Improving the osteointegration and bone-implant interface by incorporation of bioactive particles in sol-gel coatings of stainless steel implants.
ABSTRACT In this study, we report a hybrid organic-inorganic TEOS-MTES (tetraethylorthosilicate-methyltriethoxysilane) sol-gel-made coating as a potential solution to improve the in vivo performance of AISI 316L stainless steel, which is used as permanent bone implant material. These coatings act as barriers for ion migration, promoting the bioactivity of the implant surface. The addition of SiO(2) colloidal particles to the TEOS-MTES sol (10 or 30 mol.%) leads to thicker films and also acts as a film reinforcement. Also, the addition of bioactive glass-ceramic particles is considered responsible for enhancing osseointegration. In vitro assays for bioactivity in simulated body fluid showed the presence of crystalline hydroxyapatite (HA) crystals on the surface of the double coating with 10mol.% SiO(2) samples on stainless steel after 30 days of immersion. The HA crystal lattice parameters are slightly different from stoichiometric HA. In vivo implantation experiments were carried out in a rat model to observe the osteointegration of the coated implants. The coatings promote the development of newly formed bone in the periphery of the implant, in both the remodellation zone and the marrow zone. The quality of the newly formed bone was assessed for mechanical and structural integrity by nanoindentation and small-angle X-ray scattering experiments. The different amount of colloidal silica present in the inner layer of the coating slightly affects the material quality of the newly formed bone but the nanoindentation results reveal that the lower amount of silica in the coating leads to mechanical properties similar to cortical bone.
- [Show abstract] [Hide abstract]
ABSTRACT: The evolution of implant stability in bone tissue remains difficult to assess because remodeling phenomena at the bone-implant interface are still poorly understood. The characterization of the biomechanical properties of newly formed bone tissue in the vicinity of implants at the microscopic scale is of importance in order to better understand the osseointegration process. The objective of this study is to investigate the potentiality of micro-Brillouin scattering techniques to differentiate mature and newly formed bone elastic properties following a multimodality approach using histological analysis. Coin-shaped Ti-6Al-4V implants were placed in vivo at a distance of 200 μm from rabbit tibia leveled cortical bone surface, leading to an initially empty cavity of 200 μm×4.4 mm. After 7 weeks of implantation, the bone samples were removed, fixed, dehydrated, embedded in methyl methacrylate, and sliced into 190 μm thick sections. Ultrasonic velocity measurements were performed using a micro-Brillouin scattering device within regions of interest (ROIs) of 10 μm diameter. The ROIs were located in newly formed bone tissue (within the 200 μm gap) and in mature bone tissue (in the cortical layer of the bone sample). The same section was then stained for histological analysis of the mineral content of the bone sample. The mean values of the ultrasonic velocities were equal to 4.97×10(-3) m/s in newly formed bone tissue and 5.31×10(-3) m/s in mature bone. Analysis of variance (p=2.42×10(-4)) tests revealed significant differences between the two groups of measurements. The standard deviation of the velocities was significantly higher in newly formed bone than in mature bone. Histological observations allow to confirm the accurate locations of the velocity measurements and showed a lower degree of mineralization in newly formed bone than in the mature cortical bone. The higher ultrasonic velocity measured in newly formed bone tissue compared with mature bone might be explained by the higher mineral content in mature bone, which was confirmed by histology. The heterogeneity of biomechanical properties of newly formed bone at the micrometer scale may explain the higher standard deviation of velocity measurements in newly formed bone compared with mature bone. The results demonstrate the feasibility of micro-Brillouin scattering technique to investigate the elastic properties of newly formed bone tissue.Journal of Biomechanical Engineering 02/2011; 133(2):021006. · 1.52 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The aim of this study was to design implant surfaces that attach less to bone but at the same time improve osseous healing for use as temporary bone fracture plates. The strategy was to combine the nonadhesive properties of smooth titanium (Ti) surfaces with the differentiative and anti-inflammatory properties of eicosapentaenoic acid (EPA). Machined Ti implant surfaces coated with a layer of EPA, with or without UV irradiation, were characterized by X-ray photoelectron spectroscopy, and their in vivo performance was evaluated in New Zealand White rabbits. The performance of the functionalised implants was analyzed after 10 weeks of healing by mechanical pull-out testing, molecular biology, and histological and microcomputed tomography analysis. The results indicate that surface functionalization with UV light can reduce bone attachment and volumetric bone mineral density in the peri-implant bone tissue. The presence of EPA on the surfaces enhanced this effect further. Gene expression of bone formation markers showed a trend toward higher mRNA levels in all EPA treated groups. The histological analyses demonstrated lower inflammation in the UV-irradiated group and immature bone formation in all the groups. In conclusion, surface functionalization of Ti implants with UV light and EPA could be a biocompatible coating for reduced bone bonding ability of Ti while promoting bone formation.Journal of Biomedical Materials Research Part A 01/2011; 96(1):83-92. · 2.83 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The evaluation of elastic properties of bone matrix has been investigated using several techniques such as nanoindentation and scanning acoustic microscopy (SAM). These techniques make use of good spatial resolution, which can prevent effects due to microstructures at the level of several hundreds of microns. In this paper, micro-Brillouin scattering (μ-BR) is introduced as another possible technique to characterize the elastic properties of bone. This technique is well known as a non-contact and non-destructive method to evaluate viscoelastic properties of transparent materials in the GHz range. Using thin, translucent bone specimens with thicknesses of around 100μm, and the reflection induced optical geometry, ultrasonic wave velocities in the GHz range were obtained. Because this technique optically measures thermal phonons in the specimen, we can easily measure in-plane anisotropy of wave velocities by rotating the specimen. In a single trabecula, the site matched data between SAM and μ-BR showed good correlation, revealing the applicability of this technique to characterize material properties of bone. Some recent results on the anisotropy in a trabecula and the elasticity evaluation of newly and matured bones are also introduced.Ultrasonics 09/2013; · 2.03 Impact Factor
Improving the osteointegration and bone–implant interface by incorporation
of bioactive particles in sol–gel coatings of stainless steel implants
