Introduction

Wheel-and-axle host compounds were first described by Toda and Akagi in 1968 [1]. These compounds all consist of a long linear axis, referred to as the “axle”, the ends of which are characterised by the presence of bulky, usually rigid, substituents, the “wheels”. Their examples included the compounds 1,1,6,6-tetraphenylhexa-2,4-diyne-1,6-diol (1) and 1,1,4,4-tetraphenylbut-2-yne-1,4-diol (2) (Scheme 1), which were reported to form complexes with an extremely wide variety of guest compounds, including ketones, amines, amides and sulfoxides, amongst numerous others.

Caira and coworkers further investigated the host behaviour of 1 in guest exchange experiments (with THF, furan and thiophene as the guest species) and for the separation of isomeric mixtures [2,3,4]. In mixed lutidines, 1 was significantly more selective for 3,5-lutidine while disfavouring the 2,4-isomer. Furthermore, the preference of 1 was for 2-aminobenzonitrile when crystallization experiments were carried out in isomeric aminobenzonitriles.

Scheme 1
scheme 1

The general structures of various wheel-and-axle host compounds

Full size image

Other host compounds that have this generalized wheel-and-axle geometry that have been reported were synthesized from starting materials based on the fluorenyl tricyclic fused system (3) (Scheme 1). The R group was varied extensively, from multiple thiophene units [5] and acetylene moieties, to one or more aromatic rings [6], to provide host compounds with exceptional host abilities.

Host-guest chemistry is a subfield of supramolecular chemistry, this branch of science having extraordinarily useful applications in many industries. This type of chemistry has been employed to resolve racemates into their constituent enantiomers [7] and to separate positional isomers from one another [8]. Furthermore, host compounds have further applications in chemosensing, removing hazardous substances from the environment and in chromatography [9,10,11]. In the pharmaceutical industry, cyclodextrins, more especially, have been found to be extremely useful for stabilizing active pharmaceutical ingredients and to assist in drug transport and delivery to the desired target site [12,13,14].

Supramolecular chemistry relies on the presence of weak noncovalent interactions (hydrogen bonding, π stacking and other van der Waals forces) between two different species, which then facilitate the formation of complexes between these interacting species [15,16,17]. The search for host compounds that have greater efficiencies and complementary selectivity behaviours relative to compounds that are known is ongoing. To illustrate, the roof-shaped host compounds dimethyl trans-9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboxylate, trans-α,α,α′,α′-tetraphenyl-9,10-dihydro-9,10-ethanoanthracene-11,12-dimethanol and trans-α,α,α′,α′-tetra(p-chlorophenyl)-9,10-dihydro-9,10-ethanoanthracene-11,12-dimethanol were assessed for their selectivity behaviours in pyridine and 2-, 3- and 4-methylpyridine (PYR, 2MP, 3MP and 4MP) by means of crystallization experiments from mixtures of these organic solvents, and it was demonstrated that these host compounds possessed complementary selectivities: the first preferred PYR, the second 3MP and the last 2MP [18]. The correct selection of the host compound for the separation of the MPs, therefore, enables the extraction of the desired pyridyl species into the cavities of the so-formed complex.

In the present work, the selectivity behaviour of the wheel-and-axle host compound 1,4-phenylene-bis(di-p-tolylmethanol) (H) was investigated in these mixed pyridines with the view to determining whether H has the ability to separate, more especially, the MPs through host-guest chemistry strategies (Scheme 2). Current separation technologies employing fractional distillations are wholly unsatisfactory, these being highly energy-intensive and economically unfavourable. This is as a result of the very narrow boiling range of these MPs (2MP 129, 3MP 144 and 4MP 145 °C). In an earlier investigation, a related host compound, 1,4-bis(diphenylhydroxymethyl)benzene, was demonstrated to have the capability to separate very many pyridyl guest mixtures that were used in that work [19]. This host compound favoured either PYR or 4MP dependent upon the type and amount of each guest species present in the solution. Hence the focus of the present investigation is to determine whether H possesses selectivity when presented with mixed pyridines and to observe whether the minor change in host structure (from unsubstituted aromatic rings to p-methyl-substituted moieties) will affect this preferential behaviour. All novel complexes produced in this work were analysed by means of both single crystal X-ray diffraction (SCXRD) and thermal analyses, and these crystal structures and results are now reported herein.

