Simultaneous Voltammetric Determination of Uric Acid, Xanthine, and Hypoxanthine Using CoFe2O4/Reduced Graphene Oxide-Modified Electrode

In the present paper, the synthesis of cobalt ferrite/reduced graphene oxide (Co2Fe2O4/rGO) composite and its use for the simultaneous determination of uric acid (UA), xanthine (XA), and hypoxanthine (HX) is demonstrated. Cobalt ferrite hollow spheres were synthesized by using the carbonaceous polysaccharide microspheres prepared from a D-glucose solution as templates, followed by calcination. The CoFe2O4/rGO composite was prepared with the ultrasound-assisted method. The obtained material was characterized by using X-ray diffraction, scanning electron microscopy, X-ray photoelectron spectroscopy, EDX elemental mapping, and nitrogen adsorption/desorption isotherms. The electrochemical behavior of UA, XA, and HX on the CoFe2O4/rGO-modified electrode was studied with cyclic voltammetry and differential pulse voltammetry (DPV). The modified electrode exhibits excellent electrocatalytic activity towards the oxidation of the three compounds. The calibration curves for UA, XA, and HX were obtained over the range of 2.0–10.0 μM from DPV. The limits of detection for UA, XA, and HX are 0.767, 0.650, and 0.506μM, respectively. The modified electrode was applied to the simultaneous detection of UA, XA, and HX in human urine, and the results are consistent with those obtained from the high-performance liquid chromatography technique.

The synthesis and design of new electrode modifiers with numerous electrochemical sensing properties have been a great concern to many scientists. One of these modifiers is cobalt ferrite (CoFe 2 O 4 ). Although cobalt ferrite has excellent magnetic properties, high coactivity and hardness, and moderate saturation magnetization and is used in magnetic devices, gas sensor application [10,11], and surface-active Co(II), it has not attracted much attention in the electrochemical analysis [12]. Cobalt ferrite has a low surface area because aggregation usually occurs owing to the high surface energy of the nanoparticles, deteriorating the electrochemical activity of the material. To limit its aggregation, two approaches have usually been employed: (i) the synthesis of CoFe 2 O 4 in hierarchical structures (e.g., rods, urchins, and flower-like structures) and (ii) dispersion of CoFe 2 O 4 nanoparticles in carriers with a large surface area. Related to the latter, cobalt ferrite has usually been dispersed in organic or inorganic substrates. The cobalt ferrite-based materials have been employed to develop the novel electrodes for voltammetric determination of some compounds. Han et al. reported the synthesis of β-cyclodextrin-cobalt-ferrite nanocomposite to modify an electrode for catechol determination [13], Yardımcı et al. used cobalt ferrite/chitosan nanocomposite for H 2 O 2 sensing [14], and Ensafi et al. determined H 2 O 2 and nicotinamide adenine dinucleotide by using a cobalt ferrite/graphene oxide-modified electrode [15].
Besides graphene oxide, a derivative of this materialreduced graphene oxide (rGO)-is also an excellent substrate to disperse cobalt ferrite. The rGO with smaller oxygen content is produced from graphene oxide via chemical, thermal, or other approaches. The rGO possesses good conductivity and thermal and chemical stability [16,17] that makes it to be used as a novel material to develop the electrodes for electrochemical sensing biomolecules [18,19], metal ions [20], and toxic chemicals [21]. Therefore, a combination of CoFe 2 O 4 with rGO is expected to result in a composite with a high surface area and electrical conductivity and a possibility of application in electrochemistry. To the best of our knowledge, the use of the CoFe 2 O 4 /rGO composite as an electrode modifier for the determination of UA, XA, and HA by using the voltammetry method is very limited in the literature.
In the present work, we prepared cobalt ferrite hollow spheres to fabricate a CoFe 2 O 4 /rGO-modified electrode. Then, we used this electrode to study the electrochemical behaviors on UA, XA, and HA oxidations by using cyclic and differential pulse voltammetry. We also addressed the analysis of real samples.

