Dc and ac magnetic properties of the two-dimensional molecular-based ferrimagnetic materials A2M2[Cu(opba)]3 nsolv [A+=cation, MII=MnII or CoII, opba=orthophenylenebis( oxamato) and solv=solvent molecule] Olivier Cador,a Daniel Price,a Joulia Larionova,a Corine Mathonie`re,a Olivier Kahn*a and J. V. Yakhmib aL aboratoire des SciencesMole�culaires, Institut de Chimie de la Matie`re Condense�e de Bordeaux, UPR CNRS no 9048, 33608 Pessac, France bChemistry Division, Bhaba Atomic Research Center, 400 085 Bombay, India This paper is devoted to a thorough study of the magnetic properties, both in the dc and ac modes, of two-dimensional ferrimagnetic materials.The general formula of the compounds, abbreviated as A2M2Cu3, is A2M2[Cu(opba)]3 nsolv where A+ stands for a counter cation (radical cation, alkali-metal or tetraalkylammonium), opba is ortho-phenylenebis(oxamato), MII is MnII or CoII , and solv is a solvent molecule (Me2SO or H2O).The dc magnetic susceptibility data for all compounds down to ca. 50 K are characteristic of two-dimensional ferrimagnets. In the A2Mn2Cu3 series, three classes of compounds have been distinguished. Class I has a unique representative, AV2Mn2Cu3, where AV+ is the radical cation 2-(1-methylpyridinium-4-yl)-4,4,5,5- tetramethylimadozolin-1-oxyl-3-oxide, the structure of which has already been solved.This compound is a magnet with Tc= 22.5 K. Class II corresponds to compounds with large cations (tetraethyl- and n-tetrabutylammonium). They also behave as magnets with Tc around 15 K. Class III corresponds to compounds with small cations (alkali-metal ions and tetramethylammonium).They behave as metamagnets with a long-range antiferromagnetic ordering in zero field around 15 K, and a field-induced ferromagnetic state. The critical fields are of the order of 0.15 kOe. All the A2Co2Cu3 compounds are magnets with Tc around 30 K. Furthermore, the cobalt derivatives show a very strong coercivity, with coercive fields of several kOe at 5 K.They also display a pronounced magnetic after-eect in the ordered phase. In the course of this work several peculiar features have been observed. In particular, the A2Mn2Cu3 compounds have been found to present weak but significant negative out-of-phase ac magnetic signals at temperatures just above Tc. All the observed phenomena are discussed and, in particular, a mechanism for the long-range ordering in these two-dimensional compounds has been proposed.For more than a decade, one of our main fields of research where opba is ortho-phenylenebis(oxamato), with a divalent metal ion, MII, in 151 stoichiometry. Two-dimensional com- has dealt with the design of molecular-based magnets and the pounds may be also obtainedby reactingthe dianionic precursor study of their physical properties.The very first compounds of with a divalent metal ion, but in 253 stoichiometry and in the this kind were described in 1986 by Miller et al.,1,2 and by our presence of two equivalents of a non-coordinating cation A+. group.3,4 Since these pioneering results, quite a few research This aords compounds of formula A2M2[Cu(opba)]3 nsolv groups have initiated some activity along this line, and a large where solv is a solvent molecule.Many types of cations A+ number of molecular-based materials exhibiting a spontaneous may be utilized, including alkali-metal, simple organic cations magnetization below a certain critical temperature have been such as tetraalkylammonium, as well as radical cations.In one reported.3–10 In many cases these compounds contain two case, involving the radical cation 2-(1-methylpyridinium-4-yl)- kinds of spin carriers, either two dierent metal ions,11–20 or a 4,4,5,5-tetramethylimidazolin-1-oxyl-3-oxide (AV+), the crystal metal ion and an organic radical.21–28 In at least two cases structure of the compound was solved.29,30 This structure, shown three spin carriers were involved.29–31 Other very interesting in Fig. 1, consists of two perpendicular and fully interlocked magnetic compounds, at the frontier between molecular and networks of honeycomb-like layers. In this compound the solid-state chemistry, have also been reported.32–34 interlocking arises from the bridging ability of the radical cation Furthermore, several conferences have been devoted to this which bridges two CuII ions belonging to perpendicular layers.subject in recent years.35–37 Since all the other A+ cations used in this work lack any The syntheticapproach we have advocated to design molecu- coordination ability, their structure is probably that of a single lar-based magnets is via heterobimetallic compounds.This network of parallel