Handbook of Cellular Metals: Production, Processing, Applications

Handbook of Cellular Metals: Production, Processing, Applications. Edited by H.-P. Degischer and B. Kriszt Copyright c ...

Author: Hans-Peter Degischer | Brigitte Kriszt

321 downloads 2231 Views 8MB Size Report

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

DOWNLOAD PDF

Handbook of Cellular Metals: Production, Processing, Applications. Edited by H.-P. Degischer and B. Kriszt Copyright c 2002 Wiley-VCH Verlag GmbH & Co. KGaA ISBNs: 3-527-30339-1 (Hardback); 3-527-60055-8 (Electronic)

Hans-Peter Degischer, Brigitte Kriszt (Editors) Handbook of Cellular Metals Production, Processing, Applications


Related titles from Wiley-VCH

W. Schneider, K. Ehrke (Eds.)

Continuous Casting ISBN 3-527-30283-2

G. Santhoff

Intermetallics ISBN 3-527-29320-5


Handbook of Cellular Metals Production, Processing, Applications

Edited by Hans-Peter Degischer, Brigitte Kriszt


Editors Prof. Dr. Hans-Peter Degischer Technische UniversitaÈt Wien Institut fuÈr Werkstoffkunde und MaterialpruÈfung Karlsplatz 13 A-1040 Wien Austria Dr. Brigitte Kriszt Technische UniversitaÈt Wien Institut fuÈr Werkstoffkunde und MaterialpruÈfung Karlsplatz 13 A-1040 Wien Austria

This book was carefully produced. Nevertheless, authors and publisher do not warrant the information contained therein 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. Die Deutsche Bibliothek ± CIP-Cataloguing-in-Publication-Data A catalogue record for this book is available from Die Deutsche Bibliothek c WILEY-VCH Verlag GmbH 69469 Weinheim, 2002 All rights reserved (including those of translation in 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 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. printed in the Federal Republic of Germany printed on acid-free paper Composition Hagedorn Kommunikation, Viernheim Printing Strauss Offsetdruck GmbH, MoÈrlenbach Bookbinding J. SchaÈffer GmbH & Co. KG, GruÈnstadt ISBN

3-527-30339-1


Preface B. Kriszt and H. P. Degischer

There have been a few international and national research and development programs on cellular metals in recent years: the US Multidisciplinary Research Initiative (MURI) on ultralight metal structures since 1996; a few European research projects funded within the 4th and 5th EU Framework Program; and a focussed research program funded by the German Research Council, which began in 1998. These cooperative activities add to the original research activities at the Fraunhofergesellschaft IFAM in Bremen, the University of Cambridge, Ranshofen and Vienna in Austria, the Slovac Academy of Science in Bratislava, and several other places in Europe. A considerable quantity of data has been produced and presented at international conferences. One of these meetings was organized by the German Society of Materials (DGM) on 28/29 February 2000 at the Vienna University of Technology. The proceedings are mainly in German [1] and the authors wanted another opportunity to publish their results in context with present knowledge. The reader of proceedings might be overwhelmed by very specific research results on selected topics, obscuring a general understanding of this expanding field of research. Besides the classical books dealing with cellular metals [2,3], an overview of the state of the art in the year 2000 was needed, covering primary and secondary processing, characterization of cellular metals, properties, modeling, and exploitation. As a result, some of the contributors to the Vienna symposium ªMetallschaÈume 2000º have been asked to extend their papers, by referring to related results of other researchers and giving a review of their particular topic, whilst maintaining the detailed specialist knowledge of the author. These contributions are introduced by coordinators who describe the state of the art in that field. Foamed metals are described more extensively than other cellular metals because of the actual research activities prevalent in Europe. The European development goal is the application of cellular metals in components for motor vehicles, a field in which price limits and consistent high quality are essential. This handbook aims to give a more detailed overview of the present state of the art in research and development on cellular metals with specific emphasis on processing and characterization of foamed metals than the recent survey [4]. The handbook gives a starting point

V

VI

Preface

for researchers new to the field, and references to topics adjacent to their own specialty for experts already engaged. Engineers and potential users are encouraged to consider the application of cellular metals, taking into account the specific peculiarities of each material to avoid failures due to miscalculation of processing requirements and performance. A guide to the multitude of different types of cellular metals is provided with an indication of particular differences in properties. Research is ongoing and it is to be hoped that experience with applications, that the book intends to promote and stimulate, is expanding. The editors thank all contributors and acknowledge their assistance in adjusting the content of their contributions in cooperation with each other to produce a concise overview of the state of the art in cellular metals from a European viewpoint.

References

1. H. P. Degischer (ed), ªMetallschaÈumeº 3. M. F. Ashby, A. G. Evans, N. A. Fleck, L. J. Special Issue, Materwiss. Werkstofftechn. 2000, Gibson, J. W. Hutchinson, H. N. G. Wadley, 31(6). Metal Foams: a Design Guide, Butterworth2. L. J. Gibson, M. F. Ashby, Cellular Solids: Heinemann, Oxford 2000. Structure and Properties, 2nd edn., Cambridge 4. J. Banhart, ªManufacture, Characterisation University Press, Cambridge 1997. and Application of Cellular Metals and Metal Foamsº, Progr. Mater. Sci. 2001, 46, 559 632.


Foreword T. W. Clyne

Interest in metallic foams dates from the 1940s, when Sosnick filed a patent on a production method involving vaporization of low melting point constituents of metallic alloys [1]. Other publications and patents followed in the subsequent two decades, covering concepts such as that of injecting metal into the interstices around ªspace holderº particles, which are subsequently removed, and the dispersion of particles, which will release gas via chemical reaction or thermal decomposition. However, in general, research on the production and performance of cellular metals remained at a low level until the 1990s, when a substantial acceleration took place. It is now a research topic receiving a high level of attention and various industrial applications are currently being explored. Updated information about these activities is available from certain websites [2] and a comprehensive review has recently been published [3]. Further research is certainly necessary into the development of improved processing methods, since much of the material produced hitherto has been of relatively poor quality and/or inherently rather expensive. However, in this context it is very important to understand the processing-microstructure-property inter-relationships for cellular metals and the relevance of these to the property combinations required for various applications. Making a foam from a metal, as opposed to a polymer, boosts the stiffness, range of operating temperature and resistance to many (organic) solvents, while, in comparison with a ceramic foam, important advantages are expected in terms of toughness, (thermal and electrical) conductivity and formability. However, when considering the detailed characteristics, a clear distinction should be drawn between open-cellular and closed-cellular metal, since, not only are these two materials made by different processing routes, but in general rather separate types of application can be identified for them. Of course, all types of cellular materials tend to be relatively light, and to have a high specific stiffness, but these features usually depend primarily on the pore content, whereas many other properties are much more sensitive to microstructure (cell structure and nature of cell walls).

VII

VIII

Foreword

Closed-Cell Foams Since the most promising methods for cheap production of bulk material tend to generate closed-cell foams (commonly of aluminum), attention has been concentrated on these for many structural applications. These include various components designed to absorb energy progressively under relatively low applied (compressive) loads. In principle, the reduced constraint on cell walls, compared with solid metal, should mean that large plastic strains can arise throughout via concertina-like deformation, with substantial absorption of energy. However, the performance of such components has often been a little disappointing hitherto, with premature failure commonly occurring within shear bands and such poor tensile ductility exhibited that components tend to fracture readily if any tensile stresses arise, for example under bending moments. It is now becoming clear that these problems are substantially reduced if the cell size can be kept fine and uniform preferably to sub-mm levels. There is thus a strong incentive to develop processing techniques capable of producing such material in bulk. Furthermore, it is commonly the case that the cell walls in these foams contain severely embrittling constituents, such as large ceramic particles and thick oxide films. Some such constituents are often deliberately introduced during processing in order to raise the melt viscosity and thus inhibit cell coarsening and drainage [4]. Recent work has confirmed that such constituents can have highly deleterious effects on the mechanical characteristics of the foam and has outlined the mechanisms responsible for this [5]. Further work is required both on understanding these effects in more detail and on developing processing routes in which these constituents are suitably modified or eliminated. Closed-cell foams are also of interest for other types of application, such as thermal barriers, although ceramic materials would often be preferred for these. Open-Cellular Metals Suggested and actual applications include filters, catalyst supports, heat exchangers, fluid flow damping conduits (including various types of shock wave dissipation devices), biomedical prostheses, internally-cooled shape memory actuators, air batteries, and protective permeable membranes and sheathes. Such functional components tend to incorporate higher added value than those in purely structural applications, which is appropriate in view of the generally higher costs of producing open-cellular metal. The scale of the cell structure is often important for the functional characteristics. This will clearly be a basic specification for filters and fluid flow limiting devices, but a fine cell size would often be preferred for heat exchangers etc. (subject to limitations imposed by any danger of pore clogging), while bone in-growth into prosthetic implants might require relatively coarse pores. However, these properties would often be needed in combination with a minimum strength and ductility requirement, so a relatively fine, uniform cell structure and a defectfree cell struts' microstructure might be beneficial from that point of view. A wide range of metals is being investigated for applications of open-cellular metals, bringing a requirement for improved understanding of process optimization issues in various alloy systems.

Foreword

References

1. B. Sosnick, US Patent 2 434 775, 1948. 4. V. Gergely, H. P. Degischer, T. W. Clyne, 2. http://www.metalfoam.net ªRecycling of MMCs and Production of http://www.npl.co.uk/npl/cmmt/metalMetallic Foamsº in Comprehensive Composite foams/index.html Materials, Vol. 3: Metal Matrix Composites, http://www.msm.cam.ac.uk/mmc T. W. Clyne (ed.), Elsevier, Amsterdam 2000, 3. J. Banhart, ªManufacture, Characterisation p. 797 820. and Application of Cellular Metals and Metal 5. A. E. Markaki, T. W. Clyne, ªThe effect of cell Foamsº Progr. Mater. Sci. 2001, 46, 559 632. wall microstructure on the deformation and fracture of aluminium-based foamsº Acta Mater. 2001, 49, 1677 1686.

IX


Contents 1 2 2.1 2.1.1 2.1.2

2.1.2.1 2.1.2.2 2.2 2.2.1 2.2.2 2.2.2.1 2.2.2.2 2.2.2.3 2.2.2.4 2.2.2.5 2.2.3 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.4.1 2.3.4.2 2.4 2.4.1 2.4.2 2.4.3 2.5 2.5.1 2.5.2

Introduction: The Strange World of Cellular Metals 1 Material Definitions, Processing, and Recycling 5

Foaming Processes for Al 8 Gas Injection: the Cymat/Alcan and Norsk Hydro Process 8 In-situ Gas Generation: the Shinko Wire Process and the FORMGRIP process 10 The Shinko Wire Process [2] 10 The FORMGRIP Process 12 Industrialization of Powder-Compact Foaming Technique 14 Principles of Foam Production 14 Practical Aspects of Foam Production 17 Powder selection 17 Mixing 18 Densification 18 Further processing of foamable material 19 Foaming 19 State of Commercialization 20 Making Cellular Metals from Metals other than Aluminum 21 Zinc 22 Lead 22 Titanium 22 Steel 25 Powder-Compact Foaming Technique 25 Steel Foams from Powder Filler Mixtures 27 Recycling of Cellular Metals 28 The Remelting of Cellular Metals 28 Recycling of Cellular Metal Matrix Composites 29 Conclusions 32 The Physics of Foaming: Structure Formation and Stability 33 Isolated Gas Bubble in a Melt 34 Agglomeration of Bubbles: Foam 35

XI

XII

Contents

2.6 2.6.1 2.6.2 2.6.2.1 2.6.2.2 2.6.2.3 2.6.3 2.6.3.1 2.6.3.2 2.6.3.3 2.6.3.4 2.6.4 2.7 2.7.1 2.7.1.1 2.7.1.2 2.7.2 2.7.2.1 2.7.2.2 2.7.2.3 2.7.3 2.7.3.1 2.7.3.2 2.7.3.3 2.7.3.4 2.7.4 2.7.4.1 2.7.4.2 3 3.1 3.1.1 3.1.1.1 3.1.1.2

3.1.1.3 3.1.1.4 3.1.2 3.1.3 3.1.3.1 3.1.3.2

Infiltration and the Replication Process for Producing Metal Sponges 43 Replication 44 The Replication Process: General Principles 46 Pattern Preparation 46 Infiltration 48 Pattern Removal 49 Physical and Mechanical Properties of Metal Sponge 51 Continuous Refractory Patterns 51 Discontinuous Refractory Patterns 51 Burnable Patterns 53 Leachable Patterns 53 Conclusions 55 Solid-State and Deposition Methods 56 Formation from Single Cells: Coreless Methods 58 Hollow-Sphere Structures made from Gas Atomized Hollow Powders 58 Hollow-Sphere Structures made from Coaxially Sprayed Slurries 59 Formation from Single Cells: Lost Core Methods 60 Hollow-Sphere Structures made by Cementation and Sintering 60 Hollow-Sphere Structures made from Galvanically Coated Styrofoam Spheres 61 Hollow-Sphere Structures made from Fluidized Bed Coated Styrofoam Spheres 61 Bulk Formation: Coreless Methods 63 Sintered Metal Powders and Fibers 63 Methods Utilizing Special Sintering Phenomena 64 Foaming of Solids 65 Foaming of Slurries 67 Bulk Formation: Lost Core Methods 67 Powder Metallurgical Space Holder Method 67 Deposition Methods 68 Secondary Treatment of Cellular Metals 71

Forming, Machining, and Coating 75 High-Temperature Forming 75 Specific Problems in Foam Forming 75 Process Sequence for Manufacturing 3D Composites with Aluminum Foam Cores 76 Material Behavior at the Solidus 77 Forming of Cellular Metals at High Temperatures 78 Machining 79 Coating 79 Mechanical Properties of Spray Deposits 80 Specific Difficulties in Foam Coating 80

Contents

3.1.3.3 3.2 3.2.1 3.2.2 3.2.2.1 3.2.2.2 3.2.2.3 3.2.2.4 3.2.3 3.2.4 3.2.4.1 3.2.4.2 3.2.5 3.2.5.1 3.2.5.2 3.2.6 3.3 3.3.1 3.3.1.1 3.3.1.2 3.3.1.3 3.3.1.4 3.3.2 3.3.2.1 3.3.2.2 3.4 3.4.1 3.4.2 3.4.2.1 3.4.2.2 3.4.3 3.4.3.1 3.4.3.2 3.4.3.3 3.4.3.4 3.4.3.5 3.4.3.6 3.4.3.7 3.4.3.8 3.4.4 3.4.4.1 3.4.4.2

Thermal Sprayed Composites from Metal Foams 81 Joining Technologies for Structures Including Cellular Aluminum 83 Introduction 83 Feasible Joining Technologies 83 Mechanical Fastening Elements 83 Gluing 84 Welding 84 Soldering and Brazing 85 Foam Foam Joints 87 Foam Sheet Joints 87 Microstructural Investigations 88 Mechanical Properties of Foam Sheet Joints 90 Transferability to Structural Parts 98 Tubes 98 Hat-Profiles 99 Summary 100 Encasing by Casting 103 Foam Cores for Encasing by Casting 103 Core Production 103 Core Attachment 104 Mechanical Properties 105 Coating of the Foam Cores 106 Shell Casting Processes 107 High-Pressure Casting Processes 107 Bonding Between Shell and Foam Core 112 Sandwich Panels 113 Sandwich Foaming Process 114 Industrial Application 116 Technological Benefits 117 Technical Limitations 118 Joining Technology of AFS 119 Laser Welding 120 TIG/MIG Welding 120 Bolt/Pin Welding 121 Punch Riveting 122 Riveting Nuts and Screws 123 Flow Drilling 123 Riveting 123 Bonding 123 Cutting of AFS 124 Laser Beam Cutting 124 Water Jet Cutting 124

XIII

XIV

Contents

4 4.1 4.1.1

4.1.1.1 4.1.1.2 4.1.2 4.1.2.1 4.1.2.2 4.1.2.3 4.1.2.4 4.1.2.5 4.1.2.6 4.1.3 4.1.4 4.2 4.2.1 4.2.1.1 4.2.1.2 4.2.2 4.2.2.1 4.2.2.2 4.2.2.3 4.2.2.4 4.2.3 4.2.3.1 4.2.3.2 4.2.4 4.2.4.1 4.2.4.2 4.2.5 4.3 4.3.1 4.3.2 4.3.3 4.3.3.1 4.3.3.2 4.3.3.3 4.3.3.4 4.3.4 4.3.4.1 4.3.4.2 4.3.4.3 4.3.5

Characterization of Cellular Metals 127

Characterization of Cellular and Foamed Metals 130 Definition of Structural Features of a Cellular Metal and Influence on Property Profile 130 Density and Volume Fraction of Pores 131 Shape and Size of Pores 133 Characterization Methods and Quantities of Geometric Architecture of Real Metallic Foams 136 Sample Preparation 136 Pore Size 137 Pore Shape 139 Pore Orientation 140 Thickness of Cell Edges and Walls 140 Topological Features 141 Characterization of Microstructure of Massive Cell Material 141 Conclusions 143 Computed X-ray Tomography 145 Principle of the Technique 145 X-ray Radiography 145 X-ray Tomography 146 Set-ups 147 Medium-Resolution Microtomography 147 High-Resolution Microtomography 147 Resolution Required for the Study of Metallic Foams 148 Reconstruction Method 148 Experimental Results 148 Initial Cell Structure 148 Evolution of the Structure During a Compression Test 150 Micromodeling of a Foam by Finite Elements 151 Direct Meshing of the Actual Microstructure 151 Results 152 Conclusions 155 Considerations on Quality Features 156 Introduction 156 Non-Uniformity of Cellular Metals 156 Macroscopic Parameters 159 Type of Cellular Metal 159 Surface and Dimensions 159 Apparent Density 161 Properties 161 Microscopic Features 161 Microstructure of the Metal 162 Geometrical Features 162 Microdefects 164 Mesoscopic Features 165

Contents

4.3.5.1 4.3.5.2 4.3.6 4.3.7 4.3.7.1 4.3.7.2 4.3.7.3 4.3.8

Geometry of Cellular Structure 165 Density Distribution 166 Systematics of Quality Features 166 Approximation of a Cellular Structure by a Continuum 168 Calculation of Density Maps 168 Representation of Non-Uniformity of Densities 172 Mesoscopic Basis for Material Modeling 174 Proposal of Quality Criteria 174

5 5.1 5.1.1 5.1.1.1 5.1.1.2 5.1.1.3 5.1.2 5.1.3 5.1.3.1 5.1.3.2 5.1.4 5.1.4.1 5.1.4.2 5.1.4.3 5.1.5 5.1.6 5.1.6.1 5.1.6.2 5.1.6.3 5.2 5.2.1 5.2.2 5.2.2.1 5.2.2.2 5.2.3 5.2.4 5.3 5.3.1 5.3.2 5.3.3 5.3.3.1 5.3.3.2 5.3.3.3

Mechanical Properties and Determination 183 Young's Modulus 183 Influence of the Foam Structure 184 Influence of the Foam Density 185 Influence of Deformation 186 Compression Behavior 187 Energy-Absorbing and Impact Behavior 190 Energy Absorbing Capability 190 Impact Behavior 191 Tension Behavior 193 General Tensile Behavior 193 The Influence of Notches 195 Foam-Specific Test Problems in Tension 196 Torsion Behavior 196 Fracture Behavior 197 Crack Initiation and Crack Propagation 197 Fracture Toughness 198 Foam-Specific Test Problems 201 Fatigue Properties and Endurance Limit of Aluminum Foams 203 Literature Survey of Endurance Data 203 High Cycle Fatigue Properties and Endurance Limit 208 Material and Procedure 208 Results 209 Mechanism of Crack Initiation 210 Summary 214 Electrical, Thermal, and Acoustic Properties of Cellular Metals 215 Electrical Properties 215 Thermal Properties 221 Acoustic Properties 225 Materials for Sound Insulation 226 Sound Absorbing Materials 228 Structural Damping 236

Material Properties 179

XV

XVI

Contents

6 6.1 6.1.1 6.1.2 6.1.2.1 6.1.2.2 6.1.2.3 6.1.3 6.1.3.1 6.1.3.2 6.1.3.3 6.1.3.4 6.1.3.5 6.1.3.6 6.1.3.7 6.1.3.8 6.1.4 6.1.5 6.1.5.1 6.1.5.2 6.1.6 6.1.7 6.2 6.2.1 6.2.2 6.2.2.1 6.2.2.2 6.2.3 6.2.3.1 6.2.3.2 6.2.4 6.2.5 7 7.1

