neutron

The crystal structure of low temperature NaNiO 2 has been refined by Rietveld methods using powder X-ray diffraction and neutron scattering data. The starting model was based on parameters that had been obtained earlier by X-ray film methods. At room temperature NaNiO 2 is monoclinic, C2lm, a = 0.53192(2), £ = 0.28451(1), c = 0.55826(4) nm, 0= 110.449(2)°. NaNiO 2 has a layered structure. The Ni-0 layer is formed by edge sharing of Jahn-Teller elonganted NiO 6 octahedra with Ni-0 distances of 0.1911(2) nm and 0.2144(4) nm. The Na ions between these layers also exhibit a distorted octahedral coordination with Na-O distances of 0.2328(2) nm and 0.2369(4) nm. The final R values were R wp = 0.069, R,= 0.059, R exp = 0.059 for the neutron and R wp = 0.032, R, = 0.034, R exp = 0.017 for the X-ray data. Diffraction Data. [S0885-7156(97)00503-4]


I. INTRODUCTION
In the ternary system Na-Ni-0 the structures of only three compounds are known hitherto: Na 2 NiO 2 , a Naoxoniccolate(II), and Na 5 Ni0 4 were described by single crystal X-ray structure determinations (Zentgraf, 1980a;Zentgraf, 1980b). The structure of the monoclinic low temperature NaNiO 2 was first determined by Dyer et al. (1954) using estimated intensities from precession photographs of a single crystal. DTA experiments showed that NaNiO 2 transforms to a high temperature modification at about 220 °C; according to Debye-Scherrer photographs, this modification is rhombohedral and isotypic with LiNiO 2 and NaFeO 2 .
During our studies concerning the kinetics of NaNiO 2 hydrolysis under various conditions and crystallographic characterization of its products we felt that a reliable structure determination of monoclinic NaNiO 2 is required. In order to evaluate more accurate atomic positions, bond lengths and angles in NaNiO 2 , the structure was refined using X-ray and neutron scattering data of NaNiO 2 powder at room temperature.

II. EXPERIMENT A. Preparation
A powder sample of NaNiO 2 was prepared by the reaction of 25 g NiO and 25 g Na 2 O 2 in a Ni crucible (height 6 cm, diameter 5 cm) at 730 °C for 2 days. The hot mixture was poured on a Ni sheet; after cooling the powder was carefully washed with a small amount of water (in order to remove excess Na 2 O 2 ) in a pressure filtration apparatus, immediately dried in a stream of N 2 and stored over KOH. When the contact of the NaNiO 2 with water was only short no hydrolysis was detected by X-ray diffraction. Washing the product with absolute ethanol, as suggested by Glemser and Einerhand (1950), Dyer et al. (1954), Bade et al. (1965) and Bode et al. (1969), was tried but rejected; this procedure took a rather long time after which reaction products of the strongly oxidizing NaNiO 2 with ethanol were present.

B. X-ray data collection
Powder X-ray diffraction data were collected on a computer controlled Guinier X-ray powder diffractometer (Huber G644), using CuK a radiation (quartz monochromator). The sample was sealed in a glass capillary of 0.3 mm diameter. The diffraction pattern was collected in the range of 8°«26'«48° with steps of 0.04° and counting times of 10 s for every step. Background intensities were modelled graphically and subtracted from the raw data. Finally, the diffractogram was corrected for absorption of the sample and scaled to 1000 for the strongest reflection.

C. Neutron scattering data collection
Powder neutron scattering data of NaNiO 2 were collected at the Hahn-Meitner-Institut Berlin on the diffractometer E2 (research proposal CHE-01-543). A wavelength of about 0.120 nm was selected with a Ge monochromator. The exact wavelength of X = 0.12165 nm was calibrated against a recorded diffractogram of Si at room temperature and also against the known lattice constant of that sample. Scattering data were recorded with a 10 BF 3 multidetector in a 2 6 range from 10.0 to 89.0° until a monitor had accumulated 2.5 X 10 6 neutron counts (measuring time approximately 1.5 h). The data were corrected for detector cell efficiencies after measuring the scattered intensity of a vanadium sample. Background intensities were estimated graphically and modelled by five background points during the structure refinement.

III. STRUCTURE REFINEMENT
Rietveld analyses of the obtained X-ray and neutron diffractograms were performed with the program PROFIL (Cockroft, 1993). As a starting model cell constants and positional parameters of NaNiO 2 according to Dyer et al. (1954) were used. After refinement of scale factors, diffractometer shifts, with y ohs = y, otal -background and w = y obs .
cell constants, asymmetry and mixing parameters for pseudo-Voigt peak shapes and resolution parameters (U, V, W) the positional parameters of O and isotropic displacement parameters of all atoms were allowed to vary in a joint refinement of X-ray and neutron data. A summary of the crystallographic and refinement data can be found in Table I, the refined positional and displacement parameters in Table II. A refinement of anisotropic displacement parameters was tried but not found useful. Although a significant decrease of R wp was observed with still positively defined U t j, U 2 2 was rather small for all atoms. However, this result seems not to have any physical meaning but rather to be an artifact due to the poorer resolution along the long b* axis (see Fig. 1).  Figure 2 shows a projection of the NaNiO 2 structure along the b axis. The structure can be described as a layered structure, derived from the rhombohedral NaFeO 2 structure. The oxygen ions form an array similar to a distorted hexagonal close packed arrangement. The nickel and sodium ions lie in alternate layers in the distorted octahedral interstices of that array.