Scheme 2
scheme 2

The molecular structures of 1,4-phenylene-bis(di-p-tolylmethanol) (H), the host compound, and the potential guest species pyridine (PYR) and 2-, 3- and 4-methylpyridine (2MP, 3MP, 4MP)

Full size image

Materials and methods

General

All required materials for the synthesis of 1,4-phenylene-bis(di-p-tolylmethanol) (H) as well as the four pyridyl guest species were purchased from Merck (South Africa) and were used without further modification.

Nuclear magnetic resonance (NMR) experiments were carried out in CDCl3 by means of a Bruker Ultrashield Plus 400 MHz spectrometer, and data were analysed by means of Topspin 4.2 software.

Gas chromatography (GC) analysis was required to quantify the pyridyl guests in any mixed complexes produced in this work. An Agilent 7890 A gas chromatograph coupled to an Agilent 5975 C VL mass spectrometer was the applicable instrument, which was fitted with an Agilent J&W Cyclosil-B column. The method involved an initial 2 min hold time which was followed by a 30 °C·min‒1 heating ramp until 100 °C was reached. A second ramp at 1.5 °C·min‒1 was then implemented to reach 102 °C, followed by another 0.5 °C·min‒1 ramp to reach 103 °C. The flow rate was 1.5 mL·min‒1 and the split ratio 1:80. Dichloromethane was the dissolution solvent in all of these experiments.

The crystal structure data for any complexes with suitable crystal quality were obtained by means of a Bruker Kappa Apex II diffractometer with graphite-monochromated MoKα radiation (λ = 0.71073 Å). APEXII was used for data collection and SAINT for cell refinements and data reduction. SADABS was employed for absorption corrections [20]. SHELXT-2018/2 [21] was used to solve the structures and refinements were carried out by means of least-squares procedures using SHELXL-2018/3 [22] together with SHELXLE [23] as a graphical interface. In some cases, a Bruker D8 Quest diffractometer equipped with a Photon II CPAD detector was used instead with an IµS 3.0 Mo source (Kα, λ = 0.71073 Å). SAINT was employed for data reduction and cell refinement, and APEX4 for data collection; the SADABS numerical approach was used to adjust data for absorption effects [24]. Using a dual-space approach, SHELXT–2018/2 [21] was employed to solve the structure. SHELXL–2019/3 [22] and SHELXLE [23], as a graphical interface, through least-squares procedures, were used to improve the solution. ORTEP diagrams were prepared with ORTEP-3 for Windows (version 2023) [25]. All non-hydrogen atoms were refined anisotropically. Carbon- and oxygen-bound hydrogen atoms were added in idealized geometrical positions in a riding model. The nitrogen-bound hydrogens were located on the difference Fourier maps and included in the refinement using riding models. Data were corrected for absorption effects using the numerical method implemented in SADABS [20]. These crystal structures were deposited at the Cambridge Crystallographic Data Centre, and their CCDC numbers are 2,340,378‒2,340,381.

Thermal analyses were carried out by means of a TA SDT Q600 module system. The complexes were first isolated from their solutions using suction filtration. The crystals were then washed (while under suction) with low boiling petroleum ether and then padded dry in folded filter paper. The data were analysed using either Pyris or TA Universal data analysis software. Samples were heated in ceramic pans with an empty pan serving as the reference; the heating rate was 10 °C·min‒1 from approximately 40 to 400 °C. The purge gas was high purity nitrogen.

Synthesis of 1,4-phenylene-bis(di-p-tolylmethanol) (H)

The host compound, 1,4-phenylene-bis(di-p-tolylmethanol) (H), was synthesized by means of a standard Grignard addition reaction with p-bromotoluene as the halogenated compound according to a known procedure [26].

