Ternary Holey Carbon Nanohorn/Potassium Chloride/Polyvinylpyrrolidone Nanohybrid as Sensing Film for Resistive Humidity Sensor

: The study presents findings on the relative humidity (R


Introduction
Humidity has become an essential environmental parameter in many residential, industrial, and commercial applications.Monitoring and controlling the humidity in offices, homes, and medical settings using gas supply infrastructure, respiratory care systems, incubators, infusion pumps, ventilators, and sterilizers benefit human comfort.They measure humidity in the pharmaceutical industry (i.e., for packaging and storage), in the quality control of food and beverage production, and in the cosmetics industry.The control of humidity is essential in many industries, including electronics manufacturing (e.g., semiconductor production facilities), chemical processing (e.g., dehumidifiers, dryers), metallurgy, textiles, and papermaking.It is also important in agriculture (e.g., soil moisture detection), climatology, the automotive industry (e.g., engine test beds), and many others [1][2][3][4][5].
As a result, the humidity sensor market has experienced a significant growth over the past few decades and is projected to exceed USD 8.7 billion by 2029 [6,7].Various types of humidity sensors are currently utilized across different applications.Among these, resistive R.H. sensors offer several advantages, including cost-effectiveness, simple design, a broad operating range, rapid response times, and low power consumption.However, they also have limitations such as lower accuracy and stability, non-linear response, and relatively shorter lifespan.
In addition to the sensing principle (electrochemical, optical, gravimetric, capacitive, piezoresistive, or resistive), fabrication techniques, and sensor design, the selection of materials for the sensing layer is crucial for developing humidity sensors with superior performance [8].To date, numerous materials have been investigated for use as sensing layers in humidity sensor designs, as outlined in Table 1.
Table 1.Classification of sensing humidity materials based on sensitive response.

Type of Humidity Sensor
Sensitive Response Reference capacitive humidity sensors the dielectric constant of the material changes, leading to a change in capacitance polymers, carbon-based materials, metal oxide-based materials, composites, mesoporous silica, macroporous silica [9][10][11] resistive humidity sensors changes in the electrical resistance of the sensing material polymers, carbon-based materials, metal oxide-based materials, composites, mesoporous silica [9,10,12] thermal conductivity humidity sensors thermal conductivity of air increases, leading to a change in temperature or heat flow within the sensor polymers, ceramics, nanocomposites, carbon-based materials [13,14] gravimetric humidity sensors changes in mass or weight of a humidity-sensitive material hygroscopic polymers or salts, porous materials, quartz crystal microbalance sensors [15] optical humidity sensors changes in optical properties, such as refractive index or absorbance fiber optics coated with humidity-sensitive materials (photonic crystals, hygroscopic coatings, hydrogels, nanoparticles) [16] surface acoustic wave (SAW) humidity sensors the shift in frequency or phase of the acoustic waves on the surface of a piezoelectric substrate polymers, carbon-based materials, metal oxide-based materials, composites [17] electrolytic humidity sensors the change of conductivity or impedance of the solution hydrogel polymers, solid electrolytes, conductive polymer composites [18][19][20][21] Furthermore, numerous carbon-based materials find widespread application as sensing layers in the construction of R.H. sensors.Their large specific surface area, capacity to function at room temperature, robust mechanical properties, notable chemical inertness, Coatings 2024, 14, 517 3 of 17 and ease of hydrophilization through covalent functionalization render these materials particularly appealing as sensitive coatings for resistive humidity monitoring.Carbon nanofibers [22], carbon nanosheets [23], mesoporous carbon [24], amorphous carbon [25], carbon dots [26], N-doped hydrogenated amorphous carbon [27], nanodiamond [28], carbon nanocoil [29], carbon nanotubes and their nanocomposites [30,31], graphene oxide [32,33], reduced graphene oxide [34], fullerenol [35], and fullerenes [36] are some of the carbon nanomaterials studied for R.H. sensing applications.
Several types of devices are available for R.H. monitoring, and these operate on a wide range of principles, such as capacitive [49], optical [50], surface acoustic wave [51], resistive [52], and electrochemical [53].Resistive sensors have garnered heightened interest in recent decades owing to their straightforward design, cost-effectiveness, heightened sensitivity, and minimal energy usage.In this type of device, relative humidity (R.H.) measurement relies on the alteration in resistance within a humidity-sensitive layer upon interaction with water molecules in the gas phase.The introduction of water vapor triggers modifications in the resistance of the thin film through processes such as adsorption, absorption, swelling, and chemical reactions like dissociation and diffusion.These phenomena take place either on the surface or within the bulk of the thin film, serving as the sensing element [54,55].
This paper discusses the relative humidity (R.H.) sensing capabilities of a resistive sensor utilizing a sensing layer composed of a ternary nanohybrid containing CN-Hox/KCl/PVP at mass ratios of 7/1/2, 6.5/1.5/2, and 6/2/2 (w/w/w).The study highlights the R.H. sensing performance of the newly developed CNHox-based nanohybrid operating at room temperature (R.T.).Incorporating oxidized nanohorns with PVP and KCl aims to enhance sensitivity and selectivity in humidity sensing, addressing several research gaps and areas needing improvement in the field, such as sensitivity, response and recovery times, and sensor stability.The proposed mixture as a sensing substrate tackles these challenges by leveraging the unique properties of nanohorns, KCl, and PVP.Oxidized nanohorns are expected to boost the adsorption of water molecules, potentially broadening the sensors' sensitivity and dynamic range.Conversely, the combination of oxidized nanohorns with PVP and KCl may create a more accessible surface area for humidity interaction, thereby potentially accelerating the sensor's response to humidity changes.Additionally, the chemical stability of PVP, combined with the robust structural properties of nanohorns, could contribute to devices capable of maintaining performance over time and under varying environmental conditions.

