Electrocatalytic detection of estradiol at a carbon nanotube|Ni(Cyclam) composite electrode fabricated based on a two-factorial design

Electrocatalytic detection of estradiol at a carbon nanotube|Ni(Cyclam) composite electrode fabricated based on a two-factorial design

Analytica Chimica Acta 594 (2007) 184–191 Electrocatalytic detection of estradiol at a carbon nanotube|Ni(Cyclam) composite electrode fabricated base...

310KB Sizes 12 Downloads 32 Views

Analytica Chimica Acta 594 (2007) 184–191

Electrocatalytic detection of estradiol at a carbon nanotube|Ni(Cyclam) composite electrode fabricated based on a two-factorial design Xiaoqiang Liu, Danny K.Y. Wong ∗ Department of Chemistry and Biomolecular Sciences, Macquarie University, Sydney, NSW 2109, Australia Received 7 February 2007; received in revised form 20 April 2007; accepted 23 May 2007 Available online 29 May 2007

Abstract In this work, we have developed a carbon nanotube|Ni(cyclam)-coated glassy carbon electrode to achieve minimal fouling effects and to catalyse the oxidation of the oestrogen, estradiol, during voltammetric detection. This electrode was fabricated by initially applying a Nafion–carbon nanotube mixture, and then electropolymerising Ni(cyclam) complexes on the electrode. During this process, a two-level factorial design was used to optimise experimental parameters including the amount of carbon nanotubes, the concentration of Nafion and the surface coverage of Ni(cyclam). A linear calibration plot between 0.5 and 40 ␮M estradiol was then obtained in synthetic laboratory standard solutions. Based on a signal-to-noise ratio of 3, a detection limit of 60 nM was estimated, which is below the typical estradiol level measured in a normal menstrual cycle. The electrodes were subsequently applied to the detection of estradiol in protein-free human serum samples. Comparable sensitivity between synthetic laboratory standard solutions and serum samples was obtained, indicating minimal interference effects from the serum matrix. © 2007 Elsevier B.V. All rights reserved. Keywords: Electrocatalytic oxidation; Estradiol; Carbon nanotubes; Ni(cyclam) complexes

1. Introduction Estradiol is the most potent and biologically active oestrogen produced by the ovary. It is converted to the less potent oestrone in the liver, and then metabolised to estriol that has limited oestrogenic activity [1,2]. In a normal menstrual cycle, a typical level of 50 ng mL−1 estradiol is measured [3] (compared to a range of 25–400 pg mL−1 estradiol during non-menstrual times [4,5]). A deficiency of estradiol is often related to such diseases as menopausal symptoms or heart diseases and osteoporosis [6]. Therefore, quantitative detection of estrodiol is vitally important so that an appropriate replacement level can be administered for alleviation of menopausal symptoms or to prevent heart diseases and osteoporosis. High performance liquid chromatography and immunoassays have hitherto been used to detect estradiol [7–9]. Although these techniques are sensitive and specific, they are also expensive and time-consuming. On the other hand, electrochemistry offers a rapid and economical means of detection because of the



Corresponding author. Tel.: +61 2 9850 8300; fax: +61 2 9850 8313. E-mail address: [email protected] (D.K.Y. Wong).

0003-2670/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2007.05.043

ease of oxidising the phenolic group of estradiol to a phenoxonium ion, as illustrated in Scheme I [2,10]. However, coupling of phenoxonium ions will form a dimer or quinone, which is known to passivate or foul the electrode surface, leading to a weak, transient loss of detection signal of estradiol at naked electrodes [1,2,6,11,12]. This has limited applications of electrochemical techniques to the detection of estradiol. Meanwhile, Ni(III) has been demonstrated to show a high catalytic effect (approximately 75%) on the oxidation of phenolic compounds to quinones, while itself is being reduced to Ni(II) [13–15]. Unfortunately, strong hydration shells around nickel ions prevent a close contact between the ions and an electrode surface for electron transfer reactions, making aqueous Ni(II)|Ni(III) redox reactions ineffective in the catalysed oxidation of phenols [16]. In contrast, by coordinating a nickel ion in a hydrophobic complex, the planar structure of the resultant nickel complex would align the nickel ion close to the electrode surface for electron transfer reactions. Therefore, macrocyclic compounds such as Ni(cyclam) (cyclam = 1,4,8,11,tetraazacyclotetradecane) have often been used in the electrocatalytic reactions of phenolic compounds [13,14,17]. For example, Manriquez et al. [13] demonstrated that the Ni(cyclam) film was an efficient material for the electrocatalytic oxidation of p-nitrophenol in aqueous

