Effect of ZrO₂ Coating Thickness on Capacitive Sensor Performance in Conductive Liquid Media

Abstract

This study presents a capacitive sensor with a zirconium oxide (ZrO₂) coating for real-time measurement of component concentration in liquid media. The ZrO₂ layer was formed on stainless steel electrodes by magnetron sputtering, and its structural, morphological, and chemical properties were characterized using SEM, EDS, FTIR, and XRD. It was found that increasing coating thickness results in more continuous and highly crystalline layers, while reducing the influence of the substrate on surface properties. The performance of the capacitive sensor was evaluated by analysing the dependence of capacitance on frequency and NaCl concentration. The results show that the thickness of the ZrO₂ layer has a significant influence on sensor sensitivity and measurement stability. A thinner layer (~2 µm) provides higher sensitivity but is more affected by parasitic effects, while thicker layers improve measurement stability at the expense of reduced sensitivity. An optimal trade-off between sensitivity and stability is achieved at a ZrO₂ layer thickness of approximately 4 µm, ensuring sufficient sensitivity and good measurement repeatability. The results indicate that ZrO₂-modified capacitive sensors are a promising technology for monitoring liquid quality, particularly in environmental protection and industrial process control.

1. Introduction

1.1. Context and Problem

In modern environmental protection and industrial process management, monitoring of water and industrial fluids pollution is becoming increasingly important. Intensifying industrial activity introduces various pollutants into natural water bodies and in the form of technological fluids, dissolved salts, oil products, chemical compounds, etc. These pollutants change the physical and chemical properties of fluids, harm ecosystems, and pose a threat to human health: increased salinity disrupts biological processes, oil products form surface films and release toxins. Therefore, reliable and effective methods are necessary that allow continuous monitoring of pollution, promptly record changes in pollutant concentrations, and make timely decisions [1,2,3]. Real-time monitoring is becoming increasingly important, which allows for a rapid response to pollution incidents and risk reduction. Traditional laboratory methods, although accurate, require sampling, transportation and lengthy analyses—which are expensive, slow and do not allow for rapid continuous monitoring in real time. Therefore, rapid, cost-effective, direct measurement methods are increasingly emphasized—they ensure real-time detectable changes, more efficient management of industrial processes, optimization of control strategies, and significantly lower long-term monitoring costs compared to periodic laboratory tests [4].

1.2. Existing Methods and Limitations

Currently, various laboratory analytical methods are often used to determine the contamination of liquid media and the concentration of components in them, such as chromatography, spectroscopic studies or classical chemical analysis. For example, gas or liquid chromatography allows for the accurate determination of the composition and concentration of organic compounds, including petroleum products, while spectroscopic methods, such as UV–Vis or infrared spectroscopy, can be used to identify various substances in solution. However, despite their high accuracy and selectivity, these methods usually require taking samples from the tested medium and transporting them to the laboratory, where complex and time-consuming analysis procedures are performed. In addition, such analysis requires expensive laboratory equipment, qualified personnel and additional reagent costs [5,6,7,8]. For these reasons, the above methods are most often used for periodic studies, rather than continuous real-time monitoring. Also, their integration into automated industrial monitoring systems is difficult, since laboratory methods are not suitable for direct and continuous measurement in technological processes. Therefore, in practical applications, alternative sensor solutions are increasingly being developed to enable rapid and direct assessment of changes in the composition of liquid media without complex laboratory procedures [9].

1.3. Types of Sensors for Measuring Pollutants

Various types of sensors are increasingly used to determine the contamination of liquid media and the concentration of components in them, allowing for direct measurements in real time. The most commonly used are electrochemical, optical and capacitive sensors [2,10,11,12,13]. Electrochemical sensors operate by measuring electrical parameters, such as potential, current or conductivity, which vary depending on the concentration of ions or chemical compounds in the solution. Such sensors have fairly good sensitivity, but often require specific electrode materials and can be sensitive to changes in the composition of the medium and electrode degradation [3,14,15,16,17,18]. Optical sensors, based on the principles of light absorption, fluorescence or scattering, allow for the determination of the concentrations of certain compounds based on their optical properties. Although these sensors can provide high accuracy and selectivity, their design is usually more complex, and the accuracy of measurements may depend on the turbidity of the medium or other optical interference. An alternative to these methods is capacitive sensors, the operating principle of which is based on changes in electrical capacitance resulting from changes in the dielectric properties of the medium with changes in the concentration of pollutants [5,6,19,20,21,22,23]. Capacitive sensors are characterized by simple design, fast response, low energy consumption and the ability to easily integrate them into automated monitoring systems. In addition, they can be used in various liquid media without requiring complex reagents or additional preparation procedures. Due to these properties, capacitive sensors are considered a promising technology suitable for continuous monitoring of environmental and industrial fluid contamination [24,25].

1.4. Research and Literature Gaps

Although ZrO₂ coatings have been widely studied and applied in various sensing and electronic systems for many years due to their favourable dielectric and chemical properties, their role as a controllable design parameter in capacitive sensors operating in conductive liquid media remains insufficiently understood. In particular, the influence of coating thickness on sensor sensitivity and stability has not been systematically quantified. Despite the rapidly growing interest in sensor technologies for monitoring contamination in liquid media, several important scientific and technological gaps still remain in this area. Most existing studies are mainly focused on the study of specific sensor principles or materials in laboratory conditions, but there is still a lack of comprehensive studies examining the performance of sensors in media close to real industrial or environmental conditions, which may contain various mixtures of contaminants and varying physical parameters, such as temperature or conductivity [1,2,3,26,27,28]. In addition, relatively little attention is paid in the literature to increasing the sensitivity of capacitive sensors using additional functional electrode coatings, which could improve the electrical properties and stability of the sensor in aggressive media [1,2,3,4]. Although metal oxide layers are often used in various sensing devices due to their good dielectric and chemical properties, systematic studies examining the influence of such coatings, such as zirconium oxide (ZrO₂), on the performance of capacitive sensors in liquid media pollution measurements are still limited. Also, the literature lacks detailed experimental data on the calibration characteristics, sensitivity and stability of such sensors in the range of different pollutant concentrations. Therefore, there is a need to conduct additional studies aimed at developing advanced capacitive sensor designs, modifying their electrode surfaces and evaluating their performance characteristics in systems close to real environmental or industrial conditions.

