About this Journal Submit a Manuscript Table of Contents
Journal of Spectroscopy
Volume 2013 (2013), Article ID 123519, 10 pages
Research Article

Raman Spectroscopy, X-Ray, SEM, and DTA Analysis of Alkali-Phosphate Glasses Containing and Nb2O5

1Equipe Sciences de Matériaux, FST-Errachidia, Errachidia, Morocco
2Equipe Matériaux et Environnement, Laboratoire de Chimie Appliquée et Environnement, Université Hassan 1er, Settat, Morocco
3Physics Department (I3N), Aveiro University, Campus Universitário de Santiago, Aveiro, Portugal

Received 11 June 2012; Revised 9 August 2012; Accepted 29 August 2012

Academic Editor: Christoph Krafft

Copyright © 2013 L. Bih et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


New phosphate glasses in the quaternary system (50-x) A2O-x WO3-10 Nb2O5-40 P2O5, with x = 0; 30 and A = Li or Na were prepared by the melt quenching method. The effect on the crystallization behaviour of the glass due to the introduction of WO3 into the glass composition and, consequently, the diminishing of the molar amount of the alkaline oxide and the decreasing of the molar ratio between network modifiers and network formers (M/F) was studied. The prepared glasses were heat-treated in air, at 550°C, 600°C, and 650°C for 4 hours. The structure, of the obtained samples, was studied by differential thermal analysis (DTA), X-ray powder diffraction (XRD), and Raman spectroscopy and the morphology by scanning electron microscopy (SEM). It was found that the replacement of Li2O or Na2O by WO3 reduces the number of the crystallised phases. In the lithium-niobiophosphate glasses, the presence of WO3 promotes the formation of NbOPO4 instead of the LiNbO3 phase and reduces the formation of ortho- and pyro-phosphate phases. The thermal treatments affect the arrangements of the network structure of the AW40-glasses.

1. Introduction

Glass-ceramic composite materials have been a subject of intensive interest during the last decades. The goal for the development of such materials has been to achieve a combination of properties not achievable by any of the elemental materials alone.

Phosphate glasses have received considerable attention in the past few years due to the synthesis of new glass compositions with high chemical stability. The improvement of chemical stability [13] stimulated the application of phosphate glasses in several fields of materials science, such as fast ionic conductors [4], semiconductors [5], photonic materials [69], hermetic seals [10], rare-earth ion host solid state lasers [11], and biomedical materials [12]. The structure of these glasses consists of PO4 tetrahedra, which can be attached to a maximum of three neighbouring tetrahedra forming a three-dimensional network as in vitreous P2O5. The addition of metal oxides to the glass leads to a depolymerisation of the network, with the breaking of P–O–P linkages and the creation of nonbridging oxygen. The modifying cations can provide ionic cross-linking between the nonbridging oxygen of two phosphate chains, thus increasing the bond strength of this ionic cross-link and improving the mechanical strength and chemical durability of the glasses [1]. Niobium phosphate glasses have been investigated because they might be used in a wide range of applications, such as waste form for radioactive waste immobilization [13], rare-earth ion hosts for laser glasses [14], glass fibers and optical lenses [15, 16], hermetic sealing [17], electrodes [18], and agricultural applications [19]. Besides that, phosphate glasses are easy to produce at relatively low temperatures (900–1200°C).

Niobiophosphate glass compositions such as P2O5-Nb2O5-V2O5-TiO2 and P2O5-Nb2O5-V2O5-Fe2O3, investigated by vibrational spectroscopy, exhibit structural features, characterized by NbO4 and NbO6 units [20]. Since several niobiophosphate glasses are Nb-rich and Li-rich, their application as precursors to glass composites containing embedded LiNbO3 ferroelectric crystals has been proposed [21]. Recently, Graça et al. [22] prepared LiNbO3 glass-ceramic materials from the Li2O-Nb2O5-SiO2 glasses by controlled crystallization. The dielectric properties of these glass-ceramics materials show better properties than their parent glasses [22]. More recently, Graça et al. [23] showed the effect of the network former on the nature of the crystallized phases in the glass-ceramics matrix, by analysing the crystallization of the Li2O-Nb2O5-P2O5 glasses inducing the formation of multiple phases in opposite trends than that seen in silicate-based glasses.

In the present work, niobium phosphate glasses containing tungsten oxide belonging to the (50-x) A2O-(x) WO3-10Nb2O5-40P2O5 (x = 0 and 30; A = Li, Na) systems were prepared by the melt quenching technique. These glasses contain three types of components. P2O5 is a well-known network former with PO4 structural units with one of the four oxygen atoms in the PO4 tetrahedron being doubly bonded to the phosphorus atom with the substantial π-bond character to account for the pentavalency of phosphorus. Nb2O5 and WO3 are incipient glass network formers and as such do not readily form glass but do so in the presence of modifiers like A2O (A = Li, Na) with MO4 and MO6 (M = Nb, W) structural units. Generally, niobium and tungsten are located in octahedral sites in phosphate glasses. The tungsten oxide structural groups are expected to mix easily in a A2O-Nb2O5-P2O5 network because some of the Raman vibrational bands of the structural groups of these ions lie in the same region as those of NbO6 and P-O-P structural units and are expected to occupy a variety of sites with different crystal field strengths due to site variability. According to these structural considerations, it is expected that the interaction between phosphate and niobate/tungsten anions in niobium-tungsten-containing phosphate glasses would be simpler because the coordination number of phosphorous atoms for oxygen is usually constant at four in phosphate glasses. In this study, the as-prepared glasses were heat-treated at different temperatures in order to promote crystallization and they were investigated using differential thermal analysis, X-ray diffraction, and Raman spectroscopy to provide a structural characterization and SEM microscopy to analyse their morphology.

