Development of a Titanium Dioxide-Coated Microfluidic-Based

Sep 25, 2013 - ... Chromatography with Inductively Coupled. Plasma-Mass Spectrometry for Determination of Inorganic Selenium. Species. Tsung-Ting Shih...
0 downloads 0 Views 401KB Size
Download PDF
Article pubs.acs.org/ac

Development of a Titanium Dioxide-Coated Microfluidic-Based Photocatalyst-Assisted Reduction Device to Couple HighPerformance Liquid Chromatography with Inductively Coupled Plasma-Mass Spectrometry for Determination of Inorganic Selenium Species Tsung-Ting Shih,† Cheng-Hsing Lin,† I-Hsiang Hsu,‡ Jian-Yi Chen,† and Yuh-Chang Sun*,† †

Department of Biomedical Engineering and Environmental Sciences, National Tsing Hua University, 30013 Hisnchu, Taiwan Center for Measurement Standards, Industrial Technology Research Institute, 30011 Hsinchu, Taiwan

‡

S Supporting Information *

ABSTRACT: We developed a selective and sensitive hyphenated system employing a microfluidic-based vapor generation (VG) system in conjunction with high-performance liquid chromatography (HPLC) separation and inductively coupled plasma-mass spectrometry (ICPMS) detection for the determination of trace inorganic selenium (Se) species. The VG system exploited poly(methyl methacrylate) (PMMA) substrates of high optical quality to fabricate a microfluidic-based photocatalyst-assisted reduction device (microfluidic-based PCARD). Moreover, to reduce the consumption of photocatalysts during analytical procedures, a microfluidic-based PCARD coated with titanium dioxide nanoparticles (nano-TiO2) was employed to avoid consecutive loading. Notably, to simplify the coating procedure and improve the stability of the coating materials, a dynamic coating method was utilized. Under the optimized conditions for the selenicals of interest, the online HPLC/TiO2-coated microfluidic-based PCARD/ICPMS system enabled us to achieve detection limits (based on 3σ) of 0.043 and 0.042 μg L−1 for Se(IV) and Se(VI), respectively. Both Se(IV) and Se(VI) could be efficiently vaporized within 15 s, while a series of validation experiments indicated that our proposed method could be satisfactorily applied to the determination of inorganic Se species in the environmental water samples.

S

the environmental water samples is extremely desirable. Because the concentration of Se species in typical environmental samples can be as low as subμg L−1, high sensitivity and selectivity are necessary requirements for the development of a suitable analytical technique. To meet these requirements, a combination of separation techniques involving atomic

elenium is a trace element that occurs naturally in the environment and is released through both natural processes and human activities.1 Notably, because Se can act as both an essential and a highly toxic element, depending on its oxidation state,2 the field of Se speciation is receiving increasing attention since the past few decades. Generally, the predominant forms of Se in water are selenite (i.e., Se(IV)) and selenate (i.e., Se(VI)).3 Inorganic Se species are reportedly more toxic than the common organo-Se compounds, with Se(IV) being considered more harmful to aquatic invertebrates and fish than Se(VI).4,5 Hence, the speciation of inorganic Se species in © 2013 American Chemical Society

Received: March 28, 2013 Accepted: September 25, 2013 Published: September 25, 2013 10091

dx.doi.org/10.1021/ac400934e | Anal. Chem. 2013, 85, 10091−10098

Analytical Chemistry

Article

sensitivity.36,37 In 2012, Li et al. further used noble metalmodified nano-TiO2 and bare zirconium dioxide nanoparticles (nano-ZrO 2 ) to improve the detection sensitivity for subsequent Se speciation.38 However, additional demands for operation, such as fabrication of tangle-prone tubing/fragile chamber photoreactors, preparation of nano-TiO2 suspensions/ films, and so on, hamper a more widespread use of the nanoTiO2-enhanced photoinduced VG technique. The predicament of the nano-TiO2-enhanced photoinduced VG technique in terms of additional operation demands may be rescued until the advent of suitable microfluidic-based systems.39,40 Because unique advantages such as integration for multifunctionality, short reaction time, and low reagent/ sample requirement benefit from downsizing the dimensions of analytical systems,40 a strategy based on the concept of microfluidics for the construction of nano-TiO2-enhanced photoinduced catalytic reactors seems attractive to enhance the analytical performance. In this study, we developed a selective and sensitive hyphenated system employing a microfluidic-based VG system in conjunction with HPLC separation and ICPMS detection for the determination of trace inorganic Se species in the environmental water samples. Because poly(methyl methacrylate) (PMMA) exhibits excellent properties, including ease of fabrication, optimal chemical/ mechanical properties, and optical clarity,41−43 the VG system exploited high optical-grade PMMA substrates to fabricate a microfluidic-based photocatalyst-assisted reduction device (PCARD). In addition, to reduce the consumption of photocatalysts during analytical procedures, a microfluidicbased PCARD coated with TiO2 catalysts was employed instead of resorting to consecutive loading, while a simple coating technique was developed to simplify the coating procedure and improve the stability of the coating materials. By irradiating the TiO2-coated microfluidic-based PCARD with high-flux ultraviolet (UV) light, both Se(IV) and Se(VI) can be efficiently vaporized. After optimizing the operating conditions for the separation process, a gas−liquid separator (GLS) was placed in series to function as the interface between the developed TiO2-coated microfluidic-based PCARD and the ICPMS system. Thus, we established a simple and sensitive HPLC/TiO2-coated microfluidic-based PCARD/ICPMS hyphenated system, which facilitated the analytical differentiation of inorganic Se species in the water samples. To the best of our knowledge, this is the first study that combines a microfluidicbased platform with an online speciation system for highly sensitive and selective determination of inorganic Se species, i.e., a chromatographic column for Se(IV)/Se(VI) separation, a TiO2-coated microfluidic-based PCARD for VG, and a GLS for removing liquid components from gaseous products.

