A series of Ln₄^III clusters: Dy₄ single molecule magnet and Tb₄ multi-responsive luminescent sensor for Fe³⁺, CrO₄²⁻/Cr₂O₇²⁻ and 4-nitroaniline

Abstract

Five tetranuclear lanthanide clusters of compositions [Ln₄L₄(NO₃)₂(Piv)₂]·2CH₃OH (Ln = Gd (1), Tb (2), Dy (3), Ho (4), Er (5); H₂L = 2-(((2-hydroxy-3-methoxybenzyl)imino)methyl)-6-methoxyphenol; Piv = pivalic acid) were synthesized under solvothermal conditions. The structures of 1–5 were characterized by single-crystal X-ray crystallography. Complexes 1–5 possess a zig-zag topology with [Ln₄O₆] cores being formed by the fusion of oxygen atom-bridged two [Ln₂O₂] moieties. Direct-current magnetic susceptibility studied in the 2–300 K range revealed weak antiferromagnetic interactions in 1, 2, 4, 5 and ferromagnetic interactions in 3. Complex 3 exhibits single molecule magnet (SMM) behavior. The luminescence studies indicated that complex 2 can serve as highly sensitive and selective luminescent materials for Fe³⁺, CrO₄²⁻, Cr₂O₇²⁻ and 4-nitroaniline (4-NA), demonstrating that complex 2 should be a potential candidate for multi-responsive luminescent sensor.

Five tetranuclear lanthanide clusters were synthesized. Dy₄ complex exhibits single molecule magnet (SMM) behavior and Tb₄ compound shows sensing properties towards Fe³⁺, CrO₄²⁻, Cr₂O₇²⁻ and 4-nitroaniline (4-NA).

Introduction

Polynuclear lanthanide clusters have been attracting considerable attention due to their fascinating structures¹ as well as potential applications in single molecule magnets (SMMs),² luminescent devices³ and magnetocaloric materials.⁴ SMMs are of considerable promise as molecular spintronic devices for high-density data storage.⁵ The significant anisotropy of lanthanides arising from large unquenched orbital angular momentum^2a,2u has made lanthanides to be attractive candidates for SMMs. Recently, the research efforts towards lanthanide SMMs have been devoted to polynuclear complexes with variable nuclearities.² Moreover, owing to the characteristic sharp emission peaks, a large Stokes shift and a wide emission range, the utilization of lanthanide complexes as luminescent materials, particularly luminescent sensor,⁶ also attracts intensive interest.

Fe³⁺ ion plays an important role in industry and in various metabolic processes.⁷ However, excessive Fe³⁺ ion probably leads to some diseases, such as Alzheimer's disease, due to production of reactive oxygen species (ROS).⁸ CrO₄²⁻ and Cr₂O₇²⁻ ions can cause severe environment pollution,⁹ gemmulation, renal failure, lung cancer, and skin allergy by prolonged exposure.¹⁰ Nitroaromatic compounds (NACs), e.g., nitroaromatic (NB), 4-nitrophenol (4-NP), 4-nitroaniline (4-NA), 4-nitrotoluene (4-NT), 4-nitrochlorobenzene (4-Cl-NB), 2,4-dinitrotoluene (2,4-DNT), 2,4-dinitrophenol (2,4-DNP), etc., are threatening human life.¹¹ Thus, it is of high importance to develop fast and facile methods for detecting those ions and compounds. A number of fluorescence-based sensing materials based on metal–organic frameworks (MOFs),¹² coordination polymers,¹³ and mononuclear compounds¹⁴ have been utilized to detect specific target inorganic anions, cations or organic molecules, due to high efficiency, selectivity and simplicity. Whereas, luminescent sensors base on lanthanide polynuclear clusters were rarely studied. The recent advances in this area indicated that they are potential sensors for nitroaromatic explosive compounds¹⁵ and F⁻ ion.¹⁶

We are interested in designing and synthesizing Ln^III clusters supported by o-vanillin-containing Schiff base ligand 2-(((2-hydroxy-3-methoxybenzyl)imino)methyl)-6-methoxyphenol (H₂L) and pivalic acid. The selection of both H₂L and pivalic acid as the ligands is based on the following considerations. (i) The combination of the o-vanillin-based ligand and pivalic acid fosters the formation of polynuclear clusters owing to the incorporation of the phenolic oxygen atoms from the H₂L ligand with the carboxylate group from pivalic acid. (ii) H₂L is an electron rich π-conjugated multidentate organic ligand, which can act as antennae to sensitize the weak luminescent metal centres, such as Tb(iii). (iii) Although the single molecule magnet behaviours of a series of Ln^III clusters have been studied, their multiple fluorescence detection abilities towards Fe³⁺, CrO₄²⁻/Cr₂O₇²⁻, and NACs have not been reported. In this work, five tetranuclear lanthanide clusters of compositions [Ln₄L₄(NO₃)₂(Piv)₂]·2CH₃OH (Ln = Gd (1), Tb (2), Dy (3), Ho (4), Er (5)) supported by H₂L were synthesized via solvothermal method. Complexes 1–5 possess [Ln₄O₆] cores formed by the fusion of two phenoxide oxygen-bridged two [Ln₂O₂] moieties. The magnetic studies indicated antiferromagnetic interactions in 1, 2, 4 and 5. Complex 3 exhibits typical single molecule magnet behaviour. The luminescent sensing properties of 2 for selective and sensitive detections of Fe³⁺, CrO₄²⁻/Cr₂O₇²⁻ ions in aqueous solutions and 4-NA in ethanol solutions were also investigated. The low detection limits and high quenching constants K_SV indicate that complex 2 may potentially be considered as a multi-responsive luminescence-based sensor for quantitative and highly sensitive detections of 4-NA, Fe³⁺ and CrO₄²⁻/Cr₂O₇²⁻ ions.

