An Ion-Imprinted Functionalized Silica Gel Sorbent Prepared by a

Bhavya Srivastava , Milan K. Barman , Mousumi Chatterjee , Bhabatosh Mandal ..... Visual monitoring and removal of divalent copper, cadmium, and mercu...
0 downloads 0 Views 149KB Size
Download PDF
Anal. Chem. 2005, 77, 1734-1739

An Ion-Imprinted Functionalized Silica Gel Sorbent Prepared by a Surface Imprinting Technique Combined with a Sol-Gel Process for Selective Solid-Phase Extraction of Cadmium(II) Guo-Zhen Fang,†,‡ Jin Tan,† and Xiu-Ping Yan*,†

Research Center for Analytical Sciences, Department of Chemistry, Nankai University, Tianjin 300071, China and Department of Chemistry, Yanbei Normal College, Datong 03700, China

A new ion-imprinted thiol-functionalized silica gel sorbent was synthesized by a surface imprinting technique in combination with a sol-gel process for selective on-line, solid-phase extraction of Cd(II). The Cd(II)-imprinted thiol-functionalized silica sorbent was characterized by FTIR, the static adsorption-desorption experiment, and the dynamic adsorption-desorption method. The maximum static adsorption capacity of the ion-imprinted functionalized sorbent was 284 µmol g-1. The largest selectivity coefficient for Cd(II) in the presence of Pb(II) was over 220. The static uptake capacity and selectivity coefficient of the ion-imprinted functionalized sorbent are higher than those of the nonimprinted sorbent. The breakthrough capacity and dynamic capacity of the imprinted functionalized silica gel sorbent for 4 mg L-1 of Cd(II) at 5.2 mL min-1 of sample flow rate were 11.7 and 64.3 µmol g-1, respectively. No remarkable effect of sample flow rate on the dynamic capacity was observed as the sample flow rate increased from 1.7 to 6.8 mL min-1. The imprinted functionalized silica gel sorbent offered a fast kinetics for the adsorption and desorption of Cd(II). The prepared ionimprinted functionalized sorbent was shown to be promising for on-line, solid-phase extraction coupled with flame atomic absorption spectrometry for the determination of trace cadmium in environmental and biological samples. All competitive ions studied did not interfere with the determination of Cd(II). With a sample loading flow rate of 8.8 mL min-1 for 45-s preconcentration, an enhancement factor of 56, and a detection limit (3σ) of 0.07 µg L-1 were achieved at a sampling frequency of 55 h-1. The precision (RSD) for 11 replicate on-line sorbent extractions of 8 µg L-1 Cd(II) was 0.9%. The sorbent also offered good linearity (r ) 0.9997) for on-line, solid-phase extraction of trace Cd(II). Heavy metal ion contamination represents a significant threat to the ecosystem and especially to people due to the severe * Corresponding author. Fax: (86)22-23506075. E-mail: [email protected]. † Nankai University. ‡ Yanbei Normal College.

1734 Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

toxicological effects on living organisms.1,2 Cadmium is one of the most toxic heavy metal elements for animals and humans, even at low concentrations. Cd(II) is listed as the sixth most poisonous substance jeopardizing human’s health.3 Through the food chain system of soil-plant-animal-human, Cd(II) is transferred into animals and human beings, causing severe contamination.4 Consequently, the development of reliable methods for the removal and determination of cadmium in environmental and biological samples is of particular significance. Selective removal of toxic metal ions from aqueous solutions is usually achieved by solvent extraction and solid-phase extraction. For metal extractions from dilute solutions, the adsorption technique has greater applicability than traditional solvent extraction processes.5 An efficient adsorbing material should consist of a stable and insoluble porous matrix having suitable active groups (typically organic groups) that interact with heavy metal ions. Silica gel is an ideal support for organic groups because it is an inorganic material, stable under acidic conditions and nonswelling, and has high mass exchange characteristics and very high thermal resistance.6 The immobilization of thiol ligand onto the surface of silica gel has been reported.7,8 The effectiveness of such materials in binding metal ions has been attributed to the complexation chemistry between the ligand and the metal. The specificity of a particular ligand toward target metal ions is the result of a conventional acid-base interaction between the ligand and the metal ion. Although some of these thiol-functionalized adsorbents can exhibit specific interactions with soft Lewis acids (such as Hg(II), Cd(II), Cu(II), and Ag(I)), the selectivity of these materials is usually unremarkable because many metals have the ability to bind with thiol ligands without consideration of the (1) Hultberg, B.; Andersson, A.; Isaksson, A. . Toxicology 1998, 126 (3), 203212. (2) Antochshuk, V.; Jaroniec, M. Chem. Commun. 2002, 258-259. (3) IARC Monographs on the evaluation of carcinogenic risks to humans. International Agency for Research on Cancer: Lyon, France 1993, Vol. 58, pp 41-117. (4) Liu, C. H.; Li, Y. L.; Zhang, D. G.; Zhu, H. G.; Yang, Y. Chin. J. Soil Sci. 2003, 34 (4), 326-329. (5) Lee, J. S.; Gomez-Salazar, S.; Tavlarides, L. L. React. Funct. Polym. 2001, 49, 159-172. (6) Jal, P. K.; Patel, S.; Mishra, B. K. Talanta 2004, 62, 1005-1028. (7) Howard, A. G. Analyst 1987, 112, 159-162. (8) Vieira, E. F. S.; Simony, J. A.; Airoldi, C. J. Mater. Chem. 1997, 7, 22492252. 10.1021/ac048570v CCC: $30.25

