Analytica Chimica Acta 402 (1999) 211–221 Effect of metal/fulvic acid mole ratios on the binding of Ni(II), Pb(II), Cu(II), Cd(II), and Al(III) by two well-characterized fulvic acids in aqueous model solutions Amina L.R. Sekaly a , R. Mandal a , Nouri M. Hassan a , J. Murimboh a , C.L. Chakrabarti a,∗ , M.H. Back a , D.C. Grégoire b , W.H. Schroeder c a Ottawa-Carleton Chemistry Institute, Department of Chemistry, Carleton University, 1125 Colonel By Drive, Ottawa, K1S 5B6, Ont., Canada b Geological Survey of Canada, 601 Booth Street, Ottawa, K1S 0E8, Ont., Canada c Atmospheric Environment Service, Environment Canada, 4905 Dufferin Street, Downsview M3H 5T4, Ont., Canada Received 22 March 1999; received in revised form 25 June 1999; accepted 7 July 1999 Abstract The kinetics of dissociation of Ni(II), Pb(II), Cu(II), and Cd(II)–fulvic acid (FA, Armadale), and Al(III)–FA (Suwannee River) complexes in aqueous model solutions was studied by the competing ligand exchange method (CLEM). Chelex-100 cation exchange resin was used as the competing ligand and inductively-coupled plasma mass spectrometry (ICP-MS) or graphite furnace atomic absorption spectrometry (GFAAS) was used to determine the rate of dissociation of the metal–FA complexes in the model solutions. At low metal concentrations and at very low [metal]/[FA] mole ratios, the metals formed inert complexes (dissociation rate coefficient kd ≈ 10−5 s−1 ) with the FAs. The percentage of inert complexes increased as the [metal]/[FA] mole ratio was decreased at a constant concentration of the metal. The percentage of labile metal–FA complexes (dissociation rate coefficient kd ≈ 10−2 –10−3 s−1 ) increased as the [metal]/[FA] mole ratio in the model solutions was increased. The results were similar for all the metals studied. The two well-characterized FAs showed similar behavior despite their marked differences in the binding capacity and in the source and nature of the two FAs — the Armadale FA is pedogenic, and the Suwannee River FA is aquogenic. The results have special significance for the natural environment in which potentially toxic metals are often present at trace levels. In such a situation, experiments at the environmentally relevant [metal]/[FA] mole ratios are necessary to determine the effect of the [metal]/[FA] mole ratios on dissociation of metal–FA complexes and release of metal–aquo complexes which are reported to be toxic. ©1999 Elsevier Science B.V. All rights reserved. Keywords: Fulvic acid complexes of Ni(II), Pb(II), Cu(II), Cd(II), Al(III); Armadale soil fulvic acid; Suwannee River fulvic acid; Dissociation rate coefficients 1. Introduction ∗ Corresponding author. Tel.: +1-613-520-2600 x3839; fax: +1-613-520-3749 x3830 E-mail address: chuni [email protected] (C.L. Chakrabarti) Complexation reactions of potentially toxic metals by naturally-occurring organic complexants such as humic substances, which are ubiquitous in fresh- 0003-2670/99/$ – see front matter ©1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 3 - 2 6 7 0 ( 9 9 ) 0 0 5 3 4 - 6 212 A.L.R. Sekaly et al. / Analytica Chimica Acta 402 (1999) 211–221 waters, are increasingly being recognized as important factors in many natural systems because these reactions determine, to a large extent, the metal speciation and bioavailability of the metal species [1]. Metal speciation also determines the mobility of trace metals in the natural environment. For example, complexation of metal ions onto insoluble organic compounds strongly reduces their mobility, whereas the formation of soluble metal complexes with dissolved organic compounds enhances their mobility in the natural environment [2–4]. Dissolved organic compounds may have terrestrial or aquatic origin, and may degrade or condense in the water column. Their composition varies spatially and temporally, and differences in their elemental composition and spectroscopic properties have been well-documented [4]. Humic substances, such as fulvic acids (FA) and humic acids (HA), represent a major fraction of dissolved organic compounds present in freshwaters. FAs generally have a lower molecular weight than HAs, higher oxygen content, higher carboxylic content and higher acidity [5]. Hence, FA is the more water-soluble of the two. HA is the fraction that is soluble in only alkaline solutions. Humic substances are polyfunctional (i.e. each molecule may have a large number of different complexing sites, e.g. carboxylic, phenolic, phthalic, salicylic and amine functional groups) [4,6]. They are also oligoelectrolytic [6]. The total charge on the molecule depends on the pH and environmental conditions [6]. At low pH, the functional groups are protonated and uncharged; at higher pH, the functional groups dissociate and become negatively charged. Around the charged molecules, a diffuse double layer develops [7]. The double layer screens the charge so that the effect of electrostatic interactions is decreased. In the case of a negatively charged molecule, the concentration of the positively charged metal ions is larger in the double layer than in the bulk of the solution. Consequently, the amount of background electrotrolyte ions