Powder neutron diffraction of SrNbO₂N at room temperature and 1.5 K

The structure of strontium niobium dioxygen nitride, SrNbO₂N, has been solved by powder neutron diffraction at room temperature and 1.5 K. SrNbO₂N crystallizes in the tetragonal space group I4/mcm, with a = 5.7056 (4) and c = 8.1002 (9) Å at room temperature, and a = 5.6938 (4) and c = 8.0974 (8) Å at 1.5 K. The crystal structure is derived from the cubic perovskite archetype by a slight rotation of the Nb(O,N)₆ octahedra with respect to the tetragonal axis. A partially ordered distribution of oxygen and nitrogen on the anionic sites was found.

S1. Comment

Perovskite-related oxides of transition metals with a d⁰ configuration (e.g. Ti⁺⁴, Nb⁺⁵ or Ta⁺⁵) are usually colourless insulators. The corresponding oxynitrides, on the other hand, often exhibit bright colours and a significantly lower resistivity. The origin for this phenomenon lies in the more covalent character of the M—N bond compared with the corresponding M—O bond, leading to a reduction of the electronic band gap. Their colour and electrical properties make oxynitride perovskites interesting candidates as pigments (Jansen & Letschert, 2000) or as photocatalysts, e.g. for the light-induced water-splitting reaction (Kasahara et al., 2002).

Due to their very similar atomic form factors, O and N cannot be distinguished by X-ray diffraction. As an additional disadvantage, X-ray patterns of perovskites are usually dominated by the contribution of the heavier elements (i.e. the cations), making it difficult to gain detailed information about the anions. Neutron diffraction, on the other hand, is well suited for the investigation of oxynitride perovskites: O and N differ significantly in their neutron scattering lengths and can therefore be easily distinguished. Additionally, the scattering lengths of O and N are comparable with those of the cationic species, making it possible to determine the exact positions, (anisotropic) displacement parameters and site occupation factors of these anions.

On the basis of powder X-ray diffraction measurements, SrNbO₂N was originally reported to possess a cubic perovskite structure (Marchand et al., 1986). Very recently, Kim et al. (2004) published a structural model in the correct space group, I4/mcm. As these authors also used powder X-ray diffraction, no information about the anionic ordering could be provided. Furthermore, the small scattering contribution of the anions makes it impossible to distinguish between space group I4/mcm and other related space groups, such as I42m. Against this background, we present here the crystal structure determination of the title compound, (I), obtained from powder neutron diffraction data at room- and low-temperature. \sch

The present neutron diffraction data were recorded on the High-Resolution Powder Diffractometer for Thermal Neutrons (HRPT) of the Swiss Neutron Spallation Source (SINQ) at the Paul Scherrer Institute (PSI) in Switzerland. Data collection was performed in the high-intensity mode of the instrument.

The powder neutron patterns for the measurements at 293 K and 1.5 K, together with the fit results and their differences, are shown in Figs. 1 and 2. A l l peaks could be indexed with a tetragonal unit cell, with a = b ≈ a_c^1/2, and c ≈ 2a_c (a_c being the cell parameter of the cubic perovskite). It is to be noted that, although 2^1/2a ≈ c (with differences of only −0.031 Å at 293 K and −0.045 Å at 1.5 K), the unit cell is not cubic. Using a LeBail fit of the neutron data with the corresponding F-centred cubic supercell, only a poor agreement was achieved. Additionally, the powder X-ray pattern showed a pronounced splitting of various peaks, clearly proving the tetragonal distortion.

As a starting model for the refinement of (I), the structure of SrTaO₂N (Günther et al., 2000) was used. Other structural models found for related oxynitride perovskites, such as CaTaO₂N and LaTaON₂ (Günther et al., 2000) and LaTiO₂N (Clarke Guinot et al., 2002), were also considered, but led to rather poor agreements. Since Pors et al. (1988) reported SrTaO₂N to crystallize in space group I42m, we additionally attempted to refine our neutron data in this space group, which indeed led to reasonable results. The residual parameters, on the other hand, remained slightly larger than those for the calculations in I4/mcm, despite the greater number of refinable parameters. For this reason, we concluded that the higher symmetric space group I4/mcm is better suited to describing the structure of SrNbO₂N.

