Challenges in Fluorescence Detection of Chemical Warfare Agent Vapors Using Solid‐State Films

Challenges in Fluorescence Detection of Chemical Warfare Agent Vapors Using Solid-State Films

Shengqiang Fan, Guanran Zhang, Genevieve H. Dennison, Nicholas Fitzgerald, Paul L. Burn, Ian R. Gentle*, Paul E. Shaw*

Dr S. Fan, Dr G. Zhang, Prof. P. L. Burn, Prof. I. R. Gentle, Dr P. E. Shaw

Centre for Organic Photonics and Electronics, School of Chemistry and Molecular Biosciences, The University of Queensland, Brisbane, QLD, 4072, Australia

Email: [email protected]; [email protected]

Dr G. H. Dennison, Dr N. Fitzgerald

Defence Science & Technology Group, Land Division, Fishermans Bend, Vic 3207, Australia

Keywords: nerve agent simulant, hydrolysis, fluorescence, sensing, solid-state

Organophosphorus (OP)-based nerve agents are extremely toxic and potent acetylcholinesterase inhibitors and recent attacks involving nerve agents highlight the need for fast detection and intervention. Fluorescence-based detection, where the sensing material undergoes a chemical reaction with the agent causing a measurable change in the luminescence, is one method for sensing and identifying nerve agents. Most studies use the simulants diethylchlorophosphate (DCP) and di- iso-propylfluorophosphate (DFP) to evaluate the performance of sensors due to their reduced toxicity relative to the OP nerve agents. While detection of nerve agent simulants in solution is relatively widely reported there are fewer reports on vapor detection using solid-state

This is the author manuscript accepted for publication and has undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/adma.201905785.

sensors Herein progress in organic semiconductor sensing materials developed for solid-state detection of OP-based nerve agent vapours is reviewed. Also the effect of acid impurities arising from the hydrolysis of simulants and nerve agents on the efficacy and selectivity of the reported sensing materials is discussed. Indeed, in some cases it is unclear whether it is the simulant that is detected or the acid hydrolysis products. Finally, we highlight that while analyte diffusion into the sensing film is critical in the design of fast, responsive sensing systems, it is an area that is currently not well studied.

1. Introduction

Chemical warfare agents (CWAs) can be classified by their physiological effects into several groups, including nerve, blister, blood, choking, harassing, and incapacitation agents, and toxins.[1] Whilst the use of CWAs is banned and access to precursors tightly controlled, the relative ease of synthesis of some CWAs means they are potentially accessible to terrorist groups or rogue nation states, and still pose a real threat to public safety and national security.[2]

Among CWAs, nerve agents are a family of highly toxic organophosphorus (OP) compounds.[3] The main routes of exposure are via inhalation or the skin and their mode of action is inhibition of the enzyme acetylcholinesterase (AChE), which is critical for hydrolysing the neurotransmitter acetylcholine to control its concentration.[4] The inhibition of AChE leads to an accumulation of acetylcholine and results in muscle overstimulation.

Nerve agents are generally divided into two series, G and V (Figure 1). G-series including tabun (GA), sarin (GB), soman (GD) and cyclosarin (GF) were first developed before and during World War II by

German chemists. The V-series were synthesised later (1950s) by the British and weaponised by the

American military. Compared to the G-series, the V-series are more persistent in the environment

due to them being slower to hydrolyse and less volatile.[5] In addition, the V-series are more toxic,

[5] with LD50 (lethal dose, 50%) values 1-2 orders of magnitude lower than the G-series. Sarin was reportedly used in the recent Syrian civil war several times and VX was used in the assassination of

Kim Jong-Nam in 2017,[6] and hence sadly there is still a need to be able to rapidly detect nerve agents and CWAs more generally.

Figure 1 Structures of nerve agents (G- and V-series) and simulants.

Nerve agents are odourless and colorless when pure and development of detection technologies that can meet the needs of military, first responders (national security), healthcare and environmental monitoring agencies is a continuing challenge. A variety of detection methods have been developed including gas chromatography-mass spectroscopy (GC-MS),[7] enzyme-based biosensors,[8-10] chemiresistors,[11-13] and surface acoustic wave (SAW) sensors.[14,15] These detection systems are either slow, complex, nonselective, or the equipment is too cumbersome for field use.

An alternative detection method is optical-based sensing, that is, to use a material whose absorption or emission changes by reaction with the nerve agents. Potential advantages of optical detection include portability of the equipment, real-time monitoring, and rapid and selective detection. If the functionality on the sensing material is carefully chosen to specifically react with a nerve agent, optical detection will be less susceptible to interferents. Fluorescence-based detection has several potential advantages over colorimetric detection in terms of speed, sensitivity, and mode of detection. Fluorescence-based detection can be achieved using a change in the steady-state photoluminescence (PL) intensity (“turn on” or “turn off”), PL color (as shown in Figure 2), or PL lifetime. Therefore, it is possible to monitor multiple characteristics of the PL simultaneously (e.g., color and intensity) to improve selectivity.

Figure 2 Illustration of fluorescence-based detection of CWAs through a PL color change, (i)(ii), or a PL intensity change, (ii)(iii).

It will be appreciated that due to the toxicity of nerve agents and their restricted availability, testing sensing materials against real agents can only be undertaken in OPCW declared facilities. To aid the development of new sensing materials their performance is typically evaluated using simulants that

are reported to have lower toxicity. This perspective will focus on the detection of the G-series nerve agents for which diethylchlorophosphate (DCP) and di-iso-propylfluorophosphate (DFP) are the two widely used simulants for Sarin (GB), Soman (GD) and cyclosarin (GF) that contain a P-F bond, with diethylcyanophosphate (DCNP) being the simulant for Tabun (GA) which is comprised of a P-CN bond. It should be noted that although DCP is a much more popular simulant than DFP, whether it is appropriate to simulate the P-F bond of the G-agents (sarin, soman and cyclosarin) is questionable as the P-Cl (326 kJ/mol) and P-F (490 kJ/mol) bond strengths are markedly different.

