Hostname: page-component-745bb68f8f-hvd4g Total loading time: 0 Render date: 2025-02-04T03:40:48.216Z Has data issue: false hasContentIssue false
  • English
  • Français

Structural and Raman spectroscopic studies of the two M0.50SbFe(PO4)3 (M = Mg, Ni) NASICON phases

Published online by Cambridge University Press:  02 May 2017

Abderrahim Aatiq ,
Asmaa Marchoud ,
Hajar Bellefqih  and
My Rachid Tigha
Abderrahim Aatiq*
Affiliation:
Département de Chimie, Laboratoire de Physico-Chimie des Matériaux Appliqués, Université Hassan II de Casablanca, Faculté des Sciences Ben M'Sik, Avenue Idriss El harti, B.P. 7955, Casablanca, Morocco
Asmaa Marchoud
Affiliation:
Département de Chimie, Laboratoire de Physico-Chimie des Matériaux Appliqués, Université Hassan II de Casablanca, Faculté des Sciences Ben M'Sik, Avenue Idriss El harti, B.P. 7955, Casablanca, Morocco
Hajar Bellefqih
Affiliation:
Département de Chimie, Laboratoire de Physico-Chimie des Matériaux Appliqués, Université Hassan II de Casablanca, Faculté des Sciences Ben M'Sik, Avenue Idriss El harti, B.P. 7955, Casablanca, Morocco
My Rachid Tigha
Affiliation:
Département de Chimie, Laboratoire de Physico-Chimie des Matériaux Appliqués, Université Hassan II de Casablanca, Faculté des Sciences Ben M'Sik, Avenue Idriss El harti, B.P. 7955, Casablanca, Morocco
*
a)Author to whom correspondence should be addressed. Electronic mail: a_aatiq@yahoo.fr
Rights & Permissions [Opens in a new window]

Abstract

Structures of the two M0.50SbFe(PO4)3 (M = Mg, Ni) phases, abbreviated as [Mg0.50] and [Ni0.50], were determined at room temperature from X-ray diffraction (XRD) powder data using the Rietveld analysis. Both compounds belong to the NASICON structural family. XRD patterns of [Mg0.50] and [Ni0.50] phases were easily indexed with a primitive hexagonal unit cell [P$\overline 3 $ space group, Z = 6] similar to that already obtained for La0.33Zr2(PO4)3. Obtained unit cells parameters are [a = 8.3443(1) Å, c = 22.3629(1) Å], and [a = 8.3384(1), c = 22.3456(1) Å], respectively, for [Mg0.50] and [Ni0.50] phosphates. In both samples, the [Sb(Fe)(PO4)3]− NASICON framework is preserved and a partially-ordered distribution of Sb5+ and Fe3+ ions is observed. Raman spectroscopic study was used to obtain further structural information about the nature of bonding in [Mg0.50] and [Ni0.50] phases.

Keywords

iron and antimony phosphateNASICON structureRaman spectroscopyRietveld analysis
Type
Technical Articles
Information
Powder Diffraction , Volume 32 , Supplement S1 , September 2017 , pp. S40 - S51
Check for updates
Copyright
Copyright © International Centre for Diffraction Data 2017 

