Article pubs.acs.org/ac
Ultratrace Determination of Silver, Gold, and Iron Oxide Nanoparticles by Micelle Mediated Preconcentration/Selective BackExtraction Coupled with Flow Injection Chemiluminescence Detection George Z. Tsogas, Dimosthenis L. Giokas,* and Athanasios G. Vlessidis* Laboratory of Analytical Chemistry, Department of Chemistry, University of Ioannina, Ioannina, Epirus 45110, Greece S Supporting Information *
ABSTRACT: A new method has been developed for the ultrasensitive determination of silver, gold, and iron oxide nanoparticles in environmental samples. Cloud point extraction was optimized and used as a means to extract and preconcentrate all nanoparticle species simultaneously from the same sample. The extracted nanoparticles were sequentially isolated from the surfactant-rich phase by a new selective back-extraction procedure and dissociated into their precursor metal ions. Each ion solution was injected in a flow injection analysis (FIA) manifold, accommodating the chemiluminogenic oxidation of luminol, in order to amplify chemiluminescence (CL) emission in a manner proportional to its concentration. Under the optimum experimental conditions, the detection limits were brought down to the picomolar and femtomolar concentration levels with satisfactory analytical features in terms of precision (2.0−13.0%), selectivity against dissolved ions, and recoveries (74−114%). The method was successfully applied to the determination of iron oxide, silver, and gold nanoparticles in environmental samples of different complexity, ranging from unpolluted river water to raw sewage. The developed method could also serve as a basis for future deployment of molecular spectrometry detectors for the selective determination and speciation analysis of nanoparticles in environmental applications.
E
suffer from poor sensitivity. In addition, hyphenated techniques such as FFF-ICPMS require costly and complex instrumentation as well as a high degree of expertise which complicates their application in routine analysis. A less demanding and more versatile approach is microliquid liquid extraction in an ionic liquid.17 This method provides a good alternative to the analysis of nanomaterials, but some limitations on sensitivity and accuracy restrict its adaptation, at least not without more advanced instrumentation and statistical data treatment. Improved sensitivity and selectivity can be accomplished by ligand-assisted solid phase extraction/ligand-assisted liquid reextraction in a nonpolar solvent at the expense of extensive labor and lengthy turnaround times.18 Reversible noncovalent adsorption on an ionic exchange resin seems apparently less demanding, but it is still time-consuming and requires extensive analytical resources.19 A simple and effective approach to the extraction and enrichment of NPs from environmental samples is cloud point extraction (CPE). This method relies on temperature mediated aggregation of amphiphilic (surfactant) molecules at concentrations above their critical micellar concentration (cmc).20−22 The surfactant aggregates separate from water creating a
ngineered nanoparticles have provided a unique basis for innovation in a wide variety of fields such as medicine and pharmacology, analysis and diagnostics, food safety and packaging, energy conversion and storage, textiles, etc.1−4 As the manufacturing and use of nanoparticles are progressively rising, there is an increasing concern regarding their potential impact on the environment, which is inevitably the receiving end of human activities.5−7 To gain better insight into the fate, transportation, and hazards of engineered nanoparticles in the environment, it is necessary to monitor their occurrence in various environmental compartments, especially waste disposal sites and surface waters.5−8 Recent modeling and experimental studies point out that the concentration of man-made nanoparticles lie at the ng L−1 to the low μg L−1 levels.7,9−11 Therefore, separation and enrichment of nanomaterials from environmental samples is a crucial step in their analytical processing prior to their determination.12,13 A variety of approaches to the separation and concentration of nanometer sized particles such as liquid−liquid extraction,14 centrifugation, ultrafiltration, field flow fractionation (FFF), chromatography (hydrodynamic, counter-current, and sizeexclusion) has been described, and detailed reviews have appeared on the topic.15 However, in most occasions, these methods offer low enrichment factors, and unless combined with highly sensitive detectors such as ICPMS,16 they usually © 2014 American Chemical Society
