An in situ, reversible fluorescent paper sensor for selective detection of ambient CO₂

Abstract

Analyzing the components and contents of environmental gases is of great importance for industrial production, human health, and environmental safety. Carbon dioxide (CO₂) has drawn great concern for its adverse effects on climate change and physical health. Simple, in situ, and selective methods for CO₂ detection are thus highly demanded. This work presents a versatile and easily fabricated paper sensor for the selective detection of ambient CO₂. The sensor consists of two components with a fluorescent molecule ANT-PPh₃ as the sensing element and an ionic liquid [DBUH]⁺[Im]⁻ for the absorption of CO₂. Based on the photo-induced electron transfer (PET) between ANT-PPh₃ and [DBUH]⁺[Im]⁻, the sensor exhibited fluorescence enhancement in the presence of CO₂, and the response process is reversible. The sensor exhibits improved selectivity for CO₂ over SO₂ and NO₂ compared with most reported works, with good stability and fast response to human breath. Without complex instrumentation and gas pre-collection processes, the early notification of high indoor CO₂ concentrations was realized using this sensor. Taking advantage of its intuitive fluorescence changes, specific responsiveness, and convenience, we believe our designed CO₂ sensor could be applied practically in monitoring and indicating high indoor CO₂ concentrations in living scenarios.

---

1. Introduction

The components and contents of the surrounding air are related to many aspects of industrial production, human health, and environmental safety.^1,2 There is a growing demand for sensors that can detect and monitor various gases in both environmental and biological systems. Their applications include monitoring air quality, detecting hazardous gases, and clinical diagnosis. The target gases vary between volatile organic compounds³ (VOCs), explosive gases,⁴ highly-toxic gases,⁵ and the other common gases,^6,7 depending on different scenarios. The analytical methods behind these sensors can be primarily divided into mass spectrometry,⁸ electrochemical methods,⁹ IR absorption,¹⁰ and colorimetric/fluorescent methods.¹¹ Generally, mass spectrometry is often the golden standard for detecting most gases, but it requires expensive instrumentation and complicated operations.¹² Electrochemical gas sensors can reach an ultra-high sensitivity. However, they usually need high operating temperatures, causing energy consumption, and are unsuitable for in situ monitoring. Sensors based on IR absorption are also bothered by poor selectivity and limited sensitivity. By contrast, fluorescent gas sensors rely on highly selective reactions between target gases and sensors, which are especially appreciated for their versatility, low cost, good selectivity, and in situ sensing abilities.^13,14

Recently, growing concerns over global warming have generated widespread discussion.¹⁵ As one of the most important greenhouse gases, carbon dioxide (CO₂) contributes immeasurably to worldwide climate change. Excessive CO₂ emissions can bring significant environmental problems to our society.¹⁶ Additionally, high concentrations of CO₂ can cause physical effects. When the indoor concentration of CO₂ rises to 1500 ppm, long-time exposure can lead to dizziness and a lack of focus. Furthermore, several physiological dysfunctions and irreversible damage may occur after a long-time exposure when CO₂ concentration is above 3000 ppm. Barely able to be perceived, people tend to ignore this potential danger.^1,17,18 Thus, in situ monitoring of indoor concentrations of CO₂ is essential in confined or densely populated spaces like mines or classrooms. In this perspective, fluorescent CO₂ sensors become advantageous for their selective sensing and convenience.

Nevertheless, previously reported CO₂ sensors based on fluorescent methods are mainly performed in solution,^19–26 which is inconvenient for practical use (Table S1, ESI†). Jiang et al. developed a quantitative CO₂ sensor based on AIE-active organogels.²⁷ However, the detection setup required the presence of volatile amines, which are harmful to the human body. Meldrum et al. developed a paper sensor for CO₂ based on a pH-sensitive fluorescent dye.²⁸ The formation of carbonic acid in the presence of CO₂ and water vapor would trigger ratiometric fluorescence changes. Despite the high sensitivity, the sensor lacked selectivity over SO₂ and NO₂, while the detection required high humidity. Overall, existing solid-state fluorescent sensors for CO₂ usually require pre-collection processes or specific detecting environments. There is rare in situ solid-state fluorescent sensor that can detect ambient CO₂ concentrations. Portable, selective and robust sensors for ambient CO₂ detection are still highly demanded.

Ionic liquids (ILs) are a versatile class of functional materials with various physicochemical properties.²⁹ They can offer inherent advantages in terms of facile synthesis, low cost, and easy tunability, making them promising in the field of optical sensing.³⁰ IL-based fluorescent sensors were reported to detect H₂O₂,³¹ SO₂,³² and halogen anions.³³ Combined with polymer membranes, ILs can serve as solid-state sensors for in situ, simple, and handy detection.³⁴ More importantly, ILs have been successfully utilized as highly efficient CO₂ absorbents due to their intriguing reaction performance, such as efficient binding, good stability, and rapid response between ILs and CO₂.^35,36 These good metrics have made ILs competitive candidates for manufacturing fluorescent CO₂ sensors.

