Green Composite Sensor for Monitoring Hydroxychloroquine in Different Water Matrix

Hydroxychloroquine (HCQ), a derivative of 4-aminoquinolone, is prescribed as an antimalarial prevention drug and to treat diseases such as rheumatoid arthritis, and systemic lupus erythematosus. Recently, Coronavirus (COVID-19) treatment was authorized by national and international medical organizations by chloroquine and hydroxychloroquine in certain hospitalized patients. However, it is considered as an unproven hypothesis for treating COVID-19 which even itself must be investigated. Consequently, the high risk of natural water contamination due to the large production and utilization of HCQ is a key issue to overcome urgently. In fact, in Brazil, the COVID-19 kit (hydroxychloroquine and/or ivermectin) has been indicated as pre-treatment, and consequently, several people have used these drugs, for longer periods, converting them in emerging water pollutants when these are excreted and released to aquatic environments. For this reason, the development of tools for monitoring HCQ concentration in water and the treatment of polluted effluents is needed to minimize its hazardous effects. Then, in this study, an electrochemical measuring device for its environmental application on HCQ control was developed. A raw cork–graphite electrochemical sensor was prepared and a simple differential pulse voltammetric (DPV) method was used for the quantitative determination of HCQ. Results indicated that the electrochemical device exhibited a clear current response, allowing one to quantify the analyte in the 5–65 µM range. The effectiveness of the electrochemical sensor was tested in different water matrices (in synthetic and real) and lower HCQ concentrations were detected. When comparing electrochemical determinations and spectrophotometric measurements, no significant differences were observed (mean accuracy 3.0%), highlighting the potential use of this sensor in different environmental applications.


Introduction
COVID-19 is caused by a new strain of the coronavirus, which is associated with the same family of viruses as severe acute respiratory syndrome (SARS) and some types of the common cold [1]. As of May 19, 2021, the COVID-19 pandemic has caused more than 3.4 million deaths worldwide [2], and to date, no drug has been proven to target this virus. However, hydroxychloroquine (HCQ) has received remarkable attention as a treatment option for COVID-19 (for example, in the United Sates, Brazil, China, and India) [3]. HCQ is a halogenated aminoquinoline that exhibits wide biological activity and is often used as an antimalarial drug [4]. In Brazil, hydroxychloroquine and ivermectin are included in COVID-19 kits, as a pre-treatment option. However, no scientific results support the effectiveness of these drugs in preventing COVID-19 infection or as a treatment

Reagents and Materials
HCQ sulfate (purity 99%) and graphite powder were purchased from Sigma-Aldrich (São Paulo, Brazil). NaCl, NaOH, CH 3 COOH, HCl, Na 2 SO 4 , CH 3 COONa, and H 2 SO 4 were purchased from Merck (São Paulo, Brazil). All solutions were prepared using ultra-purified water obtained from a Milli-Q system (Millipore, Natal, Brazil). The raw cork (RAC) that was used in the experimental studies was provided by Corticeira Amorim S.G.P.S., S.A. (Porto, Portugal). The RAC granules were washed twice with distilled water for 2 h at 60 • C to remove impurities and other water extractable components that could interfere with the electrochemical analysis. Before use, the RAC was dried at 60 • C in an oven for 24 h [16]. The HCQ standard solutions were prepared daily using ultra-purified water to prevent photodegradation, to which HCQ is susceptible [13].

Preparation of Cork-Graphite Sensor
According our previous work [16], SEM micrographs of raw cork showed a honeycomb structure, which is associated with the macropores. After the pre-washing approach of cork, there are small internal impurities. Based on the results obtained by FTIR, the presence of -OH and -CH 3 signals, at 3440-3400 and 2920-2850 cm −1 , and the disappearance of the C=O stretch bands at 1745-1715 cm −1 (characteristic of ester groups, originating mainly from suberin), were the most important modifications of the material, which could give significant insights in terms of the cork surface to interact with the contaminants dissolved in water, as already reported by other authors.

