Nanostructured Carbon and Gold Screen-Printed Electrodes for Sensitive Detection of Benzisothiazolinone in Environmental Water Samples

Abstract

Benzisothiazolinone (BIT) is a commonly used biocide in water-based products, which can enter the environment from household and personal care products, as well as from leaching off building facades and roofs due to rainfall, eventually reaching rivers through stormwater runoff and raising ecological concerns due to its high aquatic toxicity. Detecting benzisothiazolinone, particularly in the environment is crucial due to health and regulatory requirements. This study explores electrochemical techniques and conductive nanomaterials for detecting BIT in environmental samples. Carbon- and gold-based screen-printed electrodes (SPEs) with distinct morphologies were investigated: carbon electrodes as nanoparticles (SPE-C) and single-wall carbon nanotubes (SPE-SWCNTs), and gold electrodes as nanoparticles (SPE-Au-BT) and thin films (SPE-Au-AT). Cyclic voltammetry and square-wave voltammetry (SWV) were optimized, with SWV demonstrating superior sensitivity—showing a two-order improvement with carbon-based electrodes and a 30-fold enhancement with gold-based electrodes. The lowest detection limits were 40 nM for carbon and 80 nM for gold nanoparticle-based electrodes. SPE-C achieved good recovery in river water, confirming its effectiveness for BIT monitoring with minimal interference from common ions or saccharin. These sensors can be easily used for everyday detection and monitoring of BIT in river water, ensuring a screening programme that supports the development of adequate regulatory guidelines.

1. Introduction

Isothiazolinones, which belong to the family of heterocyclic organic compounds, are known for their biocidal properties that derive from a nitrogen and sulphur aromatic ring. These compounds function effectively as preservatives, protecting products from microbial contamination [1]. The most used isothiazolinone is methylisothiazolinone (2-methyl-4-isothiazolin-3-one, MIT or MI), often employed in combination with another isothiazolinone, such as chloromethylisothiazolinone (5-chloro-2-methyl-4-isothiazolin-3-one, CMIT or MCI or MCIT) (Figure 1). This combination, known as Kathon^TM, is highly effective at low concentrations and is prevalent in water-based products such as paints, adhesives, and personal care items [2,3,4,5,6].

Benzisothiazolinone (1,2-Benzisothiazolin-3-one, BIT) is another important member of the isothiazolinone family, widely utilized for its antimicrobial properties (Figure 1). Like other isothiazolinones (MIT and CMIT), BIT disrupts microbial enzyme activity, leading to cell death. It is particularly popular in applications requiring longer-term preservation due to its stability and broad-spectrum efficacy [6,7]. Despite their effectiveness, isothiazolinones have raised concerns because they provide a weak-to-strong sensitizing potential. It can cause skin irritation and allergic reactions, particularly in sensitive individuals [4,5]. Besides skin sensitizing, it has also raised concerns relating to environmental safety because it poses risks to aquatic organisms [8]. Recent directives from the European Chemicals Agency have highlighted the acute toxicity of isothiazolinones to aquatic life, primarily attributed to their reduced sulfur content [9]. Isothiazolinones can enter the environment through home and personal care products or by leaching from buildings’ facades and roofs through rainfall and reaching rivers through rainwater runoff [8]. It was found that MIT can enter riverbank filtration systems, persisting and migrating for decades if the biogeochemical conditions are stable [10]. During the COVID-19 pandemic, the use of MIT and BIT drastically increased, and so did their presence in the environment [11]. The concentration of MIT and BIT in the Paris conurbation wastewater was found to be 1.7–7.8 nM [12]. Accordingly, the concentrations measured in urban discharges indicate the potential contamination of the receiving aquatic environment and underscore the need to implement source control measures.

The safe concentration of BIT in products is regulated under the Biocidal Product Regulation (BPR). According to the BPR, industrial and retail products (excluding paints) with more than 0.036% BIT must be labelled as harmful upon contact with skin, eyes, or lungs [13]. However, the regulatory requirements of safe levels of BIT in drinking and groundwater are still not established, as there are no specific guidelines for this chemical from organizations like the EPA or the WHO, but it is something that can be expected in the future [14,15]. Detecting BIT is important not only in consumer products, but also in environmental water due to the potential environmental impact.

