THE PROSPECTS OF CLAY MINERAL ELECTRODES In the context of economy and green chemistry, clay mineralchemically modified electrodes

protocols and devices, negBecause of clay’s unique properties, clay mineralprovide a suite of sensor options. ligible waste, and nontoxic materials make clays of conchemically modified electrodes (Cm-CMEs) are very suitable for many chem- siderable interest to the analytical community. The ical sensor applications (1–4). A successful chemical ultimate goal is to properly utilize the high cationsensor must have good sensitivity, selectivity, speed, exchange capacity (CEC) and intrinsic catalytic propand longevity. In economic terms, inexpensive, abun- erties of clay minerals to construct versatile chemical dant materials and cheap and easy preparation meth- sensors. This article describes basic preparation and ods certainly make clays attractive. In green chem- characteristics of Cm-CMEs in several important chemistry terms, environmentally friendly analytical ical sensor applications. JYH-MYNG ZEN and ANNAMALAI SENTHIL KUMAR NATIONAL CHUNG HSING UNIVERSITY (TAIWAN) © 2004 AMERICAN CHEMICAL SOCIETY

J U N E 1 , 2 0 0 4 / A N A LY T I C A L C H E M I S T R Y

205 A

Interlayer d

(a)

(b)

Clay mineral layer nH2O Xn+

Tetrahedral 6O 4(Al,Si) 4O+2OH 2Al,Fe,Mg 4O+2OH 4(Si,AI) 6O

H2O

d d

Cation exchange capacity Octahedral

Dry

Expanded

FIGURE 1. Clay chemistry. (a) Atomic arrangement in the unit cell of montmorillonite, a typical 2:1 layered clay mineral and (b) its plate-like structures in the presence or absence of water.

Clay minerals Clay minerals are nearly ubiquitous in soils and sediments found in good, old-fashioned mud formed millions of years ago under certain geological conditions. The predominant clay particles are often <1 µm in diameter, which places them in the colloidal size range. Primitive people used clay minerals mainly for pottery and ceramics ~25,000 years ago. Approximately 4250 clay minerals have been reported on the basis of their chemical and structural variations. These minerals form the phyllosilicate or sheet silicate family and are distinguished by layered structures composed of polymeric sheets of SiO4 tetrahedra linked to sheets of (Al, Mg, Fe)(O, OH)6 octahedra, either 1:1 or 2:1. These sheets form pseudo-3-D crystals consisting of regular stacks of parallel platelets (face-to-face association) held together by charge-balancing interlayer counterions (Na+ and Ca2+). The edge of each platelet has a hydroxide group that allows it to form water-gels in the internal structures. Smectites, such as montmorillonite, nontronite, and bentonite, are an example of the 2:1 phyllosilicate minerals. Figure 1a shows a typical crystal platelet structure of 2:1 montmorillonite. Because of their small particle size, clay minerals have a large surface area-to-mass ratio; some can swell or expand up to 20
their own volume. Clays also have good intercalation and high stability—they are chemically robust and can withstand high temperatures and highly oxidizing conditions. CEC is one of the most important properties of clay minerals and is based on the isomorphous substitution of metal ions with different valency within the crystal lattice (5). For example, substituting Mg2+, Fe2+, or Li+ for Al3+ octahedral ion results in a layer with an excessive negative charge, which is compensated for by exchangeable cations (e.g., Na+ and Ca2+) adsorbed on the surface of the layer. Apart from these characteristics, clays induce spatial constraints and surface chemical effects that can generate some selective properties, especially for catalysis or catalytic supports (1, 2).