Electrochemical Measurements
Electrochemical analyses were performed on an Autolab PGSTAT302N (Metrohm, Zurich, Switzerland) that was controlled with GPES software (4.0) and consisted of a three-electrode cell, using Ag/AgCl (3.0 M KCl), Pt wire and a cork-graphite sensor (GrRAC, (geometrical area of approximately 0.45 mm 2 )) as the reference, auxiliary, and working electrodes. The oxidation of HCQ was investigated using cyclic voltammetry (CV), which was performed with different supporting electrolytes (0.1 M NaCl, 0.1 M NaOH, 0.1 M CH 3 COOH, 0.1 M HCl, 0.1 M Na 2 SO 4 , 0.1 M CH 3 COONa, and 0.1 M H 2 SO 4 ). The effect of the potential scan rate (v) on the HCQ response was evaluated by varying this parameter in the range of 10 to 500 mV s −1 for different CV analyses. The DPV parameters were as follows-modulation time: (≥0.002 s), 0.05 s; interval time: (≥0.10 s), 0.5 s); initial potential: 1.0 V; final potential: 1.7 V; step potential: 0.00495 V; modulation amplitude: 0.01995 V; potential scan rate: 100 mV s −1 ; and agitation time: 30 s. The optimized parameters were used for all measurements. All analyses were performed in triplicate. All electrochemical analyses were conducted without deaeration, at room temperature (25 ± 2 • C). For comparative purposes, HCQ determination was also carried out spectrophotometrically using a UV-vis spectrophotometer (Bel Photonics, SP 2000 UV, Berlin, Germany) set at 342 nm, with a quartz cuvette and 0.01 mol L −1 HCl solution as the solvent [13]. For the determination of HCQ in different water matrices (river, lagoon, tap water, and ground water), the water samples were spiked with a known quantity of a standard solution of HCQ. The final HCQ concentration was determined using the standard addition method. For this purpose, the samples were doped with different concentrations of HCQ (5-69 µM). Thereafter, 10 mL aliquots of each sample were diluted in the supporting electrolyte solution, and the final solutions were analyzed.

Effects of Supporting Electrolyte
No electrochemical work corresponding to HCQ has been reported previously in the literature using a GrRAC sensor. Then, the supporting electrolyte effect was investigated based on the CV responses of 10 µM HCQ with a GrRAC sensor in 0.1 M NaCl, 0.1 M NaOH, 0.1 M CH 3 COOH, 0.1 M HCl, 0.1 M Na 2 SO 4 , 0.1 M CH 3 COONa, and 0.1 M H 2 SO 4 , as shown in Figure 1. In terms of the most favorable conditions, the GrRAC sensor in the presence of 0.1 M H 2 SO 4 showed the most well-defined anodic peak corresponding to HCQ. As can be observed in Figure 1, HCQ exhibited one well-defined oxidation peak in the presence of 0.1 M NaOH (Figure 1b (Figure 1f) were used as supporting electrolytes. Additionally, no reduction peak was observed, in all cases (see, Figure 1), in the reverse scan, confirming the irreversibility of the electro-oxidation process of HCQ in all supporting electrolytes.
It is important to remark that, when real water matrices will be analyzed, then, a preconditioning acidic strategy is used to preserve the water properties avoiding a quick biological degradation before the use of electrochemical determinations [30]. In this frame, 0.1 M H 2 SO 4 was selected as a supporting electrolyte in subsequent experiments, as already reported by Deroco et al. [31]. A comparison between GrRAC and graphite sensors was also carried out under acidic pH conditions (0.1 M H 2 SO 4 ) in order to verify the improvements on the HCQ signal when RAC was used as modifier. As can be seen in Figure 1, an important enhancement in voltammetric response was achieved, in the presence of 10 µM of HCQ, when GrRAC ( Figure 1g) was applied as electrochemical sensor in respect of a non-RAC modified electrode (Figure 1h). This result clearly evidences the potential utilization of GrRAC as HCQ monitoring device as well as demonstrating undoubtedly the advances on the current sensibility when RAC is mixed with graphite. In order to understand the different behaviors registered by CV analysis, the peak current of the GrRAC sensor, under diverse experimental conditions, was considered and estimated to the according electroactive surface area (A real ) of 0.12 mm 2 , see Table 1. It can be observed that the GrRAC sensor in the presence of H 2 SO 4 contributes most effectively for the oxidation of HCQ.  It is important to remark that, when real water matrices will be analyzed, then, a preconditioning acidic strategy is used to preserve the water properties avoiding a quick biological degradation before the use of electrochemical determinations [30]. In this frame, 0.1 M H2SO4 was selected as a supporting electrolyte in subsequent experiments, as already reported by Deroco et al. [31]. A comparison between GrRAC and graphite sensors

