Fluorenone-Based Fluorescent and Colorimetric Sensors for Selective Detection of I- Ions: Applications in HeLa Cell Imaging and Logic Gate.

Fluorenone-based fluorescent and colorimetric sensors 1 and 2 have been developed that displayed selective detection of iodide ions in the presence of interferences. Sensors displayed the fluorescence emission enhancement response toward I- with detection limits of 8.0 and 11.0 nM, respectively, which is accomplished through inhibition of intramolecular charge transfer and C=N isomerization. Excellent sensitivity and unique fluorescence enhancement response of sensors toward I- make them superior because most of the previously reported iodide sensors are based on the fluorescence quenching mechanism and are less sensitive. The sensing potential of sensors toward I- ions was investigated through 1H NMR titration, dynamic light scattering, Job's plots, and density functional theory analysis. Further, sensors displayed reversible behavior by the alternate addition of I- and Cu2+ ions that substantiate their role as recyclable sensors for the on-site detection of I- ions. Advantageously, fluorescence enhancement response of sensors was favorably used for fluorescence imaging of I- in live HeLa cells and the design of the logic gate. These sensors were successfully applied in diversified applications such as the preparation of sensors' coated paper strips and the determination of I- ions in blood serum, food, and real water samples.

Introduction

Anions are considered fundamental to many biological and chemical operations since various anionic species are involved in the progress of numerous intracellular processes.[1,2] Among essential anions, iodide ions play a vital role in the normal growth of human beings, and its recommended dose is 80–150 μg/day.[3] Iodotyrosine dehalogenase and sodium iodide cotransporter are two necessary plasma proteins that are used to balance the malfunctioning of dietary iodine.[4] In humans, malfunctioning of dietary iodine or inefficiency of thyroid hormone-binding proteins can cause serious health hazards such as hypothyroidism, goiter, and hyperthyroidism.[4] World Health Organization (WHO) has recommended 100 μg/L (∼800 nM) median urinary iodine concentrations in humans.[5] Despite its biological significance, iodine/iodide ions may also cause toxicity, as 129I is a long-lived (half-life; 1.6 × 107 years) radioactive isotope of iodine that is released into the ecosystem as a result of nuclear weapon testing.[6] Keeping this in consideration, the sensing of I– ions both inside and outside of the cells has earned paramount significance. However, the binding capacity of I– ions with organic receptors is weakest due to its less basicity, low charge density, and larger anionic size among halide ions that make it quite challenging to develop receptors for selective sensing of I– ions. In recent past, I– ions have been detected through numerous analytical methodologies such as electrochemical methods,[7−9] capillary electrophoresis,[10] gas chromatography–mass spectrometry,[11] and indirect atomic absorption spectroscopy.[12] These methods suffer from various drawbacks such as high cost, incompatibility in aqueous medium, multistep sample preparation, and lack of selectivity that made these techniques tedious, time-consuming, and cost-consuming.[13,14] Apart from these methods, the fluorescence technique provides simplicity, easy availability, routine subnanomolar detection, and sensitivity for the detection of a suitable analyte. Various fluorescence-based chemosensors have been exploited for the detection and quantification of I– ions. Silver-based ratiometric fluorescent complex was developed by Wang et al. for the sensitive detection of I– ions with the sensitivity level down to 7.16 10–6 M.[15] Similarly, Rastegarzadeh and colleagues used a redox reaction-based optical sensor for the detection of I– ions with a detection limit of 7.44 × 10–7 M.[16] Many other conjugated polymers and small organic molecules have also been reported for sensing iodide ions. However, most of the reported iodide sensors are based on the fluorescence quenching response.[17−19] Kim et al. synthesized bis-imidazolium ring-substituted naphthalene compounds with a concave cavity that was found suitable to encapsulate spherical halides, specifically iodide ions.[20] A benzimidazole-based novel tripod receptor was synthesized by Lee et al. and was found highly selective toward the detection of I– due to its trap in the pseudocavity of the fluorescent receptor.[21] Vetrichelvan and colleagues devised a series of carbazole-functionalized conjugated polymers that displayed fluorescent and colorimetric recognition of iodide due to the intermolecular charge-transfer (ICT) complex between the carbazole units of the polymer backbone and I– ions.[22] Further, the Li group modified cadmium selenide quantum dots through surface functionalization of thiourea moieties and used these fluorescent sensors for the detection of iodide ions based on hydrogen bonding between the sensor and I–.[23] In contrast to quenching, the fluorescence enhancement response is superior because it minimizes the probability of false-positive signals by co-existing quenchers and thereby increases the sensitivity of fluorophore.[24] Consequently, the design of fluorescence enhancement-based I– sensors is quite challenging because I– ions hold an intrinsic fluorescence quenching ability owing to their heavy atom effect. Usually, heavy atom effect is considered as a fundamental reason for the fluorescence quenching response in most of the reported turn-off I– ions sensors.[25] Therefore, it is imperative to develop highly sensitive fluorescent materials that can detect I– ions on the basis of fluorescence-enhanced emission response.

