UCLA Electronic Theses and Dissertations

UCLA UCLA Electronic Theses and Dissertations

Title Achieving Uniform Nanoparticle Dispersion in Metal Matrix Nanocomposites

Permalink https://escholarship.org/uc/item/91r0q1x9

Author Xu, Jiaquan

Publication Date 2015

Peer reviewed|Thesis/dissertation

eScholarship.org Powered by the California Digital Library University of California UNIVERSITY OF CALIFORNIA

Los Angeles

Achieving Uniform Nanoparticle Dispersion in Metal Matrix Nanocomposites

A dissertation submitted in partial satisfaction of the

requirements for the degree Doctor of Philosophy

in Materials Science and Engineering

Jiaquan Xu

2015

Jiaquan Xu

2015

ABSTRACT OF THE DISSERTATION

Achieving Uniform Nanoparticle Dispersion in Metal Matrix

Nanocomposites

Jiaquan Xu

Doctor of Philosophy in Materials Science and Engineering

University of California, Los Angeles, 2015

Professor Xiaochun Li, Co-Chair

Professor Jenn-Ming Yang, Co-Chair

The objective of this study is to gain fundamental knowledge on the interactions between nanoparticles to achieve a uniform dispersion of nanoparticles in metals for manufacturing metal matrix nanocomposites (MMNCs). MMNC, also known as nanoparticles reinforced metal, is an emerging class of materials exhibiting unusual mechanical, physical, and chemical properties.

However, a lack of fundamental knowledge and technology on how to achieve a uniform nanoparticle dispersion in MMNCs has hindered the rapid development of the MMNC field. In

ii this dissertation, several methods were explored to achieve a uniform nanoparticle dispersion in

MMNCs.

In-situ oxidation method were applied to fabricate Al-Al2O3 nanocomposites with a uniform dispersion of Al2O3 nanoparticles. Pure Al nanoparticles were cold compressed in a steel mold and then melted in an alumina container. Al2O3 nanoparticles were in situ synthesized through the oxidation of the Al nanoparticle surfaces to form bulk Al nanocomposites during the process. Although some Al2O3 nanoparticles were distributed along the grain boundaries of some coarse Al grains, most Al2O3 nanoparticles were evenly distributed inside ultrafine Al grains to effectively restrict their grain growth. Moreover, the microhardness of the bulk Al nanocomposites is enhanced up to about three times as high as that of pure bulk Al.

Friction stir processing (FSP) were combined with semi-solid mixing to disperse 6 vol.%

SiC nanoparticles in Mg6Zn. Semi-solid mixing was effective to incorporate SiC nanoparticle into

Mg6Zn matrix before FSP. The low temperature at the semi-solid state reduced nanoparticles burning and oxidation effectively, while a high viscosity of the metal at semi-solid state trapped the nanoparticles inside the matrix metal. Also, FSP was used to process Mg + 6vol.% HA nanocomposites with a uniform dispersion and distribution of nanoparticles after mechanical stirring. The mechanical properties of Mg nanocomposites after FSP were significantly improved.

Unfortunately these two methods discussed above are not economical for mass manufacturing of MMNCs, while solidification processing is very promising as a versatile mass manufacturing method for production of bulk MMNC parts with complex geometry and high nanoparticle loading. However, the incorporation and de-agglomeration of nanoparticles in liquid metals are extremely difficult. Thus there is a strong need to fully understand the physics of the

iii interactions between nanoparticles inside metal melts in order to develop new pathways to achieve the uniform dispersion of nanoparticles for mass solidification processing of bulk MMNCs.

A theoretical model was successfully established to reveal the essential conditions for nanoparticle dispersion in molten metal during solidification nanoprocessing of bulk MMNCs.

The interactions between nanoparticles in a molten metal include three key potentials, the interfacial energy barrier at a short range (1~2 atomic layers) to resist nanoparticles to come further into atomic contact, the attractive van der Waals potential (dominant in the longer range from

0.4~10 nm), and the Brownian potential, kT. Three possible scenarios for nanoparticles in molten metals were theoretically predicted below.

1. Clusters: when the maximum interfacial energy barrier is less than about 10kT due to a

poor wetting between nanoparticles and metal melt, the nanoparticles will come close

into atomic contact to form larger clusters in the liquid metal.

2. Pseudo-dispersion: If the maximum interfacial energy barrier is high enough (e.g. more

than 10 kT) due to a good wetting between the nanoparticle and the molten metal and

the van der Waals attraction is much larger than the Brownian potential, nanoparticles

will be trapped into a local minimum potential to form pseudo-dispersion domains

where dense nanoparticles are separated by only a few layers of metal atoms.

3. Self-dispersion: When the maximum interfacial energy barrier is high and the van der

Waals attraction is smaller than the Brownian potential, nanoparticles will move freely

inside the molten metal in a self-dispersion and self-stabilization mode.

Based on theoretic study and availability of nanoparticles in the market, two material combinations, TiC (with a radius of 25 nm) in liquid Al and SiC (with a radius of 30 nm) in liquid

Mg, were first selected for the experimental study.

To avoid oxidation and burning of TiC nanoparticles, a novel method of salt assisted nanoparticles incorporation was developed to fabricate master Al-9vol.% TiC nanocomposites. A droplet casting method was developed to avoid the nanoparticle settling down and pushing during solidification. Microstructure studies revealed that TiC nanoparticles still form domains in Al matrix, indicating a pseudo-dispersion of TiC (50 nm in diameter) in pure liquid Al. However, TiC nanoparticles were successfully dispersed in the Mg18Al eutectic alloy.

Mg6Zn-1vol.% SiC nanocomposite ingots were first obtained by ultrasonic-assisted solidification processing. A new method was developed to concentrate SiC nanoparticles by evaporating Mg and Zn away from the Mg6Zn-1vol.%SiC ingots at 6 torr in a vacuum furnace.

After evaporation and a slow cooling at approximately 0.23 K/s, a sample with about 14 vol.%

SiC nanoparticles was obtained in an Mg2Zn matrix. Material characterizations by SEM, EDS, and Vickers hardness measurements revealed that SiC nanoparticles were self-dispersed in Mg.

Micropillar compression tests showed that the Mg2Zn-14vol.% SiC nanocomposites yield at a significantly higher strength of about 410 MPa with a good plasticity, while of 50 MPa with a very poor plasticity for pure Mg2Zn.

In summary, this dissertation establishes a theoretical framework and developed experimental methodologies to achieve a uniform dispersion of dense nanoparticles in metals. The study has significantly advanced the fundamental understanding on the interactions between nanoparticles in molten metals to obtain MMNCs with a uniform dispersion of dense nanoparticles for widespread applications.

The dissertation of Jiaquan Xu is approved by

Yong Chen

King-Ning Tu

Jenn-Ming Yang, Committee Co-Chair

Xiaochun Li, Committee Co-Chair

University of California, Los Angeles

2015

Dedicated to my family for their unconditional love.

vii

TABLE OF CONTENTS

TABLE OF CONTENTS ...... viii

NOMENCLATURE...... xiii

ACKNOWLEDGEMENTS ...... xii

Chapter 1 Introduction...... 1

1.1 Motivation ...... 1

1.2 Research objectives ...... 2

1.3 Work summary ...... 2

Chapter 2 Literature review ...... 4

2.1 Metal matrix nanocomposites ...... 4

2.2 Nanoparticle effects inside grains ...... 10

2.2.1 Nucleation effect ...... 10

2.2.2 Zener pinning effect...... 11

2.2.3 Orowan strengthening effect ...... 12

2.2.4 Other functional effects ...... 13

2.3 Nanoparticle dispersion ...... 14

2.3.1 De-agglomeration of nanoparticles ...... 14

2.3.2 Stabilization of dispersed nanoparticles in liquids ...... 17

2.3.3 Challenges of nanoparticles dispersion in molten metal ...... 25

viii

Chapter 3 In-situ oxidation method to fabricate Al nanocomposites with dispersed Al2O3 nanoparticles ...... 27

3.1 Introduction ...... 27

3.2 Materials and methods ...... 28

3.3 Results and discussions ...... 30

3.3.1 Microhardness ...... 30

3.3.2 Composition Analysis by XRD ...... 31

3.3.3 Microstructure Analysis ...... 31

3.4 Summary ...... 33

Chapter 4 Dispersion of nanoparticles through friction stir processing ...... 34

4.1 Introduction ...... 34

4.2 Friction stir processing of Mg6Zn-SiC nanocomposites ...... 35

4.2.1 Materials and methods ...... 35

4.2.2 Results and discussion ...... 37

4.2.3 Summary ...... 41

4.3 Friction stir process to disperse HA nanoparticles in Mg ...... 42

4.3.1 Materials and methods ...... 42

4.3.2 Results and discussion ...... 44

4.4 Summary ...... 46

Chapter 5 Theoretical study on nanoparticle dispersion in molten metals ...... 47 ix

5.1 Introduction ...... 47

5.2 Interactions between nanoparticles in molten metals ...... 47

5.2.1 Interfacial energy ...... 48

5.2.2 Van der Waals potential ...... 50

5.2.3 Brownian potential ...... 52

5.3 Three cases for nanoparticles dispersion in molten metals ...... 53

5.3.1 Clusters ...... 53

5.3.2 Pseudo-dispersion ...... 54

5.3.3 Self-dispersion ...... 55

Chapter 6 Experimental study for nanoparticle self-dispersion in molten metals ...... 57

6.1 Experimental study of dispersion of TiC nanoparticles in Al ...... 57

6.1.1 Theoretical analysis of TiC nanoparticle dispersion in liquid Al ...... 57

6.1.2 Salt assisted nanoparticle incorporation ...... 58

6.1.3 Quantitative analysis of TiC nanoparticle loading in Al ...... 60

6.1.4 Settling down of TiC nanoparticles in Al ...... 61

6.1.5 Dispersion of TiC in Al alloys in droplet casted sample ...... 68

6.1.6 Dispersion study of TiC nanoparticles in Mg-Al alloys ...... 70

6.1.7 Summary on TiC nanoparticle dispersion in Al alloy melts ...... 76

6.2 Experimental study on SiC nanoparticle dispersion in Mg ...... 77

6.2.1 Theoretical analysis on SiC nanoparticle dispersion in Mg melt ...... 77 x

6.2.2 Fabrication of Mg2Zn with 14vol.% SiC nanocomposites ...... 79

6.2.3 Self-dispersion of 14vol.% SiC in Mg2Zn ...... 81

6.2.4 Summary on self-dispersion of SiC in Mg ...... 84

Chapter 7 Conclusions ...... 85

Chapter 8 Future work ...... 89

8.1 Scientific determination of Hamaker constants for alloys and nanoparticles ...... 89

8.2 Size effect for nanoparticle dispersion in metal melts ...... 89

8.3 Enable mass production of MMNCs with self-dispersed nanoparticles ...... 90

References ...... 90

NOMENCLATURE

Al Aluminum

Mg Magnesium

TiCN Titanium Carbonitride

SiC Silicon Carbide

TiB2 Titanium Diboride

MgO Magnesium Oxide

Ni Nickel

Al2O3 Alumina

MMNC Metal Matrix Nanocomposites

SEM Scanning Electron Microscopy

EDS Energy Dispersive X-ray Spectroscopy

TEM Transmission Electron Microscopy

HA Hydroxyapatite

FSP Friction stir process

xii

ACKNOWLEDGEMENTS

I would like to thank Professor Xiaochun Li for his remarkable support and guidance, without which this would not be possible. His diligent involvement, systematic supervision, and his wealth of creativity benefit me a lot during my Ph.D. study and for my future career.

I would also like to acknowledge Professor Yong Chen, Professor King-Ning Tu, and

Professor Jenn-Ming Yang for serving in my doctoral committee. I sincerely appreciate their valuable inputs for my research.

I am also very grateful to Dr. Lianyi Chen, who helped me a lot in discussion and in experiments. And last but not the least, I would like to acknowledge all of my colleagues in the research group. I benefited greatly from the technical discussions with them, which inspired many wonderful ideas.

xiii

VITA 2012, M.S., Materials Science Program, University of Wisconsin-Madison, US

2010, B.S., Materials Science and Engineering, Tsinghua University, Beijing, China

JOURNAL PAPERS

 J. Xu, et al., Fabrication of hierarchical metallic nanocomposite core/metal shell nanostructures by

self-assembly, Journal of Nanoscience and Nanotechnology 15.7, 5479 (2015)

 J. Xu, et al., Urchin-like AlOOH nanostructures on Al microspheres grown via in-situ oxide

template, Materials Science and Engineering B 188, 89 (2014)

 J. Xu, et al., Assembly of metals and nanoparticles into novel nanocomposite superstructures,

Scientific Reports 3, 1730 (2013)

 J. Xu, et al., Theoretical study and pathways for nanoparticle capture during solidification of metal

melt, Journal of Physics: Condensed Matter 24, 255304 (2012)

 C. Cao, L. Chen, J. Xu, H. Choi, X. Li, Strengthening Al-Bi-TiC0.7N0.3 Nanocomposites by Cu

Addition and Grain Refinement, Materials Science and Engineering: A 651, 332 (2015)

 L. Chen, J. Xu, X. C. Li, Controlling phase growth during solidification by nanoparticles,

Materials Research Letters 3, 43 (2014)

 L. Chen, J. Xu, et al., Rapid control of phase growth by nanoparticles, Nature Communications

5, 3879 (2014)

 L. Chen, J. Peng, J. Xu, et al., Achieving uniform distribution and dispersion of a high percentage

of nanoparticles in metal matrix nanocomposites by solidification processing, Scripta Materialia

69, 634 (2013)

 L. Chen, D. Weiss, J. Morrow, J. Xu, X. Li, A novel manufacturing route for production of high-

performance metal matrix nanocomposites, Manufacturing Letters 1, 62 (2013)

 L. Chen, H. Konishi, A. Fehrenbacher, C. Ma, J. Xu, et al., Novel nanoprocessing route for bulk

grapheme nanoplatelets reinforced metal matrix nanocomposites, Scripta Materialia 67, 29 (2012) xiv

 C. Ma, L. Chen, J. Xu, et al., Control of fluid dynamics by nanoparticles in laser melting, Journal

of Applied Physics 117, 114901 (2015)

 C. Ma, L. Chen, J. Xu, et al., Effect of fabrication and processing technology on the

biodegradability of magnesium nanocomposites, Journal of Biomedical Materials Research

101B, 870 (2013)

CONFERENCE PAPERS

 J. Xu, et al., High performance Mg6Zn nanocomposites fabricated through semi-solid mixing and

friction stir processing, Magnesium Technology 2015, 383

 J. Xu, et al., Bulk Al matrix nanocomposites formed through in situ oxidation and melting of

aluminum nanoparticles, Proceedings of NAMRI/SME 42 (2014)

 L.Chen, J. Peng, J. Xu, et al., Achieving uniform distribution and dispersion of a high percentage

nanoparticles in Mg18Sn matrix by solidification processing, Magnesium Technology 2014, 465

 L. Chen, H. Choi, A. Fehrenbacher, J. Xu, et al., Uniform dispersion of nanoparticles in metal

matrix nanocomposites, TMS 2012 Supplemental Proceeding: Materials Processing and

Interfaces 1, 757 (2012)

 C. Ma, L. Chen, J. Xu, et al., Biodegradability and mechanical performance of hydroxyapatite

reinforced magnesium matrix nanocomposite, TMS 2012 Supplemental Proceeding: Materials

Processing and Interfaces 1, 725 (2012)

PATENTS

 X. Li, H. Choi, L. Chen, J. Xu, 2012 “Nanomaterial-based methods and apparatuses” US

20130236352 A1

 X. Li, L. Chen, J. Xu, 2013, “Improved immiscible alloy formation with stabilizing nanoparticles,”

P120305US01

Chapter 1 Introduction

1.1 Motivation

Nanocomposites can provide exciting physical, chemical, and mechanical properties for numerous applications. Solidification processing has a great potential for an economical fabrication of bulk nanocomposites, especially for those with crystalline materials as the matrix, such as metal matrix nanocomposites (MMNCs).

A uniform nanoparticle dispersion inside MMNCs is critical to achieve enhanced properties. The final distribution of nanoparticles inside MMNCs depends on the incorporation of nanoparticles, the dispersion of nanoparticles in molten metal and the pushing of nanoparticles during solidification. Tremendous work has been carried out to incorporate nanoparticles into molten metal and to overcome the pushing of nanoparticles during solidification. However, it is still a long standing challenge to achieve a uniform dispersion of nanoparticles in molten metal for industrial production. Poor wettability between ceramic nanoparticles and molten metal often results in nanoparticle repulsion out of molten metal. Possible contamination and air bubbles attached on the nanoparticle surfaces often aggravate this tendency. Nanoparticles readily aggregate to form agglomerations in molten metal, making it difficult to obtain a stable and uniform dispersion of nanoparticles inside the molten metal. Recently, ultrasonic cavitation processing has been successfully used to disperse nanoparticle in molten Al and Mg melts. A uniform distribution and dispersion of nanoparticles inside molten metals could be achieved during the ultrasonic process. However, once the ultrasonic processing stops, the uniform dispersion of nanoparticles is not stable and nanoparticle pushing during solidification readily form nanoparticle agglomerates. Moreover, ultrasonic process is still not suitable for economical mass production of 1

MMNCs. It is thus of significance to initiate theoretical and experimental studies to provide systematic pathways to achieve a uniform distribution and dispersion of nanoparticle for large scale solidification nanoprocessing of MMNCs.

1.2 Research objectives

This study is to discover effective pathways for a uniform nanoparticle dispersion for mass manufacturing of MMNCs. Theoretical and experimental studies are especially conducted to advance the fundamental understanding of the interactions between nanoparticles in molten metals.

Two specific research objectives are included: 1) Understand the physics and establish a theoretical framework on the interactions between nanoparticles in molten metals; 2) Conduct experimental study, under the theoretical guidance, to achieve a uniform dispersion of nanoparticles during solidification nanoprocessing of MMNCs.

1.3 Work summary

An extensive literature review has been conducted. Experimental studies on in-situ oxidation and friction stir processing were conducted to obtain a uniform dispersion of nanoparticles in small samples. To realize a mass production of bulk MMNCs, solidification processing is the most economic method. Theoretical and experimental studies thus were conducted to realize a uniform dispersion of nanoparticles during solidification processing of

MMNCs. The studies are detailed the chapters of this dissertation as briefly summarized below:

 Chapter 2 reviews the state-of-art on MMNCs processing.

 Chapter 3 presents the experimental results on Al-Al2O3 nanocomposites fabricated through

in-situ oxidation.

 Chapter 4 discusses the fabrication of Mg6Zn-SiC and Mg-HA nanocomposites through

friction stir processing.

 Chapter 5 conducts a theoretical study on the interactions of nanoparticles in molten metals

and predicts potential pathways to achieve a uniform dispersion of nanoparticles during

solidification processing of MMNCs.

 In Chapter 6, experimental studies on two model material systems (i.e. TiC/Al and SiC/Mg),

based on the theoretical guidance of Chapter 5, were conducted.

 Chapter 7 draws conclusions.

 Potential future work is recommended in Chapter 8.

Chapter 2 Literature review

2.1 Metal matrix nanocomposites

The demand of an energy-lean future has significantly affect material selections for engineering systems. Composite materials combining two or more different materials to achieve novel properties have gained popularity in high performance lightweight products for energy efficiency. Composite materials are already used in various applications, e.g. plain bearing, paper, tires, and cemented carbide cutting tools. Some emerging composite materials include in situ metal matrix composites, carbon-fiber-reinforced thermoplastic composites, and toughened ceramics by use of fiber-reinforcement [1,2,3,4,5,6,7,8,9,10,11].

Recent advances in producing nanostructured materials with novel properties have stimulated research to create multi-functional engineering materials by designing structures at the nanometer scale, e.g. nanocomposites. Motivated by the recent progress in nanotechnology, nanocomposites becomes one of the rapidly evolving areas for materials research [6,9,10,12].

Based on the matrix phase, nanocomposites are classified into three categories: metal matrix nanocomposites (MMNCs), polymer matrix nanocomposites, and ceramic matrix nanocomposites.

Based on the reinforce phase with nanostructure, nanocomposites can be classified into three categories: nanoparticles reinforced nanocomposites, nanoplatelets reinforced nanocomposites, and nanofiber reinforced nanocomposites [8].

In the past decade, much work has conducted on polymer matrix nanocomposites and many such materials are already used in various applications [13,14,15]. In the meantime, MMNCs are emerging as a new class of nanocomposites that incorporate nanosized reinforcements in metal matrix for novel properties. With a similar concept to the metal matrix composites but with

4 nanoscale reinforcing particulates, MMNCs are much more promising in achieving exciting mechanical properties, such as significantly enhanced strength and ductility. Metallic nanocomposites containing nanoparticles or carbon nanotubes could offer distinct advantages over polymeric composite due to the inherent high temperature stability, high strength, high modulus, wear resistance, and thermal and electrical conductivity of the metal matrix [10,11,16,17,18].