2. Solid-state fluorescence sensing of nerve agent and simulant vapors

The majority of reported studies on the fluorescence-based detection of nerve agents have been done in solution with fewer reports describing solid-state sensors. In general, sensing materials are designed to have nucleophilic functional groups that purportedly attack the phosphorous atom leading to subsequent displacement of the cyanide moiety for GA and fluoride for GB, GD and GF and attachment of the “nerve agent” to the sensing compound. In an alternative strategy, the sensing materials may have functionality that enables co-ordination with the P=O functional group to give distinct changes in the PL characteristics. Several recent reviews discuss the myriad of different materials reported for nerve agent and simulant sensing;[16-20] hence, only those materials that are reported in the context of vapor phase detection using solid-state films will be discussed here. Vapor phase detection using solid-state films is the most practicable approach for detecting nerve agents in the field. The sensing materials and their performance in solid-state vapor detection are summarised in Table 1, with their molecular structures shown in Figures 3-6.

Table 1 Summary of sensing performance of materials that have been examined for the detection of nerve agent simulant vapors using solid-state films. The molecular structures of the materials listed

in this table are shown in Figures 3, 4, 5 and 6. (TEP – triethyl phosphate; TEA – triethylamine; PEO – polyethylene oxide; PVP – poly(4-vinylpyridine); DMAP – 4-dimethylaminopyridine; DMMP – dimethyl methylphosphonate; CA – cellulose acetate; PS – polystyrene)

Reporter Matrix or Analyte Limit of Selectivity and compound Sensor Proposed mechanism substrate vapors detection sensitivity isolated & characterised

Coordination with 1a, Not False positive Silica particle TEP La3+ and Eu3+ to free Yes 1b[21] stated from TEA the fluorescent ligands

N-protonation to give False positives 8 nM 2[22] PEO DCP strong intramolecular from AcOH and No solution charge transfer (ICT) HCl

Spin-coated 2.6 ppb False positive 3[23] DCP As 2 No on quartz vapor from HCl

Selective to DCP 1.82 ppb cf non-halogen 4[24] Filter paper DCP As 2 No vapor OP compounds such as TEP

Spin-coated 0.14 ppb False positive 5[25] DCP As 2 No on quartz vapor from HCl

Amine Films showed Polymer phosphorylation to Not selective 6[26] microbeads DCP reduce photoinduced No stated response to DCP with PVP electron transfer cf DMMP (PET)

Solution showed Filter paper selective Not Phosphorylation to 7[27] with TEA and DCP response to DCP Yes stated suppress bond rotation DMAP and DCNP cf DMMP and TEP

Phosphorylation and Films showed Spin coated DCP, Not 8[28] then N-alkylation to stronger PL No with CA DFP stated give cyclic ammonium response to DCP salt with rigid and DFP than

structure and thereby HCl vapor distinct fluorescence

Solution showed stronger PL DCP, Not 10[28] Not stated As 8 response to DCP Yes DFP stated and DFP than hydrochloric acid

Solution showed stronger PL 11a, DCP, Not PS As 8 response to DCP Yes 11b[29] DFP stated and DFP than hydrochloric acid

Phosphorylation and DCP, Not then N-alkylation to 12[30] Filter paper Not stated Yes DFP stated get cyclic ammonium salt with reduced PET

Solution: Solution showed DCNP selective DCP, 0.10 13[31] PEO As 12 response to DFP No DCNP ppm; and DCNP over DFP 0.39 HCl ppm

Solution: DCNP 14a, DCP, 4.01 PEO As 12 As 13 No 14b[31] DCNP ppm; DFP 6.10 ppm

Solution: Protonation by Polyurethane DCNP 7 HClO had hydrogel DFP, 4 15[32] ppm; As 12 similar PL with Yes matrix on DCNP DFP 8 the cyclic poly(ethylene) ppm product

Films showed 0.2 ppm selectivity to DFP, for 16[33] As 15 As 12 DCNP over other No DCNP DCNP common vapors vapor in military setting such as

gasoline, ammonia and pipe fumes

2.1 Organometallic chemosensors

Several organometallic coordination chemosensors have been developed based on La3+,[21] Eu3+,[21,34-

36] or Zn2+[37,38] complexes or Ir3+/Eu3+ diads.[39,40] Lanthanide complex-based sensing materials and mechanisms to detect nerve agents and mimics can be found in a recent review which includes both solution and solid-state detection.[34] Solid-state detection has been typically achieved by incorporating the metal complex in a polymer matrix or by being sorbed onto silica or filter paper.

Upon exposure to the nerve agent or simulant vapor, the original ligands are displaced, leading to a change of the PL of the metal complexes.[39] In an alternative approach, ligands such as 1aꞌ and 1bꞌ that emit differently when co-ordinated or not co-ordinated to the metal ion have been used to detect the presence of OP compounds (Figure 3).[21] However, PL of phosphorescent metal complexes is known to be readily quenched by oxygen, limiting the use of this type of material in real world applications. In addition, such a coordination based approach is likely to give false positives when encountering other coordinating vapors such as N,N-dimethylformamide (DMF), triethylamine (TEA), and alcohols.[21,39]

Figure 3 Sensing mechanism of organometallic chemosensor 1a and 1b developed by Rowan and Weder et al.[21]