spectrometric detectors is usually necessary. Indeed, for Se speciation studies of different samples, the direct coupling of high-performance liquid chromatography (HPLC) with inductively coupled plasma-mass spectrometry (ICPMS) is the most commonly employed analytical technique.6−9 Although ICPMS is a very powerful technique for trace and isotopic analyses, interference to the molecular ion, caused by the presence of argon (Ar) or chlorine, often disturbs the measurement of Se isotopes. In addition, poor ionization efficiency of Se in the Ar plasma (ca. 30%), resulting from its high ionization energy (9.75 eV), may lead to a relatively low sensitivity, thus limiting the detection of Se at low concentration.10 For these reasons, as well as to avoid the deteriorating performance of the ICPMS instrumentation because of excessive loading of salt- or organic material-rich effluents, the hydride-generation (HG) technique has proved to be an attractive interfacing technique for coupling HPLC and ICPMS for the determination of ultratrace Se species.11−13 Presently, sodium tetrahydroborate (NaBH4)-mediated reduction of hydride-forming elements in acidic media is almost universally employed in HG applications. Considering the reduction kinetics of different Se species by NaBH4, the hexavalent state of Se has been reported as not being reducible to the hydride state.14,15 Consequently, as a mandatory protocol, inorganic speciation by HG techniques first involves determining Se(IV) and then transforming Se(VI) to Se(IV) before determining the total concentration.16 To convert Se(VI) to Se(IV), several reducing agents such as thiourea,17−21 hydrochloric acid,22,23 and hydrobromic acid24,25 have been used under different conditions, all of which may increase potential contamination and analyte signal loss resulting from the conversion of Se species to the metallic state.26,27 Furthermore, the stability of Ar plasma may worsen when NaBH4 is used as the hydride generator because such an online HG system delivers not only hydride vapor but also a large quantity of hydrogen into the ICP.28 Therefore, it is of great interest to find new alternative vapor generation (VG) techniques. Recently, various VG techniques have been proposed, of which photoinduced VG has emerged as one of the most popular techniques. On the basis of the reaction involved, photoinduced VG techniques can be classified into two types: (i) pure photoinduced VG and (ii) titanium dioxide nanoparticles (nano-TiO2)-enhanced photoinduced VG. In 2003, Guo et al. started the research of the pure photoinduced VG for Se speciation,29 and another task for mercury speciation was then proposed by Zheng et al..30 To date, Sturgeon’s group keeps playing a pioneering role in the development of the photoinduced VG technique.12,29,31,32 Nevertheless, insufficient efficiency of the pure photoinduced VG for conversion of Se(VI) into gaseous products remains an obstacle for the speciation of inorganic Se species.33,34 In 2005, Wang et al. first combined a pure photoinduced VG technique with nano-TiO2 photocatalysts, namely, the nano-TiO2-enhanced photoinduced VG technique, for improvement on the conversion efficiency of Se(VI).35 Then, in 2006, Sun et al. further used a poly(tetrafluoroethylene) (PTFE) tubing for coupling HPLC separation with ICPMS detection as well as for vaporizing the two inorganic Se species via a nano-TiO2-enhanced photoinduced reduction reaction.33 Even so, the mismatch in terms of wavelength between the transmittance of fluoropolymer-based reactors and the photocatalytic activation reaction by nano-TiO2 still hinders the improvement of analytical

■

EXPERIMENTAL SECTION Chemicals and Materials. Unless otherwise stated, all chemicals were analytical reagent grade and used as received. High-purity water was obtained using a Milli-Q apparatus (Millipore, Bedford, MA, USA). Sodium hydroxide (NaOH), sodium chloride (NaCl), nitric acid (HNO3, 69.0−70.0%), and hydrochloric acid (HCl, 36.5−38.0%) were obtained from J. T. Baker (Phillipsburg, NJ, USA). Ammonium bicarbonate (NH 4 HCO 3 ), sodium selenate (Na 2 SeO 4 ), and poly(diallyldimethylammonium chloride) (PDADMAC) (MWav: 100 000−200 000, 20 wt % in H2O, d = 1.040) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Ammonium hydroxide (NH4OH, 33%), formic acid (HCOOH, 98− 10092

dx.doi.org/10.1021/ac400934e | Anal. Chem. 2013, 85, 10091−10098

Analytical Chemistry

Article

Figure 1. Overview of the HPLC/TiO2-coated microfluidic-based PCARD/ICPMS hyphenated system.

