Article Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX
pubs.acs.org/IC
A Robust TbIII-MOF for Ultrasensitive Detection of Trinitrophenol: Matched Channel Dimensions and Strong Host−Guest Interactions Wenfeng Xu,†,§ Hanhua Chen,‡,§ Zhengqiang Xia,*,† Chongting Ren,† Jing Han,† Wujuan Sun,‡ Qing Wei,*,† Gang Xie,† and Sanping Chen*,† †
Key Laboratory of Synthetic and Natural Functional Molecule Chemistry of Ministry of Education, College of Chemistry and Materials Science, Northwest University, Xi’an 710127, China ‡ College of Chemistry & Chemical Engineering, Xi’an Shiyou University, Xi’an 710065, China
Downloaded by BETHEL UNIV at 08:51:03:956 on June 06, 2019 from https://pubs.acs.org/doi/10.1021/acs.inorgchem.9b01008.
S Supporting Information *
ABSTRACT: Host−Guest interaction is crucial to the sensitivity of heterogeneous sensors. Here, a series of isomorphic three-dimensional lanthanide metal−organic frameworks (Ln-MOFs), [Ln(TCBA)(H 2 O) 2 ] 2 ·DMF [H3TCBA = tris(3′-carboxybiphenyl)amine; Ln = Tb (1), Eu (2), and Gd (3); DMF = dimethylformamide] was synthesized and characterized, in which the propeller-like TCBA3− ligands adopt special torsional link between Tb(III) ions to form one-dimensional triangular channels. Optical experiments show that 1 exhibits bright green luminescence with an overall quantum yield of 26%, a 5D4 lifetime of 478 μs, and can act as an excellent heterogeneous fluorescent sensor to detect 2,4,6-trinitrophenol (TNP) explosive with an extremely low detection limit of 1.64 ppb. Because the confined channels within 1 exhibit matched dimensions toward TNP and feature multiple guest-response sites including rich πconjugated groups, electron-donating N centers, and open metal nodes, strong host−guest interactions between 1 and TNP are captured and accurately determined by online microcalorimetry, which provides a distinctive thermodynamic perspective to understand the heterogeneous sensing behaviors. Additionally, the finely modulated heterometallic isomorphism [Tb0.816Eu0.184(TCBA)(H2O)2]2·DMF emits bright white light when excited at 380 nm and could potentially be used as single-phase white light-emitting diode phosphors materials.
■
INTRODUCTION
Lanthanide metal−organic frameworks (Ln-MOFs), expecially the Eu/Tb-MOFs, are considered as excellent candidates for heterogeneous fluorescence sensing due to the fascinating pore variety and outstanding luminescence properties.16−21 On one hand, the large porosity and surface area allow the quick accumulation of the target analyte inside the special cavities for fluorescence response;22,23 on the other hand, the tunable pore sizes and multiple functional sites provide a tailorable pore environment for highly selective recognition of analyte.24,25 The literature shows that numerous Ln-MOFs have been used for the fluorescence detection of solvent molecule, metal ion, explosive, or harmful gas, and many of them have achieved satisfying sensing performances.19,26−29 Taking 2,4,6-trinitrophenol (TNP) detection as an example, ∼7 of the 33 LnMOFs (searched from Web of Science) show the limit of detection (LOD) of less than 0.1 ppm (Table S1),30 and especially the [La3(PDC)3Cl3(H2O)]n compound has attained a satisfying value of 0.22 ppb.31 However, it is not difficult to find that the vast majority of reports only directly give the fluorescent sensing results or ambiguous hypothesis and do not
Fluorescence sensing techniques, based on the analyteresponse fluorescence spectra, have attracted tremendous attention due to their multiple advantages of rapid response, high selectivity, and easy operation, as well as their wide applications on the detection of the harmful species in life and environment.1−4 At the same time, numerous fluorescent sensors with novel structures and advanced functions have been developed in response to the increasingly stringent environmental protection and testing standards in recent years.5−10 The conventional organic molecule probes, as the frequently reported ones, usually exhibit low detection limit, high sensitivity, and selectivity, because the target analytes rapidly react with probe in a completely contacted homogeneous system.11−13 Unfortunately, these probes are generally unrecyclable and hard-degraded (especially for the dye-based molecules)14 and need to be first dissolved and dispersed in toxic organic solvents,15 which causes lots of inconvenience, huge waste and environmental pollution, and limits their widespread applications. Apparently, the development of recyclable solid sensing materials with excellent detection performance is urgently important and challenging. © XXXX American Chemical Society
