Tetragonal CeNbO₄ at 1073 K in air and in vacuo

The structure of the high-temperature scheelite-type polymorph of cerium niobium tetraoxide, CeNbO₄, has been determined using time-of-flight neutron powder diffraction data collected both in situ at 1073 K in air and in vacuo. In both cases, the structure was found to be tetragonal, with I4₁/a symmetry and without any significant deviation from the stoichiometric composition.

Comment

Several authors have investigated the structure of CeNbO₄, particularly under ambient conditions (Negas et al., 1977; Roth et al., 1977, 1978; Cava et al., 1978; Santoro et al., 1980; Thompson et al., 1999), and only recently has there been any attempt to characterize the structure at elevated temperatures, with one recent report of an in situ characterization of phase transformations on heating that presented minimal structural information (Skinner & Kang, 2003). The recent interest in this material has stemmed from the possible incorporation of oxy­gen interstitials that would make this material amenable to oxide ion conducting applications, such as solid electrolytes, sensors and separation membranes. The purpose of this study was to investigate the structure of the high-temperature polymorph under both static air and vacuum conditions in order to gather information regarding the likely flexibility of oxy­gen stoichiometry in CeNbO₄.

The low-temperature polymorph of CeNbO₄ has been described previously as a monoclinic distortion of the tetragonal scheelite structure, adopting a fergusonite-type structure (Santoro et al., 1980; Thompson et al., 1999). It was also predicted that CeNbO₄ will undergo oxidation on heating in air and a structural transition on heating above 847 K (Gingerich & Blair, 1964; Negas et al., 1977; Roth et al., 1977; Cava et al., 1978). Recently, in situ measurements have raised questions about the nature of these oxidation and transformation processes (Skinner & Kang, 2003). However, until the present work, there has been no determination of the structure of the tetragonal form of CeNbO₄. From our initial X-ray diffraction results, it was immediately apparent that the data recorded at 1073 K conformed to a scheelite-type structure. As a starting model for the Rietveld refinement of the neutron powder diffraction data, the structure of scheelite (CaWO₄) was therefore used [space group I4₁/a; alternate setting with origin at (0, , ); the Ce and Nb atoms in the 4b and 4a positions, respectively, and the O atom in the 16f position (Hazen et al., 1985)]. The refinements were carried out using the GSAS package (Larson & Von Dreele, 1994) and provided good fits to the 90° and backscattered neutron diffraction data collected in air and in vacuo (10⁻³ Pa) (Fig. 1).

In common with other scheelite-type compounds, the Ce atoms in CeNbO₄ are in an eightfold coordination environment, with two distinct bond lengths, while the Nb environment is tetrahedral, with equal Nb—O bonds (Fig. 2). The slightly larger unit-cell volume of the sample heated in static air is associated with slightly longer Nb—O and Ce—O distances (Tables 1 and 2). The distorted tetrahedral environment for the Nb atom, with O—Nb—O angles of 106.476 (19) and 115.64 (4)°, is comparable to that in isostructural LaNbO₄ (David, 1983; Machida et al., 1995). The transition temperature in CeNbO₄ is significantly higher than that in LaNbO₄ (773 K).

From the results of both refinements and from Fourier difference maps, it is apparent that the high-temperature CeNbO₄ polymorph is fully stoichiometric, and there is no evidence to suggest that heating it in air introduced any interstitial oxy­gen. Because the sample was heated in situ while packed in a vanadium can, it is conceivable that limited oxidation occurred at the surface only. Hence, it would be desirable either to perform a series of measurements at one temperature over a period of time to enable the study of possible oxy­gen incorporation into CeNbO₄ or to oxidize a sample before carrying out a set of measurements at temperatures up to 1073 K, allowing any oxy­gen stoichiometry variations on heating to be investigated.

Figure 1 (a) A Rietveld plot for the refinement of backscattered neutron diffraction data in vacuo. The experimental and calculated profiles are represented by crosses and a full line, respectively. The difference plot is shown underneath and tick marks indicate the positions of Bragg reflections. (b) A Rietveld plot for the refinement of the 90° bank neutron diffraction data in air. Profiles are drawn as in (a).

Figure 2 A representation of the tetragonal scheelite-type structure of CeNbO4, showing the Nb and Ce coordination environments. Displacement ellipsoids are drawn at the 50% probability level. [Symmetry code: (i) x + , y + ,  − z.]