2. Experimental Details

2.1. Sample Preparation

The glasses with composition (50-x) A2O-x WO3-10Nb2O5-40P2O5 (x = 0 and 30; A = Li, Na) (% mole) were prepared by the melt quenching method using AR-grade chemicals Li2CO3, Na2CO3, Nb2O5, WO3, and NH4H2PO4. Approximately 10 g of chemicals were thoroughly mixed and melted in a platinum crucible at 1400 K in an electric muffle furnace for about one hour. The melt was swirled occasionally to ensure homogeneity. The coin-shaped glass samples were obtained by pouring the melts onto a stainless steel plate and pressed with another plate.

In this work, the samples without W will be called Li-40 or Na-40 and the samples with W will be called LiW-40 and NaW-40.

The obtained samples were heat-treated in air at 500, 600, and 650°C in agreement with the results of the differential thermal analysis.

2.2. X-Ray Diffraction

The obtained products was examined by X-ray diffraction analysis using a Seifert XRD 3000 (Cu Kα radiation, λ = 1.5406 Å). The diffraction patterns were recorded at room temperature in the range 5–70° (6 h) in steps of 0.02°.

2.3. Density, Molar Volume, and Differential Scanning Calorimetric Measurements

The density () of all the samples was calculated by Archimedes’s method with Xylene as buoyant liquid. The molar volume of the samples was calculated using the relation where is the total molecular weight.

The glass transition temperature () of all the samples was determined using differential thermal analysis (Linseis apparatus) at the rate of 5 K/min.

2.4. Raman Spectroscopy Analysis

The Raman spectra were recorded with a HoloSpec (f/1.8i, Kaiser Optical Systems) equipped with a holographic transmission grating and thermoelectrically cooled two-dimensional multichannel CCD detector (Newton, Andor Technology, 1600 × 400 pixels, –60°C). An argon laser, with the wavelength centred at 514.5 nm, was used for the excitation. The spectrometer was calibrated by fluorescence lines of the neon lamp. Nonpolarized Raman spectra were collected in the back-scattering geometry, in the range 180–2280 cm−1, with a resolution of 3 cm−1. Accuracy of spectral measurements, resulting from the wavelength calibration procedure and experimental conditions, is estimated to be about 1.5 cm−1. Precision of the reported peak positions, as represented by standard errors obtained in peak fits is 0.04 cm−1. The acquisition time was between 2 and 5 s. All the measurements were made at room temperature.

2.5. Scanning Electron Microscopy

The microstructure and morphology of the glass and glass-ceramics were observed by scanning electron microscopy (SEM), performed in a Hitachi S4100-1 system, on the free surface and on the fracture surface of all samples covered with carbon before microscopic observation.

3. Results and Discussion

3.1. Density and Molar Volume Parameters

The density () values of the lithium (LiW-40) and sodium (NaW-40) glasses are 3.21 and 3.19 g/cm3, respectively. The values of molar volume, calculated using the formula , where is the total molecular weight, are 43.21 cm3 and 46.50 cm3 for the LiW-40 and NaW-40 glasses, respectively. These values, density and molar volume, of both tungsten-based glasses are higher than the values obtained for the Li-40 glass ( = 2.69 g/cm3 and = 36.55 cm3) [23]. The observed increase in the density when the alkaline oxide is substituted by WO3 oxide can be explained due to the replacement of an alkali cation (Li or Na) by a heavier ion (W), and possibly, denser amorphous structures are achieved due to stronger cross-linking between phosphate networks. However, the number of oxygen anions rises gradually in the glass network with the increase of WO3 content and thus the addition of WO3 can promote a relatively open structure causing an increase in the molar volume. This behaviour was observed in the present case, where the molar volume of the glass systems increases with the growth of the WO3 concentration, suggesting that the molecular volume is influenced by the number of O2− ions per unit of volume. Apparently, while substituting A2O by WO3, the molar volume should decrease because two cations per unit found for A2O are replaced by a single cation in WO3. The increase of molar volume in tungsten-glasses suggests that the excess of oxygen, due to the substitution of Li2O or Na2O by WO3, induces the formation of additional nonbridging oxygen and/or the formation of voluminous niobium and/or tungsten structural units such as WO6/NbO6 octahedron in the glass network.