Received: December 16, 2013 Accepted: February 27, 2014 Published: February 27, 2014 3484
dx.doi.org/10.1021/ac404071v | Anal. Chem. 2014, 86, 3484−3492
Analytical Chemistry
Article
bromine, EDTA disodium salt, and calcium nitrate tetrahydrate were procured from Merck (Darmstadt, Germany). C18-coated magnetic nanoparticles (nanomag-silica, C18, 250 nm) were obtained from Micromod Partikeltechnologie GmbH (Rostock, Germany). Stock solutions of silver nitrate (10.0 mM) were prepared weekly and stored at 4 °C in dark glass volumetric flasks. Aqueous standard solutions of sodium borohydride and luminol (in 0.1 M NaOH) were prepared before use. TX-114 10% (w/v) working solutions were prepared in doubly distilled water. Humic acid was used as obtained to prepare a 1000 mg L−1 stock solution in slightly alkaline distilled water. Disodium EDTA stock solution (0.4M, pH = 8.0) was prepared in distilled water and stored at room temperature. Instrumentation. Experiments were performed in a FIA manifold comprising two flow regimes connected to a peristaltic pump (MINIPULS3, Gilson, Middleton, USA). The tubing employed was PVC 1.02 mm i.d. (Anachem, UK). A six-port valve equipped with a 100 μL sample loop was used to inject the sample into the carrier solution stream. The carrier stream and the luminophor channels were merged with an external Y-piece junction shortly before entering a singleinlet flow cell. The optical flow-cell was a spiral glass unit (snailshell like cell) and was placed right in front of a light proof CL detector consisting of a Thorn EMI (Ruislip, Middlesex, England) photomultiplier tube with an internal power supply (Figure S1, Supporting Information). Details concerning the optical cell design have been reported elsewhere.36 The recorded signals were acquired by homemade software and processed in the peak height mode. PMT voltage was adjusted between 1.1 and 1.4 kV depending on CL intensity in order to avoid exceeding the saturation level of the detector (5000 mV). To prevent carryover between injections, the flow cell was rinsed with 1 M HNO3 and distilled water after each injection. A Jenway 6405 UV−vis spectrophotometer (Essex, UK) was employed for recording UV−vis spectra of nanoparticle suspensions. A Hermle Z206A centrifuge (Wehingen, Germany) and a Hettich Mikro 12−24 μL centrifuge (Tuttlingen, Germany) were used for separations. A Shimandzu AA-6800 graphite furnace atomic absorption spectrophotometer (ETAAS) with hollow cathode lamps (BGC-SR) operating at 10 mA for Au, 10 mA for Ag, and 12 mA for Fe was used for the interference study as well as for validating the results from the analysis of real samples. Measurements were performed at 242.8 nm (Au), 328.1 nm (Ag), and 248.3 nm (Fe), according to the specifications of the manufacturer, with high purity Argon gas (99.996%). A pyrolytic graphite tube was employed for the measurements. The signal intensity was calculated after integrating the background corrected absorbance for a time interval of 5 s. Atomization conditions in the presence of surfactants were the same as reported in previous works.31,32,37 The characterization of nanoparticles was performed by AFM. Measurements on dilute aqueous solutions deposited by drop casting onto silicon wafers were carried out in tapping mode by using a Bruker Multimode Nanoscope 3D equipped with Tap300-G silicon cantilevers with a tip radius of less than 10 nm and a force constant of ≈20−75 N m1. Synthesis of Nanoparticles. Gold nanoparticles (27−32 nm, Figure S2, Supporting Information) electrostatically stabilized with citrate anions were prepared as reported previously,28,38 according to the method described by Frens.39 An aqueous solution of hydrogen tetrachloroaurate (HAuCl4)
