Moxifloxacin Hydrochloride Electrochemical Detection at Gold Nanoparticles Modified Screen-Printed Electrode

It appeared that either the carbon paste or the screen-printed carbon electrodes that were modified with gold nanoparticles (AuNPs) gave rise to the largest current responses after a rapid screening of various nanomaterials as modifiers of carbon composite electrodes in view of designing an electrochemical sensor for Moxifloxacin Hydrochloride (Moxi). The screen-printed electrode (SPE) support was preferred over the carbon paste one for its ability to be used as disposable single-use sensor enabling the circumvention of the problems of surface fouling encountered in the determination of Moxi. The response of AuNPs modified SPE to Moxi was investigated by cyclic voltammetry (CV) (including the effect of the potential scan rate and the pH of the medium), chronoamperometry, and differential pulse voltammetry (DPV) after morphological and physico-chemical characterization. DPV was finally applied to Moxi detection in phosphate buffer at pH 7, giving rise to an accessible concentration window ranging between 8 µM and 0.48 mM, and the detection and quantification limits were established to be 11.6 µM and 38.6 µM, correspondingly. In order to estimate the applicability of Moxi identification scheme in actual trials, it was practiced in a human baby urine sample with excellent recoveries between 99.8 % and 101.6 % and RSDs of 1.1–3.4%, without noticeable interference.

Heterogeneous carbon-based electrodes (CPEs or carbon screen-printed electrodes, SPEs) and their modified forms are suitable for the determination of different substances due to their fairly

Carbon Paste Electrodes (CPEs)
The bare carbon paste electrode (BCPE) was fabricated from mixing 5.0 g of graphite powder with 2.0 ml of paraffin oil into mortar for 10 min. to obtain a homogenous paste. The modified CPEs were prepared through several trials by adding selected amounts of different modifiers to the paste and hand blending the resulting mixtures for 10 min. more and checking their sensitivity toward Moxi in order to achieve the optimum sensor composition in each case. On that basis, 0.05 g of zeolite powder and 0.01 g of CNTs were used to obtain zeolite/carbon nanotubes modified CPE (ZCNTMCPE), 0.05 g of NCO for the nano-Co (II, III) oxide modified CPE (NCOMCPE), 0.05 g of cellulose acetate phthalate powder +0.05 g of chitosan and 0.05 g of silver nanoparticles for the silver nanoparticles/chitosan/cellulose acetate phthalate modified CPE (SCCMCPE), 0.03 g of GO, and 0.05 g reduced glutathione for the graphene oxide/reduced glutathione modified CPE (GORGMCPE). Each CP mixture was then packed into the end of Teflon tube (with diameter of 3 mm, corresponding to 7 mm 2 electrode surface area) to prepare the corresponding sensors. The excess of composite material was removed and the electrode surface was smoothed while using ultrafine emery paper. Stainless steel ensured the electrical contact. Finally, the gold nanoparticles modified CPE (GNMCPE) was prepared by immersing a BCPE in 0.01 M of gold chloride solution and applying a deposition potential of −0.4 V for 300 s in order to deposit a nano-Au film, which was then air-dried for 5 min.

Screen-Printed Carbon Electrodes (SPEs)
Screen-printed carbon electrodes (SPEs) were fabricated through ink screen-printing over the ceramic supports, with a silver reference electrode and carbon counter electrode. The PVC insulator was handled to cover the electrodes after being dried for half an hour at 60 • C [49]. BSPE (bare screen-printed electrode) was modified with gold films by means of an electrodeposition method in order to prepare gold nanoparticles modified SPE (GNMSPE). At potential of −0.4 V for 300 s [50], gold film was deposited over SP (screen-printed) carbon surface from 6 mM gold chloride solution (pH = 2.0) and left in air to dry for five minutes in order to be subsequently utilized at the electrochemical measurements. The geometrical surface area of the working electrode was 5 mm 2 .

