- Email: [email protected]
2)O3
Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
Contents lists available at ScienceDirect
Journal of Electron Spectroscopy and Related Phenomena journal homepage: www.elsevier.com/locate/elspec
Impedance spectroscopy, electronic structure and X-ray photoelectron spectroscopy studies of Pb(Fe1/2 Nb1/2 )O3 Chandrahas Bharti a , Alo Dutta b,∗ , Santiranjan Shannigrahi c , S.N. Choudhary a , R.K. Thapa d , T.P. Sinha b a
University Department of Physics, T. M. Bhagalpur University, Bhagalpur 812007, India Department of Physics, Bose Institute, 93/1, Acharya Prafulla Chandra Road, Kolkata 700009, India c Institute of Materials Research and Engineering (IMRE), 3 Research Link, Singapore 117602, Singapore d Department of Physics, Pachhunga University College, Mizoram University, Aiwzal 796001, India b
a r t i c l e
i n f o
Article history: Received 1 August 2008 Received in revised form 14 November 2008 Accepted 3 December 2008 Available online 9 December 2008 Keywords: Ferroelectric ceramics PFN relaxor X-ray photoemission spectroscopy Electronic structure
a b s t r a c t Impedance spectroscopy is used to study the electrical behaviour of lead iron niobate, Pb(Fe1/2 Nb1/2 )O3 (PFN) in the frequency range from 100 Hz to 1 MHz and in the temperature range from 203 to 363 K. The frequency-dependent electrical data are analyzed by impedance and conductivity formalisms. The complex impedance plane plot shows that the relaxation (conduction) mechanism in PFN is purely a bulk effect arising from the semiconductive grains. The relaxation mechanism of the sample in the framework of electric modulus formalism is modelled by Davidson–Cole equation. The scaling behaviour of imaginary electric modulus suggests that the relaxation describes the same mechanism at various temperatures. We have studied the electronic structure of the PFN using X-ray photoemission spectroscopy (XPS). The density of states (DOS) obtained from the first principles full potential linearized augmented plane wave calculation of PFN shows a direct energy gap ∼0.43 eV. The XPS spectrum is compared with the calculated DOS spectra. It has been observed that the electrical properties of PFN are dominated by the interaction between transition metal and oxygen ions as its valence band consists mainly of the oxygen p-states hybridized with the iron d-states. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Ferroelectric perovskites have been the subject of extensive studies due to their promising electrical characteristic which has a potential usefulness in fundamental research and technological applications. Investigation of the electrical properties of these materials is desirable to predict their suitability for electronic applications. Various relaxation processes seem to coexist in complex perovskite ceramics, which contain a number of different energy barriers due to point defects appearing during their fabrication. Therefore, the departure of the response from an ideal Debye model in ceramic samples, resulting from the interaction between dipoles, cannot be disregarded. A method of predicting the relaxation behaviour of a perovskite is through electric modulus theory. The electric modulus spectra, therefore, provide an opportunity to investigate conductivity and its associated relaxation in complex perovskite oxides having general formula A(B1/2 B1/2 )O3 .
∗ Corresponding author. Tel.: +91 33 23031189; fax: +91 33 23506790. E-mail address: alo [email protected] (A. Dutta). 0368-2048/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.elspec.2008.12.001
Pb(Fe1/2 Nb1/2 )O3 (PFN), is known to be the most intensively investigated 1:1 type relaxor ferroelectric system [1–7]. It is a multiferroic material where the electric and magnetic polarizations are coupled directly or indirectly not only on the macroscopic but apparently also on the local level [8–15]. However, the detail information about the relaxation phenomenon and conductivity process in PFN is still lacking. In this paper, we have tried to explain the relaxation mechanism and conductivity phenomenon of PFN. The formation of single perovskite phase of PFN has been the subject of debate [10,12,15,16]. Recently, Majumder et al. [12] have also discussed the inability of the formation of single perovskite phase of PFN by using sol–gel method. Although, many methods such as co-precipitation, molten-salt, sol–gel or hydrothermal, hotpressed, etc. have been introduced from time to time [17–21] to synthesize the single perovskite phase of PFN, the mixed oxide synthetic route which includes (a) two-stage calcinations, (b) using higher purity and finer precursor powders, (c) carrying out repeated calcinations and milling cycles and (d) employing longer heat treatment, reducing the firing temperature and adjusting the cooling rate, reported by Fu and Chen [22], Lejeune and Boilet [23] and Chiu and Desu [24] and later confirmed by Ananta and Thomas [16,21] and Lee et al. [17] has demonstrated the possibility of obtaining single phase PFN powder.
C. Bharti et al. / Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
The electronic structure of the perovskite oxides consisting of 4f and 3d orbitals has been studied widely [25–31]. The electronic structure of niobates attracts attention due to their intense ferroelectric and relaxor properties. The electrical properties of these oxides are dominated by the interaction between the transition metal cation and oxygen anion. Thus, to get proper information about the transport properties, the study of electronic structure of these materials is required. In the present work, we have performed the electronic structure of PFN by first principles calculation. The electrical properties of PFN are studied by the impedance spectroscopy. The X-ray photoemission spectra (XPS) of this material have been used to get the information about the spectral density function. The density of states (DOS) are obtained by full potential linearized augmented plane wave (FLAPW) method based on density functional theory (DFT) [32] under the generalized gradient approximation (GGA) [33] as implemented in Wien 2k code [34]. The experimental XPS results are compared with the calculated electronic structure. It is to be noted that the XPS is the method to obtain the information about valence states and environment of ions in a compound. In XPS, the excitation by low energies gives direct access to the band structure of the compounds. The ability of XPS method to distinguish the different oxidation states allows an efficient study of the change of binding state of atoms in chemical compounds and the change of surface chemistry in the materials. Most of such studies are concerned with the core levels. 2. Experiment The polycrystalline sample of PFN is synthesized by Coulombite precursor method using high purity oxides Fe2 O3 (∼99.99% pure, Alfa aesar) Nb2 O5 (∼99.9% pure, Loba chemie) and PbO (∼99% pure, Loba chemie) taken in stoichiometric ratio. First, the finely mixed powder of Fe2 O3 and Nb2 O5 is calcined at 1000 ◦ C for 6 h. The calcined powder then grinded and mixed with PbO for 6 h. At this stage of mixing 4 wt% of extra PbO is taken for compensating the lead loss during calcination and sintering. The finely mixed powder is calcined at 950 ◦ C for 6 h. The calcined powder is regrinded and used to make pellet of diameter 9.87 mm and thickness 1.65 mm using polyvinyl alcohol as binder. The pellet is sintered at 1000 ◦ C for 2 h and then brought to room temperature under controlled cooling. To measure the electrical properties, gold electrodes are formed on both surfaces of sintered disk. The frequency dependence of the capacitance and conductance is measured using an LCR meter in the temperature range from 203 to 363 K and in the frequency range from 100 Hz to 1 MHz. The complex electric modulus M* (=1/ε*) and the ac electrical conductivity (=ωεo ε ) were obtained from the temperature dependence of the real (ε ) and imaginary (ε ) components of the complex dielectric constant ε* (=ε − jε ). The X-ray photoemission spectrum of PFN is taken by X-ray photoemission spectroscopy (XPS) (VG ESCALAB 2201-XL Imaging System, England). XPS profiles of the samples were obtained using Al-K␣ source (1486.6 eV). The C 1s peak was used as the reference standard. 3. Results and discussion 3.1. Impedance spectroscopy Fig. 1 shows the complex plane impedance plots for PFN at the temperatures 313 and 343 K. The values of resistance R and capacitance C can be obtained by an equivalent circuit of one parallel resistance–capacitance (RC) element. This RC element gives rise to one semicircular arc on the complex plane plot, representing the
81
Fig. 1. Complex plane impedance plots at 313 and 343 K (solid line is the fitting to the data by the RC equivalent circuit at 343 K).
