Direct Voltammetric Detection of DNA and pH Sensing on Epitaxial

Aug 16, 2010 - Chem. 82, 17, 7387-7393 ... In fact, an anodized EG voltammetric sensor can realize the simultaneous ... ACS Applied Materials & Interf...
0 downloads 0 Views 3MB Size
Download PDF
Anal. Chem. 2010, 82, 7387–7393

Direct Voltammetric Detection of DNA and pH Sensing on Epitaxial Graphene: An Insight into the Role of Oxygenated Defects Cheng Xiang Lim, Hui Ying Hoh, Priscilla Kailian Ang, and Kian Ping Loh* Department of Chemistry, National University of Singapore, 3 Science Drive 3, Singapore 117543 In this paper, we carried out detailed electrochemical studies of epitaxial graphene (EG) using inner-sphere and outer-sphere redox mediators. The EG sample was anodized systematically to investigate the effect of edge plane defects on the heterogeneous charge transfer kinetics and capacitive noise. We found that anodized EG, consisting of oxygen-related defects, is a superior biosensing platform for the detection of nucleic acids, uric acids (UA), dopamine (DA), and ascorbic acids (AA). Mixtures of nucleic acids (A, T, C, G) or biomolecules (AA, UA, DA) can be resolved as individual peaks using differential pulse voltammetry. In fact, an anodized EG voltammetric sensor can realize the simultaneous detection of all four DNA bases in double stranded DNA (dsDNA) without a prehydrolysis step, and it can also differentiate single stranded DNA from dsDNA. Our results show that graphene with high edge plane defects, as opposed to pristine graphene, is the choice platform in high resolution electrochemical sensing. Graphene has attracted intense research interests because of its massless fermions, ultrahigh charge mobility, and promising performance in spin valves and tetrahertz transistors.1,2 Most of the applications rely on lateral conduction of carriers in high quality two-dimensional sheets. Due to the in plane conjugation of the sp2 bonds in graphene, two-dimensional electrical conduction is highly efficient. As a result, heterogeneous electron transfer with redox species can take place on the edges of graphene sheets. In contrast, heterogeneous electron transfer out of the plane of the graphene sheet is expected to be relatively sluggish. The graphene basal plane is commonly associated with low reactivity, which is attributed to its low density of electronic states and the lack of functional groups and adsorption sites. Most of the recent electrochemical studies involving graphene were carried out using chemically processed, multilayer, reduced graphene oxide (GO) flakes which consist of a high density of reactive edges.3-5 However, there has been no electrochemical study of high quality crystalline * To whom correspondence should be addressed. E-mail: [email protected]. (1) Geim, A. Science 2009, 324, 1530–1534. (2) Lin, Y. M.; Dimitrakopoulos, C; Jenkins, K. A.; Farmer, D. B.; Chiu, H. Y.; Grill, A.; Avouris, Ph. Science 2010, 327, 662. (3) Shang, N. G.; Papakonstantinou, P.; McMullan, M.; Chu, M.; Stamboulis, A.; Potenza, A.; Dhesi, S. S.; Marchetto, H. Adv. Funct. Mater. 2008, 18, 3506–3514. (4) Zhou, M.; Zhai, Y.; Dong, S. Anal. Chem. 2009, 81, 5603–5613. 10.1021/ac101519v  2010 American Chemical Society Published on Web 08/16/2010

