Biomaterials 23 (2002) 1065–1072

Calcium phosphate apatites with variable Ca/P atomic ratio I.

Synthesis, characterisation and thermal stability of powders
S. Raynaud, E. Champion*, D. Bernache-Assollant, P. Thomas
Science des Proce!de!s Ce!ramiques et de Traitements de Surface, UMR 6638, 123, Avenue Albert Thomas, 87060 Limoges Cedex, France
Received 1 December 2000; accepted 11 June 2001

Abstract

Single phased apatitic calcium phosphate powders Ca10
x(PO4)6
x(HPO4)x(OH)2
x with Ca/P molar ratio ranging from 1.5 to
1.667 (0pxp1) were synthesised using wet method. Outside this compositional range the powders were biphasic mixtures composed
of a calcium phosphate apatite and CaHPO4 (Ca/Po1.5) or Ca(OH)2 (Ca/P>1.667). Temperature and pH of synthesis were the
preponderant parameters for the control of the precipitate composition. The precise determination of the chemical composition
requires the use of several complementary techniques and thermal treatments of powders. These techniques include high resolution
and high temperature X-ray diffractometry and FTIR spectroscopy and show that very small variations of the Ca/P molar ratio of
the powder lead to great changes in powder composition and characteristics after thermal treatment. r 2001 Elsevier Science Ltd.
All rights reserved.

Keywords: Hydroxyapatite; Tricalcium phosphate; Synthesis; Powder; Thermal stability

1. Introduction phosphates on powders characteristics, sintering and

properties of materials is still misunderstood.
Calcium phosphate based ceramics have proved to be On these bases, our work is concerned with a more
attractive materials for biological applications [1]. systematic approach of calcium phosphate compounds
Among these bioceramics, particular attention has been having a Ca/P molar ratio between 1.50 and 1.667. The
given to hydroxyapatite Ca10(PO4)6(OH)2 (HAP)Fdue present paper is the first part of a three-part study. It
to its bioactivity, and to tricalcium phosphate b- deals with the synthesis of apatitic calcium phosphate
Ca3(PO4)2(b-TCP)Fdue to its bioresorbability. powders using a wet method. As detailed in a previous
Whereas stoichiometric HAP and TCP have been widely work, particular attention has been given to the
studied during the last decades, few works devoted to determination and accuracy of Ca/P values [15]. The
non-stoichiometric based compounds are available thermal stability of powders is also investigated to
yet. Only some authors have reported results on precise the structural changes and stability domains of
the synthesis of calcium deficient HAP Ca10
x the different phases.
(PO4)6
x(HPO4)x(OH)2
x(Ca-dHAP), sintering or char-
acteristics of biphasic calcium phosphate (BCP) cera-
mics containing HAP and TCP or HAP and CaO [2–13].
2. Materials and methods
In such compounds, significant differences in the
characteristics may exist between materials of very close
2.1. Powder synthesis
chemical compositions [14]. But, in most of previous
studies few different compositions were investigated. As
The powders were prepared by a conventional
a result, the influence of the stoichiometry of calcium
aqueous precipitation method from the addition of a
diammonium phosphate solution (NH4)2HPO4 (Al-
*Corresponding author. Tel.: +33-5-55-45-74-60; fax: +33-5-55-45- drich, France) into a reactor containing a calcium
75-86. nitrate solution Ca(NO3)2 (Aldrich, France). The
E-mail address: champion@unilim.fr (E. Champion). synthesis device was a fully automated apparatus. The

