Anal. Chem. 2007, 79, 7431-7438
Picomolar Detection of Insulin at Renewable Nickel Powder-Doped Carbon Composite Electrode Abdollah Salimi,*,†,‡ Mahmoud Roushani,† Saied Soltanian,‡,§ and Rahman Hallaj†
Department of Chemistry, Research Center for Nanotechnology, and Department of Physics, University of Kurdistan, P.O. Box 416, Sanandaj, Iran
A sol-gel technique was used for fabrication of a renewable carbon composite electrode (CCE) modified with nickel powder. This modified electrode shows excellent catalytic activity for the oxidation of insulin in alkaline solutions. The nickel powder was then oxidized to form a nickel oxide film electrode, which was used as an amperometric detector for hydrodynamic amperometry and flow injection analysis of insulin. It was found that the calibration curve was linear up to 30 µM with a detection limit of 40 pM under the optimized conditions for hydrodynamic amperometry using a rotating disk modified CCE. Flow injection amperometric determination of insulin at this modified electrode yielded a calibration curve with the following characteristics; linear dynamic range of 15-1000 pM, sensitivity of 8659.23 pA pM-1 cm-2, and detection limit of 2 pM. This electrode shows many advantages as an insulin sensor such as simple preparation method without using any specific electron-transfer mediator or specific reagent, high sensitivity, excellent catalytic activity, short response time, long-term stability, and remarkable antifouling property toward insulin and its oxidation product. Sensitivity, detection limit, and antifouling properties of this insulin sensor are better than all of the reports in the literature. Additionally, it is promising for monitoring insulin in chromatographic effluents. Insulin is an important polypeptide that controls the glucose level in blood within a narrow concentration range. Insulin sensing is greatly important for clinical diagnostics, because it serves as a predictor of diabetes of insulinoma and trauma.1,2 Bioassay,3 immunoassay,4 and chromatographic methods5 are the main procedures for insulin determination. The assay systems are lengthy, relatively imprecise, and insensitive, and they cannot be used by clinical laboratories. Immunoassay of insulin has been widely used for the determination of insulin in biological specimens. However, cross reactions and nonspecific binding with the * Corresponding author. Fax: +98-871-6624001. E-mail:
[email protected] or
[email protected]. † Department of Chemistry. ‡ Research Center for Nanotechnology. § Department of Physics. (1) Ortmeyer, H. K.; Bodkin, N. L.; Hansen, B. C. Diabatologia 1994, 37, 127133. (2) Jenssen, T. G.; Toft, I.; Bonaa, K. H. Diabatologia 1995, 38, 127A. (3) Kowaraski, C. R.; Bado, B.; Shah, S.; Kowaraski, A. A. J. Pharm. Sci. 1983, 72, 692-693. (4) Yalow, R. S.; Berson, S. A. Nature 1959, 184, 1648-1649. (5) Ohkubo, T. Biomed. Chromatogr. 1994, 8, 301-305. 10.1021/ac0702948 CCC: $37.00 Published on Web 08/24/2007
© 2007 American Chemical Society
coexistent biomolecules to the anti-insulin antibody are the major interferences with the precise determination of insulin in biosamples.6 In contrast to immunoassays, the chromatographic methods are more reliable for selectivity since the separation steps are included. HPLC with UV-vis,5 fluorescence,7 and mass spectrometry with isotope dilution assay(IDA)8 detectors have been used for insulin determination. However, since a picomolarlevel injection on column needs to be achieved, HPLC with UVvis detection does not appear to be a useful instrument due to its poor sensitivity for insulin detection in biological samples. The fluorescence detector needs fluoregenic reagents for derivatization. The IDA method allows precise determination of human insulin at physiological concentration in blood, but labeled insulin with stable isotopes is necessary as an internal standard. It is also an expensive instrument. Direct electrooxidation of insulin is important for the development of fast and sensitive amperometric detectors coupled to flow systems or chromatographic instruments for this hormone. Direct oxidation of insulin at conventional electrodes was limited by the slow kinetics and surface fouling onto electrochemical devices. In addition, low sensitivity and reproducibility, low stability over a wide range of solution compositions, and high overpotential at which the insulin oxidation process occurred are other limitations of using the bare and unmodified electrodes as an electrochemical sensor for insulin detection. The chemical modification of inert substrate electrodes with redox-active thin films offers significant advantages in the design and development of the electrochemical insulin sensor. Different redox mediators, such as mixed-valence ruthenium oxide/cyanoruthenate (RuO/CN-Ru),9-11ruthenium oxide,12-14 Ru(II) metallodendrimer,15,16 and iridium oxide,17 have been used for electrode modification in insulin measurement. Furthermore, electrochemical oxidation of insulin at electrodes modified with carbon nanotubes18,19 and RuOx/carbon nanotube-modified carbon (6) Clark, P. M. Ann. Clin. Biochem. 