Effects of Scan Rate
As described in the previous section, the supporting electrolyte influenced the redox processes at the electrode surface. The effect of the scan rate (10-500 mV s −1 ) on the electrochemical HCQ response was also investigated using 10 µM HCQ in 0.1 M H 2 SO 4 solution. As shown in Figure 2a, increasing the scan rate, an increase on the anodic peak current (I pa ) was observed; however, the oxidation potential was also slightly shifted to more positive potential values. The peak current was determined using GPES software (version 4.0) by extrapolating from the baseline of the peak current measurement [32]. The relationship between the peak current and the square root of the scan rate (I vs. v 1/2 ), as well as the relationship between the logarithm of the peak current and the logarithm of the scan rate (log I vs. log v), both allowed us to understand the mass transport behaviors, Figure 2b. The linear relationship between the oxidation peak current and the square root of the scan rate ((I vs. v 1/2 ), see inset Figure 2a) confirmed that HCQ oxidation is a diffusion-controlled process. The relationship between I pa (µA) and log v (mV s −1 ) can be expressed by the following equations: Figure 2c. The slope estimated was about 0.5, confirming that HCQ oxidation was mainly controlled by the diffusion process, as previously observed with other organic compounds for GrRAC sensors [33,34]. Additionally, the absence of significant nonlinearity was also visually verified, as recommended by IUPAC [35,36] and the literature [37]. The number of electrons (n) is an important parameter for evaluating a completely irreversible process. E pa versus ln v was plotted in order to determine n involved in the oxidation mechanism of HCQ [38]. Figure 2c shows that E pa was shifted towards positive potentials with an increase in the scan rate. Meanwhile, the linear dependence between E pa and ln v can be expressed by the following equation: E pa /V = 0.14 − 0.10 ln v. The slope of the plot of E pa versus ln v (assuming α = 0.5) can be used to calculate n, using the Laviron theory [38,39].
where E • is the formal redox potential; K s is the standard heterogeneous reaction rate constant; α is the electron transfer coefficient that assumes a value of 0.5, owing to the irreversible electrochemical behavior in acidic media; n is the number of electrons transferred in the system, and R, T, and F have their usual meanings (R = 8.314 J mol −1 K −1 , T = 298 K, F = 96,485 C mol −1 ). E • = 0.12 V (formal standard potential) was estimated from the linear relationship between E pa and ln v by extrapolating from v = 0. From this calculation, it was found that two electrons were involved in the HCQ oxidation under acidic pH conditions ( Figure 3). In fact, the value of E • confirms the results reported by Mahnashi et al. [40], Decoro et al. [31], Khoobi et al. [41], and Arguelho et al. [11] which have suggested that the electrooxidation of HCQ involved the nitrogen atom of the alkylamino side chain and the N-heterocyclic nitrogen of the aminoquinoline moiety of hydroxychloroquine, respectively ( Figure 3). E • was obtained from the slope of E pa versus ln v (assuming α = 0.5). Consequently, the heterogeneous electron transfer rate constant (K s = 6.5 × 10 5 s −1 ) could be estimated using Equation (1).

