Introduction

Cadmium (48Cd) is a heavy metal found in different commercial items [1] as well as in soils [2,3,4]. It has a long biological half-life and is highly toxic for human and living organisms [5, 6]. Consequently, it is crucial to monitor the levels of Cd in aquatic environments, soils, and commercial products. Considering drinking water, the concentrations of Cd must remain below 3 μg L−1 [7]. In aqueous solutions, Cd2+ ions are the unique stable species at pH < 8. At pH > 8, the amount of Cd2+ drastically decreases, and Cd(OH)+ ions together with Cd(OH)2 species occur. In highly alkaline solutions (pH > 12), the amount of Cd(OH)+ ions is negligible, and hydroxyl species (Cd(OH)3−, Cd(OH)42−) becomes abundant [8]. Cd ions as well as heavy metal cations can be successfully monitored by spectrometric techniques [9, 10]. However, despite their high sensitivity, these techniques are costly and not easily deployable outside the laboratory setting [11]. Thereby, special attention has been devoted to the use of electrochemical sensors because of their advantages: on-site detection, less bulky, rapid response, high sensitivity, and low cost [9, 12]. The electrode is one of the main components of these sensing systems as the electronic signal from such a sensor is sensitive to the nature and the quantity of the target analyte. Electrochemical devices with ion-selective electrodes have been proven efficient for monitoring pH, among others [13]. Carbon paste-paraffin oil electrodes have been successfully used for heavy metal detection [14]. But, despite their good electrical conductivity and electrochemical properties, these electrodes are less selective and have limited sensitivity. In order to surmount these limitations, the carbon paste was mixed with materials characterized by good electrical, chemical, and mechanical properties [15,16,17]. Recently, it has been shown that nanomaterial-modified sensors are convenient for monitoring different chemical hazardous species, including heavy metals [18, 19]. The suitability of the nanomaterials, such as metal/metal oxide, synthesized polymers, carbon, and nanocomposites, is attributed in large part to their large surface/volume ratio and high surface reactivity. Because of their nanosized layers, chemical stability, lamellar structure, large specific surface area, and ability to exchange cations [20], smectite clay minerals are convenient materials for electrochemical sensor construction [21, 22]. As against to conventional electrodes, smectite-containing electrodes have been considered to be more efficient in electrocatalysis, electroanalysis, and preconcentration [23, 24]. Referring to the studies dealing with clay-based electrodes as sensing elements for heavy metal detection [25,26,27], the interaction between the studied metal species and these kinds of electrodes occurred by adsorption, intercalation, and/or cation exchange. Considering the published studies, which addressed the Cd cation detection by clay-modified electrode [28, 29], we noted that much less attention has been paid to the study of the operating factors such as solution pH, bentonite content, analyte concentration, electrolytes, and sweep rate on the electrochemical behavior of bentonite-containing sensors. In addition to its fundamental aspect, such a study could allow to maximize the sensor performances. On the other hand, there is a lack of studies which explained the electrochemical process in relation with the clay physico-chemical properties, i.e., cation exchange capacity (CEC). In addition, studies dealing with the regeneration and reuse of clay-modified electrodes have been ignored. Overall, no comprehensive electrochemical study involving Cd2+ ions and bentonite-carbon paste electrodes has been reported.

In light of the limitations of existing analytical techniques and the above-mentioned dearth of research on electrochemical methods for detecting cadmium (Cd), this study aimed to investigate the application of bentonite-modified carbon paste electrodes. For this purpose, the effects of the experiment operating factors, the regeneration and potential reuse of the electrodes, and their selectivity were investigated using cyclic voltammetry, square wave voltammetry, and electrochemical impedance spectroscopy.

