author's personal copyweb.eng.fiu.edu/agarwala/pdf/2011/11.pdf · author's personal copy...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 4 ( 2 0 1 1 ) 4 4 – 5 6

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/jmbbm

Research paper

Boron nitride nanotube reinforced hydroxyapatite composite:Mechanical and tribological performance and in-vitrobiocompatibility to osteoblasts

Debrupa Lahiria, Virendra Singhb, Ana Paula Benaducec, Sudipta Sealb, Lidia Kosc,Arvind Agarwala,∗

aNanomechanics and Nanotribology Lab and High Temperature Tribology Lab, Department of Mechanical and Materials Engineering,Florida International University, Miami, FL 33174, USAbAMPAC and Nanoscience Technology Center, 4000 Central FL Blvd, AMPAC, Eng 1 Room 381, University of Central Florida,Orlando 32816, USAcDepartment of Biological Sciences, Florida International University, Miami, FL 33174, USA

A R T I C L E I N F O

Article history:

Received 28 May 2010

Received in revised form

14 September 2010

Accepted 17 September 2010

Published online 21 September 2010

A B S T R A C T

This study proposes boron nitride nanotube (BNNT) reinforced hydroxyapatite (HA) as a

novel composite material for orthopedic implant applications. The spark plasma sintered

(SPS) composite structure shows higher density compared to HA. Minimal lattice mismatch

between HA and BNNT leads to coherent bonding and strong interface. HA-4 wt% BNNT

composite offers excellent mechanical properties—120% increment in elastic modulus,

129% higher hardness and 86%more fracture toughness, as compared to HA. Improvements

in the hardness and fracture toughness are related to grain refinement and crack bridging

by BNNTs. HA–BNNT composite also shows 75% improvement in the wear resistance. The

wear morphology suggests localized plastic deformation supported by the sliding of outer

walls of BNNT. Osteoblast proliferation and cell viability show no adverse effect of BNNT

addition. HA–BNNT composite is, thus, envisioned as a potential material for stronger

orthopedic implants.c⃝ 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Hydroxyapatite (HA) possesses chemical composition (Ca10(PO4)6(OH)2), crystal structure and Ca:P ratio (1.67) similarto apatite found in human skeleton (Gu et al., 2002; Whiteet al., 2007; Yu et al., 2003). These features have made HAclinically accepted orthopedic implant material. In spite of itsbioactivity, the poor fracture toughness and wear resistanceof HA limits its application in load bearing orthopedic

∗ Corresponding author. Tel.: +1 305 348 1701; fax: +1 305 348 1932.E-mail address: [email protected] (A. Agarwal).

implants. Mechanical and tribological performance of HAcould be improved by following two different approaches—grain size refinement and second phase reinforcement.

Researchers have successfully used nanocrystalline HAfor the improvement of mechanical property without havingnegative effect on its biocompatibility (Li et al., 2007; Guoet al., 2007; Que et al., 2008; Grossin et al., 2009; Wangand Shaw, 2009; Wang et al., 2009). A recent study byWang and Shaw (2009) has shown that grain size refinement

1751-6161/$ - see front matter c⃝ 2010 Elsevier Ltd. All rights reserved.doi:10.1016/j.jmbbm.2010.09.005


J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 4 ( 2 0 1 1 ) 4 4 – 5 6 45

simultaneously increase hardness (by 22%) and fracturetoughness (by 74%) of sintered HA. The increase in thefracture toughness is due to the deflection of propagatingcrack and the change of cracking mode from transgranularto intergranular (Wang and Shaw, 2009). Control on thegrain size in sintered ceramic structures could be achievedby spark plasma sintering (SPS). Due the lower sinteringtemperature and shorter time of soaking, SPS restricts thegrain growth and retains the nanocrystalline structure withimproved mechanical properties (Munir et al., 2006; Olevskyet al., 2007; Lu, 2008; Chaim et al., 2008; Ragulya, 2008). Anadded advantage of HA processing through SPS is the reducedchance of dissociation to tricalcium phosphate (TCP) due toshorter exposure at high temperature (Gu et al., 2002; Yu et al.,2003; Gu et al., 2004; Li et al., 2007; Xu et al., 2007; Li et al., 2008;Lahiri et al., 2010a).

Second phase reinforcements like, Al2O3,ZrO2,TiO2 etc.have also been used to increase the mechanical performanceof HA based ceramics (Li et al., 1995; Gautier et al., 1997; Fuet al., 2002; Wang et al., 2005; Que et al., 2008). In recentyears, carbon nanotube has appeared as a strong contenderas reinforcement to HA due to its excellent elastic modulus(200–1000 GPa (Singh et al., 2003)) and fiber-like morphologythat helps in crack bridging and improving fracture toughness(Balani et al., 2007a; Sarkar et al., 2007; Lahiri et al., 2010a).CNT also acts as grain refiner for sintered HA (Lahiri et al.,2010a), further aiding the toughening. A recent study bythe present authors have shown 25% increase in the elasticmodulus and 92% increase in the fracture toughness of sparkplasma sintered HA–4 wt% CNT composite as compared to HA(Lahiri et al., 2010a). Addition of CNT in HA is also reportedto increase the wear resistance of the composite by 66% ascompared to HA (Balani et al., 2007b; Chen et al., 2007; Lahiriet al., 2010a). But, the ongoing debate on the biocompatibilityof CNTs (Singh et al., 2006; Fioritto, 2008; Usui et al., 2008;Cheng et al., 2009) demands the search for an alternativereinforcement to HA.

Boron nitride nanotube (BNNT) is a structural analogue ofCNT—formed with tubular shaped hexagonal boron nitride(hBN) sheet. BNNTs possess elastic modulus (750–1200 GPa(Chopra and Zetll, 1998; Suryavanshi et al., 2004)) andtensile strength (>24 GPa (Shen, 2009)) similar to CNT, whichmakes it a potential reinforcement for HA. The flexibleand elastic nature of BNNT and its ability to withstandheavy deformation (Golberg et al., 2007a) could be helpful inpreventing damage to itself during high pressure applicationin SPS. High temperature oxidation resistance of BNNT isbetter than CNT (Golberg et al., 2007b), which makes it moresuitable for high temperature processing.

Another important consideration for using BNNT fororthopedic applications is its biocompatibility. Our recentstudy has shown that BNNTs are non-cytotoxic to osteoblastsand macrophages, the two most important cell lineagesrelated to orthopedic applications (Lahiri et al., 2010b). BNNTsare also found non-cytotoxic to human embryonic kidneycells (HEK-293) (Chen et al., 2009) and human neuroblastomacell line (SH-SY5Y) (Ciofani et al., 2009). In fact, no reportis available regarding adverse effect of BNNT on the livingcells.

Even though BNNTs were first synthesized in 1995 (Chopraet al., 1995), few reports are available on the use of BNNTas reinforcement to fabricate composites (Zhi et al., 2005a,2006a; Bansal et al., 2006; Huang et al., 2007; Choi et al.,2007; Ravichandran et al., 2008; Lahiri et al., 2010b). Onlyone study is available on ceramic (Al2O3 and Si3N4) basedBNNT composites (Huang et al., 2007). BNNT introduces hightemperature superplasticity in the ceramics by controlleddynamic grain growth and energy absorption mechanism(Huang et al., 2007). The hardness of the ceramic-BNNTcomposite is reported to increase, but not the elastic modulus(Huang et al., 2007). BNNT reinforcement in polymer improvesthermal conductivity, mechanical and optical properties (Zhiet al., 2005a, 2006a; Ravichandran et al., 2008; Zhi et al.,2009; Terao et al., 2010). Our recent study has shownthat BNNT reinforcement could effectively increase theelastic modulus, tensile strength and biocompatibility ofbiodegradable polymer used in orthopedic scaffold (Lahiriet al., 2010b). The potential of BNNT as reinforcement to HAhas never been studied.

