Unique Photophysical Properties of 1,8-Naphthalimide Derivatives: Generation of Semi-stable Radical Anion Species by Photo-Induced Electron Transfer from a Carboxy Group.

The development of anion sensors for selective detection of a specific anion is a crucial research topic. We previously reported a selective photo-induced colorimetric reaction of 1-methyl-3-(N-(1,8-naphthalimidyl)ethyl)imidazolium (MNEI) having a cationic receptor in the presence of molecules having multiple carboxy groups, such as succinate, citrate, and polyacrylate. However, the mechanism underlying this reaction was not clarified. Here, we investigate the photo-induced colorimetric reaction of N-[2-(trimethylammonium)ethyl]-1,8-naphthalimide (TENI), which has a different cationic receptor from MNEI and undergoes the photo-induced colorimetric reaction, and its analogues to clarify the reaction mechanism. The TENI analogues having substituents on the naphthalene ring provide important evidence, suggesting that the colorimetric chemical species were radical anions generated via photo-induced electron transfer from carboxylate to the naphthalimide derivative. The generation of the naphthalimide-based radical anion is verified by 1H NMR and cyclic voltammetry analyses, and photo-reduction of methylene blue is mediated by TENI. In addition, the role of the cationic receptor for the photo-induced colorimetric reaction is investigated with TENI analogues having different hydrophilic groups instead of the trimethylammonium group. Interestingly, the photo-induced colorimetric reaction is observed in a nonionic analogue having a polyethylene glycol group, indicating that the colorimetric reaction does not require a cationic receptor. On the other hand, we reveal that the trimethylammonium group stabilizes the radical anion species. These generation and stabilization phenomena of naphthalimide-based radical anion species will contribute to the development of sophisticated detection systems specific for carboxylate.

Introduction

Colorimetric and fluorescent anion sensors for the selective detection of specific anions are an important research topic because anions play crucial roles in biological systems and industrial applications.[1−6] Numerous probe molecules with anion receptors harnessing weak interactions such as hydrogen bonding interactions, hydrophobic interactions, and electrostatic interactions have been developed.[7−9] Their anion recognition events involve changes in their optical properties that are attributed to intra-/inter-molecular electron transfer (ET)[10] and photo-induced ET (PET)[11,12] by virtue of anion binding.

The 1,8-naphthalimide (NI) derivatives are attractive species for developing anion sensors because they have an electron-deficient naphthalene ring capable of ET and PET.[6,13−16] We previously reported the interesting photophysical property of 1-methyl-3-(N-(1,8-naphthalimidyl)ethyl)imidazolium (MNEI), which consists of the NI group as a chromophore, an imidazole group as an anion receptor, and an N-ethyl linker.[17] The fluorescence intensities decreased as the electronegativity decreased owing to charge transfer from anion to MNEI.[18,19] In addition, MNEI underwent a photo-induced colorimetric reaction, in which the MNEI aqueous solution became yellow in the presence of carboxylate derivatives by UV light irradiation.[20,21] The yellow chemical species was unstable to O2, and the yellow color gradually disappeared under an ambient condition. The colorimetric reaction of MNEI proceeded especially readily in the presence of molecules having multiple carboxy groups, such as succinate, citrate, and polyacrylate, or amphiphilic carboxylate derivatives such as laurate and phenylacetate.[20,21] In contrast, the yellowing reaction was not observed in the presence of the same amount of acetate, but it was slightly observed in the presence of a ca. 30-fold higher amount of acetate. In addition, we used machine learning analysis to extract decisive factors that dominate the fate of MNEI salts after UV light irradiation, and the results suggested that pKa, the number of carboxy groups, and the bulkiness of carboxylate derivatives were important factors for the colorimetric reaction.[21] However, the mechanism for the photo-induced colorimetric reaction has not been clarified. If the mechanism by which the carboxy groups are selectively detected could be revealed, it would provide a novel design principle for anion sensor molecules.

We expected that the photo-induced colorimetric reaction using MNEI analogues having a substituent on the naphthalene ring would provide important evidence that could help to clarify the mechanism for the selective detection of carboxy groups. However, we could not synthesize MNEI analogues because of the difficulty in the synthesis and purification of ionic molecules. Therefore, we investigated N-[2-(trimethylammonium)ethyl]-NI (TENI) having a trimethylammonium group instead of the imidazolium group (Figure A) because TENI was easily synthesized via methylation of N-[2-(dimethylamino)ethyl]-NI[22] (DENI), which is a quantitative reaction carried out at room temperature[23] (Scheme S1), and we found that TENI showed the change in fluorescence intensity as the binding anions’ electrophilic property and the photo-induced colorimetric reaction in the presence of carboxylic acid derivatives as readily as MNEI (see Figures S1 and S2 and the description below). In addition, we could synthesize TENI analogues having an electron-withdrawing group (NO2-TENI) or an electron-donating group (CH3O-TENI) (Figure A) via the methylation reaction (Scheme S1). We additionally synthesized N,N′-bis[2-(trimethylammonium)ethyl]-1,8:4,5-naphthalenetetracarboxdiimide (TENDI) as an analogue having electron-withdrawing groups because naphthalenediimide (NDI) derivatives, having the imide groups at the 1,8- and 4,5-positions on the naphthalene ring, are well-known π-electron-deficient compounds and have attracted great attention as probe molecules.[24] The photo-induced colorimetric reaction using these TENI analogues and TENDI having different highest occupied molecular orbital–lowest unoccupied molecular orbital (HOMO–LUMO) gaps would provide important evidence to clarify the mechanism underlying the colorimetric reaction. In addition, we investigated the role of the cationic receptor for the colorimetric reaction with two TENI analogues having a different substituent in the place of the trimethylammonium group—namely, DENI and N-(methoxypolyethylene glycol)-NI (PEGNI) (Figure B).

