Citation
Quesada Vilar, M. (2007, March 29). Spin-transition frameworks based on bistetrazole and triazine ligands. Retrieved from https://hdl.handle.net/1887/11463
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/11463
Note: To cite this publication please use the final published version (if applicable).
7
Spin-transition studies in a series of
triazine-based Fe II complexes: variable-
temperature structural, thermal,
magnetic and spectroscopic studies
Abstract
The crystal structure of a new mononuclear FeII spin-crossover thiocyanate complex with a polypyridyl triazine-based ligand is reported at three different temperatures. Spectroscopic, thermal and magnetic studies that have allowed the characterisation of the thermal spin- crossover of this compound and its cooperative character are presented. The temperature of the spin transition is close to 200 K, and the process is characterised by thermodynamic changes with parameters ∆trsH = 7.22(2) kJ mol-1 and ∆trsS = 36.4(2) J mol-1 K-1. Two other spin-transition compounds based on cyanoborohydride and dicyanamide are presented, and compared to the fully characterised thiocyanate system.
A part of this chapter has appeared in the literature: Quesada, M.; Monrabal, M.; Aromì, G.; de la Peña-O'Shea, V. A.; Gich, M.; Molins, E.; Roubeau, O.; Teat, S. J.; MacLean, E. J.; Gamez, P.;
Reedijk, J., J. Mater. Chem. 2006, 16, 2669-2676
7.1 Introduction.
Among molecular or molecular-based compounds with interesting magnetic properties, octahedral coordination complexes of d4 to d7 transition metal ions presenting the phenomenon of spin-crossover form a separate group. Indeed, the metal ions in these compounds exhibit a thermally-induced change of their spin-state from a ground low-spin state at low temperatures to a high-spin state at higher temperatures.1, 2 In the case of FeII, this singular property translates into a diamagnetic-to-paramagnetic transition (d6), which is often accompanied by a change in the optical response. The compounds may thusbe of interest as potential materials for magneto-optical switches.3 The temperature of the transition depends mostly on the strength of the ligand-field acting on the transition metal ion, while the sharpness of the observed thermally-induced variation in the magnetic and optical response is related to elastic long–range interactions between the SCO sites in the solid state.4, 5 Abrupt transitions, associated with the so-called cooperativity, are observed when these interactions are strong, either due to crystal packing6-9 or the presence of direct chemical links.10-14 These interaction then allow a rapid propagation through out the material of the structural changes accompanying the spin-state transition.3 The ligands that produce the required ligand-field strength around FeII have so far been restricted to certain families of azole and imine nitrogen donors, in some cases in combination with cyanate anions.15, 16 Therefore, many synthetic efforts are still necessary in order to produce new molecules fulfilling the conditions for the occurrence of spin crossover.1, 2 On the other hand, a large part of the existing knowledge on the physical basis of the phenomenon is based on the observations made on a few well- studied compounds,6-9, 17, 18 for which full spectroscopic and physical investigations have been performed at different temperatures and for which the results could be reproduced theoretically. Such comprehensive studies on new materials are thus still paramount for the evolution of this area.
A series of related FeII dinuclear complexes using a triazine-based polypyridyl multi- chelating ligand has been described in chapter 6. Small differences between these compounds resulted in very diverse magnetic properties, including that of spin-crossover.19 When studying the reactivity of this family of compounds, the reaction of one of its dinuclear species, [Fe2(dpyatriz)2Cl2](CF3SO3)2 (8), with a series of cyano-based ligands, namely thiocyanate, dicyanamide, and cyanoborohydride, was found to result in the formation of three new spin-crossover compounds. The reaction with thiocyanate leads to single crystals with the formula [Fe(dpyatriz)2(NCS)2] (11), which exhibit a spin-crossover behaviour with substantial cooperativity. Full structural, spectroscopic, magnetic and thermal data, at different temperatures, of this new spin-transition compound are presented in this chapter.
These data have allowed measuring the thermodynamic features associated with the spin- transition and its cooperativeness. Furthermore the compounds (12) [Fe(dpyatriz)2(NCBH3)2] and (13) [Fe4(dpyatriz)2(N(CN)2)6](CF3SO3)2·7MeOH are described, and this spectral and magnetic behaviour compared with the fully characterised [Fe(dpyatriz)2(NCS)2] complex.
7.2 Synthesis. a
Reactions of the ligand dpyatriz with various FeII salts under normal conditions of temperature and pressure most often leadto the formation of dinuclear species in which two triazine moieties act as bridging ligands between two FeII atoms.19 The use of higher pressure conditions results sometimes in slight structural variations of the final product, although the main coordination features are maintained. The precursor complex [Fe2(dpyatriz)2Cl2](CF3SO3)2 (8) has been prepared using autogenous pressure conditions (see chapter 6).19 This complex has been subsequently used as the FeII/dpyatriz–containing starting material since it proved to be labile enough and subject to rearrangements. The reaction of 8 with ammonium thiocyanate in methanol, under ambient pressure and temperature, yields the compound [Fe(dpyatriz)2(NCS)2] (11) in high yield (Figure 7.1). The crystallinity of 11 depends on the rate of the crystallisation process, which can be controlled with the amount of solvent used during the preparation. Samples obtained by fast precipitation are more susceptible to oxidation and more likely to contain impurities.
The stability of the precursor complex [Fe2(dpyatriz)2Cl2](CF3SO3)2 (8) in solution has been investigated first, since it is important to know whether this starting compound is altered upon its dissolution in methanol, yielding a mononuclear entity, or if instead, it is the addition of thiocyanate anions that causes the structural modification of the initial dinuclear complex.
The UV spectrum of 8 in methanol does not show any major variation compared to the diffuse reflectance spectrum of the solid-state compound, which suggests that 8 is stable in solution. Indeed, the two bands observed in the solid-state around 950 nm (d−d transitions) and 400 nm (MLCT transitions) are present in the solution spectrum, indicating that 8 is maintained in methanol. Presumably, the ligand-field strength of the thiocyanate anion is responsible for the structural rearrangement from a dinuclear to a mononuclear entity.
Figure 7.1 Labelled ORTEP representation of [Fe(dpyatriz)2(NCS)2] (11) at 150 K, at the 50% probability level. Hydrogen atoms are not shown for clarity.
a See chapter 2 for details on the experimental procedures and recipes for all compounds presented in this chapter.
7.3 Description of the crystal structure of [Fe(dpyatriz)2(NCS)2].
The single crystal X-ray diffraction structure of compound 11 was determined by means of a synchrotron radiation source at 150 (11a), 200 (11b) and 250 K (11c). This variable temperature study has revealed the structure of 11 and has also allowed following the geometrical variations accompanying the thermally activated spin-transition, centred at 200 K, undergone by this compound (see below). Crystallographic data for structures 11a, 11b, and 11c are collected in Table 7.1. Selected interatomic distances and angles for 11 at the three temperatures are listed in Table 7.2.
Table 7.1 Crystallographic parameters of complex 11. When the parameters depend on the temperature, they are entered in the order 150 K (11a), 200 K (11b) and 250 K (11c). b R = Σ||Fo| 2|Fc||/Σ|Fo|, wR2 = [(Σw(Fo2 - Fc2)2/ΣwFo4)]1/2.
