Metal binding by water-soluble polychelates and implications for agriculture

Article abstract of DOI:10.1071/CH11256

Means are developed to improve the metal ion delivery/remediation potential of polyacrylamides (PAMs), by incorporation of the co-monomer N-acryloyl-4-aminosalicylic acid. The polymers were synthesized by solution and inverse emulsion polymerization. The chemical binding of two soil micronutrients, Cu2+ and Fe3+, were investigated using atomic absorption spectroscopy. The modified PAM had an enhanced affinity for metal ions compared with conventional PAMs. This modified PAM has the potential as a delivery tool of plant micronutrients and stabilizers for agricultural soils undergoing intense irrigation. The same polymers may also provide a detoxifying effect in these applications where some micronutrient sources may be in excess and detrimental to productive agriculture. CSIRO 2011.

Full text of DOI:10.1071/CH11256

CSIRO PUBLISHING
Aust. J. Chem. 2011, 64, 1593–1601
Full Paper
http://dx.doi.org/10.1071/CH11256
Metal Binding by Water-Soluble Polychelates
and Implications for Agriculture
A
A B
,
Garry W. Warrender and Robert G. Gilbert
A
The University of Queensland, Centre for Nutrition & Food Sciences, Queensland
Alliance for Agriculture and Food Innovation, Brisbane, Qld 4072, Australia.
Corresponding author. Email: b.gilbert@uq.edu.au
B
Means are developed to improve the metal ion delivery/remediation potential of polyacrylamides (PAMs), by incorpo-
ration of the co-monomer N-acryloyl-4-aminosalicylic acid. The polymers were synthesized by solution and inverse
emulsion polymerization. The chemical binding of two soil micronutrients, Cu2þ and Fe3þ, were investigated using
atomic absorption spectroscopy. The modified PAM had an enhanced affinity for metal ions compared with conventional
PAMs. This modified PAM has the potential as a delivery tool of plant micronutrients and stabilizers for agricultural soils
undergoing intense irrigation. The same polymers may also provide a detoxifying effect in these applications where some
micronutrient sources may be in excess and detrimental to productive agriculture.
Manuscript received: 23 June 2011.
Manuscript accepted: 10 October 2011.
Published online: 16 November 2011.
affect the ability of the ligand to form strong complexes,[17–25] as
indeed shown by the research history of polymeric forms of
derivatized SA in the solid phase (gel form).[18–23,25] A func-
tional monomer that contains SA and that can be copolymerized
with acrylamide would therefore seem suitable as a candidate
for enhancing the chelation of PAMs. Such modified polymers
can be used for continued and improved uses of PAMs in
mining, agriculture, and land rehabilitation. A suitable mono-
mer for modifying the PAM chemical structure is N-acryloyl-4-
aminosalicylic acid (4-acrylamido-2-hydroxybenzoic acid),
shown in Fig. 1. In this work, a PAM with pendant SA units is
synthesized from this monomer. The chelating ability of the
modified PAM is then compared with those of a PAM contain-
ing acrylic acid and of unmodified PAM.
Introduction
Acrylamide copolymers, generically termed polyacrylamides
(PAMs), are widely used for water treatment; the monomers are
acrylamide itself, with either anionic (acrylic acid) or cationic
(e.g., N,N-dimethylaminoethylacrylate methyl chloride salt)
co-monomers. PAMs are increasingly being used in agricultural
settings in the developed world to reduce the watering time for
irrigated crops and to reduce erosive losses of soil during and
after irrigation.[1–12] However, there is a paucity of research on
soil micronutrient delivery with PAM, and also on the effect that
added PAMs might have on the sequestration of important plant-
available metal ions. This work will focus on the ability of PAM-
type polymers to bind metal ions, particularly through a tailored
synthesis with a functional acrylamide monomer chosen to
enhance this chelation.
PAMs are excellent flocculants, coagulants, and dispersants,
but they are also chelators of metal ions. Many hundreds of
different PAMs are made industrially, usually characterized by
charge and charge percent (monomer feed ratio or hydrolysis
number of pure PAM). Historical formulations still dominate the
available PAM types. Improving the chelating ability of PAM
may improve the selectivity of some types of flocculation, but
systematic syntheses have the potential to improve many floc-
culation processes. An example of selective flocculation is the
separation of silica from iron ore, where starch can be used as a
selective chelate and flocculant for iron oxides.[13–16] A syn-
thetic PAM that emulates the selectivity of starch for iron oxide
would require in its composition a pendant sugar unit such as
glucose, or another pendant chelator of iron, some examples of
which are ethylenediamine, salicylicacid(SA), and glycolicacid.
Complexes of SA, and its derivatives, are more stable than
complexes of simple acids and a-hydroxy acids. Derivatization
of SA at the 4 or 5 benzene ring positions does not adversely
Poly(N-acryloyl-4-aminosalicylic acid), P(AMSA), has been
used in gel forms as a polychelate for Fe2þ, Co2þ, Ni2þ, Cr3þ
,
UO22þ, Zn2þ, Cu2þ, Fe3þ, Mn2þ, and Cd2þ [20,23]
Patel and
.
Suthar[23] found, from electronic reflectance and IR spectra
and magnetic susceptibility measurements, that the polychelates
formed with various metal ions were bis-ligated (only two
ligands per metal ion) by the polymer (Fig. 2). Each ligand
formed one to three coordination linkages with the metal ion.
Water and/or other matrix species occupied the remainder of the
coordination sites.
Bonaccorsi et al.[18] synthesized copolymers of acrylamide
and various metal-chelating monomers, including AMSA.
These authors noted an important aspect of the polychelate
behaviour that relates to solution chemistry and binding poten-
tial, as used in this work: crosslinking of the polymer (by metal
complexation) would only occur under suitable conditions of
pH, ionic strength, and reagent concentration. Thus it is possible
to eliminate or control chelated-polymer precipitation by reduc-
ing polymer coiling and counterion condensation through the
Journal compilation Ó CSIRO 2011
www.publish.csiro.au/journals/ajc

1594
G. W. Warrender and R. G. Gilbert
O
O
M2þ to cross the membrane where the polymeric species being
tested cannot. For these dialysis-mediated experiments to be
successful, metal ions that are adsorbed to the polymer must
be removed from the equilibrium, the evidence for which would
be a significantly higher concentration of M2þ in the polymer
solution than in the bulk dialysis medium.