7.1.1 7.1.2 7.1.3 7.1.4 7.1.5 7.1.5.1 7.1.5.2 7.1.5.3 7.1.5.4

Modeling and Simulation 243

Modeling of Cellular Metals 245 Motivation 246 Micromechanical Modeling of Cellular Materials: Basics 247 Analytical and Numerical Models 247 Classification of Microgeometries 249 Information Obtainable from Micromechanics 251 Selected Results of Micromechanical Simulations 252 Influence of Material Distribution in the Cell Walls 252 Influence of Wavy and Curved Cell Walls 255 Influence of Irregular Vertex Positions 257 Microgeometries Containing Cells of Different Sizes 258 Influence of Holes and Solid-Filled Cells 260 Influence of Fractured or Removed Cell Walls 261 Yield and Collapse Surfaces 262 Fracture Simulations for Metallic Foams 266 Modeling of Mesoscopic Density Inhomogeneities 269 Macroscopic Modeling and Simulation 272 Low Energy Impact on Thin Metallic Foam Paddings 273 Crushing of Foam-Filled Crash Elements 275 Design Optimization for Cellular Metals 276 Outlook 277 Mesomodel of Real Cellular Structures 281 Introduction 281 3D Mesomodel 284 Elastic Regime 285 Plastic Regime 285 Modeling of Uniaxial Compression 288 Deformation Band 289 Mechanical Properties 292 Discussion 294 Conclusions 297 Service Properties and Exploitability 299

The Range of Applications of Structural Foams Based on Cellular Metals and Alternative Polymer Solutions 299 Introduction 299 Potential Areas of Use 300 Material Properties 300 Main Component Configurations 301 Application and Attachment Techniques 304 Casting 304 Thermal Joining Processes 305 Mechanical Joining Processes 306 Three-dimensional sandwich 306

Contents

7.1.5.5 7.1.6 7.1.6.1 7.1.6.2 7.1.6.3 7.1.6.4 7.1.7 7.2 7.2.1 7.2.2 7.2.3 7.2.4 7.2.5 7.2.6 7.2.7 7.2.8 7.2.9 7.2.10 7.2.11 7.2.12 7.2.13 7.2.14 7.3 7.3.1 7.3.1.1 7.3.1.2 7.3.2 7.3.2.1 7.3.2.2 7.3.2.3 7.3.3 7.4 7.4.1 7.4.2 7.4.2.1 7.4.2.2 7.4.2.3 7.4.2.4 7.4.3 7.4.3.1 7.4.3.2 7.4.3.3 7.4.3.4

Alternative Cellular Materials Based on Polymers 306 Effectiveness 307 Bending and Torsional Stress 307 Impact Stresses 308 Axial Load 309 Acoustics 310 Outlook 311 Functional Applications 313 General Considerations 313 Biomedical Implants 314 Filtration and Separation 315 Heat Exchangers and Cooling Machines 315 Supports for Catalysts 317 Storage and Transfer of Liquids 317 Fluid Flow Control 317 Silencers 318 Spargers 318 Battery Electrodes 318 Electrochemical Applications 319 Flame Arresters 319 Water Purification 319 Acoustic Control 319 Machinery Applications 320 Parameters 321 Thermal Behavior 321 Pull-Out Strength of Detachable Joints 322 Examples of Application 323 Foamed Steel Pipes 323 Machine Table 327 Cross-Slide 329 Conclusions 330 Prototypes by Powder Compact Foaming 330 Introduction 331 Methods, Machines, and Molds 332 Manufacturing Methods for Precursor Material 333 Foaming Process 334 Foaming Furnaces 334 Foaming Molds 336 Prototypes and their Applications 337 Automotive Applications 338 Construction and Architecture 340 Other Technical Applications 340 Improbable Applications 343

XVII

XVIII

Contents

7.5.4 7.5.4.1 7.5.4.2 7.5.5 7.5.5.1 7.5.5.2 7.5.6

Applying the Investment Methodology for Materials (IMM) to Aluminum Foams 346 Introduction: The Investment Methodology for Materials (IMM) 346 Initial Market Scan for Potential Applications for Al Foams 347 Material Assessment 347 Technical Performance 347 Cost of Production 348 Co-Minimizing Volume and Cost in Energy-Absorbing Applications 349 Market Forecast 350 Market Size for Aluminum Foam 350 Market Timing for Aluminum Foams 350 Value Capture 351 Industry Structure 351 Appropriability of Profits 352 Conclusions: Applying IMM to Aluminum Foams 353

8 8.1 8.2 8.3

Processing 355 Properties 358 Design and Application 360

7.5 7.5.1 7.5.2 7.5.3 7.5.3.1 7.5.3.2 7.5.3.3

Strengths, Weaknesses, and Opportunities 355

Index 365


List of Contributors O. Andersen Frauenhofer-Institut fuÈr Fertigungstechnik und Materialforschung (IFAM) Auûenstelle fuÈr Pulvermetallurgie und Verbundstoffe Winterberstr. 28 01277 Dresden Germany M. Arnold Lehrstuhl fuÈr Werkstoffwissenschaften UniversitaÈt Erlangen-NuÈrnberg Martensstr. 5 91058 Erlangen Germany M. F. Ashby Engineering Design Centre University of Cambridge Engineering Department, Trumpington Street Cambridge CB2 1PZ UK J. Banhart Hahn-Meitner-Institute Dept. of Materials ± SF3 Glienicker Str. 100 14109 Berlin Germany

F. BaumgaÈrtner Schunk Sinter Metalltechnik GmbH Postfach 10 09 51 35339 Gieûen Germany C. Beichelt Wilhelm KARMANN GmbH Karmannstr. 1 Postfach 26 09 49084 OsnabruÈck Germany T. Bernard Neue Materialien Bayreuth GmbH UniversitaÈtsstr. 30 94447 Bayreuth Germany H. W. Bergmann² UniversitaÈt Bayreuth Lehrstuhl Metallische Werkstoffe Ludwig-Thoma-Str. 36b 94440 Bayreuth Germany H. J. BoÈhm Institute of Lightweight Structures and Aerospace Engineering Vienna University of Technology Guûhausstr. 27±29 1040 Vienna Austria

XIX

XX

List of Contributors

R. Braune Lehrstuhl fuÈr Fertigungstechnologie UniversitaÈt Erlangen-NuÈrnberg Egerlandstr. 11 91058 Erlangen Germany

C. Haberling Firma Audi AG Abteilung Werkstoffe/Verfahren/Recycling I/EG-34 85045 Ingolstadt Germany

T. W. Clyne Department of Materials Science University of Cambridge Pembroke Street Cambridge CB2 3QZ UK

F. Heinrich Lehrstuhl fuÈr Werkstoffwissenschaften UniversitaÈt Erlangen-NuÈrnberg Martensstr. 5 91058 Erlangen Germany

T. Daxner Institute of Lightweight Structures and Aerospace Engineering Vienna University of Technology Guûhausstr. 27±29 A-1040 Vienna Austria

Th. Hipke Frauenhofer Institut fuÈr Werkzeugmaschinen und Umformtechnik (IWU) Reichenhainerstr. 88 09126 Chemnitz Germany

H. P. Degischer Institute of Materials Science and Testing Vienna University of Technology Karlsplatz 13 1040 Vienna Austria B. Foroughi Institute of Materials Science and Testing Vienna University of Technology Karlsplatz 13 1040 Vienna Austria M. C. Hahn Lehrstuhl fuÈr Fertigungstechnologie UniversitaÈt Erlangen-NuÈrnberg Egerlandstr. 11 91058 Erlangen Germany

C. KoÈrner Lehrstuhl Werkstoffkunde und Technologie der Metalle UniversitaÈt Erlangen-NuÈrnberg Martensstr. 5 91058 Erlangen Germany J. Kovacik Institute of Materials and Machine Mechanics Slovak Academy of Science Racianska 75 P.O. Box 95 83008 Bratislava Slovakia A. Kottar Institute of Materials Science and Testing Vienna University of Technology Karlsplatz 13 1040 Vienna Austria


B. Kriszt Institute of Materials Science and Testing Vienna University of Technology Karlsplatz 13 1040 Vienna Austria R. Kretz ARC Leichtmetallkompetenzzentrum Ranshofen GmbH Postfach 26 5282 Ranshofen Austria E. M. A. Maine Centre for Technology Management Engineering Design Centre Engineering Department University of Cambridge Trumpington Street Cambridge CB2 1PZ UK E. Maire CR1 CNRS GEMPPM Batiment Saint ExupeÂry 23 Avenue Capelle 69621 Villeurbanne cedex France U. Martin Institute fuÈr Metallkunde TU Bergakademie Freiberg Gustuv-Zeuner-Str. 5 09596 Freiberg Germany H. Mayer Institute of Meteorology and Physics University of Agricultural Sciences TuÈrkenschanzstr. 18 1180 Vienna Austria

A. Mortensen Laboratoire de MeÂtallurgie MeÂcanique Ecole Polytechnique FeÂdeÂrale de Lausanne 1015 Lausanne Switzerland U. Mosler Institut fuÈr Metallkunde TU Bergakademie Freiberg Gustuv-Zeuner-Str. 5 09596 Freiberg Germany C. Motz Erich-Schmid-Institute of Material Science Austrian Academy of Science Jahnstr. 12 8700 Leoben Austria R. Neugebauer Frauenhofer Institut fuÈr Werkzeugmaschinen und Umformtechnik (IWU) Reichenhainerstr. 88 09126 Chemnitz Germany A. Otto Lehrstuhl fuÈr Fertigungstechnologie UniversitaÈt Erlangen-NuÈrnberg Egerlandstr. 11 91058 Erlangen Germany R. Pippan Erich-Schmid-Institute of Material Science Austrian Academy of Science Jahnstr. 12 8700 Leoben Austria

XXI

XXII


G. Rausch Frauenhofer Institut fuÈr angewandte Materialforschung (IFAM) Lesumer Heerstr. 36 28717 Bremen Germany

R. F. Singer Institut fuÈr Werkstoffwissenschaften UnivertitaÈt Erlangen-NuÈrnberg Martensstr. 5 91058 Erlangen Germany

C. San Marchi Northwestern University Department of Materials Science and Engineering Evanston IL 60202-3108 USA

S. Stanzl-Tschegg Institute of Meteorology and Physics University of Agricultural Sciences TuÈrkenschanzstr. 18 1180 Vienna Austria

F. G. Rammerstorfer Institute of Lightweight Structures and Aerospace Engineering Vienna University of Technology Guûhausstr. 27±29 1040 Vienna Austria

G. Stephani Frauenhofer-Institut fuÈr Fertigungstechnik und Materialforschung (IFAM) Auûenstelle fuÈr Pulvermetallurgie und Verbundstoffe Winterberstr. 28 01277 Dresden Germany

W. Seeliger Wilhelm KARMANN GmbH Karmannstr. 1 Postfach 26 09 49084 OsnabruÈck Germany M. Seitzberger Institute of Lightweight Structures and Aerospace Engineering Vienna University of Technology Guûhausstr. 27±29 1040 Vienna Austria F. Simancik Institute of Materials and Machine Mechanics Slovak Academy of Science Racianska 75 P.O. Box 95 83008 Bratislava Slovakia

M. Thies Lehrstuhl fuÈr Werkstoffwissenschaften UnivertitaÈt Erlangen-NuÈrnberg Martensstr. 5 91058 Erlangen Germany B. Zettl Institute of Meteorology and Physics University of Agricultural Sciences TuÈrkenschanzstr. 18 1180 Vienna Austria


List of Abbreviations AFS AVS BET CCD CIP COD CT CTOD CVD DOF 2D or 3D EB-DVD EDM ESRF FE FORMGRIP FPZ HIP IFAM IMM INSA IP IWU LBM LDC LEFM LKR MIG MMC MURI PCF PM PSF

aluminum foam sandwich averaging volume size Brunnauer-Emmett-Teller nitrogen gas absorption method charged coupled device cold isostatic pressing crack opening displacement computer tension specimen crack tip opening dispacement chemical vapour deposition degree of freedom two- or three dimensional electron beam directed vapor deposition process electrodischarge machining European Synchrotron Radiation Facility, Grenoble finite elements foaming of reinforced metals by gas release in precursors fracture-process zone hot isostatic pressing Fraunhofer-Institute for Applied Material Research investment methodology for materials Institut National des Sciences AppliqueÂes de Lyon intellectual property Fraunhofer-Institute for Machine Tools and Forming Technology lattice block materials low density core material linear elastic fracture mechanics Leichtmetallkompetenzzentrum, Ranshofen, Austria metal inert gas arc welding metal matrix composite USA multidisciplinary research initiative powder compact foaming powder metallurgy point spread function

XXIII

XXIV

List of Abbreviations

PVD RV RVE SAS SEM TIG US XCT

physical vapour deposition representative volume representative volume element Slovak Academy of Science scanning electron microscopy tungsten inert gas arc welding ultrasonic X-ray computed tomography


1 Introduction: The Strange World of Cellular Metals F. Simancik When Nature builds large load-bearing structures, She generally uses cellular materials: wood, bone, coral. There must be good reasons for it. M. F. Ashby

It is well known that porous structures are good for insulation, packaging, or filtering, but few people believe that they can also be very effective in structural applications. Thousands of scientific publications deal with the minimization of porosity in load-bearing parts. Engineers work hard to eliminate pores from castings, powder metallurgy parts, weld joints, or coatings, thinking that a defect-free part is a pore-free one. With this attitude it is difficult for someone to accept that a loadbearing material can include pores, even quite large ones. However, large natural structures of porous materials have existed for thousands of years, demonstrating how evolution has generated cellular structures that optimize mechanical properties and structural function for minimum weight. Mankind tries to learn from nature. Understanding the benefits of natural structures gives us information to help us produce man-made cellular solids. The cellwall material has to be chosen very carefully if the structure is expected to carry loads. Polymers appear to be insufficiently rigid and ceramics are too brittle. Perhaps metals could be the right choice. Several of the engineering properties of metallic foams are superior to those of polymeric ones: they are stiffer by an order of magnitude, they are stable at elevated temperatures, they possess superior fire resistance, and they do not produce toxic fumes in a fire. Moreover, these materials are fully recyclable without any pollution or waste problems. The latter fact can no longer be ignored, because the production, disposal, and use of stronger and stiffer materials in new products often have negative environmental impacts over the product life cycle. Owing to their pores, cellular metals possess a set of unusual properties compared with bulk structural materials: they are crushable, they exhibit a plateau stress if compressed, and they exhibit a change in Poisson ratio on deformation. The excellent combination of good mechanical properties (mostly strength and stiffness) and low weight is the prime advantage. In addition, cellular metals

1

2

1 Introduction: The Strange World of Cellular Metals

absorb high impact energies regardless of the impact direction, and are very efficient in sound absorption, electromagnetic shielding, and vibration damping. Most of the mechanical properties of foam materials can be achieved with other materials, sometimes more effectively, but foams can offer a unique combination of several (apparently contradictory) properties that cannot be obtained in one conventional material at the same time (e. g., ultra-low density, high stiffness, the capability to absorb crash energy, low thermal conductivity, low magnetic permeability, and good vibration damping). Cellular metals are thus promising in applications where several of these functions can be combined. These properties depend significantly on the porosity, so that a desired portfolio of properties can be tailored by changing the foam density. This is one of the most attractive features of these remarkable materials. Cellular material properties also depend on the pore structure. This influence, imperfectly understood at present, is a topic of intense study. Various constitutive laws have been suggested for the characterization and modeling of this relationship. These laws, originally developed for polymeric foams, are usually based on the relative density of the foam, and therefore suppose uniform cellular structure, at least at a macroscopic level. However, metallic foams are dramatically different from polymeric foams: polymeric foams generally have a regular microstructure, whereas metallic foams may be highly disordered with a wide dispersion of cell size and shape. Moreover, many imperfections exist in a cell structure, such as cracks or holes in the cell walls, corrugated cells etc. These effects are inevitable due to manufacturing at significantly higher temperatures than in the case of polymers. If these features are not taken into account and the properties of the foam are characterized only in relation to apparent density, a higher scatter of properties is to be expected. This is why it is still widely believed that acceptable reproducibility of the properties of metallic foams is questionable. The structure of metallic foams is often non-uniform, especially in the case of complex 3D parts. It should be noted that a uniform structure is not necessary for obtaining acceptable and reproducible properties. Anisotropic or gradient pore structures allow the distribution of load bearing material according to load conditions (simulating the optimum bone-like structure), without a need to increase the overall weight or volume of the component. Therefore, the challenge for manufacturing is not to produce a uniform structure, but to achieve reproducible properties with a controlled non-uniform structure. If the non-uniform structure is optimal, the crucial question ªWhat is the material?º should be answered. It is really difficult to distinguish between material and structure. If a cellular metal is a material, it is very problematic to define geometryindependent material characteristics (the strength or elasticity modulus); if it is a structure made of a certain metal it is almost impossible to define its random geometry. Cellular metals can be prepared by various processing methods. They may all be called ªmetallic foamsº, but they are very different materials, depending on the manufacturing technique. The production method affects the distribution of the


cell-wall material in such a way that the properties of differently manufactured materials are not comparable. Metallic foams result from the nucleation and subsequent growth of gas bubbles in a liquid or semi-liquid metal. They usually have a non-uniform pore structure (variable pore size and sometimes preferred orientation of pores). The pores are initially closed, but some defects always appear on cooling, owing to shrinkage of solidifying metal and gas-pressure reduction in pores. Many solidified cell faces have non-uniform curvature or are corrugated and have occasional broken walls that still hang in place. These are the main features of this kind of cellular solid. Other cellular metals may be manufactured by casting or deposition of the metal onto templates or place holders, which have to be removed from the final product, thus creating a porous structure. These cellular structures usually have adjustable distribution of pore size (according to the template) and their cells are always open. The manufacturing process dictates not only the properties but also the potential applications of the foam. Thus foams prepared by the powder compact foaming (PCF) technique (usually with a dense skin) can be effectively used as net-shape components, stiffening cores in castings, or in complicated hollow profiles, whereas the foams prepared by the ªmolten metal routeº (typically large blocks or panels) can be effectively used as voluminous energy absorbers, cores for sandwiches, or for blast protection. The open cellular structures made by investment casting are good for heat exchangers, sound absorbers, or for electrodes in batteries. The properties arising from the cellular structure produced by a certain manufacturing technique cannot be effectively achieved using another method. This also means that cellular metals manufactured differently are not necessarily competitive materials. The first attempt to foam a metal was performed by B. Sosnik in 1943 [1]. In order to create pores, he added mercury to molten aluminum. In 1956 J. C. Elliot replaced mercury by foaming agents generating gas by thermal decomposition [2], so now modern scientists and engineers can develop metallic foams without having to deal with the toxicity of mercury. In 1959 B. C. Allen [3] invented the PCF route for manufacturing metallic foams and the basic processing techniques were thus completed. The success in the preparation of the first metallic foams and discovery of their remarkable properties started a euphoric enthusiasm for these materials. In 1957 J. Bjorksten stated [4]: ªFoamed metals offer great market potential and might conceivably account for 10 % of all metals produced within 20 yearsº. However he also said: ªA lot of work still remains to bring about production on a large scale, such as closer control of density and dimensions.º Unfortunately, only the second of his statements turned out to be realistic. Although it is many years since the first patents concerning the manufacture of metallic foams appeared, the material has not been put into large-scale commercial production yet. This discouraging fact can be attributed to inadequate design of components, low reproducibility of properties, a lack of testing procedures and calculation approaches, absence of concepts for secondary treatment, as well as the production technologies being too complicated and relatively expensive.

