Metal Matrix Composites: Materials, Manufacturing and Engineering 9783110315448, 9783110315417

Citation preview

J. P. Davim (Ed.) Metal Matrix Composites

Also of Interest Nanocomposites Materials, Manufacturing and Engineering Davim, Charitidis (Eds.), 2013 ISBN 978-3-11-026644-3, e-ISBN 978-3-11-026742-6

Biomedical Composites Materials, Manufacturing and Engineering Davim (Ed.), 2013 ISBN 978-3-11-026669-6, e-ISBN 978-3-11-026748-8

Nanocarbon-Inorganic Hybrids Next Generation Composites for Sustainable Energy Applications Eder, Schlögl (Eds.), 2014 ISBN 978-3-11-026971-0, e-ISBN 978-3-11-026986-4

Functional Materials For Energy, Sustainable Development and Biomedical Sciences Leclerc, Gauvin (Eds.), 2014 ISBN 978-3-11-030781-8, e-ISBN 978-3-11-030782-5

Nanotechnology Reviews Kumar, Challa (Editor-in-Chief) ISSN 2191-9089, e-ISSN 2191-9097

Science and Engineering of Composite Materials Hoa, Suong V. (Editor-in-Chief) ISSN 0792-1233, e-ISSN 2191-0359

www.degruyter.com

Metal Matrix Composites | Materials, Manufacturing and Engineering Edited by J. Paulo Davim

Editors Professor Dr. J. Paulo Davim University of Aveiro Dept. of Mechanical Engineering Campus Santiago 3810-193 Aveiro Portugal [email protected]

ISBN 978-3-11-031541-7 e-ISBN (PDF) 978-3-11-031544-8 e-ISBN (EPUB) 978-3-11-038201-3 ISSN 2192-8983 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. © 2014 Walter de Gruyter GmbH, Berlin/Munich/Boston The publisher, together with the authors and editors, has taken great pains to ensure that all information presented in this work (programs, applications, amounts, dosages, etc.) reflects the standard of knowledge at the time of publication. Despite careful manuscript preparation and proof correction, errors can nevertheless occur. Authors, editors and publisher disclaim all responsibility and for any errors or omissions or liability for the results obtained from use of the information, or parts thereof, contained in this work. The citation of registered names, trade names, trademarks, etc. in this work does not imply, even in the absence of a specific statement, that such names are exempt from laws and regulations protecting trademarks etc. and therefore free for general use. Typesetting: PTP-Berlin, Protago TEX-Produktion GmbH, www.ptp-berlin.de Printing and binding: CPI books GmbH, Leck Cover image: gettyimages/thinkstockphotos, Abalone Shell ♾ Printed on acid-free paper Printed in Germany www.degruyter.com

Preface Nowadays, metal matrix composites (MMCs) represent a great development in modern industry applications due to their excellent properties. MMCs have a high specific strength and stiffness at elevated temperatures, lower thermal expansion, high thermal conductivity and good creep, fatigue and wear resistance. Therefore, MMCs are particularly attractive as a replacement for conventional materials in many engineering applications. MMCs are metallic alloys (aluminum, titanium, copper, magnesium, etc.) with reinforcement materials (aluminum oxide, silicon carbide, boron carbide, graphite, etc.) in forms of fibers, whiskers and particles. MMCs are widely used in various fields of application such as the automotive, aerospace, aircraft, sports as well as other advanced industries. The present volume aims to provide recent information on MMCs – materials, manufacturing and engineering – in eight chapters. Chapter 1 provides information on MMCs for thermal management. Chapter 2 is dedicated to recent research and developments on the mechanical behavior of CNT-reinforced MMCs. Chapter 3 describes new preparation and mechanical properties of in situ synthesized (TiB+La2 O3 )/TiNbTaZr composites. Chapter 4 contains information on the microstructure formation of particle-reinforced MMC coatings produced by thermal spraying. Chapter 5 investigates MMC fabrication by the liquid stirring technique. Chapter 6 describes material removal processes for MMCs. Chapter 7 contains information on machining Al/SiC MMCs. Finally, chapter 8 is dedicated to the application of the response surface method and desirability function for the optimization of the machining parameters of MMCs (Al/SiC/Al2 O3 ). The present volume can be used as a research book for a final undergraduate engineering course or as a topic on composites at the postgraduate level. Also, this book can serve as a useful reference for academics and researchers, engineers for materials, mechanics and manufacturing as well as professionals in composite materials technology and related industries. The scientific interest of this book is evident for many important research centers, laboratories and universities as well as for industry as a whole. Therefore, it is hoped this book will inspire and enthuse others to undertake research in this field of metal matrix composites. The editor acknowledges De Gruyter for this opportunity and for their enthusiastic and professional support. Finally, I would like to thank all the chapters’ authors for being available to complete this work. Aveiro, Portugal, May 2014

J. Paulo Davim

Contents Preface | v List of contributing authors | xi José Miguel Molina Jordá 1 Metal matrix composites for thermal management | 1 1.1 Introduction | 1 1.2 Composite materials for thermal management | 3 1.2.1 Liquid infiltration | 4 1.2.2 Powder metallurgy | 5 1.3 Design and modeling of metal matrix composites for electronics | 6 1.3.1 Volume fraction of ceramic phase | 6 1.3.2 Thermal conductivity | 8 1.3.3 Coefficient of thermal expansion | 9 1.4 Families of advanced metal matrix composite materials for electronics | 9 1.4.1 SiC-based composites | 11 1.4.2 Carbon-based composites | 18 1.4.3 Diamond-based composites | 26 1.5 The future of metal matrix composites in electronics | 30 References | 31 Nuno Silvestre 2 Recent research and developments on the mechanical behavior of CNT-reinforced metal matrix composites | 39 2.1 Introduction | 39 2.2 CNT-Al composites | 43 2.3 CNT-Co composites | 55 2.4 CNT-Cu composites | 57 2.5 CNT-Fe composites | 61 2.6 CNT-Mg composites | 63 2.7 CNT-Ni composites | 67 2.8 CNT-Ti composites | 71 2.9 Concluding remarks | 74 References | 75

viii | Contents Yue Li, Xiaoxing Cheng, Liqiang Wang, Weijie Lu, Jining Qin, Fan Zhang, Di Zhang 3 Novel preparation and mechanical properties of in situ synthesized (TiB+La2 O3 )/TiNbTaZr composites | 81 3.1 Introduction | 81 3.1.1 The application of rare earth elements in β titanium alloys | 81 3.1.2 The influence of rare earth elements in titanium alloys | 81 3.1.3 Biosafety of rare earth elements | 82 3.2 Materials preparation and experimental procedures | 84 3.2.1 Materials preparation | 84 3.2.2 Experimental procedures | 84 3.3 Results and discussions | 85 3.3.1 Phase analysis | 85 3.3.2 Thermodynamic analysis | 85 3.3.3 Microstructure analysis | 87 3.3.4 Microstructure of reinforcements | 89 3.3.5 Analysis of the solidification mechanism | 91 3.3.6 Superelasticity | 92 3.3.7 In situ characterization of microstructure | 95 3.3.8 Mechanical properties | 97 3.4 Conclusions | 99 References | 100 Dina V. Dudina, Igor S. Batraev, Vladimir Yu. Ulianitsky 4 Microstructure formation of particle-reinforced metal matrix composite coatings produced by thermal spraying | 103 4.1 Particle-reinforced MMC coatings formed ex situ by thermal spraying of powder mixtures and composite particles | 104 4.2 MMC coatings with reinforcing particles formed in situ during thermal spraying | 108 4.3 Design of particle-reinforced MMC coatings using flexible variation of spraying parameters in computer-controlled detonation spraying | 110 4.4 Post-spray treatment of MMC coatings | 116 References | 120 Alakesh Manna 5 Fabrication of Al-metal matrix composites by liquid stirring technique | 123 5.1 Introduction | 123 5.2 Fabrication of Aluminium metal matrix composites | 125 5.2.1 Fabrication of the stirring arrangement | 126 5.2.2 Mold-making and preparation of the mold cavity | 127

Contents

| ix

5.2.3

Estimation of raw materials for Al/5, 10, 15 wt.% reinforced MMC casting | 128 5.2.4 Experimental procedure | 129 5.3 Physical, chemical and mechanical properties of stir cast samples | 129 5.3.1 Physical property of stir cast samples | 129 5.3.2 Mechanical properties of stir cast samples | 130 5.3.3 Analysis of the reinforced weight fraction | 131 5.3.4 Microstructural characterization | 131 5.4 Optimization of stir casting parameters for Al/15 wt.% SiC-MMC | 135 5.4.1 S/N Ratio for micro-hardness of prepared Al/15 wt.% SiC-MMC | 135 5.4.2 ANOVA for micro hardness of prepared Al/15 wt.% SiC-MMC | 136 5.4.3 Mathematical model for micro hardness of prepared Al/15 wt.% SiC-MMC | 137 5.4.4 S/N Ratio for tensile strength of prepared Al/15 wt.% SiC-MMC | 137 5.4.5 ANOVA for tensile strength of prepared Al/15 wt.% SiC-MMC | 138 5.4.6 Mathematical model for tensile strength of prepared Al/15 wt.% SiC-MMC | 138 5.5 Conclusion | 139 References | 139 Inderdeep Singh, Saurabh Chaitanya, Ravinder Kumar 6 Material removal processes for metal matrix composites | 141 6.1 Introduction | 141 6.2 Conventional machining processes | 142 6.2.1 Turning of PMMCs | 143 6.2.2 Milling of PMMCs | 144 6.2.3 Drilling of PMMCs | 145 6.3 Unconventional machining of MMCs | 146 6.3.1 Electrochemical machining of PMMCs | 148 6.3.2 Electric discharge machining of PMMCs | 150 6.3.3 Ultrasonic machining of PMMCs | 153 6.4 Conclusion | 155 References | 155 Vijayan Krishnaraj 7 An investigation into machining Al/SiC metal matrix composites | 159 7.1 Milling of metal matrix composites | 159 7.1.1 Introduction | 159 7.1.2 Experimental procedure | 161 7.1.3 Results and discussion | 163 7.2 Summary | 169

x | Contents 7.3 Drilling of metal matrix composites | 170 7.3.1 Introduction | 170 7.3.2 Experimental setup and procedure | 171 7.3.3 Results and discussion | 172 7.3.4 Summary | 176 References | 176 Kayaroganam Palanikumar 8 Application of response surface method and desirability function for the optimization of machining parameters of hybrid metal matrix (Al/SiC/Al2 O3 ) composites | 179 8.1 Introduction | 179 8.2 Materials and methods | 180 8.2.1 Fabrication of hybrid metal matrix composites | 180 8.2.2 Machining experiment | 182 8.3 Modeling and optimization | 184 8.3.1 Modeling of machining parameters using the response surface method | 185 8.3.2 Optimization of machining parameters using the desirability function approach (DFA) | 185 8.4 Results and discussion | 187 8.5 Conclusions | 197 References | 198 Index | 201

List of contributing authors Igor S. Batraev Lavrentiev Institute of Hydrodynamics SB RAS Lavrentiev Ave. 15 Novosibirsk 630090 Russia

Yue Li State Key Laboratory of Metal Matrix Composites Shanghai Jiao Tong University Shanghai 200240 PR China

Saurabh Chaitanya Department of Mechanical and Industrial Engineering Indian Institute of Technology Roorkee – 247667 India [email protected]

Weijie Lu State Key Laboratory of Metal Matrix Composites Shanghai Jiao Tong University Shanghai 200240 PR China

Xiaoxing Cheng State Key Laboratory of Metal Matrix Composites Shanghai Jiao Tong University Shanghai 200240 PR China Dina V. Dudina Institute of Solid State Chemistry and Mechanochemistry SB RAS Kutateladze str. 18 Novosibirsk 630128 Russia [email protected] Vijayan Krishnaraj Department of Production Engineering PSG College of Technology Coimbatore India – 641004 [email protected] Ravinder Kumar Department of Mechanical and Industrial Engineering Indian Institute of Technology Roorkee – 247667 India [email protected]

Alakesh Manna Department. of Mechanical Engineering PEC University of Technology Chandigarh India [email protected] José Miguel Molina Jordá Instituto Universitario de Materiales de Alicante Departamento de Química Inorgánica Universidad de Alicante Apdo 99 E-03690 San Vicente del Raspeig, Alicante Spain [email protected] K. Palanikumar Department of Mechanical Engineering Sri Sai Ram Institute of Technology, Chennai – 600 044 India [email protected] [email protected] Jining Qin State Key Laboratory of Metal Matrix Composites Shanghai Jiao Tong University Shanghai 200240 PR China

xii | List of contributing authors Nuno Silvestre Department of Mechanical Engineering IDMEC Instituto Superior Técnico University of Lisbon Av. Rovisco Pais 1049-001 Lisboa Portugal [email protected] Inderdeep Singh Department of Mechanical and Industrial Engineering Indian Institute of Technology Roorkee – 247667 India [email protected] Vladimir Yu. Ulianitsky Lavrentiev Institute of Hydrodynamics SB RAS Lavrentiev Ave. 15 Novosibirsk 630090 Russia

Liqiang Wang State Key Laboratory of Metal Matrix Composites Shanghai Jiao Tong University Shanghai 200240 PR China [email protected] Di Zhang State Key Laboratory of Metal Matrix Composites Shanghai Jiao Tong University Shanghai 200240 PR China Fan Zhang State Key Laboratory of Metal Matrix Composites Shanghai Jiao Tong University Shanghai 200240 PR China

José Miguel Molina Jordá

1 Metal matrix composites for thermal management 1.1 Introduction In recent years some of the high-end technological applications related to the electronics industry have seen their advances clearly limited by the inherent inability to find new materials capable of matching the increasingly stringent requirements of this important branch of human development [1, 2]. Among the technological barriers that hinder the progress in electronics, we find the inability to remove the excessive heat produced during the operation of certain equipments. The amount of heat generated has been steadily increasing during the last decades due to the miniaturization of components and the consumption of increasing electrical power in electronic circuits. This heat, if it is not conveniently removed, can cause catastrophic failures by overheating or deformation of the structures due to a mismatch in the expansion coefficients of its elements [3–6]. Finding materials capable of accomplishing these functions in sometimes extreme conditions (fast heating/cooling cycles, high humidity, etc.) is one of today’s major challenges in the electronics industry. A material that is suitable for these purposes must combine two fundamental properties: it must have a high thermal conductivity (CT) (as high as possible – the required value has increased over time according to increasing needs) and a matched value of thermal expansion coefficient (CTE) (similar to that of semiconductors used in the manufacture of electronic circuits). Moreover, for each particular application, other extra properties can be mandatory: high chemical inertia, corrosion resistance, wear resistance, etc. Nowadays, some of the most important achievements in materials science and technology come from developments of new materials capable of meeting the appropriate requirements to act as heat sinks for the electronics industry. Traditional heat sinks for electronics were made out of monolithic materials such as Al or Cu. However, these materials do not possess the right combination of properties for the current demands in electronics. The stringent requirements of high thermal conductivity and suitable CTE force the use of composite materials [7, 8]. Despite the enormous progress of composite materials in recent decades, there is still a need to develop new and better materials for these specific applications. Composite materials that have mostly been used as heat sinks in electronics are those derived from the Al/SiC family. These materials have shown, so far, a great combination of properties and represent the state of the art in heat sinking. However, the increasing property requirements make these materials disadvantageous for the near future and force the need to seek alternative materials. Among the candidates, there are some really promising materials under current development. In general, these are materials that combine a ceramic particulate reinforcement phase embedded in a metallic matrix [9]. Finely divided ceramic

2 | 1 Metal matrix composites for thermal management reinforcements have the advantages of easy handling and shaping in comparison with those formed by continuous fibers or laminar materials. The choice of a metal as the matrix seems to be beyond discussion because the temperatures of use of these heat sinks are high enough to cause the progressive degradation of the majority of polymers but, at the same time, sufficiently low in order not to choose ceramic matrices, which often leads to high manufacturing costs. Table 1.1 gathers the most important properties required for different applications in the electronics industry for the future. Table 1.1. Property requirements of the heat sinks needed for different applications in the electronics industry [10]. Applications

Opto-electronics

Power electronics

Microelectronics I

Properties/ parts

laser diodes

IGBT modules

radio frequency (RC) SiC chips & flip chips (FP)

Electronic material

GaAs and Si

AlN and Al2 O3

Si and GaAs

SiC (AlN and Al2 O3 )

λ [W/mK] (target)

≥ 400

≥ 200

≥ 600

≥ 400

α [ppm/K]

2–4 (Si) 5–7 (GaAs)

7–8

2–4 (Si) 5–7 (GaAs)

2.5 ⋅ 105 (RF)

—

E [GPa]

—

∼200

> 310

> 310

UTS/σB [MPa]

—

∼300 (σB )

—

—

Isotropy

Desirable

Desirable

Desirable

Mandatory

−40–125 (250)

−55–150

−40–125

−40–400

N−40–150 1000 N70–150 100,000

N−40–125 1000 (FC) N25–125 1000 (RF)

N−40–300 1000

∘

ΔTuse [ C]

N−40–125 1000 Cycling conditions (N = minimum num- N25–250 1000 ber of cycles for ΔT cycles)

Microelectronics II

Tsoldering [° C]

330 ∘ C/15 min

260 ∘ C/30 min

850 ∘ C/30 min

850 ∘ C/30 min

Solder system

Au/Sn

DCB

—

—

air 85 % humidity

air 85 % humidity

air 85 % humidity

air 85 % humidity

Dimensions [mm ]

1 × 1 × 0.2–0.4 (10 × 15 × 0.2–0.5)

200 × 200 × 5

75 × 75 × 3–5 (FC) 15 × 25 × 1 (RF)

75 × 75 × 1–1.5

Roughness Ra

< 0.1 μm

—

< 0.6 μm

< 1 μm

Surf. finish

Ni/Ti/Pt (front) idem + AuSn (back)

Ni-plating

homogeneous. metallic surface Ni/Au coating

homogeneous. metallic surface

Others

Cutting possible

flat. 10 μm/10 mm

flat. 10 μm/25 mm

flat. 10 μm/25 mm

Reference mat.

Cu/ diamond

Al/ SiC and Cu/ Mo

Mo/ Cu, W/ Cu

Cu

Environment 3

1.2 Composite materials for thermal management

|

3

In this chapter we review the most promising families of composite materials that are being used or are considered suitable for use in the field of heat dissipation in the electronics industry.

1.2 Composite materials for thermal management The two most important properties that are sought in heat dissipation are thermal conductivity and the thermal expansion coefficient. A material is considered or discarded for heat sinking depending, mainly, on the values of these two properties. The thermal conductivity should be the highest possible, in order to increase the capacity of heat transfer from the electronic device to the surrounding environment. The thermal expansion coefficient must lie in between very specific values (see Table 1.1). If it is too high or too low, the different elements of the electronic packaging will acquire different dimensions due to thermal expansion, and breakage caused by thermo-mechanical fatigue can occur. Ashby developed maps based on these criteria to help in the design of new materials for specific applications [11]. The maps applicable here show the property of thermal conductivity of various monolithic materials plotted as a function of their thermal expansion coefficients (see Figure 1.1).

Diamond

Thermal conductivity (W/mK)

1000

Metals Ceramics

800

600

BeO

Cu

400 SiC

Ag

Al

200

0 0

5 10 15 20 25 30 Coefficient of thermal expansion (ppm/K)

35

Fig. 1.1. Map of the thermal conductivity and coefficient of thermal expansion for different metallic and ceramic materials.

4 | 1 Metal matrix composites for thermal management In the previous map, the properties of large groups of metallic and ceramic materials are distinguished. Regarding metals, those having high values of thermal conductivity include Al, Ag and Cu (Au has an interesting thermal conductivity as well, but it is discarded due to its price). However, their thermal expansion coefficients are too high. That is why traditional heat sinks made out of monolithic metals are no longer used. On the other side, ceramics generally have lower thermal expansion coefficient values (in most cases too low) and many of them do also have high thermal conductivity. Among those having high thermal conductivity, we find diamond and SiC. BeO, despite its interesting properties, is excluded as a candidate because of its high toxicity. These ceramics could be used on their own for the production of heat sinks, but their clear drawback is the difficulty of machining them. Therefore it seems obvious that a metal-ceramic combination seems to be an appropriate proposal to solve the problem of heat dissipation in electronics. The most interesting combinations arise from the mixture of these two ceramics, SiC and/or diamond, with metals such as Al, Ag or Cu. Properties expected for these combinations fall somewhere halfway between the properties of the metal and the ceramic. Therefore, with the aim to reduce the CTE and at the same time increase the CT as much as possible, combinations with a high volume fraction of these highly thermal conducting ceramic phases must be considered. The increased presence of ceramic in the composite material has some consequences of high technological relevance. On the one hand, it is necessary to make use of techniques able to achieve proper packaging of finely divided ceramics (this has often been solved with the use of multimodal mixtures of particulate reinforcements). On the other hand, it limits the manufacturing techniques for composite materials, excluding for example compocasting, or spray deposition, which are only suitable for relatively small reinforcement contents. The most used techniques for the manufacture of composite materials with high volume fraction of ceramic reinforcement are discussed below.

1.2.1 Liquid infiltration Liquid infiltration is clearly the most versatile technique in terms of complex shapes fabrication. It involves two main steps: (i) preparation of the ceramic preform, normally by packing or pressing a finely divided ceramic; and (ii) liquid infiltration of the molten metal into the preform. Final machining is sometimes needed, although if the molds are properly designed, near net-shape fabrication is achievable. Its main drawback stems from the limited rate of penetration of the porous ceramic preforms by the molten metals, often caused by the poor wettability of the metal-ceramic systems [12–19]. This makes difficult the manufacture of large pieces. Depending on the range of values of applied pressures, liquid infiltration is divided into:

1.2 Composite materials for thermal management

|

5

• Gas pressure infiltration Pressures, assisted by gas, do not normally exceed 15 MPa, which is sufficient to ensure a complete filling of the pore structure of the ceramic architecture and at the same time avoid significant distortion of the preform. This technique is considered to be the most convenient for the fabrication of composite materials for electronics. There is, however, a clear disadvantage: the use of molten metals (thus high temperatures) accelerates the kinetics of reaction between metals and ceramics, and phases of low thermal conductivity are very often generated. Infiltration kinetics and preform chemical protection are issues to consider when using this technique. • Mechanically assisted infiltration The metal is forced to penetrate into the ceramic preform at very high pressures (in the range 50–100 MPa) by means of a mechanically driven piston. This method is called “squeeze casting” and is applied extensively in industry for the fabrication of metal matrix composites for different purposes. In many cases, special care must be taken because the high pressure may distort the ceramic preform. The solution to this problem seems to be not as easy as diminishing the working pressures, since this would cause a decrease in the infiltration rate; possible negative consequences could derive from the reactivity between metal and reinforcement, which can alter the final properties of the material or even hinder infiltration occasionally. Its main advantage is that the high pressures effectively ensure that the final composite materials are free of remaining porosity.

1.2.2 Powder metallurgy Although powder metallurgy is becoming important because of the recent advances in the manufacturing of relatively complicate shapes, it is not an extensively used fabrication method in the field of composite materials. This technique allows one to obtain high volume fractions of reinforcement (up to 75%) and offers a perfect monitoring of the whole process. Moreover, homogeneous distribution of reinforcement and reactivity control (or suppression) between metal and ceramic phases can be achieved by proper coating of the metallic powders. However, a clear drawback is the difficulty encountered regarding the control of porosity in the materials, which is a phase that inherently appears when using this technique. The relatively low control of the oxygen content (metallic oxides do exist concomitantly on metallic powders) does also seem to limit the use of this technique massively in industry for the fabrication of composite materials for electronics. Recently, another technique derived from powder metallurgy has become a matter of interest. This technique is called spark-plasma and it consists of heating a metallic mold containing the powder mixture of metal and ceramic by means of an electrical current that flows through it. This technique allows for fast processing of the materials

6 | 1 Metal matrix composites for thermal management but, nevertheless, it is not free of the disadvantages of the classical powder metallurgy route.

1.3 Design and modeling of metal matrix composites for electronics 1.3.1 Volume fraction of ceramic phase Based on Figure 1.1, it might be considered that the minimum volume fraction of reinforcement needed for composites with applicability in electronics is about 0.6. The logical reasoning to achieve high volume fractions of particles is to compact them up to their maximum density. There are three factors directly affecting the maximum attainable volume fraction in granular solid packing: size, size distribution and shape. Modeling reinforcement volume fractions proves to be an indispensable tool for predicting the final properties of materials. A considerable number of models have been developed to this aim, from the simplest by Thomson considering perfect spherical particles [20], to the most sophisticated models that take into account three parameters at the same time [21–24]. Guyon [25] studied the particle packing of identical spheres of equal size and came to the conclusion that the maximum theoretical value is 0.74. Although this sounds a priori like a good idea to attain the limit of 0.6 for electronics, it is unfeasible in practical terms, given that such regularity in packing and particle shape is almost impossible to obtain at the experimental level. The experimental values obtained when compacting particles of a single size with random dispositions are in the range 59 %–64 %, depending on the packing conditions and regularity of particles [26, 27]. The greater the regularity of particles, the greater is the volume fraction in the packed compacts [28]. Moreover, it is necessary to consider that the particles do not have a single representative size but rather a size distribution whose particular characteristics considerably affect compaction. Burminster [29] was able to carry out a detailed study of the compaction based on the distribution of sizes. His results were later gathered up by Peronius and Sweeting [30], who published their interesting findings as follows: the materials with a very small span (very close size distributions) produce the minimum compactions, whereas those with major span exhibit the maximum compactions; this effect was more marked for those size distributions with a long tail oriented towards the end of small particles [27]. Seemingly, the small particles adopt positions in the compact in between large particles, forming a complex structure where the empty spaces between particles of similar size are now filled by particles of increasingly smaller diameter. The findings of Peronius and Sweeting brought the idea that an alternative and very attractive way to increase the volume fraction of compacts consists of mixing and packing particles of largely different sizes (size ratio > 10) [31, 32]. During the past few years, diverse groups have been working on the ceramic particle packing of binary sys-

1.3 Design and modeling of metal matrix composites for electronics

|

7

tems (bimodal mixtures – two different average particle sizes), ternary systems (three different sizes) [33, 34] and quaternary systems [35]. The Yu and Standish group has also paid special attention to the modeling of these multimodal systems [36–38]. A very simple and intuitive model for the prediction of particle packing in bimodal particle mixtures was presented and validated in [39], and the main features can be summarized as follows. When a bimodal mixture is considered (two different particle sizes), the following model with two limits mathematically separated can be assumed (for details please see [39]): (a) fine particle end: V= Xcp +

1 ; 1 − Xcp

(1.1)

Vsp

(b) coarse particle end: V=

Vcp Xcp

,

(1.2)

where V is volume fraction and Xi is the fraction of an inclusion type in the bimodal particle mixture. When present, the subscripts l and s refer to large and small particles, respectively. Vl and Vs refer to the volume fraction when solely large or small particles are packed, respectively. Other mathematical schemes, like that proposed by Yu and Standish [37], make use of experimental information through a parameter with an adjustable value. This predictive scheme has been proved to accurately fit different sets of experimental data: (

v − Xl − vs Xs v − X l − vs Xs 2 v − v l Xl 2 v − vl Xl ) + 2G ( )( )+( ) = 1. vs vs vl − 1 vl − 1

(1.3)

Here v is the apparent volume occupied by unit solid volume of particles (i.e. the reciprocal of the volume fraction V), Xi is the fraction of an inclusion type and the subscripts l and s, when present, refer to large and small particles, respectively. The G parameter is related to the particle size ratio; some studies have accounted for the value of G with the following empirical formula [40]: G = 1 − (1 − vl )0.63 + (

−0.63 vl 1 −1.89 ) ( ) , vs − vl vs R

(1.4)

R being the coarse-to-small particle radius ratio. In terms of practice, it might be mentioned that mixing and packing particles of different sizes is a non-trivial exercise. Several problems may appear during compaction of particles, mainly related to segregation phenomena (they end up with nonhomogeneous distributions). In this sense, the packing procedure must be validated and optimized. While it was demonstrated for samples with a single particle size that excellent packing is obtained by alternating strokes of weight and vibrations, for bimodal particle distributions the packing procedure requires two stages: (i) mixing the

8 | 1 Metal matrix composites for thermal management particles; and (ii) packing the mixtures. Strokes and vibrations on dry mixtures have been identified as the main source of segregation. Other procedures have been developed for the packing stage in order to reduce segregation and attain the maximum compactness possible. One of these methods consists of packing the particle mixture in wet conditions (ethanol or other organic solvents might be used).

1.3.2 Thermal conductivity The thermal conductivity of composite materials can be properly predicted with the Generalized Differential Effective Medium Scheme (GDEMS). This scheme has been successfully applied to model and interpret transport properties in different composite materials [41–46]. The leading integral equation of the GDEMS approach for the thermal conductivity of a multi-phase composite material is K

dK

∫ K−m

K ∑ fi i

−(K − Kreff ) (K − Kreff )

= − ln(1 − Vp ),

(1.5)

P−K

where K is the thermal conductivity of the composite material, fi is the fraction of the i inclusion in the total amount of inclusions of the composite (hence ∑i fi = 1) and P is the polarization factor of the inclusion (equal to 0.33 for spheres, as modeled in this study). The subscripts c, m and r refer to composite, matrix and reinforcement, respectively. Kreff is the effective thermal conductivity of an inclusion, which is related to its intrinsic thermal conductivity Krin , the matrix/inclusion interface thermal conductance h and the radius of the inclusion r, by Kreff =

Krin 1+

(1.6)

Krin hr

In general, the integral on the left hand side of equation (5) has no analytical solution and has to be solved numerically by means of appropriate mathematical software. Another predictive scheme, widely used because of its simplicity, is that proposed by Hasselman–Johnson, derived for spherical particles in an infinite pore-free matrix: Kc =

Km [2Km + Kreff + 2(Kreff − Km )Vr ] 2Km + Kreff − (Kreff − Km )Vr

.

(1.7)

This model, however, is restricted to those systems where phase contrast (Kreff /Km ) is roughly below 4, as indicated in [41].

1.4 Families of advanced metal matrix composite materials for electronics

| 9

1.3.3 Coefficient of thermal expansion The simplest model for the calculation of the CTE of a composite material is the Linear Rule of Mixtures (LROM); it assumes that the interface between matrix and reinforcement is ideally elastic. This model averages over the volume fraction of the present phases: α = αr Vr + αm (1 − Vr ), (1.8) where the subscripts m and r denote matrix and reinforcement, respectively. This simple approach offers a back-of-the-envelope calculation of what the coefficient of thermal expansion must approximately be. More sophisticated treatments are all based on the thermoelasticity theory. Schapery’s model [47, 48] gives upper (+) and lower (−) bounds on the CTE. The specific expression for the former is: α (+) = αr + (αm − αr )

Km (Kr − Kc(−) Kc(−) (Kr − Km )

,

(1.9)

where K(−) is Hashin and Shtrickman’s [49] lower bound to the bulk modulus of the composite, namely, Vr Kc(−) = Km + , (1.10) Vm 1 + Kr − Km Km + 4 μm 3

where μm is the shear modulus of the matrix. The upper bound to the bulk modulus is obtained by interchanging the subscripts m and r everywhere in equation (1.10) and, when inserted in equation (1.9), gives the lower bound on the CTE. As noted by Schapery, the upper bound coincides with the expression derived by Kerner [48].

1.4 Families of advanced metal matrix composite materials for electronics Many families of materials have been proposed for their use as heat sinks in electronics. The market entry of a material for a specific application, far from being simple, is a process that is linked to developing optimized materials and finding new ones that should meet the needs in the near future. Thus, materials for electronics have been changing over time depending on the required properties, so that three generations can be distinguished according to their thermal conductivity values [1]. The first generation, also known as “traditional heat sinks”, is defined by materials with thermal conductivities up to 300 W/mK and consists of, mainly, metallic materials. These materials reach high thermal conductivities but hardly meet the requirement of a moderately low CTE. The second and third generations of heat sinks already incorporate metal matrix composites. The materials of the second generation achieve

10 | 1 Metal matrix composites for thermal management

Third-generation

Second-generation

First-generation

Table 1.2. Properties of metal matrix composites used as heat sinks of the second and third generation. The table includes some metallic materials of the first generation for comparison [1]. Reinforcement

Matrix

Thermal conductivity TC (W/mK)

Coefficient of thermal expansion CTE (ppm/K)

Density (g/cm3 )

— — — Copper Copper

Aluminum Copper Titanium Tungsten Molybdenum

237 398 7.2 160–190 180–200

23 17 9.5 5.7–8.3 7.0-7.1

2.7 8.9 4.4 15–17 9.9–10.0

SiC particles Graphite

Aluminum Epoxy

7.0–9.0 xy: −2.4

2.9–3.1 1.9

Discontinuous carbon fibers SiC particles Carbon foam

Copper

240 xy: 370 z: 6.5 xy: 300 z: 200 320 350

6.5–9.5

6.8

7.0–10.9 7.4

6.6 5.7

Continuous carbon fibers Graphite flakes

Copper

0.5–16

5.3–8.2

xy: 4.5–5.0

2.3

Aluminum Aluminum

xy: 400–420 z: 200 xy: 400–600 z: 80–110 550–600 575

7.0–7.5 5.5

3.1 —

Copper Silver Magnesium

600 400–600 550

5.8 5.8 8

5.9 5.8 —

Diamond particles Diamond + SiC particles Diamond particles Diamond particles Diamond particles

Copper Copper

Aluminum

thermal conductivities of 300–400 W/mK. The third generation is composed by materials all exceeding 400 W/mK. Many of these materials of the second and third generation are now in commercial and aerospace production systems, including power systems, servers, plasma displays, notebook computers, aircraft, spacecraft and defense electronics, and a variety of optoelectronic products. These materials include natural graphite, natural graphite/epoxy, highly ordered pyrolitic graphite (HOPG), carbon/epoxy, carbon/carbon, diamond particle-reinforced copper (diamond/Cu) and diamond particle-reinforced silicon carbide (diamond/SiC). Table 1.2 gathers some of the most important metal matrix composites organized in the three generations of heat sinks, along with their most outstanding properties [1]. Some of the composites corresponding to the families of the second and third generation that have had more impact on society (because of their significant breakthrough in its time, due to its excellent thermal properties) are presented in detail below. The text includes new findings and research currently being carried out on these materials.

1.4 Families of advanced metal matrix composite materials for electronics

| 11

1.4.1 SiC-based composites SiC is a highly versatile ceramic due to its excellent properties of high thermal conductivity and low coefficient of thermal expansion, in addition to a formidable thermal stability and extremely high hardness. It forms, in combination with aluminum, the very well-known Al/SiC composites, which have managed to be the candidates of choice for thermal control applications, and today represent the state of the art in the field of electronics. Other SiC-reinforced composite materials exist, including Ag/SiC and Cu/SiC. However, these composites present inherent problems related to their manufacture; the expected properties cannot be obtained in practice. In the case of the Cu/SiC system, calculations with the models proposed for thermal conductivity and thermal expansion coefficients provide promising results: 275 W/mK of CT and 7–9 ppm/K of CTE. However, the processing of this material by liquid infiltration leads to materials with very different properties, primarily due to the detrimental reactivity at the metalceramic interface; this leads to a solid solution of Si into the Cu matrix. There has been some attempt to improve these properties by the surface oxidation of the SiC particles [50]. While a small reduction in the extent of reaction could be observed, manifested in a slight increase of thermal conductivity, the absolute value of thermal conductivity was still very low (40–70 W/mK), which is even lower than that of a Cu/SiO2 reference sample with a similar volume fraction of ceramic phase (160 W/mK). This pre-oxidation treatment did increase the CTE up to 11 ppm/K. Another way of tackling the issue of reactivity in this system was to explore the addition of Si into the Cu matrix, which has lead to an enhanced Si activity that favors the elimination of the SiO2 film that naturally covers the SiC particle and improves wettability [51]. However, the presence of Si in the Cu matrix considerably reduces its intrinsic thermal conductivity and the final materials are rather poor in terms of this property. Moreover, the Ag/SiC system does also have problems derived from reactivity [52, 53]. Ag dissolves large quantities of oxygen that may react with the SiC. This reaction, which occurs at the interface, forms a gas that prevents direct contact between the two phases and drastically reduces the thermal properties of the final composite. The solution of oxidizing the SiC particles seems to improve, as in the case of Cu/SiC, the final properties of the material. However, for both systems, Cu/SiC and Ag/SiC, seemingly the only way to achieve the expected thermal properties by liquid state routes is to reduce the contact time between SiC particles and molten metal. Given that the problems related with reactivity in the Al/SiC system are less pronounced, admissibly high thermal properties can be easily obtained. Here we will briefly summarize two main kinds of aluminum matrix composite materials obtainable when SiC is used as reinforcement: (i) composites based on monomodal and bimodal distributions of SiC particles; and (ii) composites based on mixtures of SiC and diamond particles.

12 | 1 Metal matrix composites for thermal management (i) Al/SiC composites Traditional Al/SiC composites have had to overcome the challenge of outdoing themselves. The main challenges to overcome in this system have been its rather poor wettability [54, 55] and the undesirable interfacial reaction between Al and SiC [56–59]: 4Al(l) + 3SiC(s) → Al4 C3 (s) + 3Si(l).

(1.11)

Al4 C3 has a very poor thermal conductivity and is a highly hygroscopic product, in such a way that atmospheric moisture can degrade the thermal properties of the final composite material. Different solutions have been adopted over time to enhance wettability and/or control interfacial reactivity. A detrimental reaction can be effectively suppressed at a processing temperature of 1000 ∘ C with the incorporation of over 12 % Si into aluminum. However, this huge amount of Si degrades the thermal conductivity of the Al matrix and this solution seems to be opportune only for those applications where thermal properties are not the main concern [60]. Co-addition of 4–8 % Mg and above 6% Si, in combination with the use of N2 atmospheres, successfully improves the wettability of the system and even promotes infiltration without pressure. The resulting composites again lack decent thermal conductivity [61]. The low wettability of this system (and, in general, of most metal/ceramic systems) has proven to be a clear drawback for the use of pressure-assisted processing methods (such as, for example, pressure infiltration). Infiltration of the metal into the open channels of the preform does not take place at once, at a well-defined pressure; instead, it takes place progressively with the applied pressure when this pressure exceeds a certain minimum value (threshold pressure). Due to the microscopic configuration of the pores close to those points where particles touch, infinite pressure would be ideally necessary to obtain a complete filling of the preform and, hence, zero residual porosity. Since this is unfeasible, porosity is thus present as a phase inherently linked to the poor wettability of the metal/ceramic systems. Porosity does affect the two main properties that are important in materials management and, hence, may limit its use for this application. CTE may increase or decrease, mainly depending on the geometry of the pores. TC is strongly decreased since the pores can be viewed as a phase of zero conductivity. In [62] the authors demonstrated that for the Al-12wt.%Si/SiC system a simple application of the Hasselman-Johnson model in a two-step procedure (which accounts for the presence of two types of inclusions, SiC particles and voids, and the Al-Si matrix) offers a good approximation of the experimental results of thermal conductivity (Figure 1.2). One of the most interesting explored options to attain high values of thermal conductivity in the Al/SiC system is to increase the SiC content in the composite material. This can be achieved with the use of mixtures of two types of SiC particles largely different in size (bimodal mixtures) [39]. With these mixtures, the volume fraction of reinforcement achieves values as high as 74 % and the derived thermal properties of the resulting composites are improved in respect to those measured for composites containing a monomodal distribution of SiC particles.

1.4 Families of advanced metal matrix composite materials for electronics

| 13

Calculated Kc (W/mK)

240 220 200 180 160 140 140

160

180

200

220

Experimental Kc (W/mK)

240

Fig. 1.2. Thermal conductivity calculated with the two-step Hasselman–Johnson model vs. experimental results for samples of Al/SiC composites containing different degrees of porosity. The solid circles correspond to samples with 100 % saturation, while the empty circles correspond to samples with a given porosity content (up to 5 %). The line represents the identity function. Reproduced with permission from [62].

Table 1.3 shows the properties of various Al/SiC composite materials containing both monomodal and bimodal particle distributions. The first thing that can be observed is that the use of bimodal particle mixtures effectively increases the content of SiC in the material. As far as composites are concerned, in terms of thermal conduction, an increase in the particle content is translated into an increase of the amount of metal-ceramic interface and, given that the interface represents a thermal barrier (see equation (1.6)), there must be a positive balance between the extra contribution to heat conduction coming from the presence of more particles and the extra thermal barrier created with the new interface. Taking this into account and knowing that SiC is only slightly more thermally conductive than the aluminum matrix, it might therefore be expected that increasing the content of SiC would not result in a substantial increase in the thermal conductivity of the final maTable 1.3. Main characteristics of Al/SiC composites with monomodal and bimodal distributions of particles. Vr is the particle volume fraction of reinforcement, D is the average size determined by laser scattering, TC is the thermal conductivity (W/mK) and CTE is the coefficient of thermal expansion (ppm/K) [63–66]. Reinforcement SiC F-100 SiC F-180 SiC F-240 SiC F-320 SiC F-400 SiC F-500 SiC F-800 25% SiC F-100 + 75% SiC F-500 50% SiC F-100 + 50% SiC F-500 67% SiC F-100 + 33% SiC F-500 75% SiC F-100 + 25% SiC F-500

Vr

μm) D (4/3) (μ

TC (W/mK)

CTE (ppm/K)

0.58 0.58 0.6 0.59 0.58 0.55 0.53 0.62 0.69 0.74 0.72

167 86 57 37 23 17 9 — — — —

221 209 203 204 194 193 154 215 220 228 225

12.7 — 12.5 12.9 — 13.4 14.9 10.8 9.0 7.8 8.4

14 | 1 Metal matrix composites for thermal management terial. However, due to the much lower thermal expansion coefficient of SiC compared to that of Al, composites containing bimodal mixtures exhibit a much lower CTE than that of its analogues with monomodal content. It is highly desirable the study and prediction of the values of volume fraction of reinforcement in bimodal mixtures in order to perform a proper design of materials. In Figure 1.3, the results of volume fraction of Table 1.3 for different bimodal particle packing samples are gathered in comparison with the prediction derived from models presented in equations (1.1) and (1.2) and equations (1.3) and (1.4). 1.0

0.9

Vp

0.8

0.7

0.6

0.5

0

20

40 60 % Coarse particles

80

100

Fig. 1.3. Experimental results (solid circles) for particle volume fraction (Vp ) of bimodal mixtures of SiC compacts formed by angular particles of 17 μm and 170 μm compacts versus percentage of coarse particles in the mixture. The results obtained by means of the model discussed in the text are also shown: fine particles end (solid line: equation (1.1)) and coarse particles end (dashed line: equation (1.2)). The results of the fittings of the experimental results by means of equations (1.3) and (1.4) are depicted by means of the short-long dashed line. Reproduced with permission from [39].

Al/SiC composite materials for the different bimodal mixtures of reinforcement presented in Table 1.3 have been fabricated by liquid gas pressure infiltration [39]. These composites exhibit microstructures with homogeneous particle distributions at the level of the mesoscale (Figure 1.4a–c). Only the mixture 75 : 25 (large particles : small particles) presented problems of segregation during packaging (Figure 1.4d). Recent studies have demonstrated the possibility of improving the packing process by following a wet route compaction. The thermal properties of these composites, collected in Table 1.3, are presented for analysis in Figures 1.5 and 1.6. Figure 1.5 shows that the thermal conductivity of

1.4 Families of advanced metal matrix composite materials for electronics

100 μm (a)

| 15

200 μm (b) 200 μm

200 μm

(c)

(d)

Fig. 1.4. Optical micrographs of cross sections of different samples of Al/SiC composites containing SiC particles of: (a) 17 μm (corresponding to 500 mesh); (b) 170 μm (corresponding to 100 mesh); (c) a bimodal mixture of 67 % of 17 μm particles and 33 % of 170 μm particles; (d) a bimodal mixture of 75 % of 17 μm particles and 25 % of 170 μm particles. Reproduced with permission from [39] and [44].

these materials can conveniently be calculated using the GDEMS model, both for monomodal and bimodal mixtures. Different theoretical results for the coefficient of thermal expansion have been depicted in Figure 1.6 for composites containing mono- and bimodal distributions of SiC particles [63]. The discrepancies present between experimental results and predictions can be qualitatively understood by noting that equations (1.9) and (1.10) were developed using thermoelasticity theory. In this sense, experimental results of CTE obtained at lower temperatures might be closer to the predictive values, given that plastic deformation becomes greater at higher temperatures. These results, in consequence, illustrate the failure of the thermoelasticity framework upon which most theories are based.

240 220

230 220

200 K (W/mK)

Calculated thermal conductivity [W/mK]

16 | 1 Metal matrix composites for thermal management

180 160 140 140

200 190 180

160

180

200

220

240

0

Experimental thermal conductivity [W/mK] (a)

210

20

40 60 Xc (%)

80

100

(b)

Fig. 1.5. Thermal conductivity of the Al/SiC composites produced by infiltration of aluminum into the preforms of Table 1.3. (a) Data for composites containing monomodal distributions of SiC particles of different average size diameters in comparison with the results predicted by the GDEMS model (the values taken for the intrinsic thermal conductivity of the inclusion (Kin ) and the thermal interface conductivity (h) are 253 W/mK and 7.5 × 107 W/m2 K, respectively). Reproduced with permission from [63]. (b) Data for composites containing bimodal SiC mixtures (SiC F-500 + SiC F-100: Table 1.3) vs. the percentage of coarse particles. Modelling with the Hasselman–Johnson model, which gives the same results as the GDEMS model for this system, is also presented for comparison. Reproduced with permission from [65].

(ii) Al/SiC-diamond composites One way to increase the thermal conductivity of the Al/SiC composites, more effective than the use of bimodal mixtures of SiC particles, is to include another highly thermally conductive ceramic particulate, such as diamond particles, in combination with SiC. This proposal was first presented in [45]. A mixture of particles of SiC and diamond of similar size (average diameter around 200 μm) were used as reinforcement, packed up to their maximum volume fraction of 0.58, forming a porous ceramic preform. The preform was consolidated afterwards with Al by gas pressure infiltration. The aim of using this particular mixture of ceramics is to achieve a decrease in the material costs while maintaining the thermal conductivity at relatively high values. Figure 1.7 shows the change of the thermal conductivity in composites fabricated by liquid infiltration of aluminum into preforms containing mixtures of SiC and diamond particles as a function of the overall fraction of diamond in the reinforcement.

1.4 Families of advanced metal matrix composite materials for electronics

| 17

16

11

14

9 CTE (10˗6 K˗1)

CTE (10˗6 K˗1)

13

12

10

7

14 12 10

8

8 6

6

0.52

0.58

0.64 Vp

0.7

(a)

0.76

0.55 0.6

0.65 Vp

0.7

0.75

(b)

Fig. 1.6. (a) Thermal expansion coefficient of Al/SiC composites obtained from a linear fitting of the experimental data for the thermal expansion over the range 50–300 ∘ C vs. particle volume fraction. The results correspond to composites having either a single particle size (circles) or bimodal particle distributions (triangles). The data can be satisfactorily fitted by the straight line CTE (ppm/K) = 31.3–32.0 × Vp ; reproduced with permission from [63]. (b) Experimental and theoretical results for the coefficient of thermal expansion versus particle volume fraction in the temperature ranges 100–150 ∘ C (upper plot) and 200–250 ∘ C (lower plot). The different curves and symbols correspond to: experimental (circles), linear fitting of experimental data (dashed line), upper and lower bounds predicted by Schapery’s model (solid lines: equation (1.9)) and linear rule of mixtures (short-long dashed line: equation (1.8)). Reproduced with permission from [63].

Composite thermal conductivity [W/mK]

700 600 500 400 300 200 Experiment

100

GDEM, Eqn. (3) Km=237 W/mK GDEM, Eqn. (3), Km=185 W/mK

0

0 0.2 0.4 0.6 0.8 1 Diamond fraction of reinforcement [–]

Fig. 1.7. Thermal conductivity of composites based on mixtures of SiC and diamond particles of similar size (average diameter around 200 μm) as a function of the overall fraction of diamond in the reinforcement. The volume fraction of reinforcement for all composites was 0.58 ± 0.01. The two continuous lines are calculated with the GDEMS model for matrix conductivity of 185 and 237 W/mK for the lower (dashed) and the upper (solid) line, respectively. Reproduced with permission from [45].

18 | 1 Metal matrix composites for thermal management For modeling purposes, it is necessary to consider a relatively low thermal conductivity of the metal matrix of 185 W/mK in those materials containing a high proportion of SiC particles (the other parameters needed for modeling are collected in Table 1.4). This is because during infiltration, liquid aluminum reacts with the SiC particles to form aluminum carbide and Si. Si is then dissolved within the metal and decreases its intrinsic thermal conductivity (equation 1.11). Aluminum is also expected to react with diamond to form aluminum carbide at the interface. However, no other reaction product has been identified. In consequence, adequate processing conditions (i.e. infiltration pressure rate) must be explored in order to optimize the properties of the final materials. Table 1.4. Material parameters used in the calculation of Figure 1.7 according to the generalized differential effective medium scheme (GDEMS).

Intrinsic thermal conductivity (W/mK) Interface thermal conductance with pure aluminum (W/m2 K)

SiC

Diamond

259 [62] 1.05 × 108 [62]

1450 [45] 5 × 107 [45]

With this system, a target value of 400 W/mK can be reached with only 50–60% of diamond in the preform, the rest being inexpensive SiC powder.

1.4.2 Carbon-based composites 1.4.2.1 Composites based on graphite particles, carbon fibers or carbon nanotubes Carbon/metal composites are currently used in several applications. Since their early use in electrical contactors, such as brushes for engines and generators [5, 67, 68] or sliding contacts [5, 69], these materials have attracted the interest of many researchers due to their excellent ray attenuation in plasma-facing components and their selflubrication properties for automotive pistons [70, 71]. There has been, however, recent activity addressed to investigate their use as heat sink elements in multi-functional electronic packaging systems [72–74]. While the thermal properties of these materials have not shown to be excellent [75], there are other properties that make them very attractive for these applications in electronics: low price of the raw materials, cheap fabrication procedures and high ease of machinability. Finely divided carbon-based reinforcements (mainly graphite particles and short carbon fibers) have been mixed with metals through various fabrication techniques (infiltration, drain casting, etc.), but none of them have demonstrated enough outstanding values in order to cover the requirements for entering into the third generation of materials for thermal management [76]. The most explored metallic matrices have been Al and Cu [77, 78]. Adding alloying elements seems mandatory for both metals. In the case of Al, Si is added in order to reduce the tendency towards Al4 C3

1.4 Families of advanced metal matrix composite materials for electronics

20 μm (a)

| 19

40 μm (b)

Fig. 1.8. Optical micrographs of the composite materials obtained by infiltration with Al–12wt.%Si alloy of graphite particles (a) and carbon fibers (b). Reproduced with permission from [76].

formation at the interface between carbon and liquid aluminum [79]: 4Al(l) + 3C(s) → Al4 C3 (s).

