Role of Ethynyl-Derived Weak Hydrogen-Bond Interactions in the Supramolecular Structures of 1D, 2D, and 3D Coordination Polymers Containing 5‐Ethynyl-1,3-benzenedicarboxylate

The influence of weak hydrogen bonds on the crystal packing of a series of heavy and transition metal coordination polymers synthesized using the ligand 5-ethynyl-1,3-benzenedicarboxylic acid (H2ebdc) has been evaluated. Five coordination polymers were prepared and crystallographically characterized. These comprise two 1D chains, [Pb(ebdc)(DMSO)2] (1) and [Pb(ebdc)(DMF)] (2), two 2D nets, [Cu3(ebdc)(H2O)1.5(MeOH)0.5]·6H2O (3) and [Pb2(ebdc)2(DMF)4]·H2O (4), and a single 3D framework, [HNEt3][Zn3(μ3-OH)(μ2-H2O)(ebdc)3(MeOH)0.67(H2O)0.33]·MeOH· 1.33H2O (5). The crystal structure of the free acid ligand form, H2ebdc·H2O, is also reported. Within the lead(II) coordination structures, ethynyl-derived C− H···O interactions are consistently found to provide the dominant influence over the crystal packing, as determined by solid-state structural analysis in combination with vibrational spectroscopy. The influence of weak hydrogenbonding effects on the crystal packing of the transition metal coordination polymers that contain lattice water and methanol molecules was found to be far less prominent, which is interpreted in terms of the greater prevalence of strong hydrogen-bond donors and acceptors forming O−H···O interactions within these crystalline lattices. ■ INTRODUCTION The design of crystalline architectures employing an understanding of the intermolecular interactions available to molecular subunits is the basis for crystal engineering, a key field straddling solid-state molecular assembly and structural analysis, whose importance lies in its capability for allowing control of crystal properties and functions as well as its structure. Much progress has been made using simple systems composed of covalent or strong hydrogen bonds; however, control over weak hydrogen-bonding interactions remains one of the most challenging areas of crystal engineering. A weak hydrogen bond can broadly be defined as an electrostatic interaction formed by a hydrogen atom between two structural moieties of moderate to low electronegativity, of which C−H···X (X = O/N) interactions are a key example. The study of these interactions began with the discovery of increased polarization of haloforms with ketones, pyridines, and ethers. This observation was linked by spectroscopic interpretations to hydrogen bonding, to account for large bathochromic shifts observed in the infrared (IR) spectra of such species. While assignment of these interactions by crystallography were initially resisted in the wake of strong criticisms by Donohue toward a study by Sutor, in recent years, advancements in the field of crystallography and in computational power have demonstrated and confirmed the fundamental importance of these interactions to supramolecular self-assembly. Such interactions can vary in strength from near equivalence to van der Waals interactions to being stronger than weak covalent bonds and rivalling traditional hydrogen bonds. In many crystalline networks, weak hydrogen bonds can be thought of as having a steering effect that may preferentially favor a single solid form, for example, a particular polymorph with solid-state packing governed largely by stronger intermolecular interactions but which is supplemented by contributions from weak directional C−H···X interactions. However, there are known instances wherein weak hydrogen bonds are formed preferentially over interactions involving available strong donors or acceptors. This preference for weak hydrogen bonds ensures that structural predetermination will fail if predictive methods rely solely on the traditional hierarchy of bonding strengths. Thus, the crystal engineer must view the sum of all interactions, both weak and strong, in order to predict the intermolecular architecture and crystal packing of a material, as opposed to focusing only on the strongest few interactions. This requirement necessitates an indepth study of weak hydrogen-bonding interactions in order to understand fully, and integrate better, these stabilizing forces into the wider tenets of crystal engineering. The ethynyl group is an attractive functionality for the study of weak hydrogen bonding owing to its highly activated nature Received: October 15, 2014 Revised: November 28, 2014 Published: December 2, 2014 Article pubs.acs.org/crystal © 2014 American Chemical Society 465 dx.doi.org/10.1021/cg501535b | Cryst. Growth Des. 2015, 15, 465−474 This is an open access article published under a Creative Commons Attribution (CC-BY) License, which permits unrestricted use, distribution and reproduction in any medium, provided the author and source are cited.


