Study of glass-nanocomposite and glass–ceramic containing ferroelectric phase

Materials Chemistry and Physics 133 (2012) 69–77

Contents lists available at SciVerse ScienceDirect

Materials Chemistry and Physics journal homepage: www.elsevier.com/locate/matchemphys

Study of glass-nanocomposite and glass–ceramic containing ferroelectric phase E.K. Abdel-Khalek a,∗ , E.A. Mohamed b , Shaaban M. Salem a , F.M. Ebrahim a , I. Kashif a a b

Department of Physics, Faculty of Science, Al Azhar University, Nasr City 11884, Cairo, Egypt Department of Physics, Faculty of Science (Girl’s Branch), Al Azhar University, Nasr City, Cairo, Egypt

a r t i c l e

i n f o

Article history: Received 24 May 2011 Received in revised form 4 December 2011 Accepted 21 December 2011 Keywords: A. Glasses C. Differential scanning calorimetry (DSC) D. Optical properties D. Ferroelectricity

a b s t r a c t Transparent glass nanocomposite in the pseudo binary system (100 − x) Li2 B4 O7 –xBaTiO3 with x = 0 and 60 (in mol%) were prepared. Amorphous and glassy characteristics of the as-prepared samples were established via X-ray powder diffraction (XRD) and differential scanning calorimetry (DSC) respectively. The precipitated BaTiO3 nanocrystal phase embedded in the glass sample at x = 60 mol% was identified by transmission electron microscopic (TEM). The optical transmission bands at 598 and 660 nm were assigned to Ti3+ ions in tetragonal distorted octahedral sites. The precipitated Li2 B4 O7 , BaTi(BO3 )2 and BaTiO3 nanocrystallites phases with heat-treatment at 923 K for 6 h (HT923) in glass–ceramic were identified by XRD, TEM and infrared absorption spectroscopy. The as-prepared at x = 60 mol% and the HT923 samples exhibit broad dielectric anomalies in the vicinity of the ferroelectric-to-paraelectric transition temperature. The results demonstrate that the method presented may be an effective way to fabricate ferroelectric host and development of multifunctional ferroelectrics. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Recently, transparent glass nanocomposites and glass–ceramics comprised of ferroelectric components in the nanometer range find a countless applications [1–4]. In addition to glass–ceramics are one type of interest hosts, as the glass matrix is favorable in its capability of fiber fabrication, in the meantime nanocrystals in glass–ceramics can provide active sites for transition metal (TM) ions [5]. Thus the luminescent properties of the optical materials are directly correlated with the valence state of the active ions such as Ni2+ [6]. From previous studies it is found that the transparent glass–ceramics are of interest as hosts for Ni2+ ions to realize ultra-broadband optical amplification [7–9]. Zhou et al. [10] showed that the nanocrystalembedded hybrid materials are employed as hosts in order to take advantage of their convenience in local environment design for practical applications. Hao et al. [11] showed that the ferroelectric host (BaTiO3 ) controls the luminescent properties of the optical materials through the change in the structural symmetry. Nowadays, the preparation of numerous ferroelectric materials by glass crystallization has been described in the literature [1,3]. Barium titanate (BaTiO3 ) is a very important and interesting ferroelectric material for applications in electronic devices such as capacitors, electro-optic devices, radio communication filters [2,4]. BaTiO3 has poor glass forming ability and high melting point thus in order to obtain a glassy material, it is usually required to add a glass former.

∗ Corresponding author. E-mail address: Eid [email protected] (E.K. Abdel-Khalek). 0254-0584/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.matchemphys.2011.12.053

In the present study, we fabricate transparent glass nanocomposite and glass–ceramic in the pseudo binary system (100 − x) Li2 B4 O7 ·xBaTiO3 with x = 0 and 60 (in mol%). The reason for choosing Li2 B4 O7 is its relatively low melting point and its ability to form a glass readily by conventional melt quenching techniques [12]. In addition to Li2 B4 O7 is a non-ferroelectric and piezoelectric material [13]. In order to obtain transparent glass nanocomposite, the nanocrystal phase embedded in the glass should be much smaller than the wavelength of the visible light (200 nm) [14]. The structural, optical and dielectric properties of the present samples have been studied. 2. Experimental A transparent glasses with the molar composition (1 − x) Li2 B4 O7 + xBaTiO3 (x = 0 and 60 (in mol%)) were prepared by conventional melt quenching technique. The glass samples under investigation were prepared from reagent grade Li2 B4 O7 (99.99%) and BaTiO3 (99.99%). The batches were melted in a platinum crucible at 1100–1150 ◦ C for 1 h. The melts were poured on to a copper plate and immediately pressed into plates. The differential scanning calorimetry (DSC) for glass samples were carried out on a SETARAM LabsysTM TG-DSC16 thermal analyser in the 303–1173 K temperature range. The as-prepared sample (at x = 60 mol%) was heat-treated in air, at 923 K (HT923) during 24 h. X-ray diffraction studies were performed at room temperature (RT) from a Siemens D5000 diffractometer using Cu K␣ radiation. Rietveld analysis of the diffraction data was performed using the FULLPROF program. High resolution transmission electron microscopy (TEM) studies

70

E.K. Abdel-Khalek et al. / Materials Chemistry and Physics 133 (2012) 69–77 Table 1 Structural parameters for heat-treated (923 K/24 h) obtained from the structural refinement using X-ray powder diffraction data at room temperature.

