Supporting Electrolyte Manipulation for Simple Improvement of the Sensitivity of Trace Vanadium(V) Determination at a Lead-Coated Glassy Carbon Electrode

The paper presents a very simple way to extremely improve the sensitivity of trace V(V) determination. The application of a new supporting electrolyte composition (CH3COONH4, CH3COOH, and NH4Cl) instead of the commonly used acetate buffer (CH3COONa and CH3COOH) significantly enhanced the adsorptive stripping voltammetric signal of vanadium(V) at the lead-coated glassy carbon electrode (GCE/PbF). A higher enhancement was attained in the presence of cupferron as a complexing agent (approximately 10 times V(V) signal amplification) than in the case of chloranilic acid and bromate ions (approximately 0.5 times V(V) signal amplification). Therefore, the adsorptive stripping voltammetric system with the accumulation of V(V)–cupferron complexes at −1.1 V for 15 s in the buffer solution (CH3COONH4, CH3COOH, and NH4Cl) of pH = 5.6 ± 0.1 was selected for the development of a simple and extremely sensitive V(V) analysis procedure. Under optimized conditions, the sensitivity of the procedure was 6.30 µA/nmol L−1. The cathodic peak current of V(V) was directly proportional to its concentration in the ranges of 1.0 × 10−11 to 2.0 × 10−10 mol L−1 and 2.0 × 10−10 to 1.0 × 10−8 mol L−1. Among the electrochemical procedures, the lowest detection limit (2.8 × 10−12 mol L−1) of V(V) was obtained for the shortest accumulation time (15 s). The high accuracy of the procedure was confirmed on the basis of the analysis of certified reference material (estuarine water) and river water samples.


Introduction
Vanadium is widespread in the earth's crust but in low abundance. It is widely used and released in a wide variety of industrial processes. Its trace amount is essential for normal cell growth but can be toxic if present at higher concentrations. Vanadium(V), which occurs as VO 2 + in an acidic solution and VO 4 3− in an alkaline solution, is expected to be the predominant form in waters exposed to atmospheric oxygen [1]. The concentration of vanadium(V) in natural waters ranges from 10 −9 and 10 −7 mol L −1 [2,3] and, therefore, powerful analytical techniques are required for vanadium analysis. Just a few techniques, such as high-performance liquid chromatography, calibration-free laserinduced breakdown spectroscopy, electrothermal atomic absorption spectrometry, neutron activation analysis, inductively coupled plasma atomic spectrometry, mass spectrometry, and stripping voltammetry (SV), can meet the challenge of vanadium trace analysis in environmental water samples [4][5][6][7][8][9][10][11]. Of these techniques, only the SV can be used in routine field vanadium analysis. Adsorptive stripping voltammetry (AdSV), based on the accumulation of vanadium complexes with various complexing agents (such as chloranilic acid [11][12][13][14], cupferron [15,16], gallic acid [17], alizarin red S [18], alizarin violet [19], 2,3-dihydrobenzaldehide [20], and quercetin-5-sufonic acid [21]) on the electrode surface, has proven to be especially useful for the trace determination of vanadium. As can be seen, the most commonly used complexing agents are chloranilic acid and cupferron.

Apparatus
Voltammetric studies were performed on a µAutolab electrochemical analyzer integrated with GPES 4.9 software (Eco Chemie, Utrecht, the Netherlands) and an electrode stand (M164D, MTM Anko, Krakow, Poland), in a three-electrode arrangement with a glassy carbon electrode (diameter of 1 mm) electrochemically coated with lead (GCE/PbF) as a working electrode, a platinum electrode as an auxiliary electrode, and Ag/AgCl (3 mol L −1 KCl) as a reference electrode. A µAutolab analyzer integrated with FRA 4.9 software was applied for electrochemical impedance spectroscopy (EIS) stud- ies. Silicon carbide paper (SiC-paper, #4000, Buehler, Skovlunde, Denmark), alumina particle suspension (1.0, 0.3, and 0.05 µm), and a Buehler polishing pad were used to prepare the GCE surface before a series of measurements. A UV digester, made by Mineral, Poland, was used for three-hour mineralization of the certified reference material water samples (SLEW-3, estuarine water, National Research Council Canada, Ottawa, ON, Canada) and river water samples (acidified to a pH of 2.0 with nitric acid, Vistula River, Sandomierz, Poland).