In light of the present scenario, the aim of this studyis to explore BNNT as reinforcement to HA for orthopedicapplications. HA–BNNT composite is synthesized by sparkplasma sintering. The microstructure of BNNT and HAafter SPS is investigated. Interfacial orientation relationshipbetween BNNT and HA is analyzed. Effect of BNNTaddition on the elastic modulus and fracture toughnessof the composite is investigated. Tribological performanceof HA–BNNT composite and the wear mechanism in thepresence of BNNT is also studied. The biocompatibility of theHA–BNNT composite is assessed in terms of in-vitro osteoblastproliferation and viability study.

2. Materials and methods

2.1. Powder preparation and spark plasma sintering

Boron nitride nanotubes, obtained from Nanoamor, Hous-ton, USA, comprises nodular (bamboo type), cylindrical nan-otubes, and few nano-rods which will be referred as BNNThereafter. Fig. 1 presents the TEM image of as-received BN-NTs showing nodular and cylindrical nanotubes. The size dis-tribution (refer to supporting document—Fig. S1), measuredfrom TEM and SEM images of dispersed BNNTs, shows length:0.43–5.8 µm (mean—1.98 µm); outer diameter: 10–145 nm(mean—71 nm) and aspect ratio: 15–84. HA nano-rods (length:100–325 nm, diameter: 25-50 nm) were procured from Infra-mat Corporation, (Willington, CT, USA). The powder stock forsintering was prepared with two compositions, viz. 100% HAand HA-4 wt% BNNT. The composite powder was prepared inbatches of 0.5 g using ultrasonication for uniform dispersionof BNNTs in HA, without forming agglomeration of BNNTsor HA (refer to supporting document—Fig. S2). In each batch,0.02 g of BNNT was mixed in 20 ml of acetone and ultrasoni-cated for 3 h. Subsequently, 0.48 g of HA was mixed in the dis-persion and ultrasonicated for 1 h. Finally the dispersion wasdried in an oven at 348 K for 3 h. Spark plasma sintering of HAand HA–BNNT composite powders was carried out in vacuum


46 J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 4 ( 2 0 1 1 ) 4 4 – 5 6

Fig. 1 – Transmission electron micrograph of as-receivedBNNTs reveals bamboo type and tubular nanotubes ofvarying diameter.

at 1373 K and 70 MPa pressure using the spark plasma sinter-ing facility at Thermal Technology LLC, Sana Rosa, California,USA. A fast heating rate of 360 K/min was employed with asoaking time of 5 min at maximum temperature. The HA andHA–BNNT pellets were produced with 20 mm diameter and∼5 mm thickness.

2.2. Structural characterization

The surface of sintered pellets was polished to removethe graphitic contamination from the dies. The sampleswere sectioned, mounted and metallographically polished forthe microstructural characterization. JEOL JSM-633OF fieldemission scanning electron microscope was used for thecharacterization of powders and SPS pellets. Transmissionelectron microscopy images of as-received BNNTs werecaptured using Philips PW 6061 TEM system (model CM200, Eindhoven, Netherlands). Philips/FEI Tecnai F30 highresolution transmission electron microscope (HRTEM) wasused to study HA–BNNT interface in the sintered composite.Forward and inverse Fourier transform (FFT & Inverse-FFT) analysis has been used for an accurate calculationof the lattice spacing. Density has been measured usingArchimedes principle and water as the immersing medium.X-ray diffraction (XRD) study was carried out using Cu Kα (λ =

1.542 Å) radiation in a Siemens D-5000 X-ray diffractometer.Micro-Raman spectra were obtained using a Spectra Physics(Model 3900S, California, USA) with Ti–sapphire crystal astarget and the detector from Kaiser Optical Systems, Inc.(Michigan, USA).

2.3. Evaluation of mechanical properties

Hysitron Triboindenter (Hysitron Inc., Minneapolis, MN, USA),with 100 nm Berkovich pyramidal tip, was used in quasi-static indentation mode to measure the elastic modulus

and hardness on polished cross section of the sinteredpellets. Tip-area calibration was done using a standard fusedquartz sample of known modulus (69.6 GPa). Indentationwas performed with a constant loading/unloading rate for10 s and 3 s hold at the peak load of 2500 µN. Elasticmodulus (E) had been calculated from the load–displacementcurves using the Oliver–Pharr method (Oliver and Pharr, 1992).Microhardness was measured using a microhardness tester(Shanghai Taiming Optical Instrument Co. Ltd., model HXD-1000 TMC, Shanghai) with Vickers probe applying 1 kg of loadfor 15 s of dwell time. For an accurate measurement of radialcrack length, the indents were observed under SEM. Fracturetoughness has been calculated using the model for brittleceramic proposed by Anstis et al. (Anstis et al., 1981).

2.4. Evaluation of tribological behavior

Ball-on-disc tribometer (Nanovea, CA) was used to evaluatethe wear resistance and coefficient of friction (CoF) of sinteredpellets. Surface of sintered HA and HA–BNNT pellets waspolished to a roughness (Ra) of 0.5 µm prior to wear. Analuminum oxide ball of 3 mm diameter was used as thecounter surface. The wear studies were performed at 50 RPMspeed with a circular track of 2 mm radius, for a total traveldistance of 100 m. Three wear tracks were made in eachsample to check the reproducibility. The lateral force betweenthe aluminum oxide ball and the sintered pellet surface hasbeen measured by the linear variable differential transformer(LVDT) sensor. The coefficient of friction data was acquiredat a frequency of 16.67 Hz. For wear loss measurement, thereduction in weight of the sample was measured every 25 mof the sliding in each wear track, using a precision balancewith an accuracy of 10 µg. The weight loss was converted tovolume loss using the density of the samples.

2.5. Viability study with osteoblasts

For culturing, the human osteoblasts ATCC CRL-11372 (ATCC,Manassas, VA) were seeded at a density of 1000 cellsper well in 6-well polystyrene petri dishes (Corning, NewYork) at 310 K (37 ◦C), 5% CO2 in a 1:1 mixture ofHam’s F12 Medium Dulbecco’s Modified Eagle’s Medium,with 2.5 mM L-glutamine. The phenol red-free base mediawas supplemented with 10% Fetal Bovine Serum (AtlantaBiologicals, Lawrenceville, GA), 100 UI/ml of penicillin and100 µg/ml of streptomycin (MP Biomedicals, Irvine, CA).

HA and HA–BNNT sintered pellet surfaces (5 mm × 5 mmsurface area) were washed with 95% ethanol, washed 3 timeswith fresh medium and left for 3 h in a hood under UV light.They were then placed into 6-well polystyrene petri dishes(Corning, New York). For cell viability studies, osteoblastswere seeded at a density of 5000 cells per well in 2.5 ml ofmedium and grown in an incubator at 310 K (37 ◦C), 5% CO2.After 1, 3 and 5 days, cells grown on the pellets were stainedfor 2 min with a Phosphate Buffer Saline 1X solution contain-ing 15 µg/ml of Fluorescein Di-Acetate (FDA) (MP Biomedicals,Irvine, CA) and 4.5 µg/ml of Propidium Iodide (PI) (Fisher Sci-entific, Waltham, MA) before visualization on a Leica LeitzDM RB fluorescent microscope (Leica, Bannockburn, IL). Digi-tal pictures were captured with a Leica DM 500 camera. FDA



stains the live cells in green and PI stains the dead cells in red.Live (green) versus dead (red) cells counting was performedmanually. Viability is defined as percentage of live cells. ‘Stu-dent t’ test was performed to find out the 95% confidence in-terval for the viability data.