Preparing Cm-CMEs Raw clay minerals usually contain large amounts of impurities and need to be properly cleaned before use (2). Generally, clay minerals are suspended in water to allow larger particles to settle and be removed. Clay can be converted into the desired 206 A

A N A LY T I C A L C H E M I S T R Y / J U N E 1 , 2 0 0 4

Na+-form, simply by stirring with a high-concentration electrolyte (e.g., 1 M NaCl) for about two days. Clay colloidal particles are then separated by centrifugation and repeatedly washed with water. Freeze-drying is sometimes used for purification. Iron oxide, a common and specific contaminant in clay, is removed selectively by treating with dithionite or citrate without affecting the crystalline iron. In modification procedures, colloidal clay solutions are deposited on conductive solid supports, such as glassy carbon, platinum, and optically transparent SnO2 electrodes. Bard’s group spin-coated (8000 rpm for 15 s) a thin layer (2.9–3.2 µm) of Na-montmorillonite clay colloids (12 g/L) with small amounts of poly(vinyl alcohol) and colloidal platinum on a SnO2 electrode (6, 7 ). Fitch’s group used a drop-casting procedure for Na-montmorillonite clay colloids (5 g/L) followed by drying at 100 °C for 10 min and then by cooling for 5 min (8). Our group transferred 100 µL of 0.5g/L nontronite colloid onto a glassy carbon electrode (GCE) and dried it for ~1 h at ambient conditions to create a physically stable film. Usually at hydrated conditions, the clays’ colloids have a fully expanded crystal structure with the water molecules primarily absorbed into the spaces between the clays’ plates and/or layers; the colloids become rigid as they dry on the underlying surface (Figure 1b). The strength of the dried clay exemplifies why mud has been used since ancient times to build huts and make pottery. Traditional Chinese potters combined animal urine with kaolinite clay to increase the swelling of clay art works. Our interest in clay originated from a research program at National Chung Hsing University (Taiwan). When paraquat herbicide (methyl viologen, MV2+) was sprayed onto a local dam area, it strongly adsorbed onto the underlying soil sediments. The herbicidal activity decreased significantly when incorporated into clays, which prevented the subsequent uptake by the plants because of the strong adsorption onto the silicate. Geological surveys indicate that soil sediments enriched with relatively higher amounts of natural clay minerals are the basis for such an unusual observation (9). Subsequently, six commonly used representative clay minerals with a 1:1 structure (kaolinite) and a 2:1 structure (illite, bentonite, nontronite, montmorillonite, and vermiculite) were chosen to modify GCEs for paraquat adsorption studies (10, 11). The obtained paraquat reduction peak

❍ 100

Characteristics

80 60 Current (A)

currents ipc (c, cathodic) were in the order nontronite > illite > montmorillonite > bentonite > kaolinite > vermiculite > bare GCE by simple cyclic voltammetry (Figure 2). If one takes into account the difference in swelling capability and isomorphous substitution, the large difference in adsorption of paraquat to kaolinite and nontronite was not surprising. The interesting part, however, was that the nontronite showed the highest ipc, and the order of the measured ipc seemed to be related to the CEC of these six clays (5).

