Available online at www.sciencedirect.

com

CERAMICS
INTERNATIONAL
Ceramics International 40 (2014) 5635–5643
www.elsevier.com/locate/ceramint

Improved properties of hydroxyapatite–carbon nanotube biocomposite:

Mechanical, in vitro bioactivity and biological studies
Susmita Mukherjeea,n, Biswanath Kundub,nn, Swarnendu Senc, Abhijit Chandad
a
School of Materials Science and Nanotechnology, Jadavpur University, Kolkata 700032, India
b
Bioceramics and Coating Division, CSIR-Central Glass and Ceramic Research Institute, Kolkata 700032, India
c
Department of Mechanical Engineering, Jadavpur University, Kolkata 700032, India
d
School of Bioscience and Engineering, Jadavpur University, Kolkata 700032, India
Received 2 October 2013; received in revised form 30 October 2013; accepted 31 October 2013
Available online 12 November 2013

Abstract

The present work describes a simple shear mixing technique for developing a hydroxyapatite (HAp)–carbon nanotube (CNT) nanocomposite
and the effect of reinforcement on the physical, mechanical, in vitro bioactivity and biological properties of HAp. XRD and FTIR confirmed that
the main phase of the composites is HAp. HRTEM images demonstrated the formation of a two-dimensional nanocomposite structure, whereas
FESEM images indicated the formation of nanosized HAp grains featuring sporadically distributed CNT molecules. No major phase changes in
HAp were observed with up to 5% added CNT. However, adding more than 1% CNTs caused an increase in internal crystal strain and increased
substitution of CO23
for OH
and PO34
groups in pure HAp. The average crystallite size increased from  46 nm to 100 nm with only 0.5%
added CNT, remained nearly unaffected up to 2% CNTs thereafter and suddenly decreased at 5% CNTs ( 61 nm). The FESEM and HRTEM
images clearly showed the attachment of MWCNT chains on HAp grains, which directly affected the samples' fracture toughness and flexural
strength. Of the samples, 1% showed maximum values of K1C, whereas 5% showed maximum values of HV and three-point bending flexural
strength. The in vitro bioactivity indicated increased apatite formation on the sample surface up to 1% CNTs after 24 weeks. However, adding 2%
and 5% CNTs resulted in a manifold increase in apatite formation up to 12 weeks, after which dissolution increased up to 24 weeks, possibly due
to increased substitution of CO23
for OH
and PO34
groups. This result is confirmed by the FTIR studies. For all added CNT contents, all
samples exhibited high haemocompatibility. However, there was a compromise between the observed mechanical properties and in vitro
bioactivity studied up to 24 weeks, and care must be taken before selecting any final application of the nanocomposites.
& 2013 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Keywords: Hydroxyapatite; Carbon nanotube; In vitro bioactivity; Fracture toughness; Haemocompatibility

1. Introduction used as a reinforcement material in the preparation of nanocom-

posites, which is expected to produce composites with improved
Since the discovery of carbon nanotubes by Ijima [1], CNTs mechanical properties compared to those of single-phase materials.
have emerged as a prospective candidate for materials research Ruoff et al. determined that when an advancing crack with
due to their special structural features and excellent properties, sufficient stress intensity at its tip encounters multiwalled carbon
such as their extremely small dimensions, high aspect ratio (103– nanotube (MWCNT) chains, the concentric layers collapse and
104) and low density, in addition to high tensile strength ( 60 absorb the associated energy, arresting the propagation of the
GPa), high rigidity (Young's modulus  1 TPa), high resilience, crack [5]. This mechanism was identified as a toughening
superb flexibility and excellent electrical (4105 Sm
1) and mechanism in CNT-based composites. Lupo et al. successfully
thermal properties [2–4]. Therefore, CNTs are primarily being fabricated ZrO2/CNT composites by hydrothermal crystallisation
n
[6], and Seeger et al. synthesised MWCNT/SiO2 composites via
Corresponding author. Tel.: þ91 9674459006.
nn partial matrix melting with a Nd:YAG laser [7].
Corresponding author. Tel.: þ 91 9831772081.
E-mail addresses: [email protected] (S. Mukherjee), Hydroxyapatite (HAp) has been exclusively used as a bone
[email protected], [email protected] (B. Kundu). implant material due to the similarity between its chemical

