3.1 Structure description of [Cd2(L)2(bib)(H2O)2] (1)
An asymmetric unit of 1 was proven via the single-crystal analysis of the sample, constituting one Cd(II), a completely deprotonated L2− as the linker, half bib ligand, and one coordinated water molecule (Fig. 1a). The Cd(II) center cation in 1 with a coordination number of 7 has coordinated with two N atom donors in two bib [Cd1–N1 = 2.247(2) Å] and five O atom donors in two different L2− linker [Cd1–O1 = 2.5223(19), Cd1–O2 = 2.304(2), Cd1–O1 = 2.5090(18), Cd1–O6 = 2.328(2) and Cd1–O5 = 2.460(2)Å] and one O atom donor in the coordinated water. In this structure, two imidazole arms have a dihedral angle of 13.05°, while its value with a benzene core is 16.28°. The two carboxylate groups of L2− linker adopt µ1–η1:η1 and µ2–η2:η1 coordination modes. The L2− linker has cis-conformation and generates a “V”-type building block (Fig. 1b). Two benzene cores construct a dihedral angle of 8.15°. One 1D loop involving the polymeric ribbons running along “a” direction by the “V”-type L2− linker was formed via the Cd(II) joining (Fig. 1b). The Cd(II)…Cd(II) separation through the “V”-type L2− linker is ca. 12.32 Å. The adjacent 1D polymeric ribbon motifs in the crystal packing diagram (for 1) are strongly bonded via the intermolecular hydrogen bonds, resulting in the formation of a 2D supramolecular layer (Fig. 1c). The π-π interactions of the benzene ring resulted in the formation of a detailed 3D supramolecular motif (with centroid–centroid distance of 3.58Å) of two different “V”-type L2− linkers (Fig. 1d). Also, the lattice water is fixed in the C–H⋯O channels (C3⋯O5: 2.29Å; C13⋯O4: 2.35Å and C15⋯O3: 2.50 Å) as well as the O–H⋯O channels (O7–H7A⋯O2: 2.13 Å; O7–H7B⋯O6: 2.09 Å) created via the hydrogen bonding interactions (Fig. 1e). Also, the critical role of the hydrogen-bonding interactions to form and stabilize the 3D supramolecular architecture has been confirmed [20]. In addition, the bib linkers stretch across the loop, generating a new polyrotaxane-like structure (Fig. 1f).
2.3 Luminescent sensing
The photoluminescence (PL) spectra of 1 and H2L ligand were obtained at ambient conditions. In H2L, the appearance of the emission band was observed at 356 nm (λex = 285 nm), relating to the n–π* and π–π* transitions [21]. Likewise, a similar broad emission band centered at 360 nm was also obtained for 1 (λex = 285 nm) (Fig. S2). After assessing the photoluminescence properties of 1, its sensing properties were obtained towards antibiotics (vide supra). SMZ drastically decreased the emission intensity of 1, while this diminishing effect is much less in the presence of other antibiotics (Fig. 2a and Fig. S3). Hence, the sensing/probing ability of 1 against SMZ was investigated via the fluorescence titration in detail (Scheme S1), which confirmed that the SMZ addition in the range of 0-600 ppm completely quenches the emission of 1 (Fig. 2b). In this concentration range, a good linear relationship was observed with a KSV value of 2.98 ×103 M− 1 and the limit of detection (LOD) of 1.05 ppm (Fig. 2c). Further, the high selectivity of 1 towards SMZ by 1 was confirmed by its significant quenching effect in the presence of other antibiotics (Fig. 2d).
However, a comparable LOD value was obtained for SMZ concerning the values previously reported for other materials [22]. Feizollahi et al. reported Cu–Ag/GO/GCE sensors, which offered a well-defined signal for SMZ with a LOD of 0.46 ppm. While Jin and his coworker prepared a Tb-based MOF, that detects SMZ with a LOD of 1.58 ppm [22b]. A new UCNPs@MIP was also reported that had good binding capacity, rapid response, critical selectivity as well as good specificity for SMZ with LOD 0.34 ppm [1b]. It is worthy to mention here that, although the antibiotic-sensing performances of the materials mentioned above were evaluated, no investigation into their water stabilities was done. However, high water stability for 1 can be confirmed here, because all experiments were carried out in the aqueous phase. This stability was further complemented with the well-matched PXRD pattern for 1/ SMZ that agrees well with the PXRD pattern of the fresh sample of 1 (Fig. 3b). This indicated that no change in the structure of 1 was observed during sensing experiments and 1 maintained its structural integrity during the experiments.
Also, the recyclability of sensors to detect antibiotics is one very important parameter for their real-world application. Hence, the reusing of 1 after the detection process was done by its separation via the centrifugation of the reaction mixture (10000 rpm for 5 min). It was then ethanol washed 3 times and use in the next sensing run. The results confirmed that the sensing ability of the recovered 1 for the selectively SMZ detection is excellent for at least four recycling runs (Fig. 3a).