As a part of our ongoing studies on the development of fluorescent sensors,[26−39] two fluorenone-based sensors 1 and 2 were designed and synthesized for the sensitive detection of I– ions based on fluorescence enhancement to avoid its intrinsic fluorescence quenching nature. The design of sensors 1 and 2 plays a vital role in their fluorescence enhancement response for I– ions. Donor–acceptor (D–A)-type architecture was established through the introduction of hydroxyl/s substituted aromatic rings to the fluorenone core through the C=N linker. Basically, sensors 1 and 2 were designed based on two strategies. First, the C=N linker unit may efficiently involve establishing a complex with an anion. Second, formation of pseudocavity was anticipated to encapsulate large anions such as I– ions. Moreover, D–A-type structures of sensors 1 and 2 activate the ICT and excited-state intramolecular proton transfer (ESIPT) processes that make them less/non emissive. However, the addition of I– ions inhibits the ICT and ESIPT pathways along with the restriction of the −HC=N– isomerization that provides strong rigidity to sensor molecules. Structural rigidity results in fluorescence enhancement of sensors 1 and 2 in the presence of I– ions. To the best of our knowledge, this is the first report in which fluorenone Schiff base-based sensors are used for selective recognition of I– ions based on fluorescence enhancement. Further, a large Stokes shift, detection of I– in the cellular level and in aqueous medium, excellent stability of Schiff base-based sensors, facile synthesis, less synthetic cost, and unique fluorescence enhancement pattern provide these sensors extra advantages over previously reported fluorescent molecules. Advantageously, a unique fluorescence enhancement pattern was used to explore their practical applications such as live cell imaging, design of logic circuits, preparation of sensors’ coated paper strips, and quantification of iodide ions in blood serum, food, and real water samples.

Results and Discussion

Synthesis and Structural Characterization

Fluorenone-based sensors 1 and 2 were afforded through an easily accessed three-step synthetic strategy as given in Scheme . Nitration of compound (i) followed by SnCl2-catalyzed reduction produced amino-substituted compound (iii) with an excellent 90% yield. Then, the desired sensors 1 and 2 were synthesized by treating intermediate (iii) with 2-hydroxybenzaldehyde and 2,3-dihydroxybenzaldehyde, respectively, in methanol. 1H and 13C NMR spectroscopies were carried out for structural characterization of compounds (ii) and (iii). Sensors 1 and 2 were characterized through NMR (1H and 13C), FT-IR spectroscopy, and mass spectrometry techniques. All the spectra are provided in the Supporting Information (Figures S22–S33, p. S20–S64). The nitro group in compound (ii) was validated through the appearance of de-shielded signals at δ 8.49 and 8.44 ppm in 1H NMR by its two adjacent protons. Additionally, 13C NMR displayed 13 peaks that justify the synthesis of compound (ii). Moreover, a broad singlet at δ 3.91 ppm supports the reduction of the nitro group into an amine unit to provide compound (iii). The appearance of 13 separate peaks in 13C NMR is the confirmation of compound (iii). Further, a hydroxyl proton signal at δ 12.99 ppm in 1H NMR and appearance of 20 separate peaks in the 13C NMR spectrum support successful synthesis of sensor 1. Likewise, a hydroxyl proton signal at δ 8.68 ppm in 1H NMR and well-resolved separate 20 carbon peaks in the 13C NMR spectrum provide evidence for the formation of sensor 2. Moreover, sensors 1 and 2 were stable at room temperature in their solid and solution state. Structures of easily synthesizable sensors 1 and 2 were proposed to tune their fluorescence sensing properties for easy detection of biologically important analytes. Sensors 1 and 2 possess C=N isomerization and ICT that are likely to be disturbed upon interaction with anions.

Scheme 1

Synthetic Methodology to Afford Sensors 1 and 2

Optimization of Concentration, Solvent System, pH, and Response Time

The emission (λem) and absorption (λmax) of a fluorophore could be influenced by the polarity and viscosity of the solvent.[40] In this perspective, the effect of a solvent on the absorption and emission properties of sensors 1 and 2 (40 μM) was investigated by scanning their UV–visible (UV–vis) absorption and fluorescence emission spectra in a library of solvents (Figure S1). The absorption and emission spectra of sensors exhibited excellent stability in all tested solvents (Figure S1). Any diversion noted in their absorption and emission wavelength (λmax) is referred to its dissolution difference in that particular solvent. Sensors 1 and 2 displayed exquisite absorption and emission in tetrahydrofuran (THF) solvent. Moreover, the Schiff base-based compounds are usually considered vulnerable to aqueous solution that hampers their photophysical properties. Surprisingly, the emission wavelength (λmax) and intensity of sensors 1 and 2 (40 μM) remained unaltered upon mixing different water fractions (fw, 0–70%) to their THF solution. It reveals that sensors 1 and 2 render the highest emission in H2O/THF (7:3, v/v) with no aggregation at increased water content (Figure S2). Significant emission of sensors 1 and 2 in aqueous solution (H2O/THF (7:3, v/v) prompted us to evaluate their stability in the aforementioned solvent system. Fluorescence emission of sensors 1 and 2 remained persistent at 511 and 561 nm, respectively, for 5 days as tested after regular intervals of time (Figure S3). Stable fluorescence emission of sensors 1 and 2 was also supported by their zeta potential analysis in H2O/THF (7:3, v/v) and recorded as 13.51 and 14.23 mV, respectively, that corresponds to their excellent stability in aqueous solution (Figure S4).

