A molecular recognition platform for the simultaneous sensing of diverse chemical weapons

Chemical warfare agents (CWAs) such as phosgene and nerve agents pose serious threats to our lives and public security, but no tools can simultaneously screen multiple CWAs in seconds. Here, we rationally designed a robust sensing platform based on 8-cyclohexanyldiamino-BODIPY (BODIPY-DCH) to monitor diverse CWAs in different emission channels. Trans-cyclohexanyldiamine as the reactive site provides optimal geometry and high reactivity, allowing trans-BODIPY-DCH to detect CWAs with a quick response and high sensitivity, while cis-BODIPY-DCH has much weaker reactivity to CWAs due to intramolecular H-bonding. Upon reaction with phosgene, trans-BODIPY-DCH was rapidly converted to imidazolone BODIPY (<3 s), triggering green fluorescence with good sensitivity (LOD = 0.52 nM). trans-BODIPY-DCH coupled with nerve agent mimics, affording a blue fluorescent 8-amino-BODIPY tautomer. Furthermore, a portable test kit using trans-BODIPY-DCH displayed an instant response and low detection limits for multiple CWAs. This platform enables rapid and highly sensitive visual screening of various CWAs.


Introduction
Public security is a global concern closely related to our health and safety. Today, public security around the world faces potential risks from COVID-19, AIDS, antimicrobial resistance, and chemical warfare agents (CWAs). CWAs are a class of highly toxic chemicals used as lethal weapons on battleelds such as World War I and World War II, slaughtering millions of soldiers and civilians. 1-3 Moreover, some CWAs have been used by terrorists to create panic attacks over the past few decades, posing a serious threat to public security and human health.
Phosgene, a colorless and toxic gaseous substance, has been used as a CWA in several wars. 1 Phosgene can cause serious lung and respiratory damage within 2 min, and exposure to 90 ppm for more than 30 min is fatal. 2-4 Organophosphorus nerve agents have been identied as the most dangerous CWAs, which can readily bind with acetylcholinesterase to inhibit conduction of nerve impulses, leading to organ failure and death within seconds. 2,3 Acyl chlorides can also cause serious burns to the skin and mucous membranes, leading to dermal and respiratory diseases. 5,6 Nevertheless, phosgene, organophosphorus, and acyl chlorides are indispensable raw materials for industrial production of pesticides, 7 medicines, 8 and dyes and polymers. 9 As such, they pose a serious threat to public security when used by terrorists as CWAs or released in accidental industrial leaks. It is therefore highly desirable to devise a simple and efficient tool to rapidly measure phosgene, nerve agents and acyl chlorides for both public security and manufacturing safety.
Conventional approaches for the detection of phosgene, nerve agents and volatile acyl chlorides rely on stationary equipment such as gas chromatographs 10 and liquid chromatographs. 11 These instruments are less portable and measurements take a long time, making them unsuitable for rapid security screening in subways, airports, train stations and at customs. Since uorescent sensors are powerful optical devices capable of generating uorescent signal changes in host-guest recognition patterns, they have been used to visually determine analytes with high selectivity and sensitivity, low cost, and easy operation. [12][13][14][15][16] Recently, numerous uorescent sensors have been developed that separately detect phosgene or nerve agents using various reaction sites. [17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36] However, their detection limits are unsatisfactory, or the response rate is slow because the reactivity of recognition units such as pyridine, aromatic amine, and hydroxyl groups is low. Moreover, these sensors ignore various toxic acyl chlorides, such as diphosgene and acetyl chloride, which still pose a threat to public security.
Given the wide availability, broad application, and high toxicity of phosgene and acyl chlorides, the development of sensors to thoroughly screen them is urgently needed. Unfortunately, research on uorescent sensors for acyl chlorides remains at the embryonic stage. To date, only two uorescent sensors have been reported for oxalyl chloride, 37,38 but these uorescent sensors can neither detect nerve agents nor other acyl chlorides. Although great efforts have been made to prepare various sensors to individually detect each analyte, the complex procedures result in low detection efficiency and do not meet the need of rapid security screening in subways, airports, train stations, and customs.