Received: April 8, 2019
A
DOI: 10.1021/acs.inorgchem.9b01008 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry Scheme 1. Synthetic Procedures of H3TCBA
■
provide more experimental information on why these LnMOFs show highly sensitive detectability. Obviously, current research mostly tends to be independent and scattered experimental accumulation while lacking systematic investigation on how to directionally design high-performing LnMOF sensors and what factors dominate the sensing performances. Given that the fluorescence emissions of Ln-MOFs are mainly caused by the energy transfer from chromophore to lanthanides, the interactions between the guest analyte and framework are the crucial factors to affect the energy transfer and produce a sensing signal. And these interactions can be tuned through the flexible confined microenvironments within Ln-MOFs (such as the pore size, π-conjugated groups, electron distribution, open metal sites, and H-bond sites).32−34 Therefore, the strength of the interactions has a direct effect on the detection limit of the sensors, and the tracking and capture of the interactions will undoubtedly allow us to better understand the sensing behaviors. As mentioned above, we propose a new strategy to interpret the sensing selectivity and sensitivity of Ln-MOF materials based on the direct quantification of the host−guest interactions between analyte and MOF framework. And the integration of the obtained interaction information helps to guide and modulate the original assembly proposal of LnMOFs for improving the detection capability of sensors. Herein, a novel propeller-like tricarboxylate ligand, tris(3′carboxybiphenyl)amine (H3TCBA, Figure S1), containing photofunctional triphenylamine chromophore and rigid aromatic skeleton, was prepared to construct three isostructural non-interpenetrated three-dimensional (3D) Ln-MOF materials, namely, [Ln(TCBA)(H2O)2]2·DMF [Ln = Tb (1), Eu (2), and Gd (3); DMF = dimethylformamide], under solvothermal conditions. The flexible arms of the C3-bridging H3TCBA ligand bind lanthanide ions with minimal repulsive interactions to form the robust framework with one-dimensional (1D) triangular channels, where the multiple πconjugated rings, open metal sites, and triphenylamine N Lewis basic sites are regularly distributed on the inner surfaces working as accessible sites in response to analytes. The systematic optical investigation revealed that 1 featuring high quantum yield, long fluorescence lifetime, and excellent luminous efficiency could be applied to detect TNP with a low detection limit of 1.64 ppb. The subtle intermolecular interactions between the framework and guest were monitored online and accurately determined by microcalorimetry method, and the results provide powerful thermodynamics support for understanding the sensing mechanism.
EXPERIMENTAL SECTION
Materials and Methods. All the nitroaromatic explosives are obtained from the Key Laboratory of National Defense Science and Technology for Combustion of Explosives, Xi’an Modern Chemistry Research Institute, China. Other chemicals are of reagent grade quality obtained from commercial sources and used without any further purification. 