Tcr X=0

Exo

Lattice constants

ΔT

Tg

Tm

Endo

X=60

Tcr

Tg

Tm3

Atom

Tetragonal Li2 B4 O7 (I41 cd) a = b = 9.4608 A˚ Li c = 10.2768 A˚ B1 B2 O1 O2 O3 ¯ Rhombohedral BaTi(BO3 )2 (R3) a = b = 5.0005 A˚ Ba c = 16.3744 A˚ Ti B O Tetragonal BaTiO3 (P4mm) a = b = 4.3178 A˚ Ba c = 4.1061 A˚ Ti O O Phasefraction (%) Tetragonal (I41 cd) 81.14

Site

Lattice coordinate x

y

z

16b 16b 16b 16b 16b 8a

0.21102 0.08663 0.18005 0.14193 0.17140 −0.06105

0.26737 0.16475 −0.06861 0.28504 0.01104 0.15917

0.85901 0.22896 0.11834 0.25570 0.14720 0.18006

3a 3b 6c 18f

0.00000 0.00000 0.00000 0.31320

0.00000 0.00000 0.00000 0.12710

0.00000 0.50000 −0.22520 −0.23650

1a 1b 1b 2c

0.00000 0.50000 0.50000 0.50000

0.00000 0.50000 0.50000 0.00000

0.00000 0.48530 0.00210 0.50910

Rhombohedral R3¯ 11.76

Tetragonal (P4mm) 7.10

Tm1Tm2 0

150

300

450

600

750

900

o

Temperature ( C) Fig. 1. Differential scanning calorimetry for the as-prepared samples.

were carried out on the as-prepared (at x = 60 mol%) and HT923 samples. The IR absorption spectra of the glasses in the wavenumber range of 450–1600 cm−1 were recorded at room temperature on a Bruker (Vector 22), single beam spectrometer with a resolution of 2 cm−1 . The IR absorption spectra were recorded using KBr pellets. The optical transmission spectra of the as-prepared polished samples were recorded in the 200–750 nm wavelength range using a Hitachi 6405 spectrophotometer. Ac electrical conductivity and dielectric properties of the samples were measured as a function of composition, temperature and frequency. The ac data were collected using electronic RLC bridge type SR 720. 3. Results and discussion 3.1. DSC Fig. 1 shows the DSC curves that were obtained for the glasses under investigation. The glass sample at x = 0 mol% exhibits endothermic minima which represent the glass transition temperature Tg (481 ◦ C) and confirm the glassy nature of the sample. This sample exhibits also an exothermic peak Tcr (575 ◦ C) which due to the crystal growth followed by an endothermic effect due to the remelting of the glass symbolized by Tm (874 ◦ C). The appearance of a single peak due to the glass transition temperature in DSC pattern of glass sample indicate the existence of high homogeneity in the glass sample [1]. The differential thermogram that was obtained for x = 60 mol% shows a broad endothermic which represent the glass transition temperatures Tg at 533 and confirm the presence of the borate vitreous phase and titanate vitreous phase [15]. At still higher temperatures a broad exothermic peak Tcr (678 ◦ C) due to the crystal growth, followed by three endothermic effect at 756, 820 and 866 which are related to the crystallization of different phases. The temperature difference between the Tg and Tcr (DT = Tcr − Tg ), gives a measure for the thermal stability of the glass against

crystallization [16,17]. The DT value increases with introducing BaTiO3 . It can also be seen that the width of exothermic peak increases with introducing BaTiO3 . This widening may be attributed to the slowing down of crystallization process [18]. 3.2. X-ray diffraction and TEM studies The XRD patterns that were obtained for the as-prepared at x = 60 mol% as well as HT923 samples are depicted in Fig. 2. As shown in Fig. 2(a) the XRD pattern exhibits a broad hump without any distinct peak in the as-prepared at x = 60 mol% indicates the amorphous nature. Rietveld refinement of the data Fig. 2(b) reveals that the sample crystallizes in an tetragonal (I41 cd) Li2 B4 O7 , ¯ BaTi(BO3 )2 and tetragonal (p4mm) BaTiO3 conrhombohedral (R3) sistent with the previously reported results [3,19,20]. But there is a noticeable shift in the peak positions than those of the pure powder sample of Li2 B4 O7 , BaTi(BO3 )2 and BaTiO3 . This can be attributed to the existence of uneven distribution of strain which arising out of the anisotropic growth of Li2 B4 O7 , BaTi(BO3 )2 and BaTiO3 crystals in glass matrix [3]. The quantitative phase analysis of powders and refined values of the lattice and positional parameters at room temperature are summarized in Table 1. The broad nature of the diffraction peaks in Fig. 2(b) indicates that the presence of very fine crystallites of Li2 B4 O7 , BaTi(BO3 )2 and BaTiO3 [10]. It is also noted that there are Li2 B4 O7 , BaTi(BO3 )2 and BaTiO3 crystals in glass matrix is not perfectly stoichiometric, but contains a certain amount of oxygen vacancies where vacancies are formed after the removal of oxygen atoms that bridge the adjacent tetrahedra, resulting in the formation of distorted octahedral environments [10]. The transmission electron micrographs recorded for the asprepared at x = 60 mol% as well as HT923 samples are shown in Fig. 3. The micrograph recorded for the as-prepared glass (Fig. 3(a)) shows its overall amorphous nature with the presence of small BaTiO3 nanoclusters which precipitated during glass formation. Hence the as-prepared sample at x = 60 mol% is a glassnanocomposite. These small BaTiO3 nanoclusters are not clearly detected from the XRD patterns. These results is due to their low concentrations (low for X-ray detection) embedded in the glass matrix and small size also [3,21]. These nanoclusters are not clustered in one portion of the glass but scattered all over the

