Redox-Active Metal-Organic Frameworks with Three-Dimensional Lattice Containing the m-Tetrathiafulvalene-Tetrabenzoate

Metal-organic frameworks (MOFs) constructed by tetrathiafulvalene-tetrabenzoate (H4TTFTB) have been widely studied in porous materials, while the studies of other TTFTB derivatives are rare. Herein, the meta derivative of the frequently used p-H4TTFTB ligand, m-H4TTFTB, and lanthanide (Ln) metal ions (Tb3+, Er3+, and Gd3+) were assembled into three novel MOFs. Compared with the reported porous Ln-TTFTB, the resulted three-dimensional frameworks, Ln-m-TTFTB ([Ln2(m-TTFTB)(m-H2TTFTB)0.5(HCOO)(DMF)]·2DMF·3H2O), possess a more dense stacking which leads to scarce porosity. The solid-state cyclic voltammetry studies revealed that these MOFs show similar redox activity with two reversible one-electron processes at 0.21 and 0.48 V (vs. Fc/Fc+). The results of magnetic properties suggested Dy-m-TTFTB and Er-m-TTFTB exhibit slow relaxation of the magnetization. Porosity was not found in these materials, which is probably due to the meta-configuration of the m-TTFTB ligand that seems to hinder the formation of pores. However, the m-TTFTB ligand has shown to be promising to construct redox-active or electrically conductive MOFs in future work.

Comparatively speaking, the meta-derivative of the frequently used p-TTFTB ligand is rarely reported. It can be predicted that H 4 TTFTB will show a different assembly behavior in the construction of MOFs. In 2020, Zuo et al. firstly reported a 2D MOF, In-m-TTFTB, which possesses a proton conductivity of 6.66 × 10 −4 S cm −1 at 303 K and 98% relative humidity (RH) [31]. Recently, a series of persistent radical 2D MOFs were assembled by a hexanuclear rare-earth-cluster-based 1D chains and a (m-TTFTB) 3 trimer building block [32]. These MOFs exhibit highly chemical and thermal stability. Due to efficient light absorption, intermolecular charge transfer, low thermal conductivity, and outstanding stability, Dy 6m-TTFTB-MOF shows excellent photothermal properties, an increase of 34.7 • C within 240 s under one-sun illumination [32]. Another kind of m-TTFTB-MOF was obtained by adjusting the synthetic conditions. This MOF possesses a low BET surface area of 129 m 2· g −1 with a high near-infrared (NIR) photothermal conversion [33]. Further study revealed that the photothermal conversion of this MOF could be enhanced by redox doping and plasmon resonance. Even though the photothermal conversion of m-TTFTB-based MOFs is carefully studied [31,32], the semiconductive and magnetic properties of these MOFs have not been reached.
The absence of any other phases from Tb-m-TTFTB, Er-m-TTFTB, and Gd-m-TTFTB was confirmed by powder X-ray diffraction (PXRD) measurements in which the diffraction peak positions were similar to the calculated from single-crystal X-ray data of Tb-m-TTFTB ( Figure 2). The advantages in stability originated from the interaction of TTF matrix and the protecting of the tight organic parts surrounding the rare earth centers ( Figure 1c) [33]. In this case, reducing of coordination space is an efficient strategy to enhance the stability of the frameworks. The thermal stability was investigated. The TGA data for Ln-m-TTFTB showed that the frameworks have good thermal stability up to 450 • C ( Figure S5) in the N 2 atmosphere, which is comparable to the Ln-TTFTB series (nearly 500 • C) [28,34].

Cyclic Voltammetry
Solid-state direct current (DC) cyclic voltammetry (CV) studies on Ln-m-TTFTB were conducted in 0.1 M LiBF 4 in CH 3 CN (Figures 3a, S6 and S7). Upon scanning anodically, two reversible one-electron processes at around 0.21 and 0.48 V (vs. Fc/Fc + ) were observed for all three MOFs. These processes are attributed to the TTF/TTF •+ and TTF •+ /TTF 2+ redox couples, respectively (inset of Figure 3a). In contrast to the CV of m-H 4 TTFTB (0.13 and 0.35 V (vs. Fc/Fc + )) [31], the two one-electron processes observed for Tb-m-TTFTB were shifted by ca. 0.1 V, which is attributed to both the coordination to terbium ions and the deprotonated nature of the ligand [35]. The current associated with the two redox processes were almost the same over multiple scans which is consistent with their reversible nature. In addition, faster sweep rates led to broader features because of slow diffusion kinetics through the framework (Figures 3b, S6 and S7). The locations of these redox couples in Ln-m-TTFTB were similar to those of Ln-TTFTB, meaning that these two series of MOFs possess similar redox activity [34].