Josefina Ballarrea,b,*, Inderchand Manjubalab, Wido H. Schreinerc, Juan Carlos Orellanod,
Peter Fratzlb, Silvia Ceréa
aINTEMA, Universidad Nacional del Mar del Plata – CONICET, Juan B. Justo 4302.B7608FDQ, Mar del Plata, Argentina
bMax Planck Institute of Colloids and Interfaces, Department of Biomaterials, 14476 Potsdam, Germany
cLSI-LANSEN, Departamento de Física, UFPR, CP 19081, 81531-990 Curitiba, Brazil
dTraumatología y Ortopedia, Hospital Interzonal General de Agudos ‘‘Oscar Alende”, Mar del Plata, Argentina
a r t i c l ei n f o
Received 11 June 2009
Received in revised form 7 September 2009
Accepted 9 October 2009
Available online 14 October 2009
Surgical grade stainless steel
Newly formed bone
a b s t r a c t
In this study, we report a hybrid organic–inorganic TEOS–MTES (tetraethylorthosilicate–methyltriethox-
ysilane) sol–gel-made coating as a potential solution to improve the in vivo performance of AISI 316L
stainless steel, which is used as permanent bone implant material. These coatings act as barriers for
ion migration, promoting the bioactivity of the implant surface. The addition of SiO2colloidal particles
to the TEOS–MTES sol (10 or 30 mol.%) leads to thicker films and also acts as a film reinforcement. Also,
the addition of bioactive glass–ceramic particles is considered responsible for enhancing osseointegra-
tion. In vitro assays for bioactivity in simulated body fluid showed the presence of crystalline hydroxy-
apatite (HA) crystals on the surface of the double coating with 10 mol.% SiO2samples on stainless steel
after 30 days of immersion. The HA crystal lattice parameters are slightly different from stoichiometric
HA. In vivo implantation experiments were carried out in a rat model to observe the osteointegration
of the coated implants. The coatings promote the development of newly formed bone in the periphery
of the implant, in both the remodellation zone and the marrow zone. The quality of the newly formed
bone was assessed for mechanical and structural integrity by nanoindentation and small-angle X-ray
scattering experiments. The different amount of colloidal silica present in the inner layer of the coating
slightly affects the material quality of the newly formed bone but the nanoindentation results reveal that
the lower amount of silica in the coating leads to mechanical properties similar to cortical bone.
? 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Metallic materials have found wide application as orthopedic
devices as either permanent or temporary devices. When the
orthopedic device has to remain permanently in the body, the
bonding between the implant and the living tissue deserves atten-
tion. Alloys used for implants have in common their excellent
mechanical properties in load-bearing sites. These metallic alloys
have in common a high corrosion resistance in physiological media
mainly due to the formation of a passive oxide film that reduces
the corrosion rate by blocking the transport of metallic ions and
electrons. Surface characterization of these metallic alloys is highly
important as it is a tool to evaluate the performance of the implant
through the surface film–tissue interaction and the possible migra-
tion of metallic ions from the base metal to the nearby tissue [1–7].
One of the ways of minimizing corrosion products which might re-
lease from the implant to the surrounding tissue consists in apply-
ing a protective biocompatible coating. Sol–gel coating has been
proposed as an adequate method to achieve protective biocompat-
ible films [8–12]. It has been demonstrated that inorganic-hybrid
SiO2 coatings, obtained from tetraethylorthosilicate (TEOS) and
methyltriethoxysilane (MTES) in acidic catalysis, improve the cor-
rosion behaviour of the AISI 316L stainless steel in biological envi-
ronments [13,14]. Silica is also known as a natural catalyst for
hydroxyapatite formation (HA: Ca10(PO4)6(OH)2), showing the bio-
activity in vitro . The in vitro deposited HA has been widely
studied with microscopic, diffraction or spectroscopic techniques
. The addition of silica nanoparticles improves the coating
thickness and enhances the attachment of glass–ceramic (GC) bio-
active particles to them.
The extensive use of titanium and its alloys with chemical and
mechanical surface treatments for permanent implants is sup-
ported by some osteoconductive and osteoinductive behaviour of
these surfaces [17,18]. The use of stainless steel 316L for surgical
permanent prostheses with the presence of a natural apatite depo-
1742-7061/$ - see front matter ? 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Address: INTEMA, Corrosion Division, Universidad
Nacional del Mar del Plata – CONICET, Juan B. Justo 4302.B7608FDQ, Mar del Plata,
Argentina. Tel.: +54 223 4816600; fax: +54 223 4810046.
E-mail address: email@example.com (J. Ballarre).
Acta Biomaterialia 6 (2010) 1601–1609
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/actabiomat
sition on the surface seems to be very difficult, due to the need for
high concentrate and dense HAp or A/W coatings [19,20]. The cat-
alytic effect of Si–OH group and Ti–OH groups for the apatite
nucleation has been proven to form apatite in silica and titania gels
immersed in simulated body fluid (SBF) .
The analysis of in vivo formation of new tissue at the interfaces
of bioactive implants has been reported using histological methods
and the interfacial mechanical properties, such as stiffness, have
also been studied [22,23]. Recently other techniques that focus at
the micro- and nano-structural level analysis of the newly formed
tissue, such as nanoindentation, small-angle and wide-angle X-ray
scattering, Raman microscopy and elemental analysis are being
used in the analysis of biological materials [24,25]. These methods
allow one to characterize the quality of the newly formed bone in
terms of mechanical, structural and compositional analysis. Partic-
ularly, the small-angle X-ray scattering technique (SAXS) gives
information about the mineral thickness, shape, orientation and
arrangement within the complex collagen/mineral composite in
This work proposes the protection of 316L stainless steel used
as permanent implant material. Although stainless steel is seldom
used in developed countries as permanent implants, it is still the
most used in emerging countries. Hybrid TEOS–MTES–SiO2coating
by sol–gel technique is applied onto the alloy in order to act as a
barrier to ion diffusion to potentially toxic ions. SiO2colloidal par-
ticles were added to the sol function as reinforcement to the coat-
ing, giving additional protection to the aggressive environments. In
vitro and in vivo studies of the bone–implant interfaces and the
newly formed bone layers were carried out by evaluating HA for-
mation, crystal structural parameters and mechanical strength of
the newly formed tissue at the interface to the implant and the
remodelled bone tissue in its vicinity.
2. Materials and methods
Stainless steel AISI 316L (Atlantic Stainless Co., Inc., MA, USA) in
the form of sheets of 5 ? 10 cm2and wires of 1.3 mm diameter and
2 cm length were used as substrates. The composition of the steel:
C 0.03% max, Mn 2% max, Si 1% max, P 0.045% max, S 0.03% max, Ni
10–14%, Cr 16–18%, Mo 2–3%, balance Fe. They were degreased,
washed with distilled water, and rinsed in ethanol before coating.
