Giant Heterometallic [Mn36Ni4]0/2− and [Mn32Co8] “Loops-of-Loops-and-Supertetrahedra” Molecular Aggregates

We report the synthesis, crystal structures and magnetic properties of the giant heterometallic [Mn36Ni4]2−/0 (compounds 1, 2)/[Mn32Co8] (compound 3) “loops-of-loops-and-supertetrahedra” molecular aggregates and of a [Mn2Ni6]2+ compound (cation of 4) that is structurally related with the cation co-crystallizing with the anion of 1. In particular, after the initial preparation and characterization of compound [Mn2Ni6(μ4-O)2(μ3-OH)3(μ3-Cl)3(O2CCH3)6(py)8]2+[Mn36Ni4(μ4-O)8(μ3-O)4(μ3-Cl)8Cl4(O2CCH3)26(pd)24(py)4]2− (1) we targeted the isolation of (i) both the cationic and the anionic aggregates of 1 in a discrete form and (ii) the Mn/Co analog of [Mn36Ni4]2− aggregate. Our synthetic efforts toward these directions afforded the discrete [Mn36Ni4] “loops-of-loops-and-supertetrahedra” aggregate [Mn36Ni4(μ4-O)8(μ3-O)4(μ3-Cl)8Cl2(O2CCH3)26(pd)24(py)4(H2O)2] (2), the heterometallic Mn/Co analog [Mn32Co8(μ4-O)8(μ3-O)4(μ3-Cl)8Cl2(μ2-OCH2CH3)2(O2CCH3)28(pd)22(py)6] (3) and the discrete [Mn2Ni6]2+ cation [Mn2Ni6(μ4-O)2(μ3-OH)4(μ3-Cl)2(O2CCH3)6(py)8](ClO4)(OH) (4). The structure of 1 consists of a mixed valence [Mn28IIIMn8IINi4II]2− molecular aggregate that contains two Mn8IIINi2II loops separated by two Mn6IIIMn4II supertetrahedral units and a [Mn2IIINi6II]2+ cation based on two [MnIIINi3II(μ4-O)(μ3-OH)1.5(μ3-Cl)1.5]4+ cubane sub-units connected through both mono- and tri-atomic bridges provided by the μ4-O2− and carboxylate anions. The structures of 2–4 are related to those of the compounds co-crystallized in 1 exhibiting however some differences that shall be discussed in detail in the manuscript. Magnetism studies revealed the presence of dominant ferromagnetic interactions in 1–3 that lead to large ground state spin (ST) values for the “loops-of-loops-and-supertetrahedra” aggregates and antiferromagnetic exchange interactions in 4 that lead to a low (and possibly zero) ST value. In particular, dc and ac magnetic susceptibility studies revealed that the discrete [Mn36Ni4] aggregate exhibits a large ST value ~ 26 but is not a new SMM. The ac magnetic susceptibility studies of the [Mn32Co8] analog revealed an extremely weak beginning of an out-of-phase tail indicating the presence of a very small relaxation barrier assignable to the anisotropic Co2+ions and a resulting out-of-phase ac signal whose peak is at very low T.


INTRODUCTION
High nuclearity Mn carboxylate clusters continue to attract significant attention mainly because of their structural characteristics and physical properties (Bagai and Christou, 2009;Kostakis et al., 2010;Escuer et al., 2014). In particular, such compounds often exhibit interesting magnetic properties including high spin ground state values (Ako et al., 2006;Moushi et al., 2009) and single-molecule magnetism behavior (Sessoli et al., 1993;Bagai and Christou, 2009;Inglis et al., 2012;Milios and Winpenny, 2015). The latter appears in molecules exhibiting a large spin ground state (S T ) value and significant easy axis magnetoanisotropy (Christou et al., 2000;Nakano and Oshio, 2011;Ferrando-Soria et al., 2017). In addition, Mn complexes have attracted significant attention since they are involved in the search for structural and functional analogs of the tetranuclear Mn complex that is present in the active site of photosystem II and is responsible for the photosynthetic oxidation of H 2 O to molecular O 2 (Mukherjee et al., 2012;Yano and Yachandra, 2014;Gerey et al., 2016). Thus, Mn clusters have been proposed for various applications in diverse areas including magnetic refrigeration (Zheng et al., 2014), molecular spintronics (Bogani, 2015), quantum computation (Aromí et al., 2012), and catalysis (Maayan et al., 2018).
