Comparative Examination of Effects of Binding of Different Metals on

---

Article pubs.acs.org/est

Comparative Examination of Effects of Binding of Different Metals on Chromophores of Dissolved Organic Matter Mingquan Yan*,† and Gregory V. Korshin‡ †

The Key Laboratory of Water and Sediment Sciences, Ministry of Education, Department of Environmental Engineering, Peking University, Beijing 100871, People’s Republic of China ‡ Department of Civil and Environmental Engineering, University of Washington, Box 352700, Seattle, Washington 98195-2700, United States S Supporting Information *

ABSTRACT: This study quantified the binding of dissolved organic matter (DOM) from Suwannee River with nine metals, Ca(II), Mg(II), Fe(III), Al(III), Cu(II), Cd(II), Cr(III), Eu(III), and Th(IV), using a differential absorbance approach. The differential spectra of DOM were closely fitted with six Gaussian bands that were present for all of the metals at varying pH values. Their maxima were located at ca. 200, 240, 276, 316, 385, and 547 nm (denoted as A0, A1, A2, A3, A4, and A5, respectively). The relative contributions and signs of the Gaussian bands were metal-specific and correlated to some degree with the covalent-bonding index of the ions and applicable complexation constants of the NICA-Donnan model. The intensity of band A4 was linearly proportional to the concentration of DOM-complexed metal, although these correlations formed two groups with different slopes, reflecting the nature of DOM−metal interactions. The results demonstrate that differential spectra yield results indicative of the nature and extent of metal and DOM interactions.

■

INTRODUCTION Recent research has shown that differential absorbance spectra of dissolved organic matter (DOM) measured at environmentally relevant concentrations of dissolved organic carbon (DOC) have consistently observed features associated with various processes, such as deprotonation of DOM molecules1−5 and their interactions with surfaces,6,7 halogen species and oxidants,8−10 and metal cations (e.g., Cu2+, Cd2+, Fe3+, and Al3+).3,11−16 The intensity and shape of the linear and log-transformed differential spectra [e.g., differential absorbance at wavelength 400 nm (DA400), differential logarithm of DOM absorbance at 350 nm (DLnA350), and change of the spectral slope in the range of wavelengths 325−375 nm (DSlope325−375)] have been shown to be correlated strongly with the amount of metals bound to DOM estimated from the data of NICA-Donnan modeling.11−13 Prior research has also shown that the differential spectra of standard Suwannee River fulvic acid (SRFA) recorded at different pH values and copper concentrations can be resolved into discrete Gaussian bands, whose positions and widths have commonalities for the cases of SRFA deprotonation and its interaction with copper(II) ions.12 These bands may be manifestations of interactions between H+ or Cu2+ with discrete groups (e.g., the carboxylic and/or phenolic moieties or their subgroups), or they may be indicative of charge-transfer effectiveness of interactions that cannot be ascribed to discrete functionalities (e.g., via electrostatic Donnan gel binding).5 However, only a very limited similarity was shown to exist between differential spectroscopic responses of selected model © 2014 American Chemical Society

compounds (e.g., tannic acid and sulfosalicylic acid) and SRFA, and the issue of possible similarities between discrete function groups in DOM and those in model compounds needs to be elucidated further. Accordingly, this study compared major features of the differential spectra of SRFA and other selected DOMs caused by their interactions with the proton and a wide range of metal ions, such as alkaline earth metals calcium(II) and magnesium(II), trivalent ions iron(III) and aluminum(III), heavy metals copper(II), cadmium(II), and chromium(III), and the rare earth ions the lanthanide europium(III) and the actinide thorium(IV). The aims of this study include (1) to examine the utility of the differential spectra approach to a wide range of metal cations and (2) to explore the nature of the engaged SFRA functionalities through the comparison of responses of this standard DOM to interactions with a wide range of metal cations for which extensive modeling and other relevant information is available.

