Magnetic properties and molecular structure of a binuclear alternative bridged Cu(II)Re(IV) complex containing a macrocyclic ligand

Polyhedron 75 (2014) 1–8

Contents lists available at ScienceDirect

Polyhedron journal homepage: www.elsevier.com/locate/poly

Magnetic properties and molecular structure of a binuclear alternative bridged Cu(II)Re(IV) complex containing a macrocyclic ligand Alina Bien´ko a,⇑, Rafał Kruszyn´ski b, Dariusz Bien´ko c a

Faculty of Chemistry, University of Wroclaw, 14 F. Joliot-Curie, PL-50 383 Wroclaw, Poland Department of X-ray Crystallography and Crystal Chemistry, Institute of General and Ecological Chemistry, Technical University Lodz, Zeromskiego 116, 90-924 Lodz, Poland c ´ skiego 27, 50-370 Wroclaw, Poland Faculty of Chemistry, Wroclaw University of Technology, Wybrzeze Wyspian b

a r t i c l e

i n f o

Article history: Received 20 December 2013 Accepted 25 February 2014 Available online 12 March 2014 Keywords: Building block Bimetallic complexes Alternating chain Exchange interaction Ferrimagnetic properties

a b s t r a c t Two novel macrocyclic compounds, the heterobimetallic complex {(CuLa)[ReCl4(ox)]}n (1) and the mononuclear complex [CuLa]2ClO4 (2) (where La = N-meso-5,12-Me2-7,14-Et2-[14]-4,11-dieneN4), have been synthesized and their crystal structures were determined by the single-crystal X-ray diffraction technique. Complex 1 crystallizes in the monoclinic space group P21/c, whereas 2 crystallizes in the monoclinic space group P21/n. The [CuLa]2+ macrocyclic cation in 1 is coordinated from above and below by [ReCl4(ox)]2 units through the chloro- and monodentate oxalato ligands, and this creates an alternating chloro-oxalato-bridged heterometallic ReIV–CuII one-dimensional zig-zag chain. Their magnetic measurements were carried out over the temperature range 1.8–300 K using a Quantum Design SQUID magnetometer (MPMSxL-5 type). Compound 1 behaves like a ferrimagnetic CuII–ReIV dimetallic chain with two intrachain antiferromagnetic coupling parameters and strong single-ion anisotropy, D(Re) = 109 cm1. Compound 2 shows weakly interacting paramagnetic centers in the crystal lattice. The effects of hydrogen bonds mediating the magnetic exchange interactions on the spin density have been evidenced by DFT calculations. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction In the past decade, low-dimensional magnetism of a new kind of magnetic subsystem, namely ferrimagnetic chains, has been discovered [1]. A ferrimagnetic 1-D system exhibits original physical properties, characterized by a minimum of the magnetic susceptibility at an intermediate temperature, which depends on the strength of the exchange interactions and magnetic anisotropy, and the divergence of susceptibility at lower temperatures [2]. In this respect, studies on systems involving 4d or 5d metal ions become very interesting. Greater orbital diffuseness and spin–orbit coupling may affect the nature and magnitude of the interaction between magnetic centers [3]. One example of such an ion is Re(IV). Rhenium(IV), as a 5d3 ion, usually forms octahedral compounds; numerous types of hexachloro-, hexabromo- and hexaiodorhenate complexes have been deeply studied previously [4–6]. The substitution of two chloro-ligands for an oxalate-group in the rhenium(IV) coordination sphere allows the synthesis of new heterodinuclear [M(II)–Re(IV)] (MII = Cu, Mn) [3,7,8] complexes. Because of the low affinity of [ReCl4(ox)]2 for other ⇑ Corresponding author. Tel.: +48 713757258; fax: +48 713757307. E-mail address: [email protected] (A. Bien´ko). http://dx.doi.org/10.1016/j.poly.2014.02.045 0277-5387/Ó 2014 Elsevier Ltd. All rights reserved.

divalent metal ions of the first transition series, most of these complexes concern Cu(II)–Re(IV) ions. The coordination mode of the bridging oxalate ligand from the Re(IV) species depends on the nature of the other ligands in the coordination sphere of Cu(II), e.g. [ReCl4(ox)Cu(bipy)2] [3] and [ReCl4(ox)Cu(terpy)(CH3CN)] [7]. However no significant magnetic interaction was observed in these bimetallic oxalate bridged Cu(II)–Re(IV) complexes; as the orthogonality between the dz2 magnetic orbital of Cu(II) and dxy of Re(IV) is broken, the overlap is predicted to be very small, resulting in a weak antiferromagnetic coupling between the metal ions. Analysis of literature data suggests that a quite strong magnetic coupling should be obtained if the ligand is coordinated in a symmetric bidentate way, with two short Cu(II)–O bonds. In this respect, it seemed interesting to examine macrocyclic copper(II) complexes in order to obtain heterometallic Cu(II)Re(IV) systems. Square planar macrocyclic complexes with two vacant coordination sites at the metal atom play the role of the so-called ‘‘building block’’. This situation changes with the size and character of the substituents. The position of the substituents around the equatorial nitrogen atoms of the ligands and the N-configuration of the macrocyclic ring influences the coordination mode of polymetallic complexes, for instance, a simple dimeric form (CuL)[Co(NCS)4] [9], trimeric unit [NiL]2[Mn(NCS)4](ClO4)2H2O [10] or chain structure

