Introduction

At present, lead zirconate titanate [Pb(Zr1−x Ti x )O3, known as PZT] and its modified compositions are in high demand in industries due to their versatile device applications like FeRAM, DRAM and display devices. In addition, these materials are also used in electro-optical modulators, pyroelectric and gas sensors, transducers, hydrophones, other electronic and optoelectronic applications with high efficiency and tunability. PZT is an ABO3-type perovskite oxide and a solid solution of PbZrO3 and PbTiO3. Among the compositions in the range from x = 0 to 1 for Pb(Zr1−x Ti x )O3, the composition with x = 0.48 exhibits enhanced ferroelectric, piezoelectric and dielectric properties. At this composition, (PbZr0.52Ti0.48O3 [PZT(52/48)]) exhibits the morphotropic phase boundary (MPB) between Zr-rich rhombohedral and Ti-rich tetragonal phases [1,2,3,4,5]. As researchers continue to explore and enhance the properties of PZT by the modification of A-site, as well as B-site with suitable elements, many new properties and application possibilities are evolving.

Multiferroics offer the possibilities of fast low power electrical writing with nondestructive magnetic writing operation in a combination of high-quality FeRAM and MRAMs [6]. Hence, a trend to get multiferroics at room temperature has become more interesting and popular. Room temperature multiferroic nature has been found to be limited to fewer materials like BiFeO3, CdSr2S4, LuFe2O4, RMnO3, RMn2O5 and boracites. There have been several attempts to convert ferroelectrics into multiferroics either by introducing magnetic dopants [7, 8] or by making ferroelectric–ferromagnetic composite systems [9,10,11,12,13,14,15,16]. Xu et al. [7] reported that BaTiO3 (one of the most popular ferroelectric materials) exhibits ferroelectricity as well as ferromagnetism at room temperature upon Fe doping. Around this time, Kleebe et al. [8] successfully demonstrated the preparation and characterization of 3 mol% Fe-doped PZT(60/40) which is above the solubility limit (approximately 1 mol%). Recently, room temperature multiferroic properties of Nb- and Ta-doped PZT near MPB composition are reported by Schiemer et al. [17]. In this case, both Nb and Ta are co-doped, each up to 2 mol%.

The present work deals with the preparation and characterization of Fe and Nb co-doped (up to 8 mol% each) PLZT. Rietveld refinement was performed to confirm the phase purity and understand the impact of Nb and Fe co-doping on structural changes. In the first decade of the twenty first century, intense use of Raman spectroscopy for characterization of PZT-based materials has been seen [18,19,20,21,22,23,24,25,26,27,28,29]. With the help of those earlier reports and multiple peaks fitting of Raman spectral data, the successful incorporation of Nb and Fe in PLZT lattice is confirmed. A detailed lattice vibrational analysis clearly shows that it is possible to incorporate up to 6 mol% of Fe and Nb together in PLZT. However, doping of 8 mol% of Fe and Nb exhibits marginal effect with respect to the changes observed for the sample with 6 mol% doping in Raman spectroscopy. Band gap tuning is a very important technology, and in particular, tailoring band gap of large band gap semiconductors renders them useful for several applications. In this study, Nb and Fe co-doping is found to be very effective in the reduction in band gap of PLZT.

Experimental details

Samples of Pb0.97La0.02(Zr0.52Ti0.48)1−2x (Nb0.5Fe0.5)2x O3 for x = 0.00, 0.02, 0.04, 0.06 and 0.08 (abbreviated as NF0, NF2, NF4, NF6 and NF8, respectively) are prepared by sol–gel synthesis following the steps shown in the flowchart (Fig. 1).