40 20 0 Montmorillonite/GCE

–20 Nontronite/GCE

How the variation of CEC can influence the electrocat0 –0 .5 V –40 alytic oxidation processes can be explained by simple electrochemical studies (Figure 2). A cationic mediator MV2+ E (V vs Ag/AgCI) and the mediated H2O2 reduction reactions were uniformly applied as a model system. No faradaic responses were observed on either bare montmorillonite or nontronite- FIGURE 2. Cyclic voltammetric responses of a nontronite/GCE and montmodified GCEs (designated montmorillonite/GCE and morillonite/GCE at a scan rate of 50 mV/s on the reduction of 1 mM H2O2 nontronite/GCE, respectively). In the presence of MV2+, with 0.5 mM MV2+ in pH 8 phosphate buffer solution (I = 0.1 M). GCE geoan obvious redox characteristic corresponding to the elec- metrical area is 0.07 cm2. tron transfer reaction of MV2+/1+ at an equilibrium potential of –0.70 V was noted. The observed ipc was ~2.5
higher on the nontronite/GCE than on the montmorillonite/ minerals are hindered in electrocatalysis because ~70–80% of the GCE and was in accordance with the performance of the H2O2 ion-exchanged cationic mediator becomes electroinactive, withcatalysis. It has been proposed that the free iron, which exists out any way to mediate the analyte (13). This condition is also in the tetrahedral (nontronite) and/or octahedral (montmoril- true for nontronite. The basic physical properties of nontronite and montmorillonite) exchange sites of the clay minerals, participates in the electron transfer reaction through a charge-shuttling process lonite are summarized in Table 1. Nontronite has isomorphous substitution in the tetrahedral sheet, and montmorillonite is a for the redox mediation and for catalysis. dioctahedral mineral with substitution mainly in the octahedral Fe 2 + H2O2 MV 2 + sheet. The measured tetrahedral charge distribution ratio Q tet e– for nontronite is thus ~4.7
higher than that of montmorillonite. The fraction of electroactive MV2+ adsorbed c* and the total + 3 + MV amount of adsorbed paraquat ct are also listed in Table 1 for Fe H2O comparison. The measured c*/ct for the nontronite/GCE is 2 + Clay and MV Soln. ~4.3
higher than that of the montmorillonite/GCE. The effect Oyama and Anson first suggested that the iron ions in the of pH on the electrochemistry of MV2+ and Ru(NH3)63+ exoctahedral sites within the montmorillonite possess intrinsic change cations in the nontronite/GCE and montmorillonite/ catalytic activity (4 ). They based their model on the H2O2 re- GCE provides insight into the electrochemical accessibility of duction reaction on Ru(NH3)63+-doped, montmorillonite-coat- redox active species in the electrodes (4, 10). The number of ed graphite electrodes. Meanwhile, Wang and Huang noticed surface layer-bound MV2+ ion pairs that can access the electhat, in the presence of hydroquinone mixtures, microorgan- trode surface decreases with increasing acidity (10). In contrast, isms on nontronite retained a higher degree of activity than the electroactivity of Ru(NH3)63+ exchange ions intercalated in those on montmorillonite (12). They further stated that the the gallery region of the clay film is independent of pH (4). high content of Fe3+ in the nontronite interlayer directly assists These results, together with the difference in the octahedral and tetrahedral charge distributions between montmorillonite the redox reaction in the above scheme. Nontronite is the iron-rich end member of the montmoril- and nontronite, lead to the conclusion that the active site for lonite group, and the representative chemical formulas are (5) the H2O2 reduction is the iron that resides on the tetrahedral surface layers of nontronite. The results also showed that mainly the ion-paired species of MV2+ adsorbed in excess of the Nontronite 3+ + [(Si7.25Al0.75)(Fe 2.75 Al0.85Mg0.33Ti0.05)O20(OH)4X 0.98] CEC onto the tetrahedral sites contribute to the excellent catalytic activity. Montmorillonite 2+ + [(Si7.84Al0.16)(Fe 3+ Enzyme-trapped systems 0.26 Al3.22Mg0.40Fe 0.12)O20(OH)4X 0.68] One emerging application of Cm-CMEs is a glucose biosensor According to these formulas, the amount of iron in non- in which glucose oxidase (GOx) enzyme from Aspergillus niger tronite is ~7.2
higher than that in montmorillonite. Most clay is stabilized between the clay mineral multilayers (14–18). Precise J U N E 1 , 2 0 0 4 / A N A LY T I C A L C H E M I S T R Y

207 A

(a)

(b) 0.2

10 A

0V

Glucose Clay

Blank Blank

GLA

GOx

GOx Clay

Glucose 20 A 0

BSA

–0.2 V Clay

FIGURE 3. Typical cyclic voltammetric responses of (a) first- and (b) second-generation glucose biosensors. GOx = ~205 units/cm2 using nontronite/GCE with and without 5 mM glucose in pH 7 phosphate buffer. (a) [MV2+] = 0.1 mM in solution, nontronite loading = 28.3 µg/cm2, and scan rate = 100 mV/s. (b) [dopamine] = 0.8 mM in solution, nontronite loading = 42.4 µg/cm2, and scan rate = 50 mV/s. BSA, bovine serum albumin; GLA, glutaraldehyde.

