Four Isostructural 3d-4f Mixed Metal Organic Frameworks and Their Magnetic Properties

Four new isostructural 3d-4f mixed metal organic frameworks (MOFs) Cu Gd2 (BDC)4, compound 1, Cu Ho2 (BDC)4, compound 2, Cu Eu2 (BDC)4, compound 3, and Cu Dy2 (BDC)4, compound 4 were successfully synthesized. The structure, stability, and magnetic properties were analyzed. Each MOF has two lanthanide ions and one copper ion node with terephthalic acid as the organic linker. The lanthanide ions form a dimer with each having a capped trigonal prismatic geometry while the copper ion has a square planar geometry. Each of these MOFs shows varying degrees of antiferromagnetic interactions.


Introduction
Research in the area of metal organic frameworks, or MOFs has continued to expand due to many different applications including gas storage [1], gas separation [2], catalysis [3], and sensors [4]. There is also growing interest in ionic conductivity [5] and the magnetic properties of MOFs [6]. MOF research has largely been on single transition metal systems with fewer examples of mixed metal systems, especially with transition metals and lanthanides [7]. Lanthanides are known for their unique luminescent [8] and magnetic properties [9]. When lanthanides are combined with transition metals, the weak magnetic coupling found in lanthanides [10] can strengthen into bulk magnetics [11], create strong magnetocaloric effects [12], and result in single molecular magnetics [13]. Additionally, the high coordination numbers of lanthanides combined with the lack of stereochemical preference in comparison to transition metals increases the number of topographies possible [14]. Unfortunately, the differences in the reactivity of transition metals and lanthanides makes the synthesis of such mixed metal MOFs difficult. The nature of their reactivities often leads to homometallic coordination polymers of both metals [15,16].
The lanthanide-based materials that contain Gd 3+ , Eu 3+ , and Tb 3+ often exhibit strong paramagnetic properties which have a variety of applications [17]. Lanthanides such as Gd 3+ have large spin only magnetic moments which have been used as MRI contrasting agents [18]. Throughout the years, microporous frameworks such as zeolites [19] and MOFs have been used as drug delivery systems as well as MRI contrast agents [20,21]. Paramagnetic guest molecules like oxygen can be encapsulated in MOFs resulting in magnetic coupling [22].
The present study involves the synthesis of four new isostructural bimetallic MOFs. They contain either two gadolinium, holmium, europium, or dysprosium metal ions and one copper center using terephthalic acid as the organic linker. Terephthalic acid is a common organic linker for MOFs due to its cheap, rigid and multidentate nature [23]. The thermal stability, crystalline structure, and magnetic properties of these four new MOFs are discussed below.
The powder X-ray diffraction was performed with a Rigaku Ultima IV X-ray diffractometer (Tokyo, Japan), using a Cu Kα radiation at room temperature by scanning over a range of 2θ = 5-40 • with a scan rate of 1 • /min. The simulated powder XRD patterns were generated from the CIF files using Mercury. The thermogravimetric (TGA) analysis was collected on a SDT Q600 (TA Instruments, New Castle, DE, USA) from room temperature to 1073 K at a heating rate of 10 K from 383 K till 1073 K. The magnetic data for 1, 2, and 4 was collected on a Quantum Design Physical Properties Measurement System from temperatures of 2 K to 300 K with a magnetic field of 1000 Oe. The magnetic data for 3 was collected on a Quantum Design Magnetic Properties Measurement System from temperatures of 2 K to 300 K with a magnetic field of 1000 Oe. replaced with dysprosium (III) nitrate pentahydrate (39.5 mg, 9.01 × 10 −5 mol). This resulted in an 84.5% yield.

