Unusual redshift due to selective hydrogen bonding between F (cid:2) ion and sensor motif: a naked eye colorimetric sensor for F (cid:2) ions in an aqueous environment †

Naked-eye detection of anions is a challenging but useful technique. In this study, two novel colorimetric monomeric sensors based on highly selective 2,4-dinitrophenylhydrazine were designed, synthesized, and employed as a selective ratiometric optical sensor for F (cid:2) ion detection via the naked eye. The anion recognition properties were investigated via a pattern of color change as well as changes in absorbance signaling maxima. Sensor molecules were characterized by 1 H NMR, 13 CNMR, HRMS, and FTIR spectroscopy. Selectivity towards F (cid:2) ions was attributed to the capability of H-bonding interaction of N–H bonds, which was determined by 1 H NMR titration, 19 F NMR titration, and UV-vis spectroscopic titration. The association constants of 0.263 (cid:3) 10 6 M (cid:2) 1 (compound 4 ) and 6.4988 (cid:3) 10 6 M (cid:2) 1 were calculated using the Benesi–Hildebrand equation from UV-vis spectroscopic analysis. The limit of detection (LOD) for both monomeric sensors was 68.182 nM (compound 4 ) and 72.341 nM (compound 9 ) and for the polymeric sensor, it was 18.98157 nM (compound 11 ). Job’s plot analysis revealed the existence of 1:1 complex formation of F (cid:2) ions with both monomeric sensor molecules. The same colorimetric response and selectivity properties were exhibited by the norbornene-based homopolymer of the polymeric sensor molecule towards F (cid:2) ions.


Introduction
All anions play an important role in both biological and environmental systems. 1,2 Among them, F À ions are vital to humans because their deficiency leads to osteoporosis, and their overexposure causes weak bones and dental issues. [3][4][5] Hence, F À ion detection and removal are important challenges nowadays. Some of the existing techniques for F À ion determination include using a smartphone, an ion-selective electrode or a membrane-selective electrode, solid-state combination, optical methods, etc. In most of the cases, instruments are expensive and difficult to handle. 6 Therefore, developing comparatively cheaper and easy-to-handle F À ion techniques is currently an attractive topic for researchers in different fields. Inorganic fluoride is mainly obtained from F À ion-containing rocks, e.g., apatite, fluorite, biotite, and hornblende. Other sources include volcanic ash, agricultural fertilizers, combustion of coal, etc. Improper disposal of fly ash on the ground surface contributes fluoride to groundwater. 7,8 The WHO-permitted range of F À ion concentration in groundwater is 0.6 mg mL À1 to 1.5 mg mL À1 . The presence of this anionic species beyond this range is carcinogenic. 9 Hence, the superior detection and quantification of F À ions is a new challenge for researchers. Among all the existing detection techniques of F À ions, colorimetric detection using a simple chemical sensor has recently gained research interest due to its low cost, easy handling, and easily available environment-friendly chemical characteristics for making such a kind of sensors. [10][11][12][13] Using the colorimetric technique, we can detect anions through the naked eye, which is otherwise impossible in the case of other fluorometric probes. It also facilitates the easy detection of F À ions in real water samples. So, colorimetric naked-eye recognition is an essential tool for F À ion detection at the ppm-level. 14 All monomeric colorimetric sensors contain one chromogenic site and one binding site. Upon binding the analyte to the binding site of the sensor molecule, the chromogenic unit responds by changing its color, which can be seen by the naked eye. The analyte can bind to the sensor through both covalent and non-covalent ways. Non-covalent interaction is mostly electrostatic interaction, including p interaction and H-bonding interaction. [15][16][17] For H-bonding interaction, various research groups reported various moieties such as hydrazine-based indoles, bodipy-indole conjugates, calixpyrroles, polyamines, urea, thiourea, amides, imidazole, pyrrolyl derivatives, sulphonamides, amides, etc. [18][19][20][21][22][23][24][25][26] All the above moieties form H-bonding interactions with the smallest negatively charged dense species (mostly F À ions) and are responsible for a drastic color change. [27][28][29][30][31][32][33][34] Many of the above-mentioned systems can undergo abstraction of a proton by the classical Bronsted acid-base type of interaction at a higher concentration of F À ions. [34][35][36][37][38][39][40][41][42] All the above monomeric systems can detect the analyte at the ppm level through a change in the spectroscopic behavior at a particular wavelength. Among the reported F À ion sensors, very few systems have a ratiometric sensing property, which depends on various factors like concentration, electronic environment, phototransformation, etc. A redshift of more than 100 nm is rare for most of the existing F À ion sensors, which makes our proposed sensor very effective compared to other reported systems. In the case of a polymeric sensor, the chromogenic unit increases due to an increase in the repeating unit, so the color intensity of the polymeric sensor is higher than that of the monomeric sensor. Thus, naked-eye detection becomes easy in the case of a polymer-based sensor compared to the monomeric sensor. On the other hand, in the case of the polymeric probe, side-chain interaction between probes and analyte increases the sensing efficiency by many fold. 43

