Ferroelectric PbZr 1−x Ti x O 3 by ethylene glycol-based chemical solution synthesis
FFerroelectric PbZr Ti x O by ethyleneglycol-based chemical solution synthesis Ewout van der Veer, ∗ , † Mónica Acuautla, ∗ , ‡ and Beatriz Noheda ∗ , † † Zernike Institute for Advanced Materials, Nijenborgh 4, 9747AG Groningen, TheNetherlands ‡ Engineering and Technology Institute Groningen (ENTEG), Nijenborgh 4, 9747AGGroningen, The Netherlands
E-mail: [email protected]; [email protected]; [email protected]
Abstract
We have investigated a water-stable sol-gel method based on ethylene glycol as a sol-vent and bridging ligand for the synthesis of ferroelectric lead zirconate titanate in bulkand thin film forms. This method offers lower toxicity of the solvent, higher stabilitytowards atmospheric moisture and a simplified synthetic procedure compared to tradi-tional sol-gel methods. Ceramic pellets of Nb-doped lead zirconate titanate (PNZT) inthe rhombohedral phase were produced with high density and good piezoelectric prop-erties, comparable to those reported in the literature and those found in commercialpiezoelectric elements. In addition, a nine-layer thin film stack was fabricated from thesame sol by spin coating onto platinized silicon substrates. The films were crack-freeand showed a dense perovskite grain structure with a weak (111) orientation. Piezoelec-tric measurements of the film showed a piezoelectric coefficient comparable to literaturevalues and good stability towards fatigue. a r X i v : . [ phy s i c s . c h e m - ph ] J u l ntroduction Sol-gel methods are commonly used for the fabrication of oxide materials with a wide rangeof functionalities. These methods involve the synthesis of a precursor solution (known as‘sol’) containing oligomeric chains of metal ions and oxygen atoms. Treatment of the sol,for example by the addition of water or by heating, causes the formation of a continuousmetal-oxygen network, leading to gelation of the sol. The sol may be processed into a varietyof products, such as bulk powders (by simply heating the gel), thin films by deposition of thesol onto a substrate ( e.g. by spin coating) or a multitude of other forms.
Employing a sol-gel-type synthesis for the production of oxide materials facilitates the control of compositionand doping, making high homogeneity and short fabrication cycles possible. Furthermore,when used to fabricate thin films, it can give rise to smooth films covering a large surfacearea with a wide range of film thicknesses up to several micrometers.One material which may be produced using a sol-gel method is the well-known lead zir-conate titanate solid solution (PbZr Ti x O , also known as PZT), which is ferroelectric and,thus, piezoelectric, allowing for its use as sensors and actuators. The PZT composition withx=0.48 lies at a phase boundary between two different crystal structures with tetragonal(for Ti-rich compositions) and rhombohedral (for Zr-rich compositions) symmetries, wheremonoclinic structures have been observed. At this boundary, known as as the morphotropicphase boundary (MPB), the piezoelectric coefficients are maximized. The piezoelectric pa-rameters of PZT can be further improved through chemical doping with elements such asniobium (PbNb y (Ti x Zr ) O or PNZT). Traditional sol-gel methods used for the production of thin films of PZT, first reportedby Budd, Dey and Payne, make use of the highly toxic 2-methoxyethanol as a solvent andalkoxides and acetates as the precursors for lead, zirconium and titanium. These methodsrely on hydrolysis and condensation reactions of the alkoxide precursors to form a polymericnetwork of metal-oxygen-metal bonds. These methods use water for the initiation of thehydrolysis reaction. Hence, sols produced using such an approach tend to be sensitive to the2resence of water. As a result, these sols require storage and processing in an oxygen-freeand water-free environment, such as a glovebox.More recently, there has been an interest in the development of chemical solution de-position (CSD) methods which are not based on hydrolysis-condensation reactions, insteadrelying on different types of reactions.