2.2 Nitrogen-based nucleophiles

Materials with nitrogen-based nucleophiles constitute an important class of fluorescent sensors investigated for the detection of nerve agent and simulant vapors. N-based chemosensors (Figure 4) include quinoline (e.g., 2[22]), pyridine (e.g., 3[23] and 4[24]), Schiff bases (e.g., 5[25]), amines (e.g., 6[26]), or quinoxaline.[41] Solid-state detection of the nerve agents has been reported to be achievable in different formats including drop-casting (2) in a poly(ethylene oxide) (PEO) matrix, having (6) and (4) sorbed onto microbeads or glass, or filter paper, respectively, or in the form of a thin film by spin- coating (3) and (5) onto quartz substrates. Two typical detection mechanisms have been used with these materials to detect the nerve agents: protonation to enhance intramolecular charge transfer

(ICT) (22ꞌ , 3, 4, and 5) and phosphorylation to inhibit the photoinduced electron transfer (PET) process (66ꞌ ), with both mechansims leading to a fluorescence “turn on” response.[22-26] There are other N-based sensing materials (not listed here) that give fluorescence “turn off” signals but the sensing mechanisms are unclear.[41-43] Although N-based chemosensors have been reported to achieve a very low limit of detection against DCP vapor using 3, 4 and 5, false positives from acids due to the protonation of N-based functionality is likely to be a major issue. Furthermore, these N-

based chemosensors appear not to have been used for the detection of DFP, which might be related to the lower reactivity of the P-F bond relative to the P-Cl bond in either phosphorylation or hydrolysis. This means that these materials might not be suitable for detecting P-F based nerve agents.

Figure 4 Structures of chemosensors with N-based nucleophiles.[22-26]

2.3 Oxygen-based nucleophiles

O-based nucleophiles as chemosensors for solid-state detection have contained either phenols or oximes. Phosphorylation of phenols does not normally lead to a significant change in the fluorescent properties of the sensing material.[27] However, if phosphorylation is able to alter the molecular

rotation of a chromophore, the fluorescence may change.[27,44] For example, when compound 7[27] is phosphorylated with DCP it restricts the rotation in 7ꞌ leading to enhanced fluorescence (Figure 5).

Detection using 7 employed a porous silica TLC plate as the substrate in order to evenly mix the sensing material and the organic base [e.g., triethylamine (TEA) or 4-N,N-dimethylaminopyridine

(DMAP)] that was required for rapid phosphorylation. However, phenols react with P-F based nerve agents and the simulant DFP relatively slowly,[45,46] limiting their use. In contrast, oximes (-C=N-OH) have been reported to react with sarin, soman and DFP an order of magnitude faster than phenols in aqueous solutions possibly owing to the so-called α-effect, but they rely upon basic solutions.[46] A variety of fluorescent materials based on the oxime nucelophiles have been developed for solution- based detection,[47-52] but not translated to film-based vapor detection. For example, a fluorescent polymer containing an oxime moiety exhibited fluorescence “turn on” against DCP in solution and fluorescence “turn off” in the solid state against DCP vapor, the reason for which is unclear.[53]

Figure 5 Sensing mechanism of phenol-containing chemosensor 7 developed by Bouffald and Kim et al.[27]

2.4 Nitrogen-oxygen bifunctional nucleophiles

Alkyl alcohols are more nucleophilic than phenols but phosphorylation does not change the π- conjugation of the chromophore and hence fluorescence. However, alkyl alcohols can be

incorporated in chemosensors together with a second moiety, such as pyridine (e.g., 8[28] and 10[28]) or arylamines (e.g., 11(a,b),[29] 12,[30] 13,[31] 14,[31] 15,[32] and 16[33]) so that a reaction, such as cyclization (for example, 8ꞌ 9), can in principle occur following phosphorylation (Figure 6). The subsequent cyclization is then proposed to lead to a new structure whose photoluminescence is distinct from that of the original sening compound. Chemosensors 8 and 10 are two reported examples,[28] with the proposed mechanism for 8 being that it reacts with either DCP or DFP to produce phosphate 8ꞌ , which then undergoes intramolecular N-alkylation to cyclic product 9.

Compound 9 has been reported to have stronger and longer wavelength fluorescence emission.

Films of 8 blended with cellulose acetate (CA) gave very fast fluorescent responses, in seconds, when exposed to 10 ppm DFP vapor, which is one of the fastest responses reported so far. A variety of materials utilising the cyclisation approach have been developed including 11(a,b), 12, 13, 14, 15 and

16. For compounds 8, 10, and 11(a,b), the fluorescence “turn on” caused by DCP or DFP is believed to arise from the rigid structure of the cyclised product. For compounds 12, 13, 14(a,b), 15 and 16, the amine-containing moiety can quench the PL of the chromophore through PET (or ICT) while the proposed formation of the cyclic ammonium salt when exposed to DFP or DCNP retards the PET process, resulting in fluorescence “turn on”. Interestingly, the BODIPY-based chemosensors 13-16 showed a much better fluorescent response to DCNP than DFP or DCP although the reason for this was not explained. Blends of 13 or 14(a,b) with PEO were found to change their luminescence in the presence of DFP and DCNP. A blend of 16 in a polyurethane hydrogel on a polyethylene substrate was reported to have a relatively low limit of detection (0.2 ppm) against DCNP vapor. One of the issues in detecting DCP and DFP is the presence of acid as an interferent. Studies of 10 in solution and 8 in the solid state showed a significantly higher fluorescence intensity when when they were

exposed to DCP and DFP compared to hydrogen chloride/hydrochloric acid. However, the fact 8 and

10 respond to HCl, and in principle other acids, means that it would be difficult to differentiate between a response to a low concentration of the DCP or DFP and acid. Solutions of 13 or 14a showed fluorescence “turn on” when exposed to DCNP and DFP while fluorescence “turn off” against hydrochloric acid, possibly due to the additional pyridyl group in the structure which becomes protonated. In this case, the different response provides a pathway to selectivity for the detection of the desired analyte. It should be noted that many chromophores are known to have changes in fluorescence with pH,[26,54] thereby selectivity measurements to distinguish between nerve agent simulants and acids should consider the solution pH or acid vapor concentration. The issue of acid false positives will be discussed in detail below.