0142-9612/02/$ - see front matter r 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 2 - 9 6 1 2 ( 0 1 ) 0 0 2 1 8 - 6
1066 S. Raynaud et al. / Biomaterials 23 (2002) 1065–1072

reactor was placed in an argon atmosphere under 2001C. Scanning electron microscopy (SEM) was
dynamic flow in order to prevent any presence of CO2, used for morphological observations (Hitachi S2500,
which could result in carbonate apatite formation. The Japan).
(NH4)2HPO4 addition rate was controlled using a
peristaltic pump. The pH of the solution was maintained
at a constant value by the addition of an ammonium 3. Results
hydroxide solution using a pH stat (Hanna Instru-
ments). The temperature was controlled and regulated. 3.1. Powder synthesis
The suspension was continuously stirred and refluxed.
After total addition of the (NH4)2HPO4 solution, the The synthesis parameters and Ca/P ratio of the
suspension was matured during 30 min. Then, it was precipitated powders are summarised in Table 1. In this
filtered without washing. The resulting precipitate was table, the uncertainty associated to each Ca/P value
dried at 1001C. Synthesis conditions (i.e., pH, tempera- corresponds to a single synthesis. It represents the
ture, Ca/P molar ratio of initial reagents) were adjusted precision of the determination of the chemical composi-
in order to synthesise powders with variable Ca/P tion and should not be confused with the reproducibility
atomic ratio. of syntheses that was evaluated on a total of more than
The purity of the initial powders used to prepare the 50 syntheses and will be discussed hereafter.
reagent solutions was controlled. The content of residual The synthesis conditions greatly influenced the
moisture was also determined. These parameters were composition of powders. Four cases could be differ-
taken into account in the calculation of the concentra- entiated depending on the Ca/P ratio of the precipitates:
tions of reagents solutions.
(i) 1.50oCa/Po1.667, values corresponding to the
2.2. Powder characterisation chemical formulation of Ca-dHAP Ca10
x(PO4)6
x
(HPO4)x(OH)2
x. The Ca/P ratio of the precipitate did
The Ca/P molar ratio of powders, when between 1.50 not depend directly on the Ca/P value of the initial
and 2, was determined after calcination at 10001C for reagents. Most of these powders were synthesised using
15 h from quantitative X-ray diffraction analysis accord- a Ca/P ratio of 1.667 for the initial reagents. But this
ing to a procedure based on a standard method [16]. The ratio had to be decreased down to 1.50 in order to
standard method was completed with a systematic synthesise powders with a Ca/P ratio close to 1.50. With
determination of the precision that was detailed in a these experimental conditions and a constant pH of 7,
previous paper [15]. the Ca/P molar ratio of precipitated powders increased
Powder X-ray diffraction (XRD) patterns were with the temperature increase. Using this method, it was
recorded with Cu Ka radiation on a y=2y diffractometer possible to synthesise compounds having a Ca/P ratio
(Siemens, Model D5000, Germany). The crystalline ranging from 1.511 to 1.663. For these synthesis
phases were determined from a comparison of the conditions, a good reproducibility was achieved since
registered patterns with the ICDD powder diffraction the powders could be processed with a relative deviation
file (PDF). The thermostructural evolution of powders of their Ca/P value inferior to 1%.
was investigated from temperature programmed XRD (ii) Ca/P=1.50, value corresponding to the composi-
in air using a diffractometer (Siemens D5000, Cu Ka tion of apatitic tricalcium phosphate Ca9(PO4)6(OH).
radiation) fitted out with a high temperature furnace Pure apatitic tricalcium phosphate has not been
(Anton Paar HTK10, Pt heating sample holder) and an obtained in our experiment, the closest value being
Elphyse position detector (aperture: 141). The heating 1.511. Though the relative deviation between these two
rate was 51C min
1. Each pattern was recorded after a
setting time of 10 min at the chosen temperature.
Table 1
Powder XRD patterns were also registered during Synthesis parameters and Ca/P molar ratio of precipitated powders
isothermal treatments at 12001C and 13001C under
pH Temperature Ca/P of initial reagents Ca/P of the final powder
nitrogen atmosphere after heating at a rate of 21C s
1.
(1C) (molar ratio) (molar ratio)
Infrared spectra of powders (IR) were recorded on a
Fourier-transform spectrometer (Bomen MB2) with a 6.5 30 1.50 o1.50
resolution of 2 cm
1. Thermogravimetric analysis 7 30 1.50 1.51170.001
7 60 1.50 1.53570.004
(TGA) of powders coupled with differential thermal 7 50 1.667 1.63170.004
analysis (DTA) was performed up to 15001C (Setaram 7 70 1.667 1.64370.003
Labsys TM, France). The specific surface area of 7 80 1.667 1.65570.002
powders was measured by the BET method (8 points, 7 85 1.667 1.66370.001
analyzer Micromeritics ASAP 2010, USA) after calcina- 8 95 1.667 1:667þ0:005