In contrast to previous studies, which primarily focus on the general principles of capacitive sensing or the selection of sensing materials, this work provides a systematic and quantitative investigation of the influence of zirconium oxide (ZrO₂) coating thickness on the performance of capacitive sensors operating in conductive liquid media. To the best of the authors’ knowledge, this is one of the first studies to simultaneously combine controlled variation in coating thickness (2–6 µm) with comprehensive multi-method characterization (SEM, EDS, FTIR, XRD, and profilometry) and detailed electrical performance analysis under varying solution concentrations and frequencies. The originality of this study lies in this integrated approach, which enables a direct linkage between coating structure and sensor response. Furthermore, this study establishes a direct relationship between coating thickness and key sensor performance parameters, including sensitivity, stability, and frequency-dependent behaviour, which has not been systematically quantified in the existing literature. Unlike prior works that often consider material properties in isolation, this research demonstrates that coating thickness can be used as a critical design parameter to tune sensor response and optimize performance. Importantly, the work identifies an optimal ZrO₂ coating thickness (~4 µm) that provides a balance between sensitivity and measurement stability, thereby offering practical guidelines for sensor design and implementation in real-time monitoring applications. This contribution advances the understanding of how dielectric coatings influence capacitive sensing mechanisms in conductive environments and provides a foundation for the development of more reliable and application-oriented sensor systems.

The selection of the sensor construction in this study is based on the specific requirements of detecting ionic contaminants, such as dissolved salts, in conductive liquid media. Capacitive sensing is particularly suitable for this purpose, as variations in salt concentration directly affect both the dielectric permittivity and electrical conductivity of the solution, resulting in measurable changes in capacitance. Various capacitive sensor configurations are used for liquid analysis, including coaxial (cylindrical), interdigital (comb-type), and rod-type sensors, each offering specific advantages depending on the application and measurement conditions [1,2,3,4,5]. Coaxial sensors provide stable and well-shielded measurements in flowing media, interdigital structures are suitable for miniaturized and surface-sensitive sensing, while rod-type sensors are commonly applied in industrial tanks for bulk monitoring. In contrast to these designs, a parallel-plate sensor configuration was selected in this work due to its simple and well-defined geometry, which enables the formation of a uniform electric field between the electrodes and facilitates controlled analysis of dielectric effects in conductive media. Compared to other configurations, such as coaxial or interdigital sensors, the parallel-plate design allows for a more direct and interpretable relationship between capacitance and the dielectric properties of the liquid, making it particularly suitable for systematic investigation of coating thickness effects. Furthermore, the use of a dielectric coating ZrO₂ on the electrodes is especially important for salt-containing solutions, where direct electrode–electrolyte interaction can lead to significant parasitic effects, such as electrode polarization and leakage currents. The ZrO₂ layer acts as a dielectric barrier that stabilizes the measurement, reduces unwanted electrochemical interactions, and improves repeatability. Therefore, the selected sensor construction—a ZrO₂-coated parallel-plate capacitive sensor—is well suited for the investigation of salt concentration in liquid media, as it enables controlled, stable, and physically interpretable measurements of dielectric property changes induced by ionic contaminants.

The aim of this study is to develop and investigate a capacitive sensor with a zirconium oxide (ZrO₂) coating for determining the salt concentration of liquid media, and to evaluate the influence of this coating on the sensor capacitance and measurement sensitivity.

2. Materials and Methods

Zirconium oxide (ZrO₂) layers on stainless steel electrodes were formed using reactive magnetron sputtering technology. The deposition process was carried out in the DC magnetron sputtering mode in an argon and oxygen gas mixture environment, maintaining an operating pressure of 1 Pa, which ensured stable layer formation conditions. The flow of argon and oxygen gases into the vacuum chamber was regulated using gas flow meters, maintaining a flow rate of 30 cm³/min of oxygen and 30 cm³/min of argon gas. Deposition of zirconium oxide in an oxygen-only environment is problematic, because oxygen ions bombarding the cathode not only knock out zirconium atoms but also form a thin but dense dielectric layer of zirconium oxide on the cathode. As a result, the dielectric layer is broken down during the discharge (abundant sparks are visible escaping from the cathode), and the electrical parameters of the plasma become unstable. By introducing argon into the chamber, argon ions effectively stabilize plasma parameters and ensure uniform coating growth. A 3-inch diameter metallic zirconium target with a purity of 99.95% manufactured by Testbourne Ltd. (United Kingdom) was used for deposition. A constant 7 cm distance between the magnetron target and the substrate holder was maintained in the deposition system, ensuring a uniform material transfer process. During magnetron sputtering, the target current was set to 0.75 A, and the stainless-steel substrates were not specially heated. The deposition duration of the zirconium oxide layers was 30, 60 and 90 min, which allowed controlling the thickness of the resulting layer (by changing the deposition time, the thicknesses of the obtained ZrO₂ layers reached 2, 4 and 6 µm) and structural properties. Thus, during the magnetron sputtering process, precise control of the main technological parameters—gas flow, pressure, target current and deposition time—was ensured, which allowed the formation of a stable zirconium oxide layer on stainless steel electrodes.

To ensure and evaluate the uniformity of the deposited ZrO₂ coating, the geometric relationship between the sputtering target and the substrate was considered. The diameter of the zirconium target used for magnetron sputtering was ~76 mm, while the coated electrodes had a square geometry with dimensions of approximately 3.3 × 3.3 cm, corresponding to an effective area of about 10 cm². Both electrodes had identical dimensions and were fully coated with the ZrO₂ layer. The electrode size was significantly smaller than the target diameter, which promotes a more uniform distribution of the sputtered material across the substrate surface. In addition, the electrodes were positioned centrally relative to the target during deposition to minimize thickness gradients. Under these conditions, the variation in coating thickness across the electrode surface is expected to be minimal, and the deposited ZrO₂ layer can be considered sufficiently uniform for reliable sensor performance evaluation. The coating thickness uniformity was additionally confirmed by profilometry measurements performed at different positions on the electrode surface.

A scanning electron microscope (SEM) Hitachi S-3400 N was used to evaluate the morphology and geometric properties of the formed zirconium oxide layer on the stainless-steel electrodes. The elemental composition of the electrode surface was determined by energy dispersive X-ray spectroscopy (EDS), which was performed in conjunction with SEM surface analysis studies. This method allowed the identification of the main elements and confirmation of the presence of zirconium oxide in the deposited coating. X-ray diffraction (XRD) was used to determine the crystal structure and predominant compounds of the deposited layer, using the standard Bragg–Brentan focusing geometry. In addition, Fourier transform infrared spectroscopy (FTIR) was used to analyze the chemical bonds and surface structure of the zirconium oxide layer, which allowed identifying characteristic vibrational peaks and assessing the structural properties of the formed oxide layer. Surface topography and roughness parameters were additionally evaluated using the contact profilometer “Ambios XP-200 Profiler”. The device allows for high-resolution surface profile measurements, the vertical resolution of which can reach up to 1 Å at a measurement range of 10 μm. The profilometer also has a maximum scanning length of up to 50 mm and a maximum vertical measurement scale of 400 μm, making it suitable for analyzing the surface topography and relief characteristics of thin coatings.

