Ultrasound-Assisted Synthesis of Luminescent Micro- and Nanocrystalline Eu-Based MOFs as Luminescent Probes for Heavy Metal Ions

The luminescent coarse-, micro- and nanocrystalline europium(III) terephthalate tetrahydrate (Eu2bdc3·4H2O) metal-organic frameworks were synthesized by the ultrasound-assisted wet-chemical method. Electron micrographs show that the europium(III) terephthalate microparticles are 7 μm long leaf-like plates. According to the dynamic light scattering technique, the average size of the Eu2bdc3·4H2O nanoparticles is equal to about 8 ± 2 nm. Thereby, the reported Eu2bdc3·4H2O nanoparticles are the smallest nanosized rare-earth-based MOF crystals, to the best of our knowledge. The synthesized materials demonstrate red emission due to the 5D0–7FJ transitions of Eu3+ upon 250 nm excitation into 1ππ* state of the terephthalate ion. Size reduction results in broadened emission bands, an increase in the non-radiative rate constants and a decrease in both the quantum efficiency of the 5D0 level and Eu3+ and the luminescence quantum yields. Cu2+, Cr3+, and Fe3+ ions efficiently and selectively quench the luminescence of nanocrystalline europium(III) terephthalate, which makes it a prospective material for luminescent probes to monitor these ions in waste and drinking water.

In our current study, we report the room-temperature ultrasonic-assisted wet chemical method of the synthesis of the small-sized luminescent Eu 2 bdc 3 ·4H 2 O MOFs including 8 nm nanoparticles-the smallest nanosized rare-earth-based MOF crystals, to the best of our knowledge. The luminescent properties of the coarse-, micro-and nanocrystalline europium(III) terephthalate are studied. In addition, the selective luminescence quenching by heavy metal ions is also reported.

Synthesis
The europium(III) terephthalate was obtained by mixing the EuCl 3 and Na 2 bdc solutions. Sample 1 was synthesized by a slow mixing of equal volumes of the 2 mM Na 2 bdc and 1 mM EuCl 3 solutions accompanied by vigorous stirring (Table 1). Sample 2 was synthesized by a slow mixing of the equal volumes of the 2 mM Na 2 bdc solution and the solution containing 1 mM EuCl 3 and 20% PEG-6000, accompanied by ultrasonication (40 kHz, 60 W) and vigorous stirring. The white precipitates of europium(III) terephthalate (Samples 1 and 2) were separated from the reaction mixture by centrifugation (4000× g) and washed with deionized water 5 times. Sample 3 was synthesized by a slow mixing of equal volumes of 1 mM Na 2 bdc and 0.5 mM EuCl 3 accompanied by ultrasonication (40 kHz, 60 W) and vigorous stirring. The obtained clear solution was centrifugated at 7500× g; however, no solid was precipitated. The addition of both polar (methanol and acetone) and non-polar solvents (ethanol-dichloromethane mixture) did not result in the salting-out of any solid. Therefore, we used the solution of Sample 3 in the further experiments. All experiments were performed at the temperature of 25 • C.

Characterization
The morphologies of the microstructures of the synthesized Samples 1 and 2 were characterized using scanning electron microscopy (SEM) with a Zeiss Merlin electron microscope (Zeiss, Germany) equipped with the energy-dispersive X-ray spectroscopy (EDX) module (Oxford Instruments INCAx-act, UK). X-ray powder diffraction (XRD) measurements were performed on a D2 Phaser (Bruker, USA) X-ray diffractometer using Cu Kα radiation (λ = 1.54056 Å). The particle size distribution of the aqueous solution of Sample 3 was revealed by the dynamic light scattering technique with an SZ-100 Series Nanoparticle Analyzer (Horiba Jobin Yvon, Japan) The luminescence spectra were recorded with a Fluorolog-3 fluorescence spectrometer (Horiba Jobin Yvon, Japan). Lifetime measurements were performed with the same spectrometer using a pulsed Xe lamp (pulse duration 3 µs).