© 2005 American Chemical Society Published on Web 02/02/2005

stereochemical interactions between the ligand and metal ion. Molecular imprinting has become a powerful method for the preparation of robust materials that have the ability to recognize a specific chemical species.9 Molecular imprinting involves arranging monomers around a template molecule so that complexes between the monomer and template molecules are formed. Subsequent polymerization of the monomer molecules results in trapping template molecules in a highly cross-linked amorphous polymer matrix. Extraction of the imprint molecules leaves a predetermined arrangement of ligands and a tailored binding pocket.10 Such an imprinted polymer shows an affinity for the template molecule over other structurally related compounds. For metal ions, molecular imprinting can be interpreted as ionic imprinting exactly. Surface molecular imprinting is one of the important types of molecular imprinting. Surface molecularly imprinted polymer not only possesses high selectivity but also avoids problems with mass transfer.11 In this study, a simple procedure was developed to synthesize a new ion-imprinted thiol-functionalized silica gel sorbent by combining a surface molecular imprinting technique with a solgel process for selective separation and preconcentration of Cd(II). The characterization of this imprinted thiol-functionalized silica gel sorbent and its applicability to flow injection (FI) on-line selective solid-phase extraction of Cd(II) are described and discussed in detail. EXPERIMENTAL SECTION Materials and Chemicals. Silica gel (80-120 mesh, Qingdao Ocean Chemical Co., Qingdao, China) was used as the support to prepare the ion-imprinted functionalized sorbent. All reagents used were of at least analytical grade. 3-Mercaptopropyltrimethoxysilane (MPS, Wuhan University Chemical Factory, Wuhan, China) was used in this study. All metal stock solutions (1000 mg L-1) were purchased from the National Research Center for Standard Materials (NRCSM, Beijing, China). The working solutions were prepared by series dilution of the stock solutions immediately prior to their use. Doubly deionized water (DDW, 18.2 MΩ cm-1) obtained from a WaterPro water system (Labconco Corporation, Kansas City, MO) was used throughout the experiments. The pH of the solution was adjusted using the following buffers: hydrochloric acid for pH ) 1, sodium acetate/hydrochloric acid for pH ) 2-3, and sodium acetate/acetic acid for pH ) 4-7.5. Instrumentation. A model SOLAAR S2 atomic absorption spectrometer equipped with quadline deuterium arc background correction, a universal air-cooled titanium burner, and a PTFE spray chamber with impact bead and baffle Pt/Ir PTFE nebulizer (Thermo Elemental, Franklin, MA) was used for all measurements of metal ion concentrations. The AAS instrument was controlled by the SOLAARS operation software. Hollow cathode lamps (Beijing Shuguangming Electronic Light Source Instrument Co. Ltd., Beijing) were used as the radiation sources at 228.8 (Cd) or 283.3 nm (Pb), respectively, with a current of 2.5 (Cd) or 4 mA (Pb) and a 0.5-nm slit width (Pb and Cd). The recommended (9) Yang, H. H.; Zhang, S. Q.; Yang, W.; Chen, X. L.; Zhuang, Z. X.; Xu, J. G.; Wang, X. R. J. Am. Chem. Soc. 2004, 126, 4054-4055. (10) Zhang, Z.; Dai, S.; Hunt, R. D.; Wei, Y.; Qiu, S. Adv. Mater. 2001, 13, 493496. (11) Nicholls, I. A.; Rosengren, J. P. Bioseparation 2002, 10, 301-305.

