\

{EWING ‘Llnmxuiil

vunwoswa‘ifln u" tflte
University

 

——~—4 A I"

W ——W

This is to certify that the

thesis entitled

A Comparison of Scatchard Analysis

and the Gaussian Distribution Model
to Determine a Conditional Stability
Constant Between the Uranyl Ion and

Humic Substances
presented by

Richard Alan Gei ger

has been accepted towards fulfillment

of the requirements for
Master of

Science degree inFisheries and Wildlife

 

Major professor

Date M‘? ZZL/ygb/
U U

0.7539 my.) on Aflimnm'vc Action/Equal Opportunity Institution '

——-—— w——————

 

 

MSU

LIBRARIES
“3—.-

 

 

 

 

 

 

 

 

 

 

RETURNING MATERIALS:

Piace in book drop to
remove this checkout from
your record. FINES will
be charged if book is
returned after the date
stamped beiow.

-Tflfl

 

 

 

 

 

 

 

 

 

 

A COMPARISON OF SCATCHARD ANALYSIS AND THE
GAUSSIAN DISTRIBUTION MODEL TO DETERMINE
A CONDITIONAL STABILITY CONSTANT BETWEEN THE

URANYL ION AND HUMIC SUBSTANCES

Richard Alan Geiger

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Fisheries and Wildlife

1985

ABSTRACT
A COMPARISON OF SCATCHARD ANALYSIS AND THE
GAUSSIAN DISTRIBUTION MODEL TO DETERMINE A

CONDITIONAL STABILITY CONSTANT BETWEEN THE
URANYL ION AND HUMIC SUBSTANCES

By

Richard Alan Geiger

Naturally occurring aquatic humic substances (HS) are greater
controlling factors in the mobilization and distribution of trace
metals in blackwater systems of the southeastern USA than most
inorganic ions and suspended solids. Ion exchange, photo-oxidation

and laser fluorometry were used to determine the maximum binding

2+
2

conditional stability constant (I-(') between a constant concentration

capacity (BC) of Aldrich humic acid (HA) for U0 and the average

of Aldrich HA and U0:+ concentrations varying from 5.25 x 10-8M to
2.10 x 10-7M UO§+. The Scatchard analysis over the range of UO§+zHS

ratios tested is sufficiently linear to predict E'. A continuous

distribution (Gaussian) estimate of R' was calculated to compare to

2+
2

concentrations typical of those in aquatic systems of non-uraniferous

the R' determined by the single component Scatchard analysis. At U0

areas, the estimated R' between U0:+ and Aldrich HA is not

substantially different when calculated by Scatchard analysis or

Gaussian distribution.

ACKNOWLEDGMENTS

I want to acknowledge the generous assistance of those
individuals who made this work possible. Dr. Niles R. Kevern, my
major professor, provided invaluable advice and general guidance
throughout the development and completion of this study. Drs. John P.
Giesy, Jr. and James J. Alberts oversaw the research and furnished
encouragement and direction throughout the study. I would like to
express my gratitude to two other committee members, Drs. Darrell L.
King and David T. Long, for their advice and critical evaluation,
particularly during preparation of the research proposal and writing
of the thesis. In addition, I want to thank Dr. James M. Tiedje for
his advice during the writing of the research proposal and Dr. Michael
C. Newman for his critique of the rough draft of the thesis.

I offer my sincerest thanks to the staff and students of the
Savannah River Ecology Laboratory. Their friendships over the years
contributed to my development as a scientist and an. individual.
Without their assistance this study may never have been completed.

Several people deserve special recognition. Technical assistance
in the study was provided by Ms. Susan Lesnek, Mr. John Bowling and
Mr. Fred Stone. Typing of the draft thesis was done by Ms. Jan Moran
and Ms. Kathy Tseng. Typing of the final thesis was done by the

Savannah River Ecology Laboratory secretarial staff. Ms. Jean Coleman

ii

prepared the figures. Technical editing was performed by Ms. Karen
Patterson.

Funding was provided through a teaching assistantship from
Michigan State University and through contract DE-ACO9-76-SROO819

between the University of Georgia and the U. S. Department of Energy.

iii

TABLE OF CONTENTS

LIST OF TABLES .

LIST OF FIGURES

INTRODUCTION .

MATERIALS AND METHODS.

THEORETICAL CONSIDERATIONS .
Stability Constants .
Discrete Multiligand Models .
Continuous Multiligand Models .
Binding Capacity.

RESULTS.

DISCUSSION .

CONCLUSION .

LIST OF REFERENCES .

iv

vi

13

13

14

19

23

32

42

72

75

Table

LIST OF TABLES

Page

Primary distribution of metals and ligands in test
solution (pH 4.5; 0.01 M Ac ) . . . . . . . . . . . . . . 8

Experimental flow rates of varigus uranyl concen-
trations (M) through Chelex-IOO ion exchange resin
column (0.01 M Ac and 3.5 mg C/E at pH 4.5). . . . . . . 10

Summary statistics for nonlinear least squares analysis
to estimate th§+mean (p) and standard deviation2(o) of
uncomplexed U0 (Mf) and observed values of U02
complexed by humic acid (ML). . . 34
Acid dissociation (ionization) constants (pK )

for two major types of proton sites on simple aromatic

acids or aromatic acid components of humic and fulvic

acid molecules. Two or more constants indicate the

first, second, and possibly the third dissociation

constants, respectively . . . . . . . . . . . . . . . . . 47

Organic carbon analyses of replicate samples of
dialyzed Aldrich humic acid (3.5 mg C/2). . . . . . . . . 51

Acidities of some oxy-functional groups in humic
substances (meq/g; from Giesy and Alberts 1984) . . . . . 53

Analytical techniques for low concentrations
of uranium. . . . . . . . . . . . . . . . . . . . . . . . 55

Average conditiggal stability constants (E') for aqueous

complexes of UO2 ions and humic substances . . . . . . . 67

LIST OF FIGURES

Figure Page

1. Observed versus predicted concentrations of uranium

complexed by Aldrich humic acid as a function of

total uranium added to the solution (M; pH 4.5;

o 01! Ac ). . . . . . . . . . . . . . . . . . . . . . . . 27
2. Concentration of uranium (M) eluted through Chelex-100®

ion exchange resin column as a function of flow

rate of the solution through the column (ml/min; pH 4.5;

0.01 M Ac ; 3.5 mg C/fi, dialyzed Aldrich HA) . . . . . . . 31

3. Scatchard plot of Aldrich humic acid_3inding of U02+_§t
conc§ntrations ranging from 4.2 x 10 _ M to 2.1 x IO
M U02 (pH 4.5; 3.5 mg C/2; 0.01 M Ac ) . . .

4. The Scatchard estimate of the overall average conditional
stability constant (K') in the region of interest based
on the ratio of complexed uranium to maximum number of
bigding sites for Aldrich humic acid (v) and uncomplexed
U02 (Mf) . . . . . . . . . . . . . . . . . . . . . . . . 38

36

5. Regression of the observed conditional stability constant
(K') as predicted by a Gaussian distributiag in the region
of interest as a function of the metal (U02 ) to ligand
(Aldrich humic acid) ratio. . . . . . . . . . . . . . . . 41

6. Types of metal complexation hypothesized for humic
acid and fulvic acid (redrawn from Manning and
Ramamoorthy 1973) . . . . . . . . . . . . . . . . . . . . 45

7. Emission spectrum of an aqueous solution of uranium
(from Campen and Bachmann 1979) . . . . . . . . . . . . . 57

8. Relative signal intensity of Scintrex Model U -3
fluorometer as a function of amount of Fluran added

to uranyl solution (from Campen and Bachmann 1979). . . . 60

9. Fluorescence spectrum of water samples containing
humic substances (from Robbins 1978). . . . . . . . . . . 62

vi

INTRODUCTION

Naturally occurring uranium (U) has an atomic weight of 238.03
atomic mass units. The commonly occurring valence states of U in the

6+. Although U4+ is found in reduced

natural environment are U4+ and U
zones in natural environments (Garrels and Larsen 1959; Bonatti et a1.
1971; Kolodny and Kaplan 1973; Brookins 1978; Langmuir 1978), it is
not a very soluble chemical species. The U6+ ion is generally found
in the oxidizing environment of surface waters as the more soluble
uranyl ion, UO§+, and can actively transport U in the solution phase.

36d17s2 shells can be removed by a strongly

Valence electrons in the 5f
electronegative element such as oxygen (0) to form a strongly
complexing ion.

With the passage of the Nuclear Waste Policy Act of 1982, state
governments have been encouraged to develop regional consortia and to
establish site criteria for the disposal of low-level radioactive
waste (LLW) at shallow land burial (SLB) sites. Depleted U (i.e. U
with less 235U than 0.720%) can be disposed of as LLW in a SLB.
Previously, SLB of LLW had been restricted to three companies
operating six: disposal sites (Francis et .al. 1980). Radionuclide
migration has been detected at some of these sites; microbial
degradation of synthetic or natural organic compounds in the burial
trenches may result in formation of ligands that can bind, leach and

solubilize to form organo-uranium complexes (Means et a1. 1978;

Francis et a1. 1980; Francis 1981).

2

The Environmental Protection Agency (EPA) is considering

2+
2

) for drinking water (Cothern et al. 1983). This

proposing a health guidance concentration of 10 pCi U/2 (34 ppb U0
or 1.43 x 10'7 g UO§+
small concentration may be chronically toxic to aquatic biota
continuously exposed to contaminated surface waters.

0f the two hazards associated with U or its compounds, chemical
toxicity is considered a greater hazard than radiological because the
U compounds obstruct phosphorylation and carbohydrate metabolism, thus
causing renal failure (Gindler 1973; Nechay et al. 1980). Ingestion
and assimilation of soluble species is one potential toxic pathway of
U. With an expected increase in the use of SLB sites around the
U.S.A., there is an increasing potential for U to be complexed as a
soluble compound and to migrate from these sites. Therefore, it is
imperative to understand the factors which control the mobility of
soluble U, its ultimate fate and its potential effects.

Uranium is concentrated in several types of igneous and
sedimentary rocks (McCormick and Cotter 1964; Osmond 1964; Szalay and
Samsoni 1969; Bouwer et al. 1978; Halbach et al. 1980; Johnson et al.
1980; Metzger et al. 1980). During the weathering of granite, U is
mobilized from the bedrock, primarily as UO‘:+ (Halbach et al. 1980;
Tieh et al. 1980). Uranium can also be mobilized during phosphate
mining (Osmond 1964; Bouwer et a1. 1978; Metzger et a1. 1980). During
the past twenty years U concentrations in several North American
rivers have increased, presumably due to contributions from man's
activities (Spaulding and Sackett 1972; Bloch 1980). However, Mangini
et a1. (1979) suggest that world-wide U concentrations in rivers and

lakes are primarily a function of aquatic chemistry rather than input

3
from human activities and are controlled by the inorganic chemistry of
the water, not differences in bedrock material or amounts of U leached
from phosphate fertilizers.
Several investigators have studied the :physical and chemical
2+ mobility (Moskvin et al. 1967;

2
Langmuir 1978; Dongarra and Langmuir 1980; Giblin 1980; Giblin et al.

properties that may increase IX)

1981). Langmuir (1978) reported the pH range (pH 5.0 - 8.5) of
maximum U0:+ sorption onto most natural colloidal materials, including
humic substances, ferric and manganic oxyhydroxides, and clays. This
pH range is common to the blackwater streams and swamps of the
southeastern U.S.A. (Beck et al. 1974; Reuter and Perdue 1977; Giesy
and Briese 1977, 1978a; Alberts and Evans 1979; Alberts and Giesy
1983). These investigators conclude that naturally occurring humic
substances are greater controlling factors than most inorganic ions in
the mobilization and distribution of trace metals in these soft water
systems. Complexation of UO§+ by naturally occurring aquatic humic
substances (HS) may be important in determining the bioavailability of
U0:+ (Jennings and Leventhal 1978) and is influenced by such
physico-chemical parameters as pH, Eh, ionic strength, and adsorption
to surfaces such as clays or metal oxyhydroxides (Martin et al. 1976;
Shelke and Jahagirdar 1979; Li et al. 1980; Kribek and Podlaha 1980;
Halbach et al. 1980; Nash et al. 1981).

Comprehensive investigations have been conducted on the chemical
structure and function of naturally occurring dissolved organic matter
in surface waters. These studies characterize HS, the major component

of the dissolved organic matter (Steelink 1977; Perdue 1978; Liao et

al. 1982) and describe complexation of trace metals by HS (Jackson

4

1975; Guy and Chakrabarti 1976; Jackson et al. 1978; Mantoura et al.
1978; Saar and Weber 1982). A decline in productivity can result from
the sequestering of nutrients by HS in freshwater systems (Jackson and
Hecky 1980). HS are complex organic polymers resulting from the
decomposition of plant tissues. They have variable structures which
have not been completely defined but are known to be composed of
aromatic and aliphatic components with oxygen containing functional
groups, particularly carboxyl (-COOH) and phenolic hydroxyl (Cb-OH)
(Rashid and King 1970; Schnitzer and Khan 1972; Beck et a1. 1974;
Borggaard 1974; Reuter and Perdue 1977; Choppin and Kullberg 1978; Li
et a1. 1980; Christman and Gjessing 1983). These functional groups
can protonate and dissociate in the pH range (pH 3.0 - 9.0) common in
natural surface waters (Gamble 1970; Choppen and Kullberg 1978; Saar
and Weber 1982). Aquatic HS are important in southeastern U.S.A.
surface waters because their abundance and binding capacities make
them potential controlling factors in the biogeochemical cycles of
inorganic ions in these waters (Beck et al. 1974; Schindler and
Alberts 1974; Casagrande and Erchull 1976; Giesy and Briese 1977,
1978a; Reuter and Perdue 1977; Alberts and Evans 1979).

The strong association between U and organic matter has been
known for some time (Moore 1954; Szalay 1954, 1958, 1964a,b, 1969;
Breger et al. 1955a,b; Szalay and Samsoni 1969). Langmuir (1978)
mentions the importance of the U02

2

geochemical cycling process and cites previous work on the subject

+ . .
-organic matter complexes in the

(Germanov and Panteleyev 1968; Haji-Vassiliou and Kerr 1973; Pauli

1975) .