To this end, it is highly desirable to design a versatile sensing platform for the detection of phosgene, nerve agents and various acyl chlorides using different emission channels. In this context, we propose a robust detection platform based on 8amino-BODIPY and cyclohexanyldiamine recognition groups to simultaneously sense multiple CWAs with a fast response and high sensitivity. In addition, a portable and easy-to-use test kit was made with a melt-blown nonwoven fabric and trans-BOD-IPY-DCH for the rapid and accurate detection of phosgene, diphosgene, nerve agent mimics and other volatile acyl chlorides.

Results and discussion
Most common acyl chloride and organophosphorus nerve agents have one reactive group in their molecular structure, whereas phosgene has two reactive units. Although o-phenylenediamine has been used as a recognition site to differentiate phosgene from diethyl chlorophosphate (DCP), 33 aromatic amines have poor nucleophilic substitution reactivity towards DCP and acyl chlorides. Therefore, in this study, 1,2-diaminocyclohexane (DCH) has been used in place of an aromatic amine to improve the sensitivity and response rate. Since 1,2diaminocyclohexane has two primary aliphatic NH 2 with high pK a values (pK a1 ¼ 9.60 and pK a2 ¼ 6.21), 39 it exhibits high nucleophilic reactivity towards acyl chlorides, phosgene, and nerve agent mimics ( Fig. S1 and S2 †). 1,2-Diaminocyclohexane exists as two isomers, trans-and cis-structures, which were anchored at the 8-position of BODIPY to form 8-amino-BODIPY derivatives transor cis-BODIPY-DCH with one aliphatic NH 2 and one aromatic NH 2 for differential binding with acyl chlorides and phosgene (Fig. 1). The 8-amino-BODIPY derivatives (trans-or cis-BODIPY-DCH) exist as hemicyanine tautomers with pale blue uorescence. 40,41 In the presence of phosgene, the two adjacent amines in DCH combine with phosgene to form imidazolone. As a result, the BODIPY core no longer exists as a tautomer, and photo-induced electron transfer (PET) is prohibited, resulting in a strong green uorescence along with a noticeable color change. In the presence of acyl chlorides or DCP, the primary aliphatic NH 2 is transformed into amide that inhibits the PET process and leads to a large uorescence enhancement in the blue channel. Therefore, these sensors (trans-, cis-BODIPY-DCH) enable the simultaneous detection of phosgene, DCP and some other acyl chlorides using different emission channels.
trans-BODIPY-DCH and cis-BODIPY-DCH were easily prepared by the nucleophilic substitution of 8-Cl-BODIPY with trans-cyclohexanediamine and cis-cyclohexanediamine at room temperature, respectively, as outlined in Fig. S3. † Their chemical structures and intermediates were characterized by 1 H NMR, 13 C NMR, HRMS, and X-ray diffraction of single crystals, respectively. The detailed experimental procedure and characterization data are provided in the ESI. † The capability of the sensing platform to detect phosgene and acyl chlorides was investigated in chloroform, and a hypotoxic counterpart triphosgene was used to generate phosgene in situ in the presence of triethylamine (TEA). As shown in Fig. 2a, trans-BODIPY-DCH exhibited a strong absorption at 410 nm. Then, upon the addition of (0-5.0 mM)/TEA (100.0 mM), the absorption peak gradually decreased, while a bathochromic peak at 500 nm emerged and progressively increased with an isosbestic point at 435 nm, suggesting that trans-BODIPY-DCH exhibits high reactivity and sensitivity towards phosgene. trans-BODIPY-DCH exhibits a weak uorescence band at 478 nm in CHCl 3 (F f ¼ 0.12) due to photo-electron transfer (PET) from the 1,2-diaminocyclohexane group to the BODIPY core (Fig. 2b). Upon addition of triphosgene/TEA to trans-BODIPY-DCH, a new uorescence band at 524 nm (F f ¼ 0.46) was signicantly enhanced, while a remarkable uorescence variation from faint blue to bright green was observed. The Commission International de L'Eclairage (CIE) chromaticity coordinates ( Fig. 2c and Table S1 †) also exhibited notable changes from blue to green with a coordinate variation from (0.1247, 0.1805) to (0.2046, 0.5793). Accordingly, a linear correlation (R 2 ¼ 0.994) was obtained between the emission intensity (F 524 ) of trans-BODIPY-DCH and the concentration of triphosgene in the range of 0-3.0 mM (Fig. S4 †) with a detection limit of 0.52 nM based on 3s/k. These results demonstrated that trans-BODIPY-DCH could be used as a ratiometric uorescence sensor for phosgene with ultra-sensitivity.
