I. INTRODUCTION
The crystal structures of various Na2XO3·5H2O (X = HP (hydrogen phosphorus), Te, Se) have been reported in the past (Colton and Henn, Reference Colton and Henn1971; Philippot et al., Reference Philippot, Maurin and Moret1979; Mereiter, Reference Mereiter2013) from single crystal diffraction. Among the series one composition is missing which is the sulfite counterpart Na2SO3·5H2O. The crystal structure of heptahydrate has been reported (Weil and Mereiter, Reference Weil and Mereiter2020) but not the pentahydrate. During investigation of some selenates, we have been using Na2SeO3·5H2O as a starting material. At this point, we realized that the observed experimental pattern was not fitting any known previously reported structures. Consequently, we have been investigating its crystal structure using laboratory X-ray powder diffraction.
II. EXPERIMENTAL
A. Sample preparation
Na2SeO3·5H2O was purchased from Merck Darmstadt (Lot 8538996) and used without further purification or recrystallization. The powder was ground with a pestle in an agate mortar. Upon grinding and/or prolonged exposure to air after grinding, the sample loses its water to exhibit Na2SeO3 as a predominant phase. The sample was unfortunately too coarse to not undergo any grinding to obtain any reasonable quality powder diffraction data. Consequently, the final sample always contains a large amount of Na2SeO3 impurity. Nevertheless, it was possible to solve the crystal structure of this new polymorph.
B. X-ray powder diffraction and data preparation
For the X-ray powder diffraction measurements, the sample was loosely loaded into a glass capillary of 0.5 mm diameter in order to minimize absorption (μ = 33.93 mm−1) and spun to improve particle statistics. X-ray powder data were collected at room temperature on an Empyrean Series 3 instrument using a Cu X-ray tube, a focusing mirror, Soller slits of 0.04 rad, and a 1Der strip detector used in 1D scanning line mode. A 2Θ range of 10–110° was collected in transmission geometry. Data collector version 7.3 from the Empyrean diffractometer was used for data collection and reduction.
C. Structure solution and refinement
Due to the tendency of the sample to lose water, it has not been possible to measure a pure phase sample. Na2SeO3 was always present in a rather large impurity phase (almost 30% in this case). Besides Na2SeO3, the additional reflections present could not be attributed to any known phases nor to the previously reported phase of Na2SeO3·5H2O (Mereiter, Reference Mereiter2013). This suggested the presence of either a new hydrate of Na2SeO3, or a new polymorph.
Indexing of the pattern was performed with DICVOL (Boultif and Louër, Reference Boultif and Louër2004) as implemented in the HighScore suite (Degen et al., Reference Degen, Sadki, Bron, König and Nénert2014) using 21 non-overlapping reflections. The determined unit cell is a = 15.013 Å, b = 7.0306 Å, c = 8.1347 Å, and β = 98.4° (F(21) = 1822.9 (0.0002,47)). Refinement of lattice parameters and peak-profile determination were performed by Pawley profile fitting followed by a space group search (Markvardsen et al., Reference Markvardsen, Shankland, David, Johnston, Ibberson, Tucker, Nowell and Griffin2008) using the HighScore suite (Degen et al., Reference Degen, Sadki, Bron, König and Nénert2014). The as-determined space group was P121/n1. This space group is confirmed by the good Pawley fit obtained as illustrated in Figure 1(a) where all the reflections are being indexed.
Figure 1. (a) Mixed phase fit of the Rietveld fit of Na2SeO3 plus a Pawley fit of the Na2SeO3·5H2O phase with the symmetry P21/n. (b) Using the newly implemented option within the HighScore suite, the two deconvoluted scans are used to obtain a phase pure scan of Na2SeO3·5H2O suitable for structure solution.
The challenge arises to solve the crystal structure while having many overlapping reflections between the two phases. In order to deconvolute the two contributions and extract only the experimental pattern of interest, we have performed a conversion of the observed phase profiles (here the Pawley fit of the Na2SeO3·5H2O phase and of the Rietveld fit of Na2SeO3) to individual scans. This is illustrated in Figure 1(b). Obviously, the quality of the conversion relies a lot on the quality of the initial fit, so it is important to carry out a full Rietveld refinement after the structure determination of the two phases in order to confirm the obtained structural model.