1999, 36, 541-564. (7) Toriumi, C.; Imai, K. Anal. Chem. 2002, 74, 2321-2327. (8) Kipen, D. A.; Cerini, F.; Vadas, L.; Stocklin, R.; Vu, L.; Offord, R. E.; Rose, K. J. Biol. Chem. 1997, 272, 12513-12522. (9) Cox, J. A.; Garry, T. G. Anal. Chem. 1989, 61, 2462-2464. (10) Gorski, W.; Cox, J. A. J. Electroanal. Chem. 1992, 389, 123. (11) Kennedy, R. T.; Hung, L.; Atkinson, M. A.; Dush, P. Anal. Chem. 1993, 65, 1882. (12) Gorski, W.; Aspinwall, C. A.; Lakey, J. R. T.; Kennedy, R. T. J. Electroanal. Chem. 1997, 425, 191-199. (13) Kennedy, R. T.; Hung, L.; Aspinwall, C. A. J. Am. Chem. Soc. 1996, 118, 1795-1797. (14) Wang, J.; Zhang, X. Anal. Chem. 2001, 73, 844-847. (15) Holmstrom, S. D.; Cox, J. A. Anal. Chem. 2000, 72, 3191-3195. (16) Cheng, L.; Pacey, E. G.; Cox, J. A. Anal. Chem. 2001, 73, 5607-5610. (17) Pikulski, M.; Gorski, M. Anal. Chem. 2000, 72, 2696-26702. (18) Wang, J.; Musameh, M. Anal. Chim. Acta 2004, 511, 33-36. (19) Zhang, M.; Mullens, C.; Gorski, W. Anal. Chem. 2005, 77, 6396-6401.
Analytical Chemistry, Vol. 79, No. 19, October 1, 2007 7431
electrodes have been investigated.20 Although the modified electrodes have been successfully employed for monitoring the insulin, they usually have many disadvantages such as reduced stability of mediators and electrocatalyst under physiological condition,9 poor long-term stability of insulin oxidation response,19 high detection limit,14,17-19 and complicated multistep preparation methods (using specific and expensive reagents).15,16 They are also neither renewable nor polishable after fouling the electrode surface during insulin oxidation. Therefore, preparation of a less complex insulin detector with minor limitations is desirable for insulin detection. Due to simplicity, physical rigidity, transparency, porosity, permeability, versatility, and flexibility of the sol-gel technique, it is useful for the fabrication of electrochemical sensors and biosensors.23,24 We have successfully employed carbon composite electrodes modified with [Ru(bpy)(tpy)Cl]PF6 complex23 and sol-gel-derived carbon ceramic electrode modified with potassium octacyano molybdate(IV)24 for nanomolar detection of insulin. Both pure nickel electrodes and nickel-coated electrodes have been found to have various applications in the fields of electrocatalysis,25 alkaline batteries,26 electrochromic devices,27 and electroanalytical chemistry.25,28,29 Many electrodes were modified and developed by depositing of Ni particles and nanoparticles on traditional electrode surfaces such as diamond, gold, carbon, or graphite.25,29-31 Most of these applications originated from the application of the redox pair system of Ni(OH)2/NiOOH, which is formed on the electrode surface in the alkaline medium.32 One of the problems in using nickel-based chemically modified electrodes is the gradual change in the mechanical integrity of the catalytic film for extended periods of time under applied potential, partial solubilization of the catalyst, electrode poisoning, and surface fouling. Recently we have developed chemically modified carbon composite electrodes with improved sensitivity that exhibited good resistance to surface fouling, resulting in longterm stability of the electrodes. We have also showed great improvements in sensitivity and reproducibility of these modified electrodes for determination of sulfur derivative compounds such as cysteine, cystine, glutathione, and sulfur oxoanions.33-36 Here we have described more developments in nickel powderdoped modified electrodes that offer substantial improvements in the stability of mediators and electrocatalysts and antifouling (20) Wang, J.; Tangkuaram, T.; Loyprasert, S.; Vazquez-Alvarez, T.; Veerasai, W.; Kanatharana, P.; Thavarungkul, P. Anal. Chim. Acta 2007, 581, 1-6. (21) Walcarius, A. Electroanalysis 2001, 13, 701-718. (22) Lev, O.; Wu, Z.; Bharhti, S.; Glezer, V.; Modestor Guan, J.; Rabinovich, L.; Sampath, S. Chem. Mater. 1997, 9, 2354-2375. (23) Salimi, A.; Pourbeyram, S.; Hadazadeh, H. J. Electroanal. Chem. 2003, 542, 33-49. (24) Salimi, A.; Roushani, M.; Haghighi, B.; Soltanian, S. Biosens. Bioelectron. 