Herein, we report an in situ, reversible fluorescent paper sensor to realize ambient CO₂ sensing based on a PET (photo-induced electron transfer) process using an ionic liquid ([DBUH]⁺[Im]⁻) and a functional fluorescent molecule (ANT-PPh₃). Due to the interactions between triphenylphosphine cation moieties and anions like F⁻, ANT-PPh₃ was previously reported to detect anions in an aqueous solution.³⁷ We combined ANT-PPh₃ with an amine-based ionic liquid, [DBUH]⁺[Im]⁻,³⁸ to construct a paper-based fluorescent sensor for CO₂ detection. Here, [DBUH]⁺[Im]⁻ functions as a reversible CO₂ absorbent. After exposure to CO₂, the imidazole anion [Im]⁻ of [DBUH]⁺[Im]⁻ would transfer to carbamate species [ImCOO]⁻, which subsequently weakens the PET process from [Im]⁻ to ANT-PPh₃, and finally leads to fluorescence enhancement. The proposed mechanism is verified through fluorescence spectra, ³¹P-NMR spectra, and DFT calculations. Since [DBUH]⁺ [Im]⁻ can release the adsorbed CO₂ at higher temperatures, this sensor is reusable after heating at 60 °C for 30 min. The sensor shows improved selectivity over H₂O, SO₂, and NO₂ compared to previously reported CO₂ sensors, with good stability and fast response to human breath. We successfully performed in situ notification of high concentrations of indoor CO₂, showing its convenience and practicality.

2. Experimental section

2.1 Synthesis and characterization of ANT-PPh₃

ANT-PPh₃ was synthesized following the steps shown in Scheme 1(A).³⁷ The final product was a yellow solid, collected by filtration with a total yield of 64%. ¹H-NMR (400 MHz, DMSO-d₆): δ 7.97 (dd, J = 6.8, 3.2 Hz, 4H), 7.82–7.73 (m, 6H), 7.60–7.42 (m, 24H), 6.98 (dd, J = 6.9, 3.0 Hz, 4H), 6.12 (d, J = 13.8 Hz, 4H). ¹³C-NMR (101 MHz, DMSO-d₆): δ 135.582, 134.896, 134.845, 134.795, 130.939, 130.388, 130.326, 130.265, 130.179, 125.956, 125.560, 122.348, 118.221, 117.364, 24.310, 23.819. ESI-MS (m/z): calcd for C₅₂H₄₂F₁₂P₄ [M-PF₆]⁺: 873.2398, found: 873.2351.

Scheme 1 (A) Synthetic route of ANT-PPh₃; (B) synthetic route of [DBUH]⁺[Im]⁻ and its absorption of CO₂; (C) schematic illustrations for the fabrication of the paper sensor for CO₂.

2.2 Synthesis and characterization of [DBUH]⁺[Im]⁻

The ionic liquid [DBUH]⁺[Im]⁻ was prepared through neutralization between the super base DBU and imidazole shown in Scheme 1(B).³⁸ Imidazole was dissolved in dichloromethane. Then equivalent DBU was added dropwise under vigorous stirring. After 12 h, the solvent was removed by a rotary evaporator. The remaining ionic liquid was dried at 60 °C for 24 h in a vacuum drying oven. ¹H-NMR (400 MHz, chloroform-d): δ 7.62 (s, 1H), 7.05 (s, 2H), 3.22 (ddd, J = 20.5, 10.2, 5.1 Hz, 7H), 2.37 (s, 2H), 1.78 (p, J = 6.0 Hz, 2H), 1.59 (d, J = 29.5 Hz, 7H). ¹³C-NMR (101 MHz, chloroform-d): δ 162.102, 135.389, 121.924, 77.162, 52.930, 48.426, 43.658, 36.915, 29.755, 28.505, 25.969, 22.442.

2.3 Preparation of CO₂ paper sensor

The CO₂ paper sensor was prepared as shown in Scheme 1(C). The detailed preparation process is as follows. Ethyl cellulose (EC, 5g) was dissolved in 100 mL ethanol and stirred for 3 h at 75 °C. Then 500 μL [DBUH]⁺[Im]⁻ and 1.0 mg ANT-PPh₃ were added into 10 mL ethyl cellulose solution. After 0.5 h stirring, the mixture was transferred into a petri dish (5 mL) and kept in a vacuum drying oven at 60 °C for 4 h to prepare the required paper sensor. The original sensor was cut into rectangles (1.3 × 2.5 cm) for the subsequent measurements.

