Quantum Dots Confined in an Organic Drop as Luminescent Probes

Apr 16, 2012 - The method was validated against CRM NWTM-27.2 lake water, and a recovery study was performed with synthetic seawater. The use of CdSe ...
1 downloads 20 Views 1MB Size
Download PDF
Article pubs.acs.org/ac

Quantum Dots Confined in an Organic Drop as Luminescent Probes for Detection of Selenium by Microfluorospectrometry after Hydridation: Study of the Quenching Mechanism and Analytical Performance Isabel Costas-Mora, Vanesa Romero, Francisco Pena-Pereira, Isela Lavilla, and Carlos Bendicho* Departamento de Química Analítica y Alimentaria, Á rea de Química Analítica, Facultad de Química, Universidad de Vigo, Campus As Lagoas-Marcosende s/n, 36310 Vigo, Spain S Supporting Information *

ABSTRACT: Following a preliminary work (Costas-Mora, I.; Romero, V.; Pena-Pereira, F.; Lavilla, I.; Bendicho, C. Anal. Chem. 2011, 83, 2388−2393), a quenching mechanism has been established for the selective detection of Se (as selenium hydride) by microfluorospectrometry using CdSe quantum dots (QDs) as luminescent probes stabilized with hexadecylamine and confined in an organic droplet. For this purpose, luminescence, luminescence lifetime, UV−vis absorption, total reflection X-ray fluorescence, transmission electron microscopy, and atomic force microscopy measurements were performed. The presence of stabilizing agents of QDs in the droplet was found to cause a critical effect on both extraction efficiency of selenium hydride in the drop and luminescence quenching. A self-quenching mechanism due to the aggregation of QDs is suggested. Aggregation is thought to occur as a result of the binding between selenide trapped into the organic drop as selenium hydride and Cd2+ present in the surface of QDs, which in turn, may cause the loss of stabilizing hexadecylamine groups. After full optimization of main variables influencing the luminescent response, the analytical performance was established. A detection limit as low as 0.08 μg L−1 Se(IV) and a repeatability expressed as relative standard deviation of 4.6% were obtained. The method was validated against CRM NWTM-27.2 lake water, and a recovery study was performed with synthetic seawater. The use of CdSe as luminescent probes in an organic drop may constitute an extremely selective, sensitive, and miniaturized assay for in situ detection of Se(IV) in water.

T

tunable, and having a symmetric emission spectrum, high emission quantum yields, and high stability against photobleaching.5,6 QDs functionalized with different stabilizing agents have found application in analytical chemistry for detection of ions7−9 and organic compounds.10,11 QDs are mostly applied in aqueous phase, and hence, a modification of QD surface is mandatory since usually the synthesis is performed in organic media containing hydrophobic stabilizing agents. Modification of the QD surface to disperse them in aqueous media can lead to a decrease in the emission quantum yield and stability,12 so working in organic media could avoid these problems. Stabilizing agents affect the interaction between QDs and target analytes, thereby being important to select the appropriate stabilizing agent for an effective analytical response.13 A further step toward a miniaturized approach in the application of QDs as luminescent probes for chemical species would be their confinement in a droplet of organic

he availability of sensitive, simple, and fast methods for the detection of selenium in environmental samples is of great interest since this element can behave as an essential or toxic element, the boundary between deficiency and toxicity being very narrow.2 The increase of selenium concentration in natural waters occurs as a result of its presence in soils, rocks, and industrial wastes. The toxicity of selenium depends on its chemical form. Among selenium species, the two inorganic forms are more toxic than organic ones and, in turn, Se(IV) is much more toxic than Se(VI).2 At present, there is a demand for portable, miniaturized, and ecofriendly analytical systems for in situ detection of chemical species. Miniaturization and integration of sample pretreatment stages along with the use of microdetection techniques and nanostructured materials can provide several benefits, which are still to be exploited.3 In the last years, the use of semiconductor nanoparticles, also called quantum dots (QDs), as luminescent probes has spread to several application areas such as chemistry, medicine, and biology.4 Thus, QDs overcome the limitations of conventional organic dyes as a result of their unique optical properties such as having a broad absorption spectrum, being narrow and size© 2012 American Chemical Society

Received: January 21, 2012 Accepted: April 16, 2012 Published: April 16, 2012 4452

