Dimeric C–H⋯N interactions and the crystal engineering of new inclusion host molecules

---

Abstract

Single crystal X-ray studies of (6)·(tetrahydrofuran)₂ and (6)·(benzene)_1.5 reveal the presence of two distinct types of dimeric edge–edge C–H⋯N packing motifs. All of the racemic dibromides 2–4 and 6 form clathrate compounds where their opposite enantiomers are joined edge–edge by means of C–H⋯N dimers. Hence this interaction is robust and of considerable value in crystal engineering. Modification of the benchmark centrosymmetric aryl–H⋯N dimer can result, however, in more complex modifications which better suit a given host–guest combination. Therefore a range of different C–H⋯N dimers is observed, with the new motif present in (6)·(benzene)_1.5 completing a sub-set of these variants. This structural adaptability increases the likelihood of inclusion because competition between many weak intermolecular contacts is a central part of the host design philosophy.

---

Introduction

Recently we have described a synthetic strategy which allows simple access to a range of new lattice inclusion host molecules. For example, benzylic bromination of the non-host 1 affords the exo,exo-dibromide 2 which is an effective host. While 1 can pack efficiently by itself, this is more difficult for the substituted analogue 2 which therefore prefers to include guest molecules and yields clathrate compounds.^1,2

Host properties are also observed for 3 where the quinoline rings have been replaced by quinoxalines,^3,4 and 4 where the central portion of the structure has been changed to a bicyclo[3.3.0]octane unit.⁵ The inclusion compounds produced are noteworthy in resulting from competition between a number of potential weak intermolecular contacts such as aryl face–face, aryl edge–face, C–H⋯N, π⋯halogen, and halogen–halogen interactions.^6,7 It is the most effective combination of molecular size, shape, and these supramolecular synthons⁸ which leads to the observed crystal structure. Hence, although the hosts 2–4 have very similar molecular structures, the lattice packing modes of their inclusion compounds can vary considerably.

Results and discussion

The centrosymmetric aryl–H⋯N dimer interaction

Although aromatic hydrocarbons frequently pack in the solid state through aryl offset face–face and aryl edge–face interactions there is no simple mechanism allowing edge–edge assembly.^9,10 During the early stages of our investigation we found,² however, that both 1 and 2·CHCl₃ packed in an edge–edge manner by means of a dimeric Ar–H⋯N motif.¹¹ A Cambridge Structural Database¹² search revealed that this arrangement was widespread amongst nitrogen-containing heteroaromatic compounds.² In all of the 54 examples identified both Ar–H⋯N interactions were identical thus resulting in a symmetrical dimer and, in all but one case with a twofold axis, the arrangement was centrosymmetric. Fig. 1 illustrates this benchmark dimer for the case of diquinoline 1.²

Fig. 1 The archetypal edge–edge Ar–H⋯N dimer motif present in solid 1. Association of opposite enantiomers of 1 affords a centrosymmetric structure where the C–H⋯N distance is 3.55 Å. Weak hydrogen bonds are shown as black and white dashed lines. Colour code: C green, H light blue, N dark blue.

These observations are of significance for crystal engineering since the dimer provides a third aryl assembly mode for nitrogen heteroaromatic compounds, permits packing between opposite enantiomers in a racemic mixture (or between two achiral molecules), and because the interaction occurs frequently in the solid state.

While the favoured arrangement in simple cases is this centrosymmetric dimer, it is now becoming apparent that a range of structurally different dimeric C–H⋯N motifs can occur for molecules such as the dibromides 2–4 and 6 which have been deliberately designed to pack less efficiently. This paper explores these possibilities.

Aryl–H⋯N dimer interactions for the diquinolines 2 and 4

Whereas the centrosymmetric dimer occurs in solid 1, only modified versions have been observed for the dibromide host 2. The geometrical parameters of these, and the other variants of the C–H⋯N dimer discussed in this paper, are compared in Table 1. In the inclusion compound 2·CHCl₃ (Fig. 2) the centrosymmetric motif is present (C–H⋯N 3.65 Å) but, in addition, the second nitrogen atom of each host hydrogen bonds to a chloroform molecule (C–H⋯N 3.37 Å).¹³ The structure of the inclusion compound 2·(CH₃–CCl₃)₂ involves chains of 2 molecules linked by centrosymmetric dimers (C–H⋯N 3.59 Å).¹³ Both these dimers are described as being anti- since they are located distant from the bromine atom on the same edge of the central ring.