´ ko et al. / Polyhedron 75 (2014) 1–8 A. Bien

2

[CuLb][ReCl4(ox)]DMF [11] and [CuCuL][ReCl4(ox)]22DMF [12]. Given the well known plastic coordination sphere of copper(II) macrocycles, it seemed interesting to investigate the influence of other substituents of the macrocycle on the coordination modes of [ReCl4(ox)]2. In our previous paper, we substituted six methyl radicals in a copper(II) macrocycle (like in the [CuLb][ReCl4(ox)]DMF complex) by two methyl and two ethyl substituents [Scheme 1]. As a result, we obtained a heterobimetallic Re(IV)–Cu(II) chain, {(CuLa)[ReCl4(ox)]}n, with the first intramolecular chloride bridge and a much stronger ferrimagnetic interaction [13] than in an earlier reported complex with intrachain oxalateor intermolecular chloride ligands. However, when we changed the solvent during the reaction (from CH3CN to a mixture of CH3NO2 and CH3CN in a ratio of 1:2) and the conditions of the synthesis: temperature (heating under reflux) and time (the ratio of the reactants are the same), we obtained a new alternating chain, in which the two paramagnetic centers, Cu(II) and Re(IV), are connected by oxalate and chloride ligands. The synthesis, crystal structures and variable temperature magnetic properties are reported here. 2. Experimental 2.1. General comments Caution! Perchlorate salts of metal complexes with organic ligands are potentially explosive. Only a small amount of these materials should be prepared and they should be handled with great caution. All materials used in this work were of reagent grade purity and were used as commercially obtained. 2.2. Synthesis of the complexes 2.2.1. {(CuLa)[ReCl4(ox)]}n (1) A solution of [CuLa](ClO4)2 (81.44 mg, 0.15 mmol) in hot CH3NO2 (15 cm3) was added dropwise to a solution containing [Bu4N]2[ReCl4(ox)] (135.14 mg, 0.15 mmol) in CH3CN (30 cm3). The mixture was refluxed for 1 h. After a few days, a dark-violet crystalline solid was formed by slow evaporation. It was filtered and washed using diethyl ether (10 cm3, twice). Yield: 91.20 mg (80 %). Anal. Calc. for C18H32Cl4CuN4O4Re (760.02): C, 28.41; H, 4.49; N, 7.35; Cl, 18.61. Found: C, 28.40; H, 4.58; N, 7.19; Cl, 18.56%. The copper and rhenium content was determined by the ICP method, calcd.: Re, 24.40; Cu, 8.34; found: Re, 24.22; Cu, 8.40%. IR/cm1: bands associated with the oxalato ligand appear at 1715vs, 1700vs 1689m, 1357s, 803m.

Scheme 1. Schematic view of macrocyclic copper(II) complex [CuLu] (ClO4)2.

2.2.2. [CuLa(ClO4)2] (2) N-Meso-5,12-dimethyl-7,14-diethyl-1,4,8,11-tetraazacyclotetradecane-4,11-diene dihydrogenperchlorate (111.2 mg, 0.68 mmol) was slowly added to a boiling solution of copper acetate monohydrate (139.7 mg, 0.7 mmol) in methanol (40 cm3). The mixture was refluxed for 2 h. The resulting solid was dissolved by addition of water (8 cm3) and the solution was filtered and left for crystallization. The violet crystals of [CuLa(ClO4)2] that formed were filtered off and dried under reduced pressure. Yield: 246 mg (65%). Anal. Calcd. for C16H32Cl2CuN4O8 (542.9): C, 35.4; H, 5.89; N, 10.3; Cl, 13.1. Found: C, 35.6; H, 5.78; N, 10.4; Cl, 13.25%. The copper content was determined by the ICP method, calcd.: Cu, 11.7; found: Cu, 11.5%. 2.3. Physical measurements IR spectra (400–4000 cm1) were recorded on an FT-IR spectrometer (Spectrum One, Perkin Elmer) using KBr pellets. 2.3.1. Magnetic measurements Magnetic measurements were performed using a Quantum Design SQUID-based MPMSXL-5-type magnetometer in the temperature range T = 1.8–300 K. The SQUID magnetometer was calibrated with a palladium rod sample (Materials Research Corporation, measured purity 99.9985%). Measurements were made at a magnetic field of B = 0.5 T. Corrections based on subtracting the sample-holder signal were made and a contribution to the underlying diamagnetism was estimated from Pascal constants [14]. 2.3.2. X-ray crystallographic study The violet needle crystal of compound 1 and violet prism crystal of compound 2 were mounted in turn on a KM-4-CCD automatic diffractometer equipped with a CCD detector, and used for data collection. X-ray intensity data were collected with graphite monochromated Mo Ka radiation (k = 0.71073 Å) at temperature of 291.0(3) K, with the x scan mode. Exposure times of 28 s were used for both measurements and reflections inside the Ewald sphere were collected up to 2h = 50°. The unit cell parameters were determined from 12 354 and 9583 strongest reflections, respectively, for compounds 1 and 2. Details concerning crystal data and refinement are given in Table 1. Examination of reflections on two reference frames, monitored after each 20 frames measured, showed no loss of intensity for both compounds. During the data reductions, Lorentz, polarization and numerical absorption [15] corrections were applied. The structures were solved by a partial structure expansion procedure. All the non-hydrogen atoms were refined anisotropically using the full-matrix, leastsquares technique on F2. All the hydrogen atoms were found from difference Fourier synthesis after four cycles of anisotropic refinement, and were refined as ‘‘riding’’ on the adjacent atom with geometric idealisation after each cycle of refinement, with individual isotropic displacement factors equal to 1.2 times the value of the equivalent displacement factor of the parent non-methyl carbon or nitrogen atom or 1.5 times for a parent methyl group carbon atom. The methyl groups were allowed to rotate about their local threefold axis. The SHELXS97 [16], SHELXL97 [17] and SHELXTL [18] programs were used for all the calculations. Atomic scattering factors were those incorporated in the computer programs. Selected interatomic bond distances and angles are listed in Table 2. 2.3.3. Theoretical calculations The spin densities were performed by the stand-alone Mulliken method, implemented in the GAUSSIAN 09 packages. The complex structure was fully optimized by density functional theory DFT (B3LYP/LanL2DZ). A reliable prediction of the molecular structures of transition metal complexes can be obtained by using a combined