2.4 Adsorption–desorption cycles

The CO₂ paper sensor was first exposed to gaseous CO₂ using a balloon filled with high-purity CO₂. Then, the sensor was kept in a vacuum drying oven at 60 °C for 30 min to facilitate CO₂ desorption. The adsorption–desorption process was repeated, and the fluorescence spectra were recorded accordingly.

2.5 CO₂ detection

For measurement of the limit of detection (LOD), a CO₂ gas sensing apparatus was constructed, as shown in Scheme S1A (ESI†). Argon (Ar, purity 99.999%) was used as the carrier gas, and the purity of gaseous CO₂ was 99.999%. The flow rate of Ar was fixed at 400 mL min⁻¹. The concentration of CO₂ was controlled by changing the flow rates of CO₂, which varied from 2 to 12 mL min⁻¹. Mixed gas was delivered into a glass reactor. After 25 min, the sensor was taken out, and the fluorescence spectrum was recorded.

For selective experiments, the sensor was placed inside a dish which was sealed by a rubber plug. Gaseous N₂, Ar, O₂, CO₂ (>99 vol%), and SO₂/NO₂ (2 vol%) were continuously blown into the sealed dish through balloons. After 15 min, the fluorescence spectrum was recorded. Other experiments were performed in sealed glass bottles. 1 mL of different organic solvents or aqueous solutions were placed inside the bottle in advance. Then the sensor was placed, followed by quick sealing. After 15 min, the sensor was taken out, and the fluorescence spectrum was recorded. Lower concentrations of SO₂ were prepared by adding H₂SO₄ aqueous solution (1.0 M) into NaHSO₃ aqueous solution. Lower concentrations of NO₂ were prepared by adding H₂SO₄ aqueous solution (1.0 M) into NaNO₂ aqueous solution.³⁹ Gaseous NH₃ was prepared from concentrated ammonia (12.0 M).

For the simulation of high concentrations of indoor CO₂, NaHCO₃ aqueous solutions of different concentrations as the origin of gaseous CO₂ were used. In a typical experiment, 30 mL (0.15 M) NaHCO₃ aqueous solution was placed in a sealed glass cover. A commercial CO₂ detector was applied to indicate the average CO₂ concentration inside (Scheme S1B, ESI†). After 4 h, the sensor was taken out, and the fluorescence spectra were recorded.

3. Results and discussion

As shown in Scheme 1(A), the fluorescent functional molecule ANT-PPh₃ was synthesized and characterized through ¹H-NMR, ¹³C-NMR, and ESI-MS (Fig. S1–S3, ESI†).³⁷ As shown in Fig. 1 and Fig. S6 (ESI†), ANT-PPh₃ is strongly fluorescent with bright blue emission peaked at 430 nm and shows two absorption maxima at 268 nm and 390 nm. The ionic liquid molecule [DBUH]⁺[Im]⁻ was synthesized through the neutralization reaction between DBU and imidazole⁴⁰ and was characterized through ¹H-NMR and ¹³C-NMR (Fig. S4 and S5, ESI†). As illustrated in Scheme 1(B), [DBUH]⁺[Im]⁻ could reversibly absorb CO₂, leading to the formation of the corresponding carbamate.³⁸ To confirm the absorption of CO₂ by [DBUH]⁺[Im]⁻, we measured its ¹³C-NMR and IR spectra. According to the IR spectra of [DBUH]⁺[Im]⁻, two new bands (1709 cm⁻¹, 1281 cm⁻¹) appeared after bubbling CO₂ for 1 min (Fig. S7, ESI†), which is assigned to the stretching vibrations of C = O and the C–N bonds of the carbamate.⁴¹¹³C-NMR also shows a new peak of carbonyl carbon at 163.89 ppm (Fig. S8, ESI†). These facts confirmed the CO₂ absorption ability of [DBUH]⁺[Im]⁻.

Fig. 1 (A) Photographs of ANT-PPh₃, ANT-PPh₃ + [DBUH]⁺ [Im]⁻, ANT-PPh₃ + [DBUH]⁺[Im]⁻ after bubbling CO₂ for 1 min in DMSO. Fluorescent pictures were taken under 254 nm UV light. (B) The excitation spectrum of ANT-PPh₃ and fluorescence spectra of the same set of samples in (A). (C) Fluorescence decay curves of the same set of samples in (A). [ANT-PPh₃] = 20 μM, V_{[DBUH]⁺[Im]⁻} = 50 μL, V_DMSO = 1 mL. IL = [DBUH]⁺[Im]⁻.