■

MATERIALS AND METHODS Reagents and Chemicals. All chemicals were analytical reagent (AR) grade, unless otherwise mentioned. All solutions were prepared using Milli-Q water (18.2 MΩ cm−1, Millipore Corp., MA). Suwannee River humic acid (SRHA) (1R101H) Received: Revised: Accepted: Published: 3177

October 10, 2013 February 15, 2014 February 18, 2014 February 18, 2014 dx.doi.org/10.1021/es4045314 | Environ. Sci. Technol. 2014, 48, 3177−3185

Environmental Science & Technology

Article

and SRFA (1S101F) were obtained from the International Humic Substances Society (IHSS). The concentrations of SRHA and SRFA were 5.0 mg L−1 [as dissolved organic carbon (DOC)]. Titrations. Ca(II), Mg(II), Fe(III), Al(III), Cu(II), Cd(II), Cr(III), Eu(III), and Th(IV) titrations were carried out by adding requisite volumes of metal ion stock into 100 mL aliquots. The pH was maintained constant during metal additions and controlled by adding small amounts of dilute HClO4 or NaOH. It has been described in previous literature in detail.11−13 The absorbance spectra were recorded using a Perkin-Elmer Lambda 18 UV/vis spectrophotometer (with a 5 cm cell) at wavelengths from 200 to 600 nm. Stock solutions of the metals were prepared using Ca(ClO4)2, Mg(ClO4)2, Fe(ClO4)3, Al(ClO4)3, Cu(ClO4)2, Cd(ClO4)2, Cr(ClO4)3, Eu(ClO4)3, Th(ClO4)4, HClO4, and NaOH. Th(ClO4)4 was obtained via the adjustment of pH of 1000 mg L−1 thorium nitrate inductively coupled plasma/mass spectrometry (ICP/MS) standard, separation of the precipitated thorium hydroxide on a 0.45 μm filter, and dissolution of the solid phase in high-purity HClO4. DOM− metal ion complexation was modeled using the NICA-Donnan model and complexation constants included in the Visual MINTEQ database.17−22 Data for total concentrations of metal ion, which exceeded the respective precipitation levels that were determined on the basis of Visual MINTEQ calculations for each used pH, were excluded. These parameters are compiled in Table S1 of the Supporting Information. MINTEQ calculations assumed a 0.004 M ionic strength and a 10−3.5 atm partial pressure of CO2.

The application of this concept to the modeling of the differential spectra of DOM resulted in a very close fit between the observed and modeled spectra (R2 > 0.98 in most case) (Figure 2). In almost all cases, except those mentioned below, six Gaussian components were sufficient to model the experimental data. The positions of the maxima of these bands did not change as the pH or metal concentrations change, but their intensity and in some cases their width were sensitive to the changes of the system. The maxima of the bands denoted as A5, A4, A3, A2, A1, and A0 were located at 2.27 ± 0.04 eV (547 nm), 3.22 ± 0.07 eV (385 nm), 3.92 ± 0.04 eV (316 nm), 4.50 ± 0.07 eV (276 nm), 5.19 ± 0.08 eV (240 nm), and 6.23 ± 0.12 eV (200 nm), respectively. Because the contributions of band A0 were difficult to estimate because interferences from hydroxyl and other inorganic ions, only the data for E < 5.5 eV (or wavelengths above 220 nm) will be discussed henceforth. A prior study12 that examined the differential spectra of SRFA established that bands A1, A2, and A3 have some limited similarities with the response of the model species (e.g., polystyrenesulfonic-co-maleic acid, tannic acid, and sulfosalicylic acid), while bands A4 and A5 were present in the differential spectra of SRFA only, and as such, it may be a characteristic feature of this fulvic acid. We hypothesize that bands A4 and A5 reflect changes in the charge transfer complexes that strongly affect the absorbance of DOM, as discussed in prior literature.5,25 These changes may be caused by several processes that are not necessarily mutually exclusive. Such processes may include, for instance, changes of the distribution of electronic density in DOM molecules caused by the interactions between metal cations and discrete functional groups in DOM and attendant changes of the conformations of DOM molecules. 26−28 Alternatively, conformational changes may be caused by the accumulation of the metal cation in the Donnan gel space. These changes together with the other hypothetically involved interactions are very likely to affect the interactions between the donor and acceptor groups and resultant transfer bands in the absorbance spectra of DOM. Separation of the contributions of these processes requires further extensive spectroscopic experiments (e.g., those using pH titrations at varying ionic strengths, etc.), but even without precise quantitation of the contribution of these processes, it can be hypothesized that the intensity of band A4 can be considered to be reflective of the intrinsic property of metal−DOM interactions and can be accordingly used to quantitate these interactions, as discussed in more detail in the sections that follow. Effect of pH on the Differential Spectra Recorded at Varying Metal Concentrations. Variations of pH caused substantial changes of the intensity and shape of the differential absorbance spectra for all of the metals, as demonstrated by the data for Al(III) and Cd(II) in Figures S2 and S3 of the Supporting Information. The differential spectra corresponding to the interaction of 10 μM Al(III) with SRFA at varying pH values were also modeled by Gaussian fit, and results are shown in Figure 3. They demonstrate that, while the Gaussian deconvolution allowed for the achievement of a very good fit, an additional Gaussian band denoted henceforth as A6 needed to be introduced for pH 3.5 and for pH 6.0 in the case of high Al(III) concentrations. The intensity of band A6 for Al(III) and some other metals increased at lower pH values and higher total metal concentrations. We hypothesize that this feature may be a specific response to Donnan-type interactions or replacement of H+ by metal bound to specific sites in DOM, which needs to be studied in more detail in the future.

