Trace Hydrogen Sulfide Sensing Inspired by Polyoxometalate-Mediated Aerobic Oxidation

Abstract

A high-performance chemiresistive gas sensor is described for the detection of hydrogen sulfide (H₂S), an acutely toxic and corrosive gas. The chemiresistor operates at room temperature with low power requirements potentially suitable for wearable sensors or for rapid in-field detection of H₂S in settings such as pipelines and wastewater treatment plants. Specifically, we report chemiresistors based on single-walled carbon nanotubes (SWCNTs) containing highly oxidizing platinum-polyoxometalate (Pt-POM) selectors. We show that by tuning the vanadium content and thereby the oxidation reactivity of the constituent POMs, an efficient chemiresistive sensor is obtained that is proposed to operate by modulating CNT doping during aerobic H₂S oxidation. The sensor shows exceptional sensitivity to trace H₂S in air with a ppb-level detection limit, multimonth stability under ambient conditions, and high selectivity for H₂S over a wide range of interferants, including thiols, thioethers, and thiophene. Finally, we demonstrate that the robust sensing material can be used to fabricate flexible devices by covalently immobilizing the SWCNT-P4VP network onto a polyimide substrate, further extending the potentially broad utility of the chemiresistors. The strategy presented herein highlights the applicability of concepts in molecular aerobic oxidation catalysis to the development of low-cost analyte detection technologies.

Figure 1

H₂S sensing strategy reported in the present study. Abbreviations: SWCNT, single-walled carbon nanotube; POM, polyoxometalate.

Figure 2

(a) Selector incorporation into a SWCNT-P4VP film by stepwise soaking in [Pt][SO₃OCH₃] and POMn (n = 1–4) solutions, respectively. (b) Fraction of V content in chemiresistors containing POM1–POM4 as probed by XPS. Possible POM isomers and H⁺ equivalents are omitted for clarity. (c) Averaged conductance trace (represented as ΔG/G₀, %) of SWCNT-P4VP-Pt-POMn (n = 1–4) in response to 10 ppm of H₂S. (d) Chemiresistive responses of SWCNT-P4VP-Pt-POMn (n = 1–4) to 1 min H₂S exposures (10 ppm). Shaded areas and error bars represent standard deviations (N = 4 chemiresistors); all data were collected in air at room temperature. (e) Raman spectra of chemiresistor films on glass substrates (633 nm excitation wavelength).

Figure 3

(a) Averaged conductance trace of SWCNT-P4VP-Pt-POM3 in response to three repeated 1 min exposures of 10 ppm of H₂S each. (b) Control H₂S sensing experiments omitting selector components in comparison with the sensing response of SWCNT-P4VP-Pt-POM3 to a 1 min exposure of 10 ppm of H₂S. (c) Averaged conductance traces of SWCNT-P4VP-Pt-POM3 in response to 1 min exposures of varying H₂S concentrations. (d) Chemiresistive responses of SWCNT-P4VP-Pt-POM3 to 1 min H₂S exposures of varying concentration. Shaded areas and error bars represent standard deviations (N = 4 chemiresistors); all data were collected at room temperature in air.

Figure 4

High-resolution XPS spectra showing S 2p (a) and V 2p (b) binding energies of SWCNT-P4VP-Pt-POM3 before (blue trace) and after (red trace) exposure to H₂S gas (10 ppm) for 30 min at room temperature in air. In (b), high-resolution XPS spectrum of V 2p of H₆PV₃Mo₉O₄₀ is shown for reference. (c) Proposed H₂S sensing mechanism.

Figure 5

Chemiresistive response of SWCNT-P4VP-Pt-POM3 to various (a) VOCs and (b) S-containing species in comparison with the sensor response to H₂S following interferant exposure. (c) Chemiresistive responses of SWCNT-P4VP-Pt-POM3 to 1 min H₂S exposures (10 ppm) after storage on a laboratory bench for varying time periods. Shaded areas and error bars represent standard deviations (N = 4 chemiresistors); all data were collected in air at room temperature.

Figure 6

(a) Averaged conductance trace of SWCNT-P4VP-Pt-POM3 prepared with an aged (3 month) sample of POM3 in response to 10 ppm of H₂S. (b) Averaged conductance trace of an aged SWCNT-P4VP-Pt-POM3 device (3 month) in response to 10 ppm of H₂S. Shaded areas represent standard deviations (N = 4 chemiresistors); all data were collected in air at room temperature.

Figure 7

Kapton functionalization and chemiresistor fabrication on a flexible substrate. Possible POM isomers and H⁺ equivalents are omitted for clarity.