Received 10 August 2005
Structures of Na9[SO4]4X·2H2O2, where X = Cl or Br, in which the halide anions orchestrate extended orientation sequences of H2O2 solvate molecules
Detailed structures of nonasodium tetrakis(sulfate) chloride diperhydrate, Na9[SO4]4Cl·2H2O2, and its novel bromide analogue are compared. Hydrogen peroxide could not be resolved in a previously reported Na9[SO4]4Cl·2H2O2 substructure [tetragonal, P4/mnc; Adams et al. (1978), J. Chem. Soc. Chem. Commun. p. 288; Adams & Pritchard (1978), Acta Cryst. B34, 1428-1432]. However, on lowering the symmetry to P4/n, and using reflection data based on full unit-cells, the H2O2 solvate can be clearly seen. Although H2O2 molecules are not directly bonded to the halide anions, they exert considerable influence on the eight sodium cations that constitute each halide's coordination shell so that H2O2 ordering can be linked to halide dimensions.
The title compounds belong to a class of solids in which hydrogen peroxide simultaneously coordinates to alkali metal cations and hydrogen bonds to oxy-acid anions. These include commercially important sodium percarbonate, Na2(CO3)·1.5H2O2 (Pritchard & Islam, 2003), and the extensively studied alkali metal oxalates M2(C2O4)·H2O2 (M = Li, Na, K and Rb; Pedersen, 1969; Pedersen & Pedersen, 1964; Pedersen & Kvick, 1989, 1990; Adams et al., 1980a,b). A previous crystallographic investigation of Na9[SO4]4Cl·2H2O2 yielded a subcell structure (Adams et al., 1978; Adams & Pritchard, 1978) in which the Na+, Cl- and SO42- ions can be seen, but in which the H2O2 site is obscured by disorder. This led to continued speculation (Adams et al., 1981; Cosgrove & Jones, 1998a,b,c) about the role of H2O2 in the architecture of this exceptionally stable compound (Kao Soap Co. Ltd & Nippon Peroxide Co. Ltd, 1975).
The current investigation was initiated in order to address this last point and extended to include the bromide analogue, which was discovered during preliminary crystallization studies.
The title compounds were prepared by dissolving sodium sulfate and sodium halide in 4:1 molar ratios in 30% w/w hydrogen peroxide and leaving the solution to evaporate over 48 h.
All measurements were carried out using a Nonius Kappa-CCD diffractometer with graphite-monochromated Mo K radiation ( = 0.71073 Å). Details of cell parameters, data collection and refinement are summarized in Table 1,1 together with a listing of the software employed.
Systematic absences and statistical tests clearly indicate the space group P4/n for both the chloride and bromide structures, even though the diffraction patterns have very low intensity in regions that are not governed by the subcell.
The structures were solved by direct methods and refined with all data on F2. A weighting scheme based on P = [F2o + 2F2c]/3 was employed in order to reduce statistical bias (Wilson, 1976).
It did not prove necessary to restrain or constrain the refinements in any way, despite the pseudo-symmetric appearance of the structures.
Na9[SO4]4X·2H2O2, where X = Cl or Br, crystallize as colourless squares when sodium sulfate and the appropriate sodium halide are dissolved in 30% w/w aqueous hydrogen peroxide and allowed to evaporate to dryness. Once formed the crystals are stable under ambient conditions.
Fig. 1 shows the asymmetric units of the title chloride and bromide structures as well as the previously reported chloride substructure. It clearly illustrates how the resolution of the H2O2 molecules has dramatically increased the number of parameters that are needed to describe the structures. Interestingly, on application of the space-group symmetry the resulting packing diagrams, shown in Fig. 2, are almost identical except for the H2O2 sites. The H2O2 orientation sequence leads to the chloride a and b axes being double those of the bromide and 81/2 those of the chloride substructure.
| || Figure 1 |
Asymmetric unit of Na9[SO4]4Cl·2H2O2 (top), its bromide analogue (bottom left) and subcell structure (bottom right)
| || Figure 2 |
Perspective view of half a unit cell of Na9[SO4]4Cl·2H2O2 (c/2) on to the ab plane (top), bromide analogue (bottom left) and subcell structure (bottom right). The sublattice, defined by distorted Na+ cubes, has been highlighted in each case. Na+ ions are represented by filled ellipses and the X- ions by empty ellipses.