DPAdSV Procedure Parameters
The analysis of V(V) at the GCE/PbF was carried out by differential pulse adsorptive stripping voltammetry (DPAdSV) in a solution containing 10 mL of 0.3 mol L −1 buffer solution (CH 3 COONH 4 , CH 3 COOH, and NH 4 Cl) of pH = 5.6 ± 0.1, 5.0 × 10 −4 mol L −1 Pb(NO 3 ) 2, and 7.0 × 10 −4 mol L −1 cupferron. The DPAdSV parameters under optimized conditions of V(V) analysis at the GCE/PbF are collected in Table 1. The electrochemical cleaning step was performed at the potential of −1.1 V for 15 s and 0.2 V for 15 s. Then, lead film was deposited at the potential of −1.1 V (E dep. Pb ) for 15 s (t dep. Pb ), and the V(V)-cupferron complexes at −0.6 V (E acc. ) for 15 s (t acc. ) were accumulated. After the equilibrium period of 5 s, the DPAdSV curves were recorded in the potential range from −0.6 to −0.9 V. The background curve was subtracted, and the baseline was corrected for each voltammogram.

Results and Discussion
In our previous studies, we showed a simple way to amplify the analytical signal using a lead film electrode and a new supporting electrolyte composition (CH 3 COONH 4 , CH 3 COOH, and NH 4 Cl). This electrolyte was applied instead of the commonly used acetate buffer (CH 3 COONa and CH 3 COOH) in order to significantly enhanced the adsorptive stripping voltammetric signal of vanadium(V) at the lead-coated glassy carbon electrode (GCE/PbF). In order to establish the most suitable experimental conditions, the following optimization studies were performed: selection of supporting electrolyte and complexing agent, the influence of pH value of the supporting electrolyte, cupferron and Pb(II) concentrations, potential and time of lead deposition, and accumulation time of V(V)-cupferron complexes on the analytical signal V(V), as well as the effect of the DPV parameters (scan rate, amplitude, and modulation time) on the V(V) peak current.

Selection of Supporting Electrolyte and Complexing Agent
In this study, in order to improve the V(V) analytical signal at the GCE/PbF, a new supporting electrolyte of 0.3 mol L −1 (CH 3 COONH 4 , CH 3 COOH and NH 4 Cl) and pH 5.6 ± 0.1 was for the first time applied, instead of a 0.3 mol L −1 acetate buffer solution (CH 3 COONa and CH 3 COOH) of pH 5.6 ± 0.1, in the presence of cupferron as a complexing agent [15], as well as chloranilic acid and bromate ions [8]. Figure 1 shows a comparison of the obtained DPAdSV curves. The application of the new supporting electrolyte composition is an easy way to significantly enhance the V(V) analytical signal (0.29 vs. 3.01 µA in the presence of cupferron and 0.96 vs. 1.41 µA in the presence of chloranilic acid and bromate ions). It is related to the improvement in the conductivity of the supporting electrolyte, as we wrote in an earlier work [22]. To decide which adsorptive voltammetric system to choose (in the presence of cupferron or chloranilic acid and bromate ions) for further research, measurements were performed for low concentrations of V(V) in the range of 1 × 10 −10 to 2 × 10 −9 mol L −1 . Figure 2 shows the obtained DPAdSV curves and calibration graphs in the studied range of V(V) concentrations. As can be seen, the linear range of V(V) at the GCE/PbF in the presence of cupferron is wider and the sensitivity is approximately 10 times higher (from 1 × 10 −10 to 2 × 10 −9 mol L −1 with a sensitivity of 0.72 µA/nmol L −1 ) than in the presence of chloranilic acid and bromate ions (from 1 × 10 −10 to 1 × 10 −9 mol L −1 with a sensitivity of 0.054 µA/nmol L −1 ). Therefore, to develop a simple and extremely sensitive voltammetric procedure of V(V) determination using a GCE/PbF, the adsorptive voltammetric system in the presence of cupferron and a new buffer solution composition (CH 3 COONH 4 , CH 3 COOH, and NH 4 Cl) of pH 5.6 ± 0.1 were adopted.
(CH3COONa and CH3COOH) of pH 5.6 ± 0.1, in the presence of cupferron as a complexing agent [15], as well as chloranilic acid and bromate ions [8]. Figure 1 shows a comparison of the obtained DPAdSV curves. The application of the new supporting electrolyte composition is an easy way to significantly enhance the V(V) analytical signal (0.29 vs. 3.01 µA in the presence of cupferron and 0.96 vs. 1.41 µA in the presence of chloranilic acid and bromate ions). It is related to the improvement in the conductivity of the supporting electrolyte, as we wrote in an earlier work [22]. To decide which adsorptive voltammetric system to choose (in the presence of cupferron or chloranilic acid and bromate ions) for further research, measurements were performed for low concentrations of V(V) in the range of 1 × 10 −10 to 2 × 10 −9 mol L −1 . Figure 2 shows the obtained DPAdSV curves and calibration graphs in the studied range of V(V) concentrations. As can be seen, the linear range of V(V) at the GCE/PbF in the presence of cupferron is wider and the sensitivity is approximately 10 times higher (from 1 × 10 −10 to 2 × 10 −9 mol L −1 with a sensitivity of 0.72 µA/nmol L −1 ) than in the presence of chloranilic acid and bromate ions (from 1 × 10 −10 to 1 × 10 −9 mol L −1 with a sensitivity of 0.054 µA/nmol L −1 ). Therefore, to develop a simple and extremely sensitive voltammetric procedure of V(V) determination using a GCE/PbF, the adsorptive voltammetric system in the presence of cupferron and a new buffer solution composition (CH3COONH4, CH3COOH, and NH4Cl) of pH 5.6 ± 0.1 were adopted.

