Key Engineering Materials Vols.

284-286 (2005) pp 341-344

Online available since 2005/Apr/15 at www.scientific.net
© (2005) Trans Tech Publications, Switzerland
doi:10.4028/www.scientific.net/KEM.284-286.341
Hydroxyapatite foams for bone replacement

C. M. S. Ranito1,2,3, a, F.A. Costa Oliveira2, b, J. P. Borges1, c

1
Dept. Ciências dos Materiais and CENIMAT, Faculdade de Ciências e Tecnologia, Universidade
Nova de Lisboa, Campus da Caparica, 2829-516 Caparica – Portugal
2
Departamento de Materiais e Tecnologias de Produção, Instituto Nacional de Engenharia,
Tecnologia e Inovação, Estrada do Paço do Lumiar, 1649-038 Lisboa – Portugal
3
Catimedical, Estrada do Paço do Lumiar, 22, Ed. Q, 1649-038 Lisboa – Portugal
a
cranito@sapo.pt, bfernando.oliveira@ineti.pt, cjpb@fct.unl.pt

Keywords: Bioceramics, hydroxyapatite, bone substitute, macroporous structure

Abstract. Hydroxyapatite, often in the form of synthetic porous blocks, c) has been used in the repair
of bone defects for over 20 years owing to its biocompatibility and osseoconductive behaviour.
Bone ingrowth requires the existence of open and interconnected pores with diameters larger than
150 µm for proper circulation of nutrients. Hence, currently available materials are characterised by
poor mechanical properties. Collapse of such products is therefore a major source of concern to
surgeons using these weak materials in bone surgery. There is a need to develop stronger highly
porous structures through adequate control over the size, shape and volume fraction of pores. In this
work, highly porous open-cell hydroxyapatite foams were fabricated by the polymer foam
replication process, where two types of polyurethane (PU) foams were infiltrated with optimised
slurries containing appropriate binders and ceramic particles, followed by the removal of excess
slurry, burning out of the polymer to leave a ceramic replica of the polyurethane and finally high
temperature sintering. Open-cell HAP foams with porosities of about 80% were obtained, i.e. 30%
higher than that determined for commercial ones (50%). Many of the commercial foam cells
approach 500 µm in diameter whereas the developed foam cell size ranged from 300 up to 500 µm.
The ultimate compressive strength of the developed foams (1-2 MPa) was found to be higher than
that recorded for the commercial ones (0.7 MPa) indicating that these foams can more easily be
modelled in theatre. Both the elastic moduli and the compressive strength of the developed foams
were found to increase with increasing of the relative density, in accordance with the predictions of
available micro-mechanical models.

Introduction
Hydroxyapatite (HAP) is a biocompatible and bioactive material with a similar crystal structure to
that of bone mineral. Repair of bone defects with hydroxyapatite ceramics in a porous form is
nowadays a common practice in every day clinical work. Demand for such materials is steadily
growing as a consequence of the increasing patients’ population suffering from bone cancer as well
as traumas resulting from both automotive accidents and sport activities. Synthetic ceramics are
being developed as alternatives to autogeneous bone, xenograft or allograft materials in order to
cope with the market needs. In view of the fact the mechanical properties of porous ceramics are
extremely sensitive to their relative densities and macrostructures, there is a need to establish the
relationship between key structural features (such as pore size, pore size distribution and pore
interconnectivity) and the mechanical performance of these materials. Difficulties arise when one
tries to produce porous HAP with a structure similar to human bone, since the control of the porous
structure is a critical issue. Porous HAP implants can be manufactured by using several methods
[1,2,3]. Among all, replication of polymeric sponges is a cheap, simple and flexible method to
produce porous structures, which can be tailored according to the requirements of the specific

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of TTP,
www.ttp.net. (ID: 129.187.254.47, UB der LMU Muenchen, Muenchen, Germany-27/12/14,16:09:54)
342 Bioceramics 17

implantation site. First patented by Schwartzwalder and Somers [4], this method involves coating
flexible open-cell polymer foams with slurries of ceramic particles. The polymer is removed by
pyrolysis and the ceramic is sintered to yield a replica of the original foam. The structural properties
of the ceramic foams are controlled by the characteristics of the polymer foam, such as density,
porosity and pore size distribution, as well as the amount of slurry impregnated on the original foam
and shrinkage taking place during densification. Such parameters will influence the relative density,
cell size and cell walls distribution of the ceramic foams.The aim of this work was the study of
porous HAP foams for clinical applications, produced by the polymer replication method and to
compare the density, structure, composition and mechanical properties of the developed foams with
those commercially available.