3. Results and discussion

3.1. Structural evolution

SEM image from the fracture surface of HA (Fig. 2(a))reveals partially sintered grain structure with porosity. On thecontrary, HA–BNNT shows a dense and fully sintered fracturesurface with negligible porosity (Fig. 2(b)). The density ofHA–BNNT (97.0% ± 0.3% theoretical density) was higher thanHA (92.2% ± 1.1% theoretical density). The densification inspark plasma sintering is achieved by extremely high heatingrate that activates diffusion mechanisms, like grain boundaryand lattice diffusion and power-law dislocation creep (Muniret al., 2006; Olevsky et al., 2007; Ragulya, 2008). Thus, theconduction of heat throughout the pellet during sinteringis important for achieving better densification. The heatingsource is the graphite die in contact with periphery ofthe green pellet. A thermal gradient is generated throughthe thickness of the green pellet due to the lower thermalconductivity of HA (1.25 W/mK (Gaona et al., 2007)), whichleads to incomplete sintering and porous structure. But, thethermal conductivity of BNNT is much higher than HA, witha reported value of 200–300W/mK in axial and 20–30W/mK intransverse direction (Zhi et al., 2009; Terao et al., 2010). Henceexpected increase in thermal conductivity of HA with BNNTreinforcement results in accelerated diffusion and betterdensification of HA–BNNT. BNNTs have also displayed 160%increase in the thermal conductivity of PVA-5 vol.% BNNTcomposite as compared to PVA (T. Terao 2010). BNNTs playan active role in grain refinement of the sintered structure.HA–BNNT composite has a grain size of 0.17 ± 0.1 µm, whichis nearly three times finer than HA having a grain size of0.61 ± 0.16 µm. The presence of BNNTs has also lead to grainrefinement in Al2O3 and Si3N4 based composites (Huanget al., 2007). High surface to volume ratio makes BNNT moreactive for grain boundary pinning, as found in the case of CNT,the structural analogue of BNNT (Lahiri et al., 2010a).

The application of high temperature and pressure duringspark plasma sintering necessitates investigation on thesurvival of BNNT structure in the sintered composite.Evidence of the existence of BNNTs in the sintered pelletis provided by the HRTEM images presented in Fig. 3(a) and(b). The defect-free lattice images of BNNT in the sinteredpellet rule out damage due to the application of high pressureand temperature during SPS, as observed in case of CNTs(Yang et al., 2009; Lahiri et al., 2010a). High flexibility ofBNNT along with the ability to withstand high deformationwithout getting damaged (Golberg et al., 2007a) is attributedfor their defect-free structure in SPS pellet. Fig. 3(c) presentsthe Raman spectrum for HA and HA–BNNT at powder andsintered stages. The peak at ∼964 cm−1, present in all spectra,is from the ν1 symmetric stretching vibration of phosphateanions in HA (Liu et al., 2008). The peak at ∼1367 cm−1,

Fig. 2 – Fracture surface of (a) HA showing the porosity and(b) HA–BNNT shows a dense structure.

present in HA–BNNT powder and sintered pellet, is an E2gmode peak from h-BN (Zhi et al., 2006b; Guo and Singh, 2008;Singhal et al., 2008). The presence of h-BN peak in Ramanspectrum of the pellet along with the tubular structuresin HRTEM images (Fig. 3(a) and (b)) further establishes thesurvival of BNNT through spark plasma sintering process.

X-ray diffraction (XRD) patterns of the sintered HA andHA–BNNT pellets are shown in Fig. 4. The major peaks in bothdiffraction patterns are from hydroxyapatite (JCPDS PDF No.9-432). Low intensity boron nitride (BN) peaks in HA–BNNTare generated from hexagonal (JCPDS PDF No. 34-0421) andtetragonal (JCPDS PDF No.-25-0098) crystal structures. XRDpatterns reveal the presence of some tetragonal-BN impurityin the as-received BNNTs. Some of the BN peaks overlap withHA peaks in the vicinity (Fig. 4). None of the highest peaksof β-TCP (JCPDS PDF No. 9-169) and α-TCP (JCPDS PDF No.29-359) is present. These observations prove that HA doesnot significantly dissociate into TCP during SPS processingfor both compositions. A previous study by our researchgroup on HA–CNT composite using the same SPS processingparameters has reported a similar observation (Lahiri et al.,2010a). Faster heating rate in SPS has prevented the longexposure of HA at high temperature and thus the dissociationinto TCP.



Fig. 3 – (a–b) HRTEM images of BNNT in the sintered HA–BNNT pellet retaining their defect-free structure; (c) Raman spectraof HA and HA–BNNT powders and sintered pellet.

Fig. 4 – X-ray diffraction patterns of HA and HA–BNNT sintered pellet showing HA and BN peaks.

3.2. BNNT–HA Interface

Improvement in themechanical and tribological performanceof a composite is dependent on the strength of bonding atreinforcement–matrix interface. HA and BNNT are chemicallynon-reacting species. The HRTEM image of HA–BNNTinterface reveals absence of any reaction product (Fig. 5(a)),ruling out presence of ionic or covalent bond. Thus, Vander Waal’s bond is the most probable at HA–BNNT interface.Hence the interfacial strength is mainly governed by the workof adhesion, which is dependent on the lattice arrangementat the interface. Work of adhesion is higher when thelattice strain due to mismatch is minimal. A higher lattice

mismatch, δ > 0.25 leads to incoherent interface and poorbonding (Porter and Easterling, 2001). Fourier transform (FFTand inverse-FFT) analysis of the lattice images from HRTEMmicrograph (Fig. 5(a)) reveals the crystallographic orientationat HA–BNNT interface. BNNT shows h-BN walls with inter-wall spacing of 0.33 nm, specific to boron nitride nanotubestructure (Terrones et al., 2007). HA crystals are recognizedfrom the lattice images of (211) planes with a lattice spacingof 0.282 nm. The (211) planes of HA are arranged at an angularrange of 65◦–68◦ to the outer wall of BNNT. The basal planein hexagonal HA structure also creates 65◦ angle with (211)planes. Thus, probability of alignment of HA crystals on BNNTsurface with basal planes being parallel to the outer h-BN



Fig. 5 – (a) HRTEM image of BNNT and HA interface. FFT analysis reveals the BNNT wall spacing and HA lattice (211)spacing at the interface; (b) Schematic of atomic arrangement at the interface with basal plane of HA and coinciding h-BNsheet on BNNT outer wall; (c) Schematic of BNNT open end showing alignment of h-BN walls with (211) planes of HA.