0272-8842/$ - see front matter & 2013 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
http://dx.doi.org/10.1016/j.ceramint.2013.10.158
5636 S. Mukherjee et al. / Ceramics International 40 (2014) 5635–5643

composition and crystal structure and that of the inorganic steps in argon. In the first step, the temperature was increased
component of bone [8] as well as its excellent ability to bond to 700 1C at a heating rate of 3 1C/min. Then, the temperature
with living bone tissues. However, the intrinsic brittleness and was increased to 1250 1C at a heating rate of 6 1C/min (1 h
poor mechanical properties of sintered HAp restricts its dwelling at the highest temperature).
employment in major load-bearing applications. Pure HAp
has been reported to have low toughness (0.8–1.2 MPa m1/2) 2.3. Characterisation
and poor flexural strength (o 140 MPa) compared to human
bone [8]. Therefore, HAp is being predominantly used either The apparent porosity (A.P.) and bulk density (B.D.) of
as a coating material on various implant materials, such as the sintered specimens were estimated by Archimedes'
steel, titanium and titanium alloys [9], or as a scaffold for water displacement method using the formulae [B.D. ¼ D/
tissue engineering [10]. To improve the mechanical properties (W
S) g/cm3] and [A.P. ¼ (W
D)/(W
S)100%], where D
of HAp without sacrificing its biological properties, different is the dry weight, W is the soaked weight and S is the
reinforcing phases, such as zirconia, glass, silver, carbon fibre, suspended weight, and the relative density was calculated from
alumina and TiO2, have been used [11–14]. Some advanced the theoretical density.
processing techniques (e.g., high pressure and microwave X-ray diffraction (XRD, Ultima III Rigaku, Japan) and
sintering [15]) have also been investigated recently. Research- Fourier transform infrared spectroscopy in the mid-IR range
ers are also using carbon nanotubes or nanofibres (CNT/CNF) (FTIR, IR Prestige 21, 200VCE, Schimadzu, Japan) were used
as a reinforcing material in the HAp matrix to improve its to characterise the prepared phases. The microstructure and
mechanical properties [16–18]. Different techniques, such as morphology of the nanocomposites were studied using field
the sol–gel process [17], laser surface alloying [19], spark emission scanning electron microscopy (FESEM, Hitachi,
plasma sintering [9] and high-precision plasma spraying [9], S4800, Japan) and high-resolution transmission electron
have been used to prepare HAp–CNT composites. microscopy (HRTEM, JEOL, JEM 2100, Japan).
Most of the above-mentioned processes require the functio- The hardness and fracture toughness (K1C) of the polished
nalisation of ceramics, which further requires the use of sintered samples (o0.02 μm Ra) were evaluated using
precision techniques. In the present work, we adopted an an automated Vickers hardness testing machine (LECO,
alternative and cost-effective shear mixing technique to prepare LV-700AT, MI) via the indentation method (ASTM C1327).
CNT–HAp composites with different proportions of pristine Four different indentation loads of 0.3, 1, 3 and 5 kg were applied
CNT molecules and investigated the effect of the reinforcement to each sample with a holding time of 10 s. A minimum of six
on the physical, mechanical, biological and in vitro bioactivity measurements were obtained for each sample, and the average
properties of HAp. The CNT–HAp composites were prepared was used to calculate VH. K1C was calculated using the equation
by adding different weight percentages of CNT, and a for a radial-median crack: K 1C ¼ 0:016ðE=HÞ0:5 ðP=C 1:5 Þ, where
comparative analysis of the various properties is reported. E is Young's modulus (obtained from the nano-indentation
method, not discussed here), H is the Vickers hardness, P is
2. Experimental the indentation load and C is the half-radial crack length. This
equation was used because the c/a value was greater than 2.5,
2.1. Materials where c is the average crack length and a is the average crack
diagonal [21,22].
Analytical grade ortho-phosphoric acid (H3PO4) and cal- The flexural strength of the nanocomposites was evaluated
cium hydroxide (Ca(OH)2) (S.D. Fine-Chem Ltd., India) were using a universal testing machine (Instron 4204) by three-point
used as starting reagents for the synthesis of pure hydroxya- bending following ASTM C1674, and fractography images
patite (HAp) powder via the wet chemical method. The were obtained using an FESEM. For the flexural strength
detailed methods used for this synthesis and characterisation measurements, bar samples were prepared, cold-isostatically
have been described elsewhere [20]. MWCNTs (4 95% pressed at 150 MPa and sintered under an argon atmosphere at
purity, 10–30-nm diameter and 1–10 μm average length) were 1250 1C. The flexural strength was calculated by using the
procured from Nanoshel LLC, USA. equation s ¼ 3Fl=2bd2 , where l is the specimen length, F (N) is
the total force applied to the specimen by the loading pin and b
2.2. HAp–CNT composite preparation and d are the specimen width and thickness, respectively.
The haemocompatibility of the nanocomposite samples was
Calcined (800 1C) HAp and CNT powders (with four tested with the samples in contact with mammalian blood to
different CNT contents of 0.5%, 1%, 2% and 5%) were mixed verify their cytotoxicity for their use in biomedical implants
counterclockwise in a high-energy planetary ball mill (Fritsche (following the guidelines of ASTM F756 specifications),
Pulverissette 5, Germany) at 300 rpm for 5 h in acetone. To where the physiological tolerance of the implant material
minimise contamination during milling, a zirconia pot and was evaluated by calculating the % haemolysis in blood. In
balls with a ball-to-powder weight ratio of 10:1 were used. The this process, goat blood, which was pre-treated with sodium
milled powders were sieved and used to prepare pellet samples citrate (anti-coagulant), was collected and diluted with normal
(12 mm diameter  4 mm thick) in a uniaxial press (PEECO, saline to a ratio of 8:10. The samples, which were in the form
India) at 150 MPa for 2 min. Sintering was performed in two of round pellets (with no sharp edges), were placed in standard
 S. Mukherjee et al. / Ceramics International 40 (2014) 5635–5643 5637