Formation of the single solvent complexes

In order to determine whether H possessed host ability for any of the four pyridyl guest solvents, this host compound was crystallized from each of these. This was achieved by dissolving H in an excess of each guest solvent in glass vials with gentle heating being applied to ensure complete host dissolution. The vials were left open to the atmospheric conditions which facilitated the crystallization process. The crystals were collected under vacuum, washed with low boiling petroleum ether (also under suction), and analysed by means of 1H-NMR spectroscopy. Complexation was successful if resonance signals for both the host and guest species were present on the resultant spectra. In these instances, the host: guest (H: G) ratios were calculated by considering and comparing the areas under applicable host and guest resonance signals.

Formation of mixed complexes by host crystallization experiments in equimolar mixtures of the pyridyls

The selectivity behaviour of H was investigated by means of crystallization experiments from equimolar mixed pyridyl guest solutions. As such, equimolar binary, ternary and quaternary guest solutions were prepared (0.5‒1 mmol combined amount) in glass vials and H (0.5–0.7 g; ≈ 0.1 mmol) dissolved in each of these mixtures. The vials were closed and stored in a refrigerator (4 °C), and the crystals that formed in this fashion were treated as in the single solvent crystallization experiments. The guests in these resultant mixed complexes were quantified by means of GC analyses. Experiments were conducted in duplicate, and the percentage estimated standard deviations were calculated in order to determine the repeatability of these experimental results.

Formation of mixed complexes by host crystallization experiments in binary guest mixtures where guest amounts were successively varied

Binary guest solutions were prepared by mixing two guest solvents (GA and GB) in 20:80, 40:60, 60:40 and 80:20 molar ratios (0.5‒1 mmol combined amounts). The host compound (0.5–0.7 g; ≈ 0.1 mmol) was then dissolved in these solutions in glass vials. These vessels and the resultant solids were treated in an identical manner to the equimolar guest experiments. The crystals were, once more, analysed by means of GC experiments which provided the amounts of the two guests present in the mixed complexes. Selectivity profiles were subsequently constructed by plotting the amount of Guest A (or Guest B) in the crystals (Z) against this amount in the prepared solution (X). These profiles demonstrate the selectivity behaviour of H as guest amounts in the solutions were varied. An unselective host compound would have a selectivity coefficient, K, of one, where KA: B = ZA/ZB x XB/XA (XA + XB = 1) [27]. This scenario is represented by the diagonal line in each plot which was inserted so that comparisons may be made facilely with the experimentally determined data points.

Software

Program Mercury was used to construct unit cell diagrams from the data obtained from the SCXRD experiments as well as to observe the host-guest packing, bond lengths, bond angles, and covalent and noncovalent interactions [28]. Furthermore, after removing each guest species from the packing calculation, the type of guest accommodation (whether in channels or discrete cages) could be identified by examining the remaining void areas that were able to accommodate a probe with a radius of 1.2 Å. Program Crystal Explorer 17.5 was used to generate three-dimensional surfaces around the guest molecule in each complex in Hirshfeld surface analyses [29]. From these surfaces were then constructed two-dimensional fingerprint plots, also by means of Crystal Explorer 17.5, which allowed for the quantification of the atom···atom interactions between atoms inside (di) and outside (de) each three-dimensional surface [30]. In the case of guest disorder, each disorder component was considered independently by deleting each one in turn and performing the required calculations.

Results and discussion

Formation of the single solvent complexes

When H was independently crystallized from each of PYR, 2MP, 3MP and 4MP, inclusion compounds were obtained in each instance. The H: G ratios of these are summarised in Table 1.

Table 1 H: G ratios of complexes formed by H when crystallized from each of the pyridyl solvents
Full size table

The H: G ratios of these novel complexes varied and were 1:3 (when PYR was the guest solvent), 1:2 (2MP and 3MP) and 1:1 (4MP) (Table 1). (Host compound 1,4-bis(diphenylhydroxymethyl)benzene consistently formed 1:2 H: G complexes with these pyridyl guest species [19].)