I. INTRODUCTION

The NASICON [acronym for sodium (Na) Super Ionic CONductor] type family has been the subject of intensive research owing to its potential applications as solid electrolyte, electrode material, low thermal expansion ceramics, and as storage materials for nuclear waste (Delmas et al., Reference Delmas, Viala, Olazcuaga, Le Flem, Hagenmuller, Cherkaoui and Brochu1981; Roy et al., Reference Roy, Vance and Alamo1982; Padhi et al., Reference Padhi, Nanjundaswamy, Masquelier and Goodenoogh1997; Woodcock et al., Reference Woodcock, Lightfoot and Smith1999; Aatiq et al., Reference Aatiq, Ménétrier, El Jazouli and Delmas2002; Anuar et al., Reference Anuar, Adnan and Mohamed2014). The structure of such materials with general formula M x XX′(PO4)3 consists of a three-dimensional (3D) network made up of corner-sharing X(X′)O6 octahedra and PO4 tetrahedra in such a way that each octahedron is surrounded by six tetrahedra and each tetrahedron is connected to four octahedra. During the past few years, particular attention has been paid to combining titanium or (antimony and iron) and divalent 3d transition metals M 2+. Most investigations on Nasicon phases with general formula M 0.50Ti2(PO4)3 (M = Co, Fe, Mg, Mn, Ca) (Barth et al., Reference Barth, Olazcuaga, Gravereau, le Flem and Hagenmüller1993; El Bouari et al., Reference El Bouari, El Jazouli, Dance, Le Flem and Olazcuaga1994; Aatiq et al., Reference Aatiq, Ménétrier, El Jazouli and Delmas2002; Benmokhtar et al., Reference Benmokhtar, El Jazouli, Aatiq, Chaminade, Gravereau, Wattiaux, Fournes and Grenier2007), and M 0.50SbFe(PO4)3 (M = Mn, Cd, Ca, Sr, Pb) (Aatiq et al., Reference Aatiq, Hassine, Tigha and Saadoune2005, Reference Aatiq, Tigha, Hassine and Saadoune2006, Reference Aatiq, Tigha and Benmokhtar2012, Reference Aatiq, Tigha, Fakhreddine and Marchoud2015) have employed powder X-ray diffraction (XRPD) techniques to probe the global structure of the materials using the Rietveld analysis. For both M0.50SbFe(PO4)3 and M0.50Ti2(PO4)3 NASICON phases series, the structure was already well established when the ionic radii of M 2+ cations was relatively high and more precisely equal or larger than that of Mn2+ ions [i.e., M = Mn, Ca, Cd, Pb, Sr]. Note that all M 0.50SbFe(PO4)3 (M = Mn, Ca, Cd, Pb, Sr) compounds crystallise in the R $\overline 3 $ (No. 148) space group with a practically ordered cationic distribution of M 2+ cations within the M1 sites. In fact, the same cationic distribution was observed for M 0.50Ti2(PO4)3 (M = Mn, Ca, Cd, Pb) phases. On the contrary, up to our knowledge, it is noted that an uncertainty exists about the exact symmetry for M 0.50Ti2(PO4)3 materials containing M 2+ cations with a relatively small size [e.g., M = Mg, Co]. For example, Barth et al. (Reference Barth, Olazcuaga, Gravereau, le Flem and Hagenmüller1993) proposed the R $\overline 3 $ c (No. 167) as the space group for Mg0.50Ti2(PO4)3. During their structural discussion, they pointed out a serious problem concerning the two low-intensity reflections observed, in the XRD pattern, respectively at around 12.40° and 12.66° in 2θ Cu. Indeed, the origin of these lines does not derive from the method of preparation. Particularly these two lines of XRD diffraction are not generated by the R $\overline 3 $ c space group and also cannot be well attributed to the (101) and/or (003) reflections, which are characteristic of a possible reduction of symmetry from R $\overline 3 $ c to R $\overline 3 $ or R32 one. Note that the same problem has been raised by El Bouari et al. (Reference El Bouari, El Jazouli, Dance, Le Flem and Olazcuaga1994) for Co0.50Ti2(PO4)3 phase who have suggested, in a first time, the refinement in R $\overline 3 $ c space group even if the two low-intensity reflections cited previously are not generated by this space group. In fact, 5 years later and despite the non concordance between the experimental and calculated diffraction lines at around 12.40° and 12.66° in 2θ Cu, a reinvestigation of this orthophosphate using optical and magnetic properties leads them to assign in instead the R32 or R $\overline 3 $ space groups (Derouet et al., Reference Derouet, Beaury, Porcher, Olazcuaga, Dance, Le Flem, El Bouari and El Jazouli1999; Olazcuaga et al., Reference Olazcuaga, Dance, Le Flem, Derouet, Beaury, Porcher, El Bouari and El Jazouli1999). Ten years ago, the reinvestigation of the structure of La0.33Zr2(PO4)3 NASICON phases from Neutron and XRDP shows particularly that the reflection around 15° in 2θ Cu, which was not generated in the R $\overline 3 $ c space group (Alami et al., Reference Alami Talbi, Brochu, Parent, Rabardel and Le Flem1994) is well indexed in the P $\overline 3 $ (No. 147) space group (Barré et al., Reference Barré, Crosnier-Lopez, Le Berre, Emery, Suard and Fourquet2005, Reference Barré, Le Berre, Crosnier-Lopez, Bohnke´, Emery and Fourquet2006, Reference Barré, Crosnier-Lopez, Le Berre, Suard and Fourquet2007; Crosnier-Lopez et al., Reference Crosnier-Lopez, Barré, Le Berre and Fourquet2006). Thermal study reveals that La0.33Zr2(PO4)3 exhibits at high temperature a structural transition from P $\overline 3 $ to P $\overline 3 $ c (No. 165) space groups, which was related to the La3+ ion ability to move in the Zr2(PO4)3 framework despite its ionic size (Barré et al., Reference Barré, Le Berre, Crosnier-Lopez, Bohnke´, Emery and Fourquet2006). The P $\overline 3 $ -P $\overline 3 $ c phase transition can be explained by the fact that the P $\overline 3 $ space group is a Maximal translationengleiche subgroup of P $\overline 3 $ c one (Hans and Ulrich, Reference Hans and Ulrich2004). On the other hand, a structural study of Li1−x La x/3Zr2(PO4)3 (0 ≤ x ≤ 1) solid solution has shown reversible transition, which is clearly soft and strongly depends of the x value (Barre´ et al., Reference Barré, Crosnier-Lopez, Le Berre, Suard and Fourquet2007).

Recently, structural characteristics by powder X-ray diffraction (XRD) study using the Rietveld method for M 0.50SbFe(PO4)3 (M = Mn, Cd, Ca, Sr, Pb) showed that all samples crystallise in the R $\overline 3 $ space group and the M 2+ ions occupied practically one-half of the M1 sites. The Sb5+ and Fe3+ cations are also orderly distributed within the SbFe(PO4)3 NASICON framework (Aatiq et al., Reference Aatiq, Hassine, Tigha and Saadoune2005, Reference Aatiq, Tigha, Hassine and Saadoune2006, Reference Aatiq, Tigha and Benmokhtar2012, Reference Aatiq, Tigha, Fakhreddine and Marchoud2015). Note that because of the fast diffusion of Na+ ions in the NASICON framework, the M 0.5SbFe(PO4)3 phases can be tested as Negative Electrode materials in Na-Ion and Li-Ion Batteries as was more recently reported as example for the Mg0.5Ti2(PO4)3 NASICON phase (Zhao et al., Reference Zhao, Wei, Pang, Wei, Cai, Fu, Du, Sarapulova, Ehrenberg, Liu and Chen2017).

In a continuation of our search concerning NASICON-like phases herein we report the results of the structural investigation by the Rietveld refinement using the XRD patterns of the two M0.50SbFe(PO4)3 (M = Mg, Ni) phases, abbreviated as [Mg0.50] and [Ni0.50]. During this work we examined some additional reflections, which appeared in the experimental XRD patterns. It should be noticed that the observed small additional reflections within the XRD spectra cannot be indexed successfully with either R $\overline 3 $ c or R $\overline 3 $ space groups. In order to obtain further structural information about the nature of bonding in the two [Mg0.50] and [Ni0.50] phosphates, a Raman spectroscopic study involving the factor group analysis was also undertaken.

II. EXPERIMENTAL

Syntheses of M0.50SbFe(PO4)3 (M = Mg, Ni) phases were carried out using conventional solid-state reaction techniques. Powder crystalline samples were prepared from mixtures of oxides MO (M = Mg, Ni) (Riedel-de Haën, 99%), Sb2O3 (Riedel-de Haën, 99.9 %), Fe2O3 (Prolabo, 99%), and NH4H2PO4 (Riedel-de Haën, 99%) in stoichiometric proportions. The mixture was heated progressively with intermittent grinding at 200 °C (12 h), 400 °C (6 h), 600 °C (12 h), 800 °C (24 h), 900 °C (24 h), and 950 °C (24 h) in air atmosphere. The products of reaction were characterised by XRD at room temperature with a Panalytical X′Pert-PRO (θ–2θ) diffractometer equipped with x'celerator detector; (CuKα) radiation (45 kV, 40 mA); divergence slit of 1°, receiving slit of 0.10 mm, and antiscatter slit of 1°. The data were collected from 10° to 90°2θ, in steps of 0.02°, with a counting time of 15 s per step. The Rietveld refinement of the structure was performed using the Fullprof program (Rodriguez-Carvajal, Reference Rodriguez-Carvajal1997).

