Carbon dioxide entrapment in an organic molecular host

Abstract

4-Phenoxyphenol crystallises to yield discrete ∼60 ų cavities that are capable of enclathrating small solvent molecules; the cavities are capped by constricted 6-membered hydrogen-bonded rings and these potential apertures do not appear to facilitate gated porosity when the material is subjected to static CO₂ pressure.

---

One of the main objectives in studies of networked porous materials such as zeolites,¹ metal–organic frameworks (MOFs)² and covalent organic frameworks (COFs)³ is to synthesise new materials with large permanent pores that give rise to correspondingly high surface areas. Indeed, this is also an important goal for porous molecular materials, with the highest reported surface areas currently in the order of 3000 m² g⁻¹.⁴ However, it has been shown that molecular crystals without any permanent pores can also exhibit porous behaviour if the host molecules are able to behave in a concerted fashion in order to facilitate gated diffusion of guests.⁵ We are interested in probing the lower limits of non-gated porosity in organic molecular crystals. That is, when the host system is non-cooperative as a result of possessing structural features (e.g. stable packing arrangements or strong intermolecular interactions) that prohibit the dynamic behaviour required for a gate-opening mechanism, what is the smallest aperture through which gas molecules may pass? This question has been partially addressed by the formulation of kinetic diameters¹ for various guests, but a tantalising study of Dianin's compound has challenged this concept.

Dianin's compound (1, 4-p-hydroxyphenyl-2,2,4-trimethyl-chroman; Scheme 1) was first synthesised nearly a century ago⁶ and is an iconic example of an organic host system.⁷ The racemic mixture generally crystallises in the high symmetry space group R3̄, where six molecules are arranged about a site of 3̄ symmetry to form a hydrogen-bonded ring. This ⋯O–H⋯O–H hydrogen bonded ring is formed by the phenolic groups of three pairs of enantiomers, where alternating molecules extend towards opposites sides of the plane of the slightly distorted hexagonal ring (Fig. 1a). This hydrogen-bonded motif is well-established as the dominant structure-directing feature of 1 and it persists in a large number of its known clathrates, as well as in its apohost form.⁷ These hydrogen-bonded units stack in columns to produce hourglass shaped cavities capped at two ends by the hydrogen-bonded rings (Fig. 1c). These rings appear to serve as apertures that link successive cavities along [001] to form one dimensional channels.

Scheme 1 The molecular structures of Dianin's compound (1) and 4-phenoxyphenol (2).

Fig. 1 Projections along [001] of the hydrogen-bonded hexamers of (a) 1·vac·293K and (b) 2·vac·293K. Alternating molecules about the hydrogen-bonded ring are coloured blue (down) and green (up). The cavities formed by (c) 1 and (d) 2 viewed perpendicular to [001]. The transparent yellow surfaces define contact volumes of 226 and 68 ų, respectively (using a probe of radius 1.6 Å) while the purple surfaces represent the accessible surfaces for the smallest spheres that cannot pass through the hydrogen bonded rings (1.249 Å for 1 and 1.174 Å for 2).¹⁵ The molecules are shown in capped-stick representation, with oxygen atoms shown as spheres of arbitrary radius. Hydrogen-bonds are shown as fragmented red cylinders and all hydrogen atoms not engaged in hydrogen-bonding have been omitted for clarity.

In 1976 Barrer and Shanson⁸ described the porosity of Dianin's compound towards various gases such as argon, krypton, xenon, carbon dioxide and methane, stating that the “behaviour of Dianin's compound was in these respects the same as that of a zeolite … the framework of the host structure is however less rigid than that of a zeolite and guest molecules can penetrate into the cavities, even though wide windows giving access to these cavities do not exist”. Although this study predated the sudden surge of reports of synthetic non-zeolitic porous systems that started two decades later, it has largely been ignored, having accumulated only around 37 citations to date. Indeed, to the best of our knowledge, no follow-up study of this important observation has been undertaken as yet. There exist in the literature several analogues of Dianin's compound – compounds that form cyclic (O–H)₆ hydrogen-bonded motifs and that have been shown to enclathrate solvent guest molecules in their interstitial cavities. Examples include endo-4,endo-8-dimethylbicyclo[3.3.1]-nonane-endo-2,endo-6-diol,⁹ chiral propargylic alcohols,¹⁰ β-hydroquinone,¹¹ phenol¹² and 4-phenoxyphenol.¹³

