unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Copyright copy 2006 American Scientific PublishersAll rights reservedPrinted in the United States of America

Journal ofNanoscience and Nanotechnology

Vol 6 651ndash660 2006

High Energy Density Processing of a Free FormNickelndashAlumina Nanocomposite

V Viswanathan1 A Agarwal2 V Ocelik3 J T M De Hosson3 N Sobczak4 and S Seal1lowast1Advanced Materials Process and Analysis Center (AMPAC) Department of Mechanical Materials and Aerospace

Engineering (MMAE) University of Central Florida Eng 381 4000 Central Florida Boulevard Orlando FL 32816 USA2Mechanical and Materials Engineering Florida International University Miami FL 33174 USA

3Department of Applied Physics Materials Science Center and the Netherlands Institute for Metals ResearchUniversity of Groningen Nijenborgh 4 9747 AG Groningen The Netherlands

4Foundry Research Institute Krakow Poland

The development of a free form bulk Nickel reinforced Alumina matrix nano composites using AirPlasma Spray and laser processing has been presented The process consumes less time andrequires further minimal machining and therefore is cost effective The relative differences in usingAPS over laser processing in development of bulk metal-ceramic nanocomposites have been dis-cussed The process intricacies involved during processing such as material specific mandrel selec-tion plasma-particle interaction are highlighted Electroless coating has been used to uniformlydisperse Nickel in alumina matrix as a source material The electroless Ni coated alumina particlesare subjected to both laser processing and Air Plasma Spraying with optimized parameters Con-solidation by laser processing could not be achieved as the laser beam was reflective to Nickel Onthe other hand APS Nindashalumina nanocomposite with a cylindrical shape of 12primeprime ODtimes1primeprime ID times15primeprime

length has been fabricated with minimum or no surface defects HRTEM pictures revealed thenanostructure retention thereby corroborating the fact that bulk nanostructures can be made usingAir Plasma Spray XRD analysis confirmed the phase transformation from alpha alumina to gammaalumina and oxidation of Ni to NiO Subsequent reduction of NiO to metallic nickel using hydrogenatmosphere has also been demonstrated Mechanical properties such as hardness (1025 HV) andfracture toughness (5 MPa m12) for the nanocomposite are presented herein

Keywords Plasma Electroless Coating Nanostructures Residual Stresses Hardness Frac-ture Toughness

1 INTRODUCTION

Metallic particles are often reinforced in the ceramicmatrix for developing materials with high toughness1

magnetic properties2 and fracture strength3ndash6 Al Mo aresome of the metallic particle reinforcements for reinforcingceramic matrices to improve such properties78 Of latesuch metal ceramic composites are extended towards mak-ing one of the components in nanometer dimensions to fur-ther improve the mechanical properties Nanocompositesthe term being used to describe a composite material withat least one of its phases with dimensions in nanometersNanocomposites have shown excellent catalytic properties9

with improved mechanical magnetic properties as com-pared to its bulk counterparts10ndash12 Such metal-ceramicnanocomposites are developed using processes such as

lowastAuthor to whom correspondence should be addressed

hot isostatic pressing11 spark plasma sintering13 laser-based techniques14 processes such as gel casting infiltra-tion followed by pressureless sintering15ndash17 soaking andpulse electric current sintering18 Of importance Ni par-ticulate reinforced alumina composites showed promise inboth mechanical and magnetic applications1112 Particu-lar interest lies in improving high fracture toughness Thestrengthening mechanisms can be attributed to the blunt-ing of cracks due to the presence of nanosized Ni parti-cle uniformly distributed in the matrix The reason behindusing Nickel as reinforcement is its high temperature sta-bility and chemical inertness in comparison to copper andaluminum Having realized the potential of Nindashaluminananocomposites it is essential to appreciate the advantagesand disadvantages of each of the processing techniqueslisted above and to optimize the resource parameters toachieve the desired properties So far an optimum frac-ture strength and fracture toughness value for Nindashalumina

J Nanosci Nanotechnol 2006 Vol 6 No 3 1533-488020066651010 doi101166jnn2006071 651

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

nanocomposite that has been achieved is 980 MPa and4ndash5 MPa middotm12 through a soaking and Pulse Electric Cur-rent Sintering (PECS) technique18

Researchers came across several obstacles such ascoarsening of metallic particles1920 during conventionalsintering processes Since these processes involved hightemperatures and time involved during sintering stagenickel particle coarsening to the tune of at least 5 timesits original particle size was inevitable However consoli-dation of nanocomposites to make a bulk part has alwaysremained a challenge especially in terms of material fabri-cation Chen et al15 have developed Nindashalumina nanocom-posites with a final size of 4times3times40 mm (a plate) after asintering treatment at 1400 C for 2 hours Aharon et alprepared alumina using slip casting followed by infiltrationof Nindashnitrate solution The whole operation took more than35 hours get a specimen of dimension 50times10times10 mm11

It is evident that such processes are time consuming andeven metal dispersion was challenging Besides bulk freeform part is far from reality

In this paper electroless coating a technique conven-tionally used to increase the surface energy of ceram-ics for facilitating melt infiltration of ceramics in moltenmetals has been innovatively used for uniform distribu-tion of Nickel in alumina matrix prior to processing Theeffect of laser and plasma in the processing of free formbulk metal-ceramic composites with nanostructure reten-tion and effect of processing parameters in achieving theproperties are discussed

2 EXPERIMENTAL DETAILS

The material used were high purity alumina (995538 Alundum Norton Materials) powders For nickel elec-troless coating stannous chloride palladium chloride andnickel Sulphate were purchased from the Sigma AldrichChemical Company

21 Electroless Coating of Ni of Alumina Powder

Al2O3 powder with 15ndash60 microns size range and 9995purity were used for coating Alumina powders wereimmersed into acid stannous chloride (SnCl2) solutionsfor sensitization of Al2O3 particles Following sensitiza-tion the powders were immersed into the acid palladiumchloride (PdCl2) solutions in which Pd2+ reacted withSn(OH)Cl to form Pd nuclei on the Al2O3 surfaces Theelectroless bath composition was Nickel Sulphate contain-ing sodium hypophosphite as a reducing agent A flowchart indicating the chemistry of the electroless coatingprocess is presented in Figure 1 Using the electroless coat-ing technique alumina particles were coated with a 30 nmthick nickel layer which corresponds to 3ndash4 wt of Niloading in the alumina matrix The morphologies of theas-coated powders were characterized through ScanningElectron Microscopy (SEM JEOL 6400F) The thickness

Activation of Alumina surfaceSn2+ + Pd+ Sn4++ Pd

Deposition of Ni on activated Al2O3 powder

Ni 2++ 2endash Ni

H2PO2 + 3OHndash

HPO32ndash + 2 H2O + 2endashndash

Evolution of electrons in electroless bath

Fig 1 The chemistry of Electroless Ni coating on Al2O3

of the electroless coated Nickel Layer is characterizedthrough Focused Ion Beam (FEI 200 TEM-FIB System)The chemical compositions of the coated powders weredetermined by using a PHI 5400 XPS with 12536 eVenergy of Mg-K X-rays at a power of 350 W

22 Consolidation Techniques

221 Laser Processing

Laser processing was carried out in an argon protectiveatmosphere where 1 mm thick preplaced coated powderlayer were illuminated by a continuous Rofin Sinar 2 kWNdYAG laser beam (106 m wavelength) at two dif-ferent regimes First the preplaced powder was scanned(167 mms) with a laser beam close to its focal point(+10 mm) using four different constant laser power 250300 350 and 400 W The laser beam has a circular shapewith a Gaussian power density distribution and the beamdiameter (87 of laser energy inside the particle diameter)is 20 mm In the second regime the preplaced powder wasilluminated for 03 s by the highly defocused laser beam(+50 mm 8 mm beam diameter) with a laser power of200 W Illuminated samples were observed using a PhilipsXL30 FEG scanning electron microscope using both SEand BSE detectors

222 Plasma Processing

A cylindrical mandrel made of 6061 aluminum was usedto deposit the Ni-coated aluminum oxide powder Themachined mandrel had a surface finish of 05 micron Themandrel was degreased and thoroughly cleaned in acetoneprior to plasma spraying Plasma spraying of the pow-der mixture was carried out using Praxair surface tech-nologies (Indianapolis IN) SG 100 plasma spray systemArgon was used as the primary gas and Helium was usedas a secondary gas Plasma spray parameters used for the

652 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table I Plasma spray parameters for fabrication of NindashAluminabulk part

Parameter Value

Current (amperes) 800Voltage (Volts) 35Primary gas flow rate (SCFH) 80Standoff distance (inch) 3 inchRotational speed of the substrate (rpm) 25

Temperature of the substrate C 450ndash500

process are mentioned in Table I Plasma spray param-eters were controlled to minimize the coarsening and melt-ing of electroless coated nano Ni layer The mandrel hasbeen actively cooled during plasma spraying to prevent theparticles from getting coarsened The powder was spraydeposited to a cylinder of wall thickness 3 mm Afterspray deposition mandrel was cooled with liquid nitrogento facilitate the release of the spray formed nanocompos-ite shell due to difference in the coefficient of thermalexpansion values between the mandrel and the depositCharacterization The spray deposited Nickel rein-

forced Alumina nanocomposite has been characterizedusing Scanning Electron Microscopy Focused Ion BeamX-ray Photoelectron Spectroscopy (XPS) and High Reso-lution Transmission Electron Microscopy (HRTEM) Theuniform dispersion of Nickel in Alumina is studied usingJEOL 6400F Field Emission Scanning Electron Micro-scope (SEM) Due to the possibility of formation ofNickel oxide during spray the sprayed bulk part has beenanalyzed for the presence of Nickel oxides using X-rayDiffraction (XRD) Also the residual stresses developedin the nanocomposite have been analyzed using the XRDtechnique HMV-2 Shimadzu model microhardness testerhas been used to determine the hardness and hardness ofthe sprayed part Fracture toughness was calculated bymeasuring the length of the crack at the tip of the inden-tation obtained using a load of 196 N for 15 seconds andsubstitution of the crack length measured in the equationmentioned by Evans and Charles21 Density of the bulknanocomposite has been determined using Archimedesdisplacement method A FEI 200 TEM-FIB (Focused IonBeam) equipped with a 25ndash50 KV gallium liquid metalion source (LMIS) was used for preparing thin TEM speci-mens High Resolution Transmission Electron Microscopy(HRTEM) using FEI Tecnai F30 Transmission ElectronMicroscope is used for revealing the finest details withbetter resolution

3 RESULTS AND DISCUSSION

31 Electroless Coating

It is imperative that the synthesis of good composite pow-ders needs to be efficient and cost effective as well asachieving the desired properties The control of the micro-structure of ceramic-metal composites is generally difficult

840845850855860865870875880885890

Binding Energy (eV)

Inte

nsity

(A

U)

2p32

2p12

Fig 2 XPS spectrum on Ni coating confirming metallic Ni peaks

to achieve by traditional techniques involving mechanicalmixing of ceramic and metallic powders followed by hotpressing2223 A small-scale homogeneity can be obtainedusing the solndashgel route24ndash26 However the relatively highcost of some reactants and the difficulty to control the geldrying step are some of the drawbacks to this methodElectroless coating has been chosen as it is being perceivedas an efficient method by which Nickel can be uniformlyreinforced in to the alumina matrix to increase the mechan-ical properties of the composite The chemistry of theformation of the Nickel layer uniformly over the surfaceof alumina is shown in Figure 1 SnCl2 hydrolyzes intoSn(OH)Cl and be absorbed on the Al2O3 particles followedby Pd2+ reaction with Sn(OH)Cl to form Pd nuclei on theAl2O3 surfaces This renders the alumina surface suitablefor Nickel coating The electroless bath containing Sodiumhypophosphite and ammonium hydroxide react together toleave electrons which sufficient to reduce Nickel chlorideto Nickel The Nickel thus formed gets coated on to thesensitized and activated alumina particles The thicknessof the electroless coated Ni layer over alumina particlescould be revealed using Focused Ion Beam XPS analy-sis on the layer presented in Figure 2 revealed peaks atbinding energy values of 8528 eV and 8707 eV Thesepeaks were confirmed to be of 2p32 and 2p12 of metal-lic Nickel respectively27 The binding energy shift due tothe charging was removed by referencing the adventitiousC(1s) binding energy value28 at 2846 eV

32 Concept of Wettability

The rationale behind the formation of spherical dropletsof Nickel can be understood from the point of view ofwettability As the Nickel particles attain the liquid stateinstead of complete wetting alumina it makes the pointcontact due to surface tension constraints as shown inFigures 3 4 Contact angle is decided by the energies ofthe solidndashliquid liquidndashvapor and solidndashvapor interfacesinvolved in wetting process The contact angle is highin case of pure metals because of the high solidndashliquid

J Nanosci Nanotechnol 6 651ndash660 2006 653

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Ni Layer

Laser Irradiation

Ni Spheroids

Fig 3 Schematic illustration involving transformation of Ni layer in tospheroids

interfacial energy than the solid vapor interfacial energy Itis being reported29 that the solidndashliquid interfacial energyfor NindashAlumina couple is 186 Jm2 while the solidndashvaporinterfacial energy is calculated as 09 Jm2 (Ref [30]) Ses-sile drop test conducted to measure the wettability anglerevealed the contact angle to be at 108 If the wetta-bility is less the area immediately near the droplet willhave no or little amount of the droplet composition whilefarther away from the droplet composition will be appre-ciable This is because of a combination of the so calledldquoevaporation-depositionrdquo processes occurring during theprocess31 The bonding of Nickel on the Alumina surfacecan be comprehended based on interfacial strength At theinterface a polycrystalline surface is prone to have moregrain boundaries as a result of which roughness is highDue to this high roughness factor the accommodation fac-tor for a Nickel atom to get bonded with the aluminasurface is more compared to that of a smooth interfaceRecent results show the presence of interfacial cracks gen-erated as a result of poor bonding of Nickel with a singlecrystal alumina surface compared to a polycrystalline alu-mina surface as given in Table II

321 Validation of Self-Affine Roughness Model

Saiz et al29 determined the contact angle for NindashAluminasystem to be 130 with the surface energy values oflv sl sv plugging these values in to the Youngrsquos equa-tion But the experimental values measured were found tobe 108 in this study Considering the self-affine roughnessmodel32 the apparent contact angle will be differentfrom the contact angle calculated using the Youngrsquos

solid substrate

liquid drop

σlg

σsv σ ls Θ

gaseous atmosphere

Fig 4 Nickel prefers to make a point contact in Alumina rather thancompletely wetting

Table II Contact angle measurements of metallic Nickel on Alumina

Roughness (nm)Dimensions Contact

Substrate (mm) angle (deg) Ra Rq Rt

Al2O3PC (9999) 108 100 140 1190Al2O3SC (999993) 94 2 25 8

equation29

cos = svminussl

lvThis can be explained based on the roughness of the sur-face of the sample Surface roughness is described as asine wave with an rms amplitude 13 and in plane sur-face roughness correlation length which is the averagedistance between the consecutive hills or valleys on thesurface Rough surfaces are usually a combination of cir-cular grooves and radial grooves The apparent angle iscalculated based on the root mean square value of apparentcontact angles g and w which are contributions from theradial and circular grooves respectively If the amount ofcircular grooves is much higher than the radial grooves atthe point of contact angle measurement apparent contactangle w will be dominating31

w cosminus1

[1+

(12

)minus

(38

)2

]cos

where is the root mean square local slope which canbe approximated to 13H H is the roughness exponentHigher the value of H the higher is the roughness of thesurface For theoretical values of which was 130 solv-ing the above quadratic equation gives the value of 110

for the apparent contact angle which agrees closely withthe experimental values Values assumed for 13 (ampli-tude) (distance between two consecutive hills or val-leys) H (hardness exponent) are taken as 23 nm 60 nmand 06 respectively because alumina being a polycrys-talline and the amount of grain boundaries present aremuch higher compared to a single crystalline material

As it is evident from the table the roughness factorplays a dominant role in Wettability A cross section of thesessile drop sample with nickel droplet on top of aluminasubstrate is shown in Figure 5 EDS spectra from the areajust near the Ni droplet showed that there is no appreciableNi concentration This experimentally proves that wetta-bility of nickel is low31 Such uniformly dispersed Nickelparticles are expected to inhibit crack propagation in thebrittle alumina matrix

33 Laser Processing

Thermal energy given to the sample during laser process-ing depends on the power of the laser used The inter-action time in the first regime can be estimated as thetime for which the laser beam illuminates one point dur-ing fast scanning (sim12 ms) The absorption coefficient formetallic surfaces and NdYAG laser wavelength increases

654 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

t1=0 t2

t3 t4

Ni

Alumina

Fig 5 Illustration of wettability of Nickel on Alumina with time t1 ltt2 lt t3 lt t4

with surface roughness temperature and with an angle ofincidence of the beam33 Typical values lie in an interval25ndash35 for room temperature polished surfaces and per-pendicular beam Absorption takes place in a surface layerapproximately 10minus4 mm in thickness

At different power of the laser beam the morphologyof the Nickel showed variation which is shown in SEMmicrograph in Figures 6 through 8 Due to the irradiationof laser beam electroless Ni layer covering the surface ofalumina melts instantaneously and due to surface tensiongets converted to globules The higher the amount of ther-mal energy the more is the driving force for the globulesto get coarsened The fact that each alumina particle has itsown characteristic size of Ni globules suggests that thereis not enough thermal contact between individual particlesto equilibrate different surface temperatures The sizes ofthe Ni globules observed by SEM ranges between 01 and05 m Some of the alumina particle surfaces irradiated

Fig 6 Ni particle size at a laser power of 250 W

Fig 7 Ni particle coarsens due to increase in the power of the laserbeam to 350 W

with high laser power is free of Ni globules but large Niglobules (sim5 m) are observed in their neighborhoods

Laser processing is actually advantageous from the factthat power of laser beam can be manipulated to get tothe suitable microstructure The presence of Ni particlesas small as 5 nm was revealed in HRTEM analysis of hightemperature processed Ni-coated alumina Lattice fringesof Ni nanoparticles could be resolved in the HRTEM studyas shown in Figure 9 The corresponding EDS spectrum(Fig 9b) analyzed from the selected area confirms thepresence of Ni Al and Oxygen Other elements that aredetected are either from the specimen grid or the Si waferused for specimen preparation However consolidation ofNialumina to bulk composite fabrication was not achievedsince the whole laser beam energy is absorbed in a thin Nilayer and a fast transformation of Ni layer in to spheroidsoccurred Therefore plasma processing route was chosento manufacture a free form bulk nanocomposite part

Fig 8 Ni particle further coarsens at a higher laser power of 400 W

J Nanosci Nanotechnol 6 651ndash660 2006 655

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

(a)

(b)

Fig 9 (a) HRTEM picture revealing the nanostructure retention of NindashAlumina nanocomposite (b) EDS spectrum confirming the presence ofNi in Alumina matrix

34 Plasma Processing

The rationale and science behind using plasma spray as afree form manufacturing technique and aluminum mandrelas a substrate for development of NindashAlumina bulk parthas been discussed in this section

341 PlasmandashParticle Interaction

Plasmandashparticle interaction results in the formation ofatoms in excited state ground state ions and photons inthe plasma plume As a result of these atomic dissocia-tion and recombination processes particle that is sprayedis taken to a partial molten state for a brief period of timetill it splats on to the substrate Time of flight the timerequired for the particle to reside in the flame depends onmaterial parameters such as the latent heat of fusion melt-ing temperature particle size The relation to determineresidence time can be given as34

timinusf =CplnTi minusTsminus lnTm minusTs

h middotAs

Table III Particle radius as a function of dwell time

Particle radius Dwell time (Sec)

50 micron 401times10minus4

30 micron 112times10minus4

20 micron 744times10minus5

1 micron 372times10minus6

50 nm 186times10minus7

where Cp Ti Tm h As are the density specificheat capacity volume initial temperature temperature ofthe plasma plume melting point heat transfer coefficientsurface area of the particle and Ts is the plasma plumetemperature The knowledge of residence time is essentialbefore spraying because too much of over heat may leadto evaporation losses Also higher the amount of super-heating the particles are subjected to rapid cooling as itimpinges on the substrate as a result of which shrinkagestresses will be more This might lead to cracking of thebulk part A calculation (Table III) for residence time tf fordifferent particle size NindashAlumina system has been tabledIt can be inferred that smaller particles need less residencetime which can controlled with the parameters mentionedabove To avoid evaporation and to ensure only surfacemelting appropriate gas flow combinations can be selectedto obtain the residence time of interest The residence timecalculated was 02 microseconds for a 50 nm particle topass through

342 Coating vs Free FormsmdashEngineering Aspects

The section describes the science behind the fabricationof a free form part which is different from coating Alu-minum and Copper being good conductors of heat areideal candidate materials for being used as a substrate formaterial deposition Besides they can be fabricated to anydesired shape as a result of their malleability and the highcoefficient of thermal expansion facilitates easy removalof the sprayed deposit Aluminum substrate can facilitateeasy removal by shrinking more compared to steel or cop-per Cooling and solidification of most materials is accom-panied by contraction or shrinkage As particles strike theyrapidly cool and solidify This generates a tensile stresswithin the particle and a compressive stress within the sur-face of the substrate As the coating builds up the ten-sile stresses developed in the deposit also increases Theresidual stress developed during the air plasma spray pro-cess significantly influences the integrity of the depositStresses developed during spraying can be classified into two parts viz quenching stresses and cooling stressesQuenching stresses are stresses that develop as the splatformation takes place due to particle impingement on thesubstrate The quantification of quenching stresses can bedone using3536

q =Eststs+54 middot13td

6 middot13R middot13tdand = Ed

Es

656 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

0

5

10

15

20

25

30

35

40

Alumina

Zircon

ia

Titania

MoS

i 2

Si 3N 4

Aluminu

m

Nickel

Titaniu

m

Coppe

r

Material

Str

ess

(MP

a)

Fig 10 Quenching stresses developed during the spray of differentparticles

where E is the elastic modulii of the substrate and thedeposit indicated by subscripts s and d respectively 13tis the thickness of the deposit and 13R is the change inthe radius of curvature of the coating The plots as shownin Figure 11 depict that different materials will developdifferent quenching stresses as the spray is going onBased on the thickness of the deposit achieved the quench-ing stresses will vary Among various spray materials forwhich quenching stresses have been calculated (Fig 10)The histogram shown for alumina and nickel particles withdeposit thickness of 0015 mm (15 microns) is of interestto this paper Cooling stresses are stresses that the depositproduces when the spraying is completed and the tempera-ture of the part is plummeting to room temperature It canbe accounted for using an equation37

c =EddminussTdminusTRT

1+2Ed middot tdEsts$

where EdmdashYoungrsquos modulus of the deposit TdmdashTemperature of the deposit TRTmdashRoom temperature ofthe sample after cooling dmdashCoefficient of thermalexpansion of the deposition smdashCoefficient of thermalexpansion of the substrate EsmdashYoungrsquos modulus of thesubstrate tsmdashthickness of the substrate and tdmdashthicknessof the deposition Figure 11 depicts the calculated cooling

ndash120

ndash100

ndash80

ndash60

ndash40

ndash20

0Alumina Zirconia Titania MoSi2 Si3N4

Coo

ling

Str

ess

(MP

a)

Copper

Aluminum

Fig 11 Cooling stresses developed in the deposit for Cu and Alsubstrates

stresses for different materials for two different substratesCopper and Aluminum The total amount of stresses devel-oped in the deposit is a combination of both quenchingand cooling stresses Quenching stresses which are ten-sile in nature keeps increasing with thickness build-upThe thermal stresses (cooling stresses) will be compres-sive when d lt s Selection of aluminum mandrel forNindashAlumina is justified based on the calculation that thecooling stresses are compressive in nature It will facilitatethe removal of spray deposit from the substrate This willeventually lead to free form fabrication of deposits

343 Material Characterization

The trial for spraying nano ceramic particles throughplasma flame has been successful The NindashAlumina cylin-drical bulk nanocomposite without any surface defectswas manufactured using plasma spray technique (Fig 12)Microstructure evaluation using SEM and other character-ization tools such as XRD and HRTEM is being discussedat lengthCross-Section Microscopy In an effort to study the

porosity distribution across the thickness scanning elec-tron microstructure of the internal center and outside ofthe thickness (Fig 13) has been analyzed The section is96ndash98 dense and does not show any significant cracksor porosity It also reveals the formation of splats due tohigher velocity with which the melted particles impingeson to the substrate There was no evidence of intercon-nected pores which could have been detrimental to theintegrity of the part Also the porosity distribution is morein case of internal thickness and less in outside thicknessThis can be explained from the fact that particles temper-ature gets reduced drastically during the initial stages ofspray as a result of which shrinkage is more while theparticles at the outside thickness has to undergo relativelyless temperature difference SEM micrographs of the crosssection revealed a clear demarcation between the forma-tion of folds and the formation of porosity The porositiesare revealed as dark regions compared to less intense foldsFold formation happens due to the inability of particles tofully splat on top of the existing splat

Substrate incompression

Free form partin Tension

(a) (b)

Fig 12 (a) The mandrel is subjected to compressive stresses whilethe sprayed part acquires tensile stress as the spray continues (b) A12primeprimeODtimes1primeprimeIDtimes15primeprime length cylinder was produced with full integrity

J Nanosci Nanotechnol 6 651ndash660 2006 657

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Folds

Isolated pores

Folds

(a)

(b)

(c)