and stable immobilization of enzymes on the electrode surface while maintaining the enzymes’ activity is the crucial problem in biosensor development. Immobilizing GOx between two nontronite clay coatings in a sandwich arrangement is quite efficient at trapping the enzyme layer and shuttling the electrontransfer reaction to detect glucose through the H2O2/O2 intermediate without any interference from ascorbic and uric acids (14). The sensor has detected glucose up to 6 mM with a detection limit of 3 µM and a current sensitivity of 52.3 µA/ (mM cm2) (Figure 3a). MV2+

Fe2+

MV+

Fe3+

H2O2

e–

H2O2

O2

Clay, MV2+, and GOx

In another case, a glucose sensor was constructed simply by coating a mixture of GOx, bovine serum albumin, glutaraldehyde, and nontronite onto a GCE with dopamine as a solutionphase redox mediator without any common coexisting biochemical interferences (Figure 3b). e–

Dopaquinone

GOx (red)

Ru(NH3)63+, tetrathiafulvalene, and ruthenium purple mediators (16–18). The clay minerals systems are stable and can be reused for up to two months (14–16), unlike Nafion or SiO2 systems (19, 20). Cytochrome c, cytochromes P450, myoglobin, and other oxidase enzymes are also stabilized by clay minerals (21, 22). The stabilization of an enzyme’s protein sites in clay minerals is analogous to the stabilization effect that certain biochemical compounds, such as ascorbic acid in oranges and apples, sugars in grapes, and horseradish peroxidase in bananas, have on plants and fruit tissues. In a way, the clay support can be regarded as mimickGOx (ox) Glucose ing fruit tissue, which is quite reasonable because GOx (red) Gluconic acid soil is a natural inorganic source reSoln. sponsible for the plant growth. Some clays, such as montmorillonite and kaolinite, are even used as dietary supplements in many rural areas to improve energy and stamina, reduce arthritis, and so forth (23), thereby attesting to the positive interactive effects of clay with biological systems.

Gluconic acid

Sensors combined with Nafion

Nafion is a cation exchange perfluorinated ionomer extensively used to improve resistance to interferGlucose Dopamine GOx (ox) ences from surface-active compounds and common ions and, therefore, protect the electrode’s surface. Clay, dopamine, and GOx Soln. We constructed a highly sensitive paraquat sensor by Because the reaction mechanism of the dopamine-based glu- using a Nafion-coated nontronite/GCE with a square-wave cose sensor does not involve either H2O2 or O2, the response voltammetry detection limit of 0.5 ppb, which considerably retime and detection conditions are largely improved and simpli- duced serious interferences from p-nitrophenol and surfactants fied. The calculated detection limit is 5 µM, and the current (11). The sensor’s practicality was illustrated by selective measensitivity for glucose is 59.6 µA/(mM cm2), with 3 orders of surements of paraquat in real water samples. The same Nafion-coated nontronite/GCE can also detect magnitude dynamic detection range up to ~10 mM (15). These nontronite systems, in accordance with the physicochemical uric acid and dopamine in human urine samples in the presence characteristics discussed earlier, also possess better sensitivity of high concentrations of ascorbic acid by preconcentrating the than the montmorillonite-clay modified GOx systems based on sample (24). The concept behind eliminating the ascorbic acid 208 A

A N A LY T I C A L C H E M I S T R Y / J U N E 1 , 2 0 0 4

❍

is remarkable. At physiological pH (7.4), ascorbic acid exists in the anionic form (pKa 4.1) and dopamine exists in the cationic form (pKa 8.87); hence, the anionic Nafion-coated nontronite/ GCE can effectively discriminate a high concentration of ascorbic acid (25). However, Nafion alone cannot solve the sensitivity problem and thus needs an extra “accumulation factor” that can be obtained with clay minerals. After preanodization treatment of the Nafion-coated nontronite/GCE at ~2.0 V versus Ag/AgCl for 60 s, the accumulation of uric acid and dopamine increased enormously as did the detection current signals. This extra accumulation occurred as described in Pearson’s principle: Soft acids prefer to bind to soft bases (24). Lighter transition elements in higher oxidation states, such as Fe(III), are hard acids, whereas uric acid and dopamine are hard bases. The detection limit was 0.2 µM for uric acid and 2.7 nM for dopamine.