Figure 1 ZrO₂-coated parallel-plate capacitive sensor and experimental measurement setup. The sensor consists of two flat square stainless-steel electrodes, each with an effective area of approximately 10 cm², both coated with a zirconium oxide (ZrO₂) dielectric layer. The electrodes are positioned parallel to each other with a very small and fixed gap between their facing surfaces, forming a parallel-plate capacitor configuration. The electrical characteristics of the capacitive sensors were evaluated using an LCR meter, which allows for measurements by changing the measurement current frequency in the range of 10 Hz–300 kHz. This method was chosen because when measuring capacitance in conductive media, such as salt solutions, significant leakage currents are generated, which can distort the results of true capacitance measurements using simple capacitance meters. In order to reduce the influence of leakage currents, capacitance measurements were performed over a wide frequency range, with special attention paid to higher frequencies (e.g., ≥10 kHz), where the effect of leakage currents becomes smaller. First, the dependence of the sensor capacitance on frequency at a constant solution concentration was determined in order to identify the frequency range in which the capacitance value remains stable and can be considered close to the true dielectric capacitance. Subsequently, analogous frequency measurements were performed by changing the solution concentration, thus assessing the changes in capacitance. It should be noted that in conductive solutions, capacitance changes are most often determined by the change in leakage currents due to changes in the conductivity of the medium, while in low-conductivity media, capacitance changes are mainly related to the change in the total dielectric permittivity of the medium. Therefore, for different materials, the methodology for measuring capacitance and calibrating the sensor may differ, depending on whether the material under study changes the electrical conductivity of the medium more or its dielectric properties.

3. Results and Discussion

In the context of this study, SEM microscopy images obtained at 2 µm zirconium oxide (ZrO₂) thickness and 1000× and 5000× magnification show that the resulting coating on the stainless-steel substrate has a relatively heterogeneous microstructure (Figure 2a,b). Fine linear or elongated microrelief elements oriented along the surface are observed over the entire surface, which are visible both at lower (1000×) and higher (5000×) magnification. In addition, very fine micro-indentations are identified in some places. Such surface morphology may be related to the fact that during the formation of the ZrO₂ coating, it partially replicates the microrelief of the stainless-steel substrate, which occurs due to surface structures formed during mechanical processing or polishing.

On the other hand, a similar microstructure can also be formed due to the coating growth mechanism itself during the sputtering process, when particle condensation and surface diffusion lead to the directional formation of structures and the appearance of local irregularities. Such a microrelief structure can affect the surface area of the coating and its interaction with the liquid under study, therefore it is important in assessing the sensitivity of a capacitive sensor and changes in its electrical properties when measuring the concentration of liquid components.

In the context of this study, SEM images obtained at 4 µm ZrO₂ thickness and 1000× and 5000× magnification can be interpreted as showing a more advanced stage of coating formation, when a thicker and more continuous zirconium oxide layer is formed on the stainless-steel substrate (Figure 2c,d). Compared to the 2 µm coating, a longer deposition time (when a higher thickness is obtained) usually leads to better substrate coverage, a more pronounced formation of a specific coating microstructure and a smaller influence of the substrate relief on the final surface morphology. If linear or elongated irregularities are still visible in the SEM images, this may indicate that the coating partially replicates the topography of the steel surface, but at the same time it is likely that structural elements determined by the growth of the ZrO₂ coating itself become apparent due to the longer deposition. Small surface irregularities, local thickenings or micro-indentations can be attributed to the effects of particle condensation, layer densification and internal stresses during coating formation. Such a more developed microstructure is important for the performance of a capacitive sensor, as it can increase the effective interaction surface with the liquid medium, change the local electric field at the sensor surface and thus affect the sensitivity and response of the sensor when measuring the concentration of liquid components [1,2,3].

In order to avoid damage to the thin film morphology and ensure the formation of a uniform and stable ZrO₂ coating, careful control of the deposition parameters is essential. The results of this study indicate that film morphology is strongly influenced by deposition time, plasma stability, and substrate-related factors. In particular, insufficient deposition time leads to incomplete surface coverage and the replication of substrate microrelief, while excessive growth may result in increased surface roughness and internal stresses. These considerations are consistent with the SEM observations obtained in this study, where thinner films (~2 µm) exhibited pronounced substrate-induced morphology, while thicker coatings (~4–6 µm) showed more continuous structures but increased surface roughness and microstructural irregularities. To minimize morphological defects such as micro-cracks, irregular growth structures, or surface inhomogeneities, several strategies can be applied. First, maintaining stable plasma conditions during magnetron sputtering is critical, which can be achieved by using a controlled argon–oxygen gas mixture to prevent discharge instability and arcing effects. Second, optimizing deposition time allows achieving a continuous and sufficiently dense coating without excessive roughness; in this study, a thickness of approximately 4 µm demonstrated a favourable balance between structural integrity and uniformity. Third, proper substrate preparation, including cleaning and surface polishing, reduces the influence of initial surface defects and improves coating adhesion. Additionally, controlling residual stresses in the growing film is important to prevent mechanical degradation of the coating. This can be achieved by adjusting deposition parameters such as pressure, target current, and deposition rate. A moderate deposition rate promotes more uniform film growth and reduces the likelihood of defect formation. Therefore, the combination of optimized deposition conditions and controlled film thickness is a key factor in preventing morphological damage and ensuring reliable performance of ZrO₂-coated capacitive sensors.

EDS elemental composition analysis allows us to assess the formation of a zirconium oxide (ZrO₂) coating on a stainless-steel substrate, depending on the duration of magnetron sputtering. At a thickness of 2 µm, the elements C, O, Cr, Fe, Ni and Zr were identified in the spectrum. The largest share of atomic concentrations is oxygen (28.8 at. %), which confirms the formation of an oxide layer on the surface of the coating. However, at the same time, a relatively high concentration of iron (38.4 at. %), as well as chromium (10.6 at. %) and nickel (5 at. %), which are characteristic elements of a stainless-steel substrate, were detected. This indicates that with a shorter deposition time (smaller thickness), the ZrO₂ coating formed is relatively thin; therefore, a significant signal of substrate elements is still recorded during EDS analysis. The zirconium concentration reaches 8.7 at. %, which confirms the formation of a zirconium-containing coating on the surface. Meanwhile, at a 4 µm thick magnetron sputtering time, the elemental composition changes noticeably—the largest share is made up of oxygen (47.5 at. %) and zirconium (22 at. %), which indicates a more developed and thicker zirconium oxide layer. At the same time, the concentrations of substrate elements—iron (12.9 at. %), chromium (4 at. %) and nickel (1.3 at. %)—decrease significantly, which confirms that a longer deposition time (60 min) leads to better substrate coverage and greater coating integrity. In both cases, a small amount of carbon (8.5–12.1 at. %), which is most likely related to hydrocarbons adsorbed on the surface or possible contamination during deposition or measurement, was detected. Such EDS results confirm the successful formation of the ZrO₂ coating and show that with increasing deposition time, the coating thickness and its dominance in the surface elemental composition increase, which is important in assessing the dielectric properties of the coating and its suitability for capacitive sensor applications in liquid concentration measurements [1,2,3,4].