Relying on the aforementioned experimental results, the H2O/THF (7:3, v/v) solvent system was chosen for successive sensing studies. Further, the fluorescence emission of sensors 1 and 2 was recorded at their different concentrations ranging from 0 to 200 μM. The fluorescence emission results presented in Figure S5 showed that their emission intensity remained unaffected at all concentrations except 40 μM at which sensors 1 and 2 exhibited significantly high fluorescence emission centered at 511 and 561 nm, respectively (excitation wavelength (λexc), 310 and 325 nm). Furthermore, other experimental parameters including response time, pH, and buffer system were also optimized (Figures S6 and S7). After a detailed analysis, a 0.2 M Britton–Robinson (B–R) buffer system was selected for better sensing performance of sensors 1 and 2 (Figure S6a). It is well understood that any fluctuation in pH may affect the sensing potential of a sensor. In this regard, fluorescence experiments were performed at different pH values varying from 2.0 to 12.0 in the Britton–Robinson (B–R) buffer system (0.2 M) to scrutinize suitable pH for better sensing performance of sensors 1 and 2. Appreciable fluorescence emission of sensors 1 and 2 was observed in the range of pH 7.0–10.0 (Figure S6b). An optimum working pH in the neutral range makes these sensors a suitable candidate for the detection of biologically important analytes in environmental samples. Next, response time is another significant parameter that estimates the sensitivity of a sensor. Fluorescence experiments of sensors 1 and 2 (40 μM) in the presence of I– ions (30 nM) were performed to investigate their response time. The relative fluorescence signals were observed after regular intervals of time (0–60 s), maintaining other experimental parameters constant (Figure S7). A significant enhancement in the fluorescence emission of sensors 1 and 2 was noticed at 20 and 25 s that suggests their immediate sensing response toward I– ions.

Photophysical Characteristics of Sensors 1 and 2

Before investigating sensing studies, the UV–vis and fluorescence emission spectra of sensors 1 and 2 were scanned in H2O/THF (7:3, v/v). The well-resolved absorption bands of sensors 1 and 2 were obtained at 351 and 391 nm, respectively, that are attributed to n→π* transitions (Figure S8). Similarly, sensors 1 and 2 displayed weak fluorescence emission at 511 and 561 nm with Stokes shifts of 160 and 170 nm, respectively (Figure S9). Weak fluorescence emission and high Stokes shift in sensors 1 and 2 are ascribed to the involvement of the ICT process (Figure S10). The fluorescence emission of sensor 2 is red-shifted (40 nm) with a slightly higher emission in comparison to sensor 1. The fluorescence emission at longer wavelength is presumably due to the installation of an additional hydroxyl group in sensor 2, and its enhanced emission intensity is probably due to the intramolecular hydrogen bonding in its two vicinal hydroxyl groups. It brings mild rigidity in the structure of sensor 2 that leads to marginally improved fluorescence emission as compared to sensor 1. Moreover, quantum yield (Φ) of sensors 1 and 2 was calculated as 0.58 and 0.55, respectively, which was calculated through the reported method[41] taking fluorescein as a standard. The HOMO–LUMO gaps of sensors 1 and 2 were computed through density functional theory (DFT) analysis and were found to be 0.92 and 1.49 eV. The photophysical properties of sensors 1 and 2 are summarized in Table .

Table 1

Photophysical Characteristics of Sensors 1 and 2

	UV–vis absorption profile (nm)		fluorescence emission profile (nm)
compound	λexc	λabs	λexc	λem	quantum yield (Φ)	HOMO–LUMO gap (eV)
sensor 1	255	351	310	511	0.58	0.92
sensor 2	273	391	325	561	0.55	1.49

Fluorescence Enhancement-Based Detection of I–

Selectivity Experiments of Sensors 1 and 2

Sensors 1 and 2 were designed to accomplish fluorescence emission enhancement-based detection of biologically and environmentally important anions. It is obvious that imine (C=N) and hydroxyl group/s in sensors are crucial moieties to interact with an anion. Keeping these aspects in mind, weakly emissive sensors 1 and 2 (40 μM) were treated with a variety of anions, including F–, Cl–, Br–, I–, BrO3–, HSO4–, S2O3–, S–2, SO3–2, HCO3–, NO2–, AcO–, CN–, PO4–3, H2PO4–, and OH– ions (30 nM), to evaluate their sensing potential toward a specific anion in aqueous solution (H2O/THF (7:3, v/v). The fluorescence results presented in Figure displayed a drastic increase in the fluorescence emission of sensors 1 and 2 only against I– ions. All other tested anions did not induce any appreciable change in the emission of both sensors (Figure ). The high hydration energy of other anions was incompatible to the aqueous medium employed for sensing studies, and hence, no appreciable changes in the fluorescence spectra of 1 and 2 were observed. Further, an anti-interference experiment was performed with the co-existing ions (20 equiv.) and I– ions in H2O/THF (7:3, v/v) to evaluate the practical applicability and specificity of sensors 1 and 2. It is evident from Figures S11 and S12 that no interruption in the sensing potential of both sensors for I– ions was noticed in the presence of all tested interferences. These results clearly demonstrate the excellent selectivity of sensors for I– ions.

