Article pubs.acs.org/JAFC
Redox-Ligand Complexation Controlled Chemical Fate of Ceria Nanoparticles in an Agricultural Soil Yuji Arai*,† and Jessica T. Dahle§ †
Department of Natural Resources and Environmental Sciences, University of Illinois at Urbana−Champaign, Urbana, Illinois 61801, United States § School of Agricultural, Forest and Environmental Sciences, Clemson University, Clemson, South Carolina 29634, United States S Supporting Information *
ABSTRACT: Ceria (CeO2) has received much attention in the global nanotechnology market due to its useful industrial applications. Because of its release to the environment, the chemical fate of ceria nanoparticles (NPs) becomes important in protecting the agricultural and food systems. Using experimental biogeochemistry and synchrotron-based X-ray techniques, the fate of ceria NPs (30 and 78 nm) in an agricultural soil (mildly acidic Taccoa entisols) was investigated as a function of exchangeable Ce(III) concentration (0.3 and 1.56 mM/kg in small and large NPs, respectively) under anoxic and oxic conditions. Both ceria NPs strongly adsorbed (>98%) in soils. Under the anoxic condition, the reduction of Ce(IV) was more pronounced in small NPs, whereas the greater concentration of exchangeable Ce(III) in large NPs facilitated the formation of Ce(III) phosphate/oxalate surface precipitates that suppressed the electron transfer reaction. The study shows the importance of redoxligand complexation controlled chemical fate of ceria NPs in an agricultural soil. KEYWORDS: cerium, ceria, cerium oxide, nanoparticles, redox, phosphate, dissolution, oxalate, fate, speciation, cerium phosphate
■
= 9.84 g/100 mL) except for Ce(III)PO4 (Ksp = 7.4 × 10−23 g/ 100 mL), Ce(III) (OH)3 (Ksp = 1.6 × 10−20), and Ce oxalate (Ksp = 0.0041 g/100 mL).7,8 It is likely that activity of hydroxyl, redox state of Ce, and ligand complexation will influence the solubility of Ce in the agrochemical environment. In our previous investigation, protons were only effective in dissolving ceria NP at pH 90 mV) and ∼70 h for under anoxic conditions (Eh ≤ 400 mV). The samples under anoxic conditions were loaded in an anaerobic glovebox and kept in Ar atmosphere until the analysis. An incident of X-ray beam was calibrated at the whiteline peak of Ce(III)PO4 at 5727 eV for the following synchrotron-based X-ray analyses. All paste samples were loaded in a polycarbonate sample holder covered with a 0.2 mm polyfilm on the front. Bulk XAS measurements were conducted at beamline X11A at the National Synchrotron Light Source (NSLS), Upton, NY, USA. The XAS spectra were collected in fluorescence mode at room temperature using a Ge13 detector. No beam-induced reduction was observed during the measurement. The monochromator consisted of two parallel Si(111) crystals with a vertical entrance slit of 0.4 mm and a horizontal entrance slit of 1 cm. A total of six to eight spectra were collected for each sample. Reference spectra of Ce reference compounds (ceria, Ce(III)PO4, Ce(III) oxalate, Ce(III) sulfate, and Ce(IV) carbonate) were collected in transmission mode at BL4-3 at the Stanford Synchrotron Radiation Laboratory, Menlo Park, CA, USA. Synchrotron-based micro-X-ray microprobe analysis was conducted at beamline X27A at NSLS. An incident of X-ray beam was calibrated at the whiteline peak of Ce(III)PO4 at 5727 eV. The samples under reduced conditions were used to prepare a thin section (∼30 μm) according to the method described by Arai et al.17 Micro-XAS analysis was not conducted for samples in the oxic system because we observed some reduction in the sample only under anoxic condition. Soil paste was air-dried in the anaerobic glovebag for 2 weeks. This drying process was necessary for the Epotek 301 resin (Epoxy Technology, Inc., Billerica, MA, USA) to cure at room temperature. To eliminate the background metal(loid) impurities, quartz slides and Super Glue were used for this process, as these materials have proven to be nearly metal free, which is necessary for backing materials during the