Processing and Properties of Advanced Ceramics and Composite Ceramic Transactions

Processing and Properties of Advanced Ceramics and Composites Processing and Properties of Advanced Ceramics and Comp...

Author: Narottam P. Bansal | J. P. Singh

437 downloads 2040 Views 13MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Processing and Properties of Advanced Ceramics and Composites


Ceramic Transactions, Volume 203 A Collection of Papers Presented at the 2008 Materials Science and Technology Conference (MS&T08) October 5-9, 2008 Pittsburgh, Pennsylvania

Edited by

Narottam P. Ban sal J. P. Singh

®WILEY A John Wiley & Sons, Inc., Publication

Copyright © 2009 by The American Ceramic Society. All rights reserved. Published by John Wiley & Sons, Inc., Hoboken, New Jersey. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic format. For information about Wiley products, visit our web site at www.wiley.com.

Library of Congress Cataloging-in-Publication Data is available. ISBN 978-0-470-40845-2 Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1

Contents

Preface

ix

MICROWAVE PROCESSING Continuous Microwave-Driven Polyol Process for Synthesizing Ytterbium-Doped Yttria Powder

3

M. A. Imam, A. W. Fliflet, K. L. Siebach, A. David, R. W. Bruce, S. B. Qadri, C. R. Feng and S. H. Gold

Microwave Irradiation-Assisted Method for the Rapid Synthesis of Fine Particles of α-ΑΙ 2 0 3 and a-(AI1.xCrx)203 and Their Coatings onSi(100)

15

Anshita Gairola, A. M. Umarji, and S. A. Shivashankar

CHEMICAL VAPOR DEPOSITION Synthesis and Characterization of Si/Si2N20/Si3N4 Composites from Solid-Gas Precursor System Via CVD

25

J. C. Flores-Garcia, A. L. Leal-Cruz, and M. I. Pech-Canul

Effect of Flow Rate, Nitrogen Precursor and Diluent on Si 2 N 2 0 Deposition by HYSYCVD

35

A. L. Leal-Cruz, M. I. Pech-Canul, E. Lara-Curzio, R. M. Trejo, and R. Peascoe

COMBUSTION SYNTHESIS MgAI204/SiC Composite Ceramic Material Produced by Combustion Synthesis

4

Podbolotov Kirill Borisovich and Diatlova Evgenija Mihajlovna

v

Finite Element Analysis of Self-Propagating High-Temperature Synthesis of Strontium-Doped Lanthanum Manganate

53

Sidney Lin and Jiri Selig

REACTION FORMING AND POLYMER PROCESSING Comparison of Bulk and Nanoscale Properties of Polymer Precursor Derived Silicon Carbide with Sintered Silicon Carbide

65

Arif Rahman, Suraj C. Zunjarra, and R. P. Singh

Process Design and Production of Boron Trichloride from Native Boron Carbide in Lab-Scale

77

D. Agaogullari and I. Duman

SINTERING AND HOT PRESSING Spark Plasma Sintered Alumina-Zirconia Nano-Composites by Addition of Hydroxyapatite

93

S. F. Li, H. Izui, M. Okano, W. H. Zhang, and T. Watanabe

Comparison of Slip Cast to Hot Pressed Boron Carbide

10

T. Sano, E.S.C. Chin, B. Paliwal, and M. W. Chen

AMORPHOUS CERAMICS Mechanically Driven Amorphization and Bulk Nanocrystalline Synthesis of Ultra-High Temperature Ceramics

119

H. Kimura

Preparation and Characterization of Fused Silica Based Ceramic Cores Used in Superalloy Casting

131

M. Arin, S. Sevik, and A. B. Kayihan

COATINGS AND FILMS Photon Effects in Ultra-Thin Oxide Films: Synthesis and Functional Properties

143

S. Ramanathan, M. Tsuchiya, C. L. Chang, and C. Ko

Faradayic Process for Electrophoretic Deposition of Thermal Barrier Coatings for Use in Gas Turbine Engines

153

Joseph Kell and Heather McCrabb

A Novel Method to Spray Tungsten Carbide Using Low Pressure Cold Spray Technology J. Wang and J. Villafuerte

vi

· Processing and Properties of Advanced Ceramics and Composites

161

COMPOSITES Foreign Object Damage Versus Static Indentation Damage in an Oxide/Oxide Ceramic Matrix Composite

171

Sung R. Choi, Donald J. Alexander, and David C. Faucett

Distinguished Functions Making the Best Use of the Unique Composite Structures

181

Toshihiro Ishikawa

Effects of Environment on Creep Behavior of NEXTEL™720/ Alumina-Mullite Ceramic Composite at 1200 °C

193

C. L Genelin and M. B. Ruggles-Wrenn

Performance of Composite Materials in Corrosive Conditions: Evaluation of Adhesion Loss in Polymers Via Cathodic Disbondment and a Newly Developed NDE Technique

205

Davion Hill, Colin Scott, Ayca Ertekin, and Narasi Sridhar

Effect of Variations in Process Shear on the Mixedness of an Alumina-Titania System

215

C. August, M. Jitianu, and R. Haber

MODELING Modeling of the Pressure in 1-D Green Ceramic Bodies during Depressurization from Conditions of Supercritical Extraction of Binder

229

Kumar Krishnamurthy and Stephen J. Lombardo

Models of the Strength of Green Ceramic Bodies as a Function of Binder Content and Temperature

239

Stephen J. Lombardo and Rajiv Sachanandani

Finite Element Modeling of Steel Wire Drawing through Dies Based on Encapsulated Hard Particles

249

Daniel J. Cunningham, Erik M. Byrne, Ivi Smid, John M. Keane

Author Index

255


· vii

Preface

Two international symposia "Innovative Processing and Synthesis of Ceramics, Glasses and Composites" and "Ceramic Matrix Composites" were held during Materials Science & Technology 2008 Conference & Exhibition (MS&TO8), Pittsburgh, PA, October 5-9, 2008. These symposia provided an international forum for scientists, engineers, and technologists to discuss and exchange state-of-the-art ideas, information, and technology on advanced methods and approaches for processing, synthesis and characterization of ceramics, glasses, and composites. A total of 105 papers, including 12 invited talks, were presented in the form of oral and poster presentations. Authors from 15 countries (Belarus, Canada, China, France, Germany, India, Iran, Japan, Mexico, Norway, Russia, South Korea, Taiwan, Turkey, and the United States) participated. The speakers represented universities, industries, and government research laboratories. These proceedings contain contributions on various aspects of synthesis, processing and properties of ceramics, glasses, and composites that were discussed at the symposium. Twenty two papers describing the latest developments in the areas of combustion synthesis, microwave processing, reaction forming, polymer processing, chemical vapor deposition, electrophoresis, spark plasma sintering, mechanical amorphization, thin films, composites, etc. are included in this volume. Each manuscript was peer-reviewed using The American Ceramic Society review process. The editors wish to extend their gratitude and appreciation to all the authors for their cooperation and contributions, to all the participants and session chairs for their time and efforts, and to all the reviewers for their useful comments and suggestions. Financial support from The American Ceramic Society is gratefully acknowledged. Thanks are due to the staff of the meetings and publications departments of The American Ceramic Society for their invaluable assistance. It is our earnest hope that this volume will serve as a valuable reference for the researchers as well as the technologists interested in innovative approaches for synthesis and processing of ceramics and composites as well as their properties. NAROTTAM P. BANSAL

NASA Glenn Research Center J. P. SINGH

U.S. Army International Technology, Center-Pacific (ITC-PAC) ix

Processing and Properties of Advanced Ceramics and Composites Edited by Narottam P. Bansal and J. P. Singh Copyright O 2009 The American Ceramic Society.

Microwave Processing


CONTINUOUS MICROWAVE-DRIVEN POLYOL PROCESS FOR SYNTHESIZING YTTERBIUM-DOPED YTTRIA POWDER M.A. Imam, A.W. Fliflet, K.L. Siebach*, A. David*, R.W. Bruce**, S. B. Qadri, C. R. Feng and S.H. Gold Materials Science and Component Technology Directorate, Naval Research Laboratory, Washington, DC, USA ABSTRACT The continuous microwave polyol process is a promising novel approach to the synthesis of metallic and ceramic nanopowders. Current efforts are directed toward synthesizing ytterbia-doped yttria (Yb203:Y,03) for use as a polycrystalline laser host material. The process involves pumping a mixture of yttrium nitrate and ytterbium nitrate dissolved in hydrated diethylene glycol through a pressurized quartz tube contained in an S-Band waveguide driven by a 2.45 GHz microwave source at powers up to 6 kW. As the solution moves along the waveguide, it absorbs the co-propagating microwave energy and is heated rapidly to a temperature above 200°C causing a reaction to occur. Condensation reactions then form particles with ytterbium-doped yttria crystal structure. The rapid heating and cooling serve to limit the growth of the crystals so that they are on submicron and fairly uniform in size. The production of doped yttria was confirmed by x-ray diffraction. INTRODUCTION In the polyol process, an organic solvent such as glycol or alcohol is used to reduce a dissolved metal salt to the metal [1,2J. This is commonly done in a boiling, reflux system where the glycol solution of the metal salt is heated to boiling, and the evaporating solvent is condensed and fed back into the solution. At the elevated boiling temperature, the glycol solvent acts as a reducing agent, converting the dissolved metal salt first to a metal oxide and then to the metal. The process results first in the formation of metal atoms suspended in the glycol solvent. These then aggregate, first into clusters and then into larger metallic particles. The process is capable of producing metallic particles in the nanometer size range (1-100 nm), and the particles produced are protected from oxidation or nitridation by the organic solvent and can also be further protected by organic coatings generated during the process from additives. This process has been used for about a decade in production of nanophase powders of metals and mixtures of such metals and films or coatings of these [3], and a wide range of metals can be produced in this manner. The process can also be used to produce metal oxides, sulfides and selenides [4,5]. The limiting factor is the chemical energy available from the solvent vs. the enthalpy of formation of the metal oxides. This makes it very difficult to obtain nanophase metals such as lithium, aluminum, yttrium, magnesium, zirconium, e.g., Groups I-IV, without resorting to much higher processing temperatures, although nanophase oxides of many of these metals can be produced [6]. However, most of the balance of the metals in the periodic table can be produced by this process, e.g., Fe, Co, Ni, Cu, Ru, Rh, Pt, Au, etc., as well as intimate mixtures of these-Fe/Co, Co/Pt, Fe/Pt, Ni/Ag, Cu/Ni, Co/Ni, Co/Ni/Cu, etc. The conventional polyol process, where the processing is done with a high boiling point solvent such as ethylene glycol, heated in a reflux system by a heating mantle, is adequate for production of small quantities of experimental powders, e.g., 1-10 g of product from a 1 liter batch that may take 0.5-2 hours to process. However, the process is intrinsically limited in scalability. In the heating mantle/flask system, scaling to larger volumes results in much greater product variability from varying thermal histories in the larger volume with different convection

3

Continuous Microwave-Driven Polyol Process for Synthesizing Ytterbium-Doped Yttria

cell structure, and also typically results in larger particle sizes from longer processing times. The typical approach for increasing production rates is to set up a large number of reactors running simultaneously. This approach raises issues of batch-to-batch and reactor-to-reactor product variability, and may never be capable of producing larger quantities for production uses at reasonable costs. One process variation that has been adopted to overcome the limitation on process temperature resulting from boiling of the solvent is to drive the process in sealed containers heated in a microwave cavity [5-10]. This has been done using adaptations of systems intended for microwave digestion. This permits raising the process temperature to as high as 240°C vs. a boiling point for ethylene glycol of ca. 185°C, and avoids the complications associated with solvent boiling. While this process has some advantages over the conventional flask/heating mantle systems, it also has extreme limitations in terms of process size. The typical reactors, Teflon-lined PEI, that, in theory, are limited to temperatures of 170°C, are only 100 ml in volume. This permits production of only about 0.01 g of product per reaction vessel, typically processed in about 15 min. Again, this may be useful for preliminary research purposes, but is clearly not scalable. Going to larger reactors in a more powerful microwave system would result in unacceptably high stresses in the vessels (the stresses scale with reactor diameter), that are already being operated above their rated continuous operating temperature. We have been investigating the use of microwave and millimeter-wave systems in various types of materials processing for about a decade [1,11-17]. As part of this program we have been exploring the use of microwave and millimeter-wave heating of polyol solutions for the production of nanophase metals and oxides. EXPERIMENTAL Initially, this process was demonstrated with millimeter-wave beam heating (83 GHz) of batch polyol systems [14]. This demonstrated that greatly reduced process times could be achieved with the millimeter-wave beam heating, which is capable of rapidly heating polyol solutions in bulk, but still left the problems with the economics of the batch process. Subsequently, continuous polyol processing with millimeter-wave beam heating was demonstrated with very interesting results. In this case, shown in Figure 1 an ethylene glycol/copper acetate solution was heated by the millimeter-wave beam to approx. 200°C in a few seconds as it passed through a silica reaction tube, approx. 10 mm in ID, producing nanophase copper metal particles from the copper acetate. This is an extreme case of microwave polyol processing, as the beam configuration permits very high deposited power densities in the solution, perhaps as high as 100-500 W/cm3 in the liquid, but it's very interesting here that this process can be driven to completion in a few seconds in this continuous system, rather than the hours required in the conventional process. Use of the millimeter-wave system for large-scale production of nanophase metals was considered, but several practical considerations argued otherwise. The millimeter-wave system is very costly, high operating costs, and the continuous system involved extensive plumbing and plumbing connections within the millimeter-wave processing chamber, making operations and maintenance rather difficult and system reliability questionable. A continuous processing system is only economical when it can be operated continuously for long periods at low cost. These considerations led to the decision to use a dedicated, lower frequency, S-band source (2.45 GHz)

4



Figure 1. Schematic of millimeter-wave driven continuous polyol process using 83 GHz gyrotron source operating at 2-3 kW and producing nanophase copper with less than 10 s reaction time. for a continuous microwave polyol system much more likely to be successful and economically feasible in large scale production of nanophase metals. One pleasing but fortuitous result with this change was much better coupling to the polyol solution. It happens that the peak in absorption in ethylene glycol is very close to 2.45 GHz; thus we have very efficient coupling at this frequency into small volumes of solution. The small volume consideration has significant implications for operation under pressure. We initially experimented with an S-band analogue of the millimeter-wave system, using a resonant cavity made from a shorted waveguide section, tuned to place the maximum in the field at the location of a silica reaction tube passing transversely through the waveguide section. This was found to be impractical because of the field distribution within the cavity that resulted in a heated region only 1-2 cm in length, with inadequate time for the process to occur. We subsequently went to a traveling wave applicator that has been very successful for this process and provides a much better geometry for a well-controlled continuous polyol process [18]. This configuration is shown schematically in Figure 2. In this system, the direction of propagation of the microwaves down the waveguide and the direction of flow of the polyol solution through the inner quartz tube can be in the same or opposite sense. These produce different temperature distributions in the polyol solution. Temperature distributions, based on modeling results, can also be controlled by variations in input microwave power vs. polyol flow rate. Presently, we are operating with the propagation and flow directions the same and with process parameters that result in a fairly uniform temperature distribution. The other choice, with the directions opposite, produces a more gradual temperature rise in the solution with a shorter length of fairly uniform temperature. For the case we are using, the liquid traveling down the tube is heated by absorption of the microwaves traveling down the waveguide, with the power deposited proportional to the electric


· 5


Figure 2. Schematic of microwave polyol process with solution pumped through silica tube placed along centerline of S-Band waveguide-the microwaves that propagate down the waveguide heat the solution flowing along the tube. field squared and the dielectric loss in the liquid. The absorption in the liquid causes attenuation of the microwaves; in our present system and with our current operating parameters, all of the microwave power is absorbed in the polyol solution in a distance of 20-40 cm, as the solution is heated from its initial temperature to a maximum of 180-240°C (depending on system pressure) as shown in Figure 3. Because of the relatively simple geometry and well-defined boundary In this system, the direction of propagation of the microwaves down the waveguide and the direction of flow of the polyol solution through the inner quartz tube can be in the same or opposite sense. These produce different temperature distributions in the polyol solution. Temperature distributions, based on modeling results, can also be controlled by variations in input microwave power vs. polyol flow rate. Presently, we are operating with the propagation and flow directions the same and with process parameters that result in a fairly uniform temperature distribution. The other choice, with the directions opposite, produces a more gradual temperature rise in the solution with a shorter length of fairly uniform temperature. For the case we are using, the liquid traveling down the tube is heated by absorption of the microwaves traveling down the waveguide, with the power deposited proportional to the electric field squared and the dielectric loss in the liquid. The absorption in the liquid causes attenuation of the microwaves; in our present system and with our current operating parameters, all of the microwave power is absorbed in the polyol solution in a distance of 20-40 cm, as the solution is heated from its initial temperature to a maximum of 180-240°C (depending on system pressure) as shown in Figure 3. Because of the relatively simple geometry and well-defined boundary conditions, it is possible to calculate accurately the electric fields in this system; this calculation can include the temperature dependence of the dielectric properties of the polyol solution. From the dielectric loss of the ethylene glycol/metal salt solutions, the energy deposited in the ethylene glycol and heating can be determined, as shown in Figure 4. This heating causes a rapid increase in the temperature of the glycol solution that is offset somewhat by the attenuation of the electric field in the propagating microwaves. Calculation of temperature increases in the ethylene glycol currently include corrections for thermal losses from the silica tube to the waveguide (free