2006, 22, 220-226. (25) You, T.; Niwa, O.; Chen, Z.; Hayashi, K.; Tomita, M.; Hirono, S. Anal. Chem. 2003, 75, 5191-5196. (26) Vidts, P. D.; Delgado, J.; Wu, B.; See, D.; Kosanovich, K.; White, R. E. J. Electrochem. Soc. 1998, 145, 3874. (27) Garcia-Miquel, J. L.; Zhang, Q.; Allen, S. J.; Rougier, A.; Blyr, A.; Davies, H. O.; Jones, A. C.; Leedham, J.; Williams, A.; Impey, S. A. Thin Solid Films 2003, 424, 165-170. (28) Giovanelli, D.; Lawrence, N. C.; Jiang, L.; Jones, T. G. L.; Compton, R. G. Sens. Actuators B 2003, 88, 320-328. (29) Casella, I. G.; Gatta, M. Anal. Chem. 2000, 72, 2969-2975. (30) Xiang, C. H.; Xie, Q. J.; Yao, S. Electroanalysis 2003, 15, 987-990. (31) Casella, I. G.; Gatta, M. Electroanalysis 2001, 13, 549-554. (32) Weng, Y. C.; Rick, J. F.; Chou, T. C. Biosens. Bioelectron. 2004, 20, 41-45. (33) Salimi, A.; Hallaj, R.; Amini, M. K. Anal. Chim. Acta 2005, 534, 335-344. (34) Salimi, A.; Roushani, M. Electroanalysis 2006, 18, 205. (35) Salimi, A.; Pourbeyram, S.; Amini, M. K. Analyst 2002, 127, 1649-1656. (36) Salimi, A.; Roushani, M.; Hallaj, R. Electrochim. Acta 2006, 51, 1952-1959.
7432
Analytical Chemistry, Vol. 79, No. 19, October 1, 2007
properties of carbon composite surfaces as well as sensitivity of either voltammetric or amperometric detection of insulin. Due to chemical stability and electrochemical reversibility of the Ni(II)/ Ni(III) redox couple, in this study, the three-dimensional nickeldoped carbon ceramic electrode has been successfully used for oxidation of insulin. Finally, analytical performance of reagentless, renewable sol-gel-derived carbon ceramic electrode containing Ni powder has been described as an amperometric detector for picomolar concentrations of insulin in flow system. The results suggest great promise for miniaturized sensors for monitoring trace amount of insulin. EXPERIMENTAL SECTION Chemicals. High-purity graphite powder was obtained from Merck. Methyltrimethoxysilane (MTMOS) was purchased from Fluka and used without any further purification. Bovine insulin (5800, >27 USP units mg-1) was from Sigma. A stock solution of 0.1 mM was prepared in water using HCl to adjust the pH to 2. Insulin solutions were prepared by diluting aliquots of the insulin stock solutions with a background electrolyte solution. Nickel powder (2-3 µm), Na3PO4, Na2HPO4, NaH2PO4, NaOH, and HCl were of analytical grade from Merck or Fluka and used as received. Buffer solution (0.1M) was prepared from Na3PO4, Na2HPO4, NaH2PO4, NaOH, and HCl for the pH range 7-13. All solutions were prepared using deionized water with resistance greater than 18 MΩ cm. All solutions were sparged with argon gas for at least 5 min before use. Instrumentation. Surface morphology was examined using Jeol scanning electron microscope (SEM). Amperograms were carried out with a Metrohm multipurpose instrument model 693VA processor, equipped with a 694 VA stand. Cyclic voltamograms were performed via using an Autolab modular electrochemical system (Eco Chemie, Ultrecht, The Netherlands) equipped with a PSTA 20 module and driven by GPES (Eco Chemie) in conjunction with a three-electrode system and a personal computer for data storage and processing. An Ag/AgCl electrode (in 3 M KCl) was used as the reference electrode. A platinum wire counter electrode and a Ni-modified carbon ceramic electrode (prepared as follows) were employed for the electrochemical studies. A Metrohm drive shaft to rotate working electrodes was used in amperometric detection. The electrochemical measurements were carried out at thermostated temperature of 25.0 ( 0.1 °C. Flow Injection Assembly Setup. Amperometric measurement was done using a single line flow injection manifold with a three-electrode electrochemical cell of the wall-jet type.37 An Ag/ AgCl reference electrode was present in a circular chamber filled with 3 M KCl supplied with an external syringe. A separate chamber containing the working and a Pt auxiliary electrode (encircling the chamber) was connected to the chamber containing the reference electrode through four holes concentrically surrounding the inlet. The working electrode was a carbon ceramic electrode modified with nickel powder. The distance between the inlet nozzle and working electrode was ∼2 mm. A Waltson Marlow peristaltic pump model 205CA (Delden, The Netherlands) was equipped with silicon tubing (0.76-mm i.d.) propelled carrier, a 0.1 M phosphate buffer solution, pH 13, with the flow rate of 0.8 mL min-1 into the flow line (Supelco Teflon tubing 0.5-mm i.d.). A working potential of 0.45 V was applied to the electrodes in the standard way using a Metrohm amperometric detector (791 IC-VA detector, Herisau, Switzerland). The sample (37) Appelqvist, R.; Marko-Varga, G.; Gorton, L.; Torstensson, A.; Johansson, G. Anal. Chim. Acta 1985, 169, 237-247.