Interpretation of the FTIR spectrum obtained for the ZrO₂ coating deposited under 30 min sputtering conditions (2 µm) shows that the most important absorption bands are concentrated in the low wavenumber region, which is characteristic of metal–oxygen bonds (Figure 3a). The brighter minima at around 530 and 442 cm⁻¹ can be attributed to Zr–O and Zr–O–Zr vibrations, respectively, and these results confirm that a zirconium oxide layer has formed on the stainless-steel surface. Additional, weaker bands around 766–862 cm⁻¹ can be attributed to more complex oxide lattice vibrations or surface structure irregularities formed during the growth of the coating. Meanwhile, weak signals in the higher wavenumber region, including the region around 1700 cm⁻¹ and broader changes above 3000 cm⁻¹, are most likely not related to the main ZrO₂ lattice, but to adsorbed moisture, surface hydroxyl groups or environmental impurities, which are often found in the FTIR spectra of thin oxide coatings. Since the bands characteristic of ZrO₂ were identified in the low wavenumber region, it can be stated that the sputtering time of 30 min was sufficient for the formation of the zirconium oxide coating, although the relatively low absorption intensity suggests that the layer is still relatively thin and its formation is not completely completed, which is in good agreement with the SEM and EDS results, which indicate the partial survival of the substrate signal and the still developing microstructure of the coating. The interpretation of the FTIR spectrum obtained for the ZrO₂ coating deposited for 60 min (thickness 4 µm) under sputtering conditions shows a more pronounced formation of the zirconium oxide layer than in the case of shorter sputtering (Figure 3b). The most important absorption bands are observed in the low wavenumber region, especially around 665 cm⁻¹, as well as weaker bands around 566, 480 and 457 cm⁻¹, which can be attributed to the vibrations of the Zr–O and Zr–O–Zr bonds, which are characteristic of the zirconium oxide lattice. The intensity of these bands allows us to state that after 60 min of deposition time, a more developed and thicker ZrO₂ layer has formed, therefore the signal of the oxide coating in the FTIR spectrum becomes clearer and dominant. In the higher wavenumber region, weak changes around 1654 cm⁻¹ and very slight signals in the broad region of 3700–3900 cm⁻¹ are most likely associated with adsorbed moisture and surface hydroxyl groups, which are characteristic of oxide coating surfaces after contact with the environment. Also, a small band around 2349 cm⁻¹ may be related to atmospheric CO₂ adsorption during the measurement, therefore it should not be considered a main characteristic of the coating structure. In conclusion, it can be stated that the sputtering time of 60 min (thickness 4 µm) created more favorable conditions for the formation of a more pronounced zirconium oxide coating, and the FTIR data are in good agreement with the SEM and EDS results, indicating a thicker, more continuous layer with a higher zirconium and oxygen content, which is favorable for the stability of the dielectric properties of the capacitive sensor and sensitivity in the concentration measurements of liquid components.

XRD analysis shows that at a zirconia coating thickness of 2 µm, one main peak characteristic of the coating appears in the diffractogram at approximately 28° (2θ), while the remaining intense peaks are mainly attributed to the stainless-steel substrate, which allows us to state that the formed ZrO₂ layer is still relatively thin and X-ray radiation easily reaches the substrate (Figure 4a). The peak at ~28° is characteristic of the reflex of the monoclinic zirconia phase and is usually associated with the m-ZrO₂ (−111) plane, therefore it can be stated that at a thickness of 2 µm, a crystalline monoclinic ZrO₂ phase with the predominant orientation of this plane is already forming in the coating. Since other clear ZrO₂ peaks are almost invisible or overlap with the substrate signals, it is likely that the crystallinity of the coating is not yet very high, and the amount of crystallites and the layer thickness are not sufficient for the formation of a brighter diffraction pattern of the entire ZrO₂. This result indicates the initial stage of the formation of a crystalline coating, when zirconium oxide is no longer completely amorphous, but its diffraction response is still weak due to the small layer thickness and significant substrate influence. This interpretation is in good agreement with the SEM, EDS and FTIR results, which show that after 30 min of deposition (2 µm layer thickness), a zirconium oxide layer is formed, but it is not yet completely uniform and thick enough, so the substrate signal remains bright, which is important for assessing the structural and dielectric properties of the coating for the application of a capacitive sensor in liquid component concentration measurements [2,3,4,5].

In Figure 4b, XRD analysis shows that at a zirconia coating thickness of 4 µm, the peaks of the ZrO₂ crystalline phase at approximately 24°, 28°, 32°, 34° and 55° (2θ) become more clearly visible in the diffractogram, which indicates a more formed and higher crystallinity layer than after 30 min sputtering (with a thickness of 2 µm). The peak at ~24° can be associated with the m-ZrO₂ (−110) plane of the monoclinic zirconia phase, at ~28°—with m-ZrO₂ (−111), at ~32°—with m-ZrO₂ (111), and at ~34°—with m-ZrO₂ (200) plane; the additional peak at ~55° is also characteristic of the monoclinic phase reflections and can be associated with the diffraction of higher-order m-ZrO₂ planes. Such occurrence of several characteristic peaks allows us to state that the crystalline monoclinic zirconia phase prevails in the 4 µm coating, and the most intense reflex at ~28° indicates a preferential orientation along the m-ZrO₂ (−111) plane. Since some of the ZrO₂ peaks overlap with the peaks of the stainless-steel substrate, the influence of the substrate on the diffractogram remains significant, but, compared to the case of 2 µm sputtering, the higher number of identifiable ZrO₂ reflexes indicates a thicker layer and more advanced formation of the crystalline structure. Such a result allows us to conclude that a longer sputtering time promotes not only an increase in the thickness of the coating, but also an increase in its crystallinity, resulting in a more structured ZrO₂ layer, which should be more favorable for stable dielectric properties and better response of the capacitive sensor when measuring the concentration of liquid components. XRD analysis shows that at a zirconia coating thickness of 6 µm, the same ZrO₂ characteristic reflections at approximately 24°, 28°, 32°, 34° and 55° (2θ) remain in the diffractogram (Figure 4c), and their position and overall diffraction pattern are very similar to the 60 min sputtering case. These peaks can be assigned to the monoclinic zirconia phase: at ~24°—m-ZrO₂ (−110), at ~28°—m-ZrO₂ (−111), at ~32°—m-ZrO₂ (111), at ~34°—m-ZrO₂ (200), and the peak at ~55° also corresponds to the higher-order diffraction of the monoclinic phase. The most intense reflection at ~28° allows us to conclude about the prevailing preferential orientation along the m-ZrO₂ (−111) plane. Since some of the peaks characteristic of zirconium oxide overlap with the reflections of the stainless steel substrate, the influence of the substrate on the diffractogram still remains, but several clearly identifiable ZrO₂ peaks confirm that the layer formed after 90 min (6 µm) sputtering is crystalline and sufficiently developed. The fact that the diffraction pattern is very similar to the 4 µm case indicates that with further increasing coating thickness, no significant changes in the phase composition occur, and the same monoclinic ZrO₂ phase continues to dominate in the coating. Therefore, it can be stated that in the interval between 4 and 6 µm, the crystalline structure of the coating is essentially stabilized, and the main changes are most likely more related to the increase in layer thickness or a slight improvement in crystallinity, but not to the formation of new phases. This result is important in the research context, as it shows that increasing thickness allows maintaining a structurally stable zirconia coating, which is favorable for repeatable dielectric properties and reliable capacitive sensor performance when measuring the concentration of liquid components (e.g., salt).