100%), and sodium borohydride (NaBH4) were obtained from Riedel-de Haën (Seelze, Germany). Titanium dioxide nanoparticles (nano-TiO2, Aeroxide TiO2 P25, average primary particle size: ∼21 nm; specific surface area: 50 ± 15 m2·g−1 (BET)) was purchased from Evonik Industries AG (Essen, Germany). Se(IV) (1000 mg·L−1; 999 ± 5 mg L−1 SeO2 in 0.5 mol·L−1 HNO3) was purchased from Merck (Darmstadt, Germany) and used as received. The TiO2-coating reagent was prepared by mixing an appropriate quantity of nano-TiO2 and colloidal PDADMAC with high-purity water via ultrasonication. The mobile phase for chromatographic separation of inorganic Se species was prepared by dissolving an appropriate quantity of NH4HCO3 and NaCl in high-purity water and then adjusting to the desired pH by using dilute solutions of HNO3 or NH4OH. Prior to use, the solution was filtered through a PTFE membrane (Acrodisc, 0.45 μm, 25 mm O.D., Pall Corp., USA) and degassed via ultrasonication. The hole-scavenger solution was prepared by dissolving concentrated HCOOH in high-purity water and then adjusting to the desired pH by using dilute solutions of HNO3 or NH4OH. A stock solution of Se(VI) was prepared by dissolving Na2SeO4 in high-purity water. All stock solutions were stored in high-density polyethylene bottles wrapped with aluminum foil and stored at 4 °C. Working standards for calibration were prepared by serial dilution of the stock solutions with the mobile phase and used directly. All reagent preparations were performed in a class 100 laminar flow hood. All containers and pipet tips used in this study were cleaned by immersing them overnight in concentrated HNO3, followed by rinsing with high-purity water. The tubes used to connect the components of the apparatus were flushed with high-purity water until all contaminants were eliminated. To avoid additional contamination, fully plastic Norm-Ject syringes (Henke Sass Wolf, GmbH, Tufflingen, Germany) were used throughout this study. Fabrication of the Microfluidic-Based PhotocatalystAssisted Reduction Device (Microfluidic-Based PCARD). Fabrication procedures were similar to that described elsewhere.44 Briefly, the network of the microfluidic-based PCARD was designed using basic geometric modeling software and then patterned on PMMA substrates (Kun Quan Engineering Plastics Co. Ltd., Hsinchu, Taiwan) using a commercial carbon dioxide laser micromachining system (LES-10, Laser Life Co. Ltd., Taipei, Taiwan). Figure S-1, Supporting Information, shows a schematic illustration of the layout of the developed device (150 mm (L) × 28 mm (W) × 4 mm (H)), which consisted of two separated plates: (i) the cover plate and (ii) the bottom plate. The cover plate contained an introduction

port for the column effluent, a reagent-loading channel for the hole-scavenger reagent, and an outlet for the confluent. A corresponding port for the reagent-loading channel was located on the bottom plate, which also contained a serpentine channel employed for conducting the photocatalytic reduction reaction. The channel dimensions were 619 μm wide and 784 μm deep, and the effective reaction channel, defined as the distance from the converged point of the flows of the column effluent and the hole scavenger to the confluent outlet, was 154 cm. Two separate plates were used (instead of one) to avoid structure deformation caused by double punches. The channel features were inspected using a high-resolution optical microscope (FS880ZU, Ching Hsing Computer-Tech Ltd., Taipei, Taiwan). Here, extreme care was taken in handling the substrates to prevent scratching because a damaged surface may lead to a lower light-transmittance and, hence, a lower photocatalytic efficiency. Modification of the Channel Interior of the Microfluidic-Based PCARD with TiO2 Photocatalysts. The channel interior of the microfluidic-based PCARD was modified with TiO2 photocatalysts via dynamic coating procedures (see Figure S-2, Supporting Information). The chip channel was first flushed with saturated NaOH for 12 h. Then, the reagent, containing nano-TiO2 and PDADMAC, was immediately delivered into the reaction channel and incubated for 8 h, after which the channel was flushed with high-purity water and dried under a gentle stream of air. The morphology of the modified channel was inspected by a field-emission scanning electron microscope (FESEM) (SUPRA 60VP, Carl Zeiss AG, Oberkochen, Germany). The quantity of TiO2 coated in the channel interior was evaluated by a laser ablation (LA) system (UP-213, New Wave Research, Inc., Fremont, CA, USA) coupled with ICPMS measurements (Agilent 7500a, Agilent Technologies, Inc., Tokyo, Japan). Test samples for LA analysis were prepared via a process similar to the coating protocol described above except that chip channels were replaced with PMMA substrates. Apparatus and Instrumentation. Experiments were performed using the HPLC/TiO2-coated microfluidic-based PCARD/ICPMS hyphenated system as depicted in Figure 1. An HPLC pump (Model 426, Alltech Associates Inc., Deerfield, IL, USA) equipped with a metal-free six-port injector (Rheodyne, Model 9010, IDEX Corp., Cotati, CA, USA) and a 50 μL poly(aryletherketone) (PEEK) sample loop was employed. Sample filtrates were separated using an analytical column packed with a polymer-based anion-exchange resin (PRP-X100, 10 μm, 250 × 4.1 mm i.d., Hamilton Company, Reno, NV, USA). A peristaltic pump (Minipuls 3, Gilson, 10093

dx.doi.org/10.1021/ac400934e | Anal. Chem. 2013, 85, 10091−10098

Analytical Chemistry

Article

GLS. In the HPLC/nano-TiO2-enhanced photoinduced VG/ ICPMS system, the column effluent, at a flow rate of 1.0 mL min−1, merged with 1.0 g L−1 nano-TiO2 and 100 mM HCOOH solutions at a flow rate of 0.25 mL min−1 under UV irradiation to form volatile Se products, which were carried by a stream of Ar to the ICPMS via the GLS.

