Leigh
Loots
*,
Delia A.
Haynes
and
Tanya le
Roex
Department of Chemistry and Polymer Science, University of Stellenbosch, P. Bag X1, Matieland, 7602, Republic of South Africa. E-mail: leighl@sun.ac.za; Fax: +27 21 808 3360; Tel: +27 21 808 3348
First published on 7th May 2014
A new class of potential supramolecular building blocks is presented. The synthesis and crystal structures of seven novel pyridinium-derived zwitterionic compounds are discussed, as well as two related pyridinium salts. The synthesis of these compounds was achieved by simple solution methods using mild conditions, and the products were isolated as single crystals with no further purification required. Hydrogen-bonded chains were found to be prevalent in all of the crystal structures of the zwitterions and due to the robust nature of this motif, these compounds have the potential to be exploited in future supramolecular studies. There is a large scope for the synthesis of related zwitterionic compounds using a range of pyridyl substituents with varied functional groups. These zwitterionic compounds show great potential as a new class of supramolecular building block in the formation of ionic- as well as metal–organic frameworks.
In this study, a unique series of pyridyl-derived zwitterionic compounds have been isolated, purified and structurally characterised. The zwitterions were obtained from attempts to crystallise novel organic salts using a variety of simple acids and bases. These compounds were all obtained as single crystals by the simple reaction of acetylenedicarboxylic acid (ADC) with a selected pyridyl-derivative (Scheme 1).
![]() | ||
Scheme 1 The general reaction scheme for the formation of zwitterions from acetylenedicarboxylic acid and various pyridyl derivatives. |
To date, seven novel pyridinium-derived zwitterions have been synthesised in our laboratory by following the same simple procedure. To our knowledge, the crystal structures of only four other similar compounds have been reported in the Cambridge Structural Database (CSD) (SUCPYR10,18 SUCPYR11,19 LAQGAN,20 GOMQIP21 and TUPWEH22).
Having obtained this series of compounds, together with their crystal structures, the effects of different pyridyl derivatives on the product of the reaction, as well as on the packing arrangements of these molecules in the solid state, can be examined. Because of our interest in ionic organic framework materials,23 we have investigated the possibility that extending the length of the pyridyl substituent would create more space within the crystal structure, leading to the inclusion of guest molecules. The pyridyl derivatives used in this study vary in both their size (length) as well as their functional groups. The different functional groups were chosen to give insight into the effect of these factors on (1) the reaction with the acetylene group, (2) the crystal structure of the resulting zwitterion, and (3) any intermolecular interactions. Zwitterions were successfully synthesised and crystallised using the following pyridyl derivatives: pyridine (1), isoquinoline (2), 3-hydroxypyridine (3), 4-pyridinecarbonitrile (4), isonicotinic acid (5), 4-phenylpyridine (6), 4-benzylpyridine (7) (Scheme 2). This account describes the synthesis of 1–7, several of the crystal structures obtained thus far, as well as the challenges and future prospects for utilising this series of zwitterions as building blocks for solid-state framework materials.
A similar procedure was followed for all reactions as shown in Scheme 1, varying only the solvent or solvent mixture (see Experimental section for full details). An advantage of this method of synthesis is that there is no need for any workup of the reaction, as the resulting crystalline zwitterion can simply be filtered from the solution. If necessary, the crystals can be washed with water to remove any starting material since the products are only sparingly soluble in hot water and require the addition of a base to dissolve. Most experiments were initially carried out using small quantities as test reactions, and later scaled up. Along with the seven zwitterionic compounds, two salts were also obtained during this study (5a and 6a).
Because the literature reports the trapping of zwitterions using aldehydes,3 a further reaction was attempted: an excess of benzaldehyde (10×) was added to the reaction mixture of ADC and pyridine. This did not prevent the formation of the zwitterion. Perhaps this is because a 1:
1 molar ratio of acid to pyridine was used rather than only a catalytic amount of pyridine (as it is reported in the literature24), even though benzaldehyde was used in excess.
The majority of zwitterionic products reported in this study were obtained from para-substituted pyridyl derivatives. Ortho- and meta-substituted isomers of derivatives (see Scheme S1 in ESI†) yielding 1–7 are also currently being investigated in an attempt to rationalise any directing effects that these substituents may have on the reaction.
