Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Hydrogen‐induced damage of lead‐zirconate‐titanate (PZT) Mohammadabadi, Ali Shafiei 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2013_spring_shafieimohammadabadi_ali.pdf [ 6.99MB ]
Metadata
JSON: 24-1.0073678.json
JSON-LD: 24-1.0073678-ld.json
RDF/XML (Pretty): 24-1.0073678-rdf.xml
RDF/JSON: 24-1.0073678-rdf.json
Turtle: 24-1.0073678-turtle.txt
N-Triples: 24-1.0073678-rdf-ntriples.txt
Original Record: 24-1.0073678-source.json
Full Text
24-1.0073678-fulltext.txt
Citation
24-1.0073678.ris

Full Text

Hydrogen‐Induced Damage of Lead‐Zirconate‐Titanate (PZT)  by ALI SHAFIEI MOHAMMADABADI  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY  in THE FACULTY OF GRADUATE STUDIES (Materials Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2013 © Ali Shafiei Mohammadabadi, 2013  Abstract Lead-Zirconate-Titanate Pb(Zr,Ti)O3 (PZT) based actuators are evaluated by automotive industry for advanced fuel-injection systems, including hydrogen injection. However, hydrogen can have deleterious effect on the PZT's functionality and properties. The general objective of this work is to study the interactions between PZT and hydrogen. The results of long-term (200-1200 hours) high-pressure (10 MPa) hydrogen exposure on the PZT microstructure show that hydrogen has only superficial effects on the microstructure of bare PZT. However, when an electrode is attached to PZT, the hydrogen damage increased; a porous layer developed immediately adjacent to the electrodes on the PZT surface due to hydrogen spillover. The kinetics of the PZT structural modifications due to hydrogen was investigated by online monitoring of the electrical properties of PZT above the Curie temperature, up to 650C. The results show that the structural changes can be described by the classical nucleation and growth theory. The growth of the new structure appears to be limited by the diffusion of protons into PZT, with a calculated activation energy of 0.440.09 eV, at 450-650C. Two relaxation peaks were observed in the dissipation factor curves of the hydrogen-treated PZT. While the kinetics of one of the relaxation peaks obeys the classical Arrhenius law with the activation energy of 0.66 eV, the other peak shows an unusual relaxation kinetic. The mechanisms for the formation of these relaxation peaks are determined. Low temperature (20C) diffusion of hydrogen into the PZT was also studied, using the water electrolysis technique. Based on the microstructural observations, the  ii  diffusion coefficient of hydrogen in PZT was calculated as 9×10-11 cm2/sec. The MaxwellWagner polarization mechanism is determined to be responsible for the changes in the hydrogen-affected PZT capacitance. In the last part of the project, alumina coatings were applied to PZT plates using the sol-gel technique, to explore the possibilities of decreasing H2 damage to PZT. The functionality of the coating against hydrogen damage was evaluated by the water electrolysis technique. Significant decrease of hydrogen damage was observed even for highly porous coatings. The mechanisms by which the alumina coating decreases the hydrogen damage were tentatively proposed.  iii  Preface This research work was conducted as part of a NSERC Strategic Project Grant awarded to The University of British Columbia, Simon Fraser University and Westport Innovations Inc. of Vancouver. The journal papers listed below have been prepared from the work presented in the dissertation. I am the primary contributor to all of them, and the coauthors contributions are as follows: my supervisor, Dr. Tom Troczynski, extensively commented on the experimental and analysis methods, and the results interpretation in all three papers. C. Oprea edited all three papers, and provided the SEM pictures in all three papers. T. Nickchi contributed in the second paper by commenting on the experimental setup. Dr. A. Alfantazi contributed in the second and third papers by providing and commenting on the experimental setup. 1- A. Shafiei, C. Oprea, T. Troczynski, Investigation of the effects of high-pressure hydrogen on Pb(Zr,Ti)O3 (PZT) ceramics, Journal of the American Ceramic Society, 2012, 95(2), 782– 787 2- A. Shafiei, T. Nickchi, C. Oprea, A. Alfantazi, T. Troczynski, Investigation of hydrogen effects on the properties of Pb(Zr,Ti)O3 in tetragonal phase using water electrolysis technique, Applied Physics Letters, 2011, 99 (21), 212903-212906 3- A. Shafiei, C. Oprea, A. Alfantazi, and T. Troczynski, In situ monitoring of the effects of hydrogen on Pb(Zr,Ti)O3 structure, Journal of Applied Physics, 2011, 109 (11), 114108114116  Chapter 5-1 is based on paper “1”. Chapter 5-2 is based on paper “3”. Chapter 5-3 is based on paper “2”. Please check the first pages of these chapters to see footnotes with similar information.  iv  Table of Contents  Table of Contents Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iv Table of Contents ........................................................................................................................... v List of Tables ................................................................................................................................ vii List of Figures .............................................................................................................................. viii Nomenclature .............................................................................................................................. xiii Acknowledgements .................................................................................................................... xvi Dedication .................................................................................................................................. xvii 1 Introduction ................................................................................................................................ 1 2 Literature Review ....................................................................................................................... 4 2.1 Lead Zirconate Titanate (Pb(Zr,Ti)O3) ............................................................................... 4 2.2 Hydrogen damage of PZT .................................................................................................... 9 2.3 Mechanisms of hydrogen damage of PZT ........................................................................ 12 2.3.1 Hydrogen incorporation into PZT ............................................................................. 12 2.3.2 Stable forms of hydrogen in PZT................................................................................ 22 2.3.3 Stable sites of H+ in PZT .............................................................................................. 24 2.3.4 Effect of hydrogen on PZT ......................................................................................... 26 2.4 Dielectric spectroscopy of PZT ......................................................................................... 31 2.5 High pressure hydrogen compatibility of PZT ................................................................. 33 2.6 Methods of decreasing the hydrogen damage to PZT...................................................... 35 3 Scope and Objectives ................................................................................................................ 37 4 Materials and Methods ............................................................................................................. 40 4.1 Samples ............................................................................................................................... 40 4.2 Gas hydrogen treatment .................................................................................................... 42 4.3 Water electrolysis treatment of PZT ................................................................................ 47 4.4 Alumina sol-gel coating ..................................................................................................... 51 4.5 Characterization techniques .............................................................................................. 55 4.5.1 X-ray diffraction analysis (XRD) ................................................................................ 55 4.5.2 Scanning electron microscopy (SEM) ........................................................................ 55 4.5.3 Electrical properties measurements............................................................................ 55 5 Results and Discussion .............................................................................................................. 57 v  5.1 High-Pressure conditions (T=100C, p=10 MPa, t=200-1200 hours)* ............................. 57 5.1.1 H2 effects on PZT microstructure ............................................................................... 57 5.1.2 H2 effects on the electrical properties of PZT ............................................................ 68 5.2 High-Temperature conditions (T=450-600C, p=0.013 MPa) * ....................................... 72 5.2.1 H2 effects on PZT microstructure ............................................................................... 72 5.2.2 H2 effects on PZT electrical properties....................................................................... 75 5.3 Water-electrolysis treatment of PZT* ............................................................................... 96 5.3.1 Microstructure ............................................................................................................. 96 5.3.2 Electrical properties of PZT exposed to water electrolysis ..................................... 102 5.4 Ceramic Coatings for PZT Damage Protection .............................................................. 112 5.4.1 Alumina coatings microstructure ............................................................................. 112 5.4.2 Hydrogen resistivity of alumina-coated PZT .......................................................... 120 6 Conclusions ............................................................................................................................. 129 7 Future Work............................................................................................................................ 133 References .................................................................................................................................. 136  vi  List of Tables Table 1- The heat and entropy of dissolution of hydrogen in different metals [22] ............... 17 Table 2- The EDX analysis for the bright particles in Figure 44b ............................................ 73 Table 3- Different values of exponents for the equation (28) [67]............................................ 80 Table 4- The fitting values obtained for the equation (30) ....................................................... 82 Table 5- The fitting values obtained for the HN equation ........................................................ 91 Table 6- Microstructural charachteristics of alumina coatings a a function of TC [91]....... 117 Table 7- The EDX analysis for the -alumina coating (high concentration of Au is due to the gold coating on the sample for SEM analysis) .......................................................................... 120  vii  List of Figures Figure 1- A typical polarization versus electric field (P-E) hysteresis for ferroelectrical materials ......................................................................................................................................... 2 Figure 2- A schematic of an electronic fuel injector based on PZT actuators [3] ...................... 3 Figure 3- Binary phase diagram of PbZrO3-PbTiO3 system (refer to text for explanation of the symbols) [5] .................................................................................................................................... 5 Figure 4- Unit cell for PZT (a) above and (b) below Curie temperature .................................... 6 Figure 5- TEM image showing domains in PZT [8] and a schematic of the 180 (1-1) and 90 (2-2) domain walls in PZT............................................................................................................. 8 Figure 6- A ferroelectric ceramic with differently oriented domains inside each grain (a) before and (b) after polarization, with remnant strain ................................................................ 9 Figure 7- Initiation and growth of hydrogen fissures during charging at 50 mA/cm2 for 2h (b) and 4h (c); the black bar is 50 µm [12] ....................................................................................... 12 Figure 8- Different paths for the incorporation of hydrogen into PZT: (1) through the surface of bare PZT (only at high temperatures) and (2) through the surface of electrodes (at low temperatures). .............................................................................................................................. 14 Figure 9- Switching charge as a function of H2 annealing temperature with and without upper Pt electrode [17] ................................................................................................................ 14 Figure 10- Remnant polarization as a function of the H2 annealing temperature for capacitors with Pt, Pd, Au or Ag electrodes (applied voltage of 5V) [19] .................................................. 15 Figure 11- Solubility (cc of hydrogen per 100 g of metal) of hydrogen inside metals at 1 atmosphere pressure of hydrogen [24] ....................................................................................... 18 Figure 12- Schematic cross-section of the electrode/PZT assembly in H2 gas (it is assumed that hydrogen diffuses only in the z direction).......................................................................... 19 Figure 13- Hydrogen diffusion coefficient in different metals [25].......................................... 21 Figure 14- The stable positions of protons according to the reference [32] in a) perovskites with large lattice constants, and b) perovskites with short lattice constants ........................... 25 Figure 15- PZT crystal structure showing the possible location of protons in the lattice of PZT [31]; b) stable lattice site of protons for tetragonal PbTiO3 and c) cubic phase of PbTiO3 supposed by Park and Chadi [30] ................................................................................................ 26 Figure 16- Proposed PZT damage mechanisms can be categorized by the place where damage occurs ............................................................................................................................................ 27 Figure 17- The microstructure of bare PZT plates .................................................................... 40 Figure 18- The microstructure of the PZT samples ................................................................... 41 Figure 19- Micrographs of the surface of the electrodes: silver (a) and silver-palladium (b) . 42 Figure 20- The time-temperature schedule of the ‘High-Pressure’ hydrogen treatment used in this work .................................................................................................................................. 43 Figure 21- Schematic of an actuator made from PZT plates stacked together, b) the equivalent electrical circuit of an actuator ................................................................................. 43 viii  Figure 22- The schematic of the setup used in this work for online monitoring the electrical properties ...................................................................................................................................... 44 Figure 23- The ‘High-Temperature’ hydrogen treatment used in this work ........................... 45 Figure 24- A typical Nyquist plot for the PZT at 500°C in air (the frequency is swiped between 106 and 10-3 Hz) ............................................................................................................. 47 Figure 25- Schematic of the setup for the water electrolysis experiment ................................ 49 Figure 26- Schematic of the steps used for the preparation of the alumina coating................ 52 Figure 27- Micrographs of PZT surface: a) the as-received sample; b) after 1200 h hydrogen treatment ...................................................................................................................................... 58 Figure 28- Surfaces of PZT plates at higher magnifications: (a) before and (b) after 1200 h hydrogen treatment ..................................................................................................................... 58 Figure 29- XRD results of bare PZT for as-received and after 1200 h hydrogen treatment ... 59 Figure 30- Cross-section of the as-received sample (a) in comparison to the cross section of the sample after 1200 h hydrogen treatment (b) ....................................................................... 60 Figure 31- Cross section of the hydrogen treated sample for 1200 hours, close to the surface ....................................................................................................................................................... 60 Figure 32- Low magnification (a) and high magnification (b) images of damaged layer on the PZT surface next to the Ag electrode after 400 hours hydrogen-treatment ............................ 61 Figure 33- The interface of the Ag electrode with PZT after grinding and polishing, for the as-received sample (a) and for the sample hydrogen-treated for 400 hours (b); no detachment of the electrode from the PZT and no damaged layer are visible ............................................. 63 Figure 34- The spillover mechanism of hydrogen atoms from the surface of the Ag electrode to the surface of the PZT ............................................................................................................. 63 Figure 35- The detachment of the Ag electrode from the PZT for the sample treated for 600 h ....................................................................................................................................................... 64 Figure 36- Detachment of the Ag electrode from PZT; some cracks are present on the surface of the electrode for the sample heat-treated for 1200 hours ..................................................... 64 Figure 37- Micrographs of the sample with Ag/Pd electrodes after hydrogen-treated for 200h: the 1x10 mm side face (a) and its cross-section (b) .................................................................... 65 Figure 38- Surface of the side face of the sample with Ag/Pd electrode: (a) as-received, (b) hydrogen-treated for 400 h; noticeable corroded area next to the electrode (c) ..................... 67 Figure 39- Micrograph of the cross-section of the sample shown in Figure 38 ....................... 68 Figure 40- Capacitance of PZT sample with Ag/Pd electrode in high-pressure hydrogen atmosphere (a); at point ‘1’ the heater is on, and at point ‘3’ the heater is off. (b): capacitance variation with temperature in hydrogen atmosphere ............................................................... 69 Figure 41- Capacitance of PZT sample with Ag electrode in high-pressure argon atmosphere ....................................................................................................................................................... 70 Figure 42- Capacitance of PZT sample with Ag electrode in high-pressure hydrogen atmosphere ................................................................................................................................... 71 Figure 43- Image from the side surface of the PZT plate with Ag electrode for (a) as-received and (b) after hydrogen treatment (for 2 h / 400C / p= 0.013 MPa) ......................................... 73 ix  Figure 44- Metallic lead in hydrogen treated PZT samples (for 2 h / 600C / p= 0.013 MPa) 73 Figure 45- The XRD pattern for as-received and hydrogen treated PZT with Ag electrodes 74 Figure 46- The XRD pattern for hydrogen treated PZT with Ag electrodes at 500C and 600C ............................................................................................................................................ 75 Figure 47- (a) The general trend of PZT capacitance variation with time in hydrogen atmosphere at 500°C; (b) measurements for the real part of the impedance (ZRe) and the calculated values of R according to equation (8) (the data is obtained at the constant frequency of 1 kHz) ..................................................................................................................... 76 Figure 48- The ZRe -ZIm plot for PZT plate heat treated at 550C and the resistance determined using the ZRe -ZIm plots for PZT; the noise in the ZRe-ZIm plots corresponds to the times when the heater was on. ................................................................................................... 77 Figure 49- The variation of PZT capacitance at 530°C, 550°C and 600°C ................................ 78 Figure 50- The general trend for the isothermal α - time plots having different time steps, time equal to zero shows the start of the reaction [68] ............................................................. 79 Figure 51- The results of fitting the capacitance data to equation (30) for the temperatures of 550C (a) and 600C (b) ............................................................................................................... 81 Figure 52- (a) The results of fitting the capacitance data to equation (30); (b) the activation energy of hydrogen diffusion, obtained from the fit ................................................................. 82 Figure 53- The Grotthuss mechanism for diffusion of protons in PZT, including the reorientation and hopping of protons between oxygen onions ................................................ 83 Figure 54- Hypothetical schematic of the different modes which can be assumed for the dissolution of hydrogen in PZT; (a) where the diffusion of protons into PZT occur uniformly from the surface; in this case the diffusion equation with proper initial and boundary equation could be used for determining the total amount of protons in PZT; (b) where the diffusion of protons can occur from limited places in the PZT; in this case the nucleation and growth models can be used to describe the total amount of protons in PZT........................... 84 Figure 55- Changes of capacitance C and dissipation factor DF of hydrogen-treated sample (for 24 hrs / 550C / p= 0.013 MPa) versus temperature (The thick grey line shows the changes of capacitance for as-received sample) ......................................................................... 86 Figure 56- Schematics of the dipolar polarization mechanisms, wherein direction of the dipoles changes with changing the direction of applied voltage .............................................. 87 Figure 57- Schematics of the Maxwell-Wagner polarization mechanism, wherein differences in the electrical properties of different regions cause charge accumulation at the interfaces between the different regions, leading to the increase of capacitance ..................................... 88 Figure 58- Variations of ε’ and ε’’ for hydrogen-treated samples with the frequency of applied voltage in the temperature range of 200-325C, with 25C increments .................................. 89 Figure 59- The results of fitting the ε´ and ε’’ data to the Debye equation for at T= 325C, for PZT hydrogen-treated samples (for 24 hrs / 550C / p= 0.013 MPa) ........................................ 90 Figure 60- The results of fitting the ε´ and ε’’ data to the Havriliak–Negami equation for at T= 325C, for PZT hydrogen-treated samples (for 24 hrs / 550C / p= 0.013 MPa) ................ 91 x  Figure 61- The activation energy for the ion jumping, obtained from the fits, for PZT hydrogen-treated samples (for 24 hrs / 550C / p= 0.013 MPa) ................................................ 92 Figure 62- Variation of DF with frequency in the temperature range 22-42C, for PZT hydrogen-treated samples (for 24 hrs / 550C / p= 0.013 MPa) ................................................ 94 Figure 63- Micrographs of the cross-section through PZT plate after water electrolysis: a) low-magnification image (after 48 hours water electrolysis); b) microstructure of the corroded layer (close to the electrode); c) microstructure in a region far from the corroded layer .............................................................................................................................................. 97 Figure 64- The microstructure of PZT after water electrolysis just beneath the electrode (after removing the electrode) .................................................................................................... 99 Figure 65- XRD pattern of the as-received PZT sample versus the water electrolyzed PZT sample, using the following parameters: I=100 mA/cm2, t=48 hours ..................................... 100 Figure 66- The thickness of the corroded layer versus the square root of time of water electrolysis .................................................................................................................................. 102 Figure 67- The changes of capacitance (C) and dissipation factor (DF) versus the duration of water electrolysis at the frequency of 1 kHz............................................................................ 103 Figure 68- Variations of electrical properties after 6 hours water electrolysis and subsequent aging in air: a) capacitance (C); b) dissipation factor (DF) (: as-received,: after water electrolysis, : after aging) ........................................................................................................ 104 Figure 69- Variations of electrical properties after 10 hours water electrolysis and subsequent aging in air: a) capacitance (C); b) dissipation factor (DF) (: as-received,: after water electrolysis, : after aging) ........................................................................................................ 104 Figure 70- Variations of capacitance (C) and dissipation factor (DF) after water electrolysis for 48 hrs, and subsequent aging at room temperature in air (■: as-received, : after water electrolysis, ▲:after 10 hours aging, : after 24 hours aging) ................................................ 105 Figure 71- The results of fitting the ε´´ (ε´´=DFε´) data to Debye and Havriliak–Negami equation. An iterative MATLAB code was developed and used for the fitting procedure. .. 