One example of such a non-aqueous CSD methodis based on ethylene glycol as bridging ligand and common alkoxides and acetates as reagents.This method was reported to be nontoxic, more stable to atmospheric moisture and havea more straightforward synthesis procedure. However, an investigation of the ferroelectricand piezoelectric properties of materials derived from this CSD method has, to our knowl-edge, not been reported.We have studied the properties of both bulk and thin film products fabricated using theethylene glycol-based CSD method . We show that multilayer stacks of thin films can beproduced without cracks, voids or parasitic phases by carefully designing the deposition andheat treatment procedures, despite the presence of a large amount of organic material inthe as-deposited film, which is commonly known to reduce film quality. The piezoelectricbehavior of the bulk and thin film products are comparable to reported values for filmsof similar characteristics. Finally, we have investigated the sensitivity of the solution tomoisture.
Results and discussion
Properties of the sol as well as structural and ferroelectric properties of the PNZT films andbulk ceramic pellets have been investigated.Figure 1 contains plots of the differential thermal analysis (DTA) and thermo-gravimetricanalysis (TGA) data collected from the sol dried at 230℃ on a hotplate. Initial weight lossoccurs around 300℃ and is associated with a peak in the DTA trace. This peak correspondsto loss of ethylene glycol groups. Further weight loss occurs between 320℃ and 400℃, cor-3esponding to a large exothermic peak in the DTA trace. This peak is presumably the resultof the removal of remaining organic material and the onset of crystallization of PNZT.
A total weight loss of approximately 23% was observed up to 400℃. The final peak presentat 842℃ is possibly due to melting of lead oxide in the sample. These results are in roughcorrespondence with those reported in the literature using a similar ethylene glycol-basedsolution deposition method. The sensitivity of the sol to the presence of water (for example, from the atmosphere), wasdetermined by directly adding various concentrations of water to 1 mL of the sol. The sol wasleft at room temperature in a dark location for over a month, yet no gelation occurred evenin sols to which 10 vol.% water was added. After a month, some gelation was observed, butthere was no correlation between the gelation time and the concentration of water that wasadded to the sol. Sols with up to 5 vol.% water were used to produce pellets as described inthe ‘Materials and Methods’ section below. These pellets were analyzed by x-ray diffractionand scanning electron microscopy to assess the influence of the addition of water to the sol onthe structural and microstructural properties of the product. No trends could be discernedin either the structural or the microstructural properties as the concentration of water wasincreased. These results show that the ethylene glycol-based sol is highly stable towardsmoisture.
Bulk
PNZT pellets were produced from the PNZT sol with a 20% excess of lead precursor. Thesol was dried at 230℃ and pyrolyzed at 420℃. Pellets with a nominal diameter of 10 mmwere pressed from this powder at 6.4 ton/cm . One pellet was sintered at 800℃ for 2 hours.X-ray diffraction analysis of this pellet suggested that the pellet was in the perovskite phase.Nevertheless, the peak splittings expected for either the rhombohedral or tetragonal phasesof PNZT were not present and no ferroelectric behavior was measured in this pellet. A secondpellet was sintered at 1200℃ for 2 hours. After sintering, the pellet had a diameter of 8.164igure 1: DTA and TGA traces of the sol dried at 230℃.mm, a thickness of 1.39 mm and a density of 6936 kg / m , that is 86.9% of the theoreticallypredicted density. Figure 2 shows an x-ray diffraction pattern of the pellet after sintering at 1200℃. A goodfit of this pattern was obtained using a combination of a rhombohedral PNZT phase and a β -PbO phase. This indicates that the excess of lead precursor in the sol is too high, leadingto the formation of an impurity phase. Nevertheless, good quality PNZT pellets could not beobtained using a lower lead excess. Blown-up versions of the (111), (200) and (220) peaks ofthe pattern are shown in figure 3. The splitting of the peaks indicates that the material is ina mostly rhombohedral phase, with a small admixute of a tetragonal or possibly a