Experimental section

Synthesis of 2-(((2-hydroxy-3-methoxybenzyl)imino)methyl)-6-methoxyphenol (H₂L)

All solvents and regents were of analytical grade and used as received without further purification. 3-Methoxysalicylamine was prepared according to the reported procedure.¹⁷ A solution of methoxysalicylaldehyde (20 mmol) in ethanol was added to a solution of 3-methoxysalicylamine (20 mmol) in 15 mL ethanol. After the reaction mixture was stirred at room temperature for 4 hours, yellow precipitates were formed. The yellow precipitates were isolated, washed with hexane and dried in vacuum to afford the product as a yellow solid (Scheme 1). Yield: 4.76 g (83%). ¹H NMR (600 MHz, chloroform-d) δ 8.38 (s, 1H), 6.90–6.84 (m, 3H), 6.83–6.74 (m, 3H), 4.82 (s, 2H), 3.87 (s, 3H), 3.85 (s, 3H). ¹³C NMR (151 MHz, chloroform-d) δ 165.57, 152.66, 148.63, 146.39, 143.48, 123.64, 123.01, 121.34, 119.66, 118.56, 117.50, 113.90, 109.89, 56.60, 56.06, 56.04. Elemental anal. calcd for C₁₆H₁₇NO₄: C, 66.89; H, 5.96; N, 4.88%. Found: C, 66.77; H, 6.23; N, 5.18%. Selected IR data (KBr, cm⁻¹): 3400(w), 1630(m), 1593(w), 1487(m), 1474(s), 1461(m), 1440(m), 1351(w), 1315(w), 1277(m), 1249(s), 1237(s), 1216(m), 1188(m), 1162(m), 1074(m), 993(m), 979(m), 903(m), 842(w), 830(w), 801(w), 785(w), 738(s), 760(s), 743(s), 734(s), 710(w).

Scheme 1. Synthesis of the H₂L ligand.

Physical measurement

Elemental analyses for C, H, and N were carried out with a Perkin-Elmer 2400 analyser. Fourier transform (FT) IR spectra were recorded with a Bruker VERTEX 70 FTIR spectrophotometer in the range of 600–4000 cm⁻¹. Magnetic susceptibility measurements were performed in the temperature range of 2–300 K, using a Quantum Design MPMS XL-7 SQUID magnetometer. Powder X-ray diffraction (PXRD) was obtained from a Rigaku D/Max-2500 diffractometer at 40 kV and 100 mA with a Cu-target tube and a graphite monochromator. Fluorescence spectra for the samples were recorded on a FLS-920 fluorescence spectrophotometer. The UV-Vis absorption spectra were determined with an Agilent Cary-60 spectra photometer at room temperature. X-ray photoelectron spectroscopy (XPS) experiments were carried out by a ESCALAB 250Xi spectrometer using an Al Kα source at room temperature.

X-ray single-crystal data collection and structure determination

The diffraction data for 1–5 were collected on a Bruker D8-VENTURE (120 K) CCD X-ray diffractometer equipped with a graphite monochromated Mo Kα radiation (λ = 0.71073 Å). The ω–2θ scan technique was applied. The crystal structures of all complexes were solved with the Olex2 solve solution program¹⁸ using Intrinsic Phasing and refined by full-matrix least-squares minimization using the ShelXL refinement package.¹⁹ All H atoms were placed on appropriate positions in theory and their positions were refined using the riding model. The details of the crystal parameters, data collection and refinements for the complexes are summarized in Table 1. Selected bond lengths and angles with their estimated standard deviations are listed in Table S1.†

Crystal data and structure refinements for 1–5.