honeycomb layers, with the A+ cations approach consists of designing low-dimensional heterobimetal- situated between the layers. lic species showing ferrimagnetic behaviour, and of assembling In preceding papers we reported on the magnetic properties them within the crystal lattice in a ferromagnetic fashion. The of some of these two-dimensional compounds, exclusively in ferrimagnetic behaviour arises from antiferromagnetic inter- the dc mode.12,29–31 Here, we present a detailed investigation actions between dierent spin carriers.We first focused on of the magnetic properties in both dc and ac modes. We will one-dimensional,19,38,39 then on two-dimensional ferrimag- focus on two series of compounds, with MII=MnII and CoII, nets.12,20 The one-dimensional compounds are synthesized by respectively.For convenience, all the compounds of formula reacting a dianionic copper(II) precursor such as [Cu(opba)]2-: A2M2[Cu(opba)]3 nsolv are abbreviated to A2M2Cu3 . The nature and the number of solvent molecules present in the structure, as deduced from chemical analyses, are specified in Tables 1 and 2. Experimental Syntheses Solvents and reagents were obtained commercially, and were used without further purification.The ligand ortho-phenylene- J. Mater. Chem., 1997, 7(7), 1263–1270 1263Fig. 1 Structure of AV2Mn2Cu3 . (Top) view of a honeycomb-like layer. (Bottom) view showing the interpenetration of the two perpendicular networks (reproduced with permission from ref. 29 and 30). 1264 J. Mater. Chem., 1997, 7(7), 1263–1270Table 1 Some characteristic temperatures (K) for the magnetic proper- tives the eect of concentration upon the reaction is crucial. If ties of the A2Mn2Cu3 materials the solution is too concentrated, Me2SO solvated manganese( II) acetate appears in addition to the expected compound ac mode whereas if the solution is too dilute, no compound is obtained, dc mode MnII being slightly base sensitive, with decompositions before xMT a x¾Mb xMb or during the crystallization, producing an insoluble brown class I impurity.For cobalt derivatives, the eect of concentration is A 2Mn2[Cu(opba)]3 2Me2SO 2H2O 115 23 22.5 not so crucial,although if the solutionis too dilute precipitation class II will take longer. (NEt4)Mn2[Cu(opba)]3 4Me2SO 2H2O 125 15.5 15.5 Chemical analyses of all elements, except oxygen, were (NBu4)Mn2[Cu(opba)]3 4Me2SO 2H2O 125 15 15 carried out and satisfactory results were obtained.In particular, class III the ratio Cu/M (M=Mn or Co) was found to be 1.5±0.04 Na2Mn2[Cu(opba)]3 11Me2SO 10H2O 125 15 for all the compounds. On heating the solvent molecules are 5.5 5.5 removed, which usually leads to an increase of the critical K2Mn2[Cu(opba)]3 11Me2SO 10H2O 120 14 temperatures.12 However, the resulting compounds seem to be 7.5 6 poorly crystallized, so we discuss them no further.(NMe4)2Mn2[Cu(opba)]3 4Me2SO 2H2O 125 14 8 Magnetic measurements aTemperature for minimum of xMT . bTemperature for maximum of x¾M These were carried out with a Quantum Design MPMS-5S or xM .SQUID magnetometer, working in both dc and ac modes between 2 and 300 K, and from 0 to 50 kOe. Table 2 Some characteristic temperatures (K) for the magnetic properties of the A2Co2Cu3 materials A2Mn2 [Cu(opba)]3 nsolv compounds ac mode We can distinguish three classes of A2Mn2Cu3 compounds, dc mode namely: class I containing a unique representative with AV+, xMT a x¾Mb xMb class II where A+ is a large cation (NEt4+ or NBu4+) and Na2Co2[Cba)]3 6Me2SO 4H2O —c 31 31 class III where A+ is a small cation (Na+, K+ or NMe4+).K2Co2[Cu(opba)]3 12Me2SO 3H2O —c 32.5 32.5 Independently of their magnetic properties, it became evident Cs2Co2[Cu(opba)]3 5Me2SO 4H2O —c 32.5 32.5 over the course of this work that not all these materials had (NMe4)2Co2[Cu(opba)]3 7Me2SO H2O —c 33 32.5 the same stability. While compounds of classes I and II were (NBu4 )2Co2[Cu(opba)]3 6Me2SO H2O 95 29 28.5 stable for over a year at room temperature compounds of aTemperature for minimum of xMT .bTemperature for maximum of x¾M class III were found to gradually decompose. At room tempera- or xM. cThe magnetic susceptibility data were not measured ture this decomposition is noticeable within a few hours of above 70 K.their removal from the mother-liquor, and was apparently complete within a matter of a day and was detected as a bis(ethyloxamate), H2Et2opba, was synthesized as previously colour change from blue to purple. Storing the samples in the described.12,39 The synthesis of AV2Cu(opba) H2O and cold, however, extends their lifetime suciently for physical AV2Mn2[Cu(opba)]3 2Me2SO 2H2O were also performed as studies to be