摘要:
CrystEngComm, 2000, 9 Caitlin Horn, Marcia Scudder and Ian Dance* School of Chemistry, University of New South Wales, Sydney, NSW 2052, Australia. E-mail: I. Dance@unsw.edu.au Received 2nd March 2000, Accepted 29th March 2000, Published 6th April 2000 We describe three different crystals, 1A, 1B, 1C, with the composition [Fe(phen)3] I12 (phen = 1,10-phenanthroline). They are crystallised from solutions containing [Fe(phen)3] I2 plus I2, but there is not complete reproducibility in the crystallisation experiments and it appears that kinetic crystallisation is responsible for the formation of these trimorphs. The crystal packing of each is described and analysed in detail, in terms of three types of supramolecular motif: (1) multiple aryl embraces between [Fe(phen)3]2+ cations; (2) [Fe(phen)3]2+···Ix motifs; and (3) polyiodide association. The crystallisation results indicate that the energies of these three types of motifs are closely balanced.The two main types of local [Fe(phen)3]2+···Ix interaction are facial with Ix chains along phen faces, and peripheral with C–H···I interactions in the planes of the phen ligands. The three crystals appear to have different dominant motifs: 1A is controlled by a two-dimensional polyiodide network and a closely associated pair of [Fe(phen)3]2+ cations in a parallel fourfold aryl embrace (P4AE) motif; 1B has a cyclic polyiodide which envelops weakly interacting [Fe(phen)3]2+ cations; 1C is evidently influenced by a chain of [Fe(phen)3]2+ cations linked by P4AE, with I5– and I7– as the walls of the channel around the chain of cations.The expected complementary orthogonality of the phen ligand planes and the polyiodide segments is evident: an I – clamp on one phen ligand occurs in 1B and 1C. 7 3] Crystal supramolecular motifs in trimorphs of [Fe(phen)3] I12 (a) (b) Fig. 1 (a) 6AE for { [Fe(phen)3]2+ }2 (b) P4AE for { [Fe(phen)3]2+ }2. Introduction We are investigating the crystallisation and crystal supramolecularity of compounds which contain [M(phen)3]z+ and [Cu(phen)2I]+ cations (phen = 1,10- phenanthroline) and polyiodide anions.1,2 Metal phenanthroline complexes, like metal–bipyridyl3 and metalterpyridyl4 complexes, can be involved in multiple aryl embraces comprised of concerted offset-face-to-face and/or edge-to-face interactions.5,6 Two of these embraces for [Fe(phen)3]2+ are the sixfold aryl embrace (6AE) comprised of six concerted edge to face interactions [Fig.1(a)], and the parallel fourfold aryl embrace [P4AE, Fig. 1(b)] comprised of one offset-face-to-face and two edge-to-face interactions. The structural chemistry of the variety of polyiodides is diverse7–9 and incompletely organised and understood, partly because there is not a clear differentiation of intramolecular and intermolecular distances. However, the occurrence of linear and bent segments in the larger polyiodides is well-established,7–9 and therefore there is potential for a metrical complementarity between the orthogonality of the ligand planes in [Fe(phen)3]2+ and the angularity of the polyiodides.Our investigations have revealed that the [Fe(phen)3]2+ Ix system is particularly rich, with 13 different crystalline compounds characterised so far, ranging from [Fe(phen) I4 (acetone) to [Fe(phen)3] I18.10 In this set of compounds there are three different crystalline forms of [Fe(phen)3] I12 1, which are the subject of this paper. These trimorphs are labeled 1A, 1B and 1C. We have investigated both the variable crystallisation of these trimorphs and their crystal structures, and analyse in detail the structures of the polyiodides, the crystal packing, and the supramolecular motifs which occur. Finally, we discuss some of the issues and outstanding questions for polymorphism in systems of this type.DOI: 10.1039/b001706jExperimental 4)2 in acetone. The experiments which yielded crystals [Fe(phen)3] (BF4)2 was precipitated by addition of NaBF4 to FeSO4 + 3 phen in water, and [Fe(phen)3] I2 was crystallised by addition of NaI in acetone to [Fe(phen)3] (BF of 1A, 1B and 1C are outlined in Tables 1 and 2. After crystallisation, the products were collected by filtration, and washed with diethylether. 1A, 1B and 1C are stable in air, and to loss of I2 at ambient pressure. All products were identified by single crystal diffraction, and microscopic examination of crystal habit. Crystals of 1A were obtained from various solutions of [Fe(phen)3] I2 in acetonitrile and I2 in acetonitrile, allowed to evaporate slowly at ambient temperature.The compositions are listed as experiments 1–6 in Table 1. However, repetitions of some of these experiments yielded other products. Experiment 1 on another occasion yielded instead [Fe(phen)3] I18. Experiment 2 yielded [Fe(phen)3] I7 as a second product. An experiment with stoichiometry and concentration intermediate between experiments 2 and 3 crystallised [Fe(phen)3] I7. In experiment 7 the temperature of crystallisation was investigated: a more concentrated solution at 60 °C yielded 1A on cooling to 30 °C. This product was then redissolved by returning to 60 °C, and on cooling and standing at 4 °C crystallised instead 1C (experiment 13, Table 1). This was redissolved at 60 °C, and on cooling and crystallising at 18 °C yielded 1A (experiment 8, Table 1). A further cycle of redissolution and crystallisation at 4 °C yielded 1C. Another solution, nominally with the same composition as that in experiments 7 and 8, yielded 1A on three cycles of dissolution at 60 °C followed by crystallisation at 4 °C (experiment 9).One experiment (10) with water added to [Fe(phen)3] I2 and I2 in acetonitrile yielded 1A. Crystals of 1B were obtained from a solution (experiment 11) in acetonitrile to which water had been added in volume ratio 3 H2O : 5 CH3CN. However, this solution simultaneously crystallised also [Fe(phen)3] I7 and [Fe(phen)3] I8. 1B was crystallised alone from a solution in dichloromethane (experiment 12). Crystals of 1C were obtained on two occasions when relatively concentrated solutions in acetonitrile at 60 °C were placed in a refrigerator at 4 °C (experiment 13). There is conflict between the results of experiment 13, where solutions which had previously crystallised 1A at higher temperatures crystallised only 1C at 4 °C, and experiment 9 in which a solution with the same composition was rapidly cooled from 60 to 4 °C and yielded only 1A, with repeated crystallisation of 1A on recycling between 60 and 4 °C.Seeding experiments involved preparation of solutions in acetonitrile which were 5 mM in [Fe(phen)3] I2 and 25 mM in I2, and then addition of a few small crystals of other compounds. The results are listed in Table 2. Since these solutions were very intensely coloured it was not clear whether the seed crystals had dissolved prior to the crystallisation of product.Crystal structures were determined with data collected on a CAD4 diffractometer at ambient temperature. Details are contained in Table 3. Results Crystallisation Many crystallisation experiments were undertaken in attempt to develop reproducible procedures for the crystallisation of 1A, 1B and 1C, and other [Fe(phen)3] Ix compounds where 4 £ x £ 18. The main experimental variables were the solvent system (acetonitrile, dichloromethane, and mixtures of acetonitrile with water, alcohols, dichloromethane, diethyl ether, ethyl acetate, tetrahydrofuran, dimethylformamide and hexane), the I2 / [Fe(phen)3] I2 molar ratio, concentration, and temperature.The experiments which yielded 1A, 1B and 1C are reported in Tables 1 and 2, and the experimental section. However, while each product has been obtained on more than one occasion, there are variable results, and controlled reproducible crystallisation of 1A, 1B and 1C has not yet been achieved, despite a large number of experimental investigations. All indications are that the crystallisations are under kinetic control, with products which are a consequence of the nucleation dynamics. Solutions which were apparently identical yielded different products, not only other trimorphs, but other [Fe(phen)3] Ix compositions. Some solutions crystallised more than one product, but, on the basis of careful visual inspection, no batch of [Fe(phen)3] I12 crystals contained more than one of the trimorphs. The crystalline products are stable, and do not contain solvent.Evidence of the subtlety of the crystallisation conditions, particularly for 1A and 1C, come from experiments in which crystallisation and redissolution were effected in the mixture by cycles of temperature changes. A relatively concentrated solution of [Fe(phen)3] I2 + 5 I2 in acetonitrile at 60 °C was cooled to various temperatures, the crystallisation product identified, then redissolved at 60 °C and recooled to crystallise. One solution yielded first 1A at 30 °C, then 1C at 4 °C, then 1A at 18 °C, then 1C at 4 °C. This limited correlation of product with temperature of crystallisation was not sustained by a repeat experiment (different solution, different flask) which yielded only 1A on three cycles of crystallisation at 4 °C.Solutions of the same composition gave reproduced crystallisation of 1A at 4 °C and reproduced crystallisation of 1C at 4 °C. Again the results are indicative of kinetic control of crystallisation, but further investigation is needed. Seeding experiments were performed in an attempt to control crystallisation products. These experiments used the trimorphs of [Fe(phen)3] I12 as seeds, and also other [Fe(phen)3] Ix compounds we have characterised,10 and yielded crystalline products other than [Fe(phen)3] I12. The experiments which yielded [Fe(phen)3] I12 are outlined in Table 2. In all cases the solvent was acetonitrile, and the concentrations of [Fe(phen)3] I2 and I2 were 5 and 25 mM respectively, so that the stoichiometry was correct for formation of [Fe(phen)3] I12. These seeding experiments yielded either 1A or 1C; 1B