Fig. 2 The centrosymmetric edge–edge Ar–H⋯N dimer interaction, and the hydrogen bonded chloroform guests, present in the diquinoline inclusion compound 2·CHCl₃. The C–H⋯N distances are 3.65 and 3.37 Å respectively. Colour code: Cl orange, Br brown.

Table 1 Characteristics of dimeric C–H⋯N interactions present between opposite enantiomers

Compound	C–H⋯N/Å	C–H⋯N/Å	Interaction type	Angle/°	Centrosymmetric dimer?
a The angle is that between the C⋯N vector of the intermolecular contact and the normal to the plane of the aryl ring containing this carbon atom. A value of 90° would indicate a perfectly planar array. b The C–H⋯N angle.
1	3.55	2.56	Ar–H⋯N	84.2^a	Yes
2·(CHCl3)	3.65	2.67	Ar–H⋯N	89.0^a	Yes
2·(CH3–CCl3)2	3.59	2.60	Ar–H⋯N	88.4^a	Yes (chain)
(3)2·(C2H2Cl4)	3.79	3.00	Ar–H⋯N	57.9^a	No (chain)
	3.59	2.72	BrC–H⋯N	146^b	No (chain)
(3)2·(C4H8O)	3.59	2.67	Ar–H⋯N (no. 1)	68.2^a	No (chain)
	3.49	2.59	BrC–H⋯N (no. 1)	149^b	No (chain)
	3.50	2.59	Ar–H⋯N (no. 2)	70.6^a	No (chain)
	3.41	2.54	BrC–H⋯N (no. 2)	145^b	No (chain)
(4)2·(CH3–CCl3)	3.38	2.47	Ar–H⋯N	79.4^a	Yes
(4)2·(CHCl3)	3.47	2.51	Ar–H⋯N	57.4^a	No
	3.54	2.58	Ar–H⋯N	62.0^a	No
(6)·(C4H8O)2	3.75	2.76	Ar–H⋯N	83.0^a	Yes (chain)
(6)·(C6H6)1.5	3.51	2.67	BrC–H⋯N	142^b	Yes

---

The centrosymmetric aryl dimer is present in solid (4)₂·(CH₃–CCl₃) (C–H⋯N 3.38 Å), but the closely related inclusion compound (4)₂·(CHCl₃) involves a non-centrosymmetric dimer (C–H⋯N 3.47 and 3.54 Å). This, however, still joins opposite enantiomers of 4.⁵ Whereas methyl chloroform has approximate tetrahedral symmetry and is accommodated easily within the host lattice, the less symmetrical chloroform molecule fits less efficiently. Formation of the asymmetric dimer is one of several structural adjustments adopted to overcome this difficulty while retaining the same general type of packing.

C–H⋯N dimer interactions for the diquinoxaline 3

The situation is rather different for the dibromide host 3. In (3)·(CHCl₃)₂ alternate enantiomers are linked as chains by means of centrosymmetric dimers (C–H⋯N 3.88 Å).⁴ In addition, each of the two remaining host nitrogen atoms associates with a chloroform guest through a Cl⋯N interaction^14,15 (3.04 Å). The Ar–H⋯N dimer chooses to form on the same side (syn-) as the bromine substituent of the central bicyclic ring, instead of the alternative anti-arrangement.⁴ (The dimers involving compounds 2 and 4 must adopt the latter configuration because of their molecular structures.)