Figure 1
figure 1

Flowchart of sol–gel synthesis of Nb and Fe co-doped PLZT (NF0, NF2, NF4, NF6 and NF8)

Full size image

All chemicals used for the preparation of PLZT (NF0, NF2, NF4, NF6 and NF8) are procured from Sigma-Aldrich. Lead(II) acetate trihydrate (99.999%) and Lanthanum(III) acetate hydrate (99.9%) are added to glacial acetic acid and stirred to get a homogeneous solution. In another beaker, Niobium(V) ethoxide (99.95%) is added to glacial acetic acid and stirred for an hour. Then Zirconium(IV) propoxide solution (70 wt% in 1-propanol), Titanium(IV) isopropoxide (97%) and Ferric nitrate are added to the niobium ethoxide–acetic acid solution one after the other with continuous stirring. Ferric(III) nitrate nonahydrate (99.95%) is turned into aqueous solution before adding. After stirring at room temperature for about 30 min, these two solutions are mixed. The mixture looks milky white in color at room temperature. Then, the temperature is increased slowly to 90 °C with continuous stirring, and after a few minutes, the solution gradually becomes colorless. The solution is kept at this temperature until it becomes white powder via thick gel formation. It takes almost 96 h for the thick gel to become dry powder. The white powder is ground and then calcined at 700 °C for 1 h. This has led to an orange color PLZT powder.

X-ray diffraction(XRD) studies are carried out using Rigaku SmartLab diffractometer with Cu Kα radiation (λ = 1.5406 Å), and Raman spectroscopic studies are carried out in LabRAM HR spectrometer (Horiba Jobin–Yvon, France) with the help of LabSpec 6 software. The DRS and PL experiments are performed using Bentham TMc300 monochromator and Horiba Fluorolog-3.

Results and discussion

Structural analysis

Phase purity of the samples is determined by XRD analysis, and the results are shown in Fig. 2. XRD patterns indicate that the samples exhibit perovskite structure without the presence of any secondary phases like pyrochlore or phases related to niobium and iron oxides. XRD results confirm the formation of tetragonal structure, as they are in good agreement with the previously reported tetragonal PZT. The patterns are indexed using well-matched JCPDS card #33-0784 and COD ID 1526174. The absence of any impurity phase confirms the successful incorporation of Nb and Fe in PLZT lattice. A systematic change with doping concentration of Nb and Fe is observable in the closer view (Fig. 2b–d) of the first three peaks. Due to doping of 8 mol% of Nb and Fe, the (100) peak shifted toward higher 2θ by about 0.13°. Similar shifts are also observed for (110) and (111) peaks. The shift in 2θ value for these two peaks is around 0.16°. The shifts for all the peaks are toward higher 2θ. The shift is prominent at lower 2θ values while at higher angles the shift is not so prominent due to the fact that the intensity is relatively low and peak width is high at higher 2θ. There are many reasons which could be responsible for this effect. Some of the possible technical sources which may cause peak shift are avoided/minimized by taking precautions which include “optics alignment” prior to the start of the experiment. Rietveld refinement has also been performed using FullProf suite to get an insight into the peak shift. The results reveal that the lattice parameters are reduced upon Nb and Fe doping. For NF0, the lattice parameters a and c are found to be 4.0644 and 4.0995 Å, respectively, which decrease to 4.0536 and 4.0808 Å, respectively, in NF8 (Fig. 3). Hence, the unit cell volume has been found to decrease from 67.72 to 67.05 Å3 due to 8 mol% co-doping of Nb and Fe. The Nb5+ and Fe3+ have crystal ionic radii ~78 and 69 pm, respectively (effective ionic radii ~64 and 55 pm, respectively), while the corresponding values for Zr4+ and Ti4+ are 86 and 74.5 pm, respectively (effective ionic radii ~72 and 60.5 pm, respectively). This suggests that doping of smaller ions causes shrinkage of the unit cells. Hence, the shift in peaks toward higher 2θ values can be attributed to decrease in the lattice parameters.