E.K. Abdel-Khalek et al. / Materials Chemistry and Physics 133 (2012) 69–77

71

Fig. 2. (a) XRD patterns for the as-prepared at x = 60 mol%. (b) Rietveld plot of XRD for the HT923.

glass matrix covered by glassy phase [3,21]. These results are in line with those reported for ferroelectricity in multicomponent Bi1.8 Pb0.3 Sr2 Ca2 Cu2.8 K0.2 Oı glass by Mukherjee et al. [3,21]. The micrograph recorded for HT923 (Fig. 3(b)) shows the existence of larger nanocrystals Li2 B4 O7 , BaTi(BO3 )2 and BaTiO3 dispersed in the glass matrix which confirmed by XRD. The presence of the nanocrystalline in the as-prepared glass may arise from the formation of phase separations prior to the crystallization of nanocrystals Li2 B4 O7 , BaTi(BO3 )2 and BaTiO3 crystals in the HT923 sample. 3.3. Infrared spectral studies Fig. 4 shows the infrared absorption spectra for the as-prepared as well as HT923 samples. The vibration modes of the borate network are mainly active in three IR spectral regions (i) in the region 1200–1600 cm−1 , (ii) in the region 800–1200 cm−1 and (iii) in the region 600–800 cm−1 . The first region of bands is attributed to the

stretching relaxation of the B O bond of the trigonal BO3 units while the second region of bands is attributed to BO4 units and the third region of bands is due to the bending of B O B linkages in the borate network [22,23]. From Fig. 4 it can be seen that the bands are very broad, asymmetric and presenting also some shoulders. The observed broadening of the bands in the spectra of the glasses may arise for two reasons. The first is the distribution of bond angles and bond lengths and fluctuations of the local electronic and atomic environments in the amorphous state [24]. The second reason is the overlapping of some individual bands with each other. To get quantitative information about the structural groups in samples, the deconvolution of the experimental spectra was necessary. Fig. 4 shows the deconvolution, in Gaussian bands, spectra of the samples. In order to get best fitting it is found that the two beaks (409 for x = 60 mol% and 430 for HT923) are predicted out of range measurement. The deconvolution parameters, the band centers C and the relative area A are given in Table 2 for

72

E.K. Abdel-Khalek et al. / Materials Chemistry and Physics 133 (2012) 69–77

Absorbance (arbit. u.)

x=0

x = 60

HT923

1600

1400

1200

1000

800

600

-1

Wavenumber (cm ) Fig. 4. Band deconvoluted IR spectra for the as-prepared and HT923 sample, using a Gaussian-type function.

Fig. 3. Transmission electron micrographs of: (a) the as-prepared at x = 60 mol%. (b) The HT923 sample.

the studied samples. From table it can be seen that the analysis of the deconvoluted IR bands of the sample at x = 0 mol% revealed the presence of the bands at 1330, 1416 and 1517 cm−1 on the first group and the appearance of these bands can be taken as an evidence for the presence of the nonbridging oxygen [25]. The band at 1330 cm−1 is assigned to the stretching vibrations of the B O of trigonal (BO3 )3− units in pyroborates. The band at 1416 cm−1 is assigned to antisymmetrical stretching vibrations with three NBOs of the B O B groups. The band at 1517 cm−1 is assigned to asymmetric stretching relaxation of the B O band of trigonal BO3 units [26]. The band at 1089 cm−1 is assigned to vibrations of pentaborate groups and the band around 977 cm−1 has been assigned to diborate groups. The band at 901 cm−1 is attributed to the stretching vibrations of tetrahedral BO4 units [27]. The band at 694 cm−1 is due to the bending of B O linkages in the borate network [28]. The spectrum of glass sample at x = 60 mol% exhibits new bands at 409, 625 and 722 cm−1 . The presence of a vibrational band at about 722 cm−1 due to vibrations of TiO4 groups [29,30]. The band at 625 cm−1 is assigned to the stretching vibrations of Ti O bonds in TiO6 octahedra [29,31]. The band at 409 cm−1 is attributed to

