Introduction

Anisole (ANI) and the bromoanisole isomers (2-BA, 3-BA and 4-BA, Scheme 1) find important applications in perfumes, flavouring agents, pharmaceuticals and other organic synthesis procedures [1,2,3,4,5]. The BA isomers may be synthesized by means of the bromination of ANI, where a mixture of 2-BA and 4-BA usually results (together with a small quantity of 3-BA and, also, unreacted ANI) [6, 7]. These isomers may also be prepared by employing O-methylation procedures on the respective bromophenol starting compounds [8, 9]. Additionally, bromination of ANI can be achieved using N-bromophthalimide in the presence of catalytic amounts of mercuric acetate with acetic acid as the solvent [10].

The boiling points of 2-BA, 3-BA and 4-BA are very similar (216, 211 and 223 °C, correspondingly) and, as a consequence, their separation by means of fractional distillation protocols is extremely challenging. Therefore, host‒guest chemistry is an attractive alternative separation strategy for mixtures of these guest compounds.

The present work focusses on the roof-shaped compounds bearing the rigid 9,10-dihydro-9,10-ethanoanthracene framework, H1–H3, and investigates their host selectivity and thus their ability to serve as separatory tools for mixtures of 2-BA, 3-BA and 4-BA (for the sake of interest, ANI was also included in these investigations). These compounds were specifically selected as viable host candidates for these separations largely as a result of work carried out previously with the related host compound trans-9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboxylic acid (DED) (Scheme 1, H4) and these anisoles as guest species. In that investigation [11], DED was revealed to possess an extraordinary affinity for 4-BA (93.1‒98.6%) during crystallization procedures from these guest solutions. Furthermore, remarkable results were obtained not only in binary guest mixtures, but also those ternary and even quaternary in nature. Moreover, host compounds H1‒H3 were shown in the recent past to also possess discriminatory behaviour in various mixtures such as ethylbenzene/xylenes [12, 13], dichlorobenzenes [14], anisole/methylanisoles [15] and pyridine/methylpyridines [16] and, in many instances, the observed host selectivities for one particular guest in these mixtures were significantly enhanced and, oftentimes, it was concluded that these mixtures may be separated by means of host‒guest chemistry strategies when employing H1‒H3 as the applicable host compounds. This field of science offers a significantly greener separation protocol compared with other separatory methods since the host compound, after guest release, may be recycled and reused in the process.

We report now on the results so obtained and confirm that these have not been published on a prior occasion.

Scheme 1
scheme 1

Molecular structures of host compounds H1‒H4 and the potential guest molecules ANI and 2-, 3- and 4-BA

Full size image

Materials and methods

General

All starting and guest materials were purchased from Sigma Aldrich, South Africa, and were used as received.

1H-NMR experiments were carried out by means of a Bruker Ultrashield Plus 400 MHz spectrometer; CDCl3 was the deuterated solvent.

Gas chromatographic experiments were carried out by means of an Agilent Technologies 7890 A gas chromatograph equipped with an Agilent 5975 C VL mass spectrometer. An Agilent J&W Cyclosil-B column served as the stationary phase with helium gas being the mobile phase. The split ratio was 80:1. The metal plates of the inlet and detector were maintained at 220 and 300 °C, respectively. The column was initially heated and maintained at 50 °C. The first heating ramp of 15 °C·min‒1 then raised the column temperature to 150 °C. A second heating ramp (20 °C·min‒1) allowed the final column temperature to reach 170 °C (which was maintained for 1 min). The flow rate varied between 1 and 1.3 mL·min‒1. Dependent upon availability, a Young Lin YL6500 gas chromatograph coupled to a flame ionization detector was also employed. The column was as before, while both hydrogen gas (30 mL·min‒1) and air (300 mL‧min‒1) served as the mobile phase. The split ratio was, once more, 80:1. The inlet and detector metal plates were maintained at 200 and 300 °C, respectively. The column was heated from 50 °C to 200 °C by implementing a heating ramp of 20 °C·min‒1. This final temperature was held for 1 min, and the final run time was 8.5 min. The flow rate was 2 mL·min‒1. Finally, in all GC experiments, dichloromethane served as the dissolution solvent.

