Article pubs.acs.org/est
Inhibition Mechanisms of Zn Precipitation on Aluminum Oxide by Glyphosate: A 31P NMR and Zn EXAFS Study Wei Li,*,† Yu-Jun Wang,*,†,‡ Mengqiang Zhu,§ Ting-Ting Fan,‡ Dong-Mei Zhou,‡ Brian L. Phillips,∥ and Donald L. Sparks† †
Environmental Soil Chemistry Group, Delaware Environmental Institute and Department of Plant and Soil Sciences, University of Delaware, Newark, Delaware 19717-1303 United States ‡ Key Laboratory of Soil Environment and Pollution Remediation, Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, PR China § Earth Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States ∥ Department of Geosciences and Center for Environmental Molecular Science, Stony Brook University, Stony Brook, New York, 11794-2100, United States S Supporting Information *
ABSTRACT: In this research, the effects of glyphosate (GPS) on Zn sorption/precipitation on γ-alumina were investigated using a batch technique, Zn K-edge EXAFS, and 31P NMR spectroscopy. The EXAFS analysis revealed that, in the absence of glyphosate, Zn adsorbed on the aluminum oxide surface mainly as bidentate mononuclear surface complexes at pH 5.5, whereas Zn−Al layered double hydroxide (LDH) precipitates formed at pH 8.0. In the presence of glyphosate, the EXAFS spectra of Zn sorption samples at pH 5.5 and 8.0 were very similar, both of which demonstrated that Zn did not directly bind to the mineral surface but bonded with the carboxyl group of glyphosate. Formation of γ-alumina-GPS-Zn ternary surface complexes was further suggested by 31P solid state NMR data which indicated the glyphosate binds to γ-alumina via a phosphonate group, bridging the mineral surface and Zn. Additionally, we showed the sequence of additional glyphosate and Zn can influence the sorption mechanism. At pH 8, Zn−Al LDH precipitates formed if Zn was added first, and no precipitates formed if glyphosate was added first or simultaneously with Zn. In contrast, at pH 5.5, only γ-alumina-GPS-Zn ternary surface complexes formed regardless of whether glyphosate or Zn was added first or both were added simultaneously.
1. INTRODUCTION
layered double hydroxide (LDHs) precipitates. They can be an important sink for toxic heavy metals in contaminated soils.6−9 However, in soils the presence of anionic ligands may affect precipitate formation. Yamaguchi et al.8 reported that citrate and salicylate suppressed the formation of Ni−Al LDH on gibbsite, and an α-Ni hydroxide formed instead due to the complexation of Al by these two organic ligands. Likewise, by coating kaolinite with 5% humic acids, Nachtegaal and Sparks11 suggested that Ni−Al LDH precipitation was significantly inhibited. In the past twenty years, a large amount of anthropic ligands such as herbicides were introduced to agricultural soils. One of the most widely used herbicides is glyphosate (N[phosphonomethyl]glycine, (HO) 2 -PO-CH 2 -H 2 N + -CH 2 COOH). The glyphosate molecule contains three functional groups (e.g., amine, carboxylate, and phosphonate) shown in
One of the negative impacts of industrialization was that more and more heavy metals were released from industrial sites and ultimately transported into soils, which resulted in a high accumulation of heavy metals and degradation of environmental quality.1 For example, most agricultural soils contain Zn in the range of 10−300 mg kg−1 depending on the parent material,2 and much higher concentrations of Zn (in the range of 1000−10000 mg kg−1) are found in polluted soils.3,4 Such a high concentration of Zn severely exceeds the nutrient amounts required by plants and crops, and becomes a potential threat to human beings via the food chain.5 In soils, excessive Zn can interact with soil colloids via surface adsorption reactions. Formation of surface precipitates is another important mechanism for soil Zn immobilization.6 Laboratory studies suggest that at a high concentration (>1 mM), Zn and other transition metals (e.g., Co and Ni) can react with Al-rich soil clays to form hydrotalcite-like surface precipitates.7−9 These precipitates are mixed metal/Al hydroxides and referred to as © 2013 American Chemical Society
Received: Revised: Accepted: Published: 4211
December 13, 2012 March 29, 2013 April 3, 2013 April 3, 2013 dx.doi.org/10.1021/es305120x | Environ. Sci. Technol. 2013, 47, 4211−4219
Environmental Science & Technology
Article
adsorbent for three reasons: (i) γ-alumina is an analog to naturally occurring aluminum hydroxides and Al-rich clays,21 (ii) γ-alumina has a large specific surface area,23 and (iii) the mechanism of Zn sorption on aluminum oxides is well understood.8,24,25 The objectives of this research are to test the hypothesis that glyphosate can suppress the formation of Zn−Al LDH precipitates and to address the underlying mechanisms using a combination of macroscopic methods and spectroscopic techniques.
