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

Advances in Sintering Science and Technology Ceramic Transactions, Volume 209 A Collection of Papers Presented at the I...

Author: E. A. Olevsky | Rajendra Bordia

151 downloads 2076 Views 19MB 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

Advances in Sintering Science and Technology Ceramic Transactions, Volume 209 A Collection of Papers Presented at the International Conference on Sintering November 16-20, 2009 La Mia, California Edited by

Rajendra K. Bordia Eugene A. Olevsky

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

This Page Intentionally Left Blank

Advances in Sintering Science and Technology


Advances in Sintering Science and Technology Ceramic Transactions, Volume 209 A Collection of Papers Presented at the International Conference on Sintering November 16-20, 2009 La Mia, California Edited by

Rajendra K. Bordia Eugene A. Olevsky

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

Copyright © 2010 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 107orl08ofthel976 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 978-0-470-40849-0 Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1

Contents

Preface Acknowledgements

xi xiii

APPLICATION OF SINTERING IN EMERGING ENERGY APPLICATIONS: FUEL CELLS, SOLAR CELLS, HYDROGEN STORAGE Sintering Behavior of Ce0 9 Gd 0 -,0-, 95.8 in Reducing Atmosphere

3

Hydrogen Sorption Properties of Ti-Oxide/Chloride Catalyzed Na2LiAIH6

13

High Density Green Pellets of ZrN Fabricated by Particle Processing

21

A. Kaiser, J. W. Phair, S. Foghmoes, S. Ramousse, Z. He

Enrique Martinez-Franco, Thomas Klassen, Martin Dornheim, Ruediger Bormann, and David Jaramillo-Vigueras

Thomas T. Meek, K. Gwathney, Chaitanya K. Narula, and L.R. Walker

EVOLUTION AND CONTROL OF MICROSTRUCTURE DURING SINTERING PROCESSES The Effect of Carbon Source on the Microstructure and the Mechanical Properties of Reaction Bonded Boron Carbide

29

S. Hayun, H. Dilman, M. P. Dariel, N. Frage, and S. Dub

Modification of Mass Transport during Sintering Induced by Thermal Gradient Sébastien Saunier and Frangois Valdivieso

41

FUNDAMENTAL ASPECTS OF SINTERING Effects of Crystallization and Vitrification on Sintering Properties of Bentonite Clay

53

H. Camacho, CA. Martínez, P.E. García, H J . Ochoa, J.T. Elizalde, A. García, A. Aguilar, M. Bocanegra, and C. Domínguez

Dissolution of Alumina in Silicate Glasses and the Glass Formation Boundary

61

The Effect of Volume Fraction on Grain Growth during Liquid Phase Sintering of Tungsten Heavy Alloys

71

Keith J. DeCarlo, Thomas F. Lam, and William M. Carty

John L. Johnson, Louis G. Campbell, Seong Jin Park, and Randall M. German

IN-SITU MEASUREMENTS IN SINTERING In-Situ Investigation of the Cooperative Material Transport during the Early Stage of Sintering by Synchrotron X-Ray Computed Tomography

85

R. Grupp, M. Nöthe, B. Kieback, and J. Banhart

Geopolymers Sintering by Optical Dilatometry

91

Elie Kamseu, Cristina Leonelli, and Dan S. Perera

MODELING OF SINTERING AT MULTIPLE SCALES Meso-Scale Monte Carlo Sintering Simulation with Anisotropie Grain Growth

103

Numerical Simulation of Densification and Shape Distortion of Porous Bodies in a Granular-Transmitting Medium

113

Gordon Brown, Richard Levine, Veena Tikare, and Eugene Olevsky

Junkun Ma and Eugene A. Olevsky

The Effect of a Substrate on the Microstructure of Particulate Films

125

C.L. Martin and R. K. Bordia

Modelling Constrained Sintering and Cracking

135

Atomistic Scale Study on Effect of Crystalline Misalignment on Densification during Sintering Nano Scale Tungsten Powder

149

Variations in Sintering Stress and Viscosity with Mixing Ratio of Metal/Ceramic Powders

161

Ruoyu Huang and Jingzhe Pan

Amitava Moitra, Sungho Kim, Seong-Gon Kim, Seong Jin Park, Randall German, and Mark F. Horstemeyer

Kazunari Shinagawa

vi

· Advances in Sintering Science and Technology

NOVEL SINTERING PROCESSES: FIELD-ASSISTED SINTERING TECHNIQUES Finite Element Modelling of Microwave Sintering

173

Direct and Hybrid Microwave Sintering of Yttria-Doped Zirconia in a Single-Mode Cavity

181

The Influence of Minor Additives on Densification and Microstructure of Submicrometer Alumina Ceramics Prepared by SPS and HIP

193

D. Bouvard, S. Charmond, and C.P. Carry

S. Charmond, C. P. Carry, and D. Bouvard

Jaroslav Sedlácek, Monika Michálková, Deniz Karaman, Dusan Galusek, and Michael Hoffmann

The Electro-Discharge Compaction of Powder Tungsten Carbide-Cobalt-Diamond Composite Material

205

Microwave Sintering Explored by X-Ray Microtomography

211

Evgeny G. Grigoryev and Alexander V. Rosliakov

Kotaro Ishizakl, Manjusha Battabyal, Yoko Yamada Pittini, Radu Nicula, and Sebastien Vaucher

Pulse Plasma Sintering and Applications

219

Influence of Electric Fields during the Field Assisted Sintering Technique (FAST)

227

Sintering of Combustion Synthesized TiB 2 -Zr0 2 Composite Powders in Conventional and Microwave Furnaces

237

Andrzej Michalski and Marcin Rosinski

Michaela Müller and Rolf Ciasen

Hayk Khachatryan, Alok Vats, Zachary Doorenbos, Suren Kharatyan, and Jan A. Puszynski

Production and Characterization of WC-Co Cemented Carbides by Field Assisted Sintering

249

Rafet Emre Özüdogru, Filiz Cinar Sahin, and Onuralp Yucel

Microwave Rapid Debinding and Sintering of MIM/CIM Parts

259

P. Veronesi, C.Leonelli, G. Poli, L. Denti, and A. Gatto

SINTERING OF BIOMATERIALS Analysis of Sintering of Titanium Porous Material Processed by the Space Holder Method

273

L. Reig, V. Amigó, D. Busquéis, M.D. Salvador and J.A. Calero


· vii

Effect of Sintering Temperature and Time on Microstructure and Properties of Zirconla Toughened Alumina (ZTA)

283

Sintering Zirconia for Dental CAD/CAM Technology

291

M. M. Hasan and F. Islam

Kuljira Sujirote, Sukunthakan Ngernbamrung, Kannigar Dateraksa, Tossapol Chunkiri, Marut Wongcumchang, and Kriskrai Sitthiseripratip

SINTERING OF MULTI-MATERIAL AND MULTI-LAYERED SYSTEMS Co-Sintering Behaviors of Oxide Based Bl-Materials

307

Coupling between Sintering and Liquid Migration to Process Tungsten-Copper Functionally Graded Materials

321

Laser Sintering of Nanosized Alumina Powder for Scratch Resistant Transparent Coatings

333

Optimization of Density, Microstructure and Interface Region in a Co-Sintered (Steel/Cemented Carbide) Bi-Layered Material

343

Claude Carry, Emre Yalamag, and Sedat Akkurt

J.-J. Raharijaona, J.-M. Missiaen, and R. Mitteau

Christoph Rivinius and Rolf Ciasen

A. Thomazlc, C. Pascal, J.M. Chaix

SINTERING OF NANOSTRUCTURED MATERIALS MoSi2 Formation Mechanisms during a Spark Plasma Synthesis from Mechanically Activated Powder Mixture

357

Spark Plasma Sintering of Nanocrystalline WC-12Co Cermets

367

Si3N4/SiC Materials Based on Preceramic Polymers and Ceramic Powder

379

Grain Growth during Sintering of Nanosized Particles

389

F. Bernard, G. Cabouro, S. Le Gallet, S. Chevalier, E. Gaffet, and Yu Grin

Victoria Bonache, Maria Dolores Salvador, Vicente Amigo, David Busquéis, and Alicia Castro

U. Degenhardt, G. Motz, W. Krenkel, F. Stegner, K. Berroth, W. Harrer, and R. Danzer

Z. Zak Fang, Hongtao Wang, Xu Wang, and Vineet Kumar

Atomic Investigation of Thermal Stability of Nanosized Ceria Particles on Metal Oxide Surfaces

401

Two-Step Sintering of Molybdenum Nanopowder

415

W. Jiang, M. Wong, A.R. Rammohan, Y. Jiang, and J.L. Williams

Min Suh Park, Tae Sun Jo, Se Hoon Kim, Dae-Gun Kim, and Young Do Kim

vüi


Standard and Two-Stage Sintering of a Submicrometer Alumina Powder: The Influence on the Sintering Trajectory

421

SAXS Investigation of the Sintered Niobium Powder: Method of Stabilizing Porosity and Fractal Properties

429

M. Michálková, K. Ghillányová, and D. Galusek

Leonid Skatkov

Author Index

437


· ix


Preface

This issue of the Ceramic Transactions compiles a number of papers presented at the International Conference Sintering 2008, which was held in San Diego, USA on November 16-20, 2008. The meeting was chaired by Professors Rajendra Bordia and Eugene Olevsky and was organized under the auspices of The American Ceramic Society. This was the fifth meeting in a series that started in 1995 as a continuation of a famous cycle of conferences on Sintering and Related Phenomena organized by G. Kuczynski in the period from 1967 till 1983. The first three meetings in the newly re-established series of conferences have been held at Pennsylvania State University in 1995, 1999, 2003, and the fourth has been held in Grenoble, France in 2005. In parallel to the US-based cycle of conferences on sintering, from 1969 till 2002 another important series of conferences named Round Tables on Sintering have been held in Eastern Europe and have been attended by many sintering professionals from that geographical area as well as by scientists from Western Europe and Asia. In general, over the past 50 years, there has been a series of important conferences aimed at documenting the status of sintering theory and practice. Previous meetings were also organized by the Tokyo Institute of Technology, University of Notre Dame, and the University of British Columbia. Sintering 2008 became the -largest specialized sintering forum in history by bringing together more than 200 registered participants of various research communities, which fostered the high level of scientific interaction and created atmosphere of broader international collaboration. The meeting included participants from North and Central America, Europe (both Eastern and Western), Asia, Australia and Africa. The technical program at this meeting included 203 presentations from 30 countries, which addressed the latest advances achieved in the sintering processes for the fabrication of powder-based materials in terms of fundamental understanding, technological issues and industrial applications. The conference has demonstrated a significant progress that has been made in multi-scale modeling of densification and microstructure development, better understanding of the processing of complex systems (multi-layered, composites and reactive systems). In sintering technology, innovative approaches like Field Assisted Sintering (also known as Spark Plasma Sintering) gain more attention of the materials processing community. Another very XI

timely and well represented topic was sintering and microstructure development in nanostructured materials. Papers were also presented on the sintering of bio- and energy applications-related materials. To augment the traditional technical presentations, a Plenary Round Table Discussion on the "Challenges and Opportunities in Sintering Science and Technology" identified critical avenues for research and development as well as the most exciting developments in the science and technology of sintering. This volume contains 43 papers covering a rich diversity of the sintering science and technology topics. Another 19 papers were published in a special issue of the Journal of the American Ceramic Society (July 2009). Together, a third of the papers presented during the conference were published in these two publications, Thanks go to both the conference participants and organizers who had to meet numerous deadlines to enable the timely publication of this volume and of the special issue of the Journal. We hope you will enjoy the papers assembled here and we are looking forward to see you in 2011 in Jeju Island, Korea at the International Conference on Sintering 2011. Rajendra K. Bordia University of Washington Eugene A. Olevsky San Diego State University

xii


Acknowledgements

The Co-chairs of Sintering 2008 gratefully acknowledge the advice and help given to us in organizing this meeting by colleagues on the Organizing and International Advisory Committees listed below. We also appreciate the support from the conference sponsors listed below. Support from the staff at The American Ceramic Society is gratefully acknowledged. Finally, RKB acknowledges the opportunity to work at the Whiteley Center of the University of Washington during two critical periods of editing this proceedings. Organizing Committee Iver Anderson, AMES National Laboratory Gary Messing, Pennsylvania State University Randall German, Mississippi State University Khaled Morsi, San Diego State University Ian Nettleship, University of Pittsburgh Veena Tikare, Sandia National Laboratory International Advisory Committee Debby Blaine, Stellenboch University, South Africa Aldo Boccaccini, Imperial College - London, UK Didier Bouvard, INP Grenoble, France W. Roger Cannon, Rutgers University, USA Alan Cocks, Oxford University, UK Lutgard C. DeJonghe, University of California, Berkeley, USA Terry Garino, Sandia National Laboratory, USA Joanna Groza, University of California, Davis, USA Vikaram Jayaram, Bangalore India Institute of Science, India John Johnson, Alldyne, USA Suk-Joong L. Kang of KAIST, Korea Torsten Kraft, Fraunhoffer Institute, Germany Robert McMeeking, University of California, Santa Barbara, USA

XIII

Sung-Tag Oh, Seoul National University of Technology, Korea Jingzhe Pan, Leicester University, UK Irene Peterson, Corning Inc., USA MohamedN. Rahaman, Missouri University of Science and Technology, USA Jürgen Rodel, Universität of Darmstadt, Germany Lucio Salgado, IPEN/CNEN, Brazil Graham Schaffer, Queensland University, Australia Aziz Shaikh, Ferro Corporation, USA Valery Skorohod, National Academy of Sciences, Ukraine Mark Thompson, General Electric Co., USA Francois Valdivesco, Ecole des Mines de Saint-Etienne, France Fumihiro Wakai, Tokyo Institute of Technology, Japan Antonios Zavaliangos, Drexel University, USA Conference Sponsors Sandia National Laboratories Netzsch Thermal Technology LLC Metal Processing Systems, Inc. American Elements

xiv


Application of Sintering in Emerging Energy Applications: Fuel Cells, Solar Cells, Hydrogen Storage


SINTERING BEHAVIOR OF Ce 0 . 9 Gd 0 .,O 1 . 95 . s IN REDUCING ATMOSPHERE A. Kaiser, J. W. Phair, S. Foghmoes, S. Ramousse, Z. He Rise National Laboratory, Fuel Cells and Solid State Chemistry Department 4000 Roskilde, Denmark ABSTRACT At low oxygen partial pressures and high temperatures Gd-doped ceria can be reduced and the material becomes electronic conducting (e.g. 0.08 Scm"1 at 800°C at a ρθ2 of 10"16 arm for Ceo.9Gdo.i 01.95.5 CGO 10). These properties make CGO attractive for use in oxygen membranes above 600°C. The sintering temperature of CGO ceramics might be significantly reduced, if a sintering atmosphere with very low oxygen partial pressure is applied (for example a p02 of 10"15 arm or below). In the present work, the densification behaviour of CGO 10 was investigated in reducing atmosphere and in air. Samples were prepared by tape casting and lamination of the single layers into multi-layers and by die pressing. A dilatometer was used to measure the sample shrinkages from room temperature to 1773 K with different constant heating rates. Based on the sintering results of pressed samples the activation energy for densification was determined. The activation energy for densification of CGO 10 can be reduced significantly from 770±40 to 300±40 KJ/mol by switching atmosphere from air (pO2=0.21 arm) to highly reducing conditions (pU2 down to 10" arm), which indicated enhanced densification behaviour of CGO 10 in reducing atmosphere during early stage sintering. INTRODUCTION Ceria based solid solutions have been investigated intensively as promising electrolyte materials for intermediate temperature solid oxide fuel cells (IT-SOFC), as cathode barrier layers in SOFC1 and as membrane material for oxygen separation membranes2. The doping with gadolinium leads to one of the highest ionic conductivities in ceria among different other dopants at intermediate temperatures3 (500-600 °C). Cerium Gadolinium Oxide (CGO) has therefore been proposed as electrolyte material for stainless steel supported fuel cells4. The sintering of ceria to full density requires relatively high sintering temperatures, as high as 1300°C tol600°C, depending on raw powders and processing. For many processes, such as the manufacturing of solid oxide fuel cells or membranes, it would be beneficial, if the sintering temperature to achieve dense CGO ceramics could be reduced significantly. Several studies have been under-taken on the addition of dopants to lower the sintering temperature of CGO ceramics in air5'6,7,8·9'10. The sintering behaviour of CGO is expected to be quite different in reducing atmosphere due to the reduction of Ce4+ to Ce3+ and the related change in oxygen vacancy concentration, which is expected to have a significant influence on the sintering kinetics. The influence of sintering atmosphere has so far only rarely been investigated. J.-G. Li et al." have reported an increase in density and grain growth of nano crystalline yttria doped ceria ceramics, if the atmosphere was switched to more reducing atmosphere, e.g. from oxygen to air. Grain growth and enhanced densification in CGO compared to un-doped ceria was explained by reduction of some Ce4+ to Ce3+ and a correlated formation of oxygen vacancies, which should cause rapid grain boundary migration due to changed grain boundary energies12,13. It is well known, that the CGO lattice shows a volume expansion upon reduction and volume reduction during re-oxidation13'14. This paper investigates the densification kinetics of Ceo.9Gdo.1O1.95-6 at relatively low oxygen partial pressures (ρθ2 = · Na,AlH„ + 2A1 + 3H, T

(1)

Na,AlH„ ->■ 3NaH + Al + 3/2 H, t

(2)

NaH -> Na + 1/2H, t (3) Nevertheless, irreversibility of these materials is the drawbacks to use them as hydrogen storage materials. Although there are more than one dissociation step of alkali alanates and irreversibility, these new class of materials are considered because of their low absorption temperature as well as high hydrogen content. One of the pioneers of the synthesis of different alanates using chemical methods

13


was Bogdanovic and co-workers''. However, chemical methods requires high hydrogen pressures and relative high temperatures, for example Na2LiAlH6 has been prepared from NaH+LiAlH4 in toluene under 300 bar H, at 433 K or by the reaction of NaAlH4+LiH+NaH in heptane under hydrogen pressure. Furthermore, producing sodium lithium alanate (Na2LiAlHt) results in a change of equilibrium pressure to lower value". An alternative technique such as mechanical alloying (MA) has been employed to produce nanocrystalline alanates by Zaluska and Huot". On the other hand, Ti0 2 has been successfully used5" as catalyst in Mg-based hydrogen storage materials. For sodium alanate (NaAlHJ, TiCl, was found to be one of the most effective catalysts up to now1'''8. Due to the high reactivity of Na with Cl, the general solid state reaction proposed by Sandrock8 after milling is: (1 -x)NaAlH4 + xTiCl, -> (I -4x)NaAlH4 + 3xNaCl + xTi + 3xAl + 6xH,

(4)

where x is the mole fraction of TiCl, added to the initial material. Sandrock suggested that TiCl, act just as precursor and the real catalyst could be, according to the reaction, the zero-valent Ti. In addition, the real catalyst could also be TiHx, Ti-alloy or some intermetallic compounds formed during milling. Sun et al.' used titanium n-butoxide (Ti(OBu")4) and zircon n-propoxide (Zr(Opr")4) as catalysts on NaAlH4 and supposed that Ti and Zr doping results in lattice substitution of Na-cations. In this paper, we describe the preparation of nanocrystalline Na,LiAlH6 by simple method of ball milling with TiO, and TiCl, addition in order to find out the catalytic effect of the additions on the hydrogen sorption kinetics in Na2LiAlH6.

EXPERIMENTAL The powders used were NaH (95% purity, Aldrich Germany), LiAlH4 (98% purity, Alfa Aesar Germany), TiO ; (99.5% purity, Alfa Aesar Germany) and TiCl, (99.999% purity, Aldrich Germany). The milling was carried out in a Fritsch P5 planetary ball mill using an initial ball-to-powder mass ratio of 10:1. All handling of the powders, including milling and weighting, was performed inside a glove box under continuously purified argon atmosphere (oxygen and moisture content Na,LiAlH, Na2LiAlH„ -» 2NaH + LiH + Al + 3/2 H,

(after milling)

(5)

(after desorption)

(6)

by adding Ti-based catalyst, Ti and Al have the possibility to form an intermetallic phase. However, due to the low content of catalyst used, we could not detect after milling any peaks for an Al-Ti alloy phase by XRD powder diffraction. Table 1 Reaction rates calculated between 20 and 80 % of maximum capacity Material Desorption Absorption Wt.%/s Wt.%/s Without catalyst 0.0017 0.0013 +5mol% TiO, 0.0021 0.0032 +5mol% TiCl, 0.0061 0.0081 Figure 3 shows desorption kinetics of Na,LiAlH6 without and with Ti-catalysts. Faster kinetics were observed for materials in which Ti catalysts were used, compared to the material without those additives. The time for full desorption was 7, 18 and 50 minutes for samples with TiCl,, TiO, and without catalyst addition, respectively. Desorption rates calculated, in a range of linear region of the curves (between 20 and 80 wt.% of maximum capacity), are shown in Table 1.

16

■ Advances in Sintering Science and Technology


Figure 3. Desorption kinetics of Na,LiAIH6 without and with Ti-based catalysts at 230 °C under vacuum (10"' bar) conditions. In order to analyze the phase transformations after absorption, one sample of powder without and with Ti-catalyst was charged with hydrogen at 230 °C/43 bar for 60 minutes, then the valve of sample holder was closed and cooled to room temperature. The sample holder was opened in the glove box and powders were prepared for XRD measurement. Figure 4 shows the corresponding XRD-results. The indexed patterns show that NaH and Al phases remain after absorption. For the sample with TiCl, catalyst, the presence of NaCl was detected. These results show that reaction (3) is not thoroughly reversible under these operation conditions even by catalysts additions. In the material without catalyst the proposed reversible reaction according to XRD results (Figure 4) after absorption is: 3NaH + LiH + 2A1 + 3/2 H, o · Na,LiAlHs + NaH + Al

(7)

Calculation of hydrogen content of the reaction (4) is 2.75 wt. %. From the absorption curves showed in the Figure 2, hydrogen content absorbed by the material without and with TiO, catalyst was 2.4 wt.%, which is close to the calculated. By using TiCl, the hydrogen capacity is 1.8 wt.%.

2 0, degrees

Figure 4. XRD-patterns of Na,LiAlH6 after absorption at 230 °C/43 bar.


· 17

Hydrogen Sorption Properties of Ti-Oxide/Chloride Catalyzed Na2LiAIH,

Figure 5 shows the absorption properties of Na,LiAlH6 catalyzed with TiCl, at 60 bar and different temperatures. Absorption kinetics is faster with increasing temperature. Differences in kinetics at 230 and 200 °C are small and hydrogen content of 1.8 and 1.6 wt.% is obtained after 60 minutes, respectively.

Time [min]

Figure 5. Absorption kinetics ot Na,LiAIH6 catalyzed with 5 mol% 1 rCl, under 60 bar at different temperatures. Figure 6 shows desorption curves of Na,LiAlHs catalyzed with TiCl, at different temperatures. The fastest kinetics are obtained at 230 °C and complete desorption takes about 6 minutes. The slowest kinetics measured is at 180 °C and takes about 60 minutes.

Time [mini

Figure 6. Desorption kinetics of Na,LiAIHe catalyzed with 5mol% TiCl, under vacuum (10"' bar) conditions at different temperatures. DSC-measurements Figure 7 shows DSC measurements of as milled powders after milling lOOhrs. A decomposition peak of un-catalyzed Na,LiAlH,, at a temperature range of 222 - 292 °C is in a good agreement with the results of other research groups"4. The influence of catalyst is remarkable by using TiCl,: overlapping peaks in the temperature range of 140 - 282 °C are observed. The single DSC peak for Na,LiAlH6 did not shift significantly after addition of TiO„ but shifted by 53 °C to lower temperatures after addition of TiCl,. Unknown peak is observed at 267.5 9C in the sample with addition of TiCl3. Enthalpies recorded are 54.7, 49.7 and 32 kJ/mol for materials without catalyst, with TiO, and TiCl,, respectively. A decrease of enthalpy and temperature confirm the instability of the system and eases the

18


Hydrogen Sorption Properties of Ti-Oxide/Chloride Catalyzed is^LiAIH,

decomposition and are in a good agreement with the desorption results obtained (Figure 3). Onset temperatures calculated by the program are 235, 234 and 172 °C for materials without catalyst, TiO, and TiCl,, respectively. These results confirm that TiCl, has strong influence on the decomposition, which enhances the sorption kinetics.

Figure 7. DSC measurements of as-milled at 100h of Na,LiAlHt without and with Ti-based catalysts.

CONCLUSIONS AND OUTLOOK Ti-based catalysed NaXiAlH,, was successfully synthesised by high-energy milling. Catalytic effect of TiCl, shows better absorption-desorption kinetics than TiOz, however hydrogen capacity is lowered due to NaCl formed during milling. DSC measurements confirm the strong influence on the thermal stability of Na,LiAlH„ in the presence of TiCl, catalyst by shifting the peak temperature for decomposition to about 50 °C lower value. Possible mechanisms for the catalytic effect of TiCl, were discussed by other research groups12'1 but still unclear. According to our results, elemental Ti (from TiO,) seems not have great influence on the sorption kinetics. In the case of TiCl, catalyst addition, the Na* is replaced in the Na,LiAlH6 structure by Ti/, which makes the incorporation of hydrogen into the structure easier, thereby improving the absorption and desorption kinetics. In order to overcome the capacity loss due the product of catalyst decomposition, NaCl, more efforts should to be made in the milling process and doping method. As results shown, high reactivity of Na* with Cl' is expected that different Ti-halides (F, Br, I) precursors could act as good catalyst for sodium alanate compounds. Therefore, different Ti-halides catalyst precursor such as TiCl4. TiF„ TiF4 and TiBr4 are going to be tested and will be reported ¡n a forthcoming paper. ACKNOWLEDGMENTS Authors acknowledge financial support for projects SIP-IPN 20080169 and IPN-SIP-ICYTDF 059. E. Martinez-Franco is gratefully for receiving a fellowship for Ph.D. research work by CONACyTMexico, DAAD-Germany and GKSS-Forschungszentrum.


· 19

Hydrogen Sorption Properties of Ti-Oxide/Chloride Catalyzed Na2LiAIH,

REFERENCES 'Borislav Bogdanovic, Richard A. Brandt, Ankica Marjanovic, Manfred Schwickardi and Joachim Tolle: Journal of Alloy and Compounds 302 (2000) 36-58. 2 Borislav Bogdanovic and Manfred Schwickardi: Journal of Alloy and Compounds 253-254 (1997) 1-9. 'Zaluski, L. Zaluska, J.O. Ström-Olsen: Journal of Alloy and Compounds 290 (1999) 71-78. 4 J. Hout, S. Boily, V. Güther, R. Schulz: Journal of Alloy and Compounds 283 (1999) 304-306. 'Mat. Res. Bull. Vol. 22 pp. 405-412, 1987. "W. Oelerich, Ph. D. Thesis, Technische Univertität Hamburg-Harburg, 2000. 'K.J. Gross, G.J. Thomas, CM. Jensen: Journal of Alloy and Compounds 330-332 (2002) 683-690. S G. Sandrock, K. Gross, G. Thomas: Journal of Alloy and Compounds 339 (2002) 299-308. "D Sun, T. Kiyobayashi, H.T. Takeshita, N. Kuriyama, CM. Jensen: Journal of Alloy and Compounds 337 (2002) L8-L11. "'Bogdanovic, German patent 19526434, (1995). "E. Martinez-Franco: Ph. D. Thesis, ESIQIE-IPN, Mexico City, 2006. I! K.J. Gross, E.H. Majzoub, S.W. Spangler: Journal of Alloy and Compounds 356-357 (2003) 423-428. "P.S. Rudman: Journal of Less-Common Metals, 89 (1983) 93-110. "P. Claudy, B. Bonnetot, J.-P. Bastide, J.-M. Letoffe: Mater. Res. Bull. 17 (1982) 1499.

20


HIGH DENSITY GREEN PELLETS OF ZRN FABRICATED BY PARTICLE PROCESSING Thomas T. Meek1, K. Gwathney2, Chaitanya K. Narula3, L.R. Walker4 ''2The University of Tennessee; 434 Dougherty Engineering Bid.; Knoxville, TN 37996-2200 34

' 0ak Ridge National Laboratory; MST Division, Oak Ridge, TN 37831-6115 3

Adjunct Professor, MSE Department, The University of Tennessee 434 Dougherty Engineering Bldg.; Knoxville, TN 37996-2200 ABSTRACT Fabrication of nitride nuclear fuels has generally been not successful due to instability of nitrides loss of americium nitride during the high temperature sintering process, (private communication with K. McClellan of Los Alamos National Laboratory) under hot pressing or conventional 1 arm sintering conditions to make net shapes. A low temperature-processing route that enables near net shapes in almost theoretical density is highly desirable. In order to carry out low-temperature pressing, it is necessary to synthesize nitride materials with controlled particle size. A compact made of the right mix of particles can be pressed at room-temperature to near theoretical density. In order to achieve controlled nitride powder synthesis, we are developing ceramic precursor processing that has been shown to offer unique advantages over conventional synthesis of advanced materials. In general, the precursors for metal nitrides already contain M-N bonds that are terminated into organic groups. A majority of these precursors are soluble in organic solvents enabling synthesis of powders. A major benefit of ceramic precursor processing is that the nano-particles of metal nitrides can be easily synthesized. These nano-particles can then be converted to particles of various sizes under carefully controlled thermal conditions. These powders have been blended to form bimodal distribution. Trimodal and continuous distributions are also under study. We should then be able to use these blends to isostatically press or uniaxial press samples to the desired fractional density (85%) at room temperature INTRODUCTION Metal-actinide nitride nuclear fuel is attractive because it can be used at higher operating temperatures than oxide fuels. The composite nitride fuel also has a higher thermal conductivity than does the oxide fuel. A serious drawback for the nitride fuel; however, is the loss of americium nitride during the conventional high temperature sintering cycle due to its high vapor pressure. Various conventional sintering approaches have been employed to sinter composite nitride fuel pellets. However, no conventional approach has solved the loss of americium problem. This paper investigates routes for the fabrication of high-density (>85% of theoretical) pellets of surrogate materials (eg. ZrN for UN and DyN for AmN) by using commercially available ZrN powders and by using the chemical precursor approach to synthesize the starting powders and then using either bimodal, trimodal or continuous powder distributions to process the powder into pellets using either cold isostatic pressing or uniaxial pressing at room temperature.

21


There are several traditional methods for the synthesis of zirconium nitride. Generally, a source of zirconium and a source of nitrogen are heated together at very high temperatures to obtain crystalline zirconium nitride. For example, heating of zirconium oxide and carbon in the presence of nitrogen, zirconium chloride and ammonia, zirconium chloride and sodium in the presence of nitrogen are some of the traditional methods [1-4]. Ceramic precursor technology has not been extensively developed for zirconium and dysprosium nitride materials [5]. The first ceramic precursor was reported by Brown and Maya in 1988 [6] from the reaction of [ ( C F ^ N ^ Z r with ammonia to obtain white powders. These powders, upon pyrolysis at 800°C, furnished carbon-contaminated zirconium nitride. The carbon content could be reduced by further pyrolysis in a hydrogen atmosphere. We have previously shown that ceramic precursors for zirconium nitride can be synthesized by simple reaction of hexamethyldisilazane, [(Me3Si)2NH], with zirconium tetrachloride [7]. The resulting compound is a free-flowing solid powder and can be recrystallized from dichloromethane. During pyrolysis of this compound trimethylsily chloride and hydrogen chloride are eliminated and elimination is complete by 600°C.

Figure 1: Size distribution of ZrN particles/agglomerates that formed during treatment at 1130'C for 45 minutes in vacuum.

Particles of ZrN in the 30nm - 15μηι range (Figure 1) form when sintered at 1075°c in vacuum The lattice parameter, a, for ZrN calculated from (111) and (200) is 4.567±0.0008 Á which is close to that for stoichiometric zirconium nitride (0.457756 nm from JCPDS 35-0753). This is a general approach and can be extended to synthesis of other nitrides. EDS spectra of the particles does not show Si or Cl impurities and particles are stable to oxidation under ambient conditions. There are no known ceramic precursors for Dysprosium Nitride, DyN. It is prepared from traditional methods by heating a source of Dy and a source of nitrogen at very high temperatures.

DISCUSSION During the fabrication of nitride composite nuclear fuel pellets when the green pellet is sintered conventionally at 1600°C for many hours there is observed a significant loss of americium due to its high vapor pressure. This is unacceptable and may be addressed by the fabrication of pellets using bimodal, trimodal or continuous powder distributions which when cold pressed yield densities of greater than 85%. For simulant pellets made using bimodal mixing, monosized particles of ZrN and DyN will be synthesized using the precursor technique described above. Particles of ZrN will be produced in the size ratios of 1:5 or greater (eg. 100 Á in diameter and 800 Á in diameter). DyN particles will also be synthesized in similar sizes. Figure 2 shows the effect of particle size ratio on the maximum fractional density for bimodal mixtures [8]. Note that when the size ratio is 1:1, the

22



fractional density is 63% while a ratio of 1:8 yields a fractional density greater than 80% when considering weight fractions of small versus large particles (eg. 73% large and 27% small particles in the ratio of 1:8) the optimum bimodal fractional packing density of 86% is achieved. Since packing density will increase as a function of the homogeneity of the mixture, care will be taken in how the bimodal mixture is blended.

Figure 2: The experimental results of MaGeary [ 10] showing the effect of the particle size ratio on the maximum fractional density for bimodal powder mixtures.

Figure 3: Percent of theoretical for mixtures of 10 μ ZrN powder and 44 μ ZrN powder,

Figure 3 shows bimodal mixing of two powder distributions of ZrN. One of 44μπι and one of ΙΟμιη were blended in different proportions as shown in Table 1. Maximum density achieved through dry mixing in a ball mill using 2mm zirconia media after isostatic pressing at 32.5 Ksi was 75% fractional density. Pellets of right circular cylinder geometry of OD 1.27 cm and height 1.27 cm with this density could then be sintered to 85% fractional density in a reduced time, thus resulting in less Am loss through volatilization. Table 1. ZrN Pellet Density for Bimodal Distributions Sample

Wt %44μ (G)

Wt°/c, 10μ (g)

1 9 1 2 8 2 7 3 3 4 6 4 5 5 5 *Samples 1 and 2 were broken prior to

p(g/cc) uniaxial pressed 71 70 71 69 68 isopressing.

p(g/cc) isostatic pressed

-

73 75 73


· 23


This work will also explore trimodal mixing of powders. Using data shown in Table 2 [11], a trimodal ratio 1:7:49 should yield factional densities of 87.8% to 95% depending on the volume fraction of small, medium and large particles used. Table 2 Some Optimized Trimodal Packings of Spheres [11] Relative volume % small % medium % large fractional density Size ratio 9.2 21.6 1:5:25 0.850 69.2 6.1 20.7 13.2 1:7:49 0.878 14.0 11.0 0.950 75.0 1:7:49 23.0 10.0 1:7:77 0.900 67.0 11.2 0.892 22.5 1:10:100 66.3 23.4 10.0 1:100:10000 0.916 66.6 Lastly, continuous distributions will be considered for the fabrication of simulant pellets. Using a wide distribution of particles, fractional packing densities have been reported as high as 0.95 [9, 11, 12, 13, 14, 15, 16] for powder distributions reported in the literature. Once the specified powder distributions are produced, various blends will be made by dry ball milling or wet milling in hexane for at least four hours or blended in a spex mill for a few minutes in hexane. The resultant powder will then be uniaxially pressed and cold isostatically pressed into pellets with an aspect ratio of approximately 1.0. CONCLUSION ZrN pellets of 75% fractional density have been made by blending 44μπι ZrN powder with ΙΟμπι ZrN powder in a bimodal distribution and then isostatically pressing at 37.5 Ksi. Initial attempts at trimodal distributions have not resulted in optimum density pellets (85% fractional density) because of hard aggolermates in the powders. These aggolermates have also affected the densities achieved in the bimodal distributions. ACKNOWLEDGEMENT The authors would like to acknowledge the support provided for this work from DOE (DE-FC07-06-ID14731). REFERENCES 1. Y. Kaeda, T. Oei, Jajn Kokai, 89-76,905 (1989) CA 111 (1989) 156958. 2. K. Hirano, Y. Miyamoto, M. Koizumi, Yogyo Kyokaishi, 95 (1987) 906, CA 107 (1987) 159987. 3. S. Somiya, K. Suzuki, M. Yoshimura, Adv. Ceram., 21 (1987), 279. 4. T. Ilda, T. Mitamura, Kagaku Kogyo, 37 (1986), 720.

