Advancements in wearable ammonia sensors using polypyrrole/MWCNT coated yarn

In this study, we utilized a dip coating method to modify insulating yarn with polypyrrole and multiwall carbon nanotubes (MWCNTs) to convert it into a conductive yarn. The resulting fabricated conducting yarn underwent thorough characterization through scanning electron microscope, x-ray diffraction pattern, and thermal gravimetric analysis. Subsequently, we examined the ammonia sensing properties of the modified yarn at various stages of its development. Our findings revealed that the combination of MWCNTs followed by polypyrrole modification significantly enhanced the ammonia sensing capabilities compared to using MWCNTs or polypyrrole-coated yarn individually. Specifically, the MWCNTs followed by polypyrrole modified yarn demonstrated an excellent sensing response, remarkable repeatability (up to 24 continuous cycles), quick response time (11 ± 2 s), and recovery time (34 ± 5 s). Additionally, the sensor exhibited good linearity in detecting ammonia vapor concentrations within the range of 20–100 ppm. We also assessed the sensor’s performance with diverse vapors at room temperature, revealing its high selectivity for ammonia. Furthermore, the sensor’s response correlated linearly with yarn length. Remarkably, it demonstrated minimal sensitivity to humidity and exceptional stability over fifty days. These results have the potential to lead to the development of wearable room temperature ammonia sensors, suitable for use in agricultural and industrial chemistry, as well as in environmental, automotive, and medical applications.

Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

Introduction
Ammonia, an odorless and highly toxic gas, poses a significant health hazard when inhaled [1][2][3][4].It can originate from various sources, including fossil fuel burning, metabolic pathways, chemical engineering processes, and hazardous waste, such as plastics and fertilizers [5][6][7][8][9].Industries like chemical manufacturing, industrial hygiene monitoring, and laboratories frequently utilize ammonia, but concentrations exceeding 50 ppm can lead to harmful effects on the environment and human health.Exposure to excessive ammonia can cause eye and skin irritation, as well as chronic lung diseases.Therefore, there is a growing need for cost-effective and highly sensitive ammonia sensors with quick response times, applicable in diverse industries such as fertilizer manufacturing, food technology, and various laboratories [10].
In recent years, textile-based sensors have garnered attention due to their desirable properties, such as flexibility and wearability [11][12][13][14].Single-yarn-based sensors, an advanced version of textile sensors, have gained popularity for their small diameter and stretchable properties, making them ideal for wearable technology [15,16].Although metal oxide and conductive polymer-based sensors have been thoroughly investigated, it is worth noting that metal oxides often demand elevated operating temperatures and exhibit limited selectivity, adding complexity sensor fabrication.The current trend in gas sensor development, including those for ammonia detection, focuses on using conducting polymers in combination with various metal oxide nanoparticles or carbon-based nanomaterials such as graphene or carbon nanotubes, allowing for operation at room temperature.
Carbon nanotubes offer promise as versatile nanomaterials due to their exceptional electrical properties, mechanical stability, thermal conductivity, flexibility, corrosion resistance, large surface area, and length-to-diameter ratio.These properties can be tailored through polymer and chemical treatments for specific applications [17][18][19].Among conducting polymers, polypyrrole stands out due to its stability, electrical conductivity, and ease of fabrication via oxidation processes (electrochemical or chemical polymerization) from pyrrole [20,21].Combining polypyrrole with carbon nanotubes forms a conducting nanocomposite, highly selective and sensitive to ammonia, making it suitable for numerous commercial applications.The Π-Π conjugation in between polypyrrole and carbon nanotubes helps to form a conducting nanocomposite by strong attachment [22,23].The introduction of multiwall carbon nanotubes (MWCNTs) played a pivotal role in enhancing the sensor's response, as it facilitated the easy transfer of electrons from polypyrrole to MWCNTs.This, in turn, allowed for more ammonia molecules to interact with the active surface of polypyrrole, amplifying the sensor's performance.Previous studies on polypyrrole and carbon nanotube composites for ammonia sensing primarily used rigid substrates and microfabricated electrodes, resulting in more complex and costly processes [24][25][26][27][28][29][30].
Addressing these research gaps, our study aimed to develop an ammonia sensor on a single yarn, allowing attachment to clothing (fashion and wearable sensor industries) for ambient ammonia detection and emergency response [31][32][33].We achieved this by coating the yarn's surface with polypyrrole and multi-walled carbon nanotubes through pyrrole polymerization.Characterization using scanning electron microscope (SEM), energy-dispersive x-ray spectroscopy (EDAX), x-ray diffraction (XRD), and thermal gravimetric analysis (TGA) confirmed the formation of the nanocomposite.Extensive investigation of the sensor's ammonia sensing properties included varying ammonia concentration, yarn length, humid conditions, and different bending angles to cater to commercial applications.The sensor's performance was tested over fifty days, and to the best of our knowledge, this was the first comprehensive investigation of a multi-walled carbon nanotubes and polypyrrole modified yarn for ammonia sensing.The developed ammonia sensor exhibited low cost, ease of fabrication, wearability, stretchability, high stability, reproducibility, quick response, and recovery properties, making it highly promising for field-level applications.