test tubes containing 10 mL of saline and incubated at 37 1C addition of the CNTs, the B.D. first increased up to 1% CNT,
for 30 min to allow thermal equilibration. Then, 0.2 mL of the then decreased up to 5% CNT, and the opposite trend was
diluted blood was added to the sample test tubes, mixed gently observed for A.P. The open porosity was determined to be
and incubated at 37 1C for 60 min. A positive control was  19% for the 5% CNT samples. Because the density of CNTs
prepared by adding 10 mL of a 1% Na2CO3 solution (Na2CO3 is substantially less than that of HAp, the addition of CNTs
causes large scale rupture of RBC) to 0.2 mL of the diluted reduces the overall density of the composite. However, the
blood, and for the negative control, saline (it does not rupture density of 2% samples was similar to the density of pure HAp.
RBC) was used instead of Na2CO3. Both the positive and
negative controls were incubated at 37 1C for 60 min to allow
thermal equilibration. After the incubation period, the sample
test tubes as well as the positive and negative controls were
centrifuged for 5 min (RCF 500g). The supernatant was
collected, and the OD was measured at 545 nm using a UV
spectroscopic method. The % haemolysis is calculated based
on the following equation:
ODðtestÞ
ODðnegativeÞ
%Haemolysis ¼
ODðpositiveÞ
ODðnegativeÞ
The in vitro bioactivity of the composites was investigated
by immersing disc specimens in simulated body fluid (SBF)
prepared in the laboratory according to the protocol previously
reported by Kokubo [23]. Appropriate quantities of the
reagents, including NaCl, NaHCO3, KCl, K2HPO4  3H2O,
MgCl2  6H2O, CaCl2, Na2SO4 and Tris buffer, were dissolved
in 1 L of de-ionised water under constant stirring to ensure
complete dissolution. Dilute HCl was added to maintain the Fig. 1. Variations in A.P., B.D. and relative density as a function of % CNT.
pH at 7.4 7 0.1 at 37 1C. The compositions of the SBF used in The theoretical density of HAp was 3.16 g/cm3.
the present study are provided in Table 1. A set of two samples
for each composition was immersed in 15 mL of a freshly
prepared SBF solution under static conditions in an incubator
at pH 7.4 and 37 1C and studied for up to 24 weeks. To avoid
saturation, SBF was replaced with freshly prepared solution
every three days. At the end of the 2nd, 4th, 8th, 12th and 24th
week, the samples were removed, dried at 65 1C and weighed.
The surface morphology and microstructures of the samples
were observed using an SEM, and changes in the composition
on the top surface with time were measured using FTIR.

3. Results and discussions

3.1. Physical properties

The changes in the A.P., B.D. and relative density have been
plotted as a function of the CNT content in Fig. 1. Upon the

Table 1
Approximate composition of SBF used in the study [23].