determines the efficiency of the screening of the surface charge. At low ionic strength, the electric field around the charged molecule extends relatively far in the solution and the double layer is thick. At high ionic strength, a strong screening results in a thin double layer. The presence of ions in the electrical double layer can affect the effective charge of the charged surface of humic substances, resulting in an ionic strength dependency of metal ion binding [8]. Futhermore, the effective charge and the extent of hydrogen bonding can also affect the conformational properties of humic substances. The degree of aggregation and coiling of the various molecules of humic substances increase with ionic strength, and as a result, the accessibility of some ligand sites to larger metal ions may be reduced [6,9,10]. It has been reported that, at low pH, humic substances have tightly coiled, cross-linked (via H-bonding) conformation where metal binding sites are not readily available, but at higher pH, they have a more open conformation [11,12]. Consequently, metal ion binding to humic substances is affected by both physical and chemical heterogeneity [2,8,13,14]. The chemical equilibria involved in metal complexation reactions have been studied in great detail, and numerous thermodynamic and mathematical models have been proposed [15–19]; however, the kinetics of metal complexation reactions has received less attention. Several studies investigating the kinetics of metal ion–humic substance interactions have recently been reported [20,21]. Langford et al. [22] developed a ligand exchange probe for kinetic speciation. Shuman et al. [23] used a rotating disk electrode (RDE) in combination with anodic stripping voltammetry (ASV) to estimate the dissociation rate coefficients of metal complexes of humic substances. Olson et al. [24,25] studied the dissociation of copper complexes of estuarine humic substances. They used 4-(2-pyridylazo) resorcinol (PAR) as the competing ligand and monitored the dissociation reaction by measuring the concentration of the metal–PAR complex spectrophotometrically. Lavigne et al. [26] also used PAR in their study of the dissociation of nickel complexes of a well-characterized soil FA. Hering and Morel [27] studied the rates of ligand exchange reactions of copper–nitrilotriacetic acid complexes and copper–humate complexes. In their work, Hering and Morel [27] found that the rate coefficients of ligand exchange were pH-dependent and reported conditional rate coefficients that depended not only on the properties of the humic substance but also on the experimental conditions. The complexing sites of heterogeneous, macromolecular, organic complexants can be classified into major and minor sites types [28]. The major sites are weak binding sites and are the ones present in large A.L.R. Sekaly et al. / Analytica Chimica Acta 402 (1999) 211–221 proportion [29]. There are only a few types of major sites, including carboxylate and phenolate functional groups. The range of the free energy of formation of the complexes with the major sites does not exceed the equivalent of 1–2 log K units because the sites are chemically homogeneous. The minor sites are strong binding sites and represent a small fraction of the total sites, but they consist of a very large number of sites types. These sites form strong complexes with metals. Examples of minor sites include nitrogen and sulphur bearing functional groups. The distribution of equilibrium (formation) constants for the minor sites is larger (at least several log K units) because of not only the physical factors mentioned earlier but also of the wide variety in the chemical nature of these sites [10]. This paper forms a part of our research program for the development of a comprehensive scheme of chemical speciation for metals and metalloids in freshwaters, rain and snow samples based on kinetic approaches [30–40]. The objective of this research is to investigate the progressive occupation of strong and weak binding sites of two well-characterized FAs by selected metals. This paper shows how kinetic data lead to differentiation between strong and weak binding sites in FA. The metals selected for the study were Ni(II), Pb(II), Cu(II), Cd(II) and Al(III). 2. Theory 2.1. CLEM The method used for the determination of dissociation kinetics of metal complexes was the CLEM, in which Chelex-100 cation exchange resin was used as a solid-phase competing ligand. The kinetic model proposed by Olson et al. [24,25] was adopted [36] to study the dissociation kinetics of metal complexes, MLi , where M is a metal ion and Li is the ith binding site on a polyfunctional, macromolecular, organic complexant, such as FA. For simplicity, the charges on M and Li have been omitted. Consider an aqueous mixture of n components in which each component, designated MLi , undergoes a first-order or pseudo-first-order reaction and exists in equilibrium with its dissociation products: k1 MLi M + Li (slow) k−1 213 (1) where k1 and k−1 are the rate coefficients for the forward and the backward reaction, respectively. The Chelex-100 reacts with M as follows: k2 Chelex + M