0:001
9 95 1.7–2 1.7170.01
tion at 4001C for 30 min and degassing under vacuum at
 S. Raynaud et al. / Biomaterials 23 (2002) 1065–1072 1067

values was low (about 0.7%), such compositions could surface area of powders is given in Table 2. It tends to
be differentiated between them because of the precision decrease with the increase of the Ca/P ratio. This
of the determination of Ca/P values. This powder with evolution should be associated with the temperature of
Ca/P=1.511 was precipitated at 301C, pH=7 and a synthesis. Indeed, for fixed pH and Ca/P ratio of initial
Ca/P ratio of reagents equal to 1.50. The decrease of half precursors, the increase of Ca/P ratio of powders is
a unity in pH value, from 7 to 6.5, led to a drop of the obtained by increasing the temperature. It can be
Ca/P ratio from 1.511 down to below 1.50. This shows hypothesised that the temperature increase activates
that the synthesis of powders with Ca/P=1.50 would particle growth, on account of the decrease of surface
require a very strict regulation of synthesis parameters area.
and it appeared impossible to obtain a relative deviation XRD patterns of as synthesised powders are given in
of the Ca/P molar ratio below 1%. Fig. 2. Powders exhibited broad diffraction peaks
(iii) Ca/P=1.667, value of stoichiometric HAP indicating a low crystallinity. Powders with Ca/P within
Ca10(PO4)6(OH)2. Such powders could be synthesised the range 1.50–1.667 had the same apatitic structure as
with a high reproducibility. The precision was Ca/ hydroxyapatite (PDF 9-432) and no other crystalline
P=1.667+0.005

0.001 . This is explained because this composi- phase was detected. For Ca/Po1.50, two crystalline
tion can be obtained within a wide range of variation of phases were detected that corresponded to the HAP
synthesis parameters. The pH must be basic (pHX8) structure and to the monetite structure CaHPO4 (PDF
and the temperature above 901C, the only restrictive 9-80). The powder with Ca/P=1.71 had the same
condition is a Ca/P ratio of initial reagents fixed at pattern as hydroxyapatite. No second crystalline phase
1.667. that could contain the excess calcium was detected.
(iv) Ca/P>1.667. These powders are obtained with IR spectra of powders (Fig. 3) were similar whatever
synthesis conditions similar to that used for stoichio- the composition of the powder might be. Most of the
metric HAP when the Ca/P ratio of initial reagents
exceeds 1.667. Table 2
Characteristics of powders

3.2. Characterisation of as synthesised powders Ca/P Specific surface Composition at

(molar ratio) area (m2 g
1) 10001C (wt%)
Fig. 1 shows a typical SEM image of powders. They o 1.50 8872 b-TCP+b-Ca2P2O7
are formed of agglomerated particles of a needle-like (minor phaseE10 wt%)
morphology and a size of about 100 nm long. The 1.51170.001 6071 93% b-TCP+7% HAP
1.53570.004 4770.5 78% b-TCP+22% HAP
1.63170.004 6871.5 80% HAP+20% b-TCP
1.64370.003 6271.5 87% HAP+13% b-TCP
1.65570.002 6271.5 93.5% HAP+6.5% b-TCP
1.66370.001 5971.5 98% HAP+2% b-TCP
1:667 þ0:005