Figure 2
figure 2

XRD results for Nb and Fe co-doped PLZT. a “Inset” shows the plot of the dataset with vertical log scale. b–d Show the zoomed view of the shifting of peaks with doping concentration

Full size image
Figure 3
figure 3

Change in lattice parameters due to change in the doping concentration of Nb and Fe co-doped PLZT, as calculated by Rietveld refinement using FullProf suite

Full size image

Analysis of lattice vibrations

Raman spectra recorded for NF0, NF2, NF4, NF6 and NF8 are shown in Fig. 4. The results exhibit well-distinguished phonon modes including the “soft modes.” In order to detect visible changes, which occur due to gradual incorporation of Nb and Fe into PLZT lattice, the energy range from 20 to 1000 cm−1 is divided into several regions and plotted separately with closer view (A, D, E and F in Fig. 4). The spectra for NF0 show both the soft modes in the frequency range between 50 and 100 cm−1 (A in Fig. 4) having nearly equal intensity, while the peak around 76 cm−1 gains relatively higher intensity with increasing Nb and Fe contents. The peaks around 106 and 138 cm−1 (marked as B and C) are visible in the case of NF0 and, however, seem to disappear due to Nb and Fe doping and the regions become smooth. The panel D shows another two peaks that have lesser intensity and larger FWHM due to incorporation of dopants. The broad peaks around 580 and 730 cm−1 (panel E and F in Fig. 4) show redshift for doping up to 6 mol%, and interestingly, the shift is smaller in case of NF8.

Figure 4
figure 4

Raman spectra of Nb and Fe co-doped PLZT. The four panels on the right side are the closer views of regions marked as A, D, E and F in the left panel. B and C are found to show two humps, which disappear upon Nb and Fe doping

Full size image

In order to identify the phonon modes and analyze their characteristic changes, the spectra are de-convoluted to individual peaks. Ferroelectric tetragonal phase with space group C 14v (P4 mm) has 3A 1 + 4E polar modes and B 1 nonpolar modes. Each polar mode has transverse optical (TO) and longitudinal optical (LO) branches. Hence, there are 15 modes labeled with suffix number 1–4 according to their position in frequency spectrum from low to high. All these modes along with another three modes (E(TO), E(TA) and A 1(TA)) mentioned by Freire and Katiyar [30] are listed in Table 1. For the pure PLZT(2/52/48), i.e., NF0, the modes are shown in Fig. 5. The identified and indexed modes are compared with earlier reports in the table. The mode observed at 57 cm−1 can be either E(TA) [30] or a bandpass [28]. The peak found in fit around 694 cm−1 is still unidentified and is shown with a “#” mark (Fig. 5).

Table 1 Comparison of Raman spectroscopic results of NF0 (i.e., PLZT (2/52/48)) with earlier published results of similar structures
Full size table
Figure 5
figure 5

Cumulative fit (red solid line) of Raman spectroscopic data for NF0. The colored lines (except the red line) are the de-convoluted individual peaks. “#” and “*” marks represent unidentified and unconfirmed peaks

Full size image

Cumulative fit and de-convoluted peaks for Raman spectra of NF2, NF4, NF6 and NF8 are shown in Fig. 6. From these detailed analyses, it is found that E(TO2) and A 1(TO3) modes exhibit redshift upon Nb and Fe doping, which is similar to that observed in Fig. 4. Up to 6% doping, the phonon softening happens gradually (Fig. 7). For NF0, the E(TO2) mode frequency is near 201 cm−1, which shifts to below 193 cm−1 for NF6. Similarly, the A 1(TO3) mode shifts from 596 to 555 cm−1. It is interesting to note that both the modes shifted toward higher energy on further doping. For NF8 (i.e., 8 mol% Nb- and Fe-doped PLZT), the peak positions become 195.5 and 558.5 cm−1, respectively. To understand the change, schematic two-dimensional diagrams are drawn for these two vibrations as shown in Fig. 7b. The E(TO2) mode corresponds to the vibration of central B (in ABO3 perovskite) atom (Zr/Ti/Nb/Fe) against the octahedron in which all other atoms vibrate in opposite direction. The E(TO2) mode vibration is perpendicular to the polarization. In case of A 1(TO3) mode, the vibration occurs in a similar way but parallel to polarization, i.e., along c-axis.