	1	2	3	4	5
Empirical formula	C76H86N6O28Gd4	C76H86N6O28Tb4	C76H86N6O28Dy4	C76H86N6O28Ho4	C76H86N6O28Er4
Formula weight	2160.50	2167.18	2181.50	2191.02	2200.54
Temperature/K	120	120	120	120	120
Crystal system	Monoclinic	Monoclinic	Monoclinic	Monoclinic	Monoclinic
Space group	P21/c	P21/c	P21/c	P21/c	P21/c
a/Å	13.4900(07)	13.4681(15)	13.457(03)	13.4321(16)	13.4131(7)
b/Å	11.9674(06)	11.9193(13)	11.905(02)	11.8480(14)	11.8637(7)
c/Å	24.5060(13)	24.483(3)	24.500(06)	24.450(3)	24.4580(14)
α/°	90	90	90	90	90
β/°	95.961(2)	95.984(4)	96.03	96.123(4)	96.222(2)
γ/°	90	90	90	90	90
Volume/Å3	3934.9(9)	3908.9(7)	3903.1(07)	3868.9(8)	3869.1(4)
Z	2	2	2	2	2
ρ calc g cm−3	1.823	1.841	1.856	1.881	1.889
μ/mm−1	3.413	3.661	3.872	4.133	4.381
F(000)	2128.0	2136.0	2144.0	2152.0	2160.0
Crystal size/mm3	02 × 0.04 × 0.04	0.4 × 0.05 × 0.05	0.40 × 0.05 × 0.05	0.40 × 0.05 × 0.05	0.25 × 0.24 × 0.23
θ Range	2.28 to 27.502	2.287 to 27.565	2.928 to 27.488	2.501 to 27.63	2.394 to 27.605
Index ranges	−17 ≤ h ≤ 17, −15 ≤ k ≤ 15, −31 ≤ l ≤ 29	−17 ≤ h ≤ 17, −15 ≤ k ≤ 15, −31 ≤ l ≤ 31	−17 ≤ h ≤ 17, −15 ≤ k ≤ 15, −26 ≤ l ≤ 31	−17 ≤ h ≤ 17, −15 ≤ k ≤ 15, −31 ≤ l ≤ 31	−17 ≤ h ≤ 16, −15 ≤ k ≤ 15, −28 ≤ l ≤ 31
Reflections collected	59 585	60 747	53 678	59 835	60 923
Independent reflections	9047 [R (int) = 0.0954]	8987 [R (int) = 0.0456]	8994 [R (int) = 0.0408]	8938 [R (int) = 0.0803]	8920 [R (int) = 0.0483]
Data/restraints/parameters	9047/18/533	8987/12/533	8944/12/533	8938/12/533	8920/12/533
Goodness-of-fit on F2	1.093	1.116	1.182	1.058	1.206
Final R indexes [I ≥ 2σ (I)]	R 1 = 0.0428, wR2 = 0.0616	R 1 = 0.0267, wR2 = 0.0513	R 1 = 0.0311, wR2 = 0.0531	R 1 = 0.0313, wR2 = 0.0603	R 1 = 0.0329, wR2 = 0.0546
Final R indexes [all data]	R 1 = 0.0737, wR2 = 0.0682	R 1 = 0.0405, wR2 = 0.0575	R 1 = 0.0383, wR2 = 0.0549	R 1 = 0.0500, wR2 = 0.0685	R 1 = 0.0413, wR2 = 0.0563
Largest diff. peak/hole/e Å−3	0.81/−0.88	0.90/−0.77	0.85/−1.08	0.85/−1.10	0.93/−1.41

Synthesis of 1–5

A mixture of Ln(NO₃)₃·6H₂O (Ln = Gd, Tb, Dy, Ho, Er) (0.1 mmol), H₂L (0.1 mmol), pivalic acid (0.1 mmol), triethylamine (0.2 mmol) and CH₃OH (2.5 mL) was sealed in a pyrex-tube (10 mL). The tube was heated for 24 h at 80 °C. After cooling to room temperature, yellow block crystals suitably for X-ray diffraction analysis were obtained.

[Gd₄L₄(NO₃)₂(Piv)₂]·2CH₃OH (1)

Yield of 1 is 47% (0.0254 g) based on Gd. Elemental anal. calcd for C₇₆H₈₆N₆O₂₈Gd₄: C, 42.25; H, 4.01; N, 3.89%. Found: C, 42.44; H, 3.99; N, 4.12%. Selected IR data (KBr, cm⁻¹): 2955(m), 1621(m), 1548(m), 1468(s), 1420(m), 1403(m), 1323(w), 1296(w), 1271(m), 1245(s), 1228(m), 1166(m), 1062(m), 1028(m), 1013(m), 977(w), 951(w), 914(w), 867(m), 785(m), 735(s), 612(w).

[Tb₄L₄(NO₃)₂(Piv)₂]·2CH₃OH (2)

Yield of 2 is 53% (0.0287 g) based on Tb. Elemental anal. calcd for C₇₆H₈₆N₆O₂₈Tb₄: C, 42.12; H, 4.00; N, 3.88%. Found: C, 41.74; H, 4.04; N, 4.07%. Selected IR data (KBr, cm⁻¹): 2956(m), 1622(m), 1549(m), 1469(s), 1420(m), 1402(m), 1323(w), 1288(w), 1271(m), 1245(s), 1229(m), 1166(m), 1063(m), 1028(m), 1013(m), 978(w), 952(w), 914(w), 868(m), 785(m), 735(s), 612(w).

[Dy₄L₄(NO₃)₂(Piv)₂]·2CH₃OH (3)

Yield of 3 is 54% (0.0294 g) based on Dy. Elemental anal. calcd for C₇₆H₈₆N₆O₂₈Dy₄: C, 41.84; H, 3.97; N, 3.85%. Found: C, 41.84; H, 4.05; N, 3.95%. Selected IR data (KBr, cm⁻¹): 2956(m), 1623(m), 1550(m), 1470(s), 1420(m), 1403(m), 1324(w), 1289(w), 1271(m), 1245(s), 1230(m), 1166(m), 1063(m), 1029(m), 1014(m), 979(w), 953(w), 916(w), 868(m), 785(m), 735(s), 613(w).

[Ho₄L₄(NO₃)₂(Piv)₂]·2CH₃OH (4)

Yield of 4 is 53% (0.0290 g) based on Ho. Elemental anal. calcd for C₇₆H₈₆N₆O₂₈Ho₄: C, 41.66; H, 3.96; N, 3.84%. Found: C, 41.47; H, 4.05; N, 4.00%. Selected IR data (KBr, cm⁻¹): 2956(m), 1623(m), 1551(m), 1471(s), 1421(m), 1403(m), 1325(w), 1290(w), 1272(m), 1245(s), 1231(m), 1166(m), 1063(m), 1030(m), 1014(m), 979(w), 953(w), 916(w), 868(m), 785(m), 736(s), 613(w).

