Crystal structures of spinel-type Na2MoO4 and Na2WO4 revisited using neutron powder diffraction.

Time-of-flight neutron powder diffraction data have been collected from Na2MoO4 and Na2WO4 to a resolution of sin (θ)/λ = 1.25 Å(-1), which is substanti-ally better than the previous analyses using Mo Kα X-rays, providing roughly triple the number of measured reflections with respect to the previous studies [Okada et al. (1974 ▶). Acta Cryst. B30, 1872-1873; Bramnik & Ehrenberg (2004 ▶). Z. Anorg. Allg. Chem. 630, 1336-1341]. The unit-cell parameters are in excellent agreement with literature data [Swanson et al. (1962 ▶). NBS Monograph No. 25, sect. 1, pp. 46-47] and the structural parameters for the molybdate agree very well with those of Bramnik & Ehrenberg (2004 ▶). However, the tungstate structure refinement of Okada et al. (1974 ▶) stands apart as being conspicuously inaccurate, giving significantly longer W-O distances, 1.819 (8) Å, and shorter Na-O distances, 2.378 (8) Å, than are reported here or in other simple tungstates. As such, this work represents an order-of-magnitude improvement in precision for sodium molybdate and an equally substantial improvement in both accuracy and precision for sodium tungstate. Both compounds adopt the spinel structure type. The Na(+) ions have site symmetry .-3m and are in octa-hedral coordination while the transition metal atoms have site symmetry -43m and are in tetra-hedral coordination.

Chemical context

Both Na2MoO4 and Na2WO4 have rich phase diagrams in pressure and temperature space (Pistorius, 1966 ▸). The stable form at room temperature is the β-Ag2MoO4 cubic spinel structure type, space group Fd m, which has been known for almost a century (Wyckoff, 1922 ▸). Among the alkali metal sulfates, chromates, molybdates and tungstates, only Na2MoO4 and Na2WO4 adopt the normal spinel structure at ambient pressure. Li2MoO4 forms a cubic spinel structure at high pressure (Liebertz & Rooymans, 1967 ▸). Li2WO4 forms a ‘spinel-like’ phase at high pressure (Pistorius, 1975 ▸; Horiuchi et al., 1979 ▸). Cubic sodium molybdate and sodium tungstate have been examined inter­mittently over subsequent decades using a variety of crystallographic techniques (Lindqvist, 1950 ▸; Becka & Poljak, 1958 ▸; Swanson et al., 1957 ▸, 1962 ▸; Singh Mudher et al., 2005 ▸) and vibrational spectroscopic methods (Busey & Keller, 1964 ▸; Preudhomme & Tarte, 1972 ▸; Breitinger et al., 1981 ▸; Luz Lima et al., 2010 ▸, 2011 ▸), or by nuclear magnetic resonance and quadrupole coupling (Lynch & Segel, 1972 ▸). However, the extant structural information on both phases is derived from X-ray diffraction data of low to modest precision. The first published structure refinement of Na2MoO4 was only reported recently (Bramnik & Ehrenberg, 2004 ▸) from X-ray powder diffraction data measured to sin (θ)/λ = 0.71 Å−1; the last structure refinement of Na2WO4 was reported by Okada et al. (1974 ▸) from X-ray single-crystal diffraction data to sin (θ)/λ = 0.81 Å−1. Both compounds are highly soluble in water, crystallizing at room temperature as ortho­rhom­bic dihydrates (space group Pbca, Atovmyan & D’yachenko, 1969 ▸; Farrugia, 2007 ▸). Below 283.5 K for the molybdate and 279.2 K for the tungstate, crystals grow with ten water mol­ecules per formula unit (Funk, 1900 ▸; Cadbury, 1955 ▸; Zhilova et al., 2008 ▸). The high solubility in water and propensity towards forming hydrogen-bonded hydrates (unlike the heavier alkali metal molybdates and tungstates) suggests that both compounds would be excellent candidates for formation of hydrogen-bonded complexes with water soluble organics, such as amino acids, producing metal–organic crystals with potentially useful optical properties (cf., glycine lithium molybdate; Fleck et al., 2006 ▸).