The single solvent complexes for the thermal experiments were recovered from the glass vials by means of suction filtration, washed with petroleum ether (bp 40–60 °C) and analysed without further manipulation. The appropriate instrument was a TA SDT Q600 Module system; data were analysed using TA Universal Analysis 2000 software. Samples were placed in open ceramic pans whilst an empty pan served as the reference. The purge gas was high purity nitrogen, and the samples were heated from 40 to 400 °C at a heating rate of 10 °C·min–1.

Synthesis of host compounds H1, H2 and H3

Host compounds H1‒H3 were prepared by considering previous literature reports [17,18,19].

Crystallization experiments of H1‒H3 from ANI and 2-, 3- and 4-BA

It was required at the outset to determine whether host compounds H1‒H3 possess host ability for each of ANI and 2-, 3- and 4-BA. To this end, the host compound was dissolved in an excess of each of these anisoles in glass vials which were subsequently left open to the ambient conditions. In this fashion, some solvent evaporated off and crystallization proceeded. (H1 and H2 required three to four months for crystallization to occur (if this occurred at all), while H3 crystallized rapidly, within one or two days.) The crystals were collected by suction filtration, washed with petroleum ether (bp 40–60 °C) and then analysed by means of 1H-NMR spectroscopy. This technique revealed whether complexation had been successful, in which instance the host: guest (H: G) ratio was calculated from the resultant spectra by considering the areas under the peaks of appropriate host and guest resonance signals.

Crystallization experiments of H3 from equimolar mixtures of ANI and 2-, 3- and 4-BA

Owing to the successful complexation of each of the guest anisoles using H3 as the host compound, an investigation of its host selectivity when presented with guest mixtures was subsequently undertaken. (Note that only the behaviour of H3 was investigated in mixed guests owing to its rapid crystallization from these solutions (one to two days) and the prolonged crystallization times (three to four months) required for H1 and H2 to form crystals and also since the single solvent experiments with these two host compounds mostly resulted in gels (no crystallization occurred).) Therefore, H3 (0.04 g) was dissolved in every combination of equimolar guest mixtures (combined amount 5 mmol) in glass vials. These were closed and stored at 4 °C. The crystals that formed in this way were treated as in the single solvent experiments, and analysis was then by means of 1H-NMR spectroscopy (to obtain the overall H: G ratios) and GC (to quantify the guest compounds in any mixed complexes that formed).

Crystallization of H3 from binary guest mixtures in which the molar amounts of each guest species was varied sequentially.

The host selectivity behaviour of H3 was also assessed in binary guest anisole mixtures (GA and GB) in which the molar amount of each guest was varied sequentially; applicable GA:GB molar ratios ranged between 80:20 and 20:80 mol%. (Once more, only H3 was investigated in these conditions for the reasons already described above.) This host species was crystallized from these binary guest solutions according to the method described for the experiments that employed equimolar guest mixtures (host and combined guest amounts also remained unchanged). Both the crystals that formed and the solution from which these formed were analysed by means of GC. A plot of the molar ratio of GA (or GB) in the crystals (Z) against GA (or GB) in the solution (X) therefore afforded selectivity plots which demonstrated how the selectivity behaviour of the host compound changed when the GA:GB molar ratios were varied. In order to measure this selectivity, the selectivity coefficient (K) was calculated in each instance by employing the equation KA: B = ZA/ZB × XB/XA, where XA + XB = 1 [20]. K = 1 represents an unselective host compound and this circumstance is represented by the straight-line diagonal plots that have been inserted into each of these selectivity profiles for facile comparison purposes with the experimentally obtained data points.

Results and discussion

Crystallization experiments of H1‒H3 from ANI and 2-, 3- and 4-BA

Table 1 contains the H: G ratios of complexes formed in the single solvent crystallization experiments of H1–H3 from ANI, 2-BA, 3-BA and 4-BA, as obtained from 1H-NMR experiments.

