Redetermination at 180 K of a layered lanthanide–organic framework

The asymmetric unit of the title compound, poly[(μ4-{[bis(hydrogen phosphonatomethyl)azaniumyl]methyl}phosphonato)lanthanum(III)], [La(C3H9NO9P3)]n, comprises an La3+ center and a H3nmp3− anion (where H3nmp3− is a residue of partially deprotonated nitrilotris(methylenephosphonic acid), namely {[bis(hydrogen phosphonatomethyl)azaniumyl]methyl}phosphonate). This study concerns a structural redetermination using single-crystal X-ray diffraction data, collected at the low temperature of 180 K, of a recently investigated material whose structural details have been proposed from powder X-ray diffraction studies [Silva et al. (2011 ▶). J. Am. Chem. Soc. 133, 15120–15138]. The main difference between the two models rests on the position of the H atoms. While two H atoms were modeled as attached to the same phosphonate group in the powder determination, in the current model, the same H atoms are instead distributed among two of these groups. The sample studied was an inversion twin.


Related literature
. For singlecrystal structural studies on MOFs having residues of (carboxymethyl)iminodi(methylphosphonic acid), see: Tang et al. (2006). For a description of the graph-set notation for hydrogen-bonded aggregates, see: Grell et al. (1999). For a description of the Flack parameter, see: Flack (1983).
The title material, [La(H 3 nmp)] (1) [where H 3 nmp 3is a residue of partially deprotonated nitrilotris(methylenephosphonic acid)], was recently isolated for the first time as microcrystalline powders which prevented a priori a straightforward structural elucidation using single-crystal X-ray diffraction studies (Silva et al., 2011). Structural details were, ultimately, unveiled using laboratory powder X-ray diffraction studies (PXRD) at ambient temperature. Indeed, materials belonging to this class of compounds are usually isolated as microcrystalline powders, as found using (carboxymethyl)iminodi(methylphosphonic acid) (Cunha-Silva, Ananias et al., 2009;Cunha-Silva, Lima et al., 2009) and also nitrilotris-(methylenephosphonic acid) (H 6 nmp) (Cunha-Silva et al., 2007;Silva et al., 2011). A search in the literature reveals a sole publication containing two single-crystal structural determinations of MOFs combining residues of (carboxymethyl)iminodi(methylphosphonic acid) and rare-earth elements (Tang et al., 2006). To the best of our knowledge the structural determination reported in the communication is the first based on single-crystal data for materials combining residues of H 6 nmp and rare-earth elements.
Changes in the synthetic route allowed us to obtain single crystals of 1 which were used for a detailed single-crystal Xray diffraction study. We note that differences from the original synthetic procedure are solely based on two essential features: (i) the metal precursor, which we have changed from LaCl 3 . 7H 2 O to La 2 O 3 ; (ii) the heating method, we now employ typical static conditions for the hydrothermal synthesis, instead of the two previously used techniques, a dynamic (with constant rotation) hydrothermal synthesis and microwave heating.
The asymmetric unit of the title compound (see Scheme and Figure 1) comprises a La 3+ metal center and a whole H 3 nmp 3anion. The single La 3+ center is nine-coordinated, {LaO 9 }, to a total of seven phosphonate groups arising from four symmetry-related H 3 nmp 3anionic ligands, with the coordination polyhedron resembling a highly distorted tricapped trigonal prism. Conversely, the H 3 nmp 3anion coordinates to a total of four symmetry-related La 3+ metal centers, with such connectivity leading to the formation of a two-dimensional coordination polymer perpendicular to the [010] direction of the unit cell.
The crystal structure unveiled from single-crystal X-ray diffraction resembles that previously described by us and based on powder X-ray diffraction data (Silva et al., 2011): (i) both structures were solved in the Pca2 1 orthorhombic space group; (ii) despite the differences in temperature from the two data sets, the unit cell parameters are very similar; (iii) the coordinates of the non-hydrogen atoms in the two models are highly superimposable with the differences on spatial positions being smaller than ca 0.23 Å ( Figure 2). For example, the La-O distances range from 2.466 (6) to 2.916 (6) Å in the single-crystal structural determination (Table 1) and from 2.487 (12) to 2.932 (11) Å in the powder model.
The main difference between the two models resides on the position of the hydrogen atoms. While two H atoms were modeled as attached to the same phosphonate group in the powder determination, in the current model the same hydrogen atoms are instead distributed among two of such groups. In the powder determination the location of the hydrogen atoms was inferred from NMR data of similar compounds (Cunha-Silva et al., 2007). In opposition, in the present single-crystal determination the observed P-O distances were used for the location of the same hydrogen atoms. The hydrogen bonding network present in the crystal is, thus, slightly distinct from that suggested by the previous powder X-ray studies.
N1-H1 interacts with two neighboring phosphonate groups coordinated to the metal center (O1 and O8), in a typical bifurcated motif. The O5, O6 and O9 oxygen atoms belonging to the protonated phosphonate groups (P2 and P3) are engaged in strong O-H···O hydrogen bonds (d D···A below ca 2.50 Å) forming a discrete chain represented by the graph set motif D 2 1 (4) (Grell et al., 1999). Individual two-dimensional layers close pack along the [010] direction of the unit cell as depicted in Figure 3. We note the absence of strong supramolecular interactions between adjacent layers and only weak C-H···O contacts (not shown) are present.

Experimental
Chemicals have been purchased from commercial sources and were used as received without further purification.
A reactive mixture containing nitrilotris(methylenephosphonic acid) (H 6 nmp, 0.26 g, 0.87 mmol, Fluka, 97%) and La 2 O 3 (0.14 g, 0.43 mmol, Inframat Advanced Materials, 99.995%) in ca 10 g of distilled water (molar ratios of about 2: 1: 1300) was stirred thoroughly in open air (ambient temperature) for 5 minutes. The resulting homogeneous suspension was transferred into an adapted teflon-lined Parr Instruments reaction vessel (autoclave with internal volume of ca 24 ml) which was then placed inside a preheated oven at 165 °C. The reaction took place under static conditions over a period of 72 h.
The isolated material consisted systematically of physical mixtures composed of the desired material (the twodimensional MOF structure) alongside with other products. It was possible, however, to isolate from these mixtures a crystal suitable for single-crystal X-ray diffraction data collection (Figure 4).
The reaction conditions highlighted above were fine tuned in order to find the optimal parameters which allowed the The crystal selected for data collection was found to be twinned by inversion and at the last stages of the refinement procedure the TWIN instruction was used alongside with one BASF (Flack) parameter wich refined to 0.44 (4) (Flack, 1983). A total of 1229 Friedel pairs have been measured and have been used as independent data during the refinement procedure.
The highest peak (3.85 e.A -3 ) is located at 1.36 Å from O2, which is in the middle of the bond O2-La1. The deepest hole (-1.65 e.A -3 ) is located at 0.77 Å from La1. Attempts to correct these anomalies proved to be unsuccessful.

Figure 1
Asymmetric unit of the title compound showing all non-H atoms represented as displacement ellipsoids drawn at the 50% probability level and H atoms as small spheres with arbitrary radius. The coordination sphere of the La1 center has been completed for clarity and the atomic labeling is provided for all non-H atoms. Symmetry transformations used to generate equivalent atoms: (i) 1/2-x, y, 1/2+z, (ii) -1/2+x, 1-y, z, (iii) 1-x, 1-y, 1/2+z.

poly[(µ 4 -{[bis(hydrogen phosphonatomethyl)azaniumyl]methyl}phosphonato)lanthanum(III)]
Crystal data 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.