Received 19 August 2010
Accepted 14 October 2010
Online 21 October 2010
Twinned low-temperature structures of tris(ethylenediamine)zinc(II) sulfate and tris(ethylenediamine)copper(II) sulfate
aBijvoet Center for Biomolecular Research, Crystal and Structural Chemistry, Faculty of Science, Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands
Correspondence e-mail: email@example.com
Tris(ethylenediamine)zinc(II) sulfate, [Zn(C2H8N2)3]SO4, (I), undergoes a reversible solid-solid phase transition during cooling, accompanied by a lowering of the symmetry from high-trigonal P1c to low-trigonal P and by merohedral twinning. The molecular symmetries of the cation and anion change from 32 (D3) to 3 (C3). This lower symmetry allows an ordered sulfate anion and generates in the complex cation two independent N atoms with significantly different geometries. The twinning is the same as in the corresponding Ni complex [Jameson et al. (1982). Acta Cryst. B38, 3016-3020]. The low-temperature phase of tris(ethylenediamine)copper(II) sulfate, [Cu(C2H8N2)3]SO4, (II), has only triclinic symmetry and the unit-cell volume is doubled with respect to the room-temperature structure in P1c. (II) was refined as a nonmerohedral twin with five twin domains. The asymmetric unit contains two independent formula units, and all cations and anions are located on general positions with 1 (C1) symmetry. Both molecules of the Cu complex are in elongated octahedral geometries because of the Jahn-Teller effect. This is in contrast to an earlier publication, which describes the complex as a compressed octahedron [Bertini et al. (1979). J. Chem. Soc. Dalton Trans. pp. 1409-1414].
We report here the twinned crystal structures of tris(ethylenediamine)zinc(II) sulfate, (I), and tris(ethylenediamine)copper(II) sulfate, (II), both measured at 110 (2) K. At room temperature, the corresponding complexes of vanadium (Daniels et al., 1995), cobalt (Yotnoi et al., 2010), nickel (Mazhar-ul-Haque et al., 1970), copper (Cullen & Lingafelter, 1970) and zinc (Neverov et al., 1990) are isostructural in the trigonal space group P1c and the metal complex is a racemic mixture of the and isomers. There has also been a report of an isostructural manganese complex (Lu, 2009), but because of the positive residual electron density on the site of the metal and the unusually short metal-nitrogen distance we regard this structure with due care. In P1c, the above-mentioned compounds have both the cationic metal complex and the sulfate anion on sites of 32 (D3) symmetry. The tetrahedral sulfate cannot fulfil this symmetry and must therefore be disordered. Different disorder models have been used in the literature. The octahedral metal complexes can, in principle, have 32 (D3) symmetry, but in the case of Cu2+ with a d9 electron configuration one would expect different Cu-N distances because of the Jahn-Teller theorem (Procter et al., 1968; Bertini et al., 1977, 1979). Cullen & Lingafelter (1970) describe the presence of dynamic Jahn-Teller effects in the room-temperature structure of the Cu complex with some evidence in the displacement parameters of the N atoms. It is not uncommon to see average structures of Jahn-Teller-distorted complexes in X-ray crystal structure determinations because of the long time scale of the experiment (Persson et al., 2002).
At 110 (2) K, the structure refinement of (I) based on the literature coordinates in space group P1c failed. The R1 value remains above 25% and anisotropic refinement leads to nonpositive definite displacement parameters. In the list of the most disagreeable reflections, the values of Fo2 are significantly larger than Fc2, which can be considered as a warning sign for twinning (Herbst-Irmer & Sheldrick, 1998). Because splitting of reflections was not detected in the diffraction images, only merohedral twinning is possible. A test with the TWINROTMAT routine of the PLATON software (Spek, 2009) did not find suitable twin operations. Therefore, hexagonal twinning of the P1c symmetric structure can be excluded. A closer inspection of the reflections showed that the extinction conditions related to the c-glide plane of P1c are violated. This prompted us to lower the symmetry to space group P, which is a maximal isomorphic subgroup. The Rint value improves from 0.101 to 0.028 and the R1 value improves from 0.265 to 0.183 (isotropic refinement). In the low-trigonal P, the twofold rotations perpendicular to the c axis are potential twin operations. A new test with TWINROTMAT now indeed finds the correct twin relation as a twofold rotation about hkl = (100). Its symmetry-equivalent operations can be obtained by a coset decomposition (Flack, 1987) of point group 6/mmm with respect to and they are of course also suitable (Table 1). If the correct twin operation is included in the refinement (Herbst-Irmer & Sheldrick, 1998), R1 improves to 0.015 and the anisotropic refinement is stable. The twin fraction refined to 0.3745 (9).