Fig 13 Cross sectional low magnification micrographs indicating theextent of porosity formation at the (a) external (b) middle and (c) inter-nal portions of the bulk nanocomposite

344 Structural Characterization

Phase Transformation It is being observed that thealumina matrix undergone phase transformation fromalpha to gamma structure The alpha to gamma transi-tion depends on the ratio of the gun power to the primarygas flow in Standard Cubic Feet per Hour (SCFH) It ismentioned elsewhere [shaw] that if the parent phase is -alumina based on this ratio mentioned above the product

2θ20 30 40 50 60 70 80

(c)

(b)

(a)

NickelAlpha AluminaGamma AluminaNiO

Fig 14 Phase transformations associated with plasma spray of Ni-Al2O3 (a) Feedstock powder (b) plasma sprayed component (c) hydro-gen reduction

phase might be a combination of more -alumina and less alumina if the ratio is less than or equal to 240 or a com-bination of more alumina and less -alumina if the ratiois more than or equal to 310 The density and hardness val-ues achieved38 with the above ratios show significant dif-ferences with the former showing less hardness and densitythan the latter The phase transformation achieved in ourpresent case had more alumina The spray parameterswere manipulated to keep the ratio above 310 to makesure that dense and hard nanostructures were obtained Thediffractograms before and after plasma spraying is shownin Figure 14 X-ray diffraction spectrums from the plasmaspray deposition and powder mixture prior to the spray-ing are compared Both the diffractograms reveal the pres-ence of FCC Ni and Alumina The difference betweenthe diffractograms is that of the peak broadening Dur-ing the spray particles literally gets quenched from theirmelting temperatures towards the substrate preheat temper-ature Such large temperatures gradients are expected toimpart residual stresses on the sprayed part The stressescan be calculated based on the difference in ldquodrdquo spacingof the particles prior to spraying and after spraying Fromthe modulus of elasticity values the amount of residualstress can be calculated as follows39

Eminus1hkl =

C11 +C12

C11 minusC12C11 +2C12

+(

1C44

minus 2C11 minusC12

)l2m2 +m2n2 +n2l2

The compliance constants are C11 = 2465 GPa C12 =1473 GPa and C44 = 1247 GPa The calculated residualstress values are presented in Table IV

Since the process has been carried out in atmosphericair Ni has the propensity to get converted to its surfaceoxide Gibbs free energy of formation of Nickel oxide upto melting point of Nickel is negative Although Gibbs free

658 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table IV Residual stress measurements using X-ray diffraction techniques

hkl 2 (Measured) 2 (JCPDS) d (Measured) Aring d (JCPDS) Aring Strain () E (GPa) Stress (Mpa)

NiO 200 4326 4329 209 209 0066 136 90NiO 111 3725 3728 241 241 0078 304 236NiO 220 6289 6292 148 148 0043 232 99-Alumina 440 67 67032 140 139 0042 408 172

energy curve for formation of Nickel oxide is sloped upwith temperature the partial pressure of oxygen at NTP ismuch higher for the reaction to proceed forward and formNiO APS process carried out at dehumidified air will havean oxygen partial pressure of 02 atm which is more thansufficient for the NiO formation to occur Moreover if thestand off distance is more Nickel in its molten state picksup more oxygen from the atmosphere than it would haveif it were in solid state By reducing the stand off distancethe amount of formation of Nickel oxide can be reducedbut can not be totally eliminated XRD peaks confirmedthe formation of NiO An effort has been made to bringNiO back to its metallic state The plasma spray formednanocomposite was reduced in an atmosphere of Hydrogenat 1000 C for 1 hour

345 High Resolution TransmissionElectron Microscopy

Alumina grain boundaries being high energy regions thereis more probability of the presence of Nickel along thegrain boundaries In order to determine the location ofNickel EDS has been performed along the grain bound-aries to select spots for imaging of Nickel nano parti-cles In those spots where strong Nickel signal have beendetected STEM imaging has been done to determine themorphology and the size of the Nickel particles Spheri-cal Ni particles with 30 nm size have been imaged andare shown in Figure 15 Another evidence of nanostruc-ture retention of Nickel particle is the SAD pattern whichshows a spotty ring pattern as revealed in Figure 15(a)(inset) This pattern is a result of extremely fine grainsThe possible explanations of the nano structure retentionin NindashAlumina nanocomposite could be the lack of timepermitted for the coarsening of Nickel particles travers-ing through the high temperature zone (plasma flame) On

0100200300400500600700800900

1000

0 2000 4000 6000 8000 10000

Energy (eV)

Inte

nsity

(A

U)

NiCu

Al

O

20 nm

Fig 15 STEM pictures depicting nanostructure retention of nickel par-ticles and the corresponding EDS spectrum confirming the presence ofnickel

the other hand during conventional sintering a significanttime is spent in the temperature range where the coarsen-ing effect is appreciable20 Besides in the nanostructuressurface melting is preferred in the fast translation thuskeeping the core intact From the TEM pictures of plasmasprayed NickelndashAlumina it is envisioned that plasma pro-cessing is considered to be a potential tool to fabricate freeform (eg cylindrical) bulk composites with nanostructureretention and improved mechanical properties

35 Mechanical Properties

Nickel plays a dominant role in increasing the strengthof the ceramic matrix The presence of cracks in ceram-ics makes them more brittle compared to metals At thetip of the crack as shown in Figure 16 there is an enor-mous amount of stress concentration developed as a resultof which the cracks tend to propagate easily The surfaceenergy required to propagate the crack is more than com-pensated by the release of the elastic strain energy Thepresence of Nickel spheroids dispersed uniformly through-out the matrix blunts the crack thereby making it more dif-ficult to propagate as shown in Figure 16 Nickel particleswhich are highly plastic in nature aids in crack bluntingand improvement of fracture toughness of the nanocom-posite Lieberthal et al16 proposed a model in which acrack propagating along the alumina grain boundaries aredeflected along around the Ni particle located in grainboundaries which aid in improving the fracture toughnessof the ceramic nanocomposite

The hardness and fracture toughness of the NindashAluminaplasma sprayed sample that could be achieved with thegiven set of plasma spray parameters were measured tobe 1025 HV and sim5 MPa middot m12 The hardness valuesachieved were found to be better than that achieved for the

Crack tip in a ceramicmatrix

Ni-particle

Fig 16 Nickel particles blunt the crack and arrest crack propagationimproves the fracture toughness

J Nanosci Nanotechnol 6 651ndash660 2006 659

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

monolithic alumina developed using plasma spray (940plusmn20 HV)40 The marked increase in the fracture toughnesscompared to other processing techniques can be attributedto the crack blunting capability of ductile Nickel It shouldbe noted that NindashAlumina nanocomposites made throughtechniques such as hot press sintering12 reported a frac-ture toughness of sim35 MPa middotm12 In other words the increase in fracture toughness was approximately 40

4 CONCLUSIONS

Electroless Nickel coated alumina particles have been suc-cessfully used to manufacture a free-form bulk nanocom-posite with improved hardness and fracture toughnessWettability study of Nickel on polycrystalline Aluminarevealed the contact angle which validated the self affineroughness model being used to calculate the apparent con-tact angle Laser processing had shortcomings from thepoint of view of time and the high laser required to conso-lidate the powders to bulk form especially in the casemetalceramic matrix For plasma processing substratematerial selection is important based on the material com-position to be sprayed Calculation of residual High frac-ture toughness can be attributed to the nanostructuredNickel which arrests the crack propagation Thus plasmaspray processing is an effective technique to make freeform bulk nanostructures

Acknowledgments The authors wish to thank theOffice of Naval Research Young Investigator Award (ONRYIP-N000140210591) for their financial support Theauthors would also like to thank Plasma Processes Incfor their help in product development We also appreci-ate the equipment support from Materials CharacterizationFacility at the University of Central Florida

References and Notes

1 W H Tuan and R J Brook J Eur Ceram Soc 6 31 (1990)2 A Paesano Jr C K Matsuda J B M Da Cunha M A Z

Vasconcellos B Hallouche and S L Silva J Magn Magn Mater264 264 (2003)

3 T Sekino and K Niihara J Mater Sci 32 3943 (1997)4 T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)5 A Nakahira and K Niihara Fract Mech 9 165 (1992)6 Y K Jeong and K Niihara Nanostruct Mater 9 193 (1997)7 S C Wang and W C J Wei Nanostruct Mater 10 983 (1998)

8 B Budiansky J C Amazigo and A J Evans J Mech Phys Solids36 167 (1988)

9 S Ozkar G A Ozin and R A Prokopowicz Chem Mater 4 1380(1992)

10 S T Oh T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)11 O Aharon S Bar-Ziv D Gorni T Cohen-Hyams and W D

Kaplan Scripta Mater 50 1209 (2004)12 T Sekino T Nakajima and K Niihara Mater Lett 29 165 (1996)13 S I Cha S H Hong and B K Kim Mater Sci Eng A 351 31

(2003)14 S Seal S C Kuiry P Georgieva and K Rea Scripta Mater 50

1237 (2004)15 R Z Chen and W H Tuan J Eur Ceram Soc 19 463 (1999)16 M Lieberthal and W D Kaplan Mater Sci Eng A 302 83 (2001)17 K Niihara B S Kim T Nakayama T Kusunose T Nomoto

A Hikasa and T Sekino J Eur Ceram Soc 24 3419 (2004)18 U L Adisorn T Matsunaga Y Kobayashi S M Choi and

H Awaji Ceram Intl 31 803 (2005)19 S Seal S C Kuiry P Georgieva and A Agarwal MRS Bull 29

16 (2004)20 J Michalski K Konopka and M Trzaska Mater Chem Phys 81

407 (2003)21 A G Evans and E A Charles J Am Ceram Soc 59 371 (1976)22 W H Tuan and R J Brook J Eur Ceram Soc 10 95 (1992)23 W H Tuan and W B Chou J Am Ceram Soc 80 2418 (1997)24 E Breval G Dodds and C G Pantano Mater Res Bull 20 1191

(1985)25 E Breval Z Deng S Chiou and C G Pantano J Mater Sci 27

1464 (1992)26 E D Rodeghiero O K Tse J Chisaki and E P Giannelis Mater

Sci Eng A 195 151 (1995)27 J F Moulder W F Stickle P E Sobol and K D Bomben

in Handbook of X-Ray Photoelectron Spectroscopy edited byJ Chastain and R C King Jr Physical Electronics Inc Minnesota(1995) p 71

28 T L Barr and S Seal J Vac Sci Technol A 13 1239 (1995)29 E Saiz R M Cannon and A P Tomsia Acta Mater 47 4209

(1999)30 W D Kingery J Am Ceram Soc 37 42 (1954)31 G Palasantzas and J Th M De Hosson Acta Mater 49 3533

(2001)32 N Eustathopolous Acta Mater 46 2319 (1998)33 J A Vreeling V Ocelrsquok Y T Pei D T L Van Agterveld and

J Th M De Hosson Acta Mater 48 4225 (2000)34 R Ye T Ishigaki J Jurewicz P Proulx and M I Boulos Plasma

Chem Plasma Process 24 555 (2004)35 A Brenner and S Senderoff J Res Natl Bur Stand 42 105 (1949)36 S Kuroda and T W Clyne Thin Solid Films 200 49 (1991)37 M W Hunt Guide to engineering materials Adv Mater Proc 158

6 (2000)38 L L Shaw D Goberman R Ren M Gell S Jiang Y Wang T D

Xiao and P R Strutt Surf Coat Technol 130 1 (2000)39 O D Slagle and H A Mckinsky J Appl Phys 38 437 (1967)40 M Vural S Zeytin and A H Ucisik Surf Coat Technol 97 347

(1997)

Received 10 July 2005 Accepted 12 October 2005

660 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

nanocomposite that has been achieved is 980 MPa and4ndash5 MPa middotm12 through a soaking and Pulse Electric Cur-rent Sintering (PECS) technique18

Researchers came across several obstacles such ascoarsening of metallic particles1920 during conventionalsintering processes Since these processes involved hightemperatures and time involved during sintering stagenickel particle coarsening to the tune of at least 5 timesits original particle size was inevitable However consoli-dation of nanocomposites to make a bulk part has alwaysremained a challenge especially in terms of material fabri-cation Chen et al15 have developed Nindashalumina nanocom-posites with a final size of 4times3times40 mm (a plate) after asintering treatment at 1400 C for 2 hours Aharon et alprepared alumina using slip casting followed by infiltrationof Nindashnitrate solution The whole operation took more than35 hours get a specimen of dimension 50times10times10 mm11

It is evident that such processes are time consuming andeven metal dispersion was challenging Besides bulk freeform part is far from reality

In this paper electroless coating a technique conven-tionally used to increase the surface energy of ceram-ics for facilitating melt infiltration of ceramics in moltenmetals has been innovatively used for uniform distribu-tion of Nickel in alumina matrix prior to processing Theeffect of laser and plasma in the processing of free formbulk metal-ceramic composites with nanostructure reten-tion and effect of processing parameters in achieving theproperties are discussed

2 EXPERIMENTAL DETAILS

The material used were high purity alumina (995538 Alundum Norton Materials) powders For nickel elec-troless coating stannous chloride palladium chloride andnickel Sulphate were purchased from the Sigma AldrichChemical Company

21 Electroless Coating of Ni of Alumina Powder

Al2O3 powder with 15ndash60 microns size range and 9995purity were used for coating Alumina powders wereimmersed into acid stannous chloride (SnCl2) solutionsfor sensitization of Al2O3 particles Following sensitiza-tion the powders were immersed into the acid palladiumchloride (PdCl2) solutions in which Pd2+ reacted withSn(OH)Cl to form Pd nuclei on the Al2O3 surfaces Theelectroless bath composition was Nickel Sulphate contain-ing sodium hypophosphite as a reducing agent A flowchart indicating the chemistry of the electroless coatingprocess is presented in Figure 1 Using the electroless coat-ing technique alumina particles were coated with a 30 nmthick nickel layer which corresponds to 3ndash4 wt of Niloading in the alumina matrix The morphologies of theas-coated powders were characterized through ScanningElectron Microscopy (SEM JEOL 6400F) The thickness

Activation of Alumina surfaceSn2+ + Pd+ Sn4++ Pd

Deposition of Ni on activated Al2O3 powder

Ni 2++ 2endash Ni

H2PO2 + 3OHndash

HPO32ndash + 2 H2O + 2endashndash

Evolution of electrons in electroless bath

Fig 1 The chemistry of Electroless Ni coating on Al2O3

of the electroless coated Nickel Layer is characterizedthrough Focused Ion Beam (FEI 200 TEM-FIB System)The chemical compositions of the coated powders weredetermined by using a PHI 5400 XPS with 12536 eVenergy of Mg-K X-rays at a power of 350 W

22 Consolidation Techniques

221 Laser Processing

Laser processing was carried out in an argon protectiveatmosphere where 1 mm thick preplaced coated powderlayer were illuminated by a continuous Rofin Sinar 2 kWNdYAG laser beam (106 m wavelength) at two dif-ferent regimes First the preplaced powder was scanned(167 mms) with a laser beam close to its focal point(+10 mm) using four different constant laser power 250300 350 and 400 W The laser beam has a circular shapewith a Gaussian power density distribution and the beamdiameter (87 of laser energy inside the particle diameter)is 20 mm In the second regime the preplaced powder wasilluminated for 03 s by the highly defocused laser beam(+50 mm 8 mm beam diameter) with a laser power of200 W Illuminated samples were observed using a PhilipsXL30 FEG scanning electron microscope using both SEand BSE detectors

222 Plasma Processing

A cylindrical mandrel made of 6061 aluminum was usedto deposit the Ni-coated aluminum oxide powder Themachined mandrel had a surface finish of 05 micron Themandrel was degreased and thoroughly cleaned in acetoneprior to plasma spraying Plasma spraying of the pow-der mixture was carried out using Praxair surface tech-nologies (Indianapolis IN) SG 100 plasma spray systemArgon was used as the primary gas and Helium was usedas a secondary gas Plasma spray parameters used for the

652 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table I Plasma spray parameters for fabrication of NindashAluminabulk part

Parameter Value

Current (amperes) 800Voltage (Volts) 35Primary gas flow rate (SCFH) 80Standoff distance (inch) 3 inchRotational speed of the substrate (rpm) 25

Temperature of the substrate C 450ndash500

process are mentioned in Table I Plasma spray param-eters were controlled to minimize the coarsening and melt-ing of electroless coated nano Ni layer The mandrel hasbeen actively cooled during plasma spraying to prevent theparticles from getting coarsened The powder was spraydeposited to a cylinder of wall thickness 3 mm Afterspray deposition mandrel was cooled with liquid nitrogento facilitate the release of the spray formed nanocompos-ite shell due to difference in the coefficient of thermalexpansion values between the mandrel and the depositCharacterization The spray deposited Nickel rein-

forced Alumina nanocomposite has been characterizedusing Scanning Electron Microscopy Focused Ion BeamX-ray Photoelectron Spectroscopy (XPS) and High Reso-lution Transmission Electron Microscopy (HRTEM) Theuniform dispersion of Nickel in Alumina is studied usingJEOL 6400F Field Emission Scanning Electron Micro-scope (SEM) Due to the possibility of formation ofNickel oxide during spray the sprayed bulk part has beenanalyzed for the presence of Nickel oxides using X-rayDiffraction (XRD) Also the residual stresses developedin the nanocomposite have been analyzed using the XRDtechnique HMV-2 Shimadzu model microhardness testerhas been used to determine the hardness and hardness ofthe sprayed part Fracture toughness was calculated bymeasuring the length of the crack at the tip of the inden-tation obtained using a load of 196 N for 15 seconds andsubstitution of the crack length measured in the equationmentioned by Evans and Charles21 Density of the bulknanocomposite has been determined using Archimedesdisplacement method A FEI 200 TEM-FIB (Focused IonBeam) equipped with a 25ndash50 KV gallium liquid metalion source (LMIS) was used for preparing thin TEM speci-mens High Resolution Transmission Electron Microscopy(HRTEM) using FEI Tecnai F30 Transmission ElectronMicroscope is used for revealing the finest details withbetter resolution

3 RESULTS AND DISCUSSION

31 Electroless Coating

It is imperative that the synthesis of good composite pow-ders needs to be efficient and cost effective as well asachieving the desired properties The control of the micro-structure of ceramic-metal composites is generally difficult

840845850855860865870875880885890

Binding Energy (eV)

Inte

nsity

(A

U)

2p32

2p12

Fig 2 XPS spectrum on Ni coating confirming metallic Ni peaks

to achieve by traditional techniques involving mechanicalmixing of ceramic and metallic powders followed by hotpressing2223 A small-scale homogeneity can be obtainedusing the solndashgel route24ndash26 However the relatively highcost of some reactants and the difficulty to control the geldrying step are some of the drawbacks to this methodElectroless coating has been chosen as it is being perceivedas an efficient method by which Nickel can be uniformlyreinforced in to the alumina matrix to increase the mechan-ical properties of the composite The chemistry of theformation of the Nickel layer uniformly over the surfaceof alumina is shown in Figure 1 SnCl2 hydrolyzes intoSn(OH)Cl and be absorbed on the Al2O3 particles followedby Pd2+ reaction with Sn(OH)Cl to form Pd nuclei on theAl2O3 surfaces This renders the alumina surface suitablefor Nickel coating The electroless bath containing Sodiumhypophosphite and ammonium hydroxide react together toleave electrons which sufficient to reduce Nickel chlorideto Nickel The Nickel thus formed gets coated on to thesensitized and activated alumina particles The thicknessof the electroless coated Ni layer over alumina particlescould be revealed using Focused Ion Beam XPS analy-sis on the layer presented in Figure 2 revealed peaks atbinding energy values of 8528 eV and 8707 eV Thesepeaks were confirmed to be of 2p32 and 2p12 of metal-lic Nickel respectively27 The binding energy shift due tothe charging was removed by referencing the adventitiousC(1s) binding energy value28 at 2846 eV

32 Concept of Wettability

The rationale behind the formation of spherical dropletsof Nickel can be understood from the point of view ofwettability As the Nickel particles attain the liquid stateinstead of complete wetting alumina it makes the pointcontact due to surface tension constraints as shown inFigures 3 4 Contact angle is decided by the energies ofthe solidndashliquid liquidndashvapor and solidndashvapor interfacesinvolved in wetting process The contact angle is highin case of pure metals because of the high solidndashliquid

J Nanosci Nanotechnol 6 651ndash660 2006 653

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Ni Layer

Laser Irradiation

Ni Spheroids

Fig 3 Schematic illustration involving transformation of Ni layer in tospheroids

interfacial energy than the solid vapor interfacial energy Itis being reported29 that the solidndashliquid interfacial energyfor NindashAlumina couple is 186 Jm2 while the solidndashvaporinterfacial energy is calculated as 09 Jm2 (Ref [30]) Ses-sile drop test conducted to measure the wettability anglerevealed the contact angle to be at 108 If the wetta-bility is less the area immediately near the droplet willhave no or little amount of the droplet composition whilefarther away from the droplet composition will be appre-ciable This is because of a combination of the so calledldquoevaporation-depositionrdquo processes occurring during theprocess31 The bonding of Nickel on the Alumina surfacecan be comprehended based on interfacial strength At theinterface a polycrystalline surface is prone to have moregrain boundaries as a result of which roughness is highDue to this high roughness factor the accommodation fac-tor for a Nickel atom to get bonded with the aluminasurface is more compared to that of a smooth interfaceRecent results show the presence of interfacial cracks gen-erated as a result of poor bonding of Nickel with a singlecrystal alumina surface compared to a polycrystalline alu-mina surface as given in Table II

321 Validation of Self-Affine Roughness Model

Saiz et al29 determined the contact angle for NindashAluminasystem to be 130 with the surface energy values oflv sl sv plugging these values in to the Youngrsquos equa-tion But the experimental values measured were found tobe 108 in this study Considering the self-affine roughnessmodel32 the apparent contact angle will be differentfrom the contact angle calculated using the Youngrsquos

solid substrate

liquid drop

σlg

σsv σ ls Θ

gaseous atmosphere

Fig 4 Nickel prefers to make a point contact in Alumina rather thancompletely wetting

Table II Contact angle measurements of metallic Nickel on Alumina

Roughness (nm)Dimensions Contact

Substrate (mm) angle (deg) Ra Rq Rt

Al2O3PC (9999) 108 100 140 1190Al2O3SC (999993) 94 2 25 8

equation29

cos = svminussl

lvThis can be explained based on the roughness of the sur-face of the sample Surface roughness is described as asine wave with an rms amplitude 13 and in plane sur-face roughness correlation length which is the averagedistance between the consecutive hills or valleys on thesurface Rough surfaces are usually a combination of cir-cular grooves and radial grooves The apparent angle iscalculated based on the root mean square value of apparentcontact angles g and w which are contributions from theradial and circular grooves respectively If the amount ofcircular grooves is much higher than the radial grooves atthe point of contact angle measurement apparent contactangle w will be dominating31

w cosminus1

[1+

(12

)minus

(38

)2

]cos

where is the root mean square local slope which canbe approximated to 13H H is the roughness exponentHigher the value of H the higher is the roughness of thesurface For theoretical values of which was 130 solv-ing the above quadratic equation gives the value of 110

for the apparent contact angle which agrees closely withthe experimental values Values assumed for 13 (ampli-tude) (distance between two consecutive hills or val-leys) H (hardness exponent) are taken as 23 nm 60 nmand 06 respectively because alumina being a polycrys-talline and the amount of grain boundaries present aremuch higher compared to a single crystalline material

As it is evident from the table the roughness factorplays a dominant role in Wettability A cross section of thesessile drop sample with nickel droplet on top of aluminasubstrate is shown in Figure 5 EDS spectra from the areajust near the Ni droplet showed that there is no appreciableNi concentration This experimentally proves that wetta-bility of nickel is low31 Such uniformly dispersed Nickelparticles are expected to inhibit crack propagation in thebrittle alumina matrix

33 Laser Processing

Thermal energy given to the sample during laser process-ing depends on the power of the laser used The inter-action time in the first regime can be estimated as thetime for which the laser beam illuminates one point dur-ing fast scanning (sim12 ms) The absorption coefficient formetallic surfaces and NdYAG laser wavelength increases

654 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

t1=0 t2

t3 t4

Ni

Alumina

Fig 5 Illustration of wettability of Nickel on Alumina with time t1 ltt2 lt t3 lt t4

with surface roughness temperature and with an angle ofincidence of the beam33 Typical values lie in an interval25ndash35 for room temperature polished surfaces and per-pendicular beam Absorption takes place in a surface layerapproximately 10minus4 mm in thickness

At different power of the laser beam the morphologyof the Nickel showed variation which is shown in SEMmicrograph in Figures 6 through 8 Due to the irradiationof laser beam electroless Ni layer covering the surface ofalumina melts instantaneously and due to surface tensiongets converted to globules The higher the amount of ther-mal energy the more is the driving force for the globulesto get coarsened The fact that each alumina particle has itsown characteristic size of Ni globules suggests that thereis not enough thermal contact between individual particlesto equilibrate different surface temperatures The sizes ofthe Ni globules observed by SEM ranges between 01 and05 m Some of the alumina particle surfaces irradiated

Fig 6 Ni particle size at a laser power of 250 W

Fig 7 Ni particle coarsens due to increase in the power of the laserbeam to 350 W

with high laser power is free of Ni globules but large Niglobules (sim5 m) are observed in their neighborhoods