acetylene flow rates of 1.1 L min-1 for Pb and 0.9 L min-1 for Cd were employed. The air flow rates were automatically adjusted to meet the stoichiometric air-acetylene flame conditions. A model FIA-3100 flow injection analyzer (Vital Instrumental Co. Ltd., Beijing, China) was used to evaluate the applicability of the developed sorbent to on-line, solid-phase extraction separation and preconcentration of cadmium. The FIA-3100 consists of two peristaltic pumps and a standard rotary injection valve (8-channel 16-port multifunctional injector). Tygon peristaltic pump tubings were employed to propel the sample and reagent. PTFE tubing with 0.5 mm i.d. was used for all connections. These connections were kept as short as possible to minimize the dead volume. FT-IR spectra (4000-400 cm-1) in KBr were recorded using a Magna-560 spectrometer (Nicolet, U.S.A.). A Qwave-2000 model microwave digestion system (Questron Co., Canada) was employed to digest the samples used for the validation of the developed on-line separation and preconcentration method. Preparation of the Cd(II)-Imprinted Thiol-Functionalized Silica Gel Sorbent. To activate the silica gel surfaces, 8 g of silica gel (80-120 mesh) was mixed with 60 mL of 33% methanesulfonic acid and refluxed under stirring for 8 h. The solid product was recovered by filtration, washed with DDW to neutral, and dried under vacuum at 70 °C for 8 h. To prepare the Cd(II)-imprinted thiol-functionalized silica gel sorbent, 1.92 g of CdCl2‚2.5H2O was dissolved in 80 mL of methanol under stirring and heating and then mixed with 4 mL of MPS. The solution was stirred and refluxed for 1 h, to which 6 g of activated silica gel was added. After 20 h of stirring and refluxing the mixture, the product was recovered by filtration, washed with ethanol, and stirred in 50 mL of 6 mol L-1 hydrochloric acid for 2 h. The final product was recovered by filtration, washed with DDW up to the eluent pH ) 4-5 and dried under vacuum at 80 °C for 12 h. For comparison, the nonimprinted functionalized silica gel sorbent was also prepared using an identical procedure, but without the addition of CdCl2‚2.5H2O. Static Adsorption Test. The time-dependent uptake test indicates that 30 min is sufficient for adsorption. The effect of pH on the static adsorption of Cd(II) was tested by equilibrating 100 mg of Cd(II)-imprinted thiol-functionalized silica gel sorbent with 10 mL of the buffer solutions containing 450 µmol L-1 of Cd(II) under different pH conditions. To measure the static adsorption capacity, 100 mg of Cd(II)-imprinted or nonimprinted sorbent was equilibrated with 10 mL of various concentrations of Cd(II) solutions buffered with sodium acetate/acetic acid solution at pH ) 5.5. Competitive loading of Cd(II) and Pb(II) by Cd(II)-imprinted and nonimprinted sorbents was measured at pH ) 5.5 buffered with sodium acetate/acetic acid solution. A 100-mg portion of Cd(II) imprinted or nonimprinted sorbents was equilibrated with 10 mL of the buffered solutions containing 450 µmol L-1 of Cd(II) and 450 µmol L-1/900 µmol L-1 of Pb(II), respectively. In all the above batch experiments, the mixtures were mechanically shaken for 30 min at room temperature, then centrifugally separated. The supernatants were measured for unextracted Cd(II) or Pb(II) by flame atomic absorption spectrometry (FAAS). Measurement of the Dynamic Capacity. The following procedure was employed to measure the dynamic capacity of the Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