Electrochemical methods offer a sensitive and selective option for this purpose, and the use of disposable screen-printed electrodes with a portable potentiostat enables practical on-site analysis. The amount of analyte can be very small, 100–150 μL. Techniques such as cyclic voltammetry (CV), differential pulse voltammetry (DPV), and amperometry are commonly used. Besides them, square-wave voltammetry (SWV) is favoured over other voltammetric techniques due to its shorter analysis times, reduced capacitive currents, and minimized electrode-surface blocking. SWV reduces problems with electrode surface blocking by using short, alternating potential pulses that minimize the time for adsorption and accumulation of reaction products on the electrode surface, thereby maintaining cleaner and more reproducible electrode conditions [16]. These methods rely on the electrochemical behaviour of isothiazolinones, where their oxidation or reduction at the electrode surface generates a measurable current proportional to their concentration. Jakubczyk et al. [17] demonstrated the determination of MIT using CV and DPV on a boron-doped diamond (BDD) electrode in a citrate–phosphate buffer (pH 5.6). CV was employed to study the reaction kinetics, while DPV provided a limit of detection (LOD) of 2.05 µM. They also demonstrated that the oxidation response of isothiazolinones was attributed to hydroxyl radicals generated at the BDD surface, which is independent of pH. Abad-Gil et al. [18] reported the detection of MIT using gold nanoparticles incorporated into a polymer matrix, which was drop-cast onto screen-printed electrodes. A limit of detection of 22.59 µM was achieved using cyclic voltammetry in 0.1 M NaOH (pH 13). In a subsequent study [19], the same group employed an unmodified gold-disc electrode in 0.1 M phosphate buffer (pH 6) for MIT detection. Using square-wave voltammetry (SWV), they obtained an LOD of 460 µM, while with square-wave adsorptive stripping voltammetry (SWAdSV), a significantly lower LOD of 0.069 µM was achieved. The detection efficacy of an electrochemical sensor depends on various factors, including the electrode’s composition, electrolyte solution type and concentration, and pH level. Choosing the adequate techniques (CV, DPV, SWV, SWAdS etc.) for the detection of the target analyte is also important. Optimizing all these parameters is essential for developing a highly sensitive electrochemical sensor.

Despite extensive research on the electrochemical detection of isothiazolinones such as MIT, systematic studies comparing electrochemical techniques for BIT detection are lacking. The main focus was on determining MIT and CMIT due to the increased prevalence of contact dermatitis over the years, which cast a shadow over BIT [20]. Considering that BIT is a persistent and mobile chemical that is toxic to the environment and present in wastewater, its detection and monitoring are crucial. Our research addresses this gap by exploring the potential of electrochemical sensors for BIT detection, highlighting the necessity of developing effective, real-time monitoring tools for this compound. Developing and utilizing electrochemical detection methods for BIT is essential to ensure their effective and safe use, compliance with regulatory standards, and the protection of public health. To the best of our knowledge, no research has been published to date on the electrochemical detection of BIT on carbon- and gold-based, screen-printed electrodes. In the present study, we developed a simple, cost-effective, portable, and easy-to-use electrochemical sensor for insitu BIT measurements using unmodified carbon- and gold-based commercial screen-printed electrodes. The real sample measurements were conducted with square-wave voltammetry measurements on samples of river water.

2. Materials and Methods

2.1. Materials

1,2-Benzisothiazolin-3-one (BIT), boric acid, citric acid and sodium hydroxide were purchased from Sigma-Aldrich (St. Louis, MO, USA), hydrochloric acid and ortho-phosphoric acid were purchased from Merck Serono (Aubonne, Switzerland). Potassium ferricyanide was purchased from Alfa Aesar (Kandel, Germany). All reagents were of analytical grade and used without further purification. Britton–Robinson buffer (BRB) was prepared as a solution of 40 mM boric acid, 40 mM citric acid, and 40 mM ortho-phosphoric acid in ultrapure water. The solution was then titrated to the desired pH with a 2.5 M NaOH solution. A stock solution of 2 mM BIT was prepared by dissolving an appropriate amount in the Britton-Robinson buffer.

