Polyaniline-Functionalized Nanofibers for Colorimetric Detection of HCl Vapor.

Hydrogen chloride (HCl) gas is a hazardous byproduct of industrial processes. Colorimetric approaches to facilitate portable and remote detection are especially desirable. We graft polyaniline to the surface of electrospun nylon nanofibers to minimize mass transfer. Using the resulting nanofibers, we demonstrate colorimetric detection of HCl at sub-ppm levels. We investigated the reusability of the fibers and observed a twofold increase in the limit of detection with multiple uses because of dedoping of the PANi indicated by elemental analysis. The limit of detection using visual detection was compared to spectrophotometric analysis. The ΔE from CIE LAB color space analysis via diffuse reflectance spectroscopy enhances the limit of detection by ∼fivefold when compared to visual detection. This analysis is a promising approach for remote detection using simple commercial digital cameras to achieve low limits of detection.

Introduction

Hydrogen chloride (HCl) gas is a common byproduct of industrial processes. Because it is hazardous, HCl is strictly regulated. An exposure to 5 ppm is considered a workplace hazard. Therefore, portable and remote detection of HCl at the ppm level is necessary. Colorimetric detection that does not require instrumentation is especially desirable.[1−6]

Optochemical detection of HCl has been reported using thin films of meso-tetra(4-pyridyl)porphyrin, which undergoes a color change upon protonation. Monitoring the change in UV absorbance at 470 nm, the detection limit was 0.01 ppm HCl.[4] Similar porphyrins have been incorporated into hydrophobic membranes for sensing. The response rate of the sensor was observed to be diffusion-controlled.[7]

Electrospun nanofibers are considered promising for improving sensing performance because of their high surface area-to-volume ratio. Porphyrinated polyimide (porphyrins incorporated into the polymer backbone)[5] or blends of styrene–ethylene–butylene–styrene triblock copolymer with pH-sensitive, pyridine-based dyes[3] has been electrospun for fluorescence-based HCl detection. Improved sensitivity was observed when comparing electrospun nanofibers and drop-cast polymer membranes.

To avoid the need for instrumentation, colorimetric HCl detection using polymer nanofibers has also been considered using polydiacetylenes, a pH-responsive, conjugated polymer that transitions from blue to red upon exposure to HCl because of the formation of quaternary ammonium salts. Visible detection of 2 ppm was reported. However, the nanofibers were not stable upon exposure to HCl, so reuse was not possible.[1]

Using nylon–polyaniline blend nanofibers, ppb-level detection of HCl was achieved using a quartz crystal microbalance (QCM). The limit of detection was significantly lower than polyvinyl alcohol–polyaniline films (∼10 ppm). Frequency shift response curves as a function of time were comparable for three cycles.[2] When using a QCM for detection, a color change from blue to green was observed upon exposure to HCl because of the protonation of the polymer backbone.[2,8] However, additional studies are required to establish colorimetric HCl detection.

In this work, we demonstrate colorimetric detection of HCl using polyaniline grafted to electrospun nylon-6,6 nanofibers to minimize diffusion limitations. We focus on evaluating the effect of reuse on the limit of detection. We further compare the limit of detection using visual detection to spectrophotometric analysis.

Results and Discussion

Nanofiber Preparation

Nylon nanofibers were prepared by electrospinning a solution of nylon in formic acid. In a typical experiment, nylon 6,6 was dissolved in formic acid (20 wt %). Electrospinning nylon–formic acid solutions resulted in 290 ± 80 nm nanofibers (Figure A). Some netting, that is, the presence of networks of thin nylon nanofibers, was observed. This observation is consistent with recent work using formic acid as a solvent. The formation of the “nanonets” was attributed to phase separation of the polymer during solvent evaporation.[9]

Figure 1

SEM images of (A) electrospun nylon, (B) acid-treated nylon nanofibers, (C) PANi-grafted to nylon nanofibers exposed to acid and reset with ammonia, and (D) PANi-grafted to nylon reset 15 times with a 3 μm scale bar.