1735

waste. In step 2 (Figure 1B), pump 2 started to work, whereas pump 1 was stopped, and the injection valve was turned to the inject position to introduce diluted HCl solution at a flow rate of 3.6 mL min-1 for 15 s to elute the analyte retained on the column. In step 3 (Figure 1C), pump 2 worked while pump 1 was stopped, and the injection valve was turned to the fill position so that DDW was introduced at a flow rate of 7.6 mL min-1 for 5 s to wash the microcolumn. A complete cycle of the preconcentration and elution required 65 s. Sample Preparation. The following certified reference materials (NRCSM) were analyzed to check the accuracy of the developed FI on-line, solid phase extraction FAAS using the developed ion imprinted functionalized silica gel as the packing material: GBW08571 (mussel), GBW08511 (rice flour), GBW07605 (tea), GBW07601 (human hair), and GBW07405 (soil). Real samples of river water, wastewater, and soil were collected locally. Immediately after sampling, the water samples were filtered and adjusted to the optimal pH range with ammonium hydroxide or acetic acid solution. A 0.6000-1.0000-g portion of the solid samples was mixed with 6-12 mL of concentrated HNO3 or in combination with 2 mL hydrofluoric acid (for soil) and digested in sealed PFA (Teflonperfluoralkoxy) vessels using a Qwave-2000 model microwave digestion system (Questron Co., Canada). All instrumental parameters for the digestion were chosen according to the recommendations of the EPA. The clear digest was adjusted to a pH of 5-6 with ammonium hydroxide solution, transferred into a 100mL calibrated flask, and diluted to volume with sodium acetate/ acetic acid buffer solution of pH ) 5.5. Figure 1. FI manifold and operational sequence for the on-line, solid-phase extraction coupled with FAAS. P1 and P2, peristaltic pump; MC, microcolumn; W, waste.

imprinted functionalized silica gel sorbent. A fixed concentration of Cd(II) solution was pumped at a constant flow rate through a cylindrically shaped PTFE microcolumn (2-cm long × 3-mm i.d.) packed with 46 mg of the imprinted functionalized silica gel sorbent. The concentration of Cd(II) in the effluent from the exit of the microcolumn was on-line monitored by FAAS until the cadmium absorbance of the effluent was stable. The adsorbed Cd(II) on the column was then eluted with 0.8 mol L-1 HCl, collected, and diluted to 100 mL for subsequent analysis by conventional FAAS, since the concentration of eluted Cd(II) was too high to be quantified on-line. Evaluation of the Imprinted Functionalized Silica Gel Sorbent for Selective On-Line, Solid-Phase Extraction of Cd(II). To evaluate the applicability of the imprinted functionalized silica gel sorbent for on-line, solid phase extraction of trace Cd(II), a hyphenated technique, namely, flow injection on-line microcolumn preconcentration and separation coupled with FAAS using a cylindrically shaped microcolumn (2-cm long × 3-mm i.d.) packed with 46 mg of the imprinted functionalized silica gel sorbent, was employed. The FI manifold and its operation sequence for the on-line, solid-phase extraction are shown in Figure 1. In step 1 (Figure 1A), the injection valve was in the fill position, and pump 1 was activated so that the sample solution was loaded onto the microcolumn at a flow rate of 8.8 mL min-1 for 45 s. In this step, the metal ions were absorbed onto the microcolumn while the effluent from the column was flowing to 1736

Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

RESULTS AND DISCUSSION Preparation of the Cd(II)-Imprinted Thiol-Functionalized Silica Gel Sorbent. Silica gel is an amorphous inorganic polymer having siloxane groups (Si-O-Si) in the bulk and silanol groups (Si-OH) on its surface. The latter are responsible for chemical modifications that may occur on the silica surface. Because commercial silica gel possesses a low concentration of surface silanol groups suitable for modification, the activation of silica gel surfaces is necessary. Methanesulfonic acid was used for the activation of the silica gel because of its strong catalytic action; high boiling point, which prevents it from contaminating the air as HCl;12 and its ability to be used repeatedly.13 The effect of the concentration of methanesulfonic acid was tested. As the concentration of methanesulfonic acid increased from 5 to 90%, the reflux temperature required increased from 100 to 123 °C, whereas the uptake percentage of the final functionalized silica gel sorbent for Cd(II) did not change significantly. In this work, 33% methanesulfonic acid solution was chosen for activation. The complex was formed between Cd(II) and MPS, then cohydrolyzed and co-condensed with the activated silica gel. Thus, the activated silica gel surface was grafted with the complex of Cd(II) and MPS rather than just the free MPS. After the remnant MPS and Cd(II) were removed by ethanol and 6 mol L-1 HCl, respectively, the imprinted functionalized silica gel sorbent which contained a tailor-made cavity for Cd(II) was formed. The ratio of MPS to the activated silica gel is directly related to the moles (12) Walter, A. A.; Corazo, R. H. J. Chromatogr. 1969, 42, 319-335. (13) Liu, L. J.; Zhuo, R. X. Ion Exch. Adsorpt. 1995, 11, 541-544.