The obtained deconvoluted pattern of Na2SeO3·5H2O shown in Figure 1(b) was used for structure solution. The structure was determined using FOX (Favre-Nicolin and Cerny, Reference Favre-Nicolin and Cerny2002) by a parallel tempering algorithm. The starting model was made of a SeO3 group taken from Na2SeO3 and 2 NaO6 octahedra. Rietveld refinement for the final structure was performed using Jana2006 (Petricek et al., Reference Petricek, Dusek and Palatinus2014). U iso of all oxygen atoms were restrained to be equal while the same was applied to the sodium atoms and the U iso of the selenium was freely refined. Twenty-five parameters using a Legendre polynomial were used to describe the background. Peak shape was fitted using a Pseudo-Voigt function using only two parameters (W and the ratio Voigt/Gaussian profile) plus the asymmetry. The final Rietveld fit is shown in Figure 2, while crystal data, data collection, and structure refinement details are summarized in Tables I and II.
Figure 2. Final Rietveld refinement of the phase mixture β-Na2SeO3·5H2O and Na2SeO3. The observed pattern (red dots), the Rietveld fit (black line), and the difference between observed and diffracted intensities (blue line) are shown together with the allowed Bragg reflections (green lines for both phases). The ratio of the phases β-Na2SeO3·5H2O/Na2SeO3 is 71.5%/28.5%.
TABLE I. Data collection and structure refinement details for the β form of disodium selenite pentahydrate
TABLE II. Atomic coordinates obtained from the Rietveld refinement of the β form of disodium selenite pentahydrate
III. RESULTS
The crystal structure of β-Na2SeO3·5H2O is shown in Figure 3 and can be regarded as a layered structure. While the crystal structure of β-Na2SeO3·5H2O is different from the other members of the series, we note that all of them have a layered structure. The major difference lies in the fact that while Na2XO3·5H2O (X = HP, Te) (Colton and Henn, Reference Colton and Henn1971; Philippot et al., Reference Philippot, Maurin and Moret1979) and α-Na2SeO3·5H2O (Mereiter, Reference Mereiter2013) exhibit vertices and edges sharing of their respective Na polyhedra, however β-Na2SeO3·5H2O exhibits only edge sharing polyhedra. In addition, the layers of β-Na2SeO3·5H2O are isolated, forming double 1D chains running along the b-axis (see Figure 3). The SeO3 groups are connected to the 1D chains but not connecting the chains between each other contrary to the other members of the series. This makes β-Na2SeO3·5H2O structurally significantly different from the series. For completeness, in Figure 4, we are comparing the two polymorphic forms of Na2SeO3·5H2O.
Figure 3. The crystal structure of β-Na2SeO3·5H2O projected along (a) the c-axis and along (b) the b-axis.
Figure 4. Comparison between the two polymorphic forms of Na2SeO3·5H2O. The α phase is shown in (a) while a representation of the newly reported β phase is shown in (b).
As illustrated in Figure 4, the α-Na2SeO3·5H2O phase (Mereiter, Reference Mereiter2013) exhibits clearly a 3D structure forming channels toward which the lone pair of selenium are pointing while the β-Na2SeO3·5H2O exhibits a layered structure.
Additional in-situ heat treatments show that β-Na2SeO3·5H2O readily transforms to Na2SeO3 between 35 and 40 °C without any evidence for any intermediate phase. So, the condensation of the network is very fast and that is consistent with the difficulty to obtain this β phase in pure form as it loses its water of crystallization quickly.
IV. DATA DEPOSITION
The crystallographic data of the new polymorphic form of Na2SeO3·5H2O are deposited as a crystallographic information file (CIF) in the Cambridge Structural Database (CSD) with the deposition number 2255102.
ACKNOWLEDGEMENTS
The author acknowledges the feedback from an anonymous referee and providing the results of DFT calculations. The excellent agreement between the Rietveld-refined and DFT-optimized positions of the heavy atoms are confirming the structure correctness.