Ti O Ti vibrations [32]. Hence, there is a possibility for the formation of single boron–oxygen–titanium framework in the glass network with the introducing of BaTiO3 in to the Li2 B4 O7 glass network. Thus tetragonally positioned Ti4+ ions do not induce the formation of any nonbridging oxygen ions but octahedrally positioned ions may act as modifiers [29,33]. The TiO4 tetrahedrons enter the glass network and also alternate with borate structural units and form linkages of the type B O Ti [29]. Titanium ions seem to exist mainly in Ti4+ state in the glass sample at x = 60 mol%. However, the reduction of Ti4+ to Ti3+ appears to be viable during melting process of the present glass sample [29]. The spectrum of HT923 sample exhibits new bands at 480 and 579 cm−1 . The band observed at 658 cm−1 is due to Ti O Ti symmetric stretching vibrations of such TiO6 units [29,31]. The broad band at 740 cm−1 is due to vibrations of TiO4 groups, in this case Table 2 Deconvolution parameters of the infrared spectra for the as-prepared: (a) x = 60 mol%, and (d) heat-treated (923 K/24 h). C is the component band center (cm−1 ) and A is the relative area (%) of the component band. X=0

X = 60

HT923

C

A

C

A

C

A

– – – 694 – 901 977 1089 1330 1416 1517

– – – 25.8 – 102.7 115.7 54.7 69.0 78.0 38.1

409 – 625 – 722 866 969 1074 1217 1310 1429

164.1 – 111.8 – 139.0 101.8 91.7 71.2 24.1 81.5 74.4

432 480 579 658 740 865 954 1062 1226 1371 –

48.9 54.2 32.4 42.2 14.0 74.4 63.0 98.1 64.4 139.8 –

E.K. Abdel-Khalek et al. / Materials Chemistry and Physics 133 (2012) 69–77

Table 3 The optical absorption edge cut-off , the optical band gap Eopt and Urbach energy (E0 ), for all glass samples.

80

Transmission %

70

73

x=0 x=60

60 50

X

cut-off (nm)

Eopt (eV)

E0 (eV)

0 60

275 350

3.23 3.07

0.538 0.545

40 30

x=0

30

x=60

20 25 1/2

eV )

10

200

300

400

500

600

700

800

-1/2

0 100

(αhν)(cm

λ (nm) Fig. 5. The optical transmission spectra for all glass samples.

20 15 10 5 0 1.2 1.5 1.8 2.1 2.4 2.7 3.0 3.3 3.6 3.9 4.2 4.5 4.8 5.1 5.4 5.7 6.0 6.3 6.6 6.9

hν (eV) Fig. 6. The variation of the (˛h)1/2 with h for all glass samples.

3.4

x=0 x = 60

3.2 3.0

Ln (α)

it may be assumed due to the vibrations of B O Ti linkages, and the band at 579 cm−1 is assigned to the O6 octahedra deformation mode of the BaTiO3 lattice [34]. The band at 480 cm−1 is assigned to vibration of the LiO4 group, and the band at 432 cm−1 is attributed to vibration of Li O network in Li2 B4 O7 crystal [35]. The strong modes for heat treated sample at x = 60 mol%. (see Table 2 and Fig. 4) are very indicative of the presence of the crystalline phase. The relative area of the first group bands (1517 and 1416 cm−1 ) increases and is opposite to the second group (901 and 977 cm−1 ) and are shifted to lower wavenumbers (Table 2) with the introducing of BaTiO3 on expense of Li2 B4 O7 . The replacement of Li2 B4 O7 by BaTiO3 causes a decrease in the number of oxygen ions in the network, which in turn causes a decrease in BO4 and increases the number of nonbridging oxygen. While this behavior is opposite with heat-treating the representative glass sample at x = 60 mol%. The relative area of the band at 625 decreases and are shifted to higher wavenumbers. In addition to, the appearing of the band at 579 cm−1 is very indicative of the presence of the crystalline phase.

2.8 2.6 2.4 2.2 2.0 3.1

3.2

3.4. Optical properties The optical transmission spectra for the as-prepared samples (Fig. 5) were recorded at room temperature in the wavelength range 200–750 nm. The HT923 sample is found to be opaque, this means that the optical transmission measurements was not allowed. From Fig. 5 it can been that the sample at x = 60 mol% have lower levels of transparency. The former glass at x = 60 mol% was slightly yellow colored, indicating the presence of a small amount of Ti3+ [36]. The spectrum of glass sample at x = 0 mol% does not exhibit any transmission bands. The spectrum of glass sample at x = 60 exhibits two transmission bands at 598 and 660 nm corresponding to 2 B2g → 2 B1g and 2 B2g → 2 A1g transitions of 3d1 electron of the Ti3+ ions in tetragonal distorted octahedral sites [29]. The shift of the optical band at 598 to longer wavelengths indicates a weaker ligand field of Ti3+ in this sample [37]. From these optical characteristics, we can suggest that the BaTiO3 nanocrystalline as ferroelectric host for glass nanocomposite at x = 60 mol%, which was confirmed by the TEM [11]. The cut-off is the wavelength at which the percentage transmission is zero. This cut-off shifts towards longer wavelengths of the spectra with introducing BaTiO3 on expense of Li2 B4 O7 (Table 3). The absorption coefficient, ˛, related to the light that is transmitted out of a sample of thickness t, is given by, It = I0 exp[˛t]

(1)

3.3

3.4

hν (eV)

3.5

3.6

Fig. 7. The variation of the (Ln ˛) with h for all glass samples.

where I0 and It are the intensities of incident and transmitted radiation, respectively. Absorption coefficient follows the empirical relation [38].