24



5. C.K. Narula Ceramic Precursor Technology and Its Applications (New York, NY:Marcel Dekker, 1995). 6. G.M. Brown and L. Maya, J. Am. Ceram. Soc, 71 (1988), 78. 7. C.K. Narula and L. F. Allard, J. Mater. Chem., 8 (8) (1998), 1881-1884. 8. G.M. Brown and L. Maya, J. Am. Ceram. Soc, 71 (1988), 78. 9. R.K. McGeary, "Mechanical Packing of Spherical Particles," Journal of the American Ceramic Society, 44 (1961), 513-522. 10. C.C. Furnas, "Grading Aggregates I - Mathematical Relations for Beds of Broken Solids of Maximum Density," Industrial and Engineering Chemistry, 23 (1931), 1052-1058. 11. R.F. Fedors and R. F. Landel, "An Empirical Method of Estimating the Void Fraction in Mixtures of Uniform Particles of Different Size," Powder Technology, 23 (1979), 225-231. 12. J.V. Milewski, "Packing Concepts in the Utilization of Filler and Reinforcement Combinations," Handbook of Fillers and Reinforcements for Plastics, ed. H.S. Katz and J.V. Milewski, (New York, NY: VanNostrand Reinhold, 1978), 66-78. 13. W. B. Fuller and S. E. Thompson, " The Laws of Proportioning Concrete," American Society of Civil Engineers Transactions, 59 (1907), 67-143. 14. F.O. Anderegg, "Grading Aggregates II - The Application of Mathematical Formulas to Mortars," Industrial and Engineering Chemistry, 23 (1931), 1058-1064. 15. A.H.M. Andreasen, "Ueber die Beziehung Zwischen Kornabstufung und Zwischenraum in Produkten aus losen Kornern (mit einigen Experimenten), Kolloid Zeitschrift, 50 (1930), 217-228. 16. N. Peronium and T.J. Sweeting, "On the Correlation of Minimum Porosity with Particle Size Distribution," Powder Technology, 42 (1985), 113-121.


· 25


Evolution and Control of Microstructure during Sintering Processes


THE EFFECT OF CARBON SOURCE ON THE MICROSTRUCTURE AND THE MECHANICAL PROPERTIES OF REACTION BONDED BORON CARBIDE S. Hayun, H. Dilman, M. P. Dariel, N. Frage Department of Materials Engineering, Ben-Gurion University of the Negev, P. 0. Box 653, BeerSheva 84105, Israel S. Dub Institute for Superhard Materials of the National Academy of Science of Ukraine Avtozavodskaya str. 2 Kiev 04074 Ukraine ABSTRACT The present communication is concerned with the effect of the carbon source on the microstructure and mechanical properties of reaction bonded boron carbide composites. The composites were fabricated by molten Si-infiltration of partially sintered boron carbide preforms with 20, 30, or 40 vol.% porosity with and without free carbon addition. The infiltrated composites consist of four phases namely the original boron carbide particles, the ternary Bi2(B,C,Si)3 compound, formed in the course of the infiltration process, ß-SiC and residual silicon. In the absence of initial free carbon, the ß-SiC phase appears as plate-like particles. In the presence of initial free carbon, the ß-SiC phase particles display an irregular polygonal form. The plate like morphology of the SiC phase improves significantly the strength, the fracture toughness and reliability of the infiltrated composites. For an equal volume of SiC, the high aspect ratio of the plate-like particles increases their number per unit volume and thereby, the number of boundaries that a propagating crack has to cross. Moreover, crack deflection on SiC plates was also observed. INTRODUCTION Light ceramics are particularly attractive for personal, land and airborne vehicle armor. High values of hardness are by common consensus of crucial importance for good ballistic resistance. Thus, it is not surprising that boron carbide , the hardness of which is third after that of diamond and cubic boron nitride (CBN), and which has one of the lowest mass per volume ratio, has long been considered as a choice armor material candidate. However, difficulties and costs of fabrication, mainly by hot pressing, are factors of paramount importance in determining whether a particular material is considered for armor applications and limit the extended use of boron carbide based armor. An alternative way for the fabrication of a fully dense boron carbide composite is the socalled "reaction bonding" process2"4. According to mis method, a green body of boron carbide with or without free carbon is infiltrated with molten silicon. The reaction of molten silicon with the boron carbide particles or with free carbon leads to the formation of silicon carbide. The final composite consists mostly of the initial boron carbide grains, a newly formed Bi2(B,C,Si)3 phase, which surrounds the initial boron carbide particles, ß-SiC and some unreacted residual silicon. In order to improve mechanical properties of the composites the amount of the residual silicon has to be reduced. One approach for achieving this goal is to reduce a porosity of the boron carbide preform by its pre-sintering before infiltration with molten Si. This approach was described in our previous communication5 and allows fabricating composites with and without free carbon additions. It was established that the morphology of the new formed SiC phase depends on the

29

Effect of Carbon Source on Microstructure and Mechanical Properties of Boron Carbide

carbon source5. In composites fabricated without free carbon addition the ß-SiC phase appears with a plate-like morphology, while free carbon addition to boron carbide leads to the formation of SiC particles with a polygonal shape. In addition, boron carbide particles display a core-rim structure and the rim regions, the composition of which corresponds to the ternary B]2(B,C,Si)3 carbide, interconnect to a large extent the boron carbide particles. The mechanism of the rimcore structure formation has been discussed previously5 and was attributed to the dissolutionprecipitation process in the boron carbide - silicon system. The issue of free carbon addition in the reaction-bonding process is of current interest both from fundamental and processing standpoints. The present work was performed in order to further our understanding and to evaluate the effect of the resulting microstructural features on the mechanical properties of the reaction bonded composites, fabricated with and without free carbon additions. EXPERIMENTAL PROCEDURE. Sample fabrication. Boron carbide (average size of about 5μπι, "Mudan Jiang" Chinese company) powder were uniaxially compacted at 10 MPa and then partially sintered in the 1900-2100°C temperature range for 30 min in order to obtain preforms with 20, 30 or 40 vol.% porosity. Free carbon was added to some of the preforms using a water-sugar solution (50:50). The preforms were infiltrated with this solution, dried, and heat-treated at 500°C in inert atmosphere (Ar 99.999%). Under these conditions, complete pyrolysis of the sugar takes place. The preforms with and without carbon addition were subsequently infiltrated with liquid silicon (Alfa-Aesar 98.4%) in a vacuum furnace (10~5 torr) at 1480°C for 20 min. The infiltration was carried out by placing an appropriate silicon lump on the top of the porous preform. The composites fabricated without carbon addition are denoted as (type-A) and the composites fabricated with the free carbon addition are denoted as (type-B). Microstructural investigation. The microstructure of the samples was studied by optical microscopy (OM, Zeiss Axiovert 25), scanning electron microscopy (SEM, JEOL-35) in conjunction with an energydispersive spectrometer (EDS) and a wavelength-dispersive spectrometer (WDS). The samples for the OM and SEM characterization were prepared using a standard metallographic procedure that included a last stage of polishing by 1 μηι diamond paste. Image analysis was performed using the Thixomet software in order to determine the amount of residual silicon in the composites. In order to emphasize the difference between the rim and core regions and the SiC morphology an electro-chemically etching in KOH solution and chemical etching (HF-NHO,) were used respectively. The phase composition and the structure of the samples was analyzed by X-Ray diffraction (XRD), using a Rigaku RINT 2100 diffractometer with Cu Ka radiation. Mechanical Properties Mechanical properties (hardness and Young's modulus) of the core and rim regions of boron carbide particles, were determined by the nanoindintation technique preformed using a Nano Indenter-II, MTS Systems Corporation, Oak Ridge, TN, USA. The Young's modulus of the composites was determined by the pulse-echo technique using a 5 MHz probe. The Young (E) and shear (G) moduli, were derived from the longitudinal Q and shear Cs speeds of sound and using the density values, determined by the Archimedes method. Vickers hardness of the

30



composites was measured under 20N load in a Buehler-Micromet 2100 hardness tester. The fiexural strength was determined on the basis of three-point bending tests on 3x4x20 mm bars in an LRX Plus LLOYD instrument (Lloyd Instruments, Fareham Hants, U.K.). The fracture toughness of the composites was measured by a single-edge notched beam (SENB) method. The notch on a specimen was machined by an electrical discharge machining (EDM) with a 0.1 mm thick Cu wire. The notch depth was measured by an optical microscope and was close to one quarter of the specimen width W (see Fig 1). The specimens were fractured in the three-point bending mode. The fracture toughness value was calculated using equations:6

Kxc=g{alw\^]

3[a/Wfs 2[l - a/W]

(1)

g(a/^) = 1.9109-5.1552(a/FF)+12.6880(a/fr)2 19.5736(a/Pf )3 +15.9377(a/ff)4 -5.U54(a/Wf where Pmax is the maximum load, W is the specimen thickness (4±0.lmm), b is the specimen width (3 ± 0. lmm), L is the span length between supports ( 20mm), a is the notch depth, and g(a/w) is the stress intensity shape factor. p

Figure 1. Single- edge notched beam test

RESULTS AND DISCUSSION Microstructure of the infiltrated composites The XRD spectra of type-A and type-B samples are shown in Fig.2. The diiffractograms are quasi identical and show that four phases, namely, the original boron carbide particles, the ternary Bi2(B,C,Si)3 compound, ß-SiC and residual silicon are present in both types of infiltrated composites.


· 31


- -r~

20

~T

|

25

i

1

30

■

1

35

1

1

40

'

I

45

^

50

2Θ

Figure 2. XRD pattern of type-A and type-B composites. In type-A samples, SiC is formed as the outcome of the reaction between molten silicon and carbon that originates in the boron carbide phase. In type-B samples, the SiC phase is mostly formed by the reaction between free carbon and molten silicon. A typical microstructure of the composites is presented in Fig. 3. In the type-A composite the ß-SiC phase appears as white plate-like particles (Fig. 3b), while in the type-B composite the ß-SiC phase appears in an irregular polygonal form (Fig. 3a) and only a small fraction of the particles display the plate-like form. The light-gray regions correspond to residual silicon and the dark gray areas correspond to the partially sintered boron carbide skeleton. The microstructure after etching and removal of the residual silicon clearly puts in relief the morphology differences of the SiC phase (Fig. 4).

Figure 3. The microstructure (SEM images) of type-A (a) and type-B (b) composites. The initial open porosity of the preforms was about 20 vol.%.

32



Figure 4. SEM image of the chemical etched (HF-NH03) sample of type-A (a) and type-B (b) composites. The morphology of the SiC phase is clearly apparent. The mechanism of the reactions during the infiltration of the composites, which leads to the morphology differences in the composite have been, described previously5. In addition, the interaction between silicon and boron carbide leads to the formation of the "core-rim" structure (Fig. 5), consisting of an inner boron carbide core surrounded by a ternary boron carbide (B12(B,C,Si)3) rim.

Fig. 5. SEM image (backscatter electrons) of an electro-chemically etched (KOH solution) sample where the "core-rim" structure is put in evidence. Initial boron carbide particles (black core region) surrounded by a 3-7 μηι thick envelope (lighter black rim region).


· 33


According to the image analysis7 of the microstructure (Table I), the amount of the SiC phase increases from 10 to 16 vol% with increase initial porosity of the perform from 20 to 40 vol%. It is noteworthy that the relative amount of the SiC phase depends essentially on the initial porosity of the preforms and only to a slight extent on the source of carbon. Table I - Phase composition (average values) of the composites. Boron Initial Carbon Silicon Materials porosity, addition, Carbide, vol. % vol. % Vol. % vol. % 20±3 0 80±3 10±3 type-A 70±3 30±3 0 16±3 40±3 0 60±3 24±3 20±3 3±0.5 80±3 10±3 type-B 30±3 4±0.5 70±3 15±3 40±3 6±0.5 60±3 25±3

Silicon Carbide, vol. % 10±3 14±3 16±3 10±3 15±3 15±3

MECHANICAL PROPERTIES. Mechanical properties of core (initial boron carbide particles) and the rim regions. The mechanical properties of the silicon containing boron carbide compound (Bi2(B,C,Si)3) were studied by nano-indentation. Typical force-displacement curves for B4C (core) and for Bi2(B,C,Si)3 (rim) are presented in (Fig. 6). The raw data were treated according to Oliver and Pharr8. The hardness and the Young's modulus values of the Bi2(B,C,Si)3 phase are slightly higher then those for the initial boron carbide phase (Table II). The inspection of a crack propagation path indicates that the boundary between the core and the rim regions is a relatively strong one and that no crack deflection takes place (Fig. 7). These observations are in a good agreement with the TEM analysis of the boundary between the core and the rim, which is apparently a semi coherent boundary5.

Table II. The values of Young modulus and hardness for Bi2(B,C,Si)3 and B4C Young Contact Load Hardness Displáceme diameter modulus nt ,nm ,mN ,GPa ,nm ,GPa 63.8+4.1 474+34 46.1+4.2 103.2+4.3 10+0 B12(B,C,Si)3 B4C 106.3+2.7 10+0 67.9+37.9 460+23 42.0+3.3

34



40

60 Displacement, n m

Figure 6. Force vs. displacement plot for core (B4C) and rim (B12(B,C,Si)3) regions.

Figure 7. The crack propagation path in the composite material underlines the strength of the boundary between the core of the boron carbide particles and the adjacent rim.

Mechanical properties of type-A and type-B composites The Young's modulus and the hardness values of the composites decrease with increasing the amount of residual silicon (Fig 8 and 9). It is important to stress that the hardness values refer to the average hardness of the composite and reflect the contribution of the different phases with widely varying individual hardness values4.

Figure 8. Elastic modulus of the composites as a function of the residual silicon

Figure 9. Vickers hardness of the composites as a function of the residual silicon


· 35


The experimental results for the flexural strength exhibit a different behavior (Fig. 10). The amount of the residual silicon exerts only a minor effect on the flexural strength, while the effect of the carbon source is significant. The flexural strength of the composites with carbon addition (type-B) is significantly lower then those of type-A composites. A similar tendency was observed for the fracture toughness of the composites. The values of the fracture toughness, which were obtained for the composites, fabricated by infiltration of the preforms with about 30 vol. % porosity, are 3.62 ± 0.16 MPa4m and 2.57 + 0.36 MPa4m for type-A and type-B composites, respectively.

O ■

type-A type-B

t C

250-

lOVJ-l

5

,

1

10

,

1

,

15

1

,

20

1

25

,

1

30

,

1

35

residual silicon, % vol.

Figure 10. Flexural strength of the composites as a function of the amount of residual silicon These results may be attributed to the specific plate-like morphology of the SiC phase in the type-A composites. A similar strengthening effect of plate-like SiC particles on ceramic composites was reported in9"11.The presence of the SiC particles with the plate-like morphology affects the crack propagation through the composites (Fig. 11). As was noted above, the volume fraction of SiC particles in the composites, fabricated from the preforms with a given porosity, doesn't depend on the carbon source. Moreover, the polygonal SiC particles are significantly coarser than the plate-like particles. These features stand behind the larger number of the particles with the plate-like morphology per unit volume and thus the larger number of boundaries, which are crossed by a crack (Fig. 11 a,b) resulting in larger crack energy losses. It is also noteworthy that crack deflection takes place on the interaction with the SiC plates (Fig. 11 c,d).

36



Figure 11. SEM images of crack propagation paths in the composites fabricated with and without carbon additions. The scatter of mechanical properties of the reaction bonded boron carbide composites is of relevance. The WeibuU modulus12 of the composites with 15-16 vol.% of residual silicon, as determined on two groups of 16 samples, is equal to 5.84 and 3.67 for type-A and for type-B composites, respectively (Fig. 12).


· 37


1.5-, 1.0-

_■

type-A

0.5-

■ 4

0.0-

5" -0.5-

|.,.„

■ ■

ψ ■ m

-1.5· -2.0-2.5-

>

*. m=5.Sfa0.24 R2=0.975

1

type-B

■ ■ ,r

0

'm

■

■m

1-1

a

/'

3 ' -2

'

m=3.67±0.16 R2=0.973

,,É

m

-3

.' ■

5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4 1η(σ)

4.8 5.0 5.2 5.4 5.6 5.8 6.0 6.2 1η(σ)

Figure 12. Weibull plots of flexural strength of the type-A and type-B composites. SUMMARY AND CONCLUSIONS Reaction bonded boron carbide composites, fabricated with and without free carbon additions, consist of four phases: the original boron carbide particles, the ternary Bi2(B,C,Si)3 compound, ß-SiC and residual silicon. With no free carbon added, the ß-SiC phase appears as plate-like particles. With an initial free carbon addition, the ß-SiC phase displays mostly an irregular polygonal form. The plate like morphology of the SiC phase enhances significantly the strength and the fracture toughness of the infiltrated composites. It does not affect the hardness and the stiffness of the composites. This specific morphology of SiC provides a larger number of the particles with a high aspect ratio per unit volume and thus a higher number of boundaries to be crossed by a propagating crack REFERENCES 'F. Thévenot, Boron carbide - a comprehensive review, J. Eur. Cer. Soc. , 6 205-225 (1990). 2 M. K. Aghajanian, B. N. Morgan, J. R. Singh, J. Mears and R. A. Wolffe, A New Family of Reaction Bonded Ceramics for Armor Applications, in "Ceramic Armor Material by Design" Eds. J. W. McCauley et. al. Ceramic Transactions, Vol. 134, American Ceramic Society 2001, pp. 527-539 K.M. Taylor and R.J. Palicke, Dense Carbide Composite for Armor and Abrasives, U.S. Pat. No. 3 765 300, Oct. 16 1973 4 S. Hayun, A. Weizmann, M. P. Dariel and N. Frage, The Effect of Particle Size Distribution on the Microstructure and the Mechanical Properties of Boron Carbide-Based Reaction-Bonded Composites, Int. J. Appl. Ceram. Tech., DOI: 10.1111/J.1744-7402.2008.02290.X (2008) 5 S. Hayun, N. Frage, M. P. Dariel, The morphology of ceramic phases in BxC-SiC-Si infiltrated composites, J. Sol. St. Chem., 179( 9), 2875-79 (2006) 6 ASTM E399-74. American Society for Testing Materials Standard Test method for plan Strain Fracture Toughness of Metallic Materials PP. 923-36 ASTM Philadelphia PA (1983). 'Underwood, E. E.„ Weibel. E. R. and Exner, H. E., and H. P. Hougardy, in Physical Metallurgy, edited by Robert. W. Cahn and Peter Haasen (North-Holland, Amsterdam, 1996), pp. 1000-1007.

38



W.C. Oliver and G.M. Pharr, An Improved Technique for Determining Hardness and Elastic Modulus Using Load And Displacement Sensing Indentation Experiments, J. Mater. Res, 7(6), 1564-83 (1992). K. Lee, Y.C. Kim and C.H. Kim, Microstructural development and mechanical properties of pressureless-sintered SiC with plate-like grains using AI2O3-Y2O3 additives. J. Mater. Sei., 29, 5321-26, (1994). 10 W. J. Moberlychan, J. J. Cao and L.C. DE Jonghe, In Situ Toughened Silicon Carbide with AlB-C Additions, Acta Mater., 6(5), 1625-35, (1998). "S. Hayun, D. Rittel, N. Frage and M.P. Dariel, Static and Dynamic Mechanical Properties of Infiltrated B4C-S1 Composites, Mat. Sei. Eng. A, 487(1-2), 405-409, (2008). I2 W. Weibull, "A Statistical Distribution Function of Wide Applicability. J. Appl. Mech., 18 293297(1951).


· 39


MODIFICATION OF MASS TRANSPORT DURING SINTERING INDUCED BY THERMAL GRADIENT Sébastien SAUNIER, Francois VALDIVIESO Departement MPE, Centre SMS, UMR CNRS 5146, Ecole Nationale Supérieure des Mines de Saint-Etienne 158 cours Fauriel, 42023 Saint-Etienne cedex 2, France ABSTRACT Due to the developments of the new sintering processes, the thermal cycles duration has been considerably reduced. But it turns out that this cycle time decrease can also conduct to thermal gradients in the sample. Up to day, the effects of such gradients are not well known. That's why the aim of this study is to contribute to understand phenomena induced by thermal gradients applied during the initial stage of sintering. We will observe the consequences on this stage but also on the further densification. Two important results are brought out: - For pure alpha-alumina : the comparison between conventional treatments and treatments under thermal gradients has shown that a gradient modifies the specific surface area drop and then accelerates the further densification. - For alpha-alumina containing impurities : The experiments have revealed that a gradient can modify the impurities' dissolution and change the mass transport. Finally grain growth can be favoured at the expense of densification. INTRODUCTION Compared to conventional sintering, sintering under microwaves leads to a better homogeneity and to an improvement of the microstructure with a reduction of porosity and grain size1' . Moreover, experimental data on spark plasma sintering showed a higher sintering activity3'4. In addition to these advantages, both new processes are based on very fast heating rate. In such processes, the high heating rate inevitably leads to a complex temperature distribution, and possibly to the appearance of temperature gradient in the piece. The presence of density gradients in samples sintered by field activated sintering processes is mentioned in the literature5'6. This constitutes an indirect evidence of the development of temperature gradient during this sintering process. Moreover, Zavaliangos et al.7'8 have recently studied the temperature distribution in the FAST device using the finite elements methods simulation. They have showed the development of temperature gradient in this device. Although thermal gradients are present in a lot of sintering processes, few experimental studies treat about the consequences induced by thermal gradients 9 ' 011 . This can be explained by the difficulty to dissociate the effect induced by heating rate and the effect induced by the thermal gradient. Then the consequences of a thermal gradient are not well known. Nevertheless, it seems that a thermal gradient can enhance the densification. Other studies have mentioned that a thermal gradient can induce mass transport which lead to a demixing in multi-component oxides12'13. The present study deals with the quantification of the effect induced by the existence of a thermal gradient applied during the initial stage of sintering on submicronics aluminas with different purities. 41

Modification of Mass Transport during Sintering Induced by Thermal Gradient

MATERIALS STUDIED AND EXPERIMENTAL METHODS Powder formulation and processing The powder used in this study is a submicronic a-alumina powder (named UF powder) generated by ex-alun process. This alumina contains initially 210 ppm of impurities. The main impurities are: Na, K, Si and Ca. To carry out a slurry of 60 wt.% powders, alumina is dispersed in a pH 10 deionised water using polyacrylic acid (Acros Organics, USA) of molecular weight 2000. After 24 hours in a mixer with rollers, attrition using zirconia balls doped with CaO, MgO and S1O2 has been carried out (830 rpm for 1 h). In these conditions, the contamination in zirconia due to the attrition is quantified at 0.79 weight % (measured by X-ray fluorescence). Then 2.25 wt.% of polyvinyl alcohol 4/125 (Prolabo, France) and 0.75 wt.% of plasticizer polyethylene glycol 4000 (Merck-Schuchardt, Germany) are added. After attrition milling the powders are atomized. Cylindrical samples of 6 cm lengths are obtained by uniaxially compaction at 40 MPa in a steel die of 8 mm diameter and cold isostatically compacted at 400 MPa. Finally, in order to remove the organic phases, green samples are treated during 1 hour at 600°C. The theoretical density of the material takes into account the zirconia contamination coming from attrition milling (measured by X-ray fluorescence). After debinding, the density homogeneity of the green sample is verified (equal to 53.1 % ± 1% TD). Furnace with heat gradient To carry out thermal gradients, a muffle furnace has been used (figure 1). Inside this furnace, different systems of porous refractory masks are introduced to make a thermal barrier. By the modification of the thickness of the masks, thermal gradients can be reached. To study the thermal gradient in the initial stage of sintering, the samples stay under the mask. The temperature regulation of the furnace allows to obtain a temperature difference of 100°C on the 6 cm length sample. The temperature applied on the sample is reported in figure 2. After this pre-treatment the samples are cut in sections of one centimetre length and then undergo various analyses (analyses of specific surface areas, sinterability,...). In these operating conditions, each section do not have the same thermal history. To quantify thermal gradient effects, reference samples (1 cm length) are sintered with a same thermal schedule of each section but without thermal gradient.

Figure 1. Representation of the thermal furnace.

42



Figure 2. Temperature profile in the thermal furnace Experimental analysis Relative densities and porosities are measured by using the water immersion method. The error of measurement on the density is estimated at 0,1 % in the case of the double weighing. This error can reach 0.5 % in the case of triple weighed. To study the influence of the thermal gradient in the initial stage of sintering, the specific surface area of the specimens is evaluated by the Brunauer-Emmet-Teller (BET) method. Microstructural observations are carried out via scanning electron microscopy (model 6500 F, JEOL, Japan). RESULTS Influence of a thermal gradient applied in the initial stage of sintering Analyses of specific surface area measured by the BET method are done on presintered samples (UF alumina) under thermal gradient and without thermal gradient. The specific surface area is measured at both extremities (average temperatures of 880°C and 980°C) of the samples treated under thermal gradient and compared with those obtained in conventional treatment (Table 1). It appears a more important specific surface area decrease for conventional heating than for heating under thermal gradient. This result is valid for an average temperature of 880°C and 980°C. However, this difference is less significant when the temperature increases. These results obtained suggest that a thermal gradient applied in initial stage of sintering leads to neck formation delay. Table 1 . Specific surface area of UF alumina Green samples Conventional pre-treatment at 880°C Pre-treatment under gradient at 880°C Conventional pre-treatment at 980°C Pre-treatment under gradient at 980°C

Specific surface area (m2/g) 17.3 ±0,1 AS/So (%) 14.2 ±0,1 -17.9 -12.7 15.1 ±0,1 12.8 ±0,1 -26.0 13.4 ±0,1 -22.5

To see the consequences of a thermal gradient applied in the initial stage of sintering on the future densification of the material, a dilatometric study is realised for each section. A relatively low temperature of sintering is chosen (1450°C) in order not to be nearby to the theoretical density and then to mask the effects of gradient. Besides, the heating rate used is fast (20cC/min), in order to avoid the screening of the initial treatment effects. The dilatometric results (Figure 3) show a densification gain (compared to conventional sintering) when the sample is first pre-treated under thermal gradient. This gain is more important for a pre-treatment at lower temperature.


· 43


Figure 3. Dilatometric behaviour of UF alumina after different pre-treatment. Effect thermal gradient and impurities on sintering To enhance the level of impurities present in the alumina powder, the attrition time has been increased. Thus the contamination in zirconia measured by X-ray fluorescence in this new powder (noticed UF*) is 1,81 weight %. A same pre-treatment under thermal gradient is applied on the UF* alumina. After cutting up the pre-treated samples, measures of specific surface areas have been carried out (Table 2). For a same pre-treatment under gradient, it appears a strongest specific surface area decrease for UF * samples (i.e. sample with more impurities) than for UF samples. Table 2. Specific surface area of UF and UF* alumina after pre-treatment under gradient. UF alumina SBET (m2/g)

Green samples Pre-treatment at 880°C Pre-treatment at 980°C

AS/SO

(%)

UF* alumina

Final density (1450°C) SBET(nrVg)

AS/So (%)

Final density (1450°C)

18.2 ±0,1

17.3 ±0,1 15.1

-12.7

90.1%TD

15.0

-17.6

77.1%TD

13.4

-22.5

90.9%TD

13.2

-27.4

84.8%TD

Nevertheless, after sintering cycle (20°C/min up to 1450°C), the densification rate of samples with impurities (UF*) are lower than those recorded for pure alumina UF (Table 2). This phenomenon is the more pronounced if the pre-treatment is realized at low temperatures.

44 · Advances in Sintering Science and Technology


Sintering with high heating rate UF* sample (without pre-treatment) has been sintered at 1450°C with a heating rate of 20°C/min in a muffle furnace and in a dilatometer. The obtained densification is 87.3 + 0.3 %TD in the case of the muffle furnace. For the sintering in the dilatometer the final density is only 63.3 % ± 0.8 %. The microstructure obtain in this two cases are presented in figure 4. For sintering in the muffle furnace, the grain size is smaller in spite of a better densification.

Figure 4. Microstructural observations after sintering with a heating rate of 20°C/min in a dilatometer (a) and in a muffle furnace (b).

For an identical thermal cycle (heating rate and sintering temperature), the behaviour of UF* is totally different according to the furnace type (muffle and dilatometer). Nevertheless, the temperature field is more uniform in the muffle furnace than in the dilatometer. Thus, reactivity difference experimentally observed are linked with the existence of thermal gradient in the dilatometer. Besides, in the next part, we are going to study the effects of technological parameters of the dilatometer affecting the thermal gradient value. Various parameters can affect the thermal gradient value which can be developed in the dilatometer: - the gas flow; - the sample height; - the heating rate. The influences of these parameters on the alumina UF* sintering without presintering are summarised in table 3. Increasing the heating rate or the gas flow or the sample heights leads to a significant decrease in final densification. In addition, analysis of microstructure fracture show an increase of the grains size and a faceted aspect when the gas flow, the heating rate or the sample heights increase (Figure 5).


· 45


Table 3. Effect of technological parameters on the final density. Heating Gas Sample Final rate flow height density (°C/min) (L/h) (mm) (% TD) 92,4 5 1,8 7 63,3 20 1,8 7 76,2 20 0 7 (a) 64,1 20 0,5 7 63,3 20 1,8 7 62,4 20 3 (b) 7 68,1 20 1,8 (c) 2 63,3 20 1,8 7 59,5 1,8 15 (Φ 20

Figure 5. Microstructure fracture of UF* sintered alumina according to conditions presented in table 3. Effect of intergranulares impurities At any time during sintering, it is possible to create a thermal gradient with dilatometer thanks to particular condition (fast heating rate and high gas flow). Thus we apply different thermal cycles, with high heating rate and high gas flow in the initial stage or all along sintering process. For UF* samples, the temperature gradient induced by the dilatometer is applied only in the initial stage of sintering (between 20°C and 980°C). Beyond 980°C, the heating rate is 5°C/min (minimization of thermal gradient). The dilatometric curve is reported in figure 6-(a). This sintering cycle leads to a densification of 73.7% TD.

46



For comparison, the figure 6 also presents the dilatometric behaviour of UF* alumina sintered until 1450°C at two heating rate : 5°C/min (figure 6-(b)) and 20°C/min (figure 6-(c)).

Figu re 6. densification curves of UF* alumina, from dilatometric experiments, (a : 20°C 20°ami° »980°C 5°c/mi° »1450°C / b : ?Π°Γ 5°c/min >14Sn°r / c : 20°C 20°c/mip» 1450°C) Dilatometric studies reveal that a thermal gradient applied only in the initial stage of sintering (figure 6-(a)) leads to a densification which is far less advanced than when conventional heat treatment is applied over the cycle sintering (figure 6-(b)). The densification of UF* alumina is the less advanced when a thermal cycle is present in all the thermal cycle (figure 6-(c)). On this figure, the gap is more important between curves (a) and (b) than between curves (a ) and (c). In consequence, for UF* alumina, the existence of a thermal gradient in the initial stage sintering is the principal source of the densification delay. In the case of a powder (like UF*) containing impurities coming from attrition contamination (ZrC>2, S1O2, MgO, CaO), these impurities are initially located in intergranular position. The location of these impurities is therefore different from that of impurities initially present in the powder (dissolved in grains of alumina due to process ex-alun production). The aim of this part is to establish whether or not the impurities location before the gradient application affects the densification. A long conventional pre-treatment (72 hours at 920°C) is realized. After this pretreatment, sintering cycle in a dilatometer with a heating rate of 20°C/min up to 1450°C is applied. The density obtained in this case is 91.2%TD. Thus, in the case of UF* alumina (alumina containing initially intergranular impurities), a long thermal schedule (long dwell time) at low temperature drives to a dissolution of impurities before the sintering starts. This can finally increase the densification under thermal gradient. SUMMARY AND INTERPRETATION OF RESULTS The experimental study highlights the effects of a thermal gradient applied in the initial stage of sintering. This gradient must be benefit to the further densification in the case of pure alumina. By delaying the necks formation, a temperature gradient applied in the initial stage of sintering can keep a high sintering potential and therefore lead to a better densification. Advances in Sintering Science and Technology

· 47


In other side, this gradient can also block the densification and increase the grain size. This phenomena has been observed in the case of alumina which contains intergranular impurities. Such grains growth without significant densification has previously been observed in different materials such as A120314, Zr0 2 15, Ti0 2 I6or ZnO ". The causes advanced by the authors to explain the existence of such porous structure are based on an amplification of evaporation-condensation or surface diffusion mechanisms. In the case of two particles in contact, it produces a rapid enlargement of the neck and then to a grain boundary moving towards the curvature centre, leading finally to a single particle of an important size. In addition, Petot and al.12'1 have demonstrated a modification in the distribution of magnesium impurities (phenomenon of segregation) which is induced by a temperature gradient. By analogy, it is possible to think that in our study, the temperature gradient change the transport mechanisms of matter by a modification of impurities distribution. Other studies show the influence of the impurities location on the materials sinterability. Gouvea18 has studied the densification of Sn0 2 added Mn0 2 . He has shown the effect of foreigners cations distribution before sintering. If the foreigners cations are distributed on the surface of the grains Sn0 2 (in the first surface layers of grains) sintering begins at a lower temperature than in cases of foreigners cations initially distributed outside the Sn0 2 grains. The authors have proved that in the latter case, the densification begins only when a critical concentration in foreign element is present on the surface grains of the matrix. Below this critical value, they have demonstrated that only the surface of grains is concerned by diffusion. Beyond this limit the material densifies. The authors have concluded a modification of mass transport mechanism depending on the concentration of foreign elements on the grain surface. It results a densification delay if the foreigners elements remained outside the grains. The results and observations from the present study suggest that the thermal gradient can change the surface diffusion of alumina grains according to impurities distribution within the alumina grains or in the periphery ofthat grain. For pure alumina, without intergranular impurities, a strong reactivity of the system is maintain to a higher temperature. So, that leads to a better densification. For contamined alumina powder, the present impurities modify the kinetics of matter transport. So, that can lead to grain growth at densification's expense. CONCLUSION By this analysis several points can be mentioned and emphasized : - Comparisons between thermal treatments carried out in the initial stage of sintering (with and without thermal gradient) show that the gradient induces delay of specific surface area drop. In this case, the further densification of the material is improved. The interpretation of these experimental results due to correlate delay of specific surface area drop to a delay in the necks formation. This delay helps to maintain a strong reactivity of the system to a higher temperature and thus a better densification. - Through complementary experiments, it has been demonstrated that a thermal gradient applied in the initial stage of sintering can affect the impurities dissolution. The mass transport is then modified : the grain growth is promoted at the expense of densification. Optimising the co-doping and a thermal gradient is an interesting way to obtain porous ceramic with stable and controlled microstructures. REFERENCES 1 Z. Xie, J. Yang, Y. Huang, Densification and grain growth of alumina by microwave processing, Mat. Letters, 37 (4-5), 215-220 (1998).

48



2

J.H. Yang, K.W. Song, Y.W. Lee, J.H. Kim, K.W. Kang, K.S. Kim, Y.H. Jung, Microwave process for sintering of uranium dioxide, J. Nucl. Mat., 325 (2-3), 210-216 (2004). J.R. Groza, Consolidation of atomizaed NiAl powders by plasma activated sintering process, Scr. Mater., 30, 47-52 (1994). 4 Y. Zhou, K. Hirao, Y. Yamauchi and S. Kanzaki, Densification and grain growth in pulse electric current sintering of alumina, J Eur. Ceram. Soc, 24 (12), 3465-3470 (2004). 5 B. Klotz, K. Cho, RJ. Dowding and R.D. Jr. Sisson, Boron carbide consolidated by the plasma pressure compact (P2C) Method in air, CESP, 22, 4 (2001). K. Ozaki, K. Kobayashi, T. Nishio, A. Matsumoto and A. Suyiyama, Sintering phenomena on initial stage in pulsed current sintering, J. Jpn. Soc. Powder Powder Metall, 47, 293-297 (2000). 7 A. Zavaliangos, J. Zhang, M. Krammer and J. R. Groza, Temperature evolution during field activated sintering, Materials Science and Engineering A, 379 (1-2), 218-228, (2004). 8 B. Mc Williams, a. Zavaliangos, Temperature distribution and efficiency considerations for field activated sintering (FAST), presented at sintering'05, Grenoble, (2005). 9 P. Braudeau, Transport de matiére dans les oxydes - Influence d'un gradient thermique Approche du frittage rapide, These Université Paris VI, (1983). 10 A.W. Searcy, D. Beruto, Theory and experiments for isothermal and non isothermal sintering, Science of ceramics, 14, 1-13 (1987). n R. Botter, A.W. Searcy, Influence of Temperature Gradients on Sintering : Experimental Tests of a Theory, J. Am. Ceram. Soc, 72 (2), 232-235 (1989). 12 D. Monceau, C. Petot, G. Petot-Ervas, Kinetic demixing profile calculation under a temperature gradient in multi-component oxides, J. Eur. Ceram. Soc, 9, 193-204 (1992). 13 C. Petot, G. Petot-Ervas, M. Tebtoub, J.W. Fräser, MJ Graham, G.I. Sproule, Kinetic demixing in a-alumina during cooling : Influence of the powder reactivity, Solid State Ionics, 95 (1-2), 65-72 (1997). C. Greskovich, K.W. Lay, Grain growth in very porous AI2O3 compacts, J. Am. Ceram. Soc, 55 (3), 142-146(1972). 15 M.J. Readey, D.W. Readey, Sintering of ZrÜ2 in HC1 atmospheres, /. Am. Ceram. Soc, 69 (7), 580-582(1986). 16 M.J. Readey, D.W. Readey, Sintering T1O2 in HC1 atmospheres, J. Am. Ceram. Soc, 70 (12), 358-C361 (1987). 17 T. Quadir, D.W. Readey, Microstructure development of zinc oxide in hydrogen, J. Am. Ceram. Soc, 72 (2), 297-302 (1989). 18 D. Gouvea, A. Smith, J.P. Bonnet, J.A. Várela, Densification and coarsening of SnCVbased materials containing manganese oxide, J. Euro. Ceram. Soc, 18 (4), 345-351 (1998).