performed.The study of the magnetic properties previously described.29,30 On the other hand, the syntheses of of the compounds after decomposition very strongly suggests the other precursors A2Cu(opba) nsolv and ferrimagnetic that a mixture of a chain compound MnCu(opba) nsolv12,39 materials A2M2[Cu(opba)]3 nsolv were slightly modified.A and the isolated precursor A2Cu(opba) was obtained. It is typical synthesis of a precursor, (NBu4)2Cu(opba), is as follows: sucient to point out that compounds of class III seem to be an aqueous solution of NBu4OH (8.77 g, 40% m/m, 13.5 mmol) thermodynamically unstable with respect to the breakdown of was added dropwise, with rapid stirring, to a solution of the planar structure into chains and residual non-coordinated H2Et2opba (2.92 g, 3.25 mmol) and CuCl2 2H2O (0.55 g, copper precursors.The Rb+ and Cs+ compounds were found 3.23 mmol) in methanol (100 ml). Initially, a green precipitate to be so unstable that they could not be studied. was formed, but upon continued addition of the base this redissolved to give a blue–violet solution.The reaction mixture Dc magnetic properties. Both the temperature and field dependences of the dc magnetic responses were investigated. was stirred for 5 min, filtered, and the solvent removed under reduced pressure to give the crude product as an oil, which The temperature dependences may be represented in the form of the xMT vs.T plots, where xM is the dc molar magnetic slowly crystallized. Recrystallization from acetonitrile gave violet needles of (NBu4)2Cu(opba), which were collected by susceptibility and T the temperature. The xMT vs. T plot for AV2Mn2Cu3 has already been investigated in detail.29,30 All filtration and dried under vacuum (1.95 g, 76%). The preparation of A2Cu(opba) precursors with other organic cations the other compounds behave in a similar way down to ca. 30 K. At room temperature, xMT is ca. 8.1 emu K mol-1, which is (NMe4+ and NEt4+) was the same but using the appropriate hydroxides. For the alkali-metal cations, Na+, K+ and Cs+, slightly lower than expected for isolated spin carriers. Then xMT decreases smoothly as the temperature is lowered, reaches water was added as a co-solvent to aid the solubility of the precursors (methanol–water, 80/20 v/v).a rounded minimum around 120 K, then increases very abruptly at lower T to a very large maximum value around A typical example of the synthesis of a ferrimagnetic material, that of (NBu4)2Mn2[Cu(opba)] 4Me2SO 2H2O, is as follows: 15 K. This maximum value of xMT is roughly twice as large for compounds of class II than for compounds of class III.A (NBu4)2Cu(opba) (0.2469 g, 0.449 mmol) was dissolved in Me2SO (2 ml), which gives a violet solution to which was typical example, that of (NBu4)2Mn2Cu3, is shown in Fig. 2. The minimum in the xMT vs. T plot reveals antiferromagnetic added Mn(MeCO2)2 4H2O (0.059 g, 0.241 mmol). The resulting blue solution was filtered, and allowed to stand.After interactions between nearest-neighbour metal ions without correlation length, and the huge increase of xMT at low one day a microcrystalline product had formed. This was collected by filtration, washed with Me2SO, dried under temperature is due to the increase of the correlation length with the SMn spins of MnII ions aligned along the field direction vacuum for 4 h, and stored at-18 °C.All the other derivatives were synthesized by a similar procedure. It is probably worth and the SCu spins of CuII ions aligned along the opposite direction. specifying here that for the A2Mn2[Cu(opba)]3 nsolv deriva- J. Mater. Chem., 1997, 7(7), 1263–1270 1265above ca. 0.15 kOe, which is characteristic of metamagnetic behaviour (Fig. 3). The field dependence of the magnetization, M=f (H), at 2 K for AV2Mn2Cu3 has already been reported.29,30 The magnetization first increases very rapidly, as expected for a magnet, then instead of reaching a saturation value at low field, increases smoothly as H increases, and does not reach saturation even at 200 kOe.This peculiar behaviour was attributed to the progressive decoupling of the radical spin with respect to the Mn2[Cu(opba)]3 skeleton spin.For all the compounds of classes II and III the magnetization at 2 K first increases very rapidly, then reaches a saturation value of ca. 7 mB. This value corresponds well to that expected if all the SMn spins are oriented along the field direction, and all the SCu spins oriented along the opposite direction. An interesting feature is that .M/.H is not at a maximum at exactly zero field, but at a field which depends on temperature.For instance, for (NBu4)2Mn2Cu3, the maximum