wall is very strong. The symmetric hexagonal structure for

both basal planes of HA and h-BN makes the alignment more

evitable. Fig. 5(b) represents a schematic of the basal plane

of HA superimposed on the h-BN wall. The basal plane of

HA has Ca atoms sitting at each corner of the hexagon with

a distance of 0.94 nm (Kay and Young, 1964). The distance

between two neighboring atoms in h-BN is 1.44 Å (Nag et al.,

2010). As shown in the schematic in Fig. 5(b), the distance

between two B atoms on h-BN, coinciding with the Ca atoms

of superimposed HA basal plane, is 1.04 nm. So, themismatch

between two superimposed pair of Ca (in HA) and B (in h-BN)

atoms is δ ∼ 0.11, which is much lower than the incoherence

limit of 0.25. Thus, the preferential alignment of HA crystals

on BNNT surface suggests a strong coherent interfacial bond

with minimal lattice strain. Similar observation is reported

at the interface of HA and CNT of SPS processed composite

structure (D. Lahiri 2010). HA can also form another interface

with BNNT at its open end, as shown in Fig. 5(c). BNNT has

an inter-wall spacing of 0.33 nm. The lattice spacing of (211)

set of HA plane is 0.282 nm. The lattice mismatch (δ) between

(211) plane of HA and BNNT walls is 0.17, which is also lower

than 0.25. Hence, open ends of BNNT form a semi-coherent

interface with HA crystals having (211) planes parallel to h-

BN walls of BNNT.

The strength of the HA–BNNT interface can be estimated

based on the model proposed by Chen et al. for Al2O3-CNT

system (Chen et al., 2008). BNNT is a structural analogue of

CNT with similar elastic modulus and tensile strength. The

effective area of load carrying outer layers (Aeff) of amultiwall

BNNT is calculated using the following expression:

Aeff = π

N−m=1

[RCNT − (m − 1)h − (m − 1)h′

]2

−[RCNT − mh − (m − 1)h′]2

(1)

where, RBNNT is the outer radius of BNNT (5–70 nm in thisstudy), h is the effective layer thickness (∼0.25 nm (Lan et al.,2009)), d is the spacing between each h-BN layer (∼0.33 nm(Terrones et al., 2007)), h′

= d − h, and N is the numberof outer layers carrying load. BNNTs have ∼10–50 walls asobserved from HRTEM images. For a conservative estimate ofthe strength, it has been assumed that load is borne by 5 outerwalls of BNNT. Cox model is used to compute the interfacialshear strength (τ) between BNNT and HA. According to theCox model, the fiber at the center of a coaxial cylinder of thematrix (of radius R) is used to calculate τ using the followingexpression:

τ =EBNNT × e × Aeff × β

2πRBNNT×

sinhβL2 − x

coshβ L

2

(2)

where,

β =

G′

HAEBNNT

2π

Aeffln

RRBNNT

(3)

EBNNT is the elastic modulus of BNNT, used as 750 GPa(Suryavanshi et al., 2004), which is the lower end of reportedrange of values. The applied strain ε has been taken as 0.04which is the fracture strain of HA (Pramanik et al., 2005). L is



the average length of BNNT used in this study (2 µm), x is thedistance from end of BNNT. G′

HA ∼ 45 GPa (Hellimich andUlm, 2005; Fritsch et al., 2009) is the shear modulus of HA.The radius of matrix coaxial cylinder, R, has been calculatedusing the following relationship,

RRBNNT

2=

π

4Vf(4)

Vf ∼ 0.059, is the volume fraction of BNNT in HA matrix. Thecalculated interfacial shear stress at HA–BNNT interface,τ, is0.35–3 GPa. Thus, BNNT debonding from HA matrix requiresa minimum shear stress ≥0.35 GPa. The computed τ has beenused in calculation of pullout energy (Gpullout) for HA–BNNTsystem, using Eq. (5).

Gpullout =

Vf l2τ

3RBNNT(5)

where l is the pullout length of BNNTs (100–800 nm)measuredfrom SEM images of fracture surface. The computed Gpullout

for BNNT from HA matrix is 2–100 J/m2, which is greatercompared to the fracture energy of monolithic HA of 1 J/m2

(Nakahira and Eguchi, 2001). The higher value of BNNTpullout energy from HA matrix highlights its effectiveness asreinforcement for toughening.

3.3. Elastic modulus and fracture toughness of HA–BNNTcomposite

Measurement of the elastic modulus (E) is performed onthe polished cross section of the sintered pellets usingnanoindentation technique. Since nanoindentation provideslocalized mechanical properties, more than 100 indents weremade at randomly chosen regions in each sample. Total areacovered by the indents was > 5000 µm2 in each sample.The representative load vs. displacement curves for HA andHA–BNNT are shown in Fig. 6(a). Low indentation depthindicates higher hardness for HA–BNNT (12 ± 2 GPa), whichis 100% increase over HA (6 ± 1 GPa). Elastic modulus values,calculated from the unloading part of the load–displacementcurves, shows a higher value of 205 ± 15 GPa for HA–BNNTcompared to 93±9 GPa for HA (Fig. 6(b)). BNNT reinforcementincreases the E of HA matrix by 120%. Statistical distributionof E values shows similar spread for HA and HA–BNNT(Fig. 6(b)) indicating uniform improvement in elastic modulusthroughout the HA matrix with BNNT reinforcement. Theimprovement in elastic modulus with BNNT addition isattributed to two major factors: (i) higher elastic modulusof BNNT reinforcement and (ii) strong bonding at HA–BNNTinterface. The E value of BNNT (750–1200 GPa) is muchhigher than that of HA (100 GPa (Ravaglioli and Krajewski,1992)). Further, BNNT retains the defect-free structure duringSPS, which makes its contribution more effective towardsimproving the elastic modulus. HA–BNNT exhibits strongcoherent interface, as discussed in Section 3.2. Duringapplication of stress, strong bonding at the interface helps intransferring the load effectively from HA matrix to BNNT. Asa result, with same amount of stress, the resultant strain inthe HA–BNNT composite is lower than HA. Reduction in theelastic strain causes increase in E for the composite. Apartfrom these two, the increased density in HA–BNNT composite

also contributes towards the increasing E. Density of thecomposite structure increased by 5% with BNNT additionto HA. However, the effect of density on strengthening ofcomposite structure cannot be isolated from the contributionof BNNTs. In fact, higher density of the sintered structureis caused by BNNT content, as processing conditions weresame for HA with and without BNNTs. The role of BNNTs intoughening mechanism is discussed below.

Vickers indentation method has been used to determinethe fracture toughness of the composite using radial crackmeasurement method (Anstis et al., 1981). Fig. 7(a) and(b) show Vickers indents on the polished cross-sections ofHA and HA–BNNT pellets, respectively. Indent on HA–BNNTshows significantly smaller impression with shorter radialcracks as compared to HA. Smaller indentation with sameload indicates higher hardness for HA–BNNT (5.5 ± 0.12 GPa)than HA (2.4 ± 0.05 GPa). The absolute values of hardnessare different in Vickers and nanoindentation experiments.But the direct comparison is not justified owing to thedifference in tip geometry, measurement length scale andthe vast difference in applied load. On a relative scale,Vickers’ hardness shows a 129% increase with BNNT addition,which is comparable to 100% increase in hardness fromnanoindentation experiments. Comparable improvement inthe mechanical properties at multiple length scales canbe attributed to homogeneous distribution of BNNTs inHA matrix and higher density of the composite structure.Significant improvement in the hardness for HA–BNNT couldalso be due to its finer grain size. The increase in the hardnesswith decreasing grain size in metals and ceramics, includingHA, could be explained through the Hall–Petch mechanism(Huang et al., 2007; Wang and Shaw, 2009).