Step of the Optimization Procedure
To achieve an extremely sensitive voltammetric procedure of V(V) determination at the GCE/PbF, the individual parameters were optimized. The influence of pH value of the supporting electrolyte, cupferron and Pb(II) concentrations, potential and time of lead deposition, and accumulation time of V(V)-cupferron complexes on the analytical signal V(V) was tested. The pH value of the supporting electrolyte (pH of 5.6 ± 0.1) was selected on the basis of results obtained for 2.0 × 10 −9 mol L −1 V(V) ( Figure 3A). The effect of the pH buffer solution over the range from 3.5 to 5.9 was investigated. The results show that an increase in the acidity of the supporting electrolyte contributes to a deterioration of the V(V) signal. It is related to the shifting of the V(V) peak towards the less negative potential values with the increase in the acidity of the solution and the overlapping of the V(V) peak on the reduction Pb(II) peak. The variation of cupferron concentration (from 1.0 × 10 −4 to 9.0 × 10 −4 mol L −1 ) significantly affected the peak current of 2.0 × 10 −9 mol L −1 V(V) ( Figure  3B). In the subsequent measurements, a cupferron concentration of 7.0 × 10 −4 mol L −1 was used because the highest value of the V(V) peak current was attained. The Pb(II) concentration within the range of 0 to 1.0 × 10 −3 mol L −1 significantly affected the peak current of

