Nitride Ceramics: Combustion Synthesis, Properties and Applications [1 ed.] 3527337555, 9783527337552

Citation preview

Edited by Alexander A. Gromov and Liudmila N. Chukhlomina Nitride Ceramics

Related Titles Gromov, A.A., Teipel, U. (eds.)

Morkoc, H.

Metal Nanopowders

Nitride Semiconductor Devices

Production, Characterization, and Energetic Applications 2014 ISBN: 978-3-527-33361-5 (Also available in a variety of electronic formats)

Chen, Z., Dodelet, J., Zhang, J. (eds.)

Non-Noble Metal Fuel Cell Catalysts 2014 ISBN: 978-3-527-33324-0 (Also available in a variety of electronic formats)

Kääriäinen, T., Cameron, D., Kääriäinen, M., Sherman, A.

Atomic Layer Deposition Principles, Characteristics, and Nanotechnology Applications Second Edition 2013 ISBN: 978-1-118-06277-7 (Also available in a variety of electronic formats)

Lackner, M., Palotás, Á.B., Winter, F.

Combustion From Basics to Applications 2013 ISBN: 978-3-527-33376-9 (Also available in a variety of electronic formats) ISBN: 978-3-527-66718-5

Fundamentals and Applications 2013 ISBN: 978-3-527-41101-6 (Also available in a variety of electronic formats)

Barsoum, M.

MAX Phases Properties of Machinable Ternary Carbides and Nitrides 2013 ISBN: 978-3-527-33011-9 (Also available in a variety of electronic formats)

Zhao, D., Wan, Y., Zhou, W.

Ordered Mesoporous Materials 2012 ISBN: 978-3-527-32635-8 (Also available in a variety of electronic formats)

Schubert, U., Hüsing, N.

Synthesis of Inorganic Materials Third Edition 2012 ISBN: 978-3-527-32714-0

Edited by Alexander A. Gromov and Liudmila N. Chukhlomina

Nitride Ceramics Combustion Synthesis, Properties, and Applications

The Editors Prof. Dr. Alexander A. Gromov

Energetic Institute Tomsk Polytechnic University 30 Lenin Prospekt Tomsk, 634050 Russia and Technische Hochschule Nürnberg Georg S. Ohm Fakultät Verfahrenstechnik Wassertorstr. 10 90489 Nürnberg Germany Dr. Liudmila N. Chukhlomina

Russian Academy of Science Department of Structural Macrokinetics Tomsk Scientific Center 10/3 Akademichesky Ave. Tomsk, 634021 Russia

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.d-nb.de. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Print ISBN: 978-3-527-33755-2 ePDF ISBN: 978-3-527-68452-6 ePub ISBN: 978-3-527-68454-0 Mobi ISBN: 978-3-527-68455-7 oBook ISBN: 978-3-527-68453-3 Cover-Design Bluesea Design, McLeese Lake, Canada Typesetting Laserwords Private Limited, Chennai, India Printing and Binding

Printed on acid-free paper

V

Dedicated to the great Russian Scientist Alexander Grigorevich Merzhanov.

VII

Contents Foreword XV List of Contributors XIX Preface XXIII 1

Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation 1 Inna P. Borovinskaya, Vazgen E. Loryan, and Vladimir V. Zakorzhevsky

1.1 1.2

Introduction 1 Peculiarities of Phase and Structure Formation of Metal and Nonmetal Nitrides in Combustion Mode 2 Systems of Transition Metal of the IV–V Groups of the Periodic Table with Nitrogen 2 Two Threshold Mechanisms of Combustion and Structure Formation 5 Structural and Morphological Peculiarities of Silicon, Boron, and Aluminum Nitrides – Nanosized Powders 7 Dependence of SHS Nitride Composition and Structure on Infiltration Combustion Mode 20 Two Stages of SHS Process 21 Nitride Dissociation 26 Role of Nitrogen Admission to the Reaction Zone 30 SHS Equipment for Powder Synthesis 32 Synthesis of SHS-Ceramics Based on Silicon and Aluminum Nitrides and SiAlON Powders 33 Silicon Nitride Powders and Items 33 SiAlON Powders and Items 35 Aluminum Nitride Powders and Items 37 Direct Production of Materials and Items Based on Nitride Ceramics by SHS Gasostating 38 Nitride Ceramics Based on SiAlONs 39 Nitride Ceramics Based on BN 40 Nitride Ceramics Based on AlN 42

1.2.1 1.2.1.1 1.2.1.2 1.3 1.3.1 1.3.1.1 1.3.1.2 1.4 1.5 1.5.1 1.5.2 1.5.3 1.6 1.6.1 1.6.2 1.6.3

VIII

Contents

1.7

Conclusion 44 References 44

2

Combustion Synthesis of Boron Nitride Ceramics: Fundamentals and Applications 49 Alexander S. Mukasyan

2.1 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5

Introduction 49 Background: Brief Historical Overview 49 Key Properties and Markets Values 49 Applications of h-BN 51 Method of Synthesis: Advantages and Disadvantages 52 Combustion Synthesis in Gas–Solid Systems: General Definitions 53 Combustion in Boron–Nitrogen System 57 Thermodynamic Considerations 58 Conditions for Combustion Synthesis of Boron Nitride Material 59 Mechanism of Structure Formation in CS wave 62 Methods of Investigation of Structural Transformation in Combustion Wave 62 Mechanism of BN Formation 64 Combustion Synthesis of Nitride-Based Ceramics 67 CS of Boron Nitride Ceramics 68 Properties of BN Materials Synthesized by CS Technology 70 Examples of Special Application of CS-BN Ceramics 72 Final Remarks 72 References 73

2.2 2.2.1 2.2.2 2.3 2.3.1 2.3.2 2.4 2.4.1 2.4.2 2.4.3 2.5

3

Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies 75 Zhongqi Shi, Yoshinari Miyamoto, and Hailong Wang

3.1 3.2 3.2.1 3.2.2 3.3

Introduction 75 Combustion Synthesis of Quasi-Aligned AlN Nanowhiskers 76 Experimental Methods of Approach 76 Results and Discussion 78 Enhanced Thermal Conductivity of Polymer Composites Filled with 3D Brush-Like AlN Nanowhiskers by Combustion Method 83 Experimental Methods of Approach 84 Results and Discussion 84 Growth of Flower-Like AlN by Combustion Synthesis Assisted with Mechanical Activation 86 Experimental Methods of Approach 87 Results and Discussion 87 Combustion Synthesis of AlN Porous-Shell Hollow Spheres 90

3.3.1 3.3.2 3.4 3.4.1 3.4.2 3.5

Contents

3.5.1 3.5.2 3.6

Experimental Methods of Approach Results and Discussion 91 Summary and Conclusions 93 References 94

4

Combustion Synthesis and Spark Plasma Sintering of �-SiAlON 97 Xuemei Yi, Tomohiro Akiyama, and Kazuya Kurokawa

4.1 4.1.1 4.1.2 4.1.3 4.2 4.2.1 4.2.2 4.2.2.1 4.2.2.2 4.2.2.3 4.3 4.3.1 4.3.2 4.4 4.4.1 4.4.1.1 4.4.1.2 4.4.1.3 4.4.2 4.5

Introduction 97 β-SiAlON 97 Combustion Synthesis (CS) 98 Spark Plasma Sintering (SPS) 98 CS of High-Purity β-SiAlON and Densification by SPS Reaction Mechanisms 99 Dense β-SiAlON by CS and SPS 101 Combustion Synthesis of β-SiAlON Powder 101 Spark Plasma Sintering of CSed Powders 102 Characterization of CS–SPSed β-SiAlON 103 Physical Properties of CS-SPSed β-SiAlON 108 Vickers Hardness 108 Thermal Conductivity 109 Corrosion Resistance 112 Oxidation Behavior in Air 112 Oxidation Kinetics 113 Microstructure of Oxide Scale 113 Reaction Mechanisms 116 Corrosion Resistance in Supercritical Water 118 Conclusions of This Chapter 122 References 122

5

Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application 125 Alexander A. Gromov, Filippo Maggi, Ekaterina V. Malikova, Julia I. Pautova, Alexander P. Il’ in, Elena M. Popenko, Alexey V. Sergienko, Alexander G. Korotkikh, and Ulrich Teipel

5.1

Thermochemical Features of Aluminum Particles Combustion (Theoretical Background) 125 Aluminum–Oxygen Systems 127 Aluminum–Nitrogen Systems 129 Aluminum–Air Systems 129 Chemical Features of Metals Combustion in Air (Experimental Background) 131 Combustion of Aluminum Particles in Air 131 Combustion of Boron Particles in Air 132 Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air 133

5.1.1 5.1.2 5.1.3 5.2 5.2.1 5.2.2 5.3

90

99

IX

X

Contents

5.3.1 5.3.2 5.3.2.1 5.3.2.2 5.3.3 5.3.3.1 5.3.3.2 5.3.3.3 5.3.3.4 5.3.3.5 5.3.4 5.3.4.1 5.3.4.2 5.3.5 5.4 5.4.1 5.4.2 5.4.3

Nitrides Formation at nAl/μAl and [nAl + (μAl/Zr Alloy)] Combustion in Air 134 Nitrides Formation at nAl Combustion in Air 134 CCP Microstructure 137 Effect of Additives on AlN Yield 141 AlN (Al3 O3 N), ZrN, TiN Obtained by Combustion of Metal Powders and their Mixtures in Air 143 nAl 146 μAl 148 μTi 149 μZr 149 Combustion Scenario 149 Nitrides Obtained by Combustion of μTi/μAl and μTi/μTiO2 Mixtures in Air 152 Combustion of the Mixtures I (“Ti–TiO2 ”) 154 Combustion of the Mixtures II (“Ti–Al”) 154 Combustion Synthesis of Aluminum Oxynitride in Air 154 Application of the Synthesized Nitrides and Oxynitrides in Dense Ceramics 156 Nitride Ceramics on the Base of the CCP in the System “Zr–O–N” 156 Nitride Containing Ceramics on the Base of the CCP in the System “Al–O–N” 156 Technology of Nitride Ceramics Production on the Basis of the CCP in the System “Me(Al, Ti, Zr)–O–N” 159 References 160

6

Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum 165 Chun-Liang Yeh

6.1 6.2 6.3 6.3.1 6.3.2 6.3.3 6.4

Introduction 165 Experimental Methods of Approach 166 Results and Discussion 167 Combustion Synthesis of Vanadium Nitride 167 Combustion Synthesis of Niobium Nitride 173 Combustion Synthesis of Tantalum Nitride 177 Conclusions 180 References 183

7

Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen 185 Liudmila N. Chukhlomina, Yury M. Maksimov, and Lidiya N. Skvortsova

7.1 7.2

Introduction 185 Synthesis of Silicon Nitride by Combustion of Ferrosilicon in Nitrogen 186

Contents

7.2.1 7.2.2 7.2.3 7.2.4 7.2.5 7.3 7.4 7.5 7.5.1 7.5.2 7.6 7.7 7.8 7.8.1

7.8.2

7.9

Nitriding Degree of Combustion Products and Burning Rate Versus Main SHS Parameters 186 Filtration Combustion Modes of Ferrosilicon in Nitrogen 188 Dilution of Initial Ferrosilicon by Previously Nitrided Ferrosilicon 190 Influence of Content of Silicon in Initial Fe–Si Alloys on SHS of Silicon Nitride 192 Mechanism of Structure and Phase Formation of Silicon Nitride during Combustion of Ferrosilicon in Nitrogen 195 Synthesis of Vanadium Nitride by Combustion of Ferrovanadium in Nitrogen 199 Synthesis of Niobium Nitride by Combustion of Ferroniobium in Nitrogen 204 Synthesis of Titanium Nitride by Combustion of Ferrotitanium in Nitrogen 206 Features of Ferrotitanium Nitriding 207 Phase-Formation Processes of Titanium Nitrides During Combustion of Ferrotitanium in Nitrogen 210 Combustion of Ferrochromium in Nitrogen and Synthesis of Chromium Nitride 213 Combustion of Ferroboron in Nitrogen and Synthesis of Boron Nitride 215 Application Prospects of Products of Combustion of Ferroalloys in Nitrogen 217 Application of Fe-Containing Composite Materials Based on Silicon and Boron Nitride for the Catalytic Destruction of Dissolved Organics 219 Boron Nitride-Based Composites in Combined Processes of Degradation of Dissolved Organics and Generation of Molecular Hydrogen 224 Conclusions 226 References 227

8

Halides in SHS Azide Technology of Nitrides Obtaining 229 Georgy V. Bichurov

8.1 8.2 8.3 8.4

Introduction 229 The Use of Ammonia Halides 231 The Use of Halides of Elements to be Nitrided 232 The Use of Complexing Halides of Elements to be Nitrided and Alkaline Metals 233 The Use of Complexing Halides of Ammonia and Elements to be Nitrided 234 The Use of Halides for Obtaining Refractory Compositions 234 Efficiency of Use of Halides in Azide SHS Systems 235 Chemical Stages of Formation of Nitrides in a Mode SHS-Az 236

8.5 8.6 8.7 8.8

XI

XII

Contents

8.9 8.10 8.11 8.11.1 8.11.2 8.12

Property of SHS-Az Powders 241 Property of SHS-Az Ceramics 243 The Synthesis of Nanostructural SHS-Az Powders 246 Nitride of Titanium, Boron, and Silicon 246 Aluminum Nitride 252 Conclusion 261 References 261

9

AlN Ceramics from Nanosized Plasma Processed Powder, its Properties and Application 265 Laima Trinkler, Baiba Berzina, and Eriks Palcevskis

9.1

Introduction: AlN Ceramics, its Characteristics and Application 265 Production of AlN Ceramics from Nanosized Plasma Processed Powder 266 Manufacturing of AlN Nanopowder 266 Production of AlN Ceramics with High Thermal Conductivity 269 Properties of AlN Ceramics from Nanosized Plasma Processed Powder 272 Dielectric Properties 272 Luminescence Properties 274 Experimental Details 274 Photoluminescence 275 Afterglow 278 Thermoluminescence 279 Optically Stimulated Luminescence 281 Fading of TL and OSL Signal 282 Luminescence Mechanism 282 Practical Application of Luminescence Properties of AlN Ceramics 286 Dosimetric Properties 286 TL and OSL After Exposure to Ionizing Radiation 287 TL and OSL after Exposure to UV Light 288 Estimation of AlN Ceramics as Dosimeter Material 289 Using Long-Lasting Luminescence 290 Using of Translucent Doped AlN Ceramics as Solid-State Lasers 290 Conclusions 291 References 291

9.2 9.2.1 9.2.2 9.3 9.3.1 9.3.2 9.3.2.1 9.3.2.2 9.3.2.3 9.3.2.4 9.3.2.5 9.3.2.6 9.3.2.7 9.4 9.4.1 9.4.1.1 9.4.1.2 9.4.1.3 9.4.2 9.4.3 9.5

10

An Overview of the Application of Nitrides and Oxynitrides in Photocatalysis and Electrocatalysis 295 Justin S. J. Hargreaves and Andrew R. McFarlane

10.1 10.2

Introduction 295 Preparation 297

Contents

10.2.1 10.2.2 10.2.3 10.2.4 10.3 10.4 10.5

Ammonolysis 297 Preparation Using N2 /H2 Mixtures 298 Metathesis Routes 299 Sol–Gel and Related Routes 300 Photocatalysis 302 Electrocatalysis 308 Conclusion 316 References 316

11

Conclusion 321 Alexander A. Gromov and Liudmila N. Chukhlomina

Reference Index 325

323

XIII

XV

Foreword Although we live in a world filled with nitrogen, that is, the main constituents of our atmosphere are nitrogen with 78% and oxygen with about 21%, nitrides or nitridic minerals are very rare in nature. This is caused by the fact that nitrogen is much less reactive than oxygen and even lesser than carbon. So, most of the natural minerals are more or less of oxidic nature. Nevertheless, a broad range of nitrogen compounds or nitrides exist. An overview on existing nitrogen compounds was compiled by Kieffer and Benesovsky [1] and is given in Table 1. The positions of the most interesting nitrides are accentuated in Table 1. These are the metallic nitrides of the group 4B to 6B and also the so-called diamond-like or nonmetallic nitrides of the group 3A and 4A. The nitrides of the group nos.1–3 are salt-like, while those of the groups 7B and 8 are metallic. The nitrides of the Lanthanides and Actinides have a transitional position; the actinide nitrides in Table 1 can be numbered among the metallic species. The majority of the remaining nitrogen compounds of the groups 5A–7A is gaseous. Of great technical interest are the nitrides in the accentuated fields in Table 1. The metallic nitrides, especially those of the 4B and 6B groups show very high hardness, high abrasion resistance, and are characterized by very high melting temperatures. The stability of these nitrides decreases from group 4B to 6B and within the 6B group from the relatively stable chromium nitrides via molybdenum nitride to tungsten nitride. Furthermore, it is important to state that the stability of these nitrides, especially at higher temperatures up to the melting point, is strongly dependent on the nitrogen pressure. The mononitrides of the 4B and 5B metals crystallize in an fcc lattice with one exception: TaN at ambient conditions shows an hcp lattice. Only at enhanced temperatures (>about 1500 ∘ C) and enhanced nitrogen pressures (>about 5 bar), TaN is transferred into an fcc lattice. Other interesting properties of these metallic nitrides are their excellent electrical and thermal conductivity as well as in the case of TiN, the gold-like color. So TiN coatings are used not only on cutting tools to enhance the abrasion resistance but also for decorative – and scratch-resistant – surfaces of various metallic articles.

XVI

Foreword

Table 1 Nitrogen compounds in the periodic table of elements according to Kieffer and Benesovsky [1].

1A

2A

3B

4B

5B

6B

7B

8

1B

2B

3A

4A

5A

6A

7A

N

OxN

F3N

Si3N4

PxN

SxN

Zn3N2 GaN Ge3N4

AsN

SeN

Cl3N

Cd3N2

InN

Sn3N4

SbN

TeN

Br3N

Hg3N2

TlN

Pb3N4

BiN

TbN

DyN

HoN

ErN

H3N Li3N

Be3N3

Na3N

Mg3N2

BN AlN Mn4N Fe4N Mn2N Fe2N Mn3N2

K3N

Ca3N2

ScN

Ti2N TiN

V2 N VN

Cr2N CrN

Rb3N

Sr3N2

YN

ZrN

Nb2N Nb4N3 NbN

Mo2N MoN

Cs3N

Ba2N Ba3N2

LaN

Hf2N HfN

Ta2N TaN

W2N WN

Re2N

CeN

PrN

NdN

Co3N Ni3N Co2N Ni3N2

Cu3N Ag2N

J3N

Lanthanides SmN

EuN

GdN

TmN

YbN

Actinides ThN Th2N3

UN U2N3 UN2

NpN

PuN

Regarding the three nonmetallic nitrides in Table 1, the technical importance of these is partly similar to that of the metallic nitrides but partly quite different. Boron nitride (BN) exists in two modifications with completely different properties. The hexagonal α-boron nitride is according to its crystal lattice and its properties very similar to graphite, that is, it crystallizes in a hexagonal lattice and consists of layers of a planar, hexagonal honeycomb structure with alternating B- and N-atoms. This BN modification is also called “white graphite.” It is, for example, used as a solid lubricant as well as a material for high-temperature and oxidation-resistant crucibles. Cubic boron nitride (CBN) is the second hardest material behind diamond. So CBN is predominantly used as material for cutting tools or as grinding material. The outstanding property of AlN is its high thermal conductivity up to 180–220 W m−1 K−1 and it is an electrical insulator. So AlN is used as a ceramic heat sink in microelectronics. The last of the mentioned nonmetallic nitrides is silicon nitride (Si3 N4 ). Since about 50 years, the importance of Si3 N4 as a technical ceramic increased permanently. Si3 N4 combines high hardness and abrasion resistance with a relatively low density of about 3.24 g cm−3 and good oxidation resistance up to

Foreword

high temperatures. These properties led to the application of Si3 N4 as material for ball bearings, nozzles, valves in automotive engines, and many more. For all the nitrides mentioned here, there are – besides or because of their outstanding properties – two main challenges:

• Synthesis of pure materials. • Processing and shaping to technical components. The classical ways of nitride synthesis are the following:

• • • • • •

Nitriding of metals or of hydrides. Nitriding of metal oxides in the presence of carbon. Conversion of metal chlorides and metal oxychlorides with ammonia. Decomposition of ammonia compounds. Conversion of oxides with calcium nitride. Gas-phase deposition of metal halides.

The classical processing path of shaping these refractory materials to bulk bodies is sintering. This means the conversion of a more or less porous structure into a dense body. During normal sintering, the driving force for the densification is the reduction of the surface energy. This process is initiated by applying a sufficient amount of activation energy, mainly by heating from an external source (sintering furnace) up to the sintering temperature. But especially regarding refractory materials like nitrides, the necessary temperatures are very high and require rather expensive processing equipment. One way out of this problem is to implement a process, where the necessary heat of reaction is generated by an intrinsic chemical reaction. This is performed by using special precursors as starting materials, which have a lower level of activation energy combined with an exothermic reaction, which leads to the desired material or the desired shape. The generic term for this processing method is “reaction bonding.” Within reaction bonding, one of the most interesting processes is combustion synthesis (CS), also called selfpropagating high-temperature synthesis” (SHS). As can be seen from the wording within this process, the main emphasis is laid on material synthesis, but it may also be used to produce bulk shapes. The characteristic feature of CS is, after initiation locally, the self-sustained propagation of a reaction wave through the heterogeneous mixture of reactants [2]. Since the process occurs at high temperatures, the method is ideally suited for the production of refractory materials with unusual properties such as advanced nitrides and nitrogen compounds. In its usual form, SHS is conducted starting from finely powdered reactants that are intimately mixed. The synthesis is initiated by local heating of one preferred end of the sample. Usually, heating is started from the top of the sample. After starting a wave of exothermic reaction runs through the sample. Thus, the SHS mode of reaction can be considered to be a well-organized wavelike propagation of the exothermic chemical reaction through a heterogeneous medium, followed by the synthesis of the desired condensed product. Regarding the synthesis or processing of nitrides or nitrogen compounds, the atmosphere during the CS process

XVII

XVIII

Foreword

plays an important role. By adjusting composition and pressure of the samplesurrounding atmosphere, the products’ shape and properties can be targeted and controlled. In the present book, the material’s spectrum concentrates on new or advanced synthesis methods for metallic nitrides and nitrogen compounds – TiN, VN, NbN, TaN, and CrN, as well as on the nonmetallic species – AlN, BN, and SiAlONs. Up to now, CS or SHS has become a promising choice for industrial fabrication because of its relatively low processing costs, high energy efficiency, and short processing time. These technologies shall update or extend the classical technologies and they shall lead to new materials with technologically important properties combined with more simple and more economic processing technologies Dr. Hans-Joachim Ritzhaupt-Kleissl References 1. Kieffer, R. and Benesovsky, B. (1963)

Hartstoffe, Springer-Verlag, Wien, p. 287. 2. Varma, A., Rogachev, A.S., Mukasyan, A.S., and Hwang, S. (1998) Combustion

synthesis of advanced materials: principles and applications. Adv. Chem. Eng., 24, 79–227.

XIX

List of Contributors Tomohiro Akiyama

Inna P. Borovinskaya

Hokkaido University Faculty of Engineering Center for Advanced Research of Energy and Materials Kita 13 Nishi 8, Kita-ku Sapporo 060-8628 Japan

Institute of Structural Macrokinetics and Material Sciences Russian Academy of Sciences Laboratory of Self-propagating High-temperature Synthesis Academician Osipyan Street 8 Chernogolovka Moscow Region 142432 Russia

Baiba Berzina

University of Latvia Institute of Solid State Physics Kengaraga Street 8 Riga 1063 Latvia Georgy V. Bichurov

Samara State Technical University Material Science, Powder Metallurgy, Nano Materials Physical Technologies Faculty 244 Molodogvardeyskaya Street Samara 443100 Russia

Liudmila N. Chukhlomina

Tomsk Scientific Center Siberian Branch of Russian Academy of Sciences Department of Structural Macrokinetics 10/3 Academichesky Avenue Tomsk 634021 Russia Alexander A. Gromov

Tomsk Polytechnic University Energetic Institute 30 Lenin Prospekt Tomsk, 634050 Russia and

XX

List of Contributors

Nürnberg Technical University Georg Simon Ohm Process Engineering Department Wassertorstraße 10 90489 Nürnberg Germany Justin S. J. Hargreaves

University of Glasgow School of Chemistry WestCHEM Joseph Black Building Glasgow G12 8QQ Scotland Alexander P. Il’in

Tomsk Polytechnic University Inorganic Chemistry Department Lenin prospekt 30 Tomsk 634050 Russia Alexander G. Korotkikh

Tomsk Polytechnic University Energetic Institute Atomic and Power Stations Department Lenin prospekt 30 Tomsk 634050 Russia

Vazgen E. Loryan

Institute of Structural Macrokinetics and Material Sciences Russian Academy of Sciences Laboratory of Self-propagating High-temperature Synthesis Academician Osipyan Street 8 Chernogolovka Moscow Region 142432 Russia Filippo Maggi

Politecnico di Milano SPLab Department of Aerospace Science and Technology Via G. La Masa 34 20156, Milano Italy Yury M. Maksimov

Tomsk Scientific Center Siberian Branch of Russian Academy of Sciences Department of Structural Macrokinetics 10/3 Academichesky Avenue Tomsk 634021 Russia Ekaterina V. Malikova

Kazuya Kurokawa

Hokkaido University Faculty of Engineering Center for Advanced Research of Energy and Materials Kita 13 Nishi 8, Kita-ku Sapporo 060-8628 Japan

Tomsk Polytechnic University Silicates and Nanomaterials Department Lenin prospekt 30 Tomsk 634050 Russia

List of Contributors

Andrew R. McFarlane

Elena M. Popenko

University of Glasgow School of Chemistry WestCHEM Joseph Black Building Glasgow, G12 8QQ Scotland

Altai State Technical University Bijsk Technologic Institute Department of Chemical Technology of Energy Saturated Materials Trofimova Street 27 Bijsk 659305 Russia

Yoshinari Miyamoto

Osaka University Joining and Welding Research Institute 11-1 Mihogaoka Ibaraki Osaka, 567-0047 Japan

Hans-Joachim Ritzhaupt-Kleissl

Karlruhe Institute of Technology Hermann-von-Helmholtz-Platz 1 76344 Eggenstein-Leopoldshafen Germany

Alexander S. Mukasyan

University of Notre Dame Department of Chemical and Biomolecular Engineering 210 Stinson-Remick Hall Notre Dame, IN 46556 USA Eriks Palcevskis

Riga Technical University Institute of Inorganic Chemistry Miera Street 34 Salaspils, 2169 Latvia Julia I. Pautova

Tomsk Polytechnic University Energetic Institute Atomic and Power Stations Department Lenin prospekt 30 Tomsk 634050 Russia

Alexey V. Sergienko

Altai State Technical University Bijsk Technologic Institute Department of Chemical Technology of Energy Saturated Materials Trofimova Street 27 Bijsk 659305 Russia Zhongqi Shi

Xi’an Jiaotong University State Key Laboratory for Mechanical Behavior of Materials School of Materials Science and Engineering No. 28, Xianning West Road, Xi’an Shaanxi 710049 China Lidiya N. Skvortsova

Tomsk State University Chemical Department 36 Lenina Avenue Tomsk 634050 Russia

XXI

XXII

List of Contributors

Ulrich Teipel

Xuemei Yi

Nürnberg Technical University Georg Simon Ohm Process Engineering Department Wassertorstraße 10 90489 Nürnberg Germany

Hokkaido University Faculty of Engineering Center for Advanced Research of Energy and Materials Kita 13 Nishi 8, Kita-ku Sapporo 060-8628 Japan

Laima Trinkler

University of Latvia Institute of Solid State Physics Kengaraga Street 8 Riga, 1063 Latvia Hailong Wang

Ningxia University School of Physics Electrical Information Engineering No. 217 Wencui Street (North) Yinchuan Ningxia 750021 China Chun-Liang Yeh

Feng Chia University Department of Aerospace and Systems Engineering 100 Wenhwa Road, Seatwen Taichung 40724 Taiwan

and Northwest A&F University College of Mechanical and Electronic Engineering Department of Mechanical Engineering Xinong Road 22, Yangling Shaanxi 712100 China Vladimir V. Zakorzhevsky

Institute of Structural Macrokinetics and Material Sciences Russian Academy of Sciences Laboratory of Self-propagating High-temperature Synthesis Academician Osipyan Street 8 Chernogolovka Moscow Region 142432 Russia

XXIII

Preface Widely used oxides for the production of technical ceramics are rooted in the past. In twenty-first century, non-oxide materials will play a decisive role in technical ceramics production, in particular materials based on nitrides. Nitride ceramic is characterized by stability of dielectric properties, high mechanical strength, heat resistance, chemical resistance in different environments, and so on. Oxynitride ceramics is an intermediary between the nitride and oxide ceramics, which is a cheaper and affordable technological solution. Combustion synthesis (CS) or self-propagated high-temperature synthesis (SHS), discovered in USSR, is a simple method used for nitrides and oxynitrides production because of their nonexpensive facility, high-rate exothermic reaction, and wide range of synthesized products. The cost of CS ceramic products is much lower than that of the products synthesized by “furnace” technologies. The rate of nitrides phase formation in the combustion wave “metal-nitrogen” is very fast (up to several moles per second).

Chapter

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. a)

Nitrides of element B

Al

Si

Ti

+ +

+

+

V

Cr

Oxynitrides

Zr

Nb

Hf

Ta

+

+

+

+

+

+

+

Other

SiAlON

CSa

Pra

Apa

+ + + + + + + + + +

+ +

+

+ + + + + + + + +

+

+ + +

+

+

+

+ + +

+ +

+ +

+

+

+ +

+ +

+

+

+ +

+

+

+ +

+

+

+

CS = combustion synthesis, Pr = properties of nitrides, Ap = application.

+

Chapter describe

AlON

+

+ + + +

XXIV

Preface

The formation mechanism, properties, and application of the nitrides and oxynitrides produced by CS are comprehensively studied in this book. The majority of industrially important nitrides are represented by all authors (table). October 2014

Alexander A. Gromov Liudmila N. Chukhlomina

1

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation Inna P. Borovinskaya, Vazgen E. Loryan, and Vladimir V. Zakorzhevsky

1.1 Introduction

Metal and nonmetal nitrides are remarkable compounds characterized by a great number of valuable and interesting properties such as high chemical stability in different aggressive media, heat resistance, ability to transition to superconducting state, excellent semiconducting and dielectric characteristics, high hardness which is sometimes close to that of diamond, and so on. Nitride ceramics is widely applied in various industries, for example, electronics, ferrous and nonferrous metallurgy, aerospace, and nuclear power engineering. It is not surprising that since the time when the first nitrides were obtained, active investigations have been carried out to understand the character of nitrides, their behavior and relationship of their properties with peculiarities of their structure, and chemical bonds. The increasing requirements to the quality and operation properties of nitride ceramics make the researchers improve the available synthesis methods and develop new efficient technologies. Self-propagating high-temperature synthesis (SHS) based on combustion processes [1–6] is one of the leading methods of investigation of theory and practice, structure and phase formation of nitrides in the combustion mode, development of new variations of synthesis and technology of nitrides and composite materials thereof, and items and parts based on nitride ceramics for various application purposes. By this time, a lot of articles have been published in all the spheres of the exploration. Their level is constantly being increased. In this chapter, we try to analyze the investigation results of regularities and mechanism of metal and nonmetal combustion in nitrogen as well as structure and phase formation of nitrides, which appear to be less known or unpublished in literature. Besides, we demonstrate some scientific and practical achievements in development of SHS powder technology of the most important nitrides, direct synthesis of SHS materials and items based on nitride ceramics, and some examples of their practical application. The presented information employs the experimental work carried out at the Institute of Structural Macrokinetics and Materials Science of the Russian Academy of Sciences. Nitride Ceramics: Combustion Synthesis, Properties, and Applications, First Edition. Edited by Alexander A. Gromov and Liudmila N. Chukhlomina. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

2

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation

1.2 Peculiarities of Phase and Structure Formation of Metal and Nonmetal Nitrides in Combustion Mode

The first regularities of nitride structure were established by Hagg [7]. He gave the definition of the primary structures in which nonmetal atoms penetrated into the metal crystal lattice and called them “penetration structures.” Face-centered cubic, volume-centered cubic, hexagonal compact, and hexagonal simple types of lattice are the most common for the penetration structures. According to Hagg, transition-metal nitrides (metal-like nitrides) are typical penetration structures. The penetration phases are often considered as heterodesmic compounds with complex superimposition of covalent, metal, and ionic bonds. Many scientists think that the combination of the bonds is the main reason of the extraordinary behavior and unique properties of nitrides. So-called nonmetal nitrides, for example, BN, Si3 N4 , AlN, and so on, are mainly characterized by covalent bonds and appear to be actual chemical compounds. The works in SHS studied the regularities and mechanism of phase and structure formation of both types of nitrides: metal-like and nonmetal. 1.2.1 Systems of Transition Metal of the IV–V Groups of the Periodic Table with Nitrogen

Single-phase solid solutions of nitrogen in metals are “lower” phases of the systems of metal (IV–V groups)-nitrogen formed directly in the course of SHS. The conditions and mechanism of their formation have been extensively studied in combustion of porous titanium and zirconium samples at gaseous nitrogen pressures of 0.1 up to 500 MPa [8–10]. Theoretical consideration of the possibility of combustion between metals and nonmetals forming merely solid solutions was carried out in [11]. The systems of Zr + N2 and Ti + N2 are the most highly exothermic. Homogeneity regions of α-solid solutions of nitrogen in these metals are fairly large and attain compositions of ZrN0.33 and TiN0.27 . Thermodynamic calculations and experiments have shown that the SHS process in Zr + N2 and Ti + N2 systems is possible to occur at the expense of solid solution formation. The dependence of the calculated adiabatic combustion temperature in Zr + N2 system on nitrogen content in the homogeneity region of the solid solution is given in Figure 1.1. The combustion temperatures are seen to be high. A necessary condition for the formation of single-phase solid solution of nitrogen in metals is the arrest of the reaction at the stage of the solid solution formation. The further nitriding of the solid solution can be ceased in several ways. The most common ones are the sample quenching in liquid argon and sharp gas drop immediately after the combustion front passage [1, 8]. The efficient way is also to create the conditions when after-burning (bulk after-nitriding of the samples heated in the

1.2

Peculiarities of Phase and Structure Formation of Metal and Nonmetal Nitrides

Tad 4000 3000 2000 α solid solution ZrNx

1000

0.1

0.2

0.33 m

Figure 1.1 Combustion temperature versus solid solution composition. m – N2 content (relative units).

combustion front) is unlikely to happen [1, 8]. In Zr + N2 and Ti + N2 systems, such unfavorable conditions are formed spontaneously since the sample under combustion often becomes sintered or molten (especially at nitrogen pressure of >3 MPa), and there is no access for the gas into the combustion zone. Other techniques eliminating the after-burning can be: confining the sample’s lateral surface in a gas-tight jacket, implementing the combustion process in the mixture of metals with nitrides in a pure surface combustion mode, which plays the role of a “chemical furnace” to facilitate the product homogenization or dissolving nitrogen with inert gas, for example, argon [8, 12]. Combustion of the mixtures of metals with nitrides in the “chemical furnace” mode is analogous to the furnace synthesis of nitrogen solid solutions in metals by homogenization. American scientist observed the combustion wave propagation due to the formation of TiN1−x thin layer with the formation of nitrogen solid solution in the quenched combustion products using X-ray analysis and scanning electron microscope (SEM) [13]. An important achievement in the studies of the direct combustion synthesis of solid solutions was the production of compositions with minimum nitrogen content (MeN0.13 –MeN0.22 ) and synthesis of single-phase oversaturated solid solutions of MeN0.34 –MeN0.45 with nitrogen content exceeding the value known from literature. This fact initiated the hypothesis of nitride formation in the combustion mode by means of saturating metal with nitrogen along with forming oversaturated solid solutions which are then decomposed to form a solid solution of a lower composition and nonstoichiometric nitride MeNx in contrast to commonly accepted mechanism of the reaction diffusion through the product film. In this case, the combustion rate is determined by the rate of the nonmetal dissolution in metal. At this stage, the major heat release takes place. This concept was presented in [14]. The chemical and phase analyses of the compositions from ZrN0.34 to ZrN0.57 prove that the combustion products are oversaturated solid solutions. They decompose when annealed or dissolved in some specific liquids to solid solutions

3

4

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation

of a lower composition and nonstoichiometric nitrides. ZrN0.34−0.57

dissolution

ZrN0.54−0.61

combustion −−−−−−−−→ nitrides product

in solution

ZrN0.17−0.44 −−−−→ α-solid

(1.1)

solutions precipitation

Metallographic investigations of the sample microstructure prove nitride evolution on the grain boundaries. Generalizing these concepts for various metal–nonmetal systems, one can assume that the systems with wide regions of homogeneity of solid solutions can burn, under certain conditions, involving the stage of the solid solution saturation without yielding any compounds. The systems with narrow regions of homogeneity of solid solutions under specific conditions (reactive mass melting, high pressures) can also produce solid solutions when burning with their subsequent decomposition or crystallization. As to Me–N systems with complex state diagrams (e.g., Ta–N, Nb–N) which comprise solid solutions with narrow regions of homogeneity, single-phase solid solutions in these systems are also formed during the combustion. However, it is difficult to isolate them. They can only be observed in the quenched products in the layers adjacent to the combustion front, so-called warming-up zones [15, 16]. Examples of interaction including nonmetal dissolving in metals in solid–solid systems were described by Vidavsky [17] for Zr–C systems, and by Itin and Bratchykov [18] for SHS intermetallics particularly for Ti–Co system. Aleksandrov and Boldyrev [19] and Holt with coworkers [20] performed direct observation of the dissolving stage during SHS product formation in Ni–Al and Ti–C systems by measuring the phase structure of the combustion wave by means of synchrotron radiation. Shteinberg with coworkers obtained unique results on carbon dissolving in titanium under thermal explosion. The authors thoroughly investigated the mechanism of this process and regularities of structure formation [21]. One of the interesting examples of chemical stages in the SHS processes with lower phase formation is the occurrence of the metastable phase of ε-Ti2 N during titanium combustion in gaseous nitrogen. The combustion process in this system was studied in our work [22, 23] as well as in the works carried out by Japanese scientists Hirao et al. [24] and American colleagues Munir et al. [25]. According to literature data, the phase of ε-Ti2 N is unstable and starts decomposing at temperatures exceeding 1000–1100 ∘ C. We first produced this phase when decomposing oversaturated solid solutions of nitrogen in titanium of TiN0.33 –TiN0.42 composition. Besides, ε-Ti2 N phase was discovered in the quenched products of titanium combustion in nitrogen at P = 300–450 MPa. The following studies by Karimov and Em with coworkers [26, 27] dealt with combustion products of Ti–N system. They used neutron diffractometer (� = 1.08 Å) which was installed at the thermal column of a nuclear reactor. In their experiments, the specimens were undergoing heat treatment (homogenizing annealing) in the temperature range of 1500–1300 K with quenching from 1700 K and

1.2

Peculiarities of Phase and Structure Formation of Metal and Nonmetal Nitrides

α+ε

α+δ

ε

ε+δ

Specimen

1500–1300 K

1300–600 K

Quenching 1700 K

After (800 K)

δ

δI 0.3 (a)

0.4

0.5

0.6

TiN0.45

ε+δ

ε + δ + δ′

δ

δ′

TiN0.42

ε

ε

ε

ε

TiN0.40

ε

ε

ε

ε

TiN0.38

ε

ε

ε

ε

TiN0.33

α+ε

α+ε

α+ε

α+ε+δ

0.7 (b)

Figure 1.2 Neutron diffraction study (a, b): (a) state diagram of Ti–N system and (b) phase composition of products after thermal treatment.

subsequent annealing of the products after quenching at 800 K. ε-Phase was found to be formed in the concentration range of 0.38 ≤ N/Ti ≤ 0.42. The εTi2 N1−x phase (0.24 ≥ x ≥ 0.16) is characterized by lower nitrogen content as compared to that of stoichiometric compound Ti2 N. The ordered tetragonal phase δ′ -Ti2 N was found to be formed at 800 K on the base of metastable hightemperature cubic δ-phase. Its homogeneity region is in the range of 0 ≤ x ≤ 0.1. Based on the obtained data, a more precise diagram of the Ti–N system was suggested. A fragment of this diagram and phase composition of the products after heat treatment are presented in Figure 1.2. 1.2.1.1 Two Threshold Mechanisms of Combustion and Structure Formation

A hypothesis on stepped formation of final SHS products through formation of amorphous substances being primary combustion products was suggested by Merzhanov [28] on the base of experimental studies. It should be noted that certain amorphous phases of some SHS nitrides were isolated and identified in earlier works [8]. Most of the nitrides synthesized at optimum combustion terms are polycrystals. Usually, they are in equilibrium state and have distinct, narrow characteristic peaks in X-ray patterns. Many of them can be considered as typical examples of equilibrium compounds. Nevertheless, under special or nonoptimum synthesis conditions and, particularly, during cooling (natural or forced quenching), the formation of nonequilibrium products with blurred and broadened lines in X-ray patterns is observed. It proves the appearance of amorphous phases, which later acquire a certain crystalline structure depending on the process conditions (combustion and cooling temperature and rate, presence of crystallizers, etc.). The formation of amorphous phases was frequently observed when nonmetals (boron, silicon, and phosphorus) were burnt in high-pressure nitrogen atmosphere (�N2 = 300–800 MPa). Amorphous phases can be crystallized during heating or long exposure to room temperature. The chemical analysis of the amorphous products reveals that their composition corresponds to that of stoichiometric nitrides. However, in some cases, the unknown phases were obtained, for example, pink-brownish boron nitride of

5

6

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation

BN0.75 composition. After heating or extended exposure to room temperature, amorphous stoichiometric compounds, as mentioned before, are crystallized, retaining their chemical composition. When synthesizing boron nitride from boron anhydride (SHS with a reduction stage), Mamyan [29] also observed the formation of amorphous SHS products. Amosov [30] obtained amorphous products when synthesizing silicon nitride in azides presence. In our experiments, amorphous silicon nitride was produced when silicon powder was burnt in gaseous nitrogen in the presence of ammonium chloride. In these reactions, amorphous silicon diimide was identified as an intermediate product, which is known to be a network polymer. This intermediate compound is a source of SHS nitride with a unique particle structure in the form of fibers. Silicon nitride powders with different particle structures (agglomerates and column crystals) were obtained at various combustion terms. The mechanism of their formation in the SHS mode was thoroughly studied in [31–34]. Later on, Merzhanov [28] suggested the conception of structure formation in SHS processes. The starting point was the idea of complete destruction of the phase composition of the initial components during the chemical reaction and formation of a new “primary” microstructure. Two threshold combustion and structure formation mechanisms in SHS processes were defined in connection with characteristic times of combustion t c and structure formation t s , Figure 1.3. If t s /t c ≪ 1, the equilibrium mechanism of structure formation takes place. In this case, the SHS reaction proceeds by the mechanism of the reactive diffusion SHS – two-staged process: SHS = combustion + structure formation Equilibrium mechanism

τc > τr Reagents

Combustion

Nonequilibrium product in nonstable state (amorphous) Oversaturated solid solution melt Decomposition with new phase nuclei formation and particle growth (crystallization)

Polycrystalline products (crystallites) Recrystallization

Recrystallized product

Figure 1.3

Cooling

Final SHS product

Two threshold mechanisms of final product formation in SHS processes.

1.2

Peculiarities of Phase and Structure Formation of Metal and Nonmetal Nitrides

and yields the equilibrium product, which does not undergo any structural transformations. Structure formation of the final product influences the combustion front and chemical transformation velocity. The equilibrium mechanism taking its name from the conception by Khaikin–Aldushin–Merzhanov is preferable for slow-burning systems. If t s /t c ≫ 1, the nonequilibrium mechanism of structure formation takes place. During the reaction, the products appear to be in metastable state: oversaturated solid solutions, amorphous phases, and melts. After chemical reactions in these products, phase structural transformations occur far behind the combustion front. These processes do not affect the reaction and combustion front rates. We can change the ratio of the characteristic times of combustion product structure formation t s and cooling t rel and thereby regulate the completeness of structure formation and obtain either nonequilibrium products (t rel /t s ≪ 1) or equilibrium ones (trel /t s ≫ 1). At the nonequilibrium mechanism named after Borovinskaya, local equilibrium state of the substance is absent in the combustion wave, and it is preferable for fast-burning systems. Nowadays, the subject of structure formation during SHS reactions is being developed by means of novel methods used in structural macrokinetics: X-ray microanalysis, optical spectroscopy, Auger spectroscopy, electronic microscopes, and physical and mathematical modeling. Of great importance in investigation of structure formation in SHS is local dynamic X-ray spectrum analysis using diffractometer with a special detector [35]. Processing of the diffraction patterns allows plotting the kinetic curves of phase composition changes in the combustion wave and getting information on chemical and phase transformations. The mentioned method is used to study formation mechanism of specific products and intermediate phases from the initial green mixtures to final products in the systems of solid–solid, but there are also some works demonstrating the combustion product formation in the solid–gas systems. 1.2.1.2 Structural and Morphological Peculiarities of Silicon, Boron, and Aluminum Nitrides – Nanosized Powders

Many experimental results proving the formation of various nonequilibrium products and investigation of synthesis peculiarities of nitrides formed under nonequilibrium terms (high synthesis velocities and fast crystallization of combustion products) give the go-ahead to the SHS processes in such an actual field as nano-industry and throw a sidelight upon the mechanism of nanodispersed particle formation. The conception of SHS of nanomaterials can be presented using the scheme suggested by Merzhanov, Figure 1.4 [36]. According to the figure, nanoparticles are formed and then grown under the nonequilibrium terms of SHS processes. The size and amount of nanoparticles depend on the combustion temperature and on the cooling time as well. There are some nanoparticles in the samples, which are cooled naturally. It is necessary to quench them in order to increase the nanoparticle amount. The possibility of nanoproduct formation at combustion of various SHS systems (solid–solid, solid–gas, gas-phase SHS, etc.) was considered in the work by Sytchev and Merzhanov [37]. In our days, we have obtained a lot of powder SHS compounds with nanodispersed particles, and even

7

8

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation Product cooling

Temperature

U

Combustion Recrystallization End of chemical reactions

Figure 1.4

Nucleation and growth of new phase particles

Amorphous state of combustion product

SHS product structure formation in combustion wave.

tubular nanostructures as well as nanostructured materials. Figure 1.5 demonstrates some of SHS nanopowders synthesized by combustion. The experimental investigation on the mechanism and regularities of nanoparticle formation in refractory inorganic compounds and the methods of their separation show that the following approaches are the most efficient: transfer of the combustion process to the gas-phase mode using gasifying additives or evaporation of solid reagents, fast crystallization of combustion products from melts or gas phase, and chemical dispersion of agglomerated combustion products in special liquids. Specific terms of synthesis and separation of nanodispersed powders shown in Figure 1.5 are described in many original works. This chapter suggests discussing such an α-Si3N4 (a)

(b)

Fibers, thread-like crystals d/l ~1 :1000 (up to 90%)

Fine crystals, d/l ~1

AIN

Needle-like crystals

Figure 1.5

(d)

Whiskers, tapes

BN (f)

(e)

β-Si3N4 (c)

Rods – laminated microcomposites with nanosized laminas (100%)

MoSi2 (g)

Needle-like crystals, d – up to Spherical particles, d ≤ 100 nm (up to 50%) 50 mm, transparent plates, Ssp – up to 65 m2 g−1 (up to 100%)

(a–h) Nanosized SHS powders of various classes.

WC (h)

Ultradispersed with nano-component, d ≤ 300 nm (>70%)

1.2

Peculiarities of Phase and Structure Formation of Metal and Nonmetal Nitrides

important and interesting characteristic of nanomaterial particles as their crystal morphology. Active assimilation of up-to-date investigation methods of synthesized substances such as optical and electron microscopy, atomic spectrometric methods, dynamic X-ray method, and so on allows widening the conception of the material structure, shape, and particle size and even predicting the material properties. However, in most cases, the particle morphology of compounds synthesized by combustion is only considered as an object of some interest. This chapter suggests considering particle morphology as a “tool” for understanding structure formation mechanism at combustion processes. This approach is based on the fact that morphological characteristics can define practical application of a definite material [38]. Therefore, it is more important to study the morphology of a substance than to demonstrate it as an additional approach for understanding the processes occurring during the synthesis. It should be pointed out that the conventional synthesis methods (hydrothermal, gas-phase, melted, etc.) successfully use a morphological variety of some compounds to determine the fields of their application. When studying morphological peculiarities of nitrides and their connection with combustion terms, we used nonmetals nitrides: Si3 N4 , AlN, BN, SiAlONs, and AlONs, as they are characterized by a great variety of forms in comparison with metal nitrides. So-called condensation synthesis was used for obtaining nanosized particles [39]. This method is not well studied and is rather seldom used in SHS. It is based on evaporation of a solid reagent in nitrogen. Also, we applied gas-phase synthesis with gasifying additives and fast-product crystallization from flux melts. Separation of the product was carried out by means of “chemical dispersion.” Silicon Nitride On the basis of the obtained experimental results, we can point out the following important results. Formation of final holohedral crystals of Si3 N4 can occur with formation of primary so-called skeleton crystals as hollow rods (Figure 1.6a, b) [6, 40]. They are formed under the terms of “starvation” of the forming nitride crystal, for example, in the case of gaseous nitrogen lack, and prove some infiltration difficulties of the gas supply to the combustion front. These “skeleton” rods gather into spheres. It is typical for structures forming under nonequilibrium terms. When typical β-Si3 N4 rods were synthesized under equilibrium terms of nitriding without any infiltration difficulties, the conventional scheme of crystal formation by the vapor–liquid–crystal mechanism was observed [41]. It included evaporation of the solid reagent, interaction of its vapor with gaseous nitrogen, formation of a drop with the condensation reaction product dissolved in it (in our case it was silicon nitride), and further crystal growth. In the case of infiltration difficulties of nitrogen supply, the crystal framework grows fast only at the edges because of the maximum mass transfer to them. Due to the material lack, the “starving” crystal tries to grow the “skeleton” and absorbs practically the entire nitrogen, which is only able to grow the protruding parts of the crystal under the terms of the fast growth. As a result, internal cavities, holes, channels, or gutters are formed parallel to the edges in the crystal of hexagonal configuration. This system is nonequilibrium

9

10

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation (a)

(c)

(b)

Figure 1.6

β-Si3 N4 morphology: (a, b) “skeleton” crystals and (c) holohedral crystals.

and energetically unstable, and the most favorable thermodynamic configuration of its existence is a sphere. At mechanical destruction of the sphere, the product disintegrates to separate hollow hexagons (Figure 1.6b). If β-Si3 N4 rods are treated with alkali, we shall see that the crystal has a laminated structure with a nanosized thickness of the layers, though the crystal itself can be larger than 1 μm. The layers are nitride “flakes” joined to each other by thin layers of melted silicon, which is seen in X-ray patterns while the crystal is treated with alkali. The total amount of the rods with “skeleton” configuration depends on nitrogen pressure, sample density, and initial nitrogen particle size. The number of the “skeleton” crystals grows with a decrease in nitrogen pressure and an increase in the density of the initial compact samples. Probably, it is connected with the development of infiltration difficulties of gaseous nitrogen supply to the reaction zone. The output of silicon nitride with the morphology of “skeleton” crystals is also increased when the combustion product is hardened by fast drop of gaseous nitrogen pressure after the combustion front propagation or the time of the reactive mass cooling in nitrogen medium is shortened. A decrease in the “skeleton” crystal content in the reactive mass, which is observed in the experiments without nitrogen pressure drop immediately after the combustion front propagation or with an increase in the time of the product cooling in nitrogen medium, proves the ability of β-Si3 N4 “skeleton” crystals to grow up to standard holohedral rods, Figure 1.6c. It is typical for energetically unstable systems to change in the external terms. In this case, it can be after nitriding of the hot “skeleton” crystals connected with an increase in nitrogen exposure time and better nitrogen supply to incompletely developed products. Investigation of structure formation mechanism of SiAlONs (solid solutions of Si3 N4 –AlN–Al2 O3 ) under the SHS mode with cooling of the reactive mass (quenching of intermediate products) shows that not only individual compounds

1.2 (a)

Peculiarities of Phase and Structure Formation of Metal and Nonmetal Nitrides (b)

(c)

Figure 1.7 Structural transformations of β-SiAlON at thermal treatment in nitrogen, T ≥ 1000 ∘ C: (a) crystal internal part and (b, c) second phase on crystal surface.

such as silicon nitride but also more complex systems, for example, solid solutions, can be formed through the stage of skeleton crystal formation. SiAlONs appeared to be interesting compounds for studying structural transformations, chemical and phase composition changes, and variations of properties due to external effect, for example, high-temperature annealing. At T > 1500 ∘ C, β-SiAlON of Si4 Al2 O2 N6 composition transforms to Si3 Al3 O3 N5 with disengagement of some amount of aluminum oxide (Al2 O3 ) or AlON (Al6 O3 N4 ). It can be seen in Figure 1.7a–c that the internal part of the crystal disintegrates into the finest nanofiber bunches with very little spherical formations, which can be considered as nanoparticles. The nanoparticles gather into rhomboid formations on the edges of big crystals. Probably, Si3 Al3 O3 N5 obtained after annealing is supersaturated α-SiAlON containing a large amount of impurity phases of aluminum nitride polytypoids (12H and 21R) and oxide compounds indissoluble in SiAlON crystal lattice. A rather interesting morphological pattern can be observed in the case of silicon nitride synthesis with crystallization of its particles from the melts of fusing agents, which can be presented by various alkali metal salts. The crystallization mechanism is determined by the existence of the melt, the character of the fusing agent of a simple or complicated composition, including eutectics with a hard reagent, and its behavior during the combustion process. Figure 1.8a demonstrates that initial nitride crystals in NaCl-containing melt have a shape of a sphere. They can be concerned as “antiskeleton” crystals. It can be seen that silicon nitride crystals are formed in the melt in which a great number of crystallization centers appear. Disintegration of the crystal-sphere results in the development of rather complicated branching dendrites (Figure 1.8b,c). According to the well-known mechanism of the dendrite growth of a branching crystal [42], long branches are the first to grow, then some others appear on them until they come into contact with each other and fill the interaxial space. Then the dendrite turns into the crystal (in our case it is β-Si3 N4 ) with irregular external faces (Figure 1.8d). Therefore, the dendrite appears to be an intermediate crystal form, which is changed to the acicular one (spherulite). It confirms fast growth of nitride crystals at high velocities of SHS and fast crystallization of combustion products. It is typical for nonequilibrium processes. Hardening of the reactive mass by fast drop of gaseous nitrogen from the reactor encourages the terms of accelerated crystallization of

11

12

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation (a)

(c)

(d) (b)

Figure 1.8 Si3 N4 crystal morphology in the presence of fusing agents: (a) spheres; (b, c) dendrite growth; and (d) imperfect Si3 N4 crystal.

silicon nitride and obtaining of energetically unstable products. This approach results not only in cooling but also sometimes in overcooling of synthesis products to low temperatures with formation of amorphous products. The morphological pattern of the products of silicon vapor combustion in gaseous nitrogen at condensation synthesis with “skeleton” crystal formation as well as dendrite growth of silicon nitride crystals in melted metal salts proves the existence of the nonequilibrium mechanism of structure formation in the case of SHS. The mechanism appears to be the basis of the conception of nanodispersed particle formation under the combustion mode [28]. We have discussed the formation of β-Si3 N4 crystals. However, silicon nitride has another modification – α-Si3 N4 , which exists till 1500 ∘ C. When silicon nitride is obtained by condensation synthesis, in many experiments α-modification is formed along with β-Si3 N4 . At low nitrogen pressures (≤1.0 MPa), up to 50 mass% of α-Si3 N4 are formed. The morphology of α-Si3 N4 is remarkable for its variety: cotton-like agglomerates, cobweb-like nanofibers, nanotapes, and “nanoflowers” (Figure 1.9a–e). Probably, at the terms of condensation synthesis, α-silicon nitride is formed by the gas-phase mechanism when gas-transport reactions occur with participation of silicon monoxide, which is practically always present in the reactive mass [41]. Nowadays, there are some variations of α-Si3 N4 synthesis under the SHS mode. A production technology, which allows obtaining nitride with 97 mass% of α-phase, has been developed [34]. The main parameters influencing the formation of α-Si3 N4 at combustion of silicon in gaseous nitrogen are the synthesis

1.2

Peculiarities of Phase and Structure Formation of Metal and Nonmetal Nitrides

(b)

(a)

(d)

(c)

(e)

= 7.765 μm

Figure 1.9 Morphology of α- and β-Si3 N4 nanoparticles obtained at “condensation synthesis”: (a) α-Si3 N4 threads, (b) α-Si3 N4 “flowers”, (c) β-Si3 N4 tapes, (d) β-Si3 N4 whiskers, and (e) β-Si3 N4 tubes.

temperature and presence of gasifying additives decomposing at combustion and providing propagation of gas-transport reactions or formation of intermediate products with silicon, which are precursors in nitride formation. The gasifying additives can be salts (mainly, ammonium halides), organic compounds, polymers, and so on. The morphology of α-Si3 N4 synthesized with addition of ammonium chlorides and fluorides is presented by two types of silicon nitride particles. At comparatively high contents of halides (up to 15 mass%), thin long crystals – fibers of 100–300 nm in diameter and up to 10 μm in length – are formed, Figure 1.9a. Special studies show that these crystals are formed due to the complicated chemical transformations of ammonium chloride to diimide Si(NH)2 at the reaction with silicon. The diimide is an intermediate compound at nitride synthesis and has a “polymer” structure, which is “memorized” at diimide transformation to nitride. By this time, an SHS technology of filamentary α-Si3 N4 has been developed with the product output of 100%. At lower temperatures ( 0.6), self-oscillating combustion takes place. Appearance of the second combustion front at �N2 > 1.0 MPa is of great interest. It follows the main front as a bright and wide glow downwards with the same velocity. The sample quenching in argon immediately after the first front propagation shows that at any pressures nitrogen content in the sample

(a)

(b)

(c) (d) Figure 1.15 Photographic record of hafnium combustion in nitrogen at �N = 6.0 MPa: (a) 2 stationary combustion, (b) second front, (c) self-oscillating combustion, and (d) record of spin combustion at �N = 6.0 MPa and PAr = 0.1 MPa. 2

1.3

Dependence of SHS Nitride Composition and Structure on Infiltration Combustion Mode

center is very low ( 2973 K results in formation of faceted AlN particles. But as the combustion temperature is close to AlN dissociation point, AlN is also formed from the gas phase and the condensed particles take the form of needle-like crystals. As it was mentioned in part 2, boron nitride can be obtained as melted particles and laminar single crystals. The main problems at melted boron nitride formation are those of the temperatures of boron nitride melting and dissociation. It is known that the melting point of boron nitride at nitrogen pressure is ∼3000 ∘ C, and the temperature gradually grows up to 3500 ∘ C at the pressure of about 8000 MPa. At a further increase, boron nitride dissociates to amorphous boron and nitrogen. Obviously, melting occurs simultaneously with the nitriding reaction, and the melting velocity is controlled by that of the chemical reaction. The temperature is constant and equal to the melting point (T m ). Probably, the chemical reaction takes place before the formation of the composition which can exist as a melt at these terms (�N2 and T c ), that is, to the composition of BNx , where x = f (�N2 and T c ). The melt consisting of liquid boron with nitrogen diluted in it in the amount “admitted” by the external pressure and combustion temperature forms the melting surface. The melt is homogeneous in its composition. During the melt cooling, some transformations can occur: quenching of nonstoichiometric composition as an amorphous compound must have taken place in work [8] when BN0.75 was obtained, crystallization by the scheme (BNx )(l) → BN(s) + B(l), when x < 1, occurs in most cases, or crystallization by the scheme (BN)(l) → (BNx )(s), if x = 1. Considering the results of these experiments, we can assume that boron combustion at high nitrogen pressures corresponds to the elemental combustion model of the second type, namely, the model of high-temperature melting. The experiments with other elements (metals) prove that at temperatures higher than the melting point of the solid reagent but lower than the nitride dissociation temperature and at high gaseous nitrogen pressures, nitriding can occur by the liquid-phase mechanism with formation of the solid reagent melt, development of saturated nitrogen solution in liquid nonmetal or metal with the composition “admitted” by the external nitrogen pressure and the combustion temperature of the composition, and the following cooling with the yield of melted nitride crystals. The weakest point in the discussion on the combustion model of hightemperature melting is the absence of the data of nitrogen dilution in boron, silicon, and metals in a liquid state, that is, the absence of equilibrium phase diagrams at P = 0.1 MPa as well as at P > 0.1 MPa. But if the combustion model is true, it is possible to plot the liquidus line for Al–N, B–N, and Si–N by plotting the dependence with the quenched products: T com = f (mquen ), where (mquen ) is the composition due to which T c develops in the combustion front and T c = f (�N2 ).