Figure 1

Fluorescence sensing of sensors 1 (a) and 2 (b) against different anions in H2O/THF (7:3, v/v) at excitation wavelengths (λexc) of 310 and 325 nm, respectively (slit width, 2/2).

Binding Assay of Sensors with I–

The binding interaction of sensors 1 and 2 with I– was evaluated through fluorescence and UV–vis spectroscopy. The fluorescence emissions of sensors 1 and 2 (40 μM) at 511 and 561 nm were enhanced upon progressive addition of I– ions (0–30 nM) in H2O/THF (7:3, v/v) (Figure ). The fluorescence emission enhancement in sensors 1 and 2 was accompanied with a prominent hypsochromic shift of 15 and 10 nm, respectively. The percent enhancement efficiency of sensors 1 and 2 was calculated through (I – I0/I) × 100 and found to be 100 and 78%, respectively, at the highest concentration of I– ions (30 nM). These results indicate that binding between sensors (1 and 2) and I– ions may inhibit the ICT and ESIPT processes that consequently favor significant fluorescence enhancement. Advantageously, sensors 1 and 2 exhibited light yellow emission under UV (365 nm) upon addition of I– ions (20 nM) (insets; Figure ). Further, the Benesi–Hildebrand (B–H) plots presented in Figure show that the fluorescence emission of sensors 1 and 2 changes linearly in the presence of increasing concentrations of I– ions (0–30 nM) with correlation coefficients (R2) of 0.999 and 0.9989, respectively. The strong association indicated by the linear behavior of the B–H plots was supplemented by calculating the association constant (Ka) through the Benesi–Hildebrand equation and found to be 4.21 × 106 and 4.08 × 106, respectively. The lower Ka value of sensor 2 is probably due to the intramolecular hydrogen bonding between its two neighboring hydroxyl groups that somewhat hinder its complexation with I– ions. Further, the calculated Ka values of sensors 1 and 2 are much higher than that reported for I– sensors that demonstrate their excellent specificity and sensitivity for I– (Table ). Additionally, the sensitivity of sensors 1 and 2 was estimated by calculating the limit of detection (LOD) through 3δ/S, where δ represents the standard deviation and S is the linear calibration curve in a straight-line plot (Figure S13). Consequently, the LOD of sensors 1 and 2 for I– ions was calculated to be 8 and 11 nM, respectively. The calculated LOD values are much better than that of previously reported iodide sensors as summarized in Table .

Figure 2

Variations in the fluorescence emission of sensors 1 (a) and 2 (c) upon incremental addition of I– ions (0–30 nM) in H2O/THF (7:3, v/v) (excitation wavelength (λexc), 310 and 325 nm, respectively) (slit width, 2/2). The Benesi–Hildebrand plots showing the linear relationship between sensors 1 (b) and 2 (d) and varying concentration of I– ions (0–30 nM).

Table 2

Comparison of Different Features of Sensors 1 and 2 with Previously Reported Iodide Ion (I–) Sensors

sensing technique	sensors	response time (sec)	LOD (nM)	Ka (M–1)	pH	applications	refs
fluorescence enhancement	fluorenone-based sensors 1 and 2	20 and 25	8 and 11	4.21 × 106 4.08 × 106	7–10	HeLa cell imaging, logic gate, paper strips, and others	This work
electrochemical	ion-exchange chromatography		3900			milk and wastewater	(42)
fluorescence turn-off response	carbon nanodots		180		7.2	HeLa cells and human serum	(43)
flow injection chemiluminescence	KMnO4 carbon dots system		350		3–8	food material	(44)
UV–vis	polymer-capped silver nanoparticles		1.08 × 10–4			river, tap, and pond water	(45)
fluorescence			0.3				(46)
fluorescence	silver nanoclusters and carbon dots		19.8			human urine
fluorescence turn-off	porphyrin-based sensor	20	180	1.9 × 104			(47)
fluorescence turn-off response	naphthalene-based sensor	0.8	30	0.8 × 104		testing kits	(48)
fluorescence	N-doped C-dots with turn off–on response		60			urine samples	(49)
fluorescence turn off–on	1,10-phenanthroline-based sensors		3000	1.05 × 103			(50)
fluorescence turn off–on	anthryl-appended porphyrin-based sensors	60	33			table salt	(51)
potentiometry-based iodide sensor	cetylpyridiniumtetraiodomercurate-based sensors	20	3000		3–8		(52)
UV–vis	PVC-based liquid membrane	120	5300		2		(53)
fluorescence turn off–on	naphthyl boronic acid-based iodide sensors		5200	8.0 × 103			(54)