Step of the Optimization Procedure
To achieve an extremely sensitive voltammetric procedure of V(V) determination at the GCE/PbF, the individual parameters were optimized. The influence of pH value of the supporting electrolyte, cupferron and Pb(II) concentrations, potential and time of lead deposition, and accumulation time of V(V)-cupferron complexes on the analytical signal V(V) was tested. The pH value of the supporting electrolyte (pH of 5.6 ± 0.1) was selected on the basis of results obtained for 2.0 × 10 −9 mol L −1 V(V) ( Figure 3A). The effect of the pH buffer solution over the range from 3.5 to 5.9 was investigated. The results show that an increase in the acidity of the supporting electrolyte contributes to a deterioration of the V(V) signal. It is related to the shifting of the V(V) peak towards the less negative potential values with the increase in the acidity of the solution and the overlapping of the V(V) peak on the reduction Pb(II) peak. The variation of cupferron concentration (from 1.0 × 10 −4 to 9.0 × 10 −4 mol L −1 ) significantly affected the peak current of 2.0 × 10 −9 mol L −1 V(V) ( Figure 3B). In the subsequent measurements, a cupferron con-  Accumulation potential and time are always important factors since they affect the linear range of the calibration graph and detection limit of the voltammetric procedure. The optimal potential of accumulation (Eacc.) of V(V)-cupferron complexes onto the GCE/PbF surface in the buffer solution (CH3COONH4, CH3COOH, and NH4Cl) of pH 5.6 ± 0.1 was −0.6 V. For the Eacc. lower than −0.6 V and higher than −0.6 V, the V(V) signal decreased significantly. It is connected with a narrow range of available potentials between the Pb(II) and V(V) reduction signals. Additionally, the impact of accumulation time of V(V)-cupferron complexes (tacc.., from 0 to 45 s) on the peak currents of 5.0 × 10 −10 mol L −1 V(V) was investigated ( Figure 4C). As can be seen, the highest V(V) peak current was achieved for 15 s and hence this value was chosen as optimal. The investigation of the effect of Pb deposition potential (E dep. Pb , from −1.1 to −1.4 V) on the values of 2.0 × 10 −9 mol L −1 V(V) peak current showed that the highest response was achieved at a potential of −1.1 V and therefore this value of E dep. Pb was selected for further studies ( Figure 4A). Furthermore, the influence of Pb deposition time (t dep. Pb , from 0 to 30 s) on the values of 2.0 × 10 −9 mol L −1 V(V) peak current was studied ( Figure 4B). The highest V(V) responses were obtained within t dep. Pb in the range of 10-20 s; the value of 15 s was chosen for further studies.
Accumulation potential and time are always important factors since they affect the linear range of the calibration graph and detection limit of the voltammetric procedure. The optimal potential of accumulation (E acc .) of V(V)-cupferron complexes onto the GCE/PbF surface in the buffer solution (CH 3 COONH 4 , CH 3 COOH, and NH 4 Cl) of pH 5.6 ± 0.1 was −0.6 V. For the E acc . lower than −0.6 V and higher than −0.6 V, the V(V) signal decreased significantly. It is connected with a narrow range of available potentials between the Pb(II) and V(V) reduction signals. Additionally, the impact of accumulation time of V(V)cupferron complexes (t acc.., from 0 to 45 s) on the peak currents of 5.0 × 10 −10 mol L −1 V(V) was investigated ( Figure 4C). As can be seen, the highest V(V) peak current was achieved for 15 s and hence this value was chosen as optimal. The effect of the DPV parameters (scan rate (ν), amplitude (ΔEA), and modulation time (tm)) on the 2.0 × 10 −9 mol L −1 V(V) peak current was examined. For tm of 2 ms and ν of 20 mV s −1 , the ΔEA was changed from 25 to 125 mV. The best results were obtained for the ΔEA of 100 mV. The ΔEA higher than 100 mV caused a major increase in the background current. Next, the dependence of the ν values, ranging from 10-100 mV s −1 , on the 2.0 × 10 −9 mol L −1 V(V) signal was studied. The highest V(V) signal was found at the ν value of 20 mV s −1 , so this value was used for subsequent experiments. Furthermore, the tm was varied from 2 to 10 ms. For the tm of 2 ms, the highest 2.0 × 10 −9 mol L −1 V(V) peak current was achieved and therefore this value was selected as optimal.