was not observed. However there is no evident logic, since seed 1A generated 1C, and seed 1C generated 1A.There is some uncertainty in these seeding experiments, because the intense absorbance of the solutions made it difficult to confirm that the seed crystals had not dissolved prior to crystallisation of product: the solutions used were made sufficiently concentrated in an attempt to avoid dissolution of the crystal, being effectivelyTable 1 Crystallisation experiments Solvent Experiment number 1 CH3 CN 2.8 2 CH3 CN 20 3 CH3 CN 30 4 CH3 CN 40 5 CH3 CN 50 6 CH3 CN 100 7 CH3 CN 5 8 CH3 CN 5 9 CH3 CN 5 10 11 CH3CN + 9 vol% H2O CH3CN + 39 vol% H2O CH2Cl2 CH3CN 12 13 a [Fe(phen)3] I7 and [Fe(phen)3] I8 also crystallised. Table 2 Results of seeding experiments.All experiments used an acetonitrile solution 5 mM in [Fe(phen)3] I2 and 25 mM in I2, and were crystallised at ca. 18 °C Experiment number 14 15 16 17 18 19 20 21 22 a [Fe(phen)3] I8 also crystallised. Table 3 Crystal data for 1A, 1B and 1Ca Properties Formula MCrystal system Space group a/Å b/Å c/Å a/° b/ ° g/ ° V/Å3 Dc /g cm–3 ZmMo/mm–1 2 qmax Crystal decay Min transmission factor Max transmission factor Unique reflections Observed reflections Rmerge RRw a Click here for full crystallographic data (CCDC no.1350/15). Mole ratio I2 / [Fe(phen)3] I2 Initial concentration of [Fe(phen)3]2+ / mM 6.4 2.5 2.0 1.7 1.4 0.8 5.0 5.0 5.0 4.6 5 3.1 5 4.2 5.0 55 Seed None 1A 1C [Fe(phen)3] I6·DMF [Fe(phen)3] I14 1A 1B [Fe(phen)3] I7 [Fe(phen)3] I8 1B 1A 7.726 40 55% 0.18 0.31 4618 3406 0.021 0.045 0.056 C36 H24 Fe I12 N6 C36 H24 Fe I12 N6 C36 H24 Fe I12 N6 2119.3 2119.3 2119.3 Triclinic Triclinic Triclinic P 1 P 1 P 1 11.596(9) 12.796(8) 11.219(5) 13.019(12) 13.992(8) 15.310(9) 18.107(16) 15.574(10) 16.687(9) 75.51(7) 85.56(4) 71.92(4) 74.45(7) 76.31(5) 71.85(4) 79.71(7) 66.37(6) 73.52(4) 2531(4) 2482(3) 2533(3) 2.78 2.84 2.78 2 2 2 7.573 7.569 44 50 19% None 0.34 0.20 0.49 0.35 6191 8898 4176 5942 0.034 0.014 0.064 0.038 0.077 0.050 Product Conditions 1A 1A 1A 1A 1A 1A 1A 1A 1A 20 °C, evap.20 °C, evap. 20 °C, evap. 20 °C, evap. 20 °C, evap. 20 °C, evap. Cooling from 60 to 30 °C Cooling from 60 to 18 °C Cooling quickly from 60 to 4°, recycled to 60 and 4°C 20 °C, sealed 1A 20 °C, sealed 1Ba 1B 1C 20 °C, sealed Cooling quickly from 60 to 4 °C Product 1A 1A 1A 1A 1Aa 1C 1C 1Ca 1C1CTable 4 Cell and molecular volumes, and packing coefficients k Cell volume/Å3 [Fe(phen)3] volumea/Å3 (%) 475 (37.5) 474 (38.2) 468 (36.9) 2531 2482 2533 1A 1B 1C a Calculated within the Connolly surface,19,20 with a probe radius of 1.4 Å.saturated with respect to [Fe(phen)3] I12. The unseeded control experiment crystallised 1A, indicating that it is unlikely that 1A as seed would have dissolved. Nevertheless, we cannot be sure that the seed crystals were present at the time of product crystallisation. Even if the seed crystals had dissolved, the variety of results in Table 2 and in very similar experiments which yielded products other than [Fe(phen)3] I12 confirm the lack of control over the crystallisation product. All indications are that our crystallisations of 1A, 1B and 1C were under kinetic control, and we have no unambiguous indication that thermodynamic products are being formed.We do not yet have an experiment in which one of the trimorphs as initial major product subsequently redissolved with the formation of another of the trimorphs at the same temperature. How can crystallisation control be achieved? We believe that further investigation should concentrate on establishing thermodynamic control, using longer time periods at above ambient temperatures to facilitate heterogeneous conversion to thermodynamic products. Thereafter the effects of major variables such as solvents and molar ratios can be established. Shorter-term crystallisation experiments should use clean conditions. 2: the tri-iodide ion, Crystal structures The molecular geometries of [Fe(phen)3]2+ are normal. The following descriptions focus on the details of the polyiodides and the crystal packing for each trimorph.All three structures are space-group P 1 , with [Fe(phen)3] I12 as the asymmetric unit. Cell volumes, cation and anion volumes, and packing coefficients are contained in Table 4. The different structures of the polyiodide arrays yield slightly different volumes. Interpretation of the polyiodide networks in 1A, 1B and 1C uses the I–I distances. In polyiodides there is a continuous range of I–I distances, from that of the I2 molecule (2.70– 2.75 Å) to about 4.3 Å which is the nominal sum of the van der Waals radii. This distance continuum characteristic of catenated iodine distinguishes it from all other chemical systems where there is clear distinction between short intramolecular distances and long intermolecular distances: the concepts of intra- and inter-molecular interactions are quite indefinite for polyiodide structures.Nevertheless, analysis of many polyiodide