Edinburgh Research Explorer Magneto-structural correlations in dirhenium(IV) complexes possessing magnetic pathways with even or odd numbers of atoms.

Employment of pyrazine (pyz), pyrimidine (pym) and s-triazine (triz) ligands in Re IV chemistry leads to the isolation of a family of complexes of general formula (NBu 4 ) 2 [(ReX 5 ) 2 ( μ -L)] (L = pyz, X = Cl ( 1 ) or Br ( 2 ); L = pym, X = Br ( 3 ); L = triz, X = Br ( 4 )). 1 - 4 are dinuclear compounds where two pentahalorhenium(IV) fragments are connected by bidentate pyz, pym and triz ligands. Variable-temperature magnetic measurements, in combination with detailed theoretical studies, uncover the underlying magneto-structural correlation whereby the nature of the exchange between the metal ions is dictated by the number of intervening atoms. That is, the spin-polarization mechanism present dictates that odd and even number of atoms favour ferromagnetic (F) and antiferromagnetic (AF) exchange interactions, respectively. Hence, while the pyz ligand in 1 and 2 mediates AF coupling, the pym and triz ligands in 3 and 4 promote F interactions.


Introduction
In 1963 McConnell proposed a mechanism to describe the magnetic exchange between radicals in aromatic organic molecules. 1 Antiferromagnetic and ferromagnetic interactions, with singlet and triplet ground spin states, were favoured when the magnetic pathway was made up of even or odd numbers of atoms, respectively. Such a mechanism therefore dictated that the more stable spin configuration described an alternation of spin densities on adjacent atoms on the pathway through dynamic spin polarization, i.e. an 'up-down-up-down' arrangement. This proposal created new research avenues in the field of molecular magnetism that aroused great interest over several decades. [2][3][4][5][6][7] However, doubts existed about its application to inorganic systems where the spin carriers were metal ions. For example, a dinuclear Ti III complex synthesized in the 1970s unexpectedly displayed moderate ferromagnetic exchange as a consequence of a spin polarisation mechanism of exchange. Although this was the very first example of an inorganic system exhibiting such behaviour, the result was largely ignored and remained unexploited for some time. 8,9 Two decades later, the first magnetic systems with an operative McConnell mechanism containing paramagnetic metal ions were reported; these were "dinuclear" systems where one paramagnetic centre was a metal ion and the other an organic radical. [10][11][12] In the early 1990s the first purely inorganic examples based on Mo III and Co II ions were reported. [13][14][15][16][17][18][19][20] In the latter case, the nature of the magnetic coupling was nicely tuned (ferromagnetic vs. antiferromagnetic) using diazine type ligands that provided -exchange pathways with even or odd number of atoms. To this point all known inorganic systems were based on metal ions possessing magnetic t2g orbitals interacting with a -pathway. Metal complexes with magnetic eg orbitals were not considered good candidates as they were assumed to promote spin delocalization. However, in 2001, this position was challenged when strong spin polarization between d 9 Cu II ions was observed, 21 opening the door toward further research focussing on the study of magnetic communication across long intermetallic distances, and the application of these concepts in, for example, spintronics. [22][23][24][25][26] In this paper, we have adapted this approach by employing pyrazine (pyz), pyrimidine (pym) and s-triazine (triz) ligands to examine the nature and magnitude of magnetic exchange between Re IV ions (5d 3 electronic configuration), leading to the first magneto-structural correlation of dinuclear [Re IV 2] molecules of general formula (NBu4)2[(ReX5)2(μ-L)] (L = pyz, X = Cl (1) or Br (2); L = pym, X = Br (3); L = s-triz, X = Br (4)).

Materials and methods
All chemicals were used as received. Syntheses were carried out under aerobic conditions.  31 and refined by full-matrix least-squares on F-squared using ShelXL, interfaced through Olex2. 32 All non-hydrogen atoms were refined anisotropically.
Hydrogen atom parameters were constrained. For full details see Table S1. CCDC 1557650-1557653.