29

30

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation

1.3.1.2 Role of Nitrogen Admission to the Reaction Zone

The infiltration mechanism of gaseous nitrogen admission to the reaction zone is very important for organization of the entire process and combustion product formation. Nitrogen admission to the reaction zone is another significant factor for attaining nitriding completeness and development of the terms providing formation of nitride particle phases and structures. From the viewpoint of gaseous reagent supply, we should remember that there can be some diffusion difficulties of nitrogen admission to the reaction zone. In this case, the real combustion temperature can be much lower than its thermodynamic equivalent and the combustion process can result in incomplete substance transformation to nitride. From the viewpoint of solid reagent behavior at combustion, we can expect its melting or melting and decomposition of the intermediate products, insufficient initial permeability of the samples under study, low values of nitrogen pressure, and so on. They can lead to infiltration difficulties for gaseous nitrogen to admit. The diffusion mode of gas supply to the combustion front can be established by dilution of nitrogen with other gas, for example, argon [8, 12], which can gather close to the reaction zone. Figure 1.20a shows the dependence of zirconium combustion velocity on the amount of argon in the mixture of nitrogen–argon at constant nitrogen pressure. Transformation of infiltration combustion to the diffusion mode results in a sharp slowdown of the combustion process. Even small additions of argon decrease zirconium combustion velocity ∼5 times, and addition of ∼0.26 parts results in its complete damping. The results obtained at filtration combustion transformation to the diffusion mode at nitrogen dilution with argon were very important for further development of the new direction in combustion theory and practice, the so-called “spin” combustion, which had not been known before [2, 12, 63]. Since its discovery, the nature and varieties of spin combustion have been of great interest for scientists. But as to the influence of spin combustion on product composition, this direction has not been studied thoroughly. In general, the degree U (cm s−1)

2.0 1.5 1.0 0.5

0 (a)

3

2

1

10

20

30

40 Pin (atm)

(b)

Figure 1.20 Spin combustion: (a) of zirconium and (b) dependence of Zr combustion velocity on argon content in nitrogen at spin combustion.

1.3

Dependence of SHS Nitride Composition and Structure on Infiltration Combustion Mode

of the initial solid conversion to nitride depends on the character of spin travel, that is, the hot spot where the reaction is localized and which moves in tangential and longitudinal directions round the sample spirally, Figure 1.20b. The spin wave can move either on the sample surface, and its middle part remains unreacted, or in the whole volume. These phenomena were studied experimentally in [12]. Paper [58] describes the replacement of combustion modes from stationary to spin at hafnium combustion at nitrogen pressure ranging from 2.0 to 150 85 >150

100 60 30 25 25 5

41

42

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation

Table 1.3 Application of ceramics obtained by direct synthesis. Ceramics

Items and applications

BN

Molds and crucibles for teeming metal alloys based on iron, cobalt, nickel, chromium, as well as titanium alloys Installations for centrifugal casting “Formaks,” “Degudron,” and “Minimax” Two to three times increase in operation life Stop valves, crucibles for melting complex alloyed steels, and hardeners Teeming machine “Sirius,” 50 processes instead of 2–3 (kersil) Elements of technological installations for growing semiconducting materials, thermal cycling, and so on. Dielectric characteristics are constant; stability is higher than that of quartz, alundum, and graphite Protection of parts of laser installations; high thermal stability (CO2 3 kW) Bushings for air-plasma cutting (instead of Al2 O3 ) Cases for measuring temperature of aluminum and other metal melts, T melt = 790 ∘ C, number of cycles – 72 (without destruction) Nozzles for sand- and shot-blasting machines

BN–TiB2 BN–SiO2 BNx Cy BN–Al2 O3 SiAlON–SiC–BN

Si3 N4 –SiC–TiN SiAlON–TiN TiB2 –B4 C

X-ray-amorphous silicon dioxide. We can assume that at combustion of boron with SiO2 in nitrogen, BN crystallization occurs with participation of the liquid phase containing boron and nitrogen as well as silicon dioxide. They interfere with SiO2 crystallization and encourage growth of hexagonal boron nitride crystals. Thorough investigation of the mechanism of BN–SiO2 synthesis allowed us to follow the ways of the combustion product formation and their physical state. Boron solid-phase nitriding was established to start in the warming-up zone. In connection with it, the combustion temperature attains its maximum and exceeds boron melting point [45]. Silicon dioxide also melts in the warming-up zone. In the combustion zone (maximum heat release), maximum growth of the substance volume and a decrease in the sample porosity are observed. In the after-burning zone (volume after nitriding), the temperature decreases. If the residual porosity of the sample remains the same at the level of 25–30%, the conversion degree of boron to nitride increases to the final value (� = 1), herewith dioxide melt remains. In the crystallization zone, the temperature goes on decreasing, the melt grows hard, and crystal SiO2 turns to amorphous. At low values of the sample porosity, some side reactions can occur in the reactive mass after the combustion stage with formation of impurity phases: B2 O3 , B6 O, Si, and so on. 1.6.3 Nitride Ceramics Based on AlN

The items based on AlN and its composite materials are also widely used due to its physicochemical characteristics. So, the development of direct synthesis of these materials is an actual task. As it was mentioned above, the main restriction

1.6 Direct Production of Materials and Items Based on Nitride Ceramics by SHS Gasostating

in the combustion synthesis of nitride-based materials is the insufficient initial concentration of nitrogen in the sample pores. In order to avoid this disadvantage and nitride dissociation, the process is carried out in SHS gasostats at nitrogen pressure of up to 300 MPa. At aluminum nitride synthesis, the initial component was aluminum or its mixtures with aluminum nitride. Some refractory compounds (borides, carbides, oxides, etc.) were introduced into the green mixture to obtain composite materials [87]. A photo of the fracture structure of aluminum nitride obtained from Al + AlN mixture is shown in Figure 1.27; the porosity is 20%. The structure was formed as a result of melting, and the grains of a newly formed aluminum nitride and those of AlN diluent do not differ. Within the combustion process of the mixture of aluminum and titanium diboride, AlN–TiB2 ceramics with various ratios of the components was obtained (Table 1.2). The material density is ≥92%, Rockwell hardness – 70–80, and microhardness of TiB2 grains – 25 750–31 800 MPa and that of AlN matrix – 10 720–12 250 MPa. The material is homogeneous; diboride grains are distributed in the matrix of molten aluminum nitride, and distinct TiB2 and AlN boundaries are seen. Several factors – mechanism of liquid aluminum interaction with nitrogen, the second component wettability with the melt, a possibility of chemical interaction between the second component and the melt, the second component capability of melting and the following failure of the components interface – must be taken into account for comprehension of the mechanism of the processes occurring at SHS of aluminum nitride and composites thereof in gasostats. X-ray microanalysis shows that the melt formed in the SHS zone does not interact with titanium diboride; it corresponds to the literature data on the tendency of the wetting angle of titanium diboride with aluminum to 0 and absence of their interaction at the temperatures higher than 1573 K [88]. It is also encouraged by the high melting point of diboride (T m = 3223 K) in comparison with the combustion temperature (T c = 2773 K). Therefore, titanium diboride wettability and its possible melting do not restrain (a)

(b)

200 nm

Figure 1.27 Microstructure of molten AlN: (a) fracture surface and (b) microcrystals on matrix surface.

43

44

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation

the melt spreading. The existence of molten aluminum nitride matrix and characteristic plateau on heat patterns leads us to confirm that the combustion products are in molten form in the reaction zone. Thorough investigation of the influence of temperature, porosity, shrinkage, and so on, on the product state during SHS proves that the melt of aluminum nitride is formed in the combustion zone. The experimental data allow us to describe the combustion mechanism of aluminum and its mixtures with refractory compounds as the elemental model of the II type – the model of high-temperature melting. But the concept of nonequilibrium mechanism of phase formation should be used too, as it was mentioned above. The experimental studies of production of boron and aluminum nitride compositions with various refractory compounds by SHS gas-stating prove that the main factors determining the composition of almost all the systems under study and the degree of nitriding are existence of nitriding volume effect, volume shrinkage extent, retaining (or loss) of porosity, and magnitude of the equilibrium wetting angle. The synthesis peculiarities inherent for each specific system are connected with physicochemical properties of refractory compounds making the composition with nitrides of boron, aluminum, and other elements.

1.7 Conclusion

While putting the finishing touch in this chapter, we would like to underline that it was nitrides and composites thereof that played a significant role in the development of SHS R&D and contributed greatly to a better comprehension of the processes occurring at combustion synthesis and regularities of product phase and structure formation. Investigation of nitride formation mechanism and peculiarities at combustion encouraged the origin of such a scientific field as structural macrokinetics. Theoretical studies carried out by the scientists from ISMAN: Merzhanov, Aldushin, Shkadinsky, Seplyarsky, Grachev, Ivleva, and their colleagues from different Russian institutes and abroad appeared to be remarkable in the development of this field. An important contribution was made by the researchers headed by Maksimov from Tomsk, who elaborated theoretical and practical backgrounds of metal and alloy nitriding at combustion, and their colleagues headed by Amosov from Samara, who developed the azide variation of SHS. References 1. Merzhanov, A. and Borovinskaya, I.

(1972) Dokl. Akad. Nauk SSSR, 204 (2), 366–369. 2. Merzhanov, A. and Borovinskaya, I. (1975) Combust. Sci. Technol., 10 (5–6), 195–200.

3. Merzhanov, A. (1990) in Com-

bustion and Plasma Synthesis of High-Temperature Materials (eds Z.A. Munir and J.B. Holt), VCH Publisher, New York, pp. 1–53.

References 4. Borovinskaya, I. (1992) Pure Appl. 5.

6.

7. 8.

9.

10. 11. 12.

13. 14.

15.

16.

17. 18.

19.

20.

Chem., 64 (7), 919–940. Munir, Z. and Anselmi-Tamburini, U. (1989) Mater. Sci. Reports, 3 (7–8), 277–365. (a) Merzhanov, A., Borovinskaya, I., Int. J. SHS, 2008, 17(4), 242-265; (b) Merzhanov, A., Borovinskaya, I., Forum WAC: Ceramic Materials in Energy Systems for Sustainable Development, Chianciano Terme, 2008. Hagg, G. (1931) Z. Phys. Chem., 12, 33. Borovinskaya, I. (1972) Self-propagating High-temperature Synthesis of Nitrides. PhD Thesis, Chernogolovka. Borovinskaya, I. and Loryan, V. (1976) Dokl. Akad. Nauk SSSR, 231 (4), 911–914. Borovinskaya, I. and Loryan, V. (1978) Poroshkovaya Metall., 11 (191), 42–45. Merzhanov, A. and Rumanov, E. (1977) Izvestiya AN SSSR, Met., 3, 188–193. Merzhanov, A., Filonenko, A., and Borovinskaya, I. (1973) Dokl. AN SSSR, 208 (4), 892–894. Munir, Z. and Holt, J. (1987) J. Mater. Sci., 22, 710–714. Borovinskaya, I. (1977) Gorenie i Vzryv, IV Symposium on Combustion, Nauka, Moscow, pp. 138–148. Borovinskaya, I., Merzhanov, A., Pityulin, A., and Shehtman, V. (1975) in Combustion Processes in Chemical Technology and Metallurgy (ed. A.G. Merzhanov), USSR Academy of Science, Chernogolovka, p. 133. Borovinskaya, I., Merzhanov, A., and Pityulin, A. (1975) Combustion Processes in Chemical Technology and Metallurgy, Publishing department of The Branch of Institute for Chemical Physics, Russian Academy of Sciences, Chernogolovka, pp. 113–118. Vidavsky, Y. and Konovalov, S. (1983) J. Inorg. Chem., 2412–2414. Itin, V., Bratchykov, A., Doronin, V., and Pribitkov, G. (1981) Izv. Vuzov, Fiz., 12, 75–78. Boldyrev, V., Aleksandrov, V., and Korchagin, M.A. (1981) Dokl. Akad. Nauk SSSR, 5, 1127–1129. Frahm, R., Wong, J., Holt, J., Larson, E., Rupp, B., and Waide, A. (1992) AIP Conf. Proc., 258, 615–620.

21. Knyazik, V., Merzhanov, A., Solomonov,

22.

23.

24.

25.

26.

27.

28.

29. 30.

31. 32.

33.

34.

V., and Shteinberg, A. (1985) Fiz. Goreniya Vzryva, 3, 69. Loryan, V. (1995) Self-propagating High-temperature Synthesis of Nitride Ceramics at High Pressure of Nitrogen. D.Sci. Thesis, Chernogolovka. Borovinskaya, I., Veshnyakova, G., and Loryan, V. (1991) Problems of Structural Macrokinetics, Publishing department of Institute of Structural Macrokinetics, Russian Academy of Sciences, Chernogolovka, pp. 5–21. Hirao, K., Miyamoto, Y., and Koizumi, M. (1987) J. Soc. Mater. Sci., 36 (400), 12. Munir, Z. and Anselmi-Tamburini, U. (1989) Mater Sci. Rep., 3 (7–8), 277–365. Khidirov, I., Karimov, I., Em, V., Loryan, V., and Borovinskaya, I. (1980) Dokl. Akad. Nauk SSSR, 195, 29–31. Khidirov, I., Karimov, I., Em, V.T., and Loryan, T. (1986) Izv. AN SSSR, 22 (10), 1675–1678. (a) Merzhanov, A. (2000) Solid Flame Combustion, Publishing department of Institute of Structural Macrokinetics, Russian Academy of Sciences, Chernogolovka, p. 238; (b) Merzhanov, A. and Mukasyan, A. (2007) Solid Flame Combustion, 2nd edn, Torus Press, Moscow, p. 336. Mamyan, S. and Vershinnikov, V. (1992) Int. J. SHS, 1 (3), 392–400. (a) Amosov, A., Borovinskaya, I., and Merzhanov, A. (2007) Powder Technology of Self-propagating High-temperature Synthesis of Materials, Mashinostroenie, Moscow, p. 567; (b) Amosov, A., Bichurov, N., and Bolshova, V. (1992) Int. J. SHS, 1 (2), 239–245. Mukasyan, A. and Borovinskaya, I. (1992) Int. J. SHS, 1 (1), 55–63. Mukasyan, A., Martynenko, V., Merzhanov, A., Dorovinskaya, I., and Blinov, M. (1986) Fiz. Goreniya Vzryva, 5, 43–49. Cano, J., Borovinskaya, I., Rodriguez, A., and Grachev, V. (2002) J. Am. Ceram. Soc., 85 (9), 2209–2211. Zakorzhevsky, V. (2004) Self-propagating High-temperature Synthesis of Silicon

45

46

1 Combustion Synthesis of Nitrides for Development of Ceramic Materials of New Generation

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47. 48.

Nitride, Aluminum Nitride and Composite Powders Based on these Nitrides. PhD Thesis, Chernogolovka. Merzhanov, A., Borovinskaya, I., Khomenko, I., Mukasyan, A., Ponomarev, V., Rogachev, A., and Shkiro, V. (1995) Ann. Chim. Fr., 20, 123–138. Merzhanov, A., Borovinskaya, I., and Loryn, V. (2011) Abstracts. XIX Mendeleev Forum on General and Applied Chemistry, Publishing department of Volgograd State Technical University, Volgograd, p. 76, 536, 600, 704. (a) Sytchev, A. and Merzhanov, A. (2004) Usp. Khim., 73 (2), 157; (b) Sytchev, A. and Merzhanov, A. (2004) Russ. Chem. Rev., 73 (2), 147–159. Suslik, K. and Price, G. (2010) in Nanotechnology (ed. Y.D. Tretyakov), Fizmatlit, Moscow, p. 368. Andrievsky, R. and Ragulya, A. (2005) Nanostructural Materials, Academic Press, Moscow, p. 193. Borovinskaya, I. (2012) Abstracts. XI International Symposium on SHS, 2011, Attica, Greece, p. 75. Mukasyan, A., Stepanov, B., Galchenko, Y., and Borovinskaya, I. (1990) Fiz. Goreniya Vzryva, 1, 104–108. Melikhov, I. (2006) Physico-chemical Evolution of Solids, Binom, Moscow, p. 309. Borovinskaya, I., Ignat’eva, T., Vershinnikov, V., Khurtina, G., and Sachkova, N. (2003) Neorg. Mater., 39 (6), 698–704. Loryan, V., Borovinskaya, I., and Merzhanov, A. (2009) Int. J. SHS, 18, 154–156. Loryan, V., Borovinskaya, I., and Merzhanov, A. (2011) Int. J. SHS, 20 (3), 153–155. Pirtskhalaishvili, R., Bayramishvili, I., Samkurashvili, T., Lomidze, G., Loladze, S., and Gudushauri, N. (1974) Boron, Production, Structure, Properties, Science, Moscow, pp. 23–26. Borovinskaya, I. (1974) Archiwum Procesow Spalania, 5, 145–161. Borovinskaya, I., Merzhanov, A., Dubovitsky, F., and Enman, V. (1970)

49.

50. 51.

52.

53. 54.

55.

56.

57.

58. 59. 60.

61. 62. 63.

64. 65.

Certificate USSR, No. 1393281/23-26, 1970. Zakorzhevsky, V., Borovinskaya, I., and Sachkova, N. (2002) Neorg. Mater., 38 (11), 109–115. Zakorzhevsky, V. and Borovinskaya, I. (2011) Int. J. SHS, 20 (3), 154–158. Merzhanov, A., Borovinskaya, I., and Volodin, Y. (1972) Dokl. AN SSSR, 206 (4), 905–908. Aldushin, A., Merzhanov, A., and Khaikin, B. (1974) Dokl. AN SSSR, 315 (3), 612–615. Grachev, V. and Borovinskaya, I. (2006) Adv. Sci. Technol., 45, 1011–1017. Borovinskaya, I., Merzhanov, A., Mukasyan, A., Rogachev, A., Khina, B., and Khusid, B. (1992) Dokl. AN SSSR, Phys.Chem., 322 (5), 912–917. Grachev, V., Ivleva, T., Borovinskaya, I., and Merzhanov, A. (1996) Dokl. AN SSSR, Chem., 346 (5), 626–629. Borovinskaya, I., Merzhanov, A., Butakov, A., Ryabinkin, A., and Shehtman, V. Method of producing Tantalum Nitride. (1970) Patent SU 264365, Bull. invent., 1970, 9. Molodovskaya, E., Petrunin, V., Karimov, I., Kalanov, M., Khaidarov, T., Borovinskaya, I., Pityulin, A., and Merzhanov, A. (1975) Fiz. Met. Metalloved. Akad. Nauk SSSR, 40, 202–204. Borovinskaya, I. and Pityulin, A. (1978) Fiz. Goreniya Vzryva, 1, 137–140. Merzhanov, A. (1995) Int. J. SHS, 4 (4), 323–350. Merzhanov, A., Rogachev, A., Mukasyan, A., and Khusid, B. (1990) Fiz. Goreniya Vzryva, 1, 104–114. Merzhanov, A. and Rogachev, A. (1992) Pure Appl. Chem., 64 (7), 941–953. Andrievsky, R. and Lyutikov, R. (1996) Zh. Fiz. Khim., 70 (3), 567–569. Ivleva, T., Merzhanov, A., and Shkadinsky, G. (1980) Fiz. Goreniya Vzryva, 16 (2), 3–10. Zakorzhevsky, V. and Borovinskaya, I. (2002) Inorg. Mater., 38 (11), 1140. Martirosyan, H., Dolukhanyan, S., Mkrtchyan, G., Borovinskaya, I., and Merzhanov, A. (1977) Poroshkovaya Metall., 7.

References 66. Osipova, I., Borovinskaya, I., and

67. 68. 69.

70.

71. 72.

73.

74.

75.

76.

77.

78.

Yaroshenko, V. (1996) Poroshkovaya Metall., 1/2, 98–102. Sharivker, S. and Borovinskaya, I. (1995) Mashinostroitel, 9, 20–23. Sharivker, S. (1995) Ann. Chem. Phys., 20, 157–167. Sharivker, S., Mamyan, S., Vlasov, V., and Mukasyan, A. (1994) Poroshkovaya Metall., 9/10, 107–111. Hizao, K., Miyamoto, Y., and Koizumi, M. (1987) J. Ceram. Soc. Jpn., 95 (10), 955–960. Pampuch, R. (1997) Int. J. SHS, 6 (2), 187–201. Wang, L., Sigmund, W., Roy, S., and Aldinger, F. (1999) J. Mater. Res., 14 (12), 1456–1459. Zakorzhevsky, V., Borovinskaya, I., Chevykalova, L., and Kelina, I. (2007) Poroshkovaya Metall., 1/2, 10–14. Loryan, V., Borovinskaya, I., Smirnov, K., and Titov, S. (2012) Int. J. SHS, 21 (1), 41–44. Levashov, E., Filonov, M., and Borovinskaya, I. (1995) J. Mater. Synth. Process., 3 (2), 111–114. Borovinskaya, I., Loryan, V., Grigoryan, E., Salnikova, E., and Pershikova, N. (1992) Int. J. SHS, 1 (1), 131–134. Borshch, V., Zhuk, S., Vakin, N., Smirnov, K., Borovinskaya, I., and Merzhanov, A. (2008) Dokl. AN SSSR, 420 (4), 496–499. Prokudina, V., Shestakova, T., Borovinskaya, I., Kuznetsova, I., Gracheva, N., and Nedelko, E. (1981)

79.

80.

81.

82.

83. 84. 85.

86. 87.

88.

Problems of Technological Combustion, vol. 2, Publishing department of The Branch of Institute for Chemical Physics, Russian Academy of Sciences, Chernogolovka, pp. 5–8. Zakorzhevsky, V., Sharivker, S., Borovinskaya, I., Ignatieva, T., and Sachkova, N. (1999) Int. J. SHS, 8 (2), 165–176. Borovinskaya, I., Loryan, E., Kudakhtin, V., Losikov, A., Merzhanov, A., Ratnikov, V., and Shekk, G. (1993) Installation for producing of compact items from nitride ceramics in the regime of combustion. Patent RU 1780241, Bull. Izobretenii of 27.09.1995. Ratnikov, V., Borovinskaya, I., and Prokudina, V. (2013) Izvestiya Vuzov, Poroshkovaya Metall., 1, 34–41. Borovinskaya, I., Loryan, V., Blinov, M., and Zozulya, V. (1991) Ogneupory, 9, 15–17. Borovinskaya, I. and Smirnov, K. (1998) Nauka Proizvodstvu, 8 (10), 70–74. Borovinskaya, I. and Smirnov, K. (2001) Key Eng. Mater., 217, 159–164. Borovinskaya, I. (2007) Presentation at XVIII Mendeleev Forum on General and Applied Chemistry, Moscow, 2007. Shulga, Y. and Loryan, V. (1994) Zhurn. Neorg. Khim., 39 (7), 1096–1099. Loryan, V. and Borovinskaya, I. (2003) Combust. Explos. Shock Waves, 39 (5), 525–533. Yasinskaya, G. (1966) Poroshkovaya Metall., 7, 53–56.

47

49

2 Combustion Synthesis of Boron Nitride Ceramics: Fundamentals and Applications Alexander S. Mukasyan

2.1 Introduction 2.1.1 Background: Brief Historical Overview

Boron nitride (BN) is a chemical compound consisting of equal numbers of boron and nitrogen atoms. It is rarely found in nature, only in the form of dispersed micrometer-sized inclusions of qingsongite (c-BN), named after Prof. Qingsong Fang (1939–2010) from the Institute of Geology, the Chinese Academy of Geological Sciences, in chromium-rich rocks in Tibet. Thus, BN is essentially a synthetic material, which although discovered in the early nineteenth century was not developed as a commercial material until the latter half of the twentieth century. Indeed, this compound was prepared for the first time by an English chemist Balmain [1–3]. He heated mixture of boric acid and melon or molten boric acid and potassium cyanide, but the obtained new compound was unstable. Only in the 1950s, Carborundum [4] and Union Carbide [5] companies managed to prepare boron nitride powder on an industrial scale and fabricated shaped parts of boron nitride for commercial applications. 2.1.2 Key Properties and Markets Values

Boron nitride possesses a similarity in electronic structure to the most versatile of elements, carbon, sharing the same number of electrons between adjacent atoms. The hexagonal form (h-BN, hexagonal boron nitride) corresponding to graphite is the most stable and softest among BN polymorphs. The cubic (sphalerite structure) phase analogous to diamond is called c-BN. Its hardness is slightly less than that for diamond, but its thermal and chemical stabilities are superior. The rare wurtzite w-BN modification is similar to lonsdaleite and may even be harder than the cubic form. Nanotubes of BN [6] have a structure similar to that of carbon nanotubes (CNTs), but the properties are Nitride Ceramics: Combustion Synthesis, Properties, and Applications, First Edition. Edited by Alexander A. Gromov and Liudmila N. Chukhlomina. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

50

2 Combustion Synthesis of Boron Nitride Ceramics

different: whereas CNTs can be metallic or semiconducting depending on the rolling direction and radius, a BN nanotube is an electrical insulator with a wide band gap of ∼5.5 eV, which is almost independent of tube chirality. BN nanotubes are more thermally and chemically stable than CNTs, which favors them for some applications. In this chapter, only h-BN (or α-BN or g-BN: graphitic boron nitride) modification is considered. Thus, let us briefly overview the key properties of the hexagonal form of boron nitride, which in contrast to graphite shows a white color; thus, it is often called “white graphite.” It is worth noting that different bonding – strong covalent within the basal planes (planes where boron and nitrogen atoms are covalently bonded) and weak between them – causes high anisotropy of most properties of h-BN. For example, the hardness and electrical and thermal conductivities are much higher within the planes than perpendicular to them. Other anisotropy is related to the orientation toward the pressing direction (|| or ⊥). The following boron nitride properties make it unique for variety of applications [7, 8]:

• • • • • • • • • • •

high thermal conductivity (30–100 W m−1 K−1 ); relatively low thermal expansion (0.5–12 × 10−6 /∘ C); high volume electrical resistivity (>1014 ohm cm); high dielectric strength (60–100 ac-kV mm−1 ); low dielectric constant (8–10 GHz); low loss tangent; microwave transparency; nontoxicity; easy machinability; nonabrasive and lubricious; chemical inertness (stable in inert and reducing atmospheres up to 2000 ∘ C, and in oxidizing atmospheres to 850 ∘ C); hence, the maximum application temperature is higher than that of Si3 N4 , A12 O3 , or SiC. • good thermal shock resistance; the temperature resistance of BN can be compared to MgO, ZrO2 , or CaO, but BN shows higher thermal shock resistance than these oxides. • non-wetting by most metallic (like Al, Mg, Zn, Pb, and Cu) and nonmetallic (Si, B, glass, cryolite, and halides) melts. Detailed information about the chemical and physical properties of h-BN is given in [7, 8, 9, 10]. No statistical figures are available on the production levels of boron nitride, but a reliable estimate was about 220 metric tons in 1993 for the western world, one-third each of which was produced in the United States, Europe, and Japan [11]. The 1999 production for the western world was about 300–350 metric tons [12]. It is stated that production figures should be similar for 2000 with a slight increase possible. The principal producers include Advanced Ceramics Corp. (formerly Union Carbide/Praxair Advanced Ceramics) and Carborundum Co. in the United States; Boride Ceramics & Composites Ltd. in the United Kingdom;

2.1 Introduction

Elektroschmelzwerk Kempten GmbH and H. C. Starck GmbH & Co. KG in Germany; and Denki Kagaku Kogyo, Kawasaki Steel Corp., Shin-Etsu Chemical Co. Ltd., and Showa Denko KK in Japan. Prices for h-BN range from −$75 to 120/kg for standard qualities depending on purity of BN, and may reach prices up to $200–400/kg for high-purity and tailor-made grades [10]. Although a large demand for BN exists at present, no large price increases are to be expected in the near future. 2.1.3 Applications of h-BN

The following properties – applications relations are typically outlined [7, 8]:

• High heat resistivity, stability against oxygen, extremely low friction coefficient. – Thus, h-BN powders are used as lubricants. Compared to graphite, the h-BN can be utilized in an oxidizing atmosphere up to 900 ∘ C, as well as at extremely low temperatures, for example, in space because no water inclusions between the atomic sheet layers are present. It is used for reinforcing ceramics and alloys to reduce wear and friction and thus obtain self-lubricating parts and as bearing materials for high-temperature applications and as sliding contact materials made of alloys or porous ceramics filled with Cu, Ag, and Pb. BN powder serves as an additive in paints, and it acts as a lubricating agent for casting and forming processes. BN powder is used to produce lubricants for high pressure and temperature conditions to prevent reactions between the powder mixture and the mold. • Excellent thermal conductor and an electrical insulator. – an insulator for high temperature furnaces; it has a dielectric constant of about 4 and a dielectric strength almost four times higher than that of alumina; a thermal conductor BN exceeds almost all other electrical insulators while maintaining high strength and low thermal expansion. Today, due to frequent breakdown of oxide ceramic insulators at temperatures higher than 2000 ∘ C, h-BN insulators are almost exclusively in use. – thermocouple protection tubes and insulating sleeves. Preferably, for the new B4 C/C thermocouple which can be used up to 2200 ∘ C, h-BN sleeves are used. • Chemical and thermal stability combined with the non-wetting property. – Crucible and structural material in metallurgical applications, for example, as a break ring in the casting nozzles. Cross-sections of 50–250 mm diameters are now in use. The break ring determines the zone in which the liquid steel forms a solid. – BN is used for crucibles for growing GaAs single crystals by the liquid encapsulation technique.

51

52

2 Combustion Synthesis of Boron Nitride Ceramics

2.1.4 Method of Synthesis: Advantages and Disadvantages

Several chemical routes are known for the synthesis of h-BN powders. Two of them (i, ii) are used on the industrial scale: (i) Boric acid with ammonia. In this manufacturing process, boric oxide (B2 O3 ), boric acid (H3 BO3 ), or borax (Na2 B4 O7 ) react with ammonia (NH3 ) in the presence of a catalyst (most often tricalcium orthophosphate (Ca3 (PO4 )2 , or CaCO3 , CaO, Zn-borate): 900 ∘ C B2 O3 + 2NH3 −−−−−−→ 2BN + 3H2 O (2.1) The material obtained is heat treated at temperatures up to 2000 ∘ C under nitrogen. This leads to well-formed boron nitride platelets with a thickness of 0.1–0.5 μm and a maximum diameter of 5 μm. (ii) Reaction of boric oxide with organic nitrogen compounds. In this process, boric oxide or boric acid is made to react with nitrogen compounds such as melamine, urea, dicyanamide, or guanidine: >1000 ∘ C (2.2) B2 O3 + CO(NH2 )2 −−−−−−−→ 2BN + CO2 + 2H2 O These reactions are carried out at temperatures between 1000 and 2100 ∘ C in N2 atmosphere. Before final thermal treatment, the product can be washed with methanol or diluted acids in order to remove all non-reacted products. For removing oxygen impurities, a thermal treatment at 1500 ∘ C in inert N2 or Ar atmosphere is used. The process leads to crystal lattices with a less pronounced three-dimensional order, that is, BN with turbostratic structure (t-BN) is obtained, which is characterized by partial or complete absence of three-dimensional order in the stacking of its atomic planes. Although above routes can yield cheap product, the reaction kinetics and the state of art is not given in the scientific literature. Among other conventional methods, let us outline the following: (iii) Direct nitridation of boron in a furnace according to reaction: 2B + N2 = 2BN

(2.3)

This method is limited since pure boron is an expensive element and nitridation of the element is not homogeneous even at high temperatures. (iv) Carbon reduction of boric acid or borates in nitrogen atmosphere at 1000–1500 ∘ C. The mixture is heated up to 1200–1500 ∘ C in nitrogen atmosphere, remaining at maximum temperature for 60–600 min. To enhance the yield of BN, catalysts (CaCO3 , MnO2 ) are sometimes added to the mixture. However, too high carbon content of the mixture leads to the formation of boron carbide. (v) Reacting alkali metal borides or alkaline earth metal borides with silicon and/or aluminum in a nitrogen-containing atmosphere at a temperature

2.1 Introduction

between 200 and 1200 ∘ C. The product is leached after the reaction to remove the water-soluble alkali metal salts. This method, however, requires the preparation of starting materials and a tedious leaching step in order to remove the difficultly soluble trimetal hydroxide impurities. (vi) Pyrolysis of borazine (B3 N3 H6 ) can also yield to BN, primarily in amorphous form. (vii) Fine and ultra-fine BN are used for lubricants and toners and can be produced by combustion of boron powder at 5500 ∘ C in a nitrogen plasma. h-BN materials are solely made by hot pressing (HP) or hot isostatic pressing (HIP) of BN powders, in many cases, using boric oxide as a sintering additive. Typical conditions for HP of boron nitride (h-BN)-based ceramics by reactive HP in a graphite die with BN coating: 1900 ∘ C and pressure of 50 MPa for 60 min, with B2 O3 as sintering additive in a nitrogen atmosphere. The BN powders should have a fine grain size and disordered lattice or turbostratic structure. The produced materials are soft and can easily be machined to the desired size and shape, but the big waste of material during shaping makes the products expensive. As mentioned above, the HP BN ceramics possesses a significant anisotropy in thermal expansion, thermal conductivity, strength, and Young’s modulus relative to hot-press direction. It is worth noting that HIP-BN ceramics show isotropic behavior. It is clear from the above that h-BN materials possess a variety of unique properties and demands on such ceramics consistently increase. However, the production methods are energy and time consuming and thus expensive. Indeed, being extremely refractory material, similar to graphite, with nitride dissociation, it is difficult to sinter this type of compound. Can one suggest a novel approach to produce BN ceramics? The author truly believes that direct synthesis of boron nitride net shape articles by combustion-based approach described below is an effective alternative to the conventional technologies. 2.1.5 Combustion Synthesis in Gas–Solid Systems: General Definitions

Combustion synthesis (CS) is an attractive technique to produce a wide variety of advanced materials including powders and near-net shape products of ceramics, intermetallics, composites, and functionally graded materials. This method was discovered in the former Soviet Union [13]. The development of this technique leads to the appearance of a new scientific direction that incorporates both aspects of combustion and materials science [14, 15]. CS can occur by two modes: selfpropagating high-temperature synthesis (SHS) and volume combustion synthesis (VCS). The characteristic feature of the SHS mode is, after initiation locally, the self-sustained propagation of a reaction wave through the heterogeneous mixture of reactants. The temperature of the wave front can reach high values (2000–4000 K). During VCS, the entire sample is heated uniformly in a controlled

53

54

2 Combustion Synthesis of Boron Nitride Ceramics

manner until the reaction occurs essentially simultaneously throughout the volume. This mode of synthesis is more appropriate for weakly exothermic reactions that require preheating prior to ignition. From the viewpoint of chemical nature, three main types of CS processes can be distinguished. The first is a gasless CS from elements. The second type, the so-called gas–solid combustion synthesis, involves at least one gaseous reagent in the main combustion reaction. And the third type of CS is reduction CS. In this chapter, we consider the gas–solid type of CS, which in general can be described by the following overall reaction formula: �−� � � ∑ ∑ ∑ �� (s) + �� (g) = �� (s, l) + � �=1

�=1

(2.4)

�=1

where Xi (s) are elemental reactant powders (metals or nonmetals) and Yj (g) represents the gaseous reactants (e.g., N2 , O2 , H2 , CO), which, in some cases, penetrate the sample by infiltration through its pores. This type of CS is also called infiltration (or filtration) combustion (FC) synthesis. A general schematic of FC is shown in Figure 2.1. The initial reaction medium consists of a porous matrix of solid reactants and inert diluents, where the pores are filled with gas-phase reactants. The combustion front propagates through the sample with a velocity, U, due to the chemical interaction between the gas- and condensed-phase reactants. Behind the front, the final product is formed, which in some cases may approach pore-free structure, since the volume of the final product grains is typically greater than that of reactant particles. It is necessary to distinguish between the gas initially contained in the pores of the sample (internal reagent) and in the environment (external agent). At relatively low pressures in the reactor, the amount of internal nitrogen may be insufficient to ensure adequate heat required for the propagation of the wave. In this case, combustion wave propagation may be due to infiltration of nitrogen from the environment to the combustion front through pores of the solid reaction B – gas

Vf

U

2

3

Vf

Vf C

A

Vf

A(sol) + xB(gas) = ABx(sol) Figure 2.1

Filtration of gas through the porous frame

Schematic representation of the infiltration combustion.

2.1 Introduction

media. Infiltration takes place due to the pressure difference arising between the combustion zone, in which nitrogen is absorbed (P = 0), and the surrounding atmosphere with constant pressure (P0 ). Indeed, the gas contained in the pores is quickly absorbed in the front of the combustion wave due to the reaction with the solid particles. As a result, the gas pressure in the pores sharply decreases (in the limit to zero), while the external pressure in the reactor remains constant and equal to P0 . There is a pressure gradient ΔP/Δx (where x is the characteristic scale, such as the radius of the sample), which is a driving force in the supply of gas-phase reactant from the outside to the reaction front. This process can be compared with the effect of a chemical pump, which supplies the gas from the environment to the reaction surface in the sample volume. The question naturally arises: is the infiltration of gas from outside necessary for the existence of self-sustained combustion wave? Is it possible to ensure that the gas, initially located inside the pores of the sample, is sufficient for the synthesis process? Obviously, this gas must be at high pressure, but what is the magnitude of this pressure? One can make some simple estimates. Consider the same reaction (1.1) with respect to systems with a single gas reagent: A(s) + 0.5�⋅B2 (g) = AB� (s)

(2.5)

(here, it is accounted that the most commonly used gas reagents, that is, N2 and H2 , are diatomic molecules). The relative porosity of the medium (Π) is the total pore volume divided by the total sample volume. If the mass of the gas in this volume (pore) is such that per mole of solid reactant A we have 0.5x mole of the reactant gas B2 , the reaction will be complete without the participation of the gas from the outside, but solely by the gas stored in the pores. It is easy to show that, to fulfill this condition, the density of the gas in the pores must be equal to the critical value �B : �B =

��B 1 − � �A �A �

(2.6)

where �A and �B – densities of the solid reactant A and gas reactant B2 , respectively; MA , MB – molecular (mole) weights of the reagents. The results of calculations by Equation 2.6 for the “solid reagent – N2 ” system are shown in Figure 2.2. Typically, the relative porosity of the sample does not exceed 0.5–0.6 (samples with greater porosity crumble in the hands), but the samples with a porosity of less than 0.35–0.4 are also rarely used in the combustion mode. As can be seen from Figure 2.2, in the range of actually used porosities (0.35–0.6), the value of the critical density of nitrogen is very high. For most systems, it is higher than the density of liquid nitrogen and for some is higher than the density of solid nitrogen! But is it always necessary to complete the reaction for the propagation of combustion wave? For example, in the Ti–N2 system reaction, heat is so high that the adiabatic combustion temperature reaches 3446 K, while the reaction product TiN melts completely and dissociates partially. Thermodynamic calculations show that if only one-tenth of titanium reacts with nitrogen, the temperature in the reaction zone may exceed 1300 K, and if one-fifth of titanium reacts, the temperature

55

2 Combustion Synthesis of Boron Nitride Ceramics

Hf Ta Ti Si Al V Zr Nb 2.5 Density of nitrogen (g−1 cm3)

56

B

2.0 1.5 Solid nitrogen

1.0

Liquid nitrogen

0.5 0.0

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Porosity Figure 2.2 The critical density of nitrogen in the pores of the sample, sufficient for 100% conversion of the solid reactant to the corresponding nitride. (Adapted from Ref. [16] with permission from FizMatLit.)

is over 2000 K. And it is shown that the combustion wave can self-propagate at these temperatures. The density of nitrogen stored in the pores decreases many times if only a small portion of the solid component reacts in the combustion wave. Although these critical densities and gas pressures still remain pretty high (hundreds or thousands of atmospheres), they can be achieved in high-pressure reactors. Thus, the SHS wave with a single gas phase reactant may in principle propagate due to the gas stored in the pores. However, to complete the reaction during post-combustion stage, it is required to ensure infiltration supply of the gas reagent from the external atmosphere. There is the other method to increase the degree of conversion in the reaction front of the combustion wave. Assume that in addition to solid reactant (A) and pores, filled with reagent gas (B), the original sample already contains some amount of final solid ABx , that is, the reaction equation is as follows: A(s) + 0.5�⋅B2 (g) + �⋅AB� (s) = (1 + �)AB� (s)

(2.7)

For fixed pressure and porosity, the greater the amount of ABx in the initial mixture, the more gas per solid reactant one has in pore skeleton of the sample. In this case, Equation 2.6 for the critical density of the gas takes the form �� =

��� 1−� ��� �� �� ��� + ���� �� �

(2.8)

where �AB is the density of the product ABx . Dilution of the initial mixture by the reaction product is a very effective way to increase the ratio of gaseous and solid reactants inside the sample and often is applied in practice. Its advantage is that incomplete combustion is reduced to a minimum, and the combustion product does not retain any initial component A, thus the entire product consists of

2.2

Combustion in Boron–Nitrogen System

the desired phase ABx . On the other hand, in this case, some amount of released energy is used for preheating of the inert additive, which does not participate in the reaction; therefore, this approach is applicable only to systems with large energy generation. For example, for the Ti–N2 system for y = 8 (7), the adiabatic combustion temperature is ∼1610 K, and for y = 9, it decreases to 990 K, and combustion mode becomes unachievable. Thus, in most cases, infiltration of the reactant gas from the environment into the porous sample is essential for the synthesis of the single-phase product by SHS in the solid–gas systems. It turns out that, in such systems, there are typically two stages of chemical interaction. During the first stage, the reaction proceeds with a degree of conversion that is significantly less than 100%; however, this stage controls the velocity and temperature of combustion wave. The second stage, so-called volume post-combustion stage, – under optimal conditions, may lead to complete conversion of the solid-phase reactant to the final product with desired composition. 2.2 Combustion in Boron–Nitrogen System

There are two different chemical routes to produce h-BN powder by using the combustion-based approach. One is the CS from the elements [17]: 2B + N2 + �⋅BN = 2(1 + �)⋅BN + Q1

(2.9)

The other one is CS by reduction reaction [18]: B2 O3 + 3Mg + N2 = 2BN + 3MgO + Q2

(2.10)

Both relations are exothermic enough to proceed in self-sustained way. The reduction route Equation 2.10 assumes the leaching stage for complete elimination of MgO phase and is used for the synthesis of technical and ultrapure BN powders, the characteristics of which are shown in Table 2.1. This chapter is dedicated to the direct synthesis of BN ceramic net shape articles in the Table 2.1 Some characteristics and properties of BN powders. Content (wt%)

CS product Ultra-pure

Boron nitride, BN Nitrogen, N Oxygen, O Carbon, C Metal impurities (Fe, Mg) Specific surface area (m2 g−1 ) Source: Adapted from Ref. [19].

> 99.5 > 55.7 < 0.5 < 0.01 < 0.2 11.0

Furnace product

Technically pure

ORPAC Grade 99

97.3 54.9 1.5 0.3 0.3 8–14

98–99 54–55 1.5 N/A N/A 10

Denka (Japan)

>98 54.5 1.5 N/A N/A N/A

57

2 Combustion Synthesis of Boron Nitride Ceramics

combustion wave, which can be accomplished only by using other chemical routes, that is, synthesis from elements, which corresponds to route Equation 2.9. 2.2.1 Thermodynamic Considerations

The enthalpy of formation for h-BN is on the order of 60 kcal mol−1 . This nitride melts at high pressure to avoid dissociation at temperature ∼3300 K. The equilibrium partial nitrogen pressure above the BN surface can be calculated as follows: lg�N2 (mm⋅Hg) = 4.0–6450/T. Because of the product dissociation, the combustion in B–N system can be described as a system of the second-type with adiabatic temperature (T ad ) equal to the dissociation temperature of the product [20] and the degree of conversion �=

��B

(2.11)

�0B

where ��B , �0B – mass of reacted and initial boron in the mixture correspondingly is a function of both of nitrogen pressure (P) and amount of the gaseous reagent. To illustrate the concept, Figure 2.3 represents the dependences of T ad and � as a function of P, at stoichiometric ratio (1 : 0.5 mol) between B and N2 . It can be seen that at P = 1 atm, equilibrium degree of conversion equals to ∼0.5 and only at P = 100 atm, it approaches unity. It means that, due to dissociation of the product, having sufficient amount of nitrogen in the system does not guarantee full conversion. It also can be seen that adiabatic combustion temperature is extremely 1.0

4000

0.9

3800 3600

0.7 0.6

3400

0.5 3200

0.4 0.3

3000 Tad η χ=0

2800 2600

Degree of conversion (η)

0.8 Adiabatic temperature (K)

58

0.2 0.1 0.0

0

100

200

300

400

500

Nitrogen pressure (atm)

Figure 2.3 The dependences of adiabatic combustion temperature (T ad ) and degree of conversion (�) as a function of nitrogen pressure at stoichiometric ratio between reagents in B–N2 system.

2.2

Combustion in Boron–Nitrogen System

high (>2700 K) for all considered pressures and essentially equal to the dissociation temperature of the nitride. As mentioned above, the dilution of B by BN is effective way to increase the degree of conversion. Figure 2.4 illustrates how T ad and � depend on � = MBN /MB at P = 1 and 100 atm and stoichiometric amount (�MB moles) of nitrogen. At 1 atm (Figure 2.4a), an increase of � up to 1 : 1 B/BN ratio does not change the T ad , while it significantly increases the degree of conversion. This surprising effect is again related to the dissociation of the product. Also, at some critical amount of dilution (� = 1.2), one can reach full conversion even at low gas pressure. At 100 atm, three characteristic regents can be outlined: (i) � < 0.1 – dissociation temperature defines combustion; (ii) 0.2 < � < 0.7 – another phase transition temperature, that is, BN m.p., equal to T ad ; and (iii) � > 0.7 – increase of � leads to the significant decrease of T ad ; however, it still remains high (>3000 K). 2.2.2 Conditions for Combustion Synthesis of Boron Nitride Material

Initial relative volume (� 0 ) and initial porosity (� 0 ) of the green sample are other important parameters, which, together with dilution (�), define the strategy of the CS of BN ceramics. These parameters can be introduced as follows: ( ( ) ) 1 1 + �� + �� �B∕� + �BN∕� � � � BN BN B BN = �0 ⋅ B (2.12) �0 = = 0⋅ B �� �� (1 + �) (1 + �) �� = 1 − ��

(2.13)

The experimentally obtained dependencies of the maximum combustion temperature (T m ) and the degree of conversion in the combustion front (quenched samples) as a function of the initial nitrogen pressure for different initial sample porosities (� 0 ) and mixtures with variety of BN dilutions (�) are shown in Figure 2.5. It can be seen that experimental T m , being still high (in the range 2200–2600 K), are well below the corresponding values of the adiabatic temperatures, calculated under assumption that the nitrogen pressure above the reaction surface is equal to the initial pressure in the vessel. This discrepancy can be explained accounting two issues: (i) the degree of conversion in the reaction front is below the equilibrium ones and (ii) the real average pressure above the reaction surface is below initial gas pressure in the pores. Both these effects occur due to the infiltration type of combustion in this system. In general, three synthesis stages can be outlined: (i) reaction between initial nitrogen and the boron in the pores; (ii) reaction between B and N2 infiltrated to the combustion front from the outside due to the pressure gradient; and (iii) post combustion reaction due to nitrogen infiltration to the hot reactive surface in the sample volume after combustion front has already passed through. It is worth noting that mixtures with dilutions below some critical value may react at temperature below boron melting point (Figure 2.5a). Finally, Figure 2.5b

59

2 Combustion Synthesis of Boron Nitride Ceramics

1.1

2800 BN dissoc.

2750

3500 3450

1.0

BN dissoc.

1.00

0.98

2650

0.8

2600

0.7

Tad η

3350

P=100 atm Tad

0.9

Degree of conversion

Tad

3400 2700

0.94

BN m.p.

3250

0.92

3200 Tad η

2550

0.6

3150

0.90

P=1 atm 2500 0.00 (a)

0.25

0.50

0.75

1.00

χ, BN dilution (mole)

1.25

1.50

0.5 1.75

3100 (b)

0.96

3300

Degree of conversion

60

0.0

0.2

0.4

0.6

0.8

1.0

χ, BN dilution (mole)

Figure 2.4 The dependences of adiabatic combustion temperature (T ad ) and degree of conversion (�) as a function of degree of dilution of boron by boron nitride (� = MBN /MB ) at different nitrogen pressure, P: (a) 1 atm and (b) 100 atm.

