SCIENCE CHINA Chemistry. Cyanido-bridged one-dimensional systems assembled from [Re IV Cl 4 (CN) 2 ] 2 and [M II (cyclam)] 2+ (M = Ni, Cu) precursors

Transcription

1 SCIENCE CHINA Chemistry ARTICLES June 2012 Vol.55 No.6: SPECIAL TOPIC Molecular Magnetism doi: /s z Cyanido-bridged one-dimensional systems assembled from [Re IV Cl 4 (CN) 2 ] 2 and [M II (cyclam)] 2+ (M = Ni, Cu) precursors BHOWMICK Indrani 1,2, HARRIS T. David 3, DECHAMBENOIT Pierre 1,2, HILLARD Elizabeth A. 1,2, PICHON Céline 1,2, JEON Ie-Rang 1,2 & CLÉRAC Rodolphe 1,2* 1 CNRS, CRPP, UPR 8641, F Pessac, France 2 Univ. Bordeaux, CRPP, UPR 8641, F Pessac, France 3 Department of Chemistry, University of California, Berkeley, California 94720, USA Received March 5, 2012; accepted March 26, 2012; published online May 11, 2012 Three new cyanido-bridged heterometallic Re IV Ni II and Re IV Cu II one-dimensional systems were synthesized and extensively characterized both structurally and magnetically. Single-crystal X-ray diffraction analysis revealed that these compounds display a common topology, with chains composed of alternating [Re IV Cl 4 (CN) 2 ] 2 and [M II (cyclam)] 2+ (M = Ni in 1, Cu in 2) or [Cu II (N,N -dimethylcyclam)] 2+ (in 3) building units. Two different chain orientations with a tilt angle of ca. 51 to 55 are present in the crystal packing of these compounds. The magnetic susceptibility measurements suggest the presence of intrachain ferromagnetic interactions between the S = 3/2 Re IV centers and the 3d metal ions: S = 1 Ni II or S = 1/2 Cu II. At low temperature, a three-dimensional ordered magnetic phase induced by interchain antiferromagnetic interactions (antiferromagnetic for 1 and 2; canted antiferromagnetic for 3) is detected for the three compounds. molecular magnets, cyanido-based materials, crystal structure, magnetic properties 1 Introduction Over the last two decades, low-dimensional magnetic systems have been a core interest in the field of molecular magnetism. In particular, molecular complexes such as Single-Molecule Magnets (SMMs) [1 5] and one-dimensional solids such as Single-Chain Magnets (SCMs) [6 12] have been a topic of intense research activity. Of special interest is their slow magnetization dynamics, which give rise to magnet-like behavior at the molecular level. Heterobimetallic coordination compounds featuring various transition metal ions have significantly enriched our understanding of SCM properties [7 11]. In the design of such systems, molecular building units bearing bidentate ligands such as cya- *Corresponding author ( clerac@crpp.bordeaux.cnrs.fr) nido groups (CN ) have been extensively employed [9, 11]. Besides the catalogue of well known hexacyanidometallates, chemists are particularly interested in mononuclear transcyanido transition metal complexes to build one-dimensional heterometallic compounds [13 16], mainly because the cyanido moiety allows a highly directional coordination mode and behaves as a good mediator of magnetic exchange between spin carriers [17]. We have recently reported the highly anisotropic building unit trans-[re IV Cl 4 (CN) 2 ] 2 with a spin ground state of S = 3/2 (t 2g 3 ), which can form SCMs with 3d transition metal ions (Mn II, Fe II, Co II, Ni II ) [18]. Notably, the magnetic exchange through the cyanido bridge between Re IV Ni II or Re IV Cu II was found to be ferromagnetic [18, 19], remarkably so in the Re IV Cu II case [19]. In order to further explore these bimetallic 3d-5d systems, we herein report our studies on [Re IV Cl 4 (CN) 2 ] 2 Science China Press and Springer-Verlag Berlin Heidelberg 2012 chem.scichina.com

2 Bhowmick I, et al. Sci China Chem June (2012) Vol.55 No linked to 3d transition metal macrocycles [M II (cyclam)] 2+ (M = Ni, Cu, and cyclam = 1,4,8,11-tetraazacyclotetradecane) and [Cu II (N,N -dimethylcyclam)] 2+ (N,N -dimethylcyclam = 1,8-dimethyl-1,4,8,11-tetraazacyclotetradecane). Ni II and Cu II metallocyclam units are known to form magnetic one-dimensional compounds with different cyanido-based transition metal complexes, organic radicals and other metal organic moieties [14, 20 25]. The 14-member tetradentate cyclam ligand can incorporate transition metal ions in its cavity and in the case of Ni II and Cu II, it occupies the equatorial positions of the metal ion to give a D 4h coordination sphere. With two free axial positions, these building units are therefore particularly well adapted to form one-dimensional compounds with [Re IV Cl 4 (CN) 2 ] 2. Indeed, reactions between [Re IV Cl 4 (CN) 2 ] 2 and [M II (cyclam)] 2+ (M = Ni, Cu) precursors afforded two isostructural one-dimensional compounds with the general molecular formula [M II (cyclam)][re IV Cl 4 (CN) 2 ] H 2 O (M = Ni (1), Cu (2)). A similar combination of [Re IV Cl 4 (CN) 2 ] 2 with [Cu II (N,N - dimethylcyclam)] 2+ led to another chain system, [Cu II (N,N - dimethylcyclam)][re IV Cl 4 (CN) 2 ] (3) with a closely related crystal structure. Herein, we report the synthesis of these three new compounds together with their structural and magnetic properties. 