■

RESULTS AND DISCUSSION Changes of Absorbance Spectra of DOM Caused by Metals. Zero-order absorbance spectra of DOM exhibited no easily discernible changes caused by the presence of varying concentrations of the metal cation (in Figure S1 of the Supporting Information), but their differentials calculated versus the reference spectrum corresponding to DOM without metal ion additions showed multiple features, as shown for the metal ions in Figure 1. For all metals, there were peaks located at wavelengths of ca. 200, 240, 276, 316, 385, and 547 nm. However, the intensity and shape of the differential absorbance spectra differed among the metals. For instance, the negative sign of the differential absorbance (that is, its decrease compared to that in the absence of metals) was observed in the range of wavelength 220−300 nm for Ca(II), Mg(II), and Eu(III). In contrast, the absorbance increased significantly in this range of wavelengths in the presence of Cu(II), Fe(III), and Th(IV). Deconvolution of the Differential Spectra. To examine the structure of the differential spectra, they were processed to determine the presence and contributions of distinct bands. As in prior research,12,23,24 such bands were assumed to have a Gaussian shape when represented versus photon energy (measured in eV) calculated as E (eV) =

1240 λ (nm)

(1)

Each of the Gaussian bands is characterized by the location of its maximum (E0i), width (Wi), and intensity at E = E0i (A0i). The resultant differential spectra [ΔA(E)] obtained at varying pH values or metal concentrations are represented as 2⎞ ⎛ ⎛ E − E 0i ⎞ ⎟ ⎜ ΔA(E) = ∑ ΔA 0i exp⎜ −⎜ ⎟⎟ ⎝ ⎝ Wi 2 ⎠ ⎠ i

(2) 3178

dx.doi.org/10.1021/es4045314 | Environ. Sci. Technol. 2014, 48, 3177−3185

Environmental Science & Technology

Article

Figure 1. Differential spectra of DOM recorded at varying concentrations of (a) Ca(II), (b) Mg(II), (c) Fe(III), (d) Al(III), (e) Cu(II), (f) Cd(II), (g) Cr(III), (h) Eu(III), and (i) Th(IV). The DOC concentration is 5.0 mg L−1. Metal concentrations shown in the legend are in micromolar units.

Comparison of the Differential Spectra of SRFA and SRHA. To examine the utility of the approach to different DOM samples, the interaction of metals with alternative DOM samples has been studied. Only the data of SRFA and SRHA are presented here because the relevant complexation parameters are currently available in the MINTEQ database.17−20,29,30 Metal binding was studied using 1.0 μM Cu(II) and Al(III)

Band A6 did influence to some degree the intensity of absorbance at wavelengths below 350 nm, but there was little or no overlap of this band with band A4 that largely determines the intensity of differential absorbance at 400 nm. This allowed for use of band A4 in general and the intensity of absorbance at wavelength 400 nm in particular to quantify interactions between DOM and metals, as explained in the following sections. 3179

dx.doi.org/10.1021/es4045314 | Environ. Sci. Technol. 2014, 48, 3177−3185

Environmental Science & Technology

Article

Figure 2. Gaussian band fit of the differential spectra of DOM recorded at selected pH and concentrations of (a) Ca(II), (b) Mg(II), (c) Fe(III), (d) Al(III), (e) Cu(II), (f) Cd(II), (g) Cr(III), (h) Eu(III), and (i) Th(IV). The DOC concentration is 5.0 mg L−1. Metal concentrations shown in the legend are in micromolar units.