The results of profilometric measurements show that with a 30 min ZrO₂ sputtering time, a relatively thin, but already clearly measurable zirconium oxide layer was formed on the stainless-steel substrate, the step height of which according to the profile curve is approximately in the order of several micrometers, about 2 µm (Figure 5a). The height difference between the coated and uncoated areas visible in the profilogram confirms the successful deposition of the coating, while the observed profile fluctuations and irregularities indicate that the surface is not ideally smooth. Such a shape can be associated both with the influence of the microrelief of the initial stainless-steel substrate and with the growth characteristics of the ZrO₂ coating itself during sputtering, when the layer is formed by uneven densification and repeats part of the substrate topography. The relatively small coating thickness of 2 µm is in good agreement with SEM, EDS, FTIR and XRD results, which show that after 30 min of sputtering, the zirconium oxide layer has already formed, but still remains quite thin, therefore the influence of the substrate on both the surface morphology and the elemental and structural characteristics is still noticeable. Such a coating state is important for the operation of a capacitive sensor, because even a small but already formed oxide layer can change the dielectric properties of the electrode surface and affect the sensitivity of the sensor when measuring the concentration of liquid components [1,3,11].

The profilometric measurement results show that at a ZrO₂ sputtering time of 60 min, a clearly measurable zirconium oxide layer was formed on the stainless-steel substrate, the thickness of which, according to the difference in the height of the profilogram, is approximately 4 µm (Figure 5b). The sharp transition between the coated and uncoated areas visible in the profile confirms the successful deposition of the coating, and the rather large step height indicates that the longer sputtering time led to more intensive layer growth and better substrate coverage. At the same time, the height fluctuations and surface irregularities observed in the profilogram allow us to state that the coating is not ideally smooth, and its relief is formed by both the initial topography of the stainless-steel surface and the growth characteristics of the ZrO₂ coating itself during the sputtering process. Such a thicker coating reduces the influence of the substrate on the surface properties, therefore this result is in good agreement with the SEM, EDS and XRD data, which in the case of 4 µm show a more developed ZrO₂ coating with a higher zirconium and oxygen content and a clearer crystalline structure. In the context of the publication, this allows us to state that a 60 min sputtering time is sufficient to form a relatively thick and continuous zirconium oxide layer, which should be favorable for more stable dielectric properties and more efficient operation of the capacitive sensor when measuring the concentration of liquid components. The results of the profilometric measurement show that a clearly measurable and sufficiently thick zirconium oxide layer was formed on the stainless-steel substrate with a 90 min ZrO₂ sputtering time. According to the difference in the height of the presented profile, the thickness of the coating can be estimated as approximately 6 µm (Figure 5c). The sharp transition between the coated and uncoated areas visible in the profilogram confirms the successful deposition of the coating, and the general nature of the profile indicates that the layer becomes more developed with increasing sputtering time. At the same time, quite pronounced height fluctuations and local irregularities observed in the curve allow us to state that the surface is not smooth, and its relief is determined by both the initial topography of the stainless-steel substrate and the growth mechanism of the ZrO₂ coating itself during deposition. Such morphology is characteristic of thicker oxide coatings, when the layer growth is uneven, and a more pronounced microrelief is formed on the surface. The sputtering time of 90 min allows the formation of a well-developed ZrO₂ layer, which should more effectively insulate the substrate, ensure more stable dielectric properties and be favourable for the operation of the capacitive sensor when measuring the concentration of liquid components.

Analysing the frequency characteristics of the capacitive sensor, when the thickness of ZrO₂ layer is 2 µm, it was found that the capacitance strongly depends on the excitation frequency (Figure 6a). At low frequencies (about 10–100 Hz), high capacitance values are observed, which increase with increasing NaCl concentration in the solution. In this frequency range, the effects of electrode polarization and double electric layer dominate, which lead to the formation of additional “apparent” capacitance. With increasing frequency, the capacitance gradually decreases, as the contribution of ion migration and polarization decreases. In the higher frequency range (above ~10 kHz), the capacitance values stabilize and become less dependent on frequency, which indicates that in this range the measurement is close to the true capacitance of the sensor, determined mainly by the dielectric properties. However, in this region, the curves for different NaCl concentrations converge, which reduces the sensitivity of the sensor to concentration changes. Thus, with a 2 µm thick ZrO₂ layer, the sensor exhibits high sensitivity in the low frequency range, but measurements in this range are strongly affected by parasitic electrochemical processes.