· 49


Fundamental Aspects of Sintering


EFFECTS OF CRYSTALLIZATION AND VITRIFICATION ON SINTERING PROPERTIES OF BENTONITE CLAY H. Camacho, C.A. Martinez, P.E. Garcia, H.J. Ochoa, and J.T. Elizalde Instituto de Ingeniería y Tecnología. Universidad Autónoma de Ciudad Juárez. Av. del Charro 610 norte, Ciudad Juárez, Chih., México 32310 A. García Interceramic Technological Center, Av. Carlos Pacheco No. 7200, Chihuahua, Chih., México 31060 A. Aguilar, M. Bocanegra, C. Domínguez Centro de Investigación en Materiales Avanzados, S.C. Miguel de Cervantes 120 Chihuahua, Chih., México 31109 ABSTRACT Bentonite clays are frequently used in the formulation of structural ceramics at low concentration in spite of its chemical composition which is similar to kaolinite, the most used clay for this purpose. During the study of the sintering properties of natural bentonite clay, we found that the studied bentonite can be used as main component to formulate structural ceramic products where control of dimension is fundamental. According to our study, with ceramic formulation based on bentonite clay, during thermal treatment a plateau where dimensional changes are minimal can be used to sinter the ceramic pieces. This process represents a significant advance related to the conventional process where CaC03 is used to promote the crystallization/vitrification process of several phases by reaction of amorphous silica and amorphous meta kaolinite with the CaO produced during the thermal decomposition of CaC03. INTRODUCTION Traditional ceramics generally concerns with the use of silicate-primary clay minerals and/or silicate glasses as raw material. Traditional ceramic industry have been spending great part of their efforts searching for proper clay minerals and due to requirements of production costs and exhausting of deposits under mining, this is still an issue.1"4 In ceramic industry, local minerals are usually the source for raw materials and they need to be studied before they can be incorporated to ceramic processes.5,6 Current models of ceramic sintering7 do not consider phase transformation during thermal treatment. For natural clay minerals, crystallization and vitrification is frequently a common process during ceramic processing to obtain ceramic materials.8"10 In the present work, the attention is focused on the study of sintering of a natural bentonite clay in order to characterize the processes of crystallization and vitrification during thermal treatment. This type of clay is known as SodiumCalcium Bentonite, it has a dimensional stability region between 950 - 1080 °C, the relation of this stability region with the chemical transformations is an issue of scientific interest. It has consequences for the traditional ceramic processing, because for some applications porous and light ceramic titles are of interest. In addition, this regional Bentonite clay offers an economic option. In this work, evolution of phase transformations is obtained mainly by XRD and SEM analysis. Two kind of dilatometric tests were made to study the sintering process, linear shrinkage and beam deflection. EXPERIMENTAL PROCEDURE Bentonite clay currently used to formulate ceramic bodies was taken from a mine located at 28° 57' 3 3 " lat N;105° 35'20" long W and 1269 m over sea level. Clay was milled in an alumina jar mill and then dried in order to eliminate free water. Elemental analysis was developed in a Bruker AXS S4-

53

Effects of Crystallization and Vitrification on Sintering Properties of Bentonite Clay

Pioneer Wavelength dispersive X-ray fluorescence spectrometer (WDXRF) and the samples preparation was made by the pressed pellets method, the elemental composition was obtained by means of standardless program. After drying the clay samples, alumina jar ball-mill and planetary equipment was used to wet mill the clay and drying to have the clay as a powder with particle size lower than 420 μπι. Next step was to humidify the clay in order to obtain standard conditions (6 % of water), and finally the wet powder was compacted by unidirectional pressing in steel mould in an automatic hydraulic press. For each experiment, clay discs of 5 cm of diameter and 0.6 mm of thickness were pressed in unidirectional way under 280 Kg cm" of load. Differential Thermal Analysis (DTA) and Thermo Gravimetric Analysis (TGA) were carried out to detect phase transformations in TA instrument DTA/TGA model Q 600 in static air atmosphere, heating rate 14 °/min from T=25 °C - 1180 °C, The calibration was made with STD of Indium, Zinc and Silver. These analyses were complemented using X-Ray Diffraction (XRD) with Copper radiation (λ = 1.540598Á) made to samples heated under different thermal treatments of 900, 950, 1000, 1050, 1100, 1150 and 1180 °C. XRD patterns were obtained in a PANanalytical X'Pert PRO Difractometer. As the main objective is to study sintering behavior, it is especially useful to quantify the crystal/amorphous ratio of phases generated during heating treatments. Hence, the Scanning Electron Microscope is used to characterize the amorphous and crystalline phases. Quantification of the amorphous phase is conducted by SEM image processing. SEM micrographs were obtained with a JEOL JSM-5800 LV. The TDA (Thermo dilatometric analysis), developed in a Misura-ODHT Model M3D1400/50, is applied to measure shrinkage. A green sample was prepared where the three dimensions were in the same order. Shrinkage, ha I a , is measured experimentally along the three axes to calculate the total relative volume shrinkage, Θ, which permits to follow the progress of sintering. Θ is given by 0=l-£ ^-[(l-^Kl-M/Ml-dc/q,)]

(1)

where Aa = ao - a; Ab and Ac are similarly defined. The parameters, Vo and V are the sample volume values at starting and at any time (t) corresponding to temperature T. Values are in the range 0< Θ < 1 and they correspond to shrinkage. If the mass remains constant, this shrinkage means densification. Finally, a beam was prepared to perform the bending creep test to measure the maximum deflection, δ. RESULTS AND DISCUSSION Table I shows the elemental analysis for the Bentonita herein studied showing that it may be call as a Sodium-Calcium Bentonite clay. Clay samples as obtained from mining were analyzed by DTA using a heating program from room temperature to 1180 C under heating rate of 14 C/min. Below 500 C a characteristic signal of endothermic process was observed, which can be associated with the evaporation of free and interlaminar water. Above 500 C a close sequence of endothermic and exothermic process suggest that several kinds of phase transformations are going on at the same time. One of these is the endothermic lost of chemically bonded water occurring between 650 C and 700 °C, which is typical for bentonitic clays. The typical pattern of crystallization as reported by Syam and Varma11 is not observed.

54



Table I. Elemental Analysis for the Sodium Calcium Bentonite Clay in percent for each element

10

20

30

40

50

60

2Θ

Figure. 1. X-ray diffraction patterns for untreated and thermally treated clay:(a) Raw Clay, (o) 900°C, (Δ) 950 °C, (V)1000 °C, (0) 1050 °C, («)1100 °C, (·)1150 °C, (A) 1180 °C. Phase transformations may be observed as result of the thermal treatment.


· 55


550·

Anorthite 21.97 500

■Ä

450

Π3 Φ Q.

400

> ω DT

350·

300

— i — ' — i — ■ — i — ' — i — ' — i — ' — i — ' — i —

900

950

1000

1050

1100

1150 1200

Temperature ( C) Figure. 2. Dependence of the intensity of Anorthite phase peak (2Θ= 21.97) on temperature of thermal treatment. Values were taken from Fig. 1 data. It can be inferred that for temperatures higher than 1100 C the volume fraction of Anorthite tends to decrease. Fig. 1 shows the X-ray diffraction patterns carried out for several temperatures from 950 C to 1180 C. Before any thermal treatment, besides montmorillonite, mainly detected phases are anorthite, albite, opal and quartz (CaAl 2 Si 2 0 8 , NaAlSi 3 0 8 , Si0 2 X H 2 0 and SiÓ2), being montmorillonite and quartz the main phases whose signals decrease significantly after thermal treatment. Albite or anhortite phase signal increased with the thermal treatment up to 1100 °C, above this temperature, rapidly tend to disappear, which clearly can be seen at figure 2. In the insert the sharpening of this peak is observed from 800 to 1150 °C, which indicates that crystals are growing. Above 1150 °C the dissolution of anorthite phase is indicated by the widening of peak and the lowering of intensity. Fig. 3 shows the SEM micrographs of several samples fired for 30 min at a fixed temperature from 900 °C to 1180 °C. For treatments of 900 °C to 1000 °C, the observed changes in microstructure are mainly related to sintering. For 1050 C the fraction of the amorphous phase starts to increase. For the sample treated at 1100 C, the amorphous phase becomes dominant and for higher temperatures, a lot of well defined crystals can be observed. According to XRD analyses (Fig. 1) anorthite, quartz and montmorillonte phases were dissolved to form vitreous phase. SEM image analysis is implemented to estimate the amorphous/crystal fraction. Fig. 4 shows the amorphous volume fraction as function of temperature. This fraction was calculated from Fig. 3a to 3g. At 800 C, sintering is activated and shrinkage is observed according to TDA curve (Fig. 4), which is probably due to the activation of the viscous flow in the amorphous matrix. For temperatures higher than 950 °C, the crystalline fraction tends to increase and shrinkage almost stops, thus the crystallization process increase the viscosity of the vitreous phase and as a result the sintering process is inhibited. The stabilization of volume can be qualified as unusual for most sintering materials and can be taken as an advantage, since this property can be manipulated for specific needs of the ceramic industry.

56



Fig. 3. SEM micrographs (Magnification: 2200 Voltage: 15 KV) of the Sodium-Calcium Bentonite Morrión fired at (a) 900°C, (b) 950°C, (c) 1000°C, (d) 1050°C, (e) 1100°C, (f) 1150°C, and (g) 1180°C during 30 min. The maximum for the amorphous volume fraction is observed for 1100 C.


· 57


For treatments at temperature higher than 1050 °C a tendency for the decrease of crystals phase can be observed and the sintering shrinkage is activated again. It is worthily to point out that, in this case, crystallization is related to sintering inhibition, this sintering mechanism is knowing as liquid reactive phase, since_diffusional sintering related to crystals structures occurs much slower than viscous sintering related to amorphous structures.'2'13 This sintering mechanism exhibits behavior differences in comparison to traditional mechanism of vitrification, which is well known for production of traditional ceramic bodies (wall, floor and porcelain tiles).14 The possibility to take control of shrinkage in an extended range of temperatures can be useful to produce ceramic bodies where minimum changes on dimensions are required, as in the porous ceramic body (wall tiles). At industry this kind of behavior is currently obtained using calcium carbonate for ceramic body formulations (silica, kaolin, ball clay, feldspar and calcium carbonate (CaCCh)) to promote the crystallization process of wollastonite, gehlenite and anorthite phases by reaction of amorphous silica and/or amorphous meta kaolinite with the CaO produced during the thermal decomposition of CaCCh giving thus a relatively gradual vitrification process. For the vitrification process promoted with CaCC>3, reaction is not fully achieved and many intermediary phases are presented at the end of the vitrification step, imparting some undesirables characteristics at the final product like water absorption expansion (time dependant products).14

Temperature ( C) Figure. 4. Dependence of the amorphous fraction (left axe) calculated from SEM images and relative shrinkage (right axe) on temperature for heating rate of 14 C/min determined by TDA. The connection between the amorphous volume fraction and the relative shrinkage may be observed.

58



CD

-i—'—i—'—i—'—i—'—r 1000 1050 1100 1150 1200

Temperature ( C) Figure. 5. Relative volume shrinkage (left axe) and beam deflection (right axe) as a function of temperature. About 1100 °C, the average slope for beam deflection shows a remarkable change. At this temperature, the Anorthite volume fraction start to drops and the maximum volume fraction of the amorphous phase is reached. Figure 5 shows the evolution of the Relative volume shrinkage together with Beam deflection with the temperature increase at 7 °C / min. Relative volume shrinkage and density may be related by p = p„ /(l - θ). Then, the free sintering density evolution for amorphous materials15 exhibits somehow a similar tendency to that shown in Fig. 5. Then, the effect of phase transformations on volume shrinkage is low. Beam deflection is more sensitive to crystallization and vitrification. For low temperatures up to 950, the tendency for the deflection is to increase. This tendency is interrupted and it appears again for temperatures higher than 1000 °C. It can be noticed that this tendency is close related to the amorphous fraction with temperature shown in Fig. 4 and to Relative volume shrinkage (Fig. 5). A combination of density evolution and Beam deflection may lead to estimate uniaxial viscosity1 . Uniaxial viscosity and density evolution depends on the viscosity of the sintering material15. Phase transformations play a key role on the matrix viscosity and it is expected to affect all the parameters used to describe sintering. CONCLUSIONS • The phase transformation of Sodium-Calcium Bentonite clay as function of thermal treatment is reveled. For lower temperatures, anorthite, opal and quartz are the identified phases. As temperature increases, vitrification takes place. • For this clay sintering is favored between 800 °C and 950 °C and for temperatures higher than 1100 °C. Between 950 °C and 1100 °C the dimensional changes are inhibited due to crystallization and the sintering mechanism, known as liquid reactive phase. • Beam deflection slope change (macroscopic constrained sintering) can be observed as a clear evidence of crystallization phases from bentonite clay. • Bentonite clay offers the possibility to control the crystal/amorphous fraction in ways that are beneficial to specific application in the ceramic industry.

Advances in Sintenng Science and Technology

· 59


ACKNOWLEDGEMENT This work was supported by CONACYT Chihuahua CHIH-2006-C02-58108. The support of the Interceramic Technological Center is also highly appreciated. REFERENCES 'C. Zhang, K. Komeya, J. Tatami, T. Meguro, and Y-B. Cheng, Synthesis of Mg-α SiAlON powders from talc and halloysite clay minerals. J. Eur. Ceram. Soc, 20, 1809 - 14 (2000). 2 A. J. Flynn, and Z. H. Stachurski, Microstructure and properties of stoneware clay bodies. Clay Minerals, 41, 775 - 89 (2006). 3 B. Bauluz, M. J. Mayayo, A. Yuste, C. Fernandez-Nieto, and J. M. Gonzalez-Lopez, TEM study of mineral transformations in fired carbonated clays: relevance to brick making. Clay Minerals, 39, 333 44 (2004). 4 K. Traoré, G.V. Ouédraogo, P. Blanchart, J.-P. Jernot, and M. Gomina, Influence of calcite on the microstructure and mechanical properties of pottery ceramics obtained from a kaolinite-rich clay from Burkina Faso. J. Eur. Ceram. Soc, 27, 1677-81 (2007). 5 S. Kacim, and M. Hajjaji, Firing transformations of a carbonatic clay from the High-Atlas, Morocco. Clay Minerals, 38, 3 6 1 - 6 5 (2003). 6 G. E. Christidis, P. Makri, and V. Perdikatsis, Influence of grinding on the structure and colour properties of talc, bentonite and calcite white fillers. Clay Minerals, 3 9 , 1 6 3 - 1 7 5 (2004). R. K. Bordia, and G. W. Scherer, On constrained Sintering-1. Constitutive Model for a sintering body. Acta Metall, 36,2393 - 2397 (1988). 8 M. Murat, and M. Driouche, Characterization of the crystallinity of silicates by dissolution conductimetry. J. Eur. Ceram. Soc, 6, 73 - 83 (1990). M. Roskosz, M. J. Toplis, and P. Richet, Kinetic vs. thermodynamic control of crystal nucleation and growth in molten silicates. J. Non-Crystal. Solids, 352, 180-84 (2006). °S. Chandrasekhar, and P. N. Pramada, Sintering behaviour of calcium exchanged low silica zeolites synthesized from kaolin. Ceram. Int., 27, 105 - 14 (2001). N. Syam Prasad, and K. B. R. Varma, Crystallization Kinetics of the LÍBO2-ND2O5 Glass Using Differential Thermal Analysis. J. Am. Ceram. Soc., 88, 357 - 61 (2005). 12 M. Reiterer, T. Kraft, and H. Riedel, Application of a microstructure-based model for sintering and creep. In Proc. 106th Annual Meeting of ACerS, ed. C. DiAntonio, Indianapolis, Ceram. Trans., 157, 49 - 58 (2004). 13 G.W. Scherer, and D.L. Bachman, Sintering of low-density glasses: II, Experimental study. J. Am. Ceram. Soc, 60, 239 - 43 (1977). '"Applied Ceramic Technology volume 1 SACMI, ISBN 88-88108-48-3, Edifice La Mandragora of Imola s.r.l, Chapter VIII, 245 - 54 (2002) 15 G. W. Scherer, Cell Models for Viscous Sintering. J. Am. Ceram. Soc, 74, 1523 - 31 (1991). 16 S-H. Lee, G.L. Messing, and D. J. Green, Bending creep test to measure the viscosity of porous materials during sintering. J. Am. Ceram. Soc, 86, 877 - 82 (2003).

60


DISSOLUTION OF ALUMINA JN SILICATE GLASSES AND THE GLASS FORMATION BOUNDARY Keith J. DeCarlo, Thomas F. Lam, and William M. Carty Kazuo Inamori School of Engineering, Alfred University Alfred, New York, USA ABSTRACT Chemical impurities typically segregate to the grain boundary in oxide ceramics and it is proposed that the grain boundary chemistry is dictated by the glass formation boundary in the system. For this to be possible it is necessary that the matrix grains readily dissolve into the grain boundary during liquid-phase sintering. Alumina is a model material for this study because the impurities are typical alkali and alkaline earth oxides plus silica and the glass formation ability of alumino-silicates is well known. This study examined the dissolution of alumina into a silicate glass through the use of glass-alumina diffusion couples and chemical mapping via WDS. The diffusion couples were heat treated to different temperatures, held for various times, then quenched. Samples were sectioned and polished for WDS analysis. The data shows alumina dissolution rate is rapid and that resulting chemistry of the diffusion couple interface was consistent with the glass formation boundary in the system studied. INTRODUCTION Solid state sintering is typically proposed to occur at approximately at 0.8 of the melting temperature (K). Alumina, however, has been observed to sinter to high densities at temperatures significantly lower (0.67-0.72) than those required for solid-state sintering and without magnesia additions.1' Due to these observations, it is proposed that many oxide systems, and in particular, refractory oxides such as alumina, typically sinter via liquid-phase mechanisms. This can only be true if alumina dissolves into the grain boundary phase efficiently. The dissolution rates of refractory solid oxides in a melt depend on three phenomena: 1) the solubility of the oxide in the melt, 2) the mobility of the reacting species within the melt, and 3) the mobility of the dissolved ion in the melt.3 The solubility of the refractory oxide in the melt is directly proportional to the mobility of the dissolved species. The rate-limiting step of a heterogeneous dissolution process is cither the introduction of the reacting species to the melt ("reaction-rate controlled") or the removal of the dissolved products from the interfacial region ("dissolution controlled"). The process of dissolution of alumina into the grain boundary phase is hypothesized to be "diffusion-controlled". Diffusion is modeled via Fick's second law:5 ^ = DS-± (1) a ' Sx1 where D, is the diffusion coefficient of the component in question, c¡ is the concentration of the diffusing component, and x is the direction of material flow. Application of Fick's second law frequently involves the error function. In the case of diffusion couples, the error function cannot be applied because the interdiffusion coefficient (interdiffusivity) of the system is not constant with changing concentration. The interdiffusivity takes into account the diffusivities of each diffusing component, which is known as the Darken relationship:6 D = NJ)I+NI~D;

(2)

61


where D is the interdiffusivity of component 1 and 2, /V,is the mole fraction of component /, and D, is the intrinsic diffusivity of component /. The interdiffusivity from experimental data may be determined from a graphical technique commonly referred to as the Boltzmann-Matano analysis:5' '8

^-¿(f)>

(3)

The Boltzmann-Matano solution can be applied to a diffusion couple to calculate the interdiffusivity of the components. From the interdiffusivity it is possible to calculate the activation energy, Q, of the diffusion process. The relationship between diffusivity and temperature can be understood by the subsequent empirical equation:

D = DaexV[-j^

(4)

D0 and Q vary with material composition but are independent of temperature. The values of these two variables are obtained graphically by plotting the natural logarithm of interdiffusivity versus inverse temperature, yielding: lnD = lnD - — ( - ]

" R{T)

(5)

the slope is -Q/R and the x-intercept is In D„ (and R is the gas constant). Dissolution of alumina into the grain boundary during sintering also accounts for different grain boundary chemistries observed with sintering temperature.9 In order for the grain boundary chemistry to change during sintering, alumina must dissolve into the grain boundary efficiently. The original grain boundary chemistry typically consists only of the impurities within the system. The differences in impurity ion size and charge limit impurity solubility in the alumina lattice (or refractory oxide). If the working hypothesis is correct, as the system is heated, eutectic melts form first at low temperatures in the grain boundaries. The composition of the eutectic melt formed at the grain boundaries is dependent on the initial impurity ions (R+, R2+, Si4+, etc.). Upon further heating, the impurities not melted due to compositional restrictions are incorporated into the liquid phase. At this stage of sintering, the alumina grains can either be further dissolved or some of the dissolved alumina within the grain boundary melt can be precipitated based on the reaction paths predicted by phase equilibria. Once the impurity phase is completely melted, it is hypothesized that alumina continues to dissolve into the liquid following the composition path toward increasing alumina concentration, as illustrated schematically in Figure 1. The dissolution process of the grain into the grain boundary dictates the chemistry of the grain boundaries but is ultimately controlled by the glass formation boundary (GFB). If the final composition of the grain boundary phase lies within the glass formation region, the grain boundaries will be amorphous. In contrast, if the final composition of the grain boundary is outside the glass formation region crystals will precipitate in the grain boundary. It is hypothesized that when the chemistry of the grain boundary lies outside the glass formation region, the liquid phase from which the crystals precipitate will always be at the saturation limit, with respect to the GFB. If a crystalline phase incorporating Al + precipitates in the grain boundary liquid, the crystals will serve as a "sink" and grain dissolution will continue.

62



Figure 1. The proposed dissolution path with increasing temperature (arrow) of AI2O3 into a grain boundary that contains S1O2 and CaO at a ratio of 1.4:1.0 (Si02:CaO). EXPERIMENTAL Diffusion couples were fabricated using a coarse-grained fused-cast alumina substrate (Monofrax'' M, Monofrax, Falconer, NY) and a commercial soda-lime-silicate glass sphere (Industrial Tectonics Inc., Dexter, MI). The fused-cast refractory substrate nominally consisted of 40% 01-AI2O3 and 60% β-Α1203. The typical chemistries are shown in Table I. Table I. Chemical Make-up of Diffusion Couple Materials Constituent AI2O3 Na 2 0 S1O2 CaO MgO Other

Monofrax M (weight %) 94 4.0 1.0 N/A N/A 5% for GP-Na after 200°C but remain constant as in the case of GP-K up to 940°C. The second and the large shrinkage is between 900°C and 1000°C and should be linked to the structural densification by viscous sintering.

o

-5

ε. -io

é W

-15

-20

-25

200

400

600 800 Temperature (°C)

1000

1200

1400

Figure 2: Dilatomtry curves of GP-Na and GP-K samples The XRD graphs (Figure 3) indicate that geopolymer materials are essentially amorphous. At 200°C there is no change in the phases present in different types of geopolymer. This is demonstrated by the fact apart from water loss (Figure 1), the structure of geopolymer is not affected by the temperature development up to 800°C. At 800°C the crystalline peaks already present increases in intensity for both types of geopolymers. Even at this temperature, it can be observed that potassium based geopolymer tend to develop more crystalline phases. The K form more complex crystalline phases as kaliophilite and leucite for K-geopolymer.


· 95


10

20

30

40

50

60

70

2teta (degrees)

Figure 3a): Phase evolution in Na-geopolymer with temperature (25, 200 at 800°C)

Figure 3b): Phase evolution in K-geopolymer with temperature (25,200 at 800°C) As it can be observed in Figure 1, there is metakaolin in the matrixes that did not react and which will be transformed to mullite from 950CC. This is follow by formation of liquid phase responsible for the increase in mechanical properties, density and the reduction of porosity. For above reason geopolymer materials are indicated for thermal applications in the range of temperature under 950°C.

96



Figure 4 a: SEM micrographs of geopolymer materials at 200°C (a: GP-Na, b: GPK) and 800°C (c: GP-Na, d: GP-K).

Figure 4 b: SEM micrographs of geopolymer material showing transformations at 1300°C due to the intensives liquid phase developed. In the Figure 4 it can be observed that geopolymers maintain their homogeneous and dense microstructure up 900°C but above this temperature as confirmed by dilatometry analysis, the geopolymer materials shrink considerably and surely can not maintain their high temperature properties. At 1100°C a vitrified phase was observed and at 1300°C only K-geopolymer maintain


· 97


their structure even with vitrification (Figure 4). Na-geopolymer at this temperature shows fissures and cracks. CONCLUSION The sintering behavior of the geopolymer materials were studied using DTA/TGA, DSC and optical dilatometry. The different observations were correlated to XRD and SEM observations. Geopolymers are mostly influenced by the composition of the alkali activating solution. It can be observed from the results that the sintering behavior of the geopolymer materials is also influenced by the nature of alkali activator used. The K-geopolymers presented better resistance to shrinkage, vitrification and deformation compared to Na-geopolymers. Microstructural and mineralogical investigations showed that thermal shrinkage relates to the densification by reduction of porosity during dehydroxylation and sintering. The sintering of geopolymer can then be structured with sequences of dehydration, dehydroxylation, densification and deformation. ACKNOWLEDGMENT We are particularly grateful to dr. Braga Mirko, R.&D. Laboratory, INGESSIL S.r.l., (Verona, Italy) for providing with the potassium silicate solution. REFERENCES 1. E. M. Levin, C. R. Robbins, and H.F. McMurdie, Fig. No. 407 in phase Diagrams for ceramics, Vol. 1. Edited by M. K. Reser. American Ceramic Society, Columbus, OH, 1964. 2. S.M. Johnson, J.A. Pask, and J.S. Moya, Influence of impurities on High-Temperature Reactions of Kaolinite, J. Am. Ceram. Soc. 65[1]31-35(1982). 3. L. Weng, K. Sagoe-Crentsil, T. Brown, Speciation and hydrolysis kinetics of aluminates in inorganic polymer systems, presented to geopolymer, International Conference on geopolymers 2829 October, Melbourne, Austrialia (2002). 4. M.R. Anseau, J.P. Leung, N.Sahai, T.W. Swaddle, Interactions of silicate ions with Zinc(II) and Aluminium(III) in alkali aqueous solution, Inorg. Chem., 44(22) 8023-8032 (2005). 5. M.R. North, T.W. Swaddle, Kinetics of silicate exchange in alkaline alumino-silicate solutions, Inorg. Chem., 39(12) 2661-2665 (2000).

98



6. J.S.J. Davidovits, Long term durability of hazardous toxic and nuclear waste disposals. In: Davidovits, J.S.J., Orlinski, J. (Eds.), Proceedings of the 1st International Conference on Geopolymers, Vol 1, Compiege, France, 1-3 June, PP. 125-134 (1988). 7. J. W. Phair, J.S.J.Van Deventer, Effect of the silicate activator pH on the microstructural characteristics of waste-based geopolymers. Intl. J. Min. Processing, 66 (1-4) 121-143 (2002). 8. J. A. Kostuch, G. V. Walters, T. R. Jones, High performance concrete containing metakaolin-A review. In: Dhir R.K., Jones M.R (Eds.), Proceedings of the Concrete 2000 International Conference on Economic and Durable Concrete through Excellence. University of Dundee, Scotland, UK, 7-9 September, 2, pp. 1799-1811 (2000). 9. P. Duxson, G.C. Lukey, J.S.J. Van Deventer, Physical evaluation of Na-geopolymer derived from metakaolin up to 1000°C, J. Matls Sei, 42 3044-3054 (2007).


· 99


Modeling of Sintering at Multiple Scales


MESO-SCALE MONTE CARLO SINTERING SIMULATION WITH ANISOTROPIC GRAIN GROWTH Gordon Brown1 , Richard Levine1, Veena Tikare2, & Eugene Olevsky1; '-San Diego State University, 2-Sandia National Laboratory ABSTRACT Although Monte Carlo (MC) simulations are widely used for understanding the microstructural evolution of sintering bodies, the models currently in use do not accommodate materials with anisotropic properties. Anisotropy has a significant effect on grain growth rates, which impacts material properties. Computer simulation models are used to better understand phenomena associated with sintering. A two-dimensional algorithm to simulate the evolution of granular structure with anisotropic materials using a Potts MC model, which incorporates the sintering mechanisms of grain growth, pore migration and vacancy annihilation, is presented. Limitations of this algorithm imposed by the underlying lattice structure are identified and analyzed. Solutions to mitigate these artifacts are proposed and implemented. Results are discussed and evaluated. The ability to incorporate anisotropic grain growth in our meso-scale modeling allows the investigation of anisotropic granular development under several different situations to better understand some of the observed anisotropic phenomena like patterning in sintered materials. INTRODUCTION Monte Carlo (MC) models have been used by many researchers for the simulation of grain growth and sintering 122 . The most common is the Potts Model 23 , which is an extension of the Ising 24 Model . This model discretizes the initial grain structure onto a lattice, and assigns a state to each site on the lattice. This state is assumed to be constant and uniform on a cell surrounding the site. Contiguous sites with the same grain state are considered to be parts of the same grain. Grain boundaries are then implicitly defined as existing between neighboring sites with different grain states. This discrete representation of the grain structure is a reasonable approximation of the actual morphology provided that the size of the lattice spacing is small compared to the grain sizes. Later in our discussion, it will be important to keep in mind that these states of the lattice sites are just discrete approximations of the actual material structure. The dynamic evolution of this discrete grain structure is then simulated by using a Kinetic Monte Carlo (KMC) approach 25 with the Potts model. This approach assumes a quasi-equilibrium state where many molecular exchanges occur shifting back and forth without affecting the overall distribution of states on a time scale of the atomic vibrations. Then on a much longer time scale, an infrequent event occurs which shifts the state configuration to a new quasi-equilibrium state. It is these infrequent events which are simulated in this Potts model so that the dynamic evolution of the structure will be captured. This way we can track the changes that occur in the distributions of these quasi-equilibrium states. As we are only looking at one site changing state at a time, in order to simulate all sites changing simultaneously, one time step consists of a cycle through every site. This time unit is called a Monte Carlo Step (MCS) of the simulation process.

103

Meso-Scale Monte Carlo Sintering Simulation with Anisotropie Grain Growth

THE POTTS MODEL The Potts model represents a sample of the material and thus it is a sampling from the distribution of configurations throughout the material. A fundamental result from statistical mechanics is that this distribution of configurations follows the Boltzmann distribution.

p(Q = ±e-E(c),k-T

Eq.(l)

where

/7(C) is the probability of configuration C of the state space, Z = ^ e ~ £ < c ) / * ' r i s the normalizing constant, c E(C) is the energy associated with configuration C, kB is the Boltzmann constant, and T is the temperature. The most widely used algorithm for sampling from this distribution is the Metropolis 26 algorithm . This algorithm first assumes that the energy associated with a particular site is only dependent on the state ofthat site and the states of the surrounding interacting "neighbor" sites. From this, the total energy of a configuration can then be computed as the sum of the energies associated with each site. The next assumption is that we can sample from the distribution of configurations by just looking at transitions associated with a single site and then do this repeatedly for all sites in the lattice. These conditions allow us to compute the probability that an existing configuration will transition into a new configuration. Under the Metropolis algorithm, the transition probability for such a change of state is P = {e-w,k*T if Δ£>0; 1 otherwise}

Eq. (2)

where AE is the change in energy from the existing state to the candidate state. There are various ways of calculating these energies, and the resulting microstructural evolution is determined by how these energies are calculated and the transitions under consideration. The simplest assumption for calculating these energies is to assume that the material is isotropic so that the energies associated with the states of neighboring sites are not affected by direction or orientation. Even with this assumption, there are considerations for variations in the boundary energies resulting from the anisotropy of the lattice. This was investigated by Holm et al.27 who found that using different boundary energies for first and second nearest neighbors of square or triangular lattices could result in lattice pinning and only using the first nearest neighbors also resulted in lattice pinning due to the larger anisotropy of the lattice. As a result, using the first and second nearest neighbors with the same boundary energies minimized the pinning effects of the lattice anisotropy. For this situation, the Hamiltonian for a transition associated with a site, /, is

//^/¿(l-^,,?,))

Eq.(3)

where H¡ is the Hamiltonian associated with lattice site /, J is the boundary energy with a boundary between any two neighboring sites q¡ is the state of site ¡,

104


Meso-Scale Monte Carlo Sintering Simulation with Anisotropic Grain Growth

q¡ is the state of one of the neighboring sites on the lattice adjacent to site i, n is the number of neighboring sites considered, and S(q¡,q¡) is the standard Kroenecker delta function {1 if q¡=qf, 0 otherwise). The boundary conditions for this model are determined by the need to determine shrinkage and thus have a boundary for constant mass. This means that we will have a buffer of void sites surrounding the simulation cell. As vacancies are annihilated and pores sites migrate to this buffer, the change in density can be seen visually and easily measured. Using this expression for the Hamiltonian with constant boundary energies implies isothermal, isotropic materials. It is for this reason that although the Potts model has been widely used for simulating the dynamic evolution of the meso-scale morphology, most implementations assume isotropic materials. Unfortunately, many of the materials we are trying to simulate are anisotropic. In order to understand the impact of the material anisotropy on the evolution of the grain morphology, the anisotropy must be incorporated into our model. PREVIOUS ANISOTROPIC MODELS Several researchers have developed methods to incorporate anisotropy into the Potts model " ' 21, 27-31 j ^ e m o s t c o m m o n a p p r o a c ¿ ¡ s to incorporate the material anisotropy through various modifications of the energy calculations associated with the states of neighboring sites. For these models, the boundary energy can be different for each of the neighbors, so the Hamiltonian of equation (3) becomes

^ÉW1-«^..?,)))

Ε

ι·(4>

where Jtj is the boundary energy between sites i and_/. In 2000, Morhac and Morhacova " presented an algorithm where a percentage of the grains had anisotropy introduced through one of four directions in a square lattice. The boundary energies were at a fixed ratio for each of these four directions, and were randomly assigned to the anisotropic grains. The result is shown in figure (1).

Fig. 1: Results from Morhac & Morhacova ".

Several researchers including Holm et al. (2001) 28, Grest et al. (1985) 29, and Yu and Esche 30 (2002) have approached this issue by looking at boundary energy and mobilities. Boundary energy is calculated using Read-Shockley theory and the misorientation angles across the grain boundaries. Their results on a triangular lattice are shown in figure (2). The two images shown are for the isotropic case (a) and the anisotropic case with the anisotropy parameter set to 60% of the maximum

Advances In Sintering Science and Technology

· 105


value (b). Additionally, Upmanyu et al. 31 evaluated the relative impact of boundary energy and boundary mobility and found that the boundary energy was the dominant factor in anisotropic microstructural evolution.

Fig. 2: Results from Yu & Esche

with anisotropy parameter at 0 (a) and 60% (b).

A different approach was taken by Yang et al. in 1995 21 by using Wulff plots to capture the surface energy anisotropy. An example of the model results from this approach is shown in figure (3). The Wulff plots can take a variety of complex shapes, and can be modified for specific applications.

Fig. 3: Results from Yang et al ' on triangular lattice.