Structure Analysis of 3
For the Eu analogue, compound 3, there is also a seven-coordinate rare earth local  Tables 1  and S1. The lanthanides within the structures formed dimers, which is very common in MOFs with lanthanides formed using a one-pot synthesis [26]. These dimers form through a process known as ligand-controlled hydrolysis. When hydrated lanthanide salts are reacted in the presence of a base and an appropriate ligand, e.g., containing carboxylates, the formation of lanthanide clusters is driven [27]. While this synthesis is performed with DMF and not water, the lanthanide salts are hydrated, possibly providing enough water to Crystals 2021, 11, 1547 5 of 11 create the hydroxide precursor to the dimers [28]. The hydroxides are then interchanged with the terephthalic acid and create the dimers shown in Figure 1a.
Unlike most 3d-4f MOFs formed using the reactivities of different functional groups, terephthalic acid only has carboxylic acid functional groups which bind to both the copper ion and the lanthanide ions. The terephthalic acid bridges each lanthanide dimer square planar copper center in an alternating fashion creating a 1-D chain as shown in Figure 2. A similar MOF was discovered by Luo et al. with gadolinium dimers separated by copper ions with a distorted square planar geometry partly contributing to the rod-shaped substructures [29].
The lanthanides within the structures formed dimers, which is very common in MOFs with lanthanides formed using a one-pot synthesis [26]. These dimers form through a process known as ligand-controlled hydrolysis. When hydrated lanthanide salts are reacted in the presence of a base and an appropriate ligand, e.g., containing carboxylates, the formation of lanthanide clusters is driven [27]. While this synthesis is performed with DMF and not water, the lanthanide salts are hydrated, possibly providing enough water to create the hydroxide precursor to the dimers [28]. The hydroxides are then interchanged with the terephthalic acid and create the dimers shown in Figure 1a.
Unlike most 3d-4f MOFs formed using the reactivities of different functional groups, terephthalic acid only has carboxylic acid functional groups which bind to both the copper ion and the lanthanide ions. The terephthalic acid bridges each lanthanide dimer square planar copper center in an alternating fashion creating a 1-D chain as shown in Figure 2. A similar MOF was discovered by Luo et al. with gadolinium dimers separated by copper ions with a distorted square planar geometry partly contributing to the rod-shaped substructures [29]. In the extended view of compound 1 along the a-axis, as shown in Figure 3, the potential pores of the structure are shown. Figure 3a demonstrates how the free DMF and coordinated DMF face towards the pores of the 3-D structure. Figure 3b shows the presence of potential pores within the structure with the absence of the coordinated and free DMF. The extended structure of the MOF and the 1-D pores are highlighted in Figure 3c. In the extended view of compound 1 along the a-axis, as shown in Figure 3, the potential pores of the structure are shown. Figure 3a demonstrates how the free DMF and coordinated DMF face towards the pores of the 3-D structure. Figure 3b shows the presence of potential pores within the structure with the absence of the coordinated and free DMF. The extended structure of the MOF and the 1-D pores are highlighted in Figure 3c. Crystals 2021, 11, x FOR PEER REVIEW 6 of 11

Powder X-ray Diffraction (PXRD)
3.3.1. PXRD Comparison of 1, 2, 3, and 4 Figure 4 shows the comparison of the powder XRD patterns for the experimental data and simulated from the CIF file. The lack of extra peaks in comparison to the simulated pattern suggests that the crystalline phases (1, 2, 3, and 4) are reasonably pure. There are no additional crystalline phases apparent on the PXRD.

PXRD Comparison of the Solvent Removal and Replacement in 1, 2, 3, and 4
The X-ray diffraction pattern shown in Figure 5 shows a change in crystal structure after compound 2 was heated to 423 K in a vacuum oven in order to remove the DMF. There is a small color change in the sample, going from a sky blue to a pale blue, with an associated structural change. The three peaks at 8.88, 9.48, and 10.04 degrees are reduced to two peaks at 9.48 and 10.32. However, as seen in Figure 5, the original structure can be recovered with the addition of DMF. The third peak is regained, although the three peaks are not as distinct. This suggests that the structure does not collapse completely upon removal of the coordinated DMF and can be regained. Compounds 1, 3, and 4 experience the same structural change with the loss of coordinated solvent. Compound 4 in Figure S4, shows the same recovery of the structure with the addition of solvent. While there is a loss of crystallinity as the peaks broaden when activated for 4, the three peaks at 8.96, 9.60, and 10.20 are not lost. However, Figures S2 and S3 show that 1 and 3 do not have the same recovery in structure once solvent is added to the material again. Both activated structures show a change in phase losing one of the three peaks around 10 • . Interestingly, enough, the MOFs containing the heavier lanthanides, Ho 3+ and Dy 3+ , were able to recover the original phase while the lighter lanthanides were not. This might be due to the heavier lanthanides having smaller ionic radii and greater ionic potential allowing for more stability and interaction with DMF.