monoclinicphase. Hence, the material is approaching the morphotropic phase boundary betweenthe rhombohedral and tetragonal phases known to present the best piezoelectric response. The slight deviation from the exact composition at the morphotropic phase boundary mayresult from the loss of titanium precursor during synthesis due to its high reactivity withatmospheric moisture or impurity of the precursor itself.Scanning electron microscopy (SEM) of the pellet (figure 4) shows a dense grain structure5igure 2: X-ray diffraction pattern of the pellet and a fit of the profile using rhombohedralPNZT and β -PbO phases.with PNZT grains of 500-1000 nm. Additionally, large, plate-like crystals are present inthe PNZT matrix. These crystals were determined to be lead oxide by energy dispersivespectroscopy (EDS), confirming the presence of a lead oxide phase in the pellet.The pellet was poled in a silicone oil bath at 100℃ with an electric field of 29 kV/cmfor 30 minutes to align the dipoles in the material. Ferroelectric property measurementsof the bulk ceramic pellet were performed, yielding the polarization-electric field hysteresisloops and strain-electric field ("butterfly") loops expected for a ferroelectric, as displayedin figure 5. The remnant polarization measured for this pellet is P r = 9.5 µ C / cm , thecoercive field E c = 7.78 kV/cm and the longitudinal piezoelectric coefficient d = 441 pm/V.The d coefficient obtained here is compared to literature values in table 1. Our PNZTpellet has piezoelectric properties in line with those found in literature, even competingwith commercially available piezoelectric elements. We expect that the ferroelectric andpiezoelectric parameters can be further increased by bringing the composition closer to themorphotropic phase boundary and by improved densification of the pellet by, for example,6igure 3: Blow-ups of the (a) (111), (b) (200) and (c) (220) peaks of the pattern in figure 2.Peaks originating from a tetragonal or monoclinic phase are indicated using green arrows.hot pressing. This work shows that the ethylene glycol CSD method is capable of producinga high-quality material despite the simplicity of the method.Table 1: Comparison of the longitudinal piezoelectric coefficient of the PNZT pellet fabri-cated using the ethylene glycol CSD method with literature values. d (pm/V) Doping element Production method Reference
441 Nb CSD This work500 Commercial n.a. Hinterstein et al.
475 Commercial n.a. Xu et al.
420 Undoped Sol-gel Sharma et al.
155 Undoped Wet chemical Choy et al.
300 Undoped Wet chemical Guiffard and Troccaz
569 La Sol-gel Shannigrahi et al.
269 Nd Sol-gel Shannigrahi et al.
325 La Wet chemical Sahoo and Panda
236 BiFeO /BaCu W O /CuO Solid-state Dong et al.
338 La/Nb Solid-state Singh et al.
520 Sr/Nb Solid-state Zheng et al.
255 Nb Solid-state Garcia et al. Thin films
A nine-layer PNZT film was produced from the 1.5 M PNZT sol by spinning at 5000 rpmfollowed by drying on the hotplate, with pyrolysis and annealing steps performed after everythird layer. During heat treatment of these films, lead can be lost through evaporation at the7igure 4: Scanning electron microscopy image of a pellet sintered at 1200℃ showing PNZTgrains and larger lead oxide crystals (orange arrows).film surface and through diffusion into the silicon substrate. This leads to the formation of alayer of lead-deficient pyrochlore phase at the film surface or at the film-electrode interface.An excess of lead precursor can be added to the PNZT sol to compensate for this loss.However, too large an excess can cause the formation of voids in the film due to evaporationof the excess lead species. Therefore, careful control of the excess is required.To achieve such control, an alternative method was used here. A relatively small ex-cess of lead of 10% was added to the sol, compensating for diffusion but not evaporation.Additionally, a layer of pure lead oxide sol was deposited before the final pyrolysis step,compensating for evaporation from the film surface (see ref. 28). The resulting film shows adense structure with few voids and grains with sizes from several hundred nanometers up to1 micrometer (figure 6). Using this procedure, no lead-deficient pyrochlore phase was foundand no cracks or leakage paths are visible. These observations show that the combinationof a lead excess in the sol with a lead oxide overcoat is effective at producing high qualitythin films. A columnar grain structure is commonly observed