Facile synthesis of graphene oxide/Fe3O4 nanocomposite for electrochemical sensing on determination of dopamine

Abstract Dopamine concentration abnormalities in the body can cause various disorders and diseases such as Parkinson's, Tourette's syndrome, and depression. In this study, graphene oxide (GO) was combined with Fe3O4 to sensitively and selectively detect dopamine. The performance was evaluated by cyclic voltammetry (CV) and differential pulse voltammetry (DPV) methods. The results of testing with CV on the solution [Fe(CN)6] showed that a modification with GO gave a maximum effective surface area value of 0.0127 cm2, proving that GO can increase the effective area and conductivity of the sensor. DPV testing shows that dopamine detection using GO/Fe3O4 has a linear range at a concentration of 1–10 μM with a detection limit of 0.48 μM and a quantification limit of 1.6 μM. GO/Fe3O4 also shows good selectivity where the peak current is separated by 0.245 V with ascorbic acid, which is the closest interference compound. Graphical Abstract


Introduction
Dopamine is a neurotransmitter that belongs to the monoamine and catecholamine groups, produced by the adrenal glands and other parts of the brain. Dopamine, as a neurotransmitter, is very important to help nerve cells to send messages to each other. Dopamine functions in excitatory and inhibitory classes to activate postsynaptic neurotransmitter receptors or reverse them [1]. Dopamine's low concentration level is associated with several serious disorders, such as attention deficit hyperactivity disorder (ADHD), Parkinson's disease, and schizophrenia [2][3][4][5][6], while a high level of concentration leads to cardiotoxicities, such as hypertension, heart failure, and drug addiction [7]. Therefore, dopamine measurement is critical to early diagnose these abnormalities. Dopamine concentration is commonly measured through laboratory analysis techniques, such as conventional chromatography, electrophoresis, enzymatic assay, spectrophotometry, and liquid chromatography/tandem mass spectrometry (LC-MS/MS) methods [8,9]. However, there are drawbacks to these techniques, such as complex preparation, expert personnel, long testing times and expensive cost.
Electrochemical-based sensors are a good candidate for dopamine determination because dopamine has a relatively good electrochemical activity and, importantly, its robustness, high sensitivity, and rapid analysis. Electrochemical measurements are performed using electrodes, and the current or potential responses toward the analytes will be observed. However, the analytical performance and electrode selectivity are still significant problems that need to be addressed in developing the electrochemical sensor. Many researchers overcome this by using electrode surface modification with nanomaterials to increase the detection efficiency due to their large surface-to-volume ratio, improved electrical conductivity, and fast heterogeneous electron transfer rate [10,11]. Carbon-based nanomaterials have become the material of interest for researchers, especially in biosensors [12][13][14][15][16][17][18][19][20][21]. Graphene oxide (GO), as one of the variations of graphene, is graphene decorated by a hydroxyl group and epoxy (1,2ether) on the surface [22]. GO's high active surface area and reactivity facilitate convenient compositing with other desired specific materials, such as metals, metal oxides, and some complex oxides. Metal oxide nanoparticles have been renowned for having many applications in biomedical since the 1970s [23]. One of the most used materials in this metal oxide is magnetic iron oxide, Fe 3 O 4 [24,25]. It has properties such as non-toxic, biocompatible, and highly accumulated in target tissues or specific organs. Magnetic nanoparticle materials are also usually utilized in biomolecule separation, hypothermia, drug delivery system, and magnetic resonance imaging.
GO-based nanocomposites have been used for dopamine biosensors, such as nanostructured material modified GO and polymer-modified GO for label-free detection [26,27]. Label-free detection helps solve challenges in enzyme-based biosensors, particularly in screening highly active enzymes efficiently and with high sensitivity [8]. The novelty of this study relied on synthesis GO-based nanocomposites with Fe 3 O 4 nanoparticles for dopamine sensor. Among metal oxide nanoparticles, Fe 3 O 4 nanoparticles has display strong superparamagnetism, low toxicity, good biocompatibility and high catalytic activity. In addition, compared to the previous method of syntesis of graphene based nanocomposites with Fe 3 O 4 nanoparticles using hydrothermal [28] and cold quenching in liquid nitrogen [29], GO/Fe 3 O 4 nanocomposites in current work was prepared using chemical coprecipitation technique which is facile, cheap, the most environentally friendly and promising for large scale production with the size and homogenity of the composites can be controlled. The combination of Fe 3 O 4 and GO as nanocomposites as an active material is promising for developing dopamine electrochemical sensors. Electrochemical analyses evaluated the dopamine response with GO/Fe 3 O 4 nanocomposite to obtain its detection limit and selectivity against several interferences. This synergistic effect resulted in a great dopamine determination performance while showing excellent selectivity against glucose, ascorbic acid, and uric acid as interfering compounds. Our relatively simple nanocomposites synthesis and its great sensing performance show a promising label-free biosensor development for dopamine detection.

