The effect of co-precipitation on cadmium(II) adsorption on hydrous aluminium(III) hydroxide in the presence of a range of chelates

Article abstract of DOI:10.1039/b201175a

The adsorption of cadmium(II) on freshly precipitated aluminium(III) hydroxide in the presence of a range of chelates has been investigated. By precipitating the metal, chelate and adsorbent together it is possible to change the pH variation of the metal-complex adsorption from anionic, "ligand-like", binding to cationic binding. This is a general phenomenon and is explained by the formation of a ternary Al-O-Cd-L surface species. As a consequence of the preparation method, the pH edge is found to shift to lower pH values in the presence of the chelate which gives rise to an apparent increase in adsorption of Cd²⁺. This increase is, in general, most pronounced at [chelate]/[metal] > 1. Computer modelling shows that the observed trends result from the competition between Al-O-Cd-L and Al-L for the available aluminium(III) binding sites. The enhanced adsorption in the presence of phenylenediaminetetraacetate is anomalous since it is observed at a [chelate]/[metal] ≈ 0.1 and cannot be interpreted by the simple competition model.

Full text of DOI:10.1039/b201175a

View Article Online / Journal Homepage / Table of Contents for this issue
The effect of co-precipitation on cadmium(II) adsorption on hydrous
aluminium(III) hydroxide in the presence of a range of chelates
Michael G. Burnett, Christopher Hardacre* and Heather J. Mawhinney
School of Chemistry, David Keir Building, The Queen’s University of Belfast, UK BT9 5AG.
E-mail: c.hardacre@qub.ac.uk; Fax: +44 (0)28 9038 2117; Tel: +44 (0)28 9027 4592
Received 31st January 2002, Accepted 15th April 2002
First published as an Advance Article on the web 19th June 2002
The adsorption of cadmium(II) on freshly precipitated aluminium(III) hydroxide in the presence of a range of
chelates has been investigated. By precipitating the metal, chelate and adsorbent together it is possible to change
the pH variation of the metal-complex adsorption from anionic, ‘‘ligand-like’’, binding to cationic binding.
This is a general phenomenon and is explained by the formation of a ternary Al–O–Cd–L surface species. As a
consequence of the preparation method, the pH edge is found to shift to lower pH values in the presence of the
chelate which gives rise to an apparent increase in adsorption of Cd2+. This increase is, in general, most
pronounced at [chelate]/[metal] > 1. Computer modelling shows that the observed trends result from the
competition between Al–O–Cd–L and Al–L for the available aluminium(III) binding sites. The enhanced
adsorption in the presence of phenylenediaminetetraacetate is anomalous since it is observed at a [chelate]/
[metal] ꢀ 0.1 and cannot be interpreted by the simple competition model.
equimolar mixture with EDTA, moreover, was more strongly
adsorbed as the pH10,11 increased although the trend reported
previously for g-alumina was exactly the reverse.14
In this paper, we show that the changes observed in the pre-
sence of EDTA are common to a wide range of EDTA like
chelates. The model proposed11 to explain the EDTA adsorp-
tion, is shown to be capable of predicting the behaviour of all
the systems studied apart from PhDTA. Fig. 1 shows the struc-
tures and abbreviations for all chelates used in this study.
Introduction
Inorganic coagulants, such as aluminium (III), are widely used
for cleaning potable water and a variety of aqueous effluents.1
The process depends on adding the metal salt to the water
under treatment and adjusting the pH of the mixture to the
range within which the amorphous metal hydroxide is precipi-
tated. The resulting ‘floc’ is removed carrying with it pollu-
tants such as micro-organisms, particulate materials,
inorganic species and organic substances. They are removed
by particle entrapment, colloid destabilisation and in the case
of heavy metal cations, adsorption of dissolved compounds.2
Experimental
This paper describes the results from the adsorption of Cd2+
on hydrous aluminium hydroxide in the presence of a variety
of chelates. Understanding the effect of chelates and organic
matter on the speciation of metals in solution in the presence
of hydroxides is important and has been studied extensively.3
Analytical methods
Metal ion concentrations were measured using a Perkin Elmer
Atomic Emission Spectrometer Plasma 400. Solution pH was
measured by a Metrohm 713 pH meter, calibrated before each
set of observations using standard NBS buffers at pH 4, 7 and
10. The reported hydrogen ion concentrations were calculated
from pH measurements using the standard NBS activity cor-
rection procedure.15 The effect of varying liquid junction
potentials16 calculated from the Henderson equation was neg-
ligible.
The adsorption of a variety of metals such as Cu2+, Zn2+
,
Pb2+ as well as Cd2+ on both iron and aluminium based sys-
tems has been studied in the presence of small chelates, such
as EDTA, as well as polymeric matter such as humic and fulvic
acids, for example.4–8 The presence of chelates can be proble-
matic in water treatment. For example, precipitation of heavy
metals which form insoluble hydroxides and carbonates is also
used to treat waste water; however the presence of chelates
which complex the metal prevent efficient and reliable removal
of the pollutant.9
Reagents
Sodium hydroxide, disodium ethylenediaminetetraacetate
dihydrate (EDTA), N,N,N0,N0-phenylenediaminetetraacetic
acid (PhDTA), trans-1,2-diaminocyclohexane-N,N,N0,N0-tet-
raacetic acid monohydrate (CDTA), 1,3-diaminopropane-
N,N,N0,N0-tetraacetic acid (PDTA), trisodium [S,S]-ethylene-
diamine-N,N0-disuccinate (EDDS), nitrilotriacetic acid
(NTA), aluminium nitrate nonahydrate, zirconyl chloride
octahydrate, and nitric acid were Analar quality. Cadmium
nitrate tetrahydrate was GP quality. Distilled deionised 18.2
MO water was used to prepare the stock solutions.