1H NMR and 13C NMR spectra were recorded on a Varian INOVA-400 MHz spectrometer. Elemental analyses for C, H, and N were recorded on an Elemental Vario EL III analyzer. Thermogravimetric analyses (TGA) were conducted using a NETZSCH STA 449F3 device at a rate of 10 °C min−1 from 30 to 900 °C in a nitrogen atmosphere. Fourier transform infrared (FT-IR) spectra were recorded on a Bruker Vertex 70 FI-IR spectrometer. UV absorption studies were performed with a Shimadzu UV-2450 spectrophotometer. The solid-state photoluminescence and lifetime measurements were performed on an Edinburgh FLS920 TCSPC fluorescence spectrometer. The fluorescence spectra and relative fluorescence intensity were measured with a Hitachi F-2700 fluorescence spectrometer. Powder X-ray diffraction (PXRD) patterns were obtained with a Bruker D8 Advance X-ray powder diffractometer (Cu Kα, 1.5418 Å). Inductively coupled plasma (ICP) spectroscopy was collected on an Agilent 725 ICP-OES spectrometer. The calorimetric experiment was performed using a microcalorimeter of Tian-Calvet type (C80 from Setaram) in air atmosphere. The N2 sorption isotherms were measured using a Micrometrics ASAP 2020 M adsorption instrument at 77 K. Synthesis. Synthesis of Tris(3′-carbomethoxybiphenyl)amine (Me3-TCBA). A mixture of tris(4-bromophenyl)amine (3.5 g, 7.26 mmol), (3-(methoxycarbonyl)phenyl)boronic acid (5.23 g, 29.04 mmol), palladium tetrakis(triphenylphosphine) (0.42 g, 0.36 mmol, 5 mmol %), and Cs2CO3 (9.46 g, 29.04 mmol) was stirred at reflux in anhydrous tetrahydrofuran (THF; 100 mL) for 24 h under argon atmosphere. After it cooled to room temperature, the solvent was evaporated under reduced pressure. The resulting yellow solid was dissolved in dichloromethane and washed several times with water. The organic phase was dried with anhydrous MgSO4 and concentrated under reduced pressure. The resulting crude product was purified by silica gel column chromatography using petroleum ether and dichloromethane (DCM) (1:1) as the eluent, and the pure product was harvested after removal of the solvents (4.3 g, yield: 92%). 1H NMR (400 MHz, CDCl3) δ 8.29 (t, J = 1.6 Hz, 3H), 8.03− 7.96 (m, 3H), 7.83−7.75 (m, 3H), 7.61−7.47 (m, 6H), 7.26 (t, J = 4.3 Hz, 3H), 5.30 (s, 6H), 3.96 (d, J = 3.6 Hz, 9H) (Figure S2). Synthesis of Tris(3′-carboxybiphenyl)amine (H3TCBA). KOH (15 g, 0.28 mol) was added to a suspension of Me3-TCBA (2.5 g, 3.86 mmol) in 120 mL of THF, and the mixture was stirred under reflux for 10 h (Scheme 1). After removal of THF in vacuum, 100 mL of water was added. The resulting mixture was heated, until the solid was fully dissolved, then the solution was acidified with concentrated (conc) HCl, until no precipitate formed (pH < 2). The green-yellow powder was collected by filtration, washed with water, and dried in air (2.25 g, yield: 90%). 1H NMR (400 MHz, deuterated dimethyl sulfoxide (DMSO-d6)) δ 13.05 (s, 3H), 8.18 (s, 3H), 7.91 (t, J = 8.7 Hz, 6H), 7.71 (d, J = 8.6 Hz, 6H), 7.58 (t, J = 7.7 Hz, 3H), 7.21 (d, J = 8.6 Hz, 6H) (Figure S3). 13C NMR (100 MHz, DMSO-d6) δ B
DOI: 10.1021/acs.inorgchem.9b01008 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Figure 1. Coordination environment of Tb(III) ion (a) and TCBA3− (b) in 1 (symmetry codes: A = 1 − x, 1 − y, 2 − z; B = 1 − x, 1 − y, 1 − z;). (c) The 3D framework of 1 with open channels, where the blue polyhedra represents the coordination sphere of Tb(III) in 1. (d) 3D topological of 1. 1616(m), 1549(m), 1418(m), 1062(m), 828(m), 753(s), 537(m), 462(w). Synthesis of [TbxEu1−x(TCBA)(H2O)2]2·DMF [x = 0.028 (4), 0.111 (5), 0.322 (6), 0.394 (7), 0.682 (8), 0.737 (9), and 0.816 (10)]. The compounds 4−10 were synthesized by following the same procedure as for the preparation of compound 1 but using mixed Tb(NO3)3· 6H2O and Eu(NO3)3·6H2O with different ratios. For details, see Table S3. X-ray Crystallography. Crystallographic data were obtained on a Bruker smart APEXII CCD diffractometer, equipped with graphitemonochromatized Mo Kα radiation (λ = 0.710 73 Å) by using ω and ϕ scan modes. Cell determination and data reduction were processed with the SAINT processing program. The absorption correction based on multiscan was applied in SADABS.35 All the structures were solved by direct methods with SHELXS and refined with full-matrix leastsquares refinements based on SHELXL-97.36 All non-hydrogen atoms were refined anisotropically. Hydrogen atoms were placed in geometrically calculated positions and refined as part of a riding model. The structural data and the structure refinement parameters of 1 and 2 are summarized in Table S4. The selected bond distances and angles are given in Table S5.