structures allows recognition of frequently occurring distance ranges which aid in the interpretation of new structures. Distances in the range 2.7 to 2.8 Å signify essentially uncharged I – has two distances = 3.0 Å (the average is 2.92 Å9); I3 larger polyiodides show distances in the range 3.1 to 3.5 Å, sometimes with some shorter distances of around 2.8 Å in the sequence. k I12 volumea/Å3 (%) 0.679 0.683 0.662 384 (30.3) 73 (30.0) 370 (29.2) 7 Crystal structure of 1A The principal feature is the polyiodide network, which will be described first.Fig. 2 shows the polyiodide network, and the I–I distances. The non-molecular net is comprised of chains of I atoms (types I1, I2, I3) linked by pairs of I chains to form large I16 cycles in distinctive chair conformation. There are also I2 pendant sections (atoms I6, I7) on the chains. –. The – and their neighbours are 3.33, In 1A atom I1 is surrounded at distances of 3.19, 3.43, 3.43, 3.44, 3.55 Å by five end-on I2 units (within which the distances are 2.75 or 2.76 Å) in approximately trigonal bipyramidal stereochemistry. Therefore I1 is formally I–. The "arms" of the I16 chairs (atoms I10, I11, I12) have distances of 2.88, 2.95 Å, and are effectively I3 linkages between these I3 ––I 3.38 Å, which indicate that I3 2 sequences in the chair (i.e.sequences I10, I11, I12, I9, I8, or I12, I11, I10, I5, I4) differ from standard symmetrical I5–. Fig. 2 The polyiodide network in 1A. The net is propagated by the centres of inversion at the vertices and edge-centres of the cell. Atom labels and bond distances are marked.2, 2 Therefore, the main structural units of the polyiodide network can be interpreted as chains of alternating I– and I linked by I16 chairs comprised of 2 I–, 2 I3–, and 4 I2, with the I– at the head and foot of the chair. There is another I attached to the chains at I–. A significant observation is the consistency of the I–I distances, which have an effective symmetry higher than required by the space-group: there is a small number of independent I–I distances, even though there are many atoms in this net.This supports our intepretation of the simplified formal bonding in the polyiodide net, and also signifies its important role in the crystal packing. Fig. 4 Three contiguous polyiodide layers in 1A. The network just described is two-dimensional, occurring as layers parallel to the (110) planes of the crystal. Fig. 3 shows an end view of one layer, and the substantial extension of the I16 chairs from the layer of chains. Fig. 4 shows three contiguous layers, which interpenetrate each other's space but without formal connection. We now introduce the [Fe(phen)3]2+ cations. There is a P4AE between pairs of cations, shown in Fig.5. The cations, and the P4AE pairs, are located on the faces of the polyiodide layers, as shown in Fig. 6. In fact, a centrosymmetric pair of P4AEs is located on either side of the chair in the polyiodide layer (Fig. 7). Fig. 5 The P4AE between a pair of [Fe(phen)3]2+ cations in 1A. The Fe···Fe vector (9.45 Å) is marked with orange/black candystripes in this and subsequent figures. Fig. 3 One layer of the 2D polyiodide net in 1A. This view is approximately perpendicular to that in Fig. 2. Fig. 6 The occurrence of the cations and the P4AEs (indicated by orange/black candystripes) above and below a polyiodide layer. Click image or here for a 3D view.Fig. 7 The occurrence of four cations as two P4AEs, one either side of a chair in the polyiodide layer in 1A.The two 16 membered chairs have been coloured pink. There is no significant interaction between the two phen ligands which approach each other through the centre of inversion at the centre of the chair. Fig. 8 The polyiodide surrounds of one [Fe(phen)3]2+ cation in 1A. Parts of three different polyiodide layers are coloured differently. The chain sections of the blue and purple polyiodide layers are almost parallel to the view direction, and run along the two faces of phen ligand B. Parts of three different chairs surround phen ligands A and C. The I4, I5, I10 linear section of the purple chair lies parallel to the front face of phen A, while the I10, I11, I12 segment of the brown chair is parallel to the back face of phen A. The I8, I9, I12 segment of the blue chair is parallel to the back face of phen C, but is obscured.In order to complete description of the lattice packing, it is necessary to look at the total surrounds of one cation, shown in Fig. 8. There are segments of the polyiodide network aproximately parallel to the faces of each phen ligand, as detailed in the caption to Fig. 8. There are no polyiodide segments parallel and peripheral to phen ligands, as commonly occur in other crystals of phenanthroline metal complexes with polyiodide ligands.1,2 The polyiodide segments do not appear to fit as closely around the faces of the [Fe(phen)3]2+ cation as they do in other structures we have analysed. To summarise, the quality and consistency of the polyiodide network, and the absence of optimum placement of polyiodide segments around the cation, implies that the cation did not strongly template this structure. 