3.2. Glass Transition Temperature

The DTA spectra of the glasses LiW-40 and NaW-40 are shown in Figure 1. Typical glass parameters: (glass transition temperature), (the onset of crystallization temperature), and (crystallization temperature) were extracted from these curves and are summarized in Table 1. The characteristic temperatures are indicated by arrows in Figure 1. Both glasses show two crystallisation temperatures ( and ), indicating that, at least, two crystalline phases are formed in their network by heating. The influence of the WO3 content on the thermal properties of the glass is seen by comparing the DTA spectrum of Li-40 glasses (40P2O5-50Li2O-10Nb2O5 (% mol)) [23] and the spectra of LiW-40 and NaW-40 (Figure 1). It must be noticed that in these compositions the P2O5 and Nb2O5 molar concentration were kept equal. One can note that the value for the glass LiW-40 ( = 340°C) is lower than that of Li-40 ( = 450°C) [23]. It was expected that with the substituting of Li2O by WO3, the glass transition temperature should increase because of the increase of the number of strong W-O-P and W-O-W linkages. Thus, the observed decrease of the glass transition temperature from Li-40 to LiW-40 samples (Figure 1, Table 1) indicates that the replacement of Li2O by WO3 induces the depolymerisation of the network and the creation of nonbridging oxygen ions. In fact, this substitution is associated with the increase of O/P ratio. The depolymerisation of the network is in agreement with the fact that the molar volumes of LiW-40 and NaW-40 are higher than the corresponding free-WO3 glass (Table 1).

Table 1: Physical parameters for AW-40 glasses.
Figure 1: DTA curves of the AW-40 (A = Li, Na) glasses.

For the evaluation of the thermal stability of glasses, the criterion was applied. Higher values of this criterion usually correspond to a higher thermal stability and glass-forming ability. The obtained values of the for the studied glasses are given in Table 1. It can be verified that the value of = 123, measured for LiW-40 glass, is higher than the = 70 obtained for the NaW-40 glass. It means that NaW-40 glass has a lower thermal stability and a higher crystallization tendency. However, these AW-40 glasses are thermally more stable than the WO3-free glass, which has a smaller value of = 30 [23]. The DTA of the tungsten-free glass, Li-40 [23], exhibits a broad exothermic peak at 480–600°C [23]. This broadband can be ascribed to the simultaneous crystallization of more than one phase from the glass matrix. For instance, based upon the Raman spectra and the X-ray diffraction, this broad exothermic peak can be attributed to the crystallization of LiNbO3, Li3PO4, Li4P2O7, and LiPO3 crystalline phases. Since the shape of the crystallisation peak in the DTA pattern of both LiW-40 and NaW-40 glasses is different to that of WO3-free glass, one can anticipate that the nature of the crystallised phases issued from these glasses could be different. Based on the fact that the alkali content in LiW-40 and NaW-40 glasses is lower than that of for a WO3-free glass, one can expect the formation of niobium- and/or tungsten-phosphate phases rather than alkali phosphate phases. This hypothesis is well confirmed by X-ray diffraction and Raman analysis in the next sections.

3.3. X-Ray Diffraction and SEM Micrographs

The glassy state of the LiW-40 and NaW-40 samples before heat treatments is confirmed by X-ray diffraction. XRD patterns of the crystallized samples are shown in Figure 2. The crystalline phases for the samples after heat treatment were identified by comparing the X-ray diffraction pattern shown in Figure 2 with the JCPDS file NbOPO4 (27-1316) and Li4P2O7 (77-1045). It must be mentioned that all the diffraction peaks associated with the NbOPO4 phase were detected and only the major peak of the Li4P2O7 was observed. In the Li-40 glass (WO3-free), it was also found crystallization peaks corresponding to LiPO3, Li3PO4, and LiNbO3 phases [23]. Thus, the replacement of Li2O by WO3 reduces the number of the crystallised phases and consequently, the absence of the alkali-phosphates phases (e.g., Li3PO4 and LiPO3) in the crystallised samples correlates well with the alkali content in the glasses.

Figure 2: X-Ray diffraction patterns of the AW-40 glasses submitted to different heat treatments.