The Raman spectra are recorded on RENISHAW 1000B spectrometer in the wave number range 100–1500 cm−1. All the spectra have been recorded at room temperature.

III. RESULTS AND DISCUSSION

A. Structure of M 0.50SbFe(PO4)3 (M = Mg, Ni) phases

Analysis of the XRPD spectra for the two M 0.50SbFe(PO4)3 (M = Mg, Ni) materials indicated that the principal peak positions of the XRD lines are similar to those observed for other NASICON-type phases. At first sight, it seems that the existing low-intensity lines in the XRD pattern of [Mg0.50] and [Ni0.50] phases can be indexed in the usual R $\overline 3 $ c or R $\overline 3 $ space groups but a LeBail fit (profile matching) program (LeBail et al., Reference Le Bail, Duroy and Fourquet1988), realised for both materials and in the two possible space groups, shows clearly that some additional reflections with a relatively low intensity are not indexed. A comparison between experimental, calculated and difference profile of the XRD pattern for [Ni0.50] phase, in the selected 10°–17° angular range in 2θ Cu, are given as example in Figure 1. In fact, the initial findings show clearly that in the R $\overline 3 $ c space group several lines of XRD diffraction, which are indicated by blue stars, are not indexed [Figure 1(a)]. Note also that the two usually known (101) and (003) reflections, which are characteristic of the R $\overline 3 $ space group are not a good fit to the experimental pattern [Figure 1(b)]. In regards to our tested LeBail fit refinement in both R $\overline 3 $ c and R $\overline 3 $ space group, obtained results are comparable with those already signalised for M 0.50Ti2(PO4)3 (M = Mg, Co) NASICON phases (Barth et al., Reference Barth, Olazcuaga, Gravereau, le Flem and Hagenmüller1993; El Bouari et al., Reference El Bouari, El Jazouli, Dance, Le Flem and Olazcuaga1994). Consequently, it is clear that the two rhombohedral R $\overline 3 $ c or R $\overline 3 $ space groups usually encountered for many NASICON–type phases are excluded here. On the other hand, given that the P $\overline 3 $ (No. 147) space group is a Maximal klassengleiche subgroup of R $\overline 3 $ (No. 148) as a result of the loss of centring translations (Hans and Ulrich, Reference Hans and Ulrich2004), and according to the structural study recently realised for similar phosphates with NASICON like phases such as La0.33Zr2(PO4)3 (Barré et al., Reference Barré, Crosnier-Lopez, Le Berre, Emery, Suard and Fourquet2005), the model of the LeBail refinement in the P $\overline 3 $ space group is tested. In this case, a good concordance between the experimental and calculated profile of the XRD pattern is obtained [Figure 1(c)]. In fact, the non-indexed diffraction lines in the two rhombohedral space groups R $\overline 3 $ c or R $\overline 3 $ are generated by the P $\overline 3 $ space group. In the case of Ln 0.33Zr2(PO4)3 (Ln = Ce, Eu, Yb) phases, the already obtained results from the Rietveld refinement study, using the powder XRD data, have shown that they were indexed in the P $\overline 3 $ c space group (Bykov et al., Reference Bykov, Gobechia, Kabalov, Orlova and Tomilin2006). It should be noticed that the P $\overline 3 $ space group is also a Maximal translationengleiche subgroup of P $\overline 3 $ c one. Hence, a comparison between experimental, calculated and difference profile of the XRD pattern for the selected [Mg0.50] phase, in both P $\overline 3 $ and P $\overline 3 $ c space group, are realised and given as example in the selected 18°–42° (2θ Cu) angular range (Figure 2). For more details, the experimental, calculated and difference profile of the XRD pattern for [Ni0.50], in the selected 10°–17° (2θ Cu) angular range, are also given for comparison in Figure 1(d). In fact, because of the loss of c-plane in transforming from space group P $\overline 3 $ c to the space group P $\overline 3 $ , some diffraction lines of the experimental XRD patterns indicated in the Figure 2 by blue stars are indexed only in the P $\overline 3 $ space group. It should be noticed that all obtained results, from LeBail fit, agree well with the choice of P $\overline 3 $ space group for both [Mg0.50] and [Ni0.50] materials.

Figure 1. (Color online) Comparison of the experimental (•••) calculated (—), and difference profile of the XRD pattern of Ni0.50SbFe(PO4)3. The LeBail refinements of cell parameters were made for (a) R $\overline 3 $ c, (b) R $\overline 3 $ , (c) P $\overline 3 $ , and (d) P $\overline 3 $ c space groups over the range 10°–17° (2θ Cu). The non-indexed XRD diffraction lines are indicated by blue stars.

Figure 2. (Color online) Comparison of the experimental (•••) calculated (—), and difference profile of the XRD pattern of Mg0.50SbFe(PO4)3. The LeBail refinements of cell parameters were made for (a) P $\overline 3 $ c, and (b) P $\overline 3 $ space groups over the range 18°–42° (2θ Cu). The non-indexed XRD diffraction lines are indicated by blue stars.