1A appears to be a structure determined by the polyiodide network. 12 3]2+ Crystal structure of 1B 5 Fig.9 shows this crystal lattice, viewed along the 1 1 1 direction. There appears to be segregation of the iodine atoms in layers, surrounding pairs of cations. However, there is no significance to the cation pairs in this projection, and there are other more significant motifs between cations and anions. It is first necessary to examine the organisation of the atoms in the polyiodide components. Fig. 10 shows the connection of the 12 independent iodine atoms into an almost planar ion which could be regarded as I122–.The I–I distances range from 2.73 to 3.57 Å, and can be interpreted in various ways. The sequence I1–I2–I3–I4–I5 is typical of bent I – ions, as is the I8–I9–I10–I11–I12 segment, while I6–I7 is the shortest bond and is close to the I–I distance in I2, and so the I122– ion could be deconstructed as I5–·I2· I5–. However, as will be seen, these Ix sequences are closely associated with the [Fe(phen)3]2+ cations, and the details of their geometry are considered to be influenced by that association. The I12 segments are associated in pairs by a centre of inversion, to form an I24 cycle, shown in Fig. 11. The I24 unit is comprised of two parallel but non-coplanar I segments, which account for the appearance of the polyiodide components in Fig.9. The shortest distance between different I24 cycles in the crystal is 3.9 Å, linking cycles within the layers evident in Fig. 9. Each I24 cycle has two [Fe(phen)3]2+ complexes intimately associated with it, as shown in Fig. 12. There is no significant interaction between the two [Fe(phen) complexes, but there are clearly close and favourable juxtapositions of the phen ligands with sections of the polyiodide cycle. In particular, phen ligand C protrudes right through the cycle, and the I1–I2–I3–I4–I5–I6–I7 section is tightly wrapped around this phen ligand. This also allows the I3–I4–I5–I6–I7–I8 segment to be facially parallel to phen ligand B, with atoms I6 and I7 almost centred over two of the 6-rings of phen [Fig.12(c)] at I···C(N) distances of 3.86 to 4.12 Å: recall that I6 and I7 constitute the I2-like section of the polyiodide. The third phen ligand, A, is therefore orthogonal to the planar sections of the cycle, but is sufficiently close to I1–I2 to form the peripheral C–H···I attractions (H···I 3.15 Å) which are commonly observed.1 The complementary orthogonality of Ix segments in polyiodides and phen ligands in [M(phen)3] complexes is well demonstrated in this part of crystalline 1B, and it appears that the polyiodide chain has been templated by the cation. The {[Fe(phen)3]2·I24}complex is the essential repeat unit of this crystal, and so now we describe how they are packed. Fig. 13 shows the lateral packing of three adjacent {[Fe(phen)3]2·I24} complexes.The local interactions which maintain these relationships are C–H···I, and are detailed in Figs. 14 and 15.Fig. 9 The crystal lattice of 1B, projected along the 11 1 direction. Click image or here for a 3D view. (b) (a) Fig. 10 Two views of the I12 connection in 1B. Fig. 11 (a) The centrosymmetric I24 cycle in 1B, in relation to the unit cell. The two I12 units are coloured differently. The I1—I12 distance is 3.71 Å, and the shortest contra-cyclic distance is I10— I10 at 4.8 Å. (b) Side view showing the displacement of the two I12 segments at the I1—I12 connection.(c) (b) (a) Fig. 12 Three views of the centrosymmetric {[Fe(phen)3]3·I24} unit in 1B. Part (c) shows atoms and labels for the space filling representation (b) and identifies phen ligand C which protrudes through the I24 cycle, and phen ligand B which is parallel to the I3–I4–I5–I6–I7–I8 section.Part (a) is rotated 90° about a vertical axis relative to (b). Fig. 13 Top and side view of the lateral packing of three {Fe(phen)3]2·I24} complexes in 1B: the I24 cycles are colour differentiated, for reference in the following figures.Fig. 14 Detail of the C–H···I interactions between the purple and blue {Fe(phen)3]2·I24} complexes shown in Fig. 13. The H···I distances are 3.11, 3.50 Å, and involve phen ligand C. Fig. 15 Detail of the C–H···I interactions (black/white candystripe, 3.06 Å) between the purple and orange {[Fe(phen)3]2·I24} complexes shown in Fig.13. The end-to-end stacking of {[Fe(phen)3]2·I24} complexes involves the only substantial motif between [Fe(phen)3]2+ cations, and is illustrated in Fig. 16. The phen A ligands of two cations are offset-face-to-face (OFF), separated by 3.5 Å, and the I3–I4–I5 segments at the ends of two I24 cycles are also parallel to these faces. This is not a P4AE because there are no edge-to-face interactions between ligands. Our conclusion about the lattice packing for 1B is that it is dominated by