109 Figure 72- The changes of the capacitance (C) and dissipation factor (DF) for (a) a leaky capacitor with electronic conduction, (b) for a capacitor with hopping charge carriers adapted from [38] ....................................................................................................................... 109 Figure 73- Low magnification image of the coating on the surface of PZT (a) after dip coating with pure boehmite sol (b) before and (c) after heat treatment of the coating in the furnace, in some places on the surface of the coating, detachment of the coating was observed ....... 113 Figure 74- Low magnification image of the coating on the surface of PZT after dip coating before the heat treatment of the coating in the furnace (comparing with Figure 73a, a smooth uniform coating has formed on the surface of PZT with the addition of PVA to sol) .......... 114 Figure 75- Low magnification (a) and high magnification (b) images of the coating on the surface of PZT after dip coating and after heat treatment of the coating in the furnace (comparing with Figure 73b and c, a smooth uniform coating has formed on the surface of PZT with the addition of PVA to sol) ...................................................................................... 115 xi  Figure 76- Low magnification (a) and high magnification (b) images of the cross section of the alumina coating. As it can be seen from (b), the coating had enough fluidity to fill out the pores on the surface of PZT ...................................................................................................... 116 Figure 77- High resolution image of the cross section through the alumina coating processed at 450C in air for 5 hours ......................................................................................................... 118 Figure 78- Transformation sequence of the different aluminum hydroxides with temperature (adapted from [93]). ................................................................................................................... 119 Figure 79- XRD results for the as-received boehmite powder and after heat treatment at 450C for 5 hr ............................................................................................................................. 119 Figure 80- The cross section of the sample with Au-Pd electrodes and after 24 hours water electrolysis. The thickness of the corroded layer is about 100 microns ................................. 121 Figure 81- The cross section of the sample with Au-Pd electrodes and alumina coating and after 48 hours water electrolysis ............................................................................................... 122 Figure 82- Schematic for the reaction of hydrogen atoms with -alumina particles 1) transformation of hydrogen atoms to hydrogen molecules which leave the system away from the coating (i.e. as hydrogen bubbles during water electrolysis), 2) diffusion of hydrogen atoms through the electrode and attachment to -alumina particles, followed by surface and bulk diffusion through -alumina towards PZT ...................................................................... 123 Figure 83- An image of the cross section of PZT with alumina coating on top..................... 124 Figure 84- The cross section of the sample with Au-Pd electrodes and thin alumina coating and after 144 hours water electrolysis ...................................................................................... 125 Figure 85- Schematic image for the combination of hydrogen atoms at the interface of metallic electrode with - alumina ........................................................................................... 125 Figure 86-Equivalent electrical circuit for PZT and PZT with coatings ................................ 128  xii  Nomenclature Latin Symbols A a C C0 D d DF DPb DM DPZT E Ec f fMax g G h hH ΔH ΔHM i ic JH k KIC KIH m n nis P p P0 p0 PH2 Pr Ps Q R SH ΔSM  Constant Area of electrodes Hydrogen concentration Capacitance Concentration of hydrogen Diffusion coefficient Distance between electrodes Dissipation factor Lead diffusion coefficient in PbTiO3 Diffusion coefficient of hydrogen in metallic electrode Diffusion coefficient of hydrogen in PZT Electric Field Coercive Field Frequency Frequency at which the maximum occurs fugacity of H2 Gravity Conductance Coating thickness Partial enthalpy Activation Energy Heat of dissolution of hydrogen inside metal Imaginary number Charging current density Flux of hydrogen atoms Coefficient in equation (10) Fracture toughness of the un-affected PZT Fracture toughness of hydrogen-treated PZT Constant Constant Number of interstitial sites per metal atom Polarization Pressure Constant Pressure at standard conditions Partial pressure of hydrogen Remnant Polarization Spontaneous Polarization Activation energy Ohmic resistance Non-configurational part of entropy Dissolution of hydrogen in different metals  xiii  t T Tc Y x Z ZIm ZRe  Time Temperature Curie temperature Admittance Thickness of corroded layer Impedance Imaginary part of impedance Real part of impedance  Greek Symbols θ α LV β ε ε0 τ 0 ε´ ε´´ εs ε    ( ) ( )  ( )  ( )   Constant Fraction of the volume converted to the product of reaction Sol-vapor surface energy Constant Dielectric constant Vacuum permittivity Relaxation time Constant Real part of dielectric constant Imaginary part of dielectric constant Dielectric constant when  0 Dielectric constant when   Viscosity of the sol Density of the sol Withdraw speed Chemical potential of gaseous hydrogen per molecule Chemical potential of hydrogen atom dissolved in the metallic electrode per atom Chemical potential at a given standard state Chemical potential at a given standard state Angular frequency  Abbreviation AO EDX FCC FM FR FT KTN MPB MW PVA PZT  Orthorhombic phase Energy-dispersive X-ray spectroscopy Face-centered cubic lattice Monoclinic phase Rhombohedral phase Tetragonal phase Potassium tantalate niobate Morphotropic phase boundary Maxwell-Wagner polarization mechanism Polyvinyl alcohol Lead zirconate titanate  xiv  Abbreviation (continue) Pc SEM TEM TC XRD  Cubic perovskite structure Scanning electron microscope Transmission electron microscope Curie temperature X-ray diffraction  xv  Acknowledgements  I would like to take this opportunity to express my utmost gratitude toward my supervisor, Dr. Tom Troczynski for his continuous trust, patience, and guidance throughout the course of this work. I also acknowledge Prof. Akram Alfantazi, Prof. Guangrui Xia, and Prof. Steve Cockcroft for their valuable comments. My sincere thanks go to Carmen Oprea, who has been a valuable and amazing colleague and friend, and has been continuously supporting and helping me throughout the years I have been at UBC. She taught me to write “in comparison to” instead of “in compare to”. She taught me to write “reach” instead of “reach to”. Thank you so much Carmen! I would like to thank all staff members in the Department of Materials Engineering at The University of British Columbia for their assistance with my research work. My special thanks to all colleagues and officemates for providing a friendly environment that I was always pleased to work in. Natural Sciences and Engineering Research Council of Canada (NSERC) and Westport are greatly acknowledged for financial support. Special thanks are owed to my parents, who have supported me throughout my years of education.  xvi  Dedication  To my patient parents  xvii  1 Introduction Lead Zirconate Titanate (Pb(Zr,Ti)O3) or PZT is the general name for the perovskite solid solutions between PbZrO3 and PbTiO3. PZT is well known because of its unique electrical properties such as high dielectric, piezoelectric and electro-optic coefficients. The main reason for having such interesting electrical properties is the unique crystal structure and the arrangement of ions inside unit-cell of PZT. In the unit-cell of PZT, the titanium or zirconium ions reside off-center in the octahedral interstitial positions surrounded by six oxygen ions. Therefore, the center of negative charge of oxygen ions will not coincide with the center of positive charge of Ti or Zr ions and this arrangement results in permanent dipoles inside the unit-cell of PZT. These built-in permanent dipoles in the unit-cell of the crystal structure of PZT are the origin of the superior electrical properties of PZT. Application of an electric field of sufficient magnitude will cause these built-in dipoles inside the PZT to switch to a different, stable direction in accordance to the direction of the applied electric field. Moreover, by removal of the electric field, the dipoles will not return to their original direction. This brings up one of the very interesting properties of PZT, which is the switchable Polarization (P)-Electric Field (E) hysteresis, as schematically shown in Figure 1. This property of PZT is generally known as ferroelectricity, and it has also been seen in other oxides like BaTiO3. This property of PZT has enabled the extensive use of these materials in applications such as ferroelectric random access memories (FeRAM). It is known that hydrogen treated PZT may not show this hysteresis anymore. Although the 1  main reason for this phenomena is not known very well, this has been attributed to the formation of [OH]– dipoles, which inhibits the switching of the spontaneous dipoles in PZT. This effect of hydrogen on the polarization hysteresis of PZT has been the main reason for investigating the effect of hydrogen on PZT since 1995 [1-2].  P (C/cm2)  Ps Pr  Ec  Ec E(V/m) Ps :spontaneous polarization Pr :remnant polarization Ec :coercive field  Figure 1- A typical polarization versus electric field (P-E) hysteresis for ferroelectrical materials  In addition to the ferroelectrical properties, a “poled” PZT ceramic (i.e. PZT with oriented dipoles) can also show superior piezoelectric properties, which have caused the extensive use of PZT in other applications, such as actuators and sensors. Recently, attention has been paid to the effect of hydrogen on the piezoelectric properties of PZT as well, as it has been suggested that hydrogen might also have deleterious effects on these properties [3]. For example, Figure 2 schematically shows a modern electronic fuel injector that uses PZT  2  actuators for valve opening, instead of the conventional solenoid technology. These fuel injectors have been introduced by the leading engine manufacturers in recent years. One of the issues of using such fuel injectors in a hydrogen atmosphere is the possible deleterious effects which hydrogen may have on the functionality of the PZT actuators [3].  Figure 2- A schematic of an electronic fuel injector based on PZT actuators [3]  The above points provide rationale for the investigation of the interactions of hydrogen with PZT, which is important topic both from practical and scientific points of view. The objective of the present work is to address some of the issues regarding the interaction of hydrogen and PZT. More specifically, we focused on the kinetics of degradation of PZT properties by hydrogen. Attempts have also been made to propose techniques to inhibit or decrease the hydrogen damage to PZT. In this regard, the sol-gel technique was used to develop hydrogen barrier coatings on the surface of PZT.  3  2 Literature Review 2.1 Lead Zirconate Titanate (Pb(Zr,Ti)O3) PZT is the general name for the perovskite solid solutions between PbZrO3 and PbTiO3, as shown in the binary phase diagram in Figure 3. Starting at PbTiO3 part of the diagram with a ferroelectric tetragonal (FT) phase, by increasing the amount of PbZrO3 in the solution, the composition Pb(Zr0.53Ti0.47)O3 is reached, where the ferroelectric tetragonal phase starts to transforms into another ferroelectric phase, however, with a different crystal structure (rhombohedral phase (FR)) [4]. The composition where the tetragonal phase (FT) transforms to the rhombohedral phase (FR) is considered to be the morphotropic phase boundary (MPB) in the PZT binary phase diagram (Figure 3). Instead of a sharp boundary, the MPB is often observed in real systems as a region of phase coexistence whose width depends on the compositional homogeneity and on the sample processing conditions [5]. Therefore, the location of this boundary has not been determined exactly, as its positions changed from report to report. However, the recent work by Noheda et al. [5] has changed our understanding of the MPB in PZT. Their studies show that instead of a boundary, a lowsymmetry monoclinic phase (FM) exists between the tetragonal and rhombohedral phases and the superior electrical properties of PZT around MPB are actually due to the very high polarization in this phase (Figure 3). In other words, the monoclinic phase acts as a bridge for the phase transformation from tetragonal to rombohedral and vice versa.  4  The rhombohedral part of the diagram itself consists of two other phases: a hightemperature phase (FR(HT)) and a low-temperature ferroelectric rhombohedral (FR(LT)) phase. The PZT structure close to the PbZrO3 part of the diagram up to about 10 %mole of PbTiO3 (Figure 3) has an antiferroelectric, orthorhombic (AO) structure. This phase of PZT has a very complex structure, in the sense that the displacement of cations along the [110] direction is coupled with octahedral tilts [4]. According to Figure 3, we can see that depending on its composition, PZT can have different crystal structures, and consequently PZT can show different piezoelectric, pyroelectric and electro-optic coefficients. The great technological and commercial importance of PZT is actually due to such variations in the electrical properties, which can be obtained by changing the PZT composition. Especially when the PZT is doped with other secondary ions, it can show even more interesting electrical properties [4]. The most widely used PZT ceramics today have the compositions near the MPB composition. For example, Pb(Zr0.53Ti0.47)O3 is the composition for the PZT ceramics used in this work.  Temperature (°C)  500  PC  400  MPB 300  FT  FR(HT) 200  100  AO  FM  FR(LT) 0  PbZrO3  20  40  60  Mole % PbTiO3  80  100  PbTiO3  Figure 3- Binary phase diagram of PbZrO3-PbTiO3 system (refer to text for explanation of the symbols) [5]  5  The cubic perovskite structure of PZT (Pc) above Curie temperature is shown in Figure 4a, wherein each lead ion is surrounded by 12 oxygen ions. The oxygen ions plus the lead ions form a face-centered cubic (FCC) lattice. The titanium or zirconium ions reside in the octahedral interstitial positions surrounded by six oxygen ions. Temperature has a strong effect on the cubic structure shown in Figure 4a. When the temperature decreases to about 375C, the structure changes to the tetragonal shown in Figure 4b. The octahedral site is now distorted, with the Ti or Zr in off-center positions, resulting in a permanent dipole. This "built-in" permanent dipole in the unit cell of the crystal structure of PZT is the origin of the superior dielectrical and piezoelectrical properties of PZT. This is sometimes called a  spontaneous polarization. The temperature of transformation from the cubic to the tetragonal phase is called the Curie temperature (Tc).  Pb Dipole direction  O Ti or Zr Above TC (a)  Below TC (b)  Figure 4- Unit cell for PZT (a) above and (b) below Curie temperature  6  During the transformation from the cubic to the tetragonal phase, permanent dipoles form inside the PZT crystal; however, the direction of these dipoles is not the same throughout the whole crystal or inside each PZT grain. The regions inside PZT crystals where the dipoles are aligned in the same direction are called ferroelectric domains. As soon as spontaneous dipoles start to form during the phase transformation from the cubic to the tetragonal, surface charges start to appear on the surface of PZT. Such surface charges produce very high electric fields in the order of MV m−1 [7]. Therefore, the electrostatic energy of the system increases due to the existence of such electric fields. The electrostatic energy associated with these electric fields can be minimized if the PZT crystal can split into separate ferroelectric domains, and this is one of the reasons why ferroelectric domains form inside the PZT crystals [7]. Another reason for the formation of ferroelectric domains inside the PZT crystals is to minimize the elastic energy associated with the mechanical constraints to which the PZT crystal is subjected as it is cooled, through the paraelectric–ferroelectric phase transition [7]. To better understand this, assume that a part of the PZT crystal is subjected to compression forces while cooling down from the cubic phase to the tetragonal phase. At the phase transformation temperature, spontaneous dipoles start to form inside PZT; however, they will form in the directions perpendicular to the compression forces, in order to minimize the elastic energy of the system. Therefore, the mechanical forces, and the elastic energy of the system are minimized by the fragmentation of the grains into separate domains [7]. As a result, a complex domain structure develops in each grain, according to the allowed 7  directions of the polarization in each domain. In the tetragonal phase, the so-called separated 180 and 90 domain walls form (Figure 5) [7].  2 1  1  2  Figure 5- TEM image showing domains in PZT [8] and a schematic of the 180 (1-1) and 90 (2-2) domain walls in PZT  In each ferroelectric domain inside a grain of a ferroelectric ceramic, a uniform orientation of the dipoles exists. It should be noted that since the grains and the domains contained in them are randomly oriented, the properties of the ferroelectric ceramics are isotropic both after the synthesis and after the cooling below the Curie temperature [7]. By applying an electric field with sufficient magnitude, the spontaneous polarization inside the ferroelectric ceramic can switch to a different, stable direction. By stable direction we mean that the spontaneous polarizations will not return to its original direction and magnitude when the electric field is removed, as shown in Figure 6. This brings up the most important characteristic property of PZT, which is the switchable Polarization (P)-Electric Field (E) hysteresis, as schematically shown in Figure 1. 8  remnant strain  (a)  (b)  Figure 6- A ferroelectric ceramic with differently oriented domains inside each grain (a) before and (b) after polarization, with remnant strain  2.2 Hydrogen damage of PZT  Four different kinds of degradation of the electrical properties after H2 treatment have been reported for PZT: 1) loss of polarization hysteresis [9], 2) increase in leakage current [9], 3) drop in resistivity [10], and 4) decrease in dielectric constant [9]. To confirm that this degradation is not just due to the reducing nature of the atmosphere (usually a mixture of nitrogen and hydrogen), the ferroelectric capacitor characteristics were also measured after the heat treatment in nitrogen and they were unchanged. Therefore, the degradation is due to the hydrogen in the atmosphere entering PZT and altering atomic structure of PZT. The decrease in resistivity after H2 treatment has also been reported: the resistivity of as-grown PZT drops from 5×1011 Ωcm to 2×107 Ωcm after annealing in forming gas (a mixture of up to 5.7% hydrogen and nitrogen) at 400°C for 30 min [10]. Although strong changes have been observed in the polarization hysteresis characteristics and leakage current  9  of the hydrogen-treated Pt/PZT/Pt ferroelectric capacitors, their relative dielectric constant measured for a low-voltage signal was not greatly affected; it was about half of the asreceived sample, i.e. 800. This may indicate that the PZT film was not damaged completely through its thickness [9]. The degradation mechanisms will be discussed in the next section, and it will be shown that the deterioration occurs mostly at the interface between PZT and electrode. There are just a few papers regarding the degradation of the mechanical properties of the PZT after hydrogen treatment [11-13]. The general conclusions drawn from these papers are the following: -The cohesive strength of PZT decreases due to the presence of H or H+ in the lattice. -Recombination of H or H+ at grain boundaries and in micro-voids may form molecular hydrogen; when the internal pressure of H2 exceeds the cohesive strength of the grain boundaries, fissures or micro-cracks appear. Delayed hydrogen-induced failure has been reported for PZT ceramics during charging by hydrogen under a constant load [11]. Therefore, it can be concluded that hydrogen atoms incorporated into the structure of PZT can decrease the cohesive strength of PZT, although they also have a considerable effect on ferroelectric properties. Wang et al. have studied hydrogen induced delayed fracture of PZT [11]. Their results show that the strength of PZT decreases with increasing the hydrogen concentration inside the specimen. Moreover, they have also found that the KIH/KIC decreases with the hydrogen concentration in the specimen,  C0, in the form of KIH/KIC = 0.400-0.155ln(C0) where C0 is the concentration of hydrogen 10  which can diffuse out from the samples after charging, KIH is the fracture toughness of hydrogen-treated sample and KIC is the fracture toughness of the un-affected sample. They have reported that the fracture mostly occurred in inter-granular mode, which shows that cracks were mainly initiated at the grain boundaries. Peng et al. have investigated the initiation and propagation of hydrogen fissures in a PZT ferroelectric ceramic during charging by hydrogen without loading [12]. Their results show that when hydrogen concentration in PZT exceeds a certain value (about 260 ppm), hydrogen fissures or micro-cracks form within PZT. Figure 7, reproduced from [12], shows the initiation of such micro-cracks at grain boundaries. Usually a typical sintered PZT ceramic is not 100% dense, and there are many voids and porosities at the grain boundaries. The recombination of hydrogen atoms at such porosity or voids results in increasing pressure of the molecular hydrogen (H2) inside such holes, and when the hydrogen pressure inside these voids or porosities becomes equal to the strength of PZT at the grain boundary, which has been decreased by the presence of atomic hydrogen, hydrogen fissures or microcracks form [12].  11  Figure 7- Initiation and growth of hydrogen fissures during charging at 50 mA/cm2 for 2h (b) and 4h (c); the black bar is 50 µm [12]  2.3 Mechanisms of hydrogen damage of PZT  2.3.1 Hydrogen incorporation into PZT  Hydrogen damage has been linked to structural modifications of PZT, which lead to the changes in the properties of PZT [1-2]. This could occur due to chemical reactions between hydrogen and constituents of PZT, or simply due to the presence of hydrogen (in different forms of ions, atoms, or molecules) within PZT. Before discussing these mechanisms, we first need to address two important issues: 1) the paths for the incorporation of hydrogen in PZT and 2) the stable forms of hydrogen inside the lattice of PZT (H or H+).  12  There are many uncertainties regarding the second issue, and in the following sections we will review the results reported in other studies.  2.3.1.1 Mechanism  In general two different scenarios can be considered. The first mechanism is that hydrogen molecules dissociate at the PZT surface, and as a result H atoms diffuse into the structure (Figure 8 ). Generally, such mechanism can be active in bare (no electrodes) PZT crystals, as observed in a few experiments [9, 14-15]. However, most oxides have limited hydrogen diffusivity [16], and the results show that hydrogen can be incorporated into the PZT structure only at temperatures higher than 400°C [17]. Another possible mechanism, especially at temperatures as low as 200°C, is the incorporation of H into the structure from the metallic electrode. Hydrogen molecule dissociates at the surface of the electrode, and thus produced H atoms diffuse to the electrode/PZT interface, then continue into the PZT crystal and modify the PZT structure (Figure 8). Indeed, this is the most probable mechanism, and it has been reported in many studies [9, 14, 18-21] as the degradation of PZT properties can be correlated with the properties of electrodes. For example, the PZT-electrode assembly with In2O3 electrode showed the least amount of degradation when subjected to hydrogen atmosphere, whereas Pt is the worst electrode [21], due to the highly catalytic nature of Pt helping in dissociating the hydrogen molecules into atoms. 13  2 H 1 H2  H  H2  H H  Electrode PZT  H H  Figure 8- Different paths for the incorporation of hydrogen into PZT: (1) through the surface of bare PZT (only at high temperatures) and (2) through the surface of electrodes (at low temperatures).  Figure 9 [17] shows that degradation of PZT with Pt electrodes is more severe in comparison to the bare PZT, which clearly confirms the catalytic behavior of Pt. Figure 10 [19] shows the effects of different electrodes on the remnant polarization (Pr). It is clear that capacitors with Au or Ag electrodes are much more stable during H2 annealing than those using Pt or Pd. The difference in the level of H2 damage corresponds to the difference in the catalytic activity in the hydrogenation reaction and the adsorptive properties of hydrogen by metallic electrodes [19].  