´ ko et al. / Polyhedron 75 (2014) 1–8 A. Bien Table 1 Crystal data and structure refinement.

Table 2 Selected structural data for compounds 1 and 2 (Å, °).

Compound

1

2

Empirical formula Formula weight T (K) k (Å) Crystal system Space group Unit cell dimensions a (Å) b (Å) c (Å) b (°) V (Å3) Z Dcalc (Mg/m3) Absorption coefficient (mm1) F(0 0 0) Crystal size (mm) h Range for data collection (°) Index ranges

C18H32Cl4CuN4O4Re 760.02 291.0(3) Mo Ka = 0.71073 monoclinic P21/c

C16H32Cl2CuN4O8 542.90 291.0(3) Mo Ka = 0.71073 monoclinic P21/n

13.296(11) 10.017(7) 20.915(16) 92.58(5) 2783(4) 4 1.814 5.524 1488 0.173  0.012  0.009 1.53–25.01 15 6 h 6 15, 11 6 k 6 11, 24 6 1 6 24 35243/4899 (0.0476)

7.7206(7) 9.3663(10) 16.2151(15) 99.819(9) 1155.39(19) 2 1.561 1.225 566 0.154  0.142  0.091 2.55–25.04 9 6 h 6 9, 11 6 k 6 11, 19 6 1 6 19 15622/2044 (0.0150)

99.9 full-matrix leastsquares on F2 0.926 and 0.949

99.9 full-matrix leastsquares on F2 0.830 and 0.897

Reflections collected/unique (Rint) Completeness to 2h = 50° (%) Refinement method Minimum and maximum transmission Data/restraints/parameters Goodness-of-fit (GOF) on F2 Final R indices [I > 2r(I)]

4899/0/293 1.244 R1 = 0.0506, wR2 = 0.1332 R indices (all data) R1 = 0.0561, wR2 = 0.1353 Largest difference in peak and 2.602 and 1.977 hole (e Å3)

3

2044/0/144 1.095 R1 = 0.0290, wR2 = 0.0792 R1 = 0.0294, wR2 = 0.0795 0.397 and 0.366

basis set consisting of the polarized valence double-j basis set, D95V(d,p) for all ligand atoms in conjunction with the LanL2DZ effective core potential and valence basis set for the metal atom. All calculations in this work have been performed with the combined D95V(d,p) and LanL2DZ basis sets. 3. Results and discussion 3.1. Structural description Perspective views of compounds 1 and 2 are shown in Figs. 1 and 2. All atoms of 1 lie in general positions. The copper atom of compound 2 occupies the special position b of the P21/n space group at the 0, 0, 1/2 site with 1 symmetry and a multiplicity of 2. Thus the molecule of 2 occupies two asymmetric units. The all central atoms in both compounds are six coordinated, the copper by four macrocyclic nitrogen atoms (equatorial plane), and in the case of compound 1 by one axial oxygen atom of the butanedioate ion and one axial chloride ion. In compound 2 the axial positions are occupied by oxygen atoms of perchlorate ions. The rhenium is coordinated by four chloride ions and two oxygen atoms of a chelating butanedioate ion, thus the butanedioate ion acts as tridentate bridging-chelating ion. The Cu–O distances (Table 2) are longer than the typical Cu–O coordination bond length of 1.951 Å, but it lies in the range 2.3–2.8 Å, typical for complexes with Jahn–Teller distortion (the mean axial Cu–O bond length for Jahn–Teller distorted complexes is 2.483 Å). The same can be stated for the Cu1–Cl3#(x, y + 1, z) bond, which typically adopts a length of 2.252 Å, but in Jahn–Teller distorted complexes the length of Cu–Cl bonds fall in range of 2.5–3.1 Å The coordination

1 Re1–O3 Re1–O1 Re1–Cl4 Re1–Cl2 Re1–Cl1 Re1–Cl3 O1–C17 C17–O2 C17–C18 O3–C18 C18–O4 Cu1–N2 Cu1–N4 Cu1–N1 Cu1–N3 Cu1–O2 Cu1–Cl3#1

2.079(7) 2.087(6) 2.368(3) 2.375(3) 2.387(3) 2.414(3) 1.308(10) 1.246(10) 1.572(13) 1.325(11) 1.232(11) 2.025(7) 2.033(7) 2.043(7) 2.056(7) 2.581(6) 2.997(3)