However, the development of MMNCs is still in its infancy. The MMNCs synthesized to date include Al-B4C [19], Al-AlN [20], Al-AlB2 [21], Al-MgB2 [21,22], Al-MgO [23,24], Al-TiC

[25], Al-SiC [26], Al-ZrO2 [27], Al-CNT [28], Al-Al2O3 [29,30], Al-TiB2 [31], Al-Al4C3 [32], Cu-

Al2O3 [33,34], Cu-NbC [35], Cu-TaC [36], Cu-TiC [37], Cu-CNT [38], Cu-TiO [39], Cu-CeO2

[40], Fe-Al2O3 [41], Fe-FeyO [42], Fe-SiO2 [43], Fe-NbC [44], Ni-Al2O3 [45], Ni-SiC [46], Ni-

CeO2 [47], Ni-TiO2 [48], Ni-ZrO2 [49], Ni-La2O3 [50], Ni-CNT [51], Ni-TiN [52], Mg-BN [53],

Mg-Al2O3 [54], Mg-TiC [55], Mg-Cr2O3 [56], Mg-SiC [57], Mg-Y2O3 [58], Mg-graphene [59],

Mg-CNT [60], Mg-MgO [61], Mg-TiB2 [62], Mg-TiO2 [63], Mg - hydroxyapatite [64], Zn-Al2O3

[65], Zn-TiO2 [66], Zn-ZrO2 [67], Zn-SiC [68] and many more alloys based nanocomposites, including Al alloys, Cu alloys, Fe alloys, Ni alloys Mg alloys, and Zn alloys. MMNCs offer a great potential for numerous applications, however, their cost, difficulty of mass manufacturing, and sometimes poor ductility due to processing problems hindered their widely usage in industry

[69,70].

There are a few methods available to synthesize and process bulk metal matrix nanocomposites, e.g., wet chemical methods [38,71,72], electrodeposition [45,46,47,48,49,50,51], friction stir processing [59,73,74], squeeze casting [75,76], stir casting [23,24,77,78,79,80,81,], mechanical alloying [21,22,24,25,26,27,28,30,31,34,35,36,37,42,55,65,82,83,84], laser cladding

[85,86,87,88,89,90,91] and solidification processing.

In the wet chemistry method, nano-reinforcements and the precursors of matrix materials are mixed in a solution. With delicate chemical reactions, precipitation, and sintering, the precursors were transformed into the solid matrix with a uniform distribution/dispersion of the nano-reinforcements. For example, in the “molecular-level mixing” process, CNTs were suspended in a solvent by surface functionalization and further mixed with Cu ions. After drying, calcination and reduction of the CNT suspension, CNTs are homogeneously implanted within the

Cu powders. The CNT/Cu nanocomposite, consolidated by spark plasma sintering of CNT/Cu composite powders, offered three times the strength and twice the Young’s modulus as the Cu matrix [38,71,72].

Electrodeposition was used for the fabrication of Ni nanocomposite coating, Zn nanocomposite coating, and Cu nanocomposite coating to protect the original surface from corrosion and wear [39,40,45,46,47,48,49,50,51,66,67,92,93]. Electrodeposition could offer a cost effective means to fabricate nanocomposite surface coating. Insoluble, second-phase nanoparticles are normally suspended in a conventional plating electrolyte and later integrated into the growing metal coating. For example, for the fabrication of Cu-TiO2 nanocomposites, TiO2 nanoparticles were added into the copper sulfate and sulfuric acid plating bath with an optimized composition.

The solution were under continuous agitation for 4-6 hrs to ensure a uniform dispersion of the nanoparticles in the plating bath. Electrodeposited composite coating were carried out using a copper bar and polished copper plates of high purity as the anode. Usually Gelatin was also added into the plating bath as an additive to improve the surface quality of electrodeposited layer [39].

Friction stir processing (FSP) were capable to fabricate surface metal matrix nanocomposites (SMMNC) [64,94,95,96]. FSP is one of the most capable processes for severe plastic deformation (SPD). FSP is able to modify the microstructure of the surface material

[97,98,99] and to fabricate SMMNCs [94,95,96]. FSP can achieve improvement of surface mechanical properties, enhancement of wear resistance, and microstructure refinement. The FSP method includes a high-strength rotating tool with a pin and a shoulder. The tool inserts into the surface of a flat sheet and stirs the material around the pin under a frictional heating, which is generated through severe plastic deformations. The nanoparticles or clusters were normally pre- incorporated into the matrix, or embedded in a groove inside the matrix metal. The nanoparticles were stirred and flowed with the matrix materials in the process zone [100]. The key parameters of FSP are the tool geometry, rotating and traverse speed, target depth and spindle tilt angle, which have significant effects on the microstructure and mechanical properties of the products. FSP has been successfully applied to fabricate Al and Mg based SMMNCs with various reinforcements, such as carbide particles such as TiC [97] and SiC [94,96], carbon nanotubes [98], and Al2O3[99], etc.

Squeeze casting, a liquid-state method, was experimented for MMNC fabrication. The fabrication process include the following steps: (1) fabrication and pre-heating of nanoparticles preform; (2) melting of alloy; (3) infiltration of the molten alloy into nanoparticles preform under pressure; (4) solidification of nanocomposites [11,101,102,103]. For squeeze casting, the contact angle between the nanoparticle and the liquid alloy are required to be less than 90˚ to ensure wetting. The maximum volume fraction of nanoparticle inside metal matrix could be as high as

50%. Al-CNT [101], Al-CuO [102], A356 alloy-CNT nanocomposites [103] were fabricated through squeeze casting method.

Stir casting is another liquid-state method for composite materials fabrication, in which a dispersed phase (ceramic particles or short fibers) is mixed with a molten metal by mechanical stirring. Compared with solid-state methods, melt processing by stirring ceramic particles into melt

7 has some important advantages: a better matrix-particle bonding, easier control of matrix structure, simplicity, low cost processing, nearer net shape, and a wide selection of materials. Depending on the temperature at which the particles are introduced into the melt, there are two types of melting methods for making composites. In the liquid metallurgy process the particles are incorporated above the liquid temperature of the molten alloy, while in compocasting method the particles are incorporated into the semi-solid slurry. Al-MgO [23,24], Al-Al2O3 [77], A356-Al2O3 [78],

Al6061-Al2O3 [79], A356-SiC [80], and Al4.5wt%Cu-Al2O3 [81] nanocomposites were fabricated through stir casting.

Mechanical alloying, also referred as high energy ball milling or powder metallurgy in some cases, successfully produced a variety of nanocomposites [21, 22, 24, 25, 26, 27, 28, 30, 31,

34, 35, 36, 37, 42, 55, 65, 82,83, 104]. There are ex-situ powder metallurgy method and in -situ method. In the ex situ method, metal powders and nanoparticles are mixed together and charged into a hardened bowl with grinding balls inside. In the in-situ method, nanoparticles were produced by high energy ball milling through mechanochemical reaction [32,42,44]. Mechanical alloy treatment is usually protected under argon atmosphere and lasts for tens of hours. In some cases, in order to avoid excessive temperature rise with the grinding bowl, dozens of minutes ball milling was usually followed by an interval of dozens of minutes. The resulting powders are compacted and sintered together under pressure to obtain bulk parts for application [错误!未定义书签。,错

误!未定义书签。,104].

Laser cladding is an efficient route for fabricating metal matrix nanocomposite, whereby a new layer of materials is additively deposited onto the substrate by laser melting of blown powders or preplaced powder bed. Due to the high energy density and high spatial resolution and temporal resolution of laser beam, powder layers and the beneath substrate experience rapid heating and

8 rapid cooling, forming a cladded nanocomposite layer on the substrate. The laser cladded nanocomposite layer has a metallurgical bond with the substrate while maintains basically the original composition and properties of the added materials. A great variety of powders or powder mixtures can be effectively deposited onto the substrate to obtain various properties. Laser cladding is capable of locally tailoring the surface to achieve designed microstructure and mechanical properties, such as enhanced corrosion resistance, or wear resistance or solid lubricant properties [85,86,87,88,89,90,91].

In the squeeze casting and stir casting processes, the addition of nanoparticles to the melt leads to inhomogeneity of nanoparticle distribution and detrimental residual porosity

[23,24,75,76,77,78,79,80]. Electrodeposition, friction stir process and laser cladding method are only suitable for fabricating surface metal matrix nanocomposites [11,16,39,59,64,66,68]. Powder metallurgy method is not very suitable for large parts with complex shapes [105] and is extremely time consuming. Some efforts are made to combine all the above methods together as a compromise [106].

Solidification processing, a melt based processing technology for metals (e.g. casting), is a widely used economical method for mass production of bulk metal components with complex shapes. Tremendous efforts on fabrication of MMNCs through solidification processing have been made during the last decade [18,57,59,64,69,107,108,109,110,111,112,113]. Nanoparticles are incorporated into liquid metals or semi-solid metals and dispersed by mechanical stir or ultrasonic processing. Liquid metals mixed with nanoparticles are then cast into molds to obtain bulk

MMNCs.

Incorporation and de-agglomeration of nanoparticles in liquid metals are extremely difficult. Fortunately some effective means, such as ultrasonic dispersion, were developed to tackle

9 these problems. However, the interactions among nanoparticles inside liquid metals may redistribute nanoparticles after ultrasonic process stops. For example, the nanoparticles, well dispersed in molten metals during the ultrasonic process, may re-agglomerate to form clusters, which may be pushed to grain boundaries and phase boundaries during solidification. There is strong need to fully understand the physics of the interactions between nanoparticles inside metal melts in order to develop pathways to achieve the uniform dispersion of nanoparticles for mass manufacturing of bulk MMNCs.

2.2 Nanoparticle effects inside grains

2.2.1 Nucleation effect

Nanoparticles could refine grains through nucleation. Kim and Cantor developed an adsorption model on potent inoculants for nucleation [114]. The initial layer of atoms of the nucleus embryo can adsorb onto the inoculant surface with little or no undercooling. This initial layer of atoms then serves as the foundation on which the embryo grows. Quested and Greer termed this type of model as the free growth model and identified the necessary undercooling ∆푇, for free grain initiation

4휎 ∆푇 = 푙푠 , Equation 2-1 ∆푠푉푑 where d is the diameter of the initial adsorbed layer of atoms, commonly taken as the particle diameter, 휎푙푠 are the liquid-solid interfacial energy, ∆푠푉 is the entropy of fusion per unit volume

[115]. The free growth model was applied to commercial grain refiners where the particles size of grain refiners is at micrometer scale [116]. Examination of the model at the nanoparticles in

MMNCs were also conducted, which match reasonably well between the theoretical prediction and experimental results. Using an ultrasonic mixing technique, metal matrix nanocomposite 10 samples were prepared and examined for nucleation catalysis using the droplet emulsion technique.

Nanoparticles were proven to be effective as nucleation catalysis for grain refinement. Varying sizes of SiC, TiC and Al2O3 nanoparticles were shown to be potent nucleation catalysts, have reduced undercoolings compared to the undercooling for pure Al alloy A356, although the nucleation catalyst potency is related to the crystal structure of nanoparticles [117,118]. For example, Al2O3 and TiB2 nanoparticles were reported to refine the grain size of A206 alloy

[119,120].

2.2.2 Zener pinning effect

In an alloy consisting of a matrix and dispersed particles, grain growth is impeded by the interaction between grain boundaries and particles. If the grains start with a size smaller than the inter particle distance, they grow normally until there are enough particles interaction with the boundaries to pin the grains. According to Zener [121] in a poly-crystal with randomly distributed and dispersed particles of identical size, those particles located at the grain boundaries will decrease the boundary area and create a drag force to resist grain boundary migration. The drag force 퐹푧 is

퐹푧 = 푎푓/푑, Equation 2-2 where a is a constant close to 1, and d and f are the diameter and the volume fraction of particles respectively. Take a micro particle 5 µm and a nanoparticle 50 nm for comparison, with a same volume fraction, 푓, nanoparticles will apply 100 times higher drag force on the moving grain boundaries. The grain growth should cease when the mean grain size achieves a limiting value

〈퐷〉퐿 = 푏푑/푓, Equation 2-3

11 where b is a constant close to 4/3. In other words, the maximum diameter, 푑, of pinning particles in a volume fraction, 푓, required to stabilize grains of diameter, 퐷, [122, 123] is

3 푑 = 퐷푓 Equation 2-4 4

The growth suppression by dispersed particles was studied by computer simulations using a statistical model for grain growth. In those simulation, it is assumed that all the grain boundaries are of equal and constant energy and mobility. Nanoparticles with different shapes and size distribution are considered [124,125,126]. Zener pinning effect by nanoparticles limits grain size during cooling, recovery, recrystallization and grain growth to achieve finer grain size [126,127].

2.2.3 Orowan strengthening effect

Orowan strengthening, caused by the resistance of closely spaced hard particles to the movement of dislocations, is important in MMNCs. Highly-dispersed nanoparticles (smaller than

100 nm) in a metal matrix can considerably enhance its creep resistance, even with only a small volume fraction (1%) [128,129,130]. Orowan strengthening can be determined by the Orowan-

Ashby equation [131]

0.13퐺 푏 푟 ∆휎 = 푚 푙푛 , Equation 2-5 푂푟표푤푎푛 휆 푏

where r is the particle radius, 푟 = 푑푝⁄2, and 휆 is the interparticle spacing determined by 휆 ≈

1 1 푑푝 [( )3 − 1] [129,130]. This model was validated by experiments [128]. When the 2푉푝 nanoparticles size is as small as 5 nm, 1.0 vol.% of nanoparticles could increase the yield strength of Al by more than 100 MPa.

2.2.4 Other functional effects

Nanocomposites, especially Mg based, are identified as promising structural materials with high damping capacity [132]. Damping capacity is the measure of a material capability to dissipate elastic strain energy during mechanical vibrations or wave propagations. In MMNCs, damping could be improved through the additions of reinforcing phases that possess high intrinsic damping or that dramatically modify the grain nanostructure [133]. In nanocomposites, energy dissipation is mainly determined by local plastic deformation, resulting from the bonding at the nanoparticle- matrix interface.

Nanocomposites could improve the erosion and wear resistance of pure alloys since ceramic nanoparticles have higher chemical stability than alloys. Experimental showed SiO2,

Al2O3, SiC all could reinforced Ni-Based composite alloying layers to resist erosion in all hydrodynamic conditions [134,135,136]. And Cr2O3 reinforced Ni-Based nanocomposites were identified with the best erosion resistance [137].

Besides mechanical properties, nanocomposites could also provide exciting physical and chemical properties for numerous applications, such as magnetic [138] and electrical [139,140] materials. Co nanoparticles could provide significant giant magnetoresistance in Cu matrix, which offer great potential for novel read heads for hard disk recording and magnetic field sensors.

Carbon nanotubes reinforced copper and copper alloys have been fabricated by hot-press sintering and the electrical conductivity of composites has improved up to 20% compared to that of the pure alloy [38,140].

Nanocomposites are also considered to be promising biomedical materials with its tunable bio-corrosion properties [64,141]. Mg–hydroxyapatite (HA) nanocomposites were fabricated through casting and friction stir process. The corrosion measurement showed that the addition of

HA nanoparticles in Mg could improve the corrosion resistance by five times [64]. Magnesium- fluorapatite (FA) nanocomposite was made via a blending-pressing-sintering method and the corrosion measurement showed that the addition of FA nanoparticles in Mg can reduce the corrosion rate and increase the formation of bone like apatite layer and thus protect the matrix alloy [141].

2.3 Nanoparticle dispersion

2.3.1 De-agglomeration of nanoparticles

Nanoparticles are readily agglomerated in liquids due a reduction of surface energy.

Dispersing nanoparticles in liquids is essential for numerous applications including pigments for coating applications, ceramic materials, fillers in polymers, and tailored rheology of suspensions

[142]. Dispersion and de-aggregation of nanoparticles in liquids can be carried out in ball mills, rotor-stator high shear mixers, and ultrasonic processors. Agglomeration of nanoparticles can be broken by normal or shear stress by erosion, or by bulk rupture during the milling, during high shear mixing, and during ultrasonic processing [143]. The mechanism of the breakage depends on the size of agglomeration and the input energy intensity [144]. When the size of the agglomeration become smaller during de-agglomeration, surface forces dominate over mass forces. Thus, mechanical action such as mechanical stir or high shear mix could only break the large agglomeration above 1 μm, whereas the breakage of smaller agglomeration becomes very difficult.

It has been suggested that the agglomeration of nanoparticles smaller than 10 to 100 nm cannot be broken by mechanical action [145,146].

Ball milling for de-agglomeration of nanoparticles is a form of comminution. All applied stresses during the milling result in a three-dimensional stress state with compressive, tensile and

14 shear stresses which leads to fracture or to release of contact points. The stress intensity and the frequency are critical in a dispersion process. However, stressing during dispersing not only produces fractures, but also leads to agglomeration at the same time [147,148].The process of agglomeration can be hindered by different types of stabilization, which will be discussed later

(e.g. electrostatic, steric and electrostatic). Moreover, the stressing inside agglomerates does not only result in break up but also in the consolidation of the agglomerates due to reorientation of the particles and an increase in the contact area resulting from plastic or elastic deformation. The results achieved by ball milling from 10 nm to tens of μm. The following parameters can significantly alter the results:

 stress intensity and time,

 measurement in a highly diluted or a very concentrated system,

 sequence of dispersion and dilution,

 stability of the suspension,

 steric or electrostatic double layers, dynamics of the change,

 influence of impurities.

The smallest possible particles that can be produced in the comminution process [149] is in the range 1–10 nm. However, the majority of the agglomeration are still in the micrometer scale.

Uniformly dispersed nanoparticles have not been achieved through the ball milling method [142].

The kinetic of de-agglomeration in a high shear mixer is relatively easy to control and such devices are frequently used in industry to produce different types of suspensions. De- agglomeration of nanoparticles by a high shear mixer can be divided into two steps: low energy input step (3000 to 500 rpm for 12 nm silica nanoparticles in water as an example) and high energy input step. In the first step, large agglomeration (up to hundreds of μm) are broken into a smaller,

15 secondary aggregates with a size between 1 to 100 μm. A single modal volume distribution functions could be achieved through this step. In the second step, at a sufficiently high energy input (above 8000 rpm for 12 nm silica nanoparticles in water as an example), large secondary aggregates start breaking into a small primary aggregates. A bimodal distributions functions are often achieved, with the first mode approximately ranging from tens of nm to 1 μm and the second mode similar to the secondary aggregates in the first step. A higher energy input and longer processing time could gradually reduce the volume fractions of nanoparticles in the secondary aggregates in this step. However, the nanoparticles in the primary aggregates could not be broken into single nanoparticles. The size limitation of high energy mixer is due to the energy density during the process (~ 1500 kJ kg-1) is not strong enough to overcome the bonding force between nanoparticles in agglomerations [150].

The de-agglomeration of nanoparticles by ultrasonic process is extensively studied

[151,152,153,154,155]. High-intensity ultrasonic waves generate some important non-linear effects in liquids, namely transient cavitation and acoustic streaming, which are mostly responsible for dispersive effects for homogenizing. On one hand, cyclic high intensity ultrasound wave is launched into the liquid particle-matrix-mixture, where cavitation bubbles can form, grow, pulsate during several cycles until they attain a critical diameter before their implosively collapse (in less than 10−6 s). Thousands of micro bubbles will form, expanding during the negative pressure cycles and collapsing during the positive pressure cycles. The collapse causes extreme conditions such as very high local pressure and very high temperatures in so called hot-spots. During a cavitation collapse within a few microseconds, intense heating of the bubbles occurs. These localized hot spots could reach temperatures of roughly 5000°C, pressures of more than 1000 atmospheres, and heating and cooling rates above 1010 K/S [156]. The implosive impact from a cavitation collapse

16 is sufficient to break up the clustered particles. Therefore, a splitting up of nanoparticle agglomerates can occur to disperse nanoparticles more uniformly in liquids [157]. On the other hand, acoustic streaming, a liquid flow due to acoustic pressure gradient, is very effective for stirring [158,159]. In short, the implosive bubble collapse in combination with acoustic streaming generated by cyclic high intensity ultrasonic oscillations leads to dispersion effects. The schematic of the cavitation and streaming effects for nanoparticle dispersion is shown in Figure 2-1.

Figure 2-1. Schematic of the cavitation and streaming effects for nanoparticles dispersion

in liquid metal by ultrasonic waves [159]

2.3.2 Stabilization of dispersed nanoparticles in liquids

2.3.2.1 DLVO theory

Derjaguin, Landau, Vervey, and Overbeek (DLVO) developed a theory of colloidal stability based on interactions between colloidal particles and their aggregation behavior.

Nowadays for colloid science the theory of DLVO stands at the same level of significance as does

Darwin's theory of the origin of species in biology [160]. This theory is being used to rationalize forces between interfaces and between thin liquid films. The same theory is also applied to interpret

17 particle deposition to planar substrates. Here a short summary of the relevant concepts will be discussed while a more detailed treatment can be found elsewhere [161].

DLVO theory assumes that between two colloidal particles with a surface separation, ℎ, the free energy per unit area between two surfaces, 푊 (ℎ), can be approximated by two additive contributions, van der Waals, 푊푣푑푤(ℎ), and double layer interactions, 푊푑푙(ℎ), i.e.

푊(ℎ) = 푊푣푑푤(ℎ) + 푊푑푙(ℎ) Equation 2-6

Van der Waals forces result from interactions of the rotating or fluctuating dipoles of atoms and molecules. They are always present between two surfaces. In the simplest situation, this interaction can be calculated as

퐴 푊 (ℎ) = − J m-2 Equation 2-7 푣푑푤 12휋ℎ2 where A is the Hamaker constant, which defines its strength. The van der Waals force between two identical bodies in a medium is always attractive, while between two different bodies in a medium it can be attractive or repulsive.