6



Figure 3. Schematic of basis for model for waveguide microwave heating of flowing polyol solution

Figure 4. Schematic of electric field distribution in waveguide containing silica tube with flowing polyol solution; field in polyol solution is reduced because of high permittivity (dielectric constant), but energy absorption is highly localized in polyol solution because of very high dielectric loss. convection in a closed system) and from the outside of the waveguide (free convection), though these are not significant, corrections with present conditions. Other thermal losses (radiation, conduction) are negligible. With present parameters, the outside of the waveguide, cooled by


· 7


free convection, remains below 60°C, with the polyol solution at 240°C. The result of these calculations, as confirmed by experiments with thermocouple probes at various locations in the polyol solution is a rapid increase, over about 10 cm, in temperature of the polyol solution to a plateau value as shown in Figure 5. This plateau value is maintained, for appropriate choices of microwave power

Figure 5. Schematic of temperature distribution in continuous microwave polyol system with polyol flow and microwave propagation directions coincident with experimental results shown for comparison. and solution flow rate, over a substantial distance-30-60 cm. With this present arrangement, the solution being processed can be maintained at a desired process temperature over this distance which corresponds to a residence time, in the reaction tube of 30-60 s. With this system at ambient pressure, temperatures slightly in excess of boiling can be reached, approx. 180-200°C, for ethylene glycol solutions. We currently are operating the system at an overpressure of 0.20.3 MPa, using pressure regulators on the outlet side and positive displacement pumps to move the solution, and thus can operate at about 240°C, without boiling occurring. The use of this overpressure is greatly facilitated by the relatively small diameter of our silica processing tube that should be capable of operating at pressures as high as 70 MPa. An overall picture of the actual system is shown in Figure 6. This is a somewhat earlier version of the system but shows most of the current features. The S-Band source is to the left is a self-contained unit, -1.5 m in height and 1 m in width, and requires 440V, 30A three-phase electrical power as well as cooling water. The power is taken out of the unit on the top through conventional aluminum and brass waveguide. The working portion of the waveguide, containing the silica reaction tube for the polyol, is placed vertically here. Any power not absorbed in the polyol process is collected in a water load at the upper right of the waveguide. Input power to the system is monitored as well as power reflected back into the source. If needed, power into the water load can be monitored as well either by RF power meters or by measurement of cooling water temperature load. During normal operations, temperature of the polyol solution is monitored at the outlet (at the top). A well either by RF power meters or by measurement of cooling water temperature load. During

8



Figure 6. Overall view of continuous microwave polyol system based on 6 kW maximum output S-Band, 2.45 GHz source. normal operations, temperature of the polyol solution is monitored at the outlet (at the top). A fine wire thermocouple probe can be used, if necessary to check the temperature at various points along the length of the reaction tube, but is normally not needed with stable continuous operation. Viewports are also available which can be used to observe the process visually at roughly the midpoint of the reaction tube. The system is currently being operated under overpressure; this is achieved via a positive displacement pump driving the polyol solution, currently capable of achieving about 0.7 MPa pressure. The pressure is controlled via a pressure regulator on the outlet. The product solution is rapidly cooled via a stainless steel heat exchanger in an ice water bath and collected for use or analysis. Nearly all of the parts of the present system through which the reactants and products pass are either stainless steel or silica, with thick wall silicone rubber tubing used in some areas. These components will tolerate the reactants used, the temperatures involved (currently up to 240°C), and the overpressures used to suppress boiling. The entire system will tolerate cleaning operations such as flushing with nitric acid to remove metallic residues, followed by distilled water, alcohol and ethylene glycol flushes. This capability is critical if the system is to be used for more than one type of material and purity of products is critical. In the present work, the chemicals used were diethylene glycol (Alfa Aesar), yttrium nitrate (99.999%, Metall), ytterbium nitrate (99.99%, Metall), urea (Fisher Chemical), and


· 9


distilled water. Diethylene glycol was used instead of ethylene glycol to minimize the onset of boiling during processing as it has a higher boiling point. Three thermocouples were used to monitor the process at various points—one at the intake flask, another after the waveguide to measure internal temperature, and another at the collection flask to determine how much the solution has cooled after going through the cooling coil. An S-Band 2.45-GHz Cober high power microwave generator provided the power. The pump system was made up of a peristaltic pump, quartz tubing through an s-band waveguide, a stainless-steel pressure-regulation system, and rubber tubing that led to a stainless-steel cooling coil in a water bath as shown in Figure 6. A precursor solution was created containing diethylene glycol, yttrium nitrate, and ytterbium nitrate. Dissolving agents and catalysts, urea and water, were also added to the solution. Each run contained 1200-mL of diethylene glycol, 30-mL of water, and 24-g of urea. The amount of yttrium nitrate and ytterbium nitrate was determined by weight according to the dopant concentrations of 0%, 5%, and 10%. When mixing the precursor solution, a magnetic stirrer and a hot plate were used to warm the mixture to 70°C, until homogeneous. Before the precursor solution is pumped through the system, the reaction pathway is primed with pure diethylene glycol to eliminate air from the system. Next microwave power is increased slowly until a temperature of ~ 212°C is achieved at the waveguide output at a pressure of 20-psi. This temperature and pressure is maintained as the solution is pumped at 0.37 mL/s through the system and collected in a flask that was maintained at room temperature. The solution is then cleaned in an alcohol rinsing procedure, which requires a centrifuge. The solution is first put in centrifuge tubes and centrifuged for 30-min at 10,000-rpm. The diethylene glycol is then decanted. The tubes are then filled with the alcohol reagent and spun for 15-min at 10,000-rpm two more times. The result was a highly compacted white paste. The paste was then placed on a Petri-dish and warmed to 70°C in open air to evaporate off the alcohol reagent. The powder was calcined in a furnace for 150-min at about 700°C. The resulting powder is analyzed using x-ray diffraction and scanning electron microscopy. X-ray diffraction scans on Y,03 and Yb203 powder samples were obtained using Cu Κα radiation from a rotating anode x-ray source and a high resolution powder diffractometer. Figures 7a shows the scan taken for Y,03 samples after microwave processing. The vertical lines correspond to the expected diffraction pattern for Y203 from JCPDS card (Pdf # 00-043-1036) [19]. It is clear from these figures that a pure Y : 0 3 phase was obtained although in Fig. 7a we also see some extra peaks that are not identifiable. Figures 8a and 9a correspond to a mixture of Y,03 with 5 and 10 % Yb203, respectively. Superimposed on these scans are the red vertical lines corresponding to pure Y : 0 3 phase and blue vertical lines (Pdf # 00-043 1037) [19] corresponding to Yb,03. As can be seen there is a shift between in the position of the peaks with their centroids located in between these two vertical lines. This indicates a change in lattice parameters with increasing Yb,03 composition and also the ytterbium is occupying the yttrium site substitutionally. This is expected if a solid solution is formed between Y203 and Yb,03 since both are isostructural having a space group Ia3(206) [20]. SEM micrographs, Figures 7b, 8b and 9b, show particle size ranging from 400 to 700 nm. We did not attempt to control the particle size by regulating flow rate or temperature. DISCUSSION The steps of reaction for the formation of yttrium oxide can be assessed through the color changes during processing. The solution is initially clear, precipitation occurs while heating and becomes translucent, and then slowly changes to milky-tan. These reaction steps indicate that

10



Figure 7a. X-ray diffraction scans on Y203 sample obtained by polyol processing

Figure 7b. SEM micrograph of Y20,

Figure 8a. X-ray diffraction scans on Y20, + 5 wt%Yb20, sample obtained by polyol processing

Figure 8b. SEM micrograph of Y,03 +5 wt%Yb203

the precursor first reacts to form an insoluble intermediate, which then slowly becomes yttria by means of a continuous reaction. A likely mechanism for this reaction is given below: W Y(N0 3 ) 3 + ΟΗ->Υ(Ν0 3 ) χ (ΟΗ) 3 _ χ -> Y(OH)3 (2)

Y(OH)3 -> Y(OH)x03_x -> Y 2 0 3

A similar reaction should apply for the formation of ytterbium oxide from ytterbium nitrate. Doping essentially involves introducing an active ion into the crystal lattice. In order to create a good material, there needs to be a homogeneous mixture of the ytterbium oxide within the


· 11


Figure 9a. X-ray diffraction scans on Y,03 + 10 wt%Yb20, sample obtained by polyol processing

Figure 9b. SEM micrograph of Υ,Ο, +10 wt%Yb,03

yttrium oxide crystal structure. By using diethylene glycol in the process, the ytterbium nitrate and yttrium nitrate can be suspended in solution to create a homogeneous solution. The suspension results in a dispersion that helps to create a well dispersed ytterbium-yttrium oxide structure. The resulting powder should have the proper proportions of ytterbium and be evenly distributed since the solution itself was evenly distributed. X-ray diffraction results give the indication that ytterbium was introduced into the yttrium crystal lattice as substitutional element. To date, we have demonstrated that the continuous microwave polyol system can be used in production of sizeable quantities of nanophase ytterbium doped yttrium oxide. The process is energy efficient and can produce material with the purity level determined by the precursor materials. The particle size can be controlled which may help in further processing by optimizing temperature, pump speed, and pressure to fully take advantage of the microwave-assisted polyol process. Note that one significant advantage of operation under overpressure is the ability to use lower boiling point solvents or glycols that typically have higher solubility for the metal salts used as precursors. This could be useful in increasing production rates for materials whose precursors have very limited solubility in high temperature diols such as ethylene glycol. Another approach that may be possible to increase production rates, but has not been explored by the authors, is to employ suspensions of fine particles of precursors in a diol. This would eliminate the restrictions on concentrations imposed by solubilities, but might lead to larger metal particle size in the product because of the very different reaction path. The economics of this process should be far superior to any of the batch polyol processes with much larger production quantities. We would expect that the process here could be readily scaled to higher production quantities. One obvious approach is to move to a higher power SBand microwave source. The higher power, coupled with either a larger reaction tube and/or higher flow rates, would permit production rates several times higher than possible with our prototype system. A limitation with this approach is the high loss in solvents such as ethylene glycol, which limit the penetration depth of microwave energy. For sufficiently large reaction tubes, heating would no longer be relatively uniform across the diameter of the tube and greater process variability would occur. Another, also relatively practical scaling approach, is to move to a lower frequency system, such as L-band (915 MHz) for which low cost, commercial sources and hardware are also available. In this case, with a frequency about 1/3 that of S-band, the

12



waveguide are approx. 3 times larger, the absorption is somewhat less than at S-Band, and reaction tubes 3-4 times larger than our 10 mm ID could easily be accommodated. An L-Band system, with a 40 mm ID reaction tube, and about 100 kW output, could process hundreds of liters of solution per hour and produce tens of kilograms of nanophase product per shift. There are a number of cautions that need to be cited here. While the economics of the continuous microwave polyol process are quite good, with low capital cost, low operating and labor costs, and low raw materials cost, there are some other critical issues. One is the matter of recovering the nanophase metal or other powder product from the diol solvent and other reaction products. This is done on small scales now, but economical processes (settling, centrifuging, ultrafiltration, controlled agglomeration) need to be available for the separation process on much larger scales. There are also generic problems associated with handling nanophase powders, which tend to agglomerate readily and are highly reactive with atmosphere, but these are not peculiar to polyol-derived nanophase powders. Processing hundreds or thousands of liters of glycol solution per shift also require economical techniques for recycling large quantities of usable solvents and reactants and disposing of waste products, that may or may not be hazardous. Assuming that these problems noted here could be overcome, the continuous microwave polyol process could be a viable technique for economical large-scale production of a wide range of nanophase materials-metals, metal alloys and mixtures, metal oxides and other materials such as selenides and sulfides. CONCLUSIONS 1. The continuous microwave polyol system can be used in production of sizeable quantities of nanophase ytterbium doped yttrium oxide. 2. The economics of this process should be far superior to any of the batch polyol processes. 3. X-ray diffraction results give the indication that ytterbium was introduced into the yttrium crystal lattice as substitutional element. 4. A likely mechanism for the formation of ytterbia/yttria from ytterbia/yttria nitrate is proposed. ACKNOWLEDGEMENTS This work was supported by the U.S. Office of Naval Research. FOOTNOTES * Summer student ** Summer Faculty REFERENCES 'D. Lewis III, M. A. Imam, R. W. Bruce, L. Kurihara, A. W. Fliflet and S. H. Gold, Production of Nanophase Metals via the Continuous Microwave Polyol Process, TMS Proceedings on Powder Materials, pp. 157-168 (2003) 2 D. Larcher and R. Patrice, Preparation of Metallic Powders and Alloys in Polyol Media: A Thermodynamic Approach, J. Solid State Chem. 154,405-411 (2000). 3 G. M. Chow, J. Zhang, Y. Y. Li, J. Ding and W. C. Goh, Electroless polyol synthesis and properties of nanostructured NixCo100x films, Mater. Sci. and Eng. A304-A306, 194-199 (2001). 4 D. Chen, G.Z. Shen, K. B. Tang, X. Jiang, L. Y. Huang, Y. Jin and Y. T. Qian, Polyol mediated synthesis of Nanocrystalline M3SbS3 (M=Ag, Cu), Mater. Res. Bull. 38 (3) 509-513 (2003)


· 13


Ή. Grisaru, O. Palchik, A. Gedanken, V. Palchik, M. A. Slifkin and A. M. Weiss, Preparation of the Cd(l-x)Zn(x)Se alloys in the nanophase form using microwave irradiation, J. Mater. Chem. 12 (2) 339-344 (2002) 6 T. Yamamoto, Y. Wada, H. B. Yin, T. Sakata, H. Mori, S. Yanagida, Microwave-driven polyol method for preparation of Ti02 nanocrystallites, Chemistry Letters 10, 964-965, (2002). 7 M. Tsuji, M. Hashimoto and T. Tsuji, Fast Preparation of Nano-sized Nickel Particle Under Microwave Irrad. without Using Catalyst for Nucleation, Chemistry Letters, 1232-1233 (2002). 8 W. Tu and H. Liu, Rapid synthesis of nanoscale colloidal metal clusters by microwave irradiation, J. Mater. Chem. 10, 2207-2211 (2000). 9 S. Komarneni, L. Dongsheng, B. Newalkar, H. Katsuki and A. S. Bhalla, Microwave-Polyol Process for Pt and Ag Nanoparticles, Langmuir 2002,18 5959-5962 (2002). I0 D. Li and S. Komarneni, Microwave-Assisted Polyol Process for Synthesis of Ni Nanoparticles, Journal of the American Ceramic Society. 89, 1510 - 1517(2006). "R. W. Bruce, A. W. Fliflet, D. Lewis III, R. J. Rayne et al., Microwave Sintering of Pure and Doped Nanocrystalline Alumina Compacts, Mat. Res. Soc. Symp. Proc, 430, 139-144 (1996). ,2 R. W. Bruce, A. W. Fliflet, R. P. Fischer, D. Lewis III et al., Millimeter-Wave Processing of Alumina Compacts, Ceramic Transactions Vol. 80 Microwaves: Theory and Application in Materials Processing IV, 287-294 (1997). n A.W. Fliflet, R. W. Bruce, R. P. Fischer, D. Lewis III, L. K. Kurihara, B. A. Bender, G.-M. Chow and R. J. Rayne, A Study of Millimeter-Wave Sintering of Fine-Grained Alumina Compacts, IEEE Trans. On Plasma Sci., Vol. 28 No. 3, 924-935 (2000). U L. K. Kurihara, D. Lewis, A. M. Jung, A. W. Fliflet and R. W. Bruce, Millimeter-Wave Driven Polyol Processing of Nanocrystalline Metals, Proc. MRS, Vol. 634 (2001) 15 S. H. Gold, D. Lewis, A. W. Fliflet, B. Hafizi and J. R. Penano, Interference and Guiding Effects in the Heating of Ceramic Slabs and Joints with Millimeter-Wave Radiation, J. Mater. Synth, and Proc, 9, No. 5 287-297 (2002). 16 D. Lewis, M. A. Imam, L. K. Kurihara, A. W. Fliflet, S. Gold, R. W. Bruce, Material Processing with a High Frequency Millimeter-Wave Source, Mater, and Manuf. Proc, 18, No. 2 151-167(2003). I7 M. A. Imam, D. Lewis III, R. W. Bruce, A. W. Fliflet and L. K. Kurihara, Processing of Advanced Materials with a High Frequency, Millimeter-Wave Beam Source and Other Microwave Systems, Materials Science Forum, 426-432, 4111-4116 (2003). 18 S.H. Gold, R.W. Bruce, A.W. Fliflet, D. Lewis III, L.K. Kurihara, and M.A. Imam, System for Continuous Production of Namophase Material using a Microwave-Driven Polyol Process, Rev. Sci. 78, 02309, pp. 1-6(2007). 19 MacKenzie, K.J.D., Gainsford, G.J., Ryan, M.J, J. Eur. Ceram. Soc, vl6 p553 (1996) 20 Zachariasen, W.H. Skr. Nor. Vidensk.-Akad., Kl. 1: Mat.-Naturvidensk. Kl., v 1928 pi (1928)

