Introduction

Water is essential for regulating physiological and biochemical processes in living organisms1,2,3,4,5. Water can form intermolecular hydrogen bonds with inorganic compounds, organic molecules, and biological macromolecules, such as proteins and enzymes, inducing fascinating functionalities6,7,8,9,10,11,12,13. For example, water can help remove damaged adenine residues from oligonucleotides in the presence of a DNA glycosylase14. The structural participation of bound water in biomolecular recognition provides effective predictions for drug design15. However, water also causes some problems and limits the application of certain devices, such as perovskite solar cells and organic light-emitting diodes (OLEDs). For example, the performance of perovskite solar cells under an air atmosphere dramatically decreases over time because water in the air destroys the perovskite structure. Hydrogen bonding is among the most important noncovalent intermolecular interactions16, determining molecular conformation, molecular aggregation, and the function of numerous macroscopic substances. For example, bound water molecules exhibit liquid properties rather than gas properties owing to the existence of multiple intermolecular hydrogen bonds. In addition, hydrogen bond can also be used to design smart materials, such as self-healing polymers17,18, artificial muscle19, and organic or inorganic fluorescent/phosphorescent materials20,21. Hydrogen bond-assisted emission behaviour has mostly been attributed to restricted intramolecular rotation and vibration22,23.

Smart (stimuli-responsive) organic materials are able to reversibly respond to external stimuli, allowing their properties to be controlled by environmental stimuli, including temperature, electricity, mechanical, magnetic fields, chemicals, and stress24,25,26,27,28. Organic photoluminescent (PL) materials, as a class of smart organic materials, have gained considerable attention owing to their highly sensitive signal response and practicality-no sophisticated or expensive instruments are required29,30,31,32,33,34. However, in polar solvents, such as water, many organic PL materials show weak or quenched luminescence owing to polar interactions between the electronic structure and water, limiting their practical applications. Recently, certain organic PL materials have shown strong luminescence in the solid state without any solvent. In contrast, the emission behaviour of organic PL materials in the solid state under a water atmosphere remains in its infancy, but is highly important because most practical applications, such as OLEDs, are performed under an air atmosphere. Therefore, we are interested in designing a type of molecule to achieve high emission efficiencies in the solid state under a water atmosphere as an important development in this field.

Herein, we report a series of organic PL materials that can form hydrated crystals with water to produce organic hydrate luminogens (OHLs; Fig. 1, Supplementary Figs. 1, 5, 9, and 13). These OHLs are tautomerizable Schiff-base materials (2-(((1H-benzo[d]imidazol-2-yl)imino)methyl)-4-methoxyphenol (named as 1a), 2-(((1H-benzo[d]imidazol-2-yl)imino)methyl)phenol (named as 1b), and 2-(((1H-benzo[d]imidazol-2-yl)imino)methyl)-5-methoxyphenol (named as 1c)) in which water plays a crucial role in regulating molecular stacking modes and effectively induces turn-on fluorescence. Hydrated single crystals (microcrystals) and anhydrous single crystals of the compounds were obtained in the presence and absence, respectively, of water (H2O) or deuterium oxide (D2O). The three Schiff-base compounds in anhydrous organic solvents, the anhydrous polycrystalline powder state, and the anhydrous single crystal state all showed weak fluorescence. However, their water-containing microcrystals and single hydrate crystals exhibited enhanced fluorescence due to the formation of a hydrogen-bonding network, which hindered excited state tautomerization and thus minimized energy dissipation. Intercalated water molecules altered the molecular reorganization of these small organic compounds in the hydrated crystalline state. Strong affinity between organic molecules and water blocked the in situ structural rearrangement due to the formation of intermolecular hydrogen bonds in the hydrated crystals. Furthermore, reversible “on-off” fluorescence switching of the three organic compounds was achieved by hydration and dehydration of their polycrystalline samples. Compounds 1a, 1b, and 1c in anhydrous polycrystalline and crystalline phases induced strong light emission through in situ assembly with water (Fig. 1). The mechanism of enhanced luminescence in these compounds was attributed to the inhibited tautomer formation of benzimidazoles, which altered the nπ* to ππ* electronic transition of their keto form, and increased molecular rigidity of their assemblies in the solid state.