graphene, so the intrinsic electrochemical characteristic of real graphene is not known. Such crystalline graphene is markedly different in chemical composition and structure from reduced graphene oxide flakes. To this end, we performed a detailed electrochemical study of crystalline epitaxial graphene (EG) prepared on silicon carbide. To investigate the effect of edge plane defects on the electrochemical and biosensing activities of EG, a systematic study of the heterogeneous charge transfer rate as a function of defect density on EG was carried out. It is found that the electrochemistry of EG converges with that of reduced GO flakes following its anodization treatment, and high sensitivity resolution of nucleic acids is possible. EXPERIMENTAL SECTION Chemicals and Materials. Silicon carbide substrates used for EG growth were cut from nitrogen-doped, on-axis oriented, double side polished, research grade, (0001) face (Si-terminated) 6H-SiC wafers, purchased from Cree. Potassium ferricyanide (K3Fe[CN]6), potassium ferrocyanide (K4Fe[CN]6), hexaammine ruthenium(III) chloride (Ru(NH3)6Cl3), hexaammine ruthenium(II) chloride (Ru(NH3)6Cl2), potassium chloride (KCl), guanine (G), adenine (A), thymine (T), cytosine (C), and double stranded DNA (dsDNA) from salmon testes were purchased from Sigma Aldrich. Single stranded DNA (ssDNA) of the sequence 5′-CAT-GAA-CCG-3′ and phosphate buffer solutions were obtained from 1st Base. Ascorbic acid (AA), dopamine (DA), and uric acid (UA) were purchased from Alfa Aesar. Preparation of EG. The SiC samples were diced into 10 mm by 5 mm crystals. After degreasing with ethanol, the sample was introduced into an annealing station housed in an ultrahigh vacuum chamber equipped with reflection high energy electron diffraction (RHEED). The SiC substrate was dosed with silicon from a silicon evaporation source for 3 min to enrich the surface with silicon; this was followed by a series of annealing steps (1 min) between 900 and 1300 °C. Flash annealing was performed until a (6(3)1/2 × 6(3)1/2)-R30° reconstruction pattern appeared as monitored by RHEED. The sample was removed from the UHV chamber and transferred in air to a scanning tunneling microscopy (STM) chamber for further annealing and analysis. When a long-range moire´ pattern characteristic of honeycomb graphene was observed, it can be concluded that the epitaxial (5) Tang, L.; Wang, Y.; Li, Y.; Feng, H.; Lu, J.; Li, J. Adv. Funct. Mater. 2009, 19, 1–8.

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

7387

graphene had been successfully prepared on the surface and the sample was removed for electrochemistry studies. Raman Spectroscopy. The WITEC CRM200 Raman system was used for characterization of samples. The excitation source was a 532 nm laser (2.33 eV). To prevent laser-induced local heating, it was necessary to keep the laser power on the sample below 0.1 mW. A 100 objective lens with a numerical aperture of 0.95 was used. The spot size of the 532 nm laser was estimated to be 500 nm. This system provides a spectral resolution of 1 cm-1. Sample Pretreatment-Graphene Anodization. Anodized EG was prepared by electrochemical pretreatment of as-synthesized samples using the Autolab PGSTAT30 digital potentiostat/ galvanostat with GPES 4.9 software (Eco Chemie, The Netherlands) and a lock-in amplifier (PAR EG&G model 273A). The electrochemical cell was assembled in a Teflon housing, with a conventional three-electrode system: EG working electrode, Ag|AgCl/KCl (sat.) reference electrode, and a Pt wire counter electrode. In the Teflon housing, the working electrode was well clamped at the bottom with a 2 mm diameter O ring. The exposed surface area of the graphene electrode was 0.0314 cm2. Graphene anodization was carried out by applying a potential of 2.0 V versus Ag/AgCl for 500 s in pH 7 phosphate buffer solution. Electrochemical Measurements. Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) were performed using the same three-electrode setup. Besides EG, other types of working electrodes used for comparative studies were the glassy carbon (GC) and boron-doped diamond (BDD) electrodes, which have exposed surface areas of 0.0707 cm2 and 0.0314 cm2, respectively. Measurements involving the GC electrode were conducted in a beaker while measurements on BDD were conducted in the Teflon housing. CV measurements of redox couples and background scans were repeated 3 times, and the third scans were recorded. Heterogeneous electron transfer rate constants were calculated from the anodic/cathodic peak separation using the method of Nicholson6 by assuming R ) 0.5 and using the following diffusion coefficients:7,8 Fe(CN)63-/4-, D0 ) 7.63 × 10-6 cm2s-1, DR ) 6.32 × 10-6 cm2s-1; Ru(NH3)62+/3+, D0 ) DR ) 6.5 × 10-6 cm2s-1. DPVs for biosensing were carried out with the bioanalytes in 10 mM KCl/10 mM phosphate buffer solution at pH 7. DPV measurements were conducted on the basis of the following parameters: 60 s accumulation time at 0.2 V, 50 ms modulation time, 0.5 s interval time, 25 mV modulation amplitude, and 5 mV step. Raw voltammograms were treated with the Savitzky and Golay filter and moving average baseline corrected with a peak width of 0.005. Mott-Shottky plots were recorded using the Autolab PGSTAT30 digital potentiostat/galvanostat with FRA2 module with FRA 4.9 software (Eco Chemie, The Netherlands) with a frequency of 1000 Hz. pH 2 and pH 10 solutions were prepared by adding aliquots of 0.5 M HCl or KOH to deionized water. Before all measurements, the electrolyte solutions were purged with nitrogen gas for 30 min. Care was taken to ensure no bubble formation on the electrodes. The solution was allowed to come to rest before the commencement of each scan. (6) Nicholson, R. S. Anal. Chem. 1965, 37, 1351–1355. (7) Gerhardt, G.; Adams, R. N. Anal. Chem. 1982, 54, 2618–2620. (8) Kovach, P. M.; Deakin, M. R.; Wightman, R. M. J. Phys. Chem. 1986, 90, 4612–4617.