Our group has been exploring reactions of diols with Mn-containing precursor compounds targeting to new high nuclearity Mn clusters and SMMs (Tasiopoulos and Perlepes, 2008 aggregates have been stabilized in several cases, especially with polyol-type ligands and in most cases these compounds exhibited entirely ferromagnetic exchange interactions and S T = 22 (Stamatatos et al., 2006;Manoli et al., 2007Manoli et al., , 2008Wu et al., 2011). ; M = Na + , x = 0; M = Mn 2+ , x = 1) loops linked through Na + or Mn 2+ ions (called "loops-of-loops") and have a saddle-like topology. The [Mn 44 ] analog of this family displays a spin S T = 6 ground state and SMM behavior (Moushi et al., 2007(Moushi et al., , 2010a. Further investigation of the reactions that afforded the [Mn 40 M 4 ] loops-of-loops aggregates involved the use of various 3d paramagnetic metal ions in an attempt to isolate a series of heterometallic Mn/3d analogs and/or other large aggregates composed of smaller clusters. These investigations afforded compounds [Mn 2 Ni 6 (µ 4 -O) 2 (µ 3 -OH) 3 (µ 3 -Cl) 3 (O 2 CCH 3 ) 6 (py) 8 ] 2+  Co 8 ] analog (complex 3) the existence of an out-of-phase tail is an indication of SMM behavior, however, further studies are required to confirm this conclusion. Part of this work, involving the synthesis and characterization of compound 2 has been communicated previously (Charalambous et al., 2012).

Materials and Physical Measurements
All manipulations were performed under aerobic conditions using chemicals and solvents as received, unless otherwise stated. [Mn 3 O(O 2 CCH 3 ) 6 (py) 3 ]·py was prepared as previously described (Vincent et al., 1987).
IR spectra were recorded in the solid state (KBr pellets) in the 4,000-400 cm −1 range using a Shimadzu Prestige−21 spectrometer. Elemental analysis (C, H, and N) were performed by the in-house facilities of the Chemistry Department at the University of Florida.
Variable-temperature dc magnetic susceptibility data down to 1.80 K were collected on a Quantum Design MPMS-XL SQUID magnetometer equipped with a 70 kG (7 T) dc magnet at the University of Florida. Diamagnetic corrections were applied to the observed paramagnetic susceptibilities using Pascal's constants. Samples were embedded in solid eicosane to prevent torquing. AC magnetic susceptibility data were collected on the same instrument employing a 3.5 G AC field oscillating at frequencies up to 1,500 Hz. Magnetization vs. field and temperature data were fit using the program MAGNET (Davidson, E. R.) 1 .

Experimental
To a stirred brown solution of [Mn 3 O(O 2 CCH 3 ) 6 (py) 3 ]·py (0.23 g, 0.27 mmol) in 15 ml CH 3 CN were added pdH 2 (100 µL, 0.105 g, 1.38 mmol) and solid NiCl 2 ·6H 2 O (0.066 g, 0.27 mmol). The reaction mixture was left under magnetic stirring for 10 min, filtered off and the filtrate was left undisturbed at room temperature. After 1 week brown X-ray quality crystals of 1 suitable for X-ray structural determination were formed. The crystals were isolated by filtration, washed with CH 3 CN and dried in vacuo; the yield was ∼36%. The crystals for X-ray studies were maintained in contact with mother liquor to prevent solvent loss. % C H N Anal. for C 196 To a stirred brown solution of [Mn 3 O(O 2 CCH 3 ) 6 (py) 3 ]·py (0.23 g, 0.27 mmol) in 12 ml of EtOH were added pdH 2 (300 µL, 0.316 g, 4.15 mmol) and solid CoCl 2 ·6H 2 O (0.066 g, 0.27 mmol) and the reaction mixture was left under magnetic stirring for 2 h. The resulting red-brown slurry was filtered off and the filtrate was layered with Et 2 O (1:3 v/v). After 2 weeks brown crystals of (3)·3.84 EtOH·6H 2 O were formed suitable for X-ray structural determination. The crystals were isolated by filtration, washed with EtOH and dried in vacuo; the yield was 29%. The crystals for X-ray studies were maintained in contact with mother liquor to prevent solvent loss. % C H N Anal. for [ To a stirred brown solution of [Mn 3 O(O 2 CCH 3 ) 6 (py) 3 ]·py (0.20 g, 0.24 mmol) in 15 mL EtOH were added solids NiCl 2 ·6H 2 O (0.06 g, 0.24 mmol) and NaClO 4 (0.03 g, 0.24 mmol) and the reaction mixture was left under magnetic stirring at room temperature for 2 h. The resulting brown slurry was filtered off and the dark brown filtrate was layered with Et 2 O (1:3 v/v) and left undisturbed at room temperature for a period of 1 week, upon which yellow-brown crystals of 4 suitable for Xray structural determination were formed. The crystals were isolated by filtration, washed with EtOH and dried in vacuo; the yield was ∼21%.