Efficient binding of sensors 1 and 2 with I– ions was further elaborated through UV–vis absorption studies under the same experimental parameters. In this regard, upon the addition of I– ions (0–30 nM), the UV–vis absorption of sensors 1 and 2 at 351 and 391 nm displayed a significant hyperchromic shift along with the appearance of completely new absorption bands at 384 and 424 nm, respectively (Figure ). The interaction of sensors 1 and 2 with other analytes displayed an insignificant change in their absorption spectra (Figure S14). The appearance of a new absorption peak is presumably attributed to the blockage of ICT and ESIPT processes in sensors 1 and 2 in the presence of I– ions. UV–vis absorption changes are also associated with a change in the light-yellow color of sensors 1 and 2 to reddish and light pink, respectively (insets; Figure ). Furthermore, linear fit plots presented in the insets of Figure reveal the strong association between sensors and I– ions. The binding association (Ka) of sensors 1 and 2 calculated through linear fit plots was found to be 4.11 × 106 and 3.98 × 106 M–1, respectively, which is in good agreement with the Ka values calculated through fluorescence experiments. Conclusively, the fluorescence and UV–vis absorption experiments clearly demonstrate the preferred sensitivity and selectivity of sensors for iodide ions.

Figure 3

Changes in the UV–vis absorption of sensors 1 (a) and 2 (b) upon incremental addition of I– ions (0–30 nM) in H2O/THF (7:3, v/v) (slit width, 2/2); insets: the linear fit plots showing the linear relation between sensors 1 (a) and 2 (b) and varying concentration of I– ions (0–30 nM).

The preferential selectivity and thermodynamic stability of sensors 1 and 2 toward I– ions were also validated through DFT analysis. Better interaction energy of a fluorophore with a specific analyte is rationale to its substantial selectivity and sensitivity for that particular analyte.[55] In this regard, the interaction energies of sensors and sensor–ion complexes were calculated through the Gaussian 09 program using the B3LYP/6-31G(d) methodology.[56,57] The maximum interaction energies (Eint) of sensors 1 and 2 were calculated for their complex with I– ions and were found to be −218.53 and −192.28 Kcal/mol, respectively. Hence, theoretically calculated greatest interaction energies of I–@1 and I–@2 justify excellent selectivity of sensors 1 and 2 for I– ions among all other tested anions. Optimized complexes of 1 and 2 with selective I– are presented in Figure S15.

Plausible Sensing Mechanism of Sensors with I– Ions

Sensors 1 and 2 possess two regions in their Schiff-based structures, one fluorenone core as a fluorophore and the hydroxyl-substituted aromatic unit/s as a receptor. The electron-donating hydroxyl unit/s persuade electronic transitions from the receptor to the fluorophore unit through the linker imine functionality (−C=N−). It consequently activates the ICT and ESIPT processes. However, the addition of iodide ions to sensor suppresses the ICT and ESIPT pathways along with the restriction of −CH=N– isomerization that provides strong rigidity to sensor molecules. Structural rigidity, as a consequence, results in fluorescence enhancement of sensors in the presence of I– ions.[58,59] A general representation of the proposed mechanism is given in Figure . In principle, the ICT process only occurs when the highest occupied molecular orbital (HOMO) level of the receptor lies at a higher energy in comparison to the HOMO of the fluorophore unit (Figure S16). The complexation of iodide ions with sensors is anticipated to lower the HOMO of the bound receptor in comparison to the HOMO of the excited fluorophore and inhibits the ESIPT and ICT pathways that ultimately trigger fluorescence emission response (Figure b).

Figure 4

General representation of C=N–I– and OH–I– complex between sensors and I– ions (a) and plausible mechanism of I– ion detection (b).

Generally, the formation of hydrogen bonds between sensors and I– ions is pivotal to their strong complexation. However, hydrogen bonding alone was not responsible for such an efficient and specific fluorescence enhancement response toward iodide ions as other anions with similar hydrogen bonding capabilities would behave similarly. Preferential detection of I– ions demonstrated that the pseudocavity generated in both sensors is more compatible to encapsulate only I– ions due to its unique adjustable size. Several other reports have already been published that govern that the size of the halide ions plays a potential role to adjust in the sensors’ binding cavities in contrary to their basicity and H-bonding ability.[60−62]

The DFT calculations were performed using Gaussian 09 software to support these assumptions. Frontier molecular orbital analysis and geometries of sensors 1 and 2 and sensor–I– complexes were optimized at the B3LYP/6-31G (d) level of theory.[63] The HOMO–LUMO surfaces of sensors 1 and 2 are depicted in Figures and S17. It is distinctly explained that in the HOMO level, electronic density was dispersed over the whole sensor molecule, and in the LUMO energy level, the electron density was only limited to the fluorenone core of a sensor. Spread of electronic cloud over the whole sensor corresponds to extended π conjugation. Complexation of I– to a sensor (1 or 2) restricts the electronic density to the fluorenone core that is an indication of disruption in π conjugation (ICT process). Furthermore, the H–L gap of sensors 1 and 2 decreased notably from 4.56 to 0.92 and 4.87 to 1.49 eV, respectively, in the presence of iodide, which accounts for their strong complexation with I– ions (Figures and S17).