NH2
Acrylamide
OH
Acrylic acid
O
NH
Experimental
Monomer Synthesis
The monomer, AMSA, was synthesized by adding excess
acryloyl chloride (Aldrich, 0.368 mol) to sodium 4-aminosali-
cylate (Aldrich, 0.24 mol) in 500 mL of cold distilled water.[19]
The method of Kennedy et al.[19] was followed with minor
variations, those being: the reaction mixture was stirred for 2 h at
58C and also at room temperature, making two additions of
acryloyl chloride, and precipitation of AMSA by drop-wise and
careful addition of 10 M HCl.
OH
OH
N-acryloyl-4-aminosalicylic acid
O
Fig. 1. The co-monomers used in this work, clockwise from left: acrylam-
ide, acrylic acid and N-acryloyl-4-aminosalicylic acid. The polymers made
here for solution chelation testing were 95 % acrylamide or 100 % acrylam-
ide, the remaining 5 % being one of the co-monomers.
The melting point of the product was 2238C and the decom-
1
position temperature was 2258C. The H NMR spectrum of
AMSA (Fig. 3) revealed an acid group at 10.4 ppm, three
benzene protons between 7 and 7.8 ppm, three olefinic protons
between 5.7 and 6.6 ppm, and a broad (attributable to water and
the b-hydroxy proton) peak at 4 ppm for the amido proton.
(nꢀꢁ 4)
O
NHR
Polymer Syntheses
O
The nomenclature of polymers used here is based on the
following: a 95 mol-% AM, 5 mol-% AA co-polymer will be
referred to as PAMAA; a 95 mol-% AM, 5 mol-% AMSA
co-polymer will be referred to as P(chelate); and homo-
polyacrylamide will be referred to as PAM. Copolymers of
AMSA, AA, and AM were synthesized by solution and inverse
emulsion polymerizations, as set out in Table 1. The study also
uses one of the Aqpol series of solution polymers with various
monomer ratios.
H2O
O
Fe
O
OH2
O
RHN
O
Precipitation reactions were solution polymerizations con-
ducted in distilled and dried DMF at 20 % w/w of monomers.
The initiator of choice for DMF solution reactions was 2,20-
azoisobutyronitrile (AIBN, Aldrich, recrystallized from etha-
nol). The reactions were typically left to go to completion over
20 h. Polymers were filtered from the DMF, washed with
methanol, ethanol, and diethyl ether, and vacuum dried to ensure
removal of DMF, unreacted monomer, and solvent.
Aqueous polymerizations were conducted at 10 wt-% of total
monomers. The solutions were typically pale tan in colour after
pH adjustment. Initiation in this system was by redox and
thermal polymerizations using an azo initiator, V-50, 2,20-
azobis(2-methylpropionamidine) dihydrochloride. The reac-
tions were conducted on a small scale with magnetic stirring
and oxygen was purged from the system by passing ultrapure
nitrogen through the reaction solution. Initiator, 3 mL of
0.02 wt-% V-50, was doped into the reaction over 3–5 min. In
most instances, up to 30 mol-% AMSA, polymerization began
before the initiator had been completely added. Above 30 mol-%
of AMSA in a reaction, polymerization was retarded and not
visually apparent. 1H NMR spectroscopy was used to estimate
the conversion of these reactions.
Fig. 2. Structure of a possible metal-bis chelate of N-acryloyl-4-
aminosalicylic acid (AMSA) and polyAMSA. The overall charge of the
complex is dependent on the charge, nþ, of the metal species.
limitation of the number of chelating monomer units in the
polymer chains. In the present work, the chelating monomer
content did not exceed 5 % by weight of the polymer chemical
composition.
Effect of pH on Metal Polymer Chelate Formation
]
The adsorption of Cu2þ by poly(acrylic acid) (PAA) and deri-
vatives is pH dependent.[26,27] Ionization of carboxylate groups
enhances the binding process and more copper is adsorbed up to
a pH where copper will no longer be adsorbed by an aquated
polymer.[27] Sudden declines in adsorption with increasing pH
can only be attributed to the solubility of the metal species[18,27]
as a result of excess hydroxo ligands available in solution; these
excess OHꢀ ions dominate the chelation equilibrium, and sub-
sequently the adsorption of metal ions by the polymer is
reduced.
Equilibrium dialysis experiments are used here to measure
the metal-adsorption effectiveness of the various polymers
under investigation. The principle of this method is that two
solutions separated by a permeable membrane, differing in ionic
content, tend to equalize the concentrations of all ions across the
membrane. The experimental method relies on the ability of
Inverse emulsion polymers were synthesized by typical
methods[28,29] at a solids load of 20 % by weight. The continuous
phase was Shellsol D-60 (Shell), a light mineral oil, with a
2–2.5 % by weight of reaction of a surfactant blend. Surfactants
were Eco-Teric T61 (a polyoxyethylene-sorbitan monooleate
made by Huntsman Laboratories) and Crill-4 (a sorbitan

Metal Binding by Water-Soluble Polychelates
1595
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
PPM
Fig. 3. 1H NMR spectrum of the N-acryloyl-4-aminosalicylic acid (AMSA) monomer, with trace ethanol, measured in dimethyl sulfoxide (DMSO). Benzylic
protons at 7–8 ppm, olefinic protons 5.5–6.5 ppm, acid group proton at 10.5 ppm, NH and OH protons are seen as the broadened peak at 4 ppm. DMSO appears
at 2.5 ppm.
Table 1. Conditions and ingredients in polymer synthesesA
Type
Solvent
DMF
Initiator
AIBN
Temp.