Fig. 1: Schematic of turn-on molecular emission.
figure 1

Water-induced in situ changes in molecular organisation and turn-on of molecular emission for polycrystalline powders of 1a, 1b, and 1c. Intercalation of water molecules into crystal lattices was observed to alter the surroundings of these molecules, leading to structural rearrangement in the crystalline state. These structural changes triggered significantly enhanced fluorescence emission.

Full size image

Results

Synthetic target compounds

Target compounds 1a, 1b, and 1c were prepared in high yields by one-step aldimine condensations of commercially available materials (Supplementary Figs. 1, 5, and 9). These compounds were characterized by 1H and 13C nuclear magnetic resonance (NMR) spectroscopies and mass spectrometry (Supplementary Figs. 2–4, 6–8, and 10–12). A methoxy group was introduced at different substitution positions in compounds 1b and 1c to investigate the effect of electron-donating groups on the optical properties of the materials. The presence of electronegative nitrogen and oxygen atoms, and active hydrogen atoms (N–H and O–H) promoted the formation of intramolecular and intermolecular hydrogen bonding (Supplementary Tables 7 and 8) between the organic molecules and with polar guest molecules, such as water (Fig. 2b, Supplementary Figs. 17, 19, and 20).

Fig. 2: Molecular stacking and photoluminescence results.
figure 2

Molecular stacking of prepared single crystals: a 1b and b 1b·H2O. Multiple intermolecular and intramolecular hydrogen bonds in the single crystals are denoted by dotted lines. c Emission spectra (λex = 400 nm) of 1b (1 × 10−5 M) in ethanol and 1b·H2O (D2O) microcrystal dispersion in water or deuterium oxide; d changes in relative photoluminescence intensity with respect to different ratios of H2O and D2O.

Full size image

Single crystals of target compounds

The single-crystal structures of the small organic compounds showed that all structures possessed intramolecular and intermolecular hydrogen bonds. Molecules of 1a interlinked to form a one-dimensional (1D) supramolecular chain (Supplementary Fig. 17). Two molecules of 1b interlinked to form a dimer via two N–H···O hydrogen bonds, with these dimers forming a 1D supramolecular chain (Fig. 2a). Four molecules of 1c packed together to form a supramolecular ring via N–H···O and N–H···N intermolecular hydrogen bonds (Supplementary Fig. 20). The molecular configuration of 1a single crystals showed a dihedral angle between R1 (benzimidazole ring) and R2 (phenyl ring) of about 18.56° (Supplementary Fig. 18). An intramolecular hydrogen bond between the hydrogen atom of a hydroxyl group and the nitrogen atom of a C=N bond (N∙∙∙H distance, 1.892 Å) and one intermolecular hydrogen bond between two adjacent molecules (N–H∙∙∙N distance, 2.133 Å or 2.072 Å) supported the aforementioned organization of the 1a molecules (Supplementary Fig. 17).

Single crystals of target compounds with H2O or D2O

In organic crystal growth, the presence of H2O or D2O molecules induces reorganization of the supramolecular array of organic compounds. Although the supramolecular structures of the three compounds were distinctive, the water-containing single-crystal structures of 1a and 1c were similar. The stoichiometry of water molecules in the unit cell of hydrated single crystals was 1:1. The unit cell dimensions of hydrated single crystals of 1a and 1c were larger than those of their anhydrous counterparts. For compound 1b, the unit cell dimensions of hydrated single crystals were smaller than those of their anhydrous counterparts (Supplementary Tables 1–3). Taking compound 1a as an example, a water molecule acted as a tridentate H-bonding moiety to link three surrounding molecules into a 1D ladder-shaped superstructure (Supplementary Fig. 17). The dihedral angles between aromatic segments R1 and R2 in the 1a·H2O and 1a·D2O lattices were about 10.98° and 10.99°, respectively (Supplementary Fig. 18), indicating that molecule 1a in the 1a·H2O and 1a·D2O lattices was closer to planar compared with the anhydrous form. Molecules of 1a formed a layer-by-layer structure indicative of face-to-face stacking. In each layer, molecular packing in a shifted cofacial arrangement was promoted. The molecules were stacked on top of each other and into columns along the b axis. Multiple intermolecular bonds between the organic molecules and H2O or D2O were beneficial for molecular planarity and acted as the driving force for formation of the layer-by-layer structure in the single-crystal state.