The twinned low-temperature structure of the zinc complex, (I), at 110 (2) K corresponds to the low-temperature structure of the nickel compound in space group P (Jameson et al., 1982), where twinning was also detected. The phase transition of the nickel complex occurs at 180 (1) K (Jameson et al., 1982), while in the zinc complex we still observe P1c symmetry at 140 (2) K during cooling. The phase transition in (I) is reversible and P1c symmetry is regained at 140 (2) K after warming up. In the low-temperature P structure, both the cationic complex and the sulfate anion are located on sites with 3 (C3) symmetry, which can be fulfilled by the octahedral cation and the tetrahedral anion without disorder (Fig. 1). Because of this symmetry, the zinc complex, (I), has two independent N and C atoms, and the r.m.s. deviation from 32 (D3) symmetry is 0.0619 Å, as calculated with MOLSYM (Pilati & Forni, 1998). Similar to the room-temperature structure, the N-Zn-N angle in the same ethylenediamine ligand (`bite angle') is smaller than 90° and the trans angles are smaller than 180°. The Zn1-N2 distance is 0.0311 (15) Å shorter than the Zn1-N1 distance and the C2-N2-Zn1 angle is 1.59 (10)° larger than the corresponding C1-N1-Zn1 angle (Table 2). As Zn2+ has a d10 electron configuration, electronic reasons can be excluded for this symmetry breaking. It might be a consequence of different intermolecular hydrogen-bonding geometries: the hydrogen bonds involving N2 are more linear than those involving N1 (Table 3). While the five-membered chelate ring in the room-temperature phase is in an exact C2 symmetric twist conformation, the ring puckering in (I) is a linear combination of 76.5% of the C2 symmetric twist and 23.5% of the Cs symmetric envelope conformation (Evans & Boeyens, 1989). The C2 axis in the room-temperature structure passes through the mid-point of the C-C bond. In (I) at 110 K this mid-point is shifted significantly away from this axis.
The largest change between the room-temperature and the low-temperature phase is in the sulfate anion, which is heavily disordered in P1c and well ordered in P. In P, one S-O bond of the sulfate tetrahedron is on a threefold axis, while the Zn-N bonds of the Zn(en)3 octahedron (en is ethylenediamine) form angles of 55.28 (3) and 121.00 (3)° with the threefold axis. The ordering of the sulfate now also allows a reliable description of the hydrogen-bonding interactions (Table 3), which was difficult in the room-temperature phase. All four NH hydrogens act as hydrogen-bond donors with sulfate O atoms as acceptors. Each of the two independent O atoms is an acceptor of three hydrogen bonds. Overall, this is an infinite three-dimensional hydrogen-bonding network.
The solid-solid phase transition temperature for the copper complex, (II), is given in the literature as 180 K (Bertini et al., 1977). A low-temperature crystal structure of the copper complex, (II), has been reported (Bertini et al., 1979) based on X-ray data from a diffractometer with a point detector. The authors found a single crystal with unit-cell parameters similar to the trigonal room-temperature cell, but the space-group symmetry was determined as triclinic P. The metal complex and the sulfate anion are on general positions (C1 symmetry) and three sulfate O atoms were refined with a disorder model. The coordination octahedron appeared to be compressed. This is in contrast to a [Cu(en)3]Cl2 complex described in the same paper, which has the expected elongated octahedra. The sulfate and the chloride complexes have similar g values in low-temperature ESR (electron spin resonance) spectra, which is an indication of a similar geometry of the complexes and contradicts the compression of the octahedron. The presence of three signals in single-crystal ESR spectra would require the presence of three triclinic unit cells, which were not detected in the X-ray diffraction experiment. A contradiction with the compressed geometry of the Cu complex was later also found with XAFS (X-ray absorption fine structure) (Villain et al., 1997), where equatorial distances of 2.06 and 2.04 Å were determined at room temperature and 10 K, respectively, and axial distances of 2.28 and 2.34 Å.