Chelex − M (fast) k−2 (2) where Chelex represents Chelex-100 cation exchange resin, and k2 and k−2 are first-order (or pseudo-first-order) rate coefficients for the forward and the backward reaction, respectively. The model is based on two assumptions: (i) that reaction (2) is much faster than reaction (1); and (ii) that [Chelex] [M]. As [Chelex] is large and [Chelex] [M], [Chelex] can be considered as constant and reaction (2) as pseudo-first-order. Since k2 is large, as has been determined by ICP-MS, and with Chelex added in large excess, the condition k2 [Chelex] k−1 [Li ] holds and the overall reaction Chelex + MLi → Chelex − M + Li (3) is assumed to be irreversible, and the rate of dissociation of MLi is d[MLi ] d[M − Chelex] =− = k1 [MLi ] dt dt (4) The two assumptions made above have been verified as follows: assumption no. 1, by our previous studies using ICP-MS, and assumption no. 2, by taking a large excess of Chelex-100 resin over the very low concentrations of metal ions present in the test samples. If each complex, MLi , undergoes, independently and simultaneously, a first-order or pseudo-first-order dissociation reaction, the sum, C(t), of the concentrations of all components remaining undissociated in the test solution at time t can be described as n X Ci0 exp(−ki t) C(t) = (5) i=1 where Ci0 is the initial concentration of MLi , the ith component. C(t) is determined by ICP-MS or by GFAAS, by measuring the decrease in the amount of metal ions in the test solution due to their binding by the Chelex-100, as a function of time [32]. 214 A.L.R. Sekaly et al. / Analytica Chimica Acta 402 (1999) 211–221 3. Experimental 3.1. Chemicals and reagents Stock solutions containing (1000 g ml−1 ) of Pb(II), Cu(II), Cd(II) and Ni(II) (ICP-MS-2 high-purity standard Delta Scientific) were used to prepare the model solutions. The Al(III) stock solution (1000 g ml−1 ) was prepared by dissolving an appropriate amount of aluminium metal (SPEX 99.99%) in ultrapure nitric acid (ULTREX II, J.T. Baker Inc., Phillipsburg, NJ, USA) with heating, and diluting to the appropriate volume with ultrapure water. Ultrapure water having a resistivity of 18.2 M cm was used to prepare all standards and test solutions. The ultrapure water was obtained from a Milli-Q-Plus water purification system (Millipore Corporation). Analytical grade (minimum 99% pure), Chelex-100 (100–200 mesh), a styrene-divinylbenzene co-polymer with iminodiacetate functional groups, was supplied by Bio-Rad Laboratories. The Chelex-100 cation-exchange resin was pre-treated for this work as follows: it was equilibrated with a sodium acetate–acetic acid (NaOAc–HOAc) buffer solution (pH 5.0 ± 0.1) by soaking it in the buffer solution for 24 h. The quantity of cations exchanged by the Chelex resin is a function of the pH and is very small at any pH below 2.0; the quantity is a maximum at pH 7.4. Dryness of the Chelex-100 affects its properties and changes the pH of the test solution during kinetic runs. Hence, after the above pre-treatment, the Chelex resin was kept in the buffer solution till it was used. The pH of the test solutions was adjusted to pH 5.0 ± 0.5 using 1% NaOH. The sodium hydroxide solution was purified by electrodeposition at –1.2 V (versus Esce ) for at least 48 h immediately prior to use. All the standard solutions were acidified to make them 1% (v/v) using HNO3 (ULTREX II, J.T. Baker Inc., Phillipsburg, NJ, USA). Standard buffers (supplied by Fisher Scientific) were used for the calibration of the pH-meter. 3.2. FAs used as complexants Two well-characterized FAs were used as heterogeneous, macromolecular, organic complexants in this study. The FAs used in this study were Armadale soil FA supplied by Dr. D.S. Gamble of Agriculture Canada, Ottawa, Canada [41–45], and the other was the Standard Suwannee River aquatic FA obtained from the International Humic Substance Society (IHHS) [46]. The total amounts of phenolic OH and carboxyl groups determined by potentiometric titration of the Armadale FA were 3 mmol g−1 [41,42] and 7.71 mmol g−1 [41,43,44], respectively. Hence, the bidendetate complexing capacity of the Armadale soil FA was estimated to be approximately 5.4 mmol g−1 [45]. The complexing capacity for Al(III) of the Standard Suwannee River FA was reported to be 390 mol g−1 at pH 4.7 [47]. An estimate of the metal background concentration in the Armadale FA was obtained by us by using ICP-MS. The concentrations of all detectable trace metals were estimated semi-quantitatively by using a single-standard analytical calibration curve. The background concentrations of the metals of interest were found to be Ni(II) 5.2 g g−1 , Cu(II) 18.4 g g−1 , Pb(II) 6.5 g g−1 , Al(III) 524.8 g g−1 ; Cd(II) was not detectable. Taylor and Garbarino reported the occurrence and distribution of trace metals in the IHSS’s FA isolated from Suwannee River [48]. The concentration of Al(III) was estimated semi-quantitatively to be 30 g g−1 , by using ICP-MS and a single-standard analytical calibration curve. Isotope-dilution mass spectrometry was used for accurate determination of the concentration of Ni(II), Cu(II), Pb(II) and Cd(II) and some other elements. The concentrations of Ni(II), Cu(II), Pb(II) and Cd(II) in the Standard Suwannee River FA were reported to be 0.92 ± 0.06, 1.7 ± 0.08, 1.2 ± 0.06, and 0.06 ± 0.8 g g−1 , respectively. 