Experimental
The materials used were P120 medical grade hydroxyapatite powder (Plasma Biotal Ltd., U.K.),
ammonium polycarbonate (0.4wt% of the solid load) used as dispersant and a commercial HAP
foam.
Two types of polyurethane (PU) foams with a density of 21 kg⋅m-3, manufactured by Flexipol –
Espumas Sintéticas S.A. (S. João da Madeira, Portugal) were selected as templates, hereafter
referred to as grades 20DB and 22SB4, because of their low degradation temperature and gradual
degradation (~350ºC). The mean cell diameter was determined from SEM micrographs by the Heyn
intersection method. The values obtained for the 20DB and 22SB4 foams were about 700 µm and
900 µm, respectively. Sponge blocks (130x30x15 mm3) were cut, measured and weighed to
determine the geometric density.
The slurry was made up of ceramic particles, water and other additives, preventing the collapse
of the foam structure during polymer removal. Stock-slip containing 45 vol% solid loading was
prepared. Additions of a dispersant (0.4-wt% relative to dry matter) were made. The slurries were
stabilised under rotation for 4 h using a Heidolph RZR – 2000 mixer. Once the ceramic slurry was
prepared, the sponges were impregnated with the slurry. The sponge was compressed to remove air,
immersed in the slurry and then allowed to expand. This procedure was repeated until most air
bubbles were eliminated. The excess ceramic slurry was removed by passing it through preset
rollers. The infiltrated sponge was then dried in an oven at 100ºC overnight. The dry sponge was
slowly heated in static air to 500ºC at 1ºC⋅min-1 to prevent collapse of the ceramic web soaked for 1
h and then heated to 1000ºC for 1 h. The final step was densification of the ceramic network by
sintering in a MoSi2 heated resistance furnace at 1300ºC for 1 h.
The HAP foams were characterized by means of X-ray diffraction, XRD (Rigaku,Tokyo, Japan,
CuKα, 45 kV, 20 mA), Fourier transform infrared spectroscopy, FT-IR (Nicolet Magna 560) and
scanning electron microscopy (SEM/FEG Philips XL30, The Netherlands) to evaluate the
crystalline phases and the molecular groups present and to qualitatively assess pore size and its
morphology. The evaluation of the pore distribution in the samples, was done using the Arquimedes
method. The mechanical properties of the resulting HAP foams were characterized in terms of
Young and Coulomb moduli (resonance impulse excitation method, IMCE n.v., Belgium) and
compression strengths (load cell: 1000 N, Instron Corp., USA, 0.5 mm⋅min-1).

Results and discussion

From X-ray diffraction patterns of samples sintered at 1300ºC it was evident that the final product
consists of approximately 95% HAP and 5% β-TCP (Fig. 1A). In contrast, the commercial foams
consisted of 60% HAP and 40% β-TCP (according to the manufacturer). FT-IR analysis data were
in agreement with the expectations for calcium phosphate materials (Fig. 1B). The main peaks of
hydoxyapatite are observed [1]: a) a band at 569 and 602 cm-1 assigned to the PO4-3 bending
vibration and a band at 1043 cm-1 assigned to the same group for the stretching vibration, b) a band
 Key Engineering Materials Vols. 284-286 343

at 3569 cm-1 associated to the OH stretching vibration, c) a wide band at 1997 cm-1 corresponds to
the water adsorption.
The overall porosity of the developed HAP foams (≈80%) is higher than that of the commercial
ones (≈50%). Fig. 2 shows that the morphology of the sintered HAP foam is similar to that of the
template polymer foams (cell size decreased owing to shrinkage of the ceramic upon sintering).
Some of the cell windows were found to be covered by a thin ceramic membrane. Longitudinal strut
cracks were observed at the relatively sharp edges of the hollow triangular cross-section struts of
the PU substrate. These cracks result from stresses originated from differential drying, thermal
expansion mismatch between the polymer and the ceramic coating, and the gas pressure produced
by pyrolysis of the polymeric skeleton, as suggested by Brown and Green [5]. Indeed, the replica
technique leads to struts containing a hollow cavity possessing sharp corners due to the burnout of
the polyurethane sponge coated by the ceramic slurry. Little or no porosity was visible on SEM
micrographs of the strut surfaces, implying relatively dense struts. The commercial foam showed a
more regular cell size distribution and thicker struts (about twice as thick as that observed in the
developed ones), which were found to be rather porous, suggesting that these foams contain both
macro and micro pores. The cell size of the commercial foams is ≈500 µm whereas the developed
foams cell size ranged from 300 to 500 µm.