In our redetermination of (II) at 110 (2) K, the indexing program DIRAX (Duisenberg, 1992) finds a C-centred unit cell with a = 17.6371 (9) Å, b = 30.7843 (12) Å, c = 12.9731 (6) Å, = 89.976 (2)°, = 132.408 (2)° and = 90.076 (2)° (based on 805 reflections between = 4.12° and = 29.87°). This cell is a very rough approximation because it ignores the splitting of the reflections. Additionally this cell predicts too many reflections in some reciprocal planes and too few in others (Fig. 2).
To cover all reflections and to take reflection splitting into account, a triclinic subcell must be chosen together with suitable twin operations. The corresponding cell parameters are a = 8.875 Å, b = 12.972 Å, c = 13.066 Å, = 71.96°, = 70.15° and = 70.46°. The volume of this cell is twice the trigonal room-temperature cell. For a better comparison with the room-temperature cell we decided to transform the triclinic P cell into a nonstandard triclinic A cell (Table 4). Based on this A-centred lattice, 120° rotations about the c axis are selected as twin operations resulting in three twin domains (`drilling'). This way all reflections are indexed and because of small deviations from exact hexagonal angles the splitting of the reflections can also be explained. Unfortunately, the refinement of the drilling still gives unsatisfactory results. The addition of twofold rotations perpendicular to the c axis as twin operations results in six twin domains. The inclusion of the six domains into the refinement improves results, but the twin fraction of one domain remains zero. Therefore the final intensity integration with EVAL14 (Duisenberg et al., 2003) and the final SHELXL (Sheldrick, 2008b) refinements are based on five twin domains. The corresponding orientation matrices are given in the CIF file and the refined batch scale factors were 0.402 (3), 0.0191 (7), 0.3395 (18), 0.1480 (18) and 0.0913 (18). A coset decomposition of 6/mmm with respect to results in 12 cosets and thus 12 potential twin operations. Only five of these are necessary for a satisfactory refinement.
The asymmetric unit of (II) consists of two independent formula units, and all cations and anions are on general positions with C1 symmetry (Fig. 3). The two independent cations are related by an approximate threefold rotation about the c axis and their geometries are very similar. A quaternion fit (Mackay, 1984) based on all non-H atoms is shown in Fig. 4. Each cation has four short equatorial [2.045 (3)-2.073 (3) Å] and two long axial Cu-N distances [2.373 (3) and 2.396 (4) Å] (Table 5). The equatorial distances are very similar to those in polymeric [Cu(SO4)(C2H8N2)2]n (Lutz et al., 2010), which has Cu-N distances of 2.0173 (8) and 2.0226 (8) Å. The Jahn-Teller effect, which at room temperature can only be recognized in enlarged anisotropic displacement parameters (Cullen & Lingafelter, 1970), is clearly visible in the different Cu-N bond lengths at 110 (2) K. Here, none of the Cu-N bonds in (II) fails the Hirshfeld rigid bond test (Hirshfeld, 1976) with more than 5.
Despite the absence of symmetry, all four equatorial N atoms still form a plane including the Cu atom [the maximal deviation from the least-squares plane is 0.086 (4) Å]. The N-Cu-N bite angles are smaller than the ideal octahedral 90° [79.64 (12)-83.99 (14)°] and the trans angles consequently deviate from linearity [166.73 (12)-174.13 (12)°]. The six independent five-membered chelate rings are mainly in the twist conformation. If they are considered as a linear combination of the twist and the envelope conformations, the contribution of the envelope conformation ranges from 0.0 to 37.9%.
In contrast to the room-temperature phase, the two independent sulfate anions in (II) at 110 K are ordered, but have a different orientation with respect to the cell axes (Fig. 5). In fact, the crystal structure has pseudo-translational symmetry, which is only broken by the O atoms of the sulfate. The ADDSYM routine of PLATON (Spek, 2009) finds an 83% fit for a halving of the c axis. If the O atoms are left out of the analysis, a 100% fit is found. In the standard P setting of the space group this would correspond to a pseudo-I centring. Despite this pseudosymmetry, no correlation matrix elements for coordinates or displacement parameters are larger than 0.76 in the least-squares refinement. ADDSYM does not indicate the presence of monoclinic or orthorhombic symmetry, while the cell parameters can be transformed to a pseudo-orthorhombic I cell with angles close to 90°. Merging of the calculated structure factors according to monoclinic or orthorhombic symmetries results in Rint values larger than 0.34.