When WO3 is added instead of A2O modifier oxide, it can be incorporated into the glass network as corner-sharing [WO6] octahedral units. The formula [WO6] suggests that no additional oxygen atoms are required for the coordination of tungsten since WO3 itself can almost generate [WO6/2] in the network. However, the presence of niobium in the network as corner-sharing [NbO6/2] octahedral units suggests that additional oxygen atoms are required for the coordination of niobium (since Nb2O5 itself can almost generate [NbO5/2] in the network), provided by the host network. One possible way for Nb5+ to get its six-oxygen coordination is to react with phosphate units according to the reaction: [NbO5/2] + [POO3/2] → [NbO6/2]- + [PO4/2]+. It is worth to note that the neutral phosphate unit [POO3/2] can be formed only in the ultraphosphate region where the ratio [O]/[P] < 3. Since in the AW-40 glasses the [O]/[P] = 4.25, one can state that the phosphate units with more than two nonbridging oxygen supply oxygen to niobium to form the oxygen-octahedrally coordinated NbO6 in the glasses. Thus, the formation of motifs of the niobium-phosphate NbOPO4 crystalline phase can be understood through the [O]/[P] ratio. It must be noticed that NbOPO4 belongs to a subgroup of materials that exhibit anisotropic negative thermal expansion (NTE) at high temperature [24]. Under ambient conditions, NbOPO4 has been synthesized having a monoclinic structure (a = 13.097 Å, b = 5.280 Å, c = 13.228 Å, β = 120.334°) [25] and also in a tetragonal structure (a = 6.389 Å, c = 4.105 Å), which is 18% denser than the monoclinic phase [26]. Both phases of NbOPO4 are based on NbO6 octahedra and PO4 tetrahedra, which share corners to form three-dimensional networks. The twofold coordination of oxygen atoms in the structure facilitates transverse motion of oxygen in the Nb-O-P linkages and results in negative thermal expansion [24]. It is well known that the structure of phosphate glasses is strongly dependent on the O/P ratio of the glass composition and with the increase of the O/P ratio, the phosphate network becomes more depolymerized and the ultraphosphate structure composed of cross-linked Q3 units (O/P = 2.5) changes to (a) the chain-like metaphosphate structure made up of Q2 units (O/P = 3.0) to (b) the pyrophosphate, Q1, structure (O/P = 3.5) and finally to (c) the orthophosphate structure consisting of isolated Q0 units (O/P = 4.0). In all of the AW-40 glasses of the present work, the O/P ratio is 4.25, so the network structure is expected to be dominated by orthophosphate Q0 groups. For instance, the structure of NbOPO4 consists of Q0 units sharing corners with niobium NbO6 units. In agreement with the structural O/P ratio, the XRD patterns of (Figure 2) show that in the LiW-40 sample, treated at 650°C, the NbOPO4 phase crystallizes in the orthorhombic and in the NaW-40 sample in the monoclinic structure. The formation of different NbOPO4 structures can be related with the size of the alkali ion. Thus, increasing the size of the alkali, the phase symmetry decreases from orthorhombic to monoclinic, which can suggest that the presence of sodium ions in the glass matrix acts as an additional structural constraint which favours the formation of NbOPO4 with low symmetry.

Furthermore, the decrease of the content of the alkali ions from 40P2O5-10Nb2O5-50A2O to 40P2O5-10Nb2O5-20WO3-30A2O (AW-40) glass seems to favour the formation of NbOPO4 instead of the LiNbO3 phase. One possible explanation for this behaviour can be given bellow: since the roles of Li+ and Nb5+ in the structure of phosphate-based glasses would be naturally different and all ions of Li+ and Nb5+ in the original glasses would not be included into precipitated crystals, one can state that the major content of alkali ions are inside the glassy-matrix rather than in precipitated crystalline phases. Thus, the formation of NbOPO4 phase as shown in Figure 2 demonstrates that the major content of alkali A+ (A = Li, Na) ions are probably stabilized in phosphate-based glasses and thus the presence of a small amount of alkali ion around niobium units is low enough to produce the formation of ANbO3 crystals, which could be identified by X-ray diffraction, in crystallised samples. Moreover, Graça et al. [23] highlight that the crystallisation of the glass with high Li+/Nb5+ ratio (i.e., WO3-free glass) leads to the precipitation of LiNbO3 phase.

The SEM micrographs of the samples treated at 650°C, presented in Figure 3, revealed the presence of particles with a spherical shape and linear size lower than 50 nm in both compositions. In all samples treated at temperatures below 650°C, the presence of particles dispersed in the glass network was not observed. While the NaW-40 sample presents a higher number of particles, which should be related with NbOPO4 particles, the resolution of its XRD pattern (Figure 2) is very low compared to that of LiW-40. Thus, the resolution of the X-ray pattern depends on both the size and the content of the crystalline phases. Moreover, knowing that the spectra of the LiW-40 sample, treated at 650°C, are a multiphase sample with well-defined diffraction peaks (Figure 2), it is suggested the existence of particles with a very small linear size, not detectable by SEM.

Figure 3: SEM micrographs of the LiW-40 (a) and NaW-40 (b) samples heat-treated at 650°C for 4 hours.
3.4. Raman Spectroscopy

Raman spectra of the AW-40 glasses and of the corresponding heat-treated glasses are shown in Figures 4(a) and 4(b). The spectra of AW40-glasses, with and without heat treatment, are almost similar suggesting that the arrangement of the structural units in the glass matrix is not very much affected by the heat treatments of the glass. However, for the purposes of studying the effect of heat treatment, it is useful to deconvolute these spectra.

Figure 4: Raman spectra of AW-40 glasses in the frequency range 100–1500 cm−1 submitted to heat treatments at different temperatures: (a) Li; (b) Na.

Previously reported structure [2730] of phosphate glasses showed that the basic structural element of phosphate glasses is the [PO4]-tetrahedron. Their structures are usually presented using the Qn terminology where n represents the number of bridging oxygens (BO) per tetrahedron [28]. They start with the three-dimensional network of Q3 groups, an example of which is vitreous P2O5 (v-P2O5). Changes in the structure caused by introducing other metallic oxides are most easily understood in the case of binary systems. The addition of modifying oxides leads to the scission of P-O-P bonds and to the network depolymerization [29]. Acidic phosphate glasses which contain P2O5 greater than 50 mol% have the structure of a three-dimensional network consisting of Q3 and Q2 groups. Metaphosphate glasses with 50 mol% of modifying oxides consist of metaphosphate chains or rings formed by Q2 groups. These glasses do not tend to have separated phases and thus do not fulfil the criterion required for controlled glass-ceramic crystallization. The structure of polyphosphate glasses (inverse phosphate glasses), containing <50 mol% P2O5, consists of chains formed by Q2 groups, Q1 and Q0 groups. The Raman spectrum of Li-40 shows broadbands at 925 and 265 cm−1, that can be attributed to the P-O bending and to the vibration of NbO6 units, respectively [23]. Moreover, this glass does not promote the preparation of glass-ceramics with only LiNbO3 crystallites because in the Raman spectra of the heat-treated samples it is observed the 950 cm−1 (Li3PO4), 1050 cm−1 (Li4P2O7), and 1100 cm−1 (LiPO3) vibration modes.