Figure 6 Structures of chemosensors with N-O bifunctional nucleophiles that can undergo cyclisation.[28-33]

Despite a large number of sensing materials being developed for the vapor phase detection of nerve agents and simulants, most research articles only present proposed mechanisms with limited evidence on how the sensing works. The reporter compounds, i.e., the product from the reaction between the sensing compound and the analyte such as DCP or DFP are given in some reports but many are not characterised to a sufficient level to unambiguously confirm the structures.[27,28] In some cases, the reporter compounds were synthesised from the sensing compound using more reactive reagents, e.g., tosyl chloride[30,32] and thionyl chloride, but this in itself is not sufficient evidence that the compounds would be formed during the sensing process.[28,29]

3. Stability of nerve agent simulants

At first sight it appears that significant progress has been made in developing materials that can detect nerve agents and, in particular, their simulants. However, while it is known that nerve agents and their simulants are susceptible to hydrolysis by moisture in the air, leading to acid impurities

(Figure 7), it is not clear whether the sensing community has always taken this into account when results are reported. Materials of the G-series have been shown to react much faster with water than the V-series, however, materials in both families hydrolyse.[55] The major products from sarin hydrolysis are iso-propyl methylphosphonic acid and hydrogen fluoride. In fact, the hydrolysis of G- and V-series in the presence of inorganic bases,[55,56] small-molecule organic bases,[57] polymeric bases,[58] or metal-organic frameworks[59] into non-toxic products is currently an active area of research.[60]

Figure 7 The major acid products from hydrolysis of DCP and DFP.

Thus, when testing a sensing material against an agent it is critical to know whether acid degradation products are present or not. The hydrolysis process of simulants DCP[61] and DFP[62] has been studied using NMR measurements. For example, the hydrolysis of DCP was monitored using 31P NMR and it was found that diethylphosphoric acid (DEPA) could be observed 1 minute after water addition.[61] 1H

NMR experiments also showed that addition of water to DFP leads to the formation of the

hydrolysed product, di-iso-propylphosphoric acid (DIPPA) after 2.5 hours.[62] The relatively slower hydrolysis rate of DFP compared with DCP is largely due to the less reactive P-F bond. It is also found that storage conditions are critical for the stability of DFP.[63] Commercial DFP is normally stored at 4 oC, however, most DFP samples were found to degrade significantly at this temperature during long- term storage (i.e., from 2 weeks onwards).[63] In comparison, only minor degradation was observed at -10 oC and none was observed at -80 oC.[63] In addition, glass was found to accelerate the rate of

DFP hydrolysis, which is possibly due to the attack by hydrogen fluoride on the silica in the glass.[63,64]

Indeed, trace amounts of HF have been observed in freshly purchased DFP from commercial suppliers.[65] The inconsistent quality of commercial sources of DFP is therefore a real problem for sensing measurements as well as other areas. For example, toxicological studies show that LD50 values vary significantly from 0.0027 mg/kg to 6.4 mg/kg in a mouse model, with the differences arising from the source of the DFP used.[63] Acid impurities not only exist in the simulant DFP, but they are very likely to be present in the actual nerve agents such as sarin as the latter is more readily hydrolysed than the former.[46] 31P NMR showed a nearly full conversion of sarin to its breakdown product, iso-propylmethylphosphonic acid, in the presence of 1 equivalent of water after 24 hours.[57]

Only very few reports have considered the issue of acid impurities in the sensing measurements. For example, in one case an acid scavenger, poly(4-vinylpyridine) (PVP) was added to DCP[12] immediately before the sensing measurement, and in a second report, piperidinomethyl polystyrene

(PDN-PS) was added to DFP for storage and removal of HF.[65] The former was used in a chemiresistor with ionic liquid for the detection of DCP and the latter was used in solution-based detection of DFP. However, for fluorescence-based vapor detection, acid impurities in the nerve

agent/simulant vapors have not generally been considered, which may lead to incorrect assessment of the performance of the sensing materials. It should be noted that the vapor pressures of the G- series simulants are in the range of parts per thousand and hence a sample in which 1% had hydrolysed would lead to acid impurity with a vapor pressure in the ppm range – a level that could be easily detected. Such concentrations are usually hard to observe by NMR but can be easily observed by changes in fluorescence. There has been a couple of reports that compare the fluorescent response of sensing materials with acids or nerve agents/simulants but these did not consider the fact that acid impurities might be in the analyte.[28,31] Care should therefore be taken in future reports to ensure good quality simulants are used and appropriate storage conditions applied including appropriate container type and temperature, preferably in a plastic bottle and at -10 C for DFP.[63] The addition of a non-volatile acid scavenger such as PVP for DCP and PDN-PS for DFP might also be a way to ensure that the simulant rather than an acid is causing a response for vapor phase detection.