■ INTRODUCTION
The design of crystalline architectures employing an understanding of the intermolecular interactions available to molecular subunits is the basis for crystal engineering, a key field straddling solid-state molecular assembly and structural analysis, whose importance lies in its capability for allowing control of crystal properties and functions as well as its structure. 1,2 Much progress has been made using simple systems composed of covalent 3 or strong hydrogen bonds; 4 however, control over weak hydrogen-bonding interactions remains one of the most challenging areas of crystal engineering. 5 A weak hydrogen bond can broadly be defined as an electrostatic interaction formed by a hydrogen atom between two structural moieties of moderate to low electronegativity, of which C−H···X (X = O/N) interactions are a key example. The study of these interactions began with the discovery of increased polarization of haloforms with ketones, pyridines, and ethers. 6 This observation was linked by spectroscopic interpretations to hydrogen bonding, to account for large bathochromic shifts observed in the infrared (IR) spectra of such species. 7 While assignment of these interactions by crystallography were initially resisted in the wake of strong criticisms by Donohue 8 toward a study by Sutor,9,10 in recent years, advancements in the field of crystallography and in computational power have demonstrated and confirmed the fundamental importance of these interactions to supramolecular self-assembly. Such interactions can vary in strength from near equivalence to van der Waals interactions 11 to being stronger than weak covalent bonds and rivalling traditional hydrogen bonds. 12 In many crystalline networks, weak hydrogen bonds can be thought of as having a steering effect that may preferentially favor a single solid form, for example, a particular polymorph with solid-state packing governed largely by stronger intermolecular interactions but which is supplemented by contributions from weak directional C−H···X interactions. 5 However, there are known instances wherein weak hydrogen bonds are formed preferentially over interactions involving available strong donors or acceptors. 13,14 This preference for weak hydrogen bonds ensures that structural predetermination will fail if predictive methods rely solely on the traditional hierarchy of bonding strengths. Thus, the crystal engineer must view the sum of all interactions, both weak and strong, in order to predict the intermolecular architecture and crystal packing of a material, as opposed to focusing only on the strongest few interactions. This requirement necessitates an indepth study of weak hydrogen-bonding interactions in order to understand fully, and integrate better, these stabilizing forces into the wider tenets of crystal engineering.
The ethynyl group is an attractive functionality for the study of weak hydrogen bonding owing to its highly activated nature and acidity. 15 It is also ideally suited to spectroscopic study in the solid state, owing to its prominent C(sp)−H vibrational band in terms of intensity and unique location. Indeed examples of the C(sp)−H···O hydrogen bond match moderately strong O−H···O hydrogen bonds geometrically and, in terms of trends, spectroscopically. 16 This finding has been supported by studies comparing the hydrogen-bonding potential energy curves of weak acidic hydrogen bonds to O− H···O bonds that showed matching distributions. 5,17 The ethynyl group can act cooperatively as a hydrogen-bond acceptor via the alkyne π system, as well as containing a donor site from the C(sp)−H hydrogen, enhancing hydrogenbond strength by mutual polarization. 5 Similar cooperatively is one factor that provides strength to traditional O−H···O−H··· O−H hydrogen-bonded chains.
Evaluation of the weak C−H···O hydrogen bonding observed for coordination polymers in this study has been achieved by single-crystal diffraction studies in order to obtain accurate interatomic donor−acceptor distances (D) and angular properties (θ) of the hydrogen donor approaching the acceptor atom, in conjunction with analysis of C(sp)−H hydrogen bands by vibrational spectroscopy and changes to thermal vibrations observed in the crystal structure 18 that result from hydrogenbond interactions. The coordination networks studied allow a comparison of networks containing hard acids (copper(II) and zinc(II)) that exhibit well-defined coordination geometries with those of lead(II), a softer metal center that permits a more flexible coordination environment. It is hoped that increasing the coordination flexibility about the metal center will translate into optimal ligand orientations for ethynyl-based weak hydrogen bonding, which will allow the targeted synthesis of materials containing such interactions, aiding in their study.