X. Liu, D.K.Y. Wong / Analytica Chimica Acta 594 (2007) 184–191

NaOH solution. Similarly, Xu et al. [17] employed a Ni(cyclam)coated glassy carbon electrode in the electrocatalytic oxidation of norepinephrine to achieve selective detection of the compound. Recently, Lin and Li [18] have demonstrated that Pt nanoparticles also exhibited catalytic effects on the oxidation of estradiol, allowing them to achieve a detection limit down to 180 nM estradiol at a Pt nanoparticle-coated glassy carbon electrode. In electroanalytical chemistry, carbon nanotubes (CNTs) have increasingly been used as an excellent electrode material in the detection of a range of molecules, for example, NADH [19], norepinephrine [20], H2 S [21] and catechol [22]. CNTs can offer a number of advantages including a large surface area, high electrical conductivity, good chemical stability and significant mechanical strength. By conducting adsorptive voltammetry, detection limits down to nM of estradiol were reported by a number of groups after several minutes of accumulation at CNT-modified electrodes [23–26]. Equally important, several groups have demonstrated an antifouling capability of CNTs [27,28]. For example, in the chronoamperometric determination of homocysteine, Gong et al. [27] observed a virtually constant oxidation current at a CNT|Nafion-coated glassy carbon electrode over a 50-min duration, compared to a relatively fast decrease in the response at a Nafion-coated glassy carbon electrode. Similarly, Hu et al. [29] observed an enhanced oxidation of estradiol at a Congo Red-functionalised multi-walled CNT electrode without any appreciable fouling effect. However, the issue of electrode fouling by the oxidation products of estradiol has not been adequately addressed, particularly work relying on detection methods involving an accumulation step whereby electrode fouling was expected to be more severe. In this work, we report the development of a detection strategy for estradiol without appreciable electrode fouling by initially immobilising a glassy carbon electrode with a CNT–Nafion layer to act as an antifouling barrier, followed by Ni(cyclam) for the electrocatalytic oxidation of estradiol. The oxidation current obtained was used to relate quantitatively to the concentration of estradiol. In achieving this goal, we will also employ a twofactorial design method to identify the significant operational parameters, including their interactions, for the development of the composite electrode. Finally, the performance of the electrode will be evaluated in detecting estradiol in human serum samples. There is no accumulation time required in our strategy, making it a more efficient technique for the detection of estradiol. 2. Experimental 2.1. Reagents 17␤-Estradiol (1,3,5-estratriene-3,17␤-diol), Nafion (5 wt.% in lower aliphatic alcohols) and nickel(II) perchlorate hexahydrate (99.999% purity), analytical grade Na2 HPO4 , KH2 PO4 , and KCl were all purchased from Aldrich (Sydney, Australia). Multi-walled carbon nanotubes with 95% purity were purchased from Nanolab Inc. (Newton, MA, USA). Cyclam (Fluka) was