Graft polymerization of polyaniline is initiated from amine groups on the nylon nanofiber. To increase the amine groups available for grafting, the nylon nanofibers were partially hydrolyzed by HCl. Nanofibers were cut into 0.5 in. diameter pieces, then placed in 2.5 M HCl at ambient conditions for 1 h, and finally washed with deionized water until the wash was neutral. Acid treatment of the nylon fibers cleaves the amide groups exposing a carboxylic acid and amine functional group.[10] Polyaniline was grafted to the nylon fibers via chemical oxidative polymerization at 0 °C based on previous reports.[11] Briefly, the acid-treated fibers were placed in aniline hydrochloride dissolved in HCl. Ammonium persulfate was added dropwise with constant stirring. The reaction proceeded on ice with constant stirring for 2.5 h. The PANi was converted to the emeraldine base form by sonicating the sample in ammonium hydroxide. Because the emeraldine base form is soluble in N-methyl-2-pyrrolidone (NMP), the fabrics were washed with NMP to remove any PANi that was not covalently attached to the surface.[12,13]

Scanning electron microscopy (SEM) confirms that the acid treatment did not significantly affect the morphology of the fibers (Figure B). The fiber diameter decreased slightly with the acid treatment to 190 ± 60 nm. Following the acid treatment, PANi was grafted to the nylon nanofibers. Similar to the acid treatment, the addition of PANi did not affect the nanofiber morphology (Figure C). The fiber diameter was comparable (170 ± 40 nm) after the graft polymerization. Exposing the PANi-containing fibers to HCl and resetting with aqueous ammonium hydroxide multiple times (at least 15 resets), the fiber structure remains unaffected (Figure D). The size after multiple resets was comparable 200 ± 30 nm.

Fourier Transform Infrared Spectroscopy

Minimal changes in Fourier transform infrared spectroscopy (FTIR) were observed following acid treatment, consistent with previous studies.[14] Upon grafting, features of both nylon and PANi are present and overlap. Analyzing a direct comparison of nylon with PANi to neat nylon, differences in the spectra are attributed to the presence of PANi (Figure ). Specifically, we examine changes in the relative peak intensities that increase with increasing species (PANi) concentration as an indication of the presence of PANi.[15] To facilitate comparison, spectra were baseline-corrected to the value at 1800 cm–1 and normalized to the peak at ∼1660 cm–1. We observed an increase in the ratio of amide II (C–N stretching) to amide I (C=O stretching) to bands at ∼1540 and ∼1660 cm–1,[16] respectively. We attribute this difference to an overlap of the amide II group peak of nylon with benzenoid ring stretching of PANi at ∼1540 cm–1. Further, upon grafting, we observe new peaks at ∼820 and ∼1495 cm–1 that are not present in the nylon spectra. The peak at ∼820 cm–1 has been attributed to the out-of-plane CH bending vibrations of PANi, and the peak at ∼1495 cm–1 is associated with the quinoid ring of PANi[14] that is not present in nylon. Overall, the changes in the spectra upon grafting indicate a successful modification of the nylon fibers with PANi.

Figure 2

FTIR-ATR spectra of nylon and PANi-grafted to nylon nanofibers normalized and baseline-corrected. Upon grafting, new peaks at ∼820 and ∼1495 cm–1 are observed because of the presence of PANi.

Diffuse Reflectance Spectroscopy

Diffuse reflectance spectroscopy confirms the polymerization of PANi in the emeraldine salt form. The emeraldine salt form can be converted to the emeraldine base form by exposing the sample to ammonium hydroxide. Upon exposure to ammonium hydroxide, the PANi backbone is deprotonated. There is a corresponding change in optical properties, that is, green to blue (Figure A), as can also be identified by the maximum in % reflectance shifts from ∼540 to ∼475 nm (Figure B). The changes in the spectra as a function of pH are comparable to previous reports.[17,18] Because this change is reversible and can be repeated for multiple cycles, it is the basis for the colorimetric detection of HCl vapor.

Figure 3

(A) Photograph of PANi-grafted nanofibers in the emeraldine base form (blue, bottom) and emeraldine salt form (green, top). (B) Analysis of the different forms of PANi by diffuse reflectance spectroscopy.

Colorimetric Detection of HCl Vapor

Proof of Concept

For visual detection of HCl, the PANi-containing fibers were suspended above HCl solutions in a sealed container. Half of the sample was covered, and half of the sample was exposed to the acid (Figure A). The sample was observed after a set amount of time (arbitrarily set at 5 min) to determine if a visual color change occurred. Six fabrics were tested. A color change was observed at ∼0.09 ppm HCl vapor concentration. This result is comparable to previous studies using polyaniline and QCM for detection.[19]

Reuse

For practical applications, reusability of the sample is an important consideration. Following exposure to HCl, fibers were reset to blue with 0.1 M ammonia, dried, and retested. When reset to blue with ammonia, the concentration of HCl vapor that could be detected visually within 5 min increased by ∼50% (Figure ). With additional resets, the concentration of HCl vapor that could be detected visually within 5 min increased from 0.09 to 0.16 ppm. The concentration of HCl vapor that could be detected visually after 15 resets was comparable to 6 resets (Figure ).

