Pyridine-2,6-dicarboxylic acid (pydcH2) as a very important carboxylate derivative has attracted much interest in coordination chemistry and we have utilized it widely in our studies , , , , , , , . In coordination to metals, it can act as a neutral, double ionized or single ionized ligand. Several crystal structures of mononuclear, binuclear or polynuclear Zn(II) complexes with dipicolinic acid (dipicH2) and its derivatives have been reported. Complex anion [Zn(pydc)2]2− is rather frequent in the literature and the corresponding cations can be either organic, inorganic or complex , . According to our knowledge, there is no report in the literature for any incommensurately modulated structure for complexes with pyridine-2,6-dicarboxylic acid. The crystal structure of [Zn(pydcH)2]·3H2O at 288 K has been reported by Okabe and Oya in 2000 . In the present work we concentrate on the structure analysis of the low temperature modulated phase β-[Zn(pydcH)2]·3H2O.
An aqueous solution of pyridine-2,6-dicarboxylic acid (0.167 g, 1 mmol) was added to an aqueous solution of Zn(NO3)2·6H2O (0.298 g, 1 mmol). The reaction mixture was stirred at 25°C for 4 h. Colorless crystals of the compound were obtained after 4 days of evaporation at 4°C.
Differential scanning calorimetry was carried out using Perkin-Elmer DSC7 instrument and Pyris 4.02 control and evaluation software. Melting point and phase transition temperature of cyclohexane were applied for the temperature and heat flow calibration. The sample holder of the calorimeter was cooled by liquid nitrogen and purged by helium at 20 mL/min. Small crystals of [Zn(pydcH)2]·3H2O were removed from mother liquor, dried on filter paper and closed hermetically in aluminium sample pan (sample mass 7.96 mg). The DSC measurement started by heating from 93 K to 298 K followed by cooling to 103 K, at a heating and cooling rate of 10 K/min, and repeated to see the reproducibility.
Crystal structure determination
A single crystal of the β-[Zn(pydcH)2]·3H2O was measured at 120 K on a four-circle diffractometer Gemini of Oxford Diffraction using CuKα radiation from a sealed X-ray tube collimated by mirrors of the Cu-Ultra collimator. Diffraction spots were registered with a CCD detector Atlas S1 and data were processed with the CrysAlis software. The unit cell was selected to be related to the unit cell of the room temperature phase (α[Zn(pydcH)2]·3H2O) in the standard setting, i.e. with the P21/c space group. The parameter a was doubled in order to get rid of the ½ component of the q-vector, which can be more conveniently described by an additional centering vector. The resulting unit cell a=28.2947(11) Å, b=9.711(33) Å, c=13.717(4) Å and β=116.37(3)° is related to the published  room temperature unit cell as a≈2c′, b≈b′, c≈a′, β≈β′.
Satellite reflections of the first and second order could be easily detected from the Ewald explorer of CrysAlis and indexed using the q-vector 0.4051(10)b*. Intensities of both main and satellite spots were evaluated without any preliminary symmetry assumptions except the Laue symmetry, which is needed for proper frame scaling and absorption correction by spherical harmonic functions (ABSPACK of CrysAlis). Systematic absences were then evaluated visually using the reciprocal layers (Figure 1) reconstructed from the CCD frames (UNWARP tool of CrysAlis), and by computer methods using the program Jana2006 . This analysis revealed centring vector (½, 0, 0, ½) and other systematic absences corresponding to the superspace group symbol X21/c(0b0)s0, where X stands for the non-standard centring vector. In the list of Stokes et al. , this is a non-standard setting of the superspace group 126.96.36.199. The interpretation of the rules of systematic absences is manifested in the reconstructed reciprocal space. The condition for (hklm) is h+m=2n, therefore the layer (hk0) shows first order satellites for absent main reflections (Figure 1a),the layer (0kl) shows only second order satellites (Figure 1b) and the layer (1kl) has no main reflections (Figure 1c).
