JO U R N AL O F M A TE R IA LS S CIE NC E: M ATE RI AL S IN ME DI C IN E 8 (19 9 7 ) 29 – 3 7

Sintering and characterization of HA and TCP

bioceramics with control of their strength
and phase purity
A. T A MP I E RI , G . CE L OT T I , F. S Z ON T AG H* and E . L AN DI
IRTEC-CNR, via Granarolo 64, 48018 Faenza, Italy
* Department of Silicate Chemistry and Technology, University of Veszprem, Hungary

HA and b-TCP-based ceramics were prepared using commercial powders. Powder

characteristics were defined and the processing parameters studied, aimed at the
production of samples with improved microstructural and mechanical properties. The
behaviour of HA powder subjected to various thermal treatments was investigated in order
to control the formation of secondary phases (a- and b-TCP) during sintering. The optimal
thermal treatment required to prepare pure b-TCP powder from the precursors (HA and DCP)
was determined and the sintering method required to prepare fully dense b-TCP completely
free from a-form, was identified. Translucent hot-pressed b-TCP ceramics with potential
applications in aesthetic restorative prostheses were prepared and characterized. The
interval of existence of a6 -TCP and a-TCP as secondary products was also defined.
Crystallographic analysis was carried out on the imperfectly known low-temperature a6 -TCP
phase, and a proper monoclinic unit cell determined.

1. Introduction in the transformation bPa-TCP seems to be the most

Calcium phosphates are major constituents of bone, deleterious phenomenon for the mechanical proper-
and hold great promise as biomaterials for bone im- ties [7].
plantation since they have the ability to bond to As regards b-TCP, it presents the same problem
bone. The major phase found in bone is HA caused by thermal treatment as described for HA,
(Ca (PO ) OH), however, very often defective HA and in addition the difficulty of obtaining pure
5 43
occurs with a Ca/P ratio which deviates from 1.67 compound starting from the precursors HA and
[1, 2]. Among these secondary phases, we focus atten- CaHPO [8].
4
tion on a-TCP (Ca (PO ) ). The present study aims to prepare dense and pure
3 42
Porous HA implants have been widely used in vari- bulk HA and b-TCP materials suitable for clinical
ous fields (orthopaedics, maxillofacial, etc.) and have applications and characterized by improved mech-
demonstrated bone in-growth into the open pores on anical properties, and discusses the formation of
the surface of the implant. Similar studies [3] have a-TCP and related phases at different stages of
shown that dense HA and b-TCP implants become thermal treatment.
surrounded by mature fibrous tissue with a variable
amount of new bone formation. Because of their bio-
activity calcium phosphate implants are expected to 2. Experimental procedures
create good contact with bone and to reduce the Spray-dried HA powder and a precursor powder for
problems of rejection. As a result of its peculiar char- TCP, consisting of a stoichiometric molar mixture
acteristics, HA and b-TCP could be considered prom- 1 : 1 of HA and DCP (CaHPO ), were supplied by
4
ising materials for biomedical applications. Jesse—Shirley. Both powders were characterized by
Up to now calcium phosphate structural implants X-ray diffraction (XRD), using a Rigaku diffrac-
have been characterized by poor mechanical proper- tometer operating with graphite monochromatized
ties: the tendency to develop cracks [4, 5] requires CuKa radiation. From the peak intensities and the
optimization of the material microstructure and con- scattering powers, using a-Al O as a reference stan-
2 3
trol of the compositional evolution to improve the dard, volume fractions of the various phases detected
material strength. were evaluated. Powders morphology was examined
At temperatures ¹*1200 °C, HA becomes unsta- by scanning electron microscopy (SEM) (Leica Cam-
ble and tends to eliminate OH groups and to form bridge).
decomposition products [6]. Among them, a-TCP Mean particle size was determined by Sedigraph
and b-TCP often form during sintering of the material. (Micromeritics): an ultrasound treatment to disperse
In particular the molecular volume increase occurring the particles in ethanol was applied.
0957—4530 ( 1997 Chapman & Hall 29
 Thermoanalytical investigations (TG, DTA) were
performed in the temperature range 20—1550 °C with
a heating rate of 10 °C/min using a Netzsch instru-
ment capable of an accuracy of 2]10~5 g.
The evolution of phases, with particular attention
to a-TCP, was followed directly by a special Rigaku
high temperature attachment (maximum 1500 °C) on
the XRD instrument that allows study of reversible
transformations.
The shrinkage of cold isostatically pressed
(100 MPa) bars was recorded as a function of temper-
ature by Netzsch dilatometer.
Sintering tests were carried out in the temperature
range 1220—1300 °C for HA and 1220—1260 °C for
b-TCP; in both cases the heating programme was
50 °C/h up to 600 °C, then 200 °C/h up to the final
temperature, maintained for 1 h, followed by
subsequent free cooling in the furnace. With TCP, hot-
pressing tests were performed in a graphite die at
1150 °C for 1 h with 30 MPa applied load. Density
was measured by Archimedes method and relative
values were calculated assuming, as a reference,
3.13 g/cm3 for HA and 3.07 g/cm3 for b-TCP.
Flexural strength was determined on polished bars
suitably shaped (30]3]3 mm) by a four-point
bending device (Instron) with outer span 20 mm, inner
span 10 mm and crosshead speed 0.5 mm/min.