DPV Analytical Curve
A protocol for the electrochemical determination of HCQ was developed using the DPV approach with a GrRAC electrode, and it was optimized using different electrolytes. However, based on the CV responses, H 2 SO 4 was employed as the main supporting electrolyte. Figure 4 shows the DPV data for HCQ in H 2 SO 4 ; the resulting high-intensity curve is shown in the inset of Figure 4. As previously observed for the oxidation of HCQ in the CV analysis, in terms of the voltammetric current response, acidic H 2 SO 4 conditions produced the best oxidation electroactivity signal, and it was more intense and higher at GrRAC sensor than that obtained in the graphite electrode. This behavior is associated with the morphology of the GrRAC electrode, which promoted chemical and electrochemical interactions, associated with the composition of the cork [16]. The analytical curve obtained for the GrRAC electrochemical sensor is represented and it can be observed that the peak current (Ip) increased linearly with the concentration of HCQ in the range from µM, and the linear regression equation (Ip vs. C) was obtained, Ipa (µA) = 3.0 × 10 −3 × [HCQ] − 1.6 × 10 −3 (R 2 = 0.994). The analytical curve was obtained by considering the peak intensity as a function of the HQC concentration and evaluating a range between 5-65 µM analyte concentrations. The residuals of the regression curve were also included in Figure 4 in order to confirm visually the absence of significant non-linearity, as recommended by IUPAC [35,36] and the literature [37]. The limit of detection (LOD) for the GrRAC sensor was estimated according this equation: 3.3 × S y/x /b, where S y/x is the residual standard deviation and b is the slope of the calibration plot [36]. The LOD and LQ were found to be 1.05 and 3.15 µM. Another parameter that was evaluated in this work consisted to the stability of the sensor used that shows good performance for at least two months of intensive use as well as reproducibility when four analytical curves were achieved by using the same concentration range and sensor, obtaining no significant variations.
There are few studies in the literature reporting on the quantification of HCQ. Table 2 shows these works, providing a comparison of the main analytical parameters for quantification HCQ, such as electrode composition, supporting electrolyte, method and LOD. In addition, the analytical performance such as linear range and LOD were compared with the previously reported sensor for HQC quantification. As can be confirmed, the proposed sensor has a good electroactivity performance in comparison with the methods reported previously. Ghoreishi et al. [43] reported the use of glassy carbon modified with N,Nbis[(E)-(1-pyridyl) methylidene]-1,3-propanediamine (PMPD) self-assembled monolayer (SAM) for detection of HCQ achieving a LOD of 25.8 µM. In another study, glassy carbon was modified with N,N-bis[(E)-(1-pyridyl) methylidene]-1,3-propanediamine (PMPDA) self-assembled monolayer (SAM) with which was assessed a LOD of 0.0046 µM in the presence of acetaminophen [44]. Thus, based on the results in the existing literature, the main advantages of the developed sensor are the use of a 100% natural and sustainable material as the modifier, affording low LOD values for HCQ detection. This novel sensor material seems to offer a fast, reliable, economic, and simple way for quantifying HCQ. The concentration capability of the cork modification, due to its affinity towards the analyte, produces a significant gain of sensitivity which can be used to identify and quantify HCQ in different samples without consuming time or aggressive reagents pre-treatment. There are few studies in the literature reporting on the quantification of HCQ. Table 2 shows these works, providing a comparison of the main analytical parameters for quantification HCQ, such as electrode composition, supporting electrolyte, method and LOD. In addition, the analytical performance such as linear range and LOD were compared with the previously reported sensor for HQC quantification. As can be confirmed, the proposed sensor has a good electroactivity performance in comparison with the methods reported previously. Ghoreishi et al. [43] reported the use of glassy carbon modified with N,N′bis[(E)-(1-pyridyl) methylidene]-1,3-propanediamine (PMPD) self-assembled monolayer (SAM) for detection of HCQ achieving a LOD of 25.8 µM. In another study, glassy carbon was modified with N,N-bis[(E)-(1-pyridyl) methylidene]-1,3-propanediamine (PMPDA) self-assembled monolayer (SAM) with which was assessed a LOD of 0,0046 µM in the presence of acetaminophen [44]. Thus, based on the results in the existing literature, the main advantages of the developed sensor are the use of a 100% natural and sustainable material as the modifier, affording low LOD values for HCQ detection. This novel sensor material seems to offer a fast, reliable, economic, and simple way for quantifying HCQ.