grain effect. The equivalent electrical equation for grain is Z ∗ = Z − jZ =
Z = Rg
1 Rg−1
+ jωCg
ωRg Cg
;
Z =
Rg 1 + (ωRg Cg )2
(1)
1 + (ωRg Cg )2
(2)
where Cg and Rg are the grain capacitance and grain resistance, respectively. We have fitted the experimental data using these expressions and the best fit of the data at 343 K is shown by solid line in Fig. 1 with Rg = 6.2 × 105 and Cg = 1.1 × 10−9 F. The angular frequency ω (=2) dependence of the real part (ε ) of dielectric constant, real part ( ac ) of ac conductivity and imaginary part (M ) of electric modulus of PFN are shown in Fig. 2 at various temperatures. The nature of the dielectric permittivity as shown in Fig. 2(a) can be explained by considering that the free dipoles oscillate in an alternating field. At very low frequencies (ω 1/, = relaxation time), dipoles follow the field and we have ε = εs (value of the dielectric constant at quasistatic field). With the increase of frequency (ω 1/), dipoles begin to lag behind the field and ε slightly decreases. When the applied frequency reaches the characteristic frequency of the material (i.e. ω = 1/), the dielectric constant drops and indicates the relaxation. At very high frequencies (ω 1/), dipoles can no longer follow the field and ε ≈ ε∞ (high frequency value of ε ). Qualitatively, this behaviour has been observed in Fig. 2(a). The frequency-dependent conductivity plots of PFN at various temperatures as shown in Fig. 2(b) possess a characteristic dispersion. The frequency spectra in Fig. 2(b) display the typical shape found for electronically conducting system. The conductivity spectra have the tendency to merge at higher frequencies with the increase in temperature. At a particular temperature, the conductivity decreases with decreasing frequency and becomes independent of frequency after certain value. The extrapolation of this part towards lower frequency gives dc which is attributed to the long range translational motion of the charge carriers. The basic fact about ac conductivity ( ac ) in PFN is that ac is an increasing function of frequency (any hopping model has this feature) and thus the real part of conductivity spectra can be explained by the power law defined as [35–37]:
ac = dc 1 +
ω n ωH
(3)
82
C. Bharti et al. / Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
Therefore, the modulus represents the real dielectric relaxation process. Data presented in this way exhibit a pronounced relaxation peak for imaginary part M (ω) of electric modulus that moves towards lower frequencies during cooling of the sample as shown in Fig. 2(c). Consequently, it means that the relaxation rate for this process decreases with decreasing temperature. The frequency region ) determines the range in which charge below peak maximum (Mm carriers are mobile on long distances. At frequency above peak maximum the carriers are confined to potential wells, being mobile on short distances. As a convenient measure of the characteristic relaxation time one can choose the inverse of frequency of the maximum peak position, i.e., m = ωm −1 . Thus we can determine the temperature dependence of the characteristic relaxation time as shown in the inset of Fig. 2(b) which satisfies the Arrhenius law. From the numerical fitting analysis, we have obtained the value of activation energy = 0.37 eV. The dielectric relaxation in PFN can be modelled by Cole–Cole or Davidson–Cole equation. The Davidson–Cole equation accounts for an asymmetric distribution of relaxation times resulting from a dielectric dispersion within a system by introducing an exponential parameter in the Debye dielectric function [39]. We have fitted the experimental data with Davidson–Cole equation, defined as [40]: M =
M =
M∞ Ms [Ms + (M∞ − Ms )(cos ) cos ] Ms2
+ (M∞ − Ms )(cos ) [2Ms cos + (M∞ − Ms )(cos ) ] (4)
M∞ Ms [(M∞ − Ms )(cos ) sin
Ms2
+ (M∞ − Ms )(cos ) [2Ms cos + (M∞ − Ms )(cos ) ] (5) /2 = tan( +1 ).
Fig. 2. Frequency dependence of (a) real (ε ) part of dielectric constant, (b) ac conductivity ( ac ) and (c) imaginary (M ) part of electric modulus at various temperatures for Pb(Fe1/2 Nb1/2 )O3 . The Arrhenius plots of dc conductivity ( dc ) and most probable relaxation frequency (ωm ) corresponding to M are shown in the inset of (b). The fitting of M using expression (5) is shown by solid lines in (c). The scaling behaviour of M is shown in the inset of (c).
where dc is the dc conductivity, ωH is the hopping frequency of the charge carriers and n is the dimensionless frequency exponent. The experimental conductivity spectra of PFN are fitted to Eq. (3) with dc and ωH as variable keeping in the mind that the value of parameter n is weakly temperature dependent. The best fit of the conductivity spectrum at 303 K is shown in Fig. 2(b) by solid line with dc (=9.37 × 10−6 ) and ωH (=11061 Hz) and n (=0.78). The values of dc obtained from the fitting of the experimental data at various temperatures using Eq. (3) follow the Arrhenius law with an activation energy 0.44 eV as shown in the inset of Fig. 2(b). Such a value of activation energy indicates that the conduction mechanism for PFN may be due to the polaron hopping based on the electron carriers. In the hopping process, the electron disorders its surroundings by moving its neighbouring sites in the system. The relaxation property of PFN has been demonstrated using modulus formalism [38]. From the physical point of view, the electrical modulus corresponds to the relaxation of the electric field in the material when the electric displacement remains constant.
where 0 < ≤ 1; tan = ω; ωm A good agreement between the directly measured data and the data obtained by using Eq. (5) is observed as shown by solid line in Fig. 2(c) for the temperatures 323 K with the value of = 0.56. and each frequency by ω (ω We have scaled each M by Mm m m corresponds to the frequency of the peak position of M in the M vs. log ω plots) in the inset of Fig. 2(c). The perfect overlap of the curves for all the temperatures into a single master curve indicates that the relaxation describes the same mechanism at various temperatures. A similar nature of the scaling behaviour of M was observed in Sr(Fe1/2 Nb1/2 )O3 and Ba(Al1/2 Nb1/2 )O3 [41,42]. 3.2. Electronic structure and X-ray photoemission spectroscopy The electronic band structure and the density of states (DOS) of PFN are calculated by considering an ordered double perovskite having Fm3m structure with lattice parameter 8.2 Å. No shape approximation related to either potential or the charge density in the interstitial is taken. Due to a heavy ion Pb, relativistic effect has been considered in the calculation of DOS. The product of muffin tin radius R and Kmax (maximum reciprocal space K for the plane-wave expansion in the interstitial) governs the convergence of basis set. The basis set is converged by increasing the value of this product as well as the maximum angular momentum. In our calculation, RKmax is taken to be 8 and 47 number of k points in the irreducible wedge of the Brillouin zone has been used. The iteration process has been repeated until the calculated total energy is converged to less than (0.1) mRy. Fig. 3 shows the angular momentum decomposed total DOS and partial DOS of Pb-f, Fe-d, Nb-d and O-p states for PFN. A direct energy gap (∼0.43 eV) is obtained near the Fermi level (EF ) set at 0 eV. In the valence band, the band extending from −6.5 eV to −1.56 eV is mainly composed of O-p state and Nb-3d state with a small contribution of Fe-d state and Pb-f state. The 3d-state of Fe
C. Bharti et al. / Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
83
Fig. 4. X-ray photo-emission spectrum of Pb(Fe1/2 Nb1/2 )O3 .