3. Results and discussion

3.1. Hydroxyapatite
HA powder was shown to be pure under XRD exam-
ination. Powder morphology is reported in Fig. 1,
showing nearly spherical agglomerates (18—22 lm in
diameter) of globular grains with average diameter
0.5 lm. The average particle size determined by Sedi-
graph was 0.5 lm because the ultrasound treatment
causes disaggregation of the agglomerates into pri-
mary particles, and particle size distribution results
are necessarily very large.
In order to improve powder compaction, 1.5 wt % Figure 1 SEM micrographs showing: (a) HA agglomerate; (b) mor-
of organic binder (PVA) was added. phology of HA powder.
Thermoanalytical investigations performed on
powders with suitable additives (Fig. 2) showed
elimination of the organic phase at 200)¹)420 °C,
through a corresponding weight loss in the TG curve.
A further weight loss of about 1.8% was detected at
temperatures higher than 800 °C, attributed to the
elimination of hydroxyl groups. The weight loss
detected at temperatures higher than 1400 °C is
attributed to the elimination of CO from CO2~
2 3
substituting PO3~ . Infrared analysis of the powder
4
shows the presence of carbonate ions, indicated by
bands at 880, 1420 and 1460 cm~1. The weight loss
(+0.5%) could correspond to the presence of
+0.1 mole of CO2~ in each Ca (PO ) O mole.
3 10 46
Finally the DTA peak at around 1450 °C corresponds
Figure 2 DTA and TG analysis for the HA powder (the total weight
to HA decomposition and subsequent melting of some loss was 4.12%).
decomposition products [6].
Usually the sintering temperature for HA powders boundaries, and the higher one by the advancement in
is chosen in the range 1150—1350 °C: depending on the degradation process.
starting powder features, the lower temperature is Shrinkage has been analysed as a function of tem-
defined by the excess in residual porosity at grain perature and is shown in Fig. 3. In Fig. 3a two steps
30
 Phase composition does not change during cooling,
unless this process is very rapid compared to a classi-
cal furnace (around 500 °C/h), where amorphization of
secondary phases can be expected.
The isostatically pressed bars were sintered at
temperatures in the range 1220—1280 °C: the charac-
teristics of green bodies and sintered materials are
reported in Table I.
Relative density values range from about 92 up to
98%. An increase in sintering temperature beyond
1250 °C causes an increase in secondary phase content
as determined by XRD. At 1250 °C the sintered sam-
ples comprise almost pure HA, and relative density
values are greater than 97%. On the other hand,
increase of the temperature to 1260 °C causes the
formation of about 3% a-TCP; a further increase to
1280 °C results in the appearance of about 3% a-TCP,
4% b-TCP, and sometimes, traces of Ca (PO ) O
4 42
(TeCP) and CaO.
Morphological evolution of the fracture surface of
sintered samples is shown in Fig. 4a, b, c, in which
gradual reduction of intergranular porosity occurs in
going from 1220 °C to 1280 °C, together with slight
grain growth from +2 lm to +8 lm. A peculiar
morphology can be observed starting at 1260 °C: the
typical polyhedric grains begin to coalesce, forming
Figure 3 (a) Dilatometric curve indicating the shrinkage of
debonded HA powder as a function of temperature;
large necks that at 1280 °C yield large tabular grains;
(b) densification versus time for HA material: h 1000 °C; at times an amorphous layer is observable. High
n 1100 °C; s 1200 °C; e 1300 °C. values of flexural strength (r+130 MPa) [9] were
measured in samples sintered at 1250 °C (Table I),
while at lower and higher sintering temperatures re-
can be identified; the first one of about 2% corres- duced r values were found. Two explanations can
ponds to the effect of water elimination resulting in account for these results: (i) composition almost
cell volume contraction. The second one, starting at monophasic compared to the other temperatures,
950 °C, represents the sintering period, from which stimulating the formation of impurities, mainly a- and
the best densification range can be determined b-TCP, which are known to transform into each other
(1100—1300 °C). In Fig. 3b the densification behaviour with deleterious volume variations; (ii) homogeneous
under isothermal conditions at different temperatures microstructure characterized by optimal grain size
is reported, limiting the optimal sintering temperature (4—5 lm): lower temperatures imply incomplete den-
to the interval 1200—1300 °C in which complete den- sification and diffuse porosity, higher temperatures led
sification is attained. to rapid and excessive grain growth, leaving occa-
Owing to the determinant role played by secondary sional sinks of ungrown grains (Fig. 5).
phases if formed during sintering, phase composition Both these aspects depress flexural strength; in the
evolution was followed directly by an XRD high tem- first case because of the lack of a tight link among the
perature attachment: it was found that, starting at grains, in the second case because of the presence of
1200 °C, a small amount of b-TCP (1—3%) begins to inhomogeneities acting as critical defects.
appear, remaining almost constant up to 1250 °C. Fig. 6a shows the fresh fracture surface of a high-
When the temperature exceeds 1250 °C, equivalent flexural-strength HA sample. The study of fracture
quantities of a-TCP (implying the already mentioned origin reveals that usually cracks are generated at
volume increase) are detected. Staying at 1300 °C for specimen surface defects, therefore, as previously ob-
100 min, no time-dependent compositional difference served [9], remarkable improvement can be achieved
was observed: +3% b-TCP and +3% a-TCP. by use of extremely well-polished samples. It is