˛() = B

(h − Eopt ) h

2



(2)

where B is a constant, h is the incident photon energy and Eopt is the optical band gap. The values of Eopt for the samples have been calculated by extrapolating the linear region of the curves to meet the h axis at (˛h)1/2 = 0 as shown in Fig. 6 and are presented in Table 3. The band tail associated with valence band and conduction band extend into band gap and show an exponential behavior. The band tails are characterized by the band tail parameter E0 (Urbach energy). The values of E0 can be obtained from the following relation [39]: ˛() = C + exp

 h  E0

(3)

where C is a constant. The values of E0 were found as the inverse slope of the ln ˛ versus h plot as shown in Fig. 7 and are presented in Table 3. From Table 3 it can be seen that the Eopt decreases while

74

E.K. Abdel-Khalek et al. / Materials Chemistry and Physics 133 (2012) 69–77

the Urbach energy increases with introducing BaTiO3 . These can be attributed to the increase of non-bridging oxygen ions which binds excited electrons less tightly than bridging oxygen [1]. Thus the transformation from bridging to NBO ions led to the raising the top of the valence band, resulting in the reduction of (Eopt ) [40].

106

0.12 kHz

5

10

1

kHz

104

10

kHz

103

100 kHz

x=0

102 101

3.5. Dielectric studies Usually, electronic, ionic, dipolar and space charge polarizations contribute to the dielectric constant (εr ) [1,3]. The variation of the dielectric constant as a function of temperature at various (120 Hz–100 kHz) frequencies for the as-prepared and HT923 are depicted in Fig. 8. The dielectric constant (Fig. 8) becomes larger at both lower frequencies and higher temperatures which is normal in oxide glasses and is not an indication for spontaneous polarization [1,41]. This may be due to the fact that as the frequency increases, the polarizability contribution from ionic and orientation sources decreases and finally vanishes due to the inertia of the ions [1,41]. At low temperatures, the contribution of electronic and ionic components to the total polarizability will be small. As the temperature is increased the electronic and ionic polarizability sources start to increase [1,42]. The peak in the 400–650 K temperature range at x = 0 mol% is ascribed to lithium ion hopping, as it has been reported in pure lithium borate single crystals and glasses [43,44]. It is noticed that the peak position in εr shifts towards higher temperatures as the frequency increases (Fig. 8 for x = 0 mol%), which is similar to that observed in the case of relaxor ferroelectrics [45] near Tc . However, the basic difference between typical relaxors and the present sample at x = 0 mol% is the incidence of the dielectric anomaly in the vicinity of the crystallization temperature (Tcr ) [46,47]. The as-prepared at x = 60 mol% as well as HT923 samples (Fig. 8) exhibit the anomaly around 309 and 305 K respectively which is attributed to the phase transition (Curie) temperature of the BaTiO3 phase. Fig. 9(a) shows the dielectric constant εr as a function of temperature at 1 kHz for the as-prepared at x = 60 mol% as well as HT923 samples. This curve can be divided into three regions. The first region represents a ferroelectric behavior up to the transition Tc , the second region near the transition

66

εr

1

kHz

10

kHz

x = 60

100 kHz

102

HT923

0.12 kHz

103

1

kHz

10

kHz

100 kHz

102

300

350

400

450

500

550

600

650

700

Temperature (K) Fig. 8. Variation of εr with temperature for the samples.

indicates a diffuse transition up to a temperature Tcw and the third region represents the paraelectric phase [12]. Several remarked variations are observed at x = 60 mol% as well as HT923 samples. Only one peak was observed corresponding to the phase transition of paraelectric–ferroelectric at TC , dielectric maximum decrease and no shift in TC with frequency [48,49]. Thus the as-prepared at x = 60 mol% and HT923 are normal ferroelectrics. Furthermore, the

x=60 HT923

(a)

0.12 kHz

103

0.021

(b)

64 0.020

62

Tc

60 0.019

Tcw

58

x = 60 HT923

0.018

1/εr

εr

56 54

0.017 52

Tc

50

0.016

48

Tcw 0.015

46 290

300

310

320

Temperature (K)

330

340

295

300

305

310

315

320

325

330

Temperature (K)

Fig. 9. (a) Variation of εr with temperature for the as-prepared and HT923 samples at 1 kHz. (b) Plot of 1/εr versus T and the fitting plot by Curie–Weiss law for the as-prepared and HT923 samples at 1 kHz.

E.K. Abdel-Khalek et al. / Materials Chemistry and Physics 133 (2012) 69–77 -2.4

100

x = 60 HT923

-2.6

log ( 1/εr-1/εmax )

-2.8

75

0.12 kHz

80

1

kHz

60

10

kHz

40

100 kHz

x=0

20 -3.0

0

-3.2

tan δ

-3.4 -3.6

4

0.12 kHz

3

1

kHz

10

kHz

2

x = 60

100 kHz

1 0.75 0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15 1.20 1.25 1.30

log ( T-Tc ) Fig. 10. Plot of log(1/εr − 1/εmax ) versus log(T − Tc ) for the as-prepared and HT923 samples at 1 kHz.

broadening of the εr peak can be related to a non-homogeneous and to the existence of internal stresses [48,49]. From Fig. 9(a) it can be seen that the broadening of the dielectric peak and εmax decreases with HT. This decreases can be attributed to the increase in the grain size [48]. For normal ferroelectrics, the dielectric constant above the Curie point follows the Curie–Weiss law described by εr =

C (T − T0 )

0 0.12 kHz

8 6

1

kHz

10

kHz

HT923

100 kHz

4 2 0 300

350

400

450

500

550

600

650

700

Temperature (K) Fig. 11. Variation of tan ı with temperature for the samples.

