Determination of Cocaine by Square Wave Voltammetry with Carbon Paste Electrodes

We compare the electrochemical behavior of cocaine hydrochloride (in acidic medium) and its free base form (in acetonitrile) by a simple, cheap, and fast square wave voltammetry method for cocaine analysis based on carbon paste electrodes without chemical modification. The electrodes performed better than the electrodes obtained for analysis on commercial screen-printed electrodes, which we also tested here. We conducted the analyses in aqueous solution containing 0.1 mol·L NH4ClO4 as supporting electrolyte. For cocaine in acidic medium, the linear correlation coefficient, the LOD, and the LOQ were 0.996, 4.66 10 mol·L, and 1.55 10 mol·L, respectively. For cocaine in acetonitrile medium, the linear correlation coefficient, the LOD, and the LOQ were 0.994, 9.77 10 mol·L, and 3.26 10 mol·L, respectively. The specificity of the methodology is advantageous when the response of different interfering substances analyzed in this work (lidocaine, procaine, caffeine and phenacetine) is concerned.


Introduction
Among the several currently available psychotropic substances, cocaine is one of the most often used drugs both in Brazil and worldwide. In Brazil, cocaine users correspond to 2.9% of the country's population (2005) 1 . In global terms, cocaine users make up 0.4% of the total world population (2015) 2 .
The main issues attached to drugs go beyond health: trafficking troubles and organized crime are abundant and family members engage in conflicts, not to mention problems related to addiction livelihood, violence, etc. 3 . To protect the population and to ensure safety, drug control policies exist. In Brazil, the National Anti-Drug Policy (PNAD), decreed in 2002, establishes objectives and guidelines to develop strategies for prevention, treatment, recovery and social reinsertion, health and social damage reduction, trafficking repression and studies, as well as research and evaluation due to improper drug use 4 .
Cocaine is a central nervous system stimulant that can be extracted from Erythroxylum coca leaves 5 , mostly found in South America. This drug is commonly found in two main forms ( Figure 1): as a salt (a crystalline powder that is generally inhaled or dissolved and injected) and as an alkaloid (which is volatile at low temperatures and is usually smoked; e.g., crack cocaine) 6

. Despite constant decline in
Erythroxylum coca cultivation, this illicit drug is still present in the society and causes serious damage 7 . Given the negative consequences of cocaine (and other drugs), drug control has an essential role in preserving social order. Therefore, Civil and Federal Polices need drug determination methods that are not only reliable (once the analysis result will mean someone's sentence), but also practical, fast, and cheap, allowing a larger number of analyses and meeting the high police demand.
The literature contains descriptions of many methods for analysis of cocaine [8][9][10][11][12][13][14][15][16][17][18][19][20] and its metabolites 10,12,14,18,19,21 . The reported methods have advantages and disadvantages. Voltammetry is worth highlighting: it is practical and relatively reliable; it is not a definitive test like some spectroscopic and spectrometric methods, but it is more reliable than regularly employed colorimetric tests 8 . Besides being cheaper (potentiostats cost less and demand cheaper solvents) and faster than definitive tests, voltammetry is simple, sensitive, specific, and easy to operate, and it can be adopted for qualitative and quantitative purposes, as well as for elucidation of the action mechanisms of drugs 22,23 .
Voltammetry is an electroanalytical method during which a potential variation is applied to the system; an electric current is detected as response. In this method, a three-electrode system is required: the working electrode (where the main redox reaction occurs), a reference electrode, and the auxiliary electrode (the current generated by redox reaction flows between the auxiliary electrode and the working electrode) 24 . Finally, a supporting electrolyte, which is a non-reactive salt dissolved in the medium, is necessary because it accounts for the current flow and hence decreases the solution resistance.
In the forensic area, voltammetry is widely used to analyze seized drugs, biological matrixes, contaminated water, explosives, gunshot residues (GSR), and toxins [25][26][27][28][29][30] . Chemically modified electrodes are often employed, as well. Cheaper solvents and reagents, faster methods, and portable systems (for in loco analyses) are some of the advantages of voltammetric methods. The need for modifiers, extreme pH, and toxic reagents and solvents may constitute disadvantages of these methods. All the advantages and disadvantages should be considered when choosing the best method for a certain analysis.
We propose a simple and cheap methodology to quantify cocaine in acidic solution and in organic solution by using carbon paste electrode.