Double layer interactions come from the surface charge of particles in the liquid. Particles in water or any liquid with a high dielectric constant (polarizing medium) are usually charged. The charging of a particle surface in a liquid could be attributed to three sources: first, by ionization or dissociation of surface groups; second, by the adsorption or binding of ions from the solution on a previously uncharged surface; third, protons or electrons hop across from one surface to the other through charge exchange mechanism between two dissimilar surfaces very close together. The surface charge is balanced by an equal but oppositely charged region of counterions. Some of the counterions bound to the surface within the so-called Stern layer. Others form an atmosphere of ions in rapid Brownian motion close to the surface, forming a diffuse electric double layer. Two

18 similar charged surfaces usually repel each other due to the repulsive coulomb electrostatic force of the double layer. For two planar surfaces, the interaction per unit area is

2 −휅퐷 2 −휅퐷 -2 푊푑푙 ≈ 2휀0휀휅휓0푒 = 2휎 푒 /휅휀0휀 J m Equation 2-8

where 휀0 is the vacuum dielectric constant, 휀 is the dielectric constant of the liquid, 휅 is the Debye length, 휓0 is the surface potential, 퐷 is the distance between two surfaces.

The total interaction between any two surfaces include the van der Waals attraction and double layer repulsion. Van der Waals attraction is a power law interaction (i.e., W ∝ −1/퐷푛, whereas the double layer interaction energy remains finite or rises much more slowly as D → 0.

Therefore, the van der Waals attraction always exceed the double layer repulsion at small enough distance. Figure 2-2 shows schematically the various types of interaction potentials that could occur between two similarly charged surfaces or colloidal particles in a 1:1 electrolyte solution under the combined action of these two forces. Depending on the surface charge density or potential and the electrolyte concentration, the following potential profile may occur:

For highly charged surfaces in dilute electrolyte, a strong long range repulsion that peaks at some distance (between 1-5 nm) dominates the interaction. The peak is referred as energy barrier, which is often many kT.

Figure 2-2. Schematic energy versus distance profiles of the DLVO interaction. The actual magnitude of the energy W is proportional to the particle size (radius) or interaction area

(between two planar surfaces) [161]

For highly charged surfaces in more concentrated electrolyte, beyond the distance of the energy barrier, there is significant secondary minimum. The potential energy minimum at contact is known as the primary minimum. For a colloidal system, though the thermodynamically equilibrium state is when particles in contact in the deep primary minimum, the energy barrier may be too high for the particles to overcome within a reasonable time period. In this case, when the secondary minimum is strong, particles will be trapped in the secondary minimum. When the secondary minimum is weak, particles will remain dispersed in the solution. The latter case the colloid is referred as being kinetically stable.

For low charged surfaces, the energy barrier will be low. The colloidal particles will coagulate slowly or rapidly, depending on the height of the energy barrier. As the surface charge approaches zeros, the interaction approaches the pure attractive van der Waals interaction. Two particles attract each other at all separations and are readily to approach each other and reach the primary minimum point. At the primary minimum point, two surfaces come into adhesive contact is driven by the reduction of their surface charge or potential.

2.3.2.2 Non-DLVO forces

Other than attractive van der Waals and repulsive double-layer forces, other non-DLVO forces come into play when surfaces or particles approach closer than a few nanometer. These additional forces can be much stronger than the two DLVO forces at small separation, and they could be repulsive, attractive, or oscillatory. Among those forces, there are solvation force, hydration force, steric and thermal fluctuation forces.

Solvation forces could be induced when spherical liquid molecules are forced to order into quasi-discrete layers within any highly restricted space between two surfaces, including two opposite surfaces between two colloidal particles. Due to the geometric origin, solvation force is usually oscillatory and the oscillatory period critically depends on the shapes of the liquid molecules. And the strength of solvation forces depend not only on the properties of the liquid molecules, but also on the chemical and physical properties of the surfaces. These factors, for example, those surfaces are the hydrophilic or hydrophobic, smooth or rough, amorphous or crystalline, homogeneous or heterogeneous, natural or patterned, rigid or fluid-like, will affect the solvation force, since they could affect the structure that confined liquid between those two surfaces. On the other hand, those factors also have tremendous effects on the surface forces, which

21 make it difficult to distinguish between a solvation force and a surface force. Solvation forces could be very strong at short range in some cases. And the magnitude of the adhesion between two surface, or particles in contact, or at their potential energy minimum, are determined by solvation forces. The most general type of solvation force is the oscillatory force.

Theoretical work and particularly computer simulations indicate that at solid-liquid interface, liquid density oscillation occurs. The attractive interactions between the wall and liquid molecules, as well as the geometric constraining effect of the “hard wall” on these molecules force the liquid molecules to order (or “structure”) into quasi-discrete layers, as shown in Figure 2-3b.

These oscillatory density profile extends several molecular diameters into the liquid, which is referred as out-of-plane ordering.

If the surface is smooth, the liquid molecules within each layer will be random. However, if the surface itself is structured (crystalline lattice), there will also be ordering within these layers to some extent, which is referred as in plane ordering. Using x-ray reflectivity techniques, four water layers adjacent to mica surface, with a periodicity of 2.6 ± 0.1Å, were identified [162]. Both liquid metals [163] and van der Waals liquids [164,165] were found to have similar density oscillations. Solvation forces are expected to arise whenever packing constraints limit the ordering of particles (atoms or molecules) in a confined geometry, and they have been found to exist in concentrated solutions of hard silica particles, soft spherical micelles, and lipid bilayers.

In the case of liquid between two solid surfaces, even in the absence of any attractive or structured solid-liquid interaction, the liquid molecules must reorder themselves so as to be accommodated between two surfaces, and the variation of this ordering with separation D gives rise to the solvation force between the confining surfaces. The simple case of simple spherical molecules between two hard, smooth surfaces, the structure of the liquid molecules, and the density

22 are shown in Figure 2-3c. The solvation forces is usually a decaying oscillatory (but not sinusoidal) function of distance.

Figure 2-3. (b) Liquid density profile at an isolated solid-liquid interface. (c) Liquid density

profile between two hard surfaces with a distance D apart. 흆푺(∞) is the liquid density at

the surface, and 흆풎(∞) is the bulk liquid density, 흆푺(푫) and 흆풎(푫) are liquid density at

the surface and the middle plane, which will be a function of D [161]

The main features of the oscillatory forces are summarized in details by Israelachvili [161].

The periodicity of the oscillatory force, the peak-to-peak amplitudes of the oscillations, the magnitude forces complicated depend on various factors, such as the structure of the liquid molecules, temperature, surface lattice structure and roughness, surface hardness or fluidity, surface curvature and geometry. As summarized by Israelachvili, modern theories and simulations of liquids have only shown how very complex structures and force-laws can arise across an

23 assembly of molecules that interact even via the simplest possible pair potential, both quantitatively and qualitatively.

Solvation forces in aqueous system sometimes referred as hydration forces. The short range hydration forces between surfaces in water and aqueous solutions displays some highly unusual properties. The hydration forces can be strongly monotonically repulsive, attractive, oscillatory, or a combination of these. These forces are believed to due to water structuring at surfaces gained steady popularity and are often observed exponential decay length of around 2.5 Å. It is believed to be due to some characteristic property of water such as its molecular size. However, theoretical work and computer modeling failed to confirm the relationship. There are two types of such forces, strong steric-hydration forces between fluid-like amphiphilic surfaces, and monotonically repulsive hydration forces between solid hydrophilic surfaces. The former force is due to the presence of strongly hydrophilic groups on the surfaces. Such forces are more akin to the steric and thermal fluctuation forces between two polymer covered interfaces. The latter hydration force between two hydrophilic surfaces appears to follow the simple equation

−퐷/휆0 푊(퐷) = 푊0푒 Equation 2-9 where 휆0 = 0.6 − 1.1 nm for 1:1 electrolytes, and 푊0depends on the hydration of the surfaces but

-2 is usually below 3-30 mJ m , higher 푊0 values generally being associated with lower 휆0 values.

Israelachvili and Pashley conducted a series of experiments to identify the factors that determine and regulate these hydration forces. The electrolyte concentration, the types of the cations and anions in the aqueous solutions influence the profile and the amplitude of these hydration forces.

Thus, the repulsive hydration forces can be modified or “regulated” by exchanging ions of different hydrations on surface in colloidal dispersion.

Although extensive study have been conducted, so far neither experiment nor theory has yet revealed the full nature of short-range monotonically repulsive solvation forces, including hydration forces. The solvation forces will remain controversial for more time.

Solvation forces and hydration forces are between smooth and rigid interfaces. There are many instances where interfaces are spatially diffuse ad where the forces between them depend on the diffuse boundaries overlap. There are two common types of such diffuse interfaces. First type is the interface is inherently mobile or fluid-like, as occurs at liquid-liquid, liquid-vapor, and some amphiphilic-water interfaces. Molecular-scale thermal fluctuations or protrusions exists between these surfaces and affect the forces between them. Such forces are entropic or osmotic in origin and are referred to as thermal fluctuation forces. The second type of a thermally diffusion interface occurs between chain molecules attached at some point to a surface of colloidal particles. On approach of another similar surface, these chains result in a repulsive entropic force and are referred as steric forces.

2.3.3 Challenges of nanoparticles dispersion in molten metal

Ultrasonic process for de-agglomeration of nanoparticles are effective in molten metal for solidification nanoprocessing of MMNCs. However, the stabilization of the dispersed nanoparticles in molten metal after the ultrasonic process is extremely challenging. When ultrasonic process stops, the uniform distribution of nanoparticles could not be maintained and are further pushed to grain boundary during solidification.

The common strategies for stabilization of nanoparticles are not effective in molten metals.

For example, ceramic nanoparticles in molten metals have rigid and well-defined surfaces.

Thermal fluctuation forces could thus be ruled out from the system. Meanwhile, the ambiguous

25 characteristic conditions to set up reliable solvation forces is not realistic for industry scale control.

Moreover, in molten metals, especially lightweight Al and Mg, the temperature is around 1000 K.

Chain organic molecules responsible for steric forces are not stable at such high temperature.

Moreover, molten metals are highly conductive, resulting in the failure of repulsive forces based on electrostatic interactions. Therefore, it is fundamentally difficult to disperse nanoparticles in molten metal for mass production of MMNCs based on existing theories and experimental work.

It is thus very important to initiate a rigorous theoretical study to discover pathways for experimental investigations for effectively dispersion of nanoparticles in solidification processing for mass production of MMNCs.

Chapter 3 In-situ oxidation method to fabricate Al nanocomposites with dispersed Al2O3

nanoparticles

3.1 Introduction

The mechanical properties of Al alloys can be effectively enhanced with an addition of a low volume fraction of nano-sized reinforcements for numerous applications [166,167,168,169].

Liquid processing would be economical to fabricate bulk Al matrix nanocomposites. However, it remains a great challenge due to the poor wettability between ceramic nanoparticles and liquid Al and also the difficulty to capture nanoparticles into Al matrix during solidification processing

[107,120]. Thus, Al matrix nanocomposites with ex-situ nano-sized reinforcements were mainly fabricated through solid state mechanical alloying and consolidation, which generally avoid the wetting and capture problems encountered in solidification processing [ 错误! 未定义书

签。,170,171]. However, mechanical alloying is a time consuming process and the geometry and sizes of the parts fabricated through the consolidation are limited. In situ synthesized nanocomposites are defined as multiphase materials where the nano-reinforcements are synthesized within the matrix during fabrication [172,173,174]. Due to the high reactivity and morphological diversity, nanomaterials are frequently considered as basic building blocks for the fabrication of in-situ nanocomposties through bottom-up approach [175,176]. For Al nanoparticles, a native oxygen-rich thin film of about 2-3 nm thick usually covers their surfaces [177], which could serve as the oxide nano-reinforcement if the Al nanoparticles are to be melted together. Thin

Al2O3 films could be re-distributed as dispersoids into the bulk Al material after melting of the Al matrix, thus forming bulk Al matrix nanocomposites with evenly distributed Al2O3 nano- reinforcements.

In this work, the feasibility to fabricate bulk Al matrix nanocomposites with uniformly dispersed Al2O3 nanoparticles through in-situ oxidation and melting of Al nanoparticles is investigated. Al nanoparticles were first cold compressed into bulk pellets at room temperature and then melted at 700 ˚C under a protective atmosphere. During the process, thin Al2O3 films were further formed in situ and then re-distributed into the Al matrix. The experimental results show that the microhardness of the bulk Al matrix nanocomposites is significantly enhanced to about three times as high as that of pure Al. This study paves the way to fabricate bulk Al matrix nanocomposites through in situ oxidation under liquid state processing.

3.2 Materials and methods

Spherical Al nanoparticles with an average diameter of 40 nm and a purity higher than

99.9% (from US Research Nanomaterials Inc.) were used as the starting material. The composition of the Al nanoparticles is given in Table 3-1. The X-ray diffraction (XRD) pattern of the initial Al nanoparticle is shown in Figure 3-1. About 300 mg Al nanoparticles were initially filled into a stainless steel mold and cold compressed under 500 MPa for 5 minutes at room temperature. Figure

3-2 illustrates a schematic of the compression experimental setup. The inner diameter of the mold is 6 mm. As shown in the Figure 3-2, inside the mold, nanoparticles were cold compressed between upper pestle and base. After unloading, the compressed pellet was removed from the mold, then heated to and held in an alumina container at 700 ˚C for 1 hr in a chemical vapor deposition (CVD) quartz tube under a protective atmosphere of Ar/H2 (1:1). The final samples were obtained after the sample cooled in the air down to the room temperature.

The samples were then cut, mounted, and manually ground. The microstructure of the samples was studied by XRD and Field Emission Scanning Electron Microscopy (FESEM, LEO

1530, Carl Zeiss). Ion milling was used to reveal the microstructure of Al2O3 in the Al matrix.

XRD was conducted using a Bruker D8 Advance Powder diffractometer with Cu Kα radiation.

Microhardness tests were conducted with a Wilson Tukon 1102 Knoop/Vickers Hardness tester

(with a load of 200 N and a dwelling time of 10 s).

Table 3-1. Aluminum nanoparticles certificate of analysis (unit: ppm).

Si Fe Mg Mn Pb Cr Ni Cu Ti Zn Ca Sn V Al

≤65 ≤110 ≤25 ≤10 ≤12 ≤5 ≤34 ≤280 ≤7 ≤10 ≤5 ≤5 ≤5 ≥99.9%

Figure 3-1. XRD pattern for initial Al nanoparticles and Al nanocomposites

Figure 3-2. Schematic of the cold compression experimental setup (unit: mm)

3.3 Results and discussions

3.3.1 Microhardness

The average microhardness of the Al nanocomposite samples was 103 Hv with a standard deviation of 6.4. Ten microhardness readings were taken from each sample. Compared with the microhardness of 33 Hv for the pure bulk Al [171,178], the average microhardness of the nanocomposites was enhanced by more than two times. When compared with a 92% increase in the hardness of Al matrix Al2O3 nanocomposites fabricated through high-energy ball milling, our new method has obvious advantages [179,180]. The significant enhancement in microhardness in the Al nanocomposite samples could be largely attributed to two factors: grain refinement [181,182] and Al2O3 nanoparticle reinforcement [128,129].

3.3.2 Composition Analysis by XRD

The phase compositions of the Al nanoparticles and the nanocomposites were determined by XRD, as shown in Figure 3-1. Only pure Al peaks were shown for both the starting Al nanoparticles and the resultant Al nanocomposites. It is also difficult for XRD to detect the low content of nano-sized Al2O3 film or nanoparticles.

3.3.3 Microstructure Analysis

The microstructure of the Al nanocomposites is shown in Figure 3-3. The SEM images show two different domains in the Al matrix: one without nano-sized dispersoids and the other with a high concentration of evenly distributed Al2O3 dispersoids. Energy-dispersive X-ray spectroscopy (EDS) mapping (Figure 3-4) confirmed that the nano-sized dispersoids are Al2O3.

Thus it suggests that the nanocomposite consists of nano Al2O3-reinforced Al grains and some coarse Al grains (Figure 3-3b). The nanoscale Al2O3 could come from the Al2O3 impurity on the initial Al nanoparticles and later oxidation of Al nanoparticles during the heating and melting process. On the other hand, the coarse grains may come from the coarse Al particles presented in the initial Al nanoparticles and/or coalesce of Al nano-droplets in the liquid state. Figure 3-3c shows some Al2O3 nanoparticles at the boundary of one coarse Al grain, which helped hinder the growth of the Al grains. Moreover, Al2O3 nanoparticles were well distributed inside the ultrafine

Al grains, as shown in Figure 3-3d.

Figure 3-3. Microstructures revealed by SEM, a) Al2O3 reinforced Al nanocomposite:

ultrafine Al2O3 reinforced Al matrix region and some embedded Al coarse grains; b) Al coarse grains; c) Al2O3 nanoparticles along the Al coarse grain; d) Al2O3 distribution in the

Al matrix

Figure 3-4. EDS mapping of the Al nanocomposites

3.4 Summary

In this work, pure Al nanoparticles were cold compressed in a steel mold and then melted in an alumina container under a protective under Ar/H2 (1:1) atmosphere. Al2O3 nanoparticles were in situ synthesized through oxidation of Al nanoparticle surfaces to form bulk Al nanocomposites during the process. Although some Al2O3 nanoparticles were distributed along the grain boundaries of some coarse Al grains, most Al2O3 nanoparticles were evenly distributed inside the ultrafine Al grains to effectively restrict their grain growth. Moreover, the microhardness of the bulk Al nanocomposites is enhanced up to about three times as high as that of pure bulk Al. This study provides a new and economical study to fabricate bulk Al matrix nanocomposites.

Chapter 4 Dispersion of nanoparticles through friction stir processing

4.1 Introduction

Two major routes for MMNCs processing have been established – solid-state processing

(e.g. powder metallurgy) and liquid state solidification processing. In the powder metallurgy route, high shear stress generated by high-energy ball milling is applied to enforce nanoparticles into matrix metal and shear nanoparticle clusters. However, the low nanoparticles loading (less than 2 vol.%) for effective dispersion limited this method for high performance Mg nanocomposites

[183,184]. And the strict conditions to avoid oxidations during high-energy ball milling also make this method not economic for scalable processing of Mg nanocomposites [185,186]. In the liquid- state solidification processing, firstly, non-oxide nanoparticles are easy to burn during their feeding at high temperature (above the liquidus temperature of the metal); secondly, nanoparticles form clusters in molten metal due to the attractive van der Waals force and the lack of a repulsive force; thirdly, nanoparticles or clusters tend to float or settle without effective incorporation into the molten metal due to a poor wettability with the molten metal; and fourthly, nanoparticles are normally pushed to the grain boundary by solidification fronts during solidification [107].

Recently, friction stir processing (FSP) was used to disperse particles for metal matrix composites in several studies, especially in Al and Mg based MMCs and MMNCs

[108,187,188,189,190,191,192]. The dispersion of the particles and the interfacial bonding between the particles and the Al and Mg matrix after friction stir processing were found to be very good [100,193,194,195]. The trajectory of particles moving during friction stir processing were revealed in previous studies [195]. The significant particle movement induced by friction stir processing provides a powerful tool for dispersion of particles. Compared with the liquid-state

34 solidification processing, high viscosity of matrix metal during friction stir processing effectively hindered clusters formation, floating or settling of particles in molten metal; moreover, the solid state friction stir processing avoids the pushing of particles to the grain boundary by solidification fronts during solidification.

Mg-based metal matrix nanocomposites (MMNCs) are expected to provide significant enhancement of properties by introducing thermally stable ceramic nanoparticles into the metal matrix. However, it is extremely difficult to efficiently incorporate a large amount of nanoparticles into metals and to achieve a uniform distribution of nanoparticles in the magnesium matrix for the predicted significant property enhancement [59,195,196,197].

In this chapter, high loading of SiC nanoparticles were first incorporated into Mg-6Zn matrix and form micro clusters through semi-solid mixing. And a high loading of HA nanoparticles were incorporated into Mg matrix and form micro clusters through mechanical stir. FSP was then applied to effectively disperse the nanoparticles from these micro clusters of SiC and HA nanoparticles in Mg matrix.

4.2 Friction stir processing of Mg6Zn-SiC nanocomposites

4.2.1 Materials and methods

Mg6Zn with 6 vol.% SiC nanoparticles were first fabricated by semi-solid processing. The schematic of the experimental setup is in Figure 4-1. Pure Mg (99.93%) of 365.7 g and pure Zn of

21.94 g were melt together at 700 °C in a furnace under a SF6 (99%)/CO2 (1%) flow gas protection.

The metal melt were then set at 600 °C to the semi-solid state. Mechanical stir with a four blades

(25.4 mm in diameter) at 1400 rpm were applied at the 1/3 height of the liquid melt. The vortex were generated by the rotating blade in a SiC crucible with an inner diameter of 70 mm. Then 41.3

35 g SiC nanoparticles were loaded and stirred into the melt with the help of the high viscosity of the semi-solid melt. The melt were then cooled to room temperature in the air. The solidified nanocomposites were re-melted inside a plate mold inside the furnace at 700 °C under the same gas flow protection. The nanocomposites were solidified to plate shape and cooled to room temperature inside the mold in the furnace.

The nanocomposites plates were then friction stir processed with a H13 steel pin. The pin length is 6 mm and the pin diameter is 7.8 mm with a shoulder diameter of 25.5 mm. The rotational speed is 1500 rpm. The traverse speed is 25.4 mm/min. The tilt angle is 3°. The materials were processed for four passes back and forth. Pure Mg6Zn was processed under same conditions for comparison.