in PZT thin films derivedfrom traditional sol-gel methods based on 2-methoxyethanol, due to bottom-up growth of8igure 5: Polarization and strain loops of sol-gel derived PNZT pellet sintered at 1200℃.the grains after heterogeneous nucleation at the film-electrode interface. Such structure isnot present in these films (figure 6b), indicating more homogeneous nucleation. This maybe the result of the high organic content of the as-deposited films compared to traditionalsol-gel-derived films.An x-ray diffraction pattern of the same film is shown in figure 7. The pattern showsa pure perovskite PNZT phase with no impurity peaks, except those originating from theplatinized silicon substrate. No peak splitting is observed due to the broadening of thepeaks. A small Pt(200) peak is present due to the top electrode, which is not perfectly (111)oriented. The preferential orientation of the PNZT film can be quantified by normalizing theintegrated peak intensities with the intensities of the x-ray diffraction patterns of a powderedsample using the following expression: P ( h i k i l i ) = I ( h i k i l i ) I ∗ ( h i k i l i ) (cid:88) hkl I ( hkl ) I ∗ ( hkl ) (1)9here P ( h i k i l i ) is a texture index quantifying the preferred orientation of the sample, I ( h i k i l i ) is the intensity in the thin film sample and I ∗ ( h i k i l i ) is the intensity in the powderedsample. The values in table 2 were obtained using the data in figure 7.Table 2: Texture index values of the thin film sample < h i k i l i > P ( h i k i l i ) <100> 0.25<110> 0.0086<111> 0.59<200> 0.11<211> 0.020<220> 0.016A <111> orientation is preferred in these films due to the <111> texture of the underlyingplatinum electrode, showing that at least some of the film nucleates heterogeneously at thefilm-electrode interface. However, it is evident that some of the film nucleates homogeneously,resulting in a decreased <111> texture of the film. This is in agreement with the lack ofcolumnar grains in the film.Figure 8a shows the ferroelectric hysteresis loop and strain loop of a nine-layer PNZTthin film stack. A double-beam laser interferometer, which corrects for substrate bending toextract the true deformation of the film, was used to collect these loops. The loops were col-lected by sweeping the potential applied to the top electrode between +800 kV/cm and –800kV/cm using a triangular waveform at a frequency of 100 Hz. The longitudinal piezoelectriccoefficient is extracted from the strain loop by determining its slope at zero applied field.The film showed a remnant polarization of 10.5 µ C / cm , an average coercive field of 61.3kV/cm, a longitudinal piezoelectric coefficient of 50 pm/V and a maximum deformation of1.41 nm, that is 0.3% of the thickness of the film. These values are again compared to thosefound in the literature, see table 3, which displays the wide range of piezoelectric parametervalues reported depending on the synthesis technique. Our piezoelectric coefficient is on thelow end of this range, but improvements can likely be made. For example, fabrication ofthicker films will improve piezoelectric behavior due to reduced clamping from the substrate.10able 3: Comparison of the longitudinal piezoelectric coefficient of the PNZT thin filmfabricated using the ethylene glycol CSD method with literature values. These results arefor undoped PZT, unless otherwise noted. *: 1 % Nb doping, OMCVD = organometalicchemical vapor deposition, PLD = pulsed laser deposition. d (pm/V) Production method Reference
50 CSD This work50 Sol-gel Balma et al.
77 Sol-gel Taylor and Damjanovic
200 OMCVD Lefki and Dormans
400 Sol-gel Lefki and Dormans
85 Sol-gel Ledermann et al.
200 Sol-gel Chen et al.
25 Sol-gel Wang et al.
106 PLD Goh et al. Figure 8b shows the fatigue response of the film. The film was switched at a frequencyof 200 Hz with an electric field amplitude of 114 kV/cm, that is above the coercive field.Ferroelectric hysteresis loops were collected at 3 points/decade with a field amplitude of 800kV/cm and a frequency of 100 Hz. The film is stable to fatigue for at least cycles. Theseresults are similar to those reported in the literature (see, e.g., refs. 29,39).To summarize, a sol-gel method was developed based on ethylene glycol as a solvent andbridging ligand. This sol was used for the production of pellets and thin films of ferroelectricniobium-doped lead zirconate titanate. This sol offers the advantages of a lower toxicitysolvent, improved stability during