reduction of bandwidth is due to the muffin-tin approximation as observed by us in the earlier studies for pure perovskites [43,44]. From the above analysis, we may conclude that the electrical properties of PFN are dominated by the interaction between transition metal and oxygen ions as its valence band consists mainly of the oxygen p-states hybridized with the iron d-states. The eg orbitals of Fe cation overlap with the nearby 2p orbital from the split O-2p to form -bonds (eg –p –eg bond) and the t2g orbitals overlap with 2p of O-2p states to form weaker -bonds
Fig. 3. Total density of states of Pb(Fe1/2 Nb1/2 )O3 along with the partial-density of states of Pb-f, Fe-d, Nb-d, and O-p.
and 4d-state of Nb are split up into t2g and eg states due to the presence of crystal field produced by the oxygen octahedra, with the t2g states having lower energy and the eg states at higher energy. The t2g and eg states are the eigen states to the Hamiltonian and do not hybridize with each other. In the scalar relativistic DOS, the three fold degenerate Fe-t2g states are filled. Therefore, the Fermi level ends up in the crystal-field gap between the Fe-t2g and Fe-eg states. In order to verify the electronic structure calculations experimentally, we have performed the XPS study of PFN in a wide energy range. The XPS spectrum of PFN is shown in Fig. 4, where the profiles of the spectrum are identified and indexed. Except carbon, no other contamination is found in the spectrum. Carbon contamination may arise due to long time exposure on atmosphere and/or trace of the starting materials. The total DOS is convoluted with a Lorentzian of 0.4 eV full width at half maximum and the calculated spectrum is compared with the experimental XPS valence band spectrum as shown in Fig. 5. The calculated electronic structure of PFN is qualitatively similar to that of the valence band XPS spectrum in terms of spectral features, energy positions and relative intensities. To get more insight into the valence band spectrum, we have also shown the convoluted angular momentum and site decomposed partial DOS of Fe-d and O-p in Fig. 5. It is clear from Fig. 5 that the peaks at 7.5 eV, 10.7 eV and a hump at 5.28 eV arise due to the strong hybridization between O-p states and Fe-d states. It seems in Fig. 5 that the calculated DOS exhibits sharper peaks than the experimental spectra because we have not included the lifetime broadening in our DOS calculations. The experimental bandwidth is larger than the calculated one. This
Fig. 5. The X-ray photo-emission spectrum in the valence band region is compared with the calculated DOS spectra of Pb(Fe1/2 Nb1/2 )O3 .
84
C. Bharti et al. / Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
(t2g –p –t2g ) and may be responsible for the electrical conduction in PFN. It is to be mentioned that if one considers the energy band diagram refers to the polycrystalline system such as PFN (where grain boundary is depicted having two successive grains) the donor states and the trapping states are the potential states available below the conduction band. The trapping states are presumed to have a range in the band gap below the conduction band. Upon the application of the ac small signal voltage these states will sweep through the amplitude of the small signal. These sweeping states are unlikely to be singular in nature for the defect states in polycrystalline materials. Therefore, the time constant comprises the charge mobility represented by the conductance and the charge storage represented by the capacitance for the defect states. This gives rise to the trapping contribution in the series event for the single charge that is being trapped and subsequently released to become mobile carrier. This event essentially gives the concept of R–C combination having relaxation time as a distributed parameter which in turn can be extracted from the complex plane plot. It should be noted that the activation energy ∼0.44 eV as obtained from the Arrhenius plot of dc is nearly equal to the activation energy ∼0.37 eV obtained from Arrhenius plot of M . These experimental results imply that the nature of charge carriers responsible for dielectric relaxation peaks and dc conduction belongs to the same category, which indicates that the polarization relaxation has a close relation with the conductivity in grain interior, and the polarization process probably depends on the conducting of the charge in the grain interior, which result that the two activation energies obtained are almost same. We have also compared the core level XPS spectra of Pb-4f, Fe-2p and Nb-3d with the corresponding convoluted partial-DOS data as shown in Fig. 6. The high resolution XPS spectrum of Pb4f is shown in Fig. 6(a). Only one spin–orbit doublet is observed. The first component at around 140.3 eV is attributed to the Pb4f7/2 peak, whereas the second at around 145 eV is attributed to the Pb-4f5/2 peak. The difference between Pb-4f7/2 and Pb-4f5/2 giving spin–orbit splitting is approximately 4.7 eV for the PFN compound and is comparable to that of the other lead-based relaxor compounds [45,46]. Fig. 6(b) shows the high resolution XPS spectrum of Fe-2p. The Fe-2p spectrum is split up into two components, 2p3/2 (714 eV) and 2p1/2 (725.6 eV) due to the spin–orbit effect. The spin orbit splitting of Fe-2p level is about 12 eV. The 2p region of compounds having electronic configuration d5 has been extensively studied both theoretically [47] and experimentally [48,49] and resembles with Fe-2p spectrum of PFN. The Fe-2p3/2 peak of PFN shows a distinct satellite around 718 eV. This satellite is caused by a dynamic charge-transfer during the photo-emission process, and appears in the XPS spectra of most transition metal compounds. The effective suppression of Fe2+ ions in PFN is confirmed by the XPS spectrum which suggests a negligible concentration of oxygen vacancies as shown in Fig. 6(b). As expected, peak appears around 714 eV and a satellite appears at 718 eV, indicating the existence of Fe3+ , while absence of peaks at 709.5 and 716 eV indicates an absence of Fe2+ [50]. The high-resolution XPS spectrum of Nb-3d is shown in Fig. 6(c). It is clearly seen that the Nb-3d spectrum consists of two peaks corresponding to its angular momentum of electrons. Only one spin–orbit doublet is observed for Nb-3d5/2 (208.8 eV) and Nb-3d3/2 (211.5 eV). It means that Nb ion exhibits only one chemical state characteristic for PFN. The energy positions of both peaks coincide well with previous reports on different niobium-based perovskite compounds [51]. Thus the XPS investigation allows us to determine the relative concentrations of the various constituents of PFN. The experimental core level XPS spectra are compared with the calculated corresponding spectra (dashed lines) in Fig. 6(a)–(c) for Pb-4f, Fe-2p and Nb-3d, respectively. All the calculated spectra (dashed
Fig. 6. The X-ray photo-emission spectra of the Pb-4f state (a), Fe-2p state (b) and Nb-3d state (c) are compared with the corresponding convoluted partial-DOS spectra of Pb(Fe1/2 Nb1/2 )O3 .