2.2. Methods

2.2.1. Scanning Electron Microscopy

The surface morphologies of the SPE-C and SPE-SWCNT electrodes were investigated using scanning electron microscopy (SEM). The samples for SEM were mounted on metallic sample holders using carbon adhesive tape and carbon conductive paint and observed at various magnifications in a field-emission-gun scanning electron microscope (FEG-SEM Verios G4 HP, Thermo Fisher Scientific, Waltham, MA, USA). Preliminary examinations of the samples showed that they were sufficiently electrically conductive for the SEM analysis. The experimental conditions for imaging were set to an accelerating voltage of 4 kV, a beam current between 20 pA and 50 pA, and a working distance of 3 mm. The electron micrographs were recorded with two detectors: (i) a through-the-lens detector (TLD) for secondary electrons (SEs) and (ii) a mirror detector (MD) for backscattered electrons (BSEs) positioned within the objective lens. EDS analysis was performed at 10 kV and 0.2 nA.

2.2.2. Electrochemical Measurements

Electrochemical measurements were performed via cyclic voltammetry (CV), and square-wave voltammetry (SWV) using screen-printed electrodes (Metrohm DropSens, Asturias, Spain) and a portable potentiostat (EmStat4, Palm Instruments, Amsterdam, The Netherlands), which was operated with PSTrace 5.9 software.

Four types of screen-printed electrodes were used (

Table 1. Composition of screen-printed electrodes.

Electrodes	Working Electrode	Auxiliary Electrode	Reference Electrode
SPE-C	C	Pt	Quasi-Ag
SPE-SWCNT	COOH-SWCNT	C	Quasi-Ag
SPE-Au-AT	Au	Pt	Quasi-Ag
SPE-Au-BT	Au	Pt	Quasi-Ag

): SPE-C (DropSens DRP-150) with a carbon working electrode; SPE-SWCNT (DropSens DRP-110SWCNT) with a carboxylated single-wall carbon nanotubes working electrode; SPE-Au-AT (DropSens DRP-250AT) with a high-temperature gold ink working electrode and SPE-Au-BT (DropSens DRP-250BT) with a low-temperature gold ink working electrode. The diameter of all the working electrodes is 4 mm. All the potentials reported are referred to the Ag quasi-reference electrode.

2.2.3. Electrode Pre-Treatment

The SPE-C and SPE-SWCNT electrodes were used without a pre-treatment step. The SPE-Au-AT and SPE-Au-BT were electropolished before the electrochemical measurements to activate the surface and obtain reproducible electrochemical behaviour. For this purpose, the gold-based electrodes were cycled four times between 0 and 1.2 V in 0.5 M H₂SO₄ at a scan rate of 100 mVs⁻¹ [21].

2.2.4. Determination of the Electrochemical Surface Area

The electrochemical surface area of the electrodes was determined by cycling the electrodes three times between −0.3 and 0.6 V, scan rate 50 mVs⁻¹, potential step 10 mV in a solution containing 1 mM K₃Fe(CN)₆ and 0.1 M KCl. The electrochemical surface area was calculated from the Randles-Ševčík equation for 25 °C (Equation (1)), where i_p is the anodic current maximum of the third cycle in ferricyanide, 2.69 is constant with unit C mol⁻¹V^−1/2, n is number of electrons transferred in the reaction (in this case 1), A is electrode surface area (cm²), C is the concentration (mol cm⁻³), ʋ is the scan rate (0.05 Vs⁻¹), D is the diffusion coefficient (for K₃Fe(CN)₆ is 7.3 × 10⁻⁶ cm²s⁻¹). Step-by-step calculation of electrochemical surface area for the SPE-C was represented in the Supporting Information. The same procedure was performed for all samples.

$$i_p=2.69\times {10}^5{n}^{{\displaystyle \frac{3}{2}}}AC\sqrt{\vartheta D}$$

(1)

2.2.5. Calibration Curve

CV scans were performed at a potential that was set between 0 and 1 V vs. a quasi-Ag-reference electrode with a potential step of 10 mV and a scan rate of 50 mVs⁻¹ vs. a quasi-Ag-reference electrode. After each set of measurements, the SPE electrode was cycled six times in the blank solution (BRB), as a cleaning stage of the electrode.

SWV measurements were performed between 0.2 V and 1 V with a potential step of 10 mV. For the square-wave adsorptive stripping voltammetry measurements, BIT was accumulated on the electrode surface prior to the measurement. The parameters for SWV and SWAdSV are given in Table S1’s in Supporting Information.

Unless stated otherwise, the BRB was used as a supporting electrolyte in all the electrochemical measurements. The pH was optimized before the quantitative analysis and was 4 for the gold-based electrodes and 5 for the carbon-based electrodes.