[Er₄L₄(NO₃)₂(Piv)₂]·2CH₃OH (5)

Yield of 5 is 51% (0.0280 g) based on Er. Elemental anal. calcd for C₇₆H₈₆N₆O₂₈Er₄: C, 41.48; H, 3.94; N,3.82%. Found: C, 40.83; H, 4.03; N 3.93%. Selected IR data (KBr, cm⁻¹): 2956(m), 1623(m), 1552(m), 1471(s), 1421(m), 1402(m), 1326(w), 1290(w), 1272(m), 1243(s), 1231(m), 1166(m), 1063(m), 1030(m), 1014(m), 980(w), 953(w), 916(w), 868(m), 785(m), 736(s), 613(w).

Results and discussion

Synthesis of complexes

The tetranuclear complexes 1–5 were synthesized under solvothermal conditions. When a mixture of Ln(NO₃)₃·6H₂O (Ln = Gd, Tb, Dy, Ho, Er), H₂L (the synthesis of H₂L was shown in Scheme 1), and pivalic acid, in a 1 : 1 : 1 molar ratio in MeOH/Et₃N was sealed in a Pyrex-tube and heated under solvothermal conditions, yellow block crystals were generated after being heated at 353 K for one day. The experimental powder X-ray diffraction patterns matched well with those simulated from the crystal structures, demonstrating the purity of complexes 1–5.

Description of crystal structure of 1–5

Single crystal X-ray analyses and powder X-ray diffraction analyses indicated that complexes 1–5 are isostructural. They crystalize in the monoclinic P2₁/c space group (Table 1, Fig. 1). Thus, complex 3 was chosen as a representative for 1–5 to describe the structure in detail. The molecular structure of 3 contains four Dy(iii) ions in a zig-zag topology, four L²⁻ ligands, two deprotonated pivalic acids, two NO₃⁻ anions and two CH₃OH solvate molecules (Fig. 1). Two L²⁻ ligands coordinate in a μ₂: η¹, η², η¹, η¹ fashion whereas the other two L²⁻ ligands coordinate in a μ₃: η¹, η², η¹, η², η¹ mode (Scheme 2). The structure of 3 is composed by two asymmetric [Dy2] units which are connected by phenolate bridges from the Schiff base ligands with Dy⋯Dy distances of 3.697(2) Å (Dy1⋯Dy2) and 3.839(2) Å (Dy2⋯Dy2A). Dy1 centre is coordinated by two oxygen atoms (O1 and O3) from one ligand, one nitrogen atom (N1) from the same ligand, two oxygen atoms (O5 and O6) from another ligand set, two oxygen atoms (O11 and O12) from one NO₃⁻ ion, and one oxygen atom (O10) from a deprotonated pivalic acid, forming a NO7 coordination environment. Dy2 centre was coordinated by two oxygen atoms (O7 and O9) from one ligand, one nitrogen atom (N2) from the same ligand, four oxygen atoms (O7A, O8A, O1 and O2) from another two ligand sets and one oxygen atom (O9) from deprotonated pivalic acid. It also forms a NO7 coordination environment. The Dy1–O bond lengths are in the range of 2.169(3)–2.534(3) Å and the Dy1–N bond is 2.491(3) Å. The Dy2–O bond lengths are in the range of 2.249(3)–2.635(3) Å and the Dy2–N bond is 2.443(3) Å. They are all comparable to those of reported complexes with [DyNO7] units.^2o,20 The exact coordination geometries of the octa-coordinated Ln(iii) ions were analysed by SHAPE 2.1 software²¹ and resulting data were shown in Table S2.† The resulting data from the closer analysis reveal that Ln1(iii) ion is in a square antiprism (D_4d) configuration while Ln2(iii) is in a biaugmented trigonal prism (C_2V) configuration with a minimum continuous shape measures (CShM) value.

Fig. 1. (a) Molecular structure of 3. Symmetry code: A = 2 − X, −Y, 1 − Z (hydrogen atoms and solvents have been omitted for clarity). (b) The tetranuclear [Dy₄O₆] core in 3. (c) Coordination polyhedrons of Dy(iii) ions in 3.

Scheme 2. The coordination modes of ligand (L²⁻) in complexes 1–5.

Magnetic studies

The magnetic properties of 1–5 were investigated by direct-current (DC) magnetic susceptibility studies under a magnetic field of 1000 Oe in the temperature range 2–300 K. The experimental χ_MT values at 300 K and magnetic behaviours of 1–5 are shown in Table 2. The χ_MT versus T plots for 1–5 are shown in Fig. 2. For 1, the room temperature χ_MT value of 33.41 cm³ K mol⁻¹ is slightly higher than value of 31.52 cm³ K mol⁻¹ expected for four magnetically isolated Gd^III ions (⁸S_7/2, g = 2). At 300 K, the χ_MT values of 46.91 and 58.50 cm³ K mol⁻¹ for complexes 2 and 3 are close to the expected values of 47.20 cm³ K mol⁻¹ for uncoupled Tb^III ions (⁷F₆, g = 2) and 56.68 cm³ K mol⁻¹ for uncoupled Dy^III ions (⁸H_15/2, g = 4/3). The room temperature χ_MT values for 4 and 5 are 57.47 and 44.21 cm³ K mol⁻¹, respectively, which are slightly lower than values of 58.28 cm³ K mol⁻¹ expected for four magnetically uncoupled Ho^III ions (⁵I₈, g = 5/4) and 45.92 cm³ K mol⁻¹ expected for four magnetically uncoupled Er^III ions (⁴I_15/2, g = 6/5). Upon cooling, the χ_MT product of 1 stays almost constant in the temperature range of 300–30 K and then decreases rapidly, reaching minimum value of 13.80 cm³ K mol⁻¹ at 2 K. This magnetic behaviour suggests weak antiferromagnetic exchange interactions. The application of a Curie–Weiss law gives values of θ = −1.98 K, C = 32.86 for 1 (Fig. S6†). With the decrease of temperature, the χ_MT values for 2, 4 and 5 firstly decrease slowly and then decline sharply to minimum values at 2 K. The decrease of χ_MT values for 2, 4 and 5 is typical for Ln^III ions and is due to several factors, namely, the thermal depopulation of the excited m_J sublevels of the ^2s+1Γ_J ground state of the Ln^III ion originated by a crystal field symmetry, in combination with the weak Ln^III–Ln^III antiferromagnetic interactions in 2, 4 and 5.²²

Direct-current (dc) magnetic susceptibilities of 1–5.

