Synthesis and Properties of New Molecule-Based Magnets Containing Mn ( II ) , Cu ( II ) and Nitronyl Nitroxide Radical Cation

Neste trabalho, descrevemos a obtenção de dois novos compostos de cobre(II), ou precursores, (Pr-Rad) 2 [Cu(opba)].H 2 O (1) e (Bu-Rad) 2 [Cu(opba)].2H 2 O (2), onde opba é o ligante ortofenilenobis(oxamato) e R-Rad são radicais cátions do tipo nitronil nitróxido. A partir destes compostos 1 e 2, foram sintetizados os dois novos magnetos moleculares [Pr-Rad] 2 [Mn 2 {Cu(opba)} 3 ].3,3 DMSO.5H 2 O (3) e [Bu-Rad] 2 [Mn 2 {Cu(opba)} 3 ].3DMSO.6H 2 O (4), respectivamente. As propriedades magnéticas dos precursores 1 e 2 mostram a presença de interações ferromagnéticas entre os radicais cátions e o cobre (II), na faixa de temperatura entre 20 e 300 K. Os magnetos moleculares 3 e 4 exibem magnetização espontânea com temperatura crítica, T c , de 23 K e 24 K, respectivamente.


Introduction
As evidenced by the remarkable increase in publications on the topic, the field of molecule-based magnetism is flourishing.However, the critical temperature of the molecule-based magnets remains low and up to now there have not been compounds of this type used for any technological application.2][3] One of them decomposes at 350 K, and the other three demagnetize near 315 K.The crystal structures of these compounds have not all been determined.
One of the outstanding features of molecule-based materials is that the magnetic properties may be transformed by quite small and subtle modifications in the molecular chemistry.Variations in the organic molecules may lead to modifications in the magnetism of the material.Sometimes the molecular fragments only serve as spacers and are not involved directly in the interaction between neighboring atoms. 4In other cases, the organic part, although not possessing magnetic moment or not being responsible for the transmission of the interaction is essential to the exchange process. 5In the last few years, numerous magnetic molecular units carrying spin and able to establish chemical connections with other magnetic units have been developed and investigated. 6Several magnetic studies have been performed on these polynuclear systems and theoretical models have been created to correlate the experimental results (susceptibility, magnetization) with the respective Hamiltonian. 7Comparing the theoretical models and the experimental data, it is possible to determine the exchange and/or anisotropy parameters of the system.The results confirm or contradict the type of interaction initially foreseen or desired (antiferromagnetic or ferromagnetic).From these studies, it is conceivable to elaborate new synthetic strategies to tailor materials with specific properties.
The first examples of molecular materials possessing the properties of magnets appeared in 1986.Miller 8  In the beginning of the nineties, we initiated a research in the direction to increase the dimensionality of the compounds using a more 'organic' building block: [Cu(opba)] 2-where opba stands for ortho-phenylenebis(oxamato) (see Figure 1a).We succeeded in designing two-dimensional magnets of formula [cat]  In 1993, we obtained the compound [Me-Rad] 2 [Mn 2 {Cu(opba)} 3 ](DMSO) 2 .2H 2 O, 12 where Me-Rad + is a nitronyl nitroxide radical cation (see Figure 1b).This compound, a magnet with T c of 22 K, presents a rare interlocked structure.Moreover, by the time of the description of this compound, no molecular magnets were known with a resolved structure and T c above of 15 K.
A large number of reports have appeared in the literature on this [cat] 2 [M 2 {Cu(opba)} 3 ].S system involving theoretical 13,14 and different experimental studies, [15][16][17] aiming the development of models to explain the observed magnetic behaviors.Also, numerous reports explored the potentialities of these bimetallic compounds 18,19 through chemical variations.Although several groups have been working on this system, only recently the reproduction of the interlocked structure of the compound [Me-Rad] 2 [Mn 2 {Cu(opba)} 3 ](DMSO) 2 .2H 2 O was achieved. 20,21n this work we describe the synthesis of two new precursors (Pr-Rad)  4).With the present study we intended to verify the influence of changes in the radical cation in the interlocked structure, as well as the effect of these changes on the critical temperature.The results obtained through X-ray structural analysis, magnetic properties and elemental analysis and the critical temperature of the compounds indicate the interlocked structure to be the most likely for these new molecule-based magnets.