Although these models show anisotropic grain growth, none of them incorporate the mechanisms of pore migration and vacancy annihilation necessary to simulate sintering. Additionally, they are all limited to a small number of angles for the anisotropic grains. The Wulff plot approach will be used in this paper as the function for the plot can be evaluated at any angle and can be modified in the coded software to incorporate other mechanisms affecting the boundary energy in future work. THE ANISOTROPIC MODEL As the goal in this paper is to simulate the mesoscopic morphology of a sintered material, we will modify Yang's model to include porosity and add the mechanisms of pore migration and vacancy annihilation into the model as presented by Tikare, et al. . Additionally, the orientation angles will not be limited as in Yang's model. For simplicity, an ellipse will be used for the initial implementation of an anisotropic model, and the ratio of the axes (a/b) will be referred to as the "aspect ratio" of the ellipse. The magnitude of the surface energy in any direction will be the distance from the origin to the Wulff plot function in that direction as shown in figure (4).

106



Fig. 4: Elliptical Wulff Plot

With this structure in place, the boundary energy (J¡¡) of equation (4) can be computed for any adjacent sites. Now for any two configurations, we can compute the change in energy (ΔΕ) and thus the probability of acceptance of the new candidate configuration using equation (2). As in other models, a grain growth step considers a single site and a transition of the state of that site from its current state to one of the states of a neighbor site. For a pore migration step, two neighboring sites are selected, where one is a pore and the other is a grain site. In this case, since there are two sites potentially changing state at the same time, all neighbor sites of both of these sites must be considered. Vacancy annihilation is incorporated into the model using the jump algorithm of Tikare et al. of 2003 32 and extended by Braginsky in 2005 33. All of these algorithms conserve the number of grain sites so by using a boundary around the structure the shrinkage of the simulated compact can be quantified. The initial configuration can be a bitmap transcribed from a micrograph, but for this research, a starting configuration is generated from random orientation assignments to all grain sites and then allowing the grain growth algorithm to proceed until a large enough average grain size is achieved. SIMULATION RESULTS To see the impact of the anisotropy in the Wulff plots, results are shown in figure (5) with all orientations aligned, time stopped at 40 MCS and the aspect ratio of the ellipse set to 1 (isotropic), 2, and 5. We can see that the increased anisotropy in the Wulff plots results in increased anisotropy in the grain growth and that with increased anisotropy, the average grain size is smaller. As all of these simulation results are for the same time (40 MCS) we may conclude that the grains are growing slower with increased anisotropy.

Fig. 5: Wulff Plot Anisotropy Ratios 1 (isotropic), 2, & 5

This result is consistent with the results of other models. To verify the impact of the orientation angle of aligned grains, results shown in figure (6) have the same time in MCS and the same aspect ratio for the Wulff plot ellipse, but the orientation angles are 0, 90, and 135 degrees in the CCW direction. The results are as expected for the angles.


· 107


Fig. 6: Orientation Angles CCW 0 (vertical), 90, & 135 Degrees

Additionally, the diagonal directions appear to have more anisotropy. This is an artifact which results from the square lattice as the eight neighbors are all treated as equidistant, but in fact the diagonal neighbors are farther away than the vertical and horizontal neighbors. Now looking at the results with random orientations for various times (40, 100, and 200 MCS) it is apparent that even though the initial orientation angles are uniformly distributed, the grains only seem to grow in the directions of nearest neighbors.

Fig. 7: Random Orientations at MCS of 40, 100, & 200

This is an artifact induced by the lattice as there are no neighbors in the other directions. So how can this artifact be eliminated? A new novel approach is to rotate the lattice during the microstructural evolution so that there will be neighbors in other directions. LATTICE ROTATION There are several issues to overcome with rotating the lattice. First it is important to realize that any rotation other than multiples of 90 degrees will result in distortion as the original lattice sites will not coincide with the new lattice sites. Some distortion is acceptable as the states are just discrete approximations as we discussed earlier in this paper. Secondly, when the square lattice is rotated, some sites will no longer be in the domain of the lattice. This problem can be overcome by having a larger square lattice with only active sites in a circular region inscribed in the square. Then if we want to have a square specimen, it must be fully contained in the circular active region. With this in mind, the implementation in this paper will use a circular specimen to see the maximum use of the active region of the lattice. As the specimen is rotated, the distortion mentioned earlier will cause a shift of some neighbor sites so that they are no longer neighbors. As the diagonal neighbors are only contiguous at a corner point any amount of distortion can make them no longer neighbors. In fact, some shifting of neighbors is necessary to capture the other angles; we just wish to avoid excessive distortion. We want to

108



maintain the number of grain sites and the grain size. To do this, the rotation algorithm must map the grain sites of each grain to the same number of sites after the rotation. This means that some of the sites of a grain which are neighbors before the rotation will not be neighbors after the rotation, but of course if the structure is rotated back to the original grid, the sites will be neighbors again. This means that the algorithm must have the same effect regardless of direction (CW or CCW). Additionally, this algorithm must be able to be extended to three dimensions. As the lattice geometry of a cubic lattice in the third dimension is topologically the same as the two dimensional square lattice, the algorithm is focused on the square lattice and is directly extendable to a third dimension by using two angles, azimuth and altitude. An algorithm to do this is to divide the lattice into concentric rings. To rotate the lattice, rotate the sites within these rings through the desired angle of rotation. The resulting distortion is of two types. One comes from the fact that the sites in the ring vary in radii within the ring. This distortion is reduced with smaller thicknesses of the rings. Of course as the thickness of the rings decreases, the density of sites in the rings decreases as well. This means that there will be more angular distortion. Nonetheless, this trade off is natural. As an illustration of this rotation algorithm, consider a circular specimen obtained from a sample micrograph with aligned grains so that the rotation angle is easier to visualize. Each lattice site is shown as a square, and the purpose of this figure is to demonstrate that the algorithm rotates the structure and that the grain and pore boundaries have some distortion as the structure rotates. Note that as the pores get smaller they approach the size of the square lattice site. Once they become the size of a single site, if on a grain boundary they are considered a vacancy and can be annihilated.

Fig. 8: Rotation of Grains with Circular boundary for angles of 0, 5, and 35 degrees.

The maximum distortion occurs at 45 degrees and is symmetric so that at 90 degree intervals there is no distortion at all as the ring sites line up exactly with similar sites in the rings. CONCLUSIONS The Potts model of Yang has been extended to include porosity and the mechanisms of pore migration and vacancy annihilation to simulate sintering. The lattice discretization causes directional artifacts. Rotation of the lattice is a novel approach to overcome these artifacts, but requires an algorithm for the rotation. Lattice rotation will result in some distortion. Some of distortion is necessary to remove the artifacts, but excessive distortion should be avoided. The proposed algorithm seems to meet the requirements for our concept and has a parameter (ring thickness) which can control the radial and angular distortion to find the best balance. This will be incorporated into the Potts model described above and results will be reported in a future paper. ACKNOWLEDGEMENTS The support of National Science Foundation Divisions of Materials Research (Grant DMR0705914) and of Civil and Mechanical Systems and Manufacturing Innovations (Grants DMI-0354857

Advances In Sintering Science and Technology

· 109


and CMMI-0758232) is gratefully appreciated. Sandia is a multi-program laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy National Nuclear Security Administration under the Contract DE-AC04-94AL-85000. REFERENCES 'Aldazabal, J., A. Martin-Meizoso, et al. "Simulation of liquid phase sintering using the Monte Carlo method." Materials Science and Engineering a-Structural Materials Properties Microstructure and Processing 365(1-2): 151-155. (2004). 2 Blikstein, P. "Monte Carlo Simulation of Grain G." Materials Research 2(3): 5. (1999). 3 Bordere, S. "The impact of fluctuations on the sintering kinetics of two particles demonstrated through Monte Carlo simulation." Scripte Materialia 55(10): 879-882. (2006). "Dudek, M. R., J. F. Gouyet, et al. "Q+l state Potts model of late stage sintering." Surface Science 401(2): 220-226. (1998). 5 Han, Y. S. and D. K. Kim "Monte Carlo simulation of anisotropic grain growth in liquid phase sintering." Journal of the Korean Physical Society 42: S1058-S1062. (2003). 'Hassold, G. N., I. W. Chen, et al. "MONTE-CARLO SIMULATION OF SINTERING." Journal of Metals 40(7): A44-A44. (1988). 'Huang, C. M., C. L. Joanne, et al. "Monte Carlo simulation of grain growth in polycrystalline materials." Applied Surface Science 252(11): 3997-4002. (2006). 8 Kim, Y. J., S. K. Hwang, et al. "Three-dimensional Monte-Carlo simulation of grain growth using triangular lattice." Materials Science and Engineering a-Structural Materials Properties Microstructure and Processing 408(1-2): 110-120. (2005). 'Li, H., G. H. Wang, et al. "Monte Carlo simulation of three-dimensional polycrystalline material." Materials Science and Engineering a-Structural Materials Properties Microstructure and Processing 357(1-2): 153-158. (2003). '"Luque, A., J. Aldazabal, et al. "Simulation of the microstructural evolution during liquid phase sintering using a geometrical Monte Carlo model." Modelling and Simulation in Materials Science and Engineering 13(7): 1057-1070. (2005). "Morhac, M. and E. Morhacova "Monte Carlo simulation algorithms of grain growth in polycrystalline materials." Crystal Research and Technology 35(1): 117-128. 12 01evsky, E. A., B. Kushnarev, et al. (2005). "Modelling of anisotropic sintering in crystalline ceramics." Philosophical Magazine 85(19): 2123-2146. (2000). "Potter, B. G., V. Tikare, et al. "Monte Carlo simulation of ferroelectric domain structure and applied field response in two dimensions." Journal of Applied Physics 87(9): 4415-4424. (2000). '"Radhakrishnan, B., G. B. Sarma, et al. "Modeling the kinetics and microstructural evolution during static recrystallization - Monte Carlo simulation of recrystallization." Acta Materialia 46(12): 44154433. (1998). ,5 Sault, A. G. and V. Tikare "A new Monte Carlo model for supported-catalyst sintering." Journal of Catalysis 211(1): 19-32. (2002). "Schmid, H. J., S. Tejwani, et al. "Monte Carlo simulation of aggregate morphology for simultaneous coagulation and sintering." Journal of Nanoparticle Research 6(6): 613-626. (2004). "Srolovitz, D. J., M. P. Anderson, et al. "MONTE-CARLO SIMULATION OF GRAIN-GROWTH." Journal of Metals 35(8): A60-A60. (1983). 'Teixeira, A. and R. Giudici "A Monte Carlo model for the sintering of NÍ/A1203 catalysts." Chemical Engineering Science 56(3): 789-798. (2001). "Tikare, V., M. A. Miodownik, et al. "Three-dimensional simulation of grain growth in the presence of mobile pores." Journal of the American Ceramic Society 84(6): 1379-1385. (2001). M Yamashita, T., T. Uehara, et al. "Multi-Layered Potts Model simulation of morphological changes of the neck during sintering in Cu-Ni system." Materials Transactions 46(1): 88-93. (2005).

110



2I

Yang, W., L. Q. Chen, et al. "COMPUTER-SIMULATION OF ANISOTROPIC GRAINGROWTH." Materials Science and Engineering a-Structural Materials Properties Microstructure and Processing 195(1-2): 179-187. (1995). 22 Yu, Q. and S. K. Esche "Three-dimensional grain growth modeling with a Monte Carlo algorithm." Materials Letters 57(30): 4622-4626. (2003). "Potts, R. B. "SOME GENERALIZED ORDER-DISORDER TRANSFORMATIONS." Proceedings of the Cambridge Philosophical Society 48(1): 106-109.(1952). "Ising, E. "Beitrag zur Theorie des Ferromagnetismus"; Zeitschrift für Physik. 31, 253-258. (Engl. Translation: http://www.fhaugsburg.de/~harsch/anglica/Chronologv/20thC/Ising/isi fm00.html) (1925). 25 Voter, A. F. "Introduction to the Kinetic Monte Carlo Method". IPAM 2005, UCLA. (2005). 26 Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H., and Teller, E. "Equation of state calculations by fast computing machines"; Journal of Chemical Physics 21;1087-1092. (1953). 27 Holm, E. A., Glazier, J. A., Srolovitz, D. J., Grest, G. S., "Effects of lattice anisotropy and temperature on domain growth in the two-dimensional Potts model", Physical Review A, 43(6): 26622668. (1991). 2, Tikare, V., Braginski, M. , et al. "Numerical Simulation of Anisotropie Shrinkage in a 2D Compact of Elongated Particles." Journal of the American Ceramic Society 88(1): 59-65. (2005). "Holm, E. A., G. N. Hassold, et al. "On Misorientation Distribution Evolution During Anisotropie Grain Growth" Acta Materiallia 49: 2981-2991. (2001) 2, Grest, G. S., D. J. Srolovitz, et al. "Computer simulation of grain growth. IV. Anisotropie grain boundary energies"; Acta Metallurgica. 33(3): 509-520. (1985). M Yu, Q., S. K. Esche "Modeling of grain growth kinetics with Read-Shockley grain boundary energy by a modified Monte Carlo algorithm"; Materials Letters, 56, (1): 47-52. (2002) 3 'Upmanyu, M., G.N. Hassold, et al. "Boundary Mobility and Energy Anisotropy Effects on Microstructural Evolution During Grain Growth"; Interface Science 10, (2-3): 201-216. (2002) 5! Tikare, V., M. Braginsky, et al. "Numerical Simulation of Solid-State Sintering: I. Sintering of Three Particles." Journal of the American Ceramic Society. 86, (1): 49-53. (2003). "Braginsky, M., V. Tikare, et al. "Numerical Simulation of Solid State Sintering." Intl Journal, of Solids and Structures 42: 621-636. (2005).


· 111


NUMERICAL SIMULATION OF DENSIFICATION AND SHAPE DISTORTION OF POROUS BODIES IN A GRANULAR-TRANSMITTING MEDIUM Junkun Maa, Eugene A. 01evskyb "Southeastern Louisiana University, SLU 10847, Hammond, LA 70402, USA b San Diego State University, 5500 Campanile Drive, San Diego, CA 92182, USA ABSTRACT In an attempt to achieve maximum densification and minimum distortion for near-net-shaped manufacturing of powder-based components, the process of quasi-isostatic pressing (QIP) is studied. In QIP, the self-propagating high-temperature synthesis (SHS) and the subsequent consolidation are carried out in a rigid die where a powder body subjected to SHS is surrounded by a granular pressuretransmitting medium (PTM). Following the SHS, a uniaxial load is applied to the PTM, which builds up a quasi-isostatic stress in the die. The constitutive behaviors of both the post-SHS body and the PTM are determined based on theoretical and experimental studies. The densification and shape distortion of a post-SHS powder specimen are simulated using a finite element model and the results are compared to the experimental data with satisfactory agreement. To obtain the desired final shape, the optimized initial shape of the pre-QIP porous body is determined based on a special iterative simulation approach. INTRODUCTION Self-propagating High-temperature Synthesis (SHS) has been successfully used for the fabrication of numerous advanced industrial materials since the beginning of the large-scale systemic investigation of this process in 1960s1"7. Following the SHS reaction, the consolidation of a porous specimen by applying compressive mechanical stresses is an important processing step utilized for the fabrication of final products with high relative density. Obviously, during mechanical consolidation, both the volume and shape of a post-SHS porous specimen change. Ideally, Cold or Hot Isostatic Pressing (CIP/HIP) produces high density with least shape distortion. However, high cost of equipment and long processing time associated with these processes make other cost effective alternatives more attractive. Uniaxial pressing in a rigid die with rigid particulate materials working as a PressureTransmitting Medium (PTM), such as mixture of alumina and graphite powers, has been under extensive investigation and has been used in industries as such an alternative due to its simplicity and much lower cost8"12. As shown in Fig. 1, when a uniaxial compressive load is applied, the PTM redistributes the stresses and creates Quasi-isostatic Pressing (QIP) conditions for the powder specimen located inside PTM, which is contained itself in a rigid die. The analysis of how this quasiisostatic stress mode effects the densification and shape distortion of the post-SHS porous body during QIP is important for near-net-shaped manufacturing. For this purpose, the constitutive behaviors of both the post-SHS porous body and the PTM need to be investigated. This study focuses on the densification and shape distortion of the post-SHS porous body during QIP. The constitutive properties that relate the stress, strain rate and material properties of the post-SHS porous body and the PTM have been developed based on the theory of nonlinear-viscous deformation of porous bodies13, and experimental studies. These properties have been incorporated into a Finite Element Method (FEM) code. The density development and the shape distortion evolution of the post-SHS porous body during QIP are numerically calculated and agree well with the experimental results. An optimization procedure is also proposed for determining the optimized shape of the pre-SHS green body rendering the desired final shape of the powder specimen after QIP.

113

Numerical Simulation of Densification and Shape Distortion of Porous Bodies

Fig. 1 Schematic representation of Quasi-Isostatic Pressing CONSTITUTIVE MODELING The post-SHS porous body can be considered a nonlinear-viscous porous material. Based on the continuum theory of sintering 9, its constitutive behavior follows a rheological relationship, which relates stress tensor atj and strain rate tensor éy. In the general case, a power law is used to describe the hot deformation of a nonlinear crystalline material and the stress; strain rate relationship can be expressed as13: (Sintering stress is significantly smaller than the external mechanical stress and is therefore ignored.)

¡φγ +ye

rP

φ \éSu

oo ), we simply replace R by 2Ä*, where: (6 R-=r¡r2/(r]+r2l This is in good quantitative agreement with numerical simulations [32, 33] for moderate size ratios (r2lrt < 4) and for early and intermediate configurations. Table I. Material parameters leading to r = R,/Ab/1

=1.17 106 sec and used in the simulations and

typical for fine Al 2 O 3 particles. SbDab (m 3 /s) Q„ (kJ/mole) Ω (m 3 ) 1.3 10 os 475 8.47 10 3 0

126


r

, (J/m 2 ) 2R (μνα) 1.1 0.2


s-—

h

/

;

~~^

/&

,I

'"N,

X

■ - . .

:

"s

\°s K

TS

_ - ' ' '

; substrate '■

i

Figure 1. The two types of contacts taken into account in the DEM simulation: between a particle and a plane and between two particles. Material parameters chosen for the simulations are given in Table 1. However, lengths may easily be normalized by the mean radius of the particles R, while time can be normalized by z=R4/A¡,y¡. A total sintering time of 17 minutes at 1200°C was imposed. These conditions lead to a relative density in the film of the order of 0.80-0.85 for the initial relative density of the film that we have chosen (0.62). The initial packing, before sintering, was created by first generating a gas of randomly located particles in the simulation box with no contact. The gas of particles was then slowly densified following a procedure described in [34] up to 0.60 relative density. This packing was slightly densified further to attain 0.62 relative density and to obtain small contacts between particles (a/A=0.15). We have used the following values of η: 0.-0.001-0.01-0.1 and 1, in order to simulate a wide range of interfacial conditions between the film and the substrate. The bottom surface of the film can sinter with the substrate while all other surfaces are free. However, it is not possible to generate packings with enough particles to model a whole film. Instead, packings of 4,000 to 40,000 spherical particles are generated numerically to form micro-pillars of various heights (axis z defines the vertical axis of the pillar). Both cylindrical and rectangular pillar geometries were tested with height ranging from 10 to 25 particle diameter. EFFECT OF THE SUBSTRATE ON POROSITY GRADIENT Fig. 2 shows the typical evolution of a rectangular micro-pillar with initial height of approximately 20 particle diameters, and for which both the particle-particle and particle-substrate contacts are set to a large value (r\pan = η5„(, = 0.1). The initial rectangular shape of the sample is not homothetically preserved as sintering proceeds. Instead, the sample densities much less close to the substrate than away from it. This is due to the drag that the substrate exerts on the particles. We observed that the first 6 to 7 layers of particle exhibit large porosities that are preferentially oriented in the z direction. This result is in agreement with the observations of Guillon et al. on 20 μηι, 50 μπι and 150 μπι thick AI2O3 constrained films [24, 25], However, the dip coated films studied by Guillon et al. were much thicker in terms of particle number than those simulated here (particle mean size is 0.15 μπι). Similar observations were made on YSZ/AI2O3 coatings [26]. The gradient of porosities, which is qualitatively shown in Fig. 2, may be calculated as a function of the height of the micro-pillar. Figures 3 and 4 show the evolution of the relative density with position from the substrate z, for the various micro-pillars that have been simulated for a low r\su¡, (r\sub — 0.001, Fig. 3) and an intermediate r\sub (r|su¡,= 0.01).


· 127

The Effect of a Substrate on the Microstructure of Particular« Films

Figure2: Sections showing the microstructural evolution of a rectangular micro-pillar during sintering on a substrate with r\pan= r\su/,= 0.1. Initial height: 20 particle diameters; Total number of particles: 37,000.

Relative Density

Relative Density

Figure 3. Gradient of relative density along the z axis. The relative density is calculated in the brown cylinder. η ρ ο η = 0.1 and η^ί,= 0.001. The insert views in Fig. 3 and 4 give the final shape of the pillars. When the substrate viscous parameter increases (Fig. 4), the density gradient increases as compared to the small viscous parameter case (Fig. 3). The effect of the viscous parameter η5„/, at the particle-substrate interface is clear both on the final shape of the pillars and on the porosity gradient. The results for the value rjJU¡, = 0.1 may be found elsewhere together with more details on these simulations [28]. It is essentially identical to the one shown in Fig. 4. Thus it appears that above 0.01 the effect of η^ί, saturates.

128


The Effect of a Substrate on the Microstructure of Participate Films

25

20

15

! 10

5 0 0.60

0.65

0.70

0.75

0.80

Relative Density

0.85

0.90

0.65

0.70

0.75

0.80

0.85

0.90

Relative Density

Figure 4. Same as .Fig. 3 with r\parl =0.1 and η™;, = 0.01. Symbols o and o indicate the average relative density along the whole height of the pillar. We did not observe a strong effect of the pillar geometry. However, we did observe that small chunks of particle clusters more often appear detached from the rectangular pillars than from the cylindrical ones. Also, we did not observe a strong correlation between the width to height ratio and the density gradient. We may thus infer that the present results should remain valid for films with larger width. We also measured the density gradient in the radial direction. Fig. 5 shows the typical gradient along the x and y directions (radial direction) for the pillar depicted in Fig. 2. From this figure, and from our observations on other pillars, it can be concluded that the substrate causes a gradient only along the z axis (apart from the very edge of the pillar). More interestingly, it shows that the free lateral surfaces play only a limited role in the microstructure development. This reinforces our belief that the results of the present simulations can be extended for much larger width that are typical of realistic films. Interestingly for applications, the height of the pillars influences the density gradient. Higher micro-pillars lead to larger porosity close to the substrate. The height effect is in qualitative agreement with the work of Guillon et al. who observed that, close to the substrate, 150 μηι ΑΙ2Ο3 thick films exhibit larger porosity than 50μιη AbOsthick films [24]. Fig. 4 shows that the top part of the highest pillars exhibit a relative density which is almost constant. Thus, it appears that above a height of approximately 10 particle diameters, the influence of the substrate fades. Note also that for the pillar heights studied here, the relative density averaged over the total height of the pillar does not depend on the pillar's height (as shown by the symbols in Fig. 4). EFFECT OF THE SUBSTRATE ON PARTICLE COORDINATION Additional microstructural information may be obtained from the coordination number of particles in the micro-pillar. Fig. 6 shows sections of a rectangular micro-pillar for r\sut, = 0.001 and f]mb = 0.1. The upper part of the micro-pillars shows approximately the same microstructure whatever the value of r\sub. Conversely, the particles close to the substrate have a lower coordination number especially for the η„;, = 0.1 case. More generally, observations on all tested samples have shown that the low coordination region is limited to 0.4 times the total initial height of the micro-pillar.


· 129


Relative Density Figure 5. Radial gradient of density in the x and y (radial) directions for the pillar shown in Fig.2. ηΡ, ~ 4

-

ω CC

0.6

Figure 4. Temperature and relative density changes during sintering of the partially insulated compact. The curves corresponding to Center and Bottom temperature and relative density are almost superimposed.

178


Finite Element Modelling of Microwave Sintering

Figure 5. Position of the susceptor in the cavity

Figure 6. Distribution of the electric field in a plan normal to the wave propagation direction in the whole cavity (a), in the compact and susceptor only (b)


· 179

Finite Element Modelling of Microwave Sintering

CONCLUSION The interest of finite element simulation for better understanding of microwave sintering in a monomode cavity has been demonstrated. We first calculated the distribution of electromagnetic field in the cavity and in compacts of various lengths and we found that it was easier to sinter a long compact. This effect has been confirmed by experiments. Next we showed the temperature and density gradients in a compact with complete or partial thermal insulation. Finally we investigated the electric field when a susceptor surrounded the compact and found that the compact undergoes hybrid heating. This simulation should be improved following two axes. First the dimensional changes of the compact resulting from densification should be taken into account as they certainly affect microwave propagation and heat transfer. The second requirement is more challenging and critical for a quantitative prediction of microwave sintering. It consists in finding reliable values of dielectric parameters as function of temperature and density. Both experimental and modeling approaches should be combined for this purpose. REFERENCES 1 J.D. Katz, Microwave sintering of ceramics, An. Rev. Mater. Sei., 22,153-70, (1992). 2 K.H. Brosnan, G.L. Messing, and D.K. Agrawal, Microwave sintering of alumina at 2.45 GHz, J. Am. Ceram. Soc, 86, 1307-12, (2003). 3 M.A. Willert-Porada, R. Borchet, Microwave sintering of metal-ceramic FGM, Proc. 4th Int. Symp. Functionally Graded Materials, Tsukuba, Japan, October 21-24, 1996, Elsevier, (1997). 4 J. Wang, J. Binner, B. Vaidhyanathan, N. Joomun, J. Kilner, G. Dimitrakis, Cross T.E., Evidence of the microwave effect during hybrid sintering, J. Am. Ceram. Soc., 89, 1977-1984, (2006). 5 Iskander M.D., Andrade A.O.N.M., FDTD simulation of microwave sintering of ceramics in multimode cavities, IEEE Trans. Microwave Heating and Techniques 42, 793-9, (1994). 6 J. Lasri, P.D. Ramesh, and L. Schächter, Energy conversion during microwave sintering of a multiphase ceramic surrounded by a susceptor, J. Am. Ceram. Soc, 83, 1465-68, (2000). 7 A. Birnboim, Y. Carmel, Simulation of microwave sintering of ceramic bodies with complex geometry,/ Am. Ceram. Soc., 82, 3024-30, (1999). 8 R. Riedel, and J. Svoboda, Simulation of microwave sintering with advances sintering models, Adv. in Microwave and Radio Frequency Processing, (2006). 9 S. Charmond, C.P. Carry, and D. Bouvard, Direct and hybrid microwave sintering of yttria-doped zirconia in a single-mode cavity, Ceram. Trans., this issue.

180


DIRECT AND HYBRID MICROWAVE SINTERING OF YTTRIA-DOPED ZIRCONIA IN A SINGLE-MODE CAVITY S. Charmond, C. P. Carry, D. Bouvard Laboratoire SIMaP, Grenoble-INP / CNRS / UJF BP 75, 38 402 St-Martin d'Héres, France ABSTRACT The use of microwave energy to sinter ceramic materials is an emerging technology of wide interest for improving the microstructure and reducing the time of sintering process. The microstructural development of 2 mol% yttria-stabilized zirconia nanopowders during direct and hybrid microwave sintering was studied for a constant heating rate (25 °C/min). Microwave heating experiments were achieved in a 2.45 GHz single-mode cavity instrumented with a thermal imaging camera. Insulating materials around the specimen was used to limit radiative thermal losses and prevent specimen from too high temperature gradients leading to crack development. Moreover, a cylindrical SiC susceptor was used for hybrid sintering experiments. Constant-heating rate runs were controlled by adjusting the position of a removable reflector at constant microwave forward power. The microwave temperature-time profile and the microwave power dissipated in the cavity were recorded and discussed. Next, the final densities were measured and the microstructures were observed by SEM. These results were compared to those of the conventionally-sintered materials. Microwavesintered materials always presented higher final densities for a same maximum measured temperature and, especially for hybrid-microwave sintering, homogeneous and finer microstructures. But direct microwave heating runs led to heterogeneous microstructures due to thermal gradients, which makes difficult an analysis in terms of microwave effect. Anyway, in comparison to the classical "sluggish" grain growth of TZP materials, grain growth seemed to be accelerated during microwave heating above a density estimated to be about 96 %TD. We thus conclude that microwaves are beneficial under this density and detrimental above. INTRODUCTION Microwave processing is generally characterized by uniform heating on a macroscopic scale and rapid heating rates, as compared to conventional thermal processing. It has been shown that microwave-processed ceramics led to improved densification with lower temperature and shorter times than conventional sintering with the development of finer and more uniform grain microstructures '"5. These main characteristics were attributed to rapid heating rates because the heat is generated within the materials. For instance, microwave heating of dielectric materials results from the absorption of part of the energy transported by an oscillating electric field by ionic conduction and molecular vibration. The microwave sintering of yttria-doped zirconia has been ever largely studied in various kinds of microwave furnaces : multimode and single-mode cavities with more or less complex insulation and susceptor devices, with or without pre-sintering, with temperature measurements by thermocouple, optical and IR pyrometer, thermal imaging camera, with various shapes of green samples...4"17. All of these experimental procedures make difficult to determine the microwave heating behaviors and characteristics of zirconia ceramics under pure direct electric field. A rectangular single-mode cavity at 2.45 GHz instrumented with a thermal imaging camera was designed to study both direct and susceptor-assisted microwave sintering of various kinds of materials (ceramics and metals) under either electric or magnetic field. This microwave furnace is an ideal support for understanding microwave-materials interactions. The specificity and selectivity of the microwave heating appeared the most interesting characteristics for sintering multicomponents, multilayers and functionally gradient materials. In this study, we present the first results of the

181


sintering experiments of ceramic material in our original microwave furnace. During the microwave process, constant heating rate sintering runs were achieved at a constant forward microwave power by positioning a movable reflector to yield to the desired temperature-time profile. The final densities and microstructures of microwave-sintered materials were compared to the conventionally-sintered ones. EXPERIMENTAL PROCEDURE Spray-dried 2 mol% yttria-doped zirconia powder manufactured by Tosoh was used. Its BET specific area is 16 m2/g and the primary crystallites mean size is 60 nm. The zirconia grains are agglomerated in spherical aggregates which are in the range 10-80 μπι in diameter, as observed by SEM. Green samples were prepared either by cold isostatic pressing (CIP) at 250 MPa or in two steps : first, uniaxial pressing (UP) at low pressure (50 MPa) and next, cold isostatic pressing at 250 MPa. The green density was around 2.9 g/cm (i.e. 48 % of theoretical density). Before sintering, the binder was burnt out under air in an electric furnace by heating at 5 °C/min up to 700 °C and soaking for 2 hours before cooling. The weight loss was about 0.5 %. Without debinding, we observed crack development during microwave heating due to heterogeneous temperature distribution, as mentioned by Janney 5 . Green samples were sintered by direct microwave heating, by susceptor-assisted hybrid microwave heating or by conventional heating in vertical dilatometers (SETARAM TMA92, France or LXNSEIS L75/1550, Germany). The reference thermal cycle includes heating at 25 °C/min up to the sintering temperature, without dwelling time, and fast coolmg (faster than 30 °C/min). Bulk density of the sintered samples was measured by ethanol immersion method (based on Archimede principle). Sintered samples were cut using a diamond saw in two half-cylinders. Polished sections were thermally etched in an electrical oven under air between 50 °C to 80 °C below their maximal sintering temperature during 40 min to reveal their microstructure. The average grain size was determined on SEM micrographs by a linear intercept method. EXPERIMENTAL SETUP As shown in Figure 1, the microwave furnace includes a high voltage power supply (microwave generator SAIREM GMP 20KSM, France) linked to a magnetron that delivers a variable forward power up to 2 kW at 2.45 GHz. A rectangular wave-guide (type WR340 - section 86.36 x 43.18 mm 2 ) allows the transport of the microwave radiation to a rectangular TEiop cavity. This cavity is ended on the magnetron side by a coupling iris (a vertical slot in a copper sheet) placed in the waveguide and at fixed position during experiments, and on the other side by a movable reflector, also called sliding short-circuit piston.

182



Figure 1. Schematic of the microwave furnace Samples were set in our cavity on a quartz plate inside a 40 mm-diameter quartz tube and surrounded by insulating material (alumina fiberboard box), as shown in Figure 2. In case of susceptorassisted microwave heating (hybrid sintering), a cylindrical SiC susceptor (internal diameter : 20 mm, height : 15 mm, thickness wall : 5 mm) is placed around the compact (cf. Figure 3). The temperatures of the upper surface of the specimen and of the susceptor were measured by an IR camera (FLIR SYSTEMS, ThermoVision™ A40M, Sweden) during the process to yield to the desired temperaturetime profile. This IR camera is placed just behind a ZnSe window setting about 14 cm above the center of the cavity. As drawn in Figure 2 and Figure 3, a small hole in the cap of the insulating box is needed to measure temperature of the upper section of the sample. During a test we measure the input power delivered by the magnetron (forward power) and the output power leaving the cavity (reflected power). We defined the dissipated powder as the difference between the input and output powers. All the experiments and data recordings were conducted manually.

Figure 2. Direct microwave sintering Figure 3. Hybrid microwave sintering The forward electromagnetic wave is reflected against the conductive wall of the short-circuit piston, leading to a standing wave made of a single stacking of forward and reflected waves. In the empty cavity, the points where the magnetic and electric fields are respectively maximum and minimum (near to zero) are exactly distant from each other by a quarter of the guided wave-length (43 mm). When the iris-piston distance is equal to 4 or 5 times the half guided wave-length, the iris reflects almost entirely the wave in the cavity (TE104 or TE105 resonance mode). The resonance phenomenon consists in the stacking of multiple waves reflected against the piston and the iris, which results in an increase of the electromagnetic energy in the cavity. In our single-mode cavity, the TE104 and TE105 modes are used for materials interacting respectively with the magnetic field or the electric field. Since ceramic materials as yttria-doped zirconia interact with the electric field, our specimens should be located where the electric field is maximum, i.e. theoretically in the centre of the cavity in TE105 mode. However, when materials are placed in the cavity (specimen, quartz and alumina holders, insulating materials, susceptor, etc.) the electromagnetic pattern and thus the distances between the iris, the


■ 183


specimen and the piston for getting resonance and stronger material-wave interaction are modified. In our setup the iris and the specimen are fixed during an experiment and the piston can be moved. In the tests presented here, the iris-specimen distance was chosen to be about equal to the specimen-piston distance allowing the best coupling between the material and the electric field at the resonance mode. During the test, the position of the piston was continuously adjusted so that the temperature measured by the infrared camera followed the prescribed cycle. This means that we did not stick to the position that was optimal for specimen heating. We observed in preliminary runs that this optimal position often results in overheating and cracking of the specimen. RESULTS In the following, six sintering experiments will be presented and analyzed : 3 conventional sintering tests (C), 2 direct microwave sintering runs (MWd) and 1 hybrid microwave sintering run (MWh). Table I provides the compaction mode, the dimensions of the green compact, the forward microwave power (FP), the maximum sintering temperature and the final density for each test. Note that the compacts had different sizes, ranging between 7.4 and 10.0 mm in diameter, 7.2 and 11.9 mm in height, 1.0 and 2.1 g in weight. Name

Pressing (MPa)

Cl UP50/CIP250 C2 CIP250 C3 CIP250 MWd4 CIP 250 MWd5 UP50/CIP250 MWh6 UP50/CIP250

Diameter (mm)

Height (mm)

Mass (g)

FP (W)

7.4 8.5-9.1 9.0 10.0 7.4 7.4

7.7 11.9 9.4 7.2 7.7 7.8

0.98 2.13 1.79 1.64 0.97 0.98

-

Tmax (°C)

1320 1390 1500 600 1390 1000 plasma 1000 1340

Density (g/cm3)

Relative Density

5.56 5.76 5.98 5.88 5.99 5.80

91.9 95.1 98.8 97.3 98.9 95.8

(%)

Direct microwave sintering The temperature of the sample upper surface and the dissipated power recorded during the two direct microwave sintering tests are drawn in Figure 4 (MWd4) and Figure 5 (MWd5). The temperature variations are compared to the reference thermal cycle. During the first run (MWd4), the forward power and the iris-cavity distance were fixed respectively to 600 W and 135 mm. The compacts weighted about 1.6 g.

