A Cobalt-Organic Framework with Sensitive Detection of Fe3+, Cr2O72–, and CrO42– Ions in Water

A six-coordinated 3D metal–organic framework (MOF), namely, [Co(BPDC)(3-bpdb)(H2O)2]n (1) (H2BPDC = 4,4′-biphenyldicarboxylic acid, 3-bpdb = 1,4-bis(3-pyridyl)-2,3-diaza-1,3-butadiene) was purposefully constructed and characterized by single-crystal XRD, IR, elemental analyses, PXRD, SEM, and TGA. Crystal structural analysis revealed that the complex consists of a cds-type three-fold interpenetrated framework. Hirshfeld surface analysis yielded details of the surface characteristics of 1. Significant O−H···O hydrogen bonding interactions, which could promote the stability of the framework, were found is extremely stable in aqueous solution and can resist acids and bases over an extensive pH range of 2–13. 1 shows brilliant fluorescent emission in the solid state and in aqueous solution. Fluorescence titration, cyclic, and anti-interference experiments demonstrated that 1 is an excellent probe for Fe3+, CrO42‒ and Cr2O72‒ in water. The Ksv values of 1 for Fe3+, CrO42‒, and Cr2O72‒, which were as high as 1.06 × 104 M−1, 1.50 × 104 M−1, and 1.16 × 104 M−1, respectively, were comparable with those of other sensors. The quenching mechanism of the novel probe was subsequently explained.


Introduction
Recent developments in society and significant improvements in living standards have brought increased attention to the state of the environment and its relation to human health. Following the establishment of modern industries, heavy metal ion pollution and toxic inorganic anion contamination from industrial effluents slowly became among the greatest environmental threats to humans and the ecological environment. [1][2][3][4]. Fe(III), a trivalent metal ion considered essential for all organisms, plays a vital role in a number of crucial processes. A shortage or excess of this metal in the body results in serious functional disorders [5,6]. Cr(VI) ions serve as important oxidants in industrial production but can cause acute diseases even at very low concentrations [7]. As ion pollution caused by industrial development runs counter to the current pursuit of green environmental protection, solutions to this problem must be quickly developed. However, because of portability problems and the cost of transporting and maintaining sophisticated instruments in actual field use, the current detection methods for these ions are limited [8]. Researchers are actively developing new materials for wastewater treatment and pollution prevention. For example, nanostructure semiconductor-mediated photocatalysis can effectively degrade organic pollutants in water [9]. Inorganic metal oxide semiconductor-based photocatalysts play a key role in the photocatalytic treatment of water pollution [10], and photocatalytic nanostructured semiconducting thin films can be used for the photodegradation of toxic chemicals in the environment [11]. Polymer membrane pervaporation separation for alcohol dehydration is believed to be a suitable candidate technology to replace traditional distillation methods [12]. Developing a cost-effective and selective material that can effectively detect both Fe(III) and Cr(VI) is an urgent and important undertaking. Fluorescent probes can address the limitations of the available materials and achieve simple and fast detection.
Fluorescent metal-organic frameworks (MOFs) are excellent functional materials that have been synthesized in large quantities. These frameworks usually have fascinating structures with novel topologies and can be functionally modified. Over the last few years, significant progress has been made in the detection of metal ions using fluorescent MOF materials [13][14][15][16]. In general, the fluorescent MOFs are constructed with aromatic ligands because the π electron systems of the ligands endow the MOFs with photoluminescence properties and provide potential supramolecular recognition sites between the backbones and analytes [17]. The mixed-ligands approach is implemented by the judicious choice of carboxylates with aromatic π rings and N-donor linkers can afford novel fluorescent MOFs [18,19]. Among the transition metal ions, Co 2+ is widely used in the synthesis of MOFs because of its excellent coordination ability and potential function as an electrochemically active center; moreover, it is widely abundant and inexpensive. However, compared with other d 10 electron configuration transition metal ions (e.g., Zn 2+ , Cd 2+ ), Co 2+ is rarely used in fluorescence MOF sensors [20][21][22].
Because of its large π-conjugated structure, 4,4′-biphenyldicarboxylic acid (H 2 BPDC) is considered a good mixed ligand component. The molecule presents excellent flexibility through the free rotation of two benzene rings or carboxylate groups and phenyl rings, which eventually leads to structural diversity, as well as good coordination ability and stability under acidic and basic conditions [23][24][25][26]. As a neutral bis-pyridyl-based ligand, 4-bis(3-pyridyl)-2,3-diaza-1,3-butadiene (3-bpdb) has been reported to be an efficient bidentate bridging connector for the construction of coordination architectures [27][28][29][30]. The ligand not only possesses conformational freedom but also exhibits the nonbonded functional sites of the diaza group, which could interact with metal ions. Given the aforementioned considerations, Co 2+ was assembled with the π-conjugated organic molecules H 2 BPDC and 3-bpdb to obtain a 3D luminescent Co-MOF, [Co(BPDC)(3-bpdb)(H 2 O) 2 ] n (1), which was then structurally characterized. The excellent fluorescent characteristics of 1 demonstrate its potential use as a potential fluorescent probe for detecting Fe 3+ , CrO 4 2-, and Cr 2 O 7 2-.