The friction stir processed nanocomposites were wire cut into ASTM type V tensile bars.

The tensile tests were conducted on the INSTRON 5966 tensile and compression testers. Vickers hardness tests were conducted using a 200 gf load for 10 s. The sample was grinded and ion milled to expose the microstructure and embedded nanoparticles. The microstructure of the samples was studied with powder X-ray diffraction (XRD) and Field Emission Scanning Electron Microscopy

(SEM) (Navo Nano 230) with energy dispersive spectrometry (EDS). XRD was conducted using a Bruker D8 Advance Powder diffractometer with Cu Kα radiation.

Figure 4-1. Schematic experimental setup: (a) semi-solid processing (b) friction stir

processing

4.2.2 Results and discussion

4.2.2.1 XRD of the nanocomposites

The XRD of the nanocomposites after FSP were shown in Figure 4-2. Pure Mg, SiC and

MgZn phases are identified in this sample. No reaction products from Mg6Zn and SiC or oxidation of nanoparticles were identified. The XRD pattern indicates the stability of both matrix materials and nanoparticles through the whole procedure.

Figure 4-2. XRD pattern of the nanocomposites: Mg, SiC and MgZn were identified in this

sample

4.2.2.2 Microstructures

The microstructure of the nanocomposites after semi-solid mixing, but before FSP, were checked by optical microscope, as shown in Figure 4-3. Micro clusters of SiC nanoparticles are

37 distributed in Mg6Zn matrix. The micro clusters indicate that high loading of SiC nanoparticles were successfully loaded into Mg6Zn at semi-solid state with the help of mechanical stirring.

The microstructure of both pure Mg6Zn and nanocomposites after FSP revealed by SEM were shown in Figure 4-4. For pure Mg6Zn in Figure 4-4a and b, the surface is clean with ripples induced by ion milling. For nanocomposites in Figure 4-4c and d, the surface is quite rough and

SiC nanoparticles (dark spots as represented by arrow points) were well dispersed and distributed inside Mg6Zn matrix. No clusters were identified. The surface ripple of Mg6Zn were believed to be introduced by the different milling rate of well dispersed SiC nanoparticles and Mg6Zn matrix during ion milling. And the dispersion SiC nanoparticles were from the disaggregation of micro clusters in Figure 4-3 by the shear force during friction stir.

Figure 4-3. Micro clusters of SiC in Mg6Zn after semi-solid nanoparticle incorporation

before FSP

Figure 4-4. Microstructures after FSP: (a)(b)SEM images of pure Mg6Zn after FSP; (c)(d)

SEM images of dispersed SiC nanoparticles in Mg6Zn after FSP (dark spots as represented

by arrows points are SiC nanoparticles)

4.2.2.3 Mechanical properties

The microhardness of the nanocomposites before and after FSP is shown in Figure 4-5.

The hardness (Vickers Hardness) of Mg6Zn + 6 vol.% SiC nanoparticles before FSP is 38 Hv.

After FSP, the microhardness is 107 Hv, increased by 187%. Also compared with the hardness of pure Mg6Zn after FSP, the microhardness of nanocomposites after FSP has an 84% increase. The microhardness indicates FSP dispersed SiC nanoparticles in the matrix, leading to the tremendous increase of the microhardness.

Figure 4-5. Microhardness of different samples

The tensile test results of Mg6Zn + 6 vol.% SiC (data collected from five tensile) were shown in Table 4-1. The average yield strength and the ultimate tensile strengthen are 265.6 MPa and 303.7 MPa, respectively. The average elongation is about 2.11%.

Table 4-1. Tensile test results.

Yield strength (MPa) Ultimate tensile strength (MPa) Elongation

265.6 303.7 2.11%

4.2.2.4 Discussion

Master nanocomposites with a high percentage loading of nanoparticles are achieved through the efficient incorporation at semi-solid state. At semi-solid state, the viscosity of the semi- solid metal depends on the volume fraction of solidified metal, which depends on the temperature.

A suitable viscosity can be found to prevent the nanoparticles from floating or settling by tuning the fraction of solidified metal through temperature control. At a suitable temperature, the nanoparticles were stirred into and maintained inside the slurry. Also the low temperature at semi- solid state reduce the reaction and oxidation of nanoparticles, enhancing the quality of incorporated

40 nanoparticle. XRD pattern indicates there is no detectable reaction products and oxidation. The efficient incorporation at semi-solid state provide an economic method for high loading of nanoparticle incorporation. When comparing with liquid state solidification processing, the incorporation at semi-solid state improves the quality of incorporated nanoparticles.

The strengthening effect of nanoparticles in the nanocomposites is characterized by the

Orowan equation, since the second phases are ceramics, thus are all non-shearable by dislocations.

The strengthening equation is

0.13Gbm d p  Orowan ln , Equation 4-1 1 1/3 2b d p[( ) 1] 2Vp

where Gm is the shear modulus of about 16.5 GPa and b (1/ 3)[2110] 0.32nm is the Burger’s vector

at room temperature, d p is the diameter of nanoparticles and Vp is the volume fraction of nanoparticles. Theoretical Orowan strength contributed by 6 vol.% SiC in Mg6Zn matrix is 50.6

MPa. It corresponds to 20.3 increase in microhardness [128,129].

Compared with pure Mg6Zn after FSP, Mg6Zn nanocomposites are believed to benefit from the refined microstructure of matrix materials by Zener pinning effect in addition to the

Orowan strengthening contribution. The dispersed nanoparticles will apply pining pressures on the moving grain boundaries to restrict the grain growth during friction stir to gain more refined grains.

4.2.3 Summary

In this work, we combined the semi-solid mixing for effective SiC nanoparticle incorporation and friction stir processing to achieve high performance Mg6Zn+6vol.%SiC nanocomposites. The low temperature and high viscosity at semi-solid state could effectively

41 improve the nanoparticles loading and reduce the oxidation/reaction. Moreover, friction stir processing effectively breaks micro clusters of incorporated SiC nanoparticles to achieve a uniform dispersion and distribution inside Mg6Zn matrix. The microhardness and tensile tests were conducted and the mechanical properties of Mg6Zn + 6 vol.% SiC nanocomposites are significantly enhanced. This method provides an economic pathway for processing of high performance Mg nanocomposites.

4.3 Friction stir process to disperse HA nanoparticles in Mg

4.3.1 Materials and methods

Mg with 6 vol.% HA nanocomposites were fabricated by mechanical stir process. The schematic of the experimental setup is similar as it in Figure 4-1. Pure Mg (99.93%) of 300.0 g was melt at 660 °C in a furnace under SF6/CO2 flow gas protection. Mechanical stir with a four blades (25.4 mm in diameter) at 800 rpm were applied at the 1/3 height of the liquid melt. The vortex were generated by the rotating blade in a SiC crucible with an inner diameter of 70 mm.

35.6 g HA nanoparticles were loaded and stirred into and maintained inside the melt with the help of mechanical stir. The melt was then cooled to room temperature in the air. The solidified nanocomposites were re-melted inside a plate mold inside the furnace at 660 °C under the same gas flow protection. The nanocomposites were solidified to plate shape and cooled to room temperature inside the mold in the furnace.

Figure 4-6. Friction stir process of Mg-6vol% HA nanocomposites. a) Schematic of three

step process: the white lines denote the first pass, the red line denote the second pass, and

the black lines denote the third pass; b) Rectangular processed zone

The nanocomposites plates were friction stir processed with a H13 steel pin. The pin length is 5.5 mm and the pin tip diameter is 6.40 mm with a shoulder diameter 15.19 mm. The rotational speed is 600 rpm. The traverse speed is 25.4 mm/min. The tilt angle is 2.5°. The materials were processed by three steps, schematically shown in Figure 4-6. In Figure 4-6a, the white lines denote the first pass, the red lines denote the second pass, and the black lines denote the third pass. In the end of the process, there is a rectangular area which has been processed in both vertical and horizontal directions. The rectangular area marked with black rectangle in Figure 4-6b shows the stirred zone.

The sample was grinded and ion milled to expose the microstructure and embedded nanoparticles. The microstructure of the samples was studied with powder X-ray diffraction (XRD) and Field Emission Scanning Electron Microscopy (SEM) (Navo Nano 230) with energy dispersive spectrometry (EDS). XRD was conducted using a Bruker D8 Advance Powder diffractometer with Cu Kα radiation.

4.3.2 Results and discussion

4.3.2.1 XRD of the nanocomposites

The XRD of the nanocomposites after FSP were shown in Figure 4-7. Pure Mg, MgO and

HA phases are identified in this sample. Oxidation of Mg were identified in this sample, which is believed to form during the process. HA nanoparticles were identified to successfully survive through the process. The XRD pattern indicates the stability of HA nanoparticles through the whole procedure.

Figure 4-7. XRD pattern of the nanocomposites: Mg, MgO and HA (Ca5(PO4)3(OH)) were

identified in this sample

4.3.2.2 Microstructures

The microstructure of the nanocomposites before FSP and after FSP were checked by optical microscope, as shown in Figure 4-8. In Figure 4-8a, micro clusters of HA nanoparticles are distributed in Mg matrix. In Figure 4-8b, high density of HA nanoparticles in the cluster were

44 exposed (small white spots as represented, large white spots are identified as MgO, which is also exposed in XRD pattern in Figure 4-7). The micro clusters indicate that high loading of HA nanoparticles were successfully loaded into Mg with the help of mechanical stir.

Figure 4-8. Microstructure of Mg-6 vol.% HA nanocomposites before FSP

The microstructure of nanocomposites after FSP revealed by SEM were shown in Figure

4-9. HA nanoparticles (small white spots as represented, large white spots are identified as MgO, which is also exposed in XRD pattern in Figure 4-7) were well dispersed and distributed inside

Mg matrix after friction stir process. No clusters were identified. And the dispersion HA nanoparticles were from the disaggregation of micro clusters in Figure 4-8 by the shear force during friction stir.

Figure 4-9. Microstructure of FSPed Mg-6 vol.% HA nanocomposites

4.4 Summary

In this chapter, to fabricate Mg6Zn + 6 vol.% SiC nanocomposites, semi-solid mixing for effective incorporation of SiC nanoparticle into Mg6Zn matrix was applied before FSP. The low temperature at semi-solid state could reduce the nanoparticles burning and oxidation effectively, while a high viscosity of the metal at semi-solid state would trap nanoparticles inside the matrix metal. To fabricate Mg + 6vol.% HA nanocomposites, mechanical stir was utilized to effectively incorporate HA nanoparticles into Mg melt. The solidified samples with micro clusters of incorporated SiC nanoparticles and HA nanoparticles were then friction stir processed to achieve a uniform dispersion and distribution of nanoparticles inside the solid Mg matrix. The mechanical properties of Mg nanocomposites after FSP were significantly improved.

Chapter 5 Theoretical study on nanoparticle dispersion in molten metals

5.1 Introduction

Although the two methods in Chapters 3-4 were successfully used to achieve a uniform dispersion of nanoparticles in MMNCs, they are not suitable for mass manufacturing of MMNCs.

Meanwhile, solidification processing is very promising as a versatile mass manufacturing methods to produce bulk MMNC parts with complex geometry and high nanoparticle loading. Therefore, tremendous efforts have been made on solidification processing of MMNCs during the last decade.

Nanoparticles were normally incorporated into liquid or semi-solid metals and then distributed and dispersed by mechanical stir or ultrasonic processing before casting. Kinetic dispersions of nanoparticles inside liquid metals were achieved by ultrasonic processing. But the dispersion of nanoparticles in molten metal is not stable once the ultrasonic process stops. Without effective repulsive forces between them, similar nanoparticles readily form clusters due to the attractive van der Waals force in molten metals. It is thus very important to initiate a theoretical study to examine the physics that governs the interactions between the nanoparticles in molten metals. Under the theoretical guidance, effective pathways could be discovered to disperse and stabilize nanoparticles in molten metals for large scale solidification processing.

5.2 Interactions between nanoparticles in molten metals

In a model nanoparticle-melt system, nanoparticles are assumed to be homogeneously distributed and dispersed in a static metal melt. The nanoparticle-melt system can be considered as a melt with a dilute suspension of nanoparticles when the nanoparticle loading is low. To simply the model, only the interactions between two same nanoparticles in a liquid metal are considered.

More assumptions are made to build a basic model as shown in Figure 5-1: 47

-Nanoparticles with a radius of R are spherical and D is the gap between two nanoparticles

-No macroscopic convection in the melt

-Negligible electrostatic interaction or double layers between the nanoparticle and metal melt

-Negligible buoyance and gravity forces for nanoparticles

-No severe chemical reaction between the nanoparticles and liquid metal

-No gas film or contamination on the nanoparticles’ surfaces

Figure 5-1． Model for two same nanoparticles interacting in a liquid metal

Based on the above assumptions, three major interactions, interfacial energy, van der Waals potential, and Brownian potential, are considered in this model nanoparticle-melt system:

5.2.1 Interfacial energy

When two nanoparticles interact in a liquid metal far from each other, the interfacial energy

can be described as GSb 2 pl pl , where Spl is the surface area of a nanoparticle and  pl is the interfacial energy between the nanoparticle and liquid metal. If the two nanoparticles move close to squeeze all metal atoms out to create a void between them, the interfacial energy becomes

GSSb2(S pl  pp ) pl  2 pp p , where Spp and p is the effective contact area and the surface energy

48 of the nanoparticle. If the two nanoparticles reach an adhesive contact to chemically bond together

(i.e. sintering starts), there will be no physical interface between the two nanoparticles. We will set this global energy minimum to be 0.

Therefore, when two nanoparticles move close enough to squeeze metal atoms out to create a void between them, the change in Gibbs free energy can be described as

GS1 2 (p  pl ) Equation 5-1 Where S is the effective contact area. If the nanoparticles move further close to start fusion together, the further change in Gibbs free energy can be described as

GS2  2  p Equation 5-2

Since the interfacial energy  p is always positive, G2 would be always negative. The schematic of the interfacial energies for two nanoparticles in a liquid metal is shown in Figure 5-2.

Figure 5-2. Interfacial energy W changes along the distance D between two nanoparticles

(푆: effective surface area; 휎푝푙: interfacial energy between nanoparticles and liquid metal; 휎푝:

surface energy of nanoparticles; a: characteristic atomic diameter of liquid metal) 49

 p and  pl are dominated by chemical bonds, mainly covalent bond and/or metallic bond formed between metals and the nanoparticle surface [198,199,200,201], which has a short characteristic length up to 0.2~0.4 nm. Under a thermodynamic favorable condition, when the distance between the two nanoparticles is less than the typical characteristic length of chemical bonds, adhesion would be initiated. Beyond this distance, the contribution to the interfacial energy by the chemical bonds can be considered to be negligible. There is a lack of reliable data about how the interfacial energy between liquid metal and ceramic varies with their separation distance

D. Thus a general expression for the interaction potential of two similar surfaces of a unit area along with the gap between them, D, may be applied as below [202]:

D  a (D D00)/ a 0 WDSinter  2 ( p pl )e for D00 D a Equation 5-3 Da0  0

Da/ 0 and WinterD 2 S p e (1D / D0 ) 2(S  pp  l ) for 0 DD0 . Equation 5-4

where D0 is the length of chemical bond, and a0 is the characteristic decaying length (normally

0.2~0.4 nm) for the chemical bond, i.e. two times of in this case.

5.2.2 Van der Waals potential

Van der Waals potential between two nanoparticles in a liquid metal is of a relatively long range. The attractive force induced by the van der Waals potential drives nanoparticles together.

The van der Waals potential for two spheres of radius R1 and R2 is determined by [161,203]

ARR12 WDvdw ()() Equation 5-5 6DRR12

And the force is

ARR12 Fvdw  2 () Equation 5-6 6DRR12 where A is the system Hamaker constant for the nanoparticle interactions in the liquid metal. Based on the Lifshitz theory [204], in a three-party model system the non-retarded Hamaker constant between media 1 and 2 across media 3 can be estimated by

 331  3   2  3  h 1ivn   3()() iv n  2 iv n  3 iv n A kT   [ ][ ] dv Equation 5-7 4     4  iv   ( iv )  iv   ( iv ) 1 3  2 3  v1 1n 3 n 2 n 3 n

where 1 , 2 and 3 are the static dielectric constants of the three media, ()iv is the value of 

2kT at imaginary frequencies, and vnn  . At room temperature and above, frequencies vn are h in the ultra-violet (UV) region.

The system Hamaker constant A132 , defined for material 1 and 2 interacting through

medium 3, can be approximately related to A11 and A22 through

AAAAA132( 11  33 )( 22  33 ) Equation 5-8

( A11 , A22 and A33 is material 1, 2 and 3 interaction with themselves through vacuum)

In the nanoparticle-metal system it can be described as

AAAAAsys(nanoparticle1  liquid )( nanoparticle2  li q uid ) Equation 5-9

And hence the van der Waals potential can be written as

(AAAA )( ) nanoparticle12 liquid nanoparticle liquid RR12 WDvdw    ( ) Equation 5-10 6D RR12

Therefore, the van der Waals potential between two similar nanoparticles is always negative, and thus the van der Waals force between two same nanoparticles in the liquid metal is always attractive. Table 5-1 shows some Hamaker constants available from literatures for some typical metals and ceramics in vacuum.

Table 5-1. Hamaker constant of some metals and ceramics.

Materials Hamaker constant (zJ) Reference

Al melt 266, 360 [205,206,207]

Mg melt 206 [107]

Al2O3 140 [208]

SiC 248 [209]

TiB2 256 [210]

MgO 120 [211]

TiO2 150 [212]

TiC 238 [213]

SiO2 71.6 [214]

Fe 260 [215]

5.2.3 Brownian potential

Since nanoparticles are very small and move randomly under thermal fluctuations,

Brownian potential should be considered. In this study, only the displacement along the direction that two nanoparticles approach each other will be considered. Equi-partition theorem suggests

52 that the kinetic energy/potential of the Brownian motion is kT /2 in one dimension for one particle. So the Brownian motion energy for the two nanoparticles system in on dimension is kT.

For two nanoparticles in the liquid metal at elevated temperatures, kT may be comparable to the van der Waals potential, and thus Brownian potential would play an important role on nanoparticle dispersion.

5.3 Three cases for nanoparticles dispersion in molten metals

In the model system for nanoparticle dispersion in a liquid metal, the van der Waals potential, interfacial energy, and Brownian potential intrinsically co-exist. The interfacial energy would dominate when the gap between two nanoparticles reaches one or two atomic layers. The van der Waals interaction could dominate outside this gap until a long distance (e.g. up to 10 nm or more). Three possible cases for nanoparticle dispersion could be deduced from the theoretical analysis: clusters, pseudo-dispersion, and self-dispersion.

5.3.1 Clusters

Figure 5-3 shows the interaction potentials for nanoparticles to form clusters. Wbarrier and

Wvdwmax are ∆G1 at the chemical bond characteristic length and the maximum van der Waals potential in the energy well, respectively. If Wbarrier is small (less than 10 kT, for example, due to a poor wettability between the nanoparticle and the liquid metal) and the van der Waals potential well is not deep enough, the Brownian potential kT can drive the nanoparticles to the adhesive contact to form chemical bonds, i.e. fuse together. Nanoparticle clusters will easily form in the liquid metal.

Figure 5-3. Interaction potential for nanoparticles to form clusters

5.3.2 Pseudo-dispersion

Figure 5-4 shows the interaction potentials for nanoparticle pseudo-dispersion in the liquid metal. If Wbarrier is high (i.e. good wettability between the nanoparticle and the liquid metal), the

Brownian potential kT cannot drive the nanoparticles to pass the barrier for any adhesive contact.

But if the van der Waals potential well is deep, the nanoparticles may be kinetically trapped into it, forming a cluster of separate nanoparticles, i.e. pseudo-dispersion. Nanoparticles will form pseudo-dispersion domains where dense nanoparticles are separated by only a few layers of metal atoms.

Figure 5-4. Interaction potential for nanoparticle pseudo-dispersion

5.3.3 Self-dispersion

Figure 5-5 shows the interaction potentials for nanoparticle self-dispersion in the liquid metal. If Wbarrier is high (i.e. good wettability between the nanoparticle and the liquid metal) and the van der Waals potential well is not deep, the Brownian potential kT cannot drive the nanoparticles to pass the barrier for any adhesive contact, but the nanoparticles would not be kinetically trapped either. This would allow nanoparticles to move freely without forming clusters in the liquid metal, i.e. self-dispersion. The discovery of a possible self-dispersion mechanism through our theoretical models can serve as a powerful tool for the experimental study later to realize a uniform dispersion of nanoparticles in large scale solidification processing of bulk

MMNCs.

Figure 5-5. Interaction potentials for nanoparticle self-dispersion

Chapter 6 Experimental study for nanoparticle self-dispersion in molten metals

Based on theoretic study and availability of nanoparticles in the market, two nanoparticle- metal combinations, TiC (with a radius of 25 nm) in liquid Al and SiC (with a radius of 30 nm) in liquid Mg, were first selected as promising model systems to achieve nanoparticle self-dispersion.