O3–Re1–O1 O3–Re1–Cl4 O1–Re1–Cl4 O3–Re1–Cl2 O1–Re1–Cl2 Cl4–Re1–Cl2 O3–Re1–Cl1 O1–Re1–Cl1 Cl4–Re1–Cl1 Cl2–Re1–Cl1 O3–Re1–Cl3 O1–Re1–Cl3 Cl4–Re1–Cl3 Cl2–Re1–Cl3 Cl1–Re1–Cl3 N2–Cu1–N4 N2–Cu1–N1 N4–Cu1–N1 N2–Cu1–N3 N4–Cu1–N3 N1–Cu1–N3 O2–Cu1–N1 O2–Cu1–N2 O2–Cu1–N3 O2–Cu1–N4 Cl3#1–Cu1–N1 Cl3#1–Cu1–N2 Cl3#1–Cu1–N3 Cl3#1–Cu1–N4

79.7(2) 92.0(2) 171.62(18) 173.3(2) 93.91(18) 94.40(11) 89.2(2) 88.5(2) 92.20(12) 92.47(12) 87.7(2) 86.4(2) 92.50(10) 90.09(10) 174.46(9) 176.4(3) 93.1(3) 86.1(3) 86.0(3) 94.9(3) 178.2(3) 88.7(2) 93.1(2) 92.9(2) 83.4(2) 93.4(2) 101.0 (2) 85.25(19) 82.5(2)

N1–C1–C2–C3 C1–C2–C3–N2 C2–C3–N2–C4 C3–N2–C4–C5 N2–C4–C5–N3 C4–C5–N3–C6 C5–N3–C6–C7 N3–C6–C7–C8 C6–C7–C8–N4 C7–C8–N4–C9 C8–N4–C9–C10 N4–C9–C10–N1 C9–C10–N1–C1 C10–N1–C1–C3

60.3(11) 24.6(13) 176.3(8) 163.4(8) 46.4(9) 172.6(7) 178.7(7) 73.7(10) 40.5(12) 180.0(9) 145.3(8) 51.0(10) 166.4(8) 162.5(6)

2 Cu1–N2 Cu1–N1 Cu1–O1

1.9912(15) 2.0127(16) 2.6013(19)

O1–Cu1–N1 O1–Cu1–N2 O1–Cu1–O1#2 N2–Cu1–N2#2 N2–Cu1–N1 N2–Cu1–N1#2 N1–Cu1–N1#2

88.50(7) 88.98(7) 180.00 180.000(1) 94.68(6) 85.32(6) 180.000

Symmetry transformations used to generate equivalent atoms: #1: x, y+1, z; #2: x, y+2, z+1

4

´ ko et al. / Polyhedron 75 (2014) 1–8 A. Bien

Fig. 1. The molecular conformation of compound 1 with the atom numbering, plotted with 50% probability displacement ellipsoids. The symmetry generated atoms and bonds are indicated by dashed lines (primed atoms are obtained by the x, y + 1, z symmetry transformation and double-primed by x, y  1, z). Hydrogen atoms are omitted for clarity.

Fig. 3. The coordination polymer chain of 1. Hydrogen atoms are omitted for clarity.

ap, ac, sc, ap+, ap+, sc+, sp, ap+, ac+, sc+, ap, ap, respectively for compounds 1 and 2. This proves significant changes are induced by different coordination environments. The polymer chain of 1 is stabilised by two intramolecular hydrogen bonds (Table 3), and these interactions form N1S(6) and N1S(7) motifs [19] respectively for the N–H  Cl and N–H  O bonds. The molecules of 2 are conformationally stabilised by N–H  O intramolecular bonds, forming N1S(6) motifs. In the structure of 2 intermolecular C–H  O short contacts can be found, which can be classified as weak hydrogen bonds, and they form C(6)[R22(12)] and C22(18)[R44(36)] motifs, respectively for the interactions: C2–H2B  O1#3 and C7– H7C  O3#4 (symmetry codes as in Table 3). Via these interactions, a three dimensional hydrogen bonded network is formed [Fig.4]. The polymer nearest Cu  Re (via butanedioate) and Re  Cu (via chloride) distances are respectively 6.029(4) and 4.769(3) Å. In 1 the nearest Cu  Cu and Re  Re distances also extend along the polymer chain and both are 10.017(6) Å. In 2 the shortest Cu  Cu distance is 7.7206(7) Å, propagating along the crystallographic [100] axis.

3.2. Magnetic properties Fig. 2. The molecular conformation of compound 1 with the atom numbering, plotted with 50% probability displacement ellipsoids. The symmetry generated atoms of the molecular compounds are indicated by the letter A and they are obtained by the x, y + 2, z + 1 symmetry transformation. Hydrogen atoms are omitted for clarity.

polyhedra of all the metal atoms can be described as distorted tetragonal bipyramid. The copper atom lies in the range of experimental error (compound 1) or by symmetry (compound 2) in the N4 least squares plane of the macrocycle. The coordination polymer of 1 (Fig. 3) is created via a (–Cu–butanedioate–Re–Cl–)n linkage, in opposition to the previously reported compound with the same composition: catena-(tetrakis(l2-chloro)-tetrachlorobis(oxalato)-bis(N-meso-5,12-dimethyl-7,14-diethyl-1,4,8,11-tetraazacyclotetradecane-4,11-diene)-di-copper(II)–di-rhenium(IV)) [13], in which the polymer chain was created by (–Cu–Cl–Re–Cl–)n units. The polymer chain of 1 extends along the crystallographic b axis and equivalent molecules are generated via x, y + 1, z translations. The macrocycle conformation can be described as sc, sp+, ap, ap, sc, ap+, ap, sc+, sc, ap+, ac+, sc+, ap, ap and sc, sp+,