Table 1. Competitive Loading of Cd(II) and Pb(II) by the Cd(II)-Imprinted and Nonimprinted Silica Gel Sorbent initial solution/µmol L-1

% uptake

sorbent

Cd(II)

Pb(II)

Cd(II)

Cd(II)-imprinted

450 450 450 450 450 450

0 450 900 0 450 900

99.2 98.7 98.4 98.0 97.3 96.7

nonimprinted

Pb(II) 32.6 21.9 41.6 29.7

capacity/µmol g-1 Cd(II) 44.6 44.4 44.3 44.1 43.8 43.5

Pb(II) 14.7 19.7 18.7 26.7

Kda/mL g-1 Cd(II) 12400 7592 6150 5002 3604 2930

Pb(II) 48 28 71 42

kb 158 220

k′c 3.11 3.15

50.8 69.8

a K , distribution coefficient, K ) {(C - C )/C } × {volume of solution (mL)}/{mass of gel (g)}, where C and C represent the initial and final d d i f f i f solution concentrations, respectively. b k, selectivity coefficient, Kd(Cd)/Kd(Pb). c k′, relative selectivity coefficient, k′ ) kimprinted/knonimprinted.

of the active groups with respect to the weight of the functionalized silica gel sorbent. It was found that 1 mL MPS/1.5 g activated silica gel was the optimal proportion, so this proportion was used in the present work. Characteristic of the FT-IR Spectra. To ascertain the presence of MPS in the functionalized silica gel sorbents, FT-IR spectra were obtained from activated silica gel, Cd(II)-imprinted and nonimprinted thiol-functionalized silica gel sorbents. The observed features around 1104.6 and 969.4 cm-1 indicate Si-OSi and Si-O-H stretching vibrations, respectively. The presence of adsorption water was reflected by νOH vibration at 3448.6 and 1637.0 cm-1. The bands around 800.0 and 476.0 cm-1 resulted from Si-O vibrations. A characteristic feature of the imprinted and nonimprinted sorbents when compared with activated silica gel is a S-H bond around 2560 cm-1. It shows that -SH was grafted onto the surface of the activated silica gel after modification. Imprinted and nonimprinted sorbent showed a very similar location and appearance of the major bands. It indicates that -SH was recovered after removal of Cd(II) in the imprinted sorbent. Evaluation of Static Adsorption. Study of the effect of pH on the static adsorption showed that the percentage of the Cd(II) extracted increased as the pH of the aqueous solution increased from 3 to 5, then remained constant with further increase in pH from 5 to 7.5. Below pH 3, no Cd(II) was extracted onto the imprinted functionalized sorbent mainly due to the diminution of -S- to be involved in chelate formation with the Cd(II) in aqueous solution as a result of the protonation of the thiol moiety. The optimum pH for the extraction of Cd(II) from aqueous solution ranged from 5 to 7.5. In this pH range, neither precipitation of the metal hydroxide nor the protonation of the thiol group is expected. Uptake kinetics of Cd(II) by the imprinted functionalized silica gel sorbent was examined. The results indicate that the solidphase extraction process of the imprinted sorbent is fairly rapid: 90% of Cd(II) was adsorbed by the imprinted sorbent within 2 min. The fast adsorption kinetics of the imprinted sorbent is an obvious advantage for its application in on-line, solid-phase extraction. The static adsorption capacity of the imprinted and nonimprinted sorbent for Cd(II) was 284 and 155 µmol g-1, respectively, whereas the unfunctionalized activated silica exhibited no adsorption for Cd(II). Obviously, the capacity of imprinted sorbent is larger than that of nonimprinted sorbent. Adsorption and competitive ion recognition studies were performed with Cd(II) and Pb(II) ions in order to measure the selectivity of the imprinted sorbent. The Pb(II) ion was chosen