167.31, 146.62, 142.92, 133.92, 131.57, 130.67, 129.41, 128.05, 127.95, 126.92, 124.41 (Figure S4). Synthesis of [Ln(TCBA)(H2O)2]2·DMF [Ln = Tb (1), Eu (2), and Gd (3)]. For 1, a mixture of H3TCBA (0.0305 g, 0.05 mmol) and Tb(NO3)3·6H2O (0.0227 g, 0.05 mmol) was dissolved in 6 mL of mixed solvent of DMF/CH3OH/H2O (3:2:1, v/v). Then the solution was sealed in a 10 mL Teflon-lined autoclave followed by the addition of formic acid (150 μL) and heated at 100 °C for 3 d. When it cooled to ambient temperature, yellow areatus crystals of 1 were obtained, washed with DMF and H2O, and dried in air, yield: 57% (based on H3TCBA). Elemental analysis (%) calcd for C81H63N3O17Tb2 (Mr = 1668.20): C 58.32, H 3.81, N 2.51; found: C 58.49, H 3.95, N 2.72. IR (KBr cm−1): 3378(s), 2942(w), 1694(m), 1603(m), 1544(m), 1505(w), 1423(m), 1083(m), 888(w), 825(w), 767(s), 673(w), 540(w). For 2, an identical procedure with 1 was followed to prepare 2, except Tb(NO3)3·6H2O was replaced by Eu(NO3)3·6H2O (0.0223 g, 0.05 mmol). Yellow areatus crystals of 2 were collected in 53% yield (based on H 3 TCBA). Elemental analysis (%) calcd. for C81H63N3O17Eu2 (Mr = 1654.28): C 58.81, H 3.84, N 2.54; found: C 57.88, H 3.91, N 2.57. IR (KBr cm−1): 3369(s), 2928(m), 1692(m), 1549(m), 1418(m), 1090(m), 968(m), 771(s), 678(m), 537(w), 453(w). For 3, an identical procedure with 1 was followed to prepare 3, except Tb(NO3)3·6H2O was replaced by Gd(NO3)3·6H2O (0.0226 g, 0.05 mmol). Yellow microcrystals of 3 were collected in 59% yield (based on H3TCBA), whose PXRD patterns well matched with those of 1 and 2, indicating that the complexes 1−3 were isostructural (Figure S5). Elemental analysis (%) calcd. for C81H63N3O17Gd2 (Mr = 1664.84): C 58.44, H 3.81, N 2.52; found: C 57.16, H 3.83, N 2.49. IR (KBr cm−1): 3656(s), 2930(w), 1688(w),
■
RESULTS AND DISCUSSION Structural Description and Characterization. The compounds 1−3 were obtained by the solvothermal reaction of Ln(NO3)3·6H2O and H3TCBA with modest yields. The successful construction of isostructural 1−3 was confirmed through elemental analyses, PXRD, and FT-IR spectroscopy (Figures S5 and S6). So only the structure of 1 is described in C
DOI: 10.1021/acs.inorgchem.9b01008 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Table 1. Emission Lifetime (τobs), Radiative Lifetime (τRAD), Radiative (ARAD) and Nonradiative (ANR) Decay Rates, Intrinsic Quantum Yields (ϕLn), Sensitization Efficiency (ϕsen), and Overall Quantum Yield (ϕoverall) of 1 and 2 τobs
τRAD
ARAD
ANR
ϕLn
ϕsen
compound
μs
μs
s−1
s−1
%
%
%
1 2
478 ± 1 399 ± 1
864 ± 1 2551 ± 1
2114
55 16
48 22
26 ± 1 3±1
392
ϕoverall
Photophysical Property. The solid-state emission spectra of H3TCBA and 1−3 were measured at room temperature to investigate their luminescence properties (Figures S10−S14). Upon excitation at 380 nm, the H3TCBA ligand possesses blue emission with a maximum peak of 462 nm, which belongs to the ligand-based π → π* electron transitions. Compound 3 exhibits a similar emission with slight blue-shift (∼27 nm) when compared with that of the H3TCBA ligand, which may be attributed to the metal−ligand coordinative interactions.38 Compound 1 exhibits a series of strong characteristic emissions at 491, 545, 587, and 623 nm without ligand-based emission when excited at 380 nm, corresponding to the 5D4 → 7FJ (J = 6, 5, 4, and 3) transitions of Tb(III) ions, respectively. The most prominent emission peak at 545 nm, as noticed, is attributed to the hypersensitive 5D4 → 7F5 transition, endowing the whole bulky materials with bright green luminescence.39 In 2, a weak emission band with the peak of 480 nm and a series of sharp line emissions at 593, 614, 653, and 689 nm coexist, showing that the energy absorbed by TCBA3− ligand is not completely transferred to the metal centers but is enough to excite the characteristic Eu(III) emissions. The intense electric dipole transition of 5D0 → 7F2 at 614 nm impels red luminescence of 2. Apparently, through the radiative recombination of the ligand molecular excitation,40 the H3TCBA can be used as