of .M/.H is observed at 60 and 100 Oe at 5 and 10 K, respectively. This behaviour implies that a small coercive field is required to orient the domains, and therefore that some anisotropy is present in the system. For all the A2Mn2Cu3 materials, the coercive fields are Fig. 2 xMT vs. T plot for (NBu4 )2Mn2Cu3. Black dots represent the experimental data, and the full line represents the calculated curve in extremely weak being of the order of 10 Oe. These magnets the 50–300 Ktemperature range, using the high-temperature expansion do not exhibit significant hystereses, in striking contrast with approach (see text).The insert emphasizes the minimum of xMT the A2Co2Cu3 derivatives. characterizing the ferrimagnetic behaviour. Ac magnetic properties. The ac magnetic properties also reveal dierences between the three classes of the A2Mn2Cu3 Recently, we developed two theoretical approaches of the compounds. In the following we define x¾M and xM as the in- thermodynamical properties of A2Mn2Cu3 two-dimensional and out-of-phase molar magnetic susceptibilities, respectively.compounds, the former based on a high-temperature expansion Let us first consider the unique class I compound, and the latter on a Monte Carlo simulation.40 In both AV2Mn2Cu3. Fig. 4 shows the in-phase, x¾M, and out-of-phase, approaches SMn=5/2 is treated as a classical spin and SCu= xM, responses as a function of temperature at a frequency of 1/2 is treated as a quantum spin.The high-temperature expan- 125 Hz and a drive amplitude of 3 Oe. x¾M shows a sharp sion was developed up to eleventh order in J/kT where J is maximum at 23 K, which reveals a long-range magnetic the interaction parameter occurring in the spin-Hamiltonian ordering. xM also shows a maximum near this temperature, of the form -JSSMn SCu, which led to the following at 22.5 K.The non-zero value of xM reveals that in the expression: magnetically ordered state the local momenta do not compensate. xM reflects the magnetization lag with respect to the ac xMT= Nb2 3k C2gMn2S2+ 9 4 gCu2-6gMngCuSK field, and depends on both the relaxation time and the drive frequency. Furthermore, x¾M and xM are strongly correlated and represent two components of the same quantity.Assuming +2AgMn2S2+gCu2BK2- 2 3 gMngCuSK3 that the relaxation time in the magnetically ordered phase does not vary dramatically with temperature and that the drive frequency is fixed results in similar temperature depen- - 3 5gCu2K4- 8 45 gMngCuSK5+A2 25gMn2S2+ 533 945 gCu2BK6 + 2 945 gMngCuSK7- A 4 315 gMn2S2+ 5683 14175 gCu2BK8 - 4 525 gMngCuSK9+ A 524 99225 gMn2S2+ 19912 66825 gCu2BK10 + 7108 5457375 gMngCuSK11D with: S=[SMn(SMn+1)]D and K=-JS/2kT where k is the Boltzmann constant and b the electron Bohr magneton.An excellent agreement, as shown in Fig. 2, was found between the experimental data down to 50 K and calculated data for J=-33.1 cm-1, and the local Zeeman factors gMn=2.0 and gCu=2.2. Below ca. 30 K the xM vs. T curves dier between classes I–III. AV2Mn2Cu3 shows a break in the xM vs. T plot at Tc= 22.5 K, corresponding to a long-range ferromagnetic ordering. Compounds of class II behave qualitatively in a similar way, but with Tc around 15 K. On the other hand, the magnetic behaviour of class III compounds is more complex. At very low field xM shows a maximum around 15 K, suggesting that Fig. 3 xM vs. T plot for Na2Mn2Cu3 at various fields. The full lines the long-range orderingis antiferro- rather than ferro-magnetic. are guides for the eye. H=0.05 (#), 0.10 ($), 0.15 (+), 0.20 ('), 0.40 kOe (%). This maximum of xM vanishes when applying a magnetic field 1266 J. Mater. Chem., 1997, 7(7), 1263–1270Fig. 4 In-phase, x¾M ($), and out-of-phase, xM (#), ac molar magnetic susceptibilities for AV2Mn2Cu3 Fig. 6 In-phase, x¾M ($), and out-of-phase, xM (#), ac molar magnetic susceptibilities for Na2Mn2Cu3. The full line is a guide for the eye. dences of x¾M and xM. The rapid decrease of both x¾M and xM below Tc may have two origins: (i) for microcrystalline powders the particles may form monodomains and Landau theory for and II compounds, the xM signal remains very weak, even at second-order phase transitions predicts such behaviour;41,42 5.5 K.It follows that at 15 K the material exhibits a transition (ii) measurements are carried out under a very weak ac field, toward a magnetically ordered state without spontaneous and the hysteresis loop starts to open at temperatures lower magnetization.This transition is therefore antiferromagnetic. than Tc. The very weak coercivity observed for the A2Mn2Cu3 Another magnetic transition takes place around 5.5 K toward compounds of