Figure 4

Minimum amount of HCl detected visually within 5 min initially and after being reset multiple times. The minimum amount of HCl increases after 2 and 6 resets but remains consistent from 6 to 15 resets (p < 0.05).

In contrast with our experiments, when PANi deposited on a filter paper was used for the colorimetric detection of ammonia, no change in the limit of detection was noted for up to 50 cycles.[8] To understand the observed decreased limit of detection of HCl vapor, we further analyzed the initial samples and the reset samples using X-ray photoelectron spectroscopy (XPS).

We performed elemental analysis of the fibers after 15 resets for comparison to the initial nanofibers (Table ). Comparison using XPS indicates a decrease in the chlorine content with multiple resets. This dedoping has been observed previously[20−22] and attributed to the diffusion of the small anion dopants out of the PANi.[20] In previous reports, dedoping affects the PANi backbone oxidation state and reduces the conductivity of PANi. The change in the oxidation state affects the colorimetric detection, and as the dopant level decreases, the amount of HCl required to induce a change in color increases. Given that the dopant level is 0% after 15 resets, no further decrease in the limit of detection is expected with further reuse. A consistent response would be expected in a practical setting, provided an appropriate pretreatment to dedope the fabrics.

Table 1

XPS Elemental Analysis of PANi (Emeraldine Base Form)-Containing Fibers Initially and after 15 Resets

resets	Cl (at. %)
0	0.09 ± 0.02
15	0.00 ± 0.00

Limit of Detection Comparison: Visual Detection Versus Spectrophotometric Analysis

The sub-ppm limit of detection is promising, and colorimetric detection offers an equipment-free method of analysis. With visual detection, we expect there to be a trade-off between the ease of use and limit of detection. Therefore, we compared the limit of detection using visual determination with diffuse reflectance spectroscopy after 15 resets. Using visual determination, a color change was observed at a concentration of 0.2 ppm. Using diffuse reflectance spectroscopy, we used two metrics to define the color change: (1) an increase in the wavelength at which the maximum reflectance occurred of 20 nm or (2) a ΔE of 5 in the CIE LAB color space[23] from the diffuse reflectance spectroscopy measurement. The CIE L*A*B space was defined by the International Commission on Illumination, abbreviated CIE. It uses three axes to define color, lightness (L), A-axis to differentiate red and green, and B-axis to differentiate blue and yellow. The L*A*B space was selected because it is strongly correlated with human visual perception.[24] The limit of detection using the CIE L*A*B color space was 0.04 ppm, whereas the limit using the shift in the wavelength was 0.08. As expected, spectrophotometric analysis decreased the limit of detection by ∼fivefold. Interestingly, the changes in the CIE L*A*B color space were ∼twofold more sensitive than monitoring changes in the reflectance spectra. The use of the CIE L*A*B color space to define a color change can be adapted to commercial cameras to facilitate remote monitoring as a future direction.[25]

Overall, the focus of this work was the visual detection of HCl using PANi-modified polymer nanofibers. PANi is a useful platform for evaluating visual detection and has been used for the detection of other compounds such as NH3, H2S, etc.[26] Because the mechanism of the color change is due to the protonation of the PANi backbone,[2,8] this approach is not specific to HCl. Selective detection of HCl may be possible by doping PANi with iron and aluminum[6] and will be the objective of future studies.

Conclusions

Overall, we have demonstrated the use of PANi-grafted nylon nanofibers for the colorimetric detection of HCl. Sub-ppm levels of HCl were detected visually within 5 min. We investigated the reusability of the fibers, and the limit of detection increased by ∼twofold with multiple uses. We attribute this increase in the limit of detection to dedoping of the PANi indicated by elemental analysis. After an initial decrease in the dopant level, the limit of detection remains consistent, which is important for practical applications. Complete dedoping is observed after 15 resets, so no further decrease in the limit of detection is expected with additional reuses. Finally, we quantify the trade-off between the ease of use and limit of detection by comparing visual detection to spectrophotometric analysis. The limit of detection improved by ∼fivefold using the ΔE from the CIEL*A*B color space analysis. This analytical approach is promising for remote detection using commercial digital cameras to achieve low limits of detection with simple equipment.