Each of the zwitterions has a T-shaped conformation in the solid state, and changes to the pyridyl substituent result in subtle differences in the conformations of the individual zwitterions. Twisting of these pyridyl groups to accommodate the increased steric bulk of the substituents is observed in the crystal structures of 2, 6MeOH and 7. In order to quantify/investigate these differences, the angle between a plane through the carboxylate chain and a plane through the pyridyl moiety has been measured. The results are shown in Table 1. There appears to be a slight correlation between the angle and the presence of interaction between the pyridyl substituents. The outliers 2, 6MeOH, 7 all have some weak intermolecular interactions between the pyridyl moieties (2 = π⋯π, 6MeOH = C–H⋯O and π⋯π, 7 = C–H⋯π).
Compound | Angle/° |
---|---|
1 | 81.50 |
2 | 62.72 |
3 | 70.89 |
4 | 80.99 |
5 | 77.20 |
6H2O | 75.27 |
6MeOH | 66.23 |
7 | 89.30 |
In all of the zwitterion crystal structures reported here, the carboxylate and carboxylic acid groups at either end of the alkyl chain hydrogen bond to one another to form one-dimensional chains. The major differences in this series of structures appear in the way these chains pack together. In the majority of the crystal structures obtained, the hydrogen-bonded chains interdigitate to resemble two-dimensional ladders (Fig. 1). These ladders then pack alongside one another in an offset manner resulting in layers of ladders. The crystal structures are described in detail below, so as to highlight the differences between them.
![]() | ||
Fig. 1 One-dimensional hydrogen-bonded chains in the crystal structures 1 (a), 2 (b), 3 (c), 4 (e) and 5 (d). These chains are arranged in ladder motifs in the structures of 1, 2, 3 and 5. |
1 crystallises in the chiral space group P212121 with one molecule in the asymmetric unit (ASU). The ladders of hydrogen-bonded chains in the structure of 1 run along the crystallographic a-axis (Fig. 1a). These ladders pack in such a manner that the hydrogen-bonded chains of adjacent ladders overlap slightly (like roof tiles) to form two-dimensional layers along (001) as shown in Fig. 2. These layers then pack in an offset arrangement (⋯ABA⋯) along [001].
![]() | ||
Fig. 2 Overlapping layers of unlinked ladders in the structure of 1. Different colours are used to distinguish between the individual ladder motifs. |
2 crystallises in the space group P21/n with one molecule in the ASU. Similar to the structure of 1, the isoquinolinium groups of the hydrogen-bonded chains interdigitate (Fig. 1b) but because they have a larger surface area than the pyridinium of 1, they force the one-dimensional chains to be further separated from one another (approx. 7.05 Å in 1, 9 Å in 2, Fig. 1b). In the structure of 2, ladders are organised into layers along [010]. Unlike the structure of 1, the isoquinoline moieties in 2 are arranged in pairs that participate in offset π⋯π interactions (3.375 Å, Fig. 3). It is evident from the results in Table 1 that the conformation of the molecule is quite different to the other zwitterions (more acute angle) and these π⋯π interactions may be a contributing factor.
The greater surface-area-to-volume ratio of the isoquinoline ring (compared to pyridine) may result in an increased affinity for this π-interaction. So in this instance the rungs of the ladder may be held in place by these π⋯π interactions. The two-dimensional ladders arrange in a similar manner to that observed in 1, with neighbouring hydrogen-bonded chains overlapping to give layers. However, the layers of ladders in 2 do not assemble in an offset manner but rather are stacked one on top of another along [010].
3 crystallises in the space group P21/c with one zwitterion and a water molecule in the ASU. The water molecule is situated such that it forms hydrogen bonds to both a hydroxyl group and a carboxylate group in the adjacent chain, resulting in a hydrogen-bonded ladder rung (Fig. 1c). The packing arrangement of 3 is similar to that found in 2; however, the overlapping one-dimensional chains are linked via hydrogen bonding between the water molecule in one ladder and the free oxygen atom of the carboxylate in the next ladder (Fig. 4). The inclusion of water allows further separation of the interdigitating carboxylate chains to approximately 9.5 Å, compared with that of 1 (7 Å) and 2 (9 Å).