TA BLE I Characteristics of HA samples

Sample code Relative green Sintering Relative Linear Impurity r (MPa)

density (%) temperature density (%) shrinkage (%) phases (vol %)
(°C)

H1 63.5 1220 93.0 13.0 — 50

H2 63.0 1250 97.5 14.5 — 130
H3 63.5 1260 98.0 13.5 a3 —
H4 63.0 1280 98.7 15.0 a2, b3 98
H5 62.5 1300 98.4 14.8 a3, b3 94

31
 Figure 5 Detail of fracture surface of HA sample sintered at 1260 °C
for 1 h.

Figure 4 Morphology of fracture surfaces of HA sample: Figure 6 Morphology of fracture surface of HA sample sintered at
(a) sintered at 1200 °C for 1 h; (b) sintered at 1260 °C for 1 h; 1250 °C for 1 h: (a) freshly exposed; (b) after 1 month of ageing in
(c) sintered at 1300 °C for 1 h. contact with atmospheric moisture.

32
worthwhile mentioning that such a surface, if left in
contact with atmospheric moisture for long enough,
forms degradation products (Fig. 6b). XRD analysis
of the aged surface identified these needle-like struc-
tures as CaHPO , CaHPO · 2H O and Ca(OH) .
4 4 2 2
The high r-values, comparable only with samples
prepared with powders produced by mechanochemi-
cal synthesis [10], can be attributed to the particular
powder characteristics: good packing ability, thanks
to the favourable grain size and distribution, involving
high density in the green bodies and high surface
reactivity leading to lower sintering temperature,
which limits grain growth and weakens impurity
phases formation.