4. Potential for false positive responses from acids

As discussed above, N-containing chemosensors with either N-based or N-O bifunctional nucleophiles, are reported to be the most sensitive materials for the detection of nerve agent simulants. They are also inert to non-halogen OP compounds and common solvents and are therefore not highly susceptible to false positives come from these latter analytes. However, it should be noted that fluorophores with N-containing moieties including arylamine,[66] pyridine,[67] and benzothiazole,[68] are also excellent protonation-induced fluorescence “turn on” pH sensors. This means that the N-containing materials designed for the detection of nerve agents and simulants are intrinsically sensitive to acids. Therefore, it is not surprising that acids can produce false positives for

these sensing materials.[22-24, 28, 29] Studies on the N-based compounds 2,[22] 3,[23] and 4[24] show that the protonated derivatives account for the fluorescence “turn on” responses upon exposure to DCP, which further confirms that these sensing materials cannot distinguish between nerve agent simulants and acids. Although films of the N-O bifunctional material 8,[28] and 10[28] or 11a[29] in solution show stronger PL responses to simulants than acids, the PL peak wavelengths are almost identical. It should be noted that the magnitude of PL response often depends on the concentration of the analyte, therefore, PL intensity in itself cannot be used as a parameter of “selectivity”, i.e., it is not possible to conclude that an analyte is acidic from the weak PL response as it might be a diluted simulant. Most importantly, the PL response to an analyte is not always linear. For example, in some cases, low concentration of the analyte gives fluorescence “turn on” while high concentration leads to fluorescence decrease.[53] Unfortunately, this is little considered in literature when comparing the

PL response from simulants with other interferants. Confirmation of the sensing ability of such materials is required before further effort is put into this sensor design strategy for the selective detection of sense nerve agents and simulants.

Due to the rapid hydrolysis that occurs with phosphonofluoridates it is likely that acid would be present in any CWA testing or sensing environment. Therefore, a key characteristic of developed sensing systems is to ensure they are not responsive to acid impurities. Here we review some strategies that are potentially selective, but the studies are still in the preliminary investigation stage or currently only for solution-based or using colorimetric detection. There is a need for more detailed investigation to determine whether they can be reliably used in the solid-state for vapor detection. For example, Lloyd et al[29] found that loading an acid-scavenger, silica, to the sensing films made from polymer 11b:polystyrene blends can decrease the fluorescence response from

hydrochloric acid, as shown in Figure 8. This concept is similar to the addition of Cs2CO3 particles to

PEO films containing sensing materials in a colorimetric detection system.[69] The organic base, (di- iso-propylamino)ethyl methacrylate (DPAEMA) has also been incorporated with the sensing material

11b through co-polymerisation but in this case there was a fluorescence response from hydrochloric acid.[29] This indicates that phase separation of the acid absorbent and the sensing material is critical.

It should be noted that acids have an equilibrium distribution in the two phases, the inorganic silica particles and the organic sensing composition, and it is possible that hydrochloric acid tends to be sorbed by silica and is less likely to stay in the poly(styrene) phase. However, the organophosphoric acid, HF and HCN breakdown products of the nerve agent simulants need to be examined due to their having differing characteristics to HCl.

Figure 8 Silica-dispersed polymeric sensing films show diminished PL response to hydrochloric acid.[29] The sensing films contain 0.1% polymer of 11b in PS and are coated on filter paper. Reprinted (adapted) with permission.[29] Copyright 2018, American Chemical Society.

A second strategy that might enable a selective responsive to nerve agents and simulants involves the use of supramolecules. For example, Sfrazzetto et al.[70,71] have developed chromophores with

two arms that can selectively pick DMMP through host-guest hydrogen bonding and the group is working on functionalising this organic scaffold to obtain solid devices for vapor detection (Figure 9).

Recently, Xiong and Che et al.[72] reported a nanofiber fabricated using compound 18 that showed a sharp fluorescence “turn off” before leveling off under light illumination, while showing “turn on” again when exposed to DCP (Figure 10). The proposed mechanism is related to the packing of nanofibers in which the close packing of the fibres is reduced under photoexcitation but is increased through the bonding between hydroxyl group and DCP. Considering the much slower reaction of the hydroxyl group with DFP than DCP, it would be interesting to see whether the method is applicable to DFP and hence nerve agents.

Figure 9 A supramolecular chemosensor 17 reported by Sfrazzetto et al.[70]

Figure 10 A nanofiber sensor fabricated from compound 18 showed PL decrease upon light excitation while PL increase when exposed to DCP, possibly due to the change in inter-nanofiber interactions.[72] Reprinted (adapted) with permission.[72] Copyright 2018, American Chemical Society.

A third strategy for possible selective detection is to use a sensor array. Such an approach has been examined in fluorescence detection of nerve agents in solution[73] or colorimetric detection of vapors using solid-state films.[74] At this stage there appear to be no reports of fluorescence-based solid- state detection of vapors using arrays of compounds. One of the advantages of the sensor array is that a variety of sensors such as pH probes, solvatochromic dyes, redox indicators, and nucleophilic dyes can be selected to respond to the hydrolised products such as fluoride, cyanide, phosphate, and protons.[74]

5. Analyte diffusion in solid-state sensing films

The true reactivity of the sensing materials with nerve agents and simulants is one of the major concerns for the solid-state detection of vapors. Alongside this, diffusion of the nerve agent analyte

in the sensing films is another critical factor influencing the detection performance. Studies on the fluorescence quenching-based detection of nitro-containing explosive vapors using organic fluorophores have shown that analyte diffusion is a first order parameter.[75,76] The diffusion of nerve agents or simulants in the solid-state films have not been reported. However, some reported investigations indirectly suggest that the properties of the film might be critical to the fluorescence- based vapor detection. For example, hydrophilic polymers including cellulose acetate,[28,78] PEO,[22, 31] and polyurethane hydrogel[32] have been employed as the matrix for sensing materials, which indicates that favourable diffusion of the polar analyte in the solid-state films could be important to the sensing performance. In addition, the introduction of the plasticizer, triethyl citrate, to very thick films (0.5 mm) was found to be essential for the fluorescent sensing of nerve agent simulants owing to the increased permeability and mobility.[77] However, the effect of the plasticizer has not been examined with the film thicknesses that are commonly used in fluorescence-based sensing, i.e., tens to hundreds of nanometers. The inclusion of glycolated chains, e.g., tetra(ethylene glycol) monomethyl ether, to a polymeric chromophore (19, Figure 11) was found to enable a colorimetric response to DCP vapor at a limit of detection of 6 ppm. In contrast, thin films (50 nm) that contained non-glycolated polymers did not show detectable color change even with saturated DCP vapor.[78] Increased diffusion of DCP into the glycolated polymer is a possible reason for the enhanced colorimetric change. It is also possible that the increased polarity and softness by the glycolated chains allows for fast reaction between the hydroxyl group of the sensing material and