The structure was solved by charge flipping methods incorporated in Superflip  and refined with Jana2006. The solution resulted in initial atomic positions, initial modulation waves and verification of the superspace group X21/c(0b0)s0. From the electron density maps of Superflip, it was apparent the water molecule O11 had a discontinuity in the positional modulation and that the crenel occupation function should be applied to this atom. At this stage of solution, the other atoms were refined with a positional and ADP harmonic modulation waves of second order. The further addition of modulation waves improved the agreement factors, but resulting positional modulations no longer followed the observed electron density, forming the additional maxima in residual electron density maps.
Upon further inspection of residual electron density surrounding atoms O7, O8, O10 and C14 the minima and maxima were suggesting the discontinuity of positional modulation waves. When the crenel occupation function was introduced the agreement factors improved and the residual electron density in atoms vicinity decreased. We assumed this to be the confirmation of the discontinuity in positional modulation. For the sake of model consistency, taking into the account relative rigidity of the pydcH− anion, we decided to extend the crenel occupation modulation function for the whole anion, even though some atoms could be described without discontinuity.
The discontinuities in positional modulation waves of water molecules as well as pydcH− anion were happening approximately in phase with slight discrepancies caused by data accuracy. Therefore, we decided to align the centres of the crenel intervals of all atoms around the discontinuity points the whole system of pydcH− and water molecules. This alignment resulted in insignificant change of agreement factors, confirming that it is in agreement with the refined atomic positions.
For combination of the crenel occupational modulation and modulation of positions, we used xharmonic modulation waves defined at the crenel interval . The reason for this kind of modulation waves was a slightly better convergence of the refinement, although different method like, e.g. Legendre polynomials would provide almost the same structure model. Only the first order positional modulation wave was necessary for the crenel modulated atoms. Fit of modulation functions with corresponding Fourier maxima was estimated for selected atoms in Figure 2.
The hydrogen atoms of water molecule O9 were discernible in residual electron density maps; therefore they are present in the structure model. The positions of hydrogen atoms of water molecules O10 and O11 were unclear; therefore they are absent in the structure model. The length of the vector q 0.4051 (10)b* was close to commensurate value (2/5). The commensurate refinement was tested, but it resulted in significant worsening of agreement factors. The attempts to reduce the number of refined parameters by using the rigid body refinement of the pydcH− anions had the same effect.
The basic information about the crystal, its measurement and refinement is summarized in Table 1.
The DSC curve shown in the Figure 3 reveals two endothermic peaks upon the heating, one at 200 K and the other at 270 K. The cooling curve shows two exothermic peaks at 194 K and 240 K, confirming reversibility of the process with hysteresis of temperature. The relatively sharp DSC peaks at 200 K indicates the first order transition. The hysteresis of the curve at Figure 3 is mostly instrumental one caused by relatively fast heating-cooling rate. Figure 4 shows evolution of number of indexed satellites and cell parameters. This evolution clearly reflects the phase transition at 200/194 K, but it is insensitive to the second much weaker transition at 270/240 K. This weaker effect is probably caused by transition of an impurity present inside the sample bulk and transition of adsorbed water, and cannot be attributed to a structural phase transition.
On cooling, satellites corresponding to the phase transition at 200/194 K appear abruptly, confirming the first order transition, and they also disapper abruptly on heating. On the other hand, number of satellites considerably increases when cooling continues towards lower temperatures and seems to stabilize below 120 K. The most presumable reason for this evolution is increasing of satellite intensity on cooling, which allows the peak hunting procedure to find more satellites.
The evolution of the lattice parameter b seems to be in contradiction with the DCS curve (Figure 3) as well as with evolution of satellites, because it changes gradually. In fact, this graph combines two effects: the abrupt change of b due to the first order phase transition, and gradual shrinking of the unit cell due to cooling. Typical shortening of a moderately sized unit cell parametr on cooling from the room temperature to 100 K is between 0.1 and 0.3 Å; thus the change may be of comparable magnitude to the change due to phase transition. Together with rather large experimental error inherent in fast area detector experiment, the first and second order phase transitions cannot be reliably detected from evolution of unit cell parameters.