dx.doi.org/10.1021/ac300221s | Anal. Chem. 2012, 84, 4452−4459

Analytical Chemistry

Article

CH3Hg+, Hg(II), and Cd(II) prepared from As2O3 (Merck), Na2TeO3 (Aldrich), Sn (Probus, Badalona, Spain), CH3HgCl (Riedel-de Häen), HgCl2 (Prolabo), and CdO (Aldrich) were employed for studying the selectivity of the system for detecting H2Se. To evaluate the applicability of the method, CRM NWTM-27.2 fortified water (National Water Research Institute of Canada) was used. Apparatus. A Thermo Scientific NanoDrop 3300 Fluorospectrometer was used to carry out luminescence measurements. The fluorospectrometer works without cuvettes and allows performing luminescence spectra with microsamples (1−2 μL). When the sample droplet is deposited onto the pedestal and it is slightly compressed by the sample arm, a liquid column is formed that is held in place by surface tension. The optical path length is 1 mm. Sample pedestals are made of stainless steel and quartz fiber. The fluorospectrometer is equipped with three solid-state light emitting diodes (LEDs) as excitation source, which are oriented 90° in respect to the detector. A 2048-element CCD array detector is connected by an optical fiber to the optical measurement surface. This instrument is powered through a USB connection to a computer without the need for an external power supply. Luminescence measurements were carried out at 520 nm after excitation at a wavelength of 470 nm using the blue LED. HS-SDME was performed using a commercially available 10 μL high precision microsyringe containing a guided-PTFE plunger (Hamilton model 1701 RN, 10 AL). Selenium hydride generation and HS-SDME were carried out in a 40 mL ambervial closed with a silicone rubber septum (Supelco). The needle was introduced through the septum so that it can be located above the surface of the sample solution at a fixed position. To study the quenching mechanism, UV−vis absorption measurements were performed using a NanoDrop Model ND1000 spectrophotometer. Operation procedures with this equipment have been previously described in an earlier work.16 TXRF measurements were performed using Bruker S2 Picofox, which was equipped with a Mo tube (1000 μA, 50 kV, 50 W), a multilayer monochromator, and a silicon-drift detector with an active area of 10 mm2. The resolution of the detector was better than 160 eV at 10 kcps (Mn Kα). Transmission electron microscopy (TEM) was performed with a JEOL JEM-2010F microscope operating at an acceleration voltage of 200 KV. Atomic force microscopy (AFM) measurements were performed using a Bruker MultiMode VIII with Nanoscope V controller. The images were acquired using both tapping and peak force modes. The luminescence decay measurements were carried out by time-correlated singlephoton counting (TCSPC) using a Horiba Jobin Yvon Fluoromax 3 with a FluoroHub module. The luminescence lifetimes were measured by TCSPC mode in conjunction with a pulsed diode light source (NanoLED 390) using a pulse duration of 1.2 ns. All luminescence decay curves were measured at the maximum of the luminescence peak (520 nm) after the excitation at 387 nm. Experimental Procedure. QDs-HS-SDME-μFS Procedure. The procedure for detection of Se(IV) in water after hydride generation using the QDs-HS-SDME-μFS approach was as follows: 5 mL of standard solution containing NaCl (15% m/v) and a stir bar (20 × 7 mm) were placed in a 40 mL amber vial that was thermostatted at 25 °C with a water bath. After injecting 1 mL of 0.5% (m/v) NaBH4 in order to convert Se(IV) in its hydride, a 3 μL drop of n-octane/decane (1:1)

solvent without additional functionalization combined to microfluorospectrometer (μFS) for detection. This approach would allow a reduction in the consumption of QDs about 500fold as compared to luminescence probing in aqueous phase using conventional luminescence spectrometers. The confinement of QDs in a microvolume of solvent can be carried out following a microseparation technique such as single-drop microextraction (SDME).14,15 Additionally, SDME allows the integration within one step of several operations typically required prior to detection such as sampling, derivatization, preconcentration, and cleanup. A preliminary screening carried out in our laboratory with fourteen volatile species using core−shell CdSe/ZnS QDs as luminescent probes confined in an organic droplet showed that some species such as Hg(II), H2S, MMT, CH3Hg+, and Se(IV) were able to produce an efficient quenching, especially the last one after hydridation.1 However, no mechanism for this quenching process was provided. Therefore, further research is required to fully exploit the potential of the new sensing approach. To this end, measurements made with different techniques such as luminescence, UV−vis absorption, total reflection X-ray fluorescence (TXRF), transmission electron microscopy (TEM), and atomic force microscopy (AFM) were performed in this work in order to study in depth the quenching mechanism. Features related to the drop composition and effect of stabilizing agents on the sensing ability of QDs were also investigated. Finally, the analytical performance of the headspace (HS)-SDME-μFS as a sensitive, selective, and reliable technique for in situ detection of Se(IV) in water was established.

■

EXPERIMENTAL SECTION Reagents and Chemicals. All chemicals were of analytical reagent grade. Suitable working standard solutions were prepared in deionized water obtained from a Milli-Q water purifier (Millipore, Mol Molsheim, France). A solution of CdSe quantum dots (5000 mg L−1) stabilized with hexadecylamine and dispersed in toluene was commercialized by Nanoco Technologies Ltd. (Manchester, UK) and distributed by SigmaAldrich under trademark of Lumidot. The particle size of QDs is 2.4 nm, and its quantum yield is ca. 6%. Working QD dispersions were prepared just before use by dilution in the selected solvent. Hexadecylamine (HDA) (Aldrich, St. Louis, MO, USA), trioctylphosphine oxide (TOPO) (Aldrich), and dodecyl sulfide (DS) (Aldrich) were tried as stabilizing agents of QDs. A 1000 mg L−1 stock standard solution of Se(IV) was prepared from Na2SeO3 (Aldrich, St. Louis, MO, USA) in 2 M HCl. Sodium chloride (Riedel-de Häen, Seelze, Germany) was used to assess the salting out effect. A solution of NaBH4 (Merck, Darmstadt, Germany) was prepared daily by dissolving the corresponding amount in 0.1 M NaOH (Prolabo, Paris, France) to stabilize the NaBH4 solution. n-Octane (Prolabo), decane (Aldrich), and dodecane (Aldrich) containing CdSe QDs were tried as extractant phases. MgCl 2 ·6H 2 O (Prolabo), CaCO 3 (Aldrich), Mg(NO 3 ) 3 ·6H2O (Merck), Fe(NO 3 ) 3 ·9H 2 O (Merck), CaSO4·2H2O (Merck), CuCl2 (Merck), Al2(SO4)3·18H2O (Scharlau, Barcelona, Spain), ethylenediaminetetraacetic acid (EDTA) (Carlo Erba, Italy), and humic acid (Fluka, Buchs, Switzerland) were employed for interference study. Moreover, stock standard solutions of As(III), Sb(III), Te(IV), Sn(IV), 4453