7388

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

RESULTS AND DISCUSSION Effect of Anodization on Defect Density/Surface Roughness. The EG samples were anodized in buffer solution for different lengths of time to introduce oxygenated groups on the surface of EG. Raman spectroscopy was applied to characterize the defect density on the graphene surface as a function of anodization time. The Raman spectrum of the pristine EG sheets as shown in Figure 1A is characterized by a G band assignable to the E2g phonon of sp2 carbon atoms and a small D band due to first-order scattering from a zone-boundary phonon. This zone boundary phonon which is symmetry forbidden in perfect crystal is activated by disorder associated with finite crystallite size.9,10 Following increasing time of anodization, the ratio of D:G band can be seen to increase in the order of pristine EG < anodized EG (200 s) < anodized EG (500 s). The changes in the relative ratio of the D and G band intensity (D/G peak ratio) reflect the changes in the defect density.4,10 The effect of anodization results in oxygenation of the graphene, as judged by the rise of the intensity of the O1s peak in the X-ray photoelectron spectroscopy (XPS) spectra from Figure 1B. This increase is attributed to an increase in the population of CdO and C-O-H functional groups (Figure 1C,D). Oxidation of graphene creates strain on the sp2 bonded lattice because of the transition to sp3-bonding with oxygen species. Prolonged anodization will result in fracturing of the lattice and formation of edge plane defects11 decorated by hydroxyl as well as carboxylic groups. An increase in D peak intensity as a result of anodization was similarly reported by Prasad et al. for screen-printed carbon electrodes.12 AFM characterization of the graphene surface (Figure 2) shows an increase in surface corrugation of 0.56 to 4.52 nm following anodization. Electrochemical Characterization. Ru(NH3)62+/3+ and Fe(CN)63-/4- redox couples are selected to examine the heterogeneous charge transfer mechanism of EG. Due to the different charge transfer mechanisms, these redox couples exhibit varying degrees of sensitivity to electronic properties and surface microstructure, which allow us to probe the main parameters affecting charge transfer on graphene.13 Electron transfer in the Fe(CN)63-/4- redox couple follows the inner-sphere mechanism. Electron transfer kinetics is, thus, dependent on not only the density of electronic states but also surface microstructure.14,15 Previous reports indicated that increases in electron transfer rate constant on the highly oriented pyrolytic graphite (HOPG) basal plane correlate with the appearance of edge plane defects.16 In view of this, it is proposed that electron transfer kinetics can be an effective indicator of the density of edge plane defects on EG. Therefore, CV was carried out on the anodized samples to explore the relationship between defect density and capacitive background current. (9) Ferrari, A. C.; Meyer, J. C.; Scardaci, V; Casiraghi, C.; Lazzeri, M.; Mauri, F; Piscanec, S.; Jiang, D; Novoselov, K. S.; Roth, S.; Geim, A. K. Phys. Rev. Lett. 2006, 97, 187401(1)187401(4) . (10) Ferrari, A. C.; Robertson, J. Phys. Rev. B 2000, 61, 14095–14107. (11) Bowling, R.; Packard, R. T.; McCreery, R. L. Langmuir 1989, 5, 683–688. (12) Prasad, K. S.; Muthuraman, G.; Zen, J.-M. Electrochem. Commun. 2008, 10, 559–563. (13) Paciosa, M.; del Vallea, M.; Bartrolia, J.; Esplandiu, M. J. J. Electroanal. Chem. 2008, 117, 619–620. (14) Fischer, A. E.; Show, Y.; Swain, G. M. Anal. Chem. 2004, 76, 2553–2560. (15) McCreery, R. L. Chem. Rev. 2008, 108, 2646–2687. (16) Rice, R. J.; McCreery, R. L. Anal. Chem. 1989, 61, 1637–1641.

