Partitioning of Humic Acids between Aqueous Solution and Hydrogel

Jan 12, 2015 - The hydrogel/water partitioning of the various species in the cadmium(II)/soil humic acid (HA) system is studied for two types of gel, ...
1 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/Langmuir

Partitioning of Humic Acids between Aqueous Solution and Hydrogel. 3. Microelectrodic Dynamic Speciation Analysis of Free and Bound Humic Metal Complexes in the Gel Phase Kamuran Yasadi,† Jose Paulo Pinheiro,‡ Katarzyna Zielińska,†,∥ Raewyn M. Town,*,§ and Herman P. van Leeuwen† †

Laboratory of Physical Chemistry and Colloid Science, Wageningen University, Dreijenplein 6, 6703 HB Wageningen, The Netherlands ‡ Laboratoire Interdisciplinaire des Environnements Continentaux, UMR 7360 CNRS - Université de Lorraine, 15 avenue du Charmois, 54500 Vandoeuvre-les-Nancy, France § Department of Physics, Chemistry and Pharmacy, University of Southern Denmark, Campusvej 55, 5230 Odense, Denmark ABSTRACT: The hydrogel/water partitioning of the various species in the cadmium(II)/soil humic acid (HA) system is studied for two types of gel, using in situ microelectrodic voltammetry. Under the conditions of this work, with HA particles of ca. 25 and 125 nm radius, the CdHA complex is shown to be close to nonlabile toward a 12.5 μm radius microelectrode. This implies that its kinetic contribution to Cd2+ reduction at the medium/ microelectrode interface is practically negligible. The polyacrylamide (PAAm) gels equilibrate with the aqueous medium under significant sorption of HA at the gel backbone/gel medium interface, which in turn leads to induced sorption of Cd(II) in the form of immobilized gel-bound CdHA. The rather high total Cd content of the PAAm gel suggests that the binding of Cd2+ by the hydrophobically gel-bound HA is stronger than that for dispersed HA particles. Still, the intraparticulate speciation of Cd(II) over Cd2+ and CdHA corresponds to an intrinsic stability constant similar to that for simple monocarboxylate ligands such as acetate. Alginate gels are negatively charged, and their free [Cd2+ aq ] is higher than that in the medium by the corresponding Donnan coefficient. On top of that, Cd2+ is specifically sorbed by the gel backbone/gel medium interface to reach accumulation factors as high as a few tens. HA and CdHA accumulate in the outer 20 μm film of gel at the gel/water interface of both gels, but they do not penetrate into the bulk of the alginate gel. Overall, the gel/water interface dictates drastic changes in the speciation of Cd/HA as compared to the aqueous medium, with distinct features for each individual type of gel. The results have broad significance, for example, for predictions of reactivity and bioavailability of metal species which inherently involve partitioning and diffusion into diverse gel layers such as biointerfacial cell walls, biofilm matrices, and mucous membranes.

■

the surface and the HA particles.1 Quite recently, it has been established that HA entities are able to partition between water and certain hydrogel phases.10 Depending on the nature of the hydrogel, HA may even have a distinct affinity for the gel, leading to secondary effects such as the sorptive accumulation of large quantities of metal−humic complexes in the gel phase.11 In Part 1 of the present series on partitioning of HA between water and gel,12 it was demonstrated that the spatial distribution of HA inside a gel layer with a thickness on the order of a millimeter is not homogeneous. At the gel/water interface, the outermost thin surface film of about 15−20 μm has a concentration of HA, [HA]sg, which is substantially higher than both the bulk concentration in the aqueous medium,

INTRODUCTION Humic acids (HA) form a class of ubiquitous compounds that play an indispensable role in the mobilization and transport of cations and organic compounds across environmental interfaces within aquatic bodies and soil systems.1 In particular, their ability to complex/solubilize transition metal ions is established to be a vital link in predicting the availability of metal ions to organisms.2,3 Over many years, numerous characterization studies have been carried out on various types of HA and their role in natural biogeochemical and biophysicochemical processes. The crucial properties of HA underlying such phenomena have been extensively studied, including their semiflexible structure, their mixed hydrophilicity/hydrophobicity, their subtle aggregation/disaggregation behavior, and so forth.4−7 Furthermore, HA have a propensity to be adsorbed on a broad variety of surfaces,8,9 the extent of which is generally governed by the relative hydrophobicities and charges of both © 2015 American Chemical Society

Received: December 16, 2014 Revised: January 12, 2015 Published: January 12, 2015 1737