dx.doi.org/10.1021/ac300221s | Anal. Chem. 2012, 84, 4452−4459

Analytical Chemistry

Article

Figure 1. (A) Effect of QD concentration stabilized by HDA, DS, and TOPO at a fixed concentration of 0.5 mg mL−1. (B) Effect of addition of hydrophobic stabilizing agent to extractant phase [n-octane/decane (1:1)] on the variation of the intrinsic luminescence intensity of CdSe QDs with time. (C) Effect of HDA concentration in the extractant phase on the luminescence quenching. (D) Effect of DS concentration in the extractant phase. The Se(IV) concentration in the sample vial for panels A, C, and D was 10 μg L−1. The concentrations of QDs for panels C and D were 7 and 20 mg L−1, respectively.

containing 7 mg L−1 CdSe QDs and 1 mg mL−1 hexadecylamine was exposed to the headspace from the tip of the syringe for 3 min while the sample was stirred at 1100 rpm. Then, the remaining drop was retracted back into the microsyringe and subsequently deposited onto the pedestal of the fluorospectrometer in order to obtain the corresponding analytical signal. A blank was run before each measurement. TXRF Measurements. In order to study the extractability of different metal hydrides in the organic drop, a mixture of noctane and decane (1:1) was used as extractant phase containing different stabilizing agents [i.e., hexadecylamine (HDA), dodecylsulfide (DS), and trioctylphosphine oxide (TOPO)]. When the HS-SDME process was finished, the remaining drop was deposited over a sample quartz carrier of the TXRF spectrometer. Once the solvent was evaporated, internal standard, i.e., 1 mg L−1 Ga, was added in order to perform the quantification according to the following equation: Ci =

was followed to obtain images from analytical blanks [i.e., experiments performed in the absence of Se(IV)] and standards of Se(IV) [experiments performed in the presence of 10 μg L−1 Se(IV)]. AFM Measurements. For AFM measurements, HS-SDME was performed in the presence and absence of Se(IV) under optimized conditions. Then, the extractant drop containing CdSe QDs and HDA was deposited onto a polished monocrystalline silicon substrate, and the solvent was evaporated. This procedure was followed to obtain images from analytical blanks [i.e., experiments performed in the absence of Se(IV)] and standards of Se(IV) [experiments performed in the presence of 10 μg L−1 Se(IV)]. Luminescence Lifetime Measurements. In order to perform luminescence lifetime measurements, the experiments were performed by continuous-flow hydride generation, H2Se being extracted in 5 mL of extractant phase [n-octane/decane (1:1)] containing 7 mg L−1 and 1 mg mL−1 CdSe QDs and HDA, respectively. After 5 min of extraction, the sample was placed in a cuvette and luminescence decay was monitored. This procedure was followed to measure luminescence lifetime of analytical blanks [i.e., experiments performed in the absence of Se(IV)] and standards of Se(IV) [experiments performed in the presence of 10, 25, and 50 μg L−1 Se(IV)].

C IS·Ni·SIS NIS·Si

where Ci, Ni, and Si are the concentration, the net pulse number within the measurement spectrum, and the relative sensitivity of the element i and CIS, NIS, and SIS are the concentration, the net pulse number within the measurement spectrum, and the relative sensitivity of the internal standard element. The integration time used for all measurements was 500 s. TEM Measurements. The experimental procedure and preparation of samples for transmission electron microscopy (TEM) were as follows: Due to relatively large volumes of QD dispersions required to prepare samples for TEM measurements (about 20 μL, i.e., ten times higher than the final volume of QDs dispersion after performing HS-SDME), the experiments were performed by continuous-flow hydride generation. Volatile species generated were bubbled for 5 min through the extractant phase placed in a vial and containing QDs and HDA (7 mg L−1 and 1 mg mL−1, respectively). Then, a drop of this extractant phase was deposited on carbon-coated copper grids, and the solvent [n-octane/decane (1:1)] was evaporated. Once dry, images were obtained by TEM. The described procedure

■

RESULTS AND DISCUSSION Study of Microdrop Composition. Effect of Organic Solvents Used as Extractants. In HS-SDME, the extractant phase selected must solubilize the analyte for efficient trapping and display a low-pressure vapor in order to decrease evaporation losses during the microextraction process. Since the commercial CdSe QDs are easily soluble in nonpolar solvents, n-octane, decane, dodecane, and mixtures of n-octane and decane were tested for their confinement in a drop. When commercial CdSe QDs dispersions were merely diluted in an organic solvent for its confinement into a drop, the intrinsic luminescence intensity of QDs varied quickly over time, which could be due to surface imperfections of QDs. Moreover, 4454

dx.doi.org/10.1021/ac300221s | Anal. Chem. 2012, 84, 4452−4459

Analytical Chemistry

Article

Figure 2. Luminescence (A) and absorption (B) spectra of QDs in the presence of of Se(IV) (concentration range: 0−50 μg L−1 Se) and representation of Stern−Volmer equation for Se(IV) detection after hydridation using optimized conditions.