185

used without further purification. Ni(cyclam) was synthesised by following a previously reported procedure [30]. Phosphate buffered saline (PBS, 0.1 M) containing 50 mM Na2 HPO4 , 50 mM KH2 PO4 and 100 mM KCl was adjusted to pH 7.2 by adding 0.1 M NaOH. 2.2. Apparatus and procedure A three-electrode system, consisting of a CNT–Ni(cyclam)modified glassy carbon electrode, a Ag|AgCl reference electrode (Bioanalytical Systems Inc., West Lafayette, IN, USA) and a platinum counter electrode, was accommodated in a 20-mL electrochemical cell. Electrochemical experiments were performed using a Powerlab 2120 potentiostat (eDAQ Pty Ltd., Sydney, Australia) interfaced with a PC via a v2.0 EChem software (eDAQ). For square wave voltammetry, the pulse amplitude was set at 40 mV, frequency 20 Hz and potential step 2 mV. Prior to all voltammetric experiments, the electrolyte solution was degassed with nitrogen for 10 min and a blanket of nitrogen was maintained over the solution throughout the experiment. 2.3. Electrode preparation and modification Glassy carbon electrodes of 3 mm diameter (Bioanalytical Systems Inc.) were sequentially polished in aqueous slurries of alumina of 1.0, 0.3, and 0.05 ␮m, with sonication in doubledistilled water for 1 min between each polishing step. A solution containing 1 mg CNTs and 1 mL of 1% (v/v) Nafion solution was sonicated for 20 min so that the CNTs were uniformly dispersed in the solution [27,28,31–33]. A 10-␮L aliquot of this solution was directly applied on the polished glassy carbon electrode and allowed to air dry at room temperature for 30 min. Ni(cyclam)-modified electrodes were prepared by electropolymerising the Ni complex on CNT-modified glassy carbon electrodes in a known concentration of Ni(cyclam) in 0.1 M NaOH as supporting electrolyte by cyclic scanning between 0.0 and 1.0 V at a rate of 100 mV s−1 . The glassy carbon electrode was then dried at room temperature for at least 20 min before use. Different film thickness was obtained by varying the number of cycling. 2.4. Preparation of calibration solutions Initially, 0.5 mL of a protein-free human serum sample (obtained from a Sydney hospital) was spiked in 125 ␮L of a 2.0 mM estradiol stock solution and 0.5 mL CH3 CN. After vortexing for 30 s, the mixture was centrifuged for 20 min at 1000 g to remove serum protein residues. The supernatant was then diluted to 2.5 mL with 10% CH3 CN/90% 0.1 M NaOH (v/v). This spiked protein-free sample was divided into five equal parts and each was diluted to a 10 mL solution. In these experiments, CH3 CN was deliberately used to aid in the dissolution of estradiol in aqueous solutions, and also to minimise any organic functional groups that could interact with Ni(cyclam).

186

X. Liu, D.K.Y. Wong / Analytica Chimica Acta 594 (2007) 184–191

2.5. Statistical analysis Statistical significance of correlation coefficients was evaluated at a 95% confidence level using Student’s t-test. A two-level factorial design [34] was used to identify experimental factors and their interactions that have a statistical significance on the development of a CNT|Ni(cyclam) composite electrode for detecting estradiol. 3. Results and discussion 3.1. Voltammetric behaviour of estradiol at CNT-modified glassy carbon electrodes In the first step of our development, we have immobilised CNTs on a polished glassy carbon electrode by allowing a 10 ␮L aliquot of CNT–Nafion mixture to air dry for 30 min at room temperature. Scanning electron micrographs (not shown) of CNT–Nafion-coated glassy carbon electrodes, similar to those reported by Li et al. [35] and Wu and Hu [36], revealed an irregular structure arising from randomly aligned CNTs in the Nafion film. Cyclic voltammetry of 50 ␮M of estradiol in 0.1 M PBS (pH 7.2) containing 10% CH3 CN (v/v) was then carried out at this electrode (the presence of CH3 CN caused a buffer pH change of only 0.036 units, with a standard deviation of 0.001 units). In Fig. 1(a), five successive cyclic voltammograms obtained show a distinct oxidation peak at approximately 0.9 V, but there was no reduction peak in the reverse scan, indicating an electrochemically irreversible process. This is in agreement

Fig. 1. Five successive cyclic voltammograms of 50 ␮M estradiol at (a) a CNT modified and (b) a naked glassy carbon electrode in 0.1 M PBS (pH 7.2) containing 10% CH3 CN (v/v). Scan rate: 100 mV s−1 .

with work reported previously [10]. In these voltammograms, some minor peaks attributable to redox reactions of impurities on CNTs were also observed at potentials less anodic than 0.7 V. In this work, we obtained a 21% decrease, with a standard deviation of 0.02% (N = 6), in the oxidation peak current in the fifth scan relative to the first scan. When the experiments were repeated using a naked glassy carbon electrode, a rather broad oxidation peak between 0.8 and 1.2 V was obtained in the first scan, but the magnitude of this peak decreased dramatically starting from the second scan, as shown in Fig. 1(b). This is most likely due to the formation of an insulating dimer or quinone layer on the bare electrode that inhibited the oxidation reaction, as reported by Ngundi et al. [10] and Dempsey et al. [37]. These results demonstrated the antifouling characteristics of CNT-coated electrodes, which are consistent with those reported by others [25,27,28]. 3.2. Electropolymerisation Ni(cyclam) film and its electrocatalytic behaviour The next step of our development involves the immobilisation of Ni(cyclam) on the electrode. We have commenced this study by investigating the electrocatalytic behaviour of Ni(cyclam) alone in the oxidation of estradiol. In these experiments, we have

Fig. 2. (a) Twenty successive cyclic voltammograms of 2.0 mM Ni(cyclam) at a glassy carbon electrode and (b) cyclic voltammograms of a Ni(cyclam)-coated glassy carbon electrode in 0.1 M NaOH solution. Scan rate: 100 mV s−1 .