The present study is part of a continuing research project
aimed at improving the efficiency of flocculation in removing
heavy metals from wastewater in the presence of chelates. This
research has therefore concentrated on freshly precipitated
hydrous oxides. In many of the previous studies, especially
those involving aluminium, either aged or dried samples or
rehydrated aluminium oxide were studied which behave differ-
ently from freshly precipitated systems. For example, the
adsorption of copper, nickel and cadmium was enhanced by
the presence of EDTA at pH 710,11 with freshly precipitated
aluminium hydroxide but no comparable effect was reported
in similar experiments with g-alumina.12,13 Cadmium in an
1,6-Diaminohexane-N,N,N0,N0-tetraacetic acid (HDTA) was
synthesised using the method developed by Ogino et al.17
Monochloroacetic acid (17 g, 0.18 mol) was dissolved in water
3828
Phys. Chem. Chem. Phys., 2002, 4, 3828–3834
This journal is # The Owner Societies 2002
DOI: 10.1039/b201175a

View Article Online
process in fully removing the precipitate from the supernatant
solution at pH values below 6. Therefore, in order to be able to
analyse both the precipitate and the supernatant without the
results for the latter being compromised by the presence of pre-
cipitate, filtration using 0.45 mm Millipore PVDF membrane
was employed. Comparison of the two separation techniques
using higher pH values showed little difference in the results
obtained. The volumes for each floc formed were measured
by dissolving the floc pellet in a known volume of 68% nitric
acid and measuring the difference in volume between the dis-
solved pellet and the added acid. For all experiments the ionic
strength was 0.037 M.
Modelling
The computer model used to simulate the adsorption reactions
was developed using Microsoft Excel and copies are available
from the principal author on request. Details of the calcula-
tions and the procedures used have been published pre-
viously11 and are further discussed below.
Results and discussion
Kinetics of adsorption
The results reported in this paper have all been obtained under
conditions broadly described as co-precipitation although,
strictly speaking, the mechanism is that of floc precipitation
followed by metal ion and chelate adsorption. The quasi-equi-
librium is reached rapidly in the 5–10 min before the first
observation can be made and no significant variation occurs
thereafter. This is shown in detail for cadmium with both
EDTA and EDDS in Fig. 2. Fig. 2 shows that the degree of
adsorption was constant over a period of 50 h following the
initial precipitation at either pH 6.5 or 7.0.
Fig. 1 Structures of (a) EDTA (n ¼ 2); PDTA (n ¼ 3); HDTA
(n ¼ 6), (b) EDDS, (c) NTA, (d) CDTA and (e) PhDTA.
(50 cm3) to which sodium hydroxide (14.4 g, 0.36 mol), dis-
solved in the minimum amount of water, was added drop-wise,
with stirring, at 10 ꢁC. This was followed by the addition of
hexamethylenediamine (3.5 g, 0.03 mol). The resulting solution
was heated at 100 ꢁC for 1 h and was left to stand overnight.
The pH of the cooled solution was adjusted to 2.3 with 36%
concentrated hydrochloric acid and was rotary evaporated giv-
ing a white powder containing sodium chloride. The mixture
was washed three times with distilled deionised 18.2 MO water
to remove the sodium chloride and the solid recrystallised from
hot water, resulting in the hydrate. Recrystallisation was
repeated twice. A yield of 4.66 g, 44.6%, was obtained.
NMR d1H/ppm (D2O) 1.25 (4H, s), 1.45 (4H, s), 2.55 (4H,
t), 3.20 (8H, s), CHN found: C, 47.63; H, 6.79; N, 7.76%.
Calc.: C, 48.28; H, 6.90; N, 8.05%. n(IR)/cmꢂ1: 1300–1375
(–CH2– bend), 1691 (C=O), 3000–3150 (C–H stretch), 3450
(OH).