Figure 1. (A) Raman spectra of pristine EG, anodized EG (200 s), and anodized EG (500 s). (B) XPS spectra of anodized EG and pristine EG. (C) XPS C1s spectra of pristine EG. (D) XPS C1s spectra of anodized EG.

Figure 2. AFM topographical images and cross sections for (A) pristine EG and (B) anodized EG surfaces. Edge defects are generated on the anodized EG surface, leading to high electrochemical activity.

Figure 3A shows that there is indeed a direct correlation between the electron transfer kinetics of the Fe(CN)63-/4- couple and the defect density on EG. As judged from the peak separation, heterogeneous electron transfer rate constants follow the order of anodized EG (500 s) > anodized EG (200 s) > pristine EG, as summarized in Table 1. Existing defects (such as kinks, steps, vacancies) on the edge planes of epitaxial graphene can produce localized edge states, resulting in high density of electronic states near the Fermi level. This leads to increased electrochemical reactivity and explains the observation that the sample with the highest defect density displays fastest electron transfer kinetics. In addition, it is noted that samples with

higher defect density record higher background capacitive currents associated with a nonfaradaic process. The presence of ionizable hydroxyl and carboxylic functionalities contribute to an increase in the capacitive charging current. The double layer capacitances at 0.25 V for pristine EG, anodized EG (200 s), and anodized EG (500 s) are calculated to be 0.744, 1.12, and 5.55 µF cm-2, respectively, as shown in Table 1. Using charge transfer rate as well as capacitive noise as parameters, we can see that the Fe(CN)63-/4- redox system is a reliable indicator of defect density in epitaxial graphene. The as-prepared graphene exhibits low capacitive noise but suffers from a sluggish charge transfer rate. The Ru(NH3)62+/3+ couple follows an outer-sphere mechanism. In this case, the electronic properties of the electrode, in particular the density of the electronic states near the Fermi potential, are the most important factor affecting the charge transfer rate.14,15 This system, thus, assesses whether the density of electronic states can support the electron transfer with Ru(NH3)62+/3+. The electrochemical performance of the anodized EG electrode was compared with that of GC and BDD electrodes; the latter two electrodes are commonly used in analytical chemistry. It is clear from Figure 4 that the charge transfer rate for anodized EG involving both the Ru(NH3)62+/3+ and Fe(CN)63-/4- redox systems are higher than that of GC and BDD electrodes (Table 2). The small peak separation in the CV for both redox couples is indicative of fast electron transfer rate in anodized EG. It is also apparent from Figure 5 that the cathodic and anodic currents of the anodized EG electrode shows a linear relationship with the square root of the scan rate at scan rates ranging from 25 to 150 mV s-1. This agrees with the Randles Sevcik equation Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

7389

Figure 3. CVs for pristine EG, anodized EG (200 s), and anodized EG (500 s): (A) 1.0 mM Fe(CN)63-/4- in 1 M KCl; (B) 1 M KCl. Scan rate: 100 mV s-1. Table 1. Apparent Rate Constants and Double Layer Capacitances for Pristine EG, Anodized EG (200 s), and Anodized EG (500 s) sample

defect density

k°app (cm s-1) for Fe(CN)63-/4-

capacitance at 0.25 V (µF cm-2)

pristine EG anodized EG (200s) anodized EG (500 s)

lowest moderate highest

no peaks slow 0.0101

0.744 1.12 5.55

Figure 4. CVs for two kinds of redox systems at anodized EG, GC, and BDD electrodes: (A) 1.0 mM Ru(NH3)62+/3+ in 1 M KCl; (B) 1.0 mM Fe(CN)63-/4- in 1 M KCl. Scan rate: 100 mV s-1. Table 2. CV Data and Apparent Heterogeneous Electron Transfer Rate Constants for Two Redox Systems at Anodized EG, GC, and BDD Electrodes parameter ∆Ep (mV) k°app (cm s-1)