Materials and Methods
3-bpdb was prepared according to the reported literature [31]. H 2 BPDC were obtained from Adamas company, purity 98% + , grade RG. Co(NO 3 ) 2 ·6H 2 O were obtained from Beijing Yinuokai Technology Co., Ltd., purity 99%, grade RG. All other analytical-grade chemicals and solvents were commercially available and used as received without further purification. Elemental analyses for C, H, and N were carried out with an Elementar Vario ELIII analyzer. The IR absorption spectra of all the complexes were collected using KBr pellets on a Thermo Scientific FTIR-Nicolet iS10 spectrometer in the range 4000-400 cm -1 . PXRD patterns of these compounds were recorded on a Rigaku X-ray diffractometer with Cu-Kα radiation (λ = 1.5418 Å). Simulation of the PXRD patterns were performed using single-crystal data and processed with the Mercury 3.7 program available free of charge through the Internet at http://www.iucr.org. Luminescence spectroscopy of all the compounds was performed on a Lengguang (Shanghai, China) F98 fluorescence spectrophotometer, and polarization-dependent photoluminescence studies were performed on a Hitachi F-7000 fluorescence spectrophotometer. The UV-Vis spectra were obtained using UV-240 IPC spectrophotometer. TGA data were acquired using a Mettler-Toledo simultaneous thermal analyzer from room temperature to 800 °C under an N 2 atmosphere with a heating rate of 10 °C min −1 . Molecular Hirshfeld surface calculations were performed using the CrystalExplorer 17.5 program [32]. Scanning electron microscopic (SEM) investigations were carried out using an FEI QUNTA field emission gun 650; prior to the SEM examination the specimens were sputter coated with gold. Variable-temperature susceptibility measurements of crystalline samples were performed on a Quantum Design MPMS-XL SQUID magnetometer. Diamagnetic corrections were performed from Pascal's constants, and an experimental correction for a sample holder was also applied. Point symbol and topological analyses were conducted by using the TOPOS program package [33].

Crystallographic
The selected single crystal with suitable dimensions was mounted on a glass fiber and used for X-ray diffraction analyses. Crystallographic data were collected at 293 K on a Bruker Smart APEX II CCD diffractometer with graphite-monochromated Mo-Kα radiation (λ = 0.71073 Å) using ω-scan technique. Empirical absorption corrections were applied using the SADABS program [34]. The structure was solved in each case by using the SHELXL package and refined by the full matrix least-squares method based on F 2 using the SHELX-2014 program [35][36][37][38]. All of the nonhydrogen atoms were located in different Fourier syntheses and finally refined with anisotropic thermal parameters. Hydrogen atoms attached to the organic moieties were either located from the difference Fourier map or fixed stereochemically. Details of the crystallographic data collection and refinement parameters are summarized in Table 1, selected bond distances and angles are given in Table S1, hydrogen bonds are given in Table S2. The CCDC number of 1 was assigned to be 1953740.