Syntheses
We have been systematically studying reactions of Mn salts and preformed clusters with diols as a route to new polynuclear clusters with novel structural characteristics and interesting magnetism (Tasiopoulos and Perlepes, 2008;Moushi et al., 2009Moushi et al., , 2010bSkordi et al., 2018). These studies have focused on the use of simple aliphatic diols such as pdH 2 and its derivatives which due to their alkoxide arms exhibit a high bridging capability and a fruitful coordination chemistry. We recently reported a family of large molecular aggregates consisting of four smaller clusters linked through Na + or Mn 2+ ions (Moushi et al., 2010a). These large tetrameric {[Mn 10 M(µ 3 -O) 2 (O 2 CCH 3 ) 13 (pd) 6 (py) 2 ] 4 } x+ ([Mn 40 M 4 After the isolation of compound 2 was realized, our efforts were focused on the synthesis of other Mn 40−x M x (M = a 3d metal ion) heterometallic "loops-of-loops-and-supertetrahedra" molecular aggregates and also of the [Mn 2 Ni 6 ] cation of 1 in a discrete form. The preparation of a Mn 40−x Co x analog was our initial target since the incorporation in this structure of the highly anisotropic Co 2+ ions could result in the appearance of a different magnetic behavior in the resulting compound than those shown in 1 and 2.
The molecular structure of the [Mn 36 Ni 4 ] cluster of 2·2CH 3 CN·12.30H 2 O exhibits a striking similarity to the anion of 1. The main difference between the two complexes is their overall charge since compound 1 is anionic with a 2-charge and 2 is neutral. This difference in the charges appears because a terminal Cl − ligand of 1 linked to a Mn II ion of the supertetrahedral sub-units has been replaced in 2 by a terminal H 2 O molecule. As a result, compound 2 contains two less Cl − anions and is neutral and thus, the [Mn 36 Ni 4 ] aggregate is the only metal cluster appearing in the crystal structure.
The molecular structure of (3)·3.84 EtOH·6H 2 O (Figure 2, top) is also related to the anion of 1 and to 2·2CH 3 CN·12.30H 2 O with the main difference obviously being the presence in 3 of 32 Mn/8 Co ions instead of 36 Mn/4 Ni ions that appear in the anion of 1 and in 2. These eight Co II ions are located in the decametallic [Mn III 8 Co II 2 ] loops (Figure 2, bottom left) in the same positions that the Ni II ions of the [Mn 36 Ni 4 ] aggregates are found and also in the [Mn III 6 Mn II 2 Co II 2 ] supertetrahedral sub-units (Figure 2, bottom right) The molecular structure of compound 4 is related to that of the cation of 1. In fact, there are only very minor differences between the two compounds. These include the replacement in 4 of the mixed 0.5 Cl − /0.5 OH − site that bridges 3 Ni II ions of each cubane by a µ 3 -OH − anion. In addition, in complex 4 the positive charge of the [Mn III 2 Ni II 6 ] 2+ cation is balanced by a ClO − 4 and a OH − lattice anions instead of the [Mn 36 Ni 4 ] 2− anionic aggregate.