5
In previous investigations a major problem has been the lack of a
rapid, accurate analytical technique to detect the concentration of U
bound to aquatic HS. Many techniques are not sufficiently sensitive
at small concentrations of U. Using thermodynamic simulation models
2+ in surface waters is desirable, but

2

to accomplish this the thermodynamic interactions of U0:

and inorganic ligands must be known. While the literature is replete

to predict the speciation of U0

+ . .
With organic

with information on the inorganic solubility and complexation

chemistry of U02+ (Langmuir 1978; Dongarra and Langmuir 1980), much

2

less is known about the thermodynamic interactions between U0:+ and
aquatic HS.

The hypothesis tested in this study was that environmental
concentrations of uranium (U) will bind tightly to the aquatic humic
component under conditions representative of natural surface waters of
the southeastern U.S.A. To test the hypothesis it was necessary (1)
to develop a rapid, sensitive method for quantifying U concentrations
of inland surface waters, (2) to determine maximum binding capacity
(BC) of aquatic HS for the uranyl ion (UO§+), (3) to determine the

.. .. - + .
average conditional stability constant (K') between U02 and aquatic

2

HS under simulated natural conditions in the laboratory using both the
Scatchard analysis and a Gaussian model, and (4) to evaluate which of

the two analytical methods more accurately represents the K' at

concentrations of U0:+ and HS typical of surface waters of the

southeastern U.S.A.

MATERIALS AND METHODS

Aldrich humic acid (Aldrich Chemical Co., Milwaukee, WI:
H-1675-2, Lot 082091) used in this study was purified by the methods
of Landrum and Giesy (1981). Briefly, the HA was dissolved in 0.1 M
NaOH, centrifuged auui the supernatant decanted. Supernatant pH was
reduced to 2 with 6 M HCl, then left undisturbed for 18 h. The
resulting IUX precipitate was centrifuged and the supernatant
discarded. The solid was dissolved in 0.1 M NaOH, transferred to a
beaker and stirred at room temperature for 18 h. This solution was
centrifuged and the precipitation, resolution and centrifugation were
repeated twice. The final precipitate was dissolved into 0.1 M NaOH
and the solution pH adjusted to 7 with 6 M HCl. The solution was
transferred tn) dialysis bags (Fisher Scientific, Pittsburgh, PA) and
dialysed against deionized water until the conductivity of the water
outside the bags was 10 pmhos. Following dialysis, the solution was
refrigerated in polyethylene bottles.

The dry weight and ash free dry weight of the purified HA stock
solution were determined on 5 replicate samples. Eighty (80) ml
aliquants of the extracted solution were dried at 85°C. Average dry
weight of the 5 replicates was 3.53 1 0.0002 mg/ml. Average ash free
dry weight (air atmosphere) was 1.18 i 0.003 mg/ml. Ash weight

accounted for 66.6% of the average dry weight.

7

In solutions containing low concentrations of metals, loss of
elements to the surface of glassware can result in large experimental
errors (Giesy and Paine 1977). To minimize adsorptive losses to the
glass surfaces, all glassware was thoroughly washed, rinsed and
treated with 1% Prosil-28® (PCR Research Chemicals Inc., Gainesville,
FL). Glassware was retreated after every tenth use.
2+ was separated from that complexed to Aldrich HA

2
by Chelex-lOO® chelating resin (Na+ form; BioRad Laboratories,

Uncomplexed U0

Richmond, CA). Chelex-IOO® is a weakly acidic ketoiminocarboxylic
cation exchange resin with properties similar to those of naturally
occurring HA. Separations were conducted in 10 m1 glass columns with
fritted glass bottoms (BioRad Laboratories, Richmond, CA) over which a
0.22 pm acetate filter (Millipore Corp., Bedford, MA) was fitted.

A 0.01 M acetate (Ac-) buffer solution was prepared by combining
0.01 M (0.82 g/2) sodium acetate (NaOAc) and 0.017 M (1.00 ml) acetic
acid (HOAc). The pH was adjusted to 4.5 with 1.0 M or 0.1 M NaZCOB°
Under the conditions of the study (pH 4.5 and 0.01 M Ac-) and at
4.2 x 10"6 r_1 uoi", 95.1% of the U0:

the absence of HA (Table 1). GEOCHEM, a thermodynamic equilibrium

+ should be complexed by the Ac-in

model for natural water systems (Sposito and Mattigod 1979), was used

to predict equilibrium concentrations of inorganic and simple

+ .-
UO2 -organic species. Relative to humics, Ac weakly complexes UO§+;
GEOCHEM accounts for the competition between Ac- and HA for UO2+ ions.

2

For each separation, Chelex-lOO® resin swelled in the Ac- buffer
(pH 4.5) and 2 ml (0.8 meq) was placed in the column. Overlying Ac-
buffer was removed and the column filled with the appropriate test

solution. The column was then attached to a 500 ml Mariotte flask

Table 1. Primary distribution of metals and ligands in test solution

(pH 4.5; 0.01 5 Ac_)

 

 

 

Metal/Ligand Distribution Percentage*
Na+ Free metal _ 99.5
Bound with Ac 0.5
U0:+ Free metal _ 4.5
Bound with Ac_ 95.1
Bound with OH 0.5
Ac_ Free ligand + 39.2
Bound with Na 0.2
Bound with H 60.6
N0; Free ligand 99.8
Bound with Na 0.2

* Percentage may not add to 100.0 due to rounding error

9
arrangement containing additional test solution which ensured a
consistent flow rate.
Before separations were conducted, flow rate through the exchange

2+ . .
2 was quantitatively removed

column was optimized so that unbound UO
from the solution without significantly altering the equilibrium

2+
2

solution was prepared in 0.01 M Ac-buffer

between uncomplexed and complexed U0 (Giesy 1980). To do this,

500 ml of 4.2 x 10‘6 g no:+
(pH 4.5) and 3.5 mg C/ll dialyzed Aldrich HA added. An identical
solution without the Aldrich HA was prepared as a reference solution.
The ion exchange column was filled with the UO§+-HA solution.
Approximately 10 ml (5 bed volumes or BV) of solution were eluted and
discarded. A minimum flow rate was set with an adjustable screw clamp,
and at least one BV of solute was eluted and discarded. Then a 15 ml
aliquot was collected, the flow rate was increased, measured and
recorded, and at least one more BV of solute was eluted and discarded
before the second sample aliquot was collected. This procedure was
repeated until either the flow rate was maximized or no increase in
2+

the amount of U02 eluting from the column was detected. The procedure

was repeated for the UO2+ solution without the HA.

2
U0:+ standards were prepared in pH 4.5 Ac- buffer from reference

standards (U0:+ as U02(NO3)2°6H20; Anderson Laboratories, Fort Worth,
TX). The UO§+zHA ratio was varied by maintaining a constant HA

2
2

g UO§+ (Table

concentration of 3.5 mg C/B for each test solution and adding UO + to

8 M to 2.1 x 10-5

attain concentrations from 4.2 x 10-
2). The optimal flow rate was set and 60 ml (30 BV) of the test
solution was eluted and discarded before a 15 ml aliquot was taken for

each UO§+:HA ratio.

10

Table 2. Experimental flow ates of various uranyl concentrations_(M)
through Chelex-lOO ion exchange resin column (0.01 M Ac
and 3.5 mg C/Q at pH 4.5)

2+

 

M92__(M) Flow Rate (mls/min)
4.20 x (10)”8 18.0
5.25 x (10)"8 18.0
6.30 x (10)’8 18.0
7.35 x (10)"8 18.8
8.40 x (10)"8 18.6
1.05 x (10)”7 18.5
1.26 x (10)"7 18.0
1.68 x (10)”7 18.5
2.10 x (10)"7 17.8
3.15 x (10)'7 18.2
4.20 x (10)"7 18.0
6.30 x (10)”7 18.0
8.40 x (10)“7 17.8
1.05 x (10)"6 17.8
1.26 x (10)”6 17.8
1.68 x (10)”6 17.7
2.10 x (10)’6 17.7
3.15 x (10)'6 17.7
4.20 x (10)'6 17.9
1.05 x (10)"5 17.5

2.10 x (10)”5 18.0

11

Organic matter (Aldrich HA) interferes with laser fluorometry
and was therefore eliminated by photo-oxidation. Chemical oxidation
by potassium permanganate was not used because of interference from
Mn2+ and changes in sample volume. All samples to be analyzed for U,
including those without HA, were photo-oxidized 2 11 in a xenon arc
ultraviolet photo-oxidation unit (La Jolla Scientific Co., La Jolla,
CA). Besides eliminating interference from HA, this oxidation
procedure also destroyed the C-C bonds of the Ac- buffer which also
may have interfered with the analysis. Photo-oxidized samples were
neutralized to approximately pH 7 with a small volume of Na2C03 and
diluted (5 to 10 times) so that the sample concentration fell within
the selected detection range of the instrument.

Uranium concentrations were determined with a Scintrex Model UA-3
Uranium Analyzer (Scintrex Ltd., Toronto, Canada) which uses pulsed
laser-induced fluorescence to detect ng (ppb) quantities of U. The
analyzer was zeroed and a cuvette containing 5 ml of sample was placed
in the sample compartment. The photomultiplier tube (PMT) was
activated, any residual signal was nullified and the PMT turned off.
Five hundred 1.11 (0.5 ml) of Fluran®, a proprietary pyrophosphate
buffer (Scintrex Ltd., Toronto, Canada), was then added to the 5 ml
sample which was thoroughly mixed and returned to the sample
compartment. Fluran® strongly complexes various U02+

2

may be present in a sample into a single species with a high

species which

luminescent yield. The PMT was reactivated and a sample reading
recorded when the meter deflection stabilized (10 - 20 sec). If the
meter deflection was either off-scale or minimal, a more appropriate
range was selected and the entire procedure from the initial zeroing

was repeated.

12
After a value for the sample was obtained, 25 or 50 pl of
6 2+

1.05 x 10- M U02 solution were added to the aliquot and mixed; the

volume added depended (Hi the initial meter deflection. The PMT was
reactivated and a reading taken of the meter deflection; sample UO§+
concentration was calculated (Equation 1). This standard addition
technique gives more accurate data than the semi-quantitative
calibration curve technique (Scintrex N.D.), but the principal

disadvantage of the standard addition method is that it is more time

consuming since it must be done for each sample.

2 = D1 . 2 . A (1)
D1 D2 b
where: Z = U concentration of sample aliquot
D1 = meter deflection due to sample
D2 = meter deflection due to sampling plus standard addition
a = volume of standard addition (25 or 50 pl)
b = volume of sample aliquot (5 x 103 p1)
A = U concentration of standard solution (1.05 x 10_6 M UO§+).

The total, unbound and bound U concentrations were then used to
compare conditional stability constants by' Scatchard. and. Gaussian
methods (see Theoretical Considerations).

Linear and nonlinear regressions for estimating BC and the
Scatchard parameters were computed using the general linear model
(GLM) and nonlinear least square techniques (NLIN) of SAS (Helwig and
Council 1979) and Marquardt nonlinear least squares procedures. The
Gaussian estimates of the conditional stability constants were
obtained with KINFIT4 (Dye and Nicely 1971) and a driver program

written by J. P. Giesy (Giesy, pers. comm.)

THEORETICAL CONSIDERATIONS

Stability Constants

 

The affinity of a metal for a particular ligand is represented by
a stability (formation) constant, K. By definition, K is a competi-
tive reaction between a metal ion and protons for binding sites on the

ligand, therefore, it is pH dependent (Equations 2 and 3).

[1+ +

 

M + LH <=>MLH + nH (2)
x x-n
+ n
(MLHx-n) (H ) (3)
K :
01““) (L11 )
x
where: K = stability constant
Mn+ = metal ion with n deficient electrons; M will be used
in subsequent equations to denote the unbound metal
concentration
H+ = hydrogen ion
L = organic ligand
x = number of protons released from the complexing ligand in

order to complex the metal ion; L will be used in
subsequent equations to denote ligand L with proton
deficiency x.

There are numerous HS functional groups and several acid
dissociation constants can be measured (Giesy and Alberts 1984), which
indicates that a number' of' complexation. reactions can. take ;place
simultaneously. Even if the HS functional groups involved in binding

. + .
an ion such as UO2 were very homogeneous, a range of complexes With

2

13

14
different stability constants would be expected due to polyelectric
effects of sequential site filling and the formation of both
mononuclear and polynuclear complexes. For these reasons conditional
stability constants (rather than. stability' constants) are used to

describe the strength of complexes at constant pH (Equation 4).

. [8L1] (4)
K . = -———-———
1 [M 1 [L 1
f f
where: K'i = conditional stability constant for complex type i
corrected for proton deficiency
[MLi] = molar concentration of metal complexed to ligand
type i
[Mf] = molar concentration of uncompls ed metal; this
includes both ionic species (M ) and hydrated
ionic species (Mn °yH20)
[Lf] = molar concentrations of uncomplexed ligand of type

i; all ligand types are not necessarily complexed
to any metal, whether of interest or not.

Discrete Multiligand Models
In a heterogenous mixture of binding sites, an average stability

constant (K) can be defined (Equation 5; notation of Perdue and Lytle

1983).

2i [MLi] - 2i Ki [Li] (5)

[MfliilLi] 21 [Li]

where: [13) == molar concentration of uncomplexed ligand of type i.

15

The average conditional stability constant (K') is described by
Equations 6 - 8.

 

 

 

_ 21 [ML] (6)
K' =
[Mr] 21 [Hxi Li]
where: [H . L.] = concentration of protonated ligand of
XI 1 type i (x is the number of protons
released from complexed ligand Li)°
K’ : 2i K i [HxiLi] (7)
2i [Hxi Li]
R: : [MT] - [Mf] (8)
Inf] ({LT] - [MT] + Infl)
where: [MT] = total molar concentration of metal
[LT] = total molar concentration of ligand; all types of sites.

Stability constants have been calculated for metals complexed by
ligands of known molecular weight and structure (Buffle et a1. 1977).
For the organic ligands of unknown structure which occur in natural
surface waters (Stevenson and Ardakani 1972; Gardiner 1974), stability
constants are much more difficult to calculate because of the
inability to calculate molar ligand concentrations (Stevenson 1977) or
to define the complexes as mononuclear or polynuclear and to describe
properties of’ each. type: of complex. Instead, average conditional
stability constants have been calculated by comparing the metal
binding capacity of a ligand of unknown structure to the binding
capacity of the same metal with a reference ligand of known structure

(Equation 9).