Given that there is an important number of parameter to be refined during different steps of the Rietveld refinement in the P $\overline 3 $ space group, a strategy of structural refinement was followed. Note that in La0.33Zr2(PO4)3 there is only one type of atom (i.e., one Zr atom) per X site, therefore the determination of the amount of existing cation in each X site within the [X2(PO4)3] framework was not raised. Indeed, in our case, this problem should be resolved before starting the Rietveld refinement in the P $\overline 3 $ space group. In a first time, the Rietveld refinement was carried out in R $\overline 3 $ space group and the structural parameters of Mn0.50SbFe(PO4)3 (Aatiq et al., Reference Aatiq, Hassine, Tigha and Saadoune2005) were chosen as starting model. In fact, despite the none perfect concordance for some reflections of low intensity in the XRD spectra for both [Mg0.50] and [Ni0.50] materials, the obtained results of refinements in R $\overline 3 $ space group will give particularly interesting data about the future Sb5+ and Fe3+ cationic distribution within the [SbFe(PO4)3] NASICON frameworks when the refinement is realised in the P $\overline 3 $ space group. Results of refinement, in R $\overline 3 $ space group, show that Ni2+ in [Ni0.50] and Mg2+ in [Mg0.50] are distributed over M1 sites in an ordered manner along the c-axis. They are located in 3b sites (0 0 1/2) with a complete occupancy, 3a sites (0 0 0) remain vacant. For [Mg0.50], the obtained occupancy rate of Sb5+ and Fe3+ cations, within the two possible 6c sites are, respectively, 1.20(1)/0.80(1) for Sb5+/Fe3+ in X′(1) (0 0–0.14) position, and 0.80(1)/1.20(1) for the Sb5+/Fe3+ in X′(2) (0 0–0.64) one. In the case of [Ni0.50] phase, obtained cationic distribution are 1.50(1)/0.50(1) for Sb5+/Fe3+ in the X′(1) position and 0.50(1)/1.50(1) for Sb5+/Fe3+ in the X′(2) one. Note that during the Rietveld refinement in P $\overline 3 $ space group, it is not easy to determine without ambiguity the occupancy rate of Sb5+ and Fe3+ cations within every X site position among the six existing one [i.e., X(1), X(2), X(3), X(4), X(5), X(6)]. So, realistic occupancy rates of Sb(Fe) cations within the framework can be deduced from the values obtained during the refinement in R $\overline 3 $ space group. In the following, obtained structural parameters of our Rietveld refinement, in R $\overline 3 $ space group, were used as starting structural parameters for refinements in the P $\overline 3 $ space group but after a transposition from R $\overline 3 $ to P $\overline 3 $ space groups. The corresponding splitting of possible atomic positions was given in Table I. As was shown in the Figure 3 and according to the data mentioned in Table I, this last point was also clarified by comparing projections of structures of the two NASICON-type phases, M 0.50 X′2(PO4)3 refined in the R $\overline 3 $ space group [Figure 3(a)] and the M 0.50 X 2(PO4)3, which will be refined later in the P $\overline 3 $ one [Figure 3(b)]. On the other hand, by examining the large number of structural parameters that need to be refined in the P $\overline 3 $ space group for [Mg0.50] and [Ni0.50], a realistic convergence of the refinement is reached by using soft-constrained P–O and Sb(Fe)–O distances [i.e., 1.52(1) and 2.00(1) Å for P–O and Sb(Fe)–O, respectively]. By performing the R $\overline 3 $ -P $\overline 3 $ transposition of atomic positions outlined in Table I, we can conclude that the three M 2+ cations of the unit cell are preferentially located as follows: one in the M(1b) (0, 0, 0.5) site and the two other ones are in the M(2d′) (1/3, 2/3, ~0.17) site. Both M(1b) and M(2d′) sites correspond to the M1(3b) sites frequently reported in NASICON structure with R $\overline 3 $ space group (Table I and Figure 3). As it will be shown later, this cationic distribution is in good agreement with the obtained M 2+-O (M = Mg, Ni) distance values. The final reliability factors and atomic parameters for [Mg0.50] and [Ni0.50] phases are summarised in Tables II and III, respectively. A comparison of the experimental and calculated XRD profiles, after Rietveld refinement, of [Mg0.50] and [Ni0.50] is shown in Figures 4 and 5, respectively. A view of the structure of M 0.50Sb(Fe)(PO4)3 (M = Mg, Ni) NASICON phases is shown in Figure 6. Mg–O distances in [Mg0.50] [i.e., 6 × 2.271(5) Å for Mg(1b)–O(10); 3 × 2.240(6) Å for Mg(2d′)–O(11) and 3 × 2.232(5) Å for Mg(2d′)–O(12)] and Ni–O distances in [Ni0.50] [i.e., 6 × 2.272(6) Å for Ni(1b)–O(10); 3 × 2.234(5) Å for Ni(2d′)–O(11) and 3 × 2.227(5) Å for Ni(2d′)–O(12)] are relatively consistent with the sum values of the corresponding crystal radii, which are of 2.26 Å for Mg–O and 2.23 Å for Ni–O. In contrast, analysis of distance values between the centre of each 1a and 2d sites and oxygen atoms surrounding them, confirms the fact that these two sites are normally assumed to be empty. In fact, for [Mg0.50], values of these distances are: 6 × □(1a)-O(7) = 2.524(5), 3 × □(2d)–O(8) = 2.444(6) and 3 × □(2d)-O(9) = 2.461(5) Å. The corresponding distances values for [Ni0.50] are : 6 × □(1a)-O(7) = 2.513(5) Å, 3 × □(2d)-O(8) = 2.879(6) and 3 × □(2d)-O(9) = 2.880(5) Å. X–O [X = Sb(Fe)] and P–O distances values for both materials are grouped in Tables IV and V, respectively. Sb(Fe)–O distances-types are relatively consistent with the crystal radii values in six coordination of Sb5+ and Fe3+ ions (Shannon, Reference Shannon1976). P–O distances values match well with those typically observed in NASICON-type phosphates. In order to have more structural information, the bond valence sum (BVS) based on bond strength analysis (Brown and Altermatt, Reference Brown and Altermatt1985) was also computed. As shown in Tables IV and V, the BVS values calculated for, Fe, Sb, and P sites are relatively consistent with the expected formal oxidation state of Fe3+, Sb5+, and P5+ ions. XRPD data, obtained from the “observed intensities” of the Rietveld refinement (CuKα1: 1.540 56 Å), of [Mg0.50] and [Ni0.50] phases are given in Tables VI and VII, respectively.

Figure 3. (Color online) Projections of the structure for (a) M0.50X′2(PO4)3 (R $\overline 3 $ space group), and (b) M 0.50 X 2(PO4)3 (P $\overline 3 $ space group) along the b-axis. The different M and X sites, accompanying the R $\overline 3 $ -P $\overline 3 $ transposition of atomic positions, have been labelled and located.

Figure 4. (Color online) Experimental (•••) calculated (—), and difference profile, obtained after Rietveld refinement, of the XRD pattern of Mg0.50SbFe(PO4)3.

Figure 5. (Color online) Experimental (•••) calculated (—), and difference profile, obtained after Rietveld refinement, of the XRD pattern of Ni0.50SbFe(PO4)3.

Figure 6. (Color online) View of the structure for M 0.50 X 2(PO4)3 [M = Mg, Ni; X = Sb(Fe)] NASICON-type phases.