nonpolar microenvironment where apolar species are favorably partitioned (i.e., organic compounds, hydrophobic metal chelates, etc).23−27 By virtue of their large size, these surfactant aggregates are easily separated from water by mere centrifugation. Using CPE, silver,28−31 gold,28,32 and more recently ZnO33 and nCuO,34 nanoparticles were separately extracted and determined in environmental samples including their speciation from their dissolved precursor ions. However, methods enabling the simultaneous analysis of various nanoparticles are still scant.12 Additionally, there are no reports on the extraction and determination of iron oxide magnetic nanoparticles despite the fact that their use is rapidly expanding in various applications.3 In addition to extraction, a crucial parameter in the environmental surveillance of engineered nanomaterials is the selection of the appropriate detector which must offer high sensitivity and selectivity. Atomic spectrometry has therefore been the main technique of choice because it enables the determination of various nanoparticles, as their precursor metal ions, at low and ultralow concentration levels with minimal or no interferences.29−34 On the contrary, owing to sensitivity and selectivity barriers, the deployment of molecular spectrometry for the trace analysis of nanoparticles is sparse.15 Another detection technique that holds great promise for the ultrasensitive determination of nanometer sized particles is chemiluminescence (CL).35 The benefits rising from the use of nanoparticles and their precursor metal ions in the amplification of CL intensity of various luminophor compounds and analytes are well documented.35 However, owing to selectivity limitations, the utilization of CL for the realistic quantification of nanoscale particles in environmental samples has not been reported. In this work, we present the first analytical methodology for the determination of different engineered nanoparticles in a single sample. Silver, gold, and iron oxide nanoparticles, which constitute the most widely utilized nanoparticles today,12 were effectively extracted and preconcentrated simultaneously from water samples in a single step by means of micelle mediated extraction. A sequential back-extraction procedure was developed for the selective separation of each nanoparticle from the micellar phase. Determination was carried out by dissociating each nanoparticle into its precursor metal ions followed by chemiluminescence detection exploiting the oxidative chemiluminogenic reaction of luminol. In this manner, a single luminophor compound was employed throughout the analysis of all target nanoparticles. The overall procedure was accommodated in a flow injection analysis (FIA) manifold with a single pump allowing for reproducible measurements and exceptional sensitivity that compares with that obtained with more sophisticated detectors.
■
EXPERIMENTAL SECTION Materials and Chemicals. All reagents used were of analytical grade. TX-114, luminol, monobasic and dibasic potassium phosphate, and potassium hydrogen phthalate were purchased from Acros Organics (Geel, Belgium). Silver nitrate, sodium borohydride, trisodium citrate, iron chloride, iron sulfate heptahydrate, potassium persulfate, manganese sulfate monohydrate, sodium chloride, and humic acid (HA) were obtained from Sigma-Aldrich (Steinheim, Germany). High purity mineral acids for inorganic trace analysis (TraceSELECT Ultra nitric, hydrochloric and phosphoric acids) were obtained by Sigma-Aldrich. Hydrogen peroxide (30%, Perhydrol), 3485
dx.doi.org/10.1021/ac404071v | Anal. Chem. 2014, 86, 3484−3492
Analytical Chemistry
Article
Figure 1. Schematic representation of the experimental procedure for the extraction and selective isolation (back-extraction) of Fe3O4-MNPs, AgNPs, and AuNPs.
drawing 1 mL with an automatic pipet and diluting with doubly distilled water as appropriate (100−1000-fold). Experimental Procedure. The experimental procedure for the extraction of nanoparticles and the subsequent treatment for their selective isolation and detection are illustrated in Figure 1. The overall procedure for the extraction, backextraction (speciation), and analysis of all nanoparticle species is completed within 5 h. Cloud Point Extraction. In a typical CPE experiment, 9.5 mL of an aqueous standard solution of nanoparticles or water sample was placed in a 15 mL falcon vial. The ionic strength was regulated by the addition of 0.1 mL of 150 mM Ca(NO3)2, and the solution pH was adjusted to 4.5 with dilute