In the course of preparing deuterated specimens of the dihydrated and deca­hydrated forms of Na2MoO4 and Na2WO4 for neutron diffraction analysis, the anhydrous phases were synthesised and an opportunity arose to acquire neutron powder diffraction data. The advantage of using a neutron radiation probe is that the scattering lengths of the atoms concerned are fairly similar, coherent scattering lengths being 6.715 fm for Mo, 4.86 fm for W, 3.63 fm for Na and 5.803 fm for O (Sears, 2006 ▸). Secondly, with the time-of-flight method, particularly with a very long primary flight path and high-angle backscattering detectors, one can acquire unparalleled resolution at very short flight times (i.e., small d-spacings), ensuring an order of magnitude improvement in parameter precision over the previous studies. In this work, usable data were obtained at a resolution of sin (θ)/λ = 1.25 Å−1, roughly tripling the number of measured reflections with respect to Okada et al. (1974 ▸) and Bramnik & Ehrenberg (2004 ▸). This work provides the most accurate and precise foundation on which to build future discussion of the hydrated forms of Na2MoO4 and Na2WO4. Neutron powder diffraction data for Na2MoO4 and Na2WO4 are given in Figs. 1 ▸ and 2 ▸.

Figure 1

Neutron powder diffraction data for Na2MoO4; red points are the observations, the green line is the calculated profile and the pink line beneath the diffraction pattern represents Obs − Calc. Vertical black tick marks report the expected positions of the Bragg peaks. The inset shows the data measured at very short flight times (i.e., small d-spacing).

Figure 2

Neutron powder diffraction data for Na2WO4; red points are the observations, the green line is the calculated profile and the pink line beneath the diffraction pattern represents Obs − Calc. Vertical black tick marks report the expected positions of the Bragg peaks. The inset shows the data measured at very short flight times (i.e., small d-spacing).

Structural commentary

The structure of both compounds is the normal spinel type with Na+ ions on the 16c sites in octa­hedral coordination and Mo6+/W6+ ions on 8b sites in tetra­hedral coordination. The coordinating oxygen atoms occupy the 32e general positions, their location being defined by a single variable parameter u. For ideal cubic close packing, the u coordinate adopts a value of 0.25 although for various spinels is found in the range 0.24 to 0.275. In Na2MoO4 the u parameter has a value of 0.262710 (15) and in Na2WO4 it has a value of 0.262246 (15). The practical consequence of this compared with the ‘ideal’ value of u = 0.25 is that the shared edges of the NaO6 octa­hedra are shorter than the unshared edges (Fig. 3b ▸). In the molybdate, these lengths are 3.2288 (2) and 3.5479 (2) Å, the ratio being 1.0988 (1); in the tungstate, the lengths of the two inequivalent octa­hedral edges are 3.2356 (2) Å and 3.5441 (2) Å, their ratio being 1.0953 (1). The MoO4 2− and WO4 2− tetra­hedra have perfect T symmetry with Mo—O and W—O bond lengths of 1.7716 (3) and 1.7830 (2) Å, respectively. The unit-cell parameters for both compounds are in excellent agreement with those of Swanson et al. (1962 ▸) and the structural parameters for the molybdate agree very well with those of Bramnik & Ehrenberg (2004 ▸). However, the Na2WO4 structure refinement of Okada et al. (1974 ▸) stands apart as being conspicuously inaccurate, giving significantly longer W—O distances, 1.819 (8) Å, and shorter Na—O distances, 2.378 (8) Å, than are reported here or in many other simple tungstates. Indeed the ionic radii of four-coordinated Mo6+ and W6+ obtained from analysis of a large range of crystal structures are nearly identical, being 0.41 and 0.42 Å, respectively (Shannon, 1976 ▸). The values reported here agree very well with the majority of Mo—O and W—O bond lengths in isolated MoO4 2− and WO4 2− tetra­hedral oxyanions from a range of alkali metal and alkaline earth compounds tabulated in the literature (e.g., Zachariasen & Plettinger, 1961 ▸; Gatehouse & Leverett, 1969 ▸; Koster et al., 1969 ▸; Gürmen et al., 1971 ▸; Wandahl & Christensen, 1987 ▸; Farrugia, 2007 ▸; van den Berg & Juffermans, 1982 ▸). As such, this work represents an improvement in accuracy for sodium molybdate and an improvement in both accuracy and precision for sodium tungstate.

Figure 3

(a) Arrangement of molybdate ions in the unit cell of Na2MoO4; anisotropic displacement ellipsoids are drawn at the 75% probability level. (b) Connectivity of the NaO6 octa­hedra, with shorter shared edges and longer unshared edges, to the MoO4 tetra­hedra in Na2MoO4; as in (a), the ellipsoids are drawn at the 75% probability level.

Synthesis and crystallization

Na2MoO4·2H2O (Sigma Aldrich M1003, > 99.5%) and Na2WO4·2H2O (Sigma Aldrich 14304, > 99.0%) were heated to 673 K in ceramic crucibles for 24 hr. Loss of water was confirmed by Raman spectroscopy; X-ray powder diffraction confirmed the phase identity and purity of the two anhydrous products, Na2MoO4 and Na2WO4.