500
Espectro da espum a de HAP do tipo 20DB
D if r a c to g r a m a d o H A P s in t e r iz a d o a T = 1 3 0 0 º C
100
H
1997
Transmittance (%)
400
H = HAP 80
Intensidade (cps)

WHW === Hβ-TCP

id r o x ia p a tite
Transmitância (%)

H
300 W h it lo c k ite
Intensity

60 3569

H
40
200 H 602
569
H 20
H H H
100 H H H
HH H
1043
HH H W H H
H W H 0
W W H H H H H HH H
H H H
0
20 40 60 4000 3000 2000 1000
-1
2θ C om prim ento de onda (cm
-1 )
2 θ (deg) Wavenumber (cm )

(A) (B)
Fig. 1 XRD pattern (A) and FTIR spectrum (B) of HAP foams sintered at 1300ºC for 1 h.

a b

c)

c d

Fig. 2 – SEM micrographs of the 20DB PU foam (a) sintered 20DB HAP foam at 1300ºC (b) sintered 22SB4 HAP
foam at 1300ºC (c) and commercial foam (d).
344 Bioceramics 17

Slight differences in mechanical properties were observed for the two types of sintered foams
obtained in spite of their different structures (Table 1). The ultimate compressive strength of the
developed foams was found to be higher than that for the commercial ones, suggesting that the
strut´s microporosity plays a key role in the strength of these materials. Both the elastic moduli and
the compressive strength of the developed foams were found to the closely related to the relative
density of the foams, which is in good agreement with the predictions of the models developed by
Gibson and Ashby [6]. By contrast, the compression strengths obtained for the commercial foams
are much lower than predicted. This is related to the poor resistance of the highly porous struts.

Table 1. Geometric bulk density (ρ), porosity (ε), Young modulus (E), shear modulus (G) and the
crushing strength (σcs) of the sintered hydroxyapatite foams.

Foam grade ρ [g⋅cm-3] εa [%] E [GPa] G [GPa] σcs [MPa]

20DB 0.76 76 3.0 1.2 1.7
22SB4 0.68±0.06 79 2.2 0.9 1.1
Commercial 1.60±0.03 49 - - 0.7
a
ε=1-ρ/ρs where ρs =3.18 g⋅cm-3 is the struts density
Values are presented as the mean±standard deviation for 4 samples

Conclusions
Hydroxyapatite foams were successfully fabricated by a replication method, which involved
coating of polyurethane foams with an adequate ceramic slurry followed by drying and the sintering
of the ceramic skeleton. SEM observations revealed that the developed HAP foams are of the semi-
closed cell type and retained the polymer structure.
On the other hand, the relative density (ρ/ρs) of the two foams was virtually identical (within the
experimental error), suggesting that the density of the PU sponge determines the final density of the
ceramic foams.
From a medical viewpoint, some surgeons claim that these porous HAP implants break easily
during preparation of the implants to be inserted. The challenge is therefore to be able to fabricate
bioactive and resorbable hard tissue replacement implants with mechanical properties comparable
to those of the natural bone. Innovative ways of improvement the mechanical properties of this type
of implants should be sought. The results obtained in the present work suggest that this goal can
probably be achieved, providing proper tailoring of the pore structure is accomplished.

Acknowledgements
The authors wish to thank M.T. Santos, T. Magalhães and P. Coelho at INETI for technical
assistance. Thanks are also due to Flexipol – Espumas Sintéticas S.A. for supplying the PU foams.

References

[1] Dean-Mo Liu, “Fabrication of hydroxyapatite ceramic with controlled porosity” J. Mater. Sci.:
Mater. Med. 8 227-232 (1997);
[2] T. M. G. Chu, J. W. Halloran, S. J. Hollister, S. E. Feinberg, “Hydroxyapatite implants with
designed internal architecture”, J. Mater. Sci.: Mater. Med. 12 471-478 (2001);
[3] E. Rivera-Munoz, J. R. Diaz, J. R. Rodriguez, W. Brostow, V. M. Castano, “Hydroxyapatite
spheres with controlled porosity for eye ball prosthesis: processing and characterization” J.
Mater. Sci.: Mater. Med. 12 305-311 (2001);
[4] K. Schwartzwalder, A.V. Somers: U.S. Patent nº 3 090 094, (1963);
[5] D.D. Brown, D.J. Green: J. Am. Ceram. Soc. 77 1467-1472 (1994);
[6] L.G. Gibson, M.F. Ashby, Cellular Solids: Structure & Properties, Cambridge Univ. Press,
Cambridge, U.K., 1997.
Bioceramics 17
10.4028/www.scientific.net/KEM.284-286

Hydroxyapatite Foams for Bone Replacement

10.4028/www.scientific.net/KEM.284-286.341

DOI References
[2] T. M. G. Chu, J. W. Halloran, S. J. Hollister, S. E. Feinberg, "Hydroxyapatite implants with designed
internal architecture", J. Mater. Sci.: Mater. Med. 12 471-478 (2001);
doi:10.1023/A:1011203226053