Analyst View Article Online
PAPER
View Journal
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Detection of biomarkers with graphene nanoplatelets and nanoribbons† Cite this: DOI: 10.1039/c3an01585h
Chee Shan Lim, Chun Kiang Chua and Martin Pumera* Well-defined graphene nanosheets have become increasingly popular in the electrochemical detection and quantification of small molecules. In this work, the electrochemical oxidation of biomarkers such as uric acid, ascorbic acid, dopamine, NADH and DNA bases, namely guanine and adenine, was performed using cyclic voltammetry and differential pulse voltammetry to compare the electrochemical properties of electrochemically reduced nanoplatelets (ENPs) and electrochemically reduced nanoribbons (ENRs). The graphene materials displayed better electrochemical performances than the bare glassy carbon surface. Between the two graphene materials, the oxidation of biomarkers occurred at lower oxidation potentials on the ENP surface. The sensitivities of the two graphene surfaces varied when different biomarkers were studied. The ENP surface showed enhanced sensitivities for ascorbic acid, while the ENR surface exhibited higher sensitivities for uric acid and dopamine. As for the DNA bases analysed, both guanine
Received 21st August 2013 Accepted 18th November 2013
and adenine were oxidised at lower potentials on the ENP surface than the ENR surface. The ENP DOI: 10.1039/c3an01585h
surface displayed a better sensitivity for guanine, whereas the oxidation of adenine was more sensitive
www.rsc.org/analyst
on the ENR surface.
1. Introduction A healthy human body has the ability to maintain a steady internal environment, which is also known as homeostasis. While the organ systems function to achieve homeostasis via the regulation of various positive and negative feedback mechanisms, an inability to maintain a constant internal environment can result in diseases or even death. Thus, the human body has to depend on a series of internal feedback mechanisms to achieve such homeostatic balance. Existing scientic development aims to assist in the detection of abnormalities in the human body by monitoring several signicant biomarkers. The detection of biomarkers such as uric acid, ascorbic acid, dopamine, reduced nicotinamide adenine dinucleotide (NADH) and deoxyribonucleic acid (DNA) is able to provide useful information for the prevention of genetic diseases and other health ailments, as well as their medical treatments subsequently. The electrochemical study of biomarkers and DNA bases, which rely mainly on their redox properties,1 has proven to be of signicant importance in the advancement of biosensing for its low cost, simplicity and quick response.2 Following the introduction of electrochemical detection of nucleic acids with nucleic acid-modied electrodes by Paleˇcek, various electrode
Division of Chemistry & Biological Chemistry, School of Physical and Mathematical Sciences, Nanyang Technological University, Singapore 637371, Singapore. E-mail: [email protected]; [email protected]; Fax: +65 67911961 † Electronic supplementary 10.1039/c3an01585h
information
(ESI)
available.
This journal is © The Royal Society of Chemistry 2014
See
DOI:
surfaces such as gold, mercury and carbon have been used in the development of label-free electrochemical genosensors.3–5 In recent years, studies have been carried out using nanomaterials to show improved electrochemical responses for analyses of both DNA bases and biomarkers.6,7 Graphene is a one-atom thick material containing sp2bonded carbon atoms arranged in a honeycomb structure.8 It has been known to have unique electrochemical,9 optical,10 mechanical11 and electrical12 properties since it was rst isolated in 2004. These properties, together with its low cost and little impact on the environment, have made graphene a suitable material for sensing applications.13 Fabrication of graphene materials can be achieved by a ‘top-down’ approach, via sono-chemical or thermal exfoliation of graphite,9 or a ‘bottomup’ growth using chemical vapour deposition (CVD). Of which, the former approach is most commonly used for modication of graphite to other chemically modied graphene materials. The oxidation of graphite produces graphite oxide, which can then be reduced at high temperatures to form thermally reduced graphite oxide. Graphite oxide can also undergo ultrasonication to generate graphene oxide, which can be reduced to electrochemically reduced graphene oxide and chemically reduced graphene oxide by applying a reduction potential or adding a reducing agent, respectively. Apart from the aforementioned methods, graphene can also be obtained by the lateral unzipping of stacked graphene nanobers (SGNFs) or axial unzipping of multi-walled carbon nanotubes (MWCNTs), as shown in Scheme 1.14,15 SGNFs have a perpendicular orientation of the graphene sheets to the long axis of the ber, therefore possessing a large number of Analyst
View Article Online
Analyst
Paper
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
monobasic, sodium phosphate dibasic, potassium chloride, sodium chloride, uric acid, ascorbic acid, dopamine, guanine hydrochloride, adenine hydrochloride (>99%), and reduced nicotinamide adenine dinucleotide were purchased from Sigma-Aldrich, Singapore. Stacked graphene platelet nanobers (acid washed) were obtained from Strem. Glassy carbon electrodes with a diameter of 3 mm were obtained from Autolab, The Netherlands. Milli-Q water with a resistivity of 18.2 MU cm was used throughout the experiments. 2.2 Scheme 1 Schematic representation of MWCNTs and SGNFs undergoing oxidation, resulting in axial and lateral unzipping. Subsequent ultrasonication treatment and electrochemical reduction provided graphene nanoplatelets and nanoribbons.
edge-plane sites.16 Such an arrangement gives SGNFs their unique electrochemical behaviours, as almost only the edges are exposed in SGNFs. On the other hand, MWCNTs resemble closed graphene nanobers which are rolled up perfectly, with axial orientation to the long axes of the nanobers. Lateral unzipping of stacked graphene nanobers in this work generated graphene nanoplatelet sheets of dimensions 50 50 nm, whereas axial unzipping of multi-walled carbon nanotubes gave graphene nanoribbons of dimensions 5000 300 nm.15 Studies have shown that graphene contains two types of surfaces, basaland edge-like planes upon and of which graphene structures with more exposed ends and defects (edge-like) exhibit better electrochemical properties, by providing faster rates of heterogeneous electron transfer (HET).17,18 Since the electrochemical behaviours of these graphene-based nanomaterials are ruled by the edge-plane sites,19 it is speculated that the electrochemical properties of the nanoplatelet sheets will exceed those of nanoribbon sheets. The detection methods of biomarkers and DNA bases were previously carried out using chemically modied graphenes.16,19 However, no preceding research has been done to compare the analytical performances such as sensitivity, linearity of response and the effect on the oxidation potential of electrochemically reduced graphene nanoplatelets (ENPs) and graphene nanoribbons (ENRs) sheets on biomarkers and DNA bases. Therefore, this work aims to examine the electrochemical properties, as well as the response of four key biomarkers, uric acid, ascorbic acid, dopamine and NADH; and two DNA bases, namely guanine and adenine, on the ENP and ENR surfaces. The comparative study was carried out against bare glassy carbon (GC). More importantly, simultaneous electrochemical detection was performed on a mixture of uric acid, ascorbic acid and dopamine, as it is challenging to obtain good resolution and detection sensitivity of these biomarkers using conventional electrode surfaces.
2. 2.1
Experimental Materials
Multi-walled carbon nanotubes (previously characterized in ref. 20), N,N-dimethylformamide (DMF), potassium phosphate
Analyst
Apparatus
Cyclic voltammetry and differential pulse voltammetry measurements were carried out with a mAutolab type III electrochemical analyser (Eco Chemie, The Netherlands) connected to a personal computer and controlled by General Purpose Electrochemical Systems Version 4.9 soware (Eco Chemie). The parameters used for differential pulse voltammetry are as follows: 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 potential. Raw data obtained using differential pulse voltammetry underwent a baseline correction with a peak width of 0.01 using the GPES soware. All voltammetry experiments were performed in a 5 mL electrochemical cell at room temperature using a three-electrode conguration. A platinum electrode served as an auxiliary electrode, whereas an Ag/AgCl electrode was utilised as a reference electrode. All electrochemical potentials in this report are stated vs. the Ag/AgCl reference electrode. A scanning electron microscope (JEOL JSM-7600F eld-emission, Japan) was used to obtain images under normal SEM and gentlebeam (GB) mode. The samples were attached onto an aluminum sample stub using adhesive conductive carbon tape. The XPS used for analysis is a Phoibos 100 spectrometer and a Mg X-ray radiation source (SPECS, Germany). All wide-range and high resolution scans were performed at 12.53 kV. The materials were attached onto an aluminum XPS sample holder using sticky conductive carbon tape. Special attention was paid to ensure that a homogeneous and uniform layer of the material was attached onto the tape. Raman spectroscopic measurements were carried out using a confocal micro-Raman LabRam HR instrument (Horiba Scientic) in backscattering geometry with a CCD detector. Calibration was performed at 0 cm 1 and 520 cm 1 with a silicon wafer to give a peak position resolution of less than 1 cm 1. All Raman measurements of the materials were carried out using a 514.5 nm Ar laser, an Olympus optical microscope and a 100 objective lens. 2.3
Procedure
Oxidised graphene nanoplatelets and nanoribbons were previously prepared according to the modied Hummers method.15 The materials were fully characterised by X-ray photoelectron spectroscopy, Raman spectroscopy, transmission electron microscopy and scanning electron microscopy.15 Electrochemically reduced graphene was obtained by applying a potential of 1.45 V for 900 seconds to a graphene oxide-modied GC electrode in a 50 mM phosphate buffer solution (pH 7.2). The graphene oxide-modied GC electrode exhibited strong
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Paper
Analyst
voltammetric reduction peaks between 0.8 and 1.6 V with the maximum current being displayed at 1.45 V. It is interesting to note that the electrochemical reduction potential values of some common oxygen-containing groups, peroxy, aldehyde, epoxyl and carboxyl, are around 0.7 V, 1.0 V, 1.1 V and 2.0 V, respectively. Hence, the reduction peak of 1.45 V observed is equivalent to the reduction peak of graphene oxide on various electrode surfaces with the removal of all the aforementioned oxygen-containing groups except carboxyl groups. Glassy carbon electrode surfaces were renewed by polishing with 0.05 mm alumina powder. To immobilise the carbon materials onto the working electrodes, a suspension of the desired material with a concentration of 1 mg mL 1 in DMF, followed by a 1 min sonication, was rst prepared. 1 mL aliquot of the appropriate suspension was then deposited onto the electrode surface. Evaporation of the solvent at room temperature was then allowed to take place, giving a randomly distributed lm on the glassy carbon electrode surface. The reproducibility of the measurements was obtained over 3 experiments, each time using 3 different electrode units. All voltammetry measurements were conducted with a 50 mM phosphate buffer solution (PBS) of pH 7.2.