Hydrophobic stabilizing agents tested were HDA, DS, and TOPO. As can be observed in Figure 1B, without adding a hydrophobic stabilizing agent to the extractant phase [noctane/decane (1:1)], a sharp variation in the intrinsic luminescence of QDs occurred. This drawback was minimized by the addition of a stabilizing agent (0.5 mg mL−1 of each agent). As can be observed, the use of HDA as stabilizing agent yields higher intrinsic luminescence intensity and almost no change with time. In the case of DS and TOPO, the stability of the intrinsic luminescence of QDs over time is lower than that of HDA but it can be improved using a higher concentration of stabilizing agents (i.e., 10 mg mL−1 for DS or 2 mg mL−1 for TOPO) (not shown). However, a higher intrinsic luminescence is observed for HDA regardless of the concentrations of DS and TOPO employed. On the basis of these results, the addition of the appropriate amount of stabilizing agent allows one to work with stable luminescence intensity, when QDs are confined in an organic drop for sensing. As already mentioned, when TOPO is used as stabilizing agent, no change in the QDs luminescence is observed in the presence of Se(IV). Consequently, the effect of the hydrophobic stabilizing agent concentration on the analytical response was studied for HDA and DS. As can be observed in Figure 1C, with HDA as stabilizing agent of QDs, the analytical response increased with increasing concentration up to 1.0 mg mL−1 HDA. Nevertheless, in the case of using DS (Figure 1D) to stabilize QDs, an increasing concentration caused a decrease in the analytical response apart from yielding high analytical blanks. Despite the higher analytical response provided by DS as stabilizing agent, HDA will be employed in further experiments because low analytical blanks and improved precision are obtained. On the basis of these results, working with an extractant phase containing 7 mg L−1 CdSe QDs and 1 mg mL−1 HDA is recommended to detect Se(IV). Mechanism of Luminescence Quenching. The proposed system for the detection of Se(IV) comprises several stages. First, it is necessary to volatilize the analyte so as to facilitate the mass transfer from the sample solution to the headspace above it. Derivatization by hydridation is very effective for converting Se(IV) into volatile H2Se, which is rapidly transferred to the headspace. After volatilization, H2Se should be easily solubilized in the organic drop containing stabilized CdSe QDs and, then, the presence of the analyte in the organic drop should cause a change in the optical properties of QDs. Stern−Volmer Equation. Under optimized conditions, the luminescence quenching caused by H2Se was well described by the Stern−Volmer equation (Figure 2A):

analytical blank values (variation of QD luminescence after the HS-SDME process in the absence of analyte) were too high. When core−shell CdSe/ZnS QDs are used as pointed out elsewhere,1 the ZnS shell minimized the presence of surface imperfections and no addition of stabilizing agent was required in the drop. Nevertheless, surface defects in CdSe QDs can be minimized by coating the surface with a suitable stabilizing agent. For this reason, HDA was added to extractant phases. The best results were obtained using a mixture of n-octane and decane (1:1) containing 0.5 mg mL−1 HDA as extractant phase (see Figure S-1 in the Supporting Information). Effect of QDs Concentration. The concentration of both QDs and hydrophobic stabilizing agent may influence not only the intrinsic luminescence intensity of CdSe QDs but also the quenching caused by H2Se (i.e., analytical response). The effect of QD concentration in the sensing drop [n-octane/decane (1:1)] containing different hydrophobic stabilizing agents such as HDA, DS, or TOPO at a fixed concentration of 0.5 mg mL−1 on the analytical response was studied. The results are shown in Figure 1A. For HDA, the response increased slightly up to a 5− 7 mg L−1 concentration of QDs and, then, it dropped at higher concentrations. For DS, a significant decrease in the response was observed when the concentration of QDs was higher than 20 mg L−1. The decrease in the analytical response (I0/I) with increasing QD concentration observed in both cases may be explained on the basis of self-quenching effect due to the proximity of QDs.17 This effect tends to occur when high concentrations are used and the aggregation of QDs is produced, hence causing a reduction in the analytical signal.17,18 When TOPO was employed to stabilize QDs in the organic phase, no quenching effect was observed, which points out the importance of selecting an appropriate stabilizing agent. Effect of Stabilizing Agents. Stabilizing agents of QDs have important effects on the optical properties and luminescence response of QDs, and therefore, the sensibility and selectivity can change depending on the stabilizing agent employed. By far, mercapto-containing surface-ligands have been extensively studied for dispersing QDs in aqueous solution. Nevertheless, the effects of hydrophobic ligand-surface on the sensing ability of QDs have been rarely investigated because most applications are carried out in aqueous media. Several hydrophobic stabilizing agents of QDs were tested for their use in the QDs-HS-SDME-μFS sensing approach. These stabilizing agents must meet two requirements, namely, (i) to stabilize the optical properties of QDs in order to avoid changes in their intrinsic luminescence and (ii) to assist the microextraction of analyte in the organic drop. 4455