3.3. Instrumentation The kinetics of decrease in the amounts of nickel, lead, cadmium, copper and aluminium in the model solutions due to binding by Chelex-100 as a function of time was determined by ICP-MS, and by GFAAS. The ICP-MS used was a Perkin–Elmer SCIEX Elan 5000, fitted with a pneumatic nebulizer as the sample introduction system. The instrument operating conditions and the data acquisition protocol are described in Table 1. The graphite furnace atomic absorption A.L.R. Sekaly et al. / Analytica Chimica Acta 402 (1999) 211–221 Table 1 Instrumental operating conditions and data acquisition protocol for the ICP-MS Inductively-coupled plasma RF power Coolant argon flow rate Carrier argon flow rate Auxiliary argon flow rate 1 kW 15 l min−1 0.90 l min−1 0.85 l min−1 Data acquisition protocol Dwell time Scan mode Signal measurement Points/spectral peak Resolution 100–1000 min peak hop counts per second 1 normal 215 ratios of [metal]/[FA] of 0.0018, 0.018, 0.18. It should be noted that, for naturally occurring heterogeneous complexants, the molar mass is not directly accessibile [4,10]. Nevertheless, a molar concentration has been calculated, based on the bidentate complexing capacity, mmol g−1 , of FA, as determined by Gamble et al. [45] for the Armadale FA, and by Hawke et al. [47] for the Suwannee River FA, respectively. The pH of the test solutions was then adjusted to 5.0. 3.6. Analytical procedure spectrometer used was Perkin–Elmer model 5000, fitted with a Zeeman-Background Correction System. Table 2 shows the instrumental parameters used in the GFAAS determination of the metals Ni(II), Pb(II), Cu(II), Cd(II) and Al(III). For the determination of the metal concentrations, 10 l of the solution was injected into the graphite furnace, where it was dried, ashed and atomized. The signal was measured in the peak area mode. An Accumet 20 pH/mV/conductivity meter (Fisher Scientific), fitted with an Accuplast combination glass electrode, was used for pH measurements. 3.4. Cleaning procedures for containers All containers used were made of Teflon, and they were cleaned as follows. First, they were completely filled with 10% HNO3 (AR grade) and allowed to stand at room temperature for 1 week. Then, they were rinsed with ultrapure water (having a conductivity of 18.2 m cm), and then filled with ultrapure water and allowed to stand until they were used; the filling water was renewed periodically to ensure purity of the filling water. 3.5. Samples Stock solutions of FAs (1.000 ± 0.001 g l−1 ) were prepared from the solid FAs. The model solutions were prepared by adding known amounts of the FA to a known volume of the metal test solutions to give mole The kinetics of decrease in the amounts of nickel, lead, cadmium, copper and aluminium in the model solutions due to binding by Chelex-100 as a function of time was determined by ICP-MS, and by GFAAS. One percent (w/v) of Chelex-100 resin was added to the test solution placed in a Teflon Reactor (cylindrical, 500 ml capacity) and was stirred continuously with a Teflon-coated stirring bar. Data acquisition was initiated from the instant the Chelex-100 resin was added to the Teflon Reactor. For ICP-MS, the test solution was delivered continuously to the plasma torch through the solution nebulizer using a peristaltic pump at a flow rate of 1 ml min−1 . The interval between the data points and the total time for data acquisition was set for each experiment based on the rate of decrease in the amount of metal ions due to binding by the Chelex-100 resin. In the case of rapid decrease in the amount of metal ions, both the time interval and the total time for data acquisition were short. Longer times for data acquisition were required for slowly dissociating metal–FA complexes, with corresponding longer time intervals. 3.7. Data treatment The experimental data were analyzed for discrete values of the dissociation rate coefficients by a method that was based on non-linear regression analysis. This method analyzes the data assuming the decrease in the amount of the metal, which represents dissociation of the complexes as a sum of exponential terms. All the experimental data were used in the fitting. 216 A.L.R. Sekaly et al. / Analytica Chimica Acta 402 (1999) 211–221 Table 2 Analysis lines and instrumental conditions for Zeeman GFAAS Element Wavelength (nm) Pyrolysis temperature (◦ C) Atomization temperature (◦ C) Atomization time (s) Ni Pb Cu Cd Al 232.2 217.1 325.0 228.5 217.1 1000 700 900 250 700 2300 2300 2500 2100 2300 6 6 6 6 6 4. Results and discussion 4.1. Effect of the [metal]/[FA] mole ratio on the rate of dissociation of FA complexes of Ni(II), Pb(II), Cu(II) and Cd(II) The Armadale FA was used for all studies of Ni(II), Pb(II), Cu(II) and Cd(II). Figs. 1–4 present the effects of the [metal]/[FA] mole ratios on the rates of dissociation of the FA complexes of Ni(II), Pb(II), Cu(II) and Cd(II). The [metal]/[FA] mole ratio was varied by keeping the metal concentration