storage, decreased sensitivity to atmospheric moistureand the applicability to the synthesis of both bulk and thin film products. Furthermore, thesynthesis of the sol is less complex than that of traditional, 2-methoxyethanol-based sols.DTA of the sol shows that decomposition of the gel is finished at 400℃, with crystallizationof the desired PNZT phase occurring at higher temperatures. Pellets of bulk PNZT wereproduced, having a density of 86.9% of the theoretical density and a small lead oxide impurity.These show good properties with a coercive field of 68 kV/cm, a remnant polarization of9.5 µ C / cm and a piezoelectric coefficient of 441 pm/V, in line with literature values for11imilar PZT compositions. In addition, a nine-layer stack of PNZT thin films was fabricatedfrom the sol by spin-coating with a thickness of 440 nm. An excess of lead was suppliedto the thin films to compensate for evaporation and diffusion by combining the additionof an excess of lead precursor to the sol and the application of an overcoat of pure leadoxide. This method proved effective at suppressing the appearance of lead-deficient phasesor voids in the stack. The final stack shows a dense perovskite grain structure with a weak(111) out-of-plane texture. Ferroelectric and piezoelectric characterization of the film showsferroelectric coefficients close to literature values for thin films, with a remnant polarizationof 10.5 µ C / cm , a coercive field of 61.3 kV/cm, a piezoelectric coefficient of 50 pm/V and amaximum deformation of 0.3% of the thickness of the film. Furthermore, the film shows goodstability to fatigue up to cycles. This sol-gel method provides a safer, more water-stablealternative to traditional sol-gel methods based on 2-methoxyethanol for the fabrication ofbulk and thin film products. Materials and Methods
Sol synthesis COO) , PbAc , 23 mmol, 10 mol% excess, ≥ OH) , EG) were added to a three-neckedflask under a 0.5 lpm argon flow. An excess of lead acetate was used to compensate forlosses due to evaporation and diffusion during the heat treatment steps. The suspension washeated to 90℃ while stirring to dissolve the solids, then to 110℃ to expel any remainingwater from the solution. The sol was subsequently cooled to 90℃. 2.857 mL of titaniumisopropoxide (Ti(OCH(CH ) ) , 2.743 g, 9.65 mmol, 97%, Sigma Aldrich), 0.210 mL nio-bium ethoxide (Nb(OCH CH ) , 0.266 g, 0.836 mmol, 99.95%, Sigma Aldrich) and 4.686 mLof a 70 wt.% solution of ziconium n-propoxide in 1-propanol (Zr(OCH CH CH ) , 3.425 gZr(OCH CH CH ) , 10.5 mmol, Sigma Aldrich) were dissolved in 6.1 mL 1-propanol under12nert atmosphere. The Ti/Nb/Zr solution was added to the lead sol slowly limiting exposureto air. Some precipitate formed upon addition. A further 15.7 mL of EG was added, yeilding30 mL of solution at a concentration of metal ions of 1.5 M with a nominal composition ofPb Nb (Zr Ti ) O . The suspension was stirred at 90℃ until all precipitate hadredissolved. The sol was cooled to room temperature and 4 vol.% formamide (HCONH , ≥ The sol was stored under inert atmosphere, where it is stable for atleast 3 months. In air, the lifetime of the sol is shorter, but it is still stable for 1-2 weeks.A second sol was made in the same way, using a 20% excess of lead precursor. This sol wasused for the preparation of bulk PNZT (see below).A separate lead oxide (PbO) sol was fabricated by dissolving 9.76 g of freeze dried leadacetate (Pb(CH COO) , PbAc , 23 mmol, 10 mol% excess, ≥ OH) , EG) while stirring to a final concentration of 1 M. Substrate preparation
The substrates used for deposition of the PNZT sol were prepared from a (001) oriented sili-con wafer without thermal oxide (Ted Pella) diced in 1x1 cm squares. The silicon substrateswere cleaned ultrasonically in acetone, demineralized water and ethanol for ten minutes each.The substrates were subsequently blow dried using compressed air and loaded into a KurtJ. Lesker sputtering system. The substrated were O plasma cleaned (0.15 mbar, 200 W,5 min.) after which a titanium adhesion layer of 5/10 nm was DC sputtered (200 W, 0.2nm/s) without breaking the vacuum. Subsequently, a 100 nm thick electrode of platinumwas DC sputtered (200 W, 1.61 nm/s) onto the adhesion layer. The full electrode stack wasannealed in a box furnace in air (450℃, 90 min., ramp rate 14.2 ℃/s).13 eposition procedure and heat treatment The platinized silicon substrates prepared as described above were again cleaned ultrasoni-cally in acetone, demineralized water and ethanol for 10 minutes each. The substrates wereblow dried using compressed air and UV/O treated in an Ossila UV ozone cleaner to re-move any residual organic contamination from the surface. The substrates were immediatelyplaced in the center of the vacuum chuck of a spin coater. 