Equilibrium simulation
The data are interpreted by the following set of balanced reac-
tions:
Al3þ þ edta4ꢂ Ð AlðedtaÞꢂ
ð1Þ
ð2Þ
ð3Þ
ð4Þ
ð5Þ
2ꢂ
Cd2þ þ edta4ꢂ Ð CdðedtaÞ
Ggel þ Cd2þ Ð Cd2þ; Ggel þ xHþ
Cd2þ þ edta4ꢂ þ Ggel Ð Cd2þ; edta4ꢂ; Ggel þ yHþ
Ggel þ edta4ꢂ Ð edta4ꢂ; Ggel
Al3þ þ 3H2O Ð Ggel þ 3Hþ
ð6Þ
ð7Þ
Floc preparation by co-precipitation
Ggel þ H2O Ð AlðOHÞ4ꢂ þ Hþ
The floc suspensions were prepared in 50 cm3 beakers using
stock solutions at ambient temperature, 22 ꢃ 1 ꢁC, measured
by the pH meter platinum resistance thermometer. A mixture
containing various concentrations of aluminium, cadmium
and chelate, 30 cm3 in total volume, was magnetically stirred
while the pH was adjusted to within ꢃ0.005 pH units of the
nominal value using 5 M sodium hydroxide and 0.1 M nitric
acid as required. The pH was adjusted over various time peri-
ods, up to 2 days, to ensure that equilibrium had been
achieved. It should be noted that after the initial rapid changes
of pH within the first 10 min, the maximum pH change
observed was only ꢂ0.1 pH units. Following pH adjustment,
the solution was decanted into a 50 cm3 polyethylene tube
and centrifuged for 5 min at 2500 rpm (equivalent to 1120
g). The supernatant solution was decanted and the precipitate
was dissolved in 1 cm3 68% nitric acid and made up to 50 cm3
total volume. In computing the amounts of aluminium and
cadmium in the precipitate, allowance was made for the super-
natant solution, which was unavoidably transferred with the
solid phase. Centrifugation was not found to be an efficient
2ꢂ
AlðedtaÞꢂ þ H2O Ð AlðOHÞðedtaÞ þ Hþ
ð8Þ
ð9Þ
ð10Þ
ð11Þ
Al3þ þ 3H2O Ð AlðOHÞ3ðaqÞ þ 3Hþ
Al3þ þ 2H2O Ð AlðOHÞ2þ þ 2Hþ
2þ
Al3þ þ H2O Ð AlðOHÞ þ Hþ
in which the symbol Ggel represents the hydrous floc surface
free of adsorbed edta4ꢂ or Cd2+. Throughout the text, the
equilibrium constants for each reaction (x) are denoted by
Kx . Reaction (8) was not originally included11 in the equili-
brium set because of its supposed trivial contribution to the
adsorption equilibrium. Subsequently it was realised that
Al(OH)(edta)2ꢂ did have a marked effect on the absolute equi-
librium concentrations and consequently on the absolute value
of the fitted constants K3 , K4 and K5 . The inclusion of reaction
(8), however, does not affect the overall fitting of the observed
trends to the proposed model but only the fitting parameters
required.
Phys. Chem. Chem. Phys., 2002, 4, 3828–3834
3829

View Article Online
the dissolved uncomplexed cadmium concentration is then
always negligible so that Cd(OH)+ and Cd(OH)2 need not be
included.
Equilibria (1), (2) and (6) have been studied previously over
a wide range of temperatures and ionic strengths11 and these
results have been formulated for this study using the techni-
ques developed by Daniele et al.19 in order that the constants
may be calculated for virtually any temperature and ionic
strength. Reactions (7), (9)–(11)25 and (8)27 are less thoroughly
documented and corrected only for the effects of ionic strength
using the Davies equation. The experiments modelled in this
paper have been performed at 22 ꢁC and an ionic strength of
0.037 M so that the fitted constants for equilibria (3)–(5) are
single values appropriate for that temperature and ionic
strength. The constants for reactions (1),28 (2)29 and for the
acid dissociations30 of EDDS used in modelling with the latter
are generally less well established than those for EDTA. They
have been corrected for ionic strength variation only.19
The simultaneous equations which may be derived11 for the
ion concentrations in the equilibria (1)–(11) are solved digitally
by starting from plausible estimates of the unknown concen-
trations and iterating using the procedures found in Microsoft
Excel until the final and penultimate values agree. Generally
ten or twenty cycles are sufficient for the results to converge.
The modelling studies broadly indicate that the trends in the
enhanced adsorption depend on the competition between reac-
tions (4) and (5). The numerical values of the constants K4 and
K5 are altered by changes in the postulated solution equilibria
but the final agreement of model and observation is unaffected.
Fig. 2 Variation of %Cd contained in the floc adsorbed with time
from a co-precipitated solution at pH 6.5 (circles) and pH 7.0 (squares)
containing 333 ppm Al, 3.33 ppm Cd (12.3 mM Al3+ and 29.7 mM
Cd2+) with 0.067 mM EDDS (unfilled symbols) and 0.033 mM EDTA
(filled symbols).
ˆ
The species included are only those which could play a role
in the enhanced adsorption of cadmium. A full equilibrium set
would include many other species which never attain signifi-
cant concentration and which only confuse the interpretation
of the observed adsorption.
In the case of EDTA, it is necessary to make allowance for
the effect of the various deprotonated forms in calculating the
concentration of edta4ꢂ even though the other forms do not
participate in the cadmium(II) or aluminium(III) binding. The
ion edta4ꢂ, which is the dominant complexing form of EDTA,
is the only EDTA species to appear in the simulation but the
effect of the other EDTA species is allowed for by the use of
conditional constants18 for K1 , K2 and K5 . A full allowance
is made for H4edta, H3edtaꢂ, H2edta2ꢂ, Hedta3ꢂ, as well as
the ion-paired species derived from the presence of the sodium
counter-ions19 and a minor contribution20 due to the forma-
tion of Cd(NO3)+. Each conditional constant is calculated
using a,
Cadmium adsorption in the presence of EDTA
Fig. 3 shows the variation of adsorbed Cd2+ in the presence
and absence of EDTA with pH. Both sets of data show an
increase in adsorption with pH but in the presence of EDTA
the pH edge has been shifted from 7.0 to 6.0. The shift in
the pH edge leads to an apparent enhanced adsorption at
pH 7.0 in the presence of EDTA over the system where the
chelate is not present. The lines in Fig. 3 show the predicted
trend from the model simulated using eqns. (1)–(11) above.