3.2. Tri-calcium phosphate

The morphology of spray-dried precursor powder
for TCP preparation is shown in Fig. 7, consisting of
spherical agglomerates with average size 10—15 mm
formed by primary irregular particles of size
+0.2 lm.
TG-DTA analysis, reported in Fig. 8, shows
the reactions between the precursors to form b-TCP.
The TG curve reveals a total weight loss of about
4.8%: the first step, starting at about 50 °C and com-
pleted at 500 °C, is attributed to the elimination of
organic binder (+3%) present in the commercial
precursor. At +750 °C a second step, corresponding
to a weight loss of +1.5%, indicates the elimination
of H O as a consequence of the reaction between HA
2
and DCP. The complete reaction should involve
a weight loss of 2.8%, therefore in this case it is
possible to deduce an incomplete reaction leaving HA
and DCP among the reaction products.
It is interesting to note in the DTA curve the
endothermic peak at +800 °C related to the reaction
among precursors, followed by another peak at
+1200 °C corresponding to the transformation
bPa-TCP and two sharp peaks at ¹"1415° and
1450 °C. The first peak can be attributed to the
Figure 7 SEM micrographs showing: (a) TCP powder agglom-
melting (see ICDD card no. 9-345) of pyrophosphate erate; (b) morphology of TCP powder.
(Ca P O , CPP), resulting from the condensation
2 2 7
reaction of the residual CaHPO , and the latter to the
4
decomposition of unreacted HA. From these results
the synthesis temperature of 900 °C has been defined
in order to prepare b-TCP as pure as possible, avoid-
ing the formation of any significant amount of a-TCP.
Precursor powder, reacted at 900 °C for 1 h, re-
sulted in 89% b-TCP and 11% of unreacted HA; no
a-TCP was detected.
In the particular case of TCP, whose formation in
the solid state occurs from a stoichiometric mixture of
HA and DCP, it was found necessary to perform
direct temperature observations using a XRD high
temperature attachment. Attention was focused on
three temperature ranges: ¹)750 °C, 800 °C)¹)
960 °C and ¹*1100 °C, to control the formation of Figure 8 DTA and TG analysis of TCP powder (the total weight
low-temperature a-TCP (usually called a6 -TCP) and loss was 4.80%).
high-temperature a-TCP, as well as the complete reac-
tion between the precursors. increasing amounts of a6 -TCP. Holding at 750 °C for
In the low-temperature range one can still observe about 30 min induces the temporary appearance of
the precursors (Table II) that begin to react, forming b-TCP. The total amount of TCP tends to decrease
33
TA BLE I I Percentage volume fractions of phases observed in TCP by XRD high-temperature attachment

Recording HA DCP a6 -TCP b-TCP a-TCP Total

temperature (°C) TCP

25 77 23 — — — —
650 75 19 6 — — 6
700 73 14 13 — — 13
750 76 9 15 — — 15
750 (30 min) 71 5 13 11 — 24
25 72 9 11 8 — 19
25 77 23 — — — —
800 66 7 — 27 — 27
830 48 — — 52 — 52
860 36 — — 64 — 64
900 30 — — 70 — 70
930 22 — — 72 6 78
960 16 — — 74 10 84
800 20 — — 72 8 80
600 20 — — 75 5 80
400 20 — — 77 3 80
25 24 — — 74 2 76
25 (reacted) 8 — — 92 — 92
1100 8 — — 87 5 92
1200 10 — — 81 9 90
1300 13 — — 15 72 87
1200 15 — — 20 65 85
1100 14 — — 30 56 86
1000 16 — — 31 53 84
900 17 — — 34 49 83
25 17 — — 33 50 83

TA BLE I I I Characteristics of TCP samples

Sample code Relative green Sintering Relative Linear Phase composition r (MPa)
density (%) temperature density (%) shrinkage (%) (vol %)
(°C)

T1 58.8 1220 91.0 13.6 a 62, b 29, HA9 75

T2 59.0 1250 91.6 — a 62, b 28, HA10 —
T3 59.5 1260 92.2 14.0 a 64, b 26, HA10 100
HP — 1150/30 MPa 100 — b 89, HA11 120