Characterization of microcrystals

Well-defined microcrystals (1a·H2O (D2O), 1b·H2O (D2O), and 1c·H2O (D2O)) were fabricated using a facile self-assembly method (see Supplementary Information). The crystallisation process was accelerated by adding H2O or D2O to molecular solutions in ethanol. Scanning electron microscopy (SEM) images of the as-prepared crystalline samples showed a large number of microcrystals (Supplementary Figs. 21–23). X-ray diffraction (XRD) analysis in reflection mode was performed to characterize the molecular organization in different states. XRD patterns of the bulk polycrystalline powders and hydrated microcrystals were in agreement with the simulated patterns from the corresponding single crystals (Supplementary Fig. 24) and hydrated single crystals (Supplementary Fig. 25). For example, microcrystals of 1a·H2O and 1a·D2O had similar XRD patterns (Supplementary Fig. 24a). In contrast to the XRD pattern of 1b·H2O (D2O) single crystals, 1a·H2O (D2O) microcrystals had only three Bragg reflection peaks, suggesting that 1a·H2O (D2O) microcrystals had a preferential growth orientation. Therefore, molecules in 1a·H2O (D2O) microcrystals assembled preferentially along the (100), (002), and (004) lattice directions with good crystallinity.

Optical properties

Hydrated single crystals and microcrystals of 1a, 1b, and 1c showed significantly enhanced fluorescence compared with their solutions and anhydrous powders (Fig. 2c, Supplementary Figs. 28 and 29). We investigated the photophysical properties of all synthesized compounds in different states, namely, in solution, as anhydrous crystals, and as hydrated crystals. Supplementary Figs. 26a–26c show UV-visible absorption spectra of the chromophores in different anhydrous organic solvents (methanol, ethanol, tetrahydrofuran, and dichloromethane). The absorptions of the three compounds in the region of 320–450 nm was assigned to π–π* excitation of the enol ground state (E) to the first enol excited state (E*) (S0 → S1). We also investigated the PL properties of these compounds in different organic solvents (Supplementary Fig. 27). The formation of the keto excited state (K*) occurred through the excited-state intramolecular proton transfer (ESIPT) process following photoexcitation (E → E*) of the enol ground states, as shown in Supplementary Fig. 33a. All chromophores showed almost negligible fluorescence with dual emission peaks (Supplementary Table 4) upon excitation. The position of the methoxy group significantly influenced the proton transfer event in the excited state. The short-wavelength emission of the compounds at 459 nm was attributed to the excited state enol form (E*) (enol emission), while peaks at 594 nm (1a), 620 nm (1b), and 572 nm (1c) were attributed to the excited state keto form (keto emission) through a four-level photocycle (Supplementary Fig. 33a). Methoxy group substitution at different positions had a negligible impact on enol emission wavelength, but significantly affected the keto emission wavelengths. In the K* state, the methoxy group at the 5-position in 1b further increased the electron density of the highest occupied molecular orbital (HOMO), which decreased the energy gap of the keto tautomer. Therefore, the ESIPT emission peak of 1b was red-shifted compared with that of 1a. However, when the methoxy group was located at the 4-position in 1c, the conjugation effect was hindered and an inductive effect occurred. This increased the energy level of the lowest unoccupied molecular orbital (LUMO) and increased the energy gap of the keto tautomer. Therefore, the ESIPT emission of 1c was blue-shifted compared with that of 1a. These molecules were all typical ESIPT chromophores. C=N isomerization is the predominant nonradiative decay process of excited states in compounds with an unbridged C=N structure, such that solutions of the three compounds were nonfluorescent35,36,37.