Optical properties (selectivity and sensitivity study)
Absorbance spectra of the monomeric and polymeric sensors were recorded in a DMSO : H 2 O 9 : 1 v/v solvent mixture with Tris-HCl buffer at pH 7.4. All three sensor molecules exhibited an absorbance maximum at l max 405 nm, and the color of the sensors was light yellow, which was visible to the naked eye ( Fig. 2-4). Selectivity is an important parameter for investigating the performance and further applicability of a sensing mechanism. The anion receptor properties were investigated upon the addition of various interfering anions (F À , Cl À , Br À , CH 3 CO 2 À , I À , ClO 3 À , HPO 3 2À , N 3 À , SO 4 2À and NO 3 À ), and an interesting shift in the absorbance peak was observed from 405 nm to 537 nm selectively for ion addition ( Fig. 2 and 3) to the monomeric sensor compound 4 and compound 9. The colors of the monomeric sensor solutions were changed from light yellow to dark pink instantly within seconds under the optimized conditions of Tris-HCl buffer (pH 7.4) at room temperature in a DMSO : H 2 O 9 : 1 v/v solvent mixture. The above interesting selectivity was observed only for F À ions in the presence of other interfering anions for all the monomeric and polymeric sensor molecules. The F À ion is a highly electronegative species, as a result of which a strong H-bonding interaction occurred, which is responsible for the drastic color This journal is © The Royal Society of Chemistry 2020

Materials Advances Paper
Open Access Article. Published on 24 August 2020. Downloaded on 11/6/2020 2:55:54 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

View Article Online
changes of the sensors in addition to the F À ion acceptor site 44,45 (Fig. 2-4).
A comparative UV-vis study was carried out in the presence of various interfering anions, F À , Cl À , Br À , CH 3 CO 2 À , I À , ClO 3 À , HPO 3 2À , N 3 À NO 3 À and SO 4 2À , with compound 11 (polymeric sensor). The color change upon the addition of F À ions for the polymeric sensor was investigated, taking 1 mg of the polymeric sensor and dissolving 1800 mL of DMSO solution. Upon the addition of F À ions selectively, the absorbance maximum shifted from l max 405 nm to l max 516 nm in the 9 : 1 (v/v) DMSO : H 2 O solution mixture at pH 7.4 in the Tris-HCl buffer solution. The color of the blank polymeric sensor was light yellow, as detected by the naked eye, which changes instantly to dark pink by the addition of F À ions (TBA + salt 10 À2 M stock solution in water). The color intensity of the polymeric sensor was higher than that observed for the monomeric sensors because of the greater number of repeating units in the polymeric backbone than in the monomeric sensors. It was observed that, due to the polymeric repeating unit, the spectra were broad for both blank polymeric sensors and in the presence of TBAF. The more intense and drastic naked eye and spectroscopic color changes are one of the advantages of polymeric sensors compared to the monomeric sensors with the selective addition of F À ions (Fig. 4).