16

 

 

HxiLi (9)
2. K’. ——-——-—
- 1 1 erLr
K!—
HxiLi
2.
1HL
xr r
where: erLr = concentration of protonated reference ligand.

The Schubert ion exchange method (Schubert 1948; Miller' and
Ohlrogge 1958; Randhawa and Broadbent 1965; Schnitzer and Hansen 1970)
estimates an average conditional stability constant (K') for a
metal-ligand system by measuring the distribution of a metal between a
solute and solid phase both in the absence and the presence of a
complexing agent. This method does not allow the ligand concentration
to vary; only 1:1 metal-ligand complexes can be measured. For ligand
mixtures such as HS found in the blackwater streams of the
southeastern U.S.A., an average stability constant (K) is estimated
(Clark and Turner 1969; Gamble et al. 1970; Schnitzer and Hansen 1970;
Stevenson and Ardakani 1972; Wahlgren et al. 1972; Beck et al. 1974;
Crosser and Allen 1977; Giesy et al. 1977; Giesy et al. 1978; Giesy
and Briese 1980; Giesy 1980). The Bjerrum method (van den Berg and
Kramer 1979) was developed specifically for determination of K' for
compounds such as HS (Stevenson 1977), if one type of binding site is
present, but also provides an average stability constant, K.

The Scatchard analysis, another method of estimating K' is a
discrete model which results in a straight line if only one type of
metal binding site is present. (Scatchard 1949; Mantoura and Riley
1975; Guy and Chakrabarti 1976; Mantoura et al. 1978; Giesy 1980;
Sposito 1981; Saar and Weber 1982). Scatchard analyses do not

indicate if specific types of complexes form nor if 1:1 complexes with

17
different stability constants exist. However, average stability

constants can be estimated from Equations 10 - 16.

_ [MLi] (10)

up

 

<
l

S
U‘
(D
11
(b
<
H

ratio of concentration of complexed ligand of type i to
concentration of total ligand; all types of sites

 

[LT] = [MLi] + [HxiLi]°
V. : K'i Inf) (11)
1 I
1 + K i [Mf]

When summed over all types of sites present in the mixture, 9 is the

ratio of metal bound to the total number of sites present (Equation

12).

_ Zi ViILTi] (12)

.7 _
zi [LTi]

where: [LTi] = total molar concentration of ligand type i.

 

 

 

_ K'. [M ] [L .] (13)
v = 21 i ’ f . Ti
1+Kimp In)
_ 1 I i? (14)
K, = -
[Mf] \1- v

If one type of site is assumed, the Scatchard relationship can be
expressed by Equation 15.

w .151 (15)

1 + K'i [Mf] [LT]

<1
I!

where: L1 - total molar concentration of one ligand type.

_ 18
v _ (16)
———- = K'i (n. - v)

where: ni = number of binding sites per molecule of HS.

Graphical interpretation of Equation 16 is given by a plot of

<1

Tfi—T as a function of G and can provide an estimate of ni and K'i.

It is assumed that only one type of binding site is present if the

HI

plot is linear; if the plot is curvilinear, the metal is bound by more
than one type of site. When more than one type of binding site is
indicated, values for ni and K’i cannot be resolved graphically by
extrapolating linear regions of the curve because each type of site
contributes in) the nearly linear portions of both ends of the curve
(Giesy 1980). Values for 111 and K'i can be estimated by computation
using iterative nonlinear least squares techniques. Generally two
types of metal binding sites (n1 and n2) and consequently two

conditional stability constants (K’ and K'z) result from a

1
curvilinear Scatchard plot. One type of site (n1) is less common than
the other type (n2), but forms a stronger bond with the metal ion.
The conditional stability constant (K') will be greater for the type
of site that has fewer available binding sites.

The Scatchard analyses for the interaction of metals with HS are
generally curvilinear, as seen in this study. In an attempt to more
accurately describe this relationship, a number of authors have used
Scatchard relationships which include multiple types of sites. If two

discrete types of sites are assumed, the relationship is described by

Equation 17.

19

101 Inf) . [LTll K', Inf) . [LTZI (17)
I + I
1 + K 1 [M [LT] 1 + K 2 [Hf] [LT]

 

 

 

 

<1
ll

f]

While this relationship successfully describes the empirical
data, it does not necessarily mean that there are actually two
discrete types of sites present. Perdue and Lytle (1983) show that
the average conditional stability constants derived in this way are
not constant but vary as a function of the total metal to total ligand
ratio (MT:LT). Hence the estimates of the average conditional
stability constants derived from two component Scatchard analysis are
useful only in the range of M :L for which they were determined and

T T

cannot be extrapolated to other MTzLT ratios. Because of the minimum
detection limits in the analytical techniques used by many researchers

today, the experimental meta1:ligand ratio generally must be much

greater than in natural surface waters.

Continuous Multiligand Models

 

Average stability constants (K) calculated. for £1 multiligand
mixture such as aquatic HS are not constant because the reference
ligand will most certainly have a different affinity for the metal ion
than will the experimental ligand; consequently the ratio of the two
ligands will vary as the overall metal to ligand ratio changes in the
solution. It follows that the average conditional stability constant
(K') would vary with changes in the composition of the solution, and
therefore should not be considered a constant at all (Perdue and Lytle
1983). However, with excess amounts of reference and experimental
aquatic HS or extremely low concentrations of metal, the overall metal

to ligand ratio should not change perceptibly. On the other hand, the

20
analytical technique may require larger metal concentrations to be
used which are not characteristic of those found in the natural
environment (Perdue and Lytle 1983).

The Scatchard analysis used to evaluate the metal complexing
ability of HS in this and other studies (Mantoura and Riley 1975; Guy
and Chakrabarti 1976; Mantoura et al. 1978; Giesy 1980; Kribek and
Podlaha 1980; Li et al. 1980; Alberts and Giesy 1983; Shuman et al.
1983) has been criticized for its lack of rigor in multiligand
systems. Perdue and Lytle (1983) object to the use of the Scatchard
equation for multiligand binding, principally because (1) it does not
meet the requirement for a known molar (M) concentration of the ligand
and (2) the four curve-fitting parameters can lead to the erroneous
conclusion that aquatic HS contain only two nonidentical binding
sites. Perdue and Lytle (1983) demonstrate that the average
conditional stability constant derived from a four parameter Scatchard
model, which assumes two classes of sites, is not independent of the
MT:LT ratio. They also demonstrate that parameters which are
determined by using a curve-fitting technique are average stability
constants for two classes of sites, each of which can be made up of a
number of similar but nonidentical metal binding sites. To be useful
in predicting organic-inorganic relationships in surface waters,
Perdue and Lytle (1983) indicated that stability constants must be
derived under metalzligand ratios similar to those encountered in
surface waters.

MINEQL is a computer program for the calculation of chemical
equilibria in aqueous systems. In their study, Perdue and Lytle

(1983) incorporated data from the MINEQL simulation program for a

21
hypothetical continuum, in this case the Gaussian distribution. They
noted that a continuous distribution model may be used to predict the
extent of metal complexation at metal concentrations representative of
those in the environment from laboratory results obtained at much
higher metal concentrations (Perdue and Lytle 1983).

Scheinberg (1982) noted that the measurements must be made at
equilibrium and that the molarities of both metal and ligand must be
known. If these conditions are not met, a Scatchard analysis of the
results of nmmal-ligand binding may be of heuristic value, but will
not yield the number and stability constants of the binding sites on
the binding molecule.

Hunston (1975) compared Scatchard analysis to a continuous
distribution of binding (i.e., stability) constants, which does not
require the assumption of :1 functional form for the distribution or
knowledge of the number of classes of independent binding sites. The
general continuous distribution model is not limited by the number of
parameters as is :1 more discrete distribution of binding constants.
When values of ni are plotted as a function of the binding site
constant, the resulting distribution of the binding site constants can
be used to characterize the binding reaction. The distribution of
binding constants is best characterized by an average binding constant
and a standard deviation, which serves as a measure of the
dissimilarity of the binding constants (Hunston 1975).

Posner (1966) reported that proton binding by HS was efficiently
described by a continuous multiligand distribution model. While no
data have been published that suggest HS have a Gaussian distribution

of metal-binding ligands, physical and chemical binding

22
characteristics tmf monovalent and divalent cations should be
sufficiently alike to pursue UO§+-aquatic HS binding described by a
continuous distribution model.

The Gaussian distribution is a symmetrical two parameter
frequency distribution described by a central tendency or mean and a
measure of dispersion around that mean, which is referred to as the
standard deviation. The parametric mean (p) and standard deviation
(0) are related when combined with the discrete Scatchard model

(Equation 18).

 

 

(18)
00 log K' _ _ i 2
[ML] : G - 1 [Mf]10 e 0.5 ( E 103 ). d 1 K'
calc — log K' 08
[LT] 0 2n 1+[M 110
f
-00

where: G calc = empirical ratio ML:LT at each point in a titration

m = p - 40

-m = p + 40.

This frequency distribution can be evaluated numerically by
substituting p and 0 such that the residual sums of squares (RSS) is

minimized (Equation 19).

 

_ - - 2
RSS - 21 (vcalc Vexp) (19)
\7
exp
where: G = calculated 0
calc
G = expected theoretical J.

exp

23

A Gaussian distribution model has certain advantages over a
discrete two ligand model. First, the symmetrical Gaussian function
makes it theoretically possible to define the shape of the curve at
higher metal concentrations (\7 _>_ 0.5) and to extrapolate to lower
concentrations (6 -> 0). Second, from a purely empirical point of
view, the Gaussian distribution model with two curve-fitting
parameters is less restrained than the two component Scatchard
equation with four curve-fitting parameters (Perdue and Lytle 1983).
Iterative calculations optimize several parameters (log K’ or K',
0108 K" and I7) and are necessary in both the Gaussian distribution
and Scatchard models. Like the Scatchard technique, the continuous
distribution model is not able to reduce multiligand chemical

equilibrium systems to thermodynamic equilibrium constants (Perdue and

Lytle 1983).

Binding Capacitj

 

To determine the conditional stability constants of individual
metal-ligand complexes or an average conditional stability constant
for a ligand mixture, one needs to know the amount of uncomplexed (Mf)
and complexed (ML) metal as well as the total ligand concentration
(LT) and the amount of uncomplexed (Lf) and complexed (ML) ligand.
This is relatively easy for single ligand systems, when the structure
and molecular weight of the ligand are known. In the case of HS no
exact molecular weight can be assigned. Some authors have assigned
nominal molecular sizes based on ultrafiltration (Gjessing 1970; Kwak
et al. 1977; Giesy and Briese 1977, 1978a,b; Giesy et al. 1977; Giesy

and Paine 1977; Giesy and Alberts 1984).

24
The molar concentration of the total number of metal binding site
available has to be known.

The maximum binding capacity (BC) is the metal complexing
capability by active anionic sites in a solution (Miller and Ohlrogge
1958; Zunino et al. 1972). BC is not necessarily equal to the total
number of sites because stereochemical configurations of the polymeric
humic ligand may reduce the availability of some sites for metal
binding. BC is defined for specific environmental conditions (e.g.,
pH, ionic strength, temperature) and in surface waters can be either
generic (i.e., total cationic complexation by both. inorganic {and
organic anionic components of surface waters) or component specific
(i.e., BC measured for selected inorganic or organic anionic
components).

There are several ways to determine BC. The most direct
technique is to titrate a solution of ligand with a metal and
determine the relative concentrations of complexed versus uncomplexed
metal such as mg UO§+/£ of HS. The number of metal binding sites can
be expressed as M/mg or M/B. This technique is imperfect because the
humic material can be precipitated at higher concentrations of metal,
which may cause stereochemical changes that alter the number' of
available sites measured under these conditions (Reuter and Perdue
1977).

An alternative method is to calculate the molar concentration of
binding sites (ni) from a Scatchard analysis. This can be done by
plotting v/mg HS (Alberts et al., in press). Because of the
limitations of the Scatchard technique this method was not used in

this study.

25

A third technique, which eliminates a number of problems
associated with both the previous techniques, is to titrate» the
natural water or solution of interest with U0:+ solution (Alberts et
al., in press). The maximum BC is determined by fitting a first order
saturation model to the data (Figure 1).

A fourth method of determining BC is to calculate the maximum
reduction of U0:+ concentrations by titrating U0:+ standard solutions

with natural water or solution of interest (Giesy et al. 1978;

Equation 20).

 

_ (ci)(vs)-(cf)(vs + VT) (20)
BC —
VT
where: BC = binding capacity (pg-atoms UO§+/ml solution)

. . . 2+ . .

C. = 1ni$181 free UO ion concentration (pg-atoms
1 .

U02 /ml solution)

VS = volume of UO§+ standard solution (ml)
. 2+ . . 2+

C = final free UO2 ion concentration (Pg-atoms ”U02 /ml

solution)

VT = volume of titrant added (ml of natural water or solution
of interest).

26

 

Figure 1. Observed versus predicted concentrations of uranium
complexed by Aldrich humic acid as a function of total

uranium added to the solution (M; pH 4.5; 0.01 M Ac-).

 

27

5.70.3: .23

o.»

.L ounwfim

0..

 

 

0.0 _ o .m n K 06 n f
a _ q

u -

33:33 a o
VO>bOQQO u 0
2.22.3 9:2: 2:3 u x

32:
052.3 3 .353. 03.32.
555.35. 2.02.3 052.3 u on

a... 33:95 .22 u .92
54 535:: 3.8353 u 4.2

.h!. onICUQ .42

u

0
O.

o
o
o...
0
o

i

OOwb‘DOQ’nN-

Q

 

(w,,0I X) n paxaldwoo

28

+
were separated by

+
column ion exchange techniques. Therefore, kinetics of U0: -HS
2+

2

solutions are reaction dependent, optimal flow rates must be

In this study complexed and humic-bound U0:

reactions could affect calculated BC. Because the BC of U0 HS
established for experiments which use ion exchange columns. This flow
rate should sustain equilibrium conditions in the column. In this
study, the flow rate was optimized to prevent decoupling of

humic-bound U02+ ions by Chelex-lOO® resin which has a greater

2
affinity than humics for U0:+ ions. A flow rate of greater than 16 ml
of UO§+-Aldrich HA solution/min is required to minimize potential
disequilibrium conditions as the UO§+-HS pass through the resin

column (Figure 2). This elution rate is considerably faster than that
reported (4-6 ml/min) for similar studies with divalent transition
metals (Giesy 1980). The difference may be attributed to (1) relative
affinities of different types of ion exchange resins for the same
model or (2) relative affinity of the same resin for different metals.