Materials
All the materials utilized in the fabrication of the novel R.H. sensing layers were purchased from Sigma Aldrich (Redox Lab Supplies Com, Bucharest, Romania).The holey CNHox, depicted in Figure 1a, exhibits lengths ranging from 40 nm to 50 nm, diameters between 2 nm and 5 nm, and a specific surface area of approximately 1300-1400 m 2 /g.As per the supplier, the CNHox employed in experiments is devoid of metal contamination, with graphite constituting the primary impurity at approximately 10% (w/w).The structure of the employed PVP is depicted in Figure 1b and has an average molar weight of 40,000 g mol −1 .The KCl used in the experiments has 99.99% purity (with impurities ≤15.0 ppm in the trace metal analysis).Isopropyl alcohol ((CH 3 ) 2 CHOH) is a 70% w/w solution in water.All reagents were used as-received without further purification.

Preparation of the Quaternary Organic-Inorganic Holey CNH-Based Hybrid Sensing Films and Experimental Setup
To explore the relative humidity (R.H.) sensing performance of the ternary nanohybrid, sensing films with the following chemical compositions were formulated, prepared, and evaluated: CNHox/KCl/PVP ratios of 7/1/2, 6.5/1.5/2, and 6/2/2, all expressed as mass ratios (w/w/w).The mixture was homogenized in isopropyl alcohol using gentle sonication (FS20D Fisher Scientific, Dreieich, Germany), working at 42 kHz (output power of 70 W).This process provides a relatively uniform dispersion of the CNHox and KCl in the PVP network.
The preparation of sensing films based on CNHox, KCl, and PVP with varying weight ratios followed a five-step procedure.Initially, a PVP solution was prepared by dissolving 2 mg of the hydrophilic polymer in 10 mL of isopropyl alcohol and stirring it in an ultrasonic bath for 10 min.Subsequently, holey CNHox (7 mg, 6.5 mg, and 6 mg, respectively) was dispersed in the prepared PVP solution and stirred in the ultrasonic bath for 6 h at room temperature (R.T.).Following this, KCl powder was added to the resulting suspension to obtain the desired mass ratios (1 mg, 1.5 mg, and 2 mg of KCl), and continuous stirring was maintained in the ultrasound bath for another 6 h, also at R.T. Finally, the sensing layer was obtained using the drop-casting method, depositing the mixture onto the support structure.The electrical contact areas were masked before applying the sensing dispersion.All samples were then dried at 373 K for 60 min before conducting electrical measurements.
The sensing support structure consists of a metallic interdigitated electrode (IDE) with a dual-comb structure fabricated on a Si substrate (470 µm thickness), covered by a SiO2 layer (1 µm thickness) (Figure 2).The IDE's metal stripes comprise chromium (10 nm thickness) and gold (100 nm thickness) [39][40][41]56].The relative humidity (R.H.) monitoring experiments were conducted using a suitable experimental setup (see Figure 3).Dry nitrogen was introduced through a series of bubblers containing deionized water (conductivity below 1 µS/cm) to achieve controlled variations in R.H. within the testing chamber, ranging from 0% to 100% R.H.The testing chamber accommodated two sensors: a commercial capacitive R.H. sensor (referred to as

Preparation of the Quaternary Organic-Inorganic Holey CNH-Based Hybrid Sensing Films and Experimental Setup
To explore the relative humidity (R.H.) sensing performance of the ternary nanohybrid, sensing films with the following chemical compositions were formulated, prepared, and evaluated: CNHox/KCl/PVP ratios of 7/1/2, 6.5/1.5/2, and 6/2/2, all expressed as mass ratios (w/w/w).The mixture was homogenized in isopropyl alcohol using gentle sonication (FS20D Fisher Scientific, Dreieich, Germany), working at 42 kHz (output power of 70 W).This process provides a relatively uniform dispersion of the CNHox and KCl in the PVP network.
The preparation of sensing films based on CNHox, KCl, and PVP with varying weight ratios followed a five-step procedure.Initially, a PVP solution was prepared by dissolving 2 mg of the hydrophilic polymer in 10 mL of isopropyl alcohol and stirring it in an ultrasonic bath for 10 min.Subsequently, holey CNHox (7 mg, 6.5 mg, and 6 mg, respectively) was dispersed in the prepared PVP solution and stirred in the ultrasonic bath for 6 h at room temperature (R.T.).Following this, KCl powder was added to the resulting suspension to obtain the desired mass ratios (1 mg, 1.5 mg, and 2 mg of KCl), and continuous stirring was maintained in the ultrasound bath for another 6 h, also at R.T. Finally, the sensing layer was obtained using the drop-casting method, depositing the mixture onto the support structure.The electrical contact areas were masked before applying the sensing dispersion.All samples were then dried at 373 K for 60 min before conducting electrical measurements.