Methods

Materials

Nylon-6 pellets (Ultramid B40 01, viscosity average MW = 63 kDa) were obtained from BASF Corporation (Michigan, USA). NMP and ammonium hydroxide were obtained from Fisher Scientific (Wood Dale, IL, and Arlington, TX, respectively). Aniline hydrochloride, ammonium persulfate, formic acid, and hydrochloric acid were obtained from Sigma (St. Louis, MO). All chemicals were used as received.

Electrospinning

Nylon nanofibers were prepared by electrospinning a solution of nylon in formic acid. In a typical experiment, nylon 6,6 was dissolved in formic acid (20 wt %). The mixture was stirred for 3–4 h at 60 °C and cooled at an ambient temperature for 1 h without stirring to minimize air bubbles. The solution was then electrospun onto aluminum foil using a point-plate configuration composed of a precision syringe pump (New Era NE-300) and a high-voltage power supply (Matsusada Precision Inc., model AU-40R0.75) with positive polarity. Typical electrospinning parameters were a flow rate of 0.15 mL/h, tip-to-collector distance of 12 cm, 22-gauge blunt tip (0.508 mm i.d.), and an operating voltage of ∼20 kV.

Polyaniline Graft Polymerization

To graft polymerize polyaniline onto the surface of nylon fibers, the nylon was first acid-treated for partial hydrolysis to expose additional functional end groups. Nanofibers were cut into 0.5 in. diameter pieces, then placed in 2.5 M HCl at ambient conditions for 1 h, and finally washed with deionized water until the wash was neutral. The fibers were then dried at ambient conditions.

Polyaniline was grafted to the nylon fibers via chemical oxidative polymerization at 0 °C based on previous reports.[11] Briefly, the acid-treated fibers were placed in 10 mL of 0.150 M aniline hydrochloride dissolved in 1.2 M HCl and placed on ice. Ammonium persulfate (0.150 M in 1.2 M HCl) was cooled to 0 °C and 10 mL was added dropwise with constant stirring. The reaction proceeded on ice with constant stirring for 2.5 h. The PANi–nylon samples were placed in 0.1 M ammonium hydroxide and sonicated for 30 min to convert the PANi to the emeraldine base form. Because the emeraldine base form is soluble in NMP, the fabrics were placed in NMP and sonicated for 30 min to remove PANi not covalently attached to the surface.[12,13] The NMP washes were continued until the solvent remained colorless after sonication. Finally, the fabrics were rinsed with deionized water and dried at ambient conditions before further use.

Sample Characterization

The samples were sputter-coated with gold, and the fiber morphology was examined with SEM using Hitachi SU-70 FE-SEM (accelerating voltage of 2 kV). The average fiber size and standard deviation were determined by measuring the diameter of 100 fibers using ImageJ software developed by US NIH. Nylon nanofibers and PANi-grafted samples were analyzed by attenuated total reflection (ATR) infrared spectroscopy (Thermo Scientific Nicolet iS10). Spectra were recorded from 400 to 4000 cm–1 averaging 64 scans. Diffuse reflectance spectra (350–1200 nm) of the samples were obtained with an Ocean Optics Flame spectrometer (FLAME-S-VIS-NIR-ES) equipped a tungsten–halogen light source (HL-2000-HP-FHSA, 20 W output) with a diffuse reflectance probe (RPH-1) at room temperature using poly(tetrafluoroethylene) as a standard.[27] Elemental analysis was performed using a ThermoFisher ESCALAB 250 X-ray photoelectron spectrometer with a monochromated aluminum X-ray source, survey and high-resolution spectra were recorded by 1 and 0.1 eV steps, respectively similar to previous reports.[28] Data were fit and analyzed using CasaXPS 2.3.16 V and NIST standard reference database 20, version 4.1. The dopant concentration was determined from the high-resolution spectra measuring three locations per sample.

Colorimetric Detection of HCl

Colorimetric responses to HCl vapor were carried out by visual inspection. In a sealed container, nanofibers with PANi were suspended over aqueous HCl solutions of increasing concentration until a visual color change was observed within 5 min. The HCl in the vapor phase was calculated from tabulated data.[29] Afterward, the fabric was reset by submersion in an ammonium hydroxide solution (5 mL, 0.1 M) for 1 min, dried at ambient conditions, and the colorimetric response was remeasured. To examine reusability, the colorimetric responses for up to 15 resets were measured for at least 6 samples.