Chelex-lOO® is reported to have an affinity for the U0:+ ion which is

exceeded only by the affinity for Hg2+ and Cu2+ ions (BioRad Labora-

tories 1972). This strong affinity for the U02+ ion compared to other

2

divalent cations necessitates a rapid flow rate through the resin
column in order to maintain equilibrium conditions. If the flow rate

through the column is too slow, Chelex-lOO® resin can remove U0:+ ions

from UO§+-HS complexes (Figure 2). However, a flow rate of 16 ml/min

allows quantitative removal of 'the free 'U02+

2 (Hathaway and James

1975).

29

BC is estimated from a plot of the bound U0:+ as a function of
the total U0:+ concentration. A least squares approximation of BC can
be obtained (Equation 21).

[ML] = BC(1-e “”11” (21)
where: [ML] = humic bound U0:+ (M)

BC = maximum possible number of binding sites

[MT] = total U0:+ added to solution (M)

A = curve fitting constant.

30

Figure 2. Concentration of uranium (M) eluted through a Chelex-lOO®
ion exchange resin column as a function of flow rate of the
solution through the column (ml/min; pH 4.5; 0.01 M Ac_;
3.5 mg C/2, dialyzed Aldrich HA).

 

31

3252.5 5.5.60 00. 9.8.28 ‘93:: 2.: Sci

.N munwwm

 

0. t. N. 0. o 0 ¢ N

- w a) _ . a q nfivdv

m 0.0
W
l 0.. m.
o O
m.
nu
. 30. N
n.)
m
u o N.» 1..
_W

o. o 0.?

 

 

oé

RESULTS

The maximum binding capacity (BC) of Aldrich HA was determined

with uranium concentrations from 3.15 x 10.7 M to 2.10 x 10"5 M U0:+
(Figure 1). The BC of a 3.5 mg C/2 solution was estimated to be
1.14 x 10.6 M U02+ with an asymptotic standard error of 5.0 x 10.-8 M

2
U0:+ (Table 3). The 95% confidence interval (CI) is i 1.1 x 10"7 M

U02+. A highly significant F-statistic was calculated for the regres-

2
sion of the concentration of UO?’ eluted through the ion exchange
2+

column as a function of the concentration of U02 added to the column

(F 1118; P 3 0.0001). The BC of Aldrich HA was thus

2,14(0.05)

4.8 x 10"7 E UO§+lmg C.

The Scatchard analysis was non-linear over the entire [MT]:[LT]
range titrated (Figure 3). The estimate of the average conditional

stability constant (K'), was determined from a Scatchard analysis of

data between U0:+ concentrations of 5.25 x 10-8

x 10'7 u (50 ppb) U0

M (12.5 PPb) and 2.10
3+. These concentrations fall within the range of
surface waters in.£1 non-uraniferous area (Szalay and Samsoni 1969).
The Scatchard analysis over this range of metal:ligand ratios is
sufficiently linear to allow the estimation of K' which was determined

to be 2.43 x 107 (or log K' of 7.38) from a single component Scatchard

analysis (Figure 4 and Equation 22).

32

33

7 6

G = 2.43 x 10 9 + 2.65 x 10 (22)

 

[ML]/BC

a
:r
(b
H
(b
<
II

[ML] = molar concentration of complexed metal
BC = maximim binding capacity
[Mf] = molar concentration of uncomplexed metal (UO§+).

34

 

 

 

 

 

 

 

 

 

Table 3. Summary statistics for nonlinear least squares analysis to
estimate the 2 can (p) and standard deviat' n (0) of
uncomplexed UO (Mf) and observed values of UO2 complexed
by humic acid (ML).

Source MF Sums of Squares (SS) Mean Square (MS)

Regression 2 2.85 x 10.12 1.42 x 10-12

Residual 14 1.78 x 10'14 1.27 x 10'15

Uncorrected total 16 2.87 x 10'-12

(Corrected total) 15 1.28 x 10-12

Asymptotic 95%
Asymptotic Confidence Interval
Parameter Estimate Standard Error Lower Upper
K* 2.64 x 105 2.18 x 104 2.18 x 105 3.11 x 105
BC** 1.14 x 10‘6 5.0 x 10’8 1.03 x 10"6 1.25 x 10"6
Asymptotic Correlation Matrix of
the Parameters
K BC
K 1.000 -0.877
BC -0.877 1.000
*K is a curve-fitting constant.

**BC is the maximum possible number

of binding sites.

35

2+
2 at

(pH 4.5;
5

Figure 3. Scatchard plot of Aldrich humic acid binding of U0

concentrations ranging from 4.2 x 10.8 H 1103+

3.5 mg C/2; 0.01 )1 Ac’). r_1 to 4.2 x 10'8 M to 2.1 x 10'

2+
11 1102 .

 

36

 

 

-m
i.
19
18
47
. $6
. 15
C 14.
0.
. 13
C C
o .12
O
.0. 1.!
o o o o
O O O
p . r p - p b . b . F n . . . . o
8 mm M W. O 8 6 4 2 0

70110")

Figure 3.

Figure 4.

37

The Scatchard estimate of the overall average conditional
stability constant (K') in the region of interest based on
the ratio of complexed uranium to maximum number of binding

sites for Aldrich humic acid (9) and uncomplexed UO§+ (Mf).

38

 

 

 

40)-
36+ 4 ' .
Jim-2.43 x 10’ 17 + 2.65 1110‘
32. M!
N 312'
24]-
O‘- -
.' -shpe-K'
0
IS-
0
I2- . O\
8" .\\.1
4..
N
L, 1 1 1A 1 1 1 1 1 J
00 IO 20
ii (xlO'z)

Figure 4.

39

A Gaussian estimate of K' was determined to be 7.11 x 106 (or log

8

K' of 6.85) for the U02+ concentrations of 5.25 x 10- M to 2.10 x

2
10"7 M (Figure 5). The standard deviation (OK’) of the estimated

distribution of stability constants was calculated to be 1.38 x 106

(or log (r-, of 6.14). The standard error of the estimate of K' was

4.6 x 105 (95% CI for K' = i 1.0 x 106; n=8). The logs of the observed
conditional stability constants were plotted as a function of the log

[UO§+]:BC ratio. The log of observed K' as a function of log [MT]:[LT]
ratio shows that as the U0:+ ion is added, successively weaker sites

are filled (Figure 5 and Equation 23).

[M 1 (23)
T ) + 5.4

 

log K’ = -0.6913 (log

1LT]

where: K' overall conditional stability constant

2+
2

total concentration of HS binding sites.

[MT] total concentration of U0

[L

40

Figure 5. Regression of the observed conditional stability

constant (K') as predicted by a Gaussian distribution

in the region of interest as a function of the metal

2
(U02+) to ligand (Aldrich humic acid) ratio.

41

.m muswfim

 

     

P3.
0 0.0... 0...: 0.... o.~1
- d u q
o.» - .. on
can... Ah.“ so n.mo.01 192.333 .0. 90..

to h: .
_I
o
.o
M. 0.0 r DIG/O o .. 06
o /
e ./ .
a o
u
a
nr

5:33.... 5.3.30 00.9
o... . Ea... 23:02.. .~.. - o...

 

.0... 22.22%. 2.2.2.231.

 

:3... 9225.8 .w. oo..\\‘1 on...

DISCUSSION

Typical inland surface waters of the coastal plain of the
southeastern U.S.A. include swamps with dense vegetation and streams
and rivers which drain highly leached, low relief terrain. These
waters are regionally termed blackwaters and have pH values ranging
from 3.8 to 6.8 and low ionic strengths. Their brown color is due to
naturally occurring, refractory, organic compounds, known collectively
as humic substances (HS). HS are a diverse group of polycarboxylic,
polyphenolic compounds with both aromatic and aliphatic components
which are believed to result from microbial degradation and chemical
polymerization of former vegetative components such as lignin (Flaig
1964; Trojanowski et al. 1977; Haider et al. 1978), but are resistant
to complete microbial degradation (Christman and Ghassemi 1966;
Felbeck 1971).

Humic acids (HA) have been operationally defined as naturally
occurring organics which are water soluble in basic solutions; fulvic
acids (FA) are those refractory, colored. organics that. are ‘water
soluble in acidic or basic solutions. FA are relatively small organic
molecules with nominal molecular weights (MW) ranging from 300 - 2000
(0.0009 - 0.0012pm). HA range from 5000 - 100,000 MW (0.0013 - 0.0052
pm) (Steelink 1977).

Several investigators (Schindler and Alberts 1974; Giesy' and
Briese 1978a) have reported that in low pH surface waters HS span a
wide range of MW, but southeastern blackwaters generally are

42

43

characterized by a high percentage of very small MW HS. Giesy and
Briese (1977) report that 70% of the organic carbon (C) content of the
Okefenokee Swamp water is < 0.0009 pm with an additional 27% in the
next smallest size fraction (0.0009 ima‘< F 3 0.0015 pm). These same
authors report that greater than 60% of the dissolved organic carbon
(DOC; < 0.45 pm) in two South Carolina streams is < 0.0015 pm (Giesy
and Briese 1978a). Alberts and Evans (1979) show that greater than
80% of the DOC in four southeastern coastal plain rivers is < 0.0013
pm in diameter. Because of their prominence, FA fractions of HS may
be a controlling variable in the mobility of trace metals in
southeastern surface waters.

The importance of HS in the geochemical cycling of inorganic
elements in surface waters of the southeastern U.S.A. has been
established (Beck et al. 1974; Schindler and Alberts 1974; Casagrande
and. Erchull 1976; Reuter' and ZPerdue 1977; Giesy' and IBriese 1977,
1978a). Despite the absence of descriptions of absolute structures
for HS, there is general concensus that two functional groups are
responsible for much of the complexing of metals in these freshwater
systems. The carboxylate and phenolate groups occur in both aliphatic
and aromatic configurations. While early research on HS supported a
structure with an aromatic carbon backbone (Christman and Ghassemi
1966; Langford et al. 1983), more recent investigations suggest a less
aromatic, more aliphatic structure (Stuermer and Payne 1976; Wilson
and Goh 1977; Ruggiero et al. 1979, 1980). The phthalate and
salicylate configurations of HS have been proposed as the most likely
to chelate metals (Figure 6). Metal ions may be complexed by
carboxylate and/or phenolate groups on single or two different humic

molecules.

44

Figure 6. Types of metal complexation hypothesized for humic acid
and fulvic acid (redrawn from Manning and Ramamoorthy
1973).

45

255 22 235....

S

.o wuswwm

33.0 22 212:5

46

Acid dissociation constants (Ka) for the two major types of
exchangeable proton sites on humic molecules are similar to the K8
values of phthalic acid and salicylic acid (Table 4). More recent
thermodynamic data tend to support salicylate type bonding, but not
phthalate type bonding (Choppin and Kullberg 1978). Carboxylate
groups attached to different aromatic rings may act independently of
each other, not in an ortho-carboxylic (i.e., phthalate) configuration
(Choppin and Kullberg 1978; Alberts and Giesy 1983).

Since HS are the principal ligands found in southeastern surface

waters and the complexing of U02+ ions by humics has been reported to

2

be strong relative to competing ligands (Lamar 1968; Kribek and

Podlaha 1980), the purpose of this study was to investigate the

binding of U0:+

simulate those found in inland surface waters of the Southeast (Beck

by HS under pH and ionic strength conditions that

et al. 1974; Tilly 1975).

In natural waters metal complexation or adsorption can occur by
interaction of U0:+ with different surfaces, such as ferric and
manganic oxyhydroxides, inorganic anions, clays and organics (Moskvin
et al. 1967; Langmuir 1978; Giblin 1980; Dongarra and Langmuir 1980;
Borovec 1981; Giblin et al. 1981; Tipping 1981; Tsunashima et al.

1981). Since this study focused on U02+-HS complexation, other

2
surfaces of adsorption and dissolved inorganic anions commonly found
in surface waters were intentionally not introduced into the
experimental solutions. For this reason, the results provide an

estimate of the BC under laboratory controlled conditions and should

not be interpreted as data that would result from field sampling.

47

Table 4. Acid dissociation (ionization) constants (pK ) for two major
types of proton sites on simple aromatic acids or aromatic
acidic components of humic and fulvic acid molecules. Two
or more constants indicate the first, second, and possibly
the third dissociation constants, respectively.

 

Acid pKa Study
benzoic 4.20 Perdue 1978
4.01 Choppin & Kullberg 1978
phenol 9.95 Wilson & Kinney 1977
9.78 Perdue 1978
phthalic 2.76; 4.92 Wilson & Kinney 1977
2.7; 5.0 Choppin & Kullberg 1978
o-hydroxybenzoic 3.86; 13.1 Choppin & Kullberg 1978
(salicylic) 2.40 Gamble 1970
2.97; 13.59 Wilson & Kinney 1977
2.98; 13.59 Perdue 1978
m-hydroxybenzoic 4.1; 9.9 Choppin & Kullberg 1978
4.52 Gamble 1970
p-hydroxybenzoic 4.5; 9.3 Choppin & Kullberg 1978
4.52 Gamble 1970
4.58; 9.24 Perdue 1978
humic/fulvic 3.1; 5.0; 5.0 Borggaard 1974
4.0; 9.0 Choppin & Kullberg 1978
4.23; 8kfi0H Wilson & Kinney 1977
10.5 (p ) Perdue 1978

 

48

Kribek and Podlaha (1980) reported that the stability constant
for the UO§+-HS complex did not appear to depend on ionic strength, at
least up to 0.5 M NaClOa. In the study reported here, the ionic
strength of the experimental solutions was an order of magnitude less
than the ionic strength used by Kribek and Podlaha (1980; I = 0.01 M
versus I = 0.1 M, respectively).