184



Figure 4. Direct microwave sintering run (MWd4 - 1390°C - FP : 600W) The heating rate of 25 °C/min was controlled from room temperature to about 1400 °C by adjusting the position of the sliding short-circuit piston. At low temperature, the dissipated power increased slightly up to 50 W and remained constant up to 500 °C. Next, the dissipated power decreased to 40 W while the temperature of specimen upper surface continued to increase up to 1000 °C. Above this temperature, the dissipated power increased steadily up to 170 W when the temperature reached 1400 °C. Finally, the forward power was cut off for dropping in temperature.


· 185


Figure 5. Direct microwave sintering run (MWd5 - 1350°C - FP : 1000W) In the second run (MWd5) with a smaller sample, the forward power was fixed to 1000 W. From room temperature to 500 °C, this run was similar to the previous one : the dissipated power increased slightly up to 60 W and was next constant. But above 500 °C, temperature instabilities occurred during the test. It was impossible to control the heating rate by changing the position of the piston. For this range of temperature, we believe that specific changes in dielectric properties of zirconia (sharp increasing of loss tangent above 400 °C 18) combined with the higher forward power (1000 W instead of 600 W previously) led to this unusual behavior. Therefore, during this second run it has been difficult to control the heating rate in the 500-900 °C range. Finally, the dissipated power increased up to 100 W at 1350 °C before forward power being promptly cut off because of plasma occurrence. This plasma appeared in the small hole drilled in the insulating cap. It burnt a part of insulating- materials and overheated the specimen. Hence, the maximum temperature in the specimen was unknown. Despite this overheating, the specimen was not cracked. It was characterized but the results could not be decently compared to those of the conventional experiments. Hybrid microwave sintering In the hybrid sintering run, the distance between the iris and the cavity was set to 122 mm (determined from "pre-runs") to keep a symmetrical pattern and the maximum of the electric field in the susceptor and specimen area at the resonant mode. The temperature-time profile of this specimen followed the prescribed heating rate (25 °C/min).

186



Figure 6. Hybrid microwave sintering run (MWh66 - 1350°C - 1000W) From room temperature to 600 °C, the sample remained colder than the SiC susceptor because of its poorer dielectric properties (loss tangent) at low temperature. Therefore, zirconia was mainly heated by thermal radiance from the susceptor. However above 600 °C, zirconia got hotter than the susceptor : its temperature followed more or less the prescribed thermal cycle from 600 °C to 1000 °C whereas the temperature of susceptor remained constant around 650 °C. We deduce that the compact is heated by both radiance from the susceptor and coupling with the microwaves (hybrid heating). The dissipated power varied in the 150 W - 350 W range. Above 1000 °C, the dissipated power increased steadily but sharply up to 900 W and the temperature reached 1350 °C at the upper surface of the zirconia specimen and only 950 °C in the susceptor. Moreover, we note that the evolution of the susceptor temperature vs. time shows the same trend as the dissipated power. Also this power is much higher than the one measured during direct microwave sintering runs. This proves that the main part of microwave energy dissipated in the cavity is used to heat the SiC susceptor and balance the radiative losses, especially above 700 °C. Characterization of sintered materials


· 187


Figure 7. Relative Density vs. Temperature for Direct Microwave (MWd), Hybrid Microwave (MWh) and Conventional (C) Sintering First of all, the density of the conventionally-sintered materials increases from 91.9 % at 1320 °C up to 98.8 % at 1500 °C. Next, the density of the direct microwave-sintered material heated to 1390 °C (upper surface temperature) is significantly higher than the conventionally sintered one (97.3 % compared to only 95.1 %), i.e. the porosity was reduced from 4.9 % to 2.7 %. At last, the density of the hybrid-microwave-sintered material heated to 1340 °C is 95.8 %TD. This hybrid microwave-sintered sample was then much denser than the one conventionally sintered at 1320 °C and even slightly denser than the conventionally-sintered one at 1390 °C. The microstructures (Cl, C2, MWd4, MWh6) were observed by scanning electron microscopy (SEM - Secondary Electrons detector) along the axis of the cylindrical samples on longitudinal sections.

188



Figure 8. Conventional sintering (Cl) Figure 9. Conventional sintering (C2) 1320 °C - g.s. = 400 nm 1390 °C - g.s. = 435 nm The conventional-sintered specimens presents a homogeneous microstructure of equiaxe grains, as expected, with an average size of 400 nm at 1320 °C (cf. Figure 8) and 435 nm at 1390 °C (cf. Figure 9).

g.s. = 395 nm

g.s. = 490 nm


· 189


g.s. = 540 run Figure 10. Direct-microwave-sintering (MWd4) Microstructures of the top, central and bottom zones of the specimen The microstructure is heterogeneous in the direct microwave-sintered material. The grains are smaller, i.e. less than 400 nm, in top zone, but they are larger, i.e. about 490 nm in the central zone and 540 nm in the bottom zone (cf. Figure 10). This microstructural gradient likely results from a thermal gradient.

Figure 11. Hybrid microwave sintering (MWh6) 1340 ° C - g.s. = 370 nm The final microstructure of the hybrid microwave-sintered material heated at 1340 °C is the finest (cf. Figure 11), i.e. 370 nm instead of 400 nm for the conventionally-sintered one at 1320 °C (cf. Figure 9). Moreover, the microstructure can be considered as homogeneous (average grain size : 370 nm in the upper zone ; 378 nm in the center; 357 nm in the bottom zone). DISCUSSION The external surface of the conventionally-sintered compact is heated by radiation and convection and the heat is transferred by conduction from the surface to the core of the sample. Then the temperature should be higher in the surface of than in the core. But the effects of this gradient can be neglected because homogeneous microstructures are observed. Microwave processes involve a different heating behavior. During microwave sintering, the heat is produced in the bulk of the compact. Therefore, an inverse thermal gradient takes place in materials, with higher temperature in the core. That is why thermal insulation is necessary to reduce radiative thermal losses through external surfaces and, as consequence, decrease temperature gradients in the specimen. Besides, in our device the insulation of the upper part is incomplete because we need to keep a hole for temperature measurement. Thus it is expected that the temperature measured by the IR camera is inferior to the

190



temperature in the core. During hybrid microwave sintering, heat is produced both in the bulk of the specimen and upon its surface by radiance from the susceptor. Thus it is expected that temperature gradients are less marked. As shown in Figure 7, for a given sintering temperature, both direct and hybrid microwavesintered materials reached higher final densities than the conventionally-sintered ones. As homogeneous microstructure was observed in the hybrid microwave-sintered specimen (MWh6), we can consider that the temperature measured on the specimen upper surface was representative of the bulk temperature. Therefore, the SiC susceptor acts as an efficient insulation system, which does not shield zirconia samples from microwaves. Besides, we noticed higher density (95.8 %) and finer microstructure (370 nm) than in conventional-sintered samples for comparable sintering temperature (respectively 93 % and 400 nm). Thus, we evidenced with this hybrid microwave heating experiment the existence of a favourable microwave effect on sintering behavior of zirconia, at least for relative density up to 96 %TD. In contrast with hybrid microwave sintering, the microstructural gradient observed in the direct microwave-sintered specimen (MWd4) is linked to thermal gradients. Higher grain size in the core of the specimen indicates higher temperature in the core than at the upper surface. Assuming no microwave effect for specimen MWd4, the temperature difference between the core and the upper surface could be estimated to be about 50 °C according to the density-temperature trend line of conventionally-sintered specimens (97.3 % at 1450 °C) in Figure 7. However, this temperature difference should be less if there is a positive microwave effect on densification as shown during hybrid microwave sintering. Concerning the microstructures, the average grain sizes measured in the upper zones of the direct microwave sintered specimens (395 nm) were always smaller than those observed in the conventionally-sintered ones (440 run), which is consistent with the hybrid sintering test. Nevertheless, the grain size in bottom (540 nm) and central (490 nm) zones is much higher than those observed in the conventionally-sintered ones (440 nm). In Figure 7, we noticed that the grain growth in conventional sintering is very slow, as already reported in the literature for such TZP materials in connection with the phase partitioning process ("sluggish" grain growth 19). As a consequence, the higher grain size in the core of MWd4 specimen may indicate that microwaves enhance the grain growth through the phase partitioning process. It is likely that such effect has not been observed in hybrid experiment due to lower density (95.8 % instead of 97.3 %). A hybrid microwave sintering experiment at higher temperature to reach higher density than 96 %TD should be achieved to validate that analysis. Also, as "sluggish" grain growth kinetics were not observed in 8 mol% yttria-doped zirconia materials, it will be interesting to sinter such materials with our setup and compare microwave effects with the effects evidenced in the present study. CONCLUSIONS Direct or hybrid microwave sintering experiments of 2 mol% yttria-doped zirconia were performed in a single-mode cavity instrumented with a thermal imaging camera measuring the temperature of the upper surface of the specimen. Hybrid microwave sintering experiments provided a homogeneous specimen with higher density and finer microstructure than the specimen conventionally-sintered at the same temperature. On the contrary direct microwave heating runs led to heterogeneous microstructures due to thermal gradients, which makes difficult an analysis in terms of microwave effect. Anyway, in comparison to the classical "sluggish" grain growth of TZP materials, grain growth seemed to be accelerated during microwave heating above a density estimated to be about 96 %TD. We thus conclude that microwaves are beneficial under this density and detrimental above. Further experiments will show (i) if there is enhanced grain growth during homogeneous microwave heating at densities higher than 96 %TD, (ii) if such effect is also observed during microwave sintering of 8 mol% yttria-doped zirconia.


· 191


REFERENCES 1 W. H. Sutton, Microwave Processing of Ceramic Materials, Ceramic Bulletin, 68 [2], 376-386 (1989). 2 S. S. Park, T. T. Meek, Characterization of Ζ1Ό2-ΑΙ2Ο3 Composites Sintered in a 2.45 GHz Electromagnetic Field, Journal of Materials Science, 26, 6309-6313 (1991). 3 D.E. Clark, W.H. Sutton, Microwave Processing of Materials, Annu. Rev. Mater. Sei., 26, 299-331 (1996). 4 Y.-L. Tian, Pratices of Ultra-rapid Sintering of Ceramics using Single-mode Applicators, Ceramic Trans, American Ceramic Society, 21, 283-300 (1991). 5 M.A. Janney, C.L. Calhoun, H.D. Kimrey, Microwave Sintering of Oxide Fuel Cell Materials : I, Zirconia-8 mol% Yttria,y. Am. Ceram. Soc, 75 [2], 341-46 (1992). 6 C.E. Holcombe, T.T. Meek, N.L. Dykes, Unusual Properties of Microwave-Sintered Yttria-2 wt% Zirconia, Journal of Materials Science, 7, 881-884 (1988). 7 J. Wilson, S.M. Kunz, Microwave Sintering of Partially Stabilized Zirconia, /. Am. Ceram. Soc, 71 11], C-40-C-41 (1988). S.A. Nightingale, D.P. Dunne, H.K. Worner, Sintering and Grain Growth of 3 mol% Yttria Zirconia in a Microwave Field, Journal of Materials Science, 31, 5039-5043 (1996). 9 Z. Xie, J. Li, Y. Huang, X. Kong, Microwave Sintering Behaviour of ZrO2-Y203 with Agglomerate, Journal of Materials Science, 15, 1158-1160(1996). 10 S.A. Nightingale, H.K. Worner, D.P. Dunne, Microstructural Development during the Microwave Sintering of Yttria-Zirconia Ceramics,/. Am. Ceram. Soc, 80 [2], 394-400 (1997). 1 ' A. Goldstein, N. Travitzky, A. Singurindy, M. Kravchik, Direct Microwave Sintering of YttriaStabilized Zirconia at 2.45 GHz, Journal of the European Ceramic Society, 19, 2067-2072 (1999). 12 S. Fujitsu, M. Ikegami, T. Hayashi, Sintering of Partially Stabilized Zirconia by Microwave Heating Using ΖηΟ-Μηθ2-Αΐ2θ3 Plates in a Domestic Microwave Oven, J. Am. Ceram. Soc, 83 [8], 20852087 (2000). 13 C. Zhao, J. Vleugels, C. Groffils, P.J. Luypaert, O. Van Der Biest, Hybrid Sintering with a Tubular Susceptor in a Cylindrical Single-Mode Microwave Furnace, Acta mater., 48, 3795-3801 (2000). 14 D.D. Upadhyaya, A. Ghosh, K.R. Gurumurthy, R. Prasad, Microwave Sintering of Cubic Zirconia, Ceramics International, 27, 415-418 (2001). 15 J. Wang, J. Binner, B. Vaidhyanathan, N. Joomun, J. Kilner, G. Dimitrakis, T.E. Cross, Evidence for the Microwave Effect during Hybrid Sintering, J. Am. Ceram. Soc, 89 [6], 1977-1984 (2006). 16 J. Binner, K. Annapoorani, A. Paul, I. Santacruz, B. Vaidhyanathan, Dense Nanostructured Zirconia by Two Stage Conventional/Hybrid Microwave Sintering, Journal of European Ceramic Society, 28, 973-977 (2008). 17 M. Mazaheri, A.M. Zahedi, M.M. Hejazi, Processing of nanocrystalline 8 mol% yttria-stabilized zirconia by conventional, microwave-assisted and two-step sintering, Materials Science and Engineering, A492,261-267 (2008). 18 M. Willert-Porada, T. Gerdes, Metalorganic and Microwave Processing of Eutectic Ai203-Zr02 Ceramics, Mat. Res. Soc. Symp. Proc, Vol. 347, 563-569 (1994). 19 T. Stoto, M. Nauer, C. Carry, Influence of Residual Impurities on Phase partitioning and Grain Growth Processes of Y-TZP Materials, J. Am. Ceram. Soc, 74 [10], 2615-21 (1991).

192


THE INFLUENCE OF MINOR ADDITIVES ON DENSIFICATION AND MICROSTUCTURE OF SUBMICROMETER ALUMINA CERAMICS PREPARED BY SPS AND HIP Jaroslav Sedlácek1,2, Monika Michálková1, Deniz Karaman2, Dusan Galusek1, Michael Hoffmann2 'Vitrum Laugaricio - Joint Glass Center of the Institute of Inorganic Chemistry, Slovak Academy of Sciences, Alexander Dubcek University of Trencin, and RONA, j.s.c, Trenöin, Slovakia 2

Institut fur Keramik im Maschinenbau, TU Karlsruhe, Karlsruhe, Germany

ABSTRACT Refinement of microstructure and complete elimination of residual porosity are prerequisite for improvement of mechanical properties and achievement of transparency for visible light in polycrystalline alumina (PCA). The two requirements are in obvious contradiction, as the elimination of porosity by conventional sintering requires high sintering temperatures and long dwell times, inevitably accompanied by grain growth in the final stage of densification. Spark plasma sintering (SPS) followed by hot isostatic pressing (HIP) were applied for preparation of fully dense PCA with submicrometre microstructure from high purity alumina powder Taimicron TM DAR with the initial particle size 150 nm. SPS resulted in some microstructure refinement in comparison to conventionally sintered samples. In the reference material A (pure PCA) temperatures > 1200 °C led to rapid grain coarsening without complete densification. Addition of 500 ppm of MgO, ΖΚ>2 and Y2O3 led both to microstructure refinement and deceleration of densification, and the temperatures by 50 - 100 °C higher in comparison to the reference A were required to achieve the same microstructure characteristics. The residual porosity was successfully eliminated by HIP at 1100 - 1250 °C. Fully dense reference material A with the mean grain size < 500 nm was prepared by HIP at 1150 °C. Higher HIP temperatures led to quick attainment of full density followed by rapid grain growth. The Y2O3 and Zr02-doped PCA required higher temperatures for complete densification, and fully dense materials with the mean grain size < 500 nm were prepared by HIP at 1250 °C.

INTRODUCTION Polycrystalline alumina with submicrometre grains (0.3 to 1 urn) was reported to exhibit high hardness, 1,2 good mechanical strength, 3 " 5 and improved wear resistance 6. Alumina also transmits infrared, and if sintered to high density (residual porosity < 0.01 %), also visible light with possible applications in high pressure envelopes of metal halide discharge lamps 7,8 , or transparent armours. Nevertheless, even if light scattering residual porosity is completely eliminated, polycrystalline ceramics are usually only translucent due to birefringence of alumina. The linear transmission of visible light is believed to be markedly increased by decrease of the grain size to less than the wavelength of visible light, i.e. to below ~ 300 nm . This goal is in obvious contradiction with the requirement of complete elimination of the residual porosity to less than 0.01 %, and cannot be attained by conventional sintering process, which for this purpose usually requires long soaking times at high temperatures in hydrogen atmosphere, accompanied by significant, and often abnormal, grain growth and inevitable deterioration of mechanical properties. During the standard sintering process grain growth is generally observed to take place at the final stage of sintering, or more specifically, when the last 3 % of porosity is eliminated 1 0 ' u . A number of works therefore describes the use of various, mostly pressure assisted, sintering techniques, e.g. spark plasma sintering12, hot isostatic pressing, or sinter-HIP techniques13 or their combinations with conventional sintering in order to eliminate residual porosity without excessive grain growth.

193

Influence of Minor Additives on Densification and Microstructure of Alumina Ceramics

Spark plasma sintering (or field assisted sintering technique FAST) attracted most attention in the last few years, with reported decrease of sintering temperature and shortening of required soaking times to a fraction of times usually applied in conventional sintering. Since about mid nineties a number of works dealt with spark plasma sintering of polycrystalline alumina 12,14 " 12. There seems to be general agreement that the method significantly accelerates the processes responsible for both the densification and grain growth, either by self heat generation by miroscopic discharge between particles, activation of particle surfaces 5 or by electric field accelerated grain boundary diffusion and grain boundary migration19. It is tempting to believe that grain boundary migration requires higher temperatures and pressures that densification, and that the SPS process might, under certain conditions, result in significant microstructure refinement during sintering of alumina. Indeed, some works report on refinement of the microstructure of SPS-PCA in comparison to traditional sintering, or retention of very fine grained microstructure of PCA prepared from submicrometre powders 14 ' 18 ' °. However, the densification rate and grain growth are influenced by the parameters of the SPS process, including heating rate17 " 19, applied pressure17' ", specimen thickness17, pulse sequence16, 9, and maximum temperature/time of isothermal dwell1 ' 19 and therefore no unambiguous conclusions can be drawn from the available data. In general, high pressure, high heating rate (> 50 "C.min"1) and low temperatures seem to be beneficial for retention of fine grained microstructure. Moreover, despite reported grain growth retarding action of MgO in SPS-PCA14·21, little attention has been paid to the influence of other additives/impurities, namely Z1O2 and Y2O3, whose influence on microstructure development during conventional sintering of alumina has been documented. The present work is aimed at the study of the SPS process of a submicrometre alumina powder, in particular the influence of minor, deliberately added dopants MgO, Z1O2 and Y2O3 on densification and grain growth during SPS, and post-SPS hot isostatic pressing. The results are compared to reference pure polycrystalline alumina densified by SPS and with the use of conventional pressureless sintering. EXPERIMENTAL Alumina powder (Taimicron TM-DAR, Taimei Chemicals Co., Japan) with the mean particle size 150 nm and the surface area 14.5 m2 g"1 was used as a starting material in all experiments. The reference aluminas were prepared both by pressureless sintering and by SPS. In the former green discs 12 mm in diameter and 6 mm thick, prepared by axial pressing in a steel die at 50 MPa followed by cold isostatic pressing at 500 MPa were sintered in an electrical furnace with M0SÍ2 heating elements in air at various temperatures ranging from 1000 to 1350 °C without isothermal dwell, heating rate 10 °C/min. In the latter 6 g of the alumina powder was filled into a graphite die with the diameter of 20 mm and spark plasma sintered in the temperature range between 1150 and 1250 °C, heating rate 400 °C/min, pressure 150 MPa, equilibration time at maximum temperature 1 min, and subsequent isothermal dwell between 1 and 6 minutes. The specimens are denoted as A in the following text. Doped powders were prepared by mixing 100 g of the alumina powder with respective amounts of suitable precursor: Mg(N03)2.6H20 (p.a., Lachema Brno, Czech republic), zirconium isopropoxide (p.a. Sigma Aldrich, and Y2O3 (99.9 % purity, Treibacher Industries AG, Austria) dissolved in nitric acid (p.a. Lachema Brno, Czech republic). The mixture was homogenised in a polyethylene jar in isopropanol (pure, Sigma Aldrich) with high purity alumina milling balls for 2 h. The water solution of ammonia was then added to precipitate respective hydroxides. The mixtures were then further homogenised for 2 h to complete the hydrolysis, and dried at continuous stirring under infrared lamp. The powders were gently crushed, sieved through a 100 um polyethylene sieve, calcined for 1 h at 800 °C in air, and again sieved to obtain a reasonably free flowing powder. The powders were filled into a graphite die with the diameter of 20 mm and spark plasma sintered at temperatures between 1150 and 1300 °C, under the same conditions as the reference A. The specimens containing the respective dopants MgO, Y2O3 and Z1O2 are denoted as AM, AY, and AZ.

194



The density of sintered samples was measured by the Archimedes method in water. The microstructures were examined on fracture surfaces by scanning electron microscopy (Zeiss, model EVO 40HV, Carl Zeiss SMT AG, Germany). The mean grain size was determined using the linear intercept method on fracture surfaces 23. In order to eliminate possible error introduced by measurement of the grain size on fracture surface, an empirical correction factor was applied, calculated from the comparison of the grain sizes measured on the same specimen on both fracture surfaces and the polished and thermally etched cross sections. Minimum of 200 grains were measured in order to obtain statistically robust set of data. Selected specimens, which during SPS sintered to the stage of closed porosity (or near to it) were cut in four parts, and each part was hot isostatically pressed at a different temperature 1050, 1100, 1150 or 1250 °C, with 3 h isothermal dwell at the maximum pressure of 150 MPa with Ar as the pressure medium. The HIP-ed specimens were characterized in the same manner as described above. RESULTS AND DISCUSSION Spark Plasma Sintering The results of the SPS experiments are summarized in Fig. 1 - Fig. 5. The SEM micrographs of fracture surfaces of the pure PCA (material A) and PCAs with 500 ppm of the respective dopants (AM, AY, and AZ) subjected to the same SPS regime (1 min isothermal dwell at 1200 °C) are shown in the Fig. 1. Under these conditions the material A was already highly dense, with a characteristic microstructure consisting of angular microcrystals with the mean size of 230 nm. In the material AM both the densification and grain growth were slightly suppressed. This effect was even more pronounced in case of AY and AZ, where the alumina powder maintained its original round-shaped morphology, with only necks formed among individual particles. Grain growth was negligible (180, and 190 nm grain size in comparison to 150 nm in the starting powder), and the densification was almost entirely suppressed, the relative density achieving only 73.9 and 77.6 % in AY and AZ, respectively. The results demonstrate grain growth and densification retarding action of Y2O3 and Ζ1Ό2 additives. The influence of temperature and time of isothermal dwell on the microstructure development of PCA is illustrated in Fig. 2. In the reference A both the relative density and the mean grain size increased markedly with temperature and time of isothermal dwell. Heating of 6 minutes at 1250 °C resulted in microstructure with equiaxial alumina grains with the mean size 900 nm, which is 6 times the initial size of the powder particles. The powder AY sintered 6 minutes at 1300 °C achieved comparable relative density, with finer microstructure and the mean grain size of only 300 nm. Similar results were obtained also in AM and AZ, Y2O3 and Z1O2 being the most effective additives. The time dependencies of relative densities (/>,) of all materials sintered at various temperatures are summarized in Fig. 3. A marked influence of the temperature of isothermal dwell, but especially of the additives, is obvious. After 3 minutes of sintering at 1150 °C the reference A sintered to the stage of closed porosity (> 90 % relative density). The addition of MgO shifted the required temperature by about 50 °C to higher values, and the material AY required at least 1 minute isothermal dwell at 1300 °C to achieve the relative density exceeding 90 %. The situation is similar in the materials AZ. Quite clearly the addition of as little as 500 ppm of the respective oxides significantly influences the densification rate, most likely by segregation into grain boundaries and impairing the grain boundary diffusion in sintered compacts. Along with grain boundary diffusion, the grain boundary mobility is also affected (Fig. 4): while in the reference material A all temperatures > 1150 °C resulted in rapid grain coarsening, the additives suppressed the grain growth so that the grain size did not exceed 400 nm in the whole temperature and time interval studied.


· 195

Influence of Minor Additives on Densificaron and Microstructure of Alumina Ceramics

Fig. 1 SEM micrographs of fracture surfaces of SPS densified specimens after identical heat treatment 1 min at 1200 °C: a) A, pr = 93.6 %, MGS = 230 nm, b) AM, p, = 88.5 %, MGS = 210 um, c) AY, pr = 73.9 %, MGS = 190 nm, d) AZ, pr = 77.6 %, MGS = 180 nm. Bar = 200 nm. This effect is even more obvious if the data are plotted in the form of sintering trajectories (Fig. 5), i.e. the relations between the mean grain sizes and relative densities of sintered materials. The reference material A exhibits a typical sintering curve with exponential growth of the grain size in the final stage of sintering, albeit shifted to lower values in comparison to conventionally sintered alumina. The sintering curve of doped materials is quite flat, showing almost linear dependence between the grain size and relative density up to the p, < 98 % (99.6 % for AM). Apparent decrease of density and abnormal increase of grains size of the specimen AM after 3 min SPS at 1250 °C is believed to be accidental and caused most likely by improper packing of the powder in graphite die. However, these results cannot be taken as the unambiguous evidence that the grain growth is suppressed entirely: we did not succeed in preparation of fully dense material from any doped powder, except of AM, where the ρτ = 99.6 % was achieved. In all cases the content of residual porosity was > 2 %, and majority of grain growth is known to take place when the last 3 % of porosity is eliminated 10' ". One therefore cannot exclude that additional grain growth will take place at longer sintering times, or higher temperatures. In any case, highly dense doped materials with low fraction of closed residual porosity represented a promising starting position for complete densification by HIP.

196



Fig. 2 The microstructure development of SPS densified specimens A: a) 1150°C/lmin, p, = 86.6 %, MGS = 200 um, b) 1200°C/3min, pr = 97.3 %, MGS = 480 nm, c) 1250 °C/6min, pr = 98.9 %, MGS = 900 nm. Comparison with the AY material (d) densified 6 min at 1300 °C, pr = 98.0 %, MGS = 300 nm. Hot isostatic pressing The results of HIP experiments are summarized in Fig. 6 - Fig. 9. The conditions of HIP were selected as mild as possible to attain full densification, but if possible, to avoid grain growth. The SEM micrographs of fracture surfaces of the reference materials A spark plasma sintered for 3 minutes at 1150 °C (the mildest conditions facilitating to achieve the stage of closed porosity prerequisite for successful HIP) are shown in Fig. 6. Both the relative density and grain size increased markedly with the increasing temperature of isothermal heating. Although 1050 and 1100 °C was too low for complete densification, increase of density accompanied by only negligible grain growth could be observed. At 1150 °C the mean grain size 320 nm (18 % increase in comparison to the SPS material) corresponded to the relative density 98.9 % (7.3 % increase). Fully dense specimens could be prepared at temperatures > 1150 °C, but at these temperatures also rapid grain growth was observed, so that coarse grained microstructure (the mean grain size 2 um) resulted after HIP at 1250 °C. Similarly to SPS, refinement of the microstructure was achieved in doped specimens. The sample AY spark plasma sintered for 3 minutes at 1300 °C achieved full density by HIP at 1250 °C at the mean grain size of only 490 nm. Interestingly, the grain size is similar to the material A with the mean grain size of only 470 nm HIP-ed at 1150 °C, which was the lowest temperature facilitating complete elimination of residual porosity. The results indicate that both in doped and undoped PCA the grain boundary diffusion (densification) and grain boundary mobility (grain growth) are thermally activated processes, and a temperature interval can be found where densification is already activated, while grain growth is not. The dopants influence the temperatures where either process is activated, and shift them to higher values.


· 197


Fig. 3 Time dependence of relative density of sintered specimens after SPS. 1000

*

'

800. -B-1150°C -*-1200°C 600 -A-1250°C 400 200 000

'

1

T

!-: I " -

*

0-

-■

i

, ^ - · — = AZ:

•

-#-1200°C ▲ 1250°C -ψ-1300°Ο

-A-1250°C 600 -ψ-1300°Ο

200-

.

-#-1200°C -A-1250°C

1

AY: 800 -#--1200°C

400

AM:

. -■-1150 °C

*sllll^= i

PÜ-

t/min

---i .

Fig. 4 Time dependence of the mean grain size of SPS densified specimens.

198



800

600.

A: ■ • A D

■ 1150°C 1200 °C 1250°C Pressureless sintering

'

. ■

A A· λ

"°. '*\\

200.

600

■

■

400.

800.

tíA AM: jj. 1150°C S • 1200°C A 1250°C CM * ▲

.

AY: • 1200°C ▲ 1250°C ψ 1300°C

■

•

*·

AZ:

•

A T

1200°C 1250°C 1300°C

400 200

0H 70

.

· ,

• * , 80

* * .

1 90

T

T ,

• 1 100

,

.

« , 80

. , —

■■,.

s.

• * ♦♦ .

1 90

.

1 100

Pr/% Pr/% Fig. 5 Sintering trajectories of SPS densified aluminas. Fig. 7 summarizes the relative densitities of HIP-ed specimens as the function of the temperature of isothermal heating during HIP. Densification is fastest in the reference material A, where 1100 °C is sufficient to attain relative densities close to the theoretical value. Naturally, the initial density after SPS plays a significant role. Although the addition of MgO required by about 50 °C higher temperatures to attain the stage of closed porosity during the SPS in comparison to the reference A, no such effect was observed during HIP, and fully dense material AM could be prepared at 1100 °C. Materials AY and AZ behaved in a different way. The two curves shown in Fig. 7 for each composition represent the densification during HIP of materials with residual open porosity (SPS 1250 °C / 3 min) and those sintered to the stage of closed porosity (SPS 1300 °C / 3 min). The former could not be, quite obviously, sintered to full density due to the counter-pressure of Ar gas in gradually closing pores, the latter shows marked influence of applied hydrostatic pressure on densification. Complete elimination of residual porosity was achieved after 3 h HIP at 1300 °C. The grain size development with the temperature of HIP is summarized in Fig. 8. In the reference A and in MgO-doped material AM only limited growth of grains is observed at temperatures < 1100 °C. Above this temperature the grain growth is more pronounced, achieving the values 4 times higher than in the respective SPS samples and up to 1 0 - 1 2 times higher than the initial particle size of the alumina powder. Quite obviously the temperature 1250 °C is unnecessary high and results in excessive growth of alumina grains. The Y2O3 and ZrCh-doped specimens in turn exhibit nearly no grain growth, and even after 3 h at 1250 °C the mean grain size is at about the same level as in the respective SPS specimens. The data summarized in the previous two paragraphs are easier to understand if they are plotted in the form of sintering trajectories (Fig. 9). In the reference A full density can be achieved without significant grain growth (the points are marked by the rectangle in the Fig. 9). After that, if the temperature of HIP is unnecessarily high, grain growth is observed, which results in fully dense, but coarse grained microstructure. Fully dense specimens with the grain size less than 500 nm can be prepared under the following conditions:


· 199


Fig. 6 Microstructure development of material A (spark plasma sintered 3 min at 1150 °C) with the temperature of HIP. a) 1050 °C, pr = 96.2 %, MGS = 270 nm, b) 1100 °C, pr = 98.9 %, MGS = 320 nm c) 1150 °C, pr = 100 %, MGS = 470 nm d) 1250 °C pr = 100 %, MGS = 2 μπι. A comparison with the material AY (SPS, 3 min at 1300 °C) e) 1050 °C, pr = 97.9 %, MGS = 380 nm f) 1250 °C, pr = 100 %, MGS = 490 nm.

200



100.

fl

O

t

a■

■

95.

■

1150°C/3\ 1150°C/6' 1200 ° C / 3 ' ' 1200°C/6' 1250°C/3'

D

• O ▲

90.

100A

A

A

•

•

■

•

■

•

■

■

AM: ■ 1200 ° C / 3 ' • 1250 °C/ 3'

A

A

■

■

•

A:

■

•

B

B

A

A

A

A

A

•

•

■

•

95.

90

•

85- . i / /

•

•

AZ: • _ A

AY:

• ▲

1100

1150

1250 ° C / 3 ' 1300 °C/ 3'

1200

1250

—-τ//—

1050

1100

,/°c

T

HIP

1150

1250°C/3'· 1300 ° C / 3 '

1200

1250

1300

/°C

Fig. 7 Relative density of studied specimens as a function of the temperature of hot isostatic pressing. The legend shows the conditions of SPS for the respective specimens. A:

■

D • O A

2500 Q

AM:

1150 C / 3 ' 1150°C/6' 1200 °C/ 3' 1200 °C/ 6' 1250 °C/ 3' A

*

e

2

1200 °C/3' 1250 °C/ 3'

-TIÍC

■Hl·-

AZ: • A

AY: 2000 • 1250 °C/ 3' A 1300 °C/ 3' 1500

■

1

.

1—

1250 °C/ 3' 1300 ° C / 3 '

1000-

1050

1100

1150

T

/°C

HIP

1200

1250

1050

1100

1150

T

/°C

HIP

1200

1250

1300

Fig. 8 The mean grain size of studied specimens as a function of the temperature of hot isostatic pressing. The legend in the figure summarizes the applied conditions of the SPS process. The points before the x-axis break represent grains size of the respective SPS sample.


· 201

Influence of Minor Additives on Densificaron and Microstructure of Alumina Ceramics

4 AM: Ά •■

A: 20001500. 1000.

■ D • O ▲

1150 ° C / 3 ' 1150°C/6' 1200 °C/ 3' 1200 ° C / 6 ' 1250 ° C / 3 '

*.

■

1500

AY: • A

■

■

■

500.

2000'

1200 ° C / 3 ' 1250 °C/ 3'

- i

•

I 6 i

•

■ ■ AZ:

• A

1250 °C/ 3' 1300 °C/ 3'

-

1250 °C/ 3' 1300 °C/ 3'

•

1000500-

0-

HI»» 86

88

···

90

·

92

94

96

98

1

»

100 86

pr/%

« * ♦ *

./ 88

90

92

94

96

98

100

pr/%

Fig. 9 Sintering trajectories of HIP-ed materials. (1) Closed porosity must be attained after the SPS, but the temperature of SPS must not exceed 1200 °C. Above this temperature the SPS itself results in undesirable microstructure coarsening. (2) The temperature of HIP must not exceed 1150 °C. Above this-temperature rapid grain growth takes place. Complete densification is quickly achieved, and after that the surface energy of polycrystalline material can be further reduced only by grain growth. (3) For the respective system the temperature interval between 1150 and 1200 °C seems to be the border where the mechanisms responsible for grain growth take the control over the mechanisms responsible for densification. The situation is similar in the MgO-doped system, with some suppression of the grain growth observed at 1250 °C in comparison to the A reference. In the Y2O3 and ZrCb-doped aluminas this temperature interval is shifted to higher values. For all temperatures applied during HIP only negligible grain growth is observed, and full density is attained at the mean grain size of about 500 nm. The dopants might possibly also extend the interval where alumina can be sintered without, or with only limited grain growth, but to verify this hypothesis further experiments are required. It is also quite possible that at higher HIP temperatures (e.g. 1400 °C) similar microstructure development will be observed as in the reference A, i.e. significant grain growth once full density is achieved. CONCLUSIONS Densification and microstructure development of reference pure polycrystalline alumina A and 500 ppm MgO, Ζ1Ό2 and Y203-doped aluminas AM, AZ and AY by spark plasma sintering, followed by hot isostatic pressing was studied in the present work. Although SPS resulted in some microstructure refinement in comparison to conventionally sintered samples, the temperatures above 1200 °C resulted in grain coarsening without complete densification in the reference material A.