DOI: 10.1021/la504885v Langmuir 2015, 31, 1737−1745

Article

Langmuir [HA]*m, and the concentration in the bulk of the gel, [HA]*g . From the two types of hydrogel investigated, that is, a synthetic PAAm gel and a natural alginate gel, for a pH of 6 and an ionic strength of 10 mol m−3, the PAAm gel shows a [HA]g* of about two times [HA]*m, whereas the alginate gel has a much lower concentration of HA in its bulk ( 1, and f Cd = 1, while for the alginate gel ΠD,Cd > 1, ΠD,CdHA ≈ ΠD,HA < 1, f *HA ≪ 1, and f Cd ≫ 1. A parallel table can be constructed for the various HA species by substituting the HA specific ΠD,i and f i values.21 We distinguish between speciation in the bulk of the gel and speciation in the surface film of the gel because the complexation conditions may be very different for these two gel domains. Donnan potentials, ψD, between gel phases and aqueous solutions have been measured directly by electrokinetics,22,23 and indirectly via a variety of ion concentration measurements in the two phases. Their interpretation or computation usually rides on the Oshima−Kondo expression:24

■

THEORETICAL BACKGROUND Let us consider the situation of an aqueous medium containing the probe metal ion, Cd2+, and the target complexing agent, HA, with the concentration of complexing sites on the HA by a factor of 10 or more in excess over the total Cd(II) concentration. In addition, the aqueous medium contains an indifferent 1:1 background electrolyte, NaNO3, at a concentration which in turn is in ca. 100-fold excess over the HA binding site concentration. The equilibration of this aqueous system with a hydrogel phase, in the form of a free-floating sheet with a thickness on the order of a millimeter, is involved with the transport and partitioning of all chemical species to which the gel phase is permeable. We shall assume that background electrolyte ions and Cd2+ aq ions are free to distribute between water and gel phase, whereas the partitioning of HA entities will be subject to distinct profiling as measured by confocal laser scanning microscopy, CLSM.12 The equilibrium concentration profile of HA in the two gels, in the presence of Cd2+, is given in Fig. 1. In both gels there is a marked accumulation within a thin surface film, while the bulk gel equilibrium concentrations of HA vary from one type of gel to another. Gel/Water Phase Partitioning and Speciation. The primary physicochemical properties that generally govern the speciation of Cd(II) in a gel phase equilibrated with a given

ψD = (RT /zF )arcsinh(ρg /2zFc*)

(1)

where z is the valency of the background electrolyte (equal to one in this work), c* its bulk concentration, ρg is the volume density of the immobile charges in the gel phase, and the other symbols have their usual meaning. Chemodynamics of Cd(II) Species. A solution/dispersion of transition metal ions such as Cd2+ and humic acids is composed of nanoparticulate HA entities that are partially complexed. The chemodynamic features of such soft nanoparticulate complex species may be fairly involved in the sense that complex formation and dissociation rate constants vary 1738

DOI: 10.1021/la504885v Langmuir 2015, 31, 1737−1745

Article

Langmuir

Table 1. Cd(II) Species and Their Concentrations in the Two-Phase System Aqueous Medium/Hydrogel in the Presence of Soil Humic Acida Cd(II) species in the Cd/HA water/gel system hydrated free ion, Cd2+ aq Cd2+ bound by dispersed HA nanoparticles, CdHA Cd2+ bound to the gel backbone, Cdgb Cd2+ bound to gel-immobilized HA, CdHAgb

aqueous medium, bulk [Cd2+ * aq ]m [CdHA]*m

gel surface film

gel bulk phase

governing physicochemical features

s 2+ [Cd2+ * aq ]g = ΠD,Cd[Cdaq ]m [CdHA]sg = ΠD,CdHA[CdHA]*m

2+ [Cd2+ * aq ]g* = ΠD,Cd[Cdaq ]m [CdHA]*g = ΠD,CdHA[CdHA]*m

Donnan Donnan + complexation

[Cd]sgb = fsCd ̅ ΠD,Cd[Cd2+ aq ]* m

[Cd]*gb = f CdΠD,Cd[Cd2+ aq ]* m

[CdHA]sgb = fsCdHA ̅ ΠD,CdHA[CdHA]m *

[CdHA]gb * = f CdHA * ΠD,CdHA[CdHA]m *

Donnan + complexation + binding by gel backbone Donnan + complexation + binding by gel backbone