X. Liu, D.K.Y. Wong / Analytica Chimica Acta 594 (2007) 184–191

electropolymerised Ni(cyclam) by continuous cyclic voltammetry of 2.0 mM Ni(cyclam) in 0.1 M NaOH at a glassy carbon electrode between 0.0 and 1.0 V. Fig. 2(a) shows the successive cyclic voltammograms obtained for Ni(cyclam) film growth on the electrode, with the peak between 0.49 and 0.6 V corresponds to the oxidation of Ni(II) and that between 0.41 V and 0.51 V the reduction of Ni(III). The gradual increase in the magnitude of the voltammetric peaks is indicative of the formation of a film as a result of the anodic electropolymerisation of the Ni(cyclam) complex. The Ni(cyclam) modified electrode was then rinsed with water and transferred to a 0.1 M NaOH aqueous solution. The cyclic voltammogram obtained at the Ni(cyclam)-modified electrode in 0.1 M NaOH, shown in Fig. 2(b), exhibits similar Ni(II)|Ni(III) redox peaks to that depicted in Fig. 2(a). Note that there was no distinct decrease in the Ni(II)|Ni(III) redox peaks after storing the Ni(cyclam) modified electrode at 4 ◦ C for 2 weeks, indicating the stability of the Ni(cyclam) film on the glassy carbon electrode. Next, we have conducted cyclic voltammetry of 50 ␮M estradiol at the Ni(cyclam) modified electrode in 0.1 M NaOH containing 10% CH3 CN (v/v). For comparison, the same experiment was repeated in a blank solution. The results are shown in Fig. 3. As expected, the redox couple between 0.0 and 1.0 V observed in the cyclic voltammogram of the blank (trace A) are attributed to the Ni(II)/Ni(III) redox reaction. However, in the presence of estradiol (trace B), the oxidation peak was enhanced by at least a factor of 2, while the reduction peak was decreased. Such voltammetric behaviour is consistent with an electrocatalytic reaction in which estradiol reduced the electrochemically generated Ni(III) to Ni(II), giving rise to an enhanced oxidation peak when Ni(II) was re-oxidised at the electrode surface [13]. There is a small positive shift of the oxidation peak in trace B relative to trace A. This shift has possibly arisen from the additional energy needed for oxidation at an electrode surface adsorbed with the oxidation products of estradiol from the catalytic reactions with Ni(III). It is highly unlikely that a direct oxidation of estradiol took place in this voltammetric scan because such a reaction would require a potential more positive than 0.8 V (see Fig. 1).

Fig. 3. Cyclic voltammetric response of a Ni(cyclam)-coated glassy carbon electrode in the (A) absence and (B) presence of 50 ␮M estradiol in 0.1 M NaOH containing 10% CH3 CN (v/v). Scan rate: 100 mV s−1 .

187

Fig. 4. Twenty successive cyclic voltammograms of 2.0 mM Ni(cyclam) at a CNT-coated glassy carbon electrode in 0.1 M NaOH solution. Scan rate: 100 mV s−1 .