Next, we explored the spectral response of trans-BODIPY-DCH to acetyl chloride. Upon the addition of acetyl chloride, trans-BODIPY-DCH displayed a negligible change in the UV-vis absorption (Fig. 2d), but its initial uorescence band at 478 nm was signicantly enhanced (F f ¼ 0.62). These observations might be attributed to the stronger nucleophilicity of the free aliphatic NH 2 (pK a ¼ 9.61) in DCH over the aromatic amine. An excellent linear relationship (R 2 ¼ 0.994) was established between the emission intensity at 478 nm and the concentration of acetyl chloride over a range from 0-5.0 mM (Fig. S4 †) resulting in an LOD of 0.77 nM by the principle of 3s/k. These results conrmed that trans-BODIPY-DCH could serve as a sensitive uorescence sensor to detect acetyl chloride. We also examined the spectral response of trans-BODIPY-DCH towards DCP and a family of other acyl chlorides including diphosgene, triphosgene, oxalyl chloride and benzoyl chloride. As shown in Fig. S5 and S6, † addition of DCP (5.0 mM) to trans-BODIPY-DCH (5.0 mM) induced a negligible change in UV-vis absorption spectra of the sensor, but triggered obvious uorescence enhancement at 478 nm (F f ¼ 0.31). Likewise, trans-BODIPY-DCH (5.0 mM) displayed a large uorescence enhancement at 478 nm aer the addition of various acyl chlorides (Table S2 †). Therefore, the trans-BODIPY-DCH could serve as a versatile sensor for the high throughput detection of phosgene, DCP and some other acyl chlorides.
A rapid response is critical for the on-site detection of toxic phosgene and acyl chlorides. To examine the sensing rate, the time-dependent uorescence response of trans-BODIPY-DCH (5.0 mM) was recorded aer the addition of analytes (10.0 mM). For phosgene (Fig. S7 †), the uorescence intensity ratio (F 524 / F 478 ) of trans-BODIPY-DCH (5.0 mM) reached a plateau within 3 s. Likewise, the uorescence intensity at 478 nm reached maxima within a few seconds ($3 s) aer addition of each analyte including DCP, acetyl chloride, diphosgene, oxalyl chloride, thionyl chloride, benzoyl chloride and triphosgene ( Fig. S7 and Table S2 †). In addition, a live video and time-lapse uorescence images of trans-BODIPY-DCH (5.0 mM) upon the addition of analytes were recorded ( Fig. 2g and Movie 1 †). When the triphosgene was added, obvious green uorescence emerged in the local area within 0.3 s. As triphosgene diffused and decomposed into phosgene, the green uorescence became much brighter and diffused over the whole solution within 3 s, suggesting that trans-BODIPY-DCH is a rapidly responding sensor for phosgene. Upon the addition of acetyl chloride, the blue uorescence of trans-BODIPY-DCH (5.0 mM) lit up immediately and reached a maximum within 3 s as well ( Fig. 2h and Movie 2 †). Those observations conrmed that trans-BODIPY-DCH could achieve a rapid response toward both phosgene and volatile acyl chlorides.
Next, we evaluated the uorescence response of trans-BOD-IPY-DCH (5.0 mM) towards other species (10.0 mM) including diphosgene, dimethyl methylphosphonate (DMMP), DCP, tosyl chloride (TsCl), POCl 3 , SOCl 2 , HCl, benzoyl chloride (BzCl), oxalyl chloride (OC), acetyl chloride (AC), formaldehyde (FA), methylglyoxal (MGO), acrolein, triphosgene and phosgene. As shown in Fig. 3a-f, phosgene brought about an obvious color and ratiometric uorescence change of trans-BODIPY-DCH solution. By contrast, a signicant uorescence enhancement was observed for diphosgene, DCP and other acyl chlorides. The potential interfering species like HCl, DMMP, formaldehyde, methylglyoxal and acrolein induced negligible color or uorescence changes, conrming that this sensor could avoid interference from HCl, NO, and aldehydes, and achieve accurate and specic detection of acyl chlorides and phosgene. Therefore, trans-BODIPY-DCH is superior to most of the reported sensors for phosgene or DCP. [24][25][26][27][28][29][30] To further check the selectivity of trans-BODIPY-DCH without any false positive uorescence signal for acid and aldehydes, competitive selectivity experiments were performed. Aer addition of a mixed solution of triphosgene and HCl to the trans-BODIPY-DCH solution containing TEA (100 mM), a newly formed emission band at 524 nm was observed, which was assigned to the uorescence band of trans-BODIPY-ICO. The same observation was found when the trans-BODIPY-DCH solution was treated with triphosgene and formaldehyde. When the mixture solution of acetyl chloride and HCl was added to the trans-BODIPY-DCH solution containing TEA (100 mM), the sensor only showed an enhanced uorescence band at 478 nm (Fig. S8 †). Therefore, we can conclude that trans-BODIPY-DCH is a robust sensor for the discriminative detection of acyl chlorides and phosgene.
