Magnetostructural study of substituted a-nitronyl aminoxyl radicals with chlorine and hydroxy groups as crystalline design elements Oriol Ju�rgens, Joan Cirujeda, Montse Mas, Ignasi Mata, Araceli Cabrero, Jose� Vidal-Gancedo, Concepcio� Rovira, Elies Molins and Jaume Veciana* Institut de Cie`ncia deMaterials de Barcelona (CSIC), Campus de la UAB, 08193-Bellaterra, Barcelona, Spain We present a new family of phenyl substituted a-nitronyl aminoxyl radicals which contain hydroxy- and chlorine-substituents as crystal engineering tools.The magnetic behaviour of these radicals strongly diers in dimensionality and strength, showing in all cases antiferromagnetic interactions. We have determined the X-ray crystal structure and analysed the crystal packings of these radicals. From this analysis all the observed magnetic properties can be conveniently rationalized by only considering the close contacts of NO groups of neighbouring molecules according to the generally accepted mechanisms for intermolecular magnetic interactions.Beside the strong OMH,O(MN) hydrogen bonds and the weaker CMH,O(MN) hydrogen bonds, weak Cl,H bonds also seem to play a significant role in determining the molecular arrangement in the solid state and, therefore, the magnetic properties.In only one case are close Cl,Cl contacts observed, pointing to attractive interactions between chlorine atoms. Magnetic properties of molecular solids depend both on the molecular electronic properties and on the intermolecular electronic interactions present in the solid state.Since the discovery of the b-phase of 4-nitrophenyl a-nitronyl aminoxyl,† the first example of a purely organic free radical with a bulk ferromagnetic transition,1 much work has been devoted to the design of new substituted a-nitronyl aminoxyl radicals as building blocks for new molecular magnetic materials and bonding through OH substituents has been demonstrated to be a powerful crystal engineering element of a-nitronyl aminoxyl radicals providing new supramolecular architectures with relevant magnetic properties.2 One of these new molecular solids was obtained with 4-hydroxyphenyl a-nitronyl aminoxyl radical.3 This radical in the solid state forms a two-dimensional network built up by OMH,O(MN) and CMH,O(MN) hydrogen bonds that explains its quasi-two-dimensional ferromagnetic behaviour.On the other hand the 2-hydroxyphenyl and 2,5-dihydroxyphenyl a-nitronyl aminoxyl radicals4,5 undergo bulk ferromagnetic transitions at 0.45 and 0.5 K, respectively, being two of the rare examples of purely organic Cl N N+ O• O– N N+ O• O– Cl N N+ O• O– N N+ O• O– OH OH HO N N+ O• O– 1 2 3 4 5 Cl Cl Cl ferromagnets. Beside their crystal packing, which is controlled by a complex network of weak CMH,O(MN) hydrogen Radicals with only one Cl atom at the ortho and meta bonds, the twist angles between the phenyl rings and the mean positions, radicals 1 and 2, have also been studied here as planes defined by the ONCNO groups seem to play a signifi- reference compounds in order to evaluate the role played by cant role in determining this unusual magnetic property. this bulky and electroactive atom in crystal packing.Therefore such results clearly show that the control of the Interestingly, the radical with one Cl atom at the ortho position molecular conformation and the crystal packing are key points also provides an opportunity to evaluate the eect of a large torsion between the two rings of the radical without introduc- in molecular magnetism. ing any strong OMH,O(MN) hydrogen bonds.Chlorine atoms have long been known for their steering ability in crystal engineering of molecular solids. Attractive Cl,Cl interactions and CMH,Cl hydrogen bonds have been described as responsible for this ability.6,7 Herein we report a Experimental detailed magnetostructural study of a new family of phenyl anitronyl aminoxyl radicals 3–5 that combine OH groups and General procedures Cl atoms, attached to dierent positions of the phenyl rings, Radicals 1–5 were prepared using the procedure described by as crystalline design elements.The combination of these two Ullman and co-workers.10 The 2,3-(dihydroxylamino)-2,3- crystalline design elements has provided an eective way to dimethylbutane used as a precursor was obtained by following modulate the crystal packing of several chlorine substituted the reported procedure.11 Melting points were determined by phenols,8,9 leading to interesting applications to magnetic dierential scanning calorimetry (Perkin-Elmer, DSC-7 calormolecular solids.imeter) and are given as the maxima of the observed peaks.All radicals containing OH groups, radicals 3–5, melt with decomposition. IR (Nicolet 710 FT-IR spectrometer) and UV–VIS spectra (Cary 5 UV–VIS-NIR spectrometer) of the † a-Nitronyl aminoxyl is used throughout to indicate 4,5-dihydro- 4,4,5,5-tetramethyl-3-oxido-1H-imidazol-3-ium-1-oxyl. synthesized radicals were also recorded. J. Mater. Chem., 1997, 7(9), 1723–1730 1723Synthesis of free radicals Superconducting SQUID susceptometer and using microcrystalline samples (80–115 mg) of the radicals 1–5.The diamag- 2-(2-Chlorophenyl)-4,5-dihydro-4,4,5,5-tetramethyl-3-oxidonetic contributions of the sample holder and the radicals were 1H-imidazol-3-ium-1-oxyl 1. 