Compound 11 is a mononuclear FeII complex (Figure 7.1) with formula [Fe(dpyatriz)2(NCS)2], featuring an FeN6 chromophore, which lies on a crystallographic centre of symmetry. The metal ion is chelated equatorially by the dipyridylamine moieties of two dpyatriz ligands, while the SCN− anions occupy the axial positions, bound through the nitrogen atom. The polydentate ligands span the direction of the unique axis of the central ion (thus, roughly parallel to the SCN− ligands), pointing opposite to each other, each leaving four pyridyl moieties free from coordination. This confers to this species the potential of acting as
“complex-ligand” or building block in the construction of more complicated supramolecular structures. On the other hand, this disposition allows the SCN− ligands to be positioned directly above the triazine rings, for the establishment of S−π interactions between electron- rich sulfur atoms and the electron-deficient rings (distance S···ring centroid; 3.5663(7),
Formula C68H52FeN26O2S2
FW/g mol-1 1385.33
Crystal system Triclinic
Space group P–1
a/Å 11.7103(5);11.7151(7); 11.6660(7) b/ Å 12.8096(6);12.8256(8); 12.8283(7) c/ Å 12.9408(6);13.0725(8); 13.4211(8) α/º 95.223(2);95.3500(10); 95.9420(10) β/º 111.650(2);111.7160(10); 111.8980(10) γ/º 116.420(2);116.3790(10); 115.8500(10) V/Å3 1537.98(12); 1554.48(16); 1589.67(16)
Z 1
ρcalcd/g cm-3 1.496; 1.480; 1.447
T/K 150(2); 200(K); 250(K) Crystal shape Block
Colour Yellow Dimensions/mm 0.06 × 0.04 × 0.04
Unique data 9033; 9158; 9385 Unique data with I > 2σ(I) 7565; 7343; 7126
R 0.0434; 0.0452; 0.0477
wR2b 0.1164; 0.1268; 0.1333
3.5897(9) and 3.6595(10) Å at 150, 200 and 250 K, respectively). Each mononuclear entity is weakly bound to such two equivalent species through S···S interactions (S···S distance of 3.558 Å at 150 K) giving rise to 1D chains (see Figure 7.2). In these chains, the central part consists of a succession of [Fe−N−C−S···S−C−N−Fe] units constituting a sort of “inorganic wire”, protected by the organic shell provided by the dpyatriz ligands. The intermetallic distance within this axis is 12.951 Å, whereas the shortest inter-chain Fe···Fe separation is 11.710 Å. These chains are connected to each other by means of weak π−π interactions established between some of the free pyridyl rings from dpyatriz ligands of adjacent complexes. Both types of intermolecular interactions (S···S and π−π) undoubtedly contribute to the significant degree of cooperativity observed within complex 8 upon the spin crossover (see below).
Figure 7.2 Representation of the molecular structure of [Fe(dpyatriz)2(NCS)2] (11), emphasizing the [S···S] contacts between individual molecules, leading to “inorganic wires”
protected by an organic shell of dpyatriz ligands.
Table 7.2 Selected inter-atomic distances (Å) and angles (º) of the complex [Fe(dpyatriz)2(NCS)2] (11) at 150 K (11a), 200 K (11b) and 250 K (11c).a Denotes symmetry operation [1 - x, 1 - y, 1 - z].
Fe1–N5 1.9927(12) 2.0310(14) 2.2105(14) Fe1–N6 2.0011(15) 2.0380(16) 2.2113(17) Fe1–N13 1.9572(14) 1.9803(16) 2.0971(17) S1–C34 1.6498(17) 1.6455(19) 1.636(2) N13–C34 1.167(2) 1.169(2) 1.160(3) N5–Fe1–N6 86.24(6) 85.51(6) 82.71(6) N5–Fe1–N6a 93.76(6) 94.50(6) 97.29(6) N13–Fe1–N5a 88.70(5) 88.77(6) 88.87(6) N5–Fe1–N13 91.30(5) 91.23(6) 91.13(6) N5–Fe1–N13a 88.70(5) 88.77(6) 88.87(6)
N6–Fe1–N6a 180 180 180
N13–Fe1–N6a 88.29(6) 88.41(7) 88.80(7)
N5–Fe1–N5a 180 180 180
N6–Fe1–N13 91.71(6) 91.59(7) 91.20(7)
N13–Fe1–N13a 180 180 180
N13–C34–S1 179.41(15) 179.66(15) 179.46(16)
Collection of diffraction data at various temperatures have revealed that complex 11 undergoes significant structural changes concomitant with the process of spin crossover that takes place. The most obvious variation is that of the average Fe−N bond distance (Figure 7.3); the low temperature (150 K, 11a) value is 1.984 Å, whereas at the highest temperature (250 K, 11c), it is 2.173 Å, in line with the typical values expected for LS and HS FeII systems, respectively, in this coordination environment. Thus, the increase of the average distance (0.189 Å) with the temperature is within the range expected for the spin crossover of an FeIIN6 system.20, 21 Other interesting changes deal with the contraction of the [−S···S−]
contact distances (0.025 Å) and the decrease of the Fe1−N13−S1 angle (3.58º) with decreasing temperature. The latter variations allow in part the absorption of the strain caused by the increase of Fe−N bond lengths accompanying the spin-transition. This increase is otherwise reflected on the overall changes of the unit cell of the network upon increase of the temperature from 150 to 250 K (∆a = − 0.044 Å, ∆b = 0.019 Å, ∆c = 0.48 Å, ∆V = 52 Å3).
The overall increase of the unit-cell volume (3.4%) takes place anisotropically, occurring primarily along the c axis.
Figure 7.3 ORTEP representations at the 50% probability level of the core of [Fe(dpyatriz)2(NCS)2] (11) at 150 K (left), 200 K (middle) and 250 K (right) emphasizing the change of bond distances with temperature. The labels in the left figure apply also for the others.
7.4 Variable Temperature Infrared Spectroscopy
IR spectra of 11 in a KBr pellet have been recorded at temperatures between 40 and 360 K (see Figure7.4 a), through two successive cycles. At room temperature, typical νCN
vibrations of the thiocyanate ion, with very different intensities, are observed at 2058 (s) and 2080 (w) cm−1. The fingerprint area of the spectrum is dominated by bands of dpyatriz, while a weak band at 434 cm−1 is ascribed to the stretching of the Fe−N bonds. Upon lowering the temperature, a slight blue shift of the thiocyanate bands is observed while their intensity first increases. From 250 K the intensity of the bands at 2060 and 2080 cm−1 start to decrease while a similar pair of peaks appear, with increasing intensities, at 2100 and 2117 cm−1. At 40 K, a broad band around 2070 cm-1 remains, while the band at 2100 cm-1 is the main feature in this region of the IR spectrum. The band at 434 cm−1 decreases in intensity with the lowering of the temperature, as expected for the disappearance of the Fe(HS)−N stretch. In the same manner, as observed for the thiocyanate band, a residue of this band remains with very weak
intensity even at 40 K. A weak band detected near 497 cm−1 at lower temperatures is ascribed to the corresponding Fe(LS)−N stretching, which, as expected, occurs at higher energy as a result of the increased bond order of the low-spin species. Small variations are also observed in the thiocyanate νCS region (690−840 cm−1), while the rest of the spectrum remains practically unchanged. These variations are perfectly reversible and reproducible over two complete cooling-warming cycles. Portions of the spectra including only the thiocyanate νCN
bands for selected temperatures in the second warming step are shown in Figure 7.4 b, inset).
a) b)
Figure 7.4a) Infrared spectra of a ground sample of [Fe(dpyatriz)2(NCS)2] (11) and 41 K and 300 K, emphasizing the differences and similarities between the HS and LS forms of the complex. b) Fraction of HS species of [Fe(dpyatriz)2(NCS)2] (11) upon increasing the temperature from 70 to 300 K, normalised to 1, as calculated from the absorbance of the HS and LS infrared bands. The inset shows the changes in the IR spectra of 11 during the temperature increase at the region of the CN– stretching, employed to calculate the HS vs. LS ratio at each temperature.