Raman spectra were acquired using a B&WTek i-Raman plus portable spectrometer; this device uses a 532 nm laser (37 mW power at the fiber-optic probe tip) to stimulate Raman scattering, which is measured in the range 170–4000 cm−1 with a spectral resolution of 3 cm−1. Data were collected in a series of 20 x 9 sec integrations for Na2MoO4 and 20 x 7 sec integrations for Na2WO4; after summation, the background was removed and peaks fitted using Pseudo-Voigt functions in OriginPro (OriginLab, Northampton, MA) (Fig. 4 ▸). These data are provided as an electronic supplement in the form of an ASCII file.

Figure 4

Raman spectra of Na2MoO4 (left) and Na2WO4 (right) in the range 0–1200 cm−1 (the full range of data to 4000 cm−1 is given in the electronic supplement). Band positions and vibrational assignments are indicated. For the tungstate these agree very well with literature values (e.g., Busey & Keller, 1964 ▸) whereas for the molybdate, these data show a systematic shift to lower frequencies by 3–4 wavenumbers with respect to published values (Luz Lima et al., 2010 ▸, 2011 ▸).

Refinement

Crystal data, data collection and structure refinement details are summarized in Table 1 ▸. For the neutron scattering experiments, each specimen was loaded into a vanadium tube of 11 mm internal diameter to a depth of approximately 25 mm. The exact sample volume and mass were measured in order to determine the number density for correction of the specimen self-shielding. The samples were mounted on the HRPD beamline (Ibberson, 2009 ▸) at the ISIS neutron spallation source and data were collected in the 10–110 ms time-of-flight window for 2.5 h (Na2MoO4) and 3.5 h (Na2WO4). Data were corrected for self-shielding, focussed to a common scattering angle and normalized to the incident spectrum by reference to a V:Nb null-scattering standard before being output in a format suitable for Rietveld refinement with GSAS/Expgui (Larsen & Von Dreele, 2000 ▸: Toby, 2001 ▸).

Table 1

Experimental details

	Na2MoO4	Na2WO4
Crystal data
Chemical formula	Na2MoO4	Na2WO4
M r	205.92	293.83
Crystal system, space group	Cubic, F d m	Cubic, F d m
Temperature (K)	298	298
a ()	9.10888(3)	9.12974(4)
V (3)	755.78(1)	760.98(1)
Z	8	8
Radiation type	Neutron	Neutron
(mm1)	0.014 + 0.0018 *	0.014 + 0.0097 *
Specimen shape, size (mm)	Cylinder, 25 11	Cylinder, 27 11
Data collection
Diffractometer	HRPD, high-resolution neutron powder	HRPD, high-resolution neutron powder
Specimen mounting	Vanadium tube	Vanadium tube
Data collection mode	Transmission	Transmission
Scan method	Time of flight	Time of flight
Absorption correction	Analytical	Analytical
2 values ()	2fixed = 168.329	2fixed = 168.329
Distance from source to specimen (mm)	95000	95000
Distance from specimen to detector (mm)	965	965
Refinement
R factors and goodness of fit	R p = 0.037, R wp = 0.043, R exp = 0.022, R(F 2) = 0.06364, 2 = 3.842	R p = 0.037, R wp = 0.044, R exp = 0.024, R(F 2) = 0.06245, 2 = 3.423
No. of data points	7716	7716
No. of parameters	24	24

Computer programs: HRPD control software, GSAS/Expgui (Larsen Von Dreele, 2000 ▸; Toby, 2001 ▸), MANTID (Arnold et al., 2014 ▸; Mantid, 2013 ▸), DIAMOND (Putz Brandenburg, 2006 ▸) and publCIF (Westrip, 2010 ▸).

Crystal structure: contains datablock(s) Na~2~MoO~4~, Na~2~WO~4~, New_Global_Publ_Block. DOI: 10.1107/S2056989015008774/wm5152sup1.cif

Rietveld powder data: contains datablock(s) Na2MoO4. DOI: 10.1107/S2056989015008774/wm5152Na2MoO4sup2.rtv

Rietveld powder data: contains datablock(s) Na2WO4. DOI: 10.1107/S2056989015008774/wm5152Na2WO4sup3.rtv

Supporting information file. DOI: 10.1107/S2056989015008774/wm5152sup4.txt

CCDC references: 1063434, 1063433

Additional supporting information: crystallographic information; 3D view; checkCIF report