solution (75 µL) was injected into a carrier stream via a Rheodyne six-way Teflon rotary valve type 50, and the output signal was transferred to a Pentium II 400-MHZ computer at 0.1-s intervals via RS-232 port using a 16-bit data acquisition module (Axiomtek 1-7018, Taiwan) operated with a graphical Lab VIEW program under Windows 2000. All measurements were performed at room temperature. Preparation of the Bare and Ni Powder-Modified Carbon Composite Electrode (CCE). The blank carbon ceramic electrode was prepared according to the procedure described by Lev and co-workers,38 by mixing 0.2 mL of MTMOS, 0.3 mL of methanol, and 10 µL of hydrochloric acid (11 M). This mixture was magnetically stirred for 2 min, after adding 0.3 g of graphite powder. Then the resultant mixture was shaken for an additional 1 min. A piece of Teflon tube with 5-mm length and 3-mm inner diameter was filled with the sol-gel carbon mixture and dried under ambient conditions (25 °C) for 48 h. The modified CCEs were also prepared as follows. A solution of 0.3 mL of methanol containing 16 mg of Ni powder, 0.2 mL of MTMOS, and 10 µL of hydrochloric acid (11 M) was mixed and stirred for 2 min to ensure uniform mixing; 0.284 g of graphite powder was then added to the mixture and the resultant mixture was shaken for additional 3 min. The mixture was inserted into a Teflon tube (with 2-mm inner diameter and 5-cm length and the length of composite material in the tube, which was ∼0.3 cm) and dried for 48 h at room temperature. Electric contact was made with a copper wire through the back of the electrode. The electrodes were polished with polishing paper and subsequently rinsed with distilled water. In the amprometric experiment, the modified CCE contact was made by a Metrohm drive shaft. RESULTS AND DISCUSSION Nickel Hydroxide Film formation. The nickel hydroxide layer was first generated using cyclic voltammetry (CV). A cyclic voltammogram of the modified electrode at potential range -0.6 to +0.65 V in pH 13 buffer solution was recorded (not shown). The oxidation of bulk nickel was studied in two different regions, Ni(0) to Ni(II) and Ni(II) to higher valence oxides. At potential 0.3 V, nickel is oxidized to R-nickel hydroxide according to the following reaction
M NaOH is shown in Figure 1. As a characteristic of conducting film formation, the cathodic and anodic waves grew with the number of scans up to the 200th cycle and then a current plateau and stable voltammetric response was obtained. A large anodic current is observed at the beginning of the first cyclic voltammetric scan and is due to the oxidation of Ni to R-Ni(OH)2. The modified electrode exhibits significant oxidation currents starting around 0.42 V versus the reference electrode and a peak was observed at 0.55 V, indicating oxidation of R-Ni(OH)2 and β-Ni(OH)2 to β-NiOOH or γ-NiO2.39 With increase of the cycle number, the R-Ni(OH)2 has been transferred to β-Ni(OH)2, and β-Ni(OH)2 is oxidized to β-NiOOH at less positive potential (Ep ) 0.45).39 Furthermore, more than one cathodic CV peak was observed for the reduction of these oxides/hydroxides back to Ni(OH)2.28,32 The shoulder peak observed at 0.3 V during the cathodic process originates from the reduction of molecular oxygen, which was produced and adsorbed on the electrode surface during the anodic scan. The growth of the nickel hydroxide film was monitored as a function of potential cycle number by measurement of the surface concentration of nickel hydroxide sites (Γ) obtained for each growth cycle. The relevant plot, total surface concentration of β-Ni(OH)2 versus cycle number, calculated in the range 2-200 cycle numbers, give a straight line with a correlation coefficient of 0.9987. In the present case, the calculated value of Γc is 9.5 nmol‚cm-2, corresponding to ∼200 cycles. The pH effect on the electrochemical behavior of the nickel powder-doped carbon ceramic electrode has been investigated. The cyclic voltammograms of the modified electrode at different pH solutions were recorded (not shown). Results indicated that the currents response is increased with increasing pH values. At the same time, both reduction and oxidation peak potentials of the β-Ni(OH)2/βNiOOH redox couple were shifted to less positive values and no peak currents were observed at pH values below 9. This response is indicates the redox reaction process coupled with proton transfer, which is depicted as follows.
