SC-SC Anion-Assisted Linker Exchange within a Three-Dimensional Cu(II)-Triazole Framework: A Luminescent Probe for S2.

A three-dimensional (3D) binodal 3,5-connected net, {[Cu(MTP)(H2O)](NO3)}n (1) with the Schläfli symbol of {3·72}{32·75·83} can be transformed into a two-dimensional (2D) kagóme network with the Schlafli symbol of {32·62·72} in an irreversible single crystal-single crystal (SC-SC) guest-assisted linker exchange process. The product of this SC-SC represents the first luminescent probe for S2- based on triazole ligand.

Introduction

Single crystal–single crystal (SC–SC) linker exchange in metal–organic framework (MOF) materials is a topic of high interest due to the ability of the process to add in new functionalities or perform structural changes advantageous for potential applications.[1−3] Within the general field of SC–SC processes, guest molecule modification and metal ion exchange are commonly observed methods to trigger these advantageous changes in material structure.[4] Within the MOF field, there is a push to produce materials that exhibit interesting luminescent properties, providing opportunities for the development of new chemosensors.[5] These chemosensors can be designed for the selective sensing of species such as cations, anions, small molecules, volatile organic compounds (VOCs), biomolecules, and nitroaromatic explosives.[6−8]

Environmental sulfide buildup, sourced from both man-made waste and anaerobic bacteria, is quickly becoming a concern for human water supplies. The presence of sulfide is especially problematic in warm tropical climates, where sulfide biofilms are known to form on wastewater pipelines.[9] The formation of biological sulfides in water pipes is directly related to the corrosion of sewage pipes, with both the corrosion and sulfide buildup being known to affect human health.[10] A common and effective way to treat sulfide in pipelines is to add iron salts to form FeS precipitates.[11] These FeS compounds are utilized due to their low solubility, small size, and the difficulty associated with decomposing or redissolving them. However, the production of sulfides is significantly affected by temperature, with the amount of iron added needing to be varied to control the sulfide production. The excess loading of iron ions that results from this can be easily converted to zero-valent iron nanoparticles (Fe(0)-NPs). Increasing attention is being paid to the toxicology of nanoparticles, with Fe(0)-NPs, in particular, having noticeable toxicity toward bacteria, algae, fish, plants, and animals.[12] Therefore, to ensure the safety of drinking water, there is a growing need for the real-time monitoring of sulfur and iron ions in water. The traditional detection method for iron and sulfur relies on inductively coupled plasma-mass spectrometry (ICP-MS); however, the analysis procedure is tedious and brings with it a high monetary cost. Consequently, it is necessary to establish a simple, rapid, selective, and highly sensitive detection method.

Results and Discussion

Inspired by our previous experience in metal ion detection,[13] we have started investigating a MOF based upon a new ligand, 3-methyl-5-(4H-1,2,4-triazol-4-yl)pyridine (MTP). Using the tridentate MTP ligand, we constructed a binodal 3,5-connected three-dimensional (3D) net {[Cu(MTP)(H2O)](NO3)} (1) with the Schläfli symbol of {3·72}{32·75·83}. Interestingly, an anion-induced SC–SC transformation resulted in {[Cu0.5(MTP)(H2O)](NO3)} (1a), which exhibited a kagóme network (kgm) with a Schlafli symbol of {32·62·72}. 1a can also be generated via mixing this ligand and Cu(NO3)2·2.5H2O in a 20 mL capped vessel containing water and EtOH mixed solution and evaporating at room temperature for 1 week.[14] Our initial experiments on the luminescent properties indicated that 1a represented the first example of a triazole-based ligand showing selective luminescence-based Fe3+ detection.