3.
Fig. 2 Raman spectra of graphene related materials. (NR) unzipped nanoribbon graphene oxide; (ENRs) electrochemically reduced NRs; (NPs) exfoliated SGNF graphene oxide nanoplatelets; (ENPs) electrochemically reduced nanoplatelets.
Results and discussion
Prior to electrochemical and analytical studies, we characterized the materials used. Fig. 1 shows the SEM images of the materials. The morphology of the materials did not signicantly change upon electrochemical reduction. Raman spectroscopy revealed that the density of defects increased upon electrochemical reduction, the intensity of the defect (sp3) related peak at 1350 cm 1 (D-peak) vs. the intensity of the peak related to sp2 hybridization at 1560 cm 1 (G-peak) increased from 0.84 to 1.27 in the case of graphene oxides made from carbon nanotubes (D/G ratio). The D/G ratio of graphene oxide nanoplatelets increased from 0.86 to 1.05 upon electrochemical reduction (Fig. 2). X-ray photoelectron spectroscopy (XPS) was used to shed light on the efficiency of electrochemical reduction. From wide scan XPS spectra (Fig. 3), it is obvious that the C/O ratio increased from 3.05 to 6.88 upon electrochemical reduction of NRs to ENRs and from 3.17 to 4.99 for reduction of NPs to ENPs, reecting the fact that the electrochemical reduction is indeed a successful method for preparation of reduced graphene oxide
materials. These values are comparable to those found previously.21 Core C1s XPS spectra showed a dramatic decrease of the C–O bond for electrochemically reduced materials when compared to graphene oxide precursors, likely related to the reduction of epoxides on graphene oxide surfaces (Fig. 4). The electrochemical detection of the biomarkers was rst studied using the cyclic voltammetry (CV) technique. Cyclic voltammetry is a simple and rapid technique which is commonly used to study electrochemical performances of new electrode surfaces using analytes which can be readily oxidised or reduced. Hence, a series of CV scans was performed on the four biomarkers in phosphate buffer solution with the three surfaces (bare GC, ENP and ENR electrodes). Aer which, the differential pulse voltammetry (DPV) technique was carried out on the biomarkers to probe the sensitivity of the graphene materials. DPV is a more sensitive technique, where the redox properties of extremely small amounts of analytes can be detected and quantied.
Fig. 1 SEM images of graphene related materials. (A) Unzipped nanoribbon graphene oxide (NR); (B) electrochemically reduced NRs (ENRs); (C) exfoliated SGNF graphene oxide nanoplatelets (NPs); (D) electrochemically reduced nanoplatelets (ENPs). 30 000 magnification, scale bar of 100 nm.
This journal is © The Royal Society of Chemistry 2014
Analyst
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Analyst
XPS survey spectra of graphene related materials. (NR) unzipped nanoribbon graphene oxide; (ENRs) electrochemically reduced NRs; (NPs) exfoliated SGNF graphene oxide nanoplatelets; (ENPs) electrochemically reduced nanoplatelets.
Fig. 3
Fig. 4 XPS C1s core spectra of graphene related materials. (NR) unzipped nanoribbon graphene oxide; (ENRs) electrochemically reduced NRs; (NPs) exfoliated SGNF graphene oxide nanoplatelets; (ENPs) electrochemically reduced nanoplatelets.
3.1
Biomarkers
3.1.1 Cyclic voltammetry study. Fig. 5A illustrates the cyclic voltammograms of uric acid on bare GC, ENP and ENR electrodes. The oxidation peak of uric acid on the ENP surface began from 218 mV, before reaching a maximum of 379 mV,
Analyst
Paper
which proved to be slightly lower than the oxidation peak potential of the ENR surface starting from 234 mV and peaking at 386 mV. The bare GC surface showed the highest oxidation potential among the three surfaces studied, spanning from 247 mV to a maximum at 410 mV. The results showed that the ENP surface was able to provide the lowest oxidation potential, indicating that oxidation can occur with a smaller amount of applied voltage. The oxidation currents of the ENP and ENR surfaces were seen to be much higher than that of the bare GC. This observation could possibly be due to the presence of common metallic particles such as nickel and copper which most likely remained in the SGNF and MWCNT sheets during the CVD production, as elucidated by Griffiths.22 A more signicant difference was observed among the oxidation peaks of ascorbic acid on the three materials as compared to that of uric acid, as shown in Fig. 5B. The ENP surface displayed the lowest oxidation peak, which started from 132 mV and peaked at 25 mV. This peak potential was signicantly lower than that displayed by the bare GC and ENR surface. The ENR surface gave an oxidation peak which initiated at 121 mV and reached a maximum at 162 mV whereas the oxidation peak of ascorbic acid on the bare GC originated from 169 mV before the peak was attained at 450 mV. As mentioned earlier, ENP and ENR surfaces contain edge-like and basal-like planes, respectively, and the larger number of edge-like planes on the ENP surface is likely to result in a lower oxidation peak potential than that on the ENR surface. The observed variation in oxidation peaks of ascorbic acid is in line with a previous study, where the edge-like and basal-like planes provided oxidation potentials of around 0.1 V and 0.3 V, respectively, for the oxidation of ascorbic acid.23 The electrochemical characteristics of the electrochemically reduced graphene were further investigated using dopamine, a vital neurotransmitter, as well. From the cyclic voltammograms shown in Fig. 5C, it can be observed that both ENP and ENR surfaces have identical oxidation peaks at 223 mV, with the only difference being the oxidation peak for the ENP surface started from 38 mV, whereas that of the ENR surface started from 6 mV. Although there was no clear differentiation between the electrochemical behaviours of these two materials, both surfaces displayed a lower oxidation peak than the bare GC, which was recorded to begin at 99 mV and peak at 332 mV. Similar to the aforementioned surfaces, a high concentration of edge-like planes on the graphene-modied electrodes was able to exhibit better electrochemical behaviours and characteristics. Interestingly, two reduction peaks were observed on the CV graphs, at around 0.3 V and 0.1 V, as well. Among the three surfaces studied, the ENP surface displayed the largest peak current, followed by the ENR surface and the bare GC. The large oxidation current on the ENP surface can be attributed to the extensive amount of edge-like planes on the ENP sheets. Other than the three biomarkers discussed, it is furthermore important to carry out an analysis with another signicant biomarker, NADH, which is a key coenzyme in all living cells. The oxidation of NADH on the ENP surface began from 139 mV and reached a maximum at 384 mV, as illustrated in Fig. 5D. The oxidation peak potential was lower than that of
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Paper
Analyst
Cyclic voltammograms of 5 mM uric acid, ascorbic acid, dopamine and NADH on bare GC, ENP and ENR surfaces. Conditions: 50 mM PBS background electrolyte, pH 7.2, scan rate 100 mV s 1.