classes I and II might be sucient to give rise a magnetically ordered state, apparently with spontaneous to the observed behaviour.41,43 magnetization. A weak negative xM signal is observed in the We can consider (NBu4)2Mn2Cu3 as a typical example of a temperature range 9–20 K, with surprisingly a minimum value compound belonging to class II.The temperature dependences around 15 K. The metamagnetic behaviour of Na2Mn2Cu3 is of the x¾M and xM responses are shown in Fig. 5. x¾ shows a nicely demonstrated by the temperature dependences of xM very sharp peak at 15 K, and xM an even sharper peak at the under various dc fields.While xM is very slightly negative same temperature, which can be regarded as the critical around 15 K in zero field, it shows a positive peak at 100 Oe. temperature. In other respects, the high-temperature side of the xM curve reveals a weak but significant negative out-of- A2Co2 [Cu(opba)]3 nsolv compounds phase response.This behaviour was seen for dierent samples arising from dierent synthetic procedures, and may be con- All the cobalt derivatives present similar magnetic properties sidered as an intrinsic property. We note that negative out-of- and are more stable than the manganese derivatives, those phase responses have already been observed by several with large cations (NEt4+ and NBu4+) being perhaps slightly authors.44,45 more stable than those with small cations (Na+, K+, Cs+ Finally, the ac response for Na2Mn2Cu3 as an archetype of and NMe4+).class III compounds is examined (Fig. 6). The in-phase signal shows two peaks, a well pronounced one at 15 K and a Dc magnetic responses. Fig. 7 shows a typical example of rounded one around 5.5 K. The out-of-phase signal, on the a xMT vs.T curve for an A2Co2Cu3 material; i.e. of other hand, shows a single peak at 5.5 K; compared to class I (NBu4)2Co2Cu3. At room temperature xMT is ca. 7 emu K mol-1, and decreases as the temperature is lowered, revealing again a ferrimagnetic behaviour. A minimum of xMT is observed around 95 K. As the temperature is lowered further, xMT increases dramatically, and reaches a maximum around 30 K.In the present case, there is no adequate theoretical model to interpret quantitatively the magnetic susceptibility data. The field dependences of the magnetization are particularly interesting. First, the maximum of .M/.H in the first magnetization curve is observed for a rather large field, and is temperature dependent. For K2Co2Cu3 this maximum occurs at 2 kOe at 5 K, and 1 kOe at 25 K.This behaviour reveals a large magnetic anisotropy. Secondly, even at 50 kOe, saturation is not reached. The magnetization value for K2Co2Cu3 at 50 kOe is 1.9 mB while a saturation value >3 mB may be anticipated. Most remarkably, the A2Co2Cu3 compounds present extremely broad magnetic hysteresis loops, as shown in Fig. 8 for K2Co2Cu3. The coercive field was 4.20 kOe at 5 K, 2.58 kOe at 15 K, and 0.58 kOe at 25 K. For the A2Co2Cu3 materials the coercive field tends to decrease slightly as the size of A+ increases. For instance, the coercive field for (NBu4 )2Co2Cu3 Fig. 5 In-phase, x¾M ($), and out-of-phase, xM (#), ac molar magnetic susceptibilities for (NBu4 )2Mn2Cu3. The full line is a guide for the eye.is found to be 1.40 kOe at 5 K. J. Mater. Chem., 1997, 7(7), 1263–1270 1267Fig. 7 xMT vs. T plot for (NBu4)2Co2Cu3. The insert emphasizes the minimum of xMT characterizing the ferrimagnetic behaviour. Fig. 8 Magnetic hysteresis loops for K2Co2Cu3 at 5 (2), 15 ($) and Fig. 9 Time dependences of the magnetization under 50 Oe for 25 K (#). The full lines are guides for the eye.(NBu4)2Co2Cu3 at (a) T=5 K, (b) T=15 K, (c) T=25 K. The full lines represent a fitting of the experimental data, using logarithmic laws. In the course of this study we made an interesting observation concerning the magnetic relaxation in the ordered state latter temperature is taken as the critical temperature. No as illustrated for (NBu4)2Co2Cu3. The sample was first cooled negative out-of-phase signal was observed.below the critical temperature Tc=29 K in zero field, then a In order to fully characterize the nature of the long-range field of 50 Oe was applied and the time dependence of the magnetic ordering in the A2Co2Cu3 materials, we checked that magnetization was measured. Fig. 9 shows the results at three the maxima of both x¾M and xM did not depend on the dierent temperatures, 5, 15 and 25 K.In order to compare frequency within experimental uncertainty between 10 and the three curves, the magnetization was set at zero at t=0. 