(1.12)

Al4 C3 is a detrimental product in terms of thermal conductivity and material stability (as already seen in Section 1.4.1). Microstructures of two of these materials can be seen in Figure 1.8. Some representative values of their thermal properties for composites with Al-Si matrices are gathered in Table 1.5. Table 1.5. Thermal properties of composites obtained with different carbon reinforcements and Al12wt.%Si alloy. Vr is the total volume fraction of reinforcement [76]. Reinforcement Graphite particles Short carbon fibers

Vr

metal

CT (W/mK)

CTE (ppm/K)

0.67 0.55

Al-12Si Al-12Si

118 131

11 2.8

Cu is alloyed with carbide-forming elements in order to improve interfacial bonding, which is otherwise so weak that thermal conductivity goes down to really poor values (less than 100 W/mK). The addition of 0.5–0.8vol.% Ti to Cu results in the formation of TiC on carbon fibers (CuTi2 remains in copper alloy), which improves the interfacial cohesion [80]. Even with such an effect, the resulting thermal conductivities are in the range 110–200 W/mK. The use of continuous carbon fibers and a novel processing technique based on sputtering produces anisotropic materials with a thermal conductivity in the direction of the aligned fibers of 800 W/mK (the intrinsic conductivity of the carbon fibers in that direction is about 1100 W/mK) [81]. Using nanoscale reinforcements in metal matrices has been considered an entrance into the fourth generation of materials for thermal control. Estimates of the thermal conductivity of carbon nanotubes predicted values as high as 6600 W/mK.

20 | 1 Metal matrix composites for thermal management Values above 3000 W/mK have been measured. Other promising reinforcements are graphite nanoplatelets and diamond nanoparticles [1]. These reinforcements are attractive due to their high thermal conductivity but, dispersed in a metal matrix, they do generate a lot of interface that reduces the thermal conductivity of the material. It is for this reason that the reinforcements on the nanometer scale, despite the high intrinsic thermal conductivity of some, are likely to be better designed in combination with some other highly conductive reinforcing material such as hybrid composites, where nanoreinforcements provide other properties such as a reduced CTE or increased stiffness [82–84].

1.4.2.2 Composites based on graphite flakes A new family of high thermal conductivity composites has recently been developed; it is produced by infiltration of a metallic alloy into preforms of mixtures of graphite flakes and either ceramic or carbon materials (in the form of particles or short fibers), [76, 85–87]. The microstructure of these materials roughly consists of alternating layers of flakes and metal-particles composite (Figure 1.9). The underlying idea of these composites is to exploit the anisotropic characteristics of graphite (trough graphite flakes) in a material that has a layered structure and keeps a certain degree of anisotropy; it is consequently useful for thermal guides or other bi-dimensional heat extraction applications. The layered structure is achieved by ordering the graphite flakes in planar arrangements. When packed, flakes do not pose problems related to orientation, as they naturally tend to lie on top of each other. However, they leave almost no space between them; this makes metal infiltration an almost unfeasible task. To solve this problem, the authors of [76] used SiC particles or carbon fibers premixed with graphite flakes and pressed uniaxially up to 40 MPa. The particles or fibers have the function of keeping the graphite flakes separated so that the preform can be infiltrated with liq-

40 μm (a)

80 μm (b)

Fig. 1.9. Optical micrographs of the composite materials obtained by infiltration with Al-12wt.%Si alloy of 90 % graphite flakes + 10 % carbon fibers (a), and 60% graphite flakes + 40 % SiC particles (b). Reproduced with permission from [76].

1.4 Families of advanced metal matrix composite materials for electronics

| 21

uid metal. Graphite flakes are responsible for the high thermal conductivity of these materials measured in the two directions of the graphite flakes planar alignment. The particles or fibers, meanwhile, apart from being necessary in order to facilitate infiltration, do also help in maintaining low values of CTE in the perpendicular direction to the graphite flakes planes. Table 1.6 shows the most significant results of thermal properties reached for some of these materials. These composites exhibit exceptional high values of thermal conductivity in the two directions of the graphite flakes planar alignment (note that they reach 548 W/mK for Ag-3%Si/graphite flakes + carbon fibers). These values are susceptible to increase with an optimization of the infiltration processing conditions and/or with heat treatments that can modify the metallurgical state of the matrix and, at the same time, improve the quality of the metal/carbon interphase. The possibility of ease of machinability allows one to obtain complex shapes; this advantage has been a determining factor in spurring industrial interest. Table 1.6. Thermal properties of composites obtained with graphite flakes and SiC particles or, alternatively, carbon fibers and Al-Si and Ag-Si alloys. Vp is the total volume fraction of reinforcement. xy refers to the graphitic planes, while the direction perpendicular to it is denoted by z [76] Reinforcement

Vp

metal

CT (W/mK)

CTE (ppm/K)

90 % graphite flakes + 10 % carbon fibers

0.88

Al-12Si

90 % graphite flakes + 10 % carbon fibers

0.88

Ag-3Si

60 % graphite flakes + 40 % SiC particles

0.88

Al-12Si

63 % graphite flakes + 37 % SiC particles

0.88

Ag-3Si

z: 50 xy: 367 z: 80 xy: 548 z: 45 xy: 368 z: 48 xy: 360

z: 24 xy: 3.0 z: 21 xy: 3.0 z: 11 xy: 7.0 z: 11 xy: 8.0

The possibilities of use of anisotropic materials in electronics are generally limited because the industry prefers materials with isotropic characteristics (see Table 1.1). In such cases, it is advisable to conduct a detailed characterization of their properties and use appropriate design techniques in order to study how to implement these materials in devices that, a priori, would be reserved for isotropic materials. In the present case, the materials are good thermal conductors in two dimensions and can be used in applications in which heat removal must be done bi-dimensionally (with almost no heat transfer in a certain direction); they can thus be perfect thermal guides for several applications in electronics. Their microstructure lies in between two idealized over-simplified models: (a) alternating layers of flakes and metal-particles composite; or (b) oriented discs (flakes) randomly distributed in a metal-particles composite. The experimental results for the thermal conductivity of the composites were fitted by means of expressions corre-

22 | 1 Metal matrix composites for thermal management sponding to those two microstructures using the thermal conductivity of the flakes as the only fitting parameter (Figure 1.10). Both models led to similar results with thermal conductivities of approximately 500 and 30 W/mK in the directions parallel and perpendicular to the graphite planes, respectively [88]. These values are lower and higher (respectively for the parallel and perpendicular directions of graphite flakes planes) than those of crystalline graphite, as one may have easily inferred by taking into account the presence of structural defects and impurities. However, the interesting point in the comparison of these two modeling schemes is that graphite flakes form quasi-continuous layers that are responsible for the high thermal conductivity of these materials; the other phases are helpful in consolidating the material and controling other properties such as the coefficient of thermal expansion.

1.4.2.3 Composites based on graphite foams The foamability of materials is very interesting because it has extended the technological use and applications of many traditional solids. In recent decades, many methods have been developed for manufacturing metal foams. The foaming process determines the final properties of the foam, which mainly depends on size, size distribution and shape of the pores contained therein. One of the most versatile methods is the replication method, whereby open-pore foams with strict control on pore characteristics are obtained. This method is based on a liquid metal infiltration of a preform made up of compacted particles of a leachable material that can be removed afterwards when the metal is in solid state (for example, NaCl particles, which can be removed by water dissolution). The replication method has been conveniently used for the manufacture of foams of Al and Mg with applications as heat sinks for thermal management in electronics. This method has been recently applied to the fabrication of graphite foams by infiltrating a bed of particles of NaCl with mesophase pitch (see Figure 1.11) [89]. Mesophase pitch is a material that can carbonize and graphitize with successive heat treatments. This results in high thermal conductive graphitic foams that can be subsequently used as porous bodies to be infiltrated with liquid metals. The derived metal matrix composite materials present continuity in the reinforcement phase and do have interesting thermal properties for their use in thermal management. Figure 1.12 shows the fracture surfaces of different samples of foams, evolving from stabilized pitch foam to graphite foam, with the intermediate carbonized state. In view of the micrographs, it is clearly noted that while the stabilized pitch foam presents a compact texture in the struts (Figures 1.12a and 1.12b), increasingly ordered structures are achieved after the heat treatments of carbonization (Figures 1.12c and 1.12d), graphitization at 2500 ∘ C (Figures 1.12e and 1.12f) and graphitization at 2750 ∘ C (Figures 1.12g and 1.12h). These ordered structures are responsible for the superb thermal conductivity of these materials. The arrangement and orientation of the graphitic planes depends mainly on how the mesophase pitch has flown during

| 23

1.4 Families of advanced metal matrix composite materials for electronics

Layered composite

Randomly distributed oriented discs

Serie arrangement

Graphite flake

Al-12wt%Si x SiC

y

x

z z

y

Al-12wt%Si

x

SiC

y

x

z

y

(b)

(a)

400

380

380

Kc,cal (W/(m.K))

400 360 340 320 300

360 340 320 300 280

280

280 300 320 340 360 380 400

280 300 320 340 360 380 400

80

80

70

70

Kc,cal (W/(m.K))

Kc,cal (W/(m.K))

Ƙc,exp (W/(m.K))

(d)

Ƙc,exp (W/(m.K))

(c)

60 50 40 30 30

(e)

Parallel arrangement Direction

Direction

Direction

Direction

Graphite flake z

Kc,cal (W/(m.K))

Serie arrangement

Parallel arrangement

40

50

60 Ƙc,exp (W/(m.K))

70

60 50 40 30 30

80 (f)

40

50 60 Ƙc,exp (W/(m.K))

70

80

Fig. 1.10. Results obtained by means of a model based on a layered structure (a) and a model based on random distribution of oriented discs (b) vs. experimental results for the thermal conductivity of the ternary Gf /Al-12Si-SiCp composites. The results were obtained for two values of the particle volume fraction in the Al-12Si-SiCp composite, namely, 0.52 and 0.58, which actually bound the particle content in the ternary composite, without noting appreciable differences. (c)–(d) and (e)–(f) correspond to the TC along and transversally to the graphite planes, respectively. Reproduced with permission from [88].

24 | 1 Metal matrix composites for thermal management

(a)

(b)

(c)

(d)

(e)

Fig. 1.11. Different steps of the replication processing for pitch-based carbon foams fabrication: (a) powder packing of leachable NaCl particles; (b) infiltration with molten pitch (typical conditions: 400 ∘ C and 5 bar); (c) directional solidification after infiltration; (d) NaCl removal in a stirring water bath; (e) drying at low temperature to get the final pitch-based material which is thermally stabilized and subsequently carbonized or graphitized. Reproduced with permission from [89].

infiltration. In liquid state and under low pressurization rates, the pitch enters the preform flowing around the NaCl particles, following its contour, in a way that the pitch shows a three-dimensional ordering structure. This makes the final foam have locally organized stacked graphitic planes arranged at the mesoscale in all directions, so that the foam shows completely isotropic properties. It is the possibility of achieving a high degree of graphitization and, accordingly, good thermal and electrical conductivities that makes these materials good candidates for electrodes and thermal management applications [90–92]. In fact, for those materials graphitized at the maximum temperature (heat treatment at 2750 ∘ C for 30 min) the thermal conductivity was characterized to be as high as 50 W/mK in the three spatial directions for a total open porosity of 81.4 %. This value corresponds to an intrinsic thermal conductivity for the graphitic structure of the foam wall cells of about 630 W/mK (calculated with the GDEMS model, conveniently adapted to foams); this is close to or even higher than those values obtained for commercial foams with the advantage of this material being isotropic [89]. Different metal matrix composites, derived from graphite foam infiltration, have been developed [91, 93]. In [93], anisotropic commercial graphite foams of the POCOHTC type, from Poco Graphite, Inc. (Decatur, Texas, EEUU), were successfully gas pres-

1.4 Families of advanced metal matrix composite materials for electronics

25μm

100μm (a)

(b)

2μm

25μm (c)

(d)

100μm (e)

2μm (f)

100μm (g)

| 25

20μm (h)

Fig. 1.12. Fracture images obtained with scanning electron microscopy (SEM) showing the structural details of mesophase pitch-derived foams fabricated via the replication method: stabilized pitch foam (a, b); carbonized pitch foam (c, d); graphitized pitch foam at 2500 ∘ C (e, f); and graphitized pitch foam at 2750 ∘ C (g, h). Reproduced with permission from [89].

26 | 1 Metal matrix composites for thermal management Table 1.7. Thermal properties of composites obtained with POCO-HTC graphite foam gas pressure infiltrated with Al-Si and Ag-Si alloys [93]. System

TC (W/mK)

CTE (ppm/K)

Al-12wt.%Si/POCO-HTC graphite foam

z: 140 xy: 270–330 z: 220 xy: 305–387

z: 11.6 xy: 9.9 z: 11.6 xy: 10.1

Ag-3wt.%Si/POCO-HTC graphite foam

sure infiltrated with commercial eutectic Al-12wt.%Si and Ag-3wt.%Si alloys. The main thermal properties of the infiltrated samples are summarized in Table 1.7. The anisotropy of the composites is due to the anisotropy of the commercial graphite foam preforms, which has been attributed to preferential alignment of the crystallographic planes as a consequence of the manufacturing process. These commercial foams are mainly fabricated by the blowing technique, based on the liberation of volatile species contained in the mesophase pitch; this orientates the pitch domains along the direction of pressure release. The results of thermal conductivity obtained for the foams presented in Figure 1.12, in combination with the thermal properties obtained for the composites developed in Table 1.7, allows one to point out that composites derived from highly thermal conductive isotropic foams, which are currently under development, can be perfect candidates for different applications in thermal management.

1.4.3 Diamond-based composites Diamond is the natural material that exhibits the highest known thermal conductivity (as high as 2200 W/mK [94]). The price of artificial diamond, which has been one of its main drawbacks for its use, has been falling over time and that is why it is currently a perfect candidate as a reinforcement of composite materials for electronic purposes [1, 6, 94–96]. Because diamond has such a high thermal conductivity, it is tempting to combine it with silver or copper as metal matrices in order to produce composites with the highest possible thermal conductivities. However, the physical limits calculated with modeling have not been easily achieved and researchers have adopted different solutions to mitigate the scarce interaction between these metals and diamond. Thermal conductivity is dependent on many variables that concern the nature of phases and interfacial thermal connectivity between phases. With the aim of achieving the highest possible thermal conductivity value in metal/diamond composites, research groups have based their activities on: i) exploring the effects of metal alloying; ii) increasing the diamond volume fraction by the use of bimodal mixtures; and iii) investigating the possible effects of different processing conditions.

1.4 Families of advanced metal matrix composite materials for electronics

| 27

(i) Effects of metal alloying Some authors have varied the metal composition by adding carbide-forming active elements like B or Cr in small amounts (< 10at%). Materials prepared by liquid metal infiltration with Cu-B exhibited optimized results of > 700 W/mK [97]. The need of carbides formation on the surface of the diamonds is demonstrated in [97], where the dependence of the TC and the CTE of diamond/Cu-Cr composites on the concentration of chromium in copper is shown. The authors observed a transition from weak to strong bonding at the metal-diamond interface. As a consequence, a clear increase of the TC over 600 W/mK and a significantly decrease of the CTE down to less than 10 ppm/K follows. Despite the positive effect of carbides presence at the metal/diamond interface, their thicknesses need to be properly controlled, given that the thermal conductivities of these refractory carbides usually lies in the range 7–170 W/mK. The optimum thickness of these carbides is confined to the 50–200 nm range [94]. (ii) Use of bimodal mixtures of particles One of the most interesting explored options to attain high values of thermal conductivity is to increase the reinforcement content in the composite material by using combinations of diamond particles with particles of the same [96, 98–100] or another nature [101, 102] of different average sizes (bimodal mixtures). With preforms combining diamonds of roughly 9 : 1 size ratios, maximum thermal conductivities of 725 W/mK and 970 W/mK have been reported for pure Al and Ag-3wt.%Si matrices, respectively [96]. The CTE at room temperature varied as a function of the diamond volume fraction in between 3.3–7.0 ppm/K and 3.1–5.7 ppm/K for the Al-based and the Ag-3Si-based composites, respectively. Aluminum gas-pressure infiltrated combinations of coarse diamond and small SiC particles with a size ratio of 7, 8 : 1 have shown that the thermal conductivity (with a maximum of 540 W/mK for 100 % diamond content) is decreased by the presence of SiC, while the coefficient of thermal expansion is simultaneously positively reduced [101]. (iii) Effect of processing conditions Processing conditions have been demonstrated to be very important at the level of interface quality and metallurgical states. Various research groups have produced composite materials with diamond and metal matrices through different processing techniques, namely gas pressure infiltration, squeeze casting and spark plasma. An interesting body of research has been devoted to comparing composites fabricated with squeeze casting and pressure assisted-infiltration methods [103–105]. Squeeze casted Al-7wt.%Si/diamond composites, for which metal was rapidly solidified, exhibited lower thermal conductivities than those obtained with gas pressure-assisted infiltration [105]. Those authors already concluded that contact time between the liquid metal and diamond promoted interfacial bonding relying on partition of the elements involved within the interface layer. Other authors have explored diamond-based materials processed through vacuum hot pressing methods; they found values of ther-

28 | 1 Metal matrix composites for thermal management mal conductivity in the range 320–567 W/mK, depending on sintering temperature and time [106]. Values ranging from 400 W/mK to 550 W/mK were achieved for the Al/diamond system by spark plasma processing [107, 108]. The effect of diamond surface characteristics on the thermal conductance of the diamond-metal interface and the thermal conductivity of the final composite materials has also been proved to be important. Oxygenated diamond samples have a thermal interface conductance with aluminum roughly four times higher than H-treated samples [109]. Different metallic coatings have been used to effectively increase the interfacial thermal conductance [110–112]. While some of the progresses detailed above are ultimately positive, some authors have identified the need to enhance the intrinsic thermal conductivity of diamond as the most promising direction of progress [97]. Very recent studies that attempt to be exhaustive enough about processing characteristics and interface evolution are those published in [113]; here the effect of processing conditions on the thermal conductivity of aluminum/diamond composites fabricated by means of gas-pressure infiltration have been identified [113, 114]. A close view of a diamond particle of ISD 1700 (40/50 mesh) is shown in Figure 1.13.

(111) (111) (100) (100) 10 um

250 um SE

(a)

WD15.1mm 20.OkV x180 250um

SE

WD14.7mm 20.OkV x3.5k 10um

(b)

Fig. 1.13. Scanning electron microscopy (SEM) image of a diamond particle of ISD1700 40/50 quality; (b) is a magnification of one of the sharp edges of the particle shown in (a). Reproduced with permission from [113].

With the aim of obtaining reliable information on the aluminum/diamond interface in composite materials, a sample preparation method, based on electrochemical etching and spray cleaning, has been proposed recently in [114]. The method allows the reaction products to be characterized, and thus makes a reliable assessment of the sensitivity of the diamond faces for reactions with aluminum to form aluminum carbide. Figure 1.14 shows the evolution of the thermal conductivity versus contact time; contact time is defined here as the time during which diamond is in contact with liquid metal and reaction can take place. At negligible contact time, TC has a value that strongly depends on infiltration temperature, being rather low at an infiltration temperature of 760 ∘ C (around

1.4 Families of advanced metal matrix composite materials for electronics

| 29

Thermal conductivity (W/mK)

700 650 600 550 500 450 400 0 1

2 3 4 5

10 15 20 25 30 35 40 45

Contact time (min) Fig. 1.14. Thermal conductivity of aluminum/diamond composites plotted as a function of the contact time for two infiltration temperatures: 760 ∘ C (empty circles) and 850 ∘ C (solid circles). Dotted lines are guides to the eye. Reproduced with permission from [113].

400 W/mK). As the contact time increases, TC also increases, steeply for 850 ∘ C and slowly for the lower temperature, showing a maximum at a certain time that decreases with temperature. Specifically, the maximum occurs at 15 and 1 min for 760 ∘ C and 850 ∘ C, respectively. The width of the maximum decreases as the temperature increases, while the maximum TC is higher for higher infiltration temperatures. A maximum TC of 676 W/mK was attained for an infiltration temperature of 850 ∘ C and a contact time around 1 min, a rather high value for a composite based on pure aluminum. Micrographs of the interfacial reaction products are gathered in Figure 1.15 for different infiltration conditions. X-ray diffraction analysis on the Al/diamond composites illustrated that Al4 C3 is the only crystalline product that is formed in the range of temperatures studied. The joining of Figure 1.14 and Figure 1.15 in a common explanation is not yet clear. The role of the Al4 C3 crystals on the thermal performance of these materials is still an open question that needs deeper and fundamental studies. At 760 ∘ C and contact times close to zero, the surface of the {100} planes appears covered by a large number of small Al4 C3 crystals. Contrarily, the {111} faces are almost free of any reaction product. A possible explanation behind these findings seems to be related to the fact that Al4 C3 free regions on the diamond surfaces may seemingly establish a strong interaction between diamond and Al (with dissolved carbon) that can contribute very positively to the heat transport across the interface, hence inducing a higher thermal conductance. In conclusion, the control key for the TC of these composites lies on metal-diamond interface chemical interactions, which may be properly controlled by optimization

30 | 1 Metal matrix composites for thermal management

10 um

10 um SE

(a)

WD12.4mm 20.OkV x3.5k 10μm

SE

(b)

SE

(d)

WD11.6mm 20.OkV x3.5k 10μm

10 um SE

SE

(e)

WD14.6mm 20.OkV x3.5k 10μm

(c)

10 um

10 um SE

WD18.4mm 15.OkV x3.5k 10μm

WD12.9mm 20.OkV x3.5k 10μm

10 um SE

WD11.8mm 20.OkV x3.5k 10μm

(f)

Fig. 1.15. Scanning electron microscopy (SEM) images of composites fabricated at 760 ∘ C [(a), (c) and (e)] and 850 ∘ C (b, d, f) with different contact times: tc = 0 min (a, b); maximum values of TC, tc = 15 min (c) and tc = 1 min (d); tc = 45 min (e, f). The upper left side of each micrograph corresponds to the (111) face and the lower right side to the (100) face. Reproduced with permission from [113].

of contact time and temperature during their processing. The interface conductance of any reactive metal/diamond system (like the Al/diamond system presented here), which mainly determines the final thermal conductivity of its composites, is strongly influenced by the conditions used for the material processing. In consequence, a unique value of thermal conductance or thermal conductivity does not characterize, per se, any of the reactive metal/diamond systems, unless a careful and systematic study on how these parameters evolve with processing conditions is noted.

1.5 The future of metal matrix composites in electronics We can say that we are probably only at the beginning of a revolution in the development of thermal management materials. For the last three decades – since electronics began to sense a great need of finding new materials for thermal control – the development of new and better materials has followed a historic evolution. The first Al/SiC materials were sold in the 1980s, and third-generation materials were commercialized only less than a decade ago. The future of this sector is therefore very promising, given the electronics industry’s current needs. Since the very beginning of this research path, it has been established that the metal matrix composites with reinforcements at the micro/meso-scale were most suit-

References

| 31

able. To achieve the properties of high thermal conductivity and matched thermal expansion coefficient, apart from combining the materials in the proper ratio, it is necessary to properly control processing conditions in order to get a proper matrixreinforcement interface. This interface is typically formed by interfacial products, which are less thermally conductive than the reinforcing material and sometimes less than the matrix itself. That is why the design of the nature and thicknesses of the interfaces is so important. Specifically, it has been observed that interfaces of the order of some nanometers in thickness are optimal. In the current era of nanotechnology, it seems obvious to explore the use of certain very highly conductive reinforcements (such as carbon nanotubes) in metal matrix composites. However, their limited size makes the dispersion of an acceptable volume fraction in a metal matrix difficult, and moreover it generates a great amount of metal-reinforcement interface that is detrimental to the thermal conductivity of the final material. Nanotechnology has much to say in the future development of metal matrix composite materials for thermal control – not so much in the use of nano-sized materials, but rather in the nano-dimensional control of the metal-reinforcement interfaces.

References [1] [2] [3] [4] [5] [6] [7]

[8] [9]

[10] [11] [12]

Zweben C. Thermal materials solve power electronics challenges. Power Electronics Technology 2006;32:40–47. Zweben C. Electronics packaging: heat sink materials. Encyclopaedia of Materials Science and Technology 2001;3:2676–2683. Frear DR. Materials Issues in area-array microelectronic packaging. Journal of Metals 1999;51:22–27. Herard L. Packaging reliability. Microelectronic Engineering 1999;49:17–26. Kaczmar JW, Pietrzak K and Wiosinki W. The production and application of metal matrix composite materials. Journal of Materials Processing Technology 2000;106:58–67. Mallik S, Ekere N, Best C, Bhatti R. Investigation of thermal management materials for automotive electronic control units. Applied Thermal Engineering 2011;31:355–362. Clyne TW. An introductory overview of MMC systems, types and developments, In: Comprehensive Composite Materials, Kelly A and Zweben C (Eds.),1–26, Elsevier Science 2000;3. Clyne TW. Thermal and electrical conduction in MMCs, In: Comprehensive Composite Materials, Kelly A and Zweben C (Eds.),447–468, Elsevier Science 2000;3. Qu X-H, Zhang L, Wu M and Ren S-B. Review of metal matrix composites with high thermal conductivity for thermal management applications. Progress in Natural Science: Materials International 2011;21:189–197. Jiang G, Diao L, Kuang K. Advanced Thermal Management Materials 2013. Ashby MF. Criteria for selecting the components of composites. Acta Metallurgica et Materialia 1993;41:1313–1335. Eustathopoulos N, Nicholas MG, Drevet B. Wettability at high temperatures. Pergamon Materials Series 1999.

32 | 1 Metal matrix composites for thermal management [13] Michaud V, Mortensen A. Infltration processing of fibre reinforced composites: governing phenomena. Composites Part A 2001;32:981–996. [14] García-Cordovilla C, Louis E, Narciso J. Pressure infiltration of packed ceramic particulates by liquid metals. Acta materialia 1999;47:4461–4479. [15] Mortensen A, Jin I. Solidification processing of metal matrix composites. International Materials Reviews 1992;37:101–128. [16] Bahraini M, Molina JM, Kida M, Weber L, Narciso J, Mortensen A. Measuring and tailoring capillary forces during liquid metal infiltration. Current Opinion in Solid State and Materials Science 2005;9:196–201. [17] Molina JM, Narciso J and Louis E. On the triple line in infiltration of liquid metals into porous preforms. Scripta Materialia 2010;62:961–965. [18] Molina JM, Prieto R, Duarte M, Narciso J, Louis E. On the estimation of threshold pressures in infiltration of liquid metals into particle preforms. Scripta Materialia 2008;59:243–246. [19] Piñero E, Molina JM, Narciso J, Louis E. Liquid metal infiltration into particle compacts chemically and morphologically heterogeneous. Materials Science and Engineering A 2008;495:288–291. [20] Fuller WB, Thompson SE. The laws of proportioning concrete. Transactions of the American Society of Civil Engineers 1907;59:67–172. [21] Cumberland DJ, Crawford RJ. The packing of particles. Amsterdam: Elsevier Science;1987. [22] Dodds JA. The porosity and contact points in multicomponent random sphere packings calculated by a simple statistical geometric model. Journal of Colloid and Interface Science 1980;77:317–327. [23] Ouchiyama N, Tanaka T. Porosity estimation for random packings of spherical particles. Industrial and Engineering Chemistry Fundamentals 1984;23:490–493. [24] Ouchiyama N, Tanaka T. Porosity estimation from particle size distribution. Industrial and Engineering Chemistry Fundamentals 1986;25:125–129. [25] Guyon E. The physics of random matter, in J. Souliete, J. Vannimenus and R. Stora (eds.). Le hasard et la matière, Les Houches 1986 1987;46:71–156. [26] Spaepen F. A structural model for the solid-liquid interface in monoatomic systems. Acta Metallurgica 1975;23:729–743. [27] Latham JP, Munjiza A, Lu Y. On the prediction of void porosity and packing of rock particulates. Powder Technology 2002;125:10–27. [28] Zou RP, Yu AB. Evaluation of the packing characteristics of mono-sized non-spherical particles. Powder Technology 1996;88:71–79. [29] Burminster DM. The grading density relations of granular materials. Proceedings-American Society for Testing Materials 1938;38:587–596. [30] Peronius N, Sweeting TJ. On the correlation of minimum porosity with particle size distribution, Powder Technology 1985;42:113–121. [31] Bideau D, Troadec JP, Lemaitre J, Oger L. Physics of Finely Divided Matter, N. Boccara and M. Daoud (Eds.). Berlin: Springer-Verlag 1985. [32] Premkumar MK, Hunt WH, Sawtell RR. Aluminum composite materials for multichip modules. Journal of Metals 1992;44:24–28. [33] Lee JH, Lackey WJ, Benzel JF. Ternary packing of SiC and diamond particles in ethanol. Journal of Materials Research 1996;11:2804–2810. [34] Yu AB, Zou RP, Standish N. Modifying the linear packing model for predicting the porosity of nonspherical particle mixtures. Industrial Engineering Chemical Research 1996;35:3730–3741. [35] Zou RP,Yu AB, Lin XY. Packing of quaternary mixtures of fibrous particles. Journal of the American Chemical Society 1999;82:933–938.

References

|

33

[36] Standish N, Yu AB. Limitation of the simplex-centroid design for the porosity calculation of ternary mixtures. Powder Technology 1987;53:69–72. [37] Yu AB, Standish N. An analytical-parametric theory of the random packing of particles. Powder Technology 1988;55:171–186. [38] Yu AB, Zou RP, Standish N. Packing of ternary mixtures of nonspherical particles. Journal of the American Chemical Society 1992;75:2765–2772. [39] Molina JM, Saravanan RA, Arpon R, Narciso J, Garcia-Cordovilla C, Louis E. Pressure infiltration of liquid aluminium into packed SiC particulares with a bimodal size distribution. Acta Materialia 2002;50:247–257. [40] Finkers HJ, Hoffmann AC. Structural ratio for predicting the voidage of binary particle mixtures. Aiche Journal 1998;44:495–498. [41] Tavangar R, Molina JM, Weber L. Assessing predictive schemes for thermal conductivity against diamond-reinforced silver matrix composites at intermediate phase contrast. Scripta Materialia 2007;56:357–360. [42] Weber L, Dorn J, Mortensen A. On the electrical conductivity of metal matrix composites containing high volumen fractions of non-conducting inclusions. Acta Materialia 2003;51:3199–3211. [43] Weber L, Fischer C, Mortensen A. On the influence of the shape of randomly oriented, non-conducting inclusions in a conducting matrix on the effective electrical conductivity. Acta Materialia 2003;51:495–505. [44] Molina JM, Narciso J, Weber L, Mortensen A, Louis E. Thermal conductivity of Al-SiC composites with monomodal and bimodal particle distributions. Materials Science and Engineering A 2008;480:483–488. [45] Molina JM, Rhême M, Carron J, Weber L. Thermal conductivity of aluminium matrix composites reinforced with mixtures of diamond and SiC particles. Scripta Materialia 2008;58:393–396. [46] Weber L, Sinicco G, Molina JM. Influence of processing route on electrical and thermal conductivity of Al/SiC composites with bimodal particle distribution. Journal of Materials Science 2010;45:2203–2209. [47] Schapery RA. Thermal expansión coefficients of composite materials base don energy principles. Journal of Composite Materials 1968;2:380–404. [48] Kerner EH. The elastic and thermo-elastic properties of composite media. Proceedings of the Physical Society. Section B 1956;B69:808–813. [49] Hashin Z, Shtrickman S. On some variational principles in anisotropic and nonhomogeneous elasticity. Journal of the Mechanics and Physics of Solids 1962;10:335–342. [50] Narciso J, Weber L, Molina JM, Mortensen A, Louis E. Reactivity and thermal behaviour of Cu-Si/SiC composites: effects of SiC oxidation. Materials Science and Technology 2006;22:1464–1468. [51] Zhang L, Qu XH, Duan B, He X, Ren S, Qin M. Microstructure and thermo-mechanical properties of pressureless infiltrated SiCp/Cu composites. Composites Science and Technology 2008;68(13):2731–2738. [52] Narciso J, García-Cordovilla C, Louis E. Pressure infiltration of packed SiC and alumina particulates by liquid silver with disolved oxygen. Proceedings of the International Conference High Temperature Capillarity (ed. by Eustathopoulos E, Sobczak N), 29 June – 2 July 1997, Cracow, Poland. [53] Molina JM, Garcia-Cordovilla C, Louis E, Narciso J. Pressure infiltration of silver into compacts of oxidized SiC. Materials Science Forum 2003;426–432:2181–2186. [54] Bahraini M, Molina JM, Weber L, Mortensen A. Direct measurement of drainage curves in infiltration of SiC particle preforms. Materials Science and Engineering A 2008;495:203–207.

34 | 1 Metal matrix composites for thermal management [55] Molina JM, Tian JT, Garcia-Cordovilla C, Louis E, Narciso J. Wettability in pressure infiltration of SiC and oxidized SiC particle compacts by molten Al and Al-12wt.%Si alloy. Journal of Materials Research2007;22:2273–2278. [56] Molina JM, Arpon R, Saravanan RA, Garcia-Cordovilla C, Louis E, Narciso J. Threshold pressure for infiltration and particle specific surface area of particle compacts with bimodal size distributions. Scripta Materialia 2004;51:623–627. [57] Molina JM, Bahraini M, Weber L, Mortensen A. Direct measurement of drainage curves in infiltration of SiC particle preforms: influence of interfacial reactivity. Journal of Materials Science 2008;43:5061–5067. [58] Tian JT, Molina JM, Narciso J, Garcia-Cordovilla C, Louis E. Pressure infiltration of Al and Al-12wt.%Si alloy into compacts of SiC and oxidized SiC particles. Journal of Materials Science 2005;40:2537–2540. [59] Ren SB, He XB, Qu XH. Effect of controlled interfacial reaction on the microstructure and properties of the SiCp/Al composites prepared by pressureless infiltration. Journal of Alloys and Compounds 2008;455:423–431. [60] He XB, Qu XH, Ren SB, Jia C. Net-shape forming of composite packages with high thermal conductivity. Science in China Series E: Technological Sciences 2009;52:238–242. [61] Ren SB, He XB, Qu XH. Effect of Mg and Si in the aluminum on the thermo-mechanical propertis of pressureless infiltrated SiCp/Al composites. Composites Science and Technology 2007;67:2103–2113. [62] Molina JM, Prieto R, Narciso J, Louis E. The effect of porosity on the thermal conductivity of Al-12wt.%Si/SiC composites. Scripta Materialia 2009;60:582–585. [63] Arpon R, Molina JM, Saravanan RA, Garcia-Cordovilla C, Louis E, Narciso J. Thermal expansion behaviour of aluminium/SiC composites with bimodal particle distributions. Acta Materialia 2003;51:3145–3156. [64] Arpon R, Molina JM, Saravanan RA, Garcia-Cordovilla C, Louis E, Narciso J. Thermal expansion coefficient and thermal hysteresis of Al/SiC composites with bimodal particle distributions. Materials Science Forum 2003;426–432:2187–2192. [65] Molina JM, Piñero E, Narciso J, Garcia-Cordovilla C, Louis E. Liquid metal infiltration into compacts of ceramic particles with bimodal size distributions. Current Opinion in Solid State and Materials Science 2005;9:202–210. [66] Molina JM, Arpon R, Saravanan RA, Garcia-Cordovilla C, Louis E, Narciso J. Thermal expansion coefficient and wear performance of aluminium/SiC composites with bimodal particle distributions. Materials Science and Technology 2003;19:491–496. [67] Mantell CH. Carbon and Graphite Handbook.New York:Interscience;1968. [68] Orumwense FFO, Okorie BA, Okeakpu EO, Obiora EN, Onyeki LI. Sintered copper-graphite powder compacts for industrial applications. Powder Metallurgy 2001;44:62–66. [69] Emmer S, Bielek J, Havalda A. Copper-graphite composite-material for application to sliding electrical contacts. Journal of Physics IV 1993;3:1799–1804. [70] Etter T, Papakyriacou M, Schulz P, Uggowitzer PJ. Physical properties of graphite/aluminium composites produced by gas pressure infiltration method. Carbon 2003;41:1017–1024. [71] Rodriguez-Guerrero A, Sanchez SA, Narciso J, Louis E, Rodriguez-Reinoso F. Pressure infiltration of Al-12wt.%Si-X (X=Cu, Ti, Mg) alloys into graphite particle preforms. Acta Materialia 2006;54:1821–1831. [72] Oku T, Kurumada A, Sogabe AT, Oku T, Hiraoka T, Kuroda K. Effects of titanium impregnation on the thermal conductivity of carbon/copper composite materials. Journal of Nuclear Materials 1998;257:59–66.

References

|

35

[73] DeVincent SM, Michal GM. Improvement of thermal and mechanical properties of graphite/copper composites through interfacial modification. Journal of Materials Engineering and Performance 1993;2:323–331. [74] McCoy JW, Vrable DL. Metal-matrix composites from graphitic foams and copper. SAMPE Journal 2004;40:7–15. [75] Seong HG, Lopez HF, Robertson DP, Rohatgi PK. Interface structure in carbon and graphite fiber reinforced 2014 aluminum alloy processed with active fiber cooling. Materials Science and Engineering A 2008;487:201–209. [76] Prieto R, Molina JM, Narciso J, Louis E. Fabrication and properties of graphite flakes/metal composites for thermal management applications. Scripta Materialia 2008;59:11–14. [77] Molina JM, Rodriguez-Guerrero A, Bahraini M, Weber L, Narciso J, Rodriguez-Reinoso F, Louis E, Mortensen A. Infiltration of graphite preforms with Al-Si eutectic alloy and mercury. Scripta Materialia 2007;56:991–994. [78] Rodriguez-Guerrero A, Molina JM, Rodriguez-Reinoso F, Narciso J, Louis E. Pore filling in graphite particle compacts infiltrated with Al-12wt.%Si and Al-12wt.%Si-1wt.%Cu alloys. Materials Science and EngineeringA2008;495:276–281. [79] Etter T, Schulz P, Weber M, Metz J, Wimmler M, Löffler JF, Uggowitzer PJ. Aluminium carbide formation in interpenetrating graphite/aluminium composites. Materials Science and Engineering A 2007;448:1–6. [80] Korab J, Stefanik P, Kavecky S, Sebo P, Korb G. Thermal conductivity of unidirectional copper matrix carbon fiber composites. Composites: Part A 2002;33:577–581. [81] Stoessel CH, Withers JC, Pan C, Wallace D, Loutfy RO. Improved hollow cathode magnetrón deposition for producing high thermal conductivity graphite-copper composite. Surface Coatings Technology 1995;76–77:640–644. [82] Aryasomayajula L, Wolter KJ. Carbon nanotube composites for electronic packaging applications: a review. Journal of Nanotechnology 2013;2013:1–6. [83] Bakshi SR, Lahiri D, Agarwal A. Carbon nanotube reinforced metal matrix composites – a review. International Materials Reviews 2010;55:41–64. [84] Chu K, Guo H, Jia C, Yin F, Zhang X, Liang X, Chen H. Thermal properties of carbon nanotube-copper composites for thermal management applications. Nanoscale Research Letters 2010;5:868–874. [85] Narciso J, Prieto R, Molina JM, Louis E. Producción de Materiales Compuestos con Alta Conductividad Térmica, Patente de invención, Oficina Española de Patentes y Marcas. Application number: P002700804 2007. [86] Narciso J, Prieto R, Molina JM, Louis E. Three phase composite material with high thermal conductivity and its production, European Application Patent, Application number: EP2130932-B1; 2007. [87] Narciso J, Prieto R, Molina JM, Louis E. Production of Composite Materials with High Thermal Conductivity, US Application Patent, Application number: US 20100143690-A1;2010. [88] Prieto R, Molina JM, Narciso J, Louis E. Thermal conductivity of graphite flakes-SiC particles/metal composites. Composites: Part A 2011;42:1970–1977. [89] Prieto R, Louis E, Molina JM. Fabrication of mesophase pitch-derived open-pore carbon foams by replication processing. Carbon 2012;50:1904–1912. [90] Gallego NC, Klett JW. Carbon foams for thermal management. Carbon 2003;41:1461–1466. [91] Queipo P, Granda M, Santamaría R, Menéndez R. Preparation of pitch-based carbon-copper composites for electrical applications. Fuel 2003;83:1625–1634. [92] Klett J, Hardy R, Romine E, Walls C, Burchell T. High-thermal-conductivity, mesophase pitch-derived carbon foams: effect of precursor on structure and properties. Carbon 2000;38:953–973.

36 | 1 Metal matrix composites for thermal management [93] Prieto R, Narciso J, Bahraini M, Molina JM, Weber L, Louis E. Graphite foam infiltrated with metal alloys. Electronic programme and extended abstracts – The international carbon conference – Carbon 2006. [94] Kidalov SV, Shakhov FM. Thermal conductivity of diamond composites. Materials 2009;2:2467–2495. [95] Schelling PK, Shi L, Goodson KE. Managing heat for electronics. Materials Today 2005;8:30–35. [96] Weber L, Tavangar R. Diamond-based metal matrix composites for thermal management made by liquid metal infiltration-potential and limits. Advanced Materials Research 2009;59:111–115. [97] Weber L, Tavangar R. On the influence of active element content on the thermal conductivity and thermal expansion of Cu-X (X = Cr, B) diamond composites. Scripta Materialia 2007;57:988–991. [98] Chen C, Guo H, Chu K, Yin F, Zhang X, Han Y, Fan Y. Thermal conductivity of diamond/copper composites with a bimodal distribution of diamond particle sizes prepared by pressure infiltration method. Rare Metals 2011;30:408–413. [99] Nadler JH, Ledford KA, Harris HM, Wagner K. Ag/diamond composite shims for HPA thermal management. CS MANTECH Conference, May 13th-16th, 2013, New Orleans, Louisiana, USA. [100] Davis LC, Artz BE. Thermal conductivity of metal matrix composites. Journal of Applied Physics 1995;77:4954–4960. [101] Xue C, Yu JK, Zhu XM. Thermal properties of diamond/SiC/Al composites with high volume fractions. Materials Design 2011;32:4225–4229. [102] Han Y, Guo H, Yin F, Zhang X, Chu K, Fan Y. Microstructure and thermal conductivity of copper matrix composites reinforced with mixtures of diamond and SiC particles. Rare Metals 2012;31,58–63. [103] Beffort O, Khalid FA, Weber L, Ruch P, Klotz UE, Meier S, Kleiner S. Interface formation in infiltrated Al(Si)/diamond composites. Diamond and Related Materials 2006;15:1250–60. [104] Kleiner S, Khalid FA, Ruch PW, Meier S, Beffort O. Effect of diamond crystallographic orientation on dissolution and carbide formation in contact with liquid aluminium. Scripta Materialia 2006;55:291–294. [105] Ruch PW, Beffort O, Kleiner S, Weber L, Uggowitzer PJ. Selective interfacial bonding in Al(Si)-diamond composites and its effect on thermal conductivity. Composites Science and Technology 2006;66:2677–2685. [106] Tan Z, Li Z, Fan G, Kai X, Ji G, Zhang L, Zhang D. Fabrication of diamond/aluminum composites by vacuum hot pressing: Process optimization and thermal properties. Composites: Part B 2013;47:173–80. [107] Mizuuchi K, Inoue K, Agari Y, Morisada Y, Sugioka M, Tanaka M, Takeuchi T, Tani J, Kawahara M, Makino Y. Processing of diamond particle dispersed aluminum matrix composites in continuous solid-liquid co-existent state by SPS and their thermal properties. Composites: Part B 2011;42,825–831. [108] Mizuuchi K, Inoue K, Agari Y, Morisada Y, Sugioka M, Tanaka M, Takeuchi T, Tani J, Kawahara M, Makino Y. Thermal conductivity of diamond particle dispersed aluminum matrix composites fabricated in solid-liquid co-existent state by SPS. Composites: Part B 2011;42:1029–1034. [109] Collins KC, Chen S, Chen G. Effects of surface chemistry on thermal conductance at aluminum-diamond interfaces. Applied Physics Letters 2010;97:083102_1–083102_3. [110] Abyzov AM, Kidalov SV, Shakhov FM. High thermal conductivity composite of diamond particles with tungsten coating in a copper matrix for heat sink application. Applied Thermal Engineering 2012;48:72–80.

References

| 37

[111] Zhang Y, Zhang HL, Wu JH, Wang XT. Enhanced thermal conductivity in copper matrix composites reinforced with titanium-coated diamond particles. Scripta Materialia 2011;65:1097–1100. [112] Chu K, Liu Z, Jia C, Chen H, Liang X, Gao W, Tian W, Guo H. Thermal conductivity of SPS consolidated Cu/diamond composites with Cr-coated diamond particles. Journal of Alloys and Compounds 2010;490:453–458. [113] Monje IE, Louis E, Molina JM. Optimizing thermal conductivity in gas-pressure infiltrated aluminum/diamond composites by precise processing control. Composites: Part A 2013;48:9–14. [114] Monje IE, Louis E, Molina JM. Aluminum/diamond composites: a preparative method to characterize reactivity and selectivity at the interface. Scripta Materialia 2012;66:789–792.

Nuno Silvestre

2 Recent research and developments on the mechanical behavior of CNT-reinforced metal matrix composites Abstract: The unceasing upgrade of techniques and processes to fabricate high purity carbon nanotubes (CNTs) and the improvement of the available techniques to produce high performance matrix materials have fostered the way to enhance composite materials and their properties, either mechanical, thermal, electrical, optical or magnetic. CNTs reinforcements have been introduced into polymers, ceramics, cementbased materials and metals. Polymers were the first material to be exploited as matrix material reinforced by CNTs. Up to now other materials have tentatively been investigated for that purpose, including metals. Today, many applications of CNT-reinforced composites exist, but CNT-reinforced metals are still scarce and only found in very specific applications. Several reasons can be identified, but the increasing demand for lighter and stronger metals paved the way to more fundamental research on the topic of CNT-reinforced metal matrix composites (MMCs). This review describes the state-of-the art in this field and highlights the excellent and promising mechanical properties of CNT-reinforced MMCs.

2.1 Introduction Since their discovery by Ijiima [1], carbon nanotubes (CNTs) have been considered as an ideal reinforced material to improve the mechanical performance of many materials. CNTs are promising reinforcements for lightweight and high-strength composites. This is due to their exceptional small diameters and high Young’s modulus, high tensile strength and chemical stability. Still, the main obstacle is to obtain a homogenous dispersion of the CNTs in the desired material matrix. Several methods and processes have been developed to improve the dispersion of CNTs in polymer matrices. The state-of-the-art report made by Andrews and Weisenberger [2] on CNT-polymer composites emphasized the problem of CNT dispersion. These authors focused particularly on interfacial bonding between CNTs and polymer matrices, as well as to potential topics of interest at that time. More recently, Chou et al. [3] examined the recent advancements in the science and technology of CNT-based fibers for polymer composites. Their assessment was made according to the hierarchical structural levels of CNTs used in composites, ranging from 1-D to 2-D to 3-D. At the 1-D level, fibers composed of pure CNTs or CNTs embedded in a polymeric matrix produced by various techniques were reviewed. At the 2-D level, the focus was on CNT-modified advanced fibers, CNT-

40 | 2 CNT-reinforced MMCs modified interlaminar surfaces and highly oriented CNTs in planar form. At the 3-D level, they examined the mechanical and physical properties CNT-polymer composites, CNT-based damage sensing, and textile assemblies of CNTs. The review by Yan et al. [4] mainly focused on functional CNTs and their applications in property enhancement of various polymer composites. First, the general methods for CNT preparation were briefly introduced. Second, the functionalization of CNTs, particularly via chemical approaches, was summarized and the application of these functionalized CNTs was discussed. Finally, the interaction of CNTs with various polymers, the formation of CNT-polymer composites and their property and applications were discussed by Yan et al. [4]. Currently, CNT-reinforced polymer matrix composites are being intensively investigated. However, scarce to moderate research has been conducted to improve the CNT dispersion in metal matrices. Despite the fact that CNTs are an effective reinforcement to improve the mechanical and thermal responses of metal matrix composite (MMCs), very few successful attempts have been made at commercial applications due to the difficulties of incorporating CNTs in metals. It became obvious that segregation of CNTs due to their strong van der Waals forces often produces material defects, which decreases the material’s properties. Despite CNT-reinforced MMCs having received the least attention, they are being considered for use in structural applications because of their high specific strength as well as functional materials for their exciting thermal and electrical characteristics. Light MMCs are of great interest due to their potential for reducing CO2 emission in lightweight design. Using CNT’s exceptional properties in MMCs for macroscopic applications still constitutes a big challenge for the scientific and technological communities. The observation of Figure 2.1 makes it possible to draw some conclusions about the research evolution in the field of CNT-reinforced composites. It shows the variation in the number of publications in SCI Web of Knowledge per year since 2000. Each column corresponds to a different combination of keywords in a searched topic: (i) the blue column corresponds to “Nanotube and Composite”; (ii) the green column corresponds to “Nanotube and Composite and Polymer”; and (iii) the red column corresponds to “Nanotube and Composite and Metal”. It is easily seen that the works on CNT-based composites rapidly increased from 81 in 2000 to over 7703 in 2012. The rate of increase steadily grew from 2000 to 2007, but from 2007 to 2012 it showed a huge rise mainly due to the advent of nanomaterials. Since 2007, several types of materials fabricated at the nanoscale emerged as promising matrices and nanodevices to which CNTs can effectively be added. Figure 2.1 also shows that the number of publications indexed in ISI Web of Knowledge [5] and focused on CNT-Polymer composites also increased substantially from 2000 with 31 publications to 2012 with 1593. Regarding the number of publications on CNT-Metal composites, it has also increased since 2000 but to a minor extent: 19 publications in 2000 and 468 in 2012. The data shown in Figures 2.2 and 2.3 is related to the year 2013 and its 613 publications with keywords “Nanotube and Composite and Metal”. Regarding Figure 2.2, it shows the percentage

2.1 Introduction

|

41

10000 Nanotube and composite Nanotube and composite and polymer Nanotube and composite and metal

9000

Number of publications

8000 7000 6000 5000 4000 3000 2000 1000

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

0

Year Fig. 2.1. Evolution of the number of publications indexed in the Web of Knowledge per year (period 2000–2013).

4

2 2 2 22

4

Materials Science Chemistry Physics

11

Engineering Science Technology other Topics Energy Fuels Electrochemistry Polymer Science Metallurgy Metallurgical Engineering Instruments Instrumentation

18 21

16 Fig. 2.2. Percentage distribution of the 2013 publications on CNT-metal matrix composites among different “Research Areas” indexed in the ISI Web of Knowledge.