75 µ L of the 1.5 M sol was de-posited onto the substrates. The spin coater was subsequently ramped up to the desiredspeed at 1000 rpm/s. It was held at this speed for 30 s, then slowed to a stop at 1000rpm/s. The film was then placed on a hotplate at 230℃ for drying. Additional layers weredeposited after drying for the production of multilayer films. After the deposition of up tothree layers, the films were pyrolyzed at 380℃ on the hotplate and annealed by placing themin a preheated box furnace at 650℃ for 10 minutes. Multilayer stacks of up to nine single de-posited layers were produced, for which pyrolysis and annealing steps were performed everythird layer. Multiple annealed layers are required to prevent the formation of leakage pathsthrough the film. A 4x4 grid of circular top electrodes of 100 nm of platinum was sputterdeposited onto the films using a hard mask. Pellet preparation
For the preparation of bulk pellets of PNZT, 10 mL of the sol without an excess of leadwas stirred and heated at 230℃ on a hotplate. Some of the resulting gel was used forthermogravimetric analysis (TGA) and differential thermal analysis (DTA). The remaininggel was heated in a box furnace to 420℃ to remove the organic groups. The resultingpowder was ground using a pestle and mortar and heated again at 450℃ for 30 minutes.The amorphous PNZT powder was pressed into 10 mm pellets under a load 6.4 ton/cm .These pellets were sintered in a box furnace at 800℃ or 1200℃ for two hours.14 haracterization Grain structures of the films and pellets and film thicknesses were studied using an FEI NovaNanoSEM 650 scanning electron microscope. X-ray diffraction data was collected using aPanAnalytical X’Pert Pro MRD or a Bruker D8 Advance diffractometer (both in Bragg-Brentano geometry) for the films and pellets respectively. DTA-TGA data was collected inargon from 200℃ to 1200℃ at a heating rate of 10℃/minute using a TA instruments SDT2960 differential scanning calorimeter. Finally, ferroelectric and piezoelectric properties ofthe films and pellets were measured using a state-of-the-art AixACCT TF analyzer 2000ferroelectric-piezoelectric characterization system with an AixACCT double beam (films)or a Sios single beam (pellets) interferometer. The use of a double beam interferometereliminates the contribution of the bending of the substrate to the measured deformation ofthe film.
Acknowledgement
We gratefully acknowledge the invaluable help of Jacob Baas and Henk Bonder in the lab.M.A. acknowledges financial support of a FOM-f Fellowship of the Dutch Research Council(NWO).
Supporting Information Available
The following files are available free of charge.A listing of the contents of each file supplied as Supporting Information should be in-cluded. For instructions on what should be included in the Supporting Information as well ashow to prepare this material for publications, refer to the journal’s Instructions for Authors.The following files are available free of charge.• Filename: brief description 15 Filename: brief description
References (1) Danks, A. E.; Hall, S. R.; Schnepp, Z. The evolution of ’sol-gel’ chemistry as a techniquefor materials synthesis.
Mater. horizons , , 91–112.(2) Bassiri-Gharb, N.; Bastani, Y.; Bernal, A. Chemical solution growth of ferroelectricoxide thin films and nanostructures. Chem. Soc. Rev. , , 2125–2140.(3) Noheda, B.; Cox, D. E.; Shirane, G.; Gonzalo, J. A.; Cross, L. E.; Park, S.-E.; ., Amonoclinic ferroelectric phase in the Pb(Zr1-xTix)O3 solid solution. Appl. Phys. Lett. , , 2059–2061.(4) Jaffe, B.; Cook, W. R.; Jaffe, H. L. Piezoelectric ceramics ; Academic Press, 1971.(5) Damjanovic, D. Ferroelectric, dielectric and piezoelectric properties of ferroelectric thinfilms and ceramics.
Reports Prog. Phys. , , 1267.(6) Budd, K. D.; Dey, S. Y.; Payne, D. A. Sol-gel processing of PbTiO3, PbZrO3, PZT,and PLZT thin films. Br. Ceram. Soc. Proc. , , 107–121.(7) Niederberger, M. Nonaqueous Sol-Gel Routes to Metal Oxide Nanoparticles. Acc.Chem. Res. , , 793–800.(8) Vioux, A. Nonhydrolytic Sol-Gel Routes to Oxides. Chem. Mater. , 2292–2299.(9) Debecker, D. P.; Mutin, P. H. Non-hydrolytic sol-gel routes to heterogeneous catalysts.