constant (∼10−8 M) and varying the FA concentration. For the sake of simplicity of calculating the [metal]/[FA] mole ratio, the FA will be considered to have ∼1% strong binding sites, and ∼99% weak binding sites. If the FA had ∼1% strong binding sites, then the [metal]/[FA] mole ratio required for the saturation of strong sites of the FA was ∼0.01. To be on the safe side (i.e. to be sure of the saturation of strong sites), we started with a [metal]/[FA] mole ratio of 0.001. The kinetic data were obtained using the CLEM Chelex-100 cation exchange resin in the batch mode. The kinetics of dissociation of the metal–FA complexes was measured by ICP-MS and GFAAS for quantitative determination of the metals. In Figs. 1–4, the top curves represent the [metal]/[FA] mole ratio of 0.0018. At this mole ratio, the rate of dissociation of the FA complexes is very slow and hardly detectable over the entire period of the measurement. All of the metal complexes are bound to strong binding sites and the complexes may be described as inert. When the [metal]/[FA] mole ratio is increased to 0.018, the rate of dissociation is initially rapid, but becomes slow again at longer times. At the highest mole ratio of [metal]/[FA] of 0.18, the initial rapid decrease is more evident, but again, a leveling off to a slow rate of dissociation at longer times Fig. 1. Nickel remaining bound to Armadale FA as a function of time in model solutions, measured by ICP-MS. pH = 5.0; ionic strength ∼0. Chelex-100 cation exchange resin was the competing ligand. Concentration of nickel was held constant and equal for all curves. 䊐 : [Ni(II) 4.3 × 10−8 M]/[FA 2.3 × 10−5 M] = 0.0018; 䊊: [Ni(II) 4.3 × 10−8 M]/[FA 2.3 × 10−7 M] = 0.18. was observed. These curves are in conformity with the postulation that as the [metal]/[FA] mole ratio is increased, the strong binding sites become saturated and the excess metal then binds to the weaker sites, forming weak complexes which are labile. This observation was common for all the metals studied. Results of the dissociation rate coefficients of the metal–FA complexes are presented in Table 3. For the lowest mole ratio of 0.0018, the rate of metal–complex dissociation was very slow; hence, only an estimate of the rate coefficient was possible (kd < 10−5 s−1 ). The kinetic data at the mole ratios 0.018 and 0.18 were analyzed using the method described above. The fraction of labile metal species (Figs. 1–4, the second and the third curve from the top) increase with increasing [metal]/[FA] mole ratios. Two rates of dissociation appear to be involved for the [metal]/[FA] mole ratios of A.L.R. Sekaly et al. / Analytica Chimica Acta 402 (1999) 211–221 Fig. 2. Lead remaining bound to Armadale FA as a function of time in model solutions, measured by ICP-MS. pH 5.2; ionic strength ∼0. Chelex-100 cation exchange resin was the competing ligand. Concentration of lead was held constant and equal for all curves. 䊊: [Pb(II) 4.3 × 10−8 M]/[FA 2.3 × 10−5 M] = 0.0018; 5: [Pb(II) 4.3 × 10−8 M]/[FA 2.3 × 10−6 M] = 0.018; 䊐: [Pb(II) 4.3 × 10−8 M]/[FA 2.3 × 10−7 M] = 0.18. Fig. 3. Copper remaining bound to Armadale FA as a function of time in model solutions, measured by Zeeman GFAAS. pH = 5.1; ionic strength ∼0. Chelex-100 cation exchange resin was the competing ligand. Concentration of copper was held constant and equal for all curves. 䊊: [Cu (II) 4.8 × 10−8 M]/[FA 2.6 × 10−5 M] = 0.0018; 䊐: [Cu(II) 4.8 × 10−8 M]/[FA 2.6 × 10−6 M] = 0.018; 4: [Cu(II) 4.8 × 10−8 M]/[FA 2.6 × 10−7 M] = 0.18. 0.018 and 0.18. In addition, the fraction of the metal bound to the strong sites decreases, and the dissociation rate coefficients of the strongly-bound complexes increase with increasing [metal]/[FA] mole ratios. This 217 Fig. 4. Cadmium remaining bound to Armadale FA as a function of time in model solutions, measured by Zeeman GFAAS. pH 5.2, ionic strength ∼0. Chelex-100 cation exchange resin was the competing ligand. Concentration of cadmium was held constant and equal for all curves. 䊊: [Cd(II) 3.1 × 10−8 M]/[FA 1.7 × 10−5 M] = 0.0018; 䊐: [Cd(II) 3.1 × 10−8 M]/[FA 1.7 × 10−6 M] = 0.018; 4: [Cd(II) 3.1 × 10−8 M]/[FA 1.7 × 10−7 M] = 0.18. is in conformity with the postulation that the strong binding sites are occupied first by the metal ions. After the strong binding sites are filled completely by the metal, the remaining metal binds to the weak sites, forming weak complexes that are labile. The rate of dissociation of Ni–FA complexes was measured by both ICP-MS and GFAAS, and the rate coefficients for dissociation of the strongly-bound complexes were in reasonable agreement (Table 3). The relatively poor agreement between the dissociation rate coefficients given by GFAAS and ICP-MS for the weakly bound complexes was probably due to the much shorter time scale of measurement by ICP-MS. This shorter time scale of measurement by ICP-MS allowed much better resolution of the quickly dissociating complexes. The measurement by GFAAS takes a longer time, and hence, GFAAS cannot measure the rapidly dissociating metal complexes which can be measured by ICP-MS. Hence, the dissociation rate coefficients of the first labile complexes observed by GFAAS probably represent a weighted average dissociation rate coefficients of the labile complexes and of the moderately labile complexes observed by ICP-MS. Only one set of values (those of Ni(II) complexes) by ICP-MS have been presented in Table 3 to show that these values are similar to those obtained by GFAAS. 