Figure 9- Switching charge as a function of H2 annealing temperature with and without upper Pt electrode [17]  14  Figure 10- Remnant polarization as a function of the H2 annealing temperature for capacitors with Pt, Pd, Au or Ag electrodes (applied voltage of 5V) [19]  2.3.1.1 Interactions between hydrogen and metallic electrodes The incorporation of hydrogen in PZT through the metallic electrode includes the steps of (i) hydrogen absorption into the metallic electrode and dissociation of hydrogen molecules on the surface of the metallic electrode, and (ii) diffusion of hydrogen atoms through the metallic electrode and into PZT (Figure 8). The absorption of hydrogen molecules (H2) into metallic electrode via the gas phase can be described by the following chemical reaction [22]: ( )  (1)  where [H] refers to hydrogen atoms dissolved in the metallic electrode. At equilibrium, the chemical potential of hydrogen in the gas phase is equal to the chemical potential of hydrogen dissolved in the metallic electrode, therefore: ( )  ( )  (2)  15  ( ) is the chemical potential of gaseous hydrogen per molecule and  where  ( ) is the  chemical potential of hydrogen atom dissolved in the metallic electrode per atom. The chemical potential of gaseous hydrogen can be written as follows: ( ) where constant,  ( )  (  )  (3)  ( ) refers to the chemical potential at a given standard state, is the fugacity of H2 and  is the Boltzmann  is the pressure at standard conditions (1 bar);  is  ( ). At pressure of hydrogen below 10 MPa [22-23]  defined as the activity of hydrogen  hydrogen can be considered as an ideal gas; therefore,  will be equal to  , the partial  pressure of hydrogen. Therefore, the chemical potential of gaseous hydrogen can be written as follows: ( )  ( )  (  )  (4)  On the other hand, the chemical potential of hydrogen in the metallic electrode can be written as [22]: ( ) where  ( )  (  ( ))  (5)  ( ) is the chemical potential at a given standard state, and  ( ) is the activity of  hydrogen in the metal. The chemical potential of atomic hydrogen in metals can also be written as [23]: ( ) where  is the partial enthalpy,  (  (  ))  is the non-configurational part of entropy,  number of interstitial sites per metal atom, and  (6) is the  is the number of hydrogen atoms per 16  metal atom (i.e. hydrogen solubility, expressed as atomic fraction). Taking into account equations (2), (4) and (6), the following relations can be obtained for the hydrogen solubility into the metallic electrode: (  ) ( )  (7)  or (  ) (8)  where  is the heat of dissolution of hydrogen inside the metal, and  mixing. The values of  and  is the entropy of  for the dissolution of hydrogen in different metals can be  obtained experimentally, and the literature data are reported in Table 1. The value of depends on the electronic structure of the metal in which hydrogen is being dissolved, and the  is predicted by theoretical studies to be around -7.8 [22]. Table 1- The heat and entropy of dissolution of hydrogen in different metals [22]  Metal Fe (bcc) Al Ni Pd Pt Cu Ag Au U(α)  ([eV per atom]) 0.25 0.70 0.17 0.1 0.48 0.44 0.71 0.37 0.1  / -6 −6 −6 −7 −7 −6 −5 −9 −6  T(C) <900 500 350–1400 −78–75 − <1080 550–961 700–900 <668 17  For small hydrogen concentrations ( (  ), equation (7) can be written as:  )  (  )  (9)  or (10) This relation is known as Sieverts’ law [22-23], which states that the solubility  of  hydrogen in metals is proportional to the square root of the partial pressure of the hydrogen in equilibrium with the metal.  is known as the solubility coefficient. Sieverts’ law is also  applicable for other diatomic gases (eg. N2, O2) [22]. Figure 11 shows the experimental data on hydrogen solubility in various metals as a function of temperature [24]. T (C) 1000 600 400  8  200  Pd  4 Ln cH  Fe  Ni  0  Pt Ag  -4 Al Cu  -8 4  8  12 16 20 10000/T K-1  24  Figure 11- Solubility (cc of hydrogen per 100 g of metal) of hydrogen inside metals at 1 atmosphere pressure of hydrogen [24]  Let us consider the role of the metallic electrode on the level of hydrogen absorption by PZT. The case under consideration is the metallic electrode attached to PZT, and both are 18  in the hydrogen atmosphere (Figure 12). In this case, the hydrogen adsorption in PZT includes the steps of (i) the hydrogen absorption into the metallic electrode and its dissociation, and (ii) diffusion of hydrogen atoms through the metallic electrode and into PZT (it is assumed that hydrogen diffusion is just in the z direction). The chemical potential of hydrogen in the metallic electrode  ( ) and PZT (  ( ) where and  (  (  )) can be described by [25]: )  (11)  is the hydrogen concentration at the partial enthalpy of stands for  (metal) or  ,  is the gas constant  . H2 (g)  A B  electrode z PZT  Figure 12- Schematic cross-section of the electrode/PZT assembly in H2 gas (it is assumed that hydrogen diffuses only in the z direction)  During hydrogen diffusion, the following boundary equilibrium conditions can be considered [25]: ( )  at interface A : at interface B : It should be noted that both  ( ) and  ( ) (  (  ( )  (12)  )  (13)  ) are functions of time and space; that is  because during the hydrogen diffusion, when the system is not in equilibrium,  is changing  19  ( ) and  with time and space, and therefore both  (  ) are changing with time and  space [25]. Comparing equations (11) and (13) the following relation can be written: ( By adding the term  )  (  )  (14)  ( ) to both sides of equation (14), the following relation is obtained  [25]: ( where  and  )  )  (15)  are the heat of hydrogen dissolution in metallic electrode and PZT,  respectively. Equation (15) shows that the depends on  (  ,  and  , the hydrogen concentration inside PZT,  . According to this equation, we expect that with changing  the metallic electrode, different amounts of hydrogen would dissolve in PZT. The different level of hydrogen dissolution in PZT due to the different electrodes is therefore the reason for the dependency of the damage affected by hydrogen on the metallic electrode (Figure 10). Another important boundary condition which must be satisfied in order to conserve the H atoms is the equality of the flux of hydrogen atoms which leave the metallic electrode to the flux of hydrogen atoms which enter the PZT at the interface B ( ) [25]. In other words, at the interface B, the following equation must be satisfied: (  where  )  (  )  (16)  is the diffusion coefficient of hydrogen in the metallic electrode, and  diffusion coefficient of hydrogen in PZT. According to equation (16),  and  is the are other  parameters which control the hydrogen entry to PZT. Therefore, based on equations (15) 20  and (16), one can conclude that the hydrogen incorporation, hence the damage to PZT, indeed depends on  ,  and  , i.e. on the type of the electrode used.  According to equation (16), the rate of hydrogen atoms which leave the metallic electrode increases with  . Moreover, the time needed for the hydrogen atoms to reach the  interface between the electrode and PZT from the surface of the electrode depends also on . However, often the electrodes used in making the PZT capacitors, e.g. for FeRAMs, are very thin (about 10 nm). Therefore, we expect that these diffusion times (<1 sec) are much smaller than the typical duration (30 min) of hydrogen treatments (the depth of hydrogen diffusion ( ) from the surface of the electrode can be estimated by nm, and if we assume  √  . If  = 10  to be equal to 10-5 cm2/sec at 150C Figure 13, then the time for  hydrogen atoms to reach the interface between the electrode and PZT will be on the order of microseconds). Figure 13 shows the hydrogen diffusion coefficient for different metals as a function of temperature [25]. T (C) 1000 600 400 200  0.8 D (cm2/sec) 10-4  Pd  0.6 Pt  0.4 Ni  0.2  Au Ag Cu  0 4  8  12 16 20 10000/T K-1  24 28  Figure 13- Hydrogen diffusion coefficient in different metals [25]  21  In summary, the role of metallic electrodes on the damaging effect of hydrogen on PZT can be evaluated by considering the system parameters such as hydrogen diffusivity heat of hydrogen dissolution in the metal  ,  , and hydrogen concentration in the metal  . If we want to predict the amount of hydrogen damage with different electrodes, not only the above parameters, but their interplay should be considered as well; moreover, the respective PZT parameters (  ,  ,  ) should be considered. For example, by  considering Figure 11 and Figure 13, one can see that the hydrogen diffusion and hydrogen solubility are higher in Pd than in Pt. Therefore, one may expect that Pd may cause more degradation to PZT in comparison to Pt. However, greater damage has been observed in the samples with Pt electrode (Figure 10), possibly due to the higher heat of hydrogen solubility in Pt than in Pd.  2.3.2 Stable forms of hydrogen in PZT  In general there are three possible forms of hydrogen in oxides; it can exist as a hydrogen atom (H), as a hydrogen ion (H+ or H-) or as a hydrogen molecule (H2). As an atom H, it will just fill the interstitial sites of the lattice, with no interaction with other elements of the structure, especially with oxygen anions. These hydrogen atoms can diffuse freely through the structure without changing the electrical properties of the oxide [27]. On the other hand, hydrogen atom can ionize to H+ and release one electron in the lattice. The produced proton (H+) cannot exist as such, because it is very unstable in this form [28], so it 22  will react with the oxygen anions of the lattice and form an O-H bond. This bond is directional, and depending on the interatomic distances, the coordination number of the proton would be one or two [28]. When there are large interatomic distances, the coordination number of the proton is one. On the other hand, a proton can be shared between two oxygen anions when the distance between the oxygen anions is short enough [28] (the interatomic distances in the crystalline lattice strictly depend on the radii of cations and anions which formed the structure). Finally, hydrogen can also exist as a molecule within the structure of oxides. In this form, hydrogen might only exist at grain boundaries, voids, or pores, where there is enough space for a hydrogen molecule. It has been suggested that hydrogen atoms may form hydrogen molecules at grain boundaries, and this may cause the formation of cracks in these regions [12]. The important question is “what is the stable form of hydrogen in the crystalline  lattice of PZT?” Xiong and Robertson [29] have investigated stable forms of hydrogen in the structure of PbTiO3 and PbZrO3 using the first-principle of quantum mechanics or ab-initio calculations. Their results show that a hydrogen atom will form a donor state in these structures, i.e. the stable form of hydrogen was the proton (H+) [29]. This result can also be applicable for PZT, as the formation of [OH]– bonds in PbTiO3 has also been confirmed by Park and Chadi [30]. Using first-principles calculations, they have shown that hydrogen impurities will act as shallow donors in the structure of PbTiO3. Raman spectroscopy results of hydrogen treated PZT samples also confirmed the existence of O-H bonds in the structure of PZT [31]. Therefore, it is reasonable to presume that the stable form of hydrogen in the 23  crystalline PZT is H+. However, based on the above results, one cannot conclude that hydrogen cannot exist in other forms (H or H2), but it is certain that some of the hydrogen atoms will be ionized inside the PZT lattice.  2.3.3 Stable sites of H+ in PZT  Determining the stable sites of protons in PZT matters, as protons form directional bonding with oxygen ions, and this could affect the built-in dipoles in PZT. Generally the stable sites of H+ in the crystalline lattice of perovskite oxides depend on the binding interactions between the H+ and oxygen anions [32]. In other words, since H+ will form an OH- bond in the crystalline lattice, it is reasonable to expect that the position of H+ will depend on the oxygen sites in the structure. Furthermore, one can say that “protons are localized within the valence electron density of the oxygen” [32]. The positions of OH- bond also will be determined by its interaction with cations existing in the structure [32]. Kreuer has suggested two stable sites for protons [32], depending on the lattice constants of the perovskite oxides. For the perovskites with large lattice constants, the stable sites of protons will be the edges of octahedrals (Figure 14a). On the other hand, for the perovskites with short lattice constants, the protons also could be shared between two oxygen ions of two adjacent octahedrals (Figure 14b) [32].  24  (a)  (b) Figure 14- The stable positions of protons according to the reference [32] in a) perovskites with large lattice constants, and b) perovskites with short lattice constants  Aggarwal et al. have also investigated the possible sites of protons in the structure of PZT [31]. They considered four possible sites for protons, and based on the Raman spectra obtained for hydrogen treated PZT samples, concluded that the more probable site for H+ is between the apical oxygen ions and Ti (Figure 15a) [31]. They concluded that just one of the apical oxygen ions will react with a proton [26]. Park and Chadi [30] have investigated the stable sites of protons in the crystalline lattice of PbTiO3 using first-principles calculations. For the tetragonal phase of PbTiO3 they have suggested different stable positions for protons, as depicted in Figure 15b. They have also predicted that depending on the position of protons, the [OH]– will either destroy, or enhance the polarization of the spontaneous dipoles in PZT [30]. However, they have predicted that in the presence of the [OH]– dipoles, the spontaneous dipoles cannot be switched by applying the electric field. On the other 25  hand, for a cubic PbTiO3 they have found the most stable site to be the site shown in Figure 15c.  (a)  (b)  (c) Figure 15- PZT crystal structure showing the possible location of protons in the lattice of PZT [31]; b) stable lattice site of protons for tetragonal PbTiO3 and c) cubic phase of PbTiO3 supposed by Park and Chadi [30]  2.3.4 Effect of hydrogen on PZT We can now focus on the possible mechanisms responsible for the degradation of PZT by hydrogen. As seen in Figure 16, the structural degradation can be categorized by the place 26  where it occurs: a) at the grain boundaries and at the interfaces between PZT and electrodes, and b) in the crystalline lattice of PZT.  O and Pb vacancies (mostly at grain boundaries)  degraded layer due to [OH]– (inside the crystalline lattice)  Figure 16- Proposed PZT damage mechanisms can be categorized by the place where damage occurs  Damage at the grain boundaries and interfaces of PZT and electrodes When H atoms diffuse from the electrode and reach the interface between the electrode and PZT, some may diffuse fast through the grain boundaries. Such hydrogen atoms may affect the structure and properties of the PZT in different ways. The first mechanism is that hydrogen atoms may undergo the following reaction [33]: 2H + O2-  H2O + Where  (17)  is the oxygen vacancy with two electrons. Oxygen vacancies produced by the  reaction (17) can further ionize and produce two free electrons; this can decrease resistivity of PZT. However, because H2O cannot exist inside the PZT ceramic body, reaction (17) can only occur on the surface of the PZT ceramic grains [11]. Therefore, for such a reaction to continue at the interfaces, oxygen atoms must diffuse from the bulk of the grains to the interfaces, and water molecules should also diffuse out from the sample along the grain  27  boundaries. Generally, such diffusion, especially the oxygen diffusion in PZT, is slow. As such, the above damage by H can only occur at the grain boundaries of PZT and the damage is limited to the grain boundaries [11, 34]. Therefore, the damage that occurs by reaction (17) is only at the surface of the ceramic particles, which may include electrode/PZT and grain boundaries (Figure 16). It should be noted that the oxygen vacancies might also be produced by the following reaction, without a direct reaction with hydrogen:   + ½ O2 (g)  (18)  The above reaction occurs in reducing atmospheres. Experiments show that treatment of PZT under a hydrogen deficient atmosphere (such as pure N2) has no noticeable effects on the properties of PZT, while similar treatment in hydrogen has considerable effects on the properties of PZT [16, 34]. It appears that the rate of the reaction (17) is too low to have noticeable effects on the properties of PZT. Shimakawa and Kubo have proposed that PZT color change (after exposure to H2) from white to black is due to the formation of such oxygen vacancies [35]. It was also supposed that oxygen defects would produce donor levels within the PZT band gap, which would account for the change in color and would increase the leakage current in capacitors [35]. Another mechanism, by which H atoms could change the PZT structure, is the formation of Pb vacancies. Indeed, the presence of metallic Pb at the electrode/PZT interface and at grain boundaries after the hydrogen treatment was previously reported [34-36]. The electrical properties of PZT may change due to the formation of Pb vacancies in PZT, 28  because they can produce free holes in the valence band of PZT. The formation of lead vacancies is limited by the diffusion of lead, which is very slow in PZT at temperatures less than 600C (DPb in PbTiO3 =7.210-4exp(-181800/RT) (cm2/sec), equivalent to 9.9910-15 cm2/sec at 600C and to 2.710-11 cm2/sec at 1000C [37]) and is therefore likely limited only to the interfaces [35]. As the amount of metallic Pb is low in the H2-treated PZT (less than 0.3% [35]), Pb vacancies cannot be considered the responsible mechanism for the alteration of ferroelectric properties and color change from white to black [35]. On the other hand, Ikarashi [36] has concluded that the reduction of PbO in PZT to metallic lead is the main reason for the degradation of the ferroelectric properties of PZT [36]. The author has suggested that the degradation of the properties of PZT by hydrogen annealing could be avoided by using electrode materials that prevent the Pb diffusion from PZT [36]. It should be noted that the reduction of PbO in PZT to metallic lead does not occur only due to reducing atmosphere. H can diffuse into the PZT structure and change its atomic bonding [34], leading to reduction of PbO in PZT to metallic lead. No reduction to metallic lead was observed in the bare PZT plates up to 600C, while for PZT samples with Pt electrode the reduction to lead was observed at temperatures as low as 320C [34]. The changes in the atomic bonding of PZT can be due to the existence of H+ within the lattice and its bonding with O2-. This is discussed in the next section. Damage of the crystalline lattice of PZT Another mechanism proposed for the changes of the electrical properties of PZT is the ionization of hydrogen atoms inside the lattice of PZT [31]: 29  H  H+ + e-  (19)  Upon ionization, hydrogen releases an electron, which decreases the resistivity of the films, and the produced hydrogen ion will interact with an oxygen ion to form a polar hydroxyl bond [OH]- [31]: H+ + O2-  [OH]-  (20)  Aggarwal et al. presumed that the above mechanism is the main reason for the changes in the ferroelectric properties of PZT [31]. They have concluded this from the fact that oxygen diffusion coefficient in PZT at 200°C is not high enough to cause significant damage. Another reason to support this idea is that the annealing of PZT films in oxygen deficient atmospheres does not affect their ferroelectric properties [31]. It was concluded that the reactions (3) and (4) during PZT annealing with forming gas are the primary avenues for the degradation of ferroelectric properties [31]. The [OH]− ion acts as a fixed dipole, which does not allow the switching of the ferroelectric domains [17]; this idea has also been supported by theoretical studies [29-30]. As said before, it is supposed that the bonding of hydrogen with oxygen anions may change the atomic bonding of oxygen with the other elements of PZT (Pb, Zr, Ti) and as a result, it might cause some changes in the atomic bonding of PZT elements [35]. Among the reactions mentioned above, probably the reactions (19) and (20), occurring in the crystalline lattice of PZT, are most likely responsible for the changes in the electrical properties of PZT. That is because they affect the crystalline lattice of PZT, while the other reactions just modify the interfaces.  30  2.4 Dielectric spectroscopy of PZT Dielectric spectroscopy has been used for studying the properties of a wide range of materials, such as glasses, polymers, and ceramics [38]. This technique measures the polarization response of a dielectric medium to an applied electric field; when an electric field is applied to a dielectric medium, charges, (including ions and electrons), molecules, and dipoles are displaced according to the applied electric field, which causes polarization in the dielectric medium. By measuring the polarization of the dielectric medium, different parameters of the dielectric medium, such as impedance (Z), admittance (Y), and dielectric constant (ε) can be obtained. By analyzing these parameters, useful information can be obtained regarding the charges, dipoles, and molecules inside the dielectric medium. The parameters commonly used for analyzing insulators include dissipation factor DF and dielectric constant ε, which are measured in a wide range of frequencies (from mHz to MHz) and temperature, and are then used to analyze the data by fitting the results to one of the available mathematical models [38]. The classical model of dielectric relaxation of a dielectric medium containing dipoles is the Debye model. According to this model, the polarization ( ) of a dielectric medium changes in accordance to following equation: ( )  (  )  (21)  31  where  is a constant, t is time, and τ shows the relaxation time of the dipoles. Based on the  equation (21) for polarization, the complex dielectric constant ε (where ε= ε´+iε´´) can be expressed as ( )    (22)  where ε is the dielectric constant when  , εs is the dielectric constant when  0,  is the angular frequency, and i is the imaginary number [38]. A more flexible model, commonly used for modeling the dielectric constant data, is the Havriliak–Negami equation [38-39]. According to this model, the complex dielectric constants ε of a dielectric can be evaluated by: ( )  (  (  ) )  (23)  where θ and β are constants between 0 and 1 [39]. By fitting the dielectric data to one of the above equations, one can obtain different information about the dipoles inside the dielectric medium, such as the number of dipoles and the activation energy for moving (or hopping) of ions. For example, Kamishima et al. [40] investigated the dielectric properties of proton conductor Yb-doped SrZrO3 after hydration in water, and using this technique they were able to anticipate the position of the [OH] bonds inside the SrZrO3. In this work, we have attempted (for the first time according to our knowledge) to use this technique to assess the effect of hydrogen on the properties of PZT.  32  2.5 High pressure hydrogen compatibility of PZT The high pressure hydrogen compatibility of PZT is an issue which has recently been raised due to the possible application of piezoelectric actuators in the hydrogen fuel injectors. Modern electronic fuel injectors that use lead-zirconate-titanate (PZT) based actuators for valve opening, instead of the conventional solenoid technology, have been introduced by the leading engine manufacturers in recent years (Figure 2). Since the valve is actuated quicker (i.e. about 5 times faster [41]) by the piezoelectric actuators than the conventional solenoid technology, very precise injection intervals become possible between the pre- and main injection. Consequently, fuel consumption and emissions are noticeably reduced (by up to 15 percent) [42]. Piezoelectric actuators also facilitate an increase in the injection pressure, up to 250 MPa; the higher the pressure and the more accurate the dosing and timing of the injection, the more efficient (and therefore less polluting) the combustion event becomes [42]. It should be noted that whereas the valve needle stroke was fixed in the previous electromagnetic injection systems, in injectors where the piezoelectric actuator acts directly on the needle, the needle stroke can be varied by changing the magnitude of the applied voltage, thus enabling better control over the valve opening. It was reported previously that PZT thin films lose their ferroelectrical properties after hydrogen treatment [2]. However, the hydrogen environment used in the published studies [2, 9-10] is not comparable to the hydrogen environment conditions in an engine, where the pressure of the gas is up to 30 MPa, and the maximum operating temperature of  33  the PZT ceramic is about 100C. Alvine et al., have investigated the effect of high pressure hydrogen on the properties of PZT films [43]. They studied the structural and compositional changes of PZT thin films (50 nm) after 24 hours of high pressure hydrogen treatment (p=13.8 MPa, T=100C). The most important structural changes which they observed was the hydrogen induced blistering on the surface of bare PZT films and PZT film with Pd electrodes [38]. They have also observed “significant mixing of the Pd layer into the PZT film along with migration of Pb into the Pd layer” [43]. The hydrogen absorption for bare PZT films was about 10 at%. The Pd layer on the surface of PZT films had a considerable effect on the amount of hydrogen absorption: due to the presence of the Pd layer the hydrogen concentration in the PZT ceramic is increased to nearly 20 at% [43]. Other results from their experiments are as followings: (a) Piezo actuators can degrade in high-pressure hydrogen environments due to the hydrogen uptake in the PZT plates. (b) The amount of hydrogen absorption is a function of the metallic electrode and it increases for different electrodes in the following order: Pd> Al> W> Ti> Cu. (c) Lead migrates into all the above mentioned electrodes, with the possible exception of Ti. One of the main issues not addressed in the work of Alvine et al. is the possible effects of the high pressure hydrogen environment on the electrical properties of PZT. Therefore, in the present work, in addition to investigation the effect of high-pressure hydrogen environment on the microstructure of PZT, we also investigated the effect of hydrogen on the electrical properties of PZT. 34  2.6 Methods of decreasing the hydrogen damage to PZT One such method is based on the experimental observations that the degradation of properties of PZT is related to the hydrogen reactivity and absorption of the metallic electrodes. For example, PZT-electrode assembly with Au electrodes showed the least amount of degradation when subjected to an hydrogen atmosphere, whereas Pt was the worst electrode [20], due to the catalytic nature of Pt in dissociating the hydrogen molecules into atoms. Therefore it has been suggested to use electrodes with less catalytic activity such as Au, instead of electrodes such as Pt [9]. Moreover, it was suggested that oxygen plasma treatment of the Pt electrode can be used to reduce its catalytic activity [44], as this procedure modifies the surface of the electrode. Conductive metal oxide electrodes like IrO2, LaNiO3 have been tried, and it has been found that annealing in a hydrogen containing atmosphere will not degrade the properties of PZT with such electrodes [20]. Abdolghafar et. al. proposed IrO2 as the top electrode to prevent the hydrogen damage to the PZT during hydrogen treatment [20]. Their results show that PZT capacitors with IrO2 electrodes have poor hydrogen resistivity because of the reduction of the IrO2 to metallic Ir during hydrogen treatment; however, after oxygen pre-annealing at 600C, the PZT capacitors with IrO2 electrode showed excellent hydrogen damage resistivity. The hydrogen damage resistivity after oxygen annealing is attributed to the enhancement of the IrO2 structure by oxidation. Another technique designed to improve the hydrogen resistivity of the PZT capacitors is using a hydrogen diffusion-barrier layer on top of the capacitor. This technique  35  prohibits hydrogen diffusion by encapsulating the whole capacitor (electrode/PZT/electrode) in hydrogen barrier layer(s), sometimes called a hard mask. It has even been shown that even sidewall diffusion barriers can dramatically enhance the hydrogen resistivity [40]. Both conductive (e.g. TiAlN) and non-conductive layers (e.g. Al2O3, SiO2) can be used as a hard masks. Saito et al. have investigated different oxides (Al2O3, HfO2, Bi3Ti4O12, ZrO2 and SiO2) as encapsulation layers [46]. They suggested Al2O3 and SiO2 as promising hydrogen diffusion barrier layers. While the above techniques try to prohibit hydrogen damage during the hydrogen treatment process, another idea is to recover the properties of PZT capacitors after hydrogen annealing [21]. This could be done by high temperature (600-700C) annealing of the PZT capacitor in atmospheres like N2 or air after the hydrogen gas treatment. It has been reported that the Pt/PZT/Pt assembly will recover its properties after treatment in O2 gas [17 and 21]. The recovery of the properties has been suggested to be due to the diffusion of hydrogen atoms out of the sample during the post-annealing treatment. These results confirm the recently published theoretical and experimental results by Bjorheim et al., who investigated the hydration thermodynamics of PbZrO3 and concluded that protons which are absorbed inside PZT from the hydrogen gas can be removed from PZT by heat treatment in air at temperatures higher than 700C, or at lower temperatures in dry oxidizing or inert atmospheres [47].  