Received 1 May 2013
1,4-Bis[2-(4-ferrocenylphenyl)ethynyl]anthraquinone from synchrotron X-ray powder diffraction
Maki Sachiko,a Eiji Nishibori,a* Shinobu Aoyagi,b Makoto Sakata,c Masaki Takata,a Mio Kondo,d Masaki Murata,d Ryota Sakamotod and Hiroshi Nishiharad
aRIKEN SPring-8 Center, 1-1-1 Kouto, Sayo-cho, Sayo-gun, Hyogo 679-5148, Japan,bDepartment of Information and Biological Sciences, Nagoya City University, Nagoya 467-8501, Japan,cJapan Synchrotron Radiation Research Institute, SPring-8, 1-1-1 Kouto, Sayo-cho, Sayo-gun, Hyogo 679-5198, Japan, and dDepartment of Chemistry, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
The title compound, [Fe2(C5H5)2(C40H22O2)] or 1,4-(FcPh)2Aq [where FcPh is 2-(4-ferrocenylphenyl)ethynyl and Aq is anthraquinone], was synthesized in an attempt to obtain a new solvent-incorporating porous material with a large void space. Thermodynamic data for 1,4-(FcPh)2Aq show a phase transition at approximately 430 K. The crystal structure of solvent-free 1,4-(FcPh)2Aq was determined at temperatures of 90, 300 and 500 K using synchrotron powder diffraction data. A direct-space method using a genetic algorithm was employed for structure solution. Charge densities calculated from observed structure factors by the maximum entropy method were employed for model improvement. The final models were obtained through multistage Rietveld refinements. In both phases, the structures of which differ only subtly, the planar Aq fragments are stacked alternately in opposite orientations, forming a one-dimensional column. The FcPh arms lie between the stacks and fill the remaining space, leaving no voids. C-H interactions between the Ph and Fc fragments mediate crystal packing and stabilization.
The concept of joining electron-donor (D) and -acceptor (A) fragments into single molecules has attracted much attention, owing to the structures and characteristic properties of the resulting compounds (Alberola et al., 2003; Ferraris et al., 2003). One group of D-A conjugated complexes, the ferrocenylethynylanthraquinones (FcAq), show guest-molecule absorption and valence tautomerization, among other propeties. T-shaped 1,4-bis(2-ferrocenylethynyl)anthraquinone, 1,4-(Fc)2Aq, exists in both a solvent-containing porous structure and a solvent-free nonporous structure, which are reversibly transformed, with a change of the D-A arrangement, upon desorption and adsorption of guest molecules. Many organic solvents, such as CH2Cl2, o-dichlorobenzene, tetrahydrofuran, trichloroethylene and hexane, can be entrapped in the porous phase (Kondo et al., 2006). In an attempt to expand the pore size, we synthesized the title compound, 1,4-(FcPh)2Aq, (I), by inserting phenylene rings between the ethynyl and ferrocenyl fragments. We report here the crystal structures of (I) determined at three temperatures using synchrotron radiation (SR) X-ray powder diffraction.
The brown crystals of (I) were small, less than 1 µm in a linear dimension. Using synchrotron radiation, powder profiles were measured at ten temperatures from 300 to 500 K with short (5 min) X-ray exposure times (Fig. 1a). For temperatures from 440 to 500 K, a lower-temperature peak at 3.5° in 2 disappeared and a new peak at 9° in 2 appeared. The reduced unit cells derived from the 300-420 K data were triclinic type II (all angles obtuse). The reduced cells from 440 to 500 K were triclinic type I (all angles acute), indicating that there is a phase transition at approximately 430 K. The lattice constants of the reduced cell at T = 500 K are a = 7.7447 (10) Å, b = 9.9593 (15) Å, c = 24.101 (3) Å, = 87.526 (5)°, = 83.225 (6)° and = 84.867 (4)°. The reduced cells at T = 300 K (Fig. 1b, black) and T = 500 K (Fig. 1b, blue in the electronic version of the paper) are clearly different.