Formation of mixed complexes by host crystallization experiments in equimolar mixtures of the pyridyls

Table 2 summarises the results that were obtained from GC analyses of the crystals obtained when H was crystallized from the various equimolar mixtures of the pyridyl guest solvents. Preferred guests are in bold text and %e.s.d.s are also provided here.

Table 2 Mixed complexes obtained from crystallization experiments of H from equimolar mixed pyridyls
Full size table

From Table 2, PYR was most favoured by H in the equimolar binary guest mixtures in which it was present: in each of the PYR/2MP, PYR/3MP and PYR/4MP experiments, H formed complexes with 68.7, 66.5 and 57.9% PYR, respectively. When PYR was absent but 4MP present, 4MP was then significantly favoured: 2MP/4MP and 3MP/4MP solutions furnished crystals with 71.0 and 84.7% 4MP. In the absence of both PYR and 4MP (2MP/3MP), 2MP was overwhelmingly preferred (83.4%). The ternary experiments demonstrated that the host selectivity was subject to change dependent upon which solvents were present. In PYR/2MP/3MP and PYR/2MP/4MP, PYR was the favoured guest species (63.7 and 49.9%), while in PYR/3MP/4MP, 4MP was then most preferred (48.7%). The ternary experiment in the absence of PYR (2MP/3MP/4MP) demonstrated that, once more, the 4-substituted guest compound was favoured, and significantly so (79.8%). Finally, the quaternary mixture resulted in a mixed guest complex with 41.1% 4MP. In general, therefore, it may be concluded that both PYR and 4MP were favoured guests of H, while 2MP and 3MP were, more usually, not preferred. In fact, 3MP was never preferentially selected by H in any of these crystallization experiments.

Interestingly, experiments in analogous conditions using related host compound 1,4-bis(diphenylhydroxymethyl)benzene demonstrated significantly enhanced host selectivities (54.2‒98.6%) compared with H (41.1‒84.7%, Table 2) as a result of the minor modification in the molecular structure of the host species [19]. However, both compounds were usually selective for PYR and 4MP.

Formation of mixed complexes by host crystallization experiments in binary guest mixtures where guest amounts were successively varied

The selectivity profiles that were constructed using the equation KA: B = ZA/ZB x XB/XA (XA + XB = 1) [27] are provided in Fig. 1a–f.

Fig. 1
figure 1

Selectivity profiles for H in (a) PYR/2MP, (b) PYR/3MP, (c) PYR/4MP, (d) 2MP/3MP, (e) 4MP/2MP and (f) 4MP/3MP binary mixtures; the dotted straight diagonal lines represent an unselective host compound

Full size image

When presented with the PYR/2MP binary mixtures (Fig. 1a), H remained selective for PYR across the concentration range. However, the highest K value was only 3.9, and this was calculated for the solution that contained 20% PYR, while the averaged K value for this set of experiments was 1.8. In PYR/3MP mixtures (Fig. 1b), H also remained selective for PYR; here, the highest K value was 5.9 (the solution then contained 80% PYR), while the averaged K value was 2.0. Experiments in PYR/4MP mixtures (Fig. 1c) demonstrated that the selectivity behaviour of H depended upon the concentrations of each guest species present: curiously, PYR was preferred when it was present in low concentrations only. The highest selectivity coefficient was 2.4, in favour of PYR, which was calculated from the experiment in which 20% of PYR was present. The 2MP/3MP experiments (Fig. 1d) showed 2MP to be constantly the preferred guest species: K was highest when the solution contained 20% 2MP (6.4), with the averaged K value for this set of experiments being 4.4. The behaviour of H in 4MP/2MP mixtures (Fig. 1e), on the other hand, depended once more on the amounts of the two guest species present. Here, 2MP was preferred when it was present in the greater amount. However, in a 20:80 4MP/2MP mixture, the host was, for all intents and purposes, unselective. The highest selectivity coefficient was 2.4, in favour of 4MP, and this was calculated for the solution containing 60% 4MP. Finally, in 4MP/3MP solutions (Fig. 1f), H was always selective for 4MP, but only moderately so. The highest K value was 3.3 (for a solution with 40% 4MP) and the averaged K value was 1.9.