Apparatus and chemicals
All chemical reagents were of analytical grade, and all aqueous solutions were  After each voltammetric analysis, the working and the auxiliary electrodes were polished with paper. To perform the analyses, a potential range from 0 to 1.5 V was applied; the amplitude and the frequency were 0.025 V and 25 Hz (no preconcentration), respectively, for cocaine in acetonitrile. For cocaine in acidic medium, a potential range from 0 to 1.5 V was applied; the amplitude and the frequency were 0.6 V and 30 Hz, respectively. Pre-concentration was accomplished at 0.8 V for 10 s.
The cocaine standard and the cocaine samples were obtained from a scientific partnership between this research group and the laboratory of toxicological analysis -Institute of Criminalisticscity of Ribeirão Pretostate of São Paulo, Brazil.
Analysis of interfering substances involved addition of standard caffeine, lidocaine, procaine and phenacetine solutionsall the standards, in the powder form, were purchased from Sigma-Aldrich. These interferents were dissolved in the same solvents as cocaine.

Samples and solutions
A 2.5 10 -3 mol·L -1 cocaine hydrochloride solution in acidic medium was prepared by addition of HCl, until pH 3. A 2.5 10 -3 mol·L -1 cocaine free base solution in acetonitrile was also prepared. Cocaine standard, 96% purity, was employed in both cases. The linear dependence of cocaine detection was carried out in concentrations from 9.96 10 -6 mol·L -1 to 7.43 10 -5 mol·L -1 .
Cocaine solutions of confiscated samples were prepared by pre-treating small aliquots of the water-soluble cocaine chlorhydrate samples with aqueous sodium bicarbonate solution to remove HCl from its chlorhydrate form. Then, chlorhydrate-free cocainewhich is insoluble in waterwas removed by filtration, rinsed with deionized water, dried, and dissolved in acetonitrile.
For both samples of cocaine (in acid medium and in acetonitrile), a 0.1 mol·L -1 NH4ClO4 aqueous solution was utilized as supporting electrolyte with a 0.05 mol·L -1 K2HPO4 buffer solution to carry out the analysis. The difference between the analysis of cocaine in acid medium and in acetonitrile was the pH of 11 and 6, respectively.

Manufacture of carbon paste electrodes
The carbon paste electrode was constructed with a disposable syringe containing a copper piston (as electrical contact). The syringe tip was cut and sanded, to provide a smooth and circular surface. Then, the carbon paste was inserted into the syringe and compacted by pressing the piston against a planar surface.
To prepare the carbon paste, the desired graphite and paraffin ratio was weighed and mixed in a porcelain mortar with a pestle. The mortar was heated at 50 ºC for 20 min on a hot plate, under agitation, until the mixture melted and became homogeneous.

Calculation of analytical parameters
The analytical curve provided the analytical parameters that allowed us to evaluate method sensitivity and limits. We obtained the LOD (limit of detection) and LOQ (limit of quantification) values by using 3*SD/m and 10*SD/m, respectively, where SD is the standard deviation, and m is the curve amperometric sensitivity 33 .
From the analytical curve equation, we have the sensibility as the angular coefficient and the standard deviation (SD) as the linear coefficient.

Results and discussion
According to the literature, cocaine is an electroactive substance with oxidation peak around 1.0 V (depending on the electrode). This peak is associated with oxidation of the tertiary amine present in its structure 34,35 .

Cocaine in acidic medium
We recorded the response of the cocaine solution in acidic medium during square wave voltammetry in different electrolytes, at pH 11. We tested perchlorate salts such as lithium, potassium, and ammonium perchlorates as well as potassium chlorate and potassium nitrate as supporting electrolytes. Ammonium perchlorate afforded the most intense peak, followed by potassium and sodium perchlorates, which gave similar results, and then lithium perchlorate, potassium chlorate, and potassium nitrate. We therefore selected ammonium perchlorate as supporting electrolyte for the subsequent analyses.