2600 ‘‘Liquid-phase’’ mode 2500

2400 B m.p. 2300 χ=0 χ = 0.7 χ = 1.5 θ0 = 0.6

'solid-phase' mode

2200

2100 0.25

0.50

(a)

0.75

1.00

1.25

1.50

1.75

0.8 0.7 0.6 0.5 0.4

θο = 0.6; χ = 0.7

0.3

θο = 0.6; χ = 0 θο = 0.55; χ = 0 θο = 0.45; χ = 0 1

2.00

Pressure (kBar)

2

3 Pressure (kBar)

(b)

Figure 2.5 (a, b) Dependencies of Tm and h as a function of initial N2 pressure for samples with different initial porosity (� 0 ) and degree of mixture dilution by BN (c).

also demonstrates that dilution indeed leads to the increase of the degree of conversion in the combustion front (compare data for � 0 = 0.6 and � = 0 and � = 0.7). In the case of FC, this effect can be explained not only by larger amount of nitrogen in the initial pore skeleton but also by the fact that higher � leads to the decrease of combustion temperature and thus decrease of the reaction rate, allowing more time for nitrogen to infiltrate to the combustion front [21]. Figure 2.6 supports the latter conclusion, showing that indeed combustion front velocity decreases with the increase of amount of BN in the initial reactive mixture. P = 2000 atm θ0 = 0.6

Combustion velocity (mm s−1)

6.5 6.0 5.5 5.0 4.5 4.0 0.0

0.2

0.4 0.6 BN dilution ( χ)

61

Combustion in Boron–Nitrogen System

Degree of conversion in combustion front

Maximum combustion tempertaure (K)

2.2

0.8

1.0

Figure 2.6 Dependence of combustion front velocity as a function of dilution by BN (�) at initial gas pressure 2000 atm and initial sample porosity � 0 = 0.6.

4

5

62

2 Combustion Synthesis of Boron Nitride Ceramics

One can also estimate the duration of the synthesis process (t s ), which depends on the sample size (L) and is on the order of t s (s) ≈ 2⋅L (cm); for example, for the 10 cm long cylinder, almost regardless of its diameter (D < L), the synthesis time is about 20 s.

2.3 Mechanism of Structure Formation in CS wave 2.3.1 Methods of Investigation of Structural Transformation in Combustion Wave

A variety of physical, chemical, and structural transformations, which take place during CS, makes it necessary to use a wide range of methods for their investigation. The application of a particular method depends primarily on the structural level to be studied. Thus, the evolution of the crystal structure and phase composition can be investigated by the diffraction of the X-ray, neutrons, and electrons. The evolution of the microstructure can be studied, for example, by the methods of optical and electron microscopy. The chemical composition and distribution of the elements are typically determined by the methods of local energy dispersive X-ray spectroscopy (EDS). The transformation on the macro-level can be observed by conventional video-recording technique with a relatively small magnification. To these three structural levels, which are commonly recognized in materials science, in CS, it is necessary to add the thermal structure of the combustion wave. This thermal structure can be examined by the application of different techniques including sets of micro-thermocouples [22] or by IR-imaging [23]. Depending on the characteristic time of the investigated process of structure transformation, the so-called dynamic or the static methods are used. Fast processes and short-lived structures are investigated by in situ (operando) dynamic methods such as high-speed micro-video recording (HSMVR) [24], time-resolved X-ray phase analysis (TRXRA) [25], high-speed optical spectroscopy, and so on. The slower processes that may be quenched during CS can be studied by the static methods (XRD, SEM, TEM, EDS, etc.). The later approaches, on the one hand, are more precise, but on the other hand, the quenching process itself may introduce undesired distortions to the heterogeneous media. Thus, an accurate interpretation of the results is critical. For the boron–nitrogen system, because the high gas pressure is required for the synthesis, it is difficult to apply any dynamic method for investigation of the microstructural transformations, which occur in the combustion front. Thus, the “static” quenching technique was used [26, 23, 27]. The idea of this method is to extinguish the combustion wave and quickly cool the sample; it is necessary to “freeze” all zones with the characteristic microstructure, chemical and phase structure of the reactants, intermediates, and final products. For quenching to take place, the heat loss from the reaction front at some point must exceed the critical

2.3 Mechanism of Structure Formation in CS wave

level above which self-sustained reaction is possible. There are several ways of implementing these conditions in the experiments, which can be divided into two groups. The first group includes quenching methods, in which combustion starts in the low heat loss conditions, and then the heat loss increases abruptly, leading to the extinction of the combustion process. A sharp increase in the heat loss can be achieved, for example, by dropping the burning sample in liquid argon [26] or by spraying it with a strong stream of water [28]. However, the cooling rate when dropping into liquid argon is relatively low due to the formation of an insulating gas layer around the sample, and typically does not exceed 200–300 K s−1 . The gas layer is not formed in quenching using a high-speed jet of a cooling fluid (water), and the cooling rate up to 104 K s−1 is reached in the center of the specimen with 2 mm thickness. This value is the highest cooling rate achieved experimentally for CS systems, but this method is not widely used. Apparently, this is due to the small size of the cooled region and the risk of fracturing the sample by the high-speed water jet. The second type of quenching method is based on the gradual increase of the heat loss with movement of the combustion front through the sample. The prototype of these methods was considered for experiments by Belyaev and coworkers [29] to determine the critical diameter of combustion of condensed matter. During movement of the combustion front from the base of the cone to its top, the specific heat loss increased, reaching the critical value in some cross-section of the cone where the extinction occurred. The most widely used method is the quenching of combustion waves in a wedge-shaped cutout of a massive copper block [27]. The reaction mixture is pressed to the desired porosity in the wedge-shaped cutout of the massive copper block. Lateral compaction using punches was applied to ensure uniform density over the height of the wedge. Then reaction initiates at the upper, wide side of the wedge and the combustion wave propagates from top to bottom and extinguishes before reaching the bottom of the wedge. Figure 2.7 shows a set of IR-video frames of the quenching process. It is evident that the cooling occurs faster in the narrow part of the sample. It is important that the objective is not simply to extinguish the reaction but to organize the processes in a way that the effect of heat losses on the structure of the combustion wave is minimal until the moment of the stoppage of the front, and the cooling rate is maximal at the point of the extinction. Wedge-shaped geometry allows one to achieve such 2100 K

0.33 c 0.67 c 1.00 c

1.33 c

1.67 c

3.33 c

2.17 c 2.67 c

300 K

Figure 2.7 Video footage of the quenching process for the CS wave in a wedge-shaped sample, obtained by using a conventional video camera (left) and IR-thermal video technique. (Adapted from Ref. [16] with permission from FizMatLit.)

63

64

2 Combustion Synthesis of Boron Nitride Ceramics

conditions. As shown by theoretical analysis and computer modeling, the combustion of wedged samples is accompanied by the concentration of excess enthalpy, which carries with it the heating zone [30]. By changing the angle of the wedge and the density of the sample, we can ensure that the combustion will slow down, speed up, or propagate at a constant speed. The last mode is usually optimum for quenching, as it introduces minimum distortion to the thermal structure of the combustion wave and the quenched products adequately reflect the dynamics of structure formation. This method was applied to investigate the mechanism of microstructural transformation during BN formation in the combustion wave. 2.3.2 Mechanism of BN Formation

As mentioned above, there is one known chemical compound of boron and nitrogen with the chemical formula BN, which can exist in three crystalline modifications: hexagonal (graphite), cubic (diamond), and rarely occurring wurtzite. The solubility of nitrogen in the boron is very small and there are no regions of homogeneity of the hexagonal phase of BN. Studies of the evolution of structure formation of the boron nitride in the SHS wave showed that synthesis is accompanied by a sharp change of the scale of heterogeneity of the medium, even if the maximum process temperature was below the melting point of boron [26]. The initial boron powder consisted of agglomerates (specific surface ∼16 m2 g−1 ) of spherical shape particles with a typical size of ∼0.1 μm (Figure 2.8a). The final product, that is, h-BN, has the morphology of plate crystals with the characteristic size of 20 μm and a thickness of less than 100 nm (Figure 2.8b). It is important that the morphology of the product does not depend on the combustion mode and is observed in both the “solid-phase” (T c < T m (B)) and “liquid-phase” (T g > T m (B)) reaction conditions (see Figure 2.5a). In experiments with quenching of the combustion wave, it was found that the periphery of the growing BN plate crystals always has narrow regions (2–3 μm) which have different microstructure and elemental composition as compared to those for the parent crystal (Figure 2.8c). The EDS microanalysis showed that all characteristic impurities of initial boron (mainly oxygen and magnesium) segregate at these “edges.” Round shape crystals, which can be observed only in these peripheral regions, indicate the liquid-phase state of the “edges” in the combustion wave. It is also shown that a small increase in the magnesium content in initial boron results in a significant (two to five times) increase of the combustion rate in the system [31]. These and other results have been used to formulate a mechanism of the formation of boron nitride in the SHS wave [26, 31], which is based on the concept of crystal growth by the vapor–liquid–crystal (VLC) mechanism. Indeed, the sudden change in the morphology of the medium – from submicron spherical boron particles to 20 μm plates of the nitride at temperatures below the melting point of boron – indicates the gas phase growth mechanism of the BN crystals, as the

2.3 Mechanism of Structure Formation in CS wave

5 μm

1.5 μm (a)

(b)

2.5 μm (c) Figure 2.8

Microstructures of particles: initial B (a), final h-BN (b), and quenched h-BN (c).

solid-phase mass transfer mechanisms cannot provide the observed growth rates (up to 10 m s−1 ) of the plate-shaped microstructures in the combustion wave. But how can boron, having a high boiling point (4200 K) and relatively low partial vapor pressure at the synthesis temperature (1−10 Pa), provide sufficiently high rates of transition to the gas phase to explain the observed crystal growth rates? The answer to the question is simple enough: evaporation plus “chemical pump,” similar to what we discussed for FC, working together in the SHS wave, can “deliver” the required amount of gas-phase boron. First, we present simple macrokinetic assessments. According to the kinetic theory of gases [32], the number of molecules vaporized per unit time from unit area (flow J) of a droplet with radius r into an inert gas medium, with the sink of evaporated molecules (e.g., chemical reaction), located at a distance R from its center, is given by: � = −� ⋅

(� (�) − �� )∕(2���� )1∕2 ) �� 1∕2 2 ⋅� (� − � − l�� )∕��(� + lnp ) 1+� 2�� (

(2.14)

65

66

2 Combustion Synthesis of Boron Nitride Ceramics

where P(R) and Pe – the equilibrium vapor pressure over the surfaces of the sink and the evaporating droplet, respectively, lp – the mean free path of the molecule, � – the accommodation coefficient, and D – the diffusion coefficient of the molecules of the evaporated reagent in the gas phase. Assuming that the sinks of boron molecules are the nuclei of boron nitride located in the bulk of the nearest pore (R ∼ 2r), and assuming the diffusion mode of the reaction, P(R) = 0, we can estimate the rate of evaporation in the combustion wave: �e ∕(2����s )1∕2 ) ( ��s 1∕2 1+ ⋅(�∕2�(�s )) 2��

�=

(2.15)

The numerator gives the rate of evaporation of boron particles in a vacuum; the second term in the denominator takes into account the diffusion of evaporated molecules of boron in the gas. The equilibrium vapor pressure of boron at the synthesis temperature of 2300 K is about Pe ∼ 0.1 Pa, r ∼ 1 μm, and D ∼ 10−5 m2 s−1 we obtain the flow of the order of 2 × 10−3 kg (m2 s)−1 , which converted to the diameter of the evaporated particle in 1 s of the process provides a value of ∼1 μm. The critical particle diameter of boron at which the stationary combustion mode can be implemented is dcr ∼ 0.1 μm. The typical combustion rates for the B–N2 system is ∼1–5 mm s−1 ; thus, the characteristic times of the crystal growth processes are on the order of 1–10 s. Thus, it is obvious that 0.1 μm B particle can be easily evaporated through the considered above mechanism. Figure 2.9 schematically represents mechanism of BN a formation in the combustion wave. For the boron–nitrogen system, the liquid-phase edge in which boron and nitrogen dissolve acts as a “chemical pump.” Apparently, in case of

C

O

N

D E N

S A

T

I

O

Tc < Tm.p.(B)

BMgOx - gas

B - gas phase

BMg - gas phase

B-gas phase

BOx-gas phase

BN - thin plates E

N2

N

N2

V

A

P

O R A

T

I

O

B

Liquid reaction controlling edge

N BOx-gas phase

N2

B

Figure 2.9

Schematic representation of the growth mechanism of BN in the SHS wave.

2.4

Combustion Synthesis of Nitride-Based Ceramics

the “solid-phase” combustion mode (T c < T m (B)), in the absence of impurities (e.g., magnesium), it is not possible to implement the self-sustaining combustion mode in the system. Boron, dissolved on the outer boundary of the active liquidphase edge, and nitrogen then crystallize at the inner boundary of this edge with the formation of the nitride phase. Also, note that, in the boron–nitrogen system, the active role in the process of mass transfer can be played by all kinds of boron oxides, which possess high equilibrium pressures at combustion temperatures. As shown in [33], under such conditions, we can expect a closed-loop self-sustaining mechanism of mass transfer of boron through gas-phase even in such a rapidly reacted system as molybdenum–boron, in which the characteristic combustion time is an order of magnitude smaller as compared to that for boron–nitrogen system. The role of impurities or small amounts of additives in the processes of structure formation of SHS products is of the same importance as it is recognized in the case of conventional powder metallurgy techniques.

2.4 Combustion Synthesis of Nitride-Based Ceramics

The benefits of CS are fully exploited in the direct production of ceramic materials in the combustion wave. Three main features should be outlined.

• The first aspect is purely technological and is related to the method for achieving the high temperatures. As mentioned above, the conventional method for production of BN-based ceramics includes several steps, that is, (i) synthesis of boron nitride powders; (ii) milling this powder before preparing the initial compacts; and (iii) using high-temperature furnaces for long-term sintering of BN powders. In CS, high temperature is achieved in the reaction front due to self-heating of the medium and it is logical to use this temperature not only for synthesis of BN phase but also for sintering of the obtained nitride. In this case, we exclude energy consumption, for the milling of BN cakes to produce a homogenized powder and secondly we combine synthesis and sintering process in one stage. Thus, instead of the three-stage technology (BN powder production + milling + traditional sintering), we have a one-stage technology (SHS sintering of materials/products). • The second and third features are fundamental. Many nonmetal ceramic powders, including BN, due to the high melting points, are difficult to sinter. Very high temperatures are required to intensify the mass transfer and achieve the desired consolidation. Currently, there are no furnaces that allow sintering at temperatures above 2400 ∘ C. As shown above, temperature of about 3000 ∘ C can be achieved in the combustion wave. But that is not all. • Due to the characteristics of the crystal structure, it appears that under equilibrium conditions, many ceramic powders cannot be sintered at high temperatures because the corresponding phases (BN, Si3 N4 ) dissociate under such conditions. Typically, to overcome this issue, special additives are used that

67

68

2 Combustion Synthesis of Boron Nitride Ceramics

accelerate mass transfer processes and allow one to consolidate the material at lower temperatures. But, as usual, gaining in one parameter (i.e., lower sintering temperature), we are losing in the other, that is, activating additives significantly deteriorates the physical and high-temperature properties of the ceramics. It turns out that the nonequilibrium conditions of the combustion wave may result in temperatures equal to the dissociation temperature of the products and synthesis can be carried out under such conditions. Indeed, the combustion temperatures in the boron–nitrogen can be equal to the dissociation temperatures of the BN at corresponding gas pressure. Thus, the CS may provide extremely high temperatures for sintering, which are considerably higher than the temperatures one can reach in the advanced conventional furnaces. It is more important that CS also provides unique conditions in the combustion wave, allowing sintering of the powders at the parameters, which in principle cannot be attained by conventional powder metallurgy techniques. Thus, it may be concluded that for some ceramics, such as BN-based materials, CS is the technology, which allows to effectively produce materials that cannot be obtained by traditional methods. 2.4.1 CS of Boron Nitride Ceramics

The following strategy for the CS of ceramic materials in gas–solid systems was suggested [34]: 1) Green samples or net shape articles should have the optimum initial porosity (� 0 ). On the one hand, � 0 should be as small as possible, which makes it easy to produce pore-free material after combustion sintering. On the other hand, as shown above, � 0 is one of the main parameters that defines the combustion regime and the degree of final conversion; for example, too small � 0 could make impossible to accomplish the steady-state combustion regime. 2) Initial reaction mixture is an important factor, which mainly defines the degree of conversion (�). Indeed, as it was mentioned above, by dilution of the boron by boron nitride (�) leads to the significant increase of �. Again, optimization is required because too high dilution results in low combustion temperature and thus hinders the sintering process. 3) In general, higher initial gas pressure (P) favors the properties of the CS ceramics. Indeed, high P allows: (i) higher amount of initial nitrogen in porous skeleton; (ii) increased infiltration rates; and (iii) may lead to the self-compression of the material on the post-combustion stage (see below). The main mechanism for the decrease of the material porosity during CS in gas–solid system is significant increase of sample mass due to the reaction of the solid reactant with the gaseous one. Note that, for BN ceramics, the macroscopic geometrical sizes and shapes remain essentially unchanged after combustion

2.4

Combustion Synthesis of Nitride-Based Ceramics

reaction. The above statement can be expressed in the form of the following equation [34]: [ ( 0/ ) ] � �� = 1 − (1 − �0 )⋅ 1 + � �ℎ � ⋅�� ⋅� (2.16) ��ℎ where � = �0 ∕�� is relative change of the geometrical volume of the sample; �(�0�ℎ ∕���ℎ ) – is a constant for the selected initial mixture; and �0�ℎ ��� ���ℎ are theoretical densities of the initial and final materials, respectively. For (� ∕� )(1+�� ∕�� )−1 B–χBN–N2 system, � = � �� ; �� ∕��� ≈ 1.1; �� ∕�� ≈ 1.3; 1+(� ∕� )� �

��

A ≈ 1.53/(1 + 1.1⋅�), where � is a degree of dilution of the initial boron by boron nitride. It can be seen that higher degree of conversion (� f ) leads to the lower final porosity, that is, more pores are closed by reacted nitrogen (increase of mass); lower initial porosity (� 0 ) under the same other conditions leads to lower final porosity; and higher � results in higher final porosity (less mass can be converted to nitride). However, for the last two parameters, optimization issues, mentioned above (influences of � 0 and � on �), should be taken under consideration. Thus, to optimize materials properties, Equation 2.16 should be solved with equation which defines degree of conversion as a function of initial sample parameters (� 0 , �) and synthesis conditions (P, T 0 ): �f = � (�0 , �, � , �0 )

(2.17)

Solving Equations 2.16 and 2.17 for critical conditions: �f = �fcr

and

�f = �fcr

where �fcr and �fcr are values of the final porosity and degree of conversion, respectively, below which the properties of ceramics do not satisfy the operational requirements. Let us consider the above strategy for the CS of the boron nitride ceramics. It was shown that boron + boron nitride mixture can be easily compacted to the initial porosity less than 0.4 [34]. It can be accomplished because BN powder works as a lubricator allowing one to reach high relative density without sample damaging (crack formation due to the over-pressing effect). The value for A coefficient in Equation 2.16 for � = 0 equals 1.5. The latter means that for B–N2 system, a significant decrease in porosity can be achieved by increase of degree of conversion. Let us assume that initial porosity of the sample is 0.6. In this case, if one can reach the full conversion of B to BN, the pore-free ceramics can be synthesized, that is, 60% of porosity will be eliminated due to increase of sample mass during gas–solid reaction. Experiments also indicated that if CS occurs under constant gas pressure in chemical reactor then geometrical dimensions and shape of the sample remain essentially unchanged, thus parameter >� in Equation 2.16 equals unity. However, if after combustion, wave just passed along the whole sample (which is still under very high temperature), one rapidly increases the gas pressure in the reactor, which leads to the decrease of sample volume (>� > 1).

69

2 Combustion Synthesis of Boron Nitride Ceramics

One more factor should be account. At high gas pressure (>500 atm), when the steady-state combustion regime can be achieved, the combustion front propagates primarily due to reaction between boron and nitrogen, which initially filled the pore skeleton. However, it is also shown that there exist wide post-combustion reaction zones, which significantly increase final degree of conversion. Duration of the post-combustion nitridation depends on the heat loss conditions, which define the duration of the high-temperature zone in the sample volume. Indeed, the more adiabatic the process is the higher final degree of conversion can be achieved under the otherwise similar conditions. Based on the above considerations one can optimize the synthesis parameters to produce ceramics with the desired properties. Figure 2.10 shows the parametric (�fcr –�fcr ) regions for the materials which were synthesized by CS method. 2.4.2 Properties of BN Materials Synthesized by CS Technology

From Figure 2.10, it can be seen that ceramics with the lowest final porosity (� f = 0.1–0.07) and thus highest mechanical properties (� b ∼ 25 MPa; see Table 2.2) contain ∼7–10 wt% of unreacted boron. Other limiting case is that ceramics with highest BN content (∼99 wt%) have final porosity of � f ∼ 0.3 and corresponding bending strength of ∼10 MPa. Some properties of the CS BN-based ceramics are summarized in Table 2.2. For special applications (of magnetohydrodynamic (MHD) generator; erosion-residing sleeve in cutting torches for air-plasma cutting of metals and alloys), which allow the present SiO2 phase, one may significantly increase the mechanical strength of the ceramics without changing the material’s thermal conductivity (see right column Table 2.2) 1.0

Final relative density, vf = 1– θf

70

0.8

0.6

0.4

0.2

B-BN-N2 system

0.0 0.0

0.2

0.4

0.6

0.8

1.0

Final degree of conversion (ηf)

Figure 2.10 Parametric map of the properties for the BN ceramics produced by CS.

2.4

Combustion Synthesis of Nitride-Based Ceramics

Table 2.2 Some properties of BN-based ceramics. Properties

BN

BN + SiO2

Chemical composition (wt%)

N = 55; O < 0.5; Bfree < 0.5; C < 0.5 1.5 25 3.1 0.0034 2 × 108

BN = 74; SiO2 = 26 1.85 19 6.0 0.036 1.3 × 108

20 8 25

N/A 8 50

Density (g cm−3 ) Dielectric strength (kV mm−1 ) Dielectric permittivity Tangent of dielectric loss at 1 MHz Resistivity (Ω) Thermal conductivity (W m−1 K−1 ) T = 400 K T = 900 K Bend strength, � b (MPa) Source: Adapted Ref. [14].

As mentioned above, the dimensions of the samples do not change in the combustion wave. Thus, it is possible to synthesized net-shape articles directly by CS method. This approach is more important for the extremely hard silicon nitridebased ceramics, while h-BN materials are easily machinable. Figure 2.11 illustrates different net-shape articles (tubes, cylinders, screws, blocks, etc.) produced directly in the chemical reactor after minor finishing treatment.

Figure 2.11 Examples of net-shape articles from h-BN ceramics produced by direct CS method.

71

72

2 Combustion Synthesis of Boron Nitride Ceramics

Figure 2.12 High-temperature insulators from BN-ceramics after 30 days of operation in the high temperature oven: product on the left originally produced by conventional sintering with additives; product on the right – insulator made by direct synthesis in the SHS mode.

2.4.3 Examples of Special Application of CS-BN Ceramics

Let us consider one specific example for application of ceramics based on h-BN. As mentioned above, the hexagonal phase of boron nitride has the melting point/dissociation temperature of ∼3000 ∘ C, and performs efficiently in reducing atmospheres at temperatures above 2000 ∘ C. Sintered BN ceramic has semiconductor properties (with a band gap of about 3.7 eV); it is used, for example, as a high-temperature electrical insulator in directional solidification furnaces. It was also discussed that such BN components are usually produced by high-temperature (∼2000 ∘ C) sintering of boron nitride powder with additives of oxygen-bearing oxides. The combustion temperature of boron samples in nitrogen is in the range of 2400–2600 ∘ C. Thus, in producing low-porosity products (∼10%) it is not necessary to use any activators. Figure 2.12 shows high-temperature ceramic insulators after 30 days of work in a furnace: the product on the left was originally produced by the traditional method of sintering with additives, while the insulator on the right was made by direct synthesis in the SHS mode. It can be seen that sample, consolidated by the traditional technology, could not resist long-term (30 days) high-temperature (2400 ∘ C) service in an environment of metal vapors, while the SHS-insulator retained all of its basic properties and may be subsequently used in the company producing single-crystal components. 2.5 Final Remarks

In [16], several issues that are related to the industrial application of CS technologies are outlined.

References

For example, the first reason why the powder metallurgy did not widely use this combustion-based technology is that some SHS-products cannot compete with conventional furnace products due to the high cost of initial precursors. The second obstacle, which occurs in the way of implementation of SHS powders in industry, is due to the fact that their properties are different from that for the conventional counterparts. This does not mean that the properties of SHS products are worse. The point is that the technologies of manufacture materials and articles by sintering or HP were developed for the traditional furnace-made powders. The optimum sintering conditions for the “different” SHS-powders must in many cases be determined again. But the industrial companies are not always interested in such research and change the working conditions of production equipment. Possible solutions to this problem are: (i) use the SHS method to produce powders, which can dramatically reduce the cost of sintering and (ii) produce sintered materials with unique properties directly in the combustion wave. As shown above, direct synthesis of boron nitride-based ceramics is just such a case. Indeed, it is difficult and expensive to produce such ceramics by conventional technology [7, 8]. Moreover, for some special application, which, for example, requires high material purity, it is essentially impossible to produce such ceramics by ordinary approaches. CS-based method, which is characterized by unique synthesis + sintering conditions, which cannot be achieved in the conventional furnaces, is extremely attractive for the industrial use. To make it reality, one has to overcome one more obstacles, which often takes place during the transition of advanced approaches from laboratory to industrial scales, that is, the necessity to keep high level of technological discipline. And not every industrial partner in the field of powder metallurgy is ready for such high-tech novelties. References 1. Balmain, W.H. (1842) Observation on

2.

3. 4.

5.

the formation of compounds of boron and silicon with nitrogen and certain metals. J. Prakt. Chem., 27, 422. Balmain, W.H. (1842) Observation on the formation of compounds of boron and silicon with nitrogen and certain metals. Philos. Mag., 21, 170. Balmain, W.H. (1844) Ueber das Aethogen. J. Prakt. Chem., 32, 494. Inventors Taylor Kenneth, M (1957) Methods for making boron nitride materials. US Patent 3120998 A Dec 5, 1957, Original Assignee, Carborundum Co., Niagara Falls, NY. Inventors Robert, A., Westlake, M., and Lltz, L.M. (1958) Purification of boron nitride by treatment with carbonaceous agents. US Patent 3058809, A Dec 15, 1958, Original Assignee, to Union

6. 7.

8.

9.

Carbide Corporation, a corporation of New York. Chopra, N.G. et al. (1995) Boron nitride nanotubes. Science, 269 (5226), 966–7. Haubner, P., Wilhelm, M., Weissenbacher, R., and Lux, B. (2002) Boron Nitrides – Properties, Synthesis and Applications, in the Book: Structure and Bonding, Vol. 102, Springer-Verlag, Berlin, Heidelberg. Ertu˘g, B. (2013) Powder preparation, properties and industrial applications of hexagonal boron nitride, in Sintering Applications (ed. B. Ertu˘g), InTech, http://dx.doi.org/10.5772/ (accessed 16 May 2014). Bickel, M. and Wedemeyer, H. (1991) Handbook of Inorganic and Organometallic Chemistry, Boron Compounds, 4th Suppl, 8th edn, vol. 3a, Springer, Berlin, Heidelberg, New York.

73

74

2 Combustion Synthesis of Boron Nitride Ceramics 10. Lipp, A., Schwetz, K.A., and Hunold, K.

11. 12. 13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

(1989) Hexagonal boron nitride: fabrication, properties and applications. J. Eur. Ceram. Soc., 5, 3–9. Rudolph, S. (1994) Boron Nitride. Am. Ceram. Soc. Bull., 73 (6), 89–90. Rudolph, S. (2000) Boron nitride (BN). Am. Ceram. Soc. Bull., 79 (6), 50. Merzhanov, A.G. (1990) in Combustion and Plasma Synthesis of HighTemperature Materials (eds Z.A. Munir and J.B. Holt), VCH Publishers, New York, pp. 1–90. Vrama, A., Rogachev, A.S., Mukasyan, A.S., and Hwang, S. (1998) Combustion synthesis of advanced materials: principles and applications. Adv. Chem. Eng., 24, 79–226. Aruna, S.T. and Mukasyan, A.S. (2008) Combustion synthesis and nanomaterials. Curr. Opin. Solid State Mater. Sci., 12, 44–50. Rogachev, A.S. and Mukasyan, A.S. (2013) Book: Combustion for Material Synthesis, Fizmatlit, Moscow. Merzhanov, A.G. and Borovinskaya, I.P. (1972) Self-propagating hightemperature synthesis of refractory inorganic compounds. Dokl. Chem., 204, 429. Merzhanov, A.G. (2012) Book: 40 Years of SHS: A Lucky Star of a Scientific Discovery, Bentham e-book, p. 85. Merzhanov, A.G. (1992) Self-propagating high temperature synthesis and powder metallurgy: unity of goals and competition of principles in: Particulate Materials and Processes: Advances in Powder Metallurgy and Particulate Materials, Metal Powder and Industries Federation, Princeton, NJ, p. 341. Merzhanov, A.G. (1977) New elementary models of the second kind. Dokl. Phys. Chem., 223, 430. Aldushin, A.P., Sepliarskii, B.S., and Shkadinskii, K.G. (1980) Theory of filtration combustion. Combust. Explos. Shock Waves, 16, 33. Zenin, A.A., Merzhanov, A.G., and Nersisyan, G.A. (1981) Thermal wave structure in SHS processes. Combust. Explos. Shock Waves, 17, 63. Rogachev, A.S., Mukasyan, A.S., and Varma, A. (2003) Quenching of gasless combustion wave: time-resolved thermal

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

vision studies. Combust. Sci. Technol., 175 (2), 357–372. Varma, A., Rogachev, A.S., Mukasyan, A.S., and Hwang, S. (1998) Complex behavior of self-propagating reaction waves in heterogeneous media. Proc. Natl. Acad. Sci. U.S.A., 95, 11053–11058. Merzhanov, A.G., Borovinskaya, I.P., Khomenko, I.O., Mukasyan, A.S., Ponomarev, V.I., Rogachev, A.S., and Shkiro, V.M. (1995) Dynamics of phase formation during SHS processes. Ann. Chim., 20, 123. Mukasyan, A.S. and Borovinskaya, I.P. (1992) Structure formation in SHS nitrides. Int. J. SHS, 1, 55. Merzhanov, A.G., Rogachev, A.S., Mukasyan, A.S., and Khusid, B.M. (1990) Macrokinetics of structural transformation during the gasless combustion of a titanium and carbon powder mixture. Combust. Explos. Shock Waves, 26, 92. Khusid, B.M., Khina, B.B., Kuang, V.Z., and Bashtovaya, E.A. (1992) Numerical investigation of quenching of the state of a material in a wave of SHS during a two-step reaction. Combust. Explos. Shock Waves, 28, 389. Bakhman, N.N. and Belyaev, A.F. (1967) Combustion of Heterogeneous Condensed Systems, Nauka, Moscow. Stepanov, B.V. and Rogachev, A.S. (1992) Quenching of solid-phase combustion front of a symmetric sample by supercritical heat loss. Int. J. Self Propag. High Temp. Synth., 1 (3), 409–416. Mukasyan, A.S. (2005) Combustion synthesis of nitrides: mechanistic studies. Proc. Combust. Inst., 30, 2529–2535. Hirth, J.P. and Pound, G.M. (1963) Condensation and Evaporation, Pergamon Press, Oxford. Kashireninov, O.E., Fomin, A.A., and Yuranov, I.A. (1991) Thermodynamics of the formation of gas-phase intermediate compounds in the molybdenum-boron Reacting SHS-systems. Khim. Fiz., 10 (4), 524–533. Mukasyan, A.S. (1986) Characteristics and mechanism of silicon and boron combustion in nitrogen. PhD thesis. Institute of Chemical Physics, USSR Academy of Science.

75

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies Zhongqi Shi, Yoshinari Miyamoto, and Hailong Wang

3.1 Introduction

Aluminum nitride (AlN) is an important III–V group semiconductor material, which has aroused great attention due to many outstanding features including high thermal conductivity (∼320 W m−1 K−1 ), high electrical resistance (>1010 Ω cm), low dielectric constant (8.6), wide band gap (6.2 eV), and low thermal expansion coefficient (4.2 × 10−6 ∘ C)−1 match to both Si and GaN semiconductors [1]. All these properties make AlN an excellent candidate for use in electronic, light-, and field-emission devices [2]. Since the properties of inorganic materials are highly dependent on their crystalline size, shape, and structure [3], many efforts have been devoted to synthesis and assembly of various inorganic materials with specific morphologies and structures in nanometer and micrometer size. Up to now, AlN particles with various one-dimensional (1D), two-dimensional (2D), or three-dimensional (3D) morphologies have been prepared by different methods such as arc-discharge, chemical vapor deposition, carbothermal reduction, and nitridation, as well as other chemical routes [4]. However, these methods usually require high temperature, catalysts, substrates, and long-term production cycle. In addition, the yield of products is still low. Therefore, it is imperative to further exploit new synthetic routes for preparation of AlN powders with controlled grain morphologies in large scale. Combustion synthesis (CS), also known as self-propagating high-temperature synthesis (SHS), has become a promising choice for industrial fabrication because of its low processing cost, high energy efficiency, and short reaction period. To date, many CS processes have been developed for synthesis of pure AlN powders [5]. However, because of the high thermal gradient and fast reaction speed in the CS process, as well as the low melting temperature of metal Al (∼660 ∘ C), the morphology of the AlN product is difficult to control and often consists of various grain morphologies such as agglomerated particles, whiskers, faceted particles, rods, pyramids, and so on [6]. Although a uniform morphology is very important to engineer the properties of AlN-based materials or devices, Nitride Ceramics: Combustion Synthesis, Properties, and Applications, First Edition. Edited by Alexander A. Gromov and Liudmila N. Chukhlomina. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

76

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies

a morphology-controlled synthesis condition is still difficult to realize, which hinders the combustion-synthesized AlN products for wide applications. Recently, we have reported the synthesis of 1D AlN nanowhiskers, 3D flowerlike AlN microstructure, and AlN porous-shell hollow spheres with uniform morphologies by the CS method, where the morphologies of products were controlled by manipulating combustion parameters such as nitrogen pressure, ball milling process, diluent ratio, and type and content of the catalyst [7]. These results demonstrate that the AlN grain morphologies fabricated by the CS process can be controlled, which could give a chance to extend the application areas of AlN. The structure of this chapter is as follows. Section 3.2 describes the fabrication of 1D quasi-aligned AlN nanowhiskers by CS process. Section 3.3 illustrates the application of the as-synthesized AlN nanowhiskers as inorganic fillers for polymer-matrix packaging materials. Sections 3.4 and 3.5 show the growth of 3D flower-like AlN by CS assisted with mechanical activation (MA) and CS of AlN porous-shell hollow spheres, respectively.

3.2 Combustion Synthesis of Quasi-Aligned AlN Nanowhiskers

CS process represents a cost-effective alternative for the production of pure AlN powders [5]. A lot of literatures have reported the formation of AlN nanostructure by CS [8]. However, due to the high thermal gradient and fast reaction speed in the CS process, the morphology of the AlN product is difficult to control and often consists of various grain morphologies, which prevents the combustionsynthesized AlN from wide applications. Until very recently, we [7a,b] reported novel observation of the growth of 1D quasi-aligned AlN nanowhiskers inside the reacting Al particles during the combustion under a low nitrogen pressure (≤1.0 MPa). Through this CS process, the high content of the AlN nanowhiskers with uniform morphology in the products has been achieved. The present process for synthesis of AlN nanowhiskers is simple, productive, reproducible, energy saving, and can be applied to produce fillers as reinforcement for electrical packaging, structural composites, and nanodevices. 3.2.1 Experimental Methods of Approach

The starting materials were high-purity Al (>99.9%, ∼23 μm, Toyo Aluminum K.K., Tokyo, Japan), AlN diluent powder (type H, >99.9%, ∼0.5 μm, Tokuyama K.K., Hino, Tokyo, Japan), and NH4 Cl additive (>99.0%, Nacalai Tesque, Inc., Kyoto, Japan). The morphological characteristics of Al powder and AlN diluent are shown in Figure 3.1. In a typical experimental procedure, Al and AlN powders were mixed with a molar ratio of 4 : 6, which was chosen according to a previous study to achieve looser and full AlN product [9]. Additionally, 6 wt% NH4 Cl was also added as a promoting additive to control the morphology of the product. The

3.2 Combustion Synthesis of Quasi-Aligned AlN Nanowhiskers

20 μm bk0002 25 kV

50 μm

(a) Figure 3.1

500X

A+B

2008/06/03

77

20 μm bk0007

25 kV

2 μm

10000X

A+B

2008/06/03

(b) Morphological characteristics of starting powders: (a) Al and (b) AlN diluent.

powders were lightly mixed using mortar for 10 min, and then sieved through a 212-mesh sieve to disperse any large agglomerates. Fifty grams of the mixture was poured into a porous graphite container (Φ 42 mm × 90 mm H) at a tapping density of 0.6 g cm−3 . Then, the container was placed into a combustion chamber, and two W–Re thermocouples protected by alumina tubes were inserted into the center of the mixture (one at the middle and the other near the top surface) at a fixed distance of 30 mm to record the temperature–time pattern of the combustion and determine the combustion speed by measuring the time lapsed for the wave passage between the two thermocouples. The chamber was evacuated and then filled with high-purity N2 (99.99%) at a pressure of 1.0 MPa. The mixture was ignited from the bottom with an ignition pellet (2 g, Al/AlN = 1/1 wt%) by passing an electric current of 60 A × 20 V for 10 s through a carbon ribbon under the pellet. The schematic diagram of the CS autoclave is shown in Figure 3.2. After the first combustion reaction, the as-synthesized AlN product, which was a loose cake composed of quasi-aligned nanowhiskers (QANs) and original AlN diluent, was broken lightly by a mortar. Then, the product was sieved through 212-mesh sieve to use as the diluent for the second combustion reaction with similar conditions as the first time. The procedures were repeated two times. Finally, AlN nanowhiskers with high content in the final product were achieved. The phase purity of the as-synthesized products was examined using powder X-ray diffraction (XRD; JDX-3530, JEOL, Tokyo, Japan) with Cu K� radiation. The morphology of the as-synthesized products was observed by field emission scanning electron microscopy (FESEM; ERA-8800, ELIONIX, Tokyo, Japan) equipped with energy dispersive spectrum (EDS). Samples for scanning electron microscopy (SEM) observations and EDS analyses were coated with thin films of sputtered gold to reduce electrical charge-up. Transmission electron microscopy (TEM; JEM-2010, JEOL, Tokyo, Japan) was used for further characterization of the products, where both TEM images and selected area electron diffraction (SAED) patterns were acquired.

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies

Carbon felt Reactant mixture

Graphite container

Optical pyrometer 90mmH

W/Re thermocouples 42mm �

Ignition heater

N2 gas

Vac. Igniter Figure 3.2

A schematic diagram of the CS autoclave.

3.2.2 Results and Discussion

(112) (200)

30

40

50

60

2θ (°)

Figure 3.3

XRD pattern of the as-synthesized AlN product.

70

(201)

(110)

(002)

(102)

(103)

(101)

(100)

Figure 3.3 shows a typical XRD pattern of the as-synthesized product. All the diffraction peaks can be indexed to the hexagonal wurtzite structure of AlN crystal (JCPDS No.25-1133). No characteristic peaks of impurities were detected in the pattern. The sharp diffraction peaks indicated the good crystallinity of the product. The typical morphologies of the product are shown in Figure 3.4. Figure 3.4a is a low-magnification FESEM image which clearly shows that the AlN powders

Intensity (a.u.)

78

3.2 Combustion Synthesis of Quasi-Aligned AlN Nanowhiskers

(a)

(b)

Al N

10 μm

Au 1.00

5 μm

2.00

(c)

200 nm

5 μm

0.5 μm (002)

(e)

(d)

(100)

(f) [001]

[010]

0.5 μm Figure 3.4 (a) A low-magnification FESEM image of the as-synthesized AlN product with many anemone-like particles; (b–d) detailed FESEM images of a single anemonelike particle, a cross-sectional view and a broken part of the particle, respectively;

100 nm (e) high-magnification FESEM image of the as-synthesized AlN nanowhiskers; and (f ) representative TEM image of the AlN nanowhiskers and corresponding SAED pattern. (Reproduced with permission from Ref. [7a]. Copyright © 2009, Elsevier Limited.)

in the product highly dispersed in the space without any aggregations, and two major types of morphologies can be clearly seen: One is irregular particles, which are of similar size and shape as original AlN diluent (see Figure 3.1b); the other is anemone-like particles, which are of similar size and shape as original Al particles (see also Figure 3.1a). Figure 3.4b–d shows the detailed FESEM images of a single anemone-like particle, a cross-sectional view, and a broken part of the particle, respectively. It can be observed that the particle is covered with a thin AlN

79

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies

shell (∼200 nm), and numerous AlN QANs grow from the shell into the interior. The as-synthesized nanowhiskers are shown more clearly in high-magnification FESEM image in Figure 3.4e, which shows that the AlN nanowhiskers have the diameters in the range of 80–170 nm and a length of several to several tens of micrometers. In FESEM observation, there was no grain growth or sintering for both the formed and original AlN diluent particles, and the uniform AlN QANs were formed mainly inside the reacting Al particles. Figure 3.4f shows a representative TEM image of AlN nanowhiskers and the corresponding SAED pattern. It indicates that the diameters of AlN nanowhiskers are about 80–170 nm in accordance with the FESEM image (Figure 3.4e) and the nanowhiskers are quite straight with uniform diameters along their lengths. In addition, since no droplets are observed at the tip of the nanowhiskers, the growth of the AlN nanowhiskers is probably governed by a VS (vapor–solid) mechanism. The corresponding SAED pattern (Figure 3.4f, inset) indicates that the AlN nanowhiskers are indeed a single crystal with hexagonal wurtzite structure and grow along the [0 0 1] direction. Figure 3.5 is the schematic illustration of the effect of different amount of NH4 Cl on the combustion behavior, nitridation mechanism, and products’ morphology. As can be seen, the combustion behavior can be controlled from an explosive mode and brought to mild by NH4 Cl. With the addition of NH4 Cl

AI particle

Combustion

Synthesis Irregular Quasi-aligned AIN AIN micro-rods nanowhiskers

Integrated AIN hollow

Broken AIN hollow

Direct nitridation 1.2

2100

1.0

2000

0.8

1900

0.6

1800

0.4

1700

0.2

1600

0.0

1500 0

1

2

3

4

5

Combustion temperature (°C)

Micro–reactor model Spontaneous chlorination–nitridation

Core–shell model

Combustion speed (mm • s–1)

80

6

Amount of NH4Cl (wt%)

Figure 3.5 Schematic illustration of the effect of different amount of NH4 Cl on the combustion behavior, nitridation mechanism, and products’ morphology. (Reproduced with permission from Ref. [7a]. Copyright © 2009, Elsevier Limited.)

3.2 Combustion Synthesis of Quasi-Aligned AlN Nanowhiskers

from 0 to 6 wt%, both combustion temperature and combustion speed decrease. When combustion temperature decreases from 2038 to 1707 ∘ C, the corresponding combustion speed decreases drastically from 1.12 to 0.14 mm s−1 . All the combustion temperatures are much higher than the melting point of Al particles. Therefore, the Al particles melt, followed by coalescing to a ball-like form due to the surface tension [10]. The nitridation takes place at the surfaces of these molten Al particles forming nitride shells surrounding the molten Al [8b]. For the case of the experiment without NH4 Cl additive, once the combustion reaction triggers, the temperature increases rapidly close to its apex. Due to the high thermal stress, the inner molten Al expands enough to break the new-formed AlN shell and then pours out rapidly as Al vapor to react with nitrogen gas in a direct nitridation pathway. The reaction can be expressed according to the following equation: Al(g) + N2 (g) → AlN(s)

(3.1)

As a result of high supersaturation of the Al vapor, the as-formed in situ hole is covered with a shell consisted of many fine, smooth, and homogeneous spherical AlN particles [10]. Therefore, the product with a porous broken-hollow morphology is obtained. However, by adding NH4 Cl into the starting materials and increasing its amount, the combustion process gradually changes to a mild one with no explosive mode and has a relatively low heating rate. This change is due to the sublimation and dissociation of NH4 Cl according to the following reaction: NH4 Cl(s, g) → HCl(g) + NH3 (g)

(3.2)

The reaction (3.2) is endothermic and absorbs sufficient heat from the samples to disturb the direct nitridation of Al particles with N2 gas, which decreases the combustion temperature and retards the wave propagation. Hence, with the amount of NH4 Cl increases, integrated hollow, ball-like particles with irregular micro-rods, or QANs grown inside were obtained in the products with the addition of 3, 5, or 6 wt% NH4 Cl, respectively. The morphological change has strong relationship with the supersaturation of the Al vapor. With the amount of NH4 Cl increases, more and more heat is absorbed. So the combustion temperature decreases, which results in the decrease of supersaturation of the Al vapor. And we found that if the amount of NH4 Cl in the starting materials is higher than 6 wt% in the present work, the combustion cannot be triggered. In other words, it reaches the combustion limit. The combustion limit promotes the formation of ball-like particles with an undeveloped surface as fabricated in the present work, which has been approved by Zakorzhevskii and Borovinskaya [5c]. Moreover, the addition of NH4 Cl to starting materials offers a different reaction pathway than the direct nitridation mechanism. This different nitridation proceeds via spontaneous chlorination–nitridation sequences similar to the process of direct nitridation of an Al/NH4 Cl mixture as reported by Radwan et al. [11]. The encountered reactions can be described according to the following reactions along with the reaction (3.2): Al(l) + 3HCl(g) → AlCl3 (�) + 3∕2H2 (g)

(3.3)

81

82

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies

AlCl3 (g) + 1∕2N2 (g) + 3∕2H2 (g) → AlN(s) + 3HCl(g)

(3.4)

The spontaneous chlorination–nitridation reactions were facile for the growth of 1D AlN nanostructures [11, 12]. Therefore, we can understand the growth mechanism of the AlN QANs as follows: After the ignition, a thin AlN shell is formed on the surface of Al particle, which can be seen as a microreactor and function as a self-catalyzed substrate for the growth of nanowhiskers. Then, various gaseous species such as HCl(g), NH3 (g), and N2 (g) slowly infiltrate through the shell into the molten Al core. The molten Al core is spontaneously halogenated to AlCl3 (g) and then nitrided to AlN embryos. Because the nitridation of Al core is suppressed by the AlN shell, the supersaturation degree is low in the microreactor. In this case, the embryos preferentially deposit on nucleated sites and tend to epitaxial growth along [0 0 1] direction by VS mechanism. Therefore, the AlN nanowhiskers are grown with a unique morphology that oriented growth toward the center of the reacting particle. The probable growth models of AlN QANs by the CS process are present in Figure 3.6. However, due to the relatively higher supersaturation degree [13] (caused by the higher combustion temperature and speed) in the microreactor for the sample with 5 wt% NH4 Cl, many irregular micro-rods are grown inside the reacting particle. However, the content of AlN nanowhiskers in the sample with 6 wt% NH4 Cl is not high enough for the large-scale application (at most 40 vol%, same as the mass content of reacting Al powders). In order to increase the content of nanowhiskers in the product, we repeated the combustion reaction two times with similar conditions just using the as-synthesized AlN powders obtained from the former product as diluent. Finally, high content (maximum ∼80 vol%) of AlN nanowhiskers in the final product were achieved. The approximate content of nanowhiskers in the final product was estimated according to the total content of reacting Al powders added into the mixture. In order to open the anemone-like AlN particles and utilize the QANs, the two products with different aspect ratio (40 and 80 vol%) of anemonelike particles were ground by ball milling for 10–20 min in alcohol and then dried, as shown in Figure 3.7. Then, 3D brush-like AlN particles with 40 and 80 vol% in the synthesized AlN powders were obtained (Figure 3.7a,d).

A micro-reactor N2(g), NH3(g), HCI(g) infiltration

Al particle

Al core

Surface defect

Figure 3.6

AIN crust

Growth of AIN QANs

AIN QANs

Schematic illustration of the growth model of AlN QANs by the CS method.

3.3 Enhanced Thermal Conductivity of Polymer Composites Filled with 3D Brush

(a)

(b)

10 μm (c)

5 μm (d)

5 μm

5 μm

Figure 3.7 FESEM images of the AlN products with 40 and 80 vol% of anemone-like particles ground for 10–20 min: (a) 40 vol%, ground for 20 min and (b–d) 80 vol%, ground for 10, 15, 20 min, respectively.

3.3 Enhanced Thermal Conductivity of Polymer Composites Filled with 3D Brush-Like AlN Nanowhiskers by Combustion Method

Polymer-matrix composites have been used as one of the most common packaging materials for encapsulating a variety of electronic components for dissipating heat [14]. In this section, 3D AlN nanowhiskers with brush-like structure were filled into the polymer matrix to enhance its thermal conductivity. The 3D brushlike AlN fillers were fabricated by CS process [7a], as illustrated in Section 3.2. The use of AlN as a filler candidate to enhance the thermal conductivity of the polymer is attributed to its attractive properties such as high thermal conductivity, high electrical resistivity, and good chemical stability with polymers [1]. To explore the promoting effect of the 3D brush-like AlN fillers on thermal conductivity, three types of AlN fillers with different brush-like filler aspect ratio were added into polymer matrix to fabricate a series of composites and their thermal conductivities were measured. The results demonstrated that the 3D brush-like AlN nanowhiskers fillers can effectively enhance the thermal conductivity of the polymer composite.

83

84

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies

3.3.1 Experimental Methods of Approach

Three types of AlN fillers were used. One was commercially available AlN powders (equiaxed particle, >99.9%, ∼0.5 μm, Tokuyama K.K., Japan), as shown in Figure 3.1b. The other two types of AlN fillers with different aspect ratio of 3D brush-like particles (40 and 80 vol%, respectively) were fabricated by CS method, as shown in Figure 3.7a,d. For simplicity, according to their morphological characteristics, the three types of AlN fillers are named as EP, BP40, and BP80. The polymer/AlN composites were fabricated according to the following steps. First, the epoxy resin (O-Cresol Novolac), curing agent (Phenol-Novolac), and catalyst (triphenylphosphine) with a weight ratio of 60 : 40 : 0.5 were dissolved in acetone and then mixed with each of the AlN fillers (0–47 vol%) by using a centrifugal mix-defoaming machine (AR-250, THINKY Co., Japan) to form a homogeneous slurry. Then the slurry was dried in a vacuum oven at 80 ∘ C and ground to pass through a 50-mesh sieve. Finally, the powder mixture was cured at 180 ∘ C for 2 h under a pressure of 10 MPa to obtain the composites. The thermal conductivities � of the composites were calculated by equation � = � ⋅ �� ⋅ �, where �, �� , and � are the thermal diffusivity, specific heat, and density of the composites, respectively. � of the composites was measured by laser flash method (TC-7000, Sinku-Riko, Japan); � was calculated by the density of AlN of 3.26 g cm−3 and the measured density of the polymer matrix (1.31 g cm−3 ); �� was determined by the specific heat of AlN of 0.74 J g−1 K−1 and the measured specific heat of the polymer matrix (1.27 J g−1 K−1 ). 3.3.2 Results and Discussion

Figure 3.8 shows the thermal conductivity as a function of the volume fraction of different AlN fillers. The thermal conductivities increase with the increase in filler content. For the composites filled with 47 vol% (70 wt%) of BP40 and BP80, the thermal conductivities are 3.3 and 4.2 W m−1 K−1 , respectively. These are 1.8–2.3 times higher than that of the polymer composite filled with the same content of EP. In addition, the measured thermal conductivity of the polymer composite filled with the different content of EP is matched with the predicted value by the Bruggeman model [15], which was thought to be the Bruggeman model was based on spherical particles suspended in a diluent matrix, similar in this composite. The Bruggeman model can be given by: 1 − �� =

(�� − �)(�� ∕�)1∕3 �� − ��

(3.5)

where �, �� , �� , and �� represent thermal conductivities of the composite, filler, matrix, and volume fraction of the filler in the composite, respectively. The values of 200 and 0.25 W m−1 K−1 were used for �� and �� , respectively.

3.3 Enhanced Thermal Conductivity of Polymer Composites Filled with 3D Brush

Thermal conductivity (W mK−1)

5.0 4.5

EP

4.0

BP40

3.5

BP80 Bruggeman model

3.0 2.5 2.0 1.5 1.0 0.5 0.0 0

10

30 40 20 Volume fraction of AIN (%)

50

Figure 3.8 Thermal conductivity of the composites as a function of filler content. (Reproduced with permission from Ref. [7e]. Copyright © 2009, AIP Publishing LLC.)

Remarkably, the thermal conductivities of the composites filled with BP40 and BP80 are much higher than the prediction of Bruggeman equation. This demonstrates that the brush-like AlN particles enhance the thermal conductivity of the polymer matrix significantly. The intrinsic reason can be explored by Agari model [16], which considers the effect of dispersion state by introducing factors C 1 and C 2 : log � = �� �2 log �� + (1 − �� ) log(�1 �� )

(3.6)

where �, �� , �� , and �� are defined as the same as before; C 1 is a factor relating to the effect of the filler on the secondary structure of the polymer and C 2 is a factor relating to the ease in forming conductive chains of the filler. The values of C 1 and C 2 should be between 0 and 1, the closer C 2 values are to 1, the more easily conductive chains are formed in composite. So, if the dispersion system is different, the thermal conductivity of the composites may be different even if the components in the composites are the same. Through data fitting, C 1 and C 2 for the composites containing 47 vol% of the three different types of AlN fillers are obtained and shown in Table 3.1. The AlN fillers with different aspect ratio of 3D brush-like particles affect the C 2 values more than the C 1 values. This indicates that brush-like particles do not change the secondary structure of the polymer significantly. However, the C 2 value increases with the aspect ratio of brush-like particles, which means the formation of thermal conductivity paths in the composites strongly enhanced by 3D brush-like AlN particles. This result can also be approved by their FESEM images (Figure 3.9). In the case of the composites filled with EP, each particle could be insulated easily, which prevents the formation of

85

86

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies

Table 3.1 C 1 and C 2 of Agari model for the composites containing 47 vol% of different types of AlN fillers. Filler type

EP BP40 BP80

C1

C2

Thermal conductivity (W m−1 K−1 )

1.046 1.012 1.004

0.540 0.797 0.889

1.8 3.3 4.2

Source: Reproduced with permission from Ref. [7e]. Copyright © 2009, AIP Publishing LLC.

(a)

(b)

2 μm Equiaxed particle

(c)

2 μm

2 μm

3D brush-like nanowhiskers

Epoxy matrix

Figure 3.9 FESEM images and corresponding schematic morphologies of the polymer/AlN composites filled with 47 vol% of the EP (a), BP40 (b), and BP80 (c) fillers, respectively. (Reproduced with permission from Ref. [7e]. Copyright © 2009, AIP Publishing LLC.)

thermal conductivity paths. However, with 3D brush-like particles filled in, and increase of their aspect ratio, the thermal conductivity paths can be formed and enhanced significantly. The results demonstrate that the 3D brush-like AlN fillers are effective for packaging materials with high thermal conductivity.