Absorption Spectra and Semiconducting Properties
Prior to studying the conductivities of these materials, the solid-state absorption spectra were obtained to gain insight into the influence of coordination on the optical and conducting properties. In Ln-m-TTFTB, there are three main absorption bands located in the region 250-550 nm (Figure 4a). These higher energy absorption bands are attributed to the n→π* or π→π* transition of the free ligand, which is similar to the absorption of the ligand m-H 4 TTFTB [33]. The absorption band around 760 nm can be assigned to the neutral (TTF) 2 in the framework [36,37]. The peak located at 690 nm of the free ligand can be attributed to a small number of auto-oxidized TTF •+ . Using these UV-vis-NIR adsorption data, we approximated the band gaps of the ligand and Ln-m-TTFTB through the Kubelka-Munk function. From the Tauc   To better understand the conducting behaviors of these MOFs, conductivity studies were undertaken on the single crystal samples of Ln-m-TTFTB by the two-contact probe method. The room temperature electrical conductivities in the long horizontal direction of the single crystal ( Figure S8) were 5.4 × 10 −7 , 9.6 × 10 −7 , and 1.0 × 10 −7 S·cm −1 for Tb-m-TTFTB, Er-m-TTFTB, and Gd-m-TTFTB, respectively (Figures S9-S11 and Table S2). These values are nearly ten times of the reported Ln-TTFTB series (10 −8 S·cm −1 in powder state), and this can be attributed to the lower contact resistance in single crystal (also~10 −8 S·cm −1 in powder state). The lower conducting performances of Ln-m-TTFTB originate from the lack of band formation or the poor electron transfer. Notably, the powder electrical conductivity is reported as 2.74 × 10 −8 S·cm −1 for the free m-H 4 TTFTB [31]. The similar electrical conductivity between the free ligand and the assembled MOFs are predicted by the similar band gaps.  [34]. As the temperature was lowered, χ M T decreased in each case as expected from antiferromagnetic interaction and/or the depopulation of excited states [39,40]. Focusing on the crystal structures, we can define three exchange interactions ( Figure S12) within one chain due to the three different kinds of lanthanide dimer. Given that the Gd 3+ ion is magnetically isotropic, the χ M T of Gd-m-TTFTB ( Figure S13) decreased in the high-temperature region due to the antiferromagnetic exchange interaction with J = −1 cm −1 (Ĥ EX = −2J 1 S Gd1 S Gd2 ) in dimers. In the low-temperature region, the slight upturn might be attributed to the ferromagnetic interaction and/or dipolar-dipolar interaction between dimers, or ferromagnetic intra-chain interaction. For Er-m-TTFTB, an upturn of the moment below ca. 6 K indicates one possible weak nearest neighbor ferromagnetic interaction and/or dipolar-dipolar interaction [41]. At 1.8 K, the isothermal magnetizations (M) versus field (H) reached 10.84, 11.71, and 11.08 Nβ in 70 kOe for Tb-m-TTFTB, Dy-m-TTFTB, and Er-m-TTFTB, respectively (Figures S14-S16). The low saturation of magnetization values maybe ascribed to the effects of low-lying excited states and/or magnetic anisotropy [42]. The temperature dependence of the alternating current (ac) susceptibilities (2 Oe) at different fixed frequencies (1.0-999 Hz) were measured for all MOFs. Under zero direct current (dc) field, Dy-m-TTFTB exhibited both in-phase (χ ) and out-of-phase (χ ) ac-susceptibilities (Figure 6a), but no peaks were observed. The ac-susceptibilities of Erm-TTFTB and Tb-m-TTFTB exhibited no χ in zero dc field (Figures S17a and S18a). The frequency-dependent out-of-phase (χ ) ac-susceptibility below 6 K for Dy-m-TTFTB revealed the slow relaxation of the magnetization. However, no peak of χ was observed even at 999 Hz, likely due to a lower anisotropic energy barrier. The χ susceptibility increased with the increase of the frequency, suggesting that the peak maxima are to be found at lower temperatures or higher frequencies of the SQUID instrument. Even an increase in the dc field to 1000 Oe showed no peak above 2 K in the ac-susceptibilities of Dy-m-TTFTB and Er-m-TTFTB, while Tb-m-TTFTB exhibited no χ signals (Figures 6b, S17b and S18b). With low energy barriers, no peaks above 1.8 K in the out-of-phase of ac susceptibility observed in the frequency region of 1.0-999 Hz were reasonable.