Time
20 h
Recovery/Purification
Equipment
Precipitation
608C
Filtration/washing with
MeOH, EtOH and
Sealed round bottom flask in oil
bath on magnetic stirrer/heater
EtOEt
Aqueous
Water, pH 8
3 mL of 0.02 wt-% 608C
V-50 in 3–5 min
5–30 min
12–24 h
Used as-is or diluted and
dialyzed in SpectraPore
4, MWCO ¼ 12 kDa
Polymer precipitated in
MeOH or EtOH and
washed with acetone
Semi-sealed jars, with stirrer,
needle bleed-valve and
injection needle
Inverse
Shellsol D60 with
Ecoteric T61
and SMO
AIBN
40–608C
Moore Systems, temperature-
controlled reactor, with
orthogonal multiplate impeller
emulsion
Water, monomers, Accelerated by
EDTA, NaOH
1–3 mL of
0.1 wt-% SMBS
ADMF ¼ dimethylformamide; AIBN ¼ azoisobutyronitrile; MeOH ¼ methanol; EtOH ¼ ethanol; EtOEt ¼ diethylether; V-50 ¼ 4,4-azobis(4-cyanovaleric
acid); Shellsol D60 ¼ mineral oil distillate; SMO ¼ sorbitanmonooleate; SMBS ¼ sodium metabisulfite; EDTA ¼ ethylenediaminetetraacetate.
monooleate made by Croda Surfactants), used roughly in
proportion with each other. The disperse phase contained the
monomers (100 g in total), ethylenediaminetetraacetate (EDTA)
for metal chelation (0.2 g), NaOH for pH modification (pH 8),
and MilliQ water (equivalent to doubly distilled). Disperse
phases were stirred well before being set to pH 8, and then
filtered, emulsified with the continuous phase (Ultra-turrax
emulsifier 4000–8000 rpm, 1–3 min), and polymerized with
continuous nitrogen sparging for the first 12 h. Initiator (AIBN,
Waco, recrystallized from ethanol), pre-dissolved in DMF/oil
(3:1), was added after thermal equilibration of the reaction
mixture and after a minimum of 15 min N2 sparging. Thermal
initiators can be added as a single reagent and at a single stage of
the reaction, whereas a redox couple cannot (there must be two
components, one of which is usually fed gradually through much
of the polymerization process). Polymerization can be acceler-
ated by the addition of 0.1 % w/w sodium metabisulfite (SMBS)
solution, which both reductively and thermally initiates
polymerization.
Adsorption
The dialysis experiments were performed at 308C in screw-cap
glass bottles, which contained teflon-coated magnetic stirrer
bars. A dialysis membrane was prewetted for 0.5–3 h before use.
The membrane type was Spectrapore-4, which has a molecular
weight cut-off (MWCO), or exclusion limit, of 12 kDa. Mem-
brane tubes were made by first tying one end of a membrane
sleeve, the sleeve was then filled with polymer solution and the
opposite end tied.
Solutions of metal species were made by dilution of AAS
standard solutions (1000 ppm nitrate salts in 1 M HNO3, from
Reagecon) into 100 ꢁ 10ꢀ3 M KNO3/20 ꢁ 10ꢀ3 M N-[2-
hydroxyethyl]piperazine-N0-[2-ethanesulfonic acid] (HEPES)

1596
G. W. Warrender and R. G. Gilbert
solution. Metal solutions (,20 ppm) were pH adjusted by
addition of small volumes of 1 M KOH or 1 M HNO3, and then
stabilized by placing the screw-cap bottles into an oven at 408C
overnight (it was found that gentle heating helped reduce metal
losses from the bulk solution during the dialysis experiment,
which indicated that solutions were stabilized that way, or that
small, metal-rich colloidal particles were prevented from form-
ing). Dialysis bottles were transferred from the oven to a multi-
point magnetic stirrer equipped with a water bath. When the
bottles had cooled to 308C, the dialysis bags were placed into
their corresponding bottles and the bottles were tightly capped.
Recovered solutions were assayed by flame atomic absorption
spectrophotometry (FAAS, Varian SpectrAA 220 FS spectrom-
eter) using standards made by dilution of the same 1000 ppm
stock standard described above, and ion-selective light sources.
Metal concentrations in solution were measured and reported in
parts per million. The analysis utilized an integration measure-
ment procedure, where three measurements were taken with a
read delay of 5 s. A reslope rate of 15 or less was used for all
experiments and the calibration algorithm was ‘rational’. Linear
calibration was not used during analysis, but as part of the
reslope procedure. Standards were assayed systematically
through any set of samples to enable calibration checking. Cal
Zero was taken as the background, for which the dialysis
medium, free of metal ions, was used. Samples that assayed
above calibration were immediately diluted with the dialysis
medium as diluent and re-assayed; recalibration was performed
before re-analysis. Where recalibration was performed, average
calibration values were used for assigning assay values.
Table 2. Examples of dn/dc results for determination from light
scattering data
Polymer ID
Charge MonomersA
[mol-%]
dn/dc [mL gꢀ1
]
AN956
ꢀ50
ꢀ3
0
AA, AM
AA, AM
AM
0.247
0.300
0.292
0.253
0.280
0.153
0.240
AN905
FA920
FO4107
0.75 DMAEA ꢄ MeCl, AM
FO4240
15
DMAEA ꢄ MeCl, AM
DMAEA ꢄ MeCl, AM
AM
FO4650
60
0
PAM (precipitated
from DMF)
P(chelate)
Aqpol-12
ꢀ5
ꢀ5
AMSA, AM
AMSA, AM
0.386
0.4536
AAA ¼ acrylic
acid;
AM ¼ acrylamide;
DMAEA:MeCl ¼ N,N-
dimethylaminoethylacrylate methyl chloride; AMSA ¼ N-acryloyl-4-
aminosalicylic acid.
measurements were made. Data for dn/dc are presented in
Table 2.
The MALLS instrument was calibrated using bovine serum
albumin (BSA), which has a radius of gyration (Rg) of 2 nm.
Normalization coefficients were found for each angle using
BSA as an isotropic scatterer. Typical high polymer concentra-
tions for dn/dc are linear up to 0.1 mg mLꢀ1. The actual polymer
concentration used is only important in the experimental stages
of measuring dn/dc.
The dn/dc values for the same polymers made by different
techniques are not the same. This can be ascribed to the
sensitivity of this quantity to structural features such as branch-
ing fraction (e.g. see Cave et al.[30]).
The molecular weights observed were all less than 105, so
there will be no shear-scission artefacts, which can occur for
much larger molecules in SEC.[30]
The metal ion concentration (ppm) of Mnþ in various solu-
tions will be written as follows: in polymer dialysis solutions,
Cp; in blank dialysis solutions, Cw; and in bulk solutions as Bx,
where x can be an identifying number 1–4.