Table 1 The H: G ratios of complexes of H1–H3 when crystallized from ANI and the BA isomers
Full size table

Table 1 shows that attempts to crystallize H1 and H2 from 2-BA, 3-BA and 4-BA were unsuccessful, and gels remained in the glass vials. However, both H1 and H2 enclathrated ANI, forming 3:1 H: G complexes with this guest in both instances. In contrast, H3 successfully included all four guest compounds, with ratios of 1:1 (ANI and 3-BA), 2:1 (2-BA) and 1:2 (4-BA).

Due to the prolonged time required for H1 and H2 to form complexes with ANI and the emergence of gels in the remaining experiments, these host compounds were excluded from further investigation. H3, on the other hand, was subjected to further scrutiny due to its capability of forming complexes with each of the four guest compounds within a two-day timeframe.

Crystallization experiments of H3 from equimolar mixtures of ANI and 2-, 3- and 4-BA

The H: G and G: G ratios resulting from the equimolar binary, ternary and quaternary ANI and BA guest mixtures are presented in Table 2 after H3 was crystallized from these mixtures. The preferred guests in each of these guest competition experiments are indicated in red bold font face, and the percentage estimated standard deviations (%e.s.d.s) are provided in parentheses as a result of the experiments being carried out in duplicate, where possible, as evidence of replicability.

Table 2 The overall H: G and G: G ratios of complexes of H3 when crystallized from equimolar ANI and BA mixturesa,b
Full size table

Host compound H3 possessed selectivity for ANI in the binary crystallization experiments; this selectivity was near-complete (97.5%, in favour of ANI) in the ANI/2-BA experiment, and this preferential behaviour for ANI persisted in the ANI/3-BA (67.5%) and ANI/4-BA (90.7%) experiments. In the absence of ANI, 3-BA was then favoured in these binary mixtures; once more, the selectivity behaviour of H3 was near-complete for this guest compound in both instances (94.5% in 2-BA/3-BA and 90.4% in 3-BA/4-BA).

In the ternary and quaternary experiments, 3-BA emerged as the preferred guest, yielding crystals with up to 93.9% 3-BA in the 2-BA/3-BA/4-BA experiment. However, H3 crystallized with only moderate amounts of 3-BA in competition experiments with solutions ANI/2-BA/3-BA (57.5%), ANI/3-BA/4-BA (60.0%) and ANI/2-BA/3-BA/4-BA (57.1%). The addition of ANI to the 2-BA/4-BA experiment (ANI/2-BA/4-BA) produced a solid with up to 84.1% ANI. The overall H: G ratio remained 1:1 for all experiments (except in the 2-BA/4-BA experiment which did not crystallize), consistent with the ratios observed in the single solvent experiments with the preferred ANI and 3-BA guests (1:1 H: G ratio in each instance).

Unfortunately, therefore, H3 proved to be an unlikely host candidate for the separation of 2-BA/4-BA mixtures, but it proved to be an extremely capable host candidate for isolating ANI and 3-BA from various equimolar binary and ternary mixtures of these guest compounds.

Crystallization of H3 from binary guest mixtures in which the molar amounts of each guest species was varied sequentially.

Five selectivity profiles (Fig. 1a–e) for the binary guest experiments with H3 were constructed by plotting ZA (or ZB) against XA (or XB). Note that the 2-BA/4-BA selectivity profile could not be obtained here since these solutions furnished gels with the host compound, and crystallization was not successful.

Fig. 1
figure 1

Selectivity profiles of H3 when presented with the a) ANI/2-BA, b) ANI/3-BA, c) ANI/4-BA, d) 3-BA/2-BA and e) 3-BA/4-BA binary mixtures

Full size image

From these figures, it is evident that H3 was selective for either ANI (in ANI/2-BA and ANI/4-BA mixtures) (Fig. 1a and c) or 3-BA (in the 3-BA/2-BA and 3-BA/4-BA experiments) (Fig. 1d and e) across the concentration range. Nassimbeni and coworkers reported that, for feasible separation processes on an industrial platform, K values are required to be equal to or greater than 10 [21]. Therefore, due to the high selectivity coefficients calculated in the ANI/4-BA (Fig. 1c, Kave = 17.8), 3-BA/2-BA (Fig. 1d, Kave = 18.7) and 3-BA/4-BA (Fig. 1e, Kave = 21.5) experiments (where Kave is the average of the K values obtained for each data point in a set of data points), H3 may be considered for applications in the separation and isolation of either ANI or 3-BA from these mixtures. Furthermore, the host selectivity for ANI in ANI/2-BA mixtures, where the concentration of ANI was 40% or greater, was also extremely high (Fig. 1a): for example, the complex contained 95.9% ANI when the solution comprised only 41.0% of this guest solvent (K, here, was 34.0, and hence this separation would also be a feasible one). However, the selectivity of H3 was guest concentration-dependent when presented with ANI/3BA mixtures (Fig. 1b) (ANI was only favoured when the solution contained more than 50% ANI), and K values were low in these experiments (1.4–3.2).