Middleton, USA) was used to introduce HCOOH solution into the laboratory-built TiO2-coated microfluidic-based PCARD and to mix the solution with the column effluent. PEEK tubes (Upchurch Scientific, Oak Harbor, WA, USA) were used to connect all the components of the system. After the analytes were separated by HPLC, the Se species in the effluent were delivered directly into the TiO2-coated microfluidic-based PCARD and then reduced in the presence of HCOOH (25 mM) under UV irradiation (UV-A lamp, 40 W, maximum emission at 365 nm, Great Lighting Corp., Taipei, Taiwan). (Caution! Throughout the experiment, the photocatalytic reaction was carried out in an opaque box to block stray UV rays. An appropriate exhaust system is recommended because of the production of ozone during UV irradiation.) After the VG step, the gaseous Se products were carried into the ICPMS system by a stream of Ar for measurement via the GLS. The GLS packed with glass beads was made of glass materials, and the dimensions of the GLS were around 60 mm high and 10 mm i.d.. Adjustment of the sampling position and ion lenses for the optimum signal for Se at m/z 82 was performed using the commercial Se standard solution. Detailed descriptions of the instrumental system and conditions are provided in Table 1.

■

RESULTS AND DISCUSSION Characterization of the TiO2-Coated MicrofluidicBased PCARD. So far, even though considerable effort has been devoted to the development of nano-TiO2 coating methods (e.g., impregnation45,46 or sol−gel methods28,47−49), problems such as complicated and lengthy preparation procedures and unstable modification are still frequently encountered. Therefore, to simplify the coating procedure and improve the stability of the coating materials, a dynamic coating technique was developed in this study. By using this alternative method, TiO2 catalysts were expected to tightly embed on the selected substrate via strong electrostatic attraction between the treated substrate and the modified TiO2 catalyst. Figure 2A displays the SEM image of a channel cross-section of the TiO2-coated microfluidic-based PCARD. The coating material was characterized by energy dispersive Xray analysis (EDAX), which indicated that the PDADMACmodified TiO2 catalyst was indeed embedded in the channel interior. The magnified SEM images in Figure 2B−D further reveal that the numerous clusters of PDADMAC-modified TiO2 catalysts formed a continuous bed inside the channel. Notably, the surface of the continuous bed was fairly rough and the large reaction area on the channel surface, which resulted from the uneven coating, was considered to promote the photocatalyst-assisted reduction reaction. Furthermore, no significant change in the appearance of the coated TiO2 was observed after flushing it with strongly acidic and/or basic reagents (e.g., HNO3 and NaOH), thus demonstrating the TiO2-coated microfluidic-based PCARD to be applicable in practice. Optimization of Operating Conditions. Because the TiO2-coated microfluidic-based PCARD is used for vaporizing inorganic Se species after the chromatographic separation, the composition of the eluent plays a key role in both the separation and VG efficiency of analyte Se species. As reported by Sun et al.,33 the presence of undesired anions such as phosphate ions can retard the generation of volatile Se species. On the other hand, carbonate, chloride, and nitrate ions exert no adverse effects on the analytical signals of both Se(IV) and Se(VI), suggesting that these ions can be employed for chromatographic elution of the two selenicals of interest. To achieve rapid baseline separation for the two selenicals, a binary mobile phase containing 50 mM NH4HCO3 and 30 mM NaCl at pH 7.0 was selected for chromatographic elution in this study. Moreover, to optimize the analytical performance of the HPLC/TiO2-coated microfluidic-based PCARD/ICPMS system, the influences of the operating conditions such as the composition of the column effluent and the presence of HCOOH, as well as the composition of the TiO2-coating reagent on the resulting signals, were evaluated. Influence of the pH on the Conversion Efficiency of Se Species. It has been widely accepted that the adsorption of the selenicals of interest on the surface of nano-TiO2 by electrostatic interactions is critical for the performance of the photocatalyst-assisted reduction reaction.29,50,51 Therefore, the pH of the reaction environment, which may affect the surface

Table 1. Operating Conditions for the HPLC/TiO2-Coated Microfluidic-Based PCARD/ICPMS System HPLC Separation analytical column Hamilton PRP-X100, 10 μm, 250 × 4.1 mm mobile-phase solution 50 mM NH4HCO3 + 30 mM NaCl, pH 7.0 separation flow rate 0.5 mL min−1 sample volume 50 μL TiO2-Coated Microfluidic-Based PCARD dimensions of reaction channel TiO2-coating reagent hole-scavenger reagent resulting mixture for photoreduction

619 μm (W) × 784 μm (D) × 154 cm (L); void volume ∼317 μL 500 mg L−1 nano-TiO2 + 0.5% (w/v) PDADMAC 25 mM HCOOH, 0.5 mL min−1 pH 3.0, 1.0 mL min−1 ICPMS

plasma forward power outer gas flow rate auxiliary gas flow rate carrier gas flow rate sampling cone skimmer cone

