An unexpected network in guanidinium rhodizonate

Abstract

The highly symmetrical rhodizonate ion, C₆O₆²⁻, would appear to be an ideal candidate to form double hydrogen bonds to six guanidinium cations as part of an infinite (3,6)-connected, 2D network of formula [C(NH₂)₃]₂[C₆O₆], in which each guanidium cation would link to three rhodizonate anions. As anticipated, the crystal structure reveals that the rhodizonate and the guanidinium ions serve as 6-connecting and 3-connecting nodes respectively; however, instead of a 2D network, a 3D network is formed. The 3D net is one that can be expanded or contracted by concerted rotation around connections lying in one particular direction. The relationship between the anticipated 2D net and the actual 3D net is discussed.

---

Introduction

The guanidinium cation is internally very well disposed to associate with three oxy-anions by three pairs of strong hydrogen bonds as in I. Ward and co-workers have made elegant use of this hydrogen bonding motif in an extensive family of guanidinium sulfonates in which not only does the guanidinium component act as a 3-connector as in I, but so also does the sulfonate to generate 2D sheet structures with the (6,3) topology (or hexagonal grid topology).^1–4 These guanidinium sulfonates provide outstanding examples of true crystal engineering of network structures because they contain a common, predictable yet pliable underlying hexagonal grid sheet structure for a wide range of R in the RSO₃⁻ component. We have embarked on an exploration of the sorts of structures and topologies that can be generated as a result of the association represented in I, in the course of which a number of guanidine-based families with highly symmetrical 3D structures have been discovered. In an extensive series of carbonate-bridged coordination polymers the guanidinium cation plays the role seen in I (Z = C), promoting thereby the generation of highly symmetrical metal/carbonate networks with cubic symmetry and sodalite-like topology.^5,6 Cubic symmetry and (10,3)-a topology are observed in a family of hydrogen bonded frameworks of composition [C(NH₂)₃][N(CH₃)₄][XO₄] (X = S, Cr or Mo) in which, again, the guanidinium ion plays the 3-connecting role seen in I.⁷ A series of borates of composition [C(NH₂)₃]₄[B(OCH₃)₄]₃X (X = a range of anions) have structures with cubic symmetry and boracite-like topology in which, yet again, the guanidinium ion acts as in I.⁸

In the work reported here we explored the possibility that the potentially highly symmetrical oxy-anion C₆O₆²⁻, derived from rhodizonic acid, might be able to associate with six guanidinium cations in the manner represented in Fig. 1a, whilst the guanidinium cation might associate with three C₆O₆²⁻ units, once again as in I (Z = CC of C₆O₆) leading to the 2D network shown. The essence of the connectivity underlying the network is shown in Fig. 1b. The internal geometries and symmetries of the two components in the hypothetical sheet appear to complement each other beautifully. We also note that this type of 2D network is apparent in rubidium rhodizonate,⁹ where Rb₂C₆O₆ sheets stack on top of each other, but is not present in the structure of potassium rhodizonate.¹⁰

Fig. 1 a) The hypothetical [C(NH₂)₃]₂[C₆O₆] sheet structure. The sheet can be considered to consist of strips, of composition [C(NH₂)₃]₂[C₆O₆], one pink and two blue shown here, each strip being attached to two neighbours along its edges. b) The (3,6)-connected topology underlying the structure.

Results and discussion

Aqueous solutions containing potassium rhodizonate and guanidinium chloride at room temperature deposit black needles of composition [C(NH₂)₃]₂[C₆O₆] whose structure was determined by single crystal X-ray diffraction. The C₆O₆²⁻ component does indeed participate in six pairs of hydrogen bonds, playing a 6-connecting role closely similar to that represented in Fig. 1a, and the guanidinium ion does indeed participate in three pairs of hydrogen bonds acting as a 3-connecting node as in I, but the network formed has a 3D connectivity, not the 2D connectivity in Fig. 1. All C₆O₆²⁻ units are equivalent, each being hydrogen bonded as in Fig. 2a to six guanidinium units. Four of the guanidinium units (labelled A in Fig. 2a) are very close to coplanar with the C₆O₆²⁻ unit, but the other two (labelled B in Fig. 2a) are very significantly twisted out of the plane, as can be seen. The three C₆O₆²⁻ units associated with each guanidinium cation, all of which are equivalent, are shown in Fig. 2b; as can be seen, two of the C₆O₆²⁻ units are very close to coplanar with the guanidinium cation but the third is twisted out of the plane.

Fig. 2 a) A C₆O₆²⁻ anion hydrogen bonded to six guanidinium cations. Cations A are close to coplanar with the C₆O₆²⁻ unit and cations B are significantly twisted out of the plane. b) A guanidinium cation hydrogen bonded to three C₆O₆²⁻ anions, two of which are close to coplanar with the guanidinium unit, the third being significantly twisted out of the plane. Click /ej/ce/2005/b513300a/2b.htm to access a 3D image of Fig. 2b.