It is assumed that the metal to ligand ratio is so small that complexation by Cd2+ does not affect the partitioning behavior of HA. PAAm gel ΠD,Cd = * > 1; alginate gel ΠD,Cd > 1, ΠD,CdHA ≈ ΠD,HA < 1, f Cd ≫ 1, f HA * ≪ 1. ΠD,CdHA ≈ ΠD,HA = 1, f Cd = 1, f HA a

ka′ = ka[S]) and inert if any change in species concentrations is not followed by significant re-equilibration (ka′t, kdt ≪ 1). In this context, the applicable ka and kd values are those for the rate-limiting steps in the considered system, that is, the smaller of ka,p and kisa , and kd,p and kisd , respectively. The above consideration of rate constants shows that, even at the level of a volume reaction, nanoparticulate metal complexes may exhibit reactivity that is significantly different from their molecular or colloidal counterparts. The voltammetric measurements undertaken in the present work are concerned with the dynamic nature of the interfacial reactions in a dispersion of nanoparticulate CdHA complexes in contact with an electrode surface that converts the Cd(II) into Cd0 after its release from the HA. The concept of lability quantifies the extent to which the complexes dissociate to release the free metal ion on the time scale of the diffusion of the CdHA entity toward the reactive electrode interface. The lability of the CdHA system toward a macroscopic surface reaction of Cd2+ can be estimated on the basis of the ratio, L, between the rate of dissociation and the rate of diffusion,27,28 for example, in the form of the expression:

with parameters such as particle size, pH, and ionic strength in a rather subtle manner.25,26 The leading kinetic parameter is the rate constant ka for the association between the Cd2+ aq ions from the medium and the HA particle to form the inner-sphere complex CdHA with the humic metal binding sites. As shown before,26 this overall ka is generally governed by the rate constant for diffusion of Cd2+ aq ions toward the HA particles, ka,p, and by the rate constant for the actual conversion of the outersphere Cd·HA ion pair into the eventual inner−sphere CdHA complex, kisa . For the present high charge density case where DCd/DHA ≫ 1, with the particle radius rp being much larger than the metal ion radius rCd, the two individual rate constants are given by26 ka,p = 4πNAvrpDCd fel,a ̅ /NS [m 3(mol of S)−1s−1]

(2)

kais = k wV osfB,Cd ̅ NAv [m 3(mol of S)−1s−1]

(3)

where fel,a ̅ is the coefficient for conductive acceleration of the positively charged metal ion by the negative electrostatic field of the particle (assumed to be unity in the present case), DCd is the diffusion coefficient of Cd2+ aq , NAv is Avogadro’s number, NS is the number of Cd2+ binding sites, S, per HA particle, kw is the dehydration rate constant of Cd(H2O)2+ ̅ is the average 6 , fB,Cd Boltzmann partitioning factor for Cd2+ inside the HA particle aq body, and Vos is the outer-sphere volume. In the so-called high charge density regime, where local electrostatic fields around individual charges have merged into a smeared-out overall field,25 Vos ≈ Vp/NS, where Vp is the particle volume. The fellow expressions for the dissociation rate constant kd, applicable to the same conditions, reflect the kinetics of release of Cd2+ from the HA particle and the inner sphere dissociation reaction, respectively. They read:27 kd,p = 3DCd (1 + K int[S])/rp2NS fB,Cd ̅ [s−1]

(4)

kdis = k wV osNAv /K int [s−1]

(5)

L = kd(ka′/DCd + kd /DHA )−1/2 δ ̅D̅ −1

(6)

where again ka and kd are the values for rate-limiting steps in the considered system, δ̅ is the mean diffusion layer thickness, and D̅ is the average diffusion coefficient applicable for the speciation conditions. For a spherical microelectrode with radius rel under steady-state conditions (times well beyond r2el/D̅ ), eq 6 still holds if δ̅ is replaced by rel.29,30

■

EXPERIMENTAL SECTION

Humic Acid Dispersions. The humic acid was a soil humic acid from the Tongbersven forest (Oisterwijk, The Netherlands). Details on its purification, elemental content, concentrations of acidic functional groups, and diffusion coefficients are available in the literature.31,32 Its average molar mass is reported to be 22 kDa,31 and significant aggregation has been observed by dynamic light scattering under the conditions used in the present work.12,21 pH titration data yield a negative charge at pH 6 corresponding to 2.5 mmol of COO− per g of HA.32 Stock HA dispersions of 1 kg m−3 were prepared at pH 9.6. Subsequent working solutions (0.04 kg m−3) were prepared by dilution in 10 mol m−3 NaNO3, and the pH was adjusted to 6 with NaOH and HNO3. The smeared-out concentration of carboxyl groups in the working solutions is 0.1 mol m−3, corresponding to an equivalent molar mass of 400 per carboxylate group. The volume fraction of HA particles in the 0.04 kg m−3 dispersion can be computed on the basis of the density of the dry HA and the approximate water content of the particles, which are 1600 kg m−3 and 80%, respectively.33 The corresponding volume fraction ϕHA is

