Silver-loaded carbon nanofibers for ammonia sensing

Abstract Carbon nanofibers (CNFs) were prepared by electrospinning, and silver (Ag) ions were grown on the surface of the CNFs by in situ solution synthesis. The structure and morphology of obtained Ag-doped CNFs (Ag-CNFs) were characterized by X-ray diffraction (XRD) and scanning electron microscopy (SEM). The gas sensibility of the composite fiber was investigated by ammonia (NH3) obtained by natural volatilization from 1 to 4 mL of NH3 solution at room temperature. It was found that the fibers exhibited a sensitive current corresponding to different NH3 concentrations and a greater response at high concentrations. The sensing mechanism was discussed, and the good absorptivity was demonstrated. The results show that Ag-CNF is a promising material for the detection of toxic NH3.

Nowadays, metal oxide semiconductors and solid electrolytes sensors occupy most of the markets for gas sensors (31,32). However, both of them need to work at higher temperatures (hundreds to more than 1,000°C), consume large power, and have low sensitivity, poor anti-interference ability, and inconvenient use. With the development of nanotechnology, a large number of research reports on carbon-based gas sensors have been published in recent years, which show good analytical sensitivity at room temperature (33,34), such as carbon nanotubes (CNTs), graphene, graphene oxide and activated carbon (35)(36)(37)(38)(39). 1D CNFs have high surface adsorption capacity, good electrical conductivity, electronic ballistic transmission characteristics, and other excellent properties, becoming one of the ideal materials for the fabrication of nanoscale gas sensors with high sensitivity, fast response, small size, and low energy consumption (40)(41)(42). Among the many CNFs, the CNFs prepared based on the polyacrylonitrile (PAN) are the most common ones with high tensile strength (43)(44)(45), low production cost, and suitable for large-scale production.
Ammonia is a common irritating gas. It is widely used in chemical and agricultural fields and has corrosive and irritating effects on human skin and mucous membranes. This study found that carbon materials have good response characteristics to NH 3 . Silver and its compounds are one of the most important antibacterial materials. They have high bactericidal activity and biocompatibility and have antibacterial effects against bacteria, fungi, and even viruses. Many studies have modified nanomaterials based on this property of Ag, such as Ag-CNT, Ag/ZNO, and so on, see Table 1 for details (46)(47)(48)(49)(50)(51)(52)(53)(54). This study describes an antibacterial NH 3 sensor based on the CNFs. Ag-CNFs were prepared by electrospinning and impregnation. We measured and analyzed the response characteristics to NH 3 by focusing on the change in resistance of Ag-CNFs and found that it exhibits good NH 3 sensing performance in 1-4 mL of NH 3 solution. Ag-CNFs can be used as an NH 3 gas sensor at room temperature that is inexpensive to produce, flexible, sensitive, and antibacterial.

Preparation of pure PAN fibers by electrospinning
The PAN-DMF solution having a polymer mass ratio of 12% was prepared by stirring at room temperature for 4 h by a magnetic stirrer. Self-assembled equipment was used for electrospinning, including a high-voltage power supply, propulsion pumps, and aluminum foil as the receiver. The spinning solution was placed in a plastic syringe and mounted on a propulsion pump. The aluminum foil was placed perpendicular to the horizontal plane and connected to the ground. The volume feed rate, applied voltage, and tip-to-collector distance were 1 mL/h, 15 kV, and 10 cm, respectively. The spinning temperature was 20°C and the humidity was 50%.

Preparation of Ag-CNFs
The pure PAN fiber obtained by electrospinning was subjected to a two-stage heating process. First, the preoxidation process was carried out by heating at 260°C for 140 min and the heating rate was 2°C/min. Then, the temperature was raised to 900°C for 120 min at a rate of 5°C/min in an argon atmosphere. Finally, natural cooling is performed to obtain CNFs. Since the Ag nanostructure has antibacterial properties, it is compounded onto the CNF to protect it (55,56). AgNO 3 (10 mM) was mixed with 20 mL water to soak the CNFs for 3 h. CNFs were taken out and washed with DMBA until no bubbles generated. Then, it was dried at 60°C for 4 h to obtain Ag-CNFs.

Sensor assembly
A simple gas sensor was assembled by sandwiching Ag-CNFs between two copper electrodes (copper sheets), and the size of the composite was 20 × 60 × 2 mm 3 . Two copper wires were fixed to the copper electrodes by soldering. Copper wire was used to connect an electrical performance measurement system that responds to changes in current in an NH 3 environment.

Characterization
The samples were characterized by XRD at room temperature. Scanning electron microscopy (SEM, JSM-6390) was used to observe the surface morphology of pure PAN fibers, CNFs and Ag-CNFs. Electrical properties were tested by a Keithley 6487 high resistance meter system (Washington, USA) at room temperature.

Simulation by COMSOL Multiphysics
COMSOL Multiphysics has been increasingly used in the process of simulating gas sensing mechanism. And in this article, COMSOL is applied to demonstrate the distribution of NH 3 in Ag-CNFs. The model of transport of diluted species has been used and the boundary concentration of NH 3 around Ag-CNFs has been set as 0.5 mol/m 3 to make simulation results more obvious. The temperature has been set as 20°C and the porosity of Ag-CNFs is 0.1.