Experimental
Syntheses: The radical 2-(pyridine-4-yl)-4,4,5,5tetramethylimidazoline-1-oxyl-3-oxide, abbreviated as Rad, 22 the ligand opba, and the tetrabutylammonium salt of the copper(II) precursor (Bu 4 N) 2 [Cu(opba)] were prepared as already described. 10r-Rad + I -: A mixture of 1.00 g (4.3 mmol) of Rad in 4.5 mL (7.85 g, 46 mmol) of 1-iodopropane and 1.0 mL of THF was stirred for 48 h, at 45 ºC.The green precipitate was collected by filtration, washed with THF and dried under vacuum.Yield: Bu-Rad + I -: A mixture of 1.0 g (4.3 mmol) of Rad in 5.0 mL (8.1 g, 44 mmol) of 1-iodobutane and 1.0 mL of THF was kept under stirring for 72 h, at 45 °C.The yellowgreen precipitate was isolated by filtration, washed with THF and dried under vacuum.Yield: 1.12 g (62 %).Anal.Physical Techniques: X-ray powder diffraction studies were performed in a Rigaku-Geigerflex equipment.Data were collected in the Bragg-Bretano mode with 2q varying from 4.0 to 40 and using a 0.020833° step; monochromatic CuKα radiation was used.Elemental analyses (C, H, N, Mn, Cu) were performed on a Perkin-Elmer 2400 apparatus.The magnetic properties were studied in the Quantum Design MPMS XL7 SQUID magnetometer, working in the dc mode between 2 and 300 K.The raw susceptibility data were corrected for the core diamagnetism estimated as: -397x10 -6 and -433x10 -6 emu mol -1 for the precursors 1 and 2, respectively and -790x10 -6 and -766x10 -6 emu mol -1 for the magnets 3 and 4, respectively.

Results and Discussion
The synthesis of the magnets is very sensitive to the purity of the precursors and some other parameters.To form materials of stoichiometry Cu:Mn 3:2, the copper building block must be present in excess.If the Cu:Mn ratio in the solution is low, formation of MnCu(opba) chains 23 occurs instead.The effect of reagents' concentration upon the reaction and on the formation of crystals is crucial.If the solution is over diluted no product is obtained, as Mn(II) is slightly sensitive and decomposes before or during the crystallization, producing an insoluble brown impurity.The use of water as a solvent favors formation of chains and the use of chloride is preferable to other salts.
A variety of experiments were performed to find the best crystallization conditions.Beyond variations of concentration and Cu:Mn ratio, other parameters were considered such as the temperature, the type of crystallization flask (glass or polystyrene) and mixture of solvents.Regarding the stability of the magnets, it should be pointed out that, after removal from the mother liquor, they could be stored in a freezer for several months.

Precursors
The magnetic susceptibility data for 1 and 2 are shown in Figure 2 in the form of the χ M T versus T plot, χ M being the molar magnetic susceptibility and T the temperature.At room temperature χ M T is equal to 1.05 emu K mol -1 for 1 and 1.21 emu K mol -1 for 2. These values are close to the theoretical value expected for 3 isolated spins S = 1/2 (two spins referring to the radicals and one to the copper) that can be calculated by the Curie law: Substituting the Bohr magneton (β) and Boltzmann constant (k) values into equation ( 1): and substituting the S value in equation ( 2) and considering g = 2: As T is lowered, χ M T first increases, reaches a maximum at 10.6 K and 6.1 K with χ M T = 1.56 emu K mol -1 and 1.59 emu K mol -1 for 1 and 2, respectively, and finally decreases rapidly.The profile of this curve indicates that in the two compounds some ferromagnetic interactions are operative.The decrease of χ M T in the low-temperature region may be attributed to intermolecular interactions.
The magnetic properties for the analogous precursor (Et-Rad) 2 [Cu(opba)].CH 3 CN.H 2 O has already been described. 20Its χ M T versus T curve shows the same outline as those of 1 and 2 and corresponds to a dimer ferromagnetically coupled and a spin S = 1/2 isolated.Gatteschi and co-workers have already shown that the orthogonality of a copper d x 2 -y 2 -type and a nitronyl nitroxide radical p πtype magnetic orbitals affords ferromagnetic coupling. 24his orthogonality can be obtained, for example, if the radical occupies an apical position of the Cu(II) ion, as in (Et-Rad X-ray powder diffraction were performed for 1 and 2 and the Rietveld method has been employed in structural analysis.For the sake of comparison, diffraction data were also obtained for (Et-Rad) 2 [Cu(opba)].CH 3 CN.H 2 O and its crystal structure 20 was used as a reference.The results obtained for the structure refinements for the three compounds are not of best quality.Nevertheless, as in the case of the magnetic results, the similarity in the obtained parameters indicates they positively have a similar structure.