Figure 1

Synthesized TENI analogues having a substituent on the NI group (A) or having a different substituent in the place of the trimethylammonium group (B).

Materials and Methods

Materials

N,N-Dimethylethylenediamine and succinic acid disodium salt were purchased from Nacalai Tesque (Kyoto, Japan). 1,8-Naphthalic anhydride, 3-nitro-1,8-naphthalic anhydride, iodomethane, and methoxypolyethylene glycol amine 750 were purchased from Sigma-Aldrich Japan (Tokyo, Japan). 4-Bromo-1,8-naphthalic anhydride, naphthalene-1,4,5,8-tetracarboxylic dianhydride, and sodium methoxide were purchased from Tokyo Chemical Industries (Tokyo, Japan). Copper(II) sulfate pentahydrate and sodium polyacrylate (DP: 22,000–70,000) were purchased from Wako Pure Chemical Industries (Osaka, Japan). Other reagents were of commercial grade and used without further purification.

Instrumentation

Nuclear magnetic resonance (NMR) spectra were recorded on a JNM-ECP500 (JEOL, Tokyo, Japan). Infrared (IR) spectra of the samples were recorded using a Spectrum 65 (PerkinElmer Japan, Tokyo, Japan) equipped with an attenuated total reflection attachment. Elemental analysis data were measured on a PerkinElmer 2400 II CHNS/O (PerkinElmer, Franklin Lakes, NJ, USA). Mass spectra were recorded using an Exactive plus Orbitrap MASS spectrometer (Thermo Fischer Scientific, Waltham, MA, USA). UV–vis spectra were recorded on a Multiskan GO (Thermo Fischer Scientific, Waltham, MA, USA). Fluorescence spectra were recorded on an RF-6000 (Shimadzu, Kyoto, Japan). Cyclic voltammetry (CV) was performed on a BAS 700E electrochemical analyzer (BAS, Tokyo, Japan) using the closed-type electrolysis cell equipped with a glassy carbon as the working electrode, a platinum wire as the counter electrode, and a saturated calomel electrode as the reference electrode.

Synthesis of Probe Molecules

N-[2-(Dimethylamino)ethyl]-NI (DENI)[22]

1,8-Naphthalic anhydride (2.03 g, 10.24 mmol) and N,N-dimethylethylenediamine (4.41 g, 50.03 mmol) were added to dimethylformamide (DMF) (30 mL), and the solution was stirred at 100 °C for 5 h. The reaction solution was cooled to room temperature and poured into a large volume of water (250 mL). The precipitated product was corrected by filtration, followed by washing with water. The product was dried under reduced pressure. Yield: 87.3% (2.40 g, 8.94 mmol). 1H NMR (500 MHz, DMSO-d6): δ 2.17 (s, 6H, CH3−), 4.13 (t, 2H, J = 6.30 Hz, −CH2CH2−), 7.84 (t, 2H, J = 7.73 Hz, naphthalimide), 8.43 (d, 2H, J = 8.02 Hz, naphthalimide), 8.47 (d, 2H, J = 6.87 Hz, naphthalimide) ppm.

TENI Iodide

N-[2-(Dimethylamino)ethyl]-NI (0.51 g, 1.90 mmol) and methyliodide (0.84 g, 5.91 mmol) were added to DMF (20 mL) and stirred for 1 h at room temperature. The product was gradually precipitated out during the reaction. The product was corrected by filtration, washed with water, and dried under reduced pressure. Yield: 88.4% (0.69 g, 1.68 mmol). 1H NMR (500 MHz, DMSO-d6): δ 3.20 (s, 9H, CH3−), 3.62 (t, 2H, J = 7.16 Hz, −CH2CH2−), 4.46 (t, 2H, J = 7.16 Hz, −CH2CH2−), 7.91 (t, 2H, J = 7.73 Hz, naphthalimide), 8.51 (d, 2H, J = 8.02 Hz, naphthalimide), 8.53 (d, 2H, J = 7.45 Hz, naphthalimide) ppm. ESI–MS m/z: calcd for [TENI]+ = 283.1441; found, 283.1436. Elemental analysis calcd for C17H19IN2O2 = C, 49.77; H, 4.67; N, 6.83. Found: C, 49.50; H, 4.69; N, 6.92. FT-IR (neat, cm–1) 776.92 (ArH), 1237.12 (C–N), 1653.98 (C=O), 2996.00 (C–H).