As part of our ongoing studies of the solid-state supramolecular chemistry of Dianin's compound,¹⁴ we have extended our investigations to include 4-phenoxyphenol (2, see Scheme 1), which bears a strikingly close structural resemblance to 1 despite being achiral. Compound 2 also crystallises in the space group R3̄, with six molecules arranging themselves about a site of 3̄ symmetry to form a hydrogen-bonded ring (Fig. 1b). These units also stack to form cavities, but as a result of the decreased size and somewhat different shape of the molecules, the cavities are significantly smaller (Fig. 1d) relative to those of 1. A recent report¹³ on the enclathration of small solvents of crystallisation (i.e. acetonitrile and methanol) by 2 showed that larger solvents such as diethyl-ether are not included but rather yield the “empty” apohost form. The study also showed that methanol could only be removed from the cavity upon melting or sublimation of the host.

Table 1 Summary of relevant crystallographic data

Crystallisation conditions	Sublimed (vacuum)	Sublimed (vacuum)	Sublimed (200 mbar CO2)	Sublimed (2 bar CO2)	Supercritical CO2
a/Å	29.2571(16)	29.0785(18)	29.000(4)	28.906(10)	28.908(6)
c/Å	5.9378(3)	5.8540(4)	5.8505(8)	5.862(2)	5.8899(12)
Unit cell volume/Å^3	4401.7(4)	4286.7(5)	4260.9(10)	4242(3)	4262.6(15)
Temperature/K	293(2)	100(2)	100(2)	100(2)	100(2)
Electron count/e^−	4	2	7	16	23
%CO2	0	0	32	73	100

---

As an extension of that study we have investigated the permeability of 2 to various gases (carbon dioxide in particular) (Table 1). To this end, diffraction-quality single crystals of 2 were grown by subliming pure host material at 78 °C under vacuum. A suitable crystal was placed under static vacuum in an environmental gas cell¹⁶ and single-crystal diffraction (SCD) data were collected at 293 K. After elucidation of the structure (2·vac·293K), summing electron densities using the SQUEEZE routine of PLATON¹⁷ indicated the presence of only four electrons within the guest-accessible cavity. The electron count decreased to two electrons when intensity data were recollected at 100 K (2·vac·100K). Using widely accepted van der Waals radii¹⁸ and the equilibrium atomic coordinates of the structure determined at 293 K, we calculate¹⁵ that the largest sphere that can pass through the hydrogen bonded ring has a radius of 1.174 Å. Similar analysis of the cavity (using a probe of radius 1.60 Å) reveals a free-space volume of 68 ų for 2·vac·293K or 64 ų for 2·vac·100K. Considering the many examples of gate-opening mechanisms in molecular crystals, as well as the reported porosity of 1, we continued to investigate the possibility that 2 may exhibit gas permeability despite its very narrow “pore”. Gas sorption experiments were carried out gravimetrically at room temperature and pressures up to 20 bar with nitrogen, carbon dioxide, hydrogen and helium as potential sorbates. In all cases, uptake of these gases under the specified conditions (see ESI†) was negligible.