Refinement runs converged smoothly, yielding the results listed below. To reduce the number of free parameters, isotropic displacement parameters were used for Sr and Nb. This approximation is reasonable because, in perovskites, the displacements of the cations are known to be usually almost spherical. Rietveld programs tend to underestimate heavily the errors of the cell parameters. For this reason, we multiplied the standard uncertainties of a and c by a factor of 10.

A graphical presentation of the crystal structure of (I) at 293 K is shown in Fig. 2. Surprisingly, the structure barely changes upon cooling. While the cell parameter a decreases by only 0.2%, c remains almost unchanged (Δc/c = −0.03%). In addition, the Nb—O,N distances remain very similar. As can be seen in Fig. 2, the Nb(O,N)₆ octahedra are slightly rotated around the crystallographic c axis. The rotation angle increases from 5.5 (1) to 6.6 (1)° on cooling to 1.5 K.

The experimentally determined N contents of 0.93 (3) and 0.96 (3) at 293 and 1.5 K, respectively, are (within a 2σ tolerance range) in good agreement with the expected value of 1. Comparing the O/N distribution, we observed a partial order of the anions. The axial position (4a) is occupied by 84% O/16% N, while for the equatorial position (8 h), an occupancy of 61% O/39% N was obtained. Within experimental error, identical values were found for the room- and low-temperature measurements. It is interesting to compare these data with the results given for the closely related oxynitride SrTaO₂N. In contrast with the O enrichment on the axial position which we have observed in SrNbO₂N, Pors et al. (1988) and Clarke Hardstone et al. (2002) reported a slight enrichment of O on the equatorial position. Günther et al. (2000), on the other hand, found complete O/N order, with the O ions exclusively occupying the equatorial sites. It was assumed that the different results for SrTaO₂N originate from variations in sample preparation. Further investigations are currently in progress to study the influence of the preparation technique on anionic order in SrNbO₂N.

S2. Experimental

Polycrystalline SrNbO₂N was prepared by ammonolysis of thoroughly ground stoichiometric mixtures of SrCO₃ and Nb₂O₅ in alumina boats at 1223 K. An equimolar amount of KCl was added as mineralizer. Phase purity was checked by X-ray diffraction measurements. A single-phase product was achieved after two ammonolysis cycles of 18 h each with an intermediate regrinding.

S3. Refinement

For both anion positions, a complete site occupation was assumed and only the ratios O1:N1 and O2:N2, respectively, were allowed to vary.

Computing details

For both compounds, data collection: SINQ Instrument Control System (SICS); cell refinement: FULLPROF2000 (Rodríguez-Carvajal, 1990); data reduction: SINQ Instrument Control System (SICS); program(s) used to refine structure: FULLPROF2000; molecular graphics: ATOMS (Dowty, 1995); software used to prepare material for publication: Please provide missing details.

Figures

	Fig. 1. A Rietveld refinement plot of SrNbO2N at 293 K.
	Fig. 2. A Rietveld refinement plot of SrNbO2N at 1.5 K.
	Fig. 3. The structure of SrNbO2N. Displacement parameters are shown at the 90% probability level.

(RT) strontium niobium dioxygen nitride

Crystal data

SrNbO2N	Z = 1
Mr = 906.72	Dx = 5.710 Mg m^−3
Tetragonal, I4/mcm	Neutron radiation, λ = 1.49328 Å
Hall symbol: -I 4 2c	µ = 0.01 mm^−1
a = 5.7056 (4) Å	T = 293 K
c = 8.1002 (9) Å	brown
V = 263.69 (4) Å^3	cylinder, 50 × 10 mm

Data collection

SINQ HRPT diffractometer	Data collection mode: transmission
Ge (533) monochromator	Scan method: Stationary detector
Specimen mounting: vanadium can

Refinement

Refinement on Inet	Profile function: Thompson-Cox-Hastings pseudo-Voigt
Least-squares matrix: full with fixed elements per cycle	20 parameters
Rp = 0.025	0 restraints
Rwp = 0.033	0 constraints
Rexp = 0.020	Weighting scheme based on measured s.u.'s
χ^2 = 2.657	(Δ/σ)max = 0.0001
3206 data points	Background function: linear interpolation between 37 points

Special details

Geometry. Bond distances, angles etc. have been calculated using the rounded fractional coordinates. All e.s.d.'s are estimated from the variances of the (full) variance-covariance matrix. The cell e.s.d.'s are taken into account in the estimation of distances, angles and torsion angles