3

4


In spite of disappointments and mistrust, interest in cellular metals is growing. A new era started, at least in Europe, at the end of the 80s thanks to the activities of the Fraunhofer Institute for Advanced Materials in Bremen (IFAM). When J. Baumeister showed his floating aluminum on German TV in the ªKnoff-Hoff showº this remarkable material found a lot of enthusiastic fans. Stabilization of the melt via viscosity-increasing additions significantly improved the quality of the foam structure [5,6]. The development of new foaming techniques enabled reasonable manufacturing costs, and so metallic foams became very attractive to the transport industry, especially for lightweight stiff body structures and crash absorbing elements. The first industrial companies (Shinko-Wire, Cymat, Alulight, Schunk, Karman, Neuman-Alufoam) have already established a group of ªmetfoamº producers and further companies will join them soon. At present metallic foams are still insufficiently characterized, and understanding of the process is incomplete, leading to inadequate control and, hence, variable properties. This gives an impetus for large multilateral scientific activities like the USA Multidisciplinary Research Initiative on ultralight metal structures (MURI) [7] or the German Reasearch Council's focussed research program (DFG) [8]. Better understanding leads to better process control and improved properties. The producers themselves have aggressive development programs for their materials. The next generation of metallic foams will certainly be better. With all these shortcomings, even the present generation has shown excellent performance in many case studies. The structures, though still apparently imperfect, can be acceptable if they are applied properly in a foam-familiar design, because cellular metals have to be used in non-traditional ways. The problem-solving approach, instead of the trial-and-error method, will definitely accelerate the implementation of cellular metals in real products. An intensive and close collaboration among scientists, engineers, producers, and end users is crucial for success. Dr. Bjorksten's: ªNow it's up to industry to decide what to do with itº [4] would have fatal consequences for the future of cellular metals.

References

1. B. Sosnik, US Patent 2 434 775, 1948. 6. J. Iljoon et al., US Patent 5 115 697, 1992. 2. J. C. Elliot, US Patent 2 751 289, 1956. 7. Ultralight Metal Structure Project, MURI 3. B. C. Allen, US Patent 3 087 807, 1963. Grant No. N00014-1-96-1O28, Washington 1996. 4. Modern Metals, 1957, October. 5. S. Akiyama et al., European Patent 0 210 803, 8. Zellulare metallische Werkstoffe, DFG-Priority 1986. Program 1075, Bonn 1999.


2 Material Definitions, Processing, and Recycling H. P. Degischer

A great variety of cellular metals is produced by research laboratories and industrial development departments. Established industrial products are Duocel, Incofoam, Alporas, and some others are on the brink of market introduction. The different types of products and prototypes are the result of various combinations of processing, architecture, and metal matrix. The architecture of the cellular structure is the result of the processing technique, which can be classified according to Fig. 2-1, but each process has special features typical of the producing company or laboratory. None of the manufacturing techniques can be applied to any metal; each is appropriate for one or other base metal. Cellular structures are those with a relative

Processing techniques for cellular metals classified according to the state of the metal, the formation of the cellular architecture, and the pore-forming ingredients.

Figure 2-1.

5

6

2 Material Definitions, Processing, and Recycling

density of less than 0.3 [1]. Materials with higher relative densities are called porous materials (for instance, powder compact greens), most of which also can be produced by the processing techniques listed in Fig. 2-1. The most widely developed and investigated cellular metals are based on aluminum and its alloys. The general aspects of processing are described elsewhere [1 3] and specific presentations can be found in the proceedings of the conferences dealing with cellular metals [4 10]. The most important processing techniques are described in detail in subsequent chapters. Cellular metals are heterogeneous materials formed by a 3D metallic matrix with gas-containing pores occupying more than 70 vol.-%. Cellular metals are classified according to the following criteria (Fig. 2-1). x

x

x

The metal condition during production of porosity: liquid, solution or emulsion, solid. The forming process involved: casting, foaming, deposition, sintering (including precursor slurry). The method of pore formation: incorporating hollow substrates, removable substrates, or gas (either directly, dissolved, or by means of a dissociating agent).

Open porosity structures can be formed by replication (Duocel [6]), deposition (Incofoam [6]) or by construction of solid ingredients with space between (for instance, the 3D networks of lattice block materials (LBM) [5,6] or those prepared by rapid prototyping techniques [11]). Open-cell solid structures may be called sponges. Closed cells are produced by embedding or cementation of hollow ingredients (ªsyntactic foamsº), or by foaming in the liquid state. The expression ªmetal foamsº, strictly valid only for the liquid phase, is often used to describe the solid product. Mixtures of metal powders and blowing agents are compacted by extrusion or hot pressing providing a precursor material foamed above the solidus temperature, a method called ªpowder compact foamingº (Alulight [5 10], IFAM-Foaminal [3,5 10], Alufoam [5,7]). The Formgrip material [3,7] is made by remelting a stir-cast foamable precursor metal matrix composite. Low-density-core material (LDC [3,6]) is produced by pore formation in the solid state by the high gas pressure of entrapped dissociating blowing agents. A blowing agent powder is mixed into the melt either in a crucible (Alporas [1±7]) or in the gate in pressure die casting (Buehler [10]). Foaming of particle-reinforced metals takes advantage of the stabilization of gas bubbles by the ceramic ingredients producing a cellular metal matrix composite from the melt (Cymat [1 7], COMBAL [10]). The formation of a gas metal eutectic is the principle for the production of Gasar foams [1,2,6]. A slip reaction foam technique based on foaming of precursor slurry by chemical reaction and a reaction sintering process for aluminides has been described [10]. The originally closed cells of foamed metals may not be gas tight after solidification owing to cracks in their walls. The resulting cellular metal products can be differentiated by their structural features. The term ªstructureº is used for the description of cellular materials at different levels of observation (structology): the geometric architecture of the solid (skeleton) in the individual cells and their 3D arrangement, the variations of that


architecture within a considered sample or part (degree of uniformity), and the microstructure of the solid itself and its surface. The cells are formed by plateau borders (edges and nodes) and, in the case of closed cells, by walls connecting them [12] with a particular microstructure of the metal, eventually containing the remnants of the foaming additives and other microstructural inhomogeneities. The multitude of structural features (see Chapter 4) and their spatial distribution are still subject to investigation in order to correlate them with processing parameters and functional properties, aiming for the development of quality-relevant specifications. Cellular metals are inherently heterogeneous: real samples usually exhibit local non-uniformity like variations in cell architecture and mass distribution. The simplest description of a cellular metal is given by stating the production process, the composition of the metal, and the apparent density: producer/process alloy composition apparent density [g/cm3]. The established symbol for the corresponding heat treatment condition of light metals could be added; for example: Alulight AlSiMgCu 0.45 T1.

References

1. L. J. Gibson, M. F. Ashby, Cellular Solids: Structure and Properties, 2nd ed., Cambridge University Press, Cambridge 1997. 2. M. F. Ashby, A. G. Evans, N. A. Fleck, L. J. Gibson, J. W. Hutchinson, H. N. G. Wadley, Metal Foams: a Design Guide, ButterworthHeinemann, Woburn 2000. 3. J. Banhart, N. A. Fleck, M. F. Ashby (eds), Metal Foams Special Issue, Adv. Eng. Mater. 2000, 2(4). 4. J. Banhart (ed.), Proc. MetallschaÈume, MIT, Bremen 1997. 5. J. Banhart, H. Eifert (eds), Proc. Metal Foam USA Symposium, MIT, Bremen 1998. 6. D. S. Schwartz, D. S. Shih, A. G. Evans, H. N. G. Wadley (eds), Proc. Porous and Cellular Materials for Structural Applications, MRS Symp. Proc. Vol. 521, MRS, Warendale, PA 1998.

7. J. Banhart, M. F. Ashby, N. A. Fleck (eds), Proc. Metal Foams and Porous Metal Structures, MIT, Bremen 1999. 8. T. W. Clyne, F. Simancik (eds) Proc. Metal Matrix Composites and Metallic Foams, Euromat 1999, Vol. 5, Wiley-VCH, Weinheim 2000. 9. H. P. Degischer (ed) Proc. MetallschaÈume, Special Issue, Mater. Wissenschaft Werkstofftechn. 2000, 31(6). 10. J. Banhart, M. F. Ashby, N. A. Fleck (eds), Proc. Cellular Metals and Metal Foaming Technology (MetFoam 2001), MIT, Bremen 2001. 11. A. BuÈhrig-Polaczek, in Proc. Materialsweek 2000, Symposium H3, http://www.materialsweek.org/proceedings. 12. D. Weaire, S. Hutzler, The Physics of Foams, Clarendon Press, Oxford 1999.

7

8

2.1 Foaming Processes for Al

2.1

Foaming Processes for Al

C. KoÈrner and R. F. Singer

The melt route for processing closed-cell metal foams is very attractive since this approach allows economic handling of large quantities of material. Melt-route processes are also well suited to the use of scrap as feedstock. In order to foam the melt properly, gas must be introduced. This can be either done by gas injection or by in-situ gas generation by a chemical decomposition of a foaming agent. For the production of homogenous foams some prerequisites have to be fulfilled. If foaming is by in-situ gas generation uniform dispersion of the foaming agent in the melt within a time that is short compared to the decomposition or reaction time of the additive is required. In addition, the escape of gas during the foaming process has to be prevented. One important step to meet these requirements is to increase the viscosity of the melt. There are several approaches to do this: foaming in the semi-liquid state, incorporation of ceramic particles [1] or oxidation [2]. The effect of the particles added is always twofold: beside increasing the melt viscosity they stabilize the cell walls. Independent of the specific foaming process it is found that porosity, quantified by density, and mean cell size are intimately related. Generally, for a particular alloy or composite the mean cell size can not be chosen independently of the porosity. 2.1.1

Gas Injection: the Cymat/Alcan and Norsk Hydro Process

The Cymat/Alcan and Norsk Hydro melt-foaming process is a continuous, gas-injection method developed simultaneously and independently by Alcan [1] and Norsk Hydro [3] in the late 1980s and 1990s. A sketch of the process developed by Alcan is given in Fig. 2.1-1. The patent is now licensed and exploited by the Cymat Aluminum Corporation [www.cymat.com]. The process employed by Hydro Aluminum, Norway is analogous. A metal matrix composite (Al-wrought or Al-casting alloy matrix 10 30 vol.-% SiC or Al2O3

Figure 2.1-1. Principle of the melt-foaming route employed by Cymat. The foam casting process for producing flat panels consists of melting and holding furnaces, the foaming box and foaming equipment, and a twin-belt caster [1].


particles) is used as a starting material. The starting material is molten with conventional foundry equipment and transferred to a tundish where gas, typically air, is injected via small nozzles incorporated into a rotating impeller, thus forming a dispersion of small gas bubbles. The bubble size can be controlled by adjusting the gas flow rate, the impeller design (number of nozzles and their size), and the speed of rotation of the impeller. The gas bubbles rise to the surface where they accumulate. The ceramic particles are trapping gas bubbles owing to the favorable interface energy and serve as stabilizer of the cell walls and delay their coalescence. They also reduce the velocity of the rising bubbles by increasing the viscosity of the melt. That is, the particles reduce the kinetic energy of the rising bubbles and hence the danger of mechanical rupture when they arrive at the surface. The resulting metal foam, which is still liquid, is carried away by means of a conveyor belt where it solidifies and cools. The relative density is predominantly controlled by the process parameters, such as rotor speed, gas flow, and the amount of particles in the melt, and finally the solidification condition. This process of casting aluminum foam is capable of producing slabs with a relative density in the range 2 20 % (0.05 0.55 g/cm3). The average cell size is inversely related to the density (Fig. 2.1-2a) and is in the range 2.5 30 mm [1].

a)

Average cell size as a function of the reciprocal density for: a) Cymat foam [1]; b) FORMGRIP [7] products.

Figure 2.1-2.

b)

9

10


Figure 2.1-3. Optical micrographs of Cymat foam produced by the gas-injection method: left) cell structure (density about 0.3 g/cm3), inhomogeneous and anisotropic; right) foam surface (density about 0.05 g/cm3).

The production facility set up by Cymat is capable of casting foam panels in continuous length at an average rate of 900 kg/h up to 1.5 m wide with a thickness range of 25 150 mm. This shows that the process is relatively straightforward and economical. Challenges that may require more work in the future include the variation in cell size, the density gradient, and the anisotropy of the cell structure, which results from mechanical forces from the conveyor belt (Fig. 2.1-3). Especially earlier prototype have been compressed by the conveyor belt producing flattened pores causing poor stiffness and strength values along the thickness of the slabs. It should also be noted that the preparation of the composite feedstock requires relatively tedious long time stirring processes to achieve the proper homogenous distribution of the particles in the melt. In principle, this foam generation technology also allows the casting of nonrectangular, 2D profiles as well as 3D shapes. 2.1.2

In-situ Gas Generation: the Shinko Wire Process and the FORMGRIP process

The gas-injection method suffers from the fact that a relatively small number of large bubbles is generated, which leads to rather coarse and irregularly shaped pore distribution. Two techniques are described below, in which the foaming gas results from a thermal decomposition of solid ingredient. In this way a huge number of bubble nuclei is created throughout the melt.

The Shinko Wire Process [2] The manufacturing process of the Alporas foam is a batch casting process patented by Shinko Wire Company Ltd., Japan (see Fig. 2.1-4) [2]. The installed manufacturing plant is capable of making large sized blocks of foamed aluminum. For adjusting the viscosity of the molten aluminum 1.5 % Ca is added at 680 hC and stirred for 6 min in an ambient atmosphere. The addition and subsequent agitation of an element with a high oxygen affinity facilitates an oxidation process on the surface of the molten metal and leads to an increase of 2.1.2.1


Figure 2.1-4.

Manufacturing process for Alporas foams [2].

the viscosity by the formation of oxides: CaO, Al2O3, CaAl2O4. There is an appropriate stirring resistance for optimizing the foaming ratio [4]. The thickened aluminum is poured into a casting mold and stirred with an admixture of 1.6 % TiH2 as a foaming agent. While vigorously stirring, the TiH2 dissociates and H2 -bubbles are formed causing molten material to expand and to fill the mold. Then, the foamed material is cooled by fans to solidify in the casting mold. After removal from the casting mold. An Alporas block 450 mm wide, 2050 mm long, and 650 mm high is sliced into plates. Alporas is an ultra-light material with a closed-cell architecture (Fig. 2.1-5). The density of the product is 0.18 0.24 g/cm3, the mean cell size is about 4.5 mm. Alporas is regarded as the best commercially available aluminum foam in terms of regular cell micro structure. This is to a certain extent because transfer and deformation of the still liquid foam, as in the Cymat/Norsk Hydro process, is avoided. The cell architecture is the outcome of a growing process where bubble expand and coalesce due to cell wall rupture.

Figure 2.1-5. Typical cell structure of an Alporas foam. As a result of the growth process most cells are far from equiaxed. The global homogeneity is superior to other commercially available aluminum foams.

11

12


The FORMGRIP Process The FORMGRIP (Foaming of Reinforced Metals by Gas Release in Precursors) process integrates some of the advantages of melt- and powder-route approaches for foam production into one processing technique [5 7]. It comprises ªbakingº a precursor material with the foaming agent entrapped in a mold to generate foam in-situ by dissociation, which is very similar to the established powder routes. However, the precursor is prepared via melt processing [7]. A diagram of the FORMGRIP process is depicted in Fig. 2.1-6. The precursor preparation comprises dispersion of a mixture of AlSi12 powder and the pre-treated gas-generating TiH2 powder in an Al-9Si/SiCp (particle size 12.8 mm) composite melt by conventional mechanical stirring 1200 rpm for 50 70 s. The critical point is that during this processing step only a limited portion of hydrogen is released from the foaming agent. This is achieved by the following precautions: a low temperature of the melt at the moment of the hydride introduction (T 620 hC), a high melt viscosity due to SiC particles and a pre-oxidized, retarded foaming agent [8]. The thermal pre-treatment of the foaming agent consists of a two-step thermal oxidation sequence (24 h at 400 hC 1 h at 500 hC), which slightly reduces the hydrogen concentration of the TiH2. The oxide barrier layer formed on the powder surface slows down the kinetics of gas evolution. The amount of the incorporated hydride is about 1.5 wt.-% of the melt mass. The mixed melt is cast into a mold by cooling and pore growth is suppressed by cooling. The SiCp particles are also important for foam stabilization, which becomes plausible from their distribution in Fig. 2.1-7. The resulting precursor material already exhibits a porosity of 14 24 %. The second stage of the FORMGRIP process comprises heating of the precursor material above the solidus temperature. As a result, hydrogen released from the TiH2 diffuses to the bubble nuclei already present and expands them further. Typical cell structures are depicted in Fig. 2.1-8 [7]. 2.1.2.1

Figure 2.1-6. Diagram of the melt-based FORMGRIP process for production of near net-shape metal foam parts [7].


Optical micrographs illustrating the distribution of SiC particles in Al-9Si alloy based FORMGRIP foams, showing sections through: a) cell wall, b) a node. A significant fraction of particles is located at the gas/melt interface [5].

Figure 2.1-7.

Figure 2.1-8. Examples of cross sections of aluminum alloy FORMGRIP foams baked under different conditions. Porosity, P, levels and the mean cell sizes, d, are: a) P 69 %, d 1.1 mm; b) P 79 %, d 1.9 mm; c) P 88 %, d 3.1 mm [7].

The relation between relative foam density and average cell diameter fits that derived for Cymat foams. Fig. 2.1-2b shows the mean cell diameter as a function of the reciprocal density (see Section 4.1). The mean cell diameter is inversely proportional to the density indicating that foam expansion is governed by cell coalescence and the mean cell wall thickness is constant. The dependence of the cell diameter on the density is the same for both contents of SiC particles. The influence of particle size has not yet been investigated. Theoretical work of Kaptay [9] and experimental work of Weigand [10] indicate that a reduction of particle size will not lead to a higher stabilization and therefore to a smaller critical cell wall thickness. In terms of geometrical complexity and simultaneous microstructural control, the FORMGRIP process surpasses the Shinko Wire one. There is no need to transfer material from a mold into a die while the foaming process is going on. In addition, the precursor material can be shaped before the final baking step. The economics of the FORMGRIP process, however, are clearly inferior to the other processes discussed in this chapter. This is due to the discontinuous nature of the process as well as its various number of processing steps.

13

14

2.2 Industrialization of Powder-Compact Foaming Technique

References

1. J. T. Wood, ªProduction and Application of Continuously Cast, Foamed Aluminumº in Proc. Fraunhofer USA Metal Foam Symposium, 7 8 October 1997, Stanton, Delaware. 2. T. Miyoshi, M. Itoh, S. Akiyama, A. Kitahara, ªAluminum Foam, Alporas: The Production Process, Properties and Applicationsº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 125. 3. P. Asholt, ªAluminium Foam Produced by the Melt Foaming Route Process, Properties and Applicationsº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 133 140. 4. L. Ma, Z. Song, ªCellular structure control of aluminium foams during foaming process of aluminium meltº Scripta Mater. 1998, 39(11), 1523 1528. 5. V. Gergely, T. W. Clyne, ªThe FORMGRIP process: foaming of reinforced metals by gas release in precursorsº Adv. Eng. Mater. 2000, 2(4), 175 178.

6. V. Gergely, T. W. Clyne, ªA Novel Melt-Based Route to Aluminium Foam Productionº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 83 89. 7. V. Gergely, ªMelt Route Processing for Production of Metallic Foamsº, Department of Materials Science and Metallurgy, Cambridge 2000. 8. A. San-Martin, F. D. Manchester, ªThe H Ti (hydrogen titanium) systemº Bull. Alloy Phase Diagrams 1987, 8(1), 30 43. 9. G. Kaptay, ªInterfacial Criteria for Ceramic Particle Stabilised Metallic Foamsº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 141 146. 10. P. Weigand, ªUntersuchung der Einfluûfaktoren auf die pulvermetallurgische Herstellung von AluminiumschaÈumenº, FakultaÈt fuÈr Bergbau, HuÈttenwesen und Geowissenschaften, RWTH, Aachen, MIT Verlag, Bremen 1999.

2.2

Industrialization of Powder-Compact Foaming Technique

J. Banhart and F. BaumgaÈrtner

There are many different ways to manufacture cellular materials [1]. One of the available processes has become increasingly popular in the past few years and is at the stage of industrial implementation now. The method is sometimes loosely called the ªpowder-metallurgical routeº, but the term ªpowder-compact foaming techniqueº seems more appropriate. 2.2.1

Principles of Foam Production

The technique consists of mixing aluminum or aluminum alloy powders with appropriate foaming agents, which get entrapped by compacting this mix to a dense product called ªfoamable precursor materialª. The powder mix can be compacted directly by hot pressing, conform extrusion, or powder rolling. Alternatively the pow-


Figure 2.2-1.