Materials and Methods
All the reagents and solvents were commercially available and used as received. FT-IR spectra were recorded on a Vector 27 Bruker Spectrophotometer by transmission through KBr pellets containing the ground crystals in the range 4000-400 cm −1 . The powder X-ray diffraction patterns (PXRD) were collected at room temperature using a scan speed of 0.1 s/step on a Bruker Advance D8 diffractometer (40 kV, 40 mA) (Bruker, Karlsruhe, Germany) equipped with Cu radiation. Calculated PXRD patterns were generated using Mercury 3.0 [43]. Elemental analyses (EA) for C, H, and N were performed on a Perkin-Elmer 240C analyzer (PerkinElmer, Waltham, MA, USA). TGA data were obtained on a STA 449C thermal analysis system at a heating rate of 10 • C min −1 under N 2 atmosphere. Magnetization measurements were performed using a Quantum Design SQUID VSM magnetometer (Quantum Design, Darmstadt, Germany) on polycrystalline samples for all compounds.

X-ray Structure
Single-crystal X-ray diffraction intensity data for Tb-m-TTFTB, Er-m-TTFTB, and Gdm-TTFTB were collected on a Bruker D8 Venture diffractometer fitted with a PHOTON-100 CMOS detector, monochromatized microfocus Mo K α radiation (λ = 0.71073 Å), and a nitrogen flow controlled by a KRYOFLEX II low-temperature attachment operating at 153 K. Raw data collection and reduction were controlled using APEX3 software (version 2016.9-0; Bruker, 2016) [44]. Absorption corrections were applied using the SADABS routine. The structures were solved by direct methods and refined by full-matrix leastsquares on F 2 using the SHELXTL software package (version-2018/3) [45]. Non-hydrogen atoms were refined with anisotropic displacement parameters during the final cycles. Hydrogen atoms of m-TTFTB, formate, and dimethylformamide (DMF) molecules were placed at calculated ideal positions and isotropic displacement parameters were used. Except for the coordinated DMF molecule, those free solvent molecules of DMF or water were highly disordered and were unable to be located and refined. The diffuse electron densities resulting from these residual molecules were removed from the data set using the SQUEEZE routine of PLATON and refined further using the data generated [46].

Synthesis of m-H 4 TTFTB
m-H 4 TTFTB ( Figure S19) was prepared according to the reported method [31]. The method is briefly described in the supporting information.

Solid CV
Solid-state cyclic voltammetry measurements were performed in LiBF 4 /CH 3 CN as electrolyte using a Corrtest 4-channel electrochemical workstation and a three-electrode system. The CVs were recorded using a glassy carbon working electrode (3.0 mm diameter), a platinum wire auxiliary electrode, and an Ag wire quasi-reference electrode with the solutions of 0.1 M LiBF 4 dissolved in distilled CH 3 CN. The sample was mounted on the glassy carbon working electrode by dipping the electrode into a paste made of the powder sample in ethanol. Ferrocene was added as an internal standard upon completion of each experiment. All potentials are reported in milli-Volts (mV) versus the Fc/Fc + couple.

Solid-State Diffuse Reflectance Spectra
The UV-Vis-NIR data were obtained using a Harrick Praying Mantis attachment over the range 200-900 nm. Spectra are reported as the Kubelka-Munk transform.

Electrical Conductivity
The electrical conductivities of the needle-like single crystal samples using the twoprobe method were obtained using a Keithley 2400 source meter (Keithley 2400, Tektronix, Beaverton, OR, USA) on CRX-4K Closed Cycle Refrigerator-based Probe Station at room temperature. The single crystal samples were lined in the vertical direction, which were connected by conductive carbon adhesive. All of the current-voltage (I-V) measurements were performed under ambient conditions by sweeping the voltage from −1.5 V to 1.5 V.