Figure 5

Optimized structure of sensor 1 (a), HOMO–LUMO orbitals of sensor 1 (b), and complex of sensor 1 with I– (c).

Complexation of sensors with I– ions was confirmed through 1H-NMR titration. In this regard, 1H-NMR of sensor 2 (as a model example) in DMSO-d6 indicates that a singlet peak appeared at δ 9.0594 corresponds to its hydroxyl (−OH) proton (Ha). Similarly, a singlet peak displayed at δ 7.6957 ppm corresponds to the imine proton (−N=C–H, Hd). Other signals at δ 7.8814, 7.8278, 7.6405, 7.2967, 7.1342, 6.9660, and 6.8081 ppm represent the aromatic protons Hb, Hc, He, Hf, Hg, Hh, and Hi, respectively. The addition of 0.5 equiv of TBAI to sensor 2 induced a notable downfield shift in the hydroxyl proton (Ha) from δ 9.0594 to 9.0634 ppm and an up-field shift in the imine proton (Hd) from δ 7.6957 to 7.6913 ppm (Figure ). Moreover, the mixing of 1.0–2.0 equiv of TBAI to sensor 2 prompted further shift in −OH proton from 9.0634 to 9.0894 ppm. Similarly, a significant up-field shift in the imine proton (Hd) from δ 7.6913 to 7 7.6718 ppm was recorded in the presence of 2 equiv of iodide salt (TBAI). Furthermore, significant shifts in the other aromatic protons (Hb, Hc, He, Hf, Hg, Hh, and Hi) due to the addition of 0.5–2.0 equiv of TBAI are summarized in Table S1. Hence, in the 1H-NMR spectrum of sensor 2, the downfield shift in −OH proton (Ha) and the up-field shift in imine proton (−C=N–H) proton (Hd) upon addition of 0–2.0 equiv of TBAI confirm the obvious complexation between sensor 2 and I– ions.

Figure 6

1H-NMR titration spectrum of sensor 2 with progressive addition of TBAI (0–2.0 equiv) in DMSO-d6 (400 MHz).

The 1H-NMR titration assay proposed the maximal interaction between the sensor and 2.0 equiv of TBAI, which validates their 1:2 binding stoichiometry. Host–guest stoichiometric complexation plays a crucial part to probe out the sensitivity of sensors for I– ions. In this context, binding stoichiometry of sensors 1 and 2 with I– ions was further evaluated through Job’s plot study (Figure ). The maximum enhancement in the emission intensity was observed at the 0.6 mol fraction of I– ions, which implies the 1:2 binding association of sensors 1 and 2 with I– ion (Figure ). Spectral shifts in the 1H-NMR titration experiment and 1:2 complexation ions clearly demonstrate stable and strong complex formation between sensors and I– ions.

Figure 7

Job’s plot variation exhibiting 1:2 binding stoichiometry of sensors 1 (a) and 2 (b) with I– ions.

The binding interaction was further supported by dynamic light scattering (DLS) analysis. It is likely that the size of sensors 1 and 2 would alter upon complexation with I–. In this regard, the size of sensors (1 and 2) and sensor–I– was determined in THF/H2O (3:7, v/v) (Figure ). The estimated size of sensors 1 (40 μM, PDI = 0.451) and 2 (40 μM, PDI = 0.566) was found to be 122 and 138 nm, respectively. Figure reveals that the particle size of sensors 1 and 2 distinctly increased to 191 nm (PDI = 0.586) and 223 nm (PDI = 0.673), respectively, which undoubtedly favors sensor–I– complexation.

Figure 8

Average particle size of sensors 1 and 2 before (a,c, respectively) and after treating with I– ions (b,d, respectively).

Reversibility Response of Sensors 1 and 2

Reversibility is an exceptional feature of a fluorophore that substantiates its real-time applications. The reversibility of sensors 1 and 2 was investigated by introducing Cu2+ ions to sensor–I– solution. It is likely that the mixing of Cu2+ ions to the sensor–I– solution may break this strong complex, resulting in the formation of the CuI2 molecule.[48] Initially, the emission intensity of sensors 1 and 2 (40 μM) progressively increased at 511 and 561 nm, respectively, upon the addition of I– ions (10 nM) along with the appearance of light yellow emission for both sensors under UV radiations (365 nm). Surprisingly, the addition of Cu2+ (10 nM) to the sensor–I– complex resulted in the attainment of original emission of sensors 1 and 2 at 511 and 561 nm, respectively, along with the disappearance of light yellow emission under UV (365 nm). The addition of I– ions (10 nM) to the solution again produced enhanced emission of sensors 1 and 2 along with an illumination of yellow emission under UV radiations (Figures and S18). The reversibility pattern of sensors 1 and 2 was attained thrice with 95% reversibility efficiency. This complexation/decomplexation pattern reveals that sensors 1 and 2 can be potentially employed for practical sensing of I– ions as “ON–OFF–ON” fluorescence reversal units.