1500 W 15 L min−1 Ar 0.9 L min−1 Ar 1.15 L min−1 Ar nickel, 1.0 mm orifice nickel, 0.4 mm orifice

Micromist Nebulizer (MN)/ICPMS, HG/ICPMS and Nano-TiO2-Enhanced Photo-Induced VG/ICPMS Determination. To compare the analytical performance of our system with that of conventional systems, we also analyzed samples using an MN/ICPMS, an HG/ICPMS, and the original nano-TiO2-enhanced photoinduced VG/ICPMS,33 all coupled to the same HPLC system. In the HPLC/MN/ICPMS system, an MN (AR35-1-EM04EX, Glass Expansion, Victoria, Australia) fitted to a Scott-type quartz double-pass spray chamber was used to introduce the column effluent into the ICPMS at a flow rate of 1.0 mL min−1. In the HPLC/HG/ ICPMS system, the column effluent, at a flow rate of 1.0 mL min−1, merged with 1 M HCl and 1% (w/v) NaBH4 solutions at a flow rate of 0.25 mL min−1 to form volatile Se products, which were then carried by a stream of Ar to the ICPMS via the 10094

dx.doi.org/10.1021/ac400934e | Anal. Chem. 2013, 85, 10091−10098

Analytical Chemistry

Article

Figure 2. (A) SEM micrograph of the channel cross-section of the TiO2-coated microfluidic-based PCARD. (B−D) Magnified views of the circular areas denoted in A.

charge of nano-TiO2 for analyte adsorption, was investigated. Figure 3 displays the variation in the analytical signals of inorganic Se species in a binary eluent (50 mM NH4HCO3 and 30 mM NaCl) mixed with HCOOH (25 mM). As indicated in Figure 3, an increasing trend in the intensity of both Se(IV) and Se(VI) appeared when the pH was increased from 2.0 to 3.0, presumably owing to strong interactions between the negatively

charged Se species and the positively charged TiO2 catalyst. In contrast, a dramatic decrease in the signal intensity for both the selenicals was observed when the pH became more than 3.0. This phenomenon can be explained in terms of the increased dissociation of HCOOH,51,52 leading to competitive adsorption between the analyte species and HCOOH on TiO2. On the other hand, the change of charge status of the TiO2 catalyst with the increase in pH also influenced the adsorption of the selenicals of interest on TiO2 and thus decreased the performance of the photocatalyst-assisted reduction reaction. Hence, for optimal quantification, a pH value of 3.0, which provided the highest-intensity signals for Se(IV) and Se(VI), was selected for the subsequent experiments. Influence of HCOOH Concentration on the Conversion Efficiency of Se Species. The presence of HCOOH is considered to improve the photocatalytic reduction efficiency of inorganic Se species because of its relatively high hole-scavenging efficiency.32 To determine the optimal HCOOH concentration for reduction of the Se species, its effect on the analytical signals of the analyte Se species was evaluated. Figure S-3, Supporting Information, displays the variation in the analyte signals in a binary eluent (50 mM NH4HCO3 and 30 mM NaCl) at pH 3.0, as a function of HCOOH concentration. As shown in Figure S-3, Supporting Information, a significant enhancement in the signals of the tested selenicals was obtained when HCOOH was added. The signal intensity of Se(IV) reached a plateau over the concentration range of 25−200 mM, while a distinct decrease in the intensity of Se(VI) was observed when the concentration

Figure 3. Variation in the signal intensity of Se(IV) and Se(VI) with respect to the pH of the reaction environment. All the data were normalized to the maximal value. 10095

dx.doi.org/10.1021/ac400934e | Anal. Chem. 2013, 85, 10091−10098

Analytical Chemistry

Article

speculation was confirmed by LA analysis (see Figure S-4, Supporting Information). To attain the optimal quantification limits, a PDADMAC concentration of 0.5% (w/v) was selected for the subsequent experiments, providing the highest-intensity signals for Se(IV) and Se(VI). Performance of the HPLC/TiO2-Coated MicrofluidicBased PCARD/ICPMS System. The inset of Figure 3 displays a typical chromatogram of Se(IV) and Se(VI) standards under the optimized conditions (presented in Table 1). Table 2

of HCOOH was higher than 25 mM. Because adsorption of Se species onto the surface of TiO2 catalysts is essential for the occurrence of the photocatalyst-assisted VG reaction, the competition between the adsorption of excess formate ions and the Se species to the coated TiO2 was expected to inhibit the generation of volatile Se species.51 Interestingly, the inhibition of Se(VI) signal intensity caused by the competitive adsorption of excess HCOOH was found to be more significant than that of Se(IV). As described by Tan et al.,52 this phenomenon is suggested to result from an increased ionic repulsion of Se(VI) on the surface of TiO2 compared to that of Se(IV), owing to the higher electron density of the four oxygen atoms. To attain the optimal quantification limits for both the selenicals, as a compromise, a HCOOH concentration of 25 mM was adopted for the subsequent experiments. Influence of PDADMAC Concentration on the Conversion Efficiency of Se Species. Because the photocatalyst predominantly determines the rate of photoreduction of Se species,51 the quantity of nano-TiO2 embedded on the channel surface is considered to be an important parameter affecting conversion efficiency of the Se species. Because PDADMAC acts as a mediator for TiO2 deposition, the influence of PDADMAC concentration on the analyte signals of the Se species was investigated. The experiments were conducted using a coating reagent containing various concentration of PDADMAC in the presence of 500 mg·L−1 TiO2 catalyst, followed by assessing the obtained signals of both Se(IV) and Se(VI) for each. Figure 4 displays the variation in the analyte

Table 2. Analytical Characteristics of the Proposed HPLC/ TiO2-Coated Microfluidic-Based PCARD/ICPMS Hyphenation System Se(IV) Se(VI) a

correlation coefficient, R2

detection limit,a μg L−1

RSD,b %

0.9998 0.9999

0.043 0.042

1.3 1.9

Sample volume: 50 μL. bStandard concentration: 0.5 μg L−1; n = 3.