202



Addition of dopants led in all cases to some microstructure refinement, especially in case of Ζ1Ό2 and Y2O3, but this effect was accompanied by slower densification. In general, the dopants decelerated both the grain growth and densification, and the temperatures by about 50 - 100 °C higher in comparison to the reference A were required to achieve the same microstructure characteristics. Complete densification could not be achieved by SPS and in all cases the residual porosity > 2 % was present. The porosity was successfully eliminated by subsequent HIP in the temperature range 1100 - 1250 °C. Fully dense specimens with the mean grain size < 500 nm were prepared by HIP of the reference A at 1150 °C. The HIP temperatures > 1200 °C led to quick attainment of full density followed by rapid grain growth (the mean grain size up to 2 urn). The temperature interval between 1150 and 1200 °C seems to be the border where the mechanisms responsible for grain growth take the control over the mechanisms responsible for densification in the reference material A. In the Y2O3 and ZrCvdoped aluminas this temperature interval was shifted to higher values. Only negligible grain growth was observed, and full density was attained at the mean grain size of about 500 nm by HIP at 1250 °C in materials AY and AZ. ACKNOWLEDGEMENT The financial support of this work by the grant A P W 0485-06, the Slovak National Grant Agency VEGA, grant No 2/6181/26, the Alexander von Humboldt Foundation (D. Galusek), and by Maria Curie Research Fellowship Program (J. Sedlácek) is gratefully acknowledged. REFERENCES ' A. Krell, S. Schädlich, Nanoindentation hardness of submicrometer alumina ceramics, Mater. Sei. Eng., Α30Ί, 172-181 (2001). 2 A. Krell, P. Blank, Grain size dependence of hardness in dense submicrometer alumina, J.Am.Ceram.Soc, 78, 1118-1120 (1993). 3 K. Morinaga, T. Torikai, K. Nakagawa, S. Fujino, Fabrication of fine α-alumina powders by thermal decomposition of ammonium aluminum carbonate hydroxide (AACH), Acta Mater., 48, 4735-4741 (2000). 4 A. Krell, P. Blank, The influence of shaping method on the grain size dependence of strength in dense submicrometre alumina, J.Eur.Ceram.Soc, 16, 1189-1200(1996). 5 Y.T. O, J. Koo, K.J. Hong, J.S. Park, D.C. Shin, Effect of grain size on transmittance and mechanical strength of sintered alumina, Mater.Sci.&Eng., A374, 191-195 (2004). 6 A. Krell, D. Klaffke, Effects of grain size and humidity on fretting wear in fine grained alumina, AI2O3/T1C and zirconia,/.Λ/η. Ceram.Soc, 79, 1139-1146 (1996). 7 G.C. Wei, Transparent ceramic lamp envelope materials, J. Phys. D: Appl. Phys., 38, 3057-3065 (2005). 8 A. Krell, P. Blank, H. Ma, T. Hutzler, M.P.B. van Bruggen, R. Apetz, Transparent sintered corundum with high hardness and strength, J.Am.Ceram.Soc, 86, 12-18 (2003). 9 R. Apetz, M.P.B. van Bruggen, Transparent Alumina: A Light-Scattering Model, J.Am.Ceram.Soc, 86,480-486 (2003). 1 L.C. Lim, P.M. Wong, M.A. Jan, Microstructural evolution during sintering of near-monosized agglomerate-free submicron alumina powder compacts, Acta Mater., 48, 2263-2275 (2000). 1 ' C. Nivot, F. Valdivieso, P. Goeuriot, Nitrogen pressure effects on non-isothermal alumina sintering, J.Eur.Ceram.Soc, 26, 9-15 (2006). 12 Z. Shen, H. Peng, J. Liu, M. Nygren, Conversion from nano- to micron-sized structures: experimental observations, J.Eur. Ceram.Soc, 24, 3447-3452 (2004). 13 J. Echeberria, J. Tarazona, J.Y. He, T. Butler, F. Castro, Sinter-HIP of alpha-alumina powders with sub-micron grain sizes, J.Eur.Ceram.Soc, 22, 1801-1809 (2002).


· 203


14

S. H. Risbud, C.-H. Shan, A.K. Mukherjee, M.J. Kim, J.S. Brow, R.A.Holl, Retention of nanostructure in aluminum oxide by very rapid sintering at 1150 °C, J.Mater.Res., 10 [2] 237-239 (1995) 15 L. Gao, J.S. Hong, H. Miamoto, S.DD.L. Torre, Bending strength and microstructure of AI2O3 ceramics densified by spark plasma sintering, J.Eur.Ceram.Soc, 20 2149-52 (2000) 16 R.S. Mishra, A.K. Mukherjee, Electric pulse assisted rapid consolidation of ultrafine grained alumina matrix composites, Mater.Sci. & Eng., A287 178-182 (2000) 17 S.W. Wang, L.D. Chen, T. Hirai, Densification of AI2O3 powder using spark plasma sintering, J.Mater.Res., 15 [4] 982-987 (2000) 18 L.A. Stanciu, V.Y. Kodash, J.R. Groza, Effects of heating rate on densification and grain growth during field-assisted sintering of (1-AI2O3 and M0SÍ2 powders, Metall & Mater. Trans. A, 32A 26332638 (2000) " Z. Shen, M. Johnsson, Z. Zhao, M. Nygren, Spark plasma sintering of alumina, J.Am.Ceram.Soc, 85 [8] 1921-27 (2002) 20 Y. Zhu, K. Hirao, Y. Yamauchi, S. Kanzaki, Densification and grain growth in pulse electric current sintering of alumina, J.Eur.Ceram.Soc, 24 3465-3470 (2004) 21 D.T. Jiang, D.M. Hulbert, U. Anselmi-Tamburini, T. Ng, D. Land, A.K. Mukherjee, J.Am.Ceram.Soc, 91 [1] 151-154(2008) 11 D. Chakravarty, S. Bysakh, K. Muraleedharan, T.N. Rao, R. Sundaresan, Spark plasma sintering of magnesia-doped alumina with high hardness and fracture toughness, J.Am.Ceram.Soc, 91 [1] 203208 (2008) 23 M.I. Mendelson, Average Grain Size in Polycrystalline Ceramics. J. Am. Ceram. Soc, 52, T39 -T42 (1969).

204


THE ELECTRO-DISCHARGE COMPACTION OF POWDER TUNGSTEN CARBIDE - COBALT DIAMOND COMPOSITE MATERIAL Evgeny G. Grigoryev and Alexander V. Rosliakov The manufacturing of high strength structure of powder tungsten carbide - cobalt - diamond composite material is investigated and optimal operating parameters are defined. The structure of tungsten carbide - cobalt - diamond composite material reaches top strength at certain magnitudes of applied pressure and high voltage electrical discharge parameters. We studied densification process of powder material during electro-discharge compaction by means of high-velocity filming. Pulse current parameters were measured by Rogovsky coil. The temperature evolution during electro-discharge compaction process was measured by means of thermocouple method. We have installed that the powder densification process has wave nature in electro-discharge compaction. We defined wave front velocity of densification process and pressure amplitude in wave front subject to parameters of electro-discharge compaction. INTRODUCTION Tungsten carbide and diamond are well known for its exceptional hardness and wear/erosion resistance. Cemented carbides and diamond are used extensively in the industry involving high wear, abrasive applications. Apart from their exceptional hardness, WC has other unique properties such as high melting point, high wear resistance, good thermal shock resistance, thermal conductivity and good oxidation resistance [1,2]. WC with ductile metals such as cobalt as a binding medium, which is known to be helpful in cementing fine WC particles, is used in bulk sintered forms. Matrices of ductile metals, such as cobalt, greatly improve its toughness, hence elimination the possibility of brittle fracture during operation. WC-Co-diamond composites are extensively used to enhance the wear resistance of various engineering components, e.g. cutting tools and dies. There have been a large number of works performed on producing composite materials from powders by nonconventional powder consolidation methods in which densification is enhanced by the application of a pulsed electric current combined with resistance heating and pressure: plasma pressure compaction (P2C), spark plasma sintering (SPS), plasma activated sintering (PAS) [3, 4]. The interest in these methods was motivated by their ability to consolidate a large variety of powder materials to high densities within short periods of time and without having to increase grain sizes. In this paper, we focused on the consolidation of WC-Co and WC-Co-diamond powders into a solid bulk without increasing their crystallite sizes by electro-discharge compaction (EDC) [5, 6, and 7]. The principle of EDC is to discharge a high-voltage (up to 30 kV), high-density current (-100 kA/cm2) pulse (for less than 300 μβ) from a capacitor bank through the powders under external pressure. In this way, a full or near full densification may be achieved with minimal undesirable microstructural changes due to short consolidation time. Furthermore, WC-Co-diamond powders could be consolidated into solid bulks by electro-discharge compaction (EDC) with densities close to theoretical density. This method has demonstrated the possibility of consolidating difficult-to-sinter powders to provide distinct technological and economical benefits. EXPERIMENTAL PROCEDURES Dense WC-Co and WC-Co-diamond composites were fabricated by an EDC method. In this process, the WC-Co and WC-Co-diamond mixed powders were poured into an electrically non-conducting ceramic die. The ceramic die was plugged at two ends with molybdenum electrodes-punches and an external pressure up to 400 MPa was applied to the powder on air-operated press. A high voltage

205

Electro-Discharge Compaction of Powder Tungsten Carbide-Cobalt-Diamond Composite

capacitor bank was discharged through the powder. The schematic of the Electro-Discharge Compaction (EDC) system is shown in Figure 1.

Figure 1. Schematic of the Electro-Discharge Compaction (EDC) system 1 — powder, 2 — die, and 3 — the upper and lower punches electrodes Electro-discharge compaction (EDC) apparatus for compacting the powders consists basically of a bank of capacitors; charging unit and trigatron switch to connect a powder column suddenly across the charged capacitor bank. The capacitor bank consists of thirty 200 μΡ capacitors that can store up to 6 kV. EDC uses the pulse current generated from the capacitor bank to quickly heat a powder column subjected to constant pressure. During electro-discharge compaction, we recorded, using the Rogowski coil, the parameters of the current pulse through the powder column. An oscillograph showing a typical output from the Rogowski coil is shown in Figure 2. The temperature evolution during electro discharge compaction process was measured by means of thermocouple method. We measured the temperature on the powder sample surface, using ChromelAlumel and tungsten-rhenium thermocouples. To study the powder consolidation kinetics, we used a high-speed SKS-1M camera, which makes it possible to record movies with a frequency up to 4* 103 frames s~K The high-speed consolidation kinetics of the WC-Co powder was experimentally investigated on samples with a diameter of 10 mm and a mass of 12.0 χ 10~3kg for different values of the current pulse parameters. The following commercial WC-Cc—diamond powder was used as starting material for electrodischarge compaction: WC - -80 wt.%, Co - 20 wt.%, free carbon 0.101 wt.%, total oxygen 0.13 wt.%, grain size WC < 5μπι, grain size of diamond < 40μιη. Powder column was a circular cross-section rod with a diameter of 10 mm and a length ranging from 10 to 15 mm. X-ray diffraction (XRD) was performed on the as-received powder using a DRON-3 diffractometer with a Cu target for 2Θ from 20° to 120° at a scan speed of 17min. XRD was repeated on the consolidated specimens, followed by density measurements. Density measurements were taken after EDC process using the Archimedes principle in distilled water.

206



Figure 2 Typical pulse current traces from registration system (Rogowski coil) (Peak currents: 1 - 50 kA, 2 - 80 kA, 3 - 110 kA) EDC applies a high-voltage, high-density current pulse to the WC-Co-diamond powders under external pressure for a very short period of time. This method uses the passage of the pulse electric current to provide the resistive heating of the powder by the Joule effect. Joule heating occurs at the inter-particle contact to instantaneously weld WC-Co powder particles, resulting in densification. The achieved WC-Co-diamond composite density as a result of EDC process depends on applied external air-operated pressure, magnitude and waveform of pulse current that depends on RLC - parameters of the electrical discharge circuit. RESULTS AND DISCUSSION The measurements of the current pulse parameters with a Rogowski coil showed that the discharge current pulse width in all experiments did not exceed 300 μβ. This value determines the time of energy injection into the powder. Measurement of the time dependence of the surface temperature of the sample showed that the characteristic time of its cooling of about 2 s. Analysis of the problem parameters made it possible to reveal the hierarchy of the characteristic times of the processes occurring under electric-pulse pressing. The approximate general scheme is as follows. A current pulse passing through a powder and punches strongly heats only the powder material without significant punches heating, because the powder resistivity greatly exceeds that of the punches material. The intense heating of the powder significantly decreases its resistance to plastic deformation, and, under the action of an external mechanical pressure, it is consolidated with a characteristic rate, dependent on the pneumatic system type. Simultaneously, heat sink from the powder to punches and die occurs due to the thermal conduction. The time of energy injection to the powder is determined by the current pulse width: το < 10"3 s. The time of formation of a consolidated material from the powder, τι, depends on the loading system and lies in the range 2 x 10~3 < Xi < 2 x 10"2 s. The cooling time of the pressed material, τι, is determined by the thermal conductivity of the material and the characteristic size of the pressed sample: t2 = 2.5 s. In this case, the time scales of the processes obey the following relation: το < τ\ « Τ2. The most important factor which determines the success of the EDC process is the instantaneous current density in the powder column [7]. We studied influence of pulse current density and external


· 207


pressure on kinetics of powder material densification at electro discharge compaction. Consolidation of powder material takes place at constant pressure P, created pressure system during all process of the electro discharge compaction. Time dependences of powder density during EDC process are resulted in Figure 3. P, % 100.0 92.5

85.0 77.5

70.0

62.5

0

4

8

12

16

t, ins

Figure 3 Variations of relative density as a function of time in EDC process at constant external pressure (external air-operated pressure 200 MPa) On a Figure 3 curves: 1 (75 kA/cm2), 2 (85 kA/cm2), 3 (97 kA/cm2), were got at constant pressure P = 200 MPa. Duration of process of powder material densification is range from 6 ms to 16 ms for all our experiments. The results of experiments show that motion of punches in the process of the electro discharge compaction takes place with steady speed. The value of speed depends on amplitude of pulse current and external pressure. The magnitude of external pressure determines initial specific resistance of powder column and, accordingly, amount of heat, selected in powder material. With the increase of pressure the specific resistance of powder column goes down sharply, that results in the less heating of powder material. The densification of the compacted material takes place due to an intensive plastic strain which depends on external pressure P and yield stress of powder matter - στ(Τ) (Τ temperature). Therefore speed of plastic flow of the compacted powder material and consequently speed of change of density of the compacted powder column is determined by the temperature at EDC process. The speed of densification depends on a dimensionless parameter β = στ(Τ)/Ρ. At constant amplitude of pulse current a densification speed is determined by temperature dependence of yield strength at different pressures, put to the powder column. The finishing density (Fig. 3) of a compacted material after EDC is defined by size of the external pressure and parameter β. Electro discharge compaction of a powder material represents steady state wave in the powder, created by a punch moving with constant speed. When the compaction wave reaches the second stationary punch, process of powder compaction is completed. It is shown on Fig. 3 by horizontal lines of density dependence vs the time. Duration of this compaction process is less than 20 milliseconds. Cooling of powder compact occurs during 2 second due to thermal conductivity of the punches and the die. The conventional powder consolidation processes (hot pressing, etc.) have no place because of short duration of EDC.

208



The XRD results show that the main phases of WC-Co powder were WC and α-Co, but the main phases of WC-Co compact sample after EDC were WC and ß-Co. This is associated with a rapid cooling of WC-Co compact sample in the time of EDC process. CONCLUSIONS The response of WC-Co-diamond powder composites (loaded external pressure) to high energy electrical discharge was studied and has been described and understood in terms of peak current density and external pressure. It was found that the density of composite material reach its maximum values at certain magnitudes of applied pressure and high voltage electrical discharge parameters. It should be noted that kinetics of process of densification of powder material at EDC substantially differs from kinetics at P2C, PAS and SPS [4, 5]. It is related to distinction in speed of input energy and in the parameters of influence by external pressure on powder material. It is higher speed of input energy for EDC process and volume distribution of input energy dependence on external pressure. The lowest value of pressure provides the greatest allocation of energy and rise in temperature of compacted material because of the loose powder has the worsest metallic contact and with it the most sparking - local melting. Attempts to compact WC-Co-diamond powders by EDC method presage future fruitful results of development types of tools. REFERENCES [1] Erik Lassner and Wolf-Dieter Schubert: Tungsten-Properties, Chemistry, Technology of the Element, Alloys, and Chemical Compounds, 2000, Kluwer Academic/Plenum Publishers. [2] Tungsten and Tungsten Alloys-Recent Advances, Edited by Andrew Crowson and Edward S. Chen, 1991, Minerals, Metals and Materials Society. [3] V. Mamedov, "Spark Plasma Sintering as Advanced PM Sintering Method" Powder Metallurgy, 2002, vol. 45, no. 4, pp. 322-328. [4] Z. A. Munir, U. Anselmi-Tamburini, M. Ohyanagi, "The effect of electric field and pressure on the synthesis and consolidation of materials: A review of the spark plasma sintering method" J. Mater. Sei., 2006, vol.41, pp. 763-777. [5] J. R. Groza, A. Zavaliangos, "Nanostructured Bulk Solids by Field Activated Sintering", Rev. Adv. Mater. Sei., 2003, vol. 5, pp. 24-33. [6] M. Shakery, S. T. S. Al-Hassani, T. J. Davies, "Electrical Discharge Powder Compaction", Powder Met. Int., 1979, vol. 11, no. 3, pp. 120-124. [7] E. Grigoriev, A. Rosliakov, "Electro Discharge Compaction of WC-Co Composite Material Containing Particles of Diamond", Materials Science Forum, 2007, Vols. 534-536, p. 1181-1184.


· 209


MICROWAVE SINTERING EXPLORED BY X-RAY MICROTOMOGRAPHY Kotaro Ishizaki, Manjusha Battabyal, Yoko Yamada Pittini, Radu Nicula and Sebastien Vaucher EMPA, Swiss Federal Laboratories for Materials Science and Technology, Laboratory for Materials Technology Feuerwerkerstrasse 39, Thun, CH-3602 Switzerland ABSTRACT Microwave energy is seen as offering potential energy and time savings for the processing of materials. For metal-diamond composite for example, selective heating of the metallic grain, could preserve the diamonds from allotropic conversion to graphite while the matrix could be brought to its sintering temperature. However, the interactions of microwave with random metal-dielectric mixtures are complex and not yet adequately described by any model, and the control of the process is difficult. Here we investigate the effect of microwave irradiation on a diamond-aluminium powder mixture. Using 3D X-ray microtomographic analysis, we show that for the same starting material, treated in similar conditions, very different microstructural changes inside of the powder beds can be found. The possibility to access extensively 3D microstructural information opens up new opportunities for the understanding and control of microwave processing of metal-dielectric systems. INTRODUCTION Due to its extremely high thermal conductivity and hardness properties, together with present availability and affordability of synthetic diamonds, the use of diamond is considered as first choice material in many industrial fields. Its low coefficient of thermal expansion (CTE) makes diamond the ideal filling material for metal matrix composites featuring high thermal conductivity and reasonably low CTE. Such composites are considered for use in heat sinks and heat spreaders. Recently also diamond metal composites have drawn attention due to their enhanced thermal conductivities, up to 670 W/mK, makes the composites attractive for demanding applications in microelectronics and semiconductors1. Metal diamond contacts are essential for almost any type of electromechanical sensors and electronic devices.1"3 The excellent thermophysical properties of diamond can only be exploited to the whole extent when an optimal interface between the diamond particles and the metal matrix is obtained. However, due to its high hardness property the bonding of diamond particles with metal is an extremely difficult task and has become an open and challenging area for researchers. Aluminum-diamond composites have been obtained by the metal infiltration method.1"3 According to our knowledge; the present work is the first attempt to produce aluminum-diamond composites using microwave irradiation. Microwave heating is fundamentally different from conventional sintering in that the energy is deposited volumetrically rather than relying in thermal conduction from the surface. Properly manipulated this may lead to a number of benefits, including improved product properties and reductions in manufacturing costs due to energy savings and shorter processing times.4 The reactivity and wettability of carbon with aluminum have been widely studied in the past decades since carbon fibers reinforced aluminum and aluminum alloys are an important group of metalmatrix composites. " When using diamonds however, cross sectioning becomes extremely difficult, and microtomography is of great help to visualize detailed morphological changes. 13,14 In the present study, synthesis of aluminum-diamond composites are prepared by microwave heating. The striking difference of morphology obtained for similar samples experimental conditions is analyzed using synchrotron- based X-ray tomographic microscopy.

211


EXPERIMENTAL Diamond (synthetic uncoated diamond monocrystals, MBD8 grade, 140/170 mesh, Qming Superhard Material Co. Ltd) and Aluminum powder (65 μιη, GTV GmbH,) were placed in a 60:40 weight ratio for 64 hrs in a Turbula™ to obtain a homogenous distribution of powders. Two similar samples were prepared by filling 3 cm of a 012 mm quartz crucible embedded in a porous alumina thermal insulation (packing density 1.52-1.77 g/cm3). The crucibles were introduced in the center of a resonating TEioi microwave cavity and irradiated with a constant incident power of 600 W/2.45 GHz (Dipolar AB). The incident and reflected power was recorded using a high power impedance analyzer (S-TEAM STHT) .Resonant conditions were maintained using 3 stubs impedance matching as well as sliding short adjustments. The temperature was measured with an infrared pyrometer (Raytek 3.62-3.94 μπι) facing the top of the crucible. The schematic illustration of the microwave assembly is shown in figure 1.

Figure 1. Schematic illustration of the microwave chamber. After microwave irradiation, the samples were analyzed by synchrotron-based X-ray microtomography (X02DA-TOMCAT beamline, Swiss Light Source (SLS), Switzerland) with a radiograph resolution of 5.6 x 5.6 μιη2. After 3D-reconstruction, Amira™ was used for data visualization. RESULTS AND DISCUSSION The evolution of temperature and absorbed microwave power are shown in figures 2 and 3 for two similar samples (A and B) submitted to a constant incident power of 600W. Although the microwave power supplied to both samples is identical, the reaction of the two samples is largely different. For sample A the temperature rises in the first 150 seconds to reach 1450°C then falls to 1200°C (350 s). With a constant values in the interval 50-150 seconds and a regular decay until 350 seconds, the microwave absorption profile is consistent with the pyrometer signal. The variations are found smooth and continuous for both signals. The situation of sample B is completely different. The temperature and the absorption were unstable during all the heating period. During the initial 350 seconds the temperature do not exceeds 600°C while after 500 seconds, the averaged temperature finally reaches about 1100°C due to the development of plasma above the sample top.

212



Figure 2. Time-temperature profile of sample A showing supplied and absorbed microwave powers.

Figure 3. Time-temperature profile of sample B showing supplied and absorbed microwave powers.


· 213


Figure 4. Tomographie reconstruction of various cross sections obtained for aluminum-diamond composite sample A.

Figure 5. Tomographie reconstruction of various cross sections obtained for aluminum-diamond composite sample B.

214



Tomography images of aluminum-diamond composites samples A and B are shown in figures 4 and 5, respectively. The figures include an over view of the sample after heating and the cross section views taken by X-ray microtomography. The over view image shows in which angle the cross section views were taken. Images a and b are vertical sections and c to g are horizontal sections. Black areas of these images represent porosities, dark gray, diamond and light gray aluminum. In the lower part of both samples (Fig 4 e-g and 5 e-g), localized aluminum coalescence often surrounding a closed pore suggests that melting of initial aluminum particles occurred all trough the samples. This change seems dominated by capillary forces leading to a dewetting of the diamond surface by the liquid metal. In the upper part of sample A (Fig 4a and 4b) a domain is present in the center within which diamond are completely wetted by aluminum. This localized densification has resulted in the creation of large porosities due to outwards mass transports from neighboring areas. Higher magnifications of this area (figure 6) underline the different microstructure of the composite in close vicinity. The 3D image I shows that the molten aluminum is not perfectly attached with the diamond. The small particles of aluminum were molten and assembled in a larger drop. During this agglomeration diamonds were expelled away, either due to the presence of alumina or to a too low temperature. In image II the distribution of diamonds inside of the molten aluminum is homogeneous and there are no gaps between both materials. Image III is similar as the image of initial situation of this composite. Aluminum particles are in the same size as at the beginning. The formation of dense aluminum-diamond composite is only produced in the top of sample A. The different response of sample A and B to the same incident microwave power lead to two different heating patterns and different microstructure. The stable microwave absorption in sample A, may be due to a higher and stable local electric field inside the sample. The instable behavior of sample B leading to a plasma outside of the sample is unpredictable.

Figure 6. 3D images of three different portions of sample A from figure 4 a.


· 215


A possible explanation for the observed results may be that if the heating area is different inside of the sample, the size of molten aluminum is different. If the aluminum is melting slowly at low temperature the particle of aluminum will increase in size but without wetting the diamond because of the surface tension of molten aluminum is high; as it is the case in sample B. On the other hand, with a temperature profile like the one of sample A, the heating rate was fast and probably produced a large area of localized high temperature. If the temperature is high enough to reduce the surface tension of molten aluminum, it will make possible to wet the diamond surface by aluminum and a composite as the one obtained in portion II of sample A is possible to be produced. Once the diamond and aluminum are in contact, the interface AI4C3 is possible to be created at this temperature range. Another important factor is that diamond is transparent for microwave heating. To make this type of composite it will take long time by conventional heating. CONCLUSION For these two samples the composition is the same but localized geometry and distribution of local intern fields are not equal. This difference of local intern fields may induce a different way of absorption in the sample. Field distribution assumes continuum approximation, however; a mixture of non conductive diamond and metallic aluminum makes the discreet determination of the local field defiled and singular cases 3D X-ray microtomography is a powerful method to observed diamond composite materials. With this method we could find that in one portion of this composite; the diamond was completely wetted with aluminum and homogeneously distributed in the matrix. The findings open the possibility of optimizing the microwave processing conditions to obtain a desired kind of composite and homogeneous in the entire matrix. This heating can be used to heat volumetrically a metal-dielectric mixture. However, the control of local events remains to be solved to evolve a reproducible process. X-ray microtomography makes possible to reach this level of understanding which opens the possibility of optimizing the microwave sintering conditions to obtain a desired kind of composite. But the propagation and interaction of electro magnetic waves in a metal-dielectric mixture is not yet analyzed solved for random mixture or at high metallic volumetric fraction close and above percolation threshold. It would be desirable to develop a model of microwave interaction with random metal-dielectric mixtures. ACKNOWLEDGEMENT Kotaro Ishizaki would like to thank to COST (European Cooperation in the field of Scientific and Technical Research) for financial support to attend the international conference; Sintering 2008. The authors would like to thank Dr. Samuel McDonald for the technical support of TOMCAT at SLS. REFERENCES 1 P. W. Ruch, O. Beffort, S. Kleiner, L. weber, Selective Interfacial Bonding in Al(Si)-Diamond Composites and its Effect on Thermal Conductivity, Comp. Sei. Tech., 66, 2677-85 (2006). 2 0 . Beffort, F. A. Khalid, L. Weber, P. Ruch, U. E. Klotz, S. Meier, S. Kleiner, Interface Formation in. Infiltrated Al(Si)/Diamond Composites, Diamond and Related Mater., 15, 1250-60 (2006). 3 S. Kleiner, F. A. Khalid, P. W Ruch, S. Meier and O. Beffort, Effect of Diamond Crystallographic Orientation on Dissolution and Carbide Formation in Contact With Liquid Aluminium, Scripta Materialia, 55, 291-4 (2006). 4 J.G.P. Binner, B. Vaidhyanathan, Microwave Sintering of Ceramics: What does it offer?, Key Eng.

216



Mater., 264-268, 725-30 (2004). 5 M. Yang, V.D. Scott, Carbide Formation in a Carbon Fibre Reinforced Aluminium Composite, Carbon, 29,877-9(1991). 6 M. Desanctis, S. Pelletier, Y. Bienvenu, M. Guigon, On the Formation of Interfacial Carbides in a Carbon Fibre-Reinforced Aluminium Composite, Carbon, 32, 925-30 (1994). 7 C.P. Ju, K.J. Chen, J.H.C. Lin, Process, Microstructure and Properties of Squeeze-Cast Short-CarbonFibre-Reinforced Aluminium-Matrix Composites, J. Mater. Sei., 29, 5127-34 (1994). 8 H.D. Steffens, B. Reznik, V. Kruzhanov, W. Dudzinski, Carbide Formation in Aluminium-Carbon Fibre-Reinforced Composites, J. Mater. Sei., 32, 5413-7 (1997). 9 M. Gu, G. Zhang, R. Wu, Interfacial Bondings in Gr(f)/Al Composites, Prog. Nat. Sei., 7, 600-6 (1997). 10 K. Landry, S. Kalogeropoulou, N. Eustathopoulos, Wettability of Carbon by Aluminum and Aluminum Alloys, Mater. Sei. Eng., A 254, 99-111(1998). 11 R. Asthana, Review Reinforced Cast Metals, J. Mater. Sei., 33, 1959-80 (1998). 12 E. Pippel, J. Woltersdorf, M. Doktor, J. Blücher, H.P. Degischer, Interlayer Structure of Carbon Fibre Reinforced Aluminium Wires, J. Mater. Sei., 35, 2279-89 (2000). 13 D. Bernard, 3D Quantification of Pore Scale Geometrical Changes Using Synchrotron Computed Microtomograghy, Oil Gas Sei. TechnoL, 60, 747-62 (2005). 14 S. Vaucher, P. Unifantowicz, C. Ricard, L Dubois, M. Kuball, J.-M. Catala-Civera, D. Bernard, M. Stampanoni and R. Nicula, On-line Tools for Microscopic and Macroscopic Monitoring of Microwave Processing, Physica B, 398, 191-5 (2007).


· 217


PULSE PLASMA SINTERING AND APPLICATIONS Andrzej Michalski, Marcin Rosinski Warsaw University of Technology, Faculty of Materials Science and Engineering, Warsaw, Poland ABSTRACT The paper presents a new Pulse Plasma Sintering method (PPS), developed at WUT. Unlike other electric-field assisted sintering methods, the PPS method employs pulse high-current electric discharges for heating and activating the material to be sintered. The phenomena, taking place during the high-current pulses, which heat the powder during the PPS treatment and activate the sintering process, are similar to those occurring in other new techniques, but, in PPS, thanks to the much higher energy of the pulse discharge, these phenomena run much more intensively. The presentation will also contain some examples of application of PPS technique for producing sintered parts from a wide variety of materials. INTRODUCTION In modern sintering methods, such as PAS (Plasma Assisted Sintering) [1], SPS (Spark Plasma Sintering) [2, 3, 4] and FAST (Field Assisted Sintering) [5, 6], the sintering is carried out at lower temperatures and the process lasts for a shorter time than is the case in the conventional methods. A characteristic feature of these techniques is that the pulse current is used for heating powders to be sintered. During a current pulse, spark discharges are ignited in the pores. These discharges remove adsorbed gases and oxides from the powder particle surfaces, thereby facilitating the formation of active contacts between them. In effect, the process time can be shortened and the sintering temperature can be reduced. The paper gives some examples of producing sintered composite materials by the new PPS (Pulse Plasma Sintering) method. PULSE PLASMA SINTERING METHOD (PPS) The apparatus used for consolidating powders by the PPS method has been designed and constructed at the Faculty of Materials Science and Engineering, Warsaw University of Technology (Fig. 1). The PPS method utilizes pulsed high electric current discharges for heating the powder whilst it is being pressed. The current pulses are generated by discharging a 300 μΡ capacitor, charged to a voltage of maximum 10 kV.

Figure 1. Apparatus for PPS.

Figure 2. Schematic diagram of the PPS apparatus.

219


The material to be sintered is placed in a graphite die in-between two graphite punches (Fig. 2). The pulses have amplitudes of tens of kilo-amperes and periods of several hundred microseconds (Fig. 3). It is worth noting that the PPS process has a very high thermal efficiency. Because of the short duration of the electric pulses (0.6 ms) compared to the interval between them (1 s) the temperature achieved during the current pulses is higher than the temperature measured during the sintering process. Fig. 4 shows schematically the temperature variation during the PPS process.

Figure 3. Examples of the electric current Figure 4. Temperature variations during the highand voltage waveforms during the PPS, current PPS. a) current, b) voltage. The high-current pulses, which heat the powder during sintering by the PPS method, induce similar sintering process-activating mechanisms to those occurring in the PAS, SPS, and FAST processes, except that, thanks to the high energy delivered in a pulse (of the order of 320 MVA = 40 kA-8 kV), i.e. manifold greater than that achieved in the other techniques, the PPS processes are much more intensive. As seen in Table 1, the electric parameters Table 1. Electric Parameters of the PPS and SPS of the PPS processes differ substantially Sintering Processes from those used in the SPS processes. The Parametr PPS SPS [71 differences include the electric current Current (A) 40 000 5000 intensity, electric voltage, pulse duration, Voltage (V) 8 000 10 and the pulse repetition frequency. In PPS, Pulse duration (ms) 0.6 36* the instantaneous electric current (I) is Pulse repetition frequency (ms) 500 6* several times higher than in SPS, whereas the voltage (U) is several hundred times as high, so that the instantaneous electric power in PPS is about 320 MVA (UT) compared with 50 kVA in SPS. The PPS pulse duration is also about 60 times shorter and the inter-pulse interval is about 80 times as long as those in SPS. APPLICATIONS OF THE PPS METHOD The PPS method has been used for sintering a wide variety of materials, such as e.g. WC/diamond [8, 9] and Cu/diamond [10, 11] composites, nanocrystalline sintered parts [12-15] and with the participation of the SHS reaction for fabricating high-melting ceramics [16-18]. Diamond/WC Composites Intended for Cutting Tool Edges Diamond/cemented carbide containing 30 vol% of diamond particles was produced using a mixture of a 6 wt% Co added-WC powder, with a WC grain size of 0.8 μπι and a diamond powder with a grain

220



size ranging from 40 to 60 μπι. The mixture was sintered to produce a two-layer plate, 20mm in diameter, in which a layer (0.5 mm thick) composed of WC6C0 composite + 30% of diamond was formed on a WC6C0 substrate. The sintering process was conducted at a temperature of 1100 °C under a load of 60 MPa for 5 min. Figure 5 shows a diffraction pattern obtained for the diamond/cemented carbide sintered by PPS.

Figure. 5. The XRD patterns for a cemented carbide with diamond particles. The phases identified in the sintered composite were tungsten carbide, diamond, and a cobalt phase. Figure 6 shows an SEM micrograph of the surface of the sintered diamond/cemented carbide composite and Fig. 7 is a SEM image of the surface of a fracture through the cemented carbide/diamond composite.

Figure 6. SEM image of surface of the composite.

Fi

8- 7. SEM image of the surface of a fracture through the diamond/WC6Co composite.


■ 221


The diamond particles distributed uniformly within the cemented carbide matrix do not form a skeleton because they only constitute 30 % of the matrix volume. No graphite precipitates can be seen around the diamond particles and these particles are in very good contact with the WC6C0 matrix. We can see in the figure that, on fracture, only few diamond particles were torn out from the matrix, whereas most of them clove along the fracture plane. Table 2 compares the hardness of the diamond/cemented carbide composites produced by PPS and their cemented carbide matrix with the hardness of the composites and their matrix produced by the SPS method as reported by H. Moriguchi et al. [19]. Table 2. Properties of the Sintered Composites Produced by PPS and SPS Method Temperature Time Composition [min] 5 30 vol% diamond/WC-6 wt% Co PPS 1100 PPS 1100 5 WC-6 wt% Co 1400 3 20 vol% diamond/WC-10 wt% Co SPS 3 SPS 1400 WC-10wt%Co

m

Hardness [GPa] 23 20 18 17

Ref. 9 9 19 19

The hardness of the diamond/WC6Co composite consolidated by PPS is higher about 3 GPa than that of the cemented carbide without diamond particles and also higher than the hardness of the diamond/carbide sintered by H. Moriguchi et al. who used the SPS method [19]. The lower hardness values of the sintered composites produced by H. Moriguchi et al. may be explained in terms of the smaller diamond content and the lower hardness of the WClOCo matrix. It is however worth noting that, in order to avoid graphitization, H. Moriguchi et al. also pre-covered the composite with a thin SiC film before its sintering. Wear of the milling cutter edge has been studied through machining of melamine faced boards. Figure Tabele 3. Parameters of machining 8 shows milling cutter with edges of diamond/W6Co composite used for investigations of durability. Table 3 Speed of rotation 18000 rpm shows the parameters of machining of the milling cutter Speed of machining 1130m/min edge. Figure 9 shows decrement of the milling cutter Feed per tooth 0,25 mm/tooth edge as a function of a machining distance.

Figure 8. A milling cutter ended with a diamond/cemented carbide. The critical wear value was taken to be ΙΟΟμπι. The wear of the sintered carbide milling cutter measured after cutting along a length of 32m was 126μηι which is greater than the critical wear value, whereas the wear of the milling cutter with edges made of a diamond/W6Co composite reached the critical value after the cutting length exceeded about 128m. Hence, at the critical wear assumed to be

222



ΙΟΟμπι, the milling cutter edges made of a diamond/W6Co sinter can be run along the cutting distance 4 times as long as that available for the sintered carbide cutting edge.