Fig. 5
ENR and bare GC surfaces, which originated from 331 and 417 mV, and peaked at 538 and 645 mV, respectively. The lower oxidation potentials could be attributed to the density of edgelike planes on the graphene surfaces. A notable observation from the oxidation of NADH using the two graphene surfaces was the early onset of oxidation potentials. Such a pre-peak formation could be related to the adsorption of NADH molecules onto the graphene surfaces.24 It should be noted that the NADH are known to now undergo any electrocatalysis on metallic impurities.25 3.1.2 Differential pulse voltammetry study. Quantication of the sensitivities of the graphene materials for the detection of these biomarkers was equally paramount in this study. Hence, DPV analyses were conducted in a range of biomarker concentrations (1 mM to 5 mM) as shown in Fig. 6. The oxidation current heights were observed to increase with increasing concentration of biomarkers. Subsequent comparisons were performed with calibration plots (insets of Fig. 6) generated from the oxidation current heights obtained from the DPV analyses. From the DPV curves and calibration plots of uric acid as shown in Fig. 6A and the extracted data in Table 1, it was evident that the sensitivity to oxidation of uric acid was the highest on the ENR surface, at 7.52 mA mM 1, followed by the ENP and bare GC surfaces at 4.65 and 3.12 mA mM 1, respectively. The linearity of the response, which was determined by the correlation
This journal is © The Royal Society of Chemistry 2014
coefficient values, R2, was however the highest for the ENP surface at 0.9937. The ENR surface exhibited a poorer linearity, with an R2 value of 0.9789, followed by the bare GC with an R2 value of 0.9707. The oxidation current was also compared among the three surfaces. It was shown that the ENP surface provided the largest oxidation current, indicating that the presence of uric acid can be most easily detected on an ENP surface. As for the calibration results of the oxidation of ascorbic acid on the three surfaces, the ENP surface displayed the highest sensitivity, as illustrated in Fig. 6B. Its sensitivity was 8.66 mA mM 1, followed by the ENR surface 5.94 mA mM 1 and nally the bare GC 2.73 mA mM 1. This implied that a small change in concentration of ascorbic acid on the ENP surface can induce a greater change in peak height of the DPV curves as compared to the other two surfaces. All the three surfaces studied displayed similar linearity based on the R2 values, which were 0.9675, 0.9601 and 0.9744 for the bare GC, ENP and ENR surfaces, respectively. As for the oxidation current, oxidation of ascorbic acid on the ENP surface gave the largest oxidation current in comparison with the other two surfaces. Similar to the oxidation of uric acid, the ENR surface proved to be the most sensitive material for detecting the oxidation of dopamine, followed by the ENP surface and lastly the bare GC as exemplied in Fig. 6C. The ENR surface exhibited a sensitivity of 5.75 mA mM 1; the ENP surface and the bare GC exhibited Analyst
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Analyst
Paper
Fig. 6 DPV profiles of 5 mM uric acid, ascorbic acid, dopamine and NADH on GC, ENP and ENR surfaces. Insets: respective calibration plots at concentrations of 1, 2, 3, 4 and 5 mM. Conditions: 50 mM PBS background electrolyte, pH 7.2.
Table 1 Sensitivity and correlation coefficient of DPV determinations of uric acid, ascorbic acid, dopamine and NADH on bare GC, ENP and ENR surfaces based on calibration curves from 1 mM to 5 mM with 1 mM steps. Standard deviations are given in parentheses
Analyte
Material
Slope/mA mM
Uric acid
Bare GC ENP ENR Bare GC ENP ENR Bare GC ENP ENR Bare GC ENP ENR
3.12 (0.270) 4.65 (0.185) 7.52 (0.551) 2.73 (0.250) 8.66 (0.878) 5.94 (0.480) 3.19 (0.364) 4.89 (1.15) 5.75 (0.795) 2.79 (0.103) 0.87 (0.048) 1.05 (0.083)
Ascorbic acid
Dopamine
NADH
1
R2 0.9707 0.9937 0.9787 0.9675 0.9601 0.9744 0.9498 0.8087 0.9276 0.9946 0.9877 0.9757
sensitivities of 4.89 and 3.19 mA mM 1, respectively. However, the linearity of the response was observed to be the highest for the bare GC surface, with a correlation coefficient of 0.9498. The ENR surface displayed an R2 value of 0.9276 whereas the ENP
Analyst
surface was 0.8087. The graphene surfaces exhibited poorer linearity for the oxidation of dopamine, unlike that of uric acid and ascorbic acid. Nonetheless, the oxidation current observed for oxidation of dopamine was still the highest on the ENP surface, followed by the ENR and bare GC surfaces. The sensitivities of the three surfaces towards the oxidation of NADH were also examined and the results are illustrated in Fig. 6D and Table 1. Contrary to the other three biomarkers analysed previously, the oxidation of NADH on the bare GC gave the highest sensitivity of 2.79 mA mM 1. The ENP and ENR surfaces revealed lower sensitivities of 0.87 and 1.05 mA mM 1, respectively. In addition, the bare GC surface showed the highest linearity of response, as supported by its correlation coefficient of 0.9946. However, this value did not differ signicantly from that of the ENP and ENR surfaces, which were 0.9877 and 0.9757, respectively. This indicated that the three surfaces displayed similar linearity of response for the oxidation of NADH. The comparison of oxidation current was in agreement with the sensitivities of the surfaces, with the bare GC surface giving the largest oxidation current. This discrepancy in trends from the previous three biomarkers can be attributed to the passivation effect of the oxidised product of NADH, which
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Paper
readily attached to the carboxyl groups on edge-like planes of the electrochemically reduced graphene since the reduction potential of 1.45 V was insufficient to reduce carboxyl groups. Such an attachment of NAD+ onto the edges of the graphene inhibited the access of additional NADH species for electrochemical oxidation. Oxidation can only occur to a smaller extent due to the effects of passivation, resulting in lower peak currents. From the differential pulse voltammetry study of the four biomarkers, it was evident that the graphene materials provided enhanced sensitivity for the detection of uric acid, ascorbic acid and dopamine. Oxidation of NADH proved to be most sensitive on the bare GC on the other hand. Following the individual electrochemical detection of the biomarkers, a simultaneous detection of uric acid, ascorbic acid and dopamine was performed by the differential pulse voltammetry technique as shown in Fig. 7. The impact of shis in oxidation potentials with the usage of different electrode surfaces was apparent and signicant. The bare GC was unable to provide a good resolution for the three biomarkers as seen from the overlapped oxidation peaks of ascorbic acid and dopamine. It provided two oxidation peaks at 253 and 455 mV which corresponded to the oxidation of ascorbic acid/dopamine and uric acid, respectively. On the other hand, three distinct peaks were detected when the mixture underwent oxidation on the ENP and ENR surfaces, at 19, 163 and 303 mV. These peaks corresponded to the oxidation of ascorbic acid, dopamine and uric acid, respectively, with a difference of 182 and 140 mV between each set of adjacent peaks. The differences provided good resolutions among the oxidation peaks, which averaged at 1.42 and 1.09, respectively. Hence, multiple compounds with similar oxidation potentials can be more distinctly identied when graphene surfaces were used due to shiing in oxidation potentials. The oxidation current is also noteworthy in this analysis, where oxidation of the mixture on the ENP surface gave the largest oxidation current. In addition, the ascorbic acid peak was also more distinct on the ENP surface in comparison with the ENR surface. Peak potentials and detection limits of
Fig. 7 DPV profiles of a 5 mM mixture of uric acid, ascorbic acid and dopamine on GC, ENP and ENR surfaces. Conditions: 50 mM PBS background electrolyte, pH 7.2.
This journal is © The Royal Society of Chemistry 2014
Analyst
the studied biomarkers at different surfaces as reported in the literature are given in Table S1† (ESI) for comparison.
3.2
DNA bases
The redox properties of DNA bases have been increasingly important in the advancement of nanomaterials in the eld of biosensing. Among the DNA bases, there has been increasing awareness on purine bases, mainly guanine and adenine in the eld of electrochemistry as they are more readily oxidised than the pyrimidine bases.1 The faster rates of oxidation are benecial for quick and easy detection of DNA samples. As such, the three surfaces were probed with purine bases to examine their electrochemical behaviours via cyclic voltammetry and differential pulse voltammetry studies. 3.2.1 Cyclic voltammetry study. Fig. 8A and B display the oxidation peaks of the three surfaces on guanine and adenine, respectively. The oxidation peak of guanine on the ENP surface originated from 579 mV and occurred at 720 mV, which was lower than the oxidation peaks observed for ENR and bare GC surfaces, with onsets from 589 and 602 mV, and reaching a maximum at 752 and 840 mV, respectively. As for adenine, the oxidation potentials for bare GC, ENP and ENR surfaces were 1113, 1093 and 1096 mV, respectively. The origins of their oxidation potentials were also similar at 929, 898 and 880 mV, respectively. While the difference in the oxidation potentials
Fig. 8 Cyclic voltammograms of 5 mM guanine and adenine on bare
GC, ENP and ENR surfaces. Conditions: 50 mM PBS background electrolyte, pH 7.2, scan rate 100 mV s 1.