presents the analytical features of merit of the HPLC/TiO2coated microfluidic-based PCARD/ICPMS system operating under optimum conditions. Satisfactory linearities for both the selenicals of interest were observed, with correlation coefficients higher than 0.9995, with procedural detection limits for the determination of Se(IV) and Se(VI) in aqueous solutions being 0.043 and 0.042 μg L−1, respectively (the detection limits were reached based on the 3σ criterion, where the standard deviation was obtained from the results of seven repeated measurements of column effluent mixtures). The precision for each analyte, based on three replicate injections of 0.5 μg L−1 samples of each species and measurement of the peak areas, was better than 2.0% of relative standard deviation (RSD). Because the plasma might be disturbed by the introduction of the gaseous sample, the stability of the online HPLC/TiO2-coated microfluidic-based PCARD/ICPMS system was evaluated by performing a continuous 12 h measurement for both Se(IV) and Se(VI). The repeatability of the continuous 12 h measurement for the test selenicals was less than 6% of the coefficient of variation (CV), demonstrating the practical utility of this method. The applicability of our thus-established system was examined by determining the Se species of interest in the certified reference material (NIST 1643e, artificial saline solution, National Institute for Standards and Technology, Gaithersburg, MD, USA) and irrigation water samples collected from a local region. Because no certified value for the individual Se species in the reference material was available, the analytical reliability of our proposed method could only be validated by comparing the total Se concentration obtained (by summation of the individually determined Se species concentration) with that of the certified standard. Table 3 indicates that the analytical results agreed reasonably well with the certified total Se concentration, with the Se(IV) content being considerably higher than the Se(VI) content in the reference material. The applicability of the developed method was further assessed by analyzing irrigation water samples that had been spiked with different quantities of Se(IV) and Se(VI). As shown in Table 3, Se species that naturally existed in the irrigation water could be clearly identified by our system even though the concentration of both the selenicals of interest was below the μg L−1 level. In addition, the acceptable recovery obtained, ranging between 96% and 106%, for both the spiked samples indicated that our

Figure 4. Variation in the signal intensity of Se(IV) and Se(VI) with respect to the concentration of PDADMAC in TiO2-coating reagents. All the data were normalized to the maximal value.

signals in a binary eluent (50 mM NH4HCO3 and 30 mM NaCl) mixed with HCOOH (25 mM) at pH 3.0. The analytical signals of both Se(IV) and Se(VI) increased upon increasing the PDADMAC concentration from 0.01 to 0.5% (w/v), suggesting that a sufficient quantity of embedded nano-TiO2 was available on the channel surface to promote the formation of volatile products. However, a dramatic decrease in the signal intensity was observed when PDADMAC concentration exceeded 0.5% (w/v). The competition between PDADMAC and PDADMAC-modified TiO2 catalysts for the channel surface was expected to reduce the quantity of nano-TiO2 embedded in the channel interior, leading to a poor conversion of either of the Se species into gaseous products. This 10096

dx.doi.org/10.1021/ac400934e | Anal. Chem. 2013, 85, 10091−10098

Analytical Chemistry

Article

Table 3. Analysis of Water Samples observed value, μg L−1 sample NIST SRM 1643e irrigation water spiked sample A spiked sample B a

certified value, μg L

−1

11.97 ± 0.14 0.50/0.30b 0.30/0.50b

Se(IV)

Se(VI)

11.56 ± 0.90 0.109 ± 0.019 0.639 0.420

a

ND 0.090 ± 0.013 0.390 0.570

spike recovery, % 94 106/100c 104/96c

ND: Not detected. bAdded concentration: Se(IV)/Se(VI). cSe(IV)/Se(VI).

integrated into a microfluidic device, while the results furthermore demonstrated that unique characteristics such as short reaction time and integration for multifunctionality benefit from the employment of the microfluidic-based configuration. Finally, the TiO2-coated microfluidic-based PCARD prepared from PMMA was also more cost-effective with regard to fabrication costs than conventional photoreactors (approximately 0.5 USD per TiO2-coated microfluidicbased PCARD). Because excessive consumption of photocatalysts resulting from consecutive loading was absent, the developed microfluidic-based PCARD coated with nano-TiO2 can also be considered to fulfill the goals of green nanotechnology.54,55

developed procedure could accurately determine Se(IV) and Se(VI) concentration in the environmental water samples. Comparison of Various Sample Introduction Systems. Finally, we compared the analytical features of the TiO2-coated microfluidic-based PCARD with three other sample introduction systems: MN, HG, and the nano-TiO2-enhanced photoinduced VG using PTFE tubing as a reactor.33 Figure 5

■

CONCLUSIONS We developed a selective and sensitive hyphenated system employing a microfluidic-based VG system in conjunction with HPLC separation and ICPMS detection for the determination of trace inorganic Se species. Because PMMA provides excellent characteristics with respect to optical analysis, the VG system exploited high optical-grade PMMA substrates to fabricate a microfluidic-based PCARD. Moreover, to reduce the consumption of photocatalysts during analytical procedures, a microfluidic-based PCARD coated with TiO2 catalysts was employed to avoid the consecutive loading of the device. To simplify the coating procedure and improve the stability of the coating materials, a dynamic coating technique was developed. Because the operational functionalities such as catalyst loading and tangle-prone tubing (as utilized in the original nano-TiO2enhanced photoinduced VG system) were successfully integrated into a microfluidic device, the proposed technique was convenient not only in terms of system construction but also in its operation. By coupling HPLC separation and ICPMS detection with the TiO2-coated microfluidic-based PCARD, the hyphenated system was made suitable for the determination of trace inorganic Se species in the environmental water samples. On the basis of the analytical results, the HPLC/TiO2-coated microfluidic-based PCARD/ICPMS system efficiently volatilized Se(IV) and Se(VI) and demonstrated excellent reliability in the Se speciation of the environmental water samples. Furthermore, the developed microfluidic-based PCARD coated with nano-TiO2 is considered to fulfill the goals of green nanotechnology as the consumption of photocatalysts was significantly reduced. To the best of our knowledge, this is the first study that combines a microfluidic-based platform with an online speciation system for highly sensitive and selective determination of inorganic Se species.