Laser processing is actually advantageous from the factthat power of laser beam can be manipulated to get tothe suitable microstructure The presence of Ni particlesas small as 5 nm was revealed in HRTEM analysis of hightemperature processed Ni-coated alumina Lattice fringesof Ni nanoparticles could be resolved in the HRTEM studyas shown in Figure 9 The corresponding EDS spectrum(Fig 9b) analyzed from the selected area confirms thepresence of Ni Al and Oxygen Other elements that aredetected are either from the specimen grid or the Si waferused for specimen preparation However consolidation ofNialumina to bulk composite fabrication was not achievedsince the whole laser beam energy is absorbed in a thin Nilayer and a fast transformation of Ni layer in to spheroidsoccurred Therefore plasma processing route was chosento manufacture a free form bulk nanocomposite part

Fig 8 Ni particle further coarsens at a higher laser power of 400 W

J Nanosci Nanotechnol 6 651ndash660 2006 655

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

(a)

(b)

Fig 9 (a) HRTEM picture revealing the nanostructure retention of NindashAlumina nanocomposite (b) EDS spectrum confirming the presence ofNi in Alumina matrix

34 Plasma Processing

The rationale and science behind using plasma spray as afree form manufacturing technique and aluminum mandrelas a substrate for development of NindashAlumina bulk parthas been discussed in this section

341 PlasmandashParticle Interaction

Plasmandashparticle interaction results in the formation ofatoms in excited state ground state ions and photons inthe plasma plume As a result of these atomic dissocia-tion and recombination processes particle that is sprayedis taken to a partial molten state for a brief period of timetill it splats on to the substrate Time of flight the timerequired for the particle to reside in the flame depends onmaterial parameters such as the latent heat of fusion melt-ing temperature particle size The relation to determineresidence time can be given as34

timinusf =CplnTi minusTsminus lnTm minusTs

h middotAs

Table III Particle radius as a function of dwell time

Particle radius Dwell time (Sec)

50 micron 401times10minus4

30 micron 112times10minus4

20 micron 744times10minus5

1 micron 372times10minus6

50 nm 186times10minus7

where Cp Ti Tm h As are the density specificheat capacity volume initial temperature temperature ofthe plasma plume melting point heat transfer coefficientsurface area of the particle and Ts is the plasma plumetemperature The knowledge of residence time is essentialbefore spraying because too much of over heat may leadto evaporation losses Also higher the amount of super-heating the particles are subjected to rapid cooling as itimpinges on the substrate as a result of which shrinkagestresses will be more This might lead to cracking of thebulk part A calculation (Table III) for residence time tf fordifferent particle size NindashAlumina system has been tabledIt can be inferred that smaller particles need less residencetime which can controlled with the parameters mentionedabove To avoid evaporation and to ensure only surfacemelting appropriate gas flow combinations can be selectedto obtain the residence time of interest The residence timecalculated was 02 microseconds for a 50 nm particle topass through

342 Coating vs Free FormsmdashEngineering Aspects

The section describes the science behind the fabricationof a free form part which is different from coating Alu-minum and Copper being good conductors of heat areideal candidate materials for being used as a substrate formaterial deposition Besides they can be fabricated to anydesired shape as a result of their malleability and the highcoefficient of thermal expansion facilitates easy removalof the sprayed deposit Aluminum substrate can facilitateeasy removal by shrinking more compared to steel or cop-per Cooling and solidification of most materials is accom-panied by contraction or shrinkage As particles strike theyrapidly cool and solidify This generates a tensile stresswithin the particle and a compressive stress within the sur-face of the substrate As the coating builds up the ten-sile stresses developed in the deposit also increases Theresidual stress developed during the air plasma spray pro-cess significantly influences the integrity of the depositStresses developed during spraying can be classified into two parts viz quenching stresses and cooling stressesQuenching stresses are stresses that develop as the splatformation takes place due to particle impingement on thesubstrate The quantification of quenching stresses can bedone using3536

q =Eststs+54 middot13td

6 middot13R middot13tdand = Ed

Es

656 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

0

5

10

15

20

25

30

35

40

Alumina

Zircon

ia

Titania

MoS

i 2

Si 3N 4

Aluminu

m

Nickel

Titaniu

m

Coppe

r

Material

Str

ess

(MP

a)

Fig 10 Quenching stresses developed during the spray of differentparticles

where E is the elastic modulii of the substrate and thedeposit indicated by subscripts s and d respectively 13tis the thickness of the deposit and 13R is the change inthe radius of curvature of the coating The plots as shownin Figure 11 depict that different materials will developdifferent quenching stresses as the spray is going onBased on the thickness of the deposit achieved the quench-ing stresses will vary Among various spray materials forwhich quenching stresses have been calculated (Fig 10)The histogram shown for alumina and nickel particles withdeposit thickness of 0015 mm (15 microns) is of interestto this paper Cooling stresses are stresses that the depositproduces when the spraying is completed and the tempera-ture of the part is plummeting to room temperature It canbe accounted for using an equation37

c =EddminussTdminusTRT

1+2Ed middot tdEsts$

where EdmdashYoungrsquos modulus of the deposit TdmdashTemperature of the deposit TRTmdashRoom temperature ofthe sample after cooling dmdashCoefficient of thermalexpansion of the deposition smdashCoefficient of thermalexpansion of the substrate EsmdashYoungrsquos modulus of thesubstrate tsmdashthickness of the substrate and tdmdashthicknessof the deposition Figure 11 depicts the calculated cooling

ndash120

ndash100

ndash80

ndash60

ndash40

ndash20

0Alumina Zirconia Titania MoSi2 Si3N4

Coo

ling

Str

ess

(MP

a)

Copper

Aluminum

Fig 11 Cooling stresses developed in the deposit for Cu and Alsubstrates

stresses for different materials for two different substratesCopper and Aluminum The total amount of stresses devel-oped in the deposit is a combination of both quenchingand cooling stresses Quenching stresses which are ten-sile in nature keeps increasing with thickness build-upThe thermal stresses (cooling stresses) will be compres-sive when d lt s Selection of aluminum mandrel forNindashAlumina is justified based on the calculation that thecooling stresses are compressive in nature It will facilitatethe removal of spray deposit from the substrate This willeventually lead to free form fabrication of deposits

343 Material Characterization

The trial for spraying nano ceramic particles throughplasma flame has been successful The NindashAlumina cylin-drical bulk nanocomposite without any surface defectswas manufactured using plasma spray technique (Fig 12)Microstructure evaluation using SEM and other character-ization tools such as XRD and HRTEM is being discussedat lengthCross-Section Microscopy In an effort to study the

porosity distribution across the thickness scanning elec-tron microstructure of the internal center and outside ofthe thickness (Fig 13) has been analyzed The section is96ndash98 dense and does not show any significant cracksor porosity It also reveals the formation of splats due tohigher velocity with which the melted particles impingeson to the substrate There was no evidence of intercon-nected pores which could have been detrimental to theintegrity of the part Also the porosity distribution is morein case of internal thickness and less in outside thicknessThis can be explained from the fact that particles temper-ature gets reduced drastically during the initial stages ofspray as a result of which shrinkage is more while theparticles at the outside thickness has to undergo relativelyless temperature difference SEM micrographs of the crosssection revealed a clear demarcation between the forma-tion of folds and the formation of porosity The porositiesare revealed as dark regions compared to less intense foldsFold formation happens due to the inability of particles tofully splat on top of the existing splat

Substrate incompression

Free form partin Tension

(a) (b)

Fig 12 (a) The mandrel is subjected to compressive stresses whilethe sprayed part acquires tensile stress as the spray continues (b) A12primeprimeODtimes1primeprimeIDtimes15primeprime length cylinder was produced with full integrity

J Nanosci Nanotechnol 6 651ndash660 2006 657

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Folds

Isolated pores

Folds

(a)

(b)

(c)

Fig 13 Cross sectional low magnification micrographs indicating theextent of porosity formation at the (a) external (b) middle and (c) inter-nal portions of the bulk nanocomposite

344 Structural Characterization

Phase Transformation It is being observed that thealumina matrix undergone phase transformation fromalpha to gamma structure The alpha to gamma transi-tion depends on the ratio of the gun power to the primarygas flow in Standard Cubic Feet per Hour (SCFH) It ismentioned elsewhere [shaw] that if the parent phase is -alumina based on this ratio mentioned above the product

2θ20 30 40 50 60 70 80

(c)

(b)

(a)

NickelAlpha AluminaGamma AluminaNiO

Fig 14 Phase transformations associated with plasma spray of Ni-Al2O3 (a) Feedstock powder (b) plasma sprayed component (c) hydro-gen reduction

phase might be a combination of more -alumina and less alumina if the ratio is less than or equal to 240 or a com-bination of more alumina and less -alumina if the ratiois more than or equal to 310 The density and hardness val-ues achieved38 with the above ratios show significant dif-ferences with the former showing less hardness and densitythan the latter The phase transformation achieved in ourpresent case had more alumina The spray parameterswere manipulated to keep the ratio above 310 to makesure that dense and hard nanostructures were obtained Thediffractograms before and after plasma spraying is shownin Figure 14 X-ray diffraction spectrums from the plasmaspray deposition and powder mixture prior to the spray-ing are compared Both the diffractograms reveal the pres-ence of FCC Ni and Alumina The difference betweenthe diffractograms is that of the peak broadening Dur-ing the spray particles literally gets quenched from theirmelting temperatures towards the substrate preheat temper-ature Such large temperatures gradients are expected toimpart residual stresses on the sprayed part The stressescan be calculated based on the difference in ldquodrdquo spacingof the particles prior to spraying and after spraying Fromthe modulus of elasticity values the amount of residualstress can be calculated as follows39

Eminus1hkl =

C11 +C12

C11 minusC12C11 +2C12

+(

1C44

minus 2C11 minusC12

)l2m2 +m2n2 +n2l2

The compliance constants are C11 = 2465 GPa C12 =1473 GPa and C44 = 1247 GPa The calculated residualstress values are presented in Table IV

Since the process has been carried out in atmosphericair Ni has the propensity to get converted to its surfaceoxide Gibbs free energy of formation of Nickel oxide upto melting point of Nickel is negative Although Gibbs free

658 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table IV Residual stress measurements using X-ray diffraction techniques

hkl 2 (Measured) 2 (JCPDS) d (Measured) Aring d (JCPDS) Aring Strain () E (GPa) Stress (Mpa)

NiO 200 4326 4329 209 209 0066 136 90NiO 111 3725 3728 241 241 0078 304 236NiO 220 6289 6292 148 148 0043 232 99-Alumina 440 67 67032 140 139 0042 408 172

energy curve for formation of Nickel oxide is sloped upwith temperature the partial pressure of oxygen at NTP ismuch higher for the reaction to proceed forward and formNiO APS process carried out at dehumidified air will havean oxygen partial pressure of 02 atm which is more thansufficient for the NiO formation to occur Moreover if thestand off distance is more Nickel in its molten state picksup more oxygen from the atmosphere than it would haveif it were in solid state By reducing the stand off distancethe amount of formation of Nickel oxide can be reducedbut can not be totally eliminated XRD peaks confirmedthe formation of NiO An effort has been made to bringNiO back to its metallic state The plasma spray formednanocomposite was reduced in an atmosphere of Hydrogenat 1000 C for 1 hour

345 High Resolution TransmissionElectron Microscopy

Alumina grain boundaries being high energy regions thereis more probability of the presence of Nickel along thegrain boundaries In order to determine the location ofNickel EDS has been performed along the grain bound-aries to select spots for imaging of Nickel nano parti-cles In those spots where strong Nickel signal have beendetected STEM imaging has been done to determine themorphology and the size of the Nickel particles Spheri-cal Ni particles with 30 nm size have been imaged andare shown in Figure 15 Another evidence of nanostruc-ture retention of Nickel particle is the SAD pattern whichshows a spotty ring pattern as revealed in Figure 15(a)(inset) This pattern is a result of extremely fine grainsThe possible explanations of the nano structure retentionin NindashAlumina nanocomposite could be the lack of timepermitted for the coarsening of Nickel particles travers-ing through the high temperature zone (plasma flame) On

0100200300400500600700800900

1000

0 2000 4000 6000 8000 10000

Energy (eV)

Inte

nsity

(A

U)

NiCu

Al

O

20 nm

Fig 15 STEM pictures depicting nanostructure retention of nickel par-ticles and the corresponding EDS spectrum confirming the presence ofnickel

the other hand during conventional sintering a significanttime is spent in the temperature range where the coarsen-ing effect is appreciable20 Besides in the nanostructuressurface melting is preferred in the fast translation thuskeeping the core intact From the TEM pictures of plasmasprayed NickelndashAlumina it is envisioned that plasma pro-cessing is considered to be a potential tool to fabricate freeform (eg cylindrical) bulk composites with nanostructureretention and improved mechanical properties

35 Mechanical Properties

Nickel plays a dominant role in increasing the strengthof the ceramic matrix The presence of cracks in ceram-ics makes them more brittle compared to metals At thetip of the crack as shown in Figure 16 there is an enor-mous amount of stress concentration developed as a resultof which the cracks tend to propagate easily The surfaceenergy required to propagate the crack is more than com-pensated by the release of the elastic strain energy Thepresence of Nickel spheroids dispersed uniformly through-out the matrix blunts the crack thereby making it more dif-ficult to propagate as shown in Figure 16 Nickel particleswhich are highly plastic in nature aids in crack bluntingand improvement of fracture toughness of the nanocom-posite Lieberthal et al16 proposed a model in which acrack propagating along the alumina grain boundaries aredeflected along around the Ni particle located in grainboundaries which aid in improving the fracture toughnessof the ceramic nanocomposite

The hardness and fracture toughness of the NindashAluminaplasma sprayed sample that could be achieved with thegiven set of plasma spray parameters were measured tobe 1025 HV and sim5 MPa middot m12 The hardness valuesachieved were found to be better than that achieved for the

Crack tip in a ceramicmatrix

Ni-particle

Fig 16 Nickel particles blunt the crack and arrest crack propagationimproves the fracture toughness

J Nanosci Nanotechnol 6 651ndash660 2006 659

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

monolithic alumina developed using plasma spray (940plusmn20 HV)40 The marked increase in the fracture toughnesscompared to other processing techniques can be attributedto the crack blunting capability of ductile Nickel It shouldbe noted that NindashAlumina nanocomposites made throughtechniques such as hot press sintering12 reported a frac-ture toughness of sim35 MPa middotm12 In other words the increase in fracture toughness was approximately 40

4 CONCLUSIONS

Electroless Nickel coated alumina particles have been suc-cessfully used to manufacture a free-form bulk nanocom-posite with improved hardness and fracture toughnessWettability study of Nickel on polycrystalline Aluminarevealed the contact angle which validated the self affineroughness model being used to calculate the apparent con-tact angle Laser processing had shortcomings from thepoint of view of time and the high laser required to conso-lidate the powders to bulk form especially in the casemetalceramic matrix For plasma processing substratematerial selection is important based on the material com-position to be sprayed Calculation of residual High frac-ture toughness can be attributed to the nanostructuredNickel which arrests the crack propagation Thus plasmaspray processing is an effective technique to make freeform bulk nanostructures

Acknowledgments The authors wish to thank theOffice of Naval Research Young Investigator Award (ONRYIP-N000140210591) for their financial support Theauthors would also like to thank Plasma Processes Incfor their help in product development We also appreci-ate the equipment support from Materials CharacterizationFacility at the University of Central Florida

References and Notes

1 W H Tuan and R J Brook J Eur Ceram Soc 6 31 (1990)2 A Paesano Jr C K Matsuda J B M Da Cunha M A Z

Vasconcellos B Hallouche and S L Silva J Magn Magn Mater264 264 (2003)

3 T Sekino and K Niihara J Mater Sci 32 3943 (1997)4 T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)5 A Nakahira and K Niihara Fract Mech 9 165 (1992)6 Y K Jeong and K Niihara Nanostruct Mater 9 193 (1997)7 S C Wang and W C J Wei Nanostruct Mater 10 983 (1998)

8 B Budiansky J C Amazigo and A J Evans J Mech Phys Solids36 167 (1988)

9 S Ozkar G A Ozin and R A Prokopowicz Chem Mater 4 1380(1992)

10 S T Oh T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)11 O Aharon S Bar-Ziv D Gorni T Cohen-Hyams and W D

Kaplan Scripta Mater 50 1209 (2004)12 T Sekino T Nakajima and K Niihara Mater Lett 29 165 (1996)13 S I Cha S H Hong and B K Kim Mater Sci Eng A 351 31

(2003)14 S Seal S C Kuiry P Georgieva and K Rea Scripta Mater 50

1237 (2004)15 R Z Chen and W H Tuan J Eur Ceram Soc 19 463 (1999)16 M Lieberthal and W D Kaplan Mater Sci Eng A 302 83 (2001)17 K Niihara B S Kim T Nakayama T Kusunose T Nomoto

A Hikasa and T Sekino J Eur Ceram Soc 24 3419 (2004)18 U L Adisorn T Matsunaga Y Kobayashi S M Choi and

H Awaji Ceram Intl 31 803 (2005)19 S Seal S C Kuiry P Georgieva and A Agarwal MRS Bull 29

16 (2004)20 J Michalski K Konopka and M Trzaska Mater Chem Phys 81

407 (2003)21 A G Evans and E A Charles J Am Ceram Soc 59 371 (1976)22 W H Tuan and R J Brook J Eur Ceram Soc 10 95 (1992)23 W H Tuan and W B Chou J Am Ceram Soc 80 2418 (1997)24 E Breval G Dodds and C G Pantano Mater Res Bull 20 1191

(1985)25 E Breval Z Deng S Chiou and C G Pantano J Mater Sci 27

1464 (1992)26 E D Rodeghiero O K Tse J Chisaki and E P Giannelis Mater

Sci Eng A 195 151 (1995)27 J F Moulder W F Stickle P E Sobol and K D Bomben

in Handbook of X-Ray Photoelectron Spectroscopy edited byJ Chastain and R C King Jr Physical Electronics Inc Minnesota(1995) p 71

28 T L Barr and S Seal J Vac Sci Technol A 13 1239 (1995)29 E Saiz R M Cannon and A P Tomsia Acta Mater 47 4209

(1999)30 W D Kingery J Am Ceram Soc 37 42 (1954)31 G Palasantzas and J Th M De Hosson Acta Mater 49 3533

(2001)32 N Eustathopolous Acta Mater 46 2319 (1998)33 J A Vreeling V Ocelrsquok Y T Pei D T L Van Agterveld and

J Th M De Hosson Acta Mater 48 4225 (2000)34 R Ye T Ishigaki J Jurewicz P Proulx and M I Boulos Plasma

Chem Plasma Process 24 555 (2004)35 A Brenner and S Senderoff J Res Natl Bur Stand 42 105 (1949)36 S Kuroda and T W Clyne Thin Solid Films 200 49 (1991)37 M W Hunt Guide to engineering materials Adv Mater Proc 158

6 (2000)38 L L Shaw D Goberman R Ren M Gell S Jiang Y Wang T D

Xiao and P R Strutt Surf Coat Technol 130 1 (2000)39 O D Slagle and H A Mckinsky J Appl Phys 38 437 (1967)40 M Vural S Zeytin and A H Ucisik Surf Coat Technol 97 347

(1997)

Received 10 July 2005 Accepted 12 October 2005

660 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table I Plasma spray parameters for fabrication of NindashAluminabulk part

Parameter Value

Current (amperes) 800Voltage (Volts) 35Primary gas flow rate (SCFH) 80Standoff distance (inch) 3 inchRotational speed of the substrate (rpm) 25

Temperature of the substrate C 450ndash500

process are mentioned in Table I Plasma spray param-eters were controlled to minimize the coarsening and melt-ing of electroless coated nano Ni layer The mandrel hasbeen actively cooled during plasma spraying to prevent theparticles from getting coarsened The powder was spraydeposited to a cylinder of wall thickness 3 mm Afterspray deposition mandrel was cooled with liquid nitrogento facilitate the release of the spray formed nanocompos-ite shell due to difference in the coefficient of thermalexpansion values between the mandrel and the depositCharacterization The spray deposited Nickel rein-

forced Alumina nanocomposite has been characterizedusing Scanning Electron Microscopy Focused Ion BeamX-ray Photoelectron Spectroscopy (XPS) and High Reso-lution Transmission Electron Microscopy (HRTEM) Theuniform dispersion of Nickel in Alumina is studied usingJEOL 6400F Field Emission Scanning Electron Micro-scope (SEM) Due to the possibility of formation ofNickel oxide during spray the sprayed bulk part has beenanalyzed for the presence of Nickel oxides using X-rayDiffraction (XRD) Also the residual stresses developedin the nanocomposite have been analyzed using the XRDtechnique HMV-2 Shimadzu model microhardness testerhas been used to determine the hardness and hardness ofthe sprayed part Fracture toughness was calculated bymeasuring the length of the crack at the tip of the inden-tation obtained using a load of 196 N for 15 seconds andsubstitution of the crack length measured in the equationmentioned by Evans and Charles21 Density of the bulknanocomposite has been determined using Archimedesdisplacement method A FEI 200 TEM-FIB (Focused IonBeam) equipped with a 25ndash50 KV gallium liquid metalion source (LMIS) was used for preparing thin TEM speci-mens High Resolution Transmission Electron Microscopy(HRTEM) using FEI Tecnai F30 Transmission ElectronMicroscope is used for revealing the finest details withbetter resolution

3 RESULTS AND DISCUSSION

31 Electroless Coating

It is imperative that the synthesis of good composite pow-ders needs to be efficient and cost effective as well asachieving the desired properties The control of the micro-structure of ceramic-metal composites is generally difficult

840845850855860865870875880885890

Binding Energy (eV)

Inte

nsity

(A

U)

2p32

2p12

Fig 2 XPS spectrum on Ni coating confirming metallic Ni peaks

to achieve by traditional techniques involving mechanicalmixing of ceramic and metallic powders followed by hotpressing2223 A small-scale homogeneity can be obtainedusing the solndashgel route24ndash26 However the relatively highcost of some reactants and the difficulty to control the geldrying step are some of the drawbacks to this methodElectroless coating has been chosen as it is being perceivedas an efficient method by which Nickel can be uniformlyreinforced in to the alumina matrix to increase the mechan-ical properties of the composite The chemistry of theformation of the Nickel layer uniformly over the surfaceof alumina is shown in Figure 1 SnCl2 hydrolyzes intoSn(OH)Cl and be absorbed on the Al2O3 particles followedby Pd2+ reaction with Sn(OH)Cl to form Pd nuclei on theAl2O3 surfaces This renders the alumina surface suitablefor Nickel coating The electroless bath containing Sodiumhypophosphite and ammonium hydroxide react together toleave electrons which sufficient to reduce Nickel chlorideto Nickel The Nickel thus formed gets coated on to thesensitized and activated alumina particles The thicknessof the electroless coated Ni layer over alumina particlescould be revealed using Focused Ion Beam XPS analy-sis on the layer presented in Figure 2 revealed peaks atbinding energy values of 8528 eV and 8707 eV Thesepeaks were confirmed to be of 2p32 and 2p12 of metal-lic Nickel respectively27 The binding energy shift due tothe charging was removed by referencing the adventitiousC(1s) binding energy value28 at 2846 eV

32 Concept of Wettability

The rationale behind the formation of spherical dropletsof Nickel can be understood from the point of view ofwettability As the Nickel particles attain the liquid stateinstead of complete wetting alumina it makes the pointcontact due to surface tension constraints as shown inFigures 3 4 Contact angle is decided by the energies ofthe solidndashliquid liquidndashvapor and solidndashvapor interfacesinvolved in wetting process The contact angle is highin case of pure metals because of the high solidndashliquid

J Nanosci Nanotechnol 6 651ndash660 2006 653

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Ni Layer

Laser Irradiation

Ni Spheroids

Fig 3 Schematic illustration involving transformation of Ni layer in tospheroids

interfacial energy than the solid vapor interfacial energy Itis being reported29 that the solidndashliquid interfacial energyfor NindashAlumina couple is 186 Jm2 while the solidndashvaporinterfacial energy is calculated as 09 Jm2 (Ref [30]) Ses-sile drop test conducted to measure the wettability anglerevealed the contact angle to be at 108 If the wetta-bility is less the area immediately near the droplet willhave no or little amount of the droplet composition whilefarther away from the droplet composition will be appre-ciable This is because of a combination of the so calledldquoevaporation-depositionrdquo processes occurring during theprocess31 The bonding of Nickel on the Alumina surfacecan be comprehended based on interfacial strength At theinterface a polycrystalline surface is prone to have moregrain boundaries as a result of which roughness is highDue to this high roughness factor the accommodation fac-tor for a Nickel atom to get bonded with the aluminasurface is more compared to that of a smooth interfaceRecent results show the presence of interfacial cracks gen-erated as a result of poor bonding of Nickel with a singlecrystal alumina surface compared to a polycrystalline alu-mina surface as given in Table II

321 Validation of Self-Affine Roughness Model

Saiz et al29 determined the contact angle for NindashAluminasystem to be 130 with the surface energy values oflv sl sv plugging these values in to the Youngrsquos equa-tion But the experimental values measured were found tobe 108 in this study Considering the self-affine roughnessmodel32 the apparent contact angle will be differentfrom the contact angle calculated using the Youngrsquos

solid substrate

liquid drop

σlg

σsv σ ls Θ

gaseous atmosphere

Fig 4 Nickel prefers to make a point contact in Alumina rather thancompletely wetting