The extended 3D structure is conveniently envisaged in terms of planar strips of composition [C(NH₂)₃]₂[C₆O₆], one of which is shown in Fig. 3a. Strips identical to those seen in the real structure can be discerned in the imaginary sheet structure shown in Fig. 1, where three adjacent ones are highlighted. All strips in the real structure are identical, half of them lying parallel with each other, the other half, again parallel with each other being, inclined at an angle of approximately 42° to the first. Each strip is linked on both of its edges to an infinite number of inclined strips, as shown in part in Fig. 3b. For clarity Fig. 3b shows only one strip of one orientation (strip X in Fig. 3b) and two others (strips Y and Z) equally inclined to the first. Strip X uses one of its guanidinium units to link to strip Y, then uses a C₆O₆ unit to make the link to the adjacent inclined strip (strip Z).

Fig. 3 a) One of the [C(NH₂)₃]₂[C₆O₆] strips in the structure of guanidinium rhodizonate. b) One strip of one orientation (strip X) with two attached inclined strips (Y and Z).

The topology of the underlying 3D (3,6)-connected net, namely (4¹8²)(4²8¹¹10²), is represented in Fig. 4. The relationship of the actual 3D net to the hypothetical 2D net can be clearly appreciated in terms of the arrangements of the planar strips as follows. In the hypothetical 2D structure each strip is attached to only two neighbours, which are parallel to the first and are attached along the entire edges of the strips by an infinite number of links, as can be seen in Fig. 1. In the observed 3D structure each strip is attached, on both its edges, to an infinite number of inclined strips by only a single link for each attached strip. The (4¹8²)(4²8¹¹10²) net is one of those interesting articulated systems in which the whole net can be expanded or contracted by concerted rotation of one set of strips relative to the other set around certain links in the structure. Thus the observed configuration, with one set of strips oriented at approximately 42° to the other set, as seen in Fig. 3b and 4, can in principle be converted to the most open, most symmetrical geometrical arrangement of the net, seen in Fig. 5, by concerted rotation of one set of strips relative to the other set around the links between strips, until the two sets are orthogonal. At the other extreme of this imaginary process, where the rotation is in the opposite direction and where the two sets of strips ultimately become parallel, the 3D net collapses and condenses into the 2D net seen in Fig. 1b. Throughout this entire range of configurations of the (4¹8²)(4²8¹¹10²) net the 120° angles at the 3-connecting nodes and the 60° angles at the 6-connecting nodes are maintained, without any introduction of angle strain. We are not aware of this intrinsically very simple net and its articulated, expandable and contractable nature having been discussed previously nor are we aware of any previously reported real chemical examples.

Fig. 4 a) The (3,6)-connected topology, (4¹8²)(4²8¹¹10²), underlying the structure of guanidinium rhodizonate, with the geometry as observed. The 3-connecting nodes represent the guanidinium units and the 6-connecting nodes the C₆O₆ units. b) The same observed net viewed edge on to both sets of strips.

Fig. 5 a) The (4¹8²)(4²8¹¹10²) net in its geometrically most symmetrical, most open configuration in which the strips are orthogonal. b) The same (4¹8²)(4²8¹¹10²) net viewed edge on to both sets of strips.

In the observed structure the angle of inclination of one half of the strips relative to the other half is presumably that affording the most comfortable face-to-face contact between neighbouring strips of the same orientation. As a consequence, the C₆O₆²⁻ units are found in face-to-face columns, shown in Fig. 6a. The separation between the average planes of the C₆O₆²⁻ units is 3.33 Å. A view of two neighbouring strips seen perpendicular to them is shown in Fig. 6b, which reveals how the individual components are laterally displaced relative to their neighbours. This displacement is such that a nitrogen atom of one guanidinium cation is located almost exactly vertically above the carbon atom of its neighbour (C⋯N 3.42(2) Å)—a close contact for species of the same charge.

Fig. 6 a) Face-to-face stacks of parallel strips. b) A view of two neighbouring face-to-face strips (one blue, one red) seen perpendicular to the strips revealing the degree of offset of the C₆O₆²⁻ units and the guanidinium units.

The black appearance of the solid is consistent with some degree of electronic communication along the length of the column resulting from the overlap of the π electronic clouds of adjacent C₆O₆²⁻ units which may give rise to interesting physical properties.

A possible reason why the hypothetical sheet structure shown in Fig. 1 is not observed is that, although the sheet in isolation may be internally stable, it is unable to stack in an energetically acceptable manner with others: since guanidinium cations are twice as numerous as C₆O₆²⁻ anions, the stacking may lead to unacceptable guanidinium–guanidinium repulsive interactions. However, in contradiction to this notion, we note firstly that close pairs of guanidinium cations are observed with a C⋯C separation of as little as 3.2 Å in the metal–carbonate sodalite-like nets referred to earlier⁵ and secondly that the guanidinium cations, as considered above, in any case make close contacts with each other. Furthermore, given the sheet structure of rubidium rhodizonate in which Rb cations form cationic piles it is even more surprising that we don't observe the same topology.