DCP. It should be noted that while the critical role of the glycolated chain is seen in a colorimetric detection system, this finding could also be relevant to fluorescent detection due to the similarity between the materials used in these two methods. In short, it seems that the use of a hydrophilic

polymer matrix or plasticizer, and/or the glycolation of the sensing material is important. However, whether the success of the approach is due to favourable analyte diffusion and/or fast reaction between the sensing material and the analyte is as yet unclear. To clarify this question, further investigation on the analyte diffusion using specific measurements is necessary, e.g, as quartz crystal microbalance (QCM) measurements.[75]

Figure 11 A polymeric sensor 19 developed by Swager[78] shows that in the solid state the glycolated chain is critical to the colorimetric response to DCP vapor.

6. Summary and perspectives

Public security is a continuing topic of discussion. With the increasing number of events related to the use of chemical warfare nerve agents, intensive efforts by the scientific research community have been undertaken in order to find a reliable method for recognizing such threats and deliver an early warning. Fluorescence-based detection using organic semiconductor materials have attracted much interest due to their potential for high sensitivity, rapid response, and the portability of the equipment. Despite the vast number of novel sensing materials reported there are relatively few of these that have been used in films for the detection of nerve agents and simulant vapors. Herein, we

have discussed the progress on fluorescence-based vapor detection of nerve agents and their corresponding simulants, including the sensing materials, detection mechanisms, limit of detection, and selectivity. Sensing materials with N-based or N-O bifunctional nucleophiles appear to represent the most sensitive materials. However, these materials are also prone to protonation by acids and give fluorescence responses similar to that from the nerve agents and simulants. The presence of acid is an issue in sensing measurements, and it clear that acid impurities in the G-series and simulants are often ignored or the results misinterpreted. As a general principle, care needs to be taken to correctly store the nerve agent/simulants and the addition of an acid scavenger should be adopted to avoid simply measuring the response of the sensing material to acid impurities. We strongly recommend that the sensing materials with N- or N-O bifunctional nucleophiles should be remeasured using simulants whose acid impurities have been removed to make sure that they are not simply pH sensors and that new sensing materials should be tested with both pure and impure simulants. We have also assessed the currently small number of strategies that have the potential to distinguish nerve agents or simulants from acids. These strategies include the incorporation of inorganic acid scavenger to the solid-state films, supramolecular methods and sensor arrays. Finally, we have highlighted the potential role of a hydrophilic matrix, plasticizer or glycol chains on the sensing performance and believe these are interesting avenues for further investigation into the role of analyte diffusion into the sensing films. Such an approach is important to be able to gain mechanistic insights into the factors that are important for vapor sensing and provide information for the design of fast, responsive sensing systems.

Acknowledgements

PES is supported by a University of Queensland Amplify Fellowship. PLB is an Australian Research Council Laureate Fellow (FL160100067). This research was supported by funding from the Australian Research Council under the Discovery Program (DP170102072). The provision of financial or research support does not constitute an express or implied endorsement of the results or conclusions presented here by DST Group or the Australian Department of Defence.

Received: ((will be filled in by the editorial staff)) Revised: ((will be filled in by the editorial staff)) Published online: ((will be filled in by the editorial staff))

References

[1] M. Kloske, Z. Witkiewic, Chemosphere 2019, 221, 672.

[2] T. C. C. Franca, D. A. S. Kitagawa, S. D. de. A. Cavalcante, J. A. V. da Silva, E. Nepovimova, K. Kuca, Int. J. Mol. Sci. 2019, 20, 1222.

[3] S. Costanzi, J.-H. Machado, M. Mitchell, ACS Chem. Neurosci. 2018, 9, 873

[4] R. T. Delfino, T. S. Ribeiro, J. D. Figueroa-Villar, J. Braz. Chem. Soc. 2009, 20, 407

[5] S. W. Wiener, R. S. Hoffman, J. Intensive Care Med. 2004, 19, 22

[6] E. Nepovimova, K. Kuca, Arch. Toxicol. 2019, 93, 11.

[7] W. E. Steiner, S. J. Klopsch, W. A. English, B. H. Clowers, H. H. Hill, Anal. Chem. 2005, 77, 4792.

[8] L. Matějovský, V.IPitschmann, Biosensors 2018, 8, 51.

[9] J. P. Walker, S. A. Asher, Anal. Chem. 2005, 77, 1596.

[10] G. Liu, Y. Lin, Anal. Chem. 2006, 78, 835.

[11] J. F. Fennell, H. Hamaguchi, B. Yoon, T. M. Swager, Sensors 2017, 17, 982.

[12] R. Zhu, J. M. Azzarelli, T. M. Swager, Angew. Chem. Int. Ed. 2016, 55, 9662.

[13] K. Saetia, J. M. Schnorr, M. W. Mannarino, S. Y. Kim, G. C. Rutledge, T. M. Swager, P. T. Hammond, Adv. Func. Mater. 2014, 24, 492.

[14] Y. Yang, H.-F. Ji, T. Thundat, J. Am. Chem. Soc. 2003, 125, 1124.

[15] C. Hartmann-Thompson, J. Hu, S. N. Kaganove, S. E. Keinath, D. L. Keeley, P. R. Dvornic, Chem. Mater. 2004, 16, 5357.