The absolute values of the photoluminescence quantum yields were recorded using a Fluorolog 3 Quanta-phi device. All measurements were performed at the temperature of 25 • C.

Morphology
A scanning electron microscope was used to observe the shape and the size of the particles in the synthesized materials. Sample 1, which was synthesized by a slow mixing of equal volumes of sodium terephthalate (2 mM) and europium chloride (1 mM) aqueous solutions, precipitated in the form of a polycrystalline solid with the average particle size of 120 ± 30 µm ( Figure 1). The observed species consisted of smaller particles stacked together forming dendrimer-like microparticle assemblies. The addition of the non-ionic surfactant (PEG-6000) to the reaction mixture and ultrasonication without a change in the Eu 3+ and bdc 2− concentrations (Sample 2) prevented the aggregation of the microparticles and resulted in the formation of individual microparticles (Figure 2a-c). The particles had a leaf-like shape with ratio length:width:height of about 13:5:1. The particles size was obtained from SEM images, the particle size distribution is shown in Figure 2d,e. The average length and width were calculated from these distributions and are equal to 7.1 ± 1.6 and 2.8 ± 0.8 µm, respectively. We found that under ultrasonication the solution remained clear to the eye when the concentration of Eu 3+ and bdc 2− was decreased twofold (1 mM Na 2 bdc and 0.5 mM EuCl 3 ) both in the absence and the presence of the surfactant (PEG-6000). We could not precipitate the solid from the reaction mixture using high-speed centrifugation or by salting-out using organic solvents. Therefore, the formation of the nano-sized particles of europium(III) terephthalate was supposed. In order to exclude the contribution of the PEG micelles to the experimental data, in further experiments we carefully studied the aqueous suspension of Sample 3 obtained by a slow mixing of equal volumes of the 1 mM Na 2 bdc and 0.5 mM EuCl 3 accompanied by ultrasonication and vigorous stirring without a PEG-6000 addition. The particle size distribution was revealed by a dynamic light scattering technique, resulting in the average particle size equal to about 8 ± 2 nm ( Figure 3). The SEM-EDX study of 3 aggregates formed by drying the reaction mixture on the silicon plate revealed the presence of Eu in the sample but did not determine the particle size due to the insufficient spatial resolution of the used SEM microscope. The direct observation of the species using TEM was also problematic because the high-energy electron radiation (>100 kV) burned out the sample due to the decomposition of an organic linker (terephthalate ion).
In our study, we have found that ultrasonication and PEG-6000 addition significantly decreases the particle size and prevents aggregation. The increase in particle size can be achieved via continuous growth of a particle or gradual aggregation of various particles or seed crystals. The contact of particles can reduce the total surface area in the aggregation process resulting in overall energy reduction. Ultrasonication can encourage surface tension between the species caused by the acoustic radiation force on a compressible particle [26]. The effect of PEG addition on the particle size can be explained by the well-known properties of surfactants including polyethylene glycol to be adsorbed on the particles or seed crystals that decrease their surface energy and prevents aggregation [27][28][29]. Surprisingly, we revealed that the twofold decrease in the reagents' concentration leads to size reduction for several orders. A recent kinetic study of zinc-2-methylimidazole MOF ZIF-8 [30] reported that nucleation and crystal growth rates non-monotonously depend on the concentration of the reagents. During the low concentrations of the metal ions and the organic linker, the 1:1 M:L complex dominates. This state is called the "pre-equilibrium".
Further nucleation is associative and fast because the central atom has several weakly coordinated solvent molecules and can easily react with other 1:1 complexes resulting in the formation of oligomeric secondary building units (SBUs) [31]. Increasing the concentration of the metal ions and the organic linker leads to the domination of 1:2 and 1:3 M:L complexes. The aggregation of 1:2 and 1:3 M:L complexes into SBUs is slower than 1:1 complexes, which results in slower nucleation. Therefore, nucleation is faster than the growth process in solutions containing low concentrations of the metal ions and the organic linker, which explains the formation of the smaller particle size of the MOFs crystallizing at low concentrations.