As the sulfate is ordered in (II), a detailed analysis of the hydrogen bonding is now possible (Table 6). While in the room-temperature phase there are only two independent donor H atoms, (II) at 110 K has 24 independent donor H atoms. There is a large variation in the donor-acceptor distances [between 2.900 (5) and 3.552 (5) Å]. The longest distances belong to the bifurcated hydrogen bonds at H2N and H22N. The hydrogen bonds lead to a densely packed three-dimensional network with a packing index of 76.0% (Kitajgorodskij, 1973).
| Figure 1 |
The molecular structure of (I), showing the atom-numbering scheme. Displacement ellipsoids are drawn at the 50% probability level and H atoms are drawn with arbitrary radii. [Symmetry codes: (i) 1 - y, 1 + x - y, z; (ii) -x + y, 1 - x, z; (iii) 1 - y, x - y, z; (iv) 1 - x + y, 1 - x, z.]
| Figure 2 |
Schematic representations of the hk0 (top) and hk1 planes (bottom) based on the A-centred supercell with V = 5200.7 (3) Å3 (Table 4). The indexing with the supercell (left) predicts too many reflections in hk0 and too few in hk1. Twinned indexing with three subcells related by 120° rotations about c (right) provides the correct predictions in both planes. Splitting of reflections is ignored in the drawings.
| Figure 3 |
The asymmetric unit of (II), showing the atom-numbering scheme. Displacement ellipsoids are drawn at the 50% probability level and H atoms have been omitted for clarity.
| Figure 4 |
Quaternion fit (Mackay, 1984) of the two independent cations in (II) based on all non-H atoms. The r.m.s. deviation of the fit is 0.041 Å.
| Figure 5 |
Orientations of the ordered sulfate anions in the unit cell of (II). In the electronic version of the paper, the second residue is drawn in red. The Cu(en)3 cations are located at the same position in the a, b direction, but the height in the c direction differs by c/2.
Compounds (I) and (II) were prepared by slowly adding an excess of ethylenediamine at room temperature to a solution of the corresponding metal salt in water. Single crystals were obtained by evaporating the solutions at room temperature.
Mr = 341.74
a = 8.89410 (11) Å
c = 9.68482 (13) Å
V = 663.48 (2) Å3
Z = 2
Mo K radiation
= 2.03 mm-1
T = 110 K
0.30 × 0.09 × 0.09 mm
Nonius KappaCCD diffractometer
Absorption correction: analytical (SADABS; Sheldrick, 2008a) Tmin = 0.612, Tmax = 0.872
11327 measured reflections
1355 independent reflections
1338 reflections with I > 2(I)
Rint = 0.028
|Zn1-N1 ||2.2035 (11) |
|Zn1-N2 ||2.1724 (10) |
|N1-C1 ||1.4718 (15) |
|N2-C2 ||1.4732 (16) |
|C1-C2 ||1.5246 (17) |
|N1-Zn1-N2 ||80.66 (4) |
|N1-Zn1-N2i ||170.58 (4) |
|C1-N1-Zn1 ||106.76 (7) |
|C2-N2-Zn1 ||108.35 (7) |
|Symmetry code: (i) -y+1, x-y, z. |
|D-HA ||D-H ||HA ||DA ||D-HA |
|N1-H1NO1 ||0.869 (17) ||2.308 (17) ||3.0593 (14) ||144.8 (14) |
|N1-H2NO2ii ||0.846 (17) ||2.220 (17) ||2.9711 (16) ||148.0 (15) |
|N2-H3NO1iii ||0.824 (17) ||2.205 (17) ||3.0257 (13) ||173.7 (19) |