Computational Details
In order to estimate the nature and magnitude of the intramolecular magnetic exchange interactions in 1-4, calculations were performed with the Gaussian09 package using the CAM-B3LYP functional (a long range corrected version of B3LYP) and the quadratic convergence approach. [33][34][35][36][37] Double-ζ and Los Alamos effective core potentials, as proposed by Hay and Wadt, were used for the Re IV , Cland Brions. [38][39][40] Ahlrichs double-ζ basis set was used for the remaining atoms. 41 Two-electron integrals and their derivatives were computed from Douglas-Kroll-Hess (DKH) 2nd order scalar relativistic calculations. 42,43 An approach based on the use of brokensymmetry (BS) functions built from localized orbitals was used to evaluate the energies of several spin states. 44 The BS functions, which provide positive or negative spin densities on the paramagnetic centres, were obtained from the guess functions generated with the fragment tool implemented in Gaussian09. Intermolecular magnetic couplings for the shortest contacts were calculated from the experimental structures. Parameters corresponding to the acetonitrile solvent were included to simulate the electronic effects of the surrounding molecules. 45 Calculations of the zero-field splitting (zfs) parameters were performed with version 3.0 of the ORCA program. 46 The TZVP basis set proposed by Ahlrichs, and tight SCF criteria were used in all cases. 41 Relativistic effects for the Re IV ions were introduced from a zero-order regular approximation (ZORA). 47 For complete active space (CAS) calculations, this auxiliary basis set was replaced by TZV/C. 48 (Table S3). identical coordination geometries, with the first coordination sphere of the Re IV ions similar to that described for 2 ( Figure 1c and Table S2). The values of the C-C and C-N bond lengths of the bridging pyrimidine ligand agree with those of previously published structures containing this ligand. 58 The two Re IV ions are disposed at an angle of approximately 118° with respect to the centre of the pyrimidine moiety, with the organic ligand positioned equidistant between the two perpendicular [ReCl4] planes. The intradimer Re···Re distance is ca. 6.1 Å, with the shortest intramolecular Br···Br contact being 4.006(2) Å (Br (12)···Br (18)) (Table S3). Numerous C-H···Br type interactions are present between the complex anions and organic cations in the crystal lattice, causing each [(ReBr5)2(μ-pym)] 2moiety to be well isolated from the neighbouring onesthe shortest intermolecular Br···Br contact being 4.707(2) Å ( Figure S3).
Complex 4 crystallises in the monoclinic space group I2/a. The asymmetric unit contains one NBu4 + cation and half a [(ReBr5)2(μ-triz)] 2anion, due to a two-fold rotation axis intersecting the C(2) and N(2) atoms on the s-triazine ligand. The Re IV ion sits in a coordination sphere similar to the one described for complex 2 (Figure 1d and Table S2). The triazine molecule contains C-C and C-N bond lengths which correlate well to previously published results describing this ligand bridging two metal ions. 59 Table 1. It is clear from this Table that the intermolecular exchange interactions are very small for 2 and 4 and negligible for 1 and 3. together with their corresponding calculated magnetic coupling constants (J).  Table 3. These models were built by (a) removing one of the Re IV ions together with its ligands,  (Table 3).

Compound d(X···X) / Å
With the theoretical study in mind, we modelled the experimental magnetic data with spin- Hamiltonian (1)   This approach causes no problem in 1 and 2 because their local zfs tensors are parallel or near parallel ( Figure 8). However, this is not the case in 3 and 4 where the z-axes of the local zfs tensors are almost perpendicular (Figure 8). In the ferromagnetic state, they therefore conform to an easy-magnetization plane, i.e., the D parameter takes a positive value (Figure 9). A more detailed/rigorous analysis of the magnetic data would demand the inclusion of the relative orientation of the local zfs tensors, but it would likely over-parameterize our model.