Experimental

Samples were produced through the mixing of CeO₂ (99.9%, Sigma Aldrich) and Nb₂O₅ (99.99%, Sigma Aldrich) in an agate mortar and pestle. These starting materials were mixed thoroughly under acetone and allowed to dry in air before the stoichiometric mixture was transferred to an alumina crucible for heat treatment. The mixture was prefired at 1273 K for 10 h and allowed to cool. After further mixing, the material was heated at 1673 K for 16 h and quenched. This procedure produced a bright-green material. X-ray powder diffraction data were recorded on a Philips PW1700 series diffrac­tometer using Cu Kα₁ radiation and a graphite crystal as secondary monochromator. Neutron powder diffraction data were collected from 6 g samples contained in vanadium cans on the POLARIS diffractometer using both 90° and backscatter detectors at ISIS, CCLRC Rutherford Appleton Laboratory, Oxfordshire, England. Each data set was collected over a period of 90 min, and then normalized and corrected in the usual way. Two separate samples were ex­amined during these measurements. The first sample was heated in situ under vacuum with data recorded on heating and cooling. For the second sample, the same in situ heating regime was used but the can was exposed to static air. The data were collected at 296 K, every 100 K from 473 to 973 K, and every 50 K from 973 to 1123 K. On cooling, data were recorded at 1023, 823 and 623 K. Only the data recorded at 1073 K from the samples heated in air and in vacuo are included here.

In vacuo

Crystal data

• CeNbO₄

• M_r = 297.02

• Tetragonal, I4₁/a, origin choice 2 at (0, , ) from

• a = 5.37119 (8) Å

• c = 11.58104 (18) Å

• V = 334.109 (9) ų

• Z = 4

• D_x = 5.905 Mg m⁻³

• Neutron radiation

• λ = 0.1–6.0 Å

• T = 1073 K

• Light green

Data collection

• Polaris diffractometer at ISIS (England)

• Specimen mounting: 6 mm diameter vanadium can

• 2884 independent reflections

Refinement

• Refinement on F²

• R_p = 0.0196

• R_wp = 0.0116

• R_exp = 0.0065

• S = 1.81

• Profile function: exponential pseudo-Voigt convolution

• 2884 reflections

• 49 parameters

• (Δ/σ)_max = 0.02

Table 1
Selected interatomic distances (Å) for vacuum data

Ce⋯Ce	3.94904 (4)
Ce—O^ii	2.5100 (6)
Ce—O^iii	2.4847 (6)
Nb—O	1.8537 (5)
Symmetry codes: (ii) -x,1-y,1-z; (iii) .

In air

Crystal data

• CeNbO₄

• M_r = 297.02

• Tetragonal, I4₁/a, origin choice 2 at (0, , ) from

• a = 5.37692 (8) Å

• c = 11.59514 (18) Å

• V = 335.230 (8) ų

• Z = 4

• D_x = 5.835 Mg m⁻³

• Neutron radiation

• λ = 0.1–6.0 Å

• T = 1073 K

• Light green

Data collection

• Polaris diffractometer at ISIS (England)

• Specimen mounting: 6 mm diameter vanadium can

• 2884 independent reflections

Refinement

• Refinement on F²

• R_p = 0.0203

• R_wp = 0.0118

• R_exp = 0.0065

• S = 1.83

• Profile function: exponential pseudo-Voigt convolution

• 2884 reflections

• 49 parameters

• (Δ/σ)_max = 0.03

Table 2
Selected interatomic distances (Å) for air data

Ce⋯Ce^iv	3.95358 (4)
Ce—O^ii	2.5128 (6)
Ce—O^iii	2.4882 (6)
Nb—O	1.8553 (5)
Symmetry codes: (ii) -x,1-y,1-z; (iii) ; (iv) .

Initial refinement cycles showed no significant deviation from stoichiometry, and therefore the occupancies of the Ce- and Nb-atom sites were fixed to unity in the final cycles for both the air and the vacuum data sets. The occupancy of the O-atom site was allowed to vary but did not deviate significantly from unity either. Fourier difference maps indicated no significant residual scattering density within the unit cell, with maxima of 0.49 Å⁻³ at (0.2465, 0.5004, 0.1234) for the vacuum data and 0.46 Å⁻³ at (0.1545, 0.9652, 0.1803) for the air data.