Crystal Structure
The X-ray powder diffraction (XRD) patterns were measured ( Figure 4) for Samples 1 and 2 to discover the crystalline phase of the obtained materials. We could not precipitate Sample 3 from the solution; therefore, the XRD pattern of Sample 3 was not measured. Analysis of XRD patterns demonstrated that synthesized materials 1 and 2 are isostructural with the Tb 2 bdc 3 ·4H 2 O [32], the typical crystalline phase of lanthanide terephthalates [22], which indicated that materials 1 and 2 were obtained in a form of Eu 2 bdc 3 ·4H 2 O. This structure is a three-dimensional metal-organic framework (MOF), where octacoordinated Eu 3+ -ions are bound to the two water molecules and six terephthalate ions through the oxygen atoms ( Figure 4). XRD peaks of Eu 2 bdc 3 ·4H 2 O in Samples 1 and 2 slightly diverge from their counterparts measured for Tb 2 bdc 3 ·4H 2 O reported previously [32]. To compare the structures of Eu 2 bdc 3 ·4H 2 O and Tb 2 bdc 3 ·4H 2 O materials, the refinement of unit cell parameters was performed for the Eu 2 bdc 3 ·4H 2 O samples ( Table 2). One can observe that the structure of coarse-crystalline Eu 2 bdc 3 ·4H 2 O (1) is slightly different from that of Tb 2 bdc 3 ·4H 2 O. The ionic radius of the octacoordinated Eu 3+ ion (1.066 Å) is slightly larger than that of the octacoordinated Tb 3+ ion (1.040 Å) [33], which most likely results in minor differences between Eu 2 bdc 3 ·4H 2 O and Tb 2 bdc 3 ·4H 2 O structures. The unit cell parameters of microcrystalline Eu 2 bdc 3 ·4H 2 O (2) are somewhat different, both from that of coarse-crystalline Eu 2 bdc 3 ·4H 2 O (1) and Tb 2 bdc 3 ·4H 2 O [32], which is likely caused by the surface defects due to the relatively small particle size of several micrometers. (1) is slightly different from that of Tb2bdc3·4H2O. The ionic radius of the octacoordinated Eu 3+ ion (1,066Å) is slightly larger than that of the octacoordinated Tb 3+ ion (1,040Å) [33], which most likely results in minor differences between Eu2bdc3·4H2O and Tb2bdc3·4H2O structures. The unit cell parameters of microcrystalline Eu2bdc3·4H2O (2) are somewhat different, both from that of coarse-crystalline Eu2bdc3·4H2O (1) and Tb2bdc3·4H2O [32], which is likely caused by the surface defects due to the relatively small particle size of several micrometers.