Table I. R $\overline 3 $ -P $\overline 3 $ transposition of the atomic positions from M 0.50 X′2(PO4)3 (R $\overline 3 $ space group) to M 0.50 X 2(PO4)3 (P $\overline 3 $ space group) NASICON phases.

Table II. Results of the Rietveld refinement of Mg0.50SbFe(PO4)3.

Table III. Results of the Rietveld refinement of Ni0.50SbFe(PO4)3.

Table IV. Selected interatomic distances (Å) and calculated bond valence sum (BVS) for Mg0.50SbFe(PO4)3.

Table V. Selected interatomic distances (Å) and calculated bond valence sum (BVS) for Ni0.50SbFe(PO4)3.

Table VI. Powder diffraction data of Mg0.50SbFe(PO4)3 (CuKα 1; λ = 1.540 56 Å).

Table VII. Powder diffraction data of Ni0.50SbFe(PO4)3 (CuKα 1; λ = 1.540 56 Å).

It is well known that in the R $\overline 3 $ space group there is an ordered distribution, along the c-axis, between the occupied M1(3b) sites and the empty □M1(3a) ones [i.e., M(3b)-□(M(3a)-M(3b)]. Similar cationic distribution of M cations in both [M 0.50] phases (M = Ni, Mg) between the vacancies □M(1a) sites and the occupied M(1b) one [i.e., M(1b)-□(M(1a)-M(1b)] is obtained during the structural refinements in the P $\overline 3 $ space group (Figure 7). In fact, in addition to this last-ordered cationic distribution, results of refinement show also a second-ordered distribution along the c-axis between M cations in the fully occupied M(2d′) sites and the vacancies □M(2d) ones [i.e., M(2d′)-□(M(2d)–M(2d′)] (Figure 7). Note that one can switch from the network formed by the M(1a) and M(1b) sites to the other, which is formed by the M(2d) and M(2d′) sites by a translation vector of (1/3, 2/3, ~1/6). Taking into account the above remarks, we can admit that the presence of a double-ordered cationic distribution between the M cations and vacancies, within the four possible □M(1a), M(1b), □M(2d) and M(2d′) positions of M1 sites, should be considered as the driving force for the unusual distortion observed when the space group switch from the R $\overline 3 $ space group to the P $\overline 3 $ one.

Figure 7. (Color online) Structure of M 0.50SbFe(PO4)3 (M = Mg, Ni) phases showing the ordered cationic distribution along the c-axis, between M cations and vacancies, of the four possible □M(1a), M(1b), □M(2d), and M(2d′) positions of the M1 sites. Sb(Fe) atoms are blue circles (•), P(1)O4 are yellow, P(2)O4 are blue, and the P(3)O4 are green.

B. Raman spectroscopic study

In order to obtain further structural information about the nature of bonding in [Mg0.50] and [Ni0.50] materials, in this part of the paper, a Raman spectroscopic study was undertaken. The Raman spectrum of Ca0.50SbFe(PO4)3 , abbreviated as [Ca0.50], is given here for comparison. Given that the NASICON structure contains both isolated PO4 groups and isolated Sb(Fe)O6 groups, the vibrational pattern is obviously typical of an orthophosphate. The vibrational modes of tetrahedral PO4 molecules are well known (Nakamoto, Reference Nakamoto1986) and generally the IR and Raman spectroscopic study of orthophosphate shows that phosphate group vibrations are strong compared with the lattice modes and metal-oxygen vibrations. The two [Mg0.50] and [Ni0.50] Raman spectra are similar but both are slightly different from that of the [Ca0.50] Raman spectrum (Figure 8). The relatively broadening Raman bands observed in [Mg0.50] and [Ni0.50] spectra [Figures 8(a) and 8(b)], in comparison with that found in the [Ca0.5] Raman one [Figure 8(c)], may be easily related to the difference between the cationic distribution within both NASICON framework-types. In fact, both Ca2+ ions in M1 sites and Sb5+(Fe3+) within the Ca0.5SbFe(PO4)3 NASICON framework are relatively orderly distributed (Aatiq et al., Reference Aatiq, Tigha, Hassine and Saadoune2006). Consequently, in this last material, every PO4 group is surrounded by two FeO6 octahedra and two SbO6 octahedra. In the case of [Mg0.50] and [Ni0.50], the large Sb5+(Fe3+) cationic distribution within every X site of the M 0.50 X 2(PO4)3 (M = Mg, Ni) NASICON frameworks (Tables II and III) leads to several combinations for the PO4 group environments. So, within [Mg0.50] and [Ni0.50] structures, the PO4 groups are more disturbed and consequently in addition to the broadening of the Raman bands, a shift of the modes to higher or lower wavenumbers are normally expected. In fact, the same phenomenon has already been observed in others Raman spectroscopic study of NASICON-phases (Pikl et al., Reference Pikl, de Waal, Aatiq and El Jazouli1998; Junaid et al., Reference Junaid Bushiri, Antony and Aatiq2008; Aatiq et al., Reference Aatiq, Tigha and Benmokhtar2012). The Raman band positions of the four (ν 1, ν 2, ν 3, and ν 4) PO4 modes observed in spectra are close to those expected for NASICON-type phosphates. Their assignments were made based on data from the literature concerning some crystalline compounds (Figure 8). Thus, the symmetric non degenerate PO stretching modes (ν 1) are observed in the range 900–1020 cm−1, while antisymmetric doubly degenerate PO stretching modes (ν 2) are located in the 380–480 cm−1 range. The symmetric, triply degenerate OPO bending (ν 3) is observed between 1050 and 1300 cm−1 and the triply degenerate, antisymmetric and harmonic OPO bending (ν 4) is observed in the range 530–680 cm−1.

Figure 8. (Color online) Raman spectra for (a) Mg0.50SbFe(PO4)3, (b) Ni0.50SbFe(PO4)3, and (c) Ca0.50SbFe(PO4)3 phases.