2,3-(Dihydroxylamino)-2,3- determined by extrapolation from the xT vs. T plots in the dimethylbutane (2.11 g; 14.2 mmol) was added to a stirred high-temperature range and were used later to correct the solution of 2-chlorobenzaldehyde (2 g; 14.2 mmol) in 30 ml of SQUID outputs.methanol. Stirring at room temp. was continued for 20 h and the resulting white precipitate was filtered o and dried in X-Ray measurements vacuo. This solid was oxidized with a solution of NaIO4 (1.5 g; 7.1 mmol) in 50 ml of water at 5 °C and extracted with X-Ray data for single crystals of 1, 2, 4 and 5 were collected at 293 K on an Enraf-Nonius CAD 4 FR-590 diractometer dichloromethane.The solution was evaporated and the crude product was purified by column chromatography (SiO2) with working at 1 kW with monochromatic Mo-Ka (l=0.71069 A ° ) radiation. Data were collected by using an v/2h scan method.ethyl acetate dichloromethane (151) as eluent (2.55 g; yield, 67% from the aldehyde). Single crystals of 1 were grown by The structures were all refined by a full-matrix least squares method which minimized Sw(DF)2.‡ The presence of dierent evaporation at room temp. from a toluene solution. Mp 168.3 °C (Found: C, 58.27; H, 6.02; N, 10.42.Calc. for polymorphs in each crystalline material used for magnetic measurements was ruled out by means of powder X-ray C13H16N2O2Cl: C, 58.32; H, 6.02; N, 10.46%); nmax/cm-1 (KBr) 1595w, 1449m, 1404s, 1367s, 1211w, 1171m, 1133m, 1055m, diraction spectra by comparing the experimental spectra with the simulated ones based on the single crystal X-ray diraction 766m; UV–VIS (CH2Cl2) lmax/nm (e): 354 (18 000), 554 (780); MS (EI) m/z: 267 (M+), 179, 138, 114, 84, 69, 56.structure. These spectra were simulated by using the CERIUS2 2.0 program (Molecular Simulations Inc.). Powder diraction 2-(3-Chlorophenyl)-4,5-dihydro-4,4,5,5-tetramethyl-3-oxido- spectra were collected on a Rigaku Dimax RC-200 1H-imidazol-3-ium-1-oxyl 2. Radical 2 was synthesized by the diractometer with a 12 kW rotating anode generator and a same procedure as 1.Crystals were grown by slow evaporation monochromator of single crystalline graphite for Cu-Ka of a heptane–dichloromethane (1051) solution at room temp. radiation. Mp 123.5 °C (Found: C, 58.30; H, 6.01; N, 10.40. Calc. for C13H16N2O2Cl: C, 58.32; H, 6.02; N, 10.46%); yield, 92% from EPR spectroscopic measurements the aldehyde; nmax/cm-1 (KBr) 1580m, 1418m, 1395m, 1364s, The EPR spectra of radicals 1–5 in toluene solutions under 1134m, 795m; UV–VIS (CH2Cl2) lmax/nm (e): 271 (16 000), 367 free tumbling conditions were recorded on a Bruker ESP-300E (18 000), 584 (480); MS (EI) m/z: 267 (M+), 179, 138, 114, spectrometer operating in the X-band (9.3 GHz) with a rec- 84, 69.tangular TE1ity and equipped with a field-frequency (F/F) lock accessory and a built-in NMR gaussmeter.Signal- 2-(2-Chloro-4-hydroxyphenyl)-4,5-dihydro-4,4,5,5- to-noise ratio was increased by accumulation of scans using tetramethyl-3-oxido-1H-imidazol-3-ium-1-oxyl 3. Radical 3 was the F/F lock accessory to guarantee a high-field reproducibility. obtained in a similar way to 1, but instead of stirring the Precautions to avoid undesirable spectral line broadening such reactants, they were refluxed in benzene for 19 h.All attempts as that arising from microwave power saturation and magnetic to grow large single crystals of radical 3 failed, and so its X- field overmodulation were taken. In order to avoid dipolar ray structure could not be determined. Mp 166.2 °C (decomp.) broadening, the radical solutions were carefully degassed by (Found: C, 55.21; H, 5.75; N, 9.70.Calc. for C13H16N2O3Cl: C, bubbling with pure argon. 55.03; H, 5.68; N, 9.87%); yield, 22% from the aldehyde; nmax/cm-1 (KBr): 1604s, 1456m, 1364m, 1106m, 858w; UV–VIS (CH2Cl2) lmax/nm (e): 269 (13 000), 328 (12 000), 561 (1040); Results and Discussion MS (EI) m/z: 283 (M+), 195, 153, 114, 84, 69.Spin density distribution of the radicals 2-(3-Chloro-4-hydroxyphenyl)-4,5-dihydro-4,4,5,5- The most widely accepted mechanism for rationalizing the tetramethyl-3-oxido-1H-imidazol-3-ium-1-oxyl 4. Radical 4 was intermolecular magnetic interactions in organic molecular synthesized by the same procedure as for 3. Crystals were solids is the so-called McConnell I mechanism based on the grown by a slow diusion of pentane into a concentrated overlap of the orbitals on atoms with large spin densities of toluene solution at room temp.Mp 158.4 °C (decomp.) (Found: neighbouring molecules.12,13 According to this mechanism, C, 55.31; H, 5.75; N, 9.81. Calc. for C13H16N2O3Cl: C, 55.03; dominant contacts of atoms with spin densities having the H, 5.68; N 