The synthesis of MTP ligand is given in the Supporting Information. The solvothermal reaction of MTP and Cu(NO3)2·3H2O in a CH3CN and H2O mixed solvent system at 90 °C for 3 days led to the formation of blue block-shaped crystals of 1 in 70% yield (Supporting Information). X-ray structure analysis reveals that complex 1 crystallizes in the orthorhombic space group Pna21. The asymmetric unit of 1 consists of one crystallographically unique Cu2+ ion, one MTP ligand, one bridging water molecule, and one NO3– ion. As shown in Figure S3, Cu1 is coordinated to two Ntriazole atoms (N5 and N1A) from two distinct MTP ligands, alongside two oxygen atoms from water molecules, forming a square-planar motif. Additionally, one Npyrine atom from a MTP ligand coordinates with Cu1 in the axial position, producing the final distorted tetragonal pyramidal geometry on the atom. The Cu–N bond lengths are in the range of 1.9982(2)–2.4168(2) Å, while the Cu–O bond lengths are 1.9412(1) Å and 1.9148(1) Å, respectively, all of which are within the normal range for known copper complexes.[13e] The dihedral angles between the pyridine and triazole rings in 1 are 43.9, 44.2, and 44.6°, respectively, thus indicating different spatial-distortion effect to provide for the curvature required for coordination with the CuII ions and to dictate the direction of framework extension. The most striking feature of 1 is that the water molecules bridge two Cu(II) atoms alongside the N–N group of the MTP ligands to form an infinite one-dimensional (1D) chain (Figure S4). These 1D chains are linked together by four MTP ligands, giving rise to a three-dimensional (3D) framework, which has 1D channels with parallelogram-shaped windows along the c-axis, which are then filled with NO3– anions (Figure a). To get a better insight into the 3D framework of 1, topological analysis is carried out via the freely available computer program TOPOS.[15] Each Cu(II) atom connects to three MTP ligands as well as to two other Cu(II) atoms via two separate water molecules. Therefore, Cu(II) can be regarded as a five-connected node. Each MTP ligand in turn connects three Cu atoms and can thus be treated as a three-connected node. The resulting structure of 1 is a binodal 3,5-connected 3D net with the Schläfli symbol of {3·72}{32·75·83} (Figure b).

Figure 1

(a) View of the 3D framework of 1 along the c -axis with NO3– anions encapsulated inside the 1D channels. Red, O; blue, N; gray, C; cyan polyhedron, coordination polyhedron of the Cu2+ ions. (b) Schematic illustrating the topology of the binodal (3,5)-connected network with the Schläfli symbol {3·72}{32·75·83}. The blue spheres represent the CuII centers and the yellow spheres represent the MTP ligands.

Encouraged by our previous progress in the study of SC–SC transformations,[13]1 was immersed in an aqueous solution of NaNO3, yielding blue block crystals of 1a. Its crystallinity was maintained throughout the process, as shown by powder X-ray diffraction (PXRD) (Figures S5 and S6). Single-crystal X-ray diffraction analysis indicates that 1a crystallizes in the trigonal space group R-3c:H crystal system. The asymmetric unit of 1a consists of one Cu2+ ion, two MTP ligands, two coordinated water molecules, and two free nitrate ions (Figure S7). Each Cu2+ ion is six-coordinated, with its coordination sphere containing two Ntriazole atoms from two distinct MTP ligands, two Npyridine atoms from another two MTP ligands, and two water molecules, resulting in a CuN4O2 distorted octahedral coordination sphere. The Cu–O bond distance is 2.4230(2) Å, with the Cu–N distances varying between 2.0168(2) and 2.0318(1) Å. The dihedral angles between the pyridine and triazole rings in 1a are 56.7 and 57.8°, respectively, indicating a stronger spatial-distortion effect than that observed in 1a, providing the curvature required for coordination.

As illustrated in Figure , the MTP ligands link alternating CuII ions through its Npyridine and Ntriaozle atoms. In 1a, this arrangement of Cu–MTP interactions results in the formation of a two-dimensional (2D) layered material possessing two types of triangular windows. The larger windows are surrounded by six CuII ions and six MTP ligands and have a side length of 16.61 Å. Meanwhile, the comparatively small windows are surrounded by three CuII ions and three MTP ligands with a side length of 8.57 Å. The NO3– groups act as counteranions and occupy the void space of the smaller windows to stabilize the lattice. The CuII ions are regarded as four-connected nodes, while the MTP ligands act as bidentate connectors. Thus, a kagóme network (kgm) with the Schlafli symbol {32·62·72} can be formed. The whole structure of 1a can be viewed as a stacking of the kgm layers with the sequence of −ABCDEF– (Figure S8).

Figure 2

View of the two windows in the layers of 1a. Cyan, Cu; red, O; blue, N; and gray, C.