Figure 4 shows much smaller hysteresis for the evolution of parameter b because here each cooling-heating step was followed by ~1 h of diffraction experiment.
The phase transition at 200 K is accompanied with considerable change of the diffraction pattern. The satellites are observed up to the second order, which can be described using modulation vector q=0.4051(10)b* in the doubled unit cell compared to the room temperature phase with P21/c symmetry and unit cell parameters a=14.0534(2) Å, b=10.05441(12) Å, c=13.7653(2) Å and β=116.580(2)°. The presence of higher order satellites is compatible with possible discontinuous modulation functions or strong positional modulation.
The asymmetric unit of the average structure of β[Zn(pydcH)2]·3H2O (Figure 5) contains two anions of hydrogen pyridine-2,6-dicarboxylate (pydcH−) and one zinc(II) cation. The zinc(II) cation is bonded to nitrogen of the pyridine group and two ketonic oxygens of the carboxyl group, similarly like in α-[Zn(pydcH)2]·3H2O. One of the pydcH− anions is disordered between two positions with occupancies 0.5852(65) (indicated with green bonds in Figure 5) and 0.4148(42) (red bonds). Five oxygen positions belonging to water molecules can be distinguished in the average structure: fully occupied O9; O10 and O11 with occupancy 0.5852(65); O10′ and O11′ with occupancy 0.4148(42).
The disorder of the pydcH− anion observed in the average structure indicated possible crenel modulation function in the (3+1)D modulated structure. Indeed, disordered positions become separated modulated structure, each variant belonging to different crenel interval, and the length of these intervals corresponded to the previously refined occupancies of disordered atoms. Figure 6 shows schematically this separation using an approximant 1×5×1 viewed along b. The anion (here grey line) without crenel occupation is connected to every Zn cation and changes its position slightly according to its harmonic position modulation; the anions with crenel occupation (here red and green lines) alternate according to their crenel functions, and in addition they also change position because crenel function was combined with position modulation function. This approximant is large enough to see there is more green variants than the red ones because the crenel interval for green molecules is larger than the one for red molecules.
Analysis of hydrogen bonds was complicated by the fact that not all hydrogen atom positions could be determined for the β phase, especially the ones with parent atoms described by crenel functions. In order to make figures clear, we inserted artificial hydrogen atoms labelled H? into expected missing hydrogen positions. These atoms are not included in the final CIF because they do not follow from the structure analysis and their positions might be slightly deviating in the reality.
Both phases have two systems of hydrogen bonds. The system 1 (Figure7a and c) is composed of oxygen atoms O3, O9 and O6. Although O6 is described by crenel modulation function in the β phase, the change of its position associated with the discontinuity point is rather small. Therefore, the system 1 is almost the same for both phases and cannot be considered for discussion about driving forces for modulation.
The system 2, on the other hand, differs for both phases. In the α phase this system (Figure 7b) is formed by a ring of six hydrogen bonds (O10–H1O10···O8, O10–H2O10···O11, O11–H1O11?···O7, and three symmetry equivalents), which is attached to other molecules by two bonds O2–H1O2···O10. The most striking feature of system 2 in the α phase is disorder of O11. Taking into the account all possible positions of O11 within this disorder, O11 participates not only in hydrogen bonds towards O7 and O10, as follows from its reference position, but it can also interact with O1 and with the neighbouring O11. However, it cannot participate in these competing interactions at the same time because the geometry does not allow the hydrogen-bonded chain O1···O11···O11i···O1i (i: 2−x, −y, 1−z).
In the β phase the system 2 of hydrogen bonds is formed by oxygen atoms O1, O2, O7, O7′, O8 and O8′ from organic anions, and oxygen atoms O10, O10′, O11 and O11′ from water molecules. The above mentioned competing possibilities occur separately for various t sections owing the fact that O11 position from the α phase splits to independent crenel-separated O11 and O11′ in the β phase.