Calibration curves were obtained from the CVs and SWVs data. The limit of detection (LOD) was calculated according to Equation (2):

$$\mathrm{L}\mathrm{O}\mathrm{D}=3{\displaystyle \frac{\delta }{k}}$$

(2)

where k is the slope of the calibration curve (sensitivity) and δ is the standard deviation of the lowest-detected concentration.

The limit of quantification (LOQ) is calculated according to Equation (3):

$$\mathrm{L}\mathrm{O}\mathrm{Q}=3.3\mathrm{L}\mathrm{O}\mathrm{D}$$

(3)

2.2.6. Interference and Real Sample Studies

For the real sample studies, river water was collected from the nearby Gradaščica River, Ljubljana, and refrigerated before use. Before analysis, the pH was adjusted from around 8 to 5 using 1 M HCl. The sample was allowed to settle, and the supernatant was decanted.

Recovery values were set to determine the sensor’s suitability for analyzing BIT in real samples. The resulting solution was spiked with different concentrations of BIT (0.25, 0.75, 1.5 and 5.00 µM). The spiked samples were analyzed using SWV at SPE-C, and recovery values were calculated according to Equation (4):

$$Recovery={\displaystyle \frac{c(BIT,measured)}{c(BIT,spiked)}}$$

(4)

Interference studies were performed using compounds typically detected in river water, including NaCl, Pb(NO₃)₂, Cr(NO₃)₃, paracetamol, saccharin and glyphosate. To evaluate the influence of inorganic ions (Na⁺, Cr³⁺, Pb²⁺, Cl⁻, and NO₃⁻), NaCl, Pb(NO₃)₂, and Cr(NO₃)₃ were introduced into a 20 µM BIT solution to a final concentration of 200 µM, 20 µM, and 20 µM, respectively. In addition, 20 µM of saccharin and glyphosate were also added to the 20 µM BIT solution. Paracetamol was added in concentrations ranging from 20 µM to 200 µM.

All samples were analyzed by SWV using SPE-C electrodes.

3. Results and Discussion

3.1. Morphology

The SEM analysis was conducted to evaluate the surface morphology of the working electrodes, as electrochemical detection occurs at their surfaces. Figure 2 clearly demonstrates the differences between the two gold electrodes, as well as the variations in morphology between carbon electrodes. The SPE-C consists of distinct layers made from a mixture of carbon nanoparticles and binder, while the SPE-SWCNT electrode includes the addition of single-wall carbon nanotubes. The carbon particle size is approximately 30 nm, while the single-wall carbon nanotube diameter is around 10 nm. Both the nanoparticles and nanotubes are randomly distributed. The cross-sections show approximately 20 µm carbon material layers. EDS spectra (Figure S1) confirm that both carbon-based electrodes contain only carbon. Two types of gold electrodes, SPE-Au-AT and SPE-Au-BT, were also examined, with EDS spectra (Figure S1) indicating the presence of gold exclusively. The top view of SPE-Au-AT (Figure 2e) shows a smooth surface, while SPE-Au-BT (Figure 2g) exhibits a distinctly different morphology, consisting of agglomerated gold particles ranging from 0.2 to 2 µm. The distribution of these agglomerates is shown in Figure S2, demonstrating that most fall within the 0.2–1 µm range, with a mean diameter of 0.76 µm. The side view of the electrodes further highlights notable differences in the morphology and thickness. The AT electrode forms a dense, compact, thin layer, whereas the BT electrode contains agglomerated particles that create thicker and more compact structures. The thickness of the BT electrode is similar to that of the carbon-based electrodes, i.e., around 16 µm, whereas the AT electrode is much thinner, approximately 150 nm. This discrepancy can be attributed to using low-temperature curing ink for the BT, in contrast to the high-temperature curing ink used for the AT, which likely caused sintering and resulted in the absence of particles in the AT electrode.