1000 Hz and firmly excluded superparamagnetic behaviour.47 The magnetization exhibits an after-eect for >1 h. At the three temperatures the M=f (t) curve was found to follow Discussion logarithmic behaviour, which suggests that the system presents a distribution of energy barriers.46 In this section we discuss (i) the possible structures of the The temperature dependence of the relaxation demonstrates compounds, (ii) the dierences between the magnetic properties that the process is thermally activated.The magnetization of A2Mn2Cu3 compounds belonging to class I and II on the increases faster as the system is closer to the critical tempera- one hand, and to class III on the other, (iii) the mechanism of ture.This behaviour is similar to the temperature dependence the long-range magnetic ordering, and finally (iv) the origin of of the M=f (H) hysteresis loops. In all probability, the hyster- the huge coercivity observed for the A2Co2Cu3 compounds.esis loops for the A2Co2Cu3 compounds are time dependent. Structures of the compounds Ac magnetic responses. A typical example of an ac magnetic response is shown in Fig. 10 for (NBu4)2Co2Cu3. x¾M displays Only the crystal structure of AV2Mn2Cu3 has been solved.29,30 However, for all the A2Mn2Cu3 compounds the dc magnetic a peak at 29 K, and xM displays a peak at 28.5 K and this 1268 J.Mater. Chem., 1997, 7(7), 1263–1270The crucial point concerning the dierences between the class II and III compounds is that the long-range magnetic ordering in zero field is ferromagnetic for the former and antiferromagnetic for the latter, which brings important insights on the mechanism of the long-range ordering in these compounds (see below). Mechanism of long-range ordering The long-range magnetic ordering in two-dimensional compounds may have two origins, namely in-plane magnetic anisotropy and/or interplane interactions.It is well established that there is no long-range ordering for a pure two-dimensional array of isotropic spins.48 The fact that Tc is roughly twice as large for A2Co2Cu3 than for A2Mn2Cu3 compounds suggests that magnetic anisotropy plays a significant role.Even for the manganese derivatives some magnetic anisotropy is present as revealed by the fact that .M/.H is at a maximum for a nonzero applied field. This weak anisotropy which is not sucient to give a hysteresis loop might contribute to long-range ordering. Fig. 10 In-phase, x¾M ($), and out-of-phase, xM (#), ac molar Let us now examine the role of the interlayer interactions.magnetic susceptibilities for (NBu4 )2Co2Cu3. The full line is a guide For A2Mn2Cu3 we have seen that apparently the nature of for the eye. these interactions depends on the size of A+. Such a crossover between antiferro- (AF) and ferromagnetic (F) interactions susceptibility data above ca. 30 K are very similar. These data might result from a competition between exchange (or orbital) are very well described by a two-dimensional model of edge- eects favouring the AF interactions when the layers are close sharing hexagons with MnII ions at the corners of the hexagons to each other and dipolar eects favouring F interactions when and CuII ions at the middles of the edges.It follows that it is the layers are farther apart from each other.If so, then except quite reasonable to assume that these honeycomb like layers for A2Mn2Cu3 compounds of class III, dipolar eects would are present in all compounds. What, as yet, remains unknown dominate. The dipolar interactions are expected to decrease as is the relative position of the layers as well as the position of the interlayer distance becomes very large.This is observed A+ cations between the layers. The A2Co2Cu3 compounds are when using very bulky cations such as [Ru(bipy)3]2+ or obtained as microcrystalline powders with very small grains, [Ru(phen)3]2+.49,50 and attempts to grow single crystals suitable for X-ray dirac- To summarize, we can conclude that long-range ordering tion appear to be fruitless. On the other hand, the situation is arises from the combined eect of in-plane anisotropy and not so severe for A2Mn2Cu3 compounds.The gel technique interplane interactions. Those latter have both an exchange aords hexagonal plate-shaped single crystals albeit too small. and dipolar origins. However, for (NBu4 )2Mn2Cu3 we succeeded in obtaining the lattice parameters, a=9.65 A° , b=34.54 A° , c=15.21 A° , b= Origin of the coercivity in A2Co2Cu3 compounds 101.3°, V=4970 A° 3 in a monoclinic crystal system.It is worth The A2Co2Cu3 compounds to the best of our knowledge are noting that the cell volume is roughly half that of AV2Mn2Cu3, molecular-based magnets exhibiting the strongest coercivity and that the lattice