Electrochemical Characteristics of the Sensor
In the solution of 0.1 mol L −1 KCl containing 5 × 10 −3 mol L −1 K3[Fe(CN)6] and 0 (for GCE) and 3.0 × 10 −4 mol L −1 Pb(II) (for GCE/PbF), electrochemical characteristics were investigated for the lead-coated and bare glassy carbon electrode using cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS). The CV curves were registered in the υ range of 5.0-500 mV s −1 and they showed a pair of well-shaped redox peaks of (Fe(CN)6) 3−/4− at the bare GCE and GCE/PbF ( Figure 5A for υ of 100 mV s −1 ). However, the signal intensity at the modified electrode was much better (anodic: 3.52 vs. 5.95 µA, cathodic: 3.88 vs. 5.49 µA, respectively). Moreover, in the case of the GCE/PbF a new peak appeared at a potential of −0.42 V, which is related to the oxidation of the deposited lead. Covering electrochemically the electrode surface with lead causes an acceleration of electron transfer kinetics and contributes to increasing the active surface of the electrode. The relative peak separation (χ°) was equal to 1.66 for the GCE/PbF and 5.97 for the GCE. The χ° value for the GCE/PbF was closer to the theoretical value of 1.0. On the other hand, the electrochemically active electrode area (As) of the GCE and GCE/PbF was calculated to be 0.19 and 0.46 mm 2 , respectively ( Figure 5B). Moreover, the GCE/PbF is characterized by a lower charge transfer resistance (Rct) (28.7 vs. 9.3 Ω) and good conductivity, which was found on the basis of the EIS measurements ( Figure 5C). The Nyquist plots were registered at a potential of 0.2 V and in a frequency range from 50 kHz to 1 Hz. The effect of the DPV parameters (scan rate (ν), amplitude (∆E A ), and modulation time (t m )) on the 2.0 × 10 −9 mol L −1 V(V) peak current was examined. For t m of 2 ms and ν of 20 mV s −1 , the ∆E A was changed from 25 to 125 mV. The best results were obtained for the ∆E A of 100 mV. The ∆E A higher than 100 mV caused a major increase in the background current. Next, the dependence of the ν values, ranging from 10-100 mV s −1 , on the 2.0 × 10 −9 mol L −1 V(V) signal was studied. The highest V(V) signal was found at the ν value of 20 mV s −1 , so this value was used for subsequent experiments. Furthermore, the t m was varied from 2 to 10 ms. For the t m of 2 ms, the highest 2.0 × 10 −9 mol L −1 V(V) peak current was achieved and therefore this value was selected as optimal.

Electrochemical Characteristics of the Sensor
In the solution of 0.1 mol L −1 KCl containing 5 × 10 −3 mol L −1 K 3 [Fe(CN) 6 ] and 0 (for GCE) and 3.0 × 10 −4 mol L −1 Pb(II) (for GCE/PbF), electrochemical characteristics were investigated for the lead-coated and bare glassy carbon electrode using cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS). The CV curves were registered in the υ range of 5.0-500 mV s −1 and they showed a pair of well-shaped redox peaks of (Fe(CN) 6 ) 3−/4− at the bare GCE and GCE/PbF ( Figure 5A for υ of 100 mV s −1 ). However, the signal intensity at the modified electrode was much better (anodic: 3.52 vs. 5.95 µA, cathodic: 3.88 vs. 5.49 µA, respectively). Moreover, in the case of the GCE/PbF a new peak appeared at a potential of −0.42 V, which is related to the oxidation of the deposited lead. Covering electrochemically the electrode surface with lead causes an acceleration of electron transfer kinetics and contributes to increasing the active surface of the electrode. The relative peak separation (χ • ) was equal to 1.66 for the GCE/PbF and 5.97 for the GCE. The χ • value for the GCE/PbF was closer to the theoretical value of 1.0. On the other hand, the electrochemically active electrode area (A s ) of the GCE and GCE/PbF was calculated to be 0.19 and 0.46 mm 2 , respectively ( Figure 5B). Moreover, the GCE/PbF is characterized by a lower charge transfer resistance (R ct ) (28.7 vs. 9.3 Ω) and good conductivity, which was found on the basis of the EIS measurements ( Figure 5C). The Nyquist plots were registered at a potential of 0.2 V and in a frequency range from 50 kHz to 1 Hz.

Calibration Graph, Repeatability, and Reproducibility
In the optimal conditions (0.  Figure 6. The detection (LOD) and quantification (LOQ) limits were estimated to be 2.8 × 10 −12 and 9.3 × 10 −12 mol L −1 , respectively, using the LOD = 3SDa/b and LOQ = 10 SDa/b equations (SDa-standard deviation of intercept (n = 3); bslope of calibration curve) [25]. The peak current standard deviation values for all concentrations of V(V) from the calibration graph in the range of 0.14-4.2 % (n = 3) confirmed satisfactory signal repeatability. Moreover, three GCE/PbF were prepared independently and used for the determination of 2.0 × 10 −9 mol L −1 V(V). The RSD of 5.5% (n = 9) confirms the acceptable reproducibility of a new sensor. Table 2 shows comparison of voltammetric procedures for the analysis of V(V). It can be seen that the voltammetric procedure described in this article offers the lowest detection limit for the lowest accumulation time.