When analysing the frequency characteristics of the capacitive sensor with ZrO₂ layer thickness of 4 µm, a similar overall capacitance dependence on frequency is observed as in the case of a thinner layer, but with clear differences (Figure 6b). During the experiments, the amount of salt was varied from 0.1 to 3 g, using a volume of 125 mL of water. At low frequencies, the capacitance still increases with increasing NaCl concentration, but the absolute capacitance values are smaller, and the separation of the curves between different concentrations is less pronounced than in the 2 µm case. This indicates a reduced effect of electrode polarization and the electric double layer due to the thicker dielectric layer. As the frequency increases, the capacitance decreases and stabilizes in the higher frequency range (~10 kHz and above), reflecting the true dielectric properties of the sensor. In this frequency range, better stability of the curves and lower noise are observed, but at the same time the sensitivity to concentration changes decreases. Thus, a 4 µm thick ZrO₂ layer provides a better compromise between measurement stability and sensitivity, reducing the influence of parasitic electrochemical processes. The role of the ZrO₂ coating in the capacitive sensor can be explained based on its dielectric and insulating properties, which fundamentally modify the electrode–electrolyte interface. In the absence of such a coating, the stainless-steel electrode is in direct contact with the conductive liquid, leading to significant charge transfer processes, including ion migration, electrode polarization, and the formation of an electrical double layer. These effects introduce substantial leakage currents and result in an additional apparent capacitance that distorts the true dielectric response of the system. The introduction of a ZrO₂ layer acts as a dielectric barrier between the electrode and the electrolyte, effectively suppressing faradaic processes and limiting direct charge exchange at the interface. As a result, the contribution of leakage currents is significantly reduced, and the measured capacitance becomes increasingly governed by the dielectric properties of the coating and the surrounding medium rather than by electrochemical phenomena. From a physical perspective, the ZrO₂ layer increases the effective impedance of the electrode–electrolyte interface, particularly at low frequencies, where polarization effects are typically dominant. The thickness of the coating plays a critical role in this mechanism: a thinner layer (e.g., ~2 µm) only partially suppresses ionic conduction pathways, allowing residual leakage currents and interfacial polarization effects to persist. In contrast, a thicker layer (e.g., ≥4 µm) provides more effective insulation, significantly attenuating these parasitic processes and stabilizing the capacitive response. However, increasing the coating thickness also reduces the overall electric field penetration into the liquid medium, which decreases the sensitivity of the sensor to changes in the dielectric properties of the solution.

Therefore, the observed trade-off between sensitivity and stability can be directly attributed to the balance between leakage current suppression and electric field coupling through the dielectric layer. This mechanism explains the experimental observations obtained in this study, where thinner coatings exhibit higher apparent sensitivity but are strongly affected by parasitic electrochemical effects, while thicker coatings provide more stable and repeatable measurements at the expense of reduced sensitivity. Consequently, the ZrO₂ layer functions not only as a protective coating but also as a key element controlling the interfacial physics and overall sensing behaviour of the capacitive system.

When the thickness of the ZrO₂ layer is increased to 6 µm, the frequency characteristics of the capacitive sensor become even more stable, but at the same time the sensitivity decreases noticeably (Figure 6c). In the low-frequency range, the capacitance values are significantly lower than in the 2 µm and 4 µm cases, and the curves for different NaCl concentrations are closer to each other, which indicates a weaker effect of electrode polarization. With increasing frequency, the capacitance decreases evenly and quickly reaches a stable level in the higher frequency range (~50 kHz), where the real dielectric capacitance dominates [1,2,3,4,5]. In this range, the curves almost coincide, so the possibilities for determining the concentration become limited. This indicates that an excessively large ZrO₂ layer thickness reduces the sensitivity of the sensor, although it improves measurement stability and reduces the influence of parasitic effects.

The selection of NaCl as the test medium in this study is based on both its well-defined physicochemical properties and its relevance to practical applications. NaCl is a commonly used model electrolyte, as it fully dissociates in aqueous solutions into Na⁺ and Cl⁻ ions, providing a stable and predictable change in electrical conductivity and dielectric properties with concentration. This makes it particularly suitable for evaluating the sensitivity and response characteristics of capacitive sensors under controlled conditions. In addition, NaCl solutions are widely used as a reference system in studies of liquid conductivity and salinity, which enables comparison of the obtained results with existing literature. From an application perspective, NaCl is also representative of real-world scenarios, as dissolved salts are among the most common contaminants in natural and industrial water systems. Therefore, the use of NaCl allows both reliable characterization of sensor performance and practical relevance of the obtained results. When analysing the calibration characteristic of a capacitive sensor with ZrO₂ layer thickness of 2 µm, it was found that the normalized capacitance (nF/cm²) increases markedly with increasing NaCl concentration in the solution (Figure 7a). The obtained dependence is characterized by a large slope, which indicates a high sensitivity of the sensor to concentration changes. However, the calibration curve is not linear, and the spread between experimental points is greater than in the case of thicker layers. In this case, the most stable calibration curve is obtained only at a frequency of 10 Hz. This is due to the fact that with a thin ZrO₂ layer, the measurement signal is strongly affected by the electrode polarization and double electric layer effects, which are especially significant in the low-frequency range. Due to these parasitic phenomena, the measured capacitance includes not only the true dielectric capacitance, but also an additional “apparent” capacitance, therefore the calibration characteristic becomes less stable and less repeatable. Nevertheless, the 2 µm thick ZrO₂ layer provides the highest sensitivity and may therefore be suitable for applications where high signal is important but lower requirements are placed on measurement accuracy.

When analyzing the calibration characteristic of the capacitive sensor with ZrO₂ layer thickness of 4 µm, it was found that the normalized capacitance (nF/cm²) increases with increasing NaCl concentration, and the dependence is close to linear (Figure 7b). Compared to the 2 µm case, the calibration curve is characterized by significantly lower dispersion and better linearity, which indicates a more stable sensor response. This can be explained by the reduced influence of electrode polarization and double electric layer due to the thicker ZrO₂ layer, which more effectively isolates the electrode from the solution. As a result, the measured capacitance more closely reflects the true dielectric properties of the sensor, rather than parasitic electrochemical phenomena [4,5,6,7]. Although the sensitivity decreases compared to the 2 µm case, it remains high enough for practical applications. In addition, the most stable calibration curve is obtained at a frequency of 10 kHz, which means that in this case the sensitivity of the sensor is reflected by the change in the dielectric capacitance of the medium and not by parasitic capacitive phenomena that are characteristic of low measurement frequencies (e.g., 10 Hz and above). Therefore, a 4 µm thick ZrO₂ layer can be considered an optimal compromise between sensitivity and measurement accuracy. When the thickness of the ZrO₂ layer is increased to 6 µm, the calibration characteristic remains monotonic, but has a smaller slope, which indicates a decreased sensitivity of the sensor to changes in NaCl concentration (Figure 7c). The dependence of the normalized capacitance (nF/cm²) on the concentration is quite linear, and the dispersion of the experimental points is small, which indicates good measurement stability and repeatability. A thicker ZrO₂ layer further reduces the influence of electrode polarization and other parasitic processes, so the measured capacitance is closer to the true dielectric capacitance. However, due to the increased thickness of the dielectric layer, the interaction of the electric field with the solution decreases, so changes in concentration cause smaller changes in capacitance. For this reason, the 6 µm thick ZrO₂ layer is more suitable for applications where measurement stability is the most important, but lower sensitivity may limit its use for detecting small concentration changes.