E

E

32 -Sintered carbides

64 96 Machining distance Cm] -Diamond/WCSCo

160 — — The critical wea

Fig.9. Decrement of the milling cutter edge as a function of a machining distance. Diamond/Cu Composite Intended for Heat Sink Applications The diamond/Cu composite was produced from a mixture of the Cu0.8Cr (wt%) alloy powder with the grain size from 10 to 15μπι and a synthetic diamond the particle diameter between 177 and 210 μιη. The mixture was subjected to sintering to form a plate, 30 mm in diameter and 3 mm thick, built of a Cu0.8Cr alloy matrix with the 50% diamond content. The sintering process was conducted at a temperature of 900 °C under a load of 50 MPa for 20 min.

τ~· < n n ™ , ,-. ,-.Fig.10 SEM image of the surface of a fracture through the diamond/Cu composite.

Fig. 10 shows the surface of a fracture through the diamond/Cu0.8Cr composite. As can be seen, the diamond particles torn out from the copper matrix are only few, whereas most of them have cleaved along the fracture plane. This gives evidence that the bond at the diamond/Cu0.8Cr interface is mechanically strong. Chemical analysis of these pellets (Fig. 11) performed using Energy Dispersive Spectroscope, revealed chromium. This suggests formation of chromium compounds at the interface, in particular formation of carbides. The formation of interfacial carbides in this type of materials was observed by Schubert et al. [10] who identified the ^ ^ ^ ag


· 223


Figure 11 .Pellets at the diamond particle surface The density of the Cu0.8Cr-diamond composite was equal to 99.7% of the theoretical and its thermal conductivity amounted to 640W/mK. Nanocrystalline Composites The nanocrystalline NiAl-TiC composites containing 25 wt% or 40 wt% of TiC were sintered using a mixture of a nano-crystalline NiAl-TiC powder. The powder mixture was prepared from an NiAl powder milled to reduce its grain size from about 50 μιη to 25 nm and a TiC powder milled to reduce the grain size from about 2-5 μιη to 60 nm. The powders were milled in a Fritsch Pulverisette P5 planar mill. The density of the composites sintered at a temperature of 1130 °C under a load of 60 MPa was 99.9 %. Fig. 12 shows TEM images of the NiAl-TiC composite. The NiAl-TiC composites containing 25 wt.% of TiC had a hardness of 750 HV1 and a stress intensity factor KiC of 7 MPa-m"2, whereas those containing 40 wt.% of TiC had a hardness of 1070 HV1 and Kic of 11.8 MPa-m"2.

Figure 12. Microstructure of the composites: a) NiAl+25wt%TiC, b) NiAl+40%TiC.

224



Sintering with the Participation of the SHS Reaction An original energy-saving technique of producing ceramic materials from elementary powders is the self-propagating high-temperature synthesis (SHS). This method was used for producing N1AI-AI2O3 composites with 13, 38 and 55 vol. % contents of AI2O3. The starting materials were aluminum, nickel and AI2O3 powders. The synthesis combined with consolidation was conducted at a temperature of 1250 °C under a load of 60 MPa for lOmin. Irrespective of the volumetric content of AI2O3, the relative density of the composites was 99 %. By way of example, Fig. 13 shows the microstructures of the composites with 13 and 55 vol.% of AI2O3. The hardness of the composites was 480 HV10 with 13vol% of AI2O3 and 680 HV10 with 55 vol% of A1203. As to the fracture toughness, it markedly depends on the content of the dispersed AI2O3 phase. The composite containing 13 % of AI2O3, showed no cracking at all.

Figure 13. SEM images of the N1AI2O3 composites with various AI2O3 contents: a) 13 vol.%, b) 55 vol.%. In the composite with 38 % of AI2O3, the stress intensity factor KJC was 9.1 MPam"2 and in the composite with 55 % of A1203 it was 8.2 MPam1'2. CONCLUSION Pulse Plasma Sintering (PPS) is a new sintering technique involving the action of an electric field. This method utilizes pulsed high electric current discharges to heat the powder subjected to pressing. The PPS process has a very high thermal efficiency, since the powder is heated directly by the pulse arc discharges. The use of capacitors as the source of energy permits generating periodic current pulses, with a duration of several hundred microseconds with energy of several kJ. These specific conditions permit producing sintered parts with a density close to the theoretical value in a short time. The results have shown that the PPS sintering technique can be used for producing dense sintered parts from a wide variety of materials such as e.g.: diamond/cemented carbide, diamond/cooper, nanocrystalline intermetallic phases. This method can also be used for producing refractory alloys by reactive processes with the participation of the SHS reaction. ACKNOWLEDGEMENT The work was supported by the grant no N507 017 32/0586 from the Polish Ministry of Science and Higher Education. REFERENCES 'S.H. Risbud, Ch.H. Shan, Fast Consolidation of Ceramics Powders, Mater. Sei. Eng., A204,1461-151 (1995). 2 S.L, Cha, S.H. Hong, and B.K. Kim, Spark Plasma Sintering Behavior of Nanocrystalline WC-lOCo Cemented Carbides Powders, Mater. Sei. Eng., A351,31-8(2003).


· 225


3

P. Feng, W. Xiong, L. Yu, Y. Zheng, and Y. Xia, Phase Evolution and Microstructure Characteristics of Ultrafine Ti(C,N)-Based Cermets by Spark Plasma Sintering, Inter. J. Refractory Met. Hard Mater., 22, 133-38 (2004). 4 L.H. Zhu, Q.W. Huang, and H.F. Zhao, Preparation of nanocrystalline WC-IOC0-O.8VC by Spark Plasma Sintering. J. Mater. Sei. Lett. 22, 1631-33 (2003). 5 J.R. Groza, A. Zavaliangos, Nanostructured Bulk Solids by Field Activated Sintering, Adv. Mater. Sei., 5, 24-33 (2003). 6 J.R. Groza, A. Zavaliangos, Sintering Activation by External Electrical Field, Mater. Sei Eng., A287, 171-77(2000). 7 W. Chen, U. Anselmi-Tamburini, J.E. Garay, J.R. Groza, and Z.A. Munir, Fundamental Investigations on the Spark Plasma Sintering/Synthesis Process I. Effect of DC Pulsing on Reactivity, Mater. Sei. Eng., A394, 132-138(2005). 8 A. Michalski, D. Siemaszko, M. Rosinski, and J. Jaroszewicz, Fabrication of Nanocrystalline WC5Co Carbide with a WC-5Co/Diamond Composite Surface Layer Using the Impulse-plasma Sintering Method, Inzynieria Materialowa, 5, 327-329 (2005). (Polish). 9 A. Michalski, M. Rosinski, Sintering Diamond/Cemented Carbides by the Pulse Plasma Sintering Method, J. Am. Ceram. Soc, 1-6, (2008) (in press). 10 T. Schubert, L. Ciupinski, W. Zielinski, A. Michalski, T. Weißgarbera, and B. Kieback, Interfacial Characterization of Cu/Diamond Composites Prepared by Powder Metallurgy for Heat Sink Applications, Sprita Mater., 58 263-266,(2008). n L . Ciupinski, D. Siemiaszko, M. Rosinski, A. Michalski, and K.J. Kurzydlowski, Heat Sink Materials Processing by Pulse Plasma Sintering, Advanced Materials Research, (2008) (in press) 12 A. Michalski, D. Siemaszko, Nanocrystalline Cemented Carbides Sintered by the Pulse Plasma Metod, Inter. J. Refractory Met. Hard Mater., 25, 153-158 (2007). 13 A. Michalski, J. Jaroszewicz, M. Rosinski, D. Siemaszko, K.J. Kurzydlowski, Nanocrystalline CuAI2O3 Composites Sintered by the Pulse Plasma Technique, Solid State Phenomena, 114, 227-232 (2006). 14 A. Michalski, M. Rosinski, D. Siemaszko, and J. Jaroszewicz, K.J. Kurzydlowski, Pulse Plasma Sintering of Nano-Crystalline Cu Powder, Solid State Phenomena 114, 239-244 (2006). 15 M. Rosinski, A. Michalski, Nanocrystalline NiAl-TiC Composites Sintered by the Pulse Plasma Metod, Solid State Phenomena, 114,233-238 (2006). 16 A. Michalski, J. Jaroszewicz, M. Rosinski, and D. Siemaszko, N1AI-AI2O3 Composites Produced by Pulse Plasma Sintering with the Participation of the SHS Reaction, Intermetallics, 14,603-606 (2006). 17 A. Michalski, J. Jaroszewicz, and M. Rosinski, The Synthesis of NiAl Using the Pulse Plasma Method with the Participation of the SHS Reaction, Int. J. Self-Propagation High-Temp. Synth., 12, 237-246 (2003). 18 J. Jaroszewicz, A. Michalski, Pulse Plasma Sintering Combined with a Combustion Synthesis of a TiB2 Composite with a Nickel Matrix, J. Europ. Ceram. Soc, 26, 247-2430 (2006). 1 H. Moriguchi, K. Tsuduki, A. Ikegaya, Y. Miyamoto, and Y. Morisada, Sintering Behavior and Properties of Diamond/Cemented Carbides," Int. J. Refract. Met. Hard Mater., 25,237-243 (2007).

226


INFLUENCE OF ELECTRIC TECHNIQUE (FAST)

FIELDS

DURING

THE

FIELD

ASSISTED

SINTERING

Michaela Müller and Rolf Ciasen Saarland University, Department of Powder Technology Geb. C6 3, 66123 Saarbruecken, Germany ABSTRACT The Field Assisted Sintering Technique (FAST) has become popular since the 1990s. Since then, many investigations have been effected on materials that can be sintered by this technique, numerical simulations of the processes have also been performed. Several theories about the influence of the electric current during sintering have been established, but up to now the activation due to the pulsed electric current in the powder still remains uncertain. In this paper a modified FAST equipment has been developed allowing to separate the electric circuit from the heating loop. With this setup we were able to investigate the amount of current passing through the powder and its influence on sintering. INTRODUCTION Since the late 1990s there have been many investigations on the Field Assisted Sintering Technique (FAST) which is also known as Spark Plasma Sintering (SPS). This powder technology process involves pressing loose powders uniaxially in a graphite die connected to a pulsed electric field allowing the flow of high current densities. The application of an electric current to the graphite die results in high heating rates up to 600 K/min due to the heat developed by Joule effect. The main advantage of this technique is the rapid densification of loose powders, which can be consolidated without additives. FAST has become very popular because it permits to sinter dense materials (near 100 % of theoretical density (%TD)) at lower temperatures than those normally used. Soak times are also reduced. As a consequence FAST has been applied to a wide range of materials ' and special attention has been drawn to ceramic and composite materials. Dense AI2O3, Z1O2 and T1O2 2-6 have already been successfully sintered. The role of the electric current during this process has not been clarified yet. Several theories have been set up. One of them claims that electric current builds up some sparks that form a plasma between the powder particles, promoting the transport of matter. Other theories state that the current activates the surface of the powder particles by means of removing surface impurities which may be present. Some numerical simulations have been performed7"9 stating that the current's path through the powder or die depends mostly on the material. If the powder is an insulating ceramic material, most of the current will pass through the die whereas in the case of metallic powders a considerable fraction of the current will flow through the particles. The problem of the experimental verification of these results lies in the way the FAST equipment is built. There is no isolation between the powder bed and the surrounding graphite die. Therefore the amount of current passing through the die - used for heating and the amount passing through the powder - activating the sintering process - can only be estimated by theoretical formulations. In this paper the role of the electric current during FAST is investigated by separating the heating loop from the electric circuit. To this end, a new equipment was built which guarantees current flow only through the powder. The employed heating rates are relatively low (25 K/min) compared to those normally used. But in this way, the thermal effects due to heating can be widely suppressed.

227


EXPERIMENTAL PROCEDURE The experiments were carried out in a specially designed equipment depicted in figure 1. A laboratory press (Flexi-Press 16 t, Stenhoj, Denmark) was employed along with two pressing punches made of dense silicon-infiltrated SiC. This material was selected because of its good electrical conductivity at high temperatures and its high compressive strength. The supports for both punches are made of high-temperature-resistant steel. They act as electrodes once connected to the current. As pressing die a dense, high purity (99.7 %) alumina pipe was employed to guarantee that the applied current passes only through the powder. The electrical resistivity of this pipe is higher than 10 14 Ωαη. The samples consisted of Titania nano-powder (P25, Evonic Degussa, Germany). The material was selected for the following reasons: Its electrical resistivity is about 1010 Ωαη and sinks with increasing temperature, because T1O2 is a semiconductor. In this way the sample resistivity is lower than that of the alumina, guaranteeing the flow of current through the powder and not through the die. Besides the high ratio of surface to volume of the powder particles helps to activate the surface. The pressing equipment was positioned inside a metallic heating tube that could be independently regulated. Doing so, the separation of the heating circuit from the electric circuit was achieved. The current passing through the die was measured with a multimeter (METRAtop 52, Gossen-Metrawatt, Germany) and also mapped with an oscilloscope (Oscilloscope OS-9020G, LG Precision Company, South Korea).

Figure 1. Scheme of built FAST-equipment The powder was filled in the 10 mm inner diametre die and heated up to the sintering temperature with a heating rate of 25 K/min. For the sake of comparison, sintering temperatures of 575, 725, 750, 825 and 850 CC were chosen (5 samples at each temperature). Former investigations showed that at a sintering temperature of 775 °C with a soak time of 45 min, a density of 79 %TD (%TD = % of the theoretical density) could be reached 10. Therefore the soak time was extended to

228



60 min. After reaching the sintering temperature, a pressure of 15 MPa was applied to the samples and then held during the whole soak. After finishing dwell time, the pressure was released. During heating and soak, a pulsed voltage of 4 V was applied and the current passing through the sample was measured and registerd. The duration of one pulse was 3 ms and the ratio pulse-on-time/pulse-off-time was 5/1. Additional samples were sintered with a soak time of 5 min to compare with the samples sintered in the commercially available FAST-equipment. These samples were labeled "A-i-j", where i stands for the sintering temperature and y for the holding time. To study the influence of the electric current, 10 samples were processed similarly but without application of electric current. They were labeled "B-i-j". In order to study the performance of our equipment, experiments were carried out in a commercially available FAST equipment (FCT HP D 25/1, FCT Systeme, Germany). The powder was filled in the die with an inner diameter of 20 mm, prepressed with 15 MPa for 30 s and then put into the furnace. While applying a pressure of 15 MPa, it was heated up to 450 °C during 1 min, then heated up to the sintering temperature at 25 or 100 K/min. Sintering temperatures of 650, 700, 750 and 850 °C were chosen. Samples were held at the sintering temperature during 5 min. A pressure of 15 MPa was applied during the soak time. The applied pulses had a duration of 3 ms with a pulse-on-time/pulse-offtime ratio of 5:1. Those samples were labeled "C-i" ( heating rate of 25 K/min) and "D-i" (heating rate 100 K/min). To study the role of sintering temperature, heating rate, pressure and electric current during FAST a Design-of-Experiments method (DOE) called "comparison of variables for process optimization" was used. A high density was chosen as target value. Estimated "good" and "bad" levels were then given to the four factors as can be seen in table I. The statistical spreading of the target value due to the action of all factors was estimated during preliminary tests. During these tests all factors were set on the level "good" and afterwards all factors were set on the level "bad". Each preliminary test was repeated once. Based on this estimation, single experiments were carried out in which one factor was set to the "good" level while the others were set to the "bad" level. With this method the influence of each factor on the target value can be estimated looking out for dominances or interactions between factors. More details on this method can be found elsewhere 12. The sample density was measured by the Archimedes' method and the microstructure of fractured surfaces was investigated by SEM (Jeol SEM-7000, Japan). Table I. Factor steps for Design of Experiments examination Level Level Factor "good" "bad" Sintering 800 °C 600 °C temperature Heating rate 25 K/min 10 K/min Pressure 5 MPa 15 MPa Current on off RESULTS The current passing through the powder as a function of time during the type A experiments (see table II) is shown in figure 2. At the beginning the current increases slightly to 0.1 mA. When a temperature of 400 °C is reached, the current decreases to a value of 0.008 mA followed by an increase up to several mA. From 600 to 800 °C fluctuating current values can be observed. After 40 min, when reachmg 850 °C, the current goes up from 0.1 mA and reaches about 300 mA at the end of the dwell time. At this point the applied voltage was removed.


· 229


Figure 2. Temperature profile during sintering cycle "with current" and current passing through the sample against time Table II. Characteristics of samples produced using different sintering techniques Sintering Heating Rate Dwell Time Density Grain Size Sample Temperature (min) (K/min) (%TD) (nm) (°C) 60 49 32,5 575 25 60 93 725 25 77 60 90 112 750 25 A (with current) 825 60 92 160 25 850 60 95 176 25 5 62 750 25 77 5 130 850 87 25 77 725 60 80 25 60 89 103 750 25 825 60 92 161 B (without 25 current) 60 93 171 850 25 5 81 79 750 25 132 5 86 25 850 5 112 750 77 C 25 5 86 (FAST 25 K/min) 850 271 25 5 69 69 100 650 5 74 100 700 89 D (FAST 100 K/min) 123 5 80 750 100 5 87 268 850 100

230



The densities and grain sizes of the samples prepared using the sintering techniques A and B in the new-built equipment, and C and D in the commercially available FAST equipment are shown in table II. Applying electric currents results in densities increasing from 49 %TD to 95 %TD when sintering at temperatures increasing from 575 °C to 850 °C, respectively (samples A-575-60 to A-85060). The grain sizes increase in the same way from 33 nm to 176 nm. The samples B-725-60 to B-850-60 reach densities from 80 %TD to 93 %TD and grain sizes between 77 nm and 171 nm. Shorter soak times result in lower densities and grain sizes for the samples sintered in our equipment with and without applied electric current (samples A-750-5, A-850-5, B-750-5 and B-850-5). The obtained densities vary from 77 %TD to 87 %TD, the grain sizes from 62 nm to 132 nm. The samples sintered in the FAST equipment (C-i and D-i) have densities from 69 %TD to 87 %TD and grain sizes between 69 nm and 271 nm. Figure 3 plots the comparison of the densities obtained with the different techniques. It shows that the sample A-850-60 has the highest sintered density. The samples A-750-60 to A-850-60 and B750-60 to B-850-60 reach the highest densities. The densities for the samples sintered in the conventional FAST equipment and in our new equipment for 5 min are in the range of 69 to 87 %TD and at equivalent temperatures in the same order of magnitude. Grain sizes of the sintered samples are shown in figure 4. All samples sintered at temperatures equal to or lower than 725 °C present grains with average diameters smaller than 100 nm. At a sintering temperature of 750 °C the samples A-750-60, B-750-60, C-750 and D-750 show grains with average diameters between 110 nm and 140 nm and a standard deviation of 45 nm (same order of magnitude). The samples A-750-5 and B-750-5 present smaller grains with mean diameters of 62 nm and 79 nm respectively. Significant differences are found at 850 °C, the samples C-850 and D-850 present grain sizes of around 270 nm, whereas die samples A-850 and B-850 show grains with average diameters of around 170 nm. The SEM micrographs shown in figure 6 depict this difference.

Figure 3. Measured densities against sintering temperature obtained by A) sintering with applied current, B) sintering without applied current and C) and D) in a conventional FAST equipment


■ 231


Figure 4. Measured grain size against sintering temperature obtained by A) sintering with applied current, B) sintering without applied current and C) and D) in a conventional FAST equipment SEM micrographs of fractured surfaces of the preliminary tests performed during DOEmeasurements are shown in the in figure 5. Figure 5.a presents a quite dense microstructure when sintering with all factors set on step "good" in contrast to a loose powder packing with no visible sintering progression for the "bad" test shown in figure 5.b.

Figure 5. SEM micrographs of fractured surfaces of samples sintered during preliminary tests of Design of Experiments measurements a) all factors on factor step "good", 800 °C, 25 K/min, 15 MPa, with applied current b) all factors on factor step "bad", 600 °C, 10 K/min, 5 MPa, without current application

232



Figure 6. SEM micrographs of fractured surfaces, samples sintered a) with applied current at 850 °C, 25 K/min, 60 min, b) without applied current at 850 °C, 25 K/min, 60 min, c) conventional FAST at 850 °C, 25 K/min, 5 min, d) conventional FAST at 850 °C, 100 K/min, 5 min Micrographs of fractured surfaces of the samples where the variation of one factor was performed are shown in figure 7 (One factor is set on the step "good" the others stay on step "bad"). Figure 7.a shows larger grains than the others, 7.b shows neck formation between the small particles. Figure 7.c shows a quite dense particle arrangement and the apparition of "molten zones". Figure 7.d presents a quite dense compaction without any formation of interparticle connections like "sinter necks". The influence of the factors on the density during DOE-measurements is shown in figure 8. During preliminary tests Gl, G2, Bl and B2 the range of the target value for "good" and "bad" test is evaluated. The "good" tests result in target values of 78 %TD, the "bad" tests in 40 %TD. The variation of the factor temperature leads to a complete inversion in the target value (samples TgRb and TbRg). Pressure variation results in an approximation of the values (samples PgRb and PbRg). DISCUSSION With this newly developed equipment it is possible to determine the amount of current flowing through the powder because its electrical conductivity is several orders of magnitude higher than that of the alumina die, as shown in figure 2. The initial increase in current can be attributed to the


· 233


semiconducting properties of titania. The second major increase starting at 600 °C followed by a slight decrease is due to particles rearrangement and densification during the phase transformation from anatase to rutile. Literature values describing the transformation are in good agreement with this result ". For this reason all samples sintered at temperatures higher than 800 °C consist exclusively of rutile, which has a theoretical density of 4.26 g/cm3. Such results have already been presented in , where the X-ray diffractograms, used to analyze the samples, are shown. The increase in the flowing current during the soak step can then be ascribed to densification processes. The electrical resistance sinks with a denser particle arrangement. With our equipment, the highest densities are attained at temperatures 750 °C and 850°C for the type-A and type-B samples and for soak times of 60 min whether current is applied or not (figure 3). The slightly higher densites determined for the samples A-750-60 to A-850-60 (compared to samples B-750-60 to B-850-60) lie in the range of measurement error and this difference is too small to be ascribed to a current effect. At sintering temperatures lower than 750 °C, the commercially FAST equipment (100 K/min heating rate, samples D-650 to D-750) and the newly developed equipment lead to almost the same densities. The variation of dwell time seems to have more influence on the density

Figure 7. SEM micrographs of fractured surfaces of samples sintered during Design of Experiments measurements a) high sintering temperature: grain growth, b) high heating rates: formation of "sinter-necks", c) high pressure: "molten zones", d) applied current: no visible influence

234



Figure 8. Influence of factors on target value "density" during DOE. Bl, B2: Preliminary "bad" test, Gl, G2: preliminary "good"test, b = bad, g = good, T = Temperature, H = Heating rate, P = Pressure, C = Current than applying current or varying heating rates, because all samples sintered with holding times of 5 min achieve comparable densities, whereas the samples sintered with 60 min holding time present about a 10 % increase in density. The longer holding times seem to favor densification by diffusion mechanisms. Some hints that may support this conclusion are observed by comparing the grain sizes of fractured surfaces (diagramm in figure 4). Up to 750 °C all grain sizes are smaller than 150 nm. At 850 °C, samples C-850-5 and D-850-5 result in a mean grain size of 270 nm. Samples A-850-60 and B-850-60 show smaller grains (175 nm) and samples A-850-5 and B-850-5 still smaller ones (130 nm). It seems that longer soak time lead to longer diffusion times and therefore to grain growth. The samples sintered with the conventional FAST equipment show quite large grains even at short soak times what can be explained by the optimal sintering temperature being exceeded in this furnace. Grain growth takes place rapidly, because the heat generated by the current flow through the graphite die leads to an overheating of the sample due to difficult temperature when those high heating rates are applied. The results of the DOE-method allowed evaluating the effects of the different factors on the density and microstructure of the samples. With all factors at the "good" level, the samples can be sintered to densities around 78 %TD, meaning that sintering proceeds favorably (Figure 8). A loose compaction without any sintering takes place when all factors are kept at the "bad" level. The final densities are close to the green densities, that is around 40 %TD (figure 6,b and figure 8). When comparing the microstructures obtained with the single experiments (figure 7) to that of figure 6.b the effects of each "manipulated" factor can be determined. High sintering temperatures lead to an enhanced grain growth by favoring diffusion. Figure 8 supports this theory, because a complete inversion in the target values due to the variation of one factor means that this factor is dominant u. Higher heating rates accelerate sintering by formation of "sinter-necks" between the particles, but have no direct influence on the final density as shown in figure 8. The application of a pulsed current shows


· 235


no influence on the microstructure of the fractured surfaces, as well as on the sintered density. When applying high pressures, there is an approximation in the target value shown in figure 8, meaning this factor is important, but not independent of the others. "Molten zones" are found in the microstructure (figure 7.c), but this phenomenon could not be related necessarily to the high pressure. Further investigations have to be done on this subject to quantify the influence of the factors and to generalize the results for less conducting materials. CONCLUSION AND OUTLOOK Investigations on the effect of the applied electric current during the Field Assisted Sintering Technique were carried out in a newly developed equipment. The equipment allows to measure the amount of current passing through the powder independently from that used for heating the sample. Initial tests showed no visible effect of the current on the sintered densities and grain sizes. A Designof-Experiments method was used to investigate the effect of sintering temperature, heating rate, pressure and current on density and microstructure. Sintering temperature was found to be the dominant factor on density and led to higher grain growth. Heating rate influences the formation of sinter necks, while pressure affects mostly the sample density and is also an important factor to promote sintering. The electric current passing through the sample, had no apparent influence on either microstructure or density. Further investigations will be performed to support these results and to study the influence of electric current pulses on insulating powders. REFERENCES 'M. Tokita, Mechanism of Spark Plasma Sintering. Proceedings of the International Symposium on Microwave, Plasma and Thermochemical Processing of Advanced Materials, ed. S. Miyake and M. Samandi. (1997), Osaka Universities Japan. 69-67. 2 Z. Shen, M. Johnsson, Z. Zhao and M. Nygren, Spark Plasma Sintering of Alumina, J. Am. Ceram. Soc, 85, 1921-1927(2002). 3 L. Gao, Z. Shen, H. Miyamoto and M. Nygren, Superfast Densification of Oxide/Oxide Ceramic Composites, J. Am. Ceram. Soc, 82, 1061-1063 (1999). 4 R. Chaim, Z. Shen and M. Nygren, Transparent nanocrystalline MgO by rapid and low-temperature spark plasma sintering,/. Mater. Res., 19, 2527-2531 (2004). 5 G. D. Zhan, J. Kuntz, J. Wan, J. Garay and A. K. Mukherjee, A novel processing route to develop a dense nanocrystalline alumina matrix (< 100 nm) nanocomposite material, J. Am. Ceram. Soc, 86, 200-202 (2003). 6 Y. I. Lee, J.-H. Lee, S.-H. Hong and D.-Y. Kim, Preparation of nanostructured T1O2 ceramics by spark plasma sintering, Mat. Res. Bull., 38, 925-930 (2003). 7 K. Vanmeensel, A. Laptev, J. Hennicke, J. Vleugels and O. V. d. Biest, Modelling of the temperature distribution during field assisted sintering, Acta Mater., 53,4379-4388 (2005). 8 K. Vanmeensel, A. Laptev, O. V. d. Biest and J. Vleugels, Field assisted sintering of electroconductive Zr02-based composites, J. Eur. Ceram. Soc, 27, 979-985 (2006). 9 U. Anselmi-Tamburini, S. Gennari, J. E. Garay and Z. A. Munir, Fundamental investigations on the spark plasma sintering/sythesis process II. Modeling of current and temperature distributions, Mat. Sei. Eng., 394, 139-148(2005). I0 M. E. Müller, Untersuchungen zum feldunterstützten Sintern, Diplomarbeit, Universität des Saarlandes, Saarbrücken (2007). "N. Masahashi, Fabriction of bulk anatase T1O2 by the spark plasma sintering method, Mat. Sei. Eng., 452-453, 721-726 (2007). 12 W. Kleppmann, Taschenbuch Versuchsplanung, Carl Hanser Verlag, ISBN: 3-446-40617-4 (2006)

236


SINTERING OF COMBUSTION SYNTHESIZED TIB2-ZR02 COMPOSITE POWDERS IN CONVENTIONAL AND MICROWAVE FURNACES Hayk Khachatryan2, Alok Vats', Zachary Doorenbos1, Suren ICharatyan2, and Jan A. Puszynski1 'South Dakota School of Mines and Technology, Rapid City, SD, USA, Yerevan State University, Yerevan, Armenia ABSTRACT Comparative pressureless sintering studies of Τ1Β2-Ζ1Ό2 composite powders with different compositions of Y2O3 stabilized Z1O2 were conducted in an inert atmosphere using conventional and microwave sintering furnaces. Composite powders were prepared in one step combustion synthesis (CS) also called self-propagating high-temperature synthesis (SHS) process using two different initial reactant compositions: i) Ti- 2B- 0.2 Z1O2 (unstabilized) and ii) Zr-Ti02-2B. In both cases, Y2O3 powder was added as a stabilizer. Attrition milled submicron product powders were dried and pressed to the desired green density in uniaxial or cold isostatic presses. For comparison commercial T1B2 and Z1O2 (yttria stabilized) powders of the same composition were subjected to attrition milling and green samples were sintered together with those obtained from combustion synthesized powders. Sintering experiments were done in a conventional pressureless graphite furnace for one hour at 1500°C, 1700°C, and 1850°C under the flowing argon gas. The sintering results have shown that combustion synthesized composite powders have better sinterability, CS-T1B2 rich composite sinters up to 99% of relative density at 1700°C. Sintering studies in a microwave furnace have shown that the same results can be achieved at much faster rates and lower temperatures as compared to the conventional furnace. INTRODUCTION Ceramic materials offer many excellent properties, such as high hardness, high melting point, chemical inertness, high compression strength, and impact resistance, etc. [1,2]. Mechanical properties of ceramic materials, especially fracture toughness, can be improved by the addition of other phases into a material's matrix to form a composite structure [3]. Ceramic parts are generally produced by cold die compaction with subsequent sintering and finishing or by hot pressing or hot isostatic pressing and finishing. Among these procedures, sintering or densification is the most important process for fabricating ceramics [4-6].

237

Sintering of Combustion Synthesized TiB 2 -Zr0 2 Composite Powders

Specific surface area of the powder among other properties plays a crucial role during sintering, as it provides additional driving force for densification, smaller initial size of the powders normally results in higher densities or in some cases may lead to lower firing temperatures due to faster sintering kinetics [3,5,6]. The friction between the powder and the die wall during powder compaction typically leads to residual stresses and nonuniform green density in ceramic compacts. The variation in density distribution in powder compacts results in nonuniform shrinkage during sintering process. These residual stresses cause micro cracks in the powder compacts during the sintering process and thus affects the mechanical properties of the compacts. The density that can be achieved during a cold compaction of ceramic powder in a die is lower compared to metal powders. Therefore, the effects of nonuniform density distribution and residual stresses in ceramic compacts are more serious than those associated with sintering of metal powders [5,6]. This work was focused on i) investigation of pressureless sintering of combustion synthesized TiB2-Zr02 composites [7-9]; ii) feasibility study of sintering TiB2-Zr02 composites in a microwave furnace. EXPERIMENTAL PROCEDURE In this sintering study, four different TiB2-Zr02 composite powders were used as shown in Table 1. The first two composites powders were synthesized using combustion synthesis technique with Zr0 2 concentration being 25wt% and 75wt% [6, 7]. The first composite (PI) was synthesized from titanium and boron reactants with the addition of unstabilized zirconia. In this case, the amount of magnesium present in boron was sufficient to stabilize zirconia during the combustion process [7]. The second composite powder (P2) was formed from Zr, B and T1O2 as reactants with the addition of unstabilized zirconia. In this case, yttria was added in addition to the already present magnesium in boron powder in order to fully stabilized zirconia in the T1B2Z1O2 composite [7]. For comparison, identical compositions (P3 and P4) were also prepared by mixing T1B2 and fully stabilized zirconia powders. Combustion synthesized titanium diboride powder was used in P3 and P4 composite compositions. The average particle size of combustion synthesized T1B2 was 5 μηι. Fully stabilized zirconia used in P3 and P4 composite compositions was obtained from the Zircar Company with the average particle size less than 5μπι. All four compositions were subjected to mechanical milling in an attritor mill (01-HDDM Union Process, Szegvari Attritor System). Dense zirconia grinding media (1.2-1.4 mm in dia)

238



was used with distilled water. A constant shaft speed of 3,200 rpm and a total milling time of 40 min was used to mill each composition. Table 1. Compositions of composite powders used in sintering experiments. Name

Composition

PI

75%wt T1B2 - 25wt% ZrC"2 (stabilized) combustion synthesized powder

P2

25%wt T1B2 - 75wt% ZrC>2 (stabilized) combustion synthesized powder

P3

75%wt TiB2 - 25wt% Z1O2 (stabilized) mixed powder

P4

25%wt TiB2 - 75wt% Zr0 2 (stabilized) mixed powder

The ratio of the grinding media to milled powder was 10:1 by weight. After milling, the composites powders were separated from media and dried. The final drying was done at 8090°C in a vacuum oven. Next, cylindrical pellets with a diameter of 12 mm and 7-8 mm height were prepared using both cold isostatic and uniaxial presses. Samples in the cold isostatic press (CIP 32260, Avure Autoclave Systems, Inc.) were pressed at 60, 137, 210, 275, and 345MPa. The pressing in an uniaxial press was done at 87.5MPa pressure. Pressureless sintering studies were conducted in two different sintering furnaces: i) conventional sintering furnace with a graphite heating element (FP20 from Astro Industries, Inc.) ii) microwave furnace (VIS300001B from CPI Inc.).The sintering atmosphere in a graphite furnace was argon, while the inert gas used during the sintering in microwave was high purity helium. The reason for using helium instead of argon in microwave furnace was to inhibit gas ionization which led to a plasma formation. Sintering studies in the conventional graphite furnace were carried out at 1500°C, 1700°C, and 1850°C. Heating rate was set at 5°C/min in the graphite furnace and soak time was 1 hr. In order to determine sintering kinetics, additional experiments were conducted with different soak times of 15, 30, 45, 60, and 75 min. Densities of sintered samples were determined using Archimedes method. In order to minimize experimental error, three samples of each composition were used in the analysis. Sintering studies in the microwave furnace were curried out at 1300°C, 1400°C, and 1500°C. Heating rate was 70-100°C/min and soak time was 15 min. Samples were placed in a specially designed highly porous alumina box. The box cap had a small hole to determine samples temperature using a pyrometer.


· 239


The Vickers hardness of the sintered specimens was measured on a hardness tester (model Micrometer 4 micro, Buehler Ltd.) with an indentation load of 1,000 kg. Specimens were polished on a diamond-grinding wheel (type Cameo disk Platinum 1, 3 and 4 purchased from Leco Co.). Phase composition analyses were carried out on the sintered test specimen using electron dispersive X ray analyzer (EDX) on SEM (Supra40VP from Zeiss Co.) and X-ray diffractometer (Rigaku Ultima Plus). The stability of these composites to oxidation was measured by heating the samples in air at temperature of 800 and 1000°C for 4 hours. RESULTS AND DISCUSSION SEM photographs of TiB2-Zr02 composite powder (PI) before and after attrition milling are shown in Figure 1. It is clear from these micrographs that the combustion synthesized composite powder was fused during synthesis. According to EDX analysis, gray and well defined crystals are titanium diboride, while light gray phase is zirconia (Figure la). After attrition milling, the average particle sizes of composite powders was reduced below 2 μηι (Figure lb). X-ray analyses did not show any changes in phase composition due to the milling and drying of composite powders.

Figure 1: Microstructure of PI powders; a) before milling, b) after 40 min milling. Figure 2 shows the final relative densities of the pellets as a function soak time (1500°C) and powder pressing technique. The PI powder was pressed using both CIP and uniaxial press. The samples were pressed at 210 MPa and 87.5 MPa, respectively. The pressed pellets were sintered in the graphite sintering furnace at 0.1 MPa argon pressure.