The two pairs of bands observed in that area are ascribed to the thiocyanate vibrations in the two different phases of compound 11, i.e. the HS and LS states of FeII. Indeed, as previously observed,22 a change from HS to LS states should shift the vibrations of a coordinated thiocyanate anion towards larger energies of ca. 40 cm−1. A plot of the ratio between the absorbances of the HS and LS main bands versus temperature yields a typical spin-crossover curve (see Figure 7.4 b), which is similar but not identical to the corresponding curve from bulk magnetometry (see below). A similar curve is obtained when plotting the normalised intensity of the Fe−N band against temperature. The residual HS band at low temperature should be associated with a small amount of complexes with a temperature- independent behaviour, being either impurities or defects that could be produced by grinding the sample during the pellet preparation (vide infra).
7.5 Bulk magnetic properties
The temperature dependence of the χmT product of 11 (χm being the molar paramagnetic susceptibility) is sketched in Figure 7.5 a. The 300 K value of 3.078 cm3 mol–1 K is in agreement with an S = 2 HS state for the FeII ion, with g close to 2. Upon cooling, χmT starts to decrease rather sharply around 250 K to reach a plateau at 0.177 cm3 mol–1 K below 150 K. Only at much lower temperatures, a slight additional decrease is observed down to 0.124 cm3 mol–1 K at 1.8 K. This behaviour indicates that compound 11 undergoes a thermal spin crossover from HS (S = 2) to LS (S = 0), centred at ca. 200 K, with 80% of the transition from HS to LS completed within 40 K (∆T80 = 40 K). This behaviour is perfectly reproducible over two cycles, and no indication of hysteresis was observed. The low-temperature behaviour is in agreement with the presence of some remaining HS species, presenting zero- field splitting, which causes the further lowering of the χmT near 10 K. Magnetisation versus field measurements allow an estimation of this remaining fraction of ca. 3.4 % (Figure 7.5 b).
Such a value is much lower than that calculated from the IR data. Therefore, the magnetisation of a sample subjected to the same grinding as that suffered by the IR sample has been measured. The resulting χmT vs. T curve (Figure 7.5) shows indeed a significantly higher residual fraction below the transition with a plateau at ca. 0.5 cm3 mol–1 K, and slightly less sharp transition, in better agreement with the IR results. The estimation of the residual HS fraction of this ground sample from M vs. H data is ca. 6.5 % (Figure 7.5 b).
a) b)
Figure 7.5 a) Plot of χmT vs. T per mole of a polycrystalline sample of 11 (open circles) and a ground sample (full circles). b) Isofield magnetisation vs. field curves of a polycrystalline (open circles) and ground (full circles) sample of [Fe(dpyatriz)2(NCS)2] (11) at 2 K, allowing to estimate the residual fraction of the HS form of the complex at this temperature as ca. 3.4 and 6.5 %, respectively. The solid lines are fits to the Brillouin equation.
7.6 Mössbauer spectroscopy
Mössbauer spectra (see Figure 7.6) have been recorded at various temperatures in the range 80–250 K on a sample of 11 obtained as a crystalline powder. At all temperatures, two doublet signals are observed, one with a large quadrupole splitting (∆EQ ≈ 2.75 mm·s−1) centred at an isomeric shift of ca. 1.1 mm s–1, which is ascribed to FeII ions in the HS state and the other one with a small ∆EQ (≈ 0.36 mm·s−1) centred at an isomeric shift of ca. 0.4-0.5 mm/s, which is ascribed to FeII ions in the LS state. The results of the fitting of these spectra are gathered in Table 7.3. The variation of the relative fraction of both doublets with temperature confirms the presence of a spin-crossover in 11. Nevertheless, Mössbauer data centre the crossover at a lower value, i.e. around 190 K, and indicate a more gradual and incomplete crossover than observed in magnetic measurements (see Figure 7.6 b). A 30% of the iron centres present are still in the HS state at low temperatures (see Table 7.3).
a) b)
Figure 7.6 a)Mössbauer spectra of [Fe(dpyatriz)2(NCS)2] (11) at 250 K and 80 K. b) Relative area of the HS (squares) versus LS (circles) forms of [Fe(dpyatriz)2(NCS)2] (11) at different temperatures, as revealed by Mössbauer spectroscopy, showing a gradual and ca. 70 % complete spin transition.
Table 7.3 Summary of the results from Mössbauer spectroscopy of [Fe(dpyatriz)2(NCS)2] (11) at different temperatures, presented in the HS/LS format. a IS = isomer shift. b ∆EQ = quadrupolar splitting.
T/K ISa/mm s-1 ∆EQb/ mm s-1 Ratio (%) Width/mm s-1 80 1.17(1)/0.495(3) 2.89(2)/0.353(5) 30.6/69.4 0.41/0.32 150 1.116(2)/0.475(1) 2.779(3)/0.350(1) 43.4/56.6 0.35/0.23 180 1.101(2)/0.466(2) 2.752(5)/0.340(6) 51.5/48.5 0.33/0.25 190 1.077(5)/0.451(3) 2.745(2)/0.362(4) 53.3/46.7 0.34/0.32 200 1.078(2)/0.447(5) 2.736(4)/0.355(8) 62.7/37.3 0.32/0.32 210 1.075(3)/0.454(3) 2.721(5)/0.36(1) 70.2/29.8 0.32/0.32 250 1.061(1)/0.416(5) 2.649(2)/0.379(8) 84.8/15.2 0.31/0.43
7.7 Thermal properties
The molar heat capacity under constant pressure, Cp, of 11 has been derived from Differential Scanning Calorimetry measurements at temperatures between 160 and 260 K. A broad heat capacity anomaly is observed between 180 and 220 K culminating at ca. 200 K.
Since this temperature matches the crossover point observed from magnetic and IR data, this heat capacity peak must be associated to the spin-crossover phenomenon. A lattice heat capacity curve has been estimated from the data below 170 K and above 230 K in order to determine the excess heat capacity ∆Cp corresponding to the transition, which is plotted in Figure 7.7 a. The excess enthalpy ∆trsH = 7.22(2) kJ mol–1 and entropy ∆trsS = 36.4(2) J mol−1 K−1, arising from the spin-crossover heat capacity peak, are derived by integration of ∆Cp with respect to T and lnT, respectively. Taking into account the residual HS fraction detected by magnetic measurements (from an identical sample), the figures for a complete transition would be ∆trsH = 7.45(2) kJ mol−1 and ∆trsS = 37.7(2) J mol−1 K−1.From a thermodynamic point of view, the HS fraction at a given temperature can be obtained from the enthalpy gain
∆H(T) with respect to the enthalpy of the LS state by
γ
HS = ∆H(T)/ ∆trsH. The resulting curve is plotted in Figure 7.7 b, together with that derived from magnetic measurements. Given the overall good agreement, the separation of the excess heat capacity from the observed values has been performed reasonably well.a) b)
Figure 7.7 a) Plot of the excess heat capacity, ∆Cp, of compound 11 caused by the process of spin-crossover, calculated from the estimated lattice heat capacities of the LS and HS states.