In an ideal case Bx ¼ Cw; however, if the dialysis membrane
adsorbs Mnþ, Bx can be less than Cw. To establish the amount of
metal adsorbed to polymer, Eqn 1 was used:
nþ
½M
ꢂ
¼ Cp ꢀ Cw
ð1Þ
bound
Results and Discussion
Polymer Chemistry versus Adsorption
Molecular Weight
In this section, comparison is made between two low-charge
polyelectrolytes and PAM. The polyelectrolytes (P(chelate),
PAMAA) were made to have a low charge such that conden-
sation of cations onto the polymer chains was less likely to
occur[31] and only the direct binding of metal cations from
dialysis uptake experiments occurred in solution and was sub-
sequently measured.
Time-dependent uptake experiments of Cu2þ by P(chelate)
and PAM solutions were conducted to establish what the dialysis
equilibration time would be. It was thought that over an
extended period, acid-washed glassware kept at a relatively
neutral pH might adsorb metal ions from solution. The time
experiments (Fig. 4) enabled the dialysis process to be tested for
repeatability, metal-ion loss from bulk solution, and for the
ability of non-ionic PAM to adsorb Cu2þ. Trial experiments
revealed that the stock copper solution concentration slowly
decreased when kept at a pH of 5. All metal solutions were made
up within 2 days of an experiment, and then heat treated in
bottles at 408C for 12 h.
The weight-average molecular weight (MW) of the synthesized
polyacrylamides was found using aqueous size-exclusion
chromatography (SEC) coupled with differential refractive
index detection (690 nm, Optilab DSP interferometric refrac-
tometer), UV/vis detection, and multi-angle laser light scatter-
ing (MALLS) detection (188 angle, 683 nm, Wyatt DAWN
EOS). Data were treated using ASTRA software.
Operating parameters were as follows: Column, Ultrahydro-
Gel Linear or PSS Suprema 30000; Mobile phase, 0.5 M NaCl;
flow rate, 0.5 mL minꢀ1; column temperature, 308C; injection
volume, 25–100 mL; needle rinsing volume, 450 mL; auto sam-
pler storage temperature, 208C, Optilab temperature, 308C. Plate
count for the column was N ¼ 8310 ꢃ 548, measured using
ethylene glycol elution volume (peak maximum ¼ Vp) and
half peak width (W1/2) in the equation N ¼ 5.54(Vp/W1/2)2.
Each polymer has a unique dn/dc (n ¼ refractive index,
c ¼ concentration), which was measured using a DAWN DRI
spectrometer by injecting various concentrations of a polymer
and measuring the refractive index. The change in refractive
index with respect to concentration was linear. Polymer was
dissolved in matrix solution and dialyzed against matrix solution
to remove excess counter ions and monomers before dn/dc
PAM (800 ppm) was dialyzed against Cu2þ solution at pH
6.5, as set out in the methodology, and Cp was found to be
marginally less than Cw. Metal uptake by this method, Cp – Cw,
was apparently negative (e.g., at 24 h Cp – Cw ¼ 9.3 – 10.2 ppm).

Metal Binding by Water-Soluble Polychelates
1597
40
35
30
25
20
15
10
5
Thus the Cu2þ concentration change for bottles with dialysis
tubes, DBmax, should be of the order of 1.1 ppm. The DBt of
2.3 ppm for bottles that contained dialysis tubes (DBt . DBmax
)
indicated there was an unanticipated loss of Cu2þ and this can be
assigned to one of three causes: (1) elevated adsorption to glass
bottles, (2) adsorption to cellulose membrane, or (3) precipita-
tion inside the dialysis tubes. The likely answer is membrane
adsorption, because adsorption of metal species onto glass was
assumed relatively constant for such similar bottles. Precipita-
tion of metal ions inside blank dialysis tubing was unlikely,
because no precipitating agent was present in the dialysis tubing
and any CO2 in solution is likely to improve the solubility of,
rather than precipitate, Cu .
2þ [32] Thus membrane adsorption of
Cu2þ was the main possibility and should be of the order of
DBt ꢀ DBmax ¼ 1.2 ppm. As this result comes from only two
experiments there cannot be a great deal of confidence in it;
however, the value of 1.2 ppm for membrane adsorption is a
reasonable first estimate, because reproducibility of the experi-
ments was found in other areas (Cw for the two experiments was
the same, 10.2 ppm). This could be checked through chelating
Cu2þ in the membrane with EDTA, but our final inferences
would be unaffected by the increased precision that this would
afford. In dialysis it was found that B1 ¼ B3 (a 4 % difference
where they should be the same), meaning human and machine
errors cannot be ruled out as contributors to the variation.
0
0
4
8
12
16
20
24
Time [h]
P (chelate)-800 ppm
P (chelate)-200 ppm
P (chelate)-400 ppm
PAM-800 ppm
Fig. 4. Total concentration of Cu2þ ([Cu2þ]T) in dialysis uptake experi-
ments with relation to polymer content and time. Bulk Cu2þ solutions were
made up at 17 ppm with 20 mM HEPES and 100 mM KNO3 at pH 5.5 and
dialyzed at 308C. Bulk [Cu2þ] drifted over time, 12.5–11.2 ppm, and Cw
registered 10.2 ppm Cu2þ after 24 h.
The result was unexpected and showed that great care was
essential in handling and preparing these experiments. Blank
dialysis runs were sampled before and after each experiment and
duplicate Cw values were obtained. Cw was very reproducible
in the experiments giving 0 ppm Cu2þ at time ¼ 0, and 10.2
after 24 h.