3.2. Electrochemical Behaviour of Electrodes

3.2.1. Electrochemical Behaviour in Ferro-Ferricyanide

The research started by investigating the electrochemical behaviour of the electrodes in a 1 mM ferro-ferricyanide solution prepared with 0.1 M KCl using cyclic voltammetry. The results are shown in Figure 3. The anodic and cathodic peak currents with distance between them were obtained from CVs and presented in Table S2. Among the carbon-based electrodes, the SPE-SWCNT (Figure 3a—red curve) exhibits higher anodic (I_pa) and cathodic peak currents (I_pc) compared to the SPE-C (Figure 3a—black curve). The peak-to-peak separation (ΔE) for SPE-C is 260 mV, while for SPE-SWCNT, it is 90 mV, which is higher than the ideal peak-to-peak separation of 57 mV (according to the Nernst equation). Ideal peak-to-peak separation is observed for a rapid, one-electron, reversible electrochemical process. Reversible processes enable electroactive species to undergo reduction and subsequent reoxidation. In reality, the peak-to-peak separation for screen-printed electrodes is higher than 57 mV in the ferri/ferrocyanide couple due to higher resistance to electron transfer, which makes the redox processes quasi-reversible. Having in mind that the working electrode of SPE-C is carbon nanoparticles dispersed in a polymeric matrix, creating a heterogeneous surface that can hinder electron transfer, which slows the reaction. However, the presence of SWCNT probably increases the conductivity of WE, which decreases ΔE. In addition, it is desirable that the relationship between anodic and cathodic current (I_pa/I_pc) is close to one. It is observed that for the SWCNT electrode it is one, while for the carbon-based electrode it is much lower than one.

For gold-based electrodes, it is well-established that a surface pretreatment, such as electropolishing in acid, is beneficial before electrochemical measurements [22,23], as this removes organic surface impurities. Therefore, the impact of electropolishing on the electrochemical performance of gold-based SPEs was assessed. Figure 3 and Table S2 indicate that electropolishing did not alter the voltammograms for the SPE-Au-AT (Figure 3b). However, electropolishing enhanced the peak currents and reduced the peak-to-peak separation from 160 to 90 mV for the SPE-Au-BT. Electropolishing with H₂SO₄ improved the electrochemical behavior of the SPE-Au-BT, whereas it did not affect the SPE-Au-AT electrode.

For further research, SPE-Au-BT was used after polishing, while other ones (SPE-C, SPE-SWCNT and SPE-Au-AT) were used without pretreatment. Figure S3 compares the unpolished, carbon-based SPE, the unpolished SPE-Au-AT, and the polished SPE-Au-BT. Among these, SPE-C displays the lowest peak currents and the highest ΔE, while the unpolished SPE-Au-AT and polished SPE-Au-BT are very similar and show the highest current peaks.

Electroactive surface areas (A) were estimated using Equation (1) (step-by-step calculation was represented in the Supporting Information), yielding values of 3.9 mm² for SPE-C and 7.4 mm² for SPE-SWCNT, with the gold-based SPE exhibiting approximately 8 mm² (Table S2). The actual electrochemically active surface area, denoted as A_real, can be determined to evaluate the active surface area and inactive regions across the total macroscopic area using the following method: A_real = A/A_geo, where A_geo is the geometric area based on the 4 mm diameter of the working electrode. SPE-Au-AT and SPE-Au-BT-polished showed higher active surface areas, around 65%, compared to 59% for SPE-SWCNT and 31% for SPE-C. Geometry and electroactive surface area are closer for the gold-based electrodes [24].

According to these results, it can be concluded that the fastest kinetic reaction is obtained on gold-based electrodes compared to carbon-based electrodes.

3.2.2. Electrochemical Behaviour in BIT Solution

A further analysis was conducted by examining the electrochemical response of carbon- and gold-based SPEs to detect the BIT. Figure 4a,b show the cyclic voltammograms for the carbon-based SPE in a buffer solution, with and without the BIT present in the solution. BIT oxidation, observed at a potential of 0.65 V, occurs on both carbon-based electrodes. The SPE-SWCNT displays a peak current twice as high as the SPE-C; however, it also shows an elevated baseline response in the buffer solution, indicating higher non-faradaic currents.

Figure 4c,d present the voltammograms representing BIT oxidation on the two gold-based SPEs. The baseline responses for SPE-Au-AT and SPE-Au-BT differ significantly.

SPE-Au-BT in buffer solution showing two typical peaks for the oxidation and reduction of gold around +0.8 V and +0.4 V, respectively (Figure 4d) [19,25]. However, these peaks were not detected for the SPE-Au-AT electrode, which we attribute to the distinct surface composition and morphology of this electrode compared to SPE-Au-BT. We believe that variations in crystallinity and surface structure may influence the electrochemical behaviour of SPE-Au-AT in the ferricyanide system.