Compared with UV-visible absorption spectra of solutions, the absorption spectra of microcrystals dispersed in aqueous solution were blue-shifted due to the formation of hydrated microcrystals (Supplementary Figs. 26d–26f). This blue-shift in the absorption spectra of hydrated microcrystals indicated the formation of H-aggregation, which was caused by cofacial stacking of the molecules. Subsequently, the emission spectra of hydrated microcrystals dispersed in water were investigated. A dramatic increase in the fluorescence intensity of hydrated microcrystals (microcrystals of 1a·H2O (D2O), 1b·H2O (D2O), and 1c·H2O (D2O)) was observed. Water (and D2O) greatly increased the fluorescence quantum yield (QY) of these microcrystals (Supplementary Table 5), and the emission wavelengths of the six microcrystals were consistent with the respective molecular keto emissions in solution. The emission observed for the enol form in solution was absent in the microcrystals. The QY (Φf = 15.3%) of 1a·H2O microcrystals was about 55 times higher than that in ethanol solution (Φf = 0.28%) and other organic solvents (Supplementary Table 5), accompanied with a red-shift to 593 nm. Steady-state fluorescence emission spectra of the hydrated single crystals (Supplementary Fig. 28) showed similar emission behaviour to the corresponding microcrystals. Steady-state fluorescence emission spectra of the molecular polycrystalline powders were also investigated (Supplementary Fig. 28). All compounds were weakly emissive with fairly low fluorescence QYs. The emission wavelengths of polycrystalline powders of 1a, 1b, and 1c were blue-shifted compared with their individual hydrated single crystals and microcrystals.

As 1b showed the fastest water-binding rate among the three molecules, 1b was used to explore the in situ formation of new hydrogen bonds with guest water molecules. To avoid the influence of the OH-stretching vibration in H2O, D2O was used for in situ XRD and Fourier transform infrared (FT-IR) spectroscopy analyses. When the polycrystalline powder of anhydrous 1b was dipped in D2O, in situ XRD measurements showed new Bragg reflection peaks assigned to 1b·D2O (Fig. 3a). For example, (002), (012), (122), and (131) lattice directions appeared when 1b polycrystalline powder was dipped in D2O for 1 h. When the sample was dipped for almost 24 h, the XRD pattern matched well with that of the aforementioned 1b·D2O microcrystals, indicating successful in situ transformation from the anhydrous state to the hydrated state. Interestingly, new peaks for the (002) lattice direction occurred when anhydrous 1a and 1b polycrystalline powders were dipped in D2O. The (002) lattice direction is the orientation plane in which water and organic molecules form new intermolecular hydrogen bonds in the anhydrous crystal lattices. In situ FT-IR spectra (Fig. 3c) of 1b polycrystalline powder with D2O showed a stretching band at 1577 cm−1 that diminished gradually and a new stretching band at 1560 cm−1 that appeared simultaneously. This change suggested that the N–H chemical environment in the benzimidazole ring changed after being dipped in D2O. Combined with single-crystal data analysis, the disappearance of the stretching band at 1577 cm−1 was attributed to breakage of the N–H···O (–OH) hydrogen bond between organic molecules in the anhydrous structure (Fig. 3e), while the new stretching band at 1560 cm−1 suggested the formation of N–H···O (–OCH3) hydrogen bonds between molecules in the hydrated structure (Fig. 3c). Another new stretching band at 1025 cm−1 was attributed to the formation of a N···D (D2O) hydrogen bond (Fig. 3f). The location of the methoxy group of 1b in Fig. 3e and f has changed due to the formation new hydrogen bonds between 1b molecules.

Fig. 3: Characterization of 1b and 1b·D2O.
figure 3

In situ XRD (a), FT-IR (c) and emission (d) spectra of 1b polycrystalline powder soaked in D2O. b Molecular arrangement in the (002) lattice direction of 1b·D2O single crystals; e N–H···O (–OD) hydrogen bond between molecules in the anhydrous structure; f N–H···O (–OCH3) hydrogen bond between molecules, N···D (D2O) hydrogen bond between molecules and water.