Synthesis of Reduced Graphene
Oxide. Graphite oxide (GrO) was prepared by using Hummers' process [22]. A mixture of 2 g of GrO and 500 mL of double distilled water was stirred under ultrasonication for 5 h to get a graphene oxide suspension (GO) (4 mg/mL GO). The GO suspension (12.5 mL GO in 250 mL distilled water) was adjusted to pH 9-10 with a 25% NH 3 aqueous solution. Then, add 0.012 g of N 2 H 4 ·H 2 O to the GO suspension and keep the mixture at 90°C for 60 min. To remove the residual N 2 H 4 ·H 2 O, the mixture was washed with a 30% H 2 O 2 aqueous solution several times. The resulting mixture was neutralized until pH 7, with a 5% HCl solution. The mixture was rinsed with distilled water five times (30 mL each time). The solid, which is a graphene oxide (rGO), was collected by centrifugation and dried at 60°C for 24 h. 2.2. Apparatus. The crystal structure of the material was identified by using X-ray powder diffraction (XRD) on a Bruker D8 equipped with Cu Kα radiation (λ = 1:5406 Å). Infrared spectra were recorded on a Fourier mid-IR InfraLUM FT-08 between 4000 and 150 cm −1 . Scanning electron microscopy (SEM) images were recorded on an SEM JMS-5300LV (Japan), equipped with energy-dispersive X-ray microanalysis Nova Nano SEM 450. TEM images were obtained on an FEI 2 Journal of Nanomaterials spirit instrument (120 kV) electron microscope. The thermal properties were measured on a Micromeritics Tristar 3000 (USA). Magnetic hysteresis loops were measured on a Vibrating Sample Magnetometer (Micro Sence Easy VSM 20130321-02) at room temperature. Energy-dispersive X-ray elemental mapping (EDX-elemental mapping) was conducted on a Horiba, Japan. Ultrasonic treatment was performed in a Cole-Parmer 8890. Electrochemistry was studied by using a CPA-HH5 in which the three-electrode system consisted of a glassy carbon electrode (GCE, a working electrode), an Ag/AgCl reference electrode (Model RE-5, BAS), and a platinum wire auxiliary electrode. The UA, XA, and HX determinations with high-performance liquid chromatography (HPLC) were performed on a Shimadzu 2030 HPLC system, with a UV-Vis detector set at 273 nm. An AC18 (6:0 mm × 150 mm, 5 μm) chromatographic column was employed. The mobile phase is an acetonitrile/water mixture (25/75, v/v) at a flow rate of 1.5 mL·min −1 , while the injection volume was 5 mL·min −1 .

Analytical
Procedures. The cyclic voltammetry (CV) technique was used for the preliminary studies on the electrochemical behavior of UA, XA, and HX. The differential pulse voltammetry (DPV) method was employed for the development of the electroanalytical method for the simultaneous determination of UA, XA, and HX in real samples. Before modification, the GCE was polished with 0.05 μm alumina powder on a polishing pad, followed by sonication treatment for about two minutes in double distilled water and dried at room temperature and immediately used for modification. Two milligrams of CoFe 2 O 4 /rGO was added to 1 mL of methanol under ultrasonic agitation for 60 min, resulting in a homogeneous black suspension. Five microlitres of CoFe 2 O 4 /rGO suspension was dropped on the electrode surface. Then, the modified electrode was dried at ambient temperature to obtain a CoFe 2 O 4 /rGO glassy carbon electrode.

Real Sample
Analysis. Three samples of human urine, provided by a clinical laboratory, were employed to test the method. In detail, 1.0 mL of the urine sample was spiked with UA, XA, and HX and mixed with 2 mL of the B-RBS buffer solution to make a 10.0 mL test solution. The proposed different pulse voltammetry method was used to detect UA, XA, and HX in the spiked solution.

Results and Discussion
3.1. Characterization of CoFe 2 O 4 /rGO. Figure 1(a) shows the TG/dTG curves of CoFe 2 O 4 /carbon hollow spheres recorded in airflow from 40 to 800°C. Clearly, a weight decrease of about 2 wt.% is observed from 30 to 120°C, which is ascribed to the desorption of physically adsorbed water in the precursor. Another weight loss of about 50 wt.% at around 400°C is assigned to the combustion of rGO in the CoFe 2 O 4 /rGO nanohybrid. This weight loss is close to the initial CoFe 2 O 4 /rGO ratio of 1 : 1. The carbon residues are completely removed at temperatures higher than 400°C.