After evaluating the calibration characteristics of the capacitive sensor for different thicknesses of ZrO₂ layers, it was found that the measurement accuracy and error significantly depend on the layer thickness. For a 2 µm thick ZrO₂ layer, the calibration data are characterized by the largest dispersion, and the coefficient of determination reaches about R² ≈ 0.72, which corresponds to a relative error of about 10–20% depending on the concentration range. By increasing the layer thickness to 4 µm, a significant improvement in accuracy is observed—the calibration curve becomes almost linear, and the coefficient of determination increases to R² ≈ 0.97, respectively, reducing the relative error to ~3–5%. Meanwhile, for a 6 µm thick ZrO₂ layer, the calibration dependence remains quite linear (R² ≈ 0.96), and the error remains at a similar level (~4–6%), but due to lower sensitivity, the absolute changes in capacitance are smaller, which may increase the measurement uncertainty at low concentrations. Thus, although thicker ZrO₂ layers ensure lower dispersion and better repeatability, the optimal ratio of accuracy and sensitivity in this work is achieved at a layer thickness of ~4 µm.

To quantitatively evaluate the performance of the sensor, the sensitivity was determined from the slope of the calibration curves, defined as the change in normalized capacitance with respect to NaCl concentration (dC/dc). The sensitivity was calculated by performing linear fitting of the experimental data in the most stable frequency range for each coating thickness. The calculated sensitivity values are approximately:

-

0.85 nF/(cm²·g/L) for the 2 µm ZrO₂ layer,

-

0.62 nF/(cm²·g/L) for the 4 µm ZrO₂ layer,

-

0.41 nF/(cm²·g/L) for the 6 µm ZrO₂ layer.

The results show that the sensor with a ZrO₂ layer thickness of 2 µm exhibits the highest sensitivity, characterized by the steepest slope of the calibration curve. However, this increased sensitivity is accompanied by greater data dispersion and reduced linearity, indicating the influence of parasitic electrochemical effects. For the sensor with a 4 µm thick ZrO₂ layer, the sensitivity is slightly lower, but the calibration curve demonstrates significantly improved linearity and reduced measurement uncertainty. This indicates that the measured capacitance more accurately reflects the dielectric properties of the solution rather than parasitic contributions. As a result, the effective sensitivity in practical applications can be considered more reliable. In the case of the 6 µm thick ZrO₂ layer, the sensitivity decreases further, as evidenced by the reduced slope of the calibration curve. Although the measurements show excellent stability and repeatability, the lower sensitivity limits the sensor’s ability to detect small changes in concentration. Overall, the sensitivity analysis confirms that there is a trade-off between sensitivity and measurement stability. While thinner coatings provide higher nominal sensitivity, thicker coatings improve measurement reliability. The optimal performance is achieved at a ZrO₂ thickness of approximately 4 µm, where sufficient sensitivity is retained while ensuring stable and reproducible measurements. The sensitivity of ZrO₂-based capacitive sensors can be improved by optimizing the thickness of the dielectric coating and the measurement conditions. The results of this study show that thinner ZrO₂ layers (~2 µm) provide higher sensitivity due to a stronger influence of the liquid medium on the effective capacitance. However, this comes at the expense of increased parasitic effects. Therefore, sensitivity enhancement can be achieved by selecting a relatively thin coating while operating at higher frequencies (≥10 kHz), where the influence of electrode polarization and leakage currents is reduced. In addition, increasing the effective interaction area between the sensor surface and the liquid, as well as improving the uniformity of the coating, can further enhance the sensor response. Thus, an optimal balance between coating thickness, frequency range, and structural quality is essential for maximizing sensitivity while maintaining measurement stability.

To quantitatively evaluate the repeatability and long-term stability of the proposed sensor, additional measurements were performed under identical experimental conditions. Repeatability was assessed by performing multiple consecutive measurements (n = 5) at fixed NaCl concentrations. The relative standard deviation (RSD) of the normalized capacitance values was calculated for each coating thickness. The obtained RSD values were approximately: 6.8% for the 2 µm ZrO₂ layer, 3.2% for the 4 µm ZrO₂ layer, 2.5% for the 6 µm ZrO₂ layer. These results indicate that repeatability improves with increasing coating thickness, which can be attributed to the reduced influence of parasitic electrochemical processes and more stable dielectric behavior. Long-term stability was evaluated by monitoring the sensor response over a period of 24 h under constant conditions. The capacitance drift was found to be: ~5.5% for the 2 µm layer, ~2.1% for the 4 µm layer, ~1.6% for the 6 µm layer. The reduced drift observed for thicker coatings confirms that the ZrO₂ layer effectively stabilizes the electrode–electrolyte interface and minimizes time-dependent effects such as surface polarization and ionic accumulation. Overall, the results demonstrate that while thinner coatings provide higher sensitivity, thicker coatings significantly improve repeatability and long-term stability. The 4 µm ZrO₂ layer offers an optimal balance, ensuring sufficiently low measurement variability while maintaining adequate sensitivity for practical applications.

Temperature is a critical factor influencing the performance of capacitive sensors, as it affects both the dielectric properties of the liquid medium and the electrical behavior of the sensing layer. In conductive solutions such as NaCl, an increase in temperature leads to enhanced ionic mobility and higher electrical conductivity, which in turn can increase leakage currents and alter the apparent capacitance of the sensor. Furthermore, the dielectric permittivity of the liquid medium is inherently temperature-dependent and typically decreases with increasing temperature, directly influencing the measured capacitance. In similar capacitive sensing systems, temperature variations can lead to capacitance changes on the order of several percent per 10 °C, depending on the properties of the medium and the sensor configuration. The ZrO₂ coating itself exhibits relatively stable dielectric properties over moderate temperature ranges; however, temperature variations may still affect interfacial processes between the coating and the electrolyte. In this study, all measurements were carried out under controlled laboratory conditions at a constant temperature of approximately 20–22 °C to minimize thermal effects. This ensures that the observed changes in capacitance are primarily attributed to variations in NaCl concentration and coating thickness rather than temperature-induced influences. The presence of the ZrO₂ layer is also expected to mitigate temperature-related parasitic effects by acting as a dielectric barrier, limiting direct charge transfer and stabilizing the electrode–electrolyte interface. As a result, sensors with thicker dielectric coatings may exhibit improved robustness against temperature fluctuations. For practical applications, particularly in industrial and environmental monitoring where temperature variations are unavoidable, the implementation of temperature compensation or calibration strategies is essential to ensure accurate and reliable measurements. Future work will focus on a systematic investigation of temperature-dependent sensor behavior and the development of compensation methods to enhance performance under varying environmental conditions. Capacitive sensors for component concentration measurement in liquid media operate based on changes in the dielectric properties of the medium. The capacitance of such sensors depends on the dielectric permittivity (ε) of the liquid, which is strongly influenced by the presence of dissolved ions and impurities. For example, pure water has a high dielectric constant (~80 at room temperature), while the addition of salts, such as NaCl, alters both the effective permittivity and electrical conductivity of the solution, leading to measurable changes in capacitance. The selection of sensor design is directly related to the method of interaction with the liquid medium and the specific parameter being measured, such as dielectric permittivity, conductivity, or contamination level. In practical applications, electrode surfaces are often modified with dielectric coatings to improve sensor performance. Such coatings reduce direct charge transfer, suppress parasitic electrochemical processes, and protect the electrodes from corrosion and contamination, thereby enhancing measurement stability and repeatability. In this study, zirconium oxide (ZrO₂) is used as a dielectric coating due to its chemical stability and favorable dielectric properties. Various capacitive sensor configurations are used for liquid analysis, including coaxial (cylindrical), interdigital (comb-type), and rod-type sensors, each offering specific advantages depending on the application and measurement conditions. Coaxial sensors provide stable and well-shielded measurements in flowing media, interdigital structures are suitable for miniaturized and surface-sensitive sensing, while rod-type sensors are commonly applied in industrial tanks for bulk monitoring. In contrast to these designs, the sensor developed in this work is based on a parallel-plate configuration, consisting of two flat electrodes positioned opposite each other at a small distance. This geometry provides a well-defined and uniform electric field distribution in the measurement volume and enables controlled investigation of the influence of ZrO₂ coating thickness on sensor sensitivity, stability, and overall performance in conductive liquid media.

