Actuator-Driven, Purge-Free Formaldehyde Gas Sensor Based on Single-Walled Carbon Nanotubes

Abstract

Formaldehyde vapor (HCHO) is a harmful chemical substance and a potential air contaminant, with a permissible level in indoor spaces below 0.08 ppm (80 ppb). Thus, highly sensitive gas sensors for the continuous monitoring of HCHO are in demand. The electrical conductivity of semiconducting nanomaterials (e.g., single-walled carbon nanotubes (SWCNTs)) makes them sensitive to chemical substances adsorbed on their surfaces, and a variety of portable and highly sensitive chemiresistive gas sensors, including those capable of detecting HCHO, have been developed. However, when monitoring low levels of vapors (<1 ppm) found in ambient air, most chemiresistive sensors face practical issues, including false responses to interfering effects (e.g., fluctuations in room temperature and humidity), baseline drift, and the need to apply a purge gas. Here, we report an actuator-driven, purge-free chemiresistive gas sensor that is capable of reliably detecting 0.05 ppm of HCHO in the air. This sensor is composed of an HCHO→HCl converter (powdery hydroxylamine salt, HA), an HCl detector (a SWCNT-based chemiresistor), and an HCl blocker (a thin plastic plate). Upon exposure to HCHO, the HA emits HCl vapor, which diffuses onto the adjacent SWCNTs, increasing their electrical conductivity through p-doping. Meanwhile, inserting a plastic plate between HA and SWCNTs makes the conductivity of SWCNTs insensitive to HCHO. Thus, via periodic actuation (insertion and removal) of the plastic plate, HCHO can be detected reliably over a wide concentration range (0.05-15 ppm) with excellent selectivity over other volatile organic compounds. This actuator-driven system is beneficial because it does not require a purge gas for sensor recovery or baseline correction. Moreover, since the response to HCHO is synchronized with the actuation timing of the plate, even small (~0.8%) responses to 0.05 ppm of HCHO can be clearly separated from larger noise responses (>1%) caused by interfering effects and baseline drift. We believe that this work provides substantial insights into the practical implementation of nanomaterial-based chemiresistive gas sensors.

Keywords: carbon nanotube; chemiresistor; formaldehyde; gas sensor; semiconductor.

Figure 1

Actuator-driven, purge-free chemiresistive HCHO sensor used in this work. (a) HCHO sensor composed of a powdered hydroxylamine salt sandwiched between hydrophobic polytetrafluoroethylene (PTFE) membrane filters (HA), an SWCNT network embedded on a comb-shaped gold electrode (SWCNT-based chemiresistor), and a thin plastic plate. The HA and SWCNT-based chemiresistor are spatially isolated, and the intervening plastic plate is periodically removed to monitor HCHO in the air. Note that the sensor was only exposed to analyte gases without switching to purge gas. (b) Condensation reaction between HCHO and HA, generating HCl vapor.

Figure 2

Sensing responses for HCHO at different concentrations: (a) blank air (50% RH), (b) 0.05 ppm, (c) 0.27 ppm, (d) 1.2 ppm, (e) 3.8 ppm, and (f) 15 ppm. Temporary removal (30 s) of the thin plastic plate was repeated every 1000 s (three times per run as indicated by the blue arrows). (g) Calibration curve: HCHO concentration versus normalized sensing response (mean of three repeated measurements with standard deviations). See Figure S2 and the corresponding test in SI for details on the calibration curve’s construction. (h) Normalized responses for blank air and 0.05 ppm of HCHO used for LoD estimation.

Figure 3

Sensing responses for non-carbonyl VOC vapors: (a) blank air (50% RH), (b) water, (c) methanol, (d) ethanol, (e) ethyl acetate, (f) tetrahydrofuran, (g) chloroform-d, (h) toluene, and (i) hexane. Temporary removal (30 s) of the thin plastic plate was repeated every 1000 s (four times per run as indicated by the blue arrows). Values (in ppm) quoted at the bottom right in each panel indicate the concentration of water or each VOC added into the airflow (50% RH). (j) Normalized sensing responses for each vapor. The average of four repeated measurements, along with their standard deviation, is shown. Since conventional chloroform (CHCl₃) contains ethanol as a stabilizer, chloroform-d (CDCl₃, NMR grade) was used in this study.

Figure 4

Sensing responses for carbonyl-containing VOC vapors. (a–c) acetaldehyde, (d,e) acetone, and (f,g) 2-butanone. (h) Normalized sensing responses for carbonyl vapors. The average of three repeated measurements, along with their standard deviation, is shown.

Figure 5

Continuous monitoring of HCHO in the air. Temporal removal (30 s) of the thin plastic plate was repeated every 1000 s in the presence and absence of HCHO.