0:001 5871.5 >99.5% HAP
1.7170.01 5571 98.5% HAP+1.5% CaO

Fig. 2. XRD patterns of as synthesised powders. (*) Hydroxyapatite;

Fig. 1. SEM micrograph of as synthesised powders (Ca/P=1.667). (1) monetite CaHPO4.
1068 S. Raynaud et al. / Biomaterials 23 (2002) 1065–1072

bands were characteristic of phosphate groups (at 460, ment of this band to HPO4 is ambiguous because the
550–600, 960, 1020–1120 cm
1). The bands at 630 and band should not be observed on the spectrum of
3540 cm
1 were assigned to OH. The bands at 820 and stoichiometric HAP (i.e., x ¼ 0; Ca/P=1.667) that does
1380 cm
1 were attributed to residual nitrate groups not contain hydrogenophosphates. This band may also
resulting from synthesis precursors. The band at result from residual species (NO
+
3 , NH4 ) in the as
1650 cm
1 could be assigned to carbonate groups or synthesised powder. A more precise characterisation
water. A band at 875 cm
1 could be attributed to requires thermal treatments.
HPO4 which is characteristic of Ca-dHAP
Ca10
x(HPO4)x(PO4)6
x(OH)2
x [17]. But, the assign- 3.3. Thermal stability of powders

Fig. 4 shows typical IR spectra of powders after

heating at different temperatures. At 6001C, the bands
assigned to residual species at 820 and 1380 cm
1 had
disappeared whereas the band at 875 cm
1 partly
assigned to hydrogenophosphates was always present.
From 4001C, a band at 720 cm
1 attributed to pyropho-
sphate groups was observed. This band had disappeared
after heating at 10001C for powders with 1.5oCa/Po
1.667 (Fig. 4a) and was not observed in powders with
Ca/PX1.667. This result agrees with the condensation of
hydrogenophosphate groups in Ca-dHAP that occurs in
the range 350–7201C according to a mechanism
proposed by Mortier et al. [18]
Ca10
z ðHPO4 Þ2z ðPO4 Þ6
2z ðOHÞ2 -Ca10
z ðP2 O7 Þz
s
ðPO4 Þ6
2zþ2s ðOHÞ2ð1
sÞ þ ðz þ sÞH2 O
followed by the decomposition of the apatite into a
mixture of stoichiometric HAP and b-TCP between
7001C and 9001C
Ca10
z ðP2 O7 Þz
s ðPO4 Þ6
2zþ2s ðOHÞ2ð1
sÞ þ ðz þ sÞH2 O
Fig. 3. IR spectra of as synthesised powders. -ð1
zÞCa10 ðPO4 Þ6 ðOHÞ2 þ 3zCa3 ðPO4 Þ2 þ ðz
sÞH2 O:

Fig. 4. IR spectra of powders after thermal treatment. (a) Ca/P=1.631; (b) Ca/Po1.50.
 S. Raynaud et al. / Biomaterials 23 (2002) 1065–1072 1069

For Ca/Po1.50 (Fig. 4b), several bands at 720, 920

and 1200 cm
1, attributed to pyrophosphates were
present after heating at 10001C and the bands at 630
and 3540 cm
1 assigned to OH were not observed. This
agrees with the conversion of CaHPO4 into b-Ca2P2O7
and with the dehydroxylation of apatitic TCP into
b-Ca3(PO4)2 above 7001C. Temperature programmed
XRD patterns of this powder (Fig. 5) confirmed the
transformations. b-Ca2P2O7 was difficult to detect
because its most intense diffraction peak at 2y ¼
29:551 (relative intensity 100%) is superposed with one
of the b-TCP phase. Nevertheless, the peak at 2y ¼
28:91; which is the second most intense peak of b-
Ca2P2O7 (relative intensity 45%), was present from
6001C. This was confirmed on high resolution XRD
patterns (see Fig. 7). b-TCP diffraction peaks appeared
at 7001C; in parallel, the peaks of the initial apatite
disappeared totally at 7501C. Then, b-TCP transformed
into a-TCP from 11501C. This transformation was not
totally reversible since both a and b phases were present
after cooling. Fig. 6. Temperature programmed XRD patterns of powder with Ca/
High temperature XRD patterns of Ca-dHAP pow- P=1.535. (*) HAP; (~) b-TCP; (y) a-TCP.
ders with 1.50oCa/Po1.667 showed similar evolution
between them, an example is given in Fig. 6. At 7001C,
the initial apatite dissociated into a mixture of HAP and
b-TCP according to the global reaction