In the presence of BIT, the SPE-Au-BT voltammogram shows a new peak at approximately 0.65 V, indicating BIT oxidation. It is also observed that both the oxidation and reduction peaks for gold are shifted towards more positive potentials. Additionally, the SPE-Au-AT electrode displays an oxidation peak at 0.7 V, corresponding to BIT oxidation.

Square-wave voltammetry was further used to detect the BIT on the carbon- and gold-based electrodes. This approach minimizes the contribution of the non-faradaic current, thereby increasing the sensitivity of the analyte’s detection. Moreover, SWV is the fastest pulse-voltammetry technique [26,27]. SWV voltammograms of BIT oxidation on carbon- and gold-based electrodes are presented in Figure 5. It shows that the peak current increases with increasing BIT concentration for all four electrodes. Among them, SPE-Au-AT exhibits the lowest peak current (Figure 5c), where, for instance, at a concentration of 50 µM, the peak current is approximately 1.75 µA. In contrast, for SPE-SWCNT (Figure 5b), the peak current reaches 5.5 µA, which is three times higher. The highest peak current is observed for the SPE-Au-BT electrode (Figure 5d), where a current of approximately 8 µA is already recorded at a concentration of 10 µM. Comparing the SWV with CV, a higher sensitivity is observed for the SWV than the CV, except for the SPE-Au-AT sample.

When the peak currents obtained by CV and SWV are compared at the same analyte concentration, SWV consistently yields a higher and more defined peak. This difference stems from the fundamental differences between the two techniques. In CV, the linearly swept potential allows both faradaic and capacitive currents to contribute to the overall signal, which typically produces broader peaks and slightly lower peak currents. In contrast, SWV employs a pulsed waveform and measures the difference between forward and reverse currents, effectively suppressing capacitive contributions and enhancing faradaic response, resulting in sharper and higher peaks. The slight shift in peak potential observed between the two methods is primarily due to the specific waveform characteristics and operating parameters of SWV—including pulse amplitude, frequency, and step potential—which influence electron-transfer kinetics and the potential at which maximum current is achieved. Consequently, small deviations in peak potential between CV and SWV are expected and are well aligned with reported electrochemical behaviour.

3.2.3. Effect of pH on the Working Electrode

SWV was employed to investigate the effect of pH on the oxidation of BIT on carbon- and gold-based SPE. A concentration of 50 µM BIT was investigated across a pH range of 2 to 10 in the Britton–Robinson buffer. As shown in Figure S4, increasing the pH from 2 to 7 causes a shift in the oxidation peak toward lower potentials for both electrodes, with stabilization under basic conditions.

For the carbon-based SPE (Figure 6a), the highest peak currents are observed between pH 4 and 6, indicating that slightly acidic conditions are optimal for BIT detection. The optimal pH value is 5, which is lower than the pKa of BIT. According to ref. [28], the pKa for BIT is 7.3, while in accordance with ref. [11] pKa is predicted to be 10.2. In any case, the neutral form of BIT is present in the pH range 4–6, which is easier for protonation (oxidation) than in basic pH, where the deprotonated form of BIT is present. In contrast, the gold-based SPE exhibits distinct pH-dependent behavior (Figure 6b), showing the highest peak currents under acidic conditions (pH 2–5) and the lowest at neutral pH. The maximum current is obtained at pH 3; however, peaks at pH 2 and 3 are less symmetric and appear as additional peaks, suggesting a possible two-step oxidation process (Figure S4b). A well-defined and symmetric peak is observed at pH 4, which was therefore selected for further studies.

Figure 6c,d show a linear relationship between pH and oxidation potential. For the carbon-based SPE, linearity is observed from pH 2 to 8 with a slope of 56 mV/pH, while for the gold-based SPE linearity is observed from pH 2 to 7, with a slope of 54 mV/pH, closely resembling the theoretical Nernstian value of 59.2 mV/pH, indicating a proton-coupled electron transfer involving an equal number of electrons and protons [29,30].

The voltametric response can be described according to the Nernstian Equation (Equation (5)):

$$E=E_0-0.0592{\displaystyle \frac{h}{n}}pH$$

(5)

where E₀ is the standard electrode potential, n is the number of electrons, and h is the number of protons involved in a redox reaction. Equation (5) for the voltametric response of the BIT on the carbon-based SPE in the pH range 2–8 is: E_a = 0.843 V–56 mV × pH, while for the gold-based electrode in the pH range 2–7 is: E_a = 0.895 V–54 mV × pH.