The solid line is a guide to the eye. b) Plots of the HS vs. LS ratio of [Fe(dpyatriz)2(NCS)2] (11) as a function of T, as determined from magnetic susceptibility (squares) or calorimetric (circles) measurements. The solid lines are best fits to a model using eqn (7.1) (see text, section 7.9).
7.8 Solid-State Absorption Spectroscopy
Compound 11 shows an obvious change of colour on cooling from yellow to red when undergoing the transition. Thus, a solid-state absorption spectrum was measured at both room
temperature (293 K) and liquid nitrogen temperature (~100 K). As can be observed from Figure 7.8, both spectra show two main bands at around 280 nm and about 430 nm. These bands are assigned to MLCT transitions and are responsible for the colour of the compound in its high-spin state. The spectrum at room temperature shows a weak band at around 810 nm that can be assigned to the 5T2g→5E transition, characteristic of the high-spin state. This band decreases in intensity, almost disappearing, at around 100 K. At this temperature, a shoulder at 540 nm is observed which can be assigned to the 1A1 → 1T1 transition. This band is already observed at room temperature, which indicates the presence of a small fraction of LS centres at room temperature, which has indeed been detected by 57Fe Mössbauer spectroscopy. The other expected transition of a LS FeII complex, namely the 1A1 → 1T2 transition, is completely masked by the MLCT charge transfer bands. From the 5T2g → 5E band, the ligand field parameter can be derived as 10DqHS = 12350 cm–1, which is in the range of values commonly observed for a spin-transition compound.3 The overlap between the metal-to-ligand charge transfer band and the 1A1 → 1T1 and 1A1 → 1T2, hampers the calculation of the corresponding 10DqLS.
Figure 7.8. UV-Vis spectrum for compound 11 at 100 K (dotted line) and 293 K (full line).
7.8 General discussion
Compound 11 represents a new addition to the family of mononuclear thiocyanate complexes exhibiting spin-crossover behaviour. Its crossover temperature, ca. 200 K, indeed lies in the range observed for most previously reported similar complexes.15 This new example further supports that the ligand-field strength brought about by pyridyl moieties and thiocyanate ions is in the range required for the occurrence of a thermal spin-crossover phenomenon in an octahedrally coordinated FeII ion. Nevertheless, only a few cases have been reported where the thiocyanate ions are located trans to each other as in 11.23, 24 The structural changes of the coordination sphere observed in 11 upon the spin-crossover are similar to those previously observed in comparable complexes, e.g. cis-[Fe(phen)2(NCS)2],25 or cis- [Fe(bpy)2(NCS)2],26 with a mean ∆Fe–N of 0.189 Å, and a more regularly shaped octahedron
in the LS state. The larger variation of Fe–N(py) distances with respect to the Fe–N(NCS) distances (see Figure 7.3) is likely due to a stronger π-acceptor character of the pyridine ring compared with the NCS– anion. The electron back-donation from filled metal π orbitals (t2g in Oh approximation) to vacant ligand π* orbitals occurring in the LS state is thus expected to be more important for pyridine, resulting in an increased strengthening of the Fe–N bond during the transition from HS to LS.
The discrepancy observed between the spin-crossover curve obtained by means of IR measurements and the magnetic data can be explained by the effect of grinding the sample.
This phenomenon has been already observed with other NCS− –containing spin-crossover compounds,22 and has been confirmed for compound 11 by measuring the variable temperature magnetic susceptibility of a ground sample. This experiment has revealed a higher residual HS fraction at low temperature, and a more gradual curve than observed for a polycrystalline sample. Indeed, this effect has been documented,27, 28 and even studied in detail using ball milling on [Fe(phen)2(NCS)2].29 Such effect, often ascribed to an increased number of defects and/or to smaller crystals,30 may also be related to the existence of various polymorphs with slightly different behaviours, as is indeed the case in [Fe(phen)2(NCS)2].25 The way a sample is prepared and further treated may yield different crossover behaviours.
The discrepancies in both shape and HS/LS fractions of the crossover curve, as observed from Mössbauer data on the one hand and thermal and magnetic measurements on the other hand are also important. One possible origin is the different recoilless fraction of 57Fe nuclei between the HS isomer and the LS isomer. The relative HS fraction has been taken from the area fractions of the HS species assuming equal Lamb-Mössbauer factors for the HS and the LS isomers at a given temperature. If these values are different, the area fraction does not reflect the actual isomer fraction. Although this effect may have an influence on the percentages, the most probable explanation for the discrepancies is the fact that various experiments have been performed on different batches. As observed for the [Fe(phen)2(NCS)2] and [Fe(bpy)2(NCS)2] systems,25, 26 different preparation conditions may lead to several polymorphs, resulting in various degrees of cooperativeness and completion of the spin-crossover phenomenon. A similar effect (see above) is observed for a compound made in different crystals sizes, or with various amounts of defects. The batch used for Mössbauer measurements has been indeed obtained in much larger amounts, and not as large single crystals.
The cooperative nature of a spin-transition arises from significant coupling between the electronic states of the metal ion and the lattice phonons through vibrations, mostly of the metal-ligand bonds. Consequently, the phenomenon proceeds via the formation of domains of like-spin molecules. Calorimetric data on spin-crossover compounds have often been obtained using a domain model, first developed by Sorai and Seki,31 which takes into account the cooperative character of the spin-transition through the number of molecules per interacting domain n. According to this model, the HS fraction,
γ
HS,is given by equation 7.1.(7.1)
1
2 / 1
1 exp 1
1
−
⎪⎭
⎪⎬
⎫
⎪⎩
⎪⎨
⎧ ⎥
⎦
⎢ ⎤
⎣
⎡
⎟⎟⎠
⎜⎜ ⎞
⎝
⎛ − + ∆
= R T T
H n trs γHS
For a non-cooperative system, n is expected to be close to unity, since equation 7.1 becomes then equivalent to the Van ’t Hoff equation, and thus describes a solution-like behaviour. On the other hand, for cooperative compounds, n is found to be large, for example n = 95 in [Fe(phen)2(NCS)2].31 Fitting the HS fraction curve deduced from the calorimetric data to equation 7.1 with ∆trsH = 7.22 kJ mol–1 yields n = 11.6(2) and T1/2 = 198.6(2) K (Figure 7.7, b). Similar treatment of the data obtained from measurement of bulk magnetisation on a polycrystalline sample gives n = 5.7(3) and T1/2 = 201.9(3) K. The smaller n value corresponds to the curve showing a more gradual spin-crossover, although both sets of data are in mutual agreement. Overall, 11 presents a weakly cooperative behaviour, in agreement with the actual value of ∆trsH, which is substantially smaller than that found for more cooperative spin-transition compounds, such as [Fe(btr)2(NCS)2] (btr = 4,4'-bis-1,2,4- triazole)(10 kJ mol−1)32 or [Fe(Htrz)2(trz)](BF4) (Htrz = 1,2,4-triazole, 27 kJ mol–1).33 The entropy gain associated with the spin-transition in 11 (37.7 J mol−1 K−1) is about 2.5 times higher than that arising from a change in spin manifold from S = 0 to S = 2 (Rln5 = 13.4 J mol−1 K−1). The excess entropy gain of 24.3 J mol−1 K−1 results from the changes in internal vibrations, as indicated by the variations observed in IR spectra. Such a small vibrational entropy is usually found in non-cooperative complexes such as [Fe(acpa)2](PF6) (Hacpa = N−(1-acetyl-2-propylidene)(2-pyridylmethyl)amine), [Fe(2-pic)3]Cl2⋅EtOH (2-pic = 2- picolylamine) or [Fe(Hpt)3](BF4)2⋅2H2O (Hpt = 3-(pyrid-2-yl)-1,2,4-triazole) (respectively 28.6, 28.0 and 26.1 J mol–1 K–1),34-36 again consistent with the observations of a weakly cooperative crossover in 11.