the [Fe(phen)3]2+–polyiodide interactions. The connection of the polyiodide segments into planar I12 is due to their association with the phen ligand planes, and the complementary orthogonality of the ligands in [Fe(phen)3]2+ and the extended and bent sections of the polyiodide chain.A characteristic of this lattice is the occurrence of only one substantive interaction between [Fe(phen)3]2+ cations, and that interaction involves only two phen ligands and is less than the standard P4AE. Crystal structure of 1C The crystal structure of 1C is dominated by well-developed chains of [Fe(phen)3]2+ cations engaged in P4AEs. These chains occur along the (011) direction of the lattice, and are surrounded by the polyiodide ions, as shown in Fig. 17. The P4AE chain is shown in Fig. 18. The two crystallographically independent P4AE interactions are very similar, both centrosymmetric, and have P4AE geometries. 5 7 7 The iodine atoms occur as a standard bent I – ion, and as a centrosymmetric pair of U-shaped I – ions which approach co-planarity (Fig.19). These ions form the walls of the channels which contain the [Fe(phen)3]2+ ions: the I5– ions are on one set of opposite walls, and the (I –)2 ions on the other set, as shown in Fig. 20. We now examine the closer relationships of the polyiodide ions with the [Fe(phen)3]2+ cations. Fig. 21 is a view of one cation approximately along its pseudo-threefold axis, showing all nearby sections of the polyiodides, and with the closest C–H···I interactions. It appears from Fig. 21 that most of the polyiodides are peripheral to the cation and its phen ligands, but there is an I7 segment which wraps around the phen ligand C. Details of this "I7– clamp" are shown in Fig. 22. It is analogous to the wrapping of an I7 segment around a phen ligand in 1B (see Fig. 12).Fig.16 Detail of the centrosymmetric offset-face-to-face interaction between the phen A ligands of two [Fe(phen)3]2+ cations, separating the ends of two I24 cycles in 1B. The I3–I4–I5 segments are also facial to phen B, and there is a C–H···I interaction (H···I 3.31 Å) from phen C, marked as black/white candystripe. Fig. 17 The crystal lattice of 1C, projected along the (011) direction, that is along the chains of cations which are surrounded by the polyiodides. Click image or here for a 3D view.Fig. 18 The chain of [Fe(phen)3]2+ cations in 1C, linked by P4AE interactions: the two views are rotated by 90° about the horizontal axis. The separations of Fe atoms along the chain are alternatively 9.38 and 9.45 Å.(c) (b) (a) Fig. 19 The polyiodide ions in 1C. (a) I5–. (b) The centrosymmetric pair of I7– ions. (c) The approach to planarity of the (I7–)2 ion. (b) (a) Fig. 20 I5– ions (blue) and (I7–)2 ions (purple) in the 1C lattice, forming the channels for the chains of [Fe(phen)3]2+ cations (shown only as the Fe sites in orange). (a) View along the channels showing the I5– ions near the (011) planes and the (I7–)2 ions near the (100) planes. (b) View almost perpendicular to (a) with the cation chains running horizontally.Fig. 21 One [Fe(phen)3]2+ cation, viewed approximately along its pseudo-threefold axis, with all nearby segments of polyiodide ions included. The closest C–H···I interactions are marked as black/white candystripes.Note that many of these approach linearity: the H···I distances range from 3.10 to 3.20 Å. (b) (a) Fig. 22 (a) The I7– ion forms a tight clamp around phen ligand C in 1C. (b) Rotation of (a) 90° about a vertical axis shows how one I3 segment of the I7– is also parallel to the face of phen ligand A. There is also an obscured C–H···I from phen B to I7–. Fig. 23 Detail of the P4AE (Fe···Fe 9.38 Å) between [Fe(phen)3]2+ cations in 1C, with the I7– clamp included on one of the ligands. The view direction is exactly perpendicular to the planes of phen C, with the overlapping planes being distinguished by colour. The offset-face-to-face interaction between the phen ligands involves the rim carbon atoms C1, C2, C3, C4, C5, C6, and their bonded H atoms.The I7– clamp is mainly around phen atoms C8, C9, C10. 7 Since the phen ligands have P4AE interactions which engage their faces, and the chain of P4AEs implies that all phen ligands of each cation are so involved, it might seem that none of the ligands could be involved in the I – clamp. Phen ligand C, which is clamped, is the ligand involved in the offset-face-to-face interaction in the P4AE with Fe···Fe separation 9.38 Å. Fig. 23 shows the detail of this P4AE and the clamp on one phen C ligand. It can be seen that in the central section of the P4AE there is overlap only of the outer rim of C and H atoms, and that the clamp is on the other section of phenC. Our conclusion about the lattice packing of 1C is that the structure is organised (perhaps determined) by the chain of embracing cations, and contains standard polyiodide ions, – and I –, which are organised as the walls of the channel 7 I5 7 around the chain of cations.The