9.87%); yield, 93% from the aldehyde; nmax/cm-1 same sign produce an antiferromagnetic interaction between (KBr): 1605m, 1490m, 1387m, 1340s, 1300m, 1273m, 1214m, the two neighbouring molecular units.In contrast, ferromag- 1169m, 1133m, 831m, 702w, 541w; UV–VIS (CH2Cl2) lmax/nm netic interactions are favoured if opposite signs in these (e): 282 (16 000), 369 (12 000), 618 (720); MS (EI) m/z: 283 contacts are predominant.For this reason it is important in (M+), 195, 153, 114, 84, 69. magnetic molecular materials to know how the spin density of the unpaired electron is distributed within the building block 2-(5-Chloro-2-hydroxyphenyl)-4,5-dihydro-4,4,5,5- molecules. tetramethyl-3-oxido-1H-imidazol-3-ium-1-oxyl 5. Radical 5 was Free tumbling solution EPR spectra provide the necessary obtained similarly to 3.Crystals were grown by slow evapor- information about such spin density distributions in organic ation of a heptane–dichloromethane (1051) solution at room free radicals, through the determination of the coupling contemp. Mp 118.7 °C (decomp.) (Found: C, 55.78; H, 5.94; N, stants with the magnetically active nuclei of the molecules. 9.05. Calc. for C13H16N2O3Cl: C, 55.03; H, 5.68; N, 9.87%); The EPR spectra of radicals 1–5 show basically five main yield, 25% from the aldehyde; nmax/cm-1 (KBr): 1573w, 1471s, groups of lines with relative intensities of 152535251, resulting 1375m, 1341m, 1278m, 1135m, 825m, 646w; UV–VIS (CH2Cl2) from the coupling of the unpaired electron with two equivalent lmax/nm (e): 349 (4400), 581 (440); MS (EI) m/z: 283 (M+), nitrogen nuclei (I=1), as shown in Fig. 1(a) for radical 4. The 195, 153, 114, 84, 69. ‡ Atomic coordinates, thermal parameters, and bond lengths and angles have been deposited at the Cambridge Crystallographic Data Magnetic measurements Centre (CCDC). See Information for Authors, J. Mater. Chem., 1997, DC magnetic susceptibility data from 2 to 300 K, in a magnetic Issue 1.Any request to the CCDC for this material should quote the full literature citation and the reference number 1145/40. field of 1 T, were collected using a ‘Quantum Design’ MPMS 1724 J. Mater. Chem., 1997, 7(9), 1723–1730agreement with those previously reported by other authors,16 who have determined by NMR measurements that the N and O atoms carry a large and positive spin density while the acarbon atom has a significant negative spin density.The methyl groups and the aromatic ring also carry a spin density, as inferred from the observed EPR hyperfine couplings with all these hydrogen atoms. The coupling constants for the aromatic hydrogen atoms are smaller in radicals 1, 3 and 5 than in radicals 2 and 4. The loss of planarity between the five- and the six-membered rings, due to the presence of a substituent at the ortho position in compounds 1, 3 and 5, yields a satisfactory explanation for this fact.Assignments of the coupling constants of aromatic hydrogen atoms for radicals 1–5 have been performed by comparison with a whole series of radicals with non-magnetically active substituents located at dierent positions of the phenyl rings.17 Molecular and crystal structures of radicals General crystallographic information for radicals 1, 2, 4 and 5 is summarized in Table 2.Atomic numbering schemes used for these radicals are shown in Fig. 2 together with their molecular conformations. Structure of radical 1. Radical 1 crystallizes in the orthorhombic system with an asymmetric unit which contains one radical molecule.The most relevant feature of the solid state molecular conformation of radical 1 is the large angle (62°) formed by the phenyl ring and the mean plane of the OMNMCMNMO unit. This twist angle is larger than that reported for the unsubstituted phenyl a-nitronyl aminoxyl radical (29°)18 as well as for the p-chloro- (24°),19 the o-hydroxy- (40°)4 and the p-hydroxyphenyl- (30°)3a substituted examples. Therefore, this result suggests that the large angle in 1 is merely due to steric hindrance of the bulky Cl atom at the ortho position.The intramolecular Cl,O distance in radical 1 (3.24 A ° ) is quite similar to intermolecular distances found between halogen and nucleophile atoms by Murray-Rust and co-workers.20 Fig. 1 (a) Complete EPR spectrum of radical 4 in toluene at 293 K. This fact suggests that, in spite of the strong steric hindrance (b) Experimental (upper) and simulated (lower) central groups of EPR between the Cl and the O atom, a slightly attractive interaction lines. The computer simulation was carried out using a Lorentzian between both atoms cannot be excluded. The measured dis- line shape with DH1/2 of 0.13 G and the hfcc valves given in Table 1. tance would be the resulting equilibrium position between the steric repulsion and such halogen–nucleophile attracting forces.values of the isotropic hyperfine coupling constants in all In our case, the CMCl,O angle is obviously much smaller cases are between 7.2 and 7.8 G; i.e. aN=7.5(3) G; which is