4-Bromo-N-[2-(dimethylamino)ethyl]-NI[25]

4-Bromo-1,8-naphthalic anhydride (1.39 g, 5.02 mmol) and N,N-dimethylethylenediamine (0.50 g, 5.70 mmol) were added to DMF (25 mL), and the solution was stirred at 120 °C for 2 h. The reaction solution was cooled to room temperature, and then the product was precipitated out. The precipitated product was corrected by filtration, washed with water, and dried under reduced pressure. Yield: 82.7% (1.44 g, 4.15 mmol). 1H NMR (500 MHz, DMSO-d6): δ 2.35 (s, 6H, −CH3), 2.65 (t, 2H, J = 7.49 −CH2CH2−), 4.32 (t, 2H, J = 6.30 Hz, −CH2CH2−), 7.85 (t, 1H, J = 7.45 Hz, naphthalimide), 8.04 (d, 1H, J = 7.45 Hz, naphthalimide), 8.42 (d, 1H, J = 7.45 Hz, naphthalimide), 8.58 (d, 2H, J = 8.02 Hz, naphthalimide), 8.66 (d, 2H, J = 6.30 Hz, naphthalimide) ppm.

4-Methoxy-N-[2-(dimethylamino)ethyl]-NI[26]

4-Bromo-N-[2-(dimethylamino)ethyl]-NI (0.50 g, 1.44 mmol) was added to methanol (20 mL). Then, sodium methoxide (0.62 g, 11.5 mmol) and copper(II) sulfate pentahydrate (0.04 g, 0.17 mmol) were added, and the solution was stirred at 80 °C for 12 h. After removal of methanol under reduced pressure, the residue was washed using a small volume of water, and the product was dried under reduced pressure. Yield: 12.5% (50 mg, 0.18 mmol). 1H NMR (500 MHz, CDCl3): δ 2.36 (s, 6H, CH3−), 2.65 (t, 2H, J = 6.30 Hz, −CH2CH2−), 4.13 (s, 3H, CH3O−), 4.32 (t, 2H, J = 6.87, −CH2CH2−), 7.04 (d, 1H, J = 8.02 Hz, naphthalimide), 7.70 (t, 1H, J = 7.73, naphthalimide), 8.56 (d, 2H, J = 8.02 Hz, naphthalimide), 8.60 (d, 1H, J = 6.88 Hz, naphthalimide) ppm.

4-Methoxy-TENI Iodide (CH3O-TENI-I)

4-Methoxy-N-[2-(dimethylamino)ethyl]-NI (0.050 g, 0.18 mmol) and methyliodide (0.090 g, 0.63 mmol) were added to DMF (10 mL) and stirred for 1 h at room temperature. After removal of DMF under reduced pressure, the residue was washed with water. The product was dried under reduced pressure. Yield: 22.2% (17.6 mg, 0.040 mmol). 1H NMR (500 MHz, D2O): δ 3.19 (s, 9H, CH3−), 3.50 (t, 2H, J = 7.44 Hz, −CH2CH2−), 4.40 (t, 2H, J = 7.44, −CH2CH2−), 4.02 (s, 3H, CH3O−), 7.02 (d, 1H, J = 8.02 Hz, naphthalimide), 7.55 (t, 1H, J = 8.02, naphthalimide), 8.17 (d, 1H, J = 7.98 Hz, naphthalimide), 8.24 (d, 1H, J = 7.45 Hz, naphthalimide), 8.34 (d, 1H, J = 8.59 Hz, naphthalimide) ppm. ESI–MS m/z: calcd for [CH3O-TENI]+ = 313.1547; found, 313.1541. Elemental analysis calcd for C18H21IN2O3 = C, 49.10; H, 4.81; N, 6.36. Found: C, 49.26; H, 4.86; N, 6.08. FT-IR (neat, cm–1) 780.99 (ArH), 1085.35 (C–O), 1268.50 (C–N), 1654.65 (C=O), 2973.00 (CH3).

3-Nitro-N-[2-(dimethylamino)ethyl]-NI

3-Nitro-1,8-naphthalic anhydride (0.51 g, 2.10 mmol) and N,N-dimethylethylenediamine (0.38 g, 4.11 mmol) were added to DMF (20 mL), and the solution was stirred at 90 °C for 3.5 h. The reaction solution was cooled to room temperature, and then the product was precipitated out. The precipitated product was corrected by filtration, followed by washing with water. The product was dried under reduced pressure. Yield: 45.7% (0.30 g, 0.96 mmol). 1H NMR (500 MHz, CDCl3): δ 2.34 (s, 6H, CH3−), 2.67 (t, 2H, J = 6.59 Hz, −CH2CH2−), 4.35 (t, 2H, J = 6.59 Hz, −CH2CH2−), 7.94 (t, 1H, J = 7.44 Hz, naphthalimide), 8.42 (d, 1H, J = 8.02 Hz, naphthalimide), 8.78 (d, 1H, J = 6.88 Hz, naphthalimide), 9.13 (s, 1H, naphthalimide), 9.32 (s, 1H, naphthalimide) ppm.