Na2WO4	Melting point: 969 K
Mr = 293.83	Neutron radiation
Cubic, Fd3m	µ = 0.01+ 0.0097 * λ mm−1
Hall symbol: -F 4vw 2vw 3	T = 298 K
a = 9.12974 (4) Å	white
V = 760.98 (1) Å3	cylinder, 27 × 11 mm
Z = 8	Specimen preparation: Prepared at 673 K and 100 kPa
Dx = 5.129 Mg m−3

HRPD, High resolution neutron powder diffractometer	Absorption correction: analytical Data were corrected for self shielding using σscatt = 28.088 barns and σab(λ) = 19.361 barns at 1.798 Å during the normalisation procedure. The linear absorption coefficient is wavelength dependent and is calculated as: µ = 0.014 + 0.0097 * λ [mm-1]
Radiation source: ISIS Facility, Neutron spallation source	Tmin = 1.000, Tmax = 1.000
Specimen mounting: vanadium tube	2θfixed = 168.329
Data collection mode: transmission	Distance from source to specimen: 95000 mm
Scan method: time of flight	Distance from specimen to detector: 965 mm

Least-squares matrix: full	Excluded region(s): Data at d-spacings smaller than 0.4 Å were excluded since the counting statistics became progressively poorer at very short flight times due to the lower neutron flux at the shortest wavelengths.
Rp = 0.037	Profile function: TOF profile function #3 (21 terms). Profile coefficients for exp pseudovoigt convolution [Von Dreele, 1990 (unpublished)] (α) = 0.1603, (β0) = 0.026115, (β1) = 0.004558, (σ0) = 0, (σ1) = 237.2, (σ2) = 45.0, (γ0) = 0, (γ1) = 14.21, (γ2) = 0, (γ2s) = 0, (γ1e) = 0, (γ2e) = 0, (εi) = 0, (εa) = 0, (εA) = 0, (γ11) = 0, (γ22) = 0, (γ33) = 0, (γ12) = 0, (γ13) = 0, (γ23) = 0. Peak tails ignored where intensity <0.0005x peak. Aniso. broadening axis 0.0 0.0 1.0
Rwp = 0.044	24 parameters
Rexp = 0.024	0 restraints
R(F2) = 0.06245	0 constraints
χ2 = 3.423	(Δ/σ)max = 0.01
7716 data points	Background function: GSAS Background function # 1 (10 terms). Shifted Chebyshev function of 1st kind 1: 0.884779, 2: 4.212470x10-2, 3: 4.210950x10-2, 4: -4.489520x10-2, 5: -2.683690x10-2, 6: -1.892450x10-2, 7: -2.248710x10-2, 8: -2.821970x10-3, 9: 6.467340x10-3, 10: 6.167050x10-3

	x	y	z	Uiso*/Ueq
O	0.262246 (15)	0.262246 (15)	0.262246 (15)	0.01312
W	0.375	0.375	0.375	0.00903
Na	0.0	0.0	0.0	0.01538

	U11	U22	U33	U12	U13	U23
O	0.01312 (6)	0.01312 (6)	0.01312 (6)	−0.00161 (5)	−0.00161 (5)	−0.00161 (5)
W	0.00903 (11)	0.00903 (11)	0.00903 (11)	0.0	0.0	0.0
Na	0.01538 (11)	0.01538 (11)	0.01538 (11)	−0.00045 (12)	−0.00045 (12)	−0.00045 (12)

W—O	1.7830 (2)	Naiv—Ovii	2.3995 (1)
W—Oi	1.7830 (2)	Naiv—Oviii	2.3995 (1)
W—Oii	1.7830 (2)	Naiv—Oix	2.3995 (1)
W—Oiii	1.7830 (2)	Na—Naiv	3.2279 (1)
Naiv—O	2.3995 (1)	W—Naiv	3.7850 (1)
Naiv—Ov	2.3995 (1)	O—Ovi	3.2356 (2)
Naiv—Ovi	2.3995 (1)	O—Ov	3.5441 (4)
O—W—Oi	109.4712 (3)	O—Naiv—Ov	95.211 (6)
O—W—Oii	109.4712 (3)	O—Naiv—Oix	180.000 (1)
O—W—Oiii	109.4712 (3)	O—Naiv—Oviii	84.789 (6)
Oi—W—Oii	109.4712 (3)	O—Naiv—Ovii	95.211 (6)
Oi—W—Oiii	109.4712 (3)	Ov—Naiv—Ovi	180.000 (1)
Oii—W—Oiii	109.4712 (3)	W—O—Naiv	129.043 (4)
O—Naiv—Ovi	84.789 (6)