The thermogravimetric analysis (TGA) curves of 1 (Figure S9) show almost no mass loss event between 20 and 249 °C, followed by a sudden loss of about 61% of the total mass at 249 °C caused by the decomposition of the ligand. The resulting leftover hydrated copper nitrate then dehydrates gradually after reaching 395 °C, losing another 19% of its initial mass. Interestingly, the kagóme framework of 1a begins to decompose at about 216 °C, a significantly lower decomposition temperature compared to 1 (Figure S10). We believe this reduced thermal stability of 1a is due to the weaker interactions between the material 2D sheets as compared to the 3D connected framework of 1. The decomposition event of both materials is driven by the destruction of uncoordinated organic ligand, and the weaker interlayer connectivity of 1a results in a greater degree of organic linker isolation, enabling the linker to more easily thermolyze compared to the highly connected and interpenetrated ligand of 1. After ligand combustion, the material exhibits a copper-based dehydration event starting around 400 °C, similar to that observed in the TGA curve for 1.

Fluorescence probe techniques are a method based upon a material’s light absorption properties that produce both qualitative and quantitative information regarding the photochemical and photophysical properties of a material. These photochemical and photophysical properties are strongly correlated to the steric and electronic environments of a fluorescent species, as these alter both the wavelengths of light absorption and the relaxation pathways observed by the molecule after absorption. MOFs, as porous coordination complexes, can oftentimes have a strong, guest-dependent luminescent response based on the host−guest interactions. By analyzing this response, which is typically dependent on the analyte concentration, the sensing of a target analyte can be realized. Thus, MOFs are ideal fluorescent probes due to their high spectral stability; they can provide a simple, fast, highly sensitive, and highly selective real-time detection method.

To study the guest-dependent fluorescence response of 1a, solutions of Fe3+, Zn2+, Ba2+, Mg2+, Al3+, Li+, Cd2+, Na+, K+, La3+, Cu2+, Co2+, Mn2+, Ca2+, and Ni2+ were added to the suspensions of 1a, with the final concentration of the metal cations being kept at 200 ppm. As illustrated in Figure S11, only Fe3+ had a significant fluorescence quenching effect, with the other metal ions only slightly altering the fluorescence intensity of 1a. The results show that complex 1a has high selective recognition for Fe3+, suggesting that 1a can be used as a material for the selective luminescent detection of Fe3+. Figure a shows the fluorescence spectra of a series of 1a suspensions with different concentrations of Fe3+. When the Fe3+ concentration is gradually increased, the fluorescence intensity of 1a gradually decreases. As shown in Figures and S12, the fluorescence intensity decreased linearly within the 1–200 ppm concentration range.

Figure 3

(a) Luminescence spectra of 1a in the presence of different concentrations of Fe3+. (b) Linear relationship between the luminescence intensity and Fe3+ concentration in the 1–200 ppm range.

The Fe3+ fluorescence detection limit was calculated using the following equationF0 and F are the fluorescence intensities before and after the addition of Fe3+ to 1a, respectively. Ksv is the quenching rate constant; and [Fe3+] is the concentration of Fe3+. The detection limit is 3σ/k (k represents the slope between the fluorescence intensity and the concentration and σ stands for the standard deviation of blank runs).To further study the correlation between the quenching results, we drew a linear relationship between different Fe3+ concentrations and the relative values of fluorescence intensity (Figure b). The limit of detection for 1a in regards to Fe3+ is 0.34 μM (S/N = 3), showing this material as a potentially useful Fe3+ sensor. To study the effect of anions on the fluorescence reactions, solutions of PO43–, Cl–, SiO32–, IO3–, CO32–, BF4–, NO3–, OAc–, SCN–, SO42–, Br–, I–, F–, IO3–, BrO3–, SO32–, and ClO3– were added to a suspension of 1a. As shown in Figure a, S2– exhibited a significant enhancement to the material’s fluorescence, while the other anions only had a slight impact on fluorescence intensity. Figure b shows the fluorescence spectra of a series of 1a suspensions and with different concentrations of S2–. These data show that as the concentration of S2– gradually increases, the fluorescence intensity of 1a also increases. As illustrated in Figure S13, the fluorescence intensity increases linearly within the 1–200 ppm range for S2–.