Magnetic Properties
Solid-state dc magnetic susceptibility measurements were performed on polycrystalline samples of complexes 1·6 H 2 O, 2·10 H 2 O, 3·20 H 2 O, and 4·2 H 2 O under a magnetic field of 0.1 T in the temperature range 5-300 K. The obtained data are shown as χ M T vs. T plot in Figure 3. For complexes 1·6 H 2 O, 2·10 H 2 O, and 3·20 H 2 O the χ M T value increases continuously from 172.1, 118.6, and 130.5 cm 3 mol −1 K at 300 K to a maximum value of 492.0 (at 20 K), 325.6 (at 15 K), and 273.7 cm 3 mol −1 K (at 25 K) and then decreases at low T to 465.1, 304.3, and 226.1 cm 3 mol −1 K at 5 K, respectively. For 4·2 H 2 O, the χ M T value at 300 K is 18.03 cm 3 mol −1 K and decreases continuously with decreasing temperature reaching a value of ∼15.03 cm 3 mol −1 K at 100 K and then rapidly to 2.94 cm 3 mol −1 K at 5.0K (Figure 3 and Figure S6). The increase of the χ M T values with decreasing T in 1·6 H 2 O-3·20 H 2 O indicates the presence of dominant ferromagnetic exchange interactions. The maximum χ M T values of 2·10 H 2 O and 3·20 H 2 O suggest S T values of ∼ 26 ± 1 and 22 ± 1, respectively. However, we note that in the case of 3·20 H 2 O it may not be safe to exclude any conclusions for the spin ground state using the spin-only formula due to the presence of Co II ions which is well known that exhibit strong spin-orbit coupling. The decrease in the χ M T value at the lowest temperatures is attributed to zero-field splitting (ZFS), Zeeman effects from the applied field, and/or any weak intermolecular antiferromagnetic exchange interactions. In the case of 4·2 H 2 O, the continuous decrease of χ M T with decreasing temperature, the small χ M T value at 5 K and the fact that the curve heads to 0 at 0 K suggests the presence of antiferromagnetic exchange interactions possibly leading to a diamagnetic ground state. This may be rationalized assuming that the two [Mn III Ni II 3 ] are antiferromagnetically coupled leading to a diamagnetic ground state.
Given the size and the complexity of the structures of 1-3, it is not possible to apply the Kambe method to determine the individual pairwise Mn-Mn, Mn-Ni, and Mn-Co exchange interaction parameters. In addition, the existence of two complexes co-crystallizing together in the structure of 1·6 H 2 O does not allow to obtain information for the D and S T values of these complexes. Furthermore, in 3·20 H 2 O, the presence of Co II ions exhibiting strong spin-orbit coupling and of many low lying excited states appearing due to the complexity of the giant [Mn 32 Co 8 ] aggregate did not allow to obtain reliable S and D values from reduced magnetization fitting.
In the case of compound 2·10 H 2 O magnetization vs. dc field measurements at applied magnetic fields and temperatures in the 1-10 kG and 1.8-4.0 K ranges, respectively were performed. Low field data (≤1.0 T) were used, as we have previously done for many Mn clusters containing Mn II atoms, to avoid problems from low-lying excited states. The data for complex 2·10 H 2 O are shown in Figure S7 as reduced magnetization (M/Nµ B ) vs. H/T plot, where M is the magnetization, N is Avogadro's number, µ B is the Bohr magneton, and H is the magnetic field.
The M/Nµ B vs. H/T data were fit by assuming that only the ground state is populated and by including axial zerofield splitting (DŜ 2 z ) and isotropic Zeeman interactions. The corresponding spin Hamiltonian is given by Equation (5), where D is the axial ZFS parameter,Ŝ z is the easy-axis spin operator, µ 0 is the vacuum permeability, and H is the applied field. Equal quality fits, shown as the solid lines in Figure S7, were obtained for S = 25, 26, and 27 with parameters g = 2.03(1)/D = −0.007(1) cm −1 , g = 1.96(1)/D = −0.004(1) cm −1 , and g = 1.91(1)/D = −0.004(1) cm −1 , respectively. Based on the obtained fits, we conclude that 2·10 H 2 O has a ground state of S T = 26 ± 1, and a very small D value. Alternating current (ac) magnetic susceptibility data were collected for compounds 1-3 to obtain additional information about their S T values and the possibility to exhibit slow relaxation of the magnetization phenomena indicative of SMM behavior. The temperature dependence of the in-phase (χ ′ M ), shown as χ ′ M T and out-of-phase (χ ′′ M ) ac signals for 1-3 is shown in Figures S8-S10). These studies revealed that there are not any out-of-phase ac signals in 1 and 2 suggesting that these compounds are not new SMMs. In the case of 3, there is a barely visible beginning in the 1.8 K data of a frequency dependence in the in-phase plot and a concomitant very weak tail of an out-ofphase signal (χ ′′ M /χ ′ M ∼ 1.5% at 1.8 K) representing the beginning of a χ ′′ M signal whose peak is clearly far below the 1.8 K limit of our SQUID instrument, i.e., the anisotropy barrier is extremely small. In addition, ac data are in line with the conclusions obtained from dc studies concerning the S T values of 2 and 3. In particular, extrapolation of the χ M 'T signal of compounds 2 and 3 to 0 K from above ∼8 K (to avoid the effects of intermolecular interactions at lower temperatures) gives values of ∼340 and 255 cm 3 mol −1 K, respectively. These values are consistent with S T in the range 25-27 (S T = 25, g = 2.05; S T = 26, g = 1.97; S T = 27, g = 1.90) for 2 and 21-23 (S T = 21, g = 2.10; S T = 22, g = 2.01; S T = 23, g = 1.92) for 3.