Fracture toughness of the composite structures has beenevaluated using Eq. (6) (Anstis et al., 1981) expressed as:

KIC = 0.016EH

1/2 P

c3/2(6)

where, P is the applied load, E is the elastic modulus, H is theVickers hardness and c is the radial crack length (measuredfrom the center of the indent using SEM images). E valuesmeasured by nanoindentation were used to compute KIC ofthe composite. KIC for HA–BNNT is 1.6 (±0.3) MPa m0.5, whichis 86% higher than HA with 0.85 (±0.3) MPa m0.5. Due tosimilar amount of improvement in E and H, the E/H ratiodoes not have significant contribution in the improvementof KIC in HA–BNNT. Hence, the radial crack length remainsthe determining factor. Shorter length of radial crack inHA–BNNT is attributed to two major factors — (i) grain sizerefinement and (ii) crack bridging by BNNTs. Wang and Shawhave reported simultaneous improvement in hardness andtoughness in sintered HA pellet due to refinement in grainsize (Wang and Shaw, 2009). Deflection of crack and transitionof cracking mode from transgranular to intergranular are thereasons for the improvement in fracture toughness of HAwith refined grain size. The presence of BNNT causes grainsize refinement in HA matrix as discussed in Section 3.1.Further, due to higher pullout energy of BNNT fromHAmatrix(refer to Section 3.2) cracks propagate through HA, but getsrestricted in the vicinity of BNNT, as more energy is requiredfor interface debonding. Fig. 8 shows BNNT bridges in a



a b

Fig. 6 – (a) Load vs. displacement plot for HA and HA–BNNT composite obtained by nanoindentation; (b) Statisticaldistribution of E value in HA and HA–BNNT composites measured for more than 100 nanoindents in each sample.

a

b

Fig. 7 – Vickers indent impressions showing radial crackgeneration in (a) HA and (b) HA–BNNT sintered pellets.

radial crack generated by the indentation. Deflection in the

cracking path at each BNNT bridge reveals the absorption

of fracture energy at strong HA–BNNT interface and thus

increases fracture toughness.

Fig. 8 – BNNTs bridging the radial crack generated from theindent.

3.4. Tribological behavior of HA–BNNT composite

Tribological behavior of HA and HA–BNNT composite isquantified in terms of wear volume loss and coefficient offriction (CoF). Wear volume loss is inversely related to thewear resistance. The CoF and cumulative volume loss for HAand HA–BNNT sintered pellets are shown in Fig. 9(a) and (b),respectively. Each point on Fig. 9(a) shows an average value ofCoF for 25 m interval. The error bars on Fig. 9 are based onthree wear tracks studied for each composition.

The CoF increases by ∼25% with BNNT reinforcement inHA (Fig. 9(a)). Higher KIC and E of HA–BNNT composite causemore resistance to mass removal and as a result increasein the lateral (transverse) force. The increase in the lateralforce with a constant normal force causes higher CoF forHA–BNNT. Hexagonal boron nitride sheet is known as a goodlubricator and its presence is reported to decrease the CoFof the system (Pawlak et al., 2008, 2009). Hence, increasein CoF in HA–BNNT could be due to absence of peeled offh-BN sheet on wear track. This observation is in contraryto the presence of graphene sheets providing lubrication inwear track of HA–CNT composite (Lahiri et al., 2010a). Thisdiscrepancy can be explained in terms of the capability of



a b

Fig. 9 – (a) Coefficient of friction and (b) wear volume loss for HA and HA–BNNT plotted against sliding distance duringball-on-disc wear.

a b

c d

Fig. 10 – Ball-on-disc wear tracks on (a, c) HA shows flat morphology due to abrasive wear whereas and (b, d) HA–BNNTshows localized plastic deformation and pile-up.

BNNT to withstand high amount of deformation withoutgetting damaged (Golberg et al., 2007a). Shear force appliedby wear probe may not be sufficient to peel h-BN layer fromthe BNNT surface.

The presence of BNNT decreases the wear volume lossof HA matrix by 75% (Fig. 9(b)). Similar amount of wear lossfor different tracks (indicated as error bars in plots) showsthe homogeneous tribological behavior of the compositestructure. Larger error bars in wear loss for HA could be dueits higher porosity and inhomogeneous microstructure. Theincrease in the wear resistance of HA–BNNT is the result ofits improved mechanical property (E, H and KIC). Toughenedmatrix of BNNT reinforced structure inhibits loss of massdue to fracture and chipping during wear. In order to findout the effect of E, H and KIC on the wear volume loss, amodel for brittle ceramic proposed by Evans and Marshall hasbeen employed (Evans and Marshall, 1981). The relationshipbetween wear loss volume (V) and mechanical properties

(E,H,KIC), defined by the model, is as follows:

V = P1.125K−0.5IC H−0.625

EH

0.8S (7)

where, P is the normal load, and S is the total travellingdistance on wear track. The computed volume loss for thepresent study shows a 65% reduction in the wear volumewith BNNT addition. Comparable outcomes of wear lossimprovement from experiment (75%) and computation (65%)supports the hypothesis of wear resistance being governedby the mechanical property improvement in HA–BNNTcomposite structure.

Further insight into differential wear behavior of HA andHA–BNNT sintered structure is obtained by investigatingthe morphology of wear track. Fig. 10(a) and (c) show flatmorphology in the HA wear track which is an indicator ofabrasive wear mechanism with mass being totally removed.The wear track on HA–BNNT shows displacement of mass



a

b

Fig. 11 – (a) Sword in sheathe type structure shown byBNNT; (b) BNNT bridging of mass in the wear track ofHA–BNNT.

towards the outer edge of the track, resulting in pile-up(Fig. 10(b) and (d)). Such behavior is specific to plasticdeformation, which is not common in brittle ceramics likeHA. Reinforcement of BNNT in HA matrix is responsible forsuch behavior. Huang et al. have also reported superplasticityintroduced in Al2O3 and Si3N4 ceramics as a result of BNNTreinforcement (Huang et al., 2007). Though the study byHuang et al. reported the high temperature property of thecomposite, the superplastic behavior is attributed to mainlytwo factors — (i) obstacle in dynamic grain growth at highertemperature and (ii) the ‘sword in sheathe’ phenomenon ofload transfer in BNNT (Huang et al., 2007). Although weartests in this study were conducted at room temperatureand in dry condition, localized high temperature at thepoint of contact between two surfaces may exist due tohigh friction. Such localized high temperature may aid theplastic deformation as seen in Fig. 10(d). ‘Sword in sheathe’behavior of BNNT is possible at room temperature and isobserved in Fig. 11(a). Sword in sheathe’ indicates effectiveload transfer from the matrix to the outermost wall ofBNNT. The transferred load is then carried to inner wallsin a stepwise manner upon breakdown of the outer walls,leading to the ‘sword in sheathe’ structure formation. The

gradual sliding of BNNT layers converts the applied forceto strain energy. This energy absorption mechanism causesplastic deformation in HA–BNNT composite. As a result,the mass is not totally removed from the track, but getsdislodged towards the periphery, still being held together withBNNT bridges. Fig. 11(b) shows a BNNT bridge supporting thedislodged mass on wear track.