Crystal Structure of Complex 1
Single-crystal XRD analysis revealed that 1 is isostructural with our recently reported Ni-based MOF, [Ni(BPDC) (3-bpdb)(H 2 O) 2 ] n [30]. As illustrated in Fig. 1a, the structure of 1 consists of one Co(II) atom, one BPDC 2anion, one 3-bpdb ligand, and two coordinated water molecules. Co1 is located on a crystallographic inversion center and shows a slightly distorted octahedral geometry (Scheme S1). The equatorial plane is built by two water-oxygen atoms and two carboxylate-oxygen atoms from two independent BPDC 2ligands (Co1-O1 = 2.072(2) Å; Co1-O3 = 2.129(2) Å). Two nitrogen atoms from two 3-bpdb ligands occupy the axial positions, with the Co − N 3-bpdb bond length measuring 2.157(2) Å. These values are in agreement with those reported for Co(II) complexes [20][21][22]. The BPDC 2anions in 1 display bis-single-dentate coordination modes and link Co(H 2 O) 2 units to form 1D linear Co(H 2 O) 2 (BPDC) chains with Co···Co separations of 15.16 Å (Fig. 1b). As illustrated in Fig. 1c and d, these liner chains are bridged at Co(II) 0.266, -0.343 atoms by μ 2 -3-bpdb spanning two different extending directions in an ABAB fashion to generate a 3D architecture. In 1, the free torsions in BPDC 2between two phenyl rings and between phenyl rings and carboxyl groups are 39.83(10)° and 3.87(1)°, respectively (Scheme S2). Topologically, each six-coordinated Co(II) cation can be simplified as a 4-connect (4-c) node, and the ligands serve as linkers; thus, the whole structure of 1 can be represented as a uninodal, 4-c, cds-6 5 .8 net (Fig. 1e). The single cds net has a very large void, which is partially occupied by the two other nets to form a threefold-interpenetrated framework (Fig. 1f). Significant O−H···O hydrogen bonds between H atoms from coordinated water molecules and the uncoordinated carboxylate O atoms of BPDC 2are found in 1 (Fig.  S1). These interactions stabilize the MOF molecules and strengthen the rigidity of the compound to a great extent, thereby promoting luminescence sensitivity [39]. 1 and [Ni(1,4-bdc)(3-bpdb)(H 2 O)] n exhibit the same 3D cds topology. The main difference between these two molecules is that the H 2 BPDC ligand has a longer connection length than 1,4-benzenedicarboxylic acid (1,4-H 2 bdc), which increases the degree of interpenetration from twofold in the above compound to threefold in 1. Ligands with longer connection lengths were previously expected to yield MOFs with larger pores, but the pores generated were often occupied via the phenomenon of interpenetration.

Hirshfeld Surface (HS) Analysis
The Hirshfeld surface is a suitable tool for describing the surface characteristics of molecules [40]; it is used to analyze information about all existing weak/strong interactions in a crystal system, which is important in the crystal engineering pint of view. The molecular Hirshfeld surfaces of 1 were generated using a standard (high) surface resolution mapped with d norm (Fig. 2) and d i , d e , curvedness, shape index (Fig. S2). In Fig. 2, the regions of 1A in bright red reflect hydroxyl oxygen O−H and carboxylate O Significant intermolecular interactions mapped on the molecular surface can be clearly presented through 2D fingerprints. As shown in Figs. S3 and 4, 37.1% H···H interactions and 25.9% C···H/H···C interactions account for over half of the total number of interactions identified. The proportion of 15.7% represents typical O-H···O hydrogen bonds. The other hydrogen bonds of C-H···N may contain 8.7% N· ·H/H···N interactions. Reciprocal C···C contacts with 4.3% refer to weak π···π stacking, which is also important in the formation of the structure of 1. These analytical results are highly consistent with the crystallographic structure of 1.

PXRD, IR and Thermogravimetric Analyses of 1
The phase purity of the as-synthesized complex was confirmed by PXRD measurements. As shown in Fig. 3, the diffraction peaks of the bulk sample are consistent with the simulated pattern of the single-crystal data, which indicates that 1 mainly has one crystalline phase.
As shown in Fig. S5, the broad band of 1 in the 3300 − 3500 cm −1 region implies the existence of coordinated water molecules and hydrogen bonding interactions, consistent with the crystal structure of the sample. The absence of characteristic bands at approximately 1700 cm −1 indicates the complete deprotonation of carboxylate groups in H 2 BPDC after reaction with the metal ions [41]. The vibrations of v as (COO − ) and v a (COO − ) are observed at 1577 and 1391 cm −1 , respectively. The absorption bands observed at about 430 cm −1 could be ascribed to metal-O stretching vibrations. These results are consistent with the crystal structure of 1. Fig. S6 is the IR absorption spectra of 1 and 3-bpdb.
The thermal stability of 1 was studied by TGA (Fig. S7). The TG curve of 1 indicates no obvious weight loss from ambient temperature to 172 °C, a weight loss of 6.32% from 172 °C to 225 °C, which is consistent with the loss of two coordinated water molecules (calcd: 6.61%), then the anhydrous compound successively decomposes.