To validate our design strategy, we prepared a mirror compound, cis-BODIPY-DCH from (1R,2R)-1,2-diaminocyclohexane and 8-Cl-BODIPY. As seen in Fig. S9, † cis-BODIPY-DCH displayed a strong absorption peak at 408 nm, which gradually converted to a bathochromic absorption band at 498 nm when incremental triphosgene (0-40.0 mM) was added. The initial uorescence band at 474 nm gradually decreased, meanwhile a new uorescence band at 522 nm progressively increased until the concentration of triphosgene was 8 equiv. with respect to the sensor, affording an obvious uorescence change from faint blue to green (Fig. S9 †). cis-BODIPY-DCH could also sense acetyl chloride, diphosgene, triphosgene, OC, or SOCl 2 . Upon the addition of 40 mM acetyl chloride, diphosgene, triphosgene, OC, or SOCl 2 , cis-BODIPY-DCH exhibited a large uorescence enhancement at 474 nm ( Fig. S10-S14 †), and the LOD for acetyl chloride was determined to be 6.92 nM (Fig. S15 †). A live video and time-lapse uorescence images were also recorded. The triphosgene gave a bright green color to the cis-BODIPY-DCH solution within 4 s ( Fig. S9 and Movie 3 †), and the uorescence intensity reached a maximum within 4 s aer addition of triphosgene (Fig. S16 †). The solution could also be lit up by acetyl chloride within 4 s (Movie 4 †). These results demonstrated that cis-BODIPY-DCH could also achieve a rapid response towards phosgene and acyl chlorides with good selectivity. Nevertheless, cis-BODIPY-DCH was not as good as the trans-one in terms of sensitivity and response rate. The LODs of cis-BODIPY-DCH for phosgene and acetyl chloride were 23.49 nM and 6.92 nM, respectively, which are much higher than those of trans-BODIPY-DCH (LOD ¼ 0.52 nM and 0.77 nM for phosgene and acetyl chloride, respectively).
To get insight into this observation, we performed density functional theory (DFT) calculations at the B3LYP/6-31G* level, 42 and gained optimized geometries of cis and trans-BODIPY-DCH (Fig. 4). In cis-BODIPY-DCH, two amine groups were located at the same side with a dihedral angle of 68.13 , and the space distance between the hydrogen atom (in N1) and N4 was 3.11Å. Thus, the two neighboring amines in cis-BOD-IPY-DCH are prone to form intramolecular H-bonding (Fig. 4b), which weakened the reactivity of the amines. In contrast, two amine groups were located at different sites with a dihedral angle of 168.22 in trans-BODIPY-DCH, and the distance between the hydrogen atom (in N1) and N4 is 3.74Å (Fig. 4a, S18 and S19 †), hindering the formation of intramolecular H-bonds. Furthermore, we measured the 1 H NMR spectrum of cisand trans-BODIPY-DCH in CDCl 3 and DMSO-d 6 , respectively. In DMSO-d 6 , the proton signal of the aromatic NH in cis-BODIPY-DCH was located at 5.66 ppm (Fig. S20 †), which shied to 8.48 ppm in CDCl 3 (Fig. S21 †), conrming the stable intramolecular H-bonding (IMHB) of N-H/N. 43 In this case, IMHB in cis-BODIPY-DCH would form a stable ve member ring to weaken the reactivity of two amino groups for phosgene and acyl chlorides. In contrast, IMHB does not form for trans-BODIPY-DCH according to the 1 HNMR spectra in CDCl 3 where the proton signal of the aromatic NH has completely disappeared (Fig. S21 †), conrming that the aliphatic NH 2 in trans-  BODIPY-DCH could quickly react with both phosgene and acyl chlorides.