Materials and methods

Identification and physical characteristics of the used clay

The clay material used was from the bentonite-rich area in the vicinity of the city of Nador (Morocco). It was ground and sieved (< 100 μm). Then, it was washed with a buffer solution (pH = 4.5) and sodium-saturated. The X-ray diffraction and the Fourier-transform infrared analyses of the raw and the Na-saturated clay (Fig. 1) showed that the clay was composed of smectite di-octahedral clay minerals (montmorillonite/beidellite), plagioclase, and quartz. The clay mineral amount exceeded 80 mass%. The chemical composition, the CEC, and the external surface area (SA) are given in Table 1. The chemical composition was determined by X-ray fluorescence spectrometry using an Axios-type spectrometer. The CEC value was measured by using cobaltihexamine cation ([Co(NH3)6]3+) and following the procedure adopted by Aran et al. [29]. The external surface area was estimated by the BET (Brunauer-Emmett-Teller) method [30]. The somewhat high values of CEC and SA supported the abundance of the smectite clay minerals [31]. The iron and magnesium presence was taken as an indication of a partial substitution of the octahedral Al cations. The clay mineral charge compensation was achieved by sodium cations, among others.

Fig. 1
figure 1

X-ray diffraction patterns (A) and FT-IR spectra (B) of the raw and Na-saturated bentonite. Sme smectite, Pl plagioclase, Sil silicate, W water, Q quartz

Full size image
Table 1 Chemical composition (mass%), cation exchange capacity (CEC), and external specific surface area (SA) of the studied bentonite
Full size table

Electrode construction procedure

The main steps followed for the bentonite-modified carbon paste electrode construction are indicated in Fig. 2. In relation with the bentonite addition, it should be noted the use of high amounts (> 20 mass% bentonite) led to the electrode peeling off. For the free-bentonite electrode construction, dried graphite powder (purity, 99%) and paraffin oil (85:15 mass/mass) were mixed and introduced in a PTFE carbon tube together with a wire conductor, as shown in Fig. 2. To determine the bentonite optimal amount, preliminary cyclic voltammetry tests were performed on the electrodes containing up to 20% bentonite. As can be deduced from Fig. 3, the detection sensitivity for Cd ions was better with the electrode composed of 20 mass% bentonite.

Fig. 2
figure 2

Main steps followed for the electrode construction

Full size image
Fig. 3
figure 3

Cyclic voltammograms obtained with bentonite-carbon paste electrodes composed of different amounts of bentonite. Cd2+ ion concentration, 10−3 M; electrolyte: HCl; pH = 2; sweep rate, 10 mV/s

Full size image

Supporting electrolytes and experimental conditions

The electrochemical behavior of the constructed electrodes regarding cadmium cation detection was investigated in acid (HCl, H2SO4) media and salt (NaCl, KCl)-containing aqueous solutions. The concentration of the supporting electrolytes was kept constant (0.1 M), and the initial concentration of Cd2+ ions was 10−3 M. Cadmium sulfate (purity, 99%) was used as Cd2+-bearing compound.

The influence of the change of the analyte concentration (dynamic range, 5.10−6–10−3 M) on the electrochemical behavior of the electrodes was investigated in the following conditions: pH = 2, electrolyte: HCl. The effect of pH was investigated in the range of 2 to 9. In this respect, it could be recalled that Cd2+ ions are the unique stable species in aqueous solutions with pH < 8, and both Cd2+ and Cd(OH)+ are found at 8 < pH < 9 [32].

The effect of the sweep rate in cyclic voltammetry experiments was investigated in the 10–400 mV/s range. HCl (0.1 M) was used as a supporting electrolyte. The Cd2+ ion concentration was of 10−3 M.

The limit of detection (LOD) and the limit of quantification (LOQ) of the studied electrode were determined according to the following relations [33]:

$$\mathrm{LOD }= 3{S}_{\mathrm{B}}/{S}$$
(1)
$$\mathrm{LOQ }= 10{S}_{\mathrm{B}}/{S}$$
(2)

where SB is the standard deviation of 10 measurements taken from the signal obtained from the blank, and S is the slope of the calibration curve.

Electrochemical and characterization techniques

The electrochemical study was conducted using a potentiostat (CorrTest Electrochemical Workstation) connected to both a computer with CS Studio 5 software (for data processing) and an electrochemical cell. The cell was composed of the reference electrode (saturated calomel electrode—ECS), the working electrode (described in the “Electrode construction procedure” section) and a counter platinum electrode. The pH of the solution was monitored with a HI-2210 pH meter.