PXRD Comparison of the Solvent Removal and Replacement in 1, 2, 3, and 4
The X-ray diffraction pattern shown in Figure 5 shows a change in crystal structure after compound 2 was heated to 423 K in a vacuum oven in order to remove the DMF. There is a small color change in the sample, going from a sky blue to a pale blue, with an associated structural change. The three peaks at 8.88, 9.48, and 10.04 degrees are reduced to two peaks at 9.48 and 10.32. However, as seen in Figure 5, the original structure can be recovered with the addition of DMF. The third peak is regained, although the three peaks are not as distinct. This suggests that the structure does not collapse completely upon removal of the coordinated DMF and can be regained . Compounds 1, 3, and 4 experience the same structural change with the loss of coordinated solvent. Compound 4 in Figure  S4, shows the same recovery of the structure with the addition of solvent. While there is a loss of crystallinity as the peaks broaden when activated for 4, the three peaks at 8.96, 9.60, and 10.20 are not lost. However, Figures S2 and S3 show that 1 and 3 do not have the same recovery in structure once solvent is added to the material again. Both activated structures show a change in phase losing one of the three peaks around 10°. Interestingly, enough, the MOFs containing the heavier lanthanides, Ho 3+ and Dy 3+ , were able to recover the original phase while the lighter lanthanides were not. This might be due to the heavier lanthanides having smaller ionic radii and greater ionic potential allowing for more stability and interaction with DMF.

Thermogravimetric Analysis (TGA)
The TGA for 2 was taken in air from room temperature to 1073 K. There is a steep initial loss of 5% of the weight at 389 K, as seen in Figure 6. This indicates the loss of free DMF that is in the pores, as calculated from the molar mass. The amount of free DMF in the structure, one DMF molecule per unit cell, is 5% of the weight. After this, there is a slow loss of 20% of the weight from temperatures 389 to 588 K. Most of the weight lost is the DMF that is bound to the Gd. At this point, there is a steep drop in weight signifying the breaking of the Ho-O bonds and Cu-O bonds. As shown in Figure S5, compounds 1,  3, and 4, which have the same structure, exhibit similar thermal behavior. While 4 does not show the same steep loss of weight, it shows a more gradual loss of weight at the same temperatures. The weight loss for 4 is about 5% more than the other three structures. This is possibly due to the presence of amorphous materials. The sample shows the combination diamond-shaped single crystals and highly twinned clusters that could contain an amorphous phase. This additional material could also account for the more gradual loss of weight shown for 4 in Figure S5.

Thermogravimetric Analysis (TGA)
The TGA for 2 was taken in air from room temperature to 1073 K. There is a steep initial loss of 5% of the weight at 389 K, as seen in Figure 6. This indicates the loss of free DMF that is in the pores, as calculated from the molar mass. The amount of free DMF in the structure, one DMF molecule per unit cell, is 5% of the weight. After this, there is a slow loss of 20% of the weight from temperatures 389 to 588 K. Most of the weight lost is the DMF that is bound to the Gd. At this point, there is a steep drop in weight signifying the breaking of the Ho-O bonds and Cu-O bonds. As shown in Figure S5, compounds 1, 3, and 4, which have the same structure, exhibit similar thermal behavior. While 4 does not show the same steep loss of weight, it shows a more gradual loss of weight at the same temperatures. The weight loss for 4 is about 5% more than the other three structures. This is possibly due to the presence of amorphous materials. The sample shows the combination diamond-shaped single crystals and highly twinned clusters that could contain an amorphous phase. This additional material could also account for the more gradual loss of weight shown for 4 in Figure S5.

Luminescence
Europium (III) is known for being fluorescent. There are several bands of luminescence that result for 5 D0 to 7 FJ excitation states. However, these are not present in 3. This is due to how copper quenches this fluorescence when present in the same complex. This is seen in Sakamoto et al., who compared fluorescence spectra for Eu(NO3)3 6H2O to a nickel-europium complex and a copper-europium complex. The copper-europium complex showed complete quenching of the fluorescence of the europium by the copper. This is due to coupling of the two metal centers allowing the copper to quench the europium [30]. There is also an overlap with the absorbance band of the d-d band of Cu(II) ions and emission of the Eu(III) ions [31].