Preparation of the Quaternary Organic-Inorganic Holey CNH-Based Hybrid Sensing Films and Experimental Setup
To explore the relative humidity (R.H.) sensing performance of the ternary nanohybrid, sensing films with the following chemical compositions were formulated, prepared, and evaluated: CNHox/KCl/PVP ratios of 7/1/2, 6.5/1.5/2, and 6/2/2, all expressed as mass ratios (w/w/w).The mixture was homogenized in isopropyl alcohol using gentle sonication (FS20D Fisher Scientific, Dreieich, Germany), working at 42 kHz (output power of 70 W).This process provides a relatively uniform dispersion of the CNHox and KCl in the PVP network.
The preparation of sensing films based on CNHox, KCl, and PVP with varying weight ratios followed a five-step procedure.Initially, a PVP solution was prepared by dissolving 2 mg of the hydrophilic polymer in 10 mL of isopropyl alcohol and stirring it in an ultrasonic bath for 10 min.Subsequently, holey CNHox (7 mg, 6.5 mg, and 6 mg, respectively) was dispersed in the prepared PVP solution and stirred in the ultrasonic bath for 6 h at room temperature (R.T.).Following this, KCl powder was added to the resulting suspension to obtain the desired mass ratios (1 mg, 1.5 mg, and 2 mg of KCl), and continuous stirring was maintained in the ultrasound bath for another 6 h, also at R.T. Finally, the sensing layer was obtained using the drop-casting method, depositing the mixture onto the support structure.The electrical contact areas were masked before applying the sensing dispersion.All samples were then dried at 373 K for 60 min before conducting electrical measurements.
The sensing support structure consists of a metallic interdigitated electrode (IDE) with a dual-comb structure fabricated on a Si substrate (470 µm thickness), covered by a SiO2 layer (1 µm thickness) (Figure 2).The IDE's metal stripes comprise chromium (10 nm thickness) and gold (100 nm thickness) [39][40][41]56].The relative humidity (R.H.) monitoring experiments were conducted using a suitable experimental setup (see Figure 3).Dry nitrogen was introduced through a series of bubblers containing deionized water (conductivity below 1 µS/cm) to achieve controlled variations in R.H. within the testing chamber, ranging from 0% to 100% R.H.The testing chamber accommodated two sensors: a commercial capacitive R.H. sensor (referred to as The relative humidity (R.H.) monitoring experiments were conducted using a suitable experimental setup (see Figure 3).Dry nitrogen was introduced through a series of bubblers containing deionized water (conductivity below 1 µS/cm) to achieve controlled variations mixtures of CNHox/KCl/PVP at various w/w/w ratios.
To figure out the function of individual components in the mixture, KCl added to the mix was chosen as the component for the study (concentrations of 10, 15, and 20%, while maintaining the PVP concentration constant).
The C.O.M. reference sensor confirmed the relative humidity (R.H.) level generated by controlling the mass flow controller and facilitated an objective comparison of the data measured by the S.U.I.Both sensors were positioned closely together and near the gas inlet to ensure exposure to identical gas flow and nearly identical experimental conditions.A Keithley 6620 current source (Keithley Instruments GmbH, Germering, Germany) supplied a direct current ranging from 0.01 to 0.1 A. Data acquisition and analysis were conducted using a PicoLog data logger (PICO Technology, Neots, Cambridgeshire, United Kingdom).All measurements were conducted at room temperature (R.T.).The relative humidity (R.H.) sensing capabilities of each ternary nanohybrid-based sensing layer were investigated by applying a current across the two electrodes and recording the voltage difference while varying the R.H. from 0% to 100%.For simplicity in analysis, the following abbreviations are utilized: Sensor K1: CNHox/KCl/PVP at a mass ratio of 7/1/2 (w/w/w).
Throughout the sensor realization process, particular a ention was given to the potential variability induced by the processing steps.The IDT structures were achieved using a photolithographic process with practically zero-dimensional variations.The preparation of the sensing material included a homogenization step before the material was drop-cast on the IDTs.The process was implemented using between 25 and 30 different structures for each composition.After drying, the electrical resistance of each IDT structure was measured, and an average value (Rav) was determined; samples whose electrical resistance was outside the Rav ± 5% were excluded from the following steps (i.e., recording the electrical resistance vs. R.H. variations).For each group of samples (corresponding to different material compositions), the remaining samples were between 12 and 15.The results presented in the manuscript are typical results reflecting the behavior of the samples in each group.
The Raman spectra were acquired at room temperature (R.T.) using a Witec Raman spectrometer (Alpha-SNOM 300 S, WiTec.GmbH, Ulm, Germany) with 532 nm excitation.A 532 nm diode-pumped solid-state laser with 145 mW of power was employed.The incident laser beam, with a spot size of approximately 1.0 µm, was focused onto the sample using a 6 mm working distance objective a ached to a Thorlabs MY100X-806 microscope.Raman spectra were recorded with an exposure time of 20 s accumulation, and the scattered light was collected by the same objective in back-sca ering geometry with The C.O.M. reference sensor confirmed the relative humidity (R.H.) level generated by controlling the mass flow controller and facilitated an objective comparison of the data measured by the S.U.I.Both sensors were positioned closely together and near the gas inlet to ensure exposure to identical gas flow and nearly identical experimental conditions.A Keithley 6620 current source (Keithley Instruments GmbH, Germering, Germany) supplied a direct current ranging from 0.01 to 0.1 A. Data acquisition and analysis were conducted using a PicoLog data logger (PICO Technology, Neots, Cambridgeshire, United Kingdom).All measurements were conducted at room temperature (R.T.).
The relative humidity (R.H.) sensing capabilities of each ternary nanohybrid-based sensing layer were investigated by applying a current across the two electrodes and recording the voltage difference while varying the R.H. from 0% to 100%.For simplicity in analysis, the following abbreviations are utilized: Sensor K1: CNHox/KCl/PVP at a mass ratio of 7/1/2 (w/w/w).Sensor K2: CNHox/KCl/PVP at a mass ratio of 6.5/1.5/2(w/w/w).Sensor K3: CNHox/KCl/PVP at a mass ratio of 6/2/2 (w/w/w).Throughout the sensor realization process, particular attention was given to the potential variability induced by the processing steps.The IDT structures were achieved using a photolithographic process with practically zero-dimensional variations.The preparation of the sensing material included a homogenization step before the material was drop-cast on the IDTs.The process was implemented using between 25 and 30 different structures for each composition.After drying, the electrical resistance of each IDT structure was measured, and an average value (Rav) was determined; samples whose electrical resistance was outside the Rav ± 5% were excluded from the following steps (i.e., recording the electrical resistance vs. R.H. variations).For each group of samples (corresponding to different material compositions), the remaining samples were between 12 and 15.The results presented in the manuscript are typical results reflecting the behavior of the samples in each group.
The Raman spectra were acquired at room temperature (R.T.) using a Witec Raman spectrometer (Alpha-SNOM 300 S, WiTec.GmbH, Ulm, Germany) with 532 nm excitation.A 532 nm diode-pumped solid-state laser with 145 mW of power was employed.The incident laser beam, with a spot size of approximately 1.0 µm, was focused onto the sample using a 6 mm working distance objective attached to a Thorlabs MY100X-806 microscope.Raman spectra were recorded with an exposure time of 20 s accumulation, and the scattered light was collected by the same objective in back-scattering geometry with 600 grooves/mm grating.The Raman systems underwent calibration using the 520 cm −1 Raman line of a silicon wafer.Data acquisition and processing were carried out by a dedicated computer using WiTec Project Five software (Version: 5.1).
Scanning electron microscopy (SEM) investigated the sensing films' surface topography.A field emission gun scanning electron microscope, FEG-SEM-Nova NanoSEM 630 (Thermo Scientific, Waltham, MA, USA) (F.E.I.), was used for surface visualization and had superior low-voltage resolution and high surface sensitivity imaging.The samples were investigated directly (i.e., no sample preparation was needed).The current during the measurements was 1 nA.