These observations can be correlated to the structural parameters associated to the spin-crossover in 11. Indeed, apart from variations in Fe–N distances and the NCS– parameters, no other important modifications of the vibrations in 11 can be expected upon the thermal spin-crossover, thus yielding a rather small vibrational entropy. Although the S···S interactions allow the propagation of these structural variations within the crystal, this is true only within the resulting supramolecular chains. No strong interchain interactions are observed in 11, preventing a 3D cooperative crossover to take place. Indeed, a theoretical study37 of the spin transition in the 1D compound [Fe(Htrz)2(trz)](BF4) showed that the presence of interchain interactions is required to observe a very cooperative behaviour, even if the interactions along the chains are very strong.
7.10 [Fe(dpyatriz)2(NCBH3)2] and [Fe4(dpyatriz)2(N(CN)2)6](CF3SO3)2·7MeOH 7.10.1 Synthesisb
Based on the reactivity of [Fe2(dpyatriz)2Cl2](CF3SO3)2 (8) and on its capacity to rearrange to a distinct new spin-transition compound in the presence of NH4SCN, other cyano-based ligands have been investigated. In the spectral series of ligand-field strength of the cyanide-based ligands, the strength varies as follows:38
b See chapter 2 for details on the experimental procedures and recipes for all compounds presented in this
OCN– < SCN– < SeCN– < (CN)2N– < C(CN)3– < NCBH3–< CN–
As mentioned at the beginning of the present chapter (section 7.2), the dinuclear precursor, compound 8, is structurally not stable in the presence of SCN– anions, giving rise to a mononuclear entity, i.e. [Fe(dpyatriz)2(NCS)2] (11). The ligand-field strength of NCS– appears to be responsible for the reorganisation of the initial material. Consequently, the use of other cyanide-based ligands is expected to induce a comparable rearrangement of the precursor 8. Two cyanide derivatives have been selected, namely –NCBH3 and –N(CN)2,
following the same synthetic procedure as described for the preparation of compound 11. The crystal structure of compound 11 has been solved by means of the synchrotron X-ray diffraction technique. Unfortunately, the reactions of 8 with these other two anions have not resulted in suitable single crystals. The reaction with sodium cyanoborohydride leads to a dark-yellow/orange crystalline powder (compound 12), while sodium dicyanamide leads to a yellow crystalline powder (compound 13).
7.10.2 Structural Characterisation
Structural analyses of both crystalline powders have been carried out by means of InfraRed (IR) spectroscopy, Elemental Analysis (EA) and X-ray Powder Diffraction (XRPD).
Figure 7.9. X-ray powder diffraction patterns for compounds 11, 12 and 13. The XRPD spectrum for compound 11 is calculated from the data obtained by single-crystal X-ray determination.
The IR spectra of compounds 12 and 13 reveal that they both contain dpyatriz, since characteristic bands of this ligand are observed at 1375 and 1367 cm–1, respectively. This band around 1370–1380 cm–1, also observed for compound 11 is specific of the triazine ring- torsion. The IR spectra of compound 12 and 13 show bands at 1016 and 1031 cm–1 respectively, which are assigned to vibrations of the coordinated pyridyl groups.39 Similarly to
uncoordinated pyridyl groups of the dpyatriz ligand. Surprisingly, the IR spectrum of compound 13 does not show this band around 1000 cm–1, which suggests that, contrary to compounds 11 and 12, compound 13 does not possess free ligand-donor sites (namely uncoordinated pyridyl moieties). The IR band due to the cyanide function appears around 2000 cm–1. Compound 11 shows bands at 2060 and 2080 cm–1, which are characteristic of thiocyanate vibrations.40 Similarly, compound 12 shows bands at slightly higher frequencies,
41 i.e. 2352 and 2194 cm–1. Compound 13 shows two main bands 42-44 at 2156 and 2184 cm–1, and three less intense ones at 2222, 2261 and 2312 cm–1.
The comparison of the C, H, N analyses of compounds 11 and 13 suggests that the structure of 13 significantly differs from the one of 11. Indeed, both the carbon and the nitrogen percentages are too high to consider that 13 is structurally related to 11, with two dpyatriz ligands coordinated in the equatorial plane and two axial anions. As revealed by IR spectroscopy, contrary to 11, no free coordination sites appear to be available in 13, therefore suggesting a higher metal:dpyatriz ratio for 13. Indeed, the Fe content represents 11% of all the mass, confirming that compound 13 is a polynuclear complex. Consequently, additional anions are required to compensate the charge of the multimetallic cation, which is fulfilled by triflate anions originating from the initial complex [Fe2(dpyatriz)2Cl2](CF3SO3)2 (8). The presence of triflate anions is confirmed by the characteristic IR vibration band at 1249 cm–1 for compound 13. The C, H, N analyses for the cyanoborohydride complex 12, correspond to a mononuclear entity, analogous to 11, with two equatorial dpyatriz ligands, and two axially coordinated cyanoborohydride anions.
A comparative study of the X-ray powder diffraction (XRPD) patterns for the three compounds reveals some similarities of their solid-state structures. The XRPD pattern for compound 11, calculated from single-crystal X-ray data, is depicted in Figure 7.9. The XRPD pattern of compound 12 shows similarities with the one of compound 11, especially at higher 2θ values. This analogy in the lower distances region (namely the region corresponding to the atomic distances) indicates that the structure of compound 12 only differs at the lattice distances. The XRPD spectrum of compound 13 is less well defined, also as a result of the lower crystallinity of the material. Contrary to 12, the XRPD pattern of compound 13 shares fewer diffraction peaks with that of compound 11. The higher resemblances are found at the higher 2θ values (lower distances).
The structural/composition IR, EA and X-ray powder diffraction analyses have revealed that compound 12 is structurally closely related to compound 11, while compound 13 exhibits a distinct coordination framework. Elemental analysis and IR spectroscopy indicate that compound 12 is most likely constituted of a metal centre, coordinated by two dpyatriz ligands and two anions. The similar X-ray powder diffraction patterns of compounds 11 and 12 suggests that the mononuclear compound 12 possesses a comparable coordination environment, with two axially coordinated anions and two equatorially coordinated dpyatriz ligands. Accordingly, compound 12 can be formulated as [Fe(dpyatriz)2(NCBH3)2].
Compound 13 obviously exhibits a different solid-state structure. The IR and EA analyses suggest that 13 is a polynuclear complex. However, based on the analyses currently
best fit of the elemental analysis is for the following molecular formula:
[Fe4(dpyatriz)2(N(CN)2)6](CF3SO3)2·7MeOH (see Chapter 2, section 2.4).