Spectroscopic titration study
3.2.1 UV-vis spectroscopic titration. To investigate the interaction mechanism between fluoride and sensor molecules, concentration-dependent UV-vis titration studies were carried out with (2.5 Â 10 À6 M) compound 4 and 10 À4 M F À ions as a stock solution in the DMSO : H 2 O (9 : 1 v/v) solution in the presence of Tris-HCl buffer solution (pH 7.4) (Fig. 5). With the gradual addition of F À ions into the sensor (compound 4) solution, the absorbance at l max 405 nm for the monomeric sensors decreased, and a new absorbance band gradually increased at l max 537 nm. Upon the addition of 2.5 Â 10 À6 M F À ions (4 equivalent F À ion), a saturation point was attained by the monomeric sensor (for compound 4). For the concentrationdependent titration experiment, compound 9 was taken as a 2.5 Â 10 À6 M solution in DMSO solvent with 10 À3 M F À ions as a stock solution (TBA + salt). Unlike compound 4, in the case of compound 9, the absorption maximum at l max 405 nm also gradually decreased upon the addition of the F À ion solution, and a new maximum was generated at l max 537 nm in the medium. A saturation point was achieved at 3.0 Â 10 À5 M F À ion concentration. A polymeric sensor molecule (1 mg/2 mL DMSO solvent) was obtained upon the addition of F À ions (10 À3 M). A saturation point was achieved in the (6.5 Â 10 À5 M) F À ion solution. The absorbance maximum at l max 405 nm decreased, and a new absorption maximum was generated at l max 516 nm. A saturation point was achieved in the (6.5 Â 10 À5 M) F À ion solution. Due to the high redshift of absorption spectra from l max 405 nm to l max 537 nm (for the monomeric sensor), and from l max 405 nm to l max 516 nm (for the polymeric sensor), an isosbestic point was observed at about l max 444 nm (for all sensor molecules), which provides the strongest evidence for the interaction of the sensor with the analyte. So, from the concentration-dependent UV-vis titration spectra, it was observed that the analyte was interacting with the monomeric and polymeric sensors in a ratiometric fashion. During the experiment, when the F À ion concentration was increased gradually, the color of the monomeric sensor changed from light yellow to dark pink in the cuvette, as detected by the naked eye. 36,37 A 537 /A 405 (for monomeric sensors) and A 516 /A 405 (for the polymeric sensor) were linearly and positively correlated with the F À ion concentration, as shown in Fig. 5.
As the sensors contain a strong electron-withdrawing -NO 2 group, it withdraws electrons from the NH group, resulting in a partially positive charge on the H-atom. This positively charged H atom is a harder acid in nature. The F À ion is a hard base due to its high negative charge density and smaller size than the other interfering anions (F À , Cl À , Br À , CH 3 CO 2 À , I À , ClO 3 À , HPO 3 2À , N 3 À , SO 4 2À and NO 3 À ). As a result of the HSAB principle, the F À ion can make a strong H-bonding interaction with partially positively charged H-atoms of NH groups present in both the monomeric and polymeric sensors. 36 Because of this reason, the sensor molecules undergo interaction ratiometrically with F À ions.
3.2.2 1 H NMR titration study. For the further investigation of the interaction mechanism, 5 mL of F À ions (10 À2 M) was added to the monomeric sensors of concentration 10 À3 M in the 0.5 M DMSO-d 6 solvent, followed by NMR titration. The NMR spectra showed that, after the addition of TBAF, the NH peak at around 12 ppm vanished, and a triplet was formed at around 16 ppm. Besides, the splitting number of the aromatic peaks was found to decrease compared to the splitting before the addition of the analyte to the monomeric sensors. A new peak was generated at about 16 ppm, which implies the formation of HF 2 À . This observation indicates that the F À ion abstracted the H-atom from the NH group, making all the sensor molecules negatively charged. This abstraction of a H-atom generated a highly water-soluble species, HF 2 À , in the medium. The negative charge formed due to the abstraction of H-atoms can delocalize from the N atoms all over the molecule, resulting in a decreased splitting of all aromatic peaks due to ICT. 37 Here, a naked-eye color change was observed from light yellow to dark pink in the NMR tube itself upon the addition of F À ions to the sensors due to strong ICT (intermolecular charge transfer) during the titration experiment (Fig. 6).