The choice of unit cell for the 500 K structure determination merits comment. The reduced cell is normally used for primitive structures, but in the case of a phase transition it is more useful to set the new phase on a valid primitive cell that is as closely related to the original cell as possible. A judicious choice of cell and asymmetric unit can reveal possible pathways for the phase change. In the present case, for the high-temperature phase, we selected the primitive cell that was closest to the reduced cell at T = 300 K. Beginning with the reduced cell at T = 500 K, the primitive cell that we used results from the following transformation (subscript `p' means that the primitive cell was used, and subscript `r' refers to the reduced cell): ap = -ar, bp = br and cp = ar - cr. The primitive cell used for the 500 K structure is shown in Fig. 1(b) (red in the electronic version of the paper). This cell is close to the reduced cell at 300 K.
The unit-cell volumes from 300 to 500 K are shown in Fig. 1(c). Lattice constants are listed in Table 1. From 300 to 420 K, a, b, c and V increase linearly with temperature. The discontinuity in the slope at T = 440 K in Fig. 1(c) coincides with the phase transition at approximately 430 K.
For structure determination, powder diffraction profiles were measured with long X-ray exposure times at 90 (240 min), 300 (240 min) and 500 K (120 min). The structures were solved using a direct-space method with a genetic algorithm (GA; Harris et al., 1998). The structure model for the GA was constructed using similar structures, viz. 1,4-Fc2Aq, 1,5-Fc2Aq and 1,8-Fc2Aq (Kondo et al., 2006; Murata et al., 2001). Chemically equivalent distances were set equal in the model. GA analyses were carried out using both space groups P1 and P; the solutions were almost identical, so space group P was used for the remainder of the analyses. The initial stage of refinement consisted of rigid-body Rietveld refinement. Charge-density calculations using the maximum entropy method (MEM) (Sakata & Sato, 1990) were used to improve the structural model. The omit-difference-MEM maps and the difference-MEM maps were effective for revealing structural details, such as small differences in bond angles (Nishibori, Ogura et al., 2008; Nishibori, Nakamura et al., 2008).
The refinement procedures are shown schematically in Fig. 2. The difference MEM charge densities based on the model from the initial Rietveld refinement are shown as a mesh in the upper part of the figure, along with the model. There is notable residual density on the lower side of the Aq and Ph fragments, indicating a need for further adjustment of the relative dispositions of the Fc, Ph and Aq residues. Rigid-body refinement was conducted with additional internal rotation axes to permit slight deformation of the molecule while keeping reasonable interatomic distances; this is indicated in the middle part of the figure. The two internal rotation axes pass through atoms C2 and C5 and are perpendicular to the ring centred at Cg5 (see Supplementary materials ). In the lower part of Fig. 2, the difference MEM charge densities based on the revised model are shown as a mesh, along with the results of the Rietveld refinement. The figure clearly shows a reduction in the difference MEM charge densities and an improvement in the powder fittings by the addition of two rotation parameters. Restrained Rietveld analysis was employed for the final refinement.
The molecular structures of (I) at 90, 300 and 500 K (Fig. 3) can be superimposed with essentially complete overlap of the Aq groups (Fig. 4). The largest deviations are observed on the left-hand side of Fig. 4, where it is seen that the Fc-Ph-ethynyl fragment at 500 K deviates from its position and orientation at the other temperatures. In describing the structures, we shall refer to the nine carbon-containing rings in the molecule (four five-membered and five six-membered rings) by the names of their centres of gravity, Cg1-Cg9, as shown in Fig. 4.
The relative orientations of the terminal Fc groups can be characterized using their axial directions, which are the vectors joining the centres of gravity of their respective cyclopentadienyl (Cp) rings, viz. Cg1Cg2 and Cg3Cg4. The angle between the CgCg vectors of the two terminal groups is 125.2° at 90 K, 125.6° at 300 K and 104.1° at 500 K. The dihedral angles between the two Ph groups and their respective adjacent Fe groups (Cg2 and Cg9, and Cg4 and Cg8) are listed in Table 2. The angle between Cg9 and Cg2 is much smaller at T = 500 K than at 90 and 300 K. Within the ferrocenyl groups, at T = 90 and 300 K, the two Cp rings are not eclipsed, in contrast with what was found for 1,4-Fc2Aq (Kondo et al., 2006). The Cg3Cg4 Fc group is nearly staggered at these temperatures. All of the Fc groups are nearly eclipsed at T = 500 K.