Analyst
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Analyst
Paper
Fig. 9 DPV profiles of 5 mM guanine and adenine on GC, ENP and ENR surfaces. Insets: respective calibration plots at concentrations of 1, 2, 3, 4 and 5 mM. Conditions: 50 mM PBS background electrolyte, pH 7.2.
was marginal, it was still evident that the ENP surface provided the lowest peak potential. Therefore, the electrochemical behaviours of the three surfaces on the oxidation of DNA bases displayed a close resemblance to the four biomarkers studied earlier, with the ENP surface exhibiting the best performance in general, followed by ENR and bare GC surfaces. 3.2.2 Differential pulse voltammetry study. Further analyses were carried out on the two DNA bases using the DPV technique to compare the detection sensitivities of the three surfaces. The most sensitive surface to the oxidation of guanine base, according to Fig. 9A, was the ENP surface at 3.03 mA mM 1. The sensitivity dropped to 1.18 mA mM 1 for the ENR surface and declined considerably to 0.132 mA mM 1 for the bare GC. In addition, the ENP surface provided the largest magnitude of response, as shown from the peak heights, followed by ENR and then bare GC surfaces. The linearity of response for oxidation of guanine base was, however, the best for the bare GC with an R2 value of 0.9984 and poorer for the graphene-modied electrodes. The ENR surface gave an average R2 value of 0.9940 while the ENP surface displayed an R2 value of 0.9620. This implied that despite the high sensitivity and large oxidation current from the ENP surface, its performance in the linearity of response is not as outstanding as the other two surfaces. As inferred from Fig. 9B, the oxidation of adenine base proved to be most sensitive on the ENR surface at 2.02 mA mM 1. The remaining two materials, ENP and bare GC, showed lower sensitivities at 1.39 and 0.513 mA mM 1, respectively. The linearity of response is similar for the ENR and bare GC surfaces, as indicated by the R2 values in Table 2. However, the ENP surface displayed a poor linearity of response. It can be inferred from the low R2 value of 0.5787 that the calibration range of 1 to 5 mM was not the most appropriate for accurate determination of adenine concentration in living cells. On top of that, it was notable that two peaks were observed during the oxidation of adenine on the ENP surface, whereas only one peak was seen for the other surfaces. The presence of a second peak might be due to the dimerization of adenine (A), forming the
Analyst
Sensitivity and correlation coefficient of DPV determinations of uric acid, ascorbic acid, dopamine and NADH on bare GC, ENP and ENR surfaces based on calibration curves from 1 mM to 5 mM with 1 mM steps. Standard deviations are given in parentheses
Table 2
Analyte
Material
Slope/mA mM
Guanine
Bare GC ENP ENR Bare GC ENP ENR
0.13 (0.003) 3.03 (0.300) 1.18 (0.046) 0.51 (0.048) 1.39 (0.544) 2.02 (0.212)
Adenine
1
R2 0.9984 0.9620 0.9940 0.9654 0.5787 0.9576
deprotonated radical A+_ which generated a second oxidation peak.26 Despite that, the ENP surface still provided the largest oxidation current as compared to the other two surfaces, which coincided with the trend of oxidation current for most of the compounds previously discussed. Similar to the biomarkers, the graphene materials showed heightened sensitivities and oxidation currents when probed with the two DNA bases in comparison with bare GC. Peak potentials and detection limits of studied biomarkers at different surfaces as reported in the literature are stated in Table S1† (ESI) for comparison.
4. Conclusions The electrochemical performances of electrochemically reduced graphene, specically nanoplatelets and nanoribbons, were compared against bare GC using four biomarkers, mainly uric acid, ascorbic acid, dopamine, NADH, and two DNA bases, specically guanine and adenine. Effects of the materials on the oxidation potentials of the analytes were examined using CV and it was established that graphene-coated electrodes led to lower oxidation potentials for all the biomarkers and DNA bases. This indicated that a smaller amount of potential was required for oxidation to occur in the presence of graphene materials. Parameters such as sensitivity of the surfaces were analysed based on results from DPV and calibration plots. It was
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Paper
found that the graphene materials displayed better sensitivities for the detection of the biomarkers and DNA bases examined other than NADH. The inconsistency on NADH was justied by the possibility of the passivation effect. A comparison between the two graphene materials showed that, in general, the ENP surface brought about lower oxidation potentials for all biomarkers and DNA bases and higher sensitivities for oxidation of ascorbic acid and guanine than the ENR surface. On the other hand, the ENR surface provided better sensitivities for uric acid, dopamine and adenine. Though it was hypothesized that the ENP surface will generally provide a better sensitivity due to an extensive amount of exposed edgelike planes, the nature of the biomarkers and DNA bases may contribute to the sensitivity outcomes as well. In the case of selectivity, graphene materials evidently exhibited higher selectivities as compared to the bare GC. Between the two graphene materials, it was more advantageous to utilise the ENP surface to achieve lower electrochemical oxidation potentials. However, there was no clear advantage of using a specic graphene material as they displayed different sensitivities for different compounds, despite the ENP surface exhibiting a larger oxidation current for all the compounds studied. In conclusion, the graphene materials have shown better abilities in detecting biomarkers and DNA bases than the bare GC. Further investigations can be carried out with the two other DNA bases, thymine and cytosine, as well as other substantial biomarkers to differentiate the performance of the ENP and ENR surfaces more clearly to rationalise how the difference in graphene structures can affect their electrochemical properties.
Notes and references 1 E. Paleˇ cek and M. Fojta, Anal. Chem., 2001, 73, 74A. 2 P. Gai, H. Zhang, Y. Zhang, W. Liu, G. Zhu, X. Zhang and J. Chen, J. Mater. Chem. B, 2013, 1, 2742. 3 E. Paleˇ cek, Electroanalysis, 1996, 8, 7. 4 E. Paleˇ cek, Nature, 1958, 4751, 656. 5 E. Paleˇ cek, Electroanalysis, 2009, 21, 239. 6 E. P. Randviir and C. E. Banks, RSC Adv., 2012, 2, 5800. 7 M. H. Yang, A. Javadi, H. Li and S. Q. Gong, Biosens. Bioelectron., 2010, 26, 560.
This journal is © The Royal Society of Chemistry 2014
Analyst
8 A. K. Geim and K. S. Novoselov, Nat. Mater., 2007, 6, 183. 9 D. A. C. Brownson, D. K. Kampouris and C. E. Banks, Chem. Soc. Rev., 2012, 41, 6944. 10 R. R. Nair, P. Blake, A. N. Grigorenko, K. S. Novoselov, T. J. Booth, T. Stauber, N. M. R. Peres and A. K. Geim, Science, 2008, 320, 1308. 11 J. S. Bunch, A. M. Van der Zande, S. S. Verbridge, I. W. Frank, D. M. Tanenbaum, J. M. Parpia, H. G. Craighead and P. L. McEuen, Science, 2007, 315, 490. 12 K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666. 13 C. N. R. Rao, A. K. Sood, K. S. Subrahmanyam and A. Govindaraj, Angew. Chem., Int. Ed., 2009, 48, 7752. 14 D. V. Kosynkin, A. L. Higginbotham, A. Sinitskii, J. R. Lomeda, A. Dimiev, B. K. Price and J. M. Tour, Nature, 2009, 458, 872. 15 C. K. Chua, Z. Sofer and M. Pumera, Chem.–Asian J., 2012, 7, 2367. 16 T. Kim, S. Lim, K. Kwon, S.-H. Hong, W. Qiao, C. K. Rhee, S.-H. Yoon and I. Mochida, Langmuir, 2006, 22, 9086. 17 T. J. Davies, M. E. Hyde and R. G. Compton, Angew. Chem., Int. Ed., 2005, 44, 5121. 18 (a) A. F. Holloway, K. Toghill, G. G. Wildgoose, R. G. Compton, M. A. H. Ward, G. Tobias, S. A. Llewellyn, B. Ballesteros, M. L. H. Green and A. Crossley, J. Phys. Chem. C, 2008, 112, 10389; (b) A. Ambrosi and M. Pumera, Chem. Asian J., 2010, 5, 266. 19 D. K. Kampouris and C. E. Banks, Chem. Commun., 2010, 46, 8986. 20 T. Kolodiazhnyi and M. Pumera, Small, 2008, 4, 1476. 21 A. Ambrosi, A. Bonanni, Z. Sofer, J. S. Cross and M. Pumera, Chem.–Eur. J., 2011, 17, 10763. 22 M. J. Griffiths, J. Biol. Chem., 1952, 197, 399. 23 C. L. Scott and M. Pumera, Electrochem. Commun., 2011, 13, 426. 24 M. Pumera, R. Scipioni, H. Iwai, T. Ohno, Y. Miyahara and M. Boero, Chem.–Eur. J., 2009, 15, 10851. 25 M. Pumera, A. Ambrosi and E. L. K. Chng, Chem. Sci., 2012, 3, 3347. 26 K. Kobayashi, J. Phys. Chem. B, 2010, 114, 5600.