We then used trans-BODIPY-DCH to simultaneously detect phosgene and DCP, as well as some volatile acyl chlorides. When a mixture of triphosgene and acetic chloride in a 1/3 molar fraction was added to the trans-BODIPY-DCH solution, a bathochromic peak appeared at 500 nm, while the initial absorption peak at 410 nm decreased to half. By carefully examining the uorescence spectra changes (Fig. 5b), we found that trans-BODIPY-DCH displayed two emission bands at 524 and 478 nm, respectively. Similar results were also observed when the mixture of triphosgene and some other acyl chlorides including DCP, OC, BzCl, and SOCl 2 was added to trans-BOD-IPY-DCH (Fig. 5c, d and S22 †), demonstrating that trans-BOD-IPY-DCH has a good capability for the simultaneous detection of phosgene, DCP and acyl chlorides.
To investigate the interaction mode, we employed high performance liquid chromatography (HPLC) to analyze the reaction mixture of trans-BODIPY-DCH with phosgene and acetyl chloride, respectively. As shown in Fig. 6a and b, a predominant elution peak appeared at 3.9 min in the HPLC prole, attributed to the reaction of trans-BODIPY-DCH with phosgene in a specic manner. We then isolated the major product trans-BODIPY-ICO from reaction mixtures and analyzed the NMR and HR-MS spectra. As revealed in Fig. 6c, the chemical shis of NH 2 (H1) and NH (H4) in trans-BODIPY-DCH were located at 3.38 and 5.25 ppm, respectively. Aer reaction with phosgene, the proton signal of H1 shied from 3.38 to 8.05 ppm, and the proton signal of H4 disappeared. In addition, the proton signals of two adjacent CH groups at 3.00 and 3.85 ppm slightly shied downeld. Moreover, a signicant carbonyl signal at 159.4 ppm appeared in the 13 C NMR spectra (Fig. S23 †), which was assigned to the carbonyl group of trans-BODIPY-ICO. HR-MS also conrmed the structure of trans-BODIPY-ICO, where a predominant peak at m/z ¼ 331.1538 was attributed to [C 16 H 17 BF 2 N 4 O] + (calc. 331.1536) (Fig. S24 †). UV- vis absorption and uorescence spectra (Fig. 6d and e) further veried that the major product was trans-BODIPY-ICO. These results clearly demonstrated that two adjacent amino groups in trans-BODIPY-DCH have coupled with phosgene to form imidazolone, which resulted in a red-shi of the UV-vis absorption and uorescence spectra.
For acetyl chloride, a predominant elution peak emerged at 5.6 min, exhibiting 87% of the total peak area in the HPLC prole (Fig. 6b). The major product trans-BODIPY-DACO was isolated from the reaction mixture, and characterized by NMR and HR-MS spectra. The initial proton signals of NH 2 (H1) and NH (H4) located at 3.38 and 5.25 ppm, respectively, shied to 9.36 and 8.19 ppm aer reaction with acetyl chloride (Fig. 6c). Additionally, the proton signal of the adjacent CH (H2) signicantly shied downeld (Fig. 6c and S25 †). Moreover, an important carbon signal at 171.9 ppm was observed in the 13 C NMR spectrum (Fig. S26 †), corresponding to the C]O group of trans-BODIPY-DACO. The HR-MS spectra also conrmed the structure of trans-BODIPY-DACO, where a peak at m/z ¼ 347.1853 was ascribed to [C 17 H 21 BF 2 N 4 O] + (calc. 347.1849) (Fig. S27 †). These results conrmed that acetyl chloride had coupled with the aliphatic amino group to produce an obvious uorescence enhancement in the blue channel (Fig. 6e).
