Synthesis , Structural and Magnetic Characterization of a New Copper ( II )-Nitronyl Nitroxide Radical Complex

c École Polytechnique Fédérale de Lausanne, Laboratoire de Cristallographie, BSP, CH-1015 Lausanne, Switzerland d Instituto de Química, Universidade Federal Fluminense, Outeiro de São João Batista s/n, Campus do Valonguinho, 24020-150 Niterói-RJ, Brazil e Laboratory of Molecular Magnetism, Dipartimento di Chimica & UdR INSTM, Università degli Studi di Firenze, Via della Lastruccia 3, 50019, Sesto Fiorentino (FI), Italy


Introduction
The synthesis of molecule-based magnetic materials has been the object of increasing interest in the last years.In fact, the molecular nature of these compounds allows one to obtain magnetic properties 1 in association with chemical and optical properties.These molecule-based magnets are potential building-blocks for molecular electronic devices and can eventually be exploited in the so-called molecular electronics. 2n order to obtain magnetic molecular materials, 3 several approaches are followed, namely: metal-metal, purely organic, or metal-radical approach, depending on the chemical nature of the magnetic centers.The most usual building-blocks used to prepare metal-metal molecular magnets are probably the hexacyano-metalates 4 [M(CN) 6 ] 3-and the trisoxalatometalates 5 [M(C 2 O 2 ) 3 ] 3-(where M is a transition metal ion).Due to their capacity to establish bonds in three directions, these building-blocks may lead to three-dimensional magnetic networks.In the case of the purely organic approach, nitronyl nitroxide radicals (NIT) were intensively investigated 6 after the discovery of the first purely organic ferromagnet 2-(p-nitrophenyl)nitronyl nitroxide, in 1991. 7The NIT radicals are also used in the metal-radical approach, 8 where they are combined with different paramagnetic metal ions to prepare a large number of coordination complexes.With the M(hfac) 2 (hfac = hexafluoroacetylacetonate) paramagnetic unit, one-dimensional ferrimagnetic chains were obtained, leading to long range ordering around 10 K. 9 The most famous example of these chains is the Co(hfac) 2 (NITPhOMe) (where NITPhOMe is 4'-methoxyphenyl-4,4,5,5-tetramethylimidazoline-1oxyl-3-oxide) which was reported to behave as a molecular magnetic nanowire. 10Within the metalradical approach there are two goals.One is the rational design of molecular compounds exhibiting spontaneous magnetization.The other is the study of the relation between the structure and the magnetic properties of this kind of compounds. 11n this paper we present the structure and magnetic properties of a dimeric copper(II) building-block [Cu 2 (bopba)] 4-(where bopba stands for bis-o-phenylenebis(oxamate)) and the synthesis and characterization of a new [Cu 2 (bopba)] 4--NIT compound.The nitronyl nitroxide radical chosen was the cationic para-2-(4pyridil)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide, also named pEtRad + .A schematic representation of these two magnetic blocks is given in Figure 1.

Reagents
All chemicals were purchased from Fluka and Aldrich and used without further purification.

Spectroscopic measurements
Infrared spectra were measured over the region 400-4000 cm -1 with a Nicolet Magna 760 FT-IR instrument, with 4 cm -1 resolution.The samples were examined as KBr pellets.

Elemental analysis
Carbon, nitrogen and hydrogen contents were determined on a Perkin-Elmer 2400 apparatus.
The (NBu 4 ) 4 [Cu 2 (bopba)] (1) complex was synthesized as described previously. 10,13Tiny crystals were obtained by slow evaporation of a solution containing (NBu 4 ) 4 [Cu 2 (bopba)] and acetonitrile at room temperature.Single crystals obtained by this procedure are brown-red and exhibit prismatic elongated habit.All attempts to obtain larger crystals have failed.

Magnetic measurements
Variable-temperature magnetic susceptibility and isothermal magnetization measurements were performed on polycrystalline samples in the temperature range 2.5-300 K and in the field range 0-60 kOe with a Cryogenics S-600 SQUID magnetometer.