To better evaluate the performance of the proposed sensor, it is important to compare it with state-of-the-art capacitive sensors reported in the literature for liquid component concentration measurement. Existing capacitive sensing approaches are generally based on either uncoated metallic electrodes or electrodes modified with various dielectric or functional layers, such as polymer films or metal oxides [1,2,3,4]. In studies employing uncoated electrodes, high sensitivity is often achieved due to direct interaction between the electrode and the conductive liquid [2,3,4,5,6]. However, such configurations are typically strongly affected by electrode polarization, double layer formation, and leakage currents, resulting in poor measurement stability and limited repeatability, especially at low frequencies. In contrast, sensors utilizing dielectric or oxide coatings demonstrate improved stability and reduced influence of parasitic electrochemical processes [2,3,4,5,6], although often at the expense of reduced sensitivity. Compared to these approaches, the sensor developed in this work demonstrates a balanced performance by optimizing the thickness of the ZrO₂ coating. Specifically, the identified optimal thickness (~4 µm) enables sufficient suppression of leakage currents and interfacial polarization effects while maintaining adequate sensitivity to changes in liquid composition. While similar trends have been qualitatively reported in previous studies [3,4,5,6,7], they are typically not systematically quantified or correlated with coating thickness. Furthermore, many previously reported capacitive sensors focus on either qualitative detection or operate under limited frequency ranges [2,3,4,5,6]. In contrast, the present study evaluates sensor performance over a wide frequency range (10 Hz–300 kHz) and establishes clear relationships between frequency, coating thickness, and measurement behavior. This represents a key advantage of the proposed approach, as it provides both mechanistic insight and practical guidelines for sensor optimization. Overall, the proposed ZrO₂-coated capacitive sensor offers improved measurement stability, reduced parasitic effects, and tunable sensitivity, making it competitive with—and in terms of systematic optimization and frequency-domain analysis, more comprehensive than—existing capacitive sensing solutions for real-time monitoring of liquid media.

Although the experimental investigation in this study was focused on NaCl aqueous solutions, the operating principle of the ZrO₂-coated capacitive sensor is not limited to a specific type of liquid. The sensor response is governed by changes in the dielectric permittivity and electrical conductivity of the medium; therefore, it can be applied to a wide range of liquid systems where these properties vary with composition. In particular, the sensor could be used for monitoring other electrolyte solutions containing dissolved salts (e.g., KCl, CaCl₂), as well as industrial fluids with varying ionic content. Furthermore, it is suitable for detecting changes in water quality parameters, including contamination by dissolved inorganic compounds, acids, or bases, which alter the dielectric properties of the medium. The sensor may also be applicable to organic or mixed liquid systems, such as oil–water emulsions or chemical process fluids, where variations in composition lead to measurable changes in effective permittivity. Therefore, the proposed ZrO₂-based capacitive sensor demonstrates potential for broader application in environmental monitoring and industrial process control, particularly in systems where rapid, real-time detection of changes in liquid composition is required. However, for each specific liquid system, appropriate calibration is necessary due to differences in dielectric and conductive properties.

Capacitive sensors of this type, with a metal oxide layer, have great potential for applications in the field of wastewater monitoring [2,3,4,5,6]. Due to their simple design, possibility of miniaturization and fast response, they can be used for real-time monitoring of dissolved salts or total ionic concentration. In addition, as this study shows, by properly selecting the layer thickness, it is possible to combine the sensitivity and stability of the sensor, which is especially important when working with complex, changing wastewater. Such sensors can also be calibrated to detect various hazardous pollutants and are used for rapid primary monitoring of pollutants, allowing for rapid identification of potential pollution sources. This creates the prerequisites for their integration into automated water quality control systems and contributes to more efficient management of environmental protection and wastewater treatment processes.

4. Conclusions

In this work, a capacitive sensor with a zirconium oxide (ZrO₂) coating formed on stainless steel electrodes was developed and tested for measuring the concentration of components of liquid media. Structural and morphological studies (SEM, EDS, FTIR and XRD) confirmed the successful formation of the ZrO₂ coating. It was found that with increasing coating thickness, more continuous, uniform and more crystalline layers are formed, while in the case of thinner coatings, the influence of the substrate remains more pronounced. Experimental results showed that the thickness of the ZrO₂ layer has a significant impact on the performance of the capacitive sensor. At a smaller coating thickness (~2 µm), the sensor is more sensitive to changes in NaCl concentration, but the measurements are less stable due to the pronounced polarization of the electrode and the influence of parasitic phenomena. As the coating thickness increases, the influence of these phenomena decreases, and the stability of measurements improves, but at the same time the sensitivity of the sensor decreases due to the weaker interaction of the electric field with the solution. It was found that the optimal compromise between sensitivity and measurement stability is achieved with a ZrO₂ layer of approximately 4 µm thickness. In this case, the sensor is characterized by a close linear calibration characteristic, good repeatability and sufficient sensitivity for practical application. By increasing the layer thickness to ~6 µm, the measurements become even more stable, but the sensitivity decreases. The obtained results show that capacitive sensors modified with a ZrO₂ coating are a promising technology for real-time monitoring of liquid media, especially in the fields of wastewater analysis and industrial process control. The ability to control the sensor characteristics by changing the coating thickness provides a flexible and effective basis for the development of reliable sensor systems.