Analyst
PAPER
View Journal
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Detection of biomarkers with graphene nanoplatelets and nanoribbons† Cite this: DOI: 10.1039/c3an01585h
Chee Shan Lim, Chun Kiang Chua and Martin Pumera* Well-defined graphene nanosheets have become increasingly popular in the electrochemical detection and quantification of small molecules. In this work, the electrochemical oxidation of biomarkers such as uric acid, ascorbic acid, dopamine, NADH and DNA bases, namely guanine and adenine, was performed using cyclic voltammetry and differential pulse voltammetry to compare the electrochemical properties of electrochemically reduced nanoplatelets (ENPs) and electrochemically reduced nanoribbons (ENRs). The graphene materials displayed better electrochemical performances than the bare glassy carbon surface. Between the two graphene materials, the oxidation of biomarkers occurred at lower oxidation potentials on the ENP surface. The sensitivities of the two graphene surfaces varied when different biomarkers were studied. The ENP surface showed enhanced sensitivities for ascorbic acid, while the ENR surface exhibited higher sensitivities for uric acid and dopamine. As for the DNA bases analysed, both guanine
Received 21st August 2013 Accepted 18th November 2013
and adenine were oxidised at lower potentials on the ENP surface than the ENR surface. The ENP DOI: 10.1039/c3an01585h
surface displayed a better sensitivity for guanine, whereas the oxidation of adenine was more sensitive
www.rsc.org/analyst
on the ENR surface.
1. Introduction A healthy human body has the ability to maintain a steady internal environment, which is also known as homeostasis. While the organ systems function to achieve homeostasis via the regulation of various positive and negative feedback mechanisms, an inability to maintain a constant internal environment can result in diseases or even death. Thus, the human body has to depend on a series of internal feedback mechanisms to achieve such homeostatic balance. Existing scientic development aims to assist in the detection of abnormalities in the human body by monitoring several signicant biomarkers. The detection of biomarkers such as uric acid, ascorbic acid, dopamine, reduced nicotinamide adenine dinucleotide (NADH) and deoxyribonucleic acid (DNA) is able to provide useful information for the prevention of genetic diseases and other health ailments, as well as their medical treatments subsequently. The electrochemical study of biomarkers and DNA bases, which rely mainly on their redox properties,1 has proven to be of signicant importance in the advancement of biosensing for its low cost, simplicity and quick response.2 Following the introduction of electrochemical detection of nucleic acids with nucleic acid-modied electrodes by Paleˇcek, various electrode
Division of Chemistry & Biological Chemistry, School of Physical and Mathematical Sciences, Nanyang Technological University, Singapore 637371, Singapore. E-mail: [email protected]; [email protected]; Fax: +65 67911961 † Electronic supplementary 10.1039/c3an01585h
information
(ESI)
available.
This journal is © The Royal Society of Chemistry 2014
See
DOI:
surfaces such as gold, mercury and carbon have been used in the development of label-free electrochemical genosensors.3–5 In recent years, studies have been carried out using nanomaterials to show improved electrochemical responses for analyses of both DNA bases and biomarkers.6,7 Graphene is a one-atom thick material containing sp2bonded carbon atoms arranged in a honeycomb structure.8 It has been known to have unique electrochemical,9 optical,10 mechanical11 and electrical12 properties since it was rst isolated in 2004. These properties, together with its low cost and little impact on the environment, have made graphene a suitable material for sensing applications.13 Fabrication of graphene materials can be achieved by a ‘top-down’ approach, via sono-chemical or thermal exfoliation of graphite,9 or a ‘bottomup’ growth using chemical vapour deposition (CVD). Of which, the former approach is most commonly used for modication of graphite to other chemically modied graphene materials. The oxidation of graphite produces graphite oxide, which can then be reduced at high temperatures to form thermally reduced graphite oxide. Graphite oxide can also undergo ultrasonication to generate graphene oxide, which can be reduced to electrochemically reduced graphene oxide and chemically reduced graphene oxide by applying a reduction potential or adding a reducing agent, respectively. Apart from the aforementioned methods, graphene can also be obtained by the lateral unzipping of stacked graphene nanobers (SGNFs) or axial unzipping of multi-walled carbon nanotubes (MWCNTs), as shown in Scheme 1.14,15 SGNFs have a perpendicular orientation of the graphene sheets to the long axis of the ber, therefore possessing a large number of Analyst
View Article Online
Analyst
Paper
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
monobasic, sodium phosphate dibasic, potassium chloride, sodium chloride, uric acid, ascorbic acid, dopamine, guanine hydrochloride, adenine hydrochloride (>99%), and reduced nicotinamide adenine dinucleotide were purchased from Sigma-Aldrich, Singapore. Stacked graphene platelet nanobers (acid washed) were obtained from Strem. Glassy carbon electrodes with a diameter of 3 mm were obtained from Autolab, The Netherlands. Milli-Q water with a resistivity of 18.2 MU cm was used throughout the experiments. 2.2 Scheme 1 Schematic representation of MWCNTs and SGNFs undergoing oxidation, resulting in axial and lateral unzipping. Subsequent ultrasonication treatment and electrochemical reduction provided graphene nanoplatelets and nanoribbons.
edge-plane sites.16 Such an arrangement gives SGNFs their unique electrochemical behaviours, as almost only the edges are exposed in SGNFs. On the other hand, MWCNTs resemble closed graphene nanobers which are rolled up perfectly, with axial orientation to the long axes of the nanobers. Lateral unzipping of stacked graphene nanobers in this work generated graphene nanoplatelet sheets of dimensions 50 50 nm, whereas axial unzipping of multi-walled carbon nanotubes gave graphene nanoribbons of dimensions 5000 300 nm.15 Studies have shown that graphene contains two types of surfaces, basaland edge-like planes upon and of which graphene structures with more exposed ends and defects (edge-like) exhibit better electrochemical properties, by providing faster rates of heterogeneous electron transfer (HET).17,18 Since the electrochemical behaviours of these graphene-based nanomaterials are ruled by the edge-plane sites,19 it is speculated that the electrochemical properties of the nanoplatelet sheets will exceed those of nanoribbon sheets. The detection methods of biomarkers and DNA bases were previously carried out using chemically modied graphenes.16,19 However, no preceding research has been done to compare the analytical performances such as sensitivity, linearity of response and the effect on the oxidation potential of electrochemically reduced graphene nanoplatelets (ENPs) and graphene nanoribbons (ENRs) sheets on biomarkers and DNA bases. Therefore, this work aims to examine the electrochemical properties, as well as the response of four key biomarkers, uric acid, ascorbic acid, dopamine and NADH; and two DNA bases, namely guanine and adenine, on the ENP and ENR surfaces. The comparative study was carried out against bare glassy carbon (GC). More importantly, simultaneous electrochemical detection was performed on a mixture of uric acid, ascorbic acid and dopamine, as it is challenging to obtain good resolution and detection sensitivity of these biomarkers using conventional electrode surfaces.
2. 2.1
Experimental Materials
Multi-walled carbon nanotubes (previously characterized in ref. 20), N,N-dimethylformamide (DMF), potassium phosphate
Analyst
Apparatus
Cyclic voltammetry and differential pulse voltammetry measurements were carried out with a mAutolab type III electrochemical analyser (Eco Chemie, The Netherlands) connected to a personal computer and controlled by General Purpose Electrochemical Systems Version 4.9 soware (Eco Chemie). The parameters used for differential pulse voltammetry are as follows: 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 potential. Raw data obtained using differential pulse voltammetry underwent a baseline correction with a peak width of 0.01 using the GPES soware. All voltammetry experiments were performed in a 5 mL electrochemical cell at room temperature using a three-electrode conguration. A platinum electrode served as an auxiliary electrode, whereas an Ag/AgCl electrode was utilised as a reference electrode. All electrochemical potentials in this report are stated vs. the Ag/AgCl reference electrode. A scanning electron microscope (JEOL JSM-7600F eld-emission, Japan) was used to obtain images under normal SEM and gentlebeam (GB) mode. The samples were attached onto an aluminum sample stub using adhesive conductive carbon tape. The XPS used for analysis is a Phoibos 100 spectrometer and a Mg X-ray radiation source (SPECS, Germany). All wide-range and high resolution scans were performed at 12.53 kV. The materials were attached onto an aluminum XPS sample holder using sticky conductive carbon tape. Special attention was paid to ensure that a homogeneous and uniform layer of the material was attached onto the tape. Raman spectroscopic measurements were carried out using a confocal micro-Raman LabRam HR instrument (Horiba Scientic) in backscattering geometry with a CCD detector. Calibration was performed at 0 cm 1 and 520 cm 1 with a silicon wafer to give a peak position resolution of less than 1 cm 1. All Raman measurements of the materials were carried out using a 514.5 nm Ar laser, an Olympus optical microscope and a 100 objective lens. 2.3
Procedure
Oxidised graphene nanoplatelets and nanoribbons were previously prepared according to the modied Hummers method.15 The materials were fully characterised by X-ray photoelectron spectroscopy, Raman spectroscopy, transmission electron microscopy and scanning electron microscopy.15 Electrochemically reduced graphene was obtained by applying a potential of 1.45 V for 900 seconds to a graphene oxide-modied GC electrode in a 50 mM phosphate buffer solution (pH 7.2). The graphene oxide-modied GC electrode exhibited strong
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Paper
Analyst
voltammetric reduction peaks between 0.8 and 1.6 V with the maximum current being displayed at 1.45 V. It is interesting to note that the electrochemical reduction potential values of some common oxygen-containing groups, peroxy, aldehyde, epoxyl and carboxyl, are around 0.7 V, 1.0 V, 1.1 V and 2.0 V, respectively. Hence, the reduction peak of 1.45 V observed is equivalent to the reduction peak of graphene oxide on various electrode surfaces with the removal of all the aforementioned oxygen-containing groups except carboxyl groups. Glassy carbon electrode surfaces were renewed by polishing with 0.05 mm alumina powder. To immobilise the carbon materials onto the working electrodes, a suspension of the desired material with a concentration of 1 mg mL 1 in DMF, followed by a 1 min sonication, was rst prepared. 1 mL aliquot of the appropriate suspension was then deposited onto the electrode surface. Evaporation of the solvent at room temperature was then allowed to take place, giving a randomly distributed lm on the glassy carbon electrode surface. The reproducibility of the measurements was obtained over 3 experiments, each time using 3 different electrode units. All voltammetry measurements were conducted with a 50 mM phosphate buffer solution (PBS) of pH 7.2.