3.4 Growth of Flower-Like AlN by Combustion Synthesis Assisted with Mechanical Activation

Because of the high heating and cooling rates during the CS process, some nonequilibrium phases and novel microstructures can be formed in the products. In addition, recent investigations have shown that MA of the reactants by highenergy milling can strongly enhance the acuity degree of CS process as the MA

3.4

Growth of Flower-Like AlN by Combustion Synthesis Assisted with Mechanical Activation

endows the reactants with high activity, small particle-size distribution, and more strain and defects [17]. Therefore, mechanical activation assisted combustion synthesis (MACS) may have a great effect on the structure and properties of the products. In this section, we introduce the synthesis of 3D flower-like AlN microstructure by MACS process using NH4 Cl as a reacting additive. We also propose a possible mechanism for the growth of AlN microflowers. 3.4.1 Experimental Methods of Approach

The synthesis experiments were carried out in a conventional combustion chamber (see Figure 3.2). The starting materials were high-purity Al (>99.9%, ∼23 μm), NH4 Cl reacting additive (>99%), and AlN diluent powder (>99.9%, ∼0.5 μm). In a typical experimental procedure, Al, NH4 Cl, and AlN powders were mixed with a weight ratio of 70 : 20 : 10 for 2 h using Al2 O3 balls as milling media with a ball : charge weight ratio of 10 : 1 by a high-speed ball-milling method, and then sieved through a 212-mesh sieve to disperse any large agglomerates. The mixture (50 g) was ignited in a porous graphite container under a nitrogen gas pressure of 2.0 MPa. The combustion was initiated from the bottom of the mixture. The characterization of the products was same as the description in Section 3.2. 3.4.2 Results and Discussion

XRD result showed that the as-synthesized product is pure hexagonal wurtzite structure of AlN crystal. EDS pattern illustrated that the product mainly consisted of Al and N elements with a trace of O. The presence of O element probably came from the relatively low vacuum degree of the combustion chamber. The XRD and EDS patterns can be seen in our previous study [7c]. The morphologies of the as-synthesized AlN particles are shown in Figure 3.10. Figure 3.10a is a low-magnification SEM image which shows clearly that the product mainly consists of flower-like particles with a diameter of 2–5 μm, and these AlN flowers highly disperse in the space without any aggregations. Figure 3.10b,c shows the high-magnification SEM and TEM images of the AlN microflowers, respectively. These flowers are composed of many cone-like petals with the center diameter about 500 nm and length in the range of 1–3 μm extending radially from center to outside. Figure 3.10d shows a TEM image of a single petal of 500 nm in width and 2.5 μm in length. No droplets can be found at the tips of petals, indicating that the AlN flowers were grown by VS mechanism. The corresponding SAED pattern (insert of Figure 3.10d) indicates that the petals of the AlN microflowers are single crystal with hexagonal wurtzite structure and grow along the [0 0 1] direction. In order to explore the formation process of the flower-like AlN microstructure, several experiments were performed. The synthesis process was similar to the typical procedure described above, except for change in the MA time or the NH4 Cl

87

88

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies

(a)

(b)

2 μm

1 μm (d)

(c)

[010] (002)

(100)

1 μm

200 nm

Figure 3.10 (a) Low-magnification SEM image of the as-synthesized product; (b, c) high-magnification SEM and TEM images of the AlN microflowers, respectively; and (d)

TEM image of a cone-like petal and corresponding SAED pattern (inset). (Reproduced with permission from Ref. [7c]. Copyright © 2011, TTP INC.)

content. Figure 3.11a is a typical image of the product obtained by 1 h MA with 20 wt% NH4 Cl additive, which shows that the AlN spheres have rough surfaces and many little AlN columnar particles grow on it. Figure 3.11b shows a typical image of the product obtained by decreasing the NH4 Cl content to 10 wt% with 2 h MA. As can be seen, the product is composed of many imperfect flower-like particles with numerous little AlN islands sprouted from their surfaces. The above experimental results indicate that both the MA time and the NH4 Cl content play important roles for the growth of AlN microflowers. MA process can enhance greatly the activity of the reactant powders with reduced particle size, increased defect concentration and strain energy, as well as enlarged fresh surface area. The activity of the reactant powders increases with the MA time. In the case of CS process without MA, the reaction was difficult to trigger. However, after MA for 1 h, the reactant powders could be ignited, and AlN spheres covered with many little AlN columnar particles in the product were obtained (Figure 3.11a). With the MA time increased to 2 h, numerous prefect AlN microflowers were formed in the product (Figure 3.10a). Therefore, appropriate MA time was necessary for the preparation of AlN microflowers, which strongly promoted the development of the petals. The NH4 Cl content also played a key role in the growth of flower-like AlN particles. It has been reported that the addition of NH4 Cl to the reactant

3.4

Growth of Flower-Like AlN by Combustion Synthesis Assisted with Mechanical Activation

(a)

(b)

500 nm

1 μm

500 nm

2 μm

Figure 3.11 Typical morphologies of different AlN products prepared in: (a) 20 wt% NH4 Cl with 1 h MA and (b) 10 wt% NH4 Cl with 2 h MA. (Reproduced with permission from Ref. [7c]. Copyright © 2011, TTP INC.)

Al powders during CS can strongly promote the growth of 1D structures in the products [8c]. The spontaneous chlorination–nitridation mechanism was thought to be the primary reason for the growth of 1D structures during the process [11], and three main encountered intermediate reactions are considered to be involved, as described in Equations 3.2–3.4. However, in the case of the NH4 Cl content decrease to 10 wt%, AlN product with imperfect flower-like microstructure was formed (Figure 3.11b), indicating that the amount of NH4 Cl was not enough to promote the full growth of AlN microflowers. On the basis of the above analyses and characterization, a possible growth mechanism for the AlN microflowers was proposed (as shown in Figure 3.12). After appropriate MA process, the reactant Al powders were greatly active and facile for ignition. During the CS, Al particle was nitrided initially from the surface through a direct nitridation reaction (Equation 3.1), leading to a rapid increase of temperature. In this case, the un-nitrided Al particle would melt and formed a liquid Al core. Due to the low solubility of nitrogen in Al core at this temperature, the direct nitridation and supersaturation of Al core with nitrogen led to the multiple nucleation of AlN embryos on the surface of the Al core [18]. Therefore, many AlN islands were formed and covered on the surface of the particle. With the temperature increase, NH4 Cl dissociated into HCl(g) and NH3 (g) (Equation 3.2), which was an endothermic reaction and absorbed sufficient heat from the mixture to disturb the direct nitridation of

Al particle

Nucleation

Al core

Growth

Figure 3.12 Growth model of the flower-like AlN microstructure. (Reproduced with permission from Ref. [7c]. Copyright © 2011, TTP INC.)

89

90

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies

Al core. Hence, the combustion wave propagation was retarded. Then, various gaseous species such as HCl(g), NH3 (g), and N2 (g) infiltrated into the particle and reacted with the liquid Al core through the chlorination–nitridation mechanism (Equations 3.3 and 3.4), resulting in the nitridation rate of Al core slowing down. Hence, the transportation and supersaturation degree of the vapor-like AlN embryos decreased. In this condition, the embryos preferentially deposited on the as-formed AlN islands and sprouted out epitaxially via a self-catalytic mechanism [19]. With the evolution of the CS process, the supersaturation degree of the AlN embryos was gradually decreased from inside to outside of the particle due to the increase of the diffusion distance from the inner Al core, leading to the gradual attenuation of the 1D structure. Therefore, flower-like AlN microstructure with cone-like petals was formed in the final product.

3.5 Combustion Synthesis of AlN Porous-Shell Hollow Spheres

Hollow micro/nanospheres of inorganic materials (such as TiO2 and SiO2 ) are a new class of materials and have attracted immense interest owing to their unique properties and potential applications in catalysis, light filter, waste removal, energy storage, and microvessels for drug delivery [20]. However, few literatures have reported the preparation of AlN hollow micro/nanostructures. Ma et al. prepared polycrystalline AlN hollow nanospheres through a self-template synthesis approach with Al nanoparticles precursor [21]. Zhang et al. fabricated AlN hollow nanospheres by a chemical vapor deposition reaction between AlCl3 and NH3 [22]. To the best of our knowledge, the preparation of AlN hollow micro/nanostructures via a CS route has not been reported. Herein, we demonstrated that AlN hollow spheres with porous-shell could be fabricated by a facile CS process. 3.5.1 Experimental Methods of Approach

The starting materials were Al powders (spherical particles, 2–3 μm) with the oxygen content of about 0.35 wt%, AlN powders (hollow structures, self-prepared by the same procedure as this work) as diluent, and polytetrafluoroethylene (PTFE, (C2 F4 )n ) powders. The morphology of spherical Al powders is shown in Figure 3.13. In a typical procedure, Al, AlN, and PTFE powders were mixed lightly in mortar with a weight ratio of 28 : 64 : 8, and the molar ratio of Al and AlN powders were fixed as to 4 : 6. After sieving to disperse any large agglomerates, 20 g of the powder mixture was transferred into a porous graphite crucible and ignited under a nitrogen gas pressure of 0.2 MPa. The characterization of the products was similar as the description in Section 3.2. Furthermore, thermogravimetry-differential scanning calorimetry (TG-DSC, SDT Q600) of the reactants was carried out from 50 to 1300 ∘ C in N2 atmosphere.

3.5 Combustion Synthesis of AlN Porous-Shell Hollow Spheres

5 μm Figure 3.13 SEM image of the starting Al powders.

3.5.2 Results and Discussion

XRD pattern indicated that the as-synthesized product was pure hexagonal wurtzite structure of AlN crystal [7d]. No characteristic peaks of impurities were detected in the pattern. The sharp diffraction peaks indicated the good crystallinity of the product. Figure 3.14a,b shows FESEM images of the as-synthesized AlN hollow spheres. The diameter of these spheres is about 2–3 μm, similar as the diameter of the Al powders. This indicated that the Al powders acted as a template for the growth of AlN hollow spheres. Interestingly, some spheres exhibit broken sites and expose their hollow interiors, which provide direct evidence that the AlN spheres have a hollow structure, as shown in the inset of Figure 3.14b. A FESEM image of a single hollow sphere is shown in Figure 3.14c. It can be clearly observed that the shell of hollow sphere is porous and composed of a large number of nanoparticles. These nanoparticles are irregular and their sizes are in the range of 30–120 nm, and the thickness of the shells is about 120 nm (Figure 3.14d). Figure 3.15 shows the TG-DSC result of the reactants with the weight ratio as the typical procedure. The minimum endothermic peak at 322 ∘ C was the phase change of PTFE from solid to liquid. An abrupt weight loss of about 8% at the temperature of 500–600 ∘ C could be attributed to the decomposition of PTFE. However, no corresponding endothermic peak was detected and an exothermic peak at 583 ∘ C was observed. This should be caused by the reaction between PTFE and the oxide shell (Al2 O3 ) of the Al particles [23]. The reaction can be expressed as Equation 3.7: 2Al2 O3 (s) + 3(C2 F4 )� (g) → 4nAlF3 (s) + 6CO(g)

(3.7)

91

92

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies (a)

(b)

500 nm

S4800 1.5kV 4.2mm x 5.00k SE(U)

2 μm

S4800 1.5kV 4.2mm x 13.0k SE(U)

(d)

(c)

0 12

S4800 1.5kV 4.2mm x 30.0k SE(U)

1 μm

500 nm

Figure 3.14 FESEM images of (a, b) AlN porous-shell hollow spheres with different magnifications (inset of (b): a hollow sphere with broken shell), (c) a single hollow

nm

S4800 1.5kV 4.2mm x 100k SE(U)

100 nm

sphere, and (d) the surface of the spheres. (Reproduced with permission from Ref. [7d]. Copyright © 2013, Elsevier Limited.)

which is a strong exothermic reaction with ΔH = −1373 kJ mol−1 , and therefore the endothermic decomposition of PTFE was compensated. Peak at 658 ∘ C corresponded to the melting point of Al particles. Then, the weight gain started at 820 ∘ C because of the nitridation of the Al particles. A strong exothermic peak at about 1025 ∘ C indicated the combustion could be triggered. On the basis of the experimental results, the effect of PTFE on the formation of AlN porous-shell hollow spheres can be explained. Firstly, nitridation of the Al particle was activated in the presence of PTFE because it can peel off the Al2 O3 shell surrounding the Al particle (Equation 3.7). Then, the nitridation occurred rapidly at the surface of the Al particle via a direct nitridation pathway (see Equation 3.1), which was a strong exothermic reaction (ΔH = −657 kJ mol−1 ) and emitting a large amount of heat to melt the Al core. This was in agreement with the DSC (differential scanning calorimetry) result in Figure 3.15. Due to the low solubility of nitrogen in Al particle, the direct nitridation and supersaturation of Al particle with nitrogen led to the multiple nucleation of AlN crystals on the surface of the Al particle [18]. Therefore, an AlN porous-shell surrounding the molten Al core was formed. The addition of PTFE to starting materials offered an alternative reaction pathway for the nitridation of the outwardly diffused Al atoms, which involved spontaneous fluorination–nitridation sequences [23a].

3.6 130 exo

125

2 0

120

−2

115

−4

110

−6

105

−8

100

−10

DSC (mW g−1)

TG (%)

Summary and Conclusions

−12

95

−14

90

−16 200

400

600

800

1000

1200

Temperature (°C)

Figure 3.15 The TG-DSC curves of the reactants with the weight ratio as the typical procedure. (Reproduced with permission from Ref. [7d]. Copyright © 2013, Elsevier Limited.)

The encountered reactions can be described according to the following reactions along with the Equations 3.7 and 3.1: 4�Al(l) + 3(C2 F4 )� (g) → 4�AlF3 (s) + 6�C(s)

(3.8)

2AlF3 (s) + N2 (g) → 2AlN(s) + 3F2 (g)

(3.9)

2Al(l) + 3F2 (g) → 2AlF3 (s)

(3.10)

As a result, the inner Al core was gradually evacuated, and finally the AlN porousshell hollow spheres were obtained. Actually, many gray powders were observed in the combustion chamber after the reactions. XRD result (not shown here) indicated that these powders were hexagonal AlN phase, and carbon phase could not be detected, which might be due to the limitations of the measurement technique.

3.6 Summary and Conclusions

In summary, 1D AlN nanowhiskers, 3D flower-like AlN microstructure, and AlN porous-shell hollow spheres with uniform morphologies have been successfully fabricated by CS methods. The morphologies of products can be controlled by manipulating the combustion parameters. The present processes for synthesis of the AlN micro/nanostructures are facile, productive, reproducible, and energy saving. The as-synthesized AlN nanowhiskers have been applied to produce fillers as reinforcement for electrical packaging. However, in order to realize

93

94

3 Combustion Synthesis of Aluminum Nitride (AlN) Powders with Controlled Grain Morphologies

the wider applications of the combustion-synthesized AlN products in other structural/functional materials and devices, more work needs to be done, such as promoting the yield and purity, as well as achieving the morphological diversity. References 1. Sheppard, L.M. (1990) Am. Ceram. Soc. 2.

3.

4.

5.

6.

7.

Bull., 69, 1801–1812. (a) Mussler, B.H. (2000) Am. Ceram. Soc. Bull., 79, 45–47; (b) Tondare, V.N., Balasubramanian, C., Shende, S.V., Joag, D.S., Godbole, V.P., Bhoraskar, S.V., and Bhadbhade, M. (2002) Appl. Phys. Lett., 80, 4813–4815. Yuan, J.K., Li, W.N., Gomez, S., and Sui, S.L. (2005) J. Am. Chem. Soc., 127, 14184–141845. (a) Lei, W.W., Liu, D., Zhang, J., Zhu, P.W., Cui, Q.L., and Zou, G.T. (2009) Cryst. Growth Des., 9, 1489–1493; (b) Tang, Y.B., Cong, H.T., Wang, Z.M., and Cheng, H.M. (2005) Chem. Phys. Lett., 416, 171–175; (c) Wu, Q., Hu, Z., Wang, X.Z., Chen, Y., and Lu, Y.N. (2003) J. Phys. Chem. B, 107, 9726–9729. (a) Merzhanov, A.G. (1995) Ceram. Int., 21, 371–379; (b) Tanihata, K. and Miyamoto, Y. (1998) Int. J. SHS., 7, 209–217; (c) Zakorzhevskii, V.V. and Borovinskaya, I.P. (1998) Int. J. SHS., 7, 199–208; (d) Zakorzhevskii, V.V., Borovinskaya, I.P., and Sachkova, N.V. (2002) Inorg. Mater., 38, 1131–1140. (a) Moya, J.S., Iglesias, J.E., Limpo, J., Escriña, J.A., Makhonin, N.S., and Rodríguez, M.A. (1997) Acta Mater., 45, 3089–3094; (b) Jiang, G.J., Zhuang, H.R., Zhang, J., Ruan, M.L., Li, W.L., Wu, F.Y., and Zhang, B.L. (2000) J. Mater. Sci., 35, 57–62; (c) Jiang, G.J., Zhuang, H.R., Zhang, J., Ruan, M.L., Li, W.L., Wu, F.Y., and Zhang, B.L. (2000) J. Mater. Sci., 35, 63–69; (d) Fu, R., Chen, K., Agathopoulos, S., and Ferreira, J.M.F. (2006) J. Cryst. Growth, 296, 97–103. (a) Shi, Z.Q., Radwan, M., Kirihara, S., Miyamoto, Y., and Jin, Z.H. (2009) Ceram. Int., 35, 2727–2733; (b) Shi, Z.Q., Radwan, M., Kirihara, S., Miyamoto, Y., and Jin, Z.H. (2009) J. Alloys Compd., 476, 360–365; (c)

8.

9. 10.

11.

12. 13.

14.

15. 16. 17. 18.

Shi, Z.Q., Yang, W.L., Qiao, G.J., Yang, J.F., Miyamoto, Y., and Jin, Z.H. (2011) Mater. Sci. Forum, 695, 413–416. (d) Shi, Z.Q., Yang, W.L., Kang, Y.F., Qiao, G.J., and Jin, Z.H. (2013) Ceram. Int., 39, 4663–4667; (e) Shi, Z.Q., Radwan, M., Kirihara, S., Miyamoto, Y., and Jin, Z.H. (2009) Appl. Phys. Lett., 95, 224104. (a) Lin, C.N. and Chung, S.L. (2001) J. Mater. Res., 16, 2200–2208; (b) Rosenband, V. and Gany, A. (2004) J. Mater. Process. Technol., 147, 197–203; (c) Chen, H., Cao, Y.G., and Xiang, X.W. (2001) J. Cryst. Growth, 224, 187–189; (d) Wang, H.B., Northwood, D.O., Han, J.C., and Du, S.Y. (2006) J. Mater. Sci., 41, 1697–1703. Sakurai, T., Yamada, O., and Miyamoto, Y. (2006) Mater. Sci. Eng. A, 415, 40–44. Chen, K.X., Ge, C.C., Li, J.T., and Cao, W.B. (1999) J. Mater. Res., 14, 1944–1948. Radwan, M., Bahgat, M., and El-Geassy, A.A. (2006) J. Eur. Ceram. Soc., 26, 2485–2488. Radwan, M. and Bahgat, M. (2006) J. Nanosci. Nanotechnol., 6, 558–561. Campbell, W.B. (1970) in Whisker Technology, Chapter 2 (ed. A.P. Levitt), John Wiley & Sons, Inc., New York, pp. 15–46. (a) Stankovich, S., Dikin, D.A., Dommett, G.H.B., Kohlhaas, K.M., Zimney, E.J., Stach, E.A., Piner, R.D., Nguyen, S.T., and Ruoff, R.S. (2006) Nature, 442, 282–286; (b) Sundstrom, D.W. and Lee, Y.D. (1972) J. Appl. Polym. Sci., 16, 3159–3167. Bruggeman, D.G.A. (1935) Ann. Phys., 416, 636–664. Agari, Y. and Uno, T. (1986) J. Appl. Polym. Sci., 32, 5705–5712. Takacs, L. (2002) Prog. Mater. Sci., 47, 355–414. Yu, L.S., Hu, Z., Ma, Y.W., Huo, K.F., Chen, Y., Sang, H., Lin, W.W., and

References

Lu, Y.N. (2007) Diamond Relat. Mater., 16, 1636–1642. 19. Haber, J.A., Gibbons, P.C., and Buhro, W.E. (1997) J. Am. Chem. Soc., 119, 5455–5456. 20. (a) Zhang, Q., Wang, W.S., Goebl, J., and Yin, Y.D. (2009) Nano Today, 4, 494–507; (b) Lou, X.W., Archer, L.A., and Yang, Z.C. (2008) Adv. Mater., 20, 3987–4019. 21. Ma, Y.W., Huo, K.F., Wu, Q., Lu, Y.N., Hu, Y.M., Hu, Z., and Chen, Y. (2006) J. Mater. Chem., 16, 2834–2838.

22. Zhang, F., Wu, Q., Xiao, P., Ma, Y.W.,

Hu, Y.M., Wang, X.Z., Wang, C.Y., and Hu, Z. (2008) J. Phys. Chem. C, 112, 11331–11335. 23. (a) Lee, T.H., Nersisyan, H.H., Jeong, H.G., Lee, K.H., Noh, J.S., and Lee, J.H. (2011) Chem. Eng. J., 174, 461–466; (b) Osborne, D.T. and Pantoya, M.L. (2007) Combust. Sci. Technol., 179, 1467–1480.

95

97

4 Combustion Synthesis and Spark Plasma Sintering of �-SiAlON Xuemei Yi, Tomohiro Akiyama, and Kazuya Kurokawa

4.1 Introduction 4.1.1 �-SiAlON

SiAlON is a general name for a large family of the so-called ceramic alloys based on silicon nitride. Initially, they were discovered in the early 1970s [1] and have been developed actively since. SiAlON ceramics became one of the commercially produced high-tech ceramic materials. β-SiAlON is based on the atomic arrangement existing in β-Si3 N4 . In this material, Si is substituted by Al with corresponding replacement of N by O. In this way, up to two-thirds of the silicon in β-Si3 N4 can be replaced by Al without causing a change in structure. The chemical replacement is one of changing Si–N bonds for Al–O bonds. The bond lengths are about the same for the two cases, but the Al–O bond strength is significantly higher than that of Si–N. In SiAlON, the Al is coordinated as AlO4 and not as AlO6 as in alumina (Al2 O3 ). Therefore, in β-SiAlON, the bond strength is 50% stronger than in Al2 O3 . Thus, SiAlONs intrinsically have better properties than both Si3 N4 and Al2 O3 . β-SiAlON has the general formula Si6−z Alz Oz N8−z , where z varies between 0 and 4.2 [2]. β-SiAlON materials have been attracting considerable attention on account of being suitable for high-temperature applications, owing to their excellent mechanical and thermal properties, superior chemical stability, and a conspicuous thermal-shock resistance. Consequently, they are being extensively used as high-temperature engineering ceramics in cutting tools and abrasive materials [3]. The use of Eu2+ -doped β-SiAlON phosphors for down-conversion luminescent materials in white light-emitting diodes (LEDs) has been reported [4–6], which indicates that β-SiAlON-based materials have a potential for use as functional materials in numerous fields.

Nitride Ceramics: Combustion Synthesis, Properties, and Applications, First Edition. Edited by Alexander A. Gromov and Liudmila N. Chukhlomina. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

98

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

4.1.2 Combustion Synthesis (CS)

The concept of preparing compounds through the propagation of a wave in a gasless combustion process was first proposed more than three decades ago and has since found practical applications in the synthesis of a large number of metallic and ceramic phases, such as nitrides, hydrides, oxides, and many other functional materials. The underlying thermodynamic principle for this process is a large exothermic reaction enthalpy capable of sustaining the reaction in the form of a combustion wave moving through the reactant mixture. When used for the preparation of materials, such a process is termed self-propagating high-temperature synthesis (SHS) or simply combustion synthesis (CS). Interest in this process as an alternative method for the preparation of materials has been largely motivated by the potential for low-cost production and by assertions of uniqueness in the properties of materials prepared by this process. For it to be self-sustaining, the SHS process must, perforce, be associated with high-temperature reactions. An important parameter in this regard is the adiabatic temperature of combustion, T ad . This thermodynamic parameter is the temperature to which the product is raised under adiabatic conditions as a consequence of the evolution of heat from the reaction. Reactions between particulate materials in a self-propagating mode present an attractive practical alternative to conventional methods of materials preparation for a variety of reasons [7, 8]. These include: 1) 2) 3) 4)

The simplicity of the process and its relatively low energy requirement. The higher purity of products obtained by this method. The possibility of obtaining complex or metastable phases. The possibility of simultaneous formation and densification of the desired materials.

4.1.3 Spark Plasma Sintering (SPS)

Spark plasma sintering (SPS), which is a pressure-assisted pulsed direct current sintering technique, is the most employed to consolidate powders of very different natures. Compared to conventional sintering techniques such as pressureless sintering (PS), hot pressing (HP), or hot isostatic pressing (HIP), SPS allows much faster heating rates and shorter sintering times; together with commonly lower sintering temperatures. This technique extraordinarily enhances the sinterability of most of the materials and extends the possibilities for developing new advanced materials and tailoring their properties. Until now, SPS has successfully covered a wide spectrum of materials, from metals, and alloys to ceramics, including different kinds of composites and coatings, as well as graded materials. During the first decade, the acquisition of SPS equipment and the prompt fabrication of materials to compare well-known materials processed by conventional sintering techniques

4.2 CS of High-Purity β-SiAlON and Densification by SPS

was the pattern. Therefore, SPS is regarded as a rapid and effective sintering technique for a variety of materials [9]. This chapter is focused on CS and SPS of high-purity β-SiAlON, which includes the reaction mechanism of CS; the physical properties of Vickers hardness and thermal properties; and the corrosion resistance properties in different atmospheres.

4.2 CS of High-Purity �-SiAlON and Densification by SPS 4.2.1 Reaction Mechanisms

It is known that an understanding of the temperature change during the CS reaction and the combustion rate is necessary to design a large-scale CS β-SiAlON production. However, the combination of high reaction temperatures and high propagation rates makes it difficult to investigate experimentally the intermediate reactions. The lack of knowledge of the reaction mechanism makes it difficult to anticipate the behavior of the reactions under conditions different from those employed experimentally. Here, one simple method was proposed: heating the raw materials slowly to obtain the temperature curves, using which we can analyze the transition reaction during CS process, which actually works as a high-temperature and high-pressure Differential Scanning Calorimetry (DSC) [10]. Figure 4.1 gives the schematic diagram of the experimental apparatus. The device can be used to uniformly heat up to 1500 ∘ C under a high gas pressure up to 1.0 MPa. Two R-type thermocouples were used to control and to measure the temperatures inside the reactor. Two W–Re thermocouples placed inside protective magnesia sheaths were introduced into the center of the samples to measure the temperature changes during the CS process. The same weight reactants (Si, Al, and SiO2 ) and standard sample Al2 O3 were charged into the carbon crucible. Between these two powders, some carbon sponge was placed for insulation (see Figure 4.1b). Then the carbon crucible and samples were uniformly heated to 1400 ∘ C using the graphite heater with a heating rate of 20 ∘ C min−1 . Figure 4.2a shows the temperature curve of CS of β-SiAlON from 3.5 g raw mixtures of Si, Al, and SiO2 under 1.0 MPa nitrogen pressure. There are two big exothermic reaction peaks: one peak locates at ∼800 ∘ C and the other locates at ∼1350 ∘ C. The highest temperature reached was 1932 ∘ C for the last big exothermic reaction. From these, it can be deduced that the main reaction should start at ∼1350 ∘ C. Besides these two peaks, there are also several small exothermic or endothermic reaction peaks. The change of the compositions was checked using X-ray diffraction (XRD) to investigate what had happened during these reactions when heated to different temperatures (see the points 1–4 in Figure 4.2b). These points represent the raw materials before heating and when heated to 1000, 1200, and 1400 ∘ C, respectively. From this, we can see that the

99

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

Top view N2: 1 MPa

Thermocouples

P N2 outlet Sample Carbon heater

N2 inlet

Carbon sponge insulation

Al2O3

(b)

Side view

Protective magnesia sheath

Carbon crucible

(a) Figure 4.1 (a) The schematic diagram of the SHS apparatus and (b) carbon crucible. (Reproduced with permission from Ref. [10] © 2013, Elsevier.)

2000 Sample Heater

1600 Temperature (°C)

100

Si, Al, SiO2 Si, SiO2, AlN

1200

Si, SiO2, AlN, Al2O3 Si, β-SiAlON

800 400 0

0

10

20

30

40

50

60

70

80

Time (min) Figure 4.2 Temperature curve of combustion synthesis of β-SiAlON from the mixture of Si, Al, and SiO2 when heating to 1400 ∘ C under 1 MPa nitrogen pressure. (Reproduced with permission from Ref. [10] © 2013, Elsevier.)

raw reactants (see Figure 4.2a point 1) are Si, Al, and SiO2 . When these reactants were heated to 1000 ∘ C (see Figure 4.2a point 2), XRD results showed that Al peaks disappeared, and AlN was detected. When they were heated to 1200 ∘ C (see Figure 4.2a point 3), besides Si, AlN, and SiO2 , a little of Al2 O3 peaks were also detected. When the reactants were finally heated to 1400 ∘ C, β-Si5 AlON7 was synthesized, coexisting with a little of unreacted Si.

4.2 CS of High-Purity β-SiAlON and Densification by SPS

According to the above analyses, we can propose the general reaction procedure in the CS of β-SiAlON. Reactions Al (s) = Al (l) and 2Al + N2 = AlN are simultaneous, which corresponded to the first big exothermic peak located at ∼800 ∘ C. Reactions Si + N2 = Si3 N4 and Si3 N4 + AlN + Al2 O3 + SiO2 → β-SiAlON occurred synchronously, which corresponded to the last main exothermic peak located at ∼1350 ∘ C. A detailed in-depth study was performed on the other small exothermic and endothermic peaks. The first endothermic reaction was found at ∼570 ∘ C, which should be attributed to the eutectic melting of Al–Si (577 ∘ C). Another exothermic peak was found between 1000 and 1200 ∘ C. According to the XRD results, when the reactants were heated to 1200 ∘ C, Al2 O3 peaks were detected. Therefore, this exothermic peak should be due to the reaction of Al + SiO2 → Si + Al2 O3 + Q, where Al was from the melted and unreacted Al during the aforementioned reaction of Al and N2 . Finally, we conclude the reaction process of CS of β-SiAlON using the raw reactants of Si, Al, and SiO2 . This reaction process indicates that with the temperature increasing, first, Al melts and reacts with nitrogen gas to get AlN, the unreacted Al then reacts with SiO2 to get Si and Al2 O3 . It is widely considered that the reaction of Si + N2 → Si3 N4 should be the fuse for synthesizing β-SiAlON, namely the main reaction should start at the melting point of Si (1414 ∘ C). But in this study, we found that the main reaction started at ∼1350 ∘ C, which showed a melting point lower than that of Si. Therefore, this should be due to some kinds of compounds coexisting, thereby affecting the trigger temperature of the main reaction. At the present time, we still suppose that at last, Si melts and reacts with nitrogen gas to get Si3 N4 ; synchronously, β-SiAlON is synthesized. 4.2.2 Dense �-SiAlON by CS and SPS

As we know, the purity of the β-SiAlON powders obtained by CS is low due to the existence of an unreacted Si residue. The presence of the unreacted Si was attributed to the melting and subsequent conglomeration caused by the extremely high reaction temperature and the high reaction rate of the CS. The use of diluents such as α/β-Si3 N4 or β-SiAlON would help lower the combustion temperature as well as decrease the velocity of the combustion front. This would also improve the conversion rate of reactants and increase the purity of the β-SiAlON products. Therefore, the purpose of this part is to synthesize high-purity β-SiAlON (z = 1–4) powders by CS using β-SiAlON as a diluent, and to densify the combustion-synthesized β-SiAlON powders through the SPS technique. 4.2.2.1 Combustion Synthesis of �-SiAlON Powder

The raw materials of Si, Al, and SiO2 were weighed according to the mole ratios of Equation 4.1 for different z values, and β-SiAlON (z = 1) was added as the diluent in the raw mixtures. The ratio in which the starting materials were used was precisely adjusted to compensate for the lack of Al and O in the diluent used for

101

102

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

the synthesis of β-SiAlON (z = 2–4). The powder was subjected to CS at a low nitrogen pressure of 1 MPa (nitrogen purity: 99.999%). (6 − 1.5�)Si + �Al + 0.5�SiO2 + (4 − 0.5�)N2 → β − Si6−� Al� O� N8−� (� = 1, 2, 3, and 4)

(4.1)

The starting powder mixture, after milling, was loosely charged into a cylindrical graphite crucible (diameter: 40 mm; length: 65 mm) with several vents, on both the surface and bottom, for infiltrating nitrogen into all areas of the raw materials. The raw materials were usually processed in batches of ∼30 g. The combustion reaction was ignited by passing a current of 60 A at 35 V for approximately 10 s through a carbon foil, which helped burn a small quantity of the Al powder placed on top of the raw material mixture. Figure 4.3 gives the sketch drawing of CS apparatus. 4.2.2.2 Spark Plasma Sintering of CSed Powders

The combustion-synthesized (CSed) powders were first subjected to planetary ball milling for 60 min to eliminate the agglomerates formed during the CS process, and to enhance the sinterability of the powders. Then, the milled powder was compacted into a carbon die (inner diameter: 10 mm) and sintered using an SPS system (Sumitomo Coal Mining Co. Ltd., Tokyo, Japan) under vacuum conditions of lower than 4 Pa at a compressive stress of 50 MPa, shown in Figure 4.4. The resulting compacts were first heated to 600 ∘ C for 5 min, and then were heated to 1600 ∘ C at a rate of 30 ∘ C min−1 . The compacts were maintained at this temperature for 12 min before the power was turned off. Temperature monitoring during sintering between 600 and 1600 ∘ C was carried out using an optical pyrometer, focused on a hole on the surface of the graphite die.

TC

Pressure gauge 1

4

TC

Vacuum pump

N2 bottle

3

2

1 Carbon foil 2 Electric pole 3 Sample

5

Cooling water

4 Ignition agent (Al) 5 Graphite crucible

TC

High-pressure stainless reactor Figure 4.3

TC: thermocouple

~

Sketch drawing of the combustion synthesis apparatus.

4.2 CS of High-Purity β-SiAlON and Densification by SPS

P

Glass window

Powder

Carbon die

Carbon foil

Carbon punch Carbon plate

Vacuum chamber

A hole for temperature observation through the glass window Figure 4.4

Sintering DC pulse generator

Punch electrode

P

Sketch drawing of the spark plasma sintering apparatus.

The lattice parameters of the hexagonal β-SiAlON phase were calculated from the XRD peaks of the products by employing the least-squares method. The lattice parameter a was calculated from (1 0 0), (1 1 0), (2 0 0), (2 1 0), (3 0 0), and (3 1 0) peaks of β-SiAlON, and the lattice parameter c was calculated from the (1 0 1) peak using the calculated a value. The bulk density of the spark-plasma-sintered (SPSed) specimens was measured according to the Archimedean principle, using distilled water as the medium. The detailed experimental procedure was published in paper [11]. 4.2.2.3 Characterization of CS–SPSed �-SiAlON

The XRD patterns of CSed powder synthesized with the optimum quantity of diluent and of CS–SPSed fractures are shown in Figure 4.5. The CS reaction did not self-sustain when excessive diluent was added due to too low adiabatic temperature (T ad ) < 1800 K, according to Munir and Holt [8]. Pure β-SiAlON with z value of 4 was not obtained by SPS due to the low purity and short z value of CSed powder. The CSed powder showed trace amounts of unreacted Si except the dominant phase β-SiAlON in all of the products with different z values; in the product with z = 4, the presence of Al2 O3 was also detected. According to Dickon H.L. Ng et al., Si3 N4 , AlN, and Al2 O3 are the intermediate solid phases that are simultaneously consumed during the formation of β-SiAlON. Among these, the formation of Si3 N4 would be most important because it can act as a crystal seed

103

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

: β-SiAlON

Intensity (a.u.)

Z4

: Si

: Al2O3

Z3 Z2 Z1 10

20

30

40

50

60

Diffraction angle (2θ )

(a)

: β-SiAlON S3 Intensity (a.u.)

104

S2

S1 10 (b)

20

30 40 Diffraction angle, 2θ (°)

Figure 4.5 XRD patterns of (a) combustionsynthesized product powders with β-SiAlON (z = 1) as diluent (Z1: z = 1 (45 mass% diluent); Z2: z = 2 (40 mass% diluent); Z3: z = 3 (35 mass% diluent); Z4: z = 4 (30 mass% diluent) and (b) spark-plasma-sintered

50

60

β-SiAlON from CSed powders (S1: z = 1; S2: z = 2; S3: z = 3) (S4 was not obtained by SPS due to the low purity of CSed powder). (Reproduced with permission from Ref. [11] © 2010, Elsevier.)

for the formation of β-SiAlON. From Equation 4.1, we can see that an increase in the z value of β-SiAlON translates to an increase in the Al and SiO2 contents of the raw materials, and a decrease in the Si content. Thus, we can predict that for higher z values, the formation of Si3 N4 involves a two-step process. First, Si is deoxidized from SiO2 by Al, and then, the deoxidized Si is made to react with N2 . The CS process is very rapid; hence, it would culminate before some of this Si can react with N2 to form Si3 N4 . Thus, at higher z values, both Si and Al2 O3 will be left behind as residues in the product. After SPS, the Si peaks disappeared and only peaks due to β-SiAlON were observed. One possible reason for the disappearance of Si after SPS would be the vaporization or dissolution of Si into the solid during the high-temperature and high-pressure sintering process.

4.2 CS of High-Purity β-SiAlON and Densification by SPS

Figure 4.6 displays the scanning electron microscope (SEM) images of the CSed β-SiAlON (z = 1–4) powders and fracture surfaces of the SPSed β-SiAlON (z = 1–3). From Figure 4.6a, some different shapes of β-SiAlON crystals, rod-like morphology, tiny particles, and some whiskers can be seen. Because the CS process normally completes in only several or tens of seconds, it is difficult to perform in situ observations. The growth of rod-like crystals and whiskers has been reported in a vapor–liquid–solid (VLS) mechanism. The typical rod-like morphology of β-SiAlON can be clearly seen when z = 1, and with increasing z value, the thin and long rod-like crystals seem to get thicker and shorter, and simultaneously the size gets larger, which should be attributed to the expansion of crystal lattice due to more Si–N bonds replaced by Al–O bonds at higher z values. From Figure 4.6b, we can see the SPSed products were dense solid and the grain size increased obviously with increasing z value. Figure 4.7 shows a sequence of images (a, b, and c) and the corresponding selected area electron diffraction (SAED) patterns (a-1, b-1, and c-1) of different morphologies of the CSed powder. Image (a) presents the polycrystalline area of the particles, whose size shows 2–3 μm. The diffraction rings of the corresponding SAED patterns (a-1) were indexed as shown, which is certificated to be β-SiAlON. Image (b) and SAED (b-1) show a rod and the corresponding diffraction pattern. The rod has a diameter of ∼500 nm and the diffraction patterns were indexed as shown, as seen from the [−316] zone axis, which is also certificated to be β-SiAlON. Image (c) shows a single crystal fiber with the width of ∼50 nm. There is an interesting phenomenon that the inside of this fiber is crystal; however, the two sides are amorphous. The SAED (c-1) of the crystal area on this fiber was indexed as shown, as seen from the [012] zone axis. Further systematic study is needed to understand the growth mechanism and growth direction of these SiAlON crystals. Pure β-SiAlONs were obtained through SPS, though the CSed powders were not pure β-SiAlON products. The bulk density was measured by the Archimedes principle, while theoretical density values were obtained from literature. The relative densities of the SPSed β-SiAlON with z = 1–3 were more than 99% of the corresponding theoretical density. Table 4.1 lists the lattice parameters and the calculated z values of β-SiAlON before and after SPS. Ekstrom studied the lattice parameters of β-SiAlON as a function of composition and proposed the following equations: � = 0.7603 + 0.00296� (nm)

(4.2)

� = 0.2907 + 0.00255� (nm)

(4.3)

In our study, the lattice parameters a and c were calculated from the XRD data while z was calculated by using the lattice parameter a. Evidently, the calculated z values were in agreement with the expected z values in the raw mixture, except for the product with z = 4. Unfortunately, the z value did not increase to 4 even after the CSed β-SiAlON with z = 3 was used as a diluent to synthesize β-SiAlON with z = 4. From these observations, it could be inferred that the stoichiometic

105

106

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

Z1

Z2

1 μm

1 μm Z3

Z4

1 μm

1 μm

(a) S1

S2

1 μm

1 μm

S3

1 μm (b)

Figure 4.6 SEM images of (a) combustionsynthesized β-SiAlON powders (Z1: z = 1, Z2: z = 2, Z3: z = 3) and (b) fracture surfaces of the spark-plasma-sintered β-SiAlON

from CSed powders (S1: z = 1; S2: z = 2; S3: z = 3). (Reproduced with permission from Ref. [11, 12] © 2010, 2014, Elsevier.)

4.2 CS of High-Purity β-SiAlON and Densification by SPS

(a-1)

(a)

(210) (101) (200) (110) (100)

5 μm

(b-1)

(b)

(130) (-13-1)

(201)

100 nm

[-316] (c-1)

(c)

(100) (1-21)

20 nm

Figure 4.7 TEM micrographs of combustion synthesized β-Si5 AlON7 (z = 1) powder with the corresponding selected area electron diffractions (SAEDs) analysis. (a) particles,

(0-21)

[012]

(a-1) diffraction rings of the particles; (b) rod, (b-1) diffraction patterns of the rod; and (c) fibers, and (c-1) diffraction patterns of the fiber.

range of CSed β-SiAlON would become narrow when the nitrogen pressure is decreased. In other words, β-SiAlON with higher z values cannot be CSed under low nitrogen pressures. The z values calculated after SPS showed a slight variation when compared to the previous values. This could probably be attributed to the dissolution of unreacted Si into the solid and the rearrangement of the elements during sintering. Moreover, the z value of the diluent β-SiAlON (z = 1) has also been confirmed to affect the final z values of SPSed β-SiAlON.

107

108

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

Table 4.1 The lattice parameters and calculated z values of β-SiAlON before and after spark plasma sintering. (Reproduced with permission from Ref. [11] © 2010, Elsevier.) Aimed z value

Before SPS Lattice parameter (Å) a C

1 2 3 4

7.6355 7.6628 7.6916 7.7021

After SPS Calculated za

2.9201 2.9384 2.9673 2.9939

1.10 2.02 2.99 3.34

Lattice parameter (Å) A c

7.6373 7.6617 7.6889 —

2.9099 2.9303 2.9759 —

Calculated za

1.14 1.98 2.90 —

a) The value was calculated from lattice parameter a according to Ref. [13] with permission from Springer.

4.3 Physical Properties of CS-SPSed �-SiAlON 4.3.1 Vickers Hardness

Vickers hardness as a principal parameter for the mechanical characterization of materials has been commonly used as a technique to measure the mechanical properties of materials, but the microhardness commonly decreases with applied load, which is known as the indentation size effect (ISE). The Vickers hardness of polished CS–SPSed β-SiAlON compacts was measured using a Vickers microhardness tester with a diamond indenter of regular pyramid with an opposite angle of 136∘ . The experiments were performed under the loads between 0.981 N (0.1 kg) and 19.614 N (2.0 kg) at room temperature. The dwell time for each load was 20 s. An average of at least five readings at different locations of the specimen surfaces was taken for each specimen. The Vickers hardness (H v , GPa) was calculated according to Equation 4.4 �v = 18.19

� �2

(4.4)

where F is the applied test load (N) and d is the average diagonal distance of the indented impressions (mm). Figure 4.8 shows the measured H v as a function of indentation load for the CS–SPSed β-SiAlON (z = 1–3). The variation of H v with applied indentation test load for all samples shows that H v decreases with the increase of applied load at low load region. It reaches a saturation value at higher loads when F ≥ 4.903 N for z = 1 and 2. However, for z = 3, the ISE is more clear than that of z = 1 and 2. Also, as the figure shows, H v decreases with the increase in z value. At the max applying load of 19.614 N, the Vickers hardness value of 17.6 GPa was obtained when z = 1, and with the increase of z value, it became lower with values of 16.2 and 14.7 GPa when z = 2 and 3, respectively.

4.3

20

z=1 z=2 z=3

19 Vickers hardness (GPa)

Physical Properties of CS-SPSed β-SiAlON

18 17 16 15 14

0

4

8

12

16

20

Indentation load (N) Figure 4.8 Measured Vickers hardness as a function of indentation load for the β-Si6 zAlzOzN8 -z (z = 1–3). (Reproduced with permission from Ref. [14] © 2010, Ceramic Society of Japan.)

The decrease of Vickers hardness with the increase of z value can be mainly attributed to the increase of grain size shown in Figure 4.6b. In addition, the expansion of crystal lattice (shown in Table 4.1) can also play a part in the decrease of Vickers hardness. When Si–N bond is substituted by Al–O, Al–N and Si–O bonds are also formed simultaneously. As we know, the bond energy between different elements gives contribution to hardness and strength of materials. Bond energy gets higher with the decrease of bond length, and covalent bond has higher bond energy than ionic bond. Therefore, by replacing more Si–N with Al–O, the lattice constants, a and c, get larger; correspondingly, the bond energy gets weaker with the increase in z value. In addition, as we know, the chemical bond of most ceramics is a mixed bond between covalent bond and ionic bond. According to the electronegativation difference of elements, the covalent bond ratio of Si–N bond is higher than that of Al–O, Al–N, and Si–O bonds; consequently, bond energy should get weaker with the increase in z value. As the conclusion for CS–SPSed β-SiAlON, the Vickers hardness ranges between 14.7 and 17.6 GPa due to different z values under the indentation load of 19.614 N. 4.3.2 Thermal Conductivity

Specimens with a dimension of 10 mm in diameter and 2–3 mm in thickness were cut from the CS–SPSed discs, and were polished using emery paper until No. 1200. Prior to measurement, a thick layer of colloidal graphite was sputter-coated to the surface of the specimen to enhance absorption of the flash energy. The thermal diffusivity and specific heat capacity were measured by the laser-flash method

109

110

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

Table 4.2 Thermal properties of the β-SiAlONs and β-Si3 N4 at room temperature. (Reproduced with permission from Ref. [12] © 2014, Elsevier.) Samples

Relative density (%)

Specific heat (103 J kg−1 K−1 )

Thermal diffusivity (10−4 m2 s−1 )

Thermal conductivity (W m−1 K−1 ) 28 ∘ C

Z1 Z2 Z3 β-Sialona β-Si3 N4 b

98.8 99.9 99.7 >98.5 Near full density

0.728 0.760 0.722 0.561 0.67

0.0415 0.0354 0.0264 0.0700 0.6723

9.45 8.40 5.86 12.44 150

a) Sintered pressurelessly at 1800 ∘ C for 2 h from Si3 N4 , AlN, and Al2 O3 [16] with permission from AIP Publishing LLC. b) Sintered from β-Si3 N4 , Yb2 O3 , and ZrO2 at 1900 ∘ C for 36 h and subsequent annealed at 1700 ∘ C for 100 h [15] with permission from Elsevier.

(TC-7000, ULVAC Sinku Riko Co., Yokohama, Japan) from room temperature to 800 ∘ C. The thermal diffusivity was analyzed with the t 1/2 method. The bulk density was measured according to the Archimedean principle using distilled water as the medium. All the experiments were carried out under a flowing argon gas atmosphere. The thermal conductivity (K) of β-SiAlONs was determined by following equation: � = ��p �

(4.5)

where � represents the bulk density (g cm−3 ), C p is the specific heat capacity (J g−1 K−1 ), and � is the thermal diffusivity (cm2 s−1 ). Table 4.2 shows the thermal properties of some kinds of β-SiAlONs and β-Si3 N4 at room temperature. β-Si3 N4 sintered at 1900 ∘ C for 36 h and subsequently annealed at 1700 ∘ C for 100 h shows very high thermal diffusivity and thermal conductivity [15]. This was attributed to the reduction of internal defects of the β-Si3 N4 grains with sintering and annealing time as the grains grew. On the contrary, the thermal conductivity of β-SiAlONs is much less by nearly 2 orders of magnitude when compared to that of β-Si3 N4 . The β-SiAlON shows a little higher thermal conductivity of 12.44 W m−1 K−1 reported by Liu et al. [16], but we could not know the exact z value of this sample. For our present products, they showed that both thermal diffusivity and thermal conductivity decrease with an increase in the z value. The highest thermal conductivity of 9.45 W m−1 K−1 was obtained when z = 1 at room temperature. Figure 4.9 represents the temperature dependence of thermal diffusivity, heat capacity, and thermal conductivity of the β-SiAlONs from room temperature to 800 ∘ C. The thermal diffusivity decreases with increasing temperature for all of the samples. And with increase in z value, the thermal diffusivity decreases. Our data show higher than that of sintered β-SiAlON. For all of these CS–SPSed β-SiAlONs, Z2 shows the highest heat capacity, and Z3 shows the lowest data. The thermal conductivity of β-SiAlONs gradually decreases with the increase of

0.05 0.04 0.03 0.02 0.01 200

400

600

800

1000

1200

Thermal conductivity k (W m–1 K–1)

Measurement temperature T (K) 11

Physical Properties of CS-SPSed β-SiAlON

1.4 1.2 1.0 Z1 Z2 Z3

0.8 0.6 200

400 600 800 1000 1200 Measurement temperature T (K)

Z1 Z2 Z3

10 9 8 7 6 5 200

400

600

800

1000

111

1.6

Z1 Z2 Z3

Heat capacity Cp (J g–1 K–1)

Thermal diffusivity α (10–4 m2 s–1)

4.3

1200

Measurement temperature T (K)

Figure 4.9 Temperature dependences of (a) thermal diffusivity, (b) the heat capacity, (c) thermal conductivity, for the β-Si6 -zAlzOzN8 -z (z = 1–3). (Reproduced with permission from Ref. [12] © 2014, Elsevier.)

temperature up to 1000 K, which is attributed to the lattice thermal conduction. However, the thermal conductivity of Z3 increases slightly above 1000 K, which may be attributed to the increased radiation, also known as photon thermal conductivity, with the increase of temperature [17]. From this figure, we can see the thermal conductivity of β-SiAlONs gradually decreases with the increase of the z value under identical temperature conditions. From the literature, it is known that crystallinity; structural discontinuities such as pores, micro-cracking, and glassy phases; and micro-structural features, such as grain size, grain orientation, and atomic mass differences of phases can affect the thermal conductivity of materials. In this study of β-SiAlONs, the solid solution of β-Si3 N4 , where Si4+ and N3− ions get replaced by Al3+ and O2− ions, turns out to be substitutional impurities in the crystal. From the SEM images of these samples, the grain size increases with the increase in z value, but the true density decreases with the increase in z value, although the SPS conditions are the same. It has been demonstrated that the thermal conductivity of β-Si3 N4 is basically governed by the dissolved oxygen in the lattice of Si3 N4 , which causes phonon-defect scattering, thereby lowering thermal conductivity. The vacancies formed in the lattices and the mass difference due to substitution act as the phonon scattering sites,

112

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

thereby reducing the thermal conductivities. Their results revealed that the thermal conductivity at room temperature is independent of grain size, but the internal defect in β-Si3 N4 grains such as point defects and dislocations are the significant factors that affect the conductivity values. Therefore, in the β-Si6−z Alz Oz N8−z system, with z value increasing, more Si4+ and N3− ions get replaced by Al3+ and O2− ions, thereby forming more lattice defects. These defects will cause phonon scattering, and accordingly reduce the thermal conductivity. An increased thermal conductivity of Liu et al. [16] than our value might be mainly attributed to the Y2 O3 addition as a sintering aid and long sintering time (2 h) at high temperatures of 1800 ∘ C that produces large grains in turn possess less crystal defects and decrease phonon scatterings.

4.4 Corrosion Resistance 4.4.1 Oxidation Behavior in Air

As is known, under high temperature conditions, the oxidation resistance becomes important for the practical uses of β-SiAlONs. The oxidation behavior of hot-pressed β-SiAlONs (z = 0.5, 1.0, 2.5, 3.8) has been studied in only flowing oxygen atmosphere, and their oxidation kinetics has been shown to be influenced by sintering aid, such as Y2 O3 , which was added to enhance the densification process [18]. The oxidation kinetics of reaction-sintered β-SiAlON (z = 3) powder was studied under a nonisothermal condition [19]. Hot-pressed β-SiAlONs (z = 3) have been investigated in different atmospheres by Shimada and Kiyono [20]. However, to our best knowledge, there are very few reports on the oxidation behavior of CS–SPSed β-SiAlONs. Therefore, the purpose of this part was to investigate the oxidation behavior of the dense CS–SPSed β-SiAlONs at different temperatures in air. The effect of z value on the oxidation kinetics and products was systemically examined. The results obtained would give valuable information for the design of β-SiAlONs for the engineering application at high temperature. The flat surfaces and sides of the sintered discs were first ground to remove the carbon foil stuck to them during high-temperature sintering, and subsequently, the surfaces were ground and polished to be parallel to each other and the mirror. After polishing, the diameter and thickness of each disc were carefully measured to calculate its surface area. Prior to oxidation, the polished specimens were ultrasonically cleaned in ethanol, then placed in an alumina boat and then placed into the hot zone of a high-temperature tube furnace. Isothermal oxidation was performed at 1000, 1200, and 1400 ∘ C for up to 100 h in air, and at each temperature, the specimens were taken out of the furnace at intervals of 1, 24, 50, and 100 h to check the mass gain using an analytical balance with an accuracy of ±0.001 mg.

4.4 Corrosion Resistance

8

Z1_1000 °C Z2_1000 °C Z3_1000 °C Z1_1200 °C Z2_1200 °C Z3_1200 °C Z1_1400 °C Z2_1400 °C Z3_1400 °C

7

Mass gain (g m−2)

6 5 4 3 2 1 0 0

100

200 Time (s)

300

400

Figure 4.10 Mass gains as a function of oxidation time for all the specimens (Z1, Z2, and Z3) at 1000, 1200, and 1400 ∘ C. (Reproduced with permission from Ref. [23] © 2010, Elsevier.)