Continuously, we investigated the optical behavior of the ANT-PPh₃ + [DBUH]⁺[Im]⁻ system upon CO₂ stimulation. As is shown in Fig. 1(A), in the presence of [DBUH]⁺[Im]⁻, the fluorescence of ANT-PPh₃ is significantly quenched. Through bubbling CO₂ for 1 min, the blue emission gradually recovers. At the same time, the quantum yield of ANT-PPh₃ decreased from 71.8% to 6.3% with [DBUH]⁺[Im]⁻, then recovered to 11.4% in the presence of CO₂. The fluorescence lifetime shares the same trend, varying from 7.8 ns to 4.0 ns, then to 8.5 ns (Fig. 1(C) and Table S2, ESI†). These changes in quantum yields and fluorescence lifetimes indicate the occurrence of fluorescence quenching based on PET processes as expected.⁴²

To further validate the sensing mechanism and the interactions among ANT-PPh₃, [DBUH]⁺[Im]⁻, and CO₂, we investigated ³¹P-NMR spectra. As is shown in Fig. 2(A), the original chemical shift of ANT-PPh₃ lies at 21.20 ppm. Upon adding excess [DBUH]⁺[Im]⁻, this peak shifts to 26.55 ppm, and a new peak appears at −6.18 ppm. After bubbling CO₂ for 1 min, the original peak regenerates at 21.20 ppm, along with the partial recovery of the blue emission. Besides, in the presence of imidazole (Im) alone, no evident chemical shift of ANT-PPh₃ was observed. If imidazole was converted to its anion ([Im]⁻) with equivalent NaOH in advance, the ³¹P-NMR spectrum was nearly identical to the situation of adding [DBUH]⁺[Im]⁻ (Fig. S9, ESI†). Therefore, we assumed that interactions between [Im⁻] and ANT-PPh₃ caused these changes in ³¹P-NMR spectra. Next, we conducted theoretical DFT calculations with the B3LYP-d3 function and 6-31g* basis sets. The DFT-optimized structures are shown in Fig. 2(B). The yellow dashed lines show the interactions between [Im⁻] and the P atom of ANT-PPh₃, with a distance comparable to common hydrogen bonds (2.01 Å). After bubbling CO₂, the newly-formed carbamate ([ImCOO]⁻) only shows limited interactions with the P atom. Instead, O⋯C–H interactions are formed between [ImCOO]⁻ and ANT-PPh₃. These differences in the binding modes may lead to changes in the ³¹P-NMR spectra.

Fig. 2 (A) ³¹P-NMR spectra of ANT-PPh₃, ANT-PPh₃ + [DBUH]⁺[Im]⁻, ANT-PPh₃ + [DBUH]⁺ [Im]⁻ after bubbling CO₂ for 1 min in DMSO-d₆. (B) DFT optimized structures of ANT-PPh₃ in the presence of [Im]⁻ and [ImCOO]⁻. The yellow dashed lines represent N–P interactions. The red dashed lines represent O–H interactions. (C) Proposed mechanism of the CO₂-induced fluorescence enhancement of ANT-PPh₃ + [DBUH]⁺[Im]⁻ system.

The frontier molecular orbital investigations of ANT-PPh₃, ANT-PPh₃ + [Im]⁻, and ANT-PPh₃ + [ImCOO]⁻ adducts were also performed, and the results are displayed in Fig. S10 (ESI†). The highest occupied molecular orbitals (HOMOs) are located on the anthracene unit in ANT-PPh₃, whereas in the presence of [Im]⁻ and [ImCOO]⁻, the HOMOs switch to benzene and imidazole units. The corresponding bandgaps decrease from 3.29 eV to 3.12 eV and 2.83 eV, respectively. The higher HOMOs of ANT-PPh₃ + [Im]⁻ adduct and ANT-PPh₃ + [ImCOO]⁻ adduct thus facilitate the occurrence of PET processes.^43–45 The proposed mechanism is shown in Fig. 2(C). The PET process caused by [Im]⁻ is stronger than [ImCOO]⁻, leading to the transformation from weak fluorescence to medium fluorescence, namely CO₂-induced fluorescence enhancement of the ANT-PPh₃ + [DBUH]⁺[Im]⁻ system.

By changing the amount of [DBUH]⁺[Im]⁻, we investigated the response of the ANT-PPh₃ + [DBUH]⁺[Im]⁻ system toward a certain amount of CO₂ in DMSO. The fluorescence enhancement ratio ((I_t − I₀)/I₀) at 430 nm of the solution firstly increases and then decreases with adding more [DBUH]⁺[Im]⁻ (Fig. S11, ESI†). We determined the optimal volume of [DBUH]⁺[Im]⁻ in the mixed solution. As shown in Fig. 3 and Fig. S12 (ESI†), after bubbling different volumes of high-purity gaseous CO₂ through a syringe, the absorbance of the solution at 268 nm slowly decreased, while the fluorescence intensity at 430 nm dramatically increased. There is a good linear relationship (R² = 0.992) between the fluorescence intensities and the amount of CO₂ at lower levels, indicating that this system may have great potential to be a CO₂ turn-on sensor.