For both compounds, data ecollection: Polaris instrument control program; cell refinement: GSAS (Larson & Von Dreele, 1994); program(s) used to solve structure: GSAS; program(s) used to refine structure: GSAS; molecular graphics: GRETEP (Laugier & Bochu, 2004) and POVRAY (URL: www.povray.org).

Supporting information

S1. Comment

Several authors have investigated the structure of CeNbO₄, particularly under ambient conditions (Negas et al., 1977; Roth et al., 1977; Roth et al., 1978; Cava et al., 1978; Santoro et al., 1980; Thompson et al., 1999) and only recently has there been any attempt to characterize the structure at elevated temperatures, with one recent report of an in situ characterization of phase transformations on heating that presented minimal structural information (Skinner & Kang, 2003). The recent interest in this material has stemmed from the possible incorporation of oxygen interstitials that would make this material amenable to oxide ion conducting applications, such as solid electrolytes, sensors and separation membranes. The purpose of this study was to investigate the structure of the high-temperature polymorph under both static air and vacuum conditions in oreder to gather information regarding the likely flexibility of oxygen stoichiometry in CeNbO₄.

The low-temperature polymorph of CeNbO₄ has previously been described as a monoclinic distortion of the tetragonal scheelite structure, adopting a fergusonite-type structure (Santoro et al., 1980; Thompson et al., 1999). Previously, it was also predicted that CeNbO₄ will undergo oxidation upon heating in air and a structural transition upon heating above 847 K (Gingerich & Blair, 1964; Cava et al., 1976; Negas et al., 1977; Roth et al., 1977). Recently, in situ measurements have raised questions about the nature of these oxidation and transformation processes (Skinner & Kang, 2003). However, until the present work, there has been no determination of the structure of the tetragonal form of CeNbO₄. From our initial X-ray diffraction results, it was immediately apparent that the data recorded at 1073 K conformed to a scheelite-type structure. As a starting model for the Rietveld refinement of the neutron powder diffraction data, the structure of scheelite (CaWO₄) was therefore used [space group I4₁/a; alternate setting with origin at (0, 1/4, 1/8); the Ce and Nb atoms in the 4 b and 4a positions, respectively, and the O atom in the 16f position (Hazen et al., 1985)]. The refinements were carried out using the GSAS package (Larson & Von Dreele, 1994) and provided good fits to the 90° and backscattered neutron diffraction data collected in air and in vacuo (10⁻⁵ mbar) (Fig. 1).

In common with other scheelite-type compounds, the Ce atoms in CeNbO₄ are in an eightfold coordination environment, with two distinct bond lengths, while the Nb environment is tetrahedral, with equal Nb—O bonds (Fig. 2). The slightly larger unit-cell volume of the sample heated in static air is associated with slightly longer Nb—O and Ce—O distances. The distorted tetrahedral environment for the Nbatom, with O—Nb—O angles of 106.476 (19) and 115.64 (4)°, is comparable to that in isostructural LaNbO₄ (David, 1983; Machida et al., 1995). The transition temperature in CeNbO₄ is significantly higher than in LaNbO₄ (773 K).

From the results of both refinements and from Fourier difference maps, it is apparent that the high-temperature CeNbO₄ polymorph is fully stoichiometric, and there is no evidence to suggest that heating it in air introduced any interstitial oxygen. Because the sample was heated in situ while packed in a vanadium can, it is conceivable that limited oxidation occurred at the surface only. Hence it would be desirable either to perform a series of measurements at one temperature over a period of time to enable the study of possible oxygen incorporation into CeNbO₄ or to oxidize a sample before carrying out a set of measurements at temperatures up to 1073 K, allowing any oxygen stoichiometry variations on heating to be investigated.