Calibration Graph, Repeatability, and Reproducibility
In the optimal conditions (0.3 mol L −1 buffer solution (CH 3 COONH 4 , CH 3 Figure 6. The detection (LOD) and quantification (LOQ) limits were estimated to be 2.8 × 10 −12 and 9.3 × 10 −12 mol L −1 , respectively, using the LOD = 3SD a /b and LOQ = 10 SD a /b equations (SD a -standard deviation of intercept (n = 3); b-slope of calibration curve) [25]. The peak current standard deviation values for all concentrations of V(V) from the calibration graph in the range of 0.14-4.2 % (n = 3) confirmed satisfactory signal repeatability. Moreover, three GCE/PbF were prepared independently and used for the determination of 2.0 × 10 −9 mol L −1 V(V). The RSD of 5.5% (n = 9) confirms the acceptable reproducibility of a new sensor.     Table 2 shows comparison of voltammetric procedures for the analysis of V(V). It can be seen that the voltammetric procedure described in this article offers the lowest detection limit for the lowest accumulation time.

Selectivity Studies
To check the proposed procedure selectivity, the 2.0 × 10 −9 mol L −1 V(V) signals were registered in the presence of potential interferents. It was found that a 1000-fold excess of Ca(II), Mg(II), Bi(III), and Cu(II) as well as a 100-fold excess of Ni(II), Fe(III), and Cd(II) did not change the peak current of V(V) by more than 5%. According to the literature data [26], natural waters contain surfactants with a surface-active effect corresponding to 0.2-2.0 ppm of Triton X-100; therefore, the effect of the Triton X-100 presence in the solution on the 2.0 × 10 −9 mol L −1 V(V) signal was studied. It was found that 0.5 ppm of Triton X-100 caused a suppression of the V(V) signal to 90% of its original value. The presence of organic sample components that can block the electrode surface is a major limitation of the practical application of voltammetric techniques. In the case of the proposed procedure, this problem was solved by the decomposition of the organic matrix during UV mineralization of the samples.

Samples Analysis
The practical application of the elaborated DPAdSV procedure was evaluated by quantification of the total V in UV-mineralized samples of certified reference material (SLEW-3, estuarine water) and Vistula River water samples using the standard addition method. The voltammetric results are summarised in Table 3. No significant difference was found between the determined concentration and the certified quantity value (the relative error of 1.8%). As can be seen, V(V) was determined in the Vistula River samples at a concentration of 4.7 × 10 −9 mol L −1 . The recovery values were between 98.6 and 105.2%, which confirms a satisfactory degree of accuracy of the procedure.

Conclusions
The article presents a simple, sensitive, and selective voltammetric procedure for the determination of V(V) using a new supporting electrolyte composition (CH 3 COONH 4 , CH 3 COOH, and NH 4 Cl) and the lead-coated glassy carbon electrode (GCE/PbF) as a working electrode. The results showed that the application of a new buffer solution (CH 3 COONH 4 , CH 3 COOH, and NH 4 Cl), instead of the commonly used acetate buffer (CH 3 COONa and CH 3 COOH), contributed to a high increase in the V(V) signal at the GCE/PbF in the presence of cupferron as a complexing agent as well as by using the catalytic system of V-chloranilic acid-bromate ions. However, the highest sensitivity can be obtained by using the new supporting electrolyte, the GCE/PbF sensor, and cupferron as a complexing agent. Such a simple electrolyte change allows us to obtain the lowest detection limit of V(V) (2.8 × 10 −12 mol L −1 for an accumulation time of 15 s) in comparison with the previously described voltammetric procedures at mercury and less toxic electrodes. It is worth emphasizing that lead and lead salts are toxic but less toxic and less volatile than mercury and mercury salt used for the preparation of mercury electrodes. Therefore, the application of the proposed procedure of V(V) determination at the GCE/PbF sensor in the new supporting electrolyte allows us to eliminate mercury and mercury-salt waste. This study represents an additional step towards the replacement of mercury electrodes in adsorptive stripping voltammetric analysis of metal ions. The attractive behaviour of V(V) in a new supporting electrolyte composition indicates great promise for developing SV procedures of other metal ions and biologically active compound analysis in a manner similar to that reported for mercury electrodes.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

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