2.1.1. Preparation of the coating sols
Hybrid organic–inorganic sols were prepared with a silicon tet-
raethylorthosilane (TEOS, 99%, ABCR GmbH & Co., Germany),
methyltriethoxysilane (MTES, 98%, ABCR GmbH & Co., Germany)
and a water-based solution with colloidal silica (LEVASIL 200A
40 wt.%, Bayer, Germany). The molar ratio of the silanes were
maintained constant (TEOS/MTES = 40/60) and the concentration
of colloidal silica was varied between 10 and 30 mol.% with respect
to the total amount of silica, and nitric acid (0.1 mol l?1) was used
as catalyser. The final silica concentration for both sols was
4.16 mol l?1.
2.1.2. Glass–ceramic (GC) particles and suspension
The glass–ceramic particles were made from a precursor glass
system SiO2–P2O5–CaO. Silica sand (source), calcium carbonate (Al-
drich) and orthophosphoric acid (Aldrich) were used as precursors.
The ratio of components was calculated in order to obtain the
weight concentration of CaO 47.29%, SiO2 35.69% and P2O5
17.01% in the final glass. The mixture was fused in a platinum cru-
cible at 1600 ?C in air atmosphere, and then quenched in water.
The thermal treatment of glass was made at 1050 ?C for 2 h in an
electric furnace with the aim of obtaining apatite and wollastonite
as crystalline phases. The GC obtained was milled in an agate plan-
etary mill (Fritsch Pulverisette, Germany), at a rotation speed of
1500 rpm for 4 h and the powder was sieved with Tyler screens
(grades 270, 325 and 600) to obtain a diameter size distribution
of less than 20 lm. The particle suspensions were prepared by add-
ing 10 wt.% of GC particles to the TEOS–MTES–10% SiO2sol. The
suspensions were stirred by a high shear mixing in a rotor–stator
agitator (Silverson L2R, UK) for 6 min and 15 wt.% of solid surfac-
tant was added .
All the samples were prepared as a double layer system. A first
layer prepared with TEOS–MTES–SiO2sol was obtained at room
temperature by dip-coating at a withdrawal rate of 25 cm min?1
or 18 cm min?1for 10% and 30% of colloidal silica, respectively,
dried at room temperature for 30 min, and heat treated for
30 min at 450 ?C in an electric furnace. A second layer was made
applying TEOS–MTES–10% SiO2with 10 wt.% GC particles on the
top of the first one. The withdrawal and thermal treatment condi-
tions used were the same as in the first layer. The dual coating sys-
tems are named as follows: C10 corresponds to TEOS–MTES–10%
SiO2+ TEOS–MTES–10% SiO2with 10% GC coatings, and C30 corre-
sponds to TEOS–MTES–30% SiO2+ TEOS–MTES–10% SiO2with 10%
GC coatings. A scheme of coating deposition is shown in Fig. 1.
2.2. In vitro bioactivity analysis
Bioactivity of coated stainless steel substrates was analysed
with immersion in SBF solution. SBF was prepared according the
following chemical composition : NaCl (8.053 g l?1), KCl
(0.224 g l?1), CaCl2(0.278 g l?1), MgCl2.6H2O (0.305 g l?1), K2HPO4
(0.174 g l?1), NaHCO3 (0.353 g l?1), (CH2OH)3 CNH2 (6.057 g l?1).
Concentrated hydrochloric acid (HCl) was added to adjust the pH
to 7.25 ± 0.05. The samples were immersed in SBF for 30 days,
and maintained at 37 ?C in a sterilized oven.
2.2.1. X-ray photoelectron scattering (XPS) analysis
XPS assays of the SBF immersed samples were made with an
ESCA 3000 system (Microtech, UK) with a chamber pressure lower
than 10?9mbar. Thespectra
(1253.6 eV) radiation and the overall energy resolution was about
0.8 eV. Four samples were measured after 30 min of Ar+sputtering
performed with an argon ion gun under an accelerating voltage of
3 kV without showing a significant deviation from one to each
other. Survey spectra were recorded for the samples in the 0–
1100 eV kinetic energy range by 1 eV steps. High-resolution scans
with 0.1 eV steps were recorded over the following regions of
interest: Fe 2p (706–712 eV), O 1s (525–534 eV), Ca 2p (345–
349 eV) and P 2p (129.5–136 eV). The surface charging effects were
compensated by referencing the binding energy (BE) to the C 1s
line of residual carbon at 284.5 eV . Data analysis was per-
formed by a least squares fitting program (XPS XI-SDP Spectral
Data Processor v2.3) and by fitting the spectral decomposition
using mixed Gaussian–Lorenzian curves.
2.2.2. Wide-angle X-ray scattering (WAXS) analysis
To characterize the phase composition of the apatite coating
formed in vitro, wide-angle X-ray scattering experiments were
performed. The powder from the surface of the substrates, with
and without immersion in SBF for 30 days, was gently scratched
and collected on a glass substrate. The spectra were done using a
powder diffractometer (D8 X-ray diffractometer, BruckerAXS,
Meadowside, UK) with sealed tube and 1D 120? detector (Nonius).
J. Ballarre et al./Acta Biomaterialia 6 (2010) 1601–1609
Analysing the diffractogram, the average crystallite size of the
HA crystals grown in vitro were calculated from Scherrer’s formula
 using the equation,
L ¼ Kk=bcosh
where L is the average length of the crystal, k the wavelength of
CuKaradiation (0.154178 nm), K is a constant related with the crys-
tallite shape (0.9), h is the diffraction angle (radians) and b is the full
width of the peak at half of maximum intensity (radians). The dif-
fraction peak at h = 26.04? assigned to (0 0 2) plane was chosen
for calculation of the crystallite length since it is isolated from oth-
ers and represents the crystal size of the HA lattice in the c-axis
direction. The broadening of the peaks, b, is corrected for instru-
mental broadening effects as follows:
b2¼ b?2þ b2
where b* represents the HA crystallite contribution to the peak
broadening while the binstris the instrumental broadening contribu-
tion. In this study, other lattice contributions were neglected .