Metal ion sensors Although the CEC is not selective for particular metal ions, detection of target metal ions can easily be achieved by tuning the specific properties of the working parameters (deposition, preconcentration, and stripping potentials) and solution-phase parameters (pH and base electrolyte). The selective and simultaneous detection of Cu2+ and Zn2+ by anodic stripping voltammetry without any common intermetallic complications was demonstrated on a mercury-plated nontronite/GCE (26). Because of the difference in affinity, size, and shape of hydrated Zn2+ and Cu2+, the clay–mercury-modified electrode could discriminate Cu2+ against Zn2+ during the electrode deposition process to further help quantification. In another example, an efficient Cu2+ sensor can be constructed by using montmorillonite and nontronite (with cellulose) for the preconcentration and electroanalysis by square-wave voltammetry of Cu2+ in ammoniacal medium through the formation of the [Cu(NH3)4]2+ complex (27, 28). The nontronite with cellulose acetate membrane has relatively better performance and high mechanical stability because of its higher CEC properties. Such an approach is promising for trace Cu2+ analysis in different types of water samples. The rapid and reproducible production of disposable sensors using screen-printing technology, where the conducting carbon ink is printed on a suitable support (e.g., polypropylene base), has become an important, practical reality. In a typical nontronite-modified screen-printed electrode (SPE) procedure, 30 µL of 0.1 wt% clay mineral was uniformly spread on the bare SPE and air-dried overnight (29). The SPE was useful for regular potential scan voltammetry and flow-injection analysis (FIA). FIA is performed in chronoamperometric (i-t) mode at a fixed potential under hydrodynamic flow of the base electrolyte, in electroanalytical chemistry. The FIA method is fast and simple, and the volume of analyte used is typically ≤200 µL. So far, only the nontronite clay has been coupled with SPE for analytical assays.

Other applications A preanodized nontronite/SPE was first used to detect arbutin, a skin-whitening agent in cosmetic bleaching products, by squarewave voltammetry (30). The preanodization process of clay (2 V versus Ag/AgCl for 20 s) exhibits a marked enhancement in the oxidative current response of arbutin at +0.35 V versus Ag/AgCl

Current (A)

200

0

–200

–400

1200

800

400

0

–400

E (mV vs Ag/AgCl)

FIGURE 4. Typical continuous response for 20 cycles of voltammetry 3– of a nontronite-modified carbon paste electrode with 2 mM Fe(CN)6 in 0.1 M pH 2 KCl/HCl at a scan rate of 50 mV/s.