lines) appear to be very similar with the peak positions and relative intensities of XPS spectra thus implying a good agreement between the experimental results and our theoretical calculations. 4. Conclusions The frequency-dependent dielectric dispersion of polycrystalline PFN synthesized by the Coulombite technique is investigated in the frequency range from 100 Hz to 1 MHz and in the temperature range from 203 to 363 K. The frequency-dependent electrical data are analyzed in the framework of conductivity and impedance formalisms. The complex plane plot suggests that the relaxation mechanism in PFN is purely a bulk effect arising from the semiconductive grains. Modulus spectroscopic data are used to gain an insight into the electrical properties of the samples and fitted with the Davidson–Cole model. The scaling behaviour of imagi-
C. Bharti et al. / Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
nary electric modulus suggests that the relaxation shows the same mechanism at various temperatures. We have studied the electronic structure of the PFN using X-ray photoemission spectroscopy (XPS). The DOS of PFN as obtained from the first principles full potential linearized augmented plane wave calculation shows a direct energy gap ∼0.43 eV. The d-states of transition metals (Fe and Nb) split into t2g and eg states due to the presence of crystal field produced by oxygen octahedra. The valence band XPS spectrum is compared with the calculated DOS spectra and exhibits a strong hybridization between the Fe-d-states and O-p-states. It has been observed that the electrical properties of PFN are dominated by the interaction between transition metal and oxygen ions as its valence band consists mainly of the oxygen p-states hybridized with the iron d-states. Acknowledgement One of the authors (Alo Dutta) acknowledges the financial support provided by the CSIR, New Delhi in the form of SRF. References [1] G.A. Smolensky, A.I. Agranovskaya, S.N. Popov, V.A. Isupov, Sov. Phys. -Tech. Phys 3 (1958) 1981. [2] N. Yasuda, Y. Ueda, J. Phys.: Condens. Matter 1 (1989) 497; N. Yasuda, Y. Ueda, J. Phys.: Condens. Matter 1 (1989) 5179. [3] C.N.W. Darlington, J. Phys.: Condens. Matter 3 (1991) 4173. [4] N. Lampis, P. Sciau, A.G. Lehmann, J. Phys.: Condens. Matter 11 (1999) 3489. [5] S.A. Ivanov, R. Tellgren, H. Rundlof, N.W. Thomas, S. Ananta, J. Phys.: Condens. Matter 12 (2000) 2393. [6] V. Bonny, M. Bonin, P. Sciau, K.J. Schenk, G. Chapuis, Solid State Commun. 102 (1997) 347. [7] H. Schmid, Ferroelectrics 162 (1994) 317. [8] Y. Yang, J.M. Liu, H.B. Huang, W.Q. Zou, P. Bao, Z.G. Liu, Phys. Rev. B 70 (2004) 132101. [9] J.T. Wang, C. Zhang, Z.X. Shen, Y. Feng, Ceram. Int. 30 (2004) 1627. [10] V.V. Bhat, A.M. Umarji, V.B. Shenoy, U.V. Waghmare, Phys. Rev. B 72 (2005) 14104. [11] Y. Yang, S.T. Zhang, H.B. Huang, Y.F. Chen, Z.G. Liu, J.M. Liu, Mater. Lett. 59 (2005) 1767. [12] S.B. Majumder, S. Bhattacharyya, R.S. Katiyar, J. Appl. Phys. 99 (2006) 024108. [13] O. Raymond, R. Font, R. Machorro, E. Martinez, J.M. Siqueiros, MRS Proc. 1034E (2006) K3.24. [14] G. Alvarez, R. Font, J. Portelles, R. Zamarano, R. Valenzuela, J. Phys. Chem. Solids 68 (2007) 1436.
85
[15] R. Blinc, P. Cevc, A. Zorko, J. Holc, M. Kosec, Z. Trontelj, J. Pirnat, N. Dalal, V. Ramchandran, J. Krzystek, J. Appl. Phys. 101 (2007) 033901. [16] S. Ananta, N.W. Thomas, J. Eur. Ceram. Soc. 19 (1999) 2917. [17] S.B. Lee, S.H. Yoon, H. Kim, J. Euro. Ceram. Soc. 24 (2004) 2465. [18] C.C. Chiu, S.B. Desu, K. Amanuma, Mater. Sci. Eng. B 22 (1994) 133. [19] M. Yokosuka, Jpn. J. Appl. Phys. 32 (1993) 1142. [20] A.A. Bokov, L.A. Shpak, I.P. Rayevsky, J. Phys. Chem. Solids 54 (1993) 495. [21] S. Ananta, N.W. Thomas, J. Euro. Ceram. Soc. 19 (1999) 155. [22] S.L. Fu, G.F. Chen, Ferroelectrics 82 (1988) 119. [23] M. Lejeune, J.P. Boilot, Ceram. Int. 8 (1982) 99. [24] C.C. Chiu, S.B. Desu, Mater. Sci. Eng. B 21 (1993) 26. [25] M. Abbate, H. Ascolani, F. Prado, A. Caneiro, Solid State Commun. 129 (2004) 113. [26] V.R. Mastelaro, P.N. Lisboa-Filho, P.P. Neves, W.H. Schreiner, P.A.P. Nascente, J.A. Eiras, J. Electron, Spectrosc. Rel. Phenom. 156 (2007) 476. [27] S. Shannigrahi, K. Yao, J. Appl. Phys. 98 (2005) 034104. [28] V.A. Gasparov, S.N. Ermolov, G.K. Strukova, N.S. Sidorov, S.S. Khassanov, H.-S. Wang, M. Schneider, E. Glaser, W. Richter, Phys. Rev. B 63 (2001) 174512. [29] Alo Dutta, T.P. Sinha, S. Shannigrahi, Phys. Rev. B 76 (2007) 155113. [30] Alo Dutta, T.P. Sinha, S. Shannigrahi, J. Appl. Phys. 104 (2008) 064114. [31] Alo Dutta, T.P. Sinha, B. Pahari, R. Sarkar, K. Ghoshray, S. Shannigrahi, J. Phys.: Condens. Matter 20 (2008) 445206. [32] P. Hohenberg, W. Kohn, Phys. Rev. 136 (1964) B864. [33] J.P. Perdew, J.A. Chevary, S.H. Vosko, K.A. Jackson, M.R. Pederson, D.J. Singh, C. Fiolhais, Phys. Rev. B 46 (1992) 6671. [34] P. Blaha, K. Schwarz, G. Madsen, D. Kvasnicka, J. Luitz, WIEN2k, an augmented plane wave + local orbitals program for calculating crystal properties, Karlheinz Schwarz, Techn. Universität Wien, Austria, 2001. [35] A.K. Jonscher, J. Phys. D: Appl. Phys. 13 (1980) L89. [36] A.K. Jonscher, Dielectric Relaxation in Solids, Chelsea Dielectrics Press, London, 1983. [37] A.K. Jonscher, Universal Relaxation Law, Chelsea Dielectrics Press, London, 1996. [38] J. Liu, Ch.-G. Duan, W.-G. Yin, W.N. Mei, R.W. Smith, J. Chem. Phys. 119 (2003) 2812. [39] P. Debye, Polar Molecules, Chemical Catalogue Company, New York, 1929. [40] D.W. Davidson, R.H. Cole, J. Chem. Phys. 19 (1951) 1484. [41] S. Saha, T.P. Sinha, J. Appl. Phys. 99 (2006) 014109. [42] Alo Dutta, T.P. Sinha, J. Phys. Chem. Solids 67 (2006) 1484. [43] S. Saha, T.P. Sinha, A. Mookerjee, Phys. Rev. B 62 (2000) 8828. [44] S. Saha, T.P. Sinha, A. Mookerjee, J. Phys.: Condens. Matter 12 (2000) 3325. [45] V.R. Mastelaro, P.N. Lisboa-Filho, P.P. Neves, W.H. Schreiner, P.A.P. Nascente, J.A. Eiras, J. Electron. Spectrosc. Rel. Phenom. 156–158 (2007) 476. [46] F. Parmaigiani, L. Rollandi, G. Samoggia, L.E. Depero, Solid. State Commun. 100 (1996) 801. [47] R.P. Gupta, S.K. Sen, Phys. Rev. B 10 (1974) 71. [48] S.P. Kowalczyk, L. Ley, F.R. McFeely, D.A. Shirley, Phys. Rev. B 11 (1975) 1721. [49] J.A. Tossell, J. Electron. Spectrosc. 10 (1977) 169. [50] W. Eerenstein, F.D. Morrison, J. Dho, M.G. Blamire, J.F. Scott, N.D. Mathur, Science 307 (2005) 1203a. [51] M. Kruczek, E. Talik, A. Kania, Solid State Commun. 137 (2006) 469.