This work forms part of ongoing programs developing functional metal−ligand systems, including MOFs, 19 coordination polymers, 20 and organometallic clusters. 21−23 ■ EXPERIMENTAL METHODS Caution: Metal perchlorates are potentially explosive! Only a small amount of material should be prepared and handled with great care. Starting materials and solvents were purchased from commercial sources and used without further purification, with the exception of H 2 ebdc, for which single crystals were grown by slow diffusion of water into a concentrated methanol solution. X-ray diffraction data for H 2 ebdc and compounds 1−5 were collected on an Agilent Gemini A-Ultra diffractometer 24 at the University of Bath using Mo Kα radiation, with the crystal being cooled to 150 K by an Agilent Cryojet. 25 Powder X-ray diffraction patterns (PXRDs) were recorded on a Bruker AXS D8 Advance diffractometer with Cu Kα radiation of wavelength 1.5406 Å at 298 K. Samples were placed on a flat plate and measured with a 2θ range of 3−60°. Simulated X-ray powder patterns were generated from single-crystal data that were imported into PowderCell. Infrared spectra were recorded on a PerkinElmer Spectrum 100 spectrometer equipped with an ATR sampling accessory. Abbreviations for IR bands are s, strong; m, medium; w, weak. Elemental analyses (C, H, N) were performed on a CE-440 elemental analyzer (Exeter Analytical).

■ RESULTS AND DISCUSSION
Analysis of H 2 ebdc·H 2 O. The ligand used in this study, 5ethynyl-1,3-benzenedicarboxylic acid (H 2 ebdc), has not been previously crystallographically analyzed. It possesses two carboxylic acid functional groups to allow coordination of metal nodes upon deprotonation and an ethynyl group to promote weak hydrogen-bonding interactions. Nontarget interactions that are allowed by this ligand include π−π stacking derived from the central aromatic ring as well as strong hydrogen bonding from either the carboxylic acid or carboxylate functionality.
Crystalline H 2 ebdc·H 2 O was obtained by layering a concentrated methanol solution over a water layer and was found to adopt the triclinic space group P1̅ (Table 1). The asymmetric unit contains both H 2 ebdc and a single hydrogenbonded molecule of lattice water. A 1D hydrogen-bond network propagates along nodes of four H 2 ebdc molecules and two water molecules, composed of three rings of interactions, two outer rings designated R 3 3 (10), and an inner ring designated R 4 2 (8) according to Etter's notation 26  The ethynyl group is located in the vicinity of O(2′) of a symmetry-generated molecule (−x − 1, 2 − y, 1 − z) of H 2 ebdc ( Figure 2); however, any hydrogen bonding is likely to be extremely weak based on the observed distance D of 3.337(3) Å and approach angle θ of 147°. Instead, the ethynyl group appears to order the structure solely by interdigitation. Consequently, the IR vibration spectrum of the free ethynyl C− H stretching band for H 2 ebdc was assigned as 3301 cm −1 , providing a reference point for any bathochromic shifts that occur with hydrogen-bond formation in the coordination polymers of this ligand reported below. Furthermore, a decrease in the C terminal /C inner thermal parameter ratio for the terminal ethynyl group may serve as a qualitative measure of hydrogen bonding; hence, the nonbonded value for H 2 ebdc (C(10)/C(9)) was determined to be 1.37, which is within the range found for known nonbonded terminal ethynyl groups, albeit near the lower limit. 18 Five Synthetic Overview. The equimolar reaction of H 2 ebdc with hydrated divalent metal salts gives rise to coordination polymers consistent with the known affinity of meta-substituted aromatic carboxylic acids for such species. 27−29 The hierarchy of molecular interactions is ultimately governed by covalent metal−carboxylate interactions involving the deprotonated ebdc ligand, which yield the primary structure. Inclusion of the ligand guarantees the presence of the ethynyl group, ensuring the possibility of weak hydrogen bonds influencing the secondary molecular packing. However, a range of polar protic and aprotic solvents were used during the synthesis of 1−5 (Scheme 1), and no effort was made to exclude water from the synthesis or crystallization processes; hence, competition with strong solvent-mediated hydrogen bonding has been allowed  Packing of the 1D chains results from bifurcated donor C− H···O hydrogen bonding derived from the ethynyl group and links to O(2′) and O(4′) of two carboxylate groups (Figure 4). Distances for C(9)···O(2′) of 3.286(8) Å and C(9)···O(4′) of 3.213(9) Å were observed, and the angle of the ethynyl group is 170°relative to the plane of the carboxylate oxygen atoms, favoring the observed hydrogen-bonding interactions. A significant bathochromic shift of 80 cm −1 was observed in the IR spectrum, and the ratio of thermal vibration for C(9) and C(8) decreased to 1.33, suggesting inhibition as a result of hydrogen bonding. These interactions occur on either side of the 1D chain and collectively yield a 2D hydrogen-bonded network that packs in a herringbone arrangement ( Figure S1, Supporting Information).