typical for freely tumbling substituted a-nitronyl aminoxyl radicals.1,14,15 A more detailed analysis of these five main Table 2 Crystallographic data for radicals 1, 2, 4 and 5 groups of signals reveals a complex pattern of lines arising compound 1 2 4 5 from supplementary couplings with the twelve equivalent hydrogen atoms (I=1/2) of the four methyl groups and all the a/A° 10.483(3) 9.946(2) 13.735(1) 9.741(6) hydrogen atoms of the phenyl ring.b/A° 10.921(3) 11.522(1) 11.819(2) 11.663(5) Computer simulations of the experimental EPR spectra of c/A° 11.671(3) 12.700(2) 17.153(3) 13.324(14) 1–5 yield the hyperfine coupling constants summarized in b/° — 109.91(1) — 111.23(7) Table 1. All values are in agreement with those previously data/parameters 3887/165 4130/213 2441/176 2037/223 V /A° 3 1336.2(6) 1368.4(4) 2785.5(7) 1411(2) reported for other a-phenyl nitronyl aminoxyl radicals,15 indi- Dc/g cm-3 1.331 1.300 1.354 1.336 cating, therefore, that the chloro- and hydroxy-substituents do space group P212121 P21/c Pbca P21/n not alter the electron distribution in the radicals significantly. Z 4 4 8 4 Consequently, the unpaired electron is mainly distributed on R 0.0394 0.0478 0.0514 0.0488 both NO groups and the a-carbon atom.This result is in Table 1 Summary of hyperfine coupling constants (a/G; 1 G=10-4 T) obtained for radicals 1–5 by computer simulation of the EPR spectra of toluene solutions compound aN aH(methyl) aH(ortho) aH(meta) aH(para) 1 7.26 (2N) 0.20 (12H) 0.24 0.15, 0.12 a 2 7.40 (2N) 0.19 (12H) 0.52 (2H) 0.21 0.43 3 7.65 (2N) 0.18 (12H) 0.23 0.16 (2H) — 4 7.50 (2N) 0.21 (12H) 0.54, 0.50 0.17 — 5 7.81, 7.40 0.20 (12H) 0.33 0.25 0.30 aNot observed.J. Mater. Chem., 1997, 7(9), 1723–1730 1725Fig. 2 Molecular conformations of radicals 1, 2, 4 and 5 with the atomic numbering schemes used in the text and tables (75°) than in the examples described by Murray-Rust and co- perpendicular to the crystallographic [1 0 -1] direction.Between adjacent planes no H,O contacts with workers20 because of the intramolecular nature of that contact. As shown in Fig. 3, the molecules of radical 1 are packed in d(H,O)<3 A° are observed. Molecules of adjacent planes are just connected by weak CmethylMH11B,C11iii interactions such a way that one of the two NO groups of each molecule forms a hydrogen bond with a methyl group of a neighbouring [iii=-x, 1/2+y, 3/2-z; d(H11B,C11iii)=3.01 A ° ; h(C111MH11B,C11iii)=172°], and these interactions are molecule [d(H11D,O14i, i=-x, 1/2+y, 5/2-z)=2.64 A ° ; h(C112MH11D,O14i)=154°] giving rise to twisted chains propagated along the b axis.As a result of this complex crystal packing, the shortest along the b axis.Each of these chains is connected to two neighbouring chains by means of CaromMH7,O14iiMN13ii distance between the NO groups of dierent molecules [d(O14,N9iv)=4.94 A ° ; iv=-x, -1/2+y, -5/2-z; hydrogen bonds [ii=1/2+x, 1/2-y, 2-z; d(H7,O14ii)= 2.66 A ° ; h(C7MH7,O14ii)=130°] forming molecular sheets d(O14,O10iv)=5.39 A ° ] occurs between neighbouring molecules in the chains along the b axis.These kinds of contacts are responsible for the magnetic behaviour as will be discussed below. Structure of radical 2. This compound crystallizes in the monoclinic P21/c space group with an asymmetric unit containing one radical molecule. Fig. 2 shows its solid state molecular conformation. The torsion angle between the two rings of radical 2 is 26°, being very similar to those of other radicals with similar steric requirements at the ortho position. 18,19,3a The crystal structure of 2 clearly shows alternating planes perpendicular to the crystallographic [1 0 -1] direction. Within these planes, the molecules are connected to each other forming chains along the b axis [see Fig. 4(a)] by means of CaromMH14,O2iMN2i hydrogen bonds [i=1-x, -1/2+y, 1/2-z; d(H14,O2i)=2.47 A ° ; h(C14MH14,O2i)= 129°].These chains are linked to each other by weak CmethylMH44,O5iiMN5ii hydrogen bonds [ii=-x, 1/2+y, -1/2-z; d(H44,O5ii)=2.55 A ° ; h(C42MH44,O5ii)=148°]. As happens with radical 1, between adjacent planes containing the stronger CMH,O hydrogen bonds, no other H,O contacts with d(H,O)<3 A° are observed. In a similar way as Fig. 3 Crystal packing of radical 1. Crystallographic (1 0-1) plane occurs with radical 1 there are CaromMH15,C113iii interformed by chains of molecules connected through actions [iii=x, 1/2-y, 1/2-z; d(H15,C113iii)=3.12 A ° ; Cmethyl–H11D,O14i–N13i (i=-x, 1/2+y, 5/2-z) hydrogen bonds h(C15MH15,C113iii)=154°] between the planes. Also, (dashed lines) along the b axis. The dotted line connects the nearest between chains of neighbouring (1 0 -1) planes, we have NO groups of dierent molecules, N13–O14,N9iv (iv=-x, -1/2+y, -5/2-z).found short C113,C113iv contacts [iv=1-x, -y, -z; 1726 J. Mater. Chem., 1997, 7(9), 1723–1730strong intermolecular O8MH8,O15iMN14i bond [i=1-x, -1/2+y, 1/2-z; d(H8,O15i)=1.72 A ° ; h(O8MH8,O15i)= 168°] which connects each molecule to two