Full size image

Possible mechanism

To understand the fluorescence turn-on properties of the hydrated crystals, theoretical calculations were conducted. The emission wavelength of hydrated crystals (1a·H2O (D2O), 1b·H2O (D2O), and 1c·H2O (D2O)) was similar to that of the keto emission in an anhydrous organic solvent, while the enol emission was not observed. This showed that intramolecular hydrogen bonding was more effective in the solid state with the ESIPT process. The active hydrogen atom in the benzimidazole ring was responsible for tautomer formation and shuttled between its two N atoms (Fig. 4a), resulting in fluorescence quenching. Using the polycrystalline powder of 1a, the radiation rate (kr ~ E2f) in the powder state was very low owing to the low f value (f = 0.0123). Therefore, the fluorescence quantum efficiency of ESIPT emissions (K*) from polycrystalline powders of 1a, 1b, and 1c was low owing to the nonradiative deactivation pathways (nπ*) of the tautomer (Supplementary Table 7 and Supplementary Fig. 33b). Water acting as a hydrogen bond donor formed strong intermolecular hydrogen bonds (N∙∙∙H–O or N∙∙∙D–O) with the nitrogen atom of the molecule (Fig. 4a), inhibiting tautomer formation. When the tautomer was inhibited by intermolecular hydrogen bond formation in the excited states, the electronic transition of geometry-B (S1) (Supplementary Fig. 34) changed from nπ* to lower-energy ππ*, resulting in enhanced keto (K*) emission in the solid state. The radiation rate (kr ~ E2f) of 1a·H2O was higher than that of 1a polycrystalline powder owing to the higher f value (f = 0.1934) (Supplementary Table 9 and Supplementary Fig. 34). Therefore, hydrated crystals 1a·H2O (D2O), 1b·H2O (D2O), and 1c·H2O (D2O) exhibited relatively bright fluorescence in the solid state. Compared with water, deuterium oxide showed a stronger proton donating ability toward the N atom of the benzimidazole group. Therefore, the deuterium isotope effect caused the fluorescence quantum yields of D2O-containing hydrates to be higher than those of H2O-containing hydrates (Supplementary Table 5)38. The fluorescent images also showed fluorescence intensity of 1a·D2O microcrystals grown on silicon dioxide substrate was stronger than that of 1a·H2O microcrystals grown on silicon dioxide substrate (Supplementary Fig. 39).

Fig. 4: Proposed activation of emission mechanism.
figure 4

a Proposed activation of single-molecule emission mechanism for 1a·H2O; b mean square displacement of 1a and 1a·H2O microcrystals during 1-ns molecular dynamics simulation; optimised molecular crystal structures of c 1a and d 1a·H2O single crystals at the final step (sky-blue dotted lines represent the hydrogen bond network); solid-state 1H NMR spectra of e 1a and f 1a·H2O. 1H double quantum filtered (DQ-filtered) spectra (in red) were recorded for comparison with 1H one-pulse spectra (in black), where signals of protons with high local mobility were filtered.

Full size image

A 1-ns molecular dynamics simulation was also conducted using the COMPASS force field to investigate the molecular stacking mode and hydrogen-bonding network. Quantitative mean square displacements of dynamic trajectories, as shown in Fig. 4b, indicated a larger thermal fluctuation in the 1a molecular crystal than in 1a·H2O. The corresponding equilibrium structures are also shown in Fig. 4c and d, respectively. This verified that a supramolecular ring was present via N–H···O and N–H···N hydrogen bonding in the single crystal of 1a. This interaction induced a twist between R1 and R2, but also formed a dimer structure between the organic molecules that exhibited certain flexibility with high freedom. However, for the single crystal of 1a·H2O, water acted as a “glue” to connect the organic molecules, with a well-organized three-dimensional hydrogen network lowering the fluctuation displacement and protecting the planarity of the organic molecule. Therefore, aided by water, the hydrogen-bonding network in 1a·H2O single crystals was prone to forming a relatively confined space to suppress the vibration of 1a molecules, which effectively decreased loss from electronic transitions between orbitals and further enhanced molecular photoluminescence.

To further confirm the existence of hydrogen bonds in 1a·H2O, solid-state 1D 1H NMR double quantum filtered (DQ-filtered) spectra and 1H one-pulse spectra were recorded. Figure 4e shows the 1H one-pulse spectrum of 1b. The resonances at 12.8 and 14.8 ppm were associated with the –OH and –NH units, respectively. However, these two signals were not observed in the 1H DQ-filtered spectrum (Fig. 4f), indicating that the –OH and –NH protons of 1a exhibited high local mobility. In contrast to 1a, 1a·H2O showed only one peak at 12.0 ppm in the 1H one-pulse spectrum. This was in agreement with the single crystal structure, in which hydrogen bonding with water induced identical local structures for the –OH and –NH protons. The signal at 12.0 ppm was also observed in the 1H DQ-filtered spectrum, implying that –OH and –NH protons in 1a·H2O had limited local mobility, which might be induced by hydrogen bonding with water. To further verify the mechanism, compound (2-(((1-methyl-1H-benzo[d]imidazol-2-yl)imino)methyl)phenol (named as 1d)) (Supplementary Fig. 32) was prepared, in which the active hydrogen atom on the N atom present in 1b was replaced by a methyl group (Supplementary Fig. 13). Compound 1d strongly fluoresced in the solid state owing to the lack of tautomer formation after replacing the active hydrogen in the benzimidazole moiety.