X-ray diffraction (XRD) analysis was performed on powder samples with a Rigaku SmartLab diffractometer operating with a copper anode (λKα = 1.5418 Å). Fourier-transform infrared (FT-IR) analysis was conducted on thin discs (1 mg of sample and 99 mg of KBr) in the range of 4000–400 cm−1 using a Perkin-Elmer Fourier-Transform 1720-x-spectrophotometer. The microscopic examinations were carried out with a VEGA3 LM TESCAN scanning electron microscope coupled with EDS system. The samples were coated with a thin layer of carbon.

Results and discussion

Effect of supporting electrolytes

Values of relevant electrochemical parameters, determined from the cyclic voltammograms shown in Fig. 4, are given in Table 2. The anodic and cathodic current data showed that the peak ratio ipa/ipc deviated from unity (> 1) (Table 2) and the potential differences ∆Ep (mV) were > 59.2/n (n is the number of electrons transferred). These results together with the shape of the cyclic voltammograms allowed the deduction that the charge transfer process was quasi-reversible and no diffusion barrier could be conceivable [34,35,36].

Fig. 4
figure 4

Cyclic voltammograms related to the detection of Cd2+ ions with the studied electrode immersed in different electrolytes. Cd2+ ion concentration, 10−3 M; scan rate, 10 mV/s

Full size image
Table 2 Values of some relevant experimental and calculated electrochemical parameters. The experimental values were drawn from the cyclic voltammograms shown in Fig. 4
Full size table

The values of the ratio of the diffusion coefficients of the reduced (Dr), and the oxidized (Do) forms, which were calculated from Eq. 3 [37], were insignificant (Dr/Do < 10–9) whatever the electrolyte used.

$${E}_{1/2}= E^\circ + {(RT/\mathrm{nF})\mathrm{Ln}({D}_{\mathrm{r}}/{D}_{\mathrm{o}})}^{1/2}$$
(3)

where E1/2 = (Epa + Epc)/2, E° is the standard potential of Cd2+/Cd at 25 °C (− 0.646 V/ECS), R is the ideal gas constant (R = 8.31 JK−1mol−1), F is the Faraday (96485.33 C/mol), and finally, n number of exchanged electrons (n = 2).

Therefore, the diffusion coefficient of the reduced form was largely negligible. It was thought that the effect of the supporting electrolytes on the electrochemical behavior of the electrode did not have a relation with the interaction between cadmium species and electrolyte ions. It was rather tied to the extent of the clay mineral protonation and the cation exchange process. In this connection, it is worthy to note that SEM examination of the electrode tested in HCl solution revealed the presence of Cl, Na, and Cd (Fig. 5). The unexpected presence of Cl could be taken as an indication of the electrical balance of the protonated clay minerals. The detection of Cd associated with a drastic decrease of Na (about 70%) allowed us to suppose the substitution of Na+ ions by Cd cations in all its forms [8, 38].

Fig. 5
figure 5

SEM micrographs and EDX spectra of the studied electrode taken before (a, zone 1) and after (b, zone 2) use

Full size image

Effect of electrolyte solution pH

As can be drawn from the cyclic voltammograms shown in Fig. 6, the anodic and the cathodic current densities decreased with the increase of pH, and the ratio ipa/ipc fell within the interval 1–1.3. Moreover, the charge transfer was quasi-reversible in acidic solutions, but it was almost reversible in the basic medium. In acid solutions, Cd2+ ions are the unique stable species and the clay minerals are seemingly protonated as follows:

Fig. 6
figure 6

Effect of the solution pH change on the shape of the cyclic voltammograms associated with Cd2+ ion detection with the studied electrode. Cd2+ ion concentration, 10−3 M; electrolyte: HCl; scan rate, 10 mV/s

Full size image
$$\equiv \mathrm{Al}-\mathrm{OH}\hspace{0.17em}+{\hspace{0.17em}\mathrm{H}}^{+}\hspace{0.17em}\to \hspace{0.17em}\equiv \mathrm{Al}-{\mathrm{OH}}_{2}^{+} \hspace{0.17em}\to \hspace{0.17em}\equiv {\mathrm{Al}}^{+} \hspace{0.17em}+\hspace{0.17em}{\mathrm{H}}_{2}\mathrm{O}$$

Cd2+ ions could not be adsorbed on positively charged clay particles. Therefore, the main portion of the cations in solution should get involved in the electron transfer process. In such a condition, the current density increased. The change of the anodic current density against pH (≥ 4) followed the relation ipa =  − 0.23pH + 3.10 with a good correlation (R2Adj = 0.999).