42 | 2 CNT-reinforced MMCs distribution of the publications among the eight most representative “Research Areas” indexed in ISI Web of Knowledge: – Materials Science; – Chemistry; – Physics; – Engineering; – Science Technology Other Topics; – Energy Fuels; – Electrochemistry; – Polymer Science; – Metallurgy and Metallurgical Engineering; – Instruments and Instrumentation. It is easily observed that four major areas exist and correspond to 77 % of the papers published in this subject: materials science (22 %), chemistry (21 %), physics (16 %) and engineering (18 %). It is also interesting to note that the percentages of these “Research Areas” are similar (nearly 20 %). The remaining 23 % are divided in other more specific “Research Areas”, such as energy fuels, electrochemistry, polymer science, metallurgy and metallurgical engineering, and instruments and instrumentation. Figure 2.3 shows the distribution of the 247 papers published in the “Research Area” of “Engineering” per source title (international journal). Only one journal has more than 20 papers published (Journal of Materials Chemistry A), and two journals have more than 10 papers published (Chemical Engineering Journal; Applied Surface Science). About 170 publications (out of 247) are distributed among a large number of sources 28 10

Journal of Materials Chemistry A 12 8 7 6 4 4 2

166

Chemical Engineering Journal Applied Surface Science Advanced Functional Materials Science of Advanced Materials Journal of Nanomaterials Composites Part B Engineering Composites Science and Technology Composites Part A-Applied Science and Manufacturing Other Journals

Fig. 2.3. Distribution of 2013 publications on CNT-metal matrix composites (in “Engineering”) among different “Source Title” (journal) indexed in ISI Web of Knowledge.

2.2 CNT-Al composites

|

43

with less than 3 papers per journal. This evidence shows that the subject of CNT-Metal Matrix Composites is very wide, open and general; it is positioned on the frontier of several scientific areas of research and doesn’t belong to a specific technological domain. Since the review by Girot et al. [6], few reviews on CNT-reinforced MMCs have been published. In 2004, Curtin et al. [7] reviewed research on the incorporation of CNTs into ceramic and metal matrices to form composite structures. They emphasized the processing methods, mechanical performance and prospects for successful applications. They reviewed the literature on the topics of fabrication and properties of ceramic and metal matrix systems. In the case of metal matrix materials, they mentioned that enhanced mechanical properties (stiffness, wear and fatigue resistance) are desirable, but a wider range of properties (electrical, magnetic and vibrational) were also investigated. Li et al. [8] studied the behavior of CNT-reinforced light metal composites produced by melt stirring and by high-pressure die casting. The light metal composites showed significantly improved mechanical properties already at small CNT contents. Li et al. [8] also studied the influence of CNT concentration on the composites. The review by Neubauer et al. [9] gives an overview and summarizes the activities related to CNTs and carbon nanofibers used as a reinforcement in metallic matrix materials. They presented the main challenges and the potential with respect to material properties. Bakshi et al. [10] reviewed and summarized the research work carried out in the field of CNT-reinforced MMCs. They focused on the critical issues of CNT-reinforced MMCs that include processing techniques, nanotube dispersion, interface, strengthening mechanisms and mechanical properties. Processing techniques used for synthesis of the composites were critically reviewed with an objective to achieve homogeneous distribution of CNTs in the matrix. The mechanical property improvements achieved by addition of CNTs in various metal matrix systems were summarized by Bakshi et al. [10]. Several factors, such as the structural and chemical stability of CNTs in different metal matrices and the importance of the CNT-metal interface were described. The relevance of CNT dispersion and its quantification was also highlighted. The purpose of this review is two-fold: (i) to update the state-of-the-art concerning the use of CNTs for reinforcement of MMCs; and (ii) to explain the main developments in the mechanical behavior of CNT-reinforced MMCs. The recent developments and trends in research are shown separately by each type of material matrix, in alphabetic order: Aluminium (Al), Cobalt (Co), Copper (Cu), Iron (Fe), Magnesium (Mg), Nickel (Ni) and Titanium (Ti).

2.2 CNT-Al composites Aluminium (Al) is a silvery white, soft, ductile metal. Al is the third most abundant element (after oxygen and silicon), and the most abundant metal in the Earth’s crust. Al is remarkable for its low density and for its ability to resist corrosion due to the

44 | 2 CNT-reinforced MMCs phenomenon of passivation. Structural components made from Al and its alloys are vital to the aerospace industry and are important in other areas of transportation and structural materials. The most useful compounds of Al, at least on a weight basis, are oxides and sulfates. Al matrix is widely used for CNT-reinforced MMCs. Since the work by Zhong et al. [11], many achievements have been made in the development of CNT-Al composites. In this review, 24 papers are described and dedicated to this topic [12–35]. Cha et al. [12] proposed a novel process to fabricate CNT-alumina nanocomposites, consisting of a molecular level mixing process and an in situ spark plasma sintering process. The CNT/alumina nanocomposites fabricated by this proposed process showed improved hardness due to a load transfer mechanism of the CNTs and increased fracture toughness arising from the bridging mechanism of CNTs during crack propagation. In order to produce optimized composites, George et al. [13] studied the strength of CNT-Al composites and investigated the relevant strengthening mechanisms involved in CNT-Al composites. Three major mechanisms were analyzed along with experimental procedures for making CNT/Al composites. Using cold isostatic press and subsequent hot extrusion techniques, Deng et al. [14] fabricated a 1.0 wt.% carbon nanotube (CNT) reinforced 2024Al matrix composite, measured the mechanical properties of the composite by tensile tests and examined the fracture surfaces using field emission scanning electron microscopy. Their experimental results showed that CNTs are dispersed homogeneously in the composite, and that the interfaces of the Al matrix and the CNT bonded well. Deng et al. [15] observed that the tensile strength and the Young’s modulus of the composite were enhanced markedly, and the elongation didn’t decrease when compared to the matrix material fabricated under the same process. The extraordinary mechanical properties of CNTs and their pulling-out role in the Al matrix composite are factors that explain this improved behavior. Deng et al. [15] also investigated the microstructure characteristics and the distribution of CNTs in the Al matrix (see Figures 2.4 and 2.5). They showed that by adding a small amount of CNTs to the matrix, the elastic modulus and the tensile strength were increased greatly with respect to those of the 2024Al base material. Deng et al. [16] investigated the damping behavior of 2024Al reinforced with multi-walled CNTs. Figure 2.6 shows an SEM image of the fracture surface of 1.0 wt.% CNT/2024Al composite. It is seen (Figure 2.6a) that CNTs are distributed uniformly within the Al

200 nm

200 nm (a)

(b)

Fig. 2.4. TEM image of (a) raw, and (b) dispersed CNTs [15].

2.2 CNT-Al composites

SE

WD20.1mm

5.0OkV

x500

SE

100um

WD14.3mm

20.OkV

x500

|

45

100um

(b)

(a)

dof 012 6.OkV 12.3mm x 10.0k SF(M)

6.00um

Fig. 2.5. SEM micrograph of (a) the mixtures of CNTs and 2024Al alloy powders and (b) the local area magnified in (a) [15].

(c) (a)

(b)

CNTs

CNTs

dcf-044 15.OkV 11.6mm x20.Ok SE(M)

2.00um

dcf-016 15.OkV 11.8mm x50.Ok SE(M)

1.00um

Fig. 2.6. SEM image of the fracture surface of the composite with (a) low magnification and (b) local magnification in (a) [16].

matrix. The damping characteristics of the composite were investigated with frequencies of 0.5, 1.0, 5.0, 10, 30 Hz, at a temperature of 25–400 ∘ C. Figure 2.7 shows the DMA (dynamic mechanical analyzer) storage modulus curves of the composite. It can be observed that the composite’s storage module significantly decreases with increasing temperature. The experimental results showed that the frequency significantly affects the damping capacity of the composite when the temperature is above 230 ∘ C. If the temperature is above 230 ∘ C, the higher the frequency is, the bigger the storage modulus is. The damping capacity of the composite with a frequency of 0.5 Hz reached 975 × 10−3 , and the storage modulus is 82.3 GPa when the temperature is 400 ∘ C. This study proved that CNTs can greatly improve the damping properties of MMCs at elevated temperature without sacrificing their mechanical strength and stiffness.

46 | 2 CNT-reinforced MMCs 100 98

Storage modulus (GPa)

96 94 92 90 88

0.5Hz 1.0Hz 5.0Hz 10Hz 30Hz

86 84 82 0

50

100

150 200 250 300 Temperature (°C)

350

400

Fig. 2.7. DMA storage modulus curves of the composite [16].

Deng et al. [17] studied the thermal expansion behavior of Al composite reinforced with CNTs. The coefficient of thermal expansion of Al matrix composite reinforced with 1.0 wt.% multi-wall CNTs fabricated by a cold isostatic pressing and hot squeeze technique was measured between 25 and 400 ∘ C with a high-precision thermomechanical analyzer and then compared with those of pure Al and 2024Al matrix fabricated under the same processing. The results by Deng et al. [17] showed that the coefficient of thermal expansion (CTE) of the composite reduces in relation to those of pure Al and 2024Al matrix due to the introduction of CNTs (see Figure 2.8). The CTE of pure Al and 2024Al matrix increase significantly with increasing temperature, however the change of CTE of the composite is greatly slowed. The addition of 1.0 wt.% CNTs to 2024Al matrix decreased the coefficient of thermal expansion by as much as 12 % and 11 % compared with those of pure Al and 2024Al matrix at 50 ∘ C, respectively. This evidence indicated that CNT reinforcement of MMC may be a promising material with a low coefficient of thermal expansion. Pure Al 2024Al 1.0wt% CNT/2024Al

CTE (x 10–6/K)

33 30 27 24 21 0

100

200 300 Temperature (°C)

400

Fig. 2.8. Variation of CTE with temperature for 1.0 wt.%CNT/2024Al, pure Al and 2024Al [17].

2.2 CNT-Al composites

|

47

Bakshi et al. [18] prepared multi-walled CNT-reinforced Al coatings using cold gas kinetic spraying in order to obtain a good CNT dispersion in micron-sized gas atomized Al-Si eutectic powders. Spray dried powders containing 5 wt.% CNT were blended with pure Al powder to give overall nominal CNT compositions of 0.5 wt.% and 1 wt.%, respectively. They showed that the CNTs were uniformly distributed in the matrix. Nanoindentation led to elastic modulus values between 40 and 229 GPa, and this high scatter was attributed to microstructural heterogeneity of the coatings (including pure Al, Al-Si eutectic, porosity and CNTs). Using hot extrusion of ball-milled powders, Choi et al. [19] fabricated Al matrix composite rods in which tightly bonded multi-walled CNTs were separately dispersed and uniaxially aligned. They showed that the reinforcing efficiency of CNTs in the composites followed the volume fraction rule of discontinuous fibers in the grain size range down to 70 nm. Pérez-Bustamante et al. [20] produced Al-based nanocomposites reinforced with multi-walled CNTs using mechanical milling followed by sintering without pressure at 823 K under vacuum. The interface between Al matrix and the multi-walled CNTs was examined using transmission electron microscopy. This work showed that the multi-walled CNTs were not damaged during the preparation of the nanocomposite and that no reaction products were detected after sintering. The mechanical properties of sintered nanocomposite specimens were evaluated by a compression test. The yield stress and maximum strength obtained were considerably higher than those reported in the literature for pure Al prepared by the same route. The values for yield stress and ultimate strength increased about 100 % as the volume fraction of multi-walled CNTs increased from 0 to 0.75 wt.%, for 2 h of milling time. They concluded that the milling time and the concentration of CNTs had an important effect on the mechanical properties of the nanocomposite. In the paper by Esawi and El Borady [21], a powder rolling technique is used to fabricate CNT-reinforced Al strips. The Al-CNT mixtures were blended in either a mixershaker at a rotary speed of 46 rpm, or under argon in a planetary mill at a rotary speed of 300 rpm, prior to rolling. The CNT dispersion was shown to be better under the higher energy planetary action. The strength of the rolled strips was evaluated for various wt.% CNT samples. The Al-0.5 wt.% composite strips exhibited enhanced mechanical properties. The CNT-reinforced Al strips were shown to have numerous attractive applications in the aerospace, automotive and electronics industries. Later, Esawi et al. [22] used planetary ball milling to disperse 2 wt.% multi-walled CNT in Al powder. Despite the success of ball milling in dispersing CNTs in Al powder, it is often accompanied by considerable strain hardening of the Al powder, a fact that might have far reaching implications on the composite’s final properties. Thus, both un-annealed and annealed Al-2 wt.% CNT composites were investigated by Esawi et al. [22]. They found that ball-milled and extruded (un-annealed) samples of Al-2 wt.% CNT demonstrated high notch sensitivity and consistently fractured outside the gauge length during tensile testing. In contrast, extruded samples annealed at 400 and at 500 ∘ C for 10 h prior to testing, exhibited more ductile behavior and no notch sensitivity. Under

48 | 2 CNT-reinforced MMCs similar processing conditions, they showed that ball milling for 3 h followed by hot extrusion and annealing at 500 ∘ C resulted in enhancements of around 21 % in tensile strength compared with pure Al with the same process history. They argued that ball-milling conditions resulted in the creation of a nanostructure in all samples produced, a fact that was proved by means of XRD and TEM analysis. The tensile testing fracture surfaces identified by Esawi et al. [22] showed uniform dispersion and alignment of the CNTs in the Al matrix, but also showed that CNTs acted as nucleation sites for void formation during tensile testing. This effect contributed to the observation of CNT pullout due to the poor bond between the CNTs and the matrix. Since the uniform dispersion of CNTs in the Al matrix has been identified as being critical to the pursuit of enhanced properties, ball milling as a mechanical dispersion technique has proved its potential. Esawi et al. [23] used ball milling to disperse up to 5 wt.% CNT in Al matrix and investigated the effect of CNT content on the composites’ mechanical properties. Cold compaction and hot extrusion were used to consolidate the ball-milled Al-CNT mixtures. In comparison to pure Al, a 50 % increase in tensile strength and 23 % increase in stiffness were observed by Esawi et al. [23]. Some carbide formation was observed in the composite containing 5 wt.% CNT. The large aspect ratio of CNTs used by Esawi et al. [23] led to difficulties in CNT dispersion at CNT wt.% greater than 2. Therefore, the expected improvements in mechanical properties with increase in CNT weight content were not fully realized by these authors. Choi et al. [24] reported a study on the mechanical properties and wear characteristics of ultrafine-grained Al and Al-based composites. In this composite, well dispersed and Al atom-infiltrated multi-walled CNTs formed a strong interface with the matrix by mechanical interlocking. Wear characteristics, varied according to the grain size and the CNT volume, were evaluated by Choi et al. [24] under several combinations of applied load and sliding speed. They reported that strength and wear resistance were significantly enhanced and the coefficient of friction was extremely reduced for a grain size decrease and CNT volume increase. The ultrafine-grained composite containing 4.5 vol.% of CNTs exhibited more than 600 MPa in yield stress and less than 0.1 in the coefficient of friction. Both coefficients of friction and wear rate grew with increasing load, while they were reduced with increasing the sliding speed. This study demonstrated that CNTs are effective reinforcement for enhancing wear characteristics as well as mechanical properties. Choi et al. [25] produced nanocrystalline Al-Si alloy-based composites containing CNTs using hot rolling ball-milled powders. The grain size was effectively reduced and the Si element was dissolved in the Al matrix during the milling process. Using a thermo-mechanical process, CNTs were gradually dispersed into the Al powders. The composite produced by Choi et al. [25] contained 3 vol.% of CNTs and showed yield strength about 520 MPa and plastic elongation to failure of 5 %. Al and Al-based composites containing 4.5 vol.% multiwalled CNTs were fabricated by Choi and Bae [26] using hot-rolling the ball-milled powder. They studied the composite creep behavior at 523 K and it displayed much enhanced creep resistance at the applied stresses higher than 200 MPa. Figure 2.9a

2.2 CNT-Al composites

1.2

250

UFG AI/MWCNT composites 1.0

200

Strain

Stress (MPa)

135 MPa 0.8

125 MPa

0.6

110 MPa 100 MPa

0.4

85 MPa 0.2 65 MPa

0.0 0.0

5

5.0x10

(a) 0

log ε

–2

6

1.0x10 Time (s)

1.5x10

–3 –4 n ~ 12

–5 n ~ 0.05

n~3 523 K

–7 1.7 (c)

6

AI-6Mg-1 Sc-1Zr/SiCp 10 vol% [32] UFG AI UFG AI/MWCNT composites

–6

1.8

150

1.5x10–3 5x10–4 5x10–4 1x10–4 –5

1x10

3.5x10–3 5x10–4

UFG AI/MWCNT composites

100 50

523 K

–1

| 49

1.9

2.0 2.1 log σ

2.2 2.3 2.4

UFG AI 523 K 0 0.0 0.02 0.04 0.06 0.08 0.010 0.012 0.014 (b) Strain

Fig. 2.9. (a) Strain–time curve during a step load creep test; (b) stress–strain curves during step strain rate tests; and (c) creep properties with double logarithmic plots of stress as a function of minimum creep rate (obtained from step load creep tests by power-law equation) or strain rate (obtained from step strain rate tests) for UFG aluminum, UFC Al/CNT composite, and Al–6Mg– 1Sc–1Zr/SiC 10 vol.% composite in the literature [26].

shows a typical step load creep curve of the Al/CNT composites. The secondary creep stage is apparently observed under every applied stress. Figure 2.9b shows typical step strain rate curves of the aluminum and the composite. Double logarithmic plots of stress as a function of creep rate or strain rate for the UFG aluminum and the UFC Al/MWCNT composite together with Al–6Mg–1Sc–1Zr/SiC 10 vol.% composite in literature are shown in Figure 2.9c. Choi and Bae [26] also identified a negligible dependency of strain rate on stress for a relatively low applied stress region (stresses below 110 MPa). They explained this behavior due to the fact that diffusional flow of the matrix is significantly restricted by multi-walled CNTs. Figure 2.10 reveals mappings representing the axial normal stresses in the matrix for ultrafine-grained aluminum and Al/CNT composite with different applied stresses (65 and 100 MPa). The monolithic aluminum shows a uniform stress level over the specimen whereas the flow stress of the composite is significantly concentrated at the tips of CNTs. Singhal et al. [27] report the fabrication of Al-matrix composites reinforced with amino-functionalized CNTs by means of a powder metallurgy process. Functionalization of the CNTs was carried out by ball milling multi-walled CNTs in the presence of ammonium bicarbonate. They found that the mechanical properties of Alfunctionalized CNT composites were much superior to the composites fabricated using non-functionalized or acid functionalized CNTs. Using high-resolution transmission electron microscopy, they attributed the improvement of mechanical properties

50 | 2 CNT-reinforced MMCs MWCNTs S, S22 (Avg: 75%) +2.000e+02 +1.871e+02 +1.742e+02 +1.613e+02 +1.483e+02 +1.354e+02 +1.225e+02 +1.096e+02 +9.667e+01 +8.375e+01 +7.083e+01 +5.792e+01 +4.500e+01

σappl = 65 MPa (a)

σappl = 100 MPa (b)

σappl = 65 MPa (c)

σappl = 100 MPa (d)

Fig. 2.10. Mappings representing the axial normal stresses in the matrix for (a) monolithic aluminum with an applied stress of 65 MPa; (b) monolithic aluminum with an applied stress of 100 MPa; (c) the Al/CNT composite with an applied stress of 65 MPa; and (d) the Al/CNT composite with an applied stress of 100 MPa [26].

to (i) homogeneous dispersion of functionalized CNTs in Al matrix, as compared to non-functionalized or acid functionalized CNTs, and (ii) the formation of strong interfacial bonding between the functionalized CNTs and Al matrix, leading to efficient load transfer from Al matrix to functionalized CNTs. Using measurements of grain size and mechanical property changes upon annealing at various temperatures, Lipecka et al. [28] evaluated the effect of CNTs on the thermal stability of ultrafine-grained Al alloy processed by the consolidation of nano-powders obtained by mechanical alloying. They found that the grain size of the samples containing CNTs is stable up to high temperatures, and even after annealing at 450 ∘ C, no evident grain growth was observed. This was attributed to the presence of entangled networks of CNTs located at grain boundaries and to the formation of nanoscale particles of carbide Al4 C3 . They also revealed that CNTs decompose at a relatively low temperature (450 ∘ C) and form fine Al4 C3 precipitates. Lipecka et al. [28] revealed that this transformation did not affect the mechanical properties due to the carbide nanoscale size. Choi et al. [29] investigated the influence of multi-walled CNTs on the strength of Al-based composites with grain sizes ranging from 250 to 65 nm. They employed four fabrication routes to obtain a variation of grain sizes. Group A represents nanograined (the smallest grain size in this study) composites. Group B also has nanosized grains, but less small. Group C has an intermediate grain size. Group D has the largest grain size. True stress–true strain curves of the specimens in groups A–D are

2.2 CNT-Al composites

Strain

0.35

0.30

0.25

0.20

0.15

4.5 vol.% 3.0 vol.%

500 400 300

1.5 vol.%

200

Pure Al Pure Al (Unmilled)

100 0

0.00

0.35

0.30

0.25

0.15

0.10

0.05

0.00

(c)

0.20

Pure Al (Unmilled)

100

Strain rate – 1x10–4/s Group D

(d)

0.30

Pure Al

7.5 vol.% 6.0 vol.%

0.35

600

200

0

Strain 700

1.5 vol.%

300

Pure Al (Unmilled)

0.00

0.35

0.30

Strain rate – 1x10–4/s Group C

3.0 vol.%

400

100

0.25

500

4.5 vol.%

Pure Al

0

Stress (MPa)

Stress (MPa)

600

200

(b)

Strain 700

1.5 vol.%

300

0.20

(a)

0.25

0.15

0.10

0.05

0.00

0

0.20

Pure Al (Unmilled)

100

3 vol.%

0.10

200

400

0.10

300

500

0.05

400

600

3 vol.% 1.5 vol.% 1.0 vol.% Pure Al

0.05

500

Stress (MPa)

Stress (MPa)

600

Strain rate – 1x10–4/s Group B

700

0.15

Strain rate – 1x10–4/s Group A

700

| 51

Strain

Fig. 2.11. Tensile true stress–true strain curves of the specimens [29].

shown in Figure 2.11a–d, respectively. Tensile curve of the un-milled monolithic aluminum is also included in Figure 2.11 for comparison. As grain size of the matrix is reduced, yield strength is significantly enhanced and work hardening capacity is reduced, exhibiting early plastic instability. This trend is significant for nano-grained aluminum of groups A and B. As the content of CNTs increases, Young’s modulus and yield strength generally increase and ductility decreases. Choi et al. [29] found that the composite strength was significantly enhanced by the increase of the CNT volume. The strengthening efficiency of CNTs in ultrafine-grained composites was comparable with that predicted by the discontinuous fiber model. For grain size below 70 nm, the efficiency was half of the theoretical prediction. The deteriorated yield stress of the nano-grained aluminum matrix composites may come from the different deformation mode of the matrix. For coarse- or ultrafine-grained aluminum, dislocations are accumulated or tangled during plastic deformation, leading to a large strain field near CNTs as shown in Figure 2.12a. Dislocations located at grain boundaries move across the grains without any resistance and then reside on the opposite grain boundaries, producing a weak strain field within the nanoscale grains as shown in Figure 2.12b.

52 | 2 CNT-reinforced MMCs

(a)

σ∫

(b)

τ τmy

τ τmy (c)

σ∫

(d)

Fig. 2.12. A depiction of plastic strain fields around CNT for (a) ultrafine-grained and (b) nanograined aluminum matrix composites. The stress distribution along CNTs for (c) ultrafine-grained and (d) nano-grained aluminum matrix composites [29].

For nano-grained Al, Choi et al. [29] found that the activities of forest dislocations diminished and dislocations emitted from grain boundaries were dynamically annihilated during the recovery process. This provided a weak plastic strain field around CNTs. Their observation gave a basic understanding of the strengthening behavior of nano-grained MMCs. Jiang et al. [30] used flake powder metallurgy (flake PM) to achieve uniform distribution of CNTs in CNT-Al composites and realized the potential of CNTs for reinforcement. The structural integrity of the CNTs was maintained in the composites, since CNTs were protected from high-energy physics forces such as ball milling. They concluded that a strong and ductile CNT-Al composite might be fabricated using this process, leading to a tensile strength of 435 MPa and ultimate strain of 6 %. These values greatly surpassed the strength values for materials produced by conventional methods. In order to combine high tensile strength and ductile behavior, CNT/Al nanolaminated composites with alternating layers of Al (400 nm) and CNTs (50 nm) were fabricated by Jiang et al. [31] using flake PM. Compared with conventional homogeneous nanocomposites composed of the same constituents, they found that the final bulk products with high level ordered nanolaminates exhibited both a high tensile strength of 375 MPa and high strain of 12 %. They attributed this enhancement to the fact that they enabled enhanced dislocation storage capability and two-dimensional alignment of CNTs.

2.2 CNT-Al composites

|

53

CNT-reinforced Al matrix composite materials were successfully fabricated by Kwon and Leparoux [32] using mechanical ball milling followed by powder hot extrusion processes. Although only a small quantity of CNTs were added to the composite (1 vol.%), they found that the Vickers hardness and the tensile strength were significantly enhanced, with an up to three-fold increase relative to that of pure Al. The CNT-Al composites were not only strengthened by the addition of CNTs, but also enhanced by several synergistic effects. The nanoindentation stress-strain curve was successfully constructed by setting the effective zero-load and zero-displacement points and was compared with the tensile stress-strain curve. The yield strengths of the Al-CNT composites were obtained from the nanoindentation and tensile tests, and then compared. Using powder metallurgy followed by 4-pass friction stir processing, Liu et al. [33] fabricated 1.5 vol.% and 4.5 vol.% CNT-reinforced 2009Al (CNT-2009Al) composites with homogeneously dispersed CNTs and refined matrix grains. They tested the tensile properties of the composites between 293 and 573 K and the coefficient of thermal expansion from 293 to 473 K (see Figure 2.13). Figure 2.14 shows (a)

(b)

400nm

400nm (d)

(c)

400nm

400nm

Fig. 2.13. Fractographs of 1.5 vol.% CNT/2009Al composite tested at different temperatures: (a) 293 K; (b) 423 K, (c) 473 K; and (d) 573 K [33].

54 | 2 CNT-reinforced MMCs

Stress (MPa)

400 300

293K 423K 473K 573K

200

500 400 Stress (MPa)

500

100

(a)

Stress (MPa)

400 300

200 100

0 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Strain 500

300

293K 423K 473K 573K

(b)

0 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Strain

293K 423K 473K 573K

200 100

(c)

0 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Strain

Fig. 2.14. Tensile curves of (a) 2009Al, (b) 1.5 vol.% CNT/2009Al and (c) 4.5 vol.% CNT/2009Al at different temperatures [33].

the tensile stress-strain curves of the FSP 2009Al alloy and CNT/2009Al composites at 293–573 K obtained in [33]. Increasing the testing temperature led to a decreased strength level, for both matrix and CNT/2009Al composites. Liu et al. [33] indicated that the load transfer mechanism still takes place at temperatures elevated up to 573 K. Thus, the yield strength of the 1.5 vol.% CNT-2009Al composite at 423–573 K was enhanced compared with the 2009Al matrix. However, for the 4.5 vol.% CNT-2009Al composite, the yield strength at 573 K was even lower than that for the matrix, due to the quicker softening of ultrafine-grained matrix. Compared with the 2009Al matrix, Liu et al. [33] concluded that the coefficient of thermal expansion of the composites was greatly reduced for the zero thermal expansion and high modulus of the CNTs, and well predicted by Schapery’s model. Yoo et al. [34] studied the behavior of CNT-reinforced Al composites fabricated through ball milling combined with rolling. The composites exhibited high strength and high strain-hardening ability. During the ball milling process, CNTs were broken and became low aspect ratio tubes. They observed that the composites had the CNTs (a few dozen nanometers in size) randomly and uniformly dispersed in their grain interiors. They concluded that this type of CNT distribution contributed to work hardening and strengthening by the Orowan mechanism. Yang et al. [35] developed an approach to fabricate CNT-reinforced Al composites. This process allowed well-dispersed and deeply embedded CNT reinforcement in the Al powder, forming an effective inter-

2.3 CNT-Co composites

|

55

face bonding with matrix. They concluded that these CNT-Al composites containing 2.5 wt.% CNTs exhibited ultimate tensile strength of 334 MPa, which was 1.7 times higher than that of unreinforced Al, as well as good ductility of 18 % elongation to failure.

2.3 CNT-Co composites Cobalt (Co) is found naturally only in chemically combined form. The free element, produced by reductive smelting, is a hard, lustrous, silver-gray metal. Today, some cobalt is produced specifically from various metallic-lustered ores (e.g. cobaltite) but the main source of the element is as a by-product of copper and nickel mining. Cobalt is used in the preparation of magnetic, wear-resistant and high-strength alloys. Cobalt silicate and cobalt(II) aluminate give a distinctive deep blue color to glass, smalt, ceramics, inks, paints and varnishes. Co matrix is used for CNT-reinforced MMCs, but mainly for electrochemical performance in coatings and batteries. In this review, only 2 papers are dedicated to CNTCo composites [36,37]. Chen et al. [36] demonstrated that cobalt could be plated onto the surfaces of CNTs by electroless plating. In this manner, a layer of cobalt is formed as nanoparticles on the surface of the CNTs. It was found that the activation process and low deposition rate were critical for getting better coating. Additionally, the heattreatment of the coated CNTs was found to be a very effective way of improving the deposited coating layer. The results from the study by Chen et al. [36] demonstrated the technical feasibility of electroless plating for the preparation of a one-dimensional 0.40

Friction coefficient μ

0.35 0.30 0.25 0.20 Co- II Co- II/P-MWCNTs Co- II/O-MWCNTs

0.15 0.10 0.05 0

100

200 300 Sliding distance /m

400

500

Fig. 2.15. Typical friction coefficients of nanocrystalline Co-II and Co-II/MWCNT coatings with increasing sliding distance [37].

56 | 2 CNT-reinforced MMCs

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 2.16. The magnified SEM morphologies of the worn surfaces of various nanocrystalline Co and Co/MWCNTs sliding against GCr15 ball after sliding 30 in (a) Co-I; (b) Co-I/P-MWCNTs; (c) Co-I/OMWCNTs; (d) Co-II; (e) Co-II/P-MWCNTs; (f) Co-II/O-MWCNTs) [37].

nanoscale composite. Recently, Su et al. [37] characterized nanocrystalline Co and CNT-Co coatings produced by different electrodeposition techniques. Nanocrystalline Co and multi-walled CNT-Co coatings were synthesized by direct current and pulse reverse current electrodeposition from aqueous bath containing cobalt sulfate and multi-walled CNTs. The effect of the functionalization of CNTs and electrodeposition techniques on the microstructure and properties of these coatings was evaluated. The

2.4 CNT-Cu composites

|

57

friction coefficients of nanocrystalline Co and Co/MWCNT coatings do not have a great difference and vary from 0.23 to 0.29. The typical friction coefficient curves of Co-II and its corresponding composite coatings are shown in Figure 2.15. It is observed that the friction coefficients of these coatings steadily increase to 0.3 at around 300 m of the sliding and then slowly increase or maintain this level to the end of the test. There is also no significant difference for the variation of friction coefficient with increasing sliding distance for these coatings. Figure 2.16 exhibits the magnified images of the worn surface morphologies of various nanocrystalline Co and Co/CNT coatings after sliding 30 in. Su et al. [37] results showed that the incorporations of CNTs, particularly the functionalized CNTs, substantially improve the hardness and the resistance to wear and corrosion of the deposited coatings. The functionalization of CNTs favors the co-deposition of CNTs with Co ions, and then improves the hardness and the corrosion and wear resistance of the produced composite coatings. The differences in friction and wear behavior of these nanocrystalline Co and CNT-Co coatings as a function of treatment of CNTs or electrodeposition techniques were attributed to their different hardness, microstructures and the corresponding wear mechanisms.

2.4 CNT-Cu composites Copper (Cu) is a ductile metal with very high thermal and electrical conductivity. It is used as a conductor of heat and electricity, a building material, and a constituent of various metal alloys. In sufficient concentration, Cu compounds are poisonous to higher organisms and are used as bacteriostatic substances, fungicides and wood preservatives. The Cu matrix has been moderately used for CNT-reinforced MMCs. In this review, a total of nine papers are dedicated to CNT-Cu composites [38–46]. Kim et al. [38] investigated the hardness and wear resistance of CNT-reinforced Cu matrix (CNT-Cu) composites. The nanocomposite was fabricated by a molecular level process, which involved suspending CNTs in solvent by surface functionalization, mixing Cu ions with CNT suspension, drying, calcination and reduction. The hardness and sliding wear resistance of the CNT-Cu composite were enhanced by two and three times, respectively, compared to those of the Cu matrix. They attributed the enhancement of hardness to the effect of homogeneous distribution of CNTs in the Cu matrix, good bonding at CNT-Cu interfaces and high relative density of composites. The dispersed CNTs in Cu-matrix composite gave significantly enhanced wear resistance by retarding the peeling of Cu grains during the sliding wear process. Sun et al. [39] applied molecular dynamics and continuum mechanics to predict the compressive mechanical properties of CNTs that encapsulated helical copper nanowire. The helical structures of the copper nanowires were obtained using the “simulated annealing” method. The strain energy curves were shown to predict the interaction between the atoms during the compressive course. They used the model to determine the mechanical properties (critical buckling load and the Young’s mod-

58 | 2 CNT-reinforced MMCs uli) of CNTs that encapsulated helical copper nanowire with different diameters and lengths. Sun et al. [39] found that not all the maximum strengths of the composites are larger than the corresponding carbon nanotube bases, and they were related to the diameter, length and the CNTs chirality. Some excellent properties of CNTs encapsulating helical copper nanowire were revealed in this study. Trinh et al. [40] reported results on the fabrication and calculation of the friction coefficient of Cu matrix composite material reinforced by CNTs. CNT-Cu composites were fabricated by the powder metallurgy method; mechanical properties, such as the friction coefficient, were evaluated using the rule of mixtures. Trinh et al. [40] determined the coefficient of friction of the CNTs component in the CNT-Cu composites. Uddin et al. [41] investigated the hardness and electrical properties of CNTreinforced Cu and Cu alloy (bronze) composites. They were fabricated by a wellestablished hot-press sintering method of powder metallurgy. They found that the effect of the shape and size of metal particles as well as the selection of CNTs significantly influenced the mechanical and electrical properties of the composites. The hardness of the CNT-Cu composite improved up to 47 % compared to that of pure Cu, while the electrical conductivity of the bronze composite increased up to 20 % compared to that of the pure alloy. They concluded that CNTs improved the mechanical properties of highly-conductive low-strength Cu metals, whereas in low-conductivity high-strength copper alloys, the electrical conductivity was improved. Using high power lasers, Bhat et al. [42] fabricated and studied the mechanical and thermal properties of multi-walled CNT-reinforced Cu-10Sn alloy composites. Microstructural observations showed that CNTs were retained in the composite matrix after laser processing. The addition of CNTs showed improvement in the strain hardening, mechanical and thermal properties of Cu-10Sn alloy. Composites with 12 vol.% CNTs showed more than 80 % increase in the Young’s modulus and 40 % increase in the thermal conductivity of Cu-10Sn alloy. Bhat et al. [42] found that the yield strength estimates obtained from models based on strengthening mechanisms derived from the Maxwell-Garnett effective medium were shown to be very accurate. Xu et al. [43] studied the effect of electrical current on the tribological properties of the Cu matrix composite reinforced by CNTs. The CNT-Cu composites were reinforced with 10 % CNTs, and pure Cu bulk were prepared by powder metallurgy techniques under the same consolidation processing condition. They investigated the effect of electrical current on the tribological property of the materials using a pin-on-disk friction and wear tester. The results published by Xu et al. [43] showed that the friction coefficient and wear rate of the CNT-Cu composite as well as those of pure Cu bulk increased with the electrical current. They identified that the dominant wear mechanisms were arc erosion wear and plastic flow deformation, and CNTs improved their tribological properties of Cu composites with electrical current. Multi-walled CNT-Cu composites, exhibiting chromium (Cr) carbide nanostructures at the CNT-Cu interface, were prepared by Cho et al. [44] using a carbide formation using CuCr alloy powder. The tensile strengths of the CNT-CuCr composites in-

2.4 CNT-Cu composites

|

59

350

350

300

300

250

Loading

200 150 fcr = 47.25 μN

100

Unloading

Load sample (μN)

Load sample (μN)

creased with increasing CNTs content, while the tensile strength of CNT-Cu composite decreased from that of monolithic Cu. The enhanced tensile strength of the CNT-CuCr composites was a result of possible load-transfer mechanisms of the interfacial Cr carbide nanostructures. They observed the failure of multi-wall of CNTs in the fracture surface of the CNT-CuCr composites, indicating an improvement in the load-bearing capacity of the CNTs. This result proved that the Cr carbide nanostructures effectively transferred the tensile load to the CNTs during fracture through carbide nanostructure formation in the CNT-Cu composite. Cu matrix composites reinforced with 0.2, 5 and 10 vol.% single-walled CNT and 5 and 10 vol.% multi-walled CNTs were produced by Shukla et al. [45] using high energy milling of pure copper powder with CNTs and subsequent consolidation by vacuum hot pressing. Significant improvement in hardness of the single-walled CNT-Cu composite was observed with an increase in CNT content. In the case of the multi-walled CNT-Cu composite, hardness reduced for 10 vol.% CNT composites. Compression strength of the single-walled CNT-Cu samples was found by Shukla et al. [45] to be higher than the multi-walled CNT-reinforced samples. Tsai and Jeng [46] experimentally and numerically investigated the effect of CNT buckling on the reinforcement of CNT-Cu composites. The CNT-Cu composites with high strength and good damping were developed using acid treatment, sintering processes and consolidation techniques. In this study, strengthening of the composites, with influence of CNT buckling, was demonstrated by experimental nanoindentation

250

Loading

200 150 fcr = 55.67 μN

100

Unloading

50

50

0

0 0

0

50 100 150 200 250 300 350 400

(a)

Depth (nm)

(b)

50 100 150 200 250 300 350 400 Depth (nm)

350 Load sample (μN)

300 Loading

250 200 150

fcr = 92.38 μN

Unloading

100 50 0 0

(c)

50 100 150 200 250 300 350 400 Depth (nm)

Fig. 2.17. Representative load-displacement curves obtained from experimental nanoindentation tests: (a) Cu matrix with no CNT addition; (b) CNT/Cu composite containing CNTs of length 2.6 μm; and (c) CNT/Cu composite containing CNTs of length 2.1 μm [46].

60 | 2 CNT-reinforced MMCs 200

Cu/CNT composite (L=13.8nm) Cu/CNT composite (L=17.1nm) Pure Cu (III)

Force (nN)

150

(II) Local‒buckling

100

(I)

50

0

Global‒buckling

0

5

(a)

(I)

10 15 20 Indentation depth (nm)

(II)

25

30

(III)

(b) Fig. 2.18. (a) Simulated force-displacement curves for Cu matrix and CNT/Cu composites (note: Labels (I), (II) and (III) represent different stages of the deformation process and correspond to the snapshots presented in Figure 2.18b). (b) Typical snapshots of pure Cu and CNT/Cu composites during nanoindentation (note: snapshots (II) and (III) show local-buckling and global-buckling processes, respectively) [46].

tests and molecular dynamics simulations. Figure 2.17 presents the experimental loaddisplacement curves for the pure Cu sample and the CNT/Cu composite samples with CNTs of 2.6 and 2.1 μm in length, respectively. The experimental results obtained by Tsai and Jeng [46] showed that the buckling behavior of CNTs in the CNT-Cu composites varies with their slenderness ratio. Figure 2.18 presents the simulated forcedisplacement curves for the pure Cu sample and the CNT/Cu composite samples containing CNTs in lengths of 13.8 and 17.1 nm, respectively. As shown in Figure 2.18a, the discontinuous force drop in force-depth curves can represent the critical buckling

2.5 CNT-Fe composites

| 61

force (fcr ). Figure 2.18a also shows that both composite samples undergo linear elastic deformation during the initial loading stage. However, it can be seen that the critical buckling force increases with a decreasing CNT length. Figure 2.18b presents snapshots of the Cu matrix (I), the CNT/Cu composite with CNTs of length 17.1 nm (II) and the CNT/Cu composite with CNTs of length 13.8 nm (III). The results show that in the CNT/Cu composite samples, the buckling mode is sensitive to the length of the CNTs. Their study showed significant buckling behavior of CNTs in the CNT-Cu composites, where the shorter CNTs (with a smaller slenderness ratio) gave rise to a global buckling and the slender CNTs (with a larger slenderness ratio) induced local buckling. The MD simulation results revealed that the buckling behavior of the CNTs played a key role in increasing the mechanical strength of CNT-Cu composites.

2.5 CNT-Fe composites Iron (Fe) is a metal in the first transition series and the most common element (by mass) that forms the planet Earth as a whole (outer and inner core). It is the fourth most common element in the Earth’s crust. Iron metal has been used since ancient times, though copper alloys, which have lower melting temperatures, were used first in history. Pure iron is soft (softer than aluminum), but is unobtainable by smelting. The material is significantly hardened and strengthened by impurities from the smelting process, such as carbon. A certain proportion of carbon (between 0.002 % and 2.1 %) produces steel, which may be up to 1000 times harder than pure iron. Crude iron metal is produced in blast furnaces, where ore is reduced by coke to pig iron, which has high carbon content. Further refinement with oxygen reduces the carbon content to the correct proportion to make steel. Due to their great range of desirable properties and the abundance of iron, steels and low carbon iron alloys with other metals (alloy steels) are by far the most common metals in industrial use. Fe matrix has been fairly used for CNT-reinforced MMCs. In this section, a total of eight papers on CNT-Fe composites are reviewed [47-54]. Ding et al. [47] employed mechanical alloying to produce Al2 O3 /Fe, Al2 O3 /Co and Al2 O3 /Ni nanocomposites and found that high-energy mechanical milling leads not only to drastic refinement but also to good dispersion of catalyst precursors in oxide matrixes. After mechanical milling, they observed solid-state alloying and accelerated substitutional reactions between the parent oxides. These nanocomposites possessed the fine-grained and porous structures and thus high reducibility. Large-scale formation of multi-walled and single-walled CNTs were achieved by using these mechanical alloying-derived Al2 O3 /Fe, Al2 O3 /Co and Al2 O3 /Ni nanocomposites. Atomic structures of single-crystalline iron-based nanowires crystallized inside multi-walled CNTs during pyrolysis on silicon substrates with ferrocene as a precursor were analyzed by Golberg et al. [48] using high-resolution analytical transmission electron microscopy and electron diffraction. Standard crystal lattices (body-centered

62 | 2 CNT-reinforced MMCs cubic – bcc, α -Fe; face-centered cubic – fcc: 𝛾-Fe; orthorhombic cementite Fe3 C) were all found forming inside the CNTs. Both bcc and fcc nanowires display a wide variety of lattice planes that are parallel to the CNT walls, with none of the orientations being preferable. Both long-period and standard cementite nanowires exhibited welldefined transient zones in the vicinity of nanowire-CNT interfaces, where perfectly ordered carbide lattice fringes disappeared. The results by Golberg et al. [48] suggested the non-existence of metastable equilibrium in the nanoscale Fe-C system between carbide and graphite phases during iron crystallization inside graphitic tubular channels. Shpak et al. [49] used X-ray diffraction, transmission electron microscopy, ferromagnetic and electron paramagnetic resonances to investigate the Fe-filled multiwalled CNTs. The iron within the CNTs was found in three phases: the austenite 𝛾-Fe is located at the top of the CNTs, while the ferrite α -Fe and cementite θ -Fe3 C are found close to the substrate. Two ferromagnetic signals were observed by Shpak et al. [49] and identified as those belonging to ferrite and cementite. Ferromagnetic signals revealed a surprising temperature dependence: with decreasing temperature, their integral intensity decreases nearly linearly, and the signals disappear at temperatures below 70 K. Spark plasma sintering was used by Pang et al. [50] to fabricate a dense Fe3 Al/CNT composite with a CNT content of 5.0 vol.%, retaining the integrity of CNT in the matrix. They synthesized samples at pressure of 30 MPa and temperature of 1273 K, and studied the composite structure using X-ray diffraction and a transmission electron microscope (TEM). Pang et al. [50] concluded that the composites had very promising mechanical properties, such as microhardness of 8.7 GPa, compressive yield strength of 3175 MPa, which is about 95 % and 56 % higher than monolithic Fe3 Al fabricated under the same process. They indicated that CNTs can be considered as an effective nanoscale reinforcement of intermetallics matrix composites. Pang et al. [51] prepared multi-walled CNTs-Fe3 Al composite using spark plasma sintering and investigated the magnetic properties of the nanocomposite with an alternating gradient force magnetometer. They showed that the CNTs-Fe3 Al composite displays good soft magnetic properties and has similar magnetic hysteresis loops to that of Fe3 Al. The excellent magnetic properties evidenced in their study imply that multi-walled CNTs-Fe3 Al composites might have significant potential for applications in electronic-magnetic nanodevices. Pang et al. [52] performed analyses on the stress in the CNT-Fe3 Al composites. They mentioned that the biphase interface valence electron structure was established on the basis of Pauling’s nature of the chemical bond, and that the stress occurs by the huge interface electron density difference, which blocks the Fe3 Al grain agglomeration and growth. They also measured the compressive stress existing in the CNT-Fe3 Al interface with X-ray diffractions, receiving a value of 0.38 GPa. They declare that this experimental result verifies that the stress has a positive effect on the enhancement of mechanical properties of the composite. Using spark plasma sintering, Pang et al. [53] fabricated an iron aluminides (Fe3 Al)-based composite with different amounts of multi-walled CNTs and studied

2.6 CNT-Mg composites

|

63

the influence of CNTs content on the mechanical properties of the composite. They concluded that the compressive yield strength and fracture toughness of the 3 vol.% CNT-Fe3 Al composite, compared to the monolithic Fe3 Al, were enhanced 73.6 % and 40 %, respectively. The defects and alignment direction of CNTs in the matrix affected the mechanical properties of the composite. They also identified that the failure mechanisms include CNTs pulling-out, crack deflection, bridging and CNTs rupture. Madaleno et al. [54] used catalytic decomposition of ethylene over iron montmorillonite surfaces to synthesize montmorillonite-CNT hybrids. SEM and STEM analyses revealed the presence of CNTs attached to the clay layers and X-ray diffraction results and indicated that sodium montmorillonite layers were intercalated with iron species during the ion-exchange processes and further delaminated due to the growth of CNTs. Due to their pre-exfoliated internal structure and the presence of surface CNTs, Madaleno et al. [54] suggested that montmorillonite-CNT hybrids benefit the enhancement of mechanical properties in polymer nanocomposites.

