M. F. Hossain and
J. Y. Park*
Department of Electronic Engineering, Micro/Nano Devices & Packaging Lab, Kwangwoon University, Seoul, Republic of Korea. E-mail: jaepark@kw.ac.kr; Fax: +82-2-942-1502; Tel: +82-2-940-5113
First published on 20th July 2016
The progress of nanotechnology has encouraged scientists to continuously pursue new electrode materials for forming improved electrochemical platforms for sensing. In this study, an environmentally friendly reducing agent and optimum time were utilized to synthesize solvothermal-assisted reduced graphene oxide (TRGO) that can be easily exfoliated in a coating solution and substituted on the microelectrode surface. Series connected distributed sensing electrodes were fabricated and patterned with TRGO decorated platinum nanoparticles (PtNPs), chitosan–enzyme composites, and Nafion, which were integrated onto the modified surface for effective glucose detection. The developed biosensor demonstrated good current response to glucose with a high sensitivity, short response time, and low detection limit of 41.18 μA mM−1 cm−2, 5 s, and 0.0019 mM, respectively. The interference phenomena, reproducibility and the durability of the biosensor, were also examined. In addition, glucose levels in human urine were tested for a practical assessment of the proposed biosensor, and the results indicate that the sensor had superior urine glucose recognition. These results reveal that this noble nanostructured electrode with high surface area and electrocatalytic activity offers great promise for use in enzymatic biosensor applications.
Precious metal (Pt, Au, Ag, and Pd) nanoparticles have been used in electrodes owing to their high catalytic activities, excellent conductivity, large surface area and high enzyme loading capability.5,6 Among noble nanoparticles, Pt nanoparticles (PtNPs) have been widely used in the enzymatic or nonenzymatic detection of biomolecules owing to their excellent catalytic activity for the redox reaction of H2O2.7,8 The excellent catalytic activity of Pt nanoparticles is attributed to their large active area and beneficial lattice plane, which supply very high numbers of effective centers for electrochemical reactions and facilitate the electron transfer reaction of H2O2.9 The dispersion of NPs on supporting materials, such as CNT, graphene, and porous carbon substrate, dramatically improves their electrochemical activity.10,11
Graphene and chemically exfoliated graphene sheets are excellent supporting materials owing to their high electrical conductivity, high potential windows, large active surface area, and excellent mechanical properties compared to other carbon based materials.12 Reduced graphene oxide (RGO), a 2D allotrope of carbon, is chemically, hydrothermally/solvothermally, and thermally exfoliated from graphite oxide (GO). Pristine graphene is advantageous owing to its higher electric conductivity compared to RGO, whereas the oxygen functional groups in RGO are more beneficial for biosensors because they enable loading of enzymes and other catalytically active molecules. The OH groups on RGO may act as a mediator for electron transfer between the enzyme and the electrode surface.13 Moreover, a number of conceptually different methods, such as chemical, thermal, electrochemical, photocatalytic, and hydrothermal/solvothermal reductions, have been reported for the preparation of RGO.14 Solvothermal reduction of graphite oxide is a popular technique because of its simple setup, good scalability, and recovery of π conjugation networks at high temperature and pressure. Furthermore, hydrothermal pathways have been engaged for notable transformations of carbohydrate molecules to form homogeneous carbon nanospheres15,16 and nanotubes.17 However, several problems arise from this process, in which the oxide functional groups in RGO cannot be remarkably eliminated.18 Furthermore, RGO forms a hydrogel at high temperature and pressure, and thus cannot be used in microsystems. When solvothermal synthesized RGO gel dries at ambient conditions, it cannot be readily delaminated in polar or non-polar solvents. To address these issues, optimum times and non-toxic reducing agents are utilized for the feasible reduction of solvothermal reduced graphene oxide, denoted as TRGO.