P4AE chain of cations largely excludes the polyiodides, and the cation–anion interactions are mainly peripheral C–H···I. However, there is one close arrangement of a U-shaped I7– around a phen ligand, similar to that observed in 1B: this "I – clamp" comprises two I3 segments parallel to the face of one phen ligand, while one I3 segment is parallel to a second phen ligand, and there is a C–H···I interaction with the third phen ligand of [Fe(phen)3]2+. It is possible that the degree of offset-face-to-face overlap in one P4AE is diminished by the occurrence of the "I7– clamp" around part of the ligand.3] 3] 3]2+ and the polyiodide segments.7 – 7 Discussion Two of the metrical characteristics of polyiodide ions are the occurrence of linear segments with three or four I atoms (sometimes more), and the occurrence of bends with angles 70–100°. These characteristics are clearly evident in each of the structures reported here, and are well matched to the shapes present on the surface of an octahedral [M(phen) cation, where the ligands are mutually orthogonal extended planar segments. This leads to a concept underlying our investigations of the polyiodides of [M(phen)3] cations, namely the complementary orthogonality of the cations and anions, which allows good metrical matching of [M(phen) with Ix , as well as Ix with Ix, and [M(phen)3] ··· [M(phen)3] embraces.We have evidence of this complementarity in many such structures.10 Therefore, one criterion for assessing crystal packing in these compounds is the occurrence and the quality of complementary orthogonality between [Fe(phen) The three crystal structures for [Fe(phen)3] I12 are quite different. 1A is dominated by a two-dimensional polyiodide network and a closely associated pair of [Fe(phen)3]2+ cations in P4AE motif. The polyiodide network has effective symmetry higher than that of the crystal lattice, indicating that the polyiodide network is the principal influence in this structure.There is a moderate quality of complementary orthogonality. 1C is apparently dominated by a chain of [Fe(phen)3]2+ cations linked by P4AE, with good complementary orthogonality in the form of an I – clamp on one phen ligand. The third structure, 1B, has a cyclic polyiodide which envelops [Fe(phen)3]2+ cations, again with an I clamp and good complementary orthogonality (e.g., Figs. 11, 15), but with only one offset-face-to-face intermolecular attraction between phen ligands (Fig. 16). This diversity of crystal packing types (rather than lesser metrical variations within one type) in these trimorphs implies that the sums of the intermolecular energies of [Fe(phen)3]2+···[Fe(phen)3]2+, [Fe(phen)3]2+···Ix, and Ix···Ix must be closely balanced: we will report calculations of these energies in a subsequent paper.This conclusion about closely balanced intermolecular energies is reinforced by the crystallisation properties.Despite considerable investigation we are not yet able to identify crystallisation conditions or other experimental variables which will control the formation of a specific trimorph. 1A was obtained most frequently, followed by 1C, while 1B is more elusive but has been obtained a number of times under different conditions. We believe that our crystallisation experiments were probably under kinetic control, that is, the nucleation processes determined the products, moreso than the thermodynamics of equilibria in solution or the lattice energies of the crystals.We note that the elusive 1B has the highest density, suggesting a slight thermodynamic advantage. Subtle energy balances occur during the crystallisation, and this is consistent with our belief that subtle energy balances occur also in the summation of the intermolecular energies for the different motif types to give the crystal lattice energy. We are investigating the significance and energies of these standard motifs involving polyiodides by analysis of the CSD, and density functional calculations.10 Further experimental work needs to establish conditions for thermodynamic control of crystallisation in this system. One of the motifs which has been noted but not developed in this paper is the C–H···I interaction involving peripheral C–H bonds of the phen ligands.These have an H···I distance of ca. 3.1 Å, and in many cases the C–H···I sequence approaches linearity. The C–H···I interaction looks like a weak hydrogen bond, but with an attractive energy which is due more to dispersion than electrostatics. The crystallisation of trimorphs of 1 is only part of the story for polyiodides of [Fe(phen)3]2+. We have crystallised and characterised a total of 13 different polyiodides in the class [Fe(phen)3] Ix, ranging from x = 4 to x = 18.10 Our conclusions here about subtle kinetic control of the crystalline product are reinforced by these additional results. As an illustration, there is no correlation between the value of x in the solution stoichiometry of [Fe(phen)3] Ix and the