Surface Topography
Scanning electron micrographs reveal that the surface morphology of the coating mixture appears relatively uniform across all cases (Figure 4a-f).Additionally, particles of significantly larger dimensions (>100 nm) are observable, attributed to the aggregation of particles with varied sizes and crystallographic orientations.
Coatings 2024, 14, x FOR PEER REVIEW 6 of 18 600 grooves/mm grating.The Raman systems underwent calibration using the 520 cm −1 Raman line of a silicon wafer.Data acquisition and processing were carried out by a dedicated computer using WiTec Project Five software (Version: 5.1 ).Scanning electron microscopy (SEM) investigated the sensing films' surface topography.A field emission gun scanning electron microscope, FEG-SEM-Nova NanoSEM 630 (Thermo Scientific, Waltham, MA, USA) (F.E.I.), was used for surface visualization and had superior low-voltage resolution and high surface sensitivity imaging.The samples were investigated directly (i.e., no sample preparation was needed).The current during the measurements was 1 nA.

Surface Topography
Scanning electron micrographs reveal that the surface morphology of the coating mixture appears relatively uniform across all cases (Figure 4a-f).Additionally, particles of significantly larger dimensions (>100 nm) are observable, a ributed to the aggregation of particles with varied sizes and crystallographic orientations.

Raman Spectroscopy
The Raman spectra of the ternary nanohybrid CNHox/KCl/PVP are presented in Figure 5a,c.In Figure 5d, spectra p1-p3 are three acquisition points per sample (sample K3).In the p2 (gray) and p3 (black) spectra, between 0-1000 cm −1 , there are the vibration modes of Si combined with PVP; above 2700 cm −1 , there is a vibration mode specific to PVP (the Raman spectra of pure PVP is represented in blue).The Raman spectrum of powder PVP, for wavelengths below 1000 cm −1 , includes vibration lines associated with groups N-C=O (560 cm −1 ) and C-C (758 cm −1 , 851 cm −1 , and 934 cm −1 ) Three active Raman bands (D, G, 2D) were detected at wavenumbers of 1332, 1587, and 2660 cm −1 , respectively, which are characteristic of sp 2 -defected carbon nanomaterials.

Raman Spectroscopy
The Raman spectra of the ternary nanohybrid CNHox/KCl/PVP are presented in Figure 5a,c.In Figure 5d, spectra p1-p3 are three acquisition points per sample (sample K3).In the p2 (gray) and p3 (black) spectra, between 0-1000 cm −1 , there are the vibration modes of Si combined with PVP; above 2700 cm −1 , there is a vibration mode specific to PVP (the Raman spectra of pure PVP is represented in blue).The Raman spectrum of powder PVP, for wavelengths below 1000 cm −1 , includes vibration lines associated with groups N-C=O (560 cm −1 ) and C-C (758 cm −1 , 851 cm −1 , and 934 cm −1 ) Three active Raman bands (D, G, 2D) were detected at wavenumbers of 1332, 1587, and 2660 cm −1 , respectively, which are characteristic of sp 2 -defected carbon nanomaterials.Additionally, the Raman spectrum of pure PVP reveals a weak band at 2990 cm −1 , which can be a ributed to PVP [57].