We successfully obtained single crystals of trans-BODIPY-DCH, trans-BODIPY-ICO and trans-BODIPY-DACO from dichloromethane and n-hexane, respectively. In trans-BODIPY-DCH, the DCH is perpendicular to the BODIPY plane with a dihedral angle of 88.2 (Fig. 7a, Tables S3 and S4 †), and the bond length between N003 and C009 is 1.322Å, much shorter than the ordinary bond length of the N-C single bond (N003-C00C, 1.468Å, Table S5 †). This conrms that the trans-BODIPY-DCH exists as a tautomer with C]N as the dominant resonance structure (Fig. 7a). The trans-BODIPY-DCH tautomer was a cross-conjugated system, in which the push-pull interaction of the pyrrole nitrogen atoms was interrupted, which would expand the HOMO-LUMO band gap and lead to a signicant blue-shi in the UV-vis absorption and uorescence spectra. Hence, the trans-BODIPY-DCH tautomer exhibited an absorption band at 410 nm and a weak uorescence band at 478 nm. Aer reaction with acetic chloride, the aliphatic NH 2 with higher pK a in DCH was transformed to an amide in trans-BODIPY-DACO, but the aromatic NH was unaltered, which is evident from the single crystal structure in Fig. 7c. The bond length between N005 and C00D was 1.330Å (Table S6 †), shorter than the ordinary bond length of the N-C single bond (N005-C00b, 1.472Å), suggesting that the trans-BODIPY-DACO existed as a resonance tautomer (Fig. S28 †). By contrast, the two adjacent amines reacted with phosgene to form imidazolone, as can be seen in the single crystal structure in Fig. 7b. In trans-BODIPY-ICO, a bond like that of a normal N-C single bond (N006-C00K, 1.422Å) ( Fig. 7b and Table S7 †), implying that its tautomer was transformed to normal BODIPY aer reaction with phosgene (Fig. S29 †).
Density function theory (DFT) calculations revealed that the HOMOs of transand cis-BODIPY-DCH are enriched in the pyrrole ring, while the LUMO is mainly located at the BODIPY  core and part of DCH ( Fig. S30 and S31 †). Thus, the PET process from NH 2 quenched the intrinsic blue uorescence of BODIPY-DCH to some extent. Aer reacting with phosgene, the tautomer was transformed to the normal BODIPY structure, resulting in a large red shi in the absorption and uorescence spectra. In contrast, the LUMO energy of the tautomer increased along with a slight reduction of the band gap aer the trans-BODIPY-DCH reacted with acetyl chloride. In this case, the PET process of trans-BODIPY-DACO in the excited state was prohibited, and thus a signicant uorescence enhancement was observed in the blue channel.
We evaluated the cytotoxicity of trans-BODIPY-DCH by using a CCK-8 assay. L929 cells and HeLa cells exhibited 90% and 92% cell survival rates aer incubation with 20 mM trans-BODIPY-DCH for 24 h (Fig. S32 †), respectively, demonstrating that trans-BODIPY-DCH exhibits good biocompatibility. We then fabricated test strips for visualizing phosgene, DCP and gaseous acyl chlorides using trans-BODIPY-DCH. Herein, we used meltblown nonwovens as a matrix to fabricate test strips because nanobers exhibit high surface to volume ratios suitable for the adsorption of the sensor and analytes, which is especially benecial for the detection of gaseous compounds. Aer exposure to 20 ppm of phosgene or acetyl chloride, the nanobers still kept their original smooth morphology without exhibiting any damage (Fig. 8b).
We next employed them for the visual detection of phosgene, DCP and acyl chlorides in the gas phase. As shown in Fig. 8c, trans-BODIPY-DCH-loaded nonwoven test strips emitted faint blue uorescence under 365 nm light. Upon exposure to various concentrations of phosgene (0.5-20 ppm) under ambient conditions for 1 min, a signicant uorescence transition from faint blue to strong green uorescence was observed. By comparison, the test strips exhibited a noticeable blue uorescence enhancement with acetyl chloride (Fig. 8d). Likewise, DCP, diphosgene, and some other acyl chlorides (i.e. triphosgene, OC and BzCl) also stimulated the test strips to produce strong blue uorescence (Fig. S33 †). The trans-BOD-IPY-DCH-loaded nonwoven test strips exhibited good capability for the high throughput determination of phosgene, DCP and acyl chlorides, and this sensing platform based on melt-blown nonwovens is superior to some conventional systems. 44,45 To further assess the selectivity, trans-BODIPY-DCH-loaded nonwoven test strips were exposed to various analytes (20 ppm), including phosgene, diphosgene, DCP, DMMP, TsCl, POCl 3 , SOCl 2 , HCl, BzCl, OC, AC, formaldehyde, methylglyoxal, acrolein, and triphosgene. As shown in Fig. 8e, the phosgene resulted in a signicant green uorescence being observed for the test strip. On the other hand, DCP, diphosgene and other acyl chlorides caused a remarkable blue uorescence enhancement, while DMMP, HCl, formaldehyde, methylglyoxal and acrolein led to negligible uorescence changes.