14



MICROWAVE IRRADIATION-ASSISTED METHOD FOR THE RAPID SYNTHESIS OF FINE PARTICLES OF α-Α1203 AND a-(Al,. x Cr x ) 2 03 AND THEIR COATINGS ON Si(100) Anshita Gairola, A.M. Umarji, and S.A. Shivashankar* Materials Research Centre, Indian Institute of Science, Bangalore, Karnataka -560012, India ABSTRACT Chromium-substituted β-diketonate complexes of aluminium have been synthesized and employed as precursors for a novel "soft chemistry" process, wherein microwave irradiation of a solution of the complex yields, within minutes, well-crystallized needles of a-(Ali_xCrx)203 measuring 20-30 nm in diameter and 50 nm long. By varying the microwave irradiation parameters and using a surfactant such as polyvinyl pyrrolidone, the crystallite size and shape can be controlled and their agglomeration prevented. These microstructural parameters, as well as the polymorph of the Crsubstituted AI2O3 formed, may also be controlled by employing a different complex. Samples of a(Ali.xCrx)203 have been characterized by XRD, FTIR, and TEM. The technique results in material of homogeneous metal composition, as shown by EDAX, and can be adjusted as desired. The technique has been extended to obtain coatings of a-(Ali.xCrx)203 on Si(100). INTRODUCTION Solid state synthesis has been a well-studied area of research for preparing solid solutions of ceramic oxides. The use of two metal salts to obtain a solid solution of oxide is a strategy that has generally been employed [1-7]. The use of a single precursor, which is a metalorganic complex containing two metal ions, would thus be a new approach to the synthesis of substituted (or bimetallic) oxides. Such an approach has been explored in the work presented here, using microwave assisted synthesis, to obtain a-(Ali.xCrx)203. Α12Οι and Cr203 share the same crystal structure, i.e., corundum, where the metal ions occupy two-thirds of the available octahedral sites in the hexagonal close-packed oxide ion arrangement. Aluminium has an ionic radius of 0.53 A, that of chromium being 0.615 A [8], and the two differing by 13.8%. Hence, at normal temperatures, the two phases are expected to be miscible, without segregating, at low concentrations of the substituting ion. These mixed metal oxides are important refractory materials, and find applications as hard coatings and as refractory coatings. Preparation of homogeneous solid solutions by solid-state synthesis generally requires high temperatures and long processing periods. Such a requirement has prompted the development of nonconventional methods, such as sol-gel synthesis, that provide a high degree of compositional homogeneity under moderate processing conditions [4]. The microwave irradiation technique, apart from providing short reaction times and expanded reaction ranges, also provides a "non-conventional" route to deposit thin films on dielectric substrates. These non-conventional techniques can sometimes result in the production of non-thermodynamic or metastable reaction products [9], as opposed to solid state methods that generally give the thermodynamic product. The microwave irradiation technique also provides a "non-conventional" route to deposit thin films on dielectric substrates. This novel approach of preparing nanoparticles and coatings of bimetallic oxides using a "single source" metalorganic precursor is presented here. This rapid and effective method of preparing coatings has not been described previously in the literature. In this work, a mixed-metal acetylacetonate (denoted as 'acac') complex containing aluminium and chromium, Alo.9oCro.io(acac)3, is used to prepare nanoparticles, and thin films, of the mixed aluminium-chromium metal oxide, (AlxCr].x)203. The

15

Microwave Irradiation-Assisted Method for the Rapid Synthesis of Fine Particles on Si(100)

complex Al(acac)3 was used to make nanoparticles of α-Α^Ο^ by microwave irradiation. The use of such acetylacetonate complexes of aluminium and chromium as precursors for the chemical vapour deposition (CVD) of oxides is well known [10]. These complexes facilitate the formation of oxides because of the direct metal-oxygen bonds in the molecular structure. EXPERIMENTAL The chemical precursor, viz., the substituted metal complex Alo.9oCro.io(acac)3, was synthesized using the co-synthesis techniques, wherein the metal salts Al2(S04)3-16H20 and CrCh.oFbO, were dissolved in stoichiometnc amounts of water. The ligand acetylacetone was added to it, and the solution was neutralized using 1:1 NF^FLO, to obtain a precipitate of Alo.9oCro.iu(acac)3. The complex Al(acac)3 was synthesized in a similar manner. The crude was recrystallized using acetone. The formation of complex containing both metal ions was confirmed by FTIR spectroscopy (The spectra were recorded in KBr pellets within a wavelength range of 400-4000 cm'1.) Melting point determination (using the Buchi-B540 melting point apparatus) indicated that the compound has a single, sharp, and congruent melting point. This was confirmed by simultaneous thermogravimetry and differential thermal analysis (TG/DTA). The metal ion composition of the complex was determined using refinement of single crystal X-ray diffraction data obtained on the complex. The microwave irradiation-assisted synthesis of nanoparticles of (AlxCri-x)203 was carried out in a domestic-type microwave operating at 2.45 GHz, and rated at 800 W. The metalorganic complexes Al(acac)3 Al90Cr i0(acac)3 were used as precursors. For this purpose, 0.5 g of the precursor was taken and dissolved in 20 ml of chloroform. To this, 0.05 g of a capping reagent, i.e., polyvinyl pyrrolidine (PVP), and 5 ml of glycerol were added, the latter to aid dissolution. A double-necked round bottom flask fitted with a condenser (and a substrate holder fitted on the other neck) was used for the synthesis. To obtain nanoparticles, the reaction mixture in the round bottom flask was placed at the centre of the microwave oven, and subjected to irradiation at full power (800 W) for 15 minutes. To obtain coatings, a Si(100) substrate measuring ~1 cm' was placed in the substrate holder. Microwave irradiation of the solution results in a cloudy suspension, which is subjected to centrifugation at 5000 rpm for 10 minutes, leading to a precipitate, which was washed several times with acetone. It was then calcined at 500°C for 4 hours to remove the capping agent PVP. In a similar manner, the coated substrate was also calcined to remove the PVP. The crystalline phase of the calcined powder was identified by X-ray diffraction (XRD) using a Siemens model D500 diffractometer with Cu Ku source in the Bragg-Brentano geometry. The average particle size of the particles was estimated using the Scherrer formula. FTIR spectra of the powder samples were obtained using a Digilab FTS60A. These samples were made by pressing KBr pellets at a particle-to-KBr mass ratio of 1:25. The samples were examined by transmission electron microscopy (TEM) in a JEOL CX200 microscope (operated at 120 kV). The TEM specimens were prepared by dispersing the powder in cyclohexane with the aid of ultrasonic agitation. A few drops of this were poured onto a porous carbon film supported on a copper grid, then dried in air. RESULTS AND DISCUSSION The substituted complex was characterized by FT-IR spectroscopy to confirm the formation of the complex. The FT-IR spectrum is shown in figure 1. The spectrum shows the presence of Cr—O and Al—O indicating the presence of both chromium and aluminium in the complex. The peaks at 400

16



- 500 cm"1 correspond to the metal-oxygen bond, confirming that there is bonding of the metal ion with oxygen of the ligand. The peaks at 460 cm"1, 594 cm"1, 659.00 cm'1 confirm the Cr-0 bond and the peaks at 491 cm"1 and 577 cm"1 confirm the presence of Al-0 bond. In case of the pure Al(acac)3 complex, only the Al-0 bonds were observed.

Fig. 1 FTIR spectrum for the precursor Alo.9oCro.io(acac)3 The various other bands observed were matched with the standard stretching and bending modes for various bonds. The precursor showed a single congruent melting point at 195°C. This indicates that chromium is substituted into the Al(acac)3 lattice and there is no phase segregation. This is important as it ensures the uniformity, homogeneity and constant composition of the precursor. The complex Al(acac)3 showed the lower melting point of 192°C, consistent with the variation of melting points in a solid solution series.. The metal ion ratio (Al:Cr) in the substituted complex was estimated to be 90:10, on the basis of the refinement of the lattice parameters of the crystalline complex, using single crystal X-ray diffraction. Microwave irradiation of the substituted complex yielded a uniformly green precipitate, green being characteristic of a-Cr 2 03. XRD indicated that this powder was crystalline, and that the pattern (Fig. 2) could be indexed to the corundum phase. Microwave irradiation of the aluminium precursor gave a white powder whose XRD pattern could be matched to α-Α1203 ( JCPDS No. 85-1337 )


· 17


Fig.2. XRD pattern of (Α1 χ Ον χ ) 2 0 3 The lattice constants deduced from the data were close to those of a-C^Oi (JCPDS card no.850869). As the metal ion ratio in the precursor complex was Al:Cr = 90:10, and as the powder was green in color, it was surmised that the microwave irradiation of the substituted complex had led to the separation of AI2O3 and Cr 2 03. Thus, the powder XRD pattern was analyzed carefully for the presence of peaks due to a-A^C^. But, no separate peaks were found assignable to (X-AI2O3. These data indicate that, though the metal ion ratio in the precursor complex is Al:Cr=90:10, the metal ion ratio in the crystalline powder that results from microwave irradiation of the complex is "skewed" in favor of Cr. That is, Al substitutes partially for Cr in a-Cr2U3 (as confirmed by the elemental analysis related below). The width of the peaks in the x-ray pattern indicated that the particles were nano-sized. Using the Scherrer equation and the FWHM of the broad peaks, the average particle size was calculated to be 30-35 nm. FT-IR of the nanoparticles recorded in KBr showed the presence of Cr—O and Al—O. The peak at 566 cm"1 corresponds to the Al-0 bond and the one at 626 cm"1 corresponds to the Cr-0 bond. FT-IR spectra of the AI2O3 nanoparticles showed only the Al-0 bonds. TEM micrographs of the powder sample are shown in figure 3. The nanoparticles were observed to be needle like in shape, with a length of 100 nm and a diameter of 20-30 nm.

Fig.3 TEM micrographs of the Al doped Cr2C>3 nanoparticles.

18



EDS compositional analysis showed a uniform composition of 80% Cr and 20% Al in the powder. Spot EDS taken on various nanoparticle aggregates showed that the metal ion distribution was uniform and homogeneous throughout. These EDS results are consistent with the XRD results, where the pattern matches that of CX-O2O3, but with a slight shift in the 2Θ values of all the peaks. These results are counterintuitive, in that the precursor is aluminium-rich. This implies that, during the microwave processing under the conditions used here, the interaction of the complex with the microwave field "segregates" the "aluminium part" of the precursor from its "chromium part". Since the Al—O bond is smaller than the Cr—O bond, and the Al(acac)3 molecule is smaller than the Cr(acac)3 molecule, Al(acac)3 is expected to be more stable. Thus, during the microwave processing, Cr(acac)3 decomposes to form Cr2U3 while much of the Al(acac)3 remains undecomposed in the solution. This would explain the formation of chromium-rich nanoparticles. A complete understanding of the process leading to the formation of oxide nanoparticles from the metalorganic complex must take into account the interaction of the solvent (chloroform) and the capping agent (PVP) with the microwave field. The presence of water in the solution can also influence the reaction process. A detailed investigation is under way. The coating formed on Si(100) was characterized by XRD. The pattern matches that of aΟ2Ο3, with all peaks shifted by 0.5°. There are no peaks due to a second phase, viz., a-alumina (corundum), just as ion the case of the powder formed under microwave irradiation. Thus, the coating on Si( 100) was deduced to be that of Al-substituted Cr^C^. The XRD pattern is shown in fig.4.

Fig.4 XRD pattern of coating of Al doped Cr 2 0 3 on Si(100) EDS analysis showed that the composition of the coating was homogenous and uniform throughout, and that the metal ion ratio in the coating corresponds to Al:Cr::20:80. This result is consistent with the composition from EDS obtained on the nanoparticles. This is expected as the thin film and the nanoparticles were produced using similar reaction conditions. A SEM micrograph of the coating is shown in Fig.5, displaying its morphology.


· 19


Fig.5 SEM of the film on the silicon substrate It can be seen that coating is continuous, but is not of uniform thickness, and that it is spotted by aggregations. CONCLUSIONS Nanoparticles of bimetallic oxide (Ali.xCrx)203 have been synthesized by using a novel microwave irradiation technique using a single precursor of the type (AlxCri-x) (acac)3. Similarly, nanoparticles of AI2O3 have been prepared using the microwave irradiation technique using the metalorganic precursor Al(acac)3The particles of the substituted oxide have been analyzed by XRD, EDAX, FTIR and TEM. X-ray diffraction confirms the formation of α-(Οο.8Α1ο.2)2θ3, while EDS confirms the metal ion ratio and compositional homogeneity in the crystallites. TEM analysis shows that particle diameter is in the range 20-30 nm and that they are 50 nm long. The microwave irradiation technique has also been extended to obtain continuous, adherent, and compositionally uniform coatings of a-(Cro.8Alo.2)203 on Si(l00). The microwave-assisted technique offers a means for the rapid synthesis of nanoparticles of refractory oxides, and a one-step method for the synthesis of substituted metal oxides. Preliminary results indicate that the technique can be extended to obtain oxide coatings on dielectric substrates. REFERENCES 1.

A. Neuhaus, Physics and Chemistry of High Pressure, ed. Edited by the Society of Chemical Industry. 1963, New York: Gordon and Breach Publishers.

2.

D.M. Roy and R.E. Barks, Nature Phys.Sci, 235 (1972). 118.

3.

W. Sitte, Mater. Sci. Monogr., 28A (1985). 451.

4.

B. Durand, Ceramics Powders. 1983, Amsterdam,The Netherlands: Elsevier Science Publishers. 413-420.

5.

R.M. Spring and S.L. Bender, Journal of the American Cerammic Society, 45(10), (1962). 506506.

20



6.

L.R. Rossi and W.G. Lawrence, Journal of American Ceramic Society, 53(l 1), (1970). 604608.

7.

W.H. Gitzen, Alumina as a Ceramic Material. 1970, Westerville,OH: The American Chemical Society. Chapter 4.

8.

R.D. Shannon and C.T. Prewitt, Acta Cryst., B25 (1969). 925.

9.

P. Lidstorm, J. Tierney, B. Wathey, and W. Jacob, Tetrahedron, 57 (2001). 9225-9283.

10.

D.F. Bradley and A.R. Barron, Chemical Vapour Deposition, 7(2), (2001). 62-66.


· 21


Chemical Vapor Deposition


SYNTHESIS AND CHARACTERIZATION OF Si/Si2N20/Si3N4 COMPOSITES FROM SOLID-GAS PRECURSOR SYSTEM VIA CVD J. C. Flores-Garcia, A. L. Leal-Cruz and M. I. Pech-Canul* Centro de Investigacion y de Estudios Avanzados del IPN Unidad Saltillo Carr. Saltillo-Monterrey km 13, Coahuila, Mexico, 25900. ABSTRACT Si/Si2N20/Si3N4 composites were prepared via chemical vapor deposition (CVD) using a solid-gas precursor system. In this method the Si-F gas species generated from the thermal decomposition of a solid precursor react with a nitrogen precursor gas to produce in situ silicon nitride and oxynitride. Cylindrical preforms (3 cm in diameter x 1.25 cm long) with 50 % porosity - prepared by the uniaxial compaction of Si powders with average particle size of 12.4 μιη - were infiltrated in a multiple step mode with high purity nitrogen (HPN) and the Si-F gas species in a chemical vapor deposition (CVD) reactor. The specimens were heated at a rate of 15 °C/min up to a processing temperature of 1300 °C, and maintained isothermally for 70 minutes; then they were cooled down to room temperature. Results from the characterization by XRD and SEM show the deposition of Si2N20 and S13N4 on the Si particles, occupying the interstices of the porous preforms. S13N4 is typically deposited with a sponge-like structure and compact deposits, while Si2N20 is formed with a pin-like morphology and as whiskers. Depending on the number of infiltration steps, Si/Si2N20/Si3N4 composites with tailorable porosity can be prepared. INTRODUCTION The broad range of studies devoted to the development of composite materials has embraced two of the most important advanced ceramics, silicon nitride (S13N4) and oxynitride (Si2N20). This fact is not surprising because both compounds possess excellent properties ideal equally for structural and functional applications. Interestingly, application of composites' paradigm to these materials includes not only dense, but also porous materials. Individually, S13N4 has received much attention because of its high mechanical properties, excellent chemical resistance and high thermal properties. On the other hand, Si2N20 has been recognized by its excellent oxidation resistance in a variety of environments. Accordingly, Si3N4/Si2N20 composites are promising candidates for applications demanding high mechanical strength and oxidation resistance. Moreover, this trend includes incorporation of a third constituent, like in the case of Si3N4-Si2N20-TiN composites. Porous composites of these materials offer the potential to be used as hot gas and molten metal filters, support for catalysts, thermal insulation materials, and preforms for the manufacture of metal/ceramic (using further liquid metal infiltration) or dense ceramic/ceramic composites. Specific suggested applications for porous Si3N4/Si2N20 composites include: filters for the purification systems of air and water and Diesel Paniculate Filters (DPF), targeting the reduction of diesel PM (Paniculate Matter) emissions in the automotive area. In addition to compatibility with the catalyst, DPFs must have excellent thermal shock resistance, thermal and chemical stability, and high surface area. Several processing routes have been reported in the recent literature for these composites [1-11]. For instance porous Si3N4-Si2N20 bodies have been fabricated by the multi-pass extrusion process. This route consists of a nitridation process performed at 1400 °C in flowing N2 gas for 20 hours [1, 2]. From another perspective, silicon nitride and oxynitride composites are