Complex	Metal ions	Theoretical χMT/cm3 K mol−1	Experimental χMT/cm3 K mol−1 at 300 K	Magnetic behaviour
1	Gd	31.52	33.41	AF
2	Tb	47.20	46.91	AF
3	Dy	58.68	58.50	F
4	Ho	58.28	57.47	AF
5	Er	45.92	44.21	AF

Fig. 2. Temperature dependence of the χ_MT products in the range of 2–300 K under 1000 Oe for 1–5.

Upon cooling, the χ_MT value of 3 decreases imperceptibly and reaches minimum value of 55.9 cm³ K mol⁻¹ at 23 K, which is attributed by the thermal depopulation of Stark effect for a single Dy^III ion. Continuous being cooled down to 2 K, the χ_MT product sharply increases to a maximum value of 69.6 cm³ K mol⁻¹. This magnetic phenomenon may be caused by the presence of ferromagnetic interactions between the spin carriers.^2g,23

To investigate the dynamic magnetic properties of 3, alternating-current (ac) magnetic susceptibilities were determined under zero-dc field with a 2 Oe oscillating ac field (Fig. 3). The out-of-phase susceptibility (χ′′) displays frequency-dependent phenomenon at low temperatures, which suggests the presence of slow relaxation of the magnetization, typical of SMM behaviour.²⁴ However, the peak maxima are not found. Ac susceptibility measurements of 3 under 0–10 000 Oe dc field at 1000 Hz were conducted (Fig. S7†). Unfortunately, no optimum field was found. The energy barrier (Ea/K_B) and pre-exponential factor (τ₀) values were calculated from the frequency-dependent ac susceptibility data by using the Debye model based on the equation ln(χ′′/χ′) = ln(ωτ₀) + Ea/K_BT.^24a The best fitting results give Ea/K_B ≈ 1.61 K and τ₀ ≈ 7.42 × 10⁻⁶ s (Fig. 4). Apparently, the τ₀ value is comparable to the expected values 10⁻⁶ to 10⁻¹¹ for typical SMMs.²⁵

Fig. 3. Temperature dependence of the in-phase χ′ (a) and out-of-phase χ′′ (b) for 3 under zero dc field at different frequencies in the range of 2–20 K.

Fig. 4. Plots of ln (χ′′/χ′) vs. T⁻¹ for 3 in the range 2–4.0 K. The grey solid lines correspond to the best fit result.

Luminescent properties

The solid-state luminescent properties of the H₂L ligand and 2 were measured at room temperature (Fig. S8†). The H₂L ligand displays a broad emission around 544 nm (λ_ex = 467 nm), which is due to the π → π* transition between the ligand orbitals. Complex 1 exhibits a ligand-centred broad band emission at 557 nm and a shoulder at 483 nm (λ_ex = 396 nm), which is attributed to the ligand centred π–π* and charge–transfer transition between ligands and metal centers²⁶ (Fig. S8b†). As shown in Fig. S8,† the solid-state emission spectra of 3–5 (Dy₄, Ho₄, Er₄) display emission bands centred at 482, 482, 480 nm respectively under excitation at 324 nm. The blue shifts of 3–5 are tentatively assigned to the electrostatic interaction between the ligand and metal ions.^26b Complex 2 displays four characteristic emission peaks at 490, 544, 584, and 618 nm (λ_ex = 360 nm), corresponding to ⁵D₄ → ⁷F₆, ⁵D₄ → ⁷F₅, ⁵D₄ → ⁷F₄ and ⁵D₄ → ⁷F₃,^3b respectively (Fig. S9†). The strongest emission of 2 at 544 nm is responsible for the green emission, which can be observed with the naked eyes. The π-conjugated multidentate organic ligand can act as antennae to sensitize the weak luminescent of Tb(iii). Therefore, the strong and sharp luminescence emission of 2 can be visually observed, which allows it to be considered as a fluorescent sensor. The results of the luminescence sensing experiments revealed that complex 2 can be served as a luminescent senior for selectively sensing 4-NA, Fe³⁺, CrO₄²⁻ and Cr₂O₇²⁻.

In order to examine the stability, the PXRD patterns of 2 were measured after being soaked in aqueous solutions with different pH values for several days. The PXRD patterns remained almost unaltered and agreed well with the simulated pattern, revealing the excellent stability of 2 in different solutions (Fig. S10 and S11†). Twenty samples of 2 (2 mg/2 ml suspension) after detection of 4-NA (2 mL, 1 × 10⁻³), Fe³⁺ (1 mL, 3 × 10⁻³), CrO₄²⁻ (2 mL, 2 × 10⁻³) and Cr₂O₇²⁻ (2 mL, 2 × 10⁻³) were filtered and dried to recover for PXRD test.