neighbours giving rise to zigzag chains along the b axis.This arrangement is reinforced by a second hydrogen bond [d(H6,O15i)=2.65 A ° ; h(C6MH6,O15i)=128°] between the same NO group and an aromatic H atom in the meta position which is adjacent to the OH group which forms the strong hydrogen bond. Further, these chains are arranged into planes perpendicular to the crystallographic [1 0 0] direction by means of weak CmethylMH134,O11iiMN10ii hydrogen bonds [ii=1-x, -y, -z; d(H134,O11ii)=2.71 A ° ; h(C132MH134,O11ii)=160°], as shown in Fig. 5(a). This pattern resembles very much the crystal packing reported for the 4-hydroxyphenyl a-nitronyl aminoxyl radical3 but there are two basic dierences caused by the presence of the Cl atom at the meta position. The first is that the rings of radical 4 are not coplanar to the (1 0 0) plane; i.e.the molecules are alternately canted within the chains. The second dierence refers to the stacking of the planes along the crystallographic a direction, which in radical 4 takes place through CmethylMH125,Cl1iii interactions [iii=-1/2+x, 1/2-y, -z; d(H125,Cl1iii)=3.07 A ° ; h(C122MH125,Cl1iii)=167°], CmethylMH121,Cl1iv interactions [iv=-1/2+x, y, 1/2-z; d(H121,Cl1iv)=3.08 A ° ; h(C121MH121,Cl1iv)=134°] [shown in Fig. 5(b)] and CmethylMH131,O11vMN10v hydrogen bonds [v=1/2-x, 1/2+y, z; d(H131,O11v)=2.75 A ° ; h(C131MH131,O11v)=151°], while in the para-hydroxy substituted radical the stacking of planes occurs through van der Waals interactions. The lack of planarity of the molecules of radical 4, within the twisted chains in the (1 0 0) plane, means that the shortest intermolecular distance between NO groups occurs pairwise Fig. 4 Crystal packing of radical 2. (a) Crystallographic (1 0-1) plane formed by chains of molecules connected through Carom–H14,O2i–N2i (i=1-x, -1/2+y, 1/2-z) hydrogen bonds (dashed lines) along b axis. (b) Molecules of two neighbouring (1 0-1) planes showing as dashed lines C113,C113iv close contacts (iv= 1-x, -y, -z).The dotted lines connect the nearest NO groups of dierent molecules, N2,O5v–N5v (v=-x, 1-y,-z). d(C113,C113iv)=3.72 A ° ], which are depicted in Fig. 4(b). The distances of such contacts are in accordance with those described by Desiraju7 for several other compounds which clearly show non-covalent attractive Cl,Cl interactions.Therefore this result suggests that the Cl,Cl interactions are driving forces for the packing of molecules in (1 0 -1) planes. As will be discussed below, from the magnetic point of view, the most important aspect of the crystal packing of radical 2 is the presence of short intermolecular distances between the NO groups of two molecules in neighbouring (1 0 -1) planes.This kind of contact permits us to consider the molecular system as being composed of dimeric magnetic entities which are clearly shown in Fig. 4(b). Structure of radical 4. Radical 4 crystallizes in the orthorhom- Fig. 5 Crystal packing of 4. (a) Crystallographic (100) plane formed by zigzag chains of molecules connected through strong bic Pbca group and also contains one molecule in the asymmet- O8–H8,O15i–N14i (i=1-x, -1/2+y, 1/2-z) hydrogen bonds ric unit.In this case the phenyl ring is twisted with respect to along b axis depicted as dashed lines. The dotted lines connect the the mean OMNMCMNMO plane by 28.5°. This result is nearest NO groups of dierent molecules, N10,O11vi–N10vi (vi= again in accordance with the twist angles observed for other 1-x, -y, -z).(b) Molecules of neighbouring (100) planes showing radicals with similar steric requirements, that is, those without Cmethyl–H125,C11iii (iii=-1/2+x, 1/2-y, -z) and substituents at the ortho positions. Cmethyl–H121,C11iv (iv=-1/2+x, y, 1/2-z) interactions which are depicted as dashed lines. The main feature of the crystal packing of radical 4 is the J. Mater. Chem., 1997, 7(9), 1723–1730 1727between molecules of neighbouring chains as shown in Fig. 5(a) leading to a dimeric magnetic pattern, as will be discussed below. Structure of radical 5. This compound does not crystallize in the orthorhombic system, as occurs for radicals 1 and 4, but in the monoclinic one, similarly to radical 2 which also has one chlorine atom in the meta position. The presence of the OH group at the ortho position results in the formation of a strong intramolecular O51MH51,O15MN14 hydrogen bond [d(H51,O15)=1.66 A ° ; h(O51MH51,O15)=170°].Consequently, there is a twist angle of 35° between the two rings of the molecule, as was also observed for the radical with one OH group in the ortho position.4 Fig. 2 shows the molecular structure of 5, where the lack of planarity is clear.Due to the establishment of these strong intramolecular OMH,OMN hydrogen bonds, the molecules are packed in the crystal through other types of weak hydrogen bonds. Thus, the molecules are arranged into zigzag chains parallel to the [1 0 1] direction by means of rather weak CaromMH7,O11iMN10i hydrogen bonds [i=-1/2+x, 