Magnetism
The direct current (dc) magnetic susceptibilities of compounds 1, 2, 3, and 4 were studied under an applied magnetic field of 1000 Oe and a temperature range of 2 K to 300 K, as seen in Figure 7. Compound 1 has a magnetic susceptibility of 0.0308 cm 3 /mol at 300 K. There is a gradual increase in magnetic moment as the temperature decreases. Once the temperature reaches 50 K, the magnetic moment begins to sharply increase, characteristic of paramagnetic behavior. Compound 2 has a magnetic susceptibility of 0.0181 cm 3 /mol at 300 K. It has a gradual increase until 50 K where there is a sharp increase. Framework 4 has a magnetic susceptibility of 0.0326 cm 3 /mol at 300 K. It also has a gradual increase in susceptibility until around 50 K, where it begins to sharply increase. All four frameworks exhibit this characteristic behavior of paramagnetic materials where when the temperature decreases the electrons align with the magnetic field. As such, the inverse of the magnetic susceptibility and temperature follows the Curie-Weiss law. The linear line of best fit from temperatures 2 K-200 K for compound 1 gives a Weiss constant of θ = −29.179 K. The linear line of best for compound 2 from temperatures 2 K-151 K, gives a Weiss constant of θ = −10.560 K. The linear line of best fit for 4 from temperatures 2 K-250 K gives a Weiss Constant of θ = −16.330 K. As all of the Weiss constants are negative, they demonstrate varying degrees of antiferromagnetic coupling.

Luminescence
Europium (III) is known for being fluorescent. There are several bands of luminescence that result for 5 D 0 to 7 F J excitation states. However, these are not present in 3. This is due to how copper quenches this fluorescence when present in the same complex. This is seen in Sakamoto et al., who compared fluorescence spectra for Eu(NO 3 ) 3 6H 2 O to a nickeleuropium complex and a copper-europium complex. The copper-europium complex showed complete quenching of the fluorescence of the europium by the copper. This is due to coupling of the two metal centers allowing the copper to quench the europium [30]. There is also an overlap with the absorbance band of the d-d band of Cu(II) ions and emission of the Eu(III) ions [31].

Magnetism
The direct current (dc) magnetic susceptibilities of compounds 1, 2, 3, and 4 were studied under an applied magnetic field of 1000 Oe and a temperature range of 2 K to 300 K, as seen in Figure 7. Compound 1 has a magnetic susceptibility of 0.0308 cm 3 /mol at 300 K. There is a gradual increase in magnetic moment as the temperature decreases. Once the temperature reaches 50 K, the magnetic moment begins to sharply increase, characteristic of paramagnetic behavior. Compound 2 has a magnetic susceptibility of 0.0181 cm 3 /mol at 300 K. It has a gradual increase until 50 K where there is a sharp increase. Framework 4 has a magnetic susceptibility of 0.0326 cm 3 /mol at 300 K. It also has a gradual increase in susceptibility until around 50 K, where it begins to sharply increase. All four frameworks exhibit this characteristic behavior of paramagnetic materials where when the temperature decreases the electrons align with the magnetic field. As such, the inverse of the magnetic susceptibility and temperature follows the Curie-Weiss law. The linear line of best fit from temperatures 2 K-200 K for compound 1 gives a Weiss constant of θ = −29.179 K. The linear line of best for compound 2 from temperatures 2 K-151 K, gives a Weiss constant of θ = −10.560 K. The linear line of best fit for 4 from temperatures 2 K-250 K gives a Weiss Constant of θ = −16.330 K. As all of the Weiss constants are negative, they demonstrate varying degrees of antiferromagnetic coupling. Theoretically the magnetic susceptibility of ground state of Eu 3+ , 7 F0, is zero, as it is non-magnetic [32]. However, there is a slight paramagnetic susceptibility, as shown in Figure 7, with a magnetic susceptibility of 0.0047 cm 3 /mol. This non-zero susceptibility of 3 is likely due to the lone free electron of Cu 2+ , possible lanthanide impurities, or excitation of 7 F0 to the 7 F1 state through Zeeman perturbation [33]. This results in a material that has a very low magnetic susceptibility in comparison to the other three MOFs.
Many MOFs that contain similar lanthanide dimers have weak ferromagnetic or antiferromagnetic interactions at cryogenic temperatures [34][35][36]. For example, Almáši et al. studied the effect of solvent exchange had on the distance between the Gd 3+ ions and magnetic properties in MOF-76. As the distance between the Gd 3+ ions decreased, the magnetic interactions transformed from having ferromagnetic to antiferromagnetic properties [30]. Similar dimers are found with MOFs containing Ho 3+ , which are also known to have antiferromagnetic coupling [33,37]. These interactions were most likely made possible through super exchange mechanisms through the ligand.
Magnetic coupling between the copper and lanthanides is typical [38]. Ferromagnetic coupling found between lanthanides and copper ions occurs due to overlap of orbitals [39]. This ordering normally occurs at very low temperatures (2 K), with very few 3d-4f polymers having higher Tc values [40]. These interactions are often overshadowed by the magnetic interactions of the lanthanide dimer. As the lanthanide dimer has more unpaired electrons verses the copper's one unpaired electron.