electrode

Ru(NH3)62+/3+

Fe(CN)63-/4-

anodized EG GC BDD anodized EG GC BDD

81 88 101 0.00981 0.00740 0.00446

81 151 282 0.0101 0.00221 0.000922

which describes reversible electrochemical reactions controlled by semi-infinite linear diffusion. The electrochemical potential window of the electrode is an important parameter that defines the electrical energy range of anodic or cathodic reactions which can be carried out before the electrolysis of the electrolyte. For example, the anodic oxidation of nucleic acids is known to occur at high voltages that exceed the electrochemical potential window of conventional electrodes like gold and GC. Figure 6 compares the electrochemical potential window of anodized graphene, GC, and BDD electrodes with 7390

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

regard to the electrochemical potential window. From Figure 6, it can be seen that the electrochemical potential window of anodized EG is much wider than that of the GC and is comparable to BDD electrode which is widely recognized for its wide electrochemical potential window.17,18 This is a very useful attribute for the development of voltammetric sensors for DNA, where the oxidation is known to occur at high anodic potentials. The electrochemical data comparing the charge transfer kinetics of GC, BDD and anodized EG are summarized in Table 2. DNA Biosensing. To investigate the DNA biosensing ability of anodized EG, we carried out DPV on these analytes: (a) free DNA bases; (b) dsDNA from salmon testes and ssDNA with sequence 5′-CAT-GAA-CCG-3′. Direct voltammetric sensing of DNA bases is difficult on untreated graphite electrodes and usually necessitates modification with catalyst or metalloproteins in order to enhance detection. (17) Jia, J.; Kato, D.; Kurita, R.; Sato, Y.; Maruyama, K.; Suzuki, K.; Hirono, S.; Endo, T.; Niwa, O. Anal. Chem. 2007, 79, 98–105. (18) Prado, C.; Flechsig, G.-U.; Gru ¨ ndler, P.; Foord, J. S.; Marken, F.; Compton, R. G. Analyst 2002, 127, 329–332.

Figure 5. CVs recorded at an anodized EG electrode in (A1) 1.0 mM Ru(NH3)62+/3+ in 1 M KCl and (A2) 1.0 mM Fe(CN)63-/4- in 1 M KCl at various scan rates and corresponding peak current dependence on the square root of scan rate for (B1) Ru(NH3)62+/3+ and (B2) Fe(CN)63-/4-.

Figure 6. Electrochemical window of anodized EG, GC, and BDD electrodes in 1 M KCl at 100 mVs-1.

In this study, we evaluate the ability of the electrodes for the direct voltammetric sensing of nucleic bases. Figure 7A shows that, out of the four electrodes used, anodized EG possesses the best capability to achieve simultaneous detection of all DNA bases with sufficient resolution. In fact, pristine EG performed better than the GC and BDD electrodes in terms of detecting clearly the cytosine base in the (G + A + T + C) mixture. The ability of EG to give a clear oxidation peak can also be seen by performing the CV of single thymine base in Figure 7A. GC was unable to show a distinct oxidation peak.19,20 As for BDD, a flat CV was obtained, though Ivandini et al.21 did report the acquisition of a small T peak. (19) Brett, A. M. O.; Piedade, J. A. P.; Silva, L. A.; Diculescu, V. C. Anal. Biochem. 2004, 332, 321–329. (20) Niwa, O.; Jia, J. B.; Sato, Y.; Kato, D.; Kurita, R.; Maruyama, K.; Suzuki, K.; Hirono, S. J. Am. Chem. Soc. 2006, 128, 7144–7145. (21) Ivandini, T. A.; Honda, K.; Rao, T. N.; Fujishima, A.; Einaga, Y. Talanta 2007, 71, 648–655.