240



Figure 2: Relative density of pellets made from 75 wt%TiB2-25 wt%Zr02 composite (PI) at ]50(fC temperature as the function of soak time. As can be seen from Figure 2, the densification of this composite material at 1,500°C reached much higher density for the pellets pressed using cold isostatic press as expected. The final density of pellets prepared by uniaxial press was 82% and for the pellets prepared using cold isostatic press a density of 90% was obtained. - 50

Δ, % 100 -

B

Θ B

98 ■

96a

94-

Δ

Θ,%

—~—8—_

Δ ^___

a

n

-40

■ 30

Λ

Ä

■ 20

92 90 -

)

100

200

300

P, MPa

400

Figure 3: Relative density (Δ) and volumetric shrinkage degree (Θ) of composite powder (PI) as a function of a consolidation pressure in CIP. Samples were sintered in a graphite sintering furnace at 150(fC in argon atmosphere with a soak ofl hour. It is important to note here that for all future experiments a soak time of 1 hour was used. Relative density and degree of volumetric shrinkage of PI composite pellets formed in a cold isostatic press and later sintered at 1500°C for 1 hr in a pressureless graphite furnace as a function of green consolidation pressure is shown in Figure 3. This Figure indicates that the


■ 241


initial green density has a significant effect on the final density and volumetric shrinkage of the TiB2-Zr02 composite. It should be noted that the composite (P3) with the same composition prepared by mixed T1B2 and fully stabilized ZrC>2 powders resulted in significantly lower relative densities (78-85%) under identical sintering conditions. Figure 4a shows SEM photograph of a green sample of PI composite prepared in CIP at 210 MPa whereas Figure 4b shows the microstructure of the sintered body of the same pellet at 1,500°C for 1 hr in a graphite furnace in an argon atmosphere.

Figure 4: Microstructure ofPI powders a) sintered at 1500PC temperature b) initial powder after attrition milling. In order to increase the relative density of sintered composites, additional sintering experiments were conducted at 1,700°C. Figure 5 shows the relative density (Δ) and the volumetric shrinkage degree (Θ) of PI and P3 composites sintered for 1 hr in a graphite furnace in argon atmosphere as a function of green consolidation pressure in a cold isostatic press. Again in this case, the combustion synthesized powders did show much higher relative density after 1 hr soak time than the pellets made by mixed TiB2 and Zr0 2 powders, 98% vs. 88%). The faster sintering rates for combustion synthesized powders can be explained by the fact that the composite powders obtained via the SHS (self-propagating high-temperature synthesis)process have higher percentage of nonequilibrium phases which are formed due to high cooling rates during the synthesis. According to XRD analysis done on sintered samples, no phase changes in the samples were observed, only phase composition comprises of two phases T1B2 and stabilized ZrC>2. Figure 6(a) shows the microstructure of SHS synthesized composite powder which was sintered at 1700°C.

242



Δ,%

Θ, %

a

90 ■

• 60

Δ

• 50 •40 70 ■

Θ

0-

■ 30 ■ 20 ■ 10

j

()

100-1 Δ,%

200

100

300

P, MPa

400

r 85

b

98 -

&, °/

■45 "■"-"—Δ.

90 ■

85 ·

__1850°C

1700 °C

)

100

200

300

■ 40 *

P, M P a

- 35

400

Figure 7: Relative density (A) and degree of volumetric shrinkage (Θ) ofP2 composite as a function ofgreen consolidation pressure in C1P. Sintering was done in a graphite sintering furnace at 1700° C and 1850° Cfor Ihr in argon atmosphere. The relative densities of P2 samples sintered at 1700°C were between 95 to 96% and at 1850°C between 97to 98%. P4 composite which was prepared by mixing TiB2 and stabilized Z1O2 after sintering at same conditions showed lower relative densities (~ 90%). SEM image of P2 samples sintered at 1700°C is shown in Figure 8. It can be seen in this micrograph that TiB2 crystals are homogeneously distributed in a well sintered zirconia matrix. Also in this case XRD analysis, reveled no other peaks associated with phase other that T1B2 and stabilized ZrC>2.

Figure 8: Microstructure ofP2 powders sintered at 1850 °C temperature in a graphite furnace under argon atmosphere for 1 hour. In order to study the stability of TiB2-Zr02 composites in oxidizing atmosphere the sintered samples were heated at 800 and 1000°C temperature in air, Figure 9 presents weight change of


· 245


different composite samples at the 1000°C temperatures. The samples heat at 800°C showed very high oxidation resistance, no weight change was observed at this temperature range after 4 hours exposure. However, weight changes for the samples heated at 1000°C were observed as it is evident from Figure 9. The weight change of 3 ~ 3.5% was observed at these conditions. It should be noted that weight changes were more pronounced during the first stage of heating. According to these results, P3 composites, which was obtained by mixing, started oxidizing comparatively earlier then composites obtained by SHS method (PI and P2). This can be caused by lower porosity of (PI and P2) SHS synthesized composites as compared to P3.

I

4 3.5

3 -

i 1S2 |

"

1.5

1 0.5

0

0

70

140

210

280

, t, min

350

Figure 9: % Weight change associated with different composition at 100(fC. The Vickers hardness analysis has shown that the hardness strongly depend on the composition of TiB2-Zr02 composites. The PI composite has shown the highest hardness of 1750±50 kg/mm2 for samples sintered at 1500°C and 2400±50 kg/mm2 for samples sintered at 1700°C. While P3 type composite has shown hardness of 1750±50 kg/mm2 for samples sintered at 1700°C. For P2 composite compacts which are zirconia rich, the hardness of 1600±50 kg/mm2 was when sintered at 1850°C and 1200±50 kg/mm2 for samples sintered at 1700°C. CONCLUSIONS In this investigation, it was established that one stage SHS synthesized composites possess higher sinterability then composites obtained by mixing method. It has been clearly

246



shown that that higher relative densities were obtained for composites prepared using cold isostatic press as comparing to uni-axial press. The titanium diboride rich composites (75wt%) were successfully densified by pressureless sintering at 1700°C, while the samples with 75wt% zirconia needed higher sintering temperature of 1850°C. Composites samples sintered using microwave furnace showed much faster sintering kinetics at lower sintering temperatures as compared to samples sintered using conventional furnaces. The faster sintering kinetics is attributed to the diffusional drift term due to electromagnetic coupling with the samples in the microwave furnace. It was established that sintered composites are very resistant against oxidation up to 800°C. It is important to note that samples made from SHS derived composite powders exhibit higher stability towards oxidation. It has been also shown that T1B2- ZrC>2 composites exhibit higher hardness as compared to dense zirconia. ACKNOWLEDGEMENT The authors acknowledge the financial support endowed to Dr. Hayk Khachatryan by CIES/ Fulbright scholarship. REFERENCES [1]

J. M. Leger and J. Haines, "The Search for Superhard Materials", Endeavour, 21, 1997.

[2]

R. J. Brook, "Superhard Ceramics", Nature, 400, 1999.

[3]

W.E. Lee, W.M. Rainforth, Champman&Hall, London-Glasgow-Weinheim-New York, Tokyo-Melbourne, Madras, 1994.

[4]

R.M. German, "Sintering Theory and Practice", ISBN 0-471-05786-X. Wiley-VCH, January 1996.

[5]

S Kang, "Sintering-Densification, Grain Growth and Microstructure", ISBN 07506 63855, Elsevier Butterworth- Heinermann, 2005.

[6]

M.W. Barsoum, " Fundamentals of Ceramics", ISBN 0-07-005521 -1, McGraw Hill series in Materials Sciences, 1997.


■ 247


[7]

H. Khachatryan, J Puszynski, S. Kharatyan "Combustion Synthesis of Titanium Diboride and Zirconia Composite Powders. Part I" J. Am. Ceram. Soc, 1-6, 2008.

[8]

A.G. Merzhanov. "Twenty Years of Search and Findings". In: Combustion and Plasma Synthesis of High-Temperature Materials, Eds. Z.A. Muñir, J.B. Holt, N.Y.: VCH Publ. Inc., pp.1-53, 1990.

[9]

Z.A. Munir, U. Anselmi-Tamburini. "Self-propagating Exothermic Reactions: the Synthesis of High-Temperature Materials by Combustion". Mater. Sei. Reports., vol.69, No.7-8, pp. 277-365, 1989.

248


PRODUCTION AND CHARACTERIZATION OF WC-Co CEMENTED CARBIDES BY FIELD ASSISTED SINTERING Rafet Emre Özüdogru1, Filiz Cinar Sahin1'2, Onuralp Yucel1'2 'Istanbul Technical University, Metallurgical & Materials Eng.Dept. Maslak, Istanbul, 34469, Turkey Istanbul Technical University, Applied Research Center of Materials Science & Production Technologies, Maslak, Istanbul, 34469, Turkey ABSTRACT Cemented WC-Co bulk materials were obtained from mixed WC-Co powder by using Field Assisted Sintering technique. Powder mixtures were sintered at 1300°C, 1350°C and 1375°C under 50MPa pressure in order to determine the effects of sintering temperatures. Sintering behavior and mechanical properties such as density, hardness, fracture toughness and modulus of elasticity of fine-grained WC-5 wt% Co and WC-10 wt% Co hard metal were investigated. Microstructural observations were carried out by Scanning Electron Microscopy (SEM) technique. Test results show that densities of bulk materials sintered at 1375CC are higher than 99 % of theoretical density. Mechanical test results point out that hardness, fracture toughness and modulus of elasticity are more dependent on composition of cobalt content than sintering temperature. INTRODUCTION WC-Co is a well known hard material used for cutting tools and dies due to its high wear resistance and strength1. They consist of a high volume fraction of the hard WC phase embedded within a soft and tough Co binder phase. These materials can be densified by liquid phase sintering and their mechanical properties depend on their compositions and microstructure. Increasing the volume fraction of Co increases the fracture toughness, but decreases hardness and wear resistance . When the grain size of the WC particles is reduced to a range of submicrometer or nanometer, the hardness and the strength of the cemented carbides increase remarkably, and the toughness improves greatly as well, thus showing excellent combined mechanical properties of the hard materials. Extensive WC grain coarsening occurs when sintering nanometer sized WC-Co starting powder mixtures by the conventional pressureless sintering. The additions of small amounts of WC grain growth inhibitors, typically 10 microns)

294



was observed at high depth of cut. The dependence of surface roughness to machining time was smoother down to 5 microns for the 60% PS. Fig. 6 exhibits the surface contour of the 60% and 70% PS ingots machined under 236, 436, and 636 conditions. The 60% PS ingots could be machined at various feed rate to achieve similar contours. On the other hand, the cross-section profiles of the 70% PS ingots reveal surface roughness with larger wave length at higher feed rate.22 The depth of the grinding induced surface flaws and microcracks could be detected with an SEM (Fig.7). Closer inspection at the sub-surface flaws of the 60% and 70% PS ingots reveals different mode of damage. While the crack length is not significantly influenced by the grinding parameters, the type of material removed varied with the cutting depth as well as with the feed.26 In the 60% PS ingots, a number of microcracks were induced and more fracture energy could be absorbed at higher feed. On the other hand, the 70% PS exhibited lateral cracks which linked up to form surface chipping. The depth of the cracks was directly dependent on the feed for the 70%, but was relatively independent of the feed for the 60%. Restorations produced by hard machining of fully sintered 3Y-TZP blocks have been shown to contain a significant amount of monoclinic zirconia.23 This is usually associated with surface microcracking, higher susceptibility to low temperature degradation and lower reliability.24 The restorations produced by machining softer pre-sintered ingot should prevent the stress-induced transformation from tetragonal to monoclinic and leads to a final surface virtually free of monoclinic phase. However, XRD pattern of the 60% and 70% PS ingots showed that a small ratio of monoclinic to tetragonal was noticeable in the as-sintered material (approximately 3% and 12%, respectively) and an even greater amount was detected on the machined surface. In the stress field of propagating cracks the matrix pressure on the tetragonal particles of 3 Y-TZP is reduced by tensile stresses and a tetragonal (t) -> monoclinic (m) phase transformation occurs by a diffusionless shear process at near sonic velocities, similar to those of the martensite formation in quenched steel.25"26 The resulting volume expansion (35%) and the shear stresses formed in the particles affect martensitic transformation and pressure tensions on the matrix, opposing the opening of the crack and increasing the energy necessary for further crack growth. 7"28 5. CONCLUSION The present study has shown that depending on degree of pre-sintering, a wide range of values in hardness, Young's modulus, flexural strength and fracture toughness could be obtained, and some aspects of the relationship between microstructure, machining parameters and machinability were highlighted. It was found that the cutting condition and machining time depended very strongly on the pre-sintered condition. Vickers' hardness varies exponentially with the porosity. At presintering 60%, 65% and 70%, flexural strength doubled in values. The average pore size of 12 micron was comparable to the critical flaw size at 65% PS. Thus the 65% PS could be machined to fine finish surface. However, the lower hardness of the carbide burr suggested that optimized pre-sintering stage at 60% was suitable for fast machining with acceptable surface roughness and low tool wear. 6. REFERENCES 1 Anusavice,K.J. 'Recent development in restorative dental ceramics' J.Am.Dent.Assoc. 124, 728 (1993) 2 Kelly,J.R., Nishimura,L, Campbell,S.D. 'Ceramic in dentistry: historical roots and current perspectives' J.Prosthet.Dent. 1996, 75, 18-32 Advances in Sintering Science and Technology

· 295


3

Kelly,J.R. 'Dental ceramics: current thinking and trends' Dent.Clin.N.Am. 48(viii), 513-30 (2004) 4 Raigrodski,A.J. 'Contemporary materials and technologies for all-ceramic fixed partial dentures: a review of the literature' J.Prosthet.Dent. 92, 57-62 (2004) 5 Anusavice,K. J. 'Development and testing of ceramics for dental applications' Ceramic Transactions, 48, 101 (1995) 6 Filser F,Kocher P, Gauckler LJ. 'Net-shaping of ceramic components by direct ceramic machining' Assembly Autom 23, 382-90 (2003) 7 Studart.A.R., Filser,F., Kocher,P., & Gauckler,L.J. 'Fatigue of zirconia under cyclic loading in water and its implications for the design of dental bridges' Dental Materials 23, 106-114 (2007) 8 Rekow, E.D., Erdman, A.G., Riley, D.R., Klamecki, B. 'CAD/CAM for dental restorations—some of the current challenges' IEEE Trans, Biomedical Engineering 38 [4] 318-414(1991) 'Ralph G. L., Mandy S. H., Heike R., Volker H., Michael H. W. 'CAD/CAM-machining effects on Y-TZP Zirconia' Dental Materials 20, 655-662 (2004) 10 Ling Y., Song, X.F., Song, Y.L., Huang, T., Li, J. 'An overview of in vitro abrasive finishing & CAD/CAM of bioceramics in restorative dentistry' International Journal of Machine Tools & Manufacture 46, 1013-1026 (2006) 11 Rosenblum, M. and Schulman, A. 'A review of all-ceramic restorations' Journal of American Dental Association 128 [3] 297-307 (1997) 12 Giordano, R.A. 'Dental ceramic restorative systems' Compendium of Continuing Education in Dentistry 17 [8] 779-794 (1996) 13 Xu, H.H.K., Kelley, R.J., Jahanmir, S., Thompson, V., Rekow, E.D. 'Enamel subsurface damage due to tooth preparation with diamonds' Journal of Dental Research 76 [1] 16981706(1997) 14 Chantikul,P., Anstis G.R., Lawn,B.R., and Marshall,B.D.' A critical evaluation of indentation techniques for mesuring fracture toughness: II, strength method' J.Am.Cer.Soc. 64,539-43(1981) 15 American Society for Testing of Materials, Designation C 1259-94 Standard test method for dynamic Young's modulus, shear modulus, and Poisson's ratio for advanced ceramics by impulse excitation of vibration. In: Annual Book of ASTM Standards 15.01, Philadelphia: ASTM, 1994 16 American Society for Testing of Materials, Designation C 1327-99 Standard test method for Vickers indentation hardness of advanced ceramics. In: Annual Book of ASTM Standards 15.01, Philadelphia: ASTM, 1999 17 Garvie,R.C. and Nicholson, P.S. 'Phase analysis in zirconia systems' J.Am.Cer.Soc. 55, 303-5 (1972) 18 R.W. Rice, Evaluation and extension of physical property-porosity models based on minimum solid area, J. Mater. Sei. 31 (1996) 102-118. 19 R.W. Rice, in: Porosity of Ceramics, Marcel Dekker, 1998, pp.375^t21 20 Lawn BR, Marshall DB. (1979) 'Hardness, toughness and brittleness: an indentation approach'. J Am Ceram Soc. 62(7-8):347-50. 2 BansaLG.K. (1976) 'Effect of flaw shape on strength of ceramics' J.Am.Cer.Soc. 59[l-2] 87-8 22 Luthardt,R.G., Holzhueter.M.S., Rudolph,H., Herold.V., and Walter.M.H. 'CAD/CAMmachining effects on Y-TZP zirconia' Dental Materials 20, 655-662 (2004)

296



Guazzato M., Albakry M., Ringer S.P., Swain M.V. 'Strength, fracture toughness and microstructure of a selection of all-ceramic materials. Part II. Zirconia-based dental ceramics' Dent Mater 20, 449-56 (2004) 24 Huang H. 'Machining characteristics and surface integrity of yttria stabilized tetragonal zirconia in high speed deep grinding' Mater Sei Eng A: Struct. 345, 155-63 (2003) 25 Evans A.G. and Heuer A.H. 'Review-transformation toughening in ceramics: martensitic transformation in crack-tip stress fields' J Am Ceram Soc 63, 241-8 (1980) 26 Kosmac, T., Oblac, C, Jevnikar, P., Funduk, N., and Marion, L. 'The effect of surface grinding and sandblasting on flexural strength and reliability of Y-TZP zirconia ceramic' Dent Mater 15,426-33 (1999) 27 Christel, P., Meunier, A., Heller, M., Torre, J.P., and Peille, C.N. 'Mechanical properties and short-term in vivo evaluation of yttrium-oxide-partially-stabilized zirconia' J Biomed Mater Res 23, 45-61 (1989) 28 Stevens R. Zirconia, zirconia ceramics, 2nd ed. Magnesium Electron Publication No. 113, Twickenham: Litho 2000, 1986.


■ 297


a)

Figure 1.. Four axis CNC with high speed spindle. Schematic G-code files for machining showing (a) contact relation, (b) flat and (c) ellipsoid operation.

298



Figure 2. Shrinkage rate of zirconia ingot heated at 3,4, and 5°C/min to 1450°C.

Figure 3. Shrinkage rate of zirconia ingot heated at 3, 4, and 5°C/min to 1450°C.


· 299


(c) Figure 4. SEM micrographs of ingot sintered at (a) 55%, (b) 60%, and (c) 70%

300



Figure 5. Relationship between surface roughness and machining time.


· 301


1)

Figure 6. Optical micrographs showing surface contour of the (a-c) 60% and (d-f) 70% PS ingots machined under 236, 436, and 636 conditions.

302



c) f) Figure 7. Micrographs showing sub-surface microcrack of (a-c) 60% PS and (d-f) 70% PS ingots, machined under (a &d) 236, (b & e) 436, and (c &f) 636 conditions.

Figure 8. X-ray diffraction pattern analysis of relative monoclinic to tetragonal ratio.


· 303


Sintering of Multi-Material and Multi-Layered Systems


CO-SINTERING BEHAVIORS OF OXIDE BASED BI-MATERIALS Claude Carry,1 Emre Yalamac,1'2 Sedat Akkurt,2 'SIMAP, UMR5614 CNRS-INPG/UJF, ENSEEG, B.P. 75, 38402 Saint Martin D'Heres Cedex, France 2 Mechanical Engineering Department, Izmir Institute of Technology, 35430 Izmir, Turkey ABSTRACT Bi-materials have attracted attention due to combination of properties that such structures can offer. A strong bond between two co-sintered oxide ceramics can provide novel properties. This study focused on the densification and the microstructural evolution during co-sintering of alumina (Al O )zirconia (Y-ZrO ) and alumina-spinel (MgAl O ) bi-materials, produced by co-pressing of powders. High purity submicron powders were uniaxially pressed or co-pressed (150 or 250 MPa). The sintering behaviors of mono and bi-material bodies were investigated using a vertical dilatometer under constant heating rate conditions (from 1 to 10 cC/min up to 1580°C). Microstructural characterizations focused on the interface and diffusion layers of bonded bi-materials. Best bonding without cracks were observed on alumina-spinel bi-materials. Macroscopic and microscopic observations are analyzed, interpreted and discussed considering shrinkage and thermal expansion mismatches, residual stresses, diffusion kinetics and oxide phase diagrams. INTRODUCTION Bi-materials have functional properties, depending on mechanical, electrical and magnetic properties of their components. Their applications areas are ranging from electronic packaging applications such as multi-layer ceramic capacitors to thin film-substrate systems used widely in the microelectronics industry'1' . There are many types of bi-materials; metal-metal· ' , metal-oxide' and oxide-oxide'2'5,61. Die compaction of layers (powder stacking) is a simple and well established method. The disadvantages of the process are limited number of layers (not more than two or three in potential fabrication), limited size of the part ( 15 nm) submicrometre pores.

Figurel. Relation between the SAXS indicatrix invariant, Jq2 and diffraction vector q: 1- is NbH powder; 2- is Nb powder. The hydrogenation assists the growth of the sintered powder specific surface via stimulation of gas vacancy pore formation processes (by Skatkov et al). To a certain extent, it promotes the formation of the labyrinth porous structure, which is more resistant to subsequent external effects caused by the predominating volume contribution of large submicrometre pores, which are surface exposed. Evidently, this, along with alternative factors, which aid in improving the NbH based MDS capacitor system properties, is a decisive factor in the production development . However, the porous structure of hydrogenated powders is characterised by the rather high volume of small submicrometre pore low stable fractions4.

430


SAXS Investigation of the Sintered Niobium Powder

Nitriding of NbH powders brings about a decrease in the SAXS intensity, primarily in the region of super small angle scatterings. In addition, the nature of angular intensity distribution also changes, the SAXS indicatrix integral width increases, and their asymptotics appreciably change. The correlation peaks caused by the predominance of scattering inhomogeneities of a certain shape, disappear (Fig.2 ). In the specimens under study, SAXS was mainly a result of scattering on the electron density volumetric inhomogeneities (i.e. submicrometre pores), whose minimum size exceeds the irradiation wavelength ( 0,7 nm ), whereas its maximum size is limited by the largest possible degree of localization of primary X-ray beam and does not exceed 150 nm in the system used. In this connection, the observed SAXS picture may be interpreted as representing variations in submicrometre pore morphology and volumetric concentration in the given dimensional range.

! 3

lgq, (nm"1)

Figure 2. Double logarithmic J-q dependence: 1 - is un-nitrided NbH powder; 2 - is nitrided NbH powder. Viewed as a gas - solid kinetics reaction, the transformation of the powder's porous structure during its nitriding (detected by the SAXS technique ) is a result of volumetric and structural changes. These changes stem from diffusion, i.e. chemical reactions on pore surfaces, formation of gas vacancy complexes, and their mutual interactions with free surface, inner interfaces, and with structural imperfections5. The analysis of the changes in angular distribution and the SAXS intensity level, as well as the integral parameters, scattering indicatrix asymptotics, and calculations of the submicrometre pore sizes and concentrations suggest the following conclusion, namely that the porous system is noted for nitriding, which differs from the initial system by the greater number of submicrometre pores formed in different dimensional fractions6. This redistribution appears as a dramatic decrease of small submicrometre pores ( < 15 nm) (Table 1 ), probably owing to the pores being blocked by the products of interaction of sintered powder particles and the gas phase. The decrease in total submicrometre pore volumetric concentration is not important from the practical viewpoint, since the portion of large ( > 15 nm ) and mainly open submicrometre pores, which


· 431


determines the useful porosity, increases considerably (Table 1 ). The contribution of small submicrometre pores, which are easily blocked while the capacitor system operates, can be ignored for the corresponding analysis of the process. Thus, nitriding assists in generating a porous system and stabilizing it, and facilitates oxide restoration in the surface electrolyte contact areas. PARTB INTRODUCTION The present communication reports an experimental study of the surface fractal dimension D of porous solid niobium obtained by vacuum sintering of niobium powder. Here we present observation of the modified Porod law and also give stipulation of the obtained surface dimension which are based on the employment of the independent method of Hg porosimetry. RESULTS AND DISCUSSION As is known7,8' for describing describing X-ray scattering even in the Porod law the well known q law must be modified. The main formula of this theory which is useful for explaining our results is as follows I(q) ~ ( constant x q- (6 - D > )

(1)

Here I(q) is the X-ray scattering intensity; q is the wave vector and the D is the dimension of the object being irradiated (in our case D is a surface dimension) and the only discrepancy between the theory obtained by Wong 9 and that by Bale and Schmidt I0 is in the prefactor. The scattered intensity/wavevector relationship ( Fig.3 ) shows the fractal behaviour. Indeed, on the graph the angle of the slope of the curve part, which can be closely approximated by a line, is of the order of 73° which corresponds to the following power law: I(q) ~q"3'' , because the angle coefficient of line is tg 73° =3,19 and correspondingly the value of exponential index (6 - D) in formula (1) is 3,19 i.e. the surface dimension Ds = 2,81 (the subscript "s" denotes values obtained by experiments based on the SAXS ).

lgq, (lOnm") (lOnm 1 ) lgq, Figure 3. Logarithmic dependence of SAXS intensity J vs. the wave vector q for the porous solid Nb: (+) - experimental data: (-) - approximation linear range.

432



As it was pointed out by Wong , complete treatment of the data obtained by SAXS, as well as by any method based on reflected radiation, is not unique. Another independent method should be used to support the measurements of the fractal dimension. That is why we made use of Hg porosimetry. Due to the ability of the device to provide quite a high pressure we could investigate open pores of small radii reaching the region of the SAXS validity. This technique allows the distribution of open pores to be investigated by injecting mercury under pressure. As is clear, a pressure increase allows one to take into account pores of lower radii. If one considers the smallest pore radius to be appropriate the ε (smallest size of the fractal system) , for the surface area measure then a pressure rise means in fact a transition to a smaller scale and has to give as a result the power dependence of the specific surface area Sp on the smallest pore radius Rp. The surface dimension Dp (the subscript "p" stands for values obtained by Hg porosimetry) can be determined in this case by relationship: S p ~R< 2 - D p '{~e ( 2 ' D p )}

(2)

well known in fractal theory 8. We have observed the identical behaviour. It is represented in Fig.4 where the dependence of the open pore surface area of the sample is plotted as a function of the lowest pore radius registered by Hg porosimetry. The fact that the surface area dimension is predicted by the SAXS to be greater than 2 means that the cumulative surface area is actually determined by porosity. Hence Sp, in Fig.4, can be equated to the total effective surface area. The simple estimate of the angle of the slope approximated by a line yields the surface dimension D p equal to 2,84 ( in accordance with relationship (2)). This result is in full compliance with that obtained by the SAXS method and gives further experimental support to the law formula (1).

lgRp,(10"'nm)

Figure 4. Logarithmic dependence of the open pore area Sp vs pore radius Rp obtained by Hg porosimetry.


· 433


CONCLUSIONS The main results of Part A investigation are: Al. Niobium powder hydrogenation leads to the formation of porous structures with predominancy of open submicrometre pores; A2. Nitriding of NbH sintered powders stabilies the porous structure through coarsening, i.e. the volume contribution reductions of small submicrometre pores. The main results of Part B investigations are: Bl. The surface dimensions are stated to be of the order of 2,8 which is a stipulation of a highly developed porous structure; B2. Our results provide experimental support to the SAXS theory developed earlier. B3. The "combined" application of the SAXS and Hg porosimetry method to investigate the structural surface inhomogeneities has allowed not only to choose the most appropriate approximation of the porous shapes ( according to the SAXS data), but also to take into consideration the polymodality of the investigated submicropore system by revealing ( on the basis of Hg porosimetry ) correlation between the shape and the radius of the pores.

Table 1. Volume concentration of submicrometre pores with different sized fractions ( R is pore size ) R,nm Specimen

2-5

5-6

15-17,5

23-33

C,%

Un-nitridedNbH

2,5

4,5

8,1

11,9

27

NitridedNbH

0,8

1,3

0/7

10¿

13

REFERENCES 1 A. Fedorenko, V. Starikov, Y. Pozdeev, and N. Lykov, Layer Systems on the Base of NitrogenDoped Tantalum and Niobium with Enhanced Stability, Cryst.Res. Technol. 32, 843-48(1997). 2 L. Skatkov, P. Cheremskoy, V. Gomozov, and B. Bayrachny, Investigation of the Solid Surface Structure Inhomogeneities by the "Combined" Small-Angle X-Ray Scattering, Appl.Swf.Sci. 99, 367-70 (1997). 3 O.Glater and O.Kratky, Small Angle X-ray Scattering, Academic Press, LondonNew-York (1982). 4 L.Skatkov, P.Cheremskoy, V.Gomozov, and B.Bayrachny, Study of Porosity of Compacted Structures Formed by Vacuum Sintering of Niobium Hydride Powder, Fiz. i him. Obrab. Mater. ,6, 157-59 (1994).

434



5

B. Bayrachny, P. Cheremskoy, V. Gomozov, L. Murovtsev, and L. Skatkov, Preparation and Characterization of Submicropores in MnC>2 Semiconductor Films, Thin.Sol.Films, 201, L7-L8 (1991). 6 L.Skatkov and P.Cheremskoy The Ways of Porouse Structure Stabilization in Niobium Based Sintered Powder, Fiz.i him. Obrab. Mater.,5, 117 - 20 (1996). 7 J. Feder, Fractals, Plenum, New York (1988). 8 A.Hurd, D. Schaefer, D. Smith, S. Ross, A. Le'Mehaute, and S. Spooner, in: Progress in Electromagnetics Res. Symp., Cambridge (1989). 9 Po-zen Wong, Scattering by Inhomogeneous Systems with Rough Internal Surfaces: Porous Solids and Random-Field Ising Systems, Phys.Rev., B32, 7417-24 (1985). 10 H. Bale, and P.Schmidt, Small-Angle X-Ray Scattering Investigation Submicroscopic Porosity with Fractal Properties, Phys.Rev.Lett.,S3, 596-99 (1984).


· 435


Author Index

Aguilar, A., 53 Akkurt, S., 307 Amigó, V., 273, 367 Banhart, J., 85 Battabyal, M., 211 Bernard, F., 357 Berroth, K., 379 Bocanegra, M., 53 Bonache, V., 367 Bordia, R. K., 125 Bormann, R., 13 Bouvard, D., 173, 181 Brown, G., 103 Busquéis, D., 273, 367

Dariel, M. P., 29 Dateraksa, K., 291 DeCarlo, K. J., 61 Degenhardt, U., 379 Denti, L, 259 Dilman, H., 29 Domínguez, C , 53 Doorenbos, Z., 237 Dornheim, M., 13 Dub, S., 29 Elizalde, J. T., 53 Fang, Z. Z., 389 Foghmoes, S., 3 Frage, N., 29

Cabouro, G., 357 Calero, J. A., 273 Camacho, H., 53 Campbell, L. G., 71 Carry, C. P., 173, 181,307 Carty, W. M., 61 Castro, A., 367 Chaix, J. M., 343 Charmond, S., 173, 181 Chevalier, S., 357 Chunkiri, T., 291 Ciasen, R., 227, 333

Gaffet, E., 357 Galusek, D.,193, 421 Garcia, A., 53 Garcia, P. E., 53 Gatto, A., 259 German, R. M., 71, 149 Ghillányová, K., 421 Grigoryev, E. G., 205 Grin, Y., 357 Grupp, R., 85 Gwathney, K., 21

Danzer, R., 379

Harrer, W., 379

Author Index

Hasan, M. M., 283 Hayun, S., 29 He, Z., 3 Hoffmann, M., 193 Horstemeyer, M. F., 149 Huang, R., 135 Ishizaki, K.,211 Islam, F., 283 Jaramillo-Vigueras, D., 13 Jiang, W., 401 Jiang, Y., 401 Jo, T. S., 415 Johnson, J . L , 71 Kaiser, A., 3 Kamseu, E., 91 Karaman, D., 193 Kharatyan, S., 237 Khachatryan, H., 237 Kieback, B., 85 Kim, D.-G..415 Kim, S., 149 Kim, S.-G., 149 Kim, S. H., 415 Kim, Y. D., 415 Klassen, T., 13 Krenkel, W., 379 Kumar, V., 389 Lam, T. F., 61 Le Gallet, S., 357 Leonelli, C 91,259 Levine, R., 103 Ma, J., 113 Martin, C. L , 125 Martínez, C. A., 53 Martínez-Franco, E., 13 Meek, T. T., 21 Michálková, M., 193,421 Michalski, A.,219 Missiaen, J.-M., 321 Mitteau, R., 321

438

· Advances in Sintering

Moitra, A., 149 Motz, G., 379 Müller, M., 227 Narula, C. K., 21 Ngernbamrung, S., 291 Nicula, R.,211 Nöthe, M., 85 Ochoa, H. J., 53 Olevsky, E. A., 103, 113 Özüdogru, R. E., 249 Pan, J., 135 Park, M . S . , 415 Park, S. J., 71,149 Pascal, C , 343 Perera, D. S., 91 Phair, J . W., 3 Pittini, Y. Y., 211 Poli, G., 259 Puszynski, J. A., 237 Raharijaona, J.-J., 321 Rammohan, A. R., 401 Ramousse, S., 3 Reig, L , 273 Rivinius, C , 333 Rosinski, M „ 219 Rosliakov, A. V., 205 Sahin, F. Q., 249 Salvador, M. D., 273, 367 Saunier, S., 41 Sedlácek, J., 193 Shinagawa, K., 161 Skatkov, L., 429 Sitthiseripratip, K., 291 Stegner, F., 379 Sujirote, K., 291

Thomazic, A., 343 Tikare, V., 103 Valdivieso, F., 41

and Technology

Author Index

Vats, A., 237 Vaucher, S., 211 Veronesi, P., 259

Williams, J. L , 401 Wong, M., 401 Wongcumchang, M., 291

Walker, L. R., 21 Wang, H., 389 Wang, X., 389

Yalamac, E., 307 Yucel, O., 249


· 439

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

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

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

Advances in Electroceramic Materials: Ceramic Transactions, Volume 204 (Ceramic Transactions Series)

Advances in Porcelain Enamel Technology: Ceramic Transactions Volume 211 (Ceramic Transactions Series)

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

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

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

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

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

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

Ceramic Processing and Sintering

Processing and Properties of Advanced Ceramics and Composite Ceramic Transactions

Advances in Ceramic Armor VII: Ceramic Engineering and Science Proceedings

Advances in Ceramic Armor V (Ceramic Engineering and Science Proceedings)

Advances in Ceramic Armor III: Ceramic Engineering and Science Proceedings

Advances in Materials Science for Environmental and Nuclear Technology II: Ceramic Transactions

Processing of High Temperature Superconductors: Proceedings of the symposium held at the 104th Annual Meeting of The American Ceramic Society, April 28-May 1, ... Transactions (Ceramic Transactions Series)

Environmental Issuesand Waste Management Technologies in the Cermaic & Nuclear (Ceramic Transactions Series)

Ceramic Hardness

Ceramic Materials

Ceramic Fabrication Technology

Ceramic Industry February 2011

Ceramic Industry November 2011

Advances in Joining of Ceramics: Proceedings of the symposium held at the 104th Annual Meeting of The American Ceramic Society, April 28-May1, 2002 in ... Transactions (Ceramic Transactions Series)

Ceramic Industry March 2011

Handbook of Ceramic Composites

Ceramic Industry December 2011

Ceramic Industry October 2011

Ceramic Industry July 2011

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

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

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

Advances in Electroceramic Materials: Ceramic Transactions, Volume 204 (Ceramic Transactions Series)

Advances in Porcelain Enamel Technology: Ceramic Transactions Volume 211 (Ceramic Transactions Series)

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

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

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

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

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

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

Ceramic Processing and Sintering

Processing and Properties of Advanced Ceramics and Composite Ceramic Transactions

Advances in Ceramic Armor VII: Ceramic Engineering and Science Proceedings

Advances in Ceramic Armor V (Ceramic Engineering and Science Proceedings)

Advances in Ceramic Armor III: Ceramic Engineering and Science Proceedings

Advances in Materials Science for Environmental and Nuclear Technology II: Ceramic Transactions

Processing of High Temperature Superconductors: Proceedings of the symposium held at the 104th Annual Meeting of The American Ceramic Society, April 28-May 1, ... Transactions (Ceramic Transactions Series)

Environmental Issuesand Waste Management Technologies in the Cermaic & Nuclear (Ceramic Transactions Series)

Ceramic Hardness

Ceramic Materials

Ceramic Fabrication Technology

Ceramic Industry February 2011

Ceramic Industry November 2011

Advances in Joining of Ceramics: Proceedings of the symposium held at the 104th Annual Meeting of The American Ceramic Society, April 28-May1, 2002 in ... Transactions (Ceramic Transactions Series)

Ceramic Industry March 2011

Handbook of Ceramic Composites

Ceramic Industry December 2011

Ceramic Industry October 2011

Ceramic Industry July 2011

Recommend Documents