Ionic concentration SBF Blood plasma

(mmol/dm3)

Na þ 142.0 142.0
Kþ 5.0 5.0
Mg2 þ 1.5 1.5
Ca2 þ 2.5 2.5
Cl
147.8 103.0
HCO3
4.2 27.0
HPO4
1.0 1.0
SO4
0.5 0.5 Fig. 2. (A) XRD patterns and (B) FTIR spectra of HAp and HAp–CNT
composites with the pattern of the pure CNTs shown in the inset.
5638 S. Mukherjee et al. / Ceramics International 40 (2014) 5635–5643

3.2. XRD analysis and microstructure observed to occur in an agglomerated state (in bundles) along the
boundaries of the HAp grains. Based on the micrographs, the
Fig. 2A shows the XRD pattern of the HAp–CNT compo- adopted fabrication method does not ensure the uniform disper-
sites, with the inset showing that of pure CNT. No new peaks sion of CNTs. In fact, agglomerates were often formed and the
were observed in the XRD pattern due to the addition of average particle size increased sporadically with no relationship
different CNT amounts, and the 2θ values are in agreement to the addition of CNTs.
with those of pure HAp (JCPDS no. 09-432). The average The microstructure of HAp–CNT nanocomposite powders was
crystallite size (τ) was calculated from the XRD plots using the studied using HRTEM, and the results are shown in Fig. 5 before
Scherrer equation [24]: τ ¼ Kλ=β cos θ, where K is a constant (A) and after sintering (B). The green powder (Fig. 5A) shows
(shape-factor) whose value is 0.9, λ is the wavelength of the distinct HAp particulates with an average size of 100–150 nm
X-ray line used and β (in rad) is the full width at half- attached to the CNT molecules. After sintering, agglomerates of
maximum (FWHM). To investigate whether the presence of CNTs were formed, and a multi-layered sheet-like structure was
CNTs in the composite resulted in any internal strain within obtained, in which the HAp particles adhered to CNT layers.
the samples, the William–Hall and Scherrer methods were
applied. The resulting internal strain, ε, was calculated using 3.3. FTIR analysis
the Williamson–Hall equation [24]: β cos θ ¼ Kλ=τ þ4ε sin θ,
where θ is the diffraction angle, β is the FWHM at θ and ε is The FTIR spectra of the various CNT reinforced HAp are
the mean inhomogeneous strain. A graph of β cos θ and sin θ shown in Fig. 2B, and the FTIR spectra of the pristine
was plotted, and the internal strain was calculated from the MWCNTs are shown in the inset. The pristine CNT spectrum
slope of the graph. τ and the calculated internal strain were exhibits a single peak at 2359 cm
1, which is quite similar to
plotted with variation in the CNT content, as shown in Fig. 3. that reported in the literature [26]. Pure HAp exhibits FTIR
τ is the average grain size for the sintered samples. As peaks at 565, 603, 1045, 1089, 1415 (attributed to the PO34

indicated in Fig. 3, the grain size drastically increased from group), 1643 and 3445 cm
1 (corresponding to the OH

46 nm to 100 nm when 0.5% CNTs were added to pure HAp group). The composites exhibit peaks similar to those of pure
and remained constant up to 2% CNT. The CNTs facilitated HAp, except for those of the hydroxyl group (3OH), which is
rapid grain growth up to 2% CNT addition, with concomitant due to the loss of extra moisture during the sintering process.
crystal strain [25]. At 5% CNT addition, the grain size Therefore, the presence of CNTs in the composites does not
decreased to 61 nm, with very high crystal strain. Grain change the basic structure of HAp. The absence of any new
coarsening was most likely predominant when more than 2% peaks confirms that no primary bonds formed and that the
CNTs were added to pure HAp, with an increased porosity and association between HAp and CNT is purely physical.
decreased density.
The results of the XRD analysis were correlated with those 3.4. Mechanical properties
shown in Fig. 4, which shows the FESEM microstructure of pure
HAp and the HAp–CNT composites. The images show fine hair- Fig. 6 shows the microhardness and K1C of the nanocompo-
like CNT molecules (shown with arrows) attached to the HAp sites with increasing CNT content. The hardness of the
grains. The nanocomposites were observed to possess a porous composites varied between 2.5 and 3.5 GPa (within experimental
microstructure, which is highly favourable for cellular growth. scatter). From the plots, K1C of different HAp–CNT composites
For the composite containing 5% CNTs, the CNT chains were was observed to vary between 1.0 and 1.9 MPa m1/2 in contrast
to the value of 0.87 MPa m1/2 observed for pure HAp, and the
maximum value was observed for 1% CNT content. The plot
shows that the incorporation of 1% CNTs into the HAp matrix
provided a two-fold improvement in toughness due to the energy
absorbed by the MWCNT chains with the collapse of concentric
C-layers when subjected to external stress, in agreement with the
results reported by Ruoff and Lorents [5]. This energy absorption
mechanism favours the dissipation of energy, which enhances
the toughness of the HAp composite. The carbon nanotubes
most likely provided obstructions or ligaments to enhance the
fracture energy. Human bone typically produces K1C values
between 2 and 4 MPa m1/2 depending on the age of the person
[27], and in the current study, the 1% CNT composite achieved a
toughness value similar to that of human bone. HV increased
from 3.2 to 3.4 GPa when 0.5% CNTs were added to the pure
HAp matrix. However, further CNT addition up to the content of
2% degraded the HV values (2.5 GPa), and better values
Fig. 3. Variations in the average crystallite size and internal strain with % (3.4 GPa) were observed when 5% CNTs were added. This
CNT. result indicates that the hardness values were slightly affected by
 S. Mukherjee et al. / Ceramics International 40 (2014) 5635–5643 5639