In alkaline solution, especially at pH = 9, both Cd2+ and Cd(OH)+species should be present and the clay minerals were presumably deprotonated according to the following mechanism:

$$\equiv \mathrm{Al}-\mathrm{OH}\hspace{0.17em}+\hspace{0.17em}{\mathrm{OH}}^{-}\hspace{0.17em}\to \hspace{0.17em}\equiv \mathrm{Al}-{(\mathrm{OH})}_{2}\!{}^{-}\hspace{0.17em}\to \hspace{0.17em}\equiv \mathrm{Al}-{\mathrm{O}}^{-} \hspace{0.17em}+\hspace{0.17em}{\mathrm{H}}_{2}\mathrm{O}$$

Therefore, Cd2+ and Cd(OH)+ ions were immobilized by the clay particles and could not be engaged in the electrochemical process. So, the current density decreased [39, 40].

Effect of the analyte concentration

The shape of the cyclic voltammograms obtained by operating with different Cd2+-containing solutions was markedly influenced by changes in the Cd2+ion concentration (Fig. 6). Moreover, the current signal intensity increased with the increase of Cd2+ ion concentration (Fig. 7). The broad range of the ratio ipa/ipc, 0.01–1.60, and the unsignificant Dr/Do ratios (5 × 10−12 < Dr/Do < 10−9), determined from the data retrieved from Fig. 7, showed how the electron and the mass transfers were affected by the change of Cd2+ion concentration.

Fig. 7
figure 7

Effect of the Cd2+ ion concentration change on the shape of the anodic and cathodic peaks. Electrode, 20 mass% bentonite-carbon paste; electrolyte: HCl; pH = 2; sweep rate, 10 mV/s

Full size image

The curve representing the variation of the anodic current density (ipa) against Ln(Co) (Co is the Cd2+ ion initial concentration) could be well decomposed into three branches (Fig. 8). In the Cd2+concentration range of 5 × 10−4–10−3 M, the current density variation followed the equation

Fig. 8
figure 8

Variation of the anodic current density against Ln(Co). Co: Cd2+ ion initial concentration

Full size image
$${i}_{\mathrm{pa}}= 3.98\mathrm{Ln}({C}_{\mathrm{o}}) + 31.08\;({{R}^{2}}_{\mathrm{Ad}\mathbf{j}}=0.88)$$
(4)

In the interval 5 × 10−5–5 × 10−4 M, it changed according to the relation

$${i}_{\mathrm{pa}}= 0.29\mathrm{Ln}({C}_{\mathrm{o}}) + 3.29\;({{R}^{2}}_{\mathrm{Ad}\mathbf{j}}= 0.99)$$
(5)

For low Cd2+ concentrations (< 5 × 10−5 M), the ipa variation was almost insignificant. In view of this result, the limit of detection should be slightly smaller than 5 × 10−5 M.

It was believed that the limit of detection was mainly limited by the bentonite cation exchange capacity (CEC). Indeed, taking into consideration the CEC value given in Table 1 and the bentonite mass used (20 mg), the Cd2+ amount, which could be involved in the CEC of the electrode, was estimated to be 6 × 10−6 mol. Thus, the amount of the Cd2+ ions in solutions with Co < 6 × 10−6 M should be entirely immobilized by the electrode. In such a condition, the electron transfer is annihilated, and consequently, the anodic/cathodic peak current vanished.

The values of LOD and LOQ, which were determined by using Eqs. (1) and (2), were found to be 4.23 × 10−7 M and 1.41 × 10−6 M, respectively. It is noteworthy to note that LOQ was close to the expected limit of detection (6 × 10−6 M) deduced on the basis of the CEC data.