1/2-y, -1/2+z; d(H7,O11i)=2.22 A ° ; h(C7MH7,O11i)=157°].The chains are layered into planes perpendicular to the [1 0 -1] direction by CmethylMH13B,O15iiMN14ii hydrogen bonds [ii=1/2+x, 3/2-y, 1/2+z; d(H13B,O15ii)=2.82 A ° ; h(C131M H13B,O15ii)=155°] as shown in Fig. 6(a). Furthermore, the planes are connected pairwise through other CmethylMH12A,O11iiiMN10iii hydrogen bonds [iii= 1-x, 1-y, 1-z; d(H12A,O11iii)=2.68 A ° ; h(C121M H12A,O11iii)=174°] giving rise to an alternating pattern in which the radical molecules of two neighbouring planes form dimers, related by an inversion centre shown in Fig. 6(b). As will be seen below, these dimers are responsible for the magnetic behaviour of the compound. As can be inferred from geometric considerations, all CMH,Cl distances present in this radical are longer than 3.3 A ° .{The shortest ones are [d(H6,Cl1i)=3.33 A ° ; h(C6MH6,Cl1i)=156°] and [d(H123,Cl1iv)=3.36 A ° ; h(C122MH123,Cl1iv)=110.6°; iv=x, 1+y, z]}. The crystal packing of radical 5 is therefore very similar to the molecular arrangement of radical 2. This result can be Fig. 6 Crystal packing of 5. (a) Crystallographic (1 0-1) plane formed rationalized by noting that both compounds have similar by chains of molecules connected through Carom–H7,O11i–N10i (i= crystalline design elements: in fact both radicals have one Cl -1/2+x, 1/2-y, -1/2+z) hydrogen bonds parallel to the a+c atom at the meta position and the additional OH substituent direction.(b) Molecules of two neighbouring (1 0-1) planes forming at the ortho position of radical 5 seems to aect only the dimers by intermolecular Cmethyl–H12A,O11iii–N10iii (iii=1-x, intramolecular conformation and not the intermolecular inter- 1-y, 1-z) hydrogen bonds.The strong intramolecular O51–H51,O15–N14 hydrogen bonds are also depicted. actions. The fact that no Cl,Cl contacts are observed in radical 5 may be explained as a consequence of the lower planarity7 of this molecule due to the OH substituent at the ortho position.Magnetic properties of radicals Summarizing the above structural analysis, the arrangement of molecules of the radicals 1, 2, 4 and 5 in the solid state is Static magnetic susceptibility measurements of radicals 1–5 mainly governed by strong OMH,OMN (only possible in are shown in Fig. 7. Such measurements indicate that the radicals 4 and 5) and weak CMH,OMN hydrogen bonds molecular solids studied present in all cases intermolecular giving rise to molecular solids with one- or two-dimensional antiferromagnetic (AFM) interactions with dierent strengths.structural character.6,21 The CMH,Cl hydrogen bonds only As we will discuss later, the magnetic dimensionalities vary seem to play an important role in the crystal packing of the from one compound to another in close relationship with their radicals without OH substituents on the phenyl ring, as shown crystal packings.by the shorter H,Cl distances in radicals 1 and 2 compared The magnetic susceptibility data of 1 were nicely fitted to a with radicals 4 and 5. Cl,Cl interactions have only been 1D Heisenberg model of S=1/2 molecular units having weak found for radical 2.According to Desiraju et al.,7 the lack of AFM interactions with a magnetic exchange interaction of planarity, especially in radicals 1 and 5, and the presence of J/kB=-0.95 K. In the crystal structure of radical 1, all the other stronger intermolecular interactions, as occurs for 4, may intermolecular distances between atoms carrying the larger be responsible for the absence of Cl,Cl interactions in radicals spin densities are quite large.As described previously, the 1, 4 and 5. As a consequence of such considerations, it shortest one [d(O14,N9iv)=4.94 A ° ; iv=-x, -1/2+y, has been observed that the placement of an additional Cl -5/2-z] occurs among atoms having a positive sign of spin atom modulates considerably the crystal packing of phenyl density and that belong to molecules forming the previously described chains along the crystallographic b direction.In a-nitronyl aminoxyl radicals. 1728 J. Mater. Chem., 1997, 7(9), 1723–1730maximum at 9 K in the x vs. T plot strongly suggests a low magnetic dimensionality for this molecular solid. Actually, the data can be fitted to a dimer chain model23 of S=1/2 molecular units with an intradimer exchange interaction of J1/kB=-14.9 K and an interdimer exchange interaction of J2/kB=-10.5 K.For radical 4, the shortest distance between NO groups of dierent molecules [d(N10,O11ii)=3.69 A ° , d(O11,O11ii)=3.81 A ° , ii=1-x, -y, -z] occurs within the (1 0 0) layers among molecules belonging to neighbouring chains. The shortest distance between the spin-carrying units of dierent dimers [d(O15,O15vi)=4.61 A ° ; vi=1-x, 1-y, -z] takes place along the b axis within the (1 0 0) plane.Moreover, the remaining distances between dimers are much longer, the lowest one being 5.65 A ° . Therefore, all these facts explain the goodness of the experimental data to a dimer chain model for