as the competitive species with Cd(II) because these two ions have the same charge and similar ionic radius. Moreover, both ions bind well with the thiol ligand and behave as a geochemical pair.14 Table 1 summarizes the data for the percentage of the metal ion adsorbed, uptake capacity, distribution coefficient (Kd), selectivity coefficient of the sorbent toward Cd(II) (k), and the relative selectivity coefficient (k′) obtained in these competitive ion-binding experiments between Pb(II) and Cd(II) ions. Comparison of the k values for the imprinted sorbent with the corresponding nonimprinted sorbent reveals a significant increase in k for Cd(II) through imprinting. During the preparation of the imprinted sorbent, the presence of Cd(II) made the ligands arrange orderly. After the removal of Cd(II), the imprinted cavity and specific binding sites of functional groups in a predetermined orientation was formed, whereas in nonimprinted sorbent, the flexibility of -CH2-CH2-CH2-, which was joined to the -SH group, resulted in no such specificity. Dynamic Adsorption of Cd(II) by the Imprinted Functionalized Silica Gel. For dynamic column operation, the breakthrough capacity represents the exhaustion point in terms of feed volume, after which the absorbate leaks through into the effluent in gradually increasing amounts that can exceed the preset or desired value.15 The breakthrough capacity depends on a variety of factors, such as flow rate, temperature and sample concentration. Because the breakthrough point depends on the method of detection, it usually is assumed to occur when Ce/Ci attains an arbitrarily chosen value (e.g., 0.001, 0.01, or 0.05), where Ce is the analyte concentration in the effluent and Ci is the analyte concentration in the influent.15 The breakthrough curves of the imprinted functionalized silica gel for 4 mg L-1 of Cd(II) at various sample flow rates as the ratio of Ce/Ci as a function of the effluent volume are shown in Figure 2. The volume of solution percolated from the breakthrough point to the point of leveling of the breakthrough curve for a given solution flow rate also depends on the kinetics of adsorption.16 Ce/Ci was progressive and reached 1 at the leveling of the curve with the increase of sample flow rate. The slope of the ascending portion of the curve also suggests a higher exchange rate of -SH and Cd(II). The breakthrough capacity was calculated with the assumption that breakthrough occurred at Ce/Ci ) 0.01. The effect of sample flow rate on the breakthrough capacity of the imprinted functionalized silica gel was investigated using 4 mg L-1 Cd(II) at pH ) (14) Cox, P. A. The Elements on Earth; Oxford University Press: Oxford, 1995. (15) Wang, X. R.; Barnes, R. M. J. Anal. At. Spectrom. 1989, 4, 509-518. (16) Malla, M. E.; Alvarez, M. B.; Batistoni, D. A. Talanta 2002, 57, 277-287.

Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

1737

Figure 2. Breakthrough curves of the imprinted silica gel sorbent for 4 mg L-1 Cd(II) (pH ) 5.5) at various sample flow rates.

5.5 in the range of 1.7-6.8 mL min-1. It was found that the breakthrough capacity decreased as the sample flow rate increased, then remained constant with a further increase in the sample flow rate. The noticeably higher breakthrough capacity at the lower flow rate suggested that the longer the interaction between the imprinted sorbent and the Cd(II) solution was, the more retention sites in the sorbent were utilized. The breakthrough capacity for 4 mg L-1 of Cd(II) at a sample flow rate of 5.2 mL min-1 was calculated to be 11.7 µmol g-1. The dynamic capacity is an important parameter for the imprinted sorbent in on-line application. Study of the effect of sample flow rate on the dynamic capacity of the imprinted silica sorbent showed that no remarkable change in the dynamic capacity was observed as the sample flow rate increased from 1.7 to 6.8 mL min-1. The dynamic capacity (mean ( σ, n ) 5) for the five sample flow rates studied (1.7, 2.9, 4.1, 5.2 ,and 6.8 mL min-1) is 65.6 ( 1.7 µmol g-1. These results also indicate that the adsorption kinetics of the imprinted silica gel sorbent for Cd(II) is fairly fast, which is an obvious advantage of the imprinted sorbent for its application to on-line, solid-phase extraction. Selective on-Line, Solid-Phase Extraction of Cd(II) by the Imprinted Thiol-Functionalized Silica Gel Sorbent. The applicability of the imprinted thiol-functionalized silica gel to online, solid-phase extraction of Cd(II) was evaluated by a hyphenated technique, that is, FI on-line microcolumn preconcentration and separation coupled with FAAS. The chemical and flow variables, such as sample acidity, sample loading flow rate and loading time, eluent, and its concentration and flow rate, were optimized to achieve good sensitivity and precision for the extraction and elution of Cd(II). Sample acidity plays an important role in complex adsorption extraction because it affects the complexing reaction between metal ions and ligands. The influence of sample acidity on the on-line extraction of 16 µg L-1 Cd(II) was tested at a sample flow rate of 8.8 mL min-1 for 45-s extraction. Similar to static adsorption, below pH ) 3, no Cd(II) was adsorbed onto the imprinted sorbent. The absorbance increased with the pH of the sample solution up to 5, but further increase in the pH of sample solution up to 7.5 resulted in no obvious change in the absorbance. The effect of sample loading time on the on-line, solid-phase extraction of 16 µg L-1 Cd(II) was tested at a sample loading flow rate of 8.8 mL min-1. It was found that the absorbance increased 1738 Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