Figure 6
figure 6

Cumulative fit (red solid line) of experimental Raman spectroscopic data for NF2, NF4 NF6 and NF8. The colored lines (except the red line) are the de-convoluted individual peaks

Full size image
Figure 7
figure 7

a Redshift of E(TO2) and A 1(TO3) modes. b Schematic diagram of E(TO2) and A 1(TO3) mode vibrations viewed along c-axis

Full size image

Though the E(TO2) and A 1(TO3) vibrations are perpendicular to each other, they show similar characteristic changes which can be correlated with the fact of decreasing lattice parameters, a and c. This is a direct consequence of Nb and Fe co-doping.

Band gap study

In order to observe band gap and related changes due to co-doping of Nb and Fe, diffuse reflectance spectroscopy (DRS) is carried out. Figure 8 shows reflectance (R) at room temperature in the wavelength range from 325 to 1100 nm for NF0, NF2, NF4, NF6 and NF8. A distinct change in the optical properties of PLZT nanoparticles due to the addition of Nb and Fe is observed from these studies. Reflectance edge shifts toward higher wavelength, which is an outcome of the increasing ability of absorption of lower energy photons by the materials. This indicates that there is a decrement in the band gap (effective band gap) by substitution of Nb and Fe ions in PLZT lattice. For quantitative analysis, the Kubelka–Munk function \( F(R) = \frac{{(1 - R)^{2} }}{2R} \) has been calculated and plots of (F(R)hν) versus hν are shown in Fig. 9a. Here h is Planck’s constant and ν, the frequency of radiation. These curves show sharp transition between 400 and 500 nm indicating the photon energy required/emitted for electronic transitions. The energy intercept of linear portion of the curve gives the value of band gap of the material. The band gap of NF0 calculated is 3.21 eV.

Figure 8
figure 8

Diffuse reflectance spectra for NF0, NF2, NF4, NF6 and NF8

Full size image
Figure 9
figure 9

Band gap of NF0, NF2, NF4, NF6 and NF8. a Kubelka–Munk plot for determination of band gap. b Change in band gap due to increasing Nb and Fe doping concentration. c Schematic representation of band gap decrement by impurity energy band induced by Nb and Fe. d PL spectra of the samples. e Schematic of the mechanism of PL emissions as observed in (c)

Full size image

Undoped PZT shows band gap in the range around 2.6–3.5 eV [35], which may vary further depending on morphology, composition, processing, etc. [36]. Calculation shows that the band gap of PZT can vary from 3.45 to 3.72 eV [37]. For compositions PZT(52/48), i.e., around MPB, it is around 3.5 eV. Experimentally, it is observed that PZT prepared at 700 °C exhibits a band gap of 3.39 eV [38] while similar composition shows the values of band gap 3.36 and 3.38 eV, respectively, for band gaps in case of the samples prepared at 700 and 800 °C, respectively [39]. La doping could lower the band gap. As a result, it is seen that the band gap of NF0 (which is 2% La-doped PZT(52/48)) is 3.21 eV. The calculated values of band gap of NF2, NF4, NF6 and NF8 are 2.74, 2.70, 2.64 and 2.59 eV, respectively. The variation in band gap with doping concentration is plotted in Fig. 9b. The blueshift due to filling of near band gap states by heavy doping in a semiconductor is known as Burstein–Moss shift [40]. A schematic diagram (Fig. 9c) of band gap shrinkage due to the introduction of Nb and Fe in PLZT lattice has been drawn with the help of mechanism behind Burstein–Moss shift and is supported by photoluminescence (PL) experimental results (discussed below).