radical 4. This magnetostructural correlation has short intra- and inter-dimer distances between NO units that Fig. 7 Temperature dependence of the paramagnetic susceptibility x are responsible for the strong antiferromagnetic couplings of radicals (%) 1, (+) 2, (&) 3, (') 4 and ($) 5. The inset shows an enlargement of the temperature dependence for radicals 4 and 5. The within and between the magnetic dimers.Both distances are solid lines represent the fits of experimental data to the magnetic clearly shorter than those observed for radicals 1 and 2 in models explained in the text. agreement with the stronger antiferromagnetic couplings observed for 4. The x vs. T plot of radical 5 shows a broader maximum at consequence, this molecular arrangement is in agreement with ca. 50 K indicating again a low dimensionality with a very the observed weak 1D antiferromagnetic behaviour.strong antiferromagnetic interaction. The experimental data The magnetic behaviour of 2 is properly explained by a can be fitted to a simple dimer model with strong AFM (J/kB= dimer model with antiferromagnetic interactions. However, the -42.3 K) interactions.24 An additional Curie tail (C=0.0043 fitting of experimental data is considerably improved if an emu K mol-1) is necessary to fit the low temperature values, additional interdimer antiferromagnetic term is included in the which takes into account possible crystal surface eects or the Bleaney–Bowers equation for S=1/2 molecular units with an presence of dislocations in the crystals.The strength of the antiferromagnetic interaction of J/kB=-1.82 K, by means of intradimer interaction is remarkable, since it is one of the a temperature correction with a Weiss constant of h= largest reported to date for a purely organic compound.-1.39 K.22 This result means that the dominant magnetic As mentioned previously in the description of the molecular interactions take place within the dimers but such dimers also packing of radical 5, there are molecules linked by interact weakly with each other in a quite isotropic antiferro- CMH,OMN hydrogen bonds that clearly form dimeric enti- magnetic fashion.Analysing the crystal structure of 2, we have ties. The shortest distance between the spin-carrying units for found that the shortest distance between NO groups of dierent radical 5 occurs inside these dimers [d(O11,O11iii)=3.37 A ° ; molecules is slightly shorter [d(N2,O5v)=4.89 A ° ; iii=1-x, 1-y, 1-z].This short intermolecular distance is d(O2,O5v)=5.03 A ° ; v=-x, 1-y, -z] than in radical 1, the shortest one observed in this family of radicals and explains occurring in a dimer geometry instead of in chains. This the strong intermolecular magnetic interaction observed for structural pattern therefore explains the observed magnetic radical 5.characteristics of this molecular solid. The distances between From the magnetostructural study it has been possible to NO groups of dierent dimers are larger than 5.24 A ° and can establish a complete correlation between the solid state mag- be associated with the weaker interdimer antiferromagnetic netic properties and the molecular arrangement in the crystals interactions.for all the radicals studied. Table 3 summarizes the main The experimental data for radical 3 can be fitted either by features of the structure and magnetic behaviour for radicals a 1D AFM model with J/kB=-0.62 K or by the Curie–Weiss 1, 2, 4 and 5 showing in addition that the structural dimension- law with h=-0.89 K.Thus, such data clearly show the alities, based on packing motifs linked by hydrogen bonds, are antiferromagnetic nature of the intermolecular interactions dierent from the magnetic dimensionalities. even though they do not have any singularity that permits us It also seems clear that the magnitude of exchange inter- to distinguish the magnetic dimensionality of this molecular actions between radical molecules in these molecular solids system.In contrast, in the case of radical 4, the presence of a broad strongly decreases as the mean intermolecular distances Table 3 Summary of structural and magnetic dimensionalities of radicals 1, 2, 3 and 5 together with the most relevant intermolecular contacts from the structural and magnetic points of view.See text for symmetry operations compound structural dimensionalitya structurally relevant contacts magnetic dimensionalityb magnetically relevant contacts 1 2D d(H11D,O14i)=2.64 A ° 1D, regular chains d(H7,O14ii)=2.66 A ° (J/kB=-0.95 K) d(O14,O10iv)=5.39 A ° c 2 2D d(H14,O2i)=2.47 A ° 0D AFM, dimers d(H44,O5ii)=2.55 A ° (J/kB=-1.82 K) d(O2,O5v)=5.03 A ° d 4 2D d(H8,O15i)=1.72 A ° 1D, dimer chains d(H134,)15ii)=2.71 A ° (J/kB=-14.9 K) d(O15,O15vi)=4.61 A ° (J2/kB=-10.5 K) d(O11,O11ii)=3.81 A ° e 5 1D d(H7,O11i)=2.22 A ° 0D, dimers (J/kB=-42.3 K) d(O11,O11iii)=3.37 A ° aBased on the observed packing motifs linked by H,O hydrogen bonds with distances shorter than 2.8 A ° .bBased on the fits of experimental