Magnets
X-ray powder diffraction study of the magnets was performed as described previously.The results strongly suggest that 3 and 4 have similar structures to those of the magnets [Et-Rad] 2 [M II 2 {Cu(opba)} 3 ](DMSO) x .0.25H 2 O, with M II = Mn(II) or Co(II), for which the structures have been determined; 20 the Mn(II) compound was used as a reference (see Figure 4).The general architecture of the reference compound consists of two equivalent two-dimensional networks, each network being formed by parallel honeycomb layers.A layer is made up of edge-sharing hexagons with Mn(II) ions at each corner and Cu(II) ions at the middle of each edge.Two nearest neighboring metal ions are bridged by an oxamato group.The network layers stack above each other in a graphite-like fashion.The networks interpenetrate with a full interlocking of the Mn 6 Cu 6 hexagons, as shown in Figure 5.The topology may be described as a threedimensional wire netting.The hexagons are connected further through radical cations that interact with two copper atoms, thus bridging the two perpendicular networks.
The χ M T versus T plot for 4 is shown in Figure 6.This curve is characteristic of ferrimagnetic compounds 25  At room temperature, χ M T is equal to 8.8 emu K mol -1 , which is slightly lower than expected for two radical, two high-spin Mn(II) and three Cu(II) isolated ions (expected value: 10.6 emu K mol -1 ).It should be reminded that the model used in the calculation of the theoretical value of χ M T considers that, at room temperature, there is no interaction between the spins, which is an approximation.As T is lowered, χ M T first decreases smoothly, reaches a minimum around 90 K with χ M T = 8.1 emu K mol -1 and then increases very abruptly.The minimum in the χ M T versus T plot is characteristic of a ferrimagnetic material. 26t high temperature, χ M T approaches the paramagnetic limit.As T is lowered, the decrease in χ M T corresponds to a short-range order where the local spins S Cu and S Mn of adjacent metal ions are aligned antiparallel but have no correlation with neighboring MnCu units.As T is further lowered, below the χ M T minimum, the randomizing thermal effects are reduced and the correlation length increases.This leads to a spontaneous magnetization below a critical temperature.There is no theoretical model to interpret quantitatively these magnetic data due to the complexity of the structure, as well as the presence of three kinds of spin carriers.However, for compounds [Bu 4 N] 2 [Mn 2 {Cu(opba)} 3 ].S, there are studies 14 that allowed to evaluate the coupling constant between Mn(II) and Cu(II) as J = -33.1 cm -1 (with g Mn = 2.0 and g Cu = 2.2).From the similarity of the magnetic behavior of [Bu 4 N] 2 [Mn 2 {Cu(opba)} 3 ].S and the magnets 3 and 4, in this temperature range, we can affirm that the interaction between Mn(II) and Cu(II) in these compounds is antiferromagnetic with nearly the same intensity.From the χ M T curve it is not possible to determine the specific behavior of the radical cation spins.However, measurements on the compound 27 [Et-Rad] 2 [Mg 2 {Cu(opba)} 3 ].4.5DMSO.5.5H 2 O confirmed that the radical cation interacts ferromagnetically with the Cu(II).It follows that the radical spins tend to align along a direction opposite to that of the S Mn spins.
The temperature dependence of the molar magnetization M for 3 and 4 were investigated with an applied magnetic field H of 10 Oe. Figure 7 shows the results for 3.
The curve increases very rapidly starting from 24 K and reaching a maximum value of magnetization of 2600 emu Oe mol -1 .This behavior is attributed 28 to the appearance of magnetic order around 23 K.At the critical temperature, the magnetization increases abruptly, but its value is limited (the saturation magnetization is not reached) by the effects of the demagnetizing field and domains. 26Compound 4 presents a similar behavior, with a T c of 24 K.
The field dependence of the magnetization for 3, at 5 K, is shown in Figure 8.The magnetization quickly increases up to a field of 2 kOe and afterwards the slope continuously decreases.The saturation magnetization, M sat , was not reached, even with magnetic fields up to 50 kOe.With H = 2 kOe, the magnetization is 5.0 Nβ and with a field of 50 kOe the value 6.4 Nβ is found.For [Me-Rad] 2 [Mn 2 {Cu(opba)} 3 ] (DMSO) 2 .2H 2 O, 29 even with the application of a magnetic field of 200 kOe, M sat was not reached.We can interpret this curves considering that initially the spins of the radicals and the Cu(II) ions are coupled ferromagnetically.The magnetic field destroys this ferromagnetic alignment progressively and the spins of the radicals begin to align parallel to the spins of the Mn(II) ions.This is only possible if the interaction between the organic radicals and the Cu(II) ions is weak.In this way, the magnetization increases starting from a value of 5Nβ (2x5S Mn -3xS Cu -2xS Rad ) going in the direction of a limit value of 9Nβ (2x5S Mn -3xS Cu +2xS Rad ).On the other hand, to overcome the strong coupling between Mn(II) and Cu(II), magnetic fields of the order of 40 Tesla (400 kOe) would be necessary, if one supposes J Mn-Cu of about -30 cm -1 .Finally, in the field dependence of the magnetization, we observe that the hysteresis loop for 3 and 4 show a coercive field smaller than 10 Oe.These compounds are therefore soft magnets.