Effect of potential sweep rate

The increase of the sweep rate led to the intensification of the anodic peak current together with a drastic drop of the cathodic current density (Fig. 9). So, the change of the sweep rate resulted in the electron transfer alteration. The current, which was presumably controlled by both the mass transport and the charge transfer at low scan rate, became determined by the electron transfer at high sweep rate. Otherwise, the electrochemical process, which was quasi-reversible at low-to-moderate sweep rates, became irreversible at ν> 50 mV/s. The change in the electrochemical reversibility was supported by the parabolic trend of the curve: ipa vs \({\nu }^{1/2}\) (Fig. 10). The disappearance of the cathodic peak current at ν > 50 mV/s was believed to be due to the immobilization of the Cd2+ ions by the clay minerals principally by complexation (outer-sphere complexation). It was thought that at high sweep rates, the electron transfer between Cd2+ ions and the electrode was very slow as compared to the Cd2+ ion immobilization rate, and consequently, the cathodic reaction could not take place.

Fig. 9
figure 9

Effect of the change of the potential scan rate on the shape of the cyclic voltammograms plotted by using the studied electrode. Cd2+ ion initial concentration, 10−3 M; electrolyte: HCl; pH = 2

Full size image
Fig. 10
figure 10

Variation of the anodic current density versus the square root of the sweep rate

Full size image

Equivalent circuit model of the electrode-solution interface

The Nyquist plots obtained for the bentonite-modified electrode and the carbon paste electrode in the presence of Cd2+ ions are given in Fig. 11. Considering these results, the electrochemical behavior of the electrodes/solution interfaces was equivalent to the electrical circuit shown in the inset of Fig. 11. Otherwise, the electrodes/electrolyte interfaces could be represented by the ratio R1 + (Q2/(R2 + W)) (R1 is the solution resistance, R2 is the charge transfer resistance, Q2 is the constant phase element, and W is the Warburg diffusion resistance [41]. The R2 values for the bentonite-modified and the carbon paste electrodes were determined to be 14,254 and 52,756 Ω, respectively. The low R2 value obtained for the bentonite-modified electrode was taken as an indication of high electron transfer between the solution and the studied electrode [42]. The Q2 values determined for the carbon paste and the bentonite-modified electrodes were \(0.141\times10^{-3}\; \mathrm{F\cdot s}^{(\text{a}-1)}\) and \(52.76\times10^{-6}\; \mathrm{F\cdot s}^{(\text{a}-1)}\), respectively. The low Q2 value was taken as an evidence for the accumulation of Cd cations on the surface of the bentonite-modified electrode [43, 44]. The W values for the carbon paste and the bentonite-modified electrodes were found to be 123.4 and 30.03 Ω/s−1/2, respectively. The higher W value for the carbon paste electrode suggested that the ions experienced some barrier to diffuse towards the electrode. In contrast, easier transfer of the ions occurred across the bentonite-modified electrode–electrolyte interface. In this respect, it is worth noting that the diffusion resistance Warburg feature manifested in the low-frequency region (< 35 Hz) of Nyquist diagrams (inset of Fig. 11).

Fig. 11
figure 11

Nyquist diagrams obtained by operating in the following conditions: electrodes: carbon paste and 20 mass% bentonite-carbon paste; electrolyte: HCl; pH = 2; Cd2+ ion concentration, 10−3 M

Full size image

Regeneration and reuse of the electrode

The cyclic voltammogram plotted by operating with Cd2+-free solution and using the bentonite-modified electrode, which has been already used for Cd2+ ion detection and immersed in NaCl solution (0.1 M) for 48 h, did not display any current signal associated with this metal (Fig. 12a). Conversely, the cyclic voltammogram obtained by using the Cd2+-loaded bentonite-modified electrode closely resembled that obtained under normal experimental conditions (Cd2+-containing solution, pristine bentonite-modified electrode) (Fig. 12b). These findings indicated that Cd cations were loosely tied to the electrode and they could be released by using NaCl solution.

Fig. 12
figure 12

Cyclic voltammograms obtained with different electrodes and by operating in Cd2+-free and Cd2+-containing solutions. a Regenerated and reused bentonite-modified electrode (Cd2+-free aqueous solution). b Cd2+-saturated bentonite-carbon paste electrode immersed in Cd2+-free solution, and pristine bentonite-modified electrode submerged in Cd2+-containing solution. pH = 2; electrolyte: HCl; sweep rate, 10 mV/s