Ca10
x ðHPO4 Þx ðPO4 Þ6
x ðOHÞ2
x -ð1
xÞCa10 ðPO4 Þ6

ðOHÞ2 þ 3x b-Ca3 ðPO4 Þ2 þ xH2 O: ðR1Þ

Fig. 7. High resolution XRD patterns (step size 0.021, count time 19 s)
of powders. (a) Ca/P=1.71 ; (b) Ca/Po1.50.

The allotropic transformation b-TCP-a-TCP

occurred at 11501C and the reverse transformation
was not totally achieved during cooling.
The HAP powder (Ca/P=1.667) was stable up to
high temperatures. For Ca/P=1.71, the excess calcium
was difficult to characterise. Nevertheless, CaO could be
detected at 7501C on a high resolution XRD pattern
(Fig. 7) indicating that the as synthesised powder should
be constituted of HAP and Ca(OH)2 which is the stable
Fig. 5. Temperature programmed XRD patterns of powder with Ca/ form of CaO at room temperature (the reversible
Po1.50. (*) HAP; (E) b-TCP; (y) a-TCP; (1) CaHPO4; (&)b-Ca2P2O7. dehydroxylation of Ca(OH)2 occurring around 4001C).
1070 S. Raynaud et al. / Biomaterials 23 (2002) 1065–1072

The proportion of phases formed after heating at

10001C for 15 h are given in Table 2. All the powders,
but the stoichiometric HAP, are mixtures of two
crystallised phases.
For Ca/P=1.71, the heated powder was composed of
HAP and the excess calcium was 1.5 wt% of CaO. This
low amount explains that Ca(OH)2 was not detected in
the as synthesised powder because no specific band
could be observed on IR spectra and its main XRD peak
at 2y=34.08 is masked by an intense peak of HAP at
2y ¼ 34:05 (Fig. 7). For the powder with Ca/Po1.50,
the proportion of b-Ca2P2O7 was close to 10 wt%.
For Ca-dHAP, the biphasic mixture contains HAP
and b-TCP. According to reaction (R1), the Ca/P ratio
of the initial powder relates to the proportion of HAP or
TCP as

wt%HAP ¼ 100
wt%b-TCPE600Ca=P
900:

A very low variation of Ca/P value of the synthesised

powder results therefore in a great variation of the
proportions of phases formed after heating.
At high temperature, the behaviour of HAP, or that Fig. 9. Isothermal XRD patterns of HAP powder (Ca/P=1.667) at
13001C under nitrogen. (*) HAP; (y) a-TCP; (&) TCPM Ca4O(PO4)2.
of the HAP phase formed by the decomposition of Ca-
dHAP depended on the thermal cycle and on the
atmosphere.
Isothermal TGA under nitrogen (Fig. 8) showed that
at 12001C the HAP dehydroxylated partially, an
equilibrium been reached after 2 h. The dehydroxylation
occurred without decomposition of the HAP phase. This
was confirmed by high temperature XRD patterns, no
change was detected after 4 h at 12001C. At 13001C, the
TGA showed a more important weight loss and a
change in the kinetic occurred after about 40 min (arrow
in Fig. 8). Isothermal XRD patterns (Fig. 9) confirmed
the decomposition of the HAP at this temperature. The
HAP phase had totally disappeared after 2 h and the
reaction was not reversible since only TCP and