where Kint is the intrinsic stability of the inner-sphere complex. The term Kint[S] in the rhs of eq 4 arises because kd,p is formulated as a limiting rate constant; that is, it refers to the situation of instantaneous equilibrium between the different Cd(II) inside the HA particle body. The factor (1 + Kint[S]), i.e., [Cd]total/[Cd]free, accounts for this in kd,p. At the level of the bulk solution, two limiting situations for the equilibria between a metal ion and a nanoparticulate complexant can be distinguished; for a given time scale, t, a system is denoted as dynamic if there is frequent interchange between the free and complexed metal species (ka′t, kdt ≫ 1, where 1739

DOI: 10.1021/la504885v Langmuir 2015, 31, 1737−1745

Article

Langmuir

Figure 2. Microscope photograph of the voltammetric gold amalgam microelectrode immersed in PAAm gel. Thickness of gel = 0.84 mm. 1.25 × 10−4, divided into fractions of 0.25 × 10−4 for the HA backbone and 1 × 10−4 for the liquid content. Gels. Preparation of the gels followed protocols detailed in previous work.12,17,18 In brief, the alginate gels were synthesized by gelating 1% sodium alginate solutions in small cylindrical wells, ca. 1 cm in diameter and 2 cm in depth, with CaCO3 (15 mol m−3) and −3 D-glucono-δ-lactone (30 mol m ), followed by equilibration in 50 mol m−3 Ca(NO3)2 + 20 mol m−3 NaNO3. In view of the competition between Ca2+ and Cd2+ for the HA binding sites, the present work required minimization of the Ca2+ content of the alginate gels (“Ca-poor”), which therefore differ from the Ca-rich gels used in earlier studies. To remove excess calcium, the gels were successively equilibrated with 100 and 0.1 mol m−3 NaNO3 for 2 days, allowing the retention of sufficient Ca to maintain stable gels. From the resulting gels, small disks were cut with a volume of ca. 0.2−0.3 cm3. The PAAm gels were homemade and stored until use, according to the protocol provided by DGT Research, U.K.34 They were cut into hydrogel disks with 0.84 mm thickness and a radius of 12 mm. The water volume fraction for this open pore gel type is 0.95.34 Setup of Gel/Water Partitioning Experiments. Sets of gels (n = 4) were equilibrated in aqueous media comprising either (i) 10−2 mol m−3 Cd2+, at a pH of 6.0 and an ionic strength of 10 mol m−3 NaNO3, or (ii) medium (i) plus HA at a concentration of 0.04 kg m−3. The gel disks were equilibrated in 100 mL of medium for several days to ensure that gel/medium partitioning equilibrium was attained for all involved species. The gels were floating freely in the medium which was refreshed several times over the course of the equilibration to exclude any effects of depletion in the medium. Chemical Analyses. After equilibration, the total Cd content of the gels was determined by ICP-MS analysis of acid extracts obtained by immersion in 1 M HNO3 for several days. Free Cd2+ aq concentrations in the medium and in the gels were measured by microelectrode voltammetry (see below). The concentration of HA in both gels was measured in situ by CLSM utilizing the natural fluorescence of HA.12 In case of PAAm, the HA content was also determined in situ by UV absorption at 280 cm−1. Following the reported procedure,11 the gel disks were cut into pieces of ca. 1 cm2 size and placed directly into cuvettes with the gel layer aligned perpendicular to the light path. This method could not be used for alginate due to UV absorption by the gel itself in the relevant frequency region. The HA concentrations measured by UV absorption were in agreement with those from previous work,11 and up to a factor of 3 greater than those found by CLSM. The apparently lower CLSM values may be due to increased self-quenching of the HA within the gel as a result of the high local concentrations and aggregation state of the sorbed entities.35−37 Microelectrodic Voltammetry. An Ecochemie μAutolab type II potentiostat, together with a Metrohm 663 VA stand, was used for voltammetric measurements of the probe metal Cd(II). The working electrode was a gold amalgam microelectrode prepared from a gold inlaid disk with a radius of 12.5 μm in a needle type glass housing (CH instruments). Amalgamation of the gold surface was realized by