In this section, we will present the magnetic properties of compounds 1 and 2, together with an analysis of the magnetic exchange interaction. Finally, the analysis of the exchange pathways in 1 and 2 will be described through spin density calculations using the DFT methodology. The magnetic functions for the heteronuclear {(CuLa)[ReCl4 (ox)]}n (1) complex are displayed in Fig. 5. At room temperature, vMT is equal to 2.27 cm3 mol1 K, a value which is expected for Table 3 Hydrogen bonds for compounds 1 and 2 (Å, °). D—H  A

d(D–H)

d(H  A)

d(D  A)

<(DHA)

1 N1—H1N  O4 N3—H3N  Cl2#1

0.92 0.92

2.18 2.79

3.064(10) 3.676(7)

159.9 162.7

2 N1—H1N  O2#2 C2—H2B  O1#3 C7—H7C  O3#4

0.87 0.97 0.96

2.34 2.60 2.53

3.142(3) 3.538(3) 3.477(4)

153.7 163.2 169.2

Symmetry transformations used to generate equivalent atoms: #1 x, y+1, z; #2 x, y+2, z+1; #3x+1, y+2, z+1; #4 1/2+x, 5/2y, 1/2+z

´ ko et al. / Polyhedron 75 (2014) 1–8 A. Bien

5

At the first sight, the efficiency of the intra-chain exchange pathways in complex 1 to mediate magnetic interactions is quite surprising. The literature data [3,7,8,11–13] report detailed behaviour of heteropolynuclear complexes containing the [ReCl4(ox)]2 ion as a building block. These compounds exhibit different coordination modes depending on the M(II) 3d metal ion and the auxiliary ligand used to complete its coordination sphere: isolated dinuclear units in [ReCl4(l-ox)Mn(dmphen)2]CH3CN [8] and infinite chains in [CuLb][ReCl4(ox)] DMF (Lb = N-dl-5,7,7,12,14, 14-hexamethyl-1,4,8,11-tetraazacyclotetradeca-4,11-diene) [11]. However in all of these cases, the Re(IV) ion and M(II) metal ion are bridged by the oxalato ligand, being monodentate in [ReCl4 (ox)Cu(bipy)2] [3] and [ReCl4(ox)Cu(terpy)(CH3CN)] [7], asymmetric bidentate with two long Cu(II)–O bonds in [ReCl4(ox)Cu(phen)2] [7] and asymmetric bidentate with one short and one long Cu(II)–O bonds in [ReCl4(ox)Cu(terpy)(H2O)] [7]. In our previous paper, we presented the first example of a chloro-bridged Cu(II)Re(IV) chain complex, {(CuLa)[ReCl4(ox)]}n (where La = N-meso-5,12-Me2-7,14Et2-[14]-4,11-dieneN4) [13]. Using the same copper(II) macrocycle but changing the conditions of the synthesis, we obtained a chain of the same formula but with a different kind of coordination mode as a result of changing the macrocycle conformation. According to the crystal structure, this complex can be viewed as an alternating bimetallic chain with two intrachain magnetic coupling parameters J – through the chloride ion Cl(3) and j – through the oxalate oxygen atom (O1). On the other hand, the variation of vMT with T in the low temperature range could be due to the high value of the zero-field splitting of the 4A2 ground term of the Re(IV) ion, expected for two different types of ligands in the rhenium coordination sphere. Taking into account the 1D chain structure of 1, we have analyzed the magnetic behavior using an Ising spin-1/2 model with an alternating g factor through the Hamiltonian of Eq. (1) [3]

Fig. 4. A part of molecular packing of compound 2.

an uncoupled ReIV (SRe = 3/2) and CuII (SCu = 1/2) pair. With decreasing temperature, vMT continuously goes down in value and reaches its minimum (see inset) at 20.5 K, with vMT = 1.57 cm3 mol1 K, corresponding to a short-range order state where the two uncoupled spins of the adjacent ions {Cu(II)–Re(IV)} exist. When the temperature is lowered below 20.0 K, it is observed that vMT increases and reaches a maximum at 3.3 K and then it rapidly decreases from 3 to 1.80 K. For vMlT values below the minimum, a ferrimagnetic ordering [20] in the chain exists. At very low temperature, an additional exchange interaction between the spins of the nearest neighboring chains causes a transition to a three-dimensional magnetic system. The magnetization per formula unit M1 = Mmol/ (NAlB) at B = 5 T and T = 2.0 K tends to reach saturation with a value of Msat = 2.23. In such a case, the ground state equals Smax = 1/2 + 3/2 = 2 and the magnetization (per {Cu–Re} unit) should saturate to a value of Msat = 4.0 lB. The obtained value is substantially smaller and it evidences a sizable zero-field splitting for Re(IV), along with an antiferromagnetic exchange coupling between the Re(IV)  Cu(II) ions.