almost linearly as sample loading time increased up to at least 130 s. Studies on the effect of sample loading flow rate on the on-line, solid-phase extraction 16 µg L-1 Cd(II) for 45 s showed that the absorbance increased linearly as the sample loading flow rate increased, up to 9.7 mL min-1. These results also indicate that the kinetics for the adsorption of Cd(II) by the developed sorbent was very fast. The wide range of linearity for absorbance against sample loading time and sample loading flow rate in the present on-line, solid-phase extraction system offered great potentiality for achieving high enhancement factors by increasing sample loading rates and sample loading time without losing extraction efficiency. Diluted hydrochloric acid was chosen for on-line elution of the retained analyte from the microcolumn. The effect of HCl concentration on the elution of the retained Cd(II) was examined at a flow rate of 3.6 mL min-1 for 15 s of elution. The absorbance increased as the HCl concentration moved to 0.1 mol L-1, then kept almost constant with a further increase of HCl concentration to 1 mol L-1. These results indicate that 0.1-1 mol L-1 HCl at a flow rate of 3.6 mL min-1 for 15 s was quite efficient for quantitative elution of the adsorbed Cd(II) from the microcolumn packed with the ion-imprinted functionalized silica gel sorbent. After elution, the residual HCl solution in the connecting tubings and the column should be washed away for the next effective on-line, solid-phase extraction because the remnant acid was unfavorable for extraction of Cd(II). It was found that doubly deionized water at a flow rate of 7.6 mL min-1 for 5 s column rinsing was quite efficient for this purpose. The selectivity of the imprinted silica gel sorbent for on-line, solid-phase extraction of Cd(II) was demonstrated by studying the effect of coexisting metal ions on the recovery of 16 µg L-1 Cd(II). Selection of the coexisting ions was on the basis of the following consideration: ions with much smaller or larger size than Cd(II) without (Mg(II) and K(I)) or with high affinity (Fe(III), Zn(II), Ni(II) and Co(II)) to the thiol group; ions with similar size to Cd(II) but without affinity to the thiol ligand, such as Na(I), La(III), and Ca(II); and ions with similar size to Cd(II) and with high affinity to the thiol ligand, such as Hg(II), Cu(II), and Pb(II). The results showed that up to 400 mg L-1 of Na(I), 400 mg L-1 of K(I), 320 mg L-1 of Ca(II), 320 mg L-1 of Mg(II), 320 mg L-1 of La(III), 250 mg L-1 of Co(II), 250 mg L-1 of Ni(II), 150 mg L-1 of Fe(III), 120 mg L-1 of Zn(II), 20 mg L-1 of Hg(II), 15 mg L-1 of Cu(II), 10 mg L-1 of Pb(II), and 8 mg L-1 of As(III) had no significant interferences with the on-line sorbent extraction of 16 µg L-1 Cd(II). These results indicate that the Cd(II)imprinted thiol-functionalized silica gel sorbent had high selectivity for Cd(II). There are three possible factors for the selectivity environment. One is the thiol-functional group inherent selectivity. The thiol group that is a soft base would not interact with alkali metal and alkali earth metal ions that are classified as hard acids. The second is the hole-size selectivity. That is, the size of Cd(II) exactly fits the cavity of the Cd(II)-imprinted silica gel sorbent. The third is the coordination-geometry selectivity because the Cd(II)-imprinted sorbent can provide the ligand groups arranged in a suitable way required for coordination of Cd(II) ion. Although Cd(II), Hg(II), Cu(II), and Pb(II) have similar size, and all these four ions have high affinity with the thiol ligand, the developed imprinted functionalized silica gel sorbent still exhibits high