36  3 Scope and Objectives Scope The investigation of the interactions of hydrogen with PZT is an important topic both from practical and scientific points of view and still many issues need to be addressed in this regard. While a lot of attention has been paid to the effects of hydrogen on the ferroelectrical properties of PZT, there are very few publications on the effects of hydrogen on the piezoelectrical properties of PZT. This is an important issue for using the piezo-actuators in the advanced internal clean-combustion engines, and indeed there is very limited understanding of the performance and durability of such piezo-actuators in hydrogen environments. The broad scope of this project is to examine the key research issues related to the performance of the PZT-based piezo actuators exposed to hydrogen, and to develop methods for preventing or limiting the damaging effects due to hydrogen, with the aim of extending the lifetime of the actuators. Specifically, the scope of this work involves the following activities: (a) Bare PZT plates and similar plates including silver electrodes (Ag/PZT/Ag) are exposed to hydrogen atmosphere at 10 MPa pressure for 200-1200 hours. The changes in the microstructure of PZT plates are investigated using scanning electron microscopy combined with energy-dispersive X-ray spectroscopy (SEM/EDX), and X-ray diffraction (XRD). The effect of hydrogen on the electrical properties is investigated by measuring the changes in the capacitance of Ag/PZT/Ag capacitors online during the hydrogen treatment.  37  (b) The kinetics of PZT structural modification due to hydrogen exposure is investigated using online monitoring of the electrical properties of PZT. Specifically, the changes in the capacitance of the Ag/PZT/Ag capacitors are measured online during the hydrogen treatment at temperatures higher than the Curie temperature. Furthermore, we measure the dielectric constant and dissipation factor of PZT after the hydrogen treatment in a wide range of frequency (from 12 Hz to 200 kHz) and temperatures (25 to 400C); these data are then analyzed by fitting the results to one of the existing mathematical models. (c) The effect of hydrogen on the microstructure and electrical properties of PZT in the tetragonal phase (i.e. at room temperature) is investigated using the water electrolysis technique, wherein PZT plates are exposed to atomic hydrogen generated in water electrolysis for 6 and 48 hours. SEM is used for evaluating the changes in the microstructure. The capacitance of Ag/PZT/Ag capacitors is measured on-line during the water electrolysis and after finishing the hydrogen exposure (i.e. during PZT aging in air), to evaluate the effects of hydrogen on the electrical properties of PZT. (d) Alumina coatings are applied to PZT plates using sol-gel technique, to explore the possibilities of decreasing H2 damage to PZT. The functionality of the coating against hydrogen damage is evaluated by water electrolysis technique. It is anticipated that the findings of this project will contribute to the fundamental understanding of PZT-hydrogen interaction, and will also provide Canadian clean-engine technology developers with strategies for improving hydrogen technology, thus becoming more competitive in the global market. 38  Objectives The broad objective of the present work is to evaluate the interactions between the PZT-based piezoelectric materials used in piezo actuators, and hydrogen. Based on the improved understanding of such interactions and the resulting damage to PZT, we aim to propose methods to decrease the negative effects of H2 on PZT. Within this broad goal of the project, we have the following specific objectives: 1. To evaluate and quantify the effects of long-term high-pressure gaseous hydrogen exposure on electrical properties and microstructure of PZT 2. To determine the parameters describing the kinetics of the interactions between hydrogen and PZT, in particular the incorporation of hydrogen into PZT and the resulting changes in the PZT properties 3. To evaluate and quantify the effects of hydrogen on the electrical properties and microstructure of PZT below the Curie temperature, i.e. in the temperature range of the actuator's use in the engine 4. To propose and develop methods to decrease and prevent the hydrogen damage to PZT, such as through the deposition of protective coatings.  39  4 Materials and Methods 4.1 Samples  “PZT” plates, 10×10×1 mm, (PIC 255, PI Ceramic GmbH, Lindenstrasse, 07589 Lederhose - Germany) of the composition Pb(Zr0.53Ti0.47)O3 with 1% Nb2O5 dopant were used in this work. The effect of hydrogen on PZT microstructure was investigated using poled PZT plates with electrodes, as well as bare, not poled plates. Two types of electrodes were used: silver (which contained small additions of Bi) and silver-palladium alloy (76 wt% Ag + 24 wt% Pd, as given by SEM/EDX); they were screen-printed on the large faces of the plates, in 20 µm layers. Silver and their alloys are usually used in making the actuators, and therefore, we considered this type of electrodes. The microstructure of the bare PZT samples is shown in Figure 17.  bare PZT  Figure 17- The microstructure of bare PZT plates  40  The interface between the electrode and PZT is shown in Figure 18. As it can be seen in this figure, there is a good adhesion between the electrodes and the PZT for inside section (Figure 18a) and poorer on the edge section (Figure 18b). Later it will be shown how hydrogen affects this interface.  Ag or Ag/Pd electrode  Ag or Ag/Pd electrode  PZT  PZT  Figure 18- The microstructure of the PZT samples  Figure 19 shows micrographs of as-received electrodes' surfaces; the Ag-Pd electrodes were much more porous than the Ag electrodes.  41  Ag or Ag/Pd electrode  PZT  Figure 19- Micrographs of the surface of the electrodes: silver (a) and silver-palladium (b)  4.2 Gas hydrogen treatment Two hydrogen gas treatment procedures were used in this work. The one which we call “High-Pressure” hydrogen treatment is schematically shown in Figure 20. For this condition of hydrogen treatment different exposure times (200-1200 hours) were used, while the pressure and temperature were kept constant at 10 MPa (pure hydrogen), and 100C, respectively. Before point (a), the chamber was purged several times with argon to remove air. The temperature and pressure of the gas were chosen according to the practical condition of the actuators in the engine [43]. Therefore, the results of this experiment can be used to evaluate the possible effects of high-pressure hydrogen environment on the microstructure of PZT plates used in fuel injectors. After finishing the experiments the samples were taken  42  out from the chamber and microstructural analysis was done using XRD and SEM  Temperature  techniques.  d Cooling in Hydrogen  Hydrogen Vacuum On On  a  b  Heater on  e  c Time  Figure 20- The time-temperature schedule of the ‘High-Pressure’ hydrogen treatment used in this work  Fuel injection actuators are made of many (>50) single PZT plates stacked together, and the electrical properties of each single layer determine the performance of the whole actuator (Figure 21). electrodes R C  PZT plate  +  -  (a)  -  +  (b)  Figure 21- Schematic of an actuator made from PZT plates stacked together, b) the equivalent electrical circuit of an actuator  43  Therefore, measuring the capacitance of a single PZT plate is a suitable way to investigate the effect of hydrogen on the performance of the actuator assembly in an hydrogen atmosphere. To this end, two thin copper wires were attached to the electrodes, and a GW Instek LCR meter (LCR-821) was used to measure online the capacitance of PZT plates during hydrogen exposure. The capacitance measurements were performed with an internal voltage of 0.125 V at a constant frequency of 1000 Hz. Figure 22 schematically shows the experimental setup for measuring the electrical properties.  High-Pressure Vessel P= up to 10 MPa T= up to 700 °C PZT microstructure  Copper wires connected to LCR meter  H2 LCR meter or Potentiostat  Silver Electrodes PZT (10×10×1 mm)  Connected to the vacuum pump and hydrogen  Figure 22- The schematic of the setup used in this work for online monitoring the electrical properties  The other hydrogen treatment procedure, which we would call “High-Temperature” hydrogen treatment, is schematically shown in Figure 23. From point “a” to “b” the sample was heated in air up to the desired temperature in 450C-650C range. The chamber was then vacuumed (700 mm Hg (0.09 MPa)) at point “b”, and at point “c” pressurized with 0.13 MPa 44  of 90% Ar/10% H2 gas mix. A drop of about 5C was observed, but the temperature  Temperature  recovered after 15 minutes.  b Vacuum On Heating in Air  c  d Hydrogen On Cooling in Hydrogen  e  a Time  Figure 23- The ‘High-Temperature’ hydrogen treatment used in this work  To measure the resistance and capacitance of PZT plates, two thin copper wires were attached to the silver electrodes using high temperature silver paste. The same LCR meter (LCR-821) was again used to measure the capacitance and the ‘real’ part of the impedance (ZRe) of PZT plates (Figure 22). The frequency was 1000 Hz for all measurements. The electrical properties were monitored online during the hydrogen treatment: capacitance and  ZRe values were collected every 0.896 second. Knowing the relation between the ZRe and the capacitance, the total resistance of the PZT plates can be estimated. Figure 24 shows a typical ZRe-ZIm curve measured for this type of PZT plates in air. Curves with the same trend were obtained for other temperatures, and also in the hydrogen atmosphere.  45  According to Figure 24, the PZT plate can be considered as a parallel R-C circuit. Therefore, the relation between the ZRe, R and C is as follows [49]: (24) Using equation (24) the ohmic resistance (R) of PZT can be calculated, but this method is not without errors, especially at low temperatures. The reason is that with decreasing the temperature, the Nyquist diagrams at high frequencies show a depressed semicircle; in other words, equation (24) is no further valid [49]. Therefore, we used the R values obtained by the equation (24) to just qualify the trend for the changes of R during the hydrogen treatment. We cannot say that the R values obtained from the equation (24) are the exact values for R during the hydrogen treatment; however, they can show the general trend for the changes of the resistance. In order to measure the exact values for R, we used another technique, detailed in the next paragraph.  .  In order to determine the electrical resistance of PZT, the variation of resistance from the ZRe-ZIm curves was also calculated. The ZRe-ZIm curves were recorded every 5 minutes during the hydrogen treatment; all ZRe-ZIm curves were semi-circles where the diameter of the whole circle can be considered as the ohmic resistance of PZT. For this set of experiments we used smaller PZT plates (4×4×1 mm), but with the same composition and electrodes; the estimation of R for both sets of sample showed the same trend. It should be noted again that we used the R values obtained either from equation (28) or from the ZRe-ZIm curves just to identify the general trend for the changes of R during the  46  hydrogen treatment. The electrical property used to monitor the structural changes was the capacitance C, which we were able to read directly from the LCR meter with reasonable accuracy. The accuracy of the LCR meter used in this work was 0.05% for measuring the capacitance.  400 R 300  Im (k)  C 200  100  0 0  100  200 ZRek  300  400  Figure 24- A typical Nyquist plot for the PZT at 500°C in air (the frequency is swiped between 106 and 10-3 Hz)  4.3 Water electrolysis treatment of PZT  The water electrolysis technique was used to charge the PZT samples with hydrogen, following a previously reported methodology [33]. In comparison to the treatment in hydrogen gas, which needs temperatures higher than room temperature, this technique has the advantage that it can be done at room temperature; this inhibits the de-polarization of PZT samples due to elevated temperatures. The basic idea for this technique is that one of the silver electrodes attached to PZT is used as the cathode during water electrolysis. In this 47  way, the hydrogen atoms which became released on the surface of the silver electrode can further diffuse into the electrode and into the PZT attached to the electrodes. The water electrolysis was performed in a 0.1 M NaOH solution, with a current density of 100 mA/cm2; the current density was kept constant by varying the voltage, to ensure the same rate of hydrogen release. The leakage of the solution to the interface between the electrode and PZT and the release of hydrogen at the interface caused the detachment of the electrode from PZT after very short experimental time (around 5 minutes). In order to restrict the leakage of the solution to the interface between the Ag electrodes and PZT, the edges of the specimens were encased in epoxy. After the samples were encased in epoxy, the electrodes did not detach from the sample during the water electrolysis. At different times after the beginning of the water electrolysis (ie. t= 10, 60, 120, 300, 600, 1200 minutes) the samples were taken out, dried, and the capacitance was measured. To measure the capacitance of PZT plates, the same LCR meter (GW Instek-LCR-821) was used. An internal bias of 0.125 V was used for measuring the capacitance, and the frequency was kept constant at 1000 Hz during all measurements. Figure 25 schematically shows the setup which is used for the water electrolysis experiment. After water electrolysis, the samples were prepared for the microstructural analysis. In this order, they were grinded, polished and then the cross section was studied by SEM.  48  Ag H+ H+ H+  PZT  H  H  H  H  H  H  0.1 M NaOH anode (+)  cathode (-)  Figure 25- Schematic of the setup for the water electrolysis experiment  In order to estimate the equivalent pressure of the hydrogen above the electrode during the water electrolysis, and to compare the results with the gas hydrogen treatment, we must first quantify the hydrogen absorption into the Ag layer during the water electrolysis. The electrolysis is actually a common and efficient technique for charging metals with hydrogen, wherein very high equivalent pressures of hydrogen can be produced above the metallic electrode [50]. It has been reported that very high fugacity of hydrogen on the order of 106 atmospheres (corresponding to pressures of about 104 atmospheres) can be obtained with cathodic charging [50]. Because of this high efficiency in introducing the hydrogen atoms inside metals, this technique has being used for producing metal hydrides, and for studying of hydrogen embrittlement in metals [51]. Wu measured the hydrogen solubility in iron by hydrogen cathodic charging [52]. At a potential of 0.25 V and in 1M  49  H2SO4 solution at room temperature, he has found that the solubility of hydrogen in iron can be described by the following relation: ( where  )  (25)  is the charging current density. Therefore, for a typical value for  (0.1A/cm2) at  room temperature, 0.53 ppm of hydrogen would dissolve in iron. On the other hand, the solubility of gaseous hydrogen of pressure PH2 in Fe can be obtained from the following relation: [53] (  )    (  )  (26)  Therefore, at 1 atmosphere partial pressure of hydrogen and at room temperature, the solubility of hydrogen in Fe would be about 0.00076 ppm, much lower than the predicted solubility of hydrogen in Fe during cathodic charging. Therefore, we can conclude that the fugacity of hydrogen gas above the Fe electrode during water electrolysis is much higher than the fugacity of gaseous hydrogen. This simple comparison confirms that high solubility of hydrogen in metals can be obtained with cathodic charging. Danielson found that the solubility of hydrogen in AA5083 aluminum alloy in cathodic charging (3.410-7 g-atoms H/cm3) is higher than the solubility of hydrogen in gas atmosphere (110-11 g-atoms H/cm3), by about 4 orders of magnitude [54]. He concluded that the cathodic charging of hydrogen has a major effect on increasing the hydrogen solubility in Al. There is no published data on the absorption of hydrogen atoms inside silver during the cathodic charging. Therefore, it is impossible to estimate the equivalent pressure of the  50  hydrogen above the electrode during the water electrolysis. However, it will be shown later (Section 5-3) that noticeable damage occurred to PZT after water electrolysis treatment. We therefore propose that this high degree of degradation is due to the higher hydrogen solubility in silver during the water electrolysis, in comparison to the gas atmosphere.  4.4 Alumina sol-gel coating Sol preparation Boehmite (ALOOH) sol was used as precursor sol for the alumina (Al2O3) coatings. Boehmite sol was prepared by dissolving 10 gr boehmite powder (Dispersal Sol P2, Condea Chemie GmbH, Germany) into 100 ml distilled water. The solution was mixed for 20 minutes at room temperature using magnetic stirrer. To increase the viscosity of the sol, PVA solution (10 wt% PVA) was added to the Boehmite sol (for the total amount of PVA equal 4.5 wt% of sol) and then solution was again stirred for 10 minutes to homogenize the sol before using it for dip coating. Figure 26 schematically shows the procedure used for the preparation of the sol.  51  100 ml Deionized water  10 gr AlOOH  Stir at room temperature for 20 minutes  Stable sol  100 ml PVA (10 wt%)  Stir at room temperature for 20 minutes  Final sol  Dip coating with withdraw speed of 3 cm/min  Firing at 450°C for 5 hours  Figure 26- Schematic of the steps used for the preparation of the alumina coating  The coating process The dimension of PZT plates used for coating was 10×10×1 mm. Before coating, the surface of the samples was ground, and then polished with 1 micron diamond paste. The samples were washed with acetone before coating. To apply the boehmite sol to the surface of PZT plates dip coating technique is used. “Dip coating" is a common coating technique  52  used in applications such as optical coatings and membranes. During the dip coating process, the substrate is submerged into the sol and carefully withdrawn out of the sol [55]. We selected this technique for the coating as it is relatively simple, and the samples were relatively small. The dip coating was done using the SCS PL3201 Dip Coater (manufactured by SCS, 7645 Woodland Drive, Indianapolis, Indiana, 46278, USA) with a withdraw speed of 3 cm/minutes. The thickness of the coating can be evaluated by the Landau-Levich equation [56]: ( ) (  where  (27)  )  is the coating thickness,  is the viscosity of the sol,  is the withdraw speed,  the sol-vapor surface energy, above equation and using the  is the density of the sol, and and  is  is the gravity. Considering the  values for water, and considering the viscosity of the  sol equal to 20 mPa.s [57], the thickness of the single-layer of as-deposited coating (i.e. before heat treatment) was calculated to be about 7 µm.  Firing treatment Heat treatment is always needed for the densification and crystallization of the sol layer which is applied on the surface on the PZT plates. In this way, a relatively dense film can be obtained on the surface, depending on the temperature and time of the heat treatment. The heat treatment used in this work was at 450C in air for 5 hours [58]. When doing high temperature processing of the coating layer, we have to consider its interaction 53  with the PZT (i.e. change of properties because of the changing chemistry). The temperature that Boehmite starts to transforms to alumina is reported to be about 450C [58]. Therefore, we selected this temperature for our heat treatment procedure. Furthermore, the lower the heat treatment temperature, the lower the risk for cracking is during the cooling due to thermal expansion. Based on the literature data [59-60] it is expected that after this treatment the coating still contained about 40 vol% of porosity. Therefore, the presence of the coating will not prevent access of molecular hydrogen to PZT surface, but it will prevent direct contact between the metallic electrode and PZT. Thus it is expected that there will be no access of atomic hydrogen to the coated PZT surface. The coating and firing steps were repeated 3 times in order to achieve > 5 m thick coatings. As shown later in Section 5-4-1, the actual thickness of the coatings after heat treatment, as determined by SEM, varied between 5 and 10 µm. Assessment of coatings effects To assess the effects of the coatings on hydrogen penetration into PZT, water electrolysis technique was used. In this regard, thin Au-Pd electrodes were sputtered on the faces of the PZT plates coated with alumina. Au-Pd electrodes were also deposited on the faces of a PZT plate without alumina coating and that sample was used as a reference sample. The water electrolysis was performed in a 0.1 M NaOH solution, with a constant current density of 100 mA/cm2 and with the voltage between 7-10 V. After water electrolysis, the cross section of the sample was studied with SEM and the amount of hydrogen damage was studied. 54  4.5 Characterization techniques  4.5.1 X-ray diffraction analysis (XRD) X-ray diffraction analysis (XRD) was used in this work for the phase analysis of PZT plates after hydrogen treatment, using Rigaku Multifles diffractometer operated at 40kV, and 30 mA (X-ray:CuKα). To perform the XRD analysis, the metallic electrodes were detached from the PZT plates, and the remaining bare PZT plate was used for the analysis. The dimensions of the PZT plates were about 10101 mm. XRD spectrums were collected over diffraction angles between 20 and 80 at a speed of 2 /min.  4.5.2 Scanning electron microscopy (SEM) The microstructural investigation was performed with a Hitachi S 3000-N Scanning Electron Microscope with an Electron Dispersive X-Ray Spectroscopy detector and Quartz X-One X-ray post-processing software. The SEM/EDS investigations were performed under low vacuum (20 kPa) - variable pressure mode, using the back scattered electron method, which offers very good element contrast and allows the study of non-conducting specimens without applying any conductive coating to avoid charging and contamination and surface alterations of samples through additional processing and handling.  4.5.3 Electrical properties measurements To measure the capacitance of the PZT plates, a GW Instek LCR meter (LCR-821) was used. In online measurements, capacitance measurements were performed with an internal 55  voltage of 0.125 V at a constant frequency of 1000 Hz. The data were collected every 0.896 second. Two modes were used for measuring: C-DF mode for the capacitance (C) and the dissipation factor (DF) of samples, and Z-theta mode for the real part of the impedance (ZRe) of samples. The dielectric constant () of the samples calculated through the equation of the capacitance for parallel-plate capacitors:  = (dC)/( 0a) where d is the distance between the electrodes (1 mm in this work), C is the capacitance of parallel plate capacitors measured by LCR meter, 0 is the permittivity of vacuum, and a is the area of the electrodes (1 cm2 in this work). To measure the capacitance and the dissipation factor of PZT capacitors after water electrolysis, the same equipment and internal voltage was used; however the frequency was not constant, and it was changed between 12 and 200 kHz. To measure the resistance of PZT capacitors, the ZRe-ZIm curves were recorded every 5 minutes during the hydrogen treatment; all ZRe-ZIm curves were semi-circles where the diameter of the whole circle can be considered as the ohmic resistance of PZT. To obtain the  ZRe-ZIm curves (or Nyquist plot), a potentiostat VersaSTAT 4 was used, and the frequency of the applied voltage (1 V) was swiped between 106 and 10-3 Hz.  56  5 Results and Discussion 5.1 High-Pressure conditions (T=100C, p=10 MPa, t=200-1200 hours)  5.1.1 H2 effects on PZT microstructure To address the high pressure hydrogen compatibility of PZT plates, a high-pressure hydrogen treatment was considered as is described in Section 4-2. The results of this study can be used to evaluate the possible effects of high-pressure hydrogen environment on the microstructure of PZT plates used in fuel injectors. The results of this experiment are brought in Sections 5-1-1 (microstructure) and 5-1-2 (electrical properties). Bare PZT Plates Figure 27 shows micrographs of the top surface (10x10 mm) of bare PZT plates before (as-received) and after the exposure to the hydrogen atmosphere for 1,200 hrs at 10 MPa and at 100C; no sample preparation was done before taking the pictures. No noticeable structural changes were observed on the bare PZT plates in Figure 27; however, the higher magnification images of the surfaces show that the grain boundaries of the samples treated for 600 and 1200 hours form clear discontinuities, approximately 100 nm wide. As shown in Figure 28b, the PZT grains in the sample treated for 1,200 hours are separated, suggesting that the grain boundaries are affected by the hydrogen. The XRD results of the as-received and hydrogen-treated bare PZT up to 1200 hours at 100°C did not indicate any new diffraction peaks, Figure 29. However, it was observed that the XRD pattern of the hydrogen treated samples has shifted to lower two-theta angles by about 0.4 degree.  