H¼

8 z z z z z z X< JS2i1 S2i  jS2i S2iþ1 þ g IICu bS2i1 Hz þ g IIRe bS2i Hz þ i

h

i

9 =

: g ?Cu bðSx2i1 Hx þ Sy2i1 Hy Þ þ g ?Re bðSx2i Hx þ Sy2i Hy Þ þ 2D ðSz2i Þ2  54 ; ð1Þ

where the total spins are S2i = SRe and S2i-1 = SCu, J is the exchange coupling parameter between the ReIV and CuII magnetic centers

3,0 0,7 2,5

0,6

-1

[cm mol ]

3

-1

[cm mol ]

0,4

3

corr

χM

corr

0,5

0,2

0,3

1,0

0,2

-1

χM

0,6

3

1,5

0,7

0,3

T [cm mol K]

0,4

corr

2,0

χM

0,5

0,1 0,0 0

5

10

15

20

25

30

T [K]

0,5

0,1 0,0

0,0 0

50

100

150

200

250

300

T [K] Fig. 5. Thermal dependence of vM for: (d) {(CuLa)[ReCl4(ox)]}n. and vMT for (o) {(CuLa)[ReCl4(ox)]}n, complexes. The inset show: left – low temperature region magnetic susceptibility, right – field dependence of the magnetization per formula unit at 2 K. The solid line is the calculated curve.

´ ko et al. / Polyhedron 75 (2014) 1–8 A. Bien

6

by means of the chlorine atom, j is the exchange coupling parameter through the oxalate–oxygen atom, 2D is the energy gap between the ±3/2 and ±1/2 Kramers doublets of the ReIV center. The values of the exchange parameters were determined by a least-square procedure and the minimization of the function R was the criterion used to determine the best fit.

R¼

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi i uP h u n ðv TÞexp  ðv TÞcalc2 =ðv TÞexp2 t i1 M M M i i i n

The least squares fit of the experimental data by this expression was limited to the range 20.0–300 K and leads to the following results: 2D = 109 cm1, gavRe = 1.81, gavCu = 2.24, J = 14.2 cm1, j = 8.72 cm1. The calculated curve matches very well with the magnetic data. We conclude that the Re(IV) ion in 1 exhibits a large zfc together with antiferromagnetic interactions with Cu(II). Let us first analyze and discuss the values of the best-fit parameters concerning complex 1. A possible mechanism for the onedimensional ordering in this complex can be understood through orbital symmetry considerations. The magnitude of the interaction is due to overlap of the magnetic orbital on either side of the bridge. The tetrahedral distortion of the copper environment is the apparent cause for the four short equatorial bonds to N(4), N(8), N(11) and N(15), which ensure the delocalization of the unpaired electron of the dz2 magnetic orbital within the equatorial plane, and some spin density occurring at the axial chloro and oxalate–oxygen atoms results from the disruption of the orthogonality between this orbital and the dxy orbital of Re(IV). The asymmetric coordination and bite angle of the oxalate and chloro ligands within the CuIIReIV pair causes some overlap with

the copper orbital, resulting in an antiferromagnetic coupling (J = 14.2 cm1 and j = 8.72 cm1). The j value is smaller than J because of the very long Re(1)–O(1) and Cu(1)–O(2) axial distances (2.087 and 2.581 Å, respectively) and smaller spin density. We would like to point out here that this antiferromagnetic interaction between ReIV and CuII through the chloro- and oxalate bridges is much stronger than that observed in the chain compound [Cu(bipy)2(ox)ReCl4] (J = 8.3 cm1 j = 4.3 cm1) [3]. On the other , hand, the net antiferromagnetic coupling is weaker than the compounds described earlier with only a chloride bridge (J = 18.1 cm1) due to shorter Cu(1)–Cl(1) and Cu(1)–Cl(3) bond lengths (2.976 and 2.991 Å, respectively). To confirm this hypothesis, DFT calculations were carried out. The computed spin densities are given in Table 4 (the atom numbering of the real structure of 1 has been kept). The DFT calculations show an important spin density on the bridging atoms. According to McConnell’s theory [21], the larger the spin densities

Table 4 DFT-calculation atomic spin densities for 1. Atom

Re(1) O(1) O(2) O(3) O(4) Cu(1)

Spin density

Atom

1

CuII–Cl–ReIV

2.4999 0.0018 0.0208 0.0134 0.0271 0.5480

3.6399 0.0185 0.0357 0.0185 0.0357 0.3171

Cl(1) Cl(2) Cl(3) Cl(4) C(1) C(2)

Spin density 1

CuII–Cl–ReIV

0.0359 0.1337 0.1046 0.1485 0.0072 0.0058

0.07254 0.0426 0.0725 0.0256 0.0082 0.0082

Fig. 7. EPR spectrum of powdered complex 2 at the X-band at room temperature together with the spectrum simulated with S = 1/2 spin Hamiltonian parameters.

Fig. 6. Thermal dependence of vM (d) and vMT (s) for [CuL2(ClO4)2] . The inset shows the field dependence of the magnetization per formula unit at 2 K. The solid line is the calculated curve.

´ ko et al. / Polyhedron 75 (2014) 1–8 A. Bien Table 5 Theoretical atomic spin populations calculated by

GAUSSIAN

7

for one isolated molecule, for (1) and (2), and for interacting molecules.