Lead 2D Net 2. Lead network 2 was formed by solvothermal synthesis in DMF at 140°C and yields a 1D tape motif that crystallizes in the triclinic space group P1̅ . The asymmetric unit contains three unique lead(II) atoms that are ligated by three ebdc ligands and three coordinated molecules of DMF, two of which are disordered over two positions ( Figure S2, Supporting Information). These propagate by symmetry to give repeating hexagonal rings of lead(II) atoms that are fused by bridging ebdc ligands ( Figure 5). Pairs of ebdc ligands direct ethynyl groups above and below the tape, and a single ligand bridges the remaining orthogonal faces, ensuring that the periphery of the tape is shrouded by weak hydrogen donors ( Figure 6). Bridging modes exhibited by the six carboxylate groups include μ-(μO:κO′), μ 3 -(μO:μO′), and μ 4 -(μ 3 O:μO′), which result in one nine-coordinate and two seven-coordinate lead(II) centers. There are three unique ethynyl groups that have the potential to form hydrogen-bond interactions (Figure 7). Bifurcated ethynyl C−H···O hydrogen     The final ethynyl group is directly oriented toward the oxygen atom of a carboxylate group (C(39)···O(5‴)) with an angle of 170°; however, the distance is long at 3.562(17) Å. No other significant interactions form between the 1D chains, suggesting that weak hydrogen-bonding interactions are solely responsible for the crystal packing in 2. Despite the presence of three distinct ethynyl groups within the asymmetric unit of 2, only a single band with a prominent shoulder directed toward lower wavenumbers was observed in the IR spectrum. The dominant vibration exhibited a bathochromic shift of 56 cm −1 relative to the nonbonded reference and has been assigned to the bifurcated pair of terminal alkynes, owing to its intensity relative to the shoulder band that is likely a result of its proximity to an oxygen acceptor. Furthermore, the shoulder exhibits greater red-shifting, which has been observed for directly aligned, as opposed to bifurcated, C−H···O interactions (vide infra). The ratios of C terminal /C inner thermal vibrations for the bifurcated hydrogen-bond interactions are 1.31 and 1.47, which suggests that the former, C(20)···O(4′), holds more hydrogen-bonding character, likely owing to its more favorable orientation angle of 179°. The final ethynyl group has a very high C terminal /C inner thermal parameter ratio of 1.73; however, this may be a result of its proximity to a neighboring disordered DMF site. In such instances, infrared analysis should be considered the most reliable indicator for weak hydrogen-bond identification. 5 Copper 2D Net 3. The 2D copper network 3 crystallizes in the monoclinic I2/a space group and exhibits formation of the characteristic copper paddlewheel motif. The asymmetric unit contains two copper(II) atoms bridged by a carboxylate group from three ebdc ligands, one of which coordinates the third copper(II) atom via the second carboxylate group. The apical positions of the paddlewheel are occupied by coordinated water molecules, one of which exhibits 50:50 substitutional site disorder with a methanol molecule. Five molecules of water are located in the lattice, four of which are disordered over more than one position. The extended molecular structure forms a t 1 {6,3} 2D net, whereby the paddlewheels link to assemble a Kagomélattice (Figure 8). Each 2D sheet is offset relative to the sheets above and below. In this manner, the network is similar to that formed by copper(II) with other 1,3benzenedicarboxylates (bdc), such as 5-nitro-1,3-bdc and 5-(methylsulfanylmethyl)-1,3-bdc. 30 Of the three unique ebdc ligands, two are oriented toward the carboxylate groups of the layer above, and one is oriented toward a lattice water molecule below (Figure 9). Competing interactions include π−π stacking that directly link the layers and a network of strong hydrogen bonding. The former is a parallel offset interaction with an intercentroid distance of 3.67 Å (closest π−π contact, 3.60 Å), linking this ligand with its symmetry equivalent in a second net. Interaction between the two nets is strengthened by a hydrogen-bonding array between the apical solvent molecules of the paddlewheel and lattice water molecules (Figure 10). While the quality of the crystallographic data was insufficient to locate the individual hydrogen atoms using the electron density map, interoxygen distances give strong evidence for the presence of a strong hydrogen-bonding network, exemplified by distances of 2.687 (5)      close proximity of the ethynyl donor groups to the oxygen acceptors, the angles are not conducive of a significant hydrogen-bonding interaction, ranging from 143 to 152°.