3.3. A CNT|Ni(cyclam) composite electrode A weak and ill-defined oxidation peak is a significant drawback in the electrochemical detection of estradiol, as shown in Fig. 1(b). In this work, we have developed a CNT|Ni(cyclam) composite electrode that will enhance the detection response of estradiol without severe electrode fouling. This involves a glassy carbon electrode initially coated with a CNT–Nafion layer as described in Section 3.1, followed by Ni(cyclam) electropolymerised on the CNT-coated electrode by cyclic voltammetry. Initially, to ascertain that Ni(cyclam) was successfully electropolymerised on the electrode, successive cyclic voltammetry of 2.0 mM of Ni(cyclam) complex was carried out at a CNTcoated glassy carbon electrode in 0.1 M NaOH and the results are shown in Fig. 4. Compared with the cyclic voltammogram in Fig. 3, both the nickel oxidation and reduction peaks in Fig. 4 have shifted positively. Nonetheless, the gradual increase in the voltammetric results indicates the formation of Ni(cyclam) film on the surface of a CNT-coated glassy carbon electrode. In addition, there was a notable increase in charging current arising from an increased surface area from the CNTs on the glassy carbon electrode, thus providing an indication of the presence of CNTs on the electrode surface. Based on the relationship, Γ = Q/nFA (where n denotes the number of electrons involved, F the Faraday’s constant and A the electrode surface area) and integrating the charge (Q) on the cyclic voltammogram, we have estimated the apparent surface coverage (Γ ) of Ni(cyclam) on the electrode to be 1.93 × 10−5 ␮mol cm−2 , with a standard deviation of 0.32 × 10−5 ␮mol cm−2 (N = 5). Multi-cyclic voltammetry of 50 ␮M estradiol was then conducted at the CNT|Ni(cyclam) modified electrode in 0.1 M NaOH containing 10% CH3 CN (v/v). The results are shown in Fig. 5. Notably, the oxidation current of estradiol was enhanced by a factor of 4 compared to the oxidation currents in Fig. 1(b). Meanwhile, the peak current was observed to have decreased by only 6%, with a standard deviation of 0.5%, after five repeated scans. We attribute this to a weak adsorption of the oxidation products of estradiol at the modified

188

X. Liu, D.K.Y. Wong / Analytica Chimica Acta 594 (2007) 184–191

Fig. 5. Five successive cyclic voltammograms of 50 ␮M estradiol at a CNT|Ni(cyclam)|glassy carbon composite electrode in 0.1 M NaOH containing 10% CH3 CN (v/v). Scan rate: 100 mV s−1 .

glassy carbon electrode and the surface was thus not severely fouled. 3.4. Optimisation of CNT|Ni(cyclam) modified electrodes The detection signal of estradiol at a CNT|Ni(cycalm) modified electrode will be affected by the ease of transport of Ni(II) (produced from the catalytic reaction between estradiol and Ni(III)) from the modified electrode-electrolyte solution interfacial region to the electrode surface for electron transfer reactions. Factors including the amount of CNTs, the thickness of Nafion, and the thickness of the Ni(cyclam) layer are therefore expected to exhibit effects on the transport of Ni(II). Among them, the thickness of Nafion film can be monitored by applying different

concentrations to the electrode surface, while that of Ni(cyclam) layer by the number of electropolymerisation scan cycles. In this work, we have employed a two-level factorial design [34] to identify the optimum values of these three parameters and possible interactions among them to obtain an optimum detection signal. Table 1 shows the experimental factors and their levels employed in a 23 factorial design. The amount of CNTs, thickness of Nafion and the thickness of the Ni(cyclam) layer are labelled X1 , X2 and X3 , respectively. Based on earlier experimental results and practical operation conditions, a low and a high level were assigned to each factor. Table 2 was then constructed for different computation of effects and interactions (X1 X2 , X1 X3 , X2 X3 and X1 X2 X3 ) for a 23 factorial design. Note that two replicate measurements were conducted at each condition by following a randomised order of experimentation (refer to column 9 in Table 2). Also, in Table 2, tE represents a signal-to-noise ratio calculated by the expression tE =

Effect √ 2sp / nF

Here, Effect represents the average between the high (+) and the low (−) values of each experimental parameter, while sp is a pooled standard deviation:   m  sp =  Variancei i=1

and m is the number of experimental conditions (m = 8 in this work), nF the number of experimental replications (nF = 16). When the absolute value of tE is larger than the tabulated tvalue (denoted by t* ) at the 95% confidence level, the effect of a particular factor or interaction between factors is con-

Table 1 Experimental factors and their levels employed in a 23 factorial design Factor

Definition

Label

Low level (−1)

High level (+1)

1 2 3

Amount of CNT (␮L) Concentration of Nafion (v/v) Surface coverage of Ni(cyclam) (represented by scan cycles)

X1 X2 X3

5 0.25% 5

25 2% 25

Table 2 Computation of significance of effects and interactions for a 23 factorial design Experimental conditions