dx.doi.org/10.1021/ac300221s | Anal. Chem. 2012, 84, 4452−4459

Analytical Chemistry

Article

I0/I = KSV[Q ] + C

where I0 and I are the luminescence intensity in the absence and presence of quencher (analyte), respectively, KSV is the Stern−Volmer constant representing the affinity between luminopher and quencher, Q is the analyte concentration, and C is a constant close to 1. The Stern−Volmer constant (Ksv), which is usually expressed in molar concentration, was 3.0× 107 M−1. The luminescence quenching caused by other species that can be volatilized upon hydridation with NaBH4 including As(III), Sb(III), Te(IV), Sn(IV), CH3Hg+, Hg(II), and Cd(II) or other volatile specie such as H2S were studied. All of them were volatilized as described elsewhere,1,19−24 but none caused a significant analytical response despite being present at high concentration (5 mg L−1). Whereas the luminescence of core− shell CdSe/ZnS QDs is strongly affected by the presence of CH3HgH,1 this volatile Hg species does not cause any quenching effect when CdSe QDs are employed. This behavior could be ascribed to the absence of ZnS layer, which probably interacts through sulfur groups with Hg, thus decreasing the QDs luminescence. Remarkably, CdSe QDs are more selective than core−shell CdSe/ZnS QDs toward the detection of Se(IV). TXRF Measurements. In order to investigate the extractability of H2Se in nonpolar organic solvents and the high selectivity displayed by the QDs-HS-SDME-μFS approach toward Se(IV), TXRF measurements were performed. For this purpose, microextraction experiments of H2Se using organic drops [n-octane/decane (1:1)] containing HDA, DS, or TOPO at a fixed concentration of 1 mg mL−1 were carried out, and then, the amount of Se solubilized in the drop was determined by TXRF. In fact, the amount of analyte extracted when HDA was present in the extractant phase was 86 times higher than the amount extracted without addition of HDA. In regard to other hydrophobic stabilizing agents, DS and TOPO increased the amount extracted by 2.5 and 4 times, respectively, as compared to that obtained for organic drops without a stabilizing agent. In spite of the higher amount of selenium extracted using TOPO as compared to DS, no effect on the luminescence properties of QDs (i.e., no quenching) was observed when TOPO was used as stabilizing agent (Figure 1A). This indicates that the interaction between H2Se and CdSe QDs is troublesome when TOPO is present in the drop. This may be explained on the basis of the stabilizing agent structure, which stabilizes QDs properties by steric impediment, hence hindering the diffusion of H2Se in the organic drop and limiting its interaction with QDs. The latter phenomenon especially occurs when the stabilizing agent is very voluminous. These experiments demonstrated that the addition of a hydrophobic stabilizing agent to the extractant phase improved the extraction, especially when HDA was used. These results point out that the hydrophobic stabilizing agent not only plays a very important role in the extraction of H2Se but also is required for an effective interaction between H2Se and CdSe QDs. As other volatile hydrides did not cause any quenching effect in the QDs-HS-SDME-μFS technique, the microextraction of other hydride-forming elements such as As(III), Sb(III), and Te(IV) was also studied by TXRF in order to ascertain their extractability. All of them were extracted in n-octane/decane (1:1) containing 1 mg mL−1 HDA by HS-SDME. As can be observed in Figure 3, this study reveals that extraction of these

Figure 3. Comparison of concentration of analytes present in the organic drop after the HS-SDME process performed with and without addition of stabilizing agent to the extractant phase. Measurements were made by TXRF.

hydride-forming elements in a drop of octane/decane (1:1) is promoted by the presence of 1 mg mL−1 HDA, which improves the extraction by a factor 461, 150, and 170 times for As(III), Sb(III), and Te(IV), respectively. Nevertheless, As(III) and Sb(III) concentrations in the organic drop were very low despite using a 5 mg L−1 concentration in the sample vial, which may indicate that these hydrides are poorly soluble in the mixture n-octane/decane (1:1). H2Te is efficiently extracted into the organic drop containing HDA, its solubility being 30 times better than that of H2Se. Remarkably, H2Te extracted in the drop did not cause any significant effect on the luminescence properties of QDs when this was studied by μFS. These results reveal that there are two reasons why the QDs-HS-SDME-μFS technique is selective toward H2Se, namely, the poor extractability of some hydrides (e.g., H3As, H3Sb) and the poor analytical response of soluble species such as H2Te. Luminescence and UV−Vis Absorption Measurements. As can be observed in Figure 2A,B, both the luminescence intensity and absorbance of CdSe QDs decreased with increasing Se(IV) concentrations in the sample vial but bands in the spectra did not undergo any remarkable shift. This means that no changes of surface states of QDs (e.g., due to oxidation7) or QD composition (e.g., by the replacement of one of their constituents by the analyte25) occur as a result of their interaction with H2Se. Several workers have reported a quenching luminescent mechanism based on vacancies elimination of the QD surface.26−28 In this work, the luminescence quenching caused by the presence of H2Se could be due to its interaction with Cd2+ existing at the QD surface. Probably, H2Se formed by derivatization interacts with Cd2+ in the drop. This implies the elimination of the vacancies by the binding of both ions thereby facilitating nonradiative e−/h+ recombination in the QDs surface and, in turn, the luminescence quenching. On the other hand, the binding between Cd2+ and Se2− can cause the displacement of HDA, which is bound through amino terminal groups to Cd2+ present in the QD surface.28 The displacement of HDA could provoke the approximation of QDs, thereby producing the luminescence quenching.28,29 In order to confirm the proposed mechanism, luminescence decay measurements were performed in the absence and presence of Se(IV). All samples showed triexponential luminescence decay, and a decrease of luminescence lifetime with increasing Se(IV) was observed (Table S-1, Supporting Information), which indicates that the mechanism involved in the luminescence quenching is dynamic30 as is shown by the 4456