Luminescent Properties
Terephthalate ions are known to intensively absorb ultraviolet light, promoting them into the 1 ππ* singlet electronic excited state [32,34,35]. In europium(III) terephthalate, the 1 ππ* state efficiently undergoes the 3 ππ* triplet electronic excited state by intersystem crossing due to the heavy atom effect [35] followed by an energy transfer to 5 D1 level of the Eu 3+ ion, due to relatively close energy values of the lowest energy 3 ππ* excited state of terephthalate ion [35] (≈20000 cm −1 ) and 5 D1 level of Eu 3+ ion [36] (≈19000 cm −1 ). 5 D1 level of the Eu 3+ ion [36] then undergoes internal conversion followed by emission

Luminescent Properties
Terephthalate ions are known to intensively absorb ultraviolet light, promoting them into the 1 ππ* singlet electronic excited state [32,34,35]. In europium(III) terephthalate, the 1 ππ* state efficiently undergoes the 3 ππ* triplet electronic excited state by intersystem crossing due to the heavy atom effect [35] followed by an energy transfer to 5 D 1 level of the Eu 3+ ion, due to relatively close energy values of the lowest energy 3 ππ* excited state of terephthalate ion [35] (≈20,000 cm −1 ) and 5 D 1 level of Eu 3+ ion [36] (≈19,000 cm −1 ). 5 D 1 level of the Eu 3+ ion [36] then undergoes internal conversion followed by emission corresponding to 5 D 0 -7 F J (J = 0-5) transitions. Figure 5a presents the emission spectra of the europium(III) terephthalate series (1-3) upon 250 nm excitation into the 1 ππ* singlet electronic excited state of the terephthalate ion. The emission spectra include narrow lines corresponding to the transitions from excited 5 D 0 to lower 7 F J levels: 5 D 0 -7 F 0 (578 nm), 5 D 0 -7 F 1 (590 nm), 5 D 0 -7 F 2 (615 nm), 5 D 0 -7 F 3 (649 nm), and 5 D 0 -7 F 4 (697 nm). The emission spectrum of nanocrystalline 3 also contains spectrally broad band peaking at about 420 nm, which corresponds to the terephthalate phosphorescence [35]. The most prominent transitions in the emission spectra are magnetic dipole 5 D 0 -7 F 1 and forced electric dipole 5 D 0 -7 F 2 and 5 D 0 -7 F 4 transitions. The excitation spectrum (λ em = 615 nm) of nanocrystalline 3 resembles its UV-Vis absorption spectrum (Figure 5b) consisting of a 250 nm band as well as a 280 nm shoulder corresponding to the transitions into 1 ππ* singlet electronic excited states of the terephthalate ion. One can notice that emission bands corresponding to the f-f transitions of the Eu 3+ ion significantly broaden with the particle size reduction. Thus, the 5 D 0 -7 F 2 band of coarse-crystalline 1, microcrystalline 2, and nanocrystalline 3 have full width at half maximum (fwhm) equal to 48, 66, and 238 cm −1 , respectively. The smaller particles have larger surface-to-volume ratio and the number of structural defects, which results in a larger dispersion of energies of electronic levels of Eu 3+ ions caused by the larger non-uniformity of the local environment of europium ions [37,38]. The luminescence decay curves of europium(III) terephthalate (Figure 5c) are fitted by single-exponential functions: where time constant τ f corresponds to the observed lifetime of 5 D 0 level. The observed lifetime of 5 D 0 level of coarse-crystalline 1, microcrystalline 2, and nanocrystalline 3 were found to be equal to 393 ± 3, 371 ± 4, and 115 ± 2 µs, respectively. Luminescence decay is affected by the combination of radiative and nonradiative processes. Radiative decay rate is determined by dipole transition strength and localfield correction. Nonradiative processes include multi-phonon relaxation, quenching on impurities (e.g., O-H group of water molecules) and cooperative processes (cross-relaxation, energy migration). Detailed descriptions of these processes were provided in our earlier papers [39,40]. The radiative and nonradiative decay rates of Eu 3+ -doped phosphors can be calculated from the emission spectrum using 4f-4f intensity theory [41]. Magnetic dipole 5 D 0 -7 F 1 transition probability A 0-1 = A MD,0 ·n 0 3 = 14.65·1.5 3 = 49 s −1 . A MD,0 is the spontaneous emission probability of the magnetic dipole 5 D 0 -7 F 1 , 14.65 s −1 , and n 0 is the refractive index, 1.5 [34]. Radiative decay rates A 0-λ (λ = 2, 4) of the 5 D 0 -7 F λ emission transition can be obtained from this formula: where I 0-λ and ν 0-λ are the integral intensity and frequency of the 5 D 0 -7 F λ emission transition. The total radiative decay rate, A r , could be calculated by summing all the A 0-λ radiative decay rates (λ = 1, 2, 4). The total decay rate is reciprocal to the observed lifetime of 5 Table 3. Analyzing Table 2, one can see the quantum efficiencies of the 5 D 0 level and Eu 3+ luminescence quantum yields decrease in series 1-3 simultaneously with particle size, whereas nonradiative decay rate constants increase upon the size reduction. Smaller particles have larger surface-to-volume ratios, resulting in more efficient quenching of the 5 D 0 level and Eu 3+ by the water molecules in an aqueous solution [42]. Comparing the quantum efficiencies of the 5 D 0 level and Eu 3+ luminescence quantum yields values, one can notice that the η/Φ ratio is equal to 0.5-0.9, which indicates a very efficient energy transfer from initially excited terephthalate chromophore to the 5 D 0 level of Eu 3+ ion.