Contents lists available at ScienceDirect
Journal of Electron Spectroscopy and Related Phenomena journal homepage: www.elsevier.com/locate/elspec
Impedance spectroscopy, electronic structure and X-ray photoelectron spectroscopy studies of Pb(Fe1/2 Nb1/2 )O3 Chandrahas Bharti a , Alo Dutta b,∗ , Santiranjan Shannigrahi c , S.N. Choudhary a , R.K. Thapa d , T.P. Sinha b a
University Department of Physics, T. M. Bhagalpur University, Bhagalpur 812007, India Department of Physics, Bose Institute, 93/1, Acharya Prafulla Chandra Road, Kolkata 700009, India c Institute of Materials Research and Engineering (IMRE), 3 Research Link, Singapore 117602, Singapore d Department of Physics, Pachhunga University College, Mizoram University, Aiwzal 796001, India b
a r t i c l e
i n f o
Article history: Received 1 August 2008 Received in revised form 14 November 2008 Accepted 3 December 2008 Available online 9 December 2008 Keywords: Ferroelectric ceramics PFN relaxor X-ray photoemission spectroscopy Electronic structure
a b s t r a c t Impedance spectroscopy is used to study the electrical behaviour of lead iron niobate, Pb(Fe1/2 Nb1/2 )O3 (PFN) in the frequency range from 100 Hz to 1 MHz and in the temperature range from 203 to 363 K. The frequency-dependent electrical data are analyzed by impedance and conductivity formalisms. The complex impedance plane plot shows that the relaxation (conduction) mechanism in PFN is purely a bulk effect arising from the semiconductive grains. The relaxation mechanism of the sample in the framework of electric modulus formalism is modelled by Davidson–Cole equation. The scaling behaviour of imaginary electric modulus suggests that the relaxation describes the same mechanism at various temperatures. We have studied the electronic structure of the PFN using X-ray photoemission spectroscopy (XPS). The density of states (DOS) obtained from the first principles full potential linearized augmented plane wave calculation of PFN shows a direct energy gap ∼0.43 eV. The XPS spectrum is compared with the calculated DOS spectra. It has been observed that the electrical properties of PFN are dominated by the interaction between transition metal and oxygen ions as its valence band consists mainly of the oxygen p-states hybridized with the iron d-states. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Ferroelectric perovskites have been the subject of extensive studies due to their promising electrical characteristic which has a potential usefulness in fundamental research and technological applications. Investigation of the electrical properties of these materials is desirable to predict their suitability for electronic applications. Various relaxation processes seem to coexist in complex perovskite ceramics, which contain a number of different energy barriers due to point defects appearing during their fabrication. Therefore, the departure of the response from an ideal Debye model in ceramic samples, resulting from the interaction between dipoles, cannot be disregarded. A method of predicting the relaxation behaviour of a perovskite is through electric modulus theory. The electric modulus spectra, therefore, provide an opportunity to investigate conductivity and its associated relaxation in complex perovskite oxides having general formula A(B1/2 B1/2 )O3 .
∗ Corresponding author. Tel.: +91 33 23031189; fax: +91 33 23506790. E-mail address: alo [email protected] (A. Dutta). 0368-2048/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.elspec.2008.12.001
Pb(Fe1/2 Nb1/2 )O3 (PFN), is known to be the most intensively investigated 1:1 type relaxor ferroelectric system [1–7]. It is a multiferroic material where the electric and magnetic polarizations are coupled directly or indirectly not only on the macroscopic but apparently also on the local level [8–15]. However, the detail information about the relaxation phenomenon and conductivity process in PFN is still lacking. In this paper, we have tried to explain the relaxation mechanism and conductivity phenomenon of PFN. The formation of single perovskite phase of PFN has been the subject of debate [10,12,15,16]. Recently, Majumder et al. [12] have also discussed the inability of the formation of single perovskite phase of PFN by using sol–gel method. Although, many methods such as co-precipitation, molten-salt, sol–gel or hydrothermal, hotpressed, etc. have been introduced from time to time [17–21] to synthesize the single perovskite phase of PFN, the mixed oxide synthetic route which includes (a) two-stage calcinations, (b) using higher purity and finer precursor powders, (c) carrying out repeated calcinations and milling cycles and (d) employing longer heat treatment, reducing the firing temperature and adjusting the cooling rate, reported by Fu and Chen [22], Lejeune and Boilet [23] and Chiu and Desu [24] and later confirmed by Ananta and Thomas [16,21] and Lee et al. [17] has demonstrated the possibility of obtaining single phase PFN powder.
C. Bharti et al. / Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
The electronic structure of the perovskite oxides consisting of 4f and 3d orbitals has been studied widely [25–31]. The electronic structure of niobates attracts attention due to their intense ferroelectric and relaxor properties. The electrical properties of these oxides are dominated by the interaction between the transition metal cation and oxygen anion. Thus, to get proper information about the transport properties, the study of electronic structure of these materials is required. In the present work, we have performed the electronic structure of PFN by first principles calculation. The electrical properties of PFN are studied by the impedance spectroscopy. The X-ray photoemission spectra (XPS) of this material have been used to get the information about the spectral density function. The density of states (DOS) are obtained by full potential linearized augmented plane wave (FLAPW) method based on density functional theory (DFT) [32] under the generalized gradient approximation (GGA) [33] as implemented in Wien 2k code [34]. The experimental XPS results are compared with the calculated electronic structure. It is to be noted that the XPS is the method to obtain the information about valence states and environment of ions in a compound. In XPS, the excitation by low energies gives direct access to the band structure of the compounds. The ability of XPS method to distinguish the different oxidation states allows an efficient study of the change of binding state of atoms in chemical compounds and the change of surface chemistry in the materials. Most of such studies are concerned with the core levels. 2. Experiment The polycrystalline sample of PFN is synthesized by Coulombite precursor method using high purity oxides Fe2 O3 (∼99.99% pure, Alfa aesar) Nb2 O5 (∼99.9% pure, Loba chemie) and PbO (∼99% pure, Loba chemie) taken in stoichiometric ratio. First, the finely mixed powder of Fe2 O3 and Nb2 O5 is calcined at 1000 ◦ C for 6 h. The calcined powder then grinded and mixed with PbO for 6 h. At this stage of mixing 4 wt% of extra PbO is taken for compensating the lead loss during calcination and sintering. The finely mixed powder is calcined at 950 ◦ C for 6 h. The calcined powder is regrinded and used to make pellet of diameter 9.87 mm and thickness 1.65 mm using polyvinyl alcohol as binder. The pellet is sintered at 1000 ◦ C for 2 h and then brought to room temperature under controlled cooling. To measure the electrical properties, gold electrodes are formed on both surfaces of sintered disk. The frequency dependence of the capacitance and conductance is measured using an LCR meter in the temperature range from 203 to 363 K and in the frequency range from 100 Hz to 1 MHz. The complex electric modulus M* (=1/ε*) and the ac electrical conductivity (=ωεo ε ) were obtained from the temperature dependence of the real (ε ) and imaginary (ε ) components of the complex dielectric constant ε* (=ε − jε ). The X-ray photoemission spectrum of PFN is taken by X-ray photoemission spectroscopy (XPS) (VG ESCALAB 2201-XL Imaging System, England). XPS profiles of the samples were obtained using Al-K␣ source (1486.6 eV). The C 1s peak was used as the reference standard. 3. Results and discussion 3.1. Impedance spectroscopy Fig. 1 shows the complex plane impedance plots for PFN at the temperatures 313 and 343 K. The values of resistance R and capacitance C can be obtained by an equivalent circuit of one parallel resistance–capacitance (RC) element. This RC element gives rise to one semicircular arc on the complex plane plot, representing the
81
Fig. 1. Complex plane impedance plots at 313 and 343 K (solid line is the fitting to the data by the RC equivalent circuit at 343 K).