Fig. 4. Photomicrographs (FESEM) of the sintered HAp–CNT composites with various CNT contents [(A) 0%, (B) 0.5%, (C) 1%, (D) 2% and (E) 5%].

Fig. 5. HRTEM images of the HAp–CNT powders (A) before sintering and (B) after sintering.

a small addition of CNTs to pure HAp. However, further can be sustained before the sample fails, and most materials fail
addition of CNTs was detrimental to this property primarily under tensile stress before they fail under compressive stress
due to the CNT's effect on the formation of pores at the sintering [28]. Based on the results shown in Fig. 7, the flexural strength
temperature. demonstrated a manifold improvement with the addition of
Fig. 7 shows the % increase in the flexural strength of the more than 0.5% CNTs due to the CNT's superior tensile
composites compared to that of pure HAp. The nanocomposites properties being imparted to the nanocomposite.
with a 1% and 2% CNT content exhibited a 95–120% increase, Fig. 8 shows an FESEM image of the fractured surface of a
whereas the nanocomposite with a 5% CNT content exhibited pure HAp sample (Fig. 8A) and that of the 1% HAp–CNT
an approximately 175% increase in the flexural strength over composite sample (Fig. 8B). Fig. 8B clearly shows a CNT
that of pure HAp. In a three-point bend test, the concave face of chain acting as a bridge across a crack moving along the
a sample undergoes maximum compressive stress while the boundaries of the HAp grains, confirming the anticipated
other face (convex face) experiences maximum tensile stress. hindrance of intergranular crack propagation due to the
The flexural strength is the maximum tensile stress value that presence of ligaments. In pure HAp, the bridging action was
5640 S. Mukherjee et al. / Ceramics International 40 (2014) 5635–5643

Figs. 9 and 10, the CNTs added to the pure HAp matrix
improved the bioactivity of the nanocomposites. Due to dissolu-
tion and deposition phenomena during the initial stages, apatite
formation was observed to be the lowest after two weeks, which
slightly improved after four weeks (cf. Fig. 9D–F). The deposi-
tion of apatite crystals was predominant and reached its highest
level for the 5% CNT samples after 12 weeks (cf. Fig. 9G–I).
Pores in the nanocomposite microstructure could be observed on
the 0th day (prior to SBF immersion) (cf. Fig. 9A–C). The
composition of the apatite layer grown on the samples' surface
over time was studied by FTIR, and the results are shown in
Fig. 11. As the number of days in the SBF increased, the
substitution of CO2
3 for the OH
and PO3
4 groups of HAp
increased, reaching its maximum extent for the 2% and 5% CNT
samples. The increased dissolution of apatite observed at 24
weeks for these two types of samples was due to the replacement
Fig. 6. Variations in HV and K1C with the addition of CNTs to the HAp matrix.
of PO3
4 groups inside the crystal structure of HAp. Because the
PO3
4 groups provide the structural stability of HAp [29], the loss
of this group after 12 weeks results in faster dissolution. The
conversion of crystalline HAp to more CO2
3 -substituted HAp
over time (in SBF) was enhanced when more CNTs were added
to the composite, and this effect was observed to be the greatest
after the addition of 1% CNTs in the absence of structural OH
.
Additionally, the very high crystal strain of both 2% and 5%
CNT samples (0.148 and 0.256, respectively), indicating a
potentially higher free energy, promoted faster dissolution after
12 weeks.
Cellular tests confirmed that carbon nanotubes offer good
biocompatibility [30,31]. Zanello et al. [32] demonstrated that
bone cells can grow, generate apatite granules as extracellular
materials and multiply on a CNT scaffold, which makes it an
ideal candidate for bone engineering. The observations made
in the current study confirm these results.