The hydrated microcrystals were also found to be metastable structures. Using 1a·H2O microcrystals as an example, the XRD pattern changed significantly after heat treatment, becoming similar to that of bulk 1a polycrystalline powder, indicating a rearrangement of molecules in the crystal lattice (Supplementary Fig. 37). Furthermore, the surfaces of the hydrated microcrystals were covered with small dots after thermal treatment at 60 °C for 25 min, which might have been caused by water loss (Supplementary Figs. 21–23). Upon thermal treatment, a 1a·H2O microcrystal film deposited on a polytetrafluoroethylene (PTFE) membrane (Supplementary Fig. 34) showed gradually decreasing fluorescence with prolonged heating. The fluorescence quenching of the 1a·H2O film deposited on a PTFE membrane after heat treatment was turned on by adding water. Other microcrystals have shown similar fluorescence on-off switching behaviour. Owing to the large gap in quantum yield between 1b·H2O microcrystals and 1b·D2O microcrystals (Supplementary Table 5), 1b was used as a fluorescent probe to investigate different ratios of H2O and D2O (Fig. 2d).

Discussion

Organic hydrate luminogens were prepared and their water-promoted fluorescence enhancement properties were investigated. Intercalated water molecules played an important role in triggering molecular reorganization in the hydrated crystalline state by forming new intermolecular hydrogen bonds between the organic molecules and guest water molecules. This water-induced rearrangement of the organic compounds in the crystalline state caused significant enhancement of their molecular emissions in the solid state by suppressing shuttling of the active hydrogen atom between N atoms in the benzimidazole ring. Water molecules acting as hydrogen bond donors not only changed the molecular stacking mode in the hydrated crystals, but also markedly altered the electronic transition of the molecular keto form from nπ* to ππ* through the ESIPT process. This process enabled these small organic compounds to optically sense heavy water and quantify the ratio of H2O to D2O. This study provides a basis for the design and development of functional organic hydrate luminogens that exhibit water-induced fluorescence enhancement for solid-state device applications.

Methods

Synthesis of 2-(((1H-benzo[d]imidazol-2-yl)imino)methyl)-4-methoxyphenol (1a)

A mixture of 2-aminobenzimidazole (1.33 g, 10 mmol), salicylaldehyde (1.22 g, 10 mmol) and formic acid (0.1 mL) in ethanol (60 mL) was refluxed with vigorous stirring for 6 h. The precipitated compound was filtered and washed with ethanol (yield: 86 %). The product was confirmed by 1H NMR, 13C NMR, HR-MS. 1H NMR (400 MHz, DMSO-d6) δ (ppm): 12.81 (s, 1H), 12.15 (s, 1H), 9.69 (s, 1H), 7.89 (d, J = 6.9 Hz, 1H), 7.61 (s, 1H), 7.51 (dd, J = 11.2, 4.1 Hz, 2H), 7.22 (dd, J = 5.9, 3.1 Hz, 2H), 7.04 (t, J = 8.4 Hz, 2H). 13C NMR (151 MHz, DMSO-d6) δ (ppm) 166.03, 161.03, 154.34, 135.06, 132.89, 122.59, 120.06, 119.82, 117.31. HRMS (TOF MS EI+) calculated for C14H11N3O 237.0902, found 237.0905.