In the absence of hydrogen bonding, crystal packing is mainly stabilized by - interactions between Aq groups, propagating parallel to the a axis, as shown in Fig. 5 (dashed lines). Similar -stacking patterns were found in the Fc2Aq compounds 1,4-Fc2Aq, 1,5-Fc2Aq and 1,8-Fc2Aq (Murata et al., 2001). Geometric information for the - interactions is given in Table 3. The CgCg distances for these interactions change with increasing temperature, but the perpendicular distances from the Cg of one ring to the plane of the other do not change as significantly. These - interactions are maintained after the phase transition.
C-H interactions are also present, involving Cg8 and Cg9 as acceptors (Tables 4, 5 and 6). These also stabilize the packing parallel to the a axis. Congeners of the C28-H66Cg8iii and C21-H63Cg9ii interactions are shown as dashed lines in Fig. 5 (symmetry codes as in Table 4). The C28-H66Cg8iii contact is lengthened and the C21-H63Cg9ii contact becomes shorter with the phase change between 300 and 500 K.
The columns of molecules parallel to the a axis, mediated by - stacking as just described, are joined to neighbouring columns by self-complementary C-H interactions, C31-H68Cg4iv, shown in Fig. 6 for T = 300 and 500 K (symmetry code as in Tables 4, 5 and 6). The direction of this interaction alternates between successive molecules along the stack. These C-H interactions are maintained through the phase transition.
Another two C-H interactions are present at T = 90 and 300 K (Fig. 7, and Tables 4 and 5). Firstly, C14-H59Cg8i (symmetry code as in Table 4) joins a C-H group from the central Aq group and a Ph ring of the molecule related by b-axis translation. This contact is the weakest of the C-H interactions. Secondly, C52-H69Cg9v connects a Cp ring donor to a ring of a molecule in a neighbouring stack. The H69Cg9v distance increases from 2.770 (3) Å at 90 K to 2.925 (3) Å at 300 K.
The large temperature dependence of the HCg distance is related to the phase transition. The donor in the C52-H69Cg9v interaction belongs to Cg4, a Cp group of the ferrocenyl which has a staggered conformation at 90 and 300 K, as opposed to the eclipsed conformation found for the Cp rings in other FcAq materials (Kondo et al., 2006; Murata et al., 2001). At temperatures below the phase transition, it would appear that the electrons of Cg9v attract atom H69 to the point of rotating the donor Cg4 into a staggered conformation. With increasing temperature the system acquires enough energy to overcome the C-H attraction and enable the phase transition, with concomitant Cp rotation to an eclipsed conformation for Cg4 and a slight recoil at the opposite end of the molecule, where Cg9 is located. This double adjustment appears to be the driving force for the phase transition.
The phenylene ring that was inserted into the molecule with the goal of obtaining a porous structure plays several roles in the crystal packing. With the aim of characterizing the -electron density on the ring, quantum-chemical calculations were performed on the molecule based on the structure at T = 500 K, using density functional theory at the DFT-B3LYP level with 3-21 G basis sets. Fig. 8 shows the highest unoccupied molecular orbital (HOMO). This orbital is distributed among the rings centred at Cg8, Cg9 (phenylene) and Cg5 (anthraquinone), along with the ethynyl and ferrocenyl groups. Thus, the insertion of the phenylene rings has added two more fragments capable of participating in intermolecular interactions and, in principle, of promoting better packing. In order to produce a structure with large pores, it would seem that some measure of control over the potential for C-H interactions is required.
| || Figure 1 |
(a) Temperature dependence of the powder diffraction profiles. (b) The reduced cells for 300 (black) and 500 K (blue in the electronic version of the paper), and for T = 500 K (red), the nearest primitive cell to the reduced cell at 300 K. (c) Temperature dependence of the unit-cell volumes.
| || Figure 2 |
Illustration of the analytical procedure involving the MEM-assisted structure refinement. (Top) The difference MEM charge densities with the structural model are shown as a mesh (red in the electronic version of the paper). The fitting results of the Rietveld refinement are also shown. (Middle) In order to allow small deformations of the molecule during the rigid-body refinement, additional rotation axes were introduced to express the slight deformations while keeping reasonable interatomic distances. (Bottom) The difference MEM charge densities with a revised model are shown as a mesh (red). The fitting results of the Rietveld refinement are shown.
| || Figure 3 |
The molecular structures of (I) at (a) 90 K, (b) 300 K and (c) 500 K, showing the atom-numbering scheme. Displacement spheres are drawn at the 50% probability level.