parameters would be compatible with a reported so far and we have already discussed the origin of two-dimensional honeycomb like structures and a basal spac- this.The value of the coercive field at a given temperature for ing of ca. 9.5 A° . a polycrystalline magnet depends on both chemical and struc- We have already pointed out the instability of the A2Mn2Cu3 tural factors such as the size and the shape of the grains within compounds belonging to class III.This instability might be the sample. As far as chemical factors are concerned, the key due to the fact that the A+ cations in class III materials are role is played by the magnetic anisotropy of the spin carriers too small to fit properly the available space both inside the which prevents the domains from rotating freely when applying Mn6Cu6 hexagons and between the layers.the field.31 For A2Co2Cu3 compounds the large coercivity arises from the magnetic anisotropy of the CoII ion in octa- Class II vs. class III A2Mn2Cu3 compounds hedral surroundings. One of the striking results of this work concerns the dierences between class II A2Mn2Cu3 compounds with large A+ cations Conclusion and class III A2Mn2Cu3 compounds with small A+ cations.While the former behave as normal magnets with Tc around What is particularly appealing in the field of molecular materials is to observe physical properties which dier from those 15 K, the latter present a more complex magnetic behaviour, which can be analysed as follows: down to ca. 30 K they of classical materials which is usually a consequence of elaborate molecular structure.Here we have described the magnetic exhibit two-dimensional ferrimagnetic behaviour with a minimum of xMT at the same temperature as for class II compounds, properties of a series of two-dimensional ferrimagnetic compounds, focusing on the long-range ordering. In all cases, the and a rapid increase of xMT below this temperature. Around 15 K, xM in low field and x¾M show a maximum, but not xM.magnetic properties above ca. 50 K are characteristic of the ferrimagnetic regime. As the temperature is lowered, each layer This can be attributed to a long-range antiferromagnetic ordering. A magnetic field of ca. 0.15 kOe is sucient to acquires a very large magnetic momentum resulting from the non-compensation of the MII (MnII or CoII) spins on the one overcome the antiferromagnetic interlayer interactions.These class III compounds may be defined as metamagnets. What hand and the CuII spins on the other. The momenta of the ferrimagnetic layers can couple within the lattice either in an remains unclear is the nature of the transition revealed by the maximum of xM around 5.5 K. A possible hypothesis is the AF or in a F fashion.The former situation is encountered for the compounds A2Mn2Cu3 of class III. The AF interlayer onset of a canted antiferromagnetism (or weak ferromagnetism) at that temperature. interactions, however, are very weak, and can be overcome by J. Mater. Chem., 1997, 7(7), 1263–1270 126920 Y. Pei, S. S. Turner, L. Fournes, J. S. Miller and O. Kahn, J. Mater. applying a magnetic field of 0.15 kOe.These compounds are Chem., 1996, 6, 1521. therefore metamagnets. The latter situation is encountered for 21 J. S. Miller, J. C. Calabrese, D. A. Dixon, A. J. Epstein, all the other compounds. The mechanism of the long-range R. W. Bigelow, J. H. Zhang and W. M. Rei, J. Am. Chem. Soc., ordering probably involves both in-plane anisotropy and 1987, 109, 769.interplane interaction (of both exchange and dipolar origin). 22 W. E. Broderick, J. A. Thompson, E. P. Day and B. M. Homan, Science, 1990, 249, 410. The role of the magnetic anisotropy is further evidenced by 23 G. T. Yee, J. M. Manriquez, D. A. Dixon, R. S. McLean, the very strong coercivity of the CoII containing materials. A D. M. Groski, R. B. Flippen, K. S. Narayan, A.J. Epstein and puzzling result arises from this work, namely the negative out- J. S. Miller, Adv. Mater., 1991, 3, 309. of-phase signals observed for the A2Mn2Cu3 compounds. So 24 W. E. Broderick and B. M. Homan, J. Am. Chem. Soc., 1991, far, we have no convincing interpretation of this behaviour. In 113, 6334. 25 D. M. Eichhorn, D. C. Skee, W. E. Broderick and B. M. Homan, this connection some peculiarities concerning molecular-based Inorg.Chem., 1993, 32, 491. ferrimagnets have already been reported.17,51,52 26 J. S. Miller, J. C. Calabrese, R. S. McLean and A. J. Epstein, Adv. Molecular-based magnetic materials have already shown Mater., 1992, 4, 498. surprising behaviour from chemical, physical and even theoreti- 27 K. Inoue and H. Iwamura, J. Am. Chem.Soc., 1994, 116, 3173. cal perspectives. We are convinced that further new phen- 28 A. Caneschi, D. Gatteschi, R. Sessoli and P. Rey, Acc. Chem. Res., 1989, 22, 392. omena, concerning magnetic as well as other physical 29 H. O. Stumpf, L. Ouahab, Y. Pei, D. Grandjean and O. Kahn, properties, will be observed and we are pursuing our work Science, 1993, 261, 447. along this line. 30 H. O. Stumpf, L. Ouahab, Y. Pei, P. Bergerat and O. Kahn, J. Am. Chem. Soc., 1994, 116, 3866. 