4.4.1.1 Oxidation Kinetics

The mass gains per unit surface area are shown in Figure 4.10 as a function of oxidation time at different temperatures. The final mass gain increases clearly with an increase in temperature and also with the z value. The mass gains are small for the oxidation temperatures of 1000 and 1200 ∘ C. However, they are very large at 1400 ∘ C for all of the samples. At the initial stage of oxidation at 1400 ∘ C, the mass gain is slower for Z3 than for Z1 and Z2. The final mass gains were 4.21, 4.77, and 6.05 g m−2 for Z1, Z2, and Z3, respectively, at 1400 ∘ C. The oxidation kinetics of Si3 N4 and SiAlON has been previously reported to obey a parabolic rate law [21, 22]. However, some of our oxidation curves deviated from the parabolic rate law, and they varied with the z value. The mass gain can be expressed by the following rate raw: Δ� = ��� + � (4.6) � where Δm (g) is oxidation mass gain, A (m2 ) is original specimen surface area, k (g m−2 s−1 ) is oxidation rate constant, t (s) is time, n (–) is exponent, and c (–) is a numerical constant. 4.4.1.2 Microstructure of Oxide Scale

XRD patterns before and after the oxidation for 100 h are shown in Figure 4.11. For Z1, see Figure 4.11a, there are no visible changes for oxidation at 1000 ∘ C compared with the as-sintered product. However, after 1200 ∘ C oxidation, a visible silica (cristobalite) peak was detected and a very few mullite peaks were also detected. When the temperature reached to 1400 ∘ C, the intensity of mullite peaks became stronger and more peaks were visible, and the peak of cristobalite was still

113

114

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

: β-Si5AlON7 (z = 1)

1400 °C

: Al6Si2O13 (Mullite) : SiO2 (Cristobalite)

1200 °C

1000 °C Before oxidation 10

20

30

40

50

60

Diffraction angle, 2θ (°)

(a)

: β-Si4Al2O2N6 (z = 2)

1400 °C

: Al6Si2O13 (Mullite) : SiO2 (Cristobalite) 1200 °C

1000 °C Before oxidation 10

20

30

40

50

60

Diffraction angle, 2θ (°)

(b)

: β-Si3Al3O3N5 (z = 3)

1400 °C

: Al6Si2O13 (Mullite)

1200 °C

1000 °C Before oxidation 10 (c)

20

30

40

50

60

Diffraction angle, 2θ (°)

Figure 4.11 XRD patterns of the surfaces of β-SiAlON discs before and after oxidation for 100 h; (a) Z1, (b) Z2, and (c) Z3. (Mullite: JCPDS 15-0776; SiO2 : JCPDS 29-0085;

β-Si5 AlON7 : JCPDS 48-1615; β-Si4 Al2 O2 N6 : JCPDS 48-1616; β-Si3 Al3 O3 N5 : JCPDS36-1333). (Reproduced with permission from Ref. [23] © 2010, Elsevier.)

4.4 Corrosion Resistance

visible. For Z2, see Figure 4.11b, it showed similar change with Z1, but according to the intensity of the XRD peaks, the quantity of cristobalite was fewer for Z2 than Z1 at 1200 ∘ C. On the other hand, for Z3 shown in Figure 4.11c, no cristobalite peak was detected at any temperature and only mullite peaks were visible for 1200 and 1400 ∘ C oxidation. It is a general agreement that the reaction product is amorphous at initial stage of oxidation and at a low temperature, but the product tends to crystallize after a longer time and at higher temperatures. In our experiments, no visible oxide peaks were detected by XRD when the temperature was 1000 ∘ C. This is attributable to the formation of amorphous oxides. This can also explain why the aforementioned color of the specimens after oxidation changed from gray toward white with increasing temperature. From the XRD results, we can conclude that silica (cristobalite) was formed at initial stage of oxidation for β-SiAlONs with lower z values, but with an increase in the z value or the temperature, also with the oxidation time, mullite increased as the oxidation product. Using the above analyses, we can explain why the aforementioned oxidation kinetics results showed a little deviation from a parabolic rate law of β-SiAlONs with lower z values, which can be attributed to the component change from silica to mullite in the scale. SEM micrographs of the surfaces after oxidation at 1000, 1200, and 1400 ∘ C (represented as Z-1000, Z-1200, and Z-1400, respectively) are shown in Figure 4.12. The microstructures of the oxide scale surfaces varied considerably with oxidation temperature. After 1000 ∘ C oxidation (see Z-1000), no visible oxide crystals could be seen except for some bubbles on the surface of Z1-1000. For Z2-1000 as well, no obvious oxide crystals could be seen, but its surface looked more porous than the surface of Z1-1000. However, for Z3-1000, the microstructure looked very different from that of Z1-1000 and Z2-1000, and there were visible agglomerates of oxide particles on the surface, and these should be mullite according to the results of XRD. After 1200 ∘ C oxidation (see Z-1200), there were large numbers of white needlelike oxides of less than 1 μm in length on the surface of Z1-1200. Combining the EDS analysis and XRD result, we are convinced that the needle-like oxides should be mullite crystals which grew in amorphous SiO2 . Clearly, no oxide crystals were seen on the surface of Z2-1200, except for large numbers of white spots which should be mullite oxide as same as the needle-like crystals in Z1-1200. However, their size was smaller than that in Z1-1200. For Z3-1200, we can see that the surface of the specimen was oxidized considerably. Combining the element analysis with the result of XRD, we can conclude that the oxide is mullite. When the oxidation temperature was increased to 1400 ∘ C (Z-1400), the needlelike mullite crystals in Z1-1400 grew up to more than about 10 μm in length. In Z2-1400, the mullite showed two kinds of shapes, needle-like crystals and agglomerates. The needle-like crystals were same as those in Z1-1400, and the agglomerates were same as those in Z3-1400. The whole surface of Z3-1400 is covered with the scale consisting of agglomerates of mullite.

115

116

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

Z1-1000

Z2-1000

Z3-1000

1 μm Z1-1200

1 μm Z2-1200

1 μm Z1-1400

Z3-1200

1 μm Z2-1400

Mullite SiO2

1 μm

SiO2

1 μm

1 μm Z3-1400

Mullite 1 μm

1 μm

Figure 4.12 SEM images on the surfaces of specimens oxidized at 1000 ∘ C for 100 h (Z1-1000, Z2-1000, Z3-1000); 1200 ∘ C (Z1-1200, Z2-1200, Z3-1200); and 1400 ∘ C (Z1-1400, Z2-1400, Z3-1400) for 100 h. (Reproduced with permission from Ref. [23] © 2010, Elsevier.)

4.4.1.3 Reaction Mechanisms

The aforementioned results indicate that the main solid oxidation products are cristobalite and mullite for β-SiAlONs. Therefore, the oxidation reaction could be expressed by the following equation [18]: β-Si6−� Al� O� N8−� +

18 − 4� � 8−� 24 − 3� O2 → SiO2 + 3Al2 O3 ⋅ 2SiO2 + N2 4 3 6 2 (4.7)

here, in our experiments, z = 1–3. According to this equation, with an increase in the z value, the quantity of mullite will increase, and contrarily, that of silica and nitrogen gas will decrease. In the β-SiAlONs, with increasing z value, more Si–N bonds are replaced with Al–O bonds, and this means that the ratio of Si to Al decreases. Therefore, in the β-SiAlONs with lower z values, at initial stage of oxidation Si reacted with O to form SiO2 . As the oxidation proceeds, more Si is consumed and when the ratio of Si to Al decreases to a certain level, Al reacted with O to form Al2 O3 , which reacts with SiO2 to form mullite crystals. On the contrary, in the β-SiAlONs with higher z values, the ratio of Si to Al is low, so mullite is formed in the entire oxidation process and no separate SiO2 could be formed in the oxide scale. This could

4.4 Corrosion Resistance

Z1-S

Z2-S

Z3-S

Z2-C Oxide scale

Z3-C Oxide scale

10 μm

1 mm

1 mm

1 mm Z1-C

117

Outer surface Oxide scale

10 μm

Figure 4.13 SEM images on the surfaces (Z1-S, Z2-S, Z3-S) and on the cross-sections (Z1-C, Z2-C, Z3-C) of the specimens for oxidized at 1400 ∘ C for 100 h. (Reproduced with permission from Ref. [23] © 2010, Elsevier.)

explain why SiO2 was detected in Z1 and Z2 after oxidation, while only mullite was detected in Z3, cf. the XRD results. Figure 4.13 shows SEM images on the surfaces (denoted as Z-S) and crosssections (denoted as Z-C) of the discs after 100 h oxidation at 1400 ∘ C. There are many bubbles on the surfaces of the discs, and the bubbles on the surface of Z1-S are very big, and many of them cracked. The diameter of cracked holes seems bigger than 50 μm. The bubbles on surfaces of Z2-S and Z3-S get smaller with increasing z value. The thickness of the oxide scale seems close to each other for all of the specimens and looks to be about 15 μm. For the cross-sections, the oxide scale on Z1-C seems dense, but the scales on Z2-C and Z3-C have some pores with the latter having more pores than the former. There are a few large pores near the scale surface, and pores near the interface to the substrate are more numerous but smaller. Combining with the images shown in Figure 4.13, we can deduce that the bubbles on the surface and the pores in the oxide scale should be attributed to nitrogen gas generated at the interface between the scale and substrate during the oxidation, which can be expressed by Equation 4.7. With the progress of oxidation, more nitrogen gas is generated and tends to pass through the oxide scale from the reaction interface to outside of the disc, and then is concentrated in a form of bubbles to some places near the outer surface. Therefore, the pores near the outer surface are big and those near the interface are small. If the oxide scale consisted almost of SiO2 , because most of the SiO2 existed as amorphous phase, it would be so soft that all of the nitrogen gas can pass through it to reach the outer surface easily. However, if the oxide scale consisted almost of mullite, because it crystallized entirely, it would be so hard that it does not allow the nitrogen gas to pass

10 μm

118

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

Air

Oxide scale

N2 N2 molecules or N ions

O2

O2 molecules or O ions

SiAION Figure 4.14 Schematic illustration of the oxidation mechanism of β-SiAlONs. (Reproduced with permission from Ref. [23] © 2010, Elsevier.)

through it. Thus, most of the gas remained in the scale. The above consideration can explain why there were no pores in the oxide scale on Z1, but some pores were formed for Z2 and Z3, and also with increasing z value, the pores increased. The bubbles on the surfaces can be attributed to the nitrogen gas trapped in the scale before cooling down. This reaction processes can be explained by the schematic illustration of the oxidation mechanism shown in Figure 4.14. From the above analyses, the mechanism of oxidation reaction for the β-SiAlONs was a diffusion-controlled process and could be described as the following steps, which are similar to the oxidation of AlN [24]: 1) 2) 3) 4) 5) 6)

Oxygen molecules transfer to the surface of β-SiAlON. Physisorption of oxygen molecules. Dissociation of oxygen molecules and chemisorption. Diffusion of oxygen through the oxide scale to the scale/substrate interface. Chemical reaction producing oxide and nitrogen gas. Gas diffusion through the oxide scale to its outer surface.

4.4.2 Corrosion Resistance in Supercritical Water

Supercritical fluid, especially supercritical water (SCW), that is above the thermodynamic critical point of water (374 ∘ C, 22.1 MPa), has attracted increasing attention in various applications, such as in supercritical water oxidation (SCWO), in supercritical water gasification (SCWG), and for the continuous synthesis of nanoparticles. The environment of reactors presents a big challenge for structural materials used in the components. Many kinds of materials including stainless steel, alloys, and ceramics have been studied for using in SCW atmosphere. However, the details of the corrosion mechanism of each ceramic in an SCW environment were not fully clarified. In this part, we investigated the corrosion behavior of combustion synthesized β-SiAlONs (z = 1–3) in SCW. The effect of z value on the corrosion kinetics and oxide products was systemically examined. The schematic drawing of the experimental apparatus used in this study is shown in Figure 4.15. Here, the reactor is made of SUS316. The samples were exposed in SCW at a constant pressure of

4.4 Corrosion Resistance

Heat exchanger PG Back pressure regulator

Preheater

Refined water

Reactor

TC

Pump

Waste water tank

Heater TC – Thermal couple Sample

PG – Pressure gage

Figure 4.15 Schematic drawing of the experimental apparatus for supercritical water. (Reproduced with permission from Ref. [25] © 2012, Elsevier.)

30 MPa and the temperature of 400 ∘ C for 100 h. The flow rate of the fluid was controlled at 1 ml min−1 . The mass loss per unit surface area and the corroded layer of different z values are shown in Figure 4.16. All of the samples show obvious mass loss, and the final mass losses were 61.36, 36.95, and 6.51 mg cm−2 for Z1, Z2, and Z3, respectively. The corroded layer, which is estimated from the cross-section SEM images, decreases from 480 to 150 μm with the z value increasing from 1 to 3. Figure 4.17 gives the XRD patterns of the specimens before and after corrosion in SCW for 100 h, and all of the peaks have been compared with JCPDS cards. For all of the specimens, there were no impurity peaks detected except for Mass loss Corroded layer

500

60

400

40

300

20

200

0

Z1

Z2

Z3

Corroded layer (μm)

Mass loss (mg cm−2)

80

100

Samples

Figure 4.16 Mass loss and corroded layer of all the samples after exposing in 400 ∘ C/30 MPa supercritical water for 100 h. (Reproduced with permission from Ref. [25] © 2012, Elsevier.)

119

120

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

Z3-A

: β-SiAlON

: Al2SiO5

Z3-B

Z2-A

Z2-B

Z1-A

Z1-B

10

20

30

40

50

60

2θ (°) Figure 4.17 XRD patterns of the specimens before and after exposing in 400 ∘ C/30 MPa supercritical water for 100 h. Here, “A” means after corrosion and “B” means before corrosion. (Reproduced with permission from Ref. [25] © 2012, Elsevier.)

β-SiAlON peaks before corrosion. But the crystalline corrosion products were identified after corrosion. The XRD peaks of these corrosion products on the surfaces of all specimens were compared with JCPDS card, and we confirm that they are Al2 SiO5 (Aluminum silicate, JCPDS Card No. 16-0602). Al2 SiO5 , which can also be expressed as Al2 O3 ⋅ SiO2 , is the low-temperature form of mullite (3Al2 O3 ⋅ 2SiO2 ). When it is heated to over 1100 ∘ C, Al2 SiO5 begins transforming to mullite and quartz. The corrosion mechanism of β-SiAlONs in SCW will be explained later. SEM micrographs of the surfaces of all the specimens (Z1, Z2, and Z3) exposed in SCW at 400 ∘ C under a pressure of 30 MPa for 100 h are shown in Figure 4.18. The microstructures of Z1 and Z2 before corrosion are similar to Z3; therefore, we only show Z3 here. The microstructure of the specimen before corrosion showed dense surface with uniform equiaxed crystals. After corrosion, we can see that the surfaces of the specimens were oxidized considerably, and Z1 looked more accidental than Z2 and Z3. The local magnification of the microstructure on the surfaces was very different from each other. After corrosion, on the surface of Z1, there were large numbers of white needle-like oxides of less than 1 μm in length. For Z2, except for many needle-like oxides, some rod-like crystals could also be seen. But belt-like crystals of less than 0.5 μm in width and several micrometers in length were observed for Z3. Although, the microstructure of the oxides was

4.4 Corrosion Resistance

Z1

Z2

1 μm

20 μm

1 μm

20 μm

Z3

1 μm

20 μm

Figure 4.18 SEM morphologies of the surfaces of specimens (Z1, Z2, and Z3) exposed in supercritical water at 400 ∘ C under a pressure of 30 MPa for 100 h. (Reproduced with permission from Ref. [25] © 2012, Elsevier.)

very different from each other, combining the element analysis with the results of XRD, we can conclude that the oxides are Al2 SiO5 . That is to say, with z value increasing, the oxide crystals grow up from needle-like to belt-like crystals. This may be because more Al and O are included with the increase in z value, and the oxide crystals grow rapidly. Another hypothesis may be brought forward that with z value increasing, the ratio of Al/Si atoms would increase, and the growth direction of Al2 SiO5 crystal would change. It has been reported that Si3 N4 corrodes in SCW by the following reactions [26, 27] Si3 N4 + 6H2 O → 3SiO2 + 4NH3

(4.8)

SiO2 + OH− → HSiO−3 + 4NH3

(4.9)

In addition, due to Masahiro Nagae, on the basis of the composition of SiAlON, the crystalline corrosion products were identified as hydrates of the SiO2 –Al2 O3 system, such as kaolinite (Al2 Si2 O5 (OH)4 ). In our study, according to the results obtained now, we deem that the SiAlON ceramics should be basically corroded by a similar reaction with Si3 N4 in SCW.

121

122

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON

The aforementioned XRD results indicate that the main solid oxidation products are Al2 SiO5 for β-SiAlONs. Therefore, we propose the following oxidation reaction equations: Si5 AlON7 +

1 9 21 H O → Al2 SiO2 + SiO2 + 7NH3 2 2 2 2

Si4 Al2 O2 N6 + 9H2 O → Al2 SiO5 + 3SiO2 + 6NH3

(z = 1) (z = 2)

(4.10) (4.11)

15 3 3 H O → Al2 SiO5 + SiO2 + 5NH3 (z = 3) (4.12) 2 2 2 2 According to these equations, with an increase in the z value, the quantity of Al2 SiO5 will increase, and contrarily, that of silica and ammonia will decrease. Silica was not detected by XRD, which should be flowed out with the water due to reaction (4.8). This could also explain why all of the samples showed obvious mass losses and why the oxide layer was loose and porous. In conclusion, according to the results, β-SiAlONs could be severely corroded in SCW than in air and steam atmosphere. Dense and protective oxide scale could not be formed due to the oxide product of silica being flowed out by the water. Therefore, the reaction mechanism of SiAlON materials has been proposed, and the results in this study could present a valuable example for the exploitation of high corrosion resistance materials. Si3 Al3 O3 N5 +

4.5 Conclusions of This Chapter

In this chapter, pure and dense β-Si6−z Alz Oz N8−z s (z = 1–3) were synthesized by CS followed by SPS without any sintering additives. The Vickers hardness, thermal conductivity, and the corrosion resistance in different atmospheres were investigated for the application as high-temperature engineering materials. In summary, this study introduced an effective method to synthesize pure and dense β-SiAlON materials and showed many valuable property data, which could provide significant references for the application in fields as engineering materials. The content of this chapter is based on the author’s doctoral thesis [28].

References 1. Oyama, Y. and Kamigaito, O. (1971)

3. Ekström, T. and Nygren, M. (1972)

Solid solubility of some oxides in Si3 N4 . Jpn. J. Appl. Phys., 10 (11), 1673. 2. Jack, K.H. (1976) Review Sialons and related nitrogen ceramics. J. Mater. Sci., 11, 1135–1158.

SiAlON ceramics. J. Am. Ceram. Soc., 75 (2), 259–276. 4. Zhu, X.W., Masubuchi, Y., Motohashi, T., and Kikkawa, S. (2010) The z value dependence of photoluminescence in Eu2+ -doped [beta]-SiAlON

References

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

(Si6–z Alz Oz N8–z ) with 1 ≦ z ≦ 4. J. Alloys Compd., 489 (1), 157–161. Ho Ryu, J., Park, Y.-G., Sik Won, H., Hyun Kim, S., Suzuki, H., and Yoon, C. (2009) Luminescence properties of Eu2+ -doped [beta]-Si6–z Alz Oz N8–z microcrystals fabricated by gas pressured reaction. J.Cryst. Growth, 311 (3), 878–882. Hirosaki, N. and Rong-Jun, X. (2005) Characterization and properties of green-emitting beta-SiAlON: Eu2+ powder phosphors for white lightemitting diodes. Appl. Phys. Lett., 86 (21), 211905-3. Munir, Z.A. and Holt, J.B. (1987) The combustion synthesis of refractory nitrides. J. Mater. Sci., 22 (2), 710–714. Munir, Z.A. and Anselmi-Tamburini, U. (1989) Self-propagating exothermic reactions: the synthesis of high-temperature materials by combustion. Mater. Sci. Reports, 3 (7-8), 277–365. Jiang, D., Hulbert, D.M., Kuntz, J.D., Anselmi-Tamburini, U., and Mukherjee, A.K. (2007) Spark plasma sintering: a high strain rate low temperature forming tool for ceramics. Mater. Sci. Eng. A, 463 (1-2), 89–93. Yi, X., Niu, J., Nakamura, T., and Akiyama, T. (2013) Reaction mechanism for combustion synthesis of β-SiAlON by using Si, Al, and SiO2 as raw materials. J. Alloys Compd., 561, 1–4. Yi, X., Watanabe, K., and Akiyama, T. (2010) Fabrication of dense β-SiAlON by a combination of combustion synthesis (CS) and spark plasma sintering (SPS). Intermetallics, 18, 536–541. Yi, X., Zhang, W., and Akiyama, T. (2014) Thermal conductivity of βSiAlONs prepared by a combination of combustion synthesis and spark plasma sintering. Thermochim. Acta, 576, 56–59. Ekström, T., Käll, P.O., Nygren, M., and Olsson, P.O. (1989) Dense single-phase β-sialon ceramics by glass-encapsulated hot isostatic pressing. J. Mater. Sci., 24 (5), 1853–1861. Yi, X., Watanabe, K., and Akiyama, T. (2010) Vickers hardness of betaSiAlON prepared by a combination of combustion synthesis and spark

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

plasma sintering. J. Ceram. Soc. Jpn., 118, 250–252. Yokota, H., Abe, H., and Ibukiyama, M. (2003) Effect of lattice defects on the thermal conductivity of β-Si3 N4 . J. Eur. Ceram. Soc., 23, 1751–1759. Liu, D.-M., Chen, C.-J., and Lee, R.R.-R. (1995) Thermal diffusivity/conductivity in SiAlON ceramics. J. Appl. Phys., 77, 494–496. Liu, Z.-G., Ouyang, J.-H., Zhou, Y., and Xia, X.-L. (2008) Structure and thermal conductivity of Gd2 (Tix Zr1-x )2 O7 ceramics. Mater. Lett., 62, 4455–4457. Persson, J. and Nygren, M. (1994) The oxidation kinetics of beta-sialon ceramics. J. Eur. Ceram. Soc., 13, 467–484. Hou, X.M., Chou, K.C., and Li, F.S. (2008) Some new perspectives on oxidation kinetics of SiAlON materials. J. Eur. Ceram. Soc., 28 (6), 1243–1249. Kiyono, H. and Shimada, S. (2004) Thermal oxidation of sintered beta-sialon (z = 3) ceramics in atmospheres with water vapor. Key Eng. Mater., 264–268, 893–896. Shimada, S., Aoki, T., and Kiyono, H. (1998) Early-stage thermal oxidation of carbothermal beta-sialon powder. J. Am. Ceram. Soc., 81 (1), 266–268. Singhal, S.C. (1976) Thermodynamics and kinetics of oxidation of hot-pressed silicon nitride. J. Mater. Sci., 11 (3), 500–509. Yi, X., Yamauchi, A., Kurokawa, K., and Akiyama, T. (2010) Oxidation of [beta]-SiAlONs prepared by a combination of combustion synthesis and spark plasma sintering. Corros. Sci., 52 (5), 1738–1745. Hou, X. and Chou, K.-C. (2009) Investigation of isothermal oxidation of AlN ceramics using different kinetic model. Corros. Sci., 51 (3), 556–561. Yi, X., Yamauchi, A., Kurokawa, K., and Akiyama, T. (2012) Corrosion of combustion-synthesized β-SiAlONs in supercritical water. Corros. Sci., 56, 153–157. Yoshimura, M., Kase, J., and Somiya, S. (1986) Hydrothermal oxidation of Si3 N4 powder. J. Ceram. Soc. Jpn., 94, 129–134.

123

124

4 Combustion Synthesis and Spark Plasma Sintering of β-SiAlON 27. Oda, K., Yoshio, T., Miyamoto, Y., and

Koizumi, M. (1993) Hydrothermal corrosion of pure. Hot isostatically pressed silicon nitride. J. Am. Ceram. Soc., 76, 1365–1368.

28. Yi, X. (2011) Combustion synthesis and

spark plasma sintering of β-SiAlONs. Doctoral Thesis, 2011.9. Hokkaido University.

125

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application Alexander A. Gromov, Filippo Maggi, Ekaterina V. Malikova, Julia I. Pautova, Alexander P. Il’ in, Elena M. Popenko, Alexey V. Sergienko, Alexander G. Korotkikh, and Ulrich Teipel

5.1 Thermochemical Features of Aluminum Particles Combustion (Theoretical Background)

The features and the combustion behavior of micron-sized metal (μMe, meansurface particle diameter as > 1 μm) and nanometal (nMe, as ∼ 0.1 μm) particles were addressed in the literature by several authors, both experimentally and theoretically, and a restricted selection of relevant contributions is considered here. The combustion of a nonvolatile fuel was faced in Glassman’s book [1]. In this text the author described a general criterion based on thermodynamics to assess if a particle undergoes a process of heterogeneous or homogeneous oxidation. Another fundamental investigation on μAl particle combustion in different atmospheres was carried out by Beckstead [2], which was summed up in a report published by the North Atlantic Treaty Organization (NATO) Research and Technology Organization, covering both numerical simulation and experimental modeling. If we focus on nAl, Dreizin [3] published a wide review on powder production, characterization, and application in different energetic systems. Another comprehensive paper about nanoparticle combustion was recently published by Yetter et al. [4], regarding features and combustion properties. Other interesting investigations about thermodynamic aspects of nMe (mainly melting temperature and surface enthalpy) were performed by Trunov et al. [5], Lai et al. [6], Sun and Simon [7], Navrotsky [8], and Eckert et al. [9]. The complexity of the matter is huge since the oxidation of nAl particles depends on thermodynamical, physical, and chemical features of the reactants involved. In addition to metal–oxidizer combustion, the characteristic size of nAl powders (typically 100 nm or less) deserves further consideration. In the nanometric range, the particle can be composed by few thousands or even few hundreds of atoms. The cohesive energy of atomic clusters is expected to be inversely dependent by the particle radius while surface energy increases and may become non-negligible. Several particle properties such as melting temperature, reactivity, and surface tension may differ from the bulk features [4, 10, 11]. Regarding calorimetric Nitride Ceramics: Combustion Synthesis, Properties, and Applications, First Edition. Edited by Alexander A. Gromov and Liudmila N. Chukhlomina. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

properties, experimental evidence can be found, for example, in a decrease of the melting temperature and the melting enthalpy as a particle becomes finer [6, 7, 9]. Considering nAl thermodynamics, Dreizin in his review [3] stresses on the fact that the nanoscale certainly influences the thermodynamic properties and the stability of oxides and powders in general. However, on a theoretical basis, the energetic increment for nAl due to the nanostructured nature cannot exceed +39% for the combustion of a single Al atom and, according to their considerations, the value is further reduced to +4% when 1 nm particle is considered [12]. Navrotsky [8, 9, 13] analyzed the thermodynamics of nanoparticles (mainly oxides), analyzing the influence of surface effects over the bulk properties. It is clear that the contribution of the nanoscale becomes important when specific surface area turns out to be in the order of 200 m2 g−1 . The size of nAl for energetic purposes (combustion) is lower bounded by the fact that the active metal content steeply decreases as average diameter falls below 20–50 nm due to the presence of a passivating oxide layer of approximately constant thickness. This fact limits the specific surface area to a value in the order of 10–25 m2 s−1 . Insofar as the surface energy of aluminum at ambient temperature is about 0.89 J m−2 [12], the surplus energetic content of a 50–100 nm powder ranges between about 20 and 40 J (see Figure 5.1) and results to be negligible if compared to the heat release of the oxidation process itself as well as to the thermodynamics of this class of particles. The analysis of particle combustion through thermochemistry grounds on the minimization of the Gibbs free energy. The computations look at the adiabatic flame temperature and relevant mixture properties under the hypothesis of thermodynamic equilibrium [14, 15]. In the present document, the National

10 000

10 000 Specific surface

1000

1000

100

100

10

10

1 1

Figure 5.1 size.

10 100 Particle size (nm)

Surface energy (J g−1)

Surface energy Specific surface area (m2 g−1)

126

1 1000

Specific surface area and surface energy of aluminum particles as function of

5.1

Thermochemical Features of Aluminum Particles Combustion (Theoretical Background)

Aeronautics and Space Administration (NASA) Chemical Equilibrium with Applications (CEA) software is adopted as reference [16–18]. According to the Glassman’s criterion, the thermochemistry can anticipate several information about the behavior of metals burning in an oxidizing environment, including the tendency of particles either to evaporate and burn homogeneously in gaseous phase or not [1, 2]. In this sense, a fundamental role is played by the nature of a refractory metal oxide generated by the oxidation reaction and by availability of energy to gasify a metal [19]. Aluminum is classified as a volatile metal with a nonvolatile oxide and homogeneous combustion is expected in an oxygen environment. Nevertheless, Rai et al. noted that, when nAl particles are involved in these reactions, their characteristic size is smaller than the mean free path (�) of gaseous molecules and the continuum hypothesis does not hold anymore. The authors [12] suggest that, in such a case, the combustion is more likely to be a surface process ruled by the collision between the particle and oxygen molecules, then transported through the oxide shell. The Adiabatic flame temperature represents an important parameter which can be derived from thermochemical computations. The vaporization–dissociation or volatilization of the refractory oxide bounds the temperature regardless of the nanometric or micrometric particle size involved in the combustion. A quantitative criterion was established by Glassman and Yetter [1, p. 496], who stated: “The interesting observation is attributable to the fact that the heat of vaporization, dissociation or decomposition of the metal oxide formed is greater than the heat available to raise the condensed state of the oxide above its boiling point.” Basically, the enthalpy released during the combustion cannot consume the whole produced oxide by the process of vaporization or decomposition since the required enthalpy is not enough. Thus, a limiting temperature for the combustion process is established on the basis of thermochemical considerations. Glassman also observed that this upper bound might be overcome if an extra enthalpy is attributed to the initial reactants, namely, sensible enthalpy from heating or, we add, from other sources such as surface energy of nMe. However, Glassman had demonstrated that appreciable effects for aluminum were visible only for enthalpy increase in the order of 10 kJ g−1 or above, far from the surface energy of nAl particles used for energetic purposes. 5.1.1 Aluminum–Oxygen Systems

Two main stoichiometric reactions 5.1–5.2 are usually considered for aluminum oxidation for different oxidizer/fuel (O/F) mass ratios: O = 1.7, F O = 1.1. 2 Al + 3∕2O2 → Al2 O3 at F Al + 1∕2O2 → AlO at

(5.1) (5.2)

127

128

2500

0.6

3600

Al AlO Al2O Al2O2 Al2O3(l) O O2 Temperature

0.5 0.4 0.3 0.2

2900

0.7

Temperature (K)

Product mole fraction

3000 0.7

5000 4300

Product mole fraction

4000 3500

Temperature (K)

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

0.6 0.5 0.4 0.3 0.2 0.1

0.1 0 0.00 0.50 1.00 1.50 2.00 2.50 3.00

0.0 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 O/F Mass fraction

O/F Mass fraction

Figure 5.2 Aluminum–oxygen combustion at (a) 1 and (b) 10 bar. Adiabatic flame temperature and main composition of combustion products.

In Figure 5.2, equilibrium computations for the adiabatic flame temperature of Al are reported for variable O/F ratios, along with the most important aluminumbased products. The highest temperature is obtained for O/F = 0.9, slightly shifted toward the fuel-rich side of the reaction 5.2. The value of the adiabatic flame temperature derives from the relation between condensed and gaseous species, being, namely, alumina and aluminum suboxides. Its variation, at constant pressure and across the consistent range of O/F, does not present a real plateau but computed variations are very weak on the oxidizerrich side. The refractory oxide Al2 O3 was found in the condensed liquid state. The gaseous form is not detected by thermodynamics. From experiments it appears that gaseous Al2 O3 is an intermediary product, with a very short lifetime [2]. As O/F increases, the dissociation of O2 becomes important thanks to the high temperature and the excess of oxidizer. Part of the atomic oxygen comes also from the dissociation of Al2 O3 into its suboxides. These reactions determine the adiabatic flame temperature, which results to be heavily dependent upon the equilibrium between the phases and, in turn, upon the environmental pressure. In this regard, Beckstead [2] reported the relation 5.3 existing between the flame temperature and the vapor pressure of aluminum oxide: Tflame =

−6

250.5 × 10

1 . − 14.1 × 10−6 × ln(P)

(5.3)

The pressure dependence of combustion products and temperature for constant O/F = 0.9 are plotted in Figure 5.3. As expected, adiabatic flame temperature increases along with environmental pressure while the decrease in concentrations of atomic Al and O indicates a decrement in Al2 O3 dissociation, present in a liquid state. If the Glassman criterion is applied, the combustion of aluminum in oxygen is expected to occur as a homogeneous gas phase reaction in the investigated range. The boiling point of the metal is recovered from [20].

Thermochemical Features of Aluminum Particles Combustion (Theoretical Background)

5500

Product mole fraction

4500 0.40

3500

0.30 0.20

Temperature (K)

5.1

Al AlO Al2O Al2O2 Al2O3(l) O O2 Temperature

0.10 0.00 1

10 Pressure (bar)

100

Figure 5.3 Aluminum–oxygen combustion. Products and adiabatic flame temperature for varying pressure and constant O/F = 0.9.

5.1.2 Aluminum–Nitrogen Systems

The concept of the limiting temperature can be better appreciated for aluminum reaction in a nitrogen environment, where the refractory product species is aluminum nitride, instead of alumina. The stoichiometric reaction of oxidation dictates Al + 1∕2N2 → AlN, with the O/F mass ratio 0.52. However, backward reaction of AlN dissociation happens as well, resulting in a plateau of temperature which extends toward lower O/F values and depends on environmental pressure. From Figure 5.4, it is possible to recognize the abovementioned concurrent reactions that contribute toward the final equilibrium. As the O/F ratio is reduced, more aluminum and less excess of nitrogen are present. Sensible enthalpy of the products is incremented. However, around O/F = 0.7, the onset of AlN decomposition effectively limits the temperature, generates liquid aluminum, and increases the available nitrogen. This process overlaps with the O/F shift toward the oxidizer-lean region. The flame temperature is influenced by dissociationvaporization properties so it gains dependence from pressure. It progressively increases up to about 3600 K for 100 bar reaction while the decomposition onset of AlN moves on higher value of O/F mass ratio. Finally, according to the Glassman criterion [1], the reaction between Al and N2 is mainly heterogeneous since the adiabatic flame temperature fails in reaching the boiling point of aluminum. 5.1.3 Aluminum–Air Systems

When aluminum is reacted in air, the complexity of the oxidation process is increased due to the overlapping of the above mentioned chemistry processes.

129

130

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application 4000

AL AL(L) ALN(L) N2

Product mole fraction

0.80 0.60 0.40

3500 Temperature (K)

1.00

1 bar 10 bar 100 bar

3000

2500 0.20 0.00

0

0.5

1

1.5

2000 0.0

2

0.5

O/F mass ratio

1.0

1.5

2.0

O/F mass ratio

Figure 5.4 Main reaction products at 10 bar (a) and adiabatic flame temperatures (b) of aluminum reaction with nitrogen.

Thermochemical results point out that the dominant oxidation process is the reaction with oxygen despite its limited presence in the atmosphere (about 23 mass%). The species which are formed at the equilibrium are mainly represented by vapor of aluminum, its refractory oxide, and relevant suboxides (Figure 5.5). As for the cases already addressed in this chapter, the dissociation of the refractory species limits the temperature and, like the combustion with pure oxygen, only liquid alumina is present in the products. According to the results obtained so far, the nature of this oxidation in air switches from homogeneous to heterogeneous when the pressure increases. In adherence with the Glassman criterion, the adiabatic flame temperature is higher than the boiling point of aluminum for 1 and 10 bar, allowing for metal evaporation and mixing with the oxidizing atmosphere. Conversely, the situation changes at 100 bar, when the flame temperature does not enable liquid 0.20

5000 Al

0.15

1 bar 10 bar 100 bar

4500

Al2O AlO

Temperature (K)

Product mole fraction

Al2O3(l)

0.10

0.05

4000 3500 3000 2500

0.00 2.0

3.0

4.0

5.0

6.0

O/F mass ratio

7.0

8.0

2000 2.0

3.0

4.0

5.0

6.0

7.0

O/F mass ratio

Figure 5.5 Main reaction products at 10 bar (a) and adiabatic flame temperatures (b) of aluminum oxidation in air.

8.0

5.2

Chemical Features of Metals Combustion in Air (Experimental Background)

metal phase to vaporize. It follows that the oxidation of liquid aluminum can occur only through a heterogeneous process. 5.2 Chemical Features of Metals Combustion in Air (Experimental Background)

The results of μMe combustion in air with nitrogen as competing oxidizer appeared since Pokhil et al. [21], while those experimental data cannot be explained by equilibrium thermodynamic (see Section 5.1.3). The combustion of μB1) powder in air and with NH4 ClO4 was studied in [22], where it was found that hexagonal BN is the main product of combustion, formed by the interaction of B with bounded and atmospheric nitrogen. The elementary stages of particles combustion taking into account not only chemical but also thermophysical characteristics of the process [23]. μAl particles combustion even at relatively low temperatures (1000–1500 ∘ C) is characterized by the predominance of gas-phase combustion mechanism over diffusion one. By gas-phase μAl combustion “one of the three reacting Al atoms is entrained by gaseous products” [23]. Until the 1990s, it was believed that μAl powders with a particle size of more than 1–2 μm are not ignited and not burned in air at atmospheric pressure until a temperature which is close to the boiling point of Al [24] due to the significant internal heat dissipation and diffusion mechanism of oxide film growth. The ignition temperature of single Al particle depends on a particle size and an oxide film. Many authors publish contradictory information about the ignition temperature, which is apparently due to differences in experimental conditions. According to [25, 26], a single Al particle with the diameter of 400 μm ignites in air at 2090 ∘ C. Various data on the ignition temperature for Al particles are presented in [25–30], but the ignition temperature is always more than 2100 ∘ C. The ignition temperatures of the particles with the diameter of 6–40 μm were experimentally and theoretically studied in [26], where it varied in the range 995–2100 ∘ C. The temperature reaches (2650–2700) ± 50 ∘ C after ignition on the first stage of combustion, and it is in between of the boiling temperatures of Al (2520 ∘ C) and alumina (2980 ∘ C) [24]. 5.2.1 Combustion of Aluminum Particles in Air

Al particles of diameter 10–500 μm have temperature of ignition above 1900 ∘ C at heating in air or in steam [31]. The beginning of Al particles combustion occurs in vapor phase; moreover, the intensity of a luminescence zone appearing around a particle increases slowly. Stationary combustion is characterized by existence of a luminescence zone, which does not modify their sizes almost until the complete burnout of a metal. Aluminum suboxides observed in flames are hollow spheres in condensed combustion products (CCPs), which amount 1) Here and after, boron (B) is considered as a metal for combustion processes.

131

132

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

is 30–40 mass% of initial Al [23]. The appearance of a flame around a burning particle [32], rotation of a burning particle [32], and deviation of a trajectory from initial [33] indicate the presence of a gaseous phase during Al combustion [34]. Calculation of the intermediate combustion products composition for Al in air at T = 3270 ∘ C includes gas suboxides AlO and Al2 O also appearing in spectra flames during Al combustion in argon–oxygen mixtures and oxygen [32–34]. The reactions (5.4) and (5.5) should be valid at the temperature of particle surface about 2000 ∘ C [21]. But according to [35], these reactions are possible at 800–900 ∘ C. Volatility of Al2 O3 in Al presence increases in a factor ∼100 [23]: 4 Al (l) + Al2 O3 (l) → 3 Al2 O (g),

(5.4)

Al (l) + Al2 O3 (l) → 3 AlO (g).

(5.5)

There is some information about the properties of aluminum suboxides in the literature: for example, according to [36], Al2 O is formed by heating a mixture Al2 O3 with Si up to 1800 ∘ C in high vacuum; and AlO is formed by heating the metal Al powder to 3620 ∘ C. Al volatility is extremely high during Al oxidation in air at temperatures of 480–600 ∘ C [23], whereas there is almost complete absence of gaseous products of oxidation in a mixture of oxygen and argon at low pressures. Gaseous aluminum suboxides are recorded during combustion of Al in various conditions, including in air [37]. CCP of μAl in air having different morphology (mostly hollow spheres [38, 39]), depending on the combustion conditions and shape of initial particles, consist of α-Al2 O3 with minor impurities of χ-, θ-, γ-Al2 O3 according to X-ray diffraction (XRD) [40, 41]. However, CCP consist of mainly nitrogen-containing phase Al(8/3+x/3) O4−x Nx (spinel) under the conditions of shock-wave loading of μAl with apparent density ρa = 0.35 g cm−3 in air [42] whereas in nitrogen δ-Al2 O3 is formed, indicating the involvement of oxygen in the intensive nitridation of Al. The kinetic and thermal characteristics of combustion for particles of Al-Mg alloy in air were studied in [21], the CCP contained the significant amount of spinel Al2 MgO4 . Further experiments on combustion of Al-Mg alloy particles in air were repeated in [43]. The presence of trace amounts of nonstoichiometric oxynitride (Al3 O3 N) in the CCP along with an oxide spinel phase was shown in [43]. 5.2.2 Combustion of Boron Particles in Air

Boron suboxides were detected in the spectrums of flames of boron powder in air [44, 45]. BO prevails among suboxides, and then BO2 and B2 O2 follow in decreasing order of content [46, 47]. B2 O3 and unburnt boron can be identified in CCP, whose presence is probably due to refractoriness of boron (Tmelt. = 2074 ∘ C) and low melting point of its oxide. The model of a flame wave propagation in a boron–air mixture [48], corresponding to a real flame with a satisfactory accuracy, also includes gas-phase combustion of boron [49, 50] with the formation of intermediate spectral-registered boron suboxides BO and B2 O2 . There is

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

expected participating boron suboxide in carbothermic synthesis of BN in [45] and the process of nitride formation in the solid phase. Boron, boron oxides, hydroboric acid, and also nitrides and carbides of boron in small amount are detected in the CCP of the powdered boron burning in solid propellant [51]. The combustion of boron powder in air was studied by bomb calorimetry [22]. The samples were oxidized in KClO4 and NH4 ClO4 and burned in nitrogen, oxygen, and air for an independent determination of a nitrogen source. Formation of BN in combustion without KClO4 and NH4 ClO4 occurred with the same yield of nitrides as for nitrogen and air. And the content of nitrides increased with the air pressure. The content of bound nitrogen in the CCP, according to X-ray photoelectron spectroscopy (XPS), was up to 14 mass%. The CCP of Mg and Be particles in air are oxides [21], because for these metals nitrides are unstable phases at high temperatures. For Ti and Zr, on the contrary, the formation of nitrides on the surface of particles, and even on the bulky metal was recorded in the works of Glassman in the 1960s.

5.3 Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

Among all metals only lithium mostly forms nitride Li3 N (T m = 845 ∘ C) but not oxide at the storage in air [52]. Interaction between lithium and air occurs slowly at room temperature, the reaction is accelerated at T ∼ 250 ∘ C, and then reoxidation goes on to Li2 O formation [53]. Nitrides of other alkali metals are formed at high temperatures in the reaction between metal vapors and nitrogen in absence of oxygen. Mg3 N2 forms during the oxidation of powdered magnesium under the limited access of air. Bound nitrogen in the products of Mg particles combustion in air has not been recorded [54, 55]. Nitrides of beryllium and magnesium are obtained by direct reaction of elements with nitrogen [56]. The reaction of nitrogen with the formation of beryllium nitride begins at 900 ∘ C, and magnesium nitride – at 560 ∘ C. CCP of B [57], Al, Ti [58], and Zr [59] in air contained the traces of nitrogen-containing products – nitrides, oxynitrides, and oxicarbonitrides. The appearance of experimentally observed nitrogen-containing phases in CCP is in contradiction with the equilibrium thermodynamic calculations (see Section 5.1): First, metals should react with O2 as a significantly more active gas than N2 . Second, if the formation of nitrogen-containing intermediate products is possible, they must be immediately oxidized by O2 , especially at high temperatures [60–62]. Thermodynamic calculation of two-stage process of AlN formation (stage I) with its subsequent afterburning (stage II) was first performed in [63] for the system “Al–air.” Staging of the combustion of fused metal particles in air and high-temperature flames are probably due to the formation of liquid and gaseous intermediate products [64]. Experimental confirmation of AlN formation in large quantities (more than 50 mass%) at two-stage combustion of free-poured nAl in

133

134

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

air2) was first obtained in [65]. Thus, metals interaction with N2 of air with nitrides (oxynitrides) formation is possible. Nitrides formation at burning of metals significantly reduces the enthalpy of metals combustion (in factor 2.6 for AlN in comparison with Al2 O3 ). Excitation of the N2 molecule, which weakens all three bonds between nitrogen atoms, requires 941 kJ mol−1 of energy [66]. Nitrogen is still used as an inert gas in many technologies. Heating, ionizing radiation, action of catalysts, electric discharge, as well as the combination of these methods are the most developed and studied methods of activation of nitrogen molecule. The degree of dissociation reaches 0.1% at thermal activation up to 3000 ∘ C [67]. The efficiency of the other activation process does not exceed 1–2% [68–71]. Thus, the search for intermediates, catalysts, and special conditions of the reactions of nitrogen or activated nitrogen complex [72] with metals is of great practical interest also for atmospheric nitrogen fixation [73]. 5.3.1 Nitrides Formation at nAl/�Al and [nAl + (�Al/Zr Alloy)] Combustion in Air

According to the equilibrium thermodynamics and experimental data [74–76], AlN should be oxidized in air with the heat release. μAlN powder (as ∼ 63 μm) is oxidized in air, starting with ∼800 ∘ C. AlN oxidation was observed in [76] even for the sintered AlN samples at room temperature. The basic experimental data of the process of nAl (as ∼ 0.1 μm, produced by electrical explosion of wires [77]) combustion in air was obtained in [78–80]. The yield of nitrides in combustion products of powdery mixes nAl (as ∼ 0.1 μm)/μAl (as ∼ 20 μm) was varied by the composition of initial powdery mixes with other constant physical and chemical parameters: the initial room temperature and atmospheric pressure of nitriding reagent – air. Burning temperature tended to decrease with the increasing ratio nAl/μAl in the initial mixture, while AlN content in the combustion products decreased [80]. Another additive to nAl was μAl/Zr alloy powder [79] (mass ratio Zr/Al = 84/16). Burning temperature increased with the increasing of μAl/Zr alloy powder content in the mixture (Table 5.1) and the total content of AlN and ZrN in the CCP reached maximum at μAl/Zr alloy powder content in the initial mixture of 40 mass%. The weak dependence of the intensity of nitrides formation during the combustion on the particle size of the reagents, which was unusual for the metals combustion in nitrogen by combustion synthesis (CS), was found in [78–80]. 5.3.2 Nitrides Formation at nAl Combustion in Air

The efforts to burn nAl (apparent density ∼0.2 g cm−3 ) in a quartz tube showed that air supply to the combustion wave through the layer of nAl at pressure 2) Here and after the air pressure is 0.1 MPa (1 atm.), if not specially given.

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

Table 5.1 Phase composition of CCP (XRD data) of the mixtures [nAl + (μAl/Zr alloy)] burned in air at 0.1 MPa. nAl content in the mixture (mass%)

40 50 60 70 80

Phase content in CCP (mass%) AlNa

ZrNa

Al2 O3 (� + �)

ZrO2

Al

Otherb

20 33 36 40 44

28 27 25 15 5

24 27 28 33 37

18 8 3 4 4

7 2 4 5 5

3 3 4 5 5

a)

Chemical analysis on bound nitrogen, relation between nitrides by thermogravimetric analysis (TGA) data. b) Adsorbed gases, water, and other volatile products.

0.1 MPa is nearly absent, which is true also for the μMe powders [81]. Thus, gaseous reactant supply to the combustion wave at low pressures and, consequently, the temperature of the process are determined by the porosity of the CCP. The samples of nAl (Figure 5.6) of different masses were burned in air (Figure 5.7) and CCP were analyzed by common methods [82]. AlN content in CCP increases with increase of the nAl mass in initial samples, the residual Al content changed inversely to AlN content. The AlN content reaches a maximum and tends to increase with increasing of nAl mass. The content of oxides (γ + α) Al2 O3 decreases with increasing mass of the samples. At the same time, the ratio of phases in CCP of different masses shows that the aluminum at combustion in air is mostly consumed on reaction with nitrogen (Figure 5.8). [AlN/(γ + α) Al2 O3 ] ratio in CCP is ∼4–8 in a stationary region (m > 2 g) and increases with the bigger sample mass. AlN phase in the combustion products

SEI

(a)

10.0kV

X70.000

100 nm WD 10.0 mm

(b)

Figure 5.6 SEM image of nAl powder ((a) ×70 000) and transmission electron microscopy (TEM) image of nAl particle ((b) ×1 000 000).

135

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

5

1

2

3 4

Figure 5.7 Experimental setup for metal powders burning in air: 1 – igniter, 2 – sample of powder (red circle – ignition area), 3 – thermocouple, 4 – metal plate, and 5 – photo camera. (Reproduced with kind permission from Elsevier.)

AlN/Al2O3 Al2O3 / Al α (Al - AlN)

AlN/Al (AlN+Al2O3) / Al α (Al -Al2O3) 70

10

60 8 50 6

40

30

4

20

Al degree of conversion (α) (%)

Phase mass ratio in CCP of nA l in air

136

2 10

0

0

2

4 m (nAl ) (g)

6

0

Figure 5.8 The ratio of crystal phases of CCP and Al conversion degree under nAl samples of different masses burned in air.

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

I (%)

Al3O3N

100

AlN Al Al2O3

80

60

40

20

0

15

25

35

(a)

45

55

65

75

2δ, (°) (CuKα) I (%) 100

AlN

80 60

Al

40 Al3O3N

20 0

Al2O3 0

10

20

30 m (g)

(b) Figure 5.9 Typical XRD pattern of nAl CCP (a) and relative phase content of samples of different masses burned in open air (b).

is major (according to XRD), starting with the mass of the initial samples ∼1 g (Figure 5.9). Reflexes of α-Al2 O3 are presented on XRD patterns, but the maximum intensity of 100% peak does not exceed 0.3 of the 100% peak of AlN (Figure 5.9a). nAl samples of large masses burn incompletely, and the residual metal content in CCP increases again, passing through a minimum mass of the initial samples of 15 g (Figure 5.9b). 5.3.2.1 CCP Microstructure

nAl contains ∼10–20 mass% of Al2 O3 (as a passivation layer) after production and passivation in air [16]. The content of Al2 O3 in CCP is also ∼20 mass% according XRD. So, the mass content of aluminum oxides in the initial nAl and in CCP is virtually identical. One would assume that all reacted metal Al is spent on reaction with nitrogen in case the mass transfer inside the burning sample is absent. But, according to scanning electron microscope (SEM),

137

138

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

Figure 5.10 SEM images of CCP of nAl in air (×2000). (Reproduced with kind permission from Elsevier.)

there is a significant difference in the morphology of the initial nAl and CCP in air (Figures 5.6 and 5.10) that indicates the intensive mass transfer during combustion. While the initial powders are agglomerates of spherical particles whose diameter is substantially less than 1 μm (see Figure 5.6), the products of combustion form a three-dimensional framework structure (see Figure 5.10). Evidently, the needle-shaped crystals formation proceeds with participation of gaseous intermediate products considering the microstructural characteristics of nAl CCP in air [83, 84]. Needle shape of crystals is typical for nitride crystals obtained, for example, by evaporation–condensation methods [85] or CS. Anisotropic needle crystals (whiskers) of the CCP contain a significant amount of fixed nitrogen in the form of AlN (according to energy-dispersive spectrum (EDS) and XRD). AlN phase is mainly represented by needle crystals. Geometrical parameter characterizing the ratio of needle length to its diameter varies deeply: from 1 to 50. Needle crystals, formed during the combustion process, permeate a ceramic matrix, and the effect of material self-armoring achieved. There are no whisker structures on SEM images of the nAl samples quenched at T ∼ 750 ∘ C (Figure 5.11a). The content of bound nitrogen in the quenched CCP does not exceed 1.5 mass%. According to XRD data, the main product of nAl oxidation by air is γ-Al2 O3 on the first combustion stage. Thus, the reaction between aluminum and oxygen mainly takes place in the first stage of combustion [78]. Participation of the gas phase in the formation of whiskers determines the dependence of the structure of nAl CCP on the pressure. The morphology of the CCP formed during combustion at the initial P = 0.1 MPa of air was represented mainly by whiskers (Figure 5.11b). Preferential formation of whiskers in a closed volume (pressure decreasing by combustion from 0.1 to 0.05 MPa) is apparently because the intensification of aluminum evaporation (Figure 5.11c) accompanied the process of air components being consumed during combustion. Moreover, the equilibrium of reactions 5.4 and 5.5 shifts to the right, toward the formation of gaseous substances, when pressure decreases. That is why if nAl burn by high air pressure (0.5 MPa, Figure 5.12) the whiskers are not found (Figure 5.11d).

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

(a)

(b)

(c)

(d)

Figure 5.11 SEM images of nAl CCP in air on the first stage ((a) ×10 000); on the second stage ((b) ×1400); at low pressure ((c) ×10 000); and at high pressure ((d) ×4500).

1 t (°C) 3

2800 2400 2000 1600 2

1200 800 400 0

10

20

30

40

50

60

70

Figure 5.12 nAl combustion process at air pressure P = 0.6 MPa: 1 – brightness; 2 – thermocouple history; and 3 – pressure change.

τ (s)

139

140

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

Analysis of nAl combustion process in nitrogen–oxygen gas mixtures was carried out at different content of O2 in N2 from 2 to 100 vol%. The combustion chamber had a volume V = 1023 cm3 (the mass of nAl samples was constant 1 g). Ignition of the samples occurred at the content of O2 in the gaseous environment of ∼2 vol%, which indicates the significant role of oxygen in the mechanism of particles oxidation at the first stage of combustion. Dependence of the content of AlN in CCP on the gas composition of the initial gaseous environment O2 /N2 is shown in Figure 5.13. AlN content in CCP increases linearly with increasing N2 concentration in the gas mixture from 0 to 90 vol%, then reaches a maximum (53 mass% AlN) at 92–94 vol% N2 . The samples of nAl were not ignited when the concentration of N2 in the initial gas mixture was over 98 vol%. The presence of 2 vol% of O2 is necessary for ignition of nAl in O2 /N2 mixture at P = 0.1 MPa. This O2 amount provides the ignition due to the high enthalpy of the process of interaction of Al and O2 . N2 /O2 ratio in gaseous burning media also affects the microstructure of nAl CCP. SEM images of CCP obtained from the combustion of nAl in two different gas mixtures N2 /O2 = 97/3 and 3/97 (vol% in V = 1023 cm3 ) are shown in Figure 5.14. CCP of nAl in the mixture with a high content of N2 forms a 3D-structure (Figure 5.14a). The basis of the 3D-sinter is whiskers connecting the fused particles of unreacted aluminum deformed to lose their original form. The structure of the CCP cakes is the evidence of the intense processes of Al boiling and its oxide melting for N2 /O2 = 3/97. Maximum temperature during nAl combustion for that N2 /O2 ratio was 3004 ∘ C (over the Al boiling point). As a result, nearly monolithic glassy cake of Al2 O3 was formed (Figure 5.14b). αAl2 O3 was the main CCP phase. There was also unreacted aluminum (10 mass%) in nAl CCP.

Calculated enthalpy of combustion (kJ−1 mole of Al) AlN (mass%) 850 50

800

40

750 No combustion

30

700

20

650

10

600

0

0

10

20

30

40

50

60

70

80

90

100

550

N2 (vol%)

Figure 5.13 AlN content in CCP of nAl samples burned in O2 /N2 mixtures at 0.1 MPa and the corresponding calculated enthalpy of combustion depending on the initial N2 content.