[16] L. Chen, D. Wu, J. Yoon, ACS Sens. 2018, 3, 27.

[17] Y. J. Jang, K. Kim, O. G. Tsay, D. A. Atwood, D. G. Churchill, Chem. Rev. 2015, 115, PR1.

[18] S. O. Obare, C. De, W. Guo, T. L. Haywood, T. A. Samuels, C. P. Adams, N. O. Masika, D. H. Murray, G. A. Anderson, K. Campbell, K. Fletcher, Sensors 2010, 10, 7018.

[19] S. Royo, R. Martínez-Máñez, F. Sancenón, A. M. Costero, M. Parra, S. Gil, Chem. Commun. 2007, 46, 4839.

[20] M. Burnworth, S. J. Rowan, C. Weder, Chem. Eur. J. 2007, 13, 7828.

[21] D. Knapton, M. Burnworth, S. J. Rowan, C. Weder, Angew. Chem. Int. Ed. 2006, 45, 5825.

[22] Y.-C. Cai, C. Li, Q.-H. Song, ACS Sens. 2017, 2, 834.

[23] J. Yao, Y. Fu, W. Xu, T. Fan, Y. Gao, Q. He, D. Zhu, H. Cao, J. Cheng, Anal. Chem. 2016, 88, 2497.

[24] S. Huang, Y. Wu, F. Zeng, L. Sun, S. Wu, J. Mater. Chem. C 2016, 4, 10105.

[25] Y. Fu, J. Yu, K. Wang, H. Liu, Y. Yu, A. Liu, X. Peng, Q. He, H. Cao, J. Cheng, ACS Sens. 2018, 3, 1445.

[26] S. Bencic-Nagale, T. Sternfeld, D. R. Walt, J. Am. Chem. Soc. 2006, 128, 5041.

[27] T.-I Kim, S. B. Maity, J. Bouffard, Y. Kim, Anal. Chem. 2016, 88, 9259.

[28] S.-W. Zhang, T. M. Swager, J. Am. Chem. Soc. 2003, 125, 3420.

[29] E. P. Lloyd, R. S. Pilato, K. A. V. Houten, ACS Omega 2018, 3, 16028.

[30] T. J. Dale, J. Jr. Rebek, J. Am Chem. Soc. 2006, 128, 4500.

[31] A. Barba-Bon, A. M. Costero, S. Gil, A. Harriman, F. Sancenón, Chem. Eur. J. 2014, 20, 6339.

[32] R. Gotor, P. Gaviña, L. E. Ochando, K. Chulvi, A. Lorente, R. Martínez-Máñez, A. M. Costero, RSC Adv. 2014, 4, 15975.

[33] R. Gotor, A. M. Costero, P. Gaviña, S. Gil, Dyes Pigments 2014, 108, 76.

[34] G. H. Dennison, M. R. Johnston, Chem. Eur. J. 2015, 21, 6328.

[35] S. Sarkar, A. K. Tiwari, R. Shunmugam, Chem. Commun. 2012, 48, 4223.

[36] G. E. Southard, K. A. Van Houten, E. W. Jr. Ott, G. M. Murray, Anal. Chim. Acta 2007, 581, 202.

[37] A. Wild, A. Winter, M. D. Hager, U. S. Schubert, Chem. Commun. 2012, 48, 964.

[38] W. A. Maza, C. M. Vetromile, C. Kim, X. Xu, X. P. Zhang, R. W. Larsen, J. Phys. Chem. A 2013, 117, 11308.

[39] G. H. Dennison, C. Curty, A. J. Metherell, E. Micich, A. Zaugg, M. D. Ward, RSC Adv. 2019, 9, 7615.

[40] A. J. Metherell, C. Curty, A. Zaugg, S. T. Saad, G. H. Dennison, M. D. Ward, J. Mater. Chem. C 2016, 4, 9664.

[41] S. Jo, D. Kim, S.-H. Son, Y. Kim, T. S. Lee, ACS Appl. Mater. Interfaces 2014, 6, 1330.

[42] H. Jiang, P. Wu, Y. Zhang, Z. Jiao, W. Xu, X. Zhang, Y. Fu, Q. He, H. Cao, J. Cheng, Anal. Methods 2017, 9, 1748.

[43] W. Xiong, Y. Gong, Y. Che, J. Zhao, Anal. Chem. 2019, 91, 1711.

[44] Z. Lu, W. Fan, X. Shi, C. A. Black, C. Fan, F. Wang, Sens. Actuator B 2018, 255, 176.

[45] J. Epstein, R. E. Plapinger, H. O. Michel, J. R. Cable, R. A. Stephani, R. J. Hester, C. Billington, G. R. List, J. Am. Chem. Soc. 1964, 86, 3075.

[46] F. Terrier, P. Rodriguez-Dafonte, E. L. Guével, G. Moutiers, Org. Biomol. Chem. 2006, 4, 4352.

[47] T. J. Dale, J. Jr. Rebek, Angew. Chem. Int. Ed. 2009, 48, 7850.

[48] Y.-C. Cai, C. Li, Q.-H. Song, J. Mater. Chem. C 2017, 5, 7337.

[49] I. Walton, M. Davis, L. Munro, V. J. Catalano, P. J. Cragg, M. T. Huggins, K. J. Wallace, Org. Lett. 2012, 14, 2686.

[50] Y. J. Jang, O. G. Tsay, D. P. Murale, J. A. Jeong, A. Segev, D. G. Churchill, Chem. Commun. 2014, 50, 7531.

[51] Y. Kim, Y. J. Jang, S. V. Mulay, T.-T. T. Nguyen, D. G. Churchill, Chem. Eur. J 2017, 23, 7785.

[52] L. Chen, H. Oh, D. Wu, M. H. Kim, J. Yoon, Chem. Commun. 2018, 54, 2276.