218 A.L.R. Sekaly et al. / Analytica Chimica Acta 402 (1999) 211–221 Table 3 Effect of the [metal]/[FA] mole ratio on the dissociation rate coefficients of metal–FA complexes in aqueous model solutions. pH = 5.0 ± 0.5, ionic strength ∼0 [Metal]/[FA] mole ratio Kinetically distinguishable components C1 (%) k1 × 104 (s−1 ) C2 (%) k2 × 105 (s−1 ) Complexes with Ni(II); [Ni(II)] = 4.3 × 10−8 M, measured by ICP-MS 0.0018 – – 0.018 18 ± 1 8.7 ± 0.5 0.18 49 ± 2 10 ± 5 98 ± 2 82 ± 1 51 ± 1 0.6 ± 0.1 3.0 ± 0.1 18 ± 1 Complexes with Ni(II); [Ni(II)] = 5.2 × 10−8 M, measured by GFAAS 0.0018 – – 0.018 15 ± 1 0.4 ± 0.1 0.18 40 ± 2 10.2 ± 0.3 100 ± 2 85 ± 1 60 ± 1 0.2 ± 0.1 0.6 ± 0.5 23 ± 1 Complexes with Pb(II); [Pb(II)] = 2.4 × 10−8 M, measured by GFAAS 0.0018 – – 0.018 7.8 ± 0.2 0.2 ± 0.1 0.18 18 ± 0.5 8.5 ± 0.5 100 ± 2 92 ± 2 82 ± 1 0.2 ± 0.1 0.4 ± 0.5 20 ± 1 Complexes with Cu(II); [Cu(II)] = 4.8 × 10−8 M, measured by GFAAS 0.0018 – 0.018 9±2 0.2 ± 0.1 0.18 20 ± 1 9.0 ± 0.3 100 ± 1 91 ± 1 80 ± 1 inerta 0.5 ± 0.1 21 ± 1 Complexes with Cd(II); [Cd(II)] = 3.1 × 10−8 M, measured by GFAAS 0.0018 – – 0.018 9±2 6.0 ± 0.3 0.18 30 ± 1 9.5 ± 0.3 99 ± 1 91 ± 1 70 ± 1 0.3 ± 0.1 1.1 ± 0.1 22 ± 1 Inert k2 ≤ 10−5 s−1 . k1 , k2 are the dissociation rate coefficients of the first (the faster), and the second (the slower) component, respectively. Values after ± signs are standard deviations of non-linear regression analysis. a Table 3 shows that a minimum of two components were required to describe the dissociation kinetics of the metal–FA complexes at the [metal]/[FA] mole ratios of 0.018 and 0.18 for the four different metals. This suggests the presence of two distinct classes of binding sites in the Armadale FA. Sunda and Hanson, using an initial metal concentration of 10−8 M, reported, for the first time, the existence of the strong sites of humic substances, whereas the earlier workers had failed to detect their existence because they had used at least one order of magnitude higher concentrations of the metals in their titration experiments [49,50]. In Table 3, the rate coefficients (measured by GFAAS), either for the strong complexes or for the weak complexes, do not show much difference between the four metals, Ni(II), Pb(II), Cu(II) and Cd(II). Such a lack of difference in the rate coefficients raises the serious question as to whether the rate coefficients for the water exchange (k−w ) of the different metals are indeed the determinant of the strength of the metal’s affinity for a particular binding site. Since the k−w for Pb(II) and Cu(II) is larger than the k−w for Ni(II) by four orders of magnitude ([51], p. 400), similar difference in the rate coefficients for dissociation between the Pb(II)–FA complex, the Cu(II)–FA complex and the Ni(II)–FA complex should be observed if the k−w is indeed the determinant of the complex’s stability. However, no significant difference in the rate coefficients was observed (Table 3), particularly in the values obtained by GFAAS. In predicting the stability of the metal complex in Table 3, the following other factors that also contribute to the stability of metal–FA complexes should be considered: effective nuclear charge, polarizability of the metal ions and the ligand, the ligand field stabilization energy (LFSE) for the transition metals, Ni(II), Cu(II), and Cd(II), and the polyelectrolyte effect (electrostatic interactions between the charges on the FA and those on the cations and other ions). The polyelectrolyte effect is the same for all the above metal ions and can, therefore, be left out of consideration A.L.R. Sekaly et al. / Analytica Chimica Acta 402 (1999) 211–221 as a factor responsible for the above-noted lack of difference in the rate coefficients for the dissociation of the metal–FA complexes. The above factors do not apply to Pb(II), which is the only non-transition metal considered, and for which the kw is the only determinant of the Pb(II)–FA complex’s stability. For d8 (Ni2+ ) configuration, it seems that the low value for the water exchange rate coefficient (3 × 104 s−1 ), caused by its highly unfavorable change in LFSE during the complex formation ([51], p. 398) is more than compensated for by the other factors mentioned above (effective nuclear charge, polarizability of the d8 (Ni2+ ) ion and of the FA ligand), resulting in the formation of an Ni(II)–FA complex of a stability similar to that of the other three metal–FA complexes, and hence, in a relatively inert Ni(II)–FA complex. The rate coefficients for both strong and weak complexes show a significant increase at the highest mole ratio of [metal]/[FA] listed in Table 3. The binding sites in FA are known to involve a range of binding energies [52]. Even among the strong and weak binding sites, as revealed by the present experiments, a distribution of sites stronger or weaker than the experimentally measured averages really exists even though Table 3 presents the average values. As more metal becomes available for binding, a broader range of binding sites becomes occupied and the rate coefficients for dissociation should show a corresponding distribution. Caution should be exercised when considering such dissociation rate coefficients as they are derived from a minimum set of parameters to fit the decay curves within the experimental error. There may well be a more complicated distribution of dissociation rate coefficients than the minimum set [53]. No account was taken of the naturally-occurring metals in the FAs used, either in computing the experimental results or in their interpretation. 