Foam production by powder-compact foaming technique.

der may be cold compacted for better handing in conventional extrusion or rolling (Fig. 2.2-1). Heating the precursor above its solidus temperature releases the pressure on the foaming agent allowing decomposition and formation of bubbles. After cooling a low-density foam structure of originally closed cells is obtained [2,3]. The method is not restricted to aluminum and its alloys: tin, zinc, lead, gold, and some other metals and alloys can also be foamed by choosing appropriate foaming agents and process parameters (see Section 2.1.2). The most common alloys for foaming, however, are pure aluminum or wrought alloys such as aluminum 2xxx, 6xxx, or 7xxx alloys, e. g. AA 2014, 6060, 6061, 6082, or 7075. Casting alloys such as AlSi7Mg (A356) and AlSi12 are also frequently used because of their low melting point and good foaming properties, although in principle virtually any aluminum alloy can be foamed by carefully adjusting the process parameters. Quite complex-shaped metal foam parts can be manufactured by expanding the foam inside a mold, thus confining spatial expansion. An example for one such part is shown in Fig. 2.2-2. The part, developed in the framework of a feasibility study, is a novel pantograph horn for an electrical locomotive. This light-weight solution based on aluminum foam replaces traditional cast aluminum parts saving 30 % weight. A nice feature of the technique is that composite structures consisting of an aluminum foam and bulk metal parts can be made without using adhesives. Examples are foam-filled aluminum sections and sandwich panels with an aluminumfoam core and metallically bonded steel, aluminum, or even titanium face sheets. For making such composites the foamable precursor material is first bonded to the solid section or sheet by co-extrusion or roll-cladding, after which the foamable core layer is expanded by heat treatment [4,5] (see also Section 3.3).

15

16

2.2 Industrialization of Powder-Compact Foaming Technique Figure 2.2-2. Aluminum foam part (Schunk Sintermetalltechnik, Giessen).

The advantages of the powder-compact route are obvious and are listed in Table 2.2-1. Beside the first two features already mentioned, the flexibility arising from the preparation of the precursor from powders is important. Alloys can be made simply by mixing inexpensive elementary powders. No ceramic additives are needed to stabilize the foam, in contrast to some of the melt-route foaming processes in which up to 15 % silicon carbide has to be added [1,6]. However, if required, ceramic powders, metal fibers, or ceramic fibers can be added to the powder blend for special applications, such as for reinforcement or to increase wear resistance. Naturally, there are also some disadvantages that are inherent to the process. Metal powder is more expensive than bulk metal and requires effort for compaction. This rules out applications that require very cheap materials. Moreover, the size of aluminum foam parts that can be manufactured is limited by the size of the baking furnace, and is therefore smaller than for some of the competing melt-foaming processes. The largest sandwich components that have been manufactured using the powder-compact foaming technique are about 2 m q 1 m q 1 cm in size (possibly larger in future). At LKR in Ranshofen a part of similar size was produced without face sheets (Fig. 2.2-3). True 3D-volume parts are usually not thicker than 30 cm, a limit which is difficult to shift to higher values. A large aluminum foam column produced at Fraunhofer IFAM was 1 m high and 18 cm in diameter, weighing 13 kg. In contrast, the liquid-metal route allows for making panels 15 m in length [7] and 100 cm thickness [8]. However, as these processes cannot be used for near-net-shape production and only permit very simple geometries, they are appropriate for different fields of application. Continuous foaming of long products is under investigation [9]. The middle column of Table 2.2-1 lists some of the problems that are still encountered when foaming aluminum with the powder-compact melting method but which can, in principle, be solved with further research.

2 Material Definitions, Processing, and Recycling Table 2.2-1. Characteristics of powder-compact foaming method: advantages and disadvantages that are inherent are listed together with points that are presently problematic, but can be solved in principle.

Advantage

Problem

Disadvantage

Net-shape foaming possible

Uniformity of pore structure still not satisfactory

Cost of powders

Composites can be manufactured

Process control must be improved

Very large volume parts difficult to make

Parts are covered by metal skin

Permeable (holes)

Coating process requires sealing

Graded porosity can be achieved

Difficult to control

Flexibility in alloy choice No stabilising particles have to be added Ceramics and fibers can be added

Rear wall of an automobile made of aluminum foam (LKR Ranshofen and DaimlerChrysler AG, see chapter 7.4).

Figure 2.2-3.

2.2.2

Practical Aspects of Foam Production Powder selection The appropriate selection of the raw powders with respect to purity, particle size and distribution, alloying elements, and other powder properties is essential for successful foaming. Commercial air-atomized aluminum powders were shown to 2.2.2.1

17

18


be of sufficient quality. However, powders from different manufacturers led to notable differences in foaming behavior and empirical criteria have been derived to facilitate the selection of powders. The cost of powders and the ability of a manufacturer to provide sufficient quantity with a constant quality are also crucial. As already pointed out, alloys can be obtained in different ways. The frequently used alloy AlSi7, for example, can be either prepared by atomizing a AlSi7 melt, or by blending pure aluminum powder with 7 wt.-% silicon powder, or, in a third way, by mixing 58 % of standard AlSi12 powders with 42 % aluminum powder.

Mixing The mixing procedure should yield a homogeneous distribution of alloying elements and the foaming agent to ensure that high-quality foams with uniform pore-size distributions are obtained. Powders are mixed in batches of 500 kg at Schunk-Honsel in commercial large-scale tumbling mixers with parameters determined in technological tests. Alternatively, powder mixes can be obtained by aerodynamic mixing. For example, Alulight International GmbH Austria mixes aluminum and titanium hydride in large containers with 50 80 short pulses of pressurized nitrogen gas. 2.2.2.2

Densification Powder consolidation can be carried out by various techniques. It has to be ensured that the foaming agent is completely embedded in the metal matrix and no residual open porosity remains. One way to obtain foamable precursor material with nearly 100 % theoretical density is the combined use of cold isostatic pressing (CIP) and ram extrusion. CIP is first applied to consolidate the powder mix to billets with a relative density of 70 80 % and a mass of typically 50 kg. These billets are used in the subsequent extrusion step. Although CIPping is not absolutely necessary (powders have been filled into thin-walled aluminum cartouches and inserted into the extrusion machine without prior consolidation) it has additional advantages such as the prevention of powder contamination and powder de-mixing. The CIP billets themselves are not foamable because of their large content of residual porosity, which causes massive hydrogen losses when the material is heated. To obtain foamable material, the billets are preheated to 350 hC and extruded as rods or any other profile. For this a horizontal direct extrusion machine is used (25MN Schunk-Honsel). The extrusion machine is operated in cycles with a new billet inserted after each extrusion step. This way rather high outputs can be achieved. Foamable material has also been manufactured by rotary continuous extrusion in the so-called CONFORM process by Mepura (Ranshofen) [10]. Here a rotating wheel is used to drag the powder into the consolidation chamber from which it is pulled off in radial direction as a compacted wire. Foamable wires of about 8 mm diameter were manufactured from wrought alloys containing titanium hydride. 2.2.2.3


Further processing of foamable material The extruded material can be foamed as it is after consolidation or it can be worked to the required shape. By conventional rolling, foamable sheets with thicknesses down to about 2 mm are produced. Optionally, the foamable raw material can be clad to conventional sheets of metal, of steel or aluminum for example, by attaching two sheets to either side of the foamable precursor before rolling. This way a purely metallic sandwich structure is obtained. By deep drawing, the sheets and the sandwiches may be transformed to 3-D-shaped sheets for special applications. In all cases it is favorable to start from near-net-shape precursors in order to minimize foam flow [11]. 2.2.2.4

Foaming Heat treatment at temperatures above the solidus temperature of the foamable matrix is necessary to produce the foam structure. The gas released by the decomposing foaming agent may form pores in the solid state but only above the solidus are bubbles formed and the matrix expands up to a maximum volume, that is to a minimum density. The density and density distribution of the growing foam can be controlled by several parameters. The foaming agent content in the precursor material is obviously important, but furnace temperatures and heating rates also have an influence [12]. The mold material, the mold shape, and the type of furnace naturally influence the heating rate and have therefore also to be considered. A careful control of the heating conditions during foaming is essential for obtaining high-quality foams. The difficulty is that the liquid foam is thermodynamically unstable and conditions change constantly during foaming. There are various intermediate stages: at first only the mold is heated directly, whereas the foamable material receives heat only indirectly via heat conduction through the mold. Initially there are merely some point contacts between the piece of foamable material in the mold and the mold walls. However, as the temperature increases, the precursor softens and assumes the contour of the mold thus increasing the transfer of heat. Moreover, heat transfer via radiation gains importance with rising temperatures. The reflectivity of the mold and precursor surfaces may change during the process and add a further variable. Finally, after foaming has started, the thermal conductivity of the precursor rapidly decreases thus reducing heat flow. As soon as the mold has been filled with foam it has to be cooled down below its solidus temperature to stabilize the foam structure. The phenomena during cooling are also quite complex and difficult to describe for reasons similar to those mentioned for the heating phase. Typical densities of aluminum foams are in the range 0.4 0.8 g/cm3 including the closed skin around the foam body. The final density of a foamed part can be simply predicted if the volume of the hollow mold and the mass of the inserted precursor material are known. The foaming mold may be loaded with several small pieces or one single piece of precursor. Choosing the latter method (which is preferred by LKR and SAS [13]) one has to take into account that each piece of the expanding precursor material has a dense aluminum oxide layer on its sur2.2.2.5

19

20


Left) top of aluminum foam part made by inserting various pieces of foamable material into the mold (dark figures indicate original size of the precursor pieces). Right) foam part made of two pieces of precursor without achieving bonding between the two pieces.

Figure 2.2-4.

face, which has to be broken up by expansion of the individual foam pieces. Incomplete foaming may cause the foamed pieces to remain separated even after the foaming process (Fig. 2.2-4b). A relative movement of the foam pieces to each other helps to break up the oxide films. Fig. 2.2-4 shows an example of a successful formation of a foamed body from various pieces of the precursor and an example of failure. In the former case the location of the original individual foam pieces can still be identified from the contrast in gray scales between the various regions: darker gray identifies oxide layers of extruded surfaces, brighter gray is the new (expanded) surface. This effect is currently exploited to create foam panels and other foamed parts for making designer objects. 2.2.3

State of Commercialization

Currently the foaming technique described is still in the stage of industrial implementation. Nevertheless, a number of companies have already made commitments for a future production and are building up facilities [14]. The joint effort of Schunk Sintermetalltechnik (Gieûen) and Honsel GmbH&Co KG (Meschede) is one example. Owing to their collaboration with Karmann the activities are preferentially directed towards foam and foam sandwich parts with a complex 3D geometry (see Section 3.3). Alulight International GmbH is another example. It is a joint venture of SHW (Germany) and Eckart Austria. The company offers aluminum foam panels in sizes up to 625 mm q 625 mm, with thickness of 8 25 mm. Neuman Alufoam, another Austrian company, also offers foamable precursor material (extrusions) and foamed parts.


References

1. J. Banhart, ªManufacture, characterisation and application of cellular metals and metal foamsº Prog. Mater. Sci. 2001, 46, 559 632. 2. J. Baumeister, German Patent DE 40 18 360, 1990. 3. J. Banhart, ªFoam metal: the recipeº Europhysics News 1999, 30, 17. 4. J. Baumeister, J. Banhart, M. Weber, German Patent DE 44 266 27, 1994. 5. H.-W. Seeliger, ªApplication Strategies for Aluminium-Foam-Sandwich Parts (AFS)º in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 23. 6. P. Asholt, ªAluminium Foam Produced by the Melt Foaming Route Process, Properties and Applicationsº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 133. 7. Cymat Corp. (Canada), Product information sheets, http://www.cymat.com 1999.

8. T. Miyoshi, M. Itoh, S. Akiyama, A. Kitahara, ªAluminum Foam, Alporas: The Production Process, Properties and Applicationsº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 125. 9. G. Stengele, H. MuÈcke, A. SchoÈne, German Patent DE 197 34 394 A 1, 1998. 10. H. P. Degischer, H. WoÈrz, DE Patent 4206303, 1992. 11. F. BaumgaÈrtner, H. Gers, ªBauteile aus AluminiumschaÈumenº Ingenieur Werkstoffe. 1998, 3, 42. 12. I. Duarte, J. Banhart, ªA study of aluminium foam formation kinetics and microstructureº Acta. Mater. 2000, 48, 2349. 13. R. Kretz, F. Simancik, private communication. 14. http://www.schunk-group.com, http:// www.alulight.com, http://www.neuman.at

2.3

Making Cellular Metals from Metals other than Aluminum

G. Rausch and J. Banhart

The previous section was dedicated exclusively to aluminum foams. For many applications one would like to use cellular materials made from metals or alloys other than aluminum. There have been some attempts to manufacture metal foams by simply adapting the powder-compact process originally developed for aluminum to other metals by adjusting the properties of the foaming agent and the process parameters. This procedure was successful in some cases. However, for high-melting alloys the powder-compact foaming technique is difficult to implement and especially for titanium no promising results could be obtained. Here alternative routes based on advanced powder metallurgy yielded better results. Therefore, in the current section the topic will be slightly extended from ªfoamedº to ªcellular or porous metalsº in a more general sense.

21

22

2.3 Making Cellular Metals from Metals other than Aluminum

2.3.1

Zinc

Zinc can be foamed by a straight-forward modification of the powder-compact technique. The foaming agent used for aluminum (TiH2) can be used, although ZrH2 seems to yield slightly better results. Powder properties and mixing procedures are quite similar to aluminum. Only the pressing and foaming temperature has to be chosen slightly lower than for aluminum due to the melting temperature of zinc at 419 hC. Foamed zinc shows a very uniform pore structure. This can be attributed to the fact that the decomposition temperature of the foaming agents TiH2 matches with the melting temperature of the metal. Therefore, melting and pore formation occur simultaneously and round bubbles are created from the very beginning. In contrast to aluminum, there is no solid state expansion range with corresponding crack formation. Fig. 2.3-1a shows an example of a zinc foam. 2.3.2

Lead

Lead and lead alloys such as Pb Sn and Pb Sb can be foamed by another modification of the process. TiH2 and ZrH2 cannot be used as foaming agents because of the low melting temperatures of pure lead (327 hC) and even lower solidus temperature of the alloys. Quite good foams have been obtained by using lead(II) carbonate as a foaming agent: it decomposes above about 275 hC and releases CO2 and water, which act as foaming gas. Fig. 2.3-1b shows an example of a lead foam 2.3.3

Titanium

Owing to its high melting temperature (1670 hC) and relatively low density (4.51 g/ cm3), titanium and its alloys are excellent materials for lightweight applications at elevated temperatures and are widely used in aeronautical applications. Porous ti-

a) Figure 2.3-1.

b)

Zn and Pb foams (width of sample is about 5 cm).


tanium structures have an additional potential for weight reduction and could even be suitable for functional applications if the pore structure were open. In principle, there are many possible production methods for cellular materials based on titanium (see Section 2.4), most of them starting from metal powders. 1. 2. 3. 4.

Consolidation of slurry-saturated plastic foam. Foaming and sintering of powder slurries. Reaction sintering of elemental powder mixtures. Foaming of powder compacts containing foaming agents (powder-compact melting process). 5. Hot isostatic pressing and creep expansion of titanium compacts with entrapped inert gas. 6. Sintering of hollow spheres. 7. Sintering of compacted or loose powder filler mixtures. While some of these methods (1 3) have not yet been investigated very intensively, the feasibility of the foaming agent process (4) for titanium has been demonstrated [1]. However, owing to the high temperatures during foaming titanium, the reactivity of this metal with practically any non-inert gas and the lack of appropriate foaming molds, this method is not suitable for producing shaped titanium foam components. Hot isostatic pressing of titanium powder with gas entrapment (5) has been successfully developed for some aircraft applications [2]. Metal hollow spheres (6) can be produced using wet chemical methods for coating Styrofoam spheres [3]. Shaping and sintering of these hollow structures typically result in materials with very low porosity. One of the most promising methods for manufacturing open porous titanium materials is the sintering of compacted or extruded mixtures of powders and fillers that contain removable space-holder materials. The materials are mixed and shaped by conventional PM techniques. After removal of the space holder the green samples are sintered at temperatures of 1100 1400 hC. Bram and coworkers use urea and ammonium hydrogen carbonate as space holders [4], which can be removed by thermal treatment below 200 hC. Depending on the size and shape

Figure 2.3-2. Open porous titanium made by space-holder technique: left) pore size 1 4 mm; right) pore size about 500 mm, porosity 55 80 %.

23

24


Figure 2.3-3.

Pore structure of open porous titanium with 67 % porosity.

Figure 2.3-4. Strength and Young's modulus as a function of density obtained from bending and tension tests [5].


of the space-holder powder, spherical and angular pores in the range 0.1 2.5 mm can be obtained, resulting in overall porosities of 70 80 %. It was found that the sintering activity can be increased by partially substituting titanium by titanium hydride, thus yielding an increased compression strength. At Fraunhofer IFAM, polymer granules were used as the space holder. They were removed by a chemical process at temperatures around 130 hC, after pressing. After space-holder removal, samples are sintered in vacuum at temperatures of 1100 1250 hC. Depending on the particle size of the granules, average pore diameters in the range 200 3000 mm can be obtained. Fig. 2.3-2 shows some typical samples. Fig. 2.3-3 shows the typical pore structure of samples based on spherical space-holder granules. Beyond the primary pore structure, some microporosity (secondary pores) inside the sintered network is visible. It was shown that the secondary porosity has a strong influence on the overall strength of the samples and can be reduced by either changing the sintering parameters and/or partially replacing titanium powder by titanium hydride [4]. As for all porous materials, the mechanical properties of cellular titanium are a function of density. Fig. 2.3-4 shows the strength and Young's modulus obtained from bending and tension tests as a function of density. 2.3.4

Steel Powder-Compact Foaming Technique The long experience in making aluminum foams from powder metallurgy (PM) precursors encouraged researchers to transfer this process to higher-melting materials such as iron-based alloys and steels. The major requirements for an adaptation of the foaming agent process to this group of materials are the following. 2.3.4.1

x x x x

Selection of suitable foaming agents. Development of alloys qualified for good ªfoamabilityº. Evaluation of compaction methods. Adaptation of the foaming process.

The basic requirements for foaming agents are: point of gas emission above 1000 1200 hC (depending on the alloy composition), broad temperature range of gas emission (up to 1550 hC for nearly pure iron), and sufficient volume of gas release. It was found that especially metal nitrides and certain carbonates show a significant gas emission and qualify for being useful as foaming agents. Examples are manganese nitride, chromium nitride, molybdenum nitride, calcium carbonate, strontium carbonate, and barium carbonate [5,6]. Theoretical investigations [5,7] have shown that both the iron carbon and the iron boron system [7] are able to meet the basic requirements for being foamed to iron-based metallic foams, namely: a low melting point matching the decomposition temperature of the foaming agent, and a broad two-phase semisolid region in the phase diagram, thus creating a wide foaming interval. As for the production of Al based foamable pre-

25

26


Iron-based foams obtained from the foaming agent process.

Figure 2.3-5.

cursor material, extrusion has been successfully used for compacting iron powder mixtures. The resulting samples are shown in Fig. 2.3-5. Experiments with powder mixtures of iron and carbon have shown that free carbon without any additional foaming agent already leads to a certain degree of porosity. Carbon is oxidized during the foaming process and the resulting gaseous CO and CO2 creates pores. However, pore size distributions are not uniform and pore shape is usually rather irregular. The porosity mainly results from large, isolated pores (Fig. 2.3-6). Adding 0.25 % SrCO3 leads to an increase of porosity to 55.5 % (Fig. 2.3-7). The pore structure at this composition appears to be more homogeneous and the average pore size is obviously lower. Increasing the amount of SrCO3 results in a further increased porosity (64.3 %). From that it can be concluded that SrCO3 has a significant influence on the achievable porosity and the maximum expansion. The foaming agent technique has therefore been shown to be feasible for steel. However, foaming of stainless steel or even superalloys has not yet been successful and the general state-of-the-art of foaming steel with the foaming agent method is still far behind the aluminum foaming technology.

Iron-based metal foams made from extrusion-pressed powder mixtures of Fe 2.5 % C: left) 0.0 % SrCO3, middle) 0.25 % SrCO3, right) 0.50 % SrCO3.

Figure 2.3-6.


Figure 2.3-7.

Average porosity as a function of foaming agent content (SrCO3).

Steel Foams from Powder Filler Mixtures All foamed metals have essentially closed cells. For certain applications (filters, membranes, biomedical applications) open porosity is required. For this class of materials the space-holder technique (see also Section 2.3) can be used. The process used for steel is very similar to the one described in the titanium section. The process starts with a mixture of metal powders and the filler powder. The mixture is compacted, usually by axial compression in a conventional powder press. If necessary, an additional bonding agent is used in order to achieve a better strength of the green samples. After pressing an additional drying step is optional. After this the filler/bonding agent phase is removed from the samples, in a chemical (catalytic) or thermal process. After complete filler removal the samples are sintered in a furnace under hydrogen atmosphere. Either urea [4] or plastic granules [8] can be used as space holders. In Fig. 2.3-8 examples of porous 316L and Inconel 600 materials are given, developed by Forschungszentrum JuÈlich GmbH. The porosity of these materials is about 70 % with an average pore size of 1.0 1.4 mm. 2.3.4.2

SEM images of sintered specimen: left) stainless steel 316L, 1100 hC, 1 h, particle size I16 mm; right) Inconel 600, 1250 hC, 1 h, particle size 100 200 mm [4].