Fig. 4 indicates how the concentration of the species contained
within the model vary with pH for the co-precipitation of Cd2+
and Al3+ in the presence of EDTA.
a ¼ ½edta4ꢂꢄ=ð½edta4ꢂꢄ þ ½Hedta3ꢂꢄ þ ½H2edta2ꢂꢄ
The variation of adsorbed Cd2+ with increasing concentra-
tions of EDTA is shown in Fig. 5. As the concentration of
EDTA increases, a sharp rise in Cd2+ adsorption is observed
which reaches a maximum at 0.033 mM. At this point, the
þ ½H3edtaꢂꢄ þ ½H4edtaꢄÞ
Kconditional ¼ aK
where K is the thermodynamic binding constant for a reaction
involving edta4ꢂ
.
There are many other complexes which have been have been
suggested for aluminium(III) and cadmium(II) under acid or
alkaline conditions. At low pH there is a range of protonated
EDTA complex ions which have been invoked to explain equi-
librium measurements for cadmium21 and for aluminium.22
However, within the pH range from 5–10, their concentration
7+
appears to be negligible. The ion Al13O4(OH)24(H2O)12
should apparently be formed23 within this pH range but it
could not be detected by 27Al NMR spectroscopy and has been
omitted.11 It is known that it is formed only slowly and under
favourable conditions.
The concentration of the aluminium species in solution is
controlled by the solubility of the floc, equilibrium (6) and
its speciation by equilibria (1), (7)–(11).24,25 At a pH below
6, the aquo aluminium ion and the hydroxyaluminium ions
+
Al(OH)2+ and Al(OH)2 dominate the speciation, whereas
Fig. 3 Variation of %Cd adsorbed with increasing pH from a co-pre-
cipitated solution containing 333 ppm Al, 3.33 ppm Cd (12.3 mM Al3+
and 29.7 mM Cd2+) and 0.033 mM EDTA (S, dashed line) and distilled
water (K, solid line). Lines indicate predicted trend from the model and
the % Al (N) within the floc is shown for the solution containing
EDTA.
the Al(OH)3 (aq) species is never present in significant concen-
trations at any pH between 5 and 10. Above pH 6.5 the
Al(OH4)ꢂ ion is the predominant soluble aluminium species
present.24 The aquo cadmium ion is known26 to deprotonate
significantly to form Cd(OH)+ and Cd(OH)2 above pH 9 but
3830
Phys. Chem. Chem. Phys., 2002, 4, 3828–3834

View Article Online
in the majority of studies, we mixed the Cd2+, Al3+ and EDTA
prior to pH adjustment so that they were finally co-precipi-
tated, i.e. the hydroxide is formed in the presence of the
Cd2+ and EDTA. If the hydroxide is formed prior to adsorp-
tion, the adsorption of Cd2+ decreases at pH values above 6.5,
in the presence of EDTA which is consistent with the binding
of an anionic Cd–EDTA complex and is in agreement with
many previous studies.12,14 The decrease in adsorption with
decreasing pH cannot be explained via the dissolution of the
floc. For example, when the Cd2+, Al3+ and EDTA are copre-
cipitated, there is a substantial decrease in the adsorbed Cd2+
contained within the hydrous floc at pH 6.0, from 100% to
17%, yet less than 6% Al3+ dissolution is observed, as shown
in Fig. 3. It should be noted that it is only below pH 5.5 that
dissolution of Al3+ becomes significant and this is independent
of the order in which the reagents are precipitated or whether
chelate is present.31
In the absence of EDTA, the metal binds as uncomplexed
metal to the floc oxide surface, eqn. (3), and shows typical
cationic adsorption characteristics associated with the change
in surface charge of the floc surface with pH. With increasing
concentrations of EDTA, the rise in cadmium adsorption is
not due to the trapping of a cadmium ion within an EDTA
molecule, which is then bound to the aluminium ions in the
precipitate as an anion and would show a ligand like binding
pH curve, but rather to the formation of a distinct ternary
complex, as shown in eqn. (4), resulting in the opposite (catio-
nic) pH profile, as described above. This type of reaction may
perhaps depend on the reactivity of small aluminium(III) aggre-
gates formed during precipitation as opposed to the final stable
floc surface.
Fig. 4 Variation in the concentration of the dominant (a) complexed
EDTA and (b) non-EDTA species present during co-precipitation of a
solution containing 333 ppm Al, 3.33 ppm Cd (12.3 mM Al3+ and 29.7
mM Cd2ꢂ+) and 0.033 mM EDTA predicted from the model. The
Al(OH)4 concentration is divided by 50.
Cadmium adsorption in the presence of EDDS
Cationic-like binding of Cd2+ in the presence of a chelate is a
general phenomenon when Al3+, Cd2+ and chelate are co-pre-
cipitated. Fig. 6 shows the variation of adsorbed Cd2+ on alu-
minium hydroxide with pH in the presence of EDDS. EDDS is
of particular interest since it is readily biodegradable both as
the free ligand and when complexed with some heavy metals,
EDTA and cadmium present are approximately equimolar.
Further increases in EDTA result in a drop in adsorbed
Cd2+ with a concomitant decrease in the Al3+ contained
within the precipitate.