Chem. Soc. Rev. , , 3624–3650.(10) De-Qing, Z.; Hong-Shan, S.; Shao-Jun, W.; Mao-Sheng, C.; Xui-Li, W. Synthesis andmechanism research of an ethylene glycol-based sol-gel method for preparing PZTnanopowders. J. Sol-Gel Sci. Technol. , 157–161.1611) Damjanovic, D.; Demartin, M. Contribution of the irreversible displacement of domainwalls to the piezoelectric effect in barium titanate and lead zirconate titanate ceramicsContribution of the irreversible displacement of domain walls to the piezoelectric effectin barium titanate.
J. Phys. Condens. Matter , , 4943–4953.(12) Livage, C.; Safari, A.; Klein, L. C. Glycol-based sol-gel process for the fabrication offerroelectric PZT thin films. J. Sol-Gel Sci. Technol. , , 605–609.(13) Tu, Y. L.; Milne, S. J. Characterization of single layer PZT (53/47) films prepared froman air-stable sol-gel route. J. Mater. Res. , , 3222–3231.(14) Jaffe, B.; Roth, R. S.; Marzullo, S. Piezoelectric properties of Lead zirconate-Leadtitanate solid-solution ceramics. J. Appl. Phys. , , 809–810.(15) Noheda, B.; Gonzalo, J.; Cross, L.; Guo, R. Tetragonal-to-monoclinic phase transitionin a ferroelectric perovskite: The structure of PbZr(0.52)Ti(0.48)O(3). Phys. Rev. B , , 8687–8695.(16) Noheda, B.; Cox, D. E.; Shirane, G.; Guo, R.; Jones, B.; Cross, L. E. Stability of themonoclinic phase in the ferroelectric perovskite PbZr1-xTixO3. Phys. Rev. B , ,014103.(17) Hinterstein, M.; Rouquette, J.; Haines, J.; Papet, P.; Knapp, M.; Glaum, J.; Fuess, H.Structural Description of the Macroscopic Piezo- and Ferroelectric Properties of LeadZirconate Titanate. Physical Review Letters , , 077602.(18) Xu, T.-B.; Siochi, E. J.; Kang, J. H.; Zuo, L.; Zhou, W.; Tang, X.; Jiang, X. Energyharvesting using a PZT ceramic multilayer stack. Smart Materials and Structures , , 065015.(19) Sharma, P. K.; Ounaies, Z.; Varadan, V. V.; Varadan, V. K. Dielectric and piezoelectric17roperties of microwave sintered PZT. Smart Materials and Structures , , 878–883.(20) Choy, J.-H.; Han, Y.-S.; Kim, S.-J. Oxalate coprecipitation route to the piezoelectricPb(Zr,Ti)O3 oxide. Journal of Materials Chemistry , , 1807–1813.(21) Guiffard, B.; Troccaz, M. Low temperature synthesis of stoichiometric and homogeneouslead zirconate titanate powder by oxalate and hydroxide coprecipitation. MaterialsResearch Bulletin , , 1759–1768.(22) Shannigrahi, S.; Tay, F.; Yao, K.; Choudhary, R. Effect of rare earth (La, Nd, Sm, Eu,Gd, Dy, Er and Yb) ion substitutions on the microstructural and electrical propertiesof sol-gel grown PZT ceramics. Journal of the European Ceramic Society , ,163–170.(23) Sahoo, B.; Panda, P. K. Effect of lanthanum, neodymium on piezoelectric, dielectricand ferroelectric properties of PZT. Journal of Advanced Ceramics , , 37–41.(24) Dong, D.; Murakami, K.; Kaneko, S.; Xiong, M. Piezoelectric Properties of PZT Ce-ramics Sintered at Low Temperature with Complex-Oxide Additives. Journal of theCeramic Society of Japan , , 1090–1094.(25) Singh, V.; Kumar, H.; Kharat, D.; Hait, S.; Kulkarni, M. Effect of Lanthanum substi-tution on ferroelectric properties of Niobium doped PZT ceramics. Materials Letters , , 2964–2968.(26) Zheng, H.; Reaney, I.; Lee, W.; Jones, N.; Thomas, H. Effects of strontium substitutionin Nb-doped PZT ceramics. Journal of the European Ceramic Society , , 1371–1375.