Details of interactions of O11/O11′ and O10/O10′ atoms can be studied using their distances to neighbouring atoms plotted as a function of the internal coordinate t (Figure 8). The graphs reveal that O10/O10′ creates shorter interactions compared with O11/O11′, however, the most stable interactions have similar distance about 2.65 Å. This stable interaction forms almost continuous line in both plots, although composed of various contributions: O8, O8′, O11, O11′, for O10/O10′, and O7, O7′, O10, O10′ for O11/O11′. The difference between O10/O10′ and O11/O11′ consists in the fact that the stable interaction at 2.65 Å is almost the shortest one for O11/O11′, while O10/O10′ also participates in much shorter interactions, especially O10/O10′···O2. The most changing interaction is O11···O1, which becomes negligible for t>0.5.
This analysis shows that the stable interactions do not include the chain, which was found impossible in the average structure, i.e. O1···O11···O11i···O1i (i: 2−x, −y, 1−z), and predicted to be possible in the modulated structure. This chain is achieved using other interactions, and it is not realized as a simple alternation of the variants O1···O11···O11′ and O11···O11′···O1. Instead, O11′ interacts with O1 but it never interacts with some other O11 or O11′ oxygen, while O11 for t~0.45 interacts with O1 and another O11. For t above 0.5 the interaction O11···O1 vanishes but O11···O11 remains; for t above 0.75 the chain is completely broken and O11 interacts neither with O1, nor with O11. Two of these variants are shown in Figure 9.
The last difference between α and β phase is the presence of π-stacking in the β[Zn(pydc)2]·3H2O structure. Following the colour codes from Figure 5, we can distinguish two types of π-stacking (Figure 10), between two red anions and between two green anions, both of them corresponding to an off-centre parallel stacking between two electron rich aromatic rings , , . For the red anion the average distance C14′–C10′ is equal to 3.39 Å and C12′–C11 to 3.39 Å while for the green anion C14–C11 is 3.32 Å and C12–C12 is 3.30 Å. No π-stacking could be observed for the pydcH- anion without crenel occupation (the grey molecule).
The structure of a new low temperature phase β[Zn(pydcH)2]·3H2O was solved using the formalism for (3+1)D modulated structures. The phase transition caused considerable rearranging of weak interactions, especially the ones concerning the lattice water molecule O11. This atom is strongly disordered in the room temperature phase because of competing possibilities for hydrogen bonds, and these possibilities are realized in the modulated structures for particular crenel intervals. Moreover, the pydcH− anions of the modulated structure participate in πstacking interactions, which do not occur in the room temperature phase. The rearrangement of hydrogen bonding patterns and the appearance of πstacking interactions are the driving force causing modulation of the low temperature phase.
This research was supported by Yazd Branch, Islamic Azad University- project 51053910927005 and the crystallographic part was supported by the project 15-12653S of the Czech Science Foundation using instruments of the ASTRA lab established within the Operation program Prague Competitiveness - project CZ.2.16/3.1.00/24510.