During our spectroscopic investigations, our curiosity led us to consider and compare the number of observed PO4 Raman modes for [Ca0.50] [R $\overline 3 $ space group] and [Mg0.50] and/or [Ni0.50] [P $\overline 3 $ space group]. For this reason, we have carried out the factor group analysis, which predicts the theoretically number of Raman and Infrared active modes. Given that a detailed assignment of the external modes is difficult, our discussions were focused only to the PO stretching and OPO bending modes of PO4 ions. In fact, for [Ca0.50] we can expect eight Raman-active stretching vibrations of PO4 unit [2 ν 1 + 6 ν 3] and 10 Raman-active modes for the bending vibrations [4 ν 2 + 6 ν 4]. In the [Ca0.50] Raman spectrum, the bands are reasonably well resolved and their number is approximately six for the PO stretching vibrations and six for the OPO bending ones. In compounds with P $\overline 3 $ space group, where the atoms of phosphorous occupy three crystallographic positions [i.e., 3 × (6 g)], the number of expected Raman-active stretching vibrations is 24 [6 ν 1 + 18 ν 3] whereas for the Raman bending vibrations the expected one is 30 [12 ν 2 + 18 ν 4]. The number of observed bands in each Raman spectra of [Mg0.50] and [Ni0.50] is 10 for the stretching modes and 8 for the bending modes. It should be noticed that for compounds with R $\overline 3 $ space group, the factor group analysis predicts a maximum of eight Raman PO stretching modes. This remark clearly shows that the number of PO stretching modes observed in the Raman spectra of [Mg0.50] and [Ni0.50] is consistent with the crystallographic results we have obtained [i.e., P $\overline 3 $ space group].

Stretching vibrations of Sb–O are probably coupled with the P–O–P bending ν 4 mode. The frequencies found between 590 and 650 cm−1 in Raman spectra can be assigned empirically to Sb–O stretching modes involving Sb–O–P linkage. The same Sb–O vibrations are already observed at 597 and 652 cm−1 for SbOPO4 in the same range of frequency and at 580 and 639 cm−1 for Mn0.5MSb(PO4)3 (M = Al, Fe and Cr) phases (Sudarsan et al., Reference Sudarsan, Muthe, Vyas and Kulshreshtha2002; Anantharamulu et al., Reference Anantharamulu, Rao, Vithal and Prasad2009). In the lattice modes region, the translational modes of M 2+, Fe3+, Sb5+, and PO4 3− ions as well as librational modes of PO4 3− ions and FeO6, SbO6 groups should be expected. At wavenumbers below 450 cm−1 strong coupling between the different bending vibrations O–P–O, O–Sb–O, Sb–O–P, Sb–O–Sb is expected (Husson et al., Reference Husson, Genet, Lachgar and Piffard1988). The Raman bands observed around 300 and 340 cm−1 could be assigned to Fe3+–O stretching modes of vibrations in good agreement with the band positions observed in the range 300–370 cm−1 for Li3Fe2(PO4)3 (Butt et al., Reference Butt, Sammes, Tompsett, Smirnova and Yamamoto2004). The low-frequency modes observed below 270 cm−1 can be easily attributed to translational modes of the M 2+, Sb5+, and (PO4)3− ions.

IV. CONCLUSION

M 0.50SbFe(PO4)3 (M = Mg, Ni) phosphates are prepared and characterised by XRD and Raman spectroscopy. Both samples crystallise in P $\overline 3 $ , which is a relatively new space group in the NASICON-type materials. In both phases, the [SbFe(PO4)3] − NASICON framework is preserved and a partially-ordered distribution of Sb5+ and Fe3+ ions is observed. The three M 2+ cations of the unit cell are preferentially located as follows: one in the M(1b) (0, 0, 0.5) site and the two others one are in the M(2d′) (1/3, 2/3, ~0.17) site. Both M(1b) and M(2d′) sites correspond to the M1(3b) sites frequently reported in NASICON structure with R $\overline 3 $ space group. Analysis of distance values between the centre of each 1a and 2d sites, which both correspond to the M1(3a) sites of R $\overline 3 $ space group, and oxygen atoms surrounding them, confirms the fact that these two sites are normally assumed to be empty in [Mg0.50] and [Ni0.50] phases. Crystallographic results we obtained for both latter samples clearly show that structures of M 0.50Ti2(PO4)3 (M = Mg, Co) phosphates could be solved without problems in the P $\overline 3 $ space group instead of the two R $\overline 3 $ c or R $\overline 3 $ ones. Raman study is consistent with the obtained crystal structures. In fact, results from the factor group analysis have allowed us to follow the reduction of symmetry from R $\overline 3 $ in [Ca0.5] to P $\overline 3 $ space group in both [Mg0.50] and [Ni0.50] phosphates. All Raman spectra show characteristic PO4 vibrations and the stretching modes of SbO6 groups are observed at significantly higher frequency (570 and 670 cm−1) than stretching modes of FeO6 groups (~375 cm−1).

SUPPLEMENTARY MATERIAL

The supplementary material for this article can be found at https://doi.org/10.1017/S0885715617000331.

ACKNOWLEDGEMENTS

Financial support of the Moroccan Ministry of Higher Education, Scientific Research and Training of managerial staff (MESRSFC), National Center for Scientific and Technical Research (CNRST) are gratefully acknowledged. Authors are also grateful to Engineer in “SCA de l'Unités d'Appui Technique à la Recherche Scientifique (UATRS) “CNRS- Rabat, Morocco” for technical assistance; and for SCA of CNRS-Rabat for making to the disposal of our Laboratory a Panalytical X′Pert-PRO diffractometer.