magnetic data to magnetic models.Figures in parentheses are the strengths of dominant magnetic exchange interactions which are in all cases antiferromagnetic. cThe corresponding O,N contact is shorter at d(O14,N9iv)=4.94 A ° . dThe corresponding O,N contact is shorter at d(N2,O5v)=4.89 A ° . eThe corresponding O,N contact is shorter at d(N10,O119ii)=3.69 A ° . J. Mater. Chem., 1997, 7(9), 1723–1730 172910 J.H. Osiecki and E. F. Ullman, J. Am. Chem. Soc., 1968, 90, 1078; between ONCNO units increases, being insignificant for O,O E. F. Ullman, J. H. Osiecki, D. G. B. Boocock and R. J. Darcy, distances larger than 6.0 A ° . Moreover, the AFM nature of J. Am. Chem. Soc., 1972, 94, 7049. such interactions can be rationalized by the McConnell I 11 M. Lamchen and T. W. Mittag, J.Chem. Soc. C, 1966, 2300. mechanism, through the contacts of atoms with the same sign 12 (a) H. M. McConnell, J. Phys. Chem., 1963, 39, 1910; of spin density, that belong to the NO units carrying the large (b) J. B. Goodenough, Magnetism and the Chemical Bond, Interscience, New York, 1963, p. 163; (c) H. M. McConnell, Proc. spin density. An alternative but coincident rationalization of R.A. Welch Found, Conf. Chem. Res., 1967, 11, 144. such magnetic interactions can be provided by the overlap of 13 In some aspects the McConnell I mechanism is a restatement of the SOMOs of neighbouring molecules.13 concepts introduced many years before by Heitler and London. Thus, dominant contact of atoms of two neighbouring molecules This research was financially supported by the C.I.C.y T. with spin densities having the same sign is equivalent to a non- (Grant,MAT 94-0797), Spain and the Generalitat de Catalunya zero overlap integral between their singly occupied molecular orbitals (SOMO) because such SOMOs are located on those (Grant, SGR 95/00507). O. J. and J. C. thank also the atoms. Generalitat de Catalunya for the award of doctoral fellowships. 14 M. S. Davis, K. Morokuma and R. W. Kreilick, J. Am. Chem. Soc., 1974, 96, 652. 15 J. A. D’Anna and J. H. Wharton, J. Chem. Phys., 1970, 53, 4047. References 16 (a) M. S. Davis, K. Morokuma and R. W. Kreilick, J. Am. Chem. Soc., 1972, 94, 5588; (b) J. W. Neely, G. F. Hatch and 1 M. Tamaura, Y. Nakazawa, D. Shiomi, K. Nozawa, Y. Hosokoshi, M. Ishikawa, M. Takahashi and M. Kinoshita, Chem.Phys. L ett., R. W. Kreilick, J. Am. Chem. Soc., 1974, 96, 652. 17 (a) J. Veciana, J. Cirujeda and O. Ju� rgens, Manuscript in prep- 1991, 186, 401. 2 (a) J. Veciana, J. Cirujeda, C. Rovira and J. Vidal-Gancedo, Adv. aration; (b) J. Cirujeda, Ph D Thesis, 1997, Universitat Ramon Llull. Mater., 1995, 7, 221; (b) J. Cirujeda, C. Rovira and J. Veciana, Synth. Met., 1995, 71, 1799; (c) J.Cirujeda, E. Herna�ndez-Gasio� , 18 C. W. Wang and S. F. Watkins, J. Chem. Soc., Chem. Commun., 1973, 888. F. Lanfranc de Panthou, J. Laugier, M. Mas, E. Molins, C. Rovira, J. J. Novoa, P. Rey and J. Veciana, Mol. Cryst. L iq. Cryst., 1995, 19 (a) J.-L. Stagner, Doctoral Thesis, 1995, Universite� Louis Pasteur de Strasbourg; (b) Y. Hosokoshi, Doctoral Thesis, 1995, University 271, 1. 3 (a) H. Herna�ndez-Gasio� , M. Mas, E. Molins, C. Rovira and of Tokyo. 20 N. Ramasubbu, R. Parthasarathy and P. Murray-Rust, J. Am. J. Veciana, Angew. Chem., Int. Ed. Engl., 1993, 32, 882; (b) J. Cirujeda, E. Herna�ndez-Gasio� , C. Rovira, J. L. Stanger, Chem. Soc., 1986, 108, 4308. 21 G. R. Desiraju, Acc. Chem. Res., 1991, 24, 290. P. Turek and J. Veciana, J.Mater. Chem., 1995, 5, 243. 22 The equation used to fit the data corresponds to a modified 4 J. Cirujeda, M. Mas, E. Molins, F. Lanfranc de Panthou, Bleaney–Bowers equation which includes a multiplying term to J. Laugier, J. G. Park, C. Paulsen, P. Rey, C. Rovira and J. Veciana, take account of intradimer interactions: x= J. Chem. Soc., Chem. Commun., 1995, 709. [Ng2mB2/3k(T-h)][1+(1/3) exp(-2J/kT )]-1. See R. L. Carlin in 5 T. Sugawara, M. M. Matsuskita, A. Izuoka, N. Wada, N. Takeda Magnetochemistry, Springer Verlag, Berlin, 1986, p. 88. and M. Ishikawa, J. Chem. Soc., Chem. Commun., 1994, 1723. 23 T. Barnes and J. Riera, Phys. Rev. B, 1994, 50, 6817. 6 R. Taylor and O. Kennard, J. Am. Chem. Soc., 1982, 104, 5063. 24 However, the fitting of the experimental data for 5 is slightly 7 (a) J. A. R. P. Sarma and G. R. Desiraju, Acc. Chem. Res., 1986, 19, improved if the same model as for radical 2 is employed, with an 222; (b) G. R. Desiraju, in Organic Solid State Chemistry, ed., intradimer interaction of J/kB=-50.1 K and an additional inter- G. R. Desiraju, Elsevier, 1987, p. 519; (c) J. A. R. P. Sarma and dimer term of h=+10.4 K, corresponding to the magnetic inter- G. R. Desiraju, Chem. Phys. L ett., 1985, 117, 160. actions among the dimers. 8 G. R. Desiraju, in Organic Solid State Chemistry, ed., G. R. Desiraju, Elsevier, 1987, p. 539. 9 N.W. Thomas and G. R. Desiraju, Chem. Phys. L ett., 1984, 110, 99. Paper 7/00589J; Received 27th January, 1997 1730 J. Mater. Chem., 1997, 7(9),