electrochemical reduction of Hg2+ at −0.5 V (vs Ag/AgCl) from deoxygenated 5 mol m−3 Hg(CH3COO)2/ 100 mol m−3 HClO4 solution until the optimum amount of about 3 × 10‑10 moles of deposited metallic Hg was attained.17,38 Voltammetric measurements were performed starting from an initial potential of −0.35 V using a step potential of −0.5 mV and a scan rate of 1.5 mV s−1. These parameters correspond to steady-state current conditions, that is with a scan rate well below RTD/2Frel2.29 The in situ voltammetry required immersion of the gold amalgam microelectrode into the gel phases: this was realized by gentle manual immersion to ensure that the gels did not rupture and that the electrode tip maintained its full contact with the gel phase. The counter and reference electrodes remained in the aqueous medium. A microscope photograph of the microelectrode positioned in the bulk region of PAAm is shown in Figure 2. The [CdHA] and [Cd2+] can be derived from the voltammetric current if the lability of the CdHA complex is known. In the labile case, the current is proportional to {[CdHA] + [Cd2+]} and the average diffusion coefficient, D̅ , for Cd2+ and CdHA, whereas in the nonlabile case it is simply proportional to DCd[Cd2+]. The lability features of the present complexes are analyzed in the subsection ‘Lability of CdHA complexes’ below. Typical error levels in the in-gel voltammetric current signals were on the order of 2−5%.

■

RESULTS AND DISCUSSION Cd(II) Partitioning between the Aqueous Medium and Gel Phase in the Absence of HA. The partitioning of Cd(II) species between the aqueous medium and the two gel phases as derived from steady-state voltammetry is summarized in Table 2. Table 2. Partitioning of Cd(II) Species between Aqueous Medium and Gel Phase in the Absence of HAa * [Cd2+ aq ]m

−3

(mmol m )

⎡Cd 2 +⎤ (mmol m−3) ⎣ aq ⎦g

[Cd]gb (mmol m−3) ΠD,Cd fCd ̅

PAAm gel

alginate gel

10.3 10.0

10.2 19.0

0 1 1

465 1.9 25.5

a pH 6, I = 10 mol m−3. The overbar indicates average values for the entire gel phase.

PAAm Gel. For the case of PAAm gels, the blank Cd2+ voltammograms (Figure 3a) show that there is no change in position and magnitude of the wave for the gel phase as compared to those for the aqueous medium. This implies that Donnan effects are absent and hence that the gel is essentially 1740

DOI: 10.1021/la504885v Langmuir 2015, 31, 1737−1745

Article

Langmuir

comparable with the factor of ca. 20 previously found for the Ca-rich gel,17,18 where both the Ca competition and free Cd in the gel were higher. Cd Partitioning between Medium and Gel Phase in the Presence of HA. We first consider the equilibrium and chemodynamic features of the CdHA complexes in the aqueous dispersion, then use this knowledge to describe the partitioning behavior of the various Cd(II) species between medium and gel phases. Intraparticulate Speciation of CdHA Complexes. The chemodynamic analysis given below allows us to interpret the voltammetric data on the basis that the signal corresponds only 2+ to free Cd2+ * was found to aq . In the presence of HA, the [Cdaq ]m be 50% of the total Cd(II) implying that the remaining 50% is bound to HA. This corresponds with an apparent equilibrium constant, Kapp, that is, [CdHA]m */[Cd2+ *[COO−]m *, of aq ]m 3 −1 10 m mol . In an aqueous dispersion of HA nanoparticles the intraparticulate forms of Cd2+ comprise mobile Cd2+ aq at the electric potential of the HA particle body, and the inner-sphere complex CdHA. In the applicable high charge density domain, the potential typically is smeared-out and practically constant throughout the particle body; that is, the local primary Coulombic potential at an individual site is small compared to the overall smeared-out potential and thus the Eigen type outer-sphere precursor stage in the overall complexation process is essentially immaterial.25,26 The magnitude of the intra2+ 2+ 2+ particulate Cdaq , [Cdaq ]ip, is related to [Cdaq ]m* via its 2+ ̅ [Cd2+ *. The Boltzmann factor fB,Cd ̅ , i.e., [Cdaq ]ip = fB,Cd aq ]m magnitude of fB̅ ,Cd is of order 10 3;26 thus, the local −3 intraparticulate [Cd2+ aq ]ip will come to 5 mol m . Combined −4 with a particle liquid volume fraction of 1 × 10 , this implies that the smeared-out concentration of free Cd2+ aq inside the HA particles is 5 × 10−4 mol/m3 of dispersion. For the total Cd(II) concentration of 10−2 mol m−3 and 50% association with the HA particles, the result means that the local intraparticulate concentration of inner-sphere CdHA is 4.5 × 10−3 mol m−3. Accordingly, the intraparticulate ratio between bound and free Cd(II), [CdHA]/[Cd2+], equals 9. This figure actually holds for the intraparticulate complexation process where electrostatics are essentially irrelevant: these are accounted for in the transfer of the reactant Cd2+ aq from the aqueous medium to the particle body. The smeared-out carboxylate site concentration at pH 6 of about 0.1 mol m−3 corresponds to a local intraparticulate [COO−]ip of 800 mol m−3, which leads us to an intrinsic complexation constant, Kint, of 0.01 m3 mol−1. This value is quite reasonable for unidentate Cd2+/COO− binding at the applicable ionic strength, i.e., that prevailing within the HA body, being on the order of 800 mol m−3. For example, the reported stability constant for Cd2+ complexation by acetate is in the range 0.012− 0.013 m3 mol−1 (for I = 400−2000 mol m−3).41−46 Lability of CdHA Complexes. With the chemodynamic equations given in the Theoretical Background section at hand, we may estimate the bulk equilibration features of the CdHA system as well as its lability in an interfacial reaction of the free Cd2+ ion using approximate values for the various parameters. The local binding site density in the HA particles corresponds to an average charge−charge separation, Sc , of about 2 nm. The ratio of the relative magnitude of this separation as compared to the Debye screening length, κ−1, of 3 nm at ionic strength 10 mol m−3, is below unity implying that the intraparticulate electric field is strongly cooperative, i.e. the so-called high charge density regime applies.25 Thus, we may compute the