dx.doi.org/10.1021/ac300221s | Anal. Chem. 2012, 84, 4452−4459

Analytical Chemistry

Article

Figure 4. TEM (panels A, B) and AFM (panels C, D) images of CdSe QDs stabilized with HDA (1 mg mL−1) after hydridation of the sample in the absence of Se(IV) (panels A and C) and CdSe QDs stabilized with HDA (1 mg mL−1) after hydridation of the sample in the presence of Se(IV) (10 μg L−1) (panels B and D).

Stern−Volmer equation. Likewise, it is well established that a variation in the luminescence lifetime of QDs indicates the adsorption of an electron acceptor or a surface modifier onto the QD surface. These results are in agreement with the proposed mechanism, in which the binding between Cd2+ and Se2− changes the surface states and thus facilitate the nonradiative e−/h+ recombination in the QD surface. TEM and AFM Measurements. TEM and AFM measurements of QDs in the presence and absence of Se(IV) were carried out (Figure 4). Despite the difficulties to visualize QDs distributions in the samples by TEM and AFM due to high organic matter content, both provided similar results. TEM and AFM images of blank measurements (i.e., experiments performed by hydridation in the absence of selenium) showed that QDs were dispersed uniformly, while in the images corresponding to measurements of standards (i.e., experiments performed by hydridation in the presence of selenium), aggregation of QDs occur to some extent, which further confirms the mechanism outlined. Analytical Performance. Adequate derivatization of Se(IV) as H2Se by hydridation and its fast transfer to the drop containing CdSe QDs is needed for optimum analytical performance. After careful optimization of the influencing variables, the best results are reached using 5 mg of NaBH4, 1.5 M HCl concentration, 15% (m/v) NaCl concentration in the sample vial, and a stirring rate of 1100 rpm (see Figure S-2 in Supporting Information). Under these conditions, the microextraction time (i.e., time required to reach equilibrium conditions among the three phases involved, i.e., sample solution, headspace, drop) is 3 min. Study of Matrix Effects. An essential advantage of the QDsHS-SDME-μFS approach is the reduction of potential

interferences due to both the phase separation performed and the selectivity of the quenching process. In order to investigate the effect of potential matrix interferences in the sample, the luminescence quenching caused by 5 μg L−1 Se(IV) under optimized conditions was assessed. The effect of several salts (MgCl2, CaCO3, MgNO3, CaSO4, and EDTA), ionic metals (Fe(III), Cu(II), and Al(III)), and organic matter was studied. An interference effect was considered significant when the analytical response varied beyond ±10%. MgCl2 did not cause any interfering effect over the QD response at a 10 g L−1 concentration. Interfering effects were only observed from an interfering agent concentration higher than 5 mg L−1 CaCO3, 250 mg L−1 MgNO3, 450 mg L−1 CaSO4, and 1000 mg L−1 EDTA. Al(III) and Cu(II) caused a strong depressive effect, and the tolerance limits were 0.05 mg L−1 and 0.01 mg L−1, respectively. Concentrations of Fe(III) up to 10 mg L−1 can be tolerated without any change in the analytical response. The effect of natural organic matter was studied using humic acid as model compound, a positive interference effect being observed from 0.8 mg L−1 humic acid. Thus, the QDs-HS-SDME-μFS method shows a high tolerance to matrix concomitants typically present in waters. Analytical Figures of Merit. The calibration function was linear in the range of 5.0−65.0 μg L−1 Se(IV). The detection limit calculated following the 3σ IUPAC criterion was 0.08 μg L−1 Se(IV). The quantification limit calculated as 10 σ/m, where σ is the standard deviation of ten blank measurements and m is the slope of the calibration line, was 0.4 μg L−1 Se(IV). The repeatability expressed as relative standard deviation was about 4.6% (N = 7). Comparison of Analytical Techniques for Se. A wide variety of methods have been developed so far for the detection 4457

dx.doi.org/10.1021/ac300221s | Anal. Chem. 2012, 84, 4452−4459

Analytical Chemistry

Article

of Se(IV). The most common detection methods employed are UV−visible spectrophotometry,31,32 atomic absorption spectrometry,33 atomic emission spectrometry,34−37 and inductively coupled plasma-mass spectrometry.38 Some drawbacks inherent to these methods are the use of large amounts of sample and reagents, long operation times, complex and costly instrumentation, and their inability to be used for field analysis. Table 1 shows a comparison of common analytical methods for detecting traces of Se(IV). As can be observed, only a few

The binding between exogenous selenide extracted in the organic drop and Cd2+ existing at the QD surface could involve the elimination of vacancies, thus facilitating nonradiative e−/h+ recombination. The aggregation of QDs observed and the subsequent luminescence loss could also be ascribed to the displacement of the stabilizing agent (HDA) by the binding between exogenous selenide and Cd2+.