Figure 9

Reversibility cycle of sensors 1 band 2 after alternate addition of I– and Cu2+ (a and b, respectively); images captured under a UV lamp (365 nm) displaying the reversibility pattern in sensors 1 and 2 (c and d, respectively). (The reversibility pattern was repeated thrice and sensors displayed 95% repeatability efficiency).

Practical Applications of Sensors 1 and 2

Fabrication of INHIBIT Logic Gate

In the past few years, the design of an organic system to construct logic gate has drawn considerable attention of researchers owing to its extensive applications in electronic and digital devices.[64] The selective analyte detected by a sensor is considered an input signal and the resulting fluorescence response is referred to as an output signal. Usually, “1” and “0” correspond to the fluorescence on/strong emission and off/weak fluorescence emission signal, respectively. The input (I– ions) increased the fluorescence emission of sensors 1 and 2 at 511 and 561 nm, respectively, while the input Cu2+ retrieved the original fluorescence emission of sensors 1 and 2 (Figure S18). The regeneration of original fluorescence signals of sensors 1 and 2 upon treating Cu2+ helped to develop molecular logic operations by employing sensors’ reversible behavior. Such type of logic gate operates through “if A, then B” or “A implies B” principle. In the design of logic gate, I– was introduced as an input-1 and Cu2+ as an input-2. It is proposed that only the input code (1,0) induced fluorescence enhancement response (ON), while all other input combinations including (0,0), (0,1), and (1,1) result in weak emission or no enhancement (OFF) (Figure ). The truth table is generated based on “ON–OFF–ON” emission signals of sensors 1 and 2 upon subsequent addition of I– and Cu2+ versus binary translation (Figure b). Consequently, the output values suggest the construction of an INHIBIT logic gate that extends practical applicability of sensors 1 and 2.

Figure 10

Fluorescence emission changes of sensors 1 and 2 with four combinations of input signals at (λem) 511 and 561 nm, respectively (a), truth table showing the INHIBIT logic operations (b), and representation of INHIBIT logic gate (c).

Biocompatibility of Sensors and Cell Imaging of I– Ions in Live HeLa Cells

The biological application of sensor 1 (as a model example) for imaging of I– ions in living cells was investigated with the HeLa cells. The biocompatibility of a sensor in living cells is crucial for its practical application in the biological environment. In this regard, HeLa cells were incubated with changing concentrations of sensor 1 (0–40 μM) to evaluate its biocompatibility. The MTT assay results presented in Figure S19 reveal that the sensor was completely non-toxic to the cells in the treated concentration range (0–40 μM). Therefore, these results suggest that sensor 1 is biologically compatible and non-toxic to cells and thus can be used for diagnostic purposes. Next, the HeLa cells were incubated with sensor 1 (40 μM) in phosphate buffer with 0.5% dimethyl sulfoxide (DMSO) at 37 °C for 30 min and then incubated with an increasing concentration of I– ions (2.0–8.0 nM) for 2 h. The fluorescence images in Figure illustrate that sensor 1 (40 μM) pre-treated HeLa cells displayed weak light green fluorescence, which gradually turned bright when treated with increasing concentrations of I– ions (2.0–8.0 nM). These results indicate that sensor 1 has the potential for intracellular imaging and detection of I– ions.

Figure 11

Fluorescence images of sensor 1 pre-treated HeLa cells (a) and in the presence of increasing concentrations of I– (b = 2 nM, c = 4 nM, d = 6 nM, and e = 8 nM).

Preparation of Sensors’ Coated Paper Strips

The practical applicability of sensors 1 and 2 toward the detection of I– ions was extended by preparing sensors’ coated paper strips using Whatman-41 filter paper. Initially, the paper strips were immersed in the THF solution of sensors 1 and 2 (40 μM) and dried in an oven at 30 °C for 20 min. The dried paper strips coated with sensors 1 and 2 did not display fluorescence emission under UV radiations (365 nm) (Figure ). However, the sensors’ coated paper strips displayed yellow and green emission under UV radiations when dipped in the THF solution of I– ions (5 μM) and dried. Hence, sensors exhibited a convenient detection of 5 μM for I– ions using sensors’ coated paper strips, which is comparable to the permissible level of I– (0.78–3.9 μM) in human urine, specified by the World Health Organization (WHO).[65] Therefore, sensors’ coated strips can potentially be used to detect iodide levels in urine and may prove a portable, fast-track economical approach for visual detection and quantification of I– ions in human urine for timely health screening.

Figure 12

Response of sensors’ coated paper strips of sensors 1 and 2 for the detection of I– ions.

Determination of I– Ions in the Blood Serum

Sensor 1 (as a model) was successfully employed for the determination of I– ions in the human blood serum that further comprehends its practical applicability. In this regard, fluorescence studies were performed in H2O/THF (7:3, v/v) through successive addition of deproteinized blood serum (0–20 μL) to sensor 1 (40 μM). Fluorescence emission of sensor 1 steadily increased upon progressive addition of increased volume of blood serum (0–20 μL) and displayed linear fluorescence emission enhancement behavior with increasing volume of blood serum sample (Figure S20). Taking this into consideration, a standard linear plot was obtained by using a known concentration of I– ions (0–20 nM) to compare and quantify iodide levels in human blood serum. Comparison of fluorescence spectra and linear fit plot implies that the used volume of human blood serum (20 μL) contains 20 nM concentration of I– ions. Hence, it is proposed that sensors reported in this study can be used to quantify iodide concentration in biological samples.