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

(a)

(b)

Figure 5.14 SEM images of nAl CCP in N2 /O2 = 97/3 vol% ((a) ×8000) and in N2 /O2 = 3/97 vol% ((b) ×8000).

5.3.2.2 Effect of Additives on AlN Yield

The oxidation of Al powders with rare-earth metals added to embrittle the oxide shell on Al particles was studied in [28–30]. The introduction of such additives reduces the agglomeration of particles during combustion, and increases combustion efficiency and flame temperature. The use of nMe additives to nAl is one of the factors affecting the yield of nitride phase in the CCP. The highest yield of AlN (up to 66 mass%) in CCP at nAl combustion in air was observed for the additives nFe, nMo, and nW. The additive of nB (Ssp = 11 m2 g−1 ) is of the greatest interest, since boron is used in the synthesis of ceramic materials AlN-BN. nB probably does not enter into reaction on the first stage of combustion of mixtures (nAl + nB) despite comparable specific surface area with nAl. Boron particles are covered by a liquid film B2 O3 . Part of the heat of the reaction of Al combustion is spent on melting and gasification of B2 O3 , which is the reason for the increased duration of the first stage for (nAl + nB) combustion. The phase composition of CCP changes with increasing boron content in an initial mixture (Table 5.2). Many phases, known for system “Al–B–N–O”, except aluminum borides, are present in the CCP. All boron is probably burned in the gas phase for mixture

Table 5.2 Phase composition of CCP (XRD data) of mixtures (nAl + nB) burned in air at 0.1 MPa. Initial sample composition

nAl + 15 mass% nB nAl + 40 mass% nB nAl + 50 mass% nB nAl + 80 mass% nB

Combustion stages

Two Two Two One

CCP composition by XRD Main phase

Other phases

AlN B2 O3 BN B2 O3

Al3 O3 N, Al, α-Al2 O3 AlN, 9Al2 O2 ⋅2B2 O2 , α-Al2 O3 , Al, Al3 O3 N AlN, Al3 O3 N, Al Al

141

142

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

AlN

Mo

1 W

3

B C Fe

2

Sn

Al2O3

Al

Figure 5.15 Diagram of phases composition influence of additives Sn, Mo, W, Fe, C, B. Phase composition change at sample mass of nAl CCP in air: 1 – combustion on open air; 2 – AlH3 combustion; and 3 – combustion increasing is shown with dotted arrows. in closed vessel at initial pressure 0.1 MPa;

(nAl + 15% nB): its compounds are not detected by XRD. Single-stage combustion occurs when more than 50 mass% of nB is added to nAl. The common phase diagram of CCP content for nAl combustion with additives is shown in Figure 5.15. Combustion of mixtures [nAl + (μAl, as ∼ 100 μm)] and [nAl + (μAl, as ∼ 10 μm)] in air occurs in two stages as in the case of nAl combustion without additives. The duration of the first combustion stage rises with increase in the content of μAl in the initial mixtures. The combustion products of the mixtures differ in morphological characteristics and represent an aggregate of submicron needles emerging from the gas phase and the dendritic crystals formed from the melt (Figure 5.16). Dendritic crystals are formed, in particular, on the inner surface of the burnt μAl particles. The dependence of the AlN content in CCP (Figure 5.17) on the content of μAl in

10 μm (a)

10 μm (b)

Figure 5.16 SEM-images of CCP of mixture (μAl + 20 mass% nAl) in air ((a, b) ×1300). (Reproduced with kind permission from Springer.)

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

45

2 I

AlN (mass %)

35

1

25

15

5

II

0

10

20

30

40 50 60 μAl (mass %)

Figure 5.17 AlN content in the CCP in dependence on μAl (1, as ∼ 100 μm) and (2, as ∼ 10 μm) content in the initial mixtures with nAl: I – the area of sustainable

70

III

80

90

two-stage combustion; II – pulsed two-stage combustion second part; and III – no ignition. (Reproduced with kind permission from Springer.)

the initial mixture has two characteristic regions: from 10 to 70 mass% of μAl (I) and more than 70 mass% (II). The first plot in Figure 5.17 represents a plateau, that is, AlN yield is practically not reduced with increasing of μAl content in a mixture from 10 to 70 mass%. Such a dependence of AlN content on μAl quantity shows that μAl, almost on the same level with nAl, participates in the process of nitride formation. When the content of μAl in the initial mixture is more than 70 mass% the thickness of nAl interlayers becomes less than the critical value, and they do not provide sufficient internal insulation for sustainable burning. As a result of these factors, the content of AlN in the CCP is greatly reduced. 5.3.3 AlN (Al3 O3 N), ZrN, TiN Obtained by Combustion of Metal Powders and their Mixtures in Air

The disadvantage of industrial “furnace” methods of nitrides synthesis is high energy consumption, low yield of nitrides (30–50 mass%), contamination with graphite and iron, the need for grinding of received nitride-containing sinters, oxygen-free atmosphere (N2 or NH3 ), as well as high gas pressure – for the classical CS [87]. Solving this problem is possible by replacement of nitrogen and ammonia, and their mixtures, by the cheapest nitrogen-contained oxidizer – air. The combustion process of nonpressed metal powders at 0.1 MPa (atmospheric pressure) [88] and the usage of nanopowders [89] or their additions to the micron-sized ones allow to produce homogenized nitride ceramic batches. Ceramic materials in the systems “Me–O–N” [89, 90] are the most promising, that is, refractory nitrides of hexagonal modifications of AlN, BN, and cubic modifications of ZrN, TiN, and oxynitrides. The use of composites [91] (e.g.,

143

144

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

AlN–AlON, AlON–BN) is more preferable for the technologies [92]. For example, cermets based on AlN and metals (Al, Ni, W, and Cr) have good mechanical properties [93]. The characteristics of the studied μAl, μTi, and μZr powders in comparison with nAl are summarized in Table 5.3. The nAl and μZr powders contain 1.5 and 1.3 mass% of dissolved hydrogen accumulated during the storage time [94]. Dissolved hydrogen released by fast heating from the surface of nAl and μZr powders [95] can determine low ignition temperature for such powders [96]. The properties of the used μAl were comprehensively described in [97]. The experimental setup was previously proposed in [65]: Powders with the defined bulk density were poured onto a steel plate (Figure 5.7) to form a cone-like sample. The combustion experiments were carried out in air. Ignition of powders was initiated with the heated tungsten wire from the top of the cone-like samples. The ignition temperature was T ∼ 1000 ∘ C (igniting tungsten wire temperature) for sustainable self-propagating combustion wave existence. The combustion process occurred in air for the four studied powders (nAl, μAl, μTi, μZr) similar Table 5.3 Characteristics of metal powders (in brackets – methods of analysis for parameters). Characteristic

Powder nAl

�Al

�Ti

�Zr

Real density, �real (g cm−3 )

2.7

2.7

4.5

6.5

Bulk density �b (g cm−3 )

0.20

0.90

1.14

1.86

Metal content (EDX, XRD, volumetric analysis), [Me] (mass%)

91.2

98.5

Area of the specific surface (BET), Ssp (m2 g−1 )

10.5

0.2

Mean-surface particle diameter (calculated from BET), as (μm)

0.22

Oxide content (EDX, XRD, differential thermal analysis (DTA)-thermogravimetric analysis (TGA)), [Mex Oy ] (mass%)

5.1a

Particle shape (SEM)

Spherical

Oxide layer thickness on particles (transmission electron microscopy (TEM)), aox (nm)

2

Concentration of the dissolved hydrogen in the powder (melting in inert atmosphere), [H2 ] (mass%)

1.50

a) [Mex Oy ] = 100 − [Me] (%). Source: Reproduced with kind permission from Elsevier.

10

99.8 0.06 22

98.6 0.06 15

1.5

0.2

1.4

Spherical

Spheroidal

Spheroidal

18

0.45

20

0.03

22

1.30

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

to a nonstationary self-propagating process after the local ignition (Figure 5.18a). The bright combustion zone propagated through the whole volume of nAl and μAl cone-like samples after the primary burning spots had overlapped (fingering combustion regime, refer to [96]). The burning process was very fast for μTi and μZr (Figure 5.18b) without initial burning spots propagation (see Figure 5.19). The bright stage of the combustion was followed by sample cooling, when the

2800

Temperature (°C)

2400

nAl

μAl

μZr

2000 1600 1200 800 400 0

μTi 0

50

100

150

200

Burning time (s) (a) 1599.99 1516.84 1433.70 1345.78 1248.30 1137.91 1005.71 821.62 764.40 724.81 679.03 622.92 531.87 474.38 446.57 412.08 365.95 288.88 0 °C

(b) Figure 5.18 Plots of the temperature of metal powders (4 g) while burning in air (a) and thermal image of the μZr burning (b).

145

146

Burning stage

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

Powder nAl

μAl

μTi

μZr

8s

58 s

2s

2s

32 s

144 s

15 s

5s

65 s

200 s

32 s

20 s

1 cm I

II

III

Figure 5.19 Photo images of metal powders combustion in air. Each image was made at the time shown below (time after ignition). Burning stages: I – ignition and surface burning of particles (only for nAl and

μAl), II – heterogeneous combustion (at a maximal temperature), and III – cooling and crystallization of CCP. (Reproduced with kind permission from Elsevier.)

main part of the metal had already reacted. The maximum burning temperature (T b max ) achieved during the combustion process was lower than the boiling point for all metals, except Al: nAl (T b max = 2563 ∘ C), μAl (T b max = 2414 ∘ C), μTi (T b max = 1737 ∘ C), and μZr (T b max = 1982 ∘ C) (Figure 5.18). However, it should be noted that the temperature measurements were semi-quantitative because even the 100 μm-thermocouples had a delay in registration [87]. The powders combustion was accompanied by a flame torch (Figure 5.19) of the volatile products – most probably Al2 O, AlO2 and AlO [36], TiO [98], and hydrogen, accumulated in μZr powder (in the form of ZrH2 ). The peaks of nitrides were the highest one on XRD patterns for four studied powders (Figure 5.20 and Table 5.4), while nitrides can stabilize with much lower yield for other burning conditions in air [99]. 5.3.3.1 nAl

The peaks of α-Al2 O3 and spinel-phase Al3 O3 N appeared on XRD patterns along with AlN (Figure 5.20a and Table 5.4). Anisotropic needle-like crystals of the

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

combustion products contained a significant amount of bound nitrogen (AlN, 34 mass% of N, and Al3 O3 N, 6 mass% of N), according to XRD and energydispersive X-ray spectroscopy (EDX). The content of bound nitrogen for CCP of nAl was 25 mass% (EDX data), that was three times more than the content of bound oxygen. The needle-shaped crystals were mainly represented by AlN phase (Figure 5.21a), coated with oxynitride Al3 O3 N – the phase which is more stable to oxidation than AlN.

- AlN

20 000

- Al

Intensity (a.u.)

- Al3O3N - α-Al2O3

15 000

10 000

5000

0 20

30

40

(a)

50

60

70

80

2θ (°)

30 000 - AlN - Al

25 000

- Al3O3N

Intensity (a.u.)

- α-Al2O3 20 000

15 000

10 000

5000

0 20

(b)

30

40

50

60

70

80

2θ (°)

Figure 5.20 XRD patterns of the CCP of metals, burned in air (CuKα irradiation): (a) CCP of nAl and (b) CCP of μAl. (Reproduced with kind permission from Elsevier.)

147

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application 16 000

- TiN - Ti

Intensity (a.u.)

12 000

- TiO3 (rutile)

8000

4000

0 20

30

40

(c)

50

60

70

80

2θ (°)

20 000

- ZrN - Zr - ZrO2 (Tetragonal) - ZrO2 (Monoclinic)

15 000 Intensity (a.u.)

148

10 000

5000

0 20

30

(d)

40

50

60

70

80

2θ (°)

Figure 5.20 (Continued)

5.3.3.2 �Al

The same combustion scenario for both nAl and μAl was confirmed by the composition of CCP: AlN peaks on XRD patterns were the most intensive (Figure 5.20b). The amount of unburned Al was higher for μAl (17 mass%) compared to nAl (8 mass%), which was explained by 45 times larger initial particles size (see Table 5.3) and much lower burning rate (see Figure 5.18) for μAl powder in comparison with nAl. The quantity of needle-like crystals was much lower for CCP of μAl, according to SEM (Figure 5.21b).

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

Table 5.4 Composition of CCP of metals in air. Parameter

Metal powder nAl

�Al

�Ti

�Zr

8 25 67 Al

10 21 69 Al

7 12 81 Ti

5 10 85 Zr

Crystalline phases in CCP (XRD) (mass%)

AlN (72), α-Al2 O3 (6), Al3 O3 N (15), Al (7)

AlN (62), TiN (59), ZrN (38), α-Al2 O3 (10), TiO2 (rutile) ZrO2 Al3 O3 N (16), (10), Ti (31) (monocl.) (17), ZrO2 Al (12) (tetrag.) (22), Zr (23)

Bound nitrogen content (Kjeldahl method) (mass%)

22

18

n.a.

n.a.

Residual metal content (volumetric method) (mass%)

8

17

n.a.

n.a.

Phase composition of the CCP after slow heating (10 K/min) in air up to 1200 ∘ C (XRD)

α-Al2 O3

α-Al2 O3

TiO2 (rutile) ZrO2 (monoclinic and tetragonal)

Content of element in CCP (EDX) (mass%)

O N Me

Source: Reproduced with kind permission from Elsevier.

5.3.3.3 �Ti

The composition of CCP contained the following phases: Ti, TiO2 (rutile), TiN (see Table 5.4). There were no other phases of titanium oxides except TiO2 (rutile). In accordance with SEM, the CCP of μTi consisted of fused aggregates (Figure 5.21c) formed from the melt, while the initial Ti particles were spheroidal. 5.3.3.4 �Zr

The combustion products of μZr consisted of ZrN, Zr, and ZrO2 (tetragonal and monoclinic), according to XRD pattern. There was a large quantity of unburned zirconium in the CCP of μZr, while the intensity of 100% peak of ZrN was maximal (Figure 5.20d). The particles of CCP of μZr were the agglomerates of spheroidal shape (3–10 μm in diameter). Thus, high-porous sponge of CCP was evidently formed from gas flux (Figure 5.21d). 5.3.3.5 Combustion Scenario

The chemical reactions, considered for the Ti (as an example) combustion in air, are summarized in Table 5.5. This mechanism was confirmed by a SEM image of the CCP of μTi (Figure 5.21c). They were formed through a liquid phase. The dynamics of phase formation in the combustion wave of the “Ti–N2 ” and “Ti–air”

149

150

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

(a)

(b)

(c)

(d)

Figure 5.21 SEM images of the CCP of metals, burned in air: (a) CCP of nAl × 1900; (b) CCP of μAl × 1900; (c) CCP of μTi × 1000; and (d) CCP of μZr × 3000. (Reproduced with kind permission from Elsevier.)

Table 5.5 Chemical reactions of Ti combustion in air. Burning stage

Chemical reaction

No.

I

0 Ti(s) + O2 (g) → TiO2 (s); ΔH298 = −944 kJ mol−1

(1)

II

0 = +15 kJ mol−1 Ti(s) → Ti(l); ΔH298 0 Ti(l) + O2 (g) → TiO2 (s); ΔH298 = −929 kJ mol−1 1∕2Ti(l) + 1∕2TiO (s) → TiO(l); ΔH 0 = −472 kJ mol−1 2 298 0 TiO2 (s) → TiO2 (l); ΔH298 = +67 kJ mol−1 0 = −526 kJ mol−1 Ti(l) + 1∕2O2 (g) → TiO(s); ΔH298 0 = +149 kJ mol−1 TiO(l) + 1∕2N2 (g) → TiN(s) + 1∕2O2 (g); ΔH298 0 Ti(l) + 1∕2N2 (g) → TiN(s); ΔH298 = −323 kJ mol−1

(2) (3) (4) (5) (6) (7) (8)

III

0 Ti(l) → Ti(s); ΔH298 = −15 kJ mol−1 0 = −67 kJ mol−1 TiO2 (l) → TiO2 (s); ΔH298 0 = −621 kJ mol−1 TiN(s) + O2 (g) → TiO2 (s) + 1∕2N2 (g); ΔH298

Source: Reproduced with kind permission from Elsevier.

(9) (10) (11)

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

was investigated in [100]. It was shown there that TiN synthesis during Ti combustion in air seemed difficult, because an oxynitride phase appeared on the sample surface almost immediately after TiN formation. It was also claimed that even in case of TiN formation it was further oxidized to TiO2 due to high oxygen reactivity. In contradiction to [100], 45 mass% of crystalline TiN in CCP was found (see Table 5.4). The samples of μZr blazed up immediately after ignition. The bright white glow was visually observed. A low-temperature stage was very fast (less than 1 s) and the process was transferred into a high-temperature stage (thermal explosion). Then afterburning and cooling-down took place (the third stage). The hydrogen, contained in the surface layer of μZr grains (see Table 5.3), could evolve from the metal particle, and H2 could be ignited first during powder ignition. Thus, the fastest stage of μZr combustion was hydrogen burning, followed by the reaction of Zr with nitrogen and oxygen. The same process of H2 -release from the AlH3 powder and subsequent burning as a torch was described in [86]. Fragility and sponginess of CCP for μZr appeared due to the gaseous products formation. Zirconium oxynitrides are not stable and, thus, have not been registered on XRD pattern (see Figure 5.20d). The main question in the mechanism of nitrides formation in air was: Why did metal react with the “inert” nitrogen instead of reactive oxygen? The products of slow oxidation (10 K min−1 ) for the studied metals in air contained no nitrides. They are completely oxidized (see Table 5.4). During slow reactions, the oxygen diffuse in the porous powdery samples and a complete oxidation can occur. So, the theoretical mechanism of the studied metal powders combustion in air with nitrides formation was suggested after the analysis of experimental data (Figure 5.22). Stage I (Ignition and Surface Burning of Particles) After ignition, a metal particle was oxidized from the surface with almost no change in its morphology. So, the oxygen diffused through the oxide layer on the particle surface afterwards. The combustion of hydrogen accumulated in nAl and μZr particles during their storage provides very low temperatures of particles ignition (T ign ≈ 410 ∘ C for nAl and T ign ≈ 290 ∘ C for μZr). Hydrogen burning induced additional heating of particles during the first stage of combustion. Local quasi-adiabatic heat accumulation in a small plot (up to few μm3 ) heated the metal until the melting temperature was achieved. Also, reactions of metals with volatile compounds started to play a significant role. Stage II (Heterogeneous Combustion) When the particle temperature became

higher, the volume of liquid metal increased, the oxide film burst, and the melted metal flew out. Fast reaction of metal melt with air occurred, and this process was accompanied by a rapid temperature growth. The high brightness of the combustion zone and a flame torch over the sample indicated the gaseous products release (metal vapors and suboxides). Thus, oxygen is consumed mainly for the reaction with melted metal and further suboxides formation. Liquid metal

151

152

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

Temperature raise

Ignition

Metal particle

Surface burning

Hydrogen release and combustion

Metal melting

Oxide layer broken and metal melt flow out

Heterogeneous combustion, temperature increase up to maximal value

Slow process

Fast process

Metal melt combustion

Slow process

Formation and crystallization of condensed combustion products

Gas and liquid products combustion

Quasi-stationary combustion

Cooling

Figure 5.22 Physical model of the metal particle (in a powder) burning in air. (Reproduced with kind permission from Elsevier.)

and gaseous or liquid suboxides further directly react with nitrogen for solid nitrides formation. Stage III (Cooling and Crystallization of CCP) Considerable part of CCP was stabi-

lized in the form of anisotropic crystals corresponding to the “vapor–liquid– crystal” mechanism. The formation of nitrides during combustion of metal powders in air is a key feature not only for Al, Ti, and Zr but also for La [101], Si [102], Ga [103], and B [22]. It is reasonable that a sufficiently small and porous sample would oxidize completely (no nitrides could stabilize). 5.3.4 Nitrides Obtained by Combustion of �Ti/�Al and �Ti/�TiO2 Mixtures in Air

The combustion of μTi and its mixtures with “inert” and “active” (flammable) additives in air were studied. Two groups of free-poured powdered mixtures (m = 10 g of each sample) were burned in air: I – μTi/μTiO2 , where μTiO2 is “inert” diluent; II – μTi/μAl (spherical and flaked particles), where μAl is flammable diluent. Mixture I, containing less than 50 mass% of Ti, failed to ignite with the help of a heated tungsten spiral (T = 1000 ∘ C). Therefore, the μAl of 10 mass% excess of 100% were added. For the same reason, μAl with spherical particles was replaced by the highly dispersed μAl flake particles for the mixtures II, containing less than

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

50 mass% of Ti. Mass burning rate (Figure 5.23) was calculated by Equation 5.6: νM =

mo , τc

(5.6)

where mo = sample mass (g) and � c = combustion time (s). Steady combustion wave, which was characteristic for CS systems “Ti–N2 ,” was formed during the combustion of all samples. The mass burning rate for the mixtures I was weakly dependent on the content of Ti in the initial mixture in the range of 0.2–0.5 g s−1 (Figure 5.23, I). It can be explained by the high enthalpy of titanium oxide melting (67 kJ mol−1 , rutile). The curve of the mass burning rate (Figure 5.23, II) for compounds II can be divided into two parts, because Al was used by varying particle size in this group of samples. The mass burning rate for the mixtures II was weakly depended on Ti content in the initial samples of the mixtures with Ti < 60 mass%, but significantly increased when the μAl with spherical articles was replaced by highly dispersed μAl flake particles (Figure 5.23, II). The

Mass burning rate (g s−1)

1 II

Ti

I

0.8 Ti-TiO2-10 wt. % Al (spherical particles)

0.6 0.4

Ti-TiO2 Al (flake particles) Ti-Al (flake particles)

0.2 No ignition

0

0

Ti-Al (spherical particles)

20

40

60

80

100

Ti (mass %) Figure 5.23 Mass burning rate depending on Ti content in the initial samples of the mixtures “Ti–TiO2 ” (I) and “Ti–Al” (II) (see also Table 5.6). (Reproduced with kind permission from Elsevier.) Table 5.6 EDX data for the CCP of the mixtures on the basis of μTi burned in air. Sample group

I (μTi/μTiO2 )

II (μTi/μAl)

Ti content in the initial mixture (mass%)

20 60 100 0 20 60

Element content in CCP (mass%) Al

Ti

O

N

5 — — 71 54 11

68 74 81 — 18 35

13 8 7 8 15 43

14 18 12 21 13 11

Source: Reproduced with kind permission from Elsevier.

153

154

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

cakes of CCP for all samples had a golden color, indicating the presence of TiN on the surface. The exception was the CCP of the sample, consisting of μAl only. The results of XRD showed presence of TiN as a main phase in the average combustion products of the mixtures I. The XRD pattern of the CCP for the mixture (60% mass μTi/40% mass μTιO2 ) corresponded to 85% TiN. According to EDS, bound nitrogen was contained in the CCP of all samples (Table 5.6), which was part of the nitrides, detected by XRD. The formation mechanism of TiN in air was proposed on the analysis of a phase and an elemental composition and photographs made with SEM. 5.3.4.1 Combustion of the Mixtures I (“Ti–TiO2 ”)

The process of nitride formation determined the composition of combustion products if unburnt metal would not be taken into account. TiN formation on the number of consecutive reactions (Table 5.5) in the combustion wave was similar to the processes described early for AlN formation during combustion in air of the powder mixtures “Al–Al2 O3 ”. After the ignition, Ti begins to interact with atmospheric oxygen with large amount of heat release. Ti melts at higher temperatures and reacts with atmospheric nitrogen. Another possible mechanism of TiN formation is the interaction of titanium monoxide TiO, which is an unstable compound and strong reducing agent, with atmospheric nitrogen. According to SEM (Figure 5.21c), the combustion products of Ti is formed through a liquid phase. Thus, the inert diluent TiO2 increases the yield of TiN for the mixtures “Ti–TiO2 .” 5.3.4.2 Combustion of the Mixtures II (“Ti–Al”)

Probably, the particles of Ti and Al burnt separately because the burning rate for Ti in air was much more than for Al (Figure 5.23). In fact, the already burning Ti particles ignited Al ones and the total amount of the heat released during combustion, reduced, because part of them was spent on Al melting. It led to less active interaction of both metals with air nitrogen than for the mixtures “Ti–TiO2 .” The presence of active diluent – Al – at Ti combustion in air had a negative impact on output of both TiN and AlN. Thus, the yield of TiN, obtained by combustion of Ti powder in air, was higher than by the combustion of powders of the same dispersion in pure nitrogen. The method of Ti burning, described in this section is much easier – free-poured powder samples and air as a nitrogen source were used. The mechanism of TiN formation is probably determined by involving intermediate TiO, which has a high reducibility. 5.3.5 Combustion Synthesis of Aluminum Oxynitride in Air

The CCP of mixtures of μAl with nano-γ-Al2 O3 (Ssp ∼ 252 m2 g−1 ) burned in air contained significant amount of aluminum oxynitride, which underlined the diversity of the products obtained by CS in air depending on the particle size of the

5.3

Nitrides (Oxynitrides) Formation by Metal Powder Combustion in Air

Relative intensity (%)

100 80 60 AlN 40 Al3O3N 20 0

0

20

40 60 μAl in mixture (mass %)

80

100

Figure 5.24 The relative content of AlN and Al3 O3 N in CCP depending on the content of μAl powder in initial mixture with γ-Al2 O3 (XRD data). (Reproduced with kind permission from Elsevier.)

reagents. XRD and chemical analysis showed that the CCP of 10 g of μAl/nanoγ-Al2 O3 mixtures consisted of unreacted aluminum, AlN, aluminum oxynitride Al3 O3 N (Figure 5.24), and aluminum oxide α-Al2 O3 . Initial γ-Al2 O3 completely transformed into α-Al2 O3 even at single-stage combustion (T < 1200 ∘ C), even at the first stage. α-Al2 O3 was the main phase of the combustion products for μAl content 10–30 mass%. Al3 O3 N formed as the main phase when μAl content was 70–100 mass%. The sustainable AlN formation was observed for μAl content 60–70 mass%. Large aluminum particles acquire a coating layer of fine γ-Al2 O3 particles during a preliminary mixing. Insulated aluminum particles primarily ignited. Propagation of the combustion wave ignites the majority of aluminum particles, if one assumes that the combustion wave penetrates through the layers of γ-Al2 O3 and ignites new aluminum particles. In addition, the insulation layer consisting of aluminum oxide prevents Al melting in the combustion wave. Thus, the reaction with N2 (5.7) is possible for the mixture of μAl with nano-γ-Al2 O3 : Al + Al2 O3 + 1∕2N2 → Al3 O3 N.

(5.7)

The burning of the mixtures of μAl with nano-γ-Al2 O3 in air occurs in one or two stages, that is, similar to the burning of μAl powder without additives. The mechanism of AlN (Al3 O3 N) synthesis at the mixtures combustion is obviously similar to the combustion of nAl: similar combustion scenario and the composition of CCP. Thus, the new class of chemical reactions with nonequilibrium products (nitrides and oxynitrides) formation at the high-exothermal metals of III–IV groups’ combustion in air was experimentally confirmed.

155

156

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

5.4 Application of the Synthesized Nitrides and Oxynitrides in Dense Ceramics

ZrN/ZrO2 composites are used for the production of nozzles or for sprayers of solutions and melts; also they can be used for the production of technical ceramics. AlN/Al3 O3 N/Al2 O3 composites are used for the manufacturing of substrates for integrated circuits employed in microelectronics, as well as for the manufacturing of radar radiation absorbing material. TiN/Al2 O3 composites represent conductive ceramics, so they can be used for the manufacture of heating elements. TiN/AlN/Al2 O3 composites can be used for the manufacturing of cutting tools, holders of yarns in textile industry [104]. 5.4.1 Nitride Ceramics on the Base of the CCP in the System “Zr–O–N”

Two groups of ceramic powders were used to produce dense ceramic bodies on the basis of CCP of (μZr + nZrO2 ) in air: ZRI ∶ ZrN + 20 mass% Zr + 40 mass% ZrO2 , ZRII ∶ ZrN + 15 mass% Zr + 20 mass% ZrO2 . The sintering of green samples was carried out at 1550–1950 ∘ C for 1 h in nitrogen. The samples from the first group ZRI have low sinterability because of the presence of 14 mass% of monoclinic ZrO2 in the green bodies which undergoes phase transformations during sintering causing the change in volume. The addition of 1–3 mass% Y2 O3 into the starting composition did not result in significant changes in the sintering results. The maximum density (82%) of the samples from the send group was achieved at 1900 ∘ C, but ceramics characterized by high porosity (up to 5%) that suggested that the stabilization process of tetragonal ZrO2 had begun during the combustion of the primary mixture and contributed the sintering of ceramics. Additional introduction of 1–3 mass% of Y2 O3 to the second group ZRII led to an increase in density up to 91 g cm−3 and bending strength up to 380 MPa, the hardness of all samples averaged 18 GPa (Figure 5.25). The positive effect of the Y2 O3 additives was associated with the formation of a solid solution of yttrium oxide with zirconium nitride, which accelerated the diffusion processes during sintering and led to an increase in the samples density. The ceramics microstructure was characterized by a uniform grain size distribution and the presence of two phases: ZrN and ZrO2 . In accordance with the results of XRD, the ceramics (ZRII + 2 mass% Y2 O3 ) consisted of zirconium nitride and zirconia in the tetragonal modification. 5.4.2 Nitride Containing Ceramics on the Base of the CCP in the System “Al–O–N”

Two groups of ceramic powders were used to produce dense ceramic bodies on the basis of CCP of μAl + 50 mass% μAl2 O3 (Al3 O3 N as a main product) and CCP

5.4

Application of the Synthesized Nitrides and Oxynitrides in Dense Ceramics

HV1 = 19.0 GPa HV2 = 17.5 GPa

70 μm Figure 5.25 The microstructure of the ceramic sample on the basis of ZrON with 2 mass% Y2 O3 (nitrogen, 1950 ∘ C, 1 h).

of μAl + 30 mass% μAl2 O3 (AlN as a main product) in air. In order to improve the properties of ceramics, sintering aids were chosen: 1–4 mass% of Y2 O3 , C (soot), and B2 O3 . Yttrium oxide forms eutectic with alumina and so the melt appears during the sintering process at the same temperature (1850 ∘ C) [105–107]. It forms garnet Y5 Al3 O12 , dissolves refractory Al2 O3 grains, and promotes the sintering. Carbon black acts as a reducing agent [108]. Yttrium oxide and boron oxide form the melt of an yttrium borate at 1580 ∘ C, which promotes the sintering of the ceramic material. Carbon black reduces yttrium borate above 1700 ∘ C by the reaction 5.8: 2YBO3 + 3C + N2 → 2BN + Y2 O3 + 3CO.

(5.8)

Furthermore, the carbon black reduces part of Al2 O3 contained in the starting ceramic material AlN [108]. Sintered samples without additives contained two main phases: AlN and Al3 O3 N. In the samples with sintering additives, it was identified yttrium aluminum garnet Y3 Al5 O12 . The amount of Y3 Al5 O12 in ceramics depended on the amount of additive (Figure 5.26). Compounds containing carbon and boron were not detected in XRD pattern of ceramic samples. Ceramic samples with 1% Y2 O3 sintered at 1850 ∘ C in nitrogen atmosphere characterized by uniform distribution of the phases. Porosity decreased (from 13% to 3% and from 31% to 9%, respectively), the relative density and compressive strength increased (from 69% to 89% and from 355 to 827 MPa, respectively) after increase of Y2 O3 amount. Introduction of complex additives (3.5 mass% Y2 O3 + 1.5 mass% C; 2.5 mass% Y2 O3 + 0.45 mass% C + 0.7 mass% B2 O3 ) also had a positive effect on the properties of sintered samples. Optimal additive was 3.5 mass% Y2 O3 + 1.5 mass% C. Relative density of 91% and hardness of 7 GPa was achieved (Table 5.7).

157

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

100

5

7

6

4

38

35

35

38

90 80

41

43

Intensity (a.u.)

70

Y3Al5O12 Al3O3N AIN

60 50 40 30

57

59

57

58

59

58

1% Y2O3

2% Y2O3

10

12

12

9

10

52

48

44

45

45

45

48

42

44

43

46

45

Without adds

1% Y2O3

2% Y2O3

4% Y2O3

20 10 0

Without adds

4% Y2O3

3.5% Y2O3+ +1.5% C

2.5% Y2O3+ +0.45% C+0.7%B2O3

(a)

100 90 80 Y3Al5O12

70 Intensity (a.u.)

158

60

Al3O3N AIN

50 40 30 20 10 0

3.5% Y2O3+ 2.5% Y2O3+ +1.5% C +0.45% C+0.7%B2O3

(b) Figure 5.26 The relative phase composition of the ceramics material after sintering at 1850 ∘ C: (a) based on Al3 O3 N (CCP of μAl + 30 mass% μAl2 O3 ) and (b) based on AlN (CCP of μAl + 50 mass% μAl2 O3 ).

5.4

Application of the Synthesized Nitrides and Oxynitrides in Dense Ceramics

Table 5.7 Properties of ceramics obtained on the basis of the CCP μAl + μAl2 O3 with additives (nitrogen, 1850 ∘ C, 2 h). Characteristic

Porosity (%) Relative density (%) Compressive strength (MPa) Hardness (GPa)

Sintering additive (CCP of �Al + 50 mass% �Al2 O3 (Al3 O3 N as a main product))/(CCP of �Al + 30 mass% �Al2 O3 (AlN as a main product)) No additive

1% Y2 O3

2% Y2 O3

4% Y2 O3

Y2 O 3 + C

Y2 O3 + C + B2 O3

31/28 69/71 355/369

13/8 86/89 770/612

12/9 88/89 790/528

11/12 89/88 827/458

9/15 91/84 751/333

10/13 90/87 640/425

n.a./2

6/7

6/6

7/3

7/3

5/3

5.4.3 Technology of Nitride Ceramics Production on the Basis of the CCP in the System “Me(Al, Ti, Zr)–O–N”

The technological scheme of the composite ceramics based on the CCP powders (Figure 5.27) has been developed on the basis of results of studies on the synthesis materials by combustion and sintering of obtained nitride materials. The developed technology involves the use of industrial μMe as combustion reagents and air as the nitriding agent that significantly reduces the cost of the ceramics production. According to the scheme (see Figure 5.27), the proposed technology of composite ceramics from the CCP involves the use of inexpensive equipment and does not require significant energy costs. The production line of the nitride containing composite ceramics obtained by CS in air includes mixture preparation area, CS area, molding powder preparation area, formation area, and sintering area. The main component (metal) from hopper (1) by a screw (2) is fed into a mill (4) passed through a flow meter (3). Depending on the amount of the main component, a certain amount of additives is added using microdosage device (5). The metal oxide is fed into the mill from a second hopper (6) with a screw (7). Required quantity of oxides is dosed through the flow control device (8). Ball milling of the components was carried out after mixing (4). Ceramic mixture is placed in the hopper (9). From the hopper (9) the mixture is sent to a dozer (10). The dozer (10) delivers portions of mixture on a vane (11) which forms the prismatic shape of powdery sample (12) on the conveyer (13). The plate conveyor (13) is of use as the combustion installation. The plates of conveyor should be made from special treatment steel. Initiation of the combustion process is carried out by means of a fuse (14). The gaps between supplied powdery mixtures are given to ensure fire safety. The conveyor (13) is covered with a special net for extinction sparks with 2–5 mm cells. To ensure heat sink, the water-cooling of the conveyor is organized. After combustion, the CCP are placed into the massive steel rolls (15). The rolls (15) previously ground and cool combustion products. And in case of the system “Zr–O–N”, the rolls (15)

159

160

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application

6

1

5

11

8

2 3 7

4

12 9

14

10 13 11

17

15

16

Figure 5.27 The production line of the nitride containing composite ceramics obtaining by combustion synthesis in air (see comments in the text).

also carry out the quenching of combustion process. The next stage is milling of the CCP in a ball mill (16). The milled CCP sifted using a vibrating sieve (17). Screenings are utilized. Milled and sifted powder is sent to the molding powder preparation. And then formation and sintering are carried out. References 1. Glassman, I. and Yetter, R.A. (2008)

4. Yetter, R.A., Risha, G.A., and Son,

Combustion, 4th edn, Chapter 9, Academic Press, pp. 495–550. 2. Beckstead, M.W. (2004) Internal Aerodynamics in Solid Rocket Propulsion, Chapter 5, von Kármán Institute for Fluid Dynamics NATO RTO-EN-023, pp. 1–46. 3. Dreizin, E.L. (2009) Prog. Energy Combust., 35, 141–167.

S.F. (2009) Proc. Combust. Inst., 32, 1819–1838. 5. Trunov, M.A., Umbrajkar, S.M., Schoenitz, M., Mang, J.T., and Dreizin, E.W. (2006) J. Phys. Chem. B, 110, 13094–13099. 6. Lai, S.L., Guo, J.Y., Petrova, V., Ramanath, G., and Allen, L.H. (1996) Phys. Rev. Lett., 77, 99–102.

References 7. Sun, J. and Simon, S.L. (2007) Ther8. 9.

10. 11.

12.

13. 14.

15.

16.

17.

18.

19.

20.

21.

mochim. Acta, 463, 32–40. Navrotsky, A. (2003) Geochem. Trans., 4, 34–37. Eckert, J., Holzer, J.C., Ahn, C.C., Fu, Z., and Johnson, W.L. (1993) Nanostruct. Mater., 2, 407–413. Vanfleet, R.R. and Mockel, J.M. (1995) Surf. Sci., 341, 40–50. Farrell, H.H. and Van Siclen, C.D. (2007) J. Vac. Sci. Technol. B, 24, 1441–1447. Rai, A., Park, K., Zhou, L., and Zachariah, M.R. (2006) Combust. Theor. Model., 10, 843–859. Navrotsky, A. (2001) Rev. Mineral. Geochem., 44, 73–103. Glassman, I. and Sawyer, S.F. (1970) The Performance of Chemical Propellants, AGARDograph, vol. 129, Technivision, Slough, pp. 27–80. Smith, E.B. (2004) Basic Chemical Thermodynamics, Imperial College Press, pp. 36–62. NASA Chemical Equilibrium with Applications (CEA) downloadable software, www.grc.nasa.gov/WWW /CEAWeb (accessed 15 May 2014). Gordon, S. and McBride, B.J. (1994) Computer Program for Calculation of Complex Chemical Equilibrium Compositions and Applications I. Analysis, NASA RP, 1311, NASA Glenn Research Center, Cleveland, OH. McBride, B.J. and Gordon, S. (1996) Computer Program for Calculation of Complex Chemical Equilibrium Compositions and Applications II. User’s Manual and Program Description. NASA RP, 1311-P2, NASA Glenn Research Center, Cleveland, OH. Williams, A. (1997) in Physical and Chemical Aspects of Combustion: A Tribute to Irv Glassman (eds F.L. Dryer and R.F. Sawyer), Gordon and Breach, pp. 267–289. Steinberg, T.A., Wilson, D.B., and Benz, F. (1992) Combust. Flame, 91, 200–208. Pokhil, P.F., Belyaev, A.F., Frolov, Y.V., Logachev, V.S., and Korotkov, A.I. (1972) Combustion of Powdered Metals in Active Media, Science, Moscow.

22. Xi, X. and Shufen, L. (2003) Energy

Mater., 21, 63–72. 23. Petrov, Y.I. and Bibilashwili, R.S.

24. 25.

26.

27.

28.

29.

30.

31.

32.

33. 34. 35.

36. 37.

38.

(1964) Russ. J. Phys. Chem. A, 38 (11), 2614–2624. Dreizin, E.L. (2000) Prog. Energy Combust. Sci., 26 (1), 57–78. Ermakov, V.A., Razdobreev, A.A., Skorik, A.I., Pozdeev, V.V., and Smolyakov, S.S. (1982) Combust. Explos. Shock Waves, 18 (2), 256–257. Gurevich, M.A., Ozerov, E.S., and Yurinov, A.A. (1978) Combust. Explos. Shock Waves, 14 (4), 448–451. Rat’ko, A.I., Romanenkov, V.E., Bolotnikova, E.V., and Krupen’kova, G.V. (2004) Kinet. Catal., 45 (1), 154–161. Shevchenko, V.G., Kononenko, V.I., Latosh, I.N., Chupova, I.A., and Lukin, N.V. (1994) Combust. Explos. Shock Waves, 30 (5), 635–637. Shevchenko, V.G., Bulatov, M.A., and Kononenko, V.I. (1988) Powder Metall., 2, 1–5. Dossi, S., Reina, A., Maggi, F., and DeLuca, L.T. (2012) Int. J. Energetic Mater. Chem. Propul., 11 (4), 299–322. Maltsev, V.M., Maltsev, M.I., and Kashporov, L.Y. (1977) The Basic Characteristics of Combustion, Chemistry, Moscow. Bucher, P., Yetter, R.A., Dryer, F.L., Vicenzi, E.P., Parr, T.P., and Hanson-Parr, D.M. (1999) Combust. Flame, 117, 351–361. Dreizin, E.L. and Trunov, M.A. (1995) Combust. Flame, 101, 378–382. Dreizin, E.L. (1999) Combust. Flame, 117, 84–85. Yagodnikov, D.A. and Voronetskiy, A.V. (1997) Combust. Explos. Shock Waves, 33 (1), 60–68. Leo, B. and Searcy, A.W. (1951) J. Am. Chem. Soc., 73, 5308–5314. Bocanegra, P.E., Sarou-Kanian, V., Chauveau, C., and Gökalp, I. (2006) 31st International Symposium on Combustion, Heidelberg, Germany, C–03, p. 5675. Arhipov, B.A., Ermakov, B.A., and Razdobreev, A.A. (1982) Combust. Explos. Shock Waves, 18 (2), 16–19.

161

162

5 Combustion Synthesis of AlN (Al3 O3 N), BN, ZrN, and TiN in Air and Ceramic Application 39. Karasev, V.V., Onishchuk, A.A., Glotov,

40.

41.

42.

43.

44.

45.

46. 47.

48. 49.

50. 51. 52.

53.

54.

55. 56. 57.

O.G., Baklanov, A.M., Zarko, V.E., and Panfilov, V.N. (2001) Combust. Explos. Shock Waves, 37 (6), 734–736. Bakhman, N.N. and Belyaev, A.F. (1967) Combustion of Heterogeneous Condensed Systems, Science, Moscow. Orenbakh, M.S. (1973) Reaction Surface in Heterogeneous Combustion, Science, Novosibirsk. Beloshapko, A.G., Bukaemskii, A.A., and Staver, A.M. (1990) Combust. Explos. Shock Waves, 26 (4), 457–461. Dreizin, E.L., Shosin, Y.L., Mudryy, R.S., and Hoffmann, V.K. (2002) Combust. Flame, 130, 381–385. Derevyaga, M.E., Stesik, L.N., and Fedorin, E.A. (1979) Combust. Explos. Shock Waves, 15 (4), 438–446. Bartnitskaya, T.S., Vlasova, M.V., Kosolapova, T.Y., Kostyuk, N.V., and Timofeeva, I.I. (1990) Powder Metall. Met. Ceram., 29 (12), 990–994. Foelsche, R.O., Burton, R.L., and Krier, H. (1991) Combust. Flame, 117, 32–58. Spalding, M.J., Krier, H., and Burton, R.L. (2000) Combust. Flame, 120, 200–210. Yagodnikov, D.A. (1996) Combust. Explos. Shock Waves, 32 (6), 623–636. Ulas, A., Kuo, K.K., and Gotzmer, C. (2001) Combust. Flame, 127, 1935–1957. Yeh, C.L. and Kuo, K.K. (1996) Prog. Energy Combust. Sci., 22, 511–541. Rein, R., Ciezki, H.K., and Eickl, A. (1995) AIAA Pap., 3107, 1–9. Kofstad, P. (1966) High-Temperature Oxidation of Metals, John Wiley & Sons, Inc., New York. Lyutaya, M.D. and Bartnitskaya, T.S. (1970) Izvest. Akad. Nauk SSSR Neorg. Mater., 6 (10), 1753–1757. Dreizin, E.L., Berman, C.H., and Vicenzi, E.P. (2000) Combust. Flame, 122, 30–42. Dreizin, E.L. and Hoffmann, V.K. (1999) Combust. Flame, 118, 262–280. Remy, H. (1956) Treatise on Inorganic Chemistry, Elsevier. Dreizin, E.L., Keil, D.G., Felder, W., and Vicenzi, E.P. (1999) Combust. Flame, 119, 272–290.

58. Molodetsky, E., Vicenzi, E.P., Dreizin,

59.

60. 61. 62. 63.

64. 65. 66.

67.

68.

69. 70. 71.

72.

73. 74. 75. 76.

77.

E.L., and Law, C.K. (1998) Combust. Flame, 112, 522–532. Dreizin, E.L., Hoffmann, V.K., and Vicenzi, E.P. (1999) J. Mater. Res., 14 (10), 3840–3842. Swihart, M.T. and Catoire, L. (2000) Combust. Flame, 130, 210–222. Balat, M. (1996) Calphad, 20 (2), 161–170. Joshi, H.H.S. (1995) Surf. Coat. Technol., 76–77, 499–507. Boborykin, V.M., Gremyachkin, V.M., Istratov, A.G., Kolesnikov-Svinarev, V.I., Kuznetsov, G.P., Leipunskii, O.I., and Puchkov, V.M. (1983) Combust. Explos. Shock Waves, 19 (3), 270–276. Dreizin, E.L. (1996) Combust. Flame, 105, 541–556. Il’in, A.P. and Proskurovskaya, L.T. (1990) Rus. Powder Metal., 9, 32–35. Chatt, J., da Cômara Pina, L.M., and Richards, R.L. (eds) (1980) New Trends in the Chemistry of Nitrogen Fixation, Academic Press, London. Pastirk, I., Comstock, M., and Dantus, M. (2001) Chem. Phys. Lett., 349, 71–78. Rusina, O., Eremenko, A., Frank, G., Strunk, H.P., and Kisch, H. (2001) Angew. Chem. Int. Ed., 40 (21), 3993–3995. Fryzuk, M.D. and Johnson, S.A. (2000) Coord. Chem. Rev., 200–202, 379–409. Hoshino, K. (2001) Chem. Eur. J., 7 (13), 2727–2731. Tuczek, F. and Lehnert, N. (1998) Angew. Chem. Int. Ed., 37 (19), 2636–2638. Desmangles, N., Jenkins, H., Ruppa, K.B., and Gambarotta, S. (1996) Inorg. Chim. Acta, 250, 1–4. Gambarotta, S. (1995) J. Organomet. Chem., 500, 117–126. Suryanarayana, D. (1990) J. Am. Ceram. Soc., 73 (4), 1108–1110. Sternitzke, M. (1993) J. Am. Ceram. Soc., 76 (9), 2289–2294. Dutta, I., Mitra, S., and Rabenberg, L. (1992) J. Am. Ceram. Soc., 75 (11), 3149–3153. Kwon, Y.S., Jung, Y.H., Yavorovsky, N.A., Il’in, A.P., and Kim, J.S. (2001) Scr. Mater., 44, 2247–2251.

References 78. Il’in, A.P. and Proskurovskaya, L.T.

79.

80.

81.

82.

83.

84. 85.

86.

87.

88.

89.

90.

91. 92.

93.

94.

(1990) Combust. Explos. Shock Waves, 26 (2), 190–192. Il’in, A.P., An, V.V., Vereshchagin, V.I., and Yablunovskii, G.V. (2000) Combust. Explos. Shock Waves, 36 (2), 209–212. Il’in, A.P., An, V.V., Vereshchagin, V.I., and Yablunovskii, G.V. (1998) Glass Ceram., 3, 24–25. Pityulin, A.N., Shcherbakov, V.A., Borovinskaya, I.P., and Merzhanov, A.G. (1979) Combust. Explos. Shock Waves, 15 (4), 432–437. Fedotova, T.D., Glotov, O.G., and Zarko, V.E. (2000) Propellants Explos. Pyrotech., 25, 325–332. Guojian, J., Hanrui, Z., Jiong, Z., Meiling, R., Wenlan, L., Fengying, W., and Baolin, Z. (2000) J. Mater. Sci., 35 (1), 63–69. Caceres, P.G. and Schmid, H.K. (1994) J. Am. Ceram. Soc., 77 (4), 977–983. Givarkizov, E.I. (1977) Growth of FiberLike and Plate-Like Crystals from Vapor, Science, Moscow. Il’in, A.P., Bychin, N.V., and Gromov, A.A. (2001) Combust. Explos. Shock Waves, 37, 490–491. Loryan, V.E. and Borovinskaya, I.P. (2003) Combust. Explos. Shock Waves, 39 (5), 525–533. Shin, J., Ahn, D.H., Shin, M.S., and Kim, Y.S. (2000) J. Am. Ceram. Soc., 83, 230–235. Puszynski, J.A., Liebig, B., Dargag, S., and Swiatkiewicz, J. (2003) Int. J. Self Propag. High Temp. Synth., 12, 107–119. Brewal, E., Aghajanian, M.K., Biel, J.P., and Antolin, S. (1993) J. Am. Ceram. Soc., 76, 1865–1868. Atarashiya, K. (1995) J. Mater. Process. Technol., 54, 54–59. Takuya, O., Motohiro, T., and Shuzo, K.J. (2000) J. Eur. Ceram. Soc., 20, 783–787. Bukharin, E.N., Vlasov, A.S., and Alekseev, A.A. (1985) Glass Ceram., 2, 23–25. Chernov, I.P., Tyurin, Y.I., Cherdantzev, Y.P., Kröning, M., and Baumbach, H.

95.

96.

97.

98.

99.

100.

101.

102. 103.

104. 105.

106.

107.

108.

(1999) Int. J. Hydrogen Energy, 24, 359–362. Korshunov, A.V., Il’in, A.P., Radishevskaya, N.I., and Morozova, T.P. (2010) Russ. J. Phys. Chem. A, 84, 1576–1584. Mohan, S., Furet, L., and Dreizin, E.L. (2010) Combust. Flame, 157, 1356–1363. Rat’ko, A.I., Romanenkov, V.E., Bolotnikova, E.V., and Krupen’kina, Z.V. (2004) Kinet. Catal., 45 (1), 156–163. Siracusano, S., Baglio, V., D’Urso, C., Antonucci, V., and Aricò, A.S. (2009) Electrochim. Acta, 54 (26), 6292–6299. Eapen, B.Z., Hoffmann, V.K., Schoenitz, M., and Dreizin, E.L. (2004) Combust. Sci. Technol., 176, 1055–1069. Khomenko, I.O., Khomenko, A.S., Khomenko, V.I., Borovinskaya, I.P., and Merzhanov, A.G. (1993) Combust. Flame, 92 (3), 201–208. Shevchenko, V.G., Kononenko, V.I., Chupova, I.A., Latosh, I.N., and Lukin, N.V. (1999) Combust. Explos. Shock Waves, 35, 77–79. Mei, L. and Li, J.-T. (2008) Acta Mater., 56, 3543–3549. Strokova, Y.I., Gromov, A.A., and Teipel, U. (2006) Abstracts of the International Conference “Chemistry, Chemical Technology and Biotechnology on the Millenium”, Tomsk Polytechnic University, Tomsk, p. 137. Li, J., Gao, L., and Guo, J. (2003) J. Eur. Ceram. Soc., 23 (1), 69–74. Qiao, L., Zhou, H., Xue, H., and Wang, S. (2003) J. Eur. Ceram. Soc., 23 (1), 61–67. Boey, F., Cao, L., Khor, K.A., and Tok, A. (2001) Acta Mater., 49 (6), 3117–3127. Maghsoudipoura, A., Bahrevara, M.A., Heinrichb, J.G., and Moztarzadeha, F. (2005) J. Eur. Ceram. Soc., 7, 1067–1072. Tabarya, P., Servant, C., and Alary, J.A. (2000) J. Eur. Ceram. Soc., 20 (12), 1915–1921.

163

165

6 Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum Chun-Liang Yeh

6.1 Introduction

Nitrides of transition metals of the group VB have been considerably investigated due to their unique chemical and physical properties, such as high melting point, chemical inertness, superior hardness, and metallic electrical conductivity [1, 2]. Besides, vanadium nitride (VN) is a catalytically active material with the catalytic properties similar to those of noble metals [3] and is a promising electrode material for electrochemical supercapacitors [4]. Niobium nitride (NbN) was proven to be an excellent candidate of the field emitter applied to electron beam systems [5] and had exceptional superconducting properties [6, 7]. Tantalum nitrides (TaN and Ta2 N) have been successfully utilized as a stable thin-film resistor [8] and a diffusion barrier between silicon and metal overlayers [9]. Because the reaction enthalpy generated from the interaction of transition metals (especially in groups IVB and VB) with nitrogen is sufficiently high, their corresponding nitrides can be produced by means of combustion synthesis in the mode of self-propagating high-temperature synthesis (SHS) [10, 11]. In the SHS process, once initiated, the combustion front becomes self-sustaining and traverses the entire sample to transform the reactant into the final product progressively without requiring additional heat. With the merits of time and energy savings, combustion synthesis applied not only for solid–solid systems but also for solid–gas and/or solid–liquid systems has been recognized as an attractive alternative to the conventional methods of producing advanced materials, such as borides, carbides, nitrides, hydrides, silicides, intermetallics (such as aluminides and titanides), and various composites on their basis [12–15]. Formation of transition metal nitrides by the SHS process has been mostly conducted in a solid–gas combustion mode. Since there is usually far less of the gaseous reagent in the pores of the sample than is needed for complete nitridation, the gaseous reagent has to be delivered to the combustion zone by infiltration through the porous structure of the solid reactant [10]. Gas infiltration is believed to occur spontaneously on account of the pressure difference caused by absorption of gaseous reagent in the combustion front. The production Nitride Ceramics: Combustion Synthesis, Properties, and Applications, First Edition. Edited by Alexander A. Gromov and Liudmila N. Chukhlomina. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

166

6 Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum

of transition metal nitrides by combustion synthesis can either be accompanied by melting of the metal reactant or nitride product, or it can be completed without formation of any liquid phase. Melting of reactant powders reduces the permeability of the sample and thereby hinders the penetration of nitrogen. It has been suggested that dilution of the metal reactant with the final nitride product is an effective means to lower the sample temperature, to suppress the melting, and then to increase the degree of conversion [16, 17]. Moreover, the addition of nitride powders in the reactant mixture will not cause any variation in the homogeneity of the as-synthesized product. For the situation without formation of any liquid phase, the overall conversion percentage can also be increased by adding the nitride product as a diluent, since the diluent is inert and consumes no nitrogen. This article presents an experimental investigation on combustion synthesis of nitrides of the VB-group elements (vanadium, niobium, and tantalum) in the SHS mode with compacted samples of metal powders in gaseous nitrogen. Combustion characteristics, such as the propagation mode of self-sustaining combustion wave, flame-front velocity, combustion temperature, and afterburning reaction, were studied at length to establish a comprehensive knowledge of the SHS process for the production of transition metal nitrides. Effects of the sample density, nitrogen pressure, diluent concentration, and preheating temperature on the degree of conversion, combustion temperature, and flame-front propagation velocity were explored. In addition, the phase composition of final products obtained under different test conditions was examined.