Sensing Transition Metal Cations
Previous studies demonstrated that the presence of impurities such as ions of transition metals (Fe 3+ , Cu 2+ , Pb 2+ , MnO 4 − , Cr 2 O 7 2− ) [43], and organic compounds (aromatic, nitroaromatic, carbonyl compounds) can significantly quench the luminescence of the Eu-based metal-organic frameworks [1][2][3][4][5][6][7][8][9] making them prospective for the design of luminescent sensors for various pollutants and explosives. To reveal the selectivity of the europium(III) terephthalate MOF luminescence quenching to the various metal cations, 80 µL of aqueous suspensions of coarse-crystalline 1 (C(Eu 3+ ) = 8 mM) was mixed with the 100 µL of metal salt solutions (C(M n+ ) = 100 mM; M n+ = Fe 3+ , Ca 2+ , Ba 2+ , Cr 3+ , Fe 2+ , Mg 2+ , Ni 2+ , Pb 2+ , Cu 2+ , Co 2+ , Zn 2+ , Cd 2+ ) or distilled water. After 30 min, the photographs of these solutions under 254 nm illumination were recorded (Figure 6a,b). It was found that the Eu-based red emission faded only in the presence of Fe 3+ , Cr 3+ and Cu 2+ ions ( Figure 6a) starting from metal ion concentration 10-50 mM (Figure 6b). The emission spectra of aqueous solutions of nanocrystalline 3 (C(Eu 3+ ) = 5 µM) in the absence and in the presence of various concentrations of Cu 2+ , Cr 3+ , and Fe 3+ ions (λ exc = 250 nm) indicate the quenching of Eu 3+ 5 D 0 -7 F λ luminescence by the above-mentioned metal ions (Figure 6c-e). The dependence of the 615 nm emission band intensity on the Cu 2+ , Cr 3+ , and Fe 3+ concentration is given in Figure 6f. The concentration dependence resembles the step-function, where luminescence intensity sharply falls starting from the certain concentration of metal ion: 1 µM of Cu 2+ and 30 µM of Cr 3+ or Fe 3+ . Surprisingly, we revealed that the addition of Fe 3+ ions resulted in simultaneous quenching for the Eu 3+ 5 D 0 -7 F λ luminescence (591, 615, and 697 nm bands) and the terephthalate phosphorescence (420 nm), whereas the addition of Cu 2+ and Cr 3+ ions almost failed to reduce the intensity of the terephthalate phosphorescence band at 420 nm ( Figures S1-S3, Supplementary Materials). This observation indicates a different quenching mechanism of Eu 3+ 5 D 0 -7 F λ luminescence by the above-mentioned metal ions. Most likely, Cu 2+ , Cr 3+ , and Fe 3+ ions somehow coordinate with the oxygens of terephthalate ligands, but Fe 3+ ions quench the 3 ππ* triplet electronic excited state of terephthalate ion, whereas Cu 2+ and Cr 3+ ions quench the 5 D 0 level of Eu 3+ . To reveal the complete quenching mechanism, one must study the excited-state dynamics of singlet and triplet electronic states of terephthalate ion, as well as the 5 D 0 level of Eu 3+ , depending on the heavy metal ion concentration by time-resolved transient absorption and luminescence spectroscopy methods. We have found that nanocrystalline europium(III) terephthalate MOF 3 demonstrates significantly lower limits of detection on Cu 2+ , Cr 3+ , and Fe 3+ ions than coarse-crystalline 1 (10-50 mM for coarse-crystalline 1 vs 1-30 µM for nanocrystalline 3, Figure 6b,f). This observation is explained by a larger surface-to-volume ratio of nanoparticles relatively to the bulk material, resulting in a higher luminescence quenching efficiency of the later materials due to a greater number of coordination sites. The sensitivity of our materials to Cu 2+ , Cr 3+ and Fe 3+ ions is comparable with the best reported luminescent MOF-based sensors reported previously (Table 4). Despite the higher sensitivity of electrochemical MOF-based sensors (Table 4), the luminescent sensors can be used for the design of relatively inexpensive express tests on heavy metal ions.