3.6. Haemocompatibility study

Fig. 7. Percent increase in the flexural strength (after three-point bend test) Haemolysis is a measure of cytotoxicity of a biomaterial
with the addition of CNTs to the HAp–CNT nanocomposites.
towards red blood cells (RBCs). Cytotoxicity results in the
rupture of RBC membranes and the release of haemoglobin.
Spectrometry of the free haemoglobin allows the estimation of
not observed in the corresponding fractograph (Fig. 8A). This the extent of cytotoxicity. The results obtained from haemo-
result is in agreement with the higher K1C and flexural strength compatibility tests are shown in Table 2. The % haemolysis
values observed after the addition of 1% CNTs. was determined to vary between 0.3 and 5.3. As per the ASTM
standard, if the % haemolysis of any material is less than 10%,
3.5. Bioactivity study it is considered to be haemocompatible, and if it is less than
5%, it is highly haemocompatible. Therefore, the HAp–CNT
Apatite growth was determined to be the lowest after two composites are haemocompatible. In addition, HAp and the
weeks for each nanocomposite composition and the highest for HAp–CNT composites did not physically damage the blood
samples containing up to 1% CNTs after 24 weeks, as shown in cells, in contrast to the positive control. The haemocompatib-
Figs. 9 and 10. However, the 2% and 5% CNT composites lity of CNTs is often the subject of scrutiny due to potential
exhibited rapid apatite dissolution after 12 weeks and up to 24 metallic impurities, such as Ni. The results obtained in this
weeks. In these samples, apatite formation was the highest at 12 study clearly demonstrate that the HAp–CNT composites
weeks, which was also observed in the SEM microstructure exhibit very good haemocompatibility.
[Fig. 9A, D and G for pure HAp, Fig. 9B, E and H for 1% CNT No major phase changes in HAp were observed upon the
nanocomposite samples and Fig. 9C, F and I for 2% CNT- addition of up to 5% CNTs. However, when more than 1% CNTs
nanocomposite samples after being maintained in SBF for 0, 4 were added, an increase in the internal crystal strain and increased


and 12 weeks, respectively]. Based on the results presented in substitution of CO2
3

3 for the OH and PO4 groups of pure HAp
 S. Mukherjee et al. / Ceramics International 40 (2014) 5635–5643 5641

Fig. 8. FESEM images of the fractured surface of pure HAp and the 1% CNT–HAp nanocomposite.

Fig. 9. Apatite formation at 0th, 4th and 12th week for pure HAp (A, D and G), 1% HAp–CNT nanocomposite (B, E and H) and 2% HAp–CNT nanocomposite
(C, F and I).

were observed. The average crystallite size increased from and HRTEM clearly indicated the attachment of MWCNT chains
 46 nm to  100 nm with the addition of 0.5% CNTs. This to the HAp grains, which directly affected K1C and flexural
crystallite size was maintained up to a CNT content of 2%, and a strength of the samples. The 1% CNT samples exhibited maximum
sudden decrease in the crystallite size was observed at a CNT K1C values, and the 5% CNT samples exhibited maximum values
content of 5% ( 61 nm). The microstructure observed by FESEM for HV and three-point bending flexural strength. The in vitro
5642 S. Mukherjee et al. / Ceramics International 40 (2014) 5635–5643

bioactivity indicated increased apatite formation on the sample 4. Conclusion

surface up to 1% CNTs after 24 weeks, whereas the samples
containing 2% and 5% CNTs exhibited a manifold increase in We observed that pure HAp can be reinforced with pure
apatite formation up to 12 weeks. After 12 weeks, increased CNTs using a simple shear mixing technique. We studied the
dissolution was observed up to 24 weeks, which was most likely physical, mechanical and biological properties of pure HAp


due to the increased substitution of CO2

3 for the OH and PO4
3

and samples containing various CNT contents (i.e., 0.5–5%).
groups. This result was confirmed by FTIR studies. All of the The composites containing 5% CNTs primarily contained
samples were determined to be highly haemocompatible. uniformly distributed nanosized particles and exhibited good
density and approximately 19% porosity. The CNT molecules
form agglomerates along the grain boundaries of HAp. These
agglomerates can form bridges across advancing cracks and
thus retard crack propagation. The 1% CNT–HAp composite,
which showed 2% porosity and 2.93 g/c.c. B.D., exhibited a
substantial improvement in toughness (120%) without signifi-
cantly compromising its hardness (2.8 GPa) and flexural
strength (26 MPa). The sample also exhibited lower internal
crystal strain (0.14) and an average crystallite size of 102 nm.
This nanocomposite also exhibited substantially enhanced

Table 2
Haemolysis (%) of HAp–CNT composite samples.