In summary, H3 displayed a significant ability for the separation and isolation of either ANI from ANI/2-BA and ANI/4-BA mixtures or 3-BA from 3-BA/2-BA and 3-BA/4-BA mixtures.

Regrettably, SCXRD analyses were not possible on any of the H3 complexes with these anisoles owing to poor crystal quality, with the exception of the ANI-containing complex, which has been reported on an earlier and unrelated occasion [15]. (Note that the apohost (guest-free) crystal structure for H1 has been reported [22], while these for H2 and H3 could not be isolated without the presence of a guest species.)

Thermal analyses

Table 3 contains the thermal data for the four anisole complexes with H3 after heating these from approximately 40 to 400 °C at a rate of 10 °C⋅min‒1. The resulting thermogravimetric (TG), its derivative (DTG) and the differential scanning calorimetric (DSC) traces are provided in Fig. 2a–c for the inclusion compounds with 2-, 3- and 4-BA (these thermograms and associated data for the ANI-containing complex were reported on a prior occasion [15]).

Table 3 Thermal data for the H3·ANI, 2(H3)·2-BA, H3·3-BA and H3·2(4-BA) complexes
Full size table

From Table 2, the more thermally stable complexes (with the greater Ton values) were those containing the less favoured 2-BA and 4-BA guests (Ton 167.7 and 167.0 °C, respectively), while the favoured ANI and 3-BA guests were released at lower temperatures (94.7 and 139.1 °C). Therefore, curiously, the complexes with the more favoured guest species possessed lower thermal stabilities than the complexes with guests less preferred. The measured and expected mass losses of each of the complexes with 2-BA, 3-BA and 4-BA could not be measured since the guest release process ultimately coincided with the host melt/decomposition event. Unfortunately, therefore, these thermal experiments do not wholly explain the selectivity of H3 for ANI and 3-BA.

Fig. 2
figure 2

The DSC (blue), TG (red) and DTG (green) traces for a 2(H3)·2-BA, bH3·3-BA and cH3·2(4-BA) (these traces for the ANI-containing complex have been published before [15])

Full size image

Conclusions

Here, H3 was demonstrated to have the ability to form complexes with each of the anisole and bromoanisole guests, forming either 1:1 (ANI and 3-BA), 2:1 (2-BA) or 1:2 (4-BA) H: G complexes. H3 displayed a significant preference for both ANI and 3-BA as is evident from the equimolar binary guest competition experiments, where as much as 97.5% ANI (ANI/2-BA) and 94.5% 3-BA (2-BA/3-BA) were enclathrated. The equimolar quaternary guest mixture resulted in crystals containing only moderate amounts of both ANI (37.5%) and 3-BA (57.1%). Furthermore, this host compound also showed a remarkable affinity for either ANI or 3-BA in binary experiments where the molar amount of each guest present was varied. Here, H3 was able to selectively include either ANI or 3-BA in ANI/2-BA, ANI/4-BA, 3-BA/2-BA or 3-BA/4-BA mixtures; however, in the ANI/3-BA experiments, host selectivities were low. From many of these binary guest solvent experiments was calculated significant K values, alluding to the feasibility of using H3 for their separation, a greener separation methodology compared with tedious and energy intensive fractional distillations. Unfortunately, single crystal X-ray diffraction analyses were not possible on complexes with the bromoanisoles as a result of poor crystal quality. Finally, the relative thermal stabilities of the complexes from thermal analyses did not wholly explain the host selectivity of H3 for ANI and 3-BA.