A weakly acidic catonic chelating resin (Chelex-100®) was
selected for the study because of structural similarity between the
resin matrix and solute (Aldrich HA solution). Chelex-100® is highly
selective for divalent ions relative to monovalent cations. The
exchange kinetics of Chelex-100® are governed by second order
kinetics rather than diffusion. The aromatic nature of both the resin
and the solute increase the probability that ion exchange, not
adsorption, will occur. Chelex-lOdg has the ability to function in
weakly acidic (pH 3 4), neutral or basic solutions. At low pH,
®

Chelex-lOO® acts as an anion exchanger. The titration of Chelex-100

produces the following Zwitterionic forms as a function of pH:

CHZCOOH CH2 coou CH2 coo' cnzcoo'
¢-fH2NH+N0; ¢- -CH: NH+ ¢- -Cfi/HH+2 ¢-CH2N
CH coon CH2 CO0H {CH coo' ‘\\CH coo'
2 2 2 2
pH 2.21 pH 3.99 pH 7.41 pH 12.30

At low pH some of the material in the HS solute may be complexed by

the resin because of the anion affinity of the Chelex-100® resin and

49

the net negatively charged characteristics of HS. The amount of loss
of HS to the Chelex-IOO® was monitored by measuring the concentration
of HS before and after passing through the column. At pH 4.5 of this
study, Aldrich HA concentrations, as measured by fluorescence,
ultraviolet absorption and CO2 formation on combustion before and
after passing through the column, were not significantly different
(J. P. Giesy, pers. com.). The loss of HA to the resin due to
adsorption was small.

The Na+ form of Chelex-100® will swell in water because of the
hydration of functional groups. Therefore, when the resin was added
to deionized water at pH 4.5, the pH of the water increased,
indicating that the Chelex-100® resin. had ‘been. protonated.
Consequently, buffered solutions were used. to maintain. the
experimental pH. An acetate buffer prepared from acetic acid
(CH3C00H) and sodium acetate (CH3C00Na) was selected because (1) the
stability constant between U0§+ and the acetate anion (Ac-) is small
(Sillen and Martell 1964, 1971) and (2) the desired pH could be
obtained by increasing the ionic strength of the experimental solution
by less than two orders of magnitude over that reported for
southeastern U.S.A. surface waters (Beck et al. 1974; Tilly 1975).

The small stability constant between U02+ and Ac- permits the use of

2
the acetate buffer without strong competition between Ac- and aquatic
+
HS for U0: . Other simple organic acids form stronger complexes with
2+

U02 and therefore might compete with HS for potential binding sites
(Sillen and Martell 1964, 1971).
A 0.22 pm pore diameter Ac- filter was placed over the

glass-fritted section of each ion exchange column before the volume of

50
buffered resin was added to prevent fragments of the resin from
eluting into the filtrate. Van den Berg and Kramer (1979) noted that

Chelex-100® fragments pass through a 0.45 pm filter and therefore

+
could, in this study, increase the concentration of U0: in the
+
filtrate. This would produce an overestimate of humic-bound U0§

+
since it is presumed that all of the U0: in the eluate is bound to

the HS instead of to HS and resin fragments. However, with the
filters, very small pieces of Ac- filter may pass through the fritted
section of the column and into the sample; this would result in an

increased carbon concentration in the sample. This error was greater

than that resulting from fragments of U0§+ bound Chelex-lOO® resin

passing through 0.45 pm filter (Table 5).
Calculations of binding capacities may also be affected by the

kinetics of metal-ligand reactions (Giesy 1980). Because of greater

® for U03+ ions, the solute flow rate must

be sufficiently fast to prevent the resin from complexing humic bound

affinity of the Chelex-lOO

U0:+ ions. If this does not occur, the results will not reflect the

true binding potential of the HS for U02+ and consequently the HS

2

binding capacity will be underestimated. Conversely, if the flow rate
through the ion exchange column is too fast, the Chelex-100® resin

will not have sufficient opportunity to bind U02+ not complexed by HS,

2

and the binding capacity of the HS for U02+ will be overestimated. An

2
optimal flow rate of the metal-ligand solute must be established in
order to approximate equilibrium conditions. In very similar columns,
optimal flow rates of 4-6 ml/min have been reported in binding

capacity studies with divalent cations and aquatic HS (Giesy 1980).

However, because of the strong affinity of Chelex-100® resin for the

51

Table 5. Organic carbon analyses of replicate samples of dialyzed
Aldrich HA (3.5 mg C/2)

 

Carbon concentrations in prepared solutions (ml C/fi)

 

 

 

Before elu®ion through After elut'on through
Chelex-100 column Chelex-IOO column
Without 0.22 pm 3.3 3.7
Ac filter 3.4 3.6
With 0.22 pm 3.3 4.6
Ac filter 3.4 5.3

 

52

UO§+ ion, flow rates between 17.5 and 18.8 ml/min were required in

this study (Table 1).

The optimal flow rate was determined by eluting the buffered
3+ and Aldrich HA through the Chelex-lOO® resin
column at progressively faster rates. The point at which the U0:+

concentration reached a plateau as the flow rate continued to increase

solute containing U0

was defined as the optimal flow rate (Figure 2). At this flow rate

uncomplexed U0:+ was removed from solution without allowing enough
contact time for U0:+ complexed to HA to be removed. Partitioning

complexed metal species from uncomplexed metal species by ion exchange
can only be used when studying metal-ligand interactions with
relatively high stability constants.

Several studies of the surface waters in the southeastern U.S.A.
indicate HS have functional groups that are primarily carboxyl groups
with fewer phenolic groups; Aldrich HA acidity, in contrast, is
largely phenolic (Table 6; Beck et al. 1974; Reuter and Perdue 1977;
Perdue 1979; Giesy and Alberts 1983). Although not necessarily
representative of HS in all inland surface waters of the Southeast,
Aldrich Ifll was used in this study instead of natural aquatic HS for
several reasons. Aldrich HA has been used as a "reference" HA in
other studies of BC and conditional stability constants (e.g., Pott et
al., in press). Because it is commercially available and identified
by lot number, Aldrich HA allows direct comparison of results to those
of other metal-humic binding studies, whether the same or a different
metal is studied. In addition, the same HA can be used for subsequent
studies of U0:+ -HA complexing under different conditions, thereby

eliminating one potential variable. The ash content of this

53

Table 6. Acidities of some oxy-functional groups in humic substances
(meq/g; from Giesy and Alberts 1984).

 

Soil HA1
Soil FA1

Water H82

Upper Three Runs Creek3

12.0 nm 3 HS 3 0.9 nm

Aldrich humic acids

Total
Acidity

 

7.9 2 1.84

H-

12.8 1.2
10.4 i 1.3

4.2 i 0.4

2.91 i 0.25

Carboxylic
Acidity

 

3.7 i 1.4

8.9

1+

0.3
6.0 i 1.3

3.3 i 0.4

0.42 i 0.14

Phenolic
Acidity

 

4.2

H-

1.5
3.9 i 1.6
4.4 i 0.0

1.0 i 0.4

2.49 i 0.29

 

1Schnitzer and Khan, 1972

2Perdue, 1979

3HS = humic substance isolated by ultrafiltration

4: 1 S.D. (N = S, 3, 2, 2 and 17, respectively)

54
particular an; of Aldrich HA is 67%, which is comparable to the ash
content of naturally occurring aquatic HS in a blackwater stream in
South Carolina for which the ash content is 60-70% (Giesy and Briese
1977).

Despite the fact that numerous techniques are available to
quantify uranium concentrations in aqueous solutions, only a few are
sensitive enough to accurately quantify those U concentrations usually
found in surface waters (Table 7). Most. methods used a
preconcentration or separation step which improves the minimum
detection limit and eliminates interfering components in the solute
(Campen and Bachmann 1979). Fluorometry is a sensitive analytical
method for I] and several variations of the fluorometric method have
been devised (Table 7). Serious interferences in fluorometric
analyses of U from natural waters is caused by the presence of HS
(Robbins 1978; Campen and Bachmann 1979; Kaminski et al. 1981).

Recently a new fluorometric technique which overcomes DOC
interference has been developed. for’ the Idetermination. of ‘trace U
concentrations in surface waters. It has been tested extensively in
the National Uranium Resource Evaluation (NURE) Program. The

technique uses a Scintrex Model UA-3 laser fluorometer (Scintrex Ltd.,
2+
2 .
nitrogen laser emits short intense pulses of ultraviolet radiation

(UV) (A = 337 nm) that excite the U02

2
+
U0: compounds fluoresce in the green region of the spectrum with

Toronto, Canada) to measure laser-induced fluorescence of U0 A

+
ion. When excited by UV light,

characteristic emission peaks at 496, 519, and 544 nm (Figure 7 from

Campen and Bachmann 1979).

55

Table 7. Analytical techniques for low concentrations of uranium.

 

Analytical Technique

 

Titrametric, micro-
gravimetric

Potentiometric

Gamma spectroscopy

Colorometric
(photometric)

Liquid scintillation
Spectrophotometric
Delayed neutron
counting

Fluorescence
(fluorometric)

Phosphorescence
(spectrofluorometric)

Minimum Detectable
Limit (MDL)

 

1.2 mg U/ml*

mg U/ml quantities
1 mg/ml

10 ng total U

0.05 ng/ml

0.6 mg/ml

O

.1 mg/ml

2 ng/ml*

4...).
an

0.5 ng/ml
5 ng/ml
2 ng/ml
0.05 ng/ml

0.05 ng/ml
0.04 pg/ml

0.01 pg/ml

10 pg/ml

Reference

Kribek and Podlaha
(1980)

Bodnar (1980)

Tieh et al. (1980)
Gladney et a1. (1978)
Gladney et a1. (1976)

Jablonski and Leyden
(1978)

Horrocks (1974)

Halbach et a1. (1980)
Nash et al. (1981)

Brits and Das (1978)

Li et al. (1980)

Hathaway and James
(1975)

Campen and Bachmann
(1979)

Robbins (1978)

Johnson and Wright
(1981)

Perry et al. (1981)

Kaminski et a1.
(1981)

 

*MDL not listed; concentrations listed are experimental concentrations
**No experimental concentration listed

56

Figure 7. Emission spectrum of an aqueous solution of uranium

(from Campen and Bachmann 1979).

57

X .‘c ' 28°

 

 

 

 

 

 

 

 

 

 

496
H 519
- H
B
O
'3
.3
E u
,. 544
.3
.3 U
5
260 360 460 5'00 660

A (nm)

Figure 7.

58

Emission spectra of most U02+ salts are quite similar, implying

2
that the coordinating ligand has little effect (Hi the strong O-U-0

+
bonds (Robbins 1978). Intensity of the emission of various U0:

species in solution varies considerably. Therefore, to insure maximum
intensity of fluoresence, Fluran®, a proprietary reagent composed of a
chelating agent and a fluorescing agent, is added to the U0:+ solution
before laser excitation of the species occurs (Wallach et al. 1959;
Campen and Bachmann 1979). The primary function of the F1uran®
solution is the formation of a single fluorescent U0:+

also complexes other metallic ions in solutions to reduce their

species, but it

quenching effects on the intensity of the U02+ luminescence (Robbins

2

1978). Fluran® strongly complexes the various U05...

solution into a: single uranyl phosphate species and provides a high

species in the

luminescent yield. Because the stability of the complex is pH
dependent, Fluran® is sufficiently buffered to maintain a stable pH to
insure the formation of a single species. Campen and B;chmann (1979)
have reported signal intensity as a function of the volume of Fluran®
added to the sample (Figure 8 from Campen and Bachmann 1979).

When excited by a nitrogen laser at 337 nm, HS exhibits an
intense blue fluorescence with a maximum intensity near 400 nm
(Figure 9 from Robbins 1978). Since the U0:+ ion fluoresces in the
green region, any fluorescent interference from organic matter, which
is emitted in the blue region, can be removed by using a green filter
between the sample and the photomultiplier (Campen and Bachmann 1979).

Several other techniques were used to reduce interference. A

standard additions technique was used to minimize matrix interference.

In addition, a ‘photo-oxidation. procedure was used to remove

59

Figure 8. Relative signal intensity of Scintrex Model UA-3
fluorometer as a function of amount of Fluran® added to

uranyl solution (from Campen and Bachmann 1979).

60

 

 

‘1.»

800

500
Fluran added (pl)

 

 

LAT

223.83 2.2.2:.

Figure 8.

61

Figure 9. Fluorescence spectrum of water samples containing humic

substances (from Robbins 1978).

lniensity (emission)

 

 

 

,, , ,, uranium
0'00"": fluorescence
fluorescence
laser ' I
l I
n l I
H: a
l 1 A
x u w
I \
\
z '\
350 460 550 coo x (nm)
LJM I (if—ELLA

Figure 9.

 

63

interference by HS. Carbon - carbon bonds in both humic molecules and
acetate molecules are broken by high intensity UV radiation, in this
study provided DY’EI La Jo11a Scientific Model PO-14 photo-oxidation
unit. Other investigators have used photo-oxidation as a technique to
destroy organic molecules and to release metals bound by organics
(Blutstein and Shaw 1981; Sunda and Hanson 1979). This method is
preferred over a permanganate oxidation procedure which supplants one
interference (HS) with another (Mn2+).

Laser fluorometry further reduces the potential interference due
to fluorescence of HS by utilizing the differences in the lifetime of
U0§+ fluorescence and that of humics in solution. The fluorescence of
most naturally occurring organics ceases quickly. Lifetimes are, at
most, several tens of nanoseconds. In contrast, fluorescence: of
dilute concentrations of U0:+ persist typically for several tens of
microseconds (Robbins 1978). An electronic gating system, triggered

by the laser, delays accepting signals from the photomultiplier until

after the fluorescence from the humics has substantially ceased. The
2+
2 .

Instrument response is generally very rapid after the addition of

recorded signal is due almost entirely to the U0

Fluran® to a solution that contains U; one exception is in the case of
organic-rich samples. The extraction of U from organic species is
slow (Robbins 1978). Analytical times for samples with greater
concentrations of HS must be increased or samples with higher
concentrations of HS must be pretreated with Fluran® several minutes
before analysis. The fact that a significant amount of time is
required for a combined chelating-fluorescing reagent to extract U

from omganic molecules provides further evidence that the stability

64

constant between UO§+ and HS is large. A combination of laser
fluorometry and photo-oxidation procedures provides a quick method to

2+
2.