R.H. Monitoring Capability of the Ternary Nanocomposite
The present R.T. resistive R.H. sensor has an organic-inorganic sensitive layer containing oxidated carbon nanohorn (CNHox) with a concentration higher than the percolation threshold of the solid-state nanocomposite [62].This chemical design of the nanocarbon component assured a reasonably low value of the electrical conductivity of the sensing layer and simple, functional device measurability.Three active Raman bands (D, G, 2D) were detected at wavenumbers of 1332, 1587, and 2660 cm −1 , respectively, which are characteristic of sp 2 -defected carbon nanomaterials.Additionally, the Raman spectrum of pure PVP reveals a weak band at 2990 cm −1 , which can be attributed to PVP [57].

R.H. Monitoring Capability of the Ternary Nanocomposite
The present R.T. resistive R.H. sensor has an organic-inorganic sensitive layer containing oxidated carbon nanohorn (CNHox) with a concentration higher than the percolation threshold of the solid-state nanocomposite (Serban, B.-C., et al., 2021).This chemical design of the nanocarbon component assured a reasonably low value of the electrical conductivity of the sensing layer and simple, functional device measurability.
Figure 6 illustrates the relative humidity (R.H.) response of the fabricated sensors.It is evident that the resistance of the thin film based on the ternary nanohybrid increases with the rising of R.H. lation threshold of the solid-state nanocomposite [62].This chemical design of the nanocarbon component assured a reasonably low value of the electrical conductivity of the sensing layer and simple, functional device measurability.At the same time, the overall linearity of the ternary nanohybrid-based resistive sensors-in humid nitrogen, when varying R.H. from 0% to 100%-is excellent, as shown in Figure 7.In terms of linearity, all the manufactured sensors show superior performances (R 2 > 0.99) for operating cycles 2-6 and an excellent baseline drift, but for R.H. values up to 80%.It is interesting to note a different behavior in the first operating cycle compared to other operating cycles.At R.H. higher than 80%, a sharp increase in resistance with R.H. can be observed.
For all studied sensing layers, there is a decrease in thin-film resistance following the first operating cycle.This decrease is more pronounced in sensors with higher KCl concentrations in the sensing layer.The results of this experiment reinforce the hypothesis that ionic conduction is the primary mechanism during the first operating cycle.Thus, adding potassium salt can provide more active sites for humidity, increasing the local concentration of water molecules in the sensing layer.Furthermore, the results presented in Figure 4 show a baseline drift for all tested sensing devices.
This study provides strong evidence that the presence of KCl significantly impacts the device's response during the first cycle.This influence is likely due to potassium ions bonding with water molecules, thereby changing material properties, particularly the electrical conductivity, which is further reflected in the electrical resistance change.This initial absorption phenomenon offers a potential explanation for the observed baseline drift in the first cycle.
Our experimental findings suggest that the sensor's initial exposure to moisture acts At the same time, the overall linearity of the ternary nanohybrid-based resistive sensors-in humid nitrogen, when varying R.H. from 0% to 100%-is excellent, as shown in Figure 7.In terms of linearity, all the manufactured sensors show superior performances (R 2 > 0.99) for operating cycles 2-6 and an excellent baseline drift, but for R.H. values up to 80%.It is interesting to note a different behavior in the first operating cycle compared to other operating cycles.At R.H. higher than 80%, a sharp increase in resistance with R.H. can be observed.
For all studied sensing layers, there is a decrease in thin-film resistance following the first operating cycle.This decrease is more pronounced in sensors with higher KCl concentrations in the sensing layer.The results of this experiment reinforce the hypothesis that ionic conduction is the primary mechanism during the first operating cycle.Thus, adding potassium salt can provide more active sites for humidity, increasing the local concentration of water molecules in the sensing layer.Furthermore, the results presented in Figure 4 show a baseline drift for all tested sensing devices.The calculation of sensitivity was performed using the following equation: where Rx is the resistance of the sensitive layer measured in the test chamber for the [x-%] R.H. value indicated by the commercial sensor (measured with ±2% accuracy as the producer indicates it).R0 is the value of resistance calculated from the linear function established from the graph of resistance and is equal to f (relative humidity) obtained by This study provides strong evidence that the presence of KCl significantly impacts the device's response during the first cycle.This influence is likely due to potassium ions bonding with water molecules, thereby changing material properties, particularly the electrical conductivity, which is further reflected in the electrical resistance change.This initial absorption phenomenon offers a potential explanation for the observed baseline drift in the first cycle.