25

Synthesis and Characterization of Si/Si2N20/Si3N4 Composites from Solid-Gas Precursor

sought out for superplasticity related applications [3,4]. As it is now well known, silicon nitride exhibits superplastic-like behavior, offering the potential for the fabrication of complex near-netshape components, and reducing the high costs associated to machining. Preparation of superplastic nitrides typically starts from ultrafine P-Si3N4 and involves a lengthy series of operations, including ball-milling in hexane, and hot-pressing. Obviously these operations impact the processing costs; for example, hot pressing is conducted at a pressure of 20 MPa, and at 1750 °C under nitrogen at 1 atmosphere [3,4]. The hybrid precursor chemical vapor deposition method (HYSYCVD) offers the potential to produce silicon nitride and oxynitride composites with advantageous processing characteristics as compared to many previous processing routes. HYSYCVD is a method recently developed at Cinvestav-Saltillo (Cinvestav is the Spanish acronym for Center for Research and Advanced Studies; Saltillo is the location city), for the production of advanced ceramics (for example, alpha- and beta-Si3N4) in which solid-gas reaction systems or hybrid systems are used [5]. This method is based on the ability of some solids like sodium hexafluorosilicate (Na2SiF6) to produce highly reactive gas species (S1F3, S1F2, SiF and Si) which, by the reaction with nitrogen precursors allow the formation of condensed phases as stable solids [6]. As in the conventional CVD route for the processing of S13N4, a number of parameters such as nitrogen precursor, nitrogen precursor flow-rate, substrate condition, and processing time and temperature, significantly influence the occurrence and amount of phases (S12N2O, a- and P-S13N4). HYSYCVD allows reducing processing times and temperatures (70 min per infiltration step, at a processing temperature of 1300 °C), and more importantly, does not require to start from any kind of nitride seed (S13N4 or S12N2O) as is the case in other processing routes which use ultrafine a- or p-Si3N4. Of prime importance -in terms of the paradigm of composites- is the fact that the method allows producing reinforcing phases in an in situ mode, resulting in selfreinforced materials. Depending on the processing parameters, fine powders, whiskers and fibers, which can act as reinforcements, are usually obtained. Another quintessential feature of the method is the feasibility to produce either porous or dense composites by programming successive infiltrations. In this particular investigation, authors report on the microstructure characteristics of composites containing silicon nitride and oxynitride using the HYSYCVD method, obtained by the multiple infiltration approach. EXPERIMENTAL PROCEDURES Processing of composites Silicon (Si powders of 12.4 μιτι average particle size) cylindrical preforms (3 cm in diameter x 1.25 cm high) with 50 % porosity and sodium hexafluorosilicate (Na2SiF6) compacts were prepared by the uniaxial compaction of the corresponding powders in a steel-die. Both, the Si and Na2SiF6 powders were supplied by Sigma Aldrich Inc. Processing of composites was carried out in a hybrid CVD reactor, which consists of a horizontal tube furnace (high-alumina tube, internal diameter =3.175 cm and length = 76.2 cm) with end-cap fittings. The reactor is equipped with gas inlets and outlets to supply the nitrogen precursor (high purity nitrogen HPN, 99.997%, 02(g) (< 5 ppm) or Η 2 0 (?) (< 5 ppm)) as well as with devices to control flow rate, pressure and process atmosphere. Figure 1 is a schematic of the experimental set-up used in the investigation. Sip porous performs were consistently positioned within the tube at the center of the reaction chamber (in the high temperature zone) and the silicon hexafluorosilicate compacts

26



were placed in a low temperature zone nearby the gas entrance. The thermal gradients along the tube longitudinal axis are used to induce the Na2SiF6 decomposition, for the production of silicon-fluorine gaseous reacting species necessary for the formation of silicon nitride (Si3N4) and oxynitride (Si2N20), and their subsequent deposition. The Sip porous preforms (the subscript p stands for paniculate) were heated in high purity nitrogen (HPN) at a rate of 15 °C/min up to 1300 °C and then maintained at this temperature for 70 min at a constant gage pressure of 12 mbars. Simultaneously, Na2SiF6 was heated (in the temperature range 250-550 °C) and decomposed to generate the Si-F gas species. After completing the test time, the specimens were cooled down to room temperature at a rate of 20 °C/min. In order to maximize deposition, the Sip preforms were processed in four steps by alternating the preform deposition sides from one to another stage. Table I summarizes the conditions under which the trials were carried out. After the infiltration tests, the specimens were removed from the reactor for characterization by XRD, SEM, and EDS.

Fig. 1. Schematic representation of the experimental set-up. rable I. Processing parameters used in the lrihltration tests. 1 Processing parameters Temperature 1300°C - . - . - . Heating rate 15°C/min I Time 70 min Gage pressure 12 mbar Nitrogen precursor High purity nitrogen (HPN) N2 99.997 % Flow rate of HPN 15cnrVmin Porosity of porous preform (Si) 50 volume % of ceramic 1 Quantity of Na2SiF6 (mass) 20 g


] . | | 1 | 1 |

· 27


It should be noted that the processing conditions used in the current work were optimized in previous investigations by the same authors, using only one infiltration step [71. Characterization of raw materials and specimens Initially sodium hexafluorosilicate and silicon powders were characterized by X-ray diffraction (XRD), scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), and thermal analysis (TG/DTA). In order to identify the phases formed in situ, representative samples from each of the specimens were analyzed by XRD. A Philips model 3040 x-ray diffractometer was used under the following conditions: excitation voltage of the anode of 40 kV and current of 30 mA; monochromatic Cu Ka radiation (λ = 1.541838 A); scanning range of 10-80 2Θ degrees, at a scanning speed of 0.02 degrees/sec. In order to determine phase type and morphology, distribution and composition, the specimens were analyzed by SEM and EDS using a Philips XL 30 ESEM scanning electron microscope provided with an EDAX energy dispersive x-ray spectroscopy (EDS) microanalysis device. Both, secondary and backscattering electron modes were used in the analysis at an acceleration voltage between 20 and 30 kV. Prior to the analysis, the specimens were coated with gold. Thermogravimetry (TGVdifferential thermal analysis (DTA) was used only for Na2SiF6. Thermal analysis was performed in a Perkin Elmer model TGA7 thermogravimetric analyzer using a heating rate of 10 °C/min, scanning temperature range from 50 to 1300 °C, and nitrogen atmosphere (HPN) at flow rate of 15 cnrVmin. RESULTS AND DISCUSSION Sodium hexafluorosilicate (Na2SiF6) Figure 2 shows an x-ray diffraction pattern of sodium hexafluorosilicate powders. All reflections correspond to Na2SiF6 according to the Joint Committee on Powder Diffraction Standards (JCPDS) - International Centre for Diffraction Data (ICDD) file No. 72-1115.

Fig. 2. X-ray diffraction pattern of Na2SiF6.

28



Figure 3a) is a SEM photomicrograph showing sodium hexafluorosilicate with rosettelike geometry, constituted by prismatic particles with sizes in the range of 25 to 40 μπι. Analysis by EDX allows confirming the presence of Si, F and Na element peaks corresponding to Na2SiF6.

Fig. 3. a) SEM photomicrograph and b) EDS spectrum of Na2SiF6 powders. Results from the thermal analysis of Na2SiF6 are shown in Figure 4.

Fig. 4. Thermal analysis of Na2SiF6by TG/DTA. The thermogram in Fig. 4 (curve A) shows a maximum weight loss of approximately 52 % in the temperature range 543-565 °C, which is attributed to the decomposition of Na2SiF6 into silicon-fluorine gaseous species (SiFx where x=0-4) and sodium fluoride (NaF). A second weight loss of approximately 29 % which occurs in the temperature range 965-1200 °C is attributed to the superficial evaporation of molten NaF. In curve B, DTA results show three endothermic events; the first occurs at 559 °C and corresponds to the decomposition of the sodium hexafluorosilicate; the second appears at 972 °C and can be attributed to the transition from NaF(s) to NaF(D (m. p. 993 °C), while the third, which appears at 1024 °C may be associated to the evaporation of molten NaF.


· 29


Specimen phase analysis and microstructure Figure 5 corresponds to a representative X-ray diffraction pattern of specimens before the infiltrations. Before the treatment, reflections pertaining only to the silicon particles are revealed. 4000 η

Si*

3500-

ro

o o I . I .

ro

ai

§

i►

♦

I . I .

g o

Intensity ι counts)

3000-

500-

01

10

1

20

1

1 1

30

♦

.

1

40

1

1

1

1

50

1

1

60

1

1

1

70

1

80

2Θ (degrees)

Fig. 5. X-ray diffraction patterns of specimen before thermal treatment. Si ♦

1/ Γ1000+■· c3 800o o >» 600< / c> *■*

0)

c

400-

▼

10

20

30

500 °C), there was a chlorine excess of 29 g (i.e. 25 %). The impurities share approximately 15 % of the fed chlorine in form of MeCl, MeCl2, MeCl3, MeCl4 and MeCl6. The relatively low chlorination efficiency of boron is then explainable only due to the low heap height of the bed through which the unutilized chlorine gas flows. In the experiment 2 at 750 °C, the boron chlorination efficiency was 84 % with a chlorine consumption rate of 91 %. Only 10.7 g Cl? of given 114 g was unutilized by means of higher heap of the reaction bed and longer retention time of Ch. This improvement can be interpreted also in regard of the well-sloped gradient at the beginning of the reaction between less chlorine and much boron. FTIR spectrums of the gas samples produced in experiment 2 are shown in Figure 5. The product gas shows the same characteristic peaks with the reference one. In Figure 5 (b) and (d), the characteristic peak of BCI3 disappeared. This means that there is a successful gas capturing in the gas-wash-bottle with glass beads and PTFE rashing rings. It takes approximately 35 minutes that CI2 gas contacts with the B4C layer. So, the absorbance of the peak is very low at the 5th minute after the first contact. Also, at 1190 cm"1 in Figure 5, there is a peak that presents in all samples. This peak probably belongs to a gaseous compound which is much more volatile than BCI3. In Figure 5 (a)-(d), very weak peak of CCI4 is detectable at around 800 cm"1, although it seems not probable that CC14 slips away from the second heat exchanger and reaches the gas sampling chamber. This can be explained due to the relatively higher vapor pressure (400 mbar at -10 °C) of BCI3. However, the vapor pressure of CCI4 at the same temperature is only 26 mbar. It can be dragged by BCI3 in very small but detectable amounts. This case is in the favor of reaction (1) but it is unexpected according to reaction (2).

Figure 5. FTIR spectrum of the gas samples produced in experiment 2, (a) taken from 1. sampling chamber after 40 minutes, (b) from 2. sampling chamber after 50 minutes, (c) taken from 1. sampling chamber after 80 minutes, (d) from 2. sampling chamber after 90 minutes and (e) sample of gas mixture containing 95 % BCI3 and 5 % N2, (reference).

84



The SEM images of the B4C powder before and after experiment 2 are illustrated in Figure 6. There are not very exact differences between (a) and (b). However, there is reduction in particle size and some white clusters disappear. EDS analyses of the powders from the pointed regions are represented in Figure 7. The EDS spectra indicating the powder before and after the chlorination reaction show the rising graphite peak. This case is contrary to the reaction (1), but in favor of reaction (2). So, the rising graphite peak in Figure 4 and 7, also the weak peak of CC14 in Figure 5 give us pause to think that both reactions (1) and (2) occur partially and simultaneously.

(a)

(b)

Figure 6. SEM images of B4C powder (750X), (a) before reaction (b) after reaction.


· 85


Figure 7. EDS analysis of B4C powder, (a) before reaction, from a region (b) after reaction, from b region. In the second experimental set-up, different type of impurities formed as orange-reddish dendrites at the cold bottom region, dark-colored precipitates at the top region and white precipitates at the quartz cover of the reactor. The SEM images of the impurities are illustrated in Figure 8. The EDS analysis from region a, b and c pointed at Figure 8 is shown in Figure 9. Although EDS determined all the elements in the impurities, XRD analyzes only the dominant phases in (a) as A1C13.6H20 (JCPDS Card No: 44-1473); in (b) and (c) as A1C13.6H20 and H3BO3 (JCPDS Card No: 75-1127).

86



Figure 8. SEM images of the impurities, (a) orange-reddish dendrites at the cold bottom region (b) dark-colored precipitates at the top region (c) white precipitates at the quartz cover of the reactor.


· 87


The orange-reddish dendrites include W, Al and Si; dark-colored precipitates include W, Al, Fe and Si; white precipitates include Al, Si, Cr and Fe metallic impurities. Most of the

88



metallic impurities in the powder form also volatile chlorides with relative high densities. They went through the quartz frit down to the bottom of the reactor. This can be called as a "self-cleaning" effect. CONCLUSION Production of BC13 from native B4C was tested in two different originally designed set-ups. First set-up was not appropriate for the efficient production of BCI3. Because the spiral exchangers were insufficient in length, temperature and contact surface; Cl2 flow rate was too high for chosen amount of B4C. So, high amount of free Cl2 went to the ultra freezer with produced BCI3 and a yellowish-green liquid was obtained as a mixture of two gases. The second set-up was enough to produce BCI3 with boron chlorination efficiency of 84 %. The second set-up was superior in capturing of BC13. However, CCU in very small amounts dragged with BCI3. Rising of free graphite peak and dragging of CCU proves that both reactions, (1) and (2), were occurred simultaneously. For production of BCI3, the B4C powder does not need to be extremely pure. Most of the metallic impurities in the powder form volatile chlorides but they went down to the bottom of the reactor. Thus, there is a self-cleaning effect. Consequently, the experiments prove that the heap height of the bed elongates the solidgas contact time and provides effective chlorine utilization. These facts have to be considered for deciding to work in fix-bed or fluidized-bed reactor. ACKNOWLEDGMENTS This work has been supported by The Scientific and Technological Research Council of Turkey within with the project number of 106M087. REFERENCES 'Boron Report, The National Boron Research Institute of Turkey, BOREN, Ankara (2004). F. Habashi, Handbook of Extractive Metallurgy, Vol IV, WILEY-VCH, 1986-2021 (1997). 3 K. Othmer, Encyclopedia of Chemical Technology, Vol IV, Wiley-Interscience Publication, John Wiley&Sons, 129-132 (1978). 4 R.C. Hyer, S.M. Freund, A. Hartford and J.H. Atencio, Selective Removal of Phosgene Impurity from Boron Trichloride by Photochemical Dissociation, Journal of Applied Physics, Vol 52, 6944-6948 (1981). 5 C. Marks, Process for the Manufacture of Boron Nitride, United Kingdom Patent, No: 711254 dated 30.6.1954. 6 R.C. Davis, J.N. Haimsohn and J.T. Bashour, Manufacture of Boron Trichloride, united States Patent, No: 3025138 dated 13.3.1962. 7 W.P. Thompson, Improvements in and Relating to Mineral Active Carbons and to a Process for Their Preparation, united Kingdom Patent, No: 971943 dated 7.1.1964. 8 G. Kratel and G. Vogt, Process for Manufacturing Boron Halides, United States Patent, No: 3743698 dated 3.7.1973. 9 G. Anmelder, Verfahren zur Kontinuierlichen Darstellung von Borhalogeniden, German Patent, No: 2826747 dated 3.1.1980. 10 A.C. Jones, G. Williams and A.B. Leese, Process for Producing Boron Trichloride, United Kingdom Patent, No: 2304104 dated 12.3.1997. 2


· 89


11 J.A. Merritt and L.C. Robertson, Removal of Phosgene Impurity from Boron Trichloride by Laser Radiation, United States Patent, No: 4063896 dated 20.12.1977. 12 H.R. Bachmann, H. Noth and R. Rinck, Infrared Laser-Specific Reactions Involving Boron Compounds III: Decomposition of Phosgene Sensitized by Boron Trichloride, Journal of Photochemistry, 10, 433-437 (1979).

* Corresponding author. Tel.: +90 2122856893; Fax: +90 2122853357 E-mail address: [email protected] (Duygu Agaogullari).

90



Sintering and Hot Pressing


SPARK PLASMA SINTERED ALUMINA-ZIRCONIA NANO-COMPOSITES BY ADDITION OF HYDROXYAPATITE

'college of Science and Technology, Nihon University, 7-24-1, Narashinodai, Funabashi. Chiba, 274-850 1, Japan 2 ~ o l l e gof e Science and Technology. Nihon University, 1-8-14, Kandasurugadai, Chiyoda, Tokyo, 101-8308, Japan 3 ~ o l l e gof e Materials Science and Technology, Xi'an University ofTechnology, 5 South Jinhua Road, Xi'an Shaanxi, 710048, China ABSTRACT Zirconia-alumina composites with addition of different volume fraction of hydroxyapatite (HA) were fabricated successfully using spark plasma sintering (SPS). The densification behavior, microstructure and mechanical properties of composites are investigated as a function of sintering temperature and HA content respectively. The sintering temperature has a significant effect on the final densities achieved in the ZrOz-Al2OdHA compacts. The addition of HA has a barrier effect on diffusion between grains of Z r 0 2 and A1203 and thus limit the grain growth of ZrOz and A1203. Sintering the Zr02-AI2O3/HAcomposites at 1400 OC led to the decomposition of HA in the samples. Flexural strength, fracture toughness and Vickers hardness values increase with increasing sintering temperature, and show decreasing trend with increasing of HA content at the same temperature. They compared well with densities obtained at different sintering temperature. The maximum flexural strength, fracture toughness and Vickers hardness of 967.1MPa, 5.27 M ~ a . m l "and 13.26 GPa were achieved for ZrO?-Al?O, composite respectively. Flexural strength. fracture toughness and Vickers hardness values of the Zr02-A1203/HA composite fell within the value range of dense HA and of Zr02-AI20i composite.