Sample % Haemolysis

HAp 0.343
HAp–CNT (0.5%) 0.958
HAp–CNT (1%) 1.150
HAp–CNT (2%) 1.753
HAp–CNT (5%) 2.853
Positive control 100
Fig. 10. % Weight gain of nanocomposite samples as a function of the time Negative control 0
immersed in SBF for the bioactivity study.

Fig. 11. FTIR spectrum of the surface of the nanocomposite samples as a function of time immersed in SBF for the bioactivity study.
 S. Mukherjee et al. / Ceramics International 40 (2014) 5635–5643 5643

apatite formation when exposed to SBF for up to 24 weeks, [12] X. Zhang, G.H.M. Gubbels, R.A. Terpstra, R. Metselaar, Toughening of
with only 1.15% haemolysis. However, there was a compro- calcium hydroxyapatite with silver particles, J. Mater. Sci. 32 (1) (1997)
235–243.
mise between the mechanical properties observed and the
[13] M. Zakeri, E. Hasani, M. Tamizifar, Mechanical properties of TiO2–
in vitro bioactivity studied up to 24 weeks, and care must hydroxyapatite nanostructured coatings on Ti–6Al–4V substrates by APS
be taken before selecting any final applications of the method, Int. J. Miner.: Metall. Mater. 20 (4) (2013) 397–402.
nanocomposites. [14] M.A. Lopes, F.J. Monteiro, J.D. Santos, Glass-reinforced hydroxyapatite
composites: fracture toughness and hardness dependence on microstruc-
Acknowledgements tural characteristics, Biomaterials 20 (21) (1999) 2085–2090.
[15] A. Chanda, S. Dasgupta, S. Bose, A. Bandyopadhyay, Microwave
sintering of calcium phosphate ceramics, Mater. Sci. Eng. C 29 (4) (2009)
The authors wish to thank the personnel at the Bioceramics 1144–1149.
and Coating Division and Non-Oxide Ceramics Division, [16] Y. Zhang, S. Tan, Y. Yin, C-fibre reinforced hydroxyapatite bioceramics,
CSIR-CGCRI, Department of Chemical Engineering, Jadavpur Ceram. Int. 29 (1) (2003) 113–116.
University for their generous help and kind support in [17] H. Najafi, Z.A. Nemati, Z. Sadeghian, Inclusion of carbon nanotubes in a
hydroxyapatite sol–gel matrix, Ceram. Int. 35 (7) (2009) 2987–2991.
performing the necessary R&D experiments. We also wish to [18] C. Kaya, Electrophoretic deposition of carbon nanotube-reinforced
thank the Council of Scientific and Industrial Research (CSIR) hydroxyapatite bioactive layers on Ti–6Al–4V alloys for biomedical
and the University Grants Commission (UGC) for their applications, Ceram. Int. 34 (8) (2008) 1843–1847.
financial support. [19] Y. Chen, Y.Q. Zhang, T.H. Zhang, C.H. Gan, C.Y. Zheng, G. Yu,
Carbon nanotube reinforced hydroxyapatite composite coatings produced
through laser surface alloying, Carbon 44 (1) (2006) 37–45.
References [20] B. Kundu, M.K. Sinha, M.K. Mitra, D. Basu, Fabrication and character-
ization of porous hydroxyapatite ocular implant followed by an in vivo
[1] S. Iijima, Helical microtubules of graphitic carbon, Nature 354 (6348) study in dogs, Bull. Mater. Sci. 27 (2) (2004) 133–140.
(1991) 56–58. [21] K. Niihara, R. Morena, D.P.H. Hasselman, Evaluation of K1c of brittle
[2] M.-F. Yu, O. Lourie, M.J. Dyer, K. Moloni, T.F. Kelly, R.S. Ruoff, solids by the indentation method with low crack-to-indent ratios, J. Mater.
Strength and breaking mechanism of multiwalled carbon nanotubes under Sci. Lett. 1 (1) (1982) 13–16.
tensile load, Science 287 (5453) (2000) 637–640. [22] W.C. Oliver, G.M. Pharr, An improved technique for determining