a suitable antenna ligand to sensitize the lanthanide ions luminescence and displays a better photosensitization efficiency for 1. Figures S15 and S16 depict the luminescent decay profiles of 1 and 2, and the lifetime values are 478 μs (5D4 level) and 399 μs (5D0 level), respectively. As shown in Table 1, the overall fluorescence quantum yields (ϕoverall) are estimated to be 26% and 3%, respectively, according to the method described by Bril et al.41,42 (for the detailed calculation processes, see the Supporting Information). For 1, a 55% value of intrinsic quantum yield (ϕTb) is determined based on the natural radiative time τRAD1 (864 μs) by employing the phosphorescence spectrum of 3 at 77 K (Figure S17),43 while 2 features a poor value of 16% (ϕEu) obtained from a τRAD2 value of 2551 μs calculated by using the Strickler−Berg formula. 44 Accordingly, the sensitization efficiency of H3TCBA (ϕsen) for 2 (22%) is determined to be much lower than that for 1 (48%). Such large difference should be ascribed to the unsatisfactory energy gap between the lowest triplet-state energy level of H3TCBA ligand and the 5D0 state energy level of Eu(III).45 Tunable Emitting Colors. Numerous fascinating colors can be easily created by mixing red, blue, and green of the three primary colors of light. Considering the blue-light emission of the H3TCBA and the characteristic red light of Eu(III) as well as the green light of Tb(III), a series of heterometallic Ln-MOFs, namely, [Tb x Eu 1−x (TCBA)(H2O)2]2·DMF [x = 0.028 (4), 0.111 (5), 0.322 (6), 0.394 (7), 0.682 (8), 0.737 (9), and 0.816 (10)], was prepared by adjusting the initial addition ratios of TbIII/EuIII during the self-assembly process to investigate the interesting emitting
detail. Single-crystal X-ray diffraction analysis reveals that 1 crystallizes in the monoclinic crystal system with P21/c space group and exhibits a non-interpenetrating 3D microporous structure. The asymmetric unit of 1 contains one crystallographically independent Tb(III) center, one fully deprotonated TCBA3− ligand, two coordinated water molecules, and half free DMF molecule. Tb1 center is eight-coordinated by six carboxyl oxygen atoms from five different TCBA3− ligands and two oxygen atoms from two coordinated water molecules, forming a square antiprism coordination geometry (Figure 1a). Two Tb(III) centers are further connected together by two carboxylate groups of two ligands to generate a dinuclear secondary building unit (SBU) with a Tb1···Tb1* separation of 4.695(5) Å. The TCBA3− ligand features a propeller-like bridging linker to extend along three different directions. The three carboxylate groups in TCBA3− adopt bidentate chelation (η2μ1χ2) and bidentate bridging (η2μ2χ2) modes to connect five different Tb(III) centers (Figure 1b). The η2μ2χ2-bridging carboxylate groups link adjacent dinuclear SBUs to generate 1D chain, and the neighboring chains are further interlinked by TCBA3− ligands to afford a 3D framework containing 1D channels with triangular window (size: 7.8 × 11.3 Å2) (Figure 1c). Topologically, both the dinuclear SBU and TCBA3− ligand can be simplified as five-connected nodes, yielding a (5,5)-connected topology with Schlafli symbol of {44; 66} (Figure 1d). The porosity of 1 accounting for 18.7% of the total crystal volume without solvent molecules is determined by the PLATON software,37 which is available for the diffusion of target analyte. The rich nitrogen/oxygen atoms and the multiple benzene rings distributed on the inner surfaces of the channels could act as accessible response sites to capture and sense guest molecules. Furthermore, the coordinated water molecules of 1 could be easily excluded and depart to form exposed Tb(III) metal sites for anchoring and recognizing specific analyte. Noteworthily, 1−3 show excellent stability. The similar TGA curves reveal that the frameworks could be solidly maintained up to 520 °C after undergoing a transitory desolvation process, and the activated sample of 1′ was obtained by heating the acetone-exchanged 1 sample in a vacuum