in pH 10 ammonium buffer solution. Compared with the performance at a bare preanodized SPE, the clay/SPE had improved sensitivity and reproducibility. If desired, the electrode surface can be renewed with good reproducibility. Codeine, a narcotic drug derived from opium, has been determined by static and hydrodynamic techniques using a nontronite/ SPE with good sensitivity and selectivity (31). Selective preconcentration of the codeine at the nontronite/SPE at –0.6 V versus Ag/AgCl resulted in an obvious catalytic oxidation stripping peak at ~1.10 V versus Ag/AgCl by square-wave voltammetry in pH 4 phosphate buffer solution. Meanwhile, a catalyticoxidative current response of 313 µA/mM with a detection limit of 1 ng in 20-µL working solution was obtained by FIA at an applied potential of 1.4 V versus Ag/AgCl with a flow rate of 0.6 mL/min. Note that this is the first application of a clay/ SPE-based system in FIA. Practical analytical utility was demonstrated by selective measurements of codeine spiked into urine. The detection of amitrole, a well-known herbicide, by FIA was also developed by using a preanodized nontronite/SPE with excellent sensitivity (29). Its low volatility (bp = 260 °C) and high solubility in water (280 g/L) make it a food chain contaminant through plants, fruits, and water. It causes alveolar damage due to inhalation and is carcinogenic in human beings. Although amitrole should be monitored at a regulatory level of 0.1 µg/L in drinking water in European countries, no robust analytical procedure is available to measure it at this level. The absence of a method also explains the lack of data regarding its occurrence in various environmental waters. To reach the 0.1 µg/L level, most of the reported methods require an off-line preliminary extraction and concentration (32). Amitrole analysis by GC is very difficult because of its high polarity and low volatility. Direct detection by UV–vis spectroscopy is also not sufficiently sensitive because of the low extinction coefficient of amitrole. It is a challenge to achieve the 0.1 µg/L detection limit without any off-line and separation procedures. Using a preanodized nontronite/SPE is a good way to reach this goal. The intrinsic catalytic activity of the nontronite clay helps to detect amitrole by FIA at 1.4 V versus Ag/AgCl in pH 4.5 PBS with a detection limit of 0.07 ng in a 20-µL sample loop. J U N E 1 , 2 0 0 4 / A N A LY T I C A L C H E M I S T R Y

209 A

Table 1. Physical characteristics of nontronite and montmorillonite CMEs. Parameters

Nontronite

Montmorillonite

CEC/mmol per 100 g

107.0

87.0

Tetrahedral charge distribution/mol per unit cell

–0.75

–0.16

Octahedral charge distribution/mol per unit cell

–0.34

–0.52

Film thickness/µm

2.4

2.1

4.9

4.2

4.6

1.3

9.9

11.5

0.047

0.011

7.4

43

6

3

Film volume/10 cm a b

104 c*/mol cm–3 3

10 ct /mol cm

–3

c*/ct 10

2 –1

10 Dapp /cm s a b

From the peak current observed at a slow scan rate of 1 mV/s. From the decrease in concentration of the soaking solutions.

The preanodized clay system can also be applied to the simultaneous determination of purine bases, such as hypoxanthine and xanthine, and uric acid in plasma and urine samples, which are important in biochemistry and clinical diagnosis (33). The major pathway of purine nucleotide and deoxynuclotide metabolism in animals involves several intermediates and ends with uric acid; hypoxanthine and xanthine are precursors of uric acid and intermediates in purine biosynthesis (34). The various enzymatic methods for measuring these compounds usually lack stability and are rather expensive (35). The nonenzymatic electrochemical approach is relatively simple and fast, not to mention the fact that xanthine oxidase can oxidize both hypoxanthine and xanthine to uric acid, rendering the specific determination difficult. The iron-rich nontronite and montmorillonite can be converted into an efficient catalyst by electrochemical formation of hybrid Prussian blue [Fe3+Fe(CN)6] and ruthenium purple [Fe3+Ru(CN)6] directly inside the matrix (18, 36). In the typical hybrid Prussian blue preparation, a nontronite modified carbon paste electrode is electrochemically treated with Fe(CN)63– in pH 1–3 KCl/HCl solution for several continuous cycles (Figure 4). A larger quantity of Fe3+ in the interlaminar structure of the clay minerals can effectively assist hybrid Prussian blue formation; nontronite in particular has a high iron content. These new hybrid electrodes are extraordinarily stable and highly catalytic compared with classical systems. In another case, lead ruthenate pyrochlore (Pb2Ru2–xPbxO7–y) catalyst is stabilized inside the nontronite/SPE matrix via Ru3+ and Pb2+ ion exchange and in situ precipitation (37 ). This new material takes advantage of the electrocatalytic properties of both pyrochlore and nontronite. A sensitive catalytic dopamine sensor with a detection limit of 0.54 nM by square-wave voltammetry has been constructed with these materials.