Kinetic study
The cyclic voltammogram revealed an anodic peak at 1.15 V (Figure 3)

Analytical curve construction
Square Wave Voltammetry is a more sensitive technique than Cyclic Voltammetry. We observed the same oxidation peak around 1.08 V as a direct response for the presence of cocaine. Thus, we chose the Square Wave Voltammetry technique to quantify the analyte.
We constructed the analytical curve ( Figure 4A) by successively adding cocaine in acidic medium to the electrochemical cell. The anodic peak emerged at approximately 1.08 V. Figure

Analytical curve construction
Once the kinetic study provided the same results for cocaine in acidic medium, we    In organic medium, the signal observed for cocaine shifted approximately 0.1 V toward higher potential as compared to cocaine in acidic medium; the peak current remained unaltered. As previously reported in the literature 33 , the mechanism of cocaine oxidation on carbon paste surfaces and in aqueous medium with pH ≥ 6 involves oxidation of the cocaine tertiary amine group. For the cocaine standard in acidic medium, the tertiary amine group is protonated (higher energy barrier required for oxidation) 36 , but when the cocaine standard is added to the basic electrolyte medium, the acid form of cocaine is neutralized, facilitating tertiary amine oxidation.
Thus, for the cocaine standard in organic medium, the energy barrier required for tertiary amine oxidation increases, and the oxidation potential is displaced toward a higher value, as observed in this work.

Study of interfering substances
We tested the following possible interferers: lidocaine, procaine, caffeine, theobromine, and phenacetine, both in acetonitrile and in acidic medium. Analysis of the voltammograms ( Figure 6) revealed that the peaks due to the interfering did not arise at the potentials of interest, 1.20 V for cocaine in acetonitrile and 1.08 V for cocaine in acidic medium, so these compounds should not affect cocaine detection by the proposed method. We then conducted a similar study to verify possible interactions between cocaine and these interfering substances when they were all in the same solution. We detected cocaine with unaltered voltammetric profile in all tests, so we discarded interference of these substances. This test proved that the proposed method, between the tested interfering, was specific for cocaine determination in acetonitrile medium or acidic medium.

Analytical parameters
The LOD values were 1.58 mg·L -1 in acidic medium and 2.96 mg·L -1 in acetonitrile. If we consider about 0.1 g of a seized sample in one milliliter of solution, the calculated values were equivalent to detection in the order of 0.003%, which was less than the purity in the samples and enable cocaine detection in them.
The values obtained here pointed to the good sensibility of the technique: they Comparison of the analytical parameters with literature parameters (Table 2) demonstrated that our limits of detection were comparable to literature values in most cases and were in the same order of µmol·L - 1 25,26,36 . Although a few reports describe higher limits of detection, the method proposed herein was based on a cheap and easy to prepare working electrode without modification that offered low cost, highly reproducible and sensitive cocaine determination for analysis of seized samples.

Analysis of seized samples
The analysis of seized samples (Figure 7) has confirmed the presence of cocaine, what was confirmed with a GC-MS. By the presented method, the potential of cocaine oxidation shifted by 0.1 V, to 1.3 V, probably due to the presence of multiple interfering that were not tested in the interfering study section, but which are very common in such samples. Although the detection has been successful, the quantification was not possible, perhaps due the presence of multiple interfering. The values of quantification, calculated through the analytical curve, were inaccurate, but always lower than the real.

Analysis with screen-printed electrodes (SPE)
SPE have been widely explored in electrochemical methodologies due to characteristics like portability and simplicity. However, some changes in the electrochemical behavior of the target analyte might be expected as compared to conventional cells. We performed the cocaine analysis with this kind of electrode and obtained an analytical curve with poor linearity when cocaine was dissolved in acetonitrile. SPE just provided a qualitative test for cocaine detection.
For cocaine analysis in acidic medium, we obtained a linear analytical curve with SPE. However, the conventional system based on carbon paste electrode afforded better linearity and limit of detection.

Conclusion
The present methodology constitutes a low-cost system to detect and to quantify cocaine as in its two chemical forms: salt and base. In both acetonitrile and acidic medium, sample preparation proved to be a simple process. By using the square wave voltammetry technique, a specific result was available in less than three minutes.
When it comes to real samples, the detection is also possible, even with the presence of multiple interfering, confirming the effectiveness of the method; although the quantification becomes inaccurate. This effectiveness is seen as the peak at the characteristic potential.
Comparing between the solvents employed in this study and given the amperometric intensity of the voltammetric peak, we achieved better results for cocaine in acidic medium. Nevertheless, both methods are important, because the two forms of cocaine are widely seized by the police. Therefore, we can analyze both forms of cocaine by just preparing the sample solutions correctly and changing the voltammetric parameters at the postentiostat.