Synthesis of MWCNTs/Polypyrolle modified yarn
The purchased yarn underwent a thorough washing process using nonionic detergents to eliminate any macroscopic impurities from its surface.To achieve a uniform coating, MWCNTs were dispersed in N,N-dimethylformamide through probe sonication for 10 min, with a pulse on/off time of 10 and 10 s, respectively.The cleaned and dried yarn was then immersed multiple times in the MWCNTs solution.Afterward, the MWCNT-coated yarn was dipped into a pyrrole solution for 30 min, allowing the monomer to be absorbed onto the yarn's surface.The yarn coated with the monomer was then dried at 60 • C for 4 h to ensure complete drying.Subsequently, the pyrrole-coated yarn was submerged in the FeCl 3 solution for 12 h, as part of the polymerization process, where the FeCl 3 solution acted as an oxidizing agent to polymerize the monomer.Following the polymerization process, the MWCNTs and polypyrrole-modified yarn were thoroughly washed with water and subsequently dried using a vacuum dryer at 60 • C for 12 h to ensure complete drying.Finally, the yarns were cut into sizes (ranging from 1 cm to 5 cm) and connected to copper wires using silver paste for further analysis.

Characterization
The morphological and structural characterization of samples was conducted using field emission SEM (FESEM) on a SUS010 instrument.XRD measurements were obtained using an X ′ pert Pro diffractometer with CuKα radiation in the range of 5 • < 2θ < 80 • at 40 kV.TGA was performed using the STA 6000 instrument by Saturna sensor.

Sensing of vapours
The vapor sensing properties of the sensor were assessed by measuring the change in resistance using a keysight 34 972 A instrument.The vapors of ammonia, formaldehyde, n-Hexane, propanol, toluene, butanol, cyclohexane, ethanol, methanol, n-Decane, chloroform, butanol, and benzene were generated by passing a carrier gas (nitrogen) through a bubbler using a mass flow controller (MFC).The concentration of vapors was precisely controlled within the range of 20-100 ppm using a programmable MFC, following equation ( 1) [33] The sensor response was calculated by measuring the change in resistance when exposed to vapors, relative to the base resistance value, using equation (2).
where R g and R b represent the resistance values of the sensor films under vapors and fresh nitrogen, respectively.The response and recovery time of the sensor were defined as the time r for 90% of the resistance change from the initial value.A schematic diagram of the vapor sensing setup, which includes a MFC, bubbler, chamber, humidity sensor, and keysight 34 972 A, is illustrated in figure 1.