4.2. Effect of the [metal]/[FA] mole ratio on the rate of dissociation of Al(III)–FA complexes The effect of [Al(III)]/[FA] mole ratio on the lability of the Al–FA complexes is presented in Fig. 5. For aluminium, the Standard Suwannee River FA was employed using a fixed concentration of Al(III) [7.4 × 10−7 M]. In order to ascertain whether there would be any precipitation of Al(OH)3 at pH 5 which 219 Fig. 5. Aluminium remaining bound to Standard Suwannee River FA as a function of time in model solutions, measured by ICP-MS. pH 5.2; ionic strength ∼0. Chelex-100 cation exchange resin was the competing ligand. Concentration of aluminium was held constant for all curves. 4: [Al(III) 7.1 × 10−7 M]/[FA 4.1 × 10−5 M] = 0.018; 䊐: [Al(III) 7.1 × 10−7 M]/[FA 4.1 × 10−6 M] = 0.18; 䊊: [Al(III) 7.1 × 10−7 M]/[FA 4.1 × 10−7 M] = 1.8. would affect the kinetic experiments, we carried out equilibrium calculations using the MINEQL+ computer equilibrium program and found that there was no significant formation of Al(OH)3 at pH 5 at the Al (III) concentration used in the kinetic experiments. The [Al(III)]/[FA] mole ratio was calculated based on the aluminium complexation capacity for the Standard Suwannee River FA, which had been reported to be 390 mol g−1 [47]. The Al(III)–FA model solutions were studied by both ICP-MS and GFAAS, using the three mole ratios of [Al(III)]/[FA], 0.018, 0.18 and 1.8, for the experiments performed by ICP-MS, and the two mole ratios of 0.0033, and 0.33 for the experiments performed by GFAAS. The decay of Al(III) for the three ratios, as measured by ICP-MS, is presented in Fig. 5 and the data analysis is given in Table 4. At the [Al(III)]/[FA] mole ratio of 0.018, virtually all of the Al(III)–FA complexes were inert (Fig. 5, top curve), suggesting that only the strong binding sites of the FA were occupied by Al(III), forming strong complexes that are inert. The dissociation rate coefficients of the strongly-bound Al(III)–FA complexes was kd ∼10−5 s−1 , which was comparable to the dissociation rate coefficient of the strongly-bound Ni(II)–FA complexes measured by ICP-MS and GFAAS. The fraction of the labile Al-complex increased with 220 A.L.R. Sekaly et al. / Analytica Chimica Acta 402 (1999) 211–221 Table 4 Effect of the [metal]/[FA] mole ratio on the dissociation rate coefficient of Al(III)–FA complexes in model solutions. pH = 5.0 ± 0.5, ionic strength ∼0 [Metal]/[FA] mole ratio FA (M) Kinetically distinguishable components k1 × 104 (s−1 ) C2 (%) k2 × 105 (s−1 ) Complexes with Al(III); [Al (III)] = 7.1 × 10−7 M, measured by ICP-MS 0.018 4.1 × 10−5 – 0.18 4.1 × 10−6 52 ± 1 1.8 4.1 × 10−7 78 ± 1 – 5.1 ± 0.5 11 ± 1 97 ± 3 49 ± 1 32 ± 1 inerta 5.9 ± 0.7 8.9 ± 0.5 Complexes with Al(III); [Al (III)] = 7.1 × 10−7 M, measured by GFAAS – 0.0033 5.6 × 10−6 0.33 5.6 × 10−8 42 ± 4 – 27 ± 1 93 ± 7 59 ± 3 inerta 1.3 ± 0.1 C1 (%) Inert k2 ≤ 10−5 s−1 . k1 and k2 are the dissociation rate coefficients of the first (the faster), and the second (the slower) component, respectively. Values after ± signs stand are standard deviations of non-linear regression analysis. a the increasing [Al(III)]/[FA] mole ratio. In addition, the fraction of Al(III) bound to the strong sites decreased and the dissociation rate coefficients of the strongly-bound Al(III)–FA complexes increased with increasing [Al(III)]/[FA] mole ratios. These trends were observed in both GFAAS and ICP-MS results, which suggest the presence of two distinct classes of binding sites. The strong binding sites are occupied first, and the weak sites become occupied only after the strong sites are occupied fully. It seems that the Suwannee River FA and the Armadale FA have similar metal binding characteristics. [2] [3] [4] [5] [6] [7] [8] [9] Acknowledgements [10] The authors are grateful to the Nickel Producers Environmental Research Association, USA, International Nickel Company Ltd. Canada, and Falconbridge Nickel Ltd. Canada, and to the Natural Sciences and Engineering Research Council of Canada, for research contracts and research grants. The authors are grateful to Dr. D.S. Gamble for supplying the Armadale FA. One of the authors, Nouri M. Hassan, is also grateful to the Government of Libya for providing him with a graduate scholarship for the Ph.D. degree program at Carleton University. [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] References [1] W.J. Langston, G.W. Bryan, in: C.J.M. Kramer, J.C. Duinker (Eds.), Complexation of Trace Metals in Natural Waters, [21] [22] Martinus Nijhoff/Dr. W. Junk Publishers, The Hague, The Netherlands, 1984, p. 375. J. Buffle, in: H. Sigel (Ed.), Metal Ions in Biological Systems, vol. 18, Marcel Dekker, New York, 1984, Chapter 6. G. Sposito, CRC Crit. Rev. Environ. Control 16 (1986) 193. J. Buffle, Complexation Reactions in Aquatic Systems: an Analytical Approach, Ellis Horwood, Chichester, 1998, p. 195. R. Francois, Rev. Aquat. Sci. 3 (1990) 41. M. Filella, J. Buffle, H.P. van Leeuwen, Anal. Chim. Acta 232 (1990) 209. C. Tanford, Physical Chemistry of Macromolecules, Wiley Interscience, New York, 1961. J.C.M. De Wit, W.H. Van Riemsdijk, L.K. Koopal, Environ. Sci. Technol. 