Materials with narrow pores that are seemingly inaccessible to gases such as hydrogen and nitrogen have often been shown to absorb CO₂ as a result of its high polarizability.¹⁹ To further test the permeability of 2 towards CO₂, the crystal in the environmental gas cell was exposed to ca. 10 bar of CO₂ for 20 hours, after which intensity data were recollected; no appreciable increase in electron density was detected within the cavity. In order to offset possible kinetic effects, the crystal was then kept under a constant atmosphere of 10 bar of CO₂ for a further 6 days after which redetermination of the structure still yielded no evidence of gas uptake. It therefore appears that, under the conditions investigated, the “empty” apohost crystals obtained by sublimation do not absorb CO₂ molecule into their cavities. Our study was then expanded in order to explore the converse – that if CO₂ is trapped during the crystal assembly process, as is the case for the growth of a solvate, it is not possible for the guest to escape spontaneously. 2 was sublimed under a CO₂ pressure of 200 mbar to produce needle-shaped crystals that were subjected to SCD analysis. The resulting electron count of 7 e⁻ is consistent with partially occupied cavities that each contain, on average, 0.32 molecules of carbon dioxide. Repetition of the experiment using a CO₂ pressure of 2 bar yielded 16 electrons per cavity, corresponding to an occupancy of 0.73 guest molecules per cavity. In both structures the CO₂ atoms were located in difference electron density maps and refined with their site occupancies constrained according to their respective electron counts. In order to obtain crystals with fully occupied cavities, we ultimately resorted to using supercritical carbon dioxide as the crystallisation solvent. Crystals suitable for SCD analysis (2·CO₂, Fig. 2) were grown by slow cooling of a solution of 2 in supercritical CO₂, initially at 220 bar and 40 °C; summation of the electron density within the confines of the cavity yielded 23 e⁻ (22 e⁻ expected). The asymmetric unit contains half of the CO₂ molecule, with the carbon atom situated on the 3̄ site at the centre of the cavity and the oxygen atom located along the 3-fold axis. In our model the unconstrained C=O bond lengths of the 0.73 and fully occupied structures are 1.148(5) and 1.149(2) Å, respectively; these values are in good agreement with the reported length of 1.155(1) Å in a structure of pure CO₂ determined at 150 K.²⁰ Using a probe of radius 1.6 Å, the guest-accessible volume in 2·CO₂ is calculated to be 77 ų, representing a ca. 13% increase in cavity volume relative to that in 2·vac·100K. There appear to be no strong intermolecular interactions between host and guest in 2·CO₂ and the closest contact present is that between the guest oxygen atom and the hydrogen-bonded ring of 2: O⋯O = 3.251(2) Å. Thermogravimetric analysis shows that desorption of CO₂ is concomitant with melting of the host (m.p. 84 °C).²¹ An initial weight loss of 3.24% is observed below 100 °C while a weight loss of 3.79% is expected for a 6 : 1 host : guest ratio.

Fig. 2 (a) Projection of the hydrogen-bonded ring of 2·CO₂ viewed along [001]. (b) Perspective view perpendicular to the stacking axis [001] displaying the guest-filled cavities of 2·CO₂. The host molecules are shown in the capped-stick metaphor and the guest in van der Waals representation. Hydrogen-bonds are shown as fragmented red cylinders and all hydrogen atoms not engaged in hydrogen-bonding have been omitted for clarity.

We have shown that 4-phenoxyphenol, which possesses pore windows that are narrower than the kinetic diameter of the smallest gas, does not appear to be permeable to any of the gases under the conditions investigated. Neither does it display any dynamic behaviour, as is occasionally observed in other organic molecular crystals, even to the highly polarizable CO₂ (which has previously been shown to diffuse through surprisingly small pores). We attribute this non-porosity and non-cooperativity within the crystals to the rigidity of the hexameric hydrogen-bonded pore that guards the entrance to the cavity. Owing to its lack of porosity, 2 has the ability to non-reversibly trap CO₂ guest molecules within its interstices when crystals are grown in the presence of a CO₂ rich medium. The problems of substantiating claims of porosity in molecular crystals have recently been discussed,²² but this work highlights the fact that the converse is also true: how do we unequivocally establish non-porosity in a molecular crystal with lattice voids? Is it possible to prove that a material is non-porous or is it merely a question of “guilty until proven innocent”?