It is worth noting that the Raman bands of the as-prepared Li-W40 and Na-W40 glasses are less broad than those observed in the spectrum of WO3-free glass [23]. Since the broadness of the bands is an indication of the amorphous character of a glass, it can be deduced that the amorphous order of the glasses LiW-40 and NaW-40 is lower than that of a WO3-free glass. Therefore, the decreasing of the amorphous character is associated with the substitution of weak ionic alkali-oxygen bonds by stronger covalent W-O ones. This correlation is in agreement with the higher glass transition temperature of the Li-40 glass when compared to those of AW-40 (A = Li, Na) glasses. Considering the chemical compositions of the glasses (Table 1), the calculated [O]/[P] ratios are 3.75 and 4.25 for the WO3-free glass and for the AW-40 glasses, respectively and according to the [O]/[P] criterion, the AW-40 glasses should exhibit a higher tendency toward faster structure arrangement-crystallization which is in agreement with the Raman results that shows larger bands for the WO3-free glass.

The Raman spectra of the LiW-40 and NaW-40 glasses (Figure 4) are different even if they present equal [O]/[P] ratio. The width of the 800 cm−1–1000 cm−1 high-frequency vibrations of NaW-40 glass (Δ = 257 cm−1) is larger than that of the homologous LiW-40 glass (204 cm−1) as expected from the presence of a lower distribution of structural units in the LiW-40. The XRD spectra of the NaW-40 sample treated at 650°C (Figure 2) presents diffraction peaks, assigned to the NbOPO4 (JCPDS 27-1316), whose low intensity is in agreement with the width of the Raman bands in the high-frequency region (Figure 4).

The changes of the Raman spectra with composition and treatment temperature are used to deduce the structural evolution of the samples. For these purposes, it is useful to divide the spectra into three regions: the high-frequency region, from 700 to 1300 cm−1, the middle-frequency, from 500 to 700 cm−1, and the low-frequency region, below 500 cm−1. The deconvolution of the Raman spectra of the AW-40 in these three frequency domains are shown in Figures 5 and 6 and their assignments are summarized in Table 2. The Raman spectrum of the LiW-40 presents a general behaviour similar to the Raman spectrum of NaW-40 with only a slight shift of the frequency position assigned to the intrinsic atomic properties of the Li+ and Na+ ions. The intensity and the position of the Raman bands for the LiW-40 and NaW-40 glasses are clearly different compared to that seen for Li-40 [23]. Such differences suggest that WO3 has induced deep structural changes in the glass.

Table 2: Raman band positions of the LiW-40 and NaW-40 samples and their assignments.
Figure 5: Deconvolution of the LiW-40 Raman spectrum in three frequency regions: (a) low, (b) medium, and (c) high.
Figure 6: Deconvolution of the NaW-40 Raman spectrum in three frequency regions: (a) low, (b) medium, and (c) high.

Firstly, it will be considered the general features of the WO3 niobiophosphate network and secondly, the modification of the niobiophosphate network of the WO3-free glass by tungsten oxide. In a Li-40 glass-ceramic, peaks at 1300 cm−1 (P=O), 1100 cm−1 ((PO2)), 1050 cm−1 ((), 950 cm−1 ((), 905 cm−1 ((Nb-O)), 735 cm−1 ((P-O-P)), 657 and 605 cm−1 ((NbO6)), 420–225 cm−1 ((NbO6), and 186 cm−1 (δ(PO4)), are observed. The changes of the intensity and/or the position of these bands from a free-WO3 to AW-40 glasses could help to get more insight into the effect of WO3 on the niobiophosphate network. On the basis of the assignments of Raman spectra for the Li2O-P2O5 glass [28], the bands at around 600–760 cm−1 are due to P-O-P (bridging oxygen) stretching modes. The Raman bands in the region from 900 to 1300 cm−1 are associated with stretching vibrations of Q3, Q2, Q1, and Q0 species. The low-frequency region bands are due to bending modes. By analogy, the following attributions can be made [28]: the bands at 1141 cm−1 can be associated to symmetrical vibrations (PO2)-; the band in the range 1228–1275 cm−1 can be assigned to the vibrations of antisymmetric elongation (P=O); the broadband observed in 1040–1070 cm−1 probably corresponds to the vibration of symmetrical elongation (); the band near 763–766 cm−1 is ascribed to (P-O-P) vibration in the phosphate chains. The broadband in the range 590–640 cm−1 contains three components: the more intense band centered at 618 cm−1 must be ascribed to the vibrations of elongation (POM) (M = Nb, W); the shoulder observed around 590 cm−1 results, probably, from the same modes of vibration, whereas the shoulder observed around 640 cm−1 can arise from the elongation of M-O-M groups (M = Nb, W). The band around 511–566 cm−1 is probably connected to the vibrations of phosphate units deformation. Lastly, the three components at 184/154; 271/253; 373/382 cm−1 of the bands located in the frequency range 100–400 cm−1 are associated with the deformation modes of (MO6) units.