grain effect. The equivalent electrical equation for grain is Z ∗ = Z − jZ =
Z = Rg
1 Rg−1
+ jωCg
ωRg Cg
;
Z =
Rg 1 + (ωRg Cg )2
(1)
1 + (ωRg Cg )2
(2)
where Cg and Rg are the grain capacitance and grain resistance, respectively. We have fitted the experimental data using these expressions and the best fit of the data at 343 K is shown by solid line in Fig. 1 with Rg = 6.2 × 105 and Cg = 1.1 × 10−9 F. The angular frequency ω (=2) dependence of the real part (ε ) of dielectric constant, real part ( ac ) of ac conductivity and imaginary part (M ) of electric modulus of PFN are shown in Fig. 2 at various temperatures. The nature of the dielectric permittivity as shown in Fig. 2(a) can be explained by considering that the free dipoles oscillate in an alternating field. At very low frequencies (ω 1/, = relaxation time), dipoles follow the field and we have ε = εs (value of the dielectric constant at quasistatic field). With the increase of frequency (ω 1/), dipoles begin to lag behind the field and ε slightly decreases. When the applied frequency reaches the characteristic frequency of the material (i.e. ω = 1/), the dielectric constant drops and indicates the relaxation. At very high frequencies (ω 1/), dipoles can no longer follow the field and ε ≈ ε∞ (high frequency value of ε ). Qualitatively, this behaviour has been observed in Fig. 2(a). The frequency-dependent conductivity plots of PFN at various temperatures as shown in Fig. 2(b) possess a characteristic dispersion. The frequency spectra in Fig. 2(b) display the typical shape found for electronically conducting system. The conductivity spectra have the tendency to merge at higher frequencies with the increase in temperature. At a particular temperature, the conductivity decreases with decreasing frequency and becomes independent of frequency after certain value. The extrapolation of this part towards lower frequency gives dc which is attributed to the long range translational motion of the charge carriers. The basic fact about ac conductivity ( ac ) in PFN is that ac is an increasing function of frequency (any hopping model has this feature) and thus the real part of conductivity spectra can be explained by the power law defined as [35–37]:
ac = dc 1 +
ω n ωH
(3)
82
C. Bharti et al. / Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
Therefore, the modulus represents the real dielectric relaxation process. Data presented in this way exhibit a pronounced relaxation peak for imaginary part M (ω) of electric modulus that moves towards lower frequencies during cooling of the sample as shown in Fig. 2(c). Consequently, it means that the relaxation rate for this process decreases with decreasing temperature. The frequency region ) determines the range in which charge below peak maximum (Mm carriers are mobile on long distances. At frequency above peak maximum the carriers are confined to potential wells, being mobile on short distances. As a convenient measure of the characteristic relaxation time one can choose the inverse of frequency of the maximum peak position, i.e., m = ωm −1 . Thus we can determine the temperature dependence of the characteristic relaxation time as shown in the inset of Fig. 2(b) which satisfies the Arrhenius law. From the numerical fitting analysis, we have obtained the value of activation energy = 0.37 eV. The dielectric relaxation in PFN can be modelled by Cole–Cole or Davidson–Cole equation. The Davidson–Cole equation accounts for an asymmetric distribution of relaxation times resulting from a dielectric dispersion within a system by introducing an exponential parameter in the Debye dielectric function [39]. We have fitted the experimental data with Davidson–Cole equation, defined as [40]: M =
M =
M∞ Ms [Ms + (M∞ − Ms )(cos ) cos ] Ms2
+ (M∞ − Ms )(cos ) [2Ms cos + (M∞ − Ms )(cos ) ] (4)
M∞ Ms [(M∞ − Ms )(cos ) sin
Ms2
+ (M∞ − Ms )(cos ) [2Ms cos + (M∞ − Ms )(cos ) ] (5) /2 = tan( +1 ).
Fig. 2. Frequency dependence of (a) real (ε ) part of dielectric constant, (b) ac conductivity ( ac ) and (c) imaginary (M ) part of electric modulus at various temperatures for Pb(Fe1/2 Nb1/2 )O3 . The Arrhenius plots of dc conductivity ( dc ) and most probable relaxation frequency (ωm ) corresponding to M are shown in the inset of (b). The fitting of M using expression (5) is shown by solid lines in (c). The scaling behaviour of M is shown in the inset of (c).
where dc is the dc conductivity, ωH is the hopping frequency of the charge carriers and n is the dimensionless frequency exponent. The experimental conductivity spectra of PFN are fitted to Eq. (3) with dc and ωH as variable keeping in the mind that the value of parameter n is weakly temperature dependent. The best fit of the conductivity spectrum at 303 K is shown in Fig. 2(b) by solid line with dc (=9.37 × 10−6 ) and ωH (=11061 Hz) and n (=0.78). The values of dc obtained from the fitting of the experimental data at various temperatures using Eq. (3) follow the Arrhenius law with an activation energy 0.44 eV as shown in the inset of Fig. 2(b). Such a value of activation energy indicates that the conduction mechanism for PFN may be due to the polaron hopping based on the electron carriers. In the hopping process, the electron disorders its surroundings by moving its neighbouring sites in the system. The relaxation property of PFN has been demonstrated using modulus formalism [38]. From the physical point of view, the electrical modulus corresponds to the relaxation of the electric field in the material when the electric displacement remains constant.
where 0 < ≤ 1; tan = ω; ωm A good agreement between the directly measured data and the data obtained by using Eq. (5) is observed as shown by solid line in Fig. 2(c) for the temperatures 323 K with the value of = 0.56. and each frequency by ω (ω We have scaled each M by Mm m m corresponds to the frequency of the peak position of M in the M vs. log ω plots) in the inset of Fig. 2(c). The perfect overlap of the curves for all the temperatures into a single master curve indicates that the relaxation describes the same mechanism at various temperatures. A similar nature of the scaling behaviour of M was observed in Sr(Fe1/2 Nb1/2 )O3 and Ba(Al1/2 Nb1/2 )O3 [41,42]. 3.2. Electronic structure and X-ray photoemission spectroscopy The electronic band structure and the density of states (DOS) of PFN are calculated by considering an ordered double perovskite having Fm3m structure with lattice parameter 8.2 Å. No shape approximation related to either potential or the charge density in the interstitial is taken. Due to a heavy ion Pb, relativistic effect has been considered in the calculation of DOS. The product of muffin tin radius R and Kmax (maximum reciprocal space K for the plane-wave expansion in the interstitial) governs the convergence of basis set. The basis set is converged by increasing the value of this product as well as the maximum angular momentum. In our calculation, RKmax is taken to be 8 and 47 number of k points in the irreducible wedge of the Brillouin zone has been used. The iteration process has been repeated until the calculated total energy is converged to less than (0.1) mRy. Fig. 3 shows the angular momentum decomposed total DOS and partial DOS of Pb-f, Fe-d, Nb-d and O-p states for PFN. A direct energy gap (∼0.43 eV) is obtained near the Fermi level (EF ) set at 0 eV. In the valence band, the band extending from −6.5 eV to −1.56 eV is mainly composed of O-p state and Nb-3d state with a small contribution of Fe-d state and Pb-f state. The 3d-state of Fe
C. Bharti et al. / Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
83
Fig. 4. X-ray photo-emission spectrum of Pb(Fe1/2 Nb1/2 )O3 .