Observing distinct oxidation peaks corresponding to individual bases in dsDNA is much more difficult than the oxidation of free DNA bases. There have been reports of detection of bases in DNA with the help of a prehydrolysis step to release the bases into their free states.22,23 In Figure 7C, we demonstrate the ability of anodized EG to simultaneously detect all four DNA bases in dsDNA without a prehydrolysis step. This remarkable performance is unique to anodized graphene. The detection limit for dsDNA is 1 µg mL-1. In addition, as shown in Figure 7D, anodized EG is able to differentiate between dsDNA and ssDNA from the relatively higher oxidation peaks for A and C and lower energy shift of the C oxidation peak in the latter. Such a differentiation has not been reported before for voltammetric sensing. Pristine graphene, GC, and BDD as well as most other electrodes suffer from problems such as narrow electrochemical potential window (except BDD and graphene), slow electron transfer kinetics, and/or high background current,20,24,25 which preclude distinct detection of individual bases in intact DNA by voltammetric sensing. Selective Detection of DA in the Presence of AA and UA. To further highlight the exceptional electrochemical activity of anodized EG, we carried out electrocatalytic oxidation of a mixture containing DA, AA, and UA, which coexist in extracellular fluids in the human body and which are generally difficult to resolve by voltammetric sensing on common electrodes. It is clear from Figure 8A,C,D that pristine EG, GC, and BDD lack the ability to selectively determine DA in such a mixture, as (22) Zhang, R.; Wang, X.; Chen, C. Electroanalysis 2007, 15, 1623–1627. (23) Abdullin, T. I.; Nikitina, I. I.; Ishmukhametova, D. G.; Budnikov, G. K.; Konovalova, O. A.; Salakhov, M. Kh. J. Anal. Chem. 2007, 62, 599–603. (24) Kato, D.; Sekioka, N.; Ueda, A.; Kurita, R.; Hirono, S.; Suzuki, K.; Niwa, O. Angew. Chem., Int. Ed. 2008, 47, 6681–6684. (25) Kato, D.; Sekioka, N.; Ueda, A.; Kurita, R.; Hirono, S.; Suzuki, K.; Niwa, O. J. Am. Chem. Soc. 2008, 130, 3716–3717.

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

7391

Figure 7. DPV profiles for pristine EG, anodized EG, GC, and BDD electrodes in (A) 30 µM equimolar mixture of G, A, T, and C, (B) 1.0 mM thymine, (C) 30 µg mL-1 dsDNA, (D) anodized EG in 30 µg mL-1 dsDNA and 30 µg mL-1 ssDNA. Supporting electrolyte: 10 mM KCl/10 mM PBS solution at pH 7.

Figure 8. DPV profiles of the (A) pristine EG, (B) anodized EG, (C) GC, and (D) BDD electrodes in a mixture of 0.5 mM AA, 0.1 mM DA, and 0.1 mM UA and the individual biomolecules. Supporting electrolyte: 10 mM KCl/10 mM PBS solution at pH 7.

the oxidation of all three species gave rise to only two peaks. This is due to the overlapping voltammetric responses to AA and DA. On the other hand, as shown in Figure 8B, the pulsed voltammogram of anodized EG exhibits three sharp voltammertic peaks which successfully discriminate AA, DA, and UA. The marked 7392

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

improvement in resolution is due to the negative shift of the oxidation potential of AA by 0.12 V, relative to that of pristine EG. Another reason is that the oxidation signal of DA is much stronger than AA and UA. In neutral condition (pH 7), DA (pKa ) 8.87)26 exists as cations while AA (pKa ) 4.17)27 and UA (pKa ) 3.70)28

Figure 9. Mott-Schottky plot of (A) pristine EG and (B) anodized EG in pH 2 and pH 10 solutions.