3.
Fig. 2 Raman spectra of graphene related materials. (NR) unzipped nanoribbon graphene oxide; (ENRs) electrochemically reduced NRs; (NPs) exfoliated SGNF graphene oxide nanoplatelets; (ENPs) electrochemically reduced nanoplatelets.
Results and discussion
Prior to electrochemical and analytical studies, we characterized the materials used. Fig. 1 shows the SEM images of the materials. The morphology of the materials did not signicantly change upon electrochemical reduction. Raman spectroscopy revealed that the density of defects increased upon electrochemical reduction, the intensity of the defect (sp3) related peak at 1350 cm 1 (D-peak) vs. the intensity of the peak related to sp2 hybridization at 1560 cm 1 (G-peak) increased from 0.84 to 1.27 in the case of graphene oxides made from carbon nanotubes (D/G ratio). The D/G ratio of graphene oxide nanoplatelets increased from 0.86 to 1.05 upon electrochemical reduction (Fig. 2). X-ray photoelectron spectroscopy (XPS) was used to shed light on the efficiency of electrochemical reduction. From wide scan XPS spectra (Fig. 3), it is obvious that the C/O ratio increased from 3.05 to 6.88 upon electrochemical reduction of NRs to ENRs and from 3.17 to 4.99 for reduction of NPs to ENPs, reecting the fact that the electrochemical reduction is indeed a successful method for preparation of reduced graphene oxide
materials. These values are comparable to those found previously.21 Core C1s XPS spectra showed a dramatic decrease of the C–O bond for electrochemically reduced materials when compared to graphene oxide precursors, likely related to the reduction of epoxides on graphene oxide surfaces (Fig. 4). The electrochemical detection of the biomarkers was rst studied using the cyclic voltammetry (CV) technique. Cyclic voltammetry is a simple and rapid technique which is commonly used to study electrochemical performances of new electrode surfaces using analytes which can be readily oxidised or reduced. Hence, a series of CV scans was performed on the four biomarkers in phosphate buffer solution with the three surfaces (bare GC, ENP and ENR electrodes). Aer which, the differential pulse voltammetry (DPV) technique was carried out on the biomarkers to probe the sensitivity of the graphene materials. DPV is a more sensitive technique, where the redox properties of extremely small amounts of analytes can be detected and quantied.
Fig. 1 SEM images of graphene related materials. (A) Unzipped nanoribbon graphene oxide (NR); (B) electrochemically reduced NRs (ENRs); (C) exfoliated SGNF graphene oxide nanoplatelets (NPs); (D) electrochemically reduced nanoplatelets (ENPs). 30 000 magnification, scale bar of 100 nm.
This journal is © The Royal Society of Chemistry 2014
Analyst
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Analyst
XPS survey spectra of graphene related materials. (NR) unzipped nanoribbon graphene oxide; (ENRs) electrochemically reduced NRs; (NPs) exfoliated SGNF graphene oxide nanoplatelets; (ENPs) electrochemically reduced nanoplatelets.
Fig. 3
Fig. 4 XPS C1s core spectra of graphene related materials. (NR) unzipped nanoribbon graphene oxide; (ENRs) electrochemically reduced NRs; (NPs) exfoliated SGNF graphene oxide nanoplatelets; (ENPs) electrochemically reduced nanoplatelets.
3.1
Biomarkers
3.1.1 Cyclic voltammetry study. Fig. 5A illustrates the cyclic voltammograms of uric acid on bare GC, ENP and ENR electrodes. The oxidation peak of uric acid on the ENP surface began from 218 mV, before reaching a maximum of 379 mV,
Analyst
Paper
which proved to be slightly lower than the oxidation peak potential of the ENR surface starting from 234 mV and peaking at 386 mV. The bare GC surface showed the highest oxidation potential among the three surfaces studied, spanning from 247 mV to a maximum at 410 mV. The results showed that the ENP surface was able to provide the lowest oxidation potential, indicating that oxidation can occur with a smaller amount of applied voltage. The oxidation currents of the ENP and ENR surfaces were seen to be much higher than that of the bare GC. This observation could possibly be due to the presence of common metallic particles such as nickel and copper which most likely remained in the SGNF and MWCNT sheets during the CVD production, as elucidated by Griffiths.22 A more signicant difference was observed among the oxidation peaks of ascorbic acid on the three materials as compared to that of uric acid, as shown in Fig. 5B. The ENP surface displayed the lowest oxidation peak, which started from 132 mV and peaked at 25 mV. This peak potential was signicantly lower than that displayed by the bare GC and ENR surface. The ENR surface gave an oxidation peak which initiated at 121 mV and reached a maximum at 162 mV whereas the oxidation peak of ascorbic acid on the bare GC originated from 169 mV before the peak was attained at 450 mV. As mentioned earlier, ENP and ENR surfaces contain edge-like and basal-like planes, respectively, and the larger number of edge-like planes on the ENP surface is likely to result in a lower oxidation peak potential than that on the ENR surface. The observed variation in oxidation peaks of ascorbic acid is in line with a previous study, where the edge-like and basal-like planes provided oxidation potentials of around 0.1 V and 0.3 V, respectively, for the oxidation of ascorbic acid.23 The electrochemical characteristics of the electrochemically reduced graphene were further investigated using dopamine, a vital neurotransmitter, as well. From the cyclic voltammograms shown in Fig. 5C, it can be observed that both ENP and ENR surfaces have identical oxidation peaks at 223 mV, with the only difference being the oxidation peak for the ENP surface started from 38 mV, whereas that of the ENR surface started from 6 mV. Although there was no clear differentiation between the electrochemical behaviours of these two materials, both surfaces displayed a lower oxidation peak than the bare GC, which was recorded to begin at 99 mV and peak at 332 mV. Similar to the aforementioned surfaces, a high concentration of edge-like planes on the graphene-modied electrodes was able to exhibit better electrochemical behaviours and characteristics. Interestingly, two reduction peaks were observed on the CV graphs, at around 0.3 V and 0.1 V, as well. Among the three surfaces studied, the ENP surface displayed the largest peak current, followed by the ENR surface and the bare GC. The large oxidation current on the ENP surface can be attributed to the extensive amount of edge-like planes on the ENP sheets. Other than the three biomarkers discussed, it is furthermore important to carry out an analysis with another signicant biomarker, NADH, which is a key coenzyme in all living cells. The oxidation of NADH on the ENP surface began from 139 mV and reached a maximum at 384 mV, as illustrated in Fig. 5D. The oxidation peak potential was lower than that of
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Paper
Analyst
Cyclic voltammograms of 5 mM uric acid, ascorbic acid, dopamine and NADH on bare GC, ENP and ENR surfaces. Conditions: 50 mM PBS background electrolyte, pH 7.2, scan rate 100 mV s 1.