oven at 150 °C for 12 h (Figure S7). So the permanent microporous structure of 1′ was further established by the N2 adsorption experiment at 77 K. The isotherm displays a typical Type-I adsorption behavior with a Brunauer− Emmett−Teller (BET) surface area of 294.8 m2 g−1 (Figure S8). The solvent stability of 1′ was also examined by soaking the samples in different solvents for 3 d. The PXRD patterns of the immersed samples match well with that of the assynthesized one, implying that the skeleton of 1′ is wellpreserved (Figure S9). These results suggest that the pores of 1−3 exhibit high chemical stability in multiple mediums or under high temperature environment, endowing the Ln-MOFs with ideal potential as sensing materials for substrate recognition in heterogeneous systems. D
DOI: 10.1021/acs.inorgchem.9b01008 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Figure 2. (a) Emission spectra of 4−10. (b) CIE chromaticity diagram of 4−10. (c) Photographs of the LEDs. [1. A 3 mm reference ultraviolet LED (turned off). 2. The illuminating LED showing a blue tinge (turned on). 3. The same LED coated with a thin layer of the powder of 10 (turned off). 4. The obtained LED turned on and generating bright white light. 5. The same WLED placed in air for one week (continuously turned on)].
Notably, the compound 10 with an appropriate emissionintensity ratio of the TCBA3−, Eu(III), and Tb(III) reaps a bright white light with chromaticity coordinates of (0.3316, 0.3229), which could be utilized as promising white-light materials. So a simple white light-emitting-diode (WLED) assembly is fabricated by coating a thin layer of the power sample of 10 on a commercially available ultraviolet lightemitting diode (LED; Figure 2c). The generating WLED exhibits bright white light at a voltage of 3.8 V and could continuously illuminate for a week without brightness decay. These results demonstrate that the compound 10 features excellent white-light performance and photostability and is of potential practical applications in white-light-emitting devices. Solvent Effect. Generally, the solvent molecules are easy to diffuse into the channels of Ln-MOFs to produce disordered molecular disturbance and nonradiative energy loss during the process of photosensitization.27 So it is necessary to investigate the luminescence behavior of the activated sample of 1′ in different solvents to select the optimal sensing medium. Before that, the thermal behavior of the guest-adsorption process of 1′ was monitored by using a Setaram C80 type microcalorimetry to quantitatively obtain the apparent energy change of host− guest interaction. According to Table S6, the diverse values and negative signs of the heat of absorption for the 13 solvent systems indicate that all the guest-adsorption processes are
colors of the materials. The highly consistent PXRD patterns of 4−10 and 1 reveal that these Ln-MOFs materials are isomorphic (Figure S18). The actual TbIII/EuIII ratios of the freshly synthesized heterometallic Ln-MOFs were determined by inductively coupled plasma (ICP) spectroscopy and listed in Table S3. The solid-state optical properties measurements show that all the characteristic emission peaks of the TCBA3− ligand, Eu(III) and Tb(III) ions could be incorporated in each emission spectra of 4−10 (Figure 2a). With the increasing of the Tb(III) ion content, both the characteristic emissions of Eu(III) (614 nm) and TCBA3− ligand (∼450 nm) quickly grow up to saturation, whereas the characteristic emissions of Tb(III) (545 nm) are tardily revived even when the Tb(III) ion content reaches a high level. This phenomenon may be because of the fact that the 5D0 state energy level of Eu(III) is lower than the 5D4 state energy level of Tb(III), and the energy absorbed by ligands prefers to be transferred to Eu(III) and then to Tb(III) in such heterometallic systems. Additionally, the poor sensitization for Eu(III) causes the incomplete energy transfer from TCBA3− ligand, leading to the part energy release in the form of π → π* transition. Chromaticity analysis reveals that the finely tunable emission color can be achieved by modulating the ratio of Tb(III)/Eu(III), showing a shift of the chromaticity coordinate from blue to light green and to the white regions (Figure 2b). E