6.1 Experimental study of TiC nanoparticle dispersion in Al melt

6.1.1 Theoretical analysis of TiC nanoparticle dispersion in liquid Al

For two TiC nanoparticles in pure Al melt at 1093 K, the van der Waals interaction potential can be calculated by the following equation [161]:

( A  A )2 TiC Al RR1 2 Equation 6-1 WDvdw     ( ) 6D RR12 where D is the distance between the two nanoparticles (in nm scale and beyond one atomic layer);

ATiC and AAl is the Hamaker constant for TiC (238 zJ) [213] and molten Al (266 zJ~360 zJ)

[205,206,207] respectively, as listed in Table 5-1. R1 and R2 are the radius of the two nanoparticles. The van der Waals interaction between two same TiC in liquid Al is thus

()AA 2 R WD() TiC Al Equation 6-2 vdw 62D

-21 The unit of D and R is nm while that of WDvdw()is zJ (10 J). It should be noted that the above equation is effective only when two TiC nanoparticles interact through the molten Al when D is larger than about two atomic layer (i.e. 0.4 nm).

According to the Langbein approximation, the effective interaction area of two spheres is

RR12 , the interfacial energy will be SD 2 0 RR12

D  a (D D00)/ a 0 for D D a Equation 6-3 WDSinter  2 ( p pl )e 00 Da0  0 and

Da/ 0 WinterD 2 S p e (1D / D0 ) 2(S  pp  l ) for 0 DD0 Equation 6-4

Thermal energy at 1093 K is kT=15.1 zJ.

For TiC nanoparticles with a radius of 25 nm used in our experiments, if the Hamaker constant of Al is selected to be 266zJ, the attraction energy well W1 would be

2 ()ATiC A Al R12R WW11vdw d   ( ) 6.8zJ Equation 6-5 6d1 R12 R

But if the Hamaker constant of liquid Al is selected to be 360 zJ

()A  A 2 TiC Al R12R Equation 6-6 WW11vdw d   ( ) 131zJ 6d1 R12 R

The surface energy of liquid Al is 1.1 J/m2 while 1.35 J/m2 for TiC [216]. The wetting angle between Al and TiC is 50° at the processing temperature of 1093 K [217].The energy barrier

W2 due to the interfacial energy will be

Da RR12 ()/D0 D 0 a 0 00 3 W20WinterDe 4( D0  p  pl ) 26.48 10 zJ  1600 kT Equation 6-7 R1 R 2 D 0 a 0

Since W2 is far larger than the Brownian potential, it is of little chance for nanoparticles to overcome the energy barrier to reach adhesive contact. However, due to the uncertainty of the

Hamaker constant for liquid Al, it is difficult to conclude if the TiC nanoparticles would be in pseudo- or self- dispersion in the liquid Al.

6.1.2 Salt assisted nanoparticle incorporation

Severe problems existed for a direct feeding of TiC nanoparticles into Al melt as TiC nanoparticles either floated on the surface or settled down to the bottom of the Al melt, due to a

58 partial burning of TiC nanoparticles and the natural oxide films on Al melt at high temperature

[218,219]. K-Al-F based salt has been used to improve the wettability of Al on TiC at temperatures up to 1173 K [220, 221]. Thus, a salt-assisted nanoparticles incorporation method were developed to obtain master Al nanocomposites with a high loading of TiC nanocomposites.

KAlF4 salt pellets and TiC nanoparticles (with a radius of 25 nm) were mixed together in acetone (10 vol.% nanoparticles in the salt) by ultrasonic processing for 2 hrs. After the mixing, acetone was evaporated away at room temperature in a fume hood, resulting a well-mixed salt-TiC nanocomposite powders. The mixed powders were then dehydrated at 180 °C for 12 hrs. Pure Al

(99.93%) was melted at 820 °C in a graphite crucible (with an inner diameter of 40 mm and a height of 80 mm) in a furnace under Ar protection. The mixed powders were loaded onto the surface of the Al melt with a volume ratio of TiC nanoparticles to Al at about 15:85. After about

3~5 min, the salt started melting. A titanium stirrer with four blades (25.4 mm in diameter) was applied at the 2/3 height of the liquid melt to stir the melt at 200 rpm for 10 min. The melt were then cooled to room temperature in air. The solidified nanocomposites were cut open to obtain the

Al-TiC nanocomposite samples.

The nanocomposite samples were grinded and ion milled to expose the microstructure and embedded nanoparticles. The microstructure of the samples was studied with Field Emission

Scanning Electron Microscopy (SEM) (Zeiss Supra 40) with energy dispersive spectrometry

(EDS). Figure 6-1 shows the typical domains of TiC nanoparticles inside the Al matrix and the

TiC nanoparticle distribution inside the domain.

Figure 6-1. Domains (patches) of TiC nanoparticles inside Al matrix and TiC nanoparticles

(in black) inside one domain

6.1.3 Quantitative analysis of TiC nanoparticle loading in Al

TiC nanoparticle concentrations in Al were determined. Two Al-TiC nanocomposite samples were cut and cleaned by alcohol. The masses of the nanocomposite samples were measured by a precision scale to be 0.814 g and 0.779 g. The nanocomposite samples were then dissolved in a 12vol.% HCl solution in two centrifuge tubes in an ice-water base. More than 3 times of HCl solution was used for about 48 hrs to ensure a complete dissolution of the Al matrix.

The solution was then centrifuged at 5000 rpm for 10 min. The upper transparent liquid were collected for pH value check by a pH paper. More deionized water was then added into the centrifuge tube for repeated centrifuge processes until the PH value of the upper transparent liquid was 7. Then pure alcohol was added to the centrifuge tube to clean the water residue on the surface of TiC nanoparticles during two more centrifuge cycles. Finally the TiC nanoparticles inside the tubes were dehydrated at 180 °C for 48 hrs before the weight measurement. The TiC fractions in the two samples were determined to be 8.81 vol.% and 9.25 vol.% respectively. Therefore, the TiC

60 fraction in the nanocomposites is 9.03 vol.% by average. Since the initial volume ratio of TiC nanoparticles to Al was designed to be 15:85, the salt-assisted incorporation efficiency of TiC nanoparticle in Al melt was about 56%, which could be improved by process optimization if needed.

6.1.4 Settling down of TiC nanoparticles in Al

Al-9vol.% TiC nanocomposite ingots were applied as a master alloy to fabricate Al-2vol.%

TiC nanocomposites for the study of TiC nanoparticle dispersion in Al. From our previous study, the dispersion of TiC nanoparticles in solidified samples will be affected by the solidification front pushing. To mitigate the pushing effect, high cooling rates could be applied. A copper wedge mold was designed with a 5° inclination, as shown in Figure 6-3a. To check the melt filling of the mold, especially at the tip region, we cast the pure Al first. Figure 6-3b shows the sample cast in the copper wedge mold, indicating the mold can be filled well by the Al melt. Pure Al was melt at

820 °C in a graphite crucible under Ar protection. The Al-9vol.% TiC master alloy was added into the Al melt. The tip of the niobium ultrasonic probe was inserted about 6 mm into the melt. An ultrasonic vibration with a peak-to-peak amplitude of 60 μm at a frequency of 20 kHz was applied to process the melt for 15 min for nanoparticle dispersion. After the ultrasonic processing stopped, the metal melt were casted into the copper wedge mold. After solidification inside the wedge mold, samples from the sections of 2 mm, 4 mm, and 6 mm thick were cut for characterization.

Figure 6-3. (a) Copper wedge-copper-mold; (b) A sample cast in the mold

Figure 6-4 reveals the microstructure and the EDS spectrum of the sample of 2 mm thick.

No TiC nanoparticles or clusters were observed. Moreover, there is no Ti element in the EDS spectrum, further indicating that little TiC existed in the sample.

Figure 6-4. Sample of 2 mm thick section and EDS analysis of elements

Figure 6-5 shows the microstructure and EDS mapping of the sample at the 4 mm thick section. EDS mapping indicates the bright contains more Ti contents. The total weight fraction of the Ti element in the matrix is about 4.48%. Figure 6-6 reveals the microstructure inside the bright phase. As shown in Figure 6-6a, TiC nanoparticles are distributed but form networking structures.

TiC nanoparticles (dark phase) are identified in Figure 6-6b.

Figure 6-5. Sample of 4 mm thick and EDS mapping

Figure 6-6. (a) TiC distribution inside the bright phase in Figure 6-5; (b) TiC nanoparticles

(dark phase)

Figure 6-7 shows the microstructure and EDS mapping of the sample at the 6 mm thick section. TiC nanoparticles form dense networking structures. The dense networks could be attributed to the pushing of TiC nanoparticle clusters to the grain boundaries during solidification.

The total weight fraction of the Ti element in the matrix is 6.06%. Figure 6-8 reveals the microstructure and the TiC nanoparticles in this sample.

From the results discussed above, the concentration of TiC nanoparticles increases in the casted samples from the bottom to the top of the wedge mold. There could be two possible reasons: either TiC nanoparticles settled down inside the melt in the crucible before poured into the wedge mold or a thermos-capillary force pushed the TiC nanoparticles from bottom up during solidification in the copper wedge mold. It seems that the wedge mold casting was not suitable to study the TiC dispersion inside Al melt.

Figure 6-7. Microstructure and EDS mapping of the sample at the 6 mm thick section

Figure 6-8. Microstructure and TiC nanoparticles (dark phase) of the sample at the 6 mm

thick section

In order to study the potential sedimentation of TiC nanoparticles, more Al-1.2vol.% TiC nanocomposite samples were processed in the same conditions as above before the crucible was takes out to cool in air with a cooling rate of ~1K/s. The samples cut from the top and the bottom were grinded and then ion milled to expose the microstructure and embedded nanoparticles. Figure

6-9 (a) and (c) show the representative microstructures of the sample from top of the crucible.

Little TiC phases were identified. Figure 6-9 (b) and (d) show the representative microstructures of the sample from bottom of the crucible. TiC nanoparticles inside the domains are distributed along grain boundaries. It is believed the TiC nanoparticle domains sediment to the bottom.

Figure 6-9. (a) Microstructure of Al-1.2vol.% TiC sample from top of the crucible; (b) the

microstructure of Al-1.2vol.% TiC sample from bottom of the crucible; (c) higher

magnification for the sample from top of the crucible;(d) higher magnification for the

sample from bottom of the crucible

6.1.5 Dispersion of TiC in Al alloys in droplet casted sample

To avoid the sedimentation of TiC nanoparticle domains in Al, droplet casting was used.

Al-0.6vol.% TiC nanocomposites were fabricated in a similar processing conditions as the other casting experiments. With ultrasonic on, the surface oxide layer was skimmed out and a steel spoon was used to quickly take Al melt of about 2 g for casting into the copper wedge mold to avoid potential sedimentation and pushing of TiC nanoparticles during fast solidification. The sample made by the droplet method was shown in Figure 6-10. The cooling rate R at the thickness d in the copper wedge mold could be calculated by [222,223]

1000퐾푚푚2 푑 푅 ≈ ⁄( )2 Equation 6-8 푠 2

The estimated cooling rate at the 0.5mm thick section is around 16,000 K/s.

Figure 6-11 reveals the microstructure of the sample after solidification. TiC nanoparticles still formed domains inside the fast cooled Al matrix. Different domains were circled with black discontinuous lines in the left figure. Under a higher magnification, the representative domain suggests that TiC nanoparticles are most likely pseudo-dispersed in Al.

Figure 6-10. Sample made by droplet casting

Figure 6-11. (Left) different domains of TiC nanoparticles circled with black discontinuous

lines; (right) higher magnification of a representative TiC nanoparticle domain inside a

grain boundary

6.1.6 Dispersion study of TiC nanoparticles in Mg-Al alloys

Since TiC nanoparticles were only pseudo-dispersed in pure Al, a melt with a lower

Hamaker constant might help to achieve a self-dispersion of TiC nanoparticles. Semi-empirical results suggest that an alloy tend to have a lower Hamaker constant than its pure main metal element. Mg-Al alloy is expected to offer a lower Hamaker constant than the pure Al. Experimental study was thus conducted.

Al-9vol.% TiC nanocomposites (24.3 g) was diluted inside Mg (88 g) at 820 °C to make

Mg18Al-1.2vol.% TiC melt. After the ultrasonic processing for 15 mins, the sample were solidified inside the crucible (~1 K/s cooling rate) in air. Figure 6-12 reveals the microstructure of the Mg18Al-1.2vol.% TiC nanocomposites. Mg grains and eutectic phases are shown in Figure 6-

12 (a). Most TiC nanoparticles are located inside the eutectic phases, as shown in Figure 6-12 (b) and (c). While there are still some short TiC nanoparticle networks, most TiC nanoparticles are actually dispersed and separated from the others. The result suggests that a better TiC dispersion was achieved inside the eutectic phase in Mg-18Al alloy than in the pure Al.

Figure 6-12. TiC dispersion in eutectic phase in Mg18Al-1.2vol.% TiC nanocomposites

To quantitatively analyze the TiC dispersion in this sample, ImageJ was applied to measure the size distribution of TiC nanoparticles. One example of the processed images is presented in

Figure 6-13. The size distribution of TiC phases is shown in Figure 6-14.

Figure 6-13. ImageJ processed SEM image to quantitatively analyze nanoparticles

dispersion

Figure 6-14. Size distribution of TiC nanoparticles in Mg18Al-1.2vol.% TiC

nanocomposites

We also would like to study the size dispersion of TiC with various Mg concentration in

MgAl alloys. The Mg composition were reduced by evaporating Mg away. Mg18Al-1.2vol.% TiC nanocomposites were melted at 820 °C inside an induction heater under argon protection. The pressure was lowered to around 6 Torr, and the pressure and temperature were kept for 10 min to allow the Mg evaporation. Then the pressure was raised back to around 750 Torr for the sample to cool down to room temperature. The mass of the sample were reduced from 10.3 g to 6.72 g (i.e.

3.58 g Mg loss). A small piece was cut for characterization and the composition of the alloy is determined to be Mg29Al. The surface oxidation was polished off and the above evaporation procedure was repeated. After a second round of 10 min evaporation, the mass of the sample was reduced from 6.15 g to 4.39 g (i.e. 1.76 g Mg loss), making the alloy composition to be Mg42Al.

The surface oxidation was polished off and the above evaporation procedure was repeated for another 40 min. The mass of the sample was further reduced from 3.50 g to 1.81 g (i.e. 1.69 g Mg loss), making the composition of the alloy as Mg88Al. A small piece of sample was cut for characterization (1 hr sample).

The microstructure and the size distribution of TiC nanoparticles in Mg29Al-1.98vol.%TiC,

Mg42Al-3vol.%TiC, Mg88Al-7.5vol.%TiC are shown in Figure 6-15 to Figure 6-20.

Figure 6-15. Microstructure of Mg29Al-1.98vol.%TiC nanocomposites

Figure 6-16. Size distribution of TiC nanoparticles in Mg29Al-1.98vol.%TiC

nanocomposites

Figure 6-17. Microstructure of Mg42Al-3vol.%TiC nanocomposites

Figure 6-18. Size distribution of TiC nanoparticles in Mg42Al-3vol.%TiC nanocomposites

Figure 6-19. Microstructure of Mg88Al-7.5vol.%TiC nanocomposites

Figure 6-20. Size distribution of TiC nanoparticles in Mg88Al-7.5vol.%TiC

nanocomposites

After 1 hr evaporation, the fraction of the TiC size over 56000 nm2 in the Mg88Al-

7.5vol.%TiC was over 41.7% while only just 8.7% in the origin Mg18Al-1.2vol.% TiC. The size of TiC clusters increases possibly due to the higher Al content (thus higher Hamaker constant), higher nanoparticle concentration, and more attraction and coagulation among nanoparticles during the evaporation.

6.1.7 Summary on TiC nanoparticle dispersion in Al alloy melts

Due to the uncertainty of the Hamaker constants for both Al and TiC, theoretical analysis predicted the TiC dispersion as either pseudo-dispersion or self-dispersion. To avoid burning and surface oxidation problem for TiC nanoparticle incorporation into Al melts, a salt-assisted

76 incorporation method was developed to fabricate Al-9vol.% TiC master nanocomposites, which were easily used for dilution to a lower nanoparticle loading in final alloy melts. The quantitative determination of TiC volume fraction in the Al matrix was achieved through the HCl acid dissolution of Al alloys to extract TiC nanoparticles. Direct wedge mold casting of Al-TiC nanocomposites showed that TiC nanoparticle domains settled down after ultrasonic process stopped. Droplet casting method was then applied to avoid nanoparticle settling and pushing during solidification. In the droplet casted samples, TiC nanoparticles formed domains in the Al matrix, indicating a pseudo-dispersion of TiC in Al melts. However, in an effort to lower the

Hamaker constant by adding Mg, it was discovered that TiC nanoparticle disperse well in the eutectic phase in the Mg18Al alloy.

6.2 Experimental study on SiC nanoparticle dispersion in Mg

6.2.1 Theoretical analysis on SiC nanoparticle dispersion in Mg melt

For two SiC nanoparticles interacting in the pure Mg melt at 1000 K, the Hamaker constant,

ASiC and AMg , are 248 zJ and 206 zJ for SiC and Mg melt respectively [107,209]. Their Brownian potential (thermal energy) is kT=13.8 zJ at 1000 K.

The van der Waals interaction between the two same SiC with a radius of R in Mg is

( 248 206)2 R WD() Equation 6-9 vdw 62D

The unit of D and R is nm while the unit of WDvdw()is zJ. The above equation is effective only when two SiC interact through the liquid Mg when D is larger than two atomic Mg layers (i.e. 0.4 nm).

According to the Langbein approximation, the effective interaction area of two spheres is

RR12 SD 2 0 , the interfacial energy barrier will be RR12

D  a (D D00)/ a 0 WDSinter  2 ( p pl )e for D00 D a Equation 6-10 Da0  0

Da/ 0 and WinterD 2 S p e (1D / D0 ) 2(S  pp  l ) for 0 DD0 Equation 6-11

But before interfacial energy dominates, the local minimum W1 (energy well for attraction) is

( A  A )2 2 SiC Mg RR12 ( 248 206) R Wd11Wvdw    ( )  Equation 6-12 6d1 Rd12 R 621

Assume d1 is just the decaying length of interfacial energy, 0.4 nm, and when R is 30nm for the SiC nanoparticles used for this experimental study, then

( 248 206)2 R WW11 vdw d     12.17zJ   kT Equation 6-13 62d1

From the literature, the surface energy of liquid Mg is 0.599J/m2 [224] and the surface energy of SiC is 1.45 J/m2 [225]. The contact angle is 83° [226]. The interfacial energy between

2 liquid Mg and SiC will be 0.422 J/m according to Young’s equation. The local maximum W2 due to the interfacial energy will be

Da RR12 ()/D0 D 0 a 0 00 3 W20WinterDe 4( D0  p  pl ) 38.74 10 zJ  2800 kT Equation 6-14 R1 R 2 D 0 a 0

Since the local maximum W2 is far larger than 10kT, there will be little chance for the SiC nanoparticles to overcome the energy barrier to form clusters.

Thus, for two SiC nanoparticles with a radius of R1  30 nm in the pure Mg melt, the energy barrier W2 is much higher than the thermal energy while the energy well W1 is not enough to trap

SiC nanoparticles, allowing SiC nanoparticles to be self-dispersed in the pure Mg melt.

6.2.2 Fabrication of Mg2Zn with 14vol.% SiC nanocomposites

A uniform dispersion of a high volume fraction of SiC nanoparticles (with an average radius of 30 nm) in Mg matrix were achieved by liquid state processing. The schematic of the experimental setup is shown in Figure 6-21. Mg6Zn with 1 vol.% SiC were first fabricated through the well-established ultrasonic processing. Pure Mg (99.93%) and pure Zn (99.0%) were melt together at 700 °C in a furnace under SF6 (99 vol.%)/CO2 (1 vol.%) flow gas protection. The tip of the niobium ultrasonic probe was inserted about 6 mm in depth into the melt. An ultrasonic vibration with a frequency of 20 kHz and a peak-to-peak amplitude of 60 μm was generated from the transducer. The melt was ultrasonically processed for 15 minutes. The SiC nanoparticles are manually fed into the Mg6Zn (Mg+6wt.%Zn) melt, wetted and dispersed by ultrasonic processing.

After the ultrasonic process, the sample were cool down to room temperature in air.

The distribution and dispersion of nanoparticles in the as-solidified Mg6Zn matrix is characterized by SEM in Figure 6-22. With 1vol.% SiC nanoparticles, nanoparticles were pushed into intermetallic phase in the Mg6Zn alloy. Previous work shows that the pushing of nanoparticles by the solidification front can be effectively countered by a higher viscosity drag force in the melt via the introduction of a higher volume fraction of nanoparticles (e.g. >6vol.%). Therefore, to reveal the true dispersion of SiC in Mg, a high volume fraction of nanoparticles in the Mg melt would be desirable.

(a) (b)

Figure 6-21. Schematic of the experimental setup (a) Ultrasonic processing for

Mg6Zn-1vol.% SiC nanocomposites fabrication; (b) Evaporation for concentrating

nanoparticles in Mg

To achieve a high volume fraction of nanoparticles in the Mg melt, the nanoparticles were concentrated by evaporating away Mg and Zn from the Mg6Zn-1vol.% SiC samples at 6 Torr in a vacuum furnace. The evaporation process is schematically shown in Figure 6-21(b). Mg6Zn-1vol.%

SiC was first melt at 650 ℃ under Ar gas protection inside an induction heater before Mg and Zn were evaporated from the alloy melt at around 6 Torr.. After evaporation, the sample was cooled to room temperature in the furnace under a gas pressure of 760 Torr to obtain 14vol.% SiC in

Mg2Zn. The pushing of SiC nanoparticles by the solidification front was effectively countered by

80 a higher viscosity drag force in the melt. Therefore, the dispersion of SiC nanoparticles in the

Mg2Zn melt were maintained through the solidification inside the Mg2Zn matrix.