An interesting aspect of guanidinium rhodizonate is that almost all the crystals we observed formed in the precipitation from an aqueous solution are somewhat curved, resembling fine dark eye lashes. This led to considerable difficulty in finding a suitable single crystal for diffraction measurements. While the curved nature of these crystals is something that we have not encountered before, we note in a recent paper by Desiraju and co-workers¹¹ that certain types of organic crystals may be easily bent, particularly ones in which a mixture of strong and weak intermolecular interactions operate. A feature shared by many of the crystals considered by Desiraju is that one of the crystallographic axes is particularly short, less than 4 Å. The guanidinium rhodizonate crystal falls in to this category with a cell length of 3.641(5) Å.

Conclusions

The results described here provide yet another example of an extended network in which the building blocks behave more or less as expected (guanidinum acting as a 3-connecting node and rhodizonate acting as a 6-connecting node) but which adopts a 3D arrangement that would never have been predicted by us, nor by many others we suspect—a 3D arrangement, furthermore, which is simple and of considerable structural and topological interest.

Experimental

Synthesis

Crystals of guanidinium rhodizonate precipitated from an aqueous solution containing 0.043 g of guanidnium chloride and 0.053 g of potassium rhodizonate in 80 mL of water. Yield for [C(NH₂)₃]₂[C₆O₆] 0.015 g, 25%; elemental analysis (%) calc. for C₈H₁₂N₆O₆ C 33.3, H 4.2, N 29.2; found C 33.5, H 4.3, N 28.8%.

Crystallography

Crystal data and refinement details: C₈H₁₂N₆O₆, M = 288.24, T = 293 K, monoclinic, space group P2₁/c, a = 3.641(5), b = 9.363(14), c = 17.51(3) Å, β = 92.50(2)°, V = 596.5(2) ų, Z = 2, μ = 0.138 mm⁻¹, 772 unique reflections (R_int = 0.2651), wR2 (F² all data) = 0.2572, R1 [I > 2σ(I)] = 0.0986.

Data were measured on a Bruker CCD diffractometer fitted with Mo-Kα radiation. Structures were solved using direct methods (SHELXTL V5.1¹²) and refined using a full matrix least squares procedure based on F² (SHELX-97¹³). As indicated above, the unusual curved nature of the crystals made it difficult to find a crystal suitable for single crystal X-ray measurements. The crystal used was small and weakly diffracting resulting in elevated agreement values. Despite the difficulties associated with the weak data, the non-hydrogen atoms within the structure are very well-defined.

CCDC reference number 284214. For crystallographic data in CIF or other electronic format see DOI: 10.1039/b513300a

References

1. K. T. Holman, A. M. Pivovar and M. D. Ward, Science, 2001, 294, 1907.
2. K. T. Holman, S. M. Martin, D. P. Parker and M. D. Ward, J. Am. Chem. Soc., 2001, 123, 4421.
3. V. A. Russell and M. D. Ward, Chem. Mater., 1996, 8, 1654.
4. K. T. Holman, A. M. Pivovar, J. A. Swift and M. D. Ward, Acc. Chem. Res., 2001, 34, 107.
5. B. F. Abrahams, M. G. Haywood, R. Robson and D. A. Slizys, Angew. Chem., Int. Ed., 2003, 42, 1112.
6. B. F. Abrahams, A. Hawley, M. G. Haywood, T. A. Hudson, R. Robson and D. A. Slizys, J. Am. Chem. Soc., 2004, 126, 2894.
7. B. F. Abrahams, M. G. Haywood, T. A. Hudson and R. Robson, Angew. Chem., Int. Ed., 2004, 43, 6157.
8. B. F. Abrahams, M. G. Haywood and R. Robson, J. Am. Chem. Soc., 2005, 126, 816.
9. D. Braga, G. Cojazzi, L. Maini and F. Grepioni, New J. Chem., 2001, 25, 1221.
10. J. A. Cowan and J. A. K. Howard, Acta Crystallogr., Sect. E, 2004, 60, m511.
11. C. M. Reddy, R. C. Gundakaram, S. Basavoju, M. T. Kirchner, K. A. Padmanabhan and G. R. Desiraju, Chem. Commun., 2005, 3945.
12. G. M. Sheldrick, SHELXTL V5.1, Software Reference Manual, Bruker AXS, Inc., Madison, WI, USA, 1997.
13. G. M. Sheldrick, SHELX-97, Program for Crystal Structure Analysis, University of Göttingen, Germany, 1997.

---