Figure 5. Signal intensity of Se(IV) and Se(VI) obtained using various sample introduction systems. All the data were normalized to the maximal value.

illustrates the analytical sensitivity achieved by each system. As indicated in Figure 5, the signals of Se(IV) and Se(VI) obtained by the MN system were rather small because of the low transport efficiency.53 Using the conventional HG system, a significant improvement in Se(IV) signal intensity was achieved, whereas the volatile Se species reduced from Se(VI) were still absent. As for the other two photocatalytic techniques, both the selenicals of interest were efficiently vaporized, and a more than 70-fold enhancement of the signal intensity for both Se(IV) and Se(VI) was obtained using the TiO2-coated microfluidic-based PCARD, compared to that obtained using the MN system. In other words, the use of the TiO2-coated microfluidic-based PCARD, in conjunction with HPLC and ICPMS, not only addresses the shortcomings of conventional NaBH4-based HG techniques for the VG of Se(VI) but also dramatically improves the analytical sensitivity of ICPMS toward the detection of Se. Regardless of the fact that the reaction area of the TiO2 catalysts embedded in the channel interior is much lower than that in suspensions (used for nano-TiO2-enhanced photoinduced VG), efficient VG of analyte Se species via the microfluidic-based photoreactor could be rapidly accomplished within 15 s. The need for catalyst loading and tangle-prone tubing, as utilized in the nano-TiO2enhanced photoinduced VG system, was successfully eliminated because the operational procedures and accessories were

■

ASSOCIATED CONTENT

S Supporting Information *

Figures S-1−S-4. This material is available free of charge via the Internet at http://pubs.acs.org. 10097

dx.doi.org/10.1021/ac400934e | Anal. Chem. 2013, 85, 10091−10098

Analytical Chemistry

■

Article

(29) Guo, X.; Sturgeon, R. E.; Mester, Z.; Gardner, G. J. Anal. Chem. 2003, 75, 2092−2099. (30) Zheng, C.; Li, Y.; He, Y.; Ma, Q.; Hou, X. J. Anal. At. Spectrom. 2005, 20, 746−750. (31) Guo, X.; Sturgeon, R. E.; Mester, Z.; Gardner, G. J. Anal. Chem. 2004, 76, 2401−2405. (32) He, Y.; Hou, X.; Zheng, C.; Sturgeon, R. E. Anal. Bioanal. Chem. 2007, 388, 769−774. (33) Sun, Y.-C.; Chang, Y.-C.; Su, C.-K. Anal. Chem. 2006, 78, 2640− 2645. (34) Zheng, C.; Wu, L.; Ma, Q.; Lv, Y.; Hou, X. J. Anal. At. Spectrom. 2008, 23, 514−520. (35) Wang, Q.; Liang, J.; Qiu, J.; Huang, B. J. Anal. At. Spectrom. 2004, 19, 715−716. (36) Teflon AF Properties; see: http://www2.dupont.com/Teflon_ Industrial/en_US/products/product_by_name/teflon_af/properties. html (accessed May 2012). (37) Liu, C.; Tang, X.; Mo, C.; Qiang, Z. J. Solid State Chem. 2008, 181, 913−919. (38) Li, H.; Luo, Y.; Li, Z.; Yang, L.; Wang, Q. Anal. Chem. 2012, 84, 2974−2981. (39) Erickson, D.; Li, D. Anal. Chim. Acta 2004, 507, 11−26. (40) Tabeling, P. Introduction to Microfluidics; Oxford: New York, 2005. (41) Fiorini, G. S.; Chiu, D. T. Biotechniques 2005, 38, 429−446. (42) Shadpour, H.; Musyimi, H.; Chen, J.; Soper, S. A. J. Chromatogr., A 2006, 1111, 238−251. (43) Tan, H. Y.; Loke, W. K.; Nguyen, N.-T. Sens. Actuators, B: Chem. 2010, 151, 133−139. (44) Shih, T.-T.; Chen, W.-Y.; Sun, Y.-C. J. Chromatogr., A 2011, 1218, 2342−2348. (45) Tsai, M.-W.; Sun, Y.-C. Rapid Commun. Mass Spectrom. 2008, 22, 211−216. (46) Shih, T.-T.; Tseng, W.-Y.; Tsai, K.-H.; Chen, W.-Y.; Tsai, M.W.; Sun, Y.-C. Microchem. J. 2011, 99, 260−266. (47) Hu, Y.; Yuan, C. J. Mater. Sci. Technol. 2006, 22, 239−244. (48) Rizzo, L.; Koch, J.; Belgiorno, V.; Anderson, M. A. Desalination 2007, 211, 1−9. (49) Yin, Y.; Liang, J.; Yang, L.; Wang, Q. J. Anal. At. Spectrom. 2007, 22, 330−334. (50) Tan, T. T. Y.; Beydoun, D.; Amal, R. J. Photochem. Photobiol., A: Chem. 2003, 159, 273−280. (51) Tan, T. T. Y.; Beydoun, D.; Amal, R. J. Mol. Catal. A: Chem. 2003, 202, 73−85. (52) Tan, T. T. Y.; Zaw, M.; Beydoun, D.; Amal, R. J. Nanopart. Res. 2002, 4, 541−552. (53) Jarvis, K. E.; Gray, A. L.; Houk, R. S. Handbook of Inductively Coupled Plasma Mass Spectrometry; Chapman and Hall: New York, 1992. (54) Albrecht, M. A.; Evans, C. W.; Raston, C. L. Green Chem. 2006, 8, 417−432. (55) Smith, G. B.; Granqvist, C. G. Green Nanotechnology: Solutions for Sustainability and Energy in the Built Environment; CRC Press: Boca Raton, 2010.