Methods and Instrumentation
The electrochemical experiments were carried out using a 25 mL three-electrode cell comprising a platinum auxiliary electrode (CE), the modified CPE or SPE sensors as working electrode (WE), and a saturated calomel electrode (SCE) serving as reference electrode (RE) [51]. Differential Pulse Voltammetry (DPV), Cyclic Voltammetry (CV), Chronoamperometry (CA), and Electrochemical Impedance Spectroscopic (EIS) measurements were performed with the Bio-logic SAS model SP-150 potentiostat monitored by a computer-controlled EC-Lab ® electrochemical software. The EIS tests were made at 10 mV ac amplitude in the frequency range of 1.0 mHz to 100 kHz [52] and the equivalent circuit models were built from the EC-Lab ® software [50][51][52][53]. Adwa 1030 digitalized pH meter (Romania) that was linked to a glass electrode was used to adjust the solution pH.
The materials and electrodes were characterized by various physico-chemical techniques. Surface morphology was observed by scanning electron microscopy (SEM) while using the Model Quanta 250 Field Emission Gun apparatus coupled with an energy dispersive analysis of X-rays (EDX) unit (FEI Company, Netherlands). Transmission electron microscopy (TEM) analysis was made using a JEM-1400 Electron Microscope (JEOL, Japan). The PAN-analytical X-Ray Diffraction equipment model X Pert PRO with secondary monochromator, operating at 45 kV and 35 mA with Cu-radiation (λ = 1.542 Å) and scanning speed of 0.04 • s -1 , was used for structural investigations. The diffraction lines located at 2θ values between 2 • and 60 • , the corresponding spacing (d, Å), and relative intensities (I/I o ), were obtained. The diffraction charts and relative intensities are obtained and compared with ICDD files.

Preliminary Observations and Surface Characterization of the AuNPs Modified Screen-Printed Electrodes
Carbon paste electrodes modified with various nanomaterials were first prepared and characterized (see Figures A1 and A2 in Appendix A) and screened by cyclic voltammetry as a test to evidence the most appropriate additive to the carbon composite electrode for the electrochemical detection of Moxi (see Figure A3 in Appendix B). When investigated in 1 mM Moxi solution (buffered at pH 7.0), from CV, EIS and calibration data (see respectively parts A, B, and C in Figure A3 in Appendix B), and the best performance (most sensitive detection) was obtained with the composite carbon electrodes covered with gold nanoparticles. The screen-printed electrode modified with gold nanoparticles (GNMSPE) was selected here for the detection of Moxi and its surface characterization was achieved while using various techniques ( Figure 1). SEM, TEM, and XRD characterized GNMSPE ( Figure 1). Comparing the SPE surface observed before ( Figure 1A) and after gold electrodeposition ( Figure 1B) indicates the successful formation of gold nanoparticles well dispersed onto the electrode surface, with spherical shape (and sizes of 13-58 nm, as evaluated from TEM examination, see Figure 1C). As compared to the GNMCPE ( Figure A1F), it seems that the density of gold NPs is slightly larger on GNMSPE ( Figure 1B). The XRD pattern of the GNMSPE shows sharp diffraction lines at 26.6 • , 38.1 • , and 42.1 • ( Figure 1D). The two sharp diffraction lines located at 2θ = 38.1 • (with d-spacing of 2.6 Å) and 42.1 • (with d-spacing of 2.0 Å) correspond to the Au nanoparticles, while the third one that is located at 2θ = 26.6 • (with d-spacing of 3.36 Å) is attributed to carbon, the other main constituent of the modified electrode. No additional lines originating from any other crystalline elements can be noticed, which indicates the high purity of the deposited Au-nanoparticles and that their good crystallinity is supported by the sharpness of diffraction lines. The definite line broadening of the peaks suggests that Au-particles are in the nanometer size range.