The addition of WO3 promotes an increase in the number of W–O bonds, which present a higher polarizability when compared with the P–O bonds, masking the spectrum associated with the entities of phosphate structural units. These latter are in shoulder of the very intense bands associated with the vibrations of W–O and/or Nb–O. Due to the fact that the Nb–O and W–O bonds possess both similar atomic properties and high polarisability, the bands characteristic of the vibrations of Nb–O and W–O overlap and appear in the same frequency positions of the spectra.

In the high-frequency region, the substitution of Li2O by WO3, when we move from 40P2O5-10Nb2O5-50Li2O (WO3-free) to 40P2O5-10Nb2O5-20WO3-30Li2O (LiW-40), can be related with the shift to low frequencies of the band centered at 1300 cm−1 to 1228–1275 cm−1; a shift of 1100 cm−1 to 1125–1141 cm−1, the appearance of a band at 980–1017 cm−1; the increase of the intensity of the bands near 950–960 cm−1; the appearance of bands near 916 cm−1, 840 cm−1 and 766 cm−1. These evolutions can be ascribed to the incorporation of niobium and/or tungsten ions into the phosphate network and formation of mixed structural units of the type P-O-M, M-O-M (M = Nb, W) linkages. The appearance of bands near 980, 916, 900, and 1017 cm−1 indicates the presence of (NbO6) units with different distortion in the glass matrix. The bands around 980 cm−1 and 1017 cm−1 are seen in the Raman spectrum of NbOPO4 crystalline phase, so the band at 979.5 cm−1 and 1017 cm−1 for NaW-40 and LiW-40, respectively, do indicate the presence of motifs PO4 and NbO6, in the structure of AW-40 glasses, linked in the same way as in the crystalline NbOPO4. This result is in agreement with the X-ray analysis of the crystallised NaW-40 and LiW-40 samples. It is worth to note that the band at 979.5 cm−1 could be correlated to the monoclinic symmetry since the identification of such structural arrangement is made by X-ray diffraction for NaW-40 sample (Figure 2).

The crystalline α-NbOPO4 contains octahedral chains [NbO6] parallel with the c-axis. These chains are connected between them by tetrahedrons [PO4] to form a three-dimensional structure; these corner sharing polyhedrons [NbO6] do not have terminal oxygen ions but they are distorted, giving an alternation of short and long bonds [31]. This structure shows a Raman band at 800 cm−1. On the other hand, one can note that the frequency positions of NbO6 modes in LiW-40 and NaW-40 are higher than in α-NbOPO4, suggesting that the symmetry and/or the degree of connectivity of the NbO6 in the glasses are different from the existing in the α-NbOPO4 phase. Shannon [32] clearly highlighted the displacement of the bands towards higher frequencies when the degree of distortion of octahedral increases but also when the number of terminal bonds Nb–OT pointing towards modifying ions grows. Therefore, the bands observed around 950–980 cm−1, 900–917 cm−1 can be ascribed to the distorted [NbO6] octahedral having at least a short Nb–O bond pointing towards a modifying ion. The band around 840 cm−1 could be associated to the Nb–O vibrations in distorted octahedral which can present an association by their corners with at least another octahedra of niobium. The emergence of broadband near 760–770 cm−1 can be explained by the presence in the glass matrix of the Nb–O vibration in less distorted NbO6 octahedral. The increase in the relative intensity of this last band in LiW-40 glass compared to that of NaW-40 glass, which can certainly be interpreted like an increase in more symmetrical octahedral [NbO6] sites, is in agreement with the identification by X-ray diffraction of high-symmetry orthorhombic NbOPO4 phase in LiW-40 sample and a monoclinic phase in NaW-40 one. Generally, the presence of bands below 790 cm−1 suggests the existence of less distorted octahedral units [33]. The bands around 640, 616, 590–580, and 566–511 cm−1 can be associated with the Nb–O modes of less distorted octahedral. More precisely, by comparison with the spectra obtained for borophosphate of sodium and niobium, the band around 638 cm−1 can be the reflection of a beginning of association of octahedra to form a three-dimensional network [34]. Lastly, the area of 380–100 cm−1 corresponds to the movements of angular O–Nb–O deformation. The bands around 382–372 cm−1 and 271–256 cm−1 can thus be connected to the vibrations relating to the octahedral distortion.

It is worth to note that the nonidentification, by X-ray diffraction, of pyrophosphate and orthophosphate precipitate phases (Figure 2) indicates that phosphate network is more polymerised in AW-40 glasses in agreement with the Raman spectra (Figures 3 and 4), which show high-frequency bands near 1280 cm−1.

The appearance of the latter band can be explained if we consider that a part of the network modifier species A2O is transferred from the phosphate to the niobium NbO6 unit since the charge balance requires that the average degree of polymerization for the phosphate species must increase. We can formulate this oxygen redistribution in terms of a chemical equilibrium, involving the reaction of anionic phosphate units with octahedra to form Q2 units and anionic niobium species: Also, the presence of nonbridging oxygen ions on the niobium units is in agreement with the presence of several components in the Raman bands located between 800 and 950 cm−1.