reduction of bandwidth is due to the muffin-tin approximation as observed by us in the earlier studies for pure perovskites [43,44]. From the above analysis, we may conclude that the electrical properties of PFN are dominated by the interaction between transition metal and oxygen ions as its valence band consists mainly of the oxygen p-states hybridized with the iron d-states. The eg orbitals of Fe cation overlap with the nearby 2p orbital from the split O-2p to form -bonds (eg –p –eg bond) and the t2g orbitals overlap with 2p of O-2p states to form weaker -bonds
Fig. 3. Total density of states of Pb(Fe1/2 Nb1/2 )O3 along with the partial-density of states of Pb-f, Fe-d, Nb-d, and O-p.
and 4d-state of Nb are split up into t2g and eg states due to the presence of crystal field produced by the oxygen octahedra, with the t2g states having lower energy and the eg states at higher energy. The t2g and eg states are the eigen states to the Hamiltonian and do not hybridize with each other. In the scalar relativistic DOS, the three fold degenerate Fe-t2g states are filled. Therefore, the Fermi level ends up in the crystal-field gap between the Fe-t2g and Fe-eg states. In order to verify the electronic structure calculations experimentally, we have performed the XPS study of PFN in a wide energy range. The XPS spectrum of PFN is shown in Fig. 4, where the profiles of the spectrum are identified and indexed. Except carbon, no other contamination is found in the spectrum. Carbon contamination may arise due to long time exposure on atmosphere and/or trace of the starting materials. The total DOS is convoluted with a Lorentzian of 0.4 eV full width at half maximum and the calculated spectrum is compared with the experimental XPS valence band spectrum as shown in Fig. 5. The calculated electronic structure of PFN is qualitatively similar to that of the valence band XPS spectrum in terms of spectral features, energy positions and relative intensities. To get more insight into the valence band spectrum, we have also shown the convoluted angular momentum and site decomposed partial DOS of Fe-d and O-p in Fig. 5. It is clear from Fig. 5 that the peaks at 7.5 eV, 10.7 eV and a hump at 5.28 eV arise due to the strong hybridization between O-p states and Fe-d states. It seems in Fig. 5 that the calculated DOS exhibits sharper peaks than the experimental spectra because we have not included the lifetime broadening in our DOS calculations. The experimental bandwidth is larger than the calculated one. This
Fig. 5. The X-ray photo-emission spectrum in the valence band region is compared with the calculated DOS spectra of Pb(Fe1/2 Nb1/2 )O3 .
84
C. Bharti et al. / Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
(t2g –p –t2g ) and may be responsible for the electrical conduction in PFN. It is to be mentioned that if one considers the energy band diagram refers to the polycrystalline system such as PFN (where grain boundary is depicted having two successive grains) the donor states and the trapping states are the potential states available below the conduction band. The trapping states are presumed to have a range in the band gap below the conduction band. Upon the application of the ac small signal voltage these states will sweep through the amplitude of the small signal. These sweeping states are unlikely to be singular in nature for the defect states in polycrystalline materials. Therefore, the time constant comprises the charge mobility represented by the conductance and the charge storage represented by the capacitance for the defect states. This gives rise to the trapping contribution in the series event for the single charge that is being trapped and subsequently released to become mobile carrier. This event essentially gives the concept of R–C combination having relaxation time as a distributed parameter which in turn can be extracted from the complex plane plot. It should be noted that the activation energy ∼0.44 eV as obtained from the Arrhenius plot of dc is nearly equal to the activation energy ∼0.37 eV obtained from Arrhenius plot of M . These experimental results imply that the nature of charge carriers responsible for dielectric relaxation peaks and dc conduction belongs to the same category, which indicates that the polarization relaxation has a close relation with the conductivity in grain interior, and the polarization process probably depends on the conducting of the charge in the grain interior, which result that the two activation energies obtained are almost same. We have also compared the core level XPS spectra of Pb-4f, Fe-2p and Nb-3d with the corresponding convoluted partial-DOS data as shown in Fig. 6. The high resolution XPS spectrum of Pb4f is shown in Fig. 6(a). Only one spin–orbit doublet is observed. The first component at around 140.3 eV is attributed to the Pb4f7/2 peak, whereas the second at around 145 eV is attributed to the Pb-4f5/2 peak. The difference between Pb-4f7/2 and Pb-4f5/2 giving spin–orbit splitting is approximately 4.7 eV for the PFN compound and is comparable to that of the other lead-based relaxor compounds [45,46]. Fig. 6(b) shows the high resolution XPS spectrum of Fe-2p. The Fe-2p spectrum is split up into two components, 2p3/2 (714 eV) and 2p1/2 (725.6 eV) due to the spin–orbit effect. The spin orbit splitting of Fe-2p level is about 12 eV. The 2p region of compounds having electronic configuration d5 has been extensively studied both theoretically [47] and experimentally [48,49] and resembles with Fe-2p spectrum of PFN. The Fe-2p3/2 peak of PFN shows a distinct satellite around 718 eV. This satellite is caused by a dynamic charge-transfer during the photo-emission process, and appears in the XPS spectra of most transition metal compounds. The effective suppression of Fe2+ ions in PFN is confirmed by the XPS spectrum which suggests a negligible concentration of oxygen vacancies as shown in Fig. 6(b). As expected, peak appears around 714 eV and a satellite appears at 718 eV, indicating the existence of Fe3+ , while absence of peaks at 709.5 and 716 eV indicates an absence of Fe2+ [50]. The high-resolution XPS spectrum of Nb-3d is shown in Fig. 6(c). It is clearly seen that the Nb-3d spectrum consists of two peaks corresponding to its angular momentum of electrons. Only one spin–orbit doublet is observed for Nb-3d5/2 (208.8 eV) and Nb-3d3/2 (211.5 eV). It means that Nb ion exhibits only one chemical state characteristic for PFN. The energy positions of both peaks coincide well with previous reports on different niobium-based perovskite compounds [51]. Thus the XPS investigation allows us to determine the relative concentrations of the various constituents of PFN. The experimental core level XPS spectra are compared with the calculated corresponding spectra (dashed lines) in Fig. 6(a)–(c) for Pb-4f, Fe-2p and Nb-3d, respectively. All the calculated spectra (dashed
Fig. 6. The X-ray photo-emission spectra of the Pb-4f state (a), Fe-2p state (b) and Nb-3d state (c) are compared with the corresponding convoluted partial-DOS spectra of Pb(Fe1/2 Nb1/2 )O3 .