The dimensions of the HA crystal lattice, a-axis and c-axis were
calculated by considering all the peaks in the pattern that can be
indexed using a MS-DOS basic program, assuming a hexagonal
2.3. In vivo experiments and analysis
In vivo experiments were conducted in total in six Hokkaido
adult rats (weight 350 ± 50 g), according to rules of the ethical
committee of the National University of Mar del Plata (Interdisci-
plinary Committee, April 2005), taking care of surgical procedures,
pain control, standards of living and appropriated death. Coated
and uncoated wires were sterilized in autoclave for 20 min at
121 ?C. Rats were anaesthetized with fentanyl citrate and droperi-
dol (Janssen-Cilag Lab, Johnson and Johnson, Madrid, Spain)
according to their weight and the region of surgery surface was
cleaned with antiseptic soap. The animals were placed in a supine
position and the implantation site was exposed through the
superior part of the tibia’s internal face. A region of around
0.5 cm diameter was scraped in the tibia and femur plateau and
a hole was drilled using a hand drill of 0.15 cm diameter bur at
low speed. The implantation site was irrigated with physiological
saline solution during the drilling procedure for cleaning and cool-
ing proposes. The C10, C30 coated implants and uncoated wire im-
plants, as controls, were placed by press fit into tibia and/or femur
extending into the medullar canal. The animals were sacrificed
with an overdose of intraperitoneal fentanyl citrate and droperidol
after 60 days and the bone with implants was retrieved. Conven-
tional X-ray radiographs were taken before retrieving the samples
for control purposes.
2.3.2. Samples sectioning
The retrieved samples were cleaned from surrounding soft tis-
sues and fixed in neutral 10 wt.% formaldehyde for 24 h. Then they
were dehydrated in a series of acetone–water mixtures followed
by a methacrylated solution and finally embedded in methyl meth-
acrylate (PMMA) solution and polymerized. The PMMA embedded
blocks were cut with a low speed diamond blade saw (Buehler
GmbH) cooled with water. Various sections were made according
to different analysis: 300–450 lm thick sections for SAXS mea-
surements, 200 lm thick sections for histological staining, and
5 mm thick blocks for electron microscopy and nanoindentation
studies. The samples used for nanoindentation were further pol-
ished with 120, 240, 400 and 600 grid paper lubricated with water
and then fine polished with 3 lm alumina powder using an auto-
matic grinding and polishing machine (Logitech, UK). Care was
taken to keep the block surface free from scratches as much as
2.3.3. Histological analysis and environmental scanning electron
The surface morphology of the implant–bone interface was ob-
served with environmental scanning electron microscopy (ESEM
Quanta 600 FEG) in low vacuum using a back-scattered electron
(BSE) detector operated at 15 kV. The BSE images reveal the miner-
alized tissue regions and therefore lack the observation of the soft
Fig. 1. Graphic scheme of the coating procedure on AISI 316 stainless steel substrates.
J. Ballarre et al./Acta Biomaterialia 6 (2010) 1601–1609
tissue layer formed around the implant. To observe the soft tissue
and the bone lining cells, the histological sections were stained
with 20% Giemsa stain solution . The stained sample sections
were observed using an optical light microscope (Leica DM RXA2).
2.3.4. Nanoindentation measurements
The indentation modulus and hardness of the newly formed
bone layer at the bone–implant interface were measured using a
scanning nanoindenter (UB1, Hysitron, MN, USA), with a Berkovich
diamond indenter. The high-resolution optical microscope at-
tached to the piezo-scanner allows one to position the indenter
tip precisely at the region of interest. The same indenter tip is also
capable of making AFM images and thereby allows visualizing at
high magnifications the region of interest and the indents made
therein. The indenter tip was calibrated with fused silica. The re-
gions for the measurements were pre-selected from ESEM-BSE
images. In the bone–implant interface region, three lines of indents
were scanned, each line consisting of eight indents at a spacing of
3 lm, with a maximum load of 1000 lN. Few indents were made in
the cortical bone region for comparison. The contact stiffness (S)
was calculated by an automated software that takes into account
the slope only from 20% to 95% of the unloading curve. The hard-
ness (H) and reduced modulus (Er) were then calculated from the
unloading contact stiffness, S, and the indenter contact area Ac
based on the Oliver–Pharr theory [35,36] as
H ¼ Pmax=Ac
In this work, the reduced modulus is reported directly as the bone
indentation modulus, without taking into account the Poisson’s
ratio of bone. The values of indentation modulus and hardness are
reported as mean ± SD. The statistical significance between the
newly formed bone and remodellation zone was performed by
t-test using SigmaStat (SYSTAT Software Inc., Chicago, USA) with a
significance threshold of p < 0.05.
2.3.5. Small-angle X-ray scattering analysis
SAXS samples were mounted on a sample holder perpendicular
to the X-ray beam path and positioned using microcontrollers. Po-
sition-resolved measurements were enabled by a computer-con-
trolled stage moving the sample in two directions (x and y).
Before the measurements, radiography was performed by measur-
ing the X-ray transmission rate of the sample with a diode setup
and scanning the whole sample area with a step size of 0.2 mm.
The SAXS measurement points were chosen from the radiography
image, in the newly formed bone region at the periphery of the im-
plant and a few points in the old cortex for comparison. The scan-
ning SAXS measurements were performed using an instrument
equipped with rotating Cu-anode generator operating at 12 kW,
generating CuKa radiation, wavelength 1.5419 Å, an evacuated
double pinhole and a two-dimensional position-sensitive propor-
tional counter detector (Siemens AG, Karlsruhe, Germany). The size
of the X-ray beam was 200 lm at the sample and the sample-to-
detector distance was about 60 cm. All the spectra were corrected
for empty background scattering from the pinholes. The calibration
of the beam centre position in the detector was done by measuring
silver behanate (AgBh) sample. The reduction of SAXS pattern to
obtain intensity data was done by Fit2D program (2-Dimensional
Data Analyzer, ESRF, France), and further analysis for obtaining
mineral particle thickness and orientation was done using a self-
developed Pyton-based program.
The integration of the 2D SAXS data can be done either radially
or azimuthally to obtain the mineral thickness (T-parameter) and
the degree of alignment of the mineral particles (q-parameter).
The predominant orientation can be calculated from the anisotropy
of the SAXS pattern. The main direction of the long axis of the min-
eral particles and the width of their angular distribution (v) can be
obtained directly from the integration of I(q,v) with respect to q.