(27) Garcia, J.; Pérez, R.; Albareda, A.; Eiras, J. Non-linear dielectric and piezoelectric18esponse in undoped and Nb5+ or Fe3+ doped PZT ceramic system. Journal of theEuropean Ceramic Society , , 4029–4032.(28) Brennecka, G. L.; Ihlefeld, J. F.; Maria, J.-P.; Tuttle, B. A.; Clem, P. G. Process-ing Technologies for High-Permittivity Thin Films in Capacitor Applications. J. Am.Ceram. Soc. , , 3935–3954.(29) Balma, D.; Mazzalai, A.; Chidambaram, N.; Sandu, C. S.; Neels, A.; Dommann, A.;Hess, P.; Binz, D.; Muralt, P. High piezoelectric longitudinal coefficients in sol-gel PZTthin film multilayers. J. Am. Ceram. Soc. , , 2069–2075.(30) Taylor, D.; Damjanovic, D. Piezoelectric properties of rhombohedral Pb(Zr, Ti)O3 thinfilms with (100), (111), and âĂIJrandomâĂİ crystallographic orientation. Appl. Phys.Lett. , , 1615–1617.(31) Lefki, K.; Dormans, G. J. M. Measurement of piezoelectric coefficients of ferroelectricthin films. Journal of Applied Physics , , 1764–1767.(32) Ledermann, N.; Muralt, P.; Baborowski, J.; Gentil, S.; Mukati, K.; Cantoni, M.;Seifert, A.; Setter, N. {1 0 0}-textured, piezoelectric Pb(Zrx, Ti1-x)O3thin films forMEMS: Integration, deposition and properties. Sensors Actuators, A Phys. , ,162–170.(33) Zavala, G.; Fendler, J. H.; Trolier-McKinstry, S. Characterization of ferroelectric leadzirconate titanate films by scanning force microscopy. Journal of Applied Physics , , 7480–7491.(34) Chen, H. D.; Udayakumar, K. R.; Gaskey, C. J.; Cross, L. E.; Bernstein, J. J.;Niles, L. C. Fabrication and Electrical Properties of Lead Zirconate Titanate ThickFilms. Journal of the American Ceramic Society , , 2189–2192.1935) Wang, Y.; Ganpule, C.; Liu, B. T.; Li, H.; Mori, K.; Hill, B.; Wuttig, M.; Ramesh, R.;Finder, J.; Yu, Z.; Droopad, R.; Eisenbeiser, K. Epitaxial ferroelectric Pb(Zr, Ti)O3thin films on Si using SrTiO3 template layers. Applied Physics Letters , , 97–99.(36) Lian, L.; Sottos, N. R. Effects of thickness on the piezoelectric and dielectric propertiesof lead zirconate titanate thin films. Journal of Applied Physics , , 3941–3949.(37) Nguyen, M. D.; Trinh, T. Q.; Dekkers, M.; Houwman, E. P.; Vu, H. N.; Rijnders, G.Effect of dopants on ferroelectric and piezoelectric properties of lead zirconate titanatethin films on Si substrates. Ceram. Int. , , 1013–1018.(38) Goh, W.; Yao, K.; Ong, C. Effects of microstructure on the properties of ferroelectriclead zirconate titanate (PZT) thin films. Applied Physics A , , 1089–1093.(39) Klissurska, R. D.; Tagantsev, A. K.; Brooks, K. G.; Setter, N. Use of FerroelectricHysteresis Parameters for Evaluation of Niobium Effects in Lead Zirconate TitanateThin Films. J. Am. Ceram. Soc. , , 336–342.(40) Hench, L. L.; West, J. K. The Sol-Gel Process. Chem. Rev. , , 33–72.20igure 6: (a) Plan-view and (b) cross-section SEM images of a nine-layer PNZT thin filmstack with a PbO overcoat. The total thickness of the film is 440 nm.21igure 7: XRD pattern of the nine-layer thin film stack.22igure 8: (a) Ferroelectric and strain loop of the nine-layer PNZT thin film stack. Thelongitudinal piezoelectric coefficient is obtained from the slope of the strain loop at zeroelectric field, as indicated by the red tangent line. (b) Fatigue response of the film up to7