According to Nassimbeni et al. [31], K is required to be 10 or greater for efficient separations of these binary mixtures. It is therefore unfortunate to conclude that H will not be a suitable host candidate for the separation/purification of any of these binary mixtures since all K values were significantly lower than 10; Table 3 summarises these values (in favour of the guest in brackets) for each of the experiments applicable to this investigation.

Table 3 Selectivity coefficients for the data points contained in the selectivity profiles
Full size table

Notably, the unsubstituted host derivative, 1,4-bis(diphenylhydroxymethyl)benzene, was demonstrated to possess the ability to separate very many of these mixtures, as was reported on a prior occasion [19]. The mere substitution of the p-hydrogen in that work with a methyl group in the present investigation deleteriously affected the selectivity behaviour of the host species.

Single crystal X-ray diffraction (SCXRD) experiments

Table 4 contains the relevant crystallographic data for the SCXRD experiments that were carried out on all four single solvent complexes of H with these pyridyl guest solvents. The PYR- and 4MP-containing inclusion complexes were both solved in the monoclinic crystal system and space groups P21/c and P21/n, respectively, while those containing 2MP and 3MP crystallized in the triclinic crystal system and space group P\(\overline{1}\).

The host packing in each complex was unique as is evident from their very different unit cell parameters. One of the two PYR molecules in the asymmetric unit cell of H·3(PYR) was disordered around an inversion centre, while 2MP in its complex with H also experienced some minor disorder.

The host molecule possessed a centrosymmetric geometry around an inversion point that was positioned in the centre of the central phenyl ring.

The Supplementary Information section contains ORTEP plots to demonstrate the atom numbering used in all four complexes (Figures S1a‒d).

Table 4 The relevant crystallographic data for H·3(PYR), H·2(2MP), H·2(3MP) and H·4MP
Full size table

Figure 2a − d (left hand side) depict the unit cells for each of the four complexes, while the void diagrams (right hand side) are also included here. All four complexes housed their guest molecules in infinite channels with varying degrees of constriction.

All of the pyridyl guest species experienced classical hydrogen bonding interactions with the host molecule involving the guest nitrogen atom and the host hydroxyl hydrogen atom (Fig. 3). Table 5 contains a summary of the distances and angles associated with these interactions, which were all very short (1.98‒2.04 Å, H···A; 2.809(2) − 2.873(3) Å, D···A) (D, donor; A, acceptor) with angles 166‒174°. Only the ordered PYR molecule was involved in an interaction of this type with H (O1‒H1···N4) while both disorder components of 2MP experienced hydrogen bonding with this host compound (O1‒H1···N4 and O1‒H1···N5). A statistical analysis of these hydrogen bond distances revealed that the hydrogen bond of one of the preferred guest species, PYR, with the host molecule was much shorter (2.809(2) Å) than these bonds with less favoured 2MP (2.873(3), 2.850(10) Å) and 3MP (2.824(2) Å) (O1‒H1···N4). These observations may explain the preference of H for PYR in the guest competition experiments. However, 4MP, also a preferred guest compound, did not experience a shorter hydrogen bond with the host molecule (2.861(3) Å) (O1‒H1···N6) compared with less favoured 3MP (2.824(2) Å), and this distance was also comparable when considering this interaction in the 2MP-containing complex.

Table 5 The host···guest hydrogen bonding parameters in the four complexes with Ha
Full size table

Additionally, the complex H·4MP was the only one to experience intermolecular hydrogen bonds between two host molecules. Applicable distance parameters were 2.05 and 2.8750(19) Å, while the angle was 169°. This may also be observed in Fig. 3d.