selectivity for extraction of Cd(II) in the presence of Hg(II), Cu(II), and Pb(II). These results suggest that the coordinationgeometry selectivity may dominate in the selectivity enhancement. As demonstrated later, the present on-line, solid-phase extraction system allowed the interference-free extraction of trace Cd in the environmental and biological samples studied. Figures of Merit for the Present on-Line, Solid-Phase Extraction Coupled with FAAS Using the Developed Imprinted Functionalized Silica Gel Sorbent. The analytical figures of merit of the present on-line, solid-phase extraction using the imprinted functionalized silica gel sorbent coupled with FAAS for the determination of trace cadmium were evaluated under optimal experimental conditions. With a sample influent flow rate of 8.8 mL min-1 for 45 s of extraction, the enrichment factor (EF) obtained by comparing the slopes of the linear portion of the calibration curves before and after the on-line extraction was 56 and the detection limits (3σ) was 0.07 µg L-1 at a sampling frequency of 55 h-1. The precision (RSD) for 11 replicate solidphase extractions of 8 µg L-1 Cd(II) was 0.9%. The imprinted silica gel sorbent also offered good linearity (r ) 0.9997) with a calibration function of A (absorbance) ) 0.00798C (concentration in µg L-1)-0.00341 for the on-line, solid-phase extraction of trace Cd(II) in the range of 0.3-64 µg L-1Cd(II). To evaluate the accuracy of the developed on-line, solid-phase extraction coupled with FAAS for the determination of trace cadmium, five certified reference materials (CRMs), GBW08571 (mussel), GBW08511 (rice flour), GBW07605 (tea), GBW07601 (human hair), and GBW07405 (soil) were analyzed. As is shown in Table 2, the determined concentrations of cadmium in these CRMs by the present method using simple aqueous standards for calibration were in good agreement with the certified values. The developed methodology was also applied to analysis of six real samples. The analytical results are given in Table 3. The recoveries of Cd ranged from 97 to 101%. These results demonstrate the applicability of the on-line, solid-phase extraction system using the developed imprinted silica sorbent coupled with FAAS for interference-free determination of trace cadmium in the environmental and biological samples studied.

Table 2. Analytical Results for the Determination of Trace Cadmium in the Certified Reference Materials concn of Cd, µg g-1 sample

certified

determined (mean ( σ, n ) 5)

GBW08571 (mussel) GBW08511 (rice flour) GBW07605 (tea) GBW07601 (human hair) GBW07405 (soil)

4.5 ( 0.3 0.504 ( 0.018 0.057 ( 0.010 0.11 ( 0.03 0.45 ( 0.09

4.3 ( 0.1 0.523 ( 0.022 0.052 ( 0.03 0.10 ( 0.02 0.42 ( 0.04

Table 3. Analytical Results for the Determination of Trace Cadmium in Real Samples sample

concn of Cd determined (mean ( σ, n ) 5)/µg L-1

recovery of 2 µg L-1 Cd spiking, %

river water 1 river water 2 wastewater 1 wastewater 2 soil 1 soil 2

0.91 ( 0.24 0.73 ( 0.20 1.27 ( 0.13 3.24 ( 0.09 0.088 ( 0.047a 0.050 ( 0.018a

98 ( 2 97 ( 3 98 ( 1 101 ( 1 98 ( 2b 98 ( 3b

a

µg g-1. b 0.2 µg g-1 Cd spiking.

CONCLUSIONS A simple procedure for the synthesis of a Cd(II)-imprinted thiolfunctionalized silica gel sorbent was developed by a surface imprinting technique in conjunction with a sol-gel process. The prepared imprinted sorbent exhibits good characteristics, such as good stability under acidic conditions, nonswelling, high thermal resistance, fast adsorption and desorption kinetics, and high selectivity for Cd(II), making the sorbent very suitable for on-line, solid-phase extraction of trace Cd(II). ACKNOWLEDGMENT This research was supported by the National Natural Science Foundation of China (Grant No. 20275019, 20025516). Received for review September 26, 2004. Accepted December 16, 2004. AC048570V

Analytical Chemistry, Vol. 77, No. 6, March 15, 2005

1739