HNO3. Then, 0.3 mL of Triton X-114 (10% w/v) and 0.125 mL of 0.4 M EDTA were added successively. The vial was mixed by vortex agitation for 1 min and then incubated for 20 min in a thermostatted water bath at 40 °C. Separation of the micellar and the bulk aqueous phases was accomplished by centrifugation at 6000 rpm (4180g) for 10 min. The vials were then cooled in an ice-bath in order to increase the viscosity of the surfactant rich phase. The supernatant was decanted by vial inversion, and the condensed micellar phase containing the target nanoparticles was retained for further analysis. The procedure was completed in 40 min, and depending on the capacity of the centrifuge and the volume of the sample, 6−12 samples could be analyzed in each run. Back-Extraction of Nanoparticles, Dissolution to Precursor Ions, and CL Analysis. Determination of Fe3O4-MNPs. The viscous micellar phase obtained from the CPE procedure was reconstituted with 0.2 mL of 5 mM H2O2. The vial was capped to avoid loss of H2O2, and the NPs were redispersed in an ultrasound water bath for a few seconds. A permanent NdFeB magnet (0.55 T) was placed at the bottom of the vial to rapidly retrieve the Fe3O4-MNPs. With the magnet held in-place, the liquid suspension was recuperated with a glass Pasteur pipet and the solid residue of Fe3O4-MNPs was washed with 1 mL of HPLC-grade methanol to remove the remaining surfactant and any coadsorbed species. Dissolution of Fe3O4-MNPs was accomplished with 500 μL of 1.6% (v/v)
(100 mL, 0.25 mM) was boiled under stirring. Then, sodium citrate (1 mL, 38.7 mM) was mixed with the gold solution. The color of the mixture changed from pale yellow to colorless to dark red. Boiling was continued for 5 min, and the AuNP solution was left to cool at room temperature. The particles were then filtered through a 0.45 μm filter to remove larger agglomerates. The average size and molar concentration of the particles were determined from their UV−vis spectra using the extinction coefficients provided by Haiss et al.40 AgNPs (11−14 nm, Figure S2, Supporting Information) were synthesized by reducing silver nitrate as described previously.41,42 Basically, 10 mL of AgNO3 solution (1.0 mM) was added dropwise into 30 mL of freshly prepared NaBH4 solution (2.0 mM) under stirring in an ice bath. At alltimes, AgNO3 and AgNP solutions were protected from light. Formation of AgNPs was evidenced by the gradual color change from light yellow to a brighter yellow and the appearance of a new absorption band at 400 nm. The molar concentration of AgNP solution was calculated according to the equations described by Liu et al.43 and Xu et al.44 and crossvalidated by flame AAS (error: −2.35%). Magnetic iron oxide nanoparticles (Fe3O4-MNPs) (12−34 nm, Figure S2, Supporting Information) were synthesized by alkaline hydrolysis of ferric and ferrous salts.45 Briefly, 4.9 g of FeCl3 (1.6873 g of Fe3+) and 4.2 g of FeSO4·7H2O (0.84368 g of Fe2+), at a Fe2+/Fe3+ ratio of 0.5, was added to deoxygenated distilled water (100 mL) at room temperature, and the solution was stirred for 5 min under continuous flow of nitrogen. Then, 25 mL of sodium hydroxide (25%, w/v) was gradually added, and a black precipitate of magnetite NPs was formed. The Fe3O4-MNPs produced were washed repeatedly with distilled water until obtaining a neutral solution (pH = 7) and redispersed in distilled water. The concentration of Fe3O4MNPs in the initial mixture was 7.8 mg mL−1 calculated by weighting 1 mL aliquots of well dispersed Fe3O4-MNPs (standard addition method, n = 6) and subtracting the weight of water. Working solutions were prepared by dispersing the nanoparticles in an ultrasonic batch and immediately with3486
dx.doi.org/10.1021/ac404071v | Anal. Chem. 2014, 86, 3484−3492
Analytical Chemistry
Article
Table 1. Optimized Experimental Parameters for the Back Extraction and Detection of Fe3O4-MNPs, AgNPS, and AuNPs experimental step back-extraction/ separation
AuNPs
CPE supernatant
CPE surfactant rich phase
dissolution
1.6% (v/v) HCl, 2 h, 60 °C clean-up with methanol
5 mM H2O2, 20 min
9.75 mM Br2/30 mM NaCl/30 mM HCl, 20 min, 70 °C water evaporation and reconstitution in methanol 1.0 M KHP−HCl, pH = 4.0
pH conditioning of extract analysis time
no
surfactant cleanup with core−shell Fe2O3@C MNPs 0.2 M PBS, pH = 7.0