References

Aatiq, A., Ménétrier, M., El Jazouli, A., and Delmas, C. (2002). “Structural and lithium intercalation studies of Mn(0.5-x)Ca x Ti2(PO4)3 phases (0 ≤ x ≤ 0.50),” Solid State Ion. 150, 391–405.Google Scholar
Aatiq, A., Hassine, R., Tigha, R., and Saadoune, I. (2005). “Structures of two newly synthesized A0.50SbFe(PO4)3 (A = Mn, Cd) NASICON phases,” Powder Diffr. 20, 33–39.CrossRefGoogle Scholar
Aatiq, A., Tigha, R., Hassine, R., and Saadoune, I. (2006). “Crystallochemistry and structural studies of two newly CaSb0.50Fe1.50(PO4)3 and Ca0.50SbFe(PO4)3 NASICON phases,” Powder Diffr. 21, 45–51.Google Scholar
Aatiq, A., Tigha, R., and Benmokhtar, S. (2012). “Structure, infrared and Raman spectroscopic studies of new Sr0.50SbFe(PO4)3 and SrSb0.50Fe1.50(PO4)3 NASICON phases,” J. Mater. Sci. 47, 1354–1364.CrossRefGoogle Scholar
Aatiq, A., Tigha, R., Fakhreddine, R., and Marchoud, A. (2015). “Structure and spectroscopic characterization of the two PbSb0.5Fe1.5(PO4)3 and Pb0.5SbFe(PO4)3 phosphates with NASICON type-structure,” J. Mater. Environ. Sci. 6(12), 3483–3490.Google Scholar
Alami Talbi, M., Brochu, R.; Parent, C., Rabardel, L., and Le Flem, G. (1994). “The new phosphates Ln 1/3Zr2(PO4)3 (Ln = Rare Earth),” J. Solid State Chem. 110, 350–355.Google Scholar
Anantharamulu, N., Rao, K. K., Vithal, M., and Prasad, G. (2009). “Preparation, characterization, impedance and thermal expansion studies of Mn0.5MSb(PO4)3 (M = Al, Fe and Cr),” J. Alloys Compd. 479, 684–691.Google Scholar
Anuar, N. K., Adnan, S. B. R. S., and Mohamed, N. S. (2014). “Characterization of Mg0.5Zr2(PO4)3 for potential use as electrolyte in solid state magnesium batteries,” Ceram. Int. 40, 13719–13727.Google Scholar
Barré, M., Crosnier-Lopez, M. P., Le Berre, F., Emery, J., Suard, E., and Fourquet, J. L. (2005). “Room temperature crystal structure of La1/3Zr2(PO4)3, a NASICON-type compound,” Chem. Mater. 17, 6605–6610.Google Scholar
Barré, M., Le Berre, F., Crosnier-Lopez, M. P., Bohnke´, O., Emery, J., and Fourquet, J. L. (2006). “La3+ Diffusion in the NASICON-type compound La1/3Zr2(PO4)3: X-ray thermodiffraction, 31P NMR, and ionic conductivity investigations,” Chem. Mater. 18, 5486–5491.Google Scholar
Barré, M., Crosnier-Lopez, M. P., Le Berre, F., Suard, E., and Fourquet, J. L. (2007). “Synthesis and structural study of a new NASICON-type solid solution: Li1−x La x/3Zr2(PO4)3 ,” J. Solid State Chem. 180, 1011–1019.Google Scholar
Barth, S., Olazcuaga, R., Gravereau, P., le Flem, G., and Hagenmüller, P. (1993). “Mg0.5Ti2 (PO4)3: a new member of the NASICON family with low thermal expansion,” Mater. Lett. 16, 96–101.Google Scholar
Benmokhtar, S., El Jazouli, A., Aatiq, A., Chaminade, J. P., Gravereau, P., Wattiaux, A., Fournes, L., and Grenier, J. C. (2007). “Synthesis, structure and characterisation of Fe0.50Ti2(PO4)3: a new material with NASICON-like structure,” J. Solid State Chem. 180, 2004–2012.CrossRefGoogle Scholar
Brown, I. D. and Altermatt, D. (1985). “Bond-valence parameters obtained from a systematic analysis of the inorganic crystal structure database,” Acta Crystallogr. B: Struct. Sci. B41, 244–247.CrossRefGoogle Scholar
Bykov, D. M., Gobechia, E. R., Kabalov, Yu. K., Orlova, A. I., and Tomilin, S. V. (2006). “Crystal structures of lanthanide and zirconium phosphates with general formula Ln0.33Zr2(PO4)3, where Ln = Ce, Eu, Yb),” J. Solid State Chem. 179, 3101–3106.CrossRefGoogle Scholar
Butt, G., Sammes, N., Tompsett, G., Smirnova, A., and Yamamoto, O. (2004). “Raman spectroscopy of superionic Ti-doped Li3Fe2(PO4)3 and LiNiPO4 structures,” J. Power Sources 134, 72–79.CrossRefGoogle Scholar
Crosnier-Lopez, M. P., Barré, M., Le Berre, F., and Fourquet, J. L. (2006). “Electron-irradiation induced phase transformation in La1/3Zr2(PO4)3: La3+ displacement in a preserved NASICON framework,” J. Solid State Chem. 179, 2714–2720.Google Scholar
Delmas, C., Viala, J. C., Olazcuaga, R., Le Flem, G., Hagenmuller, P., Cherkaoui, F., and Brochu, R. (1981). “Ionic conductivity in Nasicon-type phases Na1+xZr2−x Lx(PO4)3 (L = Cr, In, Yb),” Solid State Ion. 3/4, 209–214.Google Scholar
Derouet, J., Beaury, L., Porcher, P., Olazcuaga, R., Dance, J. M., Le Flem, G., El Bouari, A., and El Jazouli, A. (1999). “A new NASICON-type phosphate: Co0.5Ti2(PO4)3 II. Simulation of optical and magnetic properties,” J. Solid State Chem. 143, 230–238.Google Scholar
El Bouari, A., El Jazouli, A., Dance, J. M., Le Flem, G., Olazcuaga, R. (1994). “A New NASICON-Like Phosphate Co0,5Ti2(PO4)3 ”, Adv. Mater. Res. 1–2, 173–176.Google Scholar
Hans, W. and Ulrich, M. (Eds.) (2004). International Tables for Crystallography Vol. A1: Symmetry Relation between Space Group (Kluwer Academic Publishers, Dordrecht/Boston/London), 1th ed., p. 286; 310.Google Scholar
Husson, E., Genet, F., Lachgar, A., and Piffard, Y. (1988). “The vibrational spectra of some antimony phosphates,” J. Solid State Chem. 75, 305–312.Google Scholar
Junaid Bushiri, M., Antony, C. J., and Aatiq, A. (2008). “Raman and FTIR studies of the structural aspects of NASICON-type crystals; AFeTi(PO4)3 (A = Ca, Cd),” J. Phys. Chem. Solids 69, 1985–1989.CrossRefGoogle Scholar
Le Bail, A., Duroy, H., and Fourquet, J. L. (1988). “Ab-initio structure determination of LiSbWO6 by X-ray powder diffraction,” Mater. Res. Bull. 23, 447–452.Google Scholar
Nakamoto, K. (1986). Infrared and Raman Spectra of Inorganic and Coordination Compounds (Wiley–Interscience, New York), 4th ed., p. 138.Google Scholar
Olazcuaga, R., Dance, J. M., Le Flem, G., Derouet, J., Beaury, L., Porcher, P., El Bouari, A., and El Jazouli, A. (1999). “A new NASICON-type phosphate Co0.5Ti2(PO4)3: I. Elaboration, optical and magnetic properties,” J. Solid State Chem. 143, 224–229.Google Scholar
Padhi, A. K., Nanjundaswamy, K. S., Masquelier, C., and Goodenoogh, J. B. (1997). “Mapping of transition metal redox energies in phosphates with NASICON structure by lithium intercalation,” J. Electrochem. Soc. 144(8), 2581–2586.Google Scholar
Pikl, R., de Waal, D., Aatiq, A., and El Jazouli, A. (1998). “Vibrational spectra and factor group analysis of Mn(0.5+x)Ti(2−2x)Cr2x(PO4)3 {0≤x≤0.50},” Vibr. Spectrosc. 16, 137–143.Google Scholar
Rodriguez-Carvajal, J. (1997). Fullprof, Program for Rietveld refinement (Laboratoire Léon Brillouin (CEA-CNRS), Saclay, France).Google Scholar
Roy, R., Vance, E. R., and Alamo, J. (1982). ” [NZP], A new radiophase for ceramic nuclear waste forms,” Mat. Res. Bull. 17, 585–589.Google Scholar
Shannon, R. D. (1976). “Revised effective ionic and systematic studies of interatomic distances in halides and chalcogenides,” Acta Crystallogr. A: Cryst. Phys. Diffr. Theor. Gen. Crystallogr. 32, 751–767.Google Scholar
Sudarsan, V., Muthe, K. P., Vyas, J. C., and Kulshreshtha, S. K. (2002). “PO4 3− tetrahedra in SbPO4 and SbOPO4: a 31P NMR and XPS study,” J. Alloys Compd. 336, 119–123.CrossRefGoogle Scholar
Woodcock, D. A., Lightfoot, P., and Smith, R. I. (1999). “Powder neutron studies of three low thermal expansion in the NZP family: K0.5Nb0.5Ti1.5(PO4)3, BaTi2(PO4)3 and Ca0.25Sr0.25Zr2(PO4)3 ,” J. Mater. Chem. 9, 2631–2636.Google Scholar
Zhao, Y., Wei, Z., Pang, Q., Wei, Y., Cai, Y., Fu, Q., Du, F., Sarapulova, A., Ehrenberg, H., Liu, B., and Chen, G. (2017). “NASICON-type Mg0.5Ti2(PO4)3 negative electrode material exhibits different electrochemical energy storage mechanisms in na-ion and li-ion batteries,” ACS Appl. Mater. Interfaces 9(5), 4709–4718.Google Scholar
Figure 0