Materials characterization and instrument
Morphology characterization and composition analysis was performed using scanning electron microscopy-energy dispersive x-ray spectroscopy (SEM-EDX, Hitachi SU3500, Hitachi, Japan). X-ray diffraction (XRD) was performed using x-ray diffractometer (Bruker D8 Advance, Bruker, Germany). The absorbance peak was obtained using UV-Vis spectrophotometer (Labtron LUS B-12, Labtron, United Kingdom). Electrochemical characteristics of sensor material were performed using Ana Pot EIS (ZP, Norway) with a glassy carbon electrode (GCE, 7 mm in diameter, 38.49 mm 2 in area) modified by GO/Fe 3 O 4 nanocomposite as a working electrode, along with a platinum wire for the counter electrode and silver/silver chloride (Ag/AgCl) for the reference electrode.

Synthesis of graphene oxide (GO)
Graphene oxide was synthesized from graphite powder using the modified hummer method from previous work with different stiring time duration and ratio of pure graphite with the oxidator [30,31]. In this method, 23.33 mL H 2 SO 4 was prepared in temperature conditions under 20 C using an ice bath. Subsequently, 0.5 gr NaNO 3 , 0.5 gr graphite powder, and 2.5 gr KMnO 4 were added to the solution, and then the solution was stirred for 2 hrs on a hot plate. After that, with a controlled temperature of under 40 C, the solution was stirred for 16 hrs. The color of the mixture would change into light brown, and it would become a paste. Next, 41.6 mL of deionized (DI) water was added to the mixture and optimized for 30 min, followed by gently adding H 2 O 2 to stop the reaction. The mixture color becomes pale yellow which shows a high oxidation level.
Afterward, 133.33 mL DI water was poured into it. The solution was washed using HCl 1 M ($20 mL) and treated with centrifugation for excess SO 4 2À ion removal. BaCl was used to define SO 4 2À ion and produced as supernatant. The supernatant was removed, and the residue product was neutralized by pouring DI water until the pH solution became 5. Centrifugation treatment was performed subsequently, followed by drying the residue product at 60 C for 24 hrs. Finally, the dried product was mashed using mortar to get powders form of graphite oxide. For obtaining GO powder, peeling treatment to get graphite sheets using ultrasonication was needed. Before that, DI water was added to the graphite oxide powder with a 1 mg/mL concentration. After sonication, the solution was filtered and dried at 60 C for 24 hrs. The final step was mashing the dried material using mortar to obtain GO powder.

Synthesis of GO/Fe 3 O 4 nanocomposite
The GO modification procedure was described in previous work with some modifications [32]. First step was dissolving 0.2 gr FeCl 3 .6H 2 O and 0.4 FeCl 2 .4H 2 O in 50 mL deionized water. Then 0.5 gr GO powder was added and sonicated for 30 min. Solution pH was controlled until it became 10 by adding 2 mL ammonium solution (25%) under strong stirring for 5 hrs. Afterward, the supernatant was removed, and the black precipitation was washed before the drying treatment at 80 C for 24 hrs. Mortar was then used to acquire GO/ Fe 3 O 4 powder.

Working electrode modification
We used the drop-casting method to modify the working electrode surface by dropping a certain amount of solution followed by a drying process at room temperature until a thin deposition of desired material was formed on the surface. Due to the simple process, the final amount drop-casted onto the electrode surface is the only optimization parameter [13,33]. Accordingly, we prepared 1 mg of material (GO or GO/Fe 3 O 4 ) dissolved in 1 mL of DI water. The solution was sonicated through an ultrasonic bath for 30 min until homogenized. Before dropcasted, the GCE surface was initially cleaned by polishing it on a cleaner sheet with the addition of alumina. Then, a modification was carried out by dropping each 10 mL of material onto the GCE surface and drying it at room temperature for 4 hrs. Next, Nafion solution as a binder which can facilitate and stabilize the present of the nanocomposites on the GCE surface was mixed with DI water in a 1:5 volume ratio before dropping 5 mL of the mixture onto the GCE surface. Finally, modified GCE was left alone for 1 h before use.