microXRF analysis. The micro-XRF compositional maps of Ce were conducted at 5900 eV using a Canberra SL30165 Si(Li) detector. After investigating five to eight spots per XRF map, three to four most representative spots with the cleanest spectra were featured in the reported data. Four to seven scans were collected. The XAS data were normalized according to the method described in the previous work.18 Because of the formation of insoluble Ce compounds, the XANES data were processed using the linear combination of reference compounds fitting the data range of 5700−5850 eV. In this analysis, the selfabsorption correction function in SIXpack was used and no negative fit.19 The energy shifts of reference compounds were not allowed during the fit. Oxalate Ligand Promoted Dissolution of Ceria NPs. In our previous investigation, proton and phosphate ligand promoted/ suppressed ceria NP dissolution was highlighted.9 Considering the effects of carboxylate ligands (hard base) on CeO2 NP solubility and an insoluble Ce(III) oxalate precipitate, the role of oxalate ligand might be important in predicting the fate of ceria NPs. Accordingly, the effects of oxalate on ceria NP dissolution were investigated as a function of particle size and [oxalate] = 0.5, 1.5, or 3 mM. All experiments were performed in 0.01 M NaNO3 at a solid/solution
possible inhibition mechanisms. The importance of Ce(III) surface valence states in ceria NPs in biological systems has also been reported by Pulido-Reyes et al.14 The concept, however, has not been tested in heterogeneous media such as soils and sediments. Understanding the fate of ceria NPs in agricultural soils should improve the current knowledge of environmental risk assessment to protect agriculture and food systems. The objective of this study was to evaluate the chemical fate of ceria NPs, which contain different amounts of exchangeable Ce(III), in an agricultural soil using experimental geochemistry and synchrotron-based X-ray analysis.
■
MATERIALS AND METHODS
Characterization of Ceria NPs and Soil. Ceria NPs of two different particle sizes (30 and 78 nm, hereafter referred to as small and large, respectively) were obtained from the U.S. Research Nanomaterials Inc. (Houston, TX, USA). The detailed physicochemical characterization is documented in our previous work15 (SI-1). Briefly, the isoelectric point of both ceria NPs is ∼4.5. Exchangeable Ce(III) in large and small ceria NPs is 1.56 and 0.3 mM/kg, respectively, whereas the BET surface of small ceria NPs (40 ± 10 m2/ g) is greater than that of large ceria NPs (11.5 ± 3.5 m2/g). Moderately acidic (pH water: 5.2 ± 0.2) surface soils (top 10−30 cm) of Toccoa sandy loam series (coarse-loamy, thermic typic Udifluvents) were collected from an organic farm in South Carolina, USA. The soil contains Mehlich I extractable P of 35 mg kg−1. Mineralogical and physicochemical characterization of soils were reported in our previous work.16 Briefly, the soil has ∼70% acid saturation, and Ca is a dominant exchangeable cation. Soil organic carbon is 1.53%. Mineralogy is dominated by quartz in sand and silt fractions, and kaolinite, hydroxyl interlayer vermiculite, gibbsite, hematite, and goethite were identified in the clay fraction. All reagents were prepared in deionized ultrapure water (18.2 MΩ). ACS grade salts, Ce(III) sulfate octahydrate, Ce(III) oxalate, Ce(III) phosphate, Ce(III) carbonate (Alfa Aesar, Ward Hill, MA, USA), and sodium sulfate (J. T. Baker, Phillipsburg, NJ, USA), were used in this study. Ceria NPs and Ce(III) (aq) Sorption Isotherm in Soils. All batch experiments were conducted in duplicate under ambient condition (22 ± 1.2 °C and pCO2 = 10−3.45 atm) according to the method described in our previous work.15 Approximately 6.2 (±0.02) g of air-dry soil was transferred to each 50 mL Nalgene tube. A 15 mL portion of 0.017 M Na2SO4 was added to each tube, and the tubes were sealed and placed on a turnover shaker overnight. After 24 h of shaking, 15 mL of Ce(III) sulfate solution or CeO2 NP stock solutions was