3.5. In-vitro biocompatibility of HA–BNNT to osteoblasts

Proliferation and viability of osteoblast cells are evaluatedon HA and HA–BNNT surface after in-vitro culturing for 1,3 and 5 days. Osteoblasts are the cells that attach first tothe orthopedic implant surface. They actively participate innew bone formation by forming collagen matrix and thenassisting deposition of apatite crystal on it. Thus, the growthand proliferation of osteoblast cells on an implant surface isextremely important for the bone generation and integration.Proliferation of the osteoblast cells is assessed qualitativelyby observing the population of FDA stained live cells on HAand HA–BNNT surface by fluorescence microscopy images.Viability of cells on each surface is evaluated by manuallymeasuring number of live and dead cells after differentdays of culture. Fig. 12 shows the fluorescent images of live(green) and dead (red) cells after 1 and 3 days of cultureon HA and HA–BNNT surfaces. The cells exhibit typicallens shape suggesting the normal cell growth behavior. Thepopulation of the osteoblast also increases visibly from 1to 3 days on both surfaces. This observation indicates thatHA and HA–BNNT surfaces are suitable for osteoblast cellproliferation. Population of osteoblast cells is slightly denseron HA–BNNT surface than HA after 3 days of culture.

Fig. 13 presents the percentage of live osteoblast cellson HA and HA–BNNT surface after 1, 3 and 5 daysof culture. The viability of osteoblast cells on HA–BNNTsurface is comparable to HA surface. Similar observationof osteoblast proliferation in the presence of BNNT wasreported by the present authors on a biodegradable polymersurface (Lahiri et al., 2010b). Gene expression study revealedthat the presence of BNNT positively influence osteoblastdifferentiation and proliferation (Lahiri et al., 2010b). Thereason for such behavior could be attachment of proteins(from the culture medium) on BNNT surface that assistsosteoblast proliferation. BNNT has a natural affinity forprotein attachment on its surface (Zhi et al., 2005b).Probability of the protein absorption on the BNNT surfaceis further supported by the similar behavior established forCNT, the structural analogue for BNNT (Matsuoka et al., 2009;Akasaka et al., 2010).

Apart from the bone growth and integration on implantsurface, another major concern regarding the orthopedicapplication could be the cytotoxicity of wear debris.Osteoblast and macrophages are the two major celllineages related to cytotoxicity of wear debris in orthopedicapplication (Goodman and Ma, 2010). A recent study bypresent authors has shown that bare BNNTs do not exhibitany cytotoxic response to osteoblasts and macrophages(Lahiri et al., 2010b). Thus, the presence of any loose orembedded BNNT in wear debris generated from HA–BNNT



Fig. 12 – Fluorescent images of osteoblast cells grown on HA and HA–BNNT surface for (a, b) 1 day and (c, d) 3 days. The livecells are stained in green with FDA and dead cells in red with PI. (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)

Fig. 13 – Osteoblast cell viability on HA and HA–BNNTsurfaces for 1, 3 and 5 days of culture (p < 0.05).

implant during service would not impose any negative effecton its biocompatibility.

4. Conclusions

The present study evaluates BNNT as a potential reinforce-ment to HA for improving its mechanical and tribologicalproperties. Higher thermal conductivity of BNNT reinforce-ment results in dense sintered structure due to accelerateddiffusion during SPS. BNNT structure was not damaged byhigh temperature and pressure applied during SPS. A signifi-cant 120% increase in the elastic modulus of HA–BNNT com-

posite is obtained. BNNT also acts as grain refiner by pinningthe grain boundary during sintering, which helps in simul-taneous increase in hardness (129%) and fracture toughness(86%) of the composite structure. High BNNT pullout energyfrom the matrix helps in improving the fracture toughness. A75% increase in the wear resistance of HA–BNNT is attributedto the improvement in elasticmodulus, hardness and fracturetoughness. Wear surface morphology reveals the transforma-tion of brittle abrasive fracture in HA to plastic deformationand pile-up in HA–BNNT. BNNT bridging, effective load trans-fer at matrix-reinforcement interface and sword in sheathe’phenomenon is responsible for the plastic deformation inHA–BNNT composite. The presence of BNNTs in HA does notnegatively influence the osteoblast proliferation and viability.

Acknowledgements

The authors acknowledge support from the research facilityat Advanced Materials Engineering and Research Institute(AMERI) in Florida International University and Mr. NealRicks, manager, AMERI. The authors are also thankful to theCenter for study of Matters in Extreme Conditions (CeSMEC)and Prof. S. Saxena for extending the use of Micro-RamanSpectroscopy facility for research purpose. A.A. acknowledgesfunding from the National Science Foundation CAREER Award(NSF-DMI-0547178), Office of Naval Research (N00014-08-1-0494) and DURIP program (N00014-06-0675). D.L acknowledgessupport from Dissertation Evidence Acquisition Fellowshipby University Graduate School of Florida InternationalUniversity.



Appendix. Supplementary data

Supplementary material related to this article can be foundonline at doi:10.1016/j.jmbbm.2010.09.005.

R E F E R E N C E S

Akasaka, T., Yokoyama, A., Matsuoka, M., Hashimoto, T., Watari,F., 2010. Thin films of single-walled carbon nanotubes promotehuman osteoblastic cells (Saos-2) proliferation in low serumconcentrations. Mater. Sci. Eng. C 30, 391–399.

Anstis, G.R., Chantikul, P.B., Lawn, R., Marshall, D.B., 1981. Acritical evaluation of indentation techniques for measuringfracture toughness: I, Direct crack measurements. J. Am.Ceram. Soc. 64 (9), 533–538.

Balani, K., Anderson, R., Laha, T., Andara, M., Tercero, J., Crumpler,E., Agarwal, A., 2007a. Plasma-sprayed carbon nanotubereinforced hydroxyapatite coatings and their interaction withhuman osteoblasts in vitro. Biomaterials 28, 618–624.

Balani, K., Chen, Y., Harimkar, S.B., Dahotre, N.B., Agarwal,A., 2007b. Tribological behavior of plasma-sprayed carbonnanotube-reinforced hydroxyapatite coating in physiologicalsolution. Acta Biomater. 3, 944–951.

Bansal, N.P., Hurst, J.B., Choi, S.R., 2006. Boron nitride nanotubes-reinforced glass composites. J. Am. Ceram. Soc. 89 (1), 388–390.

Chaim, R., Levin, M., Shlayer, A., Estournes, C., 2008. Sinteringand densification of nanocrystalline ceramic oxide powders:a review. Adv. Appl. Ceram. 107 (3), 159–169.

Chen, X., Wu, P., Rousseas, M., Okawa, D., Gartner, Z., Zettl, Z.A.,Bertozzi, C.R., 2009. Boron nitride nanotubes are noncytotoxicand can be functionalized for interaction with proteins andcells. J. Am. Chem. Soc. Comm. 131, 890–891.

Chen, Y., Zhang, T.H., Gun, C.H., Yu, G., 2007. Wear studiesof hydroxyapatite composite coating reinforced by carbonnanotubes. Carbon 45, 998–1004.

Chen, Y., Balani, K., Agarwal, A., 2008. Analytical modelto evaluate interface characteristics of carbon nanotubereinforced aluminum oxide nanocomposites. Appl. Phys. Lett.92, 011916(1-3).

Cheng, C., Muller, K.H., Koziol, K.K.K., Skepper, J.N., Midgley,P.A., Welland, M.E., Porter, A.E., 2009. Toxicity and imaging ofmulti-walled carbon nanotubes in human macrophage cells.Biomaterials 30, 4152–4160.

Choi, S.R., Bansal, N.P., Garg, A., 2007. Mechanical and microstruc-tural characterization of boron nitride nanotubes-reinforcedSOFC seal glass composite. Mater. Sci. Eng. A 460–461, 509–515.