Figure 3. Steady-state voltammograms for Cd(II) as measured at a gold amalgam microelectrode in the aqueous medium (blue curves) and within the gel phase (black curves): (a) PAAm gel and (b) alginate gel. [Cd(II)]*m = 10−2 mol m−3, pH 6, ionic strength 10 mol m−3.

uncharged. The slightly smaller limiting current observed in the gel phase is due to excluded volume effect (5%) and a smaller 11,34,39,40 mobility of Cd2+ The total metal aq in this medium. concentration in the gel phase determined by ICP for the blank Cd2+ was equal to that determined by in situ voltammetry, demonstrating that Cd2+ is not bound by the PAAm gel backbone, consistent with the scheme given in Table 1. Alginate Gel. In contrast with the findings for the PAAm gels, the concentration of Cd2+ in alginate gel is increased relative to that in the aqueous medium due to both Donnan partitioning and the affinity of Cd2+ for the gel backbone. The Donnan partitioning coefficient, ΠD, can be obtained from the voltammetric curves (Figure 3b): a significant potential shift of −7 mV is observed between the blank Cd2+ signal in the gel phase and that in the medium with a concomitantly significantly higher limiting current in the gel phase, yielding a ΠD of 1.9. The extent to which Cd2+ is bound to the gel backbone was computed from the total Cd(II) determined by ICP, yielding a [Cd]gb of 465 mmol m−3 and a gel accumulation factor fCd ̅ of 25.5 with respect to the already Donnan-enhanced concentration of free Cd2+ in the gel liquid. This extremely high value is expected due to the structure of the alginate gel and is 1741

DOI: 10.1021/la504885v Langmuir 2015, 31, 1737−1745

Article

Langmuir

the immobilized CdHAgb to maintain equilibrium with the free Cd2+ in the gel liquid, [Cd2+ aq ]g, the degree of binding site coverage, θCd (= [CdHA]/[HA]), should be the same as in the medium as long as the CdHA binding strength does not change. This condition means that the inherent heterogeneity in the humic complexation is immaterial, i.e. the dependency of Kapp on θCd can be ignored. Thus, from the [HA] in the surface film and that in the bulk gel, we can simply compute the concentration of CdHA bound to the gel backbone in the surface film, [CdHA]sgb, and that bound to the gel backbone in the bulk gel, [CdHA]*gb, using Kapp. Table 3 collects the results for both the PAAm and the alginate gels.