Determination of HCQ in Water Samples
To investigate the electrochemical sensor suitability for environmental monitoring, the new, low-cost, green device was used to determine the HCQ concentration in water samples from river, lagoon, and tap water sources. Table 3 shows the concentration of HCQ that was added to all samples assayed using the proposed sensor and the reference spectrophotometric method. Thus, acceptable recovery values were obtained for all types of samples, which ranged from 120 ± 0.52% for river samples, 111 ± 0.8% for lagoon samples, and 121 ± 0.35% for tap water samples ( Table 3). The HCQ signal was confirmed by intensification of the peaks associated with the addition of different volumes of standard HCQ solution to the samples ( Figure 5, (a) river, (b) lagoon, and (c) tap waters). It is important to highlight that the results were obtained with acceptable standard deviations and confidence intervals, within 95%. This information can be used to identify false positives and false negatives (α = β = 0.05), as already indicated by experts in the field [36]. Additionally, the experimental recoveries were close to 100%, although the electrochemical signal was validated in different matrices. Based on the data obtained, the proposed sensor can be used to determine the HCQ concentration in environmental samples. Table 3. HCQ concentration in water samples from river, lagoon, and tap water sources. Well-known HCQ concentration was added (standard additions of 10 mM HCQ in 0.1 M H 2 SO 4 (200, 400, and 600 µL)) to all samples assayed which were evaluated using the proposed sensor and the reference spectrophotometric method, and reporting the recovery results comparing the results with the spectrophotometric method. On the other hand, a river water sample presented a positive result to HCQ detection after spectrophotometric evaluation, registering a HCQ concentration of approximately (3.27 ± 0.04) × 10 −3 mM. After that, the same river water sample was evaluated by the DPV approach with a GrRAC sensor, achieving (2.85 ± 0.06) × 10 −3 mM. This figure evidenced the HCQ water pollution as well as the potential use of cork as a modifier of graphite for monitoring water quality, even when the difference between both measurements is about 12%, which can be considered acceptable.

Sample
It is important to highlight that the water and wastewater samples contain numerous inorganic and organic compounds present, which can interfere in the functioning of the electrochemical sensor, along with the target analyte. For this reason, the standard addition method is recommended for diminishing the matrix effect on the current-response sensibility. However, an additional study was carried out with a lagoon water sample (which was previously acidified to avoid sample decomposition) where the electrochemical sensor's response, in terms of current, was verified. As can be observed in Figure S1 in the Supplementary Material (SM), a good performance of the proposed electrochemical sensor was achieved, showing a lower matrix effect during analytical curve construction as well as good linearity in the current responses as a function of concentration (see inset in Figure S1). The LOD was found to be 2.71 µM, which is only 2.58-folds higher than the LOD obtained in this work by using distilled water (see Table 2). Then, the determination of HCQ in water matrices was adequately attained by the standard addition method, as is reported in Figure 5.

Conclusions
This study demonstrates that cork-graphite-based sensors provide a fast, reliable, cost-effective, and simple method for quantifying HCQ concentrations in different water samples. The sensor exhibited the best affinity and a higher sensitivity for HCQ when 0.1 M H 2 SO 4 was used as the electrolyte support. Based on the results, the sensor exhibited a good response for HCQ determination, despite the matrix effects in dilute solutions. The affinity of the cork for the analyte substantially improved the sensitivity for the evaluated analytes. Furthermore, the proposed approach is precise, with an LOD of 1.05 µM. Compared to other analytical methods, this approach is reproducible and less expensive, both in terms of time and materials.
The composite electrode could be also effectively used for the determination of HCQ in other media. However, as it pertains to the ideal physical and chemical properties that favor interactions with the analytes, more experiments are necessary to better understand the chemical/electrochemical processes that take place on the cork surface, the cork-absorbent interactions when a current is applied, or when the cork participates as a mediator. Finally, even if the LOD reported in this work (1.05 µM) is slightly above the detection limit established by other materials (see Table 2), there is room for improvements; for example, the size of the electrochemical sensor which could be reduced, the use of other carbon-based modifiers [21,22] which could enhance its selectivity and sensitivity, and consequently, improving the LOD. Finally, integrated environmentally-friendly electrochemical technologies [15] could be proposed for removing and detecting pollutants by using sensors and advanced oxidation/reduction processes [15,47,48] because it is possible to design small portable devices for monitoring pollutants before and after their elimination from water which benefit the use of specific strategies in real time. In fact, regarding the treatment of the river water sample in which was detected a HCQ contamination, reported in this work, it could be treated by electrochemical oxidation technology [15,48,49] and monitored by using the cork-graphite-based sensor proposed here.