pH Effect and Sensor Stability
The effect of pH on Moxi oxidation was evaluated in B-R buffer (pH 2.0-9.0) at BSPE and GNMSPE, respectively ( Figure 3A). The optimum pH corresponding to the largest currents was 7.0 for both electrodes, but the signal intensity was much larger on the modified electrode (in agreement with the results of Figure A3A in Appendix B). The best performance at neutral pH was also evidenced from impedance data ( Figure 3B). Though not easily measurable due to a peak-to-wave shape, the voltammetric signals shifted to lower potential values at higher pH, with a slope change around pH 6, consistent with the Moxi pK a value of 6.25 [55]. Such behavior also agrees with the transfer of protons that are involved in the electrochemical oxidation of structurally related compounds [56]. The stability of GNMSPE was tested in B-R buffer solution comprising 1 mM of the analyte through CV technique ( Figure 3C). It was demonstrated that the signal intensity decreased rapidly with lengthening the immersion time of sensor: starting around 100 µA, the signal intensity dropped by one half of its value after only one more CV cycle, to level down to about one-third of its initial value after some minutes of successive analyses. This is probably due to the irreversible adsorption of the Moxi oxidation products onto the electrode surface and/or the poisoning of the surface active sites (Au nanoparticles) by chelation with Moxi [15,57], justifying the need for a single-use sensor, as GNMSPE, thanks to the screen-printing technology to produce such electrodes.
where i = current, in unit A, F = Faraday constant (96485 C/mol), n = number of electrons, A = area of the electrode in cm 2 , c 0 = bulk concentration of the analyte in mol/cm 3 ; t = time in s, and D = diffusion coefficient for species in cm 2 /s. Figure 4A shows chronoamperograms that were recorded at a constant potential of +1.03 V vs. SCE utilizing GNMSPE in BR buffer (pH = 7.0) containing various Moxi concentrations. The catalytic rate constant k cat (mol -1 L -2 s −1 ) was obtained from the relation (2) [59]: with I cat /I f being the ratio between the currents sampled in the presence (I cat ) or absence of Moxi (If), C 0 the bulk concentration of Moxi (mol L −1 ), and t the elapsed time (s). The mean catalytic rate constant was evaluated from the slope of I cat /I f versus t 1/2 plot (Figure 4B), and a value of k cat = 7.91 × 10 3 mol −1 L −2 s −1 was calculated for a Moxi concentration of 1.0 mM. The high K cat value obtained indicates quite fast charge transfer processes between the analyte and the modified electrode surface and supports the interest of AuNPs for Moxi determination.

Calibration Plot and Detection Limit
Differential pulse voltammetry (DPV) was used as sensitive technique for building the calibration plot for Moxi ( Figure 5). As shown, a linear variation of peak currents with Moxi concentration was obtained in the 8.0-480 µM range (in B-R buffer, pH 7.0): Ip (µA)= 2.38 + 0.026 C (µM), r 2 = 0.984. The limit of detection (LOD = 3 σ/m) and limit of quantification (LOQ = 10 σ/m), where, σ is the standard deviation and m is the calibration plot slope were obtained as 11.6 µM and 38.6 µM, respectively. Comparing the obtained data for Moxi determination by GNMSPE against others analytical or methods [7,60,61] or electrochemical techniques [12][13][14] shows acceptable performance of the proposed novel sensor, offering opportunities for on-site analyses thanks to the possible mass production of low-cost screen-printed electrodes and the suggested electrochemical sensor for Moxi detection also compared with others [17,[62][63][64][65][66][67] in terms of detection limits, as shown in Table 1.