Atoms

One isolated molecule

Two interacting molecules

Atoms

One isolated molecule

Two interacting molecules

Cu N(1) H(1N) N(2A) O(1)

0.5736 0.1132 0.0032 0.1072 0.0002

0.5726 0.1101 0.0032 0.1032 0.0001

N(2) H(9A) C(9) N(1A) Cl(1)

0.1063 0.0003 0.0069 0.1114

0.1124 0.0002 0.0044 0.1120

at the bridging atoms are, the greater is the exchange coupling. Some examples of this are the oxalate-bridged dinuclear copper(II) complexes [22], for which a linear correlation between the atomic spin density and the exchange coupling was found. It deserves to be noted that the net spin density on a particular atomic orbital or atom is the result of two mechanisms: spin delocalization and spin polarization. As we can see in Table 4, there are positive and negative values for the spin densities in the examined compound, which is in contrast to the previously reported complex. Moreover, the value of the spin density is smaller than on the rhenium atom in the CuII(l2-chloro)ReIV chain. These results show that the spin polarization mechanism is dominant in the case of 1, whereas the spin delocalization is more important in the case of the complex with the chloro bridge. According to these results, we can conclude that in complex 1 there are two magnetic exchange pathways: the first one (an intrachain magnetic exchange interaction between the Re(IV) and Cu(II) paramagnetic centers through the chloro bridge) is much greater than the other (an intrachain magnetic exchange interaction through the oxalato ligand). The copper to chloride distance is longer than the copper to oxygen distance (from the oxalate ion), but the atomic spin density is much more important for the chloro atom. The product function vMT for complex 2 (and/or the effective magnetic moment) is nearly constant on cooling from room temperature down to T = 10 K: (leff)300 K = 1.83 lB, (leff)10 K = 1.80 lB (Fig.6). Below 10 K, a rapid drop of the vMT value is observed until (vMT)1.8 K = 0.36 [cm3 mol1 K] ((leff)2 K = 1.70 lB). This feature indicates that an exchange interaction of an antiferromagnetic nature occurs. In such a case, however, the magnetic susceptibility on cooling should pass through a maximum. As such a maximum is not seen until T = 1.8 K, the exchange coupling constant zJ0 is rather small and appears between the copper atoms in the crystal lattice. The magnetization per formula unit M1 = at B = 5 T and T = 2.0 K tends to reach the spin-only limit of M1 = 1 for the total spin S = ½. The experimental data have been fitted to the expression [23]:

vM ¼

Nb2 g 2 SðS þ 1Þ 3kT

where S = 1/2.The susceptibility data were eventually corrected for the molecular-field correction (zj) and the temperature-independent magnetism vTIM, i.e. vcorr = vmol/[1  (zj)vmol] + vTIM. The best fit parameters are g = 2.11 and zJ0 = 0.18 cm1, obtained with a good agreement factor, R = 1.72  108. The EPR spectrum of compound 2, examined at a room temperature, shows only a single line at H = 3500G (see Fig. 7). The spectroscopic splitting factor is typical for copper(II) centers, g is equal to 2.104 and has a good agreement with the theoretical value obtained from magnetic calculations. The values of the Curie and Weiss constants, determined from the relation 1/vMT = f(T) over the temperature range 50–300 K, are equal to 0.71 cm3 mol1 and 0.36 K, respectively. A negative value of the Weiss constant and intermolecular exchange parameter obtained from the calculations confirm the occurrence of a weak antiferromagnetic interaction between the nearest copper centers in complex 2. These observed interactions may be transmitted through hydrogen bonds in the crystal lattice. As mentioned in the structural discussion, complex 2 can be viewed as a pseudo

chain, created by a hydrogen bond network formed by C–H  O short contacts. These contacts may create a magnetic exchange pathway. The small magnitude of this interaction is a result of the long Cu–Cu separation (7.7206(7) Å). To confirm this view, a DFT calculation was carried out for the geometry of the crystal packing of one isolated molecule and two interacting molecules. The spin populations, obtained with the GAUSSIAN program, are reported in Table 5. Almost 60% of the spin is localized on the central Cu atoms, the rest being diluted on different atoms in the other parts of the molecule. In particular, the presence of a significant spin density on N(1) (0.1132), H(1N) (0.0032) and O(1) (0.0002) suggests that along the b axis of the monoclinic cell (2.34 Å for N1–H1N  O2) (2), the contacts between adjacent molecules propagate the antiferromagnetic interaction. However, when both molecules interact through the contacts explicated above, the spin density on the hydrogen atom H(1N) is not modified, but the nitrogen atom N(1), which is attached to it, now has a spin population which decreases by 0.0031. Furthermore, the negative spin on the oxygen atom O(1) changes from 0.0002 to 0.0001. From this comparison, we cannot conclude anything about the strength of the exchange interactions, but we confirm the magnetic interactions which are propagated through such contacts. 4. Conclusion The main conclusions from this work are: (1) it is possible to use a copper(II) macrocyclic unit as a ‘‘building block’’ to synthesize a heterometallic complex with a high anisotropy effect with the [ReCl4(ox)]2 precursor; (2) The coordination mode of the [ReCl4(ox)]2 ligand changes with the number and character of the substituents of the macrocyclic ring, the positions of the substituents around the equatorial nitrogen atoms of the ligands and also the conversion of the macrocyclic ring conformation, taking the form of a chain structure with a monodentate oxalate bridge in [CuLb][ReCl4(ox)]DMF) (Lb = N-dl-5,7,7,12,14,14-hexamethyl-1,4,8,11-tetraazacyclotetradeca-4,11-diene), through the chloride ligand in {(CuLa)[ReCl4(ox)]}n (La = N-meso-5,12Me2–7,14-Et2-[14]-4,11-dieneN4) or leading to the first bimetallic alternating chain in 1 with chloro and oxalate bridges; a dimeric structure in (CuLb)[ReCl4(ox)] (Lb = N-rac-5,12-Me2-7,14-Et2-[14]4,11-dieneN4) [11], trimeric units in [(CuLa)2Cl] [ReCl4(ox)]Cl (La = N-dl-5,7,7,12,12,14-hexamethyl-1,4,8,11-tetraazacyclotetradeca-4,14-diene) or a tetrameric form like in [CuL][ReCl4ox]22DMF (L = 6,13-bis(dodecylaminomethylidene)-1,4,8,11-tetrazacyclotetradeca-4,7,11,14-tetraene); (3) the coupling between Re(IV) and Cu(II) in all of these complexes is antiferromagnetic, according to the distortion in the coordination geometry which introduces better orbital overlap; (4) the copper(II) macrocycle can be used as a ligand toward the [ReCl4(ox)]2 ion, which results in a stronger magnetic interaction than in the Cu(II)Re(IV) complexes presented earlier. Acknowledgments This research was supported by the National Science Centre (Poland) under grant No. 2011/01/B/ST5/01624. The crystallographic part was financed by funds allocated by