Increasing metal adsorption in the presence of a chelate with
increasing pH is unusual. Commonly in the presence of che-
lates, the opposite variation is observed,14 i.e. a decrease in
adsorption with increasing pH, typifying metal–EDTA anio-
nic, ‘‘ligand-like’’, adsorption. The profile observed in Fig. 5
indicates cationic adsorption normally associated with metal
adsorption in the absence of a chelate. This change in pH pro-
file is a direct consequence of the preparation method. Unlike
for example Cd2+ 32 As with EDTA, Cd2+ adsorption
.
increases with increasing pH and the pH edge is shifted from
approximately 7.0, for the metal alone, to 6.0, when EDDS
is present. It should be noted that the pH variation of the
Al3+ contained within the floc is similar whether co-precipita-
tion occurs in the presence of EDDS or EDTA.
Fig. 7 shows the variation of adsorbed Cd2+ with increasing
concentration of EDDS. As with EDTA, a rapid increase in
adsorption is observed. With EDDS, the limit is reached at
Fig. 5 Variation of %Al (K, solid line) and %Cd (S, dashed line)
contained in the floc with increasing EDTA concentration formed
from a solution adjusted to pH 7.0 containing 333 ppm Al and 3.33
ppm Cd (12.3 mM Al3+ and 29.7 mM Cd2+) under co-precipitated con-
ditions. Lines indicate predicted trend from the model.
Fig. 6 Variation of %Cd adsorbed with increasing pH from a co-pre-
cipitated solution containing 333 ppm Al, 3.33 ppm Cd (12.3 mM Al3+
and 29.7 mM Cd2+) and 0.067 mM EDDS (S, dashed line) and distilled
water (K, solid line). Lines indicate predicted trend from the model.
Phys. Chem. Chem. Phys., 2002, 4, 3828–3834
3831

View Article Online
The decrease in Cd2+ adsorption at high EDTA and EDDS
concentrations cannot simply be explained by dissolution of
the precipitate. The decrease in adsorbed Cd2+ is too large
given the small decrease in hydroxide present. Although this
effect contributes, we believe the major effect is due to a com-
bination of the strong adsorption of the chelate on the floc,
resulting in site blocking, and the chelate stripping the metal
from the floc. In the case of EDDS, such a decrease is not
observed mainly due to the much lower complexation constant
for Cd2+ compared with EDTA. This is easily seen if eqns. (2),
(4) and (5) are combined:
Cd2þ; L4ꢂ; Ggel þ L4ꢂ þ yHþ Ð Cd2þ; L4ꢂ þ L4ꢂ; Ggel ð12Þ
logK12 ¼ logK2 þ logK5 ꢂ logK4
For EDTA, log K12 is 17.65 whereas for EDDS it is less than
5.72. EDDS has a much lower ability to strip the floc of the
Cd2+ than does EDTA and hence the balance remains in
favour of the adsorbed Cd2+–EDDS complex rather than that
in solution, even at relatively high chelate concentrations. Site
blocking is a significant but secondary effect to chelate strip-
ping and it only affects the metal adsorption at relatively low
chelate concentrations before the onset of surface saturation.
For example, the model shows that above 1 mM EDTA there
is little variation in the amount of adsorbed chelate since the
floc surface is close to saturation. Below 1 mM increased che-
late adsorption mirrors the drop in metal adsorbed. More var-
iation is observed for EDDS, however, because the floc surface
is more unsaturated due to the much lower binding constant
for the floc with EDDS compared to that with EDTA.
Fig. 7 Variation of %Al (K, solid line) and %Cd (S, dashed line)
contained in the floc with increasing EDDS concentration formed from
a solution adjusted to pH 7.0 containing 333 ppm Al and 3.33 ppm Cd
(12.3 mM Al3+ and 29.7 mM Cd2+) under co-precipitated conditions.
Lines indicate predicted trend from the model.
0.07 mM, which corresponds to a chelate concentration twice
that of the cadmium in the system. Unlike the case of EDTA,
further increases in chelate concentration result in very little
decrease in the adsorbed cadmium, even when the chelate con-
centration is 3.3 mM, i.e. one hundred-fold more concentrated
than the cadmium. Compared with EDTA, adsorption in the
presence of EDDS results in little reduction in the Al3+ con-
tained in the floc. At pH 7.0, more than 95% Al3+ was retained
within the precipitate over all EDDS concentrations studied.
The model equilibria (1)–(11), have been used to fit the var-
iation of cadmium adsorption on co-precipitation as shown in
Figs. 3 and 5. The results for Cd2+ co-precipitated with Al3+
and EDDS are also fitted well by this model. In Figs. 5 and
6, the lines show the simulated data. Table 1 summarises the
constants for each species used to model the experimental data.
As described above, EDDS increases the adsorption of Cd2+
but less effectively than does EDTA at the same concentration.
The model predicts this simply as a consequence of the very
strong affinity of the gel surface for EDTA compared with
EDDS.