Electroanalysis of dopamine
Electrochemical performances were examined using cyclic voltammetry (CV) and differential pulse voltammetry (DPV). CV measurement was performed for 10 cycles. The GCE with modification material is characterized by CV from À0.4 V to 1 V vs. Ag/ AgCl in 10 mM [Fe(CN) 6 ] 3À/4À (in 10 mM PBS) with a scan rate of 0.05 V/s. We measured the electrochemical performance of GCE with GO/Fe 3 O 4 modification, GO modification, and without modification (bare electrode). Optimization experiments were performed to get the best performance of the modified sensor. Electroanalysis of dopamine in PBS solution was performed using CV with the same measurement parameters. Static characteristics of sensors such as sensitivity, linearity, and limit of detection (LoD) were analyzed by DPV measurement with a potential range of 0.1 V to 0.5 V vs. Ag/ AgCl with a scan rate of 0.05 V/s. Finally, we have tested our sensor for selectivity test on glucose, urea, and ascorbic acid as interference analytes. Figure 1 shows the illustration of electrochemical analysis steps.

Material characterization
The optical properties of the GO and GO/Fe 3 O 4 using UV-Vis analysis are shown in Figure 2. It can be seen that GO produces an absorbance peak at a wavelength of 230 nm resulting from the p-p Ã transition of the C ¼ C aromatic bond, as well as the C ¼ O bond, which is at a range of 270-300 nm. For GO/Fe 3 O 4 , the peaks shift to the higher wavelengths at 239 nm due to the existence of Fe 3 O 4 nanoparticles, which give rise to the restoration of electronic conjugation for the carbon bond of graphene sheet [34].
The SEM images of GO and GO/Fe 3 O 4 are depicted in Figures 3A-B, with the magnification set at 40,000x. The result shows that the typical morphology of the graphene oxide layers is obtained as a pile of thin layers with wrinkled characteristics ( Figure 3A). Figure 3B shows the Fe 3 O 4 distribution on the GO surface. The deposition of Fe 3 O 4 nanoparticles on GO surface is observed due to the interaction of magnetic forces between them. The magnetic forces among Fe 3 O 4 nanoparticles and the surface energy effect led the particles to aggregate. The elemental analysis resulting from EDX analysis ( Figure 4 and Table 1) demonstrates that the content of the compound is a carbon (C, 67.86% atom) and followed by oxygen (O, 32.14% atom) with an approximately 2:1 ratio between C and O atom ( Figure 4A). The C and O atom ratio indicates a good composition result since the lower oxygen defects on the graphene surface, the better the chemical and electrochemical properties. In the case of the nanocomposite, consisting of C (62.16% atom), O (32.25% atom), and iron (Fe, 5.59% atom), no other elements are detected. This result indicates that a high purity of synthesized GO/Fe 3 O 4 was successfully obtained.
The crystal structure of the GO and GO/Fe 3 O 4 characterized by XRD is depicted in Figure 5. The obtained GO shows two distinct diffraction peaks at 2h around 10.0 and 42.3 due to the reflection of (001) and (100) plane [35], with the peak of graphite at 26 does not appear to reveal the formation of GO with oxygen-containing functional groups [36]. After decorating with Fe 3 O 4 , the peak of GO is still present with a slight shift and patterns of Raman spectroscopy is a potential technique to characterize the significant structural changes in carbonaceous materials. The Raman spectra of GO and GO/Fe 3 O 4 nanocomposites are given in Figure 6. The two prominent peaks of GO appear at 1351 cm À1 and 1598 cm À1 , corresponding to the D and G bands, respectively, with the intensity ratio of I D /I G is 0.92. The Raman spectra of GO/Fe 3 O 4 show the peaks of D and G bands at 1348 cm À1 and    1598 cm À1 , respectively, with the intensity ratio of I D /I G is 1.14, which is higher than that of GO. The increase in the peak intensity is evident in the successful functionalization of Fe 3 O 4 particles on GO sheet [29].