![]() | ||
Fig. 4 The packing arrangement of 3 showing hydrogen bonding between chains to form a hydrogen-bonded ladder motif. |
4 crystallises in the monoclinic space group P21/c with one molecule in the ASU, and forms similar one-dimensional hydrogen-bonded chains to those in the previous structures (Fig. 1e). However, the packing of this zwitterion is quite different to the rest of the compounds in the series. The two-dimensional assembly does not resemble a ladder, but rather two one-dimensional chains run alongside each other with the cyanopyridyl moieties arranged in a herringbone pattern (Fig. 5). Other than the hydrogen bonds forming the one-dimensional chains, there are C–H⋯O contacts between molecules in adjacent chains (Fig. 5). C–H⋯N contacts are also present between nitrile moieties and an aromatic hydrogen atom in the neighbouring molecules. Each set of parallel chains are arranged at an angle of 180° to the adjacent set, such that the carboxylate hydrogen-bonded chains overlap. This arrangement produces a zigzag pattern along [010] forming one layer and these layers then pack one on top of another along [100] to give the three-dimensional packing arrangement.
5 crystallises in the triclinic space group P with one whole molecule in the ASU. All three acid groups are involved in hydrogen bonding. The carboxylate and acid groups of the alkyl chain hydrogen bond to form chains along [010] (Fig. 1d). Two chains run parallel to one another (orientated 180°) and are linked through hydrogen bonding between the isonicotinate moieties and the free oxygen atoms of the carboxylate chains. This results in double rungs for the two-dimensional ladders, analogous to those seen in 3. The ladders pack in a slightly offset fashion (Fig. 6), however the carboxylate chains do not overlap as they do in all the previous structures (see above). The two-dimensional layers stack on top of one another in an offset manner such that the carboxylate chains overlap with the layer above or below, forming a staircase pattern.
![]() | ||
Fig. 6 Two-dimensional hydrogen-bonded ladders of 5 stack one on top of another to form layers along (001). |
A salt product (5a) was also obtained from the same components as 5. 5a crystallises in the triclinic space group P with one 4-carboxypyridinium cation, one carboxypropynate anion and a water molecule in the ASU. The propynate anions form one-dimensional hydrogen-bonded chains along [010] with pendant carboxypyridinium cations hydrogen bonding to the other oxygen atom of the carboxylate (Fig. 7). The water molecules connect two of these chains, which are stacked on top of one another, by hydrogen bonds to the pyridinium moiety and the acid group of the propynate anion. These connected chains resemble columns that pack in offset stacks (Fig. 8).
![]() | ||
Fig. 7 One layer of 5a viewed down (101) showing the one-dimensional hydrogen-bonded propynate anion chains with the 4-carboxypyridinium pendant cations. |
In 6 and 7 the pyridyl moiety has been extended with the addition of a phenyl (6) or a benzyl (7) substituent. This leads to less efficient packing of the zwitterions, and either small guest molecules are included or the molecules are contorted to pack more effectively. More degrees of freedom are also introduced, allowing for greater conformational flexibility.
Two crystalline solvates were obtained for compound 6, a hydrate (6H2O) as well as a methanol solvate (6MeOH). In addition, a salt was obtained using the same components (6a). 6H2O crystallises in the space group C2/c with one zwitterion molecule and three water molecules (two of which are partially occupied) in the ASU. The phenylpyridine moiety is also disordered over two positions of equal occupancy so that it appears this moiety is twisting slightly. This twisting motion is assumed to be owing to the movement of the water molecules and this is supported by thermal gravimetric analysis (TGA), which shows loss of water close to ambient temperatures (see Fig. S7, ESI†). The zwitterion packs in a similar manner to that seen in 1, forming one-dimensional hydrogen-bonded chains that interdigitate to form an unconnected ladder. The partially occupied (42% and 28%) water molecules (and their symmetry related molecules) are located between two interdigitated chains in the space alongside the phenylpyridine groups (Fig. 9a). The third water molecule is hydrogen bonded to the other two water molecules and also links two overlapping carboxylate chains.
6MeOH crystallises in space group P21/n with one zwitterion and one molecule of methanol (disordered over two positions) in the ASU (Fig. 10). The 4-phenylpyridine moiety is disordered over two positions of equal occupancy as in the structure of 6H2O, however, it appears that there is no twisting action, but rather a sweeping motion of the phenyl ring. The sweeping motion of the phenyl ring appears to mirror the disorder positions of the methanol molecules. The hydrogen atoms of the acid moieties are also disordered over two possible positions (55:
45).