57  bare PZT  Figure 27- Micrographs of PZT surface: a) the as-received sample; b) after 1200 h hydrogen treatment  bare PZT  Figure 28- Surfaces of PZT plates at higher magnifications: (a) before and (b) after 1200 h hydrogen treatment  One of the reasons for such systematic shift could be change of internal stresses during the treatment [61], particularly in the regions close to the surface of the PZT plates.  58  As it can be seen in Figure 27, the grain boundaries on the surface of PZT plates are corroded after the hydrogen treatment, which may affect the internal stress. Furthermore, changes of the lattice parameters due to the dissolution of hydrogen atoms in PZT could also cause shifts in the diffraction pattern [9, 18, 35]. While we were not able to measure the hydrogen content of our samples after hydrogen treatment, it has been reported that high-pressure hydrogen dissolves in PZT to a level of 4-10 at% [43]. (110)  as-received (211)  (220) (022)  50  (112)  40  (210) (102)  30  (200) (002)  (111)  (001)  20  (100)  10  hydrogen treated  60  70  80  90  2 Figure 29- XRD results of bare PZT for as-received and after 1200 h hydrogen treatment  The microstructure of the hydrogen-treated samples was also investigated in the 1x10 mm cross-sections of the plates. Figure 30 shows the cross-sections of the as-received sample and the sample which was hydrogen treated for 1200 hours.  59  bare PZT  Figure 30- Cross-section of the as-received sample (a) in comparison to the cross section of the sample after 1200 h hydrogen treatment (b)  According to this figure, one can conclude that no changes have occurred in the regions far from the surface (below 10 µm depth). In other words, the hydrogen damage appears to be limited to the regions close to the surface, while the center of the sample was not affected (as observed under SEM). This is better shown in Figure 31.  Figure 31- Cross section of the hydrogen treated sample for 1200 hours, close to the surface  60  The general conclusion from the SEM microstructural observations of the bare PZT plates is that the high pressure hydrogen atmosphere would affect the microstructure of the PZT only close to the surface (to 10 µm depth), and no differences were observed in the microstructure of the samples farther from the surface. PZT Plates with Ag Electrodes An image of the unprocessed side face (10x1mm) of a PZT plate with silver electrode after 400 hours of hydrogen treatment is shown in Figure 32, clearly showing a degraded layer formed just adjacent to the electrode. Similar microstructures were also observed for the 200 and 1200 hours hydrogen-treated samples; the degree of damage was proportional to the duration of exposure. The higher magnification image of the degraded layer, Figure 32b, shows a porous structure, where voids are left behind by individual PZT grains detached from the surface of the sample. Ag electrode  PZT  Figure 32- Low magnification (a) and high magnification (b) images of damaged layer on the PZT surface next to the Ag electrode after 400 hours hydrogen-treatment  61  Such a damaged layer was not observed in the cross-section of any of the treated samples with Ag electrodes, as shown for example in Figure 33. The hydrogen treatment conditions for the samples shown in Figure 32 and Figure 33 are the same; however, Figure 32 shows an image from the surface of the sample while Figure 33 shows an image from the  cross section of the sample. As this porous layer formed only on the surface of the PZT plates and only next to the electrodes, we conclude that this damaged layer is probably due to the spillover of hydrogen atoms from the surface of silver electrodes to the surface of PZT. By hydrogen spillover we mean the formation of hydrogen atoms on the surface of the metallic electrode attached to PZT and diffusion of those hydrogen atoms to the surface of the PZT. Therefore hydrogen spillover is just limited to the surface and regions of PZT close to the electrode. This is in accordance to our microstructural observations. This is schematically shown in Figure 34. The spillover of hydrogen from the surface of metals to the surface of oxides has been reported previously in systems such as Pd/SiO2 [62], Pt/Al2O3 [63], Pt/WO3 [64]. Another feature observed on these samples was the detachment of the electrodes from the PZT surface after prolonged hydrogen treatment of 600 hours and 1200 hours, as shown in Figure 35.  62  Ag electrode  PZT  Figure 33- The interface of the Ag electrode with PZT after grinding and polishing, for the as-received sample (a) and for the sample hydrogen-treated for 400 hours (b); no detachment of the electrode from the PZT and no damaged layer are visible  Electrode H2  H  H  H2  H  H  Hydrogen spillover  H H  PZT  H Figure 34- The spillover mechanism of hydrogen atoms from the surface of the Ag electrode to the surface of the PZT  63  Ag electrode  PZT  Figure 35- The detachment of the Ag electrode from the PZT for the sample treated for 600 h  This could be due to the accumulation of hydrogen molecules at the interface of silver electrode and PZT. For the sample treated for 1200 hours, the electrode was detached from the PZT on almost half of the interface. The electrode itself was damaged at the edges, and cracks in the silver electrode were also observed for the sample heat-treated for 1200 hours, Figure 36.  Ag electrode  PZT  Figure 36- Detachment of the Ag electrode from PZT; some cracks are present on the surface of the electrode for the sample heat-treated for 1200 hours  64  PZT Plates with Ag/Pd Electrodes Degradation of the properties of PZT after hydrogen treatment has been reported to correlate with the type of electrode in contact with PZT [9, 21]. It was proposed [21] that H2 dissociates at the surface of the electrode, and H atoms diffuse to the electrode/PZT interface; hydrogen atoms that reached the electrode/PZT interface could further diffuse into the PZT, or re-combine to form molecular H2, leading to local blisters [21]. To investigate the effects of electrode on the degradation of PZT, plates with Ag/Pd electrodes were also evaluated in this work. A micrograph of the unprocessed side face of a PZT plate with Ag/Pd electrodes after hydrogen treatment for 200 hours is shown in Figure 37a; a damaged layer similar to that seen in Figure 37b formed in the vicinity of the electrode. Ag/Pd electrode  Ag/Pd electrode  PZT  PZT  Figure 37- Micrographs of the sample with Ag/Pd electrodes after hydrogen-treated for 200h: the 1x10 mm side face (a) and its cross-section (b)  65  However, the layer was also limited only to the surface of the sample; no damaged layer was visible in the cross-section (Figure 37b), but the electrode itself was degraded – either detached or weakened to the extent that it was destroyed during the sample preparation. The surface of the side of the sample (1x10mm) hydrogen-treated for 400 hours is shown in Figure 38b. Evidence of extensive corrosion is seen, especially in the immediate vicinity of the electrodes (Figure 38b, c). Such a structure is an indication of damage to the grain boundaries, probably caused by the diffusion of hydrogen, possibly involving recombination of atomic to molecular H2 at the grain boundaries. The micrograph of the cross-section of the sample hydrogen-treated for 400 hours is shown in Figure 39 wherein in some parts of the cross section a noticeable portion (about 100 µm depth below the electrode) was corroded after hydrogen treatment. We have not investigated the effect of hydrogen on the PZT plates with Ag/Pd electrodes for 600 and 1200 hours. However, we expect that increasing the time of hydrogen exposure will cause more severe damages to these samples as well.  66  (b) Ag/Pd electrode  PZT  (b) (b) Ag/Pd electrode  PZT  (b) (b) Ag/Pd electrode  PZT  (b)  Figure 38- Surface of the side face of the sample with Ag/Pd electrode: (a) as-received, (b) hydrogen-treated for 400 h; noticeable corroded area next to the electrode (c)  When compared to the not-damaged sample with Ag electrodes (Figure 33b), the depth of the hydrogen damage in the cross section of the sample with Ag/Pd electrodes 67  indicates that the samples with Ag electrodes are more resistant to hydrogen damage. As palladium is well-known for its ability to absorb atomic hydrogen [65], from a practical point of view this means that it is better to use Ag electrodes instead of Ag/Pd electrodes. If the replacement of the electrodes is not possible, the amount of palladium in Ag/Pd electrodes should be decreased as much as possible or other electrodes with less hydrogen reactivity should be used in the manufacturing of the actuators.  Ag/Pd electrode  PZT  Figure 39- Micrograph of the cross-section of the sample shown in Figure 38  5.1.2 H2 effects on the electrical properties of PZT  A typical curve showing changes of capacitance of PZT plates with Ag/Pd electrodes in high pressure (10 MPa) hydrogen atmosphere is shown in Figure 40a. With increasing the temperature from 20C (point 1) to 100C (point 2), the sample capacitance also increases 68  from 1.67 nF (point 1) to 1.98 nF (point 2). This is most likely due to the fact that the switching of the dipoles becomes easier at higher temperature [66]. Figure 40b shows the variations in the capacitance and temperature; the increase of the capacitance follows the increase of the temperature. Figure 41 shows the variation of capacitance of PZT plates with Ag electrodes in argon atmosphere at the same pressure of 10 MPa. The changes of PZT capacitance are similar to those in H2 atmosphere (the same trend for the capacitance variation was also observed for the heat-treatment in air).  2  3  120  (a)  100  1.8  4 1.7  1  Temperture (°C)  C(nF)  2 1.9  (b)  2  80  1.98  60 1.96  40  1.94  20 0  1.6 1  10 Time (hour)  100  2.02  C(nF)  2.1  1.92 0  4  8 Time (hour)  12  Figure 40- Capacitance of PZT sample with Ag/Pd electrode in high-pressure hydrogen atmosphere (a); at point ‘1’ the heater is on, and at point ‘3’ the heater is off. (b): capacitance variation with temperature in hydrogen atmosphere  It seems that hydrogen has no noticeable effects on the capacitance of PZT plates under the conditions of this work (T=100C, p=10 MPa, time=200 hours). The reason why the capacitance drops after reaching the maximum value (point 2 in Figure 40a) could be the  69  increase in the temperature of the samples. The dielectric constant of ferroelectric ceramics shows an aging effect after any abrupt thermal changes, or application of strong mechanical stress [66], due to the rearrangement of the ferroelectric domains. The experiment was done just once for the samples with Ag/Pd electrodes. The error for measured capacitance values is within 0.05% according to the specification of the LCR meter used. Similar results were also seen for the sample with Ag electrodes (Figure 42), in repeated experiments. The general conclusion from the above results is that the dielectric constant of PZT plates will not change in high-pressure hydrogen condition (p=10 MPa, T=100C) after 200 hours. However, in the next section we will see that considerable changes occur in PZT microstructure and electrical properties in high temperature conditions.  2.1  C(nF)  2 1.9 1.8 1.7 1.6 1  10 Time (hour)  100  Figure 41- Capacitance of PZT sample with Ag electrode in high-pressure argon atmosphere  70  2.1  C(nF)  2 1.9 1.8 1.7 1.6 1  10 Time (hour)  100  Figure 42- Capacitance of PZT sample with Ag electrode in high-pressure hydrogen atmosphere  71  5.2 High-Temperature conditions (T=450-600C, p=0.013 MPa)  5.2.1 H2 effects on PZT microstructure To address the kinetics of interactions of hydrogen with PZT plates, a high-temperature hydrogen treatment was considered as it is described in Section 4-2. While for the samples hydrogen-treated in the low-temperature / high-pressure conditions we did not observe noticeable changes in the microstructure of PZT (except for the sample with Ag/Pd electrodes hydrogen treated for 400 h, Figure 39), we observed considerable changes in the microstructure of the samples hydrogen-treated at high temperatures. Figure 43 compares the microstructure of the side surface of the PZT plate with silver electrode in the asreceived condition and after heat treatment at 400 C for 2 hours. According to this figure, noticeable damage has occurred on the surface of the sample hydrogen-treated at high temperatures. Another aspect which was observed in some of the samples was the lead reduction on the surface of some of the samples, in the regions adjacent to the electrodes. This is shown in Figure 44 for the sample hydrogen-treated at 600C for 2 hours, wherein lead particles on the surface of PZT can be observed as the bright-contrasted sub-micron particles. It was considered that these particles were metallic lead based on the much higher ratio Pb/O (14.0), compared to the surrounding PZT particles (2.8) (Table 2). However, no lead reduction was observed in the cross- section of the samples.  72  Table 2- The EDX analysis for the bright particles in Figure 44b  Element Concentration (wt%) Bright particles Concentration (wt%) PZT grains  O  Ti  Zr  Pb  22  6  10  62  6  5  5  84  Ag electrode  PZT  Figure 43- Image from the side surface of the PZT plate with Ag electrode for (a) as-received and (b) after hydrogen treatment (for 2 h / 400C / p= 0.013 MPa) Ag electrode  PZT  Figure 44- Metallic lead in hydrogen treated PZT samples (for 2 h / 600C / p= 0.013 MPa)  73  XRD analysis was also done on the samples to study the possible phase changes in the PZT after hydrogen treatment. To do this analysis, the silver electrodes were removed from the plates after the hydrogen treatment. The results are shown in Figure 45. Comparing the XRD results, one can say that no new peaks have formed and PZT still has maintained its tetragonal structure after hydrogen treatment. (110)  as-received after treatment  50  (211) (112)  (210) (102)  40  (200)  30  (002)  20  (111)  (100) (001)  10  60  2 Figure 45- The XRD pattern for as-received and hydrogen treated PZT with Ag electrodes (for 24 h / 550°C /p=0.013 MPa)  According to Figure 45, a small shift can be observed in the XRD pattern of PZT after hydrogen treatment. Moreover, it can be seen that the relative intensity of some peaks changed after hydrogen treatment. The reason for these changes could be due to the hydrogen dissolution and the formation of lattice defects such as lead and oxygen vacancies [35]. Nevertheless, the important point is that no new peaks have formed in the XRD pattern of PZT samples after hydrogen treatment, indicating that no new phases, detectable by XRD, have formed inside PZT. We also investigated the XRD patterns for other samples treated at  74  500°C and 600°C, and again we did not observed any new peaks after hydrogen treatment (Figure 46). Indeed, other studies about the structural changes in PZT after hydrogen treatment also have not reported the formation of any new XRD peaks [9, 18, 35]. However, changes in the lattice parameters of PZT after hydrogen treatment due to hydrogen dissolution and formation of lattice defects such as lead and oxygen vacancies have been reported [9, 18, 35]. These changes also could be the reasons for the observed small shifts in the XRD pattern after hydrogen treatment (Figure 45).  (110)  600C 500C  50  (211) (112)  (210) (102)  40  (200)  30  (002)  20  (111)  (100) (001)  10  60  2 Figure 46- The XRD pattern for hydrogen treated PZT with Ag electrodes at 500C and 600C  5.2.2 H2 effects on PZT electrical properties Figure 47a shows a typical variation of the capacitance with the temperature and duration of H2 exposure. The hydrogen treatment begins at t=0 in Figure 47 (point c in Figure 23) and, as seen in the inset in Figure 47a (magnifying the effects in the first 40 min of the exposure), the sample capacitance starts changing immediately after the exposure to H2  75  began. The reference test of 1 hour heat treatment in Ar atmosphere did not result in noticeable changes in the PZT resistance or capacitance.  40  40  (a)  300  (b)  2.8  20  40  20  200 32  R (k)  0  ZRe (k)  2 1.6  C (nF)  36  2.4  30  28 100  10  24  0  20  0  400 800 Time (min)  1200  0  400 800 Time (min)  0 1200  Figure 47- (a) The general trend of PZT capacitance variation with time in hydrogen atmosphere at 500°C; (b) measurements for the real part of the impedance (ZRe) and the calculated values of R according to equation (8) (the data is obtained at the constant frequency of 1 kHz)  Figure 47b shows that the variation of the real part of the impedance ZRe (as measured by the LCR meter) corresponds to the variation of capacitance observed in Figure 47a; the secondary Y axis shows the resistance calculated from equation (8). While resistance variation with time shows the same trend as the variation of capacitance, the relative amount of the change of resistance is different from the relative change of capacitance. For example, while after 20 mins the capacitance increases from 2 nF to 2.75 nF, there is a decrease in resistivity to 60% of the initial value (i.e. from 250 k to 100 k ). To examine the trend of resistance decrease, ZRe vs. ZIm curves were re-plotted and a typical result is shown in Figure 48a for the sample hydrogen treated at 550C. The intersection of the semi-circles in Figure 76  48a with the ZRe axis shows the ohmic resistance of the PZT plates [49]. In this way, we measured the ohmic resistance of the PZT plates for different durations of hydrogen exposure and the results are shown in Figure 48b. According to this figure, the resistance decreases at almost the same rate as in Figure 47b, showing the general trend for the resistance variation. The curves in Figure 48 also show the relatively fast decrease in resistance in the first 20 min of H2 exposure. For the sample hydrogen-treated at 550C the resistance decreases from 500 k to 200 k, and for the sample hydrogen-treated at 600C the resistance decreases from 220 k to 100 k after 20 min of hydrogen exposure. The decrease in resistivity after hydrogen treatment is also reported elsewhere [33]. 500  600  400  400  5 mins  R (k)  ZIm (k)  30 mins  0 min 200  300 550 oC  200  600 oC  100  0  0  0  200 400 ZRe(k)  600  0  50  100 150 Time (min)  200  250  Figure 48- The ZRe -ZIm plot for PZT plate heat treated at 550C and the resistance determined using the ZRe -ZIm plots for PZT; the noise in the ZRe-ZIm plots corresponds to the times when the heater was on.  Figure 49 illustrates capacitance change for 530, 550 and 600C, showing a similar trend, which suggests similar structural changes in PZT, although the time until the capacitance reaches the maximum value decreases with increasing the temperature (600 min 77  at 530C, 250 min at 550C, and 40 min at 600C). Therefore, we may conclude that the structural changes affected by hydrogen in PZT are thermally activated processes, as it is proposed in [14]. 80  60  550°C  o  530 C  60 C (nF)  C (nF)  40  40  20  20  0  0 0  200  400 600 Time (min)  800  1000  1000  0  100 200 Time (min)  300  600°C  C (nF)  800 600 400 200 0 0  20 40 Time (min)  60  Figure 49- The variation of PZT capacitance at 530°C, 550°C and 600°C  Based on the above data, it is hypothesized that the variations observed in the capacitance and resistance of PZT are due to the structural modifications caused by the presence of hydrogen in PZT. These structural changes seem to conform to “Isothermal 78  Solid-State Reactions” where the process occurs by the nucleation and the growth of the product nuclei [68]. A typical plot α-time (t) for the solid state chemical reactions is shown in Figure 50 [68], where α is the fraction of the volume converted to the product of reaction. Before reaching point A, the short progress of the chemical reaction happens in the less stable sites of the media in which reaction is occurring; (A-B) step shows the incubation time needed for the development of growth nuclei; (B-C) is the much longer acceleratory period related to the development of the stable nuclei formed in the previous step; in this period new stable nuclei may also form; (C-D) is the step where the further expansion of the nuclei is not possible due to the impingement and consumption of reactant and this leads to the deceleratory or decay period [68]. Kinetics of such reactions can be written in the form of f(α)=kt where k depends exponentially on temperature, k=A exp (-Q/RT), where Q is the activation energy for the process [68-69], and A is a constant. 1.2  completion of reaction  1 0.8    C  D  AB  0.6 0.4 0.2  start of reaction  0  Time Figure 50- The general trend for the isothermal α - time plots having different time steps, time equal to zero shows the start of the reaction [68]  79  “The rate-determining step can be either (i) diffusion, i.e. the transportation of participants to, or from, a zone of preferred reaction, or (ii) a chemical reaction, i.e. one or more bond redistribution steps, generally occurring at a reaction interface” [68]. The model which is frequently used to describe the sigmoid isothermal α – time plots is the AvramiErofeev (A-E) relation, also known as Johnson-Mehl-Avrami-Kolmogorov, or JMAK, equation [68-69]: [–ln (1– α)](1/n)=kt  (28)  or α = 1 – exp (–ktn)  (28a)  where n is a constant. Depending on the nucleation and growth conditions, n can have different values, as summarized in Table 3.  Table 3- Different values of exponents for the equation (28) [68] Model Three-dimensional growth (Spherical particles of reactant) Nucleation rate 1. Constant 2. Zero (instantaneous) 3. Deceleratory Two-dimensional growth (Laminar particles of reactant) Nucleation rate 1. Constant 2. Zero (instantaneous) 3. Deceleratory One-dimensional growth (Lath-shaped particles of reactant) Nucleation rate 1. Constant 2. Zero (instantaneous) 3. Deceleratory  Phase Boundary control (n)  Diffusion control (m)  4 3 3-4  2.5 1.5 1.5-2.5  3 2 2-3  2 1 1-2  2 1 1-2  1.5 0.5 0.5-1.5  80  The above equations are applicable to our condition if we define α as follows: α= [C(t) – C()]/[Cmax – C()]  (29)  where C(t) shows the capacitance at time t, Cmax shows the maximum capacitance, and  is the incubation time (up to point B in Figure 50). Considering the equations (28) and (29) we have to fit the data to the following equation: α= [C(t) – C()]/[Cmax – C()]=1 – exp (–k(t-)n)  (30)  Therefore, to fit our data to the equation (30) we need to find the values for , n, and k. The best fits were obtained with the values reported in Table 4. The results of the fitting for the two temperatures of the 550C and 600C are shown in Figure 51 shows the results of the fitting for all temperatures investigated in this work.  1.2  1.2  (b)  1  1  0.8  0.8     (a)  0.6  0.6  0.4  0.4  0.2  0.2 0  0 0  100 200 t- (min)  300  0  20  40  60  t-(min)  Figure 51- The results of fitting the capacitance data to equation (30) for the temperatures of 550C (a) and 600C (b)  81  5  -5  (a)  (b)  0  550  -5  -10 ln(k)  ln(-ln(1-))  Q=42,433 ± 8087 J/mol  600  500  530  -15  -10  -15      ln(t-)    -20 0.001  0.0011 0.0012 0.0013 0.0014 1/T (K-1)  Figure 52- (a) The results of fitting the capacitance data to equation (30); (b) the activation energy of hydrogen diffusion, obtained from the fit  Temperature 500 C 530 C 550 C 600 C  Table 4- The fitting values obtained for the equation (30)  (sec) ln(k) 1.690.2 9000500 -17.480.72 1.360.2 3700250 -12.910.5 1.810.2 2100250 -15.760.46 1.480.2 45050 -9.990.24  n  R-squared 0.9813 0.9861 0.9863 0.9850  Based on the obtained values for k, the activation energy for the limiting process was obtained to be 0.440.09 eV, Figure 52b. This is close to the reported activation energy for diffusion of H+ in zirconate and titanate perovskite oxides, i.e. 0.44-0.50 eV for BaZrO3-(210)%Y [70], 0.83 0.62 eV for BaZrO3 [71], 0.50 0.22 eV for SrTiO3 [71], 0.420.30 eV for CaTiO3 [71]. Therefore we can hypothesis that the rate-limiting phenomenon for the structural changes observed in PZT is the diffusion of protons. As activation energy for the oxygen diffusion is about 1 eV [71], it appears that protons have the main contribution to the structural degradation of PZT. Comparing the average value of n (1.56) determined in this work with the data in 82  Table 3, one may hypothesize about the nucleation and growth mechanism by which the structural changes occur in PZT (for example, three-dimensional, two-dimensional or one-dimensional growth). During the growth, the coalescence of the developed nuclei and the ingestion of undeveloped nucleation sites may occur, as the coalescence and ingestion are characteristics of the sigmoid isothermal α – time plots [68]. It should be noted that the value of n itself is not enough to confirm the specific nucleation and growth mechanism; some independent confirmation, such as structural observations, are also needed. However, the value of n suggests the rate-limiting phenomenon is hydrogen diffusion, and not its chemical reaction with oxygen. The mechanism responsible for diffusion (or conduction) of protons in oxides, especially in perovskite oxides, is believed to be the Grotthuss mechanism [72], schematically shown in Figure 53.  Figure 53- The Grotthuss mechanism for diffusion of protons in PZT, including the reorientation and hopping of protons between oxygen onions  According to this model, the diffusion of protons consists of (i) transfer from one stable position to another stable position in the structure (hopping) and then (ii) reorientation of the proton for transferring to another site. Therefore, it is probably the proton reorientation and hopping between the different oxygen atoms which leads to the  83  expansion of the new structure formed in PZT. These protons penetrate into PZT and form a solid solution within PZT. Figure 54 schematically illustrates this process. Another explanation for the variation of the electrical properties would be that a new “phase”, with a new crystal structure is forming in PZT, where protons become a part of its chemical composition. Formation of new phases has been reported after H2 exposure for other oxides, such as Sr6Ta2O11, and Ba2In2O5 [73-76]. The formation of the new hydrate composition was reported to occur by a disorder-order phase transition leading to saturated solid solution of the oxide and protons [73]. We however did not see any differences in the XRD spectrum of the as-received and hydrogen treated samples (Figure 45), either because no phase transition has occurred or due to the relatively small volume of such phase, e.g. <0.5 vol% [75].  