Fig. 3 (A) Fluorescence spectra of ANT-PPh₃ + [DBUH]⁺[Im]⁻ in DMSO solution after the injection of different amounts of high-purity gaseous CO₂. (B) Fluorescence intensities at 430 nm after the injection of different amounts of high-purity gaseous CO₂. [ANT-PPh₃] = 20 μM, V_{[DBUH]⁺[Im]⁻} = 50 μL.

Taking advantage of the above characteristics, we fabricated paper sensors using ethyl cellulose as the substrate. For the stability test, we put the sensor in a sealed dish under ordinary room light for 7 days (Fig. S13, ESI†). The fluorescence intensities at 430 nm are nearly unchanged. In addition, photochemical stability is one of the key features determining the detecting performance of a fluorescent sensor. Under continuous excitation of white LED and 254 nm UV, the photostability of the paper sensor was investigated. As is shown in Fig. S14 (ESI†), under 254 nm UV, the fluorescence intensity quenches only 20% after 8 h. And with continuous exposure to white LED, the fluorescence intensity remains the same after 12 h. These data demonstrate that the sensor has good stability to support its daily application. After 1 min exposure to high-purity gaseous CO₂, the quantum yield of this paper sensor increased from 17.0% to 33.6%, and the fluorescence lifetime increased from 6.9 ns to 10 ns (Fig. S15, ESI†). Due to the reversible reaction between [DBUH]⁺[Im]⁻ and CO₂, this paper sensor also shows a reversible response to CO₂. As shown in Fig. 4(A), the sensor becomes strongly fluorescent by blowing high-purity CO₂ into a sealed quartz dish. Then, after being heated at 60 °C for 30 min, the sensor becomes weakly fluorescent again. This reuse cycle could repeat over eight times.

Fig. 4 (A) Fluorescence responses of the adsorption–desorption cycles for CO₂ paper sensor upon 5 min CO₂ treatment and 30 min heating alternatively. Inset: Fluorescence pictures of a typical adsorption–desorption cycle. (B) Fluorescence enhancement ratios ((I_t − I₀)/I₀) at 430 nm of CO₂ paper sensor after 15 min exposure to different gases (O₂, N₂, Ar, CO₂: >99 vol%; SO₂, NO₂: 2 vol%) or saturated vapors. ((CH₃)₂CO: acetone; CH₃COOC₂H₅: ethyl acetate; CH₃CN: acetonitrile; C₆H₅CH₃: toluene; PE: petroleum ether; CH₂Cl₂: dichloromethane; i-C₃H₇OH: isopropanol) Inset: Fluorescence pictures of CO₂ paper sensors after 15 min exposure to different gases. Pictures were taken under 254 nm UV.

The specificity of CO₂ sensors has become a long-standing concern. Based on PET-quenching or pH sensing mechanisms, previously reported CO₂ sensors are often affected by acidic gases like SO₂ and NO₂. To demonstrate the selectivity, we investigated the response of the CO₂ paper sensor toward other common gases (H₂O, N₂, O₂) and some VOCs. As is shown in Fig. 4(B), the fluorescence intensities of the sensor remain nearly unchanged after adding the interfering gases for 15 min, indicating that they exert a negligible effect on the detection of CO₂. A noticeable fluorescence enhancement occurs in the presence of CO₂, which can be observed by the naked eye. Especially, SO₂ and NO₂ cause evident fluorescence quenching, which is entirely different from CO₂. To further check the fluorescence quenching caused by SO₂ and NO₂, excessive SO₂ and NO₂ gases were respectively injected into ANT-PPh₃ + [DBUH]⁺[Im]⁻ in DMSO, and the solutions were nearly non-fluorescent (Fig. S17, ESI†). As is shown in Fig. S18 (ESI†), after bubbling SO₂ or NO₂, the original chemical shift of ANT-PPh₃ at around 21 ppm does not regenerate, accounting for the fluorescence quenching. As both acidic gases, SO₂ and NO₂ can also react with [DBUH]⁺[Im]⁻, forming addictives like [ImSOO]⁻ and [ImNOO]⁻.^41,46 Unlike [ImCOO]⁻, [ImSOO]⁻ and [ImNOO]⁻ species may exhibit a more intensive PET effect than [Im]⁻, leading to different fluorescence responses amongst CO₂ and NO₂/SO₂.