2 Experimental section 2.1 Synthesis of the compounds All chemicals and solvents used during the syntheses were of reagent grade. The compounds [Ni(cyclam)](ClO 4 ) 2, [Cu(N,N -dimethylcyclam)](clo 4 ) 2, [Cu(cyclam)](ClO 4 ) 2 and (Bu 4 N) 2 [ReCl 4 (CN) 2 ] 2DMA were synthesized according to literature methods [18, 26]. All the following procedures were carried out at ambient temperature. Caution! Perchlorate salts are potentially explosive and should be handled with care and in small quantity Synthesis of [Ni II (cyclam)][re IV Cl 4 (CN) 2 ] H 2 O (1) 4.6 mg (0.010 mmol) of [Ni II (cyclam)](clo 4 ) 2 was dissolved in 2 ml of water, placed in a 1.5 cm diameter tube and layered with 2 ml of pure acetonitrile. Over the pure solvent an acetonitrile solution of 10 mg ( mmol) of (Bu 4 N) 2 [Re IV Cl 4 (CN) 2 ] 2DMA was layered very carefully. The tube was then closed with paraffin to avoid evaporation. After 2 days, brown crystalline plates suitable for single-crystal X-ray diffraction were found at the bottom and walls of the tube. The crystals were collected through filtration and dried in air. The yield was ca. 60% based on the Ni II starting material. Elemental analysis calcd (%) for C 12 H 26 Cl 4 N 6 NiORe (M = g/mol): C 21.91, H 3.96, N 12.78; found (%) C 21.85, H 4.03, N 12.65; IR (KBr): = 2141 (C N), 1296 (C N) cm Synthesis of [Cu II (cyclam)][re IV Cl 4 (CN) 2 ] H 2 O (2) 4.6 mg ( mmol) of [Cu II (cyclam)](clo 4 ) 2 was dissolved in 2 ml of nitromethane, placed in a 1.5 cm diameter tube and layered with 2 ml of pure nitromethane. A nitromethane solution of 10 mg ( mmol) of (Bu 4 N) 2 [Re IV Cl 4 (CN) 2 ] 2DMA was layered very carefully on the top of the pure solvent layer. The tube was then closed with paraffin to avoid evaporation. After 2 days, red crystalline plates suitable for single-crystal X-ray diffraction were found at the bottom and walls of the tube. The crystals were collected through filtration and dried in air. The yield was ca. 70% based on the Cu II precursor. Elemental analysis calcd (%) for C 12 H 26 Cl 4 N 6 CuORe (M = g/mol): C 21.75, H 3.93, N 12.69; found (%) C 21.70, H 4.02, N 12.59; IR (KBr): = 2132 (C N), 1280 (C N) cm Synthesis of [Cu II (N,N -dimethylcyclam)][re IV Cl 4 (CN) 2 ] (3) This compound was synthesized following the same procedure as for compound 2 using 2 ml of a nitromethane solution of [Cu II (N,N -dimethylcyclam)](clo 4 ) 2 (4.9 mg; mmol). The yield was ca. 60% based on the Cu II precursor. Elemental analysis calcd (%) for C 14 H 28 Cl 4 N 6 CuRe (M = g/mol): C 25.00, H 4.17, N 12.50; found (%) C 25.09, H 4.28, N 12.46; IR: = 2130 (C N), 1364, 1391 (C N) cm Physical characterization Elemental analyses (C, H, and N) were carried out by FlashEATM 1112 automatic elemental analyzer. Infrared spectra of 1 and 2 were recorded in the range of cm 1 on a Nicolet 750 Magna-IR spectrometer using KBr pellets. The infrared spectrum of 3 was collected in the same range on a Thermal Scientific Nicolet 6700 FT-IR spectrometer equipped with a Smart itr diamond window. Heat capacity was measured on polycrystalline samples of 1, 2 and 3 (1.0 mg) between 150 and 1.9 K by a thermal relaxation technique with a Quantum Design Physical Property Measurement System (PPMS-9) in zero-dc field. In each measurement, the blank heat capacity including a small amount of Apiezon N grease (1.5 mg) used for adhesion was measured prior to the sample mounting. The heat capacities of the compound were determined by subtracting the blank data from the measured total heat capacity. Magnetic susceptibility measurements were performed using a Quantum Design MPMS-XL SQUID magnetometer. The measurements were performed on freshly filtered polycrystalline samples introduced in a polyethylene bag ( cm). The dc measurements were conducted from 300 to 1.8 K and between 70 and 70 koe applied dc fields. An M vs. H measurement was performed at 100 K to confirm the absence of ferromagnetic impurities. The field