Figure 1a, making it a strong chelator of metal ions (e.g., Al3+, Fe 3+ , Cu 2+ , Zn 2+ , and Cd 2+ ). Stable glyphosate-metal
2. EXPERIMENTAL SECTION 2.1. Macroscopic Sorption. Sorption isotherms of Zn on γ-Al2O3 (Degussa, Alumina C) in the presence of glyphosate at pH 5.5 and pH 8.0 were performed using a batch method. Zinc sorption isotherms were conducted in the concentration range of 0−1.0 mM, in a 0.01 M NaNO3 background electrolyte, and a solid/water ratio of 4 g/L, and in the absence and presence of 1 mM concentration of glyphosate. After reaction for 24 h at 25 ± 1 °C, the suspensions were centrifuged at 9000 g for 20 min to separate the solid and solution. The supernatant was filtered through a 0.22-μm membrane filter and then Zn and glyphosate in the supernatant were analyzed by inductively coupled plasma atomic emission spectroscopy (ICP-AES). The sorption density of Zn and glyphosate was calculated from the difference in concentrations between the initial and equilibrium solution. All experiments were performed in triplicate. Detailed procedures of the sorption experiments are provided in the Supporting Information. 2.2. EXAFS Data Collection and Analysis. We collected the EXAFS data for the sorption samples and relevant model compounds at beamline X11A at the National Synchrotron Light Source (NSLS) at Brookhaven National Laboratory (Upton, NY) and beamline 14W at the Shanghai Synchrotron Radiation Facility (Shanghai, China). All samples were mounted in thin plastic sample holders covered with Kapton tape and placed at 45° to the incident beam. Fluorescence data were collected using a Lytle detector positioned 90° to the beam, and transmission data were obtained simultaneously. A pair of Si(111) crystals for the monochromator were employed, which were detuned by 30% to suppress high order harmonic X-rays. Four to six scans were averaged for each sample to obtain a good signal-to-noise ratio after careful energy calibration using Zn foil (E0 = 9659 ev). The χ(k) function was Fourier transformed using k3 weighting, and all shell-byshell fitting was done in R-space using the program SIXPACK. 26 The amplitude reduction factor, S02, was determined to be 0.935 based on the fitting of the spectrum of a 50 mM Zn solution, and was applied to other samples. More details on the data analysis are provided in the Supporting Information. 2.3. 31P NMR. Solution-state 31P NMR spectra were obtained on a Bruker AV400 NMR spectrometer (9.4 T) equipped with a cryogenic quadruple nucleus probe (QNP) (i.e., 1H, 13C, 19F, and 31P), at an operating frequency of 161.8 MHz for 31P. The NMR parameters were 30° pulse (i.e., 3.07 μs pulse width), 0.68 s acquisition time, 2.0 s pulse delay, 25 °C, and ∼360 scans. The spectra were processed using standard Bruker software (e.g., Topspin) and plotted with a line broadening of 1 Hz. Solid-state 31P{1H} single pulse magic angle spinning (SP/ MAS) NMR spectra for several sorption samples were collected on a 400 MHz Varian Inova spectrometer (9.4 T), at an operating frequency of 161.8 MHz for 31P. A Varian/
Figure 1. (a) Molecular structure of glyphosate (GPS); (b) Speciation diagram of 1.0 mM Zn alone at different pH; and (c) of 1.0 mM Zn in the presence of 1.0 mM glyphosate (GPS), as calculated by the computer program WinSGW, which is available from http://www. chem.umu.se/dep/inorgchem/samarbeta/WinSGW_eng.stm. (ref 28).