Figure 2.3-8.

27

28

2.4 Recycling of Cellular Metals

References

1. G. Rausch, T. Hartwig, M. Weber, O. Schultz, ªHerstellung und Eigenschaften von TitanschaÈumenº Materwiss. Werkstofftechn. 2000, 31, 412 414. 2. R. L. Martin, R. J. Lederich, Metal Powder Rep. 1992, Oct, 30. 3. O. Andersen, U. Waag, L. Schneider, G. Stephani, B. Kieback, ªNovel metallic hollow sphere structuresº Adv. Eng. Mater. 2000, 2, 192 195. 4. M. Bram et al., ªPreparation and Characterization of High-Porosity Titanium, Stainless Steel and Superalloy partsº in Metal Foams and Porous Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 197 202.

5. B. Kriszt, A. Falahati, H. P. Degischer, ªMachbarkeitsstudie zur Herstellung von Eisenbasisschaumº in MetallschaÈume, J. Banhart (ed), MIT Verlag, Bremen 1997, p. 59 70. 6. C.-J. Yu, H. Eifert, M. KnuÈwer, M. Weber, ªInvestigation for the selection of foaming agents to produce steel foamsº Mater. Res. Soc. Symp. Proc. 1998, 521, 145 150. 7. M. KnuÈwer, Herstellung von Eisenschaum nach dem pulvermetallurgischen Treibmittelverfahren, Dissertation, UniversitaÈt Bremen, Fraunhofer IRB Verlag, Stuttgart 1999. 8. G. Rausch, M. Weber, M. KnuÈwer, ªNeue Entwicklungen zur Herstellung von StahlschaÈumenº Materwiss. Werkstofftechn. 2000, 31, 424 427.

2.4

Recycling of Cellular Metals

H. P. Degischer

Cellular metals compete with polymers for some applications. The recyclability of metals is one of their benefits, enabling ecologically sustainable product life cycles [1]. Compared to bulk metal products, there are two complications to be tackled in remelting cellular metals: x

x

The high surface-to-volume ratio of the order of 100/length unit increases the extent of surface adsorptions and reactions. The low average density, due to the high porosity filled with gas, makes the cellular material float on its melt.

2.4.1

The Remelting of Cellular Metals

Heating cellular aluminum in air enhances the growth of the oxide film not only in open cellular structures, but also in those foamed by formation of closed pores in the liquid, or semi-solid state, which usually become permeable to gas after solidification and cooling. When heating the cellular aluminum up to the melting point in a short time, about 1 h, the thickness of the oxide film may reach about 10 mm and any cracks formed will be covered quickly by


ongoing oxidation. The growth will slow down after prolonged heating, when the thickness reaches about 100 mm [2], which is similar to the cell-wall thickness. The dehydrated oxide is stable up to about 2000 hC, so the cellular aluminum part is slowly converted to a cellular alumina structure, which may maintain its macroscopic shape if it does not break up under its weight, or an external force. It is difficult to submerge a cellular part in a melt because it tends to float and contains a lot of air, which would have to be replaced by melt. Therefore the cellular structures have to be compressed as much as possible, as is done with packaging foils, or shredder scrap before melting [1,3]. Cellular metals can be shredded too and treated as normal shredder scrap. Nevertheless, the high content of surface oxides reduces the efficiency of metal reclamation. The only experience is from small batch laboratory trials on scrap from powder-compact foamed aluminum [4], in which efficiencies up to 80 % have been reached. Oxides and other impurities have to be removed by the usual cleaning process for melts to the extent necessary for the later production process [3]. The purity requirements may be very low for the production of cellular metals. Impurities in precursor material prepared for powder-compact foaming, as well as for foaming in the melt, might be advantageous, as long as they act as nuclei for the formation of pores. There might be even an up-grading of lowquality scrap when used for the production of foamed metals [5]. The Ti remaining from the blowing agent (usually I0.5 wt.-%) is soluble below 0.12 wt.-% at the peritectic temperature of the binary aluminum melt [1] and does not degrade the quality of the alloy, but provides a grain-refining effect during solidification. ALPORAS foam contains about 2 wt.-% Ca, which is an element to be restricted below 0.1 wt.-% in all cast alloys. Ca will be partly oxidized and thus transferred into the dross, but some care has to be taken in the mix of ingredient scrap for not exceeding the specified impurity levels for the secondary alloy. The impurity levels of secondary aluminum cast alloys are not as stringent [1,3], so that scrap of the common types of cellular aluminum can be added to the remelting furnace. 2.4.2

Recycling of Cellular Metal Matrix Composites

Cellular aluminum produced by foaming particulate-reinforced melts, like the Cymat process [6] (formerly developed by Alcan [7] and Hydro [8]), the FORMGRIP process [9], and the shape foaming technique COMBAL [10] may be collectively named MMC foams. It might be of interest to recycle the MMC matrix of the foam without significant reduction of the particle content, because of the economic value of reclaimed MMC based on the primary processing cost to overcome the nonwettability. Any reuse of MMC saves the effort necessary to bond the two components together during processing. Originally, this type of foam was a spin-off of particulate-reinforced aluminum processing, consequently it has been proposed to reuse particle-reinforced aluminum alloys in the production of aluminum foam [11]. One advantage of this recycling route is that the specification requirements

29

30


for the production of metallic foams, may not be as stringent as placed on MMC for bulk components, thus higher quantities of oxide skins and other impurities may be tolerated, or are even advantageous in foamable aluminum. Figure 2.4-1 shows the material flow cycle for discontinuously reinforced metals and cellular metals made from MMC indicating the possible interactions between these two cycles: x

x

x

Remelted MMC, with, or without cleaning treatments, can be used as a base material for the melt foaming processes and for FORMGRIP. Any secondary alloy, including that reclaimed from MMC melts, can be used for production of cellular structures by any of the known production methods. Scrap from MMC foams can be introduced into the MMC cycle, if quality requirements are met.

The same rules have to be obeyed for the recycling of MMC foams as for the MMC matrix [11] in addition to the pretreatment (compaction and drying) to reduce gas evolution. Two main problems have to be tackled when remelting MMC. x x

The reactivity of the reinforcement with the melt increases with temperature. The dewetting tendency of the reinforcement limits conventional melt cleaning methods.

Figure 2.4 -1. Recycling of discontinuously reinforced metals and of metal foams based on MMC, indicating the reuse of secondary MMC for the processing of foamed metals either by the powdercompact method, or by the melt-foaming technique.


The reaction between 6xxx type wrought alloys and alumina reinforcements is driven by the Mg content forming MgAl2O4, spinel. An original level of approximately 2 vol.-% spinel seems to stabilize at just above 3 vol.-% after several remelting cycles; this does not influence the mechanical properties significantly [12], but may increase wettability and, consequently, reduce foamability. SiC-reinforced wrought alloys cannot be remelted without severe aluminum carbide formation at the interfaces. The formation of Al4C3 would be detrimental as it significantly reduces the corrosion resistance of the component. Furthermore, it affects the foamability by increasing the viscosity of the melt and above all by providing wettability. Nonwettability of the additives is necessary for the trapping of gas bubbles. In the case of Al casting alloys containing 7 12 wt.-% Si reinforced by SiC, there is the chance to conserve the integrity of SiC by keeping the melt temperature below 750 hC as recommended for primary foundry technology [13]. A detailed study of the recycling of SiC-particle reinforced Al Si casting alloys is given elsewhere [14], where the quality criteria of the melt are given and remelting, recycling, and holding practices are described. x x

x x x

x x

Dry, pre-heated scrap can be added to the melt between 700 and 750 hC. During a rest period, usually more dross is formed than on primary material. It contains mainly oxides and SiC-particles without reducing significantly their content in the melt. The dross has to be skimmed. An impeller is introduced to produce strong movement under the surface skin. Fluxing and degassing with argon (SF6 may also be used) to remove oxides and reduce the hydrogen content. The melt is allowed to sit and then skimmed. The melt is mechanically agitated, without forming a vortex, to distribute the SiC particles homogeneously.

No loss in particle content was reported and the removal of porosity, oxide films, and hydrogen were efficient. The amount of dross generated is relatively high and may amount to more than 10 % of the total weight. If the ceramic particles should be removed from the aluminum melt, conventional salt addition, or fluxing techniques (as are executed to remove oxide films [14]) can be applied. Gravity settling allows the fluxed ceramic ingredients to float to the dross at the top of the aluminum melt. Rotary salt furnace technology is an established reclamation process to recover aluminum from various mixtures, including particle reinforced metals; however, this requires 20 50 wt.-% salt [13]. Both wrought and foundry alloys and even machining chips can be recovered using this technique. Recovery of about 80 % of the available aluminum can be expected [15]. The efficiency of particle removal by fluxing is related to the probability of contact between the ceramic constituent and the flux. Duralcan, a supplier of particulate-reinforced aluminum, proposes to incorporate the salt into a melt agitated by gas injection [16]. Thus dewetting is achieved with much smaller salt additions (I1 wt.-% for alumina and about 1.5 wt.-% for SiC) and in combination with

31

32


the adsorption of gas to the reinforcement also accelerates particle separation by floating, where it can be skimmed off. 2.4.3

Conclusions

There are the following possibilities for recycling, or reuse of cellular metals (especially cellular aluminum). x

x

x

x

Recycling of unreinforced cellular metals can be carried out as for packaging material (foils), or shredder scrap, by remelting, with an efficiency probably reduced to 80 % to produce bulk cast products from the secondary metal. In the case of recycling ALPORAS, the impurity level of Ca has to be controlled. Recycling of MMC foams can be done by extracting the matrix metal for conventional secondary products. The reinforcement may be segregated to the dross and deposited. Recycling of MMC foams can alternatively be achieved by remelting according to the precautions for MMC recycling, but at a reduced efficiency due to the increased oxide content to produce MMC foams again. The foam production itself may be based on an up-grading of secondary material with higher contents of impurities that increase foamability. In particular, the cycles for MMC and MMC foams can be closely related.

References

1. Grundlagen und Werkstoffe, Aluminium Taschenbuch, Vol. 1, Aluminium Verlag, DuÈsseldorf 1995. 2. American Society of Materials, Proc. 4th ASM Int. Conf. Recycling of Metals, 17 18 June 1999, Vienna, ASM International, Metals Park, OH. 3. K. Krone, Aluminium Recycling, Verein Deutscher Schmelzhuetten e. V., DuÈsseldorf 2000. 4. M. Strini, Private communication, ARCLeichtmetall Kompetenzzentrum Ranshofen, Austria 1996. 5. H. P. Degischer, F. Simancik, ªRecyclable Foamed Aluminium as an Alternative to Compositesº in Environmetnal Aspects in Materials Research, H.Warlimont (ed), DGM, Oberursel 1994, p. 137 140. 6. Cymat Corporation, Mississauga, Canada 2000, http://www.cymat.com.

7. I. Jin, L. D. Kenny, H. Sang, US Patent 5 112 697, 1992. 8. W. W. Ruch, B. Kirkevag, NO Patent 1989, World Patent wo 91/01387, 1991. 9. V. Gergely, T. W. Clyne, ªA Novel MeltBased Route to Aluminium Foam Productionº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 83 89. 10. D. Leitlmeier, H. Flankl, ªDevelopment of a New Processing Technique Based on the Melt Route to Produce Near Net Shape Foam Partsº in Proc. METFOAM 2001, MIT Verlag, Bremen 2001, p. 171 174. 11. V. Gergely, H. P. Degischer, T. W. Clyne, ªRecycling of MMC and Production of Metallic Foamsº in Comprehensive Composite Materials, Vol. 3, T. W. Clyne et al. (eds), Elsevier, London 2000, p. 797 820.

2 Material Definitions, Processing, and Recycling 12. D. M. Schuster, M. D. Skibo, R. S. Bruski, and Fabrication of Light Metals and Metal R. Provencher, G. Riverin, ªThe recycling and Matrix Composites, Montreal, TMS, 1992, p. 598 604. reclamation of metal-matrix compositesº 15. T. F. Klimowicz, ªThe large scale commerJ. Metals 1993, 45(May), 26 30. cialization of aluminum natrix compositesº 13. Duralcan, Duralcan Composite Casting J. Metals 1994, Nov, 49 53. Guidelines, Duralcan USA, San Diego, CA 16. Duralcan, Aluminium Recovery from Metal 1990. Matrix Composite Scrap, Duralcan USA, Novi, 14. R. Provencher, G. Riverin, C. Celik, Michigan 1996. ªRecycling of Duralcan Aluminium Metal Matrix Compositesº in Proc. Adv. Production

2.5

The Physics of Foaming: Structure Formation and Stability

C. KoÈrner, M. Arnold, M. Thies, and R. F. Singer

A foam is a dispersion of gas bubbles in a liquid in which the bubbles are deformed due to their mutual interaction. Cellular metals produced via gas bubbles in liquid metal are commonly named metal foams although they do not strictly meet the definition above after solidification. Owing to the surface energy necessary to form the metal±gas interface a foam is never in equilibrium and hence permanently trying to lower the internal energy by reducing the internal surface. That is, the cellular structural state of a foam evolves with time and its actual structure is a function of its history including all thermal and mechanical influences. The foam structure is normally strongly disordered and evolves by some combination of three basic mechanisms: bubble coalescence via film rupture; bubble coarsening via diffusion of gas from smaller to larger bubbles; and drainage downwards and out from the foam in response to gravity [1]. It is only in recent years that some progress has been made towards an understanding of the basic mechanisms governing the temporal evolution of foams [1±5]. Owing to their opacity and the high temperatures, in-situ observation of the structure evolution of metal foams is difficult [6]. Fortunately, the solidification process is in general much faster than the evolution processes at least as far as the evolution is not caused by expansion. That is, nearly full information of the foam formation process can be extracted from ex-situ investigation of foamed samples in different stages of expansion.

33

34

2.5 The Physics of Foaming: Structure Formation and Stability

2.5.1

Isolated Gas Bubble in a Melt

Since a foam consists of interacting gas bubbles it is helpful to consider in a first step an isolated gas bubble in an infinitely extended fluid. The dynamics of a spherical bubble with radius R is described by the Rayleigh equation [7,8] S RR

R_ 3 _2 1 2s (Pbubble s s PT ) R S 4v R 2 r R

(1)

where Pbubble bubble pressure, PT equilibrium pressure in the liquid, s surface energy, n kinematic viscosity, r density. The left hand side of Eq. (1) describes the inertia and viscous forces that both delay bubble growth. Neglecting viscous effects the time for the formation of a bubble in aluminum with radius R 1 mm at a bubble overpressure of DP 10 4 bar is given by [7] r r (2) 0:015s t 0:915 R DP That is, inertia effects delaying bubble expansion can be neglected if foam formation is on a time scale of seconds. The contribution of viscous forces for liquid metals is normally very small, for example about 4nr(R/RÇ) z 10 6 bar with RÇ 1 mm/ s; R 100 mm; n 1 mm2/s. On the other hand, if foaming takes place in the semi-liquid state, where the viscosity of the metal is several orders of magnitude higher than in the liquid state, viscous forces might delay bubble expansion. Bubble expansion takes generally more than one second for commercially known foaming methods for metals [9]. In this case, the viscous and inertia forces can be neglected and the Rayleigh equation reduces to pressure equilibrium at the gas±liquid interface Pbubble PT S 2

s s P0 S rgh S 2 R R

(3)

where g gravity constant, P0 ambient pressure, h depth. The pressure contribution resulting from the surface energy (s 0.2 N/m for Al) is 0.04 bar and 4 bar for bubble diameters of 100 mm and 1 mm, respectively. For h 100 mm the hydrostatic pressure is about 0.02 bar. Owing to gravity there is a pressure gradient present in the melt that deforms the bubble and makes it move. The bubble is accelerated until a stationary velocity, v, is reached where the resulting viscous forces balance the buoyant forces (Fig. 2.5-1). For nearly spherical bubbles the rising velocity v can be calculated from Stokes' law [7] s rgR2 Q v for R II (4) 3h rpg where h nr dynamic viscosity.

2 Material Definitions, Processing, and Recycling Figure 2.5-1. Velocity field around a rising bubble in a liquid. The velocity and deformation of the bubble depends on the viscosity of the melt, the surface tension, and the bubble size.

For a pure aluminum melt and bubble radii R of 100 mm and 10 mm the rising velocity is about 1 cm/s and 100 mm/s, respectively. Additives in the fluid like SiC or Al2O3 particles influence the movement of the bubbles and are able to stabilize them [10±12]. They have an effect on both the viscosity of the melt and the surface tension. How these particles actually operate and how their action can be optimized is not yet understood and still a matter of research. A gas bubble grows or shrinks due to gas exchange with the surrounding melt [13]. There is a gas flow from the liquid into the bubble if the concentration of dissolved gas in the liquid, as a result of the decomposition of a foaming agent for example, is higher than the equilibrium concentration in the liquid given by Henry's law for a given gas pressure in the bubble. Since hydrogen, which is preferentially used as foaming gas, dissociates when dissolved in aluminum the equilibrium hydrogen concentration, ceq, at the gas±liquid interface is governed by Sievert's law, a special form of Henry's law [14] ceq 1:4 q 10s3 q 10s

2760 T

p Pbubble

mol p cm3 bar

(5)

where T temperature of the melt. 2.5.2

Agglomeration of Bubbles: Foam

A foam is an agglomeration of many gas bubbles taking a polyhedral form due to their mutual interaction (Fig. 2.5-2). Usually, the volumes separated by thin walls are named cells. As mentioned before a foam is not a static structure but always evolving towards a lower internal energy. That is, already during the formation of a foam there are processes going on which try to alter the structure. In a pure liquid these processes are so fast that the development of a foam is not possible. For this reason additives have to be added to stabilize the foam. Foams made without a suitable additive boil, that is, the development of long-lasting membranes is not possible and the

35

36

2.5 The Physics of Foaming: Structure Formation and Stability Figure 2.5-2. Cell structure of an aluminum foam produced by the FORMGRIP process foamed under different conditions: a) porosity, 79 %; mean cell diameter, 1.9 mm; b) porosity, 88 %; mean cell diameter, 3.1 mm. The arrow marks a residual portion of a ruptured cell wall [30].