Luminescence Behaviors
The solid-state fluorescence of the ligands and 1 was studied. The spectrum of H 2 BPDC in Fig. 4a shows an emission band at 402 nm upon excitation at 346 nm, which can be ascribed to the ligand centered electronic transitions, that is, π* → n or π → π* transitions [42]. The spectrum of 3-bpdb shows a maximum emission peak at 394 nm and a shoulder peak at 469 nm (λ ex = 254 nm, Fig. S8) [28]. 1 shows an emission peak at 397 nm (λ ex = 343 nm), which is close to characteristic emission peak of the free H 2 BPDC ligand. Since non-d 10 metal-based coordination polymers are usually emission silent due to the d-d transition, the emission of 1 at 397 nm can be ascribed to the mixed contribution of intra-ligand and ligand-to-ligand charges [21][22][23]. To the best of our knowledge, fluorescent coordination compounds are mainly based on Cd(II) and Zn(II) with a d 10 electronic configuration; luminescent Co(II) MOFs are relatively rare. According to the literature [30], the isostructural compound [Ni(BPDC)(3-bpdb)(H 2 O) 2 ]n does not have similar luminescence properties.
The fluorescence properties of 1 in diverse solvent suspensions were investigated to explore its potential use as a probe for detecting ions. Finely ground samples of 1 were dispersed in different solvents, including H 2 O, acetone, acetonitrile (CH 3 CN), dimethyl acetamid (DMA), N, N-Dimethylformamide (DMF), dimethyl sulfoxide (DMSO), ethanol, iso-propanol, n-hexane, tetrahydrofuran (THF), and toluene, and then ultrasonicated for 30 min to form stable suspensions for the fluorescence studies. As shown in Fig. 4b, different solvents have different effects on the luminescence intensity of 1, which can be attributed to solvent effects. The slight shift in the emission peak of 1 compared with its solid-state emission spectrum may be ascribed to interactions between the homogeneously dispersible framework and solvent molecules with different polarities [43].

Stability of 1 in Acid-Based Solution
Stable existence in acidic/basic solutions is another important property of probes. 1 was immersed in acidic/basic solutions of pH 1-14 for 24 h, filtered and then subjected to PXRD. As shown in Fig. S9, the patterns of 1 before and after immersion treatment are consistent in the pH range of 2-13, which implies that the framework of 1 remains intact. These results demonstrate that 1 can exist stably in acidic/ basic solutions.