Figure 6-22. Nanoparticles were pushed into intermetallic phase in Mg6Zn

6.2.3 Self-dispersion of 14vol.% SiC in Mg2Zn

The distribution and dispersion of nanoparticles in Mg2Zn-14vol.% SiC were characterized by SEM. The chemical composition was determined by EDX (Energy dispersive X- ray spectroscopy) analysis. To reveal the nanoparticles clearly, the samples were first cleaned by low angle ion milling (10 degrees, to remove the nano-sized polishing powders) and then slightly etched by gallium ions (90 degrees, to preferentially etch magnesium matrix) by a focused ion beam. The SEM image were taken with a tilt angle of 52 degrees. From the SEM image as shown in Figure 6-23, the nanoparticles exposed on the surface of the Mg2Zn matrix distribute and disperse uniformly in the Mg matrix. Vickers hardness measurements were made under a load of

500 gram force with a dwelling time of 10 s. The microhardness and silicon concentration at different parts of the sample were measured to further confirm that the nanoparticles were distributed uniformly in the whole sample, as shown in Figure 6-23 (c).

Figure 6-23. (a)(b) SEM images of Mg-14vSiC sample acquired at 52 degrees tilt angle and

at different magnification; (c) Uniform distribution of nanoparticles across the whole

sample

To investigate the property enhancement induced by the populous dispersed nanoparticles, we first conducted in-situ micro-pillar compression testing in SEM on as-solidified samples.

Micro-compression tests on samples with diameter of about 4 μm and height of about 8 μm were conducted using a PI 85 SEM PicoIndenter (Hysitron, Inc., USA) with an 8 μm flat punch diamond probe inside an SEM (Nova 600, FEI, USA) using the displacement control mode at a strain rate of about 2x10-3 s-1. The micro-pillars were cut from bulk samples by FIB (Nova 600, FEI). The size of the micro-pillar (about 4 μm in diameter and 8 μm in length) was designed to contain only one grain to avoid the effect of grain boundaries on strengthening. This allows us to evaluate the property enhancement induced only by nanoparticles without the interference of grain boundaries for the cast samples. The size of the micro pillar was also carefully selected to avoid size-induced

82 strengthening in order to provide results comparable to macro scale tests for magnesium alloys.

To investigate the effect of nanoparticles on basal slip, the orientation of the pillars was chosen to favor basal slip.

The results from micro-compression tests show that the Mg-2Zn samples without nanoparticles yield at only about 50 MPa, then experience repeated loading-unloading cycles due to severe basal slipping, as shown in Figure 6-24(a) and (b). In contrast, the samples with nanoparticles yield at a significantly higher strength of about 410 MPa, and bear gradually increasing load smoothly to a plastic strain of over 30%, as shown in Figure 6-24 (a) and (c).

Moreover, after deformation, multiple slip traces are observed in the samples without nanoparticles

(Figure 6-24 (b)), but only one major slip trace developed at the later stage of deformation is observed in the samples with nanoparticles (Figure 6-24(c)). Even after the formation of the major slip trace, the samples with nanoparticles can still bear load smoothly. The testing results demonstrate that the populous dispersed nanoparticles not only can significantly strengthen the material but also can enable a more uniform and stable deformation by potentially suppressing basal slip.

Figure 6-24. Mechanical behavior of as-solidified samples at room temperature

(a) Engineering stress-strain curves of micro-pillar as-solidified samples without (black) and with (red) nanoparticles; (b)(c) SEM images showing the morphology of post-deformed

samples without (b) and with (c) nanoparticles

6.2.4 Summary on self-dispersion of SiC in Mg

Theoretical analysis shows SiC nanoparticles could self-dispersed in Mg melt.

Experimental study was conducted. Mg2Zn-14vol.% SiC nanocomposites were successfully fabricated through a two-step liquid processing method. Firstly, Mg6Zn-1vol.% SiC nanocomposites were fabricated by ultrasonic processing. Then, Mg2Zn-14vol.% SiC nanocomposites were achieved through an evaporation of Mg and Zn to concentrate SiC nanoparticles in the Mg melt. SEM images, EDS of Si composition, and Vickers hardness measurements clearly show that the SiC nanoparticles were self-dispersed and stabilized in Mg.

Micropillar compression test shows the Mg2Zn-14vol.% SiC nanocomposites yield at a significantly higher strength of about 410 MPa with a good plasticity, while only 50 MPa with a poor plasticity for pure Mg.

Chapter 7 Conclusions

The dissertation is motivated by the rapidly growing interests in the development of metal matrix nanocomposites (MMNCs) and a lack of fundamental knowledge of how to achieve a uniform nanoparticle dispersion in MMNCs. Several methods were explored to achieve a uniform nanoparticle dispersion in MMNCs.

Although some Al2O3 nanoparticles were distributed along the grain boundaries of some coarse

Al grains, most Al2O3 nanoparticles were evenly distributed inside the ultrafine Al grains to effectively restrict their grain growth. Moreover, the microhardness of the bulk Al nanocomposites is enhanced up to about three times as high as that of pure bulk Al.

Friction stir processing (FSP) were used to disperse nanoparticles in Mg6Zn + 6 vol.% SiC nanocomposites. Semi-solid mixing for effective incorporation of SiC nanoparticle into Mg6Zn matrix was applied before FSP. The low temperature at semi-solid state reduced the nanoparticles burning and oxidation effectively, while a high viscosity of the metal at semi-solid state trapped nanoparticles inside the matrix metal. FSP was also used to fabricate Mg + 6vol.% HA nanocomposites with a uniform dispersion and distribution of nanoparticles after HA nanoparticles were effectively incorporated into Mg melt by mechanical stirring. The mechanical properties of

Mg nanocomposites after FSP were significantly improved.

Unfortunately these two methods developed are not economical for mass manufacturing of

MMNCs, while solidification processing is very promising as a versatile mass manufacturing

85 method for production of bulk MMNC parts with complex geometry and high nanoparticle loading.

However, the incorporation and de-agglomeration of nanoparticles in liquid metals are extremely difficult.

To solve the long-standing challenge, a theoretical study was first conducted to understand the interactions in the nanoparticle-melt system. The fundamental understanding was then used to guide the experimental study.

A theoretical model have been successfully established to reveal the essential conditions for nanoparticle dispersion in molten metal during solidification nanoprocessing of bulk MMNCs.

The interactions between nanoparticles in molten metals include three key potentials, the interfacial energy barrier at a short range (1~2 atomic layers) to resist nanoparticles to come further into atomic contact, the attractive van der Waals potential (dominant in the longer range from 0.4 nm ~10 nm), and the Brownian potential, kT. Three possible cases for nanoparticle dispersion in molten metals were theoretically predicted below.

1. Clusters: when the maximum interfacial energy barrier is less than about 10kT due to a

poor wetting between nanoparticles and metal melt, the Brownian potential could allow

nanoparticles to come closer into atomic contact (sometimes for sintering at high

temperature). Nanoparticles will tend to form larger clusters in the liquid metal.

2. Pseudo-dispersion: if the maximum interfacial energy barrier is high enough (e.g. more

than 10 kT) due to a good wetting between nanoparticle and the molten metal, the

nanoparticles will not be able to come to atomic contact. However, if the van der Waals

attraction is much larger than the Brownian potential, nanoparticles will be trapped into

the energy well at a local minimum potential. Nanoparticles will form pseudo-

dispersion domains where dense nanoparticles are separated by only a few layers of

metal atoms.

3. Self-dispersion: When the maximum interfacial energy barrier is high (good wetting

between nanoparticles and the metal melt) and the van der Waals attraction is smaller

than the Brownian potential, nanoparticles will neither come to atomic contact nor be

trapped into the energy well. Nanoparticles will move freely inside the molten metal in

a self-dispersion and self-stabilization mode.

Based on the theoretic study and availability of nanoparticles in the market, two nanoparticle-material combinations, TiC (with a radius of 25 nm) in liquid Al and SiC (with a radius of 30 nm) in liquid Mg, were selected for experimental study.

Al matrix, indicating a pseudo-dispersion of TiC in liquid Al. To lower the attractive energy well,

Mg alloys (with lower Hamaker constants) were experimented. TiC nanoparticles were successfully dispersed in the Mg18Al eutectic alloy.

SiC nanoparticles (with a radius of 30 nm) are promising to be self-dispersed in liquid Mg from the theoretical analysis. Mg6Zn-1vol.%SiC nanocomposite ingots were first obtained by ultrasonic processing. The nanoparticles were mostly distributed along the grain boundary region after ingot solidification due to the pushing of nanoparticles by solidification front. To enable nanoparticle engulfment, a higher nanoparticle loading is needed. A new method was developed to concentrate SiC nanoparticles by evaporation of Mg and Zn from the Mg6Zn-1vol.%SiC ingots

87 at 6 torr in a vacuum furnace. After evaporation and slow cooling at approximately 0.23 K/s, a sample with about 14vol.% nanoparticles in an Mg2Zn matrix was obtained. Material characterizations conducted by using SEM, EDS, and Vickers hardness measurements revealed that SiC nanoparticles were self-dispersed in Mg. Micropillar compression tests were also carried out. The Mg2Zn-14vol.% SiC nanocomposites yield at a significantly higher strength of about 410

MPa with a good plasticity, while of 50 MPa with a poor plasticity for pure Mg2Zn.

In summary, this dissertation establishes a theoretical framework and developed experimental methodologies to achieve a uniform dispersion of dense nanoparticles in metals. The study has significantly advanced the fundamental understanding on the interactions between nanoparticles in molten metals. Under the guidance of the fundamental knowledge, MMNCs with a uniform dispersion of dense nanoparticles could be mass manufactured for widespread applications.

Chapter 8 Future work

While scientific and technical advances have been made to achieve a uniform dispersion of nanoparticles in some metal melts, significant work is much needed to tackle more scientific and technical challenges for mass manufacturing to harness the full potential of MMNCs in industry. Some key issues are identified for future work.

8.1 Scientific determination of Hamaker constants for alloys and nanoparticles

Hamaker constants of liquid and solid alloys and nanoparticles are essential to determine the van der Waals interaction potential between nanoparticles in alloy melts. However, little data on Hamaker constants of alloys and their melts is available from literatures. Even for those

Hamaker constants available, they are mostly experimented measured or theoretically estimated at room temperature, making them un-reliable for theoretical analysis of material systems at high temperatures. It will thus be of scientific significance to measure Hamaker constants of alloys and nanoparticles at high temperatures. One possible method is to measure the transmission electromagnetic spectrum, which can be used to estimate the Hamaker constant [161,227]. With more accurate temperature-dependent Hamaker constants available for alloys and nanoparticles, more accurate analytical or numerical models can be established to reliably determine nanoparticle dispersions in metal melts.

8.2 Size effect for nanoparticle dispersion in metal melts

Our earlier theoretical study suggests the interaction potentials depend on the size of the nanoparticles. It will be of scientific significance to investigate the size effect on nanoparticle dispersions in metal melts. Theoretical model suggests that small nanoparticles, for example less than

10 nm in diameter, could have better chances to achieve self-dispersion due to a smaller van der Waals attraction potential between them (thus smaller energy well), provided the nanoparticles are still thermodynamically stable to avoid chemical reactions and wet well in the melt to against an adhesive contact for sintering/bonding. Unfortunately, a lack of different sizes of nanoparticles in the market severely restrict a future investigation. Chemical synthesis of suitable nanoparticles, such as conductive transition metal carbides, borides and silicides potentially with high Hamaker constants, would be needed to provide various sizes for fundamental study.

8.3 Enable mass production of MMNCs with self-dispersed nanoparticles

Nanoparticle self-dispersion in molten metals opens wide pathways for mass processing of metal matrix nanocomposites for numerous applications. Under the guidance of the theoretical study, more metal-nanoparticle systems could be experimented for mass manufacturing. Master nanocomposite (alloy with a high loading of nanoparticles) could be first produced by the salt assisted method or other methods to be developed. More attention is needed to ensure a large volume production of master nanocomposites with high quality of nanoparticle dispersion. The master nanocomposites can then be used for dilution in regular solidification processing. Due to the self-dispersion and self-stabilization nanoparticles in the master nanocomposites, simple mechanical mixing can be used to allow their further dispersion in the diluted alloy melts. More fundamental studies are needed in this direction.

REFERENCES

1 Yang, M. Q., Zhang, N., Pagliaro, M., & Xu, Y. J. (2014). Artificial photosynthesis over

graphene–semiconductor composites. Are we getting better?. Chemical Society

Reviews, 43(24), 8240-8254.

2 Zhu, Q. L., & Xu, Q. (2014). Metal–organic framework composites. Chemical Society

Reviews, 43(16), 5468-5512.

3 Habisreutinger, S. N., Leijtens, T., Eperon, G. E., Stranks, S. D., Nicholas, R. J., & Snaith, H. J.

(2014). Carbon nanotube/polymer composites as a highly stable hole collection layer in

perovskite solar cells. Nano Letters, 14(10), 5561-5568.

4 Potts, J. R., Dreyer, D. R., Bielawski, C. W., & Ruoff, R. S. (2011). Graphene-based polymer

nanocomposites. Polymer, 52(1), 5-25.

5 Han, Z., & Fina, A. (2011). Thermal conductivity of carbon nanotubes and their polymer

nanocomposites: a review. Progress in Polymer Science, 36(7), 914-944.

6 Mahmood, N., Zhang, C., Yin, H., & Hou, Y. (2014). Graphene-based nanocomposites for

energy storage and conversion in lithium batteries, supercapacitors and fuel cells. Journal of

Materials Chemistry A, 2(1), 15-32.

7 Mutiso, R. M., & Winey, K. I. (2015). Electrical properties of polymer nanocomposites

containing rod-like nanofillers. Progress in Polymer Science, 40, 63-84.

8 Kelly, A. (1985). Composites in context. Composites Science and Technology, 23(3), 171-199.

9 Lu, H., Wang, X., Zhang, T., Cheng, Z., & Fang, Q. (2009). Design, fabrication, and properties

of high damping metal matrix composites—a review. Materials, 2(3), 958-977.

10 Tjong, S. C. (2007). Novel Nanoparticle‐Reinforced Metal Matrix Composites with Enhanced

Mechanical Properties. Advanced Engineering Materials, 9(8), 639-652.

11 Casati, R., & Vedani, M. (2014). Metal matrix composites reinforced by nano-particles—a

review. Metals, 4(1), 65-83.

12 Thostenson, E. T., Li, C., & Chou, T. W. (2005). Nanocomposites in context.Composites

Science and Technology, 65(3), 491-516.

13 Hussain, F., Hojjati, M., Okamoto, M., & Gorga, R. E. (2006). Review article: polymer-matrix

nanocomposites, processing, manufacturing, and application: an overview. Journal of

Composite Materials, 40(17), 1511-1575.

14 Hu, P., Shen, Y., Guan, Y., Zhang, X., Lin, Y., Zhang, Q., & Nan, C. W. (2014). Topological‐

Structure Modulated Polymer Nanocomposites Exhibiting Highly Enhanced Dielectric

Strength and Energy Density. Advanced Functional Materials, 24(21), 3172-3178.

15 Morgan, A. B., & Gilman, J. W. (2007). Polymer-clay nanocomposites: Design and application

of multi-functional materials. Materials Matters, 2, 20-25.

16 Zhong, M., Jiang, D., Zhang, H., Hong, C., Weisheit, A., & Kelbassa, I. (2014). Fabrication of

nanoparticulate reinforced metal matrix composites by laser cladding. Journal of Laser

Applications, 26(2), 022007.

17 Saheb, N., Iqbal, Z., Khalil, A., Hakeem, A. S., Al Aqeeli, N., Laoui, T., & Kirchner, R. (2012).

Spark plasma sintering of metals and metal matrix nanocomposites: a review. Journal of

Nanomaterials, 2012, 18.

18 Li, X., Yang, Y., & Cheng, X. (2004). Ultrasonic-assisted fabrication of metal matrix

nanocomposites. Journal of Materials Science, 39(9), 3211-3212.

19 Lakshmanan, P., & Sornakumar, T. (2015). Processing and performance characteristics of

aluminum-nano boron carbide metal matrix nanocomposites. Materials and Manufacturing

Processes, (just-accepted).

20 Goujon, C., & Goeuriot, P. (2001). Solid state sintering and high temperature compression

properties of Al-alloy5000/AlN nanocomposites. Materials Science and Engineering:

A, 315(1), 180-188.

21 Kubota, M., Kaneko, J., & Sugamata, M. (2008). Properties of mechanically milled and spark

plasma sintered Al–AlB2 and Al–MgB2 nano-composite materials. Materials Science and

Engineering: A, 475(1), 96-100.

22 Kubota, M., Cizek, P., & Rainforth, W. M. (2008). Properties of mechanically milled and spark

plasma sintered Al–15at.% MgB2 composite materials. Composites Science and

Technology, 68(3), 888-895.

23 Yar, A. A., Montazerian, M., Abdizadeh, H., & Baharvandi, H. R. (2009). Microstructure and

mechanical properties of aluminum alloy matrix composite reinforced with nano-particle

MgO. Journal of Alloys and Compounds, 484(1), 400-404.

24 Abdizadeh, H., Ebrahimifard, R., & Baghchesara, M. A. (2014). Investigation of microstructure

and mechanical properties of nano MgO reinforced Al composites manufactured by stir

casting and powder metallurgy methods: A comparative study. Composites Part B:

Engineering, 56, 217-221.

25 Nemati, N., Khosroshahi, R., Emamy, M., & Zolriasatein, A. (2011). Investigation of

microstructure, hardness and wear properties of Al–4.5wt.% Cu–TiC nanocomposites

produced by mechanical milling. Materials & Design, 32(7), 3718-3729.

26 Saheb, N., Aliyu, I. K., Hassan, S. F., & Al-Aqeeli, N. (2014). Matrix structure evolution and

nanoreinforcement distribution in mechanically milled and spark plasma sintered Al-SiC

nanocomposites. Materials, 7(9), 6748-6767.

27 Dutkiewicz, J., Lityńska, L., Maziarz, W., Haberko, K., Pyda, W., & Kanciruk, A. (2009).

Structure and properties of nanocomposites prepared from ball milled 6061 aluminium alloy

with ZrO2 nanoparticles. Crystal Research and Technology, 44(10), 1163-1169.

28 Pérez-Bustamante, R., Gómez-Esparza, C. D., Estrada-Guel, I., Miki-Yoshida, M., Licea-

Jiménez, L., Pérez-García, S. A., & Martínez-Sánchez, R. (2009). Microstructural and

mechanical characterization of Al–MWCNT composites produced by mechanical

milling. Materials Science and Engineering: A, 502(1), 159-163.

29 Ashida, M., Horita, Z., Kita, T., & Kato, A. (2012). Production of Al/Al2O3 Nanocomposites

through Consolidation by High-Pressure Torsion. Materials Transactions, 53(1), 13-16.

30 Poirier, D., Drew, R. A., Trudeau, M. L., & Gauvin, R. (2010). Fabrication and properties of

mechanically milled alumina/aluminum nanocomposites. Materials Science and Engineering:

A, 527(29), 7605-7614.

31 Sadeghian, Z., Enayati, M. H., & Beiss, P. (2009). In situ production of Al–TiB2 nanocomposite

by double-step mechanical alloying. Journal of Materials Science, 44(10), 2566-2572.

32 Wu, J. M., & Li, Z. Z. (2000). Nanostructured composite obtained by mechanically driven

reduction reaction of CuO and Al powder mixture. Journal of Alloys and Compounds, 299(1),

9-16.

33 Del Coso, R., Requejo-Isidro, J., Solis, J., Gonzalo, J., & Afonso, C. N. (2004). Third order

nonlinear optical susceptibility of Cu: Al2O3 nanocomposites: from spherical nanoparticles to

the percolation threshold. Journal of Applied Physics, 95, 2755-2762.

34 Ying, D. Y., & Zhang, D. L. (2000). Processing of Cu–Al2O3 metal matrix nanocomposite

materials by using high energy ball milling. Materials Science and Engineering: A, 286(1),

152-156.

35 Marques, M. T., Livramento, V., Correia, J. B., Almeida, A., & Vilar, R. (2005). Production of

copper–niobium carbide nanocomposite powders via mechanical alloying. Materials Science

and Engineering: A, 399(1), 382-386.

36 Takahashi, T., & Hashimoto, Y. (1991). Preparation of dispersion-strengthened coppers with

NbC and TaC by mechanical alloying. Materials Transactions, JIM, 32(4), 389-397.

37 Takahashi, T., Hashimoto, Y., & Koyama, K. (1989). Preparation of TiC-dispersion-

strengthened copper by the application of mechanical alloying. Journal of the Japan Society

of Powder and Powder Metallurgy, 36(7), 837-841.

38 Cha, S. I., Kim, K. T., Arshad, S. N., Mo, C. B., & Hong, S. H. (2005). Extraordinary

strengthening effect of carbon nanotubes in metal‐matrix nanocomposites processed by

molecular‐level mixing. Advanced Materials, 17(11), 1377-1381.

39 Ramalingam, S., Muralidharan, V. S., & Subramania, A. (2009). Electrodeposition and

characterization of Cu-TiO2 nanocomposite coatings. Journal of Solid State

Electrochemistry, 13(11), 1777-1783.