S2. Experimental

Samples were produced through the mixing of CeO₂ (99.9%, Sigma Aldrich) and Nb₂O₅ (99.99%, Sigma Aldrich) in an agate mortar and pestle. These starting materials were mixed thoroughly under acetone and allowed to dry in air before the stoichiometric mixture was transferred to an alumina crucible for heat treatment. The mixture was prefired at 1273 K for 10 h and allowed to cool. After further mixing, the material was heated at 1673 K for 16 h and quenched. This procedure produced a bright-green material. X-ray powder diffraction data were recorded on a Philips PW1700 series diffractometer using Cu Kα₁ radiation and a graphite crystal as secondary monochromator. Neutron powder diffraction data were collected from 6 g samples contained in vanadium cans on the POLARIS diffractometer using both 90° and backscatter detectors at ISIS, CCLRC Rutherford Appleton Laboratory, Oxfordshire, England. Each dataset was collected over a period of 90 min, and then normalized and corrected in the usual way. Two separate samples were examined during these measurements. The first sample was heated in situ under vacuum with data recorded on heating and cooling. For the second sample, the same in situ heating regime was used but the can was exposed to static air. The data were collected at 296 K, every 100 K from 473 to 973 K, and every 50 K from 973 to 1123 K. On cooling, data were recorded at 1023, 823 and 623 K. Only the data recorded at 1073 K from the samples heated in air and in vacuo are included here.

S3. Refinement

Initial refinement cycles showed no significant deviation from stoichiometry, and therefore the occupancies of the Ce and Nb sites were fixed to unity in the final cycles for both the air and the vacuum data sets. The occupancy of the O site was allowed to vary but did not deviate significantly from unity either. Fourier difference maps indicated no significant residual scattering density within the unit cell, with maxima of 0.489 Å⁻³ at (0.2465 0.5004 0.1234) for the vacuum data and 0.459 Å⁻³ at (0.1545 0.9652 0.1803) for the air data.

Computing details

For both compounds, program(s) used to solve structure: GSAS; program(s) used to refine structure: GSAS.

Figures

Fig. 1. (a) A Rietveld plot for the refinement of backscattered neutron diffraction data in vacuo. The experimental and calculated profiles are represented by crosses and a full line, respectively. The difference plot is shown underneath and tick marks indicate the positions of Bragg reflections. (b) A Rietveld plot for the refinement of the 90° bank neutron diffraction data in air. Profiles are drawn as in (a).

Fig. 2. Representation of the tetragonal scheelite-type structure of CeNbO4, showing the Nb and Ce coordination environments. Displacement ellipsoids are drawn at the 50% probablity level. [Symmetry code: (i) x + 1/2, y + 1/2, 1/2 − z.]

(vacuum) cerium niobium tetraoxide

Crystal data

CeNbO4	Z = 4
Mr = 297.02	Dx = 5.905 Mg m^−3
Tetragonal, I41/a,originchoice2at0,1/4,1/8from4	Neutron radiation
Hall symbol: -I 4a	T = 1073 K
a = 5.37119 (8) Å	light green
c = 11.58104 (18) Å	?, ? × ? × ? mm
V = 334.11 (1) Å^3

Data collection

Polaris diffractometer at ISIS	2θfixed = 145^o for backscatter detector (average)
Radiation source: spallation neutron source	Distance from source to specimen: 12.0 m mm
Specimen mounting: 6 mm diameter vanadium can	Distance from specimen to detector: 0.80 m for backscatter detector mm
Scan method: time of flight

Refinement

Refinement on F^2	4565 data points
Least-squares matrix: full	Profile function: exponential pseudo-Voigt convolution
Rp = 0.020	49 parameters
Rwp = 0.012	1/Yi
Rexp = 0.007	(Δ/σ)max = 0.02
χ^2 = 3.276	Background function: shifted Chebyschev

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

	x	y	z	Uiso*/Ueq	Occ. (<1)
Ce	0.0	0.25	0.625	0.0173 (5)
Nb	0.0	0.25	0.125	0.0174 (4)
O	0.16181 (9)	0.49329 (12)	0.21018 (5)	0.0304 (3)	1.0000 (22)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Ce	0.0195 (4)	0.0195 (4)	0.0130 (6)	0	0	0
Nb	0.0140 (3)	0.0140 (3)	0.0241 (5)	0	0	0
O	0.0326 (3)	0.0326 (3)	0.0257 (3)	−0.0142 (4)	0.0075 (3)	−0.0108 (2)

Geometric parameters (Å, º)

Ce—Ce^i	3.9490 (1)	Ce—O^iii	2.4847 (6)
Ce—O^ii	2.5100 (6)	Nb—O	1.8537 (5)
O^iv—Ce—O^iii	73.03 (1)	O^iii—Ce—O^vi	133.21 (3)
O^iv—Ce—O^ii	125.33 (2)	O—Nb—O^ix	106.45 (2)
O^iv—Ce—O^v	75.51 (2)	O—Nb—O^x	115.70 (4)
O^iv—Ce—O^vi	153.59 (3)	Ce^xi—O—Ce^xii	104.49 (2)
O^iv—Ce—O^vii	80.98 (3)	Ce^xi—O—Nb	121.51 (3)
O^iv—Ce—O^viii	69.29 (2)	Ce^xii—O—Nb	129.02 (2)
O^iii—Ce—O^v	99.06 (1)