To estimate the amount of metal ions ‘lost’ in the process of
24 h dialysis, bulk copper solution concentrations (Bx) were
measured at the beginning and end of two uptake experiments,
one in the presence of and the other free from dialysis tubing;
this was completed for the 800 ppm PAM experiment and the
800 ppm P(chelate). The changes in Cu2þ were 12.8 down to
10.5 ppm, for a bottle with dialysis tubing inside, and 12.3 to
11.9 ppm, for a bottle with no dialysis tubing. This showed a
small amount of unexplained bulk-copper loss in the experiment
containing dialysis tubing. The bottle without tubing gave an
indication of the amount of copper that adsorbed to glass over
the duration of a typical experiment (24 h). We define:
Cu21 Uptake by P(chelate)
The rate of binding of Cu2þ was elucidated from the dialysis
experiments. Pure PAM solution (800 ppm) did not appear to
bind copper and only reached the equilibrium concentration of
Cu2þ in dialysis tubes, giving the diffusion-limited uptake rate.
The dialysis of P(chelate) was also performed at 800 ppm and we
can subsequently assume a similar diffusion-limited uptake rate,
because we can assume a similar volume of polymer coils
(PAM ¼ 55000 and P(chelate) ¼ 77000) inside the membranes.
Thus, the time-dependent Cu2þ concentration (Cp) in the
P(chelate) 800 ppm uptake experiment is the diffusion-limited
amount plus the amount attributable to the binding equilibrium,
given by:
Â
Ã
Â
Ã
Cp ¼ Cu2þ
þ Cu2þ
ð5Þ
diff
binding
An assumption that the binding constant for chelation is large
leads to the approximation that ‘bound copper concentration’ ¼
Cp ꢀ Cw. However, the difference between the dialysis uptake
rates for the two polymer types is a direct result of the binding
constant of the AMSA component of P(chelate). The relation-
ship between dialysis rates of non-binding PAM (800 ppm) and
binding P(chelate) (800 ppm) is shown in Fig. 5. Data from the
dialysis experiments were used to generate the numerical
derivative of solution concentrations of copper (Cp) with time
and then used to generate Fig. 5, plotted as the natural logarithm.
The results in Fig. 5 are now used to show how the binding
potential of the polymer contributes to metal ion transport across
the membrane and make some inferences about the type of
chelate being formed. The approximate linearity of the plot of
ln(dCp/dt) against lnt in Fig. 5, i.e., ln(dCp/dt) ¼ m lnt þ lnD (for
some constants m and D), implies that one has approximately:
DBn ¼ B1 ꢀ B2
DBt ¼ B3 ꢀ B4
ð2Þ
ð3Þ
where:
B1 ¼ initial concentration of Cu2þ in bulk with no tube
(12.3 ppm);
B2 ¼ final concentration of Cu2þ in bulk with no tube
(11.9 ppm);
B3 ¼ initial concentration of Cu2þ in bulk with tube (12.8 ppm);
B4 ¼ final concentration of Cu2þ in bulk with tube (10.5 ppm);
If DBe is the expected bulk-copper change by dilution of
the bulk solution into the dialysis tube, then the maximum
bulk-copper change, DBmax, in a perfect system where there is
no metal adsorption to dialysis tubing, will be:
dCp
¼ D ꢄ tm
ð6Þ
DBmax ¼ DBt ¼ DBe þ DBn
ð4Þ
dt

1598
G. W. Warrender and R. G. Gilbert
For PAM, m ¼ ꢀ0.98 ꢃ 0.03 and D ¼ 1.03 ppm sꢀ1, or
pH is appropriate. The pH dependence of Cu2þ uptake by
P(chelate) was investigated further so that a better understand-
ing of the adsorptive power of AA versus AMSA could be made.
Molar binding ratios (Fig. 6) between Cu2þ and polymeric
AMSA were investigated in relation to pH.[20,33] Dialysis tubes
were filled with the same volume of P(chelate) solution and then
placed into different dialysis bottles that were pH adjusted to be
in the range 2.0–10.0. After 10 h the contents of the dialysis
tubes were analyzed as before and the molar binding ratio of
Cu2þ to AMSA calculated. The results are shown in Fig. 6.
The relationship between pH and binding ratio (Fig. 6)
indicates a significant connection between Cu2þ, its solubility,
and ionizable ꢀCOOH functional groups within AMSA moie-
ties. The pH dependence suggests that Cu2þ may chelate to
AMSA in greater quantities through the single carboxy func-
tional group. It is noted that the solubility product Ksp for Cu
(OH)2 is in the region of 10ꢀ17, which is very low, showing that
the ꢀOH ligand is strongly bound to Cu2þ at appropriate pH in
solution. However, the molar binding ratio went as high as two
Cu2þ atoms per AMSA ligand, which can mean:
dCp/dt is approximately inversely proportional to t. This is the
expected ‘constant rate of uptake’ seen in non-adsorbing
equilibrium dialysis systems. For P(chelate), one has
m ¼ ꢀ1.17 ꢃ 0.07 and D ¼ 26.6 ppm sꢀ1
.
Fig. 5 shows that equilibrium was reached at ,8 to 12 h.
Equilibrium was assumed at 10 h for the data in Fig. 5, the bound
copper concentration was then calculated as per Eqn 5. The same
experiment was then performed with PAMAA and a 10 h
equilibrium also assumed for that experiment. The purpose of
conducting this experiment was to see if relatively dilute,
statistically distributed AA units would bind in a 2:1 ratio with
Cu2þ, as previous research[27] had shown. In comparing Cu2þ
adsorption by PAMAA, PAM, and P(chelate), PAM adsorbed
no measurable Cu2þ in dialysis at 800 ppm; however, P(chelate)
adsorbed 24.7 ppm Cu2þ and PAMAA adsorbed 38.0–39.4 ppm
Cu2þ after equilibrium had been reached. Table 3 compares the
AMSA and AA results.
As the experiments for PAM and P(chelate) were completed
at pH 5–5.5, there may be a biased degree of binding in favour of
the uptake of Cu2þ by AA. PAA was shown to take up larger
amounts of Cu2þ with increasing pH.[26,27] Rivas et al.[26] and
Subramanian and Natarajan[27] showed that PAA could bind
Cu2þ through two carboxylic acid groups and could be tetra-
dentate to copper ions. Thus it is seen that PAMAA will adsorb
more Cu2þ than P(chelate) (and more than expected) where the
ꢅ Cu2þ can bind to AMSA in more than one way, i.e., through
carboxy groups, hydroxy groups, or amino groups; or
ꢅ Cu2þ is being co-adsorbed by the remainder of the polymer
chain; or
2
1.5
1
ꢁ3
ꢁ4
ꢁ5
ꢁ6
ꢁ7
ꢁ8
ꢁ9
ꢁ10
ꢁ11
ꢁ12
ꢁ13
0.5
0
5
6
7
8
9
10
11
12
2
4
6
8
10
In (time [sꢁ1])
pH
Fig. 5. Relationship between the dialysis rates of poly(acrylamide) (PAM)
and P(chelate) at 800 ppm. Dialysis in non-binding PAM shows that a linear
central region of the plot, where actual rate of uptake is almost constant,
dominates diffusion of ions across the membrane. Sample duplication error
was less than 1 %.