Figure 54- Hypothetical schematic of the different modes which can be assumed for the dissolution of hydrogen in PZT; (a) where the diffusion of protons into PZT occur uniformly from the surface; in this case the diffusion equation with proper initial and boundary equation could be used for determining the total amount of protons in PZT; (b) where the diffusion of protons can occur from limited places in the PZT; in this case the nucleation and growth models can be used to describe the total amount of protons in PZT  84  To further study the effect of hydrogen on the electrical properties of PZT, we used the dielectric spectroscopy technique. The technique measures the dissipation factor DF and dielectric constant ε of the material in a wide range of frequency (from mHz to MHz) and temperature, which are then used to analyze the data by fitting the results to one of the physical or mathematical models [38] (the relevant relationships between dielectric parameters of materials are listed in Chapter 2-3, equations (21-23)). We investigated the dielectric constant of PZT after hydrogen treatment in the frequency range of 12-200 kHz and in the temperature range of 25-400°C. The data obtained from these experiments were analyzed to determine the effect of hydrogen on the dielectric properties of PZT. We have also used this data to correlate the changes in dielectric properties with the effects of hydrogen on the microstructure of the PZT plates. Figure 55 shows the variations in capacitance (C) and dissipation factor (DF) after hydrogen gas treatment versus temperature at the frequency of 1 kHz. A capacitance peak is observed at 375C, which could be attributed to the phase transition of PZT from cubic to tetragonal phase at the Curie temperature. Because the high dielectric constant of PZT at the Curie temperature is due to the dipoles, it appears that these dipoles are still present in PZT after the hydrogen treatment. However, the capacitance at the peak (20 nF) was only about half of the capacitance before the hydrogen treatment (40 nF), therefore the dipoles inside the PZT are probably affected by hydrogen. The decrease in dielectric constant after hydrogen treatment has been reported before [31]. What should be pointed out is that the existence of dipoles even after hydrogen treatment does not necessarily mean that their 85  directions could be switched with changing the direction of electric field as it proposed by Aggarwal et al. [31]. They suggested that the [OH]− group acts as a fixed dipole, which does not allow switching of the ferroelectric dipoles and domains inside the PZT [31], and therefore PZT may not show polarization hysteresis after hydrogen treatment even if it has a tetragonal structure [31]. Accordingly, the lower value of the capacitance may be attributed  50  5  40  4  30  3  DF  C (nF)  to the interaction of [OH]- dipoles with the dipoles inside the PZT.  R1  20  2  10  1 R2  0  0 100  200 300 400 Temp (°C)  500  Figure 55- Changes of capacitance C and dissipation factor DF of hydrogen-treated sample (for 24 hrs / 550C / p= 0.013 MPa) versus temperature (The thick grey line shows the changes of capacitance for as-received sample)  Figure 55 also shows two relaxation peaks (R1 and R2) in the dissipation factor, while no such peaks were observed for the as-received sample. Generally, there are two conditions which can lead to such relaxation peaks in the dissipation factor. First, when there are dipoles in the dielectric medium and therefore, the dipolar polarization mechanism is active [38] (Figure 56). For this mechanism, the relaxation peak occurs when the natural vibrational 86  frequency of such dipoles coincides with the frequency of the applied voltage [38]. Therefore, one might conclude that the relaxation peaks in Figure 55 may be due to the dipoles formed in PZT after the hydrogen treatment.  +  -  -  +  -  +  -  +  a  +  -  + + + + + + + + + +  b -  -  -  -  +  +  +  +  +  _  _  _  _  _  _  _  _  _  _  _  _  _  _  _  _  +  +  -  -  + -  + -  +  c  -  -  + + + + + + + + + +  Figure 56- Schematics of the dipolar polarization mechanisms, wherein direction of the dipoles changes with changing the direction of applied voltage  Another reason for the occurrence of the relaxation peaks could be the MaxwellWagner (MW) polarization mechanism, which is active in inhomogeneous systems, where the dielectric material consists of regions with different electrical properties (Figure 57). This "extrinsic" polarization mechanism can be explained by considering the heterogeneity of the system without any microscopic polarization process inside the sample [38, 77]. When a voltage is applied across the dielectric medium, due to the differences in electrical conduction in different regions of a heterogeneous material, charge accommodation occurs at 87  the interfaces of the different regions, leading to the increase in the capacitance of the sample. The frequency response of such a system is similar to the frequency response of a Debye relaxor [77], leading to a Debye type relaxation peak. The relaxation time in such heterogenic systems depends on dielectric constant and conductivity of the different regions in that medium. To further understand the physics behind such relaxation peaks, we have changed the frequency of the voltage applied to the samples under the hydrogen atmosphere at different temperatures, ranging from 200C to 325C for the first relaxation peak and from 22C to 42C for the second relaxation peak.  a  + + + + + + + + + +  b  _ _  _ _ _ + + + + +  _  _  _  _  _  _  _  _  _  _  _  _  _  _  _  _  _ _ _ _ _ + + + + +  _ _ _ + + +  c  _ _ _ + + +  + + + + + + + + + +  Figure 57- Schematics of the Maxwell-Wagner polarization mechanism, wherein differences in the electrical properties of different regions cause charge accumulation at the interfaces between the different regions, leading to the increase of capacitance  88  Relaxation Peak #1 (R1) Figure 58 shows the changes of ε´ and ε´´ (where ε´´=DF  ε´) versus frequency in the temperature range of 200-325C, with 25C increments. 8000  3000  (a) 325C  2000  4000  325C  ''  '  6000  (b)  1000  2000 200C  200C 0  0  1  100  10000 f(Hz)  1000000  1  100  10000  1000000  f(Hz)  Figure 58- Variations of ε’ and ε’’ for hydrogen-treated samples with the frequency of applied voltage in the temperature range of 200-325C, with 25C increments  While decreasing the temperature, the frequency at which the relaxation peak occurs also decreases. This may be because the polarization process which has led to such relaxation becomes slower at lower temperatures, as the relaxation time is inversely proportional to the frequency at which the maximum occurs (fMax=1/τ, where τ is the relaxation time). This is typical condition for the dipolar polarization, where dipoles change their positions with changing the direction of the applied electric field [38]. At higher temperatures, such dipoles can change their direction faster and as such, the relaxation peak moves to higher frequencies [38]. Therefore, one may conclude that the first peak observed in Figure 55 is due to the dipoles formed in PZT after the hydrogen treatment. To further investigate the 89  nature of such dipoles (such as the activation energy for the relaxation process), the results should be fitted to one of the physical models of this polarization mechanism. The model most frequently used for the description of such relaxation peaks is the Debye model (Chapter 2-3, pages 32-33, equations (21-22)). We tried to fit our data to the Debye equation, but the fit was poor (Figure 59). This is likely because the Debye equation assumes that the dipoles do not interact with each other, but this is not usually valid for dipoles inside a dielectric medium. 3000  2000  ''  Debye model 1000  0 1  100  10000  1000000  f(Hz) Figure 59- The results of fitting the ε´ and ε’’ data to the Debye equation for at T= 325C, for PZT hydrogentreated samples (for 24 hrs / 550C / p= 0.013 MPa)  A more flexible model, commonly used for modeling dielectric constant data, is the Havriliak–Negami equation [38-39] (Chapter 2-3, page 33, equation (23)). The fit based on the Havriliak–Negami equation is much better than the fitting results based on the Debye equation, Figure 60 (Table 5 compiles the best fit parameters). An iterative MATLAB code was developed and used for fitting procedure.  90  8000  3000  (a)  (b)  6000 2000  4000  HN model  ''  '  HN model  1000  2000  0  0  1  100  10000  1000000  1  100  f(Hz)  10000  1000000  f(Hz)  Figure 60- The results of fitting the ε´ and ε’’ data to the Havriliak–Negami equation for at T= 325C, for PZT hydrogen-treated samples (for 24 hrs / 550C / p= 0.013 MPa)  Temperature 325 C 300 C 275 C 250C 225C 200C  Table 5- The fitting values obtained for the HN equation θ Δε=εs-ε  (sec) 7300 0.0019 0.8 4900 0.0029 0.8 4100 0.004 0.7 3200 0.0078 0.8 2700 0.025 0.8 2000 0.05 1  β 1 1 1 1 0.8 0.6  The activation energy for the relaxation process behind the relaxation peak #1 can be evaluated from the values obtained for . As  is the average residence time of an ion at any given site, it changes with temperature according to =0×exp(-ΔH/RT), where ΔH is the activation energy for ions jumping from one position to another [38]. This equation shows that the kinetics of the relaxation of the system follows the Arrhenius law, i.e. the relaxation time of the system decreases with increasing the temperature (as vibration frequency of ions increases, the probability of ions jumping from one position to another position increases, hence the average residence time or  decreases). The relaxation time fit to the above equation (Figure 61) yields the activation energy of about 0.66 eV. 91  0  ln()  Q=63,760 J/mol -5  -10 0.0016 0.0018 0.002 1/T (K-1)  0.0022  Figure 61- The activation energy for the ion jumping, obtained from the fits, for PZT hydrogen-treated samples (for 24 hrs / 550C / p= 0.013 MPa)  One of the microstructural changes reported for PZT after hydrogen treatment is the lead reduction and owing to that, the presence of lead vacancies in PZT [35]. Indeed we also observed lead reduction in our samples after hydrogen treatment (Figure 44). Therefore, the relaxation peak #1 may be tentatively assumed to be due to the hopping of lead cations in the PZT lattice. However, the reorientation of such dipoles should be relatively slow below 300C, where the diffusion of lead cations is very slow [37]. The activation energy for the diffusion of lead ions in PbTiO3 has been reported to be about 1.89 eV [37], which is significantly higher than the energy obtained here for the reorientation of the dipoles. Therefore it can be concluded that the first relaxation peak is not due to the hopping of lead cations. Another structural change proposed for PZT after the hydrogen treatment is the existence of oxygen vacancies in the lattice of PZT [29, 35], so the relaxation peak could be due to the hopping of oxygen anions in the PZT lattice. However, we believe that the temperature is not high enough for the oxygen ions to have enough mobility in the PZT 92  lattice. Therefore, the dipole reorientation or the relaxation peak cannot be due to the hopping of oxygen anions. Kamishima et al. [40] investigated the dielectric properties of the Yb-doped SrZrO3 after hydration in water atmosphere. While they did not observe any relaxation peaks for the pure SrZrO3, they did observe a relaxation peak for the sample with 1 wt% Yb. For samples with more than 1 wt% Yb, they observed yet another relaxation peak. The activation energy obtained for the first relaxation peak was about 0.58 eV, and they attributed this activation energy to the Yb-OH dipoles. They also attributed the second relaxation peak, observed in samples with more than 1%Yb, to Yb-OH dipoles in the Ybclusters. Therefore, we might tentatively conclude that the relaxation peak which we also observed is due to the dipoles formed by the protons with the dopant (Nb) in the PZT. We realize that more experimentation and analysis is needed to fully confirm this hypothesis.  Relaxation Peak #2 Figure 62 shows the variation of dissipation factor (DF=ε´´/ε´) with frequency in the temperature range 22-47C. When temperature increases, the frequency at which the maximum occurs moves to lower values. If we assume that this relaxation peak is due to the reorientation of dipoles inside the PZT under the applied electric field, then the movement of the peak to the higher frequency values with decreasing temperature means that the hopping of ions becomes slower with increasing temperature. However, this cannot be true, as the vibration of the ions increases at higher temperatures, hence the possibility for their hopping increases. Therefore, we can conclude that this peak is not due to the dipoles inside 93  the PZT. This unusual direct dependency of the relaxation time with the temperature was also  observed  in  other  systems.  For  example,  it  has  been  reported  for  Na58(AlO2)58(SiO2)136mH2O zeolite (NaY) of the faujasite type, and it was assigned to the relaxation of the water molecules confined inside the molecular cages of NaY [78-79]. A similar unusual relaxation process has also been observed for potassium tantalate niobate (KTN) crystal doped with copper, where relaxation occurred below the ferroelectric phase transition [80]. This relaxation process has been attributed to “the reorientation of virtual dipoles provided by the Cu ions hopping between different states of local equilibrium” [81].  0.5 0.4  DF  0.3 22C  0.2  27C 32C  0.1  42C  0 1  100  10000  1000000  f(Hz) Figure 62- Variation of DF with frequency in the temperature range 22-42C, for PZT hydrogen-treated samples (for 24 hrs / 550C / p= 0.013 MPa)  Different explanations could be considered for this non-monotonic relaxation kinetics, e.g. [81] suggests that “this situation usually occurs for ‘small’ systems where relaxing particles become able to participate in the relaxation due to the formation of some 94  ‘defects’ in ordered structure”. Therefore, we hypothesise that such a relaxation peak could indicate the formation of structural defects which hydrogen produces in PZT. Indeed, different structural defects have been proposed for PZT after hydrogen treatment, such as oxygen and lead vacancies, and the formation of [OH]- dipoles [31]. Therefore, the interaction between such defects and [OH]- dipoles in PZT might be the reason for the formation of this relaxation peak. Another explanation for the second relaxation peak and its unusual kinetics behavior could be due to the Maxwell-Wagner polarization mechanism, as first proposed in [81-82]. As mentioned before, this polarization mechanism is active in nonhomogenous systems. In the previous section, we showed that a new structure forms inside the PZT during the treatment in hydrogen atmosphere. Therefore, it can be assumed that the second relaxation peak is due to the formation of a new phase in PZT, with different electrical properties. In other words, this relaxation peak confirms our idea that a new structure with new electrical properties forms inside the PZT during the hydrogen treatment. It is very difficult to pinpoint the exact reason for the formation of the second relaxation peak in the dissipation factor curve of PZT after hydrogen treatment; further study is needed to understand the physics behind the second relaxation peak and the details of the structural changes induced by hydrogen in PZT.  95  5.3 Water-electolysis treatment of PZT  5.3.1 Microstructure The water electrolysis technique was used to charge the PZT samples with hydrogen, following previously reported methodology [33]. In comparison to the treatment in hydrogen gas (which needs elevated temperatures), water electrolysis technique can be completed even at room temperature; this inhibits the de-polarization of PZT samples due to the effects of temperature. Figure 63 shows SEM images of the cross-sections through PZT plates in the region adjacent to the electrode which functioned as the cathode. As seen in Figure 63a, the structure of PZT immediately below the electrode is different from the structure in the center, far from the electrode. Moreover, comparing Figure 63b and c, showing the same magnification images, one can see that the grain boundaries of PZT after water electrolysis are extensively corroded in comparison to the microstructure of the as-received samples.  96  Figure 63- Micrographs of the cross-section through PZT plate after water electrolysis: a) low-magnification image (after 48 hours water electrolysis); b) microstructure of the corroded layer (close to the electrode); c) microstructure in a region far from the corroded layer  According to the Figure 63, a corroded layer was formed just beneath the electrode after the water electrolysis. To see the possible changes in the crystal structure of PZT in the 97  corroded layer, the silver electrode was detached from the surface of the PZT and the XRD test was done on the top surface of as-received sample and on the corroded layer following the treatment. Figure 64 shows the microstructure of the PZT surface right below the electrode (after the electrode removal). As it can be seen, the grain boundaries are extensively corroded, which is in accordance to Figure 63b. According to Figure 65, no new peaks have formed in the XRD pattern of samples after the water electrolysis treatment; therefore, we can conclude that no new phase has formed inside PZT after the water electrolysis, and PZT still has its tetragonal structure after this treatment. However, a systematic shift to lower two-theta values in the XRD pattern of samples after water electrolysis is observed. Moreover, the tetragonal splitting of (100), (200) peaks became more significant after the treatment. Huang et al. have investigated the changes in PZT structure after water electrolysis using the XRD technique, and they observed a very small increase in the lattice parameter of PZT after this treatment [83]. Therefore, the changes observed in the XRD pattern of PZT after water electrolysis in this study can also be due to the dissolution of hydrogen inside the PZT and the changes of the lattice parameters of PZT. It should be noted that Huang et al. did not observe corrosion in grain boundaries of PZT after water electrolysis, as we did in this work. Therefore, the corrosion of the grain boundaries and changes in the microstructure of PZT could cause the difference in the XRD pattern after water electrolysis. Other parameter which should be considered here is the roughness of the surface of samples. After the water electrolysis, the interface between the electrode and PZT was 98  extensively degraded, and the electrode was easily detached from the PZT surface. After removing the electrode, we did not further polished the surface because the corroded layer was very thin, so the surface of the sample which was used for XRD was very rough, as it can be seen in Figure 63. Therefore, the roughness of the surface could be another reason for the changes in the XRD pattern after the water electrolysis [61]. The conclusion from the above discussion is that the hydrogen has diffused trough the crystal lattice and/or grain boundaries of PZT during water electrolysis, without the formation of any new phase detectable by XRD, and PZT still maintained its tetragonal structure after the water electrolysis treatment; however, the hydrogen presence inside the crystalline lattice of PZT could have caused some modifications in the lattice parameters of PZT.  Figure 64- The microstructure of PZT after water electrolysis just beneath the electrode (after removing the electrode)  99  (110)  as-received after treatment (220) (022)  50  (211) (112)  (210)  40  (102)  30  (200) (002)  20  (111)  (100) (001)  10  60  70  80  90  2 Figure 65- XRD pattern of the as-received PZT sample versus the water electrolyzed PZT sample, using the following parameters: I=100 mA/cm2, t=48 hours  The damage to the grain boundaries in the corroded layer can be due to the diffusion of hydrogen atoms from the silver electrode to the grains, or preferably grain boundaries (with more open structure) of PZT, followed by the formation of hydrogen molecules at the grain boundaries [12]. The formation of hydrogen molecules would be accompanied by local increase of pressure, thus stresses along the grain boundaries, which leads to cracks. If it is assumed that the damage in grain boundaries is due to the diffusion of atomic hydrogen predominantly along the grains boundaries, the thickness of the corroded layer can be used for estimating the diffusion coefficient of hydrogen atoms along the grain boundaries of PZT. The thickness of the corroded layer ( ) is proportional to the diffusion coefficient of atomic hydrogen (D) by the relation of  √  where t is the duration of water electrolysis [83].  Because the thickness of the corroded layer was not the same along the cross section of the sample, the area of the corrosion layer was first measured and then the average thickness of the corroded layer was calculated by dividing the area by the width of the cross section (10 100  mm). The results, Figure 66, indicate that the value for the diffusion coefficient of hydrogen in PZT at room temperature is about 9×10-11 (cm2/sec). It should again be emphasized that we believe that this value is related to the diffusion of atomic hydrogen along the grain boundaries of PZT, and not necessarily inside the crystalline lattice of PZT. This is lower than the values obtained by other researchers using the same technique: 4.9×10-8 cm2/sec [83]. This discrepancy in data might be related to the slightly different composition of the samples, or it might be related to the recombination of hydrogen atoms into hydrogen molecules along the grain boundaries. Moreover, the value of D= 4.9×10-8 cm2/sec is obtained based on the advancement of a layer of a different color (yellowish to grey) in the PZT charged in NaOH solutions at room temperatures with a current of 50 mA/cm. In our work, we did not observe any change in color in our samples and we measured the above value based on the thickness of the corroded layer. Therefore, this discrepancy in data might be related to the different interactions between hydrogen and PZT, as proposed by Alvine et al [84]. They recently investigated the hydrogen diffusion inside PZT using the proton nuclear magnetic resonance (1HNMR) and quasi-elastic neutron scattering (QENS) techniques after charging the samples with high-pressure gaseous hydrogen (T=100C, p=32 MPa for 1HNMR analysis, and T=100C, p= 17 MPa for QENS analysis). They obtained different values for the hydrogen diffusion coefficient at room temperature using these techniques; using 1HNMR they obtained D = 6×10-14 cm2/sec, and using the QENS technique they obtained D = 3×10-6 cm2/sec [84]. Because the diffusion results were several orders of magnitude different, they concluded that there were different diffusive processes for hydrogen inside the PZT [84]. 101  Therefore, the discrepancies between different hydrogen diffusion coefficient values could be due to the different interactions occurring between the hydrogen and PZT.  0.3 25°C  x (mm)  0.2  0.1  0 0  200 400 1/2 1/2 time (sec )  600  Figure 66- The thickness of the corroded layer versus the square root of time of water electrolysis  5.3.2 Electrical properties of PZT exposed to water electrolysis  It is reasonable to assume that the changes which occur in the microstructure of PZT will affect the dielectric properties of PZT as well. Figure 67 shows the variation of the capacitance and dissipation factor of PZT capacitors during water electrolyzes at the frequency of 1 kHz, showing that both parameters increase with time of water electrolysis. The increase in capacitance and dissipation factor after water electrolysis has also been reported in other studies [85-89].  102  1.8  0.024  1.7 DF  C(nF)  0.02 1.6 0.016 1.5 0.012 1.4 0  20  40  60  time(hours) Figure 67- The changes of capacitance (C) and dissipation factor (DF) versus the duration of water electrolysis at the frequency of 1 kHz  To further study the effect of water electrolysis on the electrical properties of PZT, the capacitance and dissipation factor of the treated samples were measured versus the frequency of applied voltage. Figure 68a shows the variation of capacitance (C) and dissipation factor (DF) versus frequency for the PZT samples before and immediately after water electrolysis, for 6 hours. The capacitance right after water electrolysis was higher than the initial value. The same trend is also observed for the dissipation factor above 200Hz. After aging for 24 hours in air however, the capacitance decreased significantly below the initial values (e.g. at 103 Hz, the capacitance values were 1.7 nF, 1.8 nF and 0.2 nF for the asreceived, hydrogen-treated for 6 hrs, and 24 hrs aged samples respectively), as also shown in other researchers’ studies [33]. The dissipation factor was however higher for the aged sample versus the hydrogen-treated sample, for frequencies below 20 kHz. While for the as103  received sample the dissipation factor was almost constant, after hydrogen treatment and aging, it increases with decreasing frequency, as seen in Figure 68b. 2.5  0.3  (a)  (b)  0.2  1.5 DF  C(nF)  2  1  0.1  0.5 0  0 1  100  10000  1  1000000  100  10000  1000000  f(Hz)  f(Hz)  Figure 68- Variations of electrical properties after 6 hours water electrolysis and subsequent aging in air: a) capacitance (C); b) dissipation factor (DF) (: as-received,: after water electrolysis, : after aging)  Another interesting finding is the relaxation peak observed in the sample after water electrolysis for 10 hours, and after 24 hours aging in air (Figure 69). 