7.10.3 Thermal Spin-Transitions.
Figure 7.10 shows the χmT vs. T plots for compounds 12 and 13. Similarly to 11 (see Figure 7.5), 12 and 13 display a spin-transition behaviour. Compound 12 (white circles) exhibits a one-step transition centred at around room temperature (T1/2 = 300 K), extending over 150 K (∆T80 = 91 K), and involving 75% of the iron centres of the bulk material. The χmT value at 400 K is 3.27 cm3mol-1K (
γ
HS = 1), and reaches a value of 0.81 cm3mol-1K already at 200 K (γ
HS = 0.25). This value remains constant until 6 K, where a second decrease, due to the zero-field splitting of the remaining iron(II) centres sets in. If the compound is heated, the same χmT value is obtained, which is considered as the fully populated HS state(γ
HS = 1). The magnetic behaviour is reproducible over several heating and cooling cycles, indicating that no chemical modification occurs during the heating process of compound 12. In contrast to compound 12, compound 13 shows a multi-step transition which extends over almost 300 K (black squares). The χmT value at 400 K is 12.50 cm3mol-1K (γ
HS = 1) and reaches 1.18 cm3mol-1K (γ
HS = 0.1) at 6 K. The first step of the multi-step transition is situated at around 278 K, and involves 50% of the total FeII centres of the bulk material.Interestingly, this first step shows two distinctive parts with disparate slopes. The initial part of the transition is more gradual, and involves 15 % of the total HS fraction, while the second part shows a steeper slope, and involves 35 % of the total HS fraction. At around 212 K,
γ
HS = 0.5, the transition flattens and enters in a so-called plateau, which expands over 58 K, involving 13% of the FeII centres. At temperatures lower than 154 K, the second transition sets in (T1/2 = 130 K), and around 22% of the FeII centres switch to the low-spin state. A residual HS fraction of 15% is present at low temperatures.a) b)
Figure 7.10. a)
γ
HS vs. T plot for compound 12 (white circles) and compound 13 (black squares). b) Absorbance spectrum at room temperature (full line) and 100 K (dashed line) for compound 13.7.10.4 Solid-State Absorption Spectroscopy
The absorption properties of both compounds have been investigated at RT temperature for compound 12, and at RT (~300 K) and ~100 K for compound 13. At room temperature, compound 12 shows the typical transition bands of both, the LS state (1A1 →
1T1) and the HS state (5T2g →5E). From the magnetic susceptibility data, the T1/2 has been determined near 300 K. Thus, at ambient temperature, both states are semi-populated, explaining the observation of both bands corresponding to the LS (1A1 → 1T1) and HS (5T2g
→5E) states.
The room temperature spectrum of compound 13 displays a main band around 900 nm, corresponding to the 5T2g → 5E transition of the high-spin state. A smaller band is observed around 560 nm, which is assigned to the 1A1 → 1T1 transitions. The presence of the characteristic LS band agrees with the magnetic data, and indicates that at room temperature some FeII centres are already in the low-spin state. The low-temperature spectrum shows a decrease in the intensity of the 5T2g →5E band and an increase of the 1A1 → 1T1 band. This feature suggests the transition of the FeII centres from the HS to the LS state.
In both cases, as well as for compound 11, the second transition band of the LS state, namely 1A1 → 1T2, is overlapped by the MLCT transitions. Moreover, for compound 12, the
1A1 → 1T1 transition band is as well overlapped by the MLCT transition bands, preventing the calculation of the 10DqLS parameter. From the 5T2g →5E transition band, 10DqHS values of 12500 cm–1 and 11100 cm–1 for compounds 12 and 13, respectively are determined. A spin- crossover behaviour is expected3 when the 10DqHS value lies in the range 10500–12500 cm−1.
7.10.5 Discussion
Compounds 12 and 13 are two new spin-transition compounds derived from the dinuclear system [Fe2(dpyatriz)2Cl2](CF3SO3)2 (8), whose structure is altered in the presence of thiocyanate, dicyanamide, or cyanoborohydride anions. The structure of compound 12 is related to the one of compound 11, while compound 13, being a tetranuclear compound, is obviously structurally less alike.
The spin-transition behaviour of compound 12 is analogous to the one exhibited by compound 11. The main difference is the shift of T½to higher temperatures, obviously as a result of the stronger ligand-field strength of the (NCBH3)– anions (in comparison with NCS–). Such variations have already been observed in other series of complexes involving these anions, and the corresponding deviations of the T½ are comparable to the one herein reported with compounds 11 and 12.45 In addition, the replacement of the sulfur atoms by BH3 moieties slightly alter the cooperativity of the material (decrease of the cooperativity).
Similarly, various compounds studied by Kaizaki and co-workers show the same tendency,45 as well as [Fe(phen)2(NCX)2] (phen = 1,10-phenantroline, X = S or BH3)46. As already mentioned above, intermolecular interactions occur in 11 through sulfur-sulfur contacts. The substitution of the sulfur atoms by BH3 (compound 12) annihilates these interactions, and results in the decrease of the cooperative behaviour within the ensuing material.
The behaviour of compound 13 differs from those of compounds 11 and 12. As the solid structure of 13 is not elucidated, the reason for such distinct magnetic behaviour is difficult to asses. It can be reasonably considered that 13 is a tetranuclear complex, for which each metal centre undergoes a HS/LS transition at a proper temperature, resulting in a multi-step phenomenon. Such SCO behaviour has already been observed in several compounds,47-49 albeit their crossover appears as a one-step gradual transition. Two-step transitions, which are typical of dinuclear systems,18 although not strictly limited to this type of materials,7, 50 have as a common feature the so-called plateau. Many studies have been dedicated to investigate the nature of the species present in this plateau. Most of the reports mention that [HS-LS]
pairs are present in this area,51 except for the case of [{Fe(NCBH3)(4phpy)}2(µ-bpypz)2], where a mixture of [HS-HS] and [LS-LS] intermediates co-exist.52 Theoretical models indicate that both intramolecular short-range interactions and intermolecular long-range interactions should be considered to reproduce two-step transitions.5, 53 Intramolecular short- range interactions are electrostatic or vibronic in nature, and stabilise [HS-LS] pairs.
Intermolecular long-range interactions allow the stabilisation of like-spin domains.
For compound 13, the plateau observed is quite narrow, thus indicating the occurrence of either weak intermolecular or intramolecular interactions. The presence of weak intermolecular interactions in this system would also explain the gradual transition observed, expanding over more than 200 K. In the case of compound [Fe(dpa)(NCS)2]2bpym (dpa = dipyridylamino and bpym = bipyrimidine),47 the very gradual transition and the absence of a plateau have been assigned to the lack of intermolecular interactions due to the dpa moieties.
Based on this idea, the role of the enthalpy is not crucial.47 Since the crystal structure of compound 13 is not yet available, the structural reasons for such a gradual transition are difficult to asses. It can be stated however, that both the dicyanamide and the dpyatriz ligands are flexible and thus may act as shock absorbers, and therefore generate (if any) only weak intramolecular interactions. This would imply an enthalpy close to that of [(HHS-HS + HLS- LS)/2] for the present [HS-LS] pairs.51, 54 Consequently, strong intermolecular interactions would be needed if the plateau is to be observed. Therefore, the presence of a plateau, although narrow, in 13 suggests the existence of at least some intermolecular interactions.