Table II Contact angle measurements of metallic Nickel on Alumina

Roughness (nm)Dimensions Contact

Substrate (mm) angle (deg) Ra Rq Rt

Al2O3PC (9999) 108 100 140 1190Al2O3SC (999993) 94 2 25 8

equation29

cos = svminussl

lvThis can be explained based on the roughness of the sur-face of the sample Surface roughness is described as asine wave with an rms amplitude 13 and in plane sur-face roughness correlation length which is the averagedistance between the consecutive hills or valleys on thesurface Rough surfaces are usually a combination of cir-cular grooves and radial grooves The apparent angle iscalculated based on the root mean square value of apparentcontact angles g and w which are contributions from theradial and circular grooves respectively If the amount ofcircular grooves is much higher than the radial grooves atthe point of contact angle measurement apparent contactangle w will be dominating31

w cosminus1

[1+

(12

)minus

(38

)2

]cos

where is the root mean square local slope which canbe approximated to 13H H is the roughness exponentHigher the value of H the higher is the roughness of thesurface For theoretical values of which was 130 solv-ing the above quadratic equation gives the value of 110

for the apparent contact angle which agrees closely withthe experimental values Values assumed for 13 (ampli-tude) (distance between two consecutive hills or val-leys) H (hardness exponent) are taken as 23 nm 60 nmand 06 respectively because alumina being a polycrys-talline and the amount of grain boundaries present aremuch higher compared to a single crystalline material

As it is evident from the table the roughness factorplays a dominant role in Wettability A cross section of thesessile drop sample with nickel droplet on top of aluminasubstrate is shown in Figure 5 EDS spectra from the areajust near the Ni droplet showed that there is no appreciableNi concentration This experimentally proves that wetta-bility of nickel is low31 Such uniformly dispersed Nickelparticles are expected to inhibit crack propagation in thebrittle alumina matrix

33 Laser Processing

Thermal energy given to the sample during laser process-ing depends on the power of the laser used The inter-action time in the first regime can be estimated as thetime for which the laser beam illuminates one point dur-ing fast scanning (sim12 ms) The absorption coefficient formetallic surfaces and NdYAG laser wavelength increases

654 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

t1=0 t2

t3 t4

Ni

Alumina

Fig 5 Illustration of wettability of Nickel on Alumina with time t1 ltt2 lt t3 lt t4

with surface roughness temperature and with an angle ofincidence of the beam33 Typical values lie in an interval25ndash35 for room temperature polished surfaces and per-pendicular beam Absorption takes place in a surface layerapproximately 10minus4 mm in thickness

At different power of the laser beam the morphologyof the Nickel showed variation which is shown in SEMmicrograph in Figures 6 through 8 Due to the irradiationof laser beam electroless Ni layer covering the surface ofalumina melts instantaneously and due to surface tensiongets converted to globules The higher the amount of ther-mal energy the more is the driving force for the globulesto get coarsened The fact that each alumina particle has itsown characteristic size of Ni globules suggests that thereis not enough thermal contact between individual particlesto equilibrate different surface temperatures The sizes ofthe Ni globules observed by SEM ranges between 01 and05 m Some of the alumina particle surfaces irradiated

Fig 6 Ni particle size at a laser power of 250 W

Fig 7 Ni particle coarsens due to increase in the power of the laserbeam to 350 W

with high laser power is free of Ni globules but large Niglobules (sim5 m) are observed in their neighborhoods

Laser processing is actually advantageous from the factthat power of laser beam can be manipulated to get tothe suitable microstructure The presence of Ni particlesas small as 5 nm was revealed in HRTEM analysis of hightemperature processed Ni-coated alumina Lattice fringesof Ni nanoparticles could be resolved in the HRTEM studyas shown in Figure 9 The corresponding EDS spectrum(Fig 9b) analyzed from the selected area confirms thepresence of Ni Al and Oxygen Other elements that aredetected are either from the specimen grid or the Si waferused for specimen preparation However consolidation ofNialumina to bulk composite fabrication was not achievedsince the whole laser beam energy is absorbed in a thin Nilayer and a fast transformation of Ni layer in to spheroidsoccurred Therefore plasma processing route was chosento manufacture a free form bulk nanocomposite part

Fig 8 Ni particle further coarsens at a higher laser power of 400 W

J Nanosci Nanotechnol 6 651ndash660 2006 655

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

(a)

(b)

Fig 9 (a) HRTEM picture revealing the nanostructure retention of NindashAlumina nanocomposite (b) EDS spectrum confirming the presence ofNi in Alumina matrix

34 Plasma Processing

The rationale and science behind using plasma spray as afree form manufacturing technique and aluminum mandrelas a substrate for development of NindashAlumina bulk parthas been discussed in this section

341 PlasmandashParticle Interaction

Plasmandashparticle interaction results in the formation ofatoms in excited state ground state ions and photons inthe plasma plume As a result of these atomic dissocia-tion and recombination processes particle that is sprayedis taken to a partial molten state for a brief period of timetill it splats on to the substrate Time of flight the timerequired for the particle to reside in the flame depends onmaterial parameters such as the latent heat of fusion melt-ing temperature particle size The relation to determineresidence time can be given as34

timinusf =CplnTi minusTsminus lnTm minusTs

h middotAs

Table III Particle radius as a function of dwell time

Particle radius Dwell time (Sec)

50 micron 401times10minus4

30 micron 112times10minus4

20 micron 744times10minus5

1 micron 372times10minus6

50 nm 186times10minus7

where Cp Ti Tm h As are the density specificheat capacity volume initial temperature temperature ofthe plasma plume melting point heat transfer coefficientsurface area of the particle and Ts is the plasma plumetemperature The knowledge of residence time is essentialbefore spraying because too much of over heat may leadto evaporation losses Also higher the amount of super-heating the particles are subjected to rapid cooling as itimpinges on the substrate as a result of which shrinkagestresses will be more This might lead to cracking of thebulk part A calculation (Table III) for residence time tf fordifferent particle size NindashAlumina system has been tabledIt can be inferred that smaller particles need less residencetime which can controlled with the parameters mentionedabove To avoid evaporation and to ensure only surfacemelting appropriate gas flow combinations can be selectedto obtain the residence time of interest The residence timecalculated was 02 microseconds for a 50 nm particle topass through

342 Coating vs Free FormsmdashEngineering Aspects

The section describes the science behind the fabricationof a free form part which is different from coating Alu-minum and Copper being good conductors of heat areideal candidate materials for being used as a substrate formaterial deposition Besides they can be fabricated to anydesired shape as a result of their malleability and the highcoefficient of thermal expansion facilitates easy removalof the sprayed deposit Aluminum substrate can facilitateeasy removal by shrinking more compared to steel or cop-per Cooling and solidification of most materials is accom-panied by contraction or shrinkage As particles strike theyrapidly cool and solidify This generates a tensile stresswithin the particle and a compressive stress within the sur-face of the substrate As the coating builds up the ten-sile stresses developed in the deposit also increases Theresidual stress developed during the air plasma spray pro-cess significantly influences the integrity of the depositStresses developed during spraying can be classified into two parts viz quenching stresses and cooling stressesQuenching stresses are stresses that develop as the splatformation takes place due to particle impingement on thesubstrate The quantification of quenching stresses can bedone using3536

q =Eststs+54 middot13td

6 middot13R middot13tdand = Ed

Es

656 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

0

5

10

15

20

25

30

35

40

Alumina

Zircon

ia

Titania

MoS

i 2

Si 3N 4

Aluminu

m

Nickel

Titaniu

m

Coppe

r

Material

Str

ess

(MP

a)

Fig 10 Quenching stresses developed during the spray of differentparticles

where E is the elastic modulii of the substrate and thedeposit indicated by subscripts s and d respectively 13tis the thickness of the deposit and 13R is the change inthe radius of curvature of the coating The plots as shownin Figure 11 depict that different materials will developdifferent quenching stresses as the spray is going onBased on the thickness of the deposit achieved the quench-ing stresses will vary Among various spray materials forwhich quenching stresses have been calculated (Fig 10)The histogram shown for alumina and nickel particles withdeposit thickness of 0015 mm (15 microns) is of interestto this paper Cooling stresses are stresses that the depositproduces when the spraying is completed and the tempera-ture of the part is plummeting to room temperature It canbe accounted for using an equation37

c =EddminussTdminusTRT

1+2Ed middot tdEsts$

where EdmdashYoungrsquos modulus of the deposit TdmdashTemperature of the deposit TRTmdashRoom temperature ofthe sample after cooling dmdashCoefficient of thermalexpansion of the deposition smdashCoefficient of thermalexpansion of the substrate EsmdashYoungrsquos modulus of thesubstrate tsmdashthickness of the substrate and tdmdashthicknessof the deposition Figure 11 depicts the calculated cooling

ndash120

ndash100

ndash80

ndash60

ndash40

ndash20

0Alumina Zirconia Titania MoSi2 Si3N4

Coo

ling

Str

ess

(MP

a)

Copper

Aluminum

Fig 11 Cooling stresses developed in the deposit for Cu and Alsubstrates

stresses for different materials for two different substratesCopper and Aluminum The total amount of stresses devel-oped in the deposit is a combination of both quenchingand cooling stresses Quenching stresses which are ten-sile in nature keeps increasing with thickness build-upThe thermal stresses (cooling stresses) will be compres-sive when d lt s Selection of aluminum mandrel forNindashAlumina is justified based on the calculation that thecooling stresses are compressive in nature It will facilitatethe removal of spray deposit from the substrate This willeventually lead to free form fabrication of deposits

343 Material Characterization

The trial for spraying nano ceramic particles throughplasma flame has been successful The NindashAlumina cylin-drical bulk nanocomposite without any surface defectswas manufactured using plasma spray technique (Fig 12)Microstructure evaluation using SEM and other character-ization tools such as XRD and HRTEM is being discussedat lengthCross-Section Microscopy In an effort to study the

porosity distribution across the thickness scanning elec-tron microstructure of the internal center and outside ofthe thickness (Fig 13) has been analyzed The section is96ndash98 dense and does not show any significant cracksor porosity It also reveals the formation of splats due tohigher velocity with which the melted particles impingeson to the substrate There was no evidence of intercon-nected pores which could have been detrimental to theintegrity of the part Also the porosity distribution is morein case of internal thickness and less in outside thicknessThis can be explained from the fact that particles temper-ature gets reduced drastically during the initial stages ofspray as a result of which shrinkage is more while theparticles at the outside thickness has to undergo relativelyless temperature difference SEM micrographs of the crosssection revealed a clear demarcation between the forma-tion of folds and the formation of porosity The porositiesare revealed as dark regions compared to less intense foldsFold formation happens due to the inability of particles tofully splat on top of the existing splat

Substrate incompression

Free form partin Tension

(a) (b)

Fig 12 (a) The mandrel is subjected to compressive stresses whilethe sprayed part acquires tensile stress as the spray continues (b) A12primeprimeODtimes1primeprimeIDtimes15primeprime length cylinder was produced with full integrity

J Nanosci Nanotechnol 6 651ndash660 2006 657

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Folds

Isolated pores

Folds

(a)

(b)

(c)

Fig 13 Cross sectional low magnification micrographs indicating theextent of porosity formation at the (a) external (b) middle and (c) inter-nal portions of the bulk nanocomposite

344 Structural Characterization

Phase Transformation It is being observed that thealumina matrix undergone phase transformation fromalpha to gamma structure The alpha to gamma transi-tion depends on the ratio of the gun power to the primarygas flow in Standard Cubic Feet per Hour (SCFH) It ismentioned elsewhere [shaw] that if the parent phase is -alumina based on this ratio mentioned above the product

2θ20 30 40 50 60 70 80

(c)

(b)

(a)

NickelAlpha AluminaGamma AluminaNiO

Fig 14 Phase transformations associated with plasma spray of Ni-Al2O3 (a) Feedstock powder (b) plasma sprayed component (c) hydro-gen reduction

phase might be a combination of more -alumina and less alumina if the ratio is less than or equal to 240 or a com-bination of more alumina and less -alumina if the ratiois more than or equal to 310 The density and hardness val-ues achieved38 with the above ratios show significant dif-ferences with the former showing less hardness and densitythan the latter The phase transformation achieved in ourpresent case had more alumina The spray parameterswere manipulated to keep the ratio above 310 to makesure that dense and hard nanostructures were obtained Thediffractograms before and after plasma spraying is shownin Figure 14 X-ray diffraction spectrums from the plasmaspray deposition and powder mixture prior to the spray-ing are compared Both the diffractograms reveal the pres-ence of FCC Ni and Alumina The difference betweenthe diffractograms is that of the peak broadening Dur-ing the spray particles literally gets quenched from theirmelting temperatures towards the substrate preheat temper-ature Such large temperatures gradients are expected toimpart residual stresses on the sprayed part The stressescan be calculated based on the difference in ldquodrdquo spacingof the particles prior to spraying and after spraying Fromthe modulus of elasticity values the amount of residualstress can be calculated as follows39

Eminus1hkl =

C11 +C12

C11 minusC12C11 +2C12

+(

1C44

minus 2C11 minusC12

)l2m2 +m2n2 +n2l2

The compliance constants are C11 = 2465 GPa C12 =1473 GPa and C44 = 1247 GPa The calculated residualstress values are presented in Table IV

Since the process has been carried out in atmosphericair Ni has the propensity to get converted to its surfaceoxide Gibbs free energy of formation of Nickel oxide upto melting point of Nickel is negative Although Gibbs free

658 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table IV Residual stress measurements using X-ray diffraction techniques

hkl 2 (Measured) 2 (JCPDS) d (Measured) Aring d (JCPDS) Aring Strain () E (GPa) Stress (Mpa)

NiO 200 4326 4329 209 209 0066 136 90NiO 111 3725 3728 241 241 0078 304 236NiO 220 6289 6292 148 148 0043 232 99-Alumina 440 67 67032 140 139 0042 408 172

energy curve for formation of Nickel oxide is sloped upwith temperature the partial pressure of oxygen at NTP ismuch higher for the reaction to proceed forward and formNiO APS process carried out at dehumidified air will havean oxygen partial pressure of 02 atm which is more thansufficient for the NiO formation to occur Moreover if thestand off distance is more Nickel in its molten state picksup more oxygen from the atmosphere than it would haveif it were in solid state By reducing the stand off distancethe amount of formation of Nickel oxide can be reducedbut can not be totally eliminated XRD peaks confirmedthe formation of NiO An effort has been made to bringNiO back to its metallic state The plasma spray formednanocomposite was reduced in an atmosphere of Hydrogenat 1000 C for 1 hour

345 High Resolution TransmissionElectron Microscopy

Alumina grain boundaries being high energy regions thereis more probability of the presence of Nickel along thegrain boundaries In order to determine the location ofNickel EDS has been performed along the grain bound-aries to select spots for imaging of Nickel nano parti-cles In those spots where strong Nickel signal have beendetected STEM imaging has been done to determine themorphology and the size of the Nickel particles Spheri-cal Ni particles with 30 nm size have been imaged andare shown in Figure 15 Another evidence of nanostruc-ture retention of Nickel particle is the SAD pattern whichshows a spotty ring pattern as revealed in Figure 15(a)(inset) This pattern is a result of extremely fine grainsThe possible explanations of the nano structure retentionin NindashAlumina nanocomposite could be the lack of timepermitted for the coarsening of Nickel particles travers-ing through the high temperature zone (plasma flame) On

0100200300400500600700800900

1000

0 2000 4000 6000 8000 10000

Energy (eV)

Inte

nsity

(A

U)

NiCu

Al

O

20 nm

Fig 15 STEM pictures depicting nanostructure retention of nickel par-ticles and the corresponding EDS spectrum confirming the presence ofnickel

the other hand during conventional sintering a significanttime is spent in the temperature range where the coarsen-ing effect is appreciable20 Besides in the nanostructuressurface melting is preferred in the fast translation thuskeeping the core intact From the TEM pictures of plasmasprayed NickelndashAlumina it is envisioned that plasma pro-cessing is considered to be a potential tool to fabricate freeform (eg cylindrical) bulk composites with nanostructureretention and improved mechanical properties

35 Mechanical Properties

Nickel plays a dominant role in increasing the strengthof the ceramic matrix The presence of cracks in ceram-ics makes them more brittle compared to metals At thetip of the crack as shown in Figure 16 there is an enor-mous amount of stress concentration developed as a resultof which the cracks tend to propagate easily The surfaceenergy required to propagate the crack is more than com-pensated by the release of the elastic strain energy Thepresence of Nickel spheroids dispersed uniformly through-out the matrix blunts the crack thereby making it more dif-ficult to propagate as shown in Figure 16 Nickel particleswhich are highly plastic in nature aids in crack bluntingand improvement of fracture toughness of the nanocom-posite Lieberthal et al16 proposed a model in which acrack propagating along the alumina grain boundaries aredeflected along around the Ni particle located in grainboundaries which aid in improving the fracture toughnessof the ceramic nanocomposite

The hardness and fracture toughness of the NindashAluminaplasma sprayed sample that could be achieved with thegiven set of plasma spray parameters were measured tobe 1025 HV and sim5 MPa middot m12 The hardness valuesachieved were found to be better than that achieved for the

Crack tip in a ceramicmatrix

Ni-particle

Fig 16 Nickel particles blunt the crack and arrest crack propagationimproves the fracture toughness

J Nanosci Nanotechnol 6 651ndash660 2006 659

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

monolithic alumina developed using plasma spray (940plusmn20 HV)40 The marked increase in the fracture toughnesscompared to other processing techniques can be attributedto the crack blunting capability of ductile Nickel It shouldbe noted that NindashAlumina nanocomposites made throughtechniques such as hot press sintering12 reported a frac-ture toughness of sim35 MPa middotm12 In other words the increase in fracture toughness was approximately 40

4 CONCLUSIONS

Electroless Nickel coated alumina particles have been suc-cessfully used to manufacture a free-form bulk nanocom-posite with improved hardness and fracture toughnessWettability study of Nickel on polycrystalline Aluminarevealed the contact angle which validated the self affineroughness model being used to calculate the apparent con-tact angle Laser processing had shortcomings from thepoint of view of time and the high laser required to conso-lidate the powders to bulk form especially in the casemetalceramic matrix For plasma processing substratematerial selection is important based on the material com-position to be sprayed Calculation of residual High frac-ture toughness can be attributed to the nanostructuredNickel which arrests the crack propagation Thus plasmaspray processing is an effective technique to make freeform bulk nanostructures

Acknowledgments The authors wish to thank theOffice of Naval Research Young Investigator Award (ONRYIP-N000140210591) for their financial support Theauthors would also like to thank Plasma Processes Incfor their help in product development We also appreci-ate the equipment support from Materials CharacterizationFacility at the University of Central Florida

References and Notes

1 W H Tuan and R J Brook J Eur Ceram Soc 6 31 (1990)2 A Paesano Jr C K Matsuda J B M Da Cunha M A Z

Vasconcellos B Hallouche and S L Silva J Magn Magn Mater264 264 (2003)

3 T Sekino and K Niihara J Mater Sci 32 3943 (1997)4 T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)5 A Nakahira and K Niihara Fract Mech 9 165 (1992)6 Y K Jeong and K Niihara Nanostruct Mater 9 193 (1997)7 S C Wang and W C J Wei Nanostruct Mater 10 983 (1998)

8 B Budiansky J C Amazigo and A J Evans J Mech Phys Solids36 167 (1988)

9 S Ozkar G A Ozin and R A Prokopowicz Chem Mater 4 1380(1992)

10 S T Oh T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)11 O Aharon S Bar-Ziv D Gorni T Cohen-Hyams and W D

Kaplan Scripta Mater 50 1209 (2004)12 T Sekino T Nakajima and K Niihara Mater Lett 29 165 (1996)13 S I Cha S H Hong and B K Kim Mater Sci Eng A 351 31

(2003)14 S Seal S C Kuiry P Georgieva and K Rea Scripta Mater 50

1237 (2004)15 R Z Chen and W H Tuan J Eur Ceram Soc 19 463 (1999)16 M Lieberthal and W D Kaplan Mater Sci Eng A 302 83 (2001)17 K Niihara B S Kim T Nakayama T Kusunose T Nomoto

A Hikasa and T Sekino J Eur Ceram Soc 24 3419 (2004)18 U L Adisorn T Matsunaga Y Kobayashi S M Choi and

H Awaji Ceram Intl 31 803 (2005)19 S Seal S C Kuiry P Georgieva and A Agarwal MRS Bull 29

16 (2004)20 J Michalski K Konopka and M Trzaska Mater Chem Phys 81

407 (2003)21 A G Evans and E A Charles J Am Ceram Soc 59 371 (1976)22 W H Tuan and R J Brook J Eur Ceram Soc 10 95 (1992)23 W H Tuan and W B Chou J Am Ceram Soc 80 2418 (1997)24 E Breval G Dodds and C G Pantano Mater Res Bull 20 1191

(1985)25 E Breval Z Deng S Chiou and C G Pantano J Mater Sci 27

1464 (1992)26 E D Rodeghiero O K Tse J Chisaki and E P Giannelis Mater

Sci Eng A 195 151 (1995)27 J F Moulder W F Stickle P E Sobol and K D Bomben

in Handbook of X-Ray Photoelectron Spectroscopy edited byJ Chastain and R C King Jr Physical Electronics Inc Minnesota(1995) p 71

28 T L Barr and S Seal J Vac Sci Technol A 13 1239 (1995)29 E Saiz R M Cannon and A P Tomsia Acta Mater 47 4209

(1999)30 W D Kingery J Am Ceram Soc 37 42 (1954)31 G Palasantzas and J Th M De Hosson Acta Mater 49 3533

(2001)32 N Eustathopolous Acta Mater 46 2319 (1998)33 J A Vreeling V Ocelrsquok Y T Pei D T L Van Agterveld and

J Th M De Hosson Acta Mater 48 4225 (2000)34 R Ye T Ishigaki J Jurewicz P Proulx and M I Boulos Plasma

Chem Plasma Process 24 555 (2004)35 A Brenner and S Senderoff J Res Natl Bur Stand 42 105 (1949)36 S Kuroda and T W Clyne Thin Solid Films 200 49 (1991)37 M W Hunt Guide to engineering materials Adv Mater Proc 158

6 (2000)38 L L Shaw D Goberman R Ren M Gell S Jiang Y Wang T D

Xiao and P R Strutt Surf Coat Technol 130 1 (2000)39 O D Slagle and H A Mckinsky J Appl Phys 38 437 (1967)40 M Vural S Zeytin and A H Ucisik Surf Coat Technol 97 347

(1997)

Received 10 July 2005 Accepted 12 October 2005

660 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Ni Layer

Laser Irradiation

Ni Spheroids

Fig 3 Schematic illustration involving transformation of Ni layer in tospheroids

interfacial energy than the solid vapor interfacial energy Itis being reported29 that the solidndashliquid interfacial energyfor NindashAlumina couple is 186 Jm2 while the solidndashvaporinterfacial energy is calculated as 09 Jm2 (Ref [30]) Ses-sile drop test conducted to measure the wettability anglerevealed the contact angle to be at 108 If the wetta-bility is less the area immediately near the droplet willhave no or little amount of the droplet composition whilefarther away from the droplet composition will be appre-ciable This is because of a combination of the so calledldquoevaporation-depositionrdquo processes occurring during theprocess31 The bonding of Nickel on the Alumina surfacecan be comprehended based on interfacial strength At theinterface a polycrystalline surface is prone to have moregrain boundaries as a result of which roughness is highDue to this high roughness factor the accommodation fac-tor for a Nickel atom to get bonded with the aluminasurface is more compared to that of a smooth interfaceRecent results show the presence of interfacial cracks gen-erated as a result of poor bonding of Nickel with a singlecrystal alumina surface compared to a polycrystalline alu-mina surface as given in Table II

321 Validation of Self-Affine Roughness Model

Saiz et al29 determined the contact angle for NindashAluminasystem to be 130 with the surface energy values oflv sl sv plugging these values in to the Youngrsquos equa-tion But the experimental values measured were found tobe 108 in this study Considering the self-affine roughnessmodel32 the apparent contact angle will be differentfrom the contact angle calculated using the Youngrsquos

solid substrate

liquid drop

σlg

σsv σ ls Θ

gaseous atmosphere

Fig 4 Nickel prefers to make a point contact in Alumina rather thancompletely wetting

Table II Contact angle measurements of metallic Nickel on Alumina

Roughness (nm)Dimensions Contact

Substrate (mm) angle (deg) Ra Rq Rt

Al2O3PC (9999) 108 100 140 1190Al2O3SC (999993) 94 2 25 8

equation29

cos = svminussl

lvThis can be explained based on the roughness of the sur-face of the sample Surface roughness is described as asine wave with an rms amplitude 13 and in plane sur-face roughness correlation length which is the averagedistance between the consecutive hills or valleys on thesurface Rough surfaces are usually a combination of cir-cular grooves and radial grooves The apparent angle iscalculated based on the root mean square value of apparentcontact angles g and w which are contributions from theradial and circular grooves respectively If the amount ofcircular grooves is much higher than the radial grooves atthe point of contact angle measurement apparent contactangle w will be dominating31

w cosminus1

[1+

(12

)minus

(38

)2

]cos

where is the root mean square local slope which canbe approximated to 13H H is the roughness exponentHigher the value of H the higher is the roughness of thesurface For theoretical values of which was 130 solv-ing the above quadratic equation gives the value of 110

for the apparent contact angle which agrees closely withthe experimental values Values assumed for 13 (ampli-tude) (distance between two consecutive hills or val-leys) H (hardness exponent) are taken as 23 nm 60 nmand 06 respectively because alumina being a polycrys-talline and the amount of grain boundaries present aremuch higher compared to a single crystalline material