[3] J.-P. Salvetat, A.J. Kulik, J.-M. Bonard, G.A.D. Briggs, T. Stöckli, hardness and elastic modulus using load and displacement sensing
K. Metenier, S. Bonnamy, F. Beguin, N.A. Burnham, L.s. Forró, Elastic indentation experiments, J. Mater. Res. 7 (6) (1992) 1564–1583.
modulus of ordered and disordered multiwalled carbon nanotubes, Adv. [23] T. Kokubo, H. Kushitani, S. Sakka, T. Kitsugi, T. Yamamuro, Solutions
Mater. 11 (2) (1999) 161–165. able to reproduce in vivo surface-structure changes in bioactive glass-
[4] E.W. Wong, P.E. Sheehan, C.M. Lieber, Nanobeam mechanics: elasticity, ceramic A-W3, J. Biomed. Mater. Res. 24 (6) (1990) 721–734.
strength, and toughness of nanorods and nanotubes, Science 277 (5334) [24] B.D. Cullity, Element of X-ray Diffraction, 2nd ed., Addison Wesley
(1997) 1971–1975. Pub. Co., Reading, MA, 1978.
[5] R.S. Ruoff, D.C. Lorents, Mechanical and thermal properties of carbon [25] D. Basu, B. Mukherjee, Sintering of ceramics, in: S. Kumar (Ed.),
nanotubes, Carbon 33 (7) (1995) 925–930. Handbook of Ceramics, vol. 2, Kumar and Associates, Kolkata, 1997, pp.
[6] F. Lupo, R. Kamalakaran, C. Scheu, N. Grobert, M. Ruhle, Micro- 168–181.
structural investigations on zirconium oxide–carbon nanotube composites [26] L. Niu, H. Kua, D.H.C. Chua, Bonelike apatite formation utilizing carbon
synthesized by hydrothermal crystallization, Carbon 42 (10) (2004) nanotubes as template, Langmuir. 26 (6) (2009) 4069–4073.
1995–1999. [27] T.L. Norman, D. Vashishth, D.B. Burr, Fracture toughness of human
[7] T. Seeger, G. de la Fuente, W.K. Maser, A.M. Benito, M.A. Callejas, bone under tension, J. Biomech. 28 (3) (1995) 309–320.
M.T. Martínez, Evolution of multiwalled carbon-nanotube/SiO2 compo- [28] J.M. Hodgkinson, Mechanical Testing of Advanced Fibre Composites,
sites via laser treatment, Nanotechnology 14 (2) (2003) 184–187. Woodhead Publishing Ltd., Cambridge, 2000.
[8] R.Z. Le Geros, J.P. Le Geros, Dense hydroxyapatite, in: L.L. Hench, [29] L.L. Hench, Bioceramics: from concept to clinic, J. Am. Ceram. Soc. 74
J. Wilson (Eds.), An Introduction to Bioceramics, World Scientific, (7) (1991) 1487–1510.
Singapore, 1993, p. 139. [30] J. Chlopek, B. Czajkowska, B. Szaraniec, E. Frackowiak, K. Szostak,
[9] L.G. Yu, K.A. Khor, H. Li, P. Cheang, Effect of spark plasma sintering F. Beguin, In vitro studies of carbon nanotubes biocompatibility, Carbon
on the microstructure and in vitro behavior of plasma sprayed HA 44 (6) (2006) 1106–1111.
coatings, Biomaterials 24 (16) (2003) 2695–2705. [31] P. Galvan-Garcia, E.W. Keefer, F. Yang, M. Zhang, S. Fang, A.
[10] S.K. Nandi, S. Roy, P. Mukherjee, B. Kundu, D.K. De, D. Basu, A. Zakhidov, R.H. Baughman, M.I. Romero, Robust cell migration and
Orthopaedic applications of bone graft and graft substitutes: a review, neuronal growth on pristine carbon nanotube sheets and yarns,,
Indian J. Med. Res. 132 (7) (2010) 15–30. J. Biomater. Sci., Polym. Ed. 18 (10) (2007) 1245–1261.
[11] J.A. Delgado, L. Morejońn, S. Martinez, M.P. Ginebra, N. Carlsson, [32] L.P. Zanello, B. Zhao, H. Hu, R.C. Haddon, Bone cell proliferation on
E. Fernandez, J.A. Planell, M.T. Clavaguera-Mora, J. Rodriguez-Viejo, carbon nanotubes, Nano Lett. 6 (3) (2006) 562–567.
Zirconia-toughened hydroxyapatite ceramic obtained by wet sintering,
J. Mater. Sci.: Mater. Med. 10 (12) (1999) 715–719.