27 (1993) 2005. G.D. Templeton III, N.D. Chasteen, Geochim. Cosmochim. Acta 44 (1980) 741. J. Buffle, R.S. Altmann, M. Filella, Anal. Chim. Acta 232 (1990) 225. N. Senesi, Anal. Chim. Acta 232 (1990) 77. S.A. Wood, Ore Geol. Rev. 11 (1996) 1. J.C.M. De Wit, M.M. Nederlof, W.H. Van Riemsdijk, L.K. Koopal, Water, Air, and Soil Pollution 57-58 (1991) 339. Johannes C.M. De Wit, W.H. Van Riemsdijk, L.K. Koopal, Environ. Sci. Technol. 27 (1993) 2015. I. Ružić, Marine Chem. 53 (1996) 1. M.M. Nederlof, H.W. van Riemsdijk, L.K. Koopal, Environ. Sci. Technol. 28 (1994) 1048. M.M. Nederlof, H.W. van Riemsdijk, L.K. Koopal, Environ. Sci. Technol. 28 (1994) 1037. C.H. Langford, R.L. Cook, Analyst 120 (1995) 591. P. Warwick, A. Hall, S.J. King, J. Zhu, J. Van der Lee, Radiochim. Acta 81 (1998) 215. A.W. Rate, R. McLaren, R.S. Switt, Environ. Sci. Technol. 26 (1992) 2477. A.W. Rate, R. McLaren, R.S. Switt, Environ. Sci. Technol. 27 (1993) 1408. C.H. Langford, W.H. Gutzman, Anal. Chim. Acta 256 (1992) 183. A.L.R. Sekaly et al. / Analytica Chimica Acta 402 (1999) 211–221 [23] M.S. Shuman, B.J. Collins, P.J. Fitzgerald, D.J. Olson, in: Aquatic and Terrestrial Humic Material, Ann Arbor Press, Ann Arbor, MI, 1983, Chapter 17, pp. 349–370. [24] D.L. Olson, M.S. Shuman, Anal. Chem. 55 (1983) 1103. [25] D.L. Olson, M.S. Shuman, Geochim. Cosmochim. Acta 49 (1985) 1371. [26] J.A. Lavigne, C.H. Langford, M.K.S. Mak, Anal. Chem. 59 (1987) 2616. [27] J.G. Hering, F.M.M. Morel, Environ. Sci. Technol. 24 (1990) 242. [28] J. Buffle, R.S. Altmann, M. Filella, A. Tessier, Geochim. Cosmochim. Acta 54 (1990) 1535. [29] J. Buffle, M. Filella, Anal. Chim. Acta 313 (1995) 144. [30] A.L.R. Sekaly, M.H. Back, C.L. Chakrabarti, D.C. Grégoire, J.Y. Lu, W.H. Schroeder, Spectrochim. Acta, Part B 53 (1998) 837. [31] A.L.R. Sekaly, M.H. Back, C.L. Chakrabarti, D.C. Grégoire, J.Y. Lu, W.H. Schroeder, Spectrochim. Acta, Part B 53 (1998) 847. [32] R. Mandal, A.L.R. Sekaly, J. Murimboh, N.M. Hassan, C.L. Chakrabarti, M.H. Back, Anal. Chim. Acta, in press. [33] R. Mandal, A.L.R. Sekaly, J. Murimboh, N.M. Hassan, C.L. Chakrabarti, M.H. Back, Anal. Chim. Acta, in press. [34] Y. Lu, C.L. Chakrabarti, M.H. Back, A.L.R. Sekaly, W.H. Schroeder, J. Anal. Atom. Spec. 11 (1996) 1189. [35] C.L. Chakrabarti, Y. Lu, D.C. Grégoire, M.H. Back, W.H. Schroeder, Environ. Sci. Technol. 28 (1994) 1957. [36] Y. Lu, C.L. Chakrabarti, M.H. Back, D.C. Grégoire, W.H. Schroeder, A.G. Szabo, L. Bramal, Anal. Chim. Acta 288 (1994) 131. [37] Y. Lu, C.L. Chakrabarti, M.H. Back, D.C. Grégoire, W.H. Schroeder, Int. J. Environ. Anal. Chem. 60 (1995) 313. [38] C.L. Chakrabarti, Y. Lu, D.C. Grégoire, M.H. Back, W.H. Schroeder, Anal. Chim. Acta 28 (1994) 1957. [39] Y. Lu, C.L. Chakrabarti, D.C. Grégoire, M.H. Back, W.H. Schroeder, Anal. Chim. Acta 293 (1994) 95. 221 [40] M.T. Lam, C.L. Chakrabarti, J. Cheng, V. Pavski, Electroanalysis 9 (1997) 1018. [41] D.S. Gamble, M. Schnitzer, in: P.C. Singer (Ed.), Trace Metals and Metal–Organic Interactions in Natural Waters, Ann Arbor Science Publishers, Ann Arbor, MI, 1973, Chapter 9, pp. 266–273. [42] M. Schnitzer, S.U. Khan, Humic Substances in the Environment, Marcel Dekker, New York, 1972. [43] D.S. Gamble, Can. J. Chem. 48 (1970) 2662. [44] D.S. Gamble, Can. J. Chem. 50 (1972) 2680. [45] D.S. Gamble, A.W. Underdown, C.H. Langford, Anal. Chem. 52 (1980) 1901. [46] R.C. Averet, J.A. Leenheer, D.M. McKnight, K.A. Thorn (Eds.), Humic Substances in the Suwannee River, Georgia: Interactions, Properties, and Proposed Structures, US Geological Survey, Water-Supply Paper 2373, 1994. [47] D.J. Hawke, K.J. Powell, J.E. Gregor, Marine Freshwater Res. 47 (1996) 11. [48] H.E. Taylor, J. Garbarino, in: R.C. Averet, J.A. Leenheer, D.M. McKnight, K.A. Thorn (Eds.), Humic Substances in the Suwannee River, Georgia: Interactions, Properties, and Proposed Structures, US Geological Survey, Water-Supply Paper 2373, 1994, Chapter E. [49] W.G. Sunda, P.J. Hanson, in: E.A. Jenne (Ed.), Chemical Modelling in Aqueous Systems, ACS Symposium Series 93, American Chemical Society, Washington, DC, 1979. [50] F.M.M. Morel, Principles of Aquatic Chemistry, Wiley, New York, 1983, p. 290. [51] F.M.M. Morel, J.G. Hering, Principles and Applications of Aquatic Chemistry, Wiley, New York, 1993, p. 400. [52] J. Buffle, R. Scott Altmann, in: W. Stumm (Ed.), Chemical Processes at the Particle–Water Interface, Wiley, New York, 1987. [53] R.L. Cook, C.H. Langford, Anal. Chem. 34 (1995) 174.
© Copyright 2026 Paperzz