As it is evident from the table the roughness factorplays a dominant role in Wettability A cross section of thesessile drop sample with nickel droplet on top of aluminasubstrate is shown in Figure 5 EDS spectra from the areajust near the Ni droplet showed that there is no appreciableNi concentration This experimentally proves that wetta-bility of nickel is low31 Such uniformly dispersed Nickelparticles are expected to inhibit crack propagation in thebrittle alumina matrix

33 Laser Processing

Thermal energy given to the sample during laser process-ing depends on the power of the laser used The inter-action time in the first regime can be estimated as thetime for which the laser beam illuminates one point dur-ing fast scanning (sim12 ms) The absorption coefficient formetallic surfaces and NdYAG laser wavelength increases

654 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

t1=0 t2

t3 t4

Ni

Alumina

Fig 5 Illustration of wettability of Nickel on Alumina with time t1 ltt2 lt t3 lt t4

with surface roughness temperature and with an angle ofincidence of the beam33 Typical values lie in an interval25ndash35 for room temperature polished surfaces and per-pendicular beam Absorption takes place in a surface layerapproximately 10minus4 mm in thickness

At different power of the laser beam the morphologyof the Nickel showed variation which is shown in SEMmicrograph in Figures 6 through 8 Due to the irradiationof laser beam electroless Ni layer covering the surface ofalumina melts instantaneously and due to surface tensiongets converted to globules The higher the amount of ther-mal energy the more is the driving force for the globulesto get coarsened The fact that each alumina particle has itsown characteristic size of Ni globules suggests that thereis not enough thermal contact between individual particlesto equilibrate different surface temperatures The sizes ofthe Ni globules observed by SEM ranges between 01 and05 m Some of the alumina particle surfaces irradiated

Fig 6 Ni particle size at a laser power of 250 W

Fig 7 Ni particle coarsens due to increase in the power of the laserbeam to 350 W

with high laser power is free of Ni globules but large Niglobules (sim5 m) are observed in their neighborhoods

Laser processing is actually advantageous from the factthat power of laser beam can be manipulated to get tothe suitable microstructure The presence of Ni particlesas small as 5 nm was revealed in HRTEM analysis of hightemperature processed Ni-coated alumina Lattice fringesof Ni nanoparticles could be resolved in the HRTEM studyas shown in Figure 9 The corresponding EDS spectrum(Fig 9b) analyzed from the selected area confirms thepresence of Ni Al and Oxygen Other elements that aredetected are either from the specimen grid or the Si waferused for specimen preparation However consolidation ofNialumina to bulk composite fabrication was not achievedsince the whole laser beam energy is absorbed in a thin Nilayer and a fast transformation of Ni layer in to spheroidsoccurred Therefore plasma processing route was chosento manufacture a free form bulk nanocomposite part

Fig 8 Ni particle further coarsens at a higher laser power of 400 W

J Nanosci Nanotechnol 6 651ndash660 2006 655

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

(a)

(b)

Fig 9 (a) HRTEM picture revealing the nanostructure retention of NindashAlumina nanocomposite (b) EDS spectrum confirming the presence ofNi in Alumina matrix

34 Plasma Processing

The rationale and science behind using plasma spray as afree form manufacturing technique and aluminum mandrelas a substrate for development of NindashAlumina bulk parthas been discussed in this section

341 PlasmandashParticle Interaction

Plasmandashparticle interaction results in the formation ofatoms in excited state ground state ions and photons inthe plasma plume As a result of these atomic dissocia-tion and recombination processes particle that is sprayedis taken to a partial molten state for a brief period of timetill it splats on to the substrate Time of flight the timerequired for the particle to reside in the flame depends onmaterial parameters such as the latent heat of fusion melt-ing temperature particle size The relation to determineresidence time can be given as34

timinusf =CplnTi minusTsminus lnTm minusTs

h middotAs

Table III Particle radius as a function of dwell time

Particle radius Dwell time (Sec)

50 micron 401times10minus4

30 micron 112times10minus4

20 micron 744times10minus5

1 micron 372times10minus6

50 nm 186times10minus7

where Cp Ti Tm h As are the density specificheat capacity volume initial temperature temperature ofthe plasma plume melting point heat transfer coefficientsurface area of the particle and Ts is the plasma plumetemperature The knowledge of residence time is essentialbefore spraying because too much of over heat may leadto evaporation losses Also higher the amount of super-heating the particles are subjected to rapid cooling as itimpinges on the substrate as a result of which shrinkagestresses will be more This might lead to cracking of thebulk part A calculation (Table III) for residence time tf fordifferent particle size NindashAlumina system has been tabledIt can be inferred that smaller particles need less residencetime which can controlled with the parameters mentionedabove To avoid evaporation and to ensure only surfacemelting appropriate gas flow combinations can be selectedto obtain the residence time of interest The residence timecalculated was 02 microseconds for a 50 nm particle topass through

342 Coating vs Free FormsmdashEngineering Aspects

The section describes the science behind the fabricationof a free form part which is different from coating Alu-minum and Copper being good conductors of heat areideal candidate materials for being used as a substrate formaterial deposition Besides they can be fabricated to anydesired shape as a result of their malleability and the highcoefficient of thermal expansion facilitates easy removalof the sprayed deposit Aluminum substrate can facilitateeasy removal by shrinking more compared to steel or cop-per Cooling and solidification of most materials is accom-panied by contraction or shrinkage As particles strike theyrapidly cool and solidify This generates a tensile stresswithin the particle and a compressive stress within the sur-face of the substrate As the coating builds up the ten-sile stresses developed in the deposit also increases Theresidual stress developed during the air plasma spray pro-cess significantly influences the integrity of the depositStresses developed during spraying can be classified into two parts viz quenching stresses and cooling stressesQuenching stresses are stresses that develop as the splatformation takes place due to particle impingement on thesubstrate The quantification of quenching stresses can bedone using3536

q =Eststs+54 middot13td

6 middot13R middot13tdand = Ed

Es

656 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

0

5

10

15

20

25

30

35

40

Alumina

Zircon

ia

Titania

MoS

i 2

Si 3N 4

Aluminu

m

Nickel

Titaniu

m

Coppe

r

Material

Str

ess

(MP

a)

Fig 10 Quenching stresses developed during the spray of differentparticles

where E is the elastic modulii of the substrate and thedeposit indicated by subscripts s and d respectively 13tis the thickness of the deposit and 13R is the change inthe radius of curvature of the coating The plots as shownin Figure 11 depict that different materials will developdifferent quenching stresses as the spray is going onBased on the thickness of the deposit achieved the quench-ing stresses will vary Among various spray materials forwhich quenching stresses have been calculated (Fig 10)The histogram shown for alumina and nickel particles withdeposit thickness of 0015 mm (15 microns) is of interestto this paper Cooling stresses are stresses that the depositproduces when the spraying is completed and the tempera-ture of the part is plummeting to room temperature It canbe accounted for using an equation37

c =EddminussTdminusTRT

1+2Ed middot tdEsts$

where EdmdashYoungrsquos modulus of the deposit TdmdashTemperature of the deposit TRTmdashRoom temperature ofthe sample after cooling dmdashCoefficient of thermalexpansion of the deposition smdashCoefficient of thermalexpansion of the substrate EsmdashYoungrsquos modulus of thesubstrate tsmdashthickness of the substrate and tdmdashthicknessof the deposition Figure 11 depicts the calculated cooling

ndash120

ndash100

ndash80

ndash60

ndash40

ndash20

0Alumina Zirconia Titania MoSi2 Si3N4

Coo

ling

Str

ess

(MP

a)

Copper

Aluminum

Fig 11 Cooling stresses developed in the deposit for Cu and Alsubstrates

stresses for different materials for two different substratesCopper and Aluminum The total amount of stresses devel-oped in the deposit is a combination of both quenchingand cooling stresses Quenching stresses which are ten-sile in nature keeps increasing with thickness build-upThe thermal stresses (cooling stresses) will be compres-sive when d lt s Selection of aluminum mandrel forNindashAlumina is justified based on the calculation that thecooling stresses are compressive in nature It will facilitatethe removal of spray deposit from the substrate This willeventually lead to free form fabrication of deposits

343 Material Characterization

The trial for spraying nano ceramic particles throughplasma flame has been successful The NindashAlumina cylin-drical bulk nanocomposite without any surface defectswas manufactured using plasma spray technique (Fig 12)Microstructure evaluation using SEM and other character-ization tools such as XRD and HRTEM is being discussedat lengthCross-Section Microscopy In an effort to study the

porosity distribution across the thickness scanning elec-tron microstructure of the internal center and outside ofthe thickness (Fig 13) has been analyzed The section is96ndash98 dense and does not show any significant cracksor porosity It also reveals the formation of splats due tohigher velocity with which the melted particles impingeson to the substrate There was no evidence of intercon-nected pores which could have been detrimental to theintegrity of the part Also the porosity distribution is morein case of internal thickness and less in outside thicknessThis can be explained from the fact that particles temper-ature gets reduced drastically during the initial stages ofspray as a result of which shrinkage is more while theparticles at the outside thickness has to undergo relativelyless temperature difference SEM micrographs of the crosssection revealed a clear demarcation between the forma-tion of folds and the formation of porosity The porositiesare revealed as dark regions compared to less intense foldsFold formation happens due to the inability of particles tofully splat on top of the existing splat

Substrate incompression

Free form partin Tension

(a) (b)

Fig 12 (a) The mandrel is subjected to compressive stresses whilethe sprayed part acquires tensile stress as the spray continues (b) A12primeprimeODtimes1primeprimeIDtimes15primeprime length cylinder was produced with full integrity

J Nanosci Nanotechnol 6 651ndash660 2006 657

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Folds

Isolated pores

Folds

(a)

(b)

(c)

Fig 13 Cross sectional low magnification micrographs indicating theextent of porosity formation at the (a) external (b) middle and (c) inter-nal portions of the bulk nanocomposite

344 Structural Characterization

Phase Transformation It is being observed that thealumina matrix undergone phase transformation fromalpha to gamma structure The alpha to gamma transi-tion depends on the ratio of the gun power to the primarygas flow in Standard Cubic Feet per Hour (SCFH) It ismentioned elsewhere [shaw] that if the parent phase is -alumina based on this ratio mentioned above the product

2θ20 30 40 50 60 70 80

(c)

(b)

(a)

NickelAlpha AluminaGamma AluminaNiO

Fig 14 Phase transformations associated with plasma spray of Ni-Al2O3 (a) Feedstock powder (b) plasma sprayed component (c) hydro-gen reduction

phase might be a combination of more -alumina and less alumina if the ratio is less than or equal to 240 or a com-bination of more alumina and less -alumina if the ratiois more than or equal to 310 The density and hardness val-ues achieved38 with the above ratios show significant dif-ferences with the former showing less hardness and densitythan the latter The phase transformation achieved in ourpresent case had more alumina The spray parameterswere manipulated to keep the ratio above 310 to makesure that dense and hard nanostructures were obtained Thediffractograms before and after plasma spraying is shownin Figure 14 X-ray diffraction spectrums from the plasmaspray deposition and powder mixture prior to the spray-ing are compared Both the diffractograms reveal the pres-ence of FCC Ni and Alumina The difference betweenthe diffractograms is that of the peak broadening Dur-ing the spray particles literally gets quenched from theirmelting temperatures towards the substrate preheat temper-ature Such large temperatures gradients are expected toimpart residual stresses on the sprayed part The stressescan be calculated based on the difference in ldquodrdquo spacingof the particles prior to spraying and after spraying Fromthe modulus of elasticity values the amount of residualstress can be calculated as follows39

Eminus1hkl =

C11 +C12

C11 minusC12C11 +2C12

+(

1C44

minus 2C11 minusC12

)l2m2 +m2n2 +n2l2

The compliance constants are C11 = 2465 GPa C12 =1473 GPa and C44 = 1247 GPa The calculated residualstress values are presented in Table IV

Since the process has been carried out in atmosphericair Ni has the propensity to get converted to its surfaceoxide Gibbs free energy of formation of Nickel oxide upto melting point of Nickel is negative Although Gibbs free

658 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table IV Residual stress measurements using X-ray diffraction techniques

hkl 2 (Measured) 2 (JCPDS) d (Measured) Aring d (JCPDS) Aring Strain () E (GPa) Stress (Mpa)

NiO 200 4326 4329 209 209 0066 136 90NiO 111 3725 3728 241 241 0078 304 236NiO 220 6289 6292 148 148 0043 232 99-Alumina 440 67 67032 140 139 0042 408 172

energy curve for formation of Nickel oxide is sloped upwith temperature the partial pressure of oxygen at NTP ismuch higher for the reaction to proceed forward and formNiO APS process carried out at dehumidified air will havean oxygen partial pressure of 02 atm which is more thansufficient for the NiO formation to occur Moreover if thestand off distance is more Nickel in its molten state picksup more oxygen from the atmosphere than it would haveif it were in solid state By reducing the stand off distancethe amount of formation of Nickel oxide can be reducedbut can not be totally eliminated XRD peaks confirmedthe formation of NiO An effort has been made to bringNiO back to its metallic state The plasma spray formednanocomposite was reduced in an atmosphere of Hydrogenat 1000 C for 1 hour

345 High Resolution TransmissionElectron Microscopy

Alumina grain boundaries being high energy regions thereis more probability of the presence of Nickel along thegrain boundaries In order to determine the location ofNickel EDS has been performed along the grain bound-aries to select spots for imaging of Nickel nano parti-cles In those spots where strong Nickel signal have beendetected STEM imaging has been done to determine themorphology and the size of the Nickel particles Spheri-cal Ni particles with 30 nm size have been imaged andare shown in Figure 15 Another evidence of nanostruc-ture retention of Nickel particle is the SAD pattern whichshows a spotty ring pattern as revealed in Figure 15(a)(inset) This pattern is a result of extremely fine grainsThe possible explanations of the nano structure retentionin NindashAlumina nanocomposite could be the lack of timepermitted for the coarsening of Nickel particles travers-ing through the high temperature zone (plasma flame) On

0100200300400500600700800900

1000

0 2000 4000 6000 8000 10000

Energy (eV)

Inte

nsity

(A

U)

NiCu

Al

O

20 nm

Fig 15 STEM pictures depicting nanostructure retention of nickel par-ticles and the corresponding EDS spectrum confirming the presence ofnickel

the other hand during conventional sintering a significanttime is spent in the temperature range where the coarsen-ing effect is appreciable20 Besides in the nanostructuressurface melting is preferred in the fast translation thuskeeping the core intact From the TEM pictures of plasmasprayed NickelndashAlumina it is envisioned that plasma pro-cessing is considered to be a potential tool to fabricate freeform (eg cylindrical) bulk composites with nanostructureretention and improved mechanical properties

35 Mechanical Properties

Nickel plays a dominant role in increasing the strengthof the ceramic matrix The presence of cracks in ceram-ics makes them more brittle compared to metals At thetip of the crack as shown in Figure 16 there is an enor-mous amount of stress concentration developed as a resultof which the cracks tend to propagate easily The surfaceenergy required to propagate the crack is more than com-pensated by the release of the elastic strain energy Thepresence of Nickel spheroids dispersed uniformly through-out the matrix blunts the crack thereby making it more dif-ficult to propagate as shown in Figure 16 Nickel particleswhich are highly plastic in nature aids in crack bluntingand improvement of fracture toughness of the nanocom-posite Lieberthal et al16 proposed a model in which acrack propagating along the alumina grain boundaries aredeflected along around the Ni particle located in grainboundaries which aid in improving the fracture toughnessof the ceramic nanocomposite

The hardness and fracture toughness of the NindashAluminaplasma sprayed sample that could be achieved with thegiven set of plasma spray parameters were measured tobe 1025 HV and sim5 MPa middot m12 The hardness valuesachieved were found to be better than that achieved for the

Crack tip in a ceramicmatrix

Ni-particle

Fig 16 Nickel particles blunt the crack and arrest crack propagationimproves the fracture toughness

J Nanosci Nanotechnol 6 651ndash660 2006 659

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

monolithic alumina developed using plasma spray (940plusmn20 HV)40 The marked increase in the fracture toughnesscompared to other processing techniques can be attributedto the crack blunting capability of ductile Nickel It shouldbe noted that NindashAlumina nanocomposites made throughtechniques such as hot press sintering12 reported a frac-ture toughness of sim35 MPa middotm12 In other words the increase in fracture toughness was approximately 40

4 CONCLUSIONS

Electroless Nickel coated alumina particles have been suc-cessfully used to manufacture a free-form bulk nanocom-posite with improved hardness and fracture toughnessWettability study of Nickel on polycrystalline Aluminarevealed the contact angle which validated the self affineroughness model being used to calculate the apparent con-tact angle Laser processing had shortcomings from thepoint of view of time and the high laser required to conso-lidate the powders to bulk form especially in the casemetalceramic matrix For plasma processing substratematerial selection is important based on the material com-position to be sprayed Calculation of residual High frac-ture toughness can be attributed to the nanostructuredNickel which arrests the crack propagation Thus plasmaspray processing is an effective technique to make freeform bulk nanostructures

Acknowledgments The authors wish to thank theOffice of Naval Research Young Investigator Award (ONRYIP-N000140210591) for their financial support Theauthors would also like to thank Plasma Processes Incfor their help in product development We also appreci-ate the equipment support from Materials CharacterizationFacility at the University of Central Florida

References and Notes

1 W H Tuan and R J Brook J Eur Ceram Soc 6 31 (1990)2 A Paesano Jr C K Matsuda J B M Da Cunha M A Z

Vasconcellos B Hallouche and S L Silva J Magn Magn Mater264 264 (2003)

3 T Sekino and K Niihara J Mater Sci 32 3943 (1997)4 T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)5 A Nakahira and K Niihara Fract Mech 9 165 (1992)6 Y K Jeong and K Niihara Nanostruct Mater 9 193 (1997)7 S C Wang and W C J Wei Nanostruct Mater 10 983 (1998)

8 B Budiansky J C Amazigo and A J Evans J Mech Phys Solids36 167 (1988)

9 S Ozkar G A Ozin and R A Prokopowicz Chem Mater 4 1380(1992)

10 S T Oh T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)11 O Aharon S Bar-Ziv D Gorni T Cohen-Hyams and W D

Kaplan Scripta Mater 50 1209 (2004)12 T Sekino T Nakajima and K Niihara Mater Lett 29 165 (1996)13 S I Cha S H Hong and B K Kim Mater Sci Eng A 351 31

(2003)14 S Seal S C Kuiry P Georgieva and K Rea Scripta Mater 50

1237 (2004)15 R Z Chen and W H Tuan J Eur Ceram Soc 19 463 (1999)16 M Lieberthal and W D Kaplan Mater Sci Eng A 302 83 (2001)17 K Niihara B S Kim T Nakayama T Kusunose T Nomoto

A Hikasa and T Sekino J Eur Ceram Soc 24 3419 (2004)18 U L Adisorn T Matsunaga Y Kobayashi S M Choi and

H Awaji Ceram Intl 31 803 (2005)19 S Seal S C Kuiry P Georgieva and A Agarwal MRS Bull 29

16 (2004)20 J Michalski K Konopka and M Trzaska Mater Chem Phys 81

407 (2003)21 A G Evans and E A Charles J Am Ceram Soc 59 371 (1976)22 W H Tuan and R J Brook J Eur Ceram Soc 10 95 (1992)23 W H Tuan and W B Chou J Am Ceram Soc 80 2418 (1997)24 E Breval G Dodds and C G Pantano Mater Res Bull 20 1191

(1985)25 E Breval Z Deng S Chiou and C G Pantano J Mater Sci 27

1464 (1992)26 E D Rodeghiero O K Tse J Chisaki and E P Giannelis Mater

Sci Eng A 195 151 (1995)27 J F Moulder W F Stickle P E Sobol and K D Bomben

in Handbook of X-Ray Photoelectron Spectroscopy edited byJ Chastain and R C King Jr Physical Electronics Inc Minnesota(1995) p 71

28 T L Barr and S Seal J Vac Sci Technol A 13 1239 (1995)29 E Saiz R M Cannon and A P Tomsia Acta Mater 47 4209

(1999)30 W D Kingery J Am Ceram Soc 37 42 (1954)31 G Palasantzas and J Th M De Hosson Acta Mater 49 3533

(2001)32 N Eustathopolous Acta Mater 46 2319 (1998)33 J A Vreeling V Ocelrsquok Y T Pei D T L Van Agterveld and

J Th M De Hosson Acta Mater 48 4225 (2000)34 R Ye T Ishigaki J Jurewicz P Proulx and M I Boulos Plasma

Chem Plasma Process 24 555 (2004)35 A Brenner and S Senderoff J Res Natl Bur Stand 42 105 (1949)36 S Kuroda and T W Clyne Thin Solid Films 200 49 (1991)37 M W Hunt Guide to engineering materials Adv Mater Proc 158

6 (2000)38 L L Shaw D Goberman R Ren M Gell S Jiang Y Wang T D

Xiao and P R Strutt Surf Coat Technol 130 1 (2000)39 O D Slagle and H A Mckinsky J Appl Phys 38 437 (1967)40 M Vural S Zeytin and A H Ucisik Surf Coat Technol 97 347

(1997)

Received 10 July 2005 Accepted 12 October 2005

660 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

t1=0 t2

t3 t4

Ni

Alumina

Fig 5 Illustration of wettability of Nickel on Alumina with time t1 ltt2 lt t3 lt t4

with surface roughness temperature and with an angle ofincidence of the beam33 Typical values lie in an interval25ndash35 for room temperature polished surfaces and per-pendicular beam Absorption takes place in a surface layerapproximately 10minus4 mm in thickness

At different power of the laser beam the morphologyof the Nickel showed variation which is shown in SEMmicrograph in Figures 6 through 8 Due to the irradiationof laser beam electroless Ni layer covering the surface ofalumina melts instantaneously and due to surface tensiongets converted to globules The higher the amount of ther-mal energy the more is the driving force for the globulesto get coarsened The fact that each alumina particle has itsown characteristic size of Ni globules suggests that thereis not enough thermal contact between individual particlesto equilibrate different surface temperatures The sizes ofthe Ni globules observed by SEM ranges between 01 and05 m Some of the alumina particle surfaces irradiated

Fig 6 Ni particle size at a laser power of 250 W

Fig 7 Ni particle coarsens due to increase in the power of the laserbeam to 350 W

with high laser power is free of Ni globules but large Niglobules (sim5 m) are observed in their neighborhoods

Laser processing is actually advantageous from the factthat power of laser beam can be manipulated to get tothe suitable microstructure The presence of Ni particlesas small as 5 nm was revealed in HRTEM analysis of hightemperature processed Ni-coated alumina Lattice fringesof Ni nanoparticles could be resolved in the HRTEM studyas shown in Figure 9 The corresponding EDS spectrum(Fig 9b) analyzed from the selected area confirms thepresence of Ni Al and Oxygen Other elements that aredetected are either from the specimen grid or the Si waferused for specimen preparation However consolidation ofNialumina to bulk composite fabrication was not achievedsince the whole laser beam energy is absorbed in a thin Nilayer and a fast transformation of Ni layer in to spheroidsoccurred Therefore plasma processing route was chosento manufacture a free form bulk nanocomposite part

Fig 8 Ni particle further coarsens at a higher laser power of 400 W

J Nanosci Nanotechnol 6 651ndash660 2006 655

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

(a)

(b)

Fig 9 (a) HRTEM picture revealing the nanostructure retention of NindashAlumina nanocomposite (b) EDS spectrum confirming the presence ofNi in Alumina matrix

34 Plasma Processing

The rationale and science behind using plasma spray as afree form manufacturing technique and aluminum mandrelas a substrate for development of NindashAlumina bulk parthas been discussed in this section

341 PlasmandashParticle Interaction

Plasmandashparticle interaction results in the formation ofatoms in excited state ground state ions and photons inthe plasma plume As a result of these atomic dissocia-tion and recombination processes particle that is sprayedis taken to a partial molten state for a brief period of timetill it splats on to the substrate Time of flight the timerequired for the particle to reside in the flame depends onmaterial parameters such as the latent heat of fusion melt-ing temperature particle size The relation to determineresidence time can be given as34

timinusf =CplnTi minusTsminus lnTm minusTs

h middotAs

Table III Particle radius as a function of dwell time

Particle radius Dwell time (Sec)

50 micron 401times10minus4

30 micron 112times10minus4

20 micron 744times10minus5

1 micron 372times10minus6

50 nm 186times10minus7

where Cp Ti Tm h As are the density specificheat capacity volume initial temperature temperature ofthe plasma plume melting point heat transfer coefficientsurface area of the particle and Ts is the plasma plumetemperature The knowledge of residence time is essentialbefore spraying because too much of over heat may leadto evaporation losses Also higher the amount of super-heating the particles are subjected to rapid cooling as itimpinges on the substrate as a result of which shrinkagestresses will be more This might lead to cracking of thebulk part A calculation (Table III) for residence time tf fordifferent particle size NindashAlumina system has been tabledIt can be inferred that smaller particles need less residencetime which can controlled with the parameters mentionedabove To avoid evaporation and to ensure only surfacemelting appropriate gas flow combinations can be selectedto obtain the residence time of interest The residence timecalculated was 02 microseconds for a 50 nm particle topass through