6.2 Experimental Methods of Approach

Vanadium (V) (Aldrich Chemicals, −325 mesh, 99.5% purity), niobium (Nb) (Strem Chemicals, −325 mesh, 99.8% purity), and tantalum (Ta) (Aldrich Chemical, −325 mesh, 99.9% purity) powders were used as the reactants in this study. Metal powders were pressed into cylindrical specimens with a diameter of 7 mm and a height of 12.5 mm. In order to prepare test samples with different porosities, the powder compacts were formed with a packing density between 45% and 60% relative to the theoretical maximum density (TMD). For the study of the effect of nitride dilution on the SHS reaction, VN and NbN were employed as the diluents to add into the V and Nb powder compacts, respectively. The diluent content ranged from 5% to 30% by weight of the total powder mixture. No TaN dilution was applied for the Ta powder compact, because of no anticipation of melting upon considerably high melting points of Ta (2996 ∘ C) and TaN (3090 ∘ C). The SHS reaction was carried out in a stainless-steel windowed combustion chamber under a nitrogen pressure ranging from 0.27 to 1.82 MPa. The nitrogen gas used in this study had a purity of 99.999%. The sample holder was equipped with a 600 W cartridge heater to rise the initial sample temperature prior to ignition up to 300 ∘ C. The ignition of test samples was accomplished by a heated tungsten coil with a voltage of 60 V and a current of 1.5 A. The propagation

6.3

Results and Discussion

velocity of combustion front was measured by recording the whole combustion event with a video camera. Details of the experimental setup and diagnostic techniques were reported elsewhere [16]. To facilitate the accurate determination of instantaneous locations of the combustion front, a beam splitter (Rolyn Optics), with a mirror characteristic of 75% transmission and 25% reflection, was used to optically superimpose a scale onto the image of the sample compact. The sample temperature during combustion was measured by a 125 μm thermocouple (Pt/Pt–13%Rh) attached on the sample surface. After combustion, the phase composition of burned samples was identified by an X-ray diffractometer with CuKα radiation operating at 40 kV. The conversion percentage (by mole) from the metallic reactant to nitride product was calculated from measurement of the weight change of the sample by assuming a stoichiometric nitride formed as the final product.

6.3 Results and Discussion 6.3.1 Combustion Synthesis of Vanadium Nitride

Figure 6.1a,b shows two recorded SHS sequences of vanadium powder compacts without and with VN dilution, respectively, in nitrogen of 1.14 MPa [18]. It is evident in Figure 6.1 that upon ignition a distinct combustion front forms and travels downward from the ignited top plane in a self-sustaining manner, and transforms the cold reactant into an incandescent combustion product. It was found that the end product in Figure 6.1a was noticeably shrunk. This implies the formation of a liquid phase during the SHS process, due most likely to the reaction temperature inside the sample exceeding the melting point of vanadium (1910 ∘ C). The presence of a liquid phase during the reaction improved the transfer of heat flux from the burned to unburned region, thus leading to the increase of the propagation rate of combustion front. However, it was believed that the substantial melting might diminish the sample porosity and hence could inhibit the infiltration of nitrogen gas. As a result, the nitride conversion percentage could be reduced. Another important SHS characteristic of the VN system is the afterburning phenomenon, which represents that the bulk reaction takes place after the passage of the flame front. The afterburning glow on the sample of Figure 6.1a is clearly observable during the time period from 1.07 to 1.87 s. For the VN-diluted sample of Figure 6.1b, it is interesting to note that no shrinkage and/or deformation of the end product were found; that is, the burned sample essentially retained its original shape. Moreover, when compared with the undiluted sample, combustion luminosity on the burning VN-diluted sample is less intense and the flame-front propagation rate is much lower. These observations could be largely attributed to the decrease of combustion temperature by the addition of VN powders as a diluent in the test sample.

167

168

6 Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum

t = 0.67 s

t = 0.07 s

t = 0.13 s

t = 0.20 s

t = 0.27 s

t = 0.33 s

t = 0.40 s

t = 0.87 s

t = 1.07 s

t = 1.27 s

t = 1.47 s

t = 1.67 s

t = 1.87 s

t = 0.20 s

t = 0.40 s

t = 0.60 s

t = 0.80 s

t = 1.00 s

t = 1.20 s

t = 1.80 s

t = 2.00 s

t = 2.20 s

t = 2.40 s

t = 2.60 s

t = 2.80 s

t = 0.47 s

(a)

t = 1.60 s

t = 1.40 s

(b) Figure 6.1 Recorded combustion images of vanadium powder compacts preheated at 200 ∘ C in nitrogen of 1.14 MPa (a) without VN dilution and (b) with 30-wt% VN dilution.

Figure 6.2a shows the effects of nitrogen pressure and sample density on the flame-front propagation velocity (V f ) of undiluted samples. It was found that the combustion velocity was increased by increasing the nitrogen pressure, because of the increase of the initial nitrogen concentration within the porous sample. As also revealed in Figure 6.2a, the combustion wave velocity increases with decreasing sample density, since the sample with a lower density offers a better permeability for nitrogen gas to penetrate. Moreover, the effect of sample density became more pronounced under nitrogen of a higher pressure. According to Figure 6.2a, at 1.48 MPa of nitrogen, a significant increase in the combustion velocity from about 6.9 to 17.9 mm s−1 is detected upon a decrease in the sample density from 60% to 50% TMD. The influence of diluent content and preheating temperature (T p ) on the flame-front velocity is presented in Figure 6.2b. For the undiluted samples, the flame-front speed increases considerably with preheating temperature, due primarily to excessive melting of the sample [18]. Figure 6.2b indicates the combustion velocity approaching asymptotically to 33 mm s−1 for T p = 150–200 ∘ C. Besides, Figure 6.2b points out that the addition of VN as the diluent to the sample results in a substantial decrease in the flame velocity. For the VN-diluted samples, the increase of the combustion wave velocity with preheating temperature was

6.3

Results and Discussion

20

Flame-front velocity, Vf (mm s−1)

Undiluted samples without preheating Samples density 50% TMD 55% TMD 60% TMD

15

10

5

0 0.00

0.25

0.50

(a)

1.00

1.25

1.50

1.75

Nitrogen pressure (MPa)

40

Flame-front velocity, Vf (mm s−1)

0.75

Undiluted 20 wt% VN-diluted 30 wt% VN-diluted

Samples density : 50% TMD Nitrogen pressure : 1.14 MPa

30

20

10

0 0

(b)

25

50

75

100

125

150

175

200

225

Preheating temperature, Tp (°C)

Figure 6.2 Dependence of flame-front propagation velocity of VN synthesis on (a) nitrogen pressure and sample density and (b) preheating temperature and diluent content.

also found, but the influence of sample preheating on the flame speed gradually diminished as the diluent content augmented. For example, as reported in Figure 6.2b, the flame-front velocity of the 30-wt% VN-diluted sample increases from 2.9 to 5.4 mm s−1 when the initial temperature of the sample increased from 25 to 200 ∘ C. Typical temperature profiles of burning samples are plotted in Figure 6.3, where the effects of VN dilution and preheating temperature are presented. As shown in Figure 6.3, for the undiluted sample (profile #1) the appearance of an abrupt peak signifies the arrival of the flame front propagating at a high speed. The peak temperature reaches about 1780 ∘ C, which stands for the combustion front temperature on the sample surface. It is believed that the reaction temperature

169

6 Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum 2000 1800

Combustion temperature (°C)

170

Sample density : 50% TMD N2 pressure : 1.14 MPa

#1

1600

#3 #2

1400 1200 1000 800

400 200

Tp = 25 °C Undiluted Tp = 25 °C 30 wt% VN-diluted Tp = 200 °C 30 wt% VN-diluted

#1 #2 #3

600

0

1

2

3

4

5

6

7

8

9

10

Time (s)

Figure 6.3 Effects of preheating temperature and diluent content on combustion temperature of VN synthesis.

within the sample is higher than that on the surface, because melting of the sample occurs. Subsequent to the sharp peak, a slight temperature rise followed by a slow decline represents the prolonged afterburning stage. It is obvious that the combustion front temperatures of two VN-diluted samples (i.e., 1346 and 1440 ∘ C for profiles #2 and #3, respectively) are significantly lower than that of the undiluted sample. A higher combustion temperature of profile #3 is attributed to the sample with a preheating temperature of 200 ∘ C. The absence of a sharp rise in temperature for the 30-wt% VN-diluted sample is due probably to the relatively low flame-front propagation velocity. After the passage of the flame front, the temperature profiles of two VN-diluted samples exhibit a fairly plateau region at high temperatures, implying that the reaction continues for a while behind the combustion front. Figure 6.4a shows the molar conversion percentage to form VN as a function of nitrogen pressure and sample density from undiluted samples with no prior heating. The conversion percentage increases with decreasing sample density. This is because compacted samples with a lower density not only possess a larger amount of nitrogen within the porous structure but also provide a higher permeability for nitrogen to infiltrate. However, a slight increase in the yield of VN was found with nitrogen pressure from 0.27 to 0.79 MPa, beyond which the degree of conversion decreased. This was believed to be caused by the fact that at nitrogen pressures above 0.79 MPa, a liquid phase formed in the central portion of the sample; therefore, the continuous penetration of nitrogen gas into the sample was hindered by the molten phase, resulting in the decrease of the conversion percentage. Generally speaking, a low degree of conversion between 45% and 55% by mole was obtained for the undiluted vanadium samples. The dependence of nitride yield content on the preheating temperature and diluent amount is presented in

6.3

Results and Discussion

65 Undiluted samples without preheating Samples density

Conversion to VN (%)

60

50% TMD 55% TMD 60% TMD

55

50

45

40 0.00

0.25

0.50

(a)

0.75

1.00

1.25

1.50

1.75

Nitrogen pressure (MPa)

100 90

Conversion to VN (%)

80 70 60 50 40 Undiluted 20 wt% VN-diluted 30 wt% VN-diluted

30

Samples density : 50% TMD Nitrogen pressure : 1.14 MPa

20 0 (b)

25

50

75

100

125

150

175

200

225

Preheating temperature, Tp (°C)

Figure 6.4 Variations of nitride conversion percentage with (a) nitrogen pressure and sample density, and (b) preheating temperature and diluent content.

Figure 6.4b. The addition of VN powders as the diluent in the sample was found to be very effective in increasing the degree of conversion. For the samples without preheating (T p = 25 ∘ C), the conversion percentage increases from around 51% for the undiluted condition to about 74% for the case with 30-wt% VN dilution. Preheating the undiluted sample led to a decrease in the nitride production because of the formation of excessive liquid phases during combustion. On the contrary, the preheating enhanced the extent of conversion of VN-diluted samples, primarily due to the increase of combustion temperature. As shown in Figure 6.4b, a nitride yield over 90% was obtained from a 30-wt% VN-diluted sample under preheating at 100–200 ∘ C.

171

6 Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum 500

50% TMD, pure V sample without preheating in 1.14 MPa N2

400 V δ-VN β-V2N

Intensity

300

200

100

0 20

30

40

50

60

70

80

2θ

(a)

50% TMD, 30 wt% VN-diluted sample with 100 °C preheating in 1.14 MPa N2

1000

800

Intensity

172

δ-VN

600

400

200

0 20 (b)

30

40

50

60

70

80

2θ

Figure 6.5 XRD patterns of SHS-derived products in 1.14 MPa N2 from (a) an undiluted sample without preheating and (b) a 30 wt% VN-diluted sample with preheating at 100 ∘ C.

The XRD patterns of burned samples under different operating conditions are depicted in Figure 6.5a,b. Figure 6.5a indicates that the constituents of the final product synthesized from an undiluted sample without preheating at 1.14 MPa of nitrogen include two nitrides, δ-VN and β-V2 N, as well as elemental vanadium. The β-V2 N phase was considered as an intermediate compound during the SHS reaction, and its presence was a consequence of insufficient nitrogen especially in the central portion of the sample. The strong intensity of vanadium peaks signifies a large amount of vanadium left unreacted in the end product.

6.3

Results and Discussion

For the product synthesized from a 20-wt% VN-diluted sample with preheating at 100 ∘ C, it was found that δ-VN was noticeably increased and the amounts of β-V2 N and vanadium were substantially decreased. With further increase in the diluent content up to 30 wt%, as shown in Figure 6.5b, the SHS-derived product is essentially an almost fully nitrided compound. 6.3.2 Combustion Synthesis of Niobium Nitride

Figure 6.6 illustrates an image sequence associated with the combustion wave propagation along an undiluted niobium sample in nitrogen of 0.62 MPa [19]. The combustion luminosity on the burned portion of the sample is very strong at the early stage. Subsequently, at about t = 0.8 s, the brightness of the burned portion decreases and the nearly planar combustion front forms a localized reaction zone moving along a spiral trajectory on the sample surface. It was presumed that before appearance of the spinning combustion wave, the thermal energy liberated from both the tungsten coil igniter and the Nb/N2 reaction kept up the propagation of the flame front. When the heat flux from the self-sustaining combustion is insufficient to maintain steady propagation of a planar wave, the combustion front forms one or several localized reaction zones. Besides the spinning combustion wave, the SHS process of the niobium/nitrogen system is characterized by the afterburning reaction. It was believed that the afterburning reaction should play a very important role in the overall conversion of niobium to NbN, since nitridation at the spinning combustion front was mostly confined on the sample surface. Effects of nitrogen pressure and sample density on the flame-front velocity of undiluted niobium samples are presented in Figure 6.7a. It should be noted that the data reported in Figure 6.7a,b are the combustion propagation speed determined in the longitudinal direction. As a result of an increase in the initial nitrogen concentration within the porous sample, the combustion velocity varying between 0.86 and 7.5 mm s−1 increases with nitrogen pressure. Moreover, the flame-front velocity increases with decreasing sample density, since a lower sample density means a better permeability for nitrogen gas to penetrate. When compared with

t = 1.23 s

t = 0.27 s

t = 0.53 s

t = 0.80 s

t = 1.07 s

t = 1.13 s

t = 1.17 s

t = 1.27 s

t = 1.30 s

t = 1.33 s

t = 1.37 s

t = 1.40 s

t = 1.60 s

t = 1.20 s

Figure 6.6 Recorded combustion images of an undiluted niobium powder compact in nitrogen of 0.62 MPa.

173

6 Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum 10

Flame-front velocity, Vf (mm s−1)

Sample density

Undiluted Nb Samples

50% TMD 55% TMD 60% TMD

8

6

4

2

0 0.00

0.25

0.50

(a)

0.75

1.00

1.25

1.50

1.75

Nitrogen pressure (MPa) 5

Flame-front Velocity, Vf (mm s−1)

174

Nitrogen pressure: 1.14 MPa Samples density: 55% TMD 4

3

2

1 0

(b)

5

10

15

20

25

30

NbN dilution content (wt%)

Figure 6.7 Dependence of flame-front propagation velocity of NbN synthesis on (a) nitrogen pressure and sample density and (b) NbN dilution content.

those for the V/N2 system, lower combustion velocities observed in the Nb/N2 system are ascribed to the spinning motion of the combustion wave and no sign of melting for the Nb powder compact. Figure 6.7b shows the decrease of the reaction front velocity from 4.32 to 1.84 mm s−1 with increasing diluent amount from 5 to 25 wt%. This was caused by the fact that the combustion temperature decreases with an increase in the dilution content. It should be noted that for the samples diluted with 30-wt% NbN or higher, combustion was quenched before reaching the self-sustaining regime. Figure 6.8 plots three temperature profiles measured from Nb powder compacts with and without NbN dilution at different nitrogen pressures. The rapid rise in temperature signifies the arrival of the spinning combustion wave and

6.3

Results and Discussion

2000

Combustion temperature (°C)

1800

Sample density: 55% TMD

#3

1600 #2

1400 #1

1200 1000 800 600

#1 1.14 MPa N2

25 wt% NbN-diluted

400

#2 0.45 MPa N2 #3 1.14 MPa N2

Undiluted Undiluted

200

0

1

2

3

4

5

6

7

8

9

10

Time (s)

Figure 6.8 Effects of nitrogen pressure and diluent content on combustion temperature of NbN synthesis.

the subsequent temperature increase followed by a gradual decline represents a prolonged afterburning period. It is important to mention that the sample temperatures during both flame-propagating and afterburning stages range between 1194 and 1874 ∘ C, which are much lower than the melting points of Nb (2468 ∘ C) and NbN (2300 ∘ C). This confirms no formation of molten phases in the NbN synthesis. The temperature of the reaction front, in general, was increased by increasing the nitrogen pressure and the addition of NbN reduced the combustion exothermicity. Figure 6.8 indicates a substantial decrease in combustion temperature by about 480 ∘ C for the sample diluted with 25-wt% NbN. This explains the observation that the self-sustaining combustion cannot be achieved under dilution with 30-wt% NbN or higher. The molar conversion percentage from niobium to NbN for the undiluted samples under nitrogen pressures of 0.27–1.48 MPa is presented in Figure 6.9a. The proportion of the nitride production was varied between 56% and 68%. In general, the sample with a lower density yielded a slightly higher amount of the nitride. No distinct dependence of the yield of NbN on the nitrogen pressure was found. This is mainly attributed to the after burning reaction, in which continuous nitridation takes place after the passage of the flame front. The increase of conversion percentage by the addition of NbN as a diluent is shown in Figure 6.9b for the 55% TMD samples at 1.14 MPa of nitrogen. The molar yield percentage was increased from 58.8% to 71.8% by adding NbN into the sample at around 15–20 wt%. Further increase of NbN led to a decrease in the conversion percentage, due to the decrease of combustion temperature. This means that an optimal increase of about 13% in the degree of conversion was obtained under 15–20 wt% NbN dilution. Because no sample melting occurs in the Nb/N2 system, the improvement in the degree of conversion by nitride dilution for the NbN synthesis is not as pronounced as that for the VN synthesis.

175

6 Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum 80

Sample density

Conversion to NbN (%)

Undiluted Nb Samples

50% TMD 55% TMD 60% TMD

75

70

65

60

55

50 0.00

0.25

0.50

(a)

0.75

1.00

1.25

1.50

1.75

Nitrogen pressure (MPa) 80

75

Conversion to NbN (%)

176

70

65

60

55

Nitrogen pressure: 1.14 MPa Samples density: 55% TMD

50 0 (b)

5

10

15

20

25

30

NbN dilution content (wt%)

Figure 6.9 Variations of nitride conversion percentage with (a) nitrogen pressure and sample density and (b) diluent content.

Two XRD spectra associated with SHS-derived products from the undiluted and 20 wt% NbN-diluted samples are depicted in Figure 6.10a,b, respectively. The XRD pattern of Figure 6.10a is a combination of two nitride phases, δ-NbN and βNb2 N, as well as the unreacted Nb. The δ-NbN phase is the dominant nitride and the existence of β-Nb2 N implies insufficient nitrogen. It was found that conversion of β-Nb2 N to δ-NbN was achieved in the NbN-diluted samples. For a 20 wt% NbNdiluted sample, as revealed in Figure 6.10b, the resulting product contains only one nitride phase δ-NbN along with a minor amount of Nb.

6.3 Pure Nb sample with 50% TMD in 0.79 MPa N2

2000

Nb δ-NbN β-Nb2N

1500

Intensity

Results and Discussion

1000

500

0 20

30

40

(a)

50

60

70

80

2θ 2500

20 wt% NbN-diluted sample with 55% TMD in 1.14 MPa N2

2000 Nb δ-NbN

Intensity

1500

1000

500

0 20

(b)

30

40

50

60

70

80

2θ

Figure 6.10 XRD patterns of SHS-derived products from (a) an undiluted Nb sample in 0.79 MPa N2 and (b) a 20 wt% NbN-diluted sample in 1.14 MPa N2 .

6.3.3 Combustion Synthesis of Tantalum Nitride

Figure 6.11 shows a series of recorded images illustrating propagation of the combustion front associated with an unpreheated tantalum sample reacted in nitrogen of 0.45 MPa [20]. The self-sustaining combustion wave forms a nearly

177

178

6 Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum

t = 2.37 s

t = 0.50 s

t = 0.77 s

t = 1.03 s

t = 1.30 s

t = 1.57 s

t = 1.83 s

t = 2.63 s

t = 2.90 s

t = 3.17 s

t = 3.43 s

t = 3.70 s

t = 3.97 s

t = 2.10 s

Figure 6.11 Recorded combustion images of an unpreheated tantalum powder compact in nitrogen of 0.45 MPa.

planar reaction front traveling in a steady manner. After combustion, the burned sample essentially retains its original shape. As revealed in Figure 6.12a, the flame-front propagation velocity of the unpreheated sample increases with increasing nitrogen pressure. When compared with the V and Nb powder compacts of the same density, the Ta sample exhibited a slower combustion wave. The lower reaction velocity means a longer reaction time for nitridation to proceed at the combustion front. The flame propagation velocity was also increased by decreasing the sample density. On account of the increase of initial sample temperature by preheating, Figure 6.12b indicates a considerable increase in the flame-front velocity. This was caused by the increase of the combustion temperature with sample preheating temperature. According to the measured temperature profiles depicted in Figure 6.13 for Ta samples, the combustion front temperature was found to increase with an increase in nitrogen pressure and preheating temperature. The measured combustion temperatures were in a range from 1470 to 1860 ∘ C under different operating conditions. Because the temperatures were considerably lower than the melting points of Ta (2996 ∘ C) and TaN (3090 ∘ C), there was no tendency to melt the reactant and nitride product during the SHS reaction and no change in the shape of the sample after combustion. The variation of the molar conversion percentage of tantalum to TaN for the unpreheated samples with nitrogen pressure is shown in Figure 6.14a. In contrast to that observed for the NbN synthesis in Figure 6.9a, the degree of conversion to TaN increased appreciably from 45.4% to 78.7% with increasing nitrogen pressure from 0.27 to 1.82 MPa. In view of the fact that the reaction zone of the Nb sample was confined and traveled crosswise and rapidly, it was believed that the afterburning reaction was responsible for the bulk of Nb nitridation, thus resulting in almost no influence of the nitrogen pressure on the yield of NbN. On the other hand, a significant increase in the extent of conversion with nitrogen pressure for the Ta samples suggests that nitridation was largely completed at the combustion front. Moreover, no dependence of the conversion percentage on the sample density was due perhaps to the sufficient reaction time resulting from the slow combustion wave.

6.3

Results and Discussion

8 Undiluted Ta samples without preheating

Flame-front velocity, Vf (mm s−1)

7

Sample density 6

45% TMD 50% TMD

5 4 3 2 1 0 0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

Nitrogen pressure (MPa)

(a) 10

Undiluted Ta Samples

Flame-front velocity, Vf (mm s−1)

9 8 7 6 5 4

Samples density: 55% TMD Nitrogen pressure: 1.48 MPa

3 2 0 (b)

50

100

150

200

250

300

350

Preheating temperature, Tp (°C)

Figure 6.12 Dependence of flame-front propagation velocity of TaN synthesis on (a) nitrogen pressure and sample density and (b) preheating temperature.

The effect of sample preheating on the conversion percentage under a nitrogen pressure of 1.48 MPa is presented in Figure 6.14b. A considerable increase in the yield of TaN by about 16% (from 71.2% to 87.4%) was achieved as the sample initial temperature increased from 25 to 150 ∘ C; yet further increase in the preheating temperature up to 300 ∘ C produced no additional improvement in the degree of conversion. This might be caused by the fact that the combustion front velocity increased significantly with sample preheating temperature, resulting in a shorter retention time of high temperatures behind the flame front for the nitridation to carry on.

179

6 Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum 2000

#3

1800

Combustion temperature (°C)

180

#2

1600 1400

#1

1200 1000

50% TMD Ta samples

800

#1 #2 #3

600 400

0.27 MPa N2 Tp = 25 °C 1.48 MPa N2 Tp = 25 °C 1.48 MPa N2 Tp = 300 °C

200 0

1

2

3

4

5

6

7

8

9

10

Time (s)

Figure 6.13 Effects of nitrogen pressure and preheating temperature on combustion temperature of TaN synthesis.

Two XRD patterns of the final products from Ta powder compacts are plotted in Figure 6.15a,b, identifying the presence of three species, including two nitrides, TaN and Ta2 N, as well as elemental Ta. Due to the low degree of conversion of about 57% at 0.62 MPa of nitrogen, Figure 6.15a shows a great amount of Ta left unreacted. With the increase of nitrogen pressure, a significant increase in TaN along with a decrease in both Ta2 N and Ta were evident. This suggests that TaN is the dominant nitride phase at high pressures of nitrogen. The presence of the intermediate nitride phase Ta2 N and elemental Ta was because of insufficient nitrogen. Figure 6.15b depicts the XRD spectrum of the end product synthesized under nitrogen of 1.48 MPa and preheating of 200 ∘ C, indicative of an enhanced conversion reaching about 86%. It is apparent in Figure 6.15b that TaN predominates over others.

6.4 Conclusions

Combustion synthesis of nitrides of vanadium, niobium, and tantalum in the SHS mode was investigated by adopting compacted samples of elemental powders in gaseous nitrogen. Effects of the sample density (45–60% TMD), nitrogen pressure (0.27–1.82 MPa), diluent concentration (5–30 wt%), and preheating temperature (50–300 ∘ C) on the combustion sustainability, flame-front propagation velocity, combustion temperature, and degree of nitridation were studied. For all three reaction systems, the propagation velocity of the self-sustaining combustion wave increased with nitrogen pressure, but decreased with sample compaction density. Heating the green samples prior to ignition also led to

6.4

Conclusions

90

Undiluted Ta samples without preheating Sample density

Conversion to TaN (%)

80

45% TMD 50% TMD 70

60

50

40 0.00

0.25

0.50

0.75

(a)

1.00

1.25

1.50

1.75

2.00

Nitrogen pressure (MPa) 95

Undiluted Ta samples

Conversion to TaN (%)

90

85

80

75

Sample density: 50% TMD Nitrogen pressure: 1.48 MPa

70 0

(b)

50

100

150

200

250

300

350

Preheating temperature, Tp (°C)

Figure 6.14 Variations of nitride conversion percentage with (a) nitrogen pressure and sample density and (b) preheating temperature.

an increase in the flame-front speed, but diluting the samples with nitride powders caused a decrease in the combustion velocity. Variations of combustion temperature with nitrogen pressure, sample density, preheating temperature, and diluent content were in a manner consistent with those of combustion wave velocity. For the V/N2 reaction system, a significant enhancement of the conversion percentage was achieved for the VN-diluted vanadium compact, due mainly to melting of the undiluted samples. Preheating also improved the nitridation of the VN-diluted sample, thus leading to a nitride yield over 90% obtained from a 30-wt% VN-diluted sample under preheating at 100–200 ∘ C. For the Nb/N2

181

6 Combustion Synthesis of Nitrides of Vanadium, Niobium, and Tantalum 3500 3000 Ta sample: 50% TMD N2 pressure: 0.62 MPa.

2500

Ta TaN Ta2N

Intensity

2000 1500 1000 500 0 30

40

50

60

70

80

2θ

(a) 3000

Ta sample: 50% TMD N2 pressure: 1.48 MPa Preheating: 200 °C

2500 2000 Intensity

182

Ta TaN Ta2N

1500 1000 500 0 30

(b)

40

60

50

70

80

2θ

Figure 6.15 XRD patterns of SHS-derived products from (a) an unpreheated Ta sample in 0.62 MPa N2 and (b) a 200 ∘ C-preheated Ta sample in 1.48 MPa N2 .

reaction system, the self-sustaining combustion wave was locally confined on the sample surface and propagated in a spinning mode. It was believed that the afterburning reaction contributed largely to the nitridation of Nb, because the conversion percentage varying between 56% and 68% showed almost no dependence on the nitrogen pressure and sample density. Moreover, an improvement of about 13% was obtained in the degree of conversion by nitride dilution for the NbN synthesis. For the Ta/N2 reaction system, the nitridation of Ta was mostly completed at the combustion front. As a result, the extent of conversion increased substantially with increasing nitrogen pressure. Preheating the sample

References

also enhanced the yield of TaN. A conversion percentage of 80% was attained for the unpreheated Ta sample and an increase by about 16% was achieved upon preheating the green compact at 150 ∘ C. References 1. Toth, L.E. (1971) Transition Metal Car-

2.

3.

4.

5.

6.

7.

8.

9.

bides and Nitrides, Academic Press, New York. Oyama, S.T. (ed.) (1996) The Chemistry of Transition Metal Carbides and Nitrides, Blackie Academic & Professional. Choi, J.G., Ha, J., and Hong, J.W. (1998) Synthesis and catalytic properties of vanadium interstitial compounds. Appl. Catal. Gen., 168, 47–56. Zhou, X., Chen, H., Shu, D., He, C., and Nan, J. (2009) Study on the electrochemical behavior of vanadium nitride as a promising supercapacitor material. J. Phys. Chem. Solid, 70, 495–500. Saito, Y., Kawata, S., Nakane, H., and Adachi, H. (1999) Emission characteristics of niobium nitride field emitters. Appl. Surf. Sci., 146, 177–181. Zhen, W., Kawakami, A., Uzawa, Y., and Komiyama, B. (1996) Superconducting properties and crystal structures of single-crystal niobium nitride thin films deposited at ambient substrate temperature. J. Appl. Phys., 79 (10), 7837–7842. Linde, A.V., Grachev, V.V., and Marin-Ayral, R.M. (2009) Self-propagating high-temperature synthesis of cubic niobium nitride under high pressures of nitrogen. Chem. Eng. J., 155, 542–547. Cuong, N.D., Kim, D.J., Kang, B.D., Kim, C.S., Yu, K.M., and Yoon, S.G. (2006) Characterization of tantalum nitride thin films deposited on SiO2 /Si substrates using dc magnetron sputtering for thin film resistors. J. Electrochem. Soc., 153 (2), G164–G167. Chen, G.S., Huang, S.C., Chen, S.T., Yang, T.J., Lee, P.Y., Jou, J.H., and Lin, T.C. (2000) An optimal quasisuperlattice design to further improve thermal stability of tantalum nitride diffusion barriers. Appl. Phys. Lett., 76 (20), 2895–2897.

10. Eslamloo-Grami, M. and Munir, Z.A.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

(1990) Effect of porosity on the combustion synthesis of titanium nitride. J. Am. Ceram. Soc., 73, 1235–1239. Agrafiotis, C.C., Puszynski, J.A., and Hlavacek, V. (1991) Experimental study on the synthesis of titanium and tantalum nitrides in the self-propagating regime. Combust. Sci. Technol., 76, 187–218. Merzhanov, A.G. (1996) Combustion processes that synthesize materials. J. Mater. Process. Technol., 56, 222–241. Mossino, P. (2004) Some aspects in selfpropagating high-temperature synthesis. Ceram. Int., 30, 311–332. Morsi, K. (2012) The diversity of combustion synthesis processing: a review. J. Mater. Sci., 47, 68–92. Liu, G., Li, J., and Chen, K. (2013) Combustion synthesis of refractory and hard materials: a review. Int. J. Refract. Met. Hard Mater., 39, 90–102. Yeh, C.L. and Chuang, H.C. (2004) Combustion characteristics of SHS process of titanium nitride with TiN dilution. Ceram. Int., 30 (5), 705–714. Eslamloo-Grami, M. and Munir, Z.A. (1990) Effect of nitrogen pressure and diluent content on the combustion synthesis of titanium nitride. J. Am. Ceram. Soc., 73, 2222–2227. Yeh, C.L., Chuang, H.C., Liu, E.W., and Chang, Y.C. (2005) Effects of dilution and preheating on SHS of vanadium nitride. Ceram. Int., 31 (1), 95–104. Yeh, C.L. and Chuang, H.C. (2004) Experimental studies on self-propagating combustion synthesis of niobium nitride. Ceram. Int., 30 (5), 733–743. Yeh, C.L., Liu, E.W., and Chang, Y.C. (2004) Effect of preheating on synthesis of tantalum nitride by self-propagating combustion. J. Eur. Ceram. Soc., 24, 3807–3815.

183

185

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen Liudmila N. Chukhlomina, Yury M. Maksimov, and Lidiya N. Skvortsova

7.1 Introduction

The Self-propagating high-temperature synthesis (SHS) method discovered by Merzhanov, Borovinskaya, and Shkiro [1–3] in 1967 is increasingly used for producing various compounds including nitrides [4–8]. Traditional SHS techniques assume using powders of chemical elements (Si, B, Ti, Al, Cr, Ta, V, Nb, etc.). However, ferroalloys – inexpensive, accessible, unpurified raw materials – can be used to synthesize nitrides and nitride-containing composition materials. The ferroalloys industry is one that generates multi-tonnage produce volumes. Fine dust (cyclone dust) produced in a process of crush of ferroalloy ingots is also an ideal raw material for SHS. At the same time, the research of laws of ferrosilicon combustion in nitrogen is a separate and independent scientific problem. Combustion of alloys in nitrogen can differ noticeably from combustion of metal. Alloys are characterized by higher diversity of properties since chemical elements of alloys interacting with each other form solid solutions and/or chemical compounds. Interactions in such systems have a range of specifics such as dependence of reagent contents at interphase boundaries on diffusion coefficients, presence of nonmetallic phases which undergo phase transformations in a process of combustion in alloys, different affinity of alloy components with other reagents, presence of eutectics, and so on. Physicochemical processes of nitride formation in a combustion wave of multicomponent systems, processes of forming of structure, and phase composition of composite materials remain insufficiently explored up to date and very interesting for further studies and applications.

Nitride Ceramics: Combustion Synthesis, Properties, and Applications, First Edition. Edited by Alexander A. Gromov and Liudmila N. Chukhlomina. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

186

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

7.2 Synthesis of Silicon Nitride by Combustion of Ferrosilicon in Nitrogen 7.2.1 Nitriding Degree of Combustion Products and Burning Rate Versus Main SHS Parameters

Ferrosilicon is referred to the main group of ferroalloys. Silicon contents in the alloy vary widely and cover practically the whole concentration line of the binary iron–silicon phase diagram. Silicon interfuses with iron in all proportions and forms a range of silicides: Fe3 Si (14%1) Si), Fe5 Si3 (23% Si), FeSi (33% Si), and FeSi2 (50% Si). The peculiarity of ferrosilicon is fragility and tendency to dispersion. That is why coarse crushing of ferrosilicon ingots leads to formation of fine-dispersed fraction or technical (cyclone) dust that is collected by the ventilation system. The cyclone dust has particles less than 160 μm in size and contains about 80% of silicon. Additional impurity elements are (no more than): C (0.1%), S (0.02%), P (0.03%), Al (3.0%), Mn (0.3%), and Cr (0.2%). The industrial ferrosilicon dust contains two phases: Si and FeSi2 (Figure 7.1). Combustion of ferrosilicon in nitrogen is carried out in the filtration mode since the content of reactant gas in the pores of the sample is insufficient for propagation of the combustion wave. In accordance with calculations, the nitriding degree does not exceed 0.13 for the case when combustion is carried out only due to nitrogen located in the pores of the sample. The nitriding degree is defined as the ratio of amount of chemically bounded nitrogen in a sample after a process of combustion to amount of nitrogen needed for a full conversion of nitride-forming elements into highest nitrides. Consequently, combustion of ferrosilicon is impossible at 1

2 2

16 20 24 28 32 36 40 44 48 52

2θ Figure 7.1

XRD pattern fragment of original ferrosilicon 1 – Si and 2 – FeSi2 .

1) Hereafter, % means mass percent if not specified otherwise.

7.2 Synthesis of Silicon Nitride by Combustion of Ferrosilicon in Nitrogen

pressures lower than 10 MPa. Delivery of reactant gas to the combustion wave is carried out by gas filtration. Under conditions of filtration combustion, important factors influencing the process are: nitrogen pressure, diameter of the sample, and relative density of the sample. Figure 7.2 shows dependence of nitriding degree (m) on nitrogen pressure. Nitriding degree was less than 1.0 within all studied ranges of pressures. The last circumstance is connected with the significant amount of a liquid phase produced during combustion of ferrosilicon in nitrogen at temperatures of 1210–1250 ∘ C. The liquid phase “seals” pores of the sample and the process of nitriding becomes impossible. During combustion of ferrosilicon in nitrogen, dependence of burning rate on nitrogen pressure (Figure 7.2, curve 2) qualitatively repeats the similar dependence for combustion of silicon in nitrogen. The increase in pressure leads to the increase in filtration rate of nitrogen toward the reaction zone which accelerates the reaction at the boundary “solid phase–gas.” Dependence of nitriding degree on sample diameter has an extremum (Figure 7.3). For the ferrosilicon used for the study (containing 80.1% of silicon), the maximum of nitriding degree is achieved with a sample diameter of 50 mm. The increase over this diameter reduces nitriding degree and this dependence is caused by decrease in heat loss which leads to rapid melting of the sample and decrease in reaction surface. The decrease in diameter leads to the increase in heat loss caused by radiation which leads to the decrease in temperature and reaction rate. The burning rate increases with the increase of the sample diameter due to the heat loss caused by radiation decreases. Heat loss depends on the ratio between the surface area and volume of the sample. The increase in diameter of the sample leads to the decrease in that ratio. The increase in initial density of samples leads to the decrease in both nitriding degree and burning rate (Figure 7.4). The increase in density prevents filtration delivery of nitrogen to the reaction zone. That is why we used not-pressed samples (powders) to obtain highly nitrided combustion products.

−1

m

U (mm s )

1.0

0.4

1

0.8

0.3

0.6

2

0.2

0.4 0.1

0.2 0.0

0

2

4 6 8 Nitrogen pressure (MPa)

10

0.0

Figure 7.2 Dependence of nitriding degree (m; 1) and burning rate (U; 2) of ferrosilicon on nitrogen pressure.

187

188

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

U (mm s−1)

m 1.0

0.20

0.9

0.18

2 1

0.8

0.16

No combustion

0.7

0.14

0.6

0.12

0.5

20

30

40

50

60

70

0.10

Diameter (mm) Figure 7.3 Dependence of nitriding degree (m; 1) and burning rate (U; 2) of ferrosilicon on sample diameter.

U (mm s−1)

m 1.0

0.20

2 0.18

0.9

0.8

0.16

1

0.7

0.14

0.6

0.12

0.5 0.3

0.4

0.5

0.10 0.6

Relative density of sample (δ) Figure 7.4 Dependence of nitriding degree (m; 1) and burning rate (U; 2) of ferrosilicon on relative density of sample.

7.2.2 Filtration Combustion Modes of Ferrosilicon in Nitrogen

The combustion of ferrosilicon in nitrogen is carried out in the self-oscillating mode but not in the stationary one. At the same time, harmonic oscillations are transformed into relaxation ones which consist of alternating flashes and depression. Propagation of a combustion wave is carried out as follows: a bright flash is observed after the initiation of the reaction and then intensity of the flush gradually decreases and eventually ceases. Then, at some distance downward from

7.2 Synthesis of Silicon Nitride by Combustion of Ferrosilicon in Nitrogen

Figure 7.5 A photograph of a sample of a product of combustion of ferrosilicon in nitrogen in a self-oscillation mode (longitudinal chip). Each light layer is enriched with silicon nitride; each dark layer is enriched with iron silicides.

the extinguished wave, a bright band appears. In general, the combustion process of ferrosilicon goes on according to the following scheme: combustion of heated layer → heating of new layer → ignition → combustion of heated layer, and so on. Figure 7.5 shows a longitudinal chip of a sample of product of combustion of ferrosilicon in nitrogen. The external dimensions and shape of the burnt sample remain unchanged but it is clearly seen that the sample consists of alternating layers which are easily separated from each other. Results of layer-by-layer chemical analyses of the burnt sample testify different nitriding degrees of light and dark layers. The layer enriched with silicon nitride is formed during a flash; the heat released at the flash is sufficient to melt a layer of initial ferrosilicon adjacent to the reaction zone where the reaction proceeds in a depressive mode due to reduction of the reaction surface of ferrosilicon caused by melting and coagulation of liquid particles of the initial alloy. Figure 7.6a,b shows the temperature history of ferrosilicon combustion in nitrogen: two thermocouples are located apart from each other. The maximum combustion temperatures are nearly the same (1990–2000 ∘ C). However, the speed and temperature of the process are changed differently for different points of the burning sample. Significant fluctuations of burning rate and temperature in the wave (Figure 7.6b) may be a sign that the system comes out from “depression” and subsequent combustion of the sample in the longitudinal direction resumes. At the same time, a relay (scintillation) reaction mode is observed when hot spots arise locally along the combustion wave. Then those spots relax and initiate formation of new hot spots. Combustion of the heated layer is comprised of series of increasing flashes which is reflected on the temperature profile of the combustion wave (Figure 7.6b). Thus, the unstable (self-oscillating) combustion of ferrosilicon in nitrogen is accompanied by a low nitriding degree of combustion products and leads to their chemical and phase inhomogeneity. SHS process control is required to achieve the phase and chemical homogeneity of SHS products and the maximum (m = 1)

189

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

2000

2000

Combustion temperature (°C)

Combustion temperature (°C)

190

1500 1000 500

2000

100

200

300

400

500

1795

1595

1000 560 500 507 0

600

100

200

(b)

Time (s)

1990

1500

0

(a)

1890

300

400

500

600

Time (s)

Figure 7.6 Temperature history of ferrosilicon combustion in nitrogen (PN = 4 MPa, 2 d = 50 mm). (a) Thermocouple is located in a light layer and (b) thermocouple is located in a dark layer.

nitriding degree. Dilution of the initial alloy with a final combustion product is a universal and well-known method used in furnace synthesis and SHS. 7.2.3 Dilution of Initial Ferrosilicon by Previously Nitrided Ferrosilicon

Dependence of conversion degree on the amount of additive (previously nitrided ferrosilicon) is shown in Figure 7.7. A burned sample with additive of 50% previously nitrided ferrosilicon is shown in Figure 7.8. The peculiarity of ferrosilicon combustion in nitrogen with dilution of raw material by the final product is the U (mm s−1)

m

0.7

1

1.0 0.9

0.6

2

0.8

0.5

0.7

0.4

0.6 0.3 0.5 0.2

0.4

0.1

0.3 0.2 0

10

20

30

40

50

60

70

80

0.0

Additive (mass %) Figure 7.7 Dependence of nitriding degree (m; 1) and burning rate (U; 2) of ferrosilicon on additive percentage (P = 4 MPa).

7.2 Synthesis of Silicon Nitride by Combustion of Ferrosilicon in Nitrogen

Figure 7.8

Burned sample with additive of 50% of previously nitrided ferrosilicon.

increase in burning rate with increase in the amount of additive. Studies have shown that dilution of raw material by nitrided ferrosilicon prevents the formation of large particles by melting of particles of the raw material and their coagulation. This ensures that typical size of particles participating in reaction is not increased in comparison to particles of the initial mixture. Absence of coagulation after dilution of initial ferrosilicon by the final product (previously nitride ferrosilicon) leads to increase in reaction surface and consequently burning rate. The combustion temperature measured for a sample with addition of 60% of nitrided ferrosilicon (nitrogen pressure 4 MPa) varies within the volume of the sample. The maximum combustion temperature observed inside the sample was 2100 ∘ C. The experimental studies have shown [9–11] that ammonium salts are effective additives in production of nitrides. According to [7], silicon nitride is formed through intermediate compounds (tetrachloride and silicon imide) in the presence of ammonium chloride. Nitriding of ferrosilicon (as well as silicone) is a complex multistage process which is carried out through the formation of intermediate products such as SiCl4 and Si(NH)2 . This influences kinetic characteristics of the process. Burning rate decreases with increase in the amount of ammonium chloride. The optimal rate corresponding to the maximum conversion degree at nitrogen pressure 4 MPa is 0.12 mm s−1 . Experiments have shown that the size of particles of initial ferrosilicon is a very important factor for its nitriding in presence of NH4 Cl. The small particle size helps in conversion of silicon into gas phase and directs the synthesis mechanism through intermediate products. The use of fine powders allowed extending the limits of combustion and increasing the amount of additive, which, in turn, allowed to influence temperature and burning rate within a wide range of ratios “amount of NH4 Cl/nitrogen pressure.” Nitriding degree of product of ferrosilicon combustion is greatly influenced by fluorine-containing additives: fluorides of ammonium and magnesium. The influence of ammonium fluoride on laws of ferrosilicon combustion is similar to the influence of ammonium chloride. 1.0% MgF2

191

192

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

additive to the initial ferrosilicon [12] extends the limits of ferrosilicon combustion up to 0.4 MPa (without MgF2 , minimum pressure to make combustion possible was 1 MPa). In this case, activation of the nitriding process is connected with the ability of fluorides of alkaline and alkaline-earth metals to remove oxide films from the surface of the reacting particles. The studies on influence of additives on laws of ferrosilicon nitriding have shown that the maximum degree of nitriding during combustion of ferrosilicon in nitrogen can be achieved not only by adding silicon nitride or nitrided ferrosilicon into raw materials but also by adding small amounts of ammonium salts or magnesium fluoride. 7.2.4 Influence of Content of Silicon in Initial Fe–Si Alloys on SHS of Silicon Nitride

To compare nitriding processes of Fe–Si alloys, we studied the alloys with different silicon content: 80.0% (FS 75), 66.7% (FS 65), and 45.0% (FS 45). The alloys studied have different phase compositions. According to the results of XRD analyses, industrial alloys FS 75 and FS 65 are two-phase materials consisting of silicon and leboite (FeSi2 ) and FS 45 contains only compounds of iron with silicon (FeSi2 and FeSi). According to the Fe–Si phase diagram, the FeSi2 –Si eutectic is melted at temperature of 1206 ∘ C, leboite (FeSi2 ) is melted at 1220 ∘ C, and the FeSi2 –FeSi eutectic at 1212 ∘ C. In general, the temperature of full melting of different types of ferrosilicon depends on the content of silicon. Thus, the melting temperature range for FS 65 with a silicon content of 63–68% is 1210–1250 ∘ C and for FS 75 and FS 90 with a silicon content of 74–80% and more than 89%, respectively, is 1210–1315 ∘ C. The melting temperature of silicon is 1415 ∘ C. The nitriding degree of FS 65 combustion products remain low (0.3–0.7) for the nitrogen pressure intervals of 5–7 MPa [13]. Initial ferrosilicon with the size of particles less than 160 μm do not burn at pressure less than 5 MPa in the constant-pressure apparatus. Despite the nonstationary propagation mode of combustion wave, combustion of the FS 75 alloy is realized within wider pressure range (1–10 MPa). Such significant difference in processes of SHS nitriding of these two alloys is determined by their phase composition. Despite the fact that the total content of silicon in alloy FS 65 is sufficient to achieve a maximum possible combustion temperature, its content in the free form is insufficient for the nitriding process in a self-propagating mode at nitrogen pressure up to 5 MPa. Additionally, the relative content of the melt for FS 65 is considerably higher than for FS 75, which creates additional filtration difficulties for delivery of nitrogen in the reaction zone. XRD analyses of samples burned at P > 5 MPa showed that components of the initial alloy (FeSi2 and Si) as well as iron mono-silicide and α-Fe (Figure 7.9) are present in combustion products of FS 65 along with main phases of Si3 N4 (α- and β-modifications). The results of the scanning electron microscopy (SEM) investigation and EDX of combustion products showed that the silicon nitride crystals grow on the surface of the iron–silicon melt (Figure 7.10a), and there are quite extended areas of the iron–silicon melt (Figure 7.10b) with the composition that corresponds to

7.2 Synthesis of Silicon Nitride by Combustion of Ferrosilicon in Nitrogen

1

1

1

3

(c)

22 1

1

4

1

5 4

6 (b)

2

1 4 1

2

1

1

5

3

22

1

1 2

2

1

4 4

2

1

6 6

4 4

4

4 (a)

30

35

40

45 2θ

50

55

60

Figure 7.9 XRD pattern fragment of raw ferrosilicon FS 65 (a), combustion product (b), and combustion product in the presence of ammonium fluoride (c) 1-β-Si3 N4 , 2-α-Si3 N4 , 3-α-Fe, 4-FeSi2 , 5-FeSi, and 6-Si.

the chemical composition of the initial alloy. Chemical stimulation of the process of nitride formation by means of ammonium fluoride additives increased nitriding degree of combustion products up to 0.8–0.85. The maximum degree of nitriding (m = 1.0) for FS 65 is achieved in the presence of the complex admixture “NH4 F + preliminary nitrided ferrosilicon” [12, 13]. Samples burnt in the presence of the 40–45% complex admixture are homogeneous throughout the volume and consist of silicon nitride and α-Fe. The action mechanism of the complex admixture is conditional on the elimination of filtration difficulties by adding a refractory component (preliminary nitrided FeSi) on the one hand and the acceleration of nitride formation due to the transition of reaction into the gas phase through the formation of SiF4 , on the other hand. Combustion of FS 45 alloy is possible with use of fine powders (less than 10 μm) and high nitrogen pressure. To understand the process of ferrosilicon nitriding in the combustion mode, the method of stopping the combustion wave by rapid nitrogen pressure release and filling the apparatus by argon was used. The microstructure analysis of quenched samples with the use of X-ray microanalysis and SEM analysis allowed us to

193

194

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

(a)

(b) Figure 7.10 SEM images of fracture of the combustion product of alloy FS 65 in nitrogen in the presence of ammonium fluoride. (a) Si3 N4 crystals on the surface of the iron–silicon melt and (b) iron–silicon melt.

find the areas with characteristic structural peculiarities. In the area that is most distant from the quenched wave (I), ferrosilicon particles were not subjected to significant changes and their morphology was not different from the morphology of initial ferrosilicon particles (Figure 7.11a–c). Area II is characterized by the presence of molten particles. Sizes of these particles increase toward the quenched wave. In this zone, the molten particles coalesce and form reaction cells. Sizes of the reaction cells are several times greater than initial particles sizes of initial ferrosilicon (Figure 7.11d–f ). In the same area, beginning of formation of silicon nitride (dark-gray boundary around the particle) is observed. It is evident that the reaction of the silicon nitride formation proceeds on the surface of the reaction cell and covers it with a dense layer.

7.2 Synthesis of Silicon Nitride by Combustion of Ferrosilicon in Nitrogen

*3 *1 *2

(a)

15 μm

(d)

(b)

(e)

(c)

(f)

25 μm

Figure 7.11 EDX-SEM data: (a–c) microstructure of raw ferrosilicon (area I), (b) Kα -Si, c-Kα Fe; (d–f ) microstructure of the sample at the initial stage of ferrosilicon nitriding (area II), (e) Kα -Si, and (f ) Kα -Fe; *1 – Si; *2 – FeSi2 ; and *3 – Si3 N4 .

7.2.5 Mechanism of Structure and Phase Formation of Silicon Nitride during Combustion of Ferrosilicon in Nitrogen

In system Si–N, one chemical compound exists in two modifications (α- and β-SiN4 ) that differ from each other in location of SiN4 tetrahedrons along axis C. Both modifications crystallize in the hexagonal system. The cubic modification c-Si3 N4 [14] was obtained recently [15]. After combustion of ferrosilicon in

195

196

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

Table 7.1 Crystallographic characteristics of polymorphic modifications of silicon nitride. Phase

Type of crystal lattice

α-Si3 N4 β-Si3 N4 γ-Si3 N4 δ-Si3 N4

Hexagonal Hexagonal Tetragonal Orthorhombic

Parameters of crystal lattice (nm) a

b

c

0.7753 0.7606 0.9245 1.3380

— — — 0.860

0.5618 0.2909 0.8482 0.7740

nitrogen, polymorph modifications of silicon nitride with tetragonal (γ) and orthorhombic (δ) crystal lattices along with universally recognized α and β modifications were observed in SHS-derived products by XRD [16] (Table 7.1). The results of XRD showed that the orthorhombic and tetragonal modifications of silicon nitride exist in a relationship with iron silicide of similar crystal modifications. According to XRD analyses results, initial ferrosilicon consists of tetragonal modification of iron disilicide (α-FeSi2 ) and silicon. In accordance with the data [17], the following transformations can take place between α-FeSi2 high-temperature and β-FeSi2 low-temperature modifications (Equations 7.1–7.3). α-FeSi2 → β-FeSi2 + Si

(7.1)

FeSi + α-FeSi2 → β-FeSi2

(7.2)

Si + FeSi → β-FeSi2

(7.3)

Therefore, not only the direct decomposition reaction of α-FeSi2 is possible but also reactions existing due to the formation of FeSi in the combustion wave as a result of high-temperature dissociation of FeSi2 . The formation of silicon nitride with tetragonal and orthorhombic lattices is caused by nitriding of low and high temperature modifications of iron disilicide, respectively, based on the results of transmission electronic microscopy (TEM) studies combined with EDX analysis. Analysis of literature data has shown that γ and δ modifications of silicon nitride are not formed during interaction of the silicon powder and nitrogen both in SHS mode and thermal synthesis. At the same time, the study [18] showed that silicon nitride has a structure different from α-Si3 N4 and β-Si3 N4 after nitriding of alloys of silicon with iron, nickel, or manganese. The authors associate it with the formation of complex nitrides Mex Siy Nz in alloys. Complex triple phases based on Fe–Si–N were not found in this study. The process of nitride formation during combustion of ferrosilicon in nitrogen can generally be described by four mechanisms: “solid–gas,” “liquid–gas,” “vapor–gas,” and “vapor–liquid–crystal” (VLC). Experimentally measured combustion temperatures (1900–2100 ∘ C) indicate that silicon will be either in the liquid state as an iron–silicon melt or in the gaseous state during SHS. Under the

7.2 Synthesis of Silicon Nitride by Combustion of Ferrosilicon in Nitrogen

nitride formation from an iron–silicon melt (“liquid–gas”), the solubility of nitrogen in liquid silicon is low (0.01% at 1450 ∘ C). The solubility of nitrogen in liquid iron is slightly higher than in silicon but does not exceed 0.044% at 1760 ∘ C. The increase in nitrogen pressure above liquid melt by the Sieverts law (Equation 7.4) leads to the increase in equilibrium content of nitrogen in liquid melt. Gaseous nitrogen dissolves in the Fe–Si alloy (Equation 7.5) and reacts with silicon to form silicon nitride (Equation 7.6). √ [N] = f ( PN2 ) (7.4) 1∕2 N2 → NFe−Si

(7.5)

3 Si(in alloy) + 4 N(in alloy) → Si3 N4 ↓(solid)

(7.6)

A system where a sparingly soluble compound exists in equilibrium with a saturated solution of this compound in the silicon–iron melt is formed. Under equilibrium conditions, the product of nitrogen and silicon activities in degrees equal to their stoichiometric coefficients is the solubility product (PSSi3 N4 ) of silicon nitride in iron. The solubility of Si3 N4 in ferrosilicon is low (4.5 × 10−10 at 1000 ∘ C) and the equilibrium concentration of nitrogen [N] = 0.0019%. This implies that small concentrations of dissolved nitrogen are sufficient for the formation of silicon nitride in liquid ferrosilicon. Si3 N4 forms a solid phase and is constantly removed from the system, which is confirmed by SEM investigations of the microstructure of combustion products (Figure 7.12). As can be seen from the figure, a droplet of the iron–silicon melt has a developed surface with numerous nuclei and growing crystals of silicon nitride. Condensed silicon nitride is produced when the reaction products of silicon and nitrogen in the melt exceeds PSSi3 N4 at given temperature. As soon as the reaction product becomes equal to PSSi3 N4 , the process of silicon nitride formation stops. Amount of nitrogen from a melt is compensated by gaseous nitrogen in accordance with the solubility of nitrogen in the Fe–Si alloy for a given temperature and pressure. When the amount of newly dissolved nitrogen exceeds PSSi3 N4 again,

(a)

(b)

Figure 7.12 SEM image of products of ferrosilicon combustion in nitrogen. (a) Solidified drop of iron-silicon melt with crystals of silicon nitride and (b) upper half of the same drop with “flowing down” crystals of silicon nitride forming a large crystal (in the background).