3 1006 Bhowmick I, et al. Sci China Chem June (2012) Vol.55 No.6 Table 1 Single crystal X-ray diffraction parameters for compounds Formula C 12 H 26 Cl 4 N 6 Ni C 12 H 26 Cl 4 N 6 Cu C 14 H 28 Cl 4 N 6 Cu ORe ORe Re M W (g/mol) T (K) 100(2) 100(2) 82(2) Crystal system monoclinic monoclinic monoclinic Space group C2/c C2/c C2/c a (Å) (19) (12) (9) b (Å) (8) (5) (3) c (Å) (3) (10) (8) β ( ) (3) (2) (4) Volume (Å 3 ) (4) (2) (17) Z D (g cm 3 ) μ MoKα (mm 1 ) F GooF a) R 1 (I > 2 (I)) b) wr 2 (all) a) R 1 = Σ F o F C /Σ F o ; b) wr 2 = [Σw(F o 2 F C 2 ) 2 /Σw(F o 2 ) 2 ] 1/2, w = 1/[ 2 (F o 2 ) + (ap) 2 + bp] and P = (F o 2 + 2F C 2 )/3. dependence of the magnetization was measured between 1.83 and 10 K with a dc magnetic field between 7 and 7 T. The ac susceptibility experiments were performed at various frequencies ranging from 1 to 1500 Hz with an ac field amplitude of 3 Oe in zero dc field. Experimental data were corrected for the sample holder and for the diamagnetic contribution of the sample. After the magnetic measurements, the polycrystalline samples were checked by powder X-ray diffraction and neither decomposition, dehydration nor evolution of the compounds was detected. Powder X-ray diffraction patterns were recorded using a Analytical X pert diffractometer (CuK, X'celerator detector) within the range of 6 to 80 (2 ) using 60 s exposures with 0.02 steps Single crystal X-ray diffraction For all compounds, a single crystal was coated with Paratone N oil and mounted on a fiber loop followed by data collection at 100 K for 1 and 2 and at 82 K for 3. The crystallographic data were collected with a Bruker APEX II diffractometer, equipped with a graphite monochromator centered on the path of MoKα (λ = Å). The program SAINT was used to integrate the data, which was thereafter corrected for absorption using SADABS [27]. The structure was solved by direct methods and refined by least squares on F 2 in SHELX97 [28]. All non-hydrogen atoms were refined anisotropically by full-matrix least-squares (SHELXL-97). Hydrogen atoms were assigned to ideal positions using the appropriate HFIX command in SHELXL-97. Table 1 contains the summary of the unit cell and structure refinement parameters for all the compounds. Selected bond lengths and angles are summarized in Table S1 for the three complexes. Crystallographic data (excluding structure factors) for the structures reported in this paper have been deposited at the Cambridge Crystallographic Data Centre as supplementary publication nos. CCDC for 1, CCDC for 2 and CCDC for 3. Copies of the data can be obtained free of charge on application to CCDC, 12 Union Road, Cambridge CB21EZ, UK (fax: (+44) ; deposit@ccdc.cam.ac.uk). 3 Results and discussion 3.1 Crystal structures Compounds 1 3 crystallized in the monoclinic C2/c space group and display very comparable crystal structures. Indeed, 1 and 2 are isostructural while 3 is only differentiated by the absence of the lattice water molecules. The structures of these compounds consist of one-dimensional chains (Figures 1 and S1) composed of alternating [Re IV Cl 4 (CN) 2 ] 2 and [M II (cyclam)] 2+ (M = Ni in 1, Cu in 2) or [Cu(N,N -dimethylcyclam)] 2+ (in 3) units. In these one-dimensional assemblies, the cyanido groups of the [Re IV Cl 4 (CN) 2 ] 2 unit bridge the 3d and 5d metal ions. In compound 3, the nitrogen and carbon atoms of the cyanido bridge were each modeled as disordered over two positions with relative occupancies of 0.7 and 0.3. In the major part, the bridge is essentially perpendicular to the ReCl 4 plane, (Cl1 Re1 C1A = 94.1(3) and Cl2 Re1 C1A = 92.1(3) ), as one would expect. The displacement of C1 and N1 in the minor part generates a second bridge position, which is distorted relative to the Re Cl axes (Cl1 Re1 C1B = 101.8(6) and Cl2 Re1 C1B = 98.2(6) ). The angles be- Figure 1 Crystal structures emphasizing the one-dimensional arrangement of 1 (a) and 3 (b) with selected atom-labeling schemes and thermal ellipsoids at 30% probability. The disorder of the cyanido groups and cyclam ligands is highlighted in 3 with shaded atoms corresponding to the minor part (ca. 30%) of the disordered atoms. The orange, green, blue, gray and violet ellipsoids correspond to Re, Cl, N, C and Ni/Cu atoms, respectively. Hydrogen atoms are omitted for clarity.