To visualize analytes on the test strips, we utilized laser confocal uorescence microscopy to record the uorescence images of trans-BODIPY-DCH-loaded nonwoven test strips before and aer addition of analytes. trans-BODIPY-DCHloaded nonwoven test strips themselves exhibited lamentous bers with faint uorescence in the blue channel (Fig. 9). Aer exposure to phosgene (0.5-20 ppm), an obvious uorescence enhancement was observed from the green channel, and the initial blue uorescence intensity diminished. Such obvious uorescence color changes enable us to use our naked eyes to detect toxic phosgene. While DCP, diphosgene, acetyl chloride, triphosgene, OC, BzCl or SOCl 2 resulted in a remarkable uorescence enhancement of the test strips in the blue channel (Fig. 9d, S34 and S39 †).

Conclusions
In summary, we have successfully developed a robust sensing platform for the selective detection of phosgene, DCP, and other acyl chlorides using different emission channels based on transand cis-BODIPY-DCH. From UV/uorescence titrations, single-crystal structure analysis, and DFT calculations, we found that they exhibited similar sensing mechanisms towards CWAs while trans-BODIPY-DCH exhibited better detection performance than the cis-version due to the intramolecular Hbonding of the cis-version. trans-BODIPY-DCH underwent nucleophilic substitution with phosgene, resulting in a noticeable and rapid uorescence change from weak blue uorescence to bright green uorescence (<3 s), with a low detection limit (0.52 nM). In contrast, acyl chlorides and DCP resulted in a rapid highly sensitive uorescence enhancement with trans-BODIPY-DCH with a low detection limit (0.77 nM). Furthermore, it could be conveniently loaded onto melt-blown nonwovens to serve as a device for the high throughout and visible detection of gaseous phosgene, diphosgene, DCP, and other acyl chlorides. Signicantly, trans-BODIPY-DCH-loaded nonwoven test strips could be used for the highly sensitive and visual detection of toxic acyl chlorides at the sub-ppm levels.
GC-MS is the most used technique that can identify chemical warfare agents at concentrations ranging from low ppb to ppm (w/w). HPLC can also be used to determine the derivatization products of phosgene, nerve agents and aryl chlorides at subppm levels, but requires appropriate pro-derivatization procedures. In addition, GC-MS and HPLC analyses require several minutes to complete the measurement, and the instruments are non-portable. In contrast, our portable trans-BODIPY-DCHloaded nonwoven test strips can rapidly determine sub-ppm levels of chemical warfare agents (within 3 seconds), which is particularly suitable for in-eld determination of chemical warfare agents without the need for bulky instrumentation.
In summary, this research not only results in a uorescent platform for the selective detection of phosgene and volatile acyl chlorides, but also provides a promising conguration-based sensing strategy to develop a "one-for-multiple" chemosensor strategy.

Data availability
All data supporting this study are provided as ESI, † which includes characterization data, crystal data, simultaneous sensing of diverse analytes with trans-BODIPY-DCH, the spectral response of cis-BODIPY-DCH towards some analytes, X-ray single crystallography, density functional theory calculations, cytotoxicity, test strips for detection of analyte vapor, confocal uorescence images of test strips for analyte vapor, 1 H NMR, 13

Author contributions
T. D. J. and J. S. K. supervised the project. L. Z. contributed to the project design and along with B. Z. carried out experiments on synthesis and characterization of compounds. T. C. and S. K. cultured the single crystals, analyzed the crystal structure, and investigated the sensing mechanism of the sensor and its application. W. L. made conceptualization. Y. T. conducted theoretical calculations. All authors prepared and edited the manuscript.

Conflicts of interest
There are no conicts to declare.