342 Coating vs Free FormsmdashEngineering Aspects

The section describes the science behind the fabricationof a free form part which is different from coating Alu-minum and Copper being good conductors of heat areideal candidate materials for being used as a substrate formaterial deposition Besides they can be fabricated to anydesired shape as a result of their malleability and the highcoefficient of thermal expansion facilitates easy removalof the sprayed deposit Aluminum substrate can facilitateeasy removal by shrinking more compared to steel or cop-per Cooling and solidification of most materials is accom-panied by contraction or shrinkage As particles strike theyrapidly cool and solidify This generates a tensile stresswithin the particle and a compressive stress within the sur-face of the substrate As the coating builds up the ten-sile stresses developed in the deposit also increases Theresidual stress developed during the air plasma spray pro-cess significantly influences the integrity of the depositStresses developed during spraying can be classified into two parts viz quenching stresses and cooling stressesQuenching stresses are stresses that develop as the splatformation takes place due to particle impingement on thesubstrate The quantification of quenching stresses can bedone using3536

q =Eststs+54 middot13td

6 middot13R middot13tdand = Ed

Es

656 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

0

5

10

15

20

25

30

35

40

Alumina

Zircon

ia

Titania

MoS

i 2

Si 3N 4

Aluminu

m

Nickel

Titaniu

m

Coppe

r

Material

Str

ess

(MP

a)

Fig 10 Quenching stresses developed during the spray of differentparticles

where E is the elastic modulii of the substrate and thedeposit indicated by subscripts s and d respectively 13tis the thickness of the deposit and 13R is the change inthe radius of curvature of the coating The plots as shownin Figure 11 depict that different materials will developdifferent quenching stresses as the spray is going onBased on the thickness of the deposit achieved the quench-ing stresses will vary Among various spray materials forwhich quenching stresses have been calculated (Fig 10)The histogram shown for alumina and nickel particles withdeposit thickness of 0015 mm (15 microns) is of interestto this paper Cooling stresses are stresses that the depositproduces when the spraying is completed and the tempera-ture of the part is plummeting to room temperature It canbe accounted for using an equation37

c =EddminussTdminusTRT

1+2Ed middot tdEsts$

where EdmdashYoungrsquos modulus of the deposit TdmdashTemperature of the deposit TRTmdashRoom temperature ofthe sample after cooling dmdashCoefficient of thermalexpansion of the deposition smdashCoefficient of thermalexpansion of the substrate EsmdashYoungrsquos modulus of thesubstrate tsmdashthickness of the substrate and tdmdashthicknessof the deposition Figure 11 depicts the calculated cooling

ndash120

ndash100

ndash80

ndash60

ndash40

ndash20

0Alumina Zirconia Titania MoSi2 Si3N4

Coo

ling

Str

ess

(MP

a)

Copper

Aluminum

Fig 11 Cooling stresses developed in the deposit for Cu and Alsubstrates

stresses for different materials for two different substratesCopper and Aluminum The total amount of stresses devel-oped in the deposit is a combination of both quenchingand cooling stresses Quenching stresses which are ten-sile in nature keeps increasing with thickness build-upThe thermal stresses (cooling stresses) will be compres-sive when d lt s Selection of aluminum mandrel forNindashAlumina is justified based on the calculation that thecooling stresses are compressive in nature It will facilitatethe removal of spray deposit from the substrate This willeventually lead to free form fabrication of deposits

343 Material Characterization

The trial for spraying nano ceramic particles throughplasma flame has been successful The NindashAlumina cylin-drical bulk nanocomposite without any surface defectswas manufactured using plasma spray technique (Fig 12)Microstructure evaluation using SEM and other character-ization tools such as XRD and HRTEM is being discussedat lengthCross-Section Microscopy In an effort to study the

porosity distribution across the thickness scanning elec-tron microstructure of the internal center and outside ofthe thickness (Fig 13) has been analyzed The section is96ndash98 dense and does not show any significant cracksor porosity It also reveals the formation of splats due tohigher velocity with which the melted particles impingeson to the substrate There was no evidence of intercon-nected pores which could have been detrimental to theintegrity of the part Also the porosity distribution is morein case of internal thickness and less in outside thicknessThis can be explained from the fact that particles temper-ature gets reduced drastically during the initial stages ofspray as a result of which shrinkage is more while theparticles at the outside thickness has to undergo relativelyless temperature difference SEM micrographs of the crosssection revealed a clear demarcation between the forma-tion of folds and the formation of porosity The porositiesare revealed as dark regions compared to less intense foldsFold formation happens due to the inability of particles tofully splat on top of the existing splat

Substrate incompression

Free form partin Tension

(a) (b)

Fig 12 (a) The mandrel is subjected to compressive stresses whilethe sprayed part acquires tensile stress as the spray continues (b) A12primeprimeODtimes1primeprimeIDtimes15primeprime length cylinder was produced with full integrity

J Nanosci Nanotechnol 6 651ndash660 2006 657

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Folds

Isolated pores

Folds

(a)

(b)

(c)

Fig 13 Cross sectional low magnification micrographs indicating theextent of porosity formation at the (a) external (b) middle and (c) inter-nal portions of the bulk nanocomposite

344 Structural Characterization

Phase Transformation It is being observed that thealumina matrix undergone phase transformation fromalpha to gamma structure The alpha to gamma transi-tion depends on the ratio of the gun power to the primarygas flow in Standard Cubic Feet per Hour (SCFH) It ismentioned elsewhere [shaw] that if the parent phase is -alumina based on this ratio mentioned above the product

2θ20 30 40 50 60 70 80

(c)

(b)

(a)

NickelAlpha AluminaGamma AluminaNiO

Fig 14 Phase transformations associated with plasma spray of Ni-Al2O3 (a) Feedstock powder (b) plasma sprayed component (c) hydro-gen reduction

phase might be a combination of more -alumina and less alumina if the ratio is less than or equal to 240 or a com-bination of more alumina and less -alumina if the ratiois more than or equal to 310 The density and hardness val-ues achieved38 with the above ratios show significant dif-ferences with the former showing less hardness and densitythan the latter The phase transformation achieved in ourpresent case had more alumina The spray parameterswere manipulated to keep the ratio above 310 to makesure that dense and hard nanostructures were obtained Thediffractograms before and after plasma spraying is shownin Figure 14 X-ray diffraction spectrums from the plasmaspray deposition and powder mixture prior to the spray-ing are compared Both the diffractograms reveal the pres-ence of FCC Ni and Alumina The difference betweenthe diffractograms is that of the peak broadening Dur-ing the spray particles literally gets quenched from theirmelting temperatures towards the substrate preheat temper-ature Such large temperatures gradients are expected toimpart residual stresses on the sprayed part The stressescan be calculated based on the difference in ldquodrdquo spacingof the particles prior to spraying and after spraying Fromthe modulus of elasticity values the amount of residualstress can be calculated as follows39

Eminus1hkl =

C11 +C12

C11 minusC12C11 +2C12

+(

1C44

minus 2C11 minusC12

)l2m2 +m2n2 +n2l2

The compliance constants are C11 = 2465 GPa C12 =1473 GPa and C44 = 1247 GPa The calculated residualstress values are presented in Table IV

Since the process has been carried out in atmosphericair Ni has the propensity to get converted to its surfaceoxide Gibbs free energy of formation of Nickel oxide upto melting point of Nickel is negative Although Gibbs free

658 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table IV Residual stress measurements using X-ray diffraction techniques

hkl 2 (Measured) 2 (JCPDS) d (Measured) Aring d (JCPDS) Aring Strain () E (GPa) Stress (Mpa)

NiO 200 4326 4329 209 209 0066 136 90NiO 111 3725 3728 241 241 0078 304 236NiO 220 6289 6292 148 148 0043 232 99-Alumina 440 67 67032 140 139 0042 408 172

energy curve for formation of Nickel oxide is sloped upwith temperature the partial pressure of oxygen at NTP ismuch higher for the reaction to proceed forward and formNiO APS process carried out at dehumidified air will havean oxygen partial pressure of 02 atm which is more thansufficient for the NiO formation to occur Moreover if thestand off distance is more Nickel in its molten state picksup more oxygen from the atmosphere than it would haveif it were in solid state By reducing the stand off distancethe amount of formation of Nickel oxide can be reducedbut can not be totally eliminated XRD peaks confirmedthe formation of NiO An effort has been made to bringNiO back to its metallic state The plasma spray formednanocomposite was reduced in an atmosphere of Hydrogenat 1000 C for 1 hour

345 High Resolution TransmissionElectron Microscopy

Alumina grain boundaries being high energy regions thereis more probability of the presence of Nickel along thegrain boundaries In order to determine the location ofNickel EDS has been performed along the grain bound-aries to select spots for imaging of Nickel nano parti-cles In those spots where strong Nickel signal have beendetected STEM imaging has been done to determine themorphology and the size of the Nickel particles Spheri-cal Ni particles with 30 nm size have been imaged andare shown in Figure 15 Another evidence of nanostruc-ture retention of Nickel particle is the SAD pattern whichshows a spotty ring pattern as revealed in Figure 15(a)(inset) This pattern is a result of extremely fine grainsThe possible explanations of the nano structure retentionin NindashAlumina nanocomposite could be the lack of timepermitted for the coarsening of Nickel particles travers-ing through the high temperature zone (plasma flame) On

0100200300400500600700800900

1000

0 2000 4000 6000 8000 10000

Energy (eV)

Inte

nsity

(A

U)

NiCu

Al

O

20 nm

Fig 15 STEM pictures depicting nanostructure retention of nickel par-ticles and the corresponding EDS spectrum confirming the presence ofnickel

the other hand during conventional sintering a significanttime is spent in the temperature range where the coarsen-ing effect is appreciable20 Besides in the nanostructuressurface melting is preferred in the fast translation thuskeeping the core intact From the TEM pictures of plasmasprayed NickelndashAlumina it is envisioned that plasma pro-cessing is considered to be a potential tool to fabricate freeform (eg cylindrical) bulk composites with nanostructureretention and improved mechanical properties

35 Mechanical Properties

Nickel plays a dominant role in increasing the strengthof the ceramic matrix The presence of cracks in ceram-ics makes them more brittle compared to metals At thetip of the crack as shown in Figure 16 there is an enor-mous amount of stress concentration developed as a resultof which the cracks tend to propagate easily The surfaceenergy required to propagate the crack is more than com-pensated by the release of the elastic strain energy Thepresence of Nickel spheroids dispersed uniformly through-out the matrix blunts the crack thereby making it more dif-ficult to propagate as shown in Figure 16 Nickel particleswhich are highly plastic in nature aids in crack bluntingand improvement of fracture toughness of the nanocom-posite Lieberthal et al16 proposed a model in which acrack propagating along the alumina grain boundaries aredeflected along around the Ni particle located in grainboundaries which aid in improving the fracture toughnessof the ceramic nanocomposite

The hardness and fracture toughness of the NindashAluminaplasma sprayed sample that could be achieved with thegiven set of plasma spray parameters were measured tobe 1025 HV and sim5 MPa middot m12 The hardness valuesachieved were found to be better than that achieved for the

Crack tip in a ceramicmatrix

Ni-particle

Fig 16 Nickel particles blunt the crack and arrest crack propagationimproves the fracture toughness

J Nanosci Nanotechnol 6 651ndash660 2006 659

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

monolithic alumina developed using plasma spray (940plusmn20 HV)40 The marked increase in the fracture toughnesscompared to other processing techniques can be attributedto the crack blunting capability of ductile Nickel It shouldbe noted that NindashAlumina nanocomposites made throughtechniques such as hot press sintering12 reported a frac-ture toughness of sim35 MPa middotm12 In other words the increase in fracture toughness was approximately 40

4 CONCLUSIONS

Electroless Nickel coated alumina particles have been suc-cessfully used to manufacture a free-form bulk nanocom-posite with improved hardness and fracture toughnessWettability study of Nickel on polycrystalline Aluminarevealed the contact angle which validated the self affineroughness model being used to calculate the apparent con-tact angle Laser processing had shortcomings from thepoint of view of time and the high laser required to conso-lidate the powders to bulk form especially in the casemetalceramic matrix For plasma processing substratematerial selection is important based on the material com-position to be sprayed Calculation of residual High frac-ture toughness can be attributed to the nanostructuredNickel which arrests the crack propagation Thus plasmaspray processing is an effective technique to make freeform bulk nanostructures

Acknowledgments The authors wish to thank theOffice of Naval Research Young Investigator Award (ONRYIP-N000140210591) for their financial support Theauthors would also like to thank Plasma Processes Incfor their help in product development We also appreci-ate the equipment support from Materials CharacterizationFacility at the University of Central Florida

References and Notes

1 W H Tuan and R J Brook J Eur Ceram Soc 6 31 (1990)2 A Paesano Jr C K Matsuda J B M Da Cunha M A Z

Vasconcellos B Hallouche and S L Silva J Magn Magn Mater264 264 (2003)

3 T Sekino and K Niihara J Mater Sci 32 3943 (1997)4 T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)5 A Nakahira and K Niihara Fract Mech 9 165 (1992)6 Y K Jeong and K Niihara Nanostruct Mater 9 193 (1997)7 S C Wang and W C J Wei Nanostruct Mater 10 983 (1998)

8 B Budiansky J C Amazigo and A J Evans J Mech Phys Solids36 167 (1988)

9 S Ozkar G A Ozin and R A Prokopowicz Chem Mater 4 1380(1992)

10 S T Oh T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)11 O Aharon S Bar-Ziv D Gorni T Cohen-Hyams and W D

Kaplan Scripta Mater 50 1209 (2004)12 T Sekino T Nakajima and K Niihara Mater Lett 29 165 (1996)13 S I Cha S H Hong and B K Kim Mater Sci Eng A 351 31

(2003)14 S Seal S C Kuiry P Georgieva and K Rea Scripta Mater 50

1237 (2004)15 R Z Chen and W H Tuan J Eur Ceram Soc 19 463 (1999)16 M Lieberthal and W D Kaplan Mater Sci Eng A 302 83 (2001)17 K Niihara B S Kim T Nakayama T Kusunose T Nomoto

A Hikasa and T Sekino J Eur Ceram Soc 24 3419 (2004)18 U L Adisorn T Matsunaga Y Kobayashi S M Choi and

H Awaji Ceram Intl 31 803 (2005)19 S Seal S C Kuiry P Georgieva and A Agarwal MRS Bull 29

16 (2004)20 J Michalski K Konopka and M Trzaska Mater Chem Phys 81

407 (2003)21 A G Evans and E A Charles J Am Ceram Soc 59 371 (1976)22 W H Tuan and R J Brook J Eur Ceram Soc 10 95 (1992)23 W H Tuan and W B Chou J Am Ceram Soc 80 2418 (1997)24 E Breval G Dodds and C G Pantano Mater Res Bull 20 1191

(1985)25 E Breval Z Deng S Chiou and C G Pantano J Mater Sci 27

1464 (1992)26 E D Rodeghiero O K Tse J Chisaki and E P Giannelis Mater

Sci Eng A 195 151 (1995)27 J F Moulder W F Stickle P E Sobol and K D Bomben

in Handbook of X-Ray Photoelectron Spectroscopy edited byJ Chastain and R C King Jr Physical Electronics Inc Minnesota(1995) p 71

28 T L Barr and S Seal J Vac Sci Technol A 13 1239 (1995)29 E Saiz R M Cannon and A P Tomsia Acta Mater 47 4209

(1999)30 W D Kingery J Am Ceram Soc 37 42 (1954)31 G Palasantzas and J Th M De Hosson Acta Mater 49 3533

(2001)32 N Eustathopolous Acta Mater 46 2319 (1998)33 J A Vreeling V Ocelrsquok Y T Pei D T L Van Agterveld and

J Th M De Hosson Acta Mater 48 4225 (2000)34 R Ye T Ishigaki J Jurewicz P Proulx and M I Boulos Plasma

Chem Plasma Process 24 555 (2004)35 A Brenner and S Senderoff J Res Natl Bur Stand 42 105 (1949)36 S Kuroda and T W Clyne Thin Solid Films 200 49 (1991)37 M W Hunt Guide to engineering materials Adv Mater Proc 158

6 (2000)38 L L Shaw D Goberman R Ren M Gell S Jiang Y Wang T D

Xiao and P R Strutt Surf Coat Technol 130 1 (2000)39 O D Slagle and H A Mckinsky J Appl Phys 38 437 (1967)40 M Vural S Zeytin and A H Ucisik Surf Coat Technol 97 347

(1997)

Received 10 July 2005 Accepted 12 October 2005

660 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

(a)

(b)

Fig 9 (a) HRTEM picture revealing the nanostructure retention of NindashAlumina nanocomposite (b) EDS spectrum confirming the presence ofNi in Alumina matrix

34 Plasma Processing

The rationale and science behind using plasma spray as afree form manufacturing technique and aluminum mandrelas a substrate for development of NindashAlumina bulk parthas been discussed in this section

341 PlasmandashParticle Interaction

Plasmandashparticle interaction results in the formation ofatoms in excited state ground state ions and photons inthe plasma plume As a result of these atomic dissocia-tion and recombination processes particle that is sprayedis taken to a partial molten state for a brief period of timetill it splats on to the substrate Time of flight the timerequired for the particle to reside in the flame depends onmaterial parameters such as the latent heat of fusion melt-ing temperature particle size The relation to determineresidence time can be given as34

timinusf =CplnTi minusTsminus lnTm minusTs

h middotAs

Table III Particle radius as a function of dwell time

Particle radius Dwell time (Sec)

50 micron 401times10minus4

30 micron 112times10minus4

20 micron 744times10minus5

1 micron 372times10minus6

50 nm 186times10minus7

where Cp Ti Tm h As are the density specificheat capacity volume initial temperature temperature ofthe plasma plume melting point heat transfer coefficientsurface area of the particle and Ts is the plasma plumetemperature The knowledge of residence time is essentialbefore spraying because too much of over heat may leadto evaporation losses Also higher the amount of super-heating the particles are subjected to rapid cooling as itimpinges on the substrate as a result of which shrinkagestresses will be more This might lead to cracking of thebulk part A calculation (Table III) for residence time tf fordifferent particle size NindashAlumina system has been tabledIt can be inferred that smaller particles need less residencetime which can controlled with the parameters mentionedabove To avoid evaporation and to ensure only surfacemelting appropriate gas flow combinations can be selectedto obtain the residence time of interest The residence timecalculated was 02 microseconds for a 50 nm particle topass through

342 Coating vs Free FormsmdashEngineering Aspects

The section describes the science behind the fabricationof a free form part which is different from coating Alu-minum and Copper being good conductors of heat areideal candidate materials for being used as a substrate formaterial deposition Besides they can be fabricated to anydesired shape as a result of their malleability and the highcoefficient of thermal expansion facilitates easy removalof the sprayed deposit Aluminum substrate can facilitateeasy removal by shrinking more compared to steel or cop-per Cooling and solidification of most materials is accom-panied by contraction or shrinkage As particles strike theyrapidly cool and solidify This generates a tensile stresswithin the particle and a compressive stress within the sur-face of the substrate As the coating builds up the ten-sile stresses developed in the deposit also increases Theresidual stress developed during the air plasma spray pro-cess significantly influences the integrity of the depositStresses developed during spraying can be classified into two parts viz quenching stresses and cooling stressesQuenching stresses are stresses that develop as the splatformation takes place due to particle impingement on thesubstrate The quantification of quenching stresses can bedone using3536

q =Eststs+54 middot13td

6 middot13R middot13tdand = Ed

Es

656 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

0

5

10

15

20

25

30

35

40

Alumina

Zircon

ia

Titania

MoS

i 2

Si 3N 4

Aluminu

m

Nickel

Titaniu

m

Coppe

r

Material

Str

ess

(MP

a)

Fig 10 Quenching stresses developed during the spray of differentparticles

where E is the elastic modulii of the substrate and thedeposit indicated by subscripts s and d respectively 13tis the thickness of the deposit and 13R is the change inthe radius of curvature of the coating The plots as shownin Figure 11 depict that different materials will developdifferent quenching stresses as the spray is going onBased on the thickness of the deposit achieved the quench-ing stresses will vary Among various spray materials forwhich quenching stresses have been calculated (Fig 10)The histogram shown for alumina and nickel particles withdeposit thickness of 0015 mm (15 microns) is of interestto this paper Cooling stresses are stresses that the depositproduces when the spraying is completed and the tempera-ture of the part is plummeting to room temperature It canbe accounted for using an equation37

c =EddminussTdminusTRT

1+2Ed middot tdEsts$

where EdmdashYoungrsquos modulus of the deposit TdmdashTemperature of the deposit TRTmdashRoom temperature ofthe sample after cooling dmdashCoefficient of thermalexpansion of the deposition smdashCoefficient of thermalexpansion of the substrate EsmdashYoungrsquos modulus of thesubstrate tsmdashthickness of the substrate and tdmdashthicknessof the deposition Figure 11 depicts the calculated cooling

ndash120

ndash100

ndash80

ndash60

ndash40

ndash20

0Alumina Zirconia Titania MoSi2 Si3N4

Coo

ling

Str

ess

(MP

a)

Copper

Aluminum

Fig 11 Cooling stresses developed in the deposit for Cu and Alsubstrates

stresses for different materials for two different substratesCopper and Aluminum The total amount of stresses devel-oped in the deposit is a combination of both quenchingand cooling stresses Quenching stresses which are ten-sile in nature keeps increasing with thickness build-upThe thermal stresses (cooling stresses) will be compres-sive when d lt s Selection of aluminum mandrel forNindashAlumina is justified based on the calculation that thecooling stresses are compressive in nature It will facilitatethe removal of spray deposit from the substrate This willeventually lead to free form fabrication of deposits

343 Material Characterization

The trial for spraying nano ceramic particles throughplasma flame has been successful The NindashAlumina cylin-drical bulk nanocomposite without any surface defectswas manufactured using plasma spray technique (Fig 12)Microstructure evaluation using SEM and other character-ization tools such as XRD and HRTEM is being discussedat lengthCross-Section Microscopy In an effort to study the

porosity distribution across the thickness scanning elec-tron microstructure of the internal center and outside ofthe thickness (Fig 13) has been analyzed The section is96ndash98 dense and does not show any significant cracksor porosity It also reveals the formation of splats due tohigher velocity with which the melted particles impingeson to the substrate There was no evidence of intercon-nected pores which could have been detrimental to theintegrity of the part Also the porosity distribution is morein case of internal thickness and less in outside thicknessThis can be explained from the fact that particles temper-ature gets reduced drastically during the initial stages ofspray as a result of which shrinkage is more while theparticles at the outside thickness has to undergo relativelyless temperature difference SEM micrographs of the crosssection revealed a clear demarcation between the forma-tion of folds and the formation of porosity The porositiesare revealed as dark regions compared to less intense foldsFold formation happens due to the inability of particles tofully splat on top of the existing splat

Substrate incompression

Free form partin Tension

(a) (b)

Fig 12 (a) The mandrel is subjected to compressive stresses whilethe sprayed part acquires tensile stress as the spray continues (b) A12primeprimeODtimes1primeprimeIDtimes15primeprime length cylinder was produced with full integrity

J Nanosci Nanotechnol 6 651ndash660 2006 657

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Folds

Isolated pores

Folds

(a)

(b)

(c)

Fig 13 Cross sectional low magnification micrographs indicating theextent of porosity formation at the (a) external (b) middle and (c) inter-nal portions of the bulk nanocomposite

344 Structural Characterization

Phase Transformation It is being observed that thealumina matrix undergone phase transformation fromalpha to gamma structure The alpha to gamma transi-tion depends on the ratio of the gun power to the primarygas flow in Standard Cubic Feet per Hour (SCFH) It ismentioned elsewhere [shaw] that if the parent phase is -alumina based on this ratio mentioned above the product

2θ20 30 40 50 60 70 80

(c)

(b)

(a)

NickelAlpha AluminaGamma AluminaNiO

Fig 14 Phase transformations associated with plasma spray of Ni-Al2O3 (a) Feedstock powder (b) plasma sprayed component (c) hydro-gen reduction

phase might be a combination of more -alumina and less alumina if the ratio is less than or equal to 240 or a com-bination of more alumina and less -alumina if the ratiois more than or equal to 310 The density and hardness val-ues achieved38 with the above ratios show significant dif-ferences with the former showing less hardness and densitythan the latter The phase transformation achieved in ourpresent case had more alumina The spray parameterswere manipulated to keep the ratio above 310 to makesure that dense and hard nanostructures were obtained Thediffractograms before and after plasma spraying is shownin Figure 14 X-ray diffraction spectrums from the plasmaspray deposition and powder mixture prior to the spray-ing are compared Both the diffractograms reveal the pres-ence of FCC Ni and Alumina The difference betweenthe diffractograms is that of the peak broadening Dur-ing the spray particles literally gets quenched from theirmelting temperatures towards the substrate preheat temper-ature Such large temperatures gradients are expected toimpart residual stresses on the sprayed part The stressescan be calculated based on the difference in ldquodrdquo spacingof the particles prior to spraying and after spraying Fromthe modulus of elasticity values the amount of residualstress can be calculated as follows39

Eminus1hkl =

C11 +C12

C11 minusC12C11 +2C12

+(

1C44

minus 2C11 minusC12

)l2m2 +m2n2 +n2l2

The compliance constants are C11 = 2465 GPa C12 =1473 GPa and C44 = 1247 GPa The calculated residualstress values are presented in Table IV

Since the process has been carried out in atmosphericair Ni has the propensity to get converted to its surfaceoxide Gibbs free energy of formation of Nickel oxide upto melting point of Nickel is negative Although Gibbs free

658 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table IV Residual stress measurements using X-ray diffraction techniques

hkl 2 (Measured) 2 (JCPDS) d (Measured) Aring d (JCPDS) Aring Strain () E (GPa) Stress (Mpa)