3-Nitro-TENI Iodide

3-Nitro-N-[2-(dimethylamino)ethyl]-NI (0.20 g, 0.64 mmol) and methyliodide (0.27 g, 1.90 mmol) were added to DMF (10 mL) and stirred for 1 h at room temperature. After removal of DMF under reduced pressure, the residue was washed by water. The product was dried under reduced pressure. Yield: 51.5% (0.15 g, 0.33 mmol). 1H NMR (500 MHz, DMSO-d6) δ = 3.20 (s, 9H, CH3−), 3.61 (t, 2H, J = 7.16 Hz, −CH2CH2−), 4.47 (t, 2H, J = 6.59 Hz, −CH2CH2−), 8.09 (t, 1H, J = 7.73 Hz, naphthalimide), 8.72 (d, 1H, J = 6.87 Hz, naphthalimide), 8.83 (d, 1H, J = 7.45 Hz, naphthalimide), 8.93 (s, 1H, naphthalimide), 9.52 (s, 1H, naphthalimide) ppm. ESI–MS m/z: calcd for [NO2-TENI]+ = 328.1292; found, 328.1286. Elemental analysis calcd for C17H18IN3O4 = C, 44.85; H, 3.99; N, 9.23. Found: C, 44.62; H, 3.92; N, 9.06. FT-IR (neat, cm–1) 785.82 (ArH), 1248.02 (C–N), 1542.33 (N–O), 1350.92 (N–O), 1661.82 (C=O).

N,N′-Bis[2-(dimethylamino)ethyl]-1,8:4,5-naphthalenetetracarboxdiimide

1,4,5,8-Naphthalenetetracarboxylic dianhydride (1.50 g, 5.59 mmol) and N,N-dimethylethylenediamine (4.90 g, 0.06 mol) were added to DMF (20 mL), and the solution was stirred at 100 °C for 6 h. The reaction solution was cooled to room temperature, and then the product was precipitated out. The precipitated product was corrected by filtration, followed by washing with water. The product was dried under reduced pressure. Yield: 42.4% (0.97 g, 2.37 mmol). 1H NMR (500 MHz, CDCl3) δ = 2.34 (s, 12H, CH3−), 2.67 (t, 4H, J = 6.59 Hz, −CH2CH2−), 4.35 (t, 4H, J = 6.87 Hz, −CH2CH2−), 8.76 (s, 4H, naphthalimide) ppm.

N,N′-Bis[2-(trimethylammonium)ethyl]-1,8:4,5-naphthalenetetracarboxdiimide Iodide[23]

N,N′-Bis[2-(dimethylamino)ethyl]-1,8:4,5-naphthalenetetracarboxdiimide (0.51 g, 1.25 mmol) and methyliodide (1.05 g, 7.40 mmol) were added to DMF (20 mL) and stirred for 1 h at room temperature. After removal of DMF under reduced pressure, the residue was washed by water. The product was dried under reduced pressure. Yield: 56.8% (0.49 g, 0.71 mmol). 1H NMR (500 MHz, D2O) δ = 3.35 (s, 18H, CH3−), 3.73 (t, 4H, J = 7.45 Hz, −CH2CH2−),4.69 (t, 4H, J = 7.45 Hz, −CH2CH2−), 8.81 (s, 4H, naphthalimide) ppm. ESI–MS m/z: calcd for [TENDI]2+ = 219.1128; found, 219.1120. Elemental analysis calcd for C24H30I2N4O4 = C, 41.64; H, 4.37; N, 8.09. Found: C, 41.43; H, 4.12; N, 8.08. FT-IR (neat, cm–1) 782.15 (ArH), 1187.07 (C–N), 1657.09 (C=O), 3031.61 (CH3).

N-(Methoxypolyethylene glycol)-NI

1,8-Naphthalic anhydride (13.2 mg, 66.7 μmol) and methoxypolyethylene glycol amine 750 (51.2 mg, 68.3 μmol) were added to DMF (8 mL), and the solution was stirred at 100 °C for 5 h. The reaction solution was evaporated. Water (10 mL) was added to the residue and filtered to remove unreacted 1,8-naphthalic anhydride. The filtrate was evaporated and the residue was washed with diethyl ether. The product was dried under reduced pressure. Yield: 23.0% (14.3 mg, 15.3 μmol).1H NMR (500 MHz, CDCl3): δ 3.38 (s, 3H, CH3−), 3.47-3.71 (m, 66H, PEG), 3.82 (t, 2H, J = 6.30 Hz, −CH2CH2−), 4.45 (t, 2H, J = 6.30 Hz, −CH2CH2−), 7.76 (t, 2H, J = 7.73 Hz, naphthalimide) ppm, 8.22 (d, 2H, J = 8.02 Hz, naphthalimide), 8.61 (d, 2H, J = 7.45 Hz, naphthalimide) ppm.