Selective Sensing of Cations
Because 1 has good fluorescence properties and can resist acids and bases in aqueous solution, it may be used as a probe for detecting ions. Various aqueous solutions of M(NO 3 ) x (0.1 M) containing different metal ions (M x+ = Na + , Mg 2+ , Al 3+ , K + , Ca 2+ , Cr 2+ , Mn 2+ , Co 2+ , Ni 2+ , Cu 2+ , Zn 2+ , Ag + , Cd 2+ , Pb 2+ , and Fe 3+ ) were separately dropped into a suspension of 1, which was obtained by dissolving 1 in aqueous solution and sonicating for 30 min to reach a stable state. The luminescence spectrum of the resulting solutions was obtained under same conditions. As shown in Figs. 5a and S10, only the solution containing Fe 3+ display a significant quenching effect on the emission intensity of 1, which indicates that 1 may be a potential fluorescent sensor for Fe 3+ with high selectivity and sensitivity. The sensitivity of 1 to Fe 3 + ions was explored via quantitative fluorescence titration experiments. Fe 3 + solution (50 mM) was gradually added dropwise to a suspension of 1, after which the fluorescence intensity of the solution was measured at 393 nm. As shown in Fig. 5b, the emission intensity of 1 continuously decreases with increasing Fe 3+ concentration. The quenching efficiency of Fe 3+ ion pair 1 is as high as 92% when the concentration of Fe 3+ is 0.79 mM (32 μL). The quenching efficiency of 1 was calculated by using the Stern-Volmer equation, (I 0 /I) = K sv [A] + 1. The SV plot for Fe 3+ displays a very good linear relationship at low concentrations, and the K sv value of the sensor toward Fe 3+ is calculated as 1.06 × 10 4 M −1 , which is comparable with that of other sensors for Fe 3+ (Fig. S11) [16,44,45] The recyclability of the probe determines whether it can be reused. Recovery experiments of 1 as an Fe 3+ probe were carried out. Here, 50 mg of 1 was first immersed in Fe 3+ solution (50 mM) to form Fe 3+ @1. The sample was washed five times with distilled water to remove the Fe 3+ , filtered, and then analyzed by PXRD. This procedure was repeated thrice. As shown in Fig. 6a, the fluorescence intensity of 1 was recovered compared with that of Fe 3+ @1. PXRD also demonstrated that the framework of 1 remains intact after interacting with Fe 3+ (Fig. 6b).
Because several ions often coexist in solution, the accurate detection of target metal ions under the interference of other metal ions must be ensured. Competitive experiments of Fe 3+ @1 in the presence of other metal ions were performed. Fe 3+ solution (1.6 mmol) was dropped into a suspension of 1 containing other metal ions (1.6 mmol). The fluorescence intensity of 1 was then detected and recorded, as shown in Fig. S12. The quenching effect caused by Fe 3+ ions remains unchanged in the presence of other metal ions, thereby indicating that 1 can selectively identify Fe 3+ against the interference of other analogs.  (Fig. 7b, d). The K sv values of 1 for CrO 4 2and Cr 2 O 7 2-, at 1.50 × 10 4 and 1.16 × 10 4 M −1 , respectively, are comparable with those of other sensors reported as luminescent probes for these ions (Fig. 7c, e) [4,16]. Recovery experiments of 1 as a CrO 4 2-/Cr 2 O 7 2probe was carried out using the protocol adopted for the recovery experiments with Fe 3+ . As shown in Fig. S14, the fluorescence intensity of 1 was recovered compared with those of CrO 4 2-@1 and Cr 2 O 7 2-@1. PXRD demonstrated that the framework of 1 remains intact after regeneration. The anti-interference ability of 1 was measured through competitive experiments by the same method (Fig. S15). The quenching effect caused by CrO 4 2-/Cr 2 O 7 2is fairly stable in the presence of other anions, thereby indicating that 1 can selectively recognize CrO 4 2-/Cr 2 O 7 2and resist the interference of analogues. Fig. 6 a The luminescence intensity of 1 and that found after three recycles. b The powder X-ray diffraction patterns of 1 after cyclic sensing Fe 3+ ions 1 3

Quenching Mechanism
The quenching mechanism was investigated to understand the high selectivity and sensitivity of 1 toward Fe 3+ and CrO 4 2-/Cr 2 O 7 2better. The PXRD patterns of 1 were obtained after ion identification, and patterns identical to those of the as-synthesized materials were obtained. This result indicates that the skeleton of the samples is unchanged following ion detection. (Figs. 6b, S14b and S14d). 1 (20 mg) was also immersed in 0.1 M Fe(NO 3 ) 3 , K 2 CrO 4 , and K 2 Cr 2 O 7 aqueous solutions for 2 days and then subjected to SEM analysis. The SEM micrographs obtained show no obvious changes in the surface of the material after treatment with different solutions compared with that of the  (Fig. S16). Hence, quenching is not caused by the destruction of the crystal framework. The quenching mechanism of 1 was explored by collecting the UV-Vis spectra of Fe 3+ and CrO 4 2-/Cr 2 O 7 2in water. (Fig. S17). The excitation spectrum (maximum emission peak, 346 nm) of 1 clearly overlaps with the absorption spectra of Fe 3+ and CrO 4 2-/Cr 2 O 7 2-. Thus, the luminescence quenching effect is due to competitive absorption between Fe 3+ and CrO 4 2-/ Cr 2 O 7 2and 1.

Conclusions
In summary, [Co(BPDC)(3-bpdb)(H 2 O) 2 ]n (1) was effectively synthesized under hydrothermal conditions and showed a threefold interpenetrated 3D structure and with a 4-c cds-type topology. The crystal displayed not only good thermal stability but also excellent stability in aqueous solution and could resist acids and bases over a wide pH range. Ion recognition experiments demonstrated that 1 can selectivity identify Fe 3+ and CrO 4 2+ /Cr 2 O 7 2+ with excellent regenerability and strong anti-interference ability. These results indicate that 1 is a promising candidate fluorescent sensing material for detecting Fe 3+ , CrO 4 2+ , and Cr 2 O 7 2+ .