Detection of metal ions

Based on the fact that compound 2 possesses excellent stability in the aqueous solutions with broad pH range, its potential florescent sensing properties towards metal ions were investigated. Firstly, a series of aqueous solutions of 2 were prepared by adding 2 mg of complex 2 into 2 mL of deionized water. Then, the resulting suspensions were sonicated for 8, 10 and 30 minutes, respectively. The sizes of suspensions of particles of 2 were analysed by scanning electron microscopy. The size-distribution histograms were obtained from measuring the diameters of 100 randomly selected particles. According to the histograms of the particle size distributions in Fig. S12,† the average diameters of 2 after being sonicated for 8, 15 and 30 minutes are estimated to be approximately 15.58 ± 3.02, 12.75 ± 2.30 and 7.77 ± 2.62 μm, respectively, of which the sample (2) after being sonicated for 30 minutes displays the smallest particle size. The smaller size of the 2 leading to larger surface areas and more accessible active sites on their surface would be beneficial to realize highly sensitive and fast-response luminescent sensing.²⁷

Therefore, the resulting suspensions were sonicated for 30 minutes. Next, aqueous solutions of nitrate salts (200 μL, 10⁻² M) of Na⁺, K⁺, Mg²⁺, Ca²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺, Cd²⁺, Pb²⁺, Al³⁺, Cr³⁺ and Fe³⁺ were added in the above suspensions for fluorescence testing. As shown in Fig. 5, the metal salts slightly altered the luminescent intensity of 2, whereas Fe³⁺ ion exhibited a significant influence on the luminescence of 2. Concentration-based studies were performed by adding different amount of aqueous solutions of Fe³⁺ ion into the suspensions of complex 2 (2 mg in 2 mL deionized water). As shown in Fig. 6, the luminescent intensities of 2 progressively decreased as the concentrations of Fe³⁺ ion increased. The luminescent intensities and the concentrations of Fe³⁺ ions show good linear relationships at low concentrations. The limit of detection (LOD) value is calculated to be 1 × 10⁻⁵ M by 3δ/s, where δ is the standard deviation of fluorescent test for 10 blank measurements and s is the slope of the calibration curve (Fig. S13b†).^12i,j In addition, the quenching effect can be explained by the Stern–Volmer equation: I₀/I = K_SV[Q] + 1, where I₀ and I are the luminescent intensities before and after addition of the target analyte, respectively. K_SV is the Stern–Volmer quenching constant (M⁻¹) and [Q] is the concentration of the analyte.^12p As can be seen from Fig. S13a,† the Stern–Volmer (SV) plot for Fe³⁺ ion displays nearly linear relationships at low concentration with K_SV value of 1.86 × 10⁴ M⁻¹, which is comparable to those of reported fluorescent sensors based on Ln-complexes (Table S3†). The S–V plot deviates from linearity at high concentration (Fig. 6b), which may be attributed to both the occurrence of static and dynamic quenching. Nonlinear S–V curve of 2 can be fitted well by an exponential quenching equation, I₀/I = 2.118 exp (4.327[Fe³⁺]) + 0.199. The possible interference experiments were carried out for complex 2. The selectivity of 2 towards Fe³⁺ ion over other metal cations was verified by adding Fe³⁺ ion (10⁻² M, 0.2 mL) to the suspensions of 2 (2.4 mL) in which other competitive metal ions (10⁻² M, 0.2 mL) were introduced. Then the changes of the emission intensities were recorded. No obvious changes were observed, which suggested that fluorescence detection of 2 could not be interfered by the presence of other cations. Competitive experiments indicated that 2 has great potential as a highly selective sensor for Fe³⁺ ion (Fig. 7). To study the recycling performance of the fluorescent sensor, Fe³⁺ ion (10⁻² M, 0.2 mL) was introduced to the suspension of complex 2 (2 mL), then washed by water (after Fe³⁺ ion being detected) with five cycles of sensing experiments. As shown in Fig. 8, the fluorescent intensity of the samples did not drop significantly after five cycles. The resulting K_SV, LOD value, high selectivity, stability and recyclable property reveal that complex 2 can be considered as a fluorescent sensor for Fe³⁺ ion in aqueous system.

Fig. 5. (a) The emission spectra and (b) relative intensities at 544 nm (⁵D₄ → ⁷F₅) for 2 which was dispersed in aqueous solutions of different metal ions upon excitation at 360 nm.

Fig. 6. (a) Luminescent responses of a water suspension of 2 (2 mg mL⁻¹) towards different concentrations of Fe³⁺ ions (3 × 10⁻³ M, 0–1000 μL). (b) Nonlinear Stern–Volmer (SV) plot for Fe³⁺ in the presence of a water suspension of 2 (2 mg/2 mL).

Fig. 7. Relative luminescent intensities of the suspensions of 2 at 544 nm treated with Fe³⁺ in the presence of the other metal ions in water.

Fig. 8. Relative luminescent intensity for ⁵D₄ → ⁷F₅ transition of suspension of 2 in the absence and presence of Fe³⁺ for each cyclic test.

As reported, the possible fluorescence quenching mechanism for 2 towards Fe³⁺ ion mainly rises from three aspects: (a) the collapse of the structure;²⁸ (b) the weak interactions between Fe³⁺ ion and the methoxy group at the terminal of 2;²⁹ (c) the energy transfer between 2 and Fe³⁺ ion.³⁰ In order to understand which part was the main contributor, the PXRD, X-ray photoelectron spectroscopy (XPS) and UV-Vis absorption spectra of analytes were determined. By comparing the PXRD pattern (Fig. S15†) of 2 before and after the detection of Fe³⁺ ion, we found that no obvious changes occurred in this process, which indicated that the structure of 2 did not shatter. Depending on the result of the PXRD patterns, the fluorescence quenching mechanism of structural collapse can be ruled out. As shown in XPS spectrum of 2 (Fig. S14†), after the detection of Fe³⁺ ion, the O1s peak did not change, demonstrating the nonexistence of the weak interactions between Fe³⁺ ion and complex 2. However, as shown in Fig. S14,† the absorption band of Fe³⁺ ion from 250 nm to 400 nm remarkably overlaps the excitation band of 2 from 300 to 400 nm. Consequently, the solution of Fe³⁺ may absorb the energy of the excitation wavelength, which leads to the luminescence quenching.