Real Sample Analysis, Interfering Species and Simultaneous Determination Test
A calibration plot was established utilizing the GNMSCPE in a diluted urine medium, giving a linear correlation over a concentration range of 20-450 µM, in order estimate the method applicability to real sample analysis ( Figure 6). The linear regression equation that was obtained from the variation of DPV peak currents as a function of the Moxi concentration s was: Ipa (µA) = 0.028C (µM) + 3.84, r 2 = 0.98). The LOD (3σ) was 10.7 µM, and the LOQ (10σ) was 35.6 µM. Four different concentrations were chosen on the calibration plot through standard addition method for the examination of Moxi in buffered solution of pH 7.0, and each measurement was repeated five times in order to evaluate the accuracy and precision of the approach, as represented in Table 2   In order to investigate the interference effects for other compounds that can co-exist with Moxi in urine samples such as uric acid (UA), ascorbic acid (AA), dopamine (DA) and hormones like estradiol (ED), a fixed amount of Moxi (200 µM) spiked with the same concentration of these substrates and evaluated under the same experimental conditions through DPV. The results (see bottom inset C in Figure 6) indicate the negligible effect of interfering substrates (i.e., 0.8%, 1.8%, 3.7%, and 5.2% signal intensity decrease for ascorbic acid, dopamine, uric acid, and estradiol, respectively), which ensures good sensor selectivity. However, it can be anticipated that structurally related fluoroquinolones could interfere and the sensor would be not selective to Moxi in this case.
Typically, a patient suffering from dry cough and high fever rate should receive 200 mg moxifloxacin orally each half a day and 500 mg paracetamol when necessary [68], justifying the need for simultaneous detection. Hence, the immediate identification of 200 µM Moxi with 200 µM paracetamol in pH 7.0 utilizing GNMSPE was achieved (see top left inset A in Figure 6). The curve demonstrates a sufficient separation between the two signals, i.e. paracetamol (at 0.49 V) and Moxi (at 0.98 V), with ∆E of 0.49 V, enabling the possible simultaneous detection of both Moxi and paracetamol in the same medium. To support these findings, a calibration for Moxi in the presence of a constant concentration of paracetamol and, the opposite, a calibration for paracetamol in the presence of a constant concentration of Moxi, were made (see right bottom inset B in Figure 6), confirming the interest of GNMSPE for the simultaneous detection of these two pharmaceuticals.

Reproducibility and Long-Term Stability
Five distinct electrodes were tested for the voltammetric determination of 50 µM Moxi without obvious dispersion of the data (RSD value of 2.3%), which ensures a preparation process highly trustworthy, in order to validate the reproducibility of GNMSPE in terms of RSD.
The long-term stability of the proposed electrode was evaluated through storing GNMSPE in a refrigerator (4°C) for seven days. After that, a single voltammetric measurement for Moxi was carried out, revealing a current response at 95% of the value measured directly after fresh preparation, indicating good storage stability of the electrode.

Conclusions
This work brings together a comparison between several modifiers of heterogeneous carbon electrodes, pointing out the interest of gold nanoparticles on carbon paste or screen-printed carbon electrodes for the electrochemical sensing of Moxi. SPE that was modified with simple electrochemically deposited gold nanoparticles, GNMSPE, showed a high sensitivity and applicability for detecting Moxi in aqueous solutions and urine samples under physiological conditions with acceptable LOD (11.6 µM), wide linear concentration range (8-480 µM), and excellent recoveries (ranging between 99.8% and 101.6%) and RSDs (1.1-3.4%, n = 5). Sensor optimization was achieved from data that were collected with the aid of numerous electrochemical techniques (CV, DPV, EIS, and CA). The GNMSPE electrochemical sensor can be also applied to Moxi determination in the presence of paracetamol.

Conflicts of Interest:
The authors declare no conflict of interest.