{[Fe(bztpen)]2[µ-N(CN)2]} also exhibits a two-step transition55 ranging over about 200 K.
In this case, the lack of cooperativity is ascribed to the flexibility of the ligand, to the bending of the dicyanamide anion, and to the lack of strong intermolecular interactions between the entities generating the lattice. None of these reasons can be excluded for compound 13. In any case, the most probable cause for the multistep transition observed for compound 13 is a different transition temperature for each of the FeII centres of the tetranuclear complex.
The UV-vis spectra of compounds 12 and 13 differ from that of compound 11. However, the 10DqHS parameter varies in the expected values for the three complexes, the highest 10DqHS being observed for the anion exhibiting the strongest ligand-field strength. The cyanoborohydride compound 12 presents a 10DqHS close to the limit expected for a spin- transition phenomenon, in agreement with the high transition temperature determined by magnetic susceptibility. These variations can be due to changes in the details of the octahedral geometry. For all three compounds, LIESST experiments have been carried out using a
wavelength of 560 nm.56 For compounds 11 and 12 no population of the metastable high-spin state has been detected, while an excitation to this metastable state could be achieved for compound 13. These different responses may be due to the fact that the 1A1 → 1T1 and 5T2g →
5E absorption bands for compound 13 are well separated, which is not the case for compound 11. The overlap observed in compound 11 may give rise to the depopulation of the metastable state through the reverse process, i.e. reverse-LIESST.57 Additionally, the stability of the light-induced metastable state is influenced by the cooperative nature of the materials, as well as by the transition temperature.58, 59 As evidenced by Hauser and co-workers, high transition temperatures coupled with efficient cooperativity result in fast relaxations of the metastable state back to the LS state. Compared to compound 13, 11 and 12 both possess higher transition temperatures and show steeper transitions, which would also explain the different behaviours observed for the three compounds.
7.11 Conclusions
In conclusion, the new FeII complex [Fe(dpyatriz)2(NCS)2] (11) has been found to display a process of spin-transition centred at about 200 K. The structural changes accompanying this phenomenon have been investigated by single-crystal X-ray diffraction measurements at different temperatures using a Synchrotron radiation source. These X-ray studies have revealed typical variations of the coordination sphere around the metal upon the spin-transition, as well as [S···S] intermolecular interactions within the lattice, giving rise to unidimensional supramolecular chains. The moderate cooperativity within the material, evidenced by its spin-crossover properties investigated using a variety of techniques, is attributed to these [S···S] interactions. Bulk magnetic measurements, infrared and Mössbauer spectroscopy, as well as calorimetric measurements are consistent with the thermally induced spin change. These studies have established transition ∆trsH and ∆trsS values of 7.22(2) kJ mol–1 and 36.4(2) J mol−1 K−1, respectively, and have revealed that the completeness and sharpness of the transition is dependent on the morphology of the solid. Compounds 12 and 13 are also spin-transition compounds. Compound 12 shows an incomplete one-step transition, while the spin-transition for compound 13 occurs in several steps. It appears that the distinct magnetic behaviours originate from the different nuclearities, which are most likely due to the different dentation of the respective anions.
7.12 References
1. Gütlich, P.; Goodwin, H. A., Spin crossover - An overall perspective. In Spin
Crossover in Transition Metal Compounds I, Topics in Current Chemistry, Gütlich, P.;
Goodwin, H. A., Eds. Springer: 2004; Vol. 233, pp 1-47.
2. König, E.; Ritter, G.; Dengler, J.; Nelson, S. M., Inorg. Chem. 1987, 26, 3582-3588.
3. Gütlich, P.; Hauser, A.; Spiering, H., Angew. Chem. Int. Ed. Engl. 1994, 33, 2024.
4. Spiering, H., Elastic interaction in spin crossover compounds. In Spin Crossover in Transition Metal Compounds III, Topics in Current Chemistry, Gütlich, P.; Goodwin, H. A., Eds. Springer: 2004; Vol. 235, pp 171-195.
5. Spiering, H.; Kohlhaas, T.; Romstedt, N.; Hauser, A.; Bruns-Yilmaz, C.; Kusz, J.;
Gütlich, P., Coord. Chem. Rev. 1999, 192, 629-647.
6. Hauser, A.; Jeftic, J.; Romstedt, H.; Hinek, R.; Spiering, H., Coord. Chem. Rev. 1999, 192, 471-491.
7. Köppen, H.; Müller, E. W.; Köhler, C. P.; Spiering, H.; Meissner, E.; Gütlich, P., Chem. Phys. Lett. 1982, 91, 348-352.
8. Sorai, M.; Ensling, J.; Hasselbach, K. M.; Gütlich, P., Chem. Phys. 1977, 20, 197-208.
9. Wiehl, L.; Kiel, G.; Kohler, C. P.; Spiering, H.; Gütlich, P., Inorg. Chem. 1986, 25, 1565-1571.
10. Galet, A.; Muñoz, M. C.; Gaspar, A. B.; Real, J. A., Inorg. Chem. 2005, 44, 8749- 8755.
11. Garcia, Y.; Kahn, O.; Rabardel, L.; Chansou, B.; Salmon, L.; Tuchagues, J. P., Inorg.
Chem. 1999, 38, 4663-4670.
12. Kahn, O.; Martinez, C. J., Science 1998, 279, 44-48.
13. Roubeau, O.; Haasnoot, J. G.; Codjovi, E.; Varret, F.; Reedijk, J., Chem. Mater. 2002, 14, 2559-2566.
14. Vreugdenhil, W.; van Diemen, J. H.; de Graaff, R. A. G.; Haasnoot, J. G.; Reedijk, J.;
van der Kraan, A. M.; Kahn, O.; Zarembowitch, J., Polyhedron 1990, 9, 2971-2979.
15. Guionneau, P.; Marchivie, M.; Bravic, G.; Létard, J. F.; Chasseau, D., Structural aspects of spin crossover. Example of the (FeLn)-L-II(NCS)(2) complexes. In Spin Crossover in Transition Metal Compounds II, Gütlich, P.; Goodwin, H. A., Eds. 2004;
Vol. 234, pp 97-128.
16. Real, J. A.; Gallois, B.; Granier, T.; Suezpanama, F.; Zarembowitch, J., Inorg. Chem.
1992, 31, 4972-4979.
17. Long, G. J.; Grandjean, F.; Reger, D. L., Spin crossover in pyrazolylborate and pyrazolylmethane complexes. In Spin Crossover in Transition Metal Compounds I, Topics in Current Chemistry, Gütlich, P.; Goodwin, H. A., Eds. 2004; Vol. 233, pp 91- 122.
18. Real, J. A.; Gaspar, A. B.; Muñoz, M. C.; Gütlich, P.; Ksenofontov, V.; Spiering, H., Bipyrimidine-bridged dinuclear iron(II) spin crossover compounds. In Spin Crossover in Transition Metal Compounds I, Topics in Current Chemistry, Gütlich, P.; Goodwin, H. A., Eds. Springer: 2004; Vol. 233, pp 167-193.
19. Quesada, M.; de Hoog, P.; Gamez, P.; Roubeau, O.; Aromì, G.; Donnadieu, B.;
Massera, C.; Lutz, M.; Spek, A. L.; Reedijk, J., Eur. J. Inorg. Chem. 2006, 1353-1361.