DISCUSSION
We described herein the synthesis, crystal structures and magnetic properties of a series of high nuclearity mixed metal Mn/Ni and Mn/Co clusters. The first member of this family of complexes, compound 1, was prepared by targeted modifications in the reaction procedures that afforded the family of [Mn 44 ] and [Mn 40 Na 4 ] "loops-of-loops" aggregates reported by our group previously (Moushi et al., 2010a). Since complex 1 consists of a [Mn 36 Ni 4 ] 2− anion which co-crystallized together with a [Mn 2 Ni 6 ] 2+ cation, we targeted and achieved, following synthetic procedures containing elements of rational design, the isolation of the discrete [Mn 36 Ni 4 ] (compound 2) and [Mn 2 Ni 6 ] 2+ (the cation of 4) clusters and also of another Mn/3d analog the [Mn 32 Co 8 ] aggregate of 3. Arguably most of the known giant molecular aggregates have been afforded from serendipitous assembly synthetic procedures although recently there have been a few reports about elegant, rationally designed synthetic strategies that led to such high nuclearity clusters (Papatriantafyllopoulou et al., 2016 and references therein). The isolation of 2-4 thus, represents a rare example of targeted synthesis of high nuclearity metal-organic complexes from procedures that contain elements of rational design.
The heterometallic [Mn 36 Ni 4 ] 2−/0 and [Mn 32 Co 8 ] "loopsof-loops-and-supertetrahedra" molecular aggregates are giant clusters exhibiting nanosized dimensions and large molecular weights approaching those of small proteins (for example the MW of compound 1 is ∼8,095 g/mol and is comparable to those of small proteins which are ∼10 kDa). In fact these are the second highest nuclearity heterometallic Mn x M y (M = any metal ion) clusters and M x M' y (M, M ′ = any 3d metal ion) with the highest nuclearity heterometallic Mn-containing complex and mixed 3d metal/3d metal cluster being a [Mn 28 Cu 17 ] aggregate (Wang et al., 2007). Interestingly, the [Mn 36 Ni 4 ] 2−/0 and [Mn 32 Co 8 ] aggregates are based on polynuclear sub-units that have appeared either as fragments and/or in a discrete form previously (Stamatatos et al., 2006;Manoli et al., 2007Manoli et al., , 2008Manoli et al., , 2016Moushi et al., 2010a;Wu et al., 2011). This structural feature makes them a member of a very small family of large matalorganic clusters based on known in a discrete form polynuclear repeating units (Manoli et al., 2016;Papatriantafyllopoulou et al., 2016). Their [Mn III 6 Mn II 4 (µ 4 -O) 4 ] 18+ supertetrahedral repeating unit is very well-known in Mn cluster chemistry and has attracted significant interest due to its symmetric structure and the fact that most of the compounds containing this core exhibit entirely ferromagnetic exchange interactions and the maximum possible, for a [Mn III 6 Mn II 4 ] complex, S T = 22 spin ground state value. The presence of this repeating unit in the [Mn 36 Ni 4 ] 2−/0 and [Mn 32 Co 8 ] "loops-of-loops-and-supertetrahedra" aggregates ensures the appearance of dominant ferromagnetic exchange interactions between their metal ions and of a large S T value. Notably there are no Ni II ions located in this sub-unit in the [Mn 36 Ni 4 ] 2−/0 clusters since all of them are found in the decametallic loops. However, in the [Mn 32 Co 8 ] aggregate there are two Co II ions in each decametallic supertetrahedron located in the apex positions replacing two Mn II ions. In fact, there have been reported some heterometallic Mn 10−x M x (M = any metal ion) supertetrahedra where the Mn II ions are partially or completely replaced, however, to the best of our knowledge there are no mixed metal Mn 10−x Co x analogs in the literature. Such compounds could be very attractive magnetically since they could possibly combine the ferromagnetic exchange interactions and large S T values appearing in the decametallic Mn-based supertetrahedra with a significant anisotropy due to the presence of Co II ions. The appearance of the Co II ions in the apex positions of the supertetrahedra is not surprising, not only because they replace a metal ion being in the same oxidation state (Mn II ) and