β-Ni(OH)2 + OH- f β-NiOOH + H2O + e-
Ni + 2OH- f R-Ni(OH)2 + 2eThe R-Ni(OH)2 has been transformed to β-Ni(OH)2 slowly during the potential scan and accumulated on the electrode surface.39,40 The cathodic peak observed at -0.3 V is due to reduction of unconverted R-Ni(OH)2 to Ni. Moreover, the oxidation peak current of Ni to R-Ni(OH)2 is decreased during the consequence potential cycling, indicating that β-Ni(OH)2 has now passivated the electrode surface preventing further oxidation of Ni from taking place.39,40 This initial voltammetric response is in good agreement with the literature.41,42 A typical example of cyclic voltammograms obtained during continuous potential cycling (voltage cycling between 0.2 and 0.55 V vs SCE) for nickel oxide film growth on a nickel-doped carbon composite electrode in 0.1 (38) Tsionsky, M.; Gun, G.; Glezer, V.; Lev, O. Anal. Chem. 1994, 66, 17471754. (39) Zhang, C.; Park, S. M. J. Electrochim. Soc. 1989, 136, 3333-3342. (40) Zhang, C.; Park, S. M. J. Electrochim. Soc. 1987, 134, 2966-2970. (41) Giovanelli, D.; Lawrence, N. C.; Jiang, L.; Jones, T. G. L.; Compton, R. G. Analyst 2003, 128, 173-177. (42) Barbosa, M. R.; Bastos, J. A.; Gracia-Jareno, J. J.; Vicente, F. Electrochim. Acta 1998, 44, 957-965.
The effect of potential scan rate on the responses of cyclic voltammograms was examined in the range 10-500 mV‚s-1(not shown). The ratio of cathodic to anodic peak currents is less than unity, which indicates some side reaction such as oxygen evaluation appears to contribute to the anodic current.40 The cathodic and anodic peak currents are directly proportional to the scan rate of the potential. This result indicates a surface-confined redox process corresponding to a rapid conversion of a surface film without diffusion or a kinetically controlled reaction step. In addition, the shift of peak potential is negligible at scan rates of 10-200 mV‚s-1, suggesting facile charge-transfer kinetics over this range of scan rates. The peak potentials were found to depend on the scan rate at higher sweep rates. This is the reflection of the relatively slow diffusion of hydroxide ions into the limited wetting section of electrode surfaces. In order to investigate renewal repeatability of the modified electrode, the 10th cyclic voltammograms for 8 successive polishings of the electrode at a scan rate of 100 mV‚s-1 in 0.1 M NaOH solution were recorded (not shown). No considerable change in currents and potential peaks were found after each polishing. The RSD was 3% by Analytical Chemistry, Vol. 79, No. 19, October 1, 2007
7433
Figure 1. Cyclic voltammograms of nickel hydroxide film growth on a carbon composite electrode in 0.1 M NaOH solution. The potential was continuously cycled at a scan rate of 100 mV‚s-1 between 0.2 and 0.55 V vs Ag/AgCl reference electrode.