Hydroxyapatite (HA, Calo(P04)6(OH)z)is a kind of calcium phosphate bioceramic materials. Owing to its chemical and structural similarity with natural bone mineral, it has a unique capability of binding to the natural bone through biochemical bonding, which promotes the interaction between host bone and graft material' I . :-I. Densified HA has attracted extensive attention for its excellent biocompatibility, and has been widely used for bone substitute and grafting; but the low mechanical properties restricted its application for load-bearing implants. The fracture toughness of the monolithic HA ceramics . with 2-12 am"' for human bone':. -1:. Investigations does not exceed 2 M ~ a m " ~coinpared aimed at broadening the medical application potential of implant materials based on HA are carried out in scientific research institutes around the world. The investigations mainly focus on the two aspects: one approach is to use HA as a coating on a strong metallic or ceramic substrate such as zirconia or titanium alloy. One of the limitations of this approach is susceptibility of the HA coating to de-bond from the substrate"!. Another attractive approach is to maintain the biocompatibility of HA and improve its mechanical properties by introducing thc concept of fabricating composites either with HA as substrate[(.. -!or with HA as additive[:. "1.

Spark Plasma Sintered Alumina-Zirconia Nano-Composites by Addition of Hydroxyapatite

Yttria stabilized zirconia (Zr02(Y203)), alumina (A1203) and their composites have been considered as substrate or reinforcement phases for use in implants, such as prostheses and dental materials due to their excellent biocompatibility as well as their desirable material properties, such as strength, chemical stability, and wear resistance . The aluminatoughened zirconia (ATZ) Bio-Hip*', developed by Metoxit AG (Thayngen, Switzerland), has a bend strength of up to 2000MPa , indicating that it can withstand loads that are nearly twenty times greater than densified HA. But as it is well known, Zr0 2 , ΑΙ2Ο3, and their composites are classified as bioinert materials. Therefore, in order to increase the biocompatibility, calcium phosphate such as HA are considered to be the most suitable bioactive material to be used as an additive. Shen et al. reported that the bend strength and the fracture toughness of HA-50 vol% Zr02 composite was 440 MPa and 2.5 MPa.m1/2, respectively . Miao et al. reported that the HA-40 wt.% Zr02(Y203) composites sintered at 1200 °C showed 200 MPa bend strength However, few detailed reports on the microstructure change and mechanical properties of HA added ΖΓ02(Υ20 3 )-Α1 2 03 composites have been published. Spark plasma sintering (SPS) processing of monolith and composite ceramics materials has been recognized to offer a number of advantages over conventional sintering approaches. Increase of densification rates at relatively low sintering temperatures, makes the SPS fabricating technique significantly faster than the conventional process, which means relatively shorter sintering time and less energy consumption. It was concluded by some, that SPS has the potential for enhanced densification and suppressed grain growth due to a fast heating rate and apparent low firing temperature. Based on the advantages mentioned previously, the application of SPS processing in this study may solve the serious problem of extensive reaction between HA and Zr0 2 to form tricalcium phosphate (TCP), and avoid the fully stabilized Zr0 2 . The forming of TCP would lead to the serious reduction in the biocompatibility of HA, and fully stabilized ZrC>2 would cause the decreasing of strength and toughness, which mainly depends on the phase transformation from the tetragonal phase to the monoclinic phase The aim of this study was to fabricate Zr02(Y203)-Al203 composites with addition of different volume fraction of HA at various sintering temperatures using spark plasma sintering. The mechanical properties and densification behavior of composites are investigated as a function of HA content and correlated with the compositional and structural variations. And this will be the foundation of the final aim; to fabricate the functionally graded materials (FGM) based on ATZ composites. EXPERIMENTAL PROCEDURES Preparation and Characterization of Powders The starting materials used in this study were two kinds of powders. One was a commercially available nano-composite powder, 3-mol% yttria-stabilized zirconia (3-YSZ) reinforced with 20 wt% alumina (Tosoh Co., Japan, average particle size 20μηι, specific surface area 13 m2/g and crystallite size 28nm) was used in the as-received state, which is denoted as TZP-3Y20A in this study. Another commercially available hydroxyapatite powder (HA, Taihei Chemical, Japan) was used for this work. Appropriate quantities of TZP-3Y20A and HA powders were slurry mixed in acetone, and ball milled for 24 h using zirconia balls to obtain a homogenous mixture with a composition ranging from 10-50% volume fraction of HA. Acetone was then evaporated and the powder was dried, crushed and sieved through a 500 mesh sieve (25

94

·



μιη granules). To determine the phase composition and purity, X-ray diffraction (XRD) was performed on the starting powders as well as the powder samples obtained by grinding the sintered samples. The powders were mounted on a borosilicate glass slide and scanned on a RINT 2000 X-ray diffractometer (Rigaku, Japan) with a Cu-Ka radiation source at 30 mA and 40 kV with a scan speed of 0.5°/min and steps of 0.02°. The morphology of the powder precursor was also examined both on a SSX-550 scanning electron microscope (SEM; Shimadzu, Japan) equipped with an energy dispersive X-ray spectrometer (EDX) attachment, and a transmission electron microscope (TEM; JEOL-JEM3010, Japan). Sample Preparation For sintering, the powders were loaded into a cylindrical graphite die and uniaxially pressed into green compacts with dimensions of 56 mm x 11 mm x 2 mm by using a hydraulic pressure of 10 MPa. Prior to this step, the interior surface of the die was sprayed with a layer of boron nitride to lubricate and prevent diffusion between the graphite and compact at high temperatures. The green compacts were sintered in an SPS-3.2MK-IV system (Sumitomo Coal Mining, Japan). The sintering was performed in the temperature range of 1000-1400°C with steps of 100°C at a heating rate of 200°C/min. The temperature was measured by means of an optical infrared thermometer focused on to the graphite die surface. The sintering pressure was set at 44.6 MPa. The vacuum level of the chamber was kept below 10 Pa during sintering. After heating at the desired temperature for 8 min, the power was turned off, and the sample was cooled in the chamber to less than 300°C at a cooling rate of 100°C/min. Characterization of Sintered Samples and Mechanical Testing The density of the sintered samples was determined by using Archimedes' method with distilled water. The composites relative density was determined using a theoretical density of 3.94 g/cm3 for alumina, of 6.00 g/cm3 for 3-YSZ and 5.50 g/cm3 for TZP-3Y20A (density supplied by the manufacture). Five measurements were conduct to obtain an average value. The crystal phases in the sample were identified by using XRD referenced to the standard ICDD PDF cards available in the system software. In addition, the microstructure evolution of the dense samples, including the fracture surface, with respect to various sintering temperatures was examined by using SEM. The surfaces prepared for examination in the SEM were coated with gold using a HITACHI E-1030 sputter coater operated at 15 mA for 30 s. The grain size was estimated from the full width at half maximum (FWHM) by the Scherrer equation|l6| and confirmed by TEM micrographs and by the line intercept method with SEM micrographs. Three-point bending tests were conducted on an Instron-5500R tensile tester (Instron Corp., Canton, MA, USA) with a cross-head speed of 0.5 mm/min with an inner span of 30 mm. Test samples with dimensions of 56 mm x 11 mm x 2 mm were ground and polished for the three-point bending test. Three samples sintered under the same conditions were prepared for the bending test in order to obtain an average value. The flexural strength, σ, was calculated using _ 3PL a

~2bd\

(1)


· 95


where P is the applied load, and L, b, and d are the span length, the width, and the thickness of the specimen, respectively. The Young's modulus E was calculated using E=

L3 (/>-/>) 4&/3 " ( & - $ ) ^

(2)

where P, (/ = 1 and 2) is the applied load, and Si (/ = 1 and 2) is the displacement, and L, £, and d are the same as equation (1). Fracture toughness measurements were conducted by using an indentation fracture (IF) method on a Vickers microhardness tester (Shimadzu HSV-30). Prior to indentation, the crosssectional surface of the samples was polished to a 3 μιτι surface finish. In order to evaluate the fracture toughness of the sintered samples, the median lengths of the radial cracks at the four corners of the indentation trace were measured using a Vickers microindenter at a load of 98 N applied for 20 s. Average hardness and radial crack length values were determined from 10 indentations made for each sample. Fracture toughness, Kic, was calculated based on the median crack equation: / O =0.016·

E

1 c3,2 L

J

,

where C is the radial crack length, and Hv is the Vickers hardness.

(3 )

RESULTS AND DISCUSSION Characteristics of powders The TEM image shown in Figure 1 (a) is the nano-powder of HA, it can be seen from the image that the HA particles show a needle-like morphology, the particle size of the short axis is 20 nm and shows a narrow distribution. Figure 1 (b) shows the micrograph of TZP-3Y20A/HA10vol% after ball milling for 24h. Most of the particles have a uniform size and the TZP-3Y20A particles are smaller than 100 nm. It can be observed that agglomeration and close bonding formed between the particles. The XRD pattern shows that the TZP-3Y20A composite powder consists of a mixture of monoclinic and tetragonal phases of Zr0 2 and α-Α1203 powders, as shown in Figure 2. Average crystal size for Zr0 2 was calculated to be 24.5 nm, and that for αAI2O3 was 86.4 nm using the Warren-Averbach method after correcting for instrument broadening.

96



Figure 1. TEM micrograph of the (a) HA powder and (b) TZP-3Y20A/HA-10 vol.% powder. ΤΖΡ-3Υ20Λ/ ΗΛ powder t: tetragonal ZrO; m: monoclinlc ZrO ; α: α-ΑΙ : 0,; Η: hydroxyapatite

2-Theta [degree]

Figure 2. XRD of TZP-3Y20A/HA-10 vol.% powder and TZP-3Y20A powders. Microstructure and densification behavior

Figure 3. Fractured surface images of (a) TZP-3Y20A composites sintered at 1400 °C and (b) pure HA specimen sintered at 1000°C.


· 97


Figure 3 (a) shows the fracture surface of TZP-20A nano-composite without HA sintered at 1400 °C. Compact microstructure and small grain size (about 296 nm) are well illustrated in this micrograph. The fracture proceeded mostly in an intergranular pattern and the fracture surface was very rough. Figure 3 (b) shows the fracture morphology of pure HA specimen sintered at 1000 °C, which was much different from that of TZP-3Y20A, with compact structure and the fracture mostly occurring in a transgranular pattern.

Figure 4. SEM micrographs of fractured surface of TZP-3Y20A/HA composites containing various volume fraction of HA after sintering at 1400 °C; (a) HA-10 vol.%; (b) HA-20 vol.%; (c) HA-30 vol.%; (d) HA-40 vol.%; (e) HA-50 vol.%. Figure 4 (a)-(e) show the microstructures of the TZP-3Y20A nano-composites containing different amounts of HA sintered at 1400°C. When 10 vol.% HA was added, the microstructure changed and exhibited a different morphology in comparison with samples without HA, as shown in Figure 5 (a). The large dark grains (about 300 nm) of HA appear in the microstructure, incompact morphology can be observed, and the grain size (125 nm) is smaller

98



than those without HA in Figure 3 (a). When 20 and 30 vol.% HA were added (Figure 5 (c)-(d)), nearly no changes occurred for the grain size of TZP-3Y20A. HA grains were contact with each other in the local area. As HA content was increased up to 40 and 50 vol.%, the amount of contact between HA grains increased and agglomeration occurred through the continual pores fabricated by ZrO: and AI2O3 grains. The fine grains of Zr02 and AI2O3 embedded in HA grains can be observed, as shown in Figure 4 (e). It was found that a continuously porous skeleton was fabricated by fine grains of ZrC>2 and AI2O3, while HA acts as material filling in the pores, this structure is beneficial to homogenous distribution of bioactive HA. This framework structure is considered as an important factors in improving the biocompatibility of ceramics for implants because it is closely related to cell attachment, growth behavior, and bone strength between the tissue and artificial implant in the human body[17]. The grain size and porosity keep steady, by contrast to distinct densification behavior of TZP-3Y20A composites as shown in Figure 3 (a). This means that the addition of HA has a barrier effect on diffusion between grains of ZrC>2 and AI2O3 and thus limit the grain growth of ZrC>2 and AI2O3.

Figure 5. Density of TZP-3 Y20A/HA composites containing various volume fraction of HA as a function of sintering temperature. The densities of TZP-3Y20A/HA composites containing various volume fraction of HA as a function of sintering temperature are shown in Figure 5. For TZP-3Y20A composites, the density increased steadily with temperature and a maximum of 97.8% of the theoretical value (5.5 g/cm3) is obtained after sintering at 1400°C. When different volume amounts of HA were added, the density of TZP-3 Y20A/HA composites as a function of sintering temperature showed similar variation in trend compared to the TZP-3Y20A composites. The densities increased with increasing temperatures, but showed a decreasing trend when the sintering temperature was higher than 1300°C. It was also found that the densities decreased with increasing the volume amounts of HA. This is due to the density of HA (3.16 g/mm3) is lower than the density of TZP3Y20A (5.5 g/mm); the density of the sample with high HA content is lower than those with low HA content. The theoretical densities (T.D.) for different TZP-3Y20A/HA composites shown in Figure 6 are calculated by the rule of mixtures. Phase stability


· 99


Figure 6 shows the X-ray diffraction patterns of the TZP-3Y20A/HA-50 vol.% composites sintered at the temperature range of 1000-1400°C. There are reports in the literature on the promotive effect of ZrC>2 on the thermal decomposition of hydroxyapatite, even at lower temperatures . In the presence of zirconia even more OH" ions are lost due to the reaction between oxyhydroxyapatite and ZrC>2 according to Eq. (2) Ca»{PO 3[Ca>(PO^] + CaO(ZrO) + (1 -x)H.O (2) The decomposition of oxyhydroxyapatite in the presence of ZrC>2 is reported to take place at a surprisingly low temperature (~950°C), taking into account that the oxyhydroxyapatite formed according to Eq. (1) is stable up to around 1400°C in air. That is to say that a decomposition temperature of 950°C is well below that required for densification of Zr02/HA composites by conventional processes such as pressureless sintering and hot pressing. By using the SPS technique, however, the deleterious reactions describe above could be avoided . The XRD patterns illustrate that sintering the composites at 1400°C led to the appearance of TCP in the samples. The results explain the reason why the densities decrease when the sintering temperature is increased to 1400°C, as shown in the Figure 5. This is because the density of TCP is 3.00 g/cm3 which is lower than that of HA . All these results could be attributed to the rapid sintering speed of SPS. The high heating rate and very short dwelling time prevented the reaction between ZrC>2 and HA. TZP-3Y20A/HA-50 vol.% composites

A

A A A 1

i i ' i 40

1

i

ιιΛ i

i ■ i ' 45

^w50

55

2-Theta [degree]

Figure 6. X-ray diffraction patterns of the TZP-3Y20A/HA-50 vol.% composites sintered at different temperatures; (a) TZP-3Y20A/HA powder, (b) 1000°C, (c) 1100°C, (d) 1200°C, (e) 1300°C, (0 1400°C, T: tricalcium phosphate (TCP; Ca3(P04)2). Mechanical properties Figure 7 shows the variations of the flexural strength of the TZP-3 Y20A/HA composites containing different volume fractions of HA with sintering temperature. The flexural strength increases with increasing sintering temperature and is strongly affected by the amount of HA content. In samples without HA, the flexural strength of 186.95 MPa was obtained at a temperature of 1000°C. the flexural strength increase continuously with increasing temperature. The maximum flexural strength of 967.1 MPa was attained at 1400°C, corresponding to a relative density of 97.8%, which is close to full density. The flexural strength is higher compared to those

100



reported for similar compositions processed differently; 450 MPa by isostatically pressed and sintered and a little lower than 1.1 GPa by SPS[ When 10 vol.% of HA was added to the TZP-3Y20A powder and sintered under the same conditions, the flexural strength decreased compared to TZP-3Y20A without HA additions. The curve also shows that with increasing temperature in the range of 1000-1300°C, the flexural strength increased from 144.4 to 374.8 MPa, then decrease to 316.4 MPa when the sintering temperature increased to 1400°C. As the HA content was increased up to 20-50 vol. % , the flexural strength showed a similar trend as the samples of 10 vol. % HA content, and the flexural strength decreased steadily with increasing HA content at the same temperature. In other words, the flexural strength increased steadily with decreasing HA content. For example, the flexural strength increased from 117.5 to 640.0 MPa with the HA content decreasing from 50 to 0 vol. % at 1300°C. Another distinct feature in Figure 7 is the flexural strength decreased with the temperature increase to 1400°C for different HA content samples. This phenomenon could be due to decomposition of HA at 1400 °C as shown in Figure 6.