Journal of Nanomaterials
The XRD pattern of rGO in Figure 1(b) presents a broad (002) diffraction peak between 20 and 35°, which corresponds to a short-range order in the stacked graphene sheets. The largely reduced interlayer spacing of about 0.342 nm (at 2 theta of 26°), compared with 0.780 nm (at 2 theta of 11.3°) for graphene oxide [23], indicates the formation of reduced graphene oxide, in which the oxygen functional groups are removed significantly during the reduction process. It is worth noting that the characteristic X-ray diffraction indexed as a spinel type according to JCPDS No. 00-002-1045 in CoFe 2 O 4 /carbon spheres is observed. This means that the cobalt ferrite phase is formed during the hydrothermal treatment. Cobalt ferrite with high crystallinity is formed after the removal of the carbon template through calcination at 500°C (Figure 1(b)). The characteristic peaks of the cobalt ferrite phase in CoFe 2 O 4 /rGO are observed in Figure 1(b). However, the large background indicates a large amorphous phase resulting from reduced graphene oxide. These results confirm the successful synthesis of the CoFe 2 O 4 /rGO composite.
The EDX analysis shows that the molar composition of Co, Fe, O, and C in CoFe 2 O 4 is 13.56, 26.81, 47.18, and 12.03%, respectively, and in CoFe 2 O 4 /rGO is 4.81, 9.21, 34.52, and 51.46%, respectively. Correspondingly, the Co/Fe molar ratio is 1 : 2, which is very close to the stoichiometric ratio in the two samples. In contrast, the excessive carbon content and low Co and Fe content in CoFe 2 O 4 /rGO indicate the presence of rGO. These results further confirm the presence of stoichiometric cobalt ferrite (CoFe 2 O 4 ), and this is probably assigned to the high dispersion of CoFe 2 O 4 on the rGO surface.
The formation of the CoFe 2 O 4 /rGO composite was also studied by using FT-IR spectroscopy ( Figure 2). On the FT-IR spectrum of rGO (Figure 2(a)), we can see vibration at 3444 cm −1 characteristic to OH groups and at 1641-1127 cm −1 , attributed to carbonyl (C=O) and epoxy (C-O-C) groups in rGO [24]. However, these peaks have very low intensities, implying that they are removed significantly during the reduction. The typical inverse spinel ferrite structure includes two adsorption bands: one at around 339 cm −1 , representing the stretching vibration of the tetragonal group The material is formed in a hollow spherical shape ( Figure 3). Figure 3(a) shows the hollow spheres (2-10 μm) with flocculent substances (CoFe 2 O 4 ·nH 2 O) on the carbon surface. These particles remain unaltered after removing the carbon template (Figure 3(b)). The TEM image of rGO shows a stacked and crumpled morphology due to the exfoliation and restacking process [25] (Figure 3(c)). The CoFe 2 O 4 /rGO composite consists of CoFe 2 O 4 hollow spheres embroiled homogeneously with rGO sheets to form a hierarchical structure that favors the diffusion and adsorption of the analytes (Figure 3(d)).
The EDX elemental mapping in Figure 4(a) shows the SEM bright field image of the CoFe 2 O 4 /rGO composite. The images in Figure 4 reveal that CoFe 2 O 4 clusters of around 500 nm in size are embedded in the rGO matrix.
The XPS survey curve presents the existence of Co, Fe, O, and C in CoFe 2 O 4 /rGO at 793.28, 721.48, 536.78, and 288.78 eV, respectively ( Figure 5(a)). The XPS spectrum Co2p possesses two main peaks at 780.67 and 796.28 eV, which are assigned to Co2p3/2 and Co2p1/2, respectively ( Figure 5(b)). The energy gap between the Co2p main peak  Journal of Nanomaterials and the satellite peak can be employed to decide whether cobalt exists as Co(II) or Co(III). If the gap is ca. 6.0 eV, cobalt exists as Co(II), and while the gap is 9-10 eV, the cation is Co(III) [26,27]. In our case, the energy gap of the cobalt cation is 6.62 eV for Co2p [28][29][30]. In Figure 5(d), the O1s peak is deconvoluted into two shoulder peaks at 532.56 eV and 530.46 eV, corresponding to the hydroxyl groups adsorbed on the surface and the Fe-O bond in the crystal lattice, respectively [31]. The C1s spectrum could be fitted to four carbon species at 284.58, 285.55, 286.9, and 289.46 eV, corresponding to C/C=C groups in the nonoxygenated rings, C-OH, epoxy C-O-C, and carboxyl group COOH, respectively ( Figure 5(e)) [32]. The magnetic hysteresis curves of the CoFe 2 O 4 and CoFe 2 O 4 /rGO composite, measured at 298 K with the field sweeping from −10000 to 10000 Oe, indicate that the two materials are ferromagnetic ( Figure 6). The saturation magnetization is 60.2 emu·g −1 for CoFe 2 O 4 and 48.6 emu·g −1 for CoFe 2 O 4 /rGO. These values are slightly higher than those of other cobalt-ferrite-based materials, reported previously [32][33][34].
The specific surface area and pore volume of the materials were determined by using the nitrogen sorption     (Figure 9). This also suggests a rapid electron transfer within the CoFe 2 O 4 /rGO film. However, when the volume of the CoFe 2 O 4 /rGO suspension is greater than 7.5 μL, the I a tends to decrease due to larger film thickness and the increasing mass transfer resistance against the electron transfer. In this work, the amount of CoFe 2 O 4 /rGO suspension on the GCE surface was selected at 5 μL.