This leads to the resulting function I(v) consisting of two peaks
separated by 180?. The q-parameter, which is a measure for the
portion of non-randomly oriented particles, is derived from divid-
ing the area under the two peaks in I(v) by the total area under the
entire function. A value of q = 0 means that there is no predomi-
nant orientation and all particles are randomly oriented within
the plane section, whereas q = 1 means that all particles are
aligned along a predominant orientation. The mean thickness of
the mineral particles is defined as:
T ¼ 4/ð1 ? /Þ=rÞ
where / is the total volume fraction of the mineral and r is the total
surface area of mineral particles per total tissue volume [26,27].
3. Results and discussion
3.1. Barrier effect
The hybrid organic–inorganic coatings applied by sol–gel meth-
od on surgical grade stainless steel were homogeneous and with-
out evident defects or cracks on the surface, and they also
presented good adherence and attachment to the metallic surface
. The average thickness of the coating, measured by profilom-
eter, was 2.1 ± 0.2 lm and 2.3 ± 0.2 lm for the TEOS–MTES–SiO2
10% and TEOS–MTES–SiO2 30% single coating, respectively. The
nanoparticles of SiO2added to reinforce the films filled the defects
and holes present in the silica network, thereby decreasing the
porosity and the conductivity of the surface .
XPS measurements were done in order to see the migration of
metallic ions through the coatings. The ‘‘barrier effect” created by
the coatings can be observed in Fig. 2 where results obtained for
Fe, Cr and Ni elements are shown for both the coated samples after
30 days of immersion in SBF. There is no evidence of toxic elements
like Cr and Ni in the surface of the in vitro samples after 30 days of
immersion. Also it is important to note that Fe ions that migrated
through TEOS–MTES films are somehow inhibited in the silica
nanostructured coatings, since no evidence of Fe or Fe2O3signal
is present in the XPS high-resolution spectra . The iron ions
are smaller and more mobile than Cr and Ni and there is the most
common cause of metalosis (necrosis of the tissues due to metallic
ions) . The inhibition of Fe in these coatings can be attributed
either to the difference of thickness between the coatings (TEOS–
MTES coating 1.6 lm, TEOS–MTES–SiO2
2.1 lm and TEOS–MTES–SiO2 30% single coating 2.3 lm) or to
the accumulation and deposition of silicates at the metal-coating
interface since the silica particles are more reactive at low alkaline
pH than the silanols that formed the layer, covering flaws and
existing defects .
10% single coating
3.2. In vitro apatite analysis
3.2.1. XPS analysis
The in vitro bioactivity of the implant was tested by studying
the formation of bone-like apatite on the surface in SBF  by
the measurement and quantification of the high-resolution spectra
of the characteristic elements of apatite: Ca, P and O, for the C10
and C30 samples after 30 days of immersion in SBF. A glass–cera-
mic (GC) powder sample was also analysed for comparison with
the initial quantity of Ca and P in the surface coating. By comparing
the proportion of the peaks area, the relationships between Ca, O
and P molar ratio were obtained. It was found that the value of
Ca/P ratio for the GC particles was 3.86 whereas for the C10 sample
J. Ballarre et al./Acta Biomaterialia 6 (2010) 1601–1609
immersed in SBF it was 1.43. The standard stoichiometric hydroxy-
apatite (HA) has a Ca/P ratio of 1.67 . In addition, O/P and O/Ca
ratio values for the C10 sample immersed in SBF were 4.13 and
2.88, slightly different from the stoichiometric HA that has values
of 4.33 and 2.6, respectively. These values for GC bioactive particles
are 3.86 and 3, respectively. The values obtained for C10 samples
support the HA dissolution/redeposition theory conducted by Kok-
ubo et al. , Peitl et al.  and Li et al.  which states the
beginning of dissolution of glass–ceramic materials forming par-
tially crystallized apatite (A) and wollastonite (W) phases, obtain-
ing apatite-like deposits after immersion in SBF for a period of
time. The deposition and formation mechanism of apatite-like
layer and its composition depend largely on the substrate and
the media content. The substrates in which the apatite layer is
deposited can also be activated by adding functional groups that
help in apatite nucleation. Therefore, the use of a silica-based coat-
ing supporting the GC particles is a way to improve the coating and
increase the bioactivity of the stainless steel surfaces.
3.2.2. Wide-angle X-ray scattering analysis
Wide-angle X-ray scattering (WAXS) was done on the powders
collected from C10 and C30 coatings on SS316L after 30 days of
immersion in SBF as shown in Fig. 3. The powder sample from
C30 coating without immersion in SBF was measured for compar-
ison as shown in Fig. 3. There was no difference observed in the
powders from C30 before and after immersion, suggesting there
was no formation of apatite layer in this sample. The powder from
C10 sample shows highly crystalline peaks, corresponding to
hydroxyapatite, and the enlargement of the spectrum for C10 after
30 days’ immersion, after subtracting the background scattering, is
Fig. 2. High-resolution XPS scan of Fe2p (a), Cr2p (b) and Ni2p (c) for uncoated SS316L and coated with TEOS–MTES, coated with TEOS–MTES–SiO210% + TEOS–MTES–
SiO210% with GC particles (C10) and coated with TEOS–MTES–SiO230% + TEOS–MTES–SiO210% with GC particles (C30) after immersion in simulated body fluid (SBF) for 1 day
and 30 days.
Fig. 3. (a) WAXS patterns for C10 and C30 samples after 30 days of immersion in
SBF and C30 sample without immersion in SBF for comparison. (b) Enlarged view of
the WAXS pattern of C10 with 30 days of immersion sample in the range 10–70?
(2h) comparing with standard JCPS 9-0432 pattern of HA (vertical lines).
J. Ballarre et al./Acta Biomaterialia 6 (2010) 1601–1609
shown in the right upper graph in Fig. 3. The more intense peaks
are indexed according to JCPDS 9-0432.
The indexed peaks of the C10 sample were used to calculate the
lattice parameters ‘‘a-axis” and ‘‘c-axis” of HA crystal lattice. The
obtained values of lattice parameters are listed in Table 1 in com-
parison with the standard HA (JCPDS 9-0432).
The small decrease in c-axis and the small increase of a-axis
could indicate that the deposited apatite on the C10 sample after
30 days of immersion is slightly deformed due to the hydroxyl sub-
stitution by carbonates and to the Ca2+substitution by other cat-
ions like Na+, to keep the charge balance constant , leading to
a non-stoichiometric HA . The XPS results also confirmed the
formation of non-stoichiometric HA.