4. Conclusion

Glasses of the system A2O-WO3-Nb2O5-P2O5 (A = Li, Na) were prepared by the melt quenching and their structure was studied by thermal analysis, X-ray diffraction, SEM microscopy, and Raman spectroscopy. It was found that the introduction of tungsten into these alkali niobiophosphate glasses induces the creation of additional nonbridging oxygen ions and/or the formation of larger niobium- and/or tungsten-containing structural units such as WO6/NbO6 octahedra in the glass network.

The thermal analysis showed that the AW-40 glasses are more stable than the WO3-free glass and that the NaW-40 presents the lowest thermal stability and the highest tendency towards crystallization, when compared with the LiW-40 glass.

In the lithium-niobiophosphate glasses, the XRD results indicate that the replacement of Li2O by WO3 reduces the number of the crystallised phases. The presence of WO3 in those glasses promotes the formation of NbOPO4 instead of the LiNbO3 phase and favours the network polymerisation due to the nonformation of pyrophosphate and orthophosphate phases. The decreasing of amorphous character, with the increase of the heat treatment temperature, is associated with the substitution of ionic alkali-oxygen bonds by covalent W–O. In all AW-40 glasses, the network structure is dominated by orthophosphate Q0 groups. The Raman spectra of those glasses indicate that the arrangements of the network structural units are less affected by the heat treatments conditions.


The authors are grateful to the CNRST (Morocco)-FCT (Portugal) for the financial support offered to this work.