during cooling to room temperature, in favour of the becomes dramatic at 1300 °C; an increase of b-form
more stable HA phase. A parallel tendency for a6 -TCP at the expense of a is detected during cooling, accom-
to decrease in isothermal conditions, as well as during panied by limited HA recrystallization. However,
cooling, can be inferred. from all these measurements it is possible to show an
In the intermediate temperature range, formation always-incomplete formation of b-TCP and the un-
of the maximum amount of b-TCP is observed: at avoidable presence of a non-negligible amount of HA,
¹"900 °C the b-TCP volume fraction is 70%, with and to determine the precise temperature interval
no traces of a-TCP; the formation of a-TCP is detec- of the existence of the two different phases: a6 - and
ted starting at 930 °C; the highest concentration of a-TCP. A more complete explanation of the reaction
TCP is found at 960 °C, consisting of 74% b-TCP and path and the kinetics involved in b-TCP formation
10% a-TCP. will be the subject of a forthcoming paper.
During cooling, while b-TCP remains nearly con- The characteristics of the green bodies and sintered
stant, a-TCP tends to disappear, making the total materials are reported in Table III. The isostatically
TCP concentration lower; a slight increase of unreac- pressed bars were sintered at temperatures in the
ted HA can also be detected. It seems reasonable to range 1220—1260 °C. The maximum relative density
suppose that the a-TCP, unstable in this temperature attained was +93% and the morphology of samples
range, transforms to b-TCP which in turn partially sintered at 1260 °C is shown in Fig. 9. Mechanical
converts to HA. properties improve with increasing sintering tem-
Using pellets made from reacted powder the high perature (r+80 MPa at 1220 °C; r+100 MPa at
temperature range was investigated in order to ob- 1260 °C).
serve the sintering phenomena. The gradual To improve final density and to avoid the formation
transformation of b-TCP into a-TCP up to 1200 °C of a-TCP (whose formation, involving volume expan-
34
 Figure 10 Morphology of fracture surface of TCP sample
hot-pressed at 1150 °C, 30 MPa for 1 h.

known high-temperature phase (a-TCP); (ii) form,

always in moderate concentrations, only between
+600° and +800 °C; (iii) then vanish well before the
appearance of a-TCP (¹'900 °C) [13—15].
The accuracy of this determination cannot be very
high, due to the small amounts of interesting phase
usually involved (in the range of +15 vol %) and the
unavoidable temporary presence of two or three other
phases. However, research focused on the small scat-
tering angles region (in particular 2h)10.5°), where
the main phases do not show appreciable diffraction
effects, allowed (with low scanning speed and long
counting times) detection of the many very weak
peaks reported in Table IV. The reproducibility of the
observed lines up to 2h"35° was tested on at least six
different patterns: the region of angles 2h'35° can be
considered, with the involved cells, as the scarcely
significant ‘‘easy matching’’ zone.
The recorded data were generally found sim-
ilar to the stanfieldite (Ca (Mg, Fe) (PO ) ) struc-
4 5 46
ture (ICDD card no. 20-223), existing also as
Figure 9 Morphology of fracture surfaces of TCP samples:
Ca Mg (PO ) (ICDD card no. 11-234). Accounting
(a) sintered at 1220 °C for 2 h; (b) sintered at 1260 °C for 2 h. 3 3 44
for properly increased lattice parameters due to the
actual stoichiometry Ca (PO ) , tentative indexing
3 42
was carried out on the basis of a monoclinic well with
sion, could be inhibited by pressure) the use of pres- a"1.778, b"1.033, c"2.363 nm, b"101°, which
sure sintering was considered. Hot pressed pellets results in a volume very close to a-TCP (suggesting
reached full density, and XRD patterns reveal no Z"24, q"2.90 g/cm3).
a-TCP and +10% of HA left unreacted. This indexing turned out to be completely success-
The morphology of hot pressed b-TCP is reported ful, not only in justifying the new reflections, but in
in Fig. 10, showing an average grain size +1—2 lm reproducing all the observed patterns (ICDD cards
and tight grain packing, completely porosity free. no. 9-348 and 29-359) of a-TCP also. This can be
Flexural strength results were improved (r+ considered as a possible explanation of the distinction
120 MPa) by the above factors. between XRD patterns of a6 -TCP and a-TCP: the
Owing to the incompleteness of data regarding former, even though characterized by a different cell,
low-temperature a-TCP (metastable a-TCP, some- has its main reflections just overlapping those of the
times called a6 or a@-TCP) [11—13], usually found latter. Moreover, well-defined geometrical relation-
during thermal crystallization of amorphous calcium ships exist between the two cells (Fig. 11): the unique
phosphate, a careful crystallographic investigation axes b can be taken as coincident (with lengths almost
was performed on a sample containing a significant in the ratio b(a) : b(a6 )"8 : 3), while a and c axes
amount of this form to attempt individuation of the are rotated so as to yield crystallographic plane
unit cell. The fundamental characteristics of this phase superpositions (2 0 0)a,(3 0 1)a6 , (0 0 4)a,(1 0 7)a6 ,
are: (i) practically undistinguishable from the well- (10 0 0)a6 ,(7 0 21 )a , (0 0 10)a6 ,(11 0 6)a .
35
TA BLE I V Reflections observed in powder pattern of a6 -Ca (PO ) (angular range 5°)2h)35° with CuKa radiation)
3 42
d (nm) hkl Intensity Overlapping a-Ca (PO ) reflections
3 42
d (nm) hkl