The IR spectrum of 3 shows that the band intensity for the ethynyl C−H stretching vibrations centered at 3287 cm −1 has decreased and broadened considerably, with each effect being characteristic of the absence of significant hydrogen bonding. 31 Greater thermal vibration was observed between the terminal and inner sp carbons of the alkyne, with C terminal /C inner ratios of 1.66, 1.74, and 1.40, which provides further evidence for a lack of hydrogen-bonding interactions. In this instance, the crystal packing appears to be determined by interactions other than those of weak hydrogen bonding (i.e., strong hydrogen bonding and/or π−π stacking described earlier).
Lead 2D Net 4. Crystals of 2D lead network [Pb 2 (ebdc) 2 (DMF) 4 ]·H 2 O (4) were obtained by heating a mixture of H 2 ebdc and hydrated lead acetate in DMF at 100°C . The network crystallizes in the tetragonal space group P4̅ 2 1 c, and the asymmetric unit contains a single lead(II) atom ligated via the carboxylate of a single ebdc ligand, two DMF molecules, of which one is disordered over a special position, and a single lattice water molecule with half occupancy that is also located on a special position ( Figure S3, Supporting Information). The lead(II) atoms are eight-coordinate and aggregate to form a Pb 4 O 4 cubane motif ( Figure 11) in which each cubane node linked to four others by pairs of bridging ebdc ligands to give an extended (4,4) 2D net ( Figure 12). The two carboxylate groups of ebdc have different coordination modes, whereby one carboxylate chelates a single lead(II) atom in a κO:κO′ manner, and the second carboxylate forms the corner of the cubane cluster, thereby coordinating three lead atoms in a μ 3 -(μ 3 O:κO′) fashion. Ethynyl-derived weak hydrogen bonds constitute the only interactions linking the 2D sheets, in which the ethynyl groups of bridging ebdc ligands coordinating to carboxylate groups located above or below the net in an alternating fashion. The C(10)···O(1′) hydrogen bond to the carboxylate group of a symmetry generated cluster ( 3 / 2 − x, y − 1 / 2 − y, 3 / 2 − z) has a distance of 3.152(5) Å, with a favorable angle for bonding of 171°( Figure 13).     The anionic structure is charge-balanced in the lattice by a triethylammonium cation that contains one ethyl arm disordered over two positions. A molecule of methanol and three molecules of water, each with partial occupancy, complete the asymmetric unit. Two of the zinc(II) atoms are fourcoordinate and exhibit tetrahedral coordination geometry. The final zinc(II) atom is six-coordinate with octahedral coordination geometry. The coordinated water molecule and a symmetry equivalent bridge between two Zn 3 (OH) 5+ clusters, generating a hexanuclear motif that is held together by OH···O hydrogen bonds involving the hydroxy and aqua ligands ( Figure 14). Of the six carboxylate groups, four coordinate to the cluster in a κ-COO manner, with the nonbonding oxygen atom hydrogen bonding to the μ 3 -OH group, O(1)···O (9) 6) and S(8) graph set motifs, respectively. The remaining two carboxylate groups bridge between two zinc centers in a μ-(κO:κO′) manner. Each solvent molecule present in the lattice also participates in hydrogen bonding, but it proved to be difficult to assign hydrogen positions unambiguously (for this reason, only a single methanol molecule has been shown in Figure 14). The lattice methanol molecule forms hydrogen bonds to the carboxylate and coordinated methanol, O(19)···O(9) (D = 2.684(9) Å), O(15)···O(19) (D = 2.705(7) Å), which can be designated as R 3 2 (8). Pairs of ebdc ligands further stabilize the cluster core by π−π stacking parallel to the bridging water molecules ( Figure 15). The participating ligands are symmetry-equivalent and give a parallel offset interaction with an intercentroid distance of 3.59 Å (closest π−π contact, 3.38 Å). The 3D network propagates by bridging ebdc ligands of fused [Zn 6 (μ 3 -OH) 2 (μ 2 -H 2 O) 2 ] 10+ centers, forming 1D channels that contain solvent and counterions ( Figure 16). Analogous hexanuclear zinc(II) secondary building units have been observed in a limited number of instances. 32,33 There are three unique ethynyl groups within the 3D network of 5. Two of these groups align in a manner suggestive of C−H···π hydrogen-bond interaction, with one acting solely as a donor and the other solely as an acceptor ( Figure 17). Such T-shaped dimers have been calculated to impart stabilization energies on the order of 1.0 to 2.0 kcal/mol. 34 The average C(20)−H···π distance is 2.74 Å, which is equivalent to the mean distance (2.72 Å) observed in a study of such terminal alkyne interactions. 5 The final ethynyl group is directed toward the carboxylate group of an adjacent ligand with a C(30)···O(8) distance of 3.453(7) Å and an offset angle of 153°( Figure 18).