lines) appear to be very similar with the peak positions and relative intensities of XPS spectra thus implying a good agreement between the experimental results and our theoretical calculations. 4. Conclusions The frequency-dependent dielectric dispersion of polycrystalline PFN synthesized by the Coulombite technique is investigated in the frequency range from 100 Hz to 1 MHz and in the temperature range from 203 to 363 K. The frequency-dependent electrical data are analyzed in the framework of conductivity and impedance formalisms. The complex plane plot suggests that the relaxation mechanism in PFN is purely a bulk effect arising from the semiconductive grains. Modulus spectroscopic data are used to gain an insight into the electrical properties of the samples and fitted with the Davidson–Cole model. The scaling behaviour of imagi-
C. Bharti et al. / Journal of Electron Spectroscopy and Related Phenomena 169 (2009) 80–85
nary electric modulus suggests that the relaxation shows the same mechanism at various temperatures. We have studied the electronic structure of the PFN using X-ray photoemission spectroscopy (XPS). The DOS of PFN as obtained from the first principles full potential linearized augmented plane wave calculation shows a direct energy gap ∼0.43 eV. The d-states of transition metals (Fe and Nb) split into t2g and eg states due to the presence of crystal field produced by oxygen octahedra. The valence band XPS spectrum is compared with the calculated DOS spectra and exhibits a strong hybridization between the Fe-d-states and O-p-states. It has been observed that the electrical properties of PFN are dominated by the interaction between transition metal and oxygen ions as its valence band consists mainly of the oxygen p-states hybridized with the iron d-states. Acknowledgement One of the authors (Alo Dutta) acknowledges the financial support provided by the CSIR, New Delhi in the form of SRF. References [1] G.A. Smolensky, A.I. Agranovskaya, S.N. Popov, V.A. Isupov, Sov. Phys. -Tech. Phys 3 (1958) 1981. [2] N. Yasuda, Y. Ueda, J. Phys.: Condens. Matter 1 (1989) 497; N. Yasuda, Y. Ueda, J. Phys.: Condens. Matter 1 (1989) 5179. [3] C.N.W. Darlington, J. Phys.: Condens. Matter 3 (1991) 4173. [4] N. Lampis, P. Sciau, A.G. Lehmann, J. Phys.: Condens. Matter 11 (1999) 3489. [5] S.A. Ivanov, R. Tellgren, H. Rundlof, N.W. Thomas, S. Ananta, J. Phys.: Condens. Matter 12 (2000) 2393. [6] V. Bonny, M. Bonin, P. Sciau, K.J. Schenk, G. Chapuis, Solid State Commun. 102 (1997) 347. [7] H. Schmid, Ferroelectrics 162 (1994) 317. [8] Y. Yang, J.M. Liu, H.B. Huang, W.Q. Zou, P. Bao, Z.G. Liu, Phys. Rev. B 70 (2004) 132101. [9] J.T. Wang, C. Zhang, Z.X. Shen, Y. Feng, Ceram. Int. 30 (2004) 1627. [10] V.V. Bhat, A.M. Umarji, V.B. Shenoy, U.V. Waghmare, Phys. Rev. B 72 (2005) 14104. [11] Y. Yang, S.T. Zhang, H.B. Huang, Y.F. Chen, Z.G. Liu, J.M. Liu, Mater. Lett. 59 (2005) 1767. [12] S.B. Majumder, S. Bhattacharyya, R.S. Katiyar, J. Appl. Phys. 99 (2006) 024108. [13] O. Raymond, R. Font, R. Machorro, E. Martinez, J.M. Siqueiros, MRS Proc. 1034E (2006) K3.24. [14] G. Alvarez, R. Font, J. Portelles, R. Zamarano, R. Valenzuela, J. Phys. Chem. Solids 68 (2007) 1436.
85
[15] R. Blinc, P. Cevc, A. Zorko, J. Holc, M. Kosec, Z. Trontelj, J. Pirnat, N. Dalal, V. Ramchandran, J. Krzystek, J. Appl. Phys. 101 (2007) 033901. [16] S. Ananta, N.W. Thomas, J. Eur. Ceram. Soc. 19 (1999) 2917. [17] S.B. Lee, S.H. Yoon, H. Kim, J. Euro. Ceram. Soc. 24 (2004) 2465. [18] C.C. Chiu, S.B. Desu, K. Amanuma, Mater. Sci. Eng. B 22 (1994) 133. [19] M. Yokosuka, Jpn. J. Appl. Phys. 32 (1993) 1142. [20] A.A. Bokov, L.A. Shpak, I.P. Rayevsky, J. Phys. Chem. Solids 54 (1993) 495. [21] S. Ananta, N.W. Thomas, J. Euro. Ceram. Soc. 19 (1999) 155. [22] S.L. Fu, G.F. Chen, Ferroelectrics 82 (1988) 119. [23] M. Lejeune, J.P. Boilot, Ceram. Int. 8 (1982) 99. [24] C.C. Chiu, S.B. Desu, Mater. Sci. Eng. B 21 (1993) 26. [25] M. Abbate, H. Ascolani, F. Prado, A. Caneiro, Solid State Commun. 129 (2004) 113. [26] V.R. Mastelaro, P.N. Lisboa-Filho, P.P. Neves, W.H. Schreiner, P.A.P. Nascente, J.A. Eiras, J. Electron, Spectrosc. Rel. Phenom. 156 (2007) 476. [27] S. Shannigrahi, K. Yao, J. Appl. Phys. 98 (2005) 034104. [28] V.A. Gasparov, S.N. Ermolov, G.K. Strukova, N.S. Sidorov, S.S. Khassanov, H.-S. Wang, M. Schneider, E. Glaser, W. Richter, Phys. Rev. B 63 (2001) 174512. [29] Alo Dutta, T.P. Sinha, S. Shannigrahi, Phys. Rev. B 76 (2007) 155113. [30] Alo Dutta, T.P. Sinha, S. Shannigrahi, J. Appl. Phys. 104 (2008) 064114. [31] Alo Dutta, T.P. Sinha, B. Pahari, R. Sarkar, K. Ghoshray, S. Shannigrahi, J. Phys.: Condens. Matter 20 (2008) 445206. [32] P. Hohenberg, W. Kohn, Phys. Rev. 136 (1964) B864. [33] J.P. Perdew, J.A. Chevary, S.H. Vosko, K.A. Jackson, M.R. Pederson, D.J. Singh, C. Fiolhais, Phys. Rev. B 46 (1992) 6671. [34] P. Blaha, K. Schwarz, G. Madsen, D. Kvasnicka, J. Luitz, WIEN2k, an augmented plane wave + local orbitals program for calculating crystal properties, Karlheinz Schwarz, Techn. Universität Wien, Austria, 2001. [35] A.K. Jonscher, J. Phys. D: Appl. Phys. 13 (1980) L89. [36] A.K. Jonscher, Dielectric Relaxation in Solids, Chelsea Dielectrics Press, London, 1983. [37] A.K. Jonscher, Universal Relaxation Law, Chelsea Dielectrics Press, London, 1996. [38] J. Liu, Ch.-G. Duan, W.-G. Yin, W.N. Mei, R.W. Smith, J. Chem. Phys. 119 (2003) 2812. [39] P. Debye, Polar Molecules, Chemical Catalogue Company, New York, 1929. [40] D.W. Davidson, R.H. Cole, J. Chem. Phys. 19 (1951) 1484. [41] S. Saha, T.P. Sinha, J. Appl. Phys. 99 (2006) 014109. [42] Alo Dutta, T.P. Sinha, J. Phys. Chem. Solids 67 (2006) 1484. [43] S. Saha, T.P. Sinha, A. Mookerjee, Phys. Rev. B 62 (2000) 8828. [44] S. Saha, T.P. Sinha, A. Mookerjee, J. Phys.: Condens. Matter 12 (2000) 3325. [45] V.R. Mastelaro, P.N. Lisboa-Filho, P.P. Neves, W.H. Schreiner, P.A.P. Nascente, J.A. Eiras, J. Electron. Spectrosc. Rel. Phenom. 156–158 (2007) 476. [46] F. Parmaigiani, L. Rollandi, G. Samoggia, L.E. Depero, Solid. State Commun. 100 (1996) 801. [47] R.P. Gupta, S.K. Sen, Phys. Rev. B 10 (1974) 71. [48] S.P. Kowalczyk, L. Ley, F.R. McFeely, D.A. Shirley, Phys. Rev. B 11 (1975) 1721. [49] J.A. Tossell, J. Electron. Spectrosc. 10 (1977) 169. [50] W. Eerenstein, F.D. Morrison, J. Dho, M.G. Blamire, J.F. Scott, N.D. Mathur, Science 307 (2005) 1203a. [51] M. Kruczek, E. Talik, A. Kania, Solid State Commun. 137 (2006) 469.