2.6 CNT-Mg composites Magnesium (Mg) is an alkaline earth metal and the eighth most abundant element in the Earth’s crust and ninth in the known universe as a whole. The free metal burns with a characteristic brilliant white light, making it a useful ingredient in flares. The metal is now mainly obtained by electrolysis of magnesium salts obtained from brine. Commercially, the main use for the metal is as an alloying agent to make aluminiummagnesium alloys, sometimes called magnalium or magnelium. Since magnesium is less dense than aluminum, these alloys are prized for their relative lightness and strength. Mg matrix is generally used for CNT-reinforced MMCs. In recent years, some achievements have been made in the development of CNT-Mg composites. In this section, 13 papers dedicated to this topic are summarized [55–67]. Thakur et al. [55] investigated the synthesis and mechanical behavior of CNT-magnesium composites hybridized with nanoparticles of alumina. The CNT-Mg composites were prepared by a powder metallurgy route coupled with rapid microwave sintering. Nanometer-sized particles of alumina were used to hybridize the CNT reinforcement in the Mg matrix and establish the intrinsic influence of hybridization on mechanical behavior of the composite. The yield strength, tensile strength and strain-to-failure of the CNT-Mg composites were found to increase with the addition of nanometer-sized alumina particles to the composite matrix. They performed scanning electron microscopy observations of the fracture surfaces of the samples that failed in uniaxial tension, which revealed the presence of cleavage-like features on the fracture surface indicative of the occurrence of locally brittle fracture mechanisms in the composite microstructure. Li et al. [56] applied a two-step process to produce CNT-Mg alloy composites. In the first stage, they used a block copolymer as a dispersion agent to pre-disperse multiwalled CNTs on Mg alloy chips. After that, they melted and stirred the chips with the

64 | 2 CNT-reinforced MMCs dispersed CNTs on their surface. They observed that CNTs were quite successfully dispersed on the surfaces of the Mg alloy chips. The mechanical properties of the CNT-Mg composites were measured by compression testing. They observed that the compression at failure, the compressive yield strength and ultimate compressive strength improved significantly up to 36 % by only adding 0.1 wt.% CNTs to the Mg alloy. CNT-Mg and CNT-Mg-Ni (Mg-23.5 wt.% Ni) composites were processed by Schaller et al. [57] using powder metallurgy and then charged with hydrogen by annealing at 620 K under a pressure of 0.4 MPa of hydrogen. They performed mechanical spectroscopy and concluded that such a treatment has no effect in the composites with pure Mg matrix. On the other hand, Mg-23.5 wt.% Ni alloys, unreinforced as well as reinforced with CNTs, exhibited mechanical loss spectrum, which was deeply modified by hydrogen charging. Aung et al. [58] mentioned that CNTs may be added to Mg matrix to produce composites of better mechanical properties, but their effect on corrosion behavior was not well understood. They studied the corrosion resistance of pure Mg and its composites reinforced with 0.3 and 1.3 wt.% CNTs in 3.5 wt.% NaCl solution using immersion testing and electrochemical measurements. They found that the corrosion rate was increased considerably by the presence of CNTs because of microgalvanic action between the cathodic CNTs and the anodic Mg matrix. An advanced powder metallurgy process that disperses un-bundled CNTs was developed by Kondoh et al. [59]. They applied it to fabricate a Mg matrix composite reinforced with CNTs. When approximately 1 vol.% CNTs were added, the extruded pure Mg and AZ31B alloy composites displayed an extremely large increase of the tensile yield stress of 25–40 %, compared to Mg materials containing no CNTs. They also verified the presence of MgO thin layers of 2–4 nm thickness at the interface between α -Mg and the unbundled CNTs. This resulted in an effective tensile loading transfer at the interface, which significantly improved the tensile strength and yield stress of the CNT-Mg composites. However, they also observed that the elongation was less than 5 % and the composites exhibited very poor ductility. They attributed this fact to the ductility of the MgO layers at the interface between α -Mg and CNT. Srivatsan et al. [60] studied the influence of CNTs and processing on cyclic fatigue and fracture behavior of a magnesium alloy. CNT-Mg alloy (AZ31) was fabricated using the technique of solidification processing followed by hot extrusion. They performed tests on both composite and unreinforced alloy at two different load ratios spanning tension-tension loading and fully-reversed tension-compression loading under total stress amplitude control. The comparison between the CNT-Mg alloy and the unreinforced counterpart made by Srivatsan et al. [60] revealed more than 200 % improvement in cyclic fatigue life. At all values of maximum stress, the high cycle fatigue response of both the reinforced and unreinforced magnesium alloy was found to degrade. They also presented the synergistic and interactive influences of reinforcement and processing on microstructural development, cyclic fatigue life and kinetics governing fracture behavior. Srivatsan et al. [61] studied the influence of CNTs on

2.6 CNT-Mg composites

|

65

the tensile response and fracture behavior of Mg alloy. They discussed the conjoint influence of reinforcement and processing on microstructural development, microhardness, tensile deformation and final fracture behavior of the Mg alloy composite; comparisons were made with the unreinforced alloy (AZ31). They highlighted the interactive influences of the CNT reinforcement and processing in governing engineering stress versus engineering strain response and tensile properties. Srivatsan et al. [61] discussed the macroscopic fracture mode and intrinsic microscopic mechanisms governing quasi-static deformation and fracture behavior of both the CNT-reinforced and unreinforced magnesium alloy. Srivatsan et al. [62] developed further studies on the microstructure, hardness, tensile properties, tensile fracture, high cycle fatigue characteristics and final fracture behavior of CNTs-Mg alloy composite with 1.0 vol.% CNT. They found that the elastic modulus, yield strength and tensile strength of the reinforced Mg alloy was noticeably higher compared to the unreinforced counterpart. The ductility, quantified both by elongation-to-failure and reduction in the cross-section area of the composite was higher than the monolithic counterpart. A comparison of the CNT-Mg alloy with the unreinforced Mg revealed a noticeable improvement in cyclic fatigue life. Park et al. [63] evaluated the mechanical properties of a magnesium matrix composite reinforced with Si coated CNTs. Multi-walled CNT-reinforced Si coatings were prepared using a solid reaction method between pure silicon powders and CNTs. MMC with CNTs reinforced AZ91 were fabricated by the squeeze infiltrattion method. They found that the Si-coated CNTs improved the wettability, distribution and bonding strength in the CNT-Mg composite. They also concluded that the squeeze infiltration technique was a proper method to fabricate Mg-based composites because it reduces casting defects such as pores and matrix/reinforcement interface separation. Park et al. [63] concluded that the tensile strength increased with the reinforcement of Si-coated CNT to magnesium alloys. Mg composite reinforced with CNTs were produced by Fukuda et al. [64] using both a pure Mg and AZ61 Mg alloy matrix. They highlighted the superior mechanical properties of the CNT-Mg composite. The composites were produced via a powder metallurgy route containing wet process using isopropyl alcohol-based zwitterionic surfactant solution with unbundled CNTs. The produced composites were evaluated with a tensile test and Vickers hardness test and analyzed by X-ray diffraction and field-emission scanning electron microscopy. They observed that the Al2 MgC2 compounds formed at the interface between Mg matrix and CNTs effectively reinforced the interfacial bonding and enabled tensile loading transfer from the Mg matrix to CNTs. Furthermore, they also clarified that the microstructures and grain orientations of the composite matrix were not significantly influenced by CNT addition. Mg containing 6 wt.% Al alloy composites reinforced with CNTs were fabricated with powder metallurgy-based wet-processing by Fukuda et al. [65]. They observed that both yield stress and tensile strength improved with the addition of CNTs. Field emissiontransmission electron microscopy microstructural analysis clarified that needle-like

66 | 2 CNT-reinforced MMCs ternary carbides of Al2 MgC2 were synthesized at some interfaces between magnesium matrix and CNTs, and the other interfaces were clean without any other materials or defects. The tensile loading transfer from Mg matrix to CNTs was effectively strengthened by both the production of Al2 MgC2 compounds and the clean interface between Mg matrix and CNTs. Zhou et al. [66] studied the tensile mechanical properties and strengthening mechanism of hybrid CNT and silicon carbide nanoparticle-reinforced Mg alloy composites. AZ91 Mg alloy hybrid composites reinforced with different hybrid ratios of CNTs and silicon carbide (SiC) nanoparticulates were fabricated by semisolid stirringassisted ultrasonic cavitation. Their results showed that grains of the matrix in the AZ91-(CNT + SiC) composites were refined after adding hybrid CNTs and SiC nanoparticles to the AZ91 alloy; the room-temperature mechanical properties of AZ91-(CNT + SiC) hybrid composites were improved compared to the unreinforced AZ91 matrix. Figure 2.19 shows the stress-strain curves of AZ91/(CNT + SiC) hybrid composites with different hybrid ratios obtained by Zhou et al. [66]. It can be seen that hybrid composites also show a similar type of curve as do CNT- or SiC-reinforced composites, but the change from linear to nonlinear deformation occurs at a higher stress in the former case probably because of the coexistence of CNTs and SiC nanoparticles. The yield strength, maximum tensile stress and elongation rate at break of the hybrid composites increased with the increase of the hybrid mass ratio between CNTs and SiC nanoparticles. As revealed by the fractured surfaces of the tensile specimens in Figure 2.20, significant improvements in tensile properties were accompanied by a tran350 300

Stress (MPa)

250 200 150 100 AZ91/0.7% CNT + 0.3% SiC, CNT:SiC = 7:3 AZ91/0.5% CNT + 0.5% SiC, CNT:SiC = 5:5 AZ91/0.3% CNT + 0.7% SiC, CNT:SiC = 3:7

50 0

0

2

4 6 Strain (%)

8

10

Fig. 2.19. Stress-strain curves of AZ91/(CNT + SiC) hybrid composites with different hybrid ratios [66].

2.7 CNT-Ni composites

(a)

|

67

(b)

Fig. 2.20. Representative SEM fractographs of: (a) AZ91 Mg-alloy; and (b) AZ91/(0.7 % CNT + 0.3 % SiC) [66].

sition from a brittle fracture (Figure 2.20a) for as-cast AZ91 Mg-alloy to quasi-cleavage fracture state (Figure 2.20b) for AZ91/(0.7 % CNT + 0.3 % SiC). Zhou et al. [66] showed that the tensile mechanical properties of the AZ91 alloy-based hybrid composites were considerably improved at the mass hybrid ratio of 7:3 for CNTs and SiC nanoparticles. The tensile and yield strength increased about 45–55 % after gravity-permanent mold casting. They attributed the increase in the room-temperature strength of the hybrid composites to: (i) the larger hybrid ratio of CNTs and SiC nanoparticles; (ii) the coefficient of thermal expansion mismatch between matrix and hybrid reinforcements; (iii) the dispersive strengthening effects (Orowan strengthening); and (iv) the grain refining (Hall-Petch effect). CNT-Mg composite powders were produced by Sun et al. [67] and synthesized by in situ chemical vapor deposition of acetylene in Co/Mg catalyst powder at 480 ∘ C. The factors influencing the morphology and yield of CNTs, including the compositions of the Co catalyst and the ratio of carbon source to carrier gas, were investigated by Sun et al. [67]. Their results showed that CNTs had a mean size of 15 nm and were highly graphitizing and homogeneously dispersed on the Mg powder. Ball milling of CNT-Mg composite powders for short times enabled the embedding of CNTs into Mg. This evidence resulted in grain refinement. They concluded that the tensile strength of CNT-Mg composites reached an average of 285 MPa, thus leading to an increase of 45 % compared to commercially pure Mg.

2.7 CNT-Ni composites Nickel (Ni) is a silvery-white lustrous metal with a slight golden tinge and belongs to the transition metals, being hard and ductile. Because of nickel’s slow rate of oxi-

68 | 2 CNT-reinforced MMCs dation at room temperature, it is considered corrosion-resistant. Historically this has led to its use in plating metals such as iron and brass, in chemical apparatus and in certain alloys that retain a high silvery polish, such as German silver. About 6 % of world nickel production is still used for corrosion-resistant pure-nickel plating. Nickel was once a common component in coins, but has largely been replaced by cheaper iron for this purpose. In comparison with Al and Mg, Ni matrix is not widely used for CNT-reinforced MMCs. In this section, five papers dedicated to this topic are briefly described [68–72]. Wang et al. [68] studied the friction and wear behavior of electroless Ni-based CNT composite coatings. Ni-based CNT composite coatings with different volume fractions (from 5 to 12 vol.%) of CNTs were deposited on medium carbon steel substrates by electroless plating. The friction and wear behavior of the composite coatings was investigated by Wang et al. [68] using a pin-on-disk wear tester under unlubricated conditions. Friction and wear tests were conducted by these authors at a sliding speed of 0.0623 m/s and at an applied load of 20 N. Their experimental results indicated that the friction coefficient of the composite coatings decreased with an increase in the volume fraction of CNTs due to self-lubrication and unique topological structure of CNTs. Within the range of volume fraction of CNTs from 0 to 11.2 %, the wear rate of the composite coatings showed a steadily decreasing trend with the increasing volume fraction of CNTs. However, the wear rate of the composite coatings increased with a further increase in the volume fraction of CNTs due to the conglomeration of CNTs in the matrix. A Ni matrix reinforced with multi-walled CNTs was processed by Scharf et al. [69] in a monolithic form using the laser-engineered net shape processing technique. They performed Auger electron spectroscopy maps and determined that the CNTs were well dispersed and bonded in the Ni matrix, and no interfacial chemical reaction products were determined in the as-synthesized composites. Mechanisms of solid lubrication were investigated by Scharf et al. [69] using micro-Raman spectroscopy spatial mapping of the worn surfaces to determine the formation of tribochemical products. The CNT-Ni composites exhibited self-lubricating behavior, forming an in situ, low interfacial shear strength graphitic film during sliding, resulting in a decrease in the friction coefficient compared to pure Ni. CNT-Ni composites were processed by Singh et al. [70] using a laser deposition technique known as the laser engineered net shaping process. The mechanical milling of the powder consisting of nickel powders and CNTs resulted in a more homogeneous distribution of the CNTs in the Ni matrix and two distinct scales of reinforcements within the composite. The larger reinforcement scale consisted of submicrometer to micrometer sized bundles of CNTs while the smaller scale consists of individual CNTs within the Ni matrix. High-resolution transmission electron microscopy indicated that the CNT-Ni interface was well bonded without the presence of any significant interfacial reaction product. Raman spectroscopy revealed that the degree of disorder and defects in the CNT bundles in the nickel matrix varied as a function of bundle size.

2.7 CNT-Ni composites

|

69

CNTs filled with metals can be used in capacitors, sensors, rechargeable batteries and so on. Atomic arrangement of the metals plays an important role in the function of the composites. Liu et al. [71] showed that Ni and NiO nanoparticles in CNTs are crystalline, resulting in the occurrence of lattice shrinkage. Lattice shrinkage caused a misfit of the lattice constant between the Ni and the internal CNT surface (from theoretical 1.21 % to actual 0.86 %). They observed that this variation was beneficial to a heterogeneous nucleation of the Ni crystal nucleus on the surface at a point with the lowest interfacial energy at the coherent interface. According to their findings, one-dimensional nanostructures can be changed and controlled during synthesis by use of CNTs as a template. Hwang et al. [72] performed interface analyses of ultrahigh strength CNT-nickel composites processed by molecular level mixing followed by spark plasma sintering (SPS). The schematic steps performed by Hwang et al. [72] for the molecular-level mixing process of CNT/Ni powders are presented in Figure 2.21. Figure 2.22a shows the true stress-strain curve of the CNT/Ni composite compared with monolithic Ni, which was produced by the SPS process under similar conditions. They found that CNT-Ni composites exhibited a yield strength of 710 MPa, about 3.7 times higher than monolithic Ni. The overall microstructure, revealing the homogeneous distribution of individual CNTs within the nickel matrix, is shown in the two bright-

Functionalization of CNTs CNT Dispersion Dispersion of CNTs in Ethylene glycol

Addition of Ni(ac)24H2O in Ethylene glycol Hydrazine + NaOH Direct Reduction of Ni in solution

Heating to 60°C and Stirring

Formation of CNT/Ni in Solvent

Cleaning and Drying (Ethanol, vacuum oven) Separation & Reduction of CNT/Ni

Reduction of CNT/Ni Nancomposite (H2+CO)

CNT/Ni Nanocomposite Powders Fig. 2.21. Schematic diagram showing experimental steps for molecular level mixing process of CNT/Ni powders [72].

70 | 2 CNT-reinforced MMCs

1200

Pure Ni 6 vol.% CNT/Ni

Stress (MPa)

1000 800

I

II

III

600 400 200 0 0.00 0.01 0.02 0.03 0.1

(a) (b)

200 nm

0.2

0.3

0.4

Strain (Ɛ) (c)

200 nm

Fig. 2.22. (a) Stress–strain curves of CNT/Ni composite and pure Ni obtained from tensile test. Bright filed TEM images of highlighted CNT region (b), and Ni matrix region (c) from two different areas [72].

field TEM micrographs, Figure 2.22b and c. These two bright-field images have been acquired by Hwang et al. [72] from two different CNT/Ni TEM samples, prepared in different ways, to achieve two different degrees of thinning. The low magnification of the TEM image, presented in Figure 2.22b, clearly shows that CNTs are well dispersed in the Ni matrix and also appear to be linked to form a network. They attributed the enormous strength increase in these composites to the homogeneous distribution of CNTs in the Ni matrix coupled with the formation of well-bonded, high-strength, contaminant-free CNT-Ni interfaces as revealed by high-resolution transmission electron microscopy. Such interfaces effectively transfer load between CNTs and the Ni matrix in the CNT-Ni composites.

2.8 CNT-Ti composites

|

71

2.8 CNT-Ti composites Titanium (Ti) is a lustrous transition metal with a silver color, low density and high strength. It is highly resistant to corrosion in seawater, aqua regia and chlorine. The element occurs within a number of mineral deposits, principally rutile and ilmenite, which are widely distributed in the Earth’s crust and lithosphere, and it is found in almost all living things, rocks, water bodies and soils. Titanium can be alloyed with iron, aluminum, vanadium, molybdenum, among other elements, to produce strong lightweight alloys for aerospace (jet engines, missiles and spacecraft), military, industrial process (chemicals and petro-chemicals, desalination plants, pulp and paper), automotive, agri-food, medical prostheses, orthopedic implants, dental and endodontic instruments and files, dental implants, sporting goods, jewelry, mobile phones and other applications. The two most useful properties of the metal are corrosion resistance and the highest strength-to-weight ratio of any metal. In its unalloyed condition, titanium is as strong as some steels, but 45 % lighter. Ti matrix is not widely used as Al and Mg for CNT-reinforced MMCs, but very recent works have been published. In this section, six papers [73-78] are reviewed that focus on this theme. Xue et al. [73] prepared CNT-Ti composites using the heterogeneous coacervation method followed by spark plasma sintering and studied the morphology, phase structure and elevated temperature compressive properties. They observed that the CNTs were well dispersed onto the Ti particles, and parts of them were converted into TiC along the interface after SPS due to a solid-state reaction. They concluded that the compressive yield strength declined before a slight rise with the increase in sintering temperatures, which indicated that the reinforcements and dense structure caused the combined action. CNT-reinforced TiNi matrix composites were produced by Feng et al. [74] and synthesized by employing elemental powders. Two methods have been used to control the interfacial reaction between CNTs and Ti, which heightened the quality of CNTs (graphitized multi-walled CNTs) and shortened the time of over-eutectic temperature sintering (two-stage hot pressed sintering). They found that a major TiNi parent phase with different proportions of intermetallic compound phases was obtained by varying the second-stage sintering temperatures. The specimens with second-stage sintering temperature at 1050 ∘ C exhibited predominant mechanical properties, and the CNTs retained their original size. Feng et al. [75] identified a slight decrease in the average friction coefficient and a great decline in volume loss (reaching to 63.1 %) in TiNi matrix composites prepared by sintering 1 vol.% multi-walled CNTs with Ti and Ni elemental powders. They concluded that in situ TiC and the remaining CNTs act as reinforcements and play the major role in the improvement. Their study explored the possibility of developing novel TiNi matrix tribocomposites. Cai et al. [76] prepared CNT-reinforced TiNi matrix composites using spark plasma sintering employing elemental powders and studied the phase structure, morphology and transformation behaviors. They found that thermoelastic martensitic transforma-

72 | 2 CNT-reinforced MMCs tion behaviors could be observed from the samples sintered above 800 ∘ C even with a short sintering time (5 in), and the transformation temperatures gradually increased with increasing sintering temperature because of more Ti-rich TiNi phase formation. They argued that method supplies a basis for preparing CNT-reinforced TiNi composites. Kondoh et al. [77] investigated the high-temperature properties of extruded titanium composites fabricated from CNTs coated titanium powder. The pure Ti matrix composite reinforced with CNTs was prepared by spark plasma sintering and hot extrusion via powder metallurgy process. Ti powders were coated with CNTs via a wet process using a zwitterionic surfactant solution containing 1.0, 2.0 and 3.0 wt.% of CNTs. They observed that situ TiC formation via reaction of CNTs with titanium occurred during sintering, and TiC particles were uniformly dispersed in the matrix. The tensile properties of the composites were evaluated by Kondoh et al. [77] at room temperature, 473, 573 and 673 K, respectively. They concluded that the mechanical properties of extruded CNT-Ti composites at elevated temperature remarkably improved by adding a small amount of CNTs, compared to extruded Ti matrix. They attributed this evidence to the stabilization effect of TiC dispersoids originated from CNTs on the microstructure of extruded Ti composites. CNT and graphite (Gr)-reinforced Ti metal matrix composites were fabricated by Li et al. [78] via powder metallurgy. They prepared 0–0.4 wt.% CNT/Gr and Ti mixture powders by a rocking mill. Microstructures and mechanical properties of the as-extruded Ti composites were investigated by Li et al. [78] to evaluate strengthening effects of CNT/Gr on Ti matrix. Figure 2.23a shows electron probe microanalysis (EPMA) results of Ti-0.4 % CNT composite extruded at 1273 K. As a comparison, a dispersoid in approximate same size in Ti-0.4 % Gr composite was also studied by EPMA and showed in Figure 2.23b. The results obtained by Li et al. [78] show that CNTs are much reactive to form titanium carbide with Ti than graphite at the same conditions. Room temperature tensile tests were carried out by Li et al. [78] to assess the tensile behavior of the as-extruded samples with different contents of CNT/Gr. The tensile strength, yield stress and total elongation to failure are illustrated in Figure 2.24a and b. Accordingly, the average values of the ultimate tensile strength, 0.2 % yield strength and elongation to failure calculated on three tensile test specimens are summarized in Figure 2.24c. Both the ultimate tensile strengths and the yield strengths of Ti-CNT composites were increased gradually when the content of CNT rose from 0 to 0.4 wt.%, and reached maximum values at 0.4 wt.% CNT additive. They mentioned that the mechanical strength of CNT/Gr-Ti composites grew after increasing CNT/Gr content from 0.1 to 0.4 wt.%. Compared to pure Ti, the yield strength and ultimate tensile strength of Ti-0.4 wt.% CNT composites increased 40.4 % and 11.4 %, respectively. Compared to pure Ti, the yield strength and ultimate tensile strength of Ti-0.4 wt.% Gr increased 30.5 % and 2.1 %, respectively. They also discussed the strengthening mechanism, which included grain refinement, carbon solid solution strengthening and TiC/carbon dispersion strengthening.

2.8 CNT-Ti composites

|

(a) Ti-0.4VGCF

Ti

SL

C

2 μm

CP

(b) Ti-0.4Gr

Ti

SL

C

2 μm

CP

Fig. 2.23. Electron probe microanalysis (EPMA) of Ti–CNT/Gr composite extruded at 1273 K. (a) Ti0.4 % CNT. (b) Ti-0.4 % Gr [78].

73

800

700

700

600

600

400 300 200 0

Ti-0.1VGCFr-1

Ti-0.1VGCFr-2

Ti-0.1VGCFr-3

Ti-0.2VGCF-1

Ti-0.2VGCF-2

Ti-0.2VGCF-3

Ti-0.4VGCF-1

Ti-0.4VGCF-2

Ti-0.4VGCF-3

Pure Ti

After tensile test

300 200 100

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 Strain,ε (b)

Before tensile test

500 400 Ti-0.1Gr2

Ti-0.1Gr3

Ti-0.2Gr1

Ti-0.2Gr2

Ti-0.2Gr3

Ti-0.4Gr1

Ti-0.4Gr2

Ti-0.4Gr3

Pure Ti

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 Strain,ε

800 700 600 500 400 300 200

0.2%YS Ti-VGCF UTS Ti-VGCF Elongation Ti-VGCF

100

0.2%YS Ti-Gr UTS Ti-Gr Elongation Ti-Gr

0 0

(c)

0

Ti-0.1Gr1

100 90 80 70 60 50 40 30 20 10 0

Elongation, ε(%)

500

100

(a)

Stress, σ/MPa

800

Tensile strength, σUTS,σYS (MPa)

Stress, σ/MPa

74 | 2 CNT-reinforced MMCs

0.1 0.2 0.3 0.4 Concentration of VGCF/Gr (wt%)

Fig. 2.24. Mechanical properties of as-extruded Ti-CNT/Gr composites. (a) Strain-stress curves of Ti-CNT composites. (b) Strain-stress curves of Ti-Gr composites. (c) Mechanical properties versus CNT/Gr concentrations [78].

2.9 Concluding remarks Since the discovery of carbon nanotubes (CNTs) and their excellent mechanical properties, they have been used as exceptional nanofibres. The simultaneous (i) improvement of available processes to produce pure carbon nanotubes (CNTs) and (ii) enhancement of existing techniques to fabricate better matrix materials paved the way to obtain high-performance composite materials with optimized mechanical properties. CNTs have been introduced into polymers, ceramics, cement-based materials and metals, the latter designated as metal matrix composites (MMCs). Today, many applications of CNT-reinforced composites exist, but CNT-reinforced MMCs are still scarce and only found in very specific applications. During the last decade, several works have been dedicated to improving the behavior of CNT-reinforced MMCs. This review described the state-of-the art in this field and highlighted the excellent and promising mechanical, thermal and electrical properties of CNT-reinforced MMCs.

References

|

75

Acknowledgement: The author gratefully acknowledges the consent given by authors of papers [15–17, 26, 29, 33, 37, 46, 66, 72, 78] and their publishers (Elsevier and Hindawi) to present some figures in this review. The author also acknowledges the financial support given by IDMEC/LAETA and FCT (Fundação para a Ciência e Tecnologia) in the context of the project “Modeling and Analysis of Nanostructures: Carbon Nanotubes and Nanocomposites” (PTDC/ECM/103490/2008).

References [1] [2] [3]

[4]

[5] [6] [7] [8] [9]

[10] [11] [12]

[13] [14] [15] [16] [17]

Iijima S. Helical microtubules of graphitic carbon. Nature 1991;354:56–58. Andrews R, Weisenberger MC. Carbon nanotube polymer composites. Current Opinion in Solid State and Materials Science 2004;8(1):31–37. Chou TW, Gao L, Thostenson ET, Zhang ZG, Byun JH. An assessment of the science and technology of carbon nanotube-based fibers and composites. Composites Science and Technology 2010;70(1):1–19. Yan J, Fan Z, Zhi L. Functionalized Carbon Nanotubes and Their Enhanced Polymers (chapter 8.17), in: Polymer Science: A Comprehensive Reference, Matyjaszewski K, Möller M (eds.). Amsterdam:Elsevier;2012, 439–478. ISI Web of Knowledge, 2013 (www.webofknowledge.com). Girot FA, Majidi AP, Chou TW. Metal Matrix Composites. Encyclopedia of Physical Science and Technology, Robert A. Meyers (ed.). New York, Academic Press, 2003:485-493. Curtin WA, Sheldon BW. CNT-reinforced ceramics and metals. Materials Today 2004;7:44–49. Li QQ, Rottmair CA, Singer RF. CNT reinforced light metal composites produced by melt stirring and by high pressure die casting. Composites Science and Technology 2010;70(16):2242–2247. Neubauer E, Kitzmantel M, Hulman M, Angerer P. Potential and challenges of metal-matrix-composites reinforced with carbon nanofibers and carbon nanotubes. Composites Science and Technology 2010;70(16):2228–2236. Bakshi SR, Lahiri D, Agarwal A. Carbon nanotube reinforced metal matrix composites – a review. International Materials Reviews 2010;55(1):41–64. Zhong R, Cong H, Hou PX. Fabrication of nano-Al based composites reinforced by single-walled carbon nanotubes. Carbon 2003;41(4):848–851. Cha SI, Kim KT, Lee KH, Mo CB, Hong SH. Strengthening and toughening of carbon nanotube reinforced alumina nanocomposite fabricated by molecular level mixing process. Scripta Materialia 2005;53(7):793–797. George R, Kashyap KT, Rahul R, Yamdagni S. Strengthening in carbon nanotube/aluminium (CNT/Al) composites, Scripta Materialia 2005;53(10):1159–1163. Deng CF, Zhang X, Ma Y, Wang D. Fabrication of aluminum matrix composite reinforced with carbon nanotubes, Rare Metals 2007;26(5):450–455. Deng C, Zhang XX, Wang D, Lin Q, Li A. Preparation and characterization of carbon nanotubes/aluminum matrix composites. Materials Letters 2007;61(8–9):1725–1728. Deng CF, Wang DZ, Zhang XX, Ma YX. Damping characteristics of carbon nanotube reinforced aluminum composite. Materials Letters 2007;61(14–15):3229–3231. Deng CF, Ma YX, Zhang P, Zhang XX, Wang DZ. Thermal expansion behaviours of aluminum composite reinforced with carbon nanotubes. Materials Letters 2008;62(15):2301–2303.

76 | 2 CNT-reinforced MMCs [18] Bakshi SR, Singh V, Balani K, McCartney DG, Seal S, Agarwal A. Carbon nanotube reinforced aluminum composite coating via cold spraying, Surface and Coatings Technology 2008;202(21):5162–5169. [19] Choi HJ, Kwon GB, Lee GY, Bae DH. Reinforcement with carbon nanotubes in aluminum matrix composites. Scripta Materialia 2008;59(3):360–363. [20] Pérez-Bustamante R, Estrada-Guel I, Antúnez-Flores W, Miki-Yoshida M, Ferreira PJ, Martínez-Sánchez R. Novel Al-matrix nanocomposites reinforced with multi-walled carbon nanotubes. Journal of Alloys and Compounds 2008;450(1–2):323–326. [21] Esawi AMK, El Borady MA. Carbon nanotube-reinforced aluminium strips. Composites Science and Technology 2008;68(2):486–492. [22] Esawi AMK, Morsi K, Sayed A, Abdel Gawad A, Borah P. Fabrication and properties of dispersed carbon nanotube-aluminum composites. Materials Science and Engineering 2009;A 508(1–2):167–173. [23] Esawi AMK, Morsi K, Sayed A, Taher M, Lanka S. Effect of carbon nanotube (CNT) content on the mechanical properties of CNT-reinforced aluminium composites. Composites Science and Technology 2010;70(16):2237–2241. [24] Choi HJ, Lee SM, Bae DH. Wear characteristic of aluminum-based composites containing multi-walled carbon nanotubes. Wear 2010;270(1–2):12–18. [25] Choi HJ, Shin JH, Min BH, Bae DH. Deformation behaviour of Al-Si alloy based nanocomposites reinforced with carbon nanotubes. Composites Part A: Applied Science and Manufacturing 2010;41(2):327–329. [26] Choi HJ, Bae DH. Creep properties of aluminum-based composite containing multi-walled carbon nanotubes. Scripta Materialia 2011;65(3):194–197. [27] Singhal SK, Pasricha R, Teotia S, Kumar G, Mathur RB. Fabrication and characterization of Al-matrix composites reinforced with amino-functionalized carbon nanotubes. Composites Science and Technology 2011;72(1):103–111. [28] Lipecka J, Andrzejczuk M, Lewandowska M, Janczak-Rusch J, Kurzydłowski KJ. Evaluation of thermal stability of ultrafine grained aluminium matrix composites reinforced with carbon nanotubes. Composites Science and Technology 2011; 71(16):1881–1885. [29] Choi HJ, Shin JH, Bae DH. Grain size effect on the strengthening behaviour of aluminum-based composites containing multi-walled carbon nanotubes. Composites Science and Technology 2011;71(15):1699–1705. [30] Jiang L, Li ZQ, Fan GL, Cao LL, Zhang D. The use of flake powder metallurgy to produce carbon nanotube (CNT)/aluminum composites with a homogenous CNT distribution. Carbon 2012a;50(5):1993–1998. [31] Jiang L, Li ZQ, Fan GL, Cao LL, Zhang D. Strong and ductile carbon nanotube/aluminum bulk nanolaminated composites with two-dimensional alignment of carbon nanotubes. Scripta Materialia 2012b;66(6):331–334. [32] Kwon HS, Leparoux M. Hot extruded carbon nanotube reinforced aluminum matrix composite materials. Nanotechnology 2012;23(41)415–701. [33] Liu ZY, Xiao BL, Wang WG, Ma ZY. Elevated temperature tensile properties and thermal expansion of CNT/2009Al composites. Composites Science and Technology 2012a;72(15):1826–1833. [34] Yoo SJ, Han SH, Kim WJ. Strength and strain hardening of aluminum matrix composites with randomly dispersed nanometer–length fragmented carbon nanotubes. Scripta Materialia 2013;68(9):711–714. [35] Yang XD, Liu E, Shi CS, He C, Li JJ, Zhao N, Kondoh K. Fabrication of carbon nanotube reinforced Al composites with well-balanced strength and ductility. Journal of Alloys and Compounds 2013;563:216–220.

References

| 77

[36] Chen X, Xia J, Peng J, Li W, Xie S. Carbon-nanotube metal-matrix composites prepared by electroless plating. Composites Science and Technology 2000;60(2): 301–306. [37] Su FH, Liu CS, Guo JL, Huang P. Characterizations of nanocrystalline Co and Co/MWCNT coatings produced by different electrodeposition techniques. Surface & Coatings Technology 2013;217:94–104. [38] Kim KT, Cha SI, Hong SH. Hardness and wear resistance of carbon nanotube reinforced Cu matrix nanocomposites. Materials Science and Engineering 2007;A: 449–451, 46–50. [39] Sun FW, Li H, Liew KM. Compressive mechanical properties of carbon nanotubes encapsulating helical copper nanowires. Carbon 2010;48(5):1586–1591. [40] Trinh PV, Trung TB, Thang NB, Thang BH, Tinh TX, Quang LD, Phuong DD, Minh PN. Calculation of the friction coefficient of Cu matrix composite reinforced by carbon nanotubes. Computational Materials Science 2010;49(4):S239–S241. [41] Uddin SM, Mahmud T, Wolf C, Glanz C, Kolaric I, Volkmer C, Höller H, Wienecke U, Roth S, Fecht HJ. Effect of size and shape of metal particles to improve hardness and electrical properties of carbon nanotube reinforced copper and copper alloy composites. Composites Science and Technology 2010;70(16):2253–2257. [42] Bhat A, Balla VK, Bysakh S, Basu D, Bose S, Bandyopadhyay A. Carbon nanotube reinforced Cu-10Sn alloy composites: Mechanical and thermal properties. Materials Science and Engineering 2011;A, 528(22–23):6727–6732. [43] Xu W, Hu R, Li JS, Fu HZ. Effect of electrical current on tribological property of Cu matrix composite reinforced by carbon nanotubes. Transactions of Nonferrous Metals Society of China 2011;21(10):2237–2241. [44] Cho SC, Kikuchi K, Kawasaki A, Kwon HS, Kim Y. Effective load transfer by a chromium carbide nanostructure in a multi-walled carbon nanotube/copper matrix composite. Nanotechnology 2011;23:315–705. [45] Shukla AK, Nayan N, Murty SVSN, Sharma SC, Chandran P, Bakshi SR, George KM. Processing of copper-carbon nanotube composites by vacuum hot pressing technique. Materials Science and Engineering 2013;A, 560:365–371. [46] Tsai PC, Jeng YR. Experimental and numerical investigation into the effect of carbon nanotube buckling on the reinforcement of CNT/Cu composites. Composites Science and Technology 2013;79:28–34. [47] Ding J, Liu BH, Dong ZL, Zhong ZY, Lin JY, White T. The preparation of Al2O3/M (Fe, Co, Ni) nanocomposites by mechanical alloying and the catalytic growth of carbon nanotubes. Composites Part B: Engineering 2004;35(2):103–109. [48] Golberg D, Mitome M, Müller Ch, Tang C, Leonhardt A, Bando Y. Atomic structures of iron-based single-crystalline nanowires crystallized inside multi-walled carbon nanotubes as revealed by analytical electron microscopy. Acta Materialia 2006;54(9):2567–2576. [49] Shpak AP, Kolesnik SP, Mogilny GS, Petrov YuN, Sokhatsky VP, Trophimova LN, Shanina BD, Gavriljuk VG. Structure and magnetic properties of iron nanowires encased in multiwalled carbon nanotubes. Acta Materialia 2007;55(5):1769–1778. [50] Pang LX, Sun KN, Ren S, Sun C, Fan RH, Lu Z. Fabrication and microstructure of Fe3Al matrix composite reinforced by carbon nanotube. Materials Science and Engineering 2007;A, 447(1–2):146–149. [51] Pang LX, Xing DJ, Zhang AQ, Xu J, Sun SS. Magnetic Properties of MWNTs-Fe3Al Composite. Advanced Materials Research 2009a;79–82:1351–1354. [52] Pang LX, Zhang JS, Xu J, Sun KN. Effect of adding carbon nanotubes on stress of Fe3Al intermetallics. Transactions of Nonferrous Metals Society of China 2009b;19(5):1201–1205. [53] Pang LX, Xing D, Zhang A, Xu J, Zhang JS, Fan RH. Carbon Nanotube Reinforced Intermetallic, Advanced Composite Materials 2010;19,3:261–267.

78 | 2 CNT-reinforced MMCs [54] Madaleno L, Pyrz R, Jensen LR, Pinto JJC, Lopes AB, Dolomanova V, Schjødt-Thomsen J, Rauhe JCM. Synthesis of clay-carbon nanotube hybrids: Growth of carbon nanotubes in different types of iron modified montmorillonite. Composites Science and Technology 2012;72(3):377–381. [55] Thakur SK, Srivatsan TS, Gupta M. Synthesis and mechanical behaviour of carbon nanotube-magnesium composites hybridized with nanoparticles of alumina. Materials Science and Engineering 2007;A, 466(1–2):32–37. [56] Li QQ, Viereckl A, Rottmair CA, Singer RF. Improved processing of carbon nanotube/magnesium alloy composites. Composites Science and Technology 2009;69(7–8):1193–1199. [57] Schaller R, Mari D, Marques dos Santos S, Tkalcec I, Carreño-Morelli E. Investigation of hydrogen storage in carbon nanotube-magnesium matrix composites. Materials Science and Engineering 2009;A:521–522, 147–150. [58] Aung NN, Zhou W, Goh CS, Nai SML, Wei J. Effect of carbon nanotubes on corrosion of Mg-CNT composites. Corrosion Science 2010;52(5):1551–1553. [59] Kondoh K, Fukuda H, Umeda J, Imai H, Fugetsu B, Endo M. Microstructural and mechanical analysis of carbon nanotube reinforced magnesium alloy powder composites. Materials Science and Engineering 2010;A, 527(16–17):4103–4108. [60] Srivatsan TS, Godbole C, Paramsothy M, Gupta M. Influence of Carbon Nanotubes and Processing on Cyclic Fatigue and Final Fracture Behaviour of a Magnesium Alloy. Advanced Materials Research 2011a;410:3–16. [61] Srivatsan TS, Godbole C, Paramsothy M, Gupta M. On the Influence of Carbon Nanotubes and Processing on Tensile Response and Fracture Behaviour of a Magnesium Alloy. Advanced Materials Research 2011b;410:133–141. [62] Srivatsan TS, Godbole C, Paramsothy M, Gupta M. Influence of nano-sized carbon nanotube reinforcements on tensile deformation, cyclic fatigue, and final fracture behaviour of a magnesium alloy. Journal of Materials Science 2012;47(8):3621–3638. [63] Park Y, Cho K, Park I, Park Y. Fabrication and mechanical properties of magnesium matrix composite reinforced with Si coated carbon nanotubes. Procedia Engineering 2011;10:1446–1450. [64] Fukuda H, Kondoh K, Umeda J, Fugetsu B. Fabrication of magnesium based composites reinforced with carbon nanotubes having superior mechanical properties. Materials Chemistry and Physics 2011a;127(3):451–458. [65] Fukuda H, Kondoh K, Umeda J, Fugetsu B. Interfacial analysis between Mg matrix and carbon nanotubes in Mg-6 wt.% Al alloy matrix composites reinforced with carbon nanotubes. Composites Science and Technology 2011b;71(5):705–709. [66] Zhou X, Su D, Wu CW, Liu LM. Tensile Mechanical Properties and Strengthening Mechanism of Hybrid Carbon Nanotube and Silicon Carbide Nanoparticle-Reinforced Magnesium Alloy Composites. Journal of Nanomaterials 2012:2012:851–862. [67] Sun FJ, Shi CS, Rhee KY, Zhao N. In situ synthesis of CNTs in Mg powder at low temperature for fabricating reinforced Mg composites. Journal of Alloys and Compounds 2013;551:496–501. [68] Wang LY, Tu JP, Chen WX, Wang YC., Liu XK, Olk C, Cheng DH, Zhang XB. Friction and wear behaviour of electroless Ni-based CNT composite coatings. Wear 2003; 254(12):1289–1293. [69] Scharf TW,;eira A,;wang, JY.;elf-lubricating carbon nanotube reinforced nickel matrix composites. Journal of Applied Physics 2009;106(1):013508. [70] Singh ARP,;wang JY,;charf TW.;ulk nickel-carbon nanotube nanocomposites by laser deposition. Materials Science and Technology 2010;26(11):1393–1400. [71] Liu B,;iu LR,;iu XJ. Variation of crystal structure in nickel nanoparticles filled in carbon nanotubes, Materials Science and Technology 2012b;28(11):1345–1348. [72] Hwang JY, Lim BK, Tiley J, Banerjee R, Hong SH. Interface analysis of ultra-high strength carbon nanotube/nickel composites processed by molecular level mixing. Carbon 2013;57:282–287.

References

|

79

[73] Xue F, Jiehe S, Yan F, Wei C. Preparation and elevated temperature compressive properties of multi-walled carbon nanotube reinforced Ti composites. Materials Science and Engineering 2010;A 527(6):1586–1589. [74] Feng X, Sui J, Cai W. Processing of multi-walled carbon nanotube-reinforced TiNi composites by hot pressed sintering. Journal of Composite Materials 2011a; 45(15):1553–1557. [75] Feng X, Sui JH, Cai W, Liu AL. Improving wear resistance of TiNi matrix composites reinforced by carbon nanotubes and in situ TiC. Scripta Materialia 2011b;64(9):824–827. [76] Cai W, Feng X, Sui J. Preparation of multi-walled carbon nanotube–reinforced TiNi matrix composites from elemental powders by spark plasma sintering. Rare Metals 2012;31(1):48–50. [77] Kondoh K, Threrujirapapong T, Umeda J, Fugetsu B. High-temperature properties of extruded titanium composites fabricated from carbon nanotubes coated titanium powder by spark plasma sintering and hot extrusion. Composites Science and Technology 2012;72(11):1291–1297. [78] Li SF, Sun B, Imai H, Mimoto T, Kondoh K. Powder metallurgy titanium metal matrix composites reinforced with carbon nanotubes and graphite. Composites Part A: Applied Science and Manufacturing 2013;48:57–66.

Yue Li, Xiaoxing Cheng, Liqiang Wang, Weijie Lu, Jining Qin, Fan Zhang, Di Zhang

3 Novel preparation and mechanical properties of in situ synthesized (TiB+La2 O3 )/TiNbTaZr composites Abstract: This article reports the application of rare earth elements in β titanium alloys. The mechanical properties and microstructure of (TiB+La2 O3 )/TiNbTaZr composites are studied systematically. TiB and La2 O3 reinforcements, which came out through the in situ reaction between LaB6 and Ti matrix, plays an important role in the strength and superelasticity of the (TiB+La2 O3 )/TiNbTaZr composite. Therefore, we highly recommend the novel method of preparing biomedical titanium matrix composites by adding rare earth elements.

3.1 Introduction 3.1.1 The application of rare earth elements in β titanium alloys The first attempt at using rare earth elements in titanium alloys was back in the 1950s, and the awareness of the application of rare earth elements in titanium alloys has kept increasing since then, especially in the field of high temperature titanium alloys. Table 3.1 gives some examples of applications of rare earth in titanium alloys. Because of lazy chemical properties of rare earth elements (RE), it is common that RE form stable compounds with other elements such as RE2 O3 .The study of a small amount (< 1 wt.%) of RE2 O3 in titanium alloys shows that the rare earth oxide particles, in general, are spheres and ellipsoids [1–2].

3.1.2 The influence of rare earth elements in titanium alloys One major effect of rare earth elements in titanium alloys is to refine grain size. In the study of CuZnAl with La+Ce (< 0.1 wt.%), rare earth composites were found to segregate at the CuZnAl alloy grain boundaries, and hence hindered grain growth. It was reported that the grain size decreased largely with 0.1 wt.% Y addition in Ti-6Al-4V alloy [4]. By comparing the microstructure of BT-5 with the addition of La, Ce, Pr, Nd, Gd, Y, Tb and Er respectively, they observed that La, Ce and Pr were the most effective elements in refining grains. As the concentration of rare earth elements increases, the grain boundaries are purified. With fewer impurities, the grain boundary affected zone shrinks, which is beneficial to the resistance of alloy martensitic transformation, re-

82 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr Table 3.1. Applications of rare earths in titanium alloys [3]. Titanium alloys

Rare earth additions

α

Pure Ti Ti-5Al-2.5Sn

La, Y, Ce, Nd, Er Rare earth mixture Y

Near α

Ti-633G 7715C Ti-55 IMI829

Gd Ce Nd Y

α +β

Ti-6Al-4V Ti-5Al-5Mo-2Sn Ti-6242S

Gd, Er, Y Ce Y

β

Ti-15-3 Ti-3Al-7Mo-11Cr

Y, Er Dy

verse martensitic transformation and the resistance of the alloy. At the same time, as the ordering degree of the alloy is improved, this somehow hinders plastic deformation. Consequently, the alloy presents good crystal reversibility during reverse transformation, while plasticity worsens. However, when the concentration of rare earth elements exceeds a certain value, the density of relatively fine grain boundaries increases, and the martensitic transformation consumes much more irreversible energy when the martensite phase moves across the grain. This would thus deteriorate the ordering which would decrease the shape memory effect of the alloy. Rare earth elements are strong deoxidizers. In the alloy, they effectively capture the oxygen and greatly improve the tensile strength and elongation. The type of reinforcement of RE is mainly fine grain strengthening, and the type of softening is interstitial solute strengthening. Rare earth oxides generally have a high melting point; when evenly distributed in the alloy, they could increase the elevated temperature instantaneous tensile strength and creep strength. In addition, they also have great importance for improving fatigue properties and thermal stability. Former studies on the effects of rare earth element additives on shape memory effects and superelasticity focus on FeMnSiCrNi [5] and CuZnAl [6], while few focus on the shape memory effects of β titanium alloys.

3.1.3 Biosafety of rare earth elements The toxicity of heavy metal elements is determined by their elemental physical and chemical properties. The toxicity of rare earth elements can be divided into two categories by comparing and sorting the ten physical properties of 51 heavy metal elements. Those properties include the first ionization potential, melting point, evapora-

3.1 Introduction

|

83

tion point, melting enthalpy, evaporation enthalpy, electrochemical equivalent, cohesive energy, ionic radius, density and charge-ionic radius ratio. Based on comparison, praseodymium, cerium and lutetium are elements with moderate toxicity, while most of the rare earth elements have a low toxicity. In particular, the elements La and Y have a lower biotoxicity than Zr, Nb, Ta, which are currently used to stabilize the β phase in titanium alloys [7].

3.1.3.1 TiB reinforcement in titanium alloys Suzuki et al. [8] studied the influence of B additions in Ti-Td-Ni high-temperature shape alloys. They found that TiB2 , which was the reactant of B and Ti, played an important role in improving the alloy’s strength and plasticity, while it made little contribution in improving the shape memory effect. Lv et al. [9] used a vacuum arc furnace to prepare TiB whiskers. The cross section of the whisker is hexagonal with the stacking faults on the (100) surface. In addition, the research observed the following orientation relationship: – [010] TiB//[01-10] Ti, (100) TiB//(-2110) Ti, – (001) TiB//(0002) Ti, (10-1) TiB//(4-2-21) Ti; – [001]TiB//[01-10]Ti, (010)TiB//(-2110)Ti, – (201) TiB//(0002) Ti. Currently, β titanium alloys with non-toxic elements such as Nb, Zr, Sn, etc. are widely applied in the biomedical field [10–15]. However, due to the low critical slipping stress of this type of alloy, plastic deformation occurs before the martensitic phase transformation and the shape memory effect in the alloy is not very obvious. As a result, by increasing the critical slipping stress, visible martensitic phase transformation would appear before plastic deformation with the aim of improving the shape memory effect and superelasticity of the β titanium alloy. In order to increase the critical slipping stress of the alloy, certain reinforcements have been considered. It has been reported that N, B and Ce could improve shape memory properties and superelasticity [16–18]. Rare earth elements are effective in refining grains and improving the overall mechanical properties including shape memory properties. However, rare earth elements are in general chemically reactive, and oxidize easily in air, which more or less brings difficulties in controlling the in situ reaction. To obtain microstructures suitable for practical applications, our research controlled the microstructure by changing the addition method of rare earth elements. Yang studied the mechanical properties of multi-reinforcements alloy (TiB+TiC+ La2 O3 )/Ti at room temperature and high temperature, respectively. He found that (TiB+TiC+La2 O3 )/Ti had better tensile strength and plasticity at both room and high temperature. This indicates that rare earth elements play an important role in improving the mechanical properties of titanium composites [19]. Compared to the traditional method, the in situ synthesis technique guarantees the specimen with

84 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr better mechanical properties, and it is consequently widely used in manufacturing titanium matrix composites [20]. TiB prepared during this processing has a better coherence with the titanium matrix in the interfaces. The rare earth oxide Re2 O3 also improves mechanical properties [21–23]. So far, there is little literature published that reports the influence of multi-reinforcements on the shape memory properties by in situ synthesis techniques. Based on the above discussion, we employed the in situ synthesis to produce TiB and Re2 O3 to improve the alloy’s shape memory properties and super elasticity. The matrix material is Ti-35Nb-3Zr-2Ta, as discussed earlier in this dissertation. In the experiments, the weight percent of LaB6 powders were 0.1 %, 0.2 %, 0.3 %, 0.4 % and 0.5 %, respectively. The in situ synthesis reaction is shown below as 12Ti + 2LaB6 + 3[O] = 12TiB + La2 O3 . We studied the influence of the concentration of reinforcements on the structure and properties of matrix alloy; this contributed to the application of the in situ synthesis in the β titanium alloy.

3.2 Materials preparation and experimental procedures 3.2.1 Materials preparation In the preparation, we employed the vacuum consumable arc furnace to produce TiNbZrTa alloy. LaB6 was added with a concentration of 0.1 %, 0.2 %, 0.3 %, 0.4 % and 0.5 % (wt.%) to the matrix alloy respectively. The mixed specimens were melted homogeneously in a consumable vacuum arc remelting furnace. We repeated the process four times to guarantee homogeneity. Finally, the samples were casted to 70 mm ×30 mm ×10 mm slates. The square samples had smooth surfaces, and no considerable inclusions and pores were observed. The samples were then hot rolled to form slates with the thickness of 0.5 mm at 950 ∘ C followed by heat treatment at 780 ∘ C for 0.5 h.

3.2.2 Experimental procedures 3.2.2.1 Microstructure observation An MX41M microscope was employed to observe the metallographic structure. The microstructure and the fracture were observed by JSM-6700F scanning electron microscope. EDS was employed to measure the oxygen concentration in the alloy. Twinjet electro-polishing was used to prepare TEM samples. The formula of the electrolyte was 64 % methanol + 30 % normal butanol + 6 % perchloric acid. The working voltage was 35 V, and the current was 10 mA. The temperature of the electrolyte was kept −40 ∘ C

3.3 Results and discussions

| 85

by adding liquid nitrogen. The samples were observed by a JEM-2000EX transmission electron microscope. The size of the rectangular TEM samples prepared traditionally was 0.05 μm in height, 9 mm and 3 mm in length and width, respectively, and the tensile direction was parallel to its length. Figure 3.1 illustrates the in situ tension samples. The operating voltage of TEM was 160KV, and the maximum load of tension was 500 g. h

σ

W

σ

Fig. 3.1. Sketched map of the tensile specimen used for the in situ studies performed in the TEM.

3.2.2.2 Phase analysis The phase analysis was conducted on D/max 2550 V automatic x-ray diffractometer, using Cu Kα 1 with the working voltage of 40 kV and current around 450 mA.

3.2.2.3 Mechanical properties tests Room temperature mechanical properties and superelasticity were measured using slates samples by a Zwick T1-Fr020TN drawing mill, and the drawing speed was 1.5 × 10−4 s−1 . The superelasticity was tested by the loading-unloading approach controlled by strain.

3.3 Results and discussions 3.3.1 Phase analysis Figure 3.2 shows the x-ray diffraction pattern of the (TiB+La2 O3 )/Ti composite material with different LaB6 content. Clearly, the peaks of β -Ti and TiB appear, which is due to the TiB and La2 O3 produced through the in situ reaction between LaB6 and the matrix.

3.3.2 Thermodynamic analysis In our research, the composite is mainly obtained during in situ reaction of 12Ti + 2LaB6 + 3[O] = 12TiB + La2 O3 . According to thermodynamics, enthalpy, entropy and

86 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr

Intensity/a.u

β TiB La2O3

(e) (d) (c) (b) (a) 20

30

40

50 60 2 Theta/degree

70

80

Fig. 3.2. XRD pattern of (TiB+La2 O3 )/Ti composites with different mass fractions of LaB6 : (a) 0.1 %; (b) 0.2 %; (c) 0.3 %; (d) 0.4 %; (e) 0.5 % [24].

constant pressure heat capacity are related by the equations dH = Cp dT, dS =

(3.1)

dH nCp dT = , T T

(3.2)

in which, H, S and Cp represent enthalpy, entropy and constant pressure heat capacity, respectively. Since phase transformation may occur during the heating process, the term for latent heat should be added to the calculated process (ΔHm represents latent heat). Rewrite the above two equations in the integral form as follows: Tm

T

H = H0 + ∫ Cp1 dT + ΔHm + ∫ Cp2 dt, T0

(3.3)

Tm

Tm

T

ΔHm S = S0 + ∫ Cp1 d ln T + + ∫ Cp2 d ln T, Tm T0

(3.4)

Tm

in which Cp1 and Cp2 are the heat capacity before and after phase transformation, respectively, and Tm is the transformation temperature. The Gibbs free energy could be expressed by enthalpy and entropy using the equation G = H − TS.