M. Tabatabaee, B.-M. Kukovec, M. Kazeroonizadeh, A unique example of a co-crystal of [Ag(atr)2][Cr(dipic)2] (dipic=dipicolinate; atr=3-amino-1H-1,2,4-triazole) and dinuclear [Cr(H2O)(dipic)(μ-OH)]2, with different coordination environment of Cr(III) ions. Polyhedron 2011, 30, 1114. CrossrefWeb of ScienceGoogle Scholar
M. Tabatabaee, F. Abbasi, B.-M. Kukovec, N. Nasirizadeh, Preparation and structural, spectroscopic, thermal, and electrochemical characterizations of iron(III) compounds containing dipicolinate and 2-aminopyrimidine or acridine. J. Coord. Chem. 2011, 64, 1718. Web of ScienceCrossrefGoogle Scholar
M. Tabatabaee, M. Tahriri, M. Tahriri, Y. Ozawa, B. Neumüller, H. Fujioka, K. Toriumi, Preparation, crystal structures, spectroscopic and thermal analyses of two co-crystals of [M(H2O)6][M(dipic)2] and (atrH)2[M(dipic)2] (M=Zn, Ni, dipic=dipicolinate; atr=3-amino-1H-1,2,4-triazole) with isostructural crystal systems. Polyhedron 2012, 33, 336. Web of ScienceCrossrefGoogle Scholar
M. Tabatabaee, B.-M. Kukovec, V. Razavimahmoudabadi, A rare example of a dinuclear Cobalt(II) complex with dipicolinate and bridging 2-aminopyrazine ligands. Preparation, structural, spectroscopic and thermal characterization. Z. Für Naturforschung B. 2011, 66b, 813. Google Scholar
M. Tabatabaee, S. Rashidi, M. Islaminia, M. Ghassemzadeh, K. Molčanov, B. Neumüller, Two new dinuclear complexes with dipicolinate and bridging 2-aminopyrazine ligands: preparation, structural, spectroscopic, and thermal characterizations. J. Coord. Chem. 2012, 65, 3449. CrossrefWeb of ScienceGoogle Scholar
M. Tabatabaee, H. Aghabozorg, J. Attar Gharamaleki, M. A. Sharif, Acridinium (6-carboxypyridine-2-carboxylato) (pyridine-2,6-dicarboxylato)zincate(II) pentahydrate. Acta Crystallogr. Sect. E Struct. Rep. Online. 2009, 65, m473. CrossrefWeb of ScienceGoogle Scholar
M. Tabatabaee, R. Mohamadinasab, K. Ghaini, H. R. Khavasi, Hydrothermal synthesis and characterization of a binuclear complex and a coordination polymer of copper(II). Bull. Chem. Soc. Ethiop. 2010, 24, 401. Google Scholar
R. Mohammadinasab, M. Tabatabaee, B.-M. Kukovec, H. Aghaie, The cerium(III) coordination polymer with mixed polycarboxylic acids. Preparation of the CeO2 nanoparticles by thermal decomposition of the polymer. Inorg. Chim. Acta. 2013, 405, 368. CrossrefWeb of ScienceGoogle Scholar
M. Tabatabaee, M. Tahriri, M. Tahriri, M. Dušek, K. Fejfarová, Bis(2,6-diaminopyridinium) bis(pyridine-2,6-dicarboxylato)zincate(II) monohydrate. Acta Crystallogr. Sect. E Struct. Rep. Online. 2011, 67, m769. CrossrefWeb of ScienceGoogle Scholar
Oxford Diffraction LTd: CrysAlis CCD. Oxford Diffraction, Abington, England, 2006. Google Scholar
V. Petříček, M. Dušek, L. Palatinus, Crystallographic Computing System JANA2006: General features. Z. Für Krist. - Cryst. Mater. 2014, 229, 345. Google Scholar
H. T. Stokes, B. J. Campbell, S. van Smaalen, Generation of (3+d)-dimensional superspace groups for describing the symmetry of modulated crystalline structures. Acta Crystallogr. A. 2011, 67, 45. CrossrefWeb of ScienceGoogle Scholar
L. Palatinus, G. Chapuis, SUPERFLIP – a computer program for the solution of crystal structures by charge flipping in arbitrary dimensions. J. Appl. Crystallogr. 2007, 40, 786. CrossrefWeb of ScienceGoogle Scholar
V. Petříček, V. Eigner, M. Dušek, A. Čejchan, Discontinuous modulation functions and their application for analysis of modulated structures with the computing system JANA2006. Z. Für Krist. - Cryst. Mater. 2016, 231, 301. Web of ScienceGoogle Scholar
C. A. Hunter, K. R. Lawson, J. Perkins, C. J. Urch, Aromatic interactions. J. Chem. Soc. Perkin Trans. 2001, 2, 651. Google Scholar
The online version of this article offers supplementary material (https://doi.org/10.1515/zkri-2016-2013).
About the article
Published Online: 2017-12-08
Published in Print: 2018-01-26