Figure 7. Influence of sintering temperature on flexural strength of TZP-3Y20A/HA composites containing different volume fraction of HA. The maximum flexural strength for TZP-3Y20A/50 vol.% HA is 117.5 MPa sintered at 1300°C, which is lower than the maximum flexural strength for pure HA of 131.5 MPa sintered at 950°C studied previously . When the HA content decreased to 40 vol. %, the maximum flexural strength increased to 211.6 MPa, which is higher than the strength of pure HA. Based on the results obtained, the flexural strength is increased with increasing TZP-3Y20A content, on the other hand, bioactivity is achieved by addition of HA at the expense of the flexural strength, if the TZP-3Y20A composite is considered as the substrate.


101


Figure 8. Vickers indentation observed in samples (a) TZP-3Y20A (b) TZP-3Y20A/50% HA for fracture toughness evaluation. Hardness and fracture toughness were measured in order to determine the influence of the sintering temperature and the content of HA phase on the mechanical characteristics of composite materials. Hardness and particularly fracture toughness belong to the most important parameters used in the characterization of ceramic materials, for which toughness is of primary important. The typical Vickers indentation with cracking at the tips for fracture toughness measurement is shown in Figure 8. To measure the fracture toughness accurately, the load of the hardness testing was increased from 9.8 N to 98 N in order to observe distinct crack generated at the tip of the indentation. Plastic deformation was found at the indentation edges of Figure 8 (a), which indicates that the as-sintered TZP-3Y20A composites have better fracture toughness. The indentation of Figure 8 (b) shows a different morphology for TZP-3Y20A/50% HA composite. Besides the cracks at the tips which can be clearly observed, cracks looking like annual rings also can be observed, at the inner area of the indentation. The result shows that the addition of HA deteriorated the fracture toughness of TZP-3Y20A composite.

Figure 9. Influence of sintering temperature on fracture toughness of TZP-3Y20A/HA composites containing different volume fraction of HA.

102



The effect of sintering temperature on fracture toughness of TZP-3Y20A/HA composites was shown in Figure 9. The fracture toughness increased from 2.12 MPa.m1/2 to 5.27 MPa.m172 with increasing temperature from 1000-1400°C for TZP-3Y20A composite without HA addition. The fracture toughness improvement is partially attributed to the increase in density of TZP3Y20A composite as shown in Figure 5. It turns out that the higher the density, the lower the probability of microcracks forming during cooling, due to lack of stress concentrators. This in turn reduces the contribution of crack deflection toughening mechanism. When 10 vol.% of HA was added to the TZP-3Y20A powder and sintered under the same conditions, the fracture toughness showed an obvious decrease compared with TZP-3Y20A composite at the same sintering temperature. It was observed that the fracture toughness increased with increasing temperature, from 1.62 to 3.74 MPa.m12 in the 1000-1300°C temperature range, and then slightly decreased to 2.58 MPa.m172 when the temperature increased to 1400°C. As HA content was increased from 20 to 40 vol. %, the fracture toughness showed similar trend as the samples of TZP-3Y20A/10% HA, and the fracture toughness decreased steadily with increasing HA content at the same temperature. In other words, the fracture toughness increased steadily with decreasing of HA content. Interestingly, as HA content was increased to 50%, the fracture toughness increased instead of decreasing. It was found that the fracture toughness increased with increasing temperature, from 2.24 to 2.83 MPa.m172 in the 1000-1200°C temperature range, reaching the maximum fracture toughness of 2.83 MPa.m172, and then decreased with increasing temperature. The improvement in fracture toughness in TZP-3Y20A/50% HA samples could be attributed to the decrease in porosity. When the HA content increased from 40 to 50%, more HA filled in the pores of the continuously porous skeleton fabricated by fine grains of ZrC>2 and AI2O3, as shown in Figure 5. For dense HA ceramics, the fracture toughness fell within the range of 0.79-1.40 MPa.m172 as reported in the literature [4]. For TZP-3Y20A composite studied in the present paper, the fracture toughness fell within the range of 2.12-5.27 MPa.m172. Comparing to dense HA, the improvement in fracture toughness of TZP-3Y20A/HA composites is beneficial. This is due to the excellent fracture toughness of TZP-3 Y20A.

Figure 10. Influence of sintering temperature on Vickers hardness of TZP-3 Y20A/HA composites containing different volume fraction of HA.


·

103


Results of Vickers microhardness studies are given in Figure 10. It can be seen that the Vickers hardness of TZP-3Y20A increased with increasing temperature, from 4.96 to 13.26 GPa in the temperature range of 1000-1300°C. The different amount of porosity of in samples can be an important factor leading to different Vickers microhardness value. As sintering temperature increasing, pores in the sintered compact are remarkably eliminated and the homogeneous microstructure can be acquired (see Figure 3 (a)). Hardness slightly decreased at 1400°. This is probably caused by the influence of grain size on the hardness of the sintered composite. In other words, coarser grain size lead to lower hardness [23]. When different volume fraction of HA from 10 to 50% were added to TZP-3Y20A, Vickers hardness decreased with increasing of HA content at the same temperature. Take samples sintered at 1300 °C as an example, hardness shows a decreasing trend from 10.57 to 6.17 GPa, when HA content was increased from 10 to 50%. Considering hardness of dense HA ceramics fell within the range of 5.7-6.6 GPa [24], for ZrO2(3Y)-20wt.% A1203 composites studied in the present work, the hardness shows a maximum value of 13.26 GPa sintered at 1300 °C. Vickers hardness of TZP-3Y20A/HA composite fell within the hardness value range of dense HA and of TZP-3Y20A composite, which means the merits of compatibility of HA and hard Zr02(3Y)-Al203 were intermediated by fabrication of the TZP-3Y20A/HA composite. CONCLUSIONS The present work shows that TZP-3Y20A/HA composites with the addition of different volume fraction of HA were fabricated successfully using spark plasma sintering. The densification behavior, microstructure and mechanical properties of composites are investigated as a function of sintering temperature and HA content. The sintering temperature has a significant effect on the final densities achieved in the TZP-3Y20A/HA compacts. The density of TZP-3Y20A composite increased steadily with temperature and a maximum value of 97.8% was obtained after sintering at 1400 °C. The addition of HA had a barrier effect on diffusion between grains of Zr0 2 and A1203 and thus limited the grain growth of Zr0 2 and AI2O3. Sintering the TZP-3Υ20Α/ΗΑ composites at 1400 °C led to the appearance of TCP in the samples. The TCP is from the decomposition of HA due to the limited thermal stability at high temperature, and not from the reaction between Zr0 2 and HA. Flexural strength, fracture toughness and Vickers hardness values increased with increasing sintering temperature and were strongly affected by the amount of HA, which compared well with densities obtained at different sintering temperatures. Flexural strength, fracture toughness and microhardness showed a decreasing trend with increasing of HA content at the same temperature. This indicates that bioactivity is achieved by addition of HA at the expense of the mechanical properties of TZP-3Y20A composite. The maximum flexural strength, fracture toughness and Vickers hardness of 967.1MPa, 5.27 MPa.m1,2 and 13.26 GPa were achieved respectively, at a relative density of 97.8%. Flexural strength, fracture toughness and Vickers hardness values of the TZP-3Y20A/HA composite fell within the value range of dense HA and of TZP-3 Y20A composite, indicates the advantages of bioactivity of HA and hard Zr02(3Y)-Al203 were obtained by fabrication of the ZP-3Y20A/HA composite.

104



ACKNOWLEDGEMENTS The authors sincerely thank Prof. Zhengxin Lu for performance of the TEM observations and helpful discussions. REFERENCES fll R. Murugan and S. Ramakrishna, Development of nanocomposites for bone grafting, Compos. Sci. Tech., Vol 65, 2005, p 2385-2406 [21 E. S. Thian, N. H. Loh. K. A. Khor, S. B. Tor, Effects of debinding parameters on powder injection molded Ti-6A1-4V/HA composition parts, Adv. Powd. Tech., Vol 12(3), 2001, p 361370 [3] W. Suchanek, M. Yashima, M. Kakihana and M. Yoshimaru, Hydroxyapatite ceramics with selected sintering additives, Biomaterials, Vol 18, 1997, p 923-933 f4] S.F. Li, H. Izui and M. Okano, Densification, microstructure and behavior of hydroxyapatite ceramics sintered by using spark plasma sintering, J. Mater. Eng. Tech, Vol 130(031012), 2008, DOI: 10.1115/1.2931153 [5] D.E. MacDonald, F. Betts, M. Stranick, S. Doty and A.L. Boskey, Physicochemical study of plasma-sprayed hydroxyapatite-coated implants in humans, J. Biomater. Res., Vol 54, 2001, p 480-490 [61 A. Rapacz-Kmita, A. Slosarczyk and Z. Paszkiewicz, Mechanical properties of Hap-ZrCh composites, J. Euro. Ceram. Soc, Vol 26, 2006, p 1481-1488 [71 X.G. Miao, Y.M. Chen, H.B. Guo and K.A. Khor, Spark plasma sintered hydroxyapatiteyttria stabilized zirconia composites, Ceram. Inter., Vol 30, 2004, p 1793-1796 [81 Y.M. Kong, CJ. Bae, S.H. Lee, H.W. Kim and H.E. Kim, Improvement in biocompatibility of Zr02-A1203 nano-composite by addition of HA, Biomaterials, Vol 26, 2005, p 509-517 [9] W. Li and L. Gao, Fabrication of HA-Zr02 (3Y) nano-composite by SPS, Biomaterials, Vol 24, 2003, p 937-940 [101 B.T. Lee, C.W. Lee, M.H. Youn, H.Y. Song, Relationship between microstructure and mechanical properties of fibrous Hap-(t-Zr02)/Al203-(m-Zr02) composites, Mater. Sci. Eng. A, Vol 58, 2007, p i 1-16 [11] Jerome Chevalier, What future for zirconia as a biomaterial. Biomaterials, Vol 27, 2006, p 535-543


·

105


[12] Z. Shen, E. Dolfasson, M. Nygren, L. Gao, H. Kawaoka and K. Niihara, Dense hydroxyapatite-zirconia ceramic composites with high strength for biological applications, Adv. Mater., Vol 13, 2001, p 214-216 [13] X.G. Miao, Y.M. Chen, H.B. Guo and K.A. Khor, Spark plasma sintered hydroxyapatiteyttria stabilized zirconia composites, Ceram. Inter., Vol 30, 2004, p 1793-1796 [14] Yoshimura M, Phase stability of zirconia. Am. Ceram. Soc. Bull., Vol 67, 1988, p 19501955 [15] Tosoh, Supplier's powder specifications, Japan, 1998. [16] C. Oprea, V. Ciupina and G. Prodan, Investigation of nanocrystals using TEM micrographs and electron diffraction technique, Rom. J. Phys., Vol 53, 2008, p 223-230 [17] B. T. Lee, 1. C. Kang, S. H. Cho and H. Y. Song, Fabrication of a continuously oriented porous Al 2 0 3 body and its in vitro study, J. Am. Ceram. Soc, Vol 88, 2005, p 2262-2266 [18] V.V. Silva, F.S. Lameiras and R.Z. Dominguez, Microstructural and mechanical study of zirconia-hydroxyapatite (ZH) composite ceramics for biomedical applications, Compo. Sci. Tech., Vol 61, 2001, p 301-310 [19] M. Nygren and ZJ Shen, On the preparation of bio-, nano- and structural ceramics and composites by spark plasma sintering. Solid State Sci., Vol 5, 2003, p 125-131 [20] P.E. Wang and T.K. Chaki, Sintering behavior and mechanical properties of hydroxyapatite and dicalcium phosphate. J. Mater. Sci.: Materials Medicine, Vol 4, 1993, p 150-8 [21] R. Chaim, Pressureless sintered ATZ and ZTA ceramic composites, J. Mater. Sci., Vol 27, 1992, p 5597-5602 [22] J.S. Hong, L. Gao, S.D.D.L. Torre and Hiroki Miyamoto, Key Miyamoto, Spark plasma sintering and mechanical properties of Zr0 2 (Υ 2 θ3)-Α1 2 0 3 composites, Materials Letters, Vol 43, 2000, p 27-31 [23] R.S. Mishra, C.E. Lesher. A.K. Mukherjee, High-Pressure Sintering of Nanocrystalline γΑ1 2 0 3 , J. Am. Ceram. Soc, Vol 79 (11), 1996, p 2989-2992 [24] A. Rapacz-Kmita, A. Slosarczyk and Z. Paszkiewicz, Mechanical properties of Hap-Zr0 2 composites, J. Euro. Ceram. Soc, Vol 26, 2006, pl481-1488

106



COMPARISON OF SLIP CAST TO HOT PRESSED BORON CARBIDE T. Sano and E. S.C. Chin U.S. Army Research Laboratory Aberdeen Proving Ground, Maryland, USA B. Paliwal Department of Mechanical Engineering, Johns Hopkins University Baltimore, Maryland, USA M. W. Chen Institute for Materials Research, Tohoku University, Sendai Japan Department of Mechanical Engineering, Johns Hopkins University Baltimore, Maryland USA ABSTRACT To meet the possible increase in future demand for armor materials, an increase in the throughput during manufacturing is necessary. One possibility is the use of the slip casting and sintering technique to form ceramic armor compacts as an alternative to current hot pressing techniques. Dynamic uniaxial compression tests with the Kolsky bar were conducted on two types of slip cast boron carbide, and compared with results from the standard hot pressed boron carbide. One type was slip cast, sintered, and hot isostatically pressed, while the other was only slip cast and sintered. Microstructural characterization by transmission electron microscopy showed graphite inclusions and more annealing twins than in the hot pressed boron carbide material. Examination of fragments recovered from the compression tests determined that the fracture mode of both slip cast materials was brittle transgranular cleavage. The compression test results show comparable compressive strengths between the sip cast and hot pressed boron carbide despite higher density of graphite in the slip cast material. INTRODUCTION Hot pressing boron carbide (B4C) powder is the commercial technique used to form personnel armor plates and components for various applications. B4C is used widely in abrasive, wear resistant components, and armor applications due to its high hardness and low density properties. It is possible for B4C components formed by the hot pressing technique to reach nearly full theoretical density and achieve high mechanical performance. Compared to sintering processes, hot pressing requires less additives for better densification and strength. These additives could however also form precipitates or secondary phases at the grain boundaries and be detrimental to the mechanical performance. The limitation of the hot pressing technique is the high operation cost per batch and only plates or cylindrical shapes of a limited size can be produced. Also, in addition to a larger die, to achieve the same pressure applied to a smaller specimen, a much larger hot press machine size is required for larger specimens. Recently, an alternative technique of forming B4C compacts was described by Matsumoto et al.1. In this technique, B4C powder was slip cast, sintered, and hot isostatically pressed (HIPed). The mechanical properties of these slip cast and HIPed B4C materials were reported2 to be as good as the hot pressed B4C materials. The aim here is to evaluate the dynamic

107


mechanical performance of slip cast-sintered-HIPed, and slip cast-sintered B4C and compare them with hot pressed B4C. In addition to the mechanical testing, microstructural characterization of the HIPed and SCS samples was conducted to better characterize the material pre and post compression tests. EXPERIMENTAL Two types of B4C samples were obtained. One was slip cast and sintered, and the other was slip cast, sintered, and HIPed. The slip cast and sintered samples will be referred to as the "SCS" B4C samples, and when identifying the samples that were also HIPed, described as "HIPed" to distinguishing between the two types. The third sample compared in this study is hot pressed B4C. This B4C sample is the armor grade reference benchmark material and will be referred to as the hot pressed sample. This hot pressed B4C was analyzed in a previous work3 and the data is taken from the paper on the prior analysis. The surfaces of the as received slip cast B4C (SCS and HIPed) samples were imaged with a scanning electron microscope (SEM) equipped with a field emission gun. The samples were polished on the Struers automatic polisher with diamond slurries of decreasing diamond abrasive size at each polishing step, until reaching 0.25 μιη. The samples were then polished on a vibramet polisher with 0.02 μπι colloidal silica. The post-polished slip cast samples were imaged on the SEM. The Knoop hardness of the two samples was measured with a microhardness tester, at loads from 3 N to 98 N (0.3 Kg to 10 Kg). To determine the existence of any impurities and different B-C phases, x-ray diffraction was conducted. The impurity phases were verified, and grain boundaries examined with transmission electron microscopy. The slip cast samples were also machined into the 4.0 mm x 5.2 mm x 3.0 mm geometry for dynamic mechanical testing with the Kolsky bar. Each sample was loaded onto the Kolsky bar ends with lithium grease, which also acts to minimize the friction between the sample and the titanium alloy platens. The full Kolsky bar setup is described in the work by Ramesh and Narasimhan4. Each sample was subjected to dynamic compression at the rate between 150 and 160 MPa^sec. The stress and strain rate throughout the compression test cycle was captured by a high speed camera at 2 or 3 μ8 intervals with 300 to 700 ns exposure time. The sample area was enclosed in a clean polycarbonate box with a clean sheet of paper lining the bottom of the box to collect the fragments after the compression experiment. After each experiment, the polycarbonate box was cleaned and a new sheet of paper was installed to minimize sample contamination and to collect the fragments from the next experiment. The collected fragments were labeled with the B4C processing type and experiment number then examined with the SEM and energy dispersive spectrometry, or EDS. The fracture surfaces of the HIPed B4C samples were compared with the SCS B4C as well as with the hot pressed B4C samples. RESULTS The surfaces of the as machined SCS and HIPed B4C samples were examined in the SEM and EDS. Samples were also prepared and examined with TEM. The various microscopy techniques revealed dark and pore-like areas to be graphite inclusions. Compared to the hot pressed B4C, the slip cast B4C appeared to have more, though smaller sized, graphite inclusions. The as received slip cast B4C samples were x-rayed to determine the phases and identify any impurities. The diffraction peaks for both samples were consistent with each other and can be inferred that both samples have the same phase and impurities. When the diffraction peaks in

108



both samples were identified, B4C, BnC2, and carbon (graphite) were determined to be the phases present. Figure 1 shows the peak identification of the HIPed sample. The same phases were identified in the SCS sample. The volume percents of the graphite flakes for both slip cast B4C were measured from SEM images and calculated to be 10%. Even with these graphite inclusions, the density of the HIPed B4C, calculated by the Archimedes method, was 2.50 g/cm3, or 99.2 % of the theoretical density, and 2.45 g/cm3, or 97.2 % of the theoretical density for the SCS B4C. Assuming from the lack of porosity, determined by visual examination of SEM micrographs, that the HIPed sample is fully dense (i.e. no porosity), the remaining 0.8 % or 0.02 g of the density was comprised of graphite. With this assumption, the vol % of graphite was calculated to be 9 vol %. Assuming the same amount of graphite was in the SCS sample, the remaining 2 % of the theoretical density is therefore due to pores. The total volume of the machined sample was measured to be 0.06 cm3. From the difference in the density between the HIPed and SCS samples, the approximate volume percent of pores in the SCS sample was calculated to be 2.0 vol%. Neither volume percent of pores nor graphite was calculated for the hot pressed B4C, but the density was determined previously by Chen et al/ to be 2.49 g/cm3, or 98.8 % of the theoretical density.