Crystallography
X-ray diffraction data collection on a (NBu 4 ) 4 [Cu 2 (bopba)] single crystal was performed on the Xcalibur CCD Diffractometer from the Swiss Norwegian Beam Line of the European Synchrotron Radiation Facility (SNBL-ESRF) at 120 K using radiation with λ = 0.80 Å.The sample was cooled with a nitrogen gas stream controlled by an Oxford Cryosystem apparatus. 16The Crysalis suite 17 was employed to integrate the intensities.Scaling based on the minimization of the R int parameter was applied to correct the data due to the incoming beam intensity decay.No absorption correction was applied to the data set.
The (NBu 4 ) 4 [Cu 2 (bopba)] crystal used in this experiment was quite small, about 0.5 × 0.1 × 0.5 mm 3 .It was also quite bad X-ray scatterer and the best I/σ(I) ratio verified at 0.85-0.80Å of resolution shell was 2.3.All together, these factors limited the final data quality and consequently the refinement statistics.

Description of the structure of the (NBu 4 ) 4 [Cu 2 (bopba)] complex
The (NBu 4 ) 4 [Cu 2 (bopba)] structure was solved by direct methods in the non-centrosymmetric space group P2 1 using the SHELXS program. 18The positions of all atoms of the [Cu 2 (bopba)] 4-molecule could be unambiguously assigned on consecutive difference Fourier maps; however, the positions of some carbon atoms of the NBu 4 + molecules were not easily identified due to structural disorder.Refinements were performed using SHELXL 17 based on F 2 through full-matrix least square routine.Restraints have been used to fix the phenyl geometry in the [Cu 2 (bopba)] 4- moiety.Split positions were assigned to several carbons of the NBu 4 + groups and their geometry, isotropic ADPs and occupation parameters were restrained and refined accordingly.Hydrogen atoms were added to the structure according to the riding model. 19All disordered atoms had their ADPs refined anisotropically.Despite the disorder, the final model gives acceptable results.Sample characteristics, data collection, data reduction and refinement parameters are summarized in Table 1; selected atomic parameters are listed in Table 2.
The structure of the copper(II) building block is shown in Figure 2.For the sake of simplicity, the NBu 4 + cations  1.247 (13)  structure and packing are sketched in Figure 3.The Cu 2 (bopba) packing along the c direction shows molecules rotated by about 100° in consecutive layers giving rise to a sort of zigzag arrangement represented in Figure 3b.