DOI: 10.1021/acs.inorgchem.9b01008 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Figure 3. (a) Luminescence intensity at 545 nm of 1′ dispersed in different solvents. (b) Luminescence spectra of 1′ suspension with the addition of different nitro-explosives (1.0 mM). The plots with different colors represent the sensor 1′ and different quenchers: blue, 1′; red, DTTP; magenta, NB; cyan, 2,4-DNT; green, 2,6-DNT; violet, P-DNB; dark yellow, TNP. (c) Fluorescence titration spectra of 1′ suspension with the addition of various amounts of TNP. (d) The time-dependent emission spectra of 1′ in the NB atmosphere. (inset) The self-made simple device for the detection of volatile nitroaromatic compounds.
exothermic but exhibit distinguishing heat flow change. The nitrobenzene (NB) system attains a maximum enthalpy value of −6.125 kJ·mol−1, while those for other 12 solvent systems are all less than −2.000 kJ·mol−1, suggesting an obvious solvent selectivity effect and significant interactions between NB and the framework of 1′. Meanwhile, the very weak host−guest interactions are demonstrated in EtOH system by the obtained minimum value of the heat of absorption, −0.167 kJ·mol−1, which is less than one percent of that of NB. The results reveal that the EtOH system gives the minimal interference on the fluorescence emission of 1′ and is a suitable liquid-sensing medium. The solvent selectivity effect was also confirmed by fluorescence method. As shown in Figure 3a, the emission spectra of 1′ dispersing within different solvents were conducted with an excitation wavelength of 380 nm. As expected, the fluorescence spectra of 1′ suspension are heavily dependent on the solvent molecules, and the photoluminescence intensity of 1′ exhibits the maximum in EtOH and the minimum in NB under the same conditions. Particularly in the case of NB, the nearly complete quenching reveals that 1′ has the potential for selective sensing of NB in comparison to other solvents. So the fluorescent response of 1′ to the low concentration of NB was further tested. With the increasing addition of NB into the ethanol suspension of 1′, the luminescence intensity of 1′ is gradually decreased to a quenching percentage of 52% at only 80 ppm and 83% at 200 ppm (Figure S19), which shows higher quenching efficiency than the superior UiO type Zr-MOFs.46 The consistent
selectivity results caused from the microcalorimetry and fluorescence investigations reveal that the quantification of the host−guest interactions effectively avoids the blind selection of analyst and fluorescence sensor and provides forceful thermodynamic interpretation for the confused MOF sensing mechanisms. Luminescence Sensing of Nitro-Explosives. Given the similar chemical composition to NB, material 1′ was employed to detect a series of nitroaromatic explosives (1.0 mM) such as 1,4-dinitrobenzene (p-DNB), 2,4-dinitrotoluene (2,4-DNT), 2,6-dinitrotoluene (2,6-DNT), and 2,4,6-trinitrophenol (TNP). As expected, all the nitro-explosives featured significant fluorescence quenching effects on the emission of 1′ suspension with the quenching efficiency order of TNP ≫ p-DNB > 2,6-DNT > 2,4-DNT > NB (Figure 3b), and the fluorescence intensity decreased with the increase of the number of nitro groups. So the strong quenching probably arises from the electron-withdrawing properties of the nitro groups, which drives the directional donor−acceptor electron transfer from the rich electron-donating framework to the electron-deficient analyte.47,48 However, when a hexanitroaromatic compound, 5′-(3,5-dinitrophenyl)-3,3″,5,5″-tetranitrom-terphenyl (DTTP), was used as an analyte with the same condition, only a quenching percentage of ∼37% was detected, which was less than that of the mononitro NB (Figure 3b). To figure out that, the microcalorimetry experiments were further performed by dispersing 1′ in different EtOH solvents of nitroexplosives with the same concentration (1.5 mM). As can be seen from Table S7, most of the mixed systems exhibited F