2D-lattice Boltzmann simulation of the growth of bubble nuclei by in-situ gas generation. The cell walls are not stabilized, so two bubbles coalesce if the cell wall reaches a

Figure 2.5-3.

critical lower value. The melt boils, the number of bubbles decreases, and gas is lost to the environment. The structure collapses once a critical expansion factor is reached [31].

structure is unstable (Fig. 2.5-3). In order to obtain a foaming material additives have to be added to the liquid, which slow down cell-wall rupture [15]. Consequently, a metal foam never consists of a pure metal. There must always be additives like SiC or Al2O3 present [11,16]. These additives are either added deliberately from outside [11], produced during processing by oxidation [17], or are already present in form of oxides if metal powders are used [18]. There are two different strategies to introduce the cell forming gas into the melt: by injection through a nozzle [11,19] or by in-situ gas segregation or generation [9,17]. The latter can be achieved by a chemical decomposition of a foaming agent or by creation of a supersaturation of gas in the melt. If foaming is realized by gas injection the gas bubbles are directly created without bubble nucleation and growth. The bubble size is determined by various parameters such as the nozzle geometry, the gas flow rate, and the impeller speed. Experience shows that the bubbles generated are relatively large. The bubble shape depends on the viscosity of the melt, the bubble size etc. Foaming by in-situ gas generation starts with homogeneous or heterogeneous bubble nucleation followed by bubble growth by gas diffusion into the nuclei [13] (Fig. 2.5-4). The rate at which bubbles nucleate homogeneously, NÇ0, is given by [20] N_ 0 c0 f0 es

DGhom kT

with DGhom

16ps 3 3DP2

(6)


Figure 2.5-4. Expansion stages of an AlMg1 foam with 0.4 % TiH2 produced by powder compaction. By applying a high ambient pressure of 100 bar during heating bubble nucleation was suppressed until the sample was

completely liquid. Already at the very beginning of foam expansion coalescence occurs frequently indicating that the initial number of nuclei will not have much influence on the residual foam structure.

where c0 concentration of gas molecules, f0 frequency factor of gas molecules joining the nucleus, k Boltzmann constant, DGhom activation energy for homogeneous nucleation, DP gas pressure resulting from the dissolved gas following Sievert's law. Heterogeneous nucleation occurs when a bubble forms at an interface between two phases, between a ceramic particle and the melt, for example. The rate for heterogeneous nucleation, NÇ1, is given by [20] N_ 1 c1 f1 es

DGhet kT

(7)

where c1 concentration of heterogeneous nucleation sites, f1 frequency factor of gas molecules joining the nucleus, DGhet activation energy for heterogeneous nucleation. In the presence of heterogeneous nucleation sites this type of nucleation will be favored over homogeneous nucleation because of its lower energy barrier. Potential nucleation sites are the foaming agent particles and the particles present for foam stabilization and alloying [18]. For polymers it is known that the nucleation rate can be influenced by the processing parameters [20±22]. A systematic investigation of the nucleation rate for metals as a function of the processing parameters is yet missing but it is expected that the mechanisms are very similar to those in polymers. Owing to the presence of additional particles in the melt, heterogeneous nucleation is probably the main nucleation mechanism. The shape of the starting gas bubbles changes during growth from spherical to polygonal (Fig. 2.5-4). From a pure energetic point of view one would expect the formation of Kelvin or Weaire±Phelan cells [3], which minimize the total internal surface and hence the total internal energy. Actually, most of the cells are far away from these ideal structures. That is, the complex formation process of the foam does not necessarily lead to the energetically preferred structure. When two cells coalesce due to cell-wall rupture the resulting cell is strongly deformed (Fig. 2.52b). It is hampered to take an energetically more suitable form due to the presence of the other cells. As a result, the cell geometry is in general not equilateral and the

37

38


cell walls are curved due to the pressure difference of neighboring cells. One effective way to characterize the cell structure is by the shape factor, F, calculated from a 2D cut through the foam [23] (see also Section 4.1) F

n 4p X ai n i li2

(8)

where n total number of cells, i cell number, li cell-boundary length, ai cell area. The shape factor describes the deviation of the cell geometry from a circle with F 1. It is found that F decreases much more with decreasing foam density than expected from the transition of spherical to polygonal cells (Fig. 2.5-5) [23]. That is, the deformation of the cells and the occurrence of bent cell walls increases continuously during foam expansion. Thus, polygonal cell structures for different densities are not self-similar. A further mechanism that might lead to a structural change of the foam is coarsening analogous to Ostwald ripening. As a result of the pressure difference of about 10 3 bar of neighboring cells with different size, there is a concentration gradient (see Eq. 5) and therefore a flow of dissolved gas from smaller to larger cells. Hence, small cells shrink and large cells grow, so the foam structure coarsens [24]. This effect is very pronounced for aqueous foams due to the small cell wall thickness (about 100 nm). Numerical calculations [25] confirm the experimental finding that this kind of coarsening is an effect of secondary importance on the relevant time scale for the production of aluminum foams with a typical cell wall thickness of 50±300 mm. On the other hand, gas loss by diffusion to the environment can be substantial for long holding times if the hydrogen partial pressure in the ambient atmosphere is zero (Fig. 2.5-6). The formation of an aluminum foam by in-situ gas generation is from the very beginning intimately correlated with cell coalescence (Fig. 2.5-4). If foam expansion would be a mere enlargement of the structure without cell coalescence, so the number of cells is constant during foaming, one would expect the mean pore diameter, D, to be proportional to rrel1/3 (rrel relative foam density). The

Mean shape factor, F, of the cell structure as a function of the foam density for an AlSi10Mg0.6 foam produced by powder compaction [23]. Energetically optimized cell structures would show a form factor equal to that of a pentagon.

Figure 2.5-5.

2 Material Definitions, Processing, and Recycling Figure 2.5-6. Cell structure of an Alulight foam after heating to 670 hC under an ambient pressure of 100 bar and subsequent foaming by pressure reduction to 11 bar. The thick metal layer at the surface is the result of gas loss to the environment during the 15 min in the liquid state.

last expression is deduced from a simple cubic plate model for the cells with the assumption that the volume of the cell-forming material per cell is constant. Actually, the experimental data (see Fig. 2.5-7) is much better fitted by D

1s

d 3d p for rrel II 1 3 1 s rrel rrel

(9)

which follows from a cubic plate model assuming a constant mean cell-wall thickness, d. That is, the mean cell-wall thickness is constant during expansion. It depends on the alloy and is found between 50 mm and 300 mm for aluminum foams. Consequently, given a fixed alloy cell size and relative density are intimately related and can hardly be influenced by processing parameters [23]. In order to generate aluminum foams with a lower density and smaller cells the mean cellwall thickness has to be reduced. How the cell walls of metal foams are stabilized by the particles and which properties of them (quantity, size, form) determine the mean cell-wall thickness is not yet understood and a matter of flow research [18]. Coalescence, rupture of a cell wall, is a mechanical process that depends on the mechanical stability of the membrane. It has statistical character and occurs if a local fluctuation of the film thickness leads to faster local thinning because it is not compensated by restoring forces. The different physical mechanisms that are responsible for the mechanical stability of a membrane have been extensively investigated for aqueous films [5,15]. Experimental studies of the stability of metal membranes are not known in literature. Physical forces between the two surfaces of the film like the van der Waals or electromagnetic interaction are only relevant

Figure 2.5-7. Mean pore diameter as a function of the reciprocal foam density for a wrought alloy (open symbols) and a cast alloy (full symbols) foam produced by powder compaction. The samples were foamed at various ambient pressures. The full lines give the results for a cubic plate model (Eq. 9) with constant mean cellwall thickness of 180 mm for the wrought alloy and 130 mm for the cast alloy [31].

39

40


for very thin films (I1 mm). For thicker films, as it is the case for metal foams, there are two effects that might lead to a kind of elasticity of the membrane. The Gibbs' effect [15] takes into account that a local thinning of the membrane leads to a local thinning of the surfactants. As a result, the surface energy increases at this point leading to a restoring force. The Marangoni effect [15] describes that a gradient of the surface energy leads to a flow of surfactants taking with it a thin film of fluid in the direction of lower concentration. In order to maximize both effects the surfactant should minimize the surface energy. At the moment, it is not clear whether the stability of metal membranes is based on one of these two mechanisms. In metals, melt surfaces are covered with oxides and other particles that might have the required effects on the surface energy. The origin for cell-wall instability is cell-wall thinning, which can be traced back to different mechanisms. The first one is simply related to foam expansion itself. If a cell expands the cell-wall thickness decreases due to mass conservation. During foam expansion the principal mechanism for cell-wall thinning is by growth. The second mechanism of cell-wall thinning is drainage due to gravity and capillary forces. Drainage determines the foam stability in the absence of further expansion due to gas release, that is, the time that a given foam structure is maintained without destruction. Gravity induces a melt flow from the top to the bottom of the foam and generates a density gradient [4,5]. Neglecting capillary forces, the draining velocity, vD, can be estimated in a first approximation by the balance of gravity and viscous stress from shear flow within the Plateau borders [1] vD Z

grd2edge h

(10)

That is, the draining velocity increases with the cell-edge thickness, dedge. Capillary forces arise due to the different curvature of the interfaces. They force the melt from the cell membranes to the cell edges (Fig. 2.5-8). The rate of cell-wall thinning due to this pressure gradient can be modeled using a Reynolds-type equation for the flow between two circular parallel discs with immobile wall surfaces [5] x_ s

2sx 3 3hR2 RPB

(11)

where x cell-wall thickness, 1/RPB curvature of the Plateau border, R radius of the disc. Integration of the last equation shows that cell walls of 100 mm in pure aluminum reduce to 1 mm in I1 s. This result is in contrast to experimental observations where metal foams are held in the liquid state for several minutes without a structural change. How a metal foam attains this high stability is not yet understood. From a practical point of view foam collapse is very important. Foam collapse usually proceeds from the outer surface to the interior and is closely related to the appearance of a dense metal layer at the bottom (Fig. 2.5-9).

2 Material Definitions, Processing, and Recycling Cellular automaton simulation of the material transport from regions with lower curvature to higher curvature due to capillary forces.

Figure 2.5-8.

Figure 2.5-9. Collapse of an aluminum foam (powder route precursor AlSi10Mg0.6). Collapse proceeds from the outer surface to the interior, the metal of the collapsed cells drains down to the bottom. Left) the sample was heated to 665 hC at 1 bar ambient pressure and further expanded at constant temperature by pressure reduction to 500 mbar. Right) the sample was heated to 665 hC at 1 bar ambient pressure and held at constant temperature for 30 min.

Collapse happens if the amount of gas loss to the environment by diffusion or by cell-wall rupture is substantial and not compensated by in-situ gas generation. One has to distinguish between collapse by growth and collapse by aging, due to a long holding time. Foam collapse by growth occurs when a critical expansion factor is reached. In this case, cell-wall rupture on the foam surface leads to a pronounced loss of gas to the environment. Collapse by aging results from gas loss to the environment by diffusion and cell-wall rupture due to drainage by gravity and capillarity. For the production of foam parts foam rheology plays a crucial role. The rheological bearing is important when mechanical forces are present. During foam formation there are always mechanical forces acting on the foam. These might result from a spatial restriction like a mold, from an injection molding process [26], or a conveyer belt on which the foam is transported before solidification [11]. The mechanical response of foams to applied forces is very complex. Foams show wall slip, compressibility, and non-Newtonian viscoelasticity [27]. They exhibit a nonzero shear modulus although made out of gas and liquid, which both individually display a vanishing shear modulus. Consequently, a foam shows a very complex rheological behavior including bubble deformation, rearrangement, and avalanche processes [28,29]. The resulting pore structure is in general not isotropic but orientated in foaming direction. This behavior is even more pronounced if the foam fills a complicated mold with dimensions of the same order as the mean pore dia-

41

42

2.5 The Physics of Foaming: Structure Formation and Stability Figure 2.5-10. AlSi10Mg0.6 foam produced by powder compaction. Since heating was not homogenous, the internal structure represents different stages of the foaming process. The expansion of the central region leads to a compression of the exterior cells that were formed before.

meter. As a result, pores are deformed and also destroyed during mold filling by mechanical forces or a non-uniform foaming velocity (Fig. 2.5-10). Hence, the internal foam structure is strongly influenced by the particular production procedure.

References

1. A. Saint-Jalmes, M. U. Vera, D. J. Durian, ªUniform foam production by turbulent mixing: new results on free drainage vs. liquid contentº Eur. Phys. J. B 1999, 12, 67±73. 2. S. A. Koehler, S. Hilgenfeld, H. A. Stone, ªLiquid flow through aqueous foams: The node-dominated foam drainage equationº Phys. Rev. Lett. 1999, 82(21), 4232±4235. 3. A. M. Kraynik et al. ªFoam Micromechanicsº in Proc. Foams and Emulsions, Cargese, Corsica; Kluwer, Dordrecht 1999. 4. S. Hutzler, ªThe Physics of Foamsº, Department of Physics, University of Dublin 1998. 5. A. E. Bhakta, E. Ruckenstein, ªDecay of standing foams: drainage, coalescence and collapseº Adv. Colloid Interface Sci. 1997, 70, 1±124. 6. J. Banhart, et al. ªMetal foam evolution studied by synchrotron radioscopyº Appl. Phys. Lett. 2000. 7. L. D. Landau, E. M. Lifschitz, Hydrodynamik, 3rd ed, Akademie-Verlag, Berlin 1974. 8. S. F. Edwards, K. D. Pithia, ªA model for the formation of foamsº Physica A 1995, 215, 270±276.

9. I. Duarte, J. Banhart, ªA study of aluminium foam formation ± kinetics and microstructureº Acta Met. 2000, 48, 2349±2362. 10. V. Gergely, T. W. Clyne, ªThe FORMGRIP process: foaming of reinforced metals by gas release in precursorsº Adv. Eng. Mater. 2000, 2(4), 175±178. 11. J. T. Wood, ªProduction and Application of Continuously Cast, Foamed Aluminumº in Proc. Fraunhofer USA Metal Foam Symp. 7±8 October 1997, Stanton, DW. 12. G. Kaptay, ªInterfacial Criteria for Ceramic Particle Stabilised Metallic Foamsº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 141±146. 13. A. Arefmanesh, S. G. Advani, ªDiffusioninduced growth of a gas bubble in a viscoelastic fluidº Rheologica Acta 1991, 30, 274±283. 14. P. Lutze, J. Ruge, ªWasserstoff in Aluminium und seinen Legierungenº METALL 1990, 65, 649±652. 15. H. Lange, SchaÈume und ihre StabilitaÈt, VDI-Berichte, 1972, 182, 71±77.

2 Material Definitions, Processing, and Recycling 16. S. W. Ip, J. Wang, J. M. Toguri, ªAluminium foam stabilization by solid particlesº Canad. Metall. Q. 1999, 38, 81±92. 17. T. Miyoshi et al. ªAluminum Foam, ALPORAS: The Production Process, Properties and Applicationsº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999. 18. P. Weigand, ªUntersuchung der Einfluûfaktoren auf die pulvermetallurgische Herstellung von AluminiumschaÈumenº, FakultaÈt fuÈr Bergbau, HuÈttenwesen und Geowissenschaften, RWTH, Aachen 1999. 19. P. Asholt, ªAluminium Foam Produced by the Melt Foaming Route Process, Properties and Applicationsº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999, p. 133±140. 20. J. S. Colton, N. P. Suh, ªNucleation of microcellular foam: theory and practiceº Polym. Eng. Sci. 1987, 27(7), 500±503. 21. H.-Y. Kwak, Y. W. Kim, ªHomogeneous nucleation and macroscopic growth of gas bubble in organic solutionsº Int. J. Heat Mass Transfer 1998, 41(4±5), 757±767. 22. C. B. Park, L. K. Cheung, ªA study of cell nucleation in the extrusion of polypropylene foamsº Polym. Eng. Sci. 1997, 37(1), 1±10. 23. C. KoÈrner et al., ªInfluence of processing conditions on morphology of metal foams produced from metal powderº Mater. Sci. Technol. 2000, July-August, 781±784.

24. C. Monnereau, M. Vignes-Adler, ªDynamics of 3D real foam coarseningº Phys. Rev. Lett. 1998, 80(23), 5228±5231. 25. C. KoÈrner, R. F. Singer. ªNumerical Simulation of Foam Formation and Evolution with Modified Cellular Automataº in Metal Foams and Porous Metal Structures, J. Banhart, M. F. Ashby, N. A. Fleck (eds), MIT Verlag, Bremen 1999. 26. F. Schorghuber, F. Simancik, E. Hartl, US patent 5 865 237, 1999. 27. W. Hanselmann, E. Windhab, ªÛber das Flieûen von Schaum in Rohren; Foam flow in pipesº Appl. Rheol. 1996, Dezember, 253±260. 28. D. J. Durian, ªFoam mechanics at the bubble scaleº Phys. Rev. Lett. 1995, 75(26), 4780±4783. 29. D. J. Durian, ªBubble-scale model of foam mechanics: Melting, nonlinear behaviour, and avalanchesº Phys. Rev. E 1997, 55(2), 1739±1751. 30. V. Gergely, ªMelt Route Processing for Production of Metallic Foamsº, Department of Materials Science and Metallurgy, Cambridge 2000. 31. M. Arnold et al., ªExperimental and Numerical Investigation of the Formation of Metal Foamº in Proc. Materials Week 2000 [http://www.materialsweek.org/proceedings/ index.htm].

2.6

Infiltration and the Replication Process for Producing Metal Sponges

C. San Marchi, A. Mortensen

The process of replicating a structure has been known to the metallurgist for centuries. For example, casting can be thought of as a replication procedure in which a metal is used to reproduce the negative form of the mold. In the study of porous materials, however, replication refers to a process used to replicate the open-pore architecture of a porous material.

43

44

2.6 Infiltration and the Replication Process for Producing Metal Sponges

2.6.1

Replication

More specifically, this process can be defined as a general three-step procedure for the production of highly porous materials (Fig. 2.6-1). The three steps are: 1. preparation of a removable pattern; 2. infiltration of the pattern followed by solidification (or an alternative process, such as curing or cross-linking, which rigidifies the infiltrate); and 3. removal of the pattern. In some cases, a fourth step, such as pyrolysis, may be required to transform the porous network into the desired phase. In this text, we focus on the use of the replication process for the production of cellular metals with open porosity. Several

Figure 2.6-1.

Diagram of the replication process.


authors have noted that metal (or metallic) sponge is perhaps the most appropriate terminology for these materials as both the term ªfoamº and ªcellularº imply a closed-cell architecture [1,2], while porous metals produced by replication must have an open-pore network. We will thus refer to metals with open porosity of any density I0.8, produced by replication, as metal sponge. This process is distinct from deposition techniques, such as those used by Inco Limited (Swansea SA6 5QR, Wales, UK) to produce porous nickel substrates, called Incofoam [3]. In that process, a temporary pattern is used, but the pore space is not replicated: instead, the pattern is coated and the pore space only partially filled. Sintering methods that utilize a space holder are also not considered replication for two reasons: the micro-architectural features of the pores are determined by the metal powder size in addition to the space-holding material and infiltration is not one of the processing steps. The so-called lattice block materials (LBM) are prepared using polymer patterns of truss structures and investment techniques; thus the lattice block is rather an engineered structure and cannot really be considered a metal sponge, which can be ªshapedº and ªformedº while maintaining the statistics of its properties, as long as the cell sizes are of millimeter or more dimensions. For more details on lattice block materials contact JAMCORP (17 Jonspin Road, Wilmington, MA 01887-1020 USA). Although we limit this discussion to metals, we note in passing that the polymeric sponge technique used to produce porous ceramics also does not replicate the pattern and should not be considered a replication process. In a metallurgical context, the replication technique can be used to produce metal sponge that is difficult or impossible to produce by other processing methodologies. Replicated metal sponge, for example, by necessity has an open-pore structure, which is a fundamental difference compared to most commercially available foamed metals. The structure or architecture of metal sponge produced by replication is flexible and determined by the pattern: porosity as high as 98 % [4] and as low as 55 % [2] for different pattern materials have been reported and pore sizes as small as 10 mm have been achieved [5]. In addition, metal sponge can be produced from virtually any alloy that can be cast (provided a suitable pattern material exists). The replication process has been used extensively for the production of various porous materials including carbon [6], silicon carbide [7], aluminum alloys [2,4,5,8±13], magnesium alloys [14,15] and alloys based on iron and nickel [4,16]. In the following sections we provide details about the production of metal sponges by replication techniques. We first describe the important physical phenomena that govern each of the three basic steps of the replication process (with some emphasis on sponge produced from salt patterns for which the most information exists in the literature). We then provide some physical and mechanical properties of metal sponge produced in this manner.