19 F NMR titration study.
To investigate the interaction of F À ions with the monomeric sensors (compound 4 and compound 9), a 19 F NMR titration experiment was performed using a Bruker 500 MHz instrument. Then, by fixing the concentration of monomeric sensors (10 À2 M for both), solutions with the TBAF concentration of 1 M (stock solution concentration) (for compound 4) and TBAF concentration of 10 À1 M (stock solution concentration) were added in 5 mL amounts, and spectra were recorded after every addition of F À ions. Then, all 19 F NMR spectral data were plotted (Fig. 7), and it was observed that, for compound 4, the addition of F À ions first abstracted the H from NH to produce HF 2 À , which gives a 19 F NMR peak at À150 ppm. A further increase in the concentration of F À ions resulted in one more peak at À110 ppm, which was due to the interaction of F À ions with sensor molecules. 37 From the above experiment, we can conclude that with the increasing concentration of F À ions, two different electronic-environmental fluorides existed, one is HF 2 À and the other one is due to sensor-bound fluoride species. For another sensor, i.e., compound 9, a similar experiment was performed; the 10 À2 M concentration of the sensor compound was fixed, and a 10 À1 M F À ion solution was added to it gradually. The 19 F NMR   spectra were recorded after every addition of F À ions, and all the 19 F NMR data were plotted. From the 19 F NMR titration plot, for compound 9, the first peak appeared at À75 ppm; with the gradual addition of F À ions, another peak appeared in the 19 F NMR spectra at À150 ppm. At a higher concentration of F À ions, both peaks at À75 ppm and À150 ppm existed in the 19 F NMR spectra. This observation confirmed that sensor compound 9 first formed a complex with the analyte through a strong H-bonding interaction, which corresponded to the peak at À75 ppm, and then formed HF 2 À with the gradual addition of F À ions by abstracting H from the sensor molecule, which corresponded to the other peak at À150 ppm. Based on all the 19 F NMR titration spectra, both monomeric sensors were found to result in a similar type of interaction (Fig. 7).

Job's plot
After confirmation of the sensing mechanism by UV-vis spectral analysis (both comparative and titration spectra), 1 H NMR titration, and 19 F NMR titration, we concluded the interaction mechanism for both monomeric sensors (compound 4 and compound 9) concerning F À ion concentration. For this, we considered both monomeric sensors' concentration (10 À4 and 10 À3 M as stock) and analyte concentration (TBAF 10 À4 M). Then, a series of samples was used for UV-vis spectral analysis by varying the mole fraction of both sensors and analyte in the range of 0 to 1. Then a UV-vis experiment was performed for every set of samples. After that, the UV-absorption data were plotted against the mole fraction of the F À ion to obtain a Job's plot. It was observed that at the anion mole fraction of 0.5, the absorption value for the sensor molecule was maximum. This value confirms that the monomeric sensors (compound 4 and compound 9) formed a 1 : 1 complex with the F À ion (10 À4 M and 10 À3 M) (Fig. 8).

Materials Advances Paper
Open Access Article. Published on 24 August 2020. Downloaded on 11/6/2020 2:55:54 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Binding constant and limit of detection (LOD)
From UV-vis spectroscopic concentration-dependent analysis for compound 4 (2.5 Â 10 À6 M) and compound 9 (1.5 Â 10 À5 M), it was observed that with the increase in F À ion concentration, a ratiometric change occurred in the spectra. Absorbance maxima were shifted from l max 405 nm to l max 537 nm with the generation of an isosbestic point at l max 444 nm and a drastic naked eye color change occurred from light yellow to dark pink. From the titration between sensor and analyte, the binding constant or association constant of sensors with the analyte was calculated by using the Benesi is maximum absorbance at l max 537 nm among all absorbance values at l max 537 nm. From the Benesi-Hildebrand plot, the binding constant was 0.263 Â 10 6 M À1 for compound 4 and 6.4988 Â 10 6 M À1 for compound 9 (Fig. 9). Furthermore, the corresponding absorbance vs. concentration of F À ions for compound 4 (2.5 Â 10 À6 M) (Fig. 10A), compound 9 (1.5 Â 10 À5 M) (Fig. 10B), and compound 11 (1 mg/ 2 mL solvent) (Fig. 10C) has been used for calculating the LOD. The limit of detection has been calculated using LOD = 3s/k, where k is the slope of each linear plot for the mentioned sensors, and s is the standard deviation of blank measurements. With this, the calculated LOD values for compound 4, compound 9, and compound 11 are 68.182 Â 10 À9 M, 72.341 Â 10 À9 M, and 18.98157 Â 10 À9 M, respectively.