Cadmium adsorption in the presence of PDTA, HDTA,
CDTA and NTA
The results obtained in the previous sections are also mirrored
by those obtained using a range of other chelates. Fig. 8 shows
the variation of adsorbed Cd2+ with increasing chelate concen-
tration for PDTA, HDTA, CDTA and NTA. Extending the
alkyl-chain from C2 to C3 , PDTA, resulted in similar results
to EDTA. Maximum adsorption was achieved at equimolar
proportions and, as with EDTA, the adsorbed Cd2+ is found
to decrease at high chelate concentrations. Using HDTA, the
adsorption limit for Cd2+ was only reached at a chelate con-
centration of 0.33 mM. In this case, the maximum Cd2+
removal was observed when the chelate concentration was
approximately ten-fold greater than that of the Cd2+. A simi-
lar result was observed for CDTA. In the presence of PDTA,
HDTA and CDTA, only 80–86% of the Cd2+ could be
adsorbed compared with >95% for EDTA and EDDS. Redu-
cing the denticity of the chelate only changes the values of
maximum adsorption but not the general shape of the curve.
Using NTA, a maximum adsorption of Cd2+ of 67% is found
when [NTA]/[Cd2+] is approximately 1.5. In all cases the
In order to model the retention of Cd2+ on the floc success-
fully at high chelate concentrations, it is necessary to reduce
the adsorption coefficient of the chelate on the floc. This would
not be surprising in view of the structure of EDDS compared
with EDTA. It has been postulated that chelates bind to
hydrous flocs using hydrogen bonding.6,33 In the case of
EDDS, the carboxylic acid groups are closer than in EDTA,
so that an intramolecular hydrogen bond could form a stable
six membered ring. This may explain the reduction in chelate
adsorption predicted by the model while the decreased bite
of the chelate explains the reduced stability of the dissolved
metal complexes.
Table 1 Constants used for eqns. (1)–(11) used in generating the curves in Figs. 3–7 (temperature 22 ꢁC, experimental ionic strength 0.037 M and
pH ¼ 7; x ¼ 1,a y ¼ 0,a number of surface sites per Al atom ¼ 0.08a )
104a
log K1
log K2
log K3
log K4
log K5
b
b
a
a
a
Chelate
EDTA
EDDS
8.05
2.39
13.51
11.62
13.29
7.60
ꢂ4.28
ꢂ4.28
3.91
4.28
8.27
<2.4
b
b
b
b
b
b
Chelate
log K6
log K7
log K8
log K9
log K10
log K11
EDTA
EDDS
ꢂ10.89
ꢂ10.89
Constants fitted to observed results. Constants taken from the references as described in the text.
ꢂ12.20
ꢂ12.20
ꢂ6.07
ꢂ17.25
ꢂ17.25
ꢂ10.55
ꢂ10.55
ꢂ5.30
ꢂ5.30
—
a
b
3832
Phys. Chem. Chem. Phys., 2002, 4, 3828–3834

View Article Online
Fig. 10 Variation of %Cd adsorbed with increasing pH from a co-
precipitated solution containing 333 ppm Al, 3.33 ppm Cd (12.3 mM
Al3+ and 29.7 mM Cd2+) and 3.3 mM PhDTA (S, dashed line) and dis-
tilled water (K, solid line). The lines are schematic only and have been
included to show the trend.
Fig. 8 Variation of %Cd contained in the floc with increasing HDTA
(K), PDTA (N), CDTA (S) and NTA (;) concentration formed from
a solution adjusted to pH 7.0 containing 333 ppm Al and 3.33 ppm Cd
(12.3 mM Al3+ and 29.7 mM Cd2+) under co-precipitated conditions.
The lines are schematic only and have been included to show the trend.
with EDTA, an increase in adsorption of Cd2+ increases with
pH indicating cationic-like adsorption.
model can be fitted to the experimental data without major dis-
crepancies.
Floc and adsorbate structure
Cadmium adsorption in the presence of PhDTA
It might be argued that the method of preparation should have
little effect on the adsorption characteristics since all systems
will reach a common equilibrium eventually but aluminium
hydroxide polymorph equilibria are notoriously difficult to
achieve.34 On the other hand, in the co-precipitated system stu-
died here, the results in Fig. 2 suggest that equilibrium is
rapidly established. This is in good agreement with the exten-
sive study performed by Dario and Ledin,7 where the systems
were equilibrated for various times up to 168 h and showed no
time variation. The final structures formed here are kinetically
stable even if they have not reached true thermodynamic equi-
librium. It is noticeable that where adsorption varied strongly
with time, the adsorbent had been dried prior to use. The equi-
libration times required become understandable since solid-
state diffusion within the core of the particles then becomes
the rate-limiting step.
It is worth noting that in every case except that of PhDTA,
there is a steady progression, shown in Figs. 5, 7 and 8, from
the normal adsorption to the maximum adsorption in the pre-
sence of chelate. In the case of PhDTA there is a sudden dis-
continuous increase, Fig. 9. Although it is possible to argue
that a steady progression might occur at sufficiently low con-
centrations of PhDTA, the observed change is more like the
transformation due to the nucleation and rapid growth of a
different structural form of the hydrous aluminium floc than
found in the other systems studied.
Fig. 9 shows the variation in adsorbed Cd2+ with increasing
PhDTA concentration at pH 7. As expected from the other
chelates described above, a rapid rise in the adsorbed Cd2+
is observed as the chelate concentration is increased. However,
compared with EDTA, the limiting metal adsorption is
reached at a much reduced concentration of chelate. Only
0.0033 mM PhDTA is required to cause >95% Cd2+ adsorp-
tion. This corresponds to a ten-fold molar excess of metal over
the chelate.