Fig. 2
figure 2

Unit cells (left) and voids (right) in (a) H·3(PYR) [100], (b) H·2(2MP) [100], (c) H·2(3MP) [010] and (d) H·4MP [010]; guest molecules are in spacefill and host compounds in capped stick representation; voids are coloured yellow/orange

Full size image
Fig. 3
figure 3

Host···guest hydrogen bonds retain the guest molecules in the (a) H·3(PYR), (b) H·2(2MP), (c) H·2(3MP) and (d) H·4MP complexes; the 4MP-containing complex, additionally, experienced an intermolecular host···host hydrogen bond

Full size image

The only significant π···π contacts in the four complexes were those between adjacent guest molecules in their channels in the complex H·2(2MP), as illustrated by means of Fig. 4 (here, the host molecules have been deleted). Distances were 3.713(2) and 3.896(2) Å with respective slippages of 1.540 and 1.782 Å. These contact types involved only the one guest disorder component (the other component was hence also deleted in Fig. 4), while the distance between the centres of gravity between two molecules of the second guest disorder component and between these centres of the first and second guest disorder components were too long (5.350 and 4.535 Å) to be meaningful. These contact types were not evident in the remaining three inclusion compounds.

Fig. 4
figure 4

The π···π stacking interactions in H·2(2MP) involving only the one guest disorder component in channels within the complex

Full size image

Upon further inspection, C‒H···π interactions were also observed between the host and guest species in H·2(2MP) and H·2(3MP) (the less preferred guest species) (these contact types could not be identified in the complexes containing PYR and 4MP). In H·2(2MP), a guest aromatic ring hydrogen atom of one guest disorder component (the component that did not experience π···π stacking contacts) interacted with the centre of gravity of one aromatic ring moiety of the host compound, with H···π, C···π and C‒H···π parameters of 2.75 Å, 3.606(12) Å and 151° (C54‒H54···Cg1); in the H·2(3MP) complex, on the other hand, it was the meta hydrogen atom of a free aromatic ring of the host compound that interacted in this manner with the centre of gravity of the 3MP guest species (2.82 Å, 3.603(2) Å and 140°) (C13‒H13···Cg4). Figure 5a and b are illustrations of these contacts.

Fig. 5
figure 5

The (a) (guest)C‒H···π(host) and (b) (host)C‒H···π(guest) interactions in the 2MP- and 3MP-containing complexes with H

Full size image

In the complex H·2(2MP) were also observed two short stabilizing interactions of this type between neighbouring guest molecules, and applicable parameters were 2.51 Å, 3.42(2) Å, 155° and 2.67 Å, 3.65(2) Å, 179°.

Subsequently were considered Hirshfeld surface analyses to quantify the interactions between atoms of the guest and host molecules. The G···H/H···G interactions that exist between atoms in complexes may be quantified effectively using these analyses. More specifically, the (guest)N···H(host) interactions were investigated (Table 6). The three-dimensional surfaces that were generated around the guest molecules are provided in Fig. 6a‒d (where the red areas indicate the strong hydrogen bonding interactions), while the two-dimensional plots obtained from these are illustrated in Fig. 7a‒d29,30 showing all guest···host interactions (left) and only the (guest)N···H(host) interactions (right, the sharp spikes). (In Figs. 6a and b and 7a and b are depicted the relevant illustrations for one guest component only, as examples; however, all components were considered when the percentage (guest)N···H(host) interactions were quantified.)

Fig. 6
figure 6

The three-dimensional Hirshfeld surfaces in (a) H·3(PYR) (showing just the ordered guest component), (b) H·2(2MP) (showing just one of the disorder guest components), (c) H·2(3MP) and (d) H1·4MP; the (guest)N···H − O(host) hydrogen bonding interactions are displayed by means of the red areas

Full size image
Fig. 7
figure 7

Hirshfeld two-dimensional fingerprint plots for all guest···host interactions (left) and for the (guest)N···H(host) interactions (right) for the (a) H·3(PYR), (b) H·2(2MP), (c) H·2(3MP) and (d) H·4MP complexes; in (a) and (b) are provided figures for only one guest component as examples though all were considered in the quantification calculations

Full size image
Table 6 The percentage (guest)N···H(host) atom···atom interactions in the four complexes with H
Full size table

From Table 6, the percentage of these (guest)N···H(host) interatomic interactions increased in the order H·2(2MP) (7.2%, disorder guest component 1) < H·4MP (7.4%) < H·2(2MP) (8.8%, disorder guest component 2) < H·2(3MP) (9.5%) < H·3(PYR) (10.1, ordered guest component) < H·3(PYR) (10.8, disorder guest components 1 and 2). While these data concur with the observation that H favoured PYR, its predilection for 4MP (relative to 2MP and 3MP) is less clear when considering these Hirshfeld surface analyses.