1.54 1 0 11 vw
1.29 101 vw
1.16 002 vw N 1.23 N001
1.06 1 0 21 w
1.03 010 vw
.944 011 w
.892 1 0 2, 110 vw
.874 2 0 0, 2 0 11 vw
.858 1 1 11 w
.773 0 0 3, 0 1 2, 2 0 1, 2 0 21 w
.762 1 0 31 vw
.740 1 1 21 w .731 031
.677 112 vw .685 130
.666 1 0 3, 2 1 0, 2 1 11 vw
.642 2 0 2, 2 0 31 w
.620 0 1 3, 211 w N .631 N 2 0 11
.615 1 1 31 , 2 1 21 vw .614 002
.592 3 0 11 vw
.583 0 0 4, 1 0 41 , 3 0 0 w .584 1 3 21
.562 1 1 3, 3 0 21 vw
.546 2 1 2, 2 1 31 vw
.515 0 2 0, 3 0 31 , 3 1 11 w .518 131
.507 0 1 4, 0 2 1, 1 1 41 , 3 1 0 vw
.495 1 2 0, 3 1 21 vw
.487 302 vw .485 2 1 31
.455 3 0 41 vw .455 060
.446 2 0 4, 1 2 2, 2 0 51 , 2 2 0, 2 2 11 , 4 0 11 w
.433 0 2 3, 1 0 5, 3 0 3, 4 0 0, 4 0 21 vw .433 2 3 31
.426 0 1 5, 1 2 31 vw .426 1 6 11
.417 3 1 41 , 401 vw .417 160
.402 2 2 31 , 3 0 51 , 4 1 21 w .401 201
.3970 1 1 5, 313 m
.3915 1 0 61 , 3 2 11 m .3903 1 6 21
.3809 2 0 61 , 3 2 21 vw .3805 2 0 41
.3784 304 vw .3768 2 1 41
.3740 3 2 1, 3 1 51 vw .3733 033
.3695 1 1 61 , 1 2 4, 2 2 3, 2 2 41 m .3689 2 6 11
.3645 1 0 6, 3 2 31 vw .3657 062
.3361 1 3 11 , 5 1 11 w
.3347 3 0 5, 3 1 61 , 4 1 3, 5 1 21 vw .3342 2 6 31
.3328 0 0 7, 2 0 6, 1 3 1, 4 1 51 , 4 2 21 vw .3326 3 3 41
.3304 0 3 2, 1 2 5, 323 vw
.3233 4 0 4, 4 2 31 , 421 w .3239 330
.3214 1 3 2, 1 1 71 vw .3217 4 0 31
.3199 2 3 0, 2 3 11 , 3 1 5, 4 0 61 , 5 0 2, 5 1 1 vw
.3126 1 2 61 , 5 1 41 vw .3122 3 6 21
.3054 1 3 3, 3 2 4, 512 vw
.3037 232 w .3043 3 6 31
.3034 2 3 31 vw .3033 4 3 31
.3008 3 1 71 w .3009 3 6 11
.2945 1 0 81 , 2 0 71 , 3 3 21 , 5 1 5 w .2943 1 9 11
.2920 2 0 81 , 3 2 61 , 4 2 3, 5 2 21 m .2918 2 6 41
.2908 0 0 8, 3 3 1, 4 2 51 , 6 0 31 s .2909 034
.2859 3 3 31 , 5 2 31 m .2859 3 3 51
.2765 5 2 41 , 5 1 61 vw .2757 360
.2687 1 1 8, 1 3 5, 3 2 71 , 3 3 3, 5 1 4 vw
.2666 4 0 6, 5 0 71 vw .2665 4 3 51
.2596 3 1 7, 1 3 61 m .2600 400
.2583 0 0 9, 0 4 0, 3 2 6, 4 1 6, 5 2 3, 5 1 71 m
.2565 1 2 81 , 6 0 61 w .2564 203

Another 14 calculated lines, corresponding perfectly to a-TCP spacings, missing from the list above, are undetectable because they are hidden
by dominant hydroxyapatite reflections. (s"strong, m"medium, w"weak, vw"very weak).