DOI: 10.1021/acs.inorgchem.9b01008 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
between the inherent triangular channels of 1′ and triangular TNP molecules, which shortens the separation and facilitates closer interactions between electron-withdrawing nitro groups of TNP and the electron-donating triphenylamine nitrogencenters of the framework, accelerating the electron transfer (Figure S23 and Table S8). This process is confirmed by a significant decrease of luminescence lifetime from 478 μs for as-synthesized 1′ to 23.38 μs for TNP adsorbed 1′ (Figure S24). The other one is the dynamic quenching process that electron transfer takes place between the quencher and sensor in the excited state through collisions, which is reflected on the linearity-deviation S−V plot at the higher concentrations of TNP (Figure S20).51,52 The strong confined effect of 1′ toward TNP caused by the highly matched channel dimensions and conjugated polymer backbone of 1′ further accelerates the collisions, leading to the strong quenching efficiency and high sensing sensitivity for TNP. So the much higher enthalpy value of −23.84 kJ·mol−1 for the sensing system could be ascribed to a coexistence of strong enough π−π stacking and rich dipole−dipole interactions between TNP and the π-conjugated framework. Apparently, the triphenylamine moiety of the ligand containing conjugated aromatic ring and electron-donating nitrogen center plays a crucial role in the TNP detection of 1′. Moreover, the density functional theory (DFT) calculations applying the B3LYP/6-31G* method with Gaussian 09 were further performed to elucidate the electron transfer quenching mechanism (Figure 4 and Table S9).53 Generally, the
remarkable heat release, especially the trinitro TNP system, which yielded an adsorption enthalpy of −23.84 kJ·mol−1, which almost quintupled that of NB (−4.937 kJ·mol−1) and far exceeded the energy scope of weak van der Waals interaction (0−4 kJ·mol−1), while the enthalpy value for bulky DTTP system was small, −1.264 kJ·mol−1, suggesting weak host− guest interactions. According to the obtained thermodynamic information, a proper interpretation can be conducted such that the 1′ exhibits obvious size-selectivity sensing for nitroexplosives: the ponysize nitroaromatic compounds readily diffuse inside the channels of 1′ and give strong host−guest interactions with the inner multiple response sites to disturb the energy transfer between chromophore and lanthanides, leading to obvious fluorescence quenching; while the size of DTTP is too large to be adsorbed within the channels, only weak dipole−dipole interactions can be formed between the nitro and the triphenylamine N sites on the external surface of 1′ framework, resulting in slight reduction of the fluorescence intensity. Notably, the nearly complete luminescence quenching in TNP system shows stronger sensing ability of 1′ toward TNP than the other nitro-explosives. So the corresponding sensing sensitivity was evaluated by further fluorescence-quenching titrations with the incremental addition of trace TNP into the EtOH suspension of 1′. With the increasing of the concentrations of TNP, the 5D4 → 7F5 emissive intensity of 1′ at 545 nm gradually decreased, and the initial intensity decreased by 38% only at 1.63 μM and 53% at 3.25 μM, and a high quenching efficiency of 97.2% was recorded upon the addition of 16.0 μM of TNP (Figure 3c). The quantitative relationship between the luminescence intensity and the concentration of TNP was analyzed by the Stern−Volmer (S−V) equation: I0/I = KsvM + 1, where I0 and I represent the luminescence intensity of 1′ suspension before and after the introduction of TNP, respectively; M represents the concentration of TNP; Ksv represents the quenching rate constant. A nearly linear S−V plot (correlation coefficient R2 = 0.999 54) was observed at low concentrations (