AUTHOR INFORMATION

Corresponding Author

*Tel: +886-3-5715131 ext 35596. Fax: +886-3-5723883. Email: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

■

REFERENCES

The authors would like to convey their gratitude to Prof. MoHsiung Yang for his advice. We are also grateful for the financial support provided by the National Science Council of the Republic of China (Taiwan).

(1) Chapman, P. M.; Adams, W. J.; Brooks, M. L.; Delos, C. G.; Luoma, S. N.; Maher, W. A.; Ohlendorf, H. M.; Presser, T. S.; Shaw, D. P. Ecological assessment of selenium in the aquatic environment: Summary of a Society of Environmental Toxicology and Chemistry (SETAC) Pellston Workshop; SETAC: Pensacola, 2009. (2) Fordyce, F. Ambio 2007, 36, 94−97. (3) Torres, J.; Pintos, V.; Domínguez, S.; Kremer, C.; Kremer, E. J. Solution Chem. 2010, 39, 1−10. (4) Chand, V.; Prasad, S. J. Hazard. Mater. 2009, 165, 780−788. (5) National Research Council. Mineral tolerance of animals, 2nd ed.; National Academies Press: Washington DC, 2005. (6) Tonietto, G. B.; Godoy, J. M.; Oliveira, A. C.; de Souza, M. V. Anal. Bioanal. Chem. 2010, 397, 1755−1761. (7) Mulugeta, M.; Wibetoe, G.; Engelsen, C. J.; Lund, W. J. Chromatogr., A 2010, 1217, 6186−6194. (8) Chen, B.; He, M.; Mao, X.; Cui, R.; Pang, D.; Hu, B. Talanta 2011, 83, 724−731. (9) Gammelgaard, B.; Bendahl, L.; Sidenius, U.; Jøns, O. J. Anal. At. Spectrom. 2002, 17, 570−575. (10) Llorente, I.; Gómez, M.; Cámara, C. Spectrochim. Acta, Part B 1997, 52, 1825−1838. (11) Wang, J.; Hansen, E. H. Trends Anal. Chem. 2003, 22, 836−846. (12) Sturgeon, R. E.; Guo, X.; Mester, Z. Anal. Bioanal. Chem. 2005, 382, 881−883. (13) Butler, O.; Evans, H.; Fisher, A.; Hill, S.; Harrington, C.; Taylor, A.; West, M.; Ellis, A. J. Anal. At. Spectrom. 2010, 25, 1546−1566. (14) Pohl, P. Trends Anal. Chem. 2004, 23, 87−101. (15) D’Ulivo, A.; Dědina, J.; Mester, Z.; Sturgeon, R. E.; Wang, Q.; Welz, B. Pure Appl. Chem. 2011, 83, 1283−1340. (16) Moreno, M. E.; Pérez-Conde, C.; Cámara, C. Anal. Bioanal. Chem. 2003, 375, 666−672. (17) Bye, R. Talanta 1983, 30, 993−996. (18) Chung, C.-H.; Iwamoto, E.; Yamamoto, M.; Yamamoto, Y. Spectrochim. Acta, Part B 1984, 39, 459−466. (19) Marcucci, K.; Zamboni, R.; D’Ulivo, A. Spectrochim. Acta, Part B 2001, 56, 393−407. (20) D’Ulivo, A.; Gianfranceschi, L.; Lampugnani, L.; Zamboni, R. Spectrochim. Acta, Part B 2002, 57, 2081−2094. (21) Qiu, J.; Wang, Q.; Ma, Y.; Yang, L.; Huang, B. Spectrochim. Acta, Part B 2006, 61, 803−809. (22) Burguera, J. L.; Carrero, P.; Burguera, M.; Rondon, C.; Brunetto, M. R.; Gallignani, M. Spectrochim. Acta, Part B 1996, 51, 1837−1847. (23) Uggerud, H.; Lund, W. J. Anal. At. Spectrom. 1995, 10, 405−408. (24) Pahlavanpour, B.; Pullen, J. H.; Thompson, M. Analyst 1980, 105, 274−278. (25) Brindle, I. D.; Lugowska, E. Spectrochim. Acta, Part B 1997, 52, 163−176. (26) Chen, Y.-W.; Zhou, M.-D.; Tong, J.; Belzile, N. Anal. Chim. Acta 2005, 545, 142−148. (27) Cutter, G. A. Anal. Chim. Acta 1978, 98, 59−66. (28) Liang, J.; Wang, Q.; Huang, B. Anal. Bioanal. Chem. 2005, 381, 366−372. 10098

dx.doi.org/10.1021/ac400934e | Anal. Chem. 2013, 85, 10091−10098