8

´ ko et al. / Polyhedron 75 (2014) 1–8 A. Bien

the Ministry of Science and Higher Education to the Institute of General and Ecological Chemistry, Technical University of Lodz. The theoretical calculations were financed by a statutory activity subsidy from the Polish Ministry of Science and Higher Education for the Faculty of Chemistry of Wrocław University of Technology. The authors acknowledge generous computer time from the Wroclaw Supercomputer and Networking Center. Appendix A. Supplementary data CCDC 736420 and 835895 contain the supplementary crystallographic data for compounds 1 and 2. Tables of crystal data, structure refinement, anisotropic displacement coefficients, atomic coordinates and equivalent isotropic displacement parameters for non-hydrogen atoms, H-atom coordinates and isotropic displacement parameters, bond lengths and interbond angles have been deposited with the Cambridge Crystallographic Data Centre. These data can be obtained free of charge via http://www.ccdc.cam.ac.uk/ conts/retrieving.html, or from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223-336-033; or e-mail: [email protected]. References [1] E. Coronado, J. Am. Chem. Soc. 110 (1988) 3907. [2] T. Mallah, C. Auberger, M. Verdaguer, P. Veillet, J. Chem. Soc., Chem. Commun. (1995) 61.

[3] R. Chiozzone, R. González, C. Kremer, G. De Munno, J. Cano, F. Lloret, M. Julve, J. Faus, Inorg. Chem. 38 (1999) 4745. [4] A. Tomkiewicz, A. Zygmunt, J. Mrozinski, J. Mol. Struct. 644 (2003) 97. [5] A. Tomkiewicz, F. Villain, J. Mrozinski, J. Mol. Struct. 555 (2000) 383. [6] A. Tomkiewicz, T.J. Bartczak, R. Kruszynski, J. Mrozinski, J. Mol. Struct. 595 (2001) 225. [7] R. Chiozzone, R. González, C. Kremer, G. De Munno, D. Armentano, F. Lloret, M. Julve, J. Faus, Inorg. Chem. 40 (2001) 4242. [8] R. Chiozzone, R. González, C. Kremer, G. De Munno, D. Armentano, F. Lloret, M. Julve, J. Faus, Inorg. Chem. 42 (2003) 1064. [9] A. Skorupa, B. Korybut-Daszkiewicz, J. Mrozin´ski, Inorg. Chim. Acta (2001) 1. [10] A. Bien´ko, J. Kłak, J. Mrozin´ski, R. Bocˇa, I. Brüdgam, H. Hartl, Dalton Trans. (2007) 2681. [11] A. Tomkiewicz, J. Mrozinski, I. Brüdgam, H. Hartl, Eur. J. Inorg. Chem. (2005) 1787. [12] A. Tomkiewicz, J. Mrozinski, B. Korybut-Daszkiewicz, I. Brüdgam, H. Hartl, Inorg. Chim. Acta 358 (2005) 2135. [13] A. Bien´ko, J. Kłak, J. Mrozin´ski, R. Kruszyn´ski, D. Bien´ko, R. Bocˇa, Polyhedron 27 (2008) 2464. [14] E.A. Boudreaux, L.N. Mulay, Theory and Applications of Molecular Paramagnetism, Wiley-Interscience, New York, 1976. [15] X-RED, Version 1.18, STOE & Cie GmbH, Darmstadt, Germany, 1999. [16] G.M. Sheldrick, Acta Crystallogr., Sect. A 46 (1990) 467. [17] G.M. Sheldrick, SHELXL97. Program for the Solution and Refinement of Crystal Structures, University of Göttingen, Germany, 1997. [18] G.M. Sheldrick, SHELXTL: release 4.1 for Siemens Crystallographic Research Systems, 1990. [19] J. Bernstein, R.E. Davis, L. Shimoni, Angew. Chem., Int. Ed. Engl. 34 (1995) 1555. [20] O. Kahn, Molecular Magnetism, VCH, New York, 1993. [21] H.M. McConell, J. Chem. Phys. 39 (1963) 1910. [22] J. Cano, E. Ruiz, S. Alvarez, M. Verdaquer, Comments Inorg. Chem. 20 (1998) 27. [23] J. Samuel Smart, Effective Field Theories of Magnetism, W. B. Saunders Comp, Philadelphia and London, 1966.