Gold thin film based substrates have been widely used as micro electromechanical system (MEMS) biosensor electrodes for the easy patterning of microdevices on the substrate. Moreover, substrate electrodes should have a wide inert potential window during the electrochemical redox reaction to obtain good performance from the electrochemical matrix; however, Au thin film electrodes show a narrow inert potential window during different redox reactions compared to other conductive substrates.19 In general, carbon materials coated with thin gold electrodes show improvements such as a wide inert potential window, high specific capacitance, and high catalytic activity.20 Among carbon materials, RGO is a good candidate for these purposes. In contrast, RGO coated gold thin film is not easy to pattern for fabricating MEMS-based devices on chips. Concerning this issue, a novel electrochemical technique was investigated to form patterns on the substrate. In this process, RGO was selectively etched by the electrochemical technique to be shorted among the three-electrode system.
Several studies have been reported for one working electrode on a single chip;21,22 however, double working electrodes on the same chip have been recently demonstrated for enzymatic glucose sensors. The study indicates that contraction in the geometric area of the working electrode generally leads to weaker signal density that approaches noise levels. Moreover, uniformity of enzyme immobilization is suitable in a small geometric area. As a result, the electrochemical properties of the total system increase owing to the potential uniformity of materials in a small area as well as the series connection within the electrode system.
In this study, a double working electrode system was designed and fabricated. TRGO was synthesized, and its suspension was dropped on the electrodes. Further, it was selectively etched from the counter and reference electrodes by an electrochemical method to be shorted between the working electrodes. PtNPs were then decorated on the working and counter electrodes. Furthermore, GOx–chit and Nafion were dropped on the PtNP/TRGO modified surface for the production of a glucose sensor. The developed TRGO and the as-prepared biosensor were investigated by field-emission scanning electron microscopy (FESEM), energy dispersive X-ray spectroscopy (EDX), X-ray photoelectron spectroscopy (XPS), Raman spectroscopy, and electrochemical techniques.
Afterward, TRGO was electrochemically etched away from the counter and reference electrodes. This was easily achieved by cyclic voltammetry (CV). Different potential windows were applied for etching TRGO without damaging the working electrodes and creating higher resistance between the three electrode gaps. The method is described in the subsequent section. PtNP was consecutively deposited on top of the working electrodes and counter electrode by CV in precursor solution containing 2.5 mM H2PtCl6 and 0.1 M KCl for 10 and 14 cycles, respectively. Then, the modified electrodes were cleaned with sufficient DI water and dried with N2 gas.
The Ag/AgCl reference electrode was produced by a chronopotentiometry technique at a constant current of 0.00461 A mm−2 in Ag metal solution. Furthermore, the reference electrode was dipped in 50 mM FeCl3 solution for 1 min to form the Ag/AgCl pattern. 6 μL of GOx–chitosan solution was cast on the working electrodes. Afterward, the glucose sensor was stored at 4 °C in a refrigerator for 1 day. Finally, 5 μL of Nafion, ethanol, and DI water (1:
6
:
1) solution were cast on the entire electrode and allowed to dry for 10 min at ambient temperature. The as-fabricated glucose sensor was maintained at 4 °C in a fridge at least 1 day before use.
![]() | ||
Fig. 2 (A) Digital image of the developed biosensor; FESEM micrographs of (B) TRGO and (C) TRGO/PtNP, and (D) the corresponding EDX pattern of TRGO/PtNP. |
For this reason, the PtNP was well assembled on the TRGO network, forming a mutually perforating chain for a good transportation route of electron transfer from the matrix surfaces to the electrode. Moreover, a rough electrode surface is suitable for enzyme stability, enhancing the enzyme loading.
Elemental composition analysis of the PtNP decorated TRGO electrode was performed by EDX, as shown in Fig. 2(D). The peaks of C, O, and PtNPs are seen in the EDX analyzed data. The atomic percentages of C, O, and PtNP were 85.78%, 11.04%, and 3.18%, respectively, indicating that PtNP was successfully deposited by the electrochemical method under the given conditions.