40 Mangam, V., Bhattacharya, S., Das, K., & Das, S. (2010). Friction and wear behavior of Cu–

CeO2 nanocomposite coatings synthesized by pulsed electrodeposition. Surface and Coatings

Technology, 205(3), 801-805.

41 Xue, D. S., Huang, Y. L., Ma, Y., Zhou, P. H., Niu, Z. P., Li, F. S., & Fahrner, W. R. (2003).

Magnetic properties of pure Fe-Al2O3 nanocomposites. Journal of Materials Science

Letters, 22(24), 1817-1820.

42 Mozaffari, M., Gheisari, M., Niyaifar, M., & Amighian, J. (2009). Magnetic properties of

mechanochemically prepared iron–wüstite (Fe–FeyO) nanocomposites. Journal of Magnetism

and Magnetic Materials, 321(19), 2981-2984.

43 Blackwell, J. J., Morales, M. P., O’Grady, K., González, J. M., Cebollada, F., & Alonso-Sañudo,

M. (2002). Interactions and hysteresis behavior of Fe/SiO2 nanocomposites. Journal of

Magnetism and Magnetic Materials, 242, 1103-1105.

44 Yazovskikh, K. А., & Lomayeva, S. F. (2014). Mechanosynthesis of Fe–NbC

nanocomposite. Journal of Alloys and Compounds, 586, S65-S67.

45 Shao, I., Vereecken, P. M., Chien, C. L., Searson, P. C., & Cammarata, R. C. (2002). Synthesis

and characterization of particle-reinforced Ni/Al2O3 nanocomposites. Journal of Materials

Research, 17(06), 1412-1418.

46 Zimmerman, A. F., Clark, D. G., Aust, K. T., & Erb, U. (2002). Pulse electrodeposition of Ni–

SiC nanocomposite. Materials Letters, 52(1), 85-90.

47 Qu, N. S., Zhu, D., & Chan, K. C. (2006). Fabrication of Ni–CeO2 nanocomposite by

electrodeposition. Scripta Materialia, 54(7), 1421-1425.

48 Lajevardi, S. A., & Shahrabi, T. (2010). Effects of pulse electrodeposition parameters on the

properties of Ni–TiO2 nanocomposite coatings. Applied Surface Science, 256(22), 6775-6781.

49 Hou, F., Wang, W., & Guo, H. (2006). Effect of the dispersibility of ZrO2 nanoparticles in Ni–

ZrO2 electroplated nanocomposite coatings on the mechanical properties of nanocomposite

coatings. Applied Surface Science, 252(10), 3812-3817.

50 Xue, Y. J., Zhu, D., & Zhao, F. (2004). Electrodeposition and mechanical properties of Ni–

La2O3 nanocomposites. Journal of Materials Science, 39(12), 4063-4066.

51 Jeon, Y. S., Byun, J. Y., & Oh, T. S. (2008). Electrodeposition and mechanical properties of

Ni–carbon nanotube nanocomposite coatings. Journal of Physics and Chemistry of

Solids, 69(5), 1391-1394.

52 Xia, F. F., Wu, M. H., Wang, F., Jia, Z. Y., & Wang, A. L. (2009). Nanocomposite Ni–TiN

coatings prepared by ultrasonic electrodeposition. Current Applied Physics, 9(1), 44-47.

53 Povstugar, I. V., Streletskii, A. N., Permenov, D. G., Kolbanev, I. V., & Mudretsova, S. N.

(2009). Mechanochemical synthesis of activated Me–BN (Me Al, Mg, Ti)

nanocomposites. Journal of Alloys and Compounds, 483(1), 298-301.

54 Prasad, Y. V. R. K., Rao, K. P., & Gupta, M. (2009). Hot workability and deformation

mechanisms in Mg/nano–Al2O3 composite. Composites Science and Technology, 69(7),

1070-1076.

55 Hwang, S., Nishimura, C., & McCormick, P. G. (2001). Compressive mechanical properties of

Mg-Ti-C nanocomposite synthesised by mechanical milling. Scripta Materialia, 44(10),

2457-2462.

56 Vijay, R., Sundaresan, R., Maiya, M. P., & Murthy, S. S. (2006). Hydrogen storage properties

of Mg–Cr2O3 nanocomposites: the role of catalyst distribution and grain size. Journal of

Alloys and Compounds, 424(1), 289-293.

57 Cao, G., Konishi, H., & Li, X. (2008). Mechanical properties and microstructure of Mg∕ SiC

nanocomposites fabricated by ultrasonic cavitation based nanomanufacturing. Journal of

Manufacturing Science and Engineering, 130(3), 031105.

58 Hassan, S. F., & Gupta, M. (2007). Development and characterization of ductile Mg∕ Y2O3

nanocomposites. Journal of Engineering Materials and Technology, 129(3), 462-467.

59 Chen, L. Y., Konishi, H., Fehrenbacher, A., Ma, C., Xu, J. Q., Choi, H., & Li, X. C. (2012).

Novel nanoprocessing route for bulk graphene nanoplatelets reinforced metal matrix

nanocomposites. Scripta Materialia, 67(1), 29-32.

60 Goh, C. S., Wei, J., Lee, L. C., & Gupta, M. (2006). Simultaneous enhancement in strength and

ductility by reinforcing magnesium with carbon nanotubes. Materials Science and

Engineering: A, 423(1), 153-156.

61 Goh, C. S., Gupta, M., Wei, J., & Lee, L. C. (2007). Characterization of high performance

Mg/MgO nanocomposites. Journal of Composite Materials, 41(19), 2325-2335.

62 Choi, H., Sun, Y., Slater, B. P., Konishi, H., & Li, X. (2012). AZ91D/TiB2 nanocomposites

fabricated by solidification nanoprocessing. Advanced Engineering Materials, 14(5), 291-

295.

63 Meenashisundaram, G. K., Nai, M. H., Almajid, A., & Gupta, M. (2015). Development of high

performance Mg–TiO2 nanocomposites targeting for biomedical/structural

applications. Materials & Design, 65, 104-114.

64 Ma, C., Chen, L., Xu, J., Fehrenbacher, A., Li, Y., Pfefferkorn, F. E., & Li, X. (2013). Effect

of fabrication and processing technology on the biodegradability of magnesium

nanocomposites. Journal of Biomedical Materials Research Part B: Applied

Biomaterials, 101(5), 870-877.

65 Karimzadeh, F., Enayati, M. H., & Tavoosi, M. (2008). Synthesis and characterization of

Zn/Al2O3 nanocomposite by mechanical alloying. Materials Science and Engineering:

A, 486(1), 45-48.

66 Vlasa, A., Varvara, S., Pop, A., Bulea, C., & Muresan, L. M. (2010). Electrodeposited Zn–TiO2

nanocomposite coatings and their corrosion behavior. Journal of Applied

Electrochemistry, 40(8), 1519-1527.

67 Vathsala, K., & Venkatesha, T. V. (2011). Zn–ZrO2 nanocomposite coatings: elecrodeposition

and evaluation of corrosion resistance. Applied Surface Science, 257(21), 8929-8936.

68 Rohatgi, P. K., & Schultz, B. (2007). Lightweight metal matrix nanocomposites-stretching the

boundaries of metals. Material Matters, 2(4), 16-20.

69 Saheb, N., Iqbal, Z., Khalil, A., Hakeem, A. S., Al Aqeeli, N., Laoui, T., & Kirchner, R. (2012).

Spark plasma sintering of metals and metal matrix nanocomposites: a review. Journal of

Nanomaterials, 2012, 18.

70 Cao, G., Konishi, H., & Li, X. (2008). Mechanical properties and microstructure of SiC-

reinforced Mg-(2,4) Al-1Si nanocomposites fabricated by ultrasonic cavitation based

solidification processing. Materials Science and Engineering: A, 486(1), 357-362.

71 Daoush, W. M., Lim, B. K., Mo, C. B., Nam, D. H., & Hong, S. H. (2009). Electrical and

mechanical properties of carbon nanotube reinforced copper nanocomposites fabricated by

electroless deposition process. Materials Science and Engineering: A, 513, 247-253.

72 Kang, J., Nash, P., Li, J., Shi, C., & Zhao, N. (2009). Achieving highly dispersed nanofibres at

high loading in carbon nanofibre–metal composites. Nanotechnology, 20(23), 235607.

73 Mazaheri, Y., Karimzadeh, F., & Enayati, M. H. (2011). A novel technique for development of

A356/Al2O3 surface nanocomposite by friction stir processing. Journal of Materials

Processing Technology, 211(10), 1614-1619.

74 Xu, J., Cao, C., Das, S., Chen, L., Ma, C., Mishra, R. S., & Li, X. High performance Mg6Zn

nanocomposites fabricated through friction stir processing. Magnesium Technology 2015,

383-386.

75 Zhang, Q., Chen, G., Wu, G., Xiu, Z., & Luan, B. (2003). Property characteristics of a AlNp/Al

composite fabricated by squeeze casting technology. Materials Letters, 57(8), 1453-1458.

76 Schultz, B. F., Ferguson, J. B., & Rohatgi, P. K. (2011). Microstructure and hardness of Al2O3

nanoparticle reinforced Al–Mg composites fabricated by reactive wetting and stir

mixing. Materials Science and Engineering: A, 530, 87-97.

77 Sajjadi, S. A., Ezatpour, H. R., & Beygi, H. (2011). Microstructure and mechanical properties

of Al–Al 2 O 3 micro and nano composites fabricated by stir casting. Materials Science and

Engineering: A, 528(29), 8765-8771.

78 Sajjadi, S. A., Parizi, M. T., Ezatpour, H. R., & Sedghi, A. (2012). Fabrication of A356

composite reinforced with micro and nano Al2O3 particles by a developed compocasting

method and study of its properties. Journal of Alloys and Compounds, 511(1), 226-231.

79 Ezatpour, H. R., Sajjadi, S. A., Sabzevar, M. H., & Huang, Y. (2014). Investigation of

microstructure and mechanical properties of Al6061-nanocomposite fabricated by stir

casting. Materials & Design, 55, 921-928.

80 Hamedan, A. D., & Shahmiri, M. (2012). Production of A356–1wt% SiC nanocomposite by

the modified stir casting method. Materials Science and Engineering: A, 556, 921-926.

81 Valibeygloo, N., Khosroshahi, R. A., & Mousavian, R. T. (2013). Microstructural and

mechanical properties of Al-4.5wt%Cu reinforced with alumina nanoparticles by stir casting

method. International Journal of Minerals, Metallurgy, and Materials, 20(10), 978-985.

100

82 Suryanarayana, C., & Al-Aqeeli, N. (2013). Mechanically alloyed nanocomposites. Progress

in Materials Science, 58(4), 383-502.

83 Benjamin, J. S. (1970). Dispersion strengthened superalloys by mechanical

alloying. Metallurgical Transactions, 1(10), 2943-2951.

84 Gu, D., Meiners, W., Hagedorn, Y. C., Wissenbach, K., & Poprawe, R. (2010). Structural

evolution and formation mechanisms of TiC/Ti nanocomposites prepared by high-energy

mechanical alloying. Journal of Physics D: Applied Physics, 43(13), 135402.

85 Doubenskaia, M., Gilmutdinov, A. K., & Nagulin, K. Y. (2015). Laser cladding of metal matrix

composites reinforced by cermet inclusions for dry friction applications at ambient and

elevated temperatures. Surface and Coatings Technology, 276, 696-703.

86 Zhong, M., Jiang, D., Zhang, H., Hong, C., Weisheit, A., & Kelbassa, I. (2014). Fabrication of

nanoparticulate reinforced metal matrix composites by laser cladding. Journal of Laser

Applications, 26(2), 022007.

87 Song, B., Dong, S., & Coddet, C. (2014). Rapid in situ fabrication of Fe/SiC bulk

nanocomposites by selective laser melting directly from a mixed powder of microsized Fe

and SiC. Scripta Materialia, 75, 90-93.

88 Pérez, A. T., Montealegre, M. A., Arias, J. L., Castro, G., Rey, P., González, M., & Bouzada,

F. (2012, January). Laser cladding of nickel alloy with ceramic nanopowder on steel as

coating in corrosive media. In Materials Science Forum (Vol. 706, pp. 2552-2558).

89 Gu, D., & Zhang, G. (2013). Selective laser melting of novel nanocomposites parts with

enhanced tribological performance: nanocrystalline TiC/Ti nanocomposites parts were built

via SLM technology and the densification, microstructures, microhardness and tribological

performance were investigated. Virtual and Physical Prototyping, 8(1), 11-18.

101

90 Gu, D., Wang, H., & Dai, D. (2015). Laser additive manufacturing of novel aluminum based

nanocomposite parts: tailored forming of multiple materials. Journal of Manufacturing

Science and Engineering.

91 Gu, D., Wang, H., Dai, D., Chang, F., Meiners, W., Hagedorn, Y. C., & Poprawe, R. (2015).

Densification behavior, microstructure evolution, and wear property of TiC nanoparticle

reinforced AlSi10Mg bulk-form nanocomposites prepared by selective laser melting. Journal

of Laser Applications, 27(S1), S17003.

92 Podlaha, E. J. (2001). Selective electrodeposition of nanoparticulates into metal matrices. Nano

Letters, 1(8), 413-416.

93 Lee, D., Gan, Y. X., Chen, X., & Kysar, J. W. (2007). Influence of ultrasonic irradiation on the

microstructure of Cu/Al2O3, CeO2 nanocomposite thin films during

electrocodeposition. Materials Science and Engineering: A, 447(1), 209-216.

94 Salehi, M., Farnoush, H., & Mohandesi, J. A. (2014). Fabrication and characterization of

functionally graded Al–SiC nanocomposite by using a novel multistep friction stir

processing. Materials & Design, 63, 419-426.

95 Abdolahzadeh, A., Omidvar, H., Safarkhanian, M. A., & Bahrami, M. (2014). Studying

microstructure and mechanical properties of SiC-incorporated AZ31 joints fabricated through

FSW: the effects of rotational and traveling speeds. The International Journal of Advanced

Manufacturing Technology, 75(5-8), 1189-1196.

96 Sun, K., Shi, Q. Y., Sun, Y. J., & Chen, G. Q. (2012). Microstructure and mechanical property

of nano-SiCp reinforced high strength Mg bulk composites produced by friction stir

processing. Materials Science and Engineering: A,547, 32-37.

102

97 Bauri, R., Yadav, D., & Suhas, G. (2011). Effect of friction stir processing (FSP) on

microstructure and properties of Al–TiC in situ composite. Materials Science and

Engineering: A, 528(13), 4732-4739.

98 Lim, D. K., Shibayanagi, T., & Gerlich, A. P. (2009). Synthesis of multi-walled CNT reinforced

aluminium alloy composite via friction stir processing. Materials Science and Engineering:

A, 507(1), 194-199.

99 Shafiei-Zarghani, A., Kashani-Bozorg, S. F., & Zarei-Hanzaki, A. (2009). Microstructures and

mechanical properties of Al/Al2O3 surface nano-composite layer produced by friction stir

processing. Materials Science and Engineering: A, 500(1), 84-91.

100 Morisada, Y., Fujii, H., Kawahito, Y., Nakata, K., & Tanaka, M. (2011). Three-dimensional

visualization of material flow during friction stir welding by two pairs of X-ray transmission

systems. Scripta Materialia, 65(12), 1085-1088..

101 Uozumi, H., Kobayashi, K., Nakanishi, K., Matsunaga, T., Shinozaki, K., Sakamoto, H., &

Yoshida, M. (2008). Fabrication process of carbon nanotube/light metal matrix composites

by squeeze casting. Materials Science and Engineering: A, 495(1), 282-287.

102 Ferguson, J. B., Aguirre, I., Lopez, H., Schultz, B. F., Cho, K., & Rohatgi, P. K. (2014).

Tensile properties of reactive stir-mixed and squeeze cast Al/CuO np-based metal matrix

nanocomposites. Materials Science and Engineering: A, 611, 326-332.

103 Elshalakany, A. B., Osman, T. A., Khattab, A., Azzam, B., & Zaki, M. (2014). Microstructure

and mechanical properties of MWCNTs reinforced A356 aluminum alloys cast

nanocomposites fabricated by using a combination of rheocasting and squeeze casting

techniques. Journal of Nanomaterials, 2014.

103

104 Gu, D., Meiners, W., Hagedorn, Y. C., Wissenbach, K., & Poprawe, R. (2010). Structural

evolution and formation mechanisms of TiC/Ti nanocomposites prepared by high-energy

mechanical alloying. Journal of physics D: Applied Physics, 43(13), 135402.

105 Lu, Z. J., Shen, Y. F., & Gu, D. D. (2010). Synthesis and characterisation of Si3N4p/Cu

nanocomposite powders by high energy ball milling. Powder Metallurgy, 53(1), 86-90.

106 Xiong, B., Xu, Z., Yan, Q., Cai, C., Zheng, Y., & Lu, B. (2010). Fabrication of SiC

nanoparticulates reinforced Al matrix composites by combining pressureless infiltration with

ball-milling and cold-pressing technology. Journal of Alloys and Compounds, 497(1), L1-L4.

107 Xu, J. Q., Chen, L. Y., Choi, H., & Li, X. C. (2012). Theoretical study and pathways for

nanoparticle capture during solidification of metal melt. Journal of Physics: Condensed

Matter, 24(25), 255304.

108 Xu, J., Chen, L., Choi, H., Konish, H., & Li, X. (2013). Assembly of metals and nanoparticles

into novel nanocomposite superstructures. Scientific Reports, 3.

109 Chen, L. Y., Xu, J. Q., Choi, H., Konishi, H., Jin, S., & Li, X. C. (2014). Rapid control of

phase growth by nanoparticles. Nature Communications, 5.

110 Hassan, S. F., & Gupta, M. (2004). Development of high performance magnesium

nanocomposites using solidification processing route. Materials Science and

Technology, 20(11), 1383-1388.

111 Cao, G., Kobliska, J., Konishi, H., & Li, X. (2008). Tensile properties and microstructure of

SiC nanoparticle–reinforced Mg-4Zn alloy fabricated by ultrasonic cavitation–based

solidification processing. Metallurgical and Materials Transactions A, 39(4), 880-886.

104

112 De Cicco, M., Konishi, H., Cao, G., Choi, H. S., Turng, L. S., Perepezko, J. H., & Li, X.

(2009). Strong, ductile magnesium-zinc nanocomposites. Metallurgical and Materials

Transactions A, 40(12), 3038-3045.

113 De Cicco, M., Turng, L. S., Li, X. C., & Perepezko, J. H. (2006, October). Semi-solid casting

of metal matrix nanocomposites. In Solid State Phenomena (Vol. 116, pp. 478-483).

114 Cantor, B. (2003). Heterogeneous nucleation and adsorption. Philosophical Transactions of

the Royal Society of London A: Mathematical, Physical and Engineering Sciences, 361(1804),

409-417.

115 Quested, T. E. (2004). Understanding mechanisms of grain refinement of aluminium alloys

by inoculation. Materials Science and Technology, 20(11), 1357-1369.

116 Quested, T. E., & Greer, A. L. (2004). The effect of the size distribution of inoculant particles

on as-cast grain size in aluminium alloys. Acta Materialia, 52(13), 3859-3868.

117 De Cicco, M. P., Turng, L. S., Li, X., & Perepezko, J. H. (2011). Nucleation catalysis in

aluminum alloy A356 using nanoscale inoculants. Metallurgical and Materials Transactions

A, 42(8), 2323-2330.

118 Wang, D., De Cicco, M. P., & Li, X. (2012). Using diluted master nanocomposites to achieve

grain refinement and mechanical property enhancement in as-cast Al–9Mg. Materials Science

and Engineering: A, 532, 396-400.

119 Choi, H., Cho, W. H., Konishi, H., Kou, S., & Li, X. (2013). Nanoparticle-induced superior

hot tearing resistance of A206 alloy. Metallurgical and Materials Transactions A, 44(4),

1897-1907.

105

120 Estruga, M., Chen, L., Choi, H., Li, X., & Jin, S. (2013). Ultrasonic-assisted synthesis of

surface-clean TiB2 nanoparticles and their improved dispersion and capture in Al-matrix

nanocomposites. ACS Applied Materials & Interfaces, 5(17), 8813-8819.

121 Zener C. (1948). quoted in Smith CS, vol. 175. Transactions of the Metallurgical Society of

AIME, p.15.

122 Perez, R. J., Jiang, H. G., Lavernia, E. J., & Dogan, C. P. (1998). Grain growth of

nanocrystalline cryomilled Fe-Al powders. Metallurgical and Materials Transactions

A, 29(10), 2469-2475.

123 Tellkamp, V. L., Lavernia, E. J., & Melmed, A. (2001). Mechanical behavior and

microstructure of a thermally stable bulk nanostructured Al alloy. Metallurgical and

Materials Transactions A, 32(9), 2335-2343.

124 Novikov, V. Y. (2000). On zener pinning in 3-D polycrystals. Scripta Materialia,42(5), 439-

443.

125 Couturier, G., Maurice, C., & Fortunier, R. (2003). Three-dimensional finite-element

simulation of Zener pinning dynamics. Philosophical Magazine, 83(30), 3387-3405.

126 Olguín, A., Ortiz, M., Wörner, C. H., Herrera, O., Kad, B. K., & Hazzledine, P. M. (2001).

Zener pins and needles. Philosophical Magazine B, 81(8), 731-744.