applicable rate constants from eqs 2−5 to establish which of the association and dissociation steps are overall rate-limiting. The various applicable parameters for the present Cd−HA system are DCd = 10−9 m2 s−1,47 rp = 25 nm,12 NS ≈ 104, Vos = Vp/NS ≈ 8 −1 48 ̅ = 1, fB,Cd ̅ = 103, kw for Cd2+ 5 × 10‑27 m3, fel,a aq = 4 × 10 s , − −3 smeared-out [COO ] = 0.1 mol m , and Kint = 0.01 mol m−3. For the association process, eqs 2 and 3 yield ka,p = 2 × 104 m3 mol−1 s−1 and kisa = 1.2 × 109 m3 mol−1 s−1, respectively. Thus, the diffusive supply of Cd2+ from the medium is limiting the overall rate of the complexation process, consistent with the expected chemodynamic behavior of nanoparticulate complexes of this size with a rapidly dehydrating metal ion.49 For the dissociation process, eqs 4 and 5 yield kd,p = 0.5 s−1 for the release of a Cd2+ from the HA entity and kisd = 5 × 107 s−1 for the preceding chemical dissociation of the inner-sphere CdHA complex, respectively. Hence the dissociation step is also overall diffusion limited. The chemodynamic analysis is consistent with measured equilibrium speciation: ka,p/kd,p = K = 101.6 m3 mol−1, in reasonable agreement with the experimentally derived value of 101 m3 mol−1. Accordingly, we may base our theoretical expectations of dynamics and labilities on ka,p and kd,p as the effective rate constants for association and dissociation of CdHA. On the basis of the values for ka,p and the concentration of binding sites [S], with [S] in large excess over [Cd(II)], we find an effective quasi-monomolecular k′ap (= ka,p[S]) of 2 × 103 s−1. Together with the kd,p of 0.5 s−1, it follows that bulk equilibrium between Cd2+ ′ t and kd,pt to be much larger than aq and CdHA, requiring kap unity, is typically established at time scales above a few seconds. To compute the lability characteristics of the complex CdHA under the time scale of steady-state voltammetric experiments we use eq 6 with ka,p and kd,p as the applicable rate constants. The microelectrode used had a radius of 12.5 μm and the DHA is approximately 2 × 10‑11 m2 s−1, which leads to a value for L of 0.2. This result implies that the 25 nm radius CdHA complexes are close to nonlabile; the fraction of larger 125 nm radius CdHA complexes will be effectively inert since for 3D ligand entities lability decreases in proportion to 1/rp2.25 It should also be noted that for individual soil HA entities, with radii on the order of 3 to 5 nm,4,10 L will reach values above unity corresponding with predominantly labile behavior. For the present experimental conditions and the selected microelectrodic technique, the complexation of Cd2+ just goes through the transition from diffusion controlled to reaction-controlled chemodynamics in the 10 nm particle size domain. Cd(II) Speciation in PAAm Gel. In the presence of HA, the Cd(II) voltammograms for the medium and the bulk gel phase are essentially identical except for a small volume effect correction of the gel signal of a few percent (data not shown). The shape of the voltammograms is the same in the absence and presence of HA which confirms that only the free Cd2+ aq is detected, consistent with the chemodynamic considerations above. As shown before, there are significant differences in the [HA] between the surface layer and the bulk PAAm gel: the average HA concentrations in the surface film and the bulk phase are 10[HA]*m and 2.5[HA]*m, respectively.12 The dissolved species in the gel phase are in equilibrium with those in the medium (cf. Table 1), throughout the surface layer and the bulk gel. 2+ Therefore, the species Cdaq and CdHA have the same concentrations in the two phases, only corrected for the small volume effect. On top of the Cd2+ aq and CdHA in solution, there is a significant amount of HA bound to the gel backbone, [HA]gb, and this immobilized HA also binds Cd2+. For

Table 3. Cd and HA Partitioning between Medium and Gel Phasesa PAAm gel

Alginate

aqueous medium [Cd2+ * aq ]m [CdHA]*m [HA]m * (as [COO−])

4.9 5.3 92

bulk gel [Cd2+ 4.7 aq ]* g [Cd]*gb [CdHA]g*+[CdHA]gb * 12.8 [HA]g* + [HA]gb * (as [COO−]) 230 gel surface film s [Cd2+ ] 4.7 aq g [CdHA]sgb [CdHA]sg + [CdHA]sgb 51 [HA]sg + [HA]sgb (as [COO−]) 920 entire gel layer 194 [Cd]g a

−3

5.1 5.2 92 9.6 235

9.6 235 38 370 235

[Cd2+ aq ]

Concentrations in mmol m . measured via gold amalgam microelectrode voltammetry, [HA]g + [HA]gb determined by CLSM, and [Cd]g from ICP-MS.