Of course a complete structural determination, just the marked mimicking of diffraction patterns, a6 -TCP
on the basis of a rather poor powder pattern and has a unit cell clearly distinct from a-TCP and conse-
accounting for the large low-symmetry cell, cannot be quently it would be appropriate to call this low-
proposed. However, it can be concluded that, despite temperature phase c-TCP.
36
 Fully dense and translucent b-TCP totally free from
a modification was successfully prepared by hot-press-
ing, attaining a maximum r value of 120 MPa.

Acknowledgements
The authors wish to thank Jesse-Shirley for kindly
supplying the powders used in the present investiga-
tion and for high-level contribution. Partial financial
support by the BRITE-EURAM project BE-7598-93
is also gratefully acknowledged.

Figure 11 Crystallographic relationships in (0 k 0) projection

between low (a6 ) and high-temperature (a) TCP cells. The dotted
traces of (2 0 0)a,(3 0 1)a6 (x) and (0 0 4)a,(1 0 7)a6 (z) planes are also
References
shown. 1. J. R. VA N W AZ ER , in ‘‘Phosphorus and its compounds’’,
Vols I and II (Interscience, New York, 1958 and 1961).
2. J. S . H A N K ER and B. L. G I A M MA R A , Science 242 (1988)
885.
4. Conclusions 3. H. D EN I SS E N , C . M AN G AN O and G . VE NIN I , in
Starting from commercial HA powders, isostatically ‘‘Hydroxylapatite implants’’ (Piccin, Padua, 1985).
formed and sintered at 1250 °C, materials with high 4. A . R O YE R, J. C . V IG U I E, M . HE U GH E BA ER T and J. C .
flexural strength were prepared (r+130 MPa). HE U GH E BA ER T, J. Mater. Sci. Mater. Med. 4 (1993) 76.
5. P. E . W A N G and T. K. C H AK I , ibid. 4 (1993) 150.
The thermal stability range and typical secondary 6. J. Z H O U, X. ZH A N G , J. C H E N, S. ZE N G and K. D E
phase formation in the HA powder were evaluated GR OO T , ibid. 4 (1993) 83.
by TG-DTA and subsequent room temperature XRD 7. H. M O NM A and M . G O TO , ½ogyo-Kyokai-Shi 91 (1983) 473.
analysis; these data were completed by high-temper- 8. C . RE Y , Biomaterials 11 (1990) 13.
ature XRD investigations, studying also the reversible 9. M. J AR CH O , C . H . B O LE N, M . B. T H O M AS , J . B O BI CK ,
J. F. K AY and R . H . D O RE MU S , J. Mater. Sci. 11 (1976)
reactions. 2027.
As regards b-TCP bioceramics, their preparation 10. M. T OR IYA M A, A . RA V AG L IO L I , A. K R AJE W S K I, G .
was performed starting from a commercial powder CEL O TTI and A . PIA N CA S TEL L I, J. Eur. Ceram. Soc.
containing a stoichiometric mixture of HA and DCP 16 (1966) 429.
as precursors. TG-DTA and high-temperature XRD 11. E. D . EAN E S , Calcif. ¹issue Res. 5 (1970) 133.
12. M. M AT H EW , L . W . S CH R O ED ER , B. D I C K EN S and
were used to determine the optimal temperature for W. F. B RO W N , Acta Crystallogr. B 33 (1977) 1325.
b-TCP phase formation; the data obtained allowed 13. T. K A N AZ AW A , T. U M EG AK I and N . U C H IY AM A ,
one to define the stability domains of low-temperature J. Chem. ¹echnol. Biotechnol. 32 (1982) 399.
(a6 -TCP) and high-temperature a-TCP phases. 14. T. UM EG A KI , S. S H IB A and T . K A N AZ AW A ,
Crystallographic analysis was carried out on the ½ogyo-Kyokai-Shi 92 (1984) 612.
15. S. I N O U E, M . K O BA YA S H I and A . O N O , ibid. 96 (1988)
less well-known a6 -TCP modification, with main dif-
182.
fraction data often confused with those of a-TCP,
determining a particular stanfieldite-type monoclinic
cell which, although distinct, reproduces the a-TCP
powder pattern remarkably well: the definition of this Received 8 November 1995
low-temperature phase as c-TCP seems justified. and accepted 29 May 1996

.
37