Full size image

The cyclic voltammetry experiments conducted on the regenerated electrode displayed anodic and cathodic current peaks related to the studied metal (Fig. 13). As compared with the pristine electrode signals, the peak intensities were substantially lessened and the potential positions were partially shifted. Considering, for instance, the cyclic voltammogram related to the second cycle, the variations of the currents intensities and the potentials were \({\Delta i}_{\mathrm{pc}\hspace{0.17em}}=\hspace{0.17em}41\mathrm{\%}, {\Delta i}_{\mathrm{pa}}\hspace{0.17em}=\hspace{0.17em}52\mathrm{\%},{\Delta E}_{\mathrm{pc}}\hspace{0.17em}=\hspace{0.17em}17\mathrm{\%},{\Delta E}_{\mathrm{pa}}\hspace{0.17em}=\hspace{0.17em}15\mathrm{\%}\). The ratio ipa/ipc, which was obtained by using the data associated with the reused electrode, approached 1. Therefore, it was believed that the charge transfer changed from a quasi-reversible state to a reversible one, and it became relatively fast.

Fig. 13
figure 13

Cyclic voltammograms plotted by using pristine (cycle 1) and reused (cycle 2) bentonite-modified electrode. Cd2+ ion initial concentration, 10−3 M; sweep rate, 10 mV/s; electrolyte: HCl; pH = 2

Full size image

The bentonite-modified electrode was convenient for a simultaneous detection of Pb2+ and Cd2+ ions in aqueous solution, and it was successfully regenerated by immersion in NaCl solution (Fig. 14). The regenerated electrode was still effective for the ion detection (Fig. 14). However, the anodic current intensities related to Cd2+ and Pb2+ ions decreased of about 57 and 48%, respectively. The remaining electrochemical parameter change was less than 10%. It appeared that the decrease of the anodic current was essentially due to a partial alteration of the electrode by deactivation of a portion of the bentonite active sites.

Fig. 14
figure 14

Cyclic voltammograms related to the Cd2+ and Pb2+ ion detection by using the reused bentonite-modified electrode (cycles 1 and 2). The cyclic voltammogram corresponding to the regenerated electrode immersed in cation-free solution is introduced for comparison. Ion initial concentration, 10−3 M; sweep rate, 10 mV/s; electrolyte: HCl; pH = 2

Full size image

Conclusion

The results obtained from this study allowed us to conclude that the bentonite-modified electrode sensitivity regarding Cd2+ ion detection was markedly affected by the change of the operating factors: solution pH, bentonite content, analyte concentration, electrolytes, and sweep rate. In alkaline media (up to pH = 9), a portion of cadmium-derivative cations seemed to be immobilized by the negatively charged clay particles, and consequently, the current signal diminished. In such a condition, the charge transfer was almost reversible. The current signal enhanced as the clay content increased, due to the good physical characteristics of the used clay minerals. However, because of the electrode peeling off, the bentonite content should not exceed 20 mass%. The relationship between the anodic current density and the cadmium cation initial concentration together with the cation exchange capacity of the electrode allowed us to conclude that the electrode’s limit of detection fell within the range of 6 × 10−6–5 × 10−5 M. The calculated values of LOD and LOQ (4.23 × 10–7M and 1.41 × 10–6M, respectively) supported the latter conclusion. The effect of the tested supporting electrolytes on the electrochemical performance of the electrode was essentially linked to the cation exchange process and the clay mineral protonation. In this connection, it should be noted that the use of HCl as a supporting electrolyte significantly improved the electrochemical performance of the studied electrode. The electron transfer process exhibited a transition from quasi-reversible to irreversible behavior when the sweep rate exceeded 50 mV/s. This change was accompanied by the disappearance of the cathodic peak and an intensification of the anodic current density.

The bentonite, as an electrode modifier, facilitated the electrode regeneration through leaching with a salt aqueous solution. The regeneration process seemed to occur by cation exchange. The regenerated electrode exhibited successful reusability. On the other hand, the studied electrode showed excellent sensitivity and high selectivity regarding the simultaneous detection of Cd2+ and Pb2+ ions.

The electrical equivalent circuit of the electrode-solution interface consisted of a double-layer capacitance in parallel with an electron transfer resistance and a Warburg impedance. This model accurately described the electrical behavior of the electrode during the electrochemical process.

These findings, which highlighted the suitability of the bentonite-carbon paste working electrode for cadmium detection, contribute to the advancement of electrochemical sensing techniques and can help to extend the clay-modified electrode use for heavy metal detection.