31 H. O. Stumpf, Y. Pei, C. Michaut, O. Kahn, J. P. Renard and L. Ouahab, Chem. Mater., 1994, 6, 657. References 32 T. Mallah, S. Thiebaut, M. Verdaguer and P. Veillet, Science, 1993, 262, 1554. 1 J. S. Miller,J. C. Calabrese, A. J. Epstein, R. W. Bigelow, J.H. Zang 33 W. R. Entley and G.S. Girolami, Science, 1995, 268, 397. and W. M. Rei, J. Chem. Soc., Chem. Commun., 1986, 1026. 34 S. Ferlay, T. Mallah, R. Ouahe`s, P. Veillet and M. Verdaguer, 2 J. S. Miller, J. C. Calabrese, H. Rommelman, S. R. Chittipedi, Nature (L ondon), 1995, 378, 701. J. H. Zang, W. M. Rei and A. J. Epstein, J. Am. Chem. Soc., 1987, 35 Proceedings of the Symposium on Ferromagnetic and High Spin 109, 769.Molecular Based Materials, ed. J. S. Miller and D. A. Dougherty, 3 Y. Pei, M. Verdaguer, O. Kahn, J. Sletten and J. P. Renard, J. Am. Mol. Cryst. L iq. Cryst., 1989, 176. Chem. Soc., 1986, 108, 428. 36 Proceedings of the Symposium on the Chemistry and Physics of 4 O. Kahn, Y. Pei, M. Verdaguer, J. P. Renard and J. Sletten, J. Am. Molecular Based Magnetic Materials, ed.H. Iwamura and Chem. Soc., 1988, 110, 782. J. S. Miller,Mol. Cryst. L iq. Cryst., 1993, 323–233. 5 J. S. Miller and A. J. Epstein, Angew. Chem., Int. Ed. Engl., 1994, 37 Proceedings of the Fourth International Conference on Molecule- 33, 385. Based Magnets, ed. J. S. Miller and A. J. Epstein, Mol. Cryst. L iq. Cryst., 1995, 271–274. 6 O. Kahn,MolecularMagnetism, VCH, New York, 1993. 38 H. O. Stumpf, Y. Pei, L. Ouahab, F. Le Berre, E. Codjovi and 7 O. Kahn, Adv. Inorg. Chem., 1995, 43, 179. O. Kahn, Inorg. Chem., 1993, 32, 5687. 8 Y. Nakazawa, M. Tamura, N. Shirakawa, D. Shiomi, 39 V. Baron, B. Gillon, J. Sletten, C. Mathonie`re, E. Codjovi and M. Takahashi, M. Kinoshita and M. Ishikawa, Phys. Rev. O. Kahn, Inorg. Chim. Acta, 1995, 235, 69. (L ondon), 1992, B46, 8906. 40 J. Leandri, Y. Leroyer, S. V. Meshkov, Y. Meurdesoif, O. Kahn, 9 R. Chiarelli, M. A. Nowak, A. Rassat and J. L. Tholence, Nature B. Mombelli and D. Price, J. Phys.: Condens.Matter, 1996, 8, L271. (L ondon), 1993, 363, 147. 41 D. W. Carnegie, C. J. Tranchita and H. Claus, J. Appl. Phys., 1979, 10 D. Gatteschi, Adv. Mater., 1994, 6, 635. 50, 7318. 11 K. Nakatani, J. Y. Carriat, Y.Journaux, O. Kahn, F. Lloret, 42 C. Bellitto, Mol. Cryst. L iq. Cryst., 1989, 176, 465. J. P. Renard, Y. Pei, J. Sletten and M. Verdaguer, J. Am. Chem. 43 A. J. Epstein and J. S. Miller,Mol. Cryst. L iq. Cryst., 1989, 176, 359. Soc., 1989, 111, 5739. 44 W. Fujita and K. Awaga, Inorg. Chem., 1996, 35, 1915. 12 H. O. Stumpf, Y. Pei, O. Kahn, J. Sletten and J. P. Renard, J. Am. 45 A. Bhattacharjee, S. Ijima, F. Mizutani, H. Hayakawa, Chem. Soc., 1993, 115, 6738. M. Hagiwara and K. Katsumata, Mol. Cryst. L iq. Cryst., 1996, 13 H. Tamaki, Z. J. Zhong, N. Matsumoto, S. Kida, M. Koikawa, 286, 141. N. Achiwa, Y. Hashimoto and H. Okawa, J. Am. Chem. Soc., 1992, 46 J. M. Hernandez, X. X. Zhang, F. Luis, J. Bartolome, J. Tejada and 114, 6974. R. Ziolo, Europhys. L ett., 1996, 35, 301. 14 H. Okawa, M. Mitsumi, M. Ohba, M. Kodera and N. Matsumoto, 47 F. Palacio, F. J. Lazaro and A. J. van Duyneveldt, Mol. Cryst. L iq. Bull. Chem. Soc. Jpn., 1994, 67, 2139. Cryst., 1989, 176, 289. 15 S. Decurtins, H. W. Schmalle, H. R. Oswald, A. Linden, J. Ensling, 48 R. L. Carlin, Magnetochemistry, Springer-Verlag, Berlin, 1986. P. Gu�tlich and A. Hauser, Inorg. Chim. Acta, 1994, 216, 65. 49 S. S. Turner, C. Michaut, O. Kahn, L. Ouahab, A. Lecas and 16 S. Decurtins, H. W. Schmalle, P. Schneuwly, J. Ensling and E. Amouyal, New J. Chem., 1995, 19, 773. P. Gu�tlich, J. Am. Chem. Soc., 1994, 116, 9521. 50 J. Larionova, S. S. Scott and O. Kahn, unpublished work. 17 C. Mathonie`re, C. J. Nuttall, S. G. Carling and P. Day, Inorg. 51 S. A. Chavan, R. Ganguly, V. K. Jain and J. V. Yakhmi, J. Appl. Chem., 1996, 35, 1201. Phys., 1996, 79, 5260. 18 M. Ohba, N. Maruono and H. Okawa, J. Am. Chem. Soc., 1994, 52 N. Re, E. Gallo, C. Floriani, H. Miyasaka and N. Matsumoto, Inorg. Chem., 1996, 35, 5964; 6004. 116, 11566. 19 S. Turner, O. Kahn and O. Rabardel, J. Am. Chem. Soc., 1996, 116, 6428. Paper 7/00272F; Received 13th January, 1997 1270 J. Mater. Chem., 1997, 7(7), 1263–