4 Bhowmick I, et al. Sci China Chem June (2012) Vol.55 No tween the two cyanido bridges are about 20 (Figure 1(b)). Disorder on the cyclam ring around the nitrogen atom bearing the methyl group was also modeled with the same relative occupancies. This disorder projects the minor methyl group towards the opposite face of the cyclam. To simplify the comparison between the complexes, the following discussion will only treat the major part of compound 3. Each metal ion resides on an inversion center and adopts an octahedral coordination geometry. Carbon atoms from the cyanido groups are coordinated to the Re IV axial sites with similar Re C bond distances (2.126(3) Å in 1, 2.131(2) Å in 2 and 2.124(8) Å in 3), and quite linear Re C N angles between 170.7(2) for 1 and 174.6(9) in 3 (174.1(2) in 2). The rhenium equatorial plane is occupied by four chlorido ligands (Re Cl: 2.322(2) to (5) Å) that are almost perfectly perpendicular to the C Re C axis (C Re Cl obtuse angles: 90.41(7) in 1 to 94.1(3) in 3) with Cl Re Cl angles close to 90 (obtuse angles between 90.80(2) (2) and 92.0(1) (3)). These geometric parameters are nearly identical to those observed for the [Re IV Cl 4 (CN) 2 ] 2 unit as its [Bu 4 N] + salt or in analogous one-dimensional systems [18, 19]. The Ni II and Cu II metal ions possess an axially elongated octahedral coordination sphere with four nitrogen atoms belonging to the neutral cyclam ligand occupying the equatorial positions (Ni N: 2.073(2), 2.080(2) Å in 1, Cu N: 2.021(2), 2.041(2) Å in 2 and 2.079(9), 2.005(6) Å in 3). It is worth noting that the Cu N bond distances are significantly longer for the N atoms bearing hydrogen atoms (2.079(9) Å) than for those bearing methyl groups (2.005(6) Å) in 3. The distortion of the 3d metal ion octahedral geometry is clearly seen in the axial positions, which are occupied by the N atoms of the cyanido groups, with Ni N and Cu N bond distances reaching 2.129(2) Å in 1 and 2.531(9) Å in 3 (2.482(2) Å in 2), respectively. The M N C angles deviate notably from linearity with Ni N C and Cu N C angles of 148.5(2) and 132.1(2) /130.9(8) (in 2/3), respectively. Due to this nonlinearity, the individual chains adopt a mild zig-zag pattern along their axis of propagation. In 1 and 2, water molecules fill the vacant space inside the crystal packing between chains, while in 3, the methyl groups of the macrocycle ligand play this role. Therefore, these three compounds experience very similar packing of the chains. In the crystal structures, the individual chains are parallel to the (ab) plane, forming layers of chains, which stack to form the crystal. The chain axes originating from two successive layers are tilted with an angle of ca. 51 for 1 and 2, and about 55 for 3 with respect to each other (Figure 2). In 1 and 2, the water molecules are weakly hydrogen bonded to four chlorido ligands belonging to two Re IV sites of neighboring chains in two different layers (Figure 3 and S2) with an average distance d O Cl of ca Å. Additionally, these water molecules also act as hydrogen bond acceptors with two N H groups of the cyclam ligands (d N O = Figure 2 Crystal packing perpendicular to the c axis for 1 illustrating the organization of two consecutive chain layers parallel to the (ab) plane. For visual clarity, each layer is represented in a different color. Hydrogen atoms and water molecules are omitted for clarity. Figure 3 Hydrogen bonding pattern observed in 2 with a water molecule connecting the chain layers Å). Similarly, for compound 3, close contacts between the chloride atoms and the methyl groups of the cyclam (d C Cl = 3.56 Å) are observed (Figure S3). 3.2 Magnetic properties The ac and dc susceptibility measurements were performed on a polycrystalline sample of the three one-dimensional compounds to probe their magnetic properties (Figures 4, S4 S6, S8 and S9). The temperature dependence of their T product collected in a dc magnetic field of 1000 Oe is shown in Figure 4 for 1, 2, 3 and the (Bu 4 N) 2 [ReCl 4 (CN) 2 ] 2DMA precursor. At 270 K, the T value is 2.7, 1.9 and 1.8 cm 3 K mol 1 for 1, 2 and 3, respectively, in good agreement with the expected Curie constants of 2.6, 1.8 and 1.8 cm 3 K mol 1 considering g factors of 2.2, 2.2 and 1.7 for isolated Ni II (S = 1), Cu II (S = 1/2) and Re IV (S = 3/2) [18] spin carriers, respectively. The temperature dependence of the T product for the three compounds is quite different as