exist as anions. The anodized surface has a high density of negatively charged edge planes which repel the anionic AA and UA while showing enhanced interaction with cationic DA, thus resulting in increased sensitivity for DA. It must be stated that the selective detection of DA in the presence of interfering agents cannot be achieved using bare electrodes. As a result, a significant portion of the electrochemical sensing literature concerns itself with different modification strategies to enhance the selectivity toward these biomolecules.29-32 However, such modified electrodes usually require multiple timeconsuming and complicated surface modification steps. A platform based on anodized EG, therefore, offers a much simpler and more efficient alternative for the selective detection of such biomolecules. EG as a pH Sensor. The pH sensitivity of flatband potential in pristine EG and anodized EG was investigated with the Mott-Shottky (Cp-2, V) plot. A pH dependence of the flat-band potential points to a protonation/deprotonation of surface hydroxyl groups or adsorption/desorption of hydroxyl ions, which causes a change in the Helmholtz potential drop (∆φH). From Figure 9, since it is observed that the 1/C2 value increases with anodic polarizations, both anodized and pristine EG can be considered n-type semiconductors. Similarly, Varchon et al. reported that graphene sheets formed on SiC substrates were n-typed doped due to charge transfer from the substrate.33 Flatband potentials can be obtained from the Mott-Shottky plot by extrapolation to 1/C2 ) 0. From Figure 9A,B, it is observed that the flatband potential for pristine EG is relatively pH independent since it changed by a mere 0.1 V when the pH was varied from 2 to 10. In contrast, the flatband potential of (26) Liu, A.; Wei, M.; Honma, I.; Zhou, H. Adv. Funct. Mater. 2006, 16, 371– 376. (27) Giz, M. J.; Duong, B.; Tao, N. J. J. Electroanal. Chem. 1997, 430, 205–214. (28) Richards, J.; Weinman, E. J. J. Nephrol. 1996, 9, 160–166. (29) Safavi, A.; Maleki, N.; Moradlou, O.; Tajabadi, F. Anal. Biochem. 2006, 359, 224–229. (30) Zare, H. R.; Rajabzadeh, N.; Nasirrizadeh, N.; Mazloum Ardakani, M. J. Electroanal. Chem. 2006, 589, 60–69. (31) Wang, H. S.; Li, T. H.; Jia, W. L.; Xu, H. Y. Biosens. Bioelectron. 2006, 22, 664–669. (32) Wang, P.; Li, Y. X.; Huang, X.; Wang, L. Talanta 2007, 73, 431–437. (33) Varchon, F.; Feng, R.; Hass, J.; Li, X.; Ngoc Nguyen, B.; Naud, C.; Mallet, P.; Veuillen, J.-Y.; Berger, C.; Conrad, E. H.; Magaud, L Phys. Rev. Lett. 2007, 99, 126805(1)126805(4) . (34) Rao, T. N.; Tryk, D. A.; Hashimoto, K.; Fujishima, A. J. Electrochem. Soc. 1999, 146, 680–684. (35) Shirafuji, J.; Sugino, T. Diamond Relat. Mater. 1996, 5, 706–713.

anodized EG displayed a much larger change of 0.41 V. These values correspond to a shift of 12.5 mV/pH and 51.3 mV/pH for pristine and anodized EG, respectively. The value for the anodized sample is close to the Nernst’s limit of 59.2 mV/pH. Such Nernstian-like behavior is typical of oxide semiconductors with surface hydroxyl groups that can undergo protonation and deprotonation with pH.34 The reason for the higher pH sensitivity of anodized EG could be attributed to the presence of ionizable carboxylic and hydroxyl groups, which have been verified by XPS measurements. Such groups create an acid-base equilibrium which subsequently changes the Helmholtz layer charge and magnitude of band bending, which translates to changes in the flatband potential. Such an observation is similar to the pH sensitivity of O-terminated diamond.35 Pristine EG does not exhibit such pH sensitivity due to the absence of ionizable oxygen functionalities. CONCLUSIONS We found that anodized epitaxial graphene (EG) is a more superior electroanalytical platform for a wide range of biomolecules compared to carbon electrodes like boron-doped diamond, carbon nanotubes, and glassy carbon. The anodization of graphene is a very simple process that can be applied as a pretreatment step in any electroanalytical analysis. As a voltammetric sensor, anodized EG is able to resolve the anodic peaks of all four nucleic acid bases in double stranded and single stranded nucleic acids, a performance unmatched by other electrodes. It is also able to differentiate dopamine, ascorbic acid, and uric acid at physiological pH, a test which most electrodes will fail without pretreatment of the electrode surfaces. All these enhanced electroanalytical performances are attributed to the presence of edge plane defects. The flatband voltages of pristine graphene do not respond as sensitively to pH changes as anodized EG. The Nernstian-like responses of anodized EG to pH changes can be explained by the presence of ionizable oxygen groups on its surface. ACKNOWLEDGMENT K. P. Loh thanks the funding support of NRF-CRP grant “Graphene Related Materials and Devices (R-143-000-360-281).”

Received for review June 8, 2010. Accepted August 2, 2010. AC101519V

Analytical Chemistry, Vol. 82, No. 17, September 1, 2010

7393