Fig. 5
ENR and bare GC surfaces, which originated from 331 and 417 mV, and peaked at 538 and 645 mV, respectively. The lower oxidation potentials could be attributed to the density of edgelike planes on the graphene surfaces. A notable observation from the oxidation of NADH using the two graphene surfaces was the early onset of oxidation potentials. Such a pre-peak formation could be related to the adsorption of NADH molecules onto the graphene surfaces.24 It should be noted that the NADH are known to now undergo any electrocatalysis on metallic impurities.25 3.1.2 Differential pulse voltammetry study. Quantication of the sensitivities of the graphene materials for the detection of these biomarkers was equally paramount in this study. Hence, DPV analyses were conducted in a range of biomarker concentrations (1 mM to 5 mM) as shown in Fig. 6. The oxidation current heights were observed to increase with increasing concentration of biomarkers. Subsequent comparisons were performed with calibration plots (insets of Fig. 6) generated from the oxidation current heights obtained from the DPV analyses. From the DPV curves and calibration plots of uric acid as shown in Fig. 6A and the extracted data in Table 1, it was evident that the sensitivity to oxidation of uric acid was the highest on the ENR surface, at 7.52 mA mM 1, followed by the ENP and bare GC surfaces at 4.65 and 3.12 mA mM 1, respectively. The linearity of the response, which was determined by the correlation
This journal is © The Royal Society of Chemistry 2014
coefficient values, R2, was however the highest for the ENP surface at 0.9937. The ENR surface exhibited a poorer linearity, with an R2 value of 0.9789, followed by the bare GC with an R2 value of 0.9707. The oxidation current was also compared among the three surfaces. It was shown that the ENP surface provided the largest oxidation current, indicating that the presence of uric acid can be most easily detected on an ENP surface. As for the calibration results of the oxidation of ascorbic acid on the three surfaces, the ENP surface displayed the highest sensitivity, as illustrated in Fig. 6B. Its sensitivity was 8.66 mA mM 1, followed by the ENR surface 5.94 mA mM 1 and nally the bare GC 2.73 mA mM 1. This implied that a small change in concentration of ascorbic acid on the ENP surface can induce a greater change in peak height of the DPV curves as compared to the other two surfaces. All the three surfaces studied displayed similar linearity based on the R2 values, which were 0.9675, 0.9601 and 0.9744 for the bare GC, ENP and ENR surfaces, respectively. As for the oxidation current, oxidation of ascorbic acid on the ENP surface gave the largest oxidation current in comparison with the other two surfaces. Similar to the oxidation of uric acid, the ENR surface proved to be the most sensitive material for detecting the oxidation of dopamine, followed by the ENP surface and lastly the bare GC as exemplied in Fig. 6C. The ENR surface exhibited a sensitivity of 5.75 mA mM 1; the ENP surface and the bare GC exhibited Analyst
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Analyst
Paper
Fig. 6 DPV profiles of 5 mM uric acid, ascorbic acid, dopamine and NADH on GC, ENP and ENR surfaces. Insets: respective calibration plots at concentrations of 1, 2, 3, 4 and 5 mM. Conditions: 50 mM PBS background electrolyte, pH 7.2.
Table 1 Sensitivity and correlation coefficient of DPV determinations of uric acid, ascorbic acid, dopamine and NADH on bare GC, ENP and ENR surfaces based on calibration curves from 1 mM to 5 mM with 1 mM steps. Standard deviations are given in parentheses
Analyte
Material
Slope/mA mM
Uric acid
Bare GC ENP ENR Bare GC ENP ENR Bare GC ENP ENR Bare GC ENP ENR
3.12 (0.270) 4.65 (0.185) 7.52 (0.551) 2.73 (0.250) 8.66 (0.878) 5.94 (0.480) 3.19 (0.364) 4.89 (1.15) 5.75 (0.795) 2.79 (0.103) 0.87 (0.048) 1.05 (0.083)
Ascorbic acid
Dopamine
NADH
1
R2 0.9707 0.9937 0.9787 0.9675 0.9601 0.9744 0.9498 0.8087 0.9276 0.9946 0.9877 0.9757
sensitivities of 4.89 and 3.19 mA mM 1, respectively. However, the linearity of the response was observed to be the highest for the bare GC surface, with a correlation coefficient of 0.9498. The ENR surface displayed an R2 value of 0.9276 whereas the ENP
Analyst
surface was 0.8087. The graphene surfaces exhibited poorer linearity for the oxidation of dopamine, unlike that of uric acid and ascorbic acid. Nonetheless, the oxidation current observed for oxidation of dopamine was still the highest on the ENP surface, followed by the ENR and bare GC surfaces. The sensitivities of the three surfaces towards the oxidation of NADH were also examined and the results are illustrated in Fig. 6D and Table 1. Contrary to the other three biomarkers analysed previously, the oxidation of NADH on the bare GC gave the highest sensitivity of 2.79 mA mM 1. The ENP and ENR surfaces revealed lower sensitivities of 0.87 and 1.05 mA mM 1, respectively. In addition, the bare GC surface showed the highest linearity of response, as supported by its correlation coefficient of 0.9946. However, this value did not differ signicantly from that of the ENP and ENR surfaces, which were 0.9877 and 0.9757, respectively. This indicated that the three surfaces displayed similar linearity of response for the oxidation of NADH. The comparison of oxidation current was in agreement with the sensitivities of the surfaces, with the bare GC surface giving the largest oxidation current. This discrepancy in trends from the previous three biomarkers can be attributed to the passivation effect of the oxidised product of NADH, which
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Paper
readily attached to the carboxyl groups on edge-like planes of the electrochemically reduced graphene since the reduction potential of 1.45 V was insufficient to reduce carboxyl groups. Such an attachment of NAD+ onto the edges of the graphene inhibited the access of additional NADH species for electrochemical oxidation. Oxidation can only occur to a smaller extent due to the effects of passivation, resulting in lower peak currents. From the differential pulse voltammetry study of the four biomarkers, it was evident that the graphene materials provided enhanced sensitivity for the detection of uric acid, ascorbic acid and dopamine. Oxidation of NADH proved to be most sensitive on the bare GC on the other hand. Following the individual electrochemical detection of the biomarkers, a simultaneous detection of uric acid, ascorbic acid and dopamine was performed by the differential pulse voltammetry technique as shown in Fig. 7. The impact of shis in oxidation potentials with the usage of different electrode surfaces was apparent and signicant. The bare GC was unable to provide a good resolution for the three biomarkers as seen from the overlapped oxidation peaks of ascorbic acid and dopamine. It provided two oxidation peaks at 253 and 455 mV which corresponded to the oxidation of ascorbic acid/dopamine and uric acid, respectively. On the other hand, three distinct peaks were detected when the mixture underwent oxidation on the ENP and ENR surfaces, at 19, 163 and 303 mV. These peaks corresponded to the oxidation of ascorbic acid, dopamine and uric acid, respectively, with a difference of 182 and 140 mV between each set of adjacent peaks. The differences provided good resolutions among the oxidation peaks, which averaged at 1.42 and 1.09, respectively. Hence, multiple compounds with similar oxidation potentials can be more distinctly identied when graphene surfaces were used due to shiing in oxidation potentials. The oxidation current is also noteworthy in this analysis, where oxidation of the mixture on the ENP surface gave the largest oxidation current. In addition, the ascorbic acid peak was also more distinct on the ENP surface in comparison with the ENR surface. Peak potentials and detection limits of
Fig. 7 DPV profiles of a 5 mM mixture of uric acid, ascorbic acid and dopamine on GC, ENP and ENR surfaces. Conditions: 50 mM PBS background electrolyte, pH 7.2.
This journal is © The Royal Society of Chemistry 2014
Analyst
the studied biomarkers at different surfaces as reported in the literature are given in Table S1† (ESI) for comparison.
3.2
DNA bases
The redox properties of DNA bases have been increasingly important in the advancement of nanomaterials in the eld of biosensing. Among the DNA bases, there has been increasing awareness on purine bases, mainly guanine and adenine in the eld of electrochemistry as they are more readily oxidised than the pyrimidine bases.1 The faster rates of oxidation are benecial for quick and easy detection of DNA samples. As such, the three surfaces were probed with purine bases to examine their electrochemical behaviours via cyclic voltammetry and differential pulse voltammetry studies. 3.2.1 Cyclic voltammetry study. Fig. 8A and B display the oxidation peaks of the three surfaces on guanine and adenine, respectively. The oxidation peak of guanine on the ENP surface originated from 579 mV and occurred at 720 mV, which was lower than the oxidation peaks observed for ENR and bare GC surfaces, with onsets from 589 and 602 mV, and reaching a maximum at 752 and 840 mV, respectively. As for adenine, the oxidation potentials for bare GC, ENP and ENR surfaces were 1113, 1093 and 1096 mV, respectively. The origins of their oxidation potentials were also similar at 929, 898 and 880 mV, respectively. While the difference in the oxidation potentials
Fig. 8 Cyclic voltammograms of 5 mM guanine and adenine on bare
GC, ENP and ENR surfaces. Conditions: 50 mM PBS background electrolyte, pH 7.2, scan rate 100 mV s 1.