Although analogous hydrogen-bonded chains form as in 6H2O, the structures are quite dissimilar. Rather than the hydrogen-bonded chains interdigitating as in 6H2O, in 6MeOH they form more of a herringbone pattern (Fig. 9b). The hydrogen-bonded carboxylate chains do not overlap but in this case are separated by the methanol molecules. The methanol molecules appear to be located in one-dimensional channels that run parallel to the crystallographic b-axis. TG analysis (Fig. S8, ESI†) confirms the presence of one methanol molecule and solvent loss commences at ambient conditions.
The salt 6a crystallises in the monoclinic space group P21/n with two pyridinium cations, one acetylenedicarboxylate dianion and a methanol molecule in the ASU (Fig. 11). The dianion hydrogen bonds to each of the symmetry independent pyridinium cations and the methanol hydrogen bonds to one of the free oxygen atoms to form a four-membered adduct. Phenylpyridinium cations stack one on top of another along [010]. The molecules are arranged 180° to one another such that the pyridyl moiety aligns with the phenyl moiety of the next molecule in the stack. The rings are stacked slightly offset with distances between the rings of 3.683 and 3.837 Å.
Three different crystal structures were obtained for 7 – a close-packed structure (7), a polymorph of 7 (7open) and a 1,4-dioxane solvate (7diox). The synthetic details for each can be found in the Experimental section, however only the crystal structure 7 will be discussed in detail here. 7 crystallises in the space group Pbca with one zwitterion molecule in the ASU. The packing arrangement of this zwitterion is remarkably different from the others in the series, although the carboxylate hydrogen-bonded chains are persistent. These chains do not form an interdigitated ladder as in all the structures described thus far. Rather, chains stack alongside one another to give anti-parallel arranged layers (Fig. 12). The layers stack on top of one another such that the pyridyl groups block the spaces created by the benzyl moiety in the neighbouring layer, resulting in bilayers. These bilayers then pack along [100] in an⋯ABCD⋯manner.
When ACD and isonicotinic acid were ground together with a few drops of the chosen solvent (neat, H2O or ethyl acetate) the result was always the same; the PXRD pattern (Fig. S15, ESI†) did not match the zwitterion, but rather more closely resembles the ACD-isonicotinate salt, 5a. The grinding together of acetylenedicarboxylic acid and 4-benzylpyridine with a few drops of dioxane resulted in a colourless oil, which was suspended in enough solvent (approximately 1 ml of dioxane/ethanol) that small crystals crashed out of solution. These crystals were subsequently analysed by single-crystal diffraction and determined to be 7open, a polymorph of 7 (see Experimental section). This series of structures will be discussed in detail in a future paper.
Because the same hydrogen-bonded chains of carboxylates were observed in all the compounds, it would be of interest to ascertain if it is possible to either extend these networks or find suitable candidates that will disrupt these networks to produce more interesting arrangements. It has been observed in the crystal structures of this series that we can extend the distance between carboxylate chains by using longer pyridyl derivatives. It would be interesting to see if it is possible to extend this even further using hydrogen bonding to another molecule – i.e. forming cocrystals or salts. From the results it is clear that increasing the length of the substituent increases the probability of guest inclusion. However, it appears that the identity of the guest is important in determining whether it will be included or not, although we cannot be certain what the requirements are. It may be as simple as a size/shape match or it could be more complicated such as polarity of the guest or whether there are stabilising interactions between host and guest.
It would also be of interest to have two different substituents on the same zwitterion. A number of mixed substituent reactions have been attempted with this goal in mind, but so far, none has yielded anything other than the singly substituted zwitterion or a salt. In one instance when 4-phenylpyridine and pyridine were combined with acetylenedicarboxylic acid, the reaction yielded a few crystals of the salt (6a) and a white precipitate. The white precipitate was identified by PXRD analysis as a mixture of 1, 6H2O and the salt 6a. Combining ADC with 4-phenylpyridine and pyridinecarbonitrile in a 1:
1
:
1 molar ratio in a mixture of MeOH and water yields a mixture of compounds 4 and 5. This begs the question: is there selectivity for one substituent over another? If this is indeed true, further reactions are required to determine which substituent is the more favoured and propose a reason for this selectivity.