5 1008 Bhowmick I, et al. Sci China Chem June (2012) Vol.55 No.6 shown in Figure 4. For the Ni II analogue 1, upon lowering the temperature, T decreases first down to 2.41 cm 3 K mol 1 at 19.5 K, then increases to 2.48 cm 3 K mol 1 at 9.2 K before finally decreasing abruptly down to 0.35 cm 3 K mol 1 at 1.8 K. For 2, T increases first up to 2.03 cm 3 K mol 1 at ca. 31 K indicating a dominant intrachain ferromagnetic interaction as already observed in a related cyanido-bridged Re IV /Cu II system [19]. At lower temperatures, T decreases slightly to 1.97 cm 3 K mol 1 around 7 K and finally decreases abruptly down to 0.17 cm 3 K mol 1 at 1.8 K. For 3, T noticeably decreases to 1.13 cm 3 K mol 1 at about 7.8 K, then increases up to 4.06 cm 3 K mol 1 at 3.9 K and lastly decreases down to 2.29 cm 3 K mol 1 at 1.8 K. Due to the intrinsic magnetic properties of the [ReCl 4 (CN) 2 ] 2 precursor, it is not straightforward to assign the nature of the intrachain magnetic interaction in 1 and 3. As shown previously [18, 19], the decrease of the T product upon lowering the temperature for the Re IV precursor (Figure 4) is essentially due to its magnetic anisotropy. This thermal behavior is superimposed onto the effect of the intrachain interactions and thus no simple one-dimensional magnetic model can be applied to evaluate these exchange couplings. Nevertheless, the ferromagnetic nature of the intrachain interaction in 2 is clearly established by the increase of the T product between 270 and 31 K. This is also confirmed by the magnetization value at 1.8 K and 7 T that reaches 2.73 B (Figure S6) in agreement with a ferromagnetic arrangement of the spins along the chain (expected magnetization values at 7 T and 1.8 K for the Re IV precursor: 1.5 B [18] and the Cu II moiety with g = 2.2: 1.1 B ). For 1 and 3, the same conclusion concerning the nature of the intrachain interaction can be drawn based on the M vs. H data (Figures 5, S4 and S8). At 1.8 K and 7 T, the magnetization of 1 and 3 reaches 2.8 B and 1.9 B, respectively, which are only compatible with ferromagnetic intrachain interactions. It is worth noting that antiferromagnetic intra- chain interactions would lead to small magnetization values around 0.7 B for 1 (expected magnetization value for the Ni II moiety with g = 2.2 at 7 T and 1.8 K: 2.2 B ) and 0.4 B for 2 and 3 at 7 T. For 1 and 3, it seems that the 3d/5d ferromagnetic interactions through the cyanido bridge are weaker than in 2 and thus they do not compensate or overcome the effect of the Re IV magnetic anisotropy leading to a decrease of the T product upon lowering the temperature (Figure 4). It should be mentioned that the Re IV Cu II magnetic interaction through the cyanido bridge is much weaker in the present two systems than in the (Bu 4 N) [TpCuReCl 4 (CN) 2 ] one-dimensional compound (Tp : hydrotris(pyrazol-1-yl)borate) [19]. This likely results from the different coordination sites occupied by the cyanido groups on the Cu II ion: axial in 2 and 3 versus equatorial for (Bu 4 N)[TpCuReCl 4 (CN) 2 ] associated with longer and shorter Cu N bond distances, respectively. The M vs. H data measured below 10 K (Figures 5, S4, S6 and S8) for the three compounds also reveal their respective magnetic ground state. 1 and 2 display very similar low temperature magnetic properties with magnetization curves which exhibit a typical S shape curve (i.e. with an inflexion point at H C ) at 1.8 K. This feature reveals the presence of antiferromagnetic interactions between chains compensated by the applied magnetic field at H C = 9400 Oe and 9340 Oe for 1 and 2, respectively. This characteristic field has been followed as a function of the temperature (using combined M vs. H and vs. T data and taking the maximum of the dm/dh vs. H and vs. T plots; Figures 5, S4 S6) in order to build the corresponding (T, H) magnetic phase diagram. These results demonstrate the presence of a three-dimensional antiferromagnetic order in both compounds with T N = 4.5 K and 4.1 K for 1 and 2, respectively. Heat capacity (C p ) measurements (Figure S7) further confirm the presence of the magnetic phase transitions as evidenced by a marked C p peak observed at the Néel tempera- Figure 4 Temperature dependence of the T product (where is the molar magnetic susceptibility that is equal to M/H) collected in an applied dc field of 1000 Oe for 1, 2, 3 and the (Bu 4 N) 2 [ReCl 4 (CN) 2 ] 2DMA precursor. Figure 5 Field dependence of the magnetization for a polycrystalline sample of 1 between 1.8 and 6 K with sweep-rates of Oe/min. Inset: Temperature dependence of the molar magnetic susceptibility (where is equal to M/H) at different dc fields up to 1.2 T between 1.83 and 12 K for a polycrystalline sample of 1. Solid lines are guides for the eye.

6 Bhowmick I, et al. Sci China Chem June (2012) Vol.55 No ture. The topology of the phase diagram shown in Figure 6 for 1 and 2 is characteristic of a metamagnetic behavior with only an antiferromagnetic-paramagnetic transition line that is certainly induced by the magnetic anisotropy brought by the Re IV metal ions. This strong magnetic anisotropy is also revealed by the high-field magnetization of these compounds, which does not saturate even at 1.8 K and 7 T (Figures 5, S4 and S6). In order to estimate the average interchain interactions, zj', between the effective spin (S eff = 5/2 for 1 and S eff = 2 for 2) of the Re IV -M II chain unit, the following expression was used [29, 30]: g H S 2 zj S (1) 0 2 av B C eff eff where g av is the average g factor of the Re IV -M II unit along the chain (g av = 1.95) and H C 0 is the critical field extrapolated at 0 K (9500 Oe for 1 and 9400 Oe for 2). zj'/k B is thus estimated at 0.25 K and 0.31 K for 1 and 2, which con- Figure 6 H vs. T magnetic phase diagram for 1 and 2. The black and red dots are respectively the experimental points deduced from the M vs. H and vs. T data (Figures 5, S4 S6). The solid black line is a guide for the eye. Figure 7 Field dependence of the magnetization for a polycrystalline sample of 3 between 1.83 and 6 K with sweep-rates of 180 Oe/min. Inset: Expanded view of the main figure between 0.5 and 0.5 T emphasizing the M vs. H hystereses. Solid lines are guides for the eye. firms the existence of significant interchain interactions responsible for the antiferromagnetic phase transition. From zj, the theoretical Néel temperature (T N ) for a threedimensional network of magnetic interactions treated in the mean field approximation can be estimated at 1.5 K and 1.2 K for 1 and 2, respectively, using the following well-known relation: T 2 zj S ( S 1)/3k (2) N eff eff B The marked underestimation of the experimental T N by this mean-field approach demonstrates the low-dimensional nature of the materials and thus the presence of much stronger intrachain interactions than interchain interactions. Quite surprisingly, the magnetic ground state of 3 is different from the two first compounds. The M vs. H data measured below 10 K reveal the presence of a hysteresis effect with a coercive field of 1700 Oe at 1.8 K (at about 180 Oe/min, Figure 7). An M vs. H hysteresis loop is observed up to 5.5 K. At higher field, the magnetization increases almost linearly up to 7 T and reaches 1.9 B at 1.83 K (Figures 7 and S8). This magnetic behavior and the strong increase at low fields of the susceptibility below 6 K (Figure S9) is the typical signature of a canted antiferromagnetic ground state. The small spontaneous magnetization resulting from the non-compensation of the two sublattices in the canted antiferromagnetic phase (M = 0.22 B at 1.83 K) has been deduced from the M vs. H data and plots after normalization, M/M sat, as a function of temperature. As shown in Figure 8 (top part, blue solid line), the M/M sat vs. T data compare almost perfectly with the temperature variation of the Ising order parameter (m) in the mean field approximation (that is, the solution of the self-consistent equation m = tanh (mt C /T)) [31, 32]. This theory/experiment agreement confirms the correlation between the T dependence of M/M sat and the magnetic phase transition with T C = 4.9 K. Between 4.9 K and 6 K, fluctuation effects above the magnetic phase transition are observed on the M/M sat vs. T data (Figure 8). The presence of short-range order in this temperature range is also seen by ac susceptibility measurements. While no out-of-phase component ( '') of the ac susceptibility was observed for 1 and 2, as expected for an antiferromagnetic ground state, compound 3 exhibits a marked ac response around T C with a succession of two peaks on both ' and '' components (Figure 8, bottom part and Figure S9). The comparison of the M/M sat vs. T and '' vs. T data as done in Figure 8 help to easily assign the first '' maximum to a pre-transitional effect while the second peak at 4.9 K is clearly the signature of the magnetic transition. Heat capacity (C p ) measurements (Figure S10) confirm the presence of both the magnetic phase transition at ca. 5 K and the pre-transitional effects between 5 and 6 K, as evidenced by the broad shape of the observed C p peak. A natural question is raised by the detailed study of the

7 1010 Bhowmick I, et al. Sci China Chem June (2012) Vol.55 No.6 magnetic properties of these compounds: why do 1 and 2 display an antiferromagnetic ground state and 3 a canted antiferromagnetic ground state while they have almost the same crystal structures? In 1 and 2, the layers of chains shown by different colors in Figure 2 are likely sheets of antiferromagnetically coupled chains that are themselves antiferromagnetically coupled in the third direction to lead to a three-dimensional antiferromagnetic order. Therefore the tilt angle of ca between chains of neighboring layers does not prevent the antiferromagnetic order and thus is not responsible for the canted antiferromagnetic order seen in 3. The key difference between 2 and 3 is the presence of a disorder along the chains in 3 (Figure 1(b)) with two different orientations of the anisotropic [Re IV Cl 4 (CN) 2 ] 2 moiety along the chains. The angle between these two orientations is about 20 (noted, see Figure 1(b)). Quite remarkably, an angle of around 18 between the two uncompensated sublattices of the canted antiferromagnetic phase is deduced from the spontaneous magnetization (sin( /2) = 2M/M HF, with M = 0.22 B at 1.83 K and M HF = 2.8 B being the magnetization expected at high field after saturation). This result strongly supports the hypothesis that the origin of the canted antiferromagnetic phase in 3 is found in the one-dimensional arrangement itself. As a consequence, it seems very likely that an order-disorder structural phase transition occurs between 82 K (lowest temperature of the crystal structure collection) and T C (4.9 K) in order to stabilize a well-ordered magnetic state. This hypothesis is supported by heat capacity measurements (Figure S10) that display a broad feature at ca. 52 K that could be the signature of this phase transition. 4 Conclusion Figure 8 Top part: Temperature dependence of the normalized spontaneous magnetization for a polycrystalline sample of 3 between 1.8 and 10 K deduced from the M vs. H data shown in Figure 7. The solid blue line is the theoretical spontaneous magnetization in the frame of the incompressible Ising model in the mean-field approximation with T C = 4.9 K [31, 32]. Solid black line is a guide for the eye. Bottom part: Temperature dependence of the out-of-phase component of the ac susceptibility between 1.8 and 10 K at different ac frequencies, under zero dc field and with a 3 Oe ac field modulation. Solid lines are guides for the eye. Three cyanido-bridged heterometallic one-dimensional compounds composed of alternating [Re IV Cl 4 (CN) 2 ] 2 and [M II (cyclam)] 2+ (M = Ni in 1, Cu in 2) or [Cu(N,N -dimethylcyclam)] 2+ (in 3) building units were synthesized and extensively characterized structurally and magnetically. The study of their magnetic properties suggests the presence of intrachain ferromagnetic interactions between the S = 3/2 Re IV centers and the 3d metal ions: S = 1 Ni II or S = 1/2 Cu II as already seen in related systems [18, 19]. Below 4.5 and 4.1 K, a three-dimensional ordered antiferromagnetic phase is observed for 1 and 2 respectively, while a canted antiferromagnetic state has been detected below 4.9 K for 3. Interchain antiferromagnetic interactions in the three compounds stabilize these ordered magnetic phases, but the two different orientations, disordered at 82 K, of the chain repeating unit are likely responsible for the canted antiferromagnetic phase observed in 3. This work was supported by the Centre National de la Recherche Scientifique (CNRS), the University of Bordeaux, the Conseil Régional d'aquitaine, GIS Advanced Materials in Aquitaine (COMET Project), the ANR (NT09_469563, AC-MAGnets project), and the Erasmus Mundus Mobility with Asia (EMMA) program (External Cooperation Window-ASIE) for the PhD fellowship of I. B. We would like also to thank Prof. Claude Coulon for fruitful discussions, Eric Lebraud for powder X-ray diffraction measurements, S. Calancea for the elemental analyses and Prof. Jeffrey R. Long for his inspiring work on the chemistry of cyanido-metallate complexes and his kind help with the I. B. research project. 1 Sessoli R, Tsai HL, Schake AR, Wang S, Vincent JB, Folting K, Gatteschi D, Christou G, Hendrickson DN. High-spin molecules: [Mn 12 O 12 (O 2 CR) 16 (H 2 O) 4 ]. J Am Chem Soc, 1993, 115: Sessoli R, Gatteschi D, Caneschi A, Novak MA. Magnetic bistability in a metal-ion cluster. Nature, 1993, 365: Gatteschi D, Sessoli, R. Quantum tunneling of magnetization and related phenomena in molecular materials. Angew Chem Int Ed, 2003, 42: Aromi G, Brechin EK. Synthesis of 3d metallic single-molecule magnets. Struct Bonding, 2006, 122: Rinehart JD, Long JR. Exploiting single-ion anisotropy in the design of f-element single-molecule magnets. Chem Sci, 2011, 2: Caneschi A, Gatteschi D, Lalioti N, Sangregorio C, Sessoli R, Venturi G, Vindigni A, Rettori A, Pini MG, Novak MA. Cobalt(II)-nitronyl nitroxide chains as molecular magnetic nanowires. Angew Chem Int Ed, 2001, 40: Clérac R, Miyasaka H, Yamashita M, Coulon C. Evidence for singlechain magnet behavior in a Mn III -Ni II chain designed with high spin magnetic units: A route to high temperature metastable magnets. J Am Chem Soc, 2002, 124:

8 Bhowmick I, et al. Sci China Chem June (2012) Vol.55 No Coulon C, Clérac R, Lecren L, Wernsdorfer W, Miyasaka H. Glauber dynamics in a single-chain magnet: From theory to real systems. Phys Rev B, 2004, 69: Coulon C, Miyasaka H, Clérac R. Single-chain magnets: theoretical approach and experimental systems. Struct Bonding (Berlin), 2006, 122: Bogani L, Vindigni A, Sessoli R, Gatteschi D. Single chain magnets: where to from here? J Mater Chem, 2008, 18: Miyasaka H, Julve M, Yamashita M, Clérac R. Slow dynamics of the magnetization in one-dimensional coordination polymers: Single-chain magnets. Inorg Chem, 2009, 48: Coulon C, Clérac R, Wernsdorfer W, Colin T, Miyasaka H. Realization of a magnet using an antiferromagnetic phase of single-chain magnets. Phys Rev Lett, 2009, 102: Zhang D, Zhang LF, Chen Y, Wang H, Ni ZH, Wernsdorfer W, Jiang J. Heterobimetallic porphyrin-based single-chain magnet constructed from manganese(iii)-porphyrin and trans-dicyano bis (acetylacetonato) ruthenate(iii) containing co-crystallized bulk anions and cations. Chem Commun, 2010, 46: Yeung WF, Lau PH, Lau TC, Wei HY, Sun HL, Gao S, Chen ZD, Wong WT. Heterometallic M II Ru III 2 compounds constructed from trans-[ru(salen)(cn) 2 ] and trans-[ru(acac) 2 (CN) 2 ]. Synthesis, structures, magnetic properties, and density functional theoretical study. Inorg Chem, 2005, 44: Kou HZ, Ni ZH, Liu CM, Zhang DQ, Cui AL. Cyanide-bridged 1D Mn(III) Fe(III) bimetallic complexes: Synthesis, crystal structure and magnetic properties. New J Chem, 2009, 33: Guo JF, Yeung WF, Lau PH, Wang XT, Gao S, Wong WT, Chui SSY, Che CM, Wong WY, Lau TC. Trans-[Os III (salen)(cn) 2 ] : A new paramagnetic building block for the construction of molecule-based magnetic materials. Inorg Chem, 2010, 49: Beltran LMC, Long JR. Directed assembly of metal-cyanide cluster magnets. Acc Chem Res, 2005, 38: Harris TD, Bennett MV, Clérac R, Long JR. [ReCl 4 (CN) 2 ] 2 : A high magnetic anisotropy building unit giving rise to the single-chain magnets (DMF) 4 MReCl 4 (CN) 2 (M = Mn, Fe, Co, Ni). J Am Chem Soc, 2010, 132: Harris TD, Coulon C, Clérac R, Long JR. Record ferromagnetic exchange through cyanide and elucidation of the magnetic phase diagram for a Cu II Re IV (CN) 2 chain compound. J Am Chem Soc, 2011, 133: Ballester L, Gil AM, Gutiérrez A, Perpiñán MF, Azcondo MT, Sánchez AE, Amador U, Campo J, Palacio F. Polymorphism in [Cu(cyclam)(TCNQ) 2 ](TCNQ) stacked systems (cyclam=1,4,8,11- tetraazacyclo-tetradecane, TCNQ = 7,7,8,8-tetracyanoquinodimethane). Inorg Chem, 1997, 36: Ballester L, Gutiérrez A, Perpiñán MF, Rico S, Azcondo MT, Bellitto C. Magnetic behavior and crystal structure of [Fe(cyclam) (NCS) 2 ](TCNQ) 2 : An unusual one-dimensional (TCNQ) 2 radical-ion system. Inorg Chem,1999, 38: Atanasov M, Comba P, Förster S, Linti G, Malcherek T, Miletich R, Prikhod ko AI, Wadepohl H. Tuning of inter-versus intrachain magnetic interactions in cyano-bridged Ni II /M III (M = Cr, Fe, Co) chain complexes. Inorg Chem, 2006, 45: Oshio H, Okamoto H, Kikuchi T, Ito T. The ferromagnetic chain system catena-(μ-cro 4 -O,O )[Ni II (cyclam)] 2H 2 O. Inorg Chem,1997, 36: Duimstra JA, Stern CL, Meade TJ. Magnetic and structural properties of a cyano-bridged 1D ruthenium(iii)-nickel(ii) coordination polymer. Polyhedron, 2006, 25: Jeannin O, Clérac R, Cauchy T, Fourmigué M. Ferromagnetic interactions in heterobimetallic chains formed through the secondary coordination of dithiolene complexes. Inorg Chem, 2008, 47: Bosnich B, Tobe ML, Webb GA. Complexes of nickel(ii) with a cyclic tetradentate secondary amine. Inorg Chem, 1965, 4: Sheldrick GM. SADABS, Version 2.03; Bruker Analytical X-Ray Systems, Madison, WI, Sheldrick GM. SHELXTL, Version 6.12; Bruker Analytical X-Ray Systems, Madison, WI, Jongh LJ, Miedena AR. Experiments on simple magnetic model systems. Adv In Physics, 1974, 23: 1 30 Chikazumi S. Physics of Ferromagnetism. Oxford: Clarendon Press, 1997, Coulon C, Moreau S. Physique Statistique et Thermodynamique. Dunod Eds, Paris, 2000, Chaikin PM, Lubensky TC. Principles of Condensed Matter Physics. Cambridge: Cambridge University Press, 1995,