Our experimental findings suggest that the sensor's initial exposure to moisture acts as a conditioning phase.During this phase, the sensor's sensing material undergoes physical or chemical changes that stabilize its response in subsequent humidity cycles.These changes might include the swelling of the polymer matrix, the redistribution of potassium chloride (KCl) and nanohorns within the composite material, or the alterations in the material's microstructure that affect its response to humidity.
After the initial cycle, the material might reach a state of equilibrium faster in subsequent cycles, showing less deviation in resistance measurements.This stabilization could suggest that the initial drift is a part of a necessary "settling" process for the sensor materials.
The drop in resistance at higher relative humidity levels is related to the KCl content in the mixture.The evolution of resistance presented in Figure 6 shows that the sensing device of the K1 device is "settled" after a few functioning cycles.When the amount of KCl is increased, the device resistance increases upon exposure to a higher level of R.H. because more water molecules penetrate the bulk of the sensing layer.Therefore, KCl and water make a homogenous mixture in the sensing layer at higher water uptake.This could change the sensing structure response and possibly cause non-linear behavior at higher humidity levels, especially for the sensing coatings with increased KCl content.
Figure 7e,f present the transfer function for the sensor K3.Noticeably, after the first cycle, the resistance at R.H. = 0 reduces from 710-715 Ω to 630-635 Ω.The same behavior is also registered for the sensor K1 (Figure 7a,b), with a decrease in resistance (for R.H. = 0) from 970 (first cycle) to 910 Ω and 930 Ω (subsequent cycles).
The calculation of sensitivity was performed using the following equation: where Rx is the resistance of the sensitive layer measured in the test chamber for the [x-%] R.H. value indicated by the commercial sensor (measured with ±2% accuracy as the producer indicates it).R 0 is the value of resistance calculated from the linear function established from the graph of resistance and is equal to f (relative humidity) obtained by extrapolation for the value at 0% R.H.The calculated relative sensitivity values for each humidity jump are presented in Figure 8a-c.A simple comparison between the calculated sensitivities for K1, K2, and K3 sensors and other similarly manufactured and tested sensors that employed holey CNHOx-based sensing layers reveals interesting results (Table 2).As one can see, K1, K2, and K3 sensors exhibit a superior performance in terms of sensitivity compared with other reported sensors.The high hydrophilicity of the alkali salt has a significant contribution, increasing the number of active sites towards water molecules.In addition, the enhanced sensing layer porosity and its associated high specific surface area can also contribute to this result.A simple comparison between the calculated sensitivities for K1, K2, and K3 sensors and other similarly manufactured and tested sensors that employed holey CNHOx-based sensing layers reveals interesting results (Table 2).As one can see, K1, K2, and K3 sensors exhibit a superior performance in terms of sensitivity compared with other reported sensors.The high hydrophilicity of the alkali salt has a significant contribution, increasing the number of active sites towards water molecules.In addition, the enhanced sensing layer porosity and its associated high specific surface area can also contribute to this result.CNHox/PVP 1/1 0.020-0.058[38] CNHox/PVP 1/2 0.017-0.025[38] CNHox/GO/PVP 3/1/1 0,043-0, 051 [38] CNHox/GO/SnO An important parameter, such as response time (t r ), was calculated for all manufactured R.H. resistive sensors.If R(t) is the response of the device in time, t r can be calculated as follows: where t 90 and t 10 represent the moments when the response R(t) reaches 90% and 10%, respectively, from the total variation of the sensor's resistance due to a change in the R.H. value (as in the example presented in Figure 9).An important parameter, such as response time (tr), was calculated for all manufactured R.H. resistive sensors.If R(t) is the response of the device in time, tr can be calculated as follows: where t90 and t10 represent the moments when the response R(t) reaches 90% and 10%, respectively, from the total variation of the sensor's resistance due to a change in the R.H. value (as in the example presented in Figure 9).The reference sensor has a response time of 60 s +/− 10 s for R.H. below 70%.The response time for R.H. > 70% increases to approximately 90 s +/−10 s.Water molecules permeate the hydrophilic nanohybrid sensing film.The hybrid nanocomposite film adsorbs/absorbs a substantial amount of water (related to the sensing film's mass).Finally, some water molecules can condense in the proximity of hydrophilic groups, thus blocking the active sites.