NiO 200 4326 4329 209 209 0066 136 90NiO 111 3725 3728 241 241 0078 304 236NiO 220 6289 6292 148 148 0043 232 99-Alumina 440 67 67032 140 139 0042 408 172

energy curve for formation of Nickel oxide is sloped upwith temperature the partial pressure of oxygen at NTP ismuch higher for the reaction to proceed forward and formNiO APS process carried out at dehumidified air will havean oxygen partial pressure of 02 atm which is more thansufficient for the NiO formation to occur Moreover if thestand off distance is more Nickel in its molten state picksup more oxygen from the atmosphere than it would haveif it were in solid state By reducing the stand off distancethe amount of formation of Nickel oxide can be reducedbut can not be totally eliminated XRD peaks confirmedthe formation of NiO An effort has been made to bringNiO back to its metallic state The plasma spray formednanocomposite was reduced in an atmosphere of Hydrogenat 1000 C for 1 hour

345 High Resolution TransmissionElectron Microscopy

Alumina grain boundaries being high energy regions thereis more probability of the presence of Nickel along thegrain boundaries In order to determine the location ofNickel EDS has been performed along the grain bound-aries to select spots for imaging of Nickel nano parti-cles In those spots where strong Nickel signal have beendetected STEM imaging has been done to determine themorphology and the size of the Nickel particles Spheri-cal Ni particles with 30 nm size have been imaged andare shown in Figure 15 Another evidence of nanostruc-ture retention of Nickel particle is the SAD pattern whichshows a spotty ring pattern as revealed in Figure 15(a)(inset) This pattern is a result of extremely fine grainsThe possible explanations of the nano structure retentionin NindashAlumina nanocomposite could be the lack of timepermitted for the coarsening of Nickel particles travers-ing through the high temperature zone (plasma flame) On

0100200300400500600700800900

1000

0 2000 4000 6000 8000 10000

Energy (eV)

Inte

nsity

(A

U)

NiCu

Al

O

20 nm

Fig 15 STEM pictures depicting nanostructure retention of nickel par-ticles and the corresponding EDS spectrum confirming the presence ofnickel

the other hand during conventional sintering a significanttime is spent in the temperature range where the coarsen-ing effect is appreciable20 Besides in the nanostructuressurface melting is preferred in the fast translation thuskeeping the core intact From the TEM pictures of plasmasprayed NickelndashAlumina it is envisioned that plasma pro-cessing is considered to be a potential tool to fabricate freeform (eg cylindrical) bulk composites with nanostructureretention and improved mechanical properties

35 Mechanical Properties

Nickel plays a dominant role in increasing the strengthof the ceramic matrix The presence of cracks in ceram-ics makes them more brittle compared to metals At thetip of the crack as shown in Figure 16 there is an enor-mous amount of stress concentration developed as a resultof which the cracks tend to propagate easily The surfaceenergy required to propagate the crack is more than com-pensated by the release of the elastic strain energy Thepresence of Nickel spheroids dispersed uniformly through-out the matrix blunts the crack thereby making it more dif-ficult to propagate as shown in Figure 16 Nickel particleswhich are highly plastic in nature aids in crack bluntingand improvement of fracture toughness of the nanocom-posite Lieberthal et al16 proposed a model in which acrack propagating along the alumina grain boundaries aredeflected along around the Ni particle located in grainboundaries which aid in improving the fracture toughnessof the ceramic nanocomposite

The hardness and fracture toughness of the NindashAluminaplasma sprayed sample that could be achieved with thegiven set of plasma spray parameters were measured tobe 1025 HV and sim5 MPa middot m12 The hardness valuesachieved were found to be better than that achieved for the

Crack tip in a ceramicmatrix

Ni-particle

Fig 16 Nickel particles blunt the crack and arrest crack propagationimproves the fracture toughness

J Nanosci Nanotechnol 6 651ndash660 2006 659

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

monolithic alumina developed using plasma spray (940plusmn20 HV)40 The marked increase in the fracture toughnesscompared to other processing techniques can be attributedto the crack blunting capability of ductile Nickel It shouldbe noted that NindashAlumina nanocomposites made throughtechniques such as hot press sintering12 reported a frac-ture toughness of sim35 MPa middotm12 In other words the increase in fracture toughness was approximately 40

4 CONCLUSIONS

Electroless Nickel coated alumina particles have been suc-cessfully used to manufacture a free-form bulk nanocom-posite with improved hardness and fracture toughnessWettability study of Nickel on polycrystalline Aluminarevealed the contact angle which validated the self affineroughness model being used to calculate the apparent con-tact angle Laser processing had shortcomings from thepoint of view of time and the high laser required to conso-lidate the powders to bulk form especially in the casemetalceramic matrix For plasma processing substratematerial selection is important based on the material com-position to be sprayed Calculation of residual High frac-ture toughness can be attributed to the nanostructuredNickel which arrests the crack propagation Thus plasmaspray processing is an effective technique to make freeform bulk nanostructures

Acknowledgments The authors wish to thank theOffice of Naval Research Young Investigator Award (ONRYIP-N000140210591) for their financial support Theauthors would also like to thank Plasma Processes Incfor their help in product development We also appreci-ate the equipment support from Materials CharacterizationFacility at the University of Central Florida

References and Notes

1 W H Tuan and R J Brook J Eur Ceram Soc 6 31 (1990)2 A Paesano Jr C K Matsuda J B M Da Cunha M A Z

Vasconcellos B Hallouche and S L Silva J Magn Magn Mater264 264 (2003)

3 T Sekino and K Niihara J Mater Sci 32 3943 (1997)4 T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)5 A Nakahira and K Niihara Fract Mech 9 165 (1992)6 Y K Jeong and K Niihara Nanostruct Mater 9 193 (1997)7 S C Wang and W C J Wei Nanostruct Mater 10 983 (1998)

8 B Budiansky J C Amazigo and A J Evans J Mech Phys Solids36 167 (1988)

9 S Ozkar G A Ozin and R A Prokopowicz Chem Mater 4 1380(1992)

10 S T Oh T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)11 O Aharon S Bar-Ziv D Gorni T Cohen-Hyams and W D

Kaplan Scripta Mater 50 1209 (2004)12 T Sekino T Nakajima and K Niihara Mater Lett 29 165 (1996)13 S I Cha S H Hong and B K Kim Mater Sci Eng A 351 31

(2003)14 S Seal S C Kuiry P Georgieva and K Rea Scripta Mater 50

1237 (2004)15 R Z Chen and W H Tuan J Eur Ceram Soc 19 463 (1999)16 M Lieberthal and W D Kaplan Mater Sci Eng A 302 83 (2001)17 K Niihara B S Kim T Nakayama T Kusunose T Nomoto

A Hikasa and T Sekino J Eur Ceram Soc 24 3419 (2004)18 U L Adisorn T Matsunaga Y Kobayashi S M Choi and

H Awaji Ceram Intl 31 803 (2005)19 S Seal S C Kuiry P Georgieva and A Agarwal MRS Bull 29

16 (2004)20 J Michalski K Konopka and M Trzaska Mater Chem Phys 81

407 (2003)21 A G Evans and E A Charles J Am Ceram Soc 59 371 (1976)22 W H Tuan and R J Brook J Eur Ceram Soc 10 95 (1992)23 W H Tuan and W B Chou J Am Ceram Soc 80 2418 (1997)24 E Breval G Dodds and C G Pantano Mater Res Bull 20 1191

(1985)25 E Breval Z Deng S Chiou and C G Pantano J Mater Sci 27

1464 (1992)26 E D Rodeghiero O K Tse J Chisaki and E P Giannelis Mater

Sci Eng A 195 151 (1995)27 J F Moulder W F Stickle P E Sobol and K D Bomben

in Handbook of X-Ray Photoelectron Spectroscopy edited byJ Chastain and R C King Jr Physical Electronics Inc Minnesota(1995) p 71

28 T L Barr and S Seal J Vac Sci Technol A 13 1239 (1995)29 E Saiz R M Cannon and A P Tomsia Acta Mater 47 4209

(1999)30 W D Kingery J Am Ceram Soc 37 42 (1954)31 G Palasantzas and J Th M De Hosson Acta Mater 49 3533

(2001)32 N Eustathopolous Acta Mater 46 2319 (1998)33 J A Vreeling V Ocelrsquok Y T Pei D T L Van Agterveld and

J Th M De Hosson Acta Mater 48 4225 (2000)34 R Ye T Ishigaki J Jurewicz P Proulx and M I Boulos Plasma

Chem Plasma Process 24 555 (2004)35 A Brenner and S Senderoff J Res Natl Bur Stand 42 105 (1949)36 S Kuroda and T W Clyne Thin Solid Films 200 49 (1991)37 M W Hunt Guide to engineering materials Adv Mater Proc 158

6 (2000)38 L L Shaw D Goberman R Ren M Gell S Jiang Y Wang T D

Xiao and P R Strutt Surf Coat Technol 130 1 (2000)39 O D Slagle and H A Mckinsky J Appl Phys 38 437 (1967)40 M Vural S Zeytin and A H Ucisik Surf Coat Technol 97 347

(1997)

Received 10 July 2005 Accepted 12 October 2005

660 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

0

5

10

15

20

25

30

35

40

Alumina

Zircon

ia

Titania

MoS

i 2

Si 3N 4

Aluminu

m

Nickel

Titaniu

m

Coppe

r

Material

Str

ess

(MP

a)

Fig 10 Quenching stresses developed during the spray of differentparticles

where E is the elastic modulii of the substrate and thedeposit indicated by subscripts s and d respectively 13tis the thickness of the deposit and 13R is the change inthe radius of curvature of the coating The plots as shownin Figure 11 depict that different materials will developdifferent quenching stresses as the spray is going onBased on the thickness of the deposit achieved the quench-ing stresses will vary Among various spray materials forwhich quenching stresses have been calculated (Fig 10)The histogram shown for alumina and nickel particles withdeposit thickness of 0015 mm (15 microns) is of interestto this paper Cooling stresses are stresses that the depositproduces when the spraying is completed and the tempera-ture of the part is plummeting to room temperature It canbe accounted for using an equation37

c =EddminussTdminusTRT

1+2Ed middot tdEsts$

where EdmdashYoungrsquos modulus of the deposit TdmdashTemperature of the deposit TRTmdashRoom temperature ofthe sample after cooling dmdashCoefficient of thermalexpansion of the deposition smdashCoefficient of thermalexpansion of the substrate EsmdashYoungrsquos modulus of thesubstrate tsmdashthickness of the substrate and tdmdashthicknessof the deposition Figure 11 depicts the calculated cooling

ndash120

ndash100

ndash80

ndash60

ndash40

ndash20

0Alumina Zirconia Titania MoSi2 Si3N4

Coo

ling

Str

ess

(MP

a)

Copper

Aluminum

Fig 11 Cooling stresses developed in the deposit for Cu and Alsubstrates

stresses for different materials for two different substratesCopper and Aluminum The total amount of stresses devel-oped in the deposit is a combination of both quenchingand cooling stresses Quenching stresses which are ten-sile in nature keeps increasing with thickness build-upThe thermal stresses (cooling stresses) will be compres-sive when d lt s Selection of aluminum mandrel forNindashAlumina is justified based on the calculation that thecooling stresses are compressive in nature It will facilitatethe removal of spray deposit from the substrate This willeventually lead to free form fabrication of deposits

343 Material Characterization

The trial for spraying nano ceramic particles throughplasma flame has been successful The NindashAlumina cylin-drical bulk nanocomposite without any surface defectswas manufactured using plasma spray technique (Fig 12)Microstructure evaluation using SEM and other character-ization tools such as XRD and HRTEM is being discussedat lengthCross-Section Microscopy In an effort to study the

porosity distribution across the thickness scanning elec-tron microstructure of the internal center and outside ofthe thickness (Fig 13) has been analyzed The section is96ndash98 dense and does not show any significant cracksor porosity It also reveals the formation of splats due tohigher velocity with which the melted particles impingeson to the substrate There was no evidence of intercon-nected pores which could have been detrimental to theintegrity of the part Also the porosity distribution is morein case of internal thickness and less in outside thicknessThis can be explained from the fact that particles temper-ature gets reduced drastically during the initial stages ofspray as a result of which shrinkage is more while theparticles at the outside thickness has to undergo relativelyless temperature difference SEM micrographs of the crosssection revealed a clear demarcation between the forma-tion of folds and the formation of porosity The porositiesare revealed as dark regions compared to less intense foldsFold formation happens due to the inability of particles tofully splat on top of the existing splat

Substrate incompression

Free form partin Tension

(a) (b)

Fig 12 (a) The mandrel is subjected to compressive stresses whilethe sprayed part acquires tensile stress as the spray continues (b) A12primeprimeODtimes1primeprimeIDtimes15primeprime length cylinder was produced with full integrity

J Nanosci Nanotechnol 6 651ndash660 2006 657

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Folds

Isolated pores

Folds

(a)

(b)

(c)

Fig 13 Cross sectional low magnification micrographs indicating theextent of porosity formation at the (a) external (b) middle and (c) inter-nal portions of the bulk nanocomposite

344 Structural Characterization

Phase Transformation It is being observed that thealumina matrix undergone phase transformation fromalpha to gamma structure The alpha to gamma transi-tion depends on the ratio of the gun power to the primarygas flow in Standard Cubic Feet per Hour (SCFH) It ismentioned elsewhere [shaw] that if the parent phase is -alumina based on this ratio mentioned above the product

2θ20 30 40 50 60 70 80

(c)

(b)

(a)

NickelAlpha AluminaGamma AluminaNiO

Fig 14 Phase transformations associated with plasma spray of Ni-Al2O3 (a) Feedstock powder (b) plasma sprayed component (c) hydro-gen reduction

phase might be a combination of more -alumina and less alumina if the ratio is less than or equal to 240 or a com-bination of more alumina and less -alumina if the ratiois more than or equal to 310 The density and hardness val-ues achieved38 with the above ratios show significant dif-ferences with the former showing less hardness and densitythan the latter The phase transformation achieved in ourpresent case had more alumina The spray parameterswere manipulated to keep the ratio above 310 to makesure that dense and hard nanostructures were obtained Thediffractograms before and after plasma spraying is shownin Figure 14 X-ray diffraction spectrums from the plasmaspray deposition and powder mixture prior to the spray-ing are compared Both the diffractograms reveal the pres-ence of FCC Ni and Alumina The difference betweenthe diffractograms is that of the peak broadening Dur-ing the spray particles literally gets quenched from theirmelting temperatures towards the substrate preheat temper-ature Such large temperatures gradients are expected toimpart residual stresses on the sprayed part The stressescan be calculated based on the difference in ldquodrdquo spacingof the particles prior to spraying and after spraying Fromthe modulus of elasticity values the amount of residualstress can be calculated as follows39

Eminus1hkl =

C11 +C12

C11 minusC12C11 +2C12

+(

1C44

minus 2C11 minusC12

)l2m2 +m2n2 +n2l2

The compliance constants are C11 = 2465 GPa C12 =1473 GPa and C44 = 1247 GPa The calculated residualstress values are presented in Table IV

Since the process has been carried out in atmosphericair Ni has the propensity to get converted to its surfaceoxide Gibbs free energy of formation of Nickel oxide upto melting point of Nickel is negative Although Gibbs free

658 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table IV Residual stress measurements using X-ray diffraction techniques

hkl 2 (Measured) 2 (JCPDS) d (Measured) Aring d (JCPDS) Aring Strain () E (GPa) Stress (Mpa)

NiO 200 4326 4329 209 209 0066 136 90NiO 111 3725 3728 241 241 0078 304 236NiO 220 6289 6292 148 148 0043 232 99-Alumina 440 67 67032 140 139 0042 408 172

energy curve for formation of Nickel oxide is sloped upwith temperature the partial pressure of oxygen at NTP ismuch higher for the reaction to proceed forward and formNiO APS process carried out at dehumidified air will havean oxygen partial pressure of 02 atm which is more thansufficient for the NiO formation to occur Moreover if thestand off distance is more Nickel in its molten state picksup more oxygen from the atmosphere than it would haveif it were in solid state By reducing the stand off distancethe amount of formation of Nickel oxide can be reducedbut can not be totally eliminated XRD peaks confirmedthe formation of NiO An effort has been made to bringNiO back to its metallic state The plasma spray formednanocomposite was reduced in an atmosphere of Hydrogenat 1000 C for 1 hour

345 High Resolution TransmissionElectron Microscopy

Alumina grain boundaries being high energy regions thereis more probability of the presence of Nickel along thegrain boundaries In order to determine the location ofNickel EDS has been performed along the grain bound-aries to select spots for imaging of Nickel nano parti-cles In those spots where strong Nickel signal have beendetected STEM imaging has been done to determine themorphology and the size of the Nickel particles Spheri-cal Ni particles with 30 nm size have been imaged andare shown in Figure 15 Another evidence of nanostruc-ture retention of Nickel particle is the SAD pattern whichshows a spotty ring pattern as revealed in Figure 15(a)(inset) This pattern is a result of extremely fine grainsThe possible explanations of the nano structure retentionin NindashAlumina nanocomposite could be the lack of timepermitted for the coarsening of Nickel particles travers-ing through the high temperature zone (plasma flame) On

0100200300400500600700800900

1000

0 2000 4000 6000 8000 10000

Energy (eV)

Inte

nsity

(A

U)

NiCu

Al

O

20 nm

Fig 15 STEM pictures depicting nanostructure retention of nickel par-ticles and the corresponding EDS spectrum confirming the presence ofnickel

the other hand during conventional sintering a significanttime is spent in the temperature range where the coarsen-ing effect is appreciable20 Besides in the nanostructuressurface melting is preferred in the fast translation thuskeeping the core intact From the TEM pictures of plasmasprayed NickelndashAlumina it is envisioned that plasma pro-cessing is considered to be a potential tool to fabricate freeform (eg cylindrical) bulk composites with nanostructureretention and improved mechanical properties

35 Mechanical Properties

Nickel plays a dominant role in increasing the strengthof the ceramic matrix The presence of cracks in ceram-ics makes them more brittle compared to metals At thetip of the crack as shown in Figure 16 there is an enor-mous amount of stress concentration developed as a resultof which the cracks tend to propagate easily The surfaceenergy required to propagate the crack is more than com-pensated by the release of the elastic strain energy Thepresence of Nickel spheroids dispersed uniformly through-out the matrix blunts the crack thereby making it more dif-ficult to propagate as shown in Figure 16 Nickel particleswhich are highly plastic in nature aids in crack bluntingand improvement of fracture toughness of the nanocom-posite Lieberthal et al16 proposed a model in which acrack propagating along the alumina grain boundaries aredeflected along around the Ni particle located in grainboundaries which aid in improving the fracture toughnessof the ceramic nanocomposite

The hardness and fracture toughness of the NindashAluminaplasma sprayed sample that could be achieved with thegiven set of plasma spray parameters were measured tobe 1025 HV and sim5 MPa middot m12 The hardness valuesachieved were found to be better than that achieved for the

Crack tip in a ceramicmatrix

Ni-particle

Fig 16 Nickel particles blunt the crack and arrest crack propagationimproves the fracture toughness

J Nanosci Nanotechnol 6 651ndash660 2006 659

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

monolithic alumina developed using plasma spray (940plusmn20 HV)40 The marked increase in the fracture toughnesscompared to other processing techniques can be attributedto the crack blunting capability of ductile Nickel It shouldbe noted that NindashAlumina nanocomposites made throughtechniques such as hot press sintering12 reported a frac-ture toughness of sim35 MPa middotm12 In other words the increase in fracture toughness was approximately 40

4 CONCLUSIONS

Electroless Nickel coated alumina particles have been suc-cessfully used to manufacture a free-form bulk nanocom-posite with improved hardness and fracture toughnessWettability study of Nickel on polycrystalline Aluminarevealed the contact angle which validated the self affineroughness model being used to calculate the apparent con-tact angle Laser processing had shortcomings from thepoint of view of time and the high laser required to conso-lidate the powders to bulk form especially in the casemetalceramic matrix For plasma processing substratematerial selection is important based on the material com-position to be sprayed Calculation of residual High frac-ture toughness can be attributed to the nanostructuredNickel which arrests the crack propagation Thus plasmaspray processing is an effective technique to make freeform bulk nanostructures

Acknowledgments The authors wish to thank theOffice of Naval Research Young Investigator Award (ONRYIP-N000140210591) for their financial support Theauthors would also like to thank Plasma Processes Incfor their help in product development We also appreci-ate the equipment support from Materials CharacterizationFacility at the University of Central Florida

References and Notes

1 W H Tuan and R J Brook J Eur Ceram Soc 6 31 (1990)2 A Paesano Jr C K Matsuda J B M Da Cunha M A Z

Vasconcellos B Hallouche and S L Silva J Magn Magn Mater264 264 (2003)

3 T Sekino and K Niihara J Mater Sci 32 3943 (1997)4 T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)5 A Nakahira and K Niihara Fract Mech 9 165 (1992)6 Y K Jeong and K Niihara Nanostruct Mater 9 193 (1997)7 S C Wang and W C J Wei Nanostruct Mater 10 983 (1998)

8 B Budiansky J C Amazigo and A J Evans J Mech Phys Solids36 167 (1988)

9 S Ozkar G A Ozin and R A Prokopowicz Chem Mater 4 1380(1992)

10 S T Oh T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)11 O Aharon S Bar-Ziv D Gorni T Cohen-Hyams and W D

Kaplan Scripta Mater 50 1209 (2004)12 T Sekino T Nakajima and K Niihara Mater Lett 29 165 (1996)13 S I Cha S H Hong and B K Kim Mater Sci Eng A 351 31

(2003)14 S Seal S C Kuiry P Georgieva and K Rea Scripta Mater 50

1237 (2004)15 R Z Chen and W H Tuan J Eur Ceram Soc 19 463 (1999)16 M Lieberthal and W D Kaplan Mater Sci Eng A 302 83 (2001)17 K Niihara B S Kim T Nakayama T Kusunose T Nomoto

A Hikasa and T Sekino J Eur Ceram Soc 24 3419 (2004)18 U L Adisorn T Matsunaga Y Kobayashi S M Choi and

H Awaji Ceram Intl 31 803 (2005)19 S Seal S C Kuiry P Georgieva and A Agarwal MRS Bull 29

16 (2004)20 J Michalski K Konopka and M Trzaska Mater Chem Phys 81

407 (2003)21 A G Evans and E A Charles J Am Ceram Soc 59 371 (1976)22 W H Tuan and R J Brook J Eur Ceram Soc 10 95 (1992)23 W H Tuan and W B Chou J Am Ceram Soc 80 2418 (1997)24 E Breval G Dodds and C G Pantano Mater Res Bull 20 1191

(1985)25 E Breval Z Deng S Chiou and C G Pantano J Mater Sci 27

1464 (1992)26 E D Rodeghiero O K Tse J Chisaki and E P Giannelis Mater

Sci Eng A 195 151 (1995)27 J F Moulder W F Stickle P E Sobol and K D Bomben

in Handbook of X-Ray Photoelectron Spectroscopy edited byJ Chastain and R C King Jr Physical Electronics Inc Minnesota(1995) p 71

28 T L Barr and S Seal J Vac Sci Technol A 13 1239 (1995)29 E Saiz R M Cannon and A P Tomsia Acta Mater 47 4209

(1999)30 W D Kingery J Am Ceram Soc 37 42 (1954)31 G Palasantzas and J Th M De Hosson Acta Mater 49 3533

(2001)32 N Eustathopolous Acta Mater 46 2319 (1998)33 J A Vreeling V Ocelrsquok Y T Pei D T L Van Agterveld and

J Th M De Hosson Acta Mater 48 4225 (2000)34 R Ye T Ishigaki J Jurewicz P Proulx and M I Boulos Plasma

Chem Plasma Process 24 555 (2004)35 A Brenner and S Senderoff J Res Natl Bur Stand 42 105 (1949)36 S Kuroda and T W Clyne Thin Solid Films 200 49 (1991)37 M W Hunt Guide to engineering materials Adv Mater Proc 158

6 (2000)38 L L Shaw D Goberman R Ren M Gell S Jiang Y Wang T D

Xiao and P R Strutt Surf Coat Technol 130 1 (2000)39 O D Slagle and H A Mckinsky J Appl Phys 38 437 (1967)40 M Vural S Zeytin and A H Ucisik Surf Coat Technol 97 347

(1997)

Received 10 July 2005 Accepted 12 October 2005

660 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

Folds

Isolated pores

Folds

(a)

(b)

(c)

Fig 13 Cross sectional low magnification micrographs indicating theextent of porosity formation at the (a) external (b) middle and (c) inter-nal portions of the bulk nanocomposite

344 Structural Characterization

Phase Transformation It is being observed that thealumina matrix undergone phase transformation fromalpha to gamma structure The alpha to gamma transi-tion depends on the ratio of the gun power to the primarygas flow in Standard Cubic Feet per Hour (SCFH) It ismentioned elsewhere [shaw] that if the parent phase is -alumina based on this ratio mentioned above the product

2θ20 30 40 50 60 70 80

(c)

(b)

(a)