. Effects of pH.
The pH of the solution has a remarkable influence on the UA, XA, and HX electrooxidation on the CoFe 2 O 4 /rGO-GCE. Figure 10(a) illustrates the DPV curves of UA, XA, and HX on the modified electrode in the pH range of 2-10. As shown in Figure 10(b), when the pH of the solution is lower than 4, the I a of all analytes decreases with pH. At a pH higher than 6, only the I a of XA increases with pH and reaches a maximum at pH 6. However, the I a of UA and HX tends to reduce until pH reaches 7. Afterward, they change irregularly as pH increases. The dependence of the peak current of UA, XA, and HX on pH is complicated, but a decreasing tendency is observed with all the analytes ( Figure 10(b)). This may result from the adsorption behavior of UA, XA, and HX on the CoFe 2 O 4 /rGO-GCE. UA, XA, and HX are known as protic aromatic molecules and can become deprotonated to form negatively charged species (anions) at higher pH. Simultaneously, the surface of CoFe 2 O 4 /rGO-GCE also becomes negatively charged at high pH. As a result, less intensive adsorption of the analytes on the electrode might take place, thus reducing the I a .
The pH effects on peak potentials (E p ) for UA, XA, and HX oxidations on the CoFe 2 O 4 /rGO-GCE were also studied. The E p values shift to more negative potential with pH, and the plots of E p vs. pH exhibit high linearity with high determination coefficients (>0.99) (Figure 10 7 Journal of Nanomaterials changing the scan rate from 50 to 500 mV·s −1 (Figure 11). Figure 11(a) presents the scan rate effect of electrochemical responses studied with the CV method. The linear plot of the peak current vs. the square root of the scan rate was conducted to estimate whether the electrooxidation reaction is controlled by diffusion or adsorption. If the plot of I p vs. v 1/2 passes the origin, this process is controlled by diffusion; otherwise, it is controlled by adsorption [35,36]. The plots of I p,UA , I p,XA , and I p,HX vs. v 1/2 are highly linear (r = 0:9588 ÷ 0:9866, p < 0:001) (Figure 11(b)). The number in the parentheses presents the 95% confidence interval. No lines pass the origin because all the intercepts are greater than zero (varying from 0.690 to 6.770 for UA, 10.913 to 16.767 for XA, and 1.799 to 7.843 for HX at 95% confidence-Equations (2), (3), and (4)). This indicates that the electrode process is controlled by adsorption. The relationship of the anodic potential vs. the natural logarithm of the scan rate is expressed by the Laviron equation [37].
where K s is the electron rate constant of the surfaceconfined redox couple, n is the number of electrons transferred, α is the charge transfer coefficient, v is the scan rate (V · s −1 ), E°is the formal redox potential, T = 298 K, R = 8:314 J · mol · K −1 , and F = 96480 C · mol −1 .
The plots of the anodic peak potential vs. the natural logarithm of the scan rate are presented in Figure 11