Table 1. Comparison of LODs and RSDs for the Detection of Se(IV) in Water by Different Techniques

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

methoda

LOD (μg L−1)

ETAAS HS-SDME-ETAAS FI-HG-ICPMS FI-HG-ICP-AES FI-HG-AFS FI-HG-AAS FI-HG-TXRF FI-spectrophotometry UV−visible spectrophotometry QDs-HS-SDME-μFS

1.5 0.15 0.03 0.03 0.002 0.2 0.3 0.25 0.95 0.08

RSD (%) 2.5 17.2 6.3 0.9 4.6

■

ASSOCIATED CONTENT

S Supporting Information *

■

reference 39 21 38 40 41 33 42 43 32 this work

AUTHOR INFORMATION

Corresponding Author

*Tel.: +34-986-812281. Fax: +34-986-812556. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Financial support from the Spanish Ministry of Economy and Competitiveness (Project CTQ2009-06956/BQU), the Xunta de Galicia (project 10PXIB 314030 PR), and the Vigo University (Contract for Reference Research Groups 09VIA08) is gratefully acknowledged. We thank Dr. Andrés Guerrero-Martinez and PhD student Laura Rodriguez-Lorenzo for their help to perform lifetime and AFM measurements. Also, we thank the CACTI facilities (University of Vigo) for recording the TEM and AFM images. I.C.-M. and V.R. thank the University of Vigo for a research grant.

a

ETAAS, electrothermal atomization-atomic absorption spectrometry; FI, flow injection; HG, hydride generation; ICPMS, inductively coupled plasma-mass spectrometry; ICP-AES, inductively coupled plasma-atomic emission spectrometry; AAS, atomic absorption spectrometry; AFS, atomic fluorescence spectrometry; μFS, microfluorospectrometry.

■

techniques such as FI-HG-AFS and FI-HG-ICPMS provide LODs better than that of QDs-HS-SDME-μFS. Moreover, a time as short as 3 min is required to obtain the analytical response, thus providing a high sample throughput. The proposed method is characterized by its high sensitivity and selectivity and constitutes a very simple assay for in situ Se(IV) determination. Validation and Application. In order to prove the applicability of the method, CRM NWTM-27.2 (lake water) and a synthetic seawater sample containing 1.6 μg/L and 5 μg/ L Se (IV), respectively, were analyzed. The composition of the synthetic seawater was 3% (m/v) NaCl + 0.5% (m/v) MgCl2 + 0.15% (m/v) CaSO4. Recoveries of Se after application of the standard addition method for calibration were 98.1 ± 5.6 and 101 ± 4.5% in the lake water and synthetic seawater, respectively.

REFERENCES

(1) Costas-Mora, I.; Romero, V.; Pena-Pereira, F.; Lavilla, I.; Bendicho, C. Anal. Chem. 2011, 83, 2388−2393. (2) Alaejos, M. S.; Romero, C. D. Chem. Rev. 1995, 95, 227−257. (3) Ríos, A.; Escarpa, A.; González, M. C.; Crevillén, A. G. TrAC, Trends Anal. Chem. 2006, 25, 467−479. (4) Frasco, M. F.; Chaniotakis, N. Sensors 2009, 9, 7266−7286. (5) Murphy, C. J.; Coffer, J. L. Appl. Spectrosc. 2002, 56, 16A−27A. (6) Martín-Palma, R. J.; Manso, M.; Torres-Costa, V. Sensors 2009, 9, 5149−5172. (7) Tang, C.-R.; Su, Z.-h.; Lin, B.-G.; Huang, H.-W.; Zeng, Y.-L.; Li, S.; Huang, H.; Wang, Y.-J.; Li, C.-X.; Shen, G.-L.; Yu, R.-Q. Anal. Chim. Acta 2010, 678, 203−207. (8) Butwong, N.; Noipa, T.; Burakham, R.; Srijaranai, S.; Ngeontae, W. Anal. Chim. Acta 2011, 85, 1063−1069. (9) Xu, H.; Miao, R.; Fang, Z.; Zhong, X. Anal. Chim. Acta 2011, 687, 82−88. (10) Qu, F.; Li, H. Sens. Actuators, B 2009, 135, 499−505. (11) Qu, F.; Zhou, X.; Xu, J.; Li, H.; Xie, G. Talanta 2009, 78, 1359− 1363. (12) Yu, W. W.; Chang, E.; Drezek, R.; Colvin, V. L. Biochem. Biophys. Res. Commun. 2006, 348, 781−786. (13) Chen, Y.; Rosenzweig, Z. Anal. Chem. 2002, 74, 5132−5138. (14) Pena-Pereira, F.; Lavilla, I.; Bendicho, C. TrAC, Trends Anal. Chem. 2010, 29, 617−628. (15) Pena-Pereira, F.; Lavilla, I.; Bendicho, C. Anal. Chim. Acta 2010, 669, 1−16. (16) Pena-Pereira, F.; Lavilla, I.; Bendicho, C. Anal. Chim. Acta 2009, 631, 223−228. (17) Koneswaran, M.; Narayanaswamy, R. Sens. Actuators, B 2009, 139, 91−96. (18) Liu, J.; Yang, X.; Wang, K.; Yang, R.; Ji, H.; Yang, L.; Wu, C. Chem. Commun. 2011, 47, 935−937.