Determination of I– Ions in Real Samples

Spike and recovery strategy is commonly used to determine the biologically important analytes in real samples. The sensing potential of sensors 1 and 2 was investigated in real samples, including water (seawater, distilled water, tap water, and mineral water) and food samples (milk, iodized salt, and cheese) for potent detection of I– ions. Seawater samples were collected from Charna Island Karachi, Pakistan, and other food/water samples were commercially accessible. The water samples were purified through filtration to eradicate the unwanted impurities. The selected samples were prepared by spiking 30 nM concentration of I– ions. The fluorescence emission of the prepared samples was recorded through fluorescence spectroscopy and the obtained spectra are shown in Figure S20. The obtained fluorescence results were compared with fluorescence emission studies of laboratory samples to estimate % recovery of spiked I– ions. The recovery measurements were repeated thrice to estimate the relative standard deviation (RSD). Recovery measurements summarized in Tables and S2 revealed >97% recoveries of spiked I– ions that signify excellent sensitivity of sensors 1 and 2 for I– ions even in complex real samples.

Table 3

Recovery Measurements of Spiked I– in Real Water and Food Samples by Sensor 1a

real sample	titrated sensor 1 (μM)	spiked I– (nM)	recovered [mean]x ± [SEM]y	recovery (%)	RSD (%)
sea water	40	30 nM	35.77 ± 0.01	104.6	1.2
distilled water	40	30 nM	29.25 ± 0.13	98.89	2.3
tap water	40	30 nM	29.84 ± 0.02	98.67	1.7
mineral water	40	30 nM	27.66 ± 0.11	97.05	3.4
milk	40	30 nM	32.76 ± 0.24	102.31	3.7
iodized salt	40	30 nM	33.39 ± 0.03	103.08	1.6
cheese	40	30 nM	29.57 ± 0.09	99.44	1.5

x = mean of three experiments: y = standard error of the mean.

Conclusions

In summary, fluorenone-based sensors 1 and 2 were synthesized and characterized through 1H and 13C NMR spectroscopies. Both sensors displayed fluorescence enhancement response toward selective recognition of I– ions that is attributed to the blockage of ICT and C=N isomerization. Sensors 1 and 2 exhibited a low detection limit of 8.0 and 11.0 nM, respectively, in comparison to the previously reported I– ion sensors. The detection mechanism was confirmed through 1H-NMR titration, Job’s plots, DFT calculations, and DLS analyses. Moreover, both sensors were found efficient for colorimetric detection of iodide ions under the daylight and UV radiations (365 nm). Advantageously, sensors displayed reversible fluorescence responses through alternate addition of I– and Cu2+ ions. Fluorescence enhancement behavior, colorimetric, and reversible sensing potential of sensors 1 and 2 toward I– ions provide them an additional advantage over already reported iodide sensors. Moreover, the MTT assay revealed that the sensors presented in this study are biologically compatible and non-toxic to live cells. Additionally, these sensors were successfully employed for cell imaging of I– ions in live HeLa cells. Further, the fluorescence reversible behavior of sensors 1 and 2 was used to construct logic circuits. Finally, the preparation of sensors’ coated paper strips and successful determination of I– ions in blood serum, food, and real water samples make these sensors unique and highly practical.

Experimental Section

Biological Analysis

An MTT assay of sensor 1 before and after addition of I– ions was performed before bioimaging to investigate its cytotoxicity. In the first step, 10 000 cells were seeded in 96-well plates, and a number of HeLa cell lines were calculated through a hemocytometer. Sensor 1 (50 nM) in the presence of varying concentration of I– ions (2–8 nM) was interacted with cells for 24 h and repeated thrice. Afterward, the cells were rinsed with phosphate buffer in triplicate, followed by a 4 h treatment with 150 μL of MTT (6 mg/mL). Last, DMSO was added, MTT was discarded, and measurements were performed at 570 nm in a microplate reader. The native cell lines with 100% viability were taken as a control to express the cytotoxic effects of sensor 1 and I– ions. Further, sensor 1 and 1–I– complex were interacted with a specific IC50 concentration of HeLa cells to scan the emission spectra through a fluorescent microscope. In this regard, the cells were seeded in six-well plates at the 85% confluency rate and treated with sensor 1 and 1-I– complex for 2 h. Afterward, cells were washed a couple of times and treated with phosphate buffer for microscopic imaging.

The synthetic methodology of sensors 1 and 2 along with all used reagents and instruments is provided in the Supporting Information (SI 1–4). Similarly, explanation of fluorescence assay experiments, visual detection, and detection limit (LOD) and theoretical (DFT) calculations are given in SI 5–7.