Our initial objective with this study was to construct new assemblies using organic salts and cocrystals. Having serendipitously discovered these novel stable zwitterions, it seems only natural to pursue supramolecular materials using these compounds as building blocks. With these seven new zwitterions in hand, having different pyridyl substituents ranging in size and functionality, it is likely that a wide range of supramolecular architectures will be produced, owing to the unique conformation of these ions, as well as their hydrogen bonding potential. Organic as well as metal–organic materials are currently being investigated.
PXRD patterns were collected using a PANalytical X'Pert Pro diffractometer with Bragg–Brentano geometry using CuKα radiation (λ = 1.5418 Å) at 45 kV and 40 mA. Intensity data were captured with an X'Celerator detector with 2θ scans performed in the range 5–50° with a 0.0167 step size. Samples were spun at 4 revolutions per second.
1H and 13C NMR spectra were recorded (in D2O) with an Agilent spectrometer at 600 MHz, unless otherwise specified, at 25 °C. Between 0.02–0.04 ml of triethylamine was added to each of the samples to aid dissolution in D2O – with the exception of 5a and 6b where this was not necessary – the signals for triethylamine are not indicated in the text.
Differential Scanning Calorimetry (DSC) was carried out using a TA Instruments Q20 system under a N2 gas purge (flow rate of 50.0 ml min−1) coupled to a RCS cooling unit. Samples were cycled from −20 °C to approximately 200 °C (depending on Tdec (TGA)) at a heating rate of 10 °C min−1 and a cooling rate of 5 °C min−1. This was performed using non-hermetically sealed aluminium pans.
Thermogravimetric analysis was carried out using a TA Instruments Q500 system under a N2 gas purge (flow rate 50.0 ml min−1) using aluminium sample pans at a ramp rate of 10 °C min−1.
Melting point determinations were carried out using a Stuart SMP3 melting point apparatus.
A salt of acetylenedicarboxylic acid and isonicotinic acid (5a) is obtained reliably by mechanochemical methods – ADC (0.020 g, 0.133 mmol) was ground together with INA (0.016 g, 0.130 mmol) and a suitable solvent (water or ethyl acetate) in a mortar and pestle for approximately 5–10 min. 5a can also be isolated from the solution mixture (water–EtOAc) prepared for 5 – colourless crystals are obtained from this mixture within a few hours. Minor amounts of 5 are observed in the NMR of 5a as indicated. Tmp = 132–137 °C; δH (600 MHz, D2O) 7.25 (2 H, s – 5), 7.69 (2 H, J = 6.15 Hz, d), 8.28 (2 H, J = 6.44 Hz, d – 5), 8.55 (2 H, J = 6.15 Hz, d), 8.76 (2 H, J = 6.74 Hz, d – 5); δC (600 MHz, D2O) 76.41, 123.77, 126.71 (5), 133.93 (5), 145.98, 146.31 (5), 149.85, 160.33, 173.68.
In the presence of pyridine, the reaction yields a salt (6a) of 4-phenylpyridine and acetylenedicarboxylic acid that crystallises as a methanol solvate. Tdec = 105–110 °C, δH (600 MHz, D2O) 7.46 (2 H, J = 7.62 Hz, t), 7.51 (1 H, J = 7.03 Hz, d), 7.69 (2 H, J = 7.62 Hz, d), 8.04 (2 H, J = 6.45 Hz, d), 8.56 (2 H, J = 7.03 Hz, d); δC (600 MHz, D2O) 76.29, 124.60, 128.40, 130.24, 132.69, 134.46, 141.48, 145.16, 158.08, 160.11.
If the crystals of the methanol solvate (6MeOH) are filtered off and left to dry in air, the crystals convert to the hydrate. When the filtrate obtained in this way is allowed to evaporate, a different product is obtained – the decarboxylation product, 6b (Scheme 3), which crystallises with one molecule each of the starting materials as well as water.
Footnotes |
† Electronic supplementary information (ESI) available: PXRD, TGA and DSC data, NMR spectra for 4, selected hydrogen bonding tables as well as selected crystallographic data. CCDC 987540–987549. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4nj00281d |
‡ The degree of protonation was determined by measuring the C–O bond lengths of each of the carboxylate groups [∼1.2 Å = C![]() |
§ A colour change occurred upon addition of triethylamine to the NMR sample indicating a possible reaction occurring. This is evident in the NMR with 4 being the major component along with a minor amount of an unknown compound. Only the NMR resonances for 4 are presented here, however the full spectra are available in the ESI.† |
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014 |