Characterization of modified yarn
In figure 2, SEM images of the Y-CNT-polypyrrole coated yarn (PPy) (MWCNTs followed by polypyrrole-coated yarn) are displayed at various magnifications.The SEM images reveal the irregular morphology of the polypyrrole coating, forming a densely covered matrix over the MWCNTs coated yarn.As a consequence, only a small quantity of MWCNTs is discernible in the Y-CNT-PPy sample in the SEM images.The presence of an agglomerated polypyrrole structure is noticeable due to the higher concentration of pyrrole used during the coating process.
The XRD patterns of both the pristine yarn and Y-CNT-PPy samples are presented.The pristine yarn shows distinct diffraction peaks at 2θ angles of 16.4 • , 22.7 • , and 34.7 • , corresponding to the (101), ( 101), (002), and (040) planes of the cellulose structure (see supplementary information figure S1).The characteristic peak of polypyrrole is evident at 24.6 • , attributed to the repeat unit of pyrrole rings (see supplementary information figure S1).Following the MWCNTs coating, new diffraction peaks appeared at 26 • , 42 • , 54 • , and 78 • , corresponding to the (002), (101), (004), and (110) planes of MWCNTs respectively (see supplementary information figure S1).The appearance of these new diffraction peaks indicates the presence of MWCNTs, as well as cellulose.However, in the Y-CNT-PPy sample, only an amorphous polypyrrole peak is observed, signifying the wrapping of a greater number of polypyrrole polymers on the surface of the MWCNTs coated yarn.Consequently, the crystalline properties are diminished due to the presence of polypyrrole.The thermal stability of the modified yarn, influenced by the polymer and MWCNTs, was investigated using TGA (supplementary information figure S2).The initial weight loss observed in all samples is attributed to the presence of moisture.As the temperature increased, the yarn underwent a sharp weight loss around 300 • C due to the decomposition of the cellulose matrix.The decomposition of MWCNTs commenced at approximately 550 • C. The MWCNTs coated yarn exhibited a similar weight loss pattern, likely due to the lower amount of MWCNTs present.In contrast, polypyrrole exhibited a slower decomposition process, linked to the robust structure of its polymer chain.The Y-CNT-PPy sample demonstrated a comparable degradation process, indicating a higher concentration of polypyrrole in the sample.Additionally, EDAX analysis of the Y-CNT-PPy sample is available in supplementary information figure S3.

Electrical properties of MWCNTs, polypyrrole and MWCNTs/Polypyrrole modified yarn
The electrical properties of the modified yarn were investigated through two-probe connection I-V measurements (supplementary information figure S4).The MWCNTs modified yarn exhibited a linearly varying current with applied voltage, indicating its ohmic nature.On the other hand, the PPy and the MWCNTs/PPy demonstrated nonlinear behavior due to the semiconducting nature of polypyrrole.The resistance values for MWCNTs, polypyrrole, and MWCNTs and polypyrrole modified yarn were measured to be 0.54 MΩ, 32.3 MΩ, and 7.4 MΩ, respectively.The addition of MWCNTs provided a pathway for charge carriers, effectively reducing the resistance of the modified yarn.

Characterization of vapours sensing performances
The Y-CNT (MWCNT coated yarn) sample exhibited a slight change in resistance, while the Y-PPy sample showed an increase in resistance upon exposure to ammonia vapors (figures 3(a)-(c)).However, the presence of MWCNTs further enhanced the sensor response.Remarkably, the sensor response of the Y-CNT-PPy sensor was 15 times better than that of Y-CNT alone.For 20 ppm ammonia vapors, the change in resistance values for Y-CNT, Y-PPy, and Y-CNT-PPy were 2.5 kΩ, 326 kΩ, and 1209 kΩ, respectively (as shown in figure 3(d)).
These results suggest that polypyrrole plays a significant role in ammonia sensing, and the electrical conductivity of the sensor can be improved by the presence of MWCNTs.The combination of polypyrrole and MWCNTs in the Y-CNT-PPy sensor demonstrates enhanced performance, making it a promising candidate for ammonia sensing applications.
The reproducibility and repeatability of the Y-CNT-PPy sensor towards ammonia were investigated (figure 4(a)).The sensor exhibited excellent reproducibility, with a 9% standard deviation, over 24 continuous measurement cycles of ammonia vapor.Successful recovery of resistances was achieved by exposing the sensor to nitrogen gas to replace ammonia vapors.The response and recovery times of the Y-CNT-PPy sensor for 20 ppm ammonia were (11 ± 2 s) and (34 ± 5 s), respectively (figure 4(b)).These results indicate the excellent gas sensing performance of the Y-CNT-PPy sensor for ammonia detection.The developed selectivity of the sensing platform for ammonia detection was evaluated by testing the sensor's response to major interference molecules such as formaldehyde, n-Hexane, propanol, toluene, butanol, cyclohexane, ethanol, methanol, n-Decane, chloroform, butanol, and benzene, all at a concentration of 20 ppm.The calibration of sensor response to these vapors is depicted in figure 5.The sensor response of the Y-CNT-PPy sensor towards the tested vapors resulted in values of 0.33, 0.65, 0.46, 1.55, 0.55, 1.31, 2.2, 2.7, 0.53, 0.69, 1.62, and 1.8 for formaldehyde, n-Hexane, propanol, toluene, butanol, cyclohexane, ethanol, methanol, n-Decane, chloroform, butanol, and benzene, respectively.Notably, the sensor response of the Y-CNT-PPy towards ammonia was 14.6.These results confirm that our sensor exhibits high selectivity for ammonia compared to other vapors.