(3.5)

3.3 Results and discussions

| 87

For a chemical reaction aA + bB = cC + dD, the enthalpy of formation ΔH and Gibbs free energy ΔG could be derived using equations (3.3), (3.4) and (3.5): Tm

T

ΔH = H0 + ∫ Cp1 dt + ΔHm + ∫ Cp2 dt, T0

(3.6)

Tm

ΔG = ΔH − TΔS,

(3.7)

and constant pressure heat capacity: Cp = a + b ⋅ 10−3 T + c ⋅ 10−6 T −2 + d ⋅ 10−6 T 2 .

(3.8)

According to the second law of thermodynamics, any spontaneous chemical reaction must be an entropy-increasing or free-energy-decreasing process. Since the criteria of changing entropy is only valid for an isolated system, usually the changes of free energy are used for determining whether a reaction can happen spontaneously or not. The criteria of in situ reactions are: – ΔG < 0, a reaction appears spontaneously; – ΔG = 0, a reaction reaches equilibrium; – ΔG > 0, a reaction is thermodynamically impossible. Whether the in situ reaction happens or not is the key factor that determines the possibility of the in situ synthesis of Ti matrix composite materials. Such questions could be answered by calculating Gibbs free energy. When the changes in Gibbs free energy are less than zero, the reaction happens spontaneously, and the smaller the ΔG is, the more easily the reaction will occur. Gibbs free energy calculation of the in situ reaction is shown in Figure 3.3. As the ΔG is less than zero, a reaction will happen spontaneously.

3.3.3 Microstructure analysis The changes of grain size with the varying of LaB6 content in TiNbTaZr are shown in Figure 3.4, in which the etched metallographic microstructure of composites with an addition of LaB6 at 0.1 %, 0.3 %, 0.5 % (wt.%) are compared. For the specimen with 0.1 %, the whisker form of TiB reinforcements could be observed at the grain boundaries, which is presented in the zoomed in Figure 3.4a. As the percentage of LaB6 increases to 0.3 % and 0.5 %, the grain size decreases from 32 μm to 28 μm and 12 μm. La2 O3 , the product of the in situ reaction, hinders the growth of the grains and thus effectively refines the matrix’s microstructure. Such an effect becomes more and more remarkable with the increase of LaB6 content, which results in the increase of La2 O3 content (Figure 3.4c). Figure 3.5 is the SEM image of the reinforcement phase

88 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr

˗3.0

ΔG (J/mol)

˗3.2 ˗3.4 ˗3.6 ˗3.8 ˗4.0 ˗4.2 0

500

1000 1500 2000 Temperature (K)

30 um (a)

2500

Fig. 3.3. Calculation value of Gibbs energy during in situ reactions.

30 um (b)

30 um (c)

Fig. 3.4. Optical micrographs of the composites with different mass fraction of LaB6 : (a) 0.1 %, (b) 0.3 %, (c) 0.5 % [25].

70 μm (a)

70 μm (b)

10 μm (c)

Fig. 3.5. SEM images of the distribution of reinforcements in the composite of 0.1 % and 0.5 % LaB6 additioned specimens: (a) 0.1 %, (b) and (c) 0.5 % [25].

within the composite material at 0.1 % and 0.5 % of LaB6 . Figure 3.5 (a) shows the TiB whisker at the grain boundary. Figure 3.5b shows even more of the TiB whisker at the grain boundary as the amount of LaB6 increases to 0.5 %, while at the same time more

3.3 Results and discussions

|

89

spheroidal La2 O3 particles can be observed in the zoomed-in image of Figure 3.5b, as it is specified by arrows in Figure 3.5c.

3.3.4 Microstructure of reinforcements The microstructure of TiB is B27 [26], as illustrated in Figure 3.6a; each boron atom is at the center of a prism made up of six titanium atoms, and the boron atoms are arranged in a zigzag pattern in the b axis ([010] direction). The prism is thus stacked to form a B27 structure [27–28]. Due to the low ionization potential of boron, only single covalent bonds are formed between boron atoms, thus creating a zigzag of a single chain of boron atoms parallel to the b axis, with a boron atom at the center of the prism made up of six Ti atoms, shown in Figure 3.6b. Figure 3.6c and the shaded area in Figure 3.6d show that when TiB is grown from a solution, it is always confined by the slowest growing plane. It is thus prone to forming a certain morphology: four pillars containing B atoms surrounding a parallelogram, along the [010] direction. Li [30] used the reactive hot pressing method to synthesize the TiB-reinforced Ti matrix composite material, and results show that TiB was in a needle shape with a hexagonal cross sectional plane. Stacking faults exist in the (100) direction of the TiB whisker, while the interface between the TiB and Ti alloy matrix is smooth with no interfacial reaction. Ranganath et al. [31] also found a hexagonal cross sectional plane in the TiB whisker when using the combustion assisted casting (CAC) method to prepare TiB+Ti2 C/Ti composites. Ma et al. [32] used the reactive hot pressing method to explore the Ti-B, Ti-TiB2 , Ti-B4 C and Ti-BN systems. In all of them, TiB comes in the form of whiskers. Lv et al. [33] used the vacuum-consuming arc melting method to obtain the TiB whisker, with a hexagonal cross sectional plane and stacking fault in the (100) plane. At a high temperature, La2 O3 has the structure of Mn2 O3 , the space group of Ia3, a cubic system, a = 11.056 nm and each unit cell has 16 La atoms and 24 O atoms; while at room temperature, La2 O3 has a crystal structure of D52 , space group P-3m1, a trigonal system, a = 0.393 nm, c = 0.612 nm and each unit cell contains 2 La atoms and 3 O atoms [34]. Since La2 O3 nucleates and grows at high temperature, it belongs to the cubic group with a high symmetry; its growth rate is thus similar in all directions and finally becomes spherical. Then as the temperature decreases to room temperature, La2 O3 transforms into a D52 structure. Figure 3.7 shows the bright field image and electron diffraction pattern of TiB and La2 O3 in the 0.4 % LaB6 composite. The diameter of La2 O3 reinforcement is around 100 nm with a clean interface to the matrix. The electron diffraction patterns are shown in Figure 3.7b. Stacking faults can easily form on the (100) plane, parallel to the [010] direction and extended through the whole reinforcement phase. The TiB stacking fault on the

90 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr

a b

c

– Boron atom – Metal atom

(b)

(a)

(c)

c

a

(d)

Ti:1/4

B:1/4

Ti:3/4

B:3/4

Fig. 3.6. Schematic illustration of the construction of the crystal structure of TiB: (a) primitive trigonal prism; (b) columns of prisms; (c) construction of the TiB structure; (d) projection of the TiB structure of the (010) plane [29].

3.3 Results and discussions

Fig. 3.7. TEM images of composites with 0.4 % LaB6 addition: (a) bright field of TiB and La2 O3 ; (b) SAD of β -Ti, TiB and La2 O3 [24].

100nm (a)

| 91

(b)

(100) plane is caused by a lack of B atoms in the plane, resulting in a rearrangement of atoms.

3.3.5 Analysis of the solidification mechanism In the solidification of the (TiB+La2 O3 )/Ti composite, no compound is formed between Ti and La [35]. Based on the results obtained from the Ti-B phase diagram in Figure 3.8 [36], the amount of B in our experiment is chosen to be less than the Ti/TiB eutectic concentration. As the temperature decreases, at first, the primary La2 O3 phase comes out after the reaction between La and O. Then β -Ti nucleates and grows. TiB appears below the eutectic point, and La precipitates out at the same time due to the Weight percent Boron 0

6

10

20

30

40

50 60 7080 100

3225±25°C L

2500 ~2200°C 2080±20°C

2092°C

2000 1670°C 1500

1540±10°C Ti3B4

Temperature °C

3000

(βTi) 1000

(βB) TiB2

TiB

884±2°C (αTi)

500 0

Ti

10

20

30

40

50

60

70

Atomic percent Boron

80

90

100

B

Fig. 3.8. A phase diagram of Ti-B [36].

92 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr L

La2O3(p)+L1

L2+βTi

βTi+TiB

βTi+La2O3(s)

Fig. 3.9. path of (TiB+La2 O3 )/Ti composite.

strong oxidation tendency of the La element robbing O from the Ti addition to form the nanometer scale La2 O3 phase. The sequence of reinforcement transformation in the (TiB+La2 O3 )/Ti composite is shown in Figure 3.9.

3.3.6 Superelasticity Figure 3.10 is the stress-strain curve of tested specimens. The superelasticity of the sample is characterized by the superelastic strain εSE and elastic strain εE . During the loading period, after elastic deformation, the composite undergoes martensitic transformation at the critical stress σc , which is shown in Figure 3.10a. The unloading part of the curve is made up of elastic and non-linear processes while the recovery strain of unloading includes the elastic strain εE and those caused by martensitic transformation. During the unloading process, the transformation from martensite α 󸀠󸀠 to β -Ti happens at the σr point. Figure 3.10b and (c) show the strain-stress curve of 4 % and 5.5 % loading strain, respectively. Figure 3.11 illustrates the relationship between LaB6 content and superelastic characteristic strains at 2.5 %, 4 % and 5 % loading strain. Clearly, when the content of LaB6 is 0.1 % and 0.2 %, both the superelastic strain and elastic strain are relatively large for all three loading situations. With the increase of LaB6 content, the superelastic strain gradually decreases, in the 0.4 % LaB6 sample; its value is around 0.72 % to 0.9 %. But when the LaB6 content reaches 0.5 %, in the 2.5 % and 5.5 % loading strain cases, the superelastic strain increases to 0.95 % and 1.15 % respectively. For the elastic strain’s value in Figure 3.11b, when the amount of LaB6 is greater than 0.2 %, pure elastic strain decreases; in the 0.5 % LaB6 sample, when the loading strain is 5.5 %, pure elastic strain decreases to 0.69 %. The increase in the superelastic strain comes from the increase of recovery rate of martensite when unloading, as LaB6 reacts with the matrix and produces TiB and

3.3 Results and discussions

| 93

500

Stress, σ/MPa

400 300 σc

σr

200 0.0% 0.1% 0.2% 0.3% 0.4% 0.5%

100 0 0 (a)

1

2

3

5 6 7 8 4 ƐSE ƐS Strain, Ɛ/(%)

9

500

Stress, σ/MPa

400 300 200 0.1% 0.2% 0.3% 0.4%

100 0 (b)

Strain, Ɛ/(%)

1.0%

500

Stress, σ/MPa

400 300 200 0.0% 0.1% 0.2% 0.3% 0.4% 0.5%

100 0 (c)

Strain, Ɛ/(%)

2.0%

Fig. 3.10. Stress-strain curves obtained by cyclic loading–unloading tensile tests for the composites with different mass fraction of LaB6 at strain of 2.5 %, 4 % and 5.5 %, respectively: (a) 2.5 %; (b) 4 %; (c) 5.5 % [25].

La2 O3 , which hinder the growth of grains. When the amount of LaB6 is small, the grain refinement effect is not very obvious, since not enough reinforcement is formed at the grain boundary as an impedance, but LaB6 purifies the material and strengthens the grain boundary. As the grains size is not very small, the effect of grain boundaries is limited, resulting in a low resistance to the martensitic transformation and

1.4 1.3 1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5

2.5 4.0 5.5

0.85 0.80 0.75 0.70 0.65

0.0 (a)

0.90

2.5 4.0 5.5

Strain, ε/(%)

Strain, ε/(%)

94 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr

0.1

0.2 0.3 0.4 LaB6 /(mass%)

0.60

0.5 (b)

0.0

0.1

0.2 0.3 0.4 LaB6 /(mass%)

0.5

Fig. 3.11. Relationships between the mass fraction of LaB6 and the superelastic characterization extracted from the stress–strain curves at strain of 2.5 %, 4 % and 5.5 %: (a) superelastic strain; (b) pure elastic strain [25].

its reversal, thus forming a better crystallographic reversibility of the alloy. Additionally, the TiB whisker produced by the in situ reaction increases its critical shear stress, which is beneficial for martensitic transformation before the slip deformation. When the contents of LaB6 are larger than 0.2 % – at 0.3 % and 0.4 % – the O content in matrix is reduced due to the formation of La2 O3 , undermining the alloy’s superelastic properties. Previous research has shown that oxygen content in the Ti-Nb alloy is a key factor that influences the superelastic property of the material [37]. In the case of 2.5 % and 5.5 % loading strain, the samples with 0.5 % LaB6 both reach the highest superelastic strain in the unloading process. This is based on the analysis above on the refined grain size for a 0.5 % LaB6 sample. On the one hand, refined grains strengthen the material and increase the critical shear stress as well as the extent of martensitic transformation before slip deformation. On the other hand, too much LaB6 results in more precipitation and destroys the continuity of the martensite phase. Dislocations will aggregate around the precipitate particles; this increases the pinning effect of the martensite phase interface, causing higher phase transformation resistance. In addition, more LaB6 content increases the amount of rare earth dissolved in the alloy’s lattice, which distorts the lattice and decreases the degrees of order, thus creating less shape memory effect. Xu et al. [38] studied the relationship between grain size and recovery stress in polycrystalline material. The grain boundary region has high defect density and a low degree of order, thus yielding happens at the grain boundaries first, leading to a much lower recovery rate than inside the grain. In most metallic matrix composite materials, the expansion coefficient of reinforcement is smaller than that of the matrix, resulting in higher tensile stress and yielding strength in the composites, thus improving the shape memory effect. Based on the principle of shape memory effect, particles in the particle reinforced shape

3.3 Results and discussions

Matrix

Martensite

Martensite

|

95

Austenite

Particles

(a)

(b)

(c)

(d)

Fig. 3.12. Schematic illustration of reinforced particles in a shape memory alloy [38]: (a) composite; (b) martensite; (c) deformation; (d) heat treatment.

memory alloy can be self-reinforced. Figure 3.12 illustrates the mechanism of particle self-reinforcement. Spherical particles are uniformly distributed in the Ti alloy matrix; they are connected with the matrix and form an elastic entity. Under usage temperature, the particles are in their parent phase; when temperature decreases below the martensitic transformation temperature, particles transform into the martensite phase, as is shown in Figure 3.12b. When deforming the sample by applying a tensile stress, re-straining appeared in the particles; this is shown in Figure 3.12c. If one heats the sample above austenite transformation temperature after being unloaded, the martensite phase transforms into the austenite phase, and the pre-strained particles recover to a spherical shape, as is shown in Figure 3.12d. Due to the constraints from the matrix, the particles cannot fully recover to a sphere; compressive strain exists in the applied tensile stress’ direction, thus improving the material’s ability to resist tensile strain and strengthen the material.

3.3.7 In situ characterization of microstructure Generally, the martensite in β phase comes from the shearing deformation of the parent phase, which affects the shape memory effect of the alloy. In situ TEM technique is used to observe the dynamics of dislocations and martensite morphology under external stress. Cai [39] studied the features of reverse martensite transformation during heating in the NiTi shape memory alloy, and analyzed the martensite stability of different morphologies as well as the process of reverse transformation to the parent phase. Research of martensite transformation for the Ni44.7 Ti46.3 Nb9 material shows that during heating, the parent phase nucleates at the martensite grain boundaries and martensite variants interfaces and then grows through the migrating of parent/martensite interface. This chapter shows the usage of in situ TEM technique in characterizing the changes of microstructure of TiB+La2 O3 /Ti composite material

96 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr under tensile stress, and focuses on the dislocation, twinning, martensite phase’s nucleation and growth during the deformation. Figure 3.13 shows the result of (TiB+La2 O3 )/Ti under TEM during the elongation process. Strain is used for specifying the degrees of stretching, and the microstructure when the amount of deformation equals 0.02 mm, 0.04 mm, 0.06 mm, 0.08 mm and 0.10 mm is observed. Figure 3.13a shows the sparse distribution of dislocation lines on the surface of the composite material under 0.02 mm; no martensite α 󸀠󸀠 phase could be observed. With the increase of deformation, the dislocation density

45°

400nm

400nm (a)

(b)

150nm (c)

(d)

(111)

(001) (000) GB 400nm (f)

(e)

500nm (g)

(h)

(111)

(001) (000)

500nm (i)

(j)

Fig. 3.13. TEM images taken during a strain-controlled tensile experiment at strain levels of: (a) ε = 0.02 mm, activated screw dislocations; (b) ε = 0.04 mm, the propagation of screw dislocations and stress-induced martensite phase in the deformed sample; (c) high-magnification images of the region indicated by the black arrow in (b); (d) the corresponding SAED pattern with the reflections from variants of stress-induced martensite phase;(e) the illustration pattern of (d); (f) ε = 0.06 mm, martensitic α 󸀠󸀠 -phase accompanying with rectilinear and aligned dislocations with their screw direction, indicated by the black arrows; (g) ε = 0.08 mm, the propagation of martensitic α 󸀠󸀠 -phase; (h) the corresponding SAED pattern with the reflections from variants of the α 󸀠󸀠 phase, indicated by the black arrow; (i) the illustration pattern of (h); (j) ε = 0.10 mm, dislocation loop indicated by black arrows and the propagation of the α 󸀠󸀠 -phase.

3.3 Results and discussions

|

97

increases and dislocations arch out in the direction of drawing, and parallel plate strain-induced martensite phases appear. Figures 3.13b and c are the morphology of the self-accommodated plate form martensite under higher magnification. A lot of (001) twinning composite thin plates are formed in the {111}̄ twinning plate, whose twinning interfaces are in a step pattern. The figures show that the {111}̄ twinning structure is formed through the nucleation and growth of the (001) composite twin. In the starting session, (001) composite twins nucleate through the slip along (001) plane inside the {11-1} twin. As the stress increases, more (001) composite twins nucleate and grow. During the drawing process, we take the tensile stress direction as the Y-axis, one of the horizontal directions as the X-axis and the normal direction of the sample’s thickness as Z-axis. According to the Tresca yielding criteria, the maximum stress plane is 45 degrees to the XY and XZ plane, and the stress is τmax = σy /2, as it is shown in the shear stress model in Figure 3.14. Martensitic transformation occurs under the shear stress, the extending direction of the self-accommodated martensite phase is 45 degrees to the external force’s direction. Figure 3.13f is the deformation microstructure after 0.06 mm deformation, in which dislocations become much longer and arch out in the drawing direction. The strain induced martensite phase is intertwined with the dislocation and extends further with increasing strain. When the amount of deformation reaches 0.08 mm, the morphology of martensite and the crystallographic relationship between martensite and the matrix are still the same, though the strain-induced martensite has extended through the grain boundary. Figure 3.13j is the microstructure of 0.10 mm. During the drawing process, {111}̄ type one twin martensite is formed due to the elastic effect between martensite ̄ phases. {111}type twin has higher twin-induced shear stress compared with ⟨011⟩ type two twin, thus in the {111}̄ twinning sample, the elastic energy is gradually reduced through the collision between martensite variants. The in situ experiment shows that, before the martensite transformation dislocations nucleate and grow, and then induce ̄ the martensite nucleation and self-accommodated twins. {111}twin and (001) composite twin could be seen throughout the whole drawing process. The (001) composite twin grows larger as the amount of deformation increases, and eventually has a simī lar size as the {111}twin.

3.3.8 Mechanical properties Figure 3.14 shows the relationship between the amount of LaB6 and the mechanical property of the (TiB+La2 O3 )/Ti composite material. The samples that contain LaB6 have higher strength than the normal Ti-Nb-Zr-Ta β titanium alloy. The ultimate tensile strength reaches a maximum value of 580 MPa at 0.1 % LaB6 , and gradually decreases as the content of LaB6 increases. For a content between 0.2 % and 0.4 %, the UTS value varies between 533 MPa to 554 MPa, when the LaB6 content reaches 0.5 %, the UTS decreases to 526 MPa. The strengthening mechanism is mainly affected by

98 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr

30 600 28 550 26 500 24 450

YS UTS EL

400

22 20

350

Elongation, EL / %

Yeild strength, YL / MPa

Ultimate tensile strength, UTS / MPa

650

18 300 250

16 0.0

0.1 0.2 0.3 0.4 Amount of added LaB6 / mass%

0.5

Fig. 3.14. Mechanical properties of TiB+La2 O3 /Ti composites with different mass fractions of LaB6 [24].

the following factors: load on the reinforcement, the strength of the matrix and the dispersive strengthening of La2 O3 nanoparticles. According to the models of Cox et al. [40] and Nardone, Prewo et al. [41–42], the strength of particle and whisker reinforced composite material could be calculated using the following formulas: σcw = σm ⋅ 0.5Vw (2 + σcp = σm ⋅ Vp (1 +

l ) + σm (1 − Vm ), d

(L + t)A ) + σm (l − Vp ), 4L

(3.9) (3.10)

in which σcw , σcp and σm are the strength of the whisker, the yielding strength of the particle and strength of the matrix alloy respectively, Vw and Vp are the volume fraction of whisker and particle respectively, l/d is the length radius ratio of the whisker, L is the length of the particle in the direction perpendicular to applied stress, A is the particle’s length radius ratio. The total strength of the composite material could be expressed in such way: σcomposites = Σ Vi σi = VTi σTi + VTiB σTiB + VLa2 O3 σLa2 O3 ,

(3.11)

in which Vi and σi represent the volume fraction and strength of the corresponding component. In the (TiB+La2 O3 )/Ti composite material, La2 O3 is in the form of a particle, while TiB is in the form of a whisker, thus the yield strength formula for the composite material is ΣTMCs = Vm σm + 0.5VTiB σm (2 +

(LLa2 O3 + t) ALa2 O3 1 ) ) (1 − d 4LLa2 O3

(3.12)

The matrix alloy’s strength in the above formula could be calculated using Hall– Petch equation [43]: σm = σ0 + km d−1/2 , (3.13)

3.4 Conclusions

| 99

in which σm is the yielding stress, σ0 is the true flow stress, km is the Hall–Petch coefficient and d is the size of the grain. Based on the former analysis, we know that the matrix alloy has the smallest grain size for 0.5 % content of LaB6 , thus when the grain size of (TiB +La2 O3 )/Ti composite material decreases to 12 μm, the yielding strength increases to 327 MPa; such a phenomenon is in agreement with the Hall-Petch equation. When the content of LaB6 is 0.1 %, the composite material exhibits high tensile strength due to the reinforcing effect of TiB whisker. As the amount of LaB6 increases, more La2 O3 particles cause the O concentration in the β phase matrix to decrease, which reduces the strengthening effect of O on the matrix. It is shown in the above microstructure analysis that precipitation of TiB whiskers and La2 O3 particles at the grain boundaries is unfavorable for the interfacial bonding strength while at the same time strengthening the material itself, leading to a less obvious strengthening effect. In the elongation vs. LaB6 content curve, the ones with 0.1 % and 0.4 % LaB6 have higher elongations than normal alloy. The sample of 0.1 % LaB6 has an elongation of 30 %, while for the 0.5 % LaB6 sample, the elongation is as low as 16.5 %. The in situ reaction in the composite material produced TiB and La2 O3 ; they clean up and strengthen the grain boundaries, prevent the formation of micro-cracks and also allow slips to release concentrated stress, which results in a higher elongation rate. But when too much LaB6 is added, the contact at the grain boundaries becomes weaker, which leads to a lower elongation rate.

3.4 Conclusions In the experiments, we added different amounts of LaB6 into the β titanium alloy to prepare a titanium matrix composite. After analyzing the microstructure and mechanical properties, we reached the following conclusions: (1) For the sample with 0.1 % LaB6 , the in situ TiB whiskers were mainly formed at the grain boundaries. More TiB whiskers and La2 O3 spheres were observed as more LaB6 was added. La2 O3 particles refined the grains, and the diameter of the grains in samples with 0.5 % LaB6 was 12 um. The improvement of the tensile strength in the samples with 0.5 % LaB6 was due to fine grain strengthening. (2) In the sample with 0.1 % LaB6, the fraction elongation reached 30 %. In the sample with 0.1 % and 0.2 % LaB6 , good superelasticity and pure elastic strains were obtained. In the samples with 0.5 % LaB6 , the elastic strains reached a maximum due to fine grains. This would contribute to the improvement of the shape memory effects of the biomedical titanium alloy. In the in situ tensile test, we studied the main microscopic deformation mechanism of (TiB+La2 O3 )/(Ti–35Nb–3Zr–2Ta) alloy. Under small strains, line dislocations nucleated along the external tensile force direction. With strains increasing, the dislocations underwent nucleation, actuation and interactions. With dislocations slip-

100 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr ping, stress-induced twinning self-cooperative martensites were formed around dislocations. In the entire drawing process, {111}̄ twins and (001) composite twins were observed. As the strain rate increased, the (001) composite twins extended to the neighboring {111}̄ twins matrix slats and gradually grew to the same size of the original {111}̄ twins slats.

References [1] [2] [3] [4] [5] [6] [7] [8]

[9] [10] [11] [12] [13] [14] [15]

[16] [17]

Li YG, Zhang SQ, Li CZ et al. Study on the microstructure of the RE casting titanium alloy. Acta Metallurgica Sinica 1984;20(2):131–137. Li SX, Liu RM, Ma JM et al. The microstructure and properties of neodymides strengthened the titanium based composite alloy. Journal of Aeronautical Materials 1997;17(3):51–56. Deng J, Yang GJ. The role of rare earth elements in pure titanium and titanium alloys. Rare Metal Materials and Engineering 1993;22(2):1–11. Bomberger Jr, Howard B, and Stanley R Seagle. “Method of producing a hot-worked titanium product.” U.S. Patent No. 3,963,525. 15 Jun. 1976. Zhao CX, Liang GY, Li CL, Jin ZH. Influence of cerium, titanium and nitrogen on shape memory effect of Fe-Mn-Si-Cr-Ni alloys. Scripta Materialia 1998;38(7):1163–1168. Jin PY, Zhang WQ, Xiong W, Zhao ZQ. The influences of rare earth elements on the mechanical properties of CuZnAl shape memory alloy. Chinese Rare Earths 1997;18(1):37–40. Qin JF, Chen XY, Li ZX. Rare earths toxicity. Guangdong Trace Elements Science 2002;9 (5):1–10. Suzuki Y, Xu Y, Morito S, Otsuka K, Mitose K. Effects of boron addition on microstructure and mechanical properties of Ti–Td–Ni high-temperature shape memory alloys. Materials Letters 1998;36:85–94. Lu WJ, Zhang D, Zhang XN, et al. HREM study of TiB/Ti interfaces in a Ti-TiB-TiC in situ composite. Scripta Mater 2001;44:1069–1075. Long, MJ, Rack, HJ. Titanium alloys in total joint replacement-a materials science perspective. Biomaterials 1998;19:1621–1639. Wang LQ, Lu WJ, Qin JN, Zhang F, Zhang D. Microstructure and mechanical properties of cold-rolled TiNbTaZr biomedical β titanium alloy. Mater.Sci.Eng.A 2008;490:421–426. Kim JI, Kim HY, Inamura T, Hosoda H, Miyazaki S. Shape memory characteristics of Ti–22Nb–(2–8)Zr(at.%) biomedical alloys. Mater. Sci. Eng. A 2005;403:334–339. Ho, WF, Ju, CP, Chern, Lin, JH. Structure and properties of cast binary Ti–Mo alloys. Biomaterials 1999;20;2115–2122. Ninomi, M. Mechanical properties of biomedical titanium alloys. Mater. Sci. Eng. A. 1998;243:231–236. Saito T, Furuta T, Hwang JH, Kuramoto S, Nishino K, Suzuki N, et al. Multifunctional alloys obtained via a dislocation free plastic deformation mechanism. Science USA 2003;300:464–467. Otsuka K, Shimizu K. Pseudoelasticity and shape memory effects in alloys. Int. Metals Rev 1986;31:93–114. Furuhara T, Annaka S, Maki T. Superelasticity in β titanium alloys with nitrogen addition. Journal of Materials Engineering and Performance 2005;14:761–764.

References

| 101

[18] Suzuki Y, Xu Y, Morito S, Otsuka K, Mitose K. Effects of boron addition on microstructure and mechanical properties of Ti–Td–Ni high-temperature shape memory alloys. Materials Letters 1998;36:85–94. [19] Yang ZF. Doctoral Dissertation from Shanghai Jiao Tong University 2007. [20] Lu WJ, Zhang D, Zhang XN, et al. HREM study of TiB/Ti interfaces in a Ti-TiB-TiC in situ composite. Scripta Mater 2001;44:1069–1075. [21] Lu WJ, Zhang D, Zhang XN, et al. Creep rupture life of in situ synthesized (TiB+TiC)/Ti matrix composites. Scripta Mater 2001;44:2449–2455. [22] Yang Z, Lu W, Zhao L, Lu J, Qin J, Zhang D. In situ synthesis of hybrid-reinforced titanium matrix composites. Mater 2007;61:2368–2372. [23] Jin PY, Zhang WQ, Xiong W, Zhao ZQ. The influences of rare earth elements on the mechanical properties of CuZnAl shape memory alloy. Chinese Rare Earths 1997;18(1):37–40 [24] Liqiang Wang, Weijie Lu, Jining Qin, Fan Zhang, Di Zhang. Tensile properties of in situ synthesized (TiB+La2 O3 )/β -Ti composite. Materials Science and Engineering 2009;29:1897–1900. [25] Liqiang Wang, Weijie Lu, Jining Qin, Fan Zhang, Di Zhang. Microstructure and superelasticity of in situ synthesized (TiB+La2 O3 )/Ti alloy composites with different mass fraction of LaB6 . Materials Science and Engineering A 2010;527:1058–1062. [26] Michael, VS. Structure and Properties of Ceramics (Material Science and Technology) Weinhelm, VCH Verlagsgesellschaft mbH; 1994. [27] Kobayashi, M, Funami, K, Suzuki, S, et al. Manufacturing process and mechanical properties of fine TiB dispersed Ti-6Al-4V alloy composites obtained by reaction sintering. Mater. Sci. Eng. A. 1998;243:279–284. [28] Yang ZF, Lu,WJ, Zhao L, Lu JQ, Qin JN, Zhang D. In situ synthesis of hybrid-reinforced titanium matrix composites. Materials Letters 2007;61:2368–2372. [29] Lundstrom, T. Boron and refractory borides. In: Matkovich VL, ed. Berlin: Springer; 1977:351–376. [30] Li DX, Ping DH, Lu YX, et al. Characterization of the microstructure in TiB whisker reinforced Ti alloy matrix composites. Materials Letters 1993;16:322–326. [31] Ranganath S, Roy T, Mishra RS. Microstructure and deformation of TiB+Ti2C reinforced titanium matrix composites. Mater. Sci. Tech. 1996;12(3):219–226 [32] Ma, ZY. In-situ Ti-TiB metal-matrix composite prepared by a reactive pressing process. Scripta Mater 2000;42:367–373. [33] Geng K, Lv WJ, Zhang D, Yang ZF, Zang JD. In situ synthesized TiB and Nd2 O3 strengthened titanium based composite alloy. Journal of Shanghai Jiao Tong University China 2000;42:367–373. [34] Pearson, WB. A Handbook of Lattice Spacings and Structures of Metals and Alloys. Oxford: Pergamon Press;1958:93–111. [35] Massalski TB, Murry JL, Bennett LH, et al. Binary Alloy Phase Diagrams 1986 [M]. (Metals Park, Ohio, USA: ASM). [36] Xiaoyan M, Changrong,L, Zhenmin D et al. Thermodynamic assessment of the Ti–B system. Journal of Alloys and Compounds 2004;370:149–158. [37] Ikehata H, Nagasako N, Furuta T, Fukumoto A, Saito T. First-principles calculations for development of low elastic modulus Ti alloy. Phys. Rev. B. 2004;70:174113–174120. [38] Xu ZY et al. Shape Memory Materials. Shanghai: Shanghai Jiao Tong University Press; 2000. [39] Cai W. Doctoral Dissertation of Harbin Institute of Technology; 1994.

102 | 3 Novel preparation and mechanical properties of (TiB+La2 O3 )/TiNbTaZr [40] Cox, HL et al. The elasticity and strength of paper and other fibrous materials. British J. Appl. Phys. 1952;3:72–79. [41] Nardone VC, Prewo KM. On the strength of discontinuous silicon carbide reinforced aluminum composites. Scr. Metall. Mater 1986;20:43–48. [42] Nardone VC. Assessment of models used to predict the strength of discontinuous silicon carbide reinforced aluminum alloys. Scr. Metall. Mater. 1987;21:1313–1318. [43] Nussbaum G, Sainfort P, Regazzoni G, Gjestland H. Strengthening mechanisms in the rapidly solidified AZ91 magnesium alloy. Scr. Metall. 1989;23:1079–1084.

Dina V. Dudina, Igor S. Batraev, Vladimir Yu. Ulianitsky

4 Microstructure formation of particle-reinforced metal matrix composite coatings produced by thermal spraying Abstract: This chapter is concerned with the microstructure development of particlereinforced metal matrix composite (MMC) coatings produced by thermal spraying. The features of thermal spray processes of composite systems imparted by the presence of two or more phases in the feedstock powders are presented. Depending on the nature of feedstock powders, several alternatives exist to obtain metal-ceramic composite coatings, which include separate injection of metal and ceramic components into the gas stream, spraying of the powder blends and metal-ceramic composite particles and synthesis of reinforcing particles in situ via chemical reactions. These alternatives are analyzed from the viewpoint of typical microstructural features of the coatings. Possible post-spray treatment methods and their influence on the microstructure development of MMC coatings are discussed. Studies performed by our group on the preparation of MMC coatings by computer-controlled detonation spraying using both ex situ and in situ routes are reviewed to emphasize the significance of flexible variation of thermal spraying parameters for controlling the microstructure and phase composition of MMC coatings.

Introduction In thermal spraying techniques, the powder particles are heated and accelerated toward the substrate by high-temperature, high-speed gaseous flows to form a coating. A wide range of spraying conditions provided by a variety of spraying methods (plasma spraying, arc spraying, detonation spraying, high-velocity oxy-fuel (HVOF) spraying and cold spraying) makes it possible to deposit powders with different physical and chemical properties [1–4]. Promising combinations of strength, hardness, wear resistance and electrical and thermal conductivities can be achieved by manufacturing particle-reinforced metal matrix composite (MMC) coatings [5–11]. Due to the multiphase nature of the feedstock powders used to produce MMC coatings, additional issues need to be taken into account compared to the spraying of single-phase powders; these are the uniformity of distribution of the phases in the composite powders and in the deposited material, level of mixing, interfacial interactions and differences in the spraying behavior of particles of different phases if the components of the composite are fed into the gaseous flow separately. MMC coatings either inherit the microstructural features from the feedstock powders or acquire new

104 | 4 MMC coatings by thermal spraying ones as a result of melting and cooling of the material upon deposition. The presence of ceramic particles influences the relative density and adhesive strength of the coatings. In cold spraying, the introduction of ceramic particles can have a positive effect on the coating’s relative density. Meydanoglu et al. [12] reports an increased density of the cold-sprayed Al 7075 matrix coatings containing ceramic particles relative to Al 7075 unreinforced coatings. A similar effect was observed by Li et al. [13] when spraying Al-TiN powder mixtures along with an increase in the adhesive strength due to the pinning effect of the TiN particles. When coatings are formed by deposition of partially molten particles, wettability of the solid ceramic inclusions by the molten metal becomes an important factor of the microstructure development. This chapter’s aim is to discuss the microstructure development features of particle-reinforced MMC coatings produced by thermal spraying. The chapter consists of four sections. Section 4.1 discusses particle-reinforced MMC coatings formed ex situ by thermal spraying of powder mixtures and composite particles. A possibility of separate injection of the coating components into the gas stream is addressed. Section 4.2 deals with MMC coatings containing reinforcing particles formed in situ during thermal spraying. In section 4.3, we present our studies of TiB2 -Cu, Ti3 SiC2 -Cu and Al4 C3 -Al MMC coatings produced by computer-controlled detonation spraying. The significance of flexible variation of thermal spraying parameters for controlling the microstructure and phase composition of the MMC coatings formed by reactive and non-reactive routes are discussed. Section 4.4 describes several possible postspray treatment methods of MMC coatings aimed at improving their microstructural characteristics and properties.

4.1 Particle-reinforced MMC coatings formed ex situ by thermal spraying of powder mixtures and composite particles Ex situ MMC coatings can be obtained by spraying blends consisting of metal and ceramic particles or metal-ceramic composite particles formed by high-energy mechanical milling. The former possibility has two options: injecting the powders through separate feeders [3–4, 14–15] or injecting powder blends prepared by a preliminary mixing operation [16–19]. In terms of the physical and chemical influence of the spraying environment on the material, these options are similar, as there is no previously established interface between the components. However, when the metal and ceramic powders are injected through separate feeders, it is possible to change the ratio of the components and thus make MMC coatings containing layers with differences in composition. Tillmann et al. [14] varied the location of the feeder that supplied ceramic particles (coarse Al2 O3 ) relative to the detonation gun and showed that when the ceramic particles are injected into the stream at the exit of the gun and copper particles are

4.1 Ex situ MMC coatings

|

105

injected into the barrel, the ceramic particles do not embed into the copper matrix. When the ceramic powder was fed into the barrel through its own feeder, the copper matrix composite coatings were successfully formed. Substrate pre-heating played a positive role in the formation of the composite coating; this facilitated embedding the ceramic particles into the metal matrix, which became softer upon heating. An interesting observation was made that, although the Al2 O3 particles injected at the exit of the gun did not embed into the metal matrix, they made the metallic coating denser by impinging on the substrate, bouncing off and not participating in the formation of a deposit. If two feeders work alternately, which is allowed in the computer-controlled detonation spraying facilities, different spraying parameters can be set for each shot of the detonation gun. This allows creating optimum spraying conditions for each component of the coating [3–4]. The resultant coating will be composed of layers of different phases. When a feedstock powder consists of particles of two types, coatings with a gradient structure can form. Figure 4.1 shows a coating formed by plasma spraying of Cu+SiC mixtures onto a graphite substrate, which demonstrates a concentration gradient of SiC in the copper matrix formed due to SiC particles’ bouncing off the graphite substrate in the beginning of the process.

Fig. 4.1. Cross-section of a plasma-sprayed coating with a gradient structure produced from Cu/SiC powders. Reprinted from [17] with permission from Elsevier.

Although mechanical milling is used in many studies for the preparation of feedstock powders for thermal spraying, the milling energy is low, and, therefore, the powder mixtures essentially maintain the structure of a blend containing very few, if any, metal-ceramic composite particles. In order to prepare powder mixtures for thermal spray, Arrabal et al. [18] and Torres et al. [20] used mechanical milling of Al and SiC powders, which helped in achieving uniform intermixing, but did not result in the formation of composite particles. Mechanical milling was used by Kang & Kang [17] to prepare the Cu/SiC feedstock powders for plasma spraying. The Cu+SiC powder mixtures were ball-milled; however, very little intermixing between the materials has been observed. The product of milling consisted mainly of separate particles of Cu and SiC. The chemical behavior of SiC particles during spraying was, therefore, similar to the

106 | 4 MMC coatings by thermal spraying case of the spraying of single-phase SiC. Subjected to high temperatures, SiC decomposed, which was microscopically seen as a surface roughness (pores) of the particles. The powder preparation stage of the powder metallurgy-based processing of MMCs often includes mechanical milling of the mixtures aimed at intimate mixing of the reinforcement and matrix phases [1, 10, 21–23]. This can be achieved if matrixreinforcement composite agglomerates form with a well-developed interface between the phases. As reinforcement particles should be distributed in a metal matrix as uniformly as possible to achieve better mechanical properties and isotropy of the composite, those conditions of mechanical milling are chosen that result in the plastic deformation of the matrix particles, which are usually larger than the particles of the reinforcement. In our studies, we have shown that the structure of the metal-ceramic composite mixture affects the chemical processes occurring during the coating formation [24]. We have found that the uniformity of distribution of reinforcing particles in the deposited layer is not the only consequence of the preliminary mechanical milling of the powder mixtures. We used the Ti3 SiC2 -Cu system as an example of a MMC with a ductile matrix and a reinforcement phase capable of chemically interacting with the matrix. Two composite powders were used, one of them was prepared by mixing the Ti3 SiC2 and Cu powders in a mortar and, therefore, did not contain composite agglomerates; the other was mechanically milled in a high-energy ball mill and consisted of Ti3 SiC2 -Cu composite agglomerates (Figure 4.2).

(a)

200 μm

(b)

200 μm

Fig. 4.2. Morphology of Ti3 SiC2 +Cu powders mixed in a mortar (a) and Ti3 SiC2 -Cu composite agglomerates produced by high-energy mechanical milling (b).

During detonation spraying, the reaction between Ti3 SiC2 and Cu in the mixture prepared by mixing in a mortar started in milder conditions (at a lower explosive charge) than in the mechanically milled mixture. As can be seen from Figure 4.3, titanium carbide TiCx , which is the product of the reaction between Ti3 SiC2 and Cu

4.1 Ex situ MMC coatings

|

107

+ + Cu * Ti3SiC2 + Intensity, a.u.

∙ TiCx

40% – 1.1 30% – 1.1 Ti3SiC2–Cu powder 30

35

40

45 50 2θ Cu Kα

(a)

55

60

65

+ + Cu * Ti3SiC2 ∙ TiCx

Intensity, a.u.

+

?

40% – 1.1 30% – 1.1

30 (b)

35

40

45 50 2θ Cu Kα

55

60

65

Fig. 4.3. XRD patterns of the coatings produced by detonation spraying of 20 vol.%Ti3 SiC2 -Cu powders: (a) mechanically milled mixture; (b) mixed in a mortar (marked with the values of explosive charge, % and O2 /C2 H2 ratio).

(Ti3 SiC2 +Cu → TiCx +Cu(Si)), was found in the coating sprayed at an explosive charge of 30% of the barrel volume using the mixture prepared by mixing in a mortar. In the coating produced using the same spraying parameters and the mechanically milled mixture as the feedstock powder, no TiCx was detected. This can be rationalized if overheating of the fine Ti3 SiC2 particles contained in the starting powder, and, consequently, in the mixture prepared by mixing in a mortar, is taken into account. While in the coatings obtained from the mechanically milled mixture, the reaction,

108 | 4 MMC coatings by thermal spraying

(a)

500 um

(b)

300 um

Fig. 4.4. Cross-sections of the coatings produced by detonation spraying of 20 vol.%Ti3 SiC2 -Cu powders at an explosive charge of 30 %, O2 /C2 H2 = 1.5: (a) mechanically milled mixture; (b) mixed in a mortar.

if initiated, led to full transformation, in the coatings formed by powder mixtures mixed in a mortar, untransformed Ti3 SiC2 was present due to the agglomeration of Ti3 SiC2 particles and a reduced interfacial area between Ti3 SiC2 and Cu. Coatings produced from the mechanically milled mixture show a uniform microstructure free from Ti3 SiC2 agglomerates (Figure 4.4). The structure of the powder also influences the substrate/coating interface with a well-bonded coating formed in the case of the mechanically milled powders and pronounced delamination effects observed in the case of spraying the mixtures prepared by simple mixing.

4.2 MMC coatings with reinforcing particles formed in situ during thermal spraying Reinforcing particles can be formed from the reactants contained in the feedstock powder by solid-state diffusion and synthesis involving liquid phases. An example of the in situ route is presented in a study by Legoux and Dallaire [25], in which plasma spraying of Cu-Ti alloy and boron powders resulted in the formation of titanium diboride particles in the copper matrix. Studies by Ozdemir et al. [26] show that the outcome of the in situ reaction process depends on the conditions of spraying; namely, the particle velocity and temperature, and, therefore, it will be different for the same reactants sprayed by different techniques. Comparative experiments on spraying AlMg-Si micron-sized particles coated with a Ni layer and covered by SiO2 nanopartices by the HVOF and atmospheric plasma spraying (APS) techniques showed that the reactions in the sprayed material proceeded to a greater extent in the APS coatings due to the higher temperatures involved in the process. To answer the question whether the reaction in the particles occurs in-flight or upon deposition on the substrate, the

4.2 MMC coatings with reinforcing particles formed in situ during thermal spraying

| 109

microstructures of the feedstock powder particles and sprayed particles collected before reaching the substrate were compared with that of the coatings. The observed differences between the distributions of elements in the feedstock, sprayed and coating materials allowed for the conclusion that the reactions occurred mainly upon particle layering on the substrate; however, partial reaction occurred during the particle flight in the APS. The reaction products hardly formed prior to particle deposition in the HVOF spraying. The overall extent of the reactions was greater in the APS, which resulted in higher hardness of the coatings. The in situ formation of TiC in a steel matrix was reported by Röttger et al. [27], who produced coatings by HVOF using a blend of a gas-atomized high-carbon highboron steel powder and a Fe-Ti alloy powder. No titanium carbide was found in the assprayed coatings; however, the subsequent hot isostatic pressing performed at 1000 ∘ C for 2 h provided sufficient time for the diffusion processes. Owing to the formation of a new hard phase in the composite coatings, their wear resistance increased relative to the as-sprayed state. An interesting application of plasma spraying was proposed by Lee et al. [28], who fabricated Al/Al3 Fe metal matrix composites by spraying iron powders into aluminum melt prepared before the spraying experiments by induction heating of an ingot. As a result of a reaction between Fe and Al, Al3 Fe formed as a reinforcing phase. The microstructure of the composite could be controlled by varying the Al melt temperature and the plasma spraying conditions, such as gas flow rate, input current and spraying distance. The fraction of the needle-shaped Al3 Fe increased with increasing melt temperature and decreasing spraying distance. High temperatures involved in thermal spraying may be detrimental for the materials containing metastable reinforcements. In this case, cold spraying offers a viable solution. Thus, the benefit of cold spraying in comparison with HVOF was in preservation of the quasi-crystalline phases used as reinforcements in Al matrix composite coatings, as was shown in a study by Byakova et al. [29]. Interaction of the particles of reactive metallic materials (titanium, chromium and aluminum) with gaseous species, such as oxygen and nitrogen, during thermal spraying presents another opportunity to form reinforcing particles [30–34]. Zhao&Lugscheider [30] showed that titanium nitrides formed in situ during plasma spraying of the TiAl6V4 alloy. The coatings produced by reactive plasma spraying showed much higher microhardness and wear resistance than the corresponding bulk alloy. Nitridation of chromium was observed by Tsunekawa et al. [33] during DC plasma spraying, the coating containing the CrN and Cr2 N phases. For reactive powders, the chemical and phase composition was shown to depend on the spraying distance when the sprayed particles traveled through air before reaching the substrate [32]. Deevi et al. [16] suggested reactive thermal spraying as a possibility of broadening the engineering applications of the coatings through increasing the variety of achievable compositions and microstructures. Despite significant progress in reactive ther-

110 | 4 MMC coatings by thermal spraying mal spraying in recent years, in practice it is not as widespread as non-reactive routes. When the spraying behavior of a material is complicated by a possibility of chemical interaction between the constituent phases or by reactions with the gaseous atmosphere, it is rather difficult to predict the microstructural features of the coating in each particular case without conducting a detailed experimental investigation of the process. Our studies have shown that the grains/particles of the reaction products that form in the detonation-sprayed coatings may be coarser than those of the feedstock powder [34]. Therefore, research that can help elucidate the microstructuregoverning factors is of great importance for the further advancement of reactive thermal spraying.

4.3 Design of particle-reinforced MMC coatings using flexible variation of spraying parameters in computer-controlled detonation spraying This section describes our research on MMC coatings produced by computer-controlled detonation spraying (CCDS) and is aimed to show the importance of flexible adjustment of the spraying parameters for controlling the microstructure formation processes of the coatings. Detonation spraying facilities using computer control of the process were developed in 2005 [4] and have been successfully used for various material systems since then [24, 34–41]. A key advantage of the computer-controlled detonation spraying method is the possibility of precisely controlling the quantity of the explosive gaseous mixture used for each shot of the detonation gun and the oxygen-to-fuel ratio of the explosive mixtures. During heating, the sprayed particles experience chemical action of the gaseous species, which are the detonation products of the fuel-oxygen mixtures or the carrier gas components. There is a possibility of particles’ interacting with air after leaving the detonation gun and before reaching the substrate. The presence of oxide phases in the sprayed material has been reported in reference [34]. In order to tackle this issue, explosive mixtures rich in fuel were used to obtain the detonation products of reducing chemical nature [37]. Chemical sensitivity of the powders to the spraying atmosphere and interaction between the phases of composite powders at high temperatures are extremely important for the phase and microstructure development of the coatings. Due to the pulsed nature of the detonation spraying process, chemical interactions take place in highly non-equilibrium conditions. The reaction products may be metastable in terms of phase and crystalline structure due to fast reaction and rapid cooling of the splats upon deposition on the substrate. We have found that as the spraying conditions are varied – flexibly and in a wide range allowed by the technical capabilities of the facility – new phases can appear in the coatings in substantial quantities as a result of chemical reactions of re-

4.3 MMC coatings by detonation spraying

2 Carrier gas Fuel Fuel Oxygen Oxygen Carrier gas

3

1

6

4

|

111 7

5 Spraying distance

Fig. 4.5. A schematic drawing of a CCDS2000 facility: (1) gun barrel; (2) computer-controlled precision gas distribution system; (3) explosive charge; (4) carrier gas; (5) computer-controlled powder feeder; (6) cloud of injected powder; (7) substrate.

duction, oxidation, nitridation as well as interfacial interactions between the phases of composite feedstock powders. A schematic of a CCDS2000 facility is presented in Figure 4.5 [4, 40]. A channel inside the gun barrel (1) measuring 850–1000 mm long and 20 mm in diameter is filled with gases by a computer-controlled precision gas distribution system (2); first, it is filled with a carrier gas, then with a certain portion of an explosive mixture; this results in the formation of a stratified gas medium consisting of explosive charge (3) and carrier gas (4). The feedstock powder is injected into the barrel through an orifice by a computer-controlled feeder (5) with the help of the carrier gas flow. The powder injected into the barrel forms cloud (6). The spraying distance is measured as a distance from the exit of the barrel (1) to the surface of the substrate (7). In our spraying experiments with different materials, the ratio O2 /C2 H2 was varied between 0.7 and 2.5 while the explosive charge was set at 30–60% of the barrel volume. An explosive combustion of the charge occurs within a time of the order of 1ms such that a detonation wave forms in the explosive mixture transforming into a shock wave in the carrier gas. The detonation products heated up to 3500–4500 K and the carrier gas heated by the shock wave up to 1000–1500 K move at a supersonic speed and exchange heat and interact with the powder during 2–5 ms. In this process, the powder particles can be heated up to the material’s melting temperature and accelerated up to velocities as high as 500 m/s. Depending on the initial composition of the explosive mixture, the detonation products create either a reducing or an oxidizing environment for the sprayed powders (Table 4.1). The formation of nitrides in the coatings is possible as a result of the interaction of the powders with nitrogen used as carrier gas or with nitrogen contained in air when air is the carrier gas. Additional amounts of nitrides in the coatings can form if the particles interact with the nitrogen of air after they exit the gun and move to the substrate to form a deposit. Oxide phases in the coatings can form due to oxidation by the detonation products and in a similar manner due to interaction of the particles with the oxygen of air.