Conclusions
The use of the building block [Cu(opba)] 2-presents some features that make it very attractive: i) achievement of soluble precursors in nonpolar solvents with counterions as R 4 N + or R-Rad + ; ii) flexibility to introduce groups in the phenyl ring; iii) counterion carrying other properties; iv) chelate/macrocycle stability of the copper, or other metal ion and v) ligand bis-bidentate capable of bridging a large number of metal ions.Furthermore, due to the stability of the Cu(II) within the block, the addition of the second metal ion is facilitated since there is no metalcompetition for the coordination sites of the ligand.
In this work, some features of these building blocks have been used to insert three different spins in molecular systems.The choice of the metals obeys a ferrimagnetic strategy which consists in placing Cu(II), with a small spin (S = 1/2) and another species with the greatest possible spin for a transition metal (S = 5/2).The third spin comes from a radical cation that is introduced, as well as Cu(II), from precursors 1 and 2. Already in these compounds, one observes ferromagnetic interactions between copper and the radical cation(s) as in the precursor (Et-Rad) 2 [Cu(opba)].CH 3 CN.H 2 O previously described.
The reactions of the precursors with a manganese salt led to new molecule-based magnets 3 and 4, which present magnetic order at 23 and 24 K, respectively.Our findings strongly suggest that the structure of these compounds consists of two nearly perpendicular graphite-like networks similar to that obtained for other compounds of this family; thus, the synthesis methods are perfectly reproducible.On the other hand, our work showed that the increase in the chain length of the alkyl group linked to the radical does not hinder the formation of this type of system.

Figure 2 .
Figure 2. χ M T versus T plot, at 1000 Oe, for (Pr-Rad) 2 [Cu(opba)].H 2 O (!) and (Bu-Rad) 2 [Cu(opba)].2H 2 O (").The full lines are guides for the eye ) 2 [Cu(opba)].CH 3 CN.H 2 O.The magnetic properties of this compound are very similar to those of the precursors 1 and 2, indicating that they have the same molecular structure.The structure of (Et-Rad) 2 [Cu(opba)].CH 3 CN.H 2 O has been solved by a single crystal X-ray diffraction analysis and, as shown in Figure 3, consists of one [Cu(opba)] 2-, two radical cations and molecules of CH 3 CN and H 2 O not coordinated.One of the radical cations is weakly linked to the copper while the other is relatively isolated.

Figure 5 .
Figure 5. Schematic representation of the interpenetration of the honeycomb layers of the M 6 Cu 6 hexagons.The organic radicals and the solvent molecules have been omitted for the sake of clarity