Notes and references

1. D. W. Breck, Zeolite Molecular Sieves, John Wiley & Sons Inc., New York, 1974.
2. O. M. Yaghi, M. O'Keeffe, N. W. Ockwig, H. K. Chae, M. Eddaoudi and J. Kim, Nature, 2003, 423, 705.
3. A. P. Cote, A. I. Benin, N. W. Ockwig, M. O'Keeffe, A. J. Matzger and O. M. Yaghi, Science, 2005, 310, 1166.
4. M. Mastalerz and I. M. Oppel, Angew. Chem., Int. Ed., 2012, 51, 5252.
5. (a) J. L. Atwood, L. J. Barbour and A. Jerga, Science, 2002, 296, 2367; (b) J. L. Atwood, L. J. Barbour, A. Jerga and B. L. Schottel, Science, 2002, 298, 1000; (c) S. J. Dalgarno, P. K. Thallapally, L. J. Barbour and J. L. Atwood, Chem. Soc. Rev., 2007, 36, 236.
6. A. P. Dianin, J. Russ. Phys.-Chem. Soc., 1914, 31, 1310.
7. (a) W. Baker, A. J. Floyd, J. F. W. McOmie, G. Pope, A. S. Weaving and J. H. Wild, J. Am. Chem. Soc., 1956, 2010; (b) J. L. Flippen, J. Karle and I. L. Karle, J. Am. Chem. Soc., 1970, 92, 3749; (c) F. Imashiro, M. Yoshimura and T. Fujiwara, Acta Crystallogr., Sect. C, 1998, 54, 1357.
8. R. M. Barrer and V. H. Shanson, J. Chem. Soc., Chem. Commun., 1976, 333.
9. V. T. Nguyen, R. Bishop, I. Y. H. Chan, D. C. Craig and M. L. Scudder, CrystEngComm, 2008, 10, 1810.
10. M. Hyacinth, M. Chruszcz, K. S. Lee, M. Sabat, G. Gao and L. Pu, Angew. Chem., Int. Ed., 2006, 45, 5358.
11. T. C. W. Mak and B. R. F. Bracke, in Comprehensive Supramolecular Chemistry, ed. J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vögtle, Pergamon, Oxford, 1996, vol. 6, p. 23.
12. L. Mandelcorn, Chem. Rev., 1959, 59, 827.
13. L. H. Thomas, E. Cheung, A. O. F. Jones, A. A. Kallay, M.-H. Lemee-Cailleau, G. J. McIntyre and C. C. Wilson, Cryst. Growth Des., 2012, 12, 1746.
14. (a) E. J. C. de Vries, M. W. Bredenkamp, T. Jacobs and G. O. Lloyd, Acta Crystallogr., Sect. E, 2005, 61, O2871; (b) C. Esterhuysen, M. W. Bredenkamp and G. O. Lloyd, Acta Crystallogr., Sect. C, 2005, 61, O32; (c) G. O. Lloyd and M. W. Bredenkamp, Acta Crystallogr., Sect. E, 2005, 61, O1512; (d) G. O. Lloyd, M. W. Bredenkamp and L. J. Barbour, Chem. Commun., 2005, 4053; (e) T. Jacobs, G. O. Lloyd and M. W. Bredenkamp, Acta Crystallogr., Sect. E, 2006, 62, O4400; (f) G. O. Lloyd, J. Alen, T. Jacobs, M. Bredenkamp and E. J. C. de Vries, Acta Crystallogr., Sect. E, 2006, 62, O691; (g) T. Jacobs and M. W. Bredenkamp, Acta Crystallogr., Sect. E, 2007, 63, O4444; (h) T. Jacobs, M. W. Bredenkamp and E. J. C. de Vries, Acta Crystallogr., Sect. E, 2007, 63, 03736; (i) T. Jacobs, G. O. Lloyd, M. W. Bredenkamp and L. J. Barbour, Cryst. Growth Des., 2009, 9, 1284; (j) T. Jacobs, G. O. Lloyd, M. W. Bredenkamp and L. J. Barbour, CrystEngComm, 2009, 11, 1545; (k) T. Jacobs, M. W. Bredenkamp, P. H. Neethling, E. G. Rohwer and L. J. Barbour, Chem. Commun., 2010, 46, 8341.
15. M. L. Connolly, Science, 1983, 221, 709.
16. D. S. Yufit and J. A. K. Howard, J. Appl. Crystallogr., 2005, 38, 583.
17. A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7.
18. A. Bondi, J. Phys. Chem., 1964, 68, 441.
19. (a) M. P. Suh and D. H. Hong, Chem. Commun., 2012, 48, 9168; (b) J. Rabone, Y. F. Yue, S. Y. Chong, K. C. Stylianou, J. Bacsa, D. Bradshaw, G. R. Darling, N. G. Berry, Y. Z. Khimyak, A. Y. Ganin, P. Wiper, J. B. Claridge and M. J. Rosseinsky, Science, 2010, 329, 1053.
20. A. Simon and K. Peters, Acta Crystallogr., Sect. B, 1980, 36, 2750.
21. CRC Handbook of Chemistry and Physics, ed. D. R. Lide, 88th edn, CRC Press, 2007–2008.
22. L. J. Barbour, Chem. Commun., 2006, 1163.

---

Footnote

† Electronic supplementary information (ESI) available: Crystallographic data, thermogravimetric analysis, powder X-ray diffraction and gas sorption isotherms. CCDC 959280–959284. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c3cc46784h

---