Photoluminescence (PL) emission spectra observed using an excitation wavelength of 325 nm (i.e., of energy 3.81 eV) are shown in Fig. 9d. Three broad peaks are found around 470 nm (2.64 eV), 482 nm (2.57 eV) and 492 nm (2.52 eV) which are almost consistent in energy for all the samples. This indicates that Nb and Fe doping does not affect the energy levels involved in these transitions. A broader peak observed at lower energy (2.3–2.45 eV) exhibits a redshift with increasing doping concentration. The position of the peak is 518 nm (2.39 eV), 522 nm (2.38 eV), 525 nm (2.36 eV), 528 nm (2.35 eV) and 532 nm (2.33 eV) for NF0, NF2, NF4, NF6 and NF8, respectively. The relative intensity of this particular peak with respect to the other three peaks (2.64, 2.57 and 2.52 eV) is found to increase as doping concentration increases.

Defects, such as Pb vacancies, O vacancies, Pb–O vacancy complex centers, the displacement of the Ti and distortion of TiO6 octahedra, are considered to have importance in the mechanisms responsible for PL at room temperature in ABO3-type materials [41, 42]. Pb evaporation during heat treatment of any Pb-based material is a well-known phenomenon. Pb evaporates as PbO creating Pb and O deficiency in the material. According to Eyraud’s model [36, 39, 43, 44], when Pb evaporates as PbO, it leaves Pb and O vacancies which may create acceptors and donors in the following way:

$$ V_{Pb} \to V^{\prime}_{Pb} + e^{\prime},V^{\prime}_{Pb} \to V^{\prime\prime}_{Pb} + e^{\prime},V_{O} \to V_{O}^{ \cdot } + e^{\prime}\,{\text{and}}\,V_{O}^{ \cdot } \to V_{O}^{ \cdot \cdot } + e^{\prime} $$

Energy of defects such as V ′ Pb , V ′ ′ Pb , V · O and V ·· O (symbols for single and doubly ionized lead and oxygen vacancies) stay in the forbidden region, just below the minimum of conduction band and/or just above maximum of valence band [43]. Depending on the availability of the energy levels (mid-band gap states) created by these vacancies, there can be multiple absorption processes by these defect states leading to a number of PL emissions peaks, if the excitation energy is sufficient. The four most possible transitions (3.2, 2.64, 2.57 and 2.52 eV) for an excitation of 3.81 eV (~325 nm) are shown in the schematic diagram. The vacancy defects are due to evaporation of Pb, which strongly depends on the temperature of preparation and La doping. Higher temperature creates more Pb loss, and La3+ in place of Pb2+ also creates lead vacancies to compensate extra charge. However, as the preparation temperature and the amount of La are the same for all the five samples (NF0, NF2, NF4, NF6 and NF8), the energy bands created by Pb and O vacancies are expected to remain constant for all the samples. As a result, we do not see any change in PL spectra for the three peaks around 2.64, 2.57 and 2.52 eV. The transition at 3.2 eV has not been observed in the spectra as it is beyond the experimental range. The fourth peak in PL spectra (2.30–2.45 eV) is strongly affected by Nb and Fe. The peak intensity increases, and it shifts toward lower energy as doping concentration increases. These effects indicate that this emission is associated with the impurity level induced by Nb and Fe in addition to that induced by La already. As the doping concentration increases, the width of the impurity level increases. Hence, the effective band gap reduces. As a result, a change in band gap is observed (Fig. 9a, b); an effect is known as Burstein–Moss shift as explained above. As the width increases, the effective band gap reduces and the number of transitions associated with it increases. As a result, the relative intensity increases and redshift (2.39–2.33 eV in Fig. 9c) occurs.

Conclusions

Up to 8 mol% of Nb and Fe are simultaneously doped in PLZT. The XRD study confirms the successful incorporation of Nb and Fe in PLZT lattice. The lattice parameters are found to decrease gradually with increasing doping concentration. A total of 14 modes are observed from Raman spectral analysis, which have frequencies close to the theoretically calculated frequencies. Upon doping, a regular redshift is observed in several modes. Soft phonon modes show relative intensity change due to the presence of Nb and Fe. The effective band gap change due to 2% Nb and Fe co-doping shows a decrease from 3.21 to 2.74 eV. The band gap decreases further down to 2.59 eV upon doping of 8 mol%. The PL study satisfactorily explains the mechanism behind the changes observed.