|N2-H4NO1i ||0.847 (16) ||2.242 (17) ||3.0818 (16) ||171.5 (15) |
|Symmetry codes: (i) -y+1, x-y, z; (ii) -x+1, -y+1, -z; (iii) y, -x+y, -z+1. |
|Twin matrix (hkl) ||Determinant ||Axis |
|0 0 / 0 0 / 0 0 ||+1 ||2 about (1 0) |
| 0 0 / 1 1 0 / 0 0 ||+1 ||2 about (0 1 0) |
|1 1 0 / 0 0 / 0 0 ||+1 ||2 about (1 0 0) |
|0 1 0 / 1 0 0 / 0 0 1 || ||m perpendicular to (1 0) |
|1 0 0 / 0 / 0 0 1 || ||m perpendicular to (0 1 0) |
| 0 / 0 1 0 / 0 0 1 || ||m perpendicular to (1 0 0) |
| ||Room temperature ||Approximate supercell at 110 K ||True cell at 110 K |
|Nonstandard setting ||Bravais = P ||Bravais = A ||Bravais = A |
|a = 8.966 (1) Å ||a = 17.7495 (6) Å ||a = 8.8749 (7) Å |
|b = 8.966 (1) Å ||b = 17.6371 (9) Å ||b = 17.6358 (11) Å |
|c = 9.597 (1) Å ||c = 19.1581 (10) Å ||c = 19.1559 (11) Å |
| = 90° || = 89.585 (4)° || = 89.582 (3)° |
| = 90° || = 90.239 (3)° || = 90.296 (5)° |
| = 120° || = 119.687 (2)° || = 119.845 (4)° |
|V = 668.13 (13) Å3 ||V = 5200.7 (3) Å3 ||V = 2600.5 (3) Å3 |
|Standard setting ||Bravais = P ||Bravais = C ||Bravais = P |
|a = 8.966 (1) Å ||a = 17.6371 (9) Å ||a = 8.8749 (7) Å |
|b = 8.966 (1) Å ||b = 30.7843 (12) Å ||b = 12.9715 (6) Å |
|c = 9.597 (1) Å ||c = 12.9731 (6) Å ||c = 13.0660 (9) Å |
| = 90° || = 89.976 (2)° || = 71.959 (5)° |
| = 90° || = 132.408 (2)° || = 70.148 (5)° |
| = 120° || = 90.076 (2)° || = 70.459 (5)° |
|V = 668.13 (13) Å3 ||V = 5200.7 (3) Å3 ||V = 1300.25 (16) Å3 |
|Reduced setting ||Bravais = P ||Bravais = P ||Bravais = P |
|a = 8.966 (1) Å ||a = 12.9731 (6) Å ||a = 8.8749 (7) Å |
|b = 8.966 (1) Å ||b = 13.0671 (8) Å ||b = 12.9715 (6) Å |
|c = 9.597 (1) Å ||c = 17.7292 (6) Å ||c = 13.0660 (9) Å |
| = 90° || = 109.702 (4)° || = 71.959 (5)° |
| = 90° || = 109.578 (2)° || = 70.148 (5)° |
| = 120° || = 94.735 (4)° || = 70.459 (5)° |
|V = 668.13 (13) Å3 ||V = 2600.4 (2) Å3 ||V = 1300.25 (16) Å3 |
Compound (I) was measured with a detector distance of 40 mm, a scan angle of 1° and an exposure time of 15 seconds per frame; 652 frames were collected. The intensity data of (I) were obtained using a single orientation matrix. For the refinement, a TWIN and a BASF instruction (SHELXL97; Sheldrick, 2008b) were included. All H atoms were located in difference Fourier maps and refined freely with isotropic displacement parameters.
Compound (II) was measured with a detector distance of 50 mm, a scan angle of 0.5° and an exposure time of 20 seconds per frame; 1615 frames were collected. The intensity data of (II) were obtained using five orientation matrices. Only the single non-overlapping reflections of the major domain and the overlapping reflections with the major domain were taken into account. These data were merged prior to the refinement using the TWINABS software (Sheldrick, 2008a), resulting in an HKLF-5 type of reflection file (Herbst-Irmer & Sheldrick, 1998). H atoms were included in calculated positions and refined using a riding model, with C-H = 0.99 Å and N-H = 0.92 Å, and with Uiso(H) = 1.2Uiso(C,N).