The different pH-dependent behaviours of carbon and gold electrodes arise from their distinct surface chemistries. Gold exhibits specific adsorption and weaker binding of neutral intermediates, making its electrochemical response more sensitive to pH and surface state, which can lead to peak shifts, changes in peak symmetry, and multiple oxidation steps under acidic conditions. In contrast, carbon-based electrodes are more electrochemically inert, possess fewer specific adsorption sites, and have a broader potential window, resulting in more stable and symmetric oxidation peaks, particularly under slightly acidic to neutral conditions [31,32]

3.2.4. Determination of LOD and LOQ

The effect of the BIT concentration on the current response was used to obtain calibration curves and determine the limit of quantification and the calculated limit of detection. Two different methods were employed for comparison: cyclic voltammetry and square-wave voltammetry.

The calibration curves for the carbon-based SPEs derived from the CVs are illustrated in Figure S5a,b, with the corresponding LOD and LOQ values presented in

Table 2. LOD, LOQ and sensitivity for SPE-C, SPE-SWCNT, SPE-Au-AT and SPE-Au-BT in BIT solution obtained via CV and SWV.

	SPE-C			SPE-SWCNT
	LOD, µM	LOQ, µM	Sensitivity, µA/µM	LOD, µM	LOQ, µM	Sensitivity, µA/µM
CV	0.69	2.29	0.025	2.23	7.37	0.050
SWV	0.04	0.13	0.057	0.39	1.29	0.115
	SPE-Au-AT			SPE-Au-BT-Polished
	LOD, µM	LOQ, µM	Sensitivity, µA/µM	LOD, µM	LOQ, µM	Sensitivity, µA/µM
CV	0.76	2.61	0.034	3.59	11.85	0.019
SWV	0.86	2.84	0.036	0.08	0.26	0.607

. Notably, the linear range spans from 0.25 to 100 µM for SPE-C and from 1 to 100 µM for SPE-SWCNT, demonstrating good linear regression with R² values of 0.996 and 0.998. While the SPE-SWCNT electrode exhibits higher sensitivity than SPE-C, the latter demonstrates a more favourable limit of detection of 0.69 µM compared to 2.23 µM for SPE-SWCNT. This difference is primarily attributed to the lower standard deviation obtained for the SPE-C electrode, which directly affects the LOD calculation. A linear response was observed for both electrodes using the SWV method, as presented in Figure S5c,d, with the corresponding analytical parameters summarized in Table 2. As shown in Table 2, the sensitivity of both electrodes increased, while the LOD and LOQ values decreased by up to 15-fold for the SPE-C electrode. The comparison between CV and SWV confirms that SWV provides enhanced sensitivity and lower detection limits, highlighting its suitability as a more sensitive technique for BIT determination.

The calibration curves for the gold-based electrodes obtained with CV and SWV are presented in Figure S6. Table 2 compares the two techniques for BIT detection. Using SWV, SPE-Au-AT demonstrates a linear range spanning from 5 to 100µM, while SPE-Au-BT ranges from 0.5 to 10 µM. For SPE-Au-AT, the sensitivity remains consistent, though slightly better LOD and LOQ values are observed with CV. However, the behavior differs for SPE-Au-BT, where sensitivity is markedly higher with SWV, significantly improving the detection of very low concentrations.

Although SWV generally provides enhanced sensitivity, this improvement may not occur for all electrode types. For the SPE-Au-AT electrode, the lack of sensitivity enhancement with SWV could be attributed to its distinct surface morphology and structure, which likely influence the charge transfer and electrochemical response characteristics. Lower sensitivity with SWV compared to CV for gold electrodes was also reported by Abad-Gil et al. who found that SWV yielded a significantly higher LOD for MIT on a gold electrode (460 μM) [19], whereas CV provided a much lower LOD (22.59 μM) [18] when gold nanoparticles were deposited on screen-printed electrodes.

Considering all the analyses, it can be concluded that the SPE-Au-BT demonstrated the highest sensitivity with a very low LOQ, while SPE-C, despite its lower sensitivity, exhibited the lowest LOQ. This is likely due to different processes occurring on the electrode surfaces or the minimal influence of the non-faradaic current.

Table 3. Comparison of electrochemical parameters was obtained from this study, with a literature review.