measuring the anodic peak current, which reflects the repeatability of surface renewal by polishing. SEM imaging of an electrodeposited nickel hydroxide film on a carbon composite electrode was performed in order to investigate the formation and growth of the nickel oxide particles (not shown). The SEM image shows that high concentrations of small particles are grown during potential cycling on the carbon ceramic surface. In addition to small submicrometer particles distributed on the surface, large agglomerated particles are also observed on the image. It is also found that the nickel oxide particle size varies from less than 1 µm to less than 10 µm. In order to find the composition of these particles, spot scan EDS were performed (not shown). The compositions of these particles are clearly identified as nickel and oxygen. Electrocatalytic Oxidation of Insulin at the Ni-Doped Modified Carbon Composite Electrode. Cyclic voltammograms were obtained in the presence and absence of insulin at bare and Ni powder-modified carbon composite electrodes in order to examine the electrocatalytic activity of the modified electrodes. Figure 2A shows the recorded cyclic voltammograms of modified and bare carbon ceramic electrode in the absence and presence of insulin in 0.1 M NaOH solution. A dramatic enhancement in anodic peak current as well as reduction in cathodic peak current appeared by the addition of 2.4 µM insulin, indicating a strong catalytic effect. The anodic wave for insulin oxidation starts at 0.20 V and anodic peak potential is ∼0.43 V, while at the surface of the bare CCE electrode the insulin was not oxidized at the potential range from 0.0 to 0.8 V. Thus, decreases in overpotential and enhancement peak current for insulin oxidation are achieved with the modified electrodes. In order to optimize the electrocatalytic response of modified electrodes toward insulin oxidation, the effect of pH on the catalytic oxidation behavior was investi7434 Analytical Chemistry, Vol. 79, No. 19, October 1, 2007
gated. The cyclic voltammograms of nickel oxide film-modified CCE in 0.4 µM insulin concentration at different pH values (913) were recorded (not shown). With increasing pH values, the insulin oxidation peak potential shifts to a less positive value and the peak current is increased. Since more reproducible results and high catalytic activity of the modified electrode are observed at pH 13, this pH value was selected as optimum pH for insulin determination. Electrocatalytic activity of the modified electrode was monitored as a function of the Ni concentration ratio in the modified carbon composite electrode. The cyclic voltammograms of the modified electrodes containing different concentrations of Ni (2, 4, 8, 16, and 30 mg) were recorded in phosphate buffer solution (pH 13) in the presence of insulin. The current ratio (Icat/ INi) remained practically constant for the selected concentration range of nickel in the carbon composite electrodes (Figure 2B). The effect of adding insulin to the system was investigated in the range 0.2-1.4 µM (Figure 3). A linear dependence of the catalytic currents versus concentration of insulin can be fitted in the equation I(µA) ) 6.587[insulin] µA‚µM + 0.0763 µA and R ) 0.996 62. Protein fouling of electrode surfaces represents a significant problem in diagnostic systems,43 and as a result, there has been a great deal of interest in devising strategies to minimize surface fouling. The antifouling property and stability of electrocatalytic activity of Ni-CCE for oxidation of insulin is examined by repetitive scanning at scan rate of 10 mV/s As shown in Figure 4, the electrocatalytic currents are constant with scan numbers and currents remain at 99% of the initial value after 20 cycles. Nonrecognizable reduction of the insulin oxidation currents indicates stable activity of the film surface and reflects the (43) Asuri, P.; Karajanagi, S. S.; Kane, R. S.; Dordick, J. S. Small 2007, 3, 5053.
Figure 2. (A) Cyclic voltammograms of Ni powder-doped modified CCE in 0.1 M NaOH solution at a scan rate of 10 mV‚s-1 in the absence (c) and presence of 2.4 µM insulin (d). (a) and (b) as (c) and (d) for CCE without nickel. (B) (a and b) Cyclic voltammograms of CCE modified with 8 and 12 mg of nickel powder, (a1 and b1) as (a and b) in solution contaning 1 µM insulin, pH 13, and a scan rate of 10 mV‚s-1. (c) plot of Icat/INi vs mg of nickel powder in CCE.