Conclusions
In conclusion, a new crystal structure of a mixed metal MOF was discovered with four isostructures containing four lanthanide ions, Gd 3+ , Ho 3+ , Eu 3+, and Dy 3+ , and one Cu 2+ ion. The MOF structure undergoes a reversible change upon removal and addition of DMF solvent. All four different MOFs were determined to have varying degrees of antiferromagnetic interactions.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Figure S1: The asymmetrical view of 1, 2, 3, and 4. Figure S2: The comparison of the activated and solvent replaced PXRDs of compound 1, Figure S3: The comparison of the activated and solvent replaced Theoretically the magnetic susceptibility of ground state of Eu 3+ , 7 F 0 , is zero, as it is non-magnetic [32]. However, there is a slight paramagnetic susceptibility, as shown in Figure 7, with a magnetic susceptibility of 0.0047 cm 3 /mol. This non-zero susceptibility of 3 is likely due to the lone free electron of Cu 2+ , possible lanthanide impurities, or excitation of 7 F 0 to the 7 F 1 state through Zeeman perturbation [33]. This results in a material that has a very low magnetic susceptibility in comparison to the other three MOFs.
Many MOFs that contain similar lanthanide dimers have weak ferromagnetic or antiferromagnetic interactions at cryogenic temperatures [34][35][36]. For example, Almáši et al. studied the effect of solvent exchange had on the distance between the Gd 3+ ions and magnetic properties in MOF-76. As the distance between the Gd 3+ ions decreased, the magnetic interactions transformed from having ferromagnetic to antiferromagnetic properties [30]. Similar dimers are found with MOFs containing Ho 3+ , which are also known to have antiferromagnetic coupling [33,37]. These interactions were most likely made possible through super exchange mechanisms through the ligand.
Magnetic coupling between the copper and lanthanides is typical [38]. Ferromagnetic coupling found between lanthanides and copper ions occurs due to overlap of orbitals [39]. This ordering normally occurs at very low temperatures (2 K), with very few 3d-4f polymers having higher T c values [40]. These interactions are often overshadowed by the magnetic interactions of the lanthanide dimer. As the lanthanide dimer has more unpaired electrons verses the copper's one unpaired electron.

Conclusions
In conclusion, a new crystal structure of a mixed metal MOF was discovered with four isostructures containing four lanthanide ions, Gd 3+ , Ho 3+ , Eu 3+, and Dy 3+ , and one Cu 2+ ion. The MOF structure undergoes a reversible change upon removal and addition of DMF solvent. All four different MOFs were determined to have varying degrees of antiferromagnetic interactions.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/cryst11121547/s1, Figure S1: The asymmetrical view of 1, 2, 3, and 4. Figure S2: The comparison of the activated and solvent replaced PXRDs of compound 1, Figure S3: The comparison of the activated and solvent replaced PXRDs of compound 3, Figure S4: The comparison of the activated and solvent replaced PXRDs of compound 4, Figure S5: The TGA for