The potential competing effects of NO₂ and SO₂ were also investigated. After exposure to SO₂ and NO₂ (200 ppm) for 3 h, the fluorescence spectra of the sensor are nearly identical, and no interference is observed when detecting ambient CO₂ (Fig. S19, ESI†). In living environments, the concentrations of NO₂ and SO₂ are minimal (<1–2 ppm).^32,47 Thus, the selective detection of ambient CO₂ over SO₂ and NO₂ can be realized. Compared with other reported sensors, this CO₂ paper sensor possesses excellent analytical capability for the selective detection of CO₂. Additionally, this sensor could respond to CO₂ in the exhaled air (Fig. 5(B) and Fig. S20, ESI†). The paper sensor was cut into the shape of letters “T”, “H”, and “U”. After continuous exposure to human breath for 40 s, the letters became strongly emissive. This phenomenon can be restored repeatedly by heating for half an hour. Human breath consists of a mixture of CO₂ (∼5%), water vapor (∼5%), N₂ (∼74%), O₂ (∼15%), trace gases, and various VOCs.²⁸ These results demonstrate that this paper sensor can detect CO₂ in complex atmospheres.

Fig. 5 (A) Schematic illustration of potential applications of CO₂ paper sensor. (B) Human breath response cycles of CO₂ paper sensor. (C) Fluorescence enhancement ratios ((I_t − I₀)/I₀) at 430 nm as a function of the CO₂ concentrations (vol%). (D) Fluorescence enhancement ratios ((I_t − I₀)/I₀) at 430 nm under different concentrations of ambient CO₂. Inset: Fluorescence pictures of CO₂ paper sensors under different concentrations of ambient CO₂. Pictures were taken under 254 nm UV.

To find the LOD, we use the experimental setup shown in Scheme S1A (ESI†). The flow rate of high-purity Ar was fixed at 400 mL min⁻¹. By adjusting the flow rates of CO₂, we can precisely control the concentrations of CO₂ in the gas mixture with the mass flow meter. We then investigated the fluorescence spectra of the paper sensor in the gas reactor after 25 min exposure to 0.5–3.0 vol% CO₂. The standard deviation of the noise level was obtained by continuously blowing high-purity Ar into the same reactor. The fluorescence enhancement ratios ((I_t − I₀)/I₀) at 430 nm gradually increased with CO₂ concentrations rising. An excellent linear calibration line (R² = 0.996) was attained, and according to LOD = 3S_b × m⁻¹, where S_b is the standard deviation of fluorescence enhancement ratios ((I_t − I₀)/I₀) for 5 blank samples, m is the slope of the linear correlation between (I_t − I₀)/I₀ and CO₂ concentrations, the LOD was calculated to be 0.183 vol%.

In order to verify the potential of this paper sensor for detecting ambient CO₂ in real scenarios, we used sealed glass covers and NaHCO₃ aqueous solutions to simulate indoor environments with high CO₂ concentrations (Scheme S1B, ESI†). We used a commercial CO₂ detector to determine the CO₂ concentrations inside the glass cover. As is shown in Fig. S21 (ESI†), the fluorescence intensities of CO₂ paper sensors increase over time. According to the fluorescence enhancement ratios ((I_t − I₀)/I₀) at 430 nm, we can determine and monitor higher levels of indoor CO₂ concentrations (Fig. 5(D)). The recommended indoor CO₂ concentration should be less than 1500 ppm.⁴⁸ With this paper sensor, indoor CO₂ concentrations above this level can be visualized and notified. These results support the application of this sensor in living scenarios, which may be very helpful in ensuring our physical health.

4. Conclusion

In summary, we have rationally designed an in situ, reversible fluorescent paper sensor for the selective detection of ambient CO₂ based on a novel fluorescent sensing system. Composed of a fluorescent molecule ANT-PPh₃ and a CO₂ absorbing element, the ionic liquid [DBUH]⁺[Im]⁻, the mixture of ANT-PPh₃ + [DBUH]⁺[Im]⁻ is fabricated into a paper sensor with ethyl cellulose as the substrate. Encountering [DBUH]⁺[Im]⁻, the fluorescence of ANT-PPh₃ was quenched owing to the intense PET process. After exposure to CO₂, the formation of [ImCOO]⁻ weakened the PET process and finally caused fluorescence enhancement. This proposed sensing mechanism, confirmed by ³¹P-NMR spectra and DFT calculations, endows the sensor with excellent selectivity over SO₂, NO₂, H₂O, and other VOCs. Due to the reversible reaction between [DBUH]⁺[Im]⁻ and CO₂, the sensor can be regenerated after heating and then reused up to eight times. We successfully conducted simulation experiments of early notification of ambient CO₂ using this sensor. Compared with the previously reported CO₂ sensors, this sensor exhibits improved selectivity and does not require particular detection environments or gas pre-collection processes. This PET-based strategy may promote the design of new approaches to developing more fluorescent systems with unique selectivity. We believe our designed CO₂ sensor could be applied practically in monitoring high concentrations of CO₂ for daily healthcare.