45

46


2.6.2

The Replication Process: General Principles Pattern Preparation The primary requirements of the pattern material are that it be sufficiently refractory at the casting temperature and that it be chemically stable when in contact with the melt. Additionally, after infiltration an open network must exist between the metal ligaments such that the pattern material can be removed. The choice of the pattern material necessarily determines the range of architectures that can be produced. Four types of pattern materials are described below: continuous refractory (investment), discontinuous refractory (pelleted casting sands), burnable (polystyrene and resin), and leachable (sodium chloride). Refractory investment casting material (continuous refractory pattern) is emerging as perhaps the most common space-holder for replication processing of lowdensity metal sponges [4,12±15]. A slurry of an appropriate investment is infiltrated into a preform that has the form of the desired metal sponge and that can be easily removed. Commercial polyurethane sponges (generally referred to as a foam) are typically used, as they are available with variegated properties in terms of porosity, pore size, and strut size (thickness of the ligaments or beams that make up the sponge) and can be easily burned out in air. Other preform materials can also be employed, provided that they can be removed without damaging the architecture of the investment. The main restriction here is that the investment must have a particulate size that is significantly smaller than the architecture to be replicated; also, its relative density must be sufficiently low that the solid investment grains can be removed from the metal±refractory composite. Investment casting is a well-developed technology that can be applied to almost any castable metal. Another conventional type of space holder is pelleted casting sand, or other granulated mineral (discontinuous refractory pattern) [16]. In this method casting sand is mixed with an organic binder to form comparatively coarse agglomerates that can then be filled into a mold to form the pattern. The porosity of the pattern is limited by the packing efficiency of the pellets, which are typically spherical. Maximum random packing efficiency of mono-sized spheres is about 0.64 [17] and can be greater for mixtures of different-sized pellets, although practical limits are determined by the necessity to remove the pattern after infiltration with the melt. Moreover, patterns of packed, discrete elements of this sort (as opposed to the continuous nature of investment pattern) must be infiltrated such that these space-holders can still be removed, that is they must not be completely embedded in the infiltrating metal; this is discussed below. In principle, pelleted casting sands can be used with any sand-castable alloy. A rather unusual pattern material is polystyrene [11] (burnable pattern), although any burnable space-holder can potentially be used in the same manner. Polystyrene spheres are coated with a resin and then filled into a mold before the resin is hardened. The resin acts to form the connected network between the spheres; this network is necessary for removal of the pattern. The pore size 2.6.2.1


is determined by the size of the initial spheres (or particles) and the ratio of polystyrene to resin. This type of pattern has the advantage that it is not necessary to physically remove solid grains from the structure after infiltration, as the pattern is essentially burned away at the end of the process. On the other hand, this pattern must be infiltrated relatively cold to avoid burning of the pattern during infiltration (i. e., while the metal is still liquid) and is thus limited to low-melting alloys such as aluminum, magnesium, and zinc. The fourth class of pattern material is leachable: these patterns are removed by dissolution in an appropriate solvent. The process of dissolution (or leaching) is less restrictive on size and density than insoluble refractory materials. In practice however, patterns of leachable material are created from granules or powder, thus the density is related to the packing efficiency of the granules, which itself depends on the size (and size distribution) of the granules. Sodium chloride (NaCl) has been employed most extensively as a leachable pattern, because it is inexpensive and easy to handle. Sodium chloride patterns used for the production of porous aluminum were first reported [8,9] and patented [10] in the 1960s, and have advanced more recently [5] (there are also reports of their use for carbon sponge [6] and for SiC sponge [7], both with a pyrolysis step). Salt patterns are limited by their melting point; NaCl, for example, is limited to aluminum and lower-melting alloys, while NaF could potentially be used at temperatures greater than 900 hC. Additionally, highly concentrated saline solutions are generated during dissolution and can cause significant corrosion in some alloys. Salt grains also have the advantage that they can be sintered to enhance the connectivity of the salt, change the structure of the pattern, and create a free standing preform that can be handled and subsequently infiltrated [5±7]. The sintering step is not necessary, however, for two reasons: transient bonding of the grains (perhaps due to humidity in the salt [8]) may occur during preheating the salt prior to and during infiltration, and incomplete penetration of the melt locally where individual grains meet, provides the necessary open-network for salt removal. Sintering processes depend on a number of parameters, including particle size, temperature, atmosphere, residual stresses, and time. Such numerous process parameters might explain some of the varying observations in the literature with regard to the sintering of sodium chloride. Incidentally, these varying observations point to the fact that sodium chloride may be a good choice for a pattern material because of flexibility in controlling the sintering mechanisms and thus the microarchitectural features of the pattern. There is a consensus in the literature that large sodium chloride particles sinter predominantly by an evaporation±condensation mechanism, in which the necks between the particles grow without densification of the salt compact. Small particle compacts, on the other hand, can densify significantly during sintering, due to the predominance of other mechanisms where the center-to-center distances between particles can decrease, such as bulk diffusion [18±21]. The minimum particle size that can sinter without densification has been estimated to be about 150 mm [19,20]. Sintering in vacuum may promote the evaporation±condensation mechan-

47

48


ism for smaller particles [22] as supported by the pressure dependence that was noted for the rate of neck growth when non-densifying sintering mechanisms are dominant [23]. There are advantages to both classes of sintering mechanisms with respect to producing metal sponge. Non-densifying mechanisms, such as evaporation±condensation should develop a structure dictated by capillary equilibrium: these are well known [24]. The pore architecture in such a case is expected to be similar to the regular structures of polymeric sponge. Densifying mechanisms, on the other hand, provide a means to achieve relative densities significantly greater (and thus significantly lower for the sponge) than those attainable by packing and pressing of particles. Sintered salt densities as high as 90 % have been reported with nearly all the pores as open porosity, although for a particle size of 1 mm, which would pose challenges in the subsequent processing steps [25].

Infiltration Ideally, infiltration under the influence of gravity is preferred as it does not require special equipment or procedures. In most cases, however, the metal does not wet the pattern and some additional force must be applied to the melt to promote (or assist) infiltration and effect a uniform distribution of metal within the pattern. Vacuum and low-pressure assisted casting procedures are standard foundry technologies and, in many cases, these are sufficient for infiltration. Indeed, it has been reported that spontaneous infiltration is possible when infiltrating salt grains greater than 4 mm in size. Vibration slightly reduces the size of salt grains that can be infiltrated under gravity, while vacuum applied to the pattern with a slight overpressure on the melt has also been used to infiltrate grains as small as a few hundred micrometers in diameter [8]. The specific pressure required to initiate infiltration (the threshold pressure, P0) depends to a large extent on the size of the pores to be infiltrated and the volume fraction of these pores. A detailed physical description of infiltration is beyond the scope of this work; more detailed information is available elsewhere [26±28]. To the authors' knowledge, the pressure required to initiate infiltration has not been explored with specific emphasis on patterns used for replication processes; however, expressions derived for the threshold pressure in the infiltration of packed ceramic beds with pure aluminum should be directly relevant. Garcia-Cordovilla et al. offer a simple semi-empirical relationship [27] 2.6.2.2

P0 MPa z 16

f s 0:09 (1 s f )D

where f is the volume fraction solid (or in this context the fractional density of the pattern) and D is a characteristic diameter of the solid in micrometers. This relationship is based on data for particles in the range 10±100 mm packed to a density of 50±60 % and infiltrated with pure aluminum, but it should provide reasonable magnitude estimates of the pressures required for infiltration over a broader range


of conditions since it contains the functional dependencies on f and D expected from theory [26±30]. This relation predicts that particles with an average diameter of 100 mm and packed to 75 vol.-% will require an applied pressure of a couple of atmospheres to initiate infiltration, while D 4 mm and f 0.98 requires the application of about one atmosphere of pressure to the melt, such that infiltration can be achieved by evacuating the pattern and taking advantage of atmospheric pressure. The degree to which the metal penetrates the pattern, particularly at the local level, is also an important consideration and obviously depends on the applied pressure. When the pattern material is porous, a very high applied pressure will cause the space between individual grains of investment to be infiltrated, such that these then cannot be removed. In a system with a wetting angle greater than ninety degrees, a low (yet finite) pressure will cause infiltration of larger pores within the pattern, yet cannot force the metal into small channels against adverse capillary forces. Thus, in patterns created from granulated materials such as pelleted sand, the regions near contacts between individual grains may remain uninfiltrated if the pressure is low enough. These open channels between adjoining discontinuous grains are necessary for the removal of the pattern. Control of the infiltration then is important to effect a uniform distribution of metal in the large channels of the pattern without penetrating the microstructural elements of the pattern when it is porous. Few details have been given in the literature about infiltration in the context of replication processing, although generally the techniques appear to be some form of vacuum or low-pressure assisted casting methods [4,12±16]. A modified die-casting method has also been used [11] and any other high-pressure casting methods such as squeeze casting should be appropriate. For example, gas-pressure infiltration at pressures up to 80 bar has been used to infiltrate patterns with small pores [5]. After infiltration is complete, the metal must, of course, solidify, which is an integral part of any casting process. Methods utilized to control solidification shrinkage depend on the specifics of the casting process, but generally consist of directional solidification in the case of gas-pressure infiltration or some type of riser or reservoir in general foundry methods.

Pattern Removal The pattern removal step depends on the characteristics of the pattern. Patterns from fine-grained ceramic powders such as investment casting compound and agglomerated silica pellets are typically removed by spraying water on the pattern± metal composite [12±16], such that the water physically breaks down the pattern into the small grains from which it is made. This requires penetration of the porous network by the spray, or at least that the fluid can ªwickº into the grain structure of the pattern so that it breaks down. For relatively fine (I1 mm) and dense sponges i10 %), however, it is unclear how efficient this process would be for sections many times the pore size. Additional methods such as (ultrasonic) vibration may also prove useful for removal of solid grains. 2.6.2.3

49

50


Polymer pattern materials can be removed by burning in air as is the case for polystyrene-resin patterns [11]. The cleanliness of this burning is an important issue in an industrial setting. These patterns, however, have the advantage that complex parts can be machined before removal of the pattern and without significant damage to the structure of the sponge. In some cases, leachable patterns can also be machined to complex geometries prior to leaching and without damaging the architecture of the sponge. In the case of salt, removal of the pattern is by immersion in water, causing dissolution of the salt. Dissolution is primarily a diffusive process, as the dissolved salt ions diffuse from their place-holding position into the water bath through the narrow channels in the sponge where the water is essentially stagnant. Thus, the rate and total time of dissolution depend strongly on the size of the specimen. Salt is more difficult to remove from small pores, particularly if gas is entrapped in the pores; this is the so-called ªgas lockº phenomenon, which can block the penetration of fluid into the porous network [8]. Corrosion of metal sponge during leaching of the pattern can be a problem due to the large surface area of the porous network, thus the salt concentration and the immersion time must be minimized. This is accomplished by simply changing the water periodically, however, it is important that the sponge is not drained at this point, as this will cause gas to be entrapped in pores and promote ªgas lock.º Additionally, impurities in the water should be controlled to avoid precipitation of other salts in these concentrated solutions. Aluminum alloys (the metals for which salt patterns are most applicable) tend to be relatively corrosion-resistant in salt solutions. The aluminum±copper alloys are among the least corrosion-resistant in salt solutions, but sponge can still be prepared from these alloys without degradation by corrosion. Electrochemical protection could also offer additional protection from corrosion. A large volume of water is not necessary as salt-saturated water is denser than distilled water, and hence falls to the bottom of the container where it remains in the absence of convection. As a result, during the dissolution process a large concentration gradient is maintained between the outer surface of the salt±metal composite suspended near the surface of the dissolution bath (but completely immersed) and the bottom of the container. For large pores and low-density sponge, mixing may modestly enhance dissolution if the velocity of the fluid in the porous network is affected. In most cases, however, this condition is not met and mixing the water primarily distributes the salt concentration and reduces the diffusion gradient in the water bath, in turn slowing the leaching process. The final stage of salt removal from the sponge is rinsing. This can be accomplished by flowing water through the structure, or simply by changing the water several times to incrementally remove the saturated fluid; this process is greatly facilitated by the higher density of salt solutions. Pressurized gas is also effective for removing the residual fluid in the sponge such that dissolved salts are not left in the structure when this water evaporates. Other methods may contribute to more efficient (faster) removal of salt, but these must enhance the rate of diffusion of the salt in the fluid. For example, ul-


trasonic vibration has been proposed as a useful method as it should increase the rate of diffusion with the added benefit of freeing gas bubbles entrapped in the porous network [8]. 2.6.3

Physical and Mechanical Properties of Metal Sponge

Published reports of metal sponge produced by replication techniques are summarized in Table 2.6-1. Relevant details of the processes and properties of the resulting metal sponge are briefly described below in the context of the basic pattern technologies: continuous refractory, discontinuous refractory, burnable, and leachable.

Continuous Refractory Patterns A variety of metal sponges have been produced from commercial polyurethane preforms: aluminum alloys [4,12,13], magnesium alloys [14,15] and several alloys having melting points up to 1500 hC [4]. Infiltration has been conducted with vacuum or pressure-assisted investment casting techniques. The investment is removed by means of a water spray, although additional specifics have not been supplied in the literature. Metal sponge with volumes up to 1500 cm3 have been prepared in this way; in addition, components, such as modules for heat exchangers integrally cast with plates and tubes, have been demonstrated [4]. The porous metals produced with this methodology have porosity in the range 92±98 % and pores sizes as small as 0.85 mm, Table 2.6-1. Practical limits of replication with investment patterns, however, have yet to be clearly identified and may include a wider range of structural parameters than reported to date. Although size effects (including density) have not been systematically explored, compressive strength (plateau stress) of aluminum and magnesium sponge with about 97 % porosity have been reported to be less than 0.25 MPa and in accordance with prediction based on phenomenological relationships [12±15]. Strain rate sensitivity has also been reported for these materials although a complete interpretation of these data has not yet been offered [31,32]. 2.6.3.1

Discontinuous Refractory Patterns Cast-iron sponge has been produced using pelleted silica sand [16]. The size of the pellets can be varied from about 1 mm to over 5 mm. The porosity of the sponges is reported to be about 60 %, Table 2.6-1. It appears that gravity casting methods were used, perhaps assisted with vacuum. The pattern is removed with pressurized water, but it is unclear if the pellets can be removed from a material several centimeters in depth. Preliminary mechanical data in compression have been reported and feature a significant drop in load carrying capacity after about 2 % strain. Permeability tests have also been reported though no systematic conclusions have been drawn. Incidentally, a similar sponge-like material (from aluminum) was produced as early as the 1930s in Germany, except that the pattern was not removed [33]. 2.6.3.2

51

Al (SG91A)

Al (Cu, Zn, Fe, Ni, Co)

pig Fe

Al

Al

Investment of polyurethane sponge


Pelleted casting sand

Polystyrene and resin

NaCl

Al alloys

* Represents range of properties as reported.

NaCl (presintered)

Mg (AZ91)


Al

Metal

gas-pressure infiltration

gravity, vibration assisted, pressure assisted by vacuum

modified die casting

vacuum assisted

vacuum or pressure assisted

vacuum assisted

vacuum assisted

Infiltration technique

leaching

leaching

burning

pressurized water

(not reported)

water spray

water spray

Pattern removal

Summary of replication techniques used for the production of metal sponge.

Pattern material

Table 2.6-1.

0.01±3

0.45±1.45

0.15±5

1.2±3.7

1±5

0.85±2.5

4.5

4.5

Pore size (mm)

65±81

55±67

(60±63) full range not reported

73±86

60±63

92±98

93±96

97

Porosity (%)

Metal sponge characteristics*

[5]

[2]

[8]

[11]

[15]

[4]

[12, 13]

[14, 15]

Reference

52



Burnable Patterns Aluminum sponge has been prepared from resin-bonded polystyrene spheres (Table 2.6-1) with porosity in the range 73±86 % and pore sizes in the range 1±4 mm [11]. Infiltration was conducted with a modified die-casting technique and the pattern was removed by burning in air. Properties of these metal sponges have not been reported. 2.6.3.3

Leachable Patterns Aluminum sponge was produced from NaCl as early as 1961 [8]. The patterns were prepared by packing salt grains in a mold and infiltrating by one of several techniques depending on the size of the salt grains: by gravity (i 3.4 mm), on a vibrating table (i 1.7 mm), or with an overpressure and simultaneous application of pressure (j 0.15 mm). The compressive strength of an A356 aluminum alloy sponge was observed to depend on the heat treatment and at 10 % strain varied in the range 20±30 MPa for porosity in the range 60±63 %. The porosity undoubtedly varied depending on the size (and distribution of sizes) of the salt grains. In an independent study, Banhart reports a range in porosity between 55 and 67 %; however, he does not provide any details about the processing other than that the pattern was leachable [2]. Some basic property indicators (mechanical response to compression, pressure drop and sound adsorption) are given in this latter report. Detailed analysis of properties is not given in either reference. Recently, gas-pressure infiltration techniques have been used to infiltrate sintered NaCl with grain sizes as small as 10 mm [5]. Porosity was in the range 68±81 % depending, in part, on the size of the salt grains. A SEM micrograph of an aluminum sponge prepared by replication of sintered NaCl is shown in Fig. 2.6-2. The mechanical response of a metal sponge depends on the microstructure of the metal (Fig. 2.6-3), as with any fully dense alloy; however, the properties of the metal sponge are further influenced by the level of porosity. The shape of the flow curves in Fig. 2.6-3 are somewhat different from what is generally observed with commercially available closed-cell metal foams. This is due to the inherently higher ductility of most of the base metals used to produce the sponges represented in Fig. 2.6-3 in comparison to commercial cellular metals, as well as the higher regularity of the architecture of these metal sponges, and the comparatively low porosity. The stiffness was found to be about half that expected from simple phenomenological relationships [5]. The processing limits for replication from leachable patterns, like the other pattern technologies, have not been clearly evaluated and there is some divergence in the literature. The range of available pore sizes and porosity, however, is clearly considerable. 2.6.3.4

53

54


Figure 2.6-2. Scanning electron micrograph of aluminum sponge produced by replication from sintered NaCl patterns.

Figure 2.6-3. Compressive response of several aluminum-alloy sponges produced by replication from sintered NaCl patterns. The salt has grains approximately 0.5 mm in diameter and the fraction solid is noted as f.


2.6.4

Conclusions

Replication processing provides a versatile and attractive method for the production of metal sponge (often called open-celled metal foam). Advantages of this process comprise considerable control of the pore size, density, and internal architecture of the sponge, the capacity for net or near-net shape component production, and the potential for production of components that are part dense, part porous. Also, the open-porosity that is intrinsic to metal sponge can be put to advantage in several applications that require fluid flow through the porous metal, such as filters and heat-exchangers. The replication process exists in a few variants, which differ mostly according to the nature of the pattern: continuous refractory, discontinuous refractory, burnable, and leachable patterns are four classes of patterns that have been employed to date. The range of density and pore size that can be achieved in replicated metal sponges depends largely on the pattern material and also on the infiltration process. With leachable patterns, in particular, these ranges have been shown to be comparatively wide; at present these also seem to be complementary to what can be achieved with the methods utilizing insoluble patterns. There is, though, significant room for innovation in replication technology as related to the production of metal sponge and there exists perhaps even opportunities for cross-pollination between the methods described here.

References

1. D. Weaire, M. A. Fortes, Adv. Phys. 1994, 43, 685. 2. J. Banhart, Adv. Eng. Mater. 2000, 2, 188. 3. J. Babjak, V. A. Ettel, V. Paserin, US Patent 4 957 543, 1990. 4. I. Wagner, C. Hintz, P. R. Sahm, in Metal Matrix Composites and Metallic Foams, Proc. Euromat 99, Munich, Germany, T. W. Clyne, F. Simancik (eds), DGM/Wiley-VCH, Weinheim 2000, p. 40. 5. C. San Marchi, J.-F. Despois, A. Mortensen, in Metal Matrix Composites and Metallic Foams, Proc. Euromat 99, Munich, Germany, T. W. Clyne, F. Simancik (eds), DGM/WileyVCH, Weinheim 2000, p. 34. 6. R. W. Pekala, R. W. Hopper, J. Mater. Sci. 1987, 22, 1840. 7. T. J. Fitzgerald, V. J. Michaud, A. Mortensen, J. Mater. Sci. 1995, 30, 1037.

8. L. Polonsky, S. Lipson, H. Markus, Modern Castings 1961, 39, 57. 9. H. Seliger, U. Deuther, Freiburger Forschungshefte 1965, 103, 129. 10. H. A. Kuchek, US Patent 3 236 706, 1966. 11. L. Ma, Z. Song, D. He, Scripta Mater. 1999, 41, 785. 12. Y. Yamada, K. Shimojima, Y. Sakaguchi, M. Mabuchi, M. Nakamura, T. Asahina, T. Mukai, H. Kanahashi, K. Higashi, Mater. Sci. Eng. 1999, A272, 455. 13. Y. Yamada, K. Shimojima, Y. Sakaguchi, M. Mabuchi, M. Nakamura, T. Asahina, T. Mukai, H. Kanahashi, K. Higashi, Mater. Sci. Eng. 2000, A280, 225. 14. Y. Yamada, K. Shimojima, Y. Sakaguchi, M. Mabuchi, M. Nakamura, T. Asahina, T. Mukai, H. Kanahashi, K. Higashi, J. Mater. Sci. Lett. 1999, 18, 1477.

55

56

2.7 Solid-State and Deposition Methods 15. Y. Yamada, K. Shimojima, Y. Sakaguchi, M. Mabuchi, M. Nakamura, T. Asahina, T. Mukai, H. Kanahashi, K. Higashi, Adv. Eng. Mater. 2000, 2, 184. 16. U. Mohr, W. Bleck, in Metal Matrix Composites and Metallic Foams, Proc. Euromat 99, Munich, Germany, T. W. Clyne, F. Simancik (eds), Wiley-VCH, Weinheim 2000, p. 28. 17. R. M. German, Particle Packing Characteristics, Metal Powder Industries Federation, Princeton, NJ 1989. 18. C. S. Morgan, L. L. Hall, C. S. Yust, J. Am. Ceram. Soc. 1963, 46, 559. 19. A. A. Ammar, D. W. Budworth, Proc. Br. Ceram. Soc. 1965, 3, 185. 20. W. J. Tomlinson, G. Astle, J. Mater. Sci. 1976, 11, 2162. 21. R. J. Thompson, Z. A. Munir, J. Am. Ceram. Soc. 1982, 65, 312. 22. W. D. Kingery, M. Berg, J. Appl. Phys. 1955, 26, 1205. 23. J. B. Moser, D. H. Whitmore, J. Appl. Phys. 1960, 31, 488.