T > Tc

where T0 is the Curie–Weiss temperature and C is the Curie constant. Fig. 9(b) shows plot of the reciprocal of dielectric constant as a function of temperature at 1 kHz and the fitting plot by Curie–Weiss law. The parameters, Curie–Weiss temperature (T0 ), Curie temperature (TC ), Curie constant (C) and the difference (TC − T0 ), obtained from the fitting and Tcw are listed in Table 4. Smolenskii [50] suggested the deviation of TC − T0 determined the order of phase transition, i.e. TC − T0 > 0 indicated the ferroelectric–paraelectric phase transition was of first-order type while TC − T0 = 0 with a broad dielectric constant peak was of a second-order transition. Based on the above results, the as-prepared at x = 60 mol% as well as HT923 obey the TC − T0 > 0, showing the first-order type phase transition behavior. The difference Tm = Tcw − Tc is indicative of the thermal diffuseness. In order to describe the diffuseness of the phase transition temperature in the glass nanocomposite and glass–ceramics [12,49] the reciprocal of the dielectric constant and temperature can be fitted to the relation: 1 1 − = (C ∗ )−1 (T − Tc ) εr εmax where C* is the Curie-like constant, εmax is the peak dielectric constant and  is the critical exponent which has the value 1 <  < 2 for normal ferroelectric and  ≥ 2 for an ideal relaxor ferroelectric. The plots of log(1/εr − 1/εmax ) as a function of log(T − Tc ) for the as-prepared at x = 60 mol% as well as HT923 samples are shown in Fig. 10. A linear relationship is observed for all samples. The slope of the fitting curves is used to determine the value of , the indicator of degree of diffuseness. The  value changes from 2.42 to 1.57 with HT923 (Table 4), indicating that a relaxor ferroelectric changes to a normal ferroelectric [12,49]. C* (from the intercept) were computed using linear regression. The results are shown in Table 4. The diffuseness in the dielectric property in the vicinity of the phase transition temperature for the as-prepared at x = 60 mol% as well as HT923 may be attributed to the irregular distribution of charged defects/impurities which may lead to large scale potential/polarization fluctuations [12]. The relaxor behavior

for the as-prepared at x = 60 mol% can be induced by many reasons such as microscopic regions into macropolar regions, or coupling of the order parameter and local disorder mode through the local strain [49]. The decrease in the value of  with HT923 suggests that the compounds become more ordered [49]. Fig. 11 shows the frequency response of dielectric loss (tan ı) at different temperatures for the as-prepared and HT923 samples. From this figure two types of peaks can be seen at x = 0 mol%. The first peak was observed to shift towards higher temperature with increasing frequency. This peak is attributed to the Debye-type relaxation process. The variation of tan ı for as-prepared at x = 60 mol% as well as HT923 with temperature (Fig. 11) shows an anomaly near the transition temperature Tc . From Fig. 11 it can be seen that the dielectric loss decreases with introducing BaTiO3 on expense of Li2 B4 O7 in paraelectric state. The dielectric loss in lithium borate glass is mainly attributed to the lithium ion conduction. The presence of BaTiO3 in Li2 B4 O7 glass hinders the hopping of lithium ions and hence the conductivity and the dielectric loss at x = 60 mol% are lower than that of the pure Li2 B4 O7 glass [51]. The decrease in both εr and tan ı for glass samples with increase in frequency could be explained by using Stevels and Taylor model, which was proposed to understand the dielectric relaxation and ionic conductivity behavior of alkali silicate glasses [52,53]. 3.6. AC conductivity The reciprocal temperature dependence of the ac conductivity ( ac ) as a function of temperature and frequency for the asprepared and HT923 samples are shown in Fig. 12. It is clear from this figure that  ac increases linearly with decreasing the reciprocal of absolute temperature at high temperature and non-linear at low temperature at x = 0 mol%. This suggested that the ac conductivity is a thermally activated process from different localized states in the gap or its tails at high temperature [1]. An anomaly has been observed at a particular temperature for the as-prepared at x = 60 mol% as well as HT923 samples. The change in slope at a particular temperature which corresponds to the transition

76

E.K. Abdel-Khalek et al. / Materials Chemistry and Physics 133 (2012) 69–77

Table 4 Parameters Tc , T0 , Tcw , Tm , C,  and C* for the as-prepared and heat-treated samples at 1 kHz.

60 HT923

-2 -4 -6 -8 -10 -12 -14 -16

Tcw (K)

T0 (K)

C

C*



Tm (K)

309 305

317 311

234 246

5.29 × 103 3.42 × 103

4.14 × 105 1.81 × 104

2.42 1.57

8 6

single-phase oxide glass has not yet been discovered. In essence the present investigations confirm that glass and glass–ceramics comprising ferroelectric nanocrystallites with diffused phase transitions. This transparent glass nanocomposite may have potential applications in fabricate ferroelectric host and development of multifunctional ferroelectrics.