Analysis of the Sensing Mechanism
All constituents of the ternary nanohybrid used as a sensing layer in the resistive monitoring of R.H. have some outstanding chemical and physical properties.CNHox shows increased conductivity (p-type semiconductor behavior), a remarkable and controllable surface area (1300 to 1400 m 2 /g), high porosity, hydrophilicity, and a rapid variation in the electrical resistance in contact with a water molecule in the 0% R.H. to 100% R.H. range.Furthermore, low-cost synthesis of CNHox via oxygen plasma treatment of the pristine carbon nanohorns (CNHs) is a simple, clean, and fast procedure.Finally, the functionalization of the CNHs in oxygen plasma has the advantage (by varying the exposure time and its power) that it can provide an optimal C:O ratio for superior sensitivity towards water molecules [58].The reference sensor has a response time of 60 s +/− 10 s for R.H. below 70%.The response time for R.H. > 70% increases to approximately 90 s +/−10 s.Water molecules permeate the hydrophilic nanohybrid sensing film.The hybrid nanocomposite film adsorbs/absorbs a substantial amount of water (related to the sensing film's mass).Finally, some water molecules can condense in the proximity of hydrophilic groups, thus blocking the active sites.

Analysis of the Sensing Mechanism
All constituents of the ternary nanohybrid used as a sensing layer in the resistive monitoring of R.H. have some outstanding chemical and physical properties.CNHox shows increased conductivity (p-type semiconductor behavior), a remarkable and controllable surface area (1300 to 1400 m 2 /g), high porosity, hydrophilicity, and a rapid variation in the electrical resistance in contact with a water molecule in the 0% R.H. to 100% R.H. range.Furthermore, low-cost synthesis of CNHox via oxygen plasma treatment of the pristine carbon nanohorns (CNHs) is a simple, clean, and fast procedure.Finally, the functionalization of the CNHs in oxygen plasma has the advantage (by varying the exposure time and its power) that it can provide an optimal C:O ratio for superior sensitivity towards water molecules [58].
At the same time, PVP is a hydrophilic polymer with excellent binding and dispersion properties.The addition of alkali ions, such as K + , ensures more active sites for water molecules.
The humidity-sensing film operates through three distinct mechanisms, considering the properties of its constituents.First, we analyze the p-type semiconducting behavior of the CNHox material.When exposed to water molecules, the CNHox accepts electron pairs, reducing the number of positive charge carriers (holes) within the nanocarbon material.
Consequently, the humidity-sensing film becomes more resistive.Several reviews reported in the literature confirm this interpretation [59].
The second sensing mechanism considers the dissociation of water.Due to the strong electrostatic field and high local charge density, the adsorbed water molecules on the CNHox hydrophilic surface and mesopores may dissociate into H + and OH − ions.The protons generated by the water dissociation process may tunnel from one water molecule to another through hydrogen bonding, thus increasing the overall electrical conductivity of the sensitive film [60].
PVP is a dielectric polymer with hydrophilic properties, which swells after interaction with water molecules and has a negligible change in resistance in response to R.H. variation, as shown in literature for the presented composites [61][62][63].Based on these properties, a third sensing mechanism can be discussed.The swelling of PVP leads to the displacement of CNHox, increases the distance between nanocarbon particles, and lowers the number of electrically percolating pathways.Accordingly, the device resistance increases upon exposure to a higher level of R.H. because more water molecules penetrate the bulk of the sensing layer.
Finally, the contribution of KCl to the sensing mechanism can be interpreted from two perspectives.Firstly, according to the Hard-Soft Acid-Base (HSAB) theory, the K+ ion is classified as a hard acid and can electrostatically interact with water molecules, which are classified as hard bases (electron donors) [64,65].Thus, adding potassium salt can provide more active sites for humidity, increasing the local concentration of water molecules in the sensing layer.Due to the p-type semiconducting behavior of CNHox and the PVP swelling, a higher number of water molecules leads to a significant increase in resistance.At the same time, once the local concentration of water molecules increases, KCl is dissolved, the ions can move freely, and the conductivity of the sensing layer increases.
In conclusion, as described above, the interaction of the sensing layer with water molecules has, theoretically, antagonistic effects on its resistance variation.Since the measured resistance of the sensing film increases when R.H. increases from 0% to 100%, one can assume that the cumulative effect of the p-type semiconductor behavior of the CNHox and the swelling of PVP are the dominant causes that lead to this effect.Therefore, without completely excluding the impact of both the proton-tunneling mechanism and ionic conduction through K+ and Cl-ions, the interaction of the CNHox-PVP tandem with water has a pivotal contribution to the variation in the resistance of the sensing layer with R.H.For all studied sensing layers, the thin film's resistance decreases after the sensor's first operating cycle.The higher the percentage of KCl in the sensitive layer, the more significant the decrease in resistance.This experimental result agrees with the idea that the ionic conduction mechanism is predominant in the first operating cycle.

Conclusions
The proposed study investigated the relative humidity (R.H.) sensing response of an R.H. detection structure employing a ternary nanohybrid as a sensing layer.The nanohybrid consisted of holey CNHox (with concentration levels surpassing the percolation threshold), KCl, and PVP, with mass ratios of 7/1/2, 6.5/1.5/2, and 6/2/2 (w/w/w).The sensing structure comprised a silicon substrate, a SiO 2 layer, and an interdigitated electrode (IDE).The sensing film was deposited onto the sensing structure using the drop-casting method.The morphology and composition of the sensing layers were investigated using scanning electron microscopy (SEM) and Raman spectroscopy.
The manufactured sensors exhibited a favorable room temperature response comparable to that of a commercial capacitive R.H. sensor.They demonstrated excellent linearity, rapid response times, and good sensitivity.Moreover, they showed superior sensitivity performance compared to other sensors employing holey CNHox-based sensing layers.Regarding linearity, all manufactured sensors demonstrated superior performance (R2 > 0.99) for operating cycles 2-6 and exhibited minimal baseline drift up to an R.H. of 80%.However, at R.H. levels exceeding 80%, increased water molecule permeation into the hydrophilic nanohybrid sensing film was observed.The hybrid nanocomposite film absorbed a significant amount of water, proportional to the sensing film's mass.Eventually, some water molecules condensed near the hydrophilic groups, potentially obstructing active sites.Although the K3 sensor shows a more significant variation in resistance when humidity varies between 0 and 100% R.H. (about 170 ohms between the minimum and maximum resistance values measured during the experiment), its response is not linear when humidity exceeds 60% R.H.The same phenomenon was recorded for the K2 sensing device.Therefore, the K1 sensor has the closest response to the reference sensor, presenting at the same time a resistance variation of approximately 130 ohms, which could allow its use for devices with high sensitivity for relative humidity measurement.
The role of each component in the ternary nanohybrid used as a sensing film was thoroughly explained, considering their electrical, chemical, and physical properties.Three distinct sensing mechanisms were explored and discussed.Notably, the p-type semiconductor behavior of CNHox, combined with the swelling of the hydrophilic polymer (PVP), was found to dominate.As a result, the overall resistance of the sensing films increased with relative humidity (R.H.) across all manufactured sensors.
The significant advantages of the presented sensors include their low power consumption, which is below 2 mW, their reliable sensing performance at room temperature, and their straightforward manufacturing process.The evaluated sensing coatings demonstrate exceptional performance for atmospheric sensors, facilitating tasks such as environmental monitoring, indoor air quality assessment, and industrial processes.These coatings deliver precise measurement accuracy within relative humidity (R.H.) ranges up to 80%.Interestingly, the decrease in resistance beyond 80% humidity can be utilized as a valuable feature rather than a drawback.This characteristic enables the triggering of alarm signals or the activation of control systems in critical applications.Such applications include electronic equipment rooms, storage facilities, sensitive manufacturing areas, chemical processing plants, laboratories handling moisture-sensitive materials, and battery energy storage facilities, where maintaining humidity below a specific threshold is crucial.