Figure 4

(a) Specification test of the sensing strategy for S2– (200 ppm) against other anions at 200 ppm in 1a. (b) Luminescence spectra of 1a under different concentrations of S2–.

The S2– fluorescence detection limit was calculated by referring to the following equationF0 and F are fluorescence intensity before and after the addition of S2– to 1a, respectively. [S2–] is the concentration of S2–. The detection limit is 3σ/k.To further study the correlation between the enhancement results, the linear relationship between the concentration of S2– and the corresponding value of fluorescence intensity is shown in Figure S14. Based on this, the detection limit of 1a for S2– is also 4.6 μM (S/N = 3), proving that 1a also has potential as a S2– detector.

Conclusions

An anion-driven SC–SC transformation process has been utilized to generate a kagóme framework with the Schläfli symbol {32·62·72} from a 3,5-connected 3D net with the Schläfli symbol of {3·72}{32·75·83}. To the best of our knowledge, 1a represents the first selective sensor toward both Fe3+ and S2–. 1a can act as a sensor for both above ions through its fluorescence response, with Fe3+ serving as a fluorescence quencher and S2– perfoming as a fluorescence enhancer, allowing for both quantitative analysis of ion concentration and naked-eye detection (Figure ). As both Fe3+ and S2– are known environmental pollutants, these material show promise as a useful probe, enhancing our ability to quickly and efficiently measure the concentration of these harmful ions in our water systems.

Figure 5

Naked-eye detection of S2– in 1a with observable color change. (1) 1 μM of 1a; (2) 200 ppm of S2–; (3) 1 μM of 1a with 10 ppm of S2–; (4) 1 μM of 1a with 50 ppm of S2–; (5) 1 μM of 1a with 100 ppm of S2–; (6) 1 μM of 1a with 150 ppm of S2–; and (7) 1 μM of 1a with 200 ppm of S2–. The photograph was taken by Dr. Yue Liu.

Experimental Section

Synthesis of 3-Methyl-5-(4H-1,2,4-triazol-4-yl)pyridine

A mixture of 5-methylpyridin-3-amine (6.15 g, 56.87 mmol), N,N-dimethylformamide azine (16.17 g, 113.74 mmol), and p-toluenesulfonic acid hydrate (TsOH·H2O) (1.02 g, 5.35 mmol) in xylene (61.5 mL) was heated at reflux for 55 h. The reaction mixture was then cooled to room temperature. After decantation of the solvent, the residue was washed with petroleum ether until the wash solution was colorless; the solid was then recrystallized from CH3OH/2-propanol (1:1) to afford the crystalline product (4.60 g, 50.49%). 1H nuclear magnetic resonance (NMR) (400 MHz, dimethyl sulfoxide (DMSO)) (illustrated in Figure S1): δ 9.17 (s, 2H), 8.78 (s, 1H), 8.50 (s, 1H), 8.02 (s, 1H), 2.39 (s, 3H). 1C NMR data of MTP is illustrated in Figure S2.

Synthesis of {[Cu(MTP)(H2O)](NO3)} (1)

A mixture of MTP (0.032 g, 0.2 mmol), Cu(NO3)2·3H2O (0.0966 g, 0.4 mmol), H2O (6 mL), and CH3CN (4 mL) were put into a 20 mL acid-digestion bomb and heated at 90 °C for 3 days. Blue crystals suitable for single-crystal X-ray diffraction studies were collected after washing with H2O (2 × 5 mL) and diethyl ether (2 × 5 mL). Yield: 70%. Elemental analysis calcd (%) for C8H10CuN5O4: C 31.63, H 3.32, N 23.06; found: C 31.69, H 3.27, N 23.11.

Synthesis of {[Cu0.5(MTP)(H2O)](NO3)} (1a)

A mixture of 1 (0.5 mmol) and NaNO3 (5 mmol) in H2O (10 mL) was stirred for about 6 h. The resultant mixture was evaporated at room temperature for 2 weeks. Well-shaped blue block crystals suitable for X-ray diffraction were obtained. Yield: 35%. Elemental analysis calcd (%) for C8H10Cu0.5N4.33O2: C 41.67, H 4.37, N 26.30; found: C 41.72, H 4.30, N 26.33.