0.12 kHz

x=0

1

kHz

10

kHz

100 kHz

-8

lnσac(Ω m-1)

Tc (K)

0.12 kHz

x = 60

-10

1

kHz

10

kHz

References

100 kHz

-12 -14 -16 -18 -8

0.12 kHz

HT923

-10

1

kHz

10

kHz

100 kHz

-12 -14 -16 1.5

2.0

2.5

3.0

3.5

-1

1000/T (K ) Fig. 12. Ac conductivity (ln  ac ) versus (1000/T), at different frequencies for samples.

temperature in the samples obtained from εr versus temperature [3]. The change in slope of curve will reflect a change in the conductivity phenomenon in paraelectric and ferroelectric regions [54]. From Fig. 12. it can been that the  ac at all studied frequencies decreased with introducing BaTiO3 . This is due to the decrease of the number of lithium ions and the presence of BaTiO3 nanocrystal in glass hinders the hopping of lithium ions. In addition to the presence amount of crystalline phase in HT923, whose dipoles are difficult to depolarize [1]. This decrease is related with the formation of Li2 B4 O7 , BaTi(BO3 )2 and BaTiO3 crystallites and with the consequent decrease of the lithium ions number structurally inserted in the glass network. Nevertheless, the presence of the Li2 B4 O7 crystals, which according to [1,55] presents a very low conductivity. 4. Conclusions Transparent glass nanocomposite and glass–ceramic were fabricated. The precipitated BaTiO3 nanocrystal phase in the glassmatrix at x = 60 mol% was identified by TEM. After heat-treatment (HT923), Li2 B4 O7 , BaTi(BO3 )2 and BaTiO3 nanocrystallites phases precipitated from the glass matrix at x = 60 mol%. These nanocrystallites phases in the glass–ceramic were identified by XRD and confirmed by TEM and infrared spectroscopy. The observed optical transmission bands at 598 and 660 nm could be assigned to the transition of 2 B2g → 2 B1g and 2 B2g → 2 A1g of tetragonal distorted octahedral Ti3+ ions in glass nanocomposite. Both glass nanocomposite and glass–ceramic exhibit broad dielectric anomalies in the vicinity of the ferroelectric–paraelectric transition temperature of the BaTiO3 crystalline phase. Thus the ferroelectricity in pure