Figure 2 .
Figure 2. The layout of the IDE sensing structure.

Figure 2 .
Figure 2. The layout of the IDE sensing structure.

Figure 2 .
Figure 2. The layout of the IDE sensing structure.
H. within the testing chamber, ranging from 0% to 100% R.H.The testing chamber accommodated two sensors: a commercial capacitive R.H. sensor (referred to as C.O.M.) as the reference and the new resistive sensing structure (abbreviated as S.U.I.-"sensor under investigation").The sensing layers of the S.U.I. comprised nanocomposite mixtures of CNHox/KCl/PVP at various w/w/w ratios.

Figure 3 .
Figure 3. Experimental setup employed for R.H. measurements.To figure out the function of individual components in the mixture, KCl added to the mix was chosen as the component for the study (concentrations of 10, 15, and 20%, while maintaining the PVP concentration constant).The C.O.M. reference sensor confirmed the relative humidity (R.H.) level generated by controlling the mass flow controller and facilitated an objective comparison of the data measured by the S.U.I.Both sensors were positioned closely together and near the gas inlet to ensure exposure to identical gas flow and nearly identical experimental conditions.A Keithley 6620 current source (Keithley Instruments GmbH, Germering, Germany) supplied a direct current ranging from 0.01 to 0.1 A. Data acquisition and analysis were conducted using a PicoLog data logger (PICO Technology, Neots, Cambridgeshire, United Kingdom).All measurements were conducted at room temperature (R.T.).The relative humidity (R.H.) sensing capabilities of each ternary nanohybrid-based sensing layer were investigated by applying a current across the two electrodes and recording the voltage difference while varying the R.H. from 0% to 100%.For simplicity in analysis, the following abbreviations are utilized:Sensor K1: CNHox/KCl/PVP at a mass ratio of 7/1/2 (w/w/w).Sensor K2: CNHox/KCl/PVP at a mass ratio of 6.5/1.5/2(w/w/w).Sensor K3: CNHox/KCl/PVP at a mass ratio of 6/2/2 (w/w/w).Throughout the sensor realization process, particular attention was given to the potential variability induced by the processing steps.The IDT structures were achieved using a photolithographic process with practically zero-dimensional variations.The preparation of the sensing material included a homogenization step before the material was drop-cast on the IDTs.The process was implemented using between 25 and 30 different structures for each composition.After drying, the electrical resistance of each IDT structure was measured, and an average value (Rav) was determined; samples whose electrical resistance was outside the Rav ± 5% were excluded from the following steps (i.e., recording the electrical resistance vs. R.H. variations).For each group of samples (corresponding to different material compositions), the remaining samples were between 12 and 15.The results presented in the manuscript are typical results reflecting the behavior of the samples in each group.The Raman spectra were acquired at room temperature (R.T.) using a Witec Raman spectrometer (Alpha-SNOM 300 S, WiTec.GmbH, Ulm, Germany) with 532 nm excitation.A 532 nm diode-pumped solid-state laser with 145 mW of power was employed.The incident laser beam, with a spot size of approximately 1.0 µm, was focused onto the sample using a 6 mm working distance objective attached to a Thorlabs MY100X-806 microscope.Raman spectra were recorded with an exposure time of 20 s accumulation, and the scattered light was collected by the same objective in back-scattering geometry with 600 grooves/mm

Figure
Figure 7e,f present the transfer function for the sensor K3.Noticeably, after the first cycle, the resistance at R.H. = 0 reduces from 710-715 Ω to 630-635 Ω.The same behavior is also registered for the sensor K1 (Figure 7a,b), with a decrease in resistance (for R.H. = 0) from 970 (first cycle) to 910 Ω and 930 Ω (subsequent cycles).The calculation of sensitivity was performed using the following equation:

Figure 9 .
Figure 9. Example of calculating the response time.

Figure 10 presents
Figure 10 presents the graphical representations of the investigated sensor's response time ratios relative to the reference sensor's (C.O.M.) response time calculated for each R.H. jump in humid nitrogen (R.H. = 0%-100%).

Figure 9 .
Figure 9. Example of calculating the response time.

Figure 10
Figure 10 presents the graphical representations of the investigated sensor's response time ratios relative to the reference sensor's (C.O.M.) response time calculated for each R.H. jump in humid nitrogen (R.H. = 0%-100%).

Figure 10 Figure 10 .
Figure 10 presents the graphical representations of the investigated sensor's response time ratios relative to the reference sensor's (C.O.M.) response time calculated for each R.H. jump in humid nitrogen (R.H. = 0%-100%).

Figure 10 .
Figure 10.Graphical representations of the ratios between the response time of investigated sensors and the response time of C.O.M. calculated for each humidity jump: (a) K1, (b) K2, and (c) K3 in humid nitrogen (R.H. = 0%-100%).

Table 2 .
Comparison between different CNHox-based R.H. sensing layers and sensing performances for similarly manufactured and tested sensors.