NickelAlpha AluminaGamma AluminaNiO

Fig 14 Phase transformations associated with plasma spray of Ni-Al2O3 (a) Feedstock powder (b) plasma sprayed component (c) hydro-gen reduction

phase might be a combination of more -alumina and less alumina if the ratio is less than or equal to 240 or a com-bination of more alumina and less -alumina if the ratiois more than or equal to 310 The density and hardness val-ues achieved38 with the above ratios show significant dif-ferences with the former showing less hardness and densitythan the latter The phase transformation achieved in ourpresent case had more alumina The spray parameterswere manipulated to keep the ratio above 310 to makesure that dense and hard nanostructures were obtained Thediffractograms before and after plasma spraying is shownin Figure 14 X-ray diffraction spectrums from the plasmaspray deposition and powder mixture prior to the spray-ing are compared Both the diffractograms reveal the pres-ence of FCC Ni and Alumina The difference betweenthe diffractograms is that of the peak broadening Dur-ing the spray particles literally gets quenched from theirmelting temperatures towards the substrate preheat temper-ature Such large temperatures gradients are expected toimpart residual stresses on the sprayed part The stressescan be calculated based on the difference in ldquodrdquo spacingof the particles prior to spraying and after spraying Fromthe modulus of elasticity values the amount of residualstress can be calculated as follows39

Eminus1hkl =

C11 +C12

C11 minusC12C11 +2C12

+(

1C44

minus 2C11 minusC12

)l2m2 +m2n2 +n2l2

The compliance constants are C11 = 2465 GPa C12 =1473 GPa and C44 = 1247 GPa The calculated residualstress values are presented in Table IV

Since the process has been carried out in atmosphericair Ni has the propensity to get converted to its surfaceoxide Gibbs free energy of formation of Nickel oxide upto melting point of Nickel is negative Although Gibbs free

658 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table IV Residual stress measurements using X-ray diffraction techniques

hkl 2 (Measured) 2 (JCPDS) d (Measured) Aring d (JCPDS) Aring Strain () E (GPa) Stress (Mpa)

NiO 200 4326 4329 209 209 0066 136 90NiO 111 3725 3728 241 241 0078 304 236NiO 220 6289 6292 148 148 0043 232 99-Alumina 440 67 67032 140 139 0042 408 172

energy curve for formation of Nickel oxide is sloped upwith temperature the partial pressure of oxygen at NTP ismuch higher for the reaction to proceed forward and formNiO APS process carried out at dehumidified air will havean oxygen partial pressure of 02 atm which is more thansufficient for the NiO formation to occur Moreover if thestand off distance is more Nickel in its molten state picksup more oxygen from the atmosphere than it would haveif it were in solid state By reducing the stand off distancethe amount of formation of Nickel oxide can be reducedbut can not be totally eliminated XRD peaks confirmedthe formation of NiO An effort has been made to bringNiO back to its metallic state The plasma spray formednanocomposite was reduced in an atmosphere of Hydrogenat 1000 C for 1 hour

345 High Resolution TransmissionElectron Microscopy

Alumina grain boundaries being high energy regions thereis more probability of the presence of Nickel along thegrain boundaries In order to determine the location ofNickel EDS has been performed along the grain bound-aries to select spots for imaging of Nickel nano parti-cles In those spots where strong Nickel signal have beendetected STEM imaging has been done to determine themorphology and the size of the Nickel particles Spheri-cal Ni particles with 30 nm size have been imaged andare shown in Figure 15 Another evidence of nanostruc-ture retention of Nickel particle is the SAD pattern whichshows a spotty ring pattern as revealed in Figure 15(a)(inset) This pattern is a result of extremely fine grainsThe possible explanations of the nano structure retentionin NindashAlumina nanocomposite could be the lack of timepermitted for the coarsening of Nickel particles travers-ing through the high temperature zone (plasma flame) On

0100200300400500600700800900

1000

0 2000 4000 6000 8000 10000

Energy (eV)

Inte

nsity

(A

U)

NiCu

Al

O

20 nm

Fig 15 STEM pictures depicting nanostructure retention of nickel par-ticles and the corresponding EDS spectrum confirming the presence ofnickel

the other hand during conventional sintering a significanttime is spent in the temperature range where the coarsen-ing effect is appreciable20 Besides in the nanostructuressurface melting is preferred in the fast translation thuskeeping the core intact From the TEM pictures of plasmasprayed NickelndashAlumina it is envisioned that plasma pro-cessing is considered to be a potential tool to fabricate freeform (eg cylindrical) bulk composites with nanostructureretention and improved mechanical properties

35 Mechanical Properties

Nickel plays a dominant role in increasing the strengthof the ceramic matrix The presence of cracks in ceram-ics makes them more brittle compared to metals At thetip of the crack as shown in Figure 16 there is an enor-mous amount of stress concentration developed as a resultof which the cracks tend to propagate easily The surfaceenergy required to propagate the crack is more than com-pensated by the release of the elastic strain energy Thepresence of Nickel spheroids dispersed uniformly through-out the matrix blunts the crack thereby making it more dif-ficult to propagate as shown in Figure 16 Nickel particleswhich are highly plastic in nature aids in crack bluntingand improvement of fracture toughness of the nanocom-posite Lieberthal et al16 proposed a model in which acrack propagating along the alumina grain boundaries aredeflected along around the Ni particle located in grainboundaries which aid in improving the fracture toughnessof the ceramic nanocomposite

The hardness and fracture toughness of the NindashAluminaplasma sprayed sample that could be achieved with thegiven set of plasma spray parameters were measured tobe 1025 HV and sim5 MPa middot m12 The hardness valuesachieved were found to be better than that achieved for the

Crack tip in a ceramicmatrix

Ni-particle

Fig 16 Nickel particles blunt the crack and arrest crack propagationimproves the fracture toughness

J Nanosci Nanotechnol 6 651ndash660 2006 659

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

monolithic alumina developed using plasma spray (940plusmn20 HV)40 The marked increase in the fracture toughnesscompared to other processing techniques can be attributedto the crack blunting capability of ductile Nickel It shouldbe noted that NindashAlumina nanocomposites made throughtechniques such as hot press sintering12 reported a frac-ture toughness of sim35 MPa middotm12 In other words the increase in fracture toughness was approximately 40

4 CONCLUSIONS

Electroless Nickel coated alumina particles have been suc-cessfully used to manufacture a free-form bulk nanocom-posite with improved hardness and fracture toughnessWettability study of Nickel on polycrystalline Aluminarevealed the contact angle which validated the self affineroughness model being used to calculate the apparent con-tact angle Laser processing had shortcomings from thepoint of view of time and the high laser required to conso-lidate the powders to bulk form especially in the casemetalceramic matrix For plasma processing substratematerial selection is important based on the material com-position to be sprayed Calculation of residual High frac-ture toughness can be attributed to the nanostructuredNickel which arrests the crack propagation Thus plasmaspray processing is an effective technique to make freeform bulk nanostructures

Acknowledgments The authors wish to thank theOffice of Naval Research Young Investigator Award (ONRYIP-N000140210591) for their financial support Theauthors would also like to thank Plasma Processes Incfor their help in product development We also appreci-ate the equipment support from Materials CharacterizationFacility at the University of Central Florida

References and Notes

1 W H Tuan and R J Brook J Eur Ceram Soc 6 31 (1990)2 A Paesano Jr C K Matsuda J B M Da Cunha M A Z

Vasconcellos B Hallouche and S L Silva J Magn Magn Mater264 264 (2003)

3 T Sekino and K Niihara J Mater Sci 32 3943 (1997)4 T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)5 A Nakahira and K Niihara Fract Mech 9 165 (1992)6 Y K Jeong and K Niihara Nanostruct Mater 9 193 (1997)7 S C Wang and W C J Wei Nanostruct Mater 10 983 (1998)

8 B Budiansky J C Amazigo and A J Evans J Mech Phys Solids36 167 (1988)

9 S Ozkar G A Ozin and R A Prokopowicz Chem Mater 4 1380(1992)

10 S T Oh T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)11 O Aharon S Bar-Ziv D Gorni T Cohen-Hyams and W D

Kaplan Scripta Mater 50 1209 (2004)12 T Sekino T Nakajima and K Niihara Mater Lett 29 165 (1996)13 S I Cha S H Hong and B K Kim Mater Sci Eng A 351 31

(2003)14 S Seal S C Kuiry P Georgieva and K Rea Scripta Mater 50

1237 (2004)15 R Z Chen and W H Tuan J Eur Ceram Soc 19 463 (1999)16 M Lieberthal and W D Kaplan Mater Sci Eng A 302 83 (2001)17 K Niihara B S Kim T Nakayama T Kusunose T Nomoto

A Hikasa and T Sekino J Eur Ceram Soc 24 3419 (2004)18 U L Adisorn T Matsunaga Y Kobayashi S M Choi and

H Awaji Ceram Intl 31 803 (2005)19 S Seal S C Kuiry P Georgieva and A Agarwal MRS Bull 29

16 (2004)20 J Michalski K Konopka and M Trzaska Mater Chem Phys 81

407 (2003)21 A G Evans and E A Charles J Am Ceram Soc 59 371 (1976)22 W H Tuan and R J Brook J Eur Ceram Soc 10 95 (1992)23 W H Tuan and W B Chou J Am Ceram Soc 80 2418 (1997)24 E Breval G Dodds and C G Pantano Mater Res Bull 20 1191

(1985)25 E Breval Z Deng S Chiou and C G Pantano J Mater Sci 27

1464 (1992)26 E D Rodeghiero O K Tse J Chisaki and E P Giannelis Mater

Sci Eng A 195 151 (1995)27 J F Moulder W F Stickle P E Sobol and K D Bomben

in Handbook of X-Ray Photoelectron Spectroscopy edited byJ Chastain and R C King Jr Physical Electronics Inc Minnesota(1995) p 71

28 T L Barr and S Seal J Vac Sci Technol A 13 1239 (1995)29 E Saiz R M Cannon and A P Tomsia Acta Mater 47 4209

(1999)30 W D Kingery J Am Ceram Soc 37 42 (1954)31 G Palasantzas and J Th M De Hosson Acta Mater 49 3533

(2001)32 N Eustathopolous Acta Mater 46 2319 (1998)33 J A Vreeling V Ocelrsquok Y T Pei D T L Van Agterveld and

J Th M De Hosson Acta Mater 48 4225 (2000)34 R Ye T Ishigaki J Jurewicz P Proulx and M I Boulos Plasma

Chem Plasma Process 24 555 (2004)35 A Brenner and S Senderoff J Res Natl Bur Stand 42 105 (1949)36 S Kuroda and T W Clyne Thin Solid Films 200 49 (1991)37 M W Hunt Guide to engineering materials Adv Mater Proc 158

6 (2000)38 L L Shaw D Goberman R Ren M Gell S Jiang Y Wang T D

Xiao and P R Strutt Surf Coat Technol 130 1 (2000)39 O D Slagle and H A Mckinsky J Appl Phys 38 437 (1967)40 M Vural S Zeytin and A H Ucisik Surf Coat Technol 97 347

(1997)

Received 10 July 2005 Accepted 12 October 2005

660 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

Viswanathan et al High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite

Table IV Residual stress measurements using X-ray diffraction techniques

hkl 2 (Measured) 2 (JCPDS) d (Measured) Aring d (JCPDS) Aring Strain () E (GPa) Stress (Mpa)

NiO 200 4326 4329 209 209 0066 136 90NiO 111 3725 3728 241 241 0078 304 236NiO 220 6289 6292 148 148 0043 232 99-Alumina 440 67 67032 140 139 0042 408 172

energy curve for formation of Nickel oxide is sloped upwith temperature the partial pressure of oxygen at NTP ismuch higher for the reaction to proceed forward and formNiO APS process carried out at dehumidified air will havean oxygen partial pressure of 02 atm which is more thansufficient for the NiO formation to occur Moreover if thestand off distance is more Nickel in its molten state picksup more oxygen from the atmosphere than it would haveif it were in solid state By reducing the stand off distancethe amount of formation of Nickel oxide can be reducedbut can not be totally eliminated XRD peaks confirmedthe formation of NiO An effort has been made to bringNiO back to its metallic state The plasma spray formednanocomposite was reduced in an atmosphere of Hydrogenat 1000 C for 1 hour

345 High Resolution TransmissionElectron Microscopy

Alumina grain boundaries being high energy regions thereis more probability of the presence of Nickel along thegrain boundaries In order to determine the location ofNickel EDS has been performed along the grain bound-aries to select spots for imaging of Nickel nano parti-cles In those spots where strong Nickel signal have beendetected STEM imaging has been done to determine themorphology and the size of the Nickel particles Spheri-cal Ni particles with 30 nm size have been imaged andare shown in Figure 15 Another evidence of nanostruc-ture retention of Nickel particle is the SAD pattern whichshows a spotty ring pattern as revealed in Figure 15(a)(inset) This pattern is a result of extremely fine grainsThe possible explanations of the nano structure retentionin NindashAlumina nanocomposite could be the lack of timepermitted for the coarsening of Nickel particles travers-ing through the high temperature zone (plasma flame) On

0100200300400500600700800900

1000

0 2000 4000 6000 8000 10000

Energy (eV)

Inte

nsity

(A

U)

NiCu

Al

O

20 nm

Fig 15 STEM pictures depicting nanostructure retention of nickel par-ticles and the corresponding EDS spectrum confirming the presence ofnickel

the other hand during conventional sintering a significanttime is spent in the temperature range where the coarsen-ing effect is appreciable20 Besides in the nanostructuressurface melting is preferred in the fast translation thuskeeping the core intact From the TEM pictures of plasmasprayed NickelndashAlumina it is envisioned that plasma pro-cessing is considered to be a potential tool to fabricate freeform (eg cylindrical) bulk composites with nanostructureretention and improved mechanical properties

35 Mechanical Properties

Nickel plays a dominant role in increasing the strengthof the ceramic matrix The presence of cracks in ceram-ics makes them more brittle compared to metals At thetip of the crack as shown in Figure 16 there is an enor-mous amount of stress concentration developed as a resultof which the cracks tend to propagate easily The surfaceenergy required to propagate the crack is more than com-pensated by the release of the elastic strain energy Thepresence of Nickel spheroids dispersed uniformly through-out the matrix blunts the crack thereby making it more dif-ficult to propagate as shown in Figure 16 Nickel particleswhich are highly plastic in nature aids in crack bluntingand improvement of fracture toughness of the nanocom-posite Lieberthal et al16 proposed a model in which acrack propagating along the alumina grain boundaries aredeflected along around the Ni particle located in grainboundaries which aid in improving the fracture toughnessof the ceramic nanocomposite

The hardness and fracture toughness of the NindashAluminaplasma sprayed sample that could be achieved with thegiven set of plasma spray parameters were measured tobe 1025 HV and sim5 MPa middot m12 The hardness valuesachieved were found to be better than that achieved for the

Crack tip in a ceramicmatrix

Ni-particle

Fig 16 Nickel particles blunt the crack and arrest crack propagationimproves the fracture toughness

J Nanosci Nanotechnol 6 651ndash660 2006 659

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

monolithic alumina developed using plasma spray (940plusmn20 HV)40 The marked increase in the fracture toughnesscompared to other processing techniques can be attributedto the crack blunting capability of ductile Nickel It shouldbe noted that NindashAlumina nanocomposites made throughtechniques such as hot press sintering12 reported a frac-ture toughness of sim35 MPa middotm12 In other words the increase in fracture toughness was approximately 40

4 CONCLUSIONS

Electroless Nickel coated alumina particles have been suc-cessfully used to manufacture a free-form bulk nanocom-posite with improved hardness and fracture toughnessWettability study of Nickel on polycrystalline Aluminarevealed the contact angle which validated the self affineroughness model being used to calculate the apparent con-tact angle Laser processing had shortcomings from thepoint of view of time and the high laser required to conso-lidate the powders to bulk form especially in the casemetalceramic matrix For plasma processing substratematerial selection is important based on the material com-position to be sprayed Calculation of residual High frac-ture toughness can be attributed to the nanostructuredNickel which arrests the crack propagation Thus plasmaspray processing is an effective technique to make freeform bulk nanostructures

Acknowledgments The authors wish to thank theOffice of Naval Research Young Investigator Award (ONRYIP-N000140210591) for their financial support Theauthors would also like to thank Plasma Processes Incfor their help in product development We also appreci-ate the equipment support from Materials CharacterizationFacility at the University of Central Florida

References and Notes

1 W H Tuan and R J Brook J Eur Ceram Soc 6 31 (1990)2 A Paesano Jr C K Matsuda J B M Da Cunha M A Z

Vasconcellos B Hallouche and S L Silva J Magn Magn Mater264 264 (2003)

3 T Sekino and K Niihara J Mater Sci 32 3943 (1997)4 T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)5 A Nakahira and K Niihara Fract Mech 9 165 (1992)6 Y K Jeong and K Niihara Nanostruct Mater 9 193 (1997)7 S C Wang and W C J Wei Nanostruct Mater 10 983 (1998)

8 B Budiansky J C Amazigo and A J Evans J Mech Phys Solids36 167 (1988)

9 S Ozkar G A Ozin and R A Prokopowicz Chem Mater 4 1380(1992)

10 S T Oh T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)11 O Aharon S Bar-Ziv D Gorni T Cohen-Hyams and W D

Kaplan Scripta Mater 50 1209 (2004)12 T Sekino T Nakajima and K Niihara Mater Lett 29 165 (1996)13 S I Cha S H Hong and B K Kim Mater Sci Eng A 351 31

(2003)14 S Seal S C Kuiry P Georgieva and K Rea Scripta Mater 50

1237 (2004)15 R Z Chen and W H Tuan J Eur Ceram Soc 19 463 (1999)16 M Lieberthal and W D Kaplan Mater Sci Eng A 302 83 (2001)17 K Niihara B S Kim T Nakayama T Kusunose T Nomoto

A Hikasa and T Sekino J Eur Ceram Soc 24 3419 (2004)18 U L Adisorn T Matsunaga Y Kobayashi S M Choi and

H Awaji Ceram Intl 31 803 (2005)19 S Seal S C Kuiry P Georgieva and A Agarwal MRS Bull 29

16 (2004)20 J Michalski K Konopka and M Trzaska Mater Chem Phys 81

407 (2003)21 A G Evans and E A Charles J Am Ceram Soc 59 371 (1976)22 W H Tuan and R J Brook J Eur Ceram Soc 10 95 (1992)23 W H Tuan and W B Chou J Am Ceram Soc 80 2418 (1997)24 E Breval G Dodds and C G Pantano Mater Res Bull 20 1191

(1985)25 E Breval Z Deng S Chiou and C G Pantano J Mater Sci 27

1464 (1992)26 E D Rodeghiero O K Tse J Chisaki and E P Giannelis Mater

Sci Eng A 195 151 (1995)27 J F Moulder W F Stickle P E Sobol and K D Bomben

in Handbook of X-Ray Photoelectron Spectroscopy edited byJ Chastain and R C King Jr Physical Electronics Inc Minnesota(1995) p 71

28 T L Barr and S Seal J Vac Sci Technol A 13 1239 (1995)29 E Saiz R M Cannon and A P Tomsia Acta Mater 47 4209

(1999)30 W D Kingery J Am Ceram Soc 37 42 (1954)31 G Palasantzas and J Th M De Hosson Acta Mater 49 3533

(2001)32 N Eustathopolous Acta Mater 46 2319 (1998)33 J A Vreeling V Ocelrsquok Y T Pei D T L Van Agterveld and

J Th M De Hosson Acta Mater 48 4225 (2000)34 R Ye T Ishigaki J Jurewicz P Proulx and M I Boulos Plasma

Chem Plasma Process 24 555 (2004)35 A Brenner and S Senderoff J Res Natl Bur Stand 42 105 (1949)36 S Kuroda and T W Clyne Thin Solid Films 200 49 (1991)37 M W Hunt Guide to engineering materials Adv Mater Proc 158

6 (2000)38 L L Shaw D Goberman R Ren M Gell S Jiang Y Wang T D

Xiao and P R Strutt Surf Coat Technol 130 1 (2000)39 O D Slagle and H A Mckinsky J Appl Phys 38 437 (1967)40 M Vural S Zeytin and A H Ucisik Surf Coat Technol 97 347

(1997)

Received 10 July 2005 Accepted 12 October 2005

660 J Nanosci Nanotechnol 6 651ndash660 2006

unknownIP 13217075212

Tue 21 Feb 2006 212627

Delivered by Ingenta to

RESEARCHARTICLE

High Energy Density Processing of a Free Form NickelndashAlumina Nanocomposite Viswanathan et al

monolithic alumina developed using plasma spray (940plusmn20 HV)40 The marked increase in the fracture toughnesscompared to other processing techniques can be attributedto the crack blunting capability of ductile Nickel It shouldbe noted that NindashAlumina nanocomposites made throughtechniques such as hot press sintering12 reported a frac-ture toughness of sim35 MPa middotm12 In other words the increase in fracture toughness was approximately 40

4 CONCLUSIONS

Electroless Nickel coated alumina particles have been suc-cessfully used to manufacture a free-form bulk nanocom-posite with improved hardness and fracture toughnessWettability study of Nickel on polycrystalline Aluminarevealed the contact angle which validated the self affineroughness model being used to calculate the apparent con-tact angle Laser processing had shortcomings from thepoint of view of time and the high laser required to conso-lidate the powders to bulk form especially in the casemetalceramic matrix For plasma processing substratematerial selection is important based on the material com-position to be sprayed Calculation of residual High frac-ture toughness can be attributed to the nanostructuredNickel which arrests the crack propagation Thus plasmaspray processing is an effective technique to make freeform bulk nanostructures

Acknowledgments The authors wish to thank theOffice of Naval Research Young Investigator Award (ONRYIP-N000140210591) for their financial support Theauthors would also like to thank Plasma Processes Incfor their help in product development We also appreci-ate the equipment support from Materials CharacterizationFacility at the University of Central Florida

References and Notes

1 W H Tuan and R J Brook J Eur Ceram Soc 6 31 (1990)2 A Paesano Jr C K Matsuda J B M Da Cunha M A Z

Vasconcellos B Hallouche and S L Silva J Magn Magn Mater264 264 (2003)

3 T Sekino and K Niihara J Mater Sci 32 3943 (1997)4 T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)5 A Nakahira and K Niihara Fract Mech 9 165 (1992)6 Y K Jeong and K Niihara Nanostruct Mater 9 193 (1997)7 S C Wang and W C J Wei Nanostruct Mater 10 983 (1998)

8 B Budiansky J C Amazigo and A J Evans J Mech Phys Solids36 167 (1988)

9 S Ozkar G A Ozin and R A Prokopowicz Chem Mater 4 1380(1992)

10 S T Oh T Sekino and K Niihara J Eur Ceram Soc 18 31 (1998)11 O Aharon S Bar-Ziv D Gorni T Cohen-Hyams and W D

Kaplan Scripta Mater 50 1209 (2004)12 T Sekino T Nakajima and K Niihara Mater Lett 29 165 (1996)13 S I Cha S H Hong and B K Kim Mater Sci Eng A 351 31

(2003)14 S Seal S C Kuiry P Georgieva and K Rea Scripta Mater 50

1237 (2004)15 R Z Chen and W H Tuan J Eur Ceram Soc 19 463 (1999)16 M Lieberthal and W D Kaplan Mater Sci Eng A 302 83 (2001)17 K Niihara B S Kim T Nakayama T Kusunose T Nomoto

A Hikasa and T Sekino J Eur Ceram Soc 24 3419 (2004)18 U L Adisorn T Matsunaga Y Kobayashi S M Choi and

H Awaji Ceram Intl 31 803 (2005)19 S Seal S C Kuiry P Georgieva and A Agarwal MRS Bull 29

16 (2004)20 J Michalski K Konopka and M Trzaska Mater Chem Phys 81

407 (2003)21 A G Evans and E A Charles J Am Ceram Soc 59 371 (1976)22 W H Tuan and R J Brook J Eur Ceram Soc 10 95 (1992)23 W H Tuan and W B Chou J Am Ceram Soc 80 2418 (1997)24 E Breval G Dodds and C G Pantano Mater Res Bull 20 1191

(1985)25 E Breval Z Deng S Chiou and C G Pantano J Mater Sci 27

1464 (1992)26 E D Rodeghiero O K Tse J Chisaki and E P Giannelis Mater

Sci Eng A 195 151 (1995)27 J F Moulder W F Stickle P E Sobol and K D Bomben

in Handbook of X-Ray Photoelectron Spectroscopy edited byJ Chastain and R C King Jr Physical Electronics Inc Minnesota(1995) p 71

28 T L Barr and S Seal J Vac Sci Technol A 13 1239 (1995)29 E Saiz R M Cannon and A P Tomsia Acta Mater 47 4209

(1999)30 W D Kingery J Am Ceram Soc 37 42 (1954)31 G Palasantzas and J Th M De Hosson Acta Mater 49 3533

(2001)32 N Eustathopolous Acta Mater 46 2319 (1998)33 J A Vreeling V Ocelrsquok Y T Pei D T L Van Agterveld and

J Th M De Hosson Acta Mater 48 4225 (2000)34 R Ye T Ishigaki J Jurewicz P Proulx and M I Boulos Plasma

Chem Plasma Process 24 555 (2004)35 A Brenner and S Senderoff J Res Natl Bur Stand 42 105 (1949)36 S Kuroda and T W Clyne Thin Solid Films 200 49 (1991)37 M W Hunt Guide to engineering materials Adv Mater Proc 158

6 (2000)38 L L Shaw D Goberman R Ren M Gell S Jiang Y Wang T D

Xiao and P R Strutt Surf Coat Technol 130 1 (2000)39 O D Slagle and H A Mckinsky J Appl Phys 38 437 (1967)40 M Vural S Zeytin and A H Ucisik Surf Coat Technol 97 347

(1997)

Received 10 July 2005 Accepted 12 October 2005

660 J Nanosci Nanotechnol 6 651ndash660 2006