The average crystallite size of the HA crystals growth in vitro, in
the direction perpendicular to the crystallographic plane, was esti-
mated from distinctive (0 0 2) and (3 1 0) peaks using Scherrer’s
formula. The average crystallite size of the HA crystals was calcu-
lated as L002= 29.3 nm and L310= 15.7 nm, which describes the size
of crystal predominantly defined by c- and a-axes direction.
3.3. In vivo samples
3.3.1. Histology and scanning electron microscopy analysis
Fig. 4 shows the microscopic images of Giemsa stained sections
of implanted C10 and C30 samples after 60 days of implantation in
rat tibias. The remodellation zone between the implant and the old
cortex bone is clearly visible (marked with arrow). The new lamel-
lae bone tissue growth seems to be perpendicular to the longer axis
of the nail-like implant, and perpendicular to the existing cortical
bone. The formation of an osteoblastic-rich (osteoid layer) surface
near the implant in C10 sample can be seen.
The morphology of the cross-section of bone–implant interface
as observed by ESEM is shown in Fig. 5a and b for the samples C10
and C30, respectively. The cross-section cut-level is approximately
15 mm away from the top of the epiphysis and still it is possible to
observe that the implants were well positioned with 10–20% of
their surface in contact with the old cortical bone and the remain-
ing in contact with the bone marrow. The major difference be-
tween the C10 and C30 implant is that the new bone formed in
the periphery of the implant surface of C30 is denser and appears
to cover fully the implant surface when compared to C10 where
the newly formed bone is not very dense and not continuous on
the surface. Based on these images, the regions for nanoindenta-
tion measurements were selected as shown with white boxes. A
position close to the old bone and a position near the marrow for
each sample were chosen and an old cortex area was also mea-
sured for comparison as seen from Fig. 5c.
3.3.2. Nanoindentation analysis
Nanoindentation measurements were carried out in order to
compare the stiffness and hardness of the newly formed bone at
the bone–implant interface, which was in contact with the cortex
old bone and medullar cavity. Fig. 6a shows a typical load–dis-
placement curve of the indents made in the newly formed bone
and old cortex bone. It illustrates the difference in penetration
depth with the same applied load for the old cortical bone and
for the newly formed less mineralized tissue. Also the morphology
Lattice parameters of hydroxyapatite formed during immersion in SBF for 30 days for
C30 sample and comparing with the values found in JDPCS 9-0432.
‘‘a” parameter (Å)‘‘c” parameter (Å)
HAp (JCPDS 9-0432)
C10 with 30 days of immersion
Fig. 4. Optical microscopic images of Giemsa stained histology sections showing the implant and the newly formed bone. (a and b) C10 sample and (c and d) C30 sample after
60 days of implantation in rat tibias. It can be observed that in sample C10 there is formation of osteoid layer around the implant which is not seen in sample C30.
J. Ballarre et al./Acta Biomaterialia 6 (2010) 1601–1609
of the indent: seen from the AFM image (Fig. 6b), the size of the
indent in the old bone tissue is smaller than the one made in
the remodellation zone for the C30 sample. The mean values of
the indentation modulus and the hardness are listed in Table 2
for the regions of the new bone tissue formed near the marrow
and of nearby old cortical bone for samples C10 and C30, respec-
tively. In the C30 sample, the newly formed bone tissue in contact
with the cortex (remodellation zone) is softer than the newly
formed bone in the marrow region and the indentation modulus
and hardness are much lower to the old cortical bone. This implies
that the new bone in contact with the cortex also undergoes a
remodelling process along with the cortical bone processes. It
can be noted that the newly formed mineralized bone tissue along
the periphery of the sample C10 after 60 days of implantation has
similar elastic-mechanical properties compared to the old cortical
bone. This indicates that the newly formed bone layer around the
C10 implant is equivalent to a mature bone in terms of mechan-
Fig. 5. ESEM images for overview cross-section (a) C10 and (b) C30 samples after 60 days of implantation. The regions for indentation measurements are marked with white
squares to be in the newly formed bone around the implant, one being in contact with old cortical bone and the other being in contact with bone marrow. (c) The cortex
region was chosen for comparison from the same sample.
Fig. 6. (a) Load–displacement curves with 1000 lN maximum load on sample C30 in remodellation zone, showing clear differences in the displacement values and (b) an
AFM image of the respective indents.
J. Ballarre et al./Acta Biomaterialia 6 (2010) 1601–1609
3.3.3. SAXS analysis
The scanning SAXS measurements were done in the newly
formed region around the periphery of the implant and also in cor-
tical bone for comparison. Table 3 gives the values of the mean
mineral thickness (T) in the old cortical bone, in the remodellation
zone (where the implant is in contact with the old bone tissue) and
in the zone of newly formed bone (where the implant is just in
contact with the marrow) for both C10 and C30 implants. The min-
eral particles are smaller in size in the newly formed bone com-
pared to the old cortex. In both locations, near the cortex and
near the marrow, comparing the same position for the samples
C10 and C30, the values were not significantly different (P <
0.001 with t-test).
Fig. 7 shows the radiograph of the C10 and C30 samples after
60 days’ implantation in rat tibias with a colour coding represent-
ing the X-ray transmission values through the sample. The orienta-
tion of the mineral particles in the measured spots of the newly
formed bone is represented in the same figure with small black
bars. The length of the bar denotes the degree of alignment of
the particles (q-parameter) and the direction of the bar gives the
overall preferred orientation of the mineral particles. Since the
mineral particles (HA) are aligned with the longitudinal axis of
the collagen fibrils, q gives also information about the orientation
of this organic matrix. It can be seen that the newly formed bone
around the C10 implant (Fig. 7a) has mineral crystals with prefer-
ential orientation parallel to the direction of the trabecular bone
and follows the circumference of the implant. On the other hand,
the mineral particles in the newly formed bone of the C30 implant
are randomly oriented and are mostly perpendicular to the circum-
ference of the implant (Fig. 7b). It is worth noting that in both sam-
ples, the mineral particles in the old cortex follow the architecture
of the bone and are aligned to the direction of bone growth. These
results are in agreement with the in vitro experiment where it was
observed that the C30 coating did not produce an apatite layer
whereas C10 formed crystalline apatite layer on the surface after
immersion in SBF (Fig. 3a) and the histology analysis showed an
osteoid-like layer for C10 sample. Also from the nanoindentation
results, the newly formed bone in the C10 implant showed stiff-
ness and hardness values similar to the old cortical bone, implying
that the newly formed bone behaves as mature bone. The SAXS
data also support the idea that the C10 implant results in well-or-
ganized bone structure (Fig. 7a), due to increased bioactivity at the
These results indicate that the amount of silica nanoparticles
plays a role in increasing the bioactivity of the coating in the metal-
lic implants. A smaller amount of silica (10%) in the inner layer was
found to be slightly better than 30% in the coatings. As the silica
nanoparticles added to the silica-based hybrid coatings decrease
the porosity, the denser the coatings become, as Montemor et al.