Figure 1. (Color online) Comparison of the experimental (•••) calculated (—), and difference profile of the XRD pattern of Ni0.50SbFe(PO4)3. The LeBail refinements of cell parameters were made for (a) R$\overline 3 $c, (b) R$\overline 3 $, (c) P$\overline 3 $, and (d) P$\overline 3 $c space groups over the range 10°–17° (2θ Cu). The non-indexed XRD diffraction lines are indicated by blue stars.

Figure 1

Figure 2. (Color online) Comparison of the experimental (•••) calculated (—), and difference profile of the XRD pattern of Mg0.50SbFe(PO4)3. The LeBail refinements of cell parameters were made for (a) P$\overline 3 $c, and (b) P$\overline 3 $ space groups over the range 18°–42° (2θ Cu). The non-indexed XRD diffraction lines are indicated by blue stars.

Figure 2

Figure 3. (Color online) Projections of the structure for (a) M0.50X′2(PO4)3 (R$\overline 3 $ space group), and (b) M0.50X2(PO4)3 (P$\overline 3 $ space group) along the b-axis. The different M and X sites, accompanying the R$\overline 3 $-P$\overline 3 $ transposition of atomic positions, have been labelled and located.

Figure 3

Figure 4. (Color online) Experimental (•••) calculated (—), and difference profile, obtained after Rietveld refinement, of the XRD pattern of Mg0.50SbFe(PO4)3.

Figure 4

Figure 5. (Color online) Experimental (•••) calculated (—), and difference profile, obtained after Rietveld refinement, of the XRD pattern of Ni0.50SbFe(PO4)3.

Figure 5

Figure 6. (Color online) View of the structure for M0.50X2(PO4)3 [M = Mg, Ni; X = Sb(Fe)] NASICON-type phases.

Figure 6

Table I. R$\overline 3 $-P$\overline 3 $ transposition of the atomic positions from M0.50X′2(PO4)3 (R$\overline 3 $ space group) to M0.50X2(PO4)3 (P$\overline 3 $ space group) NASICON phases.

Figure 7

Table II. Results of the Rietveld refinement of Mg0.50SbFe(PO4)3.

Figure 8

Table III. Results of the Rietveld refinement of Ni0.50SbFe(PO4)3.

Figure 9

Table IV. Selected interatomic distances (Å) and calculated bond valence sum (BVS) for Mg0.50SbFe(PO4)3.

Figure 10

Table V. Selected interatomic distances (Å) and calculated bond valence sum (BVS) for Ni0.50SbFe(PO4)3.

Figure 11

Table VI. Powder diffraction data of Mg0.50SbFe(PO4)3 (CuKα1; λ = 1.540 56 Å).

Figure 12

Table VII. Powder diffraction data of Ni0.50SbFe(PO4)3 (CuKα1; λ = 1.540 56 Å).

Figure 13

Figure 7. (Color online) Structure of M0.50SbFe(PO4)3 (M = Mg, Ni) phases showing the ordered cationic distribution along the c-axis, between M cations and vacancies, of the four possible □M(1a), M(1b), □M(2d), and M(2d′) positions of the M1 sites. Sb(Fe) atoms are blue circles (•), P(1)O4 are yellow, P(2)O4 are blue, and the P(3)O4 are green.

Figure 14

Figure 8. (Color online) Raman spectra for (a) Mg0.50SbFe(PO4)3, (b) Ni0.50SbFe(PO4)3, and (c) Ca0.50SbFe(PO4)3 phases.

Supplementary material: File

Aatiq supplementary material

Aatiq supplementary material 1

Download Aatiq supplementary material(File)
File 13.5 KB