Author contributions

Chu Zhang: investigation, conceptualization, methodology, validation, formal analysis, data curation, writing-original draft, writing-review & editing, visualization, Yiwen Ding: resources, formal analysis, Min Zhou: resources, formal analysis, Yu Xiang: supervision, methodology, Aijun Tong: supervision, funding acquisition, project administration, conceptualization, writing-review & editing. All authors discussed the results and commented on the manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Key R&D Program of China (2021YFA1200901) and the National Natural Science Foundation of China (21974076, 21621003). We sincerely appreciate Mr Qingyu Meng (Department of Chemistry, Tsinghua University) for his valuable discussion on the NMR study and contributions to DFT calculations. We thank Mr Xin Tan (Department of Chemistry, Tsinghua University) for his kind help in the construction of gas detecting equipment.

Notes and references

1. M. Tainio, Z. Jovanovic Andersen, M. J. Nieuwenhuijsen, L. Hu, A. de Nazelle, R. An, L. M. T. Garcia, S. Goenka, B. Zapata-Diomedi, F. Bull and T. H. Sá, Environ. Int., 2021, 147, 105954.
2. M. R. Heal, P. Kumar and R. M. Harrison, Chem. Soc. Rev., 2012, 41, 6606–6630.
3. C. Zhang, Y. Zheng, Y. Ding, X. Zheng, Y. Xiang and A. Tong, Talanta, 2022, 236, 122845.
4. G. Wang, M. Li, Q. Wei, Y. Xiong, J. Li, Z. Li, J. Tang, F. Wei and H. Tu, ACS Sens., 2021, 6, 1849–1856.
5. P. Zheng, W. Cao, Y. Zhang, F. Li and M. Zhang, ACS Sens., 2022, 7, 1946–1957.
6. L. Di, Z. Xia, H. Wang, Y. Xing and Z. Yang, Sens. Actuators, B, 2021, 326, 128987.
7. Y.-J. Gao, G. Romolini, H. Huang, H. Jin, R. A. Saha, B. Ghosh, M. De Ras, C. Wang, J. A. Steele, E. Debroye, J. Hofkens and M. B. J. Roeffaers, J. Mater. Chem. C, 2022, 10, 12191–12196.
8. C. Wang, A. Mbi and M. Shepherd, IEEE Sens. J., 2010, 10, 54–63.
9. S. M. Majhi, A. Mirzaei, H. W. Kim, S. S. Kim and T. W. Kim, Nano Energy, 2021, 79, 105369.
10. L. Ciaffoni, G. Hancock, J. J. Harrison, J.-P. H. van Helden, C. E. Langley, R. Peverall, G. A. D. Ritchie and S. Wood, Anal. Chem., 2013, 85, 846–850.
11. O. Green, P. Finkelstein, M. A. Rivero-Crespo, M. D. R. Lutz, M. K. Bogdos, M. Burger, J.-C. Leroux and B. Morandi, J. Am. Chem. Soc., 2022, 8717–8724.
12. X. Zhou, S. Lee, Z. Xu and J. Yoon, Chem. Rev., 2015, 115, 7944–8000.
13. Y. Zhang, S. Yuan, G. Day, X. Wang, X. Yang and H.-C. Zhou, Coord. Chem. Rev., 2018, 354, 28–45.
14. W.-J. Yoon, S. Yang, J. Jang, M. Oh, M. Rim, H. Ko, J. Koo, S.-I. Lim, Y.-J. Choi and K.-U. Jeong, J. Mater. Chem. C, 2022, 10, 11316–11322.
15. D. Leaf, H. J. H. Verolme and W. F. Hunt, Environ. Int., 2003, 29, 303–310.
16. T. Gasser, C. Guivarch, K. Tachiiri, C. D. Jones and P. Ciais, Nat. Commun., 2015, 6, 7958.
17. U. Satish, M. J. Mendell, K. Shekhar, T. Hotchi, D. Sullivan, S. Streufert and W. J. Fisk, Environ. Health Perspect., 2012, 120, 1671–1677.
18. L. de Lary, A. Loschetter, O. Bouc, J. Rohmer and C. M. Oldenburg, Int. J. Greenhouse Gas Control, 2012, 9, 322–333.
19. Z. Guo, N. R. Song, J. H. Moon, M. Kim, E. J. Jun, J. Choi, J. Y. Lee, C. W. Bielawski, J. L. Sessler and J. Yoon, J. Am. Chem. Soc., 2012, 134, 17846–17849.