Synthesis of 2-(((1H-benzo[d]imidazol-2-yl)imino)methyl)phenol (1b)

A mixture of 2-aminobenzimidazole (1.33 g, 10 mmol), 2-hydroxy-5-methoxybenzaldehyde (1.52 g, 10 mmol) and formic acid (0.01 mL) in ethanol (80 mL) was refluxed with vigorous stirring for 6 h. The precipitated compound was filtered and washed with ethanol (yield: 91%). The product was confirmed by 1H NMR, 13C NMR, HRMS (TOF MS EI+). 1H NMR (400 MHz, DMSO-d6) δ 12.77 (s, 1H), 11.47 (s, 1H), 9.67 (s, 1H), 7.60 (s, 1H), 7.47 (d, J = 3.0 Hz, 2H), 7.20 (dd, J = 5.9, 3.1 Hz, 2H), 7.12 (dd, J = 9.0, 3.1 Hz, 1H), 6.97 (d, J = 9.0 Hz, 1H), 3.77 (s, 3H). 13C NMR (151 MHz, DMSO-d6) δ (ppm) 165.17, 155.22, 154.72, 152.62, 142.83, 122.73, 122.38, 119.90, 119.05, 118.33, 114.18, 111.66, 55.98. HRMS (TOF MS EI+) calculated for C15H13N3O2 267.1008, found 267.1005.

Synthesis of 2-(((1H-benzo[d]imidazol-2-yl)imino)methyl)-5-methoxyphenol (1c)

A mixture of 2-aminobenzimidazole (1.33 g, 10 mmol), 2-hydroxy-4-methoxybenzaldehyde (1.52 g, 10 mmol) and formic acid (0.01 mL) in ethanol (50 mL) was refluxed with vigorous stirring for 6 h. The precipitated compound was filtered and washed with ethanol (yield: 83%). The product was confirmed by 1H NMR, 13C NMR, and HRMS (TOF MS EI+). 1H NMR (400 MHz, DMSO-d6) δ 12.71 (s, 1H), 12.64 (s, 1H), 9.54 (s, 1H), 7.83–7.73 (m, 1H), 7.57 (s, 1H), 7.44 (s, 1H), 7.18 (d, J = 3.6 Hz, 2H), 6.69-6.55 (m, 2H), 3.84 (s, 3H). 13C NMR (151 MHz, DMSO-d6) δ (ppm) 165.59, 165.26, 163.60, 154.45, 142.86, 135.10, 134.38, 122.44, 122.23, 118.78, 113.35, 111.51, 108.17, 101.34, 56.10. HRMS (TOF MS EI+) calculated for C15H13N3O2 267.1008, found 267.1006.

Synthesis of 2-(((1-methyl-1H-benzo[d]imidazol-2-yl)imino)methyl)phenol (1d)

A mixture of 2-amino-1-methylbenzimidazole (1.47 g, 10 mmol), salicylaldehyde (1.22 g, 10 mmol) and formic acid (0.01 mL) in ethanol (100 mL) was refluxed with vigorous stirring for 6 h. The precipitated compound was filtered and washed with ethanol (yield: 87%). The product was confirmed by 1H NMR, 13C NMR, and HRMS (TOF MS EI+). 1H NMR (400 MHz, DMSO-d6) 11.67 (s, 1H), 9.72 (s, 1H), 8.00 (d, J = 7.6 Hz, 1H), 7.59 (dd, J = 16.7, 7.2 Hz, 2H), 7.51 (t, J = 7.7 Hz, 1H), 7.33–7.16 (m, 2H), 7.03 (t, J = 8.7 Hz, 2H), 3.87 (s, 3H). 13C NMR (151 MHz, DMSO-d6) δ (ppm) 165.00, 160.84, 154.33, 141.64, 135.95, 135.40, 131.77, 122.81, 122.57, 120.49, 120.21, 119.14, 117.29, 110.72, 29.34. HRMS (TOF MS EI+) calculated for C15H13N3O 251.1059, found 251.1054.

Preparation of 1a·H2O, 1a·D2O, 1b·H2O, 1b·D2O, 1c·H2O, and 1c·D2O microcrystals

In a typical procedure, a sample of 1a in ethanol (500 μL, 1 × 10−3 M) was injected drop-wise into deionized H2O or D2O (99.8 %, 3 mL). Well-defined 1a·H2O or 1a·D2O microcrystals were prepared by a facile self-assembly method. Samples were left to stand for 6 h to allow stabilization.