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq	Occ. (<1)
Sr	0.00000	0.50000	0.25000	0.0092 (2)*
Nb	0.00000	0.00000	0.00000	0.0074 (2)*
O1	0.00000	0.00000	0.25000	0.0123 (4)	0.85 (2)
N1	0.00000	0.00000	0.25000	0.0123 (4)	0.15 (2)
O2	0.7741 (1)	0.2741 (1)	0.00000	0.0164 (2)	0.61 (1)
N2	0.7741 (1)	0.2741 (1)	0.00000	0.0164 (2)	0.39 (1)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
O1	0.0156 (7)	0.0156 (7)	0.0055 (9)	0.000	0.000	0.000
N1	0.0156 (7)	0.0156 (7)	0.0055 (9)	0.000	0.000	0.000
O2	0.0110 (3)	0.0110 (3)	0.0271 (5)	0.0061 (3)	0.000	0.000
N2	0.0110 (3)	0.0110 (3)	0.0271 (5)	0.0061 (3)	0.000	0.000

Geometric parameters (Å, º)

Nb—O1/N1	2.0250 (2)	Sr—O1/N1	2.8528 (2)
Nb—O2/N2^i	2.0266 (8)	Sr—O2/N2^ii	2.9990 (6)
Sr—O2/N2^i	2.7244 (5)
Nb^iii—O2/N2—Nb^iv	169.0 (1)

Symmetry codes: (i) x−1, y, z; (ii) −y, x, z; (iii) −x+1/2, y+1/2, −z; (iv) x+1, y, z.

(LT) strontium niobium dioxygen nitride

Crystal data

SrNbO2N	Z = 1
Mr = 906.48	Dx = 5.734 Mg m^−3
Tetragonal, I4/mcm	Neutron radiation, λ = 1.49328 Å
Hall symbol: -I 4 2c	µ = 0.01 mm^−1
a = 5.6938 (4) Å	T = 2 K
c = 8.0974 (8) Å	brown
V = 262.51 (4) Å^3	cylinder, 50 × 10 mm

Data collection

SINQ HRPT diffractometer	Data collection mode: transmission
Ge (533) monochromator	Scan method: Stationary detector
Specimen mounting: vanadium can

Refinement

Refinement on Inet	Profile function: Thompson-Cox-Hastings pseudo-Voigt
Least-squares matrix: full with fixed elements per cycle	20 parameters
Rp = 0.028	0 restraints
Rwp = 0.036	0 constraints
Rexp = 0.023	Weighting scheme based on measured s.u.'s
χ^2 = 2.465	(Δ/σ)max = 0.0001
3206 data points	Background function: linear interpolation between 42 points

Special details

Geometry. Bond distances, angles etc. have been calculated using the rounded fractional coordinates. All e.s.d.'s are estimated from the variances of the (full) variance-covariance matrix. The cell e.s.d.'s are taken into account in the estimation of distances, angles and torsion angles

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq	Occ. (<1)
Sr	0.00000	0.50000	0.25000	0.0036 (2)*
Nb	0.00000	0.00000	0.00000	0.0058 (2)*
O1	0.00000	0.00000	0.25000	0.0098 (4)	0.82 (2)
N1	0.00000	0.00000	0.25000	0.0098 (4)	0.18 (2)
O2	0.7788 (1)	0.2788 (1)	0.00000	0.0121 (2)	0.61 (1)
N2	0.7788 (1)	0.2788 (1)	0.00000	0.0121 (2)	0.39 (1)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
O1	0.0120 (6)	0.0120 (6)	0.0053 (8)	0.000	0.000	0.000
N1	0.0120 (6)	0.0120 (6)	0.0053 (8)	0.000	0.000	0.000
O2	0.0070 (3)	0.0070 (3)	0.0223 (5)	0.0029 (3)	0.000	0.000
N2	0.0070 (3)	0.0070 (3)	0.0223 (5)	0.0029 (3)	0.000	0.000

Geometric parameters (Å, º)

Nb—O1/N1	2.0243 (2)	Sr—O1/N1	2.8469 (2)
Nb—O2/N2^i	2.0263 (6)	Sr—O2/N2^ii	3.0227 (5)
Sr—O2/N2^i	2.6965 (4)
Nb^iii—O2/N2—Nb^iv	166.9 (1)

Symmetry codes: (i) x−1, y, z; (ii) −y, x, z; (iii) −x+1/2, y+1/2, −z; (iv) x+1, y, z.