The BC calculated by fitting a hyperbolic function to the

determine several binding characteristics of aquatic HS for U0

observed data is roughly 2.5 times greater than the BC estimated as a
function of flow rate through the ion exchange column. The estimated
BC is comparable to those reported for other divalent and trivalent
cations. Alberts et al. (in press) 1983 note that the range of BC
values for several metals is remarkably small for various waters of
the eastern U.S.A. Giesy et al. (1978) concluded that BC of Cu2+ in
surface waters of Maine was almost entirely controlled by organic
components. Trivalent aluminum (A13+) strongly outcompetes the

divalent ions Cd2+ and Pb2+ for binding sites on organic matter in

surface waters (Alberts and Giesy 1983). Pott et al. (in press)

observed that Aldrich HA solutions (2 mg C/R) bound 8.9 and 9.6 x 10-7
M Al 3+ at pHs of 4 and 5, respectively. In this study, the estimated
BC of 1.4 x 10.6 M UO2+ was determined at an Aldrich HA concentration

2
of 3.5 mg C/2 at pH 4.5. These data suggest that for Aldrich HA

solutions at concentrations of 2 - 3.5 mg C/2 and pHs 4 - 5, the BC

for A13+ and U03+ are similar. This would not be expected if the

comparison is based on ionic charge alone.

Several factors may contribute to the fact that the BC of A13+

and U0§+ by Aldrich HA are so similar. Effective radius of the U atom
(1.38 x 10.10 m) is much larger than the hexavalent uranium ion (8.0 x
10"11 m). Because U is a very large atom, its nucleus holds electrons
in the outer shells less tightly than those of the inner shells.

Electrons in the outer shell can be removed from the U nucleus by more

65

electronegative elements, such as oxygen (02-) in the case of uranyl
+ .-

ion (relative electronegativities: U6 , 1.7; O2 , 3.5; Brownlow 1979).

This explains the ionic character of U02+

2 .
radius of A13+ has been reported as 5.1 x 10-

3+

For comparison, the ionic

11 m (Krauskopf 1967).

Both A1 and U02+ exhibit approximately the same percentage of ionic

2
character (60 and 62%, respectively; Krauskopf 1967). Highly charged
cations deform electron clouds of anions and consequently form mixed
bonds (i.e., covalent and ionic). Because of their tendency to form
covalent bonds as well as ionic bonds in water, ions such as U6+ and
Al3+ distort electron clouds so that the relative strength of the
metal-oxide and metal-hydroxide bond is no longer a simple matter of
charge and radius (Krauskopf 1967). Electron cloud distortion by U6+

+
and Al3 may also be true for other components in surface waters such

as HS.
The conditional stability constant (K') is smaller for UO§+-FA
than for FE3+-FA. That implies that UO§+ probably forms an inner

sphere complex similar to the one between Fe3+-FA (Gamble et al. 1976;
Kerndorff and Schnitzer 1980). In an inner sphere complex, ligand
functional groups may displace strongly coordinated water molecules
and fill the vacated position in an inner sphere complex. By
contrast, in an outer sphere complex, the ion is bound
electrostatically to the ligand without displacement of coordinated
water molecules (Kerndorff and Schnitzer 1980). Mn2+ does not chelate
humic materials, but is bound in fully hydrated form with FA in an
outer sphere complex (Gamble et al. 1976, 1977). At pH 4.7, Kerndorff
and Schnitzer (1980) found the following order of sorption of metals

onto FA:

66
Hg = Fe = Pb = Cu = A1 = Cr > Cd > Ni = Zn > Co > Mn.

+
Disnar (1981) showed that U02+ outcompeted Cu2 for binding sites on a

2

humic molecule, regardless of the order of introduction. Reactions of

HA with U03+ in the presence of other divalent cations have shown that
+
HA has a higher cation exchange capability for U03 than for other

metal cations (Halbach et al. 1980). It has been demonstrated that

the selectivity of HA for several cations diminished in the order

uo§+ > Mn2+ > Ca2+ > H+ (Halbach et al. 1980).

Only a few values have been reported for the conditional

stability constant of the interaction of UO2+ with HS. Kribek and

2

Podlaha (1980) reported conditional stability constants (K') for U03+
-HS complexes estimated by a microgravimetric procedure but did not
postulate on the mode of U03:+ -HS binding. While their results are
similar to those of this study (Table 8), the freezing of extracted HS
by Kribek and Podlaha (1980) may have altered physical and chemical
properties of the organics since HS may aggregate into particulates
after freezing (Lush and Hynes 1973; Giesy and Briese 1978b).

Li et al. (1980) reported on the effect of pH and UO2+

2 :
§+ by HS. They reported two types of

ligand
ratio on the binding of U0
binding sites determined by molecular fluorescence spectrometry with a
difference of about two orders of magnitude between the stability
constants of the weaker and stronger binding sites with the stronger
binding sites representing only about 10% of the total number of
available binding sites. While their study contains useful

information, the experimental conditions do not typify the natural

67

N
~_-g..+~o:.

 

 

 

 

H.M.«.
_qmos.
Amemsflmem mooummu u
seaom mags emammsmu. mm.o .mzooo mu 2 .o.o m.e «m
Amemsamem mooommo u
sesom was“ enmeuuaom. mm.h .ezooo mu 2 .o.o m.e «m
.omeV mamaeom was menace m.o H A.“ soaoaz m m.o o.~ - ..m <:
.ommHV mamaeom new menace m.o H s.~ «caumz m ..o m.o - a.~ <=
Aomm.v maeHeom new museum q.o n w.~ soaumz m ..o m.o - ~.N «m
Aowmfiv .Ha no em om.o mozx m H.o o.o A<ev ewu< unease
.owofiv .Hm no no mq.~ mozx m ..o o.o flax. eeu< ue>Hsm
AomeV .Hm on as ~m.o mozx m ..o o.o .«mv wage ones:
mucouommm «. MOH mHV suwaouum mm camwwg
n oflaom
.mooamumnsm owes:
can meow +NOD mo moonmaoo wsoosvm you A.Mv muamumaoo >uwawnmum Hmaowuwwaoo ownuo>< .w oanmb

N

68
situation in which uranium concentration remains relatively constant

while the aquatic HS concentrations can vary greatly, depending on

2+
2

concentrations investigated by Li et al. (1980) is presented as a

season and amount of rainfall. When the entire range of total U0

Scatchard plot, a two-component plot is obtained similar to the one
presented in Figure 3. This nonlinear response suggests multiple
binding sites.

Halbach et al. (1980) discussed the uptake of U as a function of
total acidity and noted that at pH 4.5, all acid groups were
dissociated into carboxylates or activated phenolates or were at least

easily dissociable. Because HS are weak acid cation exchangers, the

2+
2

influenced by pH. Halbach et al. (1980) believe that both carboxylate

uptake of U0 by HA and FA is an ion exchange process which is

and phenolic acid groups are able in) bind equivalent quantities of

UO§+ ions. This does not agree with the findings of Li et al. (1980).

+
A Scatchard analysis of the entire range of U02 concentrations

2
used in the current study (4.2 x 10.8 M to 2.1 x 10.5 M U03...) also
produced a nonlinear plot but with significantly larger x- and
y-intercepts than Li et al. (1980). The number of total binding sites
(n1 + n2) varied significantly between this study and Li et al.
(1980), indicating that the concentration of humic-complexed U0:+ also
differed. This was probably because some of the smaller MW organics
complexed U02+, but still passed through the dialysis membrane. Li et
al. (1980) maintain that 100% of U6+ (i.e. U0§+) is uncomplexed by FA
(MW from 300 - 2000) and even smaller MW tannic acids at pH 3 and 4,

respectively. These investigators maintain that the only dialyzable

species was the uncomplexed uranyl ion, U035, since the nominal MW

69
exclusion limit for the dialysis membrane was 3500. What they report
2+ . 6+ . .
2 (dialyzable U or Mf) may in fact contain some
organically bound U0:+

smaller than about 3000 nominal MW units.

as uncomplexed UO

(ML) if the MW of the organic fraction is

In surface waters of the southeastern U.S.A. most of the DOC is
composed of the smaller MW fulvic acids (Beck et al. 1974). Nearly
half (48%) of the aquatic HS in a blackwater creek in South Carolina
has a MW of less than 500 (J. J. Alberts, pers. comm.). Halbach et

al. (1980) concluded that the U02+ ion migrates generally as dissolved

2
uranyl fulvates in flowing surface waters and that the migration is
favored by a humid climate with high annual rainfall. The coastal
plain of the southeastern U.S.A. is characterized by similar climatic
conditions.

The possibility that the U02+ inside the dialysis membrane was

2
associated with very small MW organics was not considered by Li et al.
(1980). In actuality, the stability constant that they reported may
be significantly lower than the true stability constant because the
binding potential of the very small HS was not considered. The
assumption that no UO§+ was complexed by the small MW organics also

led to a diminished estimate of the complexed U02+zuncomplexed U02+

2 2
ratio (y-axis). This ultimately produced an underestimate of the
product of the conditional stability constant (K') and the number of
binding sites (11), since n had been underestimated previously. It
should be noted that the present study was performed at a lower pH
(4.5) than that of Li et al. (1980; pH 6.0). The different ionic

strengths probably do not account for significantly different

results, however.

70
Perdue and Lytle (1983) have questioned whether experimental
results from Scatchard analyses of high metal-ligand concentrations
can be extrapolated beyond the limits of the study to very low
metal-ligand concentrations which generally typify natural waters.
They reported that values in the Gaussian model are still rising
sharply at ligand-metal concentrations ([LT]:[MT]) greater than 100

(log [LT]:[MT] Z 2). Because laser fluorometry is a very sensitive

+ .
analytical technique, unrealistically high U0: concentrations were
+ .
not necessary to measure the complexed U02 ion. Thus the uncertainty

2

expressed by Perdue and Lytle (1983) was not a consideration. In
fact, this study extended the theoretical work of Perdue and Lytle
(1983) to an actual metal-ligand system.

A continuous distribution model such as the Gaussian model does
not account for mass balance so care must be taken during experimental
design to obey laws of mass balance. Perdue and Lytle (1983) state
that in natural waters with 10 mg C/I’. of aquatic HS and an MT = 10-8
M, log LT/MT = 3, the Scatchard analysis underestimates K' by several
orders of magnitude. The gradual change of K' characterized by the
Gaussian distribution model cannot be modelled by the Scatchard
equation at high [LT]:[MT] ratios typical of natural surface waters
(Perdue and Lytle 1983). It should be noted that 10 mg C/2 of aquatic
HS is not considered a great concentration of organic material for
surface waters of the southeastern U.S.A. (Beck et al. 1974; Giesy and
Paine 1977; Reuter and Perdue 1977; Giesy and Briese 1978a; Alberts
and Evans 1979). Neither the Scatchard model nor the Gaussian model

has been reported to be very effective at describing the experimental

data at both low [LT]:[MT] values common to most laboratory studies

71
and the high [LT]:[MT] ratios that typify natural surface waters
(Perdue and Lytle 1983). However, the current study shows that when
the conditional stability constant is determined in the [LT]:[MT]
range representative of environmental concentrations, the Scatchard
estimate and the Gaussian estimate of K' are similar (2.4 x 107 and
7.1 x 10 , respectively). These findings are the reverse of those

reported by Perdue and Lytle (1983) who stated that the Scatchard

method underestimated log K' by several orders of magnitude.

CONCLUSION

Natural humic substances (HS) can complex uranium from
concentrations similar to those of surface waters of non-uraniferous
areas. Under simulated conditions of surface waters for the
southeastern U.S.A. (pH 4.5 and I = 0.01 M), a laser fluorometric
procedure was used to determine binding capacity (BC) and average
conditional stability constant (K‘) of Aldrich humic acid (HA) for the
uranyl ion (UO§+).

At pH 4.5 and I = 0.01, Aldrich HA (3.5 mg DOC/2) has a BC of 1.4
x 10'6 g no:+ comparable to the BC of Aldrich HA determined for 1113+

under similar laboratory controlled conditions (Pott et al., in

press). A highly charged cation such as U02+ may be able to deform

2
electron clouds of Aldrich HA and can consequently form mixed bonds
(i.e., covalent and ionic bonds). Aldrich HA may then chelate the
U0:+ ion rather than bind it electrostatically and subsequently
displace strongly coordinated water molecules.

Average conditional stability constants (K') for U0:+ to Aldrich
HA were estimated by a graphical method (Scatchard plot analysis) and
from a frequency distribution (Gaussian distribution). Both the
Scatchard and Gaussian methods resulted in similar estimates for log

K' (7.38 for single component Scatchard plot and 6.85 for the

continuous Gaussian distribution).

72

73

From this study it can be concluded that the Gaussian estimate is
not necessarily superior to the Scatchard estimate of the average
conditional stability constant, K', provided certain conditions are
met. If metal concentrations [MT] used in the simulation studies are
truly representative of environmental concentrations of the metal of
interest, it does not appear that a great difference exists between
the K' estimated by the two methods. However, if the Scatchard plot
is used to calculate K', one should not attach chemical significance
to the four empirical curve-fitting :parameters. If' higher
experimental metal concentrations are necessary than those typical of
natural surface waters, preliminary results suggest that a Gaussian
distribution model is more accurately representative than the
two-component Scatchard equation in its ability to predict accurate
log K' values (Perdue and Lytle 1983).

A plot of the observed log K' as a function of the [LT]:[MT]
ratio indicates a successive filling of" sites (Li et .al. 1980).
Thermodynamically more stable sites are filled first with less stable
sites filled later. An alternative but less probable interpretation
of this observation is that as sites are filled the nature of the
remaining sites changes such that they bind metal less tightly. This
observation has implications for both estimating conditional stability
constants (K') and applying them in geochemical models. Under
conditions of small [MT]:[LT] ratios common for dilute solutions of
U0§+, only the strongest sites would participate in binding.
Estimates of K' made at greater [MT]:[LT] ratios would certainly be

overestimated.

LIST OF REFERENCES

LIST OF REFERENCES

Alberts, J. J. and D. W. Evans. 1979. Comparative chemistry of
submicron sized organic fractions in four rivers of the

southeastern United States. Prog. Phys. Chem. Earth 12:463-468.

Alberts, J. J. and J. P. Giesy. 1983. Conditional stability constants
of trace metals and naturally occurring humic materials: their
application in equilibrium models and verification with field
data. In: Aquatic and Terrestrial Humic Materials. R. F. Christ-
man and E. T. Gjessing, eds. Ann Arbor Science. Ann Arbor, MI.