bands with identical positions and widths, but their intensities are somewhat different. However, the intensities of the differential spectra of SRFA and SRHA were predictive of the amount of DOM-bound metal, as will be discussed in the context of data interpretation based on the NICA-Donnan model in the following section.

concentrations and at pH 5.0 and 6.0 with SRFA and SRHA as different DOM samples for comparison (Figures 4 and 5, respectively). The figures indicate that differential absorbance spectra of SRFA and SRHA associated with the binding of Cu(II) and Al(III) have both striking similarities and some differences. In both cases, the differential spectra comprise six Gaussian 3180

dx.doi.org/10.1021/es4045314 | Environ. Sci. Technol. 2014, 48, 3177−3185

Environmental Science & Technology

Article

Figure 3. Gaussian band fit of the differential spectra of SRFA at varying pH values and 10.0 μM concentrations of Al(III): (a) pH 3.5, (b) pH 6.0, (c) pH 7.2, and (d) pH 9.5. The DOC concentration is 5.0 mg L−1.

Figure 4. Gaussian band fit of the differential spectra of fractions of DOM from Suwannee River, (a) SRFA and (b) SRHA, recorded at pH 5.0 and 1.0 μM concentrations of Cu(II). The DOC concentration is 5.0 mg L−1.

Figure 5. Gaussian band fit of the differential spectra of fractions of DOM from Suwannee River, (a) SRFA and (b) SRHA, recorded at 6.0 and 1.0 μM concentrations of Al(III). The DOC concentration is 5.0 mg L−1.

Correlations between Differential Absorbance at 400 nm and Metal Binding by DOM. To explore the features seen in the differential spectra of DOM caused by the presence of metal ions, theoretically predicted concentrations of metal ion bound to DOM were compared to the intensities of the differential absorbance at 400 nm (DA400). Relevant model

calculations were carried out using complexation constants included in the database of Visual MINTEQ.17−19,21,31 Comparisons between DA400 and MINTEQ-based estimates of the concentrations of DOM-bound metal ion are shown in Figure 6. They demonstrate that DA400 nm is strongly and generally linearly correlated (typically R2 > 0.95) with the 3181

dx.doi.org/10.1021/es4045314 | Environ. Sci. Technol. 2014, 48, 3177−3185

Environmental Science & Technology

Article

metal−DOM interactions, for instance, ionization and reduction potential of the ions and their polarizability, that ultimately result in a spectrum of interactions ranging from so-called hard−hard (mostly electrostatic) to soft−soft (mostly covalent) binding.32−36 To quantify formally the prevalence of such interactions, covalent [(χm)2rc] and ionic (Z2/ri) indexes are introduced with reference to the Nieboer and Richardson metal ion classification scheme.34,35 These indexes combine several parameters of the ions, for instance, their charge Z, covalent or ionic radii (rc/ri), and electronegativity, χm. The covalent index (χm)2rc of a metal ion can be considered as a measure of the relative importance of covalent interactions relative to ionic interactions. The latter type of interactions can be parametrized using the ionic index Z2/ri that quantifies the ability of a cation to form ionic bonds, specifically those in metal complexation in aqueous solution.32−36 An abbreviated Nieboer and Richardson metal ion classification scheme is shown in Figure S4 of the Supporting Information. The four metal ions [Fe(III), Al(III), Cu(II), and Th(IV)] in the group with high slopes in Figure 6a are of high covalent index or ionic index values; especially, Cu(II) has the highest value of the covalent index, and Fe(III), Al(III), and Th(IV) have the

concentration of DOM-bound metal ion. However, these correlations form two groups, one including Fe(III), Al(III), Cu(II), and Th(IV) and the other including Ca(II), Mg(II), Cd(II), Cr(III), and Eu(III). Figure 6b demonstrates that pH variations did not affect the correlations between DA400 and bound metal concentrations. Figure 6c shows that DA400 values were also correlated with the concentration of SRHA-bound Cu(II) and Al(III) calculated using Visual MINTEQ,17−20 and the slopes of the correlations for SRHA were virtually identical with those for SRFA. Correlations between of the Features of DOM Differential Spectra, Properties of Bound Metals, and Formal Complexation Parameters. As shown in Figures 1 and 2, DOM interactions with the metals result in varying shapes of the resultant differential spectra, especially in the wavelength range of