In order to answer the question of why the chloride and bromide structures have different H2O2 orientation sequences, it is useful to focus on the distorted cubes with corners defined by six-coordinate Na+ that form a sublattice (4 × 4 × 2 cubes for the bromide and 8 × 8 × 2 for the chloride). Each cube houses an SO42-, X-, or Na+ ion, or an H2O2 molecule. Sulfate anions form adjacent cubes from an eight-coordinate shell around each central Na+ ion (Fig. 5). The SO42- and H2O2 cubes each form stacks, generated by non-crystallographic glides down c. Those cubes containing central Na+ and X- ions also stack down c, alternating their Na+ and X- contents. Importantly, the halide and hydrogen peroxide cubes share edges parallel to c (Fig. 2). Also, each hydrogen peroxide molecule is not located centrally within its cube, but straddles the top (i.e. perpendicular to the c axis) square face where it coordinates to four sodium cations and hydrogen bonds to two sulfate anions (Fig. 3, Table 2). The side of the Na+ square over which the hydrogen bonding is directed distends and impinges on the corners of two neighbouring halide-containing cubes, distorting the halide environments. Travelling along an H2O2 stack, i.e. along c, the direction of the hydrogen bonding is reversed at each level and hence the side of the square that is elongated, and is also reversed (Fig. 3). This has two repercussions on the shape of the halide-bearing sodium cubes: firstly, if a top corner is pushed in, the corner directly below is not; secondly, only half the corners of any cube can be pushed in. The four unique chloride ions and their surrounding sodium cubes are shown in Fig. 4 and, given the above conditions, represent all the possible distortions. In contrast, only distortions of the type seen around Cl2, which has C4 symmetry, and Cl4 with S4 symmetry, are seen in the bromide structure. A comparison of the Na-X bond lengths from these two sites, presented in Table 3, show that the larger bromide anion is able to interact more effectively with all eight Na+ cations than the smaller chloride anion in the C4 site. The difference between the bromide and chloride structure seems to hinge on the larger halide's tendency to promote C4 coordination, which would, however, destroy the crystal's tetragonal symmetry if used exclusively. The combination of C4 and S4 sites seen in the bromide enable it to retain tetragonal symmetry, whilst doubling the occurrence of what is, presumably, a favourable halide environment for the larger bromide anion.
| || Figure 3 |
H2O2 environment in Na9[SO4]4X·2H2O2. Projections down (a) a and (b) c. The c axis is directed up the page in the top view.
| || Figure 4 |
The four Cl- environments in Na9[SO4]4Cl·2H2O2 viewed down c. Referring to how many Na+ are pushed in on the top and bottom faces of the cube, the four configurations are 2-2, 0-4, 1-3 and diagonal 2-2.
Unlike sodium percarbonate (Pritchard & Islam, 2003), the current structures show no evidence of disordered H2O2. In sodium percarbonate dynamic H2O2 disorder becomes complete above 240 K, however, no disorder is observed in Na9[SO4]4Cl·2H2O2, even when the temperature is raised to 300 K.
Although the sulfate anions in both structures conform to the expected tetrahedral geometry they all display minor systematic deviations, which are related to crystal packing interactions.
In the chloride the 32 crystallographically unique S-O bonds can be divided into two groups with those involved in hydrogen bonding to H2O2 being longer [1.488 (3)-1.492 (3) Å] than the remainder [1.456 (3)-1.484 (3) Å]. A similar picture is seen in the bromide where the hydrogen-bonding S-O bonds of 1.492 (2) and 1.494 (2) Å are clearly distinguished from the shorter non-hydrogen bonding variety of 1.467 (2)-1.479 (2) Å.
The O-S-O angles do not deviate substantially from the expected tetrahedral value, falling in the range 108.1 (2)-110.6 (2)°, however, there is a demarcation within this group with sulfate O atoms that participate in the eight-fold coordination of sodium cations (Fig. 5) subtending the smaller O-S-O bond angles [108.1 (2)-108.7 (2)°]. An identical situation arises in the bromide, where the O-S-O angles range from 108.5 (1) to 110.1 (1)°, but those involved in the eightfold coordination of sodium are both 108.5 (1)°.
| || Figure 5 |
Eight-fold coordination shell of central sodium cation involving four SO42- anions.
The peroxide O-O bond lengths are presented in Table 2 and show that the chloride values of 1.459 (3)-1.466 (3) Å, average 1.464 Å, are in good agreement with the single bromide O-O bond of 1.458 (2) Å. All these bonds are slightly shorter than an equivalent bond in sodium percarbonate, which was determined to be 1.4785 (8) Å at 150 K, but are well within the range, 1.439 (15)-1.509 (7) Å, defined by alkali-metal oxalate monoperhydrates.
HOOH torsion angles vary from -98 (4) to -114 (3)° in the chloride, but their average, -108°, is a good match for the bromide value of 106 (3)°. All these values coincide with the staggered minimum energy conformation that was identified from gas-phase spectroscopic measurements (Hunt et al., 1965).
The O-H bonds fall in the range 0.82 (5)-1.03 (7) Å in the chloride and 0.85 (2), 0.86 (4) Å in the bromide, showing good agreement with the analogous bonds in the oxalate perhydrates [0.83 (10)-1.0117 (5) Å].2 Also at 97 (3)-105 (3)° in the chloride and 102 (3)° in the bromide the OOH angles show excellent agreement with the oxalate values of 97 (3)-104 (5)°.