Figure 3. Voltammetric response of Ni powder-doped modified CCE placed in a 0.1M NaOH solution at a scan rate of 10 mV‚s-1 to additions of insulin, from inner to outer 0.0, 0.2, 0.4, 0.6, 0.8, 1.0, 1.2, and 1.4 µM. Inset: plot of catalytic peak current vs insulin concentration.
remarkably progressive antifouling properties of the film with the adsorption of the thyil radical’s products of disulfide oxidation process of insulin19 and stability of the mediator. Moreover, fouling of the electrode surface with albumin is negligible (not shown). The reproducibility of nickel powder-doped modified CCE for catalytic oxidation of insulin is evaluated by 6 successive polishings and with 10 potential cycles modification. Then cyclic voltammograms are recorded in 0.4 µM insulin solution (the relative standard deviation (RSD) for six measured anodic peak currents was 3%). The RSD of the peak currents of 0.4 µM insulin for eight repeated determinations is also 2.5%. Then at the surface of
modified CCE not only the overvoltage for insulin oxidation decreased but also the antifouling properties of ceramic composite improved reproducibility. Cyclic voltammograms of 0.4 µM insulin solution at different scan rates were recorded (not shown). The peak current for the anodic oxidation of insulin is proportional to the square root of scan rate, which is the ideal case for quantitative applications. The results show that by increasing the sweep rates the peak potential for the catalytic oxidation of insulin shifts to a more positive value and the plot of peak currents versus square root of scan rate deviates from linearity, suggesting kinetic limitations in the reaction between the redox sites of the nickel Analytical Chemistry, Vol. 79, No. 19, October 1, 2007
7435
Figure 4. (a) Cyclic voltammogram of Ni powder-doped modified CCE in pH 13 buffer solution at a scan rate of 10 mV‚s-1, (b and c) first and 20th recorded cyclic voltammograms of the modified electrode in solution containing 1.0 µM insulin.
Figure 5. (A) Amperometric response at rotating modified Ni-doped CCE (rotation speed 2000 rpm) held at 0.40 V constant potential in pH 13 buffer solution for successive additions of 100 pM insulin. (B) Plot of chronoamperometric current vs insulin concentration. (C) Stability of the amperometric response to 160 nM insulin during 50 min.
oxide and insulin. Based on the results, the following catalytic scheme (EC′ catalytic mechanism) describes the reaction sequence in the oxidation of insulin by nickel hydroxide film.
Ni(OH)2 + OH- f NiO(OH) + H2O + eNiO(OH) + insulin (reduced form) f Ni(OH)2 + insulin (oxidized form) Amperometric Detection of Insulin at Ni Powder-Doped Modified CCE. Since the cyclic voltammetry is not particularly 7436
Analytical Chemistry, Vol. 79, No. 19, October 1, 2007
sensitive for low concentrations, amperometry under stirred conditions or flow injection analysis with amperometric detection is employed. Therefore, amperometry with a rotated modified CCE was used to detect the lower concentrations of insulin. According to the potential dependence of the insulin electrocatalytic oxidation current under steady- state conditions, the optimum electrode potential was selected at 0.45 V versus the Ag/AgCl reference electrode in order to obtain constant and high sensitivity. Figure 5A shows a typical hydrodynamic amperometric response obtained by successfully adding insulin to a continuously stirred modified electrode (rotation speed 1500 rpm) in 0.1 M NaOH solution. As
Table 1. Analytical Parameters for Detection Insulin at Several Modified Electrodes electrode (RuOx) on carbon fiber micro electrode GCb electrode modified with Ru containing inorganic polymer film carbon fiber microelectrode modified with a polynuclear RuO/RuCN film carbon electrodes modified with RuRDMsc CCEd modified with [Ru(bpy)(tpy)Cl]PF6 GC modified IrOx GC electrode modified with CNTe GC electrode modified with chitosan and MWCNT CPf microelectrode modified with RuOx CCE modified with K4[Mo(CN)8]+Ni RuOx-CNTs modified carbon electrode CCE modified with Ni
limit of detection
sensitivity
23 nM -
0.072 nA/µM 0.42 ( 0.01 nA/ng
12 9
amperometry
100-1000 nM 8.2-81.6 ng in 7.5 µL -
500 nM
0.441 nA/µM
11
FIA amperometry amperometry FIA amperometry FIA amperometry
6-400 nM 0.5-850 nM 50-500 nM 100-1000 nM 100-3000 nM 100-1000 nM 0.5-500 nM
225 nA/µM 7600 nA/µM 35.2 nA/µM 48 nA/µM 135 nA/µM 0.875 6140 nA/µM
16 23 17 18 19 14 24
FIA FIA amperometry
100-500 pM 10-80 nM 0.1-700 nM
2 nM 0.4 nM 20 nM 14 nM 30 nM 50 nM 0.45 nM ( 0.05 40 pM 1 nM 40 pM
FIA
15-100 pM
2.6 pM
271.9 pA/pM
method FIAa FIA
dynamic range
8.1 nA/nM 541 nA/µM 29.8 pA/pM
ref
20 this work
a Flow injection analysis. b Glassy carbon. c:Ruthenium metallodendrimer multilayers. dCarbon ceramic electrode. e Carbon nanotubes. f Carbon paste.