Photo-Induced Colorimetric Reaction

An aqueous solution of 1.5 mM TENI or 1.5 mM TENI analogues and 6.0 mM sodium succinate were added to a quartz cell (path length: 1 mm) and irradiated with UV light from an SP-7 spot cure unit (Ushio, Tokyo) equipped with a deep UV lamp (main wavelength: 365 nm) at room temperature for 5 min. The length between the quartz cell and the UV lamp was 19.2 cm (8 mW/cm2). After irradiation, the irradiated solution was immediately subjected to UV–vis measurement.

Photo-Induced Reduction of Methylene Blue Mediated by TENI

Methylene blue (MB, 0.25 mM)/1.5 mM TENI/6.0 mM sodium succinate aqueous solution, 0.25 mM MB/1.5 mM TENI aqueous solution, or 0.25 mM MB/1.5 mM TENI/12.0 mM sodium acetate placed in a quartz cell (path length: 1 mm) was UV-irradiated by the same procedure as used for the photo-induced colorimetric reaction. After irradiation, the irradiated solution was immediately subjected to UV–vis measurement.

Binding Constant between TENI and Succinate/Acetate

The binding constants (M–1) between TENI and succinate/acetate were estimated from 1H NMR measurements. A Benesi–Hildebrand-type equation[27] was employedwhere δ, δ0, K, and a are the chemical shifts of a signal attributed to an aromatic proton of TENI in the presence and absence of sodium succinate/acetate, the binding constant, and a molecular extinction coefficient, respectively.

Computational Details

The computations for the complex formed between TENI and acetate were performed by the density functional theory (DFT) method as implemented in Gaussian16.[28] In order to include the non-bonding interaction, we used the X3LYP functionals with the 6-31G* basis set. We confirmed that the complex lies at a stable state by computing the normal mode of frequency. To compute the vertically excited energy, we used the time-dependent DFT, where the lowest 15 states are computed.

Results and Discussion

UV–Vis and Fluorescence Spectra of TENI, CH3O-TENI, NO2-TENI, and TENDI

Figure A shows the UV–vis spectra of 1.5 mM TENI, CH3O-TENI, NO2-TENI, and TENDI in aqueous solutions. TENI showed an absorbance peak attributable to the one-electron HOMO–LUMO transition at around 345 nm; this was identical to the result for MNEI (Figure ).[17] In CH3O-TENI, an absorbance peak was observed at around 375 nm, which was shifted to higher wavelength compared to TENI by virtue of push–pull internal charge transfer.[6] In NO2-TENI, absorbance peaks appeared at around 275 and 345 nm. This observation is consistent with the absorbance peak of 3-nitro-N-[2-(dimethylamino)ethyl]-NI (mitonafide).[29] We suppose that this appearance of two absorbance peaks is due to the position of the nitro group because 4-nitro-NI derivatives did not show such absorbance pattern[30] and similar absorbance bands were observed in 3-substituted NI derivatives.[31] The absorbance peak of TENDI below 400 nm was identical to the previously reported absorbance peak of TENDI.[24,32] These results indicate the production of TENI analogues having different HOMO–LUMO gaps.

Figure 2

UV–vis spectra of 1.5 mM TENI, CH3O-TENI, NO2-TENI, and TENDI aqueous solutions (A) and fluorescence spectra of 1.5 mM TENI (excited at 380 nm), CH3O-TENI (excited at 310 nm), and TENDI (excited 260 nm) aqueous solutions (B). Excitation spectra are shown in Figure S4.

Figure 3

Orbitals of the TENI and acetate complex at the X3LYP/6-31G* level. The values are the relative orbital energies to HOMO of the complex (the values in parentheses are in Eh). HOMO – 2, HOMO – 1, and HOMO of acetate lie between the HOMO and LUMO of TENI. The absorption wavelength from HOMO to LUMO excitation of TENI is computed as 329.93 nm through time-dependent DFT at the X3LYP/6-31G* level.

Figure B shows the fluorescence spectra of 1.5 mM TENI, CH3O-TENI, and TENDI aqueous solutions. In the spectra of TENI, a monomer fluorescence peak at around 395 nm and an excimer fluorescence peak at around 500 nm were shown; these were identical to the results for MNEI.[17] In CH3O-TENI, a strong fluorescence emission peak was detected at around 465 nm; however, an excimer fluorescence emission peak was not observed, suggesting that the NI derivative having an electron-donated naphthalene ring was not suitable for the excimer formation. In the case of NO2-TENI, the fluorescence emission peak was not detected. In TENDI, although the maximum fluorescence intensities were approximately one-tenth of those observed for TENI, the monomer fluorescence emission peak was detectable at around 410 nm.