By employing thermoanalytical experiments, the relative thermal stabilities of the four complexes were subsequently assessed.

Thermal analyses

The thermograms (overlaid DSC (blue, endo down), TG (red) and DTG (green)) after thermal analyses on the four complexes are provided in the Supplementary Information (Figures S2a‒d), while the more pertinent data from these are summarised in Table 7. Ton is the onset temperature for the guest release event and serves as a measure of the relative thermal stabilities of these complexes, and Tp is the peak temperature for the endotherm representing the host melt process and is the point at which this event is most rapid.

Table 7 Temperature and mass loss data from thermal analyses on the four pyridyl complexes with Ha
Full size table

Except for the 4MP-containing inclusion compound (Figure S2d), each complex released its guest species in, broadly speaking, a single step (Figures S2a‒c) (this process for H·4MP was characterised by two steps). Ton ranged between 43.4 and 69.9 °C (Table 7), and the measured and expected mass loss data were in close correlation (30.9, 26.6, 25.7, 16.2 and 32.3, 27.2, 27.2, 15.8 for H·3(PYR), H·2(2MP), H·2(3MP) and H·4MP, respectively, Table 7). These data explain the preference of H for 4MP since this complex was more stable (Ton 69.9 °C) than those complexes containing 2MP (51.4 °C) and 3MP (49.6 °C), but do not elucidate the reason for the observation that PYR was also favoured by H; H·3(PYR) was the least stable of the four complexes (Ton 43.4 °C). Finally, the endotherms representing the host melt event had peak temperatures that ranged between 197.3 and 198.6 °C.

Conclusions

The wheel-and-axle host compound, 1,4-phenylene-bis(di-p-tolylmethanol) H, was demonstrated to have inclusion ability for PYR, 2MP, 3MP and 4MP, forming 1:3, 1:2, 1:2 and 1:1 H: G complexes, respectively. In the presence of these mixed guest solvents, H showed a significant preference for PYR and 4MP. However, it was revealed that this host compound, unfortunately, would not be a successful candidate for the effective separation/purification, through host-guest chemistry strategies, of any of the mixed pyridyl systems considered in this work since the calculated K values were significantly less than 10 in each experiment (the highest K value was only 6.4). SCXRD data showed that all of the guest molecules experienced hydrogen bonding with the host compound. Furthermore, the hydrogen bond between H and PYR was pointedly shorter than those between this host molecule and less favoured 2MP and 3MP, possibly explaining the preference of H for PYR in the guest competition experiments. However, 4MP, another preferred guest compound, did not experience a shorter hydrogen bond with H than 3MP in H·2(3MP), and this distance was comparable to that in the 2MP-containing complex. Hirshfeld surfaces and two-dimensional fingerprint plots demonstrated that the preferred PYR guest compound experienced a greater percentage of (guest)N···H(host) interactions in its complex with H compared with the other complexes, while the affinity of the host compound for 4MP could not be explained in this manner. Unfortunately, thermal analyses showed that H·3(PYR) possessed the lowest thermal stability, despite this complex having a preferred guest compound; however, H·4MP, also having a favoured guest species, did have the highest thermal stability of the four complexes presented in this investigation. This work has shown that the unsubstituted host compound reported earlier, 1,4-bis(diphenylhydroxymethyl)benzene, preferred the same guest species as H (PYR and 4MP) when crystallized from the pyridyl mixtures but the extent of the selectivity in that investigation was significantly more enhanced than in the present work where the p-methyl substituted host compound was employed. Minor modifications to the molecular structure of the host compound may therefore affect its selectivity behaviour considerably when crystallized from mixtures of guests.