icals were of reagent grade quality or
better and deionized distilled water (Mil-
lipore Super Q) was used for preparing
solutions. The high calcium hardness-
high bicarbonate alkalinity natural water
(hardwater), used as the dilutant water
for aquatic-life bioassays, was prepared
by dissolving CaCOafrom limestone (5,7)
in Lake Superior water (softwater)
followed by filtration. It was subse-
quently aerated with 0.45 /u filtered air
until a constant pH of approximately 8.2
was maintained. Non-complexing di-
alysis retentate solutions (labelled as A
or B) were prepared to contain 4 x 10"' M
borate buffer (pH 8) and either 4 x 10~4 M
(A)or2x10"3M(B)Ca(N03)2.
Quenching Reagent Preparation
The quenching reagent was prepared
by adding 1,000 mg of humic acid
(Aldrich, H1675-2) to one liter of hard-
water and shaking the mixture for 1 hr.
Portions of this solution were then pres-
sure-filtered at 3 atm through a pre-
cleaned Millipore 0.45 fi, 47 mm di-
ameter membrane filter. The 0-10 mL
fraction was discarded and the 10-50 mL
fraction was collected and saved. This
filtration process was repeated with a
new filter until the entire 1 L volume was
processed. The collected filtrate was
diluted ten-fold with deionized distilled
water and stored in the absence of light at
4°C until use.
Results And Discussion
Reproducibility at High and
Low Alkalinity
Table 1 presents replication data for 5
h dialyses of cadmium and other cations
in the mini-diluter bioassay system, in
addition to cadmium dialysis from
retentate solution B. For these cases,
deionized distilled water was used as the
filling solution to compare the effect of
various retentate solutions on reproduci-
bility. Data are generally presented as the
mean and relative standard deviation of
ratios of total metal concentration in the
dialyzate to total metal concentration in
the unfiltered solution. In addition, for
just hardwater, the average ratio (and its
relative standard deviation) of total
cadmium concentration in the dialyzate
to total cadmium concentration in the
corresponding 0.45 n filtrate is also
given. Further, the means of the ratios of
each filtered concentration to each
corresponding total metal concentration
are given for comparison. As expected for
the softwaters, with little or no precipita-
tion occurring, reproducible ratios close
to 1 were obtained. These comparisons
show that the filtrate and dialyzate ratios
are very similar, suggesting a general
equilibrium of cadmium and other dis-
solved components between dialyzate
and retentate solutions. For hardwater,
ratios significantly less than 1 were due
to a contribution from paniculate forms
in the bioassay waters. In a typical
hardwater flow-through bioassay sys-
tem, Cd2+ concentrations of up to 1 mg/L
were added to the high bicarbonate
bearing bioassay waters. This resulted in
the solubility product of cadmium car-
bonate being greatly exceeded with re-
sultant precipitation continually oc-
curring in the water column and sticking
onto the surfaces of the tanks. Obviously,
this process could also be occurring in-
side the dialysis bags, possibly giving rise
to a significant source of imprecision and
positive bias. These reasons are likely to
explain the unusually high average R
Table 1. Reproducibility of 5 h Dialyses in Various Carbonate Alkalinity Waters Using Deionized
Distilled Water as the Dialysis Bag Filling Solution
Retentate Water Type
Metal
Percent
Relative
Standard
Deviation
Number
Retentate Solution B.
No carbonate
alkalinity
Hardwater, 212 mg/L
carbonate, alkalinity
bioassay water
Softwater, 48 mg/L
carbonate, alkalinity
bioassay water
Cd
Cd
Cd
Ca
Mg
Na
1.02
0.366
10312)'
0.878
(0.948)
0.912
(0.968)
0.757
(0.816)
0.968
(0.948)
6.1
45.4
34.0
4.4
1.0
1.1
1.0
12
10
10
14
3Mean ratio of (total [metal] in dialyzate) / (total [Metal] in retentate).
bIndicates mean ratio of (total [Cd] in dialyzate) / (total [Cd] in 0.45 u filtrate).
cParentheses indicate mean ratio of (total [metal]in 0.45 u filtrate) / (total[metal] in unfiltered
solution).
-------
value of 1.19 given for hardwaterandthe
corresponding high relative standard
deviation (34%-45%), as presented in
Table 1. Subsequent details will further
describe this problem and how it might be
avoided.
Quenching Effect ofHumicAcid
Chemical kinetics studies (8) of C02+
added to hardwater, with and without
added clay, indicated that the initial re-
action rates exhibited pseudo first-order
behavior with respect to loss of Cd2+ from
solution and that they were greatly
affected by a variety of other parameters,
but most importantly, humic acid pres-
ence (Figure 1). In the absence of humic
acid, observed pseudo first-order rate
constants (arbitrary units) ranged from
moderate to relatively high values,
particularly at high pH and high clay
concentration. However, under similar
conditions with 50 mg/L humic acid
present, observed rate constants were at
near zero values over a range of pH condi-
tions. In softwater, with or without clay or
humic acid present, loss of Cd2+ from
solution as a function of time was not
observed. These specific observations
demonstrated the problem and the likely
benefit of employing humic acid in the
dialyzate.
Effect of Variables on the
Conditional Concentration
Factor (F)
The effect of using the quenching
reagent preparation as the dialysis bag
filling solution was investigated with
regard to equilibration time, reproduci-
bility, and possible accuracy of dialysis.
Since this solution would undoubtedly
bind cadmium ions, there would ob-
viously be a natural tendency for cad-
mium to concentrate in the dialyzate.
Therefore, it was necessary to determine
this concentration factor and some
factors affecting it. This information
could permit calculation of the concen-
tration that would have normally dialyzed
in the absence of the concentration effect
of the quenching reagent.