Results and discussion
3.1 XRD X-ray diffraction is a powerful tool for characterizing the structure of materials. The diffraction peaks of the nanofibers are indexed by standard cards of C and Ag (PDF No. 80-0017 and 87-0717). A comparison of the XRD patterns of the CNFs and Ag-CNFs with the standard map is shown in Figure 1. It can be seen that the diffraction peaks appear in the spectra of both samples, indicating a good crystallinity. The main peak of the CNFs is at 44°, corresponding to the (111) plane of C. The main peak of Ag-CNFs at 44°corresponds to the (200) crystalline plane of Ag and the (111) plane of C, and the second-largest peak at 38°corresponds to the (111) plane of Ag. This confirms that the Ag-ions are effectively modified on the surface of the CNFs. Figure 2a shows the morphology of the as-spun pure PAN nanofibers prepared by electrospinning with a smooth surface and uniform diameter. Its average diameter is 320 ± 24.67 nm. To avoid fiber melting or fusion, we oxidized PAN nanofibers at 260°C to convert C-C and C^N to C]C and C]N bonds (57,58), followed by carbonization at 900°C. After carbonization, a significant color change from white to black was observed, and the morphology of the obtained CNFs is shown in Figure 2b. It can be seen that CNFs have uniform diameter distribution, have no significant change in appearance, and maintain fibrous morphology, which is related to the entanglement network and flexibility of pure PAN fibers. The average diameter of the CNFs is 301 ± 26.20 nm. It is reduced by ∼20 nm than pure PAN fiber, derived from the volatilization of DMF as a solvent and the disappearance of O and H components in PAN. The SEM image in Figure 2b shows that the nanofibers are slightly curved. This is because the defects in the PAN fiber itself will be inherited in the CNF fiber

NH 3 sensing measurement
We tested the NH 3 sensing performance of Ag-CNFs at room temperature. The Ag-CNF sensor was placed in a sealed container and the concentration of NH 3 was controlled by the natural evaporation of the NH 3 solution. The voltage across the sensor was 0.2 V. The structure of test device is shown in Figure 3. The electrical properties of the sensor were measured at different NH 3 concentrations. Figure 4 shows the electrical response of Ag-CNFs to 1, 2, 3, and 4 mL of the NH 3 solution, in a closed container, the density of 1-4 mL of NH 3 after volatilization is 0.0754, 0.1508, 0.2262, and 0.3016 mol/m 3 , indicating the effect of NH 3 on electrical transport behavior. It can be seen that a high concentration of NH 3 results in a bigger current, meaning that the resistance of the sensor decreases and the conductivity is improved. A sensor exposed to a higher concentration of NH 3 solution exhibits a larger current change, that is, as the gas concentration increases, the sensor response becomes larger. In addition, the sensor current remains stable at the same NH 3 concentration, indicating that Ag-CNFs have good NH 3 sensing properties. This is due to the adsorption of NH 3 by Ag-CNFs, as shown in Figure 4. Ag-CNFs is exposed to NH 3 (Figure 5a). The NH 3 molecule adheres to the pores on the surface of the sample that is not electrically conductive compared with the state after adsorption, increasing the cross section of the conductive path, as shown in Figure 5b. The resistance of Ag-CNFs adsorbed with NH 3 decreased and the current increased. When the surface adsorption reaches saturation, NH 3 molecules begin to diffuse from the surface to the inside. When the adsorption process is completed, the conductive regions of the Ag-CNFs are stabilized and the current is constant (Figure 5c). The higher the concentration of NH 3 , the larger the corresponding current change, indicating its better adsorption performance.
When the Ag-CNFs are in the NH 3 environment, the NH 3 concentration on the surface will increase rapidly. As the time in the NH 3 environment increases, the internal NH 3 concentration will gradually increase until it reaches saturation. The equation of gas diffusion in this simulation can be described by: where D i e, is gas diffusion coefficient in Ag-CNFs, ε P is porosity of Ag-CNFs, τ i F, is effective diffusivity constant, and D i F, is gas diffusion coefficient in free air. The effective diffusion coefficient model can be described by Millington-Quirk model or Freundlich equation:    Figure 6. This figure demonstrated the adsorption of NH 3 molecules on Ag-CNF by different colors and inward diffusion. As previously imagined, NH 3 adsorption mainly happened on the surface, and after a certain period of diffusion, NH 3 would appear in deeper parts, but with much lower concentration. It showed that diffusion on fibers with smaller diameters will reach equilibrium faster, which has a quicker response at low NH 3 concentration. In case of this report, this means that in a given period, the NH 3 concentration will be reflected more obviously by the change of resistance and this explains the high sensitivity of the sensor.
Ag-CNF-based sensors can be used not only for toxic gas alarm systems, but also for characterizing food spoilage. When animals, plants, and foods prepared by them are decomposed by enzymes produced by microorganisms, abnormal odors may occur. If the protein is decomposed, it will produce harmful gases such as NH 3 and H 2 S. Human senses may be more difficult to capture trace amounts of gas from food spoilage in a timely manner. By putting Ag-CNFs sensors together with food, we can get information on food spoilage in a sensitive and timely manner, avoiding food poisoning caused by consumption.
Gas sensors involve physical and chemical adsorption and other reasons. The chemical gas sensor itself requires a long time of adsorption and desorption processes, so the repeatability problem of the gas sensor has always been a problem in the field. When the gas is to be measured, there are differences in the same concentration response multiple times in a short time, but after a longer time interval and calibration, better repeatability can be obtained (62,63). With further research on the sensor, optimizing the conditions of electrospinning PAN, and to get finer fibers, continuous operation of the gas sensor reliability will be improved due to the larger specific surface area.

Conclusions
In summary, Ag-CNFs were successfully prepared by electrospinning and in situ solution polymerization, prepared into an NH 3 sensor at room temperature. The surface of the CNFs was modified with Ag nanoparticles having antibacterial properties to protect the fibers. It can be used to detect NH 3 naturally evolved from 1 to 4 mL of NH 3 solution in a closed vessel and has a high response at higher NH 3 concentrations. This work can provide a viable way to create a low-cost, highsensitivity, stable, detectable NH 3 concentration sensor for use in toxic gas alarm systems and characterize food spoilage.