added to each tube to constitute samples at 50, 100, 250, and 500 mg L−1 [Ce]T. The use of Ce(III) was to understand the fate of Ce(III) if Ce(VI)O2 undergoes reductive dissolution in the soil. Cerium NP stock solutions (12.29 g L−1) were freshly prepared and sonified in 0.017 M sodium sulfate. The pH of each sample was adjusted to 5.5, and the tubes were sealed and shaken for an additional 24 h. After shaking, Ce(III) samples were centrifuged at 14540g for 10 min; then a 5 mL aliquot was drawn out with a syringe and filtered through 0.2 μm Millex polyvinylidene fluoride (PVDF) syringe filters. Using modified Stokes’ law, the centrifugal force and time were calculated to separate NPs-soil from free NPs. The NPs were centrifuged at 14540g for 50 min, and aliquots were filtered through 0.2 μm PVDF filters. Filtrates were immediately acidified with 1 M HNO3 and analyzed for total Ce concentration ([Ce]T) using inductively coupled plasma (ICP)− atomic emission spectroscopy (AES). The detection limit of Ce was approximately 0.03 μg/L. Cerium sorption in soils was analyzed to evaluate the sorption capacity of different cerium species. Aliquots were acidified with 10% HNO3 and then analyzed for phosphorus using ICP-AES. On the basis of the results, the overall distribution coefficient, or Kd, was calculated taking a ratio of sorbed to unsorbed concentrations. Synchrotron-Based X-ray Analysis. To better understand the macroscopic behavior of ceria NPs in soils, both bulk (beam size ∼ 1 mm × 1.2 cm) and microfocused (∼10 μm × 7 μm) X-ray absorption B
DOI: 10.1021/acs.jafc.7b01277 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry ratio of 0.35g/L with 5 mM sodium acetate buffer at pH 5.5. The preliminary experiments showed that no appreciative dissolution occurred in 5 mM acetate buffer at pH 5.5 for up to 6 days. The reason for using pH 5.5 is to provide a similar soil pH range for the ceria NP adsorption experiments in soils. Suspensions were kept in 125 mL Nalgene polycarbonate bottles that were covered with aluminum foil. The bottles were shaken on a reciprocal shaker at 15 rpm for up to 5 days. Samples were taken daily via a high-speed centrifugation method described in our previous work.15 The concentrations of total Ce were measured using the ICP-AES.
increases because unreacted sites decrease as the sorption sites are filled. Such sorption suggests strong sorption of Ce(III) in the soil. A Freundlich model (Figure 1b) shows a Kd value of 1198 L kg−1. Whereas sorption mechanisms cannot be elucidated without spectroscopic evidence, the retention could be indirectly described by a combination of adsorption on variable-charged mineral surfaces and the formation of bulk and/or surface precipitates. When the dissolved PO4 3− concentration was monitored without any ceria NP addition (i.e., control), [phosphate] remained at 0.21−0.33 mg/L during the first 3 days. During the Ce(III) sorption, however, it decreases from ∼0.9 to 98%) of CeO2 NPs in soils was observed at all initial Ce concentrations. For this reason, sorption isotherm plots for CeO2 NPs are not shown. Instead, the distribution coefficient, Kd = (added [Ce as CeO2] − [Ce] in filtrate)/([Ce] in filtrate), was calculated. The distribution coefficient for small and large NPs ranged from 5101 to 102,039 L/kg and from 5356 to 71,427 L/kg, respectively. These values cover the Kd value previously reported by Cornelis et al.4 in work reporting the sorption of bulk CeO2 NPs (citrate-coated 8 nm ceria) in 16 different Australian soils with various physical and chemical properties. Unlike in the Ce(III) sorption isotherm, the changes in dissolved PO43− concentration are not clearly correlated with Ce addition as CeO2 NPs. However, the P concentration immediately dropped ∼0.04 mg/L for small NPs and ∼0.02 mg/L for large NPs (Figure 1c), and the dissolved P concentration gradually decreased with increasing [Ce]T. The dissolved P remained slightly greater in the small ceria NPs at lower Ce concentration ( Fe(III) > SO42−. At the redox potential