112 | 4 MMC coatings by thermal spraying Table 4.1. Molar fractions of the major components of the detonation products of O2 +C2 H2 mixtures of different O2 /C2 H2 ratios.

Components

Molar fraction O2 /C2 H2 = 1.1

O2 /C2 H2 = 1.5

O2 /C2 H2 = 2.0

0.009 — 0.215 0.165 0.013 0.014 0.579 0.005

0.064 0.015 0.160 0.094 0.070 0.060 0.503 0.033

0.112 0.065 0.102 0.053 0.106 0.083 0.411 0.069

O O2 H H2 OH H2 O CO CO2

We investigated the detonation spraying behavior of 57vol.% TiB2 -Cu [35–36] and 20vol.% Ti3 SiC2 -Cu [24, 40–41] powders (composite agglomerates) produced by combining high-energy mechanical milling and self-propagating high-temperature synthesis. The detonation-sprayed 57vol.% TiB2 -Cu coatings were obtained in three modes differing in the O2 /C2 H2 ratio and explosive charge. Titanium diboride does not react with copper; therefore, the microstructural changes that could be expected are those related to grain growth and phase redistribution. As can be seen from the XRD patterns of the coatings (Figure 4.6), the major phases of the composite were well preserved in all experiments. An unknown phase was also detected in the coatings (labeled by a question mark), which is likely to be an oxide. Higher temperatures of +

Intensity, a.u.

+ Cu * TiB2

?

+ *

?

*

+ 60% – 1.1

* * *

?

* *

20

30

*

30% – 2.5 *

*

30% – 1.1 *

*

?

40

50 60 2θ Co Kα

70

*

80

90

Fig. 4.6. XRD patterns of the 57vol.%TiB2 -Cu detonation-sprayed coatings (marked with the values of explosive charge, % and O2 /C2 H2 ratio).

4.3 MMC coatings by detonation spraying

(b)

SEI 10.0kV X150 100μm

(c)

SEI 10.0kV X30,000 100nm

(d)

SEI 10.0kV X500 10μm

(e)

113

SEI 10.0kV X30,000 100nm

SEI 10.0kV X100 100μm

(a)

|

SEI 10.0kV X30,000 100nm

(f)

Fig. 4.7. Microstructure of the detonation (a–e) and cold-sprayed (f) 57vol.%TiB2 -Cu coatings; (a,b) O2 /C2 H2 = 1.1, explosive charge 30 %; (c,d) O2 /C2 H2 = 2.5, explosive charge 30 %; (e) O2 /C2 H2 = 1.1, explosive charge 60 % (etched with FeCl3 solution).

the sprayed particles reached with increasing oxygen content in the explosive mixture led to the growth of TiB2 particles in the composite coatings, as can be seen from Figure 4.7a–d. In the coating sprayed at O2 /C2 H2 = 1.1 and an explosive charge

114 | 4 MMC coatings by thermal spraying

30 um

30 um

30 um

30 um

Fig. 4.8. Microstructure of the coatings produced by detonation-spraying of 20 vol.% Ti3 SiC2 -Cu feedstock powders prepared by high-energy mechanical milling. Conditions of spraying: (a) explosive charge 30 %, O2 /C2 H2 = 1.1; (b) explosive charge 60 %, O2 /C2 H2 = 1.1; (c) explosive charge 30 %, O2 /C2 H2 = 1.5; (d) explosive charge 30 %, O2 /C2 H2 = 2.5.

of 60%, phase redistribution occurred resulting in the formation of Cu-rich regions (Figure 4.7e). For comparison, the microstructure of the cold-sprayed 57vol.% TiB2 -Cu coating [42] is shown in Figure 4.7f. In the cold-sprayed coating, the size of TiB2 particles was well retained and corresponded to that of the particles in the feedstock powder. The microstructural differences resulted in higher hardness of the cold-sprayed coatings compared to the detonation-sprayed coatings [36]. The XRD pattern of the 20vol.% Ti3 SiC2 -Cu feedstock powder is shown in Figure 4.3a. The coating sprayed at an explosive charge of 30% and O2 /C2 H2 = 1.1 has a phase composition of the feedstock powder. Analyzing the microstructure of the coating (Figure 4.8a), one can conclude that this coating was formed by densely packed powder agglomerates that did not experience melting. Indeed, a relatively small quantity of the explosive mixture corresponding to the explosive charge of 30% allows depositing the Ti3 SiC2 -Cu in the solid state. A slightly larger size of Cu crystallites of the coating, as compared to that of the Ti3 SiC2 -Cu feedstock powder (Table 4.2), is due to the crystallite growth in the particles heated by the detonation products. As the particle temperature increases with increasing explosive charge or oxygen content in the O2 +C2 H2 mixture, partial melting of the sprayed material can be expected. As is seen from the XRD patterns of the corresponding coatings (Figure 4.3a), the reaction between Ti3 SiC2 and Cu occurred resulting in the formation of the TiCx phase. Sensitivity of the Ti3 SiC2 -Cu composite system to temperature is explained by a tendency of Si to de-intercalate from the T3 SiC2 compound at temperatures exceeding 900 ∘ C and react with copper to form a Cu-Si compound or a solid solution [42]. No Cu-Si compounds were found in the coatings. Larger lattice parameters of Cu(Si) observed in some of the TiCx -Cu(Si) coatings (Table 4.2) are due to the lattice expansion of copper upon dissolution of silicon. In some coatings, however, the Cu(Si) lattice parameter did not noticeably change, although, according to the XRD phase analysis, the TiCx formed and silicon had to dissolve in copper. In order to explain these observations, further investigation is needed to determine the fine structure of the TiCx /Cu(Si) interfaces and other possible reaction products formed in quantities that are not detectable by the conventional XRD phase analysis. In the coatings that have the Cu(Si) matrix, the

4.3 MMC coatings by detonation spraying

|

115

Table 4.2. The lattice parameter of copper and crystallite size of the phases of the 20vol.% Ti3 SiC2 Cu powder obtained by high-energy mechanical milling and those of the coatings produced by detonation spraying of this powder, the lattice parameter and crystallite size of the electrolytic copper powder (starting material) and detonation-sprayed copper and hardness of the coatings. Cu/Cu(Si) crystallite size, nm

Crystallite Hardness, size of HV Ti3 SiC2 /TiCx , nm

Phase composition determined by XRD

Explosive charge, %

O2 /C2 H2 molar ratio

20 vol.% Ti3 SiC2 -Cu milled powder Ti3 SiC2 -Cu coating TiCx -Cu(Si) coating TiCx -Cu(Si) coating TiCx -Cu(Si) coating TiCx -Cu(Si) coating TiCx -Cu(Si) coating TiCx -Cu(Si) coating Cu coating Cu (starting material)

—

—

3.6165 ± 0.0002

35

20

—

30 40 50 60 30 30 30 40 —

1.1 1.1 1.1 1.1 1.1 2.0 2.5 1.1 —

3.6158 ± 0.0002 3.6190 ± 0.0002 3.6210 ± 0.0003 3.6175 ± 0.0002 3.6159 ± 0.0001 3.6156 ± 0.0001 3.6150 ± 0.0001 3.6165 ± 0.0001 3.6154 ± 0.0001

50 40 30 47 60 60 68 40 145

20 40 50 30 35 40 25 — —

133 214 273 221 190 181 173 — —

Cu/Cu(Si) lattice parameter, A

reinforcing phase, TiCx , can be considered as formed in situ as a result of the reaction induced during detonation spraying. Both reinforcing phases (Ti3 SiC2 and TiCx ) have nano-sized crystallites. The temperatures of the sprayed particles were high enough for the interfacial interaction between Ti3 SiC2 and Cu to occur and for copper to melt, but they are still low to induce growth of the crystallites of the newly formed TiCx phase. A higher temperature of the particles reached during spraying with increased explosive charge or O2 /C2 H2 ratio causes the formation of coatings, the major fractions of which are formed as a result of melting and re-solidification (Figure 4.8c–e). In order to verify whether it is the cooling rate that is crucial for the formation of the nanocrystalline copper-based matrix or a stabilizing effect of the reinforcing TiCx particles plays a significant role, we sprayed pure electrolytic copper powder at an explosive charge of 40% and O2 /C2 H2 = 1.1. The surface structure of the deposited copper showed that a significant part of the coating was formed by re-solidified melt, which is in agreement with observations reported by Kosarev et al. [44], who studied the microstructure of the copper coatings produced by detonation spraying. The calculated crystallite size of copper in the coating was 40 nm (Table 4.2), which shows that even without a second phase, a nanostructured metallic material can form owing to rapid cooling of the melt. The crystallite size of copper in the starting powder was much larger (145 nm). The CCDS facility allowed us to create different conditions of spraying that resulted in the formation of coatings, in which the phase composition and microstructure could be changed and controlled. In mild conditions of detonation spraying, the interfacial reaction between the Ti3 SiC2 and Cu phases does not occur, and the phase

116 | 4 MMC coatings by thermal spraying composition of the sprayed coating is inherited from the powder. As the particle temperature increased with increasing oxygen content in the O2 +C2 H2 mixtures or with increasing explosive charge, partial melting of the sprayed material took place accompanied by a reaction between Ti3 SiC2 and Cu, which led to the formation of TiCx -Cu(Si) composite coatings. The crystallite size of the Cu(Si) matrix in the composite coatings determined from the XRD profiles was in the range from 30 to 68 nm. A correlation between the crystallite size of the copper matrix and hardness of the TiCx -Cu(Si) composite coatings was observed, TiCx -Cu(Si) coating with Cu(Si) crystallites of 30 nm showing a hardness of 273 HV. We are currently evaluating a possibility of producing metal-carbon and metalcarbide coatings using carbon in situ formed during detonation spraying conducted with explosive mixtures rich in C2 H2 . Different carbon structures can form when incomplete combustion of hydrocarbons occurs [45]. In detonation spraying, particles of amorphous carbon can form on the substrate in highly reducing conditions of spraying [46]. When metals are sprayed in such conditions, metal-carbon coatings can be obtained. However, the crystalline state of the carbon-containing phase is a question that needs to be answered in each particular case. We sprayed an aluminum powder at O2 /C2 H2 = 0.7 and an explosive charge of 45%. The microstructure of the coating and XRD pattern taken from the coating surface in the as-sprayed state are shown in Figure 4.9a,b. Despite the presence of carbon formed as a result of the incomplete combustion of C2 H2 , we found no aluminum carbide in the as-sprayed coatings. It is likely that the spraying time is too short for the interfacial reaction to occur. We suggest that the carbide phases in the coatings produced from metals or alloys by detonation spraying in highly reducing conditions can be synthesized during the subsequent annealing of the coatings in an inert atmosphere. Indeed, Al4 C3 was detected in the coating heated in an argon atmosphere up to the melting temperature of aluminum (Figure 4.9c). The Al2 O3 phase formed simultaneously with Al4 C3 – probably due to crystallization of poorly crystallized aluminum oxide present in the feedstock powder. Other post-spray treatment methods of thermally sprayed MMC coatings and their possible outcomes are reviewed in section 4.4.

4.4 Post-spray treatment of MMC coatings Thermally sprayed coatings may not reach their full potential in terms of protective and mechanical properties due to porosity and/or undesirable chemical changes relative to the feedstock powder. The properties of thermally sprayed coatings can be further improved by post-treatment operations applied to as-sprayed coatings. Several methods are currently being developed to reduce the porosity and improve the mechanical characteristics of the coatings. A possibility of increasing the coating hardness by friction stir processing was shown for cold-sprayed [47] and HVOF coatings [48]. Using the flame spray process to deposit Al/SiC on magnesium alloys, Arrabal

4.4 Post-spray treatment of MMC coatings

| 117

100 um

(a) *

Intensity, a.u.

* Al *

* * * 20

30

40

50 60 2θ Cu Kα

(b) *

70

80

90

*

*

Intensity, a.u.

* Al + Al4C3 ^ Al2O3

* + ?^

20 (c)

^ + + + +^

30

40

+

^

+^

* + ^^ +

50 60 2θ Cu Kα

70

80

90

Fig. 4.9. Microstructure of the Al-C detonation-sprayed coating (O2 /C2 H2 = 0.7; explosive charge 45 %) and XRD patterns taken from its surface in the as-sprayed (b) and annealed (c) state.

118 | 4 MMC coatings by thermal spraying

500μm (a)

500μm (b)

Fig. 4.10. Al/SiC as-sprayed coating on AZ31 substrate (a), the same coating after cold pressing (b). Reprinted from [18] with permission from Elsevier.

et al. [18] found that the as-sprayed coatings do not provide adequate corrosion protection due to remaining porosity. A way to increase the density of the coatings and improve their protective properties was suggested, which was based on the cold pressing of the as-sprayed layers at a relatively low pressure (32 MPa). As can be seen from Figure 4.10, cold pressing produced denser coatings, eliminated pores and cracks at the coating/substrate interface and established a more intimate bonding between the SiC particles and Al matrix. In order to tackle the issue of residual porosity of thermally sprayed coatings, Solonenko et al. [49] suggested using pulsed electron beam treatment. The electron beam pulses were applied to TiC-(Ni-Cr) plasma-sprayed coatings and resulted in melting of the metallic binder at the free surface of the coating. The molten binder filled the existing pores; the effect increased with an increasing number of pulses. A dramatic difference between the surface morphologies of the as-sprayed TiC-(Ni-Cr) coatings and coatings treated by electron beam pulses was observed: the rough and grainy microstructure observed in the former changed to a smooth one in the latter. Post-spray treatment of the coatings is not limited to increasing density; during the treatment, chemical changes in the material of the coatings can take place. In this direction, Lia et al. [50] proposed an approach to solving the problem of carbon losses (decarburization) in thermally sprayed WC-Co coatings based on the use of Spark Plasma Sintering. Thermally sprayed WC-Co coatings suffer from carbon losses, which cause a reduction in the resistance of the coatings against abrasive wear. A normal sintering run in the Spark Plasma Sintering (SPS) is conducted in a graphite die and with the use of graphite punches [51]. Post-spray SPS-treatment was conducted on plasma-sprayed WC-Co coatings and resulted in the restoration of WC in the coating through phase transformation from W2 C or reaction of carbon with W. After 6 min of the SPS-treatment at 800 ∘ C, WC was the dominant phase in the coatings. The restoration of carbon resulted in a 40% increase in the microhardness of the surface of the

4.4 Post-spray treatment of MMC coatings

|

119

SPS-treated coatings relative to the as-sprayed coatings. However, due to a short processing time in the SPS, carbon restoration was limited to a thickness of less than 10 μm.

Summary Thermal spraying of powders offers several possibilities for forming particle-reinforced MMC coatings. The microstructure formation processes of the MMC coatings are influenced by the spraying parameters and the structure of the feedstock powder. When a powder blend consisting of metal and ceramic particles not bonded to each other is sprayed, the coating microstructure and the distribution of the reinforcing phase depend on the spraying behavior of two separate powders. In this case, the thermal spraying process is associated with the issues of establishing an interface between the phases and reaching full density of the deposits. When mechanically milled mixtures, in which ceramic particles are already embedded in a metal matrix, are sprayed to form a coating, the distribution of the reinforcing phase does not alter provided the spraying occurs in the solid state or the presence of a molten phase does not cause phase separation and redistribution due to poor wettability. A promising research direction in the manufacture of MMC coatings is performing chemical reactions in situ during the spraying process to form the reinforcing phases. The in situ formation of the reinforcements offers additional flexibility of the microstructure design. However, the literature overview shows that MMC coatings are more often produced by spraying of powder mixtures as a simpler alternative. For metal-ceramic pairs sensitive to the temperature of the process and capable of chemically interacting with each other, a very careful adjustment of the spraying parameters is necessary. Further advances in the development of MMC coatings can be made by combining the capabilities of the up-to-date variations of thermal spraying techniques, such as computer-controlled detonation spraying, and efficient preparation methods of composite feedstock powders based on mechanical milling of controlled intensity. The latter would allow controlling the size and morphology of the composite powder particles. Improvement of the MMC coating quality can also be achieved by applying post-spray treatment of the deposits, such as cold pressing, annealing, Spark Plasma Sintering and electron beam treatment. Acknowledgement: This work was supported by RFBR, research project No. 14-0300164a.

120 | 4 MMC coatings by thermal spraying

References [1] [2]

[3] [4]

[5] [6] [7] [8] [9]

[10] [11] [12]

[13]

[14] [15]

[16] [17] [18]

[19]

[20]

Fauchais P, Montavon G, Bertrand G. From powders to thermally sprayed coatings. J Thermal Spray Technol 2010;19:56–80. Fauchais P, Montavon G, Lima RS, Marple BR. Engineering a new class of thermal spray nano-based microstructures from agglomerated nanostructure particles, suspensions and solutions: an invited review. J Phys D: Appl Phys 2011;44:093001 (53 p). Sova A, Pervushin D, Smurov I. Development of multimaterial coatings by cold spray and gas detonation spraying. Surf Coat Technol 2010;205:1108–1114. Ulianitsky V, Shtertser A, Zlobin S, Smurov I. Computer-controlled detonation spraying: from process fundamentals toward advanced applications. J Thermal Spray Technol 2011;20:791–801. Tjong SC, Ma ZY. Microstructural and mechanical characteristics of in situ metal matrix composites. Mater Sci Eng 2000;29:49–113. Kaczmar JW, Pietrzak K, Wlosinski W. The production and application of metal matrix composite materials. J Mater Proc Tech 2000;106:58–67. Torralba JM, da Costa CE, Velasco F. P/M aluminum matrix composites: an overview. J Mater Proc Tech 2003;133:203–206. Miracle DB. Metal matrix composites – From science to technological significance. Comp Sci Tech 2005;65: 2526–2540. Slipenyuk A, Kuprin A, Milman Y, Goncharuk V, Eckert J. Properties of P/M processed particle reinforced metal matrix composites specified by reinforcement concentration and matrix-to-reinforcement particle size ratio. Acta Mater 2006;54:157–166. Tjong SC. Novel nanoparticle-reinforced metal matrix composites with enhanced mechanical properties. Adv Eng Mater 2007; 9:639–652. Mortensen A, Llorca J. Metal Matrix Composites. Annu Rev Mater Res 2010;40:243–270. Meydanoglu O, Jodoin B, Sabri Kayali E. Microstructure, mechanical properties and corrosion performance of 7075 Al matrix ceramic particle reinforced composite coatings produced by the cold gas dynamic spraying process. Surf Coat Technol 2013;235:108–116. Li WY, Zhang G, Zhang C, Elkedim O, Liao H, Coddet C. Effect of ball milling of feedstock powder on microstructure and properties of TiN particle-reinforced Al alloy-based composites fabricated by cold spraying. J Thermal Spray Technol 2008;17:316–322. Tillmann W, Vogli E, Nebel J. Development of Detonation Flame Sprayed Cu-Base Coatings Containing Large Ceramic Particles. J Thermal Spray Technol 2007;16:751–758. Tillmann W, Hollingsworth PS, Fischer G, Nellesen J, Beckmann F. Development and Characterization of B4 C Reinforced Detonation-Sprayed Al Coatings. J Thermal Spray Technol 2014;23:289–295. Deevi SC, Sikka VK, Swindeman CJ, Seals RD. Application of reaction synthesis principles to thermal spray coatings. J Mater Sci 1997;32:3315–3325. Kang HK, Kang SB. Thermal decomposition of silicon carbide in a plasma-sprayed Cu/SiC composite deposit. Mater Sci Eng A 2006;428:336–345. Arrabal R, Pardo A, Merino MC, Mohedano M, Casajús P, Merino S. Al/SiC thermal spray coatings for corrosion protection of Mg-Al alloys in humid and saline environments. Surf Coat Technol 2010;204:2767–2774. Torres B, Campo M, Lieblich M, Rams J. Oxy-acetylene flame thermal sprayed coatings of aluminium matrix composites reinforced with MoSi2 intermetallic particles. Surf Coat Technol 2013;236:274–283. Torres B, Campo M, Rams J. Properties and microstructure of Al-11Si/SiCp composite coatings fabricated by thermal spray. Surf Coat Technol 2009;203:1947–1955.

References

| 121

[21] Jia DC. Influence of SiC particulate size on the microstructural evolution and mechanical properties of Al-6Ti-6Nb matrix composites. Mater Sci Eng A 2000;289:83–90. [22] Dudina DV, Georgarakis K, Li Y, Aljerf M, Braccini M, Yavari AR, Inoue A. Cu-based metallic glass particle additions to significantly improve overall compressive properties of an Al alloy. Composites Part A 2010;41:1551–1557. [23] Dudina DV, Georgarakis K, Yavari AR. Metal matrix composites reinforced with metallic glass particles: state of the art. In: Metal Matrix Composites, Davim JP, ed., Nova Science Publishers Inc. 2012:1–30. [24] Dudina DV, Ulianitsky VY, Batraev IS, Korchagin MA, Mali VI, Anisimov AG, Lomovsky OI. Inter-particle interactions during consolidation of Ti3 SiC2 -Cu powders influenced by preliminary mechanical milling. Chem Sust Dev 2014;22:31–37. [25] Legoux JG, Dallaire S. Copper-titanium diboride coatings obtained by plasma reactive micropellets. J Thermal Spray Technol 1993;2:283–286. [26] Ozdemir I, Hamanaka I, Tsunekawa Y, Okumiya M. In-Process Exothermic Reaction in High-Velocity Oxyfuel and Plasma Spraying with SiO2 /Ni/Al-Si-Mg Composite Powder. J Thermal Spray Technol 2005;14:321–329. [27] Röttger A, Weber SL, Theisen W, Rajasekaran B, Vaẞen R. HVOF Spraying of Fe-Based MMC Coatings with In Situ Formation of Hard Particles by Hot Isostatic Pressing. J Thermal Spray Technol 2012;21:344–354. [28] Lee JM, Kang SB, Sato T, Tezuka H, Kamio A. Fabrication of Al/Al3 Fe composites by plasma synthesis method. Mater Sci Eng A 2003;343:199–209. [29] Byakova AV, Kiz MM, Sirko AI, Yakovleva MS, Milman YuV. Cold-sprayed coatings based on high strength aluminium alloys reinforced by quasicrystalline particles: Microstructure and key properties. High Temp Mater Proc 2010;29:325–337. [30] Zhao L, Lugscheider E. Reactive Plasma Spraying of TiAl6V4 Alloy. Wear 2002;253:1214–1218. [31] Valente T, Galliano FP. Corrosion resistance properties of reactive plasma-sprayed titanium composite coatings. Surf Coat Technol 2000;127:86–92. [32] Yao Y, Wang Z, Zhou Z, Jiang S, Shao J. Study on reactive atmospheric plasma-sprayed in situ titanium compound composite coating. J Thermal Spray Technol 2013;22:509–517. [33] Tsunekawa Y, Okumiya M, Kobayashi T, Okuda M, Fukumoto M. Chromium Nitride in Situ Composites with a Compositional Gradient Formed by Reactive DC Plasma Spraying. J Thermal Spray Technol 1996;5:139–144. [34] Dudina DV, Korchagin MA, Zlobin SB, Ulianitsky VY, Lomovsky OI, Bulina NV, Bataev IA, Bataev VA. Compositional variations in the coatings formed by detonation spraying of Ti3 Al at different O2 /C2 H2 ratios. Intermetallics 2012;29:140–146. [35] Lomovsky OI, Dudina DV, Ulianitsky VY, Zlobin SB, Kosarev VF, Klinkov SV, Korchagin MA, Rozhkov IA, Kwon DH, Kim JS, Kwon YS. Formation of cold and detonation sprayed coatings from TiB2 -Cu nanocomposite powders produced by mechanical milling. Chem Sust Dev 2007;15:197–201. [36] Lomovsky OI, Dudina DV, Ulianitsky VY, Zlobin SB, Kosarev VF, Klinkov SV, Korchagin MA, Kwon DH, Kim JS, Kwon YS, Cold and detonation spraying of TiB2 -Cu nanocomposites, Mater Sci Forum 2007;534–536:1373–1376. [37] Shtertser A, Muders C, Veselov S, Zlobin S, Ulianitsky V, Jiang X, Bataev V. Computer controlled detonation spraying of WC/Co coatings containing MoS2 solid lubricant. Surf Coat Technol 2012;206:4763–4770. [38] Dudina DV, Zlobin SB, Bulina NV, Bychkov AL, Korolyuk VN, Ulianitsky VY, Lomovsky OI. Detonation spraying of TiO2 -2.5vol.% Ag powders in a reducing atmosphere. J Eur Ceram Soc 2012;32:815–821.

122 | 4 MMC coatings by thermal spraying [39] Dudina DV, Zlobin SB, Ulianitsky VY, Lomovsky OI, Bulina NV, Bataev IA, Bataev VA. Detonation spraying of TiO2 -Ag: controlling the phase composition and microstructure of the coatings. Ceram Trans 2012;237:161–169. [40] Dudina DV, Batraev IS, Ulianitsky VY, Korchagin MA. Possibilities of the computer-controlled eetonation spraying method: a chemistry viewpoint. Ceramics Intl 2014;40:3253–3260. [41] Dudina DV, Batraev IS, Ulianitsky VY, Korchagin MA, Golubkova GV, Abramov SY, Lomovsky OI. Control of interfacial interaction during detonation spraying of Ti3 SiC2 -Cu Composites. Inorg Mater 2014;50:35–39. [42] Kim JS, Kwon YS, Lomovsky OI, Dudina DV, Kosarev VF, Klinkov SV, Kwon DH, Smurov I. Cold spraying of in situ produced TiB2 -Cu nanocomposite powders. Comp Sci Tech 2007;67:2292–2296. [43] Zhou Y, Gu W. Chemical reaction and stability of Ti3 SiC2 in Cu during high-temperature processing of Cu/Ti3 SiC2 composites. Z Metallkd 2004;95:50–56. [44] Kosarev VF, Sova AA, Zlobin SB, Ulianitsky VY. Properties of copper and aluminium coatings deposited by cold spray and detonation spray, Strengthening Technologies and Coatings 2011;6:paper 4 (in Russian). [45] Mansurov ZA. Producing nanomaterials in combustion. Comb Expl Shock Waves 2012;48:561–569. [46] Shtertser AA, Ulianitsky VY, Batraev IS, Saprykin FI, Gromilov SA, Okotrub AV. Ultradispersed carbon by detonation spraying. Proc. V All-Russian Conf. “Interaction of high-intensity energy fluxes in advanced technologies and medicine”, Novosibirsk, Russia, 2013:283–286 (in Russian). [47] Hodder KJ, Izadi H, McDonald AG, Gerlich AP. Fabrication of aluminum-alumina metal matrix composites via cold gas dynamic spraying at low pressure followed by friction stir processing. Mater Sci Eng A 2012;556:114–121. [48] Morisada Y, Fujii H, Mizuno T, Abe G, Nagaoka T, Fukusumi M. Modification of thermally sprayed cemented carbide layer by friction stir processing. Surf Coat Technol 2010;204:2459–2464. [49] Solonenko OP, Ovcharenko VE, Ivanov YF, Golovin AA. Plasma Sprayed Metal-Ceramic Coatings and Modification of Their Structure with Pulsed Electron Beam Irradiation. J Thermal Spray Technol 2011;20:927–938. [50] Lia H, Khor KA, Yua LG, Cheang P. Microstructure modifications and phase transformation in plasma-sprayed WC-Co coatings following post-spray spark plasma sintering. Surf Coat Technol 2005;194:96–102. [51] Munir ZA, Anselmi-Tamburini U, Ohyanagi M. The effect of electric field and pressure on the synthesis and consolidation of materials: A review of the spark plasma sintering method. J Mater Sci 2006;41:763–777.

Alakesh Manna

5 Fabrication of Al-metal matrix composites by liquid stirring technique Abstract: This chapter presents the fabrication process of Al/SiC, Al/Grp and Al/Al2 O3 metal matrix composites with a liquid stirring technique. This chapter also presents the experimental investigation of the influence of stir casting parameters on various responses during casting of metal matrix composites (MMC). The Taguchi methodbased design of the experiment is used to optimize the stir casting process parameters for effective production of Al/SiC-MMCs. An orthogonal L27 (313 ) array has been used for 33 factorial design and analysis of variance (ANOVA) and is employed to indicate the most significant parameters affecting the casting performance characteristics such as micro hardness and tensile strength of the cast metal matrix composites samples. Analysis based on the test results indicated that the interaction of pouring temperature and stirring speed has great influence and the most significant interaction on micro hardness and tensile strength of prepared Al/15 wt.% SiC-MMC with 93.95 % and 90.93 % contribution respectively. Utilizing experimental results and the Gauss elimination method, mathematical models relating to the micro hardness and tensile strength are established to investigate the influence of casting parameters during stir casting of Al/SiC-MMCs.

5.1 Introduction Metal matrix composites have different important properties such as high specific strength and stiffness at elevated temperatures and good creep, fatigue and wear resistance. These properties are not achievable with lightweight monolithic metal or alloys individually. Particulate metal matrix composites have nearly isotropic properties when compared to long fiber-reinforced composites. An aluminum-silicon carbide metal matrix composite has low density and light weight, high temperature strength, hardness and stiffness, high fatigue strength and wear resistance, etc. in comparison to the monolithic materials as explained by Manna and Bhattacharyya [1]. Amongst various processing possibilities, stir casting is one of the promising routes available for fabrication of composites. The process is simple, flexible and applicable for largequantity production. The liquid phase methods with mechanical or electromagnetic agitation have low manufacturing costs and high productivity as explained by Seo and Kang [2], Rozak and Lewandowski [3], Hayashi and Tatsumoto [4], and Ohmi et al. [5]. Seo and Kang [2] fabricated the metal matrix composites by melt stirring. The author used the prepared composite specimens to study the mechanical properties and then compared the results with the extruded Al-specimens. Manoharan and Gupta

124 | 5 Fabrication of Al-metal matrix composites by liquid stirring technique [6] fabricated metal matrix composites by the stir casting route and concluded that the work hardening behavior of composites with different SiC volume fractiosn can be modeled using a modified continuum theory. Hashim et al. [7] studied the stir casting process in detail and concluded that stir casting is generally accepted as a particularly promising route as it is simple, flexible and has applicability to large-quantity production. Authors reported that the distribution of the reinforcement material in the matrix must be uniform for better wettability. Zhou and Xu [8] prepared two types of SiC particulate-reinforced composites by gravity casting and concluded that the SiC particles were observed to be located predominantly in interdendritic region and acted as substrates for heterogeneous nucleation of Si crystals. Naher et al.’s [9] simulation studied the stir casting process where liquid and semi-solid aluminum were replaced by other fluids. Authors concluded that a change in viscosity had a tremendous effect on SiC dispersion and settling time. Hunt et al. [10] studied the microstructure and the microstructural factors for various fracture modes for detection of fracture resistance of discontinuous reinforced aluminium metal matrix composite. They concluded that the thermo-mechanical properties could be improved through spatial distribution of reinforcement particles during thermo-mechanical processing of a SiCreinforced/ Al-metal matrix composite. Ourdjini et al. [11] studied the settling of particulate matter during casting of metal matrix composites and concluded that the settling measurements during isothermal holding showed that SiC particles settled at a much slower rate than predicted in the theoretical model. Hashim et al. [12] studied the effect of parameters on the final distribution of the particles in the matrix alloy and concluded that in order to have a homogeneous distribution of reinforcement, factors like particle density, size, shape, volume fraction and surface properties have considerable affect on mechanical properties. Nai and Gupta [13] studied the effects of different stirrer geometries (i.e. two bladed, four bladed and circular shaped) on the synthesis of Al/SiCp-particles functionally gradient materials. Authors concluded that the two-bladed stirrers yield the most satisfactory result in terms of uniform distribution, porosity levels and micro hardness when compared to other stirrer geometries. Aluminum-silicon alloys as a matrix material are characterized by light weight, good strength-to-weight ratio, ease of fabrication at reasonable cost, good thermal conductivity, excellent corrosion and wear resistance properties. However, aluminum alloy with a discontinuous ceramic reinforced MMC is rapidly replacing conventional materials in various automotive, aerospace and automotive industries as explained by Allison and Cole [14] and Manna and Bhattacharyya [15]. Amongst various processing routes, stir casting is one of the promising liquid metallurgy techniques utilized to fabricate the composites. The process is simple, flexible and applicable to large-quantity production. The liquid metallurgy technique is the most economical of all the available techniques to produce MMC as explained by Surappa [16] and Manna et al. [17]. Aluminum alloy-metal matrix composites containing 10 wt.% alumina of different mesh sizes were prepared by liquid metallurgy technique using the vortex method and explained the process by Yang et al. [18]. The ZnO whiskers 25 vol.% reinforced with

5.2 Fabrication of Aluminium metal matrix composites

|

125

Al-matrix composites were fabricated by a squeeze casting process; the process was explained by Guo et.al. [19]. The quartz-silicon dioxide particulates-reinforced LM6 alloy matrix composites were fabricated by carbon dioxide sand molding processes cited by Sulaiman et al. [20]. Various researchers have utilized conventional stir casting techniques for producing MMC such as by Srivatsan et.al. [21], Zhou and Xu [8], Gupta et al. [22], Manna et al. [17]. However, applied research on these areas is still needed for successful utilization of the process for manufacturing of MMCs. There are several fabrication techniques available for manufacturing metal matrix composites such as stir casting or compo-casting, liquid metal infiltration, squeeze casting, spray co-deposition, etc. Compo-casting involves the addition of particulate reinforcement into semi-solid metal by means of agitation. The compo-casting is usually accepted as a commercial route for producing MMCs but there is a problem in producing a homogeneous, high-density composite. This problem can be overcome by introducing a mechanical stirring and distributing the reinforced particles throughout the liquid matrix. This technique is known as liquid stirring technique. This technique is simple, flexible and applicable to produce MMCs irrespective of the quantity of production. Liquid stirring is attractive because, in principle, it allows for a conventional metal processing route and hence minimizes the total cost of the product. This liquid metallurgy technique is the most economical of all the available routes for production of metal matrix composites as explained by many authors as above. Various researchers, e.g. Nai and Gupta [13], Lloyd et. al. [23], Sulaiman et. al. [20], Meena et.al. [24], have also conducted the investigation on synthesis, characterization and analysis of various properties of metal matrix composites.

5.2 Fabrication of Aluminium metal matrix composites A stir casting setup has been designed and fabricated for casting of Al/SiC-MMC, Al/GrP -MMC and Al/Al2 O3 -MMCs. Utilizing the fabricated setup, the different sets of stir casting experiments have been carried out in a pre-planned way to investigate the effects of the various parameters on the quality and performance, e.g. physical and mechanical properties, of the casting samples. The casting experiments are performed with a variation in the percentages of weight fraction of reinforced particles, i.e. 5 wt.%, 10 wt.% and 15 wt.% during stir casting of Al/SiC-MMC, Al/GrP -MMC and Al/Al2 O3 -MMCs samples. The objective of the experimental investigation is to study the effect of the process variable, e.g. stirring speed (rpm), pouring temperature (∘ C) and time of stirring (min) during casting of MMCs samples and present the test results for optimal selection of the parameters setting, which may overcome the casting barriers. The sound casting of MMCs is very essential for widespread industrial applications of Al/SiCMMC, Al/GrP -MMC and Al/Al2 O3 -MMCs. To achieve such properties in the cast MMCs, it is essential to optimize the process parameters of the stir casting process.

126 | 5 Fabrication of Al-metal matrix composites by liquid stirring technique The fabricated stirring setup includes: (i) three heating furnaces of capacity 550 ∘ C, 1150 ∘ C and 1250 ∘ C that are simultaneously utilized for preheat and baking of fire clay-coated metal mold, to melt the aluminium metal matrix, and to preheat the reinforced particulates during stir casting of metal matrix composites (MMCs), respectively; (ii) a stirring arrangement; (iii) a speed regulator for the stirrer; (iv) a digital temperature regulator cum recorder; and (v) a tachometer. A motor is mounted on a spindle that is keyed to the impeller blades. The motor gives the rotary motion to the graphite impeller. The total arrangement is mounted on a rigid stand. A digital temperature recorder is also fitted to each of the furnaces to record the furnace temperature during preheating and melting of the aluminium matrix. This system of casting is one of the good alternatives to the stationary conventional method in casting the Al/Al2 O3 metal matrix composite. In this technique, aluminium metal matrix and hard reinforced particulates are required to melt and preheat simultaneously for casting such composites. The setup has temperature regulators with digital recorders; this helps to set the proper temperature with time as well as indicates the actual temperature of the furnaces. The setup has a speed regulator and tachometer; these are precisely used to control the stirring speed of the graphite impeller and record the rpm from the digital recorder. These are the most important in regulating and properly homogenizing the agitation of the mixture during the stirring of the molten metal matrix and hard reinforced particulates. A stopwatch is also fitted to the setup in order to measure the stirring time, time of heating, melting and mixing of the metal matrix and hard particulates.

5.2.1 Fabrication of the stirring arrangement Three different furnaces are designed and fabricated for stir casting of metal matrix composites. The Postconsumer Recycled Content percentage worksheet (PCRC sheet, 22 S.G.) with a double-coated metallic powder coating is selected for the furnace wall material. The sheet has been cut and bent to form a final box of 150 mm × 150 mm × 300 mm. Zirconia’s cerewool insulation is used and placed in between the outer cover and the heat generator, which insures a uniform distribution of heat. The basic aim of using this insulation is to avoid heat loss. Kainthal Al wire (18 S.W.G.) in coil form is used as a heating element. This design is suitable for working with a single phase, 220V, 4 kW, supply. A microprocessor-based PID digital temperature indicator cum controller is fitted to the furnaces. The accuracy of the indicator is ± 2 ∘ C. The DC motor has a capacity of 0.092 kW; a maximum speed limit of 400 rpm is chosen to rotate the graphite impeller. A variable speed controller is used to vary the speed of the impeller from 100 to 400 rpm. The graphite rod (25 mm in diameter) is used as a spindle fitted with an impeller that has three blades with a sweep capacity of 37 mm. The stirrer is mounted on an adjustable heavy stand and connected with a tachometer to measure

5.2 Fabrication of Aluminium metal matrix composites

|

127

the rpm of the graphite spindle, i.e. graphite impeller. Graphite crucibles of different capacities are chosen for melting the metals. After melting the aluminium metal matrix, it is necessary to add the hard reinforced particulates and mix these properly for proper homogenizing of the reinforced particulate in the molten metal matrix. To serve this purpose, a stirring mechanism was designed and fabricated. A speed regulator with a tachometer is also fitted to the system to regulate the stirring speed and record the speed during mixing of the hard reinforced particulate and matrix. Figure 5.1 shows a fabricated stir casting setup with a furnace.

Speed regulator for stirrer

Tachometer

Electrical motor operated stirrer

Furnace

Fig. 5.1. Fabricated stir casting setup.

5.2.2 Mold-making and preparation of the mold cavity A metallic mold is prepared and utilized for casting. IS 2002-1962/ high temperature service pipe, medium grade with a 40 mm nominal diameter × 3.15 mm wall thickness × 250 mm length each is used to prepare the cylindrical metal mold. The 40 mmdiameter pipe (inside) is longitudinally cut into two halves and a square base of 40 mm × 40 mm cut from an IS-1079-1968/5 mm-thick plate and welded at the bottom of a half piece of 40 mm-diameter pipe. The prepared two halves of the molds are clamped to make an exactly cylindrical mold cavity by means of two clamps as shown in Figure 5.2a and b. Figure 5.2b shows the complete metal mold cavity used to cast the MMC samples. Fire clay is prepared properly and coats the inside of both halves of the metal mold. Then the clay-coated metal mold is allowed to dry in the sunlight for two hours. After that, the partially dry clay-coated mold is put into the furnace for baking and completely dried before pouring the molten metal matrix and hard reinforced particles mixture for casting.

128 | 5 Fabrication of Al-metal matrix composites by liquid stirring technique

(a)

(b)

Fig. 5.2. (a) Exploded view of mould assembly. (b) Clamped mould assembly.

5.2.3 Estimation of raw materials for Al/5, 10, 15 wt.% reinforced MMC casting In Al/SiC-MMC casting requires aluminium metal as a matrix and SiC as hard reinforced particulates. It is thus essential to estimate the necessary weight of the aluminium matrix and the proportional weight of the precise amount of required SiC particles for Al/5 wt.%, Al/10 wt.% and Al/15 wt.% SiC-MMCs casting. The wt.% SiC is calculated using the following mathematical relation given below: composition (% by weight) =

wp W m + wp

,

(5.1)

where wp is the mass of reinforced particulates, and Wm is mass of metal matrix. The commercially available AA6061/Al alloy is used as a metal matrix for casting three different MMCs with different weight percentages of hard reinforced particulates. Table 5.1 shows the chemical composition of the commercially available Al alloy used as metal matrix for casting. Table 5.1. Chemical composition of AA6061/Al alloy used as metal Matrix (in wt.%). Matrix alloy

Si

Fe

Cu

Mn

Mg

Zn

Ti

Balance

AA6061

0.6

0.7

0.3

0.15

0.9

0.25

0.15

Al

Three different types of hard particulates such as Silicon carbide (SiCp ), Alumina (Al2 03 ) and Graphite (Grp ) are selected as reinforced particles and used in casting of three different metal matrix composites. The average particle size of all three types of reinforcements used for casting is 34 μm.

5.3 Physical, chemical and mechanical properties of stir cast samples

|

129

5.2.4 Experimental procedure The melting of commercially available AA6061/Al metal matrix is carried out in a graphite crucible placed in a resistance furnace. The commercially available AA6061/Al is preheated at 450 ∘ C for 3–4 hours before melting. The furnace temperature is first raised to and above the liquidus temperature of commercially available AA6061/Al in order to completely melt the aluminium ingot; it is then cooled down just below the liquidus to keep the slurry in a semi-solid state. A hexochloroethane tablet is used as a degassing agent to remove any unwanted gases generated during the melting of the Al-metal matrix. Simultaneously, the estimated amount of hard particulate reinforced particles is put in another graphite crucible into the second furnace with a capacity of 1250 ∘ C. The reinforced particles are preheated to 1050 ∘ C for 3–4 hours. Manual mixing is done during preheating to ensure proper heat distribution and equal agitation of the hard particles. The preheated reinforced particles are mixed with the molten Al matrix and stirred for proper mixing with variation of impeller speed from 150 to 300 rpm of the developed mechanical stirring arrangement. Simultaneously, a the preheating and baking of a fire clay-coated metal mold is done in the third furnace with a capacity of 550 ∘ C with the temperature maintained at 430 ∘ C for 2–3 hours. A sand bed is used to hold the baked metal mold. Finally the molten mixture of Al matrix and reinforced particulates is poured into the mold and allowed to solidify.

5.3 Physical, chemical and mechanical properties of stir cast samples 5.3.1 Physical property of stir cast samples Prepared stir cast metal matrix composite samples are tested for physical, mechanical and microstructural properties. Density of the specimen is measured using the Archimedes principle. Quantitative assessment of particulates in stir cast metal matrix composites is done with an image analyzer. Microstructural characterization of the prepared MMC samples are carried out using an optical microscope and a scanning electron microscope on the metallographically polished specimen to investigate the distribution of reinforced particulates, interfacial bonding and presence of porosity. Micro-hardness and ultimate tensile strength of the specimens are also investigated in various tests. Tables 5.2, 5.3 and 5.4 show the densities of prepared Al/SiC-MMC, Al/Al2 O3 -MMC and Al/Grp -MMC samples respectively. From Tables 5.2 and 5.3, it is clear that the density of the fabricated MMC slightly increases with an increase in the weight fraction of reinforced particles. From Table 5.2, the density of Al/5 wt.% SiC-MMC is 2.76 ±0.03, while for Al/15 wt.% SiC-MMC, the density is 2.81 ±0.04, hence with the increase in the reinforcement content the density of the stir cast MMC is found to be increased. From the above table a mathematical

130 | 5 Fabrication of Al-metal matrix composites by liquid stirring technique Table 5.2. Densities of prepared Al/SiC-MMC samples. Property Density, g/cm3

Aluminum (6061-T6)

Titanium (6Al-4V)

Steel (4340)

2.75

4.43

7.76

Reinforcement (SiC) fraction 5% 10 % 15 % 2.76 ± 0.03

2.79 ± 0.02

2.81 ± 0.04

Table 5.3. Densities of prepared Al/Al2 O3 -MMC samples. Property Density, g/cm3

Aluminum (6061-T6)

Titanium (6Al-4V)

Steel (4340)

2.75

4.43

7.76

Reinforcement (Al2 O3 ) fraction 5% 10 % 15 % 2.80 ± 0.04

2.85 ± 0.03

2.91 ± 0.04

Table 5.4. Densities of prepared Al/Grp -MMC samples. Property Density, g/cm3

Aluminum (6061-T6)

Titanium (6Al-4V)

Steel (4340)

2.75

4.43

7.76

Reinforcement (Grp) fraction 5% 10 % 15 % 2.69 ± 0.02

2.65 ± 0.04

2.62 ± 0.03

relation can be formulated as follows: ρmmc = ρrfp vrfp + ρm vm

(5.2)

where, ρmmc is the density of cast MMC, ρrfp is the density of reinforced particles, vrfp is the volume fraction of reinforced particles, ρm is the density of Al-matrix, and vm is the volume fraction of matrix material. The density of Al/Al2 O3 -MMC is slightly higher when compared to cast Al/SiCpMMC. It is because the density of alumina particulate is slightly higher than silicon carbide particulate. As average measured density of alumina and silicon carbide particulates are 3.95 and 3.21 g/cc respectively. The density of Al/GrP -MMC is the lowest when compared to the all other cast samples (Table 5.4). It is due to the lower density of graphite particles used for casting, i.e. 2.2 g/cc.

5.3.2 Mechanical properties of stir cast samples Ultimate tensile strength and micro hardness tests are carried out to evaluate the mechanical properties of the prepared stir cast specimens and test results presented in this article. In the present work, the Vickers hardness tests are carried out on the HMV micro-hardness tester using a diamond pyramid indenter. During the testing, 50g of load is applied for 20 seconds and the average of three readings have been recorded and reported here. The ultimate tensile tests are conducted on the Universal Testing Machine HT2107A (HUNGTA) with a 300T capacity. The circular cast specimen with a 40 mm di-

5.3 Physical, chemical and mechanical properties of stir cast samples

|

131

ameter and 200 mm gauge length are used for this test. Testing is performed at room temperature with constant 0.5 mm/min speed. The results of micro hardness and ultimate tensile strength tests are reported in Table 5.5. Table 5.5 represents the test results of micro-hardness (VHN) and tensile strength (MPa) of prepared Al/ 15 %wt SiC-MMC, Al/15 %wt Al2 O3 -MMC and Al/ 15 %wt. Grp-MMC.

5.3.3 Analysis of the reinforced weight fraction Quantitative assessment of reinforcement particles with their weight fractions is also analyzed in various images that utilize an eM power image analyzer (version 3.0.0.9) and the results from one test sample are shown in Figure 5.3. From Figure 5.3, it is clear that the percentage of weight fraction was found to be less than the actual particulates added. In Al/15 wt.% Al2 O3 -MMC, the actual weight fraction is 14.911 % instead of 15 %. Similarly for Al/10 wt.% SiC-MMC and Al/10 vol.% Grp-MMC: the weight fractions are 9.452 % and 9.956 %, respectively (not shown here). Some losses of reinforced particulates are always found and this loss may occur because some of them are dragged out with slugs.

5.3.4 Microstructural characterization Microstructural characterization of the stir cast specimens are carried out using an optical microscope and JSM 6100 (JEOL) Scanning Electron Microscope to study the distribution of SiC, Al2 O3 and Grp and their interfacial bonding with the Al matrix. Microstructural investigations are carried out on etched specimens. Keller reagent with a composition of 0.5 HF–1.5 HCl–2.5 HNO3 –95.5 H2 O is used as an etchent. The results of the optical micrograph together with SEM studies of some of prepared samples of Al/15 vol.% Al2 O3 -MMC; l/15 vol.% GrP -MMC; and Al/15 vol.% SiC-MMC are shown in Figures 5.4–5.6. Optical micrograph and SEM studies carried out on the stir cast composite samples in unetched and etched conditions reveal the good distribution of reinforcement particulates in the aluminum matrix. Different parametric settings produce different distributions of reinforcements, as is clearly seen in optical micrographs. In addition, at higher pouring temperature, low stirring speed and low stirring time there is a strong tendency of the particulte to aggloromate in all stir cast metal matrix composites.

Pouring temp (∘ C)

750 750 750 825 825 825 900 900 900 825 825 825 900 900 900 750 750 750 900 900 900 750 750 750 825 825 825

Expt no

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

150 150 150 225 225 225 300 300 300 300 300 300 150 150 150 225 225 225 225 225 225 300 300 300 150 150 150

Shear rate (rpm) 3 6 9 3 6 9 3 6 9 3 6 9 3 6 9 3 6 9 3 6 9 3 6 9 3 6 9

Stirring time (min) 177.9 178.2 178.8 182.2 183.3 183.9 184.1 184.8 185.1 175.3 176.7 174.9 181.6 182.0 180.8 182.8 183.5 182.2 174.3 173.2 173.4 177.7 175.7 176.0 181.4 179.0 179.2

Y1 178.1 178.3 178.9 182.6 183.0 183.2 184.0 184.2 185.6 175.1 176.8 174.7 181.9 182.4 180.5 182.7 183.2 182.4 174.2 173.9 173.6 177.2 175.4 176.3 181.0 179.3 179.1

Y2 178.0 178.5 178.7 182.8 183.1 183.7 184.3 184.4 185.4 175.0 176.5 174.4 181.8 182.3 180.7 182.9 183.8 182.0 174.8 173.0 173.1 177.4 175.9 176.1 181.3 179.9 179.4

Y3 178.0 178.3 178.8 182.5 183.1 183.6 184.1 184.5 185.4 175.1 176.7 174.7 181.8 182.2 180.7 182.8 183.5 182.2 174.4 173.4 173.3 177.4 175.7 176.1 181.2 179.4 179.2

Average

Al/15 % wtSiC-MMC Micro hardness (VHN)

Table 5.5. Micro-hardness (VHN) and tensile strength (MPa) of stir cast MMCs.