X1

X2

X3

X1 X2

X1 X3

X2 X3

X1 X2 X3

Run order

Ya

Average Y

Variance

1 2 3 4 5 6 7 8

−1 1 −1 1 −1 1 −1 1

−1 −1 1 1 −1 −1 1 1

−1 −1 −1 −1 1 1 1 1

1 −1 −1 1 1 −1 −1 1

1.75 −1 1 −1 −1 1 −1 1

−8.25 1 −1 −1 −1 −1 1 1

−1 1 1 −1 1 −1 −1 1

14, 5 1, 9 13, 2 11, 4 3, 6 8, 16 10, 7 15, 12

40, 24 15, 32 37, 18 62, 26 27, 22 57, 72 29, 51 77, 67

32 23.5 27.5 44 24.5 64.5 40 72

128 144.5 180.5 648 12.5 112.5 242 50

Sumproductb Effect tE

80 20 2.899

39 9.75 1.413

74 18.5 2.681

17 4.25 0.616

64 16 2.319

7 1.75 0.254

−33 −8.25 −1.196

Cyclic voltammetry of 50 ␮M estradiol was conducted throughout the experiments. a Y is the oxidation peak current of estradiol obtained in each voltammetric run. b Sumproduct is the sum of the products of each parameter or interaction and average Y over the eight sets of experimental conditions.

t* = 2.306

X. Liu, D.K.Y. Wong / Analytica Chimica Acta 594 (2007) 184–191

189

Table 3 Peak currents obtained for the oxidation of 50 ␮M estradiol over the range of amount of Nafion (X1 ) and surface coverage of Ni(cyclam) (X3 )

sidered significant. The results in Table 2 reveal that the two main factors, X1 and X3 , as well as their interaction X1 X3 , are statistically significant. Therefore, to obtain the optimum values of X1 and X3 , different combinations of X1 and X3 values were used in the voltammetric determination of estradiol at the CNT|Ni(cyclam) modified electrodes. As displayed in Table 3, the results show that optimum detection of estradiol was obtained when X1 = 15 ␮L, X3 = 20 cycles. Consequently, all composite electrodes used below were prepared by air drying 15 ␮L of the CNT–Nafion mixture on a glassy carbon electrode for 30 min, followed by 20 cyclic voltammetric scans in 2.0 mM Ni(cyclam) in 0.1 M NaOH. 3.5. Analytical utility In order to minimise background currents and further improve the detection sensitivity, square wave voltammetry was applied to the detection of estradiol in synthetic laboratory standard solutions in 0.1 M NaOH containing 10% CH3 CN. As shown in Fig. 6, an estradiol oxidation peak was obtained at approximately 0.84 V under the optimum experimental conditions. When these

Fig. 6. Square wave voltammograms of 0.5, 1.5, 5.0, 10, 15, 20, 30 and 40 ␮M estradiol at a CNT|Ni(cyclam)|glassy carbon composite electrode in 0.1 M NaOH containing 10% CH3 CN (v/v). The electrode was fabricated using 15 ␮L of a CNT–Nafion mixture and 20 cyclic scans in 2.0 mM Ni(cyclam). The inset shows a calibration plot obtained based on the oxidation peak current in the voltammograms.

peak currents were plotted against estradiol concentration (see inset in Fig. 6), a linear relationship represented by the expression Peak current (μA) = (1.92 ± 0.14) (μA μM−1 ) [estradiol] + 2.51 ± 0.27

(1)

was obtained for estradiol concentration between 0.5 and 40 ␮M, where the errors represent the 95% confidence intervals. Its correlation coefficient of 0.994 (N = 8) was found to be statistically significant. Based on a signal-to-noise ratio of 3, the detection limit was estimated to be 0.06 ␮M, which is just below the typical level (0.18 ␮M or 50 ng mL−1 ) measured during a menstrual cycle. This detection limit compares favourably to 12.1 ␮M estradiol determined by differential pulse voltammetry at a naked glassy carbon electrode [1] and 84.3 ␮M estradiol in an amperometric detection at a glassy carbon electrode following an electrophoretic separation [2]. The detection repeatability of the composite electrode was also evaluated by successively measuring the same 5.0 ␮M estradiol solution 12 times at a single CNT|Ni(cyclam) modified electrode. In between each measurement, the modified electrode was thoroughly rinsed with water, and then used in ten cyclic voltammetric scans in PBS solution between 0.0 to 1.3 V to remove any adsorbed materials. A relative standard deviation of 3.3% was achieved, indicating good repeatability. This has then facilitated the feasibility of a rapid renewal of electrode surface. This characteristic of the CNT|Ni(cyclam) modified electrodes is likely to be due to the irregular structure and random alignment of CNTs and the large cyclam ring, resulting in a less compact film of the oxidation product of estradiol compared to that on naked electrodes [38,39], thereby allowing interfacial Ni(III) on CNT/Nafion to interact with estradiol. Further, Ni(cyclam) would catalytically convert the oxidation products of estradiol back to estradiol, reducing the effects of electrode fouling. Following a 2-week storage at 4 ◦ C, 6.3% of original current in the response to 5.0 ␮M of estradiol was lost, suggesting that the CNT|Ni(cyclam) modified electrodes exhibit an acceptable level of stability. The optimised procedure was subsequently adopted for the determination of estradiol in protein-free human serum samples containing 5.0 ␮M estradiol using standard addition calibration and the results are shown in Fig. 7. Note that other than the centrifugal protein separation, no additional extraction steps were