| || Figure 4 |
Superposition of the structures of (I) at 90 K (blue in the electronic version of the paper), 300 K (green) and 500 K (red). Cg1-Cg9 are the centres of the carbon rings.
| || Figure 5 |
The crystal packing of (I) at (a) 90 K, (b) 300 K and (c) 500 K. - and C-H contacts are shown as dashed lines.
| || Figure 6 |
C31-H68Cg4iv contacts (dashed lines) at (a) 300 K and (b) 500 K. [Symmetry code: (iv) -x, -y + 2, -z + 1.]
| || Figure 7 |
(a) C14-H59Cg8i and (b) C52-H69Cg9v contacts (dashed lines) at 300 K. [Symmetry codes: (i) x, y - 1, z; (v) -x + 1, -y + 2, -z + 1.]
| || Figure 8 |
A view of the highest occupied molecular orbital (HOMO) for (I) at 500 K.
| || Figure 9 |
Results of the final Rietveld refinements at (a) 90 K, (b) 300 K and (c) 500 K. The experimental profiles are indicated by crosses (red in the electronic version of the paper). The calculated profiles are solid lines (blue). The difference profile is shown at the bottom of each plot as a solid line (green). The vertical black bars correspond to the positions of the Bragg peaks.
Under a nitrogen atmosphere, (4-ferrocenylphenyl)acetylene (908 mg, 3.2 mmol), 1,4-dibromoanthraquinone (605 mg, 1.6 mmol), CuI (62 mg) and Pd(PPh3)2Cl2 (119 mg) were suspended in Et3N (100 ml). The suspension was refluxed for 4 h, which caused a colour change from red to brown. After removal of Et3N under reduced pressure, the residue was dissolved in CHCl3 and washed with water. The organic layer was dried over Na2SO4 and evaporated, and the residue was purified by alumina column chromatography (activity: II-III) with a mixture of hexane and chloroform (2:1-1:2 v/v) as eluent. A brown fraction was collected and evaporated to produce (I) as a brown solid (yield: 500 mg, 39%). Brown crystals of (I) suitable for X-ray powder structure analysis were obtained by recrystallization from dichloromethane-hexane (1:1 v/v).
The powder was placed in a 0.4 mm glass capillary. The X-ray powder diffraction data were measured using a large Debye-Scherrer camera with an imaging-plate (IP) detector installed at SPring-8 on beamline BL02B2 (Nishibori et al., 2001). A CeO2 standard powder sample (NIST SRM674a) was used for wavelength calibration. In the temperature-dependence measurements, the calibrated wavelength was 0.70172 (1) Å and powder data were measured at temperatures of 300, 330, 360, 380, 400, 420, 440, 460, 480 and 500 K, and again at 300 K. In the measurements for structure determination, the calibrated wavelength was 1.00123 (1) Å and powder data were measured at 90, 300 and 500 K. Indexing was carried out using the program DICVOL04 (Boultif & Louër, 2004). The first 27 peaks of the powder pattern were indexed completely on the basis of a triclinic cell. The figures of merit, F(27), for the 90, 300 and 500 K data are 39.7, 17.9 and 15.9, respectively. The candidate space groups were P1 and P. There were three unindexed peaks in the 90 K data and there were some single-crystal spots in the two-dimensional diffraction data. These spots correspond to the unindexed peaks, which can be indexed as the 111, 220 and 311 reflections of ice Ic. These peaks were expressed using Pearson VII profile functions in the subsequent analysis.
The lattice constants derived in the temperature-dependence study were slightly different from those obtained from the data for structure determination for those temperatures that were common to the two sets of measurements. The deviations may be due to the short exposure times of the temperature-dependence measurements.
The temperature dependence of the lattice constants was determined with the Le Bail method using the program SP (Nishibori et al., 2007). The structure was initially solved by direct-space methods with a genetic algorithm (GA) using the program Crystal Profiler (Nishibori, Ogura et al., 2008). The MEM charge density was calculated using the program ENIGMA (Tanaka et al., 2002). Restrained Rietveld analysis was employed for the final refinement. In the analysis, chemically equivalent distances were restrained to be equal. The restrained C-H distances for the Aq and Fc fragments were 0.93 (1) Å, and the C-H distances for the Fc fragment were 0.90 (1) Å. A single displacement parameter was refined for all atoms of each element present, including hydrogen. The maximum deviations of the bond distances and angles from the rigid-body model were less than 0.03 Å and 0.2°, respectively, in the final refinement. The split-type pseudo-Voigt profile function (Toraya, 1990) was used with strain broadening (Stephens, 1999). The results of the Rietveld refinements are shown in Fig. 9. Density functional theory quantum-chemical calculations were performed using the program GAUSSIAN09 (Frisch et al., 2009).