complexes12−15 form which impacts the mobility of heavy metals and the degradation (abiotic or biotic) of glyphosate. A number of studies have suggested that the presence of glyphosate (GPS) can alter copper sorption mechanisms on metal oxides.16−21 Evidence from Cu K-edge extended X-ray absorption fine structure (EXAFS) and Fourier transform infrared (FTIR) spectroscopy suggested that mineral-GPS-Cu ternary surface complexes formed during the cosorption of Cu and glyphosate on either gibbsite17 or goethite.21 Fewer studies have explored the interaction between Zn and glyphosate at the mineral/water interface. Our previous research2 investigated the cosorption of Zn and glyphosate on two Chinese soils (e.g., Red soil and Wuhan soil), and revealed that the addition of glyphosate suppressed Zn sorption on both soils. We further investigated how glyphosate affected Zn sorption on goethite, suggesting the presence of glyphosate enhanced Zn sorption at pH < 7 but decreased it at pH > 7.22 However, due to a lack of spectroscopic evidence, mechanisms could not be elucidated. In this study, EXAFS spectroscopy was combined with 31P solid-state nuclear magnetic resonance (NMR) spectroscopy to understand the surface complexation of glyphosate and Zn at the mineral/water interface. We chose γ-alumina as the 4212
dx.doi.org/10.1021/es305120x | Environ. Sci. Technol. 2013, 47, 4211−4219
Environmental Science & Technology
Article
Chemagnetics T3-type probe was configured for 3.2 mm (o.d.) ZrO2 rotors. The 31P 90° pulse length was 6 μs, and the relaxation delay was 120 s. Roughly, more than 600 scans were collected for each sample to obtain a reasonable signal-to-noise ratio. The 31P chemical shifts (δP‑31) are reported relative to an external 85% H3PO4 solution, using hydroxylapatite as a secondary reference set to δP‑31 = 2.65 ppm. The 31P{1H}/27Al CP-rotational echo adiabatic passage double resonance (REAPDOR)27 NMR spectra were obtained on a 500 MHz Varian Infinity plus spectrometer (11.7 T) at operating frequencies of 202.4, 130.3, and 499.9 MHz for 31P, 27 Al, and 1H, respectively. The 3-channel Varian/Chemagnetics T3 probe assembly was configured for 4 mm (o.d.) rotors. These experiments employed excitation of a 31P signal through 31 1 P{ H} cross-polarization with a linear ramp of the 31P field and 1 ms contact time. The 1H 90° pulse length was 6 μs, and the relaxation delay 1 s. For REAPDOR, the 31P 180° pulse length was 10 μs and the duration and strength of the 27Al dephasing pulse were 1/5 of the rotor period and 75 kHz. Figure 2. 31P solution state NMR of glyphosate solution and aqueous complexes with either Al or Zn with a 1:1 concentration ratio at indicated pH.
3. RESULTS 3.1. Aqueous Speciation of Zn in the Presence of Glyphosate. The speciation diagrams for 1.0 mM Zn in the absence and presence of 1.0 mM glyphosate were calculated using WinSGW28 based on the available formation constants at 25 °C and I = 0.01 M.15,29 In the absence of glyphosate, the speciation diagram indicates aqueous Zn2+ is the predominant form at either pH 5.5 and 8.0 (Figure 1b). After addition of 1 mM glyphosate, Zn speciation changed dramatically (Figure 1c). Free aqueous Zn2+ only dominated at pH < 3.0. As pH increased, aqueous Zn-GPS complexes are present in the form of either aqueous ZnL− or ZnHL complexes (L represents the GPS ligand) due to the complexation of glyphosate. At pH 4− 6, a combination of Zn 2+ , ZnL − and ZnHL existed simultaneously, whereas at pH > 7 ZnL− became the dominant species. In the following sections, 31P NMR and Zn EXAFS analysis confirms the existence of a ZnL− complex and reveals its structure. 3.2. 31P Liquid State NMR. We employed 31P solution state NMR to analyze several solution samples including a 25 mM glyphosate solution at three different conditions, Zn/GPS at pH 5.5 and 8.0 and Al/GPS at pH 2.0 (Figure 2). At pH 5.5, the glyphosate solution yielded a peak at chemical shift of 9.5 ppm, whereas at pH 8 this peak shifted slightly upfield at 9.3 ppm due to the deprotonation process. At pH 2.0, the NMR peak of glyphosate continued shifting to the downfield at 9.7 ppm, consistent with a protonated state. Complexation with Zn at both pHs gave the same chemical shift at 22 ppm. This is a nearly 12 ppm difference from that for free glyphosate, suggesting direct covalent bonding of P−O−Zn due to the complexation of Zn and the phosphonate group of glyphosate.27 The complexation between Zn and the carboxyl group could be determined from 31P NMR spectroscopy, but was later confirmed by Zn EXAFS analysis (as described in the following sections and Table S1 in the Supporting Information). A solution sample containing 10 mM of dissolved glyphosate and Al at pH 2 was prepared, from which we observed a major peak at 9.7 ppm and several additional small peaks at −3 and −7 ppm, respectively. These results are in agreement with those reported by Purgel et al.29 where the peak at 9.7 ppm was assigned to free glyphosate and the additional peaks to Al-GPS aqueous complexes. The large chemical shift