  1. B. C. Sales and L. A. Boatner, “Physical and chemical characteristics of lead-iron phosphate nuclear waste glasses,” Journal of Non-Crystalline Solids, vol. 79, no. 1-2, pp. 83–116, 1986. View at Scopus
  2. Y. B. Peng and D. E. Day, “High thermal expansion phosphate glasses. Part 1,” Glass Technology, vol. 32, no. 5, pp. 166–173, 1991. View at Scopus
  3. N. Aranha, O. L. Alves, L. C. Barbosa, and C. L. Cesar, “The role of Nb2O5 on the chemical durability of P2O5-Nb2O5-PbO-K2O glass system,” in Proceedings of the 17th International Congress Glass, vol. 7, p. 282, Beijing, China, 1995.
  4. G. Fuxi, “New glass-forming systems and their practical application,” Journal of Non-Crystalline Solids, vol. 123, no. 1–3, pp. 385–399, 1990. View at Scopus
  5. L. D. Bogomolova, “The effects of copper impurity in vanadate and tungsten-phosphate glasses,” Journal of Non-Crystalline Solids, vol. 30, no. 3, pp. 379–383, 1979. View at Scopus
  6. S. Krimi, A. El Jazouli, L. Rabardel, M. Couzi, I. Mansouri, and G. Le Flem, “Glass Formation in the Na2O-TiO2-P2O5 System,” Journal of Solid State Chemistry, vol. 102, no. 2, pp. 400–407, 1993. View at Publisher · View at Google Scholar · View at Scopus
  7. E. M. Vogel, “Glasses as nonlinear photonic materials,” Journal of the American Ceramic Society, vol. 72, no. 5, pp. 719–724, 1989. View at Scopus
  8. L. A. Farrow and E. M. Vogel, “Raman spectra of phosphate and silicate glasses doped with the cations Ti, Nb and Bi,” Journal of Non-Crystalline Solids, vol. 143, no. C, pp. 59–64, 1992. View at Scopus
  9. N. V. Nikonorov and G. T. Petrovskii, “Ion-exchanged glasses in integrated optics: The current state of research and prospects (a review),” Glass Physics and Chemistry, vol. 25, no. 1, pp. 16–55, 1999. View at Scopus
  10. R. K. Brow and D. R. Tallant, “Structural design of sealing glasses,” Journal of Non-Crystalline Solids, vol. 222, pp. 396–406, 1997. View at Scopus
  11. D. K. Sardar, J. B. Gruber, B. Zandi, J. A. Hutchinson, and C. Ward Trussell, “Judd-Ofelt analysis of the Er3+(4f11) absorption intensities in phosphate glass: Er3+, Yb3+,” Journal of Applied Physics, vol. 93, no. 4, pp. 2041–2046, 2003. View at Publisher · View at Google Scholar · View at Scopus
  12. K. Franks, V. Salih, J. C. Knowles, and I. Olsen, “The effect of MgO on the solubility behavior and cell proliferation in a quaternary soluble phosphate based glass system,” Journal of Materials Science: Materials in Medicine, vol. 13, p. 549, 2002. View at Publisher · View at Google Scholar
  13. B. C. Sales and L. A. Boatner, “Lead-iron phosphate glass,” in Radioactive Waste Forms for the Future, W. Lutze and R. C. Ewing, Eds., p. 193, North-Holland, 1988.
  14. A. H. Khafagy, S. M. ElpRabaie, A. A. Higazy, and A. S. Eid, “Absorption spectra of zinc-phosphate glasses doped with Nd2O3 rare earth oxide,” Indian Journal of Physics A, vol. 74, p. 433, 2000.
  15. B. C. Sales and L. A. Boatner, “Optical, structural, and chemical characteristics of lead-indium phosphate and lead-scandium phosphate glasses,” Journal of the American Ceramic Society, vol. 70, no. 9, pp. 615–621, 1987. View at Scopus
  16. W. S. Key and J. C. Miller, “Phosphate glass for photonics,” ORNL Review, vol. 27, no. 3, p. 12, 1994.
  17. I. W. Donald, “Preparation, properties and chemistry of glass- and glass-ceramic-to-metal seals and coatings,” Journal of Materials Science, vol. 28, no. 11, pp. 2841–2886, 1993. View at Publisher · View at Google Scholar
  18. S. F. Anvari, C. A. Hogarth, and G. R. Moridi, “Electrical conductivity of zinc-barium phosphate glasses,” Journal of Materials Science, vol. 26, no. 13, pp. 3639–3642, 1991. View at Publisher · View at Google Scholar · View at Scopus
  19. R. Pyare, L. J. Lal, V. C. Joshi, and V. K. Singh, “Leachability of molybdenum from ternary phosphate glasses,” Journal of the American Ceramic Society, vol. 79, no. 5, pp. 1329–1334, 1996. View at Scopus
  20. G. E. Rachkovskaya and N. M. Bobkova, “Semiconducting niobate-phosphate glasses. Structure and properties,” Journal of Non-Crystalline Solids, vol. 90, no. 1–3, pp. 617–620, 1987. View at Scopus
  21. E. B. De Araujo, J. A. C. De Paiva, M. A. B. De Araujo, and A. S. B. Sombra, “Structure and optical properties of lithium niobium-phosphate glasses and glass ceramics,” Physica Status Solidi B, vol. 197, no. 1, pp. 231–240, 1996. View at Scopus
  22. M. P. F. Graça, M. G. Ferreira da Silva, A. S. B. Sombra, and M. A. Valente, “Electrical and dielectrical properties of SiO2-Li2O-Nb2O5 glass and glass-ceramics obtained by thermoelectric treatments,” Journal of Non-Crystalline Solids, vol. 352, no. 42–49, pp. 5199–5204, 2006. View at Publisher · View at Google Scholar · View at Scopus
  23. M. P. F. Graça, M. A. Valente, and M. G. Ferreira da Silva, “The electric behavior of a lithium-niobate-phosphate glass and glass-ceramics,” Journal of Materials Science, vol. 41, 2006. View at Publisher · View at Google Scholar
  24. T. G. Amos and A. W. Sleight, “Negative thermal expansion in orthorhombic NbOPO4,” Journal of Solid State Chemistry, vol. 160, no. 1, pp. 230–238, 2001. View at Publisher · View at Google Scholar · View at Scopus
  25. A. LeClaire, H. Chahboun, D. Groult, and B. Raveau, “The crystal structure of β-NbPO5,” Zeitschrift für Kristallographie, vol. 177, no. 3, pp. 277–286, 1986. View at Publisher · View at Google Scholar
  26. T. G. Amos, A. Yokochi, and A. W. Sleight, “Phase transition and negative thermal expansion in tetragonal NbOPO4,” Journal of Solid State Chemistry, vol. 141, no. 1, pp. 303–307, 1998. View at Publisher · View at Google Scholar · View at Scopus
  27. L. Bih, L. Abbas, S. Mohdachi, and A. Nadiri, “Thermal and electrical properties of mixed alkali in Li2O-Na2O-WO3-P2O5 glasses,” Journal of Molecular Structure, vol. 891, no. 1–3, pp. 173–177, 2008. View at Publisher · View at Google Scholar · View at Scopus
  28. R. K. Brow, “Review: the structure of simple phosphate glasses,” Journal of Non-Crystalline Solids, vol. 263-264, pp. 1–28, 2000. View at Publisher · View at Google Scholar
  29. U. Hoppe, “A structural model for phosphate glasses,” Journal of Non-Crystalline Solids, vol. 195, no. 1-2, pp. 138–147, 1996. View at Scopus
  30. W. Vogel and W. Höland, “The development of bioglass ceramics for medical applications,” Angewandte Chemie, vol. 26, pp. 527–544, 1987. View at Publisher · View at Google Scholar
  31. A. A. Mc. Connel, J. S. Anderson, and G. N. R. Rao, “Raman spectra of niobium oxides,” Spectrochimica Acta Part A, vol. 32, no. 5, pp. 1067–1076, 1976. View at Publisher · View at Google Scholar
  32. R. D. Shannon, “Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides,” Acta Crystallographica A, vol. 32, pp. 751–767, 1976. View at Publisher · View at Google Scholar
  33. K. Fukumi and S. Sakka, “Coordination state of Nb5+ ions in silicate and gallate glasses as studied by Raman spectroscopy,” Journal of Materials Science, vol. 23, p. 2819, 1988.
  34. T. Cardinal, E. Fargin, G. Le Flem et al., “Non linear optical properties of some niobium (V) oxide glasses,” European Journal of Solid State and Inorganic Chemistry, vol. 33, no. 7, pp. 597–605, 1996. View at Scopus