It is not possible to model this variation using the equilibria
(1)–(11) in which it is assumed that the floc structure is unaf-
fected by the presence of the chelate. The adsorption profile
on co-precipitating with PhDTA qualitatively behaves like that
of EDTA although only one tenth of the concentration of
PhDTA is required to produce an equivalent shift of the pH
edge. Fig. 10 shows the variation of adsorbed Cd2+ with pH
when co-precipitated in the presence of PhDTA. As observed
There is some evidence that this may indeed be the case. In
the presence of PhDTA, the co-precipitated floc retains much
more water on centrifugation than in the absence of a chelate
or in the presence of smaller chelates such as EDTA or EDDS.
On co-precipitation, the floc formed using PhDTA has a
volume 15% larger than in all other cases, suggesting a change
in the surface properties of the floc particles. This change in
structure may simply be a strengthening of the hydroxide lat-
tice, so as to enhance the water retention but any change in
structure will also change the surface acidity slightly and alter
the number of deprotonated adsorption sites. Structural
changes due to the presence of chelates have been described
previously. Szekeres et al. showed that adsorption of salicylate
on alumina caused a disordering of the surface.6,35,36 In that
study, the strong chelating ability of the salicylate disrupts
the surface order and changed the structure of the aluminium
hydroxide formed.
Fig. 9 Variation of %Al (K) and %Cd (S, solid line) contained in the
floc with increasing PhDTA concentration formed from a solution
adjusted to pH 7.0 containing 333 ppm Al and 3.33 ppm Cd (12.3
mM Al3+ and 29.7 mM Cd2+) under co-precipitated conditions. The
variation of %Cd (N, dashed line) with EDTA concentration is shown
for comparison. The lines are schematic only and have been included
to show the trend.
Phys. Chem. Chem. Phys., 2002, 4, 3828–3834
3833

View Article Online
J. Jiang and N. Graham, Chem. Br., 1998, 34(3), 338.
S. Goldberg, J. A. Davis and J. D. Hem, in The Environmental
Chemistry of Aluminium, ed. G. Sposito, Lewis Publishers, Boca
Raton, FL, 1995, ch. 4.
B. Nowack and L. Sigg, J. Colloid Interface Sci., 1996, 117,
106.
H. A. Elliot and C. P. Huang, J. Colloid Interface Sci., 1979, 70,
29.
D. C. Girvin, P. L. Gassman and H. Bolton, Jnr., Soil Sci. Soc.
Am. J., 1993, 57, 47.
M. Dario and A. Ledin, Chem. Speciation Bioavailability, 1997, 9,
3.
J. A. Davis, Geochim. Cosmochim. Acta, 1984, 48, 679.
2
3
The effect of PhDTA in reducing the pKa of the surface will
shift the pH edge observed for cadmium. Although a ternary
surface complex is also likely to be formed at higher chelate
concentrations, the additional enhancement will not be easily
detected since the metal adsorption is already so high. Analo-
gous changes in the adsorption constants with structural varia-
tions have been shown by Nowack et al.14 By changing from
hydrous ferric oxide to goethite, the pH edge for the adsorp-
tion of Ni–EDTA complex is shown to shift from 6.5 to 8.1.
A slightly smaller shift in the pH edge from 7.5 to 8.4 is also
observed changing from g-Al2O3 to d-Al2O3 .
4
5
6
7
8
9
The results presented in this paper are in good agreement
with a number of other studies where co-precipitation was
used to form the floc. Dario and Ledin7 observed cationic like
adsorption for Cd2+ on ferric hydroxides in the absence and
presence of fulvic acid. Bryce et al.37 also showed that the com-
ponent addition sequence for nickel adsorption of hydrous fer-
ric oxide in the presence of EDTA altered the type of binding
observed. For example, the addition of nickel to pre-formed
hydrous ferric oxide followed by EDTA led to a cationic
adsorption curve, whereas when nickel and EDTA were pre-
equilibrated before mixing with hydrous ferric oxide, classical
ligand-like binding was also observed. Cationic Cd2+ adsorp-
tion has also been observed in the presence of polyacrylic acid
O. Tunay and N. I. Kabdas¸ali, Water Res., 1994, 28, 2117.
¨
10 M. G. Burnett, C. Faherty, C. Hardacre, J. M. Mallon, R. M.
Ormerod and G. C. Saunders, J. Chem. Soc., Chem. Commun.,
1998, 2525.
11 M. G. Burnett, C. Hardacre, J. M. Mallon, H. J. Mawhinney and
R. M. Ormerod, Phys. Chem. Chem. Phys., 2000, 2, 1273.
12 A. R. Bowers and C. P. Huang, J. Colloid Interface Sci., 1986,
110, 575.
13 A. R. Bowers and C. P. Huang, Water Res., 1987, 21, 757.
14 B. Nowack, J. Lutzenkirchen, P. Behra and L. Sigg, Environ. Sci.
¨
Technol., 1996, 30, 2397.
15 J. H. Grimes, A. J. Huggard and S. P. Wilford, J. Inorg. Nucl.
Chem., 1963, 25, 1225.
16 R. G. Bates, in Electrometric pH Determinations, John Wiley,
New York, 1954.
17 H. Ogino, S. Kobayashi and N. Tanaka, Bull. Chem. Soc. Jpn.,
1970, 43, 97.
18 D. A. Skoog, D. M. West and F. J. Holler, Analytical Chemistry,
Saunders College Publishing, Fort Worth, 6th edn., 1992,
p. 291.