as a result there are no charge variations in the new compound but also these positions have proven to be the most labile in this family of [Mn III 6 Mn II 4 ] complexes. This is also supported from the isolation a series of heterometallic [[Mn III 6 Mn II x M 4−x ] (M = any metal ion) supertetrahedra, appearing in a discrete form and also as fragments of high nuclearity clusters, in which the Mn II ions have been completely or partially replaced by other metal ions (Skordi et al., 2018 and references therein). The [Mn 2 Ni 6 ] 2+ cation appears also for the first time in heterometallic cluster chemistry although there are homometallic [Mn II 6 Mn III 2 ] clusters reported exhibiting an analogous structural core (Boskovic et al., 2002 , although in the latter it is not safe to conclude for the S T value due to the presence of the anisotropic Co II ions. On the other hand, the [Mn 2 Ni 6 ] 2+ cation of 4 exhibits dominant antiferromagnetic exchange interactions leading to a diamagnetic S T value. In fact, a diamagnetic ground state has also been reported for the analogous homometallic [Mn II 6 Mn III 2 ] cluster (Boskovic et al., 2002). This behavior could result from the presence of antiferromagnetic exchange interactions between the two [Mn III Ni II 3 ] (in 4) or [Mn III Mn II 3 ] (in the [Mn II 6 Mn III 2 ]) cubane sub-units leading to diamagnetic ground states. The reported S T values for 2 and 3 are among the larger ones for heterometallic aggregates with the S T ≈ 26 being the second highest value reported for a heterometallic cluster (Chen et al., 2018). Clearly the overall magnetic behavior of 2 exhibits remarkable analogies with that of its [Mn III 6 Mn II 4 ] building block since they both display ferromagnetic exchange interactions, large S T and small D values and are not SMMs. On the other hand, in the case of 3 although it is not safe to conclude about the S T value, however, it is clear that the spin ground state value in 3 is smaller than that of 2. This can be attributed to the existence of more heterometal ions, since in 3 there are 8 Co II ions whereas in 2 only 4 Ni II ions and also to the presence of stronger antiferromagnetic exchange interactions as expected for Mn/Co and Mn/Ni heterometallic compounds. In addition, the out-of-phase ac signals at low T in 3 appear due to the presence of the anisotropic Co II ions leading to the increase of the anisotropy and possibly to SMM behavior.
Summarizing, a series of heterometallic [Mn 36 Ni 4 ] 2−/0 and [Mn 32 Co 8 ] "loops-of-loops-and-supertetrahedra" molecular aggregates and the cationic [Mn III 2 Ni II 6 ] 2+ cluster were prepared by employing synthetic procedures containing elements of rational design. The "loops-of-loops-and-supertetrahedra" molecular aggregates of 1-3 are among the largest heterometallic Mn-containing clusters exhibiting dimensions and molecular weights comparable to those of small proteins. In addition, they exhibit dominant ferromagnetic exchange interactions and very large S T values. These compounds are new additions in the very small family of giant Mn x M y (M = any metal ion) aggregates. Since this area is merely unexplored, further studies targeting to "loops-of-loops-and-supertetrahedra" analogs with various 3d and 4f metal ions and other high nuclearity heterometallic Mn/M clusters are in progress and the results will be reported in due course.

AUTHOR CONTRIBUTIONS
MC was involved on the synthesis, crystallization, and characterization of the reported complexes. EM was involved on the synthesis, crystallization, and characterization of the reported complexes and on manuscript preparation. TN was involved on the investigation of the magnetic properties of compounds 3 and 4. CP was involved on the investigation of the magnetic properties of compounds 1 and 2 and on manuscript preparation. VN was involved on the refinement of the crystal structures of 1-4 and on manuscript preparation. GC was involved on the investigation of the magnetic properties of compounds 1-2 and on manuscript preparation. AT supervised the reported work and was involved on all parts of the project.