24. P. J. Wray, Acta Metall. 1976, 24, 125. 25. R. W. Davidge, Proc. Br. Ceram. Soc. 1969, 12, 245. 26. A. Mortensen, I. Jin, Int. Mater. Rev. 1992, 37, 101. 27. C. Garcia-Cordovilla, E. Louis, J. Narciso, Acta Mater. 1999, 47, 4461. 28. A. Mortensen, in Metal Matrix Composites, A. Kelly, C. Zweben (eds), Vol. 3 of Comprehensive Composite Materials, T. W. Clyne (ed.), Pergamon, Oxford, UK 2000, Ch. 21. 29. A. Mortensen, J. A. Cornie, Metall. Trans. 1987, 18A, 1160. 30. A. Mortensen, T. Wong, Metall. Trans. 1990, 21A, 2257. 31. T. Mukai, H. Kanahashi, Y. Yamada, K. Shimojima, M. Mabuchi, T. G. Nieh, K. Higashi, Scripta Mater. 1999, 41, 365. 32. H. Kanahashi, T. Mukai, Y. Yamada, K. Shimojima, M. Mabuchi, T. G. Nieh, K. Higashi, Mater. Sci. Eng. 2000, A280, 349. 33. W. Thiele, Metal Mater. 1972, 6, 349.

2.7

Solid-State and Deposition Methods

O. Andersen and G. Stephani

In this chapter, solid-state and deposition methods for the manufacturing of cellular metals are reviewed with regard to their most important processing features. Discussion will be limited to methods producing predominantly 3D structures. Typical values of the cell size, cell wall thickness, and obtainable relative densities are cited wherever available. Earlier overviews including both methods of manufacturing are available [1,2]. The term ªsolid-state methodº means that directly before the formation of the cellular structure, the metal is in the solid state. However, in some cases limited amounts of liquid phase appear in the process: during liquid-phase sintering or brazing of single cells. Except for foaming of solids, all solid-state methods require a sintering step before the final cellular metallic structure is obtained. Naturally, powder metallurgical methods play a dominant role in this area.


Porosity in cellular structures made from powders often exhibits two different size scales: a macroporosity of the order of the targeted cell size, and a microporosity in the order of the primary particle or, in case of reactive sintering, the grain size. For all powder metallurgical processes, the cost of the powders is a major concern. Some of the methods, especially those using slurries of metal powders dispersed in a dissolved binder, are capable of using cheap oxides and thus have the potential for a significant cost reduction. Deposition methods need to mobilize the metal in order to transport it onto a porous substrate. This is normally done by transforming the metal into the ionic or vapor state. From the method of manufacture, it can be deduced that the deposition methods are only capable of generating an open-cell morphology. Usually, they also imply a heat treatment for the removal of the precursor on which the metal is deposited. A practical differentiation between the methods in question can be made with the help of two manufacturing parameters, as is shown in Fig. 2.7-1. The chosen criteria, ªformation of cellular structureº and ªformation of cellsº have direct implications for the processing itself, as well as for the resulting properties of the structure. According to this classification, the formation of the cellular structure can be accomplished by building it up from single cells, which do not necessarily have to be in the final metallic state. Alternatively, the structure can be built up from bulk. The formation of the cells themselves can either be carried out with the help of a lost core, or in a coreless manner. Within this framework the deposition methods, for example, constitute a small subset of the lost core bulk processes.

Solid state and deposition methods from single cell

from bulk

coreless methods structures from hollow powders

sintered metal powders and fibers

structures from hollow

special sintering effects

spheres made by coaxial

foaming of solids

spraying of slurries

foaming of slurries

lost core methods

Classification of solid-state and deposition methods from a manufacturing point of view. Figure 2.7-1.

cementation process

p/m space holder methods

galvanic hollow spheres

deposition methods

IFAM hollow spheres

57

58

2.7 Solid-State and Deposition Methods

Generally, formation from single cells has the potential of a very close control of the cellular properties. However, in most cases the single cell processes require more processing steps than bulk formation processes. Processes using lost cores usually are prone to introducing unwanted interstitials like carbon and oxygen into the base material, since the cores are predominantly hydrocarbons, which have to be removed thermally. Another disadvantage compared to coreless processes is the additional time required for the removal of the core. Generally speaking, there is a remarkable variety of solid-state and deposition methods for making cellular metals, which all have their specific strengths and weaknesses. It therefore depends very much on the final application, which method may be best suited in a given situation. 2.7.1

Formation from Single Cells: Coreless Methods Hollow-Sphere Structures made from Gas Atomized Hollow Powders It is known that gas atomization often generates a small amount (typically 1±5 %) of hollow particles owing to gas entrapment in the liquid ligaments. By appropriate separation methods, these hollow particles can be separated from the solid ones 2.7.1.1

a) Surface

b) Cross section

1.0 mm

1.0 mm

Figure 2.7-2. HIP-bonded hollowsphere structure made from alloy 625 (courtesy IPML, University of Virginia, Charlottesville).


and consolidated to yield hollow-sphere structures. Typical sphere diameters are 500±1000 mm with wall thicknesses about 100±300 mm [3]. Different methods for consolidation are reported: sintering, transient liquid phase sintering with the help of a powdered additive, and hot isostatic pressing (HIP). Best results with regard to specific stiffness were achieved with HIP bonding, Fig. 2.7-2. This was attributed to the predeformation of the cells and enlarging of the contact zones. Structures were made from TiAlV 64 as well as from a nickel-based superalloy (alloy 625). Relative densities of 0.3±0.12 were achieved. Other methods of producing hollow metal powders without using lost cores have been reported, such as coaxial atomization of melts [4], atomization of melts containing a dissolved gas [5], and plasma treatment of powder particles containing an absorbed gas [6]. However, nothing is known about cellular structures made thereof.

Hollow-Sphere Structures made from Coaxially Sprayed Slurries Hollow spheres can be produced by spraying a powder slurry containing a solvent and a polymer binder through the outer orifice of a coaxial nozzle [7]. By hydrodynamic interaction with the gas passing through the inner orifice, single bubbles are formed. In flight, they are spherodized by surface tension, and dried. The resulting hollow spheres can have diameters in the range 1±6 mm with close control (e4 %) of the dimensions. Wall thickness is typically around 100 mm and can be varied between 40 and 200 mm [8]. Subsequently, the powder shells are heat treated to remove the organic binder and, at a higher temperature, to sinter the metal powder particles to form a closed metallic shell. In order to obtain cellular structures, the single hollow spheres are filled in molds and either diffusion bonded or sinter bonded with the help of a metal powder slurry. The use of slurries leads to an undesirable reduction in overall void space of the structure. Nonetheless, values of the relative density of 0.12 were 2.7.1.2

Figure 2.7-3. FeCr hollow spheres bonded together with magnetite slurry and reduced (courtesy Georgia Institute of Technology).

59

60


achieved with stainless steel structures made from reduced iron and chromium oxide powders [9], Fig. 2.7-3. 2.7.2

Formation from Single Cells: Lost Core Methods Hollow-Sphere Structures made by Cementation and Sintering If spherical iron particles are immersed in a stirred CuSO4 solution, the so-called cementation reaction 2.7.2.1

Cu2 Fe () Cu Fe2 leads to deposition of pure copper on the particles while simultaneously the iron core is being dissolved. The process can be carried out until the iron is completely dissolved in the electrolyte and only hollow copper shells remain [10]. Hollow copper powders were prepared using reduced iron oxide powders as the lost core. The resulting copper particles had diameters of about 500±750 mm. The walls are very rough and porous, having a thickness of a few hundred microns. The hollow copper particles need further heat treatment to give them sufficient strength for handling. Cellular metallic structures are achieved by packing the hollow particles in ceramic molds and bonding them by sintering. Good results were achieved by coating the copper particles with tin, thus enabling liquid-phase sintering. The resulting cellular bronze structures exhibit good interparticle bonding, dense cell walls, and relative densities of about 0.2 [11], Fig. 2.7-4.

Cross section of a bronze hollow-sphere structure made by cementation and sintering (courtesy Institute for Chemical Technology of Inorganic Materials, Technical University of Vienna).

Figure 2.7-4.


Figure 2.7-5. Brazed nickel hollow-sphere structures with sphere diameter of 4.5 mm made from galvanically manufactured hollow spheres (courtesy ATECA, Montauban, France).

Hollow-Sphere Structures made from Galvanically Coated Styrofoam Spheres Hollow spheres can be made by galvanic coating of Styrofoam spheres, which act as a lost core [12]. The Styrofoam is thermally removed. Sphere diameters may range from 2 to 10 mm. The sphere walls show excellent uniformity in thickness, which is typically in the few micron range. Thick walls are very expensive to manufacture. Material choice is limited to a few metals suitable for galvanic deposition (typically copper and nickel). For making purely metallic structures, the single spheres can be joined by brazing [13] or diffusion bonding. Sphere sizes are in the range of a few millimeters and relative densities can be smaller than 0.05 since the galvanic process is able to produce very thin and uniform coatings [14]. With suitable sorting methods for the Styrofoam cores, very uniform spheres can be obtained, which allow for the manufacturing of ordered structures, Fig. 2.7-5. 2.7.2.2

Hollow-Sphere Structures made from Fluidized Bed Coated Styrofoam Spheres According to previously published work [15], commercially available Styrofoam spheres can be coated in a fluidized bed process with a suspension containing metal powder and a binder. The resulting green spheres may be debindered and sintered to obtain single metallic hollow spheres. Alternatively, the green spheres can be subjected to a suitable forming process to obtain green parts made up of single cells [16], Fig. 2.7-6. The resulting green parts 2.7.2.3

61

62

2.7 Solid-State and Deposition Methods Figure 2.7-6. Process scheme for making hollow-sphere structures (courtesy IFAM, Dept. of Powder Metallurgy and Composite Materials, Dresden).

are heat treated in the same way as the single spheres. In principle, this allows for the production of hollow-sphere structures from arbitrary metals and alloys. In order to enhance the mechanical properties of the resulting structures, large contact areas between the spheres are achieved by applying a slight pressure to the green spheres during the forming step. In this stage, the elastic Styrofoam core helps to prevent buckling and leads to a more uniform distribution of the cell pre-deformation. A typical example of the resulting structure is shown in Fig. 2.7-7. The attainable range of the sphere diameters is 0.5±10 mm. Wall thicknesses have been made from 10 to 1000 mm. Relative densities can be as low as 0.03. Recent overviews of the processing and properties of such structures are given elsewhere [17,18]. Materials currently under investigation are numerous, with a focus on stainless steel 316L [19] and low-cost hollow-sphere structures made from iron oxides [20].

Figure 2.7-7. IFAM 316L hollow-sphere structure. Left hand side: computer tomography images of cross sections (courtesy IFAM, Department of Powder Metallurgy and Composite Materials, Dresden).


2.7.3

Bulk Formation: Coreless Methods Sintered Metal Powders and Fibers Porous metals made from sintered particulates are already in widespread use for predominantly functional applications. A good overview of classical commercial processes and products relying on plain sintered metal powders is available [21]. The powders are sifted to narrow size fractions, die compacted, and sintered. Bronze, stainless steels, titanium, and nickel-based alloys are common in commercial production. Depending on the particle morphology and packing density, the relative density of the resulting parts can reach values as low as 0.4, while pore sizes are from 0.5 to 200 mm. However, usually the relative density is quite high (0.8±0.6). Much lower relative densities can easily be achieved by using particles with large length-to-diameter ratios: metal fibers. Depending on the length-to-diameter ratio and the processing route, relative densities of 0.01 can be obtained. Very fine fibers down to diameters of a few microns can be produced by the bundle drawing technique [22]. Fibers are available in many corrosion and heat resistant alloys: 316 and 302-type, Inconel 601, FeCrAlloy, Hastalloy X, Nichrome, and some titanium alloys. Diameters of bundle drawn fibers can be as small as a few microns. Using broken fiber bundles or short fibers, highly porous structures can be manufactured by applying textile or other dressing techniques with subsequent sintering to get a rigid structure. Depending on the application, the mean pore size of such structures can be made as small as a few microns. The relative density typically is in the range 0.35±0.1, but can also be much smaller. There are other mechanical technologies for making fibers, such as shaving from either wire, foils, or bars. These fibers have irregular cross sections, either triangular or rectangular with a certain surface roughness and equivalent diameters of 25±100 mm. These shaving methods have been recently improved by introduction of superimposed vibrations, which results in a more homogeneous cross section [23]. 2.7.3.1

Figure 2.7-8. Cross-sectional cut of sintered fiber structure made from melt-extracted short fibers from FeCrAl 23.15. From left to right: relative density 0.3, 0.2, 0.1 (courtesy IFAM, Department of Powder Metallurgy and Composite Materials, Dresden).

63

64


Using the melt extraction process for making short fibers directly from the melt, the materials limitations of the mechanical fiber manufacturing routes can be overcome [24,25]. Highly porous sintered structures from intermetallics such as Ni3Al [26] and FeAl, as well as from many other ferrous and nonferrous alloys have been made. Equivalent fiber diameters are 50±150 mm, fiber length may vary between 3 and 25 mm. This allows for relative densities between 0.5 and 0.03 in the final structure, Fig. 2.7-8.

Methods Utilizing Special Sintering Phenomena In common powder metallurgical production, the formation of pores is an unwanted by-product of the sintering process. Many origins of pore formation have been explained theoretically and attributed to a variety of different physical phenomena. Some of these effects are capable of generating a very large amount of porosity and thus lend themselves to making cellular metals. An early example is the so-called Kirkendall effect, in which in a diffusion couple pores are generated in the component with higher diffusivity [27]. Aldinger [28] has evaluated the potential of an extreme Kirkendall effect for making porous structures. Cellular structures were obtained by using coated particles where the core consists of the component with higher diffusivity. The powders are then pressed and sintered. Experiments with nickel and cobalt coated beryllium powders showed a maximum volume increase of 262 %, corresponding to an absolute density of 0.47 g/cm3 and a relative density in the order of 0.2. The structures exhibit a mixture of closed-cell porosity with diameters of 50 to 100 mm and wall thicknesses around 10 mm, depending on the size of the starting powders and the thickness of the coatings. Theoretical considerations lead to the conclusion that under ideal conditions, relative densities can be made arbitrarily small. In reality, this is hindered by a number of different effects, such as incomplete or non-uniform coating of the particles. In reactive sintering, elemental powders are used instead of pre-alloyed powders in order to enhance sinterability and reduce cost. Especially in intermetallic systems, an unwanted swelling of the sintered parts is noticed under certain processing conditions. A general treatment of the different mechanisms for swelling in the system Ti-Al was given elsewhere [29]. KnuÈwer exploited the swelling effect for manufacturing FeAl structures with open porosity [30]. Relative densities down to 0.5 were reported. In contrast to the cell morphology obtained via the Kirkendall effect, these structures exhibit an almost exclusively open porosity. Due to the different effects taking place during swelling, a macro- and a microporosity are generated, both being of an open type (Fig. 2.7-9). Macroscopic porosity is in the order of the primary particle sizes, typically in the range 50±150 mm. The size range of the microscopic porosity is determined by the grain size of the intermetallic phase generated, and is typically smaller than the macroscopic porosity by one or two orders of magnitude. 2.7.3.2


Porous TiAl3 structure prepared from elemental powder mixture. Large pores correspond to primary Al particles (courtesy IFAM, Department of Powder Metallurgy and Composites).

Figure 2.7-9.

Foaming of Solids Kearns et al. first described a solid-state foaming process where porosity in metals is generated by expanding a hot-isostatically pressed mixture of argon gas and metal powder at elevated temperatures and ambient pressure [31]. More recent work utilizes this principle for making TiAlV 64 sandwiches with a porous core [32]. This is achieved by canning TiAlV 64 powder in canisters of the same material. After HIP, the resulting billet can be isothermally forged or hot rolled to achieve sheets. Subsequently, an expansion annealing treatment (typically at temperatures i0.6Tm for 4±48 h) generates the porous core, Fig. 2.7-10. This method, which is known as the low-density core (LDC) process, uses creep expansion due to the internal gas pore pressure to generate the desired porosity. Queheillalt et al. have carried out a thorough analysis of the process kinetics and morphological evolution of the pores [33]. The resulting porosity in the core region is largely unconnected and shows pore diameters of 10±100 mm. The relative density lies in the range 0.65±0.5. One advantage is the capability of producing very large sheets (up to 2000 mm q 1200 mm q 4 mm). A slightly different approach described previously utilizes the transformation superplasticity found in allotropic materials [34]. Transformation superplasticity occurs at all grain sizes by biasing with a deviatoric stress (from the entrapped gas) of internal stresses (from the allotropic mismatch during thermal cycling about the allotropic temperature range). Hence, the rather lengthy expansion annealing can be replaced by a significantly shorter thermal cycling treatment. 2.7.3.3

65

66


Low Density Core Process Steps a) Powder / Can Preparation

Ti-6Al-4V can

Powder packing density D0

Ti-6Al-4V powder

Evacuate and backfill with Argon gas to pressure p 0

b) HIP Consolidation (1040oC, 100 - 200MPa, 2h) Isolated pressurized voids

Final relative density 0.85 - 0.95

p(t), T(t)

c) Hot Rolling (approx. 850oC, 6-20 passes) Thinning of facesheet Internal gas pressure

Changes in pore shape, matrix microstructure

d) Expansion Heat Treatment (900oC - 1200˚C, 4 - 48 h) Vacuum furnace Sandwich panel

Isolated porosity (

Handbook of Cellular Metals: Production, Processing, Applications

Cellular Effects of Heavy Metals

Handbook of Organic Food Processing and Production

Processing of Metals and Alloys

Cellular and Molecular Biology of Metals

Cellular and Molecular Biology of Metals

Fundamentals of Aluminium Metallurgy: Production, Processing and Applications

Base Metals Handbook

Metals Handbook: Corrosion

Smithells Light Metals Handbook

Smithells Light Metals Handbook

Smithells Light Metals Handbook

Handbook of Meat Processing

Handbook of petroleum processing

Nanostructured metals and alloys - Processing, microstructure, mechanical properties and applications

Handbook of Meat Processing

Handbook of Petroleum Processing

Polypropylene Handbook: Polymerization, Characterization, Properties, Processing, Applications

Tomato production, processing and technology

Potato production, processing, and technology

Opera Production: A Handbook

Radiation Physics of Metals and Its Applications

Radiation Physics of Metals and Its Applications

WELL PRODUCTION PRACTICAL HANDBOOK

Handbook on the Toxicology of Metals

Handbook on the toxicology of metals

Well Production Practical Handbook

Applications of Digital Signal Processing

Handbook of Signal Processing Systems

Handbook of extractive metallurgy: Primary metals; secondary metals; light metals, Volume 2

Handbook of Cellular Metals: Production, Processing, Applications

Cellular Effects of Heavy Metals

Handbook of Organic Food Processing and Production

Processing of Metals and Alloys

Cellular and Molecular Biology of Metals

Cellular and Molecular Biology of Metals

Fundamentals of Aluminium Metallurgy: Production, Processing and Applications

Base Metals Handbook

Metals Handbook: Corrosion

Smithells Light Metals Handbook

Smithells Light Metals Handbook

Smithells Light Metals Handbook

Handbook of Meat Processing

Handbook of petroleum processing

Nanostructured metals and alloys - Processing, microstructure, mechanical properties and applications

Handbook of Meat Processing

Handbook of Petroleum Processing

Polypropylene Handbook: Polymerization, Characterization, Properties, Processing, Applications

Tomato production, processing and technology

Potato production, processing, and technology

Opera Production: A Handbook

Radiation Physics of Metals and Its Applications

Radiation Physics of Metals and Its Applications

WELL PRODUCTION PRACTICAL HANDBOOK

Handbook on the Toxicology of Metals

Handbook on the toxicology of metals

Well Production Practical Handbook

Applications of Digital Signal Processing

Handbook of Signal Processing Systems

Handbook of extractive metallurgy: Primary metals; secondary metals; light metals, Volume 2

Recommend Documents