Bending effects on sensing properties of the Y-CNT-PPy ammonia sensor
The investigation of impact of bending on sensing properties is crucial for the development of wearable and flexible sensing platforms.Figures 6(a)-(d) illustrate the ammonia sensing properties of the Y-CNT-PPy ammonia sensor at 120 • (forward bending), 180 • (flat surface), and 240 • (backward bending), respectively.It was observed that the sensor responses slightly increased during backward bending, possibly due to the enlarged surface area facilitating more interactions with ammonia molecules.These results indicate that the sensing properties of Y-CNT-PPy remain consistent during bending tests, making it a suitable platform for fabricating wearable sensors.
Additionally, it is important to assess the sensor's performance by varying the length of the yarn for the development of wearable sensors.Figure 7(a) demonstrates the change in resistance over time, while figure 7(b) reveals that the resistance of the sensor increases linearly with yarn length when exposed to ammonia vapors.However, the sensitivity of the yarn decreases as the length of the sensor increases.The resistance of the sensor also increases with the length of the yarn due to the limitation in carrier transformation between yarn segments.The response of the ammonia sensor under different humidity conditions range from 19% (1st cycle), 37% (2nd cycle), 67% (3rd cycle), 75% (4th cycle), to 93% (5th cycle) was tested, as shown in figure 7(c).A slight increase in base line resistance of the sensor was observed with higher relative humidity.However, it was found that humidity had negligible effects on the ammonia sensing properties of the sensor.For stability analysis, the Y-CNT-PPy ammonia sensor was tested in ambient conditions, as depicted in figure 7(d).Remarkably, there was minimal variation in the response even after fifty days of continuous analysis, indicating excellent stability over an extended period.The polypyrrole and MWCNTs composite is considered a p-type semiconductor, where holes are the majority carriers.On the other hand, ammonia is a reducing gas due to its electron-donating nature.When exposed to ammonia vapors, the ammonia molecules are adsorbed on the surface of polypyrrole.The active -NH groups of polypyrrole capture electrons, neutralizing the charge carrier of the sensor.As a result, the reduction of majority carriers and carrier density leads to an increase in the sensor's resistance.The transfer of electrons from PPy to MWCNTs occurs at a higher rate due to the low potential barrier for electrons.The easy transfer of electrons from the surface of PPy to MWCNTs allows more ammonia molecules to interact with the active surface of PPy, resulting in an increased sensitivity of the sensor in the presence of MWCNTs (figure 8).The electron transfer phenomena can be explained by the following equation (3): The base resistance of the sensor was found to be slightly increased on exposure to air, which could be O 2 , and this increase can be attributed to the reaction between ammonia and adsorbed oxygen with the sensing material.The electrons produced from the reaction between the sensing material, oxygen, and ammonia can lead to the depletion of the hole concentrations in the PPy/MWCNTs composites.During the recovery process, the Y-CNT-PPy ammonia sensor undergoes proton     transfer phenomena, where the -NH group of the polypyrrole ring removes hydrogen atoms from ammonia ions [36].This charge transfer process is reversible, allowing the sensor to recover when exposed to nitrogen gas.Under higher humidity conditions, a shift in resistance is observed as water molecules react with ammonia vapors, creating -OH ions.The enhanced sensing response of the MWCNTs and polypyrrole modified yarn is attributed to the low potential barrier of electrons between polypyrrole and MWCNTs.This facilitates easy transfer of electrons to MWCNTs, which are then absorbed by the active side of polypyrrole.As a result, the change in resistance increases due to the reduction of the majority carrier of the sensor.Until now, there were no existing reports on yarn-based ammonia sensors.A comparison of the ammonia sensing behaviors of MWCNTs and polypyrrole composites was also made with the literature available (table 1).Our simple approach to a wearable and stretchable sensor, without the need for prefabricated electrodes and complex fabrication processes, demonstrated superior gas sensing performance compared to other studies.Thus, the MWCNTs and polypyrrole modified conducting yarn proved to be a promising candidate for developing a commercial wearable and stretchable ammonia sensor, operating efficiently at room temperature.