197

198

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

the melt becomes supersaturated relative to N and Si and the next portion of silicon nitride is formed. Since the concentration of nitrogen in the melt constantly increases due to gaseous nitrogen, supersaturation is achieved again at a certain distance from the already produced layer and the process of producing silicon nitride is repeated. Analysis of electron microscopic studies of SHS products allows asserting that the mechanism of silicon nitride formation by crystallization from an iron–silicon melt is the basic one [14]. At the T > 1500 ∘ C, the equilibrium vapor pressure of silicon becomes significant since silicon is evaporated from the eutectic melt. Additionally, the silicon particles with sizes less than 10 μm turn into the gas phase during the SHS due to self-evaporation. The process of nitride formation from the gas phase can be carried out by two mechanisms “vapor–gas” and “VLC”. In the first case, condensation from the gaseous phase goes directly to the solid surface. In the second case, it goes through a thin layer of the liquid phase in which a crystalline material is dissolved. A sufficient indication that the crystal grows in accordance with VLC mechanism is presence of hemispherical particles (“globules”) at the top of the whisker crystal. Characteristic globules are shown in Figure 7.13. In accordance with the VLC mechanism, depending on the crystallization conditions (temperature, supersaturation, etc.), it is possible to convert whisker crystals into scaly ones which are formed due to anisotropic growth of lateral faces of the whiskers (Figure 7.14a). Supersaturation increases with increase in synthesis temperature and aggregation of crystals begins at certain level of supersaturation. The formation of twinned crystals occurs on the developed faces of crystals. This was observed in combustion products of ferrosilicon (Figure 7.14b). Growth of silicon nitride crystals in accordance with the “vapor–gas” mechanism during combustion of ferrosilicon in nitrogen is quite possible as well. The SEM images (Figure 7.14c,d) show two types of crystals: prisms and spherical polyhedrons. Polyhedrons are likely formed in accordance with the “vapor–gas” mechanism after the formation of large silicon nitride crystals with correct crystallographic

(a)

(b)

Figure 7.13 SEM images of combustion products of ferrosilicon in nitrogen. (a, b) Whisker crystals of silicon nitride (arrows point to globules at tops of crystal growing by VLC mechanism).

7.3 Synthesis of Vanadium Nitride by Combustion of Ferrovanadium in Nitrogen

(a)

(b)

(c)

(d)

Figure 7.14 SEM images of combustion products of ferrosilicon in nitrogen. (a) Growth of scaly crystals of silicon nitride; (b, c) closing up reentrant angles between

individual crystals Si3 N4 ; and (d) silicon nitride in the form of spherical polyhedrons grown by the “vapor–crystal” mechanism.

faceting by interaction “liquid–gas” and VLC. Thus, experimental data on specifics of structure formation during combustion of ferrosilicon in nitrogen allow us to state that the growth of silicon nitride crystals was carried out both by the VLC mechanism and crystallization from iron–silicon melt. Growth of silicon nitride crystal from a melt is typical for combustion of ferrosilicon in nitrogen [14] while combustion of silicon in nitrogen is accompanied by growth of silicon nitride crystal in accordance with the VLC mechanism [19]. 7.3 Synthesis of Vanadium Nitride by Combustion of Ferrovanadium in Nitrogen

Vanadium and iron form a continuous range of α-solid solutions. In alloys with 52.3% of Fe, σ-phase is formed below 1225 ∘ C as a result of ordering of the solid solution structure. The homogeneity region of σ-phase varies from 45% to 65% of Fe at T = 20 ∘ C. Combustion of an alloy prepared by combination of carbonyl iron (99.8% Fe) with electrolytic vanadium VEL-1 (99.8% V) has been studied. The alloy

199

200

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

was grinded to achieve the particle size of the powder less than 0.14 mm and then samples of 30 mm in diameter and height with a relative density of 0.49–0.64 were pressed. Since the thermal effect of the VN formation is 60 kcal mol−1 and the thermal effect of the iron nitrides formation is 0.9–2.7 kcal mol−1 and taking into consideration that the dissociation temperature of the most stable iron nitride Fe4 N is 680 ∘ C, a conclusion can be made that the reaction wave will propagate due to heat release during the interaction of vanadium and nitrogen. Thus, increase in iron content in the alloy should lead to decrease in burning rate. However, the studies have shown that such dependence takes place only when the iron content is not more than 40%. Burning rate sharply increases (more than three times) with increase in iron content more than 40% (Figure 7.15) [9]. Further dilution of the alloy with iron leads to decrease in burning rate again. Depending on the ratio of components in the alloy, nitriding can occur with different contribution of aftercombustion (Figures 7.16 and 7.17). Figure 7.16 shows a graph illustrating a contribution of the stage of additional nitriding into the total nitrogen content of in the final product. Curve 1 corresponds to the total content of nitrogen in the final product; curve 2 is built without taking aftercombustion into account (samples were immediately quenched in water after combustion). Thus, when the concentration of iron in the alloy is less than 40%, aftercombustion contributes significantly in the process of nitriding. When the concentration of iron is more than 40%, aftercombustion is absent. The reason for increasing the burning rate of the σ-(FeV) composition is a phase transition when the temperature in the combustion wave reaches T f = 1500 K. The dependence of burning rate on iron content (at pressure of Po = 6 MPa) is shown U (mm s−1)

T (K)

18 15

1770 γ

α

12

1470

1 9

1170 σ

6

870 2

3

570

V

20

40

60

80

Fe, %

Figure 7.15 Phase diagram Fe–V (1) and dependence of combustion velocity on iron content in the alloy (2).

7.3 Synthesis of Vanadium Nitride by Combustion of Ferrovanadium in Nitrogen

m

1 0.65 2

0.35

0

20

40

Fe, %

Figure 7.16 Dependence of nitriding degree of combustion products of ferrovanadium in nitrogen on Fe content. 1 – non-quenched combustion products and 2 – quenched combustion products.

m σ-(Fe-V) α-(Fe-V) 0.5

0

5

10

Time (s)

Figure 7.17 Gravimetric curves of ferrovanadium nitriding.

graphically with the V-Fe phase diagram at a background (Figure 7.15) [20]. It is seen that a jump in burning rate corresponds to the boundary between phases α and σ. The results of investigation of microstructure of quenched samples showed that there is a direct connection between rapid nitriding and σ → α phase transition. From the macrokinetics standpoint, the influence-indicated factors can be interpreted as increase in effective diffusion coefficients and decrease in heterogeneity scope of the system during the phase transition. The change in diffusion environment during disordering and its influence on the combustion process was theoretically described by Smolyakov and Nekrasov for the model reaction of substitution A(N) + BC(FeV) → AB(VN) + C(Fe) that describes the macro process of interaction in the FeV–N system. The scheme of the diffusion zone is shown in Figure 7.18. The rate of interaction is determined here by diffusion processes in a product (AB) and initial alloy (BC) that is enriched with element C pushed in the center of the particle since another of its component is spent in the reaction with A diffusing to the boundary r1 . There are two qualitatively different modes of combustion depending on the ratio between the diffusion coefficients of the alloy and product d = DBC /DAB . In the area d ≪ 1 (Figure 7.19),

201

202

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

C

0

CA

CB CB

BC

AB

A

0 r1

R0

r2

Figure 7.18 Scheme of the diffusion zone.

ω 40 1

ω(d )

2

20

3 0

–6

ω( CB0 ) 0

0.2

0.4

0.6

0.8

CB

–2

2

6

10

In d

Figure 7.19 Dependence of dimensionless burning rate (�) on the ratio of diffusion coefficients (d = DBC /DAB ) at C o B = 0.5 and on alloy composition (C o B ).

the reaction wave velocity is determined by diffusion in the alloy. In the case when d ≫ 1, the reaction wave velocity is limited by diffusion in the product and corresponds to the burning rate (�AB ) of a binary system (e.g., vanadium in nitrogen) diluted with a certain amount of inert component C. In alloy BC, the increase in content of element B asymptotically brings the value of the reaction wave velocity closer to the value of �AB regardless of relationships between the parameters of diffusion. The last influences only the value of C A o . That is why the process of disordering leads to a sharp increase in the diffusion rate. Growth of DBC can influence the burning rate if the burning rate is determined exactly by coefficient DBC . Additionally, during combustion of the ordered phase, a jump in concentration occurs besides the change of DBC . For σ-FeV, α-phase formed during transition is a continuous range of solid solutions (Equation 7.7). { C1 , T ≤ Tf CV (x, r, −o) = (7.7) 0, T > Tf ,

7.3 Synthesis of Vanadium Nitride by Combustion of Ferrovanadium in Nitrogen

T f is the phase transition temperature and C 1 is the concentration of vanadium at the lower boundary of the σ-phase homogeneity area. The jump in burning rate during the σ → α transition in a framework of the examined model is determined by the change in DBC and change of diffusive environment caused by the decrease in the vanadium concentration down to zero C V (x, r, −o). This leads to significant increase in the vanadium flow to the boundary with the mononitride and, therefore, to sharp increase in heat release and nitriding degree and finally to increase in burning rate. The sharp change in scope of heterogeneity is clearly seen while investigating quenched layers of σ-FeV. Formation of nitride layers on surfaces of particles takes place at initial stages of interaction. The σ → α-FeV transition leads to nitride formation inside the particle. At the same time, diffusion of nitrogen goes through macro- and micro-cracks at the boundaries of grains and subgrains of the original particle of the powder. As soon as the temperature reaches melting point of iron, the liquid phase disperses powder particles. The particles are transformed into solid–liquid droplets consisting of the finest vanadium nitrides and liquid iron which form integrated melt. Combustion of α-FeV goes much slower. The areas with the heterogeneous structure of material remain intact in the final product along with the areas with the high degree of dispersion. Unlike α-FeV, the combustion products of σ-FeV remain two-phased and contain α-Fe and δ-VN for all values of investigated parameters. Nitrided σ-FeV is practically a nonporous material with a density of 6.2–6.5 g cm−3 (for comparison, ferrovanadium nitrided by the vacuum thermal method has a density of 4.5 g cm−2 ). It is optimal to use standard ferrovanadium with vanadium content of 46–53% to produce nitrogen-containing ferrovanadium by SHS. This allows producing alloys containing 9–11% of nitrogen with a density of 5.8–6.4 g cm−3 on a regular basis (Figure 7.20). The composite consists of vanadium nitride and iron. Other methods known at the present time are not capable of producing alloys with such characteristics.

(a)

(b)

Figure 7.20 Macrostructure of products of nitriding σ-(Fe–V) (a) and α-(Fe–V) (b).

203

204

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

7.4 Synthesis of Niobium Nitride by Combustion of Ferroniobium in Nitrogen

During combustion of both ferroniobium and ferrosilicon, nitrogen is supplied to the reaction zone by filtration from a surrounding volume. The increase in pressure from 0.2 to 8.0 MPa leads to increase in nitrogen content in combustion products reaching the stoichiometric value for niobium mononitride (8.25%). However, the nitrogen content can be higher than stoichiometric since partial nitriding of iron occurs along with nitriding of niobium. The data of XRD analysis and chemical analysis indicate the presence of the most thermally stable iron nitride Fe4 N in some samples. Nitrided ferroniobium represents quite well-sintered material at P < 1 MPa. Three layers are clearly seen in the cross-section of the sample: external – sintered, medium (dark) – fritted and high-porous, and internal – fritted and dense (Figure 7.21a). Chemical analysis of the layers showed that the nitrogen content is maximal in the external layer and minimal in the center. At high pressures, the burnt sample is a weakly sintered homogeneous material that can be easily milled (Figure 7.21b). Results of chemical analysis of the sample burned at high pressure showed that nitrogen is distributed evenly within the volume of the sample (contents of nitrogen in the center and in the volume are equal). Density of samples prepared for combustion considerably influences nitrogen content in combustion products. Decrease in density of initial samples leads to increase in nitrogen content in burned samples. Therefore, pouring of loose powder for initial samples is required to obtain highly nitrided final combustion products. Combustion of ferroniobium in nitrogen is accompanied by a long glow after the passage of the combustion wave which indicates a significant contribution of afterburning to the process of ferroalloy nitriding. The calculated temperature during combustion of ferroniobium in nitrogen for the alloy containing 60% Nb depends on the nitrogen pressure and varies in the

(a)

(b)

Figure 7.21 Structure of the samples of nitrided ferroniobium obtained at different pressures: (a) P = 0.2 MPa and (b) P = 4.0 MPa.

7.4

Synthesis of Niobium Nitride by Combustion of Ferroniobium in Nitrogen

range of 2329–2448 K, which is below the melting temperature of NbN (2740 K). In real conditions, the combustion temperature is even lower due to heat losses and incomplete conversion of the reactants but still higher than the melting temperatures of iron and ferroniobium. That is why the particles of nitrided ferroniobium in the combustion wave are solid–liquid drops corresponding in size to the original particle (Figure 7.22). The results of the EDX performed at the points shown in the figure indicate that niobium nitride is located in the surface layer of the particle and iron is localized in the central part of the particle. Nitriding of ferroniobium is likely to begin with the solid phase interaction between alloy particles and nitrogen and, as a result, a shell from high-melting niobium nitride is formed around the perimeter of the particle. The iron-based melt is located inside the niobium nitride shell, and the melt disperses a nitride layer, but an integrated solid–liquid layer is not formed around

*2

*1

20 μ (b)

(a)

(c) Figure 7.22 SEM/EDX images of combustion products of ferroniobium in nitrogen. (a) In reflected electrons; (b) in characteristic rays of iron; and (c) in characteristic rays of niobium (P = 3 MPa; sample diameter = 35 mm) *1 – Fe and *2 – NbN.

205

206

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

Table 7.2 Influence of nitrogen content on the phase composition of SHS-nitrided ferroniobium. Content of nitrogen N (%)

4.4 5.2 7.3 8.6 9.1

Phase composition

NbN (h), β-Nb2 N, γ-Nb4 N3 , Nb4−y Fe2+y N, α–Fe NbN (h), γ-Nb4 N3 , Nb4−y Fe2+y N, α-Fe NbN (h), γ-Nb4 N3 , Nb4−y Fe2+y N, α-Fe δ-NbN, α–Fe δ-NbN, α-Fe, Fe4 N

the sample. The sample retains its permeability which contributes to obtaining combustion products with high content of nitrogen. According to XRD analysis, products of combustion of ferroniobium in nitrogen contained γ-Nb4 N3 , β-Nb2 N, δ-NbN, Nb4−y Fe2+y N, and α-Fe. The lower niobium nitrides dominate in products of combustion of ferroniobium in nitrogen at low pressures of 0.2–1.0 MPa. The increase in nitrogen pressure leads to the increase in content of niobium mononitride, and composite material Fe–NbN is formed at pressure above 2 MPa. Table 7.2 shows the phase composition of nitrided ferroniobium depending on the nitrogen content in the alloy. Cubic niobium nitride is formed only with high nitrogen content in the alloy. SEM studies of morphology, particle size, and defect substructure of powders showed that the particles of nitrided ferroniobium have mainly splintery shapes with an average size of 0.53 μm (the minimum size was 0.083 μm). Particles with substructure were often observed. An average size of subgrains (subparticles) was 70 nm. Subparticles were free of defects and had regular geometrical shapes. XRD patterns obtained for some individual particles of nitrided ferroniobium showed that they often have complex phase compositions. Thus, nanosized particles (∼5 nm) of the Nb4−x Fe2+x N phase located on dislocations were found on particles of the NbN system. 7.5 Synthesis of Titanium Nitride by Combustion of Ferrotitanium in Nitrogen

Ferrotitanium is an alloying material used in production of stainless and heatresistant steels, for deoxidation and degassing of steel, in manufacturing of welding electrodes. The metallurgical industry produces a range of ferrotitanium types with titanium contents from 20% to 70%. Aluminothermic ferrotitanium consists of solid solution of titanium, silicon, and aluminum in iron composition of which is close to Fe2 Ti. High-grade ferrotitanium (65–70%) has a eutectic structure containing the FeTi intermetallic compound. In accordance with the results of electron microscopic studies and elemental microanalysis, the matrix is solid solution of Fe and Al in α-Ti as well as the phase (approximate formula Ti3 Fe) containing amount of titanium and aluminum bigger than in the matrix and the TiFe phase. The fine-grained structure corresponds to the TiFe–Ti eutectic composition.

7.5

Synthesis of Titanium Nitride by Combustion of Ferrotitanium in Nitrogen

To study combustion of ferrotitanium, two alloys were selected: eutectic with titanium content of 65.9% and melting temperature of 1085 ∘ C (TiFe–Ti) and the Fe2 Ti alloy. The Fe2 Ti intermetallic compound melts congruently at 1530 ∘ C and has a wide homogeneity region (∼10 at%). The first alloy is industrial ferrotitanium FTi 65 containing not less than 65% of Ti and impurity elements (not more than): Al (5.0%); Si (1.0%); C (0.4%); P (0.05%); S (0.05%); Cu (0.4%); V (3.0%); Mo (2.5%); Zr (2.0%); Sn (0.15%). The second (model) alloy was produced in a vacuum electric furnace by mixing ultrapure titanium and ultrapure carbonyl iron powders (titanium content 30%). The model alloy was then grinded in a ball mill to reach particle sizes less than 315 μm. The model alloy was needed because industrial types of low-grade ferrotitanium contain significant amounts of aluminum and silicon (up to 14% and 8%, respectively) which could make identification of features of nitriding of the Fe2 Ti intermetallic compound difficult. 7.5.1 Features of Ferrotitanium Nitriding

Another form of unstable combustion called “surface mode” is observed during combustion of ferrotitanium. This mode is determined by filtration rate of reacting gas. The characteristic peculiarity of samples burned in the surface mode is their significant macro-inhomogeneity. During combustion of ferrotitanium with sizes of particles less than 40 μm, burned samples retained external dimensions having a yellow-brown colored non-sintered external layer, but their internal parts were strongly deformed, melted, and impermeable. The middle of the samples was homogeneous, without voids and cracks and had a light gray metallic color. The burned samples made of ferrotitanium with particle sizes less than 315 μm also had a yellow-brown crumbling external layer, but the internal part was a melt divided by thin bright-yellow layers of powdered titanium nitride and voids with golden thin layers of titanium nitride on inside surfaces similar to the films obtained by deposition from the gas phase (Figure 7.23). The nitriding degree of external layers was significantly higher than the nitriding degree of internal layers (Figure 7.24). Dilution of initial ferrotitanium by preliminary nitrided ferrotitanium leads to the increase in nitriding degree. However, in contrast to ferrosilicon, increase in burning rate with increase in amount of additives was not detected since mechanisms of nitride formation for these ferroalloys are different. The amount of additive required for initiation of the SHS process depends on particle size of an original alloy and pressure of nitrogen. For the powder with particles sizes less than 40 μm, the maximum degree of nitriding is achieved by adding 40% of the additive. For the powder with particles sizes less than 315 μm, it is necessary to add half the value (nitrogen pressure 2 MPa). During combustion of ferrotitanium in the presence of ammonium chloride as an additive, values of nitriding degree and burning rate are much lower than corresponding values during combustion of ferrotitanium without this additive.

207

208

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

Figure 7.23 The ferrotitanium sample burned in nitrogen in surface mode.

1.0

m

0.9 0.8 0.7 0.6 0.5 0

5

10

15 20 Diameter (mm)

25

30

Figure 7.24 The curve illustrating changes in nitriding degree along the diameter of the ferrotitanium sample burned in nitrogen.

This is connected with the fact that during combustion in nitrogen, the hydrogenation reaction proceeds along with the nitriding reaction since the intermetallic compound FeTi is considered to be a promising material for hydrogen storage due to a high hydrogen capacity of corresponding hydride (1.8% for composition FeTiH2 ). The burning rate reduces due to competitive processes and products of

7.5

Synthesis of Titanium Nitride by Combustion of Ferrotitanium in Nitrogen

combustion of ferrotitanium in the presence of ammonium chloride contain twice less amount of nitrogen. That is why use of NH4 Cl is not advisable in ferrotitanium nitriding. A distinctive peculiarity of high-grade ferrotitanium is the lowest melting temperature among examined ferroalloys (1085 ∘ C). To understand the mechanism for nitriding of ferrotitanium, we stopped the combustion wave by quenching. Is has been found that the process of nitride formation begins with nitriding of a finely differentiated eutectic structure resulting in the formation of titanium nitride. Due to increase in volume caused by nitrogen absorption, dispersion of large particles of ferrotitanium along interphase boundaries occurs which leads to increase in speed of the process. The results of experiments (Figure 7.25) show that most nitrided products are formed during combustion of the alloy with 30.6% Ti. At the same time, the maximum possible yield of titanium nitride for this titanium content is achieved at nitrogen pressure of 3–4 MPa. Either decrease or increase in nitrogen pressure leads to decrease in nitriding degree since a surface combustion mode is realized at low pressures and the rate of heat release increases at pressures higher than 4 MPa which leads to increase in burning rate (Figure 7.26) and melting of the central part of the sample. In general, combustion of alloy with 30.6% Ti proceeds with a higher rate, and the products of combustion reach a maximum nitriding degree which is caused by the absence of kinetic limit of nitride formation by coagulation of molten particles and low (but still sufficient for self-propagation regime) temperature of combustion. Chemical analysis results have shown that products of combustion of original ferroalloy with 30.6% Ti have nitriding degree of 0.98 and products of combustion of original ferroalloy with 65.9% Ti have nitriding degree of 0.53 (Table 7.3).

1.0

m 1

0.9 0.8 2

0.7 0.6 0.5

0

1

2 3 4 Nitrogen pressure (MPa)

5

6

Figure 7.25 Dependence of nitriding degree on nitrogen pressure: 1 – Ti content = 30.6% and 2 – Ti content = 65.9%.

209

210

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

U (mm s−1) 1.0 0.9 0.8 0.7

1

0.6 0.5 0.4

2

0.3 0.2

1

0

2 3 4 Nitrogen pressure (MPa)

5

6

Figure 7.26 Dependence of burning rate of ferrotitanium on nitrogen pressure: 1 – Ti content = 30.6% and 2 – Ti content = 65.9%. Table 7.3 Influence of SHS conditions on nitriding degree of combustion products of different ferrotitanium types in nitrogen. Type

FTi 65a) FTi 65a) FTi 65a) FTi 65a) FTi 65a) FTi 30d) a) b) c) d)

Synthesis conditions

Additives (%)

PN2 = 0.2 MPa, d = 10 mm, r < 40 μm, TEb) — PN = 0.5–5.0 MPa, d = 20–40 mm, r < 315 μm — 2 PN = 2.0 MPa, d = 30 mm, r < 315 μm Fe-Ti-Nc) (10–20) 2 PN = 2.0 MPa, d = 30 mm, r < 40 μm Fe-Ti-Nc) (10–45) 2 PN2 = 2.0 MPa, d = 30 mm, r < 40 μm NH4 Cl (1.0–2.0) PN = 0.2 MPa, d = 10 mm, r < 40 μm, TEa) — 2

Nitriding degree

0.53 0.6–0.75 0.8–0.85 0.9–0.98 0.3 0.98

Industrial ferrotitanium with 65.9% of Ti (FTi 65). Thermal explosion (TE). Preliminary nitrided ferrotitanium (Fe–Ti–N). Model ferrotitanium with 30.6% of Ti (FTi 30).

7.5.2 Phase-Formation Processes of Titanium Nitrides During Combustion of Ferrotitanium in Nitrogen

The titanium–nitrogen system has been studied quite well. According to the literature sources, the system forms the following phases: α-solid solution of nitrogen in titanium (α-TiN) with Ti–TiN0.25 structure; β-solid solution of nitrogen in titanium (β-TiN); TiN0.25 –TiN0.45 structures consisting of nitride Ti2 N (�) combined with α-solid solution or nonstoichiometric TiN1−x and

7.5

Synthesis of Titanium Nitride by Combustion of Ferrotitanium in Nitrogen

TiN0.45 –TiN1.0 structures containing titanium mononitride (𝛿) TiN. Titanium nitride Ti2 N (�) is the low-temperature phase with a simple tetragonal cell and mononitride titanium (𝛿) TiN0.45−1.0 is the low-temperature phase with a face-centered cubic cell. The studies have shown that the phase composition of products of combustion of ferrotitanium in nitrogen is determined by the basic parameters of synthesis: nitrogen pressure, sample density, and composition of the original mixture. Table 7.4 presents results of research on phase components of the nitriding products of ferrotitanium depending on conditions for SHS process. For comparison, phase compositions of products of interaction between ferrotitanium and nitrogen in modes of thermal explosion and programmed heating have been included in the table. Low nitrogen pressure (0.2 MPa) was applied in these cases. The results showed that products of ferrotitanium nitriding in modes of programmed heating and thermal explosion consist of multiple phases. In addition to nonstoichiometric titanium nitride, these products contain lower nitrogencontaining phases Ti2 N and Ti3 N1.29 as well as intermetallic compounds with iron containing smaller amount of Ti. Nitriding of ferrotitanium by SHS leads to formation of titanium mononitride. The nitride phase of TiN0.3 is observed only at low pressures of nitrogen. In the case when a maximum nitriding degree is achieved during the SHS process, the combustion product differs in phase composition from other samples by small amounts of aluminum nitride and complex nitrides of titanium and aluminum (Table 7.4, Line 1). There is only one process way during combustion of the alloy: titanium diffuses toward the interphase boundary and is consumed during nitride formation. At the same time, the alloy is enriched with iron and aluminum pushed toward the center of the particle. Due to the high thermal effect of the formation of titanium nitride, the original alloy melts in the combustion wave, gas permeability of the sample reduces, and the nitriding reaction stops. To obtain nitriding products with the highest possible yield of titanium nitride, the original alloy was diluted by preliminary nitrided ferrotitanium. In this case, nitriding of both titanium and aluminum occurs. As noted above, only highly nitrided combustion products of ferrotitanium contain an aluminum nitride phase. Investigation on microstructure of nitriding products with a 100% yield of titanium nitride allowed finding certain morphological characteristics. The main part of the sintered material consisted of round-shaped particles. The thin net-like formations consisting of threads less than 0.3 μm in thickness and 40–50 μm in length, as well as needle-like crystals were observed in voids and blebs of the sintered material (Figure 7.27b). Spherical particles on the crystals tips (Figure 7.27a) indicate that the mechanism of their growth is VLC. This crystal growth mechanism is not typical for titanium nitride. The data of elemental microanalysis have shown that the spherical particle is a melt based on iron (54.9%) with dissolved aluminum (9.4%), titanium (17.2%), and nitrogen (18.5%). In a growing crystal, the ratio of components is significantly changed: aluminum (28.7%), titanium (32.0%), and nitrogen (38.2%). Complex nitrides of aluminum and titanium were detected by X-ray phase analysis.

211

212

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

Table 7.4 Results of XRD analysis data on products of ferrotitanium–nitrogen interaction. Original composition

Detected phases

Synthesis conditions, nitriding degree

Content Parameters Dimensions (�d/d) CSRa) (nm) × 10−3 of lattice of phases (nm) (vol%)

TiN cubic Fe alfa AlN hexagonal

81 8 0.5

Ti3 AlN cubic Ti3 Al2 N2 hexagonal Al0.4 Fe0.6 cubic

1.0 5.0

2. (T107) SHS Fe–Ti + Fe– P = 2 MPa Ti–N (10%) d = 30 mm m = 0.75

TiN cubic Fe2 Ti hexagonal

3. (T64) Fe–Ti

SHS P = 3 MPa d = 30 mm m = 0.60

4. (T01) Fe–Ti

Programmed heating up to 1200 ∘ C in nitrogen P = 0.2 MPa d = 10 mm m = 0.2

1. (T92) Fe– SHS P = 2 Ti + Fe– MPa Ti–N (40%) d = 30 mm m = 0.99

5. (T1) Fe–Ti

Thermal explosion P = 0.2 MPa d = 10 mm m = 0.2

a = 0.4242 a = 0.2879 a = 0.3113 c = 0.4961 a = 0.41121 a = 0.29875 c = 2.3350 a = 0.2896

120 >500 25

0.4 0.5 1.5

30 —

— —

>500

0.5

78 22

a = 0.4239 a = 0.4847 c = 0.7865

55 37

0.8 2.6

TiN cubic Fe2 Ti hexagonal

76 26

a = 0.4236 a = 0.4878 c = 0.7899

130 —

0.3 —

TiN cubic Fe2 Ti hexagonal

10 42

— —

— —

Fe0.2 Ti0.8 cubic Ti3 N1.29 trigonal rhombohedral Ti2 N tetragonal

35 5

— —

— —

—

—

TiN cubic Fe2 Ti hexagonal

40 15

— —

— —

Fe0.2 Ti0.8 cubic Ti3 N1.29 trigonal rhombohedral Ti2 N tetragonal

30 5

a = 0.4241 a = 0.4785 c = 0.7799 a = 0.3180 a = 0.2980 c = 2.1664 a = 0.4945 c = 0.3034 a = 0.4241 a = 0.4885 c = 0.7799 a = 0.3180 a = 0.2980 c = 2.1664 a = 0.4945 c = 0.3034

— —

— —

—

—

4.5

8

10

a) Coherent scattering region.

During combustion of ferrotitanium with low titanium content (30.6%), titanium nitride (TiN), and iron are formed in the combustion products regardless of nitrogen pressure. As opposed to ferrotitanium with high titanium content, in this case, the phase TiN0.3 is not detected even at low pressures of nitrogen which indicates that free titanium participates in the formation of this phase. Titanium nitride is crystallized from the melt in the form of spherical particles sizes of which are generally less than 200 nm and dominating sizes are 100–130 nm. In this case, iron is somewhat of a diluent preventing coalescence of TiN particles and assuring formation of particles of nanometric sizes.

7.6

Combustion of Ferrochromium in Nitrogen and Synthesis of Chromium Nitride

1

2

(a)

(b)

Figure 7.27 SEM images of combustion products of ferrotitanium in nitrogen (a, b).

7.6 Combustion of Ferrochromium in Nitrogen and Synthesis of Chromium Nitride

The calculations showed that adiabatic temperatures in the Fe–Cr–N system are much lower than combustion temperatures of ferroalloys. At the same time, the adiabatic combustion temperature is limited by the low dissociation temperature of chromium mononitride (1050 ∘ C at P = 0.1 MPa) which increases with increasing the nitrogen pressure. It is commonly supposed that an auto-wave mode of the reaction propagation can be practically realized if the adiabatic temperature of the combustion wave T ad is not less than 1500 K. Therefore, in accordance with the thermodynamic data, the interaction of ferrochromium and nitrogen in the combustion mode can be initiated when the chromium content in the alloy is higher than 60% and the nitrogen pressure is higher than 6 MPa. In practice, stable combustion of ferrochromium in nitrogen is difficult to initiate and maintain without additional measures due to heat loss into the environment. According to literature data, SHS is possible without additional preheating in low-energy systems (ferrochromium is among them). Preliminary mechanical activation of original reagents that creates energy reserve is used here to make SHS process possible. It was also concluded that grinding of original alloy is appropriate for thermal nitriding of ferrochromium and decrease in nitriding temperature should lead to increase in degree of grinding (dispersity). Taking into account the above-stated, industrial metallothermic ferrochromium containing 78.6% of chromium was grinded in a high-duty planetary mill (HPM) with water cooling with ratio powder/balls = 1/8. Mechanical activation was carried out in gasoline to avoid oxidation of ferrochromium. The powder obtained had particles sizes less than 80 μm but a dominant fraction was a fraction with particles sizes less than 15 μm. Specific surface of the powder determined by the BET method was 1.1 m2 g−1 . Figure 7.28 shows dependence of nitriding degree and burning rate of ferrochromium on nitrogen pressure [22]. Nitrogen content in combustion products increases with increase in nitrogen pressure. The burning rate also

213

214

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

U (mm s−1) 0.9

m 1.0 0.9

0.8

0.8

1

0.7

0.7

0.6

0.6

2

0.5

0.5

0.4

0.4 0.3

0.3

0.2

0.2

0.1 0

1

2

3 4 5 6 7 8 Nitrogen pressure (MPa)

9

0.1 10

Figure 7.28 Dependence of nitriding degree (m; 1) and burning rate (U; 2) of ferrochromium on nitrogen pressure.

increases with increase in nitrogen pressure since speed of supply of nitrogen to the combustion wave increases. Stable combustion of activated ferrochromium powder can be organized even at low values of nitrogen pressure (0.5 MPa) due to changed energy properties of the activated original powder. External examination of the burned samples showed that they have two volume zones: internal, densely sintered, compact and external, slightly sintered, friable. X-ray phase analysis showed that the phase composition of these zones depends on the pressure of nitrogen. When nitrogen pressure is more than 6 MPa, the main phase is mononitride of chromium (CrN) both within a densely sintered zone and friable part of the sample. When pressure is 4–6 MPa, the dominant phase in the sintered zone is mononitride, while the unsintered part of the sample mainly consists of Cr2 N. At pressure less than 4 MPa, the phase composition in both zones is represented mainly by Cr2 N. It should be noted that decrease in pressure leads to increase in volume of the densely sintered area. At a nitrogen pressure of 0.6 MPa, this zone occupies about two-third of the sample volume. The most probable cause of macrostructure heterogeneity of burned samples is a combustion surface mode when burning rate at the surface of the sample is higher than burning rate within the central part of the sample. Burned surface layers remain gas permeable and filtration of nitrogen toward the center of the sample does not stop. Nitrogen is heated when moving through the hot surface layer and temperature in the central part of the sample increased. Thus, the increased temperature in the central part of the sample contributes to sintering of particles of combustion products. Decrease in volume of the sintered zone observed with increase in pressure is caused by a diminishing role of the surface combustion mode. To obtain easily

7.7 Combustion of Ferroboron in Nitrogen and Synthesis of Boron Nitride

grindable sintered nitrided product with a high nitrogen content, it is necessary to conduct synthesis at a pressure of nitrogen higher than 8 MPa and composition “chromium mononitride–iron” can be obtained at nitrogen pressure of 10 MPa or higher. According to XRD results, the products of SHS nitriding of ferrochromium are represented by α-iron, chromium mononitride CrN, chromium subnitride Cr2 N, and complex nitride (Cr, Fe)2 N. Increase in nitrogen pressure during the synthesis leads to increase in a share of chromium mononitride and, correspondingly, decrease in share of complex nitride (Cr, Fe)2 N. The nitriding product of with a maximum yield of chromium mononitride was obtained at pressure of 10 MPa and higher. The SEM studies of the “quenched” samples with a low nitriding degree showed that there is a gray fringe around the phase of a dark gray color. XRD showed that the gray area corresponds to the complex nitride of chromium and iron (Cr, Fe)2 N. The dark area in the center of the grain is chromium mononitride. EDX and XRD results showed [22] that, when moving deep into the grain, successive substitution of iron atoms by nitrogen atoms occurs (Equation 7.8). Cr, Fe → (Cr, Fe)2 N → Cr2 N → CrN

(7.8)

The heat release rate and velocity of propagation of a combustion wave during SHS nitriding is determined by both diffusion of nitrogen to the center of the particle and crystal-chemical rearrangement at the boundary (interface) between solid phases. The decrease in particles sizes of original ferrochromium leads to acceleration of the diffusion process of nitrogen deep into grains and nitrogen transport ceases to be a limit factor of the nitriding process at a certain particles size. Therefore, use of fine-dispersed powders is required for effective SHS nitriding of ferrochromium and this should be taken into consideration when developing manufacturing processes to obtain products of ferrochromium nitriding with high nitrogen content. Thus, the most important parameters to synthesize chromium nitride by combustion of ferrochromium in nitrogen are particle size of original ferrochromium and pressure of nitrogen.

7.7 Combustion of Ferroboron in Nitrogen and Synthesis of Boron Nitride

Ferroboron is produced by the aluminothermy method using boric oxide, boric acid, or borate ore as a boron-containing component. Ferroboron FB 20 with boron content of 20.6% has been studied. Research into microstructure of ferroboron FB 20 has shown that it consists of two phases: the main phase FeB and rounded dark-colored inclusions FeBn . According to EDX results, boron content in it is higher than stoichiometric content for boron mononitride. It can be expected that phase composition of SHS products and laws of combustion of alloys containing a chemical compound will be different from those that are inherent in combustion of the corresponding chemical elements.

215

216

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

Results of experiments showed that self-sustaining combustion of ferroboron in nitrogen in the SHS mode was impossible due to low content of the nitrideforming element in the original alloy. After adding 5–10% of amorphous boron to original ferroboron, combustion products consisted of hexagonal boron nitride and α-Fe. Appearance of a reflex of α-iron on the X-ray diffraction pattern of combustion products indicates that nitriding has occurred. The experiments have shown that the optimum addition of boron is 7–10% since products of combustion practically do not contain boride phases of iron in this case. Addition of more than 10% of boron leads to appearance of phases of original ferroboron along with the main phase BN in products of nitriding. In this case, the process of nitride formation is carried out by interaction of boron additive and nitrogen and ferroboron is not involved in the formation of boron nitride. This is proved by the absence of α-iron reflexes on the X-ray diffraction patterns. To investigate the mechanism of boron nitride formation, combustion wave of was stopped by sharp pressure drop and injection of argon into the reactor (quenching). The nitrided products are shown in Figure 7.29. It is seen that a particle at initial stages retains its original shape and size. However, cracking and delamination of the nitride layer caused by increase in volume due to nitrogen absorption is observed on edges of the particle where nitriding degree is higher. Cracking of finer particles becomes extensive and they completely break into pieces losing their initial sizes. In accordance with results of local X-ray spectral microanalysis, central parts of particles contain Fe and B and external dispersed layers consist of boron nitride. Nitriding in this area is carried out by the “solid–gas” mechanism. The temperature of appearance of eutectic melt for this alloy is 1500 ∘ C. Data of the simultaneous Differential Scanning Calorimetry and Thermal Gravimetric Analysis (DSC–TGA) have shown that the stage of active nitriding of ferroboron is initiated at 1061 ∘ C which is caused by the phase transition of the low-temperature modification FeB (LT) into high-temperature modification FeB (HT). The increase in temperature leads to melting of the central non-nitrided

20 μ (a)

20 μ (b)

Figure 7.29 SEM images of original ferroboron particles (a) and products of nitriding with a stopped combustion front (quenched) (b).

7.8

Application Prospects of Products of Combustion of Ferroalloys in Nitrogen

part of the particles and nitriding is carried out by the “liquid–gas” mechanism. Furthermore, the composition of the alloy changes approaching firstly to Fe2 B and then to Fe3 B as boron is used from ferroboron to form boron nitride. This leads to melting of the still non-nitrided part of the iron–boron alloy and complicates the nitriding process because reduces gas permeability of the burning sample. That is probably why the Fe3 B phase is often observed in products of SHS nitriding of ferroboron. The microstructure of products of ferroboron combustion in nitrogen has various morphological forms. At the same time, a significant difference is visible in shape of particles of external and central parts of the burned sample. The presence of amorphous boron in the original material influences the process of nitride formation because (according to XRD analysis) amorphous boron contains boric acid which decomposes during combustion forming boric anhydride and water. It is known that gaseous compounds can be formed in the B–N–O system in addition to condensed (solid) products. The main component of the gas phase in the B2 O3 –B system at a temperature higher than 1400 K is boron suboxide B2 O2. Boron suboxide is formed during the SHS process at high temperatures in the presence of amorphous boron (Equation 7.9). 2B + 2B2 O3 → 3B2 O2

(7.9)

Then boron suboxide reacts further with nitrogen forming boron nitride (Equation 7.10). 3 B2 O3 + N2 → 2BN + O2 2

(7.10)

Suboxide moves to the colder external layers of the burning sample and forms a “knitted” layered structure of hexagonal boron nitride during the nitriding process (Figure 7.30a–d). Figure 7.30a shows transformation of a glassy structure (right) into the “knitted” and layered structure (left). According to results of elemental microanalysis, the first structure is boron oxide and the second one is boron nitride. In the central part of the sample, boron nitride is in a form of discs with thickness of not more than 0.1–0.2 μm and diameters less than 3.3 μm. The discs form a structure similar to open rosebuds by fancifully combining with each other (Figure 7.31a,b).

7.8 Application Prospects of Products of Combustion of Ferroalloys in Nitrogen

Presence of iron in products of SHS nitriding adds new functional advantages to nitride-containing material: raises strength and cutting ability abrasive granules based on composition Si3 N4 –Fe [22]; improves corrosion stability of electroconductive composition TiN–Fe(Al) by formation of heat-resistant intermetallic phases of the Fe–Al system [23]; allows producing sintered silicon nitride catalystbearing samples with porosity of 40–50% by SHS. In this case, iron is preliminarily

217

218

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

(a)

(b)

(c)

(d)

Figure 7.30 SEM images of particles of products of ferroboron nitriding with addition of 10% of amorphous boron; (a–d) outer part of the combusted sample.

(a)

(b)

Figure 7.31 SEM images of particles of products of ferroboron nitriding with addition of 10% of amorphous boron; (a, b) central part of the burned sample.

removed from the product of the synthesis (Si3 N4 –Fe) by means of acid treatment and the active substance is placed onto remaining porous silicon nitride matrix. A silver-containing catalyst system prepared on the base of granulated silicon nitride possesses high activity and stability in the process of selective oxidation of ethyleneglycol into glyoxal. Using Si3 N4 as a matrix allowed realizing

7.8

Application Prospects of Products of Combustion of Ferroalloys in Nitrogen

catalyst processes at higher temperatures (up to 650 ∘ C) and ensured improved resistance of nanosized (40–80 nm) particles of Ag to sintering and their uniform distribution during catalysis. Additionally, the main advantage of Si3 N4 as a matrix is absence of unwanted carbon deposits on its surface. Composite materials based on Si3 N4 –Fe and BN–Fe demonstrated high catalytic activity in processes of deep degradation of dissolved organic substances used for treatment of sewage and polluted water. 7.8.1 Application of Fe-Containing Composite Materials Based on Silicon and Boron Nitride for the Catalytic Destruction of Dissolved Organics

Heterogeneous catalysis generating highly reactive oxidizing hydroxyl radicals is perspective in cleaning of surface water from the dissolved organics (DOs). Oxygen-containing catalysts and reagents, UV-irradiation, and ozone treatment promote OH radicals formation. In the conditions of homogeneous catalysis, cyclic reactions of Ruff–Fenton system (Fe3+ + H2 O2 + UV) are highly effective as a renewable source of hydroxyl radicals. Homogeneous catalysis based on ferricoxalate system consisting of a soluble complex of iron-[Fe(C2 O4 )3 ]3− requiring mild acidic or neutral solutions is recently proposed. Simultaneous homogeneous and heterogeneous photocatalytic oxidation processes in wastewater treatment is rarely used. Our studies have shown [24–29] that the iron compositions based on nitrides of silicon and boron B–N–Fe and Si–N–Fe, synthesized by SHS, are donors for combined heterogeneous and homogeneous catalysis in the degradation processes of DO under the ozonation and UV treatment. During the photolysis or ozonolysis of catalytic systems in the solution, generation of hydroxyl radicals takes place, leading to deep oxidation of organic pollutants. The degradation of resistant pollutants, such as oxalic acid, formic acid, and formaldehyde has been observed. Capability of B–N–Fe composition samples to remove oxalic acid (OA) from water was investigated [24] and it was established (Table 7.5) that sorption of OA depends on surface and porosity properties of the material used and does not exceed 40%. H2 C2 O4 decomposition degree under UV in presence of each sample is rather high (80–90%). Addition of H2 O2 (photo-Fenton system) does not affect the catalytic activity of composites. Adsorption and catalytic activity of materials were investigated used by XRD and IR methods. Formation of photoactive ferricoxalate complexes explains efficiency of catalytic systems (Equations 7.11–7.15). [Fe(C2 O4 )3 ]3− + hν → Fe2+ + 2C2 O4 2− + C2 O4 2−

(7.11)

C2 O4 2− → CO2 2− + CO2

(7.12)

CO2 2− + O2 → CO2 + O2 2− Fe

2+

+ O2

2−

+

+ 2H → Fe

3+

(7.13)

+ H2 O2

(7.14)

Fe2+ + H2 O2 → Fe3+ +⋅ OH + OH− .

(7.15)

219

220

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

Table 7.5 Surface and porosity properties of B–N–Fe compositions, their sorption, and catalytic activity in H2 C2 O4 removal degree (R, %) from water under UV (catalyst mass – mcat = 100 mg; solution volume – v sol = 10 ml; OA concentration – cH C O = 5⋅10−4 M; 2 2 4 pH 3.5). Sample

Pa) (wt%)

B155 B152 B153 B154 B157 B35 B90 B92

3 5 10 15 20 w/o w/o 3%

Ssp (m2 g−1 )

V P (cm3 g−1 )

5.90 5.90 6.92 6.90 6.43 3.25 4.30 2.45

0.0295 0.0285 0.0367 0.0366 0.0308 0.0252 0.0268 0.0244

rp (nm)

16 − 14 16 − 14 17 − 14 18 − 15 16 − 13 14 − 12 15 − 13 14 − 11

R (%) Sorption

UV

UV/H2 O2

25 ± 1 29 ± 1 32 ± 2 34 ± 2 40 ± 2 12 ± 1 20 ± 1 4±1

54 ± 3 72 ± 4 86 ± 6 83 ± 5 76 ± 4 81 ± 5 88 ± 6 75 ± 4

72 ± 5 69 ± 3 79 ± 4 84 ± 4 75 ± 5 90 ± 6 72 ± 4 67 ± 3

a) Pore forming agent.

The decrease of metallic iron reflex and the appearance of FeC2 O4 ⋅2H2 O reflex can be observed in XRD patterns (Figure 7.32) of catalyst samples after contact with oxalic acid solution. This can be explained by the adsorption of oxalic acid at the composite surface accompanied by dissolution of iron and the formation of a new phase of low solubility. This supports the hypothesis of combining homogeneous and heterogeneous catalysis. The role of iron in the DO catalytic oxidation process using composites based on boron and silicon nitrides can be confirmed by the results of formaldehyde degradation under ozonization, UV radiation in the absence and in presence of oxalic acid additives [25] (Table 7.6). It can be seen that the degree of formaldehyde oxidation is independent of the materials phase composition, but is correlated with iron content. Moreover, in the presence of composite B50 with a very low iron content, formaldehyde oxidation does not occur. The use of composites with high iron content (B37) upon application of UV irradiation for 40 min allows destroying 90% of formaldehyde. The high efficiency of the samples during the oxidation of formaldehyde under UV radiation without additional reagents (H2 C2 O4 , H2 O2 ) is due to the formation of hydroxyl radicals as a result of iron(III) hydroxide photolysis (Equation 7.16). Fe(OH)2+ + hν → Fe2+ +• OH + H+

(7.16)

Catalytic ozonation in the presence of the Fe-containing composites generates hydroxyl radicals (Equations 7.17 and 7.18). Fe2+ + O3 → FeO2+ + O2 2+

FeO

3+

•

(7.17) −

+ H2 O → Fe + OH + OH .

(7.18)

The destruction of formaldehyde does not exceed 40% under UV treatment for 20 min in presence of ferric-oxalate system, which appears after oxalic acid addition. Increasing the time of UV irradiation up to 40 min leads to a significant

7.8

Application Prospects of Products of Combustion of Ferroalloys in Nitrogen 5

1

2 3 2

6

2

3

3

2

2

13

2

4

(b)

5 2 3

1 2 2

3

20

3

2

30

2

2

40

13

4

50

2θ

(a)

Figure 7.32 XRD pattern of B152 sample after H2 C2 O4 solution treatment (a) and initial sample (b):1 – BN; 2 – FeB; 3 – Fe2 B; 4 – Fe3 C; 5 – Fe; and 6 – FeC2 O4 ⋅2H2 O.

Table 7.6 Phase content of composites and formaldehyde degradation degree (R, %) during ozonation and UV treatment (catalyst mass – mcat = 200 mg; solution volume – v sol = 10 ml; formaldehyde concentration – cHCHO = 2.5 × 10−3 M; HCHO : H2 C2 O4 = 1 : 1; pH 5.7). Sample Phase content

w (%) Fe

R (%) O3 , 10 min UV, 40 min UV/H2 C2 O4 , 20 min

B50 B36 B11 B92 B153 B37 B77 B83 B85

BN, Fe, FeB, Fe2 B BN, Fe, FeB, Fe2 B BN, Fe, FeB, Fe2 B BN, Fe, FeB, Fe2 B, Fe3 C BN, Fe, FeB, Fe2 B BN, Fe, FeB, Fe2 B β-Si3 N4 , α-Si3 N4 , Fe, FeSi BN, β-Si3 N4 , Fe, FeB, Fe2 B, FeSi BN, β-Si3 N4 , Fe, FeB, Fe2 B, FeSi

4 32 37 55 61 69 36 42 28

∼0 10 ± 1 33 ± 2 38 ± 2 40 ± 2 51 ± 2 21 ± 1 32 ± 2 17 ± 1

∼0 67 ± 3 70 ± 3 71 ± 4 76 ± 3 88 ± 4 46 ± 2 53 ± 2 39 ± 2

∼0 32 ± 1 29 ± 1 37 ± 2 36 ± 2 40 ± 2 22 ± 1 23 ± 1 26 ± 2

increase (90%) of the organic pollutant degradation. When investigating the effect of the catalyst particle size on its activity, it was found that the optimum average particle size is below 2.5 mm. A strong decrease in the process efficiency when using the composite with granules sizes less than 0.5 mm can be caused by the fact that small particles form a stable suspension in solution and can extinguish (scatter) UV radiation.

221

222

7 Synthesis of Nitrides by SHS of Ferroalloys in Nitrogen

Table 7.7 Carbon content during formaldehyde destruction in presence of B37 catalyst under UV (mcat = 200 mg; vsol = 10 ml; � = 40 min). Destruction condition

Total carbon (TC)

Inorganic carbon (IC)

Organic carbon (OC)

mg l−1

w (%)

mg l−1

w (%)

mg l−1

w (%)

45.2a) 2.6 UV/NCON : H2 C2 O4 (1 : 1) 112.2a) 34.5

100.0 5.7 100.0 30.7

1.6 0.2 0.5 3.2

3.5 12.5 0.4 640

43.6 2.4 111.7 23.6

96.5 5.5 99.6 21.1

UV

a) In initial content.

To estimate the degree of formaldehyde destruction total carbon (TC), total inorganic carbon (TIC), and total organic carbon (TOC) after UV radiation in the absence and presence of oxalic acid were measured [25]. It can be seen (Table 7.7) that the bulk of the TC has organic nature. Upon application of UV irradiation only 5.5% of organic carbon remains in the solution, that is, the oxidation of formaldehyde proceeds deeply – to water and carbon dioxide. When the ferric-oxalate system was used, a complete mineralization of organic matter (HCOH, H2 C2 O4 ) does not occur; nearly 20% of the organic carbon remains in the solution. In this case, sixfold increase in the amount of inorganic carbon after the degradation of organic pollutants is observed. In this regard, the destruction of formaldehyde is best conducted without the addition of oxalic acid. The catalytic oxidation of phenols with a high oxidative capacity of hydroxyl radicals can generate stable secondary pollutants (benzoquinone, muconic acid, maleic acid, oxalic acid, catechol, hydroquinone). Iron composition of Si–N–Fe and B–N–Fe are promising materials to solve this environmental problem [26, 27, 29]. Table 7.8 shows phase compositions, specific surface areas, and catalytic activity of composite samples based on silicon and boron nitrides in the degradation of phenol under ozonation and UV irradiation. Low adsorption capacity of composites is apparently linked to their low specific surface area. It was found that the catalytic oxidation of phenol during ozonation is much more effective than the same during UV irradiation. Total phenol degradation occurs (Table 7.8) during ozonation in presence of B35 composite. Equal (to the pollutant) addition of oxalic acid (1 : 1) to create photoactive ferric-oxalate complex and under UV irradiation lead to a high degree of phenol degradation in the presence of B–N–Fe compositions. Higher catalytic activity of B–N–Fe composites can be caused by large iron content and structural features of the material. The ferric-oxalate system that is formed in the presence of oxalic acid and under UV radiation using samples based on nitrides of silicon (528) and boron (B90) is also quite promising for the destruction of phenol. However, the efficiency of the catalyst is determined not only by the fullness of the pollutant oxidation but also by the depth of its destruction. In this regard, the depth of phenol degradation (under conditions of maximal activity of composites) was assessed and products of phenol degradation determined by gas chromatography–mass spectrometry

7.8

Application Prospects of Products of Combustion of Ferroalloys in Nitrogen

Table 7.8 Si–N–Fe and B–N–Fe compositions characteristics and phenol removal degree (R, %) from the solution (pH 4.5) under ozonation and UV irradiation (cphen. = 5 × 10−4 M; mcat = 100 mg; v sol = 10 ml; � oz. = 10 min; � UV = 40 min; C6 H5 OH : H2 C2 O4 = 1 : 1; n = 3; P = 0.95). Sspec (m2 g−1 ) Phenol sorption (%)

Sample Phase content

621 629 423 528 B54 B35 B90

β-Si3 Al3 O3 N5 , Fe, FeSi β-Si3 N4 , ZrO2 , FeSi β-Si3 N4 , SiC, Fe, FeSi β-Si3 N4 , TiN, Fe, FeSi BN, Fe, FeB, Fe2 B, Fe3 C BN, Fe, FeB, Fe2 B BN, Fe, FeB, Fe2 B

4.86 3.18 2.53 3.22 6.90 3.25 4.30

3.7 5.4 6.2 5.6 3.6 6.3 0

O3

81 52 67 80 95 ∼100 88

UV UV/H2 C2 O4

9 11 0 4 10 35 24

56 21 26 93 82 65 89

Table 7.9 GC–MS data on phenol degradation products (cphen. = 0.047 mg l−1 ; mcat = 100 mg; v sol = 10 ml; � oz. = 10 min; � UV = 40 min; C6 H5 OH : H2 C2 O4 = 1 : 1). Ozonation # Product

1 1,4-Benzoquinone (para-quinone) 2 1,4-Dihydroxyphenol (hydroquinone) 3 Diphenol

UV/N2 C2 O4

MPC (mg l−1 )

w (%) # Product

5 × 10−5 1 × 10−3 Absent