Journal of Nanomaterials
The values of n × α for UA, XA, and HX are 1.222, 0.950, and 0.856, respectively. For the irreversible system, α is assumed to be 0.5 [38]. Therefore, the average values of n are 2.444, 1.900, and 1.912. Because the number of electrons transferred is an integer, this number should be 2 in this case for all analytes. Therefore, the oxidation of UA, XA, and HX involves two electrons and two protons. This is consistent with previous work [8]. The mechanism for the oxidation of the analytes is proposed in Scheme 1.
The CoFe 2 O 4 /rGO film possesses a large number of negatively charged surface functional groups (-OH) and electron-rich oxygen atoms (epoxide, OH, C=O), and they could interact with the purine derivatives (in this case, they are UA, XA, and HX). Furthermore, the CoFe 2 O 4 /rGOmodified electrode exhibits a larger real surface, π-π conjugated bonds, an abundant number of active sites, and better electronic conductivity, which leads to a strong interaction between the purine derivatives and the electrode interface. Therefore, the electrochemical oxidation of UA, XA, and HX may be substantially accelerated in the presence of the CoFe 2 O 4 /rGO film. This might result from the enhanced rate of electron transfer and electrocatalytic activity towards    Journal of Nanomaterials the oxidation of UA, XA, and HX. The argument is illustrated in Scheme 2.

Interferent Study.
For evaluating the selectivity of the modified electrode, several possible interferents were evaluated for their behavior in the determination of UA, XA, and HX (equal concentration of 2:0 × 10 −5 mol · L −1 ). The tolerance limit is the maximum concentration of a foreign substance that causes an approximately ±5% relative error in the determination. It is found that 10-fold of paracetamol and 50-fold of ascorbic acid do not interfere with the peak current of XA and HX but significantly affect that of UA. However, 1000-fold of D-glucose, 20-fold of KCl and Na 2 CO 3 , and 100-fold of MgSO 4 do not interfere with the determination of UA, XA, and HX (Table 1). These findings indicate that the present method can be used for real samples.
3.2.6. Repeatability, Linear Range, and Limit of Detection.
The repeatability of I p at the CoFe 2 O 4 /rGO-GCE was estimated by comparing the relative standard deviation (RSD) with the relative standard deviation calculated from Horwitz's equation: RSD H = 2 1−0:5×lg C , where C is the concentration in mole fraction. If the RSD is less than 1/2 × RSD H , the repeatability is acceptable [39]. Each signal of I p was obtained by conducting a series of four successive measurements at three concentrations.
where S is the standard deviation of the lowest concentration of UA, XN, and HP and b is the slope of calibration curve obtained from the DPV. The detection limits of UA, XA, and HX obtained in the first range are 0.767, 0.650, and 0.506 μM, respectively. A comparison of the proposed method with other electrochemical approaches for UA, XA, and HX is presented in Table 3. It is notable that the LOD of UA, XA, and HX from the proposed methods is compatible with those based on the modified electrodes reported previously. From the analysis above, we can conclude that CoFe 2 O 4 /rGO is a potential electrode modifier for determining UA, XA, and HX.   Table 4. The recoveries found in the range of 95%-105.3% indicate the accuracy and efficiency of the proposed method for the urine samples. The results obtained with HPLC for the real samples are given for comparison. The results from HPLC are comparable with those obtained from the DPV method proposed in this paper (paired sample t-test: α = 0:05, tð5Þ = 0:5, p = 0:637 > 0:05).

Conclusions
In this paper, the CoFe 2 O 4 /rGO-GCE is fabricated by using a simple method and employed as an electrochemical sensor for the simultaneous detection of UA, XA, and HX with the differential pulse voltammetry technique. The modified electrode has good stability and sensitivity. The results obtained with the proposed method are consistent with those from standard HPLC analysis. Therefore, CoFe 2 O 4 /rGO is expected to become a potential tool for the assay of UA, XA, and HX both in research and in clinical diagnosis owing to its good precision, high speed, and low cost of analysis.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no conflicts of interest. HX <LOD n/a n/a n/a n/a