190

X. Liu, D.K.Y. Wong / Analytica Chimica Acta 594 (2007) 184–191

Scheme I. Reaction mechanism for the oxidation of estradiol [2,10].

required prior to the assay of samples. This calibration plot can be represented by the expression

4. Conclusion

Peak current (μA) = (1.98 ± 0.29) (μA μM−1 ) [estradiol]

A major contribution of this work is the application of CNT|Ni(cyclam)-coated glassy carbon electrodes to achieve the electrocatalytic oxidation of estradiol with minimal fouling effects by the oxidation product of estradiol. In addition, a two-level factorial design was used to determine the optimum experimental parameters affecting the detection of estradiiol. Based on calibration in synthetic laboratory standard solutions, a detection limit of 60 nM estradiol was estimated, which is below the typical estradiol level in a normal menstrual cycle. In addition, comparable sensitivity in synthetic laboratory standard solutions and human serum samples was obtained, indicating minimal interference effects at the CNT|Ni(cyclam)-coated electrodes in the serum sample matrix.

+ 10.5 ± 3.5

(2)

where the errors are again the 95% confidence intervals. An average of 5.3 ␮M with a standard deviation of 0.1 ␮M (N = 6) estradiol was determined from the intercept on the abscissa. The 6% higher recovery than expected is possibly due to the oxidation of other compounds containing phenolic groups in the serum samples. On the other hand, by denoting the slope of the calibration plot as the sensitivity of the technique, a comparable level of sensitivity at 1.92 ± 0.14 ␮A ␮M−1 was achieved in synthetic laboratory standard solutions, relative to 1.98 ± 0.29 ␮A ␮M−1 in human serum samples, indicating minimal interference effects on the sensitivity of the CNT|Ni(cyclam) electrodes.

References

Fig. 7. Square wave voltammograms of 5.0, 10, 15, 20 and 25 ␮M estradiol at a CNT|Ni(cyclam)|glassy carbon composite electrode) in a protein-free human serum sample mixed with 0.1 M NaOH containing 10% CH3 CN (v/v). The electrode was fabricated using 15 ␮L of a CNT–Nafion mixture and 20 cyclic scans in 2.0 mM Ni(cyclam). The inset shows a calibration plot obtained based on the oxidation peak current in the voltammograms.

[1] B. Salci, I. Biryol, J. Pharm. Biomed. Anal. 28 (2002) 753. [2] A. Collier, J. Wang, D. Diamond, E. Dempsey, Anal. Chim. Acta 550 (2005) 107. [3] A.I. Mohammedebid, M.J. Shelton, J.M. Adams, J.A. Cordaro, L.J. Gugino, M. Maliszewski, R. Colomaio, R.G. Hewitt, G.D. Morse, Clin. Drug Invest. 21 (2001) 715. [4] A.M. Gronowski, M.L. Landau, in: C.A. Burtis, E.R. Ashwood (Eds.), Tietz Fundamentals of Clinical Chemistry, Vol. 18, W.B. Saunders, Philadelphia, London, 2001, p. 1601. [5] M. Wheeler, in: D. Wild (Ed.), The Immunoassay Handbook, Vol. 9, Stockton Press, New York, 1994, p. 366. [6] H. Kuramitz, M. Matsuda, J.H. Thomas, K. Sugawara, S. Tanaka, Analyst 128 (2003) 182. [7] A. Penalver, E. Pocurull, F. Borrull, R.M. Marce, J. Chromatogr. A 964 (2002) 153. [8] F. Valentini, D. Compagnone, A. Gentili, G. Palleschi, Analyst 127 (2002) 1333. [9] Y. Matsumoto, H. Kuramitz, S. Itoh, S. Tanaka, Talanta 69 (2006) 663. [10] M.M. Ngundi, O.A. Sadik, T. Yamaguchi, S.-I. Suye, Electrochem. Commun. 5 (2003) 61. [11] M.I. Paredes, M.C. Gonz´alez, M.D. V´azquez, M.L. Tasc´on, P.S. Batanero, G.J. Patriarche, Electroanalysis 6 (1994) 791. [12] A.L. Woodson, D.E. Smith, Anal. Chem. 42 (1970) 242. [13] J. Manriquez, J.L. Bravo, S. Gutierrez-Granados, S. Sucar Succar, C. BiedCharreton, A. Alatorre Ordaz, F. Bedioui, Anal. Chim. Acta 378 (1999) 159.