Time-dependent analysis and real sample analysis
The response of the sensors towards F À ions was within 3 minutes for compound 4 (Fig. S24, ESI †), compound 9 (Fig. S25, ESI †), and compound 11 (Fig. S26, ESI †) from the UV-vis experiment. For real sample analysis, a F À ioncontaining branded toothpaste was collected from a nearby supermarket. After that, 100 mg of that toothpaste was dissolved in 10 mL of HPLC grade water. Then the toothpaste solution was centrifuged for 30 min, and the residual solution was filtered down by syringe filtration. Then, a UV-vis spectroscopic analysis was performed using compounds 4, 9, and 11 with the toothpaste solution, respectively. For the analysis, compound 4 (2.5 Â 10 À6 M), compound 9 (1.5 Â 10 À5 M) and compound 11 were taken as 1 mg/2 mL DMSO : H 2 O 9 : 1 v/v, in Tris-HCl buffer of pH 7.4. To the solution of both monomeric and polymeric sensors of 100 mL, a toothpaste solution was added, and UV-vis analysis was done. In the recovery experiments, the spiked samples were obtained by adding standard solutions with varying F À ion concentrations into the toothpaste samples. The recovered F À ion concentration obtained from analysis was 1.2 Â 10 À6 M using compound 4 (Fig. S27, ESI †), 1.016 Â 10 À6 M using compound 9 (Fig. S28, ESI †), and 0.9213 Â 10 À6 M using compound 11 (Fig. S29, ESI †).

Mechanism of interaction of sensors with analyte
From all the above analytical studies, the interaction mechanism of all monomeric (compound 4 and compound 9) and polymeric sensors (compound 11) with F À ions was confirmed. The Job's plot confirms the 1 : 1 binding stoichiometry of the monomeric sensors with F À ions (Fig. 8). Then, 1 H NMR and 19 F NMR titrations confirmed the actual interaction mechanism (both strong H-bond formation and abstraction of H-atoms to produce HF 2 À species) of the sensors using the analyte (Fig. 6 and 7).
As the F À ion is smaller in size among all the halides, and negative charge density is also high, this makes F À ions a hard base in nature. Due to that hard nature, it can abstract the H-atom (hard acid) from -NH groups of the sensors, simultaneously forming HF 2 À in the medium and generating negative charge on the molecules (all monomeric and polymeric sensors). This negative charge on all sensors acts as a donor site, and the electron-withdrawing -NO 2 group of the DNPH moiety acted as an acceptor site. Due to the presence of donor and acceptor sites in the same sensor molecule, a strong ICT (intermolecular charge transfer) occurred from the negatively charged dense donor site (NH-group) to the electron-withdrawing acceptor site (NO 2 group), which was confirmed by both 1 H NMR and 19 F NMR titration studies ( Fig. 6 and 7). 39 Also, in the presence of two 2,4-DNPH groups, a nice cavity was produced by the sensor molecules for F À ions only (as an acetate anion does not interfere with the F À ion). So, it can be concluded that both the species ([sensors + F À ion] and [sensors] 2À ) were present in the solution for all the sensors (compound 4, compound 9, and compound 11). The pictorial representation of the interaction mechanism for F À ions with the sensors is given below in Fig. 11.

Conclusion
In conclusion, the sensors were synthesized successfully and wellcharacterized by 1 H NMR, 13 CNMR, HRMS, and FTIR spectroscopies. All these sensors show a selective colorimetric response (from light yellow to dark pink) at a micromolar level in semiaqueous medium with F À ions. The binding interaction with sensors was investigated by performing 1 H NMR, 19 F NMR, and UV-vis spectroscopic titrations. The analysis proved that the binding ratio was 1 : 1 (from the Job's plot) due to H-bonding interaction between the sensors and F À ions. The drastic color change from yellow to pink was due to H-atom abstraction by the probe and ICT from the N-H group, with the high electronwithdrawing nitro groups. Hence, we have successfully explained all the syntheses, analyses of monomeric and polymeric sensors, and the mechanism of interaction of the analyte with the sensors.

Conflicts of interest
There are no conflicts to declare.