315.67 316.36 318.29 320.72 321.26 322.27 322.52 323.84 324.28 313.58 314.62 313.22 319.55 320.16 318.89 319.77 321.55 320.47 312.82 312.61 312.28 314.83 313.82 314.34 319.24 315.23 316.59

Tensile strength (MPa) 161.4 161.6 162.5 164.6 165.3 165.4 166.4 166.6 166.5 159.5 160.4 159.3 163.4 164.3 162.6 163.7 165.6 164.4 158.6 158.4 158.2 160.6 159.5 160.3 163.3 161.4 162.9

305.44 306.42 307.24 309.43 309.74 310.21 310.84 311.44 311.83 303.52 304.52 303.21 308.26 308.68 307.55 308.46 309.92 308.85 302.81 302.52 302.26 304.78 303.84 304.28 307.76 305.72 306.74

Al/15 %wt Al2 O3 -MMC Tensile Micro hardness strength (MPa) (VHN) 90.3 90.7 91.3 93.6 94.2 94.6 95.3 95.4 95.6 88.6 89.5 88.1 92.7 93.1 91.6 92.5 94.3 93.4 87.7 87.6 87.3 89.8 88.7 89.2 92.2 90.4 91.3

230.27 231.30 232.35 233.93 234.16 234.74 235.44 235.83 236.45 228.55 229.57 228.25 233.14 233.51 232.62 233.31 235.36 233.72 227.83 227.52 227.26 229.89 228.83 229.26 232.89 230.69 231.83

Al/15 %wt Grp-MMC Tensile Micro hardness strength (MPa) (VHN)

132 | 5 Fabrication of Al-metal matrix composites by liquid stirring technique

5.3 Physical, chemical and mechanical properties of stir cast samples |

Al2O3 Particles

100

80

A F %

60

40

20

0 Alumina, Al Matrix SID:Alumina Frame Area(mm2) Frame - 1 0.2272 0.2272 Average

Alumina(%) 14.911 14.911

Al Matrix(%) 85.089 85.089

Fig. 5.3. Microstructure and reinforcement fraction of Al/15 wt.% Al2 O3 -MMC.

133

134 | 5 Fabrication of Al-metal matrix composites by liquid stirring technique

Al2O3

Fig. 5.4. Optical micrograph of 15 wt.% Al2 O3 MMC, at pouring temp 850 ∘ C, 225 rpm, 6 min.

100×

Grp

100×

Fig. 5.5. Optical micrograph of 15wt.%GrpMMC, at pouring temp 825 ∘ C, 300 rpm, 6 min.

SiC

100×

Fig. 5.6. Optical micrograph of 15wt.% SiC-MMC at pouring temp 900 ∘ C, 300 rpm, 9 min.

5.4 Optimization of stir casting parameters for Al/15 wt.% SiC-MMC |

135

5.4 Optimization of stir casting parameters for Al/15 wt.% SiC-MMC The optimizations of the casting parameter are important in order to obtain sound casting with a maximum possible micro hardness and tensile strength of the prepared composites without eliminating or controlling the uncountable parameters. Keeping in view, the exhaustive experimental study is being carried out to achieve the optimal parametric combination of maximum micro hardness and tensile strength during casting of Al/15 wt.% SiC-MMC. The Taguchi method, a powerful tool in the design of experiments, is used to optimize the casting parameters for effective casting of Al/15 wt.% SiC-MMC by the stir casting technique. An orthogonal L27 (313 ) array is used for 33 factorial design and the analysis of variance (ANOVA) is employed to investigate the influence of pouring temperature, stirring speed and stirring time on the prepared composite performance, e.g. mechanical properties like micro hardness and tensile strength. By considering casting parameters for experimentations and by means of multiple linear regressions, mathematical models relating to the micro hardness and tensile strength are established to investigate the influence of various casting parameters for casting Al/SiC-MMC. Three casting parameters such as pouring temperature, stirring speed and stirring time are considered as controlling factors; each parameter has three levels, namely small, medium and large. For further analysis they are denoted by 1, 2 and 3 respectively. Table 5.6 shows the casting parameters and their levels considered for experimentation. Table 5.6. Casting parameters and their level. Sr. No.

Casting parameters 1

1 2 3

A: Pouring temperature, ∘ C B: Stirring speed, rpm C: Stirring time, min

750 150 3

Level 2 825 225 6

3 900 300 9

5.4.1 S/N Ratio for micro-hardness of prepared Al/15 wt.% SiC-MMC Figure 5.7 shows S/N ratio graph of different factor levels for micro hardness. From the Figure 5.7, it is concluded that for higher micro hardness, the optimal parametric setting is A3 B3 C1 , i.e., at 900 ∘ C pouring temperature, 300 rpm stirring speed and 3 min stirring time a higher micro hardness of the stir cast Al/15 wt.% SiC-MMC ingot can be potentially achieved.

Mean S/N ratio (dB)

136 | 5 Fabrication of Al-metal matrix composites by liquid stirring technique 45.8 45.7 45.6 45.5 45.4 45.3 45.2 45.1 45 44.9 44.8 44.7

S/N ratio Grand mean

A1

A2

A3

B1 B2 B3 Casting parameter level

C1

C2

C3

Fig. 5.7. S/N ratio for microhardness of prepared Al/15 wt.% SiC-MMC.

5.4.2 ANOVA for micro hardness of prepared Al/15 wt.% SiC-MMC Table 5.7 shows the ANOVA and ‘F’ test values for micro hardness of a prepared Al/15 wt.% SiC-MMC cast sample. From Table 5.7, it is clear that the interaction of pouring temperature and stirring speed has great influence, i.e. the most significant interaction on micro hardness of prepared Al/15 wt.% SiC-MMC with a 93.95 % contribution. Stirring speed is the second influencing factor, i.e. the significant parameter with 19.9 ‘F’ test value on the micro hardness of the composite. Table 5.7. ANOVA and ‘F’ test for microhardness of Al/15 wt.%SiC-MMC cast sample. Parameters X1 X2 X3 X1.X2 X1.X3 X2.X3 Error Total

Degree of freedom

Sum of square

Variance

‘F’ test value

% of contribution

2 2 2 4 4 4 62 80

8.13 20.71 2.19 1008.30 0.54 1.06 32.27 1073.20

4.066 10.356 1.097 252.074 0.135 0.265 0.520

7.81 19.90 2.11 484.35 0.26 0.51

0.75 1.93 0.21 93.95 0.05 0.10 3.01 100.00

X1: Pouring temperature; X2: Stirring speed; X3: Stirring time

5.4 Optimization of stir casting parameters for Al/15 wt.% SiC-MMC | 137

5.4.3 Mathematical model for micro hardness of prepared Al/15 wt.% SiC-MMC In consideration of the most significant and significant parameters as identified from Table 5.8 and using the Gauss elimination method, the mathematical model for micro hardness has been developed with a notation of X1 , X2 and X3 , which represent the pouring temperature, stirring speed and stirring time, respectively. The mathematical model for micro hardness of prepared Al/15 vol.% SiC-MMC is as follows: Ymicro Al/15 wt.%-sic = 149.875859 − 0.0069802 X1 − 0.15919 X2 − 0.111112 X3 + 0.0002 X1 .X2 − 0.0061728 X1 .X3 + 0.06543X2 .X3 + 0.000014485 X12 + 0.00008427 X22 − 0.012551 X32 ,

(5.3)

where R2 = 0.97.

5.4.4 S/N Ratio for tensile strength of prepared Al/15 wt.% SiC-MMC Figure 5.8 shows S/N ratio graph of different factor levels for tensile strength. From Figure 5.8, it is concluded that for maximum tensile strength, the optimal parametric setting is A3 B1 C1 , i.e., at 900 ∘ C pouring temperature, 150 rpm stirring speed and 3 min stirring time. 50.07 50.06 50.05 Mean S/N ratio (dB)

50.04 50.03 50.02 50.01 50 S/N ratio Grand mean

49.99 49.98 49.97 A1

A2

A3

B1 B2 B3 Casting parameter level

C1

Fig. 5.8. S/N ratio for tensile strength of prepared Al/15.wt% SiC-MMC.

C2

C3

138 | 5 Fabrication of Al-metal matrix composites by liquid stirring technique 5.4.5 ANOVA for tensile strength of prepared Al/15 wt.% SiC-MMC Table 5.8 shows the ANOVA and ‘F’ test values for tensile strength of prepared Al/15 wt.% SiC-MMC. From Table 5.8, it is clear that the interaction of pouring temperature and stirring speed has great influence on tensile strength of prepared Al/15 wt.% SiCp-MMC with a 90.93 % contribution. Pouring temperature is the second influencing factor, i.e. significant factor with a 20.11 ‘F’ test value on the tensile strength of the prepared Al/15 wt.% SiC-MMC composite by stir cast method. Table 5.8. ANOVA and ‘F’ test for Tensile strength of Al/15 wt.% SiC-MMC cast sample. Parameters

Degree of freedom

Sum of square

Variance

‘F’ test value

% of contribution

2 2 2 4 4 4 62 80

27.55 12.73 0.64 984.85 7.97 6.78 42.46 1082.97

13.7750 6.6365 0.3180 246.2120 1.9920 1.6940 0.6850

20.11 9.29 0.46 359.52 2.91 2.47

2.54 1.18 0.06 90.93 0.74 0.63 3.92 100.00

X1 X2 X3 X1.X2 X1.X3 X2.X3 Error Total

X1: Pouring temperature; X2: Stirring speed; X3: Stirring time

5.4.6 Mathematical model for tensile strength of prepared Al/15 wt.% SiC-MMC In consideration of the most significant parameters as identified in Table 5.8 and using the Gauss elimination method, the mathematical model for tensile strength has been developed with a notation of X1 , X2 and X3 , which represent the pouring temperature, stirring speed and stirring time, respectively. The mathematical model for tensile strength of prepared Al/15 wt.% SiC-MMC is as follows YTensile-sic = 316.821727 − 0.00190651 X1 − 0.0018866X2 + 0.0056561 X3 + 0.000002872 X1 .X2 − 0.000008432 X1 .X3 + 0.0000059135 X2 .X3 + 0.0000008477 X12 − 0.0000012332 X22 + 0.00002736 X32 , where R2 = 0.96.

(5.4)

5.5 Conclusion

|

139

5.5 Conclusion The processing of Al/SiC-MMC, Al/Grp -MMC and Al/Al2 O3 -MMC can be made possible by a stir casting process. To achieve higher micro hardness and tensile strength of the cast Al/15 wt.% SiC-MMC, the optimal parametric combinations are A3 B3 C1 (i.e. at 900 ∘ C pouring temperature, 300 rpm stirring speed and 3 min stirring time) and A3 B1 C1 (i.e. at 900 ∘ C pouring temperature, 150 rpm stirring speed and 3 min stirring time), respectively. The interaction of pouring temperature and stirring speed has great influence, i.e. the most significant interaction on micro hardness and tensile strength of prepared Al/15 wt.% SiC-MMC with a 93.95 % and 90.93 % contribution, respectively. Stirring speed is the second influencing factor, i.e. significant parameter with a 19.9 ‘F’ test value on the composite’s micro hardness. The mathematical models developed for the micro hardness and tensile strength of the cast samples have been successfully proposed for the proper evolution of casting parameters during the stir casting of MMCs.

References [1]

[2]

[3]

[4] [5]

[6]

[7] [8] [9] [10] [11]

Manna A, Bhattacharyya B. Study on Different Tooling Systems during Turning for Effective Machining of Al/SiC-MMC. The Institution of Engineers (India) Journal-Production 2003;83:46–50. Seo YH, Kang CG. Effects of hot extrusion through a curved die on the mechanical properties of SiCp/Al composites fabricated by melt stirring. J. of Composites Science and Technology 1999;59:643–654. Rozak GA, Lewandowski JJ. Effects of casting conditions and deformation processing on A356 aluminium and A356-20 vol.%SiC composites. Journal of Composite Materials 1992;26(14):2076–2106. Hayashi M, Tatsumoto N. Processing and Mechanical properties of SiC particle dispersion strengthened AC8A composite by agitation casting. Journal of Casting 1993;65(11):846–852. Ohmi T, Minoguchi K, Kudoh M, Itoh Y, Matsuura K. Change in apparent viscosity of stirred slurry during production of semisolid hypereutectic Al-Si alloy by slurry-melt mixing process. Journal of Japan Institute of Light Metals 1994; 44(9):504–509. Manoharan M, Gupta M. Effect of silicon carbide volume fraction on the work hardening behaviour of thermomechanically processes aluminium based metal matrix composite. J. of Composites Part B 1999;30:107–112. Hashim J, Looney L, Hashmi MSJ. Metal matrix composites: production by stir casting method. J of Materials Processing Technology 1999;92-93:1–7. Zhou WZ, Xu M. Casting of SiC Reinforced Metal Matrix Composite. J. of Materials Processing Technology 1997;63:358–363. Naher S, Brabazon D, Looney L. Simulation of the stir casting process. J. of Materials Processing Technology 2003;143-144:567–571. Hunt, WH, Osman, TM, Lewandowski JJ. (). Micro and Macrostructural Factors in DRA Fracture Resistance. JOM 1993;45(1):30–35. Ourdjini A, Chew KC, Khoo BT. Settling of silicon carbide particles in cast metal matrix composites. J. of Materials Processing Technology 2001;116:72–76.

140 | 5 Fabrication of Al-metal matrix composites by liquid stirring technique [12] Hashim J, Looney L, Hashmi MSJ. Particle distribution in cast metal matrix composites Part I. J. of Materials Processing Technology 2002;123:251–257. [13] Nai SML, Gupta M. Synthesis and characterization of free standing, bulk Al/SiCp functionally gradient materials: effects of different stirrer geometries. Materials Research Bulletin 2003;38:1573–1589. [14] Allison JE, Cole GS. Metal Matrix Composite in the Automotive Industry: Opportunities and Challenges. Journal of Materials 1993;45(1):19–24. [15] Manna A, Bhattacharyya B. Influence of Machining parameters on the Machinability of particulate reinforced Al/SiC-MMC. Int. J. of Advanced Manufacturing Technology 2005;25:850–856. [16] Surappa MK. Microstructure Evolution during Solidification of DRMMCs: State of the Art. Journal of Materials Processing Technology 1997;63:325–333. [17] Manna A, Mahapatra PB, Bains HS. Experimental study on processing of Al-Al2 O3 /Grp -MMC by liquid stirring. Journal of Composite Materials 2010; 44(25):3069–3079. [18] Yang J, Pickard SM, Ctedy C, Evans AG, Mehrabian T. The Stress/Strain Behavior of Aluminium Matrix Composites with Discontinuous Reinforcements. Acta Metallurgica Materialia 1991;39:1863–1870. [19] Guo ;, Xiong ;, Yang M, Li W. Microstructure and Properties of Tetrapod-like ZnO Whiskers Reinforced Al Matrix Composite. Journal of Alloy and Compounds 2008;461:342–345. [20] Sulaiman S, Sayuti M, Samin R. Mechanical Properties of the As-Cast Quartz Particulate Reinforced LM6 Alloy Matrix Composites. Journal of Material Processing Technology 2008;201;731–735. [21] Srivatsan TS, Sudarshan TS, Lavernia EJ. Processing of Discontinuously-Reinforced Metal Matrix Composites by Rapid Solidification. Progress in Material Science 1995;39:317–409. [22] Gupta ;, Lai ;O, Lim CYH. Development of a Novel Hybrid Aluminium Based Composite with Enhanced Properties. Journal of Materials Processing Technology 2006;176:191–199. [23] Lloyd ;J, Lagace ;, Mcleod A, Morris PL. Micro Structural Aspect of Aluminium Silicon Carbide Particulate Composites Produced by a Casting Method. Materials Science and Engineering 1989;A107:73–79. [24] Meena KL, Manna A, Banwait SS, Jaswanti D. An analysis of mechanical properties of the developed Al/SiC-MMCs, American Journal of Mechanical Engineering 2013;1(1):14–19.

Inderdeep Singh, Saurabh Chaitanya, Ravinder Kumar

6 Material removal processes for metal matrix composites Abstract: The metal matrix composites have gained a prime importance in the industrial sector because of their superior properties such as light weight, high strength and stiffness, high toughness and impact strength, low sensitivity to thermal shocks and good fatigue resistance. With depleting petroleum resources, the need to develop lightweight parts to be used in automotive and aerospace applications has increased in order to achieve good fuel efficiency. Efforts have been made to fabricate MMCs in near net shape, but machining is still required to achieve special features, surface finish and desired profiles. This chapter concentrates on the material removal processes (conventional and unconventional) used for the machining of particulate reinforced metal matrix composites. The mechanism of material removal in the case of conventional machining (turning, milling and drilling) and unconventional machining (electrochemical machining, electric discharge machining and ultrasonic machining) processes has been studied, and the effect of various input variables on the output responses has been discussed.

6.1 Introduction The metal matrix composites (MMCs) have been a keen area of research for scientists and researchers worldwide. The depleting petroleum resources and increasing fuel prices have forced researchers to develop and use lightweight materials like MMCs in automotive and aerospace applications. MMCs offer the distinct advantages of being lightweight and having high strength and stiffness, low density and excellent corrosion and wear resistance. MMCs are used in a variety of applications ranging from automobiles, aircraft, marine applications and consumer goods. MMCs provide huge flexibility as they can be tailor made according to the design requirements by varying the size, shape and quantity of the reinforcement. The reinforcements can be added in the form of continuous fibers, whiskers or particulates. The major types of matrixes used for MMCs are aluminium/aluminium alloy, magnesium/magnesium alloy, titanium/titanium alloy, copper/copper alloy and super alloys. Particulate reinforced metal matrix composites (PMMCs) are the most widely used MMCs as they provide higher ductility, lower anisotropy, higher wear resistance and are easier to fabricate than fiber-reinforced MMCs. The reinforcements used in PMMCs are in the form of other metal, organic compound or ceramics (oxides, carbides) like silicon carbide (SiC), alumina (Al2 O3 ), boron carbide (B4 C) and titanium carbide (TiC). Of the various techniques of fabricating PMMCs, casting (stir or ultrasonic vibration

142 | 6 Material removal processes for metal matrix composites assisted) is considered to be the most economical and efficient fabrication process. The parts that are manufactured by casting often require further machining operations before they can be used or assembled. Material removal by machining PMMCs is considered an inevitable secondary process which results in a desired product to be used in engineering applications. In engineering applications, parts must comply with the specific design requirements in terms of dimensional accuracy and surface finish to ensure proper functioning and reliability during their expected service life. The material removal process helps in removing excess material (intentionally left as machining allowance) to obtain the desired dimensional accuracy as well as surface finish that is not possible in the casting process. Moreover, for the pre-assembly operations like cutting slots or making holes, PMMC machining is required. Hence, to obtain the final useful product made of PMMCs, various machining operations might be required. But in PMMCs, the presence of hard abrasive particles tends to limit the machinability of the composites. The material removal processes are broadly classified as conventional and unconventional processes. In case of conventional machining, material removal is only possible if all of the following conditions are satisfied: 1. the material of the tool should be harder than the material of the workpiece; 2. the tool should be able to penetrate into the workpiece to be cut; 3. there should be a relative motion between the tool and the workpiece. If any of these conditions is not fulfilled, then material removal is achieved by unconventional machining processes. The material removal by conventional processes involves the traditional machining processes like turning, milling, shaping, drilling, etc., while the unconventional processes include water jet machining (WJM), ultrasonic machining (USM), electrical discharge machining (EDM) and electrochemical machining (ECM).

6.2 Conventional machining processes El-Hofy (2005) classified conventional machining processes according to the process of material removal as cutting and mechanical abrasion. These are material removal processes in which material removal takes place by cutting in the form of macroscopic chips turning, milling, shaping and drilling. In these material removal processes, the tool penetrates the workpiece to the desired cut depth and a relative motion between the tool and the workpiece results in generating the shape of the product. The relative motion is in the form of main motion (for example, rotary in case of turning, and linear in case of shaping) and feed motion. The conventional machining of PMMCs is considered to be difficult, since PMMCs consist of two or more distinct phases, of which one are hard abrasive particles such as SiC, Al2 O3 , etc. These hard particles are well dispersed in a ductile matrix, resulting in a composite having high strength and hardness. Conventional machining processes

6.2 Conventional machining processes

|

143

that are used for material removal of PMMCs are the same as used for conventional metals. In conventional machining of PMMCs, the factors affecting the machinability are classified as matrix material, percentage of reinforcement, type of tool and machining parameters. These factors also influence the output parameters such as chip formation, thrust force, torque and surface roughness. Some of the problems associated with conventional machining processes such as turning, milling and drilling are high cutting forces, poor surface finish, built-up edge (BUE) formation, rapid tool wear, burr formation, out of roundness, etc. These problems can be reduced or eliminated by using optimum process parameters and suitable tool material and geometry. The researchers have studied the effect of various machining parameters and have predicted the optimum cutting conditions for employing conventional material removal processes.

6.2.1 Turning of PMMCs A turning process is carried out on a lathe machine and is mostly used for generating cylindrical profiles. In this process, the tool used is a single point cutting tool and is fixed on a tool post, whereas the workpiece rotates. The material removal takes place due to the severe plastic deformation and shearing occurring at the tool-workpiece interface. The material is removed in the form of chips. In case of PMMCs, the presence of hard abrasive particles causes severe tool wear, high surface roughness and an increase in cutting forces. Manna and Bhattacharayya (2005) conducted an experimental investigation to study the effect of processing parameters like cutting speed, feed and depth of cut on the machinability of SiC reinforced Aluminium matrix composites during turning using a rhombic uncoated carbide tool. It was reported that low cutting speed leads to high flank wear due to the generation of high cutting forces and formation of BUE. Also, higher cutting speeds lead to a significant increase in the cutting edge temperature, causing rapid tool deformation. They recommended that the cutting speed should lie between 60 m/min to 150 m/min while machining Al/SiC-PMMCs and concluded that tool wear is less sensitive to variation in feed rate as compared to the cutting speed. Yanming and Zehua (2000) conducted a detailed investigation into the factors leading to tool wear during machining of SiC-reinforced PMMCs and reported an abrasive wear behaviour for conventional tools and a brittle fracture for high hardness tools. It was concluded that the size and volume fraction of SiC particles are the most significant factors affecting tool life. Also, harder tools are recommended for machining coarse reinforcements, while conventional tools can be used for finer reinforcements. El-Gallab and Sklad (1998) studied the dry highspeed turning of 20 % SiC/Al-PMMCs for the selection of optimum machining parameters and cutting tool (material and geometry). It was reported that polycrystalline diamond tools (PCD) resulted in better tool life as compared to carbide and alumina coated drills and suggested that PCD tool with large nose radii and zero rake angle

144 | 6 Material removal processes for metal matrix composites should be used for roughing operations. Anandakrishnan and Mahamani (2011) studied the machinability of Al6061/(TiB2 ) PMMC fabricated by an in-situ method and reported that the increase in TiB2 reinforcement leads to an increase in tool wear and surface roughness. Also with an increase in the TiB2 reinforcement, the cutting forces are found to decrease due to the presence of fine and uniformly distributed reinforcement (formed in-situ). Further, it was concluded that the feed rate is directly proportional to tool wear, cutting force and surface roughness. Ding et al. (2005) reported that PCD tools performed better than PCBN tools while machining Al/SiC-PMMCs. The better performance of PCD tools was attributed to their higher abrasion and fracture resistance. Kumar et al. (2014) investigated the dry turning characteristics of in-situ Al-4.5 %Cu/TiC-MMCs using uncoated ceramic inserts and reported that cutting force and surface roughness increased while machining at higher feed rate. Also, there was a significant amount of BUE at lower cutting speeds and the formation of chips was influenced by the percentage of reinforcement. N. Muthukrishnan (2011) studied the machinability issues of hybrid MMC (Al/SiC/B4C) using a polycrystalline diamond insert of 1600 grade on a medium duty lathe and reported that higher cutting speeds resulted in a good surface finish and faster tool wear. Manna and Bhattacharayya (2005) recommended the use of high speed, a low feed rate and low depth of cut for achieving better surface finish during the turning of Al/SiC PMMCs.

6.2.2 Milling of PMMCs Milling is the most versatile material removal process and is used for generating a variety of profiles on the workpiece. A multi tooth cutter is used for material removal during the milling process. The tool used in milling operation is a multi-tooth cutter. Arokiadass et al. (2012) developed an empirical relationship to predict the tool flank wear during end milling of LM25Al/SiCP and reported that spindle speed and amount of reinforcement have a greater influence on the tool flank wear, followed by feed rate. Also the depth of cut has low influence on tool flank wear. Reddy et al. (2008) did an experimental study on machinability of Al/SiC-PMMCs during end milling using TiAlN coated carbide end mill cutters. The machinability of SiC reinforced aluminium metal matrix composites was compared with Al alloy on the basis of surface integrity and concluded that proper selection of parameters enhances the machinability of Al/SiC-PMMCs both in terms of surface roughness and tendency to clog the cutting tool. Karakas et al. (2006) investigated the wear behaviour of various tools during milling of Al-4Cu/B4 C composites. The effect of cutting speed on tool wear and the tool wear mechanism was determined at a constant feed rate; it was reported that an increase in the flank wear of the cutting tool occurs at higher cutting speeds. Also the formation of BUE was observed at all cutting speeds, but its size reduced with an increase in cutting speed. Wang et al. (2013) studied the machinability of Al/SiC (65 % volume fraction) during high-speed milling using a PCD cutting tool. The results were

6.2 Conventional machining processes

|

145

compared to the machinability of unreinforced Al 6063 alloy and it was reported that milling speed is the most significant parameter for surface roughness, followed by the feed rate. It was concluded that surface roughness improved slightly with the decrease in the feed rate, and that milling speed had a negligible effect on surface roughness.

6.2.3 Drilling of PMMCs Drilling is an important material removal process and it is used to create holes in a workpiece which is usually used for fastening during assembly or for creating a feature in the part. The drilling operation for PMMCs is similar to the drilling of un-reinforced metals, but the presence of hard particulate reinforcements imposes certain limitations on the tools and processing parameters to be used. The various parameters that effect drilling of PMMCs are cutting speed (rpm), feed (mm/min), composition of the material (type of matrix and reinforcement) and type of tool (tool material and geometry). The output process performance parameters are studied in terms of thrust force, torque, surface finish, chip formation and burr formation. Many researchers have investigated the effect of these input parameters on the drilling behaviour of PMMCs to obtain optimum processing conditions for best output characteristics. Tool wear is considered a major factor that influences the hole quality during the drilling of PMMCs. Tool wear is dependent on the drilling parameters as well as the material properties of the tool and workpiece. Davim and Baptista (2000) investigated the relationship between tool wear and cutting forces using a PCD tool to drill A356/20/SiC-T6. It was reported that the wear of the drill can be monitored by the variation of the cutting forces, using the correlation derived between the cutting forces and flank wear of the drills. Monaghan and Reilly (1992) studied the effect of using coated and uncoated high speed steel (HSS) drills while drilling SiC (25 %) reinforced aluminium matrix composites and found that the presence of coating on an HSS drill results in no significant improvement in the performance of the drill. Davim and Baptista (2001) also studied the output parameters like tool wear, torque and surface roughness while drilling A356/20/SiC-T6-MMC. The drilling was performed using PCD drills, and a superior surface finish exceeding standard values for drilling was reported. Ramulu et al. (2002) used HSS, carbide tipped and PCD drills to evaluate the drilling forces, tool wear, chip formation and drilled hole quality while drilling 10 and 20vol.% (Al2 O3 )P -reinforced Al6061. The results were compared and it was observed that holes made using PCD drills have better drilled hole quality along with minimum drilling induced forces. Mubaraki et al. (1995) concluded that solid carbide and PCD drills have a superior performance because of high abrasion resistance. Coelho et al. (1995) reported that a lower feed rate during drilling of aluminium based PMMCs resulted in rapid tool wear (flank wear), leading to an increase in the cutting forces. Morin et al. (1995) also stated that in the case of drilling MMCs, tool wear is inversely proportional to feed rate. This might be attributed to the decreased probability of interactions between re-

146 | 6 Material removal processes for metal matrix composites inforcements and the cutting tool at a higher feed rate. Davim (2003) did a parametric study of the drilling parameters and concluded that the cutting time and feed rate are the most significant factors affecting tool wear during the drilling of MMCs. Ramulu et al. (2003) concluded that cutting speed has an insignificant effect on tool wear or on drilling forces while drilling Al/(Al2 O3 )P using PCD drills. Sarbjit et al. 2012 investigated the effect of cutting speed, feed and drill point geometry on thrust force, surface roughness and specific cutting pressure. It has been reported that the drill point geometry along with different processing parameters significantly affect the output responses. Tosun Gul (2011) performed a statistical analysis for surface roughness while drilling Al/SiC-PMMCs and reported that the feed rate and tool type were the most significant factors among spindle speed, drill type, feed rate, point angle of drill and heat treatment. Donnini et al. (2011) reported that there is a significant effect of working temperature (20 ∘ C to 160 ∘ C) on drilling forces and surface roughness during hot drilling of Al2009/SiC, Al6061/SiC and Al6061/Al2 O3 . It was concluded that drilling at higher temperatures results in a better surface finish and decreased cutting forces. Basavarajappa et al. (2007) concluded that the surface roughness increases with the increase in feed rate but decreases with a decrease in the cutting edge radius. The reduction in the cutting edge radius might occur due to tool wear. At a constant cutting speed, the surface roughness of the drilled holes increases with the increasing feed rate, but there is no significant change when cutting speed is altered. The lowest surface roughness was reported at the lowest feed rate and highest cutting speed. This decrease in the surface roughness can be attributed to the burnishing and honing effect produced by the trapped abrasive particles between the tool and workpiece. The increase in the feed rate leads to the formation of severe BUE on the tip of the tool; this results in higher thrust force and higher surface roughness.

6.3 Unconventional machining of MMCs The industrial scenario keeps on growing in potential for accepting various challenges in different fields. The development of advanced materials also poses a challenge to the industrial sector to process them in accordance with concerns regarding quality and economy. Although there are many conventional processes as explained in the previous sections that can process advanced materials (metal matrix composites), a need is arising to develop advanced machining processes as engineers wish to make things cheaper, quicker and better than before. Unconventional machining processes are the processes that make use of unconventional means to process the material without making any direct contact between the workpiece and the tool. The material removal takes place in a media of liquid or gas. The MMCs are difficult to machine by conventional processes because of their superior mechanical properties and the limitations of the processes. The unconventional processes accept the challenge and are able to process the material to the required shape. Table 6.1 differentiates the con-

6.3 Unconventional machining of MMCs |

147

Table 6.1. General comparison between conventional and unconventional processes. Comparison

Conventional processes

Unconventional processes

Material removal rate Tool wear rate Surface roughness Material property dependence Wastage of material Machining time Accuracy Burr formation

High High Less More More Less More Yes

Low Low More Less Less More Less No

ventional and unconventional processes on some important issues. Material removal rate is an important factor which is considered among various machining processes as it is related to productivity. Conventional machining processes lead to higher material removal rates, but this benefit is negated by the higher tool wear and difficulty to form complex shapes. The formation of complex profiles with conventional processes requires a number of tool settings which increases the machining time; in contrast, unconventional processes can do it with a single set of tools and in a single pass. The classification of unconventional processes is shown in Figure 6.1 based on the mode of energy they use for processing the materials. Based on the research publications, among all unconventional machining processes, electrochemical machining (ECM), electric discharge machining (EDM) and ultrasonic machining (USM) are found to be the most effective methods for the machining of MMCs.

Unconventional machining processes

Mechanical processes

Thermal processes

Chemical/electrochemical processes

Ultrasonic machining (USM)

Electric discharge machining (EDM)

Electrochemical machining (ECM)

Water jet machining (WJM)

Electron beam machining (EBM)

Electrochemical grinding (ECG)

Abrasive jet machining (AJM)

Laser beam machining (LBM)

Chemical milling (CHM)

Plasma arc machining (PAM) Fig. 6.1. Classification of unconventional machining processes.

148 | 6 Material removal processes for metal matrix composites 6.3.1 Electrochemical machining of PMMCs The electrochemical machining works on the principle of electrolysis discovered by Michael Faraday (1791–1867). In electrolysis, when two conductive electrodes are placed in a media of conducting fluid (electrolyte) and a DC potential is applied, then metal from one electrode (anode) tends to deposit on the other electrode (cathode). This process is known as electroplating and has been used in industries for many years. Lynn A. Williams invented the present day ECM process based on Faraday’s law of electrolysis which stated that by making the workpiece an anode, the material can be removed from the workpiece and prevented from depositing on the tool (cathode) by the flow of electrolytes. Material removal through ECM requires a high current and low voltage power supply between the tool and the workpiece. The workpiece is connected to the positive terminal and the tool is connected to the negative terminal of the power supply. A proper gap is maintained between tool and the workpiece for the flow of electrically conductive electrolytes. Figure 6.2 shows the mechanism of material removal in ECM of MMC. The presence of the electrolyte with proper values of current and voltage causes the transfer of electrons from the workpiece towards the tool.

Tool ‒ +

Filter

Pump Electrolyte

DC power supply Insulation

MMC

Reinforcement

Fig. 6.2. Material removal mechanism of ECM.

The high velocity electrolyte takes away the removed workpiece particles and does not let them deposit on the tool. The electrolyte is continuously filtered and recirculated. Due to the ionic dissolution of the matrix material, the reinforcement particles are exposed and flow with the electrolyte. ECM produces a reverse image of the tool on to the workpiece; this makes it very applicable to the production of complex profiles. ECM removes the material regardless of the hardness of the workpiece; it can thus be used for the machining of advanced materials such as metal matrix composites that are difficult to machine by conventional methods.

6.3 Unconventional machining of MMCs | 149

Advantages: – no tool wear; – no mechanical and thermal stresses; – high surface finish; – complex shapes can be produced. Limitations: – only conductive materials can be machined; – sharp square corners are difficult to produce; – high equipment cost.

6.3.1.1 Input process parameters and output responses There are a number of input parameters as shown in Table 6.2 that affect the output responses during electrochemical machining of MMCs as indicated by Rao (2013). The output responses are MRR, SR and radial overcut (ROC). Table 6.2. Various input parameters in ECM. Parameters

Unit

Range

Current Voltage Feed rate Inter-electrode gap Electrolyte velocity Electrolyte pressure Electrolyte temperature Electrolyte concentration

Ampere (A) Volts (V) mm/min mm m/s MPa Celsius (∘ C) kg/lit

50–40,000 4–30 0.5–19 0.025–0.75 15–60 0.069–2.7 24–65 Up to 0.5

Many researchers have worked with different input parameters to optimize the process and to find out their effects on output responses. The key parameters found based on the literature are voltage, feed rate, electrolyte flow and electrolyte concentration. A low value of applied voltage causes non-uniform anodic dissolution and results in low surface finish. Increasing the voltage to a higher value causes a higher machining current within the inter-electrode gap which results in higher MRR and higher radial over cut (ROC). A high value of feed rate decreases the inter-electrode gap, and thereby increases the current density. This increased current density speeds up the anodic dissolution; this increases the MRR and decreases surface roughness. The formation of reaction products (i.e. sludge, hydrogen bubbles, etc.) decreases the ionic strength within the inter-electrode gap and provides a higher flow rate which helps to increase the MRR as stated by Senthikumar et al. (2009). In ECM, the most com-

150 | 6 Material removal processes for metal matrix composites monly used electrolyte is sodium chloride (NaCl) because of the low cost and stable conductivity. Rao and Padmanabhan (2013) reported that by increasing the concentration of the electrolyte, electrolyte conductivity is increased which in turn increases current density and thus results in an increased MRR. If the concentration level is lower, then the surface roughness is higher and vice versa. The reinforcement materials added in the matrix material may be conductors, semiconductors or insulators. The higher amount of semiconductors and insulators used as the reinforcement materials reduces the electrical conductivity of the MMC; this results in a decreased MRR. Hihara and Panquites (2002) reported that an increase of electrode potential above 1.75 V during electrochemical machining of Al/SiC-MMC provides for a faster dissolution of matrix material into the electrolyte. The removal of aluminium matrix material is caused by ionic dissolution while the SiC particles are taken away by the electrolyte flow. Liua et al. (2010) introduced an electrochemical mechanical machining (ECMM) process for the machining of particulate reinforced metal matrix in which matrix material is removed by the ECM action while the non-conducting reinforcement particles are removed by mechanical grinding. Goswami et al. (2009) conducted electrochemical grinding on Al/Al2 O3 -MMC and studied the effects of supply voltage, depth of cut, electrolyte flow rate and electrolyte concentration. The material removal mechanism and surface characteristics were explored with the help of scanning electron micrographs (SEM). Different grinding electrodes were used and their effects on the grinding force have been studied.

6.3.2 Electric discharge machining of PMMCs Electric discharge machining is the most widely used unconventional process for the machining of materials that are difficult to machine such as MMCs because it removes the metal irrespective of the hardness. EDM is a thermal process which converts electrical energy into thermal energy through a series of electric sparks across a small gap between two electrodes (anode and cathode). This thermal energy makes the material melt and evaporate, which results in the erosion of metal in the form of tiny particles. The EDM system consists of an anode and cathode that are immersed in a non-conducting dielectric fluid. The dielectric acts as an insulator between the tool and the workpiece until the potential is sufficiently high. This high potential (due to high voltage) breaks the dielectric strength and creates an ionised channel within the dielectric (Figure 6.3). The current finds a path through this channel, forming a spark, which increases the temperature of the zone in a range from 8,000 ∘ C to 12,000 ∘ C as reported by Boothroyd and Winston (1989). The localized melting and evaporation of the material takes place and leaves a crater behind. The amount of material removed per spark depends upon the spark duration and energy of the spark. When the current supply is turned off, the plasma channel disappears and the dielectric regains its strength, thereby insulating the gap.

6.3 Unconventional machining of MMCs | 151

Servo head Ionized channel

Filter Pump MMC

Fig. 6.3. Material removal mechanism of EDM.

Advantages: – applicable to hard-to-machine materials; – tool need not to be harder than the work material; – complex shapes can be produced easily. Limitations: – power consumption is higher than in conventional processes; – workpiece must be electrically conductive; – tool wear rate is high.

6.3.2.1 Input process parameters and output responses The input parameters used in electric discharge machining are discharge current, discharge frequency, voltage, pulse on-time, pulse off-time, electrode gap and flushing pressure. The output responses are MRR, TWR and surface roughness (SR); these are all significantly affected by the amount of heat energy input to the workpiece. The heat energy is dependent on the amount of discharge current and the duration of sparks. Dvivedi et al. (2008) concluded that MRR and TWR increases when increasing both discharge current and pulse on-time; the higher value of these causes more material melting and evaporation. A higher value of discharge current and pulse on-time produces larger craters; this causes a decrease in the workpiece’s surface finish. Surface finish can be improved by decreasing the current in the spark and increasing the frequency of the spark (500–500,000 Hz) which gives a small crater size by maintaining the MRR at the same level. Sufficient pulse off-time is required for the removal of debris (eroded particles) from the machining zone. These eroded particles – if not removed – rush into the electrode gap and make the sparks unstable, lowering the MRR. El-Hofy (2005) stated that there is a chance of short circuit due to the presence

152 | 6 Material removal processes for metal matrix composites of a large amount of eroded metal particles in the electrode gap. Hung et al. (1994) reported that the increased percentage of reinforcement particles decreases the spark area because of the non-conducting nature of the reinforcement. No direct heat from the spark is given to the reinforcement particles, therefore the particles come out from the matrix un-melted. Dielectric flushing is provided for the removal of debris and has a significant effect on the MRR. The MRR is lower with low and high values of flushing pressure. A lower flushing pressure is incapable of removing debris and a much higher pressure disturbs the sparks. Therefore an optimal pressure value is to be chosen for an increased MRR. Singh et al. (2013) investigated the effect of tool rotation on MRR, TWR and SR. It has been found that rotating tool electrodes gives a higher MRR, lower TWR and lower SR. The rotary motion of the electrode creates a centrifugal force in the fluid which helps in easy removal of eroded particles from the machining zone. Kumar and Davim (2011) developed a set-up for powder mixed electric discharge machining (PMEDM) in which the dielectric fluid is suspended with the silicon powder. The experimental result confirms that by mixing the silicon powder into the dielectric fluid, EDM performance is increased in view of the increased MRR and decreased SR. Karthikeyan et al. (1999) experimented on an aluminium based MMC with input parameters such as discharge current, pulse duration and volume percentage of SiC with three-level full factorial design and reported the decrease in the MRR while the TWR and SR increased with an increase in the SiC percentage. Ramulu et al. (2001) carried out a comparison of surface conditions between a polished Al/SiC-MMC and EDMed Al/SiC-MMC. Tensile tests and fatigue tests were also conducted and it was found that the fatigue strength of the EDMed specimen is lower than the unprocessed specimen. Mohan et al. (2004) investigated the effects of the SiC percentage, pulse duration and tool rotation on the MRR, TWR and SR. The increase in tool rotational speed and the SiC percentage increases the MRR but decreases the TWR and SR, while the effect of pulse duration is the opposite. Dvivedi et al. (2008) developed a Al6063/SiCp metal matrix composite using a melt stir-squeeze-quench casting method and investigated the effects of process variables (discharge current, pulse on-time, flushing pressure and electrode gap) on output responses (MRR, TWR and SR). The process was optimized in view of the increased MRR and decreased TWR and SR. Nanimina et al. (2011) carried out electric discharge machining of a Aluminium-MMC with 30 % Al2 O3 as a reinforcement and observed an increase in the MRR at low peak currents and pulse on-times. Rathod et al. (2013) developed the Al/SiC-MMCs with the addition of 3, 5 and 7 wt.% of SiC particles and carried out an EDM with the use of circular and square tool electrodes. It was concluded that a better surface finish of the machined workpiece can be achieved with square tool electrodes than with circular electrodes. It was also reported that circular tool electrodes with smaller diameters provide a higher surface finish.

6.3 Unconventional machining of MMCs |

153

6.3.3 Ultrasonic machining of PMMCs Ultrasonic machining (USM) was first developed by Balamuth in 1945 and the industrial applications started in the 1950s. Unlike ECM and EDM, USM is applicable for the machining of conductive as well as non-conductive materials. USM is viewed as an effective process for machining MMCs because it is capable of machining hard and brittle materials while neglecting their electrical and chemical characteristics. This is a mechanical energy-based process that removes the material by the impact of abrasive particles applied between a vibrating tool and the workpiece. The abrasive particles together with the containing media (e.g. water) forms slurry and is injected into the machining zone through a nozzle. The abrasive particles most often used in USM are silicon carbide (SiC), alumina (Al2 O3 ) or boron carbide (B4 C). The oscillating tool with an ultrasonic frequency 15–30 KHz and amplitude 10–40 μm applies a static force on the abrasive particles; this causes abrasion of the workpiece as stated by Pandey and Shan (1980). MMCs have abrasive particles as reinforcements (SiC, Al2 O3 ,B4 C, etc.) embedded into them that also take part in the abrasion process after coming out of the matrix during machining (Figure 6.4). Slurry pressure takes away the removed workpiece particles together with the reinforcements from the machining area. The material-removal mechanisms in USM as found by Shaw (1956) are: (1) the hammering of abrasive particles on the work surface by the tool; (2) the impact of free abrasive particles on the work surface; (3) cavitation erosion; and (4) the chemical action associated with the fluid employed. The material removal progresses in the direction of vibration; a mirror image of the tool is formed into the workpiece.

Tool

Debris Abrasive Nozzle

Slurry

MMC

Fig. 6.4. Material removal mechanism of USM.

Reinforcement

154 | 6 Material removal processes for metal matrix composites Advantages: – removes any material regardless of electrical conductivity; – provides good surface finish; – sharp edges can be obtained. Limitations – used for small depths only; – low material removal rate; – high tool wear.

6.3.3.1 Input process parameters and output responses Researchers have found that the key parameters such as the amplitude of tool vibration, slurry concentration and abrasive grit size affect the MRR, TWR and SR during ultrasonic machining of MMCs. Increase in the amplitude of tool vibration increases the hammering effect of the tool on the workpiece; this increases the rate of material removal. The increased tool vibrations also facilitate an easy flushing of removed particles from the machining zone. The MRR is also increased with the amount of static force (0.1–30 N) applied on the tool. The combined effect of static force and tool vibration leads to the increase in TWR. A low value of static force is required in the machining of micro-channels/holes, because the higher static force may cause the tool to bend. The abrasives used in the slurry should be harder than the work material, otherwise their sharp edges will not work for long, thereby reducing the MRR. TWR is also increased with the increase of the MRR, but the tool wear can be controlled by selecting a tool of ductile material. Gilmore (1995) observed that hard tools experience more wear than ductile tools. The grit size of the abrasive particles used in the USM ranges from 240 to 800. The coarser abrasive particles provide a higher MRR but lower the surface finish; therefore, they can only be used for roughing operations. Jiao et. al (2006) concluded that chip formation in the conventional machining of MMCs only occurs because of the compression deformation between the tool and the workpiece, while in the USM it is because of the impact action of the tool at high frequency. Ultrasonic machining of MMCs provides long spiral chips with small plastic deformation. Researchers also have conducted the experiments by providing rotary motion to the vibrating tool, and they reported its positive effect on the MRR. Pie et al. (1995) reported that rotary ultrasonic machining provides a 10-times faster machining than stationary tools. A drop in static forces and surface roughness values are also reported by using a rotary USM. Wang et al. (2013) conducted experiments on Al/SiC-MMC reinforced with 45 vol.% SiC on rotary ultrasonic machining (RUM) and the surface roughness, surface topography and cutting force were analyzed. The cutting tool did not show any sign of grinding burns and grinding wheel jams. Zhao et al. (2002) compared the results of common cutting (CC) with the ultrasonic vibration cutting of Al/SiC-MMC and re-

6.4 Conclusion

|

155

ported that ultrasonic cutting gives less deformed long spiral chips with a larger curl radius, while conventional cutting gives short spiral chips with a smaller curl radius.

6.4 Conclusion The development of MMCs is being carried out to take advantage of their superior physical and mechanical properties, especially in industrial sectors such as automotive and aerospace. The selection of matrix and reinforcement material effects the various properties of the developed MMCs such as specific strength and stiffness. In some applications, particularly in the aerospace industry, a lightweight MMC with superior mechanical properties is selected in place of a monolithic metal/alloy without considering the cost incurred in the processing of the MMC. Fuel efficiency and passenger safety are the prime concerns. With an increase in the application areas of MMCs, a need to develop the machining processes exists in order to achieve the cost-effective machining of MMCs. The present chapter highlights the conventional and unconventional material removal processes used for the processing of particulate reinforced metal matrix composites. The factors which are considered during the selection of a particular process (i.e. conventional or unconventional) are cutting forces, material removal rate, tool wear rate, surface finish, job profile, machining cost, machining time, etc. Both the conventional and the advanced machining processes have been used for the machining of MMCs that are difficult to machine and require complex profile. Based on the studies, it can be said that a number of research initiatives have been noted in the area of conceptualizing and developing unconventional/hybrid machining processes for the cost-effective machining of MMCs.

References [1] [2]

El-Hofy H. Advanced Machining Processes. New York: McGraw-Hill; 2005. Manna A, Bhattacharayya B. Influence of machining parameters on the machinability of particulate reinforced Al/SiC-MMC. Int J Adv Manuf Technol 2005;25:850–6. [3] Yanming Q, Zehua Z. Tool wear and its mechanism for cutting SiC particle-reinforced aluminium matrix composites. J Mater Process Technol 2000;100:194–9. [4] El-Gallab M, Sklad M. Machining of Al/SiC particulate metal-matrix composites Part I: Tool performance. J Mater Process Technol 1998;83:151–8. [5] Anandakrishnan V, Mahamani A. Investigations of flank wear, cutting force, and surface roughness in the machining of Al-6061-TiB2 in situ metal matrix composites produced by flux-assisted synthesis. Int J Adv Manuf Technol 2011;55:65–73. [6] Ding X, Liew WYH, Liu XD. Evaluation of machining performance of MMC with PCBN and PCD tools. Wear 2005;259:1225–34. [7] Kumar A, Mahapatra MM, Jha PK. Effect of machining parameters on cutting force and surface roughness of in situ Al-4.5%Cu/TiC metal matrix composites. Measurement 2014;48:325–32.

156 | 6 Material removal processes for metal matrix composites [8] Muthukrishnan N, Babu TSM, Ramanujam R. Fabrication and turning of Al/SiC/B4 C hybrid metal matrix composites optimization using desirability analysis. J Chinese Inst Ind Eng 2012;29:515–25. [9] Arokiadass R, Palaniradja K, Alagumoorthi N. Prediction and optimization of end milling process parameters of cast aluminium based MMC. Trans Nonferrous Met Soc China 2012;22:1568–74. [10] Suresh Kumar Reddy N, Kwang-Sup S, Yang M. Experimental study of surface integrity during end milling of Al/SiC particulate metal-matrix composites. J Mater Process Technol 2008;201:574–9. [11] Karakas MS, Acir A, Ubeyli M, Ogel B. Effect of cutting speed on tool performance in milling of B4 CP reinforced aluminum metal matrix composites. J Mater Process Technol 2006;178:241–6. [12] Wang T, Xie LJ, Wang XB, et al. Surface Integrity of High Speed Milling of Al/SiC/65p Aluminum Matrix Composites. Procedia CIRP 2013;8:475–80. [13] Davim JP, Baptista AM. Relationship between cutting force and PCD cutting tool wear in machining silicon carbide reinforced aluminium. J Mater Process Technol 2000;103:417–23. [14] Monaghan J, Reilly PO. The drilling of an Al/SiC metal-matrix composite. J Mater Process Technol 1992;33:469–80. [15] Davim JP, Baptista A Monteiro. Cutting force, tool wear and surface finish in drilling metal matrix composites. Proc Inst Mech Eng Part E J Process Mech Eng 2001;215:177–83. [16] Ramulu M, Rao PN, Kao H. Drilling of (Al2 O3 )P /6061 metal matrix composites. J Mater Process Technol 2002;124:244–54. [17] Mubaraki B, Bandyopadhyay S, Fowle R, Mathew P, Heath P J. Drilling studies of an Al2O3-Al metal matrix composite Part l Drill Wear Characteristics. J Mater Sci 1995;30:6273–80. [18] Coelho RT, Yamada S, Aspinwall DK, Wise MLH. The application of polycrystalline diamond (PCD) tool materials when drilling and reaming aluminium based alloys including MMC. Int J Mach Tools Manuf 1995;35:761–74. [19] Morin E, Masounave J, Laufer EE. Effect of drill wear on cutting forces in the drilling of metal-matrix composites. Wear 1995;184:11–6. [20] Davim JP. Study of drilling metal-matrix composites based on the Taguchi techniques. J Mater Process Technol 2003;132:250–4. [21] Ramulu M, Rao PN, Kao H. Experimental Study of PCD Tool performance in drilling Al2O3p/6061 metal matrix composites. In: Transactions of the North American Manufacturing Research Institution of SME 2003;31:169–75. [22] Singh S, Singh A, Singh I, Dvivedi A. Optimization of the process parameters for drilling of metal matrix composites (MMC) Using Taguchi analysis. Adv Mater Res 2012;410:249–52. [23] Tosun G. Statistical analysis of process parameters in drilling of AL/SICP metal matrix composite. Int J Adv Manuf Technol 2011;55:477–85. [24] Donnini R, Santo L, Tagliaferri V. Hot Drilling of Aluminium Matrix Composite. Mater Sci Forum 2011;678:95–104. [25] Basavarajappa S, Chandramohan G, Prabu M, Mukund K, Ashwin M. Drilling of hybrid metal matrix composites—Workpiece surface integrity. Int J Mach Tools Manuf 2007;47:92–6. [26] Rao PN. Manufacturing Technology: Metal Cutting and Machine Tools, 3e (Volume 2). New Delhi: McGraw-Hill;2013. [27] Senthilkumar C, Ganesan G, Karthikeyan R. Study of electrochemical machining characteristics of Al/SiCp composites. Int J Adv Manuf Technol 2009;43:256–63. [28] Rao SR, Padmanabhan G. Optimization of Machining Parameters in ECM of Al/B4 C Composites. J. Manuf. Sci. Prod. 2013;13:145–53. [29] Hihara LH, Panquites IV P. The potential of electro-chemical machining for silicon carbide/aluminum metal-matrix composites. Grind Abrasives 2002;11–5.