value of x in the product.The trimorphs 1 with x = 12 have been crystallised from solutions in which x ranges from 7.5 to 200. The trimorphism of an I12 compound is significant because there are only four other I12 compounds known, with the cations [K(2,2,2-crypt)]+,11 MePh3P+,12 [Ag([15]aneS5)]+,8 and [Cu(dafone)3]2+ (dafone = 4,5-diazafluoren-9-one).13 2– ions have the "sawhorse" structure, which 12 5– ions joined at the elbows by I2. The I12 These other I is a pair of bent I structures in 1A, 1B, and 1C are unprecedented. Finally, we discuss the classification of 1A, 1B and 1C as trimorphs, in the context of concepts of polymorphism. Polymorphism has been defined as the existence of more than one crystal structure for a compound,14 and has been described as crystal supramolecular isomerism.15 This is largely for organic compounds,16 or at least uncharged molecular compounds, and therefore leads to the corollary that polymorphs of a compound yield the same solution.16 The [Fe(phen)3] Ix system is more complex, firstly by the presence of ions, and secondly because there are almost certainly equilibria in solution involving ions which are not present in the crystal.The species formed by dissolution of the different crystals are not known, and are probably very dependent on solvent and temperature and concentration. Therefore, definitions of polymorphism involving solution identity are likely to be misleading and unhelpful. 2, 3 For inorganic and charged compounds we believe that polymorphism is best considered in terms of the crystal packing, as supramolecular isomerism.However, for crystalline polyiodides even this purely structure-based view is clouded by the lack of a clear distinction between intramolecular and intermolecular I––I interactions, and there is a continuum of I––I distances over a wide range. To some extent polyiodides can be deconstructed into I–, I and I – moieties, and definitions of Ix species can be questionable. + 3 For polyiodides it would be possible to have an alternative definition of polymorphs, as different crystals which grow from the same solution. Indeed, we report an experiment in which solutions containing [Fe(phen)3]2+, I– and I2 crystallise three compounds with different compositions, [Fe(phen)3] I7, [Fe(phen)3] I8, and 1B.Since the polyiodide entities in these have many similarities in terms of fundamental components, and almost certainly are in equilibrium in solution, and probably have similar lattice energies, the essential attributes of polymorphism are present. The literature contains only one other polyiodide system described as polymorphic, namely bis(ethylenedithio)tetrathiafulvalene triiodide (BEDT-TTF)2 I – where there are many polymorphs17,18 grown by electrocrystallisation over periods of weeks. It is clear that formal definitions have limited value here, and that progress requires understanding of the "similar energies" which pervade these discussions. This is a major objective of our continuing investigations. Acknowledgements This research is supported by the Australian Research Council and the University of New South Wales. References 1 C. Horn, Honours Thesis, University of New South Wales, Sydney, 1999. 2 C. Horn, B. F. Ali, I. G. Dance, M. L. Scudder and D. C. Craig, CrystEngComm, 2000, 2, 1. 3 I. Dance and M. Scudder, J. Chem. Soc., Dalton Trans., 1998, 1341. 4 M. L. Scudder, H. A. Goodwin and I. G. Dance, New J. Chem., 1999, 23, 695. 5 V. M. Russell, Honours thesis, University of New South Wales, Sydney, 1999. 6 V. M. Russell, C. Horn, D. Craig, M. L. Scudder and I. G. Dance, manuscript in preparation, 2000. 7 N. N. Greenwood and A. Earnshaw, Chemistry of the Elements, Pergamon Press, Oxford, 1985. 8 A. J. Blake, F. A. Devillanova, R. O. Gould, W.-S. Li, V. Lippolis, S. Parsons, C. Radek and M. Schroder, Chem. Soc. Rev., 1998, 27, 195. 9 P. H. Svensson, PhD Thesis, Lund University, Sweden, Lund, 1998. 10 C. Horn and I. G. Dance, manuscript in preparation, 2000. 11 B. Freckmann and K. F. Tebbe, Z. Naturforsch. B, 1993, 48B, 438. 12 K. F. Tebbe and T. Gilles, Z. Anorg. Allg. Chem., 1996, 622, 138. 13 S. Menon and M. V. Rajasekharan, Inorg. Chem., 1997, 36, 4983. 14 J. D. Dunitz and J. Bernstein, Acc. Chem. Res, 1995, 28, 193. 15 J. D. Dunitz, in Perspectives in Supramolecular Chemistry: The Crystal as a Supramolecular Entity, ed.G. R. Desiraju, Wiley, Cheichester, 1996, pp. 1–30. 16 J. Bernstein, J. Phys. D: Appl. Phys., 1993, 26, B66. 17 J. M. Williams, H. H. Wang, T. J. Emge, U. Geiser, M. A. Beno, P. C. W. Leung, K. D. Carlson, D. L. Thorn, R. J. Thorn and A. J. Schultz, Prog. Inorg. Chem., 1987, 35, 51. 18 N. Yoshimoto, S. Takayuki, H. Gamachi and M. Yoshizawa, Mol. Cryst. Liq. Cryst. A, 1999, 327, 233. 19 M. L. Connolly, Science, 1983, 221, 709. 20 M. L. Connolly, J. Am. Chem. Soc., 1985, 107, 1118. CrystEngComm © The Royal Society of Chemistry 2000
ISSN:1466-8033
DOI:10.1039/b001706j
出版商:RSC
年代:2000
数据来源: RSC