Fig. 6. Copper molar binding ratio to P(chelate) with respect to pH.
Binding ratio measured against N-acryloyl-4-aminosalicylic acid (AMSA)
monomer content at 308C in 20 mM HEPES for buffering and 100 mM
KNO3 for counterion population.
Table 3. Binding of Cu21 after 10 h of dialysis
Comparison of polymeric forms of N-acryloyl-4-aminosalicylic acid (AMSA) and acrylic acid (AA); pure poly(acrylamide) (PAM) did not adsorb any
Cu2þ because a dialysis bag filled with PAM solution contained the same [Cu2þ] as bulk water after 10 h. Uncertainties are from two or three measurements
for each sample.
Monomer
Mass (ꢁ10ꢀ4) [g]
Moles (ꢁ106)
Bound Cu2þ [mg]
Bound ratio (moles monomer/moles Cu)
AA
2.03 ꢃ 0.02
5.5 ꢃ 0.1
2.82 ꢃ 0.03
2.66 ꢃ 0.05
0.193 ꢃ 0.002
0.128 ꢃ 0.002
0.92 ꢃ 0.08
1.32 ꢃ 0.08
AMSA

Metal Binding by Water-Soluble Polychelates
1599
ꢅ Cu2þ is being co-adsorbed through bridging oxygen mole-
One means of testing the influence of the OHꢀ ligand in these
experiments is to perform the same experiments with salts of
iron. Iron will readily form complexes with OHꢀ ligands as the
pH rises, which will lead to the formation of insoluble com-
plexes above pH 4. Therefore, adsorption experiments with
Fe3þ were conducted for this work in the same way as that
described above for Cu2þ with the chelating polymers. The
difference to the methodology for Fe3þ was that of the pH at the
start of the experiment. Fe3þ is not soluble above pH 4 and in this
work dialysis (adsorption) of Fe3þ was conducted at or below
pH 3.5. The levels of bound Fe3þ were so high in the experi-
ments that it was decided to measure uptake of the metal with
respect to chelating monomer. Fe3þ adsorption data revealed
that Fe3þ bound to the polychelate in molar ratios far in excess of
that of Cu2þ (Fig. 8). Indeed Fig. 8 shows that not only can Fe3þ
cules, or aquo linkages between pH 5 to 7.
The option of co-adsorption by the polymer chain can be
disregarded, because PAM was shown earlier to be non-binding
towards Cu2þ. In regard to the possibility of more than one
active binding site on the chelating monomer, AMSA has been
shown to be bidentate (the same as for SA) to Cu2þ in aqueous
solutions.[18] Other work has also reported that carboxylates
(alone) can be bidentate,[27] suggesting the possibility for
AMSA to be tridentate purely through oxygen atoms (Fig. 1).
If SA and derivatives (such as AMSA) can be tridentate to metal
ions, then there is the possibility for them to be able to bind
significantly more than one metal ion in the carboxylate and
hydroxy region.
The last option is the most reasonable, which intimates that
hydroxy ions bind to Cu2þ in solution and interfere with AMSA
binding. Insoluble hydroxide formation results from increasing
pH in Cu2þ solutions.[27] Therefore, P(chelate) binding is
expected to decrease rapidly above pH 7 because of insoluble
hydroxides of Cu2þ being formed in the bulk solution. High
degrees of binding prevailed where mono-hydroxylated Cu2þ
was prevalent in solution at around pH 6.5. Copper monohydr-
oxide had, therefore, probably formed before dialysis with
P(chelate) began; therefore, AMSA units competed for avail-
able binding sites on copper with hydroxo ligands, and rather
than displace hydroxy ions, were able to bind two copper atoms
per salicylate site. Fig. 7 and Eqn 7 explain the chemistry of
AMSA complexation of Cu2þ at different pH values. In Eqn 7 it
is assumed that copper ions comprise a 50:50 mixture of the
4-coordinate and 6-coordinate species. Between pH 5.0 and 5.5,
adsorption was found to be in the ratio 1:1; at pH 6.5, 2:1
(Cu/AMSA); and at pH 7.5, the ratio dropped below unity.
bind in a higher ratio to the polychelate but that, unlike Cu2þ
,
Fe3þ binding may continue to increase with increasing poly-
chelate concentration.
The AMSA concentration is always low in these dialysis
experiments, and thus the main determinant for selectivity for
binding to P(chelate) is pH; the competition between AMSA and
OH as ligands heavily favours metal hydroxide formation at low
AMSA concentration. Thus, if pH were to be deliberately
7
6
5
4
3
2
1
0
Â
Ã
Â
Ãþ
Â
Ã
ð7Þ
2þ
CuðH2OÞ5
pH ¼ 5
Ð CuðOHÞðH2OÞ4
Ð
CuðOHÞ2ðH2OÞ3
pH ¼ 7:5
pH ¼ 6:5
0.0
0.1
0.2
0.3
0.4
[AMSA] [mM]
Despite the fact that at pH 7.5, Cu2þ solubility is quite low,
AMSA was able to bind Cu2þ from solution. Hydroxy groups
are strong ligands for Cu2þ, but a competitive study was not
undertaken to ascertain the relative strengths of the two ligands
OHꢀ and AMSA.
Fig. 8. Binding ratio (Fe3þ/N-acryloyl-4-aminosalicylic acid (AMSA))
versus available binding sites (moles AMSA) by dialysis of polychelate
solutions. Dialysis of polychelate was at pH 3.5 in 100 mM KNO3 taken to
equilibrium.