Detection of anions

The aqueous solutions (200 μL, 10⁻² M) of common sodium salts Na_yX (X = F⁻, Cl⁻, Br⁻, I⁻, NO₃⁻, OAc⁻, SCN⁻, SO₄²⁻, CO₃²⁻, Cr₂O₇²⁻, CrO₄²⁻, and PO₄³⁻ were added in the suspensions of 2 (2 mg complex 2 in 2 mL water, and then being sonicated for 30 minutes) to investigate the luminescence quenching effects. An alluring finding is that CrO₄²⁻ and Cr₂O₇²⁻ ions showed remarkable turnoff quenching effect on florescent intensities of 2, whereas other tested anions displayed minor effects on the luminescent intensity (Fig. 9). In order to further investigate the sensitivity of 2 toward CrO₄²⁻ and Cr₂O₇²⁻, we implemented the titration experiments by adding aqueous solutions of CrO₄²⁻ or Cr₂O₇²⁻ (2 × 10⁻³ M) to the suspension of complex 2 (2 mg in 2 mL water). The emission titration data were collected. As expected, the fluorescent intensity of 2 gradually decreased with the increase of the concentrations of CrO₄²⁻ or Cr₂O₇²⁻ (Fig. S16†). Moreover, the S–V curve of 2 (Fig. 10) can be fitted well by an exponential quenching equation, I₀/I = 0.897 exp(2.801[CrO₄²⁻]) + 0.135 for CrO₄²⁻ and I₀/I = 2.601 exp(2.807[Cr₂O₇²⁻]) − 1.747 for Cr₂O₇²⁻. The S–V plot shows good linear relationships at low concentrations (Fig. S17a and S17b†), suggesting that both static and dynamic quenching may happen simultaneously. The quenching constants, K_SV, are calculated to be 2.998 × 10³ M⁻¹ (CrO₄²⁻) and 7.44 × 10³ M⁻¹ (Cr₂O₇²⁻) for 2 (Fig. S17a and S17b†). The LOD values calculated according to 3δ/s reach 5.2 × 10⁻⁵ M (CrO₄²⁻) and 2.7 × 10⁻⁵ M (Cr₂O₇²⁻) for 2 (Fig. S17c and S17d†). Competitive experiments were carried out to study the interference of other ions on the fluorescent sensitivity for 2 towards CrO₄²⁻ or Cr₂O₇²⁻ ions. The fluorescent intensity decreased significantly after adding CrO₄²⁻ or Cr₂O₇²⁻ (0.2 mL, 10⁻² M) into suspensions of 2 (2 mg in 2.2 mL water), in which other anions (0.2 mL, 10⁻² M) were also added (Fig. 11). Moreover, recyclability experiments on the detection of CrO₄²⁻ or Cr₂O₇²⁻ were performed. After five cycles of quenching experiments and regenerations, the quenching efficiency experiences only a minor decrease (Fig. 12). These results showed that complex 2 could be a potential candidate for sensitive and selective detections of CrO₄²⁻ or Cr₂O₇²⁻ ions. A comparison of florescent sensing properties based on 2, Ln-MOF, and Ln-CP towards Fe³⁺, CrO₄²⁻ or Cr₂O₇²⁻ ions is listed in Table S3.†^{6b,12f–o,13a}

Fig. 9. (a) The emission spectra and (b) relative intensities at 544 nm (⁵D₄ → ⁷F₅) for 2 dispersed in different anion aqueous solutions upon excitation at 360 nm.

Fig. 10. Nonlinear Stern–Volmer (SV) plot for CrO₄²⁻/Cr₂O₇²⁻ in the presence of a water suspension of 2 (2 mg/2 mL).

Fig. 11. (a) Relative luminescent intensities of the suspension of 2 at 544 nm treated with CrO₄²⁻ in the presence of the other anions in water. (b) Relative luminescent intensities of the suspension of 2 at 544 nm treated with Cr₂O₇²⁻ in the presence of the other anions in water.

Fig. 12. Relative luminescence intensity for ⁵D₄ → ⁷F₅ transition of a suspension of 2 in the absence and presence of CrO₄²⁻ anions (a) and Cr₂O₇²⁻ anions (b) for each cyclic test.

After CrO₄²⁻ and Cr₂O₇²⁻ ions were detected, the PXRD patterns of 2 were measured. The obtained PXRD patterns still well correspond with the original data of 2, which demonstrate that the crystal structure of 2 is retained (Fig. S20†). On the other hand, the UV-Vis spectra of CrO₄²⁻ and Cr₂O₇²⁻ ions in aqueous solution were determined. As shown in Fig. 13, the absorption bands of CrO₄²⁻ in the range of 230–450 nm and Cr₂O₇²⁻ in the range of 275–475 nm display extensive overlap with the excitation band of 2. Therefore, luminescence quenching can be attributed to reducing the efficiency of the energy transfer from the ligand to the lanthanide ions, as the energy of the excitation light will be strongly absorbed by CrO₄²⁻ or Cr₂O₇²⁻.^6b,13e

Fig. 13. Spectral overlap between normalized absorption spectra of solutions of Fe³⁺, CrO₄²⁻, Cr₂O₇²⁻ ions and excitation spectrum of 2 in water.