References

|

157

[30] Liu GY, Guo ZN, Li YB, Huang ZG. An Investigation on PRMMC Polishing with Electrochemical Mechanical Machining Process. Adv Mater Res 2010;135:226–31. [31] Goswami RN, Mitra S, Sarkar S. Experimental investigation on electrochemical grinding (ECG) of alumina-aluminum interpenetrating phase composite. Int J Adv Manuf Technol 2009;40:729–41. [32] Boothroyd G, Winston AK. Non-conventional machining processes. in: Dekker M, Fundamals of Machining and Machine Tools. New York: Marcel Dekker Inc, 1989:491. [33] Dvivedi A, Kumar P, Singh I. Experimental investigation and optimization in EDM of Al 6063 SiCp metal matrix composite. Int J Adv Manuf Technol 2008;3:293–308. [34] Hung NP, Yang LJ, Leong KW. Electrical discharge machining of cast metal matrix composites. J Mater Process Technol 1994;44:229–36. [35] Singh A, Kumar P, Singh I. Electric discharge drilling of metal matrix composites with different tool geometries. Proc Inst Mech Eng Part B J Eng Manuf 2013;227:1245–9. [36] Kumar H, Davim JP. Role of powder in the machining of Al-10%SiCp metal matrix composites by powder mixed electric discharge machining. J Compos Mater 2011;45:133–51. [37] Karthikeyan R, Lakshmi Narayanan PR, Naagarazan RS. Mathematical modeling for electric discharge machining of aluminum-silicon carbide particulate composites. J Mater Process Technol 1999;87:59–63. [38] Ramulu M, Paul G, Patel J. EDM surface effects on the fatigue strength of a 15 vol.% SiCp/Al metal matrix composite material. Compos Struct 2001;54:79–86. [39] Mohan B, Rajadurai A, Satyanarayana KG. Electric discharge machining of Al-SiC metal matrix composites using rotary tube electrode. J Mater Process Technol 2004;153-4:978–85. [40] Nanimina AM, Abdul-Rani AM, Ahmad F, Zainuddin A, Jason Lo SH. Effects of Electro-discharge Machining on Aluminum Metal Matrix Composite. J Appl Sci 2011;11:1668–72. [41] Rathod S, Sekhar R, Jatti VKS, Singh TP. Optimization of tool shape and size in EDM of Al alloy metal matrix composites. Int J Appl Eng Res 2013;8:639–46. [42] Pandey PC, Shan HS. Modern Machining Processes. New Delhi: Tata McGraw-Hill; 1980. [43] Shaw MC. Ultrasonic grinding. Microtechnic 1956;10:257–65. [44] Gilmore R, Ultrasonic machining and polishing. In: Proceedings of international symposium for electromachining. Presses Polytechniques et Universitaires Romandes, Switzerland 1995:941–51. [45] Jiao F, Zhao B, Liu C, Zhu X. Chip formation mechanism and morphological characteristics in ultrasonic cutting of SICp/AL-MMC. In: Progress of Machining Technology – Proceedings of the 8th International Conference on Progress of Machining Technology. Japan Society for Precision Engineering, Japan 2006:141–4. [46] Pei ZJ, Khanna N, Ferreira PM. Rotary Ultrasonic Machining of Structural Ceramics: A Review. Ceram Eng Sci Proc 1995;16:259–78. [47] Wang M, Dong GJ, Zhou M. Experimental Study on Rotary Ultrasonic Machining of SiCp/Al Composite. Key Eng Mater 2013;589-90:529–33. [48] Zhao B, Liu CS, Zhu XS, Xu KW. Research on the vibration cutting performance of particle reinforced metallic matrix composites SiCp/Al. J Mater Process Technol 2002;129:380–4.

Vijayan Krishnaraj

7 An investigation into machining Al/SiC metal matrix composites Abstract: In recent years, the utilization of metal matrix composites (MMC) has increased in the automotive, aviation, aerospace, construction and microelectronics engineering fields. In conjunction with innovations in these advanced materials, machining them to obtain better dimensional accuracy and surface integrity has become a challenge. In this present work, 15 % (by weight) SiC particle reinforced aluminium is synthesized by a stir casting technique. The synthesized MMC is end milled using Ø16 mm carbide end mill in a CNC milling machine; the machining parameters – cutting speed, feed rate and depth of cut – are optimized for minimum surface roughness and force. Drilling a metal is an important operation in the final fabrication stage prior to application. This paper discusses the influence of cutting parameters on drilling characteristics of metal matrix composite (Al6061-T6/20SiCp). The experiments were conducted to study the effect of spindle speed and feed rate on thrust force and torque using solid carbide drills of 5 mm diameter. The results reveal that the dependent variables are greatly influenced by the feed rate rather than the spindle speed for the composites.

7.1 Milling of metal matrix composites 7.1.1 Introduction Because of their increased specific strength and stiffness, metal matrix composites are replacing the conventionally used materials especially in the automotive, aerospace and structural engineering fields. Due to their low cost, ceramic particles reinforced aluminium alloys are the most popular among the MMC. The ceramic particles or reinforcements such as SiC and alumina make machining the composite more difficult. This difficulty in machining opens a wide area of research in the processing of these materials. Ozben et al. [1] investigated the effects of the reinforcement fraction on the mechanical properties such as hardness and tensile strength of the Al/SiCp composite. They examined the effects of the machining parameters and reinforcement fraction on tool wear and surface roughness during turning of the Al/SiCp composite using a TiN coated carbide tool. The machinability of Al/SiC composite while turning with the use of rhombic tools was reported by Manna et al. [2]. They examined BUE and chip formation using the SEM micrographs to provide an economic machining solution through their work. Davim [3] reported the influence of cutting parameters during

160 | 7 An investigation into machining Al/SiC metal matrix composites turning Al/SiCp on surface roughness, tool wear and power requirements by employing ANOVA and multiple linear regression techniques. Quan et al. [4] indicated that the merchant’s circle equation cannot be used for the composite materials, because the yield strength of these materials is completely different and also the assumptions made during merchant circle derivation restricts their application to composite materials. They also related the coefficient of chip deformation to the fraction of reinforcement and shear angle. Liu et al. [5] used PCD turning tools to compare the cutting forces obtained from conventional turning and ultrasonic vibration turning, thereby concluding that the low speed and high depth of cuts reduces cutting forces in ultrasonic vibration turning. They developed process parameters for turning thin walled Al/SiCp -MMC. Muthukrishnan and Davim [6] turned Al/SiCp bars using coarse grade PCD inserts under different cutting conditions. The machining conditions were then optimized for minimized surface roughness by employing ANOVA and ANN. Seeman et al. [7] developed a mathematical model for the machinability evaluation in turning Al/SiCp composites. They also described the effects of process parameters on the tool flank wear and surface roughness. Ge et al. [8] used PCD tools for ultra-precision turning of Al/SiCp -MMC and revealed the effects of the cutting speed and feed rate on the tool workpiece surface integrity. Reddy et al. [9] end milled Al/SiCp -MMC using TiAlN coated carbide inserts and optimized the cutting parameters for minimized surface roughness using GA. They also found out the micro-hardness and residual stress at the optimized cutting conditions. Karakas et al. [10] studied the effects of the cutting speed on the tool performance in end milling of B4 Cp particles-reinforced aluminum MMC. They also compared the wear performance of the uncoated tool and multi-layer coated tools and found that the triple layer coated tool has better wear resistance. Oktem et al. [11] applied response surface methodology (RSM) for the optimization of machining parameters for minimized surface roughness during the milling of Al-7075 alloy parts. They also used the genetic algorithm (GA) for optimizing the parameters for desired surface roughness. Lin et al. [5] studied the chip formation in turning the Al/SiCp -MMC using PCD inserts. They analyzed the SEM micrographs of the chips formed and found a separation of the matrix and reinforcement within the chip. Joshi et al. [12] analyzed the chip formation mechanism in turning the Al/SiCp -MMC. They tried to correlate the quality of the machined surface by analyzing the SEM micrographs of the chips formed during turning. Arokiadass et al. [13] made an attempt at developing a predictive model of surface roughness in end milling of Al/SiCp -MMC. They also used RSM for necessary analysis; this resulted in the optimization of machining parameters. The present work reports the effect of machining parameters on cutting force, surface finish and behavior of chips during end milling and drilling.

7.1 Milling of metal matrix composites

|

161

7.1.2 Experimental procedure 7.1.2.1 Material synthesizing The 6061 Al alloy of the following composition (by weight) is chosen as the metal matrix of the composite and the reinforcement is 15 % (by weight) SiC powders of size 36 μm. Tables 7.1 and 7.2 present the composition of matrix and reinforcement used for the metal matrix composites. Table 7.1. Composition of Al-6061 alloy by weight. Al 97.25 %

Mg

Fe

Si

Mn

Others

1.08 %

0.17 %

0.63 %

0.52 %

0.35 %

Table 7.2. Composition of SiC powders by weight. SiC 99.64 %

SiO2

Si

Fe

Al

Free C

0.15 %

0.02 %

0.02 %

0.02 %

0.15 %

The Al/SiCp-MMC with the above composition is synthesized by the stir casting process. Figure 7.1 shows a schematic sketch of the stir casting setup, describing its parts. The temperature of operations involved in the process, time involved in stirring and the amount of metal and reinforcement used in casting are listed in the Table 7.3. The micro-hardness of the MMC is found to be as 106 VHN using tested Vickers micro-hardness tester at a load of 100 g. Using an optical microscope, the microstructure of synthesized Al/SiCp -MMC is found to be as shown in Figure 7.2. The distribution of the SiC particles and the agglomerations are presented in the micrograph.

Stirrer

Melting furnace

Fig. 7.1. Stir casting setup.

Pre heating furnace

162 | 7 An investigation into machining Al/SiC metal matrix composites Table 7.3. Experimental details of stir casting. Furnace

Induction furnace

Pre-heater Mass of Al 6061 alloy Mass of SiC reinforcement Mass of Mg (wetting agent) Melting point of Al 6061 Pre-heat temperature of SiC reinforcement SiC addition temperature Stirrer speed Time of SiC addition Post stirring time Mold preparation

Electric furnace 900 g 135 g 15 g 700 ∘ C 450 ∘ C 750 ∘ C 850 rpm 120 min 20 min Wooden pattern in green sand

SiC Reinforcement

Agglomeration Al matrix

Fig. 7.2. Microstructure of Al/SiCp MMC at 200×.

7.1.2.2 Experimental setup The end milling experiments were carried out in a three-axes CNC Makino Vertical Machining Centre Model S33. The experimental setup for end milling is as shown in Figure 7.3. The milling tool dynamometer (SYSCON) is mounted on the milling bed, and the rectangular metal matrix composite specimen is held in its fixture. The dynamometer is interfaced with a personal computer (PC) for force measurements.

Spindle End milling cutter Component Milling tool dynamometer

Milling bed

Fig. 7.3. Experimental setup.

Personal computer

7.1 Milling of metal matrix composites

|

163

Milling is a very complicated cutting process that involves many parameters such as cutting speed, feed rate, depth of cut and tool geometry, etc. The most influential factors affecting the surface finish and forces acting on the tool were studied by conducting a set of experiments. The factors considered for the experimentation are cutting speed, feed, and depth of cut. The experimental conditions are presented in Table 7.4. The experiments were conducted under dry conditions in the three-axes vertical machining centre. Table 7.4. Machining parameters. Parameter level −1 0 1

Speed (m/min)

Feed rate (mm/rev)

Depth of cut (mm)

150 200 250

0.05 0.20 0.35

0.10 0.15 0.20

The tool used for end milling is a Ø16 mm inserted cutter made by Sandvik Coromant. The specification of the insert used is R310 11 T3 08E NL-13A. One set of inserts is coated using nano-composite structured Hyperlox coating by the physical vapour deposition (PVD) magnetron sputtering technique. The coating thickness is 3 μm in order to compare it to an uncoated insert. Using a central composite design (CCD), 16 experiments are carried out in total. For each run, the forces acting on the tool are measured with the tool dynamometer. The surface roughness is measured by a Taylor Hobson surface roughness tester with a cut-off length of 2.5 mm. The surface roughness is measured at three different places in a machined surface; an average of these values is taken. The machining parameters are optimized using RSM. For the optimized machining condition, the effect of hyperlox coated inserts on the surface roughness and tool wear are analyzed by comparing the results of the experiments conducted using coated and uncoated inserts under dry conditions.

7.1.3 Results and discussion Milling can generally be classified as rough milling and finish milling. In the case of rough milling, the major factor to be considered is the material removal rate, whereas in finish milling it is surface roughness. When taking tool life into account, the tool wear caused by forces acting on the tool is also a major factor. An improvement in surface finish and tool life can be accomplished by optimizing the machining conditions or by changing the tool conditions. Here, experiments are conducted for both conditions and results are discussed below.

164 | 7 An investigation into machining Al/SiC metal matrix composites Thrust force Infeed force Crossfeed force

(a)

(b)

Fig. 7.4. (a) Forces acting on the tool. (b) Surface roughness tester.

The response surface model (RSM) , which is an analytical function to predict surface roughness and tool force values, is developed using RSM. RSM uses the statistical design of the experiment (experimental design) technique and the least-square fitting method during the model generation phase. The initial stage in creating an RS model is to design experiments using one of the RSM designs. Here, central composite design (CCD) is used for experimental design. A face-centered experimental design with three factors (speed, feed rate and depth of cut) and 20 runs is generated using CCD in the MINITAB software. Among the 20 runs, 4 center-point runs are ignored and 16 runs of experiments are conducted in total. The tool forces are measured using a Syscon milling tool dynamometer. The forces acting on the tool are shown in Figure 7.4. The tangential and feed forces are considered vital over the thrust force. The total force acting on the tool is then calculated as the vector sum of the tangential and feed force. The surface roughness (Ra) is measured using the Taylor Hobson surface roughness tester (shown in Figure 7.4b) with a cut-off length of 2.5 mm. At each slot, the surface roughness is calculated at three different spots and their average is taken as the average surface roughness. Table 7.5 shows the list of experiments conducted and the tool force and surface roughness obtained at each run. Table 7.5. Response Surface Optimizer result Speed (m/min) 250

Feed rate (mm/rev)

Depth of cut (mm)

Tool force (N)

Surface roughness (μm)

0.05

0.20

8.9291

1.2639

The obtained results are analyzed using the MINITAB software. The speed, feed rate and depth of cut are chosen as the independent variables or factors, and the surface

7.1 Milling of metal matrix composites

|

165

roughness and tool force are chosen as the dependent variables or responses. The response surface regression equation for cutting force (F) is given in equation (7.1): F = −150.253 + 3.662V + 325.23f − 2762.29d − 0.00704V 2 − 1291.422 + 11370.6d2 + 1.486Vf − 5.046Vd + 203.382fd

(R2 = 0.908).

(7.1)

Response surface regression equation for surface roughness (Ra ) is given in equation (7.2): Ra = −1.668 + 0.0144V − 5.303f + 43.254d − 0.00007V 2 − 5.20307f 2 − 194.828d2 + 0.0158333Vf + 0.0515Vd + 40.1667fd

(R2 = 88.95).

(7.2)

In plotting the above surface plots, a constant value is applied for the third independent variable. In Figure 7.5a, depth of cut is maintained as 0.15 mm, and in Figures 7.5b and 7.5c, feed rate is 0.2 mm/rev and cutting speed 200 m/min respectively. So, by fixing one of the independent variables, the other independent variables can be chosen for the required output using surface plots. This method can be utilized for machining at low tool forces and increased surface finish. Similar graphs (Figure 7.6a–c) were obtained for the effect of machining parameters on surface roughness (Ra). Apart from the above analysis, optimized values for machining parameters are obtained for a minimum tool force and a minimum surface roughness using the response surface optimizer. Those values are given in the Table 7.5.

7.1.3.1 Chip formation mechanism Chips can generally be classified as continuous, discontinuous, continuous with BUE and serrated (homogenous) chips. Due to the presence of abrasive particles in the MMC, most of the time the formed chips are discontinuous. But at higher feed rates, we are able to find saw-toothed chips with primary cracks at the outer free surface and secondary cracks at the inner surface with BUE. A few SEM micrographs will aid in visual compliance of the above. The SEM micrograph in Figure 7.7a shows the top surface of a chip. Cracks are highly formed because the 15 % of SiC particle reinforcement has induced high brittleness in the material; ductility is thus reduced. This results in the formation of discontinuous chips. The chip formation mechanism can be now revealed as an initiation of cracks from the outer free surface due to shear stress induced by tool rake. And during machining at shear zones, the separation of the particle and matrix results in voids caused by the stress concentration at the edge of particles. These 2 stress components, along with the help of voids, results in crack propagation and discontinuous chip formation. The sawtooth profile is formed due to a highly strained inner surface where the reinforcement is coarsely distributed. Figure 7.7b shows the views of the sheared (bottom) surface of a chip. It shows a completely propagated chip that is about to break. One can also see that aluminium is

Total force (N)

166 | 7 An investigation into machining Al/SiC metal matrix composites

(a)

160 140 120 100 80 60 40 20 0 –20 –40 –60 0.40 26 0 0.35 0 24 0.3 25 0 22 . 0 0 Fe 20 0.20 5 ed in) 0 ra 0.1 0 /m 180 te (m 0.1 5 d 1 e (m 60 e 0.0 0 14 m/ Sp 0 0.0 re v)

> 80 < 80 < 60 < 40 < 20 160 < 144 < 124 < 104 < 84 < 64 < 44 < 24

160 140 120 100 80 60 40 20 0 –20 4 – 0 0.22 0.4 0.20 0 0 De 18 0.3 .35 pt 0. 16 0 0 ho . 0 .25 0 ) 4 fc . 20 0.1 0 rev ut m/ 0.1 .15 (m 0.12 (m m) 0.10 8 0 0.05 0 ate r d .00 0.0 Fee

> 140 < 128 < 108 < 88 < 68 < 48 < 28 2.6 < 2.4 2.6 < 2.5 < 2.3 < 2.1 < 1.9 < 1.7 < 1.5

14

0

16 0

5 0.3 .30 Fe 0 5 ed 0.2 0 2 ra te 0. (m 0.15 m/ 10 re 0. 5 v) 0.0 0 0.0

(b)

Ra (μm)

0

v)

/re

(mm

0

5

3 0.

0

Fee

te d ra

3 0.

5

2 0.

2 0.

(c)

15 0. 10 0.

8 0.0

05 0. 00 0.

.18 De 0 6 fth 0.1 of .14 ou 0 t( .12 mm 0 0 0.1 )

4 0.

3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 2 0.2 0 0.2

> 2.8 < 2.8 < 2.4 Uij ,

(8.8)

where Tij is the target value of the jth response (Yij ). The target value is a lower bound value of the response (Tij = Lij ). In the case that “the larger, the better” is the desired performance characteristic, the desirability function can be constructed as 0 { { { s { { ̂ { Yij (Xi ) − Lij ̂ dij (Yij ) = { ( ) { Tij − Lij { { { { 1.0 {

if

Ŷ ij (Xi ) < Lij

if

Lij ≤ Ŷ ij (Xi ) ≤ Tij

if

Ŷ ij (Xi ) > Tij .

(8.9)

The target value is the upper bound value of the response (Tij = Uij ).

8.4 Results and discussion Hybrid metal matrix composites have enormous applications in engineering fields and are replacing conventional materials due to their important properties. The surface roughness and cutting force are the two important parameters that affect the machining performance and also lead to the quality of the products produced. Machining of metal matrix composites is different than doing so with conventional materials. They contain a soft aluminum matrix and hard reinforcements. The typical surface profile observed in the machining of hybrid metal matrix composites at medium cutting conditions is provided in Figure 8.3 (cutting speed: 100 m/min; feed: 0.20 mm/rev.; depth of cut: 0.25 mm). Figure 8.3 indicates the feed mark and tiny variations on the surface profile. In the manufacturing processes, modeling and optimization are the important criteria used to improve performance. In the present investigation, modeling and optimization are carried out for the machining of hybrid metal matrix composites. The modeling of machining parameters is carried out using the Response Surface Method (RSM). The second order response surface represents the surface roughness and cutting force in machining, and it is developed as a function of cutting peed (V), feed (f )

188 | 8 Optimization of machining parameters of hybrid MMC composites

200μm

Fig. 8.3. Typical machined surface observed at medium cutting condition.

and depth of cut (d). The empirical relation is expressed as follows: surface roughness (Ra ) = + 3.11000 − 0.012767 ⋅ V + 2.76111 ⋅ f − 3.72000 ⋅ d − 6.00000E−003 ⋅ V ⋅ f + 3.00000E−003 ⋅ V ⋅ d + 0.36667 ⋅ f ⋅ d + 3.06667E−005 ⋅ V 2 + 4.66667 ⋅ f 2 + 3.25333 ⋅ d2

(8.10)

cutting force (Fz ) = + 24.03704 − 0.56000 ⋅ V − 317.22222 ⋅ f + 389.33333 ⋅ d + 1.91667 ⋅ V ⋅ f + 0.64000 ⋅ V ⋅ d + 86.66667 ⋅ f ⋅ d − 1.31111E−003 ⋅ V 2 + 1122.22222 ⋅ f 2 − 348.44444 ⋅ d2 .

(8.11)

The confidence interval and adequacy of the models are checked by using analysis of variance (ANOVA). Analysis of variance consists of partitioning the total variation in an experiment into components ascribable to the controlled factors and error. In the ANOVA table, the sum of squares indicates the square of deviation from the grand mean. The mean square is estimated by dividing the sum of squares by degrees of freedom. F-ratio is used to check the adequacy of the model developed in which the calculated F-value is greater than the table F-value. Tables 8.7 and 8.8 show the analysis of variance results for surface roughness and cutting force. From the analysis of the model, it has been noted that the F-value for surface roughness is 19.6287 and the Fvalue for cutting force is 22.50808, which are greater than the F-table value and hence the developed models are said to be adequate at a 95 % confidence level. Further, the coefficient of correlation (R-Sq) is used to check the correlation between the model and experimental values. The results indicated that the R-Sq value for surface roughness is 0.9122 and R-Sq value for cutting force is 0.9226 and they are very close to 1, which indicate that the models are said to be adequate. The normal probability plot for surface roughness and cutting force with respect to the internally studentized residual is presented in Figure 8.4. The normal probability plots are used to arrive at the normality assumption in the experimental run and model. It is based on the central limit theorem [28, 29]. In Figures 8.4a,b the data are

8.4 Results and discussion

|

189

Table 8.7. ANOVA for surface roughness. Source

Model V – cutting speed f – feed d – depth of cut Vf Vd fd V2 f2 d2 Residual Cor Total

Sum of squares

Degrees of freedom

5.331906 1.805 3.192022 0.0098 0.0108 0.016875 0.001008 0.035267 0.013067 0.248067 0.513094 5.845

9 1 1 1 1 1 1 1 1 1 17 26

Mean square 0.592434 1.805 3.192022 0.0098 0.0108 0.016875 0.001008 0.035267 0.013067 0.248067 0.030182

F -value

19.6287 59.8038 105.759 0.324697 0.357829 0.559108 0.033408 1.168466 0.432929 8.21902

p-value prob > F < 0.0001 < 0.0001 < 0.0001 0.5763 0.5576 0.4648 0.8571 0.2948 0.5194 0.0107

significant

Table 8.8. ANOVA for cutting force. Source

Model V – cutting speed f – feed d – depth of cut Vf Vd fd V2 f2 d2 Residual Cor total

Sum of squares

Degrees of freedom

47233.75 636.0556 24200 16805.56 1102.083 768 56.33333 64.46296 755.6296 2845.63 3963.88 51197.63

9 1 1 1 1 1 1 1 1 1 17 26

Mean square 5248.194 636.0556 24200 16805.56 1102.083 768 56.33333 64.46296 755.6296 2845.63 233.1694

F -value

22.50808 2.727869 103.7872 72.07445 4.726535 3.293743 0.241598 0.276464 3.24069 12.20413

p-value prob > F < 0.0001 0.1170 < 0.0001 < 0.0001 0.0441 0.0872 0.6293 0.6058 0.0896 0.0028

significant

spread in a straight line which shows the correlation that exists between the experimental and predicted values. Further, Figure 8.5 shows the correlation between the actual experimental values and the predicted values by model. From the analysis of the figure, it is noted that the predicted and actual values of surface roughness are close to each other and hence the models developed for predicting surface roughness and cutting force are very useful for the prediction of surface roughness and cutting force in the machining of hybrid metal matrix composites.

99

99

95 90 80 70 50 30 20 10 5

95 90 80 70 50 30 20 10 5

Normal % probability

Normal % probability

190 | 8 Optimization of machining parameters of hybrid MMC composites

1

1 –1.54 –0.70 0.13 0.96 1.79 Internally studentized residuals

(a)

–2.70 –1.39 –0.09 1.21 2.51 Internally studentized residuals (b)

3.20

210.00

2.70

162.50 Predicted

Predicted

Fig. 8.4. Normal probability plot for residuals of surface roughness (a); and cutting force (b).

2.20

1.70

67.50

1.20

20.00 1.24

(a)

115.00

1.71

2.18 Actual

2.65

3.12

20.51

65.88 111.25 156.63 202.00 Actual

(b)

Fig. 8.5. Correlation graph for surface roughness (a) and cutting force (b) in machining hybrid metal matrix composites.

Optimization of process parameters in machining is an important concern; it is important for process engineers to obtain the best process parameters to improve their productivity and to reduce the cost of manufacturing. In the present investigation, response surface-based desirability optimization is carried out to achieve the best surface roughness and to minimize the cutting force in the machining of hybrid metal matrix composites. The desirability function used for the optimization of process parameters is explained already.

8.4 Results and discussion

| 191

The optimization analysis is carried out by using DESIGN-EXPERT® software. The desirability analysis is carried out in two ways: – at first, the desirability values are calculated for the response’s surface roughness (Ra ) and cutting force (Fz ); – the desirability value is maximized by identifying the optimal parameters. The input parameters (cutting speed, feed and depth of cut), the output parameters, goals used, the lower limit, upper limit, weights used and their importance are given in Table 8.9. Table 8.9. Optimization parameters considered. Name

Goal

Cutting speed Feed Depth of cut Surface roughness Cutting force

is in range is in range is in range minimize minimize

Lower limit

Upper limit

Lower weight

Upper weight

Importance

50 0.1 0.25 1.24 39

150 0.3 0.75 3.12 202

1 1 1 1 1

1 1 1 1 1

3 3 3 3 3

Using Design Expert software, optimization is carried out. Based on the optimization, several solutions are obtained. There are 16 solutions obtained by simulating the software. The best solution has been chosen based on the highest desirability value. The solutions obtained from the optimization are tabulated in Table 8.10. Figure 8.6 shows the contour graph for maximum desirability with respect to the different parameters in the machining of hybrid metal matrix composites. The results are obtained from the desirability function by maximizing desirability. The maximum desirability obtained by varying the parameters such as spindle speed, feed and depth of cut is 0.94934. From Figure 8.6a it has been asserted that the desirability is better at high cutting speed and at a lower feed rate. Figure 8.6b shows the effect of cutting speed and depth of cut on desirability. The results indicated that the best desirability is achieved at a low depth of cut and at a high cutting speed. Figure 8.6c shows the effect of feed and depth of cut on desirability in machining hybrid metal matrix composites. The figure has indicated that the better surface roughness and minimal thrust force are arrived at low feed and low depth of cut. Further, Figures 8.6 and 8.7 indicate the minimal surface roughness and cutting force achieved in the machining of metal matrix composites. Figure 8.7a shows the effect of cutting speed and feed on the surface roughness in the machining of hybrid metal matrix composites. The result indicates that minimal surface roughness is achieved at a high cutting speed and low feed rate. Figure 8.7b shows the effect of cutting speed and depth of cut in the machining of hy-

192 | 8 Optimization of machining parameters of hybrid MMC composites 0.30 0.324 0.449 Feed, mm/rev

0.25 0.574 0.20 0.699 0.15

0.824 Prediction 0.94934

0.10 50 (a)

75 100 125 Cutting speed, m/min

150

Depth of cut, mm

0.75

0.63 0.641 0.703

0.50

0.764 0.826

0.38

0.888 Prediction 0.94934

0.25 50 (b)

75 100 125 Cutting speed, m/min

150

Depth of cut, mm

0.75

0.305

0.63

0.434 0.691

0.50

0.562

0.820 0.38 Prediction 0.94934 0.25 0.10

(c)

0.15

0.20 Feed, mm/rev

0.25

0.30

Fig. 8.6. Contour graphs observed for cutting speed vs. feed (a), cutting speed vs. depth of cut (b) and feed vs. depth of cut (c) at maximum desirability.

8.4 Results and discussion

|

193

0.30

Feed, mm/rev

0.25

0.20

0.15 Prediction 1.42566 0.10 50

75 100 125 Cutting speed, m/min

(a)

150

Depth of cut, mm

0.75

0.63

0.50 Prediction 1.42566 0.38

0.25 50

75 100 125 Cutting speed, m/min

(b)

150

Depth of cut, mm

0.75

0.63

0.50

0.38

Prediction 1.42566

0.25 0.10 (c)

0.15

0.20 0.25 Feed, mm/rev

0.30

Fig. 8.7. Contour graphs observed for cutting speed vs. feed (a), cutting speed vs. depth of cut (b) and feed vs. depth of cut (c) at minimum predicted surface roughness.

194 | 8 Optimization of machining parameters of hybrid MMC composites Table 8.10. Best solutions obtained for surface roughness and cutting force. Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Cutting speed

Feed rate

Depth of cut

Surface roughness

Cutting force

150 150 150 148 150 149 150 150 150 137 150 132 150 150 150 91

0.1 0.1 0.1 0.1 0.1 0.11 0.1 0.1 0.11 0.1 0.1 0.1 0.13 0.1 0.1 0.1

0.31 0.31 0.31 0.31 0.31 0.31 0.33 0.28 0.3 0.29 0.25 0.28 0.29 0.57 0.56 0.33

1.425655 1.428182 1.422482 1.436275 1.436722 1.449584 1.409023 1.460737 1.482617 1.508142 1.521568 1.540939 1.549323 1.331747 1.34184 1.700651

39.00042 38.99983 39.7334 39.00038 39.00061 38.99907 42.90546 31.23576 39.00078 39.00066 21.24082 38.99891 38.99977 87.14936 87.50764 70.89652

Desirability 0.949339 0.948632 0.948089 0.94636 0.946233 0.942613 0.942491 0.939461 0.933244 0.925942 0.922079 0.916475 0.914039 0.818669 0.815081 0.779254

Selected

brid metal matrix composites. The results indicate that the increase in cutting speed reduces the surface roughness, whereas the increase in depth of cut increases; the surface roughness in the machining of hybrid metal matrix composites. The relation between the feed and depth of cut in machining hybrid metal matrix composites is presented in Figure 8.7c. As discussed above, the results are similar for surface roughness with an increase of feed and depth of cut; The minimal predicted surface roughness is 1.42565 μm. The real optimum solution is not exactly the above condition. The minimal surface roughness is achieved at a cutting speed of 150 m/min, feed of 0.10 mm/rev and at a depth of cut of 0.31 mm. Figure 8.8a shows the relation between the cutting speed and feed rate with respect to the cutting force in the machining of hybrid metal matrix composites. The result indicates that the increase of cutting speed reduces the cutting force, whereas the increase of feed increases the cutting force in the machining of hybrid metal matrix composites. The relation between the cutting speed and depth of cut in the machining of hybrid metal matrix composites is presented in Figure 8.8b. The result indicates that the minimum cutting force is arrived at a high cutting speed whereas the depth of cut is to be maintained at 0.31–0.33 mm. The effects of feed and depth of cut (Figure 8.8c) indicate that the minimal cutting force is achieved at a high cutting speed and a lower depth of cut (0.30 mm). The minimum cutting force is 39 N. From the results, it has been asserted that the minimal surface roughness and cutting force are achieved at the following conditions:

8.4 Results and discussion

|

195

0.30 118.438

Feed, mm/rev

0.25

102.55

86.6627

0.20

70.7749

0.15

54.8872 Prediction 39

0.10 50 (a)

75 100 125 Cutting speed, m/min

150

Depth of cut, mm

0.75

0.63 98.832

0.50 83.1675

0.38

67.503

Prediction 39

51.8385

36.174

0.25 50 (b)

75 100 125 Cutting speed, m/min

150

Depth of cut, mm

0.75

0.63

161.925 133.642

0.50 105.359 77.0756

0.38 Prediction 39

48.7926

0.25 0.10 (c)

0.15

0.20 0.25 Feed, mm/rev

0.30

Fig. 8.8. Contour graphs observed for cutting speed vs. feed (a), cutting speed vs. depth of cut (b) and feed vs. depth of cut (c) at minimum predicted cutting force.

196 | 8 Optimization of machining parameters of hybrid MMC composites – – –

cutting speed: 150 m/min; feed: 0.10–0.13 mm/rev.; depth of cut: 0.29–0.33 mm.

Parameters/Response

Figure 8.9 shows the desirability graph obtained at an optimal level for input parameters and output responses. The desirability is fully achieved for the input machining parameters cutting speed, feed and depth of cut. The desirability value obtained for the response surface roughness is 0.9012; the value for cutting force is 1.000. The combined desirability obtained is 0.949. Cutting speed

1.000

Feed

1.000

Depth of out

1.000

Surface roughness

0.9012

Cutting force

1.000

Combined

0.949

0.000

0.250

0.500 0.750 Desirability

1.000

Fig. 8.9. Combined desirability values.

The experimental results, model results and optimal conditions used by obtaining the results from desirability analysis are validated by means of a validation test. Figures 8.10 and 8.11 show the validation test results for surface roughness and cutting force in the machining of hybrid metal matrix composites. The validation experiments are carried out at optimal conditions, which are obtained from the desirability analysis and performed at the selected conditions. The results indicate that the experimen-

Surface roughness, μm

5 4 3 Optimal level

2

Experimental results Response surface model Verification test result

1 0 1

2

3 Trial No.

4

5

Fig. 8.10. Validation test result for surface roughness in machining hybrid metal matrix composites.

8.5 Conclusions

Cutting force, N

200 160 120 Optimal level

80

|

197

Experimental results Response surface model Verification test result

40 0 1

2

3 Trial No.

4

5

Fig. 8.11. Validation test result for cutting force in machining hybrid metal matrix composites.

tal results and model results are very close to each other and the optimal conditions obtained from the desirability analysis give better values of surface roughness and cutting force in the machining of hybrid metal matrix composites; this technique is thus effectively used for the modeling and optimization of process parameters in the machining of hybrid metal matrix composites.

8.5 Conclusions –

–

–

–

– –

The experiments are conducted to analyze surface roughness and cutting force in the machining of hybrid metal matrix composites. Response surface models are developed for predicting the surface roughness and cutting force within the considered limits. Multiple performance optimization is carried out to reduce surface roughness and cutting force in the machining of hybrid metal matrix composites. The optimal conditions improve the surface finish and reduce the cutting force in the machining of hybrid metal matrix composites. Multi-response optimization is carried out, and the optimum conditions arrived are: a cutting speed of 150 m/min; feed of 0.10–0.13 mm/rev. and depth of cut of 0.29–0.33 mm. The results indicated that the increase of feed increases the surface roughness and cutting force in the machining of hybrid MMCs, whereas the increase of cutting speed reduces the surface roughness and cutting force. The results are validated by the validation experiments; they show a better correlation between experimental and predicted results. Modeling and optimization are further improved by adopting an increased number of variables and level of parameters.

198 | 8 Optimization of machining parameters of hybrid MMC composites

References [1] [2]

[3] [4] [5] [6] [7]

[8]

[9] [10] [11]

[12] [13] [14] [15] [16] [17] [18] [19]

[20] [21]

Manna A, Bhattacharayya B. A study on different tooling systems during machining of Al=SiCMMC. Journal of Materials Processing Technology 2002;13(3):476–482. Lane CT. Machining characteristics of particle-reinforced aluminum. In: Proceedings of the Conference Fabrication of Particulate Reinforced Metal Matrix Composites, Montreal, Canada, 1990. Looney LA, Monaghan;M, O’Reilly P, Taplin DMR. The turning of an Al-SiC metalmatrix composite. Journal of Materials Processing Technology 1992;33(4):453–468. Monaghan J, O’Reilly P. Machinability of an Al=SiC metal matrix composite. Processing of Advanced Materials 1991;2:37–46. Davim JP. Application of Merchant theory in machining particulate metal matrix composites. Materials and Design 2007;28(10),:2684–2687. Davim JP, Muthukrishnan N. Optimization of machining parameters of Al/SiC-MMC with ANOVA and ANN analysis. J. Materials Processing Technology 2009;209:225–232. Pramanik A, Zhang LC, Arsecularatne JA. Prediction of cutting forces in machining of metal matrix composites. International Journal of Machine Tools and Manufacture 2006;46:1795–1803. Ravindran P, Manisekar K, Narayanasamy P, Selvakumar N, Narayanasamy R. Application of factorial techniques to study the wear of Al hybrid composites with graphite addition. Materials and Design 2012;39:42–54. Andrews C, Feng H, Lau W. Machining of an aluminum=SiC composite using diamond inserts. Journal of Materials Processing Technology 2000;102:25–29. Lin JT, Bhattacharyya D, Lane C. Machinability of silicon carbide reinforced aluminum metal matrix composites. Wear 1995;181–183:883–888. Davim JP, Baptista AM. Relationship betwen cutting force and PDC cutting tool wear in machining silicon carbide reinforced aluminum. Journal of Materials Processing Technology 2000;103:417–423. El-Gallab M, Sklad M. Machining of Al=SiC particulate metal matrix composites. Part I. Tool performance. Journal of Materials Processing Technology 1998;83:151–158. Everett RK, Arsenault RJ, 1991, “Metal Matrix composites: processing and interfaces”, Academic Press, San Diego, CA 92101. Composite Materials Handbook, Metal Matrix Composites, Department of Defense, USA, Philadelphia, PA 4;1999,19111–5094. Farhad N. Machining of aerospace titanium alloys. International Journal of Robotics and CIM 2001;17:99–106. Che-Haron CH, Jawaid A. The Effect of Machining on surface integrity of titanium alloy Ti-6 %Al-4 %V. Journal of Materials Processing Technology 2005;166:188–192. Colding BN. A predictive relationship between Forces, Surface Finish and Toll-life. Annals of the CIRP 2004;53(1):85–90. Montgomery DC. Design and Analysis of Experiments. New York: John Wiley and Sons; 1991. Kwak JS. Application of Taguchi and Response Surface Methodologies for Geometric Error in Surface Grinding Process. International Journal of Machine Tools and Manufacture 2005;45:327–334. Derringer G, Suich R. Simultaneous optimization of several response variables. J. Qual. Technol. 1980;12:214–219. Jinshan, L, Cuiqing YM,;an L,;ei Z,;ing X. Medium optimization by combination of response surface methodology and desirability function: An application in glutamine production. International Journal of Advanced Manufacturing Technology 2007;74:563–571.

References

| 199

[22] Derringer G, Suich R. Simultaneous optimization of several response variables. J. Qual. Technol. 1980;12:214–219. [23] Vining GG, Myers RH. Combining Taguchi and response Surface philosophies: A dual response approach. Journal of Quality Technology 1990;22:38–45. [24] Kwon YJ, Kim YJ, Cha MS. Desirability function modeling for dual response surface approach to robust design. IEMS 2008;7(3):197–203, [25] Rajakumar S, Balasubramanian V. Multi-response optimization of friction-stir-welded AA1100 aluminum alloy joints. Journal of Materials Engineering and Performance 2012; 21(6):809–822. [26] Palanikumar K, Karthikeyan R. Optimal machining conditions for turning of particulate metal matrix composites using Taguchi and response surface methodologies. Machining Science and Technology 2006;10(4):417–433. [27] Palanikumar K, Muthukrishnan N, Hariprasad KS. (), Surface roughness parameters optimization in machining A356/SiC/20p metal matrix composites using response surface methodology and desirability function. Machining Science and Technology 2008;12(4):529–545. [28] Lochner RH, Mater JE. Designing for Quality. London: Chapman & Hall; 1990. [29] Palanikumar K, Karthikeyan R. Assessment of factors influencing surface roughness on the machining of Al/SiC particulate composites. Materials and Design 2007;28(5):1584–1591.

Index AA6061/Al, 128 abrasion, 176 abrasive – nature, 171 – particle, 153 Ag, 4 Ag-3wt.%Si/POCO-HTC graphite foam, 26 Ag-3%Si/graphite flake + carbon fiber, 21 Ag/SiC, 11 aggloromate, 131 Al, 4 Al-12wt.%Si/POCO-HTC graphite foam, 26 Al-12wt.%Si/SiC, 12 Al-7wt.%Si/diamond, 27 Al-oxides, 44, 61 Al/diamond, 30 Al/SiC, 11–13, 17 Al/SiC-diamond, 16 Al/SiC-MMC, 123 Al4 C3 , 12, 19, 29 Al6061 aluminum alloy, 180 alignment, 48, 52, 63 Alumina (Al2 03 ), 128 aluminum carbide, 18, 28 aluminum/diamond, 28 Aluminium (Al), 43 – mechanical properties, 44, 48 analysis of variance, 188 ANOVA, 123 Ashby, 3 atoms, 57

behavior of chips, 160 β titanium alloy, 81 bimodal – mixture, 7, 14, 27 – particle distribution, 17 – particle mixture, 13 biosafety, 82 bond, 39 bonding, 118 buckling, 57, 59, 60 BUE, 165

carbide, 48, 50, 58, 66 carbon – fiber, 18 – material, 20 – nanotube, 18, 19 – nanotubes (CNTs), 39 carbon/metal, 18 casting process variable, 125 central composite design (CCD), 163 chemical vapor deposition, 67 circularity, 174, 176 classification of unconventional process, 147 clay-coated mold, 127 CNT-Co composites, 55 CNT-metal interface, 43 – copper, 57 – iron, 62 – magnesium, 64 – nickel, 68 CNT-reinforced Al, 47 CNT-reinforced metal matrix composites, 39 Co matrix, 55 coated, 175 coating, 163, 168 – CNT-reinforced Si, 65 Cobalt (Co), 55 ff. Coefficient of correlation, 188 cold spraying, 109 cold-sprayed, 114 combined desirability, 196 complex profile, 147 compo-casting, 125 composite – agglomerate, 106 – desirability, 186 – material, 160 compressive, 71 – mechanical properties, 57 – yield strength, 62, 64 computer-controlled detonation spraying (CCDS), 110 concentration gradient, 105 contact time, 28, 30 continuous carbon fiber, 19 contour graph, 191

202 | Index % contribution, 138 conventional machining, 142 cooling rate, 115 Copper (Cu), 57 ff., 115 – mechanical properties, 57 correlation, 189 corrosion, 43, 57, 64, 68, 71 creep, 48 crystalline Ni/NiO nanoparticles, 69 crystallite growth, 114 Cu, 4 Cu/SiC, 11 current density, 149 cutting – force, 183 – speed, 191 damping, 44, 59 deformation, 51, 58, 65 delamination, 108 density, 130 Design-Expert, 191 desirability based approach, 185 detonation spraying, 106 diameter, 173, 176 diamond, 4, 16, 26 – nanoparticle, 20 diamond-based composite, 26 diamond/Cu-Cr, 27 dielectric, 150 dimensional accuracy, 159 dispersion, 39, 47, 61, 63 drilling, 145, 159 ductile behavior, 47 ductility, 51, 64 dynamometer, 162, 183 electric discharge machining, 150 electrochemical machining, 148 electrolyte, 148, 150 electron – beam treatment, 118 – probe microanalysis (EPMA), 72 electronic, 1, 21 electroplating, 148 elongation, 44, 64, 66, 72 experiment result, 184 extrusion, 44, 64, 72

F-test, 136 failure, 59, 63 Faraday’s law, 148 feed force, 164 first-generation, 10 flank wear, 169 flushing, 152 fracture, 44, 59, 63, 67 friction coefficient, 48, 55, 58, 68, 71 functionalization of CNTs, 40, 49, 56 furnace temperature, 126 gas pressure infiltration, 5, 16 Gauss elimination method, 137 generalized differential effective medium scheme (GDEMS), 8 grain, 47, 50, 57, 54, 61, 65, 72 graphite – crucible, 127 – flake, 20 – foam, 22 – nanoplatelet, 20 – particle, 18, 19 gravity casting, 124 hardening, 47, 58 hardness, 44, 57, 58, 65, 115 Hashin, 9 Hasselman–Johnson, 8 homogenizing, 126 hybrid metal matrix composite, 181 hybridization, 63 image analyzer, 131 in situ – reaction, 85 – TEM, 95 interface, 13, 28, 31, 44, 71 interfacial interaction, 103 Iron (Fe), 61 ff. – mechanical properties, 62 isotropic, 24, 26 lattice parameter, 114 layered structure, 23 level of parameter, 197 linear rule of mixtures (LROM), 9 liquid infiltration, 4

Index | 203

liquidus temperature, 129 LM6 alloy matrix, 125 machinability, 170 machining, 159, 179 machining performance, 187 Magnesium (Mg), 63 ff. – mechanical properties, 64 manufacturing, 190 material removal – mechanism, 153 – process, 142 – rate, 147 mathematical – model, 137 – milling, 105, 106 mechanically assisted infiltration, 5 mechanically milled, 107, 108 melting, 114, 115 mesophase pitch, 22 mesophase pitch-derived foam, 25 metal matrix composite (MMC), 179 metal-ceramic composite, 106 metal-reinforcement interface, 31 micro-hardness, 161 microelectronics I, 2 microelectronics II, 2 micrograph, 161 microprocessor-based PID, 126 microstructural feature, 103 microstructure, 44, 56, 63, 65, 69, 72, 133 – development, 110 milling, 47, 54, 144, 159 mixing, 108 mixtures of SiC and diamond particles, 17 modeling, 184 monolithic material, 3 multi-response, 197 multiple performance optimization, 197 nano-dimensional control, 31 nanocrystalline – Al-Si, 48 – Co, 55 nanoparticles, 55, 63, 66 nanotechnology, 31 Ni-based CNT composite, 68 Nickel (Ni), 67 ff. nitridation, 109

nomenclature, 182 nonreactive, 104 normal probability plot, 188 objective function, 186 optical – micrograph, 134 – microscope, 161 optimization, 135, 168, 184 optimum solution, 194 opto-electronics, 2 oriented discs, 23 overheating, 1 packing, 7 parallel, 22 particle reinforcement, 180 particulate reinforced metal matrix composite (PMMC), 141 perpendicular, 22 phase composition, 109 planar alignment, 21 plasma spraying, 109 plastic deformation, 15 polycrystalline diamond tool (PCD tool), 182 porosity, 12, 118 powder – metallurgy, 5 – rolling, 47 power electronics, 2 prediction, 189 process parameter, 149, 151, 154, 190 processing parameter, 143 Raman spectroscopy, 68 rapid cooling, 110 resolidification, 115 reaction product, 149 reactive, 104 – metallic material, 109 – thermal spraying, 110 reactivity, 11 reinforcement, 40, 46, 52, 59, 62, 64, 68, 71, 150 relative density, 104 replication method, 22 response surface – method, 185 – model, 197 route compaction, 14

204 | Index S/N ratio, 135 scanning electron microscope, 131 Schapery’s model, 9 second-generation, 10 segregation, 7 shape, 6 short carbon fiber, 19 Shtrickman, 9 SiC, 4, 16, 21 SiC-based composites, 11 single-crystalline iron-based nanowires, 61 sintering, 44, 47, 58, 69, 71 size, 6 – distribution, 6 slurry, 153 solid-state diffusion, 108 spark plasma sintering, 118 spindle speed, 173 spraying distance, 109 stir casting, 123, 161, 180 stirring, 181 storage modulus, 45 strengthening, 44, 51, 54, 58, 66, 72 substrate, 108 substrate/coating interface, 108 superelasticity, 92 surface – finish, 160, 163 – integrity, 159 – profile, 187 – roughness, 151, 163, 183 Taguchi method, 123 tangential, 164 tensile strength, 39, 44, 48, 52, 58, 63, 67, 72 thermal – barrier, 13 – conductance, 29, 30 – conductivity, 1, 16, 17 – expansion, 46, 53, 67

– expansion coefficient, 1, 17 – guide, 21 – management, 1 – spraying, 103 thermo-mechanical fatigue, 3 thermoelasticity theory, 15 third-generation, 10 thrust force, 164, 176 Ti3 SiC2 , 114 TiB reinforcement, 83 Titanium (Ti), 71 ff. – mechanical properties, 71 tool – life, 175 – wear, 151, 159, 163, 182 torque, 173 transmission electron microscopy, 47, 49, 61, 65, 68, 70 turning, 143 two-bladed stirring, 124 ultimate tensile strength, 130 ultrasonic machining, 153 uncoated, 175 unconventional machining, 146 vacuum consumable arc furnace, 84 validation test, 196 volume fraction, 6, 7 WC-Co, 118 wear, 48, 55, 58, 68 – resistance, 109 wettability, 12, 104 XRD phase analyse, 114 yield strength, 48, 51, 53, 58, 69, 71 Young’s modulus, 39, 44, 51, 58