Cu
O
CuX
O
O
Cu
X
O
O
O
Cu
RHN
RHN
RHN
O
CuX
O
O
O
Cu
X
O
Cu
O
CuX
O
O
O
RHN
Cu
CuX
RHN
RHN
O
O
O
pH 5
pH 6.5
Fig. 7. Most probable mode of complexation of polymeric N-acryloyl-4-aminosalicylic acid (AMSA) to Cu2þ with respect to pH. X depicts probable OHꢀ
groups.

1600
G. W. Warrender and R. G. Gilbert
Orthorhombic
Tetragonal
by use of pH modification and polymeric chelating agents as
might induce flocculation of the excess metal ion of interest.
O
O
O
O
O
O
O
O
O
Conclusions
Cu2ꢀ
Cu2ꢀ
Cu2ꢀ
O
Cu2ꢀ
Through selection of a chelator species, a water-soluble polymer
based on AA with a SA-type co-monomer, AMSA, was syn-
thesized that is able to bind metal ions from solution. The dif-
ference between the current polymers and previous work
(polymer gels) is that these are also able to function as a potential
coagulant and/or flocculant. PAM was not seen to be a chelator
of Cu2þ, but polyelectrolytes containing AA and AMSA were
found to be good binders of metal ions, particularly Fe3þ and
Cu2þ, reported herein.
O
O
O
O
O
O
D2h
D4h
Fig. 9. The equilibrium between two forms of Cu2þꢀpoly(acrylic acid)
(PAA) chelate.
Copper uptake by polychelates with carboxylate centres is
highly sensitive to pH, the optimum pH for maximum Cu2þ
adsorption being 6.5. At this pH, AMSA is up to twice as
efficient than AA at adsorbing Cu2þ. As a result of OHꢀ ions
in solution, metal-ion solubility is also pH sensitive. Subrama-
nian and Natarajan[27] illustrated two modes for complex for-
mation with PAA, as shown in Fig. 9. Complexation of Cu2þ by
PAA, in those forms, shows there is inter- or intramolecular
crosslinking of polymer chains by Cu2þ. The binding trend with
pH observed here is explainable by, and consistent with, a
chemical transition of Cu2þ ions from this complexation equi-
librium with PAA below pH 6.5, to a monohydroxylated-
monocarboxylated Cu2þ species at pH 6.5, and further to the
insoluble Cu(OH)2 above pH 7, for which the formation constant
is known,[34] K ¼ 1017.3. Chelation experiments utilizing elec-
tron spin resonance spectrometry are inconclusive about the
structure of the formed complexes[34,35]; however, the quanti-
tative approach taken here is consistent with some of the
identified work.
increased in a mixed-metal solution, one would expect the
preferential chelation of particular ions in order of their binding
capacities and solubility.
There is a significant amount of copper monohydroxohy-
drate in solution at pH 6.5 and the AMSA concentration is not
high enough for AMSA to displace these hydroxo ligands from
the copper ions; this is why Cu2þ can bind in a higher molar ratio
than one with the polymeric AMSA.
This does not, however, explain the situation with Fe3þ. At
pH 3.5, Fe3þ should be highly mono-hydroxylated and still able
to form [Fe(AMSA)(OH)(H2O)3]. It is thought[25] that at pH 3.5
Fe3þ is more than monohydroxylated and that predominantly
bridged hydroxide species exist. Chelation of bridged iron(III)
hydroxides, therefore, pushed the binding ratio past a stoichio-
metric equilibrium. In Fe3þ binding experiments where poly-
chelate concentrations were high, blood-red coagulum was
visible inside the dialysis tubing. The latter indicates that
bridged Fe3þ complexes were forming a colloidal precipitate
and being flocculated by the polymer. Acidification of the
coagulum returned a clear solution.
AA segments of polymer chains would normally enhance the
solubility of the chelated polymer in solution, as was the case for
Cu2þ solutions. Binding of Fe3þ to a polychelate may have
initiated colloid formation, the bound Fe3þ acting as a collection
point for further Fe3þ through oxygen bridge formations. The
weight and reduced charge of these associated Fe3þ colloids
enabled hydrogen bonding by PAA segments of polymer chains
and, hence, flocculation. A combination of chelation, hydroxide
bridging, and flocculation is probably the cause of the observed
red coagulum and high molar binding ratios. The AMSA
monomer leads to synthetic polymeric coagulants that are ideal
polymeric couples for primary coagulants, such as iron(III), in
water purification.
Precipitation of weak polyelectrolytes is known,[31] and is
explained as an ‘yentropically driven complexation of acrylate
functional groups leading to precipitation of weak polyelec-
trolytes above a criticaly’ ratio rc ¼ [Mnþ]/[polymer]. This
finding was reflected here, in that precipitation of P(chelate) was
noted after equilibrium dialysis above ,400–600 ppm of
P(chelate).
Consideration of real systems, where metal ions other than
Cu2þ exist in solution, places Fe3þ in a position of importance
with SA-derived ligands. Iron(III) ions are well known for
forming strong complexes with salicylate derivatives.[19–23,25]
However, Fe3þ is not very soluble above pH 4, and solubility
presents a problem for conducting competitive binding studies
with P(chelate). Sequential extraction, combined with selective
separation, of metal ions is suggested for mixed-metal solutions,
Metal ions can be bound to P(chelate) in ratios far in excess of
stoichiometric ratios, but pH is the major determinant of this
new coagulation stoichiometry. Free hydroxides, from the
aqueous environment, were able to bind to metal ions and
thereby improve the metal uptake ratio.
Acknowledgements
The authors gratefully acknowledge the support of a Linkage grant
(LP0882618) from the Australian Research Council, and input from
Dr Cristy Warrender (UQ), Dr Chris Fellows (University of New England),
and Bradley Smith and Colin Mackenzie of CSBP Ltd (Kwinana, Western
Australia). GWW appreciates the award of a Gritton Scholarship while at the
University of Sydney.
References
[1] A. K. Bhardwaj, R. A. McLaughlin, I. Shainberg, G. J. Levy, Soil Sci.
Soc. Am. J. 2009, 73, 910. doi:10.2136/SSSAJ2007.0426
[2] A. K. Bhardwaj, I. Shainberg, D. Goldstein, D. N. Warrington, G. J.