Detection of nitrobenzene derivatives

Complex 2 was dispersed in ethanol and then stirred for three days. It was filtered and dried. The PXRD of 2 (Fig. S20†) after being stirred for three days matched with that of the simulated data, which indicated that complex 2 was stable in ethanol solution. The stable nature of 2 in ethanol solution affords a prerequisite for the detection of nitroaromatics (NACs). Nitrobenzene (NB), 4-nitrophenol (4-NP), 4-nitroaniline (4-NA), 4-nitrotoluene (4-NT), 4-nitrochlorobenzene (4-Cl-NB), 2,4-dinitrotoluene (2,4-DNT) and 2,4-dinitrophenol (2,4-DNP) were selected to explore the potential of 2 as a chemosensor in ethanol solution. Firstly, complex 2 (2 mg) was homogeneously dispersed in a series of ethanol solution (2 mL) and sonicated for 30 minutes to form a uniform suspension. Then, different NACs (10 μL, 10⁻¹ M) was added to the above-mentioned suspensions, respectively. Next, the fluorescent intensities of the resulting mixture were measured. As shown in Fig. 14, the order of quenching extend is as follows: 4-NA > 2,4-DNP > 4-NP > 4-NT > NB > 4-Cl-NB > 2,4-DNT. Concentration-dependent titration experiments of 4-NA were performed to test the influence of concentrations on fluorescent intensity. As the concentration of 4-NA increases, the fluorescent intensity gradually decreased (Fig. S18†). The liner fitting of the S–V plot for 4-NA at low concentrations (Fig. S19a†) gives K_SV value of 1.14 × 10⁴ M⁻¹ (4-NA). Nonlinear concentration curve (Fig. 15) can be fitted well by an exponential quenching equation: I₀/I = 3.776 exp(5.380[Q]) − 3.131 (Q = 4-NA). The detection limit value (Fig. S19b†) towards 4-NA was calculated by 3δ/s, generating 1.17 ppm (8.5 × 10⁻⁶ M) for 2. Recyclability experiments on the detection of 4-NA were performed. After five cycles of quenching experiments and regenerations, the quenching efficiency experiences only a minor decrease (Fig. 16). To our knowledge, few fluorescent complexes acted as sensors for 4-NA in ethanol system have been reported in the literature (Table S4†).^12q–s,13b Compared to luminescent sensors based on transition metal complexes, lanthanide complex 2 has the advantage of characteristic emissions and bright luminescent colour of Tb(iii) ion. No lanthanide clusters as sensors for 4-NA have been reported.

Fig. 14. (a) The emission spectra and (b) luminescence quenching at 544 nm (⁵D₄ → ⁷F₅) for the suspensions of 2 in the presence of seven different ethanol solutions of NACs (10 μL, 10⁻¹ M) upon excitation at 360 nm.

Fig. 15. Nonlinear Stern–Volmer (SV) plot for 4-NA in the presence of an ethanol suspension of 2 (2 mg/2 mL).

Fig. 16. Relative luminescence intensity for ⁵D₄ → ⁷F₅ transition of ethanol suspension of 2 in the absence and presence of 4-NA for each cyclic test.

As reported, the quenching mechanism of NACs sensors is ascribed to two factors: photo-induced electron transfer (PET) and resonance energy transfer (RET) or their cooperative effect. According to the molecular orbital theory, nitroaromatic compounds are good electron acceptors, due to the substitution of the electron-withdrawing nitro groups on the aromatic ring, which can stabilize the lowest unoccupied molecular orbital (LUMO) of the system via conjugation effect, ultimately leading to the complete or part quenching of luminescence-based sensor.^13b,31 The electron transfer originated from the phenyl rings of the ligands of 2 to excellent electron donor nitrobenzene leads to luminescence quenching upon excitation. On the other hand, the UV-Vis results reveal that there are obvious overlaps between the absorption spectrum of NACs especially 4-NA and the emission spectrum of 2 (Fig. 17). This would permit energy transfer from 2 to NACs, thus further improving the luminescence quenching efficiency towards NACs.^30b

Fig. 17. Spectral overlap between normalized absorption spectra of seven selected NACs and excitation spectra of 2 in ethanol solution.

Conclusions

In summary, five tetranuclear lanthanide clusters have been synthesized by using Ln(NO₃)₃·6H₂O (Ln = Gd, Tb, Dy, Ho, Er), pivalic acid and Schiff base ligand 2-(((2-hydroxy-3-methoxybenzyl)imino)methyl)-6-methoxyphenol (H₂L) as starting materials under solvothermal method. Complexes 1–5 possess [Ln₄O₆] cores formed by the fusion of two phenoxide oxygen bridged two [Ln₂O₂] moieties. The magnetic properties of 1–5 were investigated. Antiferromagnetic interactions were determined for 1, 2, 4 and 5. Complex 3 displayed typical single molecule magnet behaviour. Considering its low detection limits and high quenching constant K_sv, complex 2 may potentially be utilized as luminescence sensors for quantitative detection of 4-NA, Fe³⁺ and CrO₄²⁻/Cr₂O₇²⁻ ions. Although the design of chemosensors is more based on academic or research interest, the exceptional stability of Tb₄ clusters makes it applicable to extreme industrial application.

Conflicts of interest

There are no conflicts to declare.