A quite different motif is present in the compounds (3)₂·(1,1,2,2-tetrachloroethane) and (3)₂·(tetrahydrofuran) where host molecules assemble as molecular walls using an alternative type of C–H⋯N dimer. Just one of the host aromatic wings is utilised but it subtends two C–H⋯N dimers. Each dimer still joins opposite enantiomers but is built from one Ar–H⋯N and one aliphatic BrC–H⋯N interaction (which is akin to the chloroform ClC–H⋯N hydrogen bonds illustrated in Fig. 2). For (3)₂·(C₂H₂Cl₄) (Fig. 3) these two C–H⋯N distances are 3.79 and 3.59 Å respectively. In this compound both dimers are identical but in (3)₂·(C₄H₈O) they have dissimilar lengths.⁴

Fig. 3 Part of the crystal structure of solid (3)₂·(1,1,2,2-tetrachloroethane) showing a molecular wall formed by linking the aromatic wings of alternating enantiomers of 3 by means of C–H⋯N dimers. These unsymmetrical edge–edge dimers consist of Ar–H⋯N and aliphatic BrC–H⋯N linkages with C–H⋯N values of 3.79 and 3.59 Å respectively.

C–H⋯N dimer interactions for the diquinoline 6

To explore further the role of edge–edge dimers in crystal packing the new diquinoline derivative 6 was targeted for synthesis. Friedländer condensation¹⁶ between o-aminobenzaldehyde¹⁷ and bicyclo[3.3.1]nonane-3,7-dione¹⁸ in methanol solution containing a small amount of sodium hydroxide afforded 5¹⁹ in 87% yield. None of the alternative isomer, which would be expected from the 2,7-bis(enolate) intermediate, could be detected. This is in marked contrast to the modified Friedländer procedure reported by Aguado et al. where this diketone was condensed with o-aminobenzonitrile derivatives using zinc chloride or aluminium chloride catalyst. Under these conditions substantial proportions of both cyclisation products were obtained.²⁰

Benzylic bromination of the non-host 5 using N-bromosuccinimide in CCl₄ produced the required dibromide 6²¹ in 75% yield. This dibromide showed the expected clathrate host inclusion properties. Crystallisation of 6 from tetrahydrofuran or benzene gave inclusion compounds suitable for single crystal X-ray analysis. Numerical details of the solution and refinement of the two structures are presented in Table 2.

Table 2 Numerical details of the resolution and refinement of the inclusion structures formed by 6^a

Parameter	(6)·(tetrahydrofuran)2	(6)·(benzene)1.5
a Click b108032f.txt for full crystallographic data (CCDC 170677 and 170676).
Formula	(C23H16Br2N2)·(C4H8O)2	(C23H16Br2N2)·(C6H6)1.5
M	624.4	597.4
Crystal system	Monoclinic	Monoclinic
Space group	C2/c	P21/c
a/Å	17.268(3)	10.152(5)
b/Å	10.807(1)	21.493(7)
c/Å	16.728(3)	13.124(5)
β/°	120.122(6)	110.70(2)
V/Å^3	2700.1(8)	2679(2)
Z	4	4
T/°C	21(1)	21(1)
D c/g cm^−3	1.54	1.48
Radiation, λ(MoKα)/Å	0.7107	0.7107
μ/mm^−1	3.00	3.02
Scan mode	θ/2θ	θ/2θ
2θmax/°	46	46
No. of intensity measurements	1861	3710
Criterion for observed reflection	I/σ(I) > 3	I/σ(I) > 3
No. of independent observed reflections	1301	2213
No. of reflections (m) in final refinement	1301	2213
No. of variables (n) in final refinement	153	139
R = Σ^m|ΔF|/Σ^m|Fo|	0.030	0.042
R w = [Σ^mw|ΔF|^2/Σ^mw|Fo|^2]^1/2	0.037	0.053
s = [Σ^mw|ΔF|^2/(m − n)]^1/2	1.25	1.75
Crystal decay (%)	None	36
Max., min. transmission coefficient	0.76, 0.64	0.63, 0.58
R for multiple measurements	0.013	0.024
Largest peak in final difference map/e Å^−3	0.58	0.83

---

The diquinoline 6 is isomeric with compound 2 but the positions of its nitrogen atoms are such that its centrosymmetric Ar–H⋯N dimer would have to be syn- to the bromine substituent. This is what is found in the structure of (6)·(tetrahydrofuran)₂ except that each host subtends two identical dimers resulting in chains of 6 molecules of alternating handedness (Fig. 4). The dimers have C–H⋯N and C–H⋯N distances of 2.76 and 3.75 Å respectively.