Surface Description of the Modified Carbon Paste Electrodes
The SEM micrograph of the bare carbon paste electrode (BCPE) reveals a rather rough surface with EDX confirming the presence of carbon and trace of oxygen atoms ( Figure A1A). Data obtained for cobalt oxide nanoparticles (<100 nm) modified carbon paste electrode (NCOMCPE) and corresponding EDX analysis confirm the additional existence of cobalt in the form of well dispersed ice-like shaped particles on the electrode surface ( Figure A1B). The micrograph obtained for GORGMCPE sensor and the associated EDX data endorsing the existence of S and large carbon content ( Figure A1C) are consistent with dispersed graphene oxide blended with reduced glutathione of average size less than 445 nm. From the micrograph recorded for the ZCNTMCPE sensor, which comprises a zeolite and multi-walled carbon nanotubes (3-20 nm OD, 1-3 nm ID, 0.1-10 µm long), one can distinguish ( Figure A1D) the zeolite crystals along with dispersed carbon nanotubes bundles blended with the aluminosilicate particles (also evidenced from EDX confirming significant amounts of Si, Al and O). The SCCMCPE sensor contains AgNPs, chitosan and cellulose phthalate on the CPE surface, which agrees with both EDX data proving the existence of silver and SEM showing dispersed Ag nanoparticles with an average size of 50 nm ( Figure A1E). Finally, the observation of GNMCPE confirms that gold nanoparticles have been successfully deposited in a homogenous pattern onto the carbon paste surface ( Figure A1F) with a size of 14-75 nm.
EDX mapping of all the above electrode surfaces ( Figure A2) supports the quite good uniformity of all blended substances in the carbon paste matrix although it is not possible to distinguish all individual components/modifiers by this way.

Voltammetric Response of Moxifloxacin Hydrochloride at the Various Electrodes
Cyclic voltammetry was first used as a screening test to evidence the most appropriate additive to the carbon composite electrode, and this has been investigated in 1 mM Moxi solution (buffered at pH 7.0) at a potential scan rate of 50 mV s −1 ( Figure A3A). Because of ease of preparation and modification, the carbon paste electrodes were used for such screening and then the best system was extended to the screen-printed electrode, for further studies and practical application. As shown, an oxidation peak for Moxi appeared around 1.0 V and its intensity followed a trend (GNMCPE > GORGMCPE > SCCMCPE > NCOMCPE > BCPE> BSPE, with respective current values of 58.1 > 23.8 > 18.0 > 17.4 > 16.0 > 13.1 > 12.8 µA) indicating that the composite electrode covered with gold nanoparticles led to the most sensitive detection. The performance of gold nanoparticles deposited onto the screen-printed electrode was even better (signal current as high as 97.0 µA for GNMSPE, i.e. almost one order of magnitude larger than the response of an unmodified carbon electrode). This can be accredited to the enhanced conductivity brought by the gold nanoparticles but also to the possible chelation of Moxi to the gold surface (i.e., through the electron pairs on the hydroxyl (oxygen atoms) and/or amide groups (nitrogen atoms) or through the resonance of the double bonds on the six carbon atoms rings [17], as previously reported for Moxi binding to silver nanoparticles [15,57].
The above trend in current response was also consistent with impedance data (see Nyquist plots in Figure 3B). The best model [69][70][71] that suits the measurements ( Figure A3B, inset) is a two-time constants model (parallel) comprising R 1 (solution resistance), R 2 , R 3 (Charge transfer resistance), W (Warburg impedance), C 1 (capacitance), and Q 2 is a constant phase element of capacitance (representing surface non-ideality). Hence, the reaction model is contingent on both of diffusion and charge transfer processes [72]. From data fitting of Nyquist plots according to this equivalent circuit model, one can evaluate impedance values of ca. 1.9 KΩ (highly conductive) for GNMSPE, 20 KΩ for the bare CPE and 28 KΩ for the bare SPE (lowest conductivity) for the bare CPE. This confirms the choice of GNMSPE as the most appropriate sensor for the present work, which will thus be used exclusively afterwards owing to its highest current and conductivity. In order to prove our hypothesis and ensure the enhanced conductivity brought by the gold nanoparticles, we performed one EIS measurement before CV of the Figure 3A using BSPE (showing high resistance) and one EIS measurement after CV using GNSPE (showing low resistance) as represented by Figure A3B. Figure A3C represents a comparison between the analytical curves obtained from performing DPV measurements at scan rate 10 mV s −1 with sensing platforms ZCNTMCPE, NCOMCPE, SCCMCPE, GORGMCPE, GNMCPE and GNMSPE. As represented in this curve, there are a distinctive difference in the slope of each line owing to the type and amount of each used modifying agent at each sensing platform. The steeper line will represent low conductivity for Moxi by this sensor.