127 Okada, K., & Sakuma, T. (1992). The role of Zener's pinning effect on the grain growth in

Al2O3-ZrO2. Nippon Seramikkusu Kyokai Gakujutsu Ronbunshi, 100(4), 382-386.

128 Zhang, Z., & Chen, D. L. (2006). Consideration of Orowan strengthening effect in particulate-

reinforced metal matrix nanocomposites: a model for predicting their yield strength. Scripta

Materialia, 54(7), 1321-1326.

106

129 Zhang, Z., & Chen, D. L. (2008). Contribution of Orowan strengthening effect in particulate-

reinforced metal matrix nanocomposites. Materials Science and Engineering: A, 483, 148-

152.

130 Nie, J. F. (2003). Effects of precipitate shape and orientation on dispersion strengthening in

magnesium alloys. Scripta Materialia, 48(8), 1009-1015.

131 Dieter, G. E., & Bacon, D. (1986). Mechanical metallurgy (Vol. 3). New York: McGraw-Hill.

132 Lukac, P., & Trojanova, Z. (2005). Magnesium-based nanocomposites. International Journal

of Materials and Product Technology, 23(1-2), 121-137.

133 Yadollahpour, M., Ziaei-Rad, S., Karimzadeh, F., & Eskandari-Jam, J. (2011). A numerical

study on the damping capacity of metal matrix nanocomposites. Simulation Modelling

Practice and Theory, 19(1), 337-349.

134 Xu, J., Tao, J., Jiang, S., & Xu, Z. (2008). Investigation on corrosion and wear behaviors of

nanoparticles reinforced Ni-based composite alloying layer. Applied Surface Science, 254(13),

4036-4043.

135 Xu, J., Tao, J., & Jiang, S. (2008). Microstructure and properties of nano-Al 2 O 3 particle

reinforced Ni–Cr–Mo–Cu composite alloying layer. Materials Chemistry and Physics, 112(3),

966-972.

136 Xu, J., Zhuo, C., Tao, J., & Jiang, S. (2008). The effect of second-phase on the corrosion and

wear behaviors of composite alloying layer. Applied Surface Science, 255(5), 2688-2696.

137 Liu, L., & Xu, J. (2011). A study of the erosion–corrosion behavior of nano-Cr2O3 particles

reinforced Ni-based composite alloying layer in aqueous slurry environment. Vacuum, 85(6),

687-700.

107

138 Wang, W., Zhu, F., Weng, J., Xiao, J., & Lai, W. (1998). Nanoparticle morphology in a

granular Cu-Co alloy with giant magnetoresistance. Applied Physics Letters, 72, 1118.

139 Vovchenko, L., Matzui, L., Stelmakh, O., & Zakharenko, M. (2005). Electric and

Magnetoresistance of Nanocomposite Material Graphite‐Cobalt. Fullerenes, Nanotubes, and

Carbon Nanostructures, 13(S1), 491-495.

140 Uddin, S. M., Mahmud, T., Wolf, C., Glanz, C., Kolaric, I., Volkmer, C., & Fecht, H. J. (2010).

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, 70(16), 2253-2257.

141 Razavi, M., Fathi, M. H., & Meratian, M. (2010). Bio-corrosion behavior of magnesium-

fluorapatite nanocomposite for biomedical applications. Materials Letters, 64(22), 2487-

2490.

142 Müller, F., Peukert, W., Polke, R., & Stenger, F. (2004). Dispersing nanoparticles in

liquids. International Journal of Mineral Processing, 74, S31-S41.

143 Rwei, S. P., Manas‐Zloczower, I., & Feke, D. L. (1990). Observation of carbon black

agglomerate dispersion in simple shear flows. Polymer Engineering & Science, 30(12), 701-

706.

144 Kusters, K. A., Pratsinis, S. E., Thoma, S. G., & Smith, D. M. (1994). Energy—size reduction

laws for ultrasonic fragmentation. Powder technology, 80(3), 253-263.

145 Kendall, K., & Stainton, C. (2001). Adhesion and aggregation of fine particles. Powder

Technology, 121(2), 223-229.

146 Mende, S., Stenger, F., Peukert, W., & Schwedes, J. (2003). Mechanical production and

stabilization of submicron particles in stirred media mills. Powder Technology, 132(1), 64-73.

108

147 Müller, F., & Polke, R. F. (1999). From the product and process requirements to the milling

facility. Powder Technology, 105(1), 2-13.

148 Müller, F., & Polke, R. (2002). Agglomeration und strukturänderung bei der mahlung. AT.

Aufbereitungs-Technik, 43(8), 12-18.

149 Schönert, K. (2003). Bruchvorgänge und mikroprozesse des zerkleinerns.Handbuch der

Mechanischen Verfahrenstechnik, Whiley-VCH, Weinheim, 183-213.

150 Pacek, A. W., Ding, P., & Utomo, A. T. (2007). Effect of energy density, pH and temperature

on de-aggregation in nano-particles/water suspensions in high shear mixer. Powder

Technology, 173(3), 203-210.

151 Suslick, K. S. (1990). Sonochemistry. Science, 247(4949), 1439-1445.

152 Mason, T. J., & Lorimer, J. P. (2002). Applied sonochemistry. The Uses of Power Ultrasound

in Chemistry and Processing, 1-48.

153 Pohl, M. (2005). Benetzen und Dispergieren nativer und gezielt agglomerierter pyrogener

oxidischer Partikel. Cuvillier.

154 Xu, L. R., Bhamidipati, V., Zhong, W. H., Li, J., Lukehart, C. M., Lara-Curzio, E., & Lance,

M. J. (2004). Mechanical property characterization of a polymeric nanocomposite reinforced

by graphitic nanofibers with reactive linkers. Journal of Composite Materials, 38(18), 1563-

1582.

155 Yang, Y., Lan, J., & Li, X. (2004). Study on bulk aluminum matrix nano-composite fabricated

by ultrasonic dispersion of nano-sized SiC particles in molten aluminum alloy. Materials

Science and Engineering: A, 380(1), 378-383.

156 Suslick, K. S., Didenko, Y., Fang, M. M., Hyeon, T., Kolbeck, K. J., McNamara, W. B., &

Wong, M. (1999). Acoustic cavitation and its chemical consequences. Philosophical

109

Transactions of the Royal Society of London A: Mathematical, Physical and Engineering

Sciences, 357(1751), 335-353.

157 Suslick, K. S., & Price, G. J. (1999). Applications of ultrasound to materials chemistry. Annual

Review of Materials Science, 29(1), 295-326.

158 Ma, L., Chen, F., & Shu, G. (1995). Preparation of fine particulate reinforced metal matrix

composites by high intensity ultrasonic treatment. Journal of Materials Science Letters, 14(9),

649-650.

159 Lan, J., Yang, Y., & Li, X. (2004). Microstructure and microhardness of SiC nanoparticles

reinforced magnesium composites fabricated by ultrasonic method. Materials Science and

Engineering: A, 386(1), 284-290.

160 Verwey, E. J. W., & Overbeek, J. T. G. (1999). Theory of the Stability of Lyophobic Colloids.

Courier Corporation.

161 Israelachvili, J. N. (2011). Intermolecular and Surface Forces: revised third edition.

Academic press.

162 Cheng, L., Fenter, P., Nagy, K. L., Schlegel, M. L., & Sturchio, N. C. (2001). Molecular-scale

density oscillations in water adjacent to a mica surface. Physical Review Letters, 87(15),

156103.

163 Huisman, W. J., Peters, J. F., Zwanenburg, M. J., de Vries, S. A., Derry, T. E., Abernathy, D.,

& van der Veen, J. F. (1997). Layering of a liquid metal in contact with a hard

wall. Nature, 390(6658), 379-381.

164 Yu, C. J., Richter, A. G., Datta, A., Durbin, M. K., & Dutta, P. (1999). Observation of

molecular layering in thin liquid films using X-ray reflectivity. Physical Review

Letters, 82(11), 2326.

110

165 Yu, C. J., Richter, A. G., Datta, A., Durbin, M. K., & Dutta, P. (2000). Molecular layering in

a liquid on a solid substrate: an X-ray reflectivity study. Physica B: Condensed Matter, 283(1),

27-31.

166 Yazdabadi, H. G., Ekrami, A., Kim, H. S., & Simchi, A. (2013). An Investigation on the

Fatigue Fracture of P/M Al-SiC Nanocomposites. Metallurgical and Materials Transactions

A, 44(6), 2662-2671.

167 Zolriasatein, A., Khosroshahi, R. A., Emamy, M., & Nemati, N. (2013). Mechanical and wear

properties of Al-Al3Mg2 nanocomposites prepared by mechanical milling and hot

pressing. International Journal of Minerals, Metallurgy, and Materials, 20(3), 290-297.

168 So, K. P., Jeong, J. C., Park, J. G., Park, H. K., Choi, Y. H., Noh, D. H., ... & Lee, Y. H. (2013).

SiC formation on carbon nanotube surface for improving wettability with

aluminum. Composites Science and Technology, 74, 6-13.

169 Moazami-Goudarzi, M., & Akhlaghi, F. (2012). Novel approach based on in situ powder

metallurgy (IPM) method for embedding of SiC nanoparticles in aluminium

powders. Materials Science and Technology, 27(2), 158-164.

170 Mahmoud, T. S., El-Kady, E. Y., & Al-Shihri, A. (2012). Mechanical and corrosion

behaviours of Al/SiC and Al/Al2O3 metal matrix nanocomposites fabricated using powder

metallurgy route. Corrosion Engineering, Science and Technology, 47(1), 45-53.

171 Alizadeh, A., & Taheri-Nassaj, E. (2011). Wear behavior of nanostructured Al and Al–B4C

nanocomposites produced by mechanical milling and hot extrusion. Tribology Letters, 44(1),

59-66.

111

172 Sharifi, E. M., Karimzadeh, F., & Enayati, M. H. (2011). Fabrication and evaluation of

mechanical and tribological properties of boron carbide reinforced aluminum matrix

nanocomposites. Materials & Design, 32(6), 3263-3271.

173 Tjong, S. C., & Ma, Z. Y. (2000). Microstructural and mechanical characteristics of in situ

metal matrix composites. Materials Science and Engineering: R: Reports, 29(3), 49-113.

174 Simões, S., Viana, F., Reis, M. A., & Vieira, M. F. (2014). Improved dispersion of carbon

nanotubes in aluminum nanocomposites. Composite Structures, 108, 992-1000.

175 Casati, R., Bonollo, F., Dellasega, D., Fabrizi, A., Timelli, G., Tuissi, A., & Vedani, M. (2014).

Ex situ Al–Al2O3 ultrafine grained nanocomposites produced via powder metallurgy. Journal

of Alloys and Compounds, 615, S386-S388.

176 Rasouli, S., Taheri-Nassaj, E., & Hassanzadeh-Tabrizi, S. A. (2013). In-situ fabrication of

Al2O3–SiC nanocomposites using B2O3 as sintering aid. Ceramics International, 39(4), 3931-

3938.

177 Dallas, P., Niarchos, D., Vrbanic, D., Boukos, N., Pejovnik, S., Trapalis, C., & Petridis, D.

(2007). Interfacial polymerization of pyrrole and in situ synthesis of polypyrrole/silver

nanocomposites. Polymer, 48(7).

178 Ghannam, L., Parvole, J., Laruelle, G., Francois, J., & Billon, L. (2006). Surface‐initiated

nitroxide‐mediated polymerization: a tool for hybrid inorganic/organic nanocomposites ‘in

situ’synthesis. Polymer international, 55(10), 1199-1207.

179 Trunov, M. A., Schoenitz, M., Zhu, X., & Dreizin, E. L. (2005). Effect of polymorphic phase

transformations in Al 2 O 3 film on oxidation kinetics of aluminum powders. Combustion and

Flame, 140(4), 310-318.

180 Davis, J. R. (1998). Metals Handbook. ASM international.

112

181 Khorshid, M. T., Jahromi, S. J., & Moshksar, M. M. (2010). Mechanical properties of tri-

modal Al matrix composites reinforced by nano-and submicron-sized Al2O3 particulates

developed by wet attrition milling and hot extrusion. Materials & Design, 31(8), 3880-3884.

182 Thangaraju, S., Heilmaier, M., Murty, B. S., & Vadlamani, S. S. (2012). On the estimation of

true Hall–Petch constants and their role on the superposition law exponent in Al

alloys. Advanced Engineering Materials, 14(10), 892-897.

183 Bhadeshia, C. C. H. K. (2001). Manufacturing and microstructural evolution of mechanuically

alloyed oxide dispersion strengthened superalloys. Advanced Engineering Materials, 3(9),

647.

184 Ferkel, H., & Mordike, B. L. (2001). Magnesium strengthened by SiC

nanoparticles. Materials Science and Engineering: A, 298(1), 193-199.

185 Tun, K. S., & Gupta, M. (2007). Improving mechanical properties of magnesium using nano-

yttria reinforcement and microwave assisted powder metallurgy method. Composites Science

and Technology, 67(13), 2657-2664.

186 Goh, C. S., Wei, J., Lee, L. C., & Gupta, M. (2006). Development of novel carbon nanotube

reinforced magnesium nanocomposites using the powder metallurgy

technique. Nanotechnology, 17(1), 7.

187 Mishra, R. S., Ma, Z. Y., & Charit, I. (2003). Friction stir processing: a novel technique for

fabrication of surface composite. Materials Science and Engineering: A, 341(1), 307-310.

188 Mishra, R. S., & Mahoney, M. W. (Eds.). (2007). Friction stir welding and processing. ASM

international.

113

189 Xu, J., Estruga, M., Chen, L., Yahata, B., Choi, H., & Li, X. (2014). Urchin-like AlOOH

nanostructures on Al microspheres grown via in-situ oxide template. Materials Science and

Engineering: B, 188, 89-93.

190 Barmouz, M., Givi, M. K. B., & Seyfi, J. (2011). On the role of processing parameters in

producing Cu/SiC metal matrix composites via friction stir processing: investigating

microstructure, microhardness, wear and tensile behavior. Materials Characterization, 62(1),

108-117.

191 Liu, Q., Ke, L., Liu, F., Huang, C., & Xing, L. (2013). Microstructure and mechanical property

of multi-walled carbon nanotubes reinforced aluminum matrix composites fabricated by

friction stir processing. Materials & Design, 45, 343-348.

192 Dolatkhah, A., Golbabaei, P., Givi, M. B., & Molaiekiya, F. (2012). Investigating effects of

process parameters on microstructural and mechanical properties of Al5052/SiC metal matrix

composite fabricated via friction stir processing. Materials & Design, 37, 458-464.

193 Morisada, Y., Fujii, H., Nagaoka, T., & Fukusumi, M. (2006). Effect of friction stir processing

with SiC particles on microstructure and hardness of AZ31. Materials Science and

Engineering: A, 433(1), 50-54.

194 Chang, C. I., Wang, Y. N., Pei, H. R., Lee, C. J., & Huang, J. C. (2006). On the hardening of

friction stir processed Mg-AZ31 based composites with 5-20% nano-ZrO2 and nano-SiO2

particles. Materials Transactions, 47(12), 2942-2949.

195 Lee, C. J., Huang, J. C., & Hsieh, P. J. (2006). Mg based nano-composites fabricated by

friction stir processing. Scripta Materialia, 54(7), 1415-1420.

114

192 Chen, L. Y., Peng, J. Y., Xu, J. Q., Choi, H., & Li, X. C. (2013). Achieving uniform

distribution and dispersion of a high percentage of nanoparticles in metal matrix

nanocomposites by solidification processing. Scripta Materialia, 69(8), 634-637.

197 Chen, L. Y., Peng, J. Y., Xu, J. Q., Choi, H., & Li, X. C. (2014). Achieving uniform

distribution and dispersion of a high percentage nanoparticles in Mgl8Sn matrix by

solidification processing. Magnesium Technology 2014, 465-470.

198 Nabavi, M., Spalla, O., & Cabane, B. (1993). Surface chemistry of nanometric ceria particles

in aqueous dispersions. Journal of colloid and interface science, 160(2), 459-471.

199 Scheu, C., Dehm, G., Müllejans, H., Brydson, R., & Rühle, M. (1995). Electron energy-loss

near-edge structure of metal-alumina interfaces. Microscopy Microanalysis

Microstructures, 6(1), 19-31.

200 Johnson, K. H., & Pepper, S. V. (1982). Molecular‐orbital model for metal‐sapphire interfacial

strength. Journal of Applied Physics, 53(10), 6634-6637.

201 Sinnott, S. B., & Dickey, E. C. (2003). Ceramic/metal interface structures and their

relationship to atomic-and meso-scale properties. Materials Science and Engineering: R:

Reports, 43(1), 1-59.

202 Lee, L. H. (Ed.). (2013). Fundamentals of adhesion. Springer Science & Business Media.

203 Hamaker, H. C. (1937). The London—van der Waals attraction between spherical

particles. Physica, 4(10), 1058-1072.

204 Lifshitz, E. M. (1956). The theory of molecular attractive forces between solids.

205 Chen, X. J., Levi, A. C., & Tosatti, E. (1991). Hamaker constant calculations and surface

melting of metals. Surface Science, 251, 641-644.

115

206 Klimchitskaya, G. L., Mohideen, U., & Mostepanenko, V. M. (2000). Casimir and van der

Waals forces between two plates or a sphere (lens) above a plate made of real metals. Physical

Review A, 61(6), 062107.

207 Wu, C., & Zhigilei, L. V. (2014). Microscopic mechanisms of laser spallation and ablation of

metal targets from large-scale molecular dynamics simulations. Applied Physics A, 114(1),

11-32.

208 Freitas Jr, R. A. (1999). Nanomedicine, volume I: Basic Capabilities. Landes Bioscience,

Georgetown, TX.

209 Médout-Marère, V., Partyka, S., Chauveteau, G., Douillard, J. M., & Dutartre, R. (2003).

Surface heterogeneity of passively oxidized silicon carbide particles: vapor adsorption

isotherms. Journal of Colloid and Interface Science, 262(2), 309-320.

210 Han, Y. F., Dai, Y. B., Wang, J., Shu, D., & Sun, B. D. (2011). First-principles calculations

on Al/AlB2 interfaces. Applied Surface Science, 257(17), 7831-7836.

211 Egerton, R. (2011). Electron energy-loss spectroscopy in the electron microscope. Springer

Science & Business Media.

212 Knowles, K. M. (2005). Dispersion forces at planar interfaces in anisotropic ceramics. Journal

of Ceramic Processing & Research, 6(1), 10-16.

213 Lye, R. G., & Logothetis, E. M. (1966). Optical properties and band structure of titanium

carbide. Physical Review, 147(2), 622.

214 Polesel-Maris, J., Piednoir, A., Zambelli, T., Bouju, X., & Gauthier, S. (2003). Experimental

investigation of resonance curves in dynamic force microscopy. Nanotechnology, 14(9),

1036.

215 Mittal, K. L. (1988). Particles on Surfaces 2. New York: Plenum Press.

116

216 Rhee, S. K. (1970). Wetting of AlN and TiC by liquid Ag and liquid Cu. Journal of the

American Ceramic Society, 53(12), 639-641.

217 Contreras, A., Leon, C. A., Drew, R. A. L., & Bedolla, E. (2003). Wettability and spreading

kinetics of Al and Mg on TiC. Scripta Materialia, 48(12), 1625-1630.

218 Lopez, V. H., & Kennedy, A. R. (2006). Flux-assisted wetting and spreading of Al on

TiC. Journal of colloid and interface science, 298(1), 356-362.

219 Mazaheri, Y., Meratian, M., Emadi, R., & Najarian, A. R. (2013). Comparison of

microstructural and mechanical properties of Al–TiC, Al–B 4 C and Al–TiC–B4C composites

prepared by casting techniques. Materials Science and Engineering: A, 560, 278-287.

220 Kennedy, A. R., & Karantzalis, A. E. (1999). The incorporation of ceramic particles in molten

aluminium and the relationship to contact angle data. Materials Science and Engineering:

A, 264(1), 122-129.

221 Kennedy, A. R., Karantzalis, A. E., & Wyatt, S. M. (1999). The microstructure and mechanical

properties of TiC and TiB2-reinforced cast metal matrix composites. Journal of Materials

Science, 34(5), 933-940.

222 Stone, I. C., & Jones, H. (1997). Effect of cooling rate and front velocity on solidification

micro structure selection in Al-3.5wt.%Fe-0 to 8.5wt.%Si alloys. Materials Science and

Engineering: A, 226, 33-37.

223 Pryds, N. H., & Huang, X. (2000). The effect of cooling rate on the microstructures formed

during solidification of ferritic steel. Metallurgical and Materials Transactions A, 31(12),

3155-3166.

224 Hashim, J., Looney, L., & Hashmi, M. S. J. (2001). The wettability of SiC particles by molten

aluminium alloy. Journal of Materials Processing Technology, 119(1), 324-328.

117

225 Eustathopoulos, N., Nicholas, M. G., & Drevet, B. (Eds.). (1999). Wettability at High

Temperatures (Vol. 3). Elsevier.

226 Zhang, D., Shen, P., Shi, L., Lin, Q., & Jiang, Q. (2010). Wetting and evaporation behaviors

of molten Mg on partially oxidized SiC substrates. Applied Surface Science, 256(23), 7043-

7047.

227 Parsegian, V. A. (2005). Van der Waals forces: a handbook for biologists, chemists, engineers,

and physicists. Cambridge University Press.

118