It is appropriate to mention at this time
that the experimental system studied
(flow-through and fixed volume) had two
noteworthy characteristics: 1) the effec-
tive retentate volume was infinitely large
relative to the dialyzate volume, and 2)
the humic acid in the quenching reagent
contributed insignificantly to the overall
ionic content of the dialyzate. Therefore,
the phenomena commonly referred to as
Donnan membrane equilibria become
insignificant in this case and should not
be an additional factor complicating the
interpretation of data. Definition of F may
be attempted by first considering reten-
tate media containing only dissolved
cadmium species capable of equilibrating
with the dialyzate, with F given by the
following expression:
p -total [Cd] in dialyzate = [Cd]d
total [Cd] in retentate [Cd]r
Retentate solutions A or B fulfill these
requirements and, therefore, the concen-
tration of the species that are both inside
and outside the membrane should be
equal at equilibrium. For bioassay media,
the concentration of other metals and
ligands were considered negligibly low in
concentration, except for cadmium,
calcium, and bicarbonate. Calcium and
bicarbonate concentrations were kept at
a constant level, consequently concen-
trations of cadmium, pH, and dialysis
time were studied as main parameters
primarily affecting the value of F.
Figure 2 shows the effect of dialysis
time (1-20 h) on values of F for retentate
solutions A and B, all normalized to each
corresponding mean 5 h F value rep-
resenting a set of conditions. Conditions
varied by using either 1,000 MWCO or
12,000 MWCO membranes and total
cadmium concentrations in the range of
<10-200/ug/L. As a result, very similar
dialysis rates and equilibrium values
were obtained regardless of the par-
ticular membrane. It can be seen that the
value of F stabilized after 4 h and changed
about 10% or less in the 4-20 h range.
Reproducibility was in the range of 10%
or less for the plateau region.
As anticipated, experiments that were
conducted in high bicarbonate alkalinity
bioassay water produced similar results
with respect to dialysis rate and re-
producibility (Figure 3). Under actual
74,
72
10
8
*6
4
2
7.4 7.6 7.8 8.0 8.2 8.4 8.6 88 9.0
pH
Figure 1. Water quality variables affecting
the pseudo first-order rate con-
stant (arbitrary units) for loss of
free cadmium from solution.
A - hardwater + clay (90 NTU
turbidity) + 10 mg/L Ccf*. • -
hardwater + 10 mg/L Cd2*.
• - hardwater + clay (90 NTU
turbidity) + humic acid (50 mg/L)
+ 10 mg/L Cd2*.
1.0
0.8
.% 06
~
04
-------
bioassay conditions, a plateau was
reached in 4-5 h. In addition, the im-
provement in reproducibility, that can be
seen by comparing data in Figure 3 (error
bars indicate standard deviation) to cor-
responding data in Table 1, strongly in-
dicates that the high error associated
with dialysis in this type of water can be
significantly reduced by using an ap-
propriate quenching reagent in the
dialyzate.
1.2
0.4
0.2
0 123456
Dialysis Time (h)
Figure 3. Change in mean 4 h normalized
conditional concentration factor
as a function of dialysis time for
high bicarbonate alkalinity flow-
through bioassay waters. X -
data to which other data are
normalized. The number of data
points averaged for the 1, 2, 3.
5, and 6 hour periods were 7, 7,
2, 4. and 3, respectively.
Literature Cited
1) Benes, P. Semicontinuous Monitoring
of Truly Dissolved Forms of Trace
Elements in Streams Using Dialysis In
Situ I. Principle and Conditions. Water
Res. 1980, 14, 511-513.
2) Rainville, D.P. and Weber, J.H. Com-
plexing Capacity of Soil Fulvic Acid
for Cu2+, Cd2+, Mn2+, Ni2+ and Zn2+
Measured by Dialysis Titration: A
Model Based on Soil Fulvic Acid
Aggregation. Can. J. Chem. 1982,60,
1-5.
3) Truitt, R.E. and Weber, J.H. Deter-
mination of Complexing Capacity of
Fulvic Acid for Copper (II) and Cad-
mium (II) by Dialysis Titration. Anal.
Chem. 1981, 53, 337-342.
4) Truitt, R.E. and Weber, J.H. Copper(ll)-
and Cadmium(ll)-Binding Abilities of
Some New Hampshire Freshwaters
Determined by Dialysis Titration.
Environ. Sci. Technol. 1981, 15,
1204-1208.
5) Benoit, D.A. U.S. Environmental Pro-
tection Agency, Duluth, MN, Personal
Communication, 1981-1983.
6) Benoit, D.A., Mattson, V.M., and
Olson, D.M. A Continuous-Flow Mini-
Diluter System for Toxicity Testing.
Water Res. 1982, 16, 457-464.
7) Lemke, A.E. A Water Hardener for
Experimental Use. J. Am. Water
Works Assoc. 7963, 61, 415-416.
8) Poldoski, J.E. U.S. Environmental Pro-
tection Agency, Duluth, MN. Unpub-
lished work, 1981-1983.
The EPA author, John E. Poldoski (also the EPA Project Officer, see below), is
with Environmental Research Laboratory, Duluth, MN 55804.
The complete report, entitled "Control of Cadmium Carbonate Precipitation
Inteferences During the Dialysis of Cadmium in High Bicarbonate Alkalinity
Aquatic-Life Bioassay Waters," (Order No. PB 86-145 620/AS; Cost: $9.95,
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Research Laboratory
U.S. Environmental Protection Agency
Duluth, MN 55804.
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S3-86/001
------- |