Go/Fe 3 O 4 nanocomposites modified electrode electrochemical characterization
The electrochemical characteristics of the electrodes were investigated to determine the phenomenon of electron transfer that occurs using a solution of [Fe(CN) 6 ] 3À/4À as shown in Figure 7. Different electrode surfaces will produce different anodic and cathodic responses. Cyclic voltammetry experiments were carried out at a working voltage range of À0.4 V to 1 V with a scan rate of 50 mV/s in 10 mM [Fe(CN) 6 ] 3À/4À in 10 mM PBS (pH 7.4). Reduction and oxidation reactions occur due to the free electron transfer by electrolyte when potential is gradually changed. The working potential will be scanned from À0.4 V as a start point to the positive direction causing an anodic reaction on the electrode and resulting oxidation peak current. Then, the potential will be turned back towards À0.4 V, resulting in a reduction in peak current due to cathodic reaction. The redox reaction equation is shown in Eq. 1-2.
Eq. 1 shows the oxidation reaction on the electrode surface. When working potential reaches a sufficient value to oxidize [Fe(CN) 6 ] 4À , the anodic current density will significantly increase. The current escalation occurs until the current peak has saturated caused by the [Fe(CN) 6 ] 4À concentration drop on the electrode surface, while the back potential sweeping towards À0.4 V will initiate a cathodic reaction due to the reduction of [Fe(CN) 6 ] 4À to [Fe(CN) 6 ] 3À as written in Eq. 2, which involves a free electron. Figure 7 shows the redox responses for [Fe(CN) 6 ] 3À/4À using bare GCE and modified GCE. The modified GCE resulted in detected current enhancement compared to the bare GCE current. Furthermore, electrode modification using GO generates the oxidation peak (Ipa) and reduction (Ipc) currents greater than the modification using GO/Fe 3 O 4 , while the magnitude of the current value produced by the GO/Fe 3 O 4 modification electrode significantly increases compared to two other types of working electrodes. The high Ipa and Ipc on GO electrode are induced by the increased surface area by the morphological characteristics of the GO surface. The     Figure 8A. Measurement using a bare electrode did not generate any redox peak. The GO modified electrode produces a dopamine oxidation peak current density of 16.152 lA/cm 2 on a working potential of 0.29216 V. This indicates that improving the working electrode surface area by GO introduces more electron transfer. In contrast, the peak improvement is relatively low compared with the electrochemical response of the GO/Fe 3 O 4 -modified electrode. The GO/Fe 3 O 4 -modified electrode produces almost 10 times the oxidation peak current density, which is 153.09 lA/cm 2 , with the shift of the working electrode to the positive direction. It shows that adding nanohybrid magnetic graphene (GO/Fe 3 O 4 nanocomposite) could create a more conductive electron transfer path.
The GO/Fe 3 O 4 /GCE has higher electrocatalytic activity towards the reduction and oxidation of dopamine due to the synergetic effect between Fe 3 O 4 nanoparticles and GO as the supporting material, which could act as an effective electron promotor and improve the dopamine catalytic activity. On the other hand, the shifting of the working potential is induced by the Fe 3 O 4 nanoparticles that tend to increase the internal resistance of the electrode surface. Therefore, the difference between oxidation and reduction working potential (DEp) is increased because of the greater potential needed to oxidize or reduce the analyte. Moreover, the redox reaction on GO/Fe 3 O 4 /GCE is more reversible than using GO/GCE, with the anodic and cathodic peak ratio being 1.26 (1 % reversible) and 1.93, respectively. It confirms the quasi-reversibility of the dopamine redox reaction.
The reactions on the GO/Fe 3 O 4 /GCE interface of the dopamine detection are illustrated in Figure 9. At first, the Fe 3 O 4 nanoparticles are reduced directly on the electrode surface (Eq. 4), producing Fe 2þ ions. With increased working potential, oxidation occurs (Eq. 5) and induces the electron transfer used to deprotonate the Dopamine (DA) into the Dopamine-o-Quinone (DQ), as shown in Eq. 6. While the reverse-sweep potential takes place, Fe 3þ ions are reduced (Eq. 7) and followed by the reattachment of the hydrogen ions by DQ, resulting in DA molecules back (Eq. 8).

Scan rate variation
Electrochemical measurements towards scanning rate variation were conducted to determine the electrochemical reaction characteristic of the dopamine detection. It can be concluded by observing the relation of peak current density toward the square root of scan rate. As shown in Figure 8B, the measurement was performed by the scanning rate of 25, 35, 50, 75, 100, and 150 mV/s in 10 mM dopamine and 0.1 mM pH 7.4 PBS. CV measurement shows that both the peak current of reduction and oxidation is proportional linearly to the square root of scan rate by equation of Ipa ox ¼ 42.366 þ 4.2752v (n ¼ 6, R ¼ 0.9698) and Ipa red ¼ À28.85 À 5.4626v (n ¼ 6, R ¼ 0.9845). It indicates that the oxidation reaction towards dopamine on the electrode surface is a diffusion-controlled process.