Irrespective of the gel-bound CdHA in bulk gel and surface layer, the concentrations of the dissolved species Cd2+ and CdHA match their counterparts in the aqueous medium, just corrected for the volume effect in the gel. Since in the bulk gel the average HA concentration is about 2.5 times [HA]m *, the ratio [CdHA]/[Cd2+] is enlarged by approximately the same factor (see Theoretical Background section) yielding a total CdHA bulk gel concentration of 5.1 × 2.5, i.e. 12.8 mmol m−3. Comparing the full speciation of Cd(II) inside the gel with the ICP-MS determined total Cd(II) content of 194 mmol m−3 does not lead to satisfactory agreement: the averaged [CdHA] over the complete gel layer, [CdHA]g, is about 20 mmol m−3, almost a factor of 10 lower than the measured content. The explanation for this discrepancy could be related to a higher concentration of HA in the gel phase than that determined by CLSM: as already noted,12 UV transmission spectroscopic analysis of HA in the gel layer yields substantially higher average HA concentrations, up to some 3 times the CLSM data. On the basis of the UV data, the [CdHA]g would be about 40 mmol m−3, with a total Cd(II) content of around 50 mmol m−3, which is still significantly below the total Cd(II) measured. The discrepancy between the speciation-derived and measured total [Cd]g may also be due to a higher Cd-HA binding strength for the gel-bound HA: since the gel−HA interaction is hydrophobic in nature, the conformation of gel-bound HA 1742

DOI: 10.1021/la504885v Langmuir 2015, 31, 1737−1745

Article

Langmuir

However, the hydrophobically gel-bound HA also binds Cd(II), resulting in an enlarged concentration of CdHA in the gel phase. In both gels, [CdHA] is greater in the surface film at the gel/water interface than in the aqueous medium; for PAAm this is also true in the bulk of the gel, while in case of alginate HA does not penetrate the bulk gel and thus does not influence the speciation therein. The finding that, for both types of gel, the Cd speciation within the gel phase is significantly different from that in the surrounding aqueous medium has far reaching consequences for issues ranging from metal speciation analysis by sensors employing gel layers, for example, diffusive gradients in thin film (DGT), to the bioavailability of metal ions within gelatinous biological components such as biointerfacial cell walls and biofilm matrices. DGT employs PAAm gel, while alginate is somewhat biomimetic. For DGT, the result that the concentration of mobile Cd species within the equilibrated gel is comparable to that in the aqueous medium, means that the steady-state flux of the sensor will be representative of the speciation in the sample medium, albeit that the time to attain steady-state may become impractically long.51 The typical negative structural charge of the alginate gel leads to a potential difference at the gel/water interface and an increased free metal ion concentration in the gel phase. The present work shows that negatively charged HA is able to penetrate and become sorbed by such matrices, consistent with reports on the HA content of natural biofilms,52,53 and the sorption of various humic substances to phytoplankton54,55 and fish gill cells.54 It is the local dynamic speciation within biogel phases, including discrimination between free and immobile species, that is relevant for physicochemical interpretation of bioavailability and toxicity. The results reported herein enable development of mechanistic links between speciation in the aqueous medium and that in the interphasial and bulk zones of hydrogels, enabling robust predictions of bioavailability of target metal/ humic acid species. The present work utilizes a soil HA that aggregates in solution and sorbs significantly to the gel backbone in a spatially heterogeneous manner. It would be of interest to extend the study to other types/fractions of humic acids to explore the extent to which particle size and hydrophobicity of these nanoparticulate complexants govern the interphasial partitioning dynamics and spatial distribution of dispersed and gel-bound metal species in hydrogel/water systems.

may be expected to have its hydrophilic entities strongly oriented toward the gel medium. This may have the local effect of an enlarged concentration of the binding carboxylates, resulting in stronger Cd2+ binding either by increased intrinsic chemical affinity or enhanced electrostatic cooperativity effects. Indeed the available variety of data11,50 suggest that the Cd2+ binding strength of our soil HA increases with increasing f HA, that is with larger proportions of gel-bound HA. For example, at a HA concentration within PAAm of 31.5 mmol COO− m−3, the stability of the CdHA complexes was found to be the same in the aqueous medium and in the gel, while at a concentration of 150 mmol COO− m−3 in the gel the stability of CdHA was an apparent factor of 2 greater.50 The results in the present work can be explained by the log Kapp for gel-bound CdHA being ca. half a unit greater than that for the freely dispersed CdHA. Cd(II) Speciation in Alginate Gel. The alginate gel features HA concentrations in the surface film and the bulk phase of 2.5[HA]*m and