Figure 6. Flow injection analysis of insulin, 15, 30, 45, and 90 pM, operation potential 0.4 V, Insets, display the resulting calibration plot.
shown in Figure 5A, during the successive additions of 100 pM insulin, a well-defined response was observed. The measured currents increased, as the insulin concentrations in solution were increased. The calibration plot for insulin determination was linear for a wide range 100 pM to 700 nM. A linear least-squares calibration curve over the range of 100- 800 pM (8 determinations) shows a slope of 945 pA pM-1 cm-2 M (sensitivity) and a correlation coefficient of 0.9918 (Figure 5B). The detection limit was 40 pM when the signal-to-noise ratio was 3. The response time measured in amperometric determination is 1 s or less to 90% of the full signal. The detection limit, linear calibration range, and sensitivity for insulin determination at this modified electrode are comparable and even better than those obtained by using several modified electrodes (Table 1). The Ni powder-modified CCE imparts higher stability onto amperometric measurments of insulin. Figure 5C shows the amperometric response of 160 nM insulin during a prolonged 50.0-min experiment. The response
remains stable throughout the experiment (only 5% decrease in current was observed), indicating no inhibition effect of insulin and its oxidation products for modified electrode surface. Therefore, the modified carbon ceramic electrode has excellent, stable, and strong mediation properties and facilitates the low potential amperometric measurement of insulin. Flow injection analysis of insulin is used for assessing the temporal response and overall analytical performance of the sensor. The subsequent flow injection experiments were performed using the pH 13 phosphate buffer as a carrier solution and 0.45 V as a sensing potential for insulin. Figure 6 compares the amperometric response of the modified CCE to successive additions of different concentrations of insulin. As expected from the hydrodynamic-voltammetric data, the CCE-based detector exhibits a rapid increase and decrease of the current. As indicated from the resulting calibration plot (inset), the response of the Analytical Chemistry, Vol. 79, No. 19, October 1, 2007
7437
modified electrode is highly linear over the entire 15-90 pM concentration range with a slope of 8659.23 pA pM-1 cm-2 and a correlation coefficient of 0.9966. The favorable signal-to-noise characteristics of these data indicate a detection limit of 2 pM (based on S/N ) 3). Table 1 compares the analytical parameters of the Ni powder-doped CCE for insulin determination to those for other modified electrodes. To our best knowledge, these analytical parameters are the best reported values for insulin determination, with a simple, low-cost, and reproducible modified electrode without using any specific reagent (Table 1). CONCLUSIONS For the first time, the electrooxidation of picomolar insulin solutions was successfully performed using the renewable carbon composite electrode modified with nickel powder. The modified electrodes were prepared using a simple preparation technique without using any specific reagent or electron-transfer mediator. The modified CCE offers a significant decrease in overpotential for insulin oxidation and minimizes surface fouling effects of insulin and their oxidation products. Remarkable electrocatalytic activity and stable response of the modified electrode are comparable with all other modified electrodes employed as an insulin sensor. In addition to excellent electrochemical perfor(44) Salimi, A.; Roushani, M. Electrochem. Commun. 2006, 8, 205-214.
7438
Analytical Chemistry, Vol. 79, No. 19, October 1, 2007
mance of this electrode, the modification procedure is much less expensive and more convenient than those used for other insulin sensors. The analytical performance of the Ni powder-based carbon composite electrodes indicates that it can be used as a sensitive amperometric detector for picomolar detection of insulin when coupled with a flow system. Since the modified electrode showed excellent electrocatalytic activity for glucose oxidation at the same condition,44 this sensor can be used as amperometric detector for detection of two important biological compounds in chromatographic effluents and for improving the management of diabetes. ACKNOWLEDGMENT This work was financially supported by Research Office of University of Kurdistan and Ministry of Science, Research and Technology of Iran. The authors thank Dr. Behzad Hagighi and Dr. Ismaiel Shams (Institute for Advanced Studied in Basic Science, Zanjan-Iran) for flow injection amperometric measurements and Mr. Taher Sarhadi and Dr. Neil Rees (Oxford University) for their valuable discussion. Received for review February 12, 2007. Accepted July 14, 2007. AC0702948