Conclusions
In summary, three redox-active MOFs were constructed by lanthanide metal ions (Tb 3+ , Er 3+ , and Gd 3+ ) and m-TTFTB. These MOFs showed similar three-dimensional lattice with a dense stacking. It can be concluded that compared with H 4 TTFTB, the assembly of m-H 4 TTFTB tends to form a structure with almost no porosity. Magnetic study revealed that Dy-m-TTFTB and Er-m-TTFTB possess slow relaxation of the magnetization. In all, even though m-TTFTB-based MOFs seem not to be a good porous material, these dense stacking structures may enable them as good candidates for the study of charge transfer, electrical conductivity, and magnetic properties. Further studies focusing on the functional assembly of m-H 4 TTFTB and other metal building blocks are currently in progress in our group.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules27134052/s1, Table S1: Selected bond lengths (Å) and angles ( • ) of Tb-m-TTFTB, Er-m-TTFTB, and Gd-m-TTFTB; Table S2: The shape parameters of Tb-m-TTFTB, Er-m-TTFTB, and Gd-m-TTFTB single crystals used for the calculating of electrical conductivity; Figure S1: The crystal photos of Tb-m-TTFTB (a), Er-m-TTFTB (b), and Gd-m-TTFTB (c); Figure S2: The asymmetric units of Tb-m-TTFTB. Displacement ellipsoids are drawn at the 50% probability level; Figure S3: The coordination environment of L1 (top) and L2 (down) in Tb-m-TTFTB; Figure S4. The three-dimensional structures of Tb-m-TTFTB along a, b, and c directions. Figure S5: The TGA plots of Tb-m-TTFTB, Er-m-TTFTB, and Gd-m-TTFTB under an N 2 atmosphere; Figure S6: Solid state cyclic voltammograms of Er-m-TTFTB performed over three consecutive cycles (a) and at different scan rates (b). The experiments were conducted in 0.1 M LiBF 4 in CH 3 CN electrolyte; Figure S7: Solid-state cyclic voltammograms of Gd-m-TTFTB were performed over three consecutive cycles (a) and at different scan rates (b). The experiments were conducted in 0.1 M LiBF 4 in CH 3 CN electrolyte; Figure S8: The picture of the single crystal of Tb-m-TTFTB and the electric device used for the measurement of electrical conductivity; Figure S9: I-V curve of Tb-m-TTFTB; Figure S10: I-V curve of Er-m-TTFTB; Figure S11: I-V curve of Gd-m-TTFTB; Figure S12: Three kinds of exchange interactions in one-dimensional chain of Gd-m-TTFTB; Figure S13: Temperature dependence of the χ M T for Gd-m-TTFTB measured in a 1000 Oe field. The red line is the simulation of two isolated Gd ions. The blue line is the simulation of Gd2 cluster only existing magnetic coupling; Figure S14: The field-dependent magnetizations from 0 to 70 kOe at 1.8 K for Tb-m-TTFTB; Figure S15: The fielddependent magnetizations from 0 to 70 kOe at 1.8 K for Dy-m-TTFTB; Figure S16: The field-dependent magnetizations from 0 to 70 kOe at 1.8 K for Er-m-TTFTB; Figure S17: (a) Temperature-dependent in-phase χ and out-of-phase χ" ac susceptibility signals for Tb-m-TTFTB at the indicated frequencies under zero dc field. (b) Temperature-dependent in-phase χ and out-of-phase χ ac susceptibility signals for Tb-m-TTFTB at the indicated frequencies under 1000 dc field; Figure S18: (a) Temperaturedependent in-phase χ and out-of-phase χ ac susceptibility signals for Er-m-TTFTB at the indicated frequencies under zero dc field. (b) Temperature-dependent in-phase χ and out-of-phase χ ac susceptibility signals for Er-m-TTFTB at the indicated frequencies under 1000 dc field; Figure S19.  Data Availability Statement: CCDC 1914384, 1914385, and 1914387 contains the supplementary crystallographic data for this paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre www.ccdc.cam.ac.uk/structures (accessed on 8 February 2020) using the accession identifiers CCDC-1,914,384, CCDC-1,914,385 and CCDC-1,914,387, respectively. All other data can be obtained from the authors on request.