had proved , and the nucleation site for apatite is blocked. As
regards this relation between apatite formation and silica content,
some related results were presented by Li et al.  showing the
catalytic effect of Si–OH groups for the apatite nucleation that pro-
mote the formation of apatite in silica gels immersed in SBF. Also,
Peitl et al.  pointed out that highly bioactive silica-based
glasses and glass–ceramics materials promote apatite formation
in vitro, but the in vivo formation with different amounts of silica
nanoparticles is still under discussion and this work could be a
starting point for that.
The protection of surgical grade stainless steel implants against
corrosion can be achieved by applying hybrid organic–inorganic
sol–gel silica-based coatings. The addition of silica nanoparticles
and glass–ceramic particles to the sol–gel coating enables the
isolation of potentially toxic ions coming from the alloy to the sur-
rounding environment. The in vitro bioactivity and implant–tissue
Indentation modulus and hardness values in the regions of cortical, newly formed and
remodelled bone for C10 and C30 samples.
GPa (mean ± SD)
(mean ± SD)
Cortical bone 33.40 ± 2.791.64 ± 0.21
33.02 ± 4.20
33.26 ± 5.42
1.47 ± 0.19
1.58 ± 0.30
18.55 ± 3.56
24.64 ± 4.16
0.84 ± 0.15
1.01 ± 0.21
T-parameter of the mineral crystals in bone after 60 days of implantation of samples
C10 and C30 in rat tibia.
Region Sample C10 (mean ± SD)Sample C30 (mean ± SD)
2.70 ± 0.04 nm
2.47 ± 0.02 nm
2.29 ± 0.07 nm
2.65 ± 0.18 nm
2.44 ± 0.21 nm
2.41 ± 0.23 nm
Fig. 7. Radiograph of the samples (a) C10 and (b) C30 samples after 60 days of implantation measured with SAXS showing the intensity of the transmitted X-ray through the
sample section. The small white bars denote the degree of orientation of the mineral particles in the measured spots around the implant, the size of bar denotes the degree of
alignment and the direction of the bars denotes the orientation of the mineral particles. It can be noted that the white bars are oriented along the implant surface in newly
formed bone in C10 sample (a) and randomly distributed in C30 sample (b).
J. Ballarre et al./Acta Biomaterialia 6 (2010) 1601–1609
integration is promoted by the addition of such silica nanoparti-
cles. The TEOS–MTES–SiO2with GC system coatings on stainless
steel promotes the formation and growth of non-stoichiometric
hydroxyapatite during in vitro tests when the amount of silica par-
ticles is less. The different amount of colloidal silica slightly affects
the quality of the mineral crystals formed in the new bone in
in vivo conditions. The stiffness of the newly formed bone in the
implant with a lower amount of silica nanoparticles seems to be
similar to the values for old cortex bone, showing short periods
of maturation and mineralization of newly formed bone tissue
around the implant.
The authors acknowledge the SeCyT-CAPES Cooperation pro-
gram and the DuPont-CONICET award for financial support. The
authors acknowledge Dr. Chenghao Li at MPI for his self-developed
Pyton-based SAXS analysis program. J. Ballarre acknowledges the
Deutsche Akademische Austauschdienst (DAAD) for award of ex-
Appendix A. Figures with essential colour discrimination
Certain figures in this article, particularly Figs. 1–4, 6 and 7, are
difficult to interpret in black and white. The full colour images can
befound inthe on-line version,
at doi: 10.1016/j.actbio.
 Doorn PF, Campbell PA, Worrall J, Benya PD, McKellop HA, Amstutz HC. Metal
wear particle characterization from metal on metal total hip replacements:
transmission electron microscopy study of periprosthetic tissues and isolated
particles. J Biomed Mater Res 1998;42:103–11.
 Wintere G. Tissue reactions to metallic wear and corrosion products in human
patients. J Biomed Mater Res Symp 1974;5:11–26.
 Bullen J, Rogers H, Spalding P, Ward C. Iron and infection, the heart of the
matter. FEMS Immunol Med Microbiol 2005;43:325–30.
 Okazaki Y, Gotoh E, Manabe T, Kobayashi K. Comparison of metal
concentrations inrattibia tissues
 Disegi JA, Eschbach L. Stainless steel in bone surgery. Injury Int J Care Injured
 McGregor DB, Baan RA, Partensky C, Rice JM, Wilbourm JD. Evaluation of the
carcinogenic risks to humans associated with surgical implants and other
foreign bodies – a report of an IARC Monographs Programme Meeting.
International Agency for Research on Cancer. Eur J Cancer 2000;36:307–13.
 López DA, Durán A, Ceré S. Electrochemical characterization of AISI 316L
stainless steel in contact with simulated body fluid under infection conditions.
J Mater Sci: Mater Med 2008;19(5):2137–44.
 de Sanctis O, Gómez L, Pellegri N, Parodi C, Marajofsky A, Durán A. Coatings on
metallic substrates. J Non-Cryst Solids 1990;121:338–43.
 de Damborenea JJ, Pellegrini N, de Sanctis O, Durán A. Electrochemical
behaviour of SiO2sol–gel coatings on stainless steel. J Sol–Gel Sci Technol
 Naderi Zand B, Mahdavian M. Evaluation of the effect of vinyltrimethoxysilane
on corrosion resistance and adhesion strength of epoxy coated AA1050.
Electrochim Acta 2007;52:6438–42.
 Castro Y, Ferrari B, Moreno R, Durán A. Corrosion behaviour of silica hybrid
coatings produced from basic catalysed particulate sols by dipping and EPD.
Surf Coat Technol 2005;191:228–35.
 Chou TP, Chandrasekaran C, Limmer SJ, Seraji S, Wu Y, Forbess MJ, et al.