X. Liu, D.K.Y. Wong / Analytica Chimica Acta 594 (2007) 184–191 [14] V.C. Dall’orto, C. Danilowicz, S. Sobral, A.L. Balbo, I. Rezzano, Anal. Chim. Acta 336 (1996) 195. [15] S. Trevin, F. Bedioui, M.G.G. Villegas, C. Bied-Charreton, J. Mater. Chem. 7 (1997) 923. [16] C. Aleksander, Electroanalysis 7 (1995) 1132. [17] G.-R. Xu, H.-Y. Chang, H. Cho, W. Meng, I.-K. Kang, Z.-U. Bae, Electrochim. Acta 49 (2004) 4069. [18] X. Lin, Y. Li, Biosens. Bioelectron. 22 (2006) 253. [19] M. Musameh, J. Wang, A. Merkoci, Y. Lin, Electrochem. Commun. 4 (2002) 743. [20] R.R. Moore, C.E. Banks, R.G. Compton, Anal. Chem. 76 (2004) 2677. [21] N.S. Lawrence, R.P. Deo, J. Wang, Anal. Chim. Acta 517 (2004) 131. [22] F. Valentini, A. Amine, S. Orlanducci, M.L. Terranova, G. Palleschi, Anal. Chem. 75 (2003) 5413. [23] N. Terui, B. Fugetsu, S. Tanaka, Anal. Sci. 22 (2006) 895. [24] Y. Sun, K. Wu, S. Hu, Microchim. Acta 142 (2003) 49. [25] S. Hu, K. Wu, H. Yi, D. Cui, Anal. Chim. Acta 464 (2002) 209.

191

[26] Q. He, S. Yuan, C. Chen, S. Hu, Mater. Sci. Eng. C23 (2003) 621. [27] K. Gong, Y. Dong, S. Xiong, Y. Chen, L. Mao, Biosens. Bioelectron. 20 (2004) 253. [28] J. Wang, M. Musameh, Y. Lin, J. Am. Chem. Soc. 125 (2003) 2408. [29] C. Hu, C. Yang, S. Hu, Electrochem. Commun. 9 (2007) 128. [30] B. Bosnich, M.L. Tobe, G.A. Webb, Inorg. Chem. 4 (1965) 1109. [31] R.P. Deo, N.S. Lawrence, J. Wang, Analyst 129 (2004) 1076. [32] J. Tkac, T. Ruzgas, Electrochem. Commun. 8 (2006) 899. [33] C. Hu, S. Yuan, S. Hu, Electrochim. Acta 51 (2006) 3013. [34] J. Lawson, J. Erjavec, Modern Statistics for Engineering and Quality Improvement, Duxbury, CA, USA, 2001, chapter 8. [35] Y. Li, X. Lin, C. Jiang, Electroanalysis 18 (2006) 2085. [36] K. Wu, S. Hu, Microchim. Acta 144 (2004) 131. [37] E. Dempsey, D. Diamond, A. Collier, Biosens. Bioelectron. 20 (2004) 367. [38] J. Obirai, F. Bedioui, T. Nyokong, J. Electroanal. Chem. 576 (2005) 323. [39] M.S. Ureta-Zanartu, C. Berrios, J. Pavez, J. Zagal, C. Gutierrez, J.F. Marco, J. Electroanal. Chem. 553 (2003) 147.