Sensitivity, selectivity, and stability evaluation of GO/Fe 3 O 4 /GCE for dopamine detection
The GO/Fe 3 O 4 -modified electrode was tested towards dopamine using Differential Pulse Voltammetry (DPV) by the potential range of 0.05-0.5 V and a scan rate of 50 mV/s. As shown in Figure 8C, by varying the dopamine concentration from 0-20 mM, the sensors have increasing responses toward higher dopamine concentration. The linear range conforms with the detection requirement considering that dopamine concentration inside the human body is in the range of 0.3-3 mM in urine, which is within our measurement, although a higher sensitivity is needed for a range of 0-0.25 nM in blood [38]. Subsequently, the gradient of the linear equation represents the sensitivity value of the sensor. Limit of detection (LoD) is a sensor performance parameter that provides the information on the smallest measurable concentration of dopamine compared with the background noise. LoD calculation is shown in Eq. 9, with S and b representing a standard deviation from background noise and gradient of the linear curve, respectively. S was obtained by 3 times DPV measurement of background signal, the electrolyte used (PBS 10 mM) without dopamine. The multiplier factor of 3 in the equation is the standard signal-tonoise ratio. The limit of quantification (LoQ) calculation was written in Eq. 10, similar to the LoD calculation, except multiplier factor of 10.
Calibration curve in Figure 8D  Selectivity can provide information about how the interference of other ingredients affects the measurement of the target analyte. This test was conducted using the DPV method in the 0-1.2 V operating voltage range in PBS 10 mM pH 7.4. Three interference compounds were used: 4.5 mM glucose, 5 mM urea, and 2.5 mM ascorbic acid. These concentrations are selected according to the content of each compound in the blood serum.
From Figure 10, it appears that there are four anodic reaction peaks. Glucose gave an increasing response at a working voltage of 0.099 V, ascorbic acid at 0.245 V, and urea at 0.389 V. It was seen that ascorbic acid had a peak in the range of increasing dopamine. Ascorbic acid itself is a natural interference compound of dopamine both in the biological nervous environment and uric acid. Ascorbic acid has an important role as an antioxidant in the body. The closeness of the oxidation peaks is caused by the fact that the two specimens tend to have the same oxidation potential [39] (Liu Y et al., 2008). However, the difference in oxidation potential between DA and AA is distinct, so it can be stated that GO/Fe 3 O 4 has good selectivity for dopamine.
Finally, sensor stability testing is done by doing repeated DPV for a certain period of time ( Figure  11). The GO/Fe 3 O 4 -modified working electrode was stored at room temperature without special treatment. Then the test is carried out on the third day, the first week, the second week and lastly, the third week. The measurement on day 0 will be the standard for comparing how much change occurs in the response over time. The results showed that the peak value of the oxidation current decreased with increasing storage time. The sensor response still gave a response of 66% on the third week with no special treatment.
The comparison of our sensor performance with other published research on dopamine detection based on graphene, graphene oxide, reduced graphene oxide, and carbon nanotubes is shown in Table 2.
Our work shows better performance in terms of LOD than several previous works [40][41][42][43]. In terms of a nanocomposite synthesis method, our work uses a typical modified hummer method for obtaining GO, followed by straightforward synthesis to combine with Fe 3 O 4 and ended with simple measurement protocol. The development cost is also relatively cheaper than the previous works in Table 2 if we evaluate the precursor and electrode materials. Furthermore, magnetic-based iron oxide nanomaterial is very promising for many applications, such as in vivo options for drug delivery, magnetic resonance imaging (MRI), and thermoablation. As for in vitro options, magnetic iron oxide is used extensively for bioseparation, targeted receptors, and diverse techniques of biosensors development [44,45]. Thus, it opened up future biomedical applications in combining the functionality of biosensing with another purpose when magnetic iron oxide is selected. However, reducing the background value must be considered further whenever evaluation of another body fluid, such as blood, is necessary, as it is in the range of 0-0.25 nM [38]. A possible approach to be considered is to change electrodes into a microscale electrode array in combination with promising nanocomposite to help improve the detection limit of biosensors toward dopamine and, at the same time, provide portability and simplicity for on-site analysis [46,47].

Conclusion
Performance comparison between dopamine detection on GCE with modification and without modification showed that GO/Fe 3 O 4 has a great ability to detect dopamine with high peak current in both anodic and cathodic reactions. DPV measurement shows that the GO/Fe 3 O 4 sensor has a linear range of 1-10 mM concentrations which results in LoD value at 0.48 lM and the LoQ value of 1.6lM dopamine, respectively. GO/Fe 3 O 4 shows great selectivity when tested using ascorbic acid, glucose, and uric acid as interfering agents. Under storing durations, current peaks dropped into 66% performance after 3 weeks.  Figure 11. Peak current (Ip) change on day-to-day measurements (stability test).

Disclosure statement
No potential conflict of interest was reported by the authors.