20. H. Yuan, Y. Fan, C. Xing, R. Niu, R. Chai, Y. Zhan, J. Qi, H. An and J. Xu, Anal. Chem., 2016, 88, 6593–6597.
21. H. Mardani, H. Roghani-Mamaqani, S. Shahi and M. Salami-Kalajahi, ACS Appl. Polym. Mater., 2022, 4, 1816–1825.
22. S. Kang, J. Kim, J.-H. Park, C. K. Ahn, C.-H. Rhee and M. S. Han, Dyes Pigm., 2015, 123, 125–131.
23. Y.-J. Jin, B.-C. Moon and G. Kwak, Dyes Pigm., 2016, 132, 270–273.
24. Y. Liu, Y. Tang, N. N. Barashkov, I. S. Irgibaeva, J. W. Y. Lam, R. Hu, D. Birimzhanova, Y. Yu and B. Z. Tang, J. Am. Chem. Soc., 2010, 132, 13951–13953.
25. Y. Ma and L.-Y. L. Yung, Anal. Chem., 2014, 86, 2429–2435.
26. H. Wang, D. Chen, Y. Zhang, P. Liu, J. Shi, X. Feng, B. Tong and Y. Dong, J. Mater. Chem. C, 2015, 3, 7621–7626.
27. Y. Ma, M. Cametti, Z. Džolić and S. Jiang, J. Mater. Chem. C, 2018, 6, 9232–9237.
28. H. Wang, S. I. Vagin, B. Rieger and A. Meldrum, ACS Appl. Mater. Interfaces, 2020, 12, 20507–20513.
29. P. Goodrich, H. Q. N. Gunaratne, J. Jacquemin, L. Jin, Y. Lei and K. R. Seddon, ACS Sustainable Chem. Eng., 2017, 5, 5635–5641.
30. S. Zeng, X. Zhang, L. Bai, X. Zhang, H. Wang, J. Wang, D. Bao, M. Li, X. Liu and S. Zhang, Chem. Rev., 2017, 117, 9625–9673.
31. Q.-H. Zhu, W.-L. Yuan, L. Zhang, G.-H. Zhang, L. He and G.-H. Tao, Anal. Chem., 2019, 91, 6593–6599.
32. S. Che, Q. Shou, Y. Fan, X. Peng, C. Zhou, H. Fu and Y. She, ACS Sustainable Chem. Eng., 2022, 10, 2784–2792.
33. T. Mizuta, K. Sueyoshi, T. Endo and H. Hisamoto, Anal. Chem., 2021, 93, 4143–4148.
34. Z. Dai, R. D. Noble, D. L. Gin, X. Zhang and L. Deng, J. Membr. Sci., 2016, 497, 1–20.
35. G. Cui, J. Wang and S. Zhang, Chem. Soc. Rev., 2016, 45, 4307–4339.
36. C. Wang, X. Luo, H. Luo, D.-E. Jiang, H. Li and S. Dai, Angew. Chem., Int. Ed., 2011, 50, 4918–4922.
37. W. Huang, H. Lin and H. Lin, Sens. Actuators, B, 2011, 153, 404–408.
38. X. Zhu, M. Song and Y. Xu, ACS Sustainable Chem. Eng., 2017, 5, 8192–8198.
39. H. Yu, F. Dong, B. Li and F. S. Riehle, Sens. Actuators, B, 2019, 299, 126983.
40. F.-F. Chen, K. Huang, Y. Zhou, Z.-Q. Tian, X. Zhu, D.-J. Tao, D.-E. Jiang and S. Dai, Angew. Chem., Int. Ed., 2016, 55, 7166–7170.
41. S. F. R. Taylor, M. McClung, C. McReynolds, H. Daly, A. J. Greer, J. Jacquemin and C. Hardacre, Ind. Eng. Chem. Res., 2018, 57, 17033–17042.
42. K. Xu, J. Zhao, D. Escudero, Z. Mahmood and D. Jacquemin, J. Phys. Chem. C, 2015, 119, 23801–23812.
43. W. Sun, M. Li, J. Fan and X. Peng, Acc. Chem. Res., 2019, 52, 2818–2831.
44. D. Escudero, Acc. Chem. Res., 2016, 49, 1816–1824.
45. J. Harathi, R. Selva Kumar, S. K. Ashok Kumar, D. Saravanakumar, S. Senthilkumar and K. Thenmozhi, J. Mater. Chem. C, 2022, 10, 7949–7961.
46. A. J. Greer, S. F. R. Taylor, H. Daly, M. G. Quesne, N. H. de Leeuw, C. R. A. Catlow, J. Jacquemin and C. Hardacre, ACS Sustainable Chem. Eng., 2021, 9, 7578–7586.
47. H. Long, A. Harley-Trochimczyk, T. Pham, Z. Tang, T. Shi, A. Zettl, C. Carraro, M. A. Worsley and R. Maboudian, Adv. Funct. Mater., 2016, 26, 5158–5165.
48. T. Vehviläinen, H. Lindholm, H. Rintamäki, R. Pääkkönen, A. Hirvonen, O. Niemi and J. Vinha, J. Occup. Environ. Hyg., 2016, 13, 19–29.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2tc05116h

---