Conclusion
In a significant advancement in gas sensing technology, our study reports the successful development of a wearable and flexible ammonia sensor through the modification of insulating yarn with polypyrrole and MWCNTs.Utilizing dipcoated MWCNTs and polypyrrole-modified yarn, the ammonia sensor excels in sensitivity, selectivity towards ammonia, with a short response time (11 ± 2 s) and recovery time (34 ± 5 s), while also demonstrating good stability and repeatability.Ammonia's electron-donating nature increases sensor resistance through electron capture by -NH groups on polypyrrole, while MWCNTs expedite electron transfer, promoting higher sensor sensitivity and overall performance.This composite of polypyrrole and MWCNTs has emerged as a highly promising candidate for the development of wearable and flexible ammonia sensors that are capable of functioning effectively at room temperature.This study introduces an advanced wearable ammonia sensor and opens new horizons in the field of gas sensors, with vast potential applications spanning fertilizer manufacturing, food technology, and safety monitoring.
This novel sensor demonstrated not only high sensitivity but also impressive selectivity, along with remarkable repeatability observed over 24 cycles, showcasing its practical utility.The sensor displayed a linear response within the range of 20-100 ppm ammonia vapor, a critical feature for precise ammonia concentration measurements.One of the most remarkable findings of this study is the sensor's long-term stability, remaining effective over a continuous analysis period of fifty days.This attribute positions the Y-CNT-PPy sensor as an ideal choice for real-world field-level applications, such as industrial hygiene monitoring, safety in fertilizer manufacturing, food technology, and emergency response.

Figure 1 .
Figure 1.Schematic depiction of the ammonia sensing setup.

Figure 2 .
Figure 2. SEM images of the CNT-PPy coated yarn are shown in figure 1, capturing both higher magnification (a) and lower magnification (b).

Figure 4 (
c) illustrates the response and recovery curve of the Y-CNT-PPy sensor for different concentrations of ammonia gas ranging from 20-100 ppm.The resistance of the ammonia sensor increased significantly with the increase in ammonia concentration.The sensor response of the Y-CNT-PPy sensor with calibration for different concentrations of ammonia is displayed in figure 4(d).Our sensor operates effectively within the 20-100 ppm range, exhibiting ultralow hysteresis (figure 4(a)) and minimal baseline drift (figure 4(c)) when exposed to ammonia.

Figure 3 .
Figure 3. (a) Y-CNT, (b) Y-PPy, and (c) Y-CNT-PPy responses to 20 ppm ammonia at room temperature (1 cm).(d) Calibration of base resistance (left y-axis) and change in resistance (right y-axis) on exposing 20 ppm ammonia based on analysis of five samples.These response curves demonstrate the respective changes in resistance values of the sensors upon exposure to the ammonia vapor concentration.

Figure 4 .
Figure 4. (a) Y-CNT-PPy (1 cm) sensor's dynamic response to 20 ppm ammonia for 25 cycles.(b) Response and recovery time of Y-CNT-PPy (1 cm) ammonia sensor.(c) Change of resistance vs. ammonia concentration in the ranges of 20-100 ppm and 100-20 ppm ammonia in two cycles.(d) Calibration curve for ammonia concentration using the sensing response from a set of five Y-CNT-PPy (1 cm) ammonia sensors.

Figure 6 .
Figure 6.The response and recovery curves of the Y-CNT-PPy (1 cm) ammonia sensor at 120 • (forward bending), 180 • (flat surface), and 240 • (backward bending).Figure 6(d) presents a comparison of the sensor response and change of resistance at different bending angles.

Figure 6 (
Figure 6.The response and recovery curves of the Y-CNT-PPy (1 cm) ammonia sensor at 120 • (forward bending), 180 • (flat surface), and 240 • (backward bending).Figure 6(d) presents a comparison of the sensor response and change of resistance at different bending angles.

Figure 7 .
Figure 7. (a) illustrates the response and recovery curves of the Y-CNT-PPy ammonia sensor.(b) The calibration of the Y-CNT-PPy ammonia sensor at various yarn lengths is shown in.(c) The impact of humidity, spanning from 19% (1st cycle) to 93% (5th cycle), on the sensing properties of ammonia is illustrated.(d) Demonstrates the stability of the sensor at different days of measurements.

Table 1 .
Comparison of resistive-based ammonia sensing performances using PPy/MWCNTs in existing literature.