Organic–inorganic hybrid coatings for corrosion protection. J Non-Cryst Solids
 Galliano P, de Damborenea JJ, Pascual MJ, Durán A. Sol–gel coatings on 316L
steel for clinical applications. J Sol–Gel Sci Technol 1998;13:723–7.
 Gallardo J, Galliano PG, Durán A. Bioactive and protective sol–gel coatings on
metals for orthopaedic prosthesis. J Sol–Gel Sci Technol 2001;21:65–74.
 Li R, Clark AE, Hench LL. An investigation of bioactive glass powders by sol–gel
processing. J Appl Biomater 1991;2(4):231–9.
 Fathi MH, Hanifi A, Mortazavi V. Preparation and bioactivity evaluation of
bone-like hydroxyapatite nanopowder. J Mater Proc Technol 2008;202(1–
 Kokubo T, Kim H-M, Kawashita M. Novel bioactive materials with different
mechanical properties. Biomaterials 2003;24:2161–75.
 Gu YW, Khor KA, Cheang P. In vitro studies of plasma-sprayed hydroxyapatite/
Ti–6Al–4V composite coatings in simulated body fluid (SBF). Biomaterials
 Lin F-H, Hsu Y-S, Lin S-H, Sun J-S. The effect of Ca/P concentration and
temperature of simulated body fluid on the growth of hydroxyapatite coating
on alkali-treated 316L stainless steel. Biomaterials 2002;23:4029–38.
 Liu X, Ding C, Chu PK. Mechanism of apatite formation on wollastonite
coatings in simulated body fluid. Biomaterials 2004;25:1755–61.
 Li P, Ohtsuki C, Kokubo T, Nakanishi K, Soga N, Nakamura T, et al. Apatite
formation induced on silica gel in a simulated body fluid. J Non-Cryst Solids
 Gao T, Lindholm T, Kommonen B, Ragni P, Paranzini A, Lindholm TC, et al.
Enhanced healing of segmental tibial defects in sheep by a composite
bone substitute composedof
morphogenetic protein, and Type IV collagen. J Biomed Mater Res 1996;
 Coathup M, Blunn G, Flynn N, Williams C, Thomas N. A comparison of bone
replacements retrieved at post-mortem. J Bone Joint Surg (Br) 2001;82B:
 FratzlP,GuptaHS, Paschalis EP,Roschger P.Structureandmechanical quality of
 Fratzl P, Weinkamer R. Nature’s hierarchical materials. Prog Mater Sci
 Fratzl P, Gupta H, Paris O, Valenta A, Roschger P, Klaushofer K. Diffracting
stacks of cards – some thoughts about small-angle scattering from bone. Prog
Colloid Polym Sci 2005;130:33–9.
 Fratzl P, Schreiber S, Klaushofer K. Bone mineralization as studied by small-
angle X-ray scattering. Connect Tissue Res 1996;34(4):247–54.
 García C, Durán A, Moreno R. Stability of suspensions of bioactive particles
using hybrid organic–inorganic solutions as dispersing media. J Sol–Gel Sci
 Kokubo T, Kushitani H, Sakka S, Kitsugi T, Yamamuro T. Solutions able to
reproduce in vivo surface-structure changes in bioactive glass–ceramic A–W. J
Biomed Mater Res 1990;24:721–34.
 Moulder F, Stickle WF, Sobol PE, Bomben KD, editors. Handbook of X-ray
photoelectron spectroscopy. Eden Praire, MN, USA: Physical Electronics, Inc.;
 Cullity BD, Stock SR. Elements of X-ray diffraction. NJ, USA: Prentice Hall;
 Rusu VM, Ng C-H, Wilke M, Tiersch B, Fratzl P, Peter MG. Size-controlled
hydroxyapatite nanoparticles as self-organized organic–inorganic composite
materials. Biomaterials 2005;26:5414–26.
 Manjubala I, Sampath Kumar TS. Effect of TiO2–Ag2O additives on the
formation of calcium phosphate based functionally graded bioceramics.
 Bradbeer JN, Riminucci M, Bianco P. Giemsa as a fluorescent stain for
mineralized bone. J Histochem Cytochem 1994;42(5):677–80.
 Oliver W, Pharr G. An improved technique for determining hardness and
experiments. J Mater Res 1992;7(6):1564–83.
 Oliver W, Pharr G. Measurement of hardness and elastic modulus by
instrumented indentation: advances in understanding and refinements to
methodology. J Mater Res 2004;19(1):3–20.
 Ballarre J, Jimenez-Piqué E, Anglada M, Pellice SA, Cavalieri AL. Mechanical
characterization of nano-reinforced silica based sol–gel hybrid coatings on AISI
316L stainless steel using nanoindentation techniques. Surf Coat Technol
 Montemor MF, Cabral AM, Zheludkevich ML, Ferreira MGS. The corrosion
resistance of hot dip galvanized steel pretreated with Bis-functional silanes
modified with microsilica. Surf Coat Technol 2006;200(9):2875–85.
 López DA, Rosero-Navarro NC, Ballarre J, Durán A, Aparicio M, Ceré S.
Electrochemical evaluation of multilayer silica–methacrylate hybrid sol–gel
coatings containing bioactive particles on surgical grade stainless steel. Surf
Coat Technol 2008;202(10):2194–201.
 Kokubo T. A/W Glass-ceramic: processing and properties. In: Hench LL, Wilson
J, editors. An introduction to bioceramics. World Scientific; 1993.
 Ballarre J, Pellice S, Scheiner W, Ceré S. Coatings containing silica nanoparticles
and glass ceramic particles applied onto surgical grade stainless steel. Key Eng
 Peitl O, Zanotto Dutra E, Hench LL. Highly bioactive P2O5–Na2O–CaO–SiO2
glass–ceramics. J Non-Cryst Solids 2001;292:115–26.
 Li P, de Groot K, Kokubo T. Bioactive Ca10(PO4)6(OH)2–TiO2composite coating
prepared by sol–gel. J Sol–Gel Sci Technol 1996;7:27–34.
 Wilson R, Elliott J, Dowker S, Smith R. Reitveld structure refinement of
precipitated carbonate apatite using neutron diffraction data. Biomaterials
 El Feki H, Savariault J, Salah A, Jemal M. Sodium and carbonate distribution in
substituted calcium hydroxyapatite. Solid State Sci 2000;2:577–86.
J. Ballarre et al./Acta Biomaterialia 6 (2010) 1601–1609