Levy, Soil Sci. Soc. Am. J. 2007, 71, 406. doi:10.2136/
SSSAJ2006.0138
[3] G. J. Levy, I. Rapp, Aust. J. Soil Res. 1999, 37, 91. doi:10.1071/S97112
[4] A. I. Mamedov, S. Beckmann, C. Huang, G. J. Levy, Soil Sci. Soc. Am.
J. 2007, 71, 1909. doi:10.2136/SSSAJ2007.0096
[5] A. I. Mamedov, I. Shainberg, L. E. Wagner, D. N. Warrington, G. J.
Levy, Aust. J. Soil Res. 2009, 47, 788. doi:10.1071/SR09027
[6] A. I. Mamedov, L. E. Wagner, C. Huang, L. D. Norton, G. J. Levy, Soil
Sci. Soc. Am. J. 2010, 74, 1720. doi:10.2136/SSSAJ2009.0279
[7] I. Shainberg, D. Goldstein, A. I. Mamedov, G. J. Levy, Soil Sci. Soc.
Am. J. 2011, 75, 1090. doi:10.2136/SSSAJ2010.0293
[8] D. Sirjacobs, I. Shainberg, I. Rapp, G. J. Levy, Soil Sci. Soc. Am. J.
2000, 64, 1487. doi:10.2136/SSSAJ2000.6441487X

Metal Binding by Water-Soluble Polychelates
1601
[9] Z. Tang, T. Lei, J. Yu, I. Shainberg, A. I. Mamedov, M. Ben-Hur,
G. J. Levy, Soil Sci. Soc. Am. J. 2006, 70, 679. doi:10.2136/
SSSAJ2004.0395
[23] K. Patel, B. Suthar, J. Inst. Chem. 1987, 59, 272.
[24] A. L. R. Merce, A. S. Mangrich, B. Szpoganicz, N. M. Levy, J.
Felcman, J. Braz. Chem. Soc. 1996, 7, 239.
[10] J. Yu, T. Lei, I. Shainberg, A. I. Mamedov, G. J. Levy, Soil Sci. Soc.
[25] J. Suh, H. S. Park, J. Polym. Sci., Part A: Polym. Chem. 1997, 35, 1197.
doi:10.1002/(SICI)1099-0518(199705)35:7,1197::AID-POLA5.3.
0.CO;2-#
[26] B. L. Rivas, S. A. Pooley, M. Soto, K. E. Geckeler, J. Polym. Sci., Part
A: Polym. Chem. 1997, 35, 2461. doi:10.1002/(SICI)1099-0518
(19970915)35:12,2461::AID-POLA16.3.0.CO;2-5
[27] R. Subramanian, P. Natarajan, J. Polym. Sci., Part A; Polym. Chem.
1984, 22, 437. doi:10.1002/POL.1984.170220215
[28] F. Candau, C. Holtzscherer, J. Chim. Phys. 1989, 86, 2095.
[29] F. Candau, Y. S. Leong, R. M. Fitch, J. Polym. Sci., Part A: Polym.
Chem. 1985, 23, 193. doi:10.1002/POL.1985.170230120
[30] R. A. Cave, S. A. Seabrook, M. J. Gidley, R. G. Gilbert, Biomacro-
molecules 2009, 10, 2245. doi:10.1021/BM900426N
Am. J. 2003, 67, 630. doi:10.2136/SSSAJ2003.0630
[11] R. D. Lentz, R. E. Sojka, Agron. J. 2009, 101, 305. doi:10.2134/
AGRONJ2008.0100X
[12] R. E. Sojka, D. L. Bjorneberg, J. A. Entry, R. D. Lentz, W. J. Orts, Adv.
Agron. 2007, 92, 75. doi:10.1016/S0065-2113(04)92002-0
[13] C. O. Nebo, L. O. Asuquo, C. K. Nworu, M. C. Fuerstenau, J. Disper.
Sci. Technol. 1996, 17, 13. doi:10.1080/01932699608943485
[14] C. O. Nebo, H. Nwajiobi, N. A. G. Aneke, M. C. Fuerstenau, J. Disper.
Sci. Technol. 1996, 17, 1. doi:10.1080/01932699608943484
[15] B. P. Singh, R. Singh, Sep. Sci. Technol. 1997, 32, 993. doi:10.1080/
01496399708000941
[16] P. K. Weissenborn, Int. J. Miner. Process. 1996, 47, 197. doi:10.1016/
0301-7516(95)00096-8
[17] E.-G. A. A. Abd, A. M. El-Roudi, E. M. Soliman, M. R. Mahmoud,
[31] J. P. Boisvert, A. Malgat, I. Pochard, C. Daneault, Polymer 2002, 43,
141. doi:10.1016/S0032-3861(01)00603-6
J. Chin. Chem. Soc.-Taip. 1987, 34, 291.
[32] J. Leja, Surface Chemistry of Froth Flotation 1982 (Plenum:
New York, NY).
[18] F. Bonaccorsl, P. Albonico, R. Pappa, T. P. Lockhart, Int. J. Polym.
Mater. 1992, 18, 165. doi:10.1080/00914039208029318
[19] J. F. Kennedy, S. A. Barker, J. Epton, G. R. Kennedy, J. Chem. Soc.,
Perkin Trans. 1 1973, 488. doi:10.1039/P19730000488
[20] J. F. Kennedy, S. A. Barker, A. W. Nicol, A. Hawkins, J. Chem. Soc.,
Dalton Trans. 1973, 1129. doi:10.1039/DT9730001129
[21] J. F. Kennedy, J. Epton, Carbohydr. Res. 1973, 27, 11. doi:10.1016/
S0008-6215(00)82419-2
[33] A. W. Trochimczuk, J. Jezierska, Polymer 2000, 41, 3463.
doi:10.1016/S0032-3861(99)00481-4
[34] J. Francois, C. Heitz, M. M. Mestdagh, Polymer 1997, 38, 5321.
doi:10.1016/S0032-3861(97)00095-5
[35] K. Kruczala, S. Schlick, J. Phys. Chem. B 1999, 103, 1934.
doi:10.1021/JP9839697
[22] J. F. Kennedy, J. Epton, G. R. Kennedy, Antimicrob. Agents Che-
mother. 1973, 3, 29.