Alberts, J. J., J. P. Giesy and D. W. Evans. (in press). Distribution
of dissolved organic carbon and metal binding capacity among
unltrafilterable fractions isolated from selected surface waters

of the southeastern United States. Environ. Geol.

Beck, K. C., J. H. Reuter and E. M. Perdue. 1974. Organic and
inorganic geochemistry of some coastal plain rivers of the

southeastern United States: Geochim. Cosmochim. Acta 38:341-364.

van den Berg, C. M. G. and J. R. Kramer. 1979. Conditional stability
constants for copper ions with ligands in natural waters. In:
Chemical Modeling in .Aqueous Systems: Speciation, Sorption,
Solubility and Kinetics. E. A. Jenne, ed. Amer. Chem. Soc. Symp.
Series 93.

Bio-Rad Laboratories. 1972. Separating metals using Chelex-100®
chelating resin. Technical Bulletin 114. Bio-Rad Laboratories,
Richmond, CA.

75

76

Bloch, S. 1980. Some factors controlling the concentration of uranium

in the world ocean. Geochim. Cosmochim. Acta 44:373-377.

Blutstein, H. and R. F. Shaw. 1981. Characterization of’ copper
binding capacity in lake water. Environ. Sci. Technol.
15:1100-1102.

Bodnar, L. Z. 1980. Potentiometric determination of uranium in

organic extracts. Acta. Chem. 52:984-987.

Bonatti, E., D. E. Fisher, 0. Joensuu and H. S. Rydell. 1971.
Postdepositional mobility of some transition elements,

phosphorus, uranium and thorium in deep sea sediments. Geochim.
Cosmochim. Acta 35:189-201.

Borggaard, (L.IK. 1974. Titrimetric determination of acidity and pK
values of humic acid. Acta Chem. Scand. 28:121-122.

Borovec, Z. 1981. The adsoption of uranyl species by fine clay.
Chem. Geol. 32: 45-58.

Bouwer, E. J., J. W. McKlveen and W. J. McDowell. 1978. Uranium assay
of phosphate fertilizers and other phosphatic materials. Health
Phys. 34:345-352.

Breger, I. A., M. Deul and R. Meyrowitz. 1955a. Geochemistry and

mineralogy of a uraniferous subbituminous coal. Econ. Geol.
50:610-624.

Breger, I. A., M. Deul and S. Rubinstein. 1955b. Geochemistry and

mineralogy of a uraniferous lignite. Econ. Geol. 50:206-226.

Brits, R. J. N. and H. A. Das. 1978. The determination of uranium in
groundwater by preconcentrating on activated charcoal and delayed

neutron counting. Radiochem. Radioanal. Lett. 35:313-318.

77

Brookins, D. G. 1978. Retention of transuranic and actinide elements
and bismuth at the Oklo natural reactor, Gabon: Application of
Eh-pH diagrams. Chem. Geol. 23:209-323.

Brownlow, A. H. 1979. Geochemistry. Prentice Hall. Englewood
Cliffs, NJ.

Buffle, J., F. L. Greter and W. Haerdi. 1977. Measurement of
complexation properties of humic and fulvic acids in natural
waters with lead and copper’ ion-selective electrodes. Anal.
Chem. 49:216-222.

Campen, W. and K. Bachmann. 1979. Laser-induced fluorescence for the
direct determination of small concentrations of uranium in water.
Mikrochim. Acta 1979 II: 159-170.

Casagrande, D. J. and L. D. Erchull. 1976. Metals in Okefenokee peat
forming environments: relation to constituents found in coal.
Geochim. Cosmochim. Acta 40:387-393.

Choppin, G. R. and L. Kullberg. 1978. Protonation thermodynamics of
humic acid. J. Inorg. Nucl. Chem. 40:651-654.

Christman, R. F. and M. Ghassemi. 1966. Chemical nature of organic

color in water. J. Amer. Water Works Assoc. 58:723-741.

Christman, R. F. and G. T. Gjessing. 1983. Aquatic and Terrestrial
Humic Materials. Ann Arbor Science. Ann Arbor, MI.

Clark, J. S. and R. C. Turner. 1969. An examination of the resin
exchange method for the determination of stability constants of

metal-soil organic matter complexes. Soil Sci. 107: 8-11.

78

Cothern, C. R., W. L. Lappenbusch and J. A. Cotruvo. 1983. Health
effects guidance for uranium in drinking water. Health Phys.
44:377-384.

Crosser, M. L. and H. E. Allen. 1977. Determination of complexation
capacity of soluble ligands by ion exchange equilibrium. Soil
Sci. 123:176-181.

Disnar, J-R. 1981. Etude expérimentale de la fixation de métaux par

un matériau sédimentaire actuel d'origine algaire - II. Fixation

+
'in vitro' de UO§+, Cu2+, Ni2+, Zn2+, Pb2+, C02+, Mn2 , ainsi que
de V03, Moi- et Geog. Geochim. Cosmochim. Acta 45:363-379.

+
Dongarra, G. and D. Langmuir. 1980. The stability of UOZOH and
UOZIHPO4]§- complexes at 25°C. Geochim. Cosmochim. Acta

44:1747-1751.

Dye, J. L. and V. A. Nicely. 1971. A general purpose curvefitting
program for class and research use. J. Chem. Educ. 48:443-448.

Felbeck, G. T. 1971. Chemical and biological characterization of
humic matter. In: Soil Biochemistry. Vol. 2. A. D. McLoren and
J. Skeyins, eds. Marcel Dekker. NY.

Flaig, W. 1964. Effects of microorganisms in the transformation of
lignin to humic substances . Geochim . Cosmochim . Acta
28: 1523-1535 .

Francis, A. J. 1981. Microbial transformation of low level
radioactive waste. In: International Symposium on Migration in
the Terrestrial Environment of Long-lived Radionuclides from the
Nuclear Fuel Cycle. IAEA-SM-257/72. Vienna.

Francis, A. J., C. R. Iden, B. J. Nine and C. K. Chang. 1980.
Characterization of organics in leachates from low-level

radioactive waste disposal sites. Nucl. Technol. 50:185-163.

79

Gamble, D. S. 1970. Titration curves of fulvic acid: the analytical
chemsitry of a weak acid polyelectrolyte. Can. J. Chem.
48:2662-2669.

Gamble, D. S., C. H. Langford and J. P. K. Tong. 1976. The structure
and equilibria of a manganese (II) complex of fulvic acid studied
by ion exchange and nuclear magnetic resonance. Can. J. Chem.
54:1239-1245.

Gamble, D. S., M. Schnitzer and I. Hoffman. 1970. Cu2+-fulvic acid
chelation equilibrium in 0.1m KCl at 25°C. Can. J. Chem.
48:3197-3204.

Gamble, D. S., M. Schnitzer and D. S. Skinner. 1977. Mn2+-fulvic acid
complexing equilibrium measurements by electron spin resonance

spectrometry. Can. J. Soil Sci. 57:47-53.

Gardiner, J. 1974. The chemistry of cadmium in natural water - I. A

study of cadmium complex formation using the specification

electrode. Water Res. 8:22-30.

Garrels, R. M. and E. S. Larsen. 1959. Geochemistry and mineralogy of
the Colorado Plateau uranium ores. U. S. Geol. Survey Prof.

Paper 320. U.S. Govt. Printing Office. Washington, D.C.

Germanov, A. I. and V. M. Panteleyev. 1968. Behavior of organic

matter in groundwater during infiltrational epigenesis. Intl.
Geol. Rev. 10:826-832.

Giblin, A. M. 1980. The role of clay adsorption in genesis of uranium

ores. In: Proc. Intl. Uranium Symp. Pine Creek Geosyncline.

Giblin, A. M., B. D. Batts and D. J. Swaine. 1981. Laboratory
simulation studies of uranium mobility in natural waters.

Geochim. Cosmochim. Acta 45:699-709.

80

Giesy, J. P., Jr. 1980. Cadmium interactions with naturally occurring
organic ligands. In: Cadmium in the Environment. Jerome O.

Nriagu, ed. John Wiley and Sons, Inc., NY.

Giesy, J. P., Jr. and J. J. Alberts. 1984. Trace metal speciation:
the interaction of metals with organic constituents of surface
waters. In: Effects of Trace Elements on Aquatic Ecosystems.
B. .1. Ward, ed. Electric Power Research Institute. Palo Alto,
CA.

Giesy, J. P., Jr. and L. A. Briese. 1977. Metals associated with

organic carbon extracted from Okefenokee~ Swamp ‘water. Chem.
Geol. 20:109-120.

Giesy, J. P., Jr. and L. A. Briese. 1978a. Trace metal transport by
particulates and organic carbon in two South Carolina streams.

Verh. Intl. Verein. Limnol. 20:1401-1417.

Giesy, J. P. and L. A. Briese. 1978b. Particulate formation due to

freezing humic waters. Water Resour. Res. 14:542-544.

Giesy, J. P., Jr. and L. A. Briese. 1980. Metal binding capacity of
northern European surface waters for Cd, Cu, and Pb. Org.

Geochem. 2:57-67.

Giesy, .I. P., Jr. and D. Paine. 1977. Uptake of americiwm - 241 by
algae and bacteria. Prog. Water Technol. 9:845-857.

Giesy, J. P. Jr., G. J. Leversee and D. R. Williams. 1977. Effects of
naturally occurring aquatic organic fractions on cadmium toxicity
to Simocephalus serrulatus (Daphnidae) and Gambusia affinis

(Poeciliidae). Water Res. 11:1013-1020.

 

Giesy, J. P. Jr., L. A. Briese and G. J. Leversee. 1978. Metal

binding capacity of selected maine surface: waters. Environ.
Geol. 2:257-268.

81

Gindler, LL .E. 1973. Physical and chemical properties of uranium.
In: Uranium. Plutonium. Transplutonic Elements. H. C. Hodge, J.

N. Stannard, and J. B. Hursh, eds. Sprinter - Verlag, NY.

Gjessing, E. TF. 1970. lfltrafiltration of aquatic humics. Environ.
Sci. Technol. 4:437-438.

Gladney, E. S., J. W. Owens and J. W. Starner. 1976. Determination of
uranium in natural waters by neutron activation analysis. Anal.
Chem. 48:973-975.

Gladney, E. S., W. K. Hensley and M. M. Minor. 1978. Comparison of
three techniques for the measurement of depleted uranium in
soils. Anal. Chem. 50:652-653.

Guy, R. D. and C. L. Chakrabarti. 1976. Studies of metal-organic

interactions in model systems pertaining to natural waters. Can.
J. Chem. 54:2600-2611.

Haider, K., J. Trojanowski and V. Sundman. 1978. Screening for lignin
degrading bacteria by means of 14C-labelled lignins. Arch.
Microbiol. 119:103-106.

Haji-Vassiliou A. and P. F. Kerr. 1973. Analytic data on nature of
urano-organic deposits. Bull. Am. .Assoc. Petrol. Geol.
57:1291-1296.

Halbach, P., D. Von Borstel and K.- D. Gundermann. 1980. The uptake
of uranium by organic substances in a peat bog environment on a

granitic bedrock. Chem. Geol. 29:117-138.

Hathaway, L. R. and G. W. James. 1975. Use of chelating ion-exchange
resin in the determination of uranium in groundwater by X-ray
fluorescence. Anal. Chem. 47:2035-2037.

82

Helwig, J. T. and K. A. Council, eds. 1979. SAS User's Guide. SAS
Institute, Inc. Cary, NC.

Horrocks, D. L. 1974. Measurement of low-levels of normal uranium in
water and urine by liquid scintillation alpha counting. Nucl.
Inst. Meth. 117:589-595.

Hunston, D. L. 1975. Two techniques for evaluating small
molecule-macromolecule binding in a complex system. Anal.
Biochem. 63:99-109.

Jablonski, B. B. and D. E. Leyden. 1978. Flow photometric monitor for

uranium in carbonate solutions. Anal. Chem. 50:404-407.

Jackson, K. S., I. R. Jonasson and G. B. Skippen. 1978. The nature of
metals - sediment - water interactions in freshwater bodies, with

emphasis on the role of organic matter. Earth Sci. Rev.
14:97-146.

Jackson, T. A. 1975. Humic matter in natural waters and sediments.
Soil Sci. 119:56-64.

Jackson, T. A. and R. E. Hecky. 1980. Depression of primary
productivity by humic matter in lake and reservoir waters of the

boreal forest zone. Can. J. Fish. Aquat. Sci. 37:2300-2317.

Jennings, J. K. and J. S. Leventhal. 1978. A literature review of
interaction of oxidized uranium species and uranium complexes
with soluble organic matter. U. 8. Geological Survey, Open File
Report 78-179.

Johnson, R., C. M. Wai, B. McVeety, H. Lee and H. Willmes. 1980.
Uraniwm in soil around phosphate processing plants in Pocatello,
Idaho. Bull. Environ. Contam. Toxicol. 24:735-738.

83

Johnston, M. V. and J. C. Wright. 1981. Determination of ultratrace
levels of uranium by selective laser excitation of precipitates.

Anal. Chem. 53:1050-1054.

Kaminski, R., F. J. Purcell and E. Russavage. 1981. Uranyl
phosphorescence at the jparts-per-trillion. level. Anal. Chem.
53:1093-1096.

Kolodny, Y. and I. R. Kaplan. 1973. Deposition of uranium in the
sediment' and interstitial water of an anoxic fjord. In:
Hydrogeochemistry (Proceedings of a Symposium on
Hydrogeochemistry and Biogeochemistry, Tokyo, Japan, Sept. 7-9,
1970), Vol. I. The Clarke Company, Washington, D.C.

Krauskopf, K. 1967. Introduction to Geochemistry. McGraw-Hill. NY.

Kerndorff, H. and M. Schnitzer. 1980. Sorption of metals on humic
acid. Geochim. Cosmochim. Acta 44:1701-1708.

2+ _

Kribek, B. and J. Podlaha. 1980. The stability constant of the U02

humic acid complex. Org. Geochem. 2:93-97.

Kwak, J. C. T., R. W. P. Nelson and D. S. Gamble. 1977.
Ultrafiltration of fulvic and humic acids, a comparison of

stirred cell and hollow fiber techniques. Geochim. Cosmochim.
Acta 41:993-996.

Lamar, W. L. 1968. Evaluation of organic color and iron in natural
surface waters. U.S. Geol. Surv. Prof. Paper 600-D. U. S. Govt.
Printing Office. Washington, DC.