2.5  0.3  (a)  (b)  0.2  1.5 DF  C(nF)  2  1 0.1  0.5 0  0  1  100  10000 f(Hz)  1000000  1  100  10000  1000000  f(Hz)  Figure 69- Variations of electrical properties after 10 hours water electrolysis and subsequent aging in air: a) capacitance (C); b) dissipation factor (DF) (: as-received,: after water electrolysis, : after aging)  104  The formation of the relaxation peak is better seen in Figure 70, which shows the changes of dissipation factor and capacitance for the sample which was water-electrolyzed for 48 hours. According to Figure 70, no relaxation peak in DF was observed for the samples tested immediately after water electrolysis; however after aging for 10 hrs, a relaxation peak starts to form (at about 105 Hz), and can be clearly observed at about 103 Hz after ageing the samples for 24 hrs. The relaxation peaks in the dissipation factor have also been reported for other oxides [85-88], but not for PZT. 0.5 (a)  (b)  2  0.4  1.5  0.3  DF  C(nF)  2.5  1  0.2  0.5  0.1  0  0 1  100  10000 f(Hz)  1000000  1  100  10000  1000000  f(Hz)  Figure 70- Variations of capacitance (C) and dissipation factor (DF) after water electrolysis for 48 hrs, and subsequent aging at room temperature in air (■: as-received, : after water electrolysis, ▲:after 10 hours aging, : after 24 hours aging)  Considering the previously published results, a few issues need to be addressed: the first issue is why does the dielectric constant increase during water electrolysis, and then decreases after aging in air. It should be noted that the increase in the dielectric constant during water electrolysis and further changes in capacitance during aging were also reported 105  for oxides such as TiO2, SrTiO3, BaTiO3, CaCu3Ti4O12, BiFeO3, and WO3 [85, 88]. Chen et al. proposed that the increase in dielectric constant could result from the dipoles formed by the protons inside the oxide [86], related to the complexes formed by protons with structural defects such as oxygen vacancies. It was assumed in [86] that during the aging in air, such protons leave the oxide and this leads to the recovery of the electrical properties, or to decrease of the dielectric constant, as it was observed for BaTiO3. Another mechanism which could be considered for the increase of the dielectric constant after water electrolysis is in accordance to that proposed in the work of Park and Chadi [30]; they investigated stable sites of protons in PbTiO3 using first-principles calculations. For the tetragonal phase of PbTiO3, their results show that “the direction of the [OH]– dipole is favorably aligned with the host polarization” [30]. Thus [OH]– should enhance polarization of the spontaneous dipoles in PZT, and therefore this could be the reason for the increase in the dielectric constant of PZT after hydrogen treatment. As the bond between proton and oxygen is strong, protons attached to oxygen atoms cannot easily diffuse out from the PZT bulk. Therefore, it might be concluded that the [OH]– dipoles could not be the reason for the increase in the dielectric constant right after water electrolysis. The mechanism which we propose here for the increase of the dielectric constant, not mentioned in the previous studies [86-89], is the Maxwell-Wagner (MW) polarization, active in inhomogeneous systems where the dielectric material consists of regions with different electrical properties [77]. As mentioned before (Chapter 5-2-2), this "extrinsic" polarization mechanism can be explained by considering the heterogeneity of the system without any 106  microscopic polarization (dipolar polarization) process inside the sample [38, 77]. Due to the differences in electrical conduction in different regions of such heterogeneous material, charge accommodation occurs at the interfaces, leading to increase of capacitance of the sample. According to Figure 63a, although a corroded layer formed beneath the electrode, most of the sample was unaffected. Therefore, the assumption of an inhomogeneous dielectric medium after water electrolysis is reasonable for our samples. As such, we believe that the increase in capacitance right after the water electrolysis is due to the formation of the corroded layer, and to the difference in the electrical properties of this layer and the unaffected layer of PZT. Furthermore, during the aging, hydrogen atoms diffuse out from the corroded layer, and this leads to further changes in capacitance. The formation of the relaxation peaks could also be explained by the MW polarization mechanism. The microstructure of PZT after water electrolysis can be considered as a dielectric made of two different layers perpendicular to the electric field. It can be shown that the frequency response of such layered system is similar to the frequency response of a Debye relaxor [77], leading to a Debye type relaxation peak. In our opinion, this is the main reason for the observation of the relaxation peak in the dissipation factor. The relaxation time () can be evaluated as  = (C1+C2)/(G1+G2) where C1, C2 are the capacitance, and G1, G2 are the conductance of corroded and un-affected layers, respectively [77]. Accordingly, the frequency at which the relaxation occurs (f=1/) depends on the electrical properties of the different regions, and variations in these electrical properties will change the relaxation time and the frequency at which relaxation occurs [38, 77]. We 107  propose that the changes in the capacitance and the movement of the relaxation peak to lower frequencies with sample aging could be explained by the MW polarization mechanism. When hydrogen atoms diffuse out of the PZT bulk during aging, the electrical properties of the corroded layer change, and this changes the frequency at which the relaxation occurs. The relaxation peak was however not observed for the samples exposed for <10 hours to water electrolysis, likely because of the insufficient thickness of the corroded layer. Chen et al [86] have proposed that the relaxation peak is due to dipoles related to the complexes which protons form with structural defects in the oxide. If this is the case, the intensity of the relaxation peaks should decrease with aging, when the protons diffuse out of the sample and the number of dipoles inside the oxide decreases. However, the intensity of the relaxation peaks increases as more hydrogen diffuses out from the sample while aging continues (Figure 70), suggesting that the relaxation peak is not due to the dipoles, but rather is due to the MW polarization. Figure 71 shows the poor fit of the Figure 70 data to the Debye equation, possibly because the thickness of the corroded layer was not uniform, leading to a distribution of relaxation times (as it depends on the thickness for the corroded layer). Figure 71 shows that the Havriliak–Negami equation (Chapter 2-3, equation 23) fits well our experimental results, for ε = 450, εs = 2500,  = 0.012 sec-1, θ = 0.9, and β = 0.4.  108  1000 Havriliak-Negami  ''  800 600 400 200 0  Debye  10  103  105  f (Hz) Figure 71- The results of fitting the ε´´ (ε´´=DFε´) data to Debye and Havriliak–Negami equation. An iterative MATLAB code was developed and used for the fitting procedure.  Typical trends for the changes of the dielectric constant and dissipation factor of a leaky capacitor (due to the electronic conduction) are shown in Figure 72a. Figure 72b also shows the typical trend for the changes of the dielectric constant of a capacitor with mobile charges which can move by hopping.  Figure 72- The changes of the capacitance (C) and dissipation factor (DF) for (a) a leaky capacitor with electronic conduction, (b) for a capacitor with hopping charge carriers adapted from [38]  109  The trends shown in Figure 72 are similar to the trends observed for our samples, aged in room conditions after water electrolysis (i.e. compare Figure 68 and Figure 72). This trend is not obvious in Figure 70 because of the existence of the relaxation peak, and because of the limits on the frequency that can be used. Therefore, it can be concluded that the ohmic resistance of the PZT plates decreases after being charged with hydrogen. The increase in the electrical conduction of PZT after water electrolysis is also reported in other studies [33, 83]. Different mechanisms could be considered for the decrease in the resistivity of the samples. The first mechanism relates it to the formation of oxygen vacancies (2H + O2- → H2O +  ); ionization of the vacancies (  ) will contribute up to two electrons available for  conduction [89]. Another possible mechanism includes the ionization of hydrogen atoms inside the lattice of PZT (H → H+ + e-) [31], with the electrons produced available for conduction through hopping [31]. While these mechanisms can explain the increase in the dissipation factor, they cannot explain why the dielectric constant decreases after water electrolysis. The reason could be damage to the grain boundaries of PZT, or reaction of protons with PZT and [OH]– dipoles formation. These dipoles could hinder the switching of the dipoles inside PZT, and therefore affect the movement of the domain walls, and consequently decrease the dielectric constant of PZT [82]. The disappearance of switchable polarization hysteresis of PZT after hydrogen treatment has been also attributed to the formation of [OH]– dipoles, which inhibits the switching of the spontaneous dipoles in PZT [88]. Protons bonded with oxygen ions cannot easily leave the PZT bulk during aging, i.e. 110  high temperature treatment (>700C) is needed [47]. Thus the decrease in capacitance demonstrated in our experiments is likely due to the persistence of [OH] – dipoles within the PZT structure, obstructing the movement of the PZT domains, and the formation of oxygen vacancies. The decrease of dielectric constant of BaTiO3 at high frequency after water electrolysis was previously linked to interstitial hydrogen on the domain walls of BaTiO3 [88].  111  5.4 Ceramic Coatings for PZT Damage Protection  5.4.1 Alumina coatings microstructure Attempts have been made in the past few years to solve the issue of hydrogen damage in PZT during the forming gas annealing [55]. In this section, we investigated the possibility of coating the PZT with alumina using the sol gel technique. Alumina has been proposed before as a hydrogen barrier layer for PZT, and it has been shown that the alumina layer can successfully act as a hydrogen barrier layer [55]. However, the method which we propose here is the simple sol-gel method (as described in Chapter 4-4), which is different from the method used in the former works for alumina deposition. In this section we show that while the developed coating is porous, it can still significantly decrease the amount of hydrogen damage to PZT. Figure 73a shows the surface of the coating obtained by dip coating with pure Boehmite sol, before drying; some excessive sol accumulation occurred on the surface of PZT, probably because of the surface porosity on the PZT. According to Figure 73b, c, cracks and coating detachment were observed in the places where sol accumulation occurred after firing. Generally, the quality of the coating obtained with pure Boehmite sol was not sufficient to demonstrate its effect on H2 damage protection of PZT.  112  Figure 73- Low magnification image of the coating on the surface of PZT (a) after dip coating with pure boehmite sol (b) before and (c) after heat treatment of the coating in the furnace, in some places on the surface of the coating, detachment of the coating was observed  113  One of the ways to improve the quality of the coating is the addition of PVA to the sol, which increases viscosity of the sol [90]. The microstructures of the coating processed with 10wt% PVA are shown in Figure 74. Figure 74 shows the surface of the coating directly after dip coating (before drying it in the furnace); it can be observed that now a smooth uniform coating has formed on the surface of PZT. Comparing the Figure 73 with Figure 74, one can see that the quality of the coating was noticeably enhanced. The small pits which can be seen on the coating are due to the pores on the surface of the PZT. As it will be shown later, the sol had enough fluidity to fill out the pores on the surface of the PZT, but not completely, so small dimples formed the surface of the coating.  Figure 74- Low magnification image of the coating on the surface of PZT after dip coating before the heat treatment of the coating in the furnace (comparing with Figure 73a, a smooth uniform coating has formed on the surface of PZT with the addition of PVA to sol)  Figure 75 shows the surface of PZT plates after firing the coating in the furnace; a smooth crack-free coating developed on the surface of the sample. It should be mentioned that during the dip coating process, excessive accumulation of the sol occurred at the bottom 114  edge of the sample, and because of this, some cracks were observed at the very end of the sample; however, the center of the sample was crack-free. The darker areas seen in Figure 75b are pores on the surface of PZT, just covered by the very thin (1-2 m) coating.  Figure 75- Low magnification (a) and high magnification (b) images of the coating on the surface of PZT after dip coating and after heat treatment of the coating in the furnace (comparing with Figures 73b and c, a smooth uniform coating has formed on the surface of PZT with the addition of PVA to sol)  Figure 76 shows the cross section of a sample with alumina coating after three consecutive depositions: there is an evidence of good adhesion between the PZT and the coating. Furthermore, no cracks were observed in the coating, and there were no pores or cracks at the interface between the alumina layer and PZT. It should be noted that there were some cracks in the coating at the bottom end of the sample, where sol accumulation occurred. As it can be seen from Figure 76b, the sol had good fluidity and surface wettability, so it was able to fill the holes and grooves on the surface of the PZT plates.  115  Figure 76- Low magnification (a) and high magnification (b) images of the cross section of the alumina coating. As it can be seen from (b), the coating had enough fluidity to fill out the pores on the surface of PZT  Leenaars et al. have measured the amount of porosity of Boehmite coatings after treatment at different temperatures [91]. The porosity and the size of pores are presented in Table 6, which shows that even after firing as high as 1000C, the coating still contains a noticeable amount (>40 vol%) of porosity. They also have found that the prolonged heat treatment for 850 hours at 400C did not change the amount of porosity. Similar results have also been reported in other works [92], therefore, we also expect the coating to be porous although we did not measure porosity of the coatings processed in this work.  116  Table 6- Microstructural charachteristics of alumina coatings a a function of TC [91]  As seen from the high magnification image of the cross-section shown in Figure 77, the alumina coating after firing is a conglomerate of seemingly separate agglomerates in the range of about 20-50 nm, with <10 nm particles within the agglomerates. According to the information provided by the manufacturer of the commercial Boehmite powder used in this work, each of these particles are actually agglomerations of a few individual crystalline particles (the average size for agglomerated Boehmite particles is 25 nm and for individual crystallites is 4.5 nm [57] ). The important point which can be taken from Figure 77 is that the coating is porous, with inter-agglomerate pore sizes < 100 nm. As the alumina coating is a meso-porous medium, we expect that it will not prevent the access of molecular hydrogen (H2) to the PZT surface; however it may prohibit or decrease the diffusion of atomic  117  hydrogen (H). To further understand the microstructure of the coating after the heat treatment, XRD analysis was performed to identify what phases were present.  Figure 77- High resolution image of the cross section through the alumina coating processed at 450C in air for 5 hours  Figure 78 shows the transformation sequences of Boehmite with the firing temperature [93], suggesting that after heat treatment at 450C, the structure of the coating is -alumina. However, XRD analysis on PZT plates with thin alumina coatings (< 5 um), did not detected any alumina, and only PZT peaks were identified. Consequently, to understand the crystal structure of the coatings, 10 g bulk sample of Boehmite powder was heat-treated at 450C in air for 5 hours (i.e. following the same heat treatment procedure used for the coatings) and then the crystal structure was investigated by XRD, Figure 79. The peaks are 118  identified as -alumina [94] and they are not very sharp, which indicates a low degree of crystallinity and a fine particle size distribution.  Figure 78- Transformation sequence of the different aluminum hydroxides with temperature (adapted from [93]). (400) (311) (222) (220)  (440)  as-received after treatment (511)  (111)  (444)  10  20  30  40  50  60  70  80  90  2 Figure 79- XRD results for the as-received boehmite powder and after heat treatment at 450C for 5 hr  It should be pointed out here that the -alumina has a spinel structure, but the chemical formula for -alumina containing hydrogen (H) has been the subject of debate [95].  119  A recent theoretical study by Sohlberg et al. has shown that the -alumina can exist over a range of H content and the chemical formula for -alumina can be presented as H3mAl2-mO3 [95]. This theoretical chemical formula for -alumina has been confirmed by the available experimental data for the -alumina structure [95]. It should be pointed out that there are different types of hydrogen atoms positions in the structure of the -alumina particles [95]. H atoms present in the bulk of the -alumina particles occupy octahedral and tetrahedral sites. Additionally hydrogen atoms can be present on the surface of the -alumina particles, without specific preference towards lattice positions [95]. These hydrogen atoms have different mobilities and thermal stabilities in the structure of -alumina, and therefore they play different roles in the properties of -alumina [95]. According to the elemental compositions by EDX shown in Table 7, the Al/O ratio (0.75) in the coating is lower than the theoretical ratio Al/O in Al2O3 (1.12). The lower value of Al/O ratio could be because of the OH bands in the structure of the -alumina. Table 7- The EDX analysis for the -alumina coating (high concentration of Au is due to the gold coating on the sample for SEM analysis) Element O Na Al Cl Ti Au Concentration (wt%) 43.59 1.76 32.86 0.96 1.83 21.27  5.4.2 Hydrogen resistivity of alumina-coated PZT To assess the hydrogen resistivity the coated PZT, water electrolysis technique was used. Thin (10 nm) Au-Pd electrodes were sputtered over the alumina-coated PZT plates, as well as on as-received PZT, for reference tests. Figure 80 shows the cross section of the reference sample, without alumina coating, after water electrolysis at room temperature and  120  the following test parameters: I= 100 mA/cm2, V= 6-10 V, time= 24 hours, 0.1 M NaOH solution; a corroded layer has formed close to the top surface of the sample (the metallic electrode cannot be seen in this figure because it is only about 10 nm thick). It can be also seen that the grain boundaries of PZT close to the electrode are extensively corroded, which is not the case in the center of the section far from the electrode. Au-Pd electrode PZT  Figure 80- The cross section of the sample with Au-Pd electrodes and after 24 hours water electrolysis. The thickness of the corroded layer is about 100 microns  Figure 81 shows the cross section of the sputtered sample with alumina coating after 48 hours of water electrolysis (and all other parameters same as these used to produce the sample in Figure 77). The effect is quite dramatic - the corrosion beneath the electrode is limited to < 5 m surface film, which suggests that the porous alumina layer deposited by the sol-gel technique can act as an effective hydrogen barrier layer. 121  Au-Pd electrode  coating PZT  Figure 81- The cross section of the sample with Au-Pd electrodes and alumina coating and after 48 hours water electrolysis  Because the damage to PZT is due to the diffusion of hydrogen atoms (H) into PZT [9, 14, 18-21], we can conclude that the alumina coating has blocked or decreased the diffusion of hydrogen atoms from the metallic electrode to the surface of PZT. However, as it was observed in Figure 77 (and confirmed through [91]), the coating is  40 vol% porous, and thus molecular hydrogen could easily pass through it. Therefore, the question which may arise is how this coating decreases the damage to PZT, despite allowing H2 access to PZT surface. Several hypotheses could be formulated in this regard. First consider the possibilities of the reaction of hydrogen atoms with the coating (refer to Figure 82 for the schematic illustration of this possibility). Joubert et al. have investigated the reaction of - alumina dehydrated at 500C with hydrogen and concluded that hydrogen can react and be absorbed 122  on the surface of the - alumina particles at temperatures as low as 25C [96], due to the defective surface structure of -alumina [96]. On the other hand, Yu et al. have investigated the surface diffusion of hydrogen atoms on the surface of - alumina and they observed noticeable surface diffusion at temperatures higher than 250C [97]. Therefore, one mechanism by which the coating could decrease the amount of hydrogen damage could be reaction of the -alumina coating with the hydrogen atoms. In other words, - alumina coating could act like a “sponge” absorbing hydrogen atoms before they reach the surface of the PZT (Figure 82). It should be noted that the hydrogen atoms absorbed on the surface of - alumina particles can diffuse along their surfaces and it is anticipated that sooner or later damage should start to PZT when hydrogen atoms reach the surface of PZT. If this scenario is true, then we expect that after a prolonged time of water electrolysis, damage to PZT should occur. Au-Pd electrode  coating PZT  H2  H2  1 H  H  H  electrode  H  H  H  2 H H  H H  coating ɣ-Al2O3  Figure 82- Schematic for the reaction of hydrogen atoms with -alumina particles 1) transformation of hydrogen atoms to hydrogen molecules which leave the system away from the coating (i.e. as hydrogen bubbles during water electrolysis), 2) diffusion of hydrogen atoms through the electrode and attachment to -alumina particles, followed by surface and bulk diffusion through -alumina towards PZT  123  To examine this theory, we made a thinner alumina coating (1 m thick, produced through single-dip-coating process, Figure 83), and extended the time of water electrolysis to 144 hours (6 days), at conditions same as before (I= 100 mA/cm2, V= 6-10 V, room temperature, 0.1 M NaOH solution).  Figure 83- An image of the cross section of PZT with alumina coating on top  The microstructure of PZT after water electrolysis for these conditions (Figure 84) shows that the PZT is somewhat damaged in regions close to the electrode (to the depth of about 10-20 m). Therefore, it may be concluded that the reaction of the alumina coating with the hydrogen is a possible mechanism by which the presence of the coating decreases and delays the hydrogen damage. However, another explanation to be considered is the transformation of hydrogen atoms to hydrogen molecules after leaving the metallic electrode, schematically shown in Figure 85. As palladium is well-known for its hydrogen catalytic activity, hydrogen atoms (H) could easily transform into hydrogen molecules (H2) on the surface of the electrode at the interface between the electrode and coating (Figure 85).  124  Au-Pd electrode  coating PZT  Figure 84- The cross section of the sample with Au-Pd electrodes and thin alumina coating and after 144 hours water electrolysis  Au-Pd electrode  coating PZT  H2  electrode  H  H  2 H  H  H  H2 H  1  H  H  H H coating  H2 ɣ-Al2O3  Figure 85- Schematic image for the combination of hydrogen atoms at the interface of metallic electrode with - alumina  125  Such hydrogen molecules released on the interface of the alumina coating with the electrode can further diffuse through the pores of the coating and reach the surface of PZT. Therefore, it may seem that damage could occur to PZT as alumina coating cannot block the hydrogen molecules. However, as proposed in [9], and confirmed in this work (Chapter 5-11), hydrogen molecule itself cannot damage the PZT below 400C [9]. That is because hydrogen molecules cannot dissociate on the surface of PZT at low temperatures [9]. It might be concluded therefore that the transformation of hydrogen atoms to hydrogen molecule on Pd surface, and physical separation of the PZT surface from the electrode surface (by the porous alumina) is likely another mechanism by which the alumina coating (despite its porosity) decreases the amount of hydrogen damage of PZT. An additional point which should be mentioned here is that the metallic electrode could be porous. If this is the case, then water could diffuse inside the alumina coating during water electrolysis and thus fraction of the hydrogen molecules released on the electrode surface dissolve in the water (the solubility of hydrogen gas in the water at room temperature is 0.8 mole/liter [99]). Therefore hydrogen molecules would be in contact with the surface of PZT in water, which is also the case in the samples without alumina coating. However, because the damage was demonstrated to be considerably lower in the coated PZT samples, we conclude that the degradation during the water electrolysis is due to the hydrogen atoms that diffuse from the metallic electrode into PZT, and not to the presence of the molecular hydrogen in water.  126  The conclusion that can be drawn from the above experiment is that as far as there is no interaction between atomic hydrogen (H) and PZT, no damage would occur to PZT. If the direct contact between the metallic electrode and PZT can be diminished or decreased, such that no hydrogen atoms will be in contact with PZT, then and no damage would occur to PZT. Through such mechanism, even a porous coating between the electrode and PZT can noticeably decrease the amount of damage. It is possible that by applying this method in the manufacturing of PZT actuators, even with other types of porous coatings, the deleterious effect of hydrogen would be greatly diminished. It should be however remembered that a thin insulating surface layer might degrade functionality and performance of electrode/PZT/electrode assembly [100]. Such layer would act as a capacitor in series with the PZT (Figure 86), and as a result the externally applied voltage would be distributed across the sample inversely proportionally to the capacitance of each layer. Therefore, higher voltages would be needed to drive the PZT actuator. To solve these issues two solutions could be considered: (1) make the coating layer as thin as possible, and/or (2) make the coating layer from materials with dielectric constant higher than PZT. In both these cases, the capacitance of the coating layer will be higher than the capacitance of the PZT layer, and therefore the voltage across the PZT layer would be almost the same as the applied voltage to the whole assembly.  127  PZT  C PZT  coating  PZT  C coat  C PZT  C coat  Figure 86-Equivalent electrical circuit for PZT and PZT with coatings  There is very recent research supporting the idea of introducing a coating between the electrode and PZT [100]; this work showed that a thin alumina layer between the PZT and electrode will not affect the functionality of the PZT capacitors. The reason for this was not clear; however it was suggested that a thin alumina layer between the electrode and PZT will not act as an insulator and it may act as a resistor [100]. That is because when a voltage is applied across the sample with alumina coating, a considerable fraction of the applied voltage would be across the thin alumina layer, whereas the layer would become conductive under high voltages by either Schottky emission or thermionic field emission [100].  