For both compounds, data collection: COLLECT (Nonius, 1999); cell refinement: PEAKREF (Schreurs, 2005). Data reduction: EVAL15 (Schreurs et al., 2010) and SADABS (Sheldrick, 2008a) for (I); EVAL14 (Duisenberg et al., 2003) and TWINABS (Sheldrick, 2008a) for (II). For both compounds, program(s) used to solve structure: SHELXS97 (Sheldrick, 2008b); program(s) used to refine structure: SHELXL97 (Sheldrick, 2008b); molecular graphics: PLATON (Spek, 2009); software used to prepare material for publication: manual editing of the SHELXL CIF file.
Supplementary data for this paper are available from the IUCr electronic archives (Reference: GG3243 ). Services for accessing these data are described at the back of the journal.
Loes Kroon-Batenburg and Toine Schreurs are thanked for fruitful discussions about the indexing of (II).
Aroyo, M. I., Perez-Mato, J. M., Capillas, C., Kroumova, E., Ivantchev, S., Madariaga, G., Kirov, A. & Wondratschek, H. (2006). Z. Kristallogr. 221, 15-27.
Bertini, I., Dapporto, P., Gatteschi, D. & Scozzafava, A. (1979). J. Chem. Soc. Dalton Trans. pp. 1409-1414.
Bertini, I., Gatteschi, D. & Scozzafava, A. (1977). Inorg. Chem. 16, 1973-1976.
Cullen, D. L. & Lingafelter, E. C. (1970). Inorg. Chem. 9, 1858-1864.
Daniels, L. M., Murillo, C. A. & Rodriguez, K. G. (1995). Inorg. Chim. Acta, 229, 27-32.
Duisenberg, A. J. M. (1992). J. Appl. Cryst. 25, 92-96.
Duisenberg, A. J. M., Kroon-Batenburg, L. M. J. & Schreurs, A. M. M. (2003). J. Appl. Cryst. 36, 220-229.
Evans, D. G. & Boeyens, J. C. A. (1989). Acta Cryst. B45, 581-590.
Flack, H. D. (1987). Acta Cryst. A43, 564-568.
Herbst-Irmer, R. & Sheldrick, G. M. (1998). Acta Cryst. B54, 443-449.
Hirshfeld, F. L. (1976). Acta Cryst. A32, 239-244.
Jameson, G. B., Schneider, R., Dubler, E. & Oswald, H. R. (1982). Acta Cryst. B38, 3016-3020.
Kitajgorodskij, A. I. (1973). In Molecular Crystals and Molecules. New York: Academic Press.
Lu, J. (2009). Acta Cryst. E65, m1187.
Lutz, M., Smeets, S. & Parois, P. (2010). Acta Cryst. E66, m671-m672.
Mackay, A. L. (1984). Acta Cryst. A40, 165-166.
Mazhar-ul-Haque, Caughlan, C. N. & Emerson, K. (1970). Inorg. Chem. 9, 2421-2424.
Neverov, V. A., Mazus, M. D., Biyushkin, V. N. & Malinovskii, T. I. (1990). Kristallografiya, 35, 1281-1283.
Nonius (1999). COLLECT. Nonius BV, Delft, The Netherlands.
Persson, I., Persson, P., Sandström, M. & Ullström, A.-S. (2002). J. Chem. Soc. Dalton Trans. pp. 1256-1265.
Pilati, T. & Forni, A. (1998). J. Appl. Cryst. 31, 503-504.
Procter, I. M., Hathaway, B. J. & Nicholls, P. (1968). J. Chem. Soc. A, pp. 1678-1684.
Schreurs, A. M. M. (2005). PEAKREF. Utrecht University, The Netherlands.
Schreurs, A. M. M., Xian, X. & Kroon-Batenburg, L. M. J. (2010). J. Appl. Cryst. 43, 70-82.
Sheldrick, G. M. (2008a). SADABS (Version 2008/1) and TWINABS (Version 2008/4). University of Göttingen, Germany.
Sheldrick, G. M. (2008b). Acta Cryst. A64, 112-122.
Spek, A. L. (2009). Acta Cryst. D65, 148-155.
Villain, F., Verdaguer, M. & Dromzee, Y. (1997). J. Phys. IV Fr. 7, C2-659-C2-660.
Yotnoi, B., Seeharaj, A., Chimupala, Y. & Rujiwatra, A. (2010). Acta Cryst. E66, m628.