differences between the Al-GPS aqueous complexes and free glyphosate were due to the chemical shielding effect of Al through the covalent P−O−Al bonding.29 3.3. Macroscopic Sorption Behavior. Sorption isotherms were carried out for Zn sorption on γ-alumina at pH 5.5 and pH 8.0. A Langmuir-type isotherm was observed for Zn sorption at pH 5.5, where Zn sorption increased with Zn solution concentration and reached a maximum sorption at high Zn concentration (Figure 3a). In contrast, at pH 8.0, Zn sorption showed a linear relationship with Zn concentration in solution, leading to a large Zn uptake at the initial concentration of 1.0 mM, which is more than 30 times higher than that at pH 5.5. The large Zn uptake at pH 8.0 implied the formation of surface precipitates, which were confirmed by the EXAFS analysis described in section 3.5. The presence of glyphosate significantly changed the amount of Zn sorption at both pHs. All the Zn sorption isotherms can be described by a conventional Langmuir equation
Q=
Q mKC (1)
1 + KC −1
where Q is the amount of sorbed Zn (mg kg ), C is the equilibrium concentration (mg L−1), Qm is the maximum adsorption amount, and K is the equilibrium constant for the sorption reaction. A least-squares fit yields a Qm of 563 mg kg−1 in the absence of glyphosate at pH 5.5. As shown in Figure 3a, Zn uptake by γ-alumina was facilitated by the presence of glyphosate. With addition of 1.0 mM glyphosate, the Qm increased to 4505 mg kg−1, nearly eight times the Qm for Zn sorption alone. This can be explained from the aqueous speciation. As indicated by the isotherm at pH 5.5 (Figure 3a), only a very small portion of aqueous Zn was adsorbed by the mineral in either the absence or presence of glyphosate, leaving a large amount of aqueous Zn (50−60 ppm) still present in solution to complex with additional glyphosate and thus to form neutral and negatively charged Zn-GPS aqueous complexes (e.g., ZnHL and ZnL− as shown in Figure 1c) in solution. Given that the pHPZC of γ-alumina is 9.123 and 4213
dx.doi.org/10.1021/es305120x | Environ. Sci. Technol. 2013, 47, 4211−4219
Environmental Science & Technology
Article
Figure 4. (a) 31P single pulse magic angle spinning solid state NMR spectra for several sorption samples. Spectra were collected at a spinning rate of 8 kHz, a pulse delay of 120 s, and approximately 600 scans for each sample. (b) 31P{1H}/27Al REAPDOR spectra for GPS/ Zn cosorption samples at pH 5.5 and 8.0, respectively. Spectra were acquired with a spinning rate of 8 kHz, 1 s repetition delay, 1 ms CP contact time, 6 rotor cycles, and 15000 scans.
which the phosphate group complexed with γ-alumina. The chemical shift of this peak exhibited a 7 ppm difference (toward the upfield) from the NMR peak of the free glyphosate (Figure 2), which is similar to the difference observed for free glyphosate and the GPS/Al aqueous complex.31 This reflects the shielding effect of Al through a covalent Al−O−P bond linkage.32 In contrast, the peaks at δP‑31 = 16 and 18 ppm cannot be unequivocally assigned based on chemical shifts only. Therefore, 31P{27Al} REAPDOR NMR experiments23 were performed to further help assign the two peaks. The REAPDOR experiments allow one to measure the strength of 31P−27Al heteronuclear dipolar interaction, which in turn reflects interatomic distances and geometry (i.e., bidentate or monodentate complexes).32 Experimentally, we acquired sets of two 31P-observed NMR spectra, one (S) obtained with a heteronuclear dipolar dephasing sequence and a control spectrum (S0) acquired under identical conditions but without irradiation at the 27Al frequency. Variation in the difference in peak intensity between the dephased spectrum and the control (ΔS = S0 − S) can be used to quantify the 31P−27Al heteronuclear dipolar coupling. Figure 4b shows typical sets of 31 P−27Al REAPDOR NMR spectra, acquired for the pH 5.5 Zn-GPS cosorption sample with a 0.75 ms dephasing time, where we found the peak intensities for the peaks at δP‑31 = 6 and 16 ppm in the REAPDOR spectrum (S) were much lower than those in the control spectrum (S0). This suggested strong 27 Al−31P dipolar coupling, which in turn means a small P−Al distance (i.e.,