The typical AFM image and height profile of TRGO are shown in Fig. 3. The thickness of the TRGO sheet is in the range of 1.1 to 10 nm in the cross sectional view across the flat area of the TRGO sheet, and the length is in the range of several micrometers. The observed thickness of the TRGO sheet reveals that the monolayer and a few layers were present in the developed TRGO.
XPS was carried out to analyze the oxygen functional groups in GO and TRGO. Fig. 4(A) and (B) show the C1s spectra of GO, indicating a considerable degree of oxidation with other components, corresponding to different functional groups, i.e., C in the non-oxygenated ring in CC (284.6 eV), C in the C–O bonds of oxygenated rings (286.6 eV), C in the C
O bonds (288.2 eV), and C in the O–C
O bonds (289.1 eV).24 After the reduction of GO, a peak (285.8 eV) was observed in the C1s spectra, including the C–C bonds of non-oxygenated rings.25 This result reveals that the oxygen functional groups were significantly reduced during the hydrothermal reduction, successfully forming RGO.
Raman spectroscopy manifests the structural changes appearing during the reduction of GO to TRGO. The Raman spectra of GO and TRGO are shown in Fig. 5. The two prominent peaks, named the D band and G band, of GO appeared at 1350 and 1590 cm−1, respectively. Similarly, the peaks of TRGO appeared at 1349 and 1591 cm−1. The G band is involved in the in-plane vibration of sp2 bonded carbon atoms, whereas the D band is related to the vibrations of the carbon atoms with the sp3 electronic arrangement of disordered graphene. The ID/IG ratios of TRGO and GO are 1.28 and 0.89, respectively. This result reveals that reduction of GO occurred during the hydrothermal process.26 The number of layers of graphene material sheets were determined by the 2D band in the Raman spectrum. The 2D bands of GO and TRGO were located at ∼2693 and 2699 cm−1, respectively, indicating the presence of a few layers in the GO and TRGO sheets.27
CVs of the working electrodes (single- and double-electrode-based biosensors) were recorded in PBS, as shown in Fig. 6(C)(a) and (b). A peak was observed at ∼0.55 V in the cathodic curve. The peak at 0.55 V is because of the reduction of monolayer gold oxide.30 The peak intensity of the double working electrode-based sensor was nearly twice that of the single one, indicating that the two separate working electrodes were joined and acted as a series circuit.
The electroactive surface areas (ESA) of the TRGO modified electrode and the TRGO/PtNP modified electrode were evaluated by integrating the areas of the anodic peaks in the hydrogen region in 0.5 M H2SO4 solution at a scan rate of 50 mV s−1. The ESA was calculated by dividing the estimated charge by the charge per actual active area of the catalyst (0.21 mC cm−2). The geometric surface area of the pristine electrode was 0.0706 cm2, and the surface area of the pristine electrode was estimated to be 0.12 cm2. The ESAs for the TRGO modified electrode and the TRGO/PtNP electrode were estimated as 0.35 and 15.36 cm2, respectively.