Chopra, N.G., Luyken, R.J., Cherrey, K., Crespi, V.H., Cohen, M.L.,Louie, S.G., Zettl, A., 1995. Boron nitride nanotubes. Science269, 966–967.

Chopra, N.G., Zetll, A., 1998. Measurement of the elastic modulusof a multi walled boron nitride nanotubes. Solid StateCommun. 105 (5), 297–300.

Ciofani, G., Raffa, V., Menciassia, A., Cuschieria, A., 2009. Boronnitride nanotubes: an innovative tool for nanomedicine.Nanotoday 4, 8–10.

Evans, A.G., Marshall, B., 1981. Wear mechanisms in ceramics.In: Rigney, D.A. (Ed.), Fundamentals of Friction and Wear ofMaterials. ASM, Materials Park, Ohio, pp. 439–452.

Fioritto, S., 2008. Carbon Nanotubes: Angels or Demons? PanStanford, Singapore, ISBN: 9814241016, 9789814241014.

Fritsch, A., Dormieus, L., Hellimich, C., Sanahuja, J., 2009. Me-chanical behavior of hydroxyapatite biomaterials: an experi-mentally validated micromechanical model for elasticity andstrength. J. Biomed. Mater. Res. 88A, 149–161.

Fu, L., Khor, K.A., Joo, P.L., 2002. Effects of yttria-stabilized zirconiaon plasma-sprayed hydroxyapatite/yttria-stabilized zirconiacomposite coating. J. Am. Ceram. Soc. 84 (4), 800–806.

Gaona, M., Lima, R.S., Marple, B.R., 2007. Nanostructuredtitania/hydroxyapatite composite coatings deposited by highvelocity oxy-fuel (HVOF) spraying. Mater. Sci. Eng. A 458,141–149.

Gautier, S., Champion, E., Bernache Assollant, D., 1997. Processing,microstructure and toughness of Al2O3 platelet-reinforcedhydroxyapatite. J. Eur. Ceram. Soc. 17, 1361–1369.

Golberg, D., Bai, X.D., Mitome, M., Tang, C.C., Zhi, C.Y., Bando,Y., 2007a. Structural peculiarities of in situ deformation of amulti-walled BN nanotube inside a high-resolution analyticaltransmission electron microscope. Acta Mater. 55, 1293–1298.

Golberg, D., Bando, Y., Tang, C., Zhi, C., 2007b. Boron nitridenanotubes. Adv. Mater. 19, 2413–2432.

Goodman, S.B., Ma, T., 2010. Cellular chemotaxis inducedby wear particles from joint replacements. Biomaterials.doi:10.1016/j.biomaterials.2010.03.046.

Grossin, D., Rollin-Martinet, S., Estournès, C., Rossignol, F.,Champion, E., Combes, C., Rey, C., Geoffroy, C., Drouet, C.,2009. Biomimetic apatite sintered at very low temperature byspark plasma sintering: physico-chemistry andmicrostructureaspects. Acta Biomater. doi:10.1016/j.actbio.2009.08.021.

Gu, Y.W., Loha, N.H., Khor, K.A., Tor, S.B., Cheang, P., 2002. Sparkplasma sintering of hydroxyapatite powders. Biomaterials 23,37–43.

Gu, Y.W., Khor, K.A., Cheang, P., 2004. Bone-like apatite layerformation on hydroxyapatite prepared by spark plasmasintering (SPS). Biomaterials 25, 4127–4134.

Guo, L., Singh, R.N., 2008. Selective growth of boron nitridenanotubes by plasma-enhanced chemical vapor deposition atlow substrate temperature. Nanotechnology 19, 065601 (6).

Guo, X., Gough, J.E., Xiao, P., Liu, J., Shen, Z., 2007. Fabricationof nanostructured hydroxyapatite and analysis of humanosteoblastic cellular response. J. Biomed. Mater. Res. A 82,1022–1032.

Hellimich, C., Ulm, F.J., 2005. Poro-micromechanics of bone:impact loading and wave propagation. Mater. Res. Soc. Sym.Proc. 844, Y8.6.1–Y8.6.7.

Huang, Q., Bando, Y., Xu, X., Nishimura, T., Zhi, C, Tang, C.,Xu, F., Gao, L., Golberg, D., 2007. Enhancing superplasticityof engineering ceramics by introducing BN nanotubes.Nanotechnology 18, 485706 (7).

Kay, M.I., Young, R.A., 1964. Crystal structure of hydroxyapatite.Nature 204, 1050–1052.

Lahiri, D., Singh, V., Keshri, A.K., Seal, S., Agarwal, A., 2010a. Car-bon nanotube toughened hydroxyapatite by spark plasma sin-tering: microstructural evolution and multi-scale tribologicalproperties. Carbon. doi:10.1016/j.carbon.2010.04.047.

Lahiri, D., Rouzaud, F., Richard, T., Keshri, A.K., Bakshi, S.R.,Kos, L., Agarwal, A., 2010b. Boron nitride nanotube re-inforced polylactide–polycaprolactone copolymer com-posite: mechanical properties and cytocompatibility withosteoblasts and macrophages in vitro. Acta Biomater.doi:10.1016/j.actbio.2010.02.044.

Lan, H.P., Ye, L.H., Zhang, S., Peng, L.M., 2009. Transverse dielectricproperties of boron nitride nanotubes by ab initio electric fieldcalculations. Appl. Phys. Lett. 94, 183110(1-3).

Li, H., Khor, K.A., Chow, V., Cheang, P., 2007. Nanostructural char-acteristics, mechanical properties, and osteoblast response ofspark plasma sintered hydroxyapatite. J. Biomed. Mater. Res. A82, 296–303.

Li, J., Fartash, B., Hermansson, L., 1995. Hydroxyapatite-aluminacomposites and bone-bonding. Biomaterials 16, 417–422.

Li, S., Izui, H., Okano, M., 2008. Densification, microstructure andbehavior of hydroxyapatite ceramics sintered by using sparkplasma sintering. J. Eng. Mater. Technol. 130, 031012(1–7).



Liu, J., Glasmacher, U.A., Lang, M., Trautman, C., Voss, K.O.,Neumann, R., Wagner, G.A., Miletich, R., 2008. Ramanspectroscopy of apatite irradiated with swift heavy ions withand without simultaneous exertion of high pressure. Appl.Phys. A 91, 17–22.

Lu, K., 2008. Sintering of nanoceramics. Inter. Mater. Rev. 53 (1),21–37.

Matsuoka, M., Akasaka, T., Totsuka, Y., Watari, F., 2009. Strongadhesion of Saos-2 cells to multi-walled carbon nanotubes.Mater. Sci. Eng. B. doi:10.1016/j.mseb.2009.12.044.

Munir, Z.A., Tamburini, U.A., Ohyanagi, M., 2006. The effect ofelectric field and pressure on the synthesis and consolidationof materials: a review of the spark plasma sintering method.J. Mater. Sci. 41, 763–777.

Nag, A., Raidongia, K., Hembram, K.P.S.S., Datta, R., Waghmare,U.V., Rao, C.N.R., 2010. Graphene analogues of BN: novelsynthesis and properties. ACS Nano. doi:10.1021/nn9018762.

Nakahira, A., Eguchi, K., 2001. Evaluation of microstructure andsome properties of hydroxyapatite/Ti composites. J. Ceram.Proc. Res. 2 (3), 108–112.