Despite the rather elaborate pattern of H2O2 orientations within the above structures, each H2O2 site is identical (or nearly identical) as well as optimal in terms of H2O2 conformation, coordination and hydrogen bonding. In contrast to the alkali metal oxalate perhydrates, the title perhydrates do not form analogous hydrates. H2O2 must template these structures in a very specific way as crystallization from water yields a mixture of sodium halide and sodium sulfate. This is somewhat surprising, given that both sodium sulfate fluoride [Kogarkoite, Na3(SO4)F]; Fanfani et al., 1980] and sodium sulfate fluoride chloride [Na6(SO4)2FCl, sulphohalite; Sakamoto, 1968] are known. In these structures the halide is octahedrally coordinated by sodium cations, matching its geometry in NaF or NaCl. Also, as each sulfate coordinates to 12 sodium cations (three per oxygen), an octahedral geometry is maintained around each cation. The sodium octahedra that surround each halide share faces with the 12-pointed polyhedra that encase the sulfate anions, a truncated trigonal bipyramid for Kogarkoite and truncated cube for sulphohalite. In sulphohalite the halide octahedra share corners to form a three-dimensional orthogonal grid with alternating F and Cl anions so that the unit-cell dimension in this cubic structure corresponds directly to the sum of the bond distances in the Na-Cl-Na-F-Na sequence. If fluoride is replaced with chloride the unit cell and, consequently, the sodium polyhedra around the sulfate would expand, making it harder for the sulfate to span all 12 cations. This does not happen and, given the scenario of a system where full coordination of the cations becomes difficult due to steric effects, it is not surprising that hydrogen peroxide, with its excess of lone pairs, is able to create a niche for itself.
Initially, interest in perhydrates centred on their commercial application as bleaches, but more recent research has employed them as a convenient and safe method of introducing anhydrous hydrogen peroxide to chemical reactions (Jones, 1999). This work included extensive studies on their use in the presence of bromide ions to oxidize and brominate substituted toluenes (Jones et al., 1996) and suggests that it would be worthwhile to test Na9[SO4]4Br·2H2O2 as a combined H2O2/Br- source.
We would like to thank the EPSRC for purchasing the Nonius Kappa CCD diffractometer.
Adams, J. M. & Pritchard, R. G. (1978). Acta Cryst. B34, 1428-1432.
Adams, J. M., Pritchard, R. G. & Thomas, J. M. (1978). J. Chem. Soc. Chem. Commun. p. 288.
Adams, J. M., Ramdas, V. & Hewat, A. W. (1980a). Acta Cryst. B36, 570-574.
Adams, J. M., Ramdas, V. & Hewat, A.W. (1980b). Acta Cryst. B36, 1096-1098.
Adams, J. M., Ramdas, V. & Hewat, A.W. (1981). Acta Cryst. B37, 915-917.
Blessing, R. H. (1995). Acta Cryst. A51, 33-37.
Blessing, R. H. (1997). J. Appl. Cryst. 30, 421-426.
Cosgrove, S. D. & Jones, W. (1998a). J. Mater. Chem. 8, 413-417.
Cosgrove, S. D. & Jones, W. (1998b). J. Mater. Chem. 8, 419-424.
Cosgrove, S. D. & Jones, W. (1998c). J. Mater. Chem. 8, 1911-1915.
Fanfani, L., Giuseppetti, G., Tadini, C. & Zanazzi, P. F. (1980). Mineral. Mag. 43, 753-759.
Farrugia, L. J. (1997). J. Appl. Cryst. 30, 565.
Farrugia, L. J. (1999). J. Appl. Cryst. 32, 837-838.
Hunt, R. H., Leacock, A., Peters, C. W. & Hecht, K. T. (1965). J. Chem. Phys. 42, 1931-1946.
Jones, C. W. (1999). Application of Hydrogen Peroxide and Derivatives. Cambridge: The Royal Society of Chemistry.
Jones, C. W., Hackett, A., Pattinson, A. I., Johnstone, A. & Wilson, S. L. (1996). J. Chem. Res. pp. 438-439.
Kao Soap Co. Ltd & Nippon Peroxide Co. Ltd (1975). German Patent No. 2 530 539, filed 9 July 1975.
Nonius (1997). Kappa-CCD, Windows 3.11 Version. Nonius BV, Delft, The Netherlands.
Nonius (1998). COLLECT, edited by R. Hooft. Nonius BV, Delft, The Netherlands.
Otwinowski, Z. & Minor, W. (1997). Methods in Enzymology, Vol. 276, Macromolecular Crystallography, Part A, edited by C. W. Carter Jr & R. M. Sweet, pp. 307-326. New York: Academic Press.
Pedersen, B. F. (1969). Acta Chem. Scand. 23, 1871-1877.
Pedersen, B. F. & Kvick, A. (1989). Acta Cryst. C45, 1724-1727.
Pedersen, B. F. & Kvick, A. (1990). Acta Cryst. C46, 21-23.
Pedersen, B. F. & Pedersen, B. (1964). Acta Chem. Scand. 18, 1454-1468.
Pritchard, R. G. & Islam, E. (2003). Acta Cryst. B59, 596-605.
Sakamoto, Y. (1968). J. Sci. Hiroshima University Ser. A Mathemat. Phys. Chem. 32, 101-108.
Sheldrick, G. M. (1997). SHELXL97 and SHELXS97. University of Göttingen, Germany.
Wilson, A. J. C. (1976). Acta Cryst. A32, 994-996.