128  6 Conclusions High-pressure hydrogen treatment The microstructural and capacitance changes in the PZT ceramics exposed to high-pressure (10 MPa) hydrogen atmosphere at 100C were investigated in this part of the work. For bare PZT, no noticeable damage was observed to the PZT structure after hydrogen treatment for up to 1200 hours. The grain boundaries of PZT were corroded only in the regions just below the surface (about 10 µm deep) of the samples. In samples with Ag electrodes, the presence of metallic electrodes greatly increases the damaging effects of hydrogen on PZT. The structural degradation observed in the samples with Ag electrodes consisted of the development of a very porous layer adjacent to the electrodes on the surface, and the detachment of the electrodes from PZT. It was proposed that the hydrogen spillover is the responsible mechanism for the formation of the porous electrode on the surface of the samples. In PZT samples with Ag/Pd electrodes, the PZT damage noticeably increased compared to Ag-only electrode. It is therefore suggested to decrease the amount of Pd in the Ag/Pd electrodes to increase the resistance of the actuators to hydrogen damage. No considerable changes were observed in the dielectric constant of PZT after 200 hours hydrogen treatment at 100°C.  High-temperature hydrogen treatment The kinetics of the PZT structural modifications due to hydrogen exposure was investigated by online monitoring of the electrical properties of PZT above Curie  129  temperature, up to 650C. Considerable changes were observed in the microstructure of the PZT samples hydrogen-treated at high temperatures, including the detachment of single PZT grains from the surface, as well as reduction to metallic lead on the surface of the samples. It was found that the changes in PZT exposed to high-temperature H2 can be described by a simple nucleation and growth model. Assuming that the changes are controlled by protons diffusion, the resulting activation energy for the diffusion of protons in cubic PZT was determined to be 0.440.09 eV.  The dielectric spectroscopy study of PZT samples shows that even after the high temperature hydrogen treatment, Ti-O and Zr-O dipoles are still present inside the PZT. The results show two relaxation peaks in the dissipation factor curve of the hydrogen-treated PZT. While the first peak indicates that the kinetics obeys the classical Arrhenius law, with the activation energy of 0.664 eV, the second peak indicates the presence of unusual relaxation kinetics: the relaxation time increases with increasing temperature. This nonmonotonic relaxation kinetics can be attributed to the defects that hydrogen has produced inside the PZT, or it can be due to the Maxwell-Wagner polarization mechanism.  Water electrolysis treatment The interaction of hydrogen with PZT in the tetragonal phase was investigated using the water electrolysis technique. Development of a hydrogen-affected (corroded) layer adjacent to the electrode functioning as the cathode was observed during water electrolysis. 130  The thickness of the corroded layer was used to calculate the diffusion coefficient of hydrogen atoms in PZT, and the value obtained was 9×10-11 cm2/sec. A composite model was proposed for the microstructure of PZT affected by hydrogen generated during water electrolysis, and changes of the electrical properties of PZT are linked to the model. The Maxwell-Wagner polarization mechanism was proposed to be responsible for the changes in the dielectric properties of PZT after hydrogen charging. Although this polarization mechanism has been ignored in previous works by other researchers, we believe it is responsible for the variation in electrical properties of other oxide ceramics during water electrolysis as well. Furthermore, the results indicate that after aging, the resistivity and high-frequency dielectric constant of PZT decrease. The decrease in capacitance is expected to be due to [OH]– dipoles hindering the movement of PZT domains.  Alumina coatings for PZT protection from hydrogen In this part of the research we have investigated the possibility of protecting PZT from hydrogen damage by coating it with alumina using the sol gel technique; subsequently we have assessed the hydrogen resistance of the coated PZT. The results show that the quality of the alumina coatings obtained with pure boehmite sol was not very good, i.e. cracks and coating detachment were observed. However, the addition of the poly-vinyl alcohol (PVA) to the boehmite sol considerably enhanced the quality of the final coating; neither cracks nor detachment were observed. The hydrogen resistance of the alumina coated PZT was investigated using the water electrolysis technique, and the results have 131  shown that the alumina coatings noticeably decrease the level of hydrogen damage to PZT. It is anticipated that the main contributor to the decreased PZT damage is the physical separation of the metallic electrode from PZT by the coatings. The insulation of the PZT from the electrode leads to the re-combination of hydrogen atoms into molecules on the electrode surface and within the pores of the coating, which effectively prevents access of the damaging atomic hydrogen to the surface of PZT.  Impacts of the work The results of the study of long-term high-pressure gaseous hydrogen exposure of PZT could be beneficial to the design of the modern electronic fuel injectors that use PZT actuators for valve opening, instead of the conventional solenoid technology. The results show that the metallic electrode has considerable effects on the level of hydrogen damage, and it was suggested that Ag/Pd electrodes should be replaced with pure Ag electrodes in making such actuators. Kinetics of the hydrogen damage to PZT was also investigated in the present work. The results could be used for the prediction of the degradation caused by the hydrogen treatment. The results could be used for the re-design of the hydrogen treatment process of FeRAMs. We also have shown that even a porous separation layer between an electrode and PZT acts towards decreasing the PZT hydrogen damage. The results of this part of the research could be further used to better understand the mechanism of PZT degradation by hydrogen, and to design new methods for decreasing the hydrogen damage to PZT and other ceramics. 132  7 Future Work We suggest that the technique we used in this work for studying the kinetics of interactions between hydrogen and PZT can be also used for studying the kinetics of the interaction between hydrogen and other oxides, in particular oxides designated as the possible fast proton conductors replacing polymeric proton conductors in fuel cells. These oxides include BaCeO3, BaZrO3, and SrCeO3. The proposed nucleation and growth model for the structural changes in PZT could be also valid for other oxides, so a general model could be evaluated for studying the kinetics of interactions between hydrogen and oxides in hydrogen atmosphere.  The results of this work show that during the exposure of PZT to hydrogen, the capacitance of PZT capacitors changed, and these changes were due to the diffusion of protons into PZT. Further work should be done to build upon this observation to develop a quantitative relationship between the concentration of protons inside PZT and the capacitance increase. Such quantitative relationship could be generalized for other proton conductor oxides as well. The quantitative relationship between the capacitance and protons content inside oxides could be further developed into a method to evaluate the amount of protons inside the oxides, based on capacitance changes measurements.  133  Unusually large capacitance values were observed in PZT capacitors in hydrogen atmosphere at temperatures of 600-650C (unfortunately due to the limitations of the instruments, the maximum number we were able to read was approximately 100 F). Further investigation is therefore needed to determine the polarization mechanisms responsible for this high capacitance observed in the hydrogen-treated PZT capacitors. The research in this area might lead to the development of “high-temperature” super-capacitors. The very high capacitance in PZT capacitors could be then used for many practical applications, such as energy storage.  According to the dielectric spectroscopy results, hydrogen forms dipoles inside PZT. The nature of such dipoles should be better understood. In this regard, PZT with different amounts of dopants (Nb) should be prepared, and it should be investigated whether there is any correlation between the amount of dopants inside PZT and the intensity of the relaxation peaks. The important point about dipoles formed with Nb is that they may prohibit the switching of Ti-O or Zr-O dipoles. Therefore, the interactions between the dipoles that form with Nb and the Ti-O or Zr-O dipoles should be further investigated, maybe using computer modeling techniques.  We have proposed that a porous layer separating metallic electrode from PZT should be sufficient to substantially decrease the hydrogen damage of PZT. This theory should be further tested with various porous coatings other than -alumina, and the mechanisms 134  responsible for decreased hydrogen damage of PZT should be identified. Depending on the electrical properties of the insulator layer between PZT and electrode, the functionality and performance of electrode/PZT/electrode assembly could be degraded. The role of the coating between the electrode and PZT on the performance of electrode/PZT/electrode assembly should be evaluated. Insulators with dielectric constant higher or equal to the dielectric constant of PZT could be explored as the coating materials, to preserve the performance of electrode/PZT/electrode assembly.  135  References [1] R.E. Jones, P.D. Maniar, R. Moazzami, P. Zurcher, J.Z. Witowski, Y.T. Lii, P. Chu, S.J. Gillespie, Ferroelectric non-volatile memories for low-voltage, low-power applications, Thin Solid Films, 1995, 270, 584-588 [2] R. Cuppens, P.K. Larsen, G.A.C.M. Spierings, Ferroelectrics for non-volatile memories, Microelectron. Eng., 1992, 19, 245-252 [3] S. Pitman, K. Alvine, Hydrogen materials compatibility for the H-ICE, Pacific Northwest National Laboratory, 4/09/2010 [4] S. Kasap, P. Capper, Springer handbook of electronic and photonic materials, 2007, Springer [5] B. Noheda, D.E. Cox, G. Shirane, J.A. Gonzalo, L.E. Cross, S.E. Park, A monoclinic ferroelectric phase in the Pb(Zr1−xTix)O3 solid solution Appl Phys Lett, 74 (1999), pp. 2059– 2061 [6] W.M. Kriven, Displacive transformations and their applications in structural ceramics, J. Phys. IV France, 1995, 05, 101-110 [7] D. Damjanovic, Ferroelectric, dielectric and piezoelectric properties of ferroelectric thin films and ceramics, Rep. Prog. Phys., 1998, 61, 1267-1324 [8] H. Zheng, I.M. Reaney, W.E. Lee, N. Jones, H. Thomas, Effects of strontium substitution in Nb-doped PZT ceramics, J. Eur. Ceram. Soc., 2001, 21(10-11), 1371-1375 [9] K. Kushida-Abdelghafar, H. Miki, K. Torii, Y. Fujisaki, Electrode-induced degradation of Pb(ZrxTi1-x)O3 (PZT) polarization hysteresis characteristics in Pt/PZT/Pt ferroelectric thinfilm capacitors, Appl. Phys. Lett., 1996, 69, 3188-3190 [10] R. Ramesh, S. Aggarwal, O. Auciello, Science and technology of ferroelectric films and heterostructures for non-volatile ferroelectric memories, Mater. Sci. Eng. R, 2001, 32, 191236 [11] Y. Wang, W. Y. Chu, L. J. Qiao, Y. J. Su, Hydrogen-induced delayed fracture of PZT ceramics during dynamic charging under constant load, Mater. Sci. Eng. B, 2003, 98 (1), 1-5  136  [12] X. Peng, Y. J. Su, K. W. Gao, , L. J. Qiao, W. Y. Chu, Hydrogen fissure in PZT ferroelectric ceramic, Mater. Letter., 2004, 58 (15), 2073-2075 [13] X. H. Guoz, S. Q. Shiw, L. J. Qiao, Simulation of hydrogen diffusion and initiation of hydrogen-induced cracking in PZT ferroelectric ceramics using a phase field model, J. Am. Ceram. Soc., 2007, 90 (9), 2868-2872 [14] H. K. Yuh, E. Yoon, Y. T. Lee, Y. W. Park, S. I. Lee, Effects of Pb/Pt top electrode on hydrogen-induced degradation in Pb(Zr,Ti)O3, Jpn. J. Appl. Phys., 2002, 41, 42-46 [15] Y. Fujisaki, K. Kushida-Abdelghafar, Y. Shimamoto, H. Miki, The effects of the catalytic nature of capacitor electrodes on the degradation of ferroelectric Pb(Zr,Ti)O3 thin films during reductive ambient annealing, J. Appl. Phys., 1997, 82, 341-344 [16] A. R. Krauss, A. Dhote, O. Auciello, J. Im, R. Rarnesh, S. Aggarwal, Studies of hydrogeninduced degradation processes in PZT and SBT ferroelectric film-based capacitors, Integrated Ferroelectrics, Proceedings of the 11th International Symposium on Integrated Ferroelectrics, Colorado Springs, CO, March 7-10, 1999 [17] H. Miki, K. Kushida-Abdelghafar, K. Torii, Y. Fujisaki, Hydrogen-related degradation and recovery phenomena in Pb(Zr,Ti)O3 capacitors with a platinum electrode, Jpn. J. Appl. Phys., 1997, 36, 1132-1135 [18] H. Masumoto, A. Kojima, T. Iijima, T. Goto, Preparation of Ag-alloy top-electrode for ferroelectric Pb(Zr,Ti)O3 films under various atmospheres, Jpn. J. Appl. Phys., 2002, 41, 6882-6885 [19] Y. Shimamoto, K. Kushida-Abdelghafar, H. Miki, Y. Fujisaki, H2 damage of ferroelectric Pb(Zr,Ti)O3 thin-film capacitors- the role of catalytic and adsorptive activity of the top electrode, Appl. Phys. Lett., 1997, 70 (23), 3096-3097 [20] K. Kushida-Abdelghafar, M. Hiratani, Y. Fujisaki, Post-annealing effects on antireduction characteristics of IrO2/Pb(ZrxTi1-x)O3 /Pt ferroelectric capacitors, J. Appl. Phys., 1999, 85 (2), 1069-1074 [21] J. P. Han, T. P. Ma, Electrode dependence of hydrogen-induced degradation in ferroelectric Pb(Zr,Ti)O3 and SrBi2Ta2O9 thin films, Appl. Phys. Lett., 1997, 71 (9), 1267-1269 [22] Y. Fukai, The Metal Hydrogen Systems- Basic Bulk Properties, Springer Series in Materials Science Vol. 21 (Second revision), 2005 137  [23] J. C. Moreno-Pirajan, Thermodynamics - Interaction Studies - Solids, Liquids and Gases, InTech, 2011 [24] http://www.rebresearch.com/H2sol2.htm [25] M. Pasturel, R. J. Wijngaarden, W. Lohstroh, H. Schreuders, M. Slaman, B. Dam, and R. Griessen, Influence of the chemical potential on the hydrogen sorption kinetics of Mg2Ni/TM/Pd (TM=transition metal) trilayers, Chem. Mater., 2007, 19, 624-633 [26] Yu. I. Zvezdin, Yu. I. Belyakov, Hydrogen permeability of some transition metals of group 1 of the periodic system, Fiziko-Khimicheskaya Mekhanika Materialov., 1967, 3 (3), 344-351 [27] S. F. J. Cox, J. S. Lord, S. P. Cottrell, J. M. Gil, H. V. Alberto, A. Keren, D. Prabhakaran, R. Scheuermann, A. Stoykov, Oxide muonics: I. modelling the electrical activity of hydrogen in semiconducting oxides, J. Phys. Condens. Matter., 2006, 18, 1061-1078 [28] R. Hempelmann, Hydrogen diffusion mechanism in proton conducting oxides, Physica B, 1996, 226, 72-77 [29] K. Xiong, J. Robertson, Hydrogen-induced defects and degradation in oxide ferroelectrics, Appl. Phys. Lett., 2004, 85(13), 2577-2579 [30] C. H. Park, D. J. Chadi, Effect of interstitial hydrogen impurities on ferroelectric polarization in PbTiO3, Phys. Rev. Let., 2000, 84 (20), 4717-4720 [31] S. Aggarwal, S. R. Perusse, C. W. Tipton, R. Ramesh, H. D. Drew, T. Venkatesan, D. B. Romero, V. B. Podobedov, A. Weber, Effect of hydrogen on Pb(Zr,Ti)O3-based ferroelectric capacitors, Appl. Phys. Lett., 1998, 73 (14), 1973-1975 [32] K.D. Kreuer, Aspects of the formation and mobility of protonic charge carriers and the stability of perovskite-type oxides, Solid State Ionics, 1999, 125, 285-302 [33] W. Chen, L. Li, Y. Wang, Z. Gui, Effects of electrochemical hydrogen charging on leadbased relaxor ferroelectric multilayer ceramic capacitors, J. Mater. Res., 1998, 13 (5), 11101112 [34] S. Takatani, K. Kushida-Abdelghafar, H. Miki, Effect of H2 annealing on a Pt/PbZrxTi1xO3 interface studied by X-ray photoelectron spectroscopy, Jpn. J. Appl. Phys., 1997, 36, 435438  138  [35] Y. Shimakawa, Y. Kubo, Degradation of ferroelectric Pb(Zr,Ti)O3 under reducing conditions, Appl. Phys. Lett., 2000, 77 (16), 2590-2592 [36] N. Ikarashi, Analytical transmission electron microscopy of hydrogen-induced degradation in ferroelectric Pb(Zr, Ti)O3 on a Pt electrode, Appl. Phys. Lett., 1998, 73, 19551957 [37] R. Freer, Self-diffusion and impurity diffusion in oxides, J. Mater. Res., 1980, 15 (4), 803824 [38] F. Kremer, A. Schönhals, Broadband dielectric spectroscopy, Springer, Berlin, Germany, 2003 [39] S. Havriliak, S. Negami, A complex plane representation of dielectric and mechanical relaxation processes in some polymers, Polymer, 1967, 8, 161-210. [40] O. Kamishima, Y. Abe, J. Kawamura, T. Hattori, Dielectric relaxation in Yb-doped SrZrO3, J. Phys., Condens. Matter, 2004, 16, 4971-4981 [41 ] http://rbkwin.bosch.com/en/powerconsumptionemissions/dieselsysteme/dieselsystem/passengercar/technology/injection_systems/commonrailsystem/crs_pkw_types.html [42] A. Scharf, Piezo-actuators in fuel injection systems, EPCOS technical articles, 2006, (www. epcos.com) [43] K. J. Alvine, V. Shutthanandan, W. D. Bennett, C. C. Bonham, D. Skorski, S. G. Pitman, M. E. Dahl, C. H. Henager, High-pressure hydrogen materials compatibility of piezoelectric films, Appl. Phys. Lett., 2010, (97), 221911-221913 [44] Y. Park, J. K. Lee, I. Chung, J. Y. Lee, Effect of oxygen plasma treatment on hydrogendamaged Pt/Pb(ZrxTi1-x)O3/Pt ferroelectric thin-film capacitor, Jpn. J. Appl. Phys., 1999, 38, 577-579 [45] T. Sakoda, T. S. Moise, S. R. Summerfelt, L. Colombo, G. Xing, S. R. Gilbert, A. L. S. Loke, S. Ma, R. Kavari, L. A. Wills, J. Amano, Hydrogen-robust submicron IrOx/Pb(Zr,Ti)O3/Ir capacitors for embedded ferroelectric memory, Jpn. J. Appl. Phys., 2001, 40, 2911-2916 [46] T. Saito, T. Tsuji, K. Izumi, Y. Hirota, N. Okamoto, K. Kondo, T. Yoshimura, N. Fujimura, A. Kitajima, A. Oshima, Fabrication of robust PbLa(Zr,Ti)O3 capacitor structures 139  using insulating oxide encapsulation layers for FeRAM integration, Electron. Lett., 2011, 47 (8), 486-487 [47] T. S. Bjørheim, A. Kuwabara, I. Ahmed, R. Haugsrud, S. Stølen, T. Norby, A combined conductivity and DFT study of protons in PbZrO3 and alkaline earth zirconate perovskites, Solid State Ionics, 2010, 181, 130-137 [48] J43 advanced hydrogen injector, technical opportunities and challenges, Westport Innovations INC (a private presentation) [49] Ian D. Raistrick, Application of impedance spectroscopy to materials science, Ann. Rev. Mater. Sci., 1986, 16, pages 343-70 [50] R. N. Lyerl, H. W. Pickering, Mechanism and kinetics of electrochemical hydrogen entry and degradation of metallic systems, Annu. Rev. Mater. Sci., 1990, 20, pages 299-338 [51] T.P. Perng, J.K. Wu, A brief review note on mechanisms of hydrogen entry into metals, Materials Letters, 2003, 57, 3437– 3438 [52] J.K. Wu, Electrochemical method for studying hydrogen in iron, nickel and palladium, Int. J. Hydrogen Energy 17 (1992) 917– 921 [53] M.H. Armbruster, The Solubility of hydrogen at low pressure in iron, nickel and certain steels at 400 to 600°, J. Am. Chem. Soc. 65 (1943) 1043– 1054. [54] M. J. Danielson, Use of the Devanathan–Stachurski cell to measure hydrogen permeation in aluminum alloys, Corros. Sci., 2002, 44 (4), 829–840 [55] L. E. Scriven, Physics and applications of dip coating and spin coating, MRS Proceedings 1988, 121, 717-730 [56] L. Landau, B. Levich, Dragging of a liquid by a moving plate, Acta Physicochim., URSS, 1942, 17, 42-54 [57] http://www.sasoltechdata.com/tds/DISPERAL_DISPAL.pdf [58] K. Wefers, C. Misra, Oxides & hydroxides of aluminum, Technical Paper 19, Alcoa Laboratories, 1987 [59] M. Kumagai, G. L. Messing, Controlled transformation and sintering of a boehmite solgel by α-alumina seeding, J. Am. Ceram. Soc., 1985, 68 (9), 500-505 140  [60] G. Ruhi, O.P. Modi, A.S.K. Sinha, I.B. Singh, Effect of sintering temperatures on corrosion and wear properties of sol-gel alumina coatings on surface pre-treated mild steel, Corros. Sci., 2008, 50, 639-649 [61] B.D. Cullity, Elements of X-Ray Diffraction, 1956 [62] E. Baumgarten, C. L. Wagner, R. Wagner, H-D exchange and hydrogen spillover in the Pd/SiO2; system: kinetics and detection of hydrogen atoms, J. Mol. Catal., 1989, 50, 153-165 [63] E. Baumgarten, E. Denecke, Hydrogen spillover in the system Pt/Al2O3: I. fundamental observations, J. Catal., 1985, 95, 296-299 [64] J. E. Benson, H.W. Kohn, M. Boudart, On the reduction of tungsten trioxide accelerated by platinum and water, J. Catal., 1966, 5, 307-313 [65] T. B. Flanagan, W. A. Oates, The palladium-hydrogen system, Annu. Rev. Mater. Sci., 1991, 21, 269-304 [66] B. Jaffe, W. R. Cook, H. Jaffe, Piezoelectric ceramics, Academic Press, London, 1971 [67] A.C. Rastogi, S.R. Darvish, P.K. Bhatnagar, Phase evolution of electron-beam evaporated Pb(Zr,Ti)O3 thin films, Mater. Chem. Phys., 2002, 73, 135–143 [68] A. Khawam, D. R. Flanagan, Solid-state kinetic models: basics and mathematical fundamentals, J. Phys. Chem. B, 2006, 110, 17315-17328 [69] C.H. Bamford, C.F.H. Tipper, Reactions in solid state, comprehensive chemical kinetics, Vol. 22, Elsevier, Amsterdam, 1980 [70] K.D. Kreuer, St. Adams, W. Munch, A. Fuchs, U. Klock, J. Maier, Proton conducting alkaline earth zirconates and titanates for high drain electrochemical applications, Solid State Ionics, 2001, 145, 295-306 [71] W. Munch, K.D. Kreuer, G. Seifert, J. Maier, Proton diffusion in perovskites: comparison between BaCeO3, BaZrO3, SrTiO3, and CaTiO3 using quantum molecular dynamics, Solid State Ionics, 2000, 136-137, 183-189 [72] C.J.D. Von Grotthuss, "Sur la décomposition de l'eau et des corps qu'elle tient en dissolution à l'aide de l'électricité galvanique", Ann. Chim., 1806, 58, 54-73 141  [73] I. Animitsa, A. Nieman, S. Titova, N. Kochetova, E. Isaeva, A. Sharafutdinov, N. Timofeeva, Ph. Colomban, Phase relations during water incorporation in the oxygen and proton conductor Sr6Ta2O11, Solid State Ionics, 2003, 156, 95-102 [74] T. Schober, J. Friedrich, F. Krug, Phase transition in the oxygen and proton conductor Ba2In2O5 in humid atmospheres below 300C, Solid State Ionics, 1997, 99, 9-13 [75] W. Fischer, G. Reck, T. Schober, Structural transformation of the oxygen and proton conductor Ba2In2O5 in humid air: an in-situ X-ray powder diffraction study, Solid State Ionics, 1999, 116, 211-215 [76] T. Hashimoto, Y. Inagaki, A. Kishi, M. Dokiya, Absorption and secession of H2O and CO2 on Ba2In2O5 and their effects on crystal structure, Solid State Ionics, 2000, 128, 227-231 [77] P. Lunkenheimer, S. Krohns, S. Riegg, S.G. Ebbinghaus, A. Reller, A. Loidl, Colossal dielectric constants in transition-metal oxides, Eur. Phys. J. Spec. Top., 2010, 180, 61-89 [78] L. Frunza, H. Kosslick, S. Frunza, R. Fricke, A. Schönhals, Molecular dynamics of 4-noctyl-4′-cyanobiphenyl in partially filled nanoporous SBA-type molecular sieves, J. NonCryst. Solids, 2006, 90(1-6), 259-270 [79] L. Frunza, H. Kosslick, S. Frunza, R. Fricke, A. Schönhals, Rotational fluctuations of water inside the nanopores of SBA-type molecular sieves, J. Phys. Chem. B, 2005, 109 (18), 9154-9159 [80] P. B. Ishai, A. J. Agranat, Y. Feldman, Confinement kinetics in a KTN:Cu crystal, Phys. Rev. B, 2006, 73, 104104-104111 [81] Y. E. Ryabov, A. Puzenko, Y. Feldman, Nonmonotonic relaxation kinetics of confined systems, Phys. Rev. B, 2004, 69, 014204-014214 [82] J. Schüller, R. Richert, E. W. Fischer, Dielectric relaxation of liquids at the surface of a porous glass, Phys. Rev. B, 1995, 52, 15232-15238 [83] H. Y. Huang, W. Y. Chu, Y. J. Su, K. W. Gao, J. X. Li, L. J. Qiao, Hydrogen-induced semiconductor transformation of lead zirconate titanate ferroelectric ceramics, J. Am. Ceram. Soc., 2007, 90(7), 2062-2066 [84] K. J. Alvine, M. Vijayakumar, M. E. Bowden, A. L. Schemer-Kohrn, and S. G. Pitman, Hydrogen diffusion in lead zirconate titanate and barium titanate, J. Appl. Phys., 2012, 112, 043511-043517 142  [85] Y. Pei, W. Chen, Great diversity in stability of perovskite-type oxides studied by electrochemical hydrogen charging, Wuhan Univ. J. Nat. Sci., 2011, 16, 255-259 [86] W. P. Chen, Y. Wang, H. L. W. Chan, Hydrogen: a metastable donor in TiO2 single crystals, Appl. Phys. Lett., 2008, 92, 112907-112910 [87] W. P. Chen, X. F. Zhu, Z. J. Shen, J. Q. Sun, J. Shi, X. G. Qiu, Y. Wang, H. L. W. Chan, Effects of electrochemical hydrogen charging on electrical properties of WO3 ceramics, J. Mater. Sci., 2007, 42, 2524-2527 [88] W. Chen, Y. Wang, H. Chan, H. S. Luo, Hydrogen-related dynamic dielectric behavior of barium titanate single crystals, Appl. Phys. Lett., 2006, 88, 202906-202909 [89] W. Chen, J. Qi, Y. Wang, X. Jiang, H. Chan, Water-induced degradation in 0.91Pb(Zn1/3Nb2/3)O3-0.09PbTiO3 single crystals, J. Appl. Phys., 2004, 95, 5920-5922 [90] A.L. Ahmad, N.F. Idrus, M.R. Othman, Preparation of perovskite alumina ceramic membrane using sol–gel method, J. Membrane Sci., 2005, 262, 129-137 [91] A. F. M. Leenaars, K. Keizer, A.J. Burggraaf, The preparation and characterization of alumina membranes with ultra-fine pores, J. Mater. Sci., 1984, 19, 1077-1088 [92] D. S. Maciver, H. H. Tobin, R. T. Barth, Catalytic aluminas I. surface chemistry of eta and gamma alumina, J. Catal., 1963, 2, 485-497 [93] K. Wefers, C. Misra, Oxides and hydroxides of aluminum, Alcoa Technical Paper No. 19, Revised; Alcoa Laboratories: Alcoa Center, PA, 1987. [94] Y. Liu, D. Ma, X. Han, X. Bao, W. Frandsen, D. Wang, D. Su, Hydrothermal synthesis of microscale boehmite and gamma nanoleaves alumina, Mater. Lett., 2008, 62, 1297-1301 [95] K. Sohlberg, S. J. Pennycook, S. T. Pantelides, Hydrogen and the structure of the transition aluminas, J. Am. Chem. Soc., 1999, 121, 7493-7499 [96] J. Joubert, A. Salameh, V. Krakoviack, F. Delbecq, P. Sautet, C. Coperet, J. M. Basset, Heterolytic splitting of H2 and CH4 on -alumina as a structural probe for defect sites, J. Phys. Chem. B, 2006, 110, 23944-23950  143  [97] C. Y. Yu, B. K. Sea, D. W. Lee, S. J. Park, K. Y. Lee, K. H. Lee, Effect of nickel deposition on hydrogen permeation behavior of mesoporous  -alumina composite membranes, J. Colloid. Interf. Sci., 2008, 319, 470-476 [98] K. J. Sladek, E. R. Gilliland, R. F. Baddour, Diffusion on surfaces (II): correlation of diffusivities of physically and chemically adsorbed species, Ind. Eng. Chem. Fundamen., 1974, 13 (2), 100-105 [99] H. A. Pray, C. E. Schweickert, B. H. Minnich, Solubility of hydrogen, oxygen, nitrogen, and helium in water, Ind. Eng. Chem. Res., 1952, 44, 1146-1151 [100] A.Q. Jiang , H. J. Lee , G. H. Kim , C. S. Hwang , The inlaid Al2O3 tunnel switch for ultrathin ferroelectric films, Adv. Mater., 2009, 21, 2870-2875  144  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0073678/manifest

Comment

Related Items