For all compounds, data collection: local software (Nishibori et al., 2001); cell refinement: SP (Nishibori et al., 2007); data reduction: local software (Nishibori et al., 2001); program(s) used to solve structure: GAIA (Nishibori, Ogura et al., 2008); program(s) used to refine structure: SP; molecular graphics: PyMOL (DeLano, 2002) and Mercury (Macrae et al., 2008); software used to prepare material for publication: publCIF (Westrip, 2010).
Supplementary data for this paper are available from the IUCr electronic archives (Reference: FA3318 ). Services for accessing these data are described at the back of the journal.
This work was supported by a Grant-in-Aid for Young Scientists (A) (No. 17686003) and Scientific Research (B) (No. 20360006) of MEXT, Japan. The authors thank Mr Masanori Yoshida for experimental and analytical assistance. We also thank Drs K. Kato and J. E. Kim for experimental help. The synchrotron radiation experiments were performed at SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI).
Alberola, A., Coronado, E., Galán-Mascarós, J. R., Giménez-Saiz, C. & Gómez-García, C. J. (2003). J. Am. Chem. Soc. 125, 10774-10775.
Boultif, A. & Louër, D. (2004). J. Appl. Cryst. 37, 724-731.
DeLano, W. L. (2002). The PyMOL Molecular Graphics System. http://www.pymol.org .
Ferraris, J., Cowan, D. O., Walatka, V. Jr & Perlstein, J. H. (2003). J. Am. Chem. Soc. 95, 948-949.
Frisch, M. J. et al. (2009). GAUSSIAN09. Gaussian Inc., Wallingford, Connecticut, USA.
Harris, K. D. M., Johnston, R. L. & Kariuki, B. M. (1998). Acta Cryst. A54, 632-645.
Kondo, M., Murata, M., Nishihara, H., Nishibori, E., Aoyagi, S., Yoshida, M., Kinoshita, Y. & Sakata, M. (2006). Angew. Chem. Int. Ed. 45, 5461-5464.
Macrae, C. F., Bruno, I. J., Chisholm, J. A., Edgington, P. R., McCabe, P., Pidcock, E., Rodriguez-Monge, L., Taylor, R., van de Streek, J. & Wood, P. A. (2008). J. Appl. Cryst. 41, 466-470.
Murata, M., Yamada, M., Fujita, T., Kojima, K., Kurihara, M., Kubo, K., Kobayashi, Y. & Nishihara, H. (2001). J. Am. Chem. Soc., 123, 12903-12904.
Nishibori, E., Nakamura, T., Arimoto, M., Aoyagi, S., Ago, H., Miyano, M., Ebisuzaki, T. & Sakata, M. (2008). Acta Cryst. D64, 237-247.
Nishibori, E., Ogura, T., Aoyagi, S. & Sakata, M. (2008). J. Appl. Cryst. 41, 292-301.
Nishibori, E., Sunaoshi, E., Yoshida, A., Aoyagi, S., Kato, K., Takata, M. & Sakata, M. (2007). Acta Cryst. A63, 43-52.
Nishibori, E., Takata, M., Kato, K., Sakata, M., Kubota, Y., Aoyagi, S., Kuroiwa, Y., Yamakata, M. & Ikeda, N. (2001). Nucl. Instrum. Methods Phys. Res. Sect. A, 467, 1045-1048.
Sakata, M. & Sato, M. (1990). Acta Cryst. A46, 263-270.
Stephens, P. W. (1999). J. Appl. Cryst. 32, 281-289.
Tanaka, H., Takata, M., Nishibori, E., Kato, K., Iishi, T. & Sakata, M. (2002). J. Appl. Cryst. 35, 282-286.
Toraya, H. (1990). J. Appl. Cryst. 23, 485-491.
Westrip, S. P. (2010). J. Appl. Cryst. 43, 920-925.