Fig. 1. The HIP diffraction pattern with B4C, graphite, and Bi3C2 peaks labeled.


· 109


Fig. 2. Amorphous interface boundary between graphite and a B4C grain. Compared to the hot pressed B4C, which had possible precipitates and distinct chemical compound at the triple junctions3, the slip cast B4C had fewer impurities. SEM and TEM analysis of both slip cast samples show only graphite inclusions, some trapped within the grain and others at triple junctions. The interface between the graphite inclusion and B4C grains were determined by TEM to be amorphous as shown in Fig. 2. Another observation by SEM and TEM was the numerous twins; more than in the hot pressed B4C samples, as shown in Fig. 3. This is in agreement with previous work by Schwetz et al5. Examination of the 100 nm to 1 μιη grain size B4C powder used in the slip cast samples did not show any twins. However the polished surfaces of slip cast samples that did not undergo compression tests, show many grains with twins. Hence the twins observed in the slip cast samples are not deformation twins and are growth twins formed during the processing steps. After polishing, hardness measurements were conducted on the slip cast B4C samples. Vickers indentation was initially attempted. However due to the high hardness and low toughness properties of B4C, the indentation marks were not measurable. Hence Knoop indentations were performed with varying load. Table I shows the comparison of the Knoop hardness measurements at a load of 19.6 N (HK(2)), in accordance to ASTM C1326. Each sample that underwent dynamic uniaxial compression test with the Kolsky bar was imaged with a high speed camera and the stress and strain values recorded. Figure 4 (a) shows camera frame shots of the SCS sample #1 during the compression test and (b) is the plot of the stress and strain rate over time. In Fig. 4 (a), the sample exhibited cracking from the edges inward. The other SCS samples, as well as the HIPed samples, regardless of the compressive strength, showed similar trends in the stress and strain profiles. All samples displayed throughsample cracking at the time interval just past the maximum stress peak, and destruction of the sample shortly thereafter. The comparison among the SCS, HIPed, and the hot pressed B4C compressive strengths are shown in Fig. 5.

110



Fig. 3. Polished surfaces of B4C that was a) SCS, b) HIPed, and c) Hot pressed. Numerous growth twins are visible in a) and b).


· 111


Table I. Average Knoop Hardness Values at 19.6 N Load Std. Dev. (GPa) Material Ave. (GPa)

(a)

HIP

21.1

0.6

SCS Hot Pressed7

18.6 19.6

0.9 1.8

Inter-frame time 3μ5, exposure time 500ns

Fig. 4, a) High speed camera images of SCS sample #1 during the Kolsky bar compression test. The image numbers correspond to the stress and time plotted in b).

112



The fragments from the Kolsky bar experiments were carefully collected and examined in the SEM. All samples displayed fracture by brittle, transgranular fracture mode, evidenced by the cleavage fracture surfaces. Figure 6 (a) is an example of a fractured surface of a fragment from a SCS sample (Sample #1). Twins were observed on all the fractured surfaces. Graphite inclusions were also frequently observed on fractured surfaces, whether due to them actively influencing the fracture path or due to the high density of the inclusions in the material. Similar to Fig. 6 (a), Fig. 6 (b) is a SEM image of a fragment from a HIPed sample. Compared to the sintered B4C samples, the hot pressed samples revealed the same cleavage fracture, however with very few fractured surfaces with twins.

Fig. 5. Compressive strengths of SCS, HIPed, and hot pressed B4C samples.

Fig. 6. (a) An SCS sample fragment surface, (b) A HIPed sample fragment surface.


· 113


DISCUSSION All B4C samples had graphite inclusions, which for the slip cast samples, was found at the grain boundaries, triple junctions, and trapped within the grains. The graphite was already in the initial powder, and was not formed during the processing. The Vickers hardness values provided by the manufacturer for the HIPed B4C was the highest at 39 GPa (load unknown) followed by the SCS B4C with 32 GPa. The Vickers hardness HV(0.3), provided by the manufacturer for the hot pressed B4C was 26.5 GPa. When Vickers hardness tests were conducted by the authors, severe spalling occurred and the data was deemed unreliable. However Knoop hardness values were successfully obtained. The HK(2) values for the HIPed B4C was 21.1 GPa with a standard deviation of 0.6 GPa and the SCS B4C was 18.6 GPa with a standard deviation of 0.9 GPa. These values as well as the tested range of hardness values from HK(0.3) to HK(10) were in agreement with the Knoop hardness results of hot pressed B4C by Swab7. Except for at HK(10), the HIPed B4C consistently had the highest hardness values, followed by hot pressed, then SCS. This hardness trend follows the density, in which the densest B4C was the HIPed sample, followed by hot pressed, then the SCS sample. For the Kolsky bar compression test, having rough surfaces is known to lower the compressive strength, even with lubricant applied along the contact sides of the test specimen8. If too much lubricant is used, the lubricant can be detrimental by filling in the pores and other depressions on the rough surfaces and force crevices to open. Since the as machined slip cast B4C had many scratches on the surface, it can be expected that with a smoother surface finish, the compressive strengths would be higher. The way to minimize the surface finish problem would be to machine cylindrical dog-bone shaped specimens. However the machining cost for such a geometry would be expensive. Nevertheless assuming that the hot pressed B4C has similar machining difficulties affecting the surface finish as the slip cast material, the compressive stress results should still be comparable. Even with limited number of data and despite all the processing and microstrucutral differences, all three B4C samples performed within a standard deviation from each other in the dynamic uniaxial compression test. Provided that the surface finish was similar, this indicates the importance of the material properties on the compressive strength rather than the processing method to achieve that strength. This also shows that impurities and growth twins play limited roles on the compression performance. It can be deduced that as long as the initial B4C powder is at least 96 wt% pure, the grain size ranges from 5 to 15 μπι, and the sample density is above 98.8 % of the theoretical density, the compressive strength can be expected to fall within the range of 3000 to 4100 MPa. It is shown here that slip casting or slip casting with the HIP step can be used to form unconventional shapes of B4C samples with similar mechanical performance as B4C samples that are hot pressed. Also, though not realized in this study, the addition of the HIP process increases the density which, based on general trends, should increase the mechanical performance. Hence the HIP process could not only complement B4C production, but could possibly produce better B4C compacts. CONCLUSION Microstructural characterization and dynamic uniaxial compression tests with the Kolsky bar were conducted on B4C samples that were SCS and slip cast, sintered, and HIPed. Growth twins were more frequently observed on the surfaces of SCS and HIPed samples than those of hot pressed, benchmark B4C samples. The compressive strengths results were compared with those of previously tested commercially hot pressed B4C samples. The compressive strength of

114



the HIPed and non HIPed samples were in the range of compressive strength of the hot pressed B4C samples. Hence slip casting and sintering B4C is a feasible technique to obtain unconventionally shaped samples with comparable compressive strength as hot pressed B4C samples. ACKNOWLEDGEMENTS The authors would like to thank the discussions and characterization assistance provided by the following individuals: Prof. K. T. Ramesh of Johns Hopkins University, Dr. James McCauley, Dr. Ryan McCuiston, and Mr. Herbert Miller of the U. S. Army Research Laboratory. REFERENCES 1. A. Matsumoto, A. Kawakami, and T. Goto, "Slip Casting and Pressureless Sintering of Boron Carbide," Ceram. Trans., 133 M. Matsui, S. Jahanmir, H. Mostaghaci, M. Naito, K. Uematsu, R. Wasche, R. Morrell editors; pp. 223 The American Ceramics Society, Westerville, OH (2002). 2. A. Matsumoto, T. Goto, and A. Kawakami, "Slip Casting and Pressureless Sinter of Boron Carbide and Its Application to the Nuclear Field," J. Ceram. Soc. Japan, Suppl. 112-1, 112[5] S399-S402 (2004). 3. M. W. Chen, J.W. McCauley, J. C. LaSalvia, and K. J. Hemker, "Microstructural Characterization of Commercial Hot-Pressed Boron Carbide Ceramics," J. Am. Ceram. Soc.,88[7] 1935-42(2005). 4. K. T. Ramesh and S. Narasimhan, "Finite Deformation and the Dynamic Measurement of Radial Strains in Compression Kolsky Bar Experiments," Int. J. Solids structures, 33[25] 3723-38(1996). 5. K. A. Schwetz, W. Grellner, and A. Lipp, "Mechanical Properties of HIP Treated Sintered Boron Carbide," Inst. Phys. Conf. Ser. [75] 413-424 (1986). 6. ASTM C1326-96 "Standard Test Method for Knoop Indentation Hardness of Advanced Ceramics" 2003 Annual Book of ASTM Standards, Vol. 15.01. 7. J. J. Swab, "Recommendations for Determining the Hardness of Armor Ceramics," Int. J. Appl. Ceram. TechnoL, 1 [3] 219-25 (2004). 8. J. C. LaSalvia, verbal communication, 2006.


· 115


Amorphous Ceramics


MECHANICALLY DRIVEN AMORPHIZATION AND BULK NANOCRYSTALLINE SYNTHESIS OF ULTRA-HIGH TEMPERATURE CERAMICS H. Kimura Department of Mechanical Engineering, School of Systems Engineering National Defense Academy Yokosuka, Kanagawa, Japan ABSTRACT This article reports the process control methodology for the mechanically driven amorphization and the bulk nanocrystalline synthesis of ultra-high temperature ceramics (UHTCs) such as B4C, SiC, ZrB2 and B4C· SiC in non-equilibrium solid state powder processing. The amorphization by mechanical grinding UHTCs particle is formulated by a linear chemical reaction with the repeated impaction number N, d(Vr-V)/dN=-k(V[-V) where Vf and V is the amorphous volume at the final and intermediate stage respectively. Planetary ball milling is used as a model experiment by which to characterize the solid state amorphization of UHTCs; the k is found to relate to a manipulatable milling parameter; it increases with decreasing powder weight and temperature and in order of B4C-iV,) ot

ox

(2)

where p is the molar density and εν is the porosity of the sample. Darcy's Law from Eq. (1) can then be combined with Eq. (2) to obtain the following expression for describing 1-D flow in a porous medium as

£(^)._L£&i ct

dx

(3)

Since density is the dependent variable in the model, pressure values are calculated using the Peng-Robinson cubic equation of state with values of the constants adopted from elsewhere23'24. For simplifying the calculations related to the dependence of viscosity on molar density, we express the viscosity of CO2 as a quadratic equation of the form μ(ρ) = α*ρ2 +b*p + c*

(4)

where Λ*=1.7Χ10"' 3 Pa s m6/mol2,fc*=8.2xlO-11 Pa s nrVmol, and c = 1.8xlO"5Pa s. These values were obtained using regression analysis on data for the variation in viscosity of CO2 with molar density at 90°C25"26. Although Eq. (3) is the standard form of the continuity equation, we extend it in this work by allowing the viscosity to vary and by relating the density to pressure via a cubic equation of state.

230

·


Modeling of the Pressure in 1 -D Green Ceramic Bodies during Depressurization

Figure 1. One dimensional schematic of the green body showing the coordinate system and the body length. RESULTS Figure 2 shows an image of an MLC sample fabricated from an acrylic-based polymer with a polyester adipate plasticizer. The sample in Fig. 2 was first exposed to supercritical carbon dioxide for 1 h at 90°C and 10 MPa. Under these conditions, the sample experienced a weight loss of 0.8 weight%. As seen in Fig. 2, after depressurization over 8 h to ambient conditions, both horizontal and vertical cracks are present. The horizontal defects tend to be between the layers, and the largest crack is at the center of the green body.

Figure 2. Image of an MLC sample after depressurization over 8 h from conditions of supercritical extraction in carbon dioxide at 10 MPa at 90°C for 1 h. The sample lost 0.8 weight% of the initial binder content.


· 231

Modeling of the Pressure in 1-D Green Ceramic Bodies during Depressurization

To account for the defect formation evident in the image in Fig. 2, we next evaluate model results for different types of depressurization from conditions of supercritical extraction. At this time, it is not possible to unambiguously specify the value of the pressure or the pressure gradient that leads to failure; nonetheless, it is of interest to have at least a qualitative sense of the how the pressure varies within the green body during depressurization. Figure 3 shows the time evolution of the pressure of CO2 versus position in the sample for step changes in pressure of 40 to 20 MPa (left-hand side, LHS, of Fig. 3) and 40 to 35 MPa (right-hand side, RHS, of Fig. 3) at 90°C (see Table I for the parameters used in the simulations). For the larger step change, steep pressure profiles exist initially within the sample, and these diminish with increasing time until after 5 s the pressure becomes more uniform within the pore space. Although the pressure is largest at the center of the sample, the pressure gradients within the sample are largest near the body edges. For the smaller step change, the pressure profiles are less steep across the sample and the pressure gradients diminish more quickly with time. ou -

40CO Q_

2

40 - 35 MPa

40 - 20 MPa 0.1 s^^__|

0.1 s

Ts"^-^^ 5s

30-

a.

/ 20-

1s

5s

10- — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — I — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — -0.01 0.01 -0.005 0 0.005 x(m)

Figure 3. Pressure of CO? at 90°C versus position and time for step changes in the boundary value pressure. The left-hand side corresponds to a step change from 40 to 20 MPa, and the right-hand side corresponds to a step change from 40 to 35 MPa. Table I. List of parameters used in Figures 3-8. Values Parameter (Units) Figs. 3 - 5 Figs. 6 - 8 κ (m2)

lxlO17

lxlO17

Hm) M-)

0.02

0.02

0.15

0.15

Po(MPa)

40

40

20,35

-

-

6s,6h

step

P

(M?&)

DF

t (sorh)

232



In earlier work in the modeling of thermal debinding27'28, we have shown that the stress within the sample is proportional to the gradient in pressure within the sample28, and that both the maximum stress and maximum pressure occur in the center of the green body. We thus define a quantity Pnx_, which is proportional to the gradient in pressure, as the ratio of the pressure at any position in the body, Px, to the pressure at the body edges, PX=±L/2, as (5) Figure 4 thus compares the values of P„iX for the large and small step changes in pressure in Fig. 3 for different times. For the large step change of 40 to 20 MPa, Pnx is initially large in the center of body at a value of ~2 and decreases across the body thickness. With increasing time, Pn,x versus position in the body is reduced and approaches a value of unity everywhere, thus indicating a uniform pressure in the green body. For a smaller step change of 40 to 35 MPa, Pnx is much smaller versus position and time as compared to the corresponding values for the larger step change.

1.00

0.96 0.01

Figure 4. Behavior of PtltX of CO2 at 90°C versus position and time for step changes in the boundary value pressure. The left-hand side corresponds to a step change from 40 to 20 MPa, and the right-hand side corresponds to a step change from 40 to 35 MPa. Figure 5 summarizes the time evolution of the pressure at the body center, Px=o, and the normalized pressure at the body center, Pn,x=o, of C0 2 for step changes from 40 to 20 MPa and 40 to 35 MPa. Both quantities decrease monotonically with increasing time. An alternative to the step mode of depressurization is to decrease the fluid density (or pressure) in the vessel linearly with time, and this can be conducted either rapidly or slowly. Figure 6 quantifies the effect of a depressurization time, tDP=6 s (LHS of Fig. 6) versus ^=6 h (RHS of Fig. 6) for values of the simulation parameters given in Table I. For a rapid rate of linear depressurization, the pressure across the sample varies spatially more strongly with time as compared to a slower linear rate of depressurization, where the pressure profiles are more uniform for all times.