Landrum, P. F. and J. P. Giesy. 1981. Anomalous breakthrough of
benzo[a]pyrene during concentration with Amberlite® XAD -4 resin
from aqueous solutions. In: Advances in the Identification and
Analysis of Organic Pollutants in Water. Vol. 1. Ann Arbor

Science. Ann Arbor, MI.

84

Langford, C. H., D. S. Gamble, A. W. Underdown and S. Lee. 1983.
Interaction of metal ions with a well characterized fulvic acid.
In: Aquatic and Terrestrial Humic Materials. R. F. Christman and

E. T. Gjessing, eds. Ann Arbor Science. Ann Arbor, MI.

Langmuir, D. 1978. Uranium solution - mineral equilibria at low
temperatures with applications to sedimentary ore deposits.
Geochim. Cosmochim. Acta 42:547-569.

Li, W. C., D. M. Victor and C. L. Chakrabarti. 1980. Effect of pH and
uranium concentration on interaction of uranium (VI) and uranium

(IV) with. organic ligands in..aqueous solutions. Anal. Chem.
52:520-523.

Liao, W., R. F. Christman, J. D. Johnson, D. S. Millington and J. R.
Hass. 1982. Structural characterization of aquatic humic
material. Environ. Sci. Technol. 16:403-410.

Lush, D. L. and H. B. N. Hynes. 1973. The formation of particulates

in freshwater leachates of dead leaves. Limnol. Oceanogr.
18:968-977.

MacCarthy, P. 1977. An interpretation of stability constants for soil
organic matter-metal ion complexes under Schubert conditions. J.
Environ. Sci. Health. A12:43-59.

Mangini, A., C. Sonntag, G. Bertsch and E. Mfiller. 1979. Evidence for

a higher natural uranium content in world rivers. Nature
278:337-339.

Mantoura, R. F. C. and J. P. Riley. 1975. The use of gel filtration
in the study of metal binding by humic acids and related
compounds. Anal. Chim. Acta 78:193-200.

85

Mantoura, R. F. C., A. Dickson and J. P. Riley. 1978. The
complexation of metals with humic materials in natural waters.

Estuarine Coastal Mar. Sci. 6:387-408.

Martin, J.-M., V. Nijampurkur and F. Salvadori. 1979. Uranium and
thorium isotope behavior in estuarine systems . In:

Biogeochemistry of Estuarine Sediments. NESCO/SCOR.

McCormick, J. F. and D. J. Cotter. 1964. Radioactivity on

southeastern granite outcrops. Bull. Ga. Acad. Sci. 22:1-9.

Means, J. L., D. A. Crerar, M. P. Borcsik and J. P. Duguid. 1978.
Adsorption of Co and selected actinides by Mn and Fe oxides in

soils and sediments. Geochim. Cosmochim. Acta 42:1763-1773.

Metzger, R., J. W. McKlveen, R. Jenkins and W. J. McDowell. 1980.
Specific activity of uranium and thorium in marketable rock

phosphate as a function of particle size. Health Phys. 39:69-75.

Miller, M. H. and A. J. Ohlrogge. 1958. Water soluble chelating
agents in organic materials: I. characterization of chelating

agents and their reactions with trace metals in soils. Soil Sci.
Soc. Am. Proc. 22:225-228.

Moore, G. W. 1954. Extraction of uranium from aqueous solutions by

coal and some other materials. Econ. Geol. 49:35-53.

Moskvin, A. I., A. M. Shelyakina and P. S. Perminov. 1967.
Solubility product of uranyl phosphate and the composition and
dissociation constants of uranyl phosphate-complexes. Russ. J.
Inorg. Chem. 12:1756-1959.

Nash, K., S. Fried, A. M. Friedman and J. C. Sullivan. 1981. Redox
behavior, complexing and adsorption of hexavalent actinides by

humic acid and selected clays. Environ. Sci. Technol.
15:834-837.

86

Nechay, B. R., J. D. Thompson and J. P. Saunders. 1980. Inhibition by
uranyl nitrate of adenosine triphosphatases derived from animal

and human tissues. Toxicol. Appl. Pharmacol. 53:410-419.

Osmond, J. K. 1964. The distribution of the heavy radioelements in
the rocks and waters of Florida . In: The Natural Radiation
Environment. J. A. S. Adams and W. M. Lowder, eds. The

University of Chicago Press, Chicago, IL.

Pauli, F. W. 1975. Heavy metal humates and their behavior against
hydrogen sulfide. Soil Sci. 119:98-105.

Perdue, E. M. 1978. Solution thermochemistry of humic substances - I.
Acid-base eqnilibria of humic acid. Geochim. Cosmochim. Acta
42:1351-1358.

Perdue, E. M. 1979. Solution thermochemistry of humic substances -II.
Acid-base equilibria of river water humic substances. In:
Chemical Modelling in Aqueous Systems: Speciation, Sorption,
Solubility and Kinetics. E. A. Jenne, ed., Amer. Chem. Soc.
Symp. Series. Washington, D.C.

Perdue, E. M. and C. R. Lytle. 1983. A critical examination of
metal-ligand complexation models: application to defined
multiligand mixtures. In: Aquatic and Terrestrial Humic
Materials. R. F. Christman and G. T. Gjessing, eds. Ann Arbor
Science. Ann Arbor, MI.

Perry, D. L., S. M. Klainer, H. R. Bowman, F. P. Millanovich, T.
Hirschfield and S. Miller. 1981. Detection of ultratrace levels
of uranium in aqueous samples by laser-induced fluorescence

spectrometry. Anal. Chem. 53:1048-1050.

Posner, A. M. 1966. The humic acids extracted by various reagents
from a soil. Part 1. Yield, inorganic components, and titration
curves. J. Soil Sci. 17:65-78.

87

Pott, D. B., J. J. Alberts and A. W. Elzeriman. (in press). The
influence of pH on the binding capacity and conditional stability
constants of aluminum and naturally occurring organic matter.

(submitted to Chem. Geol.).

Randhawa, N. S. and F. E. Broadbent. 1965. Soil organic matter-metal
complexes, 6. Stability constants of Zn - humic acid complexes
at different pH values. Soil Sci. 99:362-365.

Rashid, M. A. and L. H. King. 1970. Major oxygen containing
functional groups present in humic and fulvic acid fractions
isolated from contrasting marine environments. Geochim.

Cosmochim. Acta 34:193-201.

Reuter, J. H. and E. M. Perdue. 1977. Importance of heavy
metal-organic matter interactions in natural waters. Geochim.

Cosmochim. Acta 41:325-334.

Robbins, J. C. 1978. Field technique for the measurement of uranium

in natural waters. Can. Inst. Mining Met. Bull. May 1978, 61-67.

Ruggiero, P., F. S. Interesse and O. Sciacovelli. 1979. [IR] and
[13C] NMR studies on the importance of aromatic structures in

fulvic and humic acids. Geochim. Cosmochim. Acta 43:1771-1775.

Ruggiero, P., F. S. Interesse and (L Sciacovelli. 1980. 3”! NMR
evidence of exchangeable aromatic protons in fulvic and humic

acids. Soil Biol. Biochem. 12:297-299.

Saar, R. A. and J. H. Weber. 1982. Fulvic acid: modifier of metal-ion

chemistry. Environ. Sci. Technol. 16:510A-517A.

Scatchard, G. 1949. The attractions of proteins for small molecules
and ions. Ann. N. Y. Acad. Sci. 51:660-672.

88

Scheinberg, I. H. 1982. Scatchard plots. Science 215:312-313.

Schindler, J. E. and J. J. Alberts. 1974. Analysis of
organic-inorganic associations of four Georgia reservoirs. Arch.
Hydrobiol. 74:429-440.

Schnitzer, M. and E. H. Hansen. 1970. Organo-metallic interactions in
soils: 8. An evaluation of methods for the determination of
stability constants of nmmal-fulvic acid complexers. Soil Sci.
109:333-340.

Schnitzer, M. and S. U. Khan. 1972. Humic Substances in the

Environment. Marcel Dekker. NY.

Schubert, J. 1948. The use of ion exchangers for the determination of
physical-chemical properties of substances, particularly
radiotracers in solution: I. Theoretical. .1. Phys. Colloid
Chem. 52:340-350.

Scintrex Instruction Manual, UA-3 Uranium Analyzer. (N.D.) Scintrex

Limited, Ontario, Canada.

Shelke, D. N. and D. V. Jahagirdar. 1979. Equilibrium study of the
mixed complexes of uranyl ion with pyridine dicarboxylic acids

and aliphatic dicarboxylic acids. J. Inorg. Nucl. Chem.
41:929-931.

Shuman, M. S., B. J. Collins, P. J. Fitzgerald and D. L. Olson. 1983.
Distributions of stability constants and dissociation rate
constants among binding sites on estuarine copper-organic
complexes: rotated disk electrode studies and an affinity
spectrum analysis of ion selective~ electrode and jphotometric
data. In: Aquatic and Terrestrial Humic Materials. R. F.
Christman and G. T. Gjessing, eds. Ann Arbor Science. Ann Arbor,
MI.

89

Sillen, L. G. and A. E. Martell. 1964. Stability constants of
metal-ion complexes. The Chemical Society Special Publication

No. 17. London.

Sillen, L. G. and A. E. Martell. 1971. Stability constants of
metal-ion complexes. The Chemical Society Special Publication
No. 25. London.

Spalding, R. F. and W. M. Sackett. 1972. Uranium in runoff from the
Gulf of Mexico distributive province: anomalous concentrations.

Science 175:629-631.

Sposito, G. 1981. Trace metals in contaminated waters. Environ. Sci.
Technol. 15:396-403.

Sposito, G. and S. V. Mattigod. 1979. Geochem: A computer program for
the calculatitul of chemical equilibria in. soil solutions and
other natural water systems. Kearney Foundation of Soil Science.

University of California. Riverside, CA.

Steelink, C. 1977. Humates and other natural organic substances in

the aquatic environment. J. Chem. Educ. 54:599-603.

Stevenson, F. J. 1977. Nature of divalent transition metal complexes
of humic acids as revealed by a modified potentiometric titration
method. Soil Sci. 123:10-17.

Stevenson, F. J. 1982. Humus Chemistry: Genesis, Composition

Reactions. John Wiley and Sons. New York, NY.

Stevenson, F. J. and M. S. Ardakani. 1972. Organic matter reactions
involving micronutrients in soils. In: Micronutrients in
Agriculture. J. J. Mortvedt, P. M. Giordano and W. L. Lindsay,
eds. Soil Sci. Soc. Am., Madison, WI.

9O

Stuermer, D. H. and J. R. Payne. 1976. Investigations of seawater and
terrestrial humic substances with carbon-13 and proton nuclear

magnetic resonance. Geochim. Cosmochim. Acta 40:1109-1114.

Sunda, W. G. and P. J. Hanson. 1979. Chemical speciation of copper in
river water: effect of total copper, pH, carbonate and dissolved
organic matter. In: Chemical Modeling in Aqueous Systems:
Speciation, Sorption, Solubility and Kinetics. E. A. Jenne, ed.

Amer. Chem. Soc. Symp. Series 93.

Szalay, A. 1954. The enrichment of uranium in some brown coals in

Hungary. Acta Geol. Hung. 2:299-310.

Szalay, A. 1958. The significance of humus in the geochemical
enrichment or uranium in coals and other bioliths. Acta Phys.
Hung. 8:25-26.

Szalay, A. 1964a. The role of humic acids in the geochemistry of
uranium and their possible role in the geochemistry of other
cations. In: Chemistry of the Earth's Crust 2. Nauka, ed. (in

Russian).

Szalay, A. 1964b. Cation exchange properties of humic acids and their
importance in the geochemical enrichment of U02+ and other

2
cations. Geochim. Cosmochim. Acta 28:1605-1614.

Szalay, A. 1969. Accumulation of uranium and other micrometals in
coal and organic shales and the role of humic acids in these

geochemical enrichments. Arkiv. Mineralogi Geologi 5:23-26.

Szalay, A. and Z. Samsoni. 1969. Investigations on the leaching of

uranimm from crushed magmatic rocks. Geochem. Intl. 6:613-623.

Tieh, T. T., E. B. Ledger and M. W. Rowe. 1980. Release of uranium
from granitic rocks during in situ weathering and initial erosion

(central Texas). Chem. Geol. 29:227-248.

91

Tilly, L. J. 1975. Changes in water chemistry and primary
productivity' of' a .reactor' cooling reservoir (Par' Pond). In:
Mineral Cycling in Southeastern Ecosystems. F. G. Howell, J. B.
Gentry and M. H. Smith, eds. ERDA Symp. Series,
CONF-740513-NT15. Springfield, VA.

Tipping, E. 1981. The adsorption of aquatic humic substances by iron
oxides. Geochim. Cosmochim. Acta 45:191-199.

Trojanowski, J., K. Haider and V. Sundman. 1977. Decomposition of
14C-labelled lignin and phenols by a Norcardia sp. Arch.
Microbiol. 114:149-153.

Tsunashima, A., G. W. Brindley and M. Bastovanov. 1981. Adsorption of
uranium from solutions by montmorillonite; compositions and

properties of uranyl montmorillonites. Clays Clay Minerals
29:10-16.

Wahlgren, M. A., D. N. Edgington, F. F. Rowlings and J. L. Rawls.
1972. Trace element determinations on Lake Michigan tributary
water samples using spark sources mass spectrometry. In: Proc.

15§M Conf. on Great Lake Res. Intl. Assoc. for Great Lake Res.

Wallach, D. F. H., D. M. Surgenor, J. Soderberg and E. Delano. 1959.
Preparation and properties of 3, 6-dihydroxy-2, 4-bis- [N,N'-di-
(carboxymethyl)-aminomethyl] fluoran. Anal. Chem. 31:456-460.

Wilson, D. E. and P. Kinney. 1977. Effects. of' polymeric charge
variation on the proton metal ion equilibria of humic materials.

Limnol Oceanogr. 22:281-289.

Wilson, M. A. and K. M. Goh. 1977. Proton decoupled pulse
Fourier-transform 13C magnetic resonance of soil organic matter.
J. Soil Sci. 28:645-652.

92

Zunino, H., P. Peirano, M. Aguilera and I. Escobar. 1972.
Determination of maximum complexing ability of water soluble

complexants. Soil Sci. 114:414-441.