Magnetic properties
The magnetic susceptibility data for compound 1 is presented in Figure 4 in the form of the χ M T versus T plot, χ M being the molar magnetic susceptibility and T the temperature.The χ M T value at room temperature is close to 0.7 emu K mol -1 , near the expected value for a two independent ½ spin system, 0.75 emu K mol -1 .When T is lowered, χ M T is almost flat with a small upward curvature reaching a maximum at 21 K with χ M T = 0.71 emu K mol -1 and then decreases abruptly below 4 K.This uncommon behavior, reproduced in three samples of different syntheses, may indicate the presence of competing short range ferromagnetic and antiferromagnetic interactions and a clear overall antiferromagnetic behavior setting in at low temperatures.The reciprocal susceptibility versus temperature, shown in the inset of the Figure 4, indicates the presence of average ferromagnetic interactions, with a positive Curie-Weiss temperature of about 0.5 K.The presence of ferromagnetic interactions was also observed in other copper(II) complexes, 20 with a similar positive value for the Curie-Weiss temperature.
The magnetic susceptibility data for 2 is presented in Figure 5 in the form of χ M T versus T. At room temperature the χ M T value is equal to 1.93 emu K mol -1 , which is smaller than the expected value for six magnetically non-interacting S= ½ spins (two copper(II) ions and four nitroxide radicals), 2.25 emu K mol -1 .As the temperature is decreased, χ M T increases to reach a maximum at 6 K with χ M T=2.21 emu K mol -1 and then decreases abruptly below 6 K.The magnetization data obtained at 1.8 K are shown in Figure 6 in the form of M versus H, where M is the magnetization and H is the applied field.The magnetization reaches a maximum value of 5.6 μ B at 60 kOe, which is close to the expected saturation value for six ½ spins.The inset of Figure 6 shows the measurement done at 1.8 K and at 4 K in the form of M versus H/T.The two curves do not superpose, which indicates the presence of antiferromagnetic interactions between the spin carriers at these temperatures.
Both the M versus H and χ M T versus T measurements allow to speculate about the kind of magnetic interactions operative between the magnetic centers present in the (pEtRad) 4 [Cu 2 (bopba)] compound.The value of the χ M T versus T at 300 K is slightly smaller than expected for six free spins.This behavior was also observed in other compounds with copper(II) and nitronyl nitroxide radicals, 14,21 and may be related to a smaller value of g or antiferromagnetic interactions.
On the other hand, the magnetic behavior between 300 K and 6 K observed in the χ M T versus T plot (upward curvature in Figure 5) clearly shows the presence of ferromagnetic interactions.In order to understand this behavior, one has to recall on how metal transition ions generally interact with nitronyl nitroxide radicals.These are stable organic radicals that have an N-O group with an unpaired electron in a π* orbital, equally shared by the nitrogen and oxygen atoms. 22When the nitronyl nitroxide binds through the oxygen atom, the four possible relative orientations of the magnetic orbitals are shown on Figure 7.
In the case a), if the M-O-N angle is equal to 180°, the two orbitals π * and d x 2 -y 2 are orthogonal to each other and in this case the spins have to be parallel, leading to a ferromagnetic coupling.This is an uncommon situation that occurs in copper(II)-nitroxide complexes when the radical is in an coordination site.In copper(II) complexes, when the Cu-O-N angle increases, the angle determined by the projection of the O-N bond on the XY plane with the X axis, as well as the dihedral angle between the Cu-O-N and the nitroxide plane, have an important influence on the exchange integral value, but not on its sign, which remains positive, leading to a ferromagnetic interaction. 23,7ase b) is expected in copper(II)-nitroxide complexes when the radical coordinates in an equatorial site.If the M-O-N angle is different from 180°, the coupling is antiferromagnetic.
Cases c) and d) are found in complexes of nickel(II), cobalt(II) and manganese(II), with the observed coupling being strongly antiferromagnetic.
The ferromagnetic coupling observed in the (pEtRad) 4 [Cu 2 (bopba)] compound may be explained by the coordination of each copper(II) ion with one nitronyl nitroxide radical as in case a) above.This was previously reported in a similar compound 14 where the magnetic data were fitted by considering a ferromagnetic interaction between the copper(II) ion and one NIT radical.The value reported for the coupling constant was J=31 cm -1 .We tried first to fit our data with this same model, i.e. one radical ferromagnetically coupled to each copper(II) and the others uncoupled.The next trial was to add an antiferromagnetic coupling between one uncoupled radical to the other free radical of the neighboring molecule.In both cases it was not possible to obtain a satisfactory fit, though we could estimate that the ferromagnetic coupling constant value is around 23 cm -1 in either case.Unfortunately the crystal structure could not be determined and other models of interacting spins would be too speculative.Our results suggest that, in the (pEtRad) 4 [Cu 2 (bopba)] compound, competing short range ferromagnetic and antiferromagnetic interactions are present with a slight predominance of ferromagnetic interaction at high temperatures.At low temperature other weak antiferromagnetic interactions, presumably of intermolecular origin, become important, leading to a decrease of χ M T below 6 K.The abrupt character of this behavior suggests that a long range order may be established and a phase transition to a three-dimensional magnetic order may occur below this temperature.

Conclusions
We presented the crystallographic structure and the magnetic properties of a promising copper(II) buildingblock (compound 1) that was already used to prepare a new molecular ferrimagnetic compound 12 by substitution of the tetrabuthylammonium cation with a manganese(II) ion.A new compound was synthesized and characterized by replacing the NBu 4 + species by a cationic nitronyl nitroxide radical.The magnetic behavior of this compound indicated the presence of competing short range ferromagnetic and antiferromagnetic interactions.In order to explain fully the magnetic behavior of this compound, the knowledge of the crystalline structure is necessary and new efforts to obtain single crystals are in progress.
and Magnetic Characterization of a New Copper(II) J.Braz.Chem.Soc.

Figure 7 .
Figure 7. Schematic representation of the interactions between the magnetic orbitals of a metal ion and of a nitroxide radical.