Analyst
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Analyst
Paper
Fig. 9 DPV profiles of 5 mM guanine and adenine on GC, ENP and ENR surfaces. Insets: respective calibration plots at concentrations of 1, 2, 3, 4 and 5 mM. Conditions: 50 mM PBS background electrolyte, pH 7.2.
was marginal, it was still evident that the ENP surface provided the lowest peak potential. Therefore, the electrochemical behaviours of the three surfaces on the oxidation of DNA bases displayed a close resemblance to the four biomarkers studied earlier, with the ENP surface exhibiting the best performance in general, followed by ENR and bare GC surfaces. 3.2.2 Differential pulse voltammetry study. Further analyses were carried out on the two DNA bases using the DPV technique to compare the detection sensitivities of the three surfaces. The most sensitive surface to the oxidation of guanine base, according to Fig. 9A, was the ENP surface at 3.03 mA mM 1. The sensitivity dropped to 1.18 mA mM 1 for the ENR surface and declined considerably to 0.132 mA mM 1 for the bare GC. In addition, the ENP surface provided the largest magnitude of response, as shown from the peak heights, followed by ENR and then bare GC surfaces. The linearity of response for oxidation of guanine base was, however, the best for the bare GC with an R2 value of 0.9984 and poorer for the graphene-modied electrodes. The ENR surface gave an average R2 value of 0.9940 while the ENP surface displayed an R2 value of 0.9620. This implied that despite the high sensitivity and large oxidation current from the ENP surface, its performance in the linearity of response is not as outstanding as the other two surfaces. As inferred from Fig. 9B, the oxidation of adenine base proved to be most sensitive on the ENR surface at 2.02 mA mM 1. The remaining two materials, ENP and bare GC, showed lower sensitivities at 1.39 and 0.513 mA mM 1, respectively. The linearity of response is similar for the ENR and bare GC surfaces, as indicated by the R2 values in Table 2. However, the ENP surface displayed a poor linearity of response. It can be inferred from the low R2 value of 0.5787 that the calibration range of 1 to 5 mM was not the most appropriate for accurate determination of adenine concentration in living cells. On top of that, it was notable that two peaks were observed during the oxidation of adenine on the ENP surface, whereas only one peak was seen for the other surfaces. The presence of a second peak might be due to the dimerization of adenine (A), forming the
Analyst
Sensitivity and correlation coefficient of DPV determinations of uric acid, ascorbic acid, dopamine and NADH on bare GC, ENP and ENR surfaces based on calibration curves from 1 mM to 5 mM with 1 mM steps. Standard deviations are given in parentheses
Table 2
Analyte
Material
Slope/mA mM
Guanine
Bare GC ENP ENR Bare GC ENP ENR
0.13 (0.003) 3.03 (0.300) 1.18 (0.046) 0.51 (0.048) 1.39 (0.544) 2.02 (0.212)
Adenine
1
R2 0.9984 0.9620 0.9940 0.9654 0.5787 0.9576
deprotonated radical A+_ which generated a second oxidation peak.26 Despite that, the ENP surface still provided the largest oxidation current as compared to the other two surfaces, which coincided with the trend of oxidation current for most of the compounds previously discussed. Similar to the biomarkers, the graphene materials showed heightened sensitivities and oxidation currents when probed with the two DNA bases in comparison with bare GC. Peak potentials and detection limits of studied biomarkers at different surfaces as reported in the literature are stated in Table S1† (ESI) for comparison.
4. Conclusions The electrochemical performances of electrochemically reduced graphene, specically nanoplatelets and nanoribbons, were compared against bare GC using four biomarkers, mainly uric acid, ascorbic acid, dopamine, NADH, and two DNA bases, specically guanine and adenine. Effects of the materials on the oxidation potentials of the analytes were examined using CV and it was established that graphene-coated electrodes led to lower oxidation potentials for all the biomarkers and DNA bases. This indicated that a smaller amount of potential was required for oxidation to occur in the presence of graphene materials. Parameters such as sensitivity of the surfaces were analysed based on results from DPV and calibration plots. It was
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 07 January 2014. Downloaded by Aston University on 25/01/2014 13:29:32.
Paper
found that the graphene materials displayed better sensitivities for the detection of the biomarkers and DNA bases examined other than NADH. The inconsistency on NADH was justied by the possibility of the passivation effect. A comparison between the two graphene materials showed that, in general, the ENP surface brought about lower oxidation potentials for all biomarkers and DNA bases and higher sensitivities for oxidation of ascorbic acid and guanine than the ENR surface. On the other hand, the ENR surface provided better sensitivities for uric acid, dopamine and adenine. Though it was hypothesized that the ENP surface will generally provide a better sensitivity due to an extensive amount of exposed edgelike planes, the nature of the biomarkers and DNA bases may contribute to the sensitivity outcomes as well. In the case of selectivity, graphene materials evidently exhibited higher selectivities as compared to the bare GC. Between the two graphene materials, it was more advantageous to utilise the ENP surface to achieve lower electrochemical oxidation potentials. However, there was no clear advantage of using a specic graphene material as they displayed different sensitivities for different compounds, despite the ENP surface exhibiting a larger oxidation current for all the compounds studied. In conclusion, the graphene materials have shown better abilities in detecting biomarkers and DNA bases than the bare GC. Further investigations can be carried out with the two other DNA bases, thymine and cytosine, as well as other substantial biomarkers to differentiate the performance of the ENP and ENR surfaces more clearly to rationalise how the difference in graphene structures can affect their electrochemical properties.
Notes and references 1 E. Paleˇ cek and M. Fojta, Anal. Chem., 2001, 73, 74A. 2 P. Gai, H. Zhang, Y. Zhang, W. Liu, G. Zhu, X. Zhang and J. Chen, J. Mater. Chem. B, 2013, 1, 2742. 3 E. Paleˇ cek, Electroanalysis, 1996, 8, 7. 4 E. Paleˇ cek, Nature, 1958, 4751, 656. 5 E. Paleˇ cek, Electroanalysis, 2009, 21, 239. 6 E. P. Randviir and C. E. Banks, RSC Adv., 2012, 2, 5800. 7 M. H. Yang, A. Javadi, H. Li and S. Q. Gong, Biosens. Bioelectron., 2010, 26, 560.
This journal is © The Royal Society of Chemistry 2014
Analyst
8 A. K. Geim and K. S. Novoselov, Nat. Mater., 2007, 6, 183. 9 D. A. C. Brownson, D. K. Kampouris and C. E. Banks, Chem. Soc. Rev., 2012, 41, 6944. 10 R. R. Nair, P. Blake, A. N. Grigorenko, K. S. Novoselov, T. J. Booth, T. Stauber, N. M. R. Peres and A. K. Geim, Science, 2008, 320, 1308. 11 J. S. Bunch, A. M. Van der Zande, S. S. Verbridge, I. W. Frank, D. M. Tanenbaum, J. M. Parpia, H. G. Craighead and P. L. McEuen, Science, 2007, 315, 490. 12 K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666. 13 C. N. R. Rao, A. K. Sood, K. S. Subrahmanyam and A. Govindaraj, Angew. Chem., Int. Ed., 2009, 48, 7752. 14 D. V. Kosynkin, A. L. Higginbotham, A. Sinitskii, J. R. Lomeda, A. Dimiev, B. K. Price and J. M. Tour, Nature, 2009, 458, 872. 15 C. K. Chua, Z. Sofer and M. Pumera, Chem.–Asian J., 2012, 7, 2367. 16 T. Kim, S. Lim, K. Kwon, S.-H. Hong, W. Qiao, C. K. Rhee, S.-H. Yoon and I. Mochida, Langmuir, 2006, 22, 9086. 17 T. J. Davies, M. E. Hyde and R. G. Compton, Angew. Chem., Int. Ed., 2005, 44, 5121. 18 (a) A. F. Holloway, K. Toghill, G. G. Wildgoose, R. G. Compton, M. A. H. Ward, G. Tobias, S. A. Llewellyn, B. Ballesteros, M. L. H. Green and A. Crossley, J. Phys. Chem. C, 2008, 112, 10389; (b) A. Ambrosi and M. Pumera, Chem. Asian J., 2010, 5, 266. 19 D. K. Kampouris and C. E. Banks, Chem. Commun., 2010, 46, 8986. 20 T. Kolodiazhnyi and M. Pumera, Small, 2008, 4, 1476. 21 A. Ambrosi, A. Bonanni, Z. Sofer, J. S. Cross and M. Pumera, Chem.–Eur. J., 2011, 17, 10763. 22 M. J. Griffiths, J. Biol. Chem., 1952, 197, 399. 23 C. L. Scott and M. Pumera, Electrochem. Commun., 2011, 13, 426. 24 M. Pumera, R. Scipioni, H. Iwai, T. Ohno, Y. Miyahara and M. Boero, Chem.–Eur. J., 2009, 15, 10851. 25 M. Pumera, A. Ambrosi and E. L. K. Chng, Chem. Sci., 2012, 3, 3347. 26 K. Kobayashi, J. Phys. Chem. B, 2010, 114, 5600.
Analyst