Fig. 4 Part of a chain of 6 molecules linked by centrosymmetric C–H⋯N dimer syn-interactions in (6)·(tetrahydrofuran)₂.

The structure of (6)·(tetrahydrofuran)₂ is illustrated in Fig. 5. Tetrahydrofuran guest molecules are located in pairs in the angular cleft of the host molecule. The oxygen atoms are directed outwards and each accepts a C–H⋯O interaction (3.38 Å) from an adjacent host molecule. In all, each guest is surrounded by three host molecules. Other notable intermolecular packing features are an offset aryl face–face interaction between pairs of host aromatic wings, and a short host–host Br⋯Br contact of 3.54 Å.

Fig. 5 Part of the lattice of (6)·(tetrahydrofuran)₂ showing the arrangement of host and guests. The layer lies along the ab cell diagonal. Colour code: guest O red and C yellow. Hydrogen atoms are omitted for clarity. Click image or 5.htm to access a 3D representation.

When 6 was crystallised from benzene the inclusion compound (6)·(benzene)_1.5 was produced. Its X-ray structure revealed the previously unobserved C–H⋯N dimer illustrated in Fig. 6. This motif links opposite enantiomers of 6 in a centrosymmetric manner through two identical aliphatic BrC–H⋯N interactions, and hence completes this sub-set of C–H⋯N dimer variants which are seen only amongst our host molecules (see Table 1).

Fig. 6 Centrosymmetric BrC–H⋯N dimer present in solid (6)·(benzene)_1.5. This new dimer has BrC–H⋯N and BrC–H⋯N distances of 2.67 and 3.51 Å respectively.

The structure of (6)·(benzene)_1.5, shown in Fig. 7, contains two independent benzene guests. One (coloured purple) is located at a centre of symmetry and makes an edge–face aromatic interaction with four surrounding host molecules. The second benzene guest (yellow) is located in channels in the crystal lattice where it takes part in one offset face–face interaction and two edge–face interactions with neighbouring host molecules. These aryl host–guest arrangements lessen the possibility of similar host–host interactions. Nonetheless there are two instances of offset aryl face–face contacts between host molecules in this structure. The shortest host–host Br⋯Br interaction is 3.98 Å.

Fig. 7 Part of the lattice structure of (6)·(benzene)_1.5 showing the host–guest arrangement. The benzenes located at centres of symmetry are coloured purple and those in the channels are coloured yellow. Hydrogen atoms are omitted for clarity. Click image or 7.htm to access a 3D representation.

Conclusions

The principal aim of our work has been to use crystal engineering concepts to develop a simple route to new organic hosts. This has been achieved with all four target dibromides 2–4 and 6 forming clathrate compounds. Furthermore, all these compounds assemble using C–H⋯N dimer motifs in their crystal packing. Hence, for appropriate compounds, the edge–edge dimer interaction represents a motif of much higher reliability than the single C–H⋯N interaction which does not appear to be a generally dependable one for crystal engineering.¹¹ Furthermore, in all cases discussed here, the dimer is formed between opposite enantiomers thus permitting efficient packing for a racemic mixture.

The above results demonstrate, however, that ‘awkward’ compounds have considerable leeway in being able to modify the benchmark centrosymmetric edge–edge dimer into alternative forms more suited to their special requirements. Although a considerable variety of these has been uncovered so far, the present position is unlikely to represent the full story. In the context of our host design this is a good situation. The concept is to present our potential hosts with a mixture of different weak interactions so that the best combination is chosen for any given host and guest. A wider range of possible interactions maximises the likelihood of molecular inclusion taking place, but reduces prediction of the exact lattice packing which will result.

Experimental

Structure determinations

Reflection data were measured with an Enraf-Nonius CAD-4 diffractometer in θ/2θ scan mode using graphite monochromated molybdenum radiation (λ⊕=⊕0.7107 Å). Data were corrected for absorption.²² Reflections with I > 3σ(I) were considered observed. The positions of all atoms in the asymmetric unit were determined by direct phasing (SIR92)²³ with hydrogen atoms included in calculated positions. There are two independent guests in the benzene compound, one of which is located on a centre of symmetry. Both molecules were refined as rigid groups.²⁴

Acknowledgements

We gratefully acknowledge financial support from the Australian Research Council.