Olevsky, E.A., Kandukuri, S., Froyen, L., 2007. Consolidationenhancement in spark-plasma sintering: impact of highheating rates. J. Appl. Phys. 102, 114913(1–12).

Oliver, W.C., Pharr, G.M., 1992. An improved technique fordetermining hardness and elastic modulus using load anddisplacement sensing indentation experiments. J Mater. Res.7 (6), 1564–1583.

Pawlak, Z., Pai, R., Bayraktar, E., Kaldonski, T., Oloyede, A., 2008.Lamellar lubrication in vivo and vitro: friction testing ofhexagonal boron nitride. BioSystems 94, 202–208.

Pawlak, Z., Kaldonski, T., Pai, R., Bayraktar, E., Oloyede, A.,2009. A comparative study on the tribological behaviourof hexagonal boron nitride (h-BN) as lubricating micro-particles—an additive in porous sliding bearings for a carclutch. Wear 267, 1198–1202.

Porter, D.A, Easterling, K.E., 2001. Phase Transformation in Metalsand Alloys, 2nd Edition. CRC Press, Cheltenham, UK, pp.142–169.

Pramanik, S., Agarwal, A.K., Rai, K.N., 2005. Development of highstrength hydroxyapatite for hard tissue replacement. TrendsBiomater. Artif. Organs 19 (1), 46–51.

Que, W., Khor, K.A., Xu, J.L., Yu, L.G., 2008. Hydroxyapatite/titaniananocomposites derived by combining high-energy ballmilling with spark plasma sintering processes. J. Eur. Ceram.Soc. 28, 3083–3090.

Ragulya, A.V., 2008. Consolidation of ceramic nanopowders. Adv.Appl. Ceram. 107 (3), 118–134.

Ravaglioli, A., Krajewski, A., 1992. Bioceramics: Materials,Properties, Application. Chapman & Hall, London, p. 44.

Ravichandran, J., Manoj, A.G., Liu, J., Manna, I., Carroll, D.L., 2008. Anovel polymer nanotube composite for photovoltaic packagingapplications. Nanotechnology 19, 085712(5).

Sarkar, S.K., Youn, M.H., Oh, I.H., Lee, B.T., 2007. Fabrication of CNTreinforced HAp composites by spark plasma sintering. Mater.Sci. Forum. 534–536, 893–896.

Shen, H., 2009. Thermal-conductivity and tensile-properties ofBN, SiC and Ge nanotubes. Comput. Mater. Sci. 47, 220–224.

Singh, R., Pantarotto, D., Lacerda, L., Pastorin, G., Klumpp,C., Prato, M., Bianco, A., Kostarelos, K., 2006. Tissuebiodistribution and blood clearance rates of intravenouslyadministered carbon nanotube radiotracers. Proc. Natl. Acad.Sci. USA 103 (9), 3357–3362.

Singh, S., Pei, Y., Miller, R., Sundararajan, P.R., 2003. Long-range,entangled carbon nanotube networks in polycarbonate. Adv.Funct. Mater. 13 (11), 868–872.

Singhal, S.K., Srivastava, A.K., Pant, R.P., Halder, S.K., Singh,B.P., Gupta, A.K., 2008. Synthesis of boron nitride nanotubesemploying mechanothermal process and its characterization.J. Mater. Sci. 43, 5243–5250.

Suryavanshi, A.P., Yu, M.F., Wen, J., Tang, C., Bando, Y., 2004. Elasticmodulus and resonance behavior of boron nitride nanotubes.Appl. Phys. Lett. 84 (14), 2527–2529.

Terao, T., Zhi, C., Bando, Y., Mitome, M., Tang, C., Golberg, D., 2010.Alignment of boron nitride nanotubes in polymeric compositefilms for thermal conductivity improvement. J. Phys. Chem.114, 4340–4344.

Terrones, M., Romo-Herrera, J.M., Cruz-Silva, E., López-Urías,F., Muñoz-Sandoval, E., Velázquez-Salazar, J.J., Terrones, H.,Bando, Y., Golberg, D., 2007. Pure and doped boron nitridenanotubes. Mater. Today 10 (5), 30–38.

Usui, Y., Aoki, K., Narita, N., Murakami, N., Nakamura, I.,Nakamura, K., Ishigaki, N., Yamazaki, H., Horiuchi, H., Kato,H., Taruta, S., Kim, Y.A., Endo, M., Saito, N., 2008. Carbonnanotubes with high bone-tissue compatibility and bone-formation acceleration effects. Small 4 (2), 240–246.

Wang, J., Shaw, L.L., 2009. Nanocrystalline hydroxyapatite withsimultaneous enhancements in hardness and toughness.Biomaterials 30, 6565–6572.

Wang, Q., Shirong, G., Dekun, Z., 2005. Nano-mechancial prop-erties and tribological behaviors of nanosized HA/partially-stabilized zirconia composites. Wear 259, 952–957.

Wang, W., Cheng, J., Agarwal, D.K., Malshe, A.P., Liu, H.,2009. Improved mechanical properties of nanocrystallinehydroxyapatite coating for dental and orthopedic implants.Mater. Res. Soc. Symp. Proc. 1140, 1140-HH03-03.

White, A.A., Best, S.M., Kinloch, I.A., 2007. Hydroxyapatite-carbonnanotube composites for biomedical applications: a review.Int. J. Appl. Ceram. Tech. 4 (1), 1–13.

Xu, J.L., Khor, K.A., Kumar, R., 2007. Physicochemical differencesafter densifying radio frequency plasma sprayed hydroxyap-atite powders using spark plasma and conventional sinteringtechniques. Mater. Sci. Eng. A 457, 24–32.

Yang, K., He, J., Su, Z., Reppert, J.B., Skove, M.J., Tritt, T.M.,Rao, A.M., 2009. Inter-tube bonding, graphene formationand anisotropic transport properties in spark plasmasintered multi-wall carbon nanotube arrays. Carbon.doi:10.1016/j.carbon.2009.10.022.

Yu, L.G., Khor, K.A., Li, H., Cheang, P., 2003. Effect ofspark plasma sintering on the microstructure and in vitrobehavior of plasma sprayed HA coatings. Biomaterials 24,2695–2705.

Zhi, C., Bando, Y., Tang, C., Honda, S., Sato, K., Kuwa-hara, H., Golberg, D., 2005a. Characteristics of boron ni-tride nanotube–polyaniline composites. Angew. Chem. 44,7929–7932.

Zhi, C., Bando, Y., Tang, C., Golberg, D., 2005b. Immobilization ofproteins on boron nitride nanotubes. J. Am. Chem. Soc. 127,17144–17145.

Zhi, C., Bando, Y., Tang, C., 2006a. Boron nitride nan-otube/polystyrene composites. J. Mater. Res. 21 (11), 2794–2800.

Zhi, C., Bando, Y., Tang, C., Honda, S., Sato, K., Kuwahara,H., Golberg, D., 2006b. Purification of boron nitride nan-otubes through polymer wrapping. J. Phys. Chem. B 110,1525–1528.

Zhi, C., Bando, Y., Terao, T., Tang, C., Kuwahara, H., Golberg,D., 2009. Towards thermoconductive, electrically insulatingpolymeric composites with boron nitride nanotubes as fillers.Adv. Funct. Mater. 19, 1857–1862.

author's personal copyweb.eng.fiu.edu/agarwala/pdf/2011/11.pdf · author's personal copy...

Documents