can generate strongly oxidizing •OH free radicals by the reaction of Fe2+ with H2O2 and formation of the [FeIV = O]+ intermediate (38), which can oxidize DNA, destroy pollutants, and participate in organic synthesis (39). By adopting the Fenton reaction, clay-modified systems, which can withstand a highly oxidizing environment, can achieve super oxidizing power. Properly tuning the process and coupling it with a separation technique can achieve effective DNA quantification. Because of the stability of clays, combining them with enzymes and suitable redox mediators—for example, clay in conjunction with hydroquinone mediator—could be a new way to quantify microbiological systems such as fungi, bacteria, and E. coli (12). Electrochemiluminescence (ECL) of Ru(bpy)32+ is well known in polymer-modified electrodes, where Ru(bpy) 32+ is oxidized to Ru(bpy) 33+, and the oxidized compound emits luminescent light at ~610 nm in the presence of suitable quenchers (40). ECL is used for amino acid quantification and preliminary sequencing (41). By incorporating Ru(bpy) 32+ into clay minerals, efficient analytical assays for amino acids are possible; for example, montmorillonite was used to enhance the ECL of luminol (42). Ilangovan and Zen pointed out that the inactive iron in nontronite activates Ru(bpy) 32+ redox chemistry (43). Properly matching such applications with natural clay minerals and SPEs promises to be a new eco-friendly analytical approach.

Future prospects

(3) (4) (5) (6) (7) (8) (9)

One important aspect of smectite clays is the essential role played by inactive iron sites in the clay network in intrinsic charge transfer in redox mediations (Figure 2), CEC, and derivatized macromolecules such as Prussian blue analogues (Figure 4). The Fenton reaction mechanism is also believed to operate at heterogeneous iron sites in clay minerals. The Fenton reaction 210 A

A N A LY T I C A L C H E M I S T R Y / J U N E 1 , 2 0 0 4

We gratefully acknowledge the National Science Council of Republic of China for financial support and Horng-Ji Chen for helpful suggestions.

References (1) (2)

Baker, M. D.; Senaratne, C. In Electrochemistry of Novel Materials; Lipkowski, J., Ross, P. N., Eds.; VCH Publisher: New York, 1994; pp 339–380. Bard, A. J.; Mallouk, T. In Molecular Design of Electrode Surfaces; Murray, R. W., Ed.; VCH Publisher: New York, 1992; pp 271–312. Fitch, A. Clays Clay Miner. 1990, 38, 391–400. Oyama, N.; Anson, F. C. J. Electroanal. Chem. 1986, 199, 467–470. Jaynes, W. F.; Bigham, J. M. Clays Clay Miner. 1987, 35, 440–448. Ghosh, P. K.; Bard, A. J. J. Am. Chem. Soc. 1983, 105, 5691–5693. Ege, D.; et al. J. Am. Chem. Soc. 1985, 107, 5644–5652. Fitch, A.; et al. J. Phys. Chem. 1988, 92, 6665–6670. Chen, H.-J. Studies of the Adsorption and Desorption of Paraquat in the

(10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27) (28) (29) (30)