The current responses of the biosensor at 0.3, 0.4, and 0.5 V upon successive additions of different concentrations of glucose in PBS solution, and the corresponding calibration plots (the inset of the figure), are shown in Fig. 7(B)(a)–(c). The current increased instantly with increasing glucose concentration and rapidly reached the steady state. The response time of the biosensor was as fast as 5 s, exhibiting the observation range of 0.5 to 12 mM glucose with a sensitivity of 28.37, 41.18, and 34.22 μA mM−1 cm−2 at 0.3 V (a), 0.4 V (b), and 0.5 V (c), respectively. From these responses, the highest response was obtained at 0.4 V; the linear range was 0.00625 to 10 mM. The sensitivity and linear range of the as-fabricated sensor are higher than those of previously published reports, including a sensitivity of 14.41 μA mM−1 cm−2 and a linear range of 0 to 0.7 mM for GC/CNTs–CdTe/GOx/Nafion,32 38 μA mM−1 cm−2 and 0.05 to 0.5 mM for ITO/CdS–Au–Ch/GOx,33 7.4 μA mM−1 cm−2 and 0.5 to 10 mM for Al2O3/Pt/PPy/GOx,34 and 39.63 μA mM−1 cm−2 and 0.01 to 4.5 mM for GC/PSS/PAni/Pt-DENs/GOx/Pt-DENs.35 Various concentrations of glucose were detected by the fabricated biosensor at 0.4 V in PBS, as shown in Fig. 8(A). The detection limit of the biosensor was determined to be 0.0019 mM (a signal three times above noise). This limit of detection is much smaller than those previously reported, including detection limits of 0.1 mM for Pt/TiO2/RGO/PtNPs/GOx,36 0.18 mM for Au/Gr/AuNPs/Ch–GOx,2 0.01 mM for PET/Ti/Au/SDS-MWCNT/PDDA/GOx/PDDA,37 and 0.017 mM for GC/Ag-Pdop@CNT/GOx/Nafion.38
The current responses of possible interfering species, such as AA and UA, were studied by amperometry. Humans cannot synthesize ascorbic acid; they obtain it through diet or from supplements. Ascorbic acid excretion in human urine is around 0 to 1.12 mM per day.39 Contrastingly, uric acid excretion in urine is around 0.32 to 2.9 mM per day.40,41 0.2 mM UA and 0.1 mM AA were used in this study. These interfering anodic currents were lower than that of 7.12 μA for 1.58 mM glucose. The anodic responses are negligible for 0.1 mM AA and 0.2 mM UA, as shown in Fig. 8(B). In future, the fabricated glucose sensor will be improved for tolerance of higher concentrations of interference species. The stability of the fabricated biosensor was also observed for a period of four weeks. The sensitivity of the biosensor decreased only by 15.2% during that time, indicating that the sensor has long-term stability. The repeatability and reproducibility of the fabricated sensor were verified by amperometric response in PBS up to 4.5 mM glucose. The glucose sensor showed a relative standard deviation (RSD) of 3.15% for five repetitive measurements. This indicates that the fabricated sensor has good repeatability. In addition, four individual sensors were fabricated using the same process that was used for the developed glucose sensor. All the different biosensor electrodes showed an RSD of 6.4% in their amperometric responses, suggesting the appreciable reproducibility of the as-fabricated glucose sensor.
The amperometric response of the biosensor upon continuous addition of different concentrations of human urine glucose is shown in Fig. 9, indicating that the biosensor responded rapidly to the addition of small amounts of urine glucose and reached the steady state (95% of the signal) within 5 s. The biosensor showed good linearity at the lower range detection of glucose in human urine. Therefore, the fabricated biosensor is suitable for detecting lower concentrations of glucose in human urine. In addition, 100 μL of the urine sample were added to 20 mL of 0.05 M PBS solution. Then, a certain concentration of standard glucose sample was injected into the urine and PBS diluted solution so that the concentration of glucose in the urine could be detected from the calibration curve, as shown in Table 1. The recovery of spiked glucose in the urine sample was observed in the range of 96.1 to 99.2%, and the RSD ranged from 3.8% to 6.1%; this reveals that the fabricated glucose sensor has good ability for glucose detection in human urine.
![]() | ||
Fig. 9 Amperometric response of the fabricated glucose sensor to various urine glucose concentrations (12.5, 25, 50, 100, and 750 μM) in PBS (pH 7.4) at an applied potential of 0.4 V. |
Sample | Spiked (μM) | Detected (μM) | Recovery (%) | RSD (%) |
---|---|---|---|---|
1 | 25 | 24.8 | 99.2 | 3.8 |
2 | 250 | 240.2 | 96.1 | 4.3 |
3 | 1500 | 1470 | 98.0 | 4.9 |
4 | 2500 | 2438 | 97.5 | 5.3 |
5 | 3000 | 2923 | 97.4 | 6.1 |
This journal is © The Royal Society of Chemistry 2016 |