[1] E.K. Abdel-Khalek, A.A. Bahgat, Phys. B 405 (2010) 1986. [2] Y.-I. Kim, K.-S. Ryu, S.-H. Nahm, J.-S. Park, Curr. Appl. Phys. 651 (2006) 266. [3] E.K. Abdel-Khalek, S.M. Salem, M. Farouk, E.A. Mohamed, I. Kashif, J. Non-Cryst. Solids 357 (2011) 864. [4] J. Chena, Y. Zhangb, C. Dengb, X. Dai, Mater. Chem. Phys. 121 (2010) 109. [5] B. Wu, N. Jiang, S. Zhou, D. Chen, C. Zhu, J. Qiu, Opt. Mater. 30 (2008) 1900. [6] P. Mingying, D. Ning, Q. Yanbo, W. Baotao, W. Chen, C. Danping, Q. Jianrong, J. Rare Earths 24 (2006) 749. [7] G. Feng, S. Zhou, J. Bao, X. Wang, S. Xu, J. Qiu, J. Alloys Compd. 457 (2008) 506. [8] T. Suzuki, Y. Ohishi, Appl. Phys. Lett. 84 (2004) 3804. [9] T. Suzuki, G.S. Murugan, Y. Ohishi, Appl. Phys. Lett. 86 (2005) 131903. [10] S. Zhou, N. Jiang, B. Wu, J. Hao, J. Qiu, Adv. Funct. Mater. 19 (2009) 2081. [11] J. Hao, Y. Zhang, X. Wei, Angew. Chem. Int. Ed. 50 (2011) 6876. [12] N.S. Prasad, K.B.R. Varma, S.B. Lang, J. Phys. Chem. Solids 62 (2001) 1299. [13] S.J. Kim, J.E. Kim, H.W. Choi, Y.H. Rim, Y.S. Yang, Mater. Sci. Eng. B 113 (2004) 149. [14] P. Prapitpongwanich, R. Harizanova, K. Pengpat, C. Rüssel, Mater. Lett. 63 (2009) 1027. [15] J.J.R. Valdes, A.V. Gorokhovsky, I.J.E. Garcia, J. Non-Cryst. Solids 351 (2005) 2036. [16] S. Sakka, J.D. Mackenzie, J. Non-Cryst. Solids 6 (1991) 145. [17] A. Bajaj, A. Khanna, B. Chen, G.J. Longstaffe, U-W. Zwanziger, J.W. Zwanziger, Y. Gmez, F. Gonzlez, J. Non-Cryst. Solids 355 (2009) 45. [18] O.P. Thakur, D. Kumara, O. Parkas, L. Pandey, Mater. Chem. Phys. 78 (2003) 751. [19] G. Arlt, D. Henning, D. de With, J. Appl. Phys. 58 (1985) 1619. [20] S.Y. Zhang, X. Wu, X.L. Chen, M. He, Y.G. Cao, Y.T. Song, D.Q. Ni, Mater. Res. Bull. 38 (2003) 783. [21] S. Mukherjee, B.K. Chaudhuri, H. Sakata, Phys. Rev. B 68 (2003) 016101. [22] F.A. Khalife, H.A. El Batal, A.A. Zooz, J. Pure Appl. Phys. 36 (1998) 314. [23] M.R. Reddy, S.B. Raju, N. Veeraiah, J. Phys. Chem. Solids 61 (2000) 1567. [24] I.S. Yahia, Y.B. Saddeek, G.B. Sakr, W. Knoff, T. Story, N. Romcevic, W. Dobrowolski, J. Magn. Magn. Mater. 321 (2009) 4039. [25] E.I. Kamitso, M.A. Karakassides, A.P. Patsis, J. Non-Cryst. Solids 111 (1989) 252. [26] G. Lakshminarayana, S. Buddhudu, Spectrochim. Acta A 63 (2006) 295. [27] E. Kamitsos, M. Karakassides, G. Chryssikos, J. Phys. Chem. 91 (1987) 1073. [28] R.B. Rao, R.A. Gerhardt, N. Veeraiah, J. Phys. Chem. Solids 69 (2008) 2813. [29] P. Nageswara Rao, C. Laxmi Kanth, D. Krishna Rao, N. Veeraiah, J. Quant. Spectrosc. Radiat. Transfer 95 (2005) 373. [30] D. Laudisio, M. Catauro, A. Aronne, P. Pernice, Thermochim. Acta 94 (1997) 173. [31] V.N. Sigaev, P. Pernice, A. Aronne, O.V. Akimova, S. Yu Stefanovich, A. Scag, J. Non-Cryst. Solids 292 (2000) 59. [32] D.M. Tobaldi, A. Tucci, A.S. Skapin, L. Esposito, J. Eur. Ceram. Soc. 30 (2010) 2481. [33] B.V. Raghavaiah, C. Laxmikanth, N. Veeraiah, Opt. Commun. 235 (2004) 341. [34] K. Suzuki, K. Kijima, J. Alloys Compd. 419 (2006) 234. [35] L. Lei, D. He, K. He, J. Qin, S. Wang, Solid State Chem. 182 (2009) 3041. [36] V.N. Sigaev, P. Pernice, A. Aronne, O.V. Akimova, S. Yu Stefanovich, A. Scaglione, J. Non-Cryst. Solids 292 (2001) 59. [37] A. Aboukais, L.D. Bogomolova, A.A. Deshkovskaya, V.A. Jachkin, N.A. Krasil’nikova, S.A. Prushinsky, O.A. Trul, S.V. Stefanovsky, E.A. Zhilinskaya, Opt. Mater. 19 (2002) 295. [38] E.A. Davis, N.F. Mott, Philos. Mag. 22 (1970) 903. [39] F. Urbach, Phys. Rev. 92 (1953) 1324. [40] I.Z. Hager, J. Phys. Chem. Solids 70 (2009) 210. [41] T. Sankarappa, M. Prashant Kumar, G.B. Devidas, N. Nagaraja, R. Ramakrishnareddy, J. Mol. Struct. 889 (2008) 308. [42] A. Mogus-Milankovic, V. Licina, S.T. Reis, D.E. Day, J. Non-Cryst. Solids 353 (2007) 2659. [43] K.J. Charles, J. Appl. Phys. 32 (1961) 1115. [44] S.R. Elliott, Solid State Ionics 27 (1988) 131. [45] T.R. Shrout, A. Halliyal, Am. Ceram. Soc. Bull. 66 (1987) 704. [46] N. Tsuya, K.I. Arai, Jpn. J. Appl. Phys. 18 (1979) 461.

E.K. Abdel-Khalek et al. / Materials Chemistry and Physics 133 (2012) 69–77 [47] K.B.R. Varma, K.S. Harshavardhan, K.J. Rao, C.N.R. Rao, Mater. Res. Bull. 20 (1985) 315. [48] Z. Yao, H. Liu, Y. Liu, Z. Wu, Z. Shen, Y. Liu, M. Cao, Mater. Chem. Phys. 109 (2008) 475. [49] N. Abdelmoula, H. Khemakhem, A. Simon, M. Maglione, Solid State Chem. 179 (2006) 4011. [50] G.A. Smolenskii, J. Phys. Soc. Jpn. 28 (1970) 26.

[51] [52] [53] [54]

77

G. Senthil Murugan, K.B.R. Varma, J. Non-Cryst. Solids 279 (2001) 1. J.M. Stevels, Handbuch der physik 20 (1957) 372. H.E. Taylor, J. Soc. Glass Technol. 41 (1957) 350T. K.S. Rao, D. Madhava Prasad, P. Murali Krishna, B. Tilak, K.C. Varadarajulu, Mater. Sci. Eng. B 133 (2006) 141. [55] C. Kim, Y. Hwang, H.K. Kim, J.N. Kim, Phys. Chem. Glasses 44 (2) (2003) 166.