Photo-Induced Colorimetric Reactions of TENI, CH3O-TENI, NO2-TENI, and TENDI

Succinate was the simplest molecule to induce the colorimetric reaction of MNEI by UV light irradiation.[21] Therefore, the photo-induced colorimetric reactions of TENI, CH3O-TENI, NO2-TENI, and TENDI in the presence of sodium succinate were investigated (Figure ). In TENI, the solution turned yellow under UV light irradiation and the appearance and decrease of absorbance peaks were observed at around 430 nm and at around 345 nm, respectively. This yellow color was completely disappeared after 72 h. However, the absorbance peak at around 345 nm was not completely restored, suggesting decomposition of TENI. In addition, the colorimetric reaction was not observed in the presence of 12 mM sodium acetate but was observed in the presence of 150 mM sodium acetate (Figure S2).[20] These observations were the same as those for MNEI described in the Introduction section. In contrast, the photo-induced colorimetric reaction was not observed on CH3O-TENI, suggesting that electron donation from a substituent prevented the colorimetric reaction. Interestingly, in NO2-TENI and TENDI, which have an electron-deficient naphthalene ring, the solution colors changed to magenta, and broad absorbance peaks were observed at 400–600 nm. The color change on TENDI was particularly remarkable. The broad absorbance peak observed in TENDI agrees with the previously reported absorbance peak due to an NDI-based radical anion.[24] In addition, the absorbance peak appearing on TENI at around 345 nm shows good agreement with that of the NI-based radical anion.[33,34] These results suggest that the photo-induced colorimetric reaction in the presence of succinate was the production of radical anion via PET from carboxylate to the electron-deficient naphthalene ring. The finding that there was no color change in CH3O-TENI was a reasonable result under this hypothesis because CH3O-TENI has an electron-rich naphthalene ring that does not readily accept an electron. Moreover, the phenomenon of the yellow color disappearing under O2 is also consistent with the property of radicals. Note that the finding that there was no color change in the absence of succinate was confirmed in the above derivatives.

Figure 4

UV–vis spectra of 1.5 mM TENI (A), CH3O-TENI (B), NO2-TENI (C), and TENDI (D) aqueous solutions in the presence of 6.0 mM sodium succinate before and after UV light irradiation. Multi-views are photo–images of the respective solutions after UV light irradiation.

Verification of the Occurrence of PET from Succinate to TENI

Although we performed an electron spin resonance (ESR) analysis of the TENI system to confirm that radical anions are generated by UV light irradiation, we could not observe any significant signals, probably due to the difficulty of ESR measurement in water and/or the low radical concentration. Therefore, we carried out verification experiments using 1H NMR analysis, with the expectation that the signals for paramagnetic molecules would disappear, broaden, and/or shift.[32]Figure shows the 1H NMR spectra of 1.5-mM TENI-D2O solution in the presence of 6.0 mM sodium succinate before and after UV light irradiation. The integration values of signals due to TENI were clearly decreased by UV light irradiation. We speculated that the signals disappeared or broadened due to conversion of TENI to a paramagnetic radical anion. The integration values for signals due to the NI group (d–f) were decreased to ca. 40% compared to those before UV light irradiation. In addition, the signals attributed to trimethylammonium protons (a) became complex, with shifted and broadened signals being observed at around 3.0–3.2 ppm. This change after UV light irradiation was not observed in 1.5 mM TENI-D2O solution in the presence of 12.0 mM sodium acetate, which did not exhibit the yellowing reaction, as shown in Figure S2A (Figure S3). These results support the idea that the radical anion was generated.

Figure 5

1H NMR spectra of 1.5 mM TENI-D2O solution in the presence of 6.0 mM sodium succinate before and after UV light irradiation. DMF was added as an internal standard to estimate the integration values.

In order to confirm that the electrochemical redox reaction of TENI occurred, (CV) measurement of an aqueous KCl (10 mM) solution of TENI (1.5 mM) was made at a 50 mV/s scan rate under an argon atmosphere (Figure ). The blank measurement was also performed at the same scan rate (50 mV/s) in the absence of TENI. The CV of TENI showed a reversible one-electron redox wave at −1.1 V, indicating that a one-electron reduction state of TENI is plausible.

Figure 6

Cyclic voltammograms of a 10 mM KCl aqueous solution of 1.5 mM TENI-I.

In order to further verify the generation of a radical anion, the photo-induced reduction of MB, which is a dye bleached by reduction, was carried out: if the radical anion was generated via the PET from the carboxy group to TENI, the MB would be immediately reduced by the radical anion (Figure A). Figure B shows the UV–vis spectra of 0.25 mM MB/1.5 mM TENI/6.0 mM sodium succinate, 0.25 mM MB/1.5 mM TENI, and 0.25 mM MB/1.5 mM TENI/12.0 mM sodium acetate aqueous solutions before and after UV light irradiation. In the presence of TENI and succinate, the absorbance peak due to MB at around 450–750 nm was remarkably reduced after UV light irradiation. In contrast, such a reduction was not observed in the absence of TENI. In addition, when 12.0 mM acetate was used instead of 6.0 mM sodium succinate, the reduction was not observed. The conditions under which the reduction of MB occurred were consistent with the photo-induced colorimetric reaction, indicating that TENI was converted to the radical anion via PET and that TENI functioned as the reductant.