Fig. 10. TGA (dashed line) and DTA in air of hydroxyapatite powder
(Ca/P=1.667).

tetracalcium phosphate monoxide (TCPM) were still

present after cooling.
In air, TGA–DTA (Fig. 10) showed a first weight loss
between 10001C and 14501C associated with a large
endothermic effect, which is characteristic of the
dehydroxylation of the HAP. At 14501C, the weight
loss accelerated strongly and an endothermic peak was
observed indicating the decomposition of the HAP into
TCP and TCPM. But, due to the presence of moisture in
the air, this decomposition was reversible. As a result, a
weight gain and an exothermic peak were registered
from 14501C during cooling. HAP was the major phase
Fig. 8. Isothermal TGA of HAP powder (Ca/P=1.667) under after cooling, only traces of TCP and TCPM were
nitrogen. detected by XRD (Fig. 11).
 S. Raynaud et al. / Biomaterials 23 (2002) 1065–1072 1071

detection of a second minor phase whose response can

be masked by the response of the major phase. This is
the case of Ca(OH)2 or b-Ca2P2O7 when present in low
amounts. The first one can be deduced from the
presence of CaO after thermal treatment of the powder.
It can be detected down to a very low proportion using
XRD (below 0.5 wt%) [15], whereas it is not observable
in the as synthesised powder neither with XRD nor with
IR spectroscopy. The second one (Ca2P2O7) can be
deduced from the detection of CaHPO4 in the as
synthesised powder from XRD or from IR spectroscopy
of the powder heated above 8001C, whereas it is hardly
directly detectable below 5 wt% in the calcined powder
Fig. 11. XRD pattern of HAP powder (Ca/P=1.667) after heating at using XRD. Finally, the use of complementary techni-
15001C in air. ques and thermal treatments appeared necessary
to insure a good precision of the characterisation
of powders. This characterisation constitutes also a
requirement for the understanding of the sintering
4. Discussion and properties of densified ceramics issued from Ca-
dHAP powders that will concern parts II and III of the
The synthesis of single phased calcium phosphate study.
apatites with variable Ca/P atomic ratio within the
range 1.50–1.667 can be performed by the precipitation
method with a good reproducibility. These powders References
(Ca-dHAP) can be processed by temperature and pH
regulation. Nevertheless, the composition of the pre- [1] Jarcho M. Calcium phosphate ceramics as hard tissue prosthetics.
cipitate depends strongly on the processing parameters Clin Orthop Rel Res 1981;157:259–78.
and a precise control of synthesis conditions is required [2] Toriyama M, Kawamura S, Shiba S. Bending strength of
hydroxyapatite ceramics containing a-tricalcium phosphate.
to ensure a good reproducibility. The synthesis of pure Yogyo Kyokai Shi 1987;95:456–8.
apatitic TCP (Ca/P=1.50) appears less reproducible. [3] Asada M, Oukami K, Nakamura S, Takahashi K. Microstructure
Biphasic powders with Ca/Po1.50 are easily obtained. and mechanical properties of non-stoichiometric apatite ceramics
When heated in air, Ca-dHAP decomposes at 7001C and sinterability of raw powder. J Ceram Soc Jpn Int Ed
into a biphasic mixture of HAP and b-TCP. For ceramic 1988;96:583–6.
[4] Osaka A, Miura Y, Takeuchi K, Asada M, Takahashi K. Calcium
part processing, the reproducibility of powder synthesis apatite prepared from calcium hydroxide and orthophosphoric
is all the more important as a relative deviation of 1% of acid. J Mater Sci: Mater Med 1991;2:51–5.
the Ca/P value of the initial powder leads to a variation [5] Royer A, Viguie JC, Heughebaert M, Heughebaert JC. Stoichio-
of about 10 wt% of phase proportions after thermal metry of hydroxyapatite: influence on the flexural strength.
J Mater Sci: Mater Med 1993;4:76–82.
treatment.
[6] Slosarczyk A, Stobierska E, Paszkiewicz Z, Gawlicki M. Calcium
Few works concerned with the mechanism of forma- phosphate materials prepared from precipitates with various
tion of apatitic calcium phosphates by precipitation calcium/phophorus molar ratios. J Am Ceram Soc 1996;79:
methods are available. Heughebaert proposed a ther- 2539–44.
mally activated mechanism that would proceed in [7] Toriyama M, Ravaglioli A, Krajewski A, Celotti G, Piancastelli
several stages [17]. These stages include the formation A. Synthesis of hydroxyapatite-based powders by mechano-
chemical method and their sintering. J Eur Ceram Soc 1996;
of an initial precipitate having a Ca/P ratio close to 1.5 16:429–36.
followed by an increase of the Ca/P ratio accompanied [8] Correia RN, Magalhaes MCF, Marques PAAP, Senos AMR.
with a decrease of the pH of the solution. Our results Wet synthesis and characterization of modified hydroxyapatite
agree with this mechanism, the decrease in pH of the powders. J Mater Sci: Mater Med 1996;7:501–5.
solution being set off by the addition of an NH4OH [9] Slosarczyk A, Paluszkiewicz C, Gawlicki M, Paszkiewicz Z. The
FTIR spectroscopy and QXRD studies of calcium phosphate
solution that maintains a constant pH during the based materials produced from the powder precursors with
synthesis. different Ca/P ratios. Ceramics Int 1997;23:297–304.
XRD and IR spectroscopy, as well as thermal [10] Vallet-Regi M, Rodriguez-Lorenzo LM, Salinas AJ. Synthesis
treatments must be used to characterise unambiguously and characterisation of calcium deficient apatite. Solid State
Ionics 1997;101–103:1279–85.
the chemical composition of as synthesised powders.
[11] Tas AC, Korkusuz F, Timucin M, Akkas N. An investigation of
Indeed, the results given by an individual technique -
the chemical synthesis and high-temperature sintering behaviour
(either XRD or IR spectroscopy) can be biased by the of calcium hydroxyapatite (HA) and tricalcium phosphate (TCP)
presence of residual species or by the absence of bioceramics. J Mater Sci: Mater Med 1997;8:91–6.
1072 S. Raynaud et al. / Biomaterials 23 (2002) 1065–1072