Soils. Master Thesis, Research Institute of Soil Science, National Chung Hsing University, Taiwan, 1988. Zen, J.-M.; Jeng, S.-H.; Chen, H.-J. J. Electroanal. Chem. 1996, 408, 157–163. Zen, J.-M.; Jeng, S.-H.; Chen, H.-J. Anal. Chem. 1996, 68, 498–502. Wang, M. C.; Huang, P. M. Nature 1986, 323, 529–531. King, R. D.; Nocera, D. G.; Pinnavaia, T. J. J. Electroanal. Chem. 1987, 236, 43–53. Zen, J.-M.; Lo, C.-W. Anal. Chem. 1996, 68, 2635–2640. Zen, J.-M.; Lo, C.-W. Anal. Chem. 1997, 69, 1669–1673. Ohsaka, T.; Yamaguchi, Y.; Oyama, N. Bull. Chem. Soc. Jpn. 1990, 63, 2646– 2652. Lei, C.; et al. J. Electroanal. Chem. 1996, 419, 93–98. Shyu, S.-C.; Wang, C.-M. J. Electrochem. Soc. 1998, 145, 154–158. Künzelmann, U.; Böttcher, H. Sens. Actuat. B 1997, 38–39, 222–228. Karyakin, A. A.; et al. Anal. Chem. 2002, 74, 1597–1603. Lei, C.; et al. Biochem. Biophys. Res. Comm. 2000, 268, 740–744. Dan, S.; et al. J. Electroanal. Chem. 2002, 537, 103–109. Aufreiter, S.; et al. Int. J. Food Sci. Nut. 1997, 48, 293–305. Zen, J.-M.; Chen, P.-J. Anal. Chem. 1997, 69, 5087–5093. Zen, J.-M.; Wang, W.-W.; Ilangovan, G. Anal. Chim. Acta 1998, 372, 315–321. Zen, J.-M.; Lin, H.-Y.; Yang, H.-H. Electroanalysis 2001, 13, 505–508. Zen, J.-M.; et al. Electroanalysis 2002, 14, 99–105. Pecoriari, M.; Bianco, P. Electroanalysis 1998, 10, 181–186. Zen, J.-M.; Chen, H.-P.; Senthil Kumar, A. Anal. Chim. Acta 2001, 449, 95–102. Shih, Y.; Zen, J.-M. Anal. Chim. Acta 2000, 412, 63–68.

(31) (32) (33) (34) (35) (36) (37) (38) (39) (40) (41) (42) (43)

Shih, Y.; Zen, J.-M.; Yang, H.-H. J. Pharm. Biomed. Anal. 2002, 29, 827–833. Dulay, J.; Hennion, M.-C. Trends Anal. Chem. 1995, 14, 407–414. Zen, J.-M.; et al. Sens. Actuator B: Chem. 2002, 84, 237–244. Voet, D.; Voet, J. G. Biochemistry, 2nd ed.; Wiley & Sons: New York, 1995. Elekes, O.; et al. J. Clin. Chim. Acta 1995, 239, 153–165. Zen, J.-M.; Senthil Kumar, A.; Chen, H.-W. Electroanalysis 2000, 12, 542– 545. Zen. J.-M.; Senthil Kumar, A.; Chen, H.-P. Electroanalysis 2003, 15, 1584– 1588. Groves, J. T.; Puy, M. V. D. J. Am. Chem. Soc. 1975, 97, 7118–7122. Henle, E. S.; et al. J. Bio. Chem. 1999, 274, 962–971. Rubinstein, I.; Bard, A. J. J. Am. Chem. Soc. 1980, 102, 6641–6642. He, L.; Cox, K. A.; Danielson, N. D. Anal. Lett. 1990, 23, 195–210. Ouyang, C.-S.; Wang, C.-M. J. Electrochem. Soc. 1998, 145, 2654–2659. Ilangovan, G.; Zen, J.-M. Chem. Lett. 1999, 153–154.

Jyh-Myng Zen is a professor at National Chung Hsing University (Taiwan). His research interests include CMEs, electrocatalysis, photoelectrochemistry, physical electrochemistry, and chemical sensors. Annamalai Senthil Kumar is a postdoctoral fellow in Zen’s laboratory and is deeply interested in interdisciplinary areas of material science; physical and analytical chemistry, in particular, the design and application of CMEs; electrocatalysis; and bio- and chemical sensors. Address correspondence about this article to Zen at Department of Chemistry, National Chung Hsing University, 250 Kuo Kuang Road, Taichung 402, Taiwan ([email protected]).

J U N E 1 , 2 0 0 4 / A N A LY T I C A L C H E M I S T R Y

211 A