19 P. G. Daniele, C. Rigano and S. Sammartano, Anal. Chem., 1979,
72, 413.
20 V. A. Fedorov, A. M. Robov, V. P. Plekhanov, V. V. Kudruk,
M. A. Kuznechikhina and G. E. Chemikova, Russ. J. Inorg.
Chem., 1974, 19, 666.
38
on g-Al2O3 and on a-alumina based systems in the presence
of the citrate anion.39 In all cases the systems were equilibrated
for extended periods of time.
Our general model for this cationic-like binding relies on the
presence of an Al–O–M–L species on the floc. This was also
proposed by Vohra and Davis40 to explain a similar observa-
tion following the adsorption of Pb2+ on TiO2 in the presence
of nitrilotriacetic acid. In that study, cationic binding was also
observed in the presence of the chelate, which could only be
fitted using a Ti-O-Pb-NTA2ꢂ surface complex.
21 A. P. Brunetti, G. H. Nancollas and P. N. Smith, J. Am. Chem.
Soc., 1969, 91, 4680.
It may be argued that the co-precipitation results presented
in this paper depend on the preservation of an early form of
adsorption possible only during the formation of the parent
microfloc preceding aggregation but there can be no doubt
that the product is robust, reproducible and stable. A detailed
study24 into the modes of aluminium hydroxide precipitation
has revealed that the initially formed water-rich precipitate
can be stabilised by anionic adsorbates. The same work has
described the metastable phases by systems of balanced equili-
bria paralleling the approach used in this paper. As shown, the
percentage of cadmium adsorbed follows a simple equilibrium
scheme in all cases save that of PhDTA.
By adding a chelate it is possible to improve the efficiency of
the removal of heavy metals by flocculation. However, it is
only in the case of EDDS that this process may be viable for
wastewater treatment. EDDS is not only biodegradable32 but
also can be used in excess without stripping the metal from
the floc. By adding such a promoter, the concentration of
heavy metals which can be removed increases without the need
to increase the aluminium concentration.
22 T. Moeller and S.-K. Chu, J. Inorg. Nucl. Chem., 1966, 26, 153.
23 C. F. Baes and R. E. Mesmer, The Hydrolysis of Cations, John
Wiley, New York, 1976.
¨
˚
24 L.-O. Ohman and L. Wagberg, J. Pulp Paper Sci., 1997, 23, J475.
25 D. K. Nordstrom and H. M. May, The Environmental Chemistry
of Aluminium, ed. G. Sposito, Lewis Publishers, Boca Raton, FL,
1995, ch. 2, pp. 46–49.
26 A. E. Martell and R. H. Smith, Stability Constants of Complexes,
Plenum Press, New York, 1976, vol. 4, p. 9.
27 T. R. Bhat, R. R. Das and J. Shankar, Indian J. Chem, 1967, 5,
324.
28 J. Majer, V. Jokl, E. Dvorakova and M. Jurcova, Chem. Zvesti,
1968, 22, 415.
29 A. E. Martell and R. H. Smith, Stability Constants of Complexes,
Plenum Press, New York, 1976, vol. 1, p. 92.
30 A. E. Martell and R. H. Smith, Stability Constants of Complexes,
Plenum Press, New York, 1989, vol. 6, p. 44.
31 M. G. Burnett, C. Hardacre and H. J. Mawhinney, paper in pre-
paration.
32 (a) D. Schowanek, T. C. J. Feijtel, C. M. Perkins, F. A. Hartman,
T. W. Federle and R. J. Larson, Chemosphere, 1997, 34, 2375;
(b) P. C. Vandervivere, H. Saveyn, W. Verstraete, T. C. J. Feijtel
and D. Schowanek, Environ. Sci. Tech., 2001, 35, 1765.
33 R. Kummert and W. Stumm, J. Colloid Interface Sci., 1980, 75,
373.
34 P. M. Jardine and L. W. Zelazny, in The Environmental Chemistry
of Aluminium, ed. G. Sposito, Lewis Publishers, Boca Raton, FL,
1995, p. 229, ch. 6.
Acknowledgement
We thank British Nuclear Fuels plc for their support, Asso-
ciated Octel for supplying the EDDS and the Department of
Education for Northern Ireland for a Distinction Award to
H. J. M. We also thank Prof. D. R. Williams, Cardiff Univer-
sity for useful discussions especially with regard to EDDS.
´
´ ´
35 M. Szekeres, E. Tombacz, K. Ferencz and I. Dekany, Colloids
Surf., A, 1998, 141, 319.
36 A. R. Bowers and C. P. Huang, J. Colloid Interface Sci., 1985,
105, 197.
37 A. L. Bryce, W. A. Kornicker, A. W. Elzerman and S. B. Clark,
Environ. Sci. Technol., 1994, 28, 2353.
38 R. M. Floroiu, A. P. Davis and A. Torrents, Environ. Sci. Tech-
nol., 2001, 35, 348.
39 J. F. Boily and J. B. Fein, Geochim. Cosmochim. Acta, 1996, 60,
2929.
40 M. S. Vohra and A. P. Davis, J. Colloid Interface Sci., 1997, 194,
59.
References
1
W. J. Eilbeck and G. Mattock, in Chemical Processes in Waste
Water Treatment, Ellis Horwood, Chichester, 1987.
3834
Phys. Chem. Chem. Phys., 2002, 4, 3828–3834