[12] Slosarczyk A, Bialoskorski J. Hardness and fracture toughness of [15] Raynaud S, Champion E, Bernache-Assollant D, Laval JP.
dense calcium-phosphate based ceramics. J Mater Sci: Mater Med Determination of calcium/phosphorus atomic ratio of calcium
1998;9:103–8. phosphate apatites using X-ray diffractometry. J Am Ceramic Soc
[13] Kivrak N, Tas AC. Synthesis of calcium hydroxy- 2001;84:355–66.
apatite–tricalcium phosphate (HA–TCP) composite bioceramic [16] AFNOR. D!etermination quantitative du rapport Ca/P de
powders and their sintering behavior. J Am Ceram Soc 1998;81: phosphates de calcium. Paris: AFNOR, 1998; NF S 94-066.
2245–52. [17] Heughebaert JC. Thesis, Institut national polytechnique, Tou-
[14] Raynaud S, Champion E, Bernache-Assollant D. Characterisa- louse, 1977.
tion of hydroxyapatite-tricalcium phosphate bioceramics issued [18] Mortier A, Lemaitre J, Rouxhet PG. Temperature-programmed
from Ca-defficient hydroxyapatite powders: influence of Ca/P characterization of synthetic calcium-deficient phosphate apatites.
ratio. Phosphorus Res Bul 1999;10:214–8. Thermochim Acta 1989;143:265–82.