Figure 7

Redaction of MB mediated by TENI (A). The UV–vis spectra of 0.25 mM MB/1.5 mM TENI/6.0 mM sodium succinate (red lines), 0.25 mM MB/1.5 mM TENI (blue lines), and 0.25 mM MB/1.5 mM TENI/12.0 mM sodium acetate aqueous solutions (yellow lines) before (solid lines) and after (dashed lines) UV light irradiation (B). Photos of 0.25 mM MB/1.5 mM TENI/6.0 mM sodium succinate aqueous solution before and after UV light irradiation are also shown (C).

Effect of Irradiation Wavelength on the Photo-Induced Colorimetric Reaction of TENI

The effect of irradiation wavelength on the PET was investigated by using the excitation light of the fluorescence spectrometer where 1.5 mM TENI aqueous solution containing 0.5 wt % sodium polyacrylate was placed in the fluorescence spectrometer and the excitation ray with controlled wavelength was irradiated for 5 min. The color change on the ray-irradiated part was observed visually because the diffusion of the yellow color was very slow in the high viscous solution. Figure shows the photos of the solutions after 250–400 nm UV light irradiation. In the experiment using 350 nm irradiation, which is close to the maximum absorbance wavelength, the light-irradiated part clearly became yellow. The yellow colors were gradually diluted with increasing/decreasing wavelengths of irradiated rays. The colorimetric reaction was not observed in the experiments using 250 and 400 nm of rays. This observation was consistent with the absorbance peak (300–380 nm) of TENI, as shown in Figure A, indicating that the PET occurred via the HOMO–LUMO transition. The same result was observed for TENDI (Figure S5).

Figure 8

Photos of the TENI solutions containing 0.5 wt % sodium polyacrylate after UV light irradiation at various wavelengths (◎: significantly change, ○: moderately change, △: slightly change, and ×: no change).

Clarification of the Role of the Cationic Receptor for the Photo-Induced Colorimetric Reaction

TENI has an open space between the cationic receptor and the NI group that can trap various anion species.[17] Therefore, we consider that the cationic receptor should play a role in capturing the carboxy group and in turn should be related to the photo-induced colorimetric reaction. In order to verify the roles of the cationic receptor, the photo-induced colorimetric reaction of the NI derivatives having different substituents instead of the cationic group (Figure B) was investigated. Panels A and D in Figure show the UV–vis spectra of 1.5 mM DENI, having a dimethylamino group, in the presence and absence of sodium succinate, respectively, before and after UV light irradiation. Although the spectra of the TENI aqueous solution exhibit almost no change by UV irradiation (Figure F), the colorimetric reaction of DENI was observed in the absence of succinate (Figure D). This result suggests the occurrence of PET from a lone pair on the dimethylamino group, as previously reported in NDI derivatives.[35] Furthermore, this result provides the important information that the internal ET was prevented by the quaternization of the dimethyl amino group by methylation. In the presence of succinate, the colorimetric reaction of DENI was enhanced, indicating the occurrence of PET from succinate, even though DENI has less positive charge than TENI. This result made us skeptical about the necessity of the cationic receptor for anion binding via electrostatic interaction to induce the colorimetric reaction. Therefore, we next investigated PEGNI bearing a polyethylene glycol (PEG) group (the number-average degree of polymerization estimated by 1H NMR analysis was 16.5) that is nonionic and hydrophilic (Figure B). Surprisingly, although the visibility of the colorimetric reaction was lower than that of TENI, the absorbance peak at around 430 nm was clearly observed. This absorbance peak was remarkably larger than that in the negative control experiment (Figure E). This result indicates that the generation of the NI-based radical anion via PET does not require the cationic receptors, but the cationic group promotes the photo-induced colorimetric reaction. In our previous work, the importance of aggregation to the photo-induced colorimetric reaction was suggested.[20,21] PEGNI is an amphiphilic molecule, and thus, hydrophobic NI groups should be aggregated in water as micelles. Thus, we conjecture that aggregation of the NI groups contributes to the generation and/or stabilization of the radical anion in the PEGNI system.

Figure 9

UV–vis spectra of 1.5 mM DENI (A) and PEGNI (B) aqueous solution in the presence of 6.0 mM sodium succinate before and after UV light irradiation and their time profiles of the absorbance at 430 nm after UV light irradiation (C). Negative controls: UV–vis spectra of 1.5 mM DENI (D), PEGNI (E), and TENI (F) aqueous solution before and after UV light irradiation.