Symmetry codes: (i) −x, −y, −z+1; (ii) −x, −y+1, −z+1; (iii) −x+1/2, −y+1, z+1/2; (iv) −y+3/4, x+1/4, z+1/4; (v) y−3/4, −x+3/4, −z+3/4; (vi) x−1/2, y−1/2, z+1/2; (vii) y−3/4, −x+1/4, z+1/4; (viii) −y+3/4, x−1/4, −z+3/4; (ix) y−1/4, −x+1/4, −z+1/4; (x) −x, −y+1/2, z; (xi) −x+1/2, −y+1, z−1/2; (xii) y−1/4, −x+3/4, z−1/4.

(air) cerium niobium tetraoxide

Crystal data

CeNbO4	Z = 4
Mr = 297.02	Dx = 5.835 Mg m^−3
Tetragonal, I41/a,originchoice2at0,1/4,1/8from4	Neutron radiation
Hall symbol: -I 4ad	T = 1073 K
a = 5.37692 (8) Å	light green
c = 11.59514 (18) Å	?, ? × ? × ? mm
V = 335.23 (1) Å^3

Data collection

Polaris diffractometer at ISIS	2θfixed = 145^o for backscatter detector (average)
Radiation source: spallation neutron source	Distance from source to specimen: 12.0 m mm
Specimen mounting: 6 mm diameter vanadium can	Distance from specimen to detector: 0.80 m for backscatter detector mm
Scan method: time of flight

Refinement

Refinement on F^2	Profile function: exponential pseudo-Voigt convolution
Least-squares matrix: full	49 parameters
Rp = 0.020	0 restraints
Rwp = 0.012	1/Yi
Rexp = 0.007	(Δ/σ)max = 0.03
χ^2 = 3.349	Background function: shifted Chebyschev
4565 data points

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

	x	y	z	Uiso*/Ueq	Occ. (<1)
Ce	0.0	0.25	0.625	0.0169 (4)
Nb	0.0	0.25	0.125	0.0173 (3)
O	0.16173 (9)	0.49318 (12)	0.21021 (5)	0.0303 (3)	0.993 (3)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Ce	0.0188 (4)	0.0188 (4)	0.0131 (7)	0	0	0
Nb	0.0139 (3)	0.0139 (3)	0.0244 (6)	0	0	0
O	0.0322 (4)	0.0319 (3)	0.0268 (3)	−0.0145 (4)	0.0075 (3)	−0.0106 (3)

Geometric parameters (Å, º)

Ce—Ce^i	3.9536 (1)	Ce—O^iii	2.4882 (6)
Ce—O^ii	2.5128 (6)	Nb—O	1.8553 (5)
O^iv—Ce—O^iii	73.03 (1)	O^iii—Ce—O^vi	133.20 (3)
O^iv—Ce—O^ii	125.32 (2)	O—Nb—O^ix	106.47 (2)
O^iv—Ce—O^v	75.53 (2)	O—Nb—O^x	115.64 (4)
O^iv—Ce—O^vi	153.60 (3)	Ce^xi—O—Ce^xii	104.47 (2)
O^iv—Ce—O^vii	80.99 (3)	Ce^xi—O—Nb	121.48 (3)
O^iv—Ce—O^viii	69.27 (2)	Ce^xii—O—Nb	129.05 (2)
O^iii—Ce—O^v	99.07 (1)

Symmetry codes: (i) −y−1/4, x+1/4, z+1/4; (ii) −x, −y+1, −z+1; (iii) −x+1/2, −y+1, z+1/2; (iv) −y+3/4, x+1/4, z+1/4; (v) y−3/4, −x+3/4, −z+3/4; (vi) x−1/2, y−1/2, z+1/2; (vii) y−3/4, −x+1/4, z+1/4; (viii) −y+3/4, x−1/4, −z+3/4; (ix) y−1/4, −x+1/4, −z+1/4; (x) −x, −y+1/2, z; (xi) −x+1/2, −y+1, z−1/2; (xii) y−1/4, −x+3/4, z−1/4.