Crystal Growth & Design
As in the case of copper network 3, the band intensity for the ethynyl C−H stretching vibration is both decreased and broadened; however, three unique bands were observable at 3301, 3271, and 3252 cm −1 that can be assigned to the three ethynyl groups. The non-hydrogen-bonding ethynyl ligand in 5 matches the frequency of the non-hydrogen-bonding reference of H 2 ebdc, at 3301 cm −1 . The band at 3271 cm −1 is red-shifted by 30 cm −1 , which is characteristic of C−H···π bonding, 4 and also possesses the strongest intensity, which likely reflects the optimal T-shaped geometry of the interaction shown in Figure  17. The final band at 3252 cm −1 has a red-shift of 49 cm −1 and likely corresponds to the ethynyl C−H···O bond in Figure 18, with its diminished intensity likely a result of poor alignment with the carboxylate.
Packing Influence of ebdc. This study has identified several lead(II) networks, 1, 2, and 4, which exhibit ethynylbased weak hydrogen-bonding interactions that directly influence crystal packing. Two factors appear to favor C(sp)− H···O interactions within these structures, the first being the flexible coordination polyhedra imparted by the lead(II) atoms in conjunction with O-donor ligands. This allows the ebdc ligands to adopt a wide range of coordination modes and thus possess more freedom to achieve optimal hydrogen-bonding geometry. The second factor, in part, may relate to the choice  By contrast, the two transition metal structures, 3 and 5, are both found to contain extensive hydrogen-bonding arrays as well as π−π stacking interactions, both of which are lacking in the lead(II) structures. Only minor ebdc-based weak hydrogenbonding interactions were observed by IR spectroscopy, suggesting that the crystal packing, and thus self-assembly process, for these structures was not governed by weak hydrogen-bonding interactions to any large extent. While methanol was found to participate in these hydrogen-bonding nets, in most cases, it was observed to be substitutionally disordered with water molecules; water was also available in the wet solvents used during synthesis of the lead(II) structures. This suggests that solvent choice is not the main reason for a lack of weak hydrogen-bonding interactions. Instead, it is likely that the more rigid coordination environment provided by the transition metals promotes the inclusion of more coordinated solvent, which in turn increases the availability of strong hydrogen donors. The apical positions of the copper paddlewheel motif, which are typically coordinated by aqua ligands, 35 demonstrates this effect. Similarly, the high Lewis acidity of zinc(II) promotes formation of hydroxo species and, in turn, cluster formation, which also likely favors strong, as opposed to weak, hydrogen bonding through the inclusion of more strong hydrogen donors into the crystalline structure. This provides the rationale for the contrasting behavior of lead  While gaining control over weak hydrogen-bond interactions during the self-assembly of coordination polymers remains a distant goal at the present time, this study has identified lead(II) and its corresponding coordination polymers as interesting species for study using ligands containing weak hydrogen-bond donors, in which the resulting weak hydrogen bonds are influential in forming the solid state-structures. This work will be expanded to include a range of other heavy metals ligated by ebdc as well as varying the weak hydrogen-bond donor on similar diacids in combination with lead(II) to extend the scope of these findings. Packing motif of 1 ( Figure S1), asymmetric units of 2 ( Figure  S2) and 4 ( Figure S3), hydrogen-bonding tables for H 2 ebdc and 5, selected bond lengths and angles, and PXRD spectra for networks 1−5. This material is available free of charge via the Internet at http://pubs.acs.org. CCDC 1025919−1025924 contain the supplementary crystallographic data for this article. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via http://www.ccdc.cam.ac.uk/ Community/Requestastructure/Pages/DataRequest.aspx?.

Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.