[53] D. Pangeni, E. E. Nesterov, Macromolecules 2013, 46, 7266.

[54] M. S. I. Khan, Y.-W. Wang, M. O. Senge, Y. Peng, J. Hazard. Mater. 2018, 342, 10.

[55] D. H. Rosenblatt, M. J. Small, T. A. Kimmell, A. W. Anderson, Background chemistry for chemical warfare agents and decontamination processes in support of delisting waste streams at the U.S. Army Dugway Proving Ground, Utah. OSTI, United States 1996, pp. 8-35. doi:10.2172/258187

[56] Y. C. Yang, Acc. Chem. Res. 1999, 32, 109.

[57] C. Wilson, N. J. Cooper, M. E. Briggs, A. I. Cooper, D. J. Adams, Org. Biomol. Chem. 2018, 16, 9285.

[58] B. Picard, I. Chataigner, J. Maddaluno, J. Legros, Org. Biomol. Chem. 2019, DOI:10.1039/C9OB00802K.

[59] T. Islamoglu, M. A. Ortuño, E. Proussaloglou, A. J. Howarth, N. A. Vermeulen, A. Atilgan, A. M. Asiri, C. J. Cramer, O. K. Farha, Angew. Chem. 2018, 130, 1967.

[60] B. Picard, I. Chataigner, J. Maddaluno, J. Legros, Org. Biomol. Chem. 2019, 17, 6528.

[61] T. M. Alam, M. K. Kinnan, B. W. Wilson, D. R. Wheeler, ChemistrySelect 2016, 1, 2698.

[62] J. Gao, S. X. Naughton, H. Wulff, V. Singh, W. D. Beck, J. Magrane, B. Thomas, N. A. Kaidery, C. M. Hernandez, A. V. Terry, J. Pharmacol. Exp. Ther. 2016, 356, 645.

[63] D. R. Heiss, D. W. Zehnder II, D. A. Jett, G. E. Platoff Jr., D. T. Yeung, B. N. Brewer, J. Chem. 2016, 3190891.

[64] G. H. Posner, J. W. Ellis, J. Ponton, J. Fluor. Chem. 1981, 19, 191.

[65] K. J. Wallace, R. I. Fagbemi, F. J. Folmer-Andersen, J. Morey, V. M. Lynth, E. V. Anslyn, Chem. Commun. 2006, 45, 3886.

[66] Y. Urano, D. Asanuma, Y. Hama, Y. Koyama, T. Barrett, M. Kamiya, T. Nagano, T. Watanabe, A. Hasegawa, P. L. Choyke, H. Kobayashi, Nat. Med. 2009,15, 104.

[67] J. Chao, H. Wang, K. Song, Z. Li, Y. Zhang, C. Yin, F. Huo, J. Wang, T. Zhang, Tetrahedron 2016, 72, 8342.

[68] J. Chao, Y. Liu, J. Sun, Li Fan, Y. Zhang, H. Tong, Z. Li, Sens. Actuator B 2015, 221, 427.

[69] S. El Sayed, L. Pascual, A. Agostini, R. Martínez-Máñez, F. Sancenõn, A. M. Costero, M. Parra, S. Gil, ChemistryOpen 2014, 3, 142.

[70] R. Puglisi, A. Pappalardo, A. Gulino, G. T. Sfrazzetto, ACS Omega 2019, 4, 7550.

[71] R. Puglisi, A. Pappalardo, A. Gulino, G. T. Sfrazzetto, Chem. Commun. 2018, 54, 11156.

[72] X. Liu, Y. Gong, Y. Zheng, W. Xiong, C. Wang, T. Wang, Y. Che, J. Zhao, Anal. Chem. 2018, 90, 1498.

[73] B. D. de Greñu, D. Moreno, T. Torroba, A. Berg, J. Gunnars, T. Nilsson, R. Nyman, M. Persson, J. Pettersson, I. Eklind, P. Wásterby, J. Am. Chem. Soc. 2014, 136, 4125.

[74] K. Chulvi, P. Gaviña, A. M. Costero, S. Gil, M. Parra, R. Gotor, S. Royo, R. Martínez-Máñez, F. Sancenón, J.-L. Vivancos, Chem. Commun. 2012, 48, 10105.

[75] M. A. Ali, S. Shoaee, S. Fan, P. L. Burn, I. R. Gentle, P. Meredith, P. E. Shaw, ChemPhysChem 2016, 17, 3350.

[76] P. E. Shaw, P. L. Burn, Phys. Chem. Chem. Phys. 2017, 19, 29714.

[77] K. A. Van Houten, D. C. Heath, R. S. Pilato, J. Am. Chem. Soc. 1998, 120, 12539.

[78] J. G. Weis, T. M. Swager, ACS Macro Lett. 2015, 4, 138.

Ian Gentle is Professor of Physical Chemistry in the School of Chemistry and Molecular Biosciences and a member of the Centre for Organic Photonics and Electronics at The University of Queensland (UQ). His previous positions include Deputy Executive Dean and Associate Dean (Research) in the Faculty of Science at UQ and Head of Science at the Australian Synchrotron. His research interests

are the morphological aspects of materials for optoelectronic devices, including sensors and organic light emitting diodes.

Paul Shaw received his PhD in Physics from the University of St Andrews in 2009 on the topic of exciton diffusion in conjugated polymers. Since then he has worked at The University of Queensland. His research interests lie in the application of optical spectroscopy to understand the properties of organic semiconductors and the development of sensing technologies based on fluorescent compounds.

TOC entry:

Fluorescence-based detection is a sensitive and rapid method for detecting and identifying potential threats such as nerve agents. Due to the highly toxic nature of the agents themselves, studies to develop sensors have normally involved simulants, however care is needed when interpreting the results of studies using common simulants.