■

CONCLUSIONS The QDs-HS-SDME-μFS has proved sensitive and selective for detection of Se(IV) in water. The new sensing approach is simple, portable, and miniaturized, thus allowing field analysis. A drop of n-octane/decane (1:1) as extractant phase containing 7 mg L−1 CdSe QDs and 1 mg mL−1 HDA is recommended for efficient sensing. Analytical performance is similar or even better as compared to that obtained with traditional instrumental techniques used typically in laboratories (e.g., atomic absorption, UV−vis spectrophotometry, etc.). A quenching mechanism has been outlined on the basis of measurements carried out by luminescence, UV−vis absorption, total reflection X-ray fluorescence, transmission electron microscopy, and atomic force microscopy. 4458

dx.doi.org/10.1021/ac300221s | Anal. Chem. 2012, 84, 4452−4459

Analytical Chemistry

Article

(19) Gil, S.; Loos-Vollebregt, M. T. C. d.; Bendicho, C. Spectrochim. Acta, Part B 2009, 64, 208−214. (20) González, J. C.; Lavilla, I.; Bendicho, C. Talanta 2003, 59, 525− 534. (21) Fragueiro, S.; Lavilla, I.; Bendicho, C. Talanta 2006, 68, 1096− 1101. (22) Cabon, J. Y.; Madec, C. L. Anal. Chim. Acta 2004, 504, 209− 215. (23) Caballo-López, A.; Castro, M. D. L. d. Talanta 2007, 71, 2074− 2079. (24) Lavilla, I.; Pena-Pereira, F.; Gil, S.; Costas, M.; Bendicho, C. Anal. Chim. Acta 2009, 647, 112−116. (25) Lai, S.; Chang, X.; Fu, C. Microchim. Acta 2009, 165, 39−44. (26) Zhang, B.-H.; Wu, F.-Y.; Wu, Y.-M.; Zhan, X.-S. J. Fluoresc. 2010, 20, 243−250. (27) Wu, C.-L.; Zhao, Y.-B. Anal. Bioanal. Chem. 2007, 388, 717− 722. (28) Wu, X. C.; Bittner, A. M.; Kern, K. J. Phys. Chem. B 2005, 109, 230−239. (29) Ingole, P. P.; Abhyankar, R. M.; Prasad, B. L. V.; Haram, S. K. Mater. Sci. Eng., B 2010, 168, 60−65. (30) Wang, X.; Qu, L.; Zhang, J.; Peng, X.; Xiao, M. Nano Lett. 2003, 3, 1103−1106. (31) Revanasiddappa, H. D.; Kumar, T. N. K. Anal. Bioanal. Chem. 2002, 374, 1121−1124. (32) Kartal, S.; Oymak, T.; Tokalioglu, S. J. Anal. Chem. 2010, 65, 1221−1227. (33) Zhang, Y.; Adeloju, S. B. Talanta 2008, 76, 724−730. (34) Tang, X.; Xu, Z.; Wang, J. Spectrochim. Acta, Part B 2005, 60, 1580−1585. (35) Smrkolj, P.; Stibilj, V. Anal. Chim. Acta 2004, 512, 11−17. (36) Carrión, N.; Murillo, M.; Montiel, E.; Díaz, D. Spectrochim. Acta, Part B 2003, 58, 1375−1389. (37) Bolea, E.; Laborda, F.; Castillo, J. R.; Sturgeon, R. E. Spectrochim. Acta, Part B 2004, 59, 505−513. (38) Menegário, A. A.; Giné, M. F. Spectrochim. Acta, Part B 2000, 55, 355−362. (39) Bozsai, G.; Schlemmer, G.; Grobenski, Z. Talanta 1990, 37, 545−553. (40) Narasaki, H.; Mayumi, K. Anal. Sci. 2000, 16, 65−68. (41) Méndez, H.; Lavilla, I.; Bendicho, C. J. Anal. At. Spectrom 2004, 19, 1379−1385. (42) Gomes, A. C. F.; Menegário, A. A.; Pellegrinotti, D. C.; Giné, M. F.; Filho, V. F. N. Spectrochim. Acta, Part B 2004, 59, 1481−1484. (43) Rekha, D.; Suvardhan, K.; Kumar, K. S.; Naidu, G. R. K.; Chiranjeevi, P. J. Anal. Chem. 2006, 61, 1177−1182.

4459

dx.doi.org/10.1021/ac300221s | Anal. Chem. 2012, 84, 4452−4459