Figure 4 shows the results obtained in the XPS for 1 to evaluate its possible weak interactions with SMZ. Comparison of XPS before and after sensing revealed that the framework is stable (Fig. 4a). After SMZ sensing, the binding energies of N1s and O1s of 1 increase from 404.58 eV to 404.84 eV and from 530.78 eV to 531.30 eV, respectively. This indicated some interactions operated between 1 and SMZ during sensing experiments.
The sensing mechanisms confirmed the various reasons for sensing any analyte by a CP. One important reason is an overlap of the analyte electronic absorption band with the CP emission band. This transfers the CP resonance energy to the analyte. In the presented study, a maximum emission quenching of 1 was reached because of the complete overlap of the SMZ absorption band with the 1 emission spectrum (Fig. S4) [23–25].
Table 1
The HOMO-LUMO energies (in eV) for 1 and antibiotics.
Compound | HOMO | LUMO |
1 | -6.18 | -0.74 |
Thiamphenicol (THI) | -7.09 | -1.52 |
Secnidazole (SIZ) | -6.95 | -2.33 |
Tinidazole (TDZ) | -6.86 | -2.36 |
Dinitrazole (DTZ) | -6.93 | -2.27 |
Metronidazole (MDZ) | -6.96 | -2.35 |
Ornidazole (ODZ) | -7.17 | -2.59 |
Nitrofurazone (NZF) | -6.49 | -2.74 |
Chloramphenical (CAP) | -7.35 | -2.46 |
Sulfamethoxazole (SMT) | -6.23 | -0.82 |
Sulfamethazine (SMZ) | -5.92 | -0.92 |
The density functional theory (DFT) was used to calculate the energies of the HOMO–LUMO of the antibiotics used and 1 for illustrating their decreased effect in the emission intensity of 1 (Table 1, Fig. 5). The most probable reason for this decreased effect is the charge transfer from the excited 1 to the antibiotics’ LUMO. But, this can be achieved for a higher LUMO of 1 than the analytes’ LUMOs. As shown in Table 1, all the antibiotics have lower LUMO energy than 1, confirming the suitable energy positions of the antibiotics to accept the transferred charges from photo-excited 1. This charge transfer phenomenon leads to the relief in PL intensity of 1 [3b]. Despite other antibiotics, the LUMO difference between the 1 and SMT and SMZ is critically low (Fig. 5). Thus, a relatively better position is present in these sulfa drugs for accepting the transferred photo-excited charge from 1. Based on the results, a critically different quenching effect towards 1 was observed in presence of SMT and SMZ, confirming that the charge transfer process may not be the sole reason for the decreased PL intensity. Thus, besides electron/energy transfer, the probable weak interactions between 1 and SMT/SMZ may be another phenomenon responsible for the declined PL intensity. Apart from this, the competitive light absorption may be the other cause of a decline in emission intensity of 1, especially in the presence of SMZ.
Based on the literature [21b], interactions of 1 and antibiotics were evaluated by the construction of the molecular Hirshfeld surfaces in the crystal structure of 1 (Fig. 6). The MOF-dnorm plot (between − 0.5 to 1.5 Å) shows the deep red circular depressions, proving that strong interactions are present (Fig. 6a) [26]. Further, a small relatively shaded area on dnorm surface confirms the presence of the longer and slightly weaker interactions. The total and the decomposed fingerprint plots for 1 in Fig. 6b and 4c) show the presence of the opposite regions where one MOF moiety is a donor center (de>di) and another moiety is an acceptor center (de<di). Also, the characteristic fingerprint plots can show the specific atom pairs with close contacts to investigate the share of various solid states' weak interactions. In 1, the O∙∙∙H/H∙∙∙O interactions reveal individual edges in the corresponding 2D fingerprint plots in the region of 0.6 Å < (de + di) < 2.2 Å as the light sky-blue pattern in full fingerprint 2D plots with 17.0% sharing. Similarly, such individual edges can be observed for the N∙∙∙H/H∙∙∙N interactions in the region of 1.4 Å < (de + di) < 2.5 Å as a light sky-blue pattern with 1.5% sharing. This proves a facile luminescence quenching phenomenon via the 1 charge transfer by some extensive weak interactions with the antibiotics investigated.
The Crystal Explorer was also used to calculate the crystal lattice void space for 1 [21b] with the standard void cluster parameter “unit cell + 5.0 Å” [22] (Fig. 6d). For this goal, pro-molecule surfaces capped within the unit cell consisting of all atoms of the molecule were generated within a crystal radius of 5.0 Å. The results showed the lattice void volume of 189.16 Å3 for 1 with a void area of 482.19 Å2 with globular indices and asphericity of 0.330 and 0.029 units, respectively. These confirm an efficient interaction of the antibiotics with the CP, resulting in an efficient charge/energy transfer from its photo-excited state to the antibiotics' LUMO position relating to the apt energy separation between both antibiotics and MOF. Also, deep red circular depressions over the MOF were proven from the dnorm (Fig. 6a). This is evidence of the strong MOF-antibiotic interactions, which are responsible for the sensing of the analytes by an efficient electron/energy transfer.