Conclusions
In summary, we reported the ultrasound-assisted wet-chemical synthesis and characterization of luminescent coarse-, micro-, and nano-crystalline Eu 2 bdc 3 ·4H 2 O MOFs. The particles of coarse-crystalline Eu 2 bdc 3 ·4H 2 O, which are synthesized by the mixing of sodium terephthalate and europium chloride aqueous solutions without ultrasound, are dendrimer-like microparticle assemblies with the average particle size of 120 ± 30 µm. The microcrystalline MOFs were prepared by mixing sodium terephthalate and europium chloride aqueous solutions with the addition of PEG-6000 in the presence of ultrasonication. The microparticles have the shape of leaf-like plates and an average size of 7.1 × 2.8 µm. The average size of Eu 2 bdc 3 ·4H 2 O nanoparticles, synthesized by the mixing of low-concentration sodium terephthalate and europium chloride aqueous solutions in the presence of ultrasonication, is equal to about 8 ± 2 nm. Thus, the reported Eu 2 bdc 3 ·4H 2 O nanoparticles are the smallest nanosized rare-earth-based MOF crystals, to the best of our knowledge. The emission spectra of synthesized materials exhibit narrow lines corresponding to transitions from excited 5 D 0 to lower 7 F J levels of Eu 3+ ion: 5 D 0 -7 F 0 (578 nm), 5 D 0 -7 F 1 (590 nm), 5 D 0 -7 F 2 (615 nm), 5 D 0 -7 F 3 (649 nm), and 5 D 0 -7 F 4 (697 nm). Size reduction resulted in a broadening of the emission bands. The Eu 3+ luminescence quantum yields, upon excitation into 1 ππ* singlet electronic excited state of terephthalate ion, were found to be of 10 ± 1%, 5 ± 1% and 1.5 ± 0.5% for coarse-, micro-and nanocrystalline Eu 2 bdc 3 ·4H 2 O MOFs, respectively. The nonradiative decay rate of nanocrystalline europium(III) terephthalate was significantly larger that the corresponding values of Eu 2 bdc 3 ·4H 2 O MOFs, which resulted from more efficient quenching of the 5 D 0 level and Eu 3+ by the water molecules in aqueous solution due to greater surface-to-volume ratio of nanocrystalline MOF. The Cu 2+ , Cr 3+ , and Fe 3+ ions efficiently and selectively quench the Eu 3+ 5 D 0 − 7 F λ luminescence of nanocrystalline Eu 2 bdc 3 ·4H 2 O MOFs starting from the relatively low concentrations of metal ion: 1 µM of Cu 2+ and 30 µM of Cr 3+ or Fe 3+ . The reported nanocrystalline europium(III) terephthalateis one of the most sensitive luminescent MOF-based sensorfor Cu 2+ , Cr 3+ and Fe 3+ ions (Table 4). Therefore, synthesized nanocrystalline Eu 2 bdc 3 ·4H 2 O MOFs can be considered promising luminescent probes for heavy metal ions in waste and drinking water.