Electrodes	Analyte	LOD	LOQ	Techniques	Ref.
SPEC/PDDA/Au	MIT	2.26 M	7.56 M	CV	[18]
Gold Disc Electrode	MIT CMIT DCOIT	24.33 µM 0.0696 µM	81.7 µM 0.23 µM	SWV SWAdSV	[19]
BDDE	MIT	2.05 µM	-	DPV	[17]
SPE-C SPE-SWCNT SPE-Au-AT SPE-Au-BT	BIT	0.04 µM 0.39 µM 0.86 µM 0.07 µM	0.13 µM 1.29 µM 2.84 µM 0.26 µM	SWV SWAdSV	This study

SPEC, screen-printed carbon electrode; PDDA, poly (diallyldimethylammonium); MIT, methylisothiazolinone; BDDE, boron-doped diamond electrode; CMIT, Chloromethylisothiazolinone; DCOIT, 4,5-dichloro-2-octyl-4-isothiazolin-3-one.

presents the electrochemical parameters for the detection of methylisothiazolinone and other isothiazolinones from our study and previous research. According to the literature survey, no systematic studies of the electrochemical detection of BIT have been reported. All the electrodes in this study demonstrated better LODs compared to those in the literature. Optimization of the electrochemical methods, with an appropriate choice of electrodes, can further improve the sensitivity and lower the LOQ for BIT detection.

3.3. Real Samples

The proposed electrochemical approach was applied to determine the BIT content in river samples and assess its potential for insitu analysis, see Figure 7. Because the SPE-C showed the lowest detection limit, it was used to analyze real samples using the SWV technique.

The recovery was determined by spiking the river water with known amounts of BIT as can be seen from

Table 4. Detection of BIT in spiked river-water samples at pH 5.

c (BIT, Spiked) (µM)	c (BIT, Found) ± SD (µM)	Recovery (%)
0.25	0.34 ± 0.01	136.00
0.75	0.78 ± 0.08	103.70
1.50	1.60 ± 0.10	106.70
5.00	4.80 ± 0.60	96.00

. The water was collected from the Gradaščica River in Ljubljana and used without further purification. All the measurements were carried out in triplicate. As seen from Table 4, the recovery values were 96–136%, indicating that the sensor is suitable for BIT monitoring in real samples.

3.4. Selectivity

To assess the selectivity of the SWV technique, the effect of various species commonly present in natural and wastewaters on the BIT analytical signal was examined. The investigated potential interferents included inorganic ions (Na⁺, Pb²⁺, Cr³⁺, Cl⁻, and NO₃⁻), small organic molecules (paracetamol, saccharin), and a pesticide (glyphosate). The spiked samples were analyzed using SWV at SPE-C. Among the tested compounds, only paracetamol exhibited electroactivity within the investigated potential window (0.0–1.0 V). Its oxidation peak appeared in proximity to the BIT signal but did not overlap, as shown in Figure 8, confirming that it does not interfere with BIT determination. It can be seen that different concentrations of paracetamol did not influence BIT detection. No significant peaks were observed for the other tested species, indicating the absence of electrochemical activity in this range. These findings demonstrate that BIT can be reliably detected in river water even in the presence of common inorganic ions and organic pollutants, highlighting the high selectivity of the developed method.

This study demonstrated that unmodified carbon- and gold-based commercial screen-printed electrodes, when used with appropriate electrochemical methods and parameters, can provide a simple, fast, and reliable approach to detecting BIT in river water. Given that the BIT concentration in wastewater from the Paris conurbation is approximately 2–8 nM [12], the LOD of SPE-C (40 nM) and SPE-Au-BT (80 nM) makes them suitable candidates for detecting low BIT concentrations in real samples. Future work will focus on modifying the working electrodes to achieve a detection limit below 2 nM.

4. Conclusions

This study developed a fast, on-site electrochemical method for the detection of BIT in water samples using carbon- and gold-based screen-printed electrodes. By optimizing the SWV method, the influence of non-faradic current was significantly reduced, enhancing the electrode sensitivity. SPE-Au-BT demonstrated exceptional performance for BIT detection under optimized SWV conditions. In contrast, while SPE-Au-AT exhibited good electrical properties, its performance did not improve significantly with SWV. SWV proved to be a suitable technique for electrodes dominated by background currents, enabling the detection of very low concentrations of BIT. Optimizing both materials and methods is crucial for developing high-performance sensors. These materials show great potential for low-cost, easy-to-use, portable sensors for field applications, eliminating the need for expensive analytical techniques.