Possible strong symmetric hydrogen bonding in disodium trihydrogen bis(2,2'-oxydiacetate) nitrate

The nature of short hydrogen-bonding interactions is still a subject of much interest and debate (Meot-Ner, 2005). It appears that, for O—H—O interactions where O O is less than about 2.50 Å, examples can be found of truly symmetric hydrogen bonds (Catti & Ferraris, 1976), most of which have crystallographic equivalence between donor–acceptor atoms. The title sodium compound, (I), displays just such a short and possibly symmetric hydrogen-bonding interaction.

In the title compound, 2Na + ÁC 8 H 11 O 10 À ÁNO 3 À , the Na I atom is heptacoordinate with an approximately pentagonal-bipyramidal geometry. A possible strong symmetric hydrogen bond, with the H atom located at an inversion centre and an OÁ Á ÁO distance of 2.450 (2) Å , is observed in the crystal structure.

Comment
The nature of short hydrogen-bonding interactions is still a subject of much interest and debate (Meot-Ner, 2005). It appears that, for O-H-O interactions where OÁ Á ÁO is less than about 2.50 Å , examples can be found of truly symmetric hydrogen bonds (Catti & Ferraris, 1976), most of which have crystallographic equivalence between donor-acceptor atoms. The title sodium compound, (I), displays just such a short and possibly symmetric hydrogen-bonding interaction.
The nitrate anion lies on the crystallographic twofold axis and links pairs of Na I ions in an anti-anti-1,3-bridging coordination mode. The carboxyl O atoms act in a bis--bridging capacity between Na I ions, forming the polymeric structure.
In addition to these ionic interactions, the crystal structure is also held together by a network of two types of hydrogen bonds. The first is formed between the O4-carboxyl group and the nitrate anion (Table 1). This hydrogen bond is asymmetric and non-linear. The nature of the second type of hydrogen bond is ambiguous. Certainly there exists a short hydrogenbond interaction between the O1-carboxyl group and its crystallographically equivalent group; the O1Á Á ÁO1 i distance of 2.450 (2) Å [symmetry code:(i) 3 2 À x, 3 2 À y, 1 À z] falls within the normal range for symmetric hydrogen bonds (Catti & Ferraris, 1976). A Fourier map section in the O1/C1/O2 plane (MAPVIEW; Farrugia, 1999) clearly indicates a peak of electron density centred on the crystallographic inversion ( Fig. 2). Two alternative structural models have been studied. Placing atom H1 on the inversion centre gives a symmetric structure. Full-matrix least-squares refinement converged to a stable solution which is reported here. The residual difference Fourier map has a largest peak and hole of 0.16 and À0.23 e Å À3 , respectively. The crystallographic symmetry constrains the hydrogen-bond angle to 180 and the O1-H1 distance to 1.23 Å . A second structural model is one with the H1-atom site half occupied and displaced from the inversion centre towards O1. Free refinement of the x, y, z and U iso parameters for the H1 atom (118 parameters in total) converged to give a sensible asymmetric hydrogen-bonding interaction; the crystallographic residuals are insignificantly different and the difference map shows essentially the same features. The limited data quality and resolution mean that we cannot unambiguously determine the nature of this hydrogenbonding interaction. The compound clearly merits further study, if only to resolve this issue. Despite this uncertainty in the H-atom position, such a linear hydrogen-bonding interaction linking two 2,2 0 -oxydiacetate molecules in a trans metal-organic papers The molecular structure and sodium environment, with 50% probability displacement ellipsoids and the atom-labelling scheme. Dashed lines indicate hydrogen bonds. [Symmetry codes:

Figure 2
Difference Fourier map from a refined model where atom H1 is absent. The O1/C1/O2 plane is shown with the crystallographic inversion in the centre of the plot.

Figure 3
Graph showing the distribution of C-OÁ Á ÁO-C torsion angles as a function of OÁ Á ÁO interatomic separations for short hydrogen-bonded interactions between two carboxylic acid groups. Data taken from 78 structures in the CSD (Version 5.25; Allen, 2002).
We also note that the first hydrogen-bonding pattern (between the carboxylic acid and the nitrate anion), although asymmetric, seems to be a strong and important structural motif. In nine out of 14 reported structures that contain both a carboxylic acid and a nitrate anion, the acid is hydrogen bonded to the nitrate, and none of the structures displays the common R 2 2 (8) dimeric carboxylic acid motif (see, for example, Sridhar et al., 2002).
Together, these two hydrogen-bonded motifs form a network. Using the nomenclature for graph theoretical analysis developed by Etter et al. (1990), we can describe this system as N 1 = D 2 2 (4)D; N 2 = C 3 3 (20), where the second-order net consists of infinite chains that zigzag through the crystal structure (Fig. 4).
All methylene H atoms were located in idealized positions and refined in riding mode. C-H distances were set at 0.97Å and U iso (H) values were constained to be 1.5U eq of the parent C atom. Both H atoms involved in hydrogen bonding were found in a Fourier difference map and were refined, subject only to the inversion centre constraint.
Data collection: MSC R-AXIS-II Control Software; cell refinement: DENZO (Otwinowski & Minor, 1997); data reduction: SCALEPACK (Otwinowski & Minor, 1997); program(s) used to solve structure: SHELXS86 (Robinson & Sheldrick, 1988); program(s) used to refine structure: SHELXL97 (Sheldrick, 1997) in WinGX (Farrugia, 1999); molecular graphics: DIAMOND (Brandenburg, 1999   Special details Geometry. All e.s.d.'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.'s are taken into account individually in the estimation of e.s.d.'s in distances, angles and torsion angles; correlations between e.s.d.'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.'s is used for estimating e.s.d.'s involving l.s. planes. Refinement. Refinement of F 2 against ALL reflections. The weighted R-factor wR and goodness of fit S are based on F 2 , conventional R-factors R are based on F, with F set to zero for negative F 2 . The threshold expression of F 2 > σ(F 2 ) is used only for calculating R-factors(gt) etc. and is not relevant to the choice of reflections for refinement. R-factors based on F 2 are statistically about twice as large as those based on F, and R-factors based on ALL data will be even larger.