20. Holland, J. M.; McAllister, J. A.; Kilner, C. A.; Thornton-Pett, M.; Bridgeman, A. J.;
Halcrow, M. A., J. Chem. Soc. Dalton Trans. 2002, 548-554.
21. Yoneda, K.; Adachi, K.; Hayami, S.; Maeda, Y.; Katada, M.; Fuyuhiro, A.; Kawata, S.; Kaizaki, S., Chem. Commun. 2006, 45-47.
22. Figg, D. C.; Herber, R. H., Inorg. Chem. 1990, 29, 2170-2173.
23. Roux, C.; Zarembowitch, J.; Gallois, B.; Granier, T.; Claude, R., Inorg. Chem. 1994,
24. Zhu, D. R.; Xu, Y.; Yu, Z.; Guo, Z. J.; Sang, H.; Liu, T.; You, X. Z., Chem. Mater.
2002, 14, 838-843.
25. Gallois, B.; Real, J. A.; Hauw, C.; Zarembowitch, J., Inorg. Chem. 1990, 29, 1152- 1158.
26. König, E.; Madeja, K.; Watson, K. J., J. Am. Chem. Soc. 1968, 90, 1146-1153.
27. Haddad, M. S.; Federer, W. D.; Lynch, M. W.; Hendrickson, D. N., Inorg. Chem.
1981, 20, 131-139.
28. Tsuchiya, N.; Tsukamoto, A.; Ohshita, T.; Isobe, T.; Senna, M.; Yoshioka, N.; Inoue, H., J. Solid State Chem. 2000, 153, 82-91.
29. Müller, E. W.; Spiering, H.; Gütlich, P., Chem. Phys. Lett. 1982, 93, 567-571.
30. Real, J. A.; Andrés, E.; Muñoz, M. C.; Julve, M.; Granier, T.; Bousseksou, A.; Varret, F., Science 1995, 268, 265-267.
31. Sorai, M.; Seki, S., J. Phys. Chem. Solids 1974, 35, 555-570.
32. Martin, J. P.; Zarembowitch, J.; Dworkin, A.; Haasnoot, J. G.; Codjovi, E., Inorg.
Chem. 1994, 33, 2617-2623.
33. Kröber, J.; Audière, J. P.; Claude, R.; Codjovi, E.; Kahn, O.; Haasnoot, J. G.; Groliere, F.; Jay, C.; Bousseksou, A.; Linarès, J.; Varret, F.; Gonthiervassal, A., Chem. Mater.
1994, 6, 1404-1412.
34. Jakobi, R.; Romstedt, H.; Spiering, H.; Gütlich, P., Angew. Chem. Int. Ed. Engl. 1992, 31, 178-180.
35. Roubeau, O.; de Vos, M.; Stassen, A. F.; Burriel, R.; Haasnoot, J. G.; Reedijk, J., J.
Phys. Chem. Solids 2003, 64, 1003-1013.
36. Sorai, M.; Maeda, Y.; Oshio, H., J. Phys. Chem. Solids 1990, 51, 941-951.
37. Linarès, J.; Spiering, H.; Varret, F., Eur. Phys. J. B 1999, 10, 271-275.
38. Murray, K. S.; Kepert, C. J., Cooperativity in spin crossover systems: Memory, magnetism and microporosity. In Spin Crossover in Transition Metal Compounds I, Topics in Current Chemistry, Gütlich, P.; Goodwin, H. A., Eds. Springer: 2004; Vol.
233, pp 195-228.
39. Casellas, N.; Gamez, P.; Reedijk, J.; Mutikainen, I.; Turpeinen, U.; Masciocchi, N.;
Galli, S.; Sironi, A., Inorg. Chem. 2005, 44, 7918-7924.
40. Kabesova, M.; Gažo, J., Chem. zvesti 1980, 34, 800-841.
41. Kodera, M.; Terasako, N.; Kita, T.; Tachi, Y.; Kano, K.; Yamazaki, M.; Koikawa, M.;
Tokii, T., Inorg. Chem. 1997, 36, 3861-3868.
42. Armentano, D.; De Munno, G.; Guerra, F.; Julve, M.; Lloret, F., Inorg. Chem. 2006, 45, 4626-4636.
43. Mohamadou, A.; van Albada, G. A.; Kooijman, H.; Wieczorek, B.; Spek, A. L.;
Reedijk, J., New J. Chem. 2003, 27, 983-988.
44. Riggio, I.; van Albada, G. A.; Ellis, D. D.; Spek, A. L.; Reedijk, J., Inorg. Chim. Acta 2001, 313, 120-124.
45. Nakano, K.; Suemura, N.; Kawata, S.; Fuyuhiro, A.; Yagi, T.; Nasu, S.; Morimoto, S.;
Kaizaki, S., Dalton Trans. 2004, 982-988.
46. Purcell, K. F.; Edwards, M. P., Inorg. Chem. 1984, 23, 2620-2625.
47. Gaspar, A. B.; Ksenofontov, V.; Real, J. A.; Gütlich, P., Chem. Phys. Lett. 2003, 373, 385-391.
48. Ruben, M.; Breuning, E.; Lehn, J. M.; Ksenofontov, V.; Renz, F.; Gutlich, P.;
Vaughan, G. B. M., Chem.-Eur. J. 2003, 9, 4422-4429.
49. Ruben, M.; Breuning, E.; Lehn, J. M.; Ksenofontov, V.; Renz, F.; Gütlich, P.;
Vaughan, G. B. M., Chem.-Eur. J. 2003, 9, 5176-5176.
50. Grunert, C. M.; Schweifer, J.; Weinberger, P.; Linert, W.; Mereiter, K.; Hilscher, G.;
Müller, M.; Wiesinger, G.; van Koningsbruggen, P. J., Inorg. Chem. 2004, 43, 155- 165.
51. Ksenofontov, V.; Gaspar, A. B.; Niel, V.; Reiman, S.; Real, J. A.; Gütlich, P., Chem.- Eur. J. 2004, 10, 1291-1298.
52. Nakano, K.; Kawata, S.; Yoneda, K.; Fuyuhiro, A.; Yagi, T.; Nasu, S.; Morimoto, S.;
Kaizaki, S., Chem. Commun. 2004, 2892-2893.
53. Gütlich, P.; van Koningsbruggen, P. J.; Renz, F., Struct. Bond. 2004, 107, 27-75.
54. Zein, S.; Borshch, S. A., J. Am. Chem. Soc. 2005, 127, 16197-16201.
55. Ortega-Villar, N.; Thompson, A. L.; Muñoz, M. C.; Ugalde-Saldivar, V. M.; Goeta, A.
E.; Moreno-Esparza, R.; Real, J. A., Chem.-Eur. J. 2005, 11, 5721-5734.
56. Decurtins, S.; Gütlich, P.; Köhler, C. P.; Spiering, H.; Hauser, A., Chem. Phys. Lett.
1984, 105, 1-4.
57. Hauser, A., Chem. Phys. Lett. 1986, 124, 543-548.
58. Hauser, A.; Gütlich, P.; Spiering, H., Inorg. Chem. 1986, 25, 4245-4248.
59. Hauser, A.; Vef, A.; Adler, P., J. Chem. Phys. 1991, 95, 8710-8717.