·

233


50 2.5

40 co

30 X

1.5

20

c Q.

10 H

0.5

0.01

0.1

10

t(s)

Figure 5. Behavior of Px=() (left-hand axis) and Pn,.x=o (right-hand axis) of C0 2 at 90°C versus time for step changes in the boundary value pressure of 40 to 20 MPa and 40 to 35 MPa. tDP = 6s

J

OS

J

ΤΓ

1-^-

1 -1

1

,

tDP = 6h

J

Oh

1h

J^—π

1

!

_!

2h

!

4h

i ,

(

ι

1

0.005

,

0.01

x (m)

Figure 6. Pressure of C0 2 at 90°C versus position and time for linear changes in the boundary value pressure. The left-hand side corresponds to a depressurization time of 6 s, and the righthand side corresponds to a depressurization time of 6 h. Figure 7 shows how Pnx varies spatially and temporally for the rapid and slow linear rates of depressurization shown in Fig. 6. Now, in contrast to the behavior seen in Fig. 4, Pnx is initially near unity throughout the sample, and then increases with increasing time during depressurization; once again, however, the maximum value of Pnx is always at the center of the body. Although the trends for the time evolution of P,ltX are different as compared to Fig. 4, the magnitude of P„tX is smaller for linear depressurization as compared to a step change in pressure.

234



For the slower rate of linear depressurization (RHS of Fig. 7), the values P„iX are much smaller than for the faster rate of depressurization (LHS of Fig. 7). 1.5 I

J

tDP-6s

!

tDP - 6 h

As

/^^\

I

f

1.0016

1.0012

1.2 1.0004

f -0.01

os -0.005

;

oh

0

0.005

I 0.01

x (m)

Figure 7. Behavior of PtliX of C0 2 at 90°C versus position and time for linear changes in the boundary value pressure. The left-hand side corresponds to a depressurization time of 6 s, and the right-hand side corresponds to a to a depressurization time of 6 h. Figure 8 summarizes the evolution of Px=0 and Pn,x=o of CO2 versus time for rapid and slow linear rates of depressurization. For the case with depressurization lasting 6 s, although there is a gradual reduction in the pressure at the center of the sample with time, the value of Pn,x=o increases. Although the same qualitative effect is observed in the same two quantities, with a slower linear rate of depressurization, the overall magnitude is much lower. CONCLUSIONS Defects have been shown to arise in green ceramic bodies following supercritical extraction of binder and such defects are postulated to occur during depressurization from conditions of supercritical extraction. The origin of the defects is ascribed to pressure gradients that arise during depressurization. The temporal and spatial evolution of these pressure gradients are modeled with Darcy's Law for flow in porous media for both step changes and linear changes in the boundary value pressure. In general, the pressure gradients are larger and take longer to dissipate for larger step changes in pressure or for more rapid linear changes in the boundary value pressure. ACKNOWLEDGEMENT This work was supported in part by the NSF/ICURC Center for Dielectric Studies.


· 235


0

1

| ,__ |

en J

ou

40 1o 30 CL

t(h) 3

2 j

|

4 ,_

5

|

,

!

f t = 6s

t DP ==

0-

| V 1.8

6rNv

,r

'

1.6 M

o

S

6

,

20 -

CL

Ρ=68 10 -

t

1

1.2

_L

\ΌΡ^Γ

0 -

0

·4 3

2

- ^

7

'

1 0.8

l

3

4

5

6

t(s)

Figure 8. Behavior οΐΡχ=ο (left-hand axis) and Pn.x=o (right-hand axis) of C 0 2 at 90°C versus time for rapid and slow linear changes in the boundary value pressure. REFERENCES [ K. V. Shende, D. S. Krueger and S. J. Lombardo, Supercritical Extraction of Binder Containing Poly(vinyl butyral) and Dioctyl Phthalate from Barium Titanate-Platinum Multilayer Ceramic Capacitors, J. Mater. Sci.: Mater. Electron., 12, 637-643 (2001). 2 R. V. Shende, D. S. Krueger and S. J. Lombardo, Binder Removal by Supercritical Extraction from BaTiU3-Pt Multilayer Ceramic Capacitors, Ceramic Transactions, Vol. 129, Innovative Processing and Synthesis: Ceramics, Glasses and Composites V, eds. J.P. Singh, N.P. Bansal, A. Bandyopadhyay, and L. Klein (American Ceramic Society, Westerville, OH) (2002) 165174. 3 R. V. Shende and S. J. Lombardo, Supercritical Extraction with Carbon Dioxide and Ethylene of Poly(vinyl butyral) and Dioctyl Phthalate from Multilayer Ceramic Capacitors, J. Supercrit. Fluids, 23, 153-162(2002). 4 R. V. Shende, M. Kline and S. J. Lombardo, Effects of Supercritical Extraction on the Plasticization of Poly(vinyl butyral) and Dioctyl Phthalate Films, J. Supercrit. Fluids, 28, 113120(2004). 5 R. V. Shende, T. R. Redfern and S. J. Lombardo, Defect Formation during Supercritical Extraction of Binder from Green Ceramic Components, J. Am. Ceram. Soc, 87, 1254-58 (2004). 6 D. W. Matson and R. D. Smith, Supercritical Fluid Technologies for Ceramic-Processing Applications, J. Am. Ceram. Soc, 72, 871-881 (1989). 7 S. Nakajima, S. Yasuhara, and M. Ishihara, Method of Removing Binder Material from a Shaped Ceramic Preform by Extracting with Supercritical Fluid, US Patent No. 4,731,208, March 15, 1988.

236



8

T. Miyashita, Y. Ueno, H. Nishio, and S. Kubodera, Method for Removing the Dispersion Medium from a Molded Pulverulent Material, US Patent No. 4,737,332, April 12, 1988. E. Nishikawa, N. Wakao and N. Nakashima, Binder Removal from Ceramic Green Body in the Environment of Supercritical Carbon Dioxide With and Without Entrainers, J. Supercrit. Fluids, 4, 265-269(1991). ,0 T. Chartier, M. Ferrato and J. F. Baumard, Supercritical Debinding of Injection Molded Ceramics, J. Am. Ceram. Soc, 78, 1787-92 (1995). U T. Chartier, E. Delhomme and J. Baumard, Mechanisms of Binder Removal Involved of Injection Molded Ceramics, J. De Physique III, 7, 291-302 (1997). 12 T. Chartier, E. Delhomme, J. F. Baumard, P. Marteau, P. Subra and R. Tufeu, Solubility in Supercritical Carbon Dioxide, of Paraffin Waxes Used as Binders for Low-Pressure Injection Molding, bid. Eng. Chem. Res., 38, 1904-10 (1999). ,3 F. Bordet, T. Chartier, J. F. Baumard, The Use of Co-Solvents in Supercritical Debinding of Ceramics, J. European Ceram. Soc, 22, 1067-72 (2002). 14 T. Chartier, F. Border, E. Delhomme, J. F. Baumard, Extraction of Binders from Green Ceramic Bodies by Supercritical Fluid: Influence of the Porosity, J. European Ceram. Soc, 22, 1403-09(2002). I5 H. Darcy, Les Fontaines Publiques de La Ville de Dijon, Victor Dalmont, Paris, 1856. 16 K. Kannan and K. R. Rajagopal, Flow Through Porous Media Due to High Pressure Gradients, App. Math. Comput., 199, 748-59 (2008). 17 G. G. Stokes, On the Theories of Internal Friction of Fluids in Motion, and of the Equilibrium and Motion of Elastic Solids, Trans. Camb. Phil. Soc, 8, 287-305 (1845). l8 C. Barus, Isotherms, Isopiestics and Isometrics Relative to Viscosity, Am. J. Sci., 45, 87-96 (1893). 19 P. W. Bridgman, The Physics of High Pressure, MacMillan, New York, 1931. 20 E. C. Andrade, Viscosity of Liquids, Nature, 125, 309-310 (1930). 21 S. J. Lombardo and Z. C. Feng, Pressure Distribution during Binder Burnout in ThreeDimensional Porous Ceramic Bodies with Anisotropic Permeability, J. Mater. Res., 17, 143440 (2002). 22 M. E. Harr, Mechanics of Paniculate Media, McGraw Hill, New York, 1997. 23 D.-Y. Peng and D. B. Robinson, A New Two-Constant Equation of State, Ind. Eng. Chem. Fundam., 15,59-64(1976). 24 J. M. Smith, H. C. Van Ness and M. M. Abbott, Introduction to Chemical Engineering Thermodynamics, Sixth Ed., McGraw Hill, New York, 2001. 25 A. Fenghour, W. A. Wakeham and V. Vesovic, The Viscosity of Carbon Dioxide, J. Phys. Chem. Ref Data, 27, 31 -44 (1998). 26 V. Vesovic, W. A. Wakeham, G. A. Olchowy, J. W. Sangers, J. T. R. Watson and J. Millat, The Transport Properties of Carbon Dioxide, J. Phys. Chem. Ref Data, 19, 763 -770 (1990). 27 J. W. Yun, S. J. Lombardo, Effect of Decomposition Kinetics and Failure Criteria on BinderRemoval Cycles from Three-Dimensional Porous Green Bodies, J. Am. Ceram. Soc, 89, 176183(2006). 28 Z. C. Feng, B. He, and S. J. Lombardo, Stress Distribution in Porous Ceramic Bodies During Binder Burnout, J. ofAppl. Mech., 69, 497-501 (2002). 9


· 237


MODELS OF THE STRENGTH OF GREEN CERAMIC BODIES AS A FUNCTION OF BINDER CONTENT AND TEMPERATURE Stephen J. Lombardo# and Rajiv Sachanandani Department of Chemical Engineering University of Missouri Columbia, MO 65211, USA ^Department of Mechanical and Aerospace Engineering University of Missouri Columbia, MO 65211, USA ABSTRACT The strength of ceramic green bodies is evaluated for binder distributed in two ways between ceramic particles: pendular state bonds and coated state bonds. Two models for the strength of the binder are then utilized. In the first, the strength of the binder is taken as a constant, and in the second, the yield strength of the binder is represented as a function of temperature and the activation energy for polymer segment motion. INTRODUCTION The main role of binder in the fabrication of ceramic components is to impart strength to green bodies in order to aid in handling. Such binder may consist of one or more organic compounds or polymers, and the binders are presumed to reside between particles in the pore space of the green body. Such binder, however, ultimately must be removed from the green body, and this most often occurs by thermal degradation of the organic species. During the thermal removal of binder [1-3], however, the pressure buildup arising from the decomposition of binder leads to stress in the green body, which may ultimately lead to failure. Earlier modeling efforts have been able to quantify both the buildup of pressure and the occurrence of stress [4-7]. A common feature of these models, however, is that little is said about how the strength of the green body changes during the binder removal heating cycle, when both the volume fraction of binder and the binder material properties are changing. As a consequence of both changes in temperature and binder volume fraction, the failure mode of the ceramic green body may change as well. At low temperatures and/or low binder volume fractions, failure may arise primarily from a brittle mode of failure. At high temperatures and/or high volume fractions of binder, the failure mode may be more plastic in nature. In this work, two models are evaluated to describe the strength of green ceramic bodies [8-10]. In the first model—which may correspond to brittle failure—the strength of the binder is taken as constant and then the effects of binder loading and solids loading on the green strength are examined. For the case of plastic failure, the first model is modified to include the effects of temperature on the yield strength of polymers. These two models are then applied to two cases for the distribution of binder between the ceramic particles: pendular state bonds and coated state bonds. The strength of the green body is then evaluated for different volume fractions of solid and binder.

239

Strength of Green Ceramic Bodies as a Function of Binder Content and Temperature

THEORY Onoda [9] has developed a model for the strength of porous bodies by postulating that failure occurs through the organic bonds between particles. The model then incorporates the idea that the tensile strength of the green body depends on the strength of individual bonds and the number of bonds. For an individual bond, the tensile strength depends on the binder content at the particle neck, and the force, b\ required to break a bond is given by [9,10]: (1)

F = aBA

where A is the cross-sectional area of binder at the bridge between the particles and OB is the cohesive or adhesive strength of the binder. At the neck between particles, the binder may be distributed in different ways depending on the degree of wetting of the binder on the surface of the particles. Figure 1(A) shows a nonwetting binder for which the contact angle between the binder and the particle is >90 ; such nonwetting behavior would presumably result in poor bonding between the particles. For the case of binder that wets the particles (a contact angle of

Processing and Properties of Advanced Ceramics and Composite Ceramic Transactions

Novel Processing of Ceramics and Composites: Ceramic Transactions (Ceramic Transactions Series, Vol. 195)

Handbook of Advanced Ceramics: Materials, Applications, Processing and Properties

Handbook of Advanced Ceramics: Materials, Applications, Processing and Properties

Ceramic Nanomaterials and Nanotechnology: Ceramic Transactions (Ceramic Transactions Series)

Structure and Properties of Ceramics

Processing of Nanoparticle Structures and Composites: Ceramic Transactions (Ceramic Transactions Series)

Advances and Applications in Electroceramics: Ceramic Transactions (Ceramic Transactions Series)

Materials Processing and Texture: Ceramic Transactions, Volume 200 (Ceramic Transactions Series)

Ceramic Processing and Sintering

Ceramic Materials and Components for Energy and Environmental Applications: Ceramic Transactions Volume 210 (Ceramic Transactions Series)

ceramic and glass materials. Structure, Properties and Processing

Mechanical Properties of Ceramics

Mechanical Properties of Ceramics

Ceramics and Ceramic Composites Materialographic Preparation

Processing and Properties of Nanocomposites

Processing and Properties of Nanocomposites

Ceramics and Ceramic Composites: Materialographic Preparation

Mechanical Properties and Performance of Engineering Ceramics and Composites IV (Ceramic Engineering and Science Proceedings)

Ceramics And Composite Materials: New Research

Chemical Processing of Ceramics

Advances in Sintering: Ceramic Transactions (Ceramic Transactions Series, Volume 209)

Microstructure, Property and Processing of Functional Ceramics

Fundamentals of ceramic powder processing and synthesis

Fundamentals of Ceramic Powder Processing and Synthesis

Cellular Ceramics: Structure, Manufacturing, Properties and Applications

Ceramic Materials: Processes, Properties and Applications

Materials for Electrochemical Energy Conversion and Storage (Ceramic Transactions, Vol. 127) (Ceramic Transactions Series)

Advances in Biomedical and Biomimetic Materials: Ceramic Transactions,206 (Ceramic Transactions Series)

Mechanical Properties of Ceramics and Composites: Grain and Particle Effects

Mechanical Properties of Ceramics and Composites: Grain and Particle Effects

Processing and Properties of Advanced Ceramics and Composite Ceramic Transactions

Novel Processing of Ceramics and Composites: Ceramic Transactions (Ceramic Transactions Series, Vol. 195)

Handbook of Advanced Ceramics: Materials, Applications, Processing and Properties

Handbook of Advanced Ceramics: Materials, Applications, Processing and Properties

Ceramic Nanomaterials and Nanotechnology: Ceramic Transactions (Ceramic Transactions Series)

Structure and Properties of Ceramics

Processing of Nanoparticle Structures and Composites: Ceramic Transactions (Ceramic Transactions Series)

Advances and Applications in Electroceramics: Ceramic Transactions (Ceramic Transactions Series)

Materials Processing and Texture: Ceramic Transactions, Volume 200 (Ceramic Transactions Series)

Ceramic Processing and Sintering

Ceramic Materials and Components for Energy and Environmental Applications: Ceramic Transactions Volume 210 (Ceramic Transactions Series)

ceramic and glass materials. Structure, Properties and Processing

Mechanical Properties of Ceramics

Mechanical Properties of Ceramics

Ceramics and Ceramic Composites Materialographic Preparation

Processing and Properties of Nanocomposites

Processing and Properties of Nanocomposites

Ceramics and Ceramic Composites: Materialographic Preparation

Mechanical Properties and Performance of Engineering Ceramics and Composites IV (Ceramic Engineering and Science Proceedings)

Ceramics And Composite Materials: New Research

Chemical Processing of Ceramics

Advances in Sintering: Ceramic Transactions (Ceramic Transactions Series, Volume 209)

Microstructure, Property and Processing of Functional Ceramics

Fundamentals of ceramic powder processing and synthesis

Fundamentals of Ceramic Powder Processing and Synthesis

Cellular Ceramics: Structure, Manufacturing, Properties and Applications

Ceramic Materials: Processes, Properties and Applications

Materials for Electrochemical Energy Conversion and Storage (Ceramic Transactions, Vol. 127) (Ceramic Transactions Series)

Advances in Biomedical and Biomimetic Materials: Ceramic Transactions,206 (Ceramic Transactions Series)

Mechanical Properties of Ceramics and Composites: Grain and Particle Effects

Mechanical Properties of Ceramics and Composites: Grain and Particle Effects

Recommend Documents