Notes and references

1. C. E. Marjo, R. Bishop, D. C. Craig, A. O'Brien and M. L. Scudder, J. Chem. Soc., Chem. Commun., 1994, 2513.
2. C. E. Marjo, M. L. Scudder, D. C. Craig and R. Bishop, J. Chem. Soc., Perkin Trans. 1, 1997, 2099.
3. C. E. Marjo, R. Bishop, D. C. Craig and M. L. Scudder, Aust. J. Chem., 1996, 49, 337.
4. C. E. Marjo, R. Bishop, D. C. Craig and M. L. Scudder, Eur. J. Org. Chem., 2001, 863.
5. A. N. M. M. Rahman, R. Bishop, D. C. Craig and M. L. Scudder, Chem. Commun., 1999, 2389.
6. G. R. Desiraju, Crystal Engineering: The Design of Organic Solids, Materials Science Monographs No. 54, Elsevier, Amsterdam, 1989.
7. G. R. Desiraju and T. R. Steiner, The Weak Hydrogen Bond in Structural Chemistry and Biology, Oxford University Press, Oxford, 1999.
8. G. R. Desiraju, Angew. Chem., Int. Ed. Engl., 1995, 34, 2311.
9. A. Gavezzotti and G. R. Desiraju, Acta Crystallogr., Sect. B, 1988, 44, 427.
10. G. R. Desiraju and A. Gavezzotti, J. Chem. Soc., Chem. Commun., 1989, 621.
11. For a comprehensive discussion of the C–H⋯N contact and its properties, see ref. 7: especially ch. 2, pp. 29–121 and ch. 4, pp. 293–342.
12. F. H. Allen, J. E. Davies, J. J. Galloy, O. Johnson, O. Kennard, C. F. Macrae, E. M. Mitchell, G. F. Mitchell, J. M. Smith and D. G. Watson, J. Chem. Inf. Comput. Sci., 1991, 31, 187.
13. R. Bishop, C. E. Marjo and M. L. Scudder, Mol. Cryst. Liq. Cryst., 1998, 313, 75.
14. H. A. Bent, Chem. Rev., 1968, 68, 587.
15. D. S. Reddy, K. Paneerselvam, T. Pilati and G. R. Desiraju, J. Chem. Soc., Chem. Commun., 1993, 661.
16. C.-C. Cheng and S.-J. Yan, Org. React., 1982, 28, 37.
17. L. I. Smith and J. W. Opie, Org. Synth., 1955, 3, 56.
18. S. H. Bertz, J. Org. Chem., 1985, 50, 3585.
19. Compound 5: ¹³C NMR (CDCl₃): δ 28.8 (CH₂), 33.1 (CH), 44.3 (CH₂), 125.8 (CH), 127.1 (CH), 127.5 (C), 128.1 (CH), 129.0 (CH), 134.5 (C), 135.9 (CH), 147.2 (C), 156.3 (C).
20. F. Aguado, A. Badía, J. E. Baños, F. Bosch, C. Bozzo, P. Camps, J. Contreras, M. Dierssen, C. Escalano, D. M. Görbig, D. Muñoz-Torrero, M. D. Pujol, M. Simón, M. T. Vázquez and N. M. Vivas, Eur. J. Med. Chem., 1994, 29, 205.
21. Compound 6: ¹³C NMR (CDCl₃): δ 20.0 (CH₂), 41.3 (CH), 55.7 (CH), 127.3 (CH), 127.5 (CH), 128.2 (C), 129.0 (CH), 129.4 (C), 130.1 (CH), 137.9 (CH), 148.1 (C), 153.5 (C).
22. J. De Meulenaer and H. Tompa, Acta Crystallogr., 1965, 19, 1014.
23. A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M. C. Burla, G. Polidori and M. Camalli, SIR92, A Program for Automatic Solution of Crystal Structures by Direct Methods, J. Appl. Crystallogr., 1994, 27, 435.
24. A. D. Rae, RAELS, A Comprehensive Constrained Least Squares Refinement Program, University of New South Wales, 1996.

---