Figure C shows the time profile of the absorbance at 430 nm in the TENI, DENI, and PEGNI systems after UV light irradiation. The decrease in the absorbance peak at around 430 nm in the PEGNI system was faster than the corresponding decrease in the TENI and DENI systems, indicating that the cationic receptor contributes to the stabilization of the radical anion. In addition, we conjecture that the radical anion was additionally stabilized via aggregation as described above because the lifetimes of the previously reported NI-based radical anions were on the order of microsecond,[33,34] and the generation of the semi-stable radical anion of PEGNI is implausible without stabilizing effects.

Mechanism of Generation of the TENI-Based Radical Anion

As described above, the appropriate electron donors for the photo-induced colorimetric reaction are molecules having multiple carboxy groups or amphiphilic carboxylate.[21] In the case of acetate, a ca. 30-fold higher acetate concentration toward TENI was necessary for the detectable colorimetric reaction.[20] The difference in binding constants between TENI and donor molecules is one of the possible reasons for the difficulty of the reaction in the case of acetate. Therefore, we estimated the binding constant between TENI and succinate or acetate from 1H NMR analysis (Figures S6). The binding constant between TENI and succinate was only ca. 3-fold higher (227 M–1) than that between TENI and acetate (78 M–1), which can be explained by the number of carboxy groups. It is unlikely that this small difference was the reason for the difficulty for PET observed in the TENI-acetate system. We conjecture that the enhancement of aggregation of TENI, which would accelerate and/or stabilize the radical anion species as described above, is a plausible reason for the importance of multiple carboxy groups.

In the presence of I– or Br–, MNEI/TENI showed strong fluorescence quenching via PET (Figure S1).[18,19] In these cases, it can be considered that the generated MNEI/TENI radical anion rapidly deactivated via ET from the MNEI/TENI radical anion to the I radical cation. Therefore, the radical anion was not generated in the presence of I– and Br– even though PET occurred. A very low level of quenching of TENI via PET occurred in the presence of acetate (Figure S1). This is consistent with the results of the computational simulation, which indicated that HOMO – 2, HOMO – 1, and HOMO of acetate lie between HOMO and LUMO of TENI (Figure ). However, the radical anion was hardly observed in the presence of 8 equiv of acetate. We consider that the reason for this finding was the same as in the cases of Br– and I–; that is, the most generated radical anion was rapidly deactivated via ET in the case of acetate (Figure A). In our previous report, the importance of the bulkiness of carboxylate molecules for the colorimetric reaction was suggested by machine leaning analysis.[21] Based on these previous results, we conjecture that the generation of the semi-stable radical anion was caused by the release of a bulky donor molecule having carboxylates; that is, the radical anion was generated via the PET arising on an unstable complex between the NI group and the carboxy group and subsequent release of the carboxy group (Figure B). In addition, the radical anion was probably stabilized via aggregation as described above. Note that the acetate system is able to generate the radical anion in very lower probability than the succinate system because the colorimetric reaction was observed in the presence of a large amount of acetate.

Figure 10

Plausible mechanism of the non-radiative deactivation in the presence of acetate (A) and of the generation of the radical anion in the presence of succinate (B).

Conclusions

In order to clarify the mechanism underlying the photo-induced colorimetric reaction observed in NI derivatives having a cationic receptor linking via an N-ethyl linker, such as MNEI and TENI, toward carboxylate derivatives, we investigated the photo-induced colorimetric reaction of TENI, CH3O-TENI, NO2-TENI, and TENDI in the presence of sodium succinate. Although CH3O-TENI having the electron-donating group did not show the colorimetric reaction, TENI, NO2-TENI, and TENDI show the colorimetric reaction. In TENI, the solution turned yellow under UV light irradiation. However, in NO2-TENI and TENDI, the solution turned magenta. The absorbance peak observed in TENDI agreed with that of the previously reported NDI-based radical anion. On the basis of this observation, we assumed that the yellow chemical species generated with the TENI system was NI-based radical anion, and this assumption was confirmed by the 1H NMR and CV analyses, and the photo-reduction of MB mediated by TENI. In addition, we revealed the roles of cationic receptor for the photo-induced colorimetric reaction. Although the colorimetric reaction does not require the cationic receptor, because it was observed on nonionic PEGNI, the cationic receptor plays a role in the stabilization of the radical anion. In addition, we revealed that a lone pair of nitrogen and oxygen in the receptor moiety induced the PET without carboxylate derivatives, that is, the cation is the best receptor for this sensing system. This mechanism would provide a sophisticated selective carboxylate detection system.

Based on the knowledge of the photo-induced colorimetric reaction obtained here and in our previous works, we conjecture that the keys for the photo-induced colorimetric reaction are appropriate acidity for the occurrence of PET and bulkiness of the carboxylate derivatives to prevent non-radiative deactivation via ET after PET. In addition, aggregation of the NI groups contributes to the generation and/or stabilization of the radical anion. A detailed study of the aggregation state is now in progress.