Structural and magnetic properties of iron(II)

FULL PAPER

Received 21st February 2005, Accepted 25th April 2005 First published as an Advance Article on the web 10th May 2005

Dalton

Mikhail Shatruk, Abdellatif Chouai, Andrey V. Prosvirin and Kim R. Dunbar* Department of Chemistry, Texas A&M University, College Station, TX, 77842, USA

www.rsc.org/dalton

Structural and magnetic properties of iron(II) complexes with 1,4,5,8,9,12-hexaazatriphenylene (HAT)†

Reactions of Fe(II) salts with the ligand 1,4,5,8,9,12-hexaazatriphenylene (HAT) led to the isolation and characterization of four new compounds: [Fe3 (HAT)(H2 O)12 ](SO4 )3 ·3.3H2 O (1), [Fe2 (HAT)(SO4 )(H2 O)5 ](SO4 )· 2H2 O·CH3 OH (2), [Fe2 (HAT)(SO4 )(H2 O)5 ](SO4 )·3H2 O (3), and [Fe3 Cl5 (HAT)(CH3 OH)4 (H2 O)]Cl (4). Compound 1 crystallizes as a trinuclear cluster in which HAT acts as a tris-chelating ligand. Compounds 2 and 3 are two polymorphs of an infinite one-dimensional structure in which the Fe atoms are coordinated to HAT and then connected into the chain through bridging sulfate anions. Compound 4 exhibits a similar chain structure, but with bridging chloride ligands. The magnetic behavior of the new compounds is indicative of weak antiferromagnetic coupling between the Fe(II) centers through the HAT ligand.

Introduction

DOI: 10.1039/b502567b

The tris-chelating ligand 1,4,5,8,9,12-hexaazatriphenylene (HAT) is an interesting candidate for coordination chemistry, but due to the laborious synthesis and its low solubility in most common solvents,1 it has not been explored as extensively as other nitrogen-containing heterocycles, e.g., bipyridine or phenanthroline. In spite of the inherent problems, chemists began to explore HAT chemistry in the later 1980’s and, since this time, complexes of HAT and its derivatives have been studied in the context of photochemical,2 supramolecular,3 magnetic,4 and biological chemistry.5 Our particular interest in the HAT molecule is to assemble triangular arrangements of paramagnetic metal ions. Such complexes represent the simplest models of geometrically frustrated magnetic systems (Scheme 1). A study of this phenomenon at the molecular level will provide additional data for modeling complex spin frustrated systems which are still poorly understood. So far, the discrete trinuclear clusters based on the homonuclear M3 (l3 -HAT) core have been reported only for diamagnetic Cr(0),6 Cu(I),3 Ru(II),2,7 and Rh(III).7 Heterometallic complexes containing Ru2 Os,8 Ru2 Rh, and RuRh2 7 cores have also been described. In the case of Co3 (HAT)[N(CN)2 ]6 (OH2 )2 ,4 the only compound that contains a paramagnetic M3 (l3 -HAT) fragment, the Co3 (l3 -HAT) units are bound into a three-dimensional framework via dicyanamide anions.

† Electronic supplementary information (ESI) available: Complete listings of atomic and thermal parameters, bond distances and bond angles, as well as the ORTEP plots of asymmetric units for compounds 1–4. See http://www.rsc.org/suppdata/dt/b5/b502567b/

©

Experimental Syntheses Starting materials. 1,4,5,8,9,12-Hexaazatriphenylene (HAT) was prepared according to the published procedure.1 The starting materials FeSO4 ·7H2 O (99%; Fisher Scientific) and anhydrous FeCl2 (99.9%; Strem) were used as received. [Fe3 (HAT)(H2 O)12 ](SO4 )3 ·3.3H2 O (1). Solid HAT (67.0 mg, 0.29 mmol) was added to a solution of FeSO4 ·7H2 O (461 mg, 1.66 mmol) in 50 mL of water to give a purple solution which was stirred for 30 min and then filtered. Purple crystals were obtained by slow evaporation of the filtrate. Yield 69.2 mg (25%). Due to the large excess of FeSO4 used in the reaction, an analytically pure sample could not be obtained in spite of the fact that the samples were washed with a copious amount of methanol which served to dissolve most of the unreacted FeSO4 . [Fe2 (HAT)(SO4 )(H2 O)5 ](SO4 )·2H2 O·0.8CH3 OH (2). The reaction between 25.0 mg (0.11 mmol) of HAT and 184 mg (0.66 mmol) of FeSO4 ·7H2 O in 25 mL of water was carried out in an analogous fashion to the procedure described for compound 1. Slow vapor diffusion of methanol into the purple solution, carried out in a closed H-tube, resulted in the growth of purple needle-shaped crystals along the walls of the tube and precipitation of a small quantity of yellow unreacted HAT at the bottom of the tube. Due to the separate physical location of the products in the glass vessel, it was possible to manually isolate the purple crystals from the rest of the mixture. Yield 14.7 mg (20%). Anal. Calc. for Fe2 S2 O15.8 N6 C12.8 H23.2 : S, 9.30; N, 12.18; C, 22.29; H, 3.39. Found: S, 9.13; N, 12.01; C, 22.27; H, 3.63%. The crystals could also be grown by layering methanol on top of the red-purple solution in a thin 5 mm i.d. tube, but in such cases a few plate-like crystals of [Fe2 (HAT)(SO4 )(H2 O)5 ](SO4 )·3H2 O (3) were also obtained as a minor by-product. The latter can be considered as a polymorph of 2 (vide infra).

Scheme 1

This journal is

We report herein the syntheses, crystal structures, and magnetic properties of four new compounds formed from reactions of iron(II) salts with HAT. The first of the reported compounds, [Fe3 (HAT)(H2 O)12 ](SO4 )3 ·3.3H2 O (1) is, to our knowledge, the only example of a discrete paramagnetic cluster based on the M3 (l3 -HAT) core. The other three complexes exhibit infinite one-dimensional structures.

[Fe3 Cl5 (HAT)(CH3 OH)4 (H2 O)]Cl (4). The synthesis was carried out under anaerobic conditions using conventional Schlenk-line techniques. A solution of anhydrous FeCl2 (198 mg,

The Royal Society of Chemistry 2005

Dalton Trans., 2005, 1897–1902

1897

Table 1 Crystal data and details of the structure determination for compounds 1–4

Formula

Fe3 S3 O27.3 N6 C12 H36.6 (1)

Fe2 S2 O15.8 N6 C12.8 H23.2 (2)

Fe2 S2 O16 N6 C12 H22 (3)

Space group ˚ a/A ˚ b/A ˚ c/A a/◦ b/◦ c /◦ ˚3 V /A Z qcalc /g cm−3 l/mm−1 Crystal color and habit Crystal size/mm T/K ˚ Radiation, k/A Min. and max. h/◦ Reflections collected Independent reflections Data/parameters/restraints R [F o > 4r(F o )]

P21 /c (No. 14) 18.10(2) 12.469(8) 15.39(1)

P21 /c (No. 14) 6.5269(8) 29.457(4) 12.500(2)

P21 /c (No. 14) 6.4492(5) 25.388(2) 14.183(1)

97.46(1)

91.396(2)

90.952(5)

3443(5) 4 1.850 1.537 Purple plate 0.40 × 0.15 × 0.02 110 Mo-Ka, 0.71073 1.99 to 23.26 8782 2456 [Rint = 0.0464] 2456/224/64 R1 = 0.0983 wR2 = 0.1992 1.246 1.033 and −1.208

2402.7(5) 4 1.908 1.471 Purple needle 0.45 × 0.13 × 0.05 110 Mo-Ka, 0.71073 1.38 to 26.37 22199 4913 [Rint = 0.0411] 4913/400/24 R1 = 0.0409 wR2 = 0.0912 1.073 1.408 and −0.642

2322.0(3) 4 1.951 12.564 Purple plate 0.36 × 0.15 × 0.02 110 Cu-Ka, 1.54178 3.48 to 59.10 18627 3285 [Rint = 0.0411] 3285/376/24 R1 = 0.0443 wR2 = 0.1023 1.062 0.520 and −0.799

G.o.f. on F 2 ˚ −3 Max. and min. resd. dens/e A

1.56 mmol) in 10 mL of methanol was quickly added to a suspension of HAT (102 mg, 0.44 mmol) in 40 mL of dichloromethane which led to the instantaneous formation of a dark blue solution. The reaction was stirred for 2 hours, filtered, and diethyl ether (50 mL) was added to the filtrate to produce a blue solid, which was removed by filtration, washed with cold methanol (10 ml) followed by diethyl ether (2 × 20 mL), and dried in vacuo for 3 h. Yield 187 mg (56%). Anal. Calc. for Fe3 Cl6 O5 N6 C16 H24 : N, 11.05; C, 25.26; H, 3.18. Found: N, 10.68; C, 25.51; H, 3.44%. Dark blue crystals of the compound were obtained by dissolving the parent solid in a CH3 OH/CH2 Cl2 mixture (1 : 4 v/v) and carefully layering diethyl ether on top in a thin 10 mm i.d. Schlenk tube. X-Ray crystallography General. The crystals that were selected for study were suspended in polybutene oil (Aldrich) and mounted on cryoloops that were placed in a N2 cold stream at 110(2) K. Single crystal X-ray data were collected on a Bruker SMART 1000 CCD diffractometer for 1, 2, and 4 and on a Bruker GADDS D8 diffractometer for 3. The data sets were recorded as three xscans of 606 frames each, at 0.3◦ step width and integrated with the Bruker SAINT9 software package. The absorption correction (SADABS10 ) was based on fitting a function to the empirical transmission surface as sampled by multiple equivalent measurements. Solution and refinements of the crystal structures were carried out using the SHELX11 suite of programs and X-SEED,12 a graphical interface. A summary of pertinent information relating to unit cell parameters, data collection, and refinements are given in Table 1. The metal-to-ligand bond distances are listed in Table 2. Complete listings of atomic and thermal parameters, bond distances and bond angles, as well as the ORTEP plots of asymmetric units, are available as ESI.† [Fe3 (HAT)(H2 O)12 ](SO4 )3 ·3.3H2 O (1). The data set was indexed in a monoclinic unit cell and systematic absences indicated the space group P21 /c. Structure solution by direct methods resolved positions of all Fe and S atoms, as well as some atoms of the lighter elements. The remaining non-hydrogen atoms and hydrogen atoms of the coordinated water molecules were located by the least squares refinement. The sulfur atoms S(3) and S(4) of two of the crystallographically independent [SO4 ]2− anions are located on special positions 2d and 2a, respectively, which leads 1898

Dalton Trans., 2005, 1897–1902

Fe3 Cl6 O9.1 N6 C20 H40.2 (4·4CH3 OH·0.12H2 O) P1¯ (No. 2) 7.9876(9) 16.123(2) 17.516(2) 72.224(3) 80.061(3) 80.739(3) 2101.7(4) 2 1.394 1.447 Dark-blue block 0.28 × 0.15 × 0.11 110 Mo-Ka, 0.71073 1.34 to 26.37 21685 8551 [Rint = 0.0566] 8551/454/13 R1 = 0.0649 wR2 = 0.1945 1.078 1.253 and −1.167

˚ ) in compounds 1–4 Table 2 Metal-to-ligand bond distances (A Fe3 S3 O27.3 N6 C12 H36.6 (1) Fe(1)–N(1) 2.137(10) Fe(1)–N(2) 2.224(9) Fe(1)–O(17) 2.177(10) Fe(1)–O(18) 2.101(8) Fe(1)–O(19) 2.031(6) Fe(1)–O(20) 2.147(8) Fe(2)–N(3) 2.249(9) Fe(2)–N(4) 2.312(9) Fe(2)–O(21) 2.041(7)

Fe(2)–O(22) Fe(2)–O(23) Fe(2)–O(24) Fe(3)–N(5) Fe(3)–N(6) Fe(3)–O(25) Fe(3)–O(26) Fe(3)–O(27) Fe(3)–O(28)

2.170(8) 2.083(8) 2.114(7) 2.237(9) 2.242(11) 1.936(12) 2.042(9) 2.072(9) 2.258(10)

Fe2 S2 O15.8 N6 C12.8 H23.2 (2) Fe(1)–N(1) 2.252(2) Fe(1)–N(2) 2.211(2) Fe(1)–O(1) 2.107(2) Fe(1)–O(2) 2.057(2) Fe(1)–O(3) 2.094(2) Fe(1)–O(4) 2.119(2)

Fe(2)–N(3) Fe(2)–N(4) Fe(2)–O(5) Fe(2)–O(6) Fe(2)–O(7) Fe(2)–O(8)

2.224(2) 2.206(2) 2.088(2) 2.127(2) 2.137(2) 2.028(2)

Fe2 S2 O16 N6 C12 H22 (3) Fe(1)–N(1) 2.261(3) Fe(1)–N(2) 2.205(3) Fe(1)–O(10) 2.050(3) Fe(1)–O(11) 2.116(3) Fe(1)–O(12) 2.086(2) Fe(1)–O(13) 2.119(3)

Fe(2)–N(3) Fe(2)–N(4) Fe(2)–O(6) Fe(2)–O(7) Fe(2)–O(8) Fe(2)–O(9)

2.205(3) 2.212(3) 2.134(2) 2.018(2) 2.107(2) 2.122(2)

Fe3 Cl6 O9.12 N6 C20 H40.24 (4·4CH3 OH·0.12H2 O) Fe(1)–Cl(1) 2.448(2) Fe(2)–O(2) Fe(1)–Cl(1) 2.504(2) Fe(2)–O(3) Fe(1)–Cl(2) 2.348(2) Fe(2)–O(4) Fe(1)–N(1) 2.256(4) Fe(3)–Cl(4) Fe(1)–N(2) 2.217(4) Fe(3)–Cl(5) Fe(1)–O(1) 2.128(5) Fe(3)–Cl(5) Fe(2)–Cl(3) 2.356(2) Fe(3)–N(5) Fe(2)–N(3) 2.231(5) Fe(3)–N(6) Fe(2)–N(4) 2.267(5) Fe(3)–O(5)

2.135(6) 2.160(7) 2.233(6) 2.360(2) 2.430(2) 2.511(2) 2.255(4) 2.210(4) 2.140(6)

to two different symmetry-related orientations for the disordered O atoms whose occupancies were fixed at 0.5. Hydrogen atoms of the HAT molecule were placed at calculated positions. The final refinement was carried out with anisotropic displacement parameters for only the Fe and S atoms due to the poor quality of the data related to twinning and loss of solvent. The final value of R1 was 0.0983. [Fe2 (HAT)(SO4 )(H2 O)5 ](SO4 )·2H2 O·0.8CH3 OH (2). The compound crystallizes in the monoclinic space group P21 /c. The

Fe and S atoms, as well as some of the other lighter atoms were located by direct methods. Subsequent cycles of least squares refinements and difference Fourier maps led to the location of the remaining non-hydrogen atoms and hydrogen atoms belonging to water molecules. Hydrogen atoms belonging to the aromatic rings were placed at calculated positions. Two different [SO4 ]2− anions were found in the crystal structure, one of which bridges three Fe(2) atoms from different asymmetric units and the other of which serves as an outer-sphere anion. The latter displays rotational disorder which was modeled by using two different orientations around the central S(2) atom; the total occupancy of both orientations was fixed at unity. The final refinement was performed with anisotropic thermal parameters for all nonhydrogen atoms, except for disordered oxygen atoms of the [SO4 ]2− counter-anion, resulting in R1 = 0.0409.

involves water which competes with HAT as a ligand. In the case of 4, CH3 OH and Cl− are competing with HAT as ligands which, unlike H2 O, do not reversibly labilize the HAT ligand. Crystal structures [Fe3 (HAT)(H2 O)12 ](SO4 )3 ·3.3H2 O, 1. The structure consists of discrete trinuclear clusters [Fe3 (l3 -HAT)(H2 O)12 ]6+ (Fig. 1) in which HAT acts as a tris-chelating ligand. [SO4 2− ] anions and uncoordinated water molecules stabilize the structure by forming an extensive network of hydrogen bonds that also involves the coordinated water molecules.

[Fe2 (HAT)(SO4 )(H2 O)5 ](SO4 )·3H2 O (3). The space group was found to be monoclinic, P21 /c. Solution and refinement of the structure was carried out as described for compound 2. No disorder was observed for any portion of the asymmetric unit of 3. All non-hydrogen atoms were introduced into the final refinement with anisotropic thermal parameters, except for atoms C(8), C(10), and C(11) which had to be refined isotropically. This led to the final value of R1 = 0.0443. [Fe3 Cl5 (HAT)(CH3 OH)4 (H2 O)]Cl·4CH3 OH·0.12H2 O (4·4CH3 OH·0.12H2 O). The crystals belong to the triclinic space ¯ The initial solution resolved most of the nongroup P1. hydrogen atoms, and the remaining non-hydrogen atoms and hydrogen atoms of water molecules were located from difference Fourier maps. Hydrogen atoms belonging to the HAT and methanol molecules were placed at calculated positions. The final refinement was carried out with all non-hydrogen atoms being refined anisotropically, except for some atoms of the disordered solvent molecules. The final R1 was 0.0649. CCDC reference numbers 264315–264318. See http://www.rsc.org/suppdata/dt/b5/b502567b/ for crystallographic data in CIF or other electronic format. Magnetic measurements Magnetic susceptibility and magnetization measurements were carried out with a Quantum Design SQUID magnetometer MPMS-XL. DC magnetic measurements were performed with an applied field of 1000 G in the 2–300 K temperature range. The data were corrected for the diamagnetic contributions calculated from the Pascal constants.13

Results and discussion Syntheses The HAT ligand reacts instantaneously with an aqueous solution of FeSO4 ·7H2 O to form the trinuclear cluster 1. The compound is stable in solution only in the presence of excess FeSO4 . Attempts to re-dissolve the crystalline product in water resulted in decomposition to the starting materials. Diffusion of methanol into the original aqueous solution of 1, stabilized by the excess FeSO4 , resulted in decomposition of 1 and formation of 2 with partial precipitation of HAT. Both compounds 1 and 2 are insoluble in organic solvents. A different source of FeII , namely FeCl2 , reacts with HAT to give a dark blue solution from which the one-dimensional polymer 4 was crystallized. The solution of 4 is air-sensitive and slowly turns green upon exposure to air. The compound dissolves in methanol, but addition of water results in decomposition to the starting materials as evidenced by precipitation of the HAT ligand. A large excess of the Fe(II) salt is necessary in order to stabilize compound 1 as compared to 4, although both products contain the Fe3 (l3 -HAT) structural fragment. Preparation of 1, however,

Fig. 1 Trinuclear cation [Fe3 (l3 -HAT)(H2 O)12 ]6+ in the crystal structure of 1 (hydrogen atoms are not shown).

˚, The Fe–N bond lengths vary from 2.137(10) to 2.312(9) A while the N–Fe–N angles are in the range of 73.3(4)–78.6(3)◦ . It must be noted that the N–Fe–N angle values reported previously for Fe(II) atoms coordinated by a mono-chelating ligand such as bipyridine (bpy) are larger and usually exceed 80◦ .14 Apparently, the greater rigidity of the HAT ligand results in a more severe distortion of the octahedral environment of the metal ions. The substantial strain of the Fe3 (l3 -HAT) core taken together with the fact that HAT is not a strong ligand explains the instability of the trinuclear complex towards replacement of HAT with water molecules. [Fe2 (HAT)(SO4 )(H2 O)6 ](SO4 )·2H2 O·0.8CH3 OH (2) and [Fe2 (HAT)(SO4 )(H2 O)6 ](SO4 )·3H2 O (3). Crystal structures of 2 and 3 reveal that both crystallize as chains. Each HAT molecule is coordinated to only two FeII ions, hence it is a bisrather than a tris-chelating ligand. The dinuclear Fe2 (l2 -HAT) moiety is less strained than the trinuclear cluster in the structure of 1, which translates into a higher stability for 2 and 3. The Fe2 (l2 -HAT) fragments are connected into an infinite chain through [SO4 ]2− anions (Fig. 2) each of which uses three of its O atoms to bridge three Fe(2) atoms from different Fe2 (l2 -HAT) moieties which leads to the Fe(2) atoms being bound to three different [SO4 ]2− anions. The Fe(1) atom is coordinated to one binding site of a HAT ligand and four water molecules.

Fig. 2 The one-dimensional chain in the crystal structure of 2 (hydrogen atoms have been omitted for the sake of clarity). Dalton Trans., 2005, 1897–1902

1899

The Fe–N distances fall into a narrow range of 2.206(2)– ˚ for 2 and 2.205(3)–2.261(3) A ˚ for 3, similar to what 2.252(2) A was found for 1. The N–Fe–N angles (75.33(7)–75.87(7)◦ for 2 and 74.7(1)–75.7(1)◦ for 3) are also more acute than the typical values reported for Fe(II) compounds with bpy ligands (vide supra). The chains are running parallel to the a axis in both 2 and 3. Additional stabilization of the structure is provided through p–p ˚ for 2 and 3.20 A ˚ interactions (with interplanar spacing of 3.19 A for 3) between the HAT molecules of different chains; these result in the formation of layers that are parallel to the ac plane and separated by [SO4 ]2− counter-ions and water molecules. The difference between the structures of 2 and 3 can be appreciated when both are viewed as projections along the a axis in the direction of chain propagation (Fig. 3). In both cases the layers run parallel to the ac plane, but in 3 these layers are flat whereas in 2 they produce a zigzag pattern when viewed down the a axis. The other stabilizing factor is extensive hydrogen bonding between the [SO4 2− ] anions and uncoordinated and coordinated water molecules. The zigzag arrangement of the layers in 2

involves more extensive hydrogen bonding between the chains and interstitial molecules and anions, thus leading to higher stability of this architecture over the structure of 3. [Fe3 Cl5 (HAT)(CH3 OH)4 (H2 O)]Cl·4CH3 OH·0.12H2 O (4·4CH3 OH·0.12H2 O) As in the case of the crystal structure of 1, the trinuclear Fe3 (l3 -HAT) fragment is present in the structure of 4. While the former is a molecular compound, in the latter the trinuclear fragments are connected into an extended 1-D structure through Cl− ions (Fig. 4a). Out of the three Fe atoms coordinated to the HAT ligand, only Fe(1) and Fe(3) participate in the formation of the extended structure via Fe(l2 -Cl)2 Fe bridging. The third atom, Fe(2), is terminal with the coordination environment being completed by one Cl− , one H2 O and two CH3 OH molecules. Presence of different types of FeII ions results in a lower dimensionality for the structure as compared to Co3 (HAT)[N(CN)2 ]6 (OH2 )2 4 in which all of the Co atoms of the Co3 (l3 -HAT) unit participate in the extended structure by forming bonds to the bridging dicyanamide anions. The result is a three-dimensional framework.

Fig. 4 Infinite chain in the crystal structure of 4 (a) and perspective view of the packing of chains (b) (green: iron, yellow: chlorine, red: oxygen, blue: nitrogen, gray: carbon; hydrogen atoms and interstitial solvent molecules have been omitted for the sake of clarity).

Fig. 3 Packing of the one-dimensional chains in the crystal structures of 2 (a) and 3 (b) viewed down the a axis (hydrogen atoms and interstitial solvent molecules have been omitted for the sake of clarity). Different configuration of the layers of the chains with p–p interactions running parallel to the ac plane can be seen. 1900

Dalton Trans., 2005, 1897–1902

As expected, the Fe–Cl bond lengths for the bridging Cl− ˚ ) are longer than for the terminal anions (2.430(2)–2.511(2) A ˚ ). The values are comparable to those ones (2.348(2)–2.360(2) A reported for the similar coordination of FeII in the literature: ˚ for terminal16 ˚ for bridging15 and 2.379–2.388 A 2.427–2.511 A Fe–Cl bonds. The distance between the Fe atoms in the Fe(l2 ˚ . The Fe– Cl)2 Fe fragment varies from 3.557(2) to 3.601(2) A ˚ are typical. The N distances in the range 2.210(4)–2.267(5) A Fe–O distances for coordinated methanol molecules (2.128(5)– ˚ ) are slightly shorter than the Fe–O distance of 2.160(7) A ˚ for the coordinated water molecule. 2.233(6) A

¯ The chains extend along the [101] direction and are stacked above each other along the b axis (Fig. 4b). The configuration of each chain resembles a zigzag staircase. The chains are held together by hydrogen bonding between coordinated solvent molecules. The non-coordinated solvent molecules also engage in hydrogen bonding providing additional stability to the structure. No p–p interactions between the HAT molecules are observed in the structure of 4. Magnetic properties Magnetic susceptibility measurements were performed on polycrystalline samples of 1, 2, and 4 at 1000 G. For all three compounds the value of vT at 300 K is higher than the expected value for a spin-only case (high-spin Fe(II), S = 2) which can be explained by the well-known orbital contribution of high spin Fe(II).13 The temperature dependence of 1/v in the temperature range 2–300 K approximates Curie–Weiss behavior. The negative Curie constants indicate antiferromagnetic coupling between metal atoms. [Fe3 (HAT)(H2 O)12 ](SO4 )3 ·3.3H2 O, (1). The room temperature vT value of 12.7 emu K mol−1 begins to decrease slowly below 100 K and then decreases faster below 40 K (Fig. 5a). The temperature dependence of 1/v follows Curie–Weiss behavior with C = 12.7 and h = −3 K. Such behavior is indicative of antiferromagnetic coupling between the three Fe(II) centers. The theoretical fit was performed using the spin Hamiltonian with isotropic exchange interaction J: H = −2J(S1 S2 + S2 S3 + S3 S1 ) where S1 , S2 , and S3 refer to the spin on the first, second, and third FeII center, respectively. A fitting of the experimental curve with the program MAGPAK17 led to the values of J = −0.22 cm−1 and g = 2.38. The low value of J indicates very weak antiferromagnetic coupling of FeII centers through the HAT molecule which is consistent with the weak exchange interactions observed for other transition metal compounds involving HAT.4 [Fe2 (HAT)(SO4 )(H2 O)6 ](SO4 )·2H2 O·CH3 OH (2). The vT value of 7.3 emu K mol−1 at room temperature remains constant down to 150 K, after which temperature it begins to decrease as the temperature is lowered (Fig. 5b). The 1/v vs. T curve follows the Curie–Weiss law with C = 7.45 and h = −7.5 K. In compound 4, which also contains the Fe2 (l2 -HAT) fragment, the coupling of the two Fe(II) centers through the HAT bridge is very weak (vide infra). Therefore, this interaction in compound 2 is assumed to be much weaker than magnetic coupling between Fe(II) centers bridged by [SO4 2− ] anions, which is considered to be the dominant magnetic interaction. Such an assumption is reasonable since the closest Fe · · · Fe distance along the chain is ˚ . Therefore, the magnetic behavior of the compound 4.803(1) A was modeled with a linear Heisenberg chain formed by Fe(2) atoms plus contribution from the non-interacting Fe(1) atoms: vT =

Ng2 l2B S (S + 1) 1 − u Ng2 l2B S (S + 1) · + 3k 1+u 3k

where u = coth[2JS(S + 1)/kT]–1/[2JS(S + 1)].18 Fitting the experimental dependence of vT using this expression led to J = −2.5 cm−1 and g = 2.24. The antiferromagnetic exchange constant for compound 2 is thus an order of magnitude higher than the J value obtained for compound 1. The stronger antiferromagnetic exchange in the former case can also be observed in the temperature dependence of the vT product which begins to decrease at higher temperature as compared to the latter. [Fe3 Cl5 (HAT)(CH3 OH)4 (H2 O)]Cl (4). The observed vT value at room temperature is 11.5 emu Kmol−1 . As the temperature is lowered, the value of vT decreases first slowly and

Fig. 5 Temperature dependence (circles) and theoretical fit (solid line) of the vT product and temperature dependence (squares) and Curie–Weiss fit (dashed line) of the inverse magnetic susceptibility 1/v for compounds 1 (a), 2 (b), and 4 (c).

then faster below 50 K (Fig. 5c). The temperature dependence of 1/v obeys the Curie–Weiss law (C = 11.8 and h = −8.0 K). Given the demonstrated weak nature of the exchange interactions through the HAT ligand and the much higher reported values for exchange parameters of two FeII centers bridged by Cl− ions,19 the magnetic model we chose to use is that of isolated Heisenberg dimers Fe(l2 -Cl)2 Fe in which antiferromagnetic exchange between the two FeII centers occurs. The expression for the total value of magnetic susceptibility20 also takes into account the contribution from the uncoupled Fe(2) atoms: v=A

2e2x + 10e6x + 28e12x + 60e20x + 6A 1 + 3e2x + 5e6x + 7e12x + 9e20x

where A = Ng2 l/kT and x = J/kT. The best fit was achieved with the values of J = −2.4 cm−1 and g = 2.30. The magnitude of the antiferromagnetic exchange in this case is comparable to that observed for compound 2. Dalton Trans., 2005, 1897–1902

1901

Conclusions The present study reports a route to new molecular materials based on the use of the HAT ligand as a tris- or bis-chelating ligand. In principle, the molecular trimer observed in the structure of 1 represents an attractive model compound which can provide deeper insight into the phenomenon of spin frustration. The magnitude of the magnetic superexchange, however, is rather small. As a result, the effects of magnetic coupling are observed only at very low temperatures and do not argue a good case for more detailed studies. Moreover, the insolubility of compound 1 in all common solvents except water and its instability in aqueous solution precludes its use as a building block in further chemistry. The preparation of other trimeric clusters with HAT that would be soluble in solvents other than water was sought, but the results with FeCl2 led to the extended one-dimensional compound 4. Future research in this area will involve the use of other metal ion precursors including those of heavier transition metals whose size is more compatible with the bite angle of the HAT ligand and whose orbital interactions with the N donors may favor increased magnetic exchange interactions. The results of these studies will be reported in due course.

Acknowledgements This research was supported by NSF (PI grant CHE-9906583) and DOE (grant DE-FG03-02ER45999). Funding of the CCD diffractometer (CHE-9807975) and the SQUID instrumentation (NSF-9974899) by NSF is gratefully acknowledged.

References 1 M. S. P. Sarma and A. W. Czarnik, Synthesis, 1988, 72–73; J. T. Rademacher, K. Kanakarajan and A. W. Czarnik, Synthesis, 1994, 378–380.

1902

Dalton Trans., 2005, 1897–1902

2 P. Didier, I. Ortmans, A. Kirsch-De Mesmaeker and R. J. Watts, Inorg. Chem., 1993, 32, 5239–5245; T. J. Rutherford, O. Van Gijte, A. Kirsch-De Mesmaeker and F. R. Keene, Inorg. Chem., 1997, 36, 4465–4474. 3 P. N. W. Baxter, J. M. Lehn, G. Baum and D. Fenske, Chem. Eur. J., 1999, 5, 102–112. ´ and K. R. Dunbar, Chem. Commun., 2001, 4 J. R. Gal´an-Mascaros 217–218; S. R. Marshall, A. L. Rheingold, L. N. Dawe, W. W. Shum, C. Kitamura and J. S. Miller, Inorg. Chem., 2002, 41, 3599–3601. 5 J. P. Lecomte, A. Kirsch-De Mesmaeker, M. M. Feeney and J. M. Kelly, Inorg. Chem., 1995, 34, 6481–6491; C. Moucheron, A. KirschDe Mesmaeker and S. Choua, Inorg. Chem., 1997, 36, 584–592. 6 R. Nasielski-Hinkens, M. Benedek-Vamos, D. Maetens and J. Nasielski, J. Organomet. Chem., 1981, 217, 179–182. 7 P. Didier, L. Jacquet, A. Kirsch-De Mesmaeker, R. Hueber and A. Van Dorsselaer, Inorg. Chem., 1992, 31, 4803–4809. 8 T. J. Rutherford and F. R. Keene, Inorg. Chem., 1997, 36, 3580–3582. 9 SMART and SAINT, Siemens Analytical X-ray Instruments Inc., Madison, WI, 1996. ¨ ¨ 10 G. M. Sheldrick, SADABS, University of Gottingen, Gottingen, Germany, 1996. 11 G. M. Sheldrick, SHELXS-97 and SHELXL-97, University of ¨ ¨ Gottingen, Gottingen, Germany, 1997. 12 L. J. Barbour, J. Supramol. Chem., 2001, 1, 189–191. 13 O. Kahn, Molecular Magnetism, Wiley-VCH, 1993. 14 M. Nieuwenhuyzen, B. Bertram, J. F. Gallagher and J. G. Vos, Acta Crystallogr., Sect. C, 1998, 54, 603–606; W. F. Liaw, J. H. Lee, H. B. Gau, C. H. Chen and G. H. Lee, Inorg. Chim. Acta, 2001, 322, 99–105. 15 N. R. Strel’tsova, A. B. Tarasov, V. K. Bel’skii and B. M. Bulychev, Zh. Obshch. Khim., 1990, 60, 2337–2342; S. I. Troyanov, M. Feist and E. Kemnitz, Z. Anorg. Allg. Chem., 1999, 625, 806–812. 16 J. S. Sun, H. Zhao, X. Ouyang, R. Cl´erac, J. A. Smith, J. M. Clemente´ Juan, C. Gomez-Garcia, E. Coronado and K. R. Dunbar, Inorg. Chem., 1999, 38, 5841–5855. 17 J. J. Borr´as-Almenar, J. M. Clemente-Juan, E. Coronado and B. S. Tsukerblat, J. Comput. Chem., 2001, 22, 985–991. 18 T. Smith and S. A. Friedberg, Phys